diff options
Diffstat (limited to 'drivers/gpu/drm/radeon')
146 files changed, 32535 insertions, 11250 deletions
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile index c3df52c1a60..dbcbfe80aac 100644 --- a/drivers/gpu/drm/radeon/Makefile +++ b/drivers/gpu/drm/radeon/Makefile @@ -72,14 +72,38 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \ radeon_cs.o radeon_bios.o radeon_benchmark.o r100.o r300.o r420.o \ rs400.o rs600.o rs690.o rv515.o r520.o r600.o rv770.o radeon_test.o \ r200.o radeon_legacy_tv.o r600_cs.o r600_blit_shaders.o \ - r600_blit_kms.o radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o \ - evergreen.o evergreen_cs.o evergreen_blit_shaders.o evergreen_blit_kms.o \ + radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o dce3_1_afmt.o \ + evergreen.o evergreen_cs.o evergreen_blit_shaders.o \ evergreen_hdmi.o radeon_trace_points.o ni.o cayman_blit_shaders.o \ atombios_encoders.o radeon_semaphore.o radeon_sa.o atombios_i2c.o si.o \ si_blit_shaders.o radeon_prime.o radeon_uvd.o cik.o cik_blit_shaders.o \ r600_dpm.o rs780_dpm.o rv6xx_dpm.o rv770_dpm.o rv730_dpm.o rv740_dpm.o \ rv770_smc.o cypress_dpm.o btc_dpm.o sumo_dpm.o sumo_smc.o trinity_dpm.o \ - trinity_smc.o ni_dpm.o si_smc.o si_dpm.o + trinity_smc.o ni_dpm.o si_smc.o si_dpm.o kv_smc.o kv_dpm.o ci_smc.o \ + ci_dpm.o dce6_afmt.o radeon_vm.o + +# add async DMA block +radeon-y += \ + r600_dma.o \ + rv770_dma.o \ + evergreen_dma.o \ + ni_dma.o \ + si_dma.o \ + cik_sdma.o \ + +# add UVD block +radeon-y += \ + radeon_uvd.o \ + uvd_v1_0.o \ + uvd_v2_2.o \ + uvd_v3_1.o \ + uvd_v4_2.o + +# add VCE block +radeon-y += \ + radeon_vce.o \ + vce_v1_0.o \ + vce_v2_0.o \ radeon-$(CONFIG_COMPAT) += radeon_ioc32.o radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c index fb441a790f3..15da7ef344a 100644 --- a/drivers/gpu/drm/radeon/atom.c +++ b/drivers/gpu/drm/radeon/atom.c @@ -1222,12 +1222,17 @@ int atom_execute_table(struct atom_context *ctx, int index, uint32_t * params) int r; mutex_lock(&ctx->mutex); + /* reset data block */ + ctx->data_block = 0; /* reset reg block */ ctx->reg_block = 0; /* reset fb window */ ctx->fb_base = 0; /* reset io mode */ ctx->io_mode = ATOM_IO_MM; + /* reset divmul */ + ctx->divmul[0] = 0; + ctx->divmul[1] = 0; r = atom_execute_table_locked(ctx, index, params); mutex_unlock(&ctx->mutex); return r; diff --git a/drivers/gpu/drm/radeon/atombios.h b/drivers/gpu/drm/radeon/atombios.h index 16b120c3f14..92be50c39ff 100644 --- a/drivers/gpu/drm/radeon/atombios.h +++ b/drivers/gpu/drm/radeon/atombios.h @@ -1711,7 +1711,9 @@ typedef struct _PIXEL_CLOCK_PARAMETERS_V6 #define PIXEL_CLOCK_V6_MISC_HDMI_BPP_MASK 0x0c #define PIXEL_CLOCK_V6_MISC_HDMI_24BPP 0x00 #define PIXEL_CLOCK_V6_MISC_HDMI_36BPP 0x04 +#define PIXEL_CLOCK_V6_MISC_HDMI_36BPP_V6 0x08 //for V6, the correct defintion for 36bpp should be 2 for 36bpp(2:1) #define PIXEL_CLOCK_V6_MISC_HDMI_30BPP 0x08 +#define PIXEL_CLOCK_V6_MISC_HDMI_30BPP_V6 0x04 //for V6, the correct defintion for 30bpp should be 1 for 36bpp(5:4) #define PIXEL_CLOCK_V6_MISC_HDMI_48BPP 0x0c #define PIXEL_CLOCK_V6_MISC_REF_DIV_SRC 0x10 #define PIXEL_CLOCK_V6_MISC_GEN_DPREFCLK 0x40 @@ -2223,7 +2225,7 @@ typedef struct _SET_VOLTAGE_PARAMETERS_V2 USHORT usVoltageLevel; // real voltage level }SET_VOLTAGE_PARAMETERS_V2; - +// used by both SetVoltageTable v1.3 and v1.4 typedef struct _SET_VOLTAGE_PARAMETERS_V1_3 { UCHAR ucVoltageType; // To tell which voltage to set up, VDDC/MVDDC/MVDDQ/VDDCI @@ -2290,15 +2292,36 @@ typedef struct _GET_LEAKAGE_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_1 #define ATOM_GET_VOLTAGE_VID 0x00 #define ATOM_GET_VOTLAGE_INIT_SEQ 0x03 #define ATOM_GET_VOLTTAGE_PHASE_PHASE_VID 0x04 -// for SI, this state map to 0xff02 voltage state in Power Play table, which is power boost state -#define ATOM_GET_VOLTAGE_STATE0_LEAKAGE_VID 0x10 +#define ATOM_GET_VOLTAGE_SVID2 0x07 //Get SVI2 Regulator Info +// for SI, this state map to 0xff02 voltage state in Power Play table, which is power boost state +#define ATOM_GET_VOLTAGE_STATE0_LEAKAGE_VID 0x10 // for SI, this state map to 0xff01 voltage state in Power Play table, which is performance state #define ATOM_GET_VOLTAGE_STATE1_LEAKAGE_VID 0x11 -// undefined power state + #define ATOM_GET_VOLTAGE_STATE2_LEAKAGE_VID 0x12 #define ATOM_GET_VOLTAGE_STATE3_LEAKAGE_VID 0x13 +// New Added from CI Hawaii for GetVoltageInfoTable, input parameter structure +typedef struct _GET_VOLTAGE_INFO_INPUT_PARAMETER_V1_2 +{ + UCHAR ucVoltageType; // Input: To tell which voltage to set up, VDDC/MVDDC/MVDDQ/VDDCI + UCHAR ucVoltageMode; // Input: Indicate action: Get voltage info + USHORT usVoltageLevel; // Input: real voltage level in unit of mv or Voltage Phase (0, 1, 2, .. ) or Leakage Id + ULONG ulSCLKFreq; // Input: when ucVoltageMode= ATOM_GET_VOLTAGE_EVV_VOLTAGE, DPM state SCLK frequency, Define in PPTable SCLK/Voltage dependence table +}GET_VOLTAGE_INFO_INPUT_PARAMETER_V1_2; + +// New in GetVoltageInfo v1.2 ucVoltageMode +#define ATOM_GET_VOLTAGE_EVV_VOLTAGE 0x09 + +// New Added from CI Hawaii for EVV feature +typedef struct _GET_EVV_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_2 +{ + USHORT usVoltageLevel; // real voltage level in unit of mv + USHORT usVoltageId; // Voltage Id programmed in Voltage Regulator + ULONG ulReseved; +}GET_EVV_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_2; + /****************************************************************************/ // Structures used by TVEncoderControlTable /****************************************************************************/ @@ -3864,6 +3887,8 @@ typedef struct _ATOM_GPIO_PIN_ASSIGNMENT #define PP_AC_DC_SWITCH_GPIO_PINID 60 //from SMU7.x, if ucGPIO_ID=VDDC_REGULATOR_VRHOT_GPIO_PINID in GPIO_LUTable, VRHot feature is enable #define VDDC_VRHOT_GPIO_PINID 61 +//if ucGPIO_ID=VDDC_PCC_GPIO_PINID in GPIO_LUTable, Peak Current Control feature is enabled +#define VDDC_PCC_GPIO_PINID 62 typedef struct _ATOM_GPIO_PIN_LUT { @@ -4169,10 +4194,10 @@ typedef struct _ATOM_COMMON_RECORD_HEADER #define ATOM_OBJECT_LINK_RECORD_TYPE 18 //Once this record is present under one object, it indicats the oobject is linked to another obj described by the record #define ATOM_CONNECTOR_REMOTE_CAP_RECORD_TYPE 19 #define ATOM_ENCODER_CAP_RECORD_TYPE 20 - +#define ATOM_BRACKET_LAYOUT_RECORD_TYPE 21 //Must be updated when new record type is added,equal to that record definition! -#define ATOM_MAX_OBJECT_RECORD_NUMBER ATOM_ENCODER_CAP_RECORD_TYPE +#define ATOM_MAX_OBJECT_RECORD_NUMBER ATOM_BRACKET_LAYOUT_RECORD_TYPE typedef struct _ATOM_I2C_RECORD { @@ -4397,6 +4422,31 @@ typedef struct _ATOM_CONNECTOR_REMOTE_CAP_RECORD USHORT usReserved; }ATOM_CONNECTOR_REMOTE_CAP_RECORD; +typedef struct _ATOM_CONNECTOR_LAYOUT_INFO +{ + USHORT usConnectorObjectId; + UCHAR ucConnectorType; + UCHAR ucPosition; +}ATOM_CONNECTOR_LAYOUT_INFO; + +// define ATOM_CONNECTOR_LAYOUT_INFO.ucConnectorType to describe the display connector size +#define CONNECTOR_TYPE_DVI_D 1 +#define CONNECTOR_TYPE_DVI_I 2 +#define CONNECTOR_TYPE_VGA 3 +#define CONNECTOR_TYPE_HDMI 4 +#define CONNECTOR_TYPE_DISPLAY_PORT 5 +#define CONNECTOR_TYPE_MINI_DISPLAY_PORT 6 + +typedef struct _ATOM_BRACKET_LAYOUT_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucLength; + UCHAR ucWidth; + UCHAR ucConnNum; + UCHAR ucReserved; + ATOM_CONNECTOR_LAYOUT_INFO asConnInfo[1]; +}ATOM_BRACKET_LAYOUT_RECORD; + /****************************************************************************/ // ASIC voltage data table /****************************************************************************/ @@ -4524,8 +4574,9 @@ typedef struct _ATOM_VOLTAGE_OBJECT_HEADER_V3{ #define VOLTAGE_OBJ_VR_I2C_INIT_SEQ 3 //VOLTAGE REGULATOR INIT sequece through I2C -> ATOM_I2C_VOLTAGE_OBJECT_V3 #define VOLTAGE_OBJ_PHASE_LUT 4 //Set Vregulator Phase lookup table ->ATOM_GPIO_VOLTAGE_OBJECT_V3 #define VOLTAGE_OBJ_SVID2 7 //Indicate voltage control by SVID2 ->ATOM_SVID2_VOLTAGE_OBJECT_V3 -#define VOLTAGE_OBJ_PWRBOOST_LEAKAGE_LUT 0x10 //Powerboost Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3 -#define VOLTAGE_OBJ_HIGH_STATE_LEAKAGE_LUT 0x11 //High voltage state Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3 +#define VOLTAGE_OBJ_EVV 8 +#define VOLTAGE_OBJ_PWRBOOST_LEAKAGE_LUT 0x10 //Powerboost Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3 +#define VOLTAGE_OBJ_HIGH_STATE_LEAKAGE_LUT 0x11 //High voltage state Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3 #define VOLTAGE_OBJ_HIGH1_STATE_LEAKAGE_LUT 0x12 //High1 voltage state Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3 typedef struct _VOLTAGE_LUT_ENTRY_V2 @@ -4552,6 +4603,10 @@ typedef struct _ATOM_I2C_VOLTAGE_OBJECT_V3 VOLTAGE_LUT_ENTRY asVolI2cLut[1]; // end with 0xff }ATOM_I2C_VOLTAGE_OBJECT_V3; +// ATOM_I2C_VOLTAGE_OBJECT_V3.ucVoltageControlFlag +#define VOLTAGE_DATA_ONE_BYTE 0 +#define VOLTAGE_DATA_TWO_BYTE 1 + typedef struct _ATOM_GPIO_VOLTAGE_OBJECT_V3 { ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader; // voltage mode = VOLTAGE_OBJ_GPIO_LUT or VOLTAGE_OBJ_PHASE_LUT @@ -4584,7 +4639,8 @@ typedef struct _ATOM_SVID2_VOLTAGE_OBJECT_V3 // 1:0 offset trim, USHORT usLoadLine_PSI; // GPU GPIO pin Id to SVID2 regulator VRHot pin. possible value 0~31. 0 means GPIO0, 31 means GPIO31 - UCHAR ucReserved[2]; + UCHAR ucSVDGpioId; //0~31 indicate GPIO0~31 + UCHAR ucSVCGpioId; //0~31 indicate GPIO0~31 ULONG ulReserved; }ATOM_SVID2_VOLTAGE_OBJECT_V3; @@ -4637,6 +4693,49 @@ typedef struct _ATOM_ASIC_PROFILING_INFO_V2_1 USHORT usElbVDDCI_LevelArrayOffset; // offset of 2 dimension voltage level USHORT array }ATOM_ASIC_PROFILING_INFO_V2_1; +typedef struct _ATOM_ASIC_PROFILING_INFO_V3_1 +{ + ATOM_COMMON_TABLE_HEADER asHeader; + ULONG ulEvvDerateTdp; + ULONG ulEvvDerateTdc; + ULONG ulBoardCoreTemp; + ULONG ulMaxVddc; + ULONG ulMinVddc; + ULONG ulLoadLineSlop; + ULONG ulLeakageTemp; + ULONG ulLeakageVoltage; + ULONG ulCACmEncodeRange; + ULONG ulCACmEncodeAverage; + ULONG ulCACbEncodeRange; + ULONG ulCACbEncodeAverage; + ULONG ulKt_bEncodeRange; + ULONG ulKt_bEncodeAverage; + ULONG ulKv_mEncodeRange; + ULONG ulKv_mEncodeAverage; + ULONG ulKv_bEncodeRange; + ULONG ulKv_bEncodeAverage; + ULONG ulLkgEncodeLn_MaxDivMin; + ULONG ulLkgEncodeMin; + ULONG ulEfuseLogisticAlpha; + USHORT usPowerDpm0; + USHORT usCurrentDpm0; + USHORT usPowerDpm1; + USHORT usCurrentDpm1; + USHORT usPowerDpm2; + USHORT usCurrentDpm2; + USHORT usPowerDpm3; + USHORT usCurrentDpm3; + USHORT usPowerDpm4; + USHORT usCurrentDpm4; + USHORT usPowerDpm5; + USHORT usCurrentDpm5; + USHORT usPowerDpm6; + USHORT usCurrentDpm6; + USHORT usPowerDpm7; + USHORT usCurrentDpm7; +}ATOM_ASIC_PROFILING_INFO_V3_1; + + typedef struct _ATOM_POWER_SOURCE_OBJECT { UCHAR ucPwrSrcId; // Power source @@ -5808,6 +5907,8 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO_V3 #define ATOM_S7_DOS_MODE_PIXEL_DEPTHb0 0x0C #define ATOM_S7_DOS_MODE_PIXEL_FORMATb0 0xF0 #define ATOM_S7_DOS_8BIT_DAC_ENb1 0x01 +#define ATOM_S7_ASIC_INIT_COMPLETEb1 0x02 +#define ATOM_S7_ASIC_INIT_COMPLETE_MASK 0x00000200 #define ATOM_S7_DOS_MODE_NUMBERw1 0x0FFFF #define ATOM_S7_DOS_8BIT_DAC_EN_SHIFT 8 @@ -6242,6 +6343,7 @@ typedef struct _ATOM_MC_INIT_PARAM_TABLE #define _128Mx32 0x53 #define _256Mx8 0x61 #define _256Mx16 0x62 +#define _512Mx8 0x71 #define SAMSUNG 0x1 #define INFINEON 0x2 @@ -6987,9 +7089,10 @@ typedef struct _ATOM_DISP_OUT_INFO_V3 UCHAR ucMaxDispEngineNum; UCHAR ucMaxActiveDispEngineNum; UCHAR ucMaxPPLLNum; - UCHAR ucCoreRefClkSource; // value of CORE_REF_CLK_SOURCE - UCHAR ucReserved[3]; - ASIC_TRANSMITTER_INFO_V2 asTransmitterInfo[1]; // for alligment only + UCHAR ucCoreRefClkSource; // value of CORE_REF_CLK_SOURCE + UCHAR ucDispCaps; + UCHAR ucReserved[2]; + ASIC_TRANSMITTER_INFO_V2 asTransmitterInfo[1]; // for alligment only }ATOM_DISP_OUT_INFO_V3; //ucDispCaps @@ -7661,618 +7764,6 @@ typedef struct _ATOM_POWERPLAY_INFO_V3 ATOM_POWERMODE_INFO_V3 asPowerPlayInfo[ATOM_MAX_NUMBEROF_POWER_BLOCK]; }ATOM_POWERPLAY_INFO_V3; -/* New PPlib */ -/**************************************************************************/ -typedef struct _ATOM_PPLIB_THERMALCONTROLLER - -{ - UCHAR ucType; // one of ATOM_PP_THERMALCONTROLLER_* - UCHAR ucI2cLine; // as interpreted by DAL I2C - UCHAR ucI2cAddress; - UCHAR ucFanParameters; // Fan Control Parameters. - UCHAR ucFanMinRPM; // Fan Minimum RPM (hundreds) -- for display purposes only. - UCHAR ucFanMaxRPM; // Fan Maximum RPM (hundreds) -- for display purposes only. - UCHAR ucReserved; // ---- - UCHAR ucFlags; // to be defined -} ATOM_PPLIB_THERMALCONTROLLER; - -#define ATOM_PP_FANPARAMETERS_TACHOMETER_PULSES_PER_REVOLUTION_MASK 0x0f -#define ATOM_PP_FANPARAMETERS_NOFAN 0x80 // No fan is connected to this controller. - -#define ATOM_PP_THERMALCONTROLLER_NONE 0 -#define ATOM_PP_THERMALCONTROLLER_LM63 1 // Not used by PPLib -#define ATOM_PP_THERMALCONTROLLER_ADM1032 2 // Not used by PPLib -#define ATOM_PP_THERMALCONTROLLER_ADM1030 3 // Not used by PPLib -#define ATOM_PP_THERMALCONTROLLER_MUA6649 4 // Not used by PPLib -#define ATOM_PP_THERMALCONTROLLER_LM64 5 -#define ATOM_PP_THERMALCONTROLLER_F75375 6 // Not used by PPLib -#define ATOM_PP_THERMALCONTROLLER_RV6xx 7 -#define ATOM_PP_THERMALCONTROLLER_RV770 8 -#define ATOM_PP_THERMALCONTROLLER_ADT7473 9 -#define ATOM_PP_THERMALCONTROLLER_EXTERNAL_GPIO 11 -#define ATOM_PP_THERMALCONTROLLER_EVERGREEN 12 -#define ATOM_PP_THERMALCONTROLLER_EMC2103 13 /* 0x0D */ // Only fan control will be implemented, do NOT show this in PPGen. -#define ATOM_PP_THERMALCONTROLLER_SUMO 14 /* 0x0E */ // Sumo type, used internally -#define ATOM_PP_THERMALCONTROLLER_NISLANDS 15 -#define ATOM_PP_THERMALCONTROLLER_SISLANDS 16 -#define ATOM_PP_THERMALCONTROLLER_LM96163 17 -#define ATOM_PP_THERMALCONTROLLER_CISLANDS 18 - -// Thermal controller 'combo type' to use an external controller for Fan control and an internal controller for thermal. -// We probably should reserve the bit 0x80 for this use. -// To keep the number of these types low we should also use the same code for all ASICs (i.e. do not distinguish RV6xx and RV7xx Internal here). -// The driver can pick the correct internal controller based on the ASIC. - -#define ATOM_PP_THERMALCONTROLLER_ADT7473_WITH_INTERNAL 0x89 // ADT7473 Fan Control + Internal Thermal Controller -#define ATOM_PP_THERMALCONTROLLER_EMC2103_WITH_INTERNAL 0x8D // EMC2103 Fan Control + Internal Thermal Controller - -typedef struct _ATOM_PPLIB_STATE -{ - UCHAR ucNonClockStateIndex; - UCHAR ucClockStateIndices[1]; // variable-sized -} ATOM_PPLIB_STATE; - - -typedef struct _ATOM_PPLIB_FANTABLE -{ - UCHAR ucFanTableFormat; // Change this if the table format changes or version changes so that the other fields are not the same. - UCHAR ucTHyst; // Temperature hysteresis. Integer. - USHORT usTMin; // The temperature, in 0.01 centigrades, below which we just run at a minimal PWM. - USHORT usTMed; // The middle temperature where we change slopes. - USHORT usTHigh; // The high point above TMed for adjusting the second slope. - USHORT usPWMMin; // The minimum PWM value in percent (0.01% increments). - USHORT usPWMMed; // The PWM value (in percent) at TMed. - USHORT usPWMHigh; // The PWM value at THigh. -} ATOM_PPLIB_FANTABLE; - -typedef struct _ATOM_PPLIB_FANTABLE2 -{ - ATOM_PPLIB_FANTABLE basicTable; - USHORT usTMax; // The max temperature -} ATOM_PPLIB_FANTABLE2; - -typedef struct _ATOM_PPLIB_EXTENDEDHEADER -{ - USHORT usSize; - ULONG ulMaxEngineClock; // For Overdrive. - ULONG ulMaxMemoryClock; // For Overdrive. - // Add extra system parameters here, always adjust size to include all fields. - USHORT usVCETableOffset; //points to ATOM_PPLIB_VCE_Table - USHORT usUVDTableOffset; //points to ATOM_PPLIB_UVD_Table - USHORT usSAMUTableOffset; //points to ATOM_PPLIB_SAMU_Table - USHORT usPPMTableOffset; //points to ATOM_PPLIB_PPM_Table -} ATOM_PPLIB_EXTENDEDHEADER; - -//// ATOM_PPLIB_POWERPLAYTABLE::ulPlatformCaps -#define ATOM_PP_PLATFORM_CAP_BACKBIAS 1 -#define ATOM_PP_PLATFORM_CAP_POWERPLAY 2 -#define ATOM_PP_PLATFORM_CAP_SBIOSPOWERSOURCE 4 -#define ATOM_PP_PLATFORM_CAP_ASPM_L0s 8 -#define ATOM_PP_PLATFORM_CAP_ASPM_L1 16 -#define ATOM_PP_PLATFORM_CAP_HARDWAREDC 32 -#define ATOM_PP_PLATFORM_CAP_GEMINIPRIMARY 64 -#define ATOM_PP_PLATFORM_CAP_STEPVDDC 128 -#define ATOM_PP_PLATFORM_CAP_VOLTAGECONTROL 256 -#define ATOM_PP_PLATFORM_CAP_SIDEPORTCONTROL 512 -#define ATOM_PP_PLATFORM_CAP_TURNOFFPLL_ASPML1 1024 -#define ATOM_PP_PLATFORM_CAP_HTLINKCONTROL 2048 -#define ATOM_PP_PLATFORM_CAP_MVDDCONTROL 4096 -#define ATOM_PP_PLATFORM_CAP_GOTO_BOOT_ON_ALERT 0x2000 // Go to boot state on alerts, e.g. on an AC->DC transition. -#define ATOM_PP_PLATFORM_CAP_DONT_WAIT_FOR_VBLANK_ON_ALERT 0x4000 // Do NOT wait for VBLANK during an alert (e.g. AC->DC transition). -#define ATOM_PP_PLATFORM_CAP_VDDCI_CONTROL 0x8000 // Does the driver control VDDCI independently from VDDC. -#define ATOM_PP_PLATFORM_CAP_REGULATOR_HOT 0x00010000 // Enable the 'regulator hot' feature. -#define ATOM_PP_PLATFORM_CAP_BACO 0x00020000 // Does the driver supports BACO state. -#define ATOM_PP_PLATFORM_CAP_NEW_CAC_VOLTAGE 0x00040000 // Does the driver supports new CAC voltage table. -#define ATOM_PP_PLATFORM_CAP_REVERT_GPIO5_POLARITY 0x00080000 // Does the driver supports revert GPIO5 polarity. -#define ATOM_PP_PLATFORM_CAP_OUTPUT_THERMAL2GPIO17 0x00100000 // Does the driver supports thermal2GPIO17. -#define ATOM_PP_PLATFORM_CAP_VRHOT_GPIO_CONFIGURABLE 0x00200000 // Does the driver supports VR HOT GPIO Configurable. - -typedef struct _ATOM_PPLIB_POWERPLAYTABLE -{ - ATOM_COMMON_TABLE_HEADER sHeader; - - UCHAR ucDataRevision; - - UCHAR ucNumStates; - UCHAR ucStateEntrySize; - UCHAR ucClockInfoSize; - UCHAR ucNonClockSize; - - // offset from start of this table to array of ucNumStates ATOM_PPLIB_STATE structures - USHORT usStateArrayOffset; - - // offset from start of this table to array of ASIC-specific structures, - // currently ATOM_PPLIB_CLOCK_INFO. - USHORT usClockInfoArrayOffset; - - // offset from start of this table to array of ATOM_PPLIB_NONCLOCK_INFO - USHORT usNonClockInfoArrayOffset; - - USHORT usBackbiasTime; // in microseconds - USHORT usVoltageTime; // in microseconds - USHORT usTableSize; //the size of this structure, or the extended structure - - ULONG ulPlatformCaps; // See ATOM_PPLIB_CAPS_* - - ATOM_PPLIB_THERMALCONTROLLER sThermalController; - - USHORT usBootClockInfoOffset; - USHORT usBootNonClockInfoOffset; - -} ATOM_PPLIB_POWERPLAYTABLE; - -typedef struct _ATOM_PPLIB_POWERPLAYTABLE2 -{ - ATOM_PPLIB_POWERPLAYTABLE basicTable; - UCHAR ucNumCustomThermalPolicy; - USHORT usCustomThermalPolicyArrayOffset; -}ATOM_PPLIB_POWERPLAYTABLE2, *LPATOM_PPLIB_POWERPLAYTABLE2; - -typedef struct _ATOM_PPLIB_POWERPLAYTABLE3 -{ - ATOM_PPLIB_POWERPLAYTABLE2 basicTable2; - USHORT usFormatID; // To be used ONLY by PPGen. - USHORT usFanTableOffset; - USHORT usExtendendedHeaderOffset; -} ATOM_PPLIB_POWERPLAYTABLE3, *LPATOM_PPLIB_POWERPLAYTABLE3; - -typedef struct _ATOM_PPLIB_POWERPLAYTABLE4 -{ - ATOM_PPLIB_POWERPLAYTABLE3 basicTable3; - ULONG ulGoldenPPID; // PPGen use only - ULONG ulGoldenRevision; // PPGen use only - USHORT usVddcDependencyOnSCLKOffset; - USHORT usVddciDependencyOnMCLKOffset; - USHORT usVddcDependencyOnMCLKOffset; - USHORT usMaxClockVoltageOnDCOffset; - USHORT usVddcPhaseShedLimitsTableOffset; // Points to ATOM_PPLIB_PhaseSheddingLimits_Table - USHORT usMvddDependencyOnMCLKOffset; -} ATOM_PPLIB_POWERPLAYTABLE4, *LPATOM_PPLIB_POWERPLAYTABLE4; - -typedef struct _ATOM_PPLIB_POWERPLAYTABLE5 -{ - ATOM_PPLIB_POWERPLAYTABLE4 basicTable4; - ULONG ulTDPLimit; - ULONG ulNearTDPLimit; - ULONG ulSQRampingThreshold; - USHORT usCACLeakageTableOffset; // Points to ATOM_PPLIB_CAC_Leakage_Table - ULONG ulCACLeakage; // The iLeakage for driver calculated CAC leakage table - USHORT usTDPODLimit; - USHORT usLoadLineSlope; // in milliOhms * 100 -} ATOM_PPLIB_POWERPLAYTABLE5, *LPATOM_PPLIB_POWERPLAYTABLE5; - -//// ATOM_PPLIB_NONCLOCK_INFO::usClassification -#define ATOM_PPLIB_CLASSIFICATION_UI_MASK 0x0007 -#define ATOM_PPLIB_CLASSIFICATION_UI_SHIFT 0 -#define ATOM_PPLIB_CLASSIFICATION_UI_NONE 0 -#define ATOM_PPLIB_CLASSIFICATION_UI_BATTERY 1 -#define ATOM_PPLIB_CLASSIFICATION_UI_BALANCED 3 -#define ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE 5 -// 2, 4, 6, 7 are reserved - -#define ATOM_PPLIB_CLASSIFICATION_BOOT 0x0008 -#define ATOM_PPLIB_CLASSIFICATION_THERMAL 0x0010 -#define ATOM_PPLIB_CLASSIFICATION_LIMITEDPOWERSOURCE 0x0020 -#define ATOM_PPLIB_CLASSIFICATION_REST 0x0040 -#define ATOM_PPLIB_CLASSIFICATION_FORCED 0x0080 -#define ATOM_PPLIB_CLASSIFICATION_3DPERFORMANCE 0x0100 -#define ATOM_PPLIB_CLASSIFICATION_OVERDRIVETEMPLATE 0x0200 -#define ATOM_PPLIB_CLASSIFICATION_UVDSTATE 0x0400 -#define ATOM_PPLIB_CLASSIFICATION_3DLOW 0x0800 -#define ATOM_PPLIB_CLASSIFICATION_ACPI 0x1000 -#define ATOM_PPLIB_CLASSIFICATION_HD2STATE 0x2000 -#define ATOM_PPLIB_CLASSIFICATION_HDSTATE 0x4000 -#define ATOM_PPLIB_CLASSIFICATION_SDSTATE 0x8000 - -//// ATOM_PPLIB_NONCLOCK_INFO::usClassification2 -#define ATOM_PPLIB_CLASSIFICATION2_LIMITEDPOWERSOURCE_2 0x0001 -#define ATOM_PPLIB_CLASSIFICATION2_ULV 0x0002 -#define ATOM_PPLIB_CLASSIFICATION2_MVC 0x0004 //Multi-View Codec (BD-3D) - -//// ATOM_PPLIB_NONCLOCK_INFO::ulCapsAndSettings -#define ATOM_PPLIB_SINGLE_DISPLAY_ONLY 0x00000001 -#define ATOM_PPLIB_SUPPORTS_VIDEO_PLAYBACK 0x00000002 - -// 0 is 2.5Gb/s, 1 is 5Gb/s -#define ATOM_PPLIB_PCIE_LINK_SPEED_MASK 0x00000004 -#define ATOM_PPLIB_PCIE_LINK_SPEED_SHIFT 2 - -// lanes - 1: 1, 2, 4, 8, 12, 16 permitted by PCIE spec -#define ATOM_PPLIB_PCIE_LINK_WIDTH_MASK 0x000000F8 -#define ATOM_PPLIB_PCIE_LINK_WIDTH_SHIFT 3 - -// lookup into reduced refresh-rate table -#define ATOM_PPLIB_LIMITED_REFRESHRATE_VALUE_MASK 0x00000F00 -#define ATOM_PPLIB_LIMITED_REFRESHRATE_VALUE_SHIFT 8 - -#define ATOM_PPLIB_LIMITED_REFRESHRATE_UNLIMITED 0 -#define ATOM_PPLIB_LIMITED_REFRESHRATE_50HZ 1 -// 2-15 TBD as needed. - -#define ATOM_PPLIB_SOFTWARE_DISABLE_LOADBALANCING 0x00001000 -#define ATOM_PPLIB_SOFTWARE_ENABLE_SLEEP_FOR_TIMESTAMPS 0x00002000 - -#define ATOM_PPLIB_DISALLOW_ON_DC 0x00004000 - -#define ATOM_PPLIB_ENABLE_VARIBRIGHT 0x00008000 - -//memory related flags -#define ATOM_PPLIB_SWSTATE_MEMORY_DLL_OFF 0x000010000 - -//M3 Arb //2bits, current 3 sets of parameters in total -#define ATOM_PPLIB_M3ARB_MASK 0x00060000 -#define ATOM_PPLIB_M3ARB_SHIFT 17 - -#define ATOM_PPLIB_ENABLE_DRR 0x00080000 - -// remaining 16 bits are reserved -typedef struct _ATOM_PPLIB_THERMAL_STATE -{ - UCHAR ucMinTemperature; - UCHAR ucMaxTemperature; - UCHAR ucThermalAction; -}ATOM_PPLIB_THERMAL_STATE, *LPATOM_PPLIB_THERMAL_STATE; - -// Contained in an array starting at the offset -// in ATOM_PPLIB_POWERPLAYTABLE::usNonClockInfoArrayOffset. -// referenced from ATOM_PPLIB_STATE_INFO::ucNonClockStateIndex -#define ATOM_PPLIB_NONCLOCKINFO_VER1 12 -#define ATOM_PPLIB_NONCLOCKINFO_VER2 24 -typedef struct _ATOM_PPLIB_NONCLOCK_INFO -{ - USHORT usClassification; - UCHAR ucMinTemperature; - UCHAR ucMaxTemperature; - ULONG ulCapsAndSettings; - UCHAR ucRequiredPower; - USHORT usClassification2; - ULONG ulVCLK; - ULONG ulDCLK; - UCHAR ucUnused[5]; -} ATOM_PPLIB_NONCLOCK_INFO; - -// Contained in an array starting at the offset -// in ATOM_PPLIB_POWERPLAYTABLE::usClockInfoArrayOffset. -// referenced from ATOM_PPLIB_STATE::ucClockStateIndices -typedef struct _ATOM_PPLIB_R600_CLOCK_INFO -{ - USHORT usEngineClockLow; - UCHAR ucEngineClockHigh; - - USHORT usMemoryClockLow; - UCHAR ucMemoryClockHigh; - - USHORT usVDDC; - USHORT usUnused1; - USHORT usUnused2; - - ULONG ulFlags; // ATOM_PPLIB_R600_FLAGS_* - -} ATOM_PPLIB_R600_CLOCK_INFO; - -// ulFlags in ATOM_PPLIB_R600_CLOCK_INFO -#define ATOM_PPLIB_R600_FLAGS_PCIEGEN2 1 -#define ATOM_PPLIB_R600_FLAGS_UVDSAFE 2 -#define ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE 4 -#define ATOM_PPLIB_R600_FLAGS_MEMORY_ODT_OFF 8 -#define ATOM_PPLIB_R600_FLAGS_MEMORY_DLL_OFF 16 -#define ATOM_PPLIB_R600_FLAGS_LOWPOWER 32 // On the RV770 use 'low power' setting (sequencer S0). - -typedef struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO -{ - USHORT usEngineClockLow; - UCHAR ucEngineClockHigh; - - USHORT usMemoryClockLow; - UCHAR ucMemoryClockHigh; - - USHORT usVDDC; - USHORT usVDDCI; - USHORT usUnused; - - ULONG ulFlags; // ATOM_PPLIB_R600_FLAGS_* - -} ATOM_PPLIB_EVERGREEN_CLOCK_INFO; - -typedef struct _ATOM_PPLIB_SI_CLOCK_INFO -{ - USHORT usEngineClockLow; - UCHAR ucEngineClockHigh; - - USHORT usMemoryClockLow; - UCHAR ucMemoryClockHigh; - - USHORT usVDDC; - USHORT usVDDCI; - UCHAR ucPCIEGen; - UCHAR ucUnused1; - - ULONG ulFlags; // ATOM_PPLIB_SI_FLAGS_*, no flag is necessary for now - -} ATOM_PPLIB_SI_CLOCK_INFO; - -typedef struct _ATOM_PPLIB_CI_CLOCK_INFO -{ - USHORT usEngineClockLow; - UCHAR ucEngineClockHigh; - - USHORT usMemoryClockLow; - UCHAR ucMemoryClockHigh; - - UCHAR ucPCIEGen; - USHORT usPCIELane; -} ATOM_PPLIB_CI_CLOCK_INFO; - -typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO - -{ - USHORT usLowEngineClockLow; // Low Engine clock in MHz (the same way as on the R600). - UCHAR ucLowEngineClockHigh; - USHORT usHighEngineClockLow; // High Engine clock in MHz. - UCHAR ucHighEngineClockHigh; - USHORT usMemoryClockLow; // For now one of the ATOM_PPLIB_RS780_SPMCLK_XXXX constants. - UCHAR ucMemoryClockHigh; // Currentyl unused. - UCHAR ucPadding; // For proper alignment and size. - USHORT usVDDC; // For the 780, use: None, Low, High, Variable - UCHAR ucMaxHTLinkWidth; // From SBIOS - {2, 4, 8, 16} - UCHAR ucMinHTLinkWidth; // From SBIOS - {2, 4, 8, 16}. Effective only if CDLW enabled. Minimum down stream width could be bigger as display BW requriement. - USHORT usHTLinkFreq; // See definition ATOM_PPLIB_RS780_HTLINKFREQ_xxx or in MHz(>=200). - ULONG ulFlags; -} ATOM_PPLIB_RS780_CLOCK_INFO; - -#define ATOM_PPLIB_RS780_VOLTAGE_NONE 0 -#define ATOM_PPLIB_RS780_VOLTAGE_LOW 1 -#define ATOM_PPLIB_RS780_VOLTAGE_HIGH 2 -#define ATOM_PPLIB_RS780_VOLTAGE_VARIABLE 3 - -#define ATOM_PPLIB_RS780_SPMCLK_NONE 0 // We cannot change the side port memory clock, leave it as it is. -#define ATOM_PPLIB_RS780_SPMCLK_LOW 1 -#define ATOM_PPLIB_RS780_SPMCLK_HIGH 2 - -#define ATOM_PPLIB_RS780_HTLINKFREQ_NONE 0 -#define ATOM_PPLIB_RS780_HTLINKFREQ_LOW 1 -#define ATOM_PPLIB_RS780_HTLINKFREQ_HIGH 2 - -typedef struct _ATOM_PPLIB_SUMO_CLOCK_INFO{ - USHORT usEngineClockLow; //clockfrequency & 0xFFFF. The unit is in 10khz - UCHAR ucEngineClockHigh; //clockfrequency >> 16. - UCHAR vddcIndex; //2-bit vddc index; - USHORT tdpLimit; - //please initalize to 0 - USHORT rsv1; - //please initialize to 0s - ULONG rsv2[2]; -}ATOM_PPLIB_SUMO_CLOCK_INFO; - - - -typedef struct _ATOM_PPLIB_STATE_V2 -{ - //number of valid dpm levels in this state; Driver uses it to calculate the whole - //size of the state: sizeof(ATOM_PPLIB_STATE_V2) + (ucNumDPMLevels - 1) * sizeof(UCHAR) - UCHAR ucNumDPMLevels; - - //a index to the array of nonClockInfos - UCHAR nonClockInfoIndex; - /** - * Driver will read the first ucNumDPMLevels in this array - */ - UCHAR clockInfoIndex[1]; -} ATOM_PPLIB_STATE_V2; - -typedef struct _StateArray{ - //how many states we have - UCHAR ucNumEntries; - - ATOM_PPLIB_STATE_V2 states[1]; -}StateArray; - - -typedef struct _ClockInfoArray{ - //how many clock levels we have - UCHAR ucNumEntries; - - //sizeof(ATOM_PPLIB_CLOCK_INFO) - UCHAR ucEntrySize; - - UCHAR clockInfo[1]; -}ClockInfoArray; - -typedef struct _NonClockInfoArray{ - - //how many non-clock levels we have. normally should be same as number of states - UCHAR ucNumEntries; - //sizeof(ATOM_PPLIB_NONCLOCK_INFO) - UCHAR ucEntrySize; - - ATOM_PPLIB_NONCLOCK_INFO nonClockInfo[1]; -}NonClockInfoArray; - -typedef struct _ATOM_PPLIB_Clock_Voltage_Dependency_Record -{ - USHORT usClockLow; - UCHAR ucClockHigh; - USHORT usVoltage; -}ATOM_PPLIB_Clock_Voltage_Dependency_Record; - -typedef struct _ATOM_PPLIB_Clock_Voltage_Dependency_Table -{ - UCHAR ucNumEntries; // Number of entries. - ATOM_PPLIB_Clock_Voltage_Dependency_Record entries[1]; // Dynamically allocate entries. -}ATOM_PPLIB_Clock_Voltage_Dependency_Table; - -typedef struct _ATOM_PPLIB_Clock_Voltage_Limit_Record -{ - USHORT usSclkLow; - UCHAR ucSclkHigh; - USHORT usMclkLow; - UCHAR ucMclkHigh; - USHORT usVddc; - USHORT usVddci; -}ATOM_PPLIB_Clock_Voltage_Limit_Record; - -typedef struct _ATOM_PPLIB_Clock_Voltage_Limit_Table -{ - UCHAR ucNumEntries; // Number of entries. - ATOM_PPLIB_Clock_Voltage_Limit_Record entries[1]; // Dynamically allocate entries. -}ATOM_PPLIB_Clock_Voltage_Limit_Table; - -typedef struct _ATOM_PPLIB_CAC_Leakage_Record -{ - USHORT usVddc; // We use this field for the "fake" standardized VDDC for power calculations; For CI and newer, we use this as the real VDDC value. - ULONG ulLeakageValue; // For CI and newer we use this as the "fake" standar VDDC value. -}ATOM_PPLIB_CAC_Leakage_Record; - -typedef struct _ATOM_PPLIB_CAC_Leakage_Table -{ - UCHAR ucNumEntries; // Number of entries. - ATOM_PPLIB_CAC_Leakage_Record entries[1]; // Dynamically allocate entries. -}ATOM_PPLIB_CAC_Leakage_Table; - -typedef struct _ATOM_PPLIB_PhaseSheddingLimits_Record -{ - USHORT usVoltage; - USHORT usSclkLow; - UCHAR ucSclkHigh; - USHORT usMclkLow; - UCHAR ucMclkHigh; -}ATOM_PPLIB_PhaseSheddingLimits_Record; - -typedef struct _ATOM_PPLIB_PhaseSheddingLimits_Table -{ - UCHAR ucNumEntries; // Number of entries. - ATOM_PPLIB_PhaseSheddingLimits_Record entries[1]; // Dynamically allocate entries. -}ATOM_PPLIB_PhaseSheddingLimits_Table; - -typedef struct _VCEClockInfo{ - USHORT usEVClkLow; - UCHAR ucEVClkHigh; - USHORT usECClkLow; - UCHAR ucECClkHigh; -}VCEClockInfo; - -typedef struct _VCEClockInfoArray{ - UCHAR ucNumEntries; - VCEClockInfo entries[1]; -}VCEClockInfoArray; - -typedef struct _ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record -{ - USHORT usVoltage; - UCHAR ucVCEClockInfoIndex; -}ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record; - -typedef struct _ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table -{ - UCHAR numEntries; - ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record entries[1]; -}ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table; - -typedef struct _ATOM_PPLIB_VCE_State_Record -{ - UCHAR ucVCEClockInfoIndex; - UCHAR ucClockInfoIndex; //highest 2 bits indicates memory p-states, lower 6bits indicates index to ClockInfoArrary -}ATOM_PPLIB_VCE_State_Record; - -typedef struct _ATOM_PPLIB_VCE_State_Table -{ - UCHAR numEntries; - ATOM_PPLIB_VCE_State_Record entries[1]; -}ATOM_PPLIB_VCE_State_Table; - - -typedef struct _ATOM_PPLIB_VCE_Table -{ - UCHAR revid; -// VCEClockInfoArray array; -// ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table limits; -// ATOM_PPLIB_VCE_State_Table states; -}ATOM_PPLIB_VCE_Table; - - -typedef struct _UVDClockInfo{ - USHORT usVClkLow; - UCHAR ucVClkHigh; - USHORT usDClkLow; - UCHAR ucDClkHigh; -}UVDClockInfo; - -typedef struct _UVDClockInfoArray{ - UCHAR ucNumEntries; - UVDClockInfo entries[1]; -}UVDClockInfoArray; - -typedef struct _ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record -{ - USHORT usVoltage; - UCHAR ucUVDClockInfoIndex; -}ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record; - -typedef struct _ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table -{ - UCHAR numEntries; - ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record entries[1]; -}ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table; - -typedef struct _ATOM_PPLIB_UVD_State_Record -{ - UCHAR ucUVDClockInfoIndex; - UCHAR ucClockInfoIndex; //highest 2 bits indicates memory p-states, lower 6bits indicates index to ClockInfoArrary -}ATOM_PPLIB_UVD_State_Record; - -typedef struct _ATOM_PPLIB_UVD_State_Table -{ - UCHAR numEntries; - ATOM_PPLIB_UVD_State_Record entries[1]; -}ATOM_PPLIB_UVD_State_Table; - - -typedef struct _ATOM_PPLIB_UVD_Table -{ - UCHAR revid; -// UVDClockInfoArray array; -// ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table limits; -// ATOM_PPLIB_UVD_State_Table states; -}ATOM_PPLIB_UVD_Table; - - -typedef struct _ATOM_PPLIB_SAMClk_Voltage_Limit_Record -{ - USHORT usVoltage; - USHORT usSAMClockLow; - UCHAR ucSAMClockHigh; -}ATOM_PPLIB_SAMClk_Voltage_Limit_Record; - -typedef struct _ATOM_PPLIB_SAMClk_Voltage_Limit_Table{ - UCHAR numEntries; - ATOM_PPLIB_SAMClk_Voltage_Limit_Record entries[1]; -}ATOM_PPLIB_SAMClk_Voltage_Limit_Table; - -typedef struct _ATOM_PPLIB_SAMU_Table -{ - UCHAR revid; - ATOM_PPLIB_SAMClk_Voltage_Limit_Table limits; -}ATOM_PPLIB_SAMU_Table; - -#define ATOM_PPM_A_A 1 -#define ATOM_PPM_A_I 2 -typedef struct _ATOM_PPLIB_PPM_Table -{ - UCHAR ucRevId; - UCHAR ucPpmDesign; //A+I or A+A - USHORT usCpuCoreNumber; - ULONG ulPlatformTDP; - ULONG ulSmallACPlatformTDP; - ULONG ulPlatformTDC; - ULONG ulSmallACPlatformTDC; - ULONG ulApuTDP; - ULONG ulDGpuTDP; - ULONG ulDGpuUlvPower; - ULONG ulTjmax; -} ATOM_PPLIB_PPM_Table; - -/**************************************************************************/ - // Following definitions are for compatibility issue in different SW components. #define ATOM_MASTER_DATA_TABLE_REVISION 0x01 @@ -8485,3 +7976,6 @@ typedef struct { #endif /* _ATOMBIOS_H */ + +#include "pptable.h" + diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index b9d3b43f19c..30d242b2507 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -209,6 +209,16 @@ static void atombios_enable_crtc_memreq(struct drm_crtc *crtc, int state) atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } +static const u32 vga_control_regs[6] = +{ + AVIVO_D1VGA_CONTROL, + AVIVO_D2VGA_CONTROL, + EVERGREEN_D3VGA_CONTROL, + EVERGREEN_D4VGA_CONTROL, + EVERGREEN_D5VGA_CONTROL, + EVERGREEN_D6VGA_CONTROL, +}; + static void atombios_blank_crtc(struct drm_crtc *crtc, int state) { struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); @@ -216,13 +226,23 @@ static void atombios_blank_crtc(struct drm_crtc *crtc, int state) struct radeon_device *rdev = dev->dev_private; int index = GetIndexIntoMasterTable(COMMAND, BlankCRTC); BLANK_CRTC_PS_ALLOCATION args; + u32 vga_control = 0; memset(&args, 0, sizeof(args)); + if (ASIC_IS_DCE8(rdev)) { + vga_control = RREG32(vga_control_regs[radeon_crtc->crtc_id]); + WREG32(vga_control_regs[radeon_crtc->crtc_id], vga_control | 1); + } + args.ucCRTC = radeon_crtc->crtc_id; args.ucBlanking = state; atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + + if (ASIC_IS_DCE8(rdev)) { + WREG32(vga_control_regs[radeon_crtc->crtc_id], vga_control); + } } static void atombios_powergate_crtc(struct drm_crtc *crtc, int state) @@ -250,8 +270,6 @@ void atombios_crtc_dpms(struct drm_crtc *crtc, int mode) switch (mode) { case DRM_MODE_DPMS_ON: radeon_crtc->enabled = true; - /* adjust pm to dpms changes BEFORE enabling crtcs */ - radeon_pm_compute_clocks(rdev); atombios_enable_crtc(crtc, ATOM_ENABLE); if (ASIC_IS_DCE3(rdev) && !ASIC_IS_DCE6(rdev)) atombios_enable_crtc_memreq(crtc, ATOM_ENABLE); @@ -269,10 +287,10 @@ void atombios_crtc_dpms(struct drm_crtc *crtc, int mode) atombios_enable_crtc_memreq(crtc, ATOM_DISABLE); atombios_enable_crtc(crtc, ATOM_DISABLE); radeon_crtc->enabled = false; - /* adjust pm to dpms changes AFTER disabling crtcs */ - radeon_pm_compute_clocks(rdev); break; } + /* adjust pm to dpms */ + radeon_pm_compute_clocks(rdev); } static void @@ -423,7 +441,17 @@ static void atombios_crtc_program_ss(struct radeon_device *rdev, int index = GetIndexIntoMasterTable(COMMAND, EnableSpreadSpectrumOnPPLL); union atom_enable_ss args; - if (!enable) { + if (enable) { + /* Don't mess with SS if percentage is 0 or external ss. + * SS is already disabled previously, and disabling it + * again can cause display problems if the pll is already + * programmed. + */ + if (ss->percentage == 0) + return; + if (ss->type & ATOM_EXTERNAL_SS_MASK) + return; + } else { for (i = 0; i < rdev->num_crtc; i++) { if (rdev->mode_info.crtcs[i] && rdev->mode_info.crtcs[i]->enabled && @@ -459,8 +487,6 @@ static void atombios_crtc_program_ss(struct radeon_device *rdev, args.v3.usSpreadSpectrumAmount = cpu_to_le16(ss->amount); args.v3.usSpreadSpectrumStep = cpu_to_le16(ss->step); args.v3.ucEnable = enable; - if ((ss->percentage == 0) || (ss->type & ATOM_EXTERNAL_SS_MASK) || ASIC_IS_DCE61(rdev)) - args.v3.ucEnable = ATOM_DISABLE; } else if (ASIC_IS_DCE4(rdev)) { args.v2.usSpreadSpectrumPercentage = cpu_to_le16(ss->percentage); args.v2.ucSpreadSpectrumType = ss->type & ATOM_SS_CENTRE_SPREAD_MODE_MASK; @@ -480,8 +506,6 @@ static void atombios_crtc_program_ss(struct radeon_device *rdev, args.v2.usSpreadSpectrumAmount = cpu_to_le16(ss->amount); args.v2.usSpreadSpectrumStep = cpu_to_le16(ss->step); args.v2.ucEnable = enable; - if ((ss->percentage == 0) || (ss->type & ATOM_EXTERNAL_SS_MASK) || ASIC_IS_DCE41(rdev)) - args.v2.ucEnable = ATOM_DISABLE; } else if (ASIC_IS_DCE3(rdev)) { args.v1.usSpreadSpectrumPercentage = cpu_to_le16(ss->percentage); args.v1.ucSpreadSpectrumType = ss->type & ATOM_SS_CENTRE_SPREAD_MODE_MASK; @@ -503,8 +527,7 @@ static void atombios_crtc_program_ss(struct radeon_device *rdev, args.lvds_ss_2.ucSpreadSpectrumRange = ss->range; args.lvds_ss_2.ucEnable = enable; } else { - if ((enable == ATOM_DISABLE) || (ss->percentage == 0) || - (ss->type & ATOM_EXTERNAL_SS_MASK)) { + if (enable == ATOM_DISABLE) { atombios_disable_ss(rdev, pll_id); return; } @@ -534,7 +557,8 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, u32 adjusted_clock = mode->clock; int encoder_mode = atombios_get_encoder_mode(encoder); u32 dp_clock = mode->clock; - int bpc = radeon_get_monitor_bpc(connector); + u32 clock = mode->clock; + int bpc = radeon_crtc->bpc; bool is_duallink = radeon_dig_monitor_is_duallink(encoder, mode->clock); /* reset the pll flags */ @@ -609,6 +633,24 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, radeon_crtc->pll_flags |= RADEON_PLL_USE_REF_DIV; } + /* adjust pll for deep color modes */ + if (encoder_mode == ATOM_ENCODER_MODE_HDMI) { + switch (bpc) { + case 8: + default: + break; + case 10: + clock = (clock * 5) / 4; + break; + case 12: + clock = (clock * 3) / 2; + break; + case 16: + clock = clock * 2; + break; + } + } + /* DCE3+ has an AdjustDisplayPll that will adjust the pixel clock * accordingly based on the encoder/transmitter to work around * special hw requirements. @@ -630,7 +672,7 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, switch (crev) { case 1: case 2: - args.v1.usPixelClock = cpu_to_le16(mode->clock / 10); + args.v1.usPixelClock = cpu_to_le16(clock / 10); args.v1.ucTransmitterID = radeon_encoder->encoder_id; args.v1.ucEncodeMode = encoder_mode; if (radeon_crtc->ss_enabled && radeon_crtc->ss.percentage) @@ -642,7 +684,7 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, adjusted_clock = le16_to_cpu(args.v1.usPixelClock) * 10; break; case 3: - args.v3.sInput.usPixelClock = cpu_to_le16(mode->clock / 10); + args.v3.sInput.usPixelClock = cpu_to_le16(clock / 10); args.v3.sInput.ucTransmitterID = radeon_encoder->encoder_id; args.v3.sInput.ucEncodeMode = encoder_mode; args.v3.sInput.ucDispPllConfig = 0; @@ -656,10 +698,6 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, args.v3.sInput.usPixelClock = cpu_to_le16(dp_clock / 10); } else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) { struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; - if (encoder_mode == ATOM_ENCODER_MODE_HDMI) - /* deep color support */ - args.v3.sInput.usPixelClock = - cpu_to_le16((mode->clock * bpc / 8) / 10); if (dig->coherent_mode) args.v3.sInput.ucDispPllConfig |= DISPPLL_CONFIG_COHERENT_MODE; @@ -839,14 +877,21 @@ static void atombios_crtc_program_pll(struct drm_crtc *crtc, args.v5.ucMiscInfo = 0; /* HDMI depth, etc. */ if (ss_enabled && (ss->type & ATOM_EXTERNAL_SS_MASK)) args.v5.ucMiscInfo |= PIXEL_CLOCK_V5_MISC_REF_DIV_SRC; - switch (bpc) { - case 8: - default: - args.v5.ucMiscInfo |= PIXEL_CLOCK_V5_MISC_HDMI_24BPP; - break; - case 10: - args.v5.ucMiscInfo |= PIXEL_CLOCK_V5_MISC_HDMI_30BPP; - break; + if (encoder_mode == ATOM_ENCODER_MODE_HDMI) { + switch (bpc) { + case 8: + default: + args.v5.ucMiscInfo |= PIXEL_CLOCK_V5_MISC_HDMI_24BPP; + break; + case 10: + /* yes this is correct, the atom define is wrong */ + args.v5.ucMiscInfo |= PIXEL_CLOCK_V5_MISC_HDMI_32BPP; + break; + case 12: + /* yes this is correct, the atom define is wrong */ + args.v5.ucMiscInfo |= PIXEL_CLOCK_V5_MISC_HDMI_30BPP; + break; + } } args.v5.ucTransmitterID = encoder_id; args.v5.ucEncoderMode = encoder_mode; @@ -861,20 +906,22 @@ static void atombios_crtc_program_pll(struct drm_crtc *crtc, args.v6.ucMiscInfo = 0; /* HDMI depth, etc. */ if (ss_enabled && (ss->type & ATOM_EXTERNAL_SS_MASK)) args.v6.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_REF_DIV_SRC; - switch (bpc) { - case 8: - default: - args.v6.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_24BPP; - break; - case 10: - args.v6.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_30BPP; - break; - case 12: - args.v6.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_36BPP; - break; - case 16: - args.v6.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_48BPP; - break; + if (encoder_mode == ATOM_ENCODER_MODE_HDMI) { + switch (bpc) { + case 8: + default: + args.v6.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_24BPP; + break; + case 10: + args.v6.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_30BPP_V6; + break; + case 12: + args.v6.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_36BPP_V6; + break; + case 16: + args.v6.ucMiscInfo |= PIXEL_CLOCK_V6_MISC_HDMI_48BPP; + break; + } } args.v6.ucTransmitterID = encoder_id; args.v6.ucEncoderMode = encoder_mode; @@ -915,6 +962,9 @@ static bool atombios_crtc_prepare_pll(struct drm_crtc *crtc, struct drm_display_ struct radeon_connector_atom_dig *dig_connector = radeon_connector->con_priv; int dp_clock; + + /* Assign mode clock for hdmi deep color max clock limit check */ + radeon_connector->pixelclock_for_modeset = mode->clock; radeon_crtc->bpc = radeon_get_monitor_bpc(connector); switch (encoder_mode) { @@ -938,11 +988,14 @@ static bool atombios_crtc_prepare_pll(struct drm_crtc *crtc, struct drm_display_ radeon_atombios_get_ppll_ss_info(rdev, &radeon_crtc->ss, ATOM_DP_SS_ID1); - } else + } else { radeon_crtc->ss_enabled = radeon_atombios_get_ppll_ss_info(rdev, &radeon_crtc->ss, ATOM_DP_SS_ID1); + } + /* disable spread spectrum on DCE3 DP */ + radeon_crtc->ss_enabled = false; } break; case ATOM_ENCODER_MODE_LVDS: @@ -993,10 +1046,17 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode struct radeon_encoder *radeon_encoder = to_radeon_encoder(radeon_crtc->encoder); u32 pll_clock = mode->clock; + u32 clock = mode->clock; u32 ref_div = 0, fb_div = 0, frac_fb_div = 0, post_div = 0; struct radeon_pll *pll; int encoder_mode = atombios_get_encoder_mode(radeon_crtc->encoder); + /* pass the actual clock to atombios_crtc_program_pll for DCE5,6 for HDMI */ + if (ASIC_IS_DCE5(rdev) && + (encoder_mode == ATOM_ENCODER_MODE_HDMI) && + (radeon_crtc->bpc > 8)) + clock = radeon_crtc->adjusted_clock; + switch (radeon_crtc->pll_id) { case ATOM_PPLL1: pll = &rdev->clock.p1pll; @@ -1031,7 +1091,7 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode radeon_crtc->crtc_id, &radeon_crtc->ss); atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id, - encoder_mode, radeon_encoder->encoder_id, mode->clock, + encoder_mode, radeon_encoder->encoder_id, clock, ref_div, fb_div, frac_fb_div, post_div, radeon_crtc->bpc, radeon_crtc->ss_enabled, &radeon_crtc->ss); @@ -1039,15 +1099,17 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode /* calculate ss amount and step size */ if (ASIC_IS_DCE4(rdev)) { u32 step_size; - u32 amount = (((fb_div * 10) + frac_fb_div) * radeon_crtc->ss.percentage) / 10000; + u32 amount = (((fb_div * 10) + frac_fb_div) * + (u32)radeon_crtc->ss.percentage) / + (100 * (u32)radeon_crtc->ss.percentage_divider); radeon_crtc->ss.amount = (amount / 10) & ATOM_PPLL_SS_AMOUNT_V2_FBDIV_MASK; radeon_crtc->ss.amount |= ((amount - (amount / 10)) << ATOM_PPLL_SS_AMOUNT_V2_NFRAC_SHIFT) & ATOM_PPLL_SS_AMOUNT_V2_NFRAC_MASK; if (radeon_crtc->ss.type & ATOM_PPLL_SS_TYPE_V2_CENTRE_SPREAD) - step_size = (4 * amount * ref_div * (radeon_crtc->ss.rate * 2048)) / + step_size = (4 * amount * ref_div * ((u32)radeon_crtc->ss.rate * 2048)) / (125 * 25 * pll->reference_freq / 100); else - step_size = (2 * amount * ref_div * (radeon_crtc->ss.rate * 2048)) / + step_size = (2 * amount * ref_div * ((u32)radeon_crtc->ss.rate * 2048)) / (125 * 25 * pll->reference_freq / 100); radeon_crtc->ss.step = step_size; } @@ -1074,9 +1136,10 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc, u32 fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_NONE); u32 tmp, viewport_w, viewport_h; int r; + bool bypass_lut = false; /* no fb bound */ - if (!atomic && !crtc->fb) { + if (!atomic && !crtc->primary->fb) { DRM_DEBUG_KMS("No FB bound\n"); return 0; } @@ -1086,8 +1149,8 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc, target_fb = fb; } else { - radeon_fb = to_radeon_framebuffer(crtc->fb); - target_fb = crtc->fb; + radeon_fb = to_radeon_framebuffer(crtc->primary->fb); + target_fb = crtc->primary->fb; } /* If atomic, assume fb object is pinned & idle & fenced and @@ -1112,62 +1175,141 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc, radeon_bo_get_tiling_flags(rbo, &tiling_flags, NULL); radeon_bo_unreserve(rbo); - switch (target_fb->bits_per_pixel) { - case 8: + switch (target_fb->pixel_format) { + case DRM_FORMAT_C8: fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_8BPP) | EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_INDEXED)); break; - case 15: + case DRM_FORMAT_XRGB4444: + case DRM_FORMAT_ARGB4444: + fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_16BPP) | + EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB4444)); +#ifdef __BIG_ENDIAN + fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN16); +#endif + break; + case DRM_FORMAT_XRGB1555: + case DRM_FORMAT_ARGB1555: fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_16BPP) | EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB1555)); +#ifdef __BIG_ENDIAN + fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN16); +#endif + break; + case DRM_FORMAT_BGRX5551: + case DRM_FORMAT_BGRA5551: + fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_16BPP) | + EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_BGRA5551)); +#ifdef __BIG_ENDIAN + fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN16); +#endif break; - case 16: + case DRM_FORMAT_RGB565: fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_16BPP) | EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB565)); #ifdef __BIG_ENDIAN fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN16); #endif break; - case 24: - case 32: + case DRM_FORMAT_XRGB8888: + case DRM_FORMAT_ARGB8888: fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_32BPP) | EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB8888)); #ifdef __BIG_ENDIAN fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN32); #endif break; + case DRM_FORMAT_XRGB2101010: + case DRM_FORMAT_ARGB2101010: + fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_32BPP) | + EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB2101010)); +#ifdef __BIG_ENDIAN + fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN32); +#endif + /* Greater 8 bpc fb needs to bypass hw-lut to retain precision */ + bypass_lut = true; + break; + case DRM_FORMAT_BGRX1010102: + case DRM_FORMAT_BGRA1010102: + fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_32BPP) | + EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_BGRA1010102)); +#ifdef __BIG_ENDIAN + fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN32); +#endif + /* Greater 8 bpc fb needs to bypass hw-lut to retain precision */ + bypass_lut = true; + break; default: - DRM_ERROR("Unsupported screen depth %d\n", - target_fb->bits_per_pixel); + DRM_ERROR("Unsupported screen format %s\n", + drm_get_format_name(target_fb->pixel_format)); return -EINVAL; } if (tiling_flags & RADEON_TILING_MACRO) { - if (rdev->family >= CHIP_BONAIRE) - tmp = rdev->config.cik.tile_config; - else if (rdev->family >= CHIP_TAHITI) - tmp = rdev->config.si.tile_config; - else if (rdev->family >= CHIP_CAYMAN) - tmp = rdev->config.cayman.tile_config; - else - tmp = rdev->config.evergreen.tile_config; + evergreen_tiling_fields(tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); - switch ((tmp & 0xf0) >> 4) { - case 0: /* 4 banks */ - fb_format |= EVERGREEN_GRPH_NUM_BANKS(EVERGREEN_ADDR_SURF_4_BANK); - break; - case 1: /* 8 banks */ - default: - fb_format |= EVERGREEN_GRPH_NUM_BANKS(EVERGREEN_ADDR_SURF_8_BANK); - break; - case 2: /* 16 banks */ - fb_format |= EVERGREEN_GRPH_NUM_BANKS(EVERGREEN_ADDR_SURF_16_BANK); - break; + /* Set NUM_BANKS. */ + if (rdev->family >= CHIP_TAHITI) { + unsigned index, num_banks; + + if (rdev->family >= CHIP_BONAIRE) { + unsigned tileb, tile_split_bytes; + + /* Calculate the macrotile mode index. */ + tile_split_bytes = 64 << tile_split; + tileb = 8 * 8 * target_fb->bits_per_pixel / 8; + tileb = min(tile_split_bytes, tileb); + + for (index = 0; tileb > 64; index++) + tileb >>= 1; + + if (index >= 16) { + DRM_ERROR("Wrong screen bpp (%u) or tile split (%u)\n", + target_fb->bits_per_pixel, tile_split); + return -EINVAL; + } + + num_banks = (rdev->config.cik.macrotile_mode_array[index] >> 6) & 0x3; + } else { + switch (target_fb->bits_per_pixel) { + case 8: + index = 10; + break; + case 16: + index = SI_TILE_MODE_COLOR_2D_SCANOUT_16BPP; + break; + default: + case 32: + index = SI_TILE_MODE_COLOR_2D_SCANOUT_32BPP; + break; + } + + num_banks = (rdev->config.si.tile_mode_array[index] >> 20) & 0x3; + } + + fb_format |= EVERGREEN_GRPH_NUM_BANKS(num_banks); + } else { + /* NI and older. */ + if (rdev->family >= CHIP_CAYMAN) + tmp = rdev->config.cayman.tile_config; + else + tmp = rdev->config.evergreen.tile_config; + + switch ((tmp & 0xf0) >> 4) { + case 0: /* 4 banks */ + fb_format |= EVERGREEN_GRPH_NUM_BANKS(EVERGREEN_ADDR_SURF_4_BANK); + break; + case 1: /* 8 banks */ + default: + fb_format |= EVERGREEN_GRPH_NUM_BANKS(EVERGREEN_ADDR_SURF_8_BANK); + break; + case 2: /* 16 banks */ + fb_format |= EVERGREEN_GRPH_NUM_BANKS(EVERGREEN_ADDR_SURF_16_BANK); + break; + } } fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_2D_TILED_THIN1); - - evergreen_tiling_fields(tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); fb_format |= EVERGREEN_GRPH_TILE_SPLIT(tile_split); fb_format |= EVERGREEN_GRPH_BANK_WIDTH(bankw); fb_format |= EVERGREEN_GRPH_BANK_HEIGHT(bankh); @@ -1180,23 +1322,18 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc, fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_1D_TILED_THIN1); if (rdev->family >= CHIP_BONAIRE) { - u32 num_pipe_configs = rdev->config.cik.max_tile_pipes; - u32 num_rb = rdev->config.cik.max_backends_per_se; - if (num_pipe_configs > 8) - num_pipe_configs = 8; - if (num_pipe_configs == 8) - fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P8_32x32_16x16); - else if (num_pipe_configs == 4) { - if (num_rb == 4) - fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P4_16x16); - else if (num_rb < 4) - fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P4_8x16); - } else if (num_pipe_configs == 2) - fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P2); + /* Read the pipe config from the 2D TILED SCANOUT mode. + * It should be the same for the other modes too, but not all + * modes set the pipe config field. */ + u32 pipe_config = (rdev->config.cik.tile_mode_array[10] >> 6) & 0x1f; + + fb_format |= CIK_GRPH_PIPE_CONFIG(pipe_config); } else if ((rdev->family == CHIP_TAHITI) || (rdev->family == CHIP_PITCAIRN)) fb_format |= SI_GRPH_PIPE_CONFIG(SI_ADDR_SURF_P8_32x32_8x16); - else if (rdev->family == CHIP_VERDE) + else if ((rdev->family == CHIP_VERDE) || + (rdev->family == CHIP_OLAND) || + (rdev->family == CHIP_HAINAN)) /* for completeness. HAINAN has no display hw */ fb_format |= SI_GRPH_PIPE_CONFIG(SI_ADDR_SURF_P4_8x16); switch (radeon_crtc->crtc_id) { @@ -1233,6 +1370,18 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc, WREG32(EVERGREEN_GRPH_CONTROL + radeon_crtc->crtc_offset, fb_format); WREG32(EVERGREEN_GRPH_SWAP_CONTROL + radeon_crtc->crtc_offset, fb_swap); + /* + * The LUT only has 256 slots for indexing by a 8 bpc fb. Bypass the LUT + * for > 8 bpc scanout to avoid truncation of fb indices to 8 msb's, to + * retain the full precision throughout the pipeline. + */ + WREG32_P(EVERGREEN_GRPH_LUT_10BIT_BYPASS_CONTROL + radeon_crtc->crtc_offset, + (bypass_lut ? EVERGREEN_LUT_10BIT_BYPASS_EN : 0), + ~EVERGREEN_LUT_10BIT_BYPASS_EN); + + if (bypass_lut) + DRM_DEBUG_KMS("Bypassing hardware LUT due to 10 bit fb scanout.\n"); + WREG32(EVERGREEN_GRPH_SURFACE_OFFSET_X + radeon_crtc->crtc_offset, 0); WREG32(EVERGREEN_GRPH_SURFACE_OFFSET_Y + radeon_crtc->crtc_offset, 0); WREG32(EVERGREEN_GRPH_X_START + radeon_crtc->crtc_offset, 0); @@ -1265,10 +1414,10 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc, tmp &= ~EVERGREEN_GRPH_SURFACE_UPDATE_H_RETRACE_EN; WREG32(EVERGREEN_GRPH_FLIP_CONTROL + radeon_crtc->crtc_offset, tmp); - /* set pageflip to happen anywhere in vblank interval */ - WREG32(EVERGREEN_MASTER_UPDATE_MODE + radeon_crtc->crtc_offset, 0); + /* set pageflip to happen only at start of vblank interval (front porch) */ + WREG32(EVERGREEN_MASTER_UPDATE_MODE + radeon_crtc->crtc_offset, 3); - if (!atomic && fb && fb != crtc->fb) { + if (!atomic && fb && fb != crtc->primary->fb) { radeon_fb = to_radeon_framebuffer(fb); rbo = gem_to_radeon_bo(radeon_fb->obj); r = radeon_bo_reserve(rbo, false); @@ -1300,9 +1449,10 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc, u32 fb_swap = R600_D1GRPH_SWAP_ENDIAN_NONE; u32 tmp, viewport_w, viewport_h; int r; + bool bypass_lut = false; /* no fb bound */ - if (!atomic && !crtc->fb) { + if (!atomic && !crtc->primary->fb) { DRM_DEBUG_KMS("No FB bound\n"); return 0; } @@ -1312,8 +1462,8 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc, target_fb = fb; } else { - radeon_fb = to_radeon_framebuffer(crtc->fb); - target_fb = crtc->fb; + radeon_fb = to_radeon_framebuffer(crtc->primary->fb); + target_fb = crtc->primary->fb; } obj = radeon_fb->obj; @@ -1337,18 +1487,30 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc, radeon_bo_get_tiling_flags(rbo, &tiling_flags, NULL); radeon_bo_unreserve(rbo); - switch (target_fb->bits_per_pixel) { - case 8: + switch (target_fb->pixel_format) { + case DRM_FORMAT_C8: fb_format = AVIVO_D1GRPH_CONTROL_DEPTH_8BPP | AVIVO_D1GRPH_CONTROL_8BPP_INDEXED; break; - case 15: + case DRM_FORMAT_XRGB4444: + case DRM_FORMAT_ARGB4444: + fb_format = + AVIVO_D1GRPH_CONTROL_DEPTH_16BPP | + AVIVO_D1GRPH_CONTROL_16BPP_ARGB4444; +#ifdef __BIG_ENDIAN + fb_swap = R600_D1GRPH_SWAP_ENDIAN_16BIT; +#endif + break; + case DRM_FORMAT_XRGB1555: fb_format = AVIVO_D1GRPH_CONTROL_DEPTH_16BPP | AVIVO_D1GRPH_CONTROL_16BPP_ARGB1555; +#ifdef __BIG_ENDIAN + fb_swap = R600_D1GRPH_SWAP_ENDIAN_16BIT; +#endif break; - case 16: + case DRM_FORMAT_RGB565: fb_format = AVIVO_D1GRPH_CONTROL_DEPTH_16BPP | AVIVO_D1GRPH_CONTROL_16BPP_RGB565; @@ -1356,8 +1518,8 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc, fb_swap = R600_D1GRPH_SWAP_ENDIAN_16BIT; #endif break; - case 24: - case 32: + case DRM_FORMAT_XRGB8888: + case DRM_FORMAT_ARGB8888: fb_format = AVIVO_D1GRPH_CONTROL_DEPTH_32BPP | AVIVO_D1GRPH_CONTROL_32BPP_ARGB8888; @@ -1365,9 +1527,20 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc, fb_swap = R600_D1GRPH_SWAP_ENDIAN_32BIT; #endif break; + case DRM_FORMAT_XRGB2101010: + case DRM_FORMAT_ARGB2101010: + fb_format = + AVIVO_D1GRPH_CONTROL_DEPTH_32BPP | + AVIVO_D1GRPH_CONTROL_32BPP_ARGB2101010; +#ifdef __BIG_ENDIAN + fb_swap = R600_D1GRPH_SWAP_ENDIAN_32BIT; +#endif + /* Greater 8 bpc fb needs to bypass hw-lut to retain precision */ + bypass_lut = true; + break; default: - DRM_ERROR("Unsupported screen depth %d\n", - target_fb->bits_per_pixel); + DRM_ERROR("Unsupported screen format %s\n", + drm_get_format_name(target_fb->pixel_format)); return -EINVAL; } @@ -1406,6 +1579,13 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc, if (rdev->family >= CHIP_R600) WREG32(R600_D1GRPH_SWAP_CONTROL + radeon_crtc->crtc_offset, fb_swap); + /* LUT only has 256 slots for 8 bpc fb. Bypass for > 8 bpc scanout for precision */ + WREG32_P(AVIVO_D1GRPH_LUT_SEL + radeon_crtc->crtc_offset, + (bypass_lut ? AVIVO_LUT_10BIT_BYPASS_EN : 0), ~AVIVO_LUT_10BIT_BYPASS_EN); + + if (bypass_lut) + DRM_DEBUG_KMS("Bypassing hardware LUT due to 10 bit fb scanout.\n"); + WREG32(AVIVO_D1GRPH_SURFACE_OFFSET_X + radeon_crtc->crtc_offset, 0); WREG32(AVIVO_D1GRPH_SURFACE_OFFSET_Y + radeon_crtc->crtc_offset, 0); WREG32(AVIVO_D1GRPH_X_START + radeon_crtc->crtc_offset, 0); @@ -1434,10 +1614,10 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc, tmp &= ~AVIVO_D1GRPH_SURFACE_UPDATE_H_RETRACE_EN; WREG32(AVIVO_D1GRPH_FLIP_CONTROL + radeon_crtc->crtc_offset, tmp); - /* set pageflip to happen anywhere in vblank interval */ - WREG32(AVIVO_D1MODE_MASTER_UPDATE_MODE + radeon_crtc->crtc_offset, 0); + /* set pageflip to happen only at start of vblank interval (front porch) */ + WREG32(AVIVO_D1MODE_MASTER_UPDATE_MODE + radeon_crtc->crtc_offset, 3); - if (!atomic && fb && fb != crtc->fb) { + if (!atomic && fb && fb != crtc->primary->fb) { radeon_fb = to_radeon_framebuffer(fb); rbo = gem_to_radeon_bo(radeon_fb->obj); r = radeon_bo_reserve(rbo, false); @@ -1672,8 +1852,9 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc) } /* otherwise, pick one of the plls */ if ((rdev->family == CHIP_KAVERI) || - (rdev->family == CHIP_KABINI)) { - /* KB/KV has PPLL1 and PPLL2 */ + (rdev->family == CHIP_KABINI) || + (rdev->family == CHIP_MULLINS)) { + /* KB/KV/ML has PPLL1 and PPLL2 */ pll_in_use = radeon_get_pll_use_mask(crtc); if (!(pll_in_use & (1 << ATOM_PPLL2))) return ATOM_PPLL2; @@ -1726,6 +1907,20 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc) return ATOM_PPLL1; DRM_ERROR("unable to allocate a PPLL\n"); return ATOM_PPLL_INVALID; + } else if (ASIC_IS_DCE41(rdev)) { + /* Don't share PLLs on DCE4.1 chips */ + if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(radeon_crtc->encoder))) { + if (rdev->clock.dp_extclk) + /* skip PPLL programming if using ext clock */ + return ATOM_PPLL_INVALID; + } + pll_in_use = radeon_get_pll_use_mask(crtc); + if (!(pll_in_use & (1 << ATOM_PPLL1))) + return ATOM_PPLL1; + if (!(pll_in_use & (1 << ATOM_PPLL2))) + return ATOM_PPLL2; + DRM_ERROR("unable to allocate a PPLL\n"); + return ATOM_PPLL_INVALID; } else if (ASIC_IS_DCE4(rdev)) { /* in DP mode, the DP ref clock can come from PPLL, DCPLL, or ext clock, * depending on the asic: @@ -1823,6 +2018,9 @@ int atombios_crtc_mode_set(struct drm_crtc *crtc, (ATOM_DEVICE_TV_SUPPORT | ATOM_DEVICE_CV_SUPPORT)) is_tvcv = true; + if (!radeon_crtc->adjusted_clock) + return -EINVAL; + atombios_crtc_set_pll(crtc, adjusted_mode); if (ASIC_IS_DCE4(rdev)) @@ -1910,6 +2108,27 @@ static void atombios_crtc_disable(struct drm_crtc *crtc) int i; atombios_crtc_dpms(crtc, DRM_MODE_DPMS_OFF); + if (crtc->primary->fb) { + int r; + struct radeon_framebuffer *radeon_fb; + struct radeon_bo *rbo; + + radeon_fb = to_radeon_framebuffer(crtc->primary->fb); + rbo = gem_to_radeon_bo(radeon_fb->obj); + r = radeon_bo_reserve(rbo, false); + if (unlikely(r)) + DRM_ERROR("failed to reserve rbo before unpin\n"); + else { + radeon_bo_unpin(rbo); + radeon_bo_unreserve(rbo); + } + } + /* disable the GRPH */ + if (ASIC_IS_DCE4(rdev)) + WREG32(EVERGREEN_GRPH_ENABLE + radeon_crtc->crtc_offset, 0); + else if (ASIC_IS_AVIVO(rdev)) + WREG32(AVIVO_D1GRPH_ENABLE + radeon_crtc->crtc_offset, 0); + if (ASIC_IS_DCE6(rdev)) atombios_powergate_crtc(crtc, ATOM_ENABLE); @@ -1934,7 +2153,9 @@ static void atombios_crtc_disable(struct drm_crtc *crtc) break; case ATOM_PPLL0: /* disable the ppll */ - if ((rdev->family == CHIP_ARUBA) || (rdev->family == CHIP_BONAIRE)) + if ((rdev->family == CHIP_ARUBA) || + (rdev->family == CHIP_BONAIRE) || + (rdev->family == CHIP_HAWAII)) atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id, 0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss); break; diff --git a/drivers/gpu/drm/radeon/atombios_dp.c b/drivers/gpu/drm/radeon/atombios_dp.c index 064023bed48..b1e11f8434e 100644 --- a/drivers/gpu/drm/radeon/atombios_dp.c +++ b/drivers/gpu/drm/radeon/atombios_dp.c @@ -44,6 +44,41 @@ static char *pre_emph_names[] = { }; /***** radeon AUX functions *****/ + +/* Atom needs data in little endian format + * so swap as appropriate when copying data to + * or from atom. Note that atom operates on + * dw units. + */ +void radeon_atom_copy_swap(u8 *dst, u8 *src, u8 num_bytes, bool to_le) +{ +#ifdef __BIG_ENDIAN + u8 src_tmp[20], dst_tmp[20]; /* used for byteswapping */ + u32 *dst32, *src32; + int i; + + memcpy(src_tmp, src, num_bytes); + src32 = (u32 *)src_tmp; + dst32 = (u32 *)dst_tmp; + if (to_le) { + for (i = 0; i < ((num_bytes + 3) / 4); i++) + dst32[i] = cpu_to_le32(src32[i]); + memcpy(dst, dst_tmp, num_bytes); + } else { + u8 dws = num_bytes & ~3; + for (i = 0; i < ((num_bytes + 3) / 4); i++) + dst32[i] = le32_to_cpu(src32[i]); + memcpy(dst, dst_tmp, dws); + if (num_bytes % 4) { + for (i = 0; i < (num_bytes % 4); i++) + dst[dws+i] = dst_tmp[dws+i]; + } + } +#else + memcpy(dst, src, num_bytes); +#endif +} + union aux_channel_transaction { PROCESS_AUX_CHANNEL_TRANSACTION_PS_ALLOCATION v1; PROCESS_AUX_CHANNEL_TRANSACTION_PARAMETERS_V2 v2; @@ -60,15 +95,18 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan, int index = GetIndexIntoMasterTable(COMMAND, ProcessAuxChannelTransaction); unsigned char *base; int recv_bytes; + int r = 0; memset(&args, 0, sizeof(args)); + mutex_lock(&chan->mutex); + base = (unsigned char *)(rdev->mode_info.atom_context->scratch + 1); - memcpy(base, send, send_bytes); + radeon_atom_copy_swap(base, send, send_bytes, true); - args.v1.lpAuxRequest = 0 + 4; - args.v1.lpDataOut = 16 + 4; + args.v1.lpAuxRequest = cpu_to_le16((u16)(0 + 4)); + args.v1.lpDataOut = cpu_to_le16((u16)(16 + 4)); args.v1.ucDataOutLen = 0; args.v1.ucChannelID = chan->rec.i2c_id; args.v1.ucDelay = delay / 10; @@ -82,19 +120,22 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan, /* timeout */ if (args.v1.ucReplyStatus == 1) { DRM_DEBUG_KMS("dp_aux_ch timeout\n"); - return -ETIMEDOUT; + r = -ETIMEDOUT; + goto done; } /* flags not zero */ if (args.v1.ucReplyStatus == 2) { DRM_DEBUG_KMS("dp_aux_ch flags not zero\n"); - return -EBUSY; + r = -EIO; + goto done; } /* error */ if (args.v1.ucReplyStatus == 3) { DRM_DEBUG_KMS("dp_aux_ch error\n"); - return -EIO; + r = -EIO; + goto done; } recv_bytes = args.v1.ucDataOutLen; @@ -102,189 +143,91 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan, recv_bytes = recv_size; if (recv && recv_size) - memcpy(recv, base + 16, recv_bytes); - - return recv_bytes; -} - -static int radeon_dp_aux_native_write(struct radeon_connector *radeon_connector, - u16 address, u8 *send, u8 send_bytes, u8 delay) -{ - struct radeon_connector_atom_dig *dig_connector = radeon_connector->con_priv; - int ret; - u8 msg[20]; - int msg_bytes = send_bytes + 4; - u8 ack; - unsigned retry; + radeon_atom_copy_swap(recv, base + 16, recv_bytes, false); - if (send_bytes > 16) - return -1; - - msg[0] = address; - msg[1] = address >> 8; - msg[2] = AUX_NATIVE_WRITE << 4; - msg[3] = (msg_bytes << 4) | (send_bytes - 1); - memcpy(&msg[4], send, send_bytes); - - for (retry = 0; retry < 4; retry++) { - ret = radeon_process_aux_ch(dig_connector->dp_i2c_bus, - msg, msg_bytes, NULL, 0, delay, &ack); - if (ret == -EBUSY) - continue; - else if (ret < 0) - return ret; - if ((ack & AUX_NATIVE_REPLY_MASK) == AUX_NATIVE_REPLY_ACK) - return send_bytes; - else if ((ack & AUX_NATIVE_REPLY_MASK) == AUX_NATIVE_REPLY_DEFER) - udelay(400); - else - return -EIO; - } + r = recv_bytes; +done: + mutex_unlock(&chan->mutex); - return -EIO; + return r; } -static int radeon_dp_aux_native_read(struct radeon_connector *radeon_connector, - u16 address, u8 *recv, int recv_bytes, u8 delay) +#define BARE_ADDRESS_SIZE 3 +#define HEADER_SIZE (BARE_ADDRESS_SIZE + 1) + +static ssize_t +radeon_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) { - struct radeon_connector_atom_dig *dig_connector = radeon_connector->con_priv; - u8 msg[4]; - int msg_bytes = 4; - u8 ack; + struct radeon_i2c_chan *chan = + container_of(aux, struct radeon_i2c_chan, aux); int ret; - unsigned retry; - - msg[0] = address; - msg[1] = address >> 8; - msg[2] = AUX_NATIVE_READ << 4; - msg[3] = (msg_bytes << 4) | (recv_bytes - 1); - - for (retry = 0; retry < 4; retry++) { - ret = radeon_process_aux_ch(dig_connector->dp_i2c_bus, - msg, msg_bytes, recv, recv_bytes, delay, &ack); - if (ret == -EBUSY) - continue; - else if (ret < 0) - return ret; - if ((ack & AUX_NATIVE_REPLY_MASK) == AUX_NATIVE_REPLY_ACK) - return ret; - else if ((ack & AUX_NATIVE_REPLY_MASK) == AUX_NATIVE_REPLY_DEFER) - udelay(400); - else if (ret == 0) - return -EPROTO; + u8 tx_buf[20]; + size_t tx_size; + u8 ack, delay = 0; + + if (WARN_ON(msg->size > 16)) + return -E2BIG; + + tx_buf[0] = msg->address & 0xff; + tx_buf[1] = msg->address >> 8; + tx_buf[2] = msg->request << 4; + tx_buf[3] = msg->size ? (msg->size - 1) : 0; + + switch (msg->request & ~DP_AUX_I2C_MOT) { + case DP_AUX_NATIVE_WRITE: + case DP_AUX_I2C_WRITE: + /* tx_size needs to be 4 even for bare address packets since the atom + * table needs the info in tx_buf[3]. + */ + tx_size = HEADER_SIZE + msg->size; + if (msg->size == 0) + tx_buf[3] |= BARE_ADDRESS_SIZE << 4; + else + tx_buf[3] |= tx_size << 4; + memcpy(tx_buf + HEADER_SIZE, msg->buffer, msg->size); + ret = radeon_process_aux_ch(chan, + tx_buf, tx_size, NULL, 0, delay, &ack); + if (ret >= 0) + /* Return payload size. */ + ret = msg->size; + break; + case DP_AUX_NATIVE_READ: + case DP_AUX_I2C_READ: + /* tx_size needs to be 4 even for bare address packets since the atom + * table needs the info in tx_buf[3]. + */ + tx_size = HEADER_SIZE; + if (msg->size == 0) + tx_buf[3] |= BARE_ADDRESS_SIZE << 4; else - return -EIO; + tx_buf[3] |= tx_size << 4; + ret = radeon_process_aux_ch(chan, + tx_buf, tx_size, msg->buffer, msg->size, delay, &ack); + break; + default: + ret = -EINVAL; + break; } - return -EIO; -} - -static void radeon_write_dpcd_reg(struct radeon_connector *radeon_connector, - u16 reg, u8 val) -{ - radeon_dp_aux_native_write(radeon_connector, reg, &val, 1, 0); -} - -static u8 radeon_read_dpcd_reg(struct radeon_connector *radeon_connector, - u16 reg) -{ - u8 val = 0; - - radeon_dp_aux_native_read(radeon_connector, reg, &val, 1, 0); + if (ret >= 0) + msg->reply = ack >> 4; - return val; + return ret; } -int radeon_dp_i2c_aux_ch(struct i2c_adapter *adapter, int mode, - u8 write_byte, u8 *read_byte) +void radeon_dp_aux_init(struct radeon_connector *radeon_connector) { - struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data; - struct radeon_i2c_chan *auxch = (struct radeon_i2c_chan *)adapter; - u16 address = algo_data->address; - u8 msg[5]; - u8 reply[2]; - unsigned retry; - int msg_bytes; - int reply_bytes = 1; int ret; - u8 ack; - /* Set up the command byte */ - if (mode & MODE_I2C_READ) - msg[2] = AUX_I2C_READ << 4; - else - msg[2] = AUX_I2C_WRITE << 4; - - if (!(mode & MODE_I2C_STOP)) - msg[2] |= AUX_I2C_MOT << 4; - - msg[0] = address; - msg[1] = address >> 8; - - switch (mode) { - case MODE_I2C_WRITE: - msg_bytes = 5; - msg[3] = msg_bytes << 4; - msg[4] = write_byte; - break; - case MODE_I2C_READ: - msg_bytes = 4; - msg[3] = msg_bytes << 4; - break; - default: - msg_bytes = 4; - msg[3] = 3 << 4; - break; - } + radeon_connector->ddc_bus->rec.hpd = radeon_connector->hpd.hpd; + radeon_connector->ddc_bus->aux.dev = radeon_connector->base.kdev; + radeon_connector->ddc_bus->aux.transfer = radeon_dp_aux_transfer; - for (retry = 0; retry < 4; retry++) { - ret = radeon_process_aux_ch(auxch, - msg, msg_bytes, reply, reply_bytes, 0, &ack); - if (ret == -EBUSY) - continue; - else if (ret < 0) { - DRM_DEBUG_KMS("aux_ch failed %d\n", ret); - return ret; - } + ret = drm_dp_aux_register(&radeon_connector->ddc_bus->aux); + if (!ret) + radeon_connector->ddc_bus->has_aux = true; - switch (ack & AUX_NATIVE_REPLY_MASK) { - case AUX_NATIVE_REPLY_ACK: - /* I2C-over-AUX Reply field is only valid - * when paired with AUX ACK. - */ - break; - case AUX_NATIVE_REPLY_NACK: - DRM_DEBUG_KMS("aux_ch native nack\n"); - return -EREMOTEIO; - case AUX_NATIVE_REPLY_DEFER: - DRM_DEBUG_KMS("aux_ch native defer\n"); - udelay(400); - continue; - default: - DRM_ERROR("aux_ch invalid native reply 0x%02x\n", ack); - return -EREMOTEIO; - } - - switch (ack & AUX_I2C_REPLY_MASK) { - case AUX_I2C_REPLY_ACK: - if (mode == MODE_I2C_READ) - *read_byte = reply[0]; - return ret; - case AUX_I2C_REPLY_NACK: - DRM_DEBUG_KMS("aux_i2c nack\n"); - return -EREMOTEIO; - case AUX_I2C_REPLY_DEFER: - DRM_DEBUG_KMS("aux_i2c defer\n"); - udelay(400); - break; - default: - DRM_ERROR("aux_i2c invalid reply 0x%02x\n", ack); - return -EREMOTEIO; - } - } - - DRM_DEBUG_KMS("aux i2c too many retries, giving up\n"); - return -EREMOTEIO; + WARN(ret, "drm_dp_aux_register() failed with error %d\n", ret); } /***** general DP utility functions *****/ @@ -349,6 +292,19 @@ static int dp_get_max_dp_pix_clock(int link_rate, /***** radeon specific DP functions *****/ +static int radeon_dp_get_max_link_rate(struct drm_connector *connector, + u8 dpcd[DP_DPCD_SIZE]) +{ + int max_link_rate; + + if (radeon_connector_is_dp12_capable(connector)) + max_link_rate = min(drm_dp_max_link_rate(dpcd), 540000); + else + max_link_rate = min(drm_dp_max_link_rate(dpcd), 270000); + + return max_link_rate; +} + /* First get the min lane# when low rate is used according to pixel clock * (prefer low rate), second check max lane# supported by DP panel, * if the max lane# < low rate lane# then use max lane# instead. @@ -358,7 +314,7 @@ static int radeon_dp_get_dp_lane_number(struct drm_connector *connector, int pix_clock) { int bpp = convert_bpc_to_bpp(radeon_get_monitor_bpc(connector)); - int max_link_rate = drm_dp_max_link_rate(dpcd); + int max_link_rate = radeon_dp_get_max_link_rate(connector, dpcd); int max_lane_num = drm_dp_max_lane_count(dpcd); int lane_num; int max_dp_pix_clock; @@ -396,7 +352,7 @@ static int radeon_dp_get_dp_link_clock(struct drm_connector *connector, return 540000; } - return drm_dp_max_link_rate(dpcd); + return radeon_dp_get_max_link_rate(connector, dpcd); } static u8 radeon_dp_encoder_service(struct radeon_device *rdev, @@ -419,12 +375,11 @@ static u8 radeon_dp_encoder_service(struct radeon_device *rdev, u8 radeon_dp_getsinktype(struct radeon_connector *radeon_connector) { - struct radeon_connector_atom_dig *dig_connector = radeon_connector->con_priv; struct drm_device *dev = radeon_connector->base.dev; struct radeon_device *rdev = dev->dev_private; return radeon_dp_encoder_service(rdev, ATOM_DP_ACTION_GET_SINK_TYPE, 0, - dig_connector->dp_i2c_bus->rec.i2c_id, 0); + radeon_connector->ddc_bus->rec.i2c_id, 0); } static void radeon_dp_probe_oui(struct radeon_connector *radeon_connector) @@ -435,11 +390,11 @@ static void radeon_dp_probe_oui(struct radeon_connector *radeon_connector) if (!(dig_connector->dpcd[DP_DOWN_STREAM_PORT_COUNT] & DP_OUI_SUPPORT)) return; - if (radeon_dp_aux_native_read(radeon_connector, DP_SINK_OUI, buf, 3, 0)) + if (drm_dp_dpcd_read(&radeon_connector->ddc_bus->aux, DP_SINK_OUI, buf, 3) == 3) DRM_DEBUG_KMS("Sink OUI: %02hx%02hx%02hx\n", buf[0], buf[1], buf[2]); - if (radeon_dp_aux_native_read(radeon_connector, DP_BRANCH_OUI, buf, 3, 0)) + if (drm_dp_dpcd_read(&radeon_connector->ddc_bus->aux, DP_BRANCH_OUI, buf, 3) == 3) DRM_DEBUG_KMS("Branch OUI: %02hx%02hx%02hx\n", buf[0], buf[1], buf[2]); } @@ -448,16 +403,18 @@ bool radeon_dp_getdpcd(struct radeon_connector *radeon_connector) { struct radeon_connector_atom_dig *dig_connector = radeon_connector->con_priv; u8 msg[DP_DPCD_SIZE]; - int ret, i; + int ret; - ret = radeon_dp_aux_native_read(radeon_connector, DP_DPCD_REV, msg, - DP_DPCD_SIZE, 0); + char dpcd_hex_dump[DP_DPCD_SIZE * 3]; + + ret = drm_dp_dpcd_read(&radeon_connector->ddc_bus->aux, DP_DPCD_REV, msg, + DP_DPCD_SIZE); if (ret > 0) { memcpy(dig_connector->dpcd, msg, DP_DPCD_SIZE); - DRM_DEBUG_KMS("DPCD: "); - for (i = 0; i < DP_DPCD_SIZE; i++) - DRM_DEBUG_KMS("%02x ", msg[i]); - DRM_DEBUG_KMS("\n"); + + hex_dump_to_buffer(dig_connector->dpcd, sizeof(dig_connector->dpcd), + 32, 1, dpcd_hex_dump, sizeof(dpcd_hex_dump), false); + DRM_DEBUG_KMS("DPCD: %s\n", dpcd_hex_dump); radeon_dp_probe_oui(radeon_connector); @@ -473,6 +430,7 @@ int radeon_dp_get_panel_mode(struct drm_encoder *encoder, struct drm_device *dev = encoder->dev; struct radeon_device *rdev = dev->dev_private; struct radeon_connector *radeon_connector = to_radeon_connector(connector); + struct radeon_connector_atom_dig *dig_connector; int panel_mode = DP_PANEL_MODE_EXTERNAL_DP_MODE; u16 dp_bridge = radeon_connector_encoder_get_dp_bridge_encoder_id(connector); u8 tmp; @@ -480,21 +438,30 @@ int radeon_dp_get_panel_mode(struct drm_encoder *encoder, if (!ASIC_IS_DCE4(rdev)) return panel_mode; + if (!radeon_connector->con_priv) + return panel_mode; + + dig_connector = radeon_connector->con_priv; + if (dp_bridge != ENCODER_OBJECT_ID_NONE) { /* DP bridge chips */ - tmp = radeon_read_dpcd_reg(radeon_connector, DP_EDP_CONFIGURATION_CAP); - if (tmp & 1) - panel_mode = DP_PANEL_MODE_INTERNAL_DP2_MODE; - else if ((dp_bridge == ENCODER_OBJECT_ID_NUTMEG) || - (dp_bridge == ENCODER_OBJECT_ID_TRAVIS)) - panel_mode = DP_PANEL_MODE_INTERNAL_DP1_MODE; - else - panel_mode = DP_PANEL_MODE_EXTERNAL_DP_MODE; + if (drm_dp_dpcd_readb(&radeon_connector->ddc_bus->aux, + DP_EDP_CONFIGURATION_CAP, &tmp) == 1) { + if (tmp & 1) + panel_mode = DP_PANEL_MODE_INTERNAL_DP2_MODE; + else if ((dp_bridge == ENCODER_OBJECT_ID_NUTMEG) || + (dp_bridge == ENCODER_OBJECT_ID_TRAVIS)) + panel_mode = DP_PANEL_MODE_INTERNAL_DP1_MODE; + else + panel_mode = DP_PANEL_MODE_EXTERNAL_DP_MODE; + } } else if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) { /* eDP */ - tmp = radeon_read_dpcd_reg(radeon_connector, DP_EDP_CONFIGURATION_CAP); - if (tmp & 1) - panel_mode = DP_PANEL_MODE_INTERNAL_DP2_MODE; + if (drm_dp_dpcd_readb(&radeon_connector->ddc_bus->aux, + DP_EDP_CONFIGURATION_CAP, &tmp) == 1) { + if (tmp & 1) + panel_mode = DP_PANEL_MODE_INTERNAL_DP2_MODE; + } } return panel_mode; @@ -540,37 +507,43 @@ int radeon_dp_mode_valid_helper(struct drm_connector *connector, return MODE_OK; } -static bool radeon_dp_get_link_status(struct radeon_connector *radeon_connector, - u8 link_status[DP_LINK_STATUS_SIZE]) -{ - int ret; - ret = radeon_dp_aux_native_read(radeon_connector, DP_LANE0_1_STATUS, - link_status, DP_LINK_STATUS_SIZE, 100); - if (ret <= 0) { - return false; - } - - DRM_DEBUG_KMS("link status %*ph\n", 6, link_status); - return true; -} - bool radeon_dp_needs_link_train(struct radeon_connector *radeon_connector) { u8 link_status[DP_LINK_STATUS_SIZE]; struct radeon_connector_atom_dig *dig = radeon_connector->con_priv; - if (!radeon_dp_get_link_status(radeon_connector, link_status)) + if (drm_dp_dpcd_read_link_status(&radeon_connector->ddc_bus->aux, link_status) + <= 0) return false; if (drm_dp_channel_eq_ok(link_status, dig->dp_lane_count)) return false; return true; } +void radeon_dp_set_rx_power_state(struct drm_connector *connector, + u8 power_state) +{ + struct radeon_connector *radeon_connector = to_radeon_connector(connector); + struct radeon_connector_atom_dig *dig_connector; + + if (!radeon_connector->con_priv) + return; + + dig_connector = radeon_connector->con_priv; + + /* power up/down the sink */ + if (dig_connector->dpcd[0] >= 0x11) { + drm_dp_dpcd_writeb(&radeon_connector->ddc_bus->aux, + DP_SET_POWER, power_state); + usleep_range(1000, 2000); + } +} + + struct radeon_dp_link_train_info { struct radeon_device *rdev; struct drm_encoder *encoder; struct drm_connector *connector; - struct radeon_connector *radeon_connector; int enc_id; int dp_clock; int dp_lane_count; @@ -580,6 +553,7 @@ struct radeon_dp_link_train_info { u8 link_status[DP_LINK_STATUS_SIZE]; u8 tries; bool use_dpencoder; + struct drm_dp_aux *aux; }; static void radeon_dp_update_vs_emph(struct radeon_dp_link_train_info *dp_info) @@ -590,8 +564,8 @@ static void radeon_dp_update_vs_emph(struct radeon_dp_link_train_info *dp_info) 0, dp_info->train_set[0]); /* sets all lanes at once */ /* set the vs/emph on the sink */ - radeon_dp_aux_native_write(dp_info->radeon_connector, DP_TRAINING_LANE0_SET, - dp_info->train_set, dp_info->dp_lane_count, 0); + drm_dp_dpcd_write(dp_info->aux, DP_TRAINING_LANE0_SET, + dp_info->train_set, dp_info->dp_lane_count); } static void radeon_dp_set_tp(struct radeon_dp_link_train_info *dp_info, int tp) @@ -626,7 +600,7 @@ static void radeon_dp_set_tp(struct radeon_dp_link_train_info *dp_info, int tp) } /* enable training pattern on the sink */ - radeon_write_dpcd_reg(dp_info->radeon_connector, DP_TRAINING_PATTERN_SET, tp); + drm_dp_dpcd_writeb(dp_info->aux, DP_TRAINING_PATTERN_SET, tp); } static int radeon_dp_link_train_init(struct radeon_dp_link_train_info *dp_info) @@ -636,33 +610,30 @@ static int radeon_dp_link_train_init(struct radeon_dp_link_train_info *dp_info) u8 tmp; /* power up the sink */ - if (dp_info->dpcd[0] >= 0x11) - radeon_write_dpcd_reg(dp_info->radeon_connector, - DP_SET_POWER, DP_SET_POWER_D0); + radeon_dp_set_rx_power_state(dp_info->connector, DP_SET_POWER_D0); /* possibly enable downspread on the sink */ if (dp_info->dpcd[3] & 0x1) - radeon_write_dpcd_reg(dp_info->radeon_connector, - DP_DOWNSPREAD_CTRL, DP_SPREAD_AMP_0_5); + drm_dp_dpcd_writeb(dp_info->aux, + DP_DOWNSPREAD_CTRL, DP_SPREAD_AMP_0_5); else - radeon_write_dpcd_reg(dp_info->radeon_connector, - DP_DOWNSPREAD_CTRL, 0); + drm_dp_dpcd_writeb(dp_info->aux, + DP_DOWNSPREAD_CTRL, 0); if ((dp_info->connector->connector_type == DRM_MODE_CONNECTOR_eDP) && (dig->panel_mode == DP_PANEL_MODE_INTERNAL_DP2_MODE)) { - radeon_write_dpcd_reg(dp_info->radeon_connector, DP_EDP_CONFIGURATION_SET, 1); + drm_dp_dpcd_writeb(dp_info->aux, DP_EDP_CONFIGURATION_SET, 1); } /* set the lane count on the sink */ tmp = dp_info->dp_lane_count; - if (dp_info->dpcd[DP_DPCD_REV] >= 0x11 && - dp_info->dpcd[DP_MAX_LANE_COUNT] & DP_ENHANCED_FRAME_CAP) + if (drm_dp_enhanced_frame_cap(dp_info->dpcd)) tmp |= DP_LANE_COUNT_ENHANCED_FRAME_EN; - radeon_write_dpcd_reg(dp_info->radeon_connector, DP_LANE_COUNT_SET, tmp); + drm_dp_dpcd_writeb(dp_info->aux, DP_LANE_COUNT_SET, tmp); /* set the link rate on the sink */ tmp = drm_dp_link_rate_to_bw_code(dp_info->dp_clock); - radeon_write_dpcd_reg(dp_info->radeon_connector, DP_LINK_BW_SET, tmp); + drm_dp_dpcd_writeb(dp_info->aux, DP_LINK_BW_SET, tmp); /* start training on the source */ if (ASIC_IS_DCE4(dp_info->rdev) || !dp_info->use_dpencoder) @@ -673,9 +644,9 @@ static int radeon_dp_link_train_init(struct radeon_dp_link_train_info *dp_info) dp_info->dp_clock, dp_info->enc_id, 0); /* disable the training pattern on the sink */ - radeon_write_dpcd_reg(dp_info->radeon_connector, - DP_TRAINING_PATTERN_SET, - DP_TRAINING_PATTERN_DISABLE); + drm_dp_dpcd_writeb(dp_info->aux, + DP_TRAINING_PATTERN_SET, + DP_TRAINING_PATTERN_DISABLE); return 0; } @@ -685,9 +656,9 @@ static int radeon_dp_link_train_finish(struct radeon_dp_link_train_info *dp_info udelay(400); /* disable the training pattern on the sink */ - radeon_write_dpcd_reg(dp_info->radeon_connector, - DP_TRAINING_PATTERN_SET, - DP_TRAINING_PATTERN_DISABLE); + drm_dp_dpcd_writeb(dp_info->aux, + DP_TRAINING_PATTERN_SET, + DP_TRAINING_PATTERN_DISABLE); /* disable the training pattern on the source */ if (ASIC_IS_DCE4(dp_info->rdev) || !dp_info->use_dpencoder) @@ -719,7 +690,8 @@ static int radeon_dp_link_train_cr(struct radeon_dp_link_train_info *dp_info) while (1) { drm_dp_link_train_clock_recovery_delay(dp_info->dpcd); - if (!radeon_dp_get_link_status(dp_info->radeon_connector, dp_info->link_status)) { + if (drm_dp_dpcd_read_link_status(dp_info->aux, + dp_info->link_status) <= 0) { DRM_ERROR("displayport link status failed\n"); break; } @@ -781,7 +753,8 @@ static int radeon_dp_link_train_ce(struct radeon_dp_link_train_info *dp_info) while (1) { drm_dp_link_train_channel_eq_delay(dp_info->dpcd); - if (!radeon_dp_get_link_status(dp_info->radeon_connector, dp_info->link_status)) { + if (drm_dp_dpcd_read_link_status(dp_info->aux, + dp_info->link_status) <= 0) { DRM_ERROR("displayport link status failed\n"); break; } @@ -864,19 +837,23 @@ void radeon_dp_link_train(struct drm_encoder *encoder, else dp_info.enc_id |= ATOM_DP_CONFIG_LINK_A; - tmp = radeon_read_dpcd_reg(radeon_connector, DP_MAX_LANE_COUNT); - if (ASIC_IS_DCE5(rdev) && (tmp & DP_TPS3_SUPPORTED)) - dp_info.tp3_supported = true; - else + if (drm_dp_dpcd_readb(&radeon_connector->ddc_bus->aux, DP_MAX_LANE_COUNT, &tmp) + == 1) { + if (ASIC_IS_DCE5(rdev) && (tmp & DP_TPS3_SUPPORTED)) + dp_info.tp3_supported = true; + else + dp_info.tp3_supported = false; + } else { dp_info.tp3_supported = false; + } memcpy(dp_info.dpcd, dig_connector->dpcd, DP_RECEIVER_CAP_SIZE); dp_info.rdev = rdev; dp_info.encoder = encoder; dp_info.connector = connector; - dp_info.radeon_connector = radeon_connector; dp_info.dp_lane_count = dig_connector->dp_lane_count; dp_info.dp_clock = dig_connector->dp_clock; + dp_info.aux = &radeon_connector->ddc_bus->aux; if (radeon_dp_link_train_init(&dp_info)) goto done; diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c index 092275d53d4..7d68203a373 100644 --- a/drivers/gpu/drm/radeon/atombios_encoders.c +++ b/drivers/gpu/drm/radeon/atombios_encoders.c @@ -183,7 +183,6 @@ void radeon_atom_backlight_init(struct radeon_encoder *radeon_encoder, struct backlight_properties props; struct radeon_backlight_privdata *pdata; struct radeon_encoder_atom_dig *dig; - u8 backlight_level; char bl_name[16]; /* Mac laptops with multiple GPUs use the gmux driver for backlight @@ -213,7 +212,7 @@ void radeon_atom_backlight_init(struct radeon_encoder *radeon_encoder, props.type = BACKLIGHT_RAW; snprintf(bl_name, sizeof(bl_name), "radeon_bl%d", dev->primary->index); - bd = backlight_device_register(bl_name, &drm_connector->kdev, + bd = backlight_device_register(bl_name, drm_connector->kdev, pdata, &radeon_atom_backlight_ops, &props); if (IS_ERR(bd)) { DRM_ERROR("Backlight registration failed\n"); @@ -222,12 +221,17 @@ void radeon_atom_backlight_init(struct radeon_encoder *radeon_encoder, pdata->encoder = radeon_encoder; - backlight_level = radeon_atom_get_backlight_level_from_reg(rdev); - dig = radeon_encoder->enc_priv; dig->bl_dev = bd; bd->props.brightness = radeon_atom_backlight_get_brightness(bd); + /* Set a reasonable default here if the level is 0 otherwise + * fbdev will attempt to turn the backlight on after console + * unblanking and it will try and restore 0 which turns the backlight + * off again. + */ + if (bd->props.brightness == 0) + bd->props.brightness = RADEON_MAX_BL_LEVEL; bd->props.power = FB_BLANK_UNBLANK; backlight_update_status(bd); @@ -464,11 +468,12 @@ atombios_tv_setup(struct drm_encoder *encoder, int action) static u8 radeon_atom_get_bpc(struct drm_encoder *encoder) { - struct drm_connector *connector = radeon_get_connector_for_encoder(encoder); int bpc = 8; - if (connector) - bpc = radeon_get_monitor_bpc(connector); + if (encoder->crtc) { + struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); + bpc = radeon_crtc->bpc; + } switch (bpc) { case 0: @@ -682,8 +687,6 @@ atombios_digital_setup(struct drm_encoder *encoder, int action) int atombios_get_encoder_mode(struct drm_encoder *encoder) { - struct drm_device *dev = encoder->dev; - struct radeon_device *rdev = dev->dev_private; struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); struct drm_connector *connector; struct radeon_connector *radeon_connector; @@ -709,24 +712,37 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) switch (connector->connector_type) { case DRM_MODE_CONNECTOR_DVII: case DRM_MODE_CONNECTOR_HDMIB: /* HDMI-B is basically DL-DVI; analog works fine */ - if (drm_detect_hdmi_monitor(radeon_connector->edid) && - radeon_audio && - !ASIC_IS_DCE6(rdev)) /* remove once we support DCE6 */ - return ATOM_ENCODER_MODE_HDMI; - else if (radeon_connector->use_digital) + if (radeon_audio != 0) { + if (radeon_connector->use_digital && + (radeon_connector->audio == RADEON_AUDIO_ENABLE)) + return ATOM_ENCODER_MODE_HDMI; + else if (drm_detect_hdmi_monitor(radeon_connector->edid) && + (radeon_connector->audio == RADEON_AUDIO_AUTO)) + return ATOM_ENCODER_MODE_HDMI; + else if (radeon_connector->use_digital) + return ATOM_ENCODER_MODE_DVI; + else + return ATOM_ENCODER_MODE_CRT; + } else if (radeon_connector->use_digital) { return ATOM_ENCODER_MODE_DVI; - else + } else { return ATOM_ENCODER_MODE_CRT; + } break; case DRM_MODE_CONNECTOR_DVID: case DRM_MODE_CONNECTOR_HDMIA: default: - if (drm_detect_hdmi_monitor(radeon_connector->edid) && - radeon_audio && - !ASIC_IS_DCE6(rdev)) /* remove once we support DCE6 */ - return ATOM_ENCODER_MODE_HDMI; - else + if (radeon_audio != 0) { + if (radeon_connector->audio == RADEON_AUDIO_ENABLE) + return ATOM_ENCODER_MODE_HDMI; + else if (drm_detect_hdmi_monitor(radeon_connector->edid) && + (radeon_connector->audio == RADEON_AUDIO_AUTO)) + return ATOM_ENCODER_MODE_HDMI; + else + return ATOM_ENCODER_MODE_DVI; + } else { return ATOM_ENCODER_MODE_DVI; + } break; case DRM_MODE_CONNECTOR_LVDS: return ATOM_ENCODER_MODE_LVDS; @@ -734,14 +750,19 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) case DRM_MODE_CONNECTOR_DisplayPort: dig_connector = radeon_connector->con_priv; if ((dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) || - (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) + (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) { return ATOM_ENCODER_MODE_DP; - else if (drm_detect_hdmi_monitor(radeon_connector->edid) && - radeon_audio && - !ASIC_IS_DCE6(rdev)) /* remove once we support DCE6 */ - return ATOM_ENCODER_MODE_HDMI; - else + } else if (radeon_audio != 0) { + if (radeon_connector->audio == RADEON_AUDIO_ENABLE) + return ATOM_ENCODER_MODE_HDMI; + else if (drm_detect_hdmi_monitor(radeon_connector->edid) && + (radeon_connector->audio == RADEON_AUDIO_AUTO)) + return ATOM_ENCODER_MODE_HDMI; + else + return ATOM_ENCODER_MODE_DVI; + } else { return ATOM_ENCODER_MODE_DVI; + } break; case DRM_MODE_CONNECTOR_eDP: return ATOM_ENCODER_MODE_DP; @@ -1297,7 +1318,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t } if (is_dp) args.v5.ucLaneNum = dp_lane_count; - else if (radeon_encoder->pixel_clock > 165000) + else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v5.ucLaneNum = 8; else args.v5.ucLaneNum = 4; @@ -1616,10 +1637,16 @@ radeon_atom_encoder_dpms_dig(struct drm_encoder *encoder, int mode) struct drm_connector *connector = radeon_get_connector_for_encoder(encoder); struct radeon_connector *radeon_connector = NULL; struct radeon_connector_atom_dig *radeon_dig_connector = NULL; + bool travis_quirk = false; if (connector) { radeon_connector = to_radeon_connector(connector); radeon_dig_connector = radeon_connector->con_priv; + if ((radeon_connector_encoder_get_dp_bridge_encoder_id(connector) == + ENCODER_OBJECT_ID_TRAVIS) && + (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) && + !ASIC_IS_DCE5(rdev)) + travis_quirk = true; } switch (mode) { @@ -1640,21 +1667,13 @@ radeon_atom_encoder_dpms_dig(struct drm_encoder *encoder, int mode) atombios_external_encoder_setup(encoder, ext_encoder, EXTERNAL_ENCODER_ACTION_V3_ENCODER_SETUP); } - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0); } else if (ASIC_IS_DCE4(rdev)) { /* setup and enable the encoder */ atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_SETUP, 0); - /* enable the transmitter */ - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0); - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE_OUTPUT, 0, 0); } else { /* setup and enable the encoder and transmitter */ atombios_dig_encoder_setup(encoder, ATOM_ENABLE, 0); atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_SETUP, 0, 0); - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0); - /* some early dce3.2 boards have a bug in their transmitter control table */ - if ((rdev->family != CHIP_RV710) && (rdev->family != CHIP_RV730)) - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE_OUTPUT, 0, 0); } if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(encoder)) && connector) { if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) { @@ -1662,73 +1681,56 @@ radeon_atom_encoder_dpms_dig(struct drm_encoder *encoder, int mode) ATOM_TRANSMITTER_ACTION_POWER_ON); radeon_dig_connector->edp_on = true; } + } + /* enable the transmitter */ + atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0); + if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(encoder)) && connector) { + /* DP_SET_POWER_D0 is set in radeon_dp_link_train */ radeon_dp_link_train(encoder, connector); if (ASIC_IS_DCE4(rdev)) atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_VIDEO_ON, 0); } if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_LCD_BLON, 0, 0); + atombios_dig_transmitter_setup(encoder, + ATOM_TRANSMITTER_ACTION_LCD_BLON, 0, 0); + if (ext_encoder) + atombios_external_encoder_setup(encoder, ext_encoder, ATOM_ENABLE); break; case DRM_MODE_DPMS_STANDBY: case DRM_MODE_DPMS_SUSPEND: case DRM_MODE_DPMS_OFF: - if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE5(rdev)) { - /* disable the transmitter */ - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0); - } else if (ASIC_IS_DCE4(rdev)) { + if (ASIC_IS_DCE4(rdev)) { + if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(encoder)) && connector) + atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_VIDEO_OFF, 0); + } + if (ext_encoder) + atombios_external_encoder_setup(encoder, ext_encoder, ATOM_DISABLE); + if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) + atombios_dig_transmitter_setup(encoder, + ATOM_TRANSMITTER_ACTION_LCD_BLOFF, 0, 0); + + if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(encoder)) && + connector && !travis_quirk) + radeon_dp_set_rx_power_state(connector, DP_SET_POWER_D3); + if (ASIC_IS_DCE4(rdev)) { /* disable the transmitter */ - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE_OUTPUT, 0, 0); - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0); + atombios_dig_transmitter_setup(encoder, + ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0); } else { /* disable the encoder and transmitter */ - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE_OUTPUT, 0, 0); - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0); + atombios_dig_transmitter_setup(encoder, + ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0); atombios_dig_encoder_setup(encoder, ATOM_DISABLE, 0); } if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(encoder)) && connector) { - if (ASIC_IS_DCE4(rdev)) - atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_VIDEO_OFF, 0); + if (travis_quirk) + radeon_dp_set_rx_power_state(connector, DP_SET_POWER_D3); if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) { atombios_set_edp_panel_power(connector, ATOM_TRANSMITTER_ACTION_POWER_OFF); radeon_dig_connector->edp_on = false; } } - if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_LCD_BLOFF, 0, 0); - break; - } -} - -static void -radeon_atom_encoder_dpms_ext(struct drm_encoder *encoder, - struct drm_encoder *ext_encoder, - int mode) -{ - struct drm_device *dev = encoder->dev; - struct radeon_device *rdev = dev->dev_private; - - switch (mode) { - case DRM_MODE_DPMS_ON: - default: - if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev)) { - atombios_external_encoder_setup(encoder, ext_encoder, - EXTERNAL_ENCODER_ACTION_V3_ENABLE_OUTPUT); - atombios_external_encoder_setup(encoder, ext_encoder, - EXTERNAL_ENCODER_ACTION_V3_ENCODER_BLANKING_OFF); - } else - atombios_external_encoder_setup(encoder, ext_encoder, ATOM_ENABLE); - break; - case DRM_MODE_DPMS_STANDBY: - case DRM_MODE_DPMS_SUSPEND: - case DRM_MODE_DPMS_OFF: - if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev)) { - atombios_external_encoder_setup(encoder, ext_encoder, - EXTERNAL_ENCODER_ACTION_V3_ENCODER_BLANKING); - atombios_external_encoder_setup(encoder, ext_encoder, - EXTERNAL_ENCODER_ACTION_V3_DISABLE_OUTPUT); - } else - atombios_external_encoder_setup(encoder, ext_encoder, ATOM_DISABLE); break; } } @@ -1739,7 +1741,6 @@ radeon_atom_encoder_dpms(struct drm_encoder *encoder, int mode) struct drm_device *dev = encoder->dev; struct radeon_device *rdev = dev->dev_private; struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); - struct drm_encoder *ext_encoder = radeon_get_external_encoder(encoder); DRM_DEBUG_KMS("encoder dpms %d to mode %d, devices %08x, active_devices %08x\n", radeon_encoder->encoder_id, mode, radeon_encoder->devices, @@ -1799,9 +1800,6 @@ radeon_atom_encoder_dpms(struct drm_encoder *encoder, int mode) return; } - if (ext_encoder) - radeon_atom_encoder_dpms_ext(encoder, ext_encoder, mode); - radeon_atombios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); } @@ -1890,8 +1888,11 @@ atombios_set_encoder_crtc_source(struct drm_encoder *encoder) args.v2.ucEncodeMode = ATOM_ENCODER_MODE_CRT; else args.v2.ucEncodeMode = atombios_get_encoder_mode(encoder); - } else + } else if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) { + args.v2.ucEncodeMode = ATOM_ENCODER_MODE_LVDS; + } else { args.v2.ucEncodeMode = atombios_get_encoder_mode(encoder); + } switch (radeon_encoder->encoder_id) { case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: @@ -2390,6 +2391,15 @@ static void radeon_atom_encoder_prepare(struct drm_encoder *encoder) /* this is needed for the pll/ss setup to work correctly in some cases */ atombios_set_encoder_crtc_source(encoder); + /* set up the FMT blocks */ + if (ASIC_IS_DCE8(rdev)) + dce8_program_fmt(encoder); + else if (ASIC_IS_DCE4(rdev)) + dce4_program_fmt(encoder); + else if (ASIC_IS_DCE3(rdev)) + dce3_program_fmt(encoder); + else if (ASIC_IS_AVIVO(rdev)) + avivo_program_fmt(encoder); } static void radeon_atom_encoder_commit(struct drm_encoder *encoder) diff --git a/drivers/gpu/drm/radeon/atombios_i2c.c b/drivers/gpu/drm/radeon/atombios_i2c.c index 082338df708..9c570fb15b8 100644 --- a/drivers/gpu/drm/radeon/atombios_i2c.c +++ b/drivers/gpu/drm/radeon/atombios_i2c.c @@ -30,7 +30,7 @@ #define TARGET_HW_I2C_CLOCK 50 /* these are a limitation of ProcessI2cChannelTransaction not the hw */ -#define ATOM_MAX_HW_I2C_WRITE 2 +#define ATOM_MAX_HW_I2C_WRITE 3 #define ATOM_MAX_HW_I2C_READ 255 static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan, @@ -42,28 +42,42 @@ static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan, PROCESS_I2C_CHANNEL_TRANSACTION_PS_ALLOCATION args; int index = GetIndexIntoMasterTable(COMMAND, ProcessI2cChannelTransaction); unsigned char *base; - u16 out; + u16 out = cpu_to_le16(0); + int r = 0; memset(&args, 0, sizeof(args)); + mutex_lock(&chan->mutex); + base = (unsigned char *)rdev->mode_info.atom_context->scratch; if (flags & HW_I2C_WRITE) { if (num > ATOM_MAX_HW_I2C_WRITE) { - DRM_ERROR("hw i2c: tried to write too many bytes (%d vs 2)\n", num); - return -EINVAL; + DRM_ERROR("hw i2c: tried to write too many bytes (%d vs 3)\n", num); + r = -EINVAL; + goto done; } - memcpy(&out, buf, num); + if (buf == NULL) + args.ucRegIndex = 0; + else + args.ucRegIndex = buf[0]; + if (num) + num--; + if (num) + memcpy(&out, &buf[1], num); args.lpI2CDataOut = cpu_to_le16(out); } else { if (num > ATOM_MAX_HW_I2C_READ) { DRM_ERROR("hw i2c: tried to read too many bytes (%d vs 255)\n", num); - return -EINVAL; + r = -EINVAL; + goto done; } + args.ucRegIndex = 0; + args.lpI2CDataOut = 0; } + args.ucFlag = flags; args.ucI2CSpeed = TARGET_HW_I2C_CLOCK; - args.ucRegIndex = 0; args.ucTransBytes = num; args.ucSlaveAddr = slave_addr << 1; args.ucLineNumber = chan->rec.i2c_id; @@ -73,13 +87,17 @@ static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan, /* error */ if (args.ucStatus != HW_ASSISTED_I2C_STATUS_SUCCESS) { DRM_DEBUG_KMS("hw_i2c error\n"); - return -EIO; + r = -EIO; + goto done; } if (!(flags & HW_I2C_WRITE)) - memcpy(buf, base, num); + radeon_atom_copy_swap(buf, base, num, false); + +done: + mutex_unlock(&chan->mutex); - return 0; + return r; } int radeon_atom_hw_i2c_xfer(struct i2c_adapter *i2c_adap, @@ -88,14 +106,14 @@ int radeon_atom_hw_i2c_xfer(struct i2c_adapter *i2c_adap, struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap); struct i2c_msg *p; int i, remaining, current_count, buffer_offset, max_bytes, ret; - u8 buf = 0, flags; + u8 flags; /* check for bus probe */ p = &msgs[0]; if ((num == 1) && (p->len == 0)) { ret = radeon_process_i2c_ch(i2c, p->addr, HW_I2C_WRITE, - &buf, 1); + NULL, 0); if (ret) return ret; else diff --git a/drivers/gpu/drm/radeon/btc_dpm.c b/drivers/gpu/drm/radeon/btc_dpm.c index 0bfd55e0882..f81d7ca134d 100644 --- a/drivers/gpu/drm/radeon/btc_dpm.c +++ b/drivers/gpu/drm/radeon/btc_dpm.c @@ -29,6 +29,7 @@ #include "cypress_dpm.h" #include "btc_dpm.h" #include "atom.h" +#include <linux/seq_file.h> #define MC_CG_ARB_FREQ_F0 0x0a #define MC_CG_ARB_FREQ_F1 0x0b @@ -49,6 +50,7 @@ struct rv7xx_ps *rv770_get_ps(struct radeon_ps *rps); struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev); struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev); +extern int ni_mc_load_microcode(struct radeon_device *rdev); //********* BARTS **************// static const u32 barts_cgcg_cgls_default[] = @@ -1168,6 +1170,23 @@ static const struct radeon_blacklist_clocks btc_blacklist_clocks[] = { 25000, 30000, RADEON_SCLK_UP } }; +void btc_get_max_clock_from_voltage_dependency_table(struct radeon_clock_voltage_dependency_table *table, + u32 *max_clock) +{ + u32 i, clock = 0; + + if ((table == NULL) || (table->count == 0)) { + *max_clock = clock; + return; + } + + for (i = 0; i < table->count; i++) { + if (clock < table->entries[i].clk) + clock = table->entries[i].clk; + } + *max_clock = clock; +} + void btc_apply_voltage_dependency_rules(struct radeon_clock_voltage_dependency_table *table, u32 clock, u16 max_voltage, u16 *voltage) { @@ -1913,7 +1932,7 @@ static int btc_set_mc_special_registers(struct radeon_device *rdev, } j++; - if (j > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE) + if (j >= SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE) return -EINVAL; tmp = RREG32(MC_PMG_CMD_MRS); @@ -1928,7 +1947,7 @@ static int btc_set_mc_special_registers(struct radeon_device *rdev, } j++; - if (j > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE) + if (j >= SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE) return -EINVAL; break; case MC_SEQ_RESERVE_M >> 2: @@ -1942,7 +1961,7 @@ static int btc_set_mc_special_registers(struct radeon_device *rdev, } j++; - if (j > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE) + if (j >= SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE) return -EINVAL; break; default: @@ -2080,6 +2099,7 @@ static void btc_apply_state_adjust_rules(struct radeon_device *rdev, bool disable_mclk_switching; u32 mclk, sclk; u16 vddc, vddci; + u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc; if ((rdev->pm.dpm.new_active_crtc_count > 1) || btc_dpm_vblank_too_short(rdev)) @@ -2121,6 +2141,39 @@ static void btc_apply_state_adjust_rules(struct radeon_device *rdev, ps->low.vddci = max_limits->vddci; } + /* limit clocks to max supported clocks based on voltage dependency tables */ + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk, + &max_sclk_vddc); + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk, + &max_mclk_vddci); + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk, + &max_mclk_vddc); + + if (max_sclk_vddc) { + if (ps->low.sclk > max_sclk_vddc) + ps->low.sclk = max_sclk_vddc; + if (ps->medium.sclk > max_sclk_vddc) + ps->medium.sclk = max_sclk_vddc; + if (ps->high.sclk > max_sclk_vddc) + ps->high.sclk = max_sclk_vddc; + } + if (max_mclk_vddci) { + if (ps->low.mclk > max_mclk_vddci) + ps->low.mclk = max_mclk_vddci; + if (ps->medium.mclk > max_mclk_vddci) + ps->medium.mclk = max_mclk_vddci; + if (ps->high.mclk > max_mclk_vddci) + ps->high.mclk = max_mclk_vddci; + } + if (max_mclk_vddc) { + if (ps->low.mclk > max_mclk_vddc) + ps->low.mclk = max_mclk_vddc; + if (ps->medium.mclk > max_mclk_vddc) + ps->medium.mclk = max_mclk_vddc; + if (ps->high.mclk > max_mclk_vddc) + ps->high.mclk = max_mclk_vddc; + } + /* XXX validate the min clocks required for display */ if (disable_mclk_switching) { @@ -2340,12 +2393,6 @@ int btc_dpm_set_power_state(struct radeon_device *rdev) return ret; } - ret = rv770_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO); - if (ret) { - DRM_ERROR("rv770_dpm_force_performance_level failed\n"); - return ret; - } - return 0; } @@ -2465,21 +2512,6 @@ int btc_dpm_enable(struct radeon_device *rdev) if (eg_pi->ls_clock_gating) btc_ls_clock_gating_enable(rdev, true); - if (rdev->irq.installed && - r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { - PPSMC_Result result; - - ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX); - if (ret) - return ret; - rdev->irq.dpm_thermal = true; - radeon_irq_set(rdev); - result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt); - - if (result != PPSMC_Result_OK) - DRM_DEBUG_KMS("Could not enable thermal interrupts.\n"); - } - rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true); btc_init_stutter_mode(rdev); @@ -2531,7 +2563,11 @@ void btc_dpm_disable(struct radeon_device *rdev) void btc_dpm_setup_asic(struct radeon_device *rdev) { struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev); + int r; + r = ni_mc_load_microcode(rdev); + if (r) + DRM_ERROR("Failed to load MC firmware!\n"); rv770_get_memory_type(rdev); rv740_read_clock_registers(rdev); btc_read_arb_registers(rdev); @@ -2548,9 +2584,6 @@ int btc_dpm_init(struct radeon_device *rdev) { struct rv7xx_power_info *pi; struct evergreen_power_info *eg_pi; - int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info); - u16 data_offset, size; - u8 frev, crev; struct atom_clock_dividers dividers; int ret; @@ -2568,6 +2601,10 @@ int btc_dpm_init(struct radeon_device *rdev) pi->min_vddc_in_table = 0; pi->max_vddc_in_table = 0; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = rv7xx_parse_power_table(rdev); if (ret) return ret; @@ -2633,16 +2670,7 @@ int btc_dpm_init(struct radeon_device *rdev) eg_pi->vddci_control = radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0); - if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size, - &frev, &crev, &data_offset)) { - pi->sclk_ss = true; - pi->mclk_ss = true; - pi->dynamic_ss = true; - } else { - pi->sclk_ss = false; - pi->mclk_ss = false; - pi->dynamic_ss = true; - } + rv770_get_engine_memory_ss(rdev); pi->asi = RV770_ASI_DFLT; pi->pasi = CYPRESS_HASI_DFLT; @@ -2659,8 +2687,7 @@ int btc_dpm_init(struct radeon_device *rdev) pi->dynamic_pcie_gen2 = true; - if (pi->gfx_clock_gating && - (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE)) + if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE) pi->thermal_protection = true; else pi->thermal_protection = false; @@ -2712,6 +2739,12 @@ int btc_dpm_init(struct radeon_device *rdev) else rdev->pm.dpm.dyn_state.sclk_mclk_delta = 10000; + /* make sure dc limits are valid */ + if ((rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.sclk == 0) || + (rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.mclk == 0)) + rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc = + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + return 0; } @@ -2728,6 +2761,37 @@ void btc_dpm_fini(struct radeon_device *rdev) r600_free_extended_power_table(rdev); } +void btc_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, + struct seq_file *m) +{ + struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev); + struct radeon_ps *rps = &eg_pi->current_rps; + struct rv7xx_ps *ps = rv770_get_ps(rps); + struct rv7xx_pl *pl; + u32 current_index = + (RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_PROFILE_INDEX_MASK) >> + CURRENT_PROFILE_INDEX_SHIFT; + + if (current_index > 2) { + seq_printf(m, "invalid dpm profile %d\n", current_index); + } else { + if (current_index == 0) + pl = &ps->low; + else if (current_index == 1) + pl = &ps->medium; + else /* current_index == 2 */ + pl = &ps->high; + seq_printf(m, "uvd vclk: %d dclk: %d\n", rps->vclk, rps->dclk); + if (rdev->family >= CHIP_CEDAR) { + seq_printf(m, "power level %d sclk: %u mclk: %u vddc: %u vddci: %u\n", + current_index, pl->sclk, pl->mclk, pl->vddc, pl->vddci); + } else { + seq_printf(m, "power level %d sclk: %u mclk: %u vddc: %u\n", + current_index, pl->sclk, pl->mclk, pl->vddc); + } + } +} + u32 btc_dpm_get_sclk(struct radeon_device *rdev, bool low) { struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev); diff --git a/drivers/gpu/drm/radeon/btc_dpm.h b/drivers/gpu/drm/radeon/btc_dpm.h index 1a15e0e4195..3b6f12b7760 100644 --- a/drivers/gpu/drm/radeon/btc_dpm.h +++ b/drivers/gpu/drm/radeon/btc_dpm.h @@ -46,6 +46,8 @@ void btc_adjust_clock_combinations(struct radeon_device *rdev, struct rv7xx_pl *pl); void btc_apply_voltage_dependency_rules(struct radeon_clock_voltage_dependency_table *table, u32 clock, u16 max_voltage, u16 *voltage); +void btc_get_max_clock_from_voltage_dependency_table(struct radeon_clock_voltage_dependency_table *table, + u32 *max_clock); void btc_apply_voltage_delta_rules(struct radeon_device *rdev, u16 max_vddc, u16 max_vddci, u16 *vddc, u16 *vddci); diff --git a/drivers/gpu/drm/radeon/btcd.h b/drivers/gpu/drm/radeon/btcd.h index 29e32de7e02..9c65be2d55a 100644 --- a/drivers/gpu/drm/radeon/btcd.h +++ b/drivers/gpu/drm/radeon/btcd.h @@ -44,6 +44,10 @@ # define DYN_SPREAD_SPECTRUM_EN (1 << 23) # define AC_DC_SW (1 << 24) +#define TARGET_AND_CURRENT_PROFILE_INDEX 0x66c +# define CURRENT_PROFILE_INDEX_MASK (0xf << 4) +# define CURRENT_PROFILE_INDEX_SHIFT 4 + #define CG_BIF_REQ_AND_RSP 0x7f4 #define CG_CLIENT_REQ(x) ((x) << 0) #define CG_CLIENT_REQ_MASK (0xff << 0) diff --git a/drivers/gpu/drm/radeon/cayman_blit_shaders.c b/drivers/gpu/drm/radeon/cayman_blit_shaders.c index 19a0114d2e3..98d009e154b 100644 --- a/drivers/gpu/drm/radeon/cayman_blit_shaders.c +++ b/drivers/gpu/drm/radeon/cayman_blit_shaders.c @@ -317,58 +317,4 @@ const u32 cayman_default_state[] = 0x00000010, /* */ }; -const u32 cayman_vs[] = -{ - 0x00000004, - 0x80400400, - 0x0000a03c, - 0x95000688, - 0x00004000, - 0x15000688, - 0x00000000, - 0x88000000, - 0x04000000, - 0x67961001, -#ifdef __BIG_ENDIAN - 0x00020000, -#else - 0x00000000, -#endif - 0x00000000, - 0x04000000, - 0x67961000, -#ifdef __BIG_ENDIAN - 0x00020008, -#else - 0x00000008, -#endif - 0x00000000, -}; - -const u32 cayman_ps[] = -{ - 0x00000004, - 0xa00c0000, - 0x00000008, - 0x80400000, - 0x00000000, - 0x95000688, - 0x00000000, - 0x88000000, - 0x00380400, - 0x00146b10, - 0x00380000, - 0x20146b10, - 0x00380400, - 0x40146b00, - 0x80380000, - 0x60146b00, - 0x00000010, - 0x000d1000, - 0xb0800000, - 0x00000000, -}; - -const u32 cayman_ps_size = ARRAY_SIZE(cayman_ps); -const u32 cayman_vs_size = ARRAY_SIZE(cayman_vs); const u32 cayman_default_size = ARRAY_SIZE(cayman_default_state); diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c new file mode 100644 index 00000000000..584090ac3eb --- /dev/null +++ b/drivers/gpu/drm/radeon/ci_dpm.c @@ -0,0 +1,5327 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include <linux/firmware.h> +#include "drmP.h" +#include "radeon.h" +#include "radeon_ucode.h" +#include "cikd.h" +#include "r600_dpm.h" +#include "ci_dpm.h" +#include "atom.h" +#include <linux/seq_file.h> + +#define MC_CG_ARB_FREQ_F0 0x0a +#define MC_CG_ARB_FREQ_F1 0x0b +#define MC_CG_ARB_FREQ_F2 0x0c +#define MC_CG_ARB_FREQ_F3 0x0d + +#define SMC_RAM_END 0x40000 + +#define VOLTAGE_SCALE 4 +#define VOLTAGE_VID_OFFSET_SCALE1 625 +#define VOLTAGE_VID_OFFSET_SCALE2 100 + +static const struct ci_pt_defaults defaults_hawaii_xt = +{ + 1, 0xF, 0xFD, 0x19, 5, 0x14, 0, 0xB0000, + { 0x84, 0x0, 0x0, 0x7F, 0x0, 0x0, 0x5A, 0x60, 0x51, 0x8E, 0x79, 0x6B, 0x5F, 0x90, 0x79 }, + { 0x1EA, 0x1EA, 0x1EA, 0x224, 0x224, 0x224, 0x24F, 0x24F, 0x24F, 0x28E, 0x28E, 0x28E, 0x2BC, 0x2BC, 0x2BC } +}; + +static const struct ci_pt_defaults defaults_hawaii_pro = +{ + 1, 0xF, 0xFD, 0x19, 5, 0x14, 0, 0x65062, + { 0x93, 0x0, 0x0, 0x97, 0x0, 0x0, 0x6B, 0x60, 0x51, 0x95, 0x79, 0x6B, 0x5F, 0x90, 0x79 }, + { 0x1EA, 0x1EA, 0x1EA, 0x224, 0x224, 0x224, 0x24F, 0x24F, 0x24F, 0x28E, 0x28E, 0x28E, 0x2BC, 0x2BC, 0x2BC } +}; + +static const struct ci_pt_defaults defaults_bonaire_xt = +{ + 1, 0xF, 0xFD, 0x19, 5, 45, 0, 0xB0000, + { 0x79, 0x253, 0x25D, 0xAE, 0x72, 0x80, 0x83, 0x86, 0x6F, 0xC8, 0xC9, 0xC9, 0x2F, 0x4D, 0x61 }, + { 0x17C, 0x172, 0x180, 0x1BC, 0x1B3, 0x1BD, 0x206, 0x200, 0x203, 0x25D, 0x25A, 0x255, 0x2C3, 0x2C5, 0x2B4 } +}; + +static const struct ci_pt_defaults defaults_bonaire_pro = +{ + 1, 0xF, 0xFD, 0x19, 5, 45, 0, 0x65062, + { 0x8C, 0x23F, 0x244, 0xA6, 0x83, 0x85, 0x86, 0x86, 0x83, 0xDB, 0xDB, 0xDA, 0x67, 0x60, 0x5F }, + { 0x187, 0x193, 0x193, 0x1C7, 0x1D1, 0x1D1, 0x210, 0x219, 0x219, 0x266, 0x26C, 0x26C, 0x2C9, 0x2CB, 0x2CB } +}; + +static const struct ci_pt_defaults defaults_saturn_xt = +{ + 1, 0xF, 0xFD, 0x19, 5, 55, 0, 0x70000, + { 0x8C, 0x247, 0x249, 0xA6, 0x80, 0x81, 0x8B, 0x89, 0x86, 0xC9, 0xCA, 0xC9, 0x4D, 0x4D, 0x4D }, + { 0x187, 0x187, 0x187, 0x1C7, 0x1C7, 0x1C7, 0x210, 0x210, 0x210, 0x266, 0x266, 0x266, 0x2C9, 0x2C9, 0x2C9 } +}; + +static const struct ci_pt_defaults defaults_saturn_pro = +{ + 1, 0xF, 0xFD, 0x19, 5, 55, 0, 0x30000, + { 0x96, 0x21D, 0x23B, 0xA1, 0x85, 0x87, 0x83, 0x84, 0x81, 0xE6, 0xE6, 0xE6, 0x71, 0x6A, 0x6A }, + { 0x193, 0x19E, 0x19E, 0x1D2, 0x1DC, 0x1DC, 0x21A, 0x223, 0x223, 0x26E, 0x27E, 0x274, 0x2CF, 0x2D2, 0x2D2 } +}; + +static const struct ci_pt_config_reg didt_config_ci[] = +{ + { 0x10, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x10, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x10, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x10, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x11, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x11, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x11, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x11, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x12, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x12, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x12, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x12, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x2, 0x00003fff, 0, 0x4, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x2, 0x03ff0000, 16, 0x80, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x2, 0x78000000, 27, 0x3, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x1, 0x0000ffff, 0, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x1, 0xffff0000, 16, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x0, 0x00000001, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x30, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x30, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x30, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x30, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x31, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x31, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x31, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x31, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x32, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x32, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x32, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x32, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x22, 0x00003fff, 0, 0x4, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x22, 0x03ff0000, 16, 0x80, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x22, 0x78000000, 27, 0x3, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x21, 0x0000ffff, 0, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x21, 0xffff0000, 16, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x20, 0x00000001, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x50, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x50, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x50, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x50, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x51, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x51, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x51, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x51, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x52, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x52, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x52, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x52, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x42, 0x00003fff, 0, 0x4, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x42, 0x03ff0000, 16, 0x80, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x42, 0x78000000, 27, 0x3, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x41, 0x0000ffff, 0, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x41, 0xffff0000, 16, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x40, 0x00000001, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x70, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x70, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x70, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x70, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x71, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x71, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x71, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x71, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x72, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x72, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x72, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x72, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x62, 0x00003fff, 0, 0x4, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x62, 0x03ff0000, 16, 0x80, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x62, 0x78000000, 27, 0x3, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x61, 0x0000ffff, 0, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x61, 0xffff0000, 16, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, + { 0x60, 0x00000001, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, + { 0xFFFFFFFF } +}; + +extern u8 rv770_get_memory_module_index(struct radeon_device *rdev); +extern void btc_get_max_clock_from_voltage_dependency_table(struct radeon_clock_voltage_dependency_table *table, + u32 *max_clock); +extern int ni_copy_and_switch_arb_sets(struct radeon_device *rdev, + u32 arb_freq_src, u32 arb_freq_dest); +extern u8 si_get_ddr3_mclk_frequency_ratio(u32 memory_clock); +extern u8 si_get_mclk_frequency_ratio(u32 memory_clock, bool strobe_mode); +extern void si_trim_voltage_table_to_fit_state_table(struct radeon_device *rdev, + u32 max_voltage_steps, + struct atom_voltage_table *voltage_table); +extern void cik_enter_rlc_safe_mode(struct radeon_device *rdev); +extern void cik_exit_rlc_safe_mode(struct radeon_device *rdev); +extern int ci_mc_load_microcode(struct radeon_device *rdev); +extern void cik_update_cg(struct radeon_device *rdev, + u32 block, bool enable); + +static int ci_get_std_voltage_value_sidd(struct radeon_device *rdev, + struct atom_voltage_table_entry *voltage_table, + u16 *std_voltage_hi_sidd, u16 *std_voltage_lo_sidd); +static int ci_set_power_limit(struct radeon_device *rdev, u32 n); +static int ci_set_overdrive_target_tdp(struct radeon_device *rdev, + u32 target_tdp); +static int ci_update_uvd_dpm(struct radeon_device *rdev, bool gate); + +static struct ci_power_info *ci_get_pi(struct radeon_device *rdev) +{ + struct ci_power_info *pi = rdev->pm.dpm.priv; + + return pi; +} + +static struct ci_ps *ci_get_ps(struct radeon_ps *rps) +{ + struct ci_ps *ps = rps->ps_priv; + + return ps; +} + +static void ci_initialize_powertune_defaults(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + switch (rdev->pdev->device) { + case 0x6649: + case 0x6650: + case 0x6651: + case 0x6658: + case 0x665C: + case 0x665D: + default: + pi->powertune_defaults = &defaults_bonaire_xt; + break; + case 0x6640: + case 0x6641: + case 0x6646: + case 0x6647: + pi->powertune_defaults = &defaults_saturn_xt; + break; + case 0x67B8: + case 0x67B0: + pi->powertune_defaults = &defaults_hawaii_xt; + break; + case 0x67BA: + case 0x67B1: + pi->powertune_defaults = &defaults_hawaii_pro; + break; + case 0x67A0: + case 0x67A1: + case 0x67A2: + case 0x67A8: + case 0x67A9: + case 0x67AA: + case 0x67B9: + case 0x67BE: + pi->powertune_defaults = &defaults_bonaire_xt; + break; + } + + pi->dte_tj_offset = 0; + + pi->caps_power_containment = true; + pi->caps_cac = false; + pi->caps_sq_ramping = false; + pi->caps_db_ramping = false; + pi->caps_td_ramping = false; + pi->caps_tcp_ramping = false; + + if (pi->caps_power_containment) { + pi->caps_cac = true; + pi->enable_bapm_feature = true; + pi->enable_tdc_limit_feature = true; + pi->enable_pkg_pwr_tracking_feature = true; + } +} + +static u8 ci_convert_to_vid(u16 vddc) +{ + return (6200 - (vddc * VOLTAGE_SCALE)) / 25; +} + +static int ci_populate_bapm_vddc_vid_sidd(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u8 *hi_vid = pi->smc_powertune_table.BapmVddCVidHiSidd; + u8 *lo_vid = pi->smc_powertune_table.BapmVddCVidLoSidd; + u8 *hi2_vid = pi->smc_powertune_table.BapmVddCVidHiSidd2; + u32 i; + + if (rdev->pm.dpm.dyn_state.cac_leakage_table.entries == NULL) + return -EINVAL; + if (rdev->pm.dpm.dyn_state.cac_leakage_table.count > 8) + return -EINVAL; + if (rdev->pm.dpm.dyn_state.cac_leakage_table.count != + rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count) + return -EINVAL; + + for (i = 0; i < rdev->pm.dpm.dyn_state.cac_leakage_table.count; i++) { + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_EVV) { + lo_vid[i] = ci_convert_to_vid(rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc1); + hi_vid[i] = ci_convert_to_vid(rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc2); + hi2_vid[i] = ci_convert_to_vid(rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc3); + } else { + lo_vid[i] = ci_convert_to_vid(rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc); + hi_vid[i] = ci_convert_to_vid((u16)rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].leakage); + } + } + return 0; +} + +static int ci_populate_vddc_vid(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u8 *vid = pi->smc_powertune_table.VddCVid; + u32 i; + + if (pi->vddc_voltage_table.count > 8) + return -EINVAL; + + for (i = 0; i < pi->vddc_voltage_table.count; i++) + vid[i] = ci_convert_to_vid(pi->vddc_voltage_table.entries[i].value); + + return 0; +} + +static int ci_populate_svi_load_line(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + const struct ci_pt_defaults *pt_defaults = pi->powertune_defaults; + + pi->smc_powertune_table.SviLoadLineEn = pt_defaults->svi_load_line_en; + pi->smc_powertune_table.SviLoadLineVddC = pt_defaults->svi_load_line_vddc; + pi->smc_powertune_table.SviLoadLineTrimVddC = 3; + pi->smc_powertune_table.SviLoadLineOffsetVddC = 0; + + return 0; +} + +static int ci_populate_tdc_limit(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + const struct ci_pt_defaults *pt_defaults = pi->powertune_defaults; + u16 tdc_limit; + + tdc_limit = rdev->pm.dpm.dyn_state.cac_tdp_table->tdc * 256; + pi->smc_powertune_table.TDC_VDDC_PkgLimit = cpu_to_be16(tdc_limit); + pi->smc_powertune_table.TDC_VDDC_ThrottleReleaseLimitPerc = + pt_defaults->tdc_vddc_throttle_release_limit_perc; + pi->smc_powertune_table.TDC_MAWt = pt_defaults->tdc_mawt; + + return 0; +} + +static int ci_populate_dw8(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + const struct ci_pt_defaults *pt_defaults = pi->powertune_defaults; + int ret; + + ret = ci_read_smc_sram_dword(rdev, + SMU7_FIRMWARE_HEADER_LOCATION + + offsetof(SMU7_Firmware_Header, PmFuseTable) + + offsetof(SMU7_Discrete_PmFuses, TdcWaterfallCtl), + (u32 *)&pi->smc_powertune_table.TdcWaterfallCtl, + pi->sram_end); + if (ret) + return -EINVAL; + else + pi->smc_powertune_table.TdcWaterfallCtl = pt_defaults->tdc_waterfall_ctl; + + return 0; +} + +static int ci_min_max_v_gnbl_pm_lid_from_bapm_vddc(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u8 *hi_vid = pi->smc_powertune_table.BapmVddCVidHiSidd; + u8 *lo_vid = pi->smc_powertune_table.BapmVddCVidLoSidd; + int i, min, max; + + min = max = hi_vid[0]; + for (i = 0; i < 8; i++) { + if (0 != hi_vid[i]) { + if (min > hi_vid[i]) + min = hi_vid[i]; + if (max < hi_vid[i]) + max = hi_vid[i]; + } + + if (0 != lo_vid[i]) { + if (min > lo_vid[i]) + min = lo_vid[i]; + if (max < lo_vid[i]) + max = lo_vid[i]; + } + } + + if ((min == 0) || (max == 0)) + return -EINVAL; + pi->smc_powertune_table.GnbLPMLMaxVid = (u8)max; + pi->smc_powertune_table.GnbLPMLMinVid = (u8)min; + + return 0; +} + +static int ci_populate_bapm_vddc_base_leakage_sidd(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u16 hi_sidd = pi->smc_powertune_table.BapmVddCBaseLeakageHiSidd; + u16 lo_sidd = pi->smc_powertune_table.BapmVddCBaseLeakageLoSidd; + struct radeon_cac_tdp_table *cac_tdp_table = + rdev->pm.dpm.dyn_state.cac_tdp_table; + + hi_sidd = cac_tdp_table->high_cac_leakage / 100 * 256; + lo_sidd = cac_tdp_table->low_cac_leakage / 100 * 256; + + pi->smc_powertune_table.BapmVddCBaseLeakageHiSidd = cpu_to_be16(hi_sidd); + pi->smc_powertune_table.BapmVddCBaseLeakageLoSidd = cpu_to_be16(lo_sidd); + + return 0; +} + +static int ci_populate_bapm_parameters_in_dpm_table(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + const struct ci_pt_defaults *pt_defaults = pi->powertune_defaults; + SMU7_Discrete_DpmTable *dpm_table = &pi->smc_state_table; + struct radeon_cac_tdp_table *cac_tdp_table = + rdev->pm.dpm.dyn_state.cac_tdp_table; + struct radeon_ppm_table *ppm = rdev->pm.dpm.dyn_state.ppm_table; + int i, j, k; + const u16 *def1; + const u16 *def2; + + dpm_table->DefaultTdp = cac_tdp_table->tdp * 256; + dpm_table->TargetTdp = cac_tdp_table->configurable_tdp * 256; + + dpm_table->DTETjOffset = (u8)pi->dte_tj_offset; + dpm_table->GpuTjMax = + (u8)(pi->thermal_temp_setting.temperature_high / 1000); + dpm_table->GpuTjHyst = 8; + + dpm_table->DTEAmbientTempBase = pt_defaults->dte_ambient_temp_base; + + if (ppm) { + dpm_table->PPM_PkgPwrLimit = cpu_to_be16((u16)ppm->dgpu_tdp * 256 / 1000); + dpm_table->PPM_TemperatureLimit = cpu_to_be16((u16)ppm->tj_max * 256); + } else { + dpm_table->PPM_PkgPwrLimit = cpu_to_be16(0); + dpm_table->PPM_TemperatureLimit = cpu_to_be16(0); + } + + dpm_table->BAPM_TEMP_GRADIENT = cpu_to_be32(pt_defaults->bapm_temp_gradient); + def1 = pt_defaults->bapmti_r; + def2 = pt_defaults->bapmti_rc; + + for (i = 0; i < SMU7_DTE_ITERATIONS; i++) { + for (j = 0; j < SMU7_DTE_SOURCES; j++) { + for (k = 0; k < SMU7_DTE_SINKS; k++) { + dpm_table->BAPMTI_R[i][j][k] = cpu_to_be16(*def1); + dpm_table->BAPMTI_RC[i][j][k] = cpu_to_be16(*def2); + def1++; + def2++; + } + } + } + + return 0; +} + +static int ci_populate_pm_base(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 pm_fuse_table_offset; + int ret; + + if (pi->caps_power_containment) { + ret = ci_read_smc_sram_dword(rdev, + SMU7_FIRMWARE_HEADER_LOCATION + + offsetof(SMU7_Firmware_Header, PmFuseTable), + &pm_fuse_table_offset, pi->sram_end); + if (ret) + return ret; + ret = ci_populate_bapm_vddc_vid_sidd(rdev); + if (ret) + return ret; + ret = ci_populate_vddc_vid(rdev); + if (ret) + return ret; + ret = ci_populate_svi_load_line(rdev); + if (ret) + return ret; + ret = ci_populate_tdc_limit(rdev); + if (ret) + return ret; + ret = ci_populate_dw8(rdev); + if (ret) + return ret; + ret = ci_min_max_v_gnbl_pm_lid_from_bapm_vddc(rdev); + if (ret) + return ret; + ret = ci_populate_bapm_vddc_base_leakage_sidd(rdev); + if (ret) + return ret; + ret = ci_copy_bytes_to_smc(rdev, pm_fuse_table_offset, + (u8 *)&pi->smc_powertune_table, + sizeof(SMU7_Discrete_PmFuses), pi->sram_end); + if (ret) + return ret; + } + + return 0; +} + +static void ci_do_enable_didt(struct radeon_device *rdev, const bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 data; + + if (pi->caps_sq_ramping) { + data = RREG32_DIDT(DIDT_SQ_CTRL0); + if (enable) + data |= DIDT_CTRL_EN; + else + data &= ~DIDT_CTRL_EN; + WREG32_DIDT(DIDT_SQ_CTRL0, data); + } + + if (pi->caps_db_ramping) { + data = RREG32_DIDT(DIDT_DB_CTRL0); + if (enable) + data |= DIDT_CTRL_EN; + else + data &= ~DIDT_CTRL_EN; + WREG32_DIDT(DIDT_DB_CTRL0, data); + } + + if (pi->caps_td_ramping) { + data = RREG32_DIDT(DIDT_TD_CTRL0); + if (enable) + data |= DIDT_CTRL_EN; + else + data &= ~DIDT_CTRL_EN; + WREG32_DIDT(DIDT_TD_CTRL0, data); + } + + if (pi->caps_tcp_ramping) { + data = RREG32_DIDT(DIDT_TCP_CTRL0); + if (enable) + data |= DIDT_CTRL_EN; + else + data &= ~DIDT_CTRL_EN; + WREG32_DIDT(DIDT_TCP_CTRL0, data); + } +} + +static int ci_program_pt_config_registers(struct radeon_device *rdev, + const struct ci_pt_config_reg *cac_config_regs) +{ + const struct ci_pt_config_reg *config_regs = cac_config_regs; + u32 data; + u32 cache = 0; + + if (config_regs == NULL) + return -EINVAL; + + while (config_regs->offset != 0xFFFFFFFF) { + if (config_regs->type == CISLANDS_CONFIGREG_CACHE) { + cache |= ((config_regs->value << config_regs->shift) & config_regs->mask); + } else { + switch (config_regs->type) { + case CISLANDS_CONFIGREG_SMC_IND: + data = RREG32_SMC(config_regs->offset); + break; + case CISLANDS_CONFIGREG_DIDT_IND: + data = RREG32_DIDT(config_regs->offset); + break; + default: + data = RREG32(config_regs->offset << 2); + break; + } + + data &= ~config_regs->mask; + data |= ((config_regs->value << config_regs->shift) & config_regs->mask); + data |= cache; + + switch (config_regs->type) { + case CISLANDS_CONFIGREG_SMC_IND: + WREG32_SMC(config_regs->offset, data); + break; + case CISLANDS_CONFIGREG_DIDT_IND: + WREG32_DIDT(config_regs->offset, data); + break; + default: + WREG32(config_regs->offset << 2, data); + break; + } + cache = 0; + } + config_regs++; + } + return 0; +} + +static int ci_enable_didt(struct radeon_device *rdev, bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + int ret; + + if (pi->caps_sq_ramping || pi->caps_db_ramping || + pi->caps_td_ramping || pi->caps_tcp_ramping) { + cik_enter_rlc_safe_mode(rdev); + + if (enable) { + ret = ci_program_pt_config_registers(rdev, didt_config_ci); + if (ret) { + cik_exit_rlc_safe_mode(rdev); + return ret; + } + } + + ci_do_enable_didt(rdev, enable); + + cik_exit_rlc_safe_mode(rdev); + } + + return 0; +} + +static int ci_enable_power_containment(struct radeon_device *rdev, bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + PPSMC_Result smc_result; + int ret = 0; + + if (enable) { + pi->power_containment_features = 0; + if (pi->caps_power_containment) { + if (pi->enable_bapm_feature) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_EnableDTE); + if (smc_result != PPSMC_Result_OK) + ret = -EINVAL; + else + pi->power_containment_features |= POWERCONTAINMENT_FEATURE_BAPM; + } + + if (pi->enable_tdc_limit_feature) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_TDCLimitEnable); + if (smc_result != PPSMC_Result_OK) + ret = -EINVAL; + else + pi->power_containment_features |= POWERCONTAINMENT_FEATURE_TDCLimit; + } + + if (pi->enable_pkg_pwr_tracking_feature) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_PkgPwrLimitEnable); + if (smc_result != PPSMC_Result_OK) { + ret = -EINVAL; + } else { + struct radeon_cac_tdp_table *cac_tdp_table = + rdev->pm.dpm.dyn_state.cac_tdp_table; + u32 default_pwr_limit = + (u32)(cac_tdp_table->maximum_power_delivery_limit * 256); + + pi->power_containment_features |= POWERCONTAINMENT_FEATURE_PkgPwrLimit; + + ci_set_power_limit(rdev, default_pwr_limit); + } + } + } + } else { + if (pi->caps_power_containment && pi->power_containment_features) { + if (pi->power_containment_features & POWERCONTAINMENT_FEATURE_TDCLimit) + ci_send_msg_to_smc(rdev, PPSMC_MSG_TDCLimitDisable); + + if (pi->power_containment_features & POWERCONTAINMENT_FEATURE_BAPM) + ci_send_msg_to_smc(rdev, PPSMC_MSG_DisableDTE); + + if (pi->power_containment_features & POWERCONTAINMENT_FEATURE_PkgPwrLimit) + ci_send_msg_to_smc(rdev, PPSMC_MSG_PkgPwrLimitDisable); + pi->power_containment_features = 0; + } + } + + return ret; +} + +static int ci_enable_smc_cac(struct radeon_device *rdev, bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + PPSMC_Result smc_result; + int ret = 0; + + if (pi->caps_cac) { + if (enable) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_EnableCac); + if (smc_result != PPSMC_Result_OK) { + ret = -EINVAL; + pi->cac_enabled = false; + } else { + pi->cac_enabled = true; + } + } else if (pi->cac_enabled) { + ci_send_msg_to_smc(rdev, PPSMC_MSG_DisableCac); + pi->cac_enabled = false; + } + } + + return ret; +} + +static int ci_power_control_set_level(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct radeon_cac_tdp_table *cac_tdp_table = + rdev->pm.dpm.dyn_state.cac_tdp_table; + s32 adjust_percent; + s32 target_tdp; + int ret = 0; + bool adjust_polarity = false; /* ??? */ + + if (pi->caps_power_containment && + (pi->power_containment_features & POWERCONTAINMENT_FEATURE_BAPM)) { + adjust_percent = adjust_polarity ? + rdev->pm.dpm.tdp_adjustment : (-1 * rdev->pm.dpm.tdp_adjustment); + target_tdp = ((100 + adjust_percent) * + (s32)cac_tdp_table->configurable_tdp) / 100; + target_tdp *= 256; + + ret = ci_set_overdrive_target_tdp(rdev, (u32)target_tdp); + } + + return ret; +} + +void ci_dpm_powergate_uvd(struct radeon_device *rdev, bool gate) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + if (pi->uvd_power_gated == gate) + return; + + pi->uvd_power_gated = gate; + + ci_update_uvd_dpm(rdev, gate); +} + +bool ci_dpm_vblank_too_short(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 vblank_time = r600_dpm_get_vblank_time(rdev); + u32 switch_limit = pi->mem_gddr5 ? 450 : 300; + + if (vblank_time < switch_limit) + return true; + else + return false; + +} + +static void ci_apply_state_adjust_rules(struct radeon_device *rdev, + struct radeon_ps *rps) +{ + struct ci_ps *ps = ci_get_ps(rps); + struct ci_power_info *pi = ci_get_pi(rdev); + struct radeon_clock_and_voltage_limits *max_limits; + bool disable_mclk_switching; + u32 sclk, mclk; + u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc; + int i; + + if (rps->vce_active) { + rps->evclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].evclk; + rps->ecclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].ecclk; + } else { + rps->evclk = 0; + rps->ecclk = 0; + } + + if ((rdev->pm.dpm.new_active_crtc_count > 1) || + ci_dpm_vblank_too_short(rdev)) + disable_mclk_switching = true; + else + disable_mclk_switching = false; + + if ((rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) == ATOM_PPLIB_CLASSIFICATION_UI_BATTERY) + pi->battery_state = true; + else + pi->battery_state = false; + + if (rdev->pm.dpm.ac_power) + max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + else + max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc; + + if (rdev->pm.dpm.ac_power == false) { + for (i = 0; i < ps->performance_level_count; i++) { + if (ps->performance_levels[i].mclk > max_limits->mclk) + ps->performance_levels[i].mclk = max_limits->mclk; + if (ps->performance_levels[i].sclk > max_limits->sclk) + ps->performance_levels[i].sclk = max_limits->sclk; + } + } + + /* limit clocks to max supported clocks based on voltage dependency tables */ + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk, + &max_sclk_vddc); + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk, + &max_mclk_vddci); + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk, + &max_mclk_vddc); + + for (i = 0; i < ps->performance_level_count; i++) { + if (max_sclk_vddc) { + if (ps->performance_levels[i].sclk > max_sclk_vddc) + ps->performance_levels[i].sclk = max_sclk_vddc; + } + if (max_mclk_vddci) { + if (ps->performance_levels[i].mclk > max_mclk_vddci) + ps->performance_levels[i].mclk = max_mclk_vddci; + } + if (max_mclk_vddc) { + if (ps->performance_levels[i].mclk > max_mclk_vddc) + ps->performance_levels[i].mclk = max_mclk_vddc; + } + } + + /* XXX validate the min clocks required for display */ + + if (disable_mclk_switching) { + mclk = ps->performance_levels[ps->performance_level_count - 1].mclk; + sclk = ps->performance_levels[0].sclk; + } else { + mclk = ps->performance_levels[0].mclk; + sclk = ps->performance_levels[0].sclk; + } + + if (rps->vce_active) { + if (sclk < rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].sclk) + sclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].sclk; + if (mclk < rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].mclk) + mclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].mclk; + } + + ps->performance_levels[0].sclk = sclk; + ps->performance_levels[0].mclk = mclk; + + if (ps->performance_levels[1].sclk < ps->performance_levels[0].sclk) + ps->performance_levels[1].sclk = ps->performance_levels[0].sclk; + + if (disable_mclk_switching) { + if (ps->performance_levels[0].mclk < ps->performance_levels[1].mclk) + ps->performance_levels[0].mclk = ps->performance_levels[1].mclk; + } else { + if (ps->performance_levels[1].mclk < ps->performance_levels[0].mclk) + ps->performance_levels[1].mclk = ps->performance_levels[0].mclk; + } +} + +static int ci_set_thermal_temperature_range(struct radeon_device *rdev, + int min_temp, int max_temp) +{ + int low_temp = 0 * 1000; + int high_temp = 255 * 1000; + u32 tmp; + + if (low_temp < min_temp) + low_temp = min_temp; + if (high_temp > max_temp) + high_temp = max_temp; + if (high_temp < low_temp) { + DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp); + return -EINVAL; + } + + tmp = RREG32_SMC(CG_THERMAL_INT); + tmp &= ~(CI_DIG_THERM_INTH_MASK | CI_DIG_THERM_INTL_MASK); + tmp |= CI_DIG_THERM_INTH(high_temp / 1000) | + CI_DIG_THERM_INTL(low_temp / 1000); + WREG32_SMC(CG_THERMAL_INT, tmp); + +#if 0 + /* XXX: need to figure out how to handle this properly */ + tmp = RREG32_SMC(CG_THERMAL_CTRL); + tmp &= DIG_THERM_DPM_MASK; + tmp |= DIG_THERM_DPM(high_temp / 1000); + WREG32_SMC(CG_THERMAL_CTRL, tmp); +#endif + + return 0; +} + +#if 0 +static int ci_read_smc_soft_register(struct radeon_device *rdev, + u16 reg_offset, u32 *value) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + return ci_read_smc_sram_dword(rdev, + pi->soft_regs_start + reg_offset, + value, pi->sram_end); +} +#endif + +static int ci_write_smc_soft_register(struct radeon_device *rdev, + u16 reg_offset, u32 value) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + return ci_write_smc_sram_dword(rdev, + pi->soft_regs_start + reg_offset, + value, pi->sram_end); +} + +static void ci_init_fps_limits(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + SMU7_Discrete_DpmTable *table = &pi->smc_state_table; + + if (pi->caps_fps) { + u16 tmp; + + tmp = 45; + table->FpsHighT = cpu_to_be16(tmp); + + tmp = 30; + table->FpsLowT = cpu_to_be16(tmp); + } +} + +static int ci_update_sclk_t(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + int ret = 0; + u32 low_sclk_interrupt_t = 0; + + if (pi->caps_sclk_throttle_low_notification) { + low_sclk_interrupt_t = cpu_to_be32(pi->low_sclk_interrupt_t); + + ret = ci_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Discrete_DpmTable, LowSclkInterruptT), + (u8 *)&low_sclk_interrupt_t, + sizeof(u32), pi->sram_end); + + } + + return ret; +} + +static void ci_get_leakage_voltages(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u16 leakage_id, virtual_voltage_id; + u16 vddc, vddci; + int i; + + pi->vddc_leakage.count = 0; + pi->vddci_leakage.count = 0; + + if (radeon_atom_get_leakage_id_from_vbios(rdev, &leakage_id) == 0) { + for (i = 0; i < CISLANDS_MAX_LEAKAGE_COUNT; i++) { + virtual_voltage_id = ATOM_VIRTUAL_VOLTAGE_ID0 + i; + if (radeon_atom_get_leakage_vddc_based_on_leakage_params(rdev, &vddc, &vddci, + virtual_voltage_id, + leakage_id) == 0) { + if (vddc != 0 && vddc != virtual_voltage_id) { + pi->vddc_leakage.actual_voltage[pi->vddc_leakage.count] = vddc; + pi->vddc_leakage.leakage_id[pi->vddc_leakage.count] = virtual_voltage_id; + pi->vddc_leakage.count++; + } + if (vddci != 0 && vddci != virtual_voltage_id) { + pi->vddci_leakage.actual_voltage[pi->vddci_leakage.count] = vddci; + pi->vddci_leakage.leakage_id[pi->vddci_leakage.count] = virtual_voltage_id; + pi->vddci_leakage.count++; + } + } + } + } +} + +static void ci_set_dpm_event_sources(struct radeon_device *rdev, u32 sources) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + bool want_thermal_protection; + enum radeon_dpm_event_src dpm_event_src; + u32 tmp; + + switch (sources) { + case 0: + default: + want_thermal_protection = false; + break; + case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL): + want_thermal_protection = true; + dpm_event_src = RADEON_DPM_EVENT_SRC_DIGITAL; + break; + case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL): + want_thermal_protection = true; + dpm_event_src = RADEON_DPM_EVENT_SRC_EXTERNAL; + break; + case ((1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL) | + (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL)): + want_thermal_protection = true; + dpm_event_src = RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL; + break; + } + + if (want_thermal_protection) { +#if 0 + /* XXX: need to figure out how to handle this properly */ + tmp = RREG32_SMC(CG_THERMAL_CTRL); + tmp &= DPM_EVENT_SRC_MASK; + tmp |= DPM_EVENT_SRC(dpm_event_src); + WREG32_SMC(CG_THERMAL_CTRL, tmp); +#endif + + tmp = RREG32_SMC(GENERAL_PWRMGT); + if (pi->thermal_protection) + tmp &= ~THERMAL_PROTECTION_DIS; + else + tmp |= THERMAL_PROTECTION_DIS; + WREG32_SMC(GENERAL_PWRMGT, tmp); + } else { + tmp = RREG32_SMC(GENERAL_PWRMGT); + tmp |= THERMAL_PROTECTION_DIS; + WREG32_SMC(GENERAL_PWRMGT, tmp); + } +} + +static void ci_enable_auto_throttle_source(struct radeon_device *rdev, + enum radeon_dpm_auto_throttle_src source, + bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + if (enable) { + if (!(pi->active_auto_throttle_sources & (1 << source))) { + pi->active_auto_throttle_sources |= 1 << source; + ci_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources); + } + } else { + if (pi->active_auto_throttle_sources & (1 << source)) { + pi->active_auto_throttle_sources &= ~(1 << source); + ci_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources); + } + } +} + +static void ci_enable_vr_hot_gpio_interrupt(struct radeon_device *rdev) +{ + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REGULATOR_HOT) + ci_send_msg_to_smc(rdev, PPSMC_MSG_EnableVRHotGPIOInterrupt); +} + +static int ci_unfreeze_sclk_mclk_dpm(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + PPSMC_Result smc_result; + + if (!pi->need_update_smu7_dpm_table) + return 0; + + if ((!pi->sclk_dpm_key_disabled) && + (pi->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK | DPMTABLE_UPDATE_SCLK))) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_SCLKDPM_UnfreezeLevel); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + if ((!pi->mclk_dpm_key_disabled) && + (pi->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_MCLKDPM_UnfreezeLevel); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + pi->need_update_smu7_dpm_table = 0; + return 0; +} + +static int ci_enable_sclk_mclk_dpm(struct radeon_device *rdev, bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + PPSMC_Result smc_result; + + if (enable) { + if (!pi->sclk_dpm_key_disabled) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_DPM_Enable); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + if (!pi->mclk_dpm_key_disabled) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_MCLKDPM_Enable); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + + WREG32_P(MC_SEQ_CNTL_3, CAC_EN, ~CAC_EN); + + WREG32_SMC(LCAC_MC0_CNTL, 0x05); + WREG32_SMC(LCAC_MC1_CNTL, 0x05); + WREG32_SMC(LCAC_CPL_CNTL, 0x100005); + + udelay(10); + + WREG32_SMC(LCAC_MC0_CNTL, 0x400005); + WREG32_SMC(LCAC_MC1_CNTL, 0x400005); + WREG32_SMC(LCAC_CPL_CNTL, 0x500005); + } + } else { + if (!pi->sclk_dpm_key_disabled) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_DPM_Disable); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + if (!pi->mclk_dpm_key_disabled) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_MCLKDPM_Disable); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + } + + return 0; +} + +static int ci_start_dpm(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + PPSMC_Result smc_result; + int ret; + u32 tmp; + + tmp = RREG32_SMC(GENERAL_PWRMGT); + tmp |= GLOBAL_PWRMGT_EN; + WREG32_SMC(GENERAL_PWRMGT, tmp); + + tmp = RREG32_SMC(SCLK_PWRMGT_CNTL); + tmp |= DYNAMIC_PM_EN; + WREG32_SMC(SCLK_PWRMGT_CNTL, tmp); + + ci_write_smc_soft_register(rdev, offsetof(SMU7_SoftRegisters, VoltageChangeTimeout), 0x1000); + + WREG32_P(BIF_LNCNT_RESET, 0, ~RESET_LNCNT_EN); + + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_Voltage_Cntl_Enable); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + + ret = ci_enable_sclk_mclk_dpm(rdev, true); + if (ret) + return ret; + + if (!pi->pcie_dpm_key_disabled) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_PCIeDPM_Enable); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + return 0; +} + +static int ci_freeze_sclk_mclk_dpm(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + PPSMC_Result smc_result; + + if (!pi->need_update_smu7_dpm_table) + return 0; + + if ((!pi->sclk_dpm_key_disabled) && + (pi->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK | DPMTABLE_UPDATE_SCLK))) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_SCLKDPM_FreezeLevel); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + if ((!pi->mclk_dpm_key_disabled) && + (pi->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_MCLKDPM_FreezeLevel); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + return 0; +} + +static int ci_stop_dpm(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + PPSMC_Result smc_result; + int ret; + u32 tmp; + + tmp = RREG32_SMC(GENERAL_PWRMGT); + tmp &= ~GLOBAL_PWRMGT_EN; + WREG32_SMC(GENERAL_PWRMGT, tmp); + + tmp = RREG32_SMC(SCLK_PWRMGT_CNTL); + tmp &= ~DYNAMIC_PM_EN; + WREG32_SMC(SCLK_PWRMGT_CNTL, tmp); + + if (!pi->pcie_dpm_key_disabled) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_PCIeDPM_Disable); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + ret = ci_enable_sclk_mclk_dpm(rdev, false); + if (ret) + return ret; + + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_Voltage_Cntl_Disable); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + + return 0; +} + +static void ci_enable_sclk_control(struct radeon_device *rdev, bool enable) +{ + u32 tmp = RREG32_SMC(SCLK_PWRMGT_CNTL); + + if (enable) + tmp &= ~SCLK_PWRMGT_OFF; + else + tmp |= SCLK_PWRMGT_OFF; + WREG32_SMC(SCLK_PWRMGT_CNTL, tmp); +} + +#if 0 +static int ci_notify_hw_of_power_source(struct radeon_device *rdev, + bool ac_power) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct radeon_cac_tdp_table *cac_tdp_table = + rdev->pm.dpm.dyn_state.cac_tdp_table; + u32 power_limit; + + if (ac_power) + power_limit = (u32)(cac_tdp_table->maximum_power_delivery_limit * 256); + else + power_limit = (u32)(cac_tdp_table->battery_power_limit * 256); + + ci_set_power_limit(rdev, power_limit); + + if (pi->caps_automatic_dc_transition) { + if (ac_power) + ci_send_msg_to_smc(rdev, PPSMC_MSG_RunningOnAC); + else + ci_send_msg_to_smc(rdev, PPSMC_MSG_Remove_DC_Clamp); + } + + return 0; +} +#endif + +static PPSMC_Result ci_send_msg_to_smc_with_parameter(struct radeon_device *rdev, + PPSMC_Msg msg, u32 parameter) +{ + WREG32(SMC_MSG_ARG_0, parameter); + return ci_send_msg_to_smc(rdev, msg); +} + +static PPSMC_Result ci_send_msg_to_smc_return_parameter(struct radeon_device *rdev, + PPSMC_Msg msg, u32 *parameter) +{ + PPSMC_Result smc_result; + + smc_result = ci_send_msg_to_smc(rdev, msg); + + if ((smc_result == PPSMC_Result_OK) && parameter) + *parameter = RREG32(SMC_MSG_ARG_0); + + return smc_result; +} + +static int ci_dpm_force_state_sclk(struct radeon_device *rdev, u32 n) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + if (!pi->sclk_dpm_key_disabled) { + PPSMC_Result smc_result = + ci_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_DPM_ForceState, n); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + return 0; +} + +static int ci_dpm_force_state_mclk(struct radeon_device *rdev, u32 n) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + if (!pi->mclk_dpm_key_disabled) { + PPSMC_Result smc_result = + ci_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_MCLKDPM_ForceState, n); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + return 0; +} + +static int ci_dpm_force_state_pcie(struct radeon_device *rdev, u32 n) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + if (!pi->pcie_dpm_key_disabled) { + PPSMC_Result smc_result = + ci_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_PCIeDPM_ForceLevel, n); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + return 0; +} + +static int ci_set_power_limit(struct radeon_device *rdev, u32 n) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + if (pi->power_containment_features & POWERCONTAINMENT_FEATURE_PkgPwrLimit) { + PPSMC_Result smc_result = + ci_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_PkgPwrSetLimit, n); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + + return 0; +} + +static int ci_set_overdrive_target_tdp(struct radeon_device *rdev, + u32 target_tdp) +{ + PPSMC_Result smc_result = + ci_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_OverDriveSetTargetTdp, target_tdp); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + return 0; +} + +static int ci_set_boot_state(struct radeon_device *rdev) +{ + return ci_enable_sclk_mclk_dpm(rdev, false); +} + +static u32 ci_get_average_sclk_freq(struct radeon_device *rdev) +{ + u32 sclk_freq; + PPSMC_Result smc_result = + ci_send_msg_to_smc_return_parameter(rdev, + PPSMC_MSG_API_GetSclkFrequency, + &sclk_freq); + if (smc_result != PPSMC_Result_OK) + sclk_freq = 0; + + return sclk_freq; +} + +static u32 ci_get_average_mclk_freq(struct radeon_device *rdev) +{ + u32 mclk_freq; + PPSMC_Result smc_result = + ci_send_msg_to_smc_return_parameter(rdev, + PPSMC_MSG_API_GetMclkFrequency, + &mclk_freq); + if (smc_result != PPSMC_Result_OK) + mclk_freq = 0; + + return mclk_freq; +} + +static void ci_dpm_start_smc(struct radeon_device *rdev) +{ + int i; + + ci_program_jump_on_start(rdev); + ci_start_smc_clock(rdev); + ci_start_smc(rdev); + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32_SMC(FIRMWARE_FLAGS) & INTERRUPTS_ENABLED) + break; + } +} + +static void ci_dpm_stop_smc(struct radeon_device *rdev) +{ + ci_reset_smc(rdev); + ci_stop_smc_clock(rdev); +} + +static int ci_process_firmware_header(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 tmp; + int ret; + + ret = ci_read_smc_sram_dword(rdev, + SMU7_FIRMWARE_HEADER_LOCATION + + offsetof(SMU7_Firmware_Header, DpmTable), + &tmp, pi->sram_end); + if (ret) + return ret; + + pi->dpm_table_start = tmp; + + ret = ci_read_smc_sram_dword(rdev, + SMU7_FIRMWARE_HEADER_LOCATION + + offsetof(SMU7_Firmware_Header, SoftRegisters), + &tmp, pi->sram_end); + if (ret) + return ret; + + pi->soft_regs_start = tmp; + + ret = ci_read_smc_sram_dword(rdev, + SMU7_FIRMWARE_HEADER_LOCATION + + offsetof(SMU7_Firmware_Header, mcRegisterTable), + &tmp, pi->sram_end); + if (ret) + return ret; + + pi->mc_reg_table_start = tmp; + + ret = ci_read_smc_sram_dword(rdev, + SMU7_FIRMWARE_HEADER_LOCATION + + offsetof(SMU7_Firmware_Header, FanTable), + &tmp, pi->sram_end); + if (ret) + return ret; + + pi->fan_table_start = tmp; + + ret = ci_read_smc_sram_dword(rdev, + SMU7_FIRMWARE_HEADER_LOCATION + + offsetof(SMU7_Firmware_Header, mcArbDramTimingTable), + &tmp, pi->sram_end); + if (ret) + return ret; + + pi->arb_table_start = tmp; + + return 0; +} + +static void ci_read_clock_registers(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + pi->clock_registers.cg_spll_func_cntl = + RREG32_SMC(CG_SPLL_FUNC_CNTL); + pi->clock_registers.cg_spll_func_cntl_2 = + RREG32_SMC(CG_SPLL_FUNC_CNTL_2); + pi->clock_registers.cg_spll_func_cntl_3 = + RREG32_SMC(CG_SPLL_FUNC_CNTL_3); + pi->clock_registers.cg_spll_func_cntl_4 = + RREG32_SMC(CG_SPLL_FUNC_CNTL_4); + pi->clock_registers.cg_spll_spread_spectrum = + RREG32_SMC(CG_SPLL_SPREAD_SPECTRUM); + pi->clock_registers.cg_spll_spread_spectrum_2 = + RREG32_SMC(CG_SPLL_SPREAD_SPECTRUM_2); + pi->clock_registers.dll_cntl = RREG32(DLL_CNTL); + pi->clock_registers.mclk_pwrmgt_cntl = RREG32(MCLK_PWRMGT_CNTL); + pi->clock_registers.mpll_ad_func_cntl = RREG32(MPLL_AD_FUNC_CNTL); + pi->clock_registers.mpll_dq_func_cntl = RREG32(MPLL_DQ_FUNC_CNTL); + pi->clock_registers.mpll_func_cntl = RREG32(MPLL_FUNC_CNTL); + pi->clock_registers.mpll_func_cntl_1 = RREG32(MPLL_FUNC_CNTL_1); + pi->clock_registers.mpll_func_cntl_2 = RREG32(MPLL_FUNC_CNTL_2); + pi->clock_registers.mpll_ss1 = RREG32(MPLL_SS1); + pi->clock_registers.mpll_ss2 = RREG32(MPLL_SS2); +} + +static void ci_init_sclk_t(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + pi->low_sclk_interrupt_t = 0; +} + +static void ci_enable_thermal_protection(struct radeon_device *rdev, + bool enable) +{ + u32 tmp = RREG32_SMC(GENERAL_PWRMGT); + + if (enable) + tmp &= ~THERMAL_PROTECTION_DIS; + else + tmp |= THERMAL_PROTECTION_DIS; + WREG32_SMC(GENERAL_PWRMGT, tmp); +} + +static void ci_enable_acpi_power_management(struct radeon_device *rdev) +{ + u32 tmp = RREG32_SMC(GENERAL_PWRMGT); + + tmp |= STATIC_PM_EN; + + WREG32_SMC(GENERAL_PWRMGT, tmp); +} + +#if 0 +static int ci_enter_ulp_state(struct radeon_device *rdev) +{ + + WREG32(SMC_MESSAGE_0, PPSMC_MSG_SwitchToMinimumPower); + + udelay(25000); + + return 0; +} + +static int ci_exit_ulp_state(struct radeon_device *rdev) +{ + int i; + + WREG32(SMC_MESSAGE_0, PPSMC_MSG_ResumeFromMinimumPower); + + udelay(7000); + + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(SMC_RESP_0) == 1) + break; + udelay(1000); + } + + return 0; +} +#endif + +static int ci_notify_smc_display_change(struct radeon_device *rdev, + bool has_display) +{ + PPSMC_Msg msg = has_display ? PPSMC_MSG_HasDisplay : PPSMC_MSG_NoDisplay; + + return (ci_send_msg_to_smc(rdev, msg) == PPSMC_Result_OK) ? 0 : -EINVAL; +} + +static int ci_enable_ds_master_switch(struct radeon_device *rdev, + bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + if (enable) { + if (pi->caps_sclk_ds) { + if (ci_send_msg_to_smc(rdev, PPSMC_MSG_MASTER_DeepSleep_ON) != PPSMC_Result_OK) + return -EINVAL; + } else { + if (ci_send_msg_to_smc(rdev, PPSMC_MSG_MASTER_DeepSleep_OFF) != PPSMC_Result_OK) + return -EINVAL; + } + } else { + if (pi->caps_sclk_ds) { + if (ci_send_msg_to_smc(rdev, PPSMC_MSG_MASTER_DeepSleep_OFF) != PPSMC_Result_OK) + return -EINVAL; + } + } + + return 0; +} + +static void ci_program_display_gap(struct radeon_device *rdev) +{ + u32 tmp = RREG32_SMC(CG_DISPLAY_GAP_CNTL); + u32 pre_vbi_time_in_us; + u32 frame_time_in_us; + u32 ref_clock = rdev->clock.spll.reference_freq; + u32 refresh_rate = r600_dpm_get_vrefresh(rdev); + u32 vblank_time = r600_dpm_get_vblank_time(rdev); + + tmp &= ~DISP_GAP_MASK; + if (rdev->pm.dpm.new_active_crtc_count > 0) + tmp |= DISP_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM); + else + tmp |= DISP_GAP(R600_PM_DISPLAY_GAP_IGNORE); + WREG32_SMC(CG_DISPLAY_GAP_CNTL, tmp); + + if (refresh_rate == 0) + refresh_rate = 60; + if (vblank_time == 0xffffffff) + vblank_time = 500; + frame_time_in_us = 1000000 / refresh_rate; + pre_vbi_time_in_us = + frame_time_in_us - 200 - vblank_time; + tmp = pre_vbi_time_in_us * (ref_clock / 100); + + WREG32_SMC(CG_DISPLAY_GAP_CNTL2, tmp); + ci_write_smc_soft_register(rdev, offsetof(SMU7_SoftRegisters, PreVBlankGap), 0x64); + ci_write_smc_soft_register(rdev, offsetof(SMU7_SoftRegisters, VBlankTimeout), (frame_time_in_us - pre_vbi_time_in_us)); + + + ci_notify_smc_display_change(rdev, (rdev->pm.dpm.new_active_crtc_count == 1)); + +} + +static void ci_enable_spread_spectrum(struct radeon_device *rdev, bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 tmp; + + if (enable) { + if (pi->caps_sclk_ss_support) { + tmp = RREG32_SMC(GENERAL_PWRMGT); + tmp |= DYN_SPREAD_SPECTRUM_EN; + WREG32_SMC(GENERAL_PWRMGT, tmp); + } + } else { + tmp = RREG32_SMC(CG_SPLL_SPREAD_SPECTRUM); + tmp &= ~SSEN; + WREG32_SMC(CG_SPLL_SPREAD_SPECTRUM, tmp); + + tmp = RREG32_SMC(GENERAL_PWRMGT); + tmp &= ~DYN_SPREAD_SPECTRUM_EN; + WREG32_SMC(GENERAL_PWRMGT, tmp); + } +} + +static void ci_program_sstp(struct radeon_device *rdev) +{ + WREG32_SMC(CG_SSP, (SSTU(R600_SSTU_DFLT) | SST(R600_SST_DFLT))); +} + +static void ci_enable_display_gap(struct radeon_device *rdev) +{ + u32 tmp = RREG32_SMC(CG_DISPLAY_GAP_CNTL); + + tmp &= ~(DISP_GAP_MASK | DISP_GAP_MCHG_MASK); + tmp |= (DISP_GAP(R600_PM_DISPLAY_GAP_IGNORE) | + DISP_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK)); + + WREG32_SMC(CG_DISPLAY_GAP_CNTL, tmp); +} + +static void ci_program_vc(struct radeon_device *rdev) +{ + u32 tmp; + + tmp = RREG32_SMC(SCLK_PWRMGT_CNTL); + tmp &= ~(RESET_SCLK_CNT | RESET_BUSY_CNT); + WREG32_SMC(SCLK_PWRMGT_CNTL, tmp); + + WREG32_SMC(CG_FTV_0, CISLANDS_VRC_DFLT0); + WREG32_SMC(CG_FTV_1, CISLANDS_VRC_DFLT1); + WREG32_SMC(CG_FTV_2, CISLANDS_VRC_DFLT2); + WREG32_SMC(CG_FTV_3, CISLANDS_VRC_DFLT3); + WREG32_SMC(CG_FTV_4, CISLANDS_VRC_DFLT4); + WREG32_SMC(CG_FTV_5, CISLANDS_VRC_DFLT5); + WREG32_SMC(CG_FTV_6, CISLANDS_VRC_DFLT6); + WREG32_SMC(CG_FTV_7, CISLANDS_VRC_DFLT7); +} + +static void ci_clear_vc(struct radeon_device *rdev) +{ + u32 tmp; + + tmp = RREG32_SMC(SCLK_PWRMGT_CNTL); + tmp |= (RESET_SCLK_CNT | RESET_BUSY_CNT); + WREG32_SMC(SCLK_PWRMGT_CNTL, tmp); + + WREG32_SMC(CG_FTV_0, 0); + WREG32_SMC(CG_FTV_1, 0); + WREG32_SMC(CG_FTV_2, 0); + WREG32_SMC(CG_FTV_3, 0); + WREG32_SMC(CG_FTV_4, 0); + WREG32_SMC(CG_FTV_5, 0); + WREG32_SMC(CG_FTV_6, 0); + WREG32_SMC(CG_FTV_7, 0); +} + +static int ci_upload_firmware(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + int i, ret; + + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32_SMC(RCU_UC_EVENTS) & BOOT_SEQ_DONE) + break; + } + WREG32_SMC(SMC_SYSCON_MISC_CNTL, 1); + + ci_stop_smc_clock(rdev); + ci_reset_smc(rdev); + + ret = ci_load_smc_ucode(rdev, pi->sram_end); + + return ret; + +} + +static int ci_get_svi2_voltage_table(struct radeon_device *rdev, + struct radeon_clock_voltage_dependency_table *voltage_dependency_table, + struct atom_voltage_table *voltage_table) +{ + u32 i; + + if (voltage_dependency_table == NULL) + return -EINVAL; + + voltage_table->mask_low = 0; + voltage_table->phase_delay = 0; + + voltage_table->count = voltage_dependency_table->count; + for (i = 0; i < voltage_table->count; i++) { + voltage_table->entries[i].value = voltage_dependency_table->entries[i].v; + voltage_table->entries[i].smio_low = 0; + } + + return 0; +} + +static int ci_construct_voltage_tables(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + int ret; + + if (pi->voltage_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) { + ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDC, + VOLTAGE_OBJ_GPIO_LUT, + &pi->vddc_voltage_table); + if (ret) + return ret; + } else if (pi->voltage_control == CISLANDS_VOLTAGE_CONTROL_BY_SVID2) { + ret = ci_get_svi2_voltage_table(rdev, + &rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk, + &pi->vddc_voltage_table); + if (ret) + return ret; + } + + if (pi->vddc_voltage_table.count > SMU7_MAX_LEVELS_VDDC) + si_trim_voltage_table_to_fit_state_table(rdev, SMU7_MAX_LEVELS_VDDC, + &pi->vddc_voltage_table); + + if (pi->vddci_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) { + ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDCI, + VOLTAGE_OBJ_GPIO_LUT, + &pi->vddci_voltage_table); + if (ret) + return ret; + } else if (pi->vddci_control == CISLANDS_VOLTAGE_CONTROL_BY_SVID2) { + ret = ci_get_svi2_voltage_table(rdev, + &rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk, + &pi->vddci_voltage_table); + if (ret) + return ret; + } + + if (pi->vddci_voltage_table.count > SMU7_MAX_LEVELS_VDDCI) + si_trim_voltage_table_to_fit_state_table(rdev, SMU7_MAX_LEVELS_VDDCI, + &pi->vddci_voltage_table); + + if (pi->mvdd_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) { + ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_MVDDC, + VOLTAGE_OBJ_GPIO_LUT, + &pi->mvdd_voltage_table); + if (ret) + return ret; + } else if (pi->mvdd_control == CISLANDS_VOLTAGE_CONTROL_BY_SVID2) { + ret = ci_get_svi2_voltage_table(rdev, + &rdev->pm.dpm.dyn_state.mvdd_dependency_on_mclk, + &pi->mvdd_voltage_table); + if (ret) + return ret; + } + + if (pi->mvdd_voltage_table.count > SMU7_MAX_LEVELS_MVDD) + si_trim_voltage_table_to_fit_state_table(rdev, SMU7_MAX_LEVELS_MVDD, + &pi->mvdd_voltage_table); + + return 0; +} + +static void ci_populate_smc_voltage_table(struct radeon_device *rdev, + struct atom_voltage_table_entry *voltage_table, + SMU7_Discrete_VoltageLevel *smc_voltage_table) +{ + int ret; + + ret = ci_get_std_voltage_value_sidd(rdev, voltage_table, + &smc_voltage_table->StdVoltageHiSidd, + &smc_voltage_table->StdVoltageLoSidd); + + if (ret) { + smc_voltage_table->StdVoltageHiSidd = voltage_table->value * VOLTAGE_SCALE; + smc_voltage_table->StdVoltageLoSidd = voltage_table->value * VOLTAGE_SCALE; + } + + smc_voltage_table->Voltage = cpu_to_be16(voltage_table->value * VOLTAGE_SCALE); + smc_voltage_table->StdVoltageHiSidd = + cpu_to_be16(smc_voltage_table->StdVoltageHiSidd); + smc_voltage_table->StdVoltageLoSidd = + cpu_to_be16(smc_voltage_table->StdVoltageLoSidd); +} + +static int ci_populate_smc_vddc_table(struct radeon_device *rdev, + SMU7_Discrete_DpmTable *table) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + unsigned int count; + + table->VddcLevelCount = pi->vddc_voltage_table.count; + for (count = 0; count < table->VddcLevelCount; count++) { + ci_populate_smc_voltage_table(rdev, + &pi->vddc_voltage_table.entries[count], + &table->VddcLevel[count]); + + if (pi->voltage_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) + table->VddcLevel[count].Smio |= + pi->vddc_voltage_table.entries[count].smio_low; + else + table->VddcLevel[count].Smio = 0; + } + table->VddcLevelCount = cpu_to_be32(table->VddcLevelCount); + + return 0; +} + +static int ci_populate_smc_vddci_table(struct radeon_device *rdev, + SMU7_Discrete_DpmTable *table) +{ + unsigned int count; + struct ci_power_info *pi = ci_get_pi(rdev); + + table->VddciLevelCount = pi->vddci_voltage_table.count; + for (count = 0; count < table->VddciLevelCount; count++) { + ci_populate_smc_voltage_table(rdev, + &pi->vddci_voltage_table.entries[count], + &table->VddciLevel[count]); + + if (pi->vddci_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) + table->VddciLevel[count].Smio |= + pi->vddci_voltage_table.entries[count].smio_low; + else + table->VddciLevel[count].Smio = 0; + } + table->VddciLevelCount = cpu_to_be32(table->VddciLevelCount); + + return 0; +} + +static int ci_populate_smc_mvdd_table(struct radeon_device *rdev, + SMU7_Discrete_DpmTable *table) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + unsigned int count; + + table->MvddLevelCount = pi->mvdd_voltage_table.count; + for (count = 0; count < table->MvddLevelCount; count++) { + ci_populate_smc_voltage_table(rdev, + &pi->mvdd_voltage_table.entries[count], + &table->MvddLevel[count]); + + if (pi->mvdd_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) + table->MvddLevel[count].Smio |= + pi->mvdd_voltage_table.entries[count].smio_low; + else + table->MvddLevel[count].Smio = 0; + } + table->MvddLevelCount = cpu_to_be32(table->MvddLevelCount); + + return 0; +} + +static int ci_populate_smc_voltage_tables(struct radeon_device *rdev, + SMU7_Discrete_DpmTable *table) +{ + int ret; + + ret = ci_populate_smc_vddc_table(rdev, table); + if (ret) + return ret; + + ret = ci_populate_smc_vddci_table(rdev, table); + if (ret) + return ret; + + ret = ci_populate_smc_mvdd_table(rdev, table); + if (ret) + return ret; + + return 0; +} + +static int ci_populate_mvdd_value(struct radeon_device *rdev, u32 mclk, + SMU7_Discrete_VoltageLevel *voltage) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 i = 0; + + if (pi->mvdd_control != CISLANDS_VOLTAGE_CONTROL_NONE) { + for (i = 0; i < rdev->pm.dpm.dyn_state.mvdd_dependency_on_mclk.count; i++) { + if (mclk <= rdev->pm.dpm.dyn_state.mvdd_dependency_on_mclk.entries[i].clk) { + voltage->Voltage = pi->mvdd_voltage_table.entries[i].value; + break; + } + } + + if (i >= rdev->pm.dpm.dyn_state.mvdd_dependency_on_mclk.count) + return -EINVAL; + } + + return -EINVAL; +} + +static int ci_get_std_voltage_value_sidd(struct radeon_device *rdev, + struct atom_voltage_table_entry *voltage_table, + u16 *std_voltage_hi_sidd, u16 *std_voltage_lo_sidd) +{ + u16 v_index, idx; + bool voltage_found = false; + *std_voltage_hi_sidd = voltage_table->value * VOLTAGE_SCALE; + *std_voltage_lo_sidd = voltage_table->value * VOLTAGE_SCALE; + + if (rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries == NULL) + return -EINVAL; + + if (rdev->pm.dpm.dyn_state.cac_leakage_table.entries) { + for (v_index = 0; (u32)v_index < rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count; v_index++) { + if (voltage_table->value == + rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[v_index].v) { + voltage_found = true; + if ((u32)v_index < rdev->pm.dpm.dyn_state.cac_leakage_table.count) + idx = v_index; + else + idx = rdev->pm.dpm.dyn_state.cac_leakage_table.count - 1; + *std_voltage_lo_sidd = + rdev->pm.dpm.dyn_state.cac_leakage_table.entries[idx].vddc * VOLTAGE_SCALE; + *std_voltage_hi_sidd = + rdev->pm.dpm.dyn_state.cac_leakage_table.entries[idx].leakage * VOLTAGE_SCALE; + break; + } + } + + if (!voltage_found) { + for (v_index = 0; (u32)v_index < rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count; v_index++) { + if (voltage_table->value <= + rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[v_index].v) { + voltage_found = true; + if ((u32)v_index < rdev->pm.dpm.dyn_state.cac_leakage_table.count) + idx = v_index; + else + idx = rdev->pm.dpm.dyn_state.cac_leakage_table.count - 1; + *std_voltage_lo_sidd = + rdev->pm.dpm.dyn_state.cac_leakage_table.entries[idx].vddc * VOLTAGE_SCALE; + *std_voltage_hi_sidd = + rdev->pm.dpm.dyn_state.cac_leakage_table.entries[idx].leakage * VOLTAGE_SCALE; + break; + } + } + } + } + + return 0; +} + +static void ci_populate_phase_value_based_on_sclk(struct radeon_device *rdev, + const struct radeon_phase_shedding_limits_table *limits, + u32 sclk, + u32 *phase_shedding) +{ + unsigned int i; + + *phase_shedding = 1; + + for (i = 0; i < limits->count; i++) { + if (sclk < limits->entries[i].sclk) { + *phase_shedding = i; + break; + } + } +} + +static void ci_populate_phase_value_based_on_mclk(struct radeon_device *rdev, + const struct radeon_phase_shedding_limits_table *limits, + u32 mclk, + u32 *phase_shedding) +{ + unsigned int i; + + *phase_shedding = 1; + + for (i = 0; i < limits->count; i++) { + if (mclk < limits->entries[i].mclk) { + *phase_shedding = i; + break; + } + } +} + +static int ci_init_arb_table_index(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 tmp; + int ret; + + ret = ci_read_smc_sram_dword(rdev, pi->arb_table_start, + &tmp, pi->sram_end); + if (ret) + return ret; + + tmp &= 0x00FFFFFF; + tmp |= MC_CG_ARB_FREQ_F1 << 24; + + return ci_write_smc_sram_dword(rdev, pi->arb_table_start, + tmp, pi->sram_end); +} + +static int ci_get_dependency_volt_by_clk(struct radeon_device *rdev, + struct radeon_clock_voltage_dependency_table *allowed_clock_voltage_table, + u32 clock, u32 *voltage) +{ + u32 i = 0; + + if (allowed_clock_voltage_table->count == 0) + return -EINVAL; + + for (i = 0; i < allowed_clock_voltage_table->count; i++) { + if (allowed_clock_voltage_table->entries[i].clk >= clock) { + *voltage = allowed_clock_voltage_table->entries[i].v; + return 0; + } + } + + *voltage = allowed_clock_voltage_table->entries[i-1].v; + + return 0; +} + +static u8 ci_get_sleep_divider_id_from_clock(struct radeon_device *rdev, + u32 sclk, u32 min_sclk_in_sr) +{ + u32 i; + u32 tmp; + u32 min = (min_sclk_in_sr > CISLAND_MINIMUM_ENGINE_CLOCK) ? + min_sclk_in_sr : CISLAND_MINIMUM_ENGINE_CLOCK; + + if (sclk < min) + return 0; + + for (i = CISLAND_MAX_DEEPSLEEP_DIVIDER_ID; ; i--) { + tmp = sclk / (1 << i); + if (tmp >= min || i == 0) + break; + } + + return (u8)i; +} + +static int ci_initial_switch_from_arb_f0_to_f1(struct radeon_device *rdev) +{ + return ni_copy_and_switch_arb_sets(rdev, MC_CG_ARB_FREQ_F0, MC_CG_ARB_FREQ_F1); +} + +static int ci_reset_to_default(struct radeon_device *rdev) +{ + return (ci_send_msg_to_smc(rdev, PPSMC_MSG_ResetToDefaults) == PPSMC_Result_OK) ? + 0 : -EINVAL; +} + +static int ci_force_switch_to_arb_f0(struct radeon_device *rdev) +{ + u32 tmp; + + tmp = (RREG32_SMC(SMC_SCRATCH9) & 0x0000ff00) >> 8; + + if (tmp == MC_CG_ARB_FREQ_F0) + return 0; + + return ni_copy_and_switch_arb_sets(rdev, tmp, MC_CG_ARB_FREQ_F0); +} + +static int ci_populate_memory_timing_parameters(struct radeon_device *rdev, + u32 sclk, + u32 mclk, + SMU7_Discrete_MCArbDramTimingTableEntry *arb_regs) +{ + u32 dram_timing; + u32 dram_timing2; + u32 burst_time; + + radeon_atom_set_engine_dram_timings(rdev, sclk, mclk); + + dram_timing = RREG32(MC_ARB_DRAM_TIMING); + dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2); + burst_time = RREG32(MC_ARB_BURST_TIME) & STATE0_MASK; + + arb_regs->McArbDramTiming = cpu_to_be32(dram_timing); + arb_regs->McArbDramTiming2 = cpu_to_be32(dram_timing2); + arb_regs->McArbBurstTime = (u8)burst_time; + + return 0; +} + +static int ci_do_program_memory_timing_parameters(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + SMU7_Discrete_MCArbDramTimingTable arb_regs; + u32 i, j; + int ret = 0; + + memset(&arb_regs, 0, sizeof(SMU7_Discrete_MCArbDramTimingTable)); + + for (i = 0; i < pi->dpm_table.sclk_table.count; i++) { + for (j = 0; j < pi->dpm_table.mclk_table.count; j++) { + ret = ci_populate_memory_timing_parameters(rdev, + pi->dpm_table.sclk_table.dpm_levels[i].value, + pi->dpm_table.mclk_table.dpm_levels[j].value, + &arb_regs.entries[i][j]); + if (ret) + break; + } + } + + if (ret == 0) + ret = ci_copy_bytes_to_smc(rdev, + pi->arb_table_start, + (u8 *)&arb_regs, + sizeof(SMU7_Discrete_MCArbDramTimingTable), + pi->sram_end); + + return ret; +} + +static int ci_program_memory_timing_parameters(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + if (pi->need_update_smu7_dpm_table == 0) + return 0; + + return ci_do_program_memory_timing_parameters(rdev); +} + +static void ci_populate_smc_initial_state(struct radeon_device *rdev, + struct radeon_ps *radeon_boot_state) +{ + struct ci_ps *boot_state = ci_get_ps(radeon_boot_state); + struct ci_power_info *pi = ci_get_pi(rdev); + u32 level = 0; + + for (level = 0; level < rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count; level++) { + if (rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[level].clk >= + boot_state->performance_levels[0].sclk) { + pi->smc_state_table.GraphicsBootLevel = level; + break; + } + } + + for (level = 0; level < rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.count; level++) { + if (rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries[level].clk >= + boot_state->performance_levels[0].mclk) { + pi->smc_state_table.MemoryBootLevel = level; + break; + } + } +} + +static u32 ci_get_dpm_level_enable_mask_value(struct ci_single_dpm_table *dpm_table) +{ + u32 i; + u32 mask_value = 0; + + for (i = dpm_table->count; i > 0; i--) { + mask_value = mask_value << 1; + if (dpm_table->dpm_levels[i-1].enabled) + mask_value |= 0x1; + else + mask_value &= 0xFFFFFFFE; + } + + return mask_value; +} + +static void ci_populate_smc_link_level(struct radeon_device *rdev, + SMU7_Discrete_DpmTable *table) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_dpm_table *dpm_table = &pi->dpm_table; + u32 i; + + for (i = 0; i < dpm_table->pcie_speed_table.count; i++) { + table->LinkLevel[i].PcieGenSpeed = + (u8)dpm_table->pcie_speed_table.dpm_levels[i].value; + table->LinkLevel[i].PcieLaneCount = + r600_encode_pci_lane_width(dpm_table->pcie_speed_table.dpm_levels[i].param1); + table->LinkLevel[i].EnabledForActivity = 1; + table->LinkLevel[i].DownT = cpu_to_be32(5); + table->LinkLevel[i].UpT = cpu_to_be32(30); + } + + pi->smc_state_table.LinkLevelCount = (u8)dpm_table->pcie_speed_table.count; + pi->dpm_level_enable_mask.pcie_dpm_enable_mask = + ci_get_dpm_level_enable_mask_value(&dpm_table->pcie_speed_table); +} + +static int ci_populate_smc_uvd_level(struct radeon_device *rdev, + SMU7_Discrete_DpmTable *table) +{ + u32 count; + struct atom_clock_dividers dividers; + int ret = -EINVAL; + + table->UvdLevelCount = + rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.count; + + for (count = 0; count < table->UvdLevelCount; count++) { + table->UvdLevel[count].VclkFrequency = + rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[count].vclk; + table->UvdLevel[count].DclkFrequency = + rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[count].dclk; + table->UvdLevel[count].MinVddc = + rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[count].v * VOLTAGE_SCALE; + table->UvdLevel[count].MinVddcPhases = 1; + + ret = radeon_atom_get_clock_dividers(rdev, + COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, + table->UvdLevel[count].VclkFrequency, false, ÷rs); + if (ret) + return ret; + + table->UvdLevel[count].VclkDivider = (u8)dividers.post_divider; + + ret = radeon_atom_get_clock_dividers(rdev, + COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, + table->UvdLevel[count].DclkFrequency, false, ÷rs); + if (ret) + return ret; + + table->UvdLevel[count].DclkDivider = (u8)dividers.post_divider; + + table->UvdLevel[count].VclkFrequency = cpu_to_be32(table->UvdLevel[count].VclkFrequency); + table->UvdLevel[count].DclkFrequency = cpu_to_be32(table->UvdLevel[count].DclkFrequency); + table->UvdLevel[count].MinVddc = cpu_to_be16(table->UvdLevel[count].MinVddc); + } + + return ret; +} + +static int ci_populate_smc_vce_level(struct radeon_device *rdev, + SMU7_Discrete_DpmTable *table) +{ + u32 count; + struct atom_clock_dividers dividers; + int ret = -EINVAL; + + table->VceLevelCount = + rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.count; + + for (count = 0; count < table->VceLevelCount; count++) { + table->VceLevel[count].Frequency = + rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries[count].evclk; + table->VceLevel[count].MinVoltage = + (u16)rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries[count].v * VOLTAGE_SCALE; + table->VceLevel[count].MinPhases = 1; + + ret = radeon_atom_get_clock_dividers(rdev, + COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, + table->VceLevel[count].Frequency, false, ÷rs); + if (ret) + return ret; + + table->VceLevel[count].Divider = (u8)dividers.post_divider; + + table->VceLevel[count].Frequency = cpu_to_be32(table->VceLevel[count].Frequency); + table->VceLevel[count].MinVoltage = cpu_to_be16(table->VceLevel[count].MinVoltage); + } + + return ret; + +} + +static int ci_populate_smc_acp_level(struct radeon_device *rdev, + SMU7_Discrete_DpmTable *table) +{ + u32 count; + struct atom_clock_dividers dividers; + int ret = -EINVAL; + + table->AcpLevelCount = (u8) + (rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.count); + + for (count = 0; count < table->AcpLevelCount; count++) { + table->AcpLevel[count].Frequency = + rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.entries[count].clk; + table->AcpLevel[count].MinVoltage = + rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.entries[count].v; + table->AcpLevel[count].MinPhases = 1; + + ret = radeon_atom_get_clock_dividers(rdev, + COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, + table->AcpLevel[count].Frequency, false, ÷rs); + if (ret) + return ret; + + table->AcpLevel[count].Divider = (u8)dividers.post_divider; + + table->AcpLevel[count].Frequency = cpu_to_be32(table->AcpLevel[count].Frequency); + table->AcpLevel[count].MinVoltage = cpu_to_be16(table->AcpLevel[count].MinVoltage); + } + + return ret; +} + +static int ci_populate_smc_samu_level(struct radeon_device *rdev, + SMU7_Discrete_DpmTable *table) +{ + u32 count; + struct atom_clock_dividers dividers; + int ret = -EINVAL; + + table->SamuLevelCount = + rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.count; + + for (count = 0; count < table->SamuLevelCount; count++) { + table->SamuLevel[count].Frequency = + rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.entries[count].clk; + table->SamuLevel[count].MinVoltage = + rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.entries[count].v * VOLTAGE_SCALE; + table->SamuLevel[count].MinPhases = 1; + + ret = radeon_atom_get_clock_dividers(rdev, + COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, + table->SamuLevel[count].Frequency, false, ÷rs); + if (ret) + return ret; + + table->SamuLevel[count].Divider = (u8)dividers.post_divider; + + table->SamuLevel[count].Frequency = cpu_to_be32(table->SamuLevel[count].Frequency); + table->SamuLevel[count].MinVoltage = cpu_to_be16(table->SamuLevel[count].MinVoltage); + } + + return ret; +} + +static int ci_calculate_mclk_params(struct radeon_device *rdev, + u32 memory_clock, + SMU7_Discrete_MemoryLevel *mclk, + bool strobe_mode, + bool dll_state_on) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 dll_cntl = pi->clock_registers.dll_cntl; + u32 mclk_pwrmgt_cntl = pi->clock_registers.mclk_pwrmgt_cntl; + u32 mpll_ad_func_cntl = pi->clock_registers.mpll_ad_func_cntl; + u32 mpll_dq_func_cntl = pi->clock_registers.mpll_dq_func_cntl; + u32 mpll_func_cntl = pi->clock_registers.mpll_func_cntl; + u32 mpll_func_cntl_1 = pi->clock_registers.mpll_func_cntl_1; + u32 mpll_func_cntl_2 = pi->clock_registers.mpll_func_cntl_2; + u32 mpll_ss1 = pi->clock_registers.mpll_ss1; + u32 mpll_ss2 = pi->clock_registers.mpll_ss2; + struct atom_mpll_param mpll_param; + int ret; + + ret = radeon_atom_get_memory_pll_dividers(rdev, memory_clock, strobe_mode, &mpll_param); + if (ret) + return ret; + + mpll_func_cntl &= ~BWCTRL_MASK; + mpll_func_cntl |= BWCTRL(mpll_param.bwcntl); + + mpll_func_cntl_1 &= ~(CLKF_MASK | CLKFRAC_MASK | VCO_MODE_MASK); + mpll_func_cntl_1 |= CLKF(mpll_param.clkf) | + CLKFRAC(mpll_param.clkfrac) | VCO_MODE(mpll_param.vco_mode); + + mpll_ad_func_cntl &= ~YCLK_POST_DIV_MASK; + mpll_ad_func_cntl |= YCLK_POST_DIV(mpll_param.post_div); + + if (pi->mem_gddr5) { + mpll_dq_func_cntl &= ~(YCLK_SEL_MASK | YCLK_POST_DIV_MASK); + mpll_dq_func_cntl |= YCLK_SEL(mpll_param.yclk_sel) | + YCLK_POST_DIV(mpll_param.post_div); + } + + if (pi->caps_mclk_ss_support) { + struct radeon_atom_ss ss; + u32 freq_nom; + u32 tmp; + u32 reference_clock = rdev->clock.mpll.reference_freq; + + if (pi->mem_gddr5) + freq_nom = memory_clock * 4; + else + freq_nom = memory_clock * 2; + + tmp = (freq_nom / reference_clock); + tmp = tmp * tmp; + if (radeon_atombios_get_asic_ss_info(rdev, &ss, + ASIC_INTERNAL_MEMORY_SS, freq_nom)) { + u32 clks = reference_clock * 5 / ss.rate; + u32 clkv = (u32)((((131 * ss.percentage * ss.rate) / 100) * tmp) / freq_nom); + + mpll_ss1 &= ~CLKV_MASK; + mpll_ss1 |= CLKV(clkv); + + mpll_ss2 &= ~CLKS_MASK; + mpll_ss2 |= CLKS(clks); + } + } + + mclk_pwrmgt_cntl &= ~DLL_SPEED_MASK; + mclk_pwrmgt_cntl |= DLL_SPEED(mpll_param.dll_speed); + + if (dll_state_on) + mclk_pwrmgt_cntl |= MRDCK0_PDNB | MRDCK1_PDNB; + else + mclk_pwrmgt_cntl &= ~(MRDCK0_PDNB | MRDCK1_PDNB); + + mclk->MclkFrequency = memory_clock; + mclk->MpllFuncCntl = mpll_func_cntl; + mclk->MpllFuncCntl_1 = mpll_func_cntl_1; + mclk->MpllFuncCntl_2 = mpll_func_cntl_2; + mclk->MpllAdFuncCntl = mpll_ad_func_cntl; + mclk->MpllDqFuncCntl = mpll_dq_func_cntl; + mclk->MclkPwrmgtCntl = mclk_pwrmgt_cntl; + mclk->DllCntl = dll_cntl; + mclk->MpllSs1 = mpll_ss1; + mclk->MpllSs2 = mpll_ss2; + + return 0; +} + +static int ci_populate_single_memory_level(struct radeon_device *rdev, + u32 memory_clock, + SMU7_Discrete_MemoryLevel *memory_level) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + int ret; + bool dll_state_on; + + if (rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries) { + ret = ci_get_dependency_volt_by_clk(rdev, + &rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk, + memory_clock, &memory_level->MinVddc); + if (ret) + return ret; + } + + if (rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries) { + ret = ci_get_dependency_volt_by_clk(rdev, + &rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk, + memory_clock, &memory_level->MinVddci); + if (ret) + return ret; + } + + if (rdev->pm.dpm.dyn_state.mvdd_dependency_on_mclk.entries) { + ret = ci_get_dependency_volt_by_clk(rdev, + &rdev->pm.dpm.dyn_state.mvdd_dependency_on_mclk, + memory_clock, &memory_level->MinMvdd); + if (ret) + return ret; + } + + memory_level->MinVddcPhases = 1; + + if (pi->vddc_phase_shed_control) + ci_populate_phase_value_based_on_mclk(rdev, + &rdev->pm.dpm.dyn_state.phase_shedding_limits_table, + memory_clock, + &memory_level->MinVddcPhases); + + memory_level->EnabledForThrottle = 1; + memory_level->EnabledForActivity = 1; + memory_level->UpH = 0; + memory_level->DownH = 100; + memory_level->VoltageDownH = 0; + memory_level->ActivityLevel = (u16)pi->mclk_activity_target; + + memory_level->StutterEnable = false; + memory_level->StrobeEnable = false; + memory_level->EdcReadEnable = false; + memory_level->EdcWriteEnable = false; + memory_level->RttEnable = false; + + memory_level->DisplayWatermark = PPSMC_DISPLAY_WATERMARK_LOW; + + if (pi->mclk_stutter_mode_threshold && + (memory_clock <= pi->mclk_stutter_mode_threshold) && + (pi->uvd_enabled == false) && + (RREG32(DPG_PIPE_STUTTER_CONTROL) & STUTTER_ENABLE) && + (rdev->pm.dpm.new_active_crtc_count <= 2)) + memory_level->StutterEnable = true; + + if (pi->mclk_strobe_mode_threshold && + (memory_clock <= pi->mclk_strobe_mode_threshold)) + memory_level->StrobeEnable = 1; + + if (pi->mem_gddr5) { + memory_level->StrobeRatio = + si_get_mclk_frequency_ratio(memory_clock, memory_level->StrobeEnable); + if (pi->mclk_edc_enable_threshold && + (memory_clock > pi->mclk_edc_enable_threshold)) + memory_level->EdcReadEnable = true; + + if (pi->mclk_edc_wr_enable_threshold && + (memory_clock > pi->mclk_edc_wr_enable_threshold)) + memory_level->EdcWriteEnable = true; + + if (memory_level->StrobeEnable) { + if (si_get_mclk_frequency_ratio(memory_clock, true) >= + ((RREG32(MC_SEQ_MISC7) >> 16) & 0xf)) + dll_state_on = ((RREG32(MC_SEQ_MISC5) >> 1) & 0x1) ? true : false; + else + dll_state_on = ((RREG32(MC_SEQ_MISC6) >> 1) & 0x1) ? true : false; + } else { + dll_state_on = pi->dll_default_on; + } + } else { + memory_level->StrobeRatio = si_get_ddr3_mclk_frequency_ratio(memory_clock); + dll_state_on = ((RREG32(MC_SEQ_MISC5) >> 1) & 0x1) ? true : false; + } + + ret = ci_calculate_mclk_params(rdev, memory_clock, memory_level, memory_level->StrobeEnable, dll_state_on); + if (ret) + return ret; + + memory_level->MinVddc = cpu_to_be32(memory_level->MinVddc * VOLTAGE_SCALE); + memory_level->MinVddcPhases = cpu_to_be32(memory_level->MinVddcPhases); + memory_level->MinVddci = cpu_to_be32(memory_level->MinVddci * VOLTAGE_SCALE); + memory_level->MinMvdd = cpu_to_be32(memory_level->MinMvdd * VOLTAGE_SCALE); + + memory_level->MclkFrequency = cpu_to_be32(memory_level->MclkFrequency); + memory_level->ActivityLevel = cpu_to_be16(memory_level->ActivityLevel); + memory_level->MpllFuncCntl = cpu_to_be32(memory_level->MpllFuncCntl); + memory_level->MpllFuncCntl_1 = cpu_to_be32(memory_level->MpllFuncCntl_1); + memory_level->MpllFuncCntl_2 = cpu_to_be32(memory_level->MpllFuncCntl_2); + memory_level->MpllAdFuncCntl = cpu_to_be32(memory_level->MpllAdFuncCntl); + memory_level->MpllDqFuncCntl = cpu_to_be32(memory_level->MpllDqFuncCntl); + memory_level->MclkPwrmgtCntl = cpu_to_be32(memory_level->MclkPwrmgtCntl); + memory_level->DllCntl = cpu_to_be32(memory_level->DllCntl); + memory_level->MpllSs1 = cpu_to_be32(memory_level->MpllSs1); + memory_level->MpllSs2 = cpu_to_be32(memory_level->MpllSs2); + + return 0; +} + +static int ci_populate_smc_acpi_level(struct radeon_device *rdev, + SMU7_Discrete_DpmTable *table) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct atom_clock_dividers dividers; + SMU7_Discrete_VoltageLevel voltage_level; + u32 spll_func_cntl = pi->clock_registers.cg_spll_func_cntl; + u32 spll_func_cntl_2 = pi->clock_registers.cg_spll_func_cntl_2; + u32 dll_cntl = pi->clock_registers.dll_cntl; + u32 mclk_pwrmgt_cntl = pi->clock_registers.mclk_pwrmgt_cntl; + int ret; + + table->ACPILevel.Flags &= ~PPSMC_SWSTATE_FLAG_DC; + + if (pi->acpi_vddc) + table->ACPILevel.MinVddc = cpu_to_be32(pi->acpi_vddc * VOLTAGE_SCALE); + else + table->ACPILevel.MinVddc = cpu_to_be32(pi->min_vddc_in_pp_table * VOLTAGE_SCALE); + + table->ACPILevel.MinVddcPhases = pi->vddc_phase_shed_control ? 0 : 1; + + table->ACPILevel.SclkFrequency = rdev->clock.spll.reference_freq; + + ret = radeon_atom_get_clock_dividers(rdev, + COMPUTE_GPUCLK_INPUT_FLAG_SCLK, + table->ACPILevel.SclkFrequency, false, ÷rs); + if (ret) + return ret; + + table->ACPILevel.SclkDid = (u8)dividers.post_divider; + table->ACPILevel.DisplayWatermark = PPSMC_DISPLAY_WATERMARK_LOW; + table->ACPILevel.DeepSleepDivId = 0; + + spll_func_cntl &= ~SPLL_PWRON; + spll_func_cntl |= SPLL_RESET; + + spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK; + spll_func_cntl_2 |= SCLK_MUX_SEL(4); + + table->ACPILevel.CgSpllFuncCntl = spll_func_cntl; + table->ACPILevel.CgSpllFuncCntl2 = spll_func_cntl_2; + table->ACPILevel.CgSpllFuncCntl3 = pi->clock_registers.cg_spll_func_cntl_3; + table->ACPILevel.CgSpllFuncCntl4 = pi->clock_registers.cg_spll_func_cntl_4; + table->ACPILevel.SpllSpreadSpectrum = pi->clock_registers.cg_spll_spread_spectrum; + table->ACPILevel.SpllSpreadSpectrum2 = pi->clock_registers.cg_spll_spread_spectrum_2; + table->ACPILevel.CcPwrDynRm = 0; + table->ACPILevel.CcPwrDynRm1 = 0; + + table->ACPILevel.Flags = cpu_to_be32(table->ACPILevel.Flags); + table->ACPILevel.MinVddcPhases = cpu_to_be32(table->ACPILevel.MinVddcPhases); + table->ACPILevel.SclkFrequency = cpu_to_be32(table->ACPILevel.SclkFrequency); + table->ACPILevel.CgSpllFuncCntl = cpu_to_be32(table->ACPILevel.CgSpllFuncCntl); + table->ACPILevel.CgSpllFuncCntl2 = cpu_to_be32(table->ACPILevel.CgSpllFuncCntl2); + table->ACPILevel.CgSpllFuncCntl3 = cpu_to_be32(table->ACPILevel.CgSpllFuncCntl3); + table->ACPILevel.CgSpllFuncCntl4 = cpu_to_be32(table->ACPILevel.CgSpllFuncCntl4); + table->ACPILevel.SpllSpreadSpectrum = cpu_to_be32(table->ACPILevel.SpllSpreadSpectrum); + table->ACPILevel.SpllSpreadSpectrum2 = cpu_to_be32(table->ACPILevel.SpllSpreadSpectrum2); + table->ACPILevel.CcPwrDynRm = cpu_to_be32(table->ACPILevel.CcPwrDynRm); + table->ACPILevel.CcPwrDynRm1 = cpu_to_be32(table->ACPILevel.CcPwrDynRm1); + + table->MemoryACPILevel.MinVddc = table->ACPILevel.MinVddc; + table->MemoryACPILevel.MinVddcPhases = table->ACPILevel.MinVddcPhases; + + if (pi->vddci_control != CISLANDS_VOLTAGE_CONTROL_NONE) { + if (pi->acpi_vddci) + table->MemoryACPILevel.MinVddci = + cpu_to_be32(pi->acpi_vddci * VOLTAGE_SCALE); + else + table->MemoryACPILevel.MinVddci = + cpu_to_be32(pi->min_vddci_in_pp_table * VOLTAGE_SCALE); + } + + if (ci_populate_mvdd_value(rdev, 0, &voltage_level)) + table->MemoryACPILevel.MinMvdd = 0; + else + table->MemoryACPILevel.MinMvdd = + cpu_to_be32(voltage_level.Voltage * VOLTAGE_SCALE); + + mclk_pwrmgt_cntl |= MRDCK0_RESET | MRDCK1_RESET; + mclk_pwrmgt_cntl &= ~(MRDCK0_PDNB | MRDCK1_PDNB); + + dll_cntl &= ~(MRDCK0_BYPASS | MRDCK1_BYPASS); + + table->MemoryACPILevel.DllCntl = cpu_to_be32(dll_cntl); + table->MemoryACPILevel.MclkPwrmgtCntl = cpu_to_be32(mclk_pwrmgt_cntl); + table->MemoryACPILevel.MpllAdFuncCntl = + cpu_to_be32(pi->clock_registers.mpll_ad_func_cntl); + table->MemoryACPILevel.MpllDqFuncCntl = + cpu_to_be32(pi->clock_registers.mpll_dq_func_cntl); + table->MemoryACPILevel.MpllFuncCntl = + cpu_to_be32(pi->clock_registers.mpll_func_cntl); + table->MemoryACPILevel.MpllFuncCntl_1 = + cpu_to_be32(pi->clock_registers.mpll_func_cntl_1); + table->MemoryACPILevel.MpllFuncCntl_2 = + cpu_to_be32(pi->clock_registers.mpll_func_cntl_2); + table->MemoryACPILevel.MpllSs1 = cpu_to_be32(pi->clock_registers.mpll_ss1); + table->MemoryACPILevel.MpllSs2 = cpu_to_be32(pi->clock_registers.mpll_ss2); + + table->MemoryACPILevel.EnabledForThrottle = 0; + table->MemoryACPILevel.EnabledForActivity = 0; + table->MemoryACPILevel.UpH = 0; + table->MemoryACPILevel.DownH = 100; + table->MemoryACPILevel.VoltageDownH = 0; + table->MemoryACPILevel.ActivityLevel = + cpu_to_be16((u16)pi->mclk_activity_target); + + table->MemoryACPILevel.StutterEnable = false; + table->MemoryACPILevel.StrobeEnable = false; + table->MemoryACPILevel.EdcReadEnable = false; + table->MemoryACPILevel.EdcWriteEnable = false; + table->MemoryACPILevel.RttEnable = false; + + return 0; +} + + +static int ci_enable_ulv(struct radeon_device *rdev, bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_ulv_parm *ulv = &pi->ulv; + + if (ulv->supported) { + if (enable) + return (ci_send_msg_to_smc(rdev, PPSMC_MSG_EnableULV) == PPSMC_Result_OK) ? + 0 : -EINVAL; + else + return (ci_send_msg_to_smc(rdev, PPSMC_MSG_DisableULV) == PPSMC_Result_OK) ? + 0 : -EINVAL; + } + + return 0; +} + +static int ci_populate_ulv_level(struct radeon_device *rdev, + SMU7_Discrete_Ulv *state) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u16 ulv_voltage = rdev->pm.dpm.backbias_response_time; + + state->CcPwrDynRm = 0; + state->CcPwrDynRm1 = 0; + + if (ulv_voltage == 0) { + pi->ulv.supported = false; + return 0; + } + + if (pi->voltage_control != CISLANDS_VOLTAGE_CONTROL_BY_SVID2) { + if (ulv_voltage > rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[0].v) + state->VddcOffset = 0; + else + state->VddcOffset = + rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[0].v - ulv_voltage; + } else { + if (ulv_voltage > rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[0].v) + state->VddcOffsetVid = 0; + else + state->VddcOffsetVid = (u8) + ((rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[0].v - ulv_voltage) * + VOLTAGE_VID_OFFSET_SCALE2 / VOLTAGE_VID_OFFSET_SCALE1); + } + state->VddcPhase = pi->vddc_phase_shed_control ? 0 : 1; + + state->CcPwrDynRm = cpu_to_be32(state->CcPwrDynRm); + state->CcPwrDynRm1 = cpu_to_be32(state->CcPwrDynRm1); + state->VddcOffset = cpu_to_be16(state->VddcOffset); + + return 0; +} + +static int ci_calculate_sclk_params(struct radeon_device *rdev, + u32 engine_clock, + SMU7_Discrete_GraphicsLevel *sclk) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct atom_clock_dividers dividers; + u32 spll_func_cntl_3 = pi->clock_registers.cg_spll_func_cntl_3; + u32 spll_func_cntl_4 = pi->clock_registers.cg_spll_func_cntl_4; + u32 cg_spll_spread_spectrum = pi->clock_registers.cg_spll_spread_spectrum; + u32 cg_spll_spread_spectrum_2 = pi->clock_registers.cg_spll_spread_spectrum_2; + u32 reference_clock = rdev->clock.spll.reference_freq; + u32 reference_divider; + u32 fbdiv; + int ret; + + ret = radeon_atom_get_clock_dividers(rdev, + COMPUTE_GPUCLK_INPUT_FLAG_SCLK, + engine_clock, false, ÷rs); + if (ret) + return ret; + + reference_divider = 1 + dividers.ref_div; + fbdiv = dividers.fb_div & 0x3FFFFFF; + + spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK; + spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv); + spll_func_cntl_3 |= SPLL_DITHEN; + + if (pi->caps_sclk_ss_support) { + struct radeon_atom_ss ss; + u32 vco_freq = engine_clock * dividers.post_div; + + if (radeon_atombios_get_asic_ss_info(rdev, &ss, + ASIC_INTERNAL_ENGINE_SS, vco_freq)) { + u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate); + u32 clk_v = 4 * ss.percentage * fbdiv / (clk_s * 10000); + + cg_spll_spread_spectrum &= ~CLK_S_MASK; + cg_spll_spread_spectrum |= CLK_S(clk_s); + cg_spll_spread_spectrum |= SSEN; + + cg_spll_spread_spectrum_2 &= ~CLK_V_MASK; + cg_spll_spread_spectrum_2 |= CLK_V(clk_v); + } + } + + sclk->SclkFrequency = engine_clock; + sclk->CgSpllFuncCntl3 = spll_func_cntl_3; + sclk->CgSpllFuncCntl4 = spll_func_cntl_4; + sclk->SpllSpreadSpectrum = cg_spll_spread_spectrum; + sclk->SpllSpreadSpectrum2 = cg_spll_spread_spectrum_2; + sclk->SclkDid = (u8)dividers.post_divider; + + return 0; +} + +static int ci_populate_single_graphic_level(struct radeon_device *rdev, + u32 engine_clock, + u16 sclk_activity_level_t, + SMU7_Discrete_GraphicsLevel *graphic_level) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + int ret; + + ret = ci_calculate_sclk_params(rdev, engine_clock, graphic_level); + if (ret) + return ret; + + ret = ci_get_dependency_volt_by_clk(rdev, + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk, + engine_clock, &graphic_level->MinVddc); + if (ret) + return ret; + + graphic_level->SclkFrequency = engine_clock; + + graphic_level->Flags = 0; + graphic_level->MinVddcPhases = 1; + + if (pi->vddc_phase_shed_control) + ci_populate_phase_value_based_on_sclk(rdev, + &rdev->pm.dpm.dyn_state.phase_shedding_limits_table, + engine_clock, + &graphic_level->MinVddcPhases); + + graphic_level->ActivityLevel = sclk_activity_level_t; + + graphic_level->CcPwrDynRm = 0; + graphic_level->CcPwrDynRm1 = 0; + graphic_level->EnabledForActivity = 1; + graphic_level->EnabledForThrottle = 1; + graphic_level->UpH = 0; + graphic_level->DownH = 0; + graphic_level->VoltageDownH = 0; + graphic_level->PowerThrottle = 0; + + if (pi->caps_sclk_ds) + graphic_level->DeepSleepDivId = ci_get_sleep_divider_id_from_clock(rdev, + engine_clock, + CISLAND_MINIMUM_ENGINE_CLOCK); + + graphic_level->DisplayWatermark = PPSMC_DISPLAY_WATERMARK_LOW; + + graphic_level->Flags = cpu_to_be32(graphic_level->Flags); + graphic_level->MinVddc = cpu_to_be32(graphic_level->MinVddc * VOLTAGE_SCALE); + graphic_level->MinVddcPhases = cpu_to_be32(graphic_level->MinVddcPhases); + graphic_level->SclkFrequency = cpu_to_be32(graphic_level->SclkFrequency); + graphic_level->ActivityLevel = cpu_to_be16(graphic_level->ActivityLevel); + graphic_level->CgSpllFuncCntl3 = cpu_to_be32(graphic_level->CgSpllFuncCntl3); + graphic_level->CgSpllFuncCntl4 = cpu_to_be32(graphic_level->CgSpllFuncCntl4); + graphic_level->SpllSpreadSpectrum = cpu_to_be32(graphic_level->SpllSpreadSpectrum); + graphic_level->SpllSpreadSpectrum2 = cpu_to_be32(graphic_level->SpllSpreadSpectrum2); + graphic_level->CcPwrDynRm = cpu_to_be32(graphic_level->CcPwrDynRm); + graphic_level->CcPwrDynRm1 = cpu_to_be32(graphic_level->CcPwrDynRm1); + + return 0; +} + +static int ci_populate_all_graphic_levels(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_dpm_table *dpm_table = &pi->dpm_table; + u32 level_array_address = pi->dpm_table_start + + offsetof(SMU7_Discrete_DpmTable, GraphicsLevel); + u32 level_array_size = sizeof(SMU7_Discrete_GraphicsLevel) * + SMU7_MAX_LEVELS_GRAPHICS; + SMU7_Discrete_GraphicsLevel *levels = pi->smc_state_table.GraphicsLevel; + u32 i, ret; + + memset(levels, 0, level_array_size); + + for (i = 0; i < dpm_table->sclk_table.count; i++) { + ret = ci_populate_single_graphic_level(rdev, + dpm_table->sclk_table.dpm_levels[i].value, + (u16)pi->activity_target[i], + &pi->smc_state_table.GraphicsLevel[i]); + if (ret) + return ret; + if (i == (dpm_table->sclk_table.count - 1)) + pi->smc_state_table.GraphicsLevel[i].DisplayWatermark = + PPSMC_DISPLAY_WATERMARK_HIGH; + } + + pi->smc_state_table.GraphicsDpmLevelCount = (u8)dpm_table->sclk_table.count; + pi->dpm_level_enable_mask.sclk_dpm_enable_mask = + ci_get_dpm_level_enable_mask_value(&dpm_table->sclk_table); + + ret = ci_copy_bytes_to_smc(rdev, level_array_address, + (u8 *)levels, level_array_size, + pi->sram_end); + if (ret) + return ret; + + return 0; +} + +static int ci_populate_ulv_state(struct radeon_device *rdev, + SMU7_Discrete_Ulv *ulv_level) +{ + return ci_populate_ulv_level(rdev, ulv_level); +} + +static int ci_populate_all_memory_levels(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_dpm_table *dpm_table = &pi->dpm_table; + u32 level_array_address = pi->dpm_table_start + + offsetof(SMU7_Discrete_DpmTable, MemoryLevel); + u32 level_array_size = sizeof(SMU7_Discrete_MemoryLevel) * + SMU7_MAX_LEVELS_MEMORY; + SMU7_Discrete_MemoryLevel *levels = pi->smc_state_table.MemoryLevel; + u32 i, ret; + + memset(levels, 0, level_array_size); + + for (i = 0; i < dpm_table->mclk_table.count; i++) { + if (dpm_table->mclk_table.dpm_levels[i].value == 0) + return -EINVAL; + ret = ci_populate_single_memory_level(rdev, + dpm_table->mclk_table.dpm_levels[i].value, + &pi->smc_state_table.MemoryLevel[i]); + if (ret) + return ret; + } + + pi->smc_state_table.MemoryLevel[0].ActivityLevel = cpu_to_be16(0x1F); + + pi->smc_state_table.MemoryDpmLevelCount = (u8)dpm_table->mclk_table.count; + pi->dpm_level_enable_mask.mclk_dpm_enable_mask = + ci_get_dpm_level_enable_mask_value(&dpm_table->mclk_table); + + pi->smc_state_table.MemoryLevel[dpm_table->mclk_table.count - 1].DisplayWatermark = + PPSMC_DISPLAY_WATERMARK_HIGH; + + ret = ci_copy_bytes_to_smc(rdev, level_array_address, + (u8 *)levels, level_array_size, + pi->sram_end); + if (ret) + return ret; + + return 0; +} + +static void ci_reset_single_dpm_table(struct radeon_device *rdev, + struct ci_single_dpm_table* dpm_table, + u32 count) +{ + u32 i; + + dpm_table->count = count; + for (i = 0; i < MAX_REGULAR_DPM_NUMBER; i++) + dpm_table->dpm_levels[i].enabled = false; +} + +static void ci_setup_pcie_table_entry(struct ci_single_dpm_table* dpm_table, + u32 index, u32 pcie_gen, u32 pcie_lanes) +{ + dpm_table->dpm_levels[index].value = pcie_gen; + dpm_table->dpm_levels[index].param1 = pcie_lanes; + dpm_table->dpm_levels[index].enabled = true; +} + +static int ci_setup_default_pcie_tables(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + if (!pi->use_pcie_performance_levels && !pi->use_pcie_powersaving_levels) + return -EINVAL; + + if (pi->use_pcie_performance_levels && !pi->use_pcie_powersaving_levels) { + pi->pcie_gen_powersaving = pi->pcie_gen_performance; + pi->pcie_lane_powersaving = pi->pcie_lane_performance; + } else if (!pi->use_pcie_performance_levels && pi->use_pcie_powersaving_levels) { + pi->pcie_gen_performance = pi->pcie_gen_powersaving; + pi->pcie_lane_performance = pi->pcie_lane_powersaving; + } + + ci_reset_single_dpm_table(rdev, + &pi->dpm_table.pcie_speed_table, + SMU7_MAX_LEVELS_LINK); + + ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 0, + pi->pcie_gen_powersaving.min, + pi->pcie_lane_powersaving.min); + ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 1, + pi->pcie_gen_performance.min, + pi->pcie_lane_performance.min); + ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 2, + pi->pcie_gen_powersaving.min, + pi->pcie_lane_powersaving.max); + ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 3, + pi->pcie_gen_performance.min, + pi->pcie_lane_performance.max); + ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 4, + pi->pcie_gen_powersaving.max, + pi->pcie_lane_powersaving.max); + ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 5, + pi->pcie_gen_performance.max, + pi->pcie_lane_performance.max); + + pi->dpm_table.pcie_speed_table.count = 6; + + return 0; +} + +static int ci_setup_default_dpm_tables(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct radeon_clock_voltage_dependency_table *allowed_sclk_vddc_table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + struct radeon_clock_voltage_dependency_table *allowed_mclk_table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk; + struct radeon_cac_leakage_table *std_voltage_table = + &rdev->pm.dpm.dyn_state.cac_leakage_table; + u32 i; + + if (allowed_sclk_vddc_table == NULL) + return -EINVAL; + if (allowed_sclk_vddc_table->count < 1) + return -EINVAL; + if (allowed_mclk_table == NULL) + return -EINVAL; + if (allowed_mclk_table->count < 1) + return -EINVAL; + + memset(&pi->dpm_table, 0, sizeof(struct ci_dpm_table)); + + ci_reset_single_dpm_table(rdev, + &pi->dpm_table.sclk_table, + SMU7_MAX_LEVELS_GRAPHICS); + ci_reset_single_dpm_table(rdev, + &pi->dpm_table.mclk_table, + SMU7_MAX_LEVELS_MEMORY); + ci_reset_single_dpm_table(rdev, + &pi->dpm_table.vddc_table, + SMU7_MAX_LEVELS_VDDC); + ci_reset_single_dpm_table(rdev, + &pi->dpm_table.vddci_table, + SMU7_MAX_LEVELS_VDDCI); + ci_reset_single_dpm_table(rdev, + &pi->dpm_table.mvdd_table, + SMU7_MAX_LEVELS_MVDD); + + pi->dpm_table.sclk_table.count = 0; + for (i = 0; i < allowed_sclk_vddc_table->count; i++) { + if ((i == 0) || + (pi->dpm_table.sclk_table.dpm_levels[pi->dpm_table.sclk_table.count-1].value != + allowed_sclk_vddc_table->entries[i].clk)) { + pi->dpm_table.sclk_table.dpm_levels[pi->dpm_table.sclk_table.count].value = + allowed_sclk_vddc_table->entries[i].clk; + pi->dpm_table.sclk_table.dpm_levels[pi->dpm_table.sclk_table.count].enabled = true; + pi->dpm_table.sclk_table.count++; + } + } + + pi->dpm_table.mclk_table.count = 0; + for (i = 0; i < allowed_mclk_table->count; i++) { + if ((i==0) || + (pi->dpm_table.mclk_table.dpm_levels[pi->dpm_table.mclk_table.count-1].value != + allowed_mclk_table->entries[i].clk)) { + pi->dpm_table.mclk_table.dpm_levels[pi->dpm_table.mclk_table.count].value = + allowed_mclk_table->entries[i].clk; + pi->dpm_table.mclk_table.dpm_levels[pi->dpm_table.mclk_table.count].enabled = true; + pi->dpm_table.mclk_table.count++; + } + } + + for (i = 0; i < allowed_sclk_vddc_table->count; i++) { + pi->dpm_table.vddc_table.dpm_levels[i].value = + allowed_sclk_vddc_table->entries[i].v; + pi->dpm_table.vddc_table.dpm_levels[i].param1 = + std_voltage_table->entries[i].leakage; + pi->dpm_table.vddc_table.dpm_levels[i].enabled = true; + } + pi->dpm_table.vddc_table.count = allowed_sclk_vddc_table->count; + + allowed_mclk_table = &rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk; + if (allowed_mclk_table) { + for (i = 0; i < allowed_mclk_table->count; i++) { + pi->dpm_table.vddci_table.dpm_levels[i].value = + allowed_mclk_table->entries[i].v; + pi->dpm_table.vddci_table.dpm_levels[i].enabled = true; + } + pi->dpm_table.vddci_table.count = allowed_mclk_table->count; + } + + allowed_mclk_table = &rdev->pm.dpm.dyn_state.mvdd_dependency_on_mclk; + if (allowed_mclk_table) { + for (i = 0; i < allowed_mclk_table->count; i++) { + pi->dpm_table.mvdd_table.dpm_levels[i].value = + allowed_mclk_table->entries[i].v; + pi->dpm_table.mvdd_table.dpm_levels[i].enabled = true; + } + pi->dpm_table.mvdd_table.count = allowed_mclk_table->count; + } + + ci_setup_default_pcie_tables(rdev); + + return 0; +} + +static int ci_find_boot_level(struct ci_single_dpm_table *table, + u32 value, u32 *boot_level) +{ + u32 i; + int ret = -EINVAL; + + for(i = 0; i < table->count; i++) { + if (value == table->dpm_levels[i].value) { + *boot_level = i; + ret = 0; + } + } + + return ret; +} + +static int ci_init_smc_table(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_ulv_parm *ulv = &pi->ulv; + struct radeon_ps *radeon_boot_state = rdev->pm.dpm.boot_ps; + SMU7_Discrete_DpmTable *table = &pi->smc_state_table; + int ret; + + ret = ci_setup_default_dpm_tables(rdev); + if (ret) + return ret; + + if (pi->voltage_control != CISLANDS_VOLTAGE_CONTROL_NONE) + ci_populate_smc_voltage_tables(rdev, table); + + ci_init_fps_limits(rdev); + + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC) + table->SystemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC; + + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC) + table->SystemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC; + + if (pi->mem_gddr5) + table->SystemFlags |= PPSMC_SYSTEMFLAG_GDDR5; + + if (ulv->supported) { + ret = ci_populate_ulv_state(rdev, &pi->smc_state_table.Ulv); + if (ret) + return ret; + WREG32_SMC(CG_ULV_PARAMETER, ulv->cg_ulv_parameter); + } + + ret = ci_populate_all_graphic_levels(rdev); + if (ret) + return ret; + + ret = ci_populate_all_memory_levels(rdev); + if (ret) + return ret; + + ci_populate_smc_link_level(rdev, table); + + ret = ci_populate_smc_acpi_level(rdev, table); + if (ret) + return ret; + + ret = ci_populate_smc_vce_level(rdev, table); + if (ret) + return ret; + + ret = ci_populate_smc_acp_level(rdev, table); + if (ret) + return ret; + + ret = ci_populate_smc_samu_level(rdev, table); + if (ret) + return ret; + + ret = ci_do_program_memory_timing_parameters(rdev); + if (ret) + return ret; + + ret = ci_populate_smc_uvd_level(rdev, table); + if (ret) + return ret; + + table->UvdBootLevel = 0; + table->VceBootLevel = 0; + table->AcpBootLevel = 0; + table->SamuBootLevel = 0; + table->GraphicsBootLevel = 0; + table->MemoryBootLevel = 0; + + ret = ci_find_boot_level(&pi->dpm_table.sclk_table, + pi->vbios_boot_state.sclk_bootup_value, + (u32 *)&pi->smc_state_table.GraphicsBootLevel); + + ret = ci_find_boot_level(&pi->dpm_table.mclk_table, + pi->vbios_boot_state.mclk_bootup_value, + (u32 *)&pi->smc_state_table.MemoryBootLevel); + + table->BootVddc = pi->vbios_boot_state.vddc_bootup_value; + table->BootVddci = pi->vbios_boot_state.vddci_bootup_value; + table->BootMVdd = pi->vbios_boot_state.mvdd_bootup_value; + + ci_populate_smc_initial_state(rdev, radeon_boot_state); + + ret = ci_populate_bapm_parameters_in_dpm_table(rdev); + if (ret) + return ret; + + table->UVDInterval = 1; + table->VCEInterval = 1; + table->ACPInterval = 1; + table->SAMUInterval = 1; + table->GraphicsVoltageChangeEnable = 1; + table->GraphicsThermThrottleEnable = 1; + table->GraphicsInterval = 1; + table->VoltageInterval = 1; + table->ThermalInterval = 1; + table->TemperatureLimitHigh = (u16)((pi->thermal_temp_setting.temperature_high * + CISLANDS_Q88_FORMAT_CONVERSION_UNIT) / 1000); + table->TemperatureLimitLow = (u16)((pi->thermal_temp_setting.temperature_low * + CISLANDS_Q88_FORMAT_CONVERSION_UNIT) / 1000); + table->MemoryVoltageChangeEnable = 1; + table->MemoryInterval = 1; + table->VoltageResponseTime = 0; + table->VddcVddciDelta = 4000; + table->PhaseResponseTime = 0; + table->MemoryThermThrottleEnable = 1; + table->PCIeBootLinkLevel = 0; + table->PCIeGenInterval = 1; + if (pi->voltage_control == CISLANDS_VOLTAGE_CONTROL_BY_SVID2) + table->SVI2Enable = 1; + else + table->SVI2Enable = 0; + + table->ThermGpio = 17; + table->SclkStepSize = 0x4000; + + table->SystemFlags = cpu_to_be32(table->SystemFlags); + table->SmioMaskVddcVid = cpu_to_be32(table->SmioMaskVddcVid); + table->SmioMaskVddcPhase = cpu_to_be32(table->SmioMaskVddcPhase); + table->SmioMaskVddciVid = cpu_to_be32(table->SmioMaskVddciVid); + table->SmioMaskMvddVid = cpu_to_be32(table->SmioMaskMvddVid); + table->SclkStepSize = cpu_to_be32(table->SclkStepSize); + table->TemperatureLimitHigh = cpu_to_be16(table->TemperatureLimitHigh); + table->TemperatureLimitLow = cpu_to_be16(table->TemperatureLimitLow); + table->VddcVddciDelta = cpu_to_be16(table->VddcVddciDelta); + table->VoltageResponseTime = cpu_to_be16(table->VoltageResponseTime); + table->PhaseResponseTime = cpu_to_be16(table->PhaseResponseTime); + table->BootVddc = cpu_to_be16(table->BootVddc * VOLTAGE_SCALE); + table->BootVddci = cpu_to_be16(table->BootVddci * VOLTAGE_SCALE); + table->BootMVdd = cpu_to_be16(table->BootMVdd * VOLTAGE_SCALE); + + ret = ci_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Discrete_DpmTable, SystemFlags), + (u8 *)&table->SystemFlags, + sizeof(SMU7_Discrete_DpmTable) - 3 * sizeof(SMU7_PIDController), + pi->sram_end); + if (ret) + return ret; + + return 0; +} + +static void ci_trim_single_dpm_states(struct radeon_device *rdev, + struct ci_single_dpm_table *dpm_table, + u32 low_limit, u32 high_limit) +{ + u32 i; + + for (i = 0; i < dpm_table->count; i++) { + if ((dpm_table->dpm_levels[i].value < low_limit) || + (dpm_table->dpm_levels[i].value > high_limit)) + dpm_table->dpm_levels[i].enabled = false; + else + dpm_table->dpm_levels[i].enabled = true; + } +} + +static void ci_trim_pcie_dpm_states(struct radeon_device *rdev, + u32 speed_low, u32 lanes_low, + u32 speed_high, u32 lanes_high) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_single_dpm_table *pcie_table = &pi->dpm_table.pcie_speed_table; + u32 i, j; + + for (i = 0; i < pcie_table->count; i++) { + if ((pcie_table->dpm_levels[i].value < speed_low) || + (pcie_table->dpm_levels[i].param1 < lanes_low) || + (pcie_table->dpm_levels[i].value > speed_high) || + (pcie_table->dpm_levels[i].param1 > lanes_high)) + pcie_table->dpm_levels[i].enabled = false; + else + pcie_table->dpm_levels[i].enabled = true; + } + + for (i = 0; i < pcie_table->count; i++) { + if (pcie_table->dpm_levels[i].enabled) { + for (j = i + 1; j < pcie_table->count; j++) { + if (pcie_table->dpm_levels[j].enabled) { + if ((pcie_table->dpm_levels[i].value == pcie_table->dpm_levels[j].value) && + (pcie_table->dpm_levels[i].param1 == pcie_table->dpm_levels[j].param1)) + pcie_table->dpm_levels[j].enabled = false; + } + } + } + } +} + +static int ci_trim_dpm_states(struct radeon_device *rdev, + struct radeon_ps *radeon_state) +{ + struct ci_ps *state = ci_get_ps(radeon_state); + struct ci_power_info *pi = ci_get_pi(rdev); + u32 high_limit_count; + + if (state->performance_level_count < 1) + return -EINVAL; + + if (state->performance_level_count == 1) + high_limit_count = 0; + else + high_limit_count = 1; + + ci_trim_single_dpm_states(rdev, + &pi->dpm_table.sclk_table, + state->performance_levels[0].sclk, + state->performance_levels[high_limit_count].sclk); + + ci_trim_single_dpm_states(rdev, + &pi->dpm_table.mclk_table, + state->performance_levels[0].mclk, + state->performance_levels[high_limit_count].mclk); + + ci_trim_pcie_dpm_states(rdev, + state->performance_levels[0].pcie_gen, + state->performance_levels[0].pcie_lane, + state->performance_levels[high_limit_count].pcie_gen, + state->performance_levels[high_limit_count].pcie_lane); + + return 0; +} + +static int ci_apply_disp_minimum_voltage_request(struct radeon_device *rdev) +{ + struct radeon_clock_voltage_dependency_table *disp_voltage_table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk; + struct radeon_clock_voltage_dependency_table *vddc_table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + u32 requested_voltage = 0; + u32 i; + + if (disp_voltage_table == NULL) + return -EINVAL; + if (!disp_voltage_table->count) + return -EINVAL; + + for (i = 0; i < disp_voltage_table->count; i++) { + if (rdev->clock.current_dispclk == disp_voltage_table->entries[i].clk) + requested_voltage = disp_voltage_table->entries[i].v; + } + + for (i = 0; i < vddc_table->count; i++) { + if (requested_voltage <= vddc_table->entries[i].v) { + requested_voltage = vddc_table->entries[i].v; + return (ci_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_VddC_Request, + requested_voltage * VOLTAGE_SCALE) == PPSMC_Result_OK) ? + 0 : -EINVAL; + } + } + + return -EINVAL; +} + +static int ci_upload_dpm_level_enable_mask(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + PPSMC_Result result; + + if (!pi->sclk_dpm_key_disabled) { + if (pi->dpm_level_enable_mask.sclk_dpm_enable_mask) { + result = ci_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_SCLKDPM_SetEnabledMask, + pi->dpm_level_enable_mask.sclk_dpm_enable_mask); + if (result != PPSMC_Result_OK) + return -EINVAL; + } + } + + if (!pi->mclk_dpm_key_disabled) { + if (pi->dpm_level_enable_mask.mclk_dpm_enable_mask) { + result = ci_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_MCLKDPM_SetEnabledMask, + pi->dpm_level_enable_mask.mclk_dpm_enable_mask); + if (result != PPSMC_Result_OK) + return -EINVAL; + } + } + + if (!pi->pcie_dpm_key_disabled) { + if (pi->dpm_level_enable_mask.pcie_dpm_enable_mask) { + result = ci_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_PCIeDPM_SetEnabledMask, + pi->dpm_level_enable_mask.pcie_dpm_enable_mask); + if (result != PPSMC_Result_OK) + return -EINVAL; + } + } + + ci_apply_disp_minimum_voltage_request(rdev); + + return 0; +} + +static void ci_find_dpm_states_clocks_in_dpm_table(struct radeon_device *rdev, + struct radeon_ps *radeon_state) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_ps *state = ci_get_ps(radeon_state); + struct ci_single_dpm_table *sclk_table = &pi->dpm_table.sclk_table; + u32 sclk = state->performance_levels[state->performance_level_count-1].sclk; + struct ci_single_dpm_table *mclk_table = &pi->dpm_table.mclk_table; + u32 mclk = state->performance_levels[state->performance_level_count-1].mclk; + u32 i; + + pi->need_update_smu7_dpm_table = 0; + + for (i = 0; i < sclk_table->count; i++) { + if (sclk == sclk_table->dpm_levels[i].value) + break; + } + + if (i >= sclk_table->count) { + pi->need_update_smu7_dpm_table |= DPMTABLE_OD_UPDATE_SCLK; + } else { + /* XXX check display min clock requirements */ + if (0 != CISLAND_MINIMUM_ENGINE_CLOCK) + pi->need_update_smu7_dpm_table |= DPMTABLE_UPDATE_SCLK; + } + + for (i = 0; i < mclk_table->count; i++) { + if (mclk == mclk_table->dpm_levels[i].value) + break; + } + + if (i >= mclk_table->count) + pi->need_update_smu7_dpm_table |= DPMTABLE_OD_UPDATE_MCLK; + + if (rdev->pm.dpm.current_active_crtc_count != + rdev->pm.dpm.new_active_crtc_count) + pi->need_update_smu7_dpm_table |= DPMTABLE_UPDATE_MCLK; +} + +static int ci_populate_and_upload_sclk_mclk_dpm_levels(struct radeon_device *rdev, + struct radeon_ps *radeon_state) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_ps *state = ci_get_ps(radeon_state); + u32 sclk = state->performance_levels[state->performance_level_count-1].sclk; + u32 mclk = state->performance_levels[state->performance_level_count-1].mclk; + struct ci_dpm_table *dpm_table = &pi->dpm_table; + int ret; + + if (!pi->need_update_smu7_dpm_table) + return 0; + + if (pi->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_SCLK) + dpm_table->sclk_table.dpm_levels[dpm_table->sclk_table.count-1].value = sclk; + + if (pi->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK) + dpm_table->mclk_table.dpm_levels[dpm_table->mclk_table.count-1].value = mclk; + + if (pi->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK | DPMTABLE_UPDATE_SCLK)) { + ret = ci_populate_all_graphic_levels(rdev); + if (ret) + return ret; + } + + if (pi->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_MCLK | DPMTABLE_UPDATE_MCLK)) { + ret = ci_populate_all_memory_levels(rdev); + if (ret) + return ret; + } + + return 0; +} + +static int ci_enable_uvd_dpm(struct radeon_device *rdev, bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + const struct radeon_clock_and_voltage_limits *max_limits; + int i; + + if (rdev->pm.dpm.ac_power) + max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + else + max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc; + + if (enable) { + pi->dpm_level_enable_mask.uvd_dpm_enable_mask = 0; + + for (i = rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.count - 1; i >= 0; i--) { + if (rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[i].v <= max_limits->vddc) { + pi->dpm_level_enable_mask.uvd_dpm_enable_mask |= 1 << i; + + if (!pi->caps_uvd_dpm) + break; + } + } + + ci_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_UVDDPM_SetEnabledMask, + pi->dpm_level_enable_mask.uvd_dpm_enable_mask); + + if (pi->last_mclk_dpm_enable_mask & 0x1) { + pi->uvd_enabled = true; + pi->dpm_level_enable_mask.mclk_dpm_enable_mask &= 0xFFFFFFFE; + ci_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_MCLKDPM_SetEnabledMask, + pi->dpm_level_enable_mask.mclk_dpm_enable_mask); + } + } else { + if (pi->last_mclk_dpm_enable_mask & 0x1) { + pi->uvd_enabled = false; + pi->dpm_level_enable_mask.mclk_dpm_enable_mask |= 1; + ci_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_MCLKDPM_SetEnabledMask, + pi->dpm_level_enable_mask.mclk_dpm_enable_mask); + } + } + + return (ci_send_msg_to_smc(rdev, enable ? + PPSMC_MSG_UVDDPM_Enable : PPSMC_MSG_UVDDPM_Disable) == PPSMC_Result_OK) ? + 0 : -EINVAL; +} + +static int ci_enable_vce_dpm(struct radeon_device *rdev, bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + const struct radeon_clock_and_voltage_limits *max_limits; + int i; + + if (rdev->pm.dpm.ac_power) + max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + else + max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc; + + if (enable) { + pi->dpm_level_enable_mask.vce_dpm_enable_mask = 0; + for (i = rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.count - 1; i >= 0; i--) { + if (rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries[i].v <= max_limits->vddc) { + pi->dpm_level_enable_mask.vce_dpm_enable_mask |= 1 << i; + + if (!pi->caps_vce_dpm) + break; + } + } + + ci_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_VCEDPM_SetEnabledMask, + pi->dpm_level_enable_mask.vce_dpm_enable_mask); + } + + return (ci_send_msg_to_smc(rdev, enable ? + PPSMC_MSG_VCEDPM_Enable : PPSMC_MSG_VCEDPM_Disable) == PPSMC_Result_OK) ? + 0 : -EINVAL; +} + +#if 0 +static int ci_enable_samu_dpm(struct radeon_device *rdev, bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + const struct radeon_clock_and_voltage_limits *max_limits; + int i; + + if (rdev->pm.dpm.ac_power) + max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + else + max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc; + + if (enable) { + pi->dpm_level_enable_mask.samu_dpm_enable_mask = 0; + for (i = rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.count - 1; i >= 0; i--) { + if (rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.entries[i].v <= max_limits->vddc) { + pi->dpm_level_enable_mask.samu_dpm_enable_mask |= 1 << i; + + if (!pi->caps_samu_dpm) + break; + } + } + + ci_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_SAMUDPM_SetEnabledMask, + pi->dpm_level_enable_mask.samu_dpm_enable_mask); + } + return (ci_send_msg_to_smc(rdev, enable ? + PPSMC_MSG_SAMUDPM_Enable : PPSMC_MSG_SAMUDPM_Disable) == PPSMC_Result_OK) ? + 0 : -EINVAL; +} + +static int ci_enable_acp_dpm(struct radeon_device *rdev, bool enable) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + const struct radeon_clock_and_voltage_limits *max_limits; + int i; + + if (rdev->pm.dpm.ac_power) + max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + else + max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc; + + if (enable) { + pi->dpm_level_enable_mask.acp_dpm_enable_mask = 0; + for (i = rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.count - 1; i >= 0; i--) { + if (rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.entries[i].v <= max_limits->vddc) { + pi->dpm_level_enable_mask.acp_dpm_enable_mask |= 1 << i; + + if (!pi->caps_acp_dpm) + break; + } + } + + ci_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_ACPDPM_SetEnabledMask, + pi->dpm_level_enable_mask.acp_dpm_enable_mask); + } + + return (ci_send_msg_to_smc(rdev, enable ? + PPSMC_MSG_ACPDPM_Enable : PPSMC_MSG_ACPDPM_Disable) == PPSMC_Result_OK) ? + 0 : -EINVAL; +} +#endif + +static int ci_update_uvd_dpm(struct radeon_device *rdev, bool gate) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 tmp; + + if (!gate) { + if (pi->caps_uvd_dpm || + (rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.count <= 0)) + pi->smc_state_table.UvdBootLevel = 0; + else + pi->smc_state_table.UvdBootLevel = + rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.count - 1; + + tmp = RREG32_SMC(DPM_TABLE_475); + tmp &= ~UvdBootLevel_MASK; + tmp |= UvdBootLevel(pi->smc_state_table.UvdBootLevel); + WREG32_SMC(DPM_TABLE_475, tmp); + } + + return ci_enable_uvd_dpm(rdev, !gate); +} + +static u8 ci_get_vce_boot_level(struct radeon_device *rdev) +{ + u8 i; + u32 min_evclk = 30000; /* ??? */ + struct radeon_vce_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table; + + for (i = 0; i < table->count; i++) { + if (table->entries[i].evclk >= min_evclk) + return i; + } + + return table->count - 1; +} + +static int ci_update_vce_dpm(struct radeon_device *rdev, + struct radeon_ps *radeon_new_state, + struct radeon_ps *radeon_current_state) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + int ret = 0; + u32 tmp; + + if (radeon_current_state->evclk != radeon_new_state->evclk) { + if (radeon_new_state->evclk) { + /* turn the clocks on when encoding */ + cik_update_cg(rdev, RADEON_CG_BLOCK_VCE, false); + + pi->smc_state_table.VceBootLevel = ci_get_vce_boot_level(rdev); + tmp = RREG32_SMC(DPM_TABLE_475); + tmp &= ~VceBootLevel_MASK; + tmp |= VceBootLevel(pi->smc_state_table.VceBootLevel); + WREG32_SMC(DPM_TABLE_475, tmp); + + ret = ci_enable_vce_dpm(rdev, true); + } else { + /* turn the clocks off when not encoding */ + cik_update_cg(rdev, RADEON_CG_BLOCK_VCE, true); + + ret = ci_enable_vce_dpm(rdev, false); + } + } + return ret; +} + +#if 0 +static int ci_update_samu_dpm(struct radeon_device *rdev, bool gate) +{ + return ci_enable_samu_dpm(rdev, gate); +} + +static int ci_update_acp_dpm(struct radeon_device *rdev, bool gate) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 tmp; + + if (!gate) { + pi->smc_state_table.AcpBootLevel = 0; + + tmp = RREG32_SMC(DPM_TABLE_475); + tmp &= ~AcpBootLevel_MASK; + tmp |= AcpBootLevel(pi->smc_state_table.AcpBootLevel); + WREG32_SMC(DPM_TABLE_475, tmp); + } + + return ci_enable_acp_dpm(rdev, !gate); +} +#endif + +static int ci_generate_dpm_level_enable_mask(struct radeon_device *rdev, + struct radeon_ps *radeon_state) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + int ret; + + ret = ci_trim_dpm_states(rdev, radeon_state); + if (ret) + return ret; + + pi->dpm_level_enable_mask.sclk_dpm_enable_mask = + ci_get_dpm_level_enable_mask_value(&pi->dpm_table.sclk_table); + pi->dpm_level_enable_mask.mclk_dpm_enable_mask = + ci_get_dpm_level_enable_mask_value(&pi->dpm_table.mclk_table); + pi->last_mclk_dpm_enable_mask = + pi->dpm_level_enable_mask.mclk_dpm_enable_mask; + if (pi->uvd_enabled) { + if (pi->dpm_level_enable_mask.mclk_dpm_enable_mask & 1) + pi->dpm_level_enable_mask.mclk_dpm_enable_mask &= 0xFFFFFFFE; + } + pi->dpm_level_enable_mask.pcie_dpm_enable_mask = + ci_get_dpm_level_enable_mask_value(&pi->dpm_table.pcie_speed_table); + + return 0; +} + +static u32 ci_get_lowest_enabled_level(struct radeon_device *rdev, + u32 level_mask) +{ + u32 level = 0; + + while ((level_mask & (1 << level)) == 0) + level++; + + return level; +} + + +int ci_dpm_force_performance_level(struct radeon_device *rdev, + enum radeon_dpm_forced_level level) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + PPSMC_Result smc_result; + u32 tmp, levels, i; + int ret; + + if (level == RADEON_DPM_FORCED_LEVEL_HIGH) { + if ((!pi->sclk_dpm_key_disabled) && + pi->dpm_level_enable_mask.sclk_dpm_enable_mask) { + levels = 0; + tmp = pi->dpm_level_enable_mask.sclk_dpm_enable_mask; + while (tmp >>= 1) + levels++; + if (levels) { + ret = ci_dpm_force_state_sclk(rdev, levels); + if (ret) + return ret; + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX) & + CURR_SCLK_INDEX_MASK) >> CURR_SCLK_INDEX_SHIFT; + if (tmp == levels) + break; + udelay(1); + } + } + } + if ((!pi->mclk_dpm_key_disabled) && + pi->dpm_level_enable_mask.mclk_dpm_enable_mask) { + levels = 0; + tmp = pi->dpm_level_enable_mask.mclk_dpm_enable_mask; + while (tmp >>= 1) + levels++; + if (levels) { + ret = ci_dpm_force_state_mclk(rdev, levels); + if (ret) + return ret; + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX) & + CURR_MCLK_INDEX_MASK) >> CURR_MCLK_INDEX_SHIFT; + if (tmp == levels) + break; + udelay(1); + } + } + } + if ((!pi->pcie_dpm_key_disabled) && + pi->dpm_level_enable_mask.pcie_dpm_enable_mask) { + levels = 0; + tmp = pi->dpm_level_enable_mask.pcie_dpm_enable_mask; + while (tmp >>= 1) + levels++; + if (levels) { + ret = ci_dpm_force_state_pcie(rdev, level); + if (ret) + return ret; + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX_1) & + CURR_PCIE_INDEX_MASK) >> CURR_PCIE_INDEX_SHIFT; + if (tmp == levels) + break; + udelay(1); + } + } + } + } else if (level == RADEON_DPM_FORCED_LEVEL_LOW) { + if ((!pi->sclk_dpm_key_disabled) && + pi->dpm_level_enable_mask.sclk_dpm_enable_mask) { + levels = ci_get_lowest_enabled_level(rdev, + pi->dpm_level_enable_mask.sclk_dpm_enable_mask); + ret = ci_dpm_force_state_sclk(rdev, levels); + if (ret) + return ret; + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX) & + CURR_SCLK_INDEX_MASK) >> CURR_SCLK_INDEX_SHIFT; + if (tmp == levels) + break; + udelay(1); + } + } + if ((!pi->mclk_dpm_key_disabled) && + pi->dpm_level_enable_mask.mclk_dpm_enable_mask) { + levels = ci_get_lowest_enabled_level(rdev, + pi->dpm_level_enable_mask.mclk_dpm_enable_mask); + ret = ci_dpm_force_state_mclk(rdev, levels); + if (ret) + return ret; + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX) & + CURR_MCLK_INDEX_MASK) >> CURR_MCLK_INDEX_SHIFT; + if (tmp == levels) + break; + udelay(1); + } + } + if ((!pi->pcie_dpm_key_disabled) && + pi->dpm_level_enable_mask.pcie_dpm_enable_mask) { + levels = ci_get_lowest_enabled_level(rdev, + pi->dpm_level_enable_mask.pcie_dpm_enable_mask); + ret = ci_dpm_force_state_pcie(rdev, levels); + if (ret) + return ret; + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX_1) & + CURR_PCIE_INDEX_MASK) >> CURR_PCIE_INDEX_SHIFT; + if (tmp == levels) + break; + udelay(1); + } + } + } else if (level == RADEON_DPM_FORCED_LEVEL_AUTO) { + if (!pi->sclk_dpm_key_disabled) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_NoForcedLevel); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + if (!pi->mclk_dpm_key_disabled) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_MCLKDPM_NoForcedLevel); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + if (!pi->pcie_dpm_key_disabled) { + smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_PCIeDPM_UnForceLevel); + if (smc_result != PPSMC_Result_OK) + return -EINVAL; + } + } + + rdev->pm.dpm.forced_level = level; + + return 0; +} + +static int ci_set_mc_special_registers(struct radeon_device *rdev, + struct ci_mc_reg_table *table) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u8 i, j, k; + u32 temp_reg; + + for (i = 0, j = table->last; i < table->last; i++) { + if (j >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) + return -EINVAL; + switch(table->mc_reg_address[i].s1 << 2) { + case MC_SEQ_MISC1: + temp_reg = RREG32(MC_PMG_CMD_EMRS); + table->mc_reg_address[j].s1 = MC_PMG_CMD_EMRS >> 2; + table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_EMRS_LP >> 2; + for (k = 0; k < table->num_entries; k++) { + table->mc_reg_table_entry[k].mc_data[j] = + ((temp_reg & 0xffff0000)) | ((table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16); + } + j++; + if (j >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) + return -EINVAL; + + temp_reg = RREG32(MC_PMG_CMD_MRS); + table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS >> 2; + table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS_LP >> 2; + for (k = 0; k < table->num_entries; k++) { + table->mc_reg_table_entry[k].mc_data[j] = + (temp_reg & 0xffff0000) | (table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff); + if (!pi->mem_gddr5) + table->mc_reg_table_entry[k].mc_data[j] |= 0x100; + } + j++; + if (j > SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) + return -EINVAL; + + if (!pi->mem_gddr5) { + table->mc_reg_address[j].s1 = MC_PMG_AUTO_CMD >> 2; + table->mc_reg_address[j].s0 = MC_PMG_AUTO_CMD >> 2; + for (k = 0; k < table->num_entries; k++) { + table->mc_reg_table_entry[k].mc_data[j] = + (table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16; + } + j++; + if (j > SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) + return -EINVAL; + } + break; + case MC_SEQ_RESERVE_M: + temp_reg = RREG32(MC_PMG_CMD_MRS1); + table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS1 >> 2; + table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS1_LP >> 2; + for (k = 0; k < table->num_entries; k++) { + table->mc_reg_table_entry[k].mc_data[j] = + (temp_reg & 0xffff0000) | (table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff); + } + j++; + if (j > SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) + return -EINVAL; + break; + default: + break; + } + + } + + table->last = j; + + return 0; +} + +static bool ci_check_s0_mc_reg_index(u16 in_reg, u16 *out_reg) +{ + bool result = true; + + switch(in_reg) { + case MC_SEQ_RAS_TIMING >> 2: + *out_reg = MC_SEQ_RAS_TIMING_LP >> 2; + break; + case MC_SEQ_DLL_STBY >> 2: + *out_reg = MC_SEQ_DLL_STBY_LP >> 2; + break; + case MC_SEQ_G5PDX_CMD0 >> 2: + *out_reg = MC_SEQ_G5PDX_CMD0_LP >> 2; + break; + case MC_SEQ_G5PDX_CMD1 >> 2: + *out_reg = MC_SEQ_G5PDX_CMD1_LP >> 2; + break; + case MC_SEQ_G5PDX_CTRL >> 2: + *out_reg = MC_SEQ_G5PDX_CTRL_LP >> 2; + break; + case MC_SEQ_CAS_TIMING >> 2: + *out_reg = MC_SEQ_CAS_TIMING_LP >> 2; + break; + case MC_SEQ_MISC_TIMING >> 2: + *out_reg = MC_SEQ_MISC_TIMING_LP >> 2; + break; + case MC_SEQ_MISC_TIMING2 >> 2: + *out_reg = MC_SEQ_MISC_TIMING2_LP >> 2; + break; + case MC_SEQ_PMG_DVS_CMD >> 2: + *out_reg = MC_SEQ_PMG_DVS_CMD_LP >> 2; + break; + case MC_SEQ_PMG_DVS_CTL >> 2: + *out_reg = MC_SEQ_PMG_DVS_CTL_LP >> 2; + break; + case MC_SEQ_RD_CTL_D0 >> 2: + *out_reg = MC_SEQ_RD_CTL_D0_LP >> 2; + break; + case MC_SEQ_RD_CTL_D1 >> 2: + *out_reg = MC_SEQ_RD_CTL_D1_LP >> 2; + break; + case MC_SEQ_WR_CTL_D0 >> 2: + *out_reg = MC_SEQ_WR_CTL_D0_LP >> 2; + break; + case MC_SEQ_WR_CTL_D1 >> 2: + *out_reg = MC_SEQ_WR_CTL_D1_LP >> 2; + break; + case MC_PMG_CMD_EMRS >> 2: + *out_reg = MC_SEQ_PMG_CMD_EMRS_LP >> 2; + break; + case MC_PMG_CMD_MRS >> 2: + *out_reg = MC_SEQ_PMG_CMD_MRS_LP >> 2; + break; + case MC_PMG_CMD_MRS1 >> 2: + *out_reg = MC_SEQ_PMG_CMD_MRS1_LP >> 2; + break; + case MC_SEQ_PMG_TIMING >> 2: + *out_reg = MC_SEQ_PMG_TIMING_LP >> 2; + break; + case MC_PMG_CMD_MRS2 >> 2: + *out_reg = MC_SEQ_PMG_CMD_MRS2_LP >> 2; + break; + case MC_SEQ_WR_CTL_2 >> 2: + *out_reg = MC_SEQ_WR_CTL_2_LP >> 2; + break; + default: + result = false; + break; + } + + return result; +} + +static void ci_set_valid_flag(struct ci_mc_reg_table *table) +{ + u8 i, j; + + for (i = 0; i < table->last; i++) { + for (j = 1; j < table->num_entries; j++) { + if (table->mc_reg_table_entry[j-1].mc_data[i] != + table->mc_reg_table_entry[j].mc_data[i]) { + table->valid_flag |= 1 << i; + break; + } + } + } +} + +static void ci_set_s0_mc_reg_index(struct ci_mc_reg_table *table) +{ + u32 i; + u16 address; + + for (i = 0; i < table->last; i++) { + table->mc_reg_address[i].s0 = + ci_check_s0_mc_reg_index(table->mc_reg_address[i].s1, &address) ? + address : table->mc_reg_address[i].s1; + } +} + +static int ci_copy_vbios_mc_reg_table(const struct atom_mc_reg_table *table, + struct ci_mc_reg_table *ci_table) +{ + u8 i, j; + + if (table->last > SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) + return -EINVAL; + if (table->num_entries > MAX_AC_TIMING_ENTRIES) + return -EINVAL; + + for (i = 0; i < table->last; i++) + ci_table->mc_reg_address[i].s1 = table->mc_reg_address[i].s1; + + ci_table->last = table->last; + + for (i = 0; i < table->num_entries; i++) { + ci_table->mc_reg_table_entry[i].mclk_max = + table->mc_reg_table_entry[i].mclk_max; + for (j = 0; j < table->last; j++) + ci_table->mc_reg_table_entry[i].mc_data[j] = + table->mc_reg_table_entry[i].mc_data[j]; + } + ci_table->num_entries = table->num_entries; + + return 0; +} + +static int ci_initialize_mc_reg_table(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct atom_mc_reg_table *table; + struct ci_mc_reg_table *ci_table = &pi->mc_reg_table; + u8 module_index = rv770_get_memory_module_index(rdev); + int ret; + + table = kzalloc(sizeof(struct atom_mc_reg_table), GFP_KERNEL); + if (!table) + return -ENOMEM; + + WREG32(MC_SEQ_RAS_TIMING_LP, RREG32(MC_SEQ_RAS_TIMING)); + WREG32(MC_SEQ_CAS_TIMING_LP, RREG32(MC_SEQ_CAS_TIMING)); + WREG32(MC_SEQ_DLL_STBY_LP, RREG32(MC_SEQ_DLL_STBY)); + WREG32(MC_SEQ_G5PDX_CMD0_LP, RREG32(MC_SEQ_G5PDX_CMD0)); + WREG32(MC_SEQ_G5PDX_CMD1_LP, RREG32(MC_SEQ_G5PDX_CMD1)); + WREG32(MC_SEQ_G5PDX_CTRL_LP, RREG32(MC_SEQ_G5PDX_CTRL)); + WREG32(MC_SEQ_PMG_DVS_CMD_LP, RREG32(MC_SEQ_PMG_DVS_CMD)); + WREG32(MC_SEQ_PMG_DVS_CTL_LP, RREG32(MC_SEQ_PMG_DVS_CTL)); + WREG32(MC_SEQ_MISC_TIMING_LP, RREG32(MC_SEQ_MISC_TIMING)); + WREG32(MC_SEQ_MISC_TIMING2_LP, RREG32(MC_SEQ_MISC_TIMING2)); + WREG32(MC_SEQ_PMG_CMD_EMRS_LP, RREG32(MC_PMG_CMD_EMRS)); + WREG32(MC_SEQ_PMG_CMD_MRS_LP, RREG32(MC_PMG_CMD_MRS)); + WREG32(MC_SEQ_PMG_CMD_MRS1_LP, RREG32(MC_PMG_CMD_MRS1)); + WREG32(MC_SEQ_WR_CTL_D0_LP, RREG32(MC_SEQ_WR_CTL_D0)); + WREG32(MC_SEQ_WR_CTL_D1_LP, RREG32(MC_SEQ_WR_CTL_D1)); + WREG32(MC_SEQ_RD_CTL_D0_LP, RREG32(MC_SEQ_RD_CTL_D0)); + WREG32(MC_SEQ_RD_CTL_D1_LP, RREG32(MC_SEQ_RD_CTL_D1)); + WREG32(MC_SEQ_PMG_TIMING_LP, RREG32(MC_SEQ_PMG_TIMING)); + WREG32(MC_SEQ_PMG_CMD_MRS2_LP, RREG32(MC_PMG_CMD_MRS2)); + WREG32(MC_SEQ_WR_CTL_2_LP, RREG32(MC_SEQ_WR_CTL_2)); + + ret = radeon_atom_init_mc_reg_table(rdev, module_index, table); + if (ret) + goto init_mc_done; + + ret = ci_copy_vbios_mc_reg_table(table, ci_table); + if (ret) + goto init_mc_done; + + ci_set_s0_mc_reg_index(ci_table); + + ret = ci_set_mc_special_registers(rdev, ci_table); + if (ret) + goto init_mc_done; + + ci_set_valid_flag(ci_table); + +init_mc_done: + kfree(table); + + return ret; +} + +static int ci_populate_mc_reg_addresses(struct radeon_device *rdev, + SMU7_Discrete_MCRegisters *mc_reg_table) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 i, j; + + for (i = 0, j = 0; j < pi->mc_reg_table.last; j++) { + if (pi->mc_reg_table.valid_flag & (1 << j)) { + if (i >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) + return -EINVAL; + mc_reg_table->address[i].s0 = cpu_to_be16(pi->mc_reg_table.mc_reg_address[j].s0); + mc_reg_table->address[i].s1 = cpu_to_be16(pi->mc_reg_table.mc_reg_address[j].s1); + i++; + } + } + + mc_reg_table->last = (u8)i; + + return 0; +} + +static void ci_convert_mc_registers(const struct ci_mc_reg_entry *entry, + SMU7_Discrete_MCRegisterSet *data, + u32 num_entries, u32 valid_flag) +{ + u32 i, j; + + for (i = 0, j = 0; j < num_entries; j++) { + if (valid_flag & (1 << j)) { + data->value[i] = cpu_to_be32(entry->mc_data[j]); + i++; + } + } +} + +static void ci_convert_mc_reg_table_entry_to_smc(struct radeon_device *rdev, + const u32 memory_clock, + SMU7_Discrete_MCRegisterSet *mc_reg_table_data) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 i = 0; + + for(i = 0; i < pi->mc_reg_table.num_entries; i++) { + if (memory_clock <= pi->mc_reg_table.mc_reg_table_entry[i].mclk_max) + break; + } + + if ((i == pi->mc_reg_table.num_entries) && (i > 0)) + --i; + + ci_convert_mc_registers(&pi->mc_reg_table.mc_reg_table_entry[i], + mc_reg_table_data, pi->mc_reg_table.last, + pi->mc_reg_table.valid_flag); +} + +static void ci_convert_mc_reg_table_to_smc(struct radeon_device *rdev, + SMU7_Discrete_MCRegisters *mc_reg_table) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 i; + + for (i = 0; i < pi->dpm_table.mclk_table.count; i++) + ci_convert_mc_reg_table_entry_to_smc(rdev, + pi->dpm_table.mclk_table.dpm_levels[i].value, + &mc_reg_table->data[i]); +} + +static int ci_populate_initial_mc_reg_table(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + int ret; + + memset(&pi->smc_mc_reg_table, 0, sizeof(SMU7_Discrete_MCRegisters)); + + ret = ci_populate_mc_reg_addresses(rdev, &pi->smc_mc_reg_table); + if (ret) + return ret; + ci_convert_mc_reg_table_to_smc(rdev, &pi->smc_mc_reg_table); + + return ci_copy_bytes_to_smc(rdev, + pi->mc_reg_table_start, + (u8 *)&pi->smc_mc_reg_table, + sizeof(SMU7_Discrete_MCRegisters), + pi->sram_end); +} + +static int ci_update_and_upload_mc_reg_table(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + + if (!(pi->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) + return 0; + + memset(&pi->smc_mc_reg_table, 0, sizeof(SMU7_Discrete_MCRegisters)); + + ci_convert_mc_reg_table_to_smc(rdev, &pi->smc_mc_reg_table); + + return ci_copy_bytes_to_smc(rdev, + pi->mc_reg_table_start + + offsetof(SMU7_Discrete_MCRegisters, data[0]), + (u8 *)&pi->smc_mc_reg_table.data[0], + sizeof(SMU7_Discrete_MCRegisterSet) * + pi->dpm_table.mclk_table.count, + pi->sram_end); +} + +static void ci_enable_voltage_control(struct radeon_device *rdev) +{ + u32 tmp = RREG32_SMC(GENERAL_PWRMGT); + + tmp |= VOLT_PWRMGT_EN; + WREG32_SMC(GENERAL_PWRMGT, tmp); +} + +static enum radeon_pcie_gen ci_get_maximum_link_speed(struct radeon_device *rdev, + struct radeon_ps *radeon_state) +{ + struct ci_ps *state = ci_get_ps(radeon_state); + int i; + u16 pcie_speed, max_speed = 0; + + for (i = 0; i < state->performance_level_count; i++) { + pcie_speed = state->performance_levels[i].pcie_gen; + if (max_speed < pcie_speed) + max_speed = pcie_speed; + } + + return max_speed; +} + +static u16 ci_get_current_pcie_speed(struct radeon_device *rdev) +{ + u32 speed_cntl = 0; + + speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL) & LC_CURRENT_DATA_RATE_MASK; + speed_cntl >>= LC_CURRENT_DATA_RATE_SHIFT; + + return (u16)speed_cntl; +} + +static int ci_get_current_pcie_lane_number(struct radeon_device *rdev) +{ + u32 link_width = 0; + + link_width = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL) & LC_LINK_WIDTH_RD_MASK; + link_width >>= LC_LINK_WIDTH_RD_SHIFT; + + switch (link_width) { + case RADEON_PCIE_LC_LINK_WIDTH_X1: + return 1; + case RADEON_PCIE_LC_LINK_WIDTH_X2: + return 2; + case RADEON_PCIE_LC_LINK_WIDTH_X4: + return 4; + case RADEON_PCIE_LC_LINK_WIDTH_X8: + return 8; + case RADEON_PCIE_LC_LINK_WIDTH_X12: + /* not actually supported */ + return 12; + case RADEON_PCIE_LC_LINK_WIDTH_X0: + case RADEON_PCIE_LC_LINK_WIDTH_X16: + default: + return 16; + } +} + +static void ci_request_link_speed_change_before_state_change(struct radeon_device *rdev, + struct radeon_ps *radeon_new_state, + struct radeon_ps *radeon_current_state) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + enum radeon_pcie_gen target_link_speed = + ci_get_maximum_link_speed(rdev, radeon_new_state); + enum radeon_pcie_gen current_link_speed; + + if (pi->force_pcie_gen == RADEON_PCIE_GEN_INVALID) + current_link_speed = ci_get_maximum_link_speed(rdev, radeon_current_state); + else + current_link_speed = pi->force_pcie_gen; + + pi->force_pcie_gen = RADEON_PCIE_GEN_INVALID; + pi->pspp_notify_required = false; + if (target_link_speed > current_link_speed) { + switch (target_link_speed) { +#ifdef CONFIG_ACPI + case RADEON_PCIE_GEN3: + if (radeon_acpi_pcie_performance_request(rdev, PCIE_PERF_REQ_PECI_GEN3, false) == 0) + break; + pi->force_pcie_gen = RADEON_PCIE_GEN2; + if (current_link_speed == RADEON_PCIE_GEN2) + break; + case RADEON_PCIE_GEN2: + if (radeon_acpi_pcie_performance_request(rdev, PCIE_PERF_REQ_PECI_GEN2, false) == 0) + break; +#endif + default: + pi->force_pcie_gen = ci_get_current_pcie_speed(rdev); + break; + } + } else { + if (target_link_speed < current_link_speed) + pi->pspp_notify_required = true; + } +} + +static void ci_notify_link_speed_change_after_state_change(struct radeon_device *rdev, + struct radeon_ps *radeon_new_state, + struct radeon_ps *radeon_current_state) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + enum radeon_pcie_gen target_link_speed = + ci_get_maximum_link_speed(rdev, radeon_new_state); + u8 request; + + if (pi->pspp_notify_required) { + if (target_link_speed == RADEON_PCIE_GEN3) + request = PCIE_PERF_REQ_PECI_GEN3; + else if (target_link_speed == RADEON_PCIE_GEN2) + request = PCIE_PERF_REQ_PECI_GEN2; + else + request = PCIE_PERF_REQ_PECI_GEN1; + + if ((request == PCIE_PERF_REQ_PECI_GEN1) && + (ci_get_current_pcie_speed(rdev) > 0)) + return; + +#ifdef CONFIG_ACPI + radeon_acpi_pcie_performance_request(rdev, request, false); +#endif + } +} + +static int ci_set_private_data_variables_based_on_pptable(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct radeon_clock_voltage_dependency_table *allowed_sclk_vddc_table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + struct radeon_clock_voltage_dependency_table *allowed_mclk_vddc_table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk; + struct radeon_clock_voltage_dependency_table *allowed_mclk_vddci_table = + &rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk; + + if (allowed_sclk_vddc_table == NULL) + return -EINVAL; + if (allowed_sclk_vddc_table->count < 1) + return -EINVAL; + if (allowed_mclk_vddc_table == NULL) + return -EINVAL; + if (allowed_mclk_vddc_table->count < 1) + return -EINVAL; + if (allowed_mclk_vddci_table == NULL) + return -EINVAL; + if (allowed_mclk_vddci_table->count < 1) + return -EINVAL; + + pi->min_vddc_in_pp_table = allowed_sclk_vddc_table->entries[0].v; + pi->max_vddc_in_pp_table = + allowed_sclk_vddc_table->entries[allowed_sclk_vddc_table->count - 1].v; + + pi->min_vddci_in_pp_table = allowed_mclk_vddci_table->entries[0].v; + pi->max_vddci_in_pp_table = + allowed_mclk_vddci_table->entries[allowed_mclk_vddci_table->count - 1].v; + + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk = + allowed_sclk_vddc_table->entries[allowed_sclk_vddc_table->count - 1].clk; + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.mclk = + allowed_mclk_vddc_table->entries[allowed_sclk_vddc_table->count - 1].clk; + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddc = + allowed_sclk_vddc_table->entries[allowed_sclk_vddc_table->count - 1].v; + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddci = + allowed_mclk_vddci_table->entries[allowed_mclk_vddci_table->count - 1].v; + + return 0; +} + +static void ci_patch_with_vddc_leakage(struct radeon_device *rdev, u16 *vddc) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_leakage_voltage *leakage_table = &pi->vddc_leakage; + u32 leakage_index; + + for (leakage_index = 0; leakage_index < leakage_table->count; leakage_index++) { + if (leakage_table->leakage_id[leakage_index] == *vddc) { + *vddc = leakage_table->actual_voltage[leakage_index]; + break; + } + } +} + +static void ci_patch_with_vddci_leakage(struct radeon_device *rdev, u16 *vddci) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_leakage_voltage *leakage_table = &pi->vddci_leakage; + u32 leakage_index; + + for (leakage_index = 0; leakage_index < leakage_table->count; leakage_index++) { + if (leakage_table->leakage_id[leakage_index] == *vddci) { + *vddci = leakage_table->actual_voltage[leakage_index]; + break; + } + } +} + +static void ci_patch_clock_voltage_dependency_table_with_vddc_leakage(struct radeon_device *rdev, + struct radeon_clock_voltage_dependency_table *table) +{ + u32 i; + + if (table) { + for (i = 0; i < table->count; i++) + ci_patch_with_vddc_leakage(rdev, &table->entries[i].v); + } +} + +static void ci_patch_clock_voltage_dependency_table_with_vddci_leakage(struct radeon_device *rdev, + struct radeon_clock_voltage_dependency_table *table) +{ + u32 i; + + if (table) { + for (i = 0; i < table->count; i++) + ci_patch_with_vddci_leakage(rdev, &table->entries[i].v); + } +} + +static void ci_patch_vce_clock_voltage_dependency_table_with_vddc_leakage(struct radeon_device *rdev, + struct radeon_vce_clock_voltage_dependency_table *table) +{ + u32 i; + + if (table) { + for (i = 0; i < table->count; i++) + ci_patch_with_vddc_leakage(rdev, &table->entries[i].v); + } +} + +static void ci_patch_uvd_clock_voltage_dependency_table_with_vddc_leakage(struct radeon_device *rdev, + struct radeon_uvd_clock_voltage_dependency_table *table) +{ + u32 i; + + if (table) { + for (i = 0; i < table->count; i++) + ci_patch_with_vddc_leakage(rdev, &table->entries[i].v); + } +} + +static void ci_patch_vddc_phase_shed_limit_table_with_vddc_leakage(struct radeon_device *rdev, + struct radeon_phase_shedding_limits_table *table) +{ + u32 i; + + if (table) { + for (i = 0; i < table->count; i++) + ci_patch_with_vddc_leakage(rdev, &table->entries[i].voltage); + } +} + +static void ci_patch_clock_voltage_limits_with_vddc_leakage(struct radeon_device *rdev, + struct radeon_clock_and_voltage_limits *table) +{ + if (table) { + ci_patch_with_vddc_leakage(rdev, (u16 *)&table->vddc); + ci_patch_with_vddci_leakage(rdev, (u16 *)&table->vddci); + } +} + +static void ci_patch_cac_leakage_table_with_vddc_leakage(struct radeon_device *rdev, + struct radeon_cac_leakage_table *table) +{ + u32 i; + + if (table) { + for (i = 0; i < table->count; i++) + ci_patch_with_vddc_leakage(rdev, &table->entries[i].vddc); + } +} + +static void ci_patch_dependency_tables_with_leakage(struct radeon_device *rdev) +{ + + ci_patch_clock_voltage_dependency_table_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk); + ci_patch_clock_voltage_dependency_table_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk); + ci_patch_clock_voltage_dependency_table_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk); + ci_patch_clock_voltage_dependency_table_with_vddci_leakage(rdev, + &rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk); + ci_patch_vce_clock_voltage_dependency_table_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table); + ci_patch_uvd_clock_voltage_dependency_table_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table); + ci_patch_clock_voltage_dependency_table_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table); + ci_patch_clock_voltage_dependency_table_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table); + ci_patch_vddc_phase_shed_limit_table_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.phase_shedding_limits_table); + ci_patch_clock_voltage_limits_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac); + ci_patch_clock_voltage_limits_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc); + ci_patch_cac_leakage_table_with_vddc_leakage(rdev, + &rdev->pm.dpm.dyn_state.cac_leakage_table); + +} + +static void ci_get_memory_type(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + u32 tmp; + + tmp = RREG32(MC_SEQ_MISC0); + + if (((tmp & MC_SEQ_MISC0_GDDR5_MASK) >> MC_SEQ_MISC0_GDDR5_SHIFT) == + MC_SEQ_MISC0_GDDR5_VALUE) + pi->mem_gddr5 = true; + else + pi->mem_gddr5 = false; + +} + +static void ci_update_current_ps(struct radeon_device *rdev, + struct radeon_ps *rps) +{ + struct ci_ps *new_ps = ci_get_ps(rps); + struct ci_power_info *pi = ci_get_pi(rdev); + + pi->current_rps = *rps; + pi->current_ps = *new_ps; + pi->current_rps.ps_priv = &pi->current_ps; +} + +static void ci_update_requested_ps(struct radeon_device *rdev, + struct radeon_ps *rps) +{ + struct ci_ps *new_ps = ci_get_ps(rps); + struct ci_power_info *pi = ci_get_pi(rdev); + + pi->requested_rps = *rps; + pi->requested_ps = *new_ps; + pi->requested_rps.ps_priv = &pi->requested_ps; +} + +int ci_dpm_pre_set_power_state(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps; + struct radeon_ps *new_ps = &requested_ps; + + ci_update_requested_ps(rdev, new_ps); + + ci_apply_state_adjust_rules(rdev, &pi->requested_rps); + + return 0; +} + +void ci_dpm_post_set_power_state(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct radeon_ps *new_ps = &pi->requested_rps; + + ci_update_current_ps(rdev, new_ps); +} + + +void ci_dpm_setup_asic(struct radeon_device *rdev) +{ + int r; + + r = ci_mc_load_microcode(rdev); + if (r) + DRM_ERROR("Failed to load MC firmware!\n"); + ci_read_clock_registers(rdev); + ci_get_memory_type(rdev); + ci_enable_acpi_power_management(rdev); + ci_init_sclk_t(rdev); +} + +int ci_dpm_enable(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps; + int ret; + + if (ci_is_smc_running(rdev)) + return -EINVAL; + if (pi->voltage_control != CISLANDS_VOLTAGE_CONTROL_NONE) { + ci_enable_voltage_control(rdev); + ret = ci_construct_voltage_tables(rdev); + if (ret) { + DRM_ERROR("ci_construct_voltage_tables failed\n"); + return ret; + } + } + if (pi->caps_dynamic_ac_timing) { + ret = ci_initialize_mc_reg_table(rdev); + if (ret) + pi->caps_dynamic_ac_timing = false; + } + if (pi->dynamic_ss) + ci_enable_spread_spectrum(rdev, true); + if (pi->thermal_protection) + ci_enable_thermal_protection(rdev, true); + ci_program_sstp(rdev); + ci_enable_display_gap(rdev); + ci_program_vc(rdev); + ret = ci_upload_firmware(rdev); + if (ret) { + DRM_ERROR("ci_upload_firmware failed\n"); + return ret; + } + ret = ci_process_firmware_header(rdev); + if (ret) { + DRM_ERROR("ci_process_firmware_header failed\n"); + return ret; + } + ret = ci_initial_switch_from_arb_f0_to_f1(rdev); + if (ret) { + DRM_ERROR("ci_initial_switch_from_arb_f0_to_f1 failed\n"); + return ret; + } + ret = ci_init_smc_table(rdev); + if (ret) { + DRM_ERROR("ci_init_smc_table failed\n"); + return ret; + } + ret = ci_init_arb_table_index(rdev); + if (ret) { + DRM_ERROR("ci_init_arb_table_index failed\n"); + return ret; + } + if (pi->caps_dynamic_ac_timing) { + ret = ci_populate_initial_mc_reg_table(rdev); + if (ret) { + DRM_ERROR("ci_populate_initial_mc_reg_table failed\n"); + return ret; + } + } + ret = ci_populate_pm_base(rdev); + if (ret) { + DRM_ERROR("ci_populate_pm_base failed\n"); + return ret; + } + ci_dpm_start_smc(rdev); + ci_enable_vr_hot_gpio_interrupt(rdev); + ret = ci_notify_smc_display_change(rdev, false); + if (ret) { + DRM_ERROR("ci_notify_smc_display_change failed\n"); + return ret; + } + ci_enable_sclk_control(rdev, true); + ret = ci_enable_ulv(rdev, true); + if (ret) { + DRM_ERROR("ci_enable_ulv failed\n"); + return ret; + } + ret = ci_enable_ds_master_switch(rdev, true); + if (ret) { + DRM_ERROR("ci_enable_ds_master_switch failed\n"); + return ret; + } + ret = ci_start_dpm(rdev); + if (ret) { + DRM_ERROR("ci_start_dpm failed\n"); + return ret; + } + ret = ci_enable_didt(rdev, true); + if (ret) { + DRM_ERROR("ci_enable_didt failed\n"); + return ret; + } + ret = ci_enable_smc_cac(rdev, true); + if (ret) { + DRM_ERROR("ci_enable_smc_cac failed\n"); + return ret; + } + ret = ci_enable_power_containment(rdev, true); + if (ret) { + DRM_ERROR("ci_enable_power_containment failed\n"); + return ret; + } + + ci_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true); + + ci_update_current_ps(rdev, boot_ps); + + return 0; +} + +int ci_dpm_late_enable(struct radeon_device *rdev) +{ + int ret; + + if (rdev->irq.installed && + r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { +#if 0 + PPSMC_Result result; +#endif + ret = ci_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX); + if (ret) { + DRM_ERROR("ci_set_thermal_temperature_range failed\n"); + return ret; + } + rdev->irq.dpm_thermal = true; + radeon_irq_set(rdev); +#if 0 + result = ci_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt); + + if (result != PPSMC_Result_OK) + DRM_DEBUG_KMS("Could not enable thermal interrupts.\n"); +#endif + } + + ci_dpm_powergate_uvd(rdev, true); + + return 0; +} + +void ci_dpm_disable(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps; + + ci_dpm_powergate_uvd(rdev, false); + + if (!ci_is_smc_running(rdev)) + return; + + if (pi->thermal_protection) + ci_enable_thermal_protection(rdev, false); + ci_enable_power_containment(rdev, false); + ci_enable_smc_cac(rdev, false); + ci_enable_didt(rdev, false); + ci_enable_spread_spectrum(rdev, false); + ci_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, false); + ci_stop_dpm(rdev); + ci_enable_ds_master_switch(rdev, true); + ci_enable_ulv(rdev, false); + ci_clear_vc(rdev); + ci_reset_to_default(rdev); + ci_dpm_stop_smc(rdev); + ci_force_switch_to_arb_f0(rdev); + + ci_update_current_ps(rdev, boot_ps); +} + +int ci_dpm_set_power_state(struct radeon_device *rdev) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct radeon_ps *new_ps = &pi->requested_rps; + struct radeon_ps *old_ps = &pi->current_rps; + int ret; + + ci_find_dpm_states_clocks_in_dpm_table(rdev, new_ps); + if (pi->pcie_performance_request) + ci_request_link_speed_change_before_state_change(rdev, new_ps, old_ps); + ret = ci_freeze_sclk_mclk_dpm(rdev); + if (ret) { + DRM_ERROR("ci_freeze_sclk_mclk_dpm failed\n"); + return ret; + } + ret = ci_populate_and_upload_sclk_mclk_dpm_levels(rdev, new_ps); + if (ret) { + DRM_ERROR("ci_populate_and_upload_sclk_mclk_dpm_levels failed\n"); + return ret; + } + ret = ci_generate_dpm_level_enable_mask(rdev, new_ps); + if (ret) { + DRM_ERROR("ci_generate_dpm_level_enable_mask failed\n"); + return ret; + } + + ret = ci_update_vce_dpm(rdev, new_ps, old_ps); + if (ret) { + DRM_ERROR("ci_update_vce_dpm failed\n"); + return ret; + } + + ret = ci_update_sclk_t(rdev); + if (ret) { + DRM_ERROR("ci_update_sclk_t failed\n"); + return ret; + } + if (pi->caps_dynamic_ac_timing) { + ret = ci_update_and_upload_mc_reg_table(rdev); + if (ret) { + DRM_ERROR("ci_update_and_upload_mc_reg_table failed\n"); + return ret; + } + } + ret = ci_program_memory_timing_parameters(rdev); + if (ret) { + DRM_ERROR("ci_program_memory_timing_parameters failed\n"); + return ret; + } + ret = ci_unfreeze_sclk_mclk_dpm(rdev); + if (ret) { + DRM_ERROR("ci_unfreeze_sclk_mclk_dpm failed\n"); + return ret; + } + ret = ci_upload_dpm_level_enable_mask(rdev); + if (ret) { + DRM_ERROR("ci_upload_dpm_level_enable_mask failed\n"); + return ret; + } + if (pi->pcie_performance_request) + ci_notify_link_speed_change_after_state_change(rdev, new_ps, old_ps); + + return 0; +} + +int ci_dpm_power_control_set_level(struct radeon_device *rdev) +{ + return ci_power_control_set_level(rdev); +} + +void ci_dpm_reset_asic(struct radeon_device *rdev) +{ + ci_set_boot_state(rdev); +} + +void ci_dpm_display_configuration_changed(struct radeon_device *rdev) +{ + ci_program_display_gap(rdev); +} + +union power_info { + struct _ATOM_POWERPLAY_INFO info; + struct _ATOM_POWERPLAY_INFO_V2 info_2; + struct _ATOM_POWERPLAY_INFO_V3 info_3; + struct _ATOM_PPLIB_POWERPLAYTABLE pplib; + struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2; + struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3; +}; + +union pplib_clock_info { + struct _ATOM_PPLIB_R600_CLOCK_INFO r600; + struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780; + struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen; + struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo; + struct _ATOM_PPLIB_SI_CLOCK_INFO si; + struct _ATOM_PPLIB_CI_CLOCK_INFO ci; +}; + +union pplib_power_state { + struct _ATOM_PPLIB_STATE v1; + struct _ATOM_PPLIB_STATE_V2 v2; +}; + +static void ci_parse_pplib_non_clock_info(struct radeon_device *rdev, + struct radeon_ps *rps, + struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info, + u8 table_rev) +{ + rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings); + rps->class = le16_to_cpu(non_clock_info->usClassification); + rps->class2 = le16_to_cpu(non_clock_info->usClassification2); + + if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) { + rps->vclk = le32_to_cpu(non_clock_info->ulVCLK); + rps->dclk = le32_to_cpu(non_clock_info->ulDCLK); + } else { + rps->vclk = 0; + rps->dclk = 0; + } + + if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) + rdev->pm.dpm.boot_ps = rps; + if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE) + rdev->pm.dpm.uvd_ps = rps; +} + +static void ci_parse_pplib_clock_info(struct radeon_device *rdev, + struct radeon_ps *rps, int index, + union pplib_clock_info *clock_info) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_ps *ps = ci_get_ps(rps); + struct ci_pl *pl = &ps->performance_levels[index]; + + ps->performance_level_count = index + 1; + + pl->sclk = le16_to_cpu(clock_info->ci.usEngineClockLow); + pl->sclk |= clock_info->ci.ucEngineClockHigh << 16; + pl->mclk = le16_to_cpu(clock_info->ci.usMemoryClockLow); + pl->mclk |= clock_info->ci.ucMemoryClockHigh << 16; + + pl->pcie_gen = r600_get_pcie_gen_support(rdev, + pi->sys_pcie_mask, + pi->vbios_boot_state.pcie_gen_bootup_value, + clock_info->ci.ucPCIEGen); + pl->pcie_lane = r600_get_pcie_lane_support(rdev, + pi->vbios_boot_state.pcie_lane_bootup_value, + le16_to_cpu(clock_info->ci.usPCIELane)); + + if (rps->class & ATOM_PPLIB_CLASSIFICATION_ACPI) { + pi->acpi_pcie_gen = pl->pcie_gen; + } + + if (rps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV) { + pi->ulv.supported = true; + pi->ulv.pl = *pl; + pi->ulv.cg_ulv_parameter = CISLANDS_CGULVPARAMETER_DFLT; + } + + /* patch up boot state */ + if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) { + pl->mclk = pi->vbios_boot_state.mclk_bootup_value; + pl->sclk = pi->vbios_boot_state.sclk_bootup_value; + pl->pcie_gen = pi->vbios_boot_state.pcie_gen_bootup_value; + pl->pcie_lane = pi->vbios_boot_state.pcie_lane_bootup_value; + } + + switch (rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) { + case ATOM_PPLIB_CLASSIFICATION_UI_BATTERY: + pi->use_pcie_powersaving_levels = true; + if (pi->pcie_gen_powersaving.max < pl->pcie_gen) + pi->pcie_gen_powersaving.max = pl->pcie_gen; + if (pi->pcie_gen_powersaving.min > pl->pcie_gen) + pi->pcie_gen_powersaving.min = pl->pcie_gen; + if (pi->pcie_lane_powersaving.max < pl->pcie_lane) + pi->pcie_lane_powersaving.max = pl->pcie_lane; + if (pi->pcie_lane_powersaving.min > pl->pcie_lane) + pi->pcie_lane_powersaving.min = pl->pcie_lane; + break; + case ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE: + pi->use_pcie_performance_levels = true; + if (pi->pcie_gen_performance.max < pl->pcie_gen) + pi->pcie_gen_performance.max = pl->pcie_gen; + if (pi->pcie_gen_performance.min > pl->pcie_gen) + pi->pcie_gen_performance.min = pl->pcie_gen; + if (pi->pcie_lane_performance.max < pl->pcie_lane) + pi->pcie_lane_performance.max = pl->pcie_lane; + if (pi->pcie_lane_performance.min > pl->pcie_lane) + pi->pcie_lane_performance.min = pl->pcie_lane; + break; + default: + break; + } +} + +static int ci_parse_power_table(struct radeon_device *rdev) +{ + struct radeon_mode_info *mode_info = &rdev->mode_info; + struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info; + union pplib_power_state *power_state; + int i, j, k, non_clock_array_index, clock_array_index; + union pplib_clock_info *clock_info; + struct _StateArray *state_array; + struct _ClockInfoArray *clock_info_array; + struct _NonClockInfoArray *non_clock_info_array; + union power_info *power_info; + int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo); + u16 data_offset; + u8 frev, crev; + u8 *power_state_offset; + struct ci_ps *ps; + + if (!atom_parse_data_header(mode_info->atom_context, index, NULL, + &frev, &crev, &data_offset)) + return -EINVAL; + power_info = (union power_info *)(mode_info->atom_context->bios + data_offset); + + state_array = (struct _StateArray *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(power_info->pplib.usStateArrayOffset)); + clock_info_array = (struct _ClockInfoArray *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(power_info->pplib.usClockInfoArrayOffset)); + non_clock_info_array = (struct _NonClockInfoArray *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset)); + + rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) * + state_array->ucNumEntries, GFP_KERNEL); + if (!rdev->pm.dpm.ps) + return -ENOMEM; + power_state_offset = (u8 *)state_array->states; + for (i = 0; i < state_array->ucNumEntries; i++) { + u8 *idx; + power_state = (union pplib_power_state *)power_state_offset; + non_clock_array_index = power_state->v2.nonClockInfoIndex; + non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *) + &non_clock_info_array->nonClockInfo[non_clock_array_index]; + if (!rdev->pm.power_state[i].clock_info) + return -EINVAL; + ps = kzalloc(sizeof(struct ci_ps), GFP_KERNEL); + if (ps == NULL) { + kfree(rdev->pm.dpm.ps); + return -ENOMEM; + } + rdev->pm.dpm.ps[i].ps_priv = ps; + ci_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i], + non_clock_info, + non_clock_info_array->ucEntrySize); + k = 0; + idx = (u8 *)&power_state->v2.clockInfoIndex[0]; + for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) { + clock_array_index = idx[j]; + if (clock_array_index >= clock_info_array->ucNumEntries) + continue; + if (k >= CISLANDS_MAX_HARDWARE_POWERLEVELS) + break; + clock_info = (union pplib_clock_info *) + ((u8 *)&clock_info_array->clockInfo[0] + + (clock_array_index * clock_info_array->ucEntrySize)); + ci_parse_pplib_clock_info(rdev, + &rdev->pm.dpm.ps[i], k, + clock_info); + k++; + } + power_state_offset += 2 + power_state->v2.ucNumDPMLevels; + } + rdev->pm.dpm.num_ps = state_array->ucNumEntries; + + /* fill in the vce power states */ + for (i = 0; i < RADEON_MAX_VCE_LEVELS; i++) { + u32 sclk, mclk; + clock_array_index = rdev->pm.dpm.vce_states[i].clk_idx; + clock_info = (union pplib_clock_info *) + &clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize]; + sclk = le16_to_cpu(clock_info->ci.usEngineClockLow); + sclk |= clock_info->ci.ucEngineClockHigh << 16; + mclk = le16_to_cpu(clock_info->ci.usMemoryClockLow); + mclk |= clock_info->ci.ucMemoryClockHigh << 16; + rdev->pm.dpm.vce_states[i].sclk = sclk; + rdev->pm.dpm.vce_states[i].mclk = mclk; + } + + return 0; +} + +static int ci_get_vbios_boot_values(struct radeon_device *rdev, + struct ci_vbios_boot_state *boot_state) +{ + struct radeon_mode_info *mode_info = &rdev->mode_info; + int index = GetIndexIntoMasterTable(DATA, FirmwareInfo); + ATOM_FIRMWARE_INFO_V2_2 *firmware_info; + u8 frev, crev; + u16 data_offset; + + if (atom_parse_data_header(mode_info->atom_context, index, NULL, + &frev, &crev, &data_offset)) { + firmware_info = + (ATOM_FIRMWARE_INFO_V2_2 *)(mode_info->atom_context->bios + + data_offset); + boot_state->mvdd_bootup_value = le16_to_cpu(firmware_info->usBootUpMVDDCVoltage); + boot_state->vddc_bootup_value = le16_to_cpu(firmware_info->usBootUpVDDCVoltage); + boot_state->vddci_bootup_value = le16_to_cpu(firmware_info->usBootUpVDDCIVoltage); + boot_state->pcie_gen_bootup_value = ci_get_current_pcie_speed(rdev); + boot_state->pcie_lane_bootup_value = ci_get_current_pcie_lane_number(rdev); + boot_state->sclk_bootup_value = le32_to_cpu(firmware_info->ulDefaultEngineClock); + boot_state->mclk_bootup_value = le32_to_cpu(firmware_info->ulDefaultMemoryClock); + + return 0; + } + return -EINVAL; +} + +void ci_dpm_fini(struct radeon_device *rdev) +{ + int i; + + for (i = 0; i < rdev->pm.dpm.num_ps; i++) { + kfree(rdev->pm.dpm.ps[i].ps_priv); + } + kfree(rdev->pm.dpm.ps); + kfree(rdev->pm.dpm.priv); + kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries); + r600_free_extended_power_table(rdev); +} + +int ci_dpm_init(struct radeon_device *rdev) +{ + int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info); + u16 data_offset, size; + u8 frev, crev; + struct ci_power_info *pi; + int ret; + u32 mask; + + pi = kzalloc(sizeof(struct ci_power_info), GFP_KERNEL); + if (pi == NULL) + return -ENOMEM; + rdev->pm.dpm.priv = pi; + + ret = drm_pcie_get_speed_cap_mask(rdev->ddev, &mask); + if (ret) + pi->sys_pcie_mask = 0; + else + pi->sys_pcie_mask = mask; + pi->force_pcie_gen = RADEON_PCIE_GEN_INVALID; + + pi->pcie_gen_performance.max = RADEON_PCIE_GEN1; + pi->pcie_gen_performance.min = RADEON_PCIE_GEN3; + pi->pcie_gen_powersaving.max = RADEON_PCIE_GEN1; + pi->pcie_gen_powersaving.min = RADEON_PCIE_GEN3; + + pi->pcie_lane_performance.max = 0; + pi->pcie_lane_performance.min = 16; + pi->pcie_lane_powersaving.max = 0; + pi->pcie_lane_powersaving.min = 16; + + ret = ci_get_vbios_boot_values(rdev, &pi->vbios_boot_state); + if (ret) { + ci_dpm_fini(rdev); + return ret; + } + + ret = r600_get_platform_caps(rdev); + if (ret) { + ci_dpm_fini(rdev); + return ret; + } + + ret = r600_parse_extended_power_table(rdev); + if (ret) { + ci_dpm_fini(rdev); + return ret; + } + + ret = ci_parse_power_table(rdev); + if (ret) { + ci_dpm_fini(rdev); + return ret; + } + + pi->dll_default_on = false; + pi->sram_end = SMC_RAM_END; + + pi->activity_target[0] = CISLAND_TARGETACTIVITY_DFLT; + pi->activity_target[1] = CISLAND_TARGETACTIVITY_DFLT; + pi->activity_target[2] = CISLAND_TARGETACTIVITY_DFLT; + pi->activity_target[3] = CISLAND_TARGETACTIVITY_DFLT; + pi->activity_target[4] = CISLAND_TARGETACTIVITY_DFLT; + pi->activity_target[5] = CISLAND_TARGETACTIVITY_DFLT; + pi->activity_target[6] = CISLAND_TARGETACTIVITY_DFLT; + pi->activity_target[7] = CISLAND_TARGETACTIVITY_DFLT; + + pi->mclk_activity_target = CISLAND_MCLK_TARGETACTIVITY_DFLT; + + pi->sclk_dpm_key_disabled = 0; + pi->mclk_dpm_key_disabled = 0; + pi->pcie_dpm_key_disabled = 0; + + /* mclk dpm is unstable on some R7 260X cards with the old mc ucode */ + if ((rdev->pdev->device == 0x6658) && + (rdev->mc_fw->size == (BONAIRE_MC_UCODE_SIZE * 4))) { + pi->mclk_dpm_key_disabled = 1; + } + + pi->caps_sclk_ds = true; + + pi->mclk_strobe_mode_threshold = 40000; + pi->mclk_stutter_mode_threshold = 40000; + pi->mclk_edc_enable_threshold = 40000; + pi->mclk_edc_wr_enable_threshold = 40000; + + ci_initialize_powertune_defaults(rdev); + + pi->caps_fps = false; + + pi->caps_sclk_throttle_low_notification = false; + + pi->caps_uvd_dpm = true; + pi->caps_vce_dpm = true; + + ci_get_leakage_voltages(rdev); + ci_patch_dependency_tables_with_leakage(rdev); + ci_set_private_data_variables_based_on_pptable(rdev); + + rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries = + kzalloc(4 * sizeof(struct radeon_clock_voltage_dependency_entry), GFP_KERNEL); + if (!rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries) { + ci_dpm_fini(rdev); + return -ENOMEM; + } + rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.count = 4; + rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].clk = 0; + rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].v = 0; + rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].clk = 36000; + rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].v = 720; + rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].clk = 54000; + rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].v = 810; + rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].clk = 72000; + rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].v = 900; + + rdev->pm.dpm.dyn_state.mclk_sclk_ratio = 4; + rdev->pm.dpm.dyn_state.sclk_mclk_delta = 15000; + rdev->pm.dpm.dyn_state.vddc_vddci_delta = 200; + + rdev->pm.dpm.dyn_state.valid_sclk_values.count = 0; + rdev->pm.dpm.dyn_state.valid_sclk_values.values = NULL; + rdev->pm.dpm.dyn_state.valid_mclk_values.count = 0; + rdev->pm.dpm.dyn_state.valid_mclk_values.values = NULL; + + if (rdev->family == CHIP_HAWAII) { + pi->thermal_temp_setting.temperature_low = 94500; + pi->thermal_temp_setting.temperature_high = 95000; + pi->thermal_temp_setting.temperature_shutdown = 104000; + } else { + pi->thermal_temp_setting.temperature_low = 99500; + pi->thermal_temp_setting.temperature_high = 100000; + pi->thermal_temp_setting.temperature_shutdown = 104000; + } + + pi->uvd_enabled = false; + + pi->voltage_control = CISLANDS_VOLTAGE_CONTROL_NONE; + pi->vddci_control = CISLANDS_VOLTAGE_CONTROL_NONE; + pi->mvdd_control = CISLANDS_VOLTAGE_CONTROL_NONE; + if (radeon_atom_is_voltage_gpio(rdev, VOLTAGE_TYPE_VDDC, VOLTAGE_OBJ_GPIO_LUT)) + pi->voltage_control = CISLANDS_VOLTAGE_CONTROL_BY_GPIO; + else if (radeon_atom_is_voltage_gpio(rdev, VOLTAGE_TYPE_VDDC, VOLTAGE_OBJ_SVID2)) + pi->voltage_control = CISLANDS_VOLTAGE_CONTROL_BY_SVID2; + + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_VDDCI_CONTROL) { + if (radeon_atom_is_voltage_gpio(rdev, VOLTAGE_TYPE_VDDCI, VOLTAGE_OBJ_GPIO_LUT)) + pi->vddci_control = CISLANDS_VOLTAGE_CONTROL_BY_GPIO; + else if (radeon_atom_is_voltage_gpio(rdev, VOLTAGE_TYPE_VDDCI, VOLTAGE_OBJ_SVID2)) + pi->vddci_control = CISLANDS_VOLTAGE_CONTROL_BY_SVID2; + else + rdev->pm.dpm.platform_caps &= ~ATOM_PP_PLATFORM_CAP_VDDCI_CONTROL; + } + + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_MVDDCONTROL) { + if (radeon_atom_is_voltage_gpio(rdev, VOLTAGE_TYPE_MVDDC, VOLTAGE_OBJ_GPIO_LUT)) + pi->mvdd_control = CISLANDS_VOLTAGE_CONTROL_BY_GPIO; + else if (radeon_atom_is_voltage_gpio(rdev, VOLTAGE_TYPE_MVDDC, VOLTAGE_OBJ_SVID2)) + pi->mvdd_control = CISLANDS_VOLTAGE_CONTROL_BY_SVID2; + else + rdev->pm.dpm.platform_caps &= ~ATOM_PP_PLATFORM_CAP_MVDDCONTROL; + } + + pi->vddc_phase_shed_control = true; + +#if defined(CONFIG_ACPI) + pi->pcie_performance_request = + radeon_acpi_is_pcie_performance_request_supported(rdev); +#else + pi->pcie_performance_request = false; +#endif + + if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size, + &frev, &crev, &data_offset)) { + pi->caps_sclk_ss_support = true; + pi->caps_mclk_ss_support = true; + pi->dynamic_ss = true; + } else { + pi->caps_sclk_ss_support = false; + pi->caps_mclk_ss_support = false; + pi->dynamic_ss = true; + } + + if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE) + pi->thermal_protection = true; + else + pi->thermal_protection = false; + + pi->caps_dynamic_ac_timing = true; + + pi->uvd_power_gated = false; + + /* make sure dc limits are valid */ + if ((rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.sclk == 0) || + (rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.mclk == 0)) + rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc = + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + + return 0; +} + +void ci_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, + struct seq_file *m) +{ + u32 sclk = ci_get_average_sclk_freq(rdev); + u32 mclk = ci_get_average_mclk_freq(rdev); + + seq_printf(m, "power level avg sclk: %u mclk: %u\n", + sclk, mclk); +} + +void ci_dpm_print_power_state(struct radeon_device *rdev, + struct radeon_ps *rps) +{ + struct ci_ps *ps = ci_get_ps(rps); + struct ci_pl *pl; + int i; + + r600_dpm_print_class_info(rps->class, rps->class2); + r600_dpm_print_cap_info(rps->caps); + printk("\tuvd vclk: %d dclk: %d\n", rps->vclk, rps->dclk); + for (i = 0; i < ps->performance_level_count; i++) { + pl = &ps->performance_levels[i]; + printk("\t\tpower level %d sclk: %u mclk: %u pcie gen: %u pcie lanes: %u\n", + i, pl->sclk, pl->mclk, pl->pcie_gen + 1, pl->pcie_lane); + } + r600_dpm_print_ps_status(rdev, rps); +} + +u32 ci_dpm_get_sclk(struct radeon_device *rdev, bool low) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_ps *requested_state = ci_get_ps(&pi->requested_rps); + + if (low) + return requested_state->performance_levels[0].sclk; + else + return requested_state->performance_levels[requested_state->performance_level_count - 1].sclk; +} + +u32 ci_dpm_get_mclk(struct radeon_device *rdev, bool low) +{ + struct ci_power_info *pi = ci_get_pi(rdev); + struct ci_ps *requested_state = ci_get_ps(&pi->requested_rps); + + if (low) + return requested_state->performance_levels[0].mclk; + else + return requested_state->performance_levels[requested_state->performance_level_count - 1].mclk; +} diff --git a/drivers/gpu/drm/radeon/ci_dpm.h b/drivers/gpu/drm/radeon/ci_dpm.h new file mode 100644 index 00000000000..93bbed977ff --- /dev/null +++ b/drivers/gpu/drm/radeon/ci_dpm.h @@ -0,0 +1,332 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#ifndef __CI_DPM_H__ +#define __CI_DPM_H__ + +#include "ppsmc.h" + +#define SMU__NUM_SCLK_DPM_STATE 8 +#define SMU__NUM_MCLK_DPM_LEVELS 6 +#define SMU__NUM_LCLK_DPM_LEVELS 8 +#define SMU__NUM_PCIE_DPM_LEVELS 8 +#include "smu7_discrete.h" + +#define CISLANDS_MAX_HARDWARE_POWERLEVELS 2 + +struct ci_pl { + u32 mclk; + u32 sclk; + enum radeon_pcie_gen pcie_gen; + u16 pcie_lane; +}; + +struct ci_ps { + u16 performance_level_count; + bool dc_compatible; + u32 sclk_t; + struct ci_pl performance_levels[CISLANDS_MAX_HARDWARE_POWERLEVELS]; +}; + +struct ci_dpm_level { + bool enabled; + u32 value; + u32 param1; +}; + +#define CISLAND_MAX_DEEPSLEEP_DIVIDER_ID 5 +#define MAX_REGULAR_DPM_NUMBER 8 +#define CISLAND_MINIMUM_ENGINE_CLOCK 800 + +struct ci_single_dpm_table { + u32 count; + struct ci_dpm_level dpm_levels[MAX_REGULAR_DPM_NUMBER]; +}; + +struct ci_dpm_table { + struct ci_single_dpm_table sclk_table; + struct ci_single_dpm_table mclk_table; + struct ci_single_dpm_table pcie_speed_table; + struct ci_single_dpm_table vddc_table; + struct ci_single_dpm_table vddci_table; + struct ci_single_dpm_table mvdd_table; +}; + +struct ci_mc_reg_entry { + u32 mclk_max; + u32 mc_data[SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE]; +}; + +struct ci_mc_reg_table { + u8 last; + u8 num_entries; + u16 valid_flag; + struct ci_mc_reg_entry mc_reg_table_entry[MAX_AC_TIMING_ENTRIES]; + SMU7_Discrete_MCRegisterAddress mc_reg_address[SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE]; +}; + +struct ci_ulv_parm +{ + bool supported; + u32 cg_ulv_parameter; + u32 volt_change_delay; + struct ci_pl pl; +}; + +#define CISLANDS_MAX_LEAKAGE_COUNT 8 + +struct ci_leakage_voltage { + u16 count; + u16 leakage_id[CISLANDS_MAX_LEAKAGE_COUNT]; + u16 actual_voltage[CISLANDS_MAX_LEAKAGE_COUNT]; +}; + +struct ci_dpm_level_enable_mask { + u32 uvd_dpm_enable_mask; + u32 vce_dpm_enable_mask; + u32 acp_dpm_enable_mask; + u32 samu_dpm_enable_mask; + u32 sclk_dpm_enable_mask; + u32 mclk_dpm_enable_mask; + u32 pcie_dpm_enable_mask; +}; + +struct ci_vbios_boot_state +{ + u16 mvdd_bootup_value; + u16 vddc_bootup_value; + u16 vddci_bootup_value; + u32 sclk_bootup_value; + u32 mclk_bootup_value; + u16 pcie_gen_bootup_value; + u16 pcie_lane_bootup_value; +}; + +struct ci_clock_registers { + u32 cg_spll_func_cntl; + u32 cg_spll_func_cntl_2; + u32 cg_spll_func_cntl_3; + u32 cg_spll_func_cntl_4; + u32 cg_spll_spread_spectrum; + u32 cg_spll_spread_spectrum_2; + u32 dll_cntl; + u32 mclk_pwrmgt_cntl; + u32 mpll_ad_func_cntl; + u32 mpll_dq_func_cntl; + u32 mpll_func_cntl; + u32 mpll_func_cntl_1; + u32 mpll_func_cntl_2; + u32 mpll_ss1; + u32 mpll_ss2; +}; + +struct ci_thermal_temperature_setting { + s32 temperature_low; + s32 temperature_high; + s32 temperature_shutdown; +}; + +struct ci_pcie_perf_range { + u16 max; + u16 min; +}; + +enum ci_pt_config_reg_type { + CISLANDS_CONFIGREG_MMR = 0, + CISLANDS_CONFIGREG_SMC_IND, + CISLANDS_CONFIGREG_DIDT_IND, + CISLANDS_CONFIGREG_CACHE, + CISLANDS_CONFIGREG_MAX +}; + +#define POWERCONTAINMENT_FEATURE_BAPM 0x00000001 +#define POWERCONTAINMENT_FEATURE_TDCLimit 0x00000002 +#define POWERCONTAINMENT_FEATURE_PkgPwrLimit 0x00000004 + +struct ci_pt_config_reg { + u32 offset; + u32 mask; + u32 shift; + u32 value; + enum ci_pt_config_reg_type type; +}; + +struct ci_pt_defaults { + u8 svi_load_line_en; + u8 svi_load_line_vddc; + u8 tdc_vddc_throttle_release_limit_perc; + u8 tdc_mawt; + u8 tdc_waterfall_ctl; + u8 dte_ambient_temp_base; + u32 display_cac; + u32 bapm_temp_gradient; + u16 bapmti_r[SMU7_DTE_ITERATIONS * SMU7_DTE_SOURCES * SMU7_DTE_SINKS]; + u16 bapmti_rc[SMU7_DTE_ITERATIONS * SMU7_DTE_SOURCES * SMU7_DTE_SINKS]; +}; + +#define DPMTABLE_OD_UPDATE_SCLK 0x00000001 +#define DPMTABLE_OD_UPDATE_MCLK 0x00000002 +#define DPMTABLE_UPDATE_SCLK 0x00000004 +#define DPMTABLE_UPDATE_MCLK 0x00000008 + +struct ci_power_info { + struct ci_dpm_table dpm_table; + u32 voltage_control; + u32 mvdd_control; + u32 vddci_control; + u32 active_auto_throttle_sources; + struct ci_clock_registers clock_registers; + u16 acpi_vddc; + u16 acpi_vddci; + enum radeon_pcie_gen force_pcie_gen; + enum radeon_pcie_gen acpi_pcie_gen; + struct ci_leakage_voltage vddc_leakage; + struct ci_leakage_voltage vddci_leakage; + u16 max_vddc_in_pp_table; + u16 min_vddc_in_pp_table; + u16 max_vddci_in_pp_table; + u16 min_vddci_in_pp_table; + u32 mclk_strobe_mode_threshold; + u32 mclk_stutter_mode_threshold; + u32 mclk_edc_enable_threshold; + u32 mclk_edc_wr_enable_threshold; + struct ci_vbios_boot_state vbios_boot_state; + /* smc offsets */ + u32 sram_end; + u32 dpm_table_start; + u32 soft_regs_start; + u32 mc_reg_table_start; + u32 fan_table_start; + u32 arb_table_start; + /* smc tables */ + SMU7_Discrete_DpmTable smc_state_table; + SMU7_Discrete_MCRegisters smc_mc_reg_table; + SMU7_Discrete_PmFuses smc_powertune_table; + /* other stuff */ + struct ci_mc_reg_table mc_reg_table; + struct atom_voltage_table vddc_voltage_table; + struct atom_voltage_table vddci_voltage_table; + struct atom_voltage_table mvdd_voltage_table; + struct ci_ulv_parm ulv; + u32 power_containment_features; + const struct ci_pt_defaults *powertune_defaults; + u32 dte_tj_offset; + bool vddc_phase_shed_control; + struct ci_thermal_temperature_setting thermal_temp_setting; + struct ci_dpm_level_enable_mask dpm_level_enable_mask; + u32 need_update_smu7_dpm_table; + u32 sclk_dpm_key_disabled; + u32 mclk_dpm_key_disabled; + u32 pcie_dpm_key_disabled; + struct ci_pcie_perf_range pcie_gen_performance; + struct ci_pcie_perf_range pcie_lane_performance; + struct ci_pcie_perf_range pcie_gen_powersaving; + struct ci_pcie_perf_range pcie_lane_powersaving; + u32 activity_target[SMU7_MAX_LEVELS_GRAPHICS]; + u32 mclk_activity_target; + u32 low_sclk_interrupt_t; + u32 last_mclk_dpm_enable_mask; + u32 sys_pcie_mask; + /* caps */ + bool caps_power_containment; + bool caps_cac; + bool caps_sq_ramping; + bool caps_db_ramping; + bool caps_td_ramping; + bool caps_tcp_ramping; + bool caps_fps; + bool caps_sclk_ds; + bool caps_sclk_ss_support; + bool caps_mclk_ss_support; + bool caps_uvd_dpm; + bool caps_vce_dpm; + bool caps_samu_dpm; + bool caps_acp_dpm; + bool caps_automatic_dc_transition; + bool caps_sclk_throttle_low_notification; + bool caps_dynamic_ac_timing; + /* flags */ + bool thermal_protection; + bool pcie_performance_request; + bool dynamic_ss; + bool dll_default_on; + bool cac_enabled; + bool uvd_enabled; + bool battery_state; + bool pspp_notify_required; + bool mem_gddr5; + bool enable_bapm_feature; + bool enable_tdc_limit_feature; + bool enable_pkg_pwr_tracking_feature; + bool use_pcie_performance_levels; + bool use_pcie_powersaving_levels; + bool uvd_power_gated; + /* driver states */ + struct radeon_ps current_rps; + struct ci_ps current_ps; + struct radeon_ps requested_rps; + struct ci_ps requested_ps; +}; + +#define CISLANDS_VOLTAGE_CONTROL_NONE 0x0 +#define CISLANDS_VOLTAGE_CONTROL_BY_GPIO 0x1 +#define CISLANDS_VOLTAGE_CONTROL_BY_SVID2 0x2 + +#define CISLANDS_Q88_FORMAT_CONVERSION_UNIT 256 + +#define CISLANDS_VRC_DFLT0 0x3FFFC000 +#define CISLANDS_VRC_DFLT1 0x000400 +#define CISLANDS_VRC_DFLT2 0xC00080 +#define CISLANDS_VRC_DFLT3 0xC00200 +#define CISLANDS_VRC_DFLT4 0xC01680 +#define CISLANDS_VRC_DFLT5 0xC00033 +#define CISLANDS_VRC_DFLT6 0xC00033 +#define CISLANDS_VRC_DFLT7 0x3FFFC000 + +#define CISLANDS_CGULVPARAMETER_DFLT 0x00040035 +#define CISLAND_TARGETACTIVITY_DFLT 30 +#define CISLAND_MCLK_TARGETACTIVITY_DFLT 10 + +#define PCIE_PERF_REQ_REMOVE_REGISTRY 0 +#define PCIE_PERF_REQ_FORCE_LOWPOWER 1 +#define PCIE_PERF_REQ_PECI_GEN1 2 +#define PCIE_PERF_REQ_PECI_GEN2 3 +#define PCIE_PERF_REQ_PECI_GEN3 4 + +int ci_copy_bytes_to_smc(struct radeon_device *rdev, + u32 smc_start_address, + const u8 *src, u32 byte_count, u32 limit); +void ci_start_smc(struct radeon_device *rdev); +void ci_reset_smc(struct radeon_device *rdev); +int ci_program_jump_on_start(struct radeon_device *rdev); +void ci_stop_smc_clock(struct radeon_device *rdev); +void ci_start_smc_clock(struct radeon_device *rdev); +bool ci_is_smc_running(struct radeon_device *rdev); +PPSMC_Result ci_send_msg_to_smc(struct radeon_device *rdev, PPSMC_Msg msg); +PPSMC_Result ci_wait_for_smc_inactive(struct radeon_device *rdev); +int ci_load_smc_ucode(struct radeon_device *rdev, u32 limit); +int ci_read_smc_sram_dword(struct radeon_device *rdev, + u32 smc_address, u32 *value, u32 limit); +int ci_write_smc_sram_dword(struct radeon_device *rdev, + u32 smc_address, u32 value, u32 limit); + +#endif diff --git a/drivers/gpu/drm/radeon/ci_smc.c b/drivers/gpu/drm/radeon/ci_smc.c new file mode 100644 index 00000000000..8debc9d4736 --- /dev/null +++ b/drivers/gpu/drm/radeon/ci_smc.c @@ -0,0 +1,280 @@ +/* + * Copyright 2011 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ + +#include <linux/firmware.h> +#include "drmP.h" +#include "radeon.h" +#include "cikd.h" +#include "ppsmc.h" +#include "radeon_ucode.h" +#include "ci_dpm.h" + +static int ci_set_smc_sram_address(struct radeon_device *rdev, + u32 smc_address, u32 limit) +{ + if (smc_address & 3) + return -EINVAL; + if ((smc_address + 3) > limit) + return -EINVAL; + + WREG32(SMC_IND_INDEX_0, smc_address); + WREG32_P(SMC_IND_ACCESS_CNTL, 0, ~AUTO_INCREMENT_IND_0); + + return 0; +} + +int ci_copy_bytes_to_smc(struct radeon_device *rdev, + u32 smc_start_address, + const u8 *src, u32 byte_count, u32 limit) +{ + unsigned long flags; + u32 data, original_data; + u32 addr; + u32 extra_shift; + int ret = 0; + + if (smc_start_address & 3) + return -EINVAL; + if ((smc_start_address + byte_count) > limit) + return -EINVAL; + + addr = smc_start_address; + + spin_lock_irqsave(&rdev->smc_idx_lock, flags); + while (byte_count >= 4) { + /* SMC address space is BE */ + data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3]; + + ret = ci_set_smc_sram_address(rdev, addr, limit); + if (ret) + goto done; + + WREG32(SMC_IND_DATA_0, data); + + src += 4; + byte_count -= 4; + addr += 4; + } + + /* RMW for the final bytes */ + if (byte_count > 0) { + data = 0; + + ret = ci_set_smc_sram_address(rdev, addr, limit); + if (ret) + goto done; + + original_data = RREG32(SMC_IND_DATA_0); + + extra_shift = 8 * (4 - byte_count); + + while (byte_count > 0) { + data = (data << 8) + *src++; + byte_count--; + } + + data <<= extra_shift; + + data |= (original_data & ~((~0UL) << extra_shift)); + + ret = ci_set_smc_sram_address(rdev, addr, limit); + if (ret) + goto done; + + WREG32(SMC_IND_DATA_0, data); + } + +done: + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); + + return ret; +} + +void ci_start_smc(struct radeon_device *rdev) +{ + u32 tmp = RREG32_SMC(SMC_SYSCON_RESET_CNTL); + + tmp &= ~RST_REG; + WREG32_SMC(SMC_SYSCON_RESET_CNTL, tmp); +} + +void ci_reset_smc(struct radeon_device *rdev) +{ + u32 tmp = RREG32_SMC(SMC_SYSCON_RESET_CNTL); + + tmp |= RST_REG; + WREG32_SMC(SMC_SYSCON_RESET_CNTL, tmp); +} + +int ci_program_jump_on_start(struct radeon_device *rdev) +{ + static u8 data[] = { 0xE0, 0x00, 0x80, 0x40 }; + + return ci_copy_bytes_to_smc(rdev, 0x0, data, 4, sizeof(data)+1); +} + +void ci_stop_smc_clock(struct radeon_device *rdev) +{ + u32 tmp = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0); + + tmp |= CK_DISABLE; + + WREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0, tmp); +} + +void ci_start_smc_clock(struct radeon_device *rdev) +{ + u32 tmp = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0); + + tmp &= ~CK_DISABLE; + + WREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0, tmp); +} + +bool ci_is_smc_running(struct radeon_device *rdev) +{ + u32 clk = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0); + u32 pc_c = RREG32_SMC(SMC_PC_C); + + if (!(clk & CK_DISABLE) && (0x20100 <= pc_c)) + return true; + + return false; +} + +PPSMC_Result ci_send_msg_to_smc(struct radeon_device *rdev, PPSMC_Msg msg) +{ + u32 tmp; + int i; + + if (!ci_is_smc_running(rdev)) + return PPSMC_Result_Failed; + + WREG32(SMC_MESSAGE_0, msg); + + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = RREG32(SMC_RESP_0); + if (tmp != 0) + break; + udelay(1); + } + tmp = RREG32(SMC_RESP_0); + + return (PPSMC_Result)tmp; +} + +PPSMC_Result ci_wait_for_smc_inactive(struct radeon_device *rdev) +{ + u32 tmp; + int i; + + if (!ci_is_smc_running(rdev)) + return PPSMC_Result_OK; + + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0); + if ((tmp & CKEN) == 0) + break; + udelay(1); + } + + return PPSMC_Result_OK; +} + +int ci_load_smc_ucode(struct radeon_device *rdev, u32 limit) +{ + unsigned long flags; + u32 ucode_start_address; + u32 ucode_size; + const u8 *src; + u32 data; + + if (!rdev->smc_fw) + return -EINVAL; + + switch (rdev->family) { + case CHIP_BONAIRE: + ucode_start_address = BONAIRE_SMC_UCODE_START; + ucode_size = BONAIRE_SMC_UCODE_SIZE; + break; + case CHIP_HAWAII: + ucode_start_address = HAWAII_SMC_UCODE_START; + ucode_size = HAWAII_SMC_UCODE_SIZE; + break; + default: + DRM_ERROR("unknown asic in smc ucode loader\n"); + BUG(); + } + + if (ucode_size & 3) + return -EINVAL; + + src = (const u8 *)rdev->smc_fw->data; + spin_lock_irqsave(&rdev->smc_idx_lock, flags); + WREG32(SMC_IND_INDEX_0, ucode_start_address); + WREG32_P(SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, ~AUTO_INCREMENT_IND_0); + while (ucode_size >= 4) { + /* SMC address space is BE */ + data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3]; + + WREG32(SMC_IND_DATA_0, data); + + src += 4; + ucode_size -= 4; + } + WREG32_P(SMC_IND_ACCESS_CNTL, 0, ~AUTO_INCREMENT_IND_0); + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); + + return 0; +} + +int ci_read_smc_sram_dword(struct radeon_device *rdev, + u32 smc_address, u32 *value, u32 limit) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&rdev->smc_idx_lock, flags); + ret = ci_set_smc_sram_address(rdev, smc_address, limit); + if (ret == 0) + *value = RREG32(SMC_IND_DATA_0); + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); + + return ret; +} + +int ci_write_smc_sram_dword(struct radeon_device *rdev, + u32 smc_address, u32 value, u32 limit) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&rdev->smc_idx_lock, flags); + ret = ci_set_smc_sram_address(rdev, smc_address, limit); + if (ret == 0) + WREG32(SMC_IND_DATA_0, value); + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); + + return ret; +} diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index ed1d9102592..c0ea66192fe 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -22,7 +22,6 @@ * Authors: Alex Deucher */ #include <linux/firmware.h> -#include <linux/platform_device.h> #include <linux/slab.h> #include <linux/module.h> #include "drmP.h" @@ -31,30 +30,27 @@ #include "cikd.h" #include "atom.h" #include "cik_blit_shaders.h" - -/* GFX */ -#define CIK_PFP_UCODE_SIZE 2144 -#define CIK_ME_UCODE_SIZE 2144 -#define CIK_CE_UCODE_SIZE 2144 -/* compute */ -#define CIK_MEC_UCODE_SIZE 4192 -/* interrupts */ -#define BONAIRE_RLC_UCODE_SIZE 2048 -#define KB_RLC_UCODE_SIZE 2560 -#define KV_RLC_UCODE_SIZE 2560 -/* gddr controller */ -#define CIK_MC_UCODE_SIZE 7866 -/* sdma */ -#define CIK_SDMA_UCODE_SIZE 1050 -#define CIK_SDMA_UCODE_VERSION 64 +#include "radeon_ucode.h" +#include "clearstate_ci.h" MODULE_FIRMWARE("radeon/BONAIRE_pfp.bin"); MODULE_FIRMWARE("radeon/BONAIRE_me.bin"); MODULE_FIRMWARE("radeon/BONAIRE_ce.bin"); MODULE_FIRMWARE("radeon/BONAIRE_mec.bin"); MODULE_FIRMWARE("radeon/BONAIRE_mc.bin"); +MODULE_FIRMWARE("radeon/BONAIRE_mc2.bin"); MODULE_FIRMWARE("radeon/BONAIRE_rlc.bin"); MODULE_FIRMWARE("radeon/BONAIRE_sdma.bin"); +MODULE_FIRMWARE("radeon/BONAIRE_smc.bin"); +MODULE_FIRMWARE("radeon/HAWAII_pfp.bin"); +MODULE_FIRMWARE("radeon/HAWAII_me.bin"); +MODULE_FIRMWARE("radeon/HAWAII_ce.bin"); +MODULE_FIRMWARE("radeon/HAWAII_mec.bin"); +MODULE_FIRMWARE("radeon/HAWAII_mc.bin"); +MODULE_FIRMWARE("radeon/HAWAII_mc2.bin"); +MODULE_FIRMWARE("radeon/HAWAII_rlc.bin"); +MODULE_FIRMWARE("radeon/HAWAII_sdma.bin"); +MODULE_FIRMWARE("radeon/HAWAII_smc.bin"); MODULE_FIRMWARE("radeon/KAVERI_pfp.bin"); MODULE_FIRMWARE("radeon/KAVERI_me.bin"); MODULE_FIRMWARE("radeon/KAVERI_ce.bin"); @@ -67,38 +63,875 @@ MODULE_FIRMWARE("radeon/KABINI_ce.bin"); MODULE_FIRMWARE("radeon/KABINI_mec.bin"); MODULE_FIRMWARE("radeon/KABINI_rlc.bin"); MODULE_FIRMWARE("radeon/KABINI_sdma.bin"); +MODULE_FIRMWARE("radeon/MULLINS_pfp.bin"); +MODULE_FIRMWARE("radeon/MULLINS_me.bin"); +MODULE_FIRMWARE("radeon/MULLINS_ce.bin"); +MODULE_FIRMWARE("radeon/MULLINS_mec.bin"); +MODULE_FIRMWARE("radeon/MULLINS_rlc.bin"); +MODULE_FIRMWARE("radeon/MULLINS_sdma.bin"); extern int r600_ih_ring_alloc(struct radeon_device *rdev); extern void r600_ih_ring_fini(struct radeon_device *rdev); extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save); extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save); extern bool evergreen_is_display_hung(struct radeon_device *rdev); +extern void sumo_rlc_fini(struct radeon_device *rdev); +extern int sumo_rlc_init(struct radeon_device *rdev); extern void si_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc); -extern void si_rlc_fini(struct radeon_device *rdev); -extern int si_rlc_init(struct radeon_device *rdev); +extern void si_rlc_reset(struct radeon_device *rdev); +extern void si_init_uvd_internal_cg(struct radeon_device *rdev); +static u32 cik_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh); +extern int cik_sdma_resume(struct radeon_device *rdev); +extern void cik_sdma_enable(struct radeon_device *rdev, bool enable); +extern void cik_sdma_fini(struct radeon_device *rdev); +extern void vce_v2_0_enable_mgcg(struct radeon_device *rdev, bool enable); static void cik_rlc_stop(struct radeon_device *rdev); +static void cik_pcie_gen3_enable(struct radeon_device *rdev); +static void cik_program_aspm(struct radeon_device *rdev); +static void cik_init_pg(struct radeon_device *rdev); +static void cik_init_cg(struct radeon_device *rdev); +static void cik_fini_pg(struct radeon_device *rdev); +static void cik_fini_cg(struct radeon_device *rdev); +static void cik_enable_gui_idle_interrupt(struct radeon_device *rdev, + bool enable); + +/* get temperature in millidegrees */ +int ci_get_temp(struct radeon_device *rdev) +{ + u32 temp; + int actual_temp = 0; + + temp = (RREG32_SMC(CG_MULT_THERMAL_STATUS) & CTF_TEMP_MASK) >> + CTF_TEMP_SHIFT; + + if (temp & 0x200) + actual_temp = 255; + else + actual_temp = temp & 0x1ff; + + actual_temp = actual_temp * 1000; + + return actual_temp; +} + +/* get temperature in millidegrees */ +int kv_get_temp(struct radeon_device *rdev) +{ + u32 temp; + int actual_temp = 0; + + temp = RREG32_SMC(0xC0300E0C); + + if (temp) + actual_temp = (temp / 8) - 49; + else + actual_temp = 0; + + actual_temp = actual_temp * 1000; + + return actual_temp; +} /* * Indirect registers accessor */ u32 cik_pciep_rreg(struct radeon_device *rdev, u32 reg) { + unsigned long flags; u32 r; + spin_lock_irqsave(&rdev->pciep_idx_lock, flags); WREG32(PCIE_INDEX, reg); (void)RREG32(PCIE_INDEX); r = RREG32(PCIE_DATA); + spin_unlock_irqrestore(&rdev->pciep_idx_lock, flags); return r; } void cik_pciep_wreg(struct radeon_device *rdev, u32 reg, u32 v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->pciep_idx_lock, flags); WREG32(PCIE_INDEX, reg); (void)RREG32(PCIE_INDEX); WREG32(PCIE_DATA, v); (void)RREG32(PCIE_DATA); + spin_unlock_irqrestore(&rdev->pciep_idx_lock, flags); } +static const u32 spectre_rlc_save_restore_register_list[] = +{ + (0x0e00 << 16) | (0xc12c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc140 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc150 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc15c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc168 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc170 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc178 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc204 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2b4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2b8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2bc >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2c0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8228 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x829c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x869c >> 2), + 0x00000000, + (0x0600 << 16) | (0x98f4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x98f8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9900 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc260 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x90e8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c000 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c00c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c1c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9700 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x4e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x5e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x6e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x7e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x8e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x9e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0xae00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0xbe00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x89bc >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8900 >> 2), + 0x00000000, + 0x3, + (0x0e00 << 16) | (0xc130 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc134 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc1fc >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc208 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc264 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc268 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc26c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc270 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc274 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc278 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc27c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc280 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc284 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc288 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc28c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc290 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc294 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc298 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc29c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2a0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2a4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2a8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2ac >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2b0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x301d0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30238 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30250 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30254 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30258 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3025c >> 2), + 0x00000000, + (0x4e00 << 16) | (0xc900 >> 2), + 0x00000000, + (0x5e00 << 16) | (0xc900 >> 2), + 0x00000000, + (0x6e00 << 16) | (0xc900 >> 2), + 0x00000000, + (0x7e00 << 16) | (0xc900 >> 2), + 0x00000000, + (0x8e00 << 16) | (0xc900 >> 2), + 0x00000000, + (0x9e00 << 16) | (0xc900 >> 2), + 0x00000000, + (0xae00 << 16) | (0xc900 >> 2), + 0x00000000, + (0xbe00 << 16) | (0xc900 >> 2), + 0x00000000, + (0x4e00 << 16) | (0xc904 >> 2), + 0x00000000, + (0x5e00 << 16) | (0xc904 >> 2), + 0x00000000, + (0x6e00 << 16) | (0xc904 >> 2), + 0x00000000, + (0x7e00 << 16) | (0xc904 >> 2), + 0x00000000, + (0x8e00 << 16) | (0xc904 >> 2), + 0x00000000, + (0x9e00 << 16) | (0xc904 >> 2), + 0x00000000, + (0xae00 << 16) | (0xc904 >> 2), + 0x00000000, + (0xbe00 << 16) | (0xc904 >> 2), + 0x00000000, + (0x4e00 << 16) | (0xc908 >> 2), + 0x00000000, + (0x5e00 << 16) | (0xc908 >> 2), + 0x00000000, + (0x6e00 << 16) | (0xc908 >> 2), + 0x00000000, + (0x7e00 << 16) | (0xc908 >> 2), + 0x00000000, + (0x8e00 << 16) | (0xc908 >> 2), + 0x00000000, + (0x9e00 << 16) | (0xc908 >> 2), + 0x00000000, + (0xae00 << 16) | (0xc908 >> 2), + 0x00000000, + (0xbe00 << 16) | (0xc908 >> 2), + 0x00000000, + (0x4e00 << 16) | (0xc90c >> 2), + 0x00000000, + (0x5e00 << 16) | (0xc90c >> 2), + 0x00000000, + (0x6e00 << 16) | (0xc90c >> 2), + 0x00000000, + (0x7e00 << 16) | (0xc90c >> 2), + 0x00000000, + (0x8e00 << 16) | (0xc90c >> 2), + 0x00000000, + (0x9e00 << 16) | (0xc90c >> 2), + 0x00000000, + (0xae00 << 16) | (0xc90c >> 2), + 0x00000000, + (0xbe00 << 16) | (0xc90c >> 2), + 0x00000000, + (0x4e00 << 16) | (0xc910 >> 2), + 0x00000000, + (0x5e00 << 16) | (0xc910 >> 2), + 0x00000000, + (0x6e00 << 16) | (0xc910 >> 2), + 0x00000000, + (0x7e00 << 16) | (0xc910 >> 2), + 0x00000000, + (0x8e00 << 16) | (0xc910 >> 2), + 0x00000000, + (0x9e00 << 16) | (0xc910 >> 2), + 0x00000000, + (0xae00 << 16) | (0xc910 >> 2), + 0x00000000, + (0xbe00 << 16) | (0xc910 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc99c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9834 >> 2), + 0x00000000, + (0x0000 << 16) | (0x30f00 >> 2), + 0x00000000, + (0x0001 << 16) | (0x30f00 >> 2), + 0x00000000, + (0x0000 << 16) | (0x30f04 >> 2), + 0x00000000, + (0x0001 << 16) | (0x30f04 >> 2), + 0x00000000, + (0x0000 << 16) | (0x30f08 >> 2), + 0x00000000, + (0x0001 << 16) | (0x30f08 >> 2), + 0x00000000, + (0x0000 << 16) | (0x30f0c >> 2), + 0x00000000, + (0x0001 << 16) | (0x30f0c >> 2), + 0x00000000, + (0x0600 << 16) | (0x9b7c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8a14 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8a18 >> 2), + 0x00000000, + (0x0600 << 16) | (0x30a00 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8bf0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8bcc >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8b24 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30a04 >> 2), + 0x00000000, + (0x0600 << 16) | (0x30a10 >> 2), + 0x00000000, + (0x0600 << 16) | (0x30a14 >> 2), + 0x00000000, + (0x0600 << 16) | (0x30a18 >> 2), + 0x00000000, + (0x0600 << 16) | (0x30a2c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc700 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc704 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc708 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc768 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc770 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc774 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc778 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc77c >> 2), + 0x00000000, + (0x0400 << 16) | (0xc780 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc784 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc788 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc78c >> 2), + 0x00000000, + (0x0400 << 16) | (0xc798 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc79c >> 2), + 0x00000000, + (0x0400 << 16) | (0xc7a0 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc7a4 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc7a8 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc7ac >> 2), + 0x00000000, + (0x0400 << 16) | (0xc7b0 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc7b4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9100 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c010 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x92a8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x92ac >> 2), + 0x00000000, + (0x0e00 << 16) | (0x92b4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x92b8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x92bc >> 2), + 0x00000000, + (0x0e00 << 16) | (0x92c0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x92c4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x92c8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x92cc >> 2), + 0x00000000, + (0x0e00 << 16) | (0x92d0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c00 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c04 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c20 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c38 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c3c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xae00 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9604 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac08 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac0c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac10 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac14 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac58 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac68 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac6c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac70 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac74 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac78 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac7c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac80 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac84 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac88 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac8c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x970c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9714 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9718 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x971c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x4e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x5e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x6e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x7e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x8e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x9e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0xae00 << 16) | (0x31068 >> 2), + 0x00000000, + (0xbe00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xcd10 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xcd14 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88b0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88b4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88b8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88bc >> 2), + 0x00000000, + (0x0400 << 16) | (0x89c0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88c4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88c8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88d0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88d4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88d8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8980 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30938 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3093c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30940 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x89a0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30900 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30904 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x89b4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c210 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c214 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c218 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8904 >> 2), + 0x00000000, + 0x5, + (0x0e00 << 16) | (0x8c28 >> 2), + (0x0e00 << 16) | (0x8c2c >> 2), + (0x0e00 << 16) | (0x8c30 >> 2), + (0x0e00 << 16) | (0x8c34 >> 2), + (0x0e00 << 16) | (0x9600 >> 2), +}; + +static const u32 kalindi_rlc_save_restore_register_list[] = +{ + (0x0e00 << 16) | (0xc12c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc140 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc150 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc15c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc168 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc170 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc204 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2b4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2b8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2bc >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2c0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8228 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x829c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x869c >> 2), + 0x00000000, + (0x0600 << 16) | (0x98f4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x98f8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9900 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc260 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x90e8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c000 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c00c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c1c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9700 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x4e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x5e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x6e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x7e00 << 16) | (0xcd20 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x89bc >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8900 >> 2), + 0x00000000, + 0x3, + (0x0e00 << 16) | (0xc130 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc134 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc1fc >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc208 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc264 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc268 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc26c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc270 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc274 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc28c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc290 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc294 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc298 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2a0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2a4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2a8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc2ac >> 2), + 0x00000000, + (0x0e00 << 16) | (0x301d0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30238 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30250 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30254 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30258 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3025c >> 2), + 0x00000000, + (0x4e00 << 16) | (0xc900 >> 2), + 0x00000000, + (0x5e00 << 16) | (0xc900 >> 2), + 0x00000000, + (0x6e00 << 16) | (0xc900 >> 2), + 0x00000000, + (0x7e00 << 16) | (0xc900 >> 2), + 0x00000000, + (0x4e00 << 16) | (0xc904 >> 2), + 0x00000000, + (0x5e00 << 16) | (0xc904 >> 2), + 0x00000000, + (0x6e00 << 16) | (0xc904 >> 2), + 0x00000000, + (0x7e00 << 16) | (0xc904 >> 2), + 0x00000000, + (0x4e00 << 16) | (0xc908 >> 2), + 0x00000000, + (0x5e00 << 16) | (0xc908 >> 2), + 0x00000000, + (0x6e00 << 16) | (0xc908 >> 2), + 0x00000000, + (0x7e00 << 16) | (0xc908 >> 2), + 0x00000000, + (0x4e00 << 16) | (0xc90c >> 2), + 0x00000000, + (0x5e00 << 16) | (0xc90c >> 2), + 0x00000000, + (0x6e00 << 16) | (0xc90c >> 2), + 0x00000000, + (0x7e00 << 16) | (0xc90c >> 2), + 0x00000000, + (0x4e00 << 16) | (0xc910 >> 2), + 0x00000000, + (0x5e00 << 16) | (0xc910 >> 2), + 0x00000000, + (0x6e00 << 16) | (0xc910 >> 2), + 0x00000000, + (0x7e00 << 16) | (0xc910 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc99c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9834 >> 2), + 0x00000000, + (0x0000 << 16) | (0x30f00 >> 2), + 0x00000000, + (0x0000 << 16) | (0x30f04 >> 2), + 0x00000000, + (0x0000 << 16) | (0x30f08 >> 2), + 0x00000000, + (0x0000 << 16) | (0x30f0c >> 2), + 0x00000000, + (0x0600 << 16) | (0x9b7c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8a14 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8a18 >> 2), + 0x00000000, + (0x0600 << 16) | (0x30a00 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8bf0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8bcc >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8b24 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30a04 >> 2), + 0x00000000, + (0x0600 << 16) | (0x30a10 >> 2), + 0x00000000, + (0x0600 << 16) | (0x30a14 >> 2), + 0x00000000, + (0x0600 << 16) | (0x30a18 >> 2), + 0x00000000, + (0x0600 << 16) | (0x30a2c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc700 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc704 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc708 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xc768 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc770 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc774 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc798 >> 2), + 0x00000000, + (0x0400 << 16) | (0xc79c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9100 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c010 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c00 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c04 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c20 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c38 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8c3c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xae00 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9604 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac08 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac0c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac10 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac14 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac58 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac68 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac6c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac70 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac74 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac78 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac7c >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac80 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac84 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac88 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xac8c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x970c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9714 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x9718 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x971c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x4e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x5e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x6e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x7e00 << 16) | (0x31068 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xcd10 >> 2), + 0x00000000, + (0x0e00 << 16) | (0xcd14 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88b0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88b4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88b8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88bc >> 2), + 0x00000000, + (0x0400 << 16) | (0x89c0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88c4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88c8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88d0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88d4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x88d8 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8980 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30938 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3093c >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30940 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x89a0 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30900 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x30904 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x89b4 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3e1fc >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c210 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c214 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x3c218 >> 2), + 0x00000000, + (0x0e00 << 16) | (0x8904 >> 2), + 0x00000000, + 0x5, + (0x0e00 << 16) | (0x8c28 >> 2), + (0x0e00 << 16) | (0x8c2c >> 2), + (0x0e00 << 16) | (0x8c30 >> 2), + (0x0e00 << 16) | (0x8c34 >> 2), + (0x0e00 << 16) | (0x9600 >> 2), +}; + static const u32 bonaire_golden_spm_registers[] = { 0x30800, 0xe0ffffff, 0xe0000000 @@ -272,7 +1105,7 @@ static const u32 spectre_golden_registers[] = 0x8a14, 0xf000003f, 0x00000007, 0x8b24, 0xffffffff, 0x00ffffff, 0x28350, 0x3f3f3fff, 0x00000082, - 0x28355, 0x0000003f, 0x00000000, + 0x28354, 0x0000003f, 0x00000000, 0x3e78, 0x00000001, 0x00000002, 0x913c, 0xffff03df, 0x00000004, 0xc768, 0x00000008, 0x00000008, @@ -482,6 +1315,208 @@ static const u32 kalindi_mgcg_cgcg_init[] = 0xd80c, 0xff000ff0, 0x00000100 }; +static const u32 hawaii_golden_spm_registers[] = +{ + 0x30800, 0xe0ffffff, 0xe0000000 +}; + +static const u32 hawaii_golden_common_registers[] = +{ + 0x30800, 0xffffffff, 0xe0000000, + 0x28350, 0xffffffff, 0x3a00161a, + 0x28354, 0xffffffff, 0x0000002e, + 0x9a10, 0xffffffff, 0x00018208, + 0x98f8, 0xffffffff, 0x12011003 +}; + +static const u32 hawaii_golden_registers[] = +{ + 0x3354, 0x00000333, 0x00000333, + 0x9a10, 0x00010000, 0x00058208, + 0x9830, 0xffffffff, 0x00000000, + 0x9834, 0xf00fffff, 0x00000400, + 0x9838, 0x0002021c, 0x00020200, + 0xc78, 0x00000080, 0x00000000, + 0x5bb0, 0x000000f0, 0x00000070, + 0x5bc0, 0xf0311fff, 0x80300000, + 0x350c, 0x00810000, 0x408af000, + 0x7030, 0x31000111, 0x00000011, + 0x2f48, 0x73773777, 0x12010001, + 0x2120, 0x0000007f, 0x0000001b, + 0x21dc, 0x00007fb6, 0x00002191, + 0x3628, 0x0000003f, 0x0000000a, + 0x362c, 0x0000003f, 0x0000000a, + 0x2ae4, 0x00073ffe, 0x000022a2, + 0x240c, 0x000007ff, 0x00000000, + 0x8bf0, 0x00002001, 0x00000001, + 0x8b24, 0xffffffff, 0x00ffffff, + 0x30a04, 0x0000ff0f, 0x00000000, + 0x28a4c, 0x07ffffff, 0x06000000, + 0x3e78, 0x00000001, 0x00000002, + 0xc768, 0x00000008, 0x00000008, + 0xc770, 0x00000f00, 0x00000800, + 0xc774, 0x00000f00, 0x00000800, + 0xc798, 0x00ffffff, 0x00ff7fbf, + 0xc79c, 0x00ffffff, 0x00ff7faf, + 0x8c00, 0x000000ff, 0x00000800, + 0xe40, 0x00001fff, 0x00001fff, + 0x9060, 0x0000007f, 0x00000020, + 0x9508, 0x00010000, 0x00010000, + 0xae00, 0x00100000, 0x000ff07c, + 0xac14, 0x000003ff, 0x0000000f, + 0xac10, 0xffffffff, 0x7564fdec, + 0xac0c, 0xffffffff, 0x3120b9a8, + 0xac08, 0x20000000, 0x0f9c0000 +}; + +static const u32 hawaii_mgcg_cgcg_init[] = +{ + 0xc420, 0xffffffff, 0xfffffffd, + 0x30800, 0xffffffff, 0xe0000000, + 0x3c2a0, 0xffffffff, 0x00000100, + 0x3c208, 0xffffffff, 0x00000100, + 0x3c2c0, 0xffffffff, 0x00000100, + 0x3c2c8, 0xffffffff, 0x00000100, + 0x3c2c4, 0xffffffff, 0x00000100, + 0x55e4, 0xffffffff, 0x00200100, + 0x3c280, 0xffffffff, 0x00000100, + 0x3c214, 0xffffffff, 0x06000100, + 0x3c220, 0xffffffff, 0x00000100, + 0x3c218, 0xffffffff, 0x06000100, + 0x3c204, 0xffffffff, 0x00000100, + 0x3c2e0, 0xffffffff, 0x00000100, + 0x3c224, 0xffffffff, 0x00000100, + 0x3c200, 0xffffffff, 0x00000100, + 0x3c230, 0xffffffff, 0x00000100, + 0x3c234, 0xffffffff, 0x00000100, + 0x3c250, 0xffffffff, 0x00000100, + 0x3c254, 0xffffffff, 0x00000100, + 0x3c258, 0xffffffff, 0x00000100, + 0x3c25c, 0xffffffff, 0x00000100, + 0x3c260, 0xffffffff, 0x00000100, + 0x3c27c, 0xffffffff, 0x00000100, + 0x3c278, 0xffffffff, 0x00000100, + 0x3c210, 0xffffffff, 0x06000100, + 0x3c290, 0xffffffff, 0x00000100, + 0x3c274, 0xffffffff, 0x00000100, + 0x3c2b4, 0xffffffff, 0x00000100, + 0x3c2b0, 0xffffffff, 0x00000100, + 0x3c270, 0xffffffff, 0x00000100, + 0x30800, 0xffffffff, 0xe0000000, + 0x3c020, 0xffffffff, 0x00010000, + 0x3c024, 0xffffffff, 0x00030002, + 0x3c028, 0xffffffff, 0x00040007, + 0x3c02c, 0xffffffff, 0x00060005, + 0x3c030, 0xffffffff, 0x00090008, + 0x3c034, 0xffffffff, 0x00010000, + 0x3c038, 0xffffffff, 0x00030002, + 0x3c03c, 0xffffffff, 0x00040007, + 0x3c040, 0xffffffff, 0x00060005, + 0x3c044, 0xffffffff, 0x00090008, + 0x3c048, 0xffffffff, 0x00010000, + 0x3c04c, 0xffffffff, 0x00030002, + 0x3c050, 0xffffffff, 0x00040007, + 0x3c054, 0xffffffff, 0x00060005, + 0x3c058, 0xffffffff, 0x00090008, + 0x3c05c, 0xffffffff, 0x00010000, + 0x3c060, 0xffffffff, 0x00030002, + 0x3c064, 0xffffffff, 0x00040007, + 0x3c068, 0xffffffff, 0x00060005, + 0x3c06c, 0xffffffff, 0x00090008, + 0x3c070, 0xffffffff, 0x00010000, + 0x3c074, 0xffffffff, 0x00030002, + 0x3c078, 0xffffffff, 0x00040007, + 0x3c07c, 0xffffffff, 0x00060005, + 0x3c080, 0xffffffff, 0x00090008, + 0x3c084, 0xffffffff, 0x00010000, + 0x3c088, 0xffffffff, 0x00030002, + 0x3c08c, 0xffffffff, 0x00040007, + 0x3c090, 0xffffffff, 0x00060005, + 0x3c094, 0xffffffff, 0x00090008, + 0x3c098, 0xffffffff, 0x00010000, + 0x3c09c, 0xffffffff, 0x00030002, + 0x3c0a0, 0xffffffff, 0x00040007, + 0x3c0a4, 0xffffffff, 0x00060005, + 0x3c0a8, 0xffffffff, 0x00090008, + 0x3c0ac, 0xffffffff, 0x00010000, + 0x3c0b0, 0xffffffff, 0x00030002, + 0x3c0b4, 0xffffffff, 0x00040007, + 0x3c0b8, 0xffffffff, 0x00060005, + 0x3c0bc, 0xffffffff, 0x00090008, + 0x3c0c0, 0xffffffff, 0x00010000, + 0x3c0c4, 0xffffffff, 0x00030002, + 0x3c0c8, 0xffffffff, 0x00040007, + 0x3c0cc, 0xffffffff, 0x00060005, + 0x3c0d0, 0xffffffff, 0x00090008, + 0x3c0d4, 0xffffffff, 0x00010000, + 0x3c0d8, 0xffffffff, 0x00030002, + 0x3c0dc, 0xffffffff, 0x00040007, + 0x3c0e0, 0xffffffff, 0x00060005, + 0x3c0e4, 0xffffffff, 0x00090008, + 0x3c0e8, 0xffffffff, 0x00010000, + 0x3c0ec, 0xffffffff, 0x00030002, + 0x3c0f0, 0xffffffff, 0x00040007, + 0x3c0f4, 0xffffffff, 0x00060005, + 0x3c0f8, 0xffffffff, 0x00090008, + 0xc318, 0xffffffff, 0x00020200, + 0x3350, 0xffffffff, 0x00000200, + 0x15c0, 0xffffffff, 0x00000400, + 0x55e8, 0xffffffff, 0x00000000, + 0x2f50, 0xffffffff, 0x00000902, + 0x3c000, 0xffffffff, 0x96940200, + 0x8708, 0xffffffff, 0x00900100, + 0xc424, 0xffffffff, 0x0020003f, + 0x38, 0xffffffff, 0x0140001c, + 0x3c, 0x000f0000, 0x000f0000, + 0x220, 0xffffffff, 0xc060000c, + 0x224, 0xc0000fff, 0x00000100, + 0xf90, 0xffffffff, 0x00000100, + 0xf98, 0x00000101, 0x00000000, + 0x20a8, 0xffffffff, 0x00000104, + 0x55e4, 0xff000fff, 0x00000100, + 0x30cc, 0xc0000fff, 0x00000104, + 0xc1e4, 0x00000001, 0x00000001, + 0xd00c, 0xff000ff0, 0x00000100, + 0xd80c, 0xff000ff0, 0x00000100 +}; + +static const u32 godavari_golden_registers[] = +{ + 0x55e4, 0xff607fff, 0xfc000100, + 0x6ed8, 0x00010101, 0x00010000, + 0x9830, 0xffffffff, 0x00000000, + 0x98302, 0xf00fffff, 0x00000400, + 0x6130, 0xffffffff, 0x00010000, + 0x5bb0, 0x000000f0, 0x00000070, + 0x5bc0, 0xf0311fff, 0x80300000, + 0x98f8, 0x73773777, 0x12010001, + 0x98fc, 0xffffffff, 0x00000010, + 0x8030, 0x00001f0f, 0x0000100a, + 0x2f48, 0x73773777, 0x12010001, + 0x2408, 0x000fffff, 0x000c007f, + 0x8a14, 0xf000003f, 0x00000007, + 0x8b24, 0xffffffff, 0x00ff0fff, + 0x30a04, 0x0000ff0f, 0x00000000, + 0x28a4c, 0x07ffffff, 0x06000000, + 0x4d8, 0x00000fff, 0x00000100, + 0xd014, 0x00010000, 0x00810001, + 0xd814, 0x00010000, 0x00810001, + 0x3e78, 0x00000001, 0x00000002, + 0xc768, 0x00000008, 0x00000008, + 0xc770, 0x00000f00, 0x00000800, + 0xc774, 0x00000f00, 0x00000800, + 0xc798, 0x00ffffff, 0x00ff7fbf, + 0xc79c, 0x00ffffff, 0x00ff7faf, + 0x8c00, 0x000000ff, 0x00000001, + 0x214f8, 0x01ff01ff, 0x00000002, + 0x21498, 0x007ff800, 0x00200000, + 0x2015c, 0xffffffff, 0x00000f40, + 0x88c4, 0x001f3ae3, 0x00000082, + 0x88d4, 0x0000001f, 0x00000010, + 0x30934, 0xffffffff, 0x00000000 +}; + + static void cik_init_golden_registers(struct radeon_device *rdev) { switch (rdev->family) { @@ -513,6 +1548,20 @@ static void cik_init_golden_registers(struct radeon_device *rdev) kalindi_golden_spm_registers, (const u32)ARRAY_SIZE(kalindi_golden_spm_registers)); break; + case CHIP_MULLINS: + radeon_program_register_sequence(rdev, + kalindi_mgcg_cgcg_init, + (const u32)ARRAY_SIZE(kalindi_mgcg_cgcg_init)); + radeon_program_register_sequence(rdev, + godavari_golden_registers, + (const u32)ARRAY_SIZE(godavari_golden_registers)); + radeon_program_register_sequence(rdev, + kalindi_golden_common_registers, + (const u32)ARRAY_SIZE(kalindi_golden_common_registers)); + radeon_program_register_sequence(rdev, + kalindi_golden_spm_registers, + (const u32)ARRAY_SIZE(kalindi_golden_spm_registers)); + break; case CHIP_KAVERI: radeon_program_register_sequence(rdev, spectre_mgcg_cgcg_init, @@ -527,6 +1576,20 @@ static void cik_init_golden_registers(struct radeon_device *rdev) spectre_golden_spm_registers, (const u32)ARRAY_SIZE(spectre_golden_spm_registers)); break; + case CHIP_HAWAII: + radeon_program_register_sequence(rdev, + hawaii_mgcg_cgcg_init, + (const u32)ARRAY_SIZE(hawaii_mgcg_cgcg_init)); + radeon_program_register_sequence(rdev, + hawaii_golden_registers, + (const u32)ARRAY_SIZE(hawaii_golden_registers)); + radeon_program_register_sequence(rdev, + hawaii_golden_common_registers, + (const u32)ARRAY_SIZE(hawaii_golden_common_registers)); + radeon_program_register_sequence(rdev, + hawaii_golden_spm_registers, + (const u32)ARRAY_SIZE(hawaii_golden_spm_registers)); + break; default: break; } @@ -558,17 +1621,17 @@ u32 cik_get_xclk(struct radeon_device *rdev) * cik_mm_rdoorbell - read a doorbell dword * * @rdev: radeon_device pointer - * @offset: byte offset into the aperture + * @index: doorbell index * * Returns the value in the doorbell aperture at the - * requested offset (CIK). + * requested doorbell index (CIK). */ -u32 cik_mm_rdoorbell(struct radeon_device *rdev, u32 offset) +u32 cik_mm_rdoorbell(struct radeon_device *rdev, u32 index) { - if (offset < rdev->doorbell.size) { - return readl(((void __iomem *)rdev->doorbell.ptr) + offset); + if (index < rdev->doorbell.num_doorbells) { + return readl(rdev->doorbell.ptr + index); } else { - DRM_ERROR("reading beyond doorbell aperture: 0x%08x!\n", offset); + DRM_ERROR("reading beyond doorbell aperture: 0x%08x!\n", index); return 0; } } @@ -577,18 +1640,18 @@ u32 cik_mm_rdoorbell(struct radeon_device *rdev, u32 offset) * cik_mm_wdoorbell - write a doorbell dword * * @rdev: radeon_device pointer - * @offset: byte offset into the aperture + * @index: doorbell index * @v: value to write * * Writes @v to the doorbell aperture at the - * requested offset (CIK). + * requested doorbell index (CIK). */ -void cik_mm_wdoorbell(struct radeon_device *rdev, u32 offset, u32 v) +void cik_mm_wdoorbell(struct radeon_device *rdev, u32 index, u32 v) { - if (offset < rdev->doorbell.size) { - writel(v, ((void __iomem *)rdev->doorbell.ptr) + offset); + if (index < rdev->doorbell.num_doorbells) { + writel(v, rdev->doorbell.ptr + index); } else { - DRM_ERROR("writing beyond doorbell aperture: 0x%08x!\n", offset); + DRM_ERROR("writing beyond doorbell aperture: 0x%08x!\n", index); } } @@ -634,6 +1697,35 @@ static const u32 bonaire_io_mc_regs[BONAIRE_IO_MC_REGS_SIZE][2] = {0x0000009f, 0x00b48000} }; +#define HAWAII_IO_MC_REGS_SIZE 22 + +static const u32 hawaii_io_mc_regs[HAWAII_IO_MC_REGS_SIZE][2] = +{ + {0x0000007d, 0x40000000}, + {0x0000007e, 0x40180304}, + {0x0000007f, 0x0000ff00}, + {0x00000081, 0x00000000}, + {0x00000083, 0x00000800}, + {0x00000086, 0x00000000}, + {0x00000087, 0x00000100}, + {0x00000088, 0x00020100}, + {0x00000089, 0x00000000}, + {0x0000008b, 0x00040000}, + {0x0000008c, 0x00000100}, + {0x0000008e, 0xff010000}, + {0x00000090, 0xffffefff}, + {0x00000091, 0xfff3efff}, + {0x00000092, 0xfff3efbf}, + {0x00000093, 0xf7ffffff}, + {0x00000094, 0xffffff7f}, + {0x00000095, 0x00000fff}, + {0x00000096, 0x00116fff}, + {0x00000097, 0x60010000}, + {0x00000098, 0x10010000}, + {0x0000009f, 0x00c79000} +}; + + /** * cik_srbm_select - select specific register instances * @@ -666,23 +1758,29 @@ static void cik_srbm_select(struct radeon_device *rdev, * Load the GDDR MC ucode into the hw (CIK). * Returns 0 on success, error on failure. */ -static int ci_mc_load_microcode(struct radeon_device *rdev) +int ci_mc_load_microcode(struct radeon_device *rdev) { const __be32 *fw_data; u32 running, blackout = 0; u32 *io_mc_regs; - int i, ucode_size, regs_size; + int i, regs_size, ucode_size; if (!rdev->mc_fw) return -EINVAL; + ucode_size = rdev->mc_fw->size / 4; + switch (rdev->family) { case CHIP_BONAIRE: - default: io_mc_regs = (u32 *)&bonaire_io_mc_regs; - ucode_size = CIK_MC_UCODE_SIZE; regs_size = BONAIRE_IO_MC_REGS_SIZE; break; + case CHIP_HAWAII: + io_mc_regs = (u32 *)&hawaii_io_mc_regs; + regs_size = HAWAII_IO_MC_REGS_SIZE; + break; + default: + return -EINVAL; } running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK; @@ -742,23 +1840,15 @@ static int ci_mc_load_microcode(struct radeon_device *rdev) */ static int cik_init_microcode(struct radeon_device *rdev) { - struct platform_device *pdev; const char *chip_name; size_t pfp_req_size, me_req_size, ce_req_size, - mec_req_size, rlc_req_size, mc_req_size, - sdma_req_size; + mec_req_size, rlc_req_size, mc_req_size = 0, + sdma_req_size, smc_req_size = 0, mc2_req_size = 0; char fw_name[30]; int err; DRM_DEBUG("\n"); - pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0); - err = IS_ERR(pdev); - if (err) { - printk(KERN_ERR "radeon_cp: Failed to register firmware\n"); - return -EINVAL; - } - switch (rdev->family) { case CHIP_BONAIRE: chip_name = "BONAIRE"; @@ -767,8 +1857,22 @@ static int cik_init_microcode(struct radeon_device *rdev) ce_req_size = CIK_CE_UCODE_SIZE * 4; mec_req_size = CIK_MEC_UCODE_SIZE * 4; rlc_req_size = BONAIRE_RLC_UCODE_SIZE * 4; - mc_req_size = CIK_MC_UCODE_SIZE * 4; + mc_req_size = BONAIRE_MC_UCODE_SIZE * 4; + mc2_req_size = BONAIRE_MC2_UCODE_SIZE * 4; sdma_req_size = CIK_SDMA_UCODE_SIZE * 4; + smc_req_size = ALIGN(BONAIRE_SMC_UCODE_SIZE, 4); + break; + case CHIP_HAWAII: + chip_name = "HAWAII"; + pfp_req_size = CIK_PFP_UCODE_SIZE * 4; + me_req_size = CIK_ME_UCODE_SIZE * 4; + ce_req_size = CIK_CE_UCODE_SIZE * 4; + mec_req_size = CIK_MEC_UCODE_SIZE * 4; + rlc_req_size = BONAIRE_RLC_UCODE_SIZE * 4; + mc_req_size = HAWAII_MC_UCODE_SIZE * 4; + mc2_req_size = HAWAII_MC2_UCODE_SIZE * 4; + sdma_req_size = CIK_SDMA_UCODE_SIZE * 4; + smc_req_size = ALIGN(HAWAII_SMC_UCODE_SIZE, 4); break; case CHIP_KAVERI: chip_name = "KAVERI"; @@ -788,13 +1892,22 @@ static int cik_init_microcode(struct radeon_device *rdev) rlc_req_size = KB_RLC_UCODE_SIZE * 4; sdma_req_size = CIK_SDMA_UCODE_SIZE * 4; break; + case CHIP_MULLINS: + chip_name = "MULLINS"; + pfp_req_size = CIK_PFP_UCODE_SIZE * 4; + me_req_size = CIK_ME_UCODE_SIZE * 4; + ce_req_size = CIK_CE_UCODE_SIZE * 4; + mec_req_size = CIK_MEC_UCODE_SIZE * 4; + rlc_req_size = ML_RLC_UCODE_SIZE * 4; + sdma_req_size = CIK_SDMA_UCODE_SIZE * 4; + break; default: BUG(); } DRM_INFO("Loading %s Microcode\n", chip_name); snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name); - err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->pfp_fw->size != pfp_req_size) { @@ -806,7 +1919,7 @@ static int cik_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name); - err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->me_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->me_fw->size != me_req_size) { @@ -817,7 +1930,7 @@ static int cik_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name); - err = request_firmware(&rdev->ce_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->ce_fw->size != ce_req_size) { @@ -828,7 +1941,7 @@ static int cik_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec.bin", chip_name); - err = request_firmware(&rdev->mec_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->mec_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->mec_fw->size != mec_req_size) { @@ -839,7 +1952,7 @@ static int cik_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name); - err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->rlc_fw->size != rlc_req_size) { @@ -850,7 +1963,7 @@ static int cik_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_sdma.bin", chip_name); - err = request_firmware(&rdev->sdma_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->sdma_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->sdma_fw->size != sdma_req_size) { @@ -860,23 +1973,43 @@ static int cik_init_microcode(struct radeon_device *rdev) err = -EINVAL; } - /* No MC ucode on APUs */ + /* No SMC, MC ucode on APUs */ if (!(rdev->flags & RADEON_IS_IGP)) { - snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); - err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->mc_fw->size != mc_req_size) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); + if (err) + goto out; + } + if ((rdev->mc_fw->size != mc_req_size) && + (rdev->mc_fw->size != mc2_req_size)){ printk(KERN_ERR "cik_mc: Bogus length %zu in firmware \"%s\"\n", rdev->mc_fw->size, fw_name); err = -EINVAL; } + DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size); + + snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name); + err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev); + if (err) { + printk(KERN_ERR + "smc: error loading firmware \"%s\"\n", + fw_name); + release_firmware(rdev->smc_fw); + rdev->smc_fw = NULL; + err = 0; + } else if (rdev->smc_fw->size != smc_req_size) { + printk(KERN_ERR + "cik_smc: Bogus length %zu in firmware \"%s\"\n", + rdev->smc_fw->size, fw_name); + err = -EINVAL; + } } out: - platform_device_unregister(pdev); - if (err) { if (err != -EINVAL) printk(KERN_ERR @@ -892,6 +2025,8 @@ out: rdev->rlc_fw = NULL; release_firmware(rdev->mc_fw); rdev->mc_fw = NULL; + release_firmware(rdev->smc_fw); + rdev->smc_fw = NULL; } return err; } @@ -934,9 +2069,232 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) num_pipe_configs = rdev->config.cik.max_tile_pipes; if (num_pipe_configs > 8) - num_pipe_configs = 8; /* ??? */ + num_pipe_configs = 16; - if (num_pipe_configs == 8) { + if (num_pipe_configs == 16) { + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { + switch (reg_offset) { + case 0: + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B)); + break; + case 1: + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B)); + break; + case 2: + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B)); + break; + case 3: + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B)); + break; + case 4: + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(split_equal_to_row_size)); + break; + case 5: + gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + break; + case 6: + gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B)); + break; + case 7: + gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(split_equal_to_row_size)); + break; + case 8: + gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16)); + break; + case 9: + gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); + break; + case 10: + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + break; + case 11: + gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_8x16) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + break; + case 12: + gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + break; + case 13: + gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); + break; + case 14: + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + break; + case 16: + gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_8x16) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + break; + case 17: + gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + break; + case 27: + gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); + break; + case 28: + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + break; + case 29: + gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_8x16) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + break; + case 30: + gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + break; + default: + gb_tile_moden = 0; + break; + } + rdev->config.cik.tile_mode_array[reg_offset] = gb_tile_moden; + WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden); + } + for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) { + switch (reg_offset) { + case 0: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + break; + case 1: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + break; + case 2: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_16_BANK)); + break; + case 3: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_16_BANK)); + break; + case 4: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_8_BANK)); + break; + case 5: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + break; + case 6: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_2_BANK)); + break; + case 8: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + break; + case 9: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + break; + case 10: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_16_BANK)); + break; + case 11: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_8_BANK)); + break; + case 12: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + break; + case 13: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_2_BANK)); + break; + case 14: + gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_2_BANK)); + break; + default: + gb_tile_moden = 0; + break; + } + rdev->config.cik.macrotile_mode_array[reg_offset] = gb_tile_moden; + WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden); + } + } else if (num_pipe_configs == 8) { for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { switch (reg_offset) { case 0: @@ -971,6 +2329,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 5: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); break; case 6: @@ -991,6 +2350,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 9: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); break; case 10: @@ -1013,6 +2373,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 13: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); break; case 14: @@ -1035,6 +2396,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 27: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); break; case 28: @@ -1152,6 +2514,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) gb_tile_moden = 0; break; } + rdev->config.cik.macrotile_mode_array[reg_offset] = gb_tile_moden; WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden); } } else if (num_pipe_configs == 4) { @@ -1190,6 +2553,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 5: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); break; case 6: @@ -1210,6 +2574,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 9: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); break; case 10: @@ -1232,6 +2597,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 13: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); break; case 14: @@ -1254,6 +2620,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 27: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); break; case 28: @@ -1316,6 +2683,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 5: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); break; case 6: @@ -1336,6 +2704,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 9: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); break; case 10: @@ -1358,6 +2727,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 13: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); break; case 14: @@ -1380,6 +2750,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 27: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); break; case 28: @@ -1498,6 +2869,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) gb_tile_moden = 0; break; } + rdev->config.cik.macrotile_mode_array[reg_offset] = gb_tile_moden; WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden); } } else if (num_pipe_configs == 2) { @@ -1535,6 +2907,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 5: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); break; case 6: @@ -1550,11 +2923,13 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) TILE_SPLIT(split_equal_to_row_size)); break; case 8: - gb_tile_moden = ARRAY_MODE(ARRAY_LINEAR_ALIGNED); + gb_tile_moden = ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + PIPE_CONFIG(ADDR_SURF_P2); break; case 9: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P2)); break; case 10: gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | @@ -1576,6 +2951,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 13: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); break; case 14: @@ -1598,7 +2974,8 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) break; case 27: gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P2)); break; case 28: gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) | @@ -1715,6 +3092,7 @@ static void cik_tiling_mode_table_init(struct radeon_device *rdev) gb_tile_moden = 0; break; } + rdev->config.cik.macrotile_mode_array[reg_offset] = gb_tile_moden; WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden); } } else @@ -1768,7 +3146,7 @@ static u32 cik_create_bitmask(u32 bit_width) } /** - * cik_select_se_sh - select which SE, SH to address + * cik_get_rb_disabled - computes the mask of disabled RBs * * @rdev: radeon_device pointer * @max_rb_num: max RBs (render backends) for the asic @@ -1779,7 +3157,7 @@ static u32 cik_create_bitmask(u32 bit_width) * Returns the disabled RB bitmask. */ static u32 cik_get_rb_disabled(struct radeon_device *rdev, - u32 max_rb_num, u32 se_num, + u32 max_rb_num_per_se, u32 sh_per_se) { u32 data, mask; @@ -1793,7 +3171,7 @@ static u32 cik_get_rb_disabled(struct radeon_device *rdev, data >>= BACKEND_DISABLE_SHIFT; - mask = cik_create_bitmask(max_rb_num / se_num / sh_per_se); + mask = cik_create_bitmask(max_rb_num_per_se / sh_per_se); return data & mask; } @@ -1810,7 +3188,7 @@ static u32 cik_get_rb_disabled(struct radeon_device *rdev, */ static void cik_setup_rb(struct radeon_device *rdev, u32 se_num, u32 sh_per_se, - u32 max_rb_num) + u32 max_rb_num_per_se) { int i, j; u32 data, mask; @@ -1820,24 +3198,35 @@ static void cik_setup_rb(struct radeon_device *rdev, for (i = 0; i < se_num; i++) { for (j = 0; j < sh_per_se; j++) { cik_select_se_sh(rdev, i, j); - data = cik_get_rb_disabled(rdev, max_rb_num, se_num, sh_per_se); - disabled_rbs |= data << ((i * sh_per_se + j) * CIK_RB_BITMAP_WIDTH_PER_SH); + data = cik_get_rb_disabled(rdev, max_rb_num_per_se, sh_per_se); + if (rdev->family == CHIP_HAWAII) + disabled_rbs |= data << ((i * sh_per_se + j) * HAWAII_RB_BITMAP_WIDTH_PER_SH); + else + disabled_rbs |= data << ((i * sh_per_se + j) * CIK_RB_BITMAP_WIDTH_PER_SH); } } cik_select_se_sh(rdev, 0xffffffff, 0xffffffff); mask = 1; - for (i = 0; i < max_rb_num; i++) { + for (i = 0; i < max_rb_num_per_se * se_num; i++) { if (!(disabled_rbs & mask)) enabled_rbs |= mask; mask <<= 1; } + rdev->config.cik.backend_enable_mask = enabled_rbs; + for (i = 0; i < se_num; i++) { cik_select_se_sh(rdev, i, 0xffffffff); data = 0; for (j = 0; j < sh_per_se; j++) { switch (enabled_rbs & 3) { + case 0: + if (j == 0) + data |= PKR_MAP(RASTER_CONFIG_RB_MAP_3); + else + data |= PKR_MAP(RASTER_CONFIG_RB_MAP_0); + break; case 1: data |= (RASTER_CONFIG_RB_MAP_0 << (i * sh_per_se + j) * 2); break; @@ -1870,7 +3259,7 @@ static void cik_gpu_init(struct radeon_device *rdev) u32 mc_shared_chmap, mc_arb_ramcfg; u32 hdp_host_path_cntl; u32 tmp; - int i, j; + int i, j, k; switch (rdev->family) { case CHIP_BONAIRE: @@ -1890,10 +3279,68 @@ static void cik_gpu_init(struct radeon_device *rdev) rdev->config.cik.sc_earlyz_tile_fifo_size = 0x130; gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; break; + case CHIP_HAWAII: + rdev->config.cik.max_shader_engines = 4; + rdev->config.cik.max_tile_pipes = 16; + rdev->config.cik.max_cu_per_sh = 11; + rdev->config.cik.max_sh_per_se = 1; + rdev->config.cik.max_backends_per_se = 4; + rdev->config.cik.max_texture_channel_caches = 16; + rdev->config.cik.max_gprs = 256; + rdev->config.cik.max_gs_threads = 32; + rdev->config.cik.max_hw_contexts = 8; + + rdev->config.cik.sc_prim_fifo_size_frontend = 0x20; + rdev->config.cik.sc_prim_fifo_size_backend = 0x100; + rdev->config.cik.sc_hiz_tile_fifo_size = 0x30; + rdev->config.cik.sc_earlyz_tile_fifo_size = 0x130; + gb_addr_config = HAWAII_GB_ADDR_CONFIG_GOLDEN; + break; case CHIP_KAVERI: - /* TODO */ + rdev->config.cik.max_shader_engines = 1; + rdev->config.cik.max_tile_pipes = 4; + if ((rdev->pdev->device == 0x1304) || + (rdev->pdev->device == 0x1305) || + (rdev->pdev->device == 0x130C) || + (rdev->pdev->device == 0x130F) || + (rdev->pdev->device == 0x1310) || + (rdev->pdev->device == 0x1311) || + (rdev->pdev->device == 0x131C)) { + rdev->config.cik.max_cu_per_sh = 8; + rdev->config.cik.max_backends_per_se = 2; + } else if ((rdev->pdev->device == 0x1309) || + (rdev->pdev->device == 0x130A) || + (rdev->pdev->device == 0x130D) || + (rdev->pdev->device == 0x1313) || + (rdev->pdev->device == 0x131D)) { + rdev->config.cik.max_cu_per_sh = 6; + rdev->config.cik.max_backends_per_se = 2; + } else if ((rdev->pdev->device == 0x1306) || + (rdev->pdev->device == 0x1307) || + (rdev->pdev->device == 0x130B) || + (rdev->pdev->device == 0x130E) || + (rdev->pdev->device == 0x1315) || + (rdev->pdev->device == 0x131B)) { + rdev->config.cik.max_cu_per_sh = 4; + rdev->config.cik.max_backends_per_se = 1; + } else { + rdev->config.cik.max_cu_per_sh = 3; + rdev->config.cik.max_backends_per_se = 1; + } + rdev->config.cik.max_sh_per_se = 1; + rdev->config.cik.max_texture_channel_caches = 4; + rdev->config.cik.max_gprs = 256; + rdev->config.cik.max_gs_threads = 16; + rdev->config.cik.max_hw_contexts = 8; + + rdev->config.cik.sc_prim_fifo_size_frontend = 0x20; + rdev->config.cik.sc_prim_fifo_size_backend = 0x100; + rdev->config.cik.sc_hiz_tile_fifo_size = 0x30; + rdev->config.cik.sc_earlyz_tile_fifo_size = 0x130; + gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; break; case CHIP_KABINI: + case CHIP_MULLINS: default: rdev->config.cik.max_shader_engines = 1; rdev->config.cik.max_tile_pipes = 2; @@ -1979,10 +3426,8 @@ static void cik_gpu_init(struct radeon_device *rdev) rdev->config.cik.tile_config |= (3 << 0); break; } - if ((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT) - rdev->config.cik.tile_config |= 1 << 4; - else - rdev->config.cik.tile_config |= 0 << 4; + rdev->config.cik.tile_config |= + ((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT) << 4; rdev->config.cik.tile_config |= ((gb_addr_config & PIPE_INTERLEAVE_SIZE_MASK) >> PIPE_INTERLEAVE_SIZE_SHIFT) << 8; rdev->config.cik.tile_config |= @@ -2003,6 +3448,15 @@ static void cik_gpu_init(struct radeon_device *rdev) rdev->config.cik.max_sh_per_se, rdev->config.cik.max_backends_per_se); + for (i = 0; i < rdev->config.cik.max_shader_engines; i++) { + for (j = 0; j < rdev->config.cik.max_sh_per_se; j++) { + for (k = 0; k < rdev->config.cik.max_cu_per_sh; k++) { + rdev->config.cik.active_cus += + hweight32(cik_get_cu_active_bitmap(rdev, i, j)); + } + } + } + /* set HW defaults for 3D engine */ WREG32(CP_MEQ_THRESHOLDS, MEQ1_START(0x30) | MEQ2_START(0x60)); @@ -2143,6 +3597,51 @@ int cik_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) } /** + * cik_hdp_flush_cp_ring_emit - emit an hdp flush on the cp + * + * @rdev: radeon_device pointer + * @ridx: radeon ring index + * + * Emits an hdp flush on the cp. + */ +static void cik_hdp_flush_cp_ring_emit(struct radeon_device *rdev, + int ridx) +{ + struct radeon_ring *ring = &rdev->ring[ridx]; + u32 ref_and_mask; + + switch (ring->idx) { + case CAYMAN_RING_TYPE_CP1_INDEX: + case CAYMAN_RING_TYPE_CP2_INDEX: + default: + switch (ring->me) { + case 0: + ref_and_mask = CP2 << ring->pipe; + break; + case 1: + ref_and_mask = CP6 << ring->pipe; + break; + default: + return; + } + break; + case RADEON_RING_TYPE_GFX_INDEX: + ref_and_mask = CP0; + break; + } + + radeon_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5)); + radeon_ring_write(ring, (WAIT_REG_MEM_OPERATION(1) | /* write, wait, write */ + WAIT_REG_MEM_FUNCTION(3) | /* == */ + WAIT_REG_MEM_ENGINE(1))); /* pfp */ + radeon_ring_write(ring, GPU_HDP_FLUSH_REQ >> 2); + radeon_ring_write(ring, GPU_HDP_FLUSH_DONE >> 2); + radeon_ring_write(ring, ref_and_mask); + radeon_ring_write(ring, ref_and_mask); + radeon_ring_write(ring, 0x20); /* poll interval */ +} + +/** * cik_fence_gfx_ring_emit - emit a fence on the gfx ring * * @rdev: radeon_device pointer @@ -2168,15 +3667,7 @@ void cik_fence_gfx_ring_emit(struct radeon_device *rdev, radeon_ring_write(ring, fence->seq); radeon_ring_write(ring, 0); /* HDP flush */ - /* We should be using the new WAIT_REG_MEM special op packet here - * but it causes the CP to hang - */ - radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | - WRITE_DATA_DST_SEL(0))); - radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); + cik_hdp_flush_cp_ring_emit(rdev, fence->ring); } /** @@ -2206,18 +3697,10 @@ void cik_fence_compute_ring_emit(struct radeon_device *rdev, radeon_ring_write(ring, fence->seq); radeon_ring_write(ring, 0); /* HDP flush */ - /* We should be using the new WAIT_REG_MEM special op packet here - * but it causes the CP to hang - */ - radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | - WRITE_DATA_DST_SEL(0))); - radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); + cik_hdp_flush_cp_ring_emit(rdev, fence->ring); } -void cik_semaphore_ring_emit(struct radeon_device *rdev, +bool cik_semaphore_ring_emit(struct radeon_device *rdev, struct radeon_ring *ring, struct radeon_semaphore *semaphore, bool emit_wait) @@ -2226,8 +3709,85 @@ void cik_semaphore_ring_emit(struct radeon_device *rdev, unsigned sel = emit_wait ? PACKET3_SEM_SEL_WAIT : PACKET3_SEM_SEL_SIGNAL; radeon_ring_write(ring, PACKET3(PACKET3_MEM_SEMAPHORE, 1)); - radeon_ring_write(ring, addr & 0xffffffff); + radeon_ring_write(ring, lower_32_bits(addr)); radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | sel); + + return true; +} + +/** + * cik_copy_cpdma - copy pages using the CP DMA engine + * + * @rdev: radeon_device pointer + * @src_offset: src GPU address + * @dst_offset: dst GPU address + * @num_gpu_pages: number of GPU pages to xfer + * @fence: radeon fence object + * + * Copy GPU paging using the CP DMA engine (CIK+). + * Used by the radeon ttm implementation to move pages if + * registered as the asic copy callback. + */ +int cik_copy_cpdma(struct radeon_device *rdev, + uint64_t src_offset, uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence **fence) +{ + struct radeon_semaphore *sem = NULL; + int ring_index = rdev->asic->copy.blit_ring_index; + struct radeon_ring *ring = &rdev->ring[ring_index]; + u32 size_in_bytes, cur_size_in_bytes, control; + int i, num_loops; + int r = 0; + + r = radeon_semaphore_create(rdev, &sem); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + return r; + } + + size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT); + num_loops = DIV_ROUND_UP(size_in_bytes, 0x1fffff); + r = radeon_ring_lock(rdev, ring, num_loops * 7 + 18); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_semaphore_sync_to(sem, *fence); + radeon_semaphore_sync_rings(rdev, sem, ring->idx); + + for (i = 0; i < num_loops; i++) { + cur_size_in_bytes = size_in_bytes; + if (cur_size_in_bytes > 0x1fffff) + cur_size_in_bytes = 0x1fffff; + size_in_bytes -= cur_size_in_bytes; + control = 0; + if (size_in_bytes == 0) + control |= PACKET3_DMA_DATA_CP_SYNC; + radeon_ring_write(ring, PACKET3(PACKET3_DMA_DATA, 5)); + radeon_ring_write(ring, control); + radeon_ring_write(ring, lower_32_bits(src_offset)); + radeon_ring_write(ring, upper_32_bits(src_offset)); + radeon_ring_write(ring, lower_32_bits(dst_offset)); + radeon_ring_write(ring, upper_32_bits(dst_offset)); + radeon_ring_write(ring, cur_size_in_bytes); + src_offset += cur_size_in_bytes; + dst_offset += cur_size_in_bytes; + } + + r = radeon_fence_emit(rdev, fence, ring->idx); + if (r) { + radeon_ring_unlock_undo(rdev, ring); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_ring_unlock_commit(rdev, ring); + radeon_semaphore_free(rdev, &sem, *fence); + + return r; } /* @@ -2269,7 +3829,7 @@ void cik_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); radeon_ring_write(ring, WRITE_DATA_DST_SEL(1)); radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff); + radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr)); radeon_ring_write(ring, next_rptr); } @@ -2316,6 +3876,7 @@ int cik_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256); if (r) { DRM_ERROR("radeon: failed to get ib (%d).\n", r); + radeon_scratch_free(rdev, scratch); return r; } ib.ptr[0] = PACKET3(PACKET3_SET_UCONFIG_REG, 1); @@ -2332,6 +3893,8 @@ int cik_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) r = radeon_fence_wait(ib.fence, false); if (r) { DRM_ERROR("radeon: fence wait failed (%d).\n", r); + radeon_scratch_free(rdev, scratch); + radeon_ib_free(rdev, &ib); return r; } for (i = 0; i < rdev->usec_timeout; i++) { @@ -2388,6 +3951,8 @@ static void cik_cp_gfx_enable(struct radeon_device *rdev, bool enable) if (enable) WREG32(CP_ME_CNTL, 0); else { + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT)); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; } @@ -2533,7 +4098,8 @@ static int cik_cp_gfx_resume(struct radeon_device *rdev) int r; WREG32(CP_SEM_WAIT_TIMER, 0x0); - WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0); + if (rdev->family != CHIP_HAWAII) + WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0); /* Set the write pointer delay */ WREG32(CP_RB_WPTR_DELAY, 0); @@ -2546,8 +4112,8 @@ static int cik_cp_gfx_resume(struct radeon_device *rdev) /* ring 0 - compute and gfx */ /* Set ring buffer size */ ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - rb_bufsz = drm_order(ring->ring_size / 8); - tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; + rb_bufsz = order_base_2(ring->ring_size / 8); + tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif @@ -2575,8 +4141,6 @@ static int cik_cp_gfx_resume(struct radeon_device *rdev) WREG32(CP_RB0_BASE, rb_addr); WREG32(CP_RB0_BASE_HI, upper_32_bits(rb_addr)); - ring->rptr = RREG32(CP_RB0_RPTR); - /* start the ring */ cik_cp_gfx_start(rdev); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true; @@ -2585,52 +4149,86 @@ static int cik_cp_gfx_resume(struct radeon_device *rdev) rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; return r; } + + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); + return 0; } -u32 cik_compute_ring_get_rptr(struct radeon_device *rdev, - struct radeon_ring *ring) +u32 cik_gfx_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) { u32 rptr; + if (rdev->wb.enabled) + rptr = rdev->wb.wb[ring->rptr_offs/4]; + else + rptr = RREG32(CP_RB0_RPTR); + + return rptr; +} + +u32 cik_gfx_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 wptr; + + wptr = RREG32(CP_RB0_WPTR); + + return wptr; +} + +void cik_gfx_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + WREG32(CP_RB0_WPTR, ring->wptr); + (void)RREG32(CP_RB0_WPTR); +} +u32 cik_compute_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 rptr; if (rdev->wb.enabled) { - rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]); + rptr = rdev->wb.wb[ring->rptr_offs/4]; } else { + mutex_lock(&rdev->srbm_mutex); cik_srbm_select(rdev, ring->me, ring->pipe, ring->queue, 0); rptr = RREG32(CP_HQD_PQ_RPTR); cik_srbm_select(rdev, 0, 0, 0, 0); + mutex_unlock(&rdev->srbm_mutex); } - rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift; return rptr; } -u32 cik_compute_ring_get_wptr(struct radeon_device *rdev, - struct radeon_ring *ring) +u32 cik_compute_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) { u32 wptr; if (rdev->wb.enabled) { - wptr = le32_to_cpu(rdev->wb.wb[ring->wptr_offs/4]); + /* XXX check if swapping is necessary on BE */ + wptr = rdev->wb.wb[ring->wptr_offs/4]; } else { + mutex_lock(&rdev->srbm_mutex); cik_srbm_select(rdev, ring->me, ring->pipe, ring->queue, 0); wptr = RREG32(CP_HQD_PQ_WPTR); cik_srbm_select(rdev, 0, 0, 0, 0); + mutex_unlock(&rdev->srbm_mutex); } - wptr = (wptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift; return wptr; } -void cik_compute_ring_set_wptr(struct radeon_device *rdev, - struct radeon_ring *ring) +void cik_compute_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) { - u32 wptr = (ring->wptr << ring->ptr_reg_shift) & ring->ptr_reg_mask; - - rdev->wb.wb[ring->wptr_offs/4] = cpu_to_le32(wptr); - WDOORBELL32(ring->doorbell_offset, wptr); + /* XXX check if swapping is necessary on BE */ + rdev->wb.wb[ring->wptr_offs/4] = ring->wptr; + WDOORBELL32(ring->doorbell_index, ring->wptr); } /** @@ -2645,8 +4243,11 @@ static void cik_cp_compute_enable(struct radeon_device *rdev, bool enable) { if (enable) WREG32(CP_MEC_CNTL, 0); - else + else { WREG32(CP_MEC_CNTL, (MEC_ME1_HALT | MEC_ME2_HALT)); + rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false; + } udelay(50); } @@ -2908,6 +4509,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev) WREG32(CP_CPF_DEBUG, tmp); /* init the pipes */ + mutex_lock(&rdev->srbm_mutex); for (i = 0; i < (rdev->mec.num_pipe * rdev->mec.num_mec); i++) { int me = (i < 4) ? 1 : 2; int pipe = (i < 4) ? i : (i - 4); @@ -2926,10 +4528,11 @@ static int cik_cp_compute_resume(struct radeon_device *rdev) /* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */ tmp = RREG32(CP_HPD_EOP_CONTROL); tmp &= ~EOP_SIZE_MASK; - tmp |= drm_order(MEC_HPD_SIZE / 8); + tmp |= order_base_2(MEC_HPD_SIZE / 8); WREG32(CP_HPD_EOP_CONTROL, tmp); } cik_srbm_select(rdev, 0, 0, 0, 0); + mutex_unlock(&rdev->srbm_mutex); /* init the queues. Just two for now. */ for (i = 0; i < 2; i++) { @@ -2969,10 +4572,6 @@ static int cik_cp_compute_resume(struct radeon_device *rdev) return r; } - /* doorbell offset */ - rdev->ring[idx].doorbell_offset = - (rdev->ring[idx].doorbell_page_num * PAGE_SIZE) + 0; - /* init the mqd struct */ memset(buf, 0, sizeof(struct bonaire_mqd)); @@ -2983,6 +4582,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev) mqd->static_thread_mgmt23[0] = 0xffffffff; mqd->static_thread_mgmt23[1] = 0xffffffff; + mutex_lock(&rdev->srbm_mutex); cik_srbm_select(rdev, rdev->ring[idx].me, rdev->ring[idx].pipe, rdev->ring[idx].queue, 0); @@ -3041,9 +4641,9 @@ static int cik_cp_compute_resume(struct radeon_device *rdev) ~(QUEUE_SIZE_MASK | RPTR_BLOCK_SIZE_MASK); mqd->queue_state.cp_hqd_pq_control |= - drm_order(rdev->ring[idx].ring_size / 8); + order_base_2(rdev->ring[idx].ring_size / 8); mqd->queue_state.cp_hqd_pq_control |= - (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8); + (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8); #ifdef __BIG_ENDIAN mqd->queue_state.cp_hqd_pq_control |= BUF_SWAP_32BIT; #endif @@ -3083,7 +4683,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev) RREG32(CP_HQD_PQ_DOORBELL_CONTROL); mqd->queue_state.cp_hqd_pq_doorbell_control &= ~DOORBELL_OFFSET_MASK; mqd->queue_state.cp_hqd_pq_doorbell_control |= - DOORBELL_OFFSET(rdev->ring[idx].doorbell_offset / 4); + DOORBELL_OFFSET(rdev->ring[idx].doorbell_index); mqd->queue_state.cp_hqd_pq_doorbell_control |= DOORBELL_EN; mqd->queue_state.cp_hqd_pq_doorbell_control &= ~(DOORBELL_SOURCE | DOORBELL_HIT); @@ -3098,8 +4698,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev) rdev->ring[idx].wptr = 0; mqd->queue_state.cp_hqd_pq_wptr = rdev->ring[idx].wptr; WREG32(CP_HQD_PQ_WPTR, mqd->queue_state.cp_hqd_pq_wptr); - rdev->ring[idx].rptr = RREG32(CP_HQD_PQ_RPTR); - mqd->queue_state.cp_hqd_pq_rptr = rdev->ring[idx].rptr; + mqd->queue_state.cp_hqd_pq_rptr = RREG32(CP_HQD_PQ_RPTR); /* set the vmid for the queue */ mqd->queue_state.cp_hqd_vmid = 0; @@ -3110,6 +4709,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev) WREG32(CP_HQD_ACTIVE, mqd->queue_state.cp_hqd_active); cik_srbm_select(rdev, 0, 0, 0, 0); + mutex_unlock(&rdev->srbm_mutex); radeon_bo_kunmap(rdev->ring[idx].mqd_obj); radeon_bo_unreserve(rdev->ring[idx].mqd_obj); @@ -3153,12 +4753,7 @@ static int cik_cp_resume(struct radeon_device *rdev) { int r; - /* Reset all cp blocks */ - WREG32(GRBM_SOFT_RESET, SOFT_RESET_CP); - RREG32(GRBM_SOFT_RESET); - mdelay(15); - WREG32(GRBM_SOFT_RESET, 0); - RREG32(GRBM_SOFT_RESET); + cik_enable_gui_idle_interrupt(rdev, false); r = cik_cp_load_microcode(rdev); if (r) @@ -3171,582 +4766,11 @@ static int cik_cp_resume(struct radeon_device *rdev) if (r) return r; - return 0; -} - -/* - * sDMA - System DMA - * Starting with CIK, the GPU has new asynchronous - * DMA engines. These engines are used for compute - * and gfx. There are two DMA engines (SDMA0, SDMA1) - * and each one supports 1 ring buffer used for gfx - * and 2 queues used for compute. - * - * The programming model is very similar to the CP - * (ring buffer, IBs, etc.), but sDMA has it's own - * packet format that is different from the PM4 format - * used by the CP. sDMA supports copying data, writing - * embedded data, solid fills, and a number of other - * things. It also has support for tiling/detiling of - * buffers. - */ -/** - * cik_sdma_ring_ib_execute - Schedule an IB on the DMA engine - * - * @rdev: radeon_device pointer - * @ib: IB object to schedule - * - * Schedule an IB in the DMA ring (CIK). - */ -void cik_sdma_ring_ib_execute(struct radeon_device *rdev, - struct radeon_ib *ib) -{ - struct radeon_ring *ring = &rdev->ring[ib->ring]; - u32 extra_bits = (ib->vm ? ib->vm->id : 0) & 0xf; - - if (rdev->wb.enabled) { - u32 next_rptr = ring->wptr + 5; - while ((next_rptr & 7) != 4) - next_rptr++; - next_rptr += 4; - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0)); - radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff); - radeon_ring_write(ring, 1); /* number of DWs to follow */ - radeon_ring_write(ring, next_rptr); - } - - /* IB packet must end on a 8 DW boundary */ - while ((ring->wptr & 7) != 4) - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0)); - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_INDIRECT_BUFFER, 0, extra_bits)); - radeon_ring_write(ring, ib->gpu_addr & 0xffffffe0); /* base must be 32 byte aligned */ - radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xffffffff); - radeon_ring_write(ring, ib->length_dw); - -} - -/** - * cik_sdma_fence_ring_emit - emit a fence on the DMA ring - * - * @rdev: radeon_device pointer - * @fence: radeon fence object - * - * Add a DMA fence packet to the ring to write - * the fence seq number and DMA trap packet to generate - * an interrupt if needed (CIK). - */ -void cik_sdma_fence_ring_emit(struct radeon_device *rdev, - struct radeon_fence *fence) -{ - struct radeon_ring *ring = &rdev->ring[fence->ring]; - u64 addr = rdev->fence_drv[fence->ring].gpu_addr; - u32 extra_bits = (SDMA_POLL_REG_MEM_EXTRA_OP(1) | - SDMA_POLL_REG_MEM_EXTRA_FUNC(3)); /* == */ - u32 ref_and_mask; - - if (fence->ring == R600_RING_TYPE_DMA_INDEX) - ref_and_mask = SDMA0; - else - ref_and_mask = SDMA1; - - /* write the fence */ - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_FENCE, 0, 0)); - radeon_ring_write(ring, addr & 0xffffffff); - radeon_ring_write(ring, upper_32_bits(addr) & 0xffffffff); - radeon_ring_write(ring, fence->seq); - /* generate an interrupt */ - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_TRAP, 0, 0)); - /* flush HDP */ - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_POLL_REG_MEM, 0, extra_bits)); - radeon_ring_write(ring, GPU_HDP_FLUSH_DONE); - radeon_ring_write(ring, GPU_HDP_FLUSH_REQ); - radeon_ring_write(ring, ref_and_mask); /* REFERENCE */ - radeon_ring_write(ring, ref_and_mask); /* MASK */ - radeon_ring_write(ring, (4 << 16) | 10); /* RETRY_COUNT, POLL_INTERVAL */ -} - -/** - * cik_sdma_semaphore_ring_emit - emit a semaphore on the dma ring - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * @semaphore: radeon semaphore object - * @emit_wait: wait or signal semaphore - * - * Add a DMA semaphore packet to the ring wait on or signal - * other rings (CIK). - */ -void cik_sdma_semaphore_ring_emit(struct radeon_device *rdev, - struct radeon_ring *ring, - struct radeon_semaphore *semaphore, - bool emit_wait) -{ - u64 addr = semaphore->gpu_addr; - u32 extra_bits = emit_wait ? 0 : SDMA_SEMAPHORE_EXTRA_S; - - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SEMAPHORE, 0, extra_bits)); - radeon_ring_write(ring, addr & 0xfffffff8); - radeon_ring_write(ring, upper_32_bits(addr) & 0xffffffff); -} - -/** - * cik_sdma_gfx_stop - stop the gfx async dma engines - * - * @rdev: radeon_device pointer - * - * Stop the gfx async dma ring buffers (CIK). - */ -static void cik_sdma_gfx_stop(struct radeon_device *rdev) -{ - u32 rb_cntl, reg_offset; - int i; - - radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); - - for (i = 0; i < 2; i++) { - if (i == 0) - reg_offset = SDMA0_REGISTER_OFFSET; - else - reg_offset = SDMA1_REGISTER_OFFSET; - rb_cntl = RREG32(SDMA0_GFX_RB_CNTL + reg_offset); - rb_cntl &= ~SDMA_RB_ENABLE; - WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl); - WREG32(SDMA0_GFX_IB_CNTL + reg_offset, 0); - } -} - -/** - * cik_sdma_rlc_stop - stop the compute async dma engines - * - * @rdev: radeon_device pointer - * - * Stop the compute async dma queues (CIK). - */ -static void cik_sdma_rlc_stop(struct radeon_device *rdev) -{ - /* XXX todo */ -} - -/** - * cik_sdma_enable - stop the async dma engines - * - * @rdev: radeon_device pointer - * @enable: enable/disable the DMA MEs. - * - * Halt or unhalt the async dma engines (CIK). - */ -static void cik_sdma_enable(struct radeon_device *rdev, bool enable) -{ - u32 me_cntl, reg_offset; - int i; - - for (i = 0; i < 2; i++) { - if (i == 0) - reg_offset = SDMA0_REGISTER_OFFSET; - else - reg_offset = SDMA1_REGISTER_OFFSET; - me_cntl = RREG32(SDMA0_ME_CNTL + reg_offset); - if (enable) - me_cntl &= ~SDMA_HALT; - else - me_cntl |= SDMA_HALT; - WREG32(SDMA0_ME_CNTL + reg_offset, me_cntl); - } -} - -/** - * cik_sdma_gfx_resume - setup and start the async dma engines - * - * @rdev: radeon_device pointer - * - * Set up the gfx DMA ring buffers and enable them (CIK). - * Returns 0 for success, error for failure. - */ -static int cik_sdma_gfx_resume(struct radeon_device *rdev) -{ - struct radeon_ring *ring; - u32 rb_cntl, ib_cntl; - u32 rb_bufsz; - u32 reg_offset, wb_offset; - int i, r; - - for (i = 0; i < 2; i++) { - if (i == 0) { - ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; - reg_offset = SDMA0_REGISTER_OFFSET; - wb_offset = R600_WB_DMA_RPTR_OFFSET; - } else { - ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]; - reg_offset = SDMA1_REGISTER_OFFSET; - wb_offset = CAYMAN_WB_DMA1_RPTR_OFFSET; - } - - WREG32(SDMA0_SEM_INCOMPLETE_TIMER_CNTL + reg_offset, 0); - WREG32(SDMA0_SEM_WAIT_FAIL_TIMER_CNTL + reg_offset, 0); - - /* Set ring buffer size in dwords */ - rb_bufsz = drm_order(ring->ring_size / 4); - rb_cntl = rb_bufsz << 1; -#ifdef __BIG_ENDIAN - rb_cntl |= SDMA_RB_SWAP_ENABLE | SDMA_RPTR_WRITEBACK_SWAP_ENABLE; -#endif - WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl); - - /* Initialize the ring buffer's read and write pointers */ - WREG32(SDMA0_GFX_RB_RPTR + reg_offset, 0); - WREG32(SDMA0_GFX_RB_WPTR + reg_offset, 0); - - /* set the wb address whether it's enabled or not */ - WREG32(SDMA0_GFX_RB_RPTR_ADDR_HI + reg_offset, - upper_32_bits(rdev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF); - WREG32(SDMA0_GFX_RB_RPTR_ADDR_LO + reg_offset, - ((rdev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC)); - - if (rdev->wb.enabled) - rb_cntl |= SDMA_RPTR_WRITEBACK_ENABLE; - - WREG32(SDMA0_GFX_RB_BASE + reg_offset, ring->gpu_addr >> 8); - WREG32(SDMA0_GFX_RB_BASE_HI + reg_offset, ring->gpu_addr >> 40); - - ring->wptr = 0; - WREG32(SDMA0_GFX_RB_WPTR + reg_offset, ring->wptr << 2); - - ring->rptr = RREG32(SDMA0_GFX_RB_RPTR + reg_offset) >> 2; - - /* enable DMA RB */ - WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl | SDMA_RB_ENABLE); - - ib_cntl = SDMA_IB_ENABLE; -#ifdef __BIG_ENDIAN - ib_cntl |= SDMA_IB_SWAP_ENABLE; -#endif - /* enable DMA IBs */ - WREG32(SDMA0_GFX_IB_CNTL + reg_offset, ib_cntl); - - ring->ready = true; - - r = radeon_ring_test(rdev, ring->idx, ring); - if (r) { - ring->ready = false; - return r; - } - } - - radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); - - return 0; -} - -/** - * cik_sdma_rlc_resume - setup and start the async dma engines - * - * @rdev: radeon_device pointer - * - * Set up the compute DMA queues and enable them (CIK). - * Returns 0 for success, error for failure. - */ -static int cik_sdma_rlc_resume(struct radeon_device *rdev) -{ - /* XXX todo */ - return 0; -} - -/** - * cik_sdma_load_microcode - load the sDMA ME ucode - * - * @rdev: radeon_device pointer - * - * Loads the sDMA0/1 ucode. - * Returns 0 for success, -EINVAL if the ucode is not available. - */ -static int cik_sdma_load_microcode(struct radeon_device *rdev) -{ - const __be32 *fw_data; - int i; - - if (!rdev->sdma_fw) - return -EINVAL; - - /* stop the gfx rings and rlc compute queues */ - cik_sdma_gfx_stop(rdev); - cik_sdma_rlc_stop(rdev); - - /* halt the MEs */ - cik_sdma_enable(rdev, false); - - /* sdma0 */ - fw_data = (const __be32 *)rdev->sdma_fw->data; - WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0); - for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++) - WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, be32_to_cpup(fw_data++)); - WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION); - - /* sdma1 */ - fw_data = (const __be32 *)rdev->sdma_fw->data; - WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0); - for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++) - WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, be32_to_cpup(fw_data++)); - WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION); - - WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0); - WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0); - return 0; -} - -/** - * cik_sdma_resume - setup and start the async dma engines - * - * @rdev: radeon_device pointer - * - * Set up the DMA engines and enable them (CIK). - * Returns 0 for success, error for failure. - */ -static int cik_sdma_resume(struct radeon_device *rdev) -{ - int r; - - /* Reset dma */ - WREG32(SRBM_SOFT_RESET, SOFT_RESET_SDMA | SOFT_RESET_SDMA1); - RREG32(SRBM_SOFT_RESET); - udelay(50); - WREG32(SRBM_SOFT_RESET, 0); - RREG32(SRBM_SOFT_RESET); - - r = cik_sdma_load_microcode(rdev); - if (r) - return r; - - /* unhalt the MEs */ - cik_sdma_enable(rdev, true); - - /* start the gfx rings and rlc compute queues */ - r = cik_sdma_gfx_resume(rdev); - if (r) - return r; - r = cik_sdma_rlc_resume(rdev); - if (r) - return r; + cik_enable_gui_idle_interrupt(rdev, true); return 0; } -/** - * cik_sdma_fini - tear down the async dma engines - * - * @rdev: radeon_device pointer - * - * Stop the async dma engines and free the rings (CIK). - */ -static void cik_sdma_fini(struct radeon_device *rdev) -{ - /* stop the gfx rings and rlc compute queues */ - cik_sdma_gfx_stop(rdev); - cik_sdma_rlc_stop(rdev); - /* halt the MEs */ - cik_sdma_enable(rdev, false); - radeon_ring_fini(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX]); - radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]); - /* XXX - compute dma queue tear down */ -} - -/** - * cik_copy_dma - copy pages using the DMA engine - * - * @rdev: radeon_device pointer - * @src_offset: src GPU address - * @dst_offset: dst GPU address - * @num_gpu_pages: number of GPU pages to xfer - * @fence: radeon fence object - * - * Copy GPU paging using the DMA engine (CIK). - * Used by the radeon ttm implementation to move pages if - * registered as the asic copy callback. - */ -int cik_copy_dma(struct radeon_device *rdev, - uint64_t src_offset, uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence **fence) -{ - struct radeon_semaphore *sem = NULL; - int ring_index = rdev->asic->copy.dma_ring_index; - struct radeon_ring *ring = &rdev->ring[ring_index]; - u32 size_in_bytes, cur_size_in_bytes; - int i, num_loops; - int r = 0; - - r = radeon_semaphore_create(rdev, &sem); - if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - return r; - } - - size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT); - num_loops = DIV_ROUND_UP(size_in_bytes, 0x1fffff); - r = radeon_ring_lock(rdev, ring, num_loops * 7 + 14); - if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - radeon_semaphore_free(rdev, &sem, NULL); - return r; - } - - if (radeon_fence_need_sync(*fence, ring->idx)) { - radeon_semaphore_sync_rings(rdev, sem, (*fence)->ring, - ring->idx); - radeon_fence_note_sync(*fence, ring->idx); - } else { - radeon_semaphore_free(rdev, &sem, NULL); - } - - for (i = 0; i < num_loops; i++) { - cur_size_in_bytes = size_in_bytes; - if (cur_size_in_bytes > 0x1fffff) - cur_size_in_bytes = 0x1fffff; - size_in_bytes -= cur_size_in_bytes; - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_COPY, SDMA_COPY_SUB_OPCODE_LINEAR, 0)); - radeon_ring_write(ring, cur_size_in_bytes); - radeon_ring_write(ring, 0); /* src/dst endian swap */ - radeon_ring_write(ring, src_offset & 0xffffffff); - radeon_ring_write(ring, upper_32_bits(src_offset) & 0xffffffff); - radeon_ring_write(ring, dst_offset & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xffffffff); - src_offset += cur_size_in_bytes; - dst_offset += cur_size_in_bytes; - } - - r = radeon_fence_emit(rdev, fence, ring->idx); - if (r) { - radeon_ring_unlock_undo(rdev, ring); - return r; - } - - radeon_ring_unlock_commit(rdev, ring); - radeon_semaphore_free(rdev, &sem, *fence); - - return r; -} - -/** - * cik_sdma_ring_test - simple async dma engine test - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Test the DMA engine by writing using it to write an - * value to memory. (CIK). - * Returns 0 for success, error for failure. - */ -int cik_sdma_ring_test(struct radeon_device *rdev, - struct radeon_ring *ring) -{ - unsigned i; - int r; - void __iomem *ptr = (void *)rdev->vram_scratch.ptr; - u32 tmp; - - if (!ptr) { - DRM_ERROR("invalid vram scratch pointer\n"); - return -EINVAL; - } - - tmp = 0xCAFEDEAD; - writel(tmp, ptr); - - r = radeon_ring_lock(rdev, ring, 4); - if (r) { - DRM_ERROR("radeon: dma failed to lock ring %d (%d).\n", ring->idx, r); - return r; - } - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0)); - radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xffffffff); - radeon_ring_write(ring, 1); /* number of DWs to follow */ - radeon_ring_write(ring, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev, ring); - - for (i = 0; i < rdev->usec_timeout; i++) { - tmp = readl(ptr); - if (tmp == 0xDEADBEEF) - break; - DRM_UDELAY(1); - } - - if (i < rdev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); - } else { - DRM_ERROR("radeon: ring %d test failed (0x%08X)\n", - ring->idx, tmp); - r = -EINVAL; - } - return r; -} - -/** - * cik_sdma_ib_test - test an IB on the DMA engine - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Test a simple IB in the DMA ring (CIK). - * Returns 0 on success, error on failure. - */ -int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) -{ - struct radeon_ib ib; - unsigned i; - int r; - void __iomem *ptr = (void *)rdev->vram_scratch.ptr; - u32 tmp = 0; - - if (!ptr) { - DRM_ERROR("invalid vram scratch pointer\n"); - return -EINVAL; - } - - tmp = 0xCAFEDEAD; - writel(tmp, ptr); - - r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256); - if (r) { - DRM_ERROR("radeon: failed to get ib (%d).\n", r); - return r; - } - - ib.ptr[0] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0); - ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc; - ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xffffffff; - ib.ptr[3] = 1; - ib.ptr[4] = 0xDEADBEEF; - ib.length_dw = 5; - - r = radeon_ib_schedule(rdev, &ib, NULL); - if (r) { - radeon_ib_free(rdev, &ib); - DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); - return r; - } - r = radeon_fence_wait(ib.fence, false); - if (r) { - DRM_ERROR("radeon: fence wait failed (%d).\n", r); - return r; - } - for (i = 0; i < rdev->usec_timeout; i++) { - tmp = readl(ptr); - if (tmp == 0xDEADBEEF) - break; - DRM_UDELAY(1); - } - if (i < rdev->usec_timeout) { - DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i); - } else { - DRM_ERROR("radeon: ib test failed (0x%08X)\n", tmp); - r = -EINVAL; - } - radeon_ib_free(rdev, &ib); - return r; -} - - static void cik_print_gpu_status_regs(struct radeon_device *rdev) { dev_info(rdev->dev, " GRBM_STATUS=0x%08X\n", @@ -3796,7 +4820,7 @@ static void cik_print_gpu_status_regs(struct radeon_device *rdev) * mask to be used by cik_gpu_soft_reset(). * Returns a mask of the blocks to be reset. */ -static u32 cik_gpu_check_soft_reset(struct radeon_device *rdev) +u32 cik_gpu_check_soft_reset(struct radeon_device *rdev) { u32 reset_mask = 0; u32 tmp; @@ -3893,6 +4917,10 @@ static void cik_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask) dev_info(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS)); + /* disable CG/PG */ + cik_fini_pg(rdev); + cik_fini_cg(rdev); + /* stop the rlc */ cik_rlc_stop(rdev); @@ -3995,6 +5023,160 @@ static void cik_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask) cik_print_gpu_status_regs(rdev); } +struct kv_reset_save_regs { + u32 gmcon_reng_execute; + u32 gmcon_misc; + u32 gmcon_misc3; +}; + +static void kv_save_regs_for_reset(struct radeon_device *rdev, + struct kv_reset_save_regs *save) +{ + save->gmcon_reng_execute = RREG32(GMCON_RENG_EXECUTE); + save->gmcon_misc = RREG32(GMCON_MISC); + save->gmcon_misc3 = RREG32(GMCON_MISC3); + + WREG32(GMCON_RENG_EXECUTE, save->gmcon_reng_execute & ~RENG_EXECUTE_ON_PWR_UP); + WREG32(GMCON_MISC, save->gmcon_misc & ~(RENG_EXECUTE_ON_REG_UPDATE | + STCTRL_STUTTER_EN)); +} + +static void kv_restore_regs_for_reset(struct radeon_device *rdev, + struct kv_reset_save_regs *save) +{ + int i; + + WREG32(GMCON_PGFSM_WRITE, 0); + WREG32(GMCON_PGFSM_CONFIG, 0x200010ff); + + for (i = 0; i < 5; i++) + WREG32(GMCON_PGFSM_WRITE, 0); + + WREG32(GMCON_PGFSM_WRITE, 0); + WREG32(GMCON_PGFSM_CONFIG, 0x300010ff); + + for (i = 0; i < 5; i++) + WREG32(GMCON_PGFSM_WRITE, 0); + + WREG32(GMCON_PGFSM_WRITE, 0x210000); + WREG32(GMCON_PGFSM_CONFIG, 0xa00010ff); + + for (i = 0; i < 5; i++) + WREG32(GMCON_PGFSM_WRITE, 0); + + WREG32(GMCON_PGFSM_WRITE, 0x21003); + WREG32(GMCON_PGFSM_CONFIG, 0xb00010ff); + + for (i = 0; i < 5; i++) + WREG32(GMCON_PGFSM_WRITE, 0); + + WREG32(GMCON_PGFSM_WRITE, 0x2b00); + WREG32(GMCON_PGFSM_CONFIG, 0xc00010ff); + + for (i = 0; i < 5; i++) + WREG32(GMCON_PGFSM_WRITE, 0); + + WREG32(GMCON_PGFSM_WRITE, 0); + WREG32(GMCON_PGFSM_CONFIG, 0xd00010ff); + + for (i = 0; i < 5; i++) + WREG32(GMCON_PGFSM_WRITE, 0); + + WREG32(GMCON_PGFSM_WRITE, 0x420000); + WREG32(GMCON_PGFSM_CONFIG, 0x100010ff); + + for (i = 0; i < 5; i++) + WREG32(GMCON_PGFSM_WRITE, 0); + + WREG32(GMCON_PGFSM_WRITE, 0x120202); + WREG32(GMCON_PGFSM_CONFIG, 0x500010ff); + + for (i = 0; i < 5; i++) + WREG32(GMCON_PGFSM_WRITE, 0); + + WREG32(GMCON_PGFSM_WRITE, 0x3e3e36); + WREG32(GMCON_PGFSM_CONFIG, 0x600010ff); + + for (i = 0; i < 5; i++) + WREG32(GMCON_PGFSM_WRITE, 0); + + WREG32(GMCON_PGFSM_WRITE, 0x373f3e); + WREG32(GMCON_PGFSM_CONFIG, 0x700010ff); + + for (i = 0; i < 5; i++) + WREG32(GMCON_PGFSM_WRITE, 0); + + WREG32(GMCON_PGFSM_WRITE, 0x3e1332); + WREG32(GMCON_PGFSM_CONFIG, 0xe00010ff); + + WREG32(GMCON_MISC3, save->gmcon_misc3); + WREG32(GMCON_MISC, save->gmcon_misc); + WREG32(GMCON_RENG_EXECUTE, save->gmcon_reng_execute); +} + +static void cik_gpu_pci_config_reset(struct radeon_device *rdev) +{ + struct evergreen_mc_save save; + struct kv_reset_save_regs kv_save = { 0 }; + u32 tmp, i; + + dev_info(rdev->dev, "GPU pci config reset\n"); + + /* disable dpm? */ + + /* disable cg/pg */ + cik_fini_pg(rdev); + cik_fini_cg(rdev); + + /* Disable GFX parsing/prefetching */ + WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT); + + /* Disable MEC parsing/prefetching */ + WREG32(CP_MEC_CNTL, MEC_ME1_HALT | MEC_ME2_HALT); + + /* sdma0 */ + tmp = RREG32(SDMA0_ME_CNTL + SDMA0_REGISTER_OFFSET); + tmp |= SDMA_HALT; + WREG32(SDMA0_ME_CNTL + SDMA0_REGISTER_OFFSET, tmp); + /* sdma1 */ + tmp = RREG32(SDMA0_ME_CNTL + SDMA1_REGISTER_OFFSET); + tmp |= SDMA_HALT; + WREG32(SDMA0_ME_CNTL + SDMA1_REGISTER_OFFSET, tmp); + /* XXX other engines? */ + + /* halt the rlc, disable cp internal ints */ + cik_rlc_stop(rdev); + + udelay(50); + + /* disable mem access */ + evergreen_mc_stop(rdev, &save); + if (evergreen_mc_wait_for_idle(rdev)) { + dev_warn(rdev->dev, "Wait for MC idle timed out !\n"); + } + + if (rdev->flags & RADEON_IS_IGP) + kv_save_regs_for_reset(rdev, &kv_save); + + /* disable BM */ + pci_clear_master(rdev->pdev); + /* reset */ + radeon_pci_config_reset(rdev); + + udelay(100); + + /* wait for asic to come out of reset */ + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(CONFIG_MEMSIZE) != 0xffffffff) + break; + udelay(1); + } + + /* does asic init need to be run first??? */ + if (rdev->flags & RADEON_IS_IGP) + kv_restore_regs_for_reset(rdev, &kv_save); +} + /** * cik_asic_reset - soft reset GPU * @@ -4013,10 +5195,17 @@ int cik_asic_reset(struct radeon_device *rdev) if (reset_mask) r600_set_bios_scratch_engine_hung(rdev, true); + /* try soft reset */ cik_gpu_soft_reset(rdev, reset_mask); reset_mask = cik_gpu_check_soft_reset(rdev); + /* try pci config reset */ + if (reset_mask && radeon_hard_reset) + cik_gpu_pci_config_reset(rdev); + + reset_mask = cik_gpu_check_soft_reset(rdev); + if (!reset_mask) r600_set_bios_scratch_engine_hung(rdev, false); @@ -4039,39 +5228,9 @@ bool cik_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); - return false; - } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); - return radeon_ring_test_lockup(rdev, ring); -} - -/** - * cik_sdma_is_lockup - Check if the DMA engine is locked up - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Check if the async DMA engine is locked up (CIK). - * Returns true if the engine appears to be locked up, false if not. - */ -bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) -{ - u32 reset_mask = cik_gpu_check_soft_reset(rdev); - u32 mask; - - if (ring->idx == R600_RING_TYPE_DMA_INDEX) - mask = RADEON_RESET_DMA; - else - mask = RADEON_RESET_DMA1; - - if (!(reset_mask & mask)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } @@ -4190,8 +5349,8 @@ static int cik_mc_init(struct radeon_device *rdev) rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0); rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0); /* size in MB on si */ - rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024; - rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024; + rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024ULL * 1024ULL; + rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024ULL * 1024ULL; rdev->mc.visible_vram_size = rdev->mc.aper_size; si_vram_gtt_location(rdev, &rdev->mc); radeon_update_bandwidth_info(rdev); @@ -4248,6 +5407,7 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev) WREG32(MC_VM_MX_L1_TLB_CNTL, (0xA << 7) | ENABLE_L1_TLB | + ENABLE_L1_FRAGMENT_PROCESSING | SYSTEM_ACCESS_MODE_NOT_IN_SYS | ENABLE_ADVANCED_DRIVER_MODEL | SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU); @@ -4260,7 +5420,8 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev) CONTEXT1_IDENTITY_ACCESS_MODE(1)); WREG32(VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS | INVALIDATE_L2_CACHE); WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY | - L2_CACHE_BIGK_FRAGMENT_SIZE(6)); + BANK_SELECT(4) | + L2_CACHE_BIGK_FRAGMENT_SIZE(4)); /* setup context0 */ WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12); WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR, rdev->mc.gtt_end >> 12); @@ -4296,6 +5457,7 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev) (u32)(rdev->dummy_page.addr >> 12)); WREG32(VM_CONTEXT1_CNTL2, 4); WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(1) | + PAGE_TABLE_BLOCK_SIZE(radeon_vm_block_size - 9) | RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT | RANGE_PROTECTION_FAULT_ENABLE_DEFAULT | DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT | @@ -4309,20 +5471,6 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev) WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT | WRITE_PROTECTION_FAULT_ENABLE_DEFAULT); - /* TC cache setup ??? */ - WREG32(TC_CFG_L1_LOAD_POLICY0, 0); - WREG32(TC_CFG_L1_LOAD_POLICY1, 0); - WREG32(TC_CFG_L1_STORE_POLICY, 0); - - WREG32(TC_CFG_L2_LOAD_POLICY0, 0); - WREG32(TC_CFG_L2_LOAD_POLICY1, 0); - WREG32(TC_CFG_L2_STORE_POLICY0, 0); - WREG32(TC_CFG_L2_STORE_POLICY1, 0); - WREG32(TC_CFG_L2_ATOMIC_POLICY, 0); - - WREG32(TC_CFG_L1_VOLATILE, 0); - WREG32(TC_CFG_L2_VOLATILE, 0); - if (rdev->family == CHIP_KAVERI) { u32 tmp = RREG32(CHUB_CONTROL); tmp &= ~BYPASS_VM; @@ -4331,6 +5479,7 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev) /* XXX SH_MEM regs */ /* where to put LDS, scratch, GPUVM in FSA64 space */ + mutex_lock(&rdev->srbm_mutex); for (i = 0; i < 16; i++) { cik_srbm_select(rdev, 0, 0, 0, i); /* CP and shaders */ @@ -4346,6 +5495,7 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev) /* XXX SDMA RLC - todo */ } cik_srbm_select(rdev, 0, 0, 0, 0); + mutex_unlock(&rdev->srbm_mutex); cik_pcie_gart_tlb_flush(rdev); DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n", @@ -4453,6 +5603,35 @@ void cik_vm_fini(struct radeon_device *rdev) } /** + * cik_vm_decode_fault - print human readable fault info + * + * @rdev: radeon_device pointer + * @status: VM_CONTEXT1_PROTECTION_FAULT_STATUS register value + * @addr: VM_CONTEXT1_PROTECTION_FAULT_ADDR register value + * + * Print human readable fault information (CIK). + */ +static void cik_vm_decode_fault(struct radeon_device *rdev, + u32 status, u32 addr, u32 mc_client) +{ + u32 mc_id; + u32 vmid = (status & FAULT_VMID_MASK) >> FAULT_VMID_SHIFT; + u32 protections = (status & PROTECTIONS_MASK) >> PROTECTIONS_SHIFT; + char block[5] = { mc_client >> 24, (mc_client >> 16) & 0xff, + (mc_client >> 8) & 0xff, mc_client & 0xff, 0 }; + + if (rdev->family == CHIP_HAWAII) + mc_id = (status & HAWAII_MEMORY_CLIENT_ID_MASK) >> MEMORY_CLIENT_ID_SHIFT; + else + mc_id = (status & MEMORY_CLIENT_ID_MASK) >> MEMORY_CLIENT_ID_SHIFT; + + printk("VM fault (0x%02x, vmid %d) at page %u, %s from '%s' (0x%08x) (%d)\n", + protections, vmid, addr, + (status & MEMORY_CLIENT_RW_MASK) ? "write" : "read", + block, mc_client, mc_id); +} + +/** * cik_vm_flush - cik vm flush using the CP * * @rdev: radeon_device pointer @@ -4507,16 +5686,7 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) radeon_ring_write(ring, VMID(0)); /* HDP flush */ - /* We should be using the WAIT_REG_MEM packet here like in - * cik_fence_ring_emit(), but it causes the CP to hang in this - * context... - */ - radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | - WRITE_DATA_DST_SEL(0))); - radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); + cik_hdp_flush_cp_ring_emit(rdev, ridx); /* bits 0-15 are the VM contexts0-15 */ radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); @@ -4534,191 +5704,123 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) } } -/** - * cik_vm_set_page - update the page tables using sDMA - * - * @rdev: radeon_device pointer - * @ib: indirect buffer to fill with commands - * @pe: addr of the page entry - * @addr: dst addr to write into pe - * @count: number of page entries to update - * @incr: increase next addr by incr bytes - * @flags: access flags - * - * Update the page tables using CP or sDMA (CIK). +/* + * RLC + * The RLC is a multi-purpose microengine that handles a + * variety of functions, the most important of which is + * the interrupt controller. */ -void cik_vm_set_page(struct radeon_device *rdev, - struct radeon_ib *ib, - uint64_t pe, - uint64_t addr, unsigned count, - uint32_t incr, uint32_t flags) -{ - uint32_t r600_flags = cayman_vm_page_flags(rdev, flags); - uint64_t value; - unsigned ndw; - - if (rdev->asic->vm.pt_ring_index == RADEON_RING_TYPE_GFX_INDEX) { - /* CP */ - while (count) { - ndw = 2 + count * 2; - if (ndw > 0x3FFE) - ndw = 0x3FFE; - - ib->ptr[ib->length_dw++] = PACKET3(PACKET3_WRITE_DATA, ndw); - ib->ptr[ib->length_dw++] = (WRITE_DATA_ENGINE_SEL(0) | - WRITE_DATA_DST_SEL(1)); - ib->ptr[ib->length_dw++] = pe; - ib->ptr[ib->length_dw++] = upper_32_bits(pe); - for (; ndw > 2; ndw -= 2, --count, pe += 8) { - if (flags & RADEON_VM_PAGE_SYSTEM) { - value = radeon_vm_map_gart(rdev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & RADEON_VM_PAGE_VALID) { - value = addr; - } else { - value = 0; - } - addr += incr; - value |= r600_flags; - ib->ptr[ib->length_dw++] = value; - ib->ptr[ib->length_dw++] = upper_32_bits(value); - } - } - } else { - /* DMA */ - if (flags & RADEON_VM_PAGE_SYSTEM) { - while (count) { - ndw = count * 2; - if (ndw > 0xFFFFE) - ndw = 0xFFFFE; - - /* for non-physically contiguous pages (system) */ - ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0); - ib->ptr[ib->length_dw++] = pe; - ib->ptr[ib->length_dw++] = upper_32_bits(pe); - ib->ptr[ib->length_dw++] = ndw; - for (; ndw > 0; ndw -= 2, --count, pe += 8) { - if (flags & RADEON_VM_PAGE_SYSTEM) { - value = radeon_vm_map_gart(rdev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & RADEON_VM_PAGE_VALID) { - value = addr; - } else { - value = 0; - } - addr += incr; - value |= r600_flags; - ib->ptr[ib->length_dw++] = value; - ib->ptr[ib->length_dw++] = upper_32_bits(value); - } - } - } else { - while (count) { - ndw = count; - if (ndw > 0x7FFFF) - ndw = 0x7FFFF; +static void cik_enable_gui_idle_interrupt(struct radeon_device *rdev, + bool enable) +{ + u32 tmp = RREG32(CP_INT_CNTL_RING0); - if (flags & RADEON_VM_PAGE_VALID) - value = addr; - else - value = 0; - /* for physically contiguous pages (vram) */ - ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_GENERATE_PTE_PDE, 0, 0); - ib->ptr[ib->length_dw++] = pe; /* dst addr */ - ib->ptr[ib->length_dw++] = upper_32_bits(pe); - ib->ptr[ib->length_dw++] = r600_flags; /* mask */ - ib->ptr[ib->length_dw++] = 0; - ib->ptr[ib->length_dw++] = value; /* value */ - ib->ptr[ib->length_dw++] = upper_32_bits(value); - ib->ptr[ib->length_dw++] = incr; /* increment size */ - ib->ptr[ib->length_dw++] = 0; - ib->ptr[ib->length_dw++] = ndw; /* number of entries */ - pe += ndw * 8; - addr += ndw * incr; - count -= ndw; + if (enable) + tmp |= (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + else + tmp &= ~(CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + WREG32(CP_INT_CNTL_RING0, tmp); +} + +static void cik_enable_lbpw(struct radeon_device *rdev, bool enable) +{ + u32 tmp; + + tmp = RREG32(RLC_LB_CNTL); + if (enable) + tmp |= LOAD_BALANCE_ENABLE; + else + tmp &= ~LOAD_BALANCE_ENABLE; + WREG32(RLC_LB_CNTL, tmp); +} + +static void cik_wait_for_rlc_serdes(struct radeon_device *rdev) +{ + u32 i, j, k; + u32 mask; + + for (i = 0; i < rdev->config.cik.max_shader_engines; i++) { + for (j = 0; j < rdev->config.cik.max_sh_per_se; j++) { + cik_select_se_sh(rdev, i, j); + for (k = 0; k < rdev->usec_timeout; k++) { + if (RREG32(RLC_SERDES_CU_MASTER_BUSY) == 0) + break; + udelay(1); } } - while (ib->length_dw & 0x7) - ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0); + } + cik_select_se_sh(rdev, 0xffffffff, 0xffffffff); + + mask = SE_MASTER_BUSY_MASK | GC_MASTER_BUSY | TC0_MASTER_BUSY | TC1_MASTER_BUSY; + for (k = 0; k < rdev->usec_timeout; k++) { + if ((RREG32(RLC_SERDES_NONCU_MASTER_BUSY) & mask) == 0) + break; + udelay(1); } } -/** - * cik_dma_vm_flush - cik vm flush using sDMA - * - * @rdev: radeon_device pointer - * - * Update the page table base and flush the VM TLB - * using sDMA (CIK). - */ -void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) +static void cik_update_rlc(struct radeon_device *rdev, u32 rlc) { - struct radeon_ring *ring = &rdev->ring[ridx]; - u32 extra_bits = (SDMA_POLL_REG_MEM_EXTRA_OP(1) | - SDMA_POLL_REG_MEM_EXTRA_FUNC(3)); /* == */ - u32 ref_and_mask; + u32 tmp; - if (vm == NULL) - return; + tmp = RREG32(RLC_CNTL); + if (tmp != rlc) + WREG32(RLC_CNTL, rlc); +} - if (ridx == R600_RING_TYPE_DMA_INDEX) - ref_and_mask = SDMA0; - else - ref_and_mask = SDMA1; +static u32 cik_halt_rlc(struct radeon_device *rdev) +{ + u32 data, orig; - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); - if (vm->id < 8) { - radeon_ring_write(ring, (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2); - } else { - radeon_ring_write(ring, (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2); - } - radeon_ring_write(ring, vm->pd_gpu_addr >> 12); + orig = data = RREG32(RLC_CNTL); - /* update SH_MEM_* regs */ - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); - radeon_ring_write(ring, SRBM_GFX_CNTL >> 2); - radeon_ring_write(ring, VMID(vm->id)); + if (data & RLC_ENABLE) { + u32 i; - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); - radeon_ring_write(ring, SH_MEM_BASES >> 2); - radeon_ring_write(ring, 0); + data &= ~RLC_ENABLE; + WREG32(RLC_CNTL, data); - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); - radeon_ring_write(ring, SH_MEM_CONFIG >> 2); - radeon_ring_write(ring, 0); + for (i = 0; i < rdev->usec_timeout; i++) { + if ((RREG32(RLC_GPM_STAT) & RLC_GPM_BUSY) == 0) + break; + udelay(1); + } - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); - radeon_ring_write(ring, SH_MEM_APE1_BASE >> 2); - radeon_ring_write(ring, 1); + cik_wait_for_rlc_serdes(rdev); + } - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); - radeon_ring_write(ring, SH_MEM_APE1_LIMIT >> 2); - radeon_ring_write(ring, 0); + return orig; +} - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); - radeon_ring_write(ring, SRBM_GFX_CNTL >> 2); - radeon_ring_write(ring, VMID(0)); +void cik_enter_rlc_safe_mode(struct radeon_device *rdev) +{ + u32 tmp, i, mask; - /* flush HDP */ - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_POLL_REG_MEM, 0, extra_bits)); - radeon_ring_write(ring, GPU_HDP_FLUSH_DONE); - radeon_ring_write(ring, GPU_HDP_FLUSH_REQ); - radeon_ring_write(ring, ref_and_mask); /* REFERENCE */ - radeon_ring_write(ring, ref_and_mask); /* MASK */ - radeon_ring_write(ring, (4 << 16) | 10); /* RETRY_COUNT, POLL_INTERVAL */ + tmp = REQ | MESSAGE(MSG_ENTER_RLC_SAFE_MODE); + WREG32(RLC_GPR_REG2, tmp); - /* flush TLB */ - radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); - radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2); - radeon_ring_write(ring, 1 << vm->id); + mask = GFX_POWER_STATUS | GFX_CLOCK_STATUS; + for (i = 0; i < rdev->usec_timeout; i++) { + if ((RREG32(RLC_GPM_STAT) & mask) == mask) + break; + udelay(1); + } + + for (i = 0; i < rdev->usec_timeout; i++) { + if ((RREG32(RLC_GPR_REG2) & REQ) == 0) + break; + udelay(1); + } +} + +void cik_exit_rlc_safe_mode(struct radeon_device *rdev) +{ + u32 tmp; + + tmp = REQ | MESSAGE(MSG_EXIT_RLC_SAFE_MODE); + WREG32(RLC_GPR_REG2, tmp); } -/* - * RLC - * The RLC is a multi-purpose microengine that handles a - * variety of functions, the most important of which is - * the interrupt controller. - */ /** * cik_rlc_stop - stop the RLC ME * @@ -4728,41 +5830,11 @@ void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm */ static void cik_rlc_stop(struct radeon_device *rdev) { - int i, j, k; - u32 mask, tmp; - - tmp = RREG32(CP_INT_CNTL_RING0); - tmp &= ~(CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); - WREG32(CP_INT_CNTL_RING0, tmp); - - RREG32(CB_CGTT_SCLK_CTRL); - RREG32(CB_CGTT_SCLK_CTRL); - RREG32(CB_CGTT_SCLK_CTRL); - RREG32(CB_CGTT_SCLK_CTRL); - - tmp = RREG32(RLC_CGCG_CGLS_CTRL) & 0xfffffffc; - WREG32(RLC_CGCG_CGLS_CTRL, tmp); - WREG32(RLC_CNTL, 0); - for (i = 0; i < rdev->config.cik.max_shader_engines; i++) { - for (j = 0; j < rdev->config.cik.max_sh_per_se; j++) { - cik_select_se_sh(rdev, i, j); - for (k = 0; k < rdev->usec_timeout; k++) { - if (RREG32(RLC_SERDES_CU_MASTER_BUSY) == 0) - break; - udelay(1); - } - } - } - cik_select_se_sh(rdev, 0xffffffff, 0xffffffff); + cik_enable_gui_idle_interrupt(rdev, false); - mask = SE_MASTER_BUSY_MASK | GC_MASTER_BUSY | TC0_MASTER_BUSY | TC1_MASTER_BUSY; - for (k = 0; k < rdev->usec_timeout; k++) { - if ((RREG32(RLC_SERDES_NONCU_MASTER_BUSY) & mask) == 0) - break; - udelay(1); - } + cik_wait_for_rlc_serdes(rdev); } /** @@ -4774,13 +5846,9 @@ static void cik_rlc_stop(struct radeon_device *rdev) */ static void cik_rlc_start(struct radeon_device *rdev) { - u32 tmp; - WREG32(RLC_CNTL, RLC_ENABLE); - tmp = RREG32(CP_INT_CNTL_RING0); - tmp |= (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); - WREG32(CP_INT_CNTL_RING0, tmp); + cik_enable_gui_idle_interrupt(rdev, true); udelay(50); } @@ -4796,8 +5864,7 @@ static void cik_rlc_start(struct radeon_device *rdev) */ static int cik_rlc_resume(struct radeon_device *rdev) { - u32 i, size; - u32 clear_state_info[3]; + u32 i, size, tmp; const __be32 *fw_data; if (!rdev->rlc_fw) @@ -4805,6 +5872,7 @@ static int cik_rlc_resume(struct radeon_device *rdev) switch (rdev->family) { case CHIP_BONAIRE: + case CHIP_HAWAII: default: size = BONAIRE_RLC_UCODE_SIZE; break; @@ -4814,16 +5882,22 @@ static int cik_rlc_resume(struct radeon_device *rdev) case CHIP_KABINI: size = KB_RLC_UCODE_SIZE; break; + case CHIP_MULLINS: + size = ML_RLC_UCODE_SIZE; + break; } cik_rlc_stop(rdev); - WREG32(GRBM_SOFT_RESET, SOFT_RESET_RLC); - RREG32(GRBM_SOFT_RESET); - udelay(50); - WREG32(GRBM_SOFT_RESET, 0); - RREG32(GRBM_SOFT_RESET); - udelay(50); + /* disable CG */ + tmp = RREG32(RLC_CGCG_CGLS_CTRL) & 0xfffffffc; + WREG32(RLC_CGCG_CGLS_CTRL, tmp); + + si_rlc_reset(rdev); + + cik_init_pg(rdev); + + cik_init_cg(rdev); WREG32(RLC_LB_CNTR_INIT, 0); WREG32(RLC_LB_CNTR_MAX, 0x00008000); @@ -4842,20 +5916,770 @@ static int cik_rlc_resume(struct radeon_device *rdev) WREG32(RLC_GPM_UCODE_DATA, be32_to_cpup(fw_data++)); WREG32(RLC_GPM_UCODE_ADDR, 0); - /* XXX */ - clear_state_info[0] = 0;//upper_32_bits(rdev->rlc.save_restore_gpu_addr); - clear_state_info[1] = 0;//rdev->rlc.save_restore_gpu_addr; - clear_state_info[2] = 0;//cik_default_size; - WREG32(RLC_GPM_SCRATCH_ADDR, 0x3d); - for (i = 0; i < 3; i++) - WREG32(RLC_GPM_SCRATCH_DATA, clear_state_info[i]); - WREG32(RLC_DRIVER_DMA_STATUS, 0); + /* XXX - find out what chips support lbpw */ + cik_enable_lbpw(rdev, false); + + if (rdev->family == CHIP_BONAIRE) + WREG32(RLC_DRIVER_DMA_STATUS, 0); cik_rlc_start(rdev); return 0; } +static void cik_enable_cgcg(struct radeon_device *rdev, bool enable) +{ + u32 data, orig, tmp, tmp2; + + orig = data = RREG32(RLC_CGCG_CGLS_CTRL); + + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_CGCG)) { + cik_enable_gui_idle_interrupt(rdev, true); + + tmp = cik_halt_rlc(rdev); + + cik_select_se_sh(rdev, 0xffffffff, 0xffffffff); + WREG32(RLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); + WREG32(RLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); + tmp2 = BPM_ADDR_MASK | CGCG_OVERRIDE_0 | CGLS_ENABLE; + WREG32(RLC_SERDES_WR_CTRL, tmp2); + + cik_update_rlc(rdev, tmp); + + data |= CGCG_EN | CGLS_EN; + } else { + cik_enable_gui_idle_interrupt(rdev, false); + + RREG32(CB_CGTT_SCLK_CTRL); + RREG32(CB_CGTT_SCLK_CTRL); + RREG32(CB_CGTT_SCLK_CTRL); + RREG32(CB_CGTT_SCLK_CTRL); + + data &= ~(CGCG_EN | CGLS_EN); + } + + if (orig != data) + WREG32(RLC_CGCG_CGLS_CTRL, data); + +} + +static void cik_enable_mgcg(struct radeon_device *rdev, bool enable) +{ + u32 data, orig, tmp = 0; + + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_MGCG)) { + if (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_MGLS) { + if (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_CP_LS) { + orig = data = RREG32(CP_MEM_SLP_CNTL); + data |= CP_MEM_LS_EN; + if (orig != data) + WREG32(CP_MEM_SLP_CNTL, data); + } + } + + orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE); + data &= 0xfffffffd; + if (orig != data) + WREG32(RLC_CGTT_MGCG_OVERRIDE, data); + + tmp = cik_halt_rlc(rdev); + + cik_select_se_sh(rdev, 0xffffffff, 0xffffffff); + WREG32(RLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); + WREG32(RLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); + data = BPM_ADDR_MASK | MGCG_OVERRIDE_0; + WREG32(RLC_SERDES_WR_CTRL, data); + + cik_update_rlc(rdev, tmp); + + if (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_CGTS) { + orig = data = RREG32(CGTS_SM_CTRL_REG); + data &= ~SM_MODE_MASK; + data |= SM_MODE(0x2); + data |= SM_MODE_ENABLE; + data &= ~CGTS_OVERRIDE; + if ((rdev->cg_flags & RADEON_CG_SUPPORT_GFX_MGLS) && + (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_CGTS_LS)) + data &= ~CGTS_LS_OVERRIDE; + data &= ~ON_MONITOR_ADD_MASK; + data |= ON_MONITOR_ADD_EN; + data |= ON_MONITOR_ADD(0x96); + if (orig != data) + WREG32(CGTS_SM_CTRL_REG, data); + } + } else { + orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE); + data |= 0x00000002; + if (orig != data) + WREG32(RLC_CGTT_MGCG_OVERRIDE, data); + + data = RREG32(RLC_MEM_SLP_CNTL); + if (data & RLC_MEM_LS_EN) { + data &= ~RLC_MEM_LS_EN; + WREG32(RLC_MEM_SLP_CNTL, data); + } + + data = RREG32(CP_MEM_SLP_CNTL); + if (data & CP_MEM_LS_EN) { + data &= ~CP_MEM_LS_EN; + WREG32(CP_MEM_SLP_CNTL, data); + } + + orig = data = RREG32(CGTS_SM_CTRL_REG); + data |= CGTS_OVERRIDE | CGTS_LS_OVERRIDE; + if (orig != data) + WREG32(CGTS_SM_CTRL_REG, data); + + tmp = cik_halt_rlc(rdev); + + cik_select_se_sh(rdev, 0xffffffff, 0xffffffff); + WREG32(RLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); + WREG32(RLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); + data = BPM_ADDR_MASK | MGCG_OVERRIDE_1; + WREG32(RLC_SERDES_WR_CTRL, data); + + cik_update_rlc(rdev, tmp); + } +} + +static const u32 mc_cg_registers[] = +{ + MC_HUB_MISC_HUB_CG, + MC_HUB_MISC_SIP_CG, + MC_HUB_MISC_VM_CG, + MC_XPB_CLK_GAT, + ATC_MISC_CG, + MC_CITF_MISC_WR_CG, + MC_CITF_MISC_RD_CG, + MC_CITF_MISC_VM_CG, + VM_L2_CG, +}; + +static void cik_enable_mc_ls(struct radeon_device *rdev, + bool enable) +{ + int i; + u32 orig, data; + + for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) { + orig = data = RREG32(mc_cg_registers[i]); + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_MC_LS)) + data |= MC_LS_ENABLE; + else + data &= ~MC_LS_ENABLE; + if (data != orig) + WREG32(mc_cg_registers[i], data); + } +} + +static void cik_enable_mc_mgcg(struct radeon_device *rdev, + bool enable) +{ + int i; + u32 orig, data; + + for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) { + orig = data = RREG32(mc_cg_registers[i]); + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_MC_MGCG)) + data |= MC_CG_ENABLE; + else + data &= ~MC_CG_ENABLE; + if (data != orig) + WREG32(mc_cg_registers[i], data); + } +} + +static void cik_enable_sdma_mgcg(struct radeon_device *rdev, + bool enable) +{ + u32 orig, data; + + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_SDMA_MGCG)) { + WREG32(SDMA0_CLK_CTRL + SDMA0_REGISTER_OFFSET, 0x00000100); + WREG32(SDMA0_CLK_CTRL + SDMA1_REGISTER_OFFSET, 0x00000100); + } else { + orig = data = RREG32(SDMA0_CLK_CTRL + SDMA0_REGISTER_OFFSET); + data |= 0xff000000; + if (data != orig) + WREG32(SDMA0_CLK_CTRL + SDMA0_REGISTER_OFFSET, data); + + orig = data = RREG32(SDMA0_CLK_CTRL + SDMA1_REGISTER_OFFSET); + data |= 0xff000000; + if (data != orig) + WREG32(SDMA0_CLK_CTRL + SDMA1_REGISTER_OFFSET, data); + } +} + +static void cik_enable_sdma_mgls(struct radeon_device *rdev, + bool enable) +{ + u32 orig, data; + + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_SDMA_LS)) { + orig = data = RREG32(SDMA0_POWER_CNTL + SDMA0_REGISTER_OFFSET); + data |= 0x100; + if (orig != data) + WREG32(SDMA0_POWER_CNTL + SDMA0_REGISTER_OFFSET, data); + + orig = data = RREG32(SDMA0_POWER_CNTL + SDMA1_REGISTER_OFFSET); + data |= 0x100; + if (orig != data) + WREG32(SDMA0_POWER_CNTL + SDMA1_REGISTER_OFFSET, data); + } else { + orig = data = RREG32(SDMA0_POWER_CNTL + SDMA0_REGISTER_OFFSET); + data &= ~0x100; + if (orig != data) + WREG32(SDMA0_POWER_CNTL + SDMA0_REGISTER_OFFSET, data); + + orig = data = RREG32(SDMA0_POWER_CNTL + SDMA1_REGISTER_OFFSET); + data &= ~0x100; + if (orig != data) + WREG32(SDMA0_POWER_CNTL + SDMA1_REGISTER_OFFSET, data); + } +} + +static void cik_enable_uvd_mgcg(struct radeon_device *rdev, + bool enable) +{ + u32 orig, data; + + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_UVD_MGCG)) { + data = RREG32_UVD_CTX(UVD_CGC_MEM_CTRL); + data = 0xfff; + WREG32_UVD_CTX(UVD_CGC_MEM_CTRL, data); + + orig = data = RREG32(UVD_CGC_CTRL); + data |= DCM; + if (orig != data) + WREG32(UVD_CGC_CTRL, data); + } else { + data = RREG32_UVD_CTX(UVD_CGC_MEM_CTRL); + data &= ~0xfff; + WREG32_UVD_CTX(UVD_CGC_MEM_CTRL, data); + + orig = data = RREG32(UVD_CGC_CTRL); + data &= ~DCM; + if (orig != data) + WREG32(UVD_CGC_CTRL, data); + } +} + +static void cik_enable_bif_mgls(struct radeon_device *rdev, + bool enable) +{ + u32 orig, data; + + orig = data = RREG32_PCIE_PORT(PCIE_CNTL2); + + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_BIF_LS)) + data |= SLV_MEM_LS_EN | MST_MEM_LS_EN | + REPLAY_MEM_LS_EN | SLV_MEM_AGGRESSIVE_LS_EN; + else + data &= ~(SLV_MEM_LS_EN | MST_MEM_LS_EN | + REPLAY_MEM_LS_EN | SLV_MEM_AGGRESSIVE_LS_EN); + + if (orig != data) + WREG32_PCIE_PORT(PCIE_CNTL2, data); +} + +static void cik_enable_hdp_mgcg(struct radeon_device *rdev, + bool enable) +{ + u32 orig, data; + + orig = data = RREG32(HDP_HOST_PATH_CNTL); + + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_HDP_MGCG)) + data &= ~CLOCK_GATING_DIS; + else + data |= CLOCK_GATING_DIS; + + if (orig != data) + WREG32(HDP_HOST_PATH_CNTL, data); +} + +static void cik_enable_hdp_ls(struct radeon_device *rdev, + bool enable) +{ + u32 orig, data; + + orig = data = RREG32(HDP_MEM_POWER_LS); + + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_HDP_LS)) + data |= HDP_LS_ENABLE; + else + data &= ~HDP_LS_ENABLE; + + if (orig != data) + WREG32(HDP_MEM_POWER_LS, data); +} + +void cik_update_cg(struct radeon_device *rdev, + u32 block, bool enable) +{ + + if (block & RADEON_CG_BLOCK_GFX) { + cik_enable_gui_idle_interrupt(rdev, false); + /* order matters! */ + if (enable) { + cik_enable_mgcg(rdev, true); + cik_enable_cgcg(rdev, true); + } else { + cik_enable_cgcg(rdev, false); + cik_enable_mgcg(rdev, false); + } + cik_enable_gui_idle_interrupt(rdev, true); + } + + if (block & RADEON_CG_BLOCK_MC) { + if (!(rdev->flags & RADEON_IS_IGP)) { + cik_enable_mc_mgcg(rdev, enable); + cik_enable_mc_ls(rdev, enable); + } + } + + if (block & RADEON_CG_BLOCK_SDMA) { + cik_enable_sdma_mgcg(rdev, enable); + cik_enable_sdma_mgls(rdev, enable); + } + + if (block & RADEON_CG_BLOCK_BIF) { + cik_enable_bif_mgls(rdev, enable); + } + + if (block & RADEON_CG_BLOCK_UVD) { + if (rdev->has_uvd) + cik_enable_uvd_mgcg(rdev, enable); + } + + if (block & RADEON_CG_BLOCK_HDP) { + cik_enable_hdp_mgcg(rdev, enable); + cik_enable_hdp_ls(rdev, enable); + } + + if (block & RADEON_CG_BLOCK_VCE) { + vce_v2_0_enable_mgcg(rdev, enable); + } +} + +static void cik_init_cg(struct radeon_device *rdev) +{ + + cik_update_cg(rdev, RADEON_CG_BLOCK_GFX, true); + + if (rdev->has_uvd) + si_init_uvd_internal_cg(rdev); + + cik_update_cg(rdev, (RADEON_CG_BLOCK_MC | + RADEON_CG_BLOCK_SDMA | + RADEON_CG_BLOCK_BIF | + RADEON_CG_BLOCK_UVD | + RADEON_CG_BLOCK_HDP), true); +} + +static void cik_fini_cg(struct radeon_device *rdev) +{ + cik_update_cg(rdev, (RADEON_CG_BLOCK_MC | + RADEON_CG_BLOCK_SDMA | + RADEON_CG_BLOCK_BIF | + RADEON_CG_BLOCK_UVD | + RADEON_CG_BLOCK_HDP), false); + + cik_update_cg(rdev, RADEON_CG_BLOCK_GFX, false); +} + +static void cik_enable_sck_slowdown_on_pu(struct radeon_device *rdev, + bool enable) +{ + u32 data, orig; + + orig = data = RREG32(RLC_PG_CNTL); + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_RLC_SMU_HS)) + data |= SMU_CLK_SLOWDOWN_ON_PU_ENABLE; + else + data &= ~SMU_CLK_SLOWDOWN_ON_PU_ENABLE; + if (orig != data) + WREG32(RLC_PG_CNTL, data); +} + +static void cik_enable_sck_slowdown_on_pd(struct radeon_device *rdev, + bool enable) +{ + u32 data, orig; + + orig = data = RREG32(RLC_PG_CNTL); + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_RLC_SMU_HS)) + data |= SMU_CLK_SLOWDOWN_ON_PD_ENABLE; + else + data &= ~SMU_CLK_SLOWDOWN_ON_PD_ENABLE; + if (orig != data) + WREG32(RLC_PG_CNTL, data); +} + +static void cik_enable_cp_pg(struct radeon_device *rdev, bool enable) +{ + u32 data, orig; + + orig = data = RREG32(RLC_PG_CNTL); + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_CP)) + data &= ~DISABLE_CP_PG; + else + data |= DISABLE_CP_PG; + if (orig != data) + WREG32(RLC_PG_CNTL, data); +} + +static void cik_enable_gds_pg(struct radeon_device *rdev, bool enable) +{ + u32 data, orig; + + orig = data = RREG32(RLC_PG_CNTL); + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_GDS)) + data &= ~DISABLE_GDS_PG; + else + data |= DISABLE_GDS_PG; + if (orig != data) + WREG32(RLC_PG_CNTL, data); +} + +#define CP_ME_TABLE_SIZE 96 +#define CP_ME_TABLE_OFFSET 2048 +#define CP_MEC_TABLE_OFFSET 4096 + +void cik_init_cp_pg_table(struct radeon_device *rdev) +{ + const __be32 *fw_data; + volatile u32 *dst_ptr; + int me, i, max_me = 4; + u32 bo_offset = 0; + u32 table_offset; + + if (rdev->family == CHIP_KAVERI) + max_me = 5; + + if (rdev->rlc.cp_table_ptr == NULL) + return; + + /* write the cp table buffer */ + dst_ptr = rdev->rlc.cp_table_ptr; + for (me = 0; me < max_me; me++) { + if (me == 0) { + fw_data = (const __be32 *)rdev->ce_fw->data; + table_offset = CP_ME_TABLE_OFFSET; + } else if (me == 1) { + fw_data = (const __be32 *)rdev->pfp_fw->data; + table_offset = CP_ME_TABLE_OFFSET; + } else if (me == 2) { + fw_data = (const __be32 *)rdev->me_fw->data; + table_offset = CP_ME_TABLE_OFFSET; + } else { + fw_data = (const __be32 *)rdev->mec_fw->data; + table_offset = CP_MEC_TABLE_OFFSET; + } + + for (i = 0; i < CP_ME_TABLE_SIZE; i ++) { + dst_ptr[bo_offset + i] = cpu_to_le32(be32_to_cpu(fw_data[table_offset + i])); + } + bo_offset += CP_ME_TABLE_SIZE; + } +} + +static void cik_enable_gfx_cgpg(struct radeon_device *rdev, + bool enable) +{ + u32 data, orig; + + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_PG)) { + orig = data = RREG32(RLC_PG_CNTL); + data |= GFX_PG_ENABLE; + if (orig != data) + WREG32(RLC_PG_CNTL, data); + + orig = data = RREG32(RLC_AUTO_PG_CTRL); + data |= AUTO_PG_EN; + if (orig != data) + WREG32(RLC_AUTO_PG_CTRL, data); + } else { + orig = data = RREG32(RLC_PG_CNTL); + data &= ~GFX_PG_ENABLE; + if (orig != data) + WREG32(RLC_PG_CNTL, data); + + orig = data = RREG32(RLC_AUTO_PG_CTRL); + data &= ~AUTO_PG_EN; + if (orig != data) + WREG32(RLC_AUTO_PG_CTRL, data); + + data = RREG32(DB_RENDER_CONTROL); + } +} + +static u32 cik_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh) +{ + u32 mask = 0, tmp, tmp1; + int i; + + cik_select_se_sh(rdev, se, sh); + tmp = RREG32(CC_GC_SHADER_ARRAY_CONFIG); + tmp1 = RREG32(GC_USER_SHADER_ARRAY_CONFIG); + cik_select_se_sh(rdev, 0xffffffff, 0xffffffff); + + tmp &= 0xffff0000; + + tmp |= tmp1; + tmp >>= 16; + + for (i = 0; i < rdev->config.cik.max_cu_per_sh; i ++) { + mask <<= 1; + mask |= 1; + } + + return (~tmp) & mask; +} + +static void cik_init_ao_cu_mask(struct radeon_device *rdev) +{ + u32 i, j, k, active_cu_number = 0; + u32 mask, counter, cu_bitmap; + u32 tmp = 0; + + for (i = 0; i < rdev->config.cik.max_shader_engines; i++) { + for (j = 0; j < rdev->config.cik.max_sh_per_se; j++) { + mask = 1; + cu_bitmap = 0; + counter = 0; + for (k = 0; k < rdev->config.cik.max_cu_per_sh; k ++) { + if (cik_get_cu_active_bitmap(rdev, i, j) & mask) { + if (counter < 2) + cu_bitmap |= mask; + counter ++; + } + mask <<= 1; + } + + active_cu_number += counter; + tmp |= (cu_bitmap << (i * 16 + j * 8)); + } + } + + WREG32(RLC_PG_AO_CU_MASK, tmp); + + tmp = RREG32(RLC_MAX_PG_CU); + tmp &= ~MAX_PU_CU_MASK; + tmp |= MAX_PU_CU(active_cu_number); + WREG32(RLC_MAX_PG_CU, tmp); +} + +static void cik_enable_gfx_static_mgpg(struct radeon_device *rdev, + bool enable) +{ + u32 data, orig; + + orig = data = RREG32(RLC_PG_CNTL); + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_SMG)) + data |= STATIC_PER_CU_PG_ENABLE; + else + data &= ~STATIC_PER_CU_PG_ENABLE; + if (orig != data) + WREG32(RLC_PG_CNTL, data); +} + +static void cik_enable_gfx_dynamic_mgpg(struct radeon_device *rdev, + bool enable) +{ + u32 data, orig; + + orig = data = RREG32(RLC_PG_CNTL); + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_DMG)) + data |= DYN_PER_CU_PG_ENABLE; + else + data &= ~DYN_PER_CU_PG_ENABLE; + if (orig != data) + WREG32(RLC_PG_CNTL, data); +} + +#define RLC_SAVE_AND_RESTORE_STARTING_OFFSET 0x90 +#define RLC_CLEAR_STATE_DESCRIPTOR_OFFSET 0x3D + +static void cik_init_gfx_cgpg(struct radeon_device *rdev) +{ + u32 data, orig; + u32 i; + + if (rdev->rlc.cs_data) { + WREG32(RLC_GPM_SCRATCH_ADDR, RLC_CLEAR_STATE_DESCRIPTOR_OFFSET); + WREG32(RLC_GPM_SCRATCH_DATA, upper_32_bits(rdev->rlc.clear_state_gpu_addr)); + WREG32(RLC_GPM_SCRATCH_DATA, lower_32_bits(rdev->rlc.clear_state_gpu_addr)); + WREG32(RLC_GPM_SCRATCH_DATA, rdev->rlc.clear_state_size); + } else { + WREG32(RLC_GPM_SCRATCH_ADDR, RLC_CLEAR_STATE_DESCRIPTOR_OFFSET); + for (i = 0; i < 3; i++) + WREG32(RLC_GPM_SCRATCH_DATA, 0); + } + if (rdev->rlc.reg_list) { + WREG32(RLC_GPM_SCRATCH_ADDR, RLC_SAVE_AND_RESTORE_STARTING_OFFSET); + for (i = 0; i < rdev->rlc.reg_list_size; i++) + WREG32(RLC_GPM_SCRATCH_DATA, rdev->rlc.reg_list[i]); + } + + orig = data = RREG32(RLC_PG_CNTL); + data |= GFX_PG_SRC; + if (orig != data) + WREG32(RLC_PG_CNTL, data); + + WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8); + WREG32(RLC_CP_TABLE_RESTORE, rdev->rlc.cp_table_gpu_addr >> 8); + + data = RREG32(CP_RB_WPTR_POLL_CNTL); + data &= ~IDLE_POLL_COUNT_MASK; + data |= IDLE_POLL_COUNT(0x60); + WREG32(CP_RB_WPTR_POLL_CNTL, data); + + data = 0x10101010; + WREG32(RLC_PG_DELAY, data); + + data = RREG32(RLC_PG_DELAY_2); + data &= ~0xff; + data |= 0x3; + WREG32(RLC_PG_DELAY_2, data); + + data = RREG32(RLC_AUTO_PG_CTRL); + data &= ~GRBM_REG_SGIT_MASK; + data |= GRBM_REG_SGIT(0x700); + WREG32(RLC_AUTO_PG_CTRL, data); + +} + +static void cik_update_gfx_pg(struct radeon_device *rdev, bool enable) +{ + cik_enable_gfx_cgpg(rdev, enable); + cik_enable_gfx_static_mgpg(rdev, enable); + cik_enable_gfx_dynamic_mgpg(rdev, enable); +} + +u32 cik_get_csb_size(struct radeon_device *rdev) +{ + u32 count = 0; + const struct cs_section_def *sect = NULL; + const struct cs_extent_def *ext = NULL; + + if (rdev->rlc.cs_data == NULL) + return 0; + + /* begin clear state */ + count += 2; + /* context control state */ + count += 3; + + for (sect = rdev->rlc.cs_data; sect->section != NULL; ++sect) { + for (ext = sect->section; ext->extent != NULL; ++ext) { + if (sect->id == SECT_CONTEXT) + count += 2 + ext->reg_count; + else + return 0; + } + } + /* pa_sc_raster_config/pa_sc_raster_config1 */ + count += 4; + /* end clear state */ + count += 2; + /* clear state */ + count += 2; + + return count; +} + +void cik_get_csb_buffer(struct radeon_device *rdev, volatile u32 *buffer) +{ + u32 count = 0, i; + const struct cs_section_def *sect = NULL; + const struct cs_extent_def *ext = NULL; + + if (rdev->rlc.cs_data == NULL) + return; + if (buffer == NULL) + return; + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1)); + buffer[count++] = cpu_to_le32(0x80000000); + buffer[count++] = cpu_to_le32(0x80000000); + + for (sect = rdev->rlc.cs_data; sect->section != NULL; ++sect) { + for (ext = sect->section; ext->extent != NULL; ++ext) { + if (sect->id == SECT_CONTEXT) { + buffer[count++] = + cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count)); + buffer[count++] = cpu_to_le32(ext->reg_index - 0xa000); + for (i = 0; i < ext->reg_count; i++) + buffer[count++] = cpu_to_le32(ext->extent[i]); + } else { + return; + } + } + } + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 2)); + buffer[count++] = cpu_to_le32(PA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START); + switch (rdev->family) { + case CHIP_BONAIRE: + buffer[count++] = cpu_to_le32(0x16000012); + buffer[count++] = cpu_to_le32(0x00000000); + break; + case CHIP_KAVERI: + buffer[count++] = cpu_to_le32(0x00000000); /* XXX */ + buffer[count++] = cpu_to_le32(0x00000000); + break; + case CHIP_KABINI: + case CHIP_MULLINS: + buffer[count++] = cpu_to_le32(0x00000000); /* XXX */ + buffer[count++] = cpu_to_le32(0x00000000); + break; + case CHIP_HAWAII: + buffer[count++] = cpu_to_le32(0x3a00161a); + buffer[count++] = cpu_to_le32(0x0000002e); + break; + default: + buffer[count++] = cpu_to_le32(0x00000000); + buffer[count++] = cpu_to_le32(0x00000000); + break; + } + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0)); + buffer[count++] = cpu_to_le32(0); +} + +static void cik_init_pg(struct radeon_device *rdev) +{ + if (rdev->pg_flags) { + cik_enable_sck_slowdown_on_pu(rdev, true); + cik_enable_sck_slowdown_on_pd(rdev, true); + if (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_PG) { + cik_init_gfx_cgpg(rdev); + cik_enable_cp_pg(rdev, true); + cik_enable_gds_pg(rdev, true); + } + cik_init_ao_cu_mask(rdev); + cik_update_gfx_pg(rdev, true); + } +} + +static void cik_fini_pg(struct radeon_device *rdev) +{ + if (rdev->pg_flags) { + cik_update_gfx_pg(rdev, false); + if (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_PG) { + cik_enable_cp_pg(rdev, false); + cik_enable_gds_pg(rdev, false); + } + } +} + /* * Interrupts * Starting with r6xx, interrupts are handled via a ring buffer. @@ -4925,7 +6749,9 @@ static void cik_disable_interrupt_state(struct radeon_device *rdev) u32 tmp; /* gfx ring */ - WREG32(CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + tmp = RREG32(CP_INT_CNTL_RING0) & + (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + WREG32(CP_INT_CNTL_RING0, tmp); /* sdma */ tmp = RREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET) & ~TRAP_ENABLE; WREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET, tmp); @@ -4953,6 +6779,19 @@ static void cik_disable_interrupt_state(struct radeon_device *rdev) WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0); WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); } + /* pflip */ + if (rdev->num_crtc >= 2) { + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, 0); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, 0); + } + if (rdev->num_crtc >= 4) { + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, 0); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, 0); + } + if (rdev->num_crtc >= 6) { + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, 0); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); + } /* dac hotplug */ WREG32(DAC_AUTODETECT_INT_CONTROL, 0); @@ -5018,7 +6857,7 @@ static int cik_irq_init(struct radeon_device *rdev) WREG32(INTERRUPT_CNTL, interrupt_cntl); WREG32(IH_RB_BASE, rdev->ih.gpu_addr >> 8); - rb_bufsz = drm_order(rdev->ih.ring_size / 4); + rb_bufsz = order_base_2(rdev->ih.ring_size / 4); ih_rb_cntl = (IH_WPTR_OVERFLOW_ENABLE | IH_WPTR_OVERFLOW_CLEAR | @@ -5066,14 +6905,14 @@ static int cik_irq_init(struct radeon_device *rdev) */ int cik_irq_set(struct radeon_device *rdev) { - u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE | - PRIV_INSTR_INT_ENABLE | PRIV_REG_INT_ENABLE; + u32 cp_int_cntl; u32 cp_m1p0, cp_m1p1, cp_m1p2, cp_m1p3; u32 cp_m2p0, cp_m2p1, cp_m2p2, cp_m2p3; u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0; u32 hpd1, hpd2, hpd3, hpd4, hpd5, hpd6; u32 grbm_int_cntl = 0; u32 dma_cntl, dma_cntl1; + u32 thermal_int; if (!rdev->irq.installed) { WARN(1, "Can't enable IRQ/MSI because no handler is installed\n"); @@ -5087,6 +6926,10 @@ int cik_irq_set(struct radeon_device *rdev) return 0; } + cp_int_cntl = RREG32(CP_INT_CNTL_RING0) & + (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + cp_int_cntl |= PRIV_INSTR_INT_ENABLE | PRIV_REG_INT_ENABLE; + hpd1 = RREG32(DC_HPD1_INT_CONTROL) & ~DC_HPDx_INT_EN; hpd2 = RREG32(DC_HPD2_INT_CONTROL) & ~DC_HPDx_INT_EN; hpd3 = RREG32(DC_HPD3_INT_CONTROL) & ~DC_HPDx_INT_EN; @@ -5106,6 +6949,13 @@ int cik_irq_set(struct radeon_device *rdev) cp_m2p2 = RREG32(CP_ME2_PIPE2_INT_CNTL) & ~TIME_STAMP_INT_ENABLE; cp_m2p3 = RREG32(CP_ME2_PIPE3_INT_CNTL) & ~TIME_STAMP_INT_ENABLE; + if (rdev->flags & RADEON_IS_IGP) + thermal_int = RREG32_SMC(CG_THERMAL_INT_CTRL) & + ~(THERM_INTH_MASK | THERM_INTL_MASK); + else + thermal_int = RREG32_SMC(CG_THERMAL_INT) & + ~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW); + /* enable CP interrupts on all rings */ if (atomic_read(&rdev->irq.ring_int[RADEON_RING_TYPE_GFX_INDEX])) { DRM_DEBUG("cik_irq_set: sw int gfx\n"); @@ -5263,6 +7113,14 @@ int cik_irq_set(struct radeon_device *rdev) hpd6 |= DC_HPDx_INT_EN; } + if (rdev->irq.dpm_thermal) { + DRM_DEBUG("dpm thermal\n"); + if (rdev->flags & RADEON_IS_IGP) + thermal_int |= THERM_INTH_MASK | THERM_INTL_MASK; + else + thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW; + } + WREG32(CP_INT_CNTL_RING0, cp_int_cntl); WREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET, dma_cntl); @@ -5290,6 +7148,25 @@ int cik_irq_set(struct radeon_device *rdev) WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, crtc6); } + if (rdev->num_crtc >= 2) { + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + } + if (rdev->num_crtc >= 4) { + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + } + if (rdev->num_crtc >= 6) { + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + } + WREG32(DC_HPD1_INT_CONTROL, hpd1); WREG32(DC_HPD2_INT_CONTROL, hpd2); WREG32(DC_HPD3_INT_CONTROL, hpd3); @@ -5297,6 +7174,11 @@ int cik_irq_set(struct radeon_device *rdev) WREG32(DC_HPD5_INT_CONTROL, hpd5); WREG32(DC_HPD6_INT_CONTROL, hpd6); + if (rdev->flags & RADEON_IS_IGP) + WREG32_SMC(CG_THERMAL_INT_CTRL, thermal_int); + else + WREG32_SMC(CG_THERMAL_INT, thermal_int); + return 0; } @@ -5321,6 +7203,29 @@ static inline void cik_irq_ack(struct radeon_device *rdev) rdev->irq.stat_regs.cik.disp_int_cont5 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE5); rdev->irq.stat_regs.cik.disp_int_cont6 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE6); + rdev->irq.stat_regs.cik.d1grph_int = RREG32(GRPH_INT_STATUS + + EVERGREEN_CRTC0_REGISTER_OFFSET); + rdev->irq.stat_regs.cik.d2grph_int = RREG32(GRPH_INT_STATUS + + EVERGREEN_CRTC1_REGISTER_OFFSET); + if (rdev->num_crtc >= 4) { + rdev->irq.stat_regs.cik.d3grph_int = RREG32(GRPH_INT_STATUS + + EVERGREEN_CRTC2_REGISTER_OFFSET); + rdev->irq.stat_regs.cik.d4grph_int = RREG32(GRPH_INT_STATUS + + EVERGREEN_CRTC3_REGISTER_OFFSET); + } + if (rdev->num_crtc >= 6) { + rdev->irq.stat_regs.cik.d5grph_int = RREG32(GRPH_INT_STATUS + + EVERGREEN_CRTC4_REGISTER_OFFSET); + rdev->irq.stat_regs.cik.d6grph_int = RREG32(GRPH_INT_STATUS + + EVERGREEN_CRTC5_REGISTER_OFFSET); + } + + if (rdev->irq.stat_regs.cik.d1grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, + GRPH_PFLIP_INT_CLEAR); + if (rdev->irq.stat_regs.cik.d2grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, + GRPH_PFLIP_INT_CLEAR); if (rdev->irq.stat_regs.cik.disp_int & LB_D1_VBLANK_INTERRUPT) WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VBLANK_ACK); if (rdev->irq.stat_regs.cik.disp_int & LB_D1_VLINE_INTERRUPT) @@ -5331,6 +7236,12 @@ static inline void cik_irq_ack(struct radeon_device *rdev) WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VLINE_ACK); if (rdev->num_crtc >= 4) { + if (rdev->irq.stat_regs.cik.d3grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, + GRPH_PFLIP_INT_CLEAR); + if (rdev->irq.stat_regs.cik.d4grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, + GRPH_PFLIP_INT_CLEAR); if (rdev->irq.stat_regs.cik.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT) WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VBLANK_ACK); if (rdev->irq.stat_regs.cik.disp_int_cont2 & LB_D3_VLINE_INTERRUPT) @@ -5342,6 +7253,12 @@ static inline void cik_irq_ack(struct radeon_device *rdev) } if (rdev->num_crtc >= 6) { + if (rdev->irq.stat_regs.cik.d5grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, + GRPH_PFLIP_INT_CLEAR); + if (rdev->irq.stat_regs.cik.d6grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, + GRPH_PFLIP_INT_CLEAR); if (rdev->irq.stat_regs.cik.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT) WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VBLANK_ACK); if (rdev->irq.stat_regs.cik.disp_int_cont4 & LB_D5_VLINE_INTERRUPT) @@ -5460,6 +7377,7 @@ static inline u32 cik_get_ih_wptr(struct radeon_device *rdev) tmp = RREG32(IH_RB_CNTL); tmp |= IH_WPTR_OVERFLOW_CLEAR; WREG32(IH_RB_CNTL, tmp); + wptr &= ~RB_OVERFLOW; } return (wptr & rdev->ih.ptr_mask); } @@ -5507,6 +7425,8 @@ int cik_irq_process(struct radeon_device *rdev) u32 ring_index; bool queue_hotplug = false; bool queue_reset = false; + u32 addr, status, mc_client; + bool queue_thermal = false; if (!rdev->ih.enabled || rdev->shutdown) return IRQ_NONE; @@ -5545,7 +7465,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[0])) - radeon_crtc_handle_flip(rdev, 0); + radeon_crtc_handle_vblank(rdev, 0); rdev->irq.stat_regs.cik.disp_int &= ~LB_D1_VBLANK_INTERRUPT; DRM_DEBUG("IH: D1 vblank\n"); } @@ -5571,7 +7491,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[1])) - radeon_crtc_handle_flip(rdev, 1); + radeon_crtc_handle_vblank(rdev, 1); rdev->irq.stat_regs.cik.disp_int_cont &= ~LB_D2_VBLANK_INTERRUPT; DRM_DEBUG("IH: D2 vblank\n"); } @@ -5597,7 +7517,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[2])) - radeon_crtc_handle_flip(rdev, 2); + radeon_crtc_handle_vblank(rdev, 2); rdev->irq.stat_regs.cik.disp_int_cont2 &= ~LB_D3_VBLANK_INTERRUPT; DRM_DEBUG("IH: D3 vblank\n"); } @@ -5623,7 +7543,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[3])) - radeon_crtc_handle_flip(rdev, 3); + radeon_crtc_handle_vblank(rdev, 3); rdev->irq.stat_regs.cik.disp_int_cont3 &= ~LB_D4_VBLANK_INTERRUPT; DRM_DEBUG("IH: D4 vblank\n"); } @@ -5649,7 +7569,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[4])) - radeon_crtc_handle_flip(rdev, 4); + radeon_crtc_handle_vblank(rdev, 4); rdev->irq.stat_regs.cik.disp_int_cont4 &= ~LB_D5_VBLANK_INTERRUPT; DRM_DEBUG("IH: D5 vblank\n"); } @@ -5675,7 +7595,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[5])) - radeon_crtc_handle_flip(rdev, 5); + radeon_crtc_handle_vblank(rdev, 5); rdev->irq.stat_regs.cik.disp_int_cont5 &= ~LB_D6_VBLANK_INTERRUPT; DRM_DEBUG("IH: D6 vblank\n"); } @@ -5691,6 +7611,15 @@ restart_ih: break; } break; + case 8: /* D1 page flip */ + case 10: /* D2 page flip */ + case 12: /* D3 page flip */ + case 14: /* D4 page flip */ + case 16: /* D5 page flip */ + case 18: /* D6 page flip */ + DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1); + radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1); + break; case 42: /* HPD hotplug */ switch (src_data) { case 0: @@ -5740,15 +7669,39 @@ restart_ih: break; } break; + case 124: /* UVD */ + DRM_DEBUG("IH: UVD int: 0x%08x\n", src_data); + radeon_fence_process(rdev, R600_RING_TYPE_UVD_INDEX); + break; case 146: case 147: + addr = RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR); + status = RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS); + mc_client = RREG32(VM_CONTEXT1_PROTECTION_FAULT_MCCLIENT); + /* reset addr and status */ + WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1); + if (addr == 0x0 && status == 0x0) + break; dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data); dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n", - RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR)); + addr); dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", - RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS)); - /* reset addr and status */ - WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1); + status); + cik_vm_decode_fault(rdev, status, addr, mc_client); + break; + case 167: /* VCE */ + DRM_DEBUG("IH: VCE int: 0x%08x\n", src_data); + switch (src_data) { + case 0: + radeon_fence_process(rdev, TN_RING_TYPE_VCE1_INDEX); + break; + case 1: + radeon_fence_process(rdev, TN_RING_TYPE_VCE2_INDEX); + break; + default: + DRM_ERROR("Unhandled interrupt: %d %d\n", src_id, src_data); + break; + } break; case 176: /* GFX RB CP_INT */ case 177: /* GFX IB CP_INT */ @@ -5853,6 +7806,19 @@ restart_ih: break; } break; + case 230: /* thermal low to high */ + DRM_DEBUG("IH: thermal low to high\n"); + rdev->pm.dpm.thermal.high_to_low = false; + queue_thermal = true; + break; + case 231: /* thermal high to low */ + DRM_DEBUG("IH: thermal high to low\n"); + rdev->pm.dpm.thermal.high_to_low = true; + queue_thermal = true; + break; + case 233: /* GUI IDLE */ + DRM_DEBUG("IH: GUI idle\n"); + break; case 241: /* SDMA Privileged inst */ case 247: /* SDMA Privileged inst */ DRM_ERROR("Illegal instruction in SDMA command stream\n"); @@ -5892,9 +7858,6 @@ restart_ih: break; } break; - case 233: /* GUI IDLE */ - DRM_DEBUG("IH: GUI idle\n"); - break; default: DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); break; @@ -5908,6 +7871,8 @@ restart_ih: schedule_work(&rdev->hotplug_work); if (queue_reset) schedule_work(&rdev->reset_work); + if (queue_thermal) + schedule_work(&rdev->pm.dpm.thermal.work); rdev->ih.rptr = rptr; WREG32(IH_RB_RPTR, rdev->ih.rptr); atomic_set(&rdev->ih.lock, 0); @@ -5937,26 +7902,19 @@ static int cik_startup(struct radeon_device *rdev) struct radeon_ring *ring; int r; - if (rdev->flags & RADEON_IS_IGP) { - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw || - !rdev->mec_fw || !rdev->sdma_fw || !rdev->rlc_fw) { - r = cik_init_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load firmware!\n"); - return r; - } - } - } else { - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw || - !rdev->mec_fw || !rdev->sdma_fw || !rdev->rlc_fw || - !rdev->mc_fw) { - r = cik_init_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load firmware!\n"); - return r; - } - } + /* enable pcie gen2/3 link */ + cik_pcie_gen3_enable(rdev); + /* enable aspm */ + cik_program_aspm(rdev); + /* scratch needs to be initialized before MC */ + r = r600_vram_scratch_init(rdev); + if (r) + return r; + + cik_mc_program(rdev); + + if (!(rdev->flags & RADEON_IS_IGP) && !rdev->pm.dpm_enabled) { r = ci_mc_load_microcode(rdev); if (r) { DRM_ERROR("Failed to load MC firmware!\n"); @@ -5964,18 +7922,26 @@ static int cik_startup(struct radeon_device *rdev) } } - r = r600_vram_scratch_init(rdev); - if (r) - return r; - - cik_mc_program(rdev); r = cik_pcie_gart_enable(rdev); if (r) return r; cik_gpu_init(rdev); /* allocate rlc buffers */ - r = si_rlc_init(rdev); + if (rdev->flags & RADEON_IS_IGP) { + if (rdev->family == CHIP_KAVERI) { + rdev->rlc.reg_list = spectre_rlc_save_restore_register_list; + rdev->rlc.reg_list_size = + (u32)ARRAY_SIZE(spectre_rlc_save_restore_register_list); + } else { + rdev->rlc.reg_list = kalindi_rlc_save_restore_register_list; + rdev->rlc.reg_list_size = + (u32)ARRAY_SIZE(kalindi_rlc_save_restore_register_list); + } + } + rdev->rlc.cs_data = ci_cs_data; + rdev->rlc.cp_table_size = CP_ME_TABLE_SIZE * 5 * 4; + r = sumo_rlc_init(rdev); if (r) { DRM_ERROR("Failed to init rlc BOs!\n"); return r; @@ -6023,16 +7989,35 @@ static int cik_startup(struct radeon_device *rdev) return r; } - r = cik_uvd_resume(rdev); + r = radeon_uvd_resume(rdev); if (!r) { - r = radeon_fence_driver_start_ring(rdev, - R600_RING_TYPE_UVD_INDEX); - if (r) - dev_err(rdev->dev, "UVD fences init error (%d).\n", r); + r = uvd_v4_2_resume(rdev); + if (!r) { + r = radeon_fence_driver_start_ring(rdev, + R600_RING_TYPE_UVD_INDEX); + if (r) + dev_err(rdev->dev, "UVD fences init error (%d).\n", r); + } } if (r) rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0; + r = radeon_vce_resume(rdev); + if (!r) { + r = vce_v2_0_resume(rdev); + if (!r) + r = radeon_fence_driver_start_ring(rdev, + TN_RING_TYPE_VCE1_INDEX); + if (!r) + r = radeon_fence_driver_start_ring(rdev, + TN_RING_TYPE_VCE2_INDEX); + } + if (r) { + dev_err(rdev->dev, "VCE init error (%d).\n", r); + rdev->ring[TN_RING_TYPE_VCE1_INDEX].ring_size = 0; + rdev->ring[TN_RING_TYPE_VCE2_INDEX].ring_size = 0; + } + /* Enable IRQ */ if (!rdev->irq.installed) { r = radeon_irq_kms_init(rdev); @@ -6050,8 +8035,7 @@ static int cik_startup(struct radeon_device *rdev) ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, - CP_RB0_RPTR, CP_RB0_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + PACKET3(PACKET3_NOP, 0x3FFF)); if (r) return r; @@ -6059,8 +8043,7 @@ static int cik_startup(struct radeon_device *rdev) /* type-2 packets are deprecated on MEC, use type-3 instead */ ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP1_RPTR_OFFSET, - CP_HQD_PQ_RPTR, CP_HQD_PQ_WPTR, - 0, 0xfffff, PACKET3(PACKET3_NOP, 0x3FFF)); + PACKET3(PACKET3_NOP, 0x3FFF)); if (r) return r; ring->me = 1; /* first MEC */ @@ -6071,8 +8054,7 @@ static int cik_startup(struct radeon_device *rdev) /* type-2 packets are deprecated on MEC, use type-3 instead */ ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP2_RPTR_OFFSET, - CP_HQD_PQ_RPTR, CP_HQD_PQ_WPTR, - 0, 0xffffffff, PACKET3(PACKET3_NOP, 0x3FFF)); + PACKET3(PACKET3_NOP, 0x3FFF)); if (r) return r; /* dGPU only have 1 MEC */ @@ -6083,17 +8065,13 @@ static int cik_startup(struct radeon_device *rdev) ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET, - SDMA0_GFX_RB_RPTR + SDMA0_REGISTER_OFFSET, - SDMA0_GFX_RB_WPTR + SDMA0_REGISTER_OFFSET, - 2, 0xfffffffc, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0)); + SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0)); if (r) return r; ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, CAYMAN_WB_DMA1_RPTR_OFFSET, - SDMA0_GFX_RB_RPTR + SDMA1_REGISTER_OFFSET, - SDMA0_GFX_RB_WPTR + SDMA1_REGISTER_OFFSET, - 2, 0xfffffffc, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0)); + SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0)); if (r) return r; @@ -6107,16 +8085,31 @@ static int cik_startup(struct radeon_device *rdev) ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; if (ring->ring_size) { - r = radeon_ring_init(rdev, ring, ring->ring_size, - R600_WB_UVD_RPTR_OFFSET, - UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, + RADEON_CP_PACKET2); if (!r) - r = r600_uvd_init(rdev); + r = uvd_v1_0_init(rdev); if (r) DRM_ERROR("radeon: failed initializing UVD (%d).\n", r); } + r = -ENOENT; + + ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX]; + if (ring->ring_size) + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, + VCE_CMD_NO_OP); + + ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX]; + if (ring->ring_size) + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, + VCE_CMD_NO_OP); + + if (!r) + r = vce_v1_0_init(rdev); + else if (r != -ENOENT) + DRM_ERROR("radeon: failed initializing VCE (%d).\n", r); + r = radeon_ib_pool_init(rdev); if (r) { dev_err(rdev->dev, "IB initialization failed (%d).\n", r); @@ -6129,6 +8122,10 @@ static int cik_startup(struct radeon_device *rdev) return r; } + r = dce6_audio_init(rdev); + if (r) + return r; + return 0; } @@ -6151,6 +8148,9 @@ int cik_resume(struct radeon_device *rdev) /* init golden registers */ cik_init_golden_registers(rdev); + if (rdev->pm.pm_method == PM_METHOD_DPM) + radeon_pm_resume(rdev); + rdev->accel_working = true; r = cik_startup(rdev); if (r) { @@ -6174,11 +8174,16 @@ int cik_resume(struct radeon_device *rdev) */ int cik_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); + dce6_audio_fini(rdev); radeon_vm_manager_fini(rdev); cik_cp_enable(rdev, false); cik_sdma_enable(rdev, false); - r600_uvd_rbc_stop(rdev); + uvd_v1_0_fini(rdev); radeon_uvd_suspend(rdev); + radeon_vce_suspend(rdev); + cik_fini_pg(rdev); + cik_fini_cg(rdev); cik_irq_suspend(rdev); radeon_wb_disable(rdev); cik_pcie_gart_disable(rdev); @@ -6252,6 +8257,30 @@ int cik_init(struct radeon_device *rdev) if (r) return r; + if (rdev->flags & RADEON_IS_IGP) { + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw || + !rdev->mec_fw || !rdev->sdma_fw || !rdev->rlc_fw) { + r = cik_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + } else { + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw || + !rdev->mec_fw || !rdev->sdma_fw || !rdev->rlc_fw || + !rdev->mc_fw) { + r = cik_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + } + + /* Initialize power management */ + radeon_pm_init(rdev); + ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; ring->ring_obj = NULL; r600_ring_init(rdev, ring, 1024 * 1024); @@ -6259,14 +8288,14 @@ int cik_init(struct radeon_device *rdev) ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; ring->ring_obj = NULL; r600_ring_init(rdev, ring, 1024 * 1024); - r = radeon_doorbell_get(rdev, &ring->doorbell_page_num); + r = radeon_doorbell_get(rdev, &ring->doorbell_index); if (r) return r; ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; ring->ring_obj = NULL; r600_ring_init(rdev, ring, 1024 * 1024); - r = radeon_doorbell_get(rdev, &ring->doorbell_page_num); + r = radeon_doorbell_get(rdev, &ring->doorbell_index); if (r) return r; @@ -6285,6 +8314,17 @@ int cik_init(struct radeon_device *rdev) r600_ring_init(rdev, ring, 4096); } + r = radeon_vce_init(rdev); + if (!r) { + ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX]; + ring->ring_obj = NULL; + r600_ring_init(rdev, ring, 4096); + + ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX]; + ring->ring_obj = NULL; + r600_ring_init(rdev, ring, 4096); + } + rdev->ih.ring_obj = NULL; r600_ih_ring_init(rdev, 64 * 1024); @@ -6299,7 +8339,7 @@ int cik_init(struct radeon_device *rdev) cik_cp_fini(rdev); cik_sdma_fini(rdev); cik_irq_fini(rdev); - si_rlc_fini(rdev); + sumo_rlc_fini(rdev); cik_mec_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); @@ -6332,16 +8372,21 @@ int cik_init(struct radeon_device *rdev) */ void cik_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); cik_cp_fini(rdev); cik_sdma_fini(rdev); + cik_fini_pg(rdev); + cik_fini_cg(rdev); cik_irq_fini(rdev); - si_rlc_fini(rdev); + sumo_rlc_fini(rdev); cik_mec_fini(rdev); radeon_wb_fini(rdev); radeon_vm_manager_fini(rdev); radeon_ib_pool_fini(rdev); radeon_irq_kms_fini(rdev); + uvd_v1_0_fini(rdev); radeon_uvd_fini(rdev); + radeon_vce_fini(rdev); cik_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_gem_fini(rdev); @@ -6352,6 +8397,70 @@ void cik_fini(struct radeon_device *rdev) rdev->bios = NULL; } +void dce8_program_fmt(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); + struct drm_connector *connector = radeon_get_connector_for_encoder(encoder); + int bpc = 0; + u32 tmp = 0; + enum radeon_connector_dither dither = RADEON_FMT_DITHER_DISABLE; + + if (connector) { + struct radeon_connector *radeon_connector = to_radeon_connector(connector); + bpc = radeon_get_monitor_bpc(connector); + dither = radeon_connector->dither; + } + + /* LVDS/eDP FMT is set up by atom */ + if (radeon_encoder->devices & ATOM_DEVICE_LCD_SUPPORT) + return; + + /* not needed for analog */ + if ((radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1) || + (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2)) + return; + + if (bpc == 0) + return; + + switch (bpc) { + case 6: + if (dither == RADEON_FMT_DITHER_ENABLE) + /* XXX sort out optimal dither settings */ + tmp |= (FMT_FRAME_RANDOM_ENABLE | FMT_HIGHPASS_RANDOM_ENABLE | + FMT_SPATIAL_DITHER_EN | FMT_SPATIAL_DITHER_DEPTH(0)); + else + tmp |= (FMT_TRUNCATE_EN | FMT_TRUNCATE_DEPTH(0)); + break; + case 8: + if (dither == RADEON_FMT_DITHER_ENABLE) + /* XXX sort out optimal dither settings */ + tmp |= (FMT_FRAME_RANDOM_ENABLE | FMT_HIGHPASS_RANDOM_ENABLE | + FMT_RGB_RANDOM_ENABLE | + FMT_SPATIAL_DITHER_EN | FMT_SPATIAL_DITHER_DEPTH(1)); + else + tmp |= (FMT_TRUNCATE_EN | FMT_TRUNCATE_DEPTH(1)); + break; + case 10: + if (dither == RADEON_FMT_DITHER_ENABLE) + /* XXX sort out optimal dither settings */ + tmp |= (FMT_FRAME_RANDOM_ENABLE | FMT_HIGHPASS_RANDOM_ENABLE | + FMT_RGB_RANDOM_ENABLE | + FMT_SPATIAL_DITHER_EN | FMT_SPATIAL_DITHER_DEPTH(2)); + else + tmp |= (FMT_TRUNCATE_EN | FMT_TRUNCATE_DEPTH(2)); + break; + default: + /* not needed */ + break; + } + + WREG32(FMT_BIT_DEPTH_CONTROL + radeon_crtc->crtc_offset, tmp); +} + /* display watermark setup */ /** * dce8_line_buffer_adjust - Set up the line buffer @@ -6369,8 +8478,8 @@ static u32 dce8_line_buffer_adjust(struct radeon_device *rdev, struct radeon_crtc *radeon_crtc, struct drm_display_mode *mode) { - u32 tmp; - + u32 tmp, buffer_alloc, i; + u32 pipe_offset = radeon_crtc->crtc_id * 0x20; /* * Line Buffer Setup * There are 6 line buffers, one for each display controllers. @@ -6380,22 +8489,37 @@ static u32 dce8_line_buffer_adjust(struct radeon_device *rdev, * them using the stereo blender. */ if (radeon_crtc->base.enabled && mode) { - if (mode->crtc_hdisplay < 1920) + if (mode->crtc_hdisplay < 1920) { tmp = 1; - else if (mode->crtc_hdisplay < 2560) + buffer_alloc = 2; + } else if (mode->crtc_hdisplay < 2560) { tmp = 2; - else if (mode->crtc_hdisplay < 4096) + buffer_alloc = 2; + } else if (mode->crtc_hdisplay < 4096) { tmp = 0; - else { + buffer_alloc = (rdev->flags & RADEON_IS_IGP) ? 2 : 4; + } else { DRM_DEBUG_KMS("Mode too big for LB!\n"); tmp = 0; + buffer_alloc = (rdev->flags & RADEON_IS_IGP) ? 2 : 4; } - } else + } else { tmp = 1; + buffer_alloc = 0; + } WREG32(LB_MEMORY_CTRL + radeon_crtc->crtc_offset, LB_MEMORY_CONFIG(tmp) | LB_MEMORY_SIZE(0x6B0)); + WREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset, + DMIF_BUFFERS_ALLOCATED(buffer_alloc)); + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset) & + DMIF_BUFFERS_ALLOCATED_COMPLETED) + break; + udelay(1); + } + if (radeon_crtc->base.enabled && mode) { switch (tmp) { case 0: @@ -6797,7 +8921,7 @@ static void dce8_program_watermarks(struct radeon_device *rdev, u32 lb_size, u32 num_heads) { struct drm_display_mode *mode = &radeon_crtc->base.mode; - struct dce8_wm_params wm; + struct dce8_wm_params wm_low, wm_high; u32 pixel_period; u32 line_time = 0; u32 latency_watermark_a = 0, latency_watermark_b = 0; @@ -6807,35 +8931,82 @@ static void dce8_program_watermarks(struct radeon_device *rdev, pixel_period = 1000000 / (u32)mode->clock; line_time = min((u32)mode->crtc_htotal * pixel_period, (u32)65535); - wm.yclk = rdev->pm.current_mclk * 10; - wm.sclk = rdev->pm.current_sclk * 10; - wm.disp_clk = mode->clock; - wm.src_width = mode->crtc_hdisplay; - wm.active_time = mode->crtc_hdisplay * pixel_period; - wm.blank_time = line_time - wm.active_time; - wm.interlaced = false; + /* watermark for high clocks */ + if ((rdev->pm.pm_method == PM_METHOD_DPM) && + rdev->pm.dpm_enabled) { + wm_high.yclk = + radeon_dpm_get_mclk(rdev, false) * 10; + wm_high.sclk = + radeon_dpm_get_sclk(rdev, false) * 10; + } else { + wm_high.yclk = rdev->pm.current_mclk * 10; + wm_high.sclk = rdev->pm.current_sclk * 10; + } + + wm_high.disp_clk = mode->clock; + wm_high.src_width = mode->crtc_hdisplay; + wm_high.active_time = mode->crtc_hdisplay * pixel_period; + wm_high.blank_time = line_time - wm_high.active_time; + wm_high.interlaced = false; if (mode->flags & DRM_MODE_FLAG_INTERLACE) - wm.interlaced = true; - wm.vsc = radeon_crtc->vsc; - wm.vtaps = 1; + wm_high.interlaced = true; + wm_high.vsc = radeon_crtc->vsc; + wm_high.vtaps = 1; if (radeon_crtc->rmx_type != RMX_OFF) - wm.vtaps = 2; - wm.bytes_per_pixel = 4; /* XXX: get this from fb config */ - wm.lb_size = lb_size; - wm.dram_channels = cik_get_number_of_dram_channels(rdev); - wm.num_heads = num_heads; + wm_high.vtaps = 2; + wm_high.bytes_per_pixel = 4; /* XXX: get this from fb config */ + wm_high.lb_size = lb_size; + wm_high.dram_channels = cik_get_number_of_dram_channels(rdev); + wm_high.num_heads = num_heads; /* set for high clocks */ - latency_watermark_a = min(dce8_latency_watermark(&wm), (u32)65535); + latency_watermark_a = min(dce8_latency_watermark(&wm_high), (u32)65535); + + /* possibly force display priority to high */ + /* should really do this at mode validation time... */ + if (!dce8_average_bandwidth_vs_dram_bandwidth_for_display(&wm_high) || + !dce8_average_bandwidth_vs_available_bandwidth(&wm_high) || + !dce8_check_latency_hiding(&wm_high) || + (rdev->disp_priority == 2)) { + DRM_DEBUG_KMS("force priority to high\n"); + } + + /* watermark for low clocks */ + if ((rdev->pm.pm_method == PM_METHOD_DPM) && + rdev->pm.dpm_enabled) { + wm_low.yclk = + radeon_dpm_get_mclk(rdev, true) * 10; + wm_low.sclk = + radeon_dpm_get_sclk(rdev, true) * 10; + } else { + wm_low.yclk = rdev->pm.current_mclk * 10; + wm_low.sclk = rdev->pm.current_sclk * 10; + } + + wm_low.disp_clk = mode->clock; + wm_low.src_width = mode->crtc_hdisplay; + wm_low.active_time = mode->crtc_hdisplay * pixel_period; + wm_low.blank_time = line_time - wm_low.active_time; + wm_low.interlaced = false; + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + wm_low.interlaced = true; + wm_low.vsc = radeon_crtc->vsc; + wm_low.vtaps = 1; + if (radeon_crtc->rmx_type != RMX_OFF) + wm_low.vtaps = 2; + wm_low.bytes_per_pixel = 4; /* XXX: get this from fb config */ + wm_low.lb_size = lb_size; + wm_low.dram_channels = cik_get_number_of_dram_channels(rdev); + wm_low.num_heads = num_heads; + /* set for low clocks */ - /* wm.yclk = low clk; wm.sclk = low clk */ - latency_watermark_b = min(dce8_latency_watermark(&wm), (u32)65535); + latency_watermark_b = min(dce8_latency_watermark(&wm_low), (u32)65535); /* possibly force display priority to high */ /* should really do this at mode validation time... */ - if (!dce8_average_bandwidth_vs_dram_bandwidth_for_display(&wm) || - !dce8_average_bandwidth_vs_available_bandwidth(&wm) || - !dce8_check_latency_hiding(&wm) || + if (!dce8_average_bandwidth_vs_dram_bandwidth_for_display(&wm_low) || + !dce8_average_bandwidth_vs_available_bandwidth(&wm_low) || + !dce8_check_latency_hiding(&wm_low) || (rdev->disp_priority == 2)) { DRM_DEBUG_KMS("force priority to high\n"); } @@ -6860,6 +9031,11 @@ static void dce8_program_watermarks(struct radeon_device *rdev, LATENCY_HIGH_WATERMARK(line_time))); /* restore original selection */ WREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset, wm_mask); + + /* save values for DPM */ + radeon_crtc->line_time = line_time; + radeon_crtc->wm_high = latency_watermark_a; + radeon_crtc->wm_low = latency_watermark_b; } /** @@ -6949,39 +9125,342 @@ int cik_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk) return r; } -int cik_uvd_resume(struct radeon_device *rdev) +int cik_set_vce_clocks(struct radeon_device *rdev, u32 evclk, u32 ecclk) { - uint64_t addr; - uint32_t size; - int r; + int r, i; + struct atom_clock_dividers dividers; + u32 tmp; - r = radeon_uvd_resume(rdev); + r = radeon_atom_get_clock_dividers(rdev, COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, + ecclk, false, ÷rs); if (r) return r; - /* programm the VCPU memory controller bits 0-27 */ - addr = rdev->uvd.gpu_addr >> 3; - size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 4) >> 3; - WREG32(UVD_VCPU_CACHE_OFFSET0, addr); - WREG32(UVD_VCPU_CACHE_SIZE0, size); + for (i = 0; i < 100; i++) { + if (RREG32_SMC(CG_ECLK_STATUS) & ECLK_STATUS) + break; + mdelay(10); + } + if (i == 100) + return -ETIMEDOUT; + + tmp = RREG32_SMC(CG_ECLK_CNTL); + tmp &= ~(ECLK_DIR_CNTL_EN|ECLK_DIVIDER_MASK); + tmp |= dividers.post_divider; + WREG32_SMC(CG_ECLK_CNTL, tmp); + + for (i = 0; i < 100; i++) { + if (RREG32_SMC(CG_ECLK_STATUS) & ECLK_STATUS) + break; + mdelay(10); + } + if (i == 100) + return -ETIMEDOUT; - addr += size; - size = RADEON_UVD_STACK_SIZE >> 3; - WREG32(UVD_VCPU_CACHE_OFFSET1, addr); - WREG32(UVD_VCPU_CACHE_SIZE1, size); + return 0; +} - addr += size; - size = RADEON_UVD_HEAP_SIZE >> 3; - WREG32(UVD_VCPU_CACHE_OFFSET2, addr); - WREG32(UVD_VCPU_CACHE_SIZE2, size); +static void cik_pcie_gen3_enable(struct radeon_device *rdev) +{ + struct pci_dev *root = rdev->pdev->bus->self; + int bridge_pos, gpu_pos; + u32 speed_cntl, mask, current_data_rate; + int ret, i; + u16 tmp16; - /* bits 28-31 */ - addr = (rdev->uvd.gpu_addr >> 28) & 0xF; - WREG32(UVD_LMI_ADDR_EXT, (addr << 12) | (addr << 0)); + if (radeon_pcie_gen2 == 0) + return; - /* bits 32-39 */ - addr = (rdev->uvd.gpu_addr >> 32) & 0xFF; - WREG32(UVD_LMI_EXT40_ADDR, addr | (0x9 << 16) | (0x1 << 31)); + if (rdev->flags & RADEON_IS_IGP) + return; - return 0; + if (!(rdev->flags & RADEON_IS_PCIE)) + return; + + ret = drm_pcie_get_speed_cap_mask(rdev->ddev, &mask); + if (ret != 0) + return; + + if (!(mask & (DRM_PCIE_SPEED_50 | DRM_PCIE_SPEED_80))) + return; + + speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); + current_data_rate = (speed_cntl & LC_CURRENT_DATA_RATE_MASK) >> + LC_CURRENT_DATA_RATE_SHIFT; + if (mask & DRM_PCIE_SPEED_80) { + if (current_data_rate == 2) { + DRM_INFO("PCIE gen 3 link speeds already enabled\n"); + return; + } + DRM_INFO("enabling PCIE gen 3 link speeds, disable with radeon.pcie_gen2=0\n"); + } else if (mask & DRM_PCIE_SPEED_50) { + if (current_data_rate == 1) { + DRM_INFO("PCIE gen 2 link speeds already enabled\n"); + return; + } + DRM_INFO("enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0\n"); + } + + bridge_pos = pci_pcie_cap(root); + if (!bridge_pos) + return; + + gpu_pos = pci_pcie_cap(rdev->pdev); + if (!gpu_pos) + return; + + if (mask & DRM_PCIE_SPEED_80) { + /* re-try equalization if gen3 is not already enabled */ + if (current_data_rate != 2) { + u16 bridge_cfg, gpu_cfg; + u16 bridge_cfg2, gpu_cfg2; + u32 max_lw, current_lw, tmp; + + pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &bridge_cfg); + pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &gpu_cfg); + + tmp16 = bridge_cfg | PCI_EXP_LNKCTL_HAWD; + pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL, tmp16); + + tmp16 = gpu_cfg | PCI_EXP_LNKCTL_HAWD; + pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, tmp16); + + tmp = RREG32_PCIE_PORT(PCIE_LC_STATUS1); + max_lw = (tmp & LC_DETECTED_LINK_WIDTH_MASK) >> LC_DETECTED_LINK_WIDTH_SHIFT; + current_lw = (tmp & LC_OPERATING_LINK_WIDTH_MASK) >> LC_OPERATING_LINK_WIDTH_SHIFT; + + if (current_lw < max_lw) { + tmp = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL); + if (tmp & LC_RENEGOTIATION_SUPPORT) { + tmp &= ~(LC_LINK_WIDTH_MASK | LC_UPCONFIGURE_DIS); + tmp |= (max_lw << LC_LINK_WIDTH_SHIFT); + tmp |= LC_UPCONFIGURE_SUPPORT | LC_RENEGOTIATE_EN | LC_RECONFIG_NOW; + WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, tmp); + } + } + + for (i = 0; i < 10; i++) { + /* check status */ + pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_DEVSTA, &tmp16); + if (tmp16 & PCI_EXP_DEVSTA_TRPND) + break; + + pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &bridge_cfg); + pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &gpu_cfg); + + pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, &bridge_cfg2); + pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &gpu_cfg2); + + tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4); + tmp |= LC_SET_QUIESCE; + WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp); + + tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4); + tmp |= LC_REDO_EQ; + WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp); + + mdelay(100); + + /* linkctl */ + pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &tmp16); + tmp16 &= ~PCI_EXP_LNKCTL_HAWD; + tmp16 |= (bridge_cfg & PCI_EXP_LNKCTL_HAWD); + pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL, tmp16); + + pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &tmp16); + tmp16 &= ~PCI_EXP_LNKCTL_HAWD; + tmp16 |= (gpu_cfg & PCI_EXP_LNKCTL_HAWD); + pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, tmp16); + + /* linkctl2 */ + pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, &tmp16); + tmp16 &= ~((1 << 4) | (7 << 9)); + tmp16 |= (bridge_cfg2 & ((1 << 4) | (7 << 9))); + pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, tmp16); + + pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &tmp16); + tmp16 &= ~((1 << 4) | (7 << 9)); + tmp16 |= (gpu_cfg2 & ((1 << 4) | (7 << 9))); + pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, tmp16); + + tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4); + tmp &= ~LC_SET_QUIESCE; + WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp); + } + } + } + + /* set the link speed */ + speed_cntl |= LC_FORCE_EN_SW_SPEED_CHANGE | LC_FORCE_DIS_HW_SPEED_CHANGE; + speed_cntl &= ~LC_FORCE_DIS_SW_SPEED_CHANGE; + WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl); + + pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &tmp16); + tmp16 &= ~0xf; + if (mask & DRM_PCIE_SPEED_80) + tmp16 |= 3; /* gen3 */ + else if (mask & DRM_PCIE_SPEED_50) + tmp16 |= 2; /* gen2 */ + else + tmp16 |= 1; /* gen1 */ + pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, tmp16); + + speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); + speed_cntl |= LC_INITIATE_LINK_SPEED_CHANGE; + WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl); + + for (i = 0; i < rdev->usec_timeout; i++) { + speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); + if ((speed_cntl & LC_INITIATE_LINK_SPEED_CHANGE) == 0) + break; + udelay(1); + } +} + +static void cik_program_aspm(struct radeon_device *rdev) +{ + u32 data, orig; + bool disable_l0s = false, disable_l1 = false, disable_plloff_in_l1 = false; + bool disable_clkreq = false; + + if (radeon_aspm == 0) + return; + + /* XXX double check IGPs */ + if (rdev->flags & RADEON_IS_IGP) + return; + + if (!(rdev->flags & RADEON_IS_PCIE)) + return; + + orig = data = RREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL); + data &= ~LC_XMIT_N_FTS_MASK; + data |= LC_XMIT_N_FTS(0x24) | LC_XMIT_N_FTS_OVERRIDE_EN; + if (orig != data) + WREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL, data); + + orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL3); + data |= LC_GO_TO_RECOVERY; + if (orig != data) + WREG32_PCIE_PORT(PCIE_LC_CNTL3, data); + + orig = data = RREG32_PCIE_PORT(PCIE_P_CNTL); + data |= P_IGNORE_EDB_ERR; + if (orig != data) + WREG32_PCIE_PORT(PCIE_P_CNTL, data); + + orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL); + data &= ~(LC_L0S_INACTIVITY_MASK | LC_L1_INACTIVITY_MASK); + data |= LC_PMI_TO_L1_DIS; + if (!disable_l0s) + data |= LC_L0S_INACTIVITY(7); + + if (!disable_l1) { + data |= LC_L1_INACTIVITY(7); + data &= ~LC_PMI_TO_L1_DIS; + if (orig != data) + WREG32_PCIE_PORT(PCIE_LC_CNTL, data); + + if (!disable_plloff_in_l1) { + bool clk_req_support; + + orig = data = RREG32_PCIE_PORT(PB0_PIF_PWRDOWN_0); + data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK); + data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7); + if (orig != data) + WREG32_PCIE_PORT(PB0_PIF_PWRDOWN_0, data); + + orig = data = RREG32_PCIE_PORT(PB0_PIF_PWRDOWN_1); + data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK); + data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7); + if (orig != data) + WREG32_PCIE_PORT(PB0_PIF_PWRDOWN_1, data); + + orig = data = RREG32_PCIE_PORT(PB1_PIF_PWRDOWN_0); + data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK); + data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7); + if (orig != data) + WREG32_PCIE_PORT(PB1_PIF_PWRDOWN_0, data); + + orig = data = RREG32_PCIE_PORT(PB1_PIF_PWRDOWN_1); + data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK); + data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7); + if (orig != data) + WREG32_PCIE_PORT(PB1_PIF_PWRDOWN_1, data); + + orig = data = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL); + data &= ~LC_DYN_LANES_PWR_STATE_MASK; + data |= LC_DYN_LANES_PWR_STATE(3); + if (orig != data) + WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, data); + + if (!disable_clkreq) { + struct pci_dev *root = rdev->pdev->bus->self; + u32 lnkcap; + + clk_req_support = false; + pcie_capability_read_dword(root, PCI_EXP_LNKCAP, &lnkcap); + if (lnkcap & PCI_EXP_LNKCAP_CLKPM) + clk_req_support = true; + } else { + clk_req_support = false; + } + + if (clk_req_support) { + orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL2); + data |= LC_ALLOW_PDWN_IN_L1 | LC_ALLOW_PDWN_IN_L23; + if (orig != data) + WREG32_PCIE_PORT(PCIE_LC_CNTL2, data); + + orig = data = RREG32_SMC(THM_CLK_CNTL); + data &= ~(CMON_CLK_SEL_MASK | TMON_CLK_SEL_MASK); + data |= CMON_CLK_SEL(1) | TMON_CLK_SEL(1); + if (orig != data) + WREG32_SMC(THM_CLK_CNTL, data); + + orig = data = RREG32_SMC(MISC_CLK_CTRL); + data &= ~(DEEP_SLEEP_CLK_SEL_MASK | ZCLK_SEL_MASK); + data |= DEEP_SLEEP_CLK_SEL(1) | ZCLK_SEL(1); + if (orig != data) + WREG32_SMC(MISC_CLK_CTRL, data); + + orig = data = RREG32_SMC(CG_CLKPIN_CNTL); + data &= ~BCLK_AS_XCLK; + if (orig != data) + WREG32_SMC(CG_CLKPIN_CNTL, data); + + orig = data = RREG32_SMC(CG_CLKPIN_CNTL_2); + data &= ~FORCE_BIF_REFCLK_EN; + if (orig != data) + WREG32_SMC(CG_CLKPIN_CNTL_2, data); + + orig = data = RREG32_SMC(MPLL_BYPASSCLK_SEL); + data &= ~MPLL_CLKOUT_SEL_MASK; + data |= MPLL_CLKOUT_SEL(4); + if (orig != data) + WREG32_SMC(MPLL_BYPASSCLK_SEL, data); + } + } + } else { + if (orig != data) + WREG32_PCIE_PORT(PCIE_LC_CNTL, data); + } + + orig = data = RREG32_PCIE_PORT(PCIE_CNTL2); + data |= SLV_MEM_LS_EN | MST_MEM_LS_EN | REPLAY_MEM_LS_EN; + if (orig != data) + WREG32_PCIE_PORT(PCIE_CNTL2, data); + + if (!disable_l0s) { + data = RREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL); + if((data & LC_N_FTS_MASK) == LC_N_FTS_MASK) { + data = RREG32_PCIE_PORT(PCIE_LC_STATUS1); + if ((data & LC_REVERSE_XMIT) && (data & LC_REVERSE_RCVR)) { + orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL); + data &= ~LC_L0S_INACTIVITY_MASK; + if (orig != data) + WREG32_PCIE_PORT(PCIE_LC_CNTL, data); + } + } + } } diff --git a/drivers/gpu/drm/radeon/cik_reg.h b/drivers/gpu/drm/radeon/cik_reg.h index d71e46d571f..ca1bb613358 100644 --- a/drivers/gpu/drm/radeon/cik_reg.h +++ b/drivers/gpu/drm/radeon/cik_reg.h @@ -24,6 +24,9 @@ #ifndef __CIK_REG_H__ #define __CIK_REG_H__ +#define CIK_DIDT_IND_INDEX 0xca00 +#define CIK_DIDT_IND_DATA 0xca04 + #define CIK_DC_GPIO_HPD_MASK 0x65b0 #define CIK_DC_GPIO_HPD_A 0x65b4 #define CIK_DC_GPIO_HPD_EN 0x65b8 diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c new file mode 100644 index 00000000000..8e9d0f1d858 --- /dev/null +++ b/drivers/gpu/drm/radeon/cik_sdma.c @@ -0,0 +1,870 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#include <linux/firmware.h> +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "radeon_trace.h" +#include "cikd.h" + +/* sdma */ +#define CIK_SDMA_UCODE_SIZE 1050 +#define CIK_SDMA_UCODE_VERSION 64 + +u32 cik_gpu_check_soft_reset(struct radeon_device *rdev); + +/* + * sDMA - System DMA + * Starting with CIK, the GPU has new asynchronous + * DMA engines. These engines are used for compute + * and gfx. There are two DMA engines (SDMA0, SDMA1) + * and each one supports 1 ring buffer used for gfx + * and 2 queues used for compute. + * + * The programming model is very similar to the CP + * (ring buffer, IBs, etc.), but sDMA has it's own + * packet format that is different from the PM4 format + * used by the CP. sDMA supports copying data, writing + * embedded data, solid fills, and a number of other + * things. It also has support for tiling/detiling of + * buffers. + */ + +/** + * cik_sdma_get_rptr - get the current read pointer + * + * @rdev: radeon_device pointer + * @ring: radeon ring pointer + * + * Get the current rptr from the hardware (CIK+). + */ +uint32_t cik_sdma_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 rptr, reg; + + if (rdev->wb.enabled) { + rptr = rdev->wb.wb[ring->rptr_offs/4]; + } else { + if (ring->idx == R600_RING_TYPE_DMA_INDEX) + reg = SDMA0_GFX_RB_RPTR + SDMA0_REGISTER_OFFSET; + else + reg = SDMA0_GFX_RB_RPTR + SDMA1_REGISTER_OFFSET; + + rptr = RREG32(reg); + } + + return (rptr & 0x3fffc) >> 2; +} + +/** + * cik_sdma_get_wptr - get the current write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon ring pointer + * + * Get the current wptr from the hardware (CIK+). + */ +uint32_t cik_sdma_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 reg; + + if (ring->idx == R600_RING_TYPE_DMA_INDEX) + reg = SDMA0_GFX_RB_WPTR + SDMA0_REGISTER_OFFSET; + else + reg = SDMA0_GFX_RB_WPTR + SDMA1_REGISTER_OFFSET; + + return (RREG32(reg) & 0x3fffc) >> 2; +} + +/** + * cik_sdma_set_wptr - commit the write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon ring pointer + * + * Write the wptr back to the hardware (CIK+). + */ +void cik_sdma_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 reg; + + if (ring->idx == R600_RING_TYPE_DMA_INDEX) + reg = SDMA0_GFX_RB_WPTR + SDMA0_REGISTER_OFFSET; + else + reg = SDMA0_GFX_RB_WPTR + SDMA1_REGISTER_OFFSET; + + WREG32(reg, (ring->wptr << 2) & 0x3fffc); +} + +/** + * cik_sdma_ring_ib_execute - Schedule an IB on the DMA engine + * + * @rdev: radeon_device pointer + * @ib: IB object to schedule + * + * Schedule an IB in the DMA ring (CIK). + */ +void cik_sdma_ring_ib_execute(struct radeon_device *rdev, + struct radeon_ib *ib) +{ + struct radeon_ring *ring = &rdev->ring[ib->ring]; + u32 extra_bits = (ib->vm ? ib->vm->id : 0) & 0xf; + + if (rdev->wb.enabled) { + u32 next_rptr = ring->wptr + 5; + while ((next_rptr & 7) != 4) + next_rptr++; + next_rptr += 4; + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0)); + radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); + radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr)); + radeon_ring_write(ring, 1); /* number of DWs to follow */ + radeon_ring_write(ring, next_rptr); + } + + /* IB packet must end on a 8 DW boundary */ + while ((ring->wptr & 7) != 4) + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0)); + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_INDIRECT_BUFFER, 0, extra_bits)); + radeon_ring_write(ring, ib->gpu_addr & 0xffffffe0); /* base must be 32 byte aligned */ + radeon_ring_write(ring, upper_32_bits(ib->gpu_addr)); + radeon_ring_write(ring, ib->length_dw); + +} + +/** + * cik_sdma_hdp_flush_ring_emit - emit an hdp flush on the DMA ring + * + * @rdev: radeon_device pointer + * @ridx: radeon ring index + * + * Emit an hdp flush packet on the requested DMA ring. + */ +static void cik_sdma_hdp_flush_ring_emit(struct radeon_device *rdev, + int ridx) +{ + struct radeon_ring *ring = &rdev->ring[ridx]; + u32 extra_bits = (SDMA_POLL_REG_MEM_EXTRA_OP(1) | + SDMA_POLL_REG_MEM_EXTRA_FUNC(3)); /* == */ + u32 ref_and_mask; + + if (ridx == R600_RING_TYPE_DMA_INDEX) + ref_and_mask = SDMA0; + else + ref_and_mask = SDMA1; + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_POLL_REG_MEM, 0, extra_bits)); + radeon_ring_write(ring, GPU_HDP_FLUSH_DONE); + radeon_ring_write(ring, GPU_HDP_FLUSH_REQ); + radeon_ring_write(ring, ref_and_mask); /* reference */ + radeon_ring_write(ring, ref_and_mask); /* mask */ + radeon_ring_write(ring, (0xfff << 16) | 10); /* retry count, poll interval */ +} + +/** + * cik_sdma_fence_ring_emit - emit a fence on the DMA ring + * + * @rdev: radeon_device pointer + * @fence: radeon fence object + * + * Add a DMA fence packet to the ring to write + * the fence seq number and DMA trap packet to generate + * an interrupt if needed (CIK). + */ +void cik_sdma_fence_ring_emit(struct radeon_device *rdev, + struct radeon_fence *fence) +{ + struct radeon_ring *ring = &rdev->ring[fence->ring]; + u64 addr = rdev->fence_drv[fence->ring].gpu_addr; + + /* write the fence */ + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_FENCE, 0, 0)); + radeon_ring_write(ring, lower_32_bits(addr)); + radeon_ring_write(ring, upper_32_bits(addr)); + radeon_ring_write(ring, fence->seq); + /* generate an interrupt */ + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_TRAP, 0, 0)); + /* flush HDP */ + cik_sdma_hdp_flush_ring_emit(rdev, fence->ring); +} + +/** + * cik_sdma_semaphore_ring_emit - emit a semaphore on the dma ring + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * @semaphore: radeon semaphore object + * @emit_wait: wait or signal semaphore + * + * Add a DMA semaphore packet to the ring wait on or signal + * other rings (CIK). + */ +bool cik_sdma_semaphore_ring_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait) +{ + u64 addr = semaphore->gpu_addr; + u32 extra_bits = emit_wait ? 0 : SDMA_SEMAPHORE_EXTRA_S; + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SEMAPHORE, 0, extra_bits)); + radeon_ring_write(ring, addr & 0xfffffff8); + radeon_ring_write(ring, upper_32_bits(addr)); + + return true; +} + +/** + * cik_sdma_gfx_stop - stop the gfx async dma engines + * + * @rdev: radeon_device pointer + * + * Stop the gfx async dma ring buffers (CIK). + */ +static void cik_sdma_gfx_stop(struct radeon_device *rdev) +{ + u32 rb_cntl, reg_offset; + int i; + + if ((rdev->asic->copy.copy_ring_index == R600_RING_TYPE_DMA_INDEX) || + (rdev->asic->copy.copy_ring_index == CAYMAN_RING_TYPE_DMA1_INDEX)) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + + for (i = 0; i < 2; i++) { + if (i == 0) + reg_offset = SDMA0_REGISTER_OFFSET; + else + reg_offset = SDMA1_REGISTER_OFFSET; + rb_cntl = RREG32(SDMA0_GFX_RB_CNTL + reg_offset); + rb_cntl &= ~SDMA_RB_ENABLE; + WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl); + WREG32(SDMA0_GFX_IB_CNTL + reg_offset, 0); + } + rdev->ring[R600_RING_TYPE_DMA_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX].ready = false; +} + +/** + * cik_sdma_rlc_stop - stop the compute async dma engines + * + * @rdev: radeon_device pointer + * + * Stop the compute async dma queues (CIK). + */ +static void cik_sdma_rlc_stop(struct radeon_device *rdev) +{ + /* XXX todo */ +} + +/** + * cik_sdma_enable - stop the async dma engines + * + * @rdev: radeon_device pointer + * @enable: enable/disable the DMA MEs. + * + * Halt or unhalt the async dma engines (CIK). + */ +void cik_sdma_enable(struct radeon_device *rdev, bool enable) +{ + u32 me_cntl, reg_offset; + int i; + + if (enable == false) { + cik_sdma_gfx_stop(rdev); + cik_sdma_rlc_stop(rdev); + } + + for (i = 0; i < 2; i++) { + if (i == 0) + reg_offset = SDMA0_REGISTER_OFFSET; + else + reg_offset = SDMA1_REGISTER_OFFSET; + me_cntl = RREG32(SDMA0_ME_CNTL + reg_offset); + if (enable) + me_cntl &= ~SDMA_HALT; + else + me_cntl |= SDMA_HALT; + WREG32(SDMA0_ME_CNTL + reg_offset, me_cntl); + } +} + +/** + * cik_sdma_gfx_resume - setup and start the async dma engines + * + * @rdev: radeon_device pointer + * + * Set up the gfx DMA ring buffers and enable them (CIK). + * Returns 0 for success, error for failure. + */ +static int cik_sdma_gfx_resume(struct radeon_device *rdev) +{ + struct radeon_ring *ring; + u32 rb_cntl, ib_cntl; + u32 rb_bufsz; + u32 reg_offset, wb_offset; + int i, r; + + for (i = 0; i < 2; i++) { + if (i == 0) { + ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; + reg_offset = SDMA0_REGISTER_OFFSET; + wb_offset = R600_WB_DMA_RPTR_OFFSET; + } else { + ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]; + reg_offset = SDMA1_REGISTER_OFFSET; + wb_offset = CAYMAN_WB_DMA1_RPTR_OFFSET; + } + + WREG32(SDMA0_SEM_INCOMPLETE_TIMER_CNTL + reg_offset, 0); + WREG32(SDMA0_SEM_WAIT_FAIL_TIMER_CNTL + reg_offset, 0); + + /* Set ring buffer size in dwords */ + rb_bufsz = order_base_2(ring->ring_size / 4); + rb_cntl = rb_bufsz << 1; +#ifdef __BIG_ENDIAN + rb_cntl |= SDMA_RB_SWAP_ENABLE | SDMA_RPTR_WRITEBACK_SWAP_ENABLE; +#endif + WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl); + + /* Initialize the ring buffer's read and write pointers */ + WREG32(SDMA0_GFX_RB_RPTR + reg_offset, 0); + WREG32(SDMA0_GFX_RB_WPTR + reg_offset, 0); + + /* set the wb address whether it's enabled or not */ + WREG32(SDMA0_GFX_RB_RPTR_ADDR_HI + reg_offset, + upper_32_bits(rdev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF); + WREG32(SDMA0_GFX_RB_RPTR_ADDR_LO + reg_offset, + ((rdev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC)); + + if (rdev->wb.enabled) + rb_cntl |= SDMA_RPTR_WRITEBACK_ENABLE; + + WREG32(SDMA0_GFX_RB_BASE + reg_offset, ring->gpu_addr >> 8); + WREG32(SDMA0_GFX_RB_BASE_HI + reg_offset, ring->gpu_addr >> 40); + + ring->wptr = 0; + WREG32(SDMA0_GFX_RB_WPTR + reg_offset, ring->wptr << 2); + + /* enable DMA RB */ + WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl | SDMA_RB_ENABLE); + + ib_cntl = SDMA_IB_ENABLE; +#ifdef __BIG_ENDIAN + ib_cntl |= SDMA_IB_SWAP_ENABLE; +#endif + /* enable DMA IBs */ + WREG32(SDMA0_GFX_IB_CNTL + reg_offset, ib_cntl); + + ring->ready = true; + + r = radeon_ring_test(rdev, ring->idx, ring); + if (r) { + ring->ready = false; + return r; + } + } + + if ((rdev->asic->copy.copy_ring_index == R600_RING_TYPE_DMA_INDEX) || + (rdev->asic->copy.copy_ring_index == CAYMAN_RING_TYPE_DMA1_INDEX)) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); + + return 0; +} + +/** + * cik_sdma_rlc_resume - setup and start the async dma engines + * + * @rdev: radeon_device pointer + * + * Set up the compute DMA queues and enable them (CIK). + * Returns 0 for success, error for failure. + */ +static int cik_sdma_rlc_resume(struct radeon_device *rdev) +{ + /* XXX todo */ + return 0; +} + +/** + * cik_sdma_load_microcode - load the sDMA ME ucode + * + * @rdev: radeon_device pointer + * + * Loads the sDMA0/1 ucode. + * Returns 0 for success, -EINVAL if the ucode is not available. + */ +static int cik_sdma_load_microcode(struct radeon_device *rdev) +{ + const __be32 *fw_data; + int i; + + if (!rdev->sdma_fw) + return -EINVAL; + + /* halt the MEs */ + cik_sdma_enable(rdev, false); + + /* sdma0 */ + fw_data = (const __be32 *)rdev->sdma_fw->data; + WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0); + for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++) + WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, be32_to_cpup(fw_data++)); + WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION); + + /* sdma1 */ + fw_data = (const __be32 *)rdev->sdma_fw->data; + WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0); + for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++) + WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, be32_to_cpup(fw_data++)); + WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION); + + WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0); + WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0); + return 0; +} + +/** + * cik_sdma_resume - setup and start the async dma engines + * + * @rdev: radeon_device pointer + * + * Set up the DMA engines and enable them (CIK). + * Returns 0 for success, error for failure. + */ +int cik_sdma_resume(struct radeon_device *rdev) +{ + int r; + + /* Reset dma */ + WREG32(SRBM_SOFT_RESET, SOFT_RESET_SDMA | SOFT_RESET_SDMA1); + RREG32(SRBM_SOFT_RESET); + udelay(50); + WREG32(SRBM_SOFT_RESET, 0); + RREG32(SRBM_SOFT_RESET); + + r = cik_sdma_load_microcode(rdev); + if (r) + return r; + + /* unhalt the MEs */ + cik_sdma_enable(rdev, true); + + /* start the gfx rings and rlc compute queues */ + r = cik_sdma_gfx_resume(rdev); + if (r) + return r; + r = cik_sdma_rlc_resume(rdev); + if (r) + return r; + + return 0; +} + +/** + * cik_sdma_fini - tear down the async dma engines + * + * @rdev: radeon_device pointer + * + * Stop the async dma engines and free the rings (CIK). + */ +void cik_sdma_fini(struct radeon_device *rdev) +{ + /* halt the MEs */ + cik_sdma_enable(rdev, false); + radeon_ring_fini(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX]); + radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]); + /* XXX - compute dma queue tear down */ +} + +/** + * cik_copy_dma - copy pages using the DMA engine + * + * @rdev: radeon_device pointer + * @src_offset: src GPU address + * @dst_offset: dst GPU address + * @num_gpu_pages: number of GPU pages to xfer + * @fence: radeon fence object + * + * Copy GPU paging using the DMA engine (CIK). + * Used by the radeon ttm implementation to move pages if + * registered as the asic copy callback. + */ +int cik_copy_dma(struct radeon_device *rdev, + uint64_t src_offset, uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence **fence) +{ + struct radeon_semaphore *sem = NULL; + int ring_index = rdev->asic->copy.dma_ring_index; + struct radeon_ring *ring = &rdev->ring[ring_index]; + u32 size_in_bytes, cur_size_in_bytes; + int i, num_loops; + int r = 0; + + r = radeon_semaphore_create(rdev, &sem); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + return r; + } + + size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT); + num_loops = DIV_ROUND_UP(size_in_bytes, 0x1fffff); + r = radeon_ring_lock(rdev, ring, num_loops * 7 + 14); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_semaphore_sync_to(sem, *fence); + radeon_semaphore_sync_rings(rdev, sem, ring->idx); + + for (i = 0; i < num_loops; i++) { + cur_size_in_bytes = size_in_bytes; + if (cur_size_in_bytes > 0x1fffff) + cur_size_in_bytes = 0x1fffff; + size_in_bytes -= cur_size_in_bytes; + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_COPY, SDMA_COPY_SUB_OPCODE_LINEAR, 0)); + radeon_ring_write(ring, cur_size_in_bytes); + radeon_ring_write(ring, 0); /* src/dst endian swap */ + radeon_ring_write(ring, lower_32_bits(src_offset)); + radeon_ring_write(ring, upper_32_bits(src_offset)); + radeon_ring_write(ring, lower_32_bits(dst_offset)); + radeon_ring_write(ring, upper_32_bits(dst_offset)); + src_offset += cur_size_in_bytes; + dst_offset += cur_size_in_bytes; + } + + r = radeon_fence_emit(rdev, fence, ring->idx); + if (r) { + radeon_ring_unlock_undo(rdev, ring); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_ring_unlock_commit(rdev, ring); + radeon_semaphore_free(rdev, &sem, *fence); + + return r; +} + +/** + * cik_sdma_ring_test - simple async dma engine test + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * + * Test the DMA engine by writing using it to write an + * value to memory. (CIK). + * Returns 0 for success, error for failure. + */ +int cik_sdma_ring_test(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + unsigned i; + int r; + void __iomem *ptr = (void *)rdev->vram_scratch.ptr; + u32 tmp; + + if (!ptr) { + DRM_ERROR("invalid vram scratch pointer\n"); + return -EINVAL; + } + + tmp = 0xCAFEDEAD; + writel(tmp, ptr); + + r = radeon_ring_lock(rdev, ring, 5); + if (r) { + DRM_ERROR("radeon: dma failed to lock ring %d (%d).\n", ring->idx, r); + return r; + } + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0)); + radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc); + radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr)); + radeon_ring_write(ring, 1); /* number of DWs to follow */ + radeon_ring_write(ring, 0xDEADBEEF); + radeon_ring_unlock_commit(rdev, ring); + + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = readl(ptr); + if (tmp == 0xDEADBEEF) + break; + DRM_UDELAY(1); + } + + if (i < rdev->usec_timeout) { + DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); + } else { + DRM_ERROR("radeon: ring %d test failed (0x%08X)\n", + ring->idx, tmp); + r = -EINVAL; + } + return r; +} + +/** + * cik_sdma_ib_test - test an IB on the DMA engine + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * + * Test a simple IB in the DMA ring (CIK). + * Returns 0 on success, error on failure. + */ +int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) +{ + struct radeon_ib ib; + unsigned i; + int r; + void __iomem *ptr = (void *)rdev->vram_scratch.ptr; + u32 tmp = 0; + + if (!ptr) { + DRM_ERROR("invalid vram scratch pointer\n"); + return -EINVAL; + } + + tmp = 0xCAFEDEAD; + writel(tmp, ptr); + + r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256); + if (r) { + DRM_ERROR("radeon: failed to get ib (%d).\n", r); + return r; + } + + ib.ptr[0] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0); + ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc; + ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr); + ib.ptr[3] = 1; + ib.ptr[4] = 0xDEADBEEF; + ib.length_dw = 5; + + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + radeon_ib_free(rdev, &ib); + DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); + return r; + } + r = radeon_fence_wait(ib.fence, false); + if (r) { + DRM_ERROR("radeon: fence wait failed (%d).\n", r); + return r; + } + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = readl(ptr); + if (tmp == 0xDEADBEEF) + break; + DRM_UDELAY(1); + } + if (i < rdev->usec_timeout) { + DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i); + } else { + DRM_ERROR("radeon: ib test failed (0x%08X)\n", tmp); + r = -EINVAL; + } + radeon_ib_free(rdev, &ib); + return r; +} + +/** + * cik_sdma_is_lockup - Check if the DMA engine is locked up + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * + * Check if the async DMA engine is locked up (CIK). + * Returns true if the engine appears to be locked up, false if not. + */ +bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) +{ + u32 reset_mask = cik_gpu_check_soft_reset(rdev); + u32 mask; + + if (ring->idx == R600_RING_TYPE_DMA_INDEX) + mask = RADEON_RESET_DMA; + else + mask = RADEON_RESET_DMA1; + + if (!(reset_mask & mask)) { + radeon_ring_lockup_update(rdev, ring); + return false; + } + return radeon_ring_test_lockup(rdev, ring); +} + +/** + * cik_sdma_vm_set_page - update the page tables using sDMA + * + * @rdev: radeon_device pointer + * @ib: indirect buffer to fill with commands + * @pe: addr of the page entry + * @addr: dst addr to write into pe + * @count: number of page entries to update + * @incr: increase next addr by incr bytes + * @flags: access flags + * + * Update the page tables using sDMA (CIK). + */ +void cik_sdma_vm_set_page(struct radeon_device *rdev, + struct radeon_ib *ib, + uint64_t pe, + uint64_t addr, unsigned count, + uint32_t incr, uint32_t flags) +{ + uint64_t value; + unsigned ndw; + + trace_radeon_vm_set_page(pe, addr, count, incr, flags); + + if (flags == R600_PTE_GART) { + uint64_t src = rdev->gart.table_addr + (addr >> 12) * 8; + while (count) { + unsigned bytes = count * 8; + if (bytes > 0x1FFFF8) + bytes = 0x1FFFF8; + + ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_COPY, SDMA_WRITE_SUB_OPCODE_LINEAR, 0); + ib->ptr[ib->length_dw++] = bytes; + ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */ + ib->ptr[ib->length_dw++] = lower_32_bits(src); + ib->ptr[ib->length_dw++] = upper_32_bits(src); + ib->ptr[ib->length_dw++] = lower_32_bits(pe); + ib->ptr[ib->length_dw++] = upper_32_bits(pe); + + pe += bytes; + src += bytes; + count -= bytes / 8; + } + } else if (flags & R600_PTE_SYSTEM) { + while (count) { + ndw = count * 2; + if (ndw > 0xFFFFE) + ndw = 0xFFFFE; + + /* for non-physically contiguous pages (system) */ + ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0); + ib->ptr[ib->length_dw++] = pe; + ib->ptr[ib->length_dw++] = upper_32_bits(pe); + ib->ptr[ib->length_dw++] = ndw; + for (; ndw > 0; ndw -= 2, --count, pe += 8) { + value = radeon_vm_map_gart(rdev, addr); + value &= 0xFFFFFFFFFFFFF000ULL; + addr += incr; + value |= flags; + ib->ptr[ib->length_dw++] = value; + ib->ptr[ib->length_dw++] = upper_32_bits(value); + } + } + } else { + while (count) { + ndw = count; + if (ndw > 0x7FFFF) + ndw = 0x7FFFF; + + if (flags & R600_PTE_VALID) + value = addr; + else + value = 0; + /* for physically contiguous pages (vram) */ + ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_GENERATE_PTE_PDE, 0, 0); + ib->ptr[ib->length_dw++] = pe; /* dst addr */ + ib->ptr[ib->length_dw++] = upper_32_bits(pe); + ib->ptr[ib->length_dw++] = flags; /* mask */ + ib->ptr[ib->length_dw++] = 0; + ib->ptr[ib->length_dw++] = value; /* value */ + ib->ptr[ib->length_dw++] = upper_32_bits(value); + ib->ptr[ib->length_dw++] = incr; /* increment size */ + ib->ptr[ib->length_dw++] = 0; + ib->ptr[ib->length_dw++] = ndw; /* number of entries */ + pe += ndw * 8; + addr += ndw * incr; + count -= ndw; + } + } + while (ib->length_dw & 0x7) + ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0); +} + +/** + * cik_dma_vm_flush - cik vm flush using sDMA + * + * @rdev: radeon_device pointer + * + * Update the page table base and flush the VM TLB + * using sDMA (CIK). + */ +void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) +{ + struct radeon_ring *ring = &rdev->ring[ridx]; + + if (vm == NULL) + return; + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + if (vm->id < 8) { + radeon_ring_write(ring, (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2); + } else { + radeon_ring_write(ring, (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2); + } + radeon_ring_write(ring, vm->pd_gpu_addr >> 12); + + /* update SH_MEM_* regs */ + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SRBM_GFX_CNTL >> 2); + radeon_ring_write(ring, VMID(vm->id)); + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SH_MEM_BASES >> 2); + radeon_ring_write(ring, 0); + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SH_MEM_CONFIG >> 2); + radeon_ring_write(ring, 0); + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SH_MEM_APE1_BASE >> 2); + radeon_ring_write(ring, 1); + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SH_MEM_APE1_LIMIT >> 2); + radeon_ring_write(ring, 0); + + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, SRBM_GFX_CNTL >> 2); + radeon_ring_write(ring, VMID(0)); + + /* flush HDP */ + cik_sdma_hdp_flush_ring_emit(rdev, ridx); + + /* flush TLB */ + radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000)); + radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2); + radeon_ring_write(ring, 1 << vm->id); +} + diff --git a/drivers/gpu/drm/radeon/cikd.h b/drivers/gpu/drm/radeon/cikd.h index 63514b95889..0c6e1b55d96 100644 --- a/drivers/gpu/drm/radeon/cikd.h +++ b/drivers/gpu/drm/radeon/cikd.h @@ -25,24 +25,386 @@ #define CIK_H #define BONAIRE_GB_ADDR_CONFIG_GOLDEN 0x12010001 +#define HAWAII_GB_ADDR_CONFIG_GOLDEN 0x12011003 -#define CIK_RB_BITMAP_WIDTH_PER_SH 2 +#define CIK_RB_BITMAP_WIDTH_PER_SH 2 +#define HAWAII_RB_BITMAP_WIDTH_PER_SH 4 + +/* DIDT IND registers */ +#define DIDT_SQ_CTRL0 0x0 +# define DIDT_CTRL_EN (1 << 0) +#define DIDT_DB_CTRL0 0x20 +#define DIDT_TD_CTRL0 0x40 +#define DIDT_TCP_CTRL0 0x60 /* SMC IND registers */ +#define DPM_TABLE_475 0x3F768 +# define SamuBootLevel(x) ((x) << 0) +# define SamuBootLevel_MASK 0x000000ff +# define SamuBootLevel_SHIFT 0 +# define AcpBootLevel(x) ((x) << 8) +# define AcpBootLevel_MASK 0x0000ff00 +# define AcpBootLevel_SHIFT 8 +# define VceBootLevel(x) ((x) << 16) +# define VceBootLevel_MASK 0x00ff0000 +# define VceBootLevel_SHIFT 16 +# define UvdBootLevel(x) ((x) << 24) +# define UvdBootLevel_MASK 0xff000000 +# define UvdBootLevel_SHIFT 24 + +#define FIRMWARE_FLAGS 0x3F800 +# define INTERRUPTS_ENABLED (1 << 0) + +#define NB_DPM_CONFIG_1 0x3F9E8 +# define Dpm0PgNbPsLo(x) ((x) << 0) +# define Dpm0PgNbPsLo_MASK 0x000000ff +# define Dpm0PgNbPsLo_SHIFT 0 +# define Dpm0PgNbPsHi(x) ((x) << 8) +# define Dpm0PgNbPsHi_MASK 0x0000ff00 +# define Dpm0PgNbPsHi_SHIFT 8 +# define DpmXNbPsLo(x) ((x) << 16) +# define DpmXNbPsLo_MASK 0x00ff0000 +# define DpmXNbPsLo_SHIFT 16 +# define DpmXNbPsHi(x) ((x) << 24) +# define DpmXNbPsHi_MASK 0xff000000 +# define DpmXNbPsHi_SHIFT 24 + +#define SMC_SYSCON_RESET_CNTL 0x80000000 +# define RST_REG (1 << 0) +#define SMC_SYSCON_CLOCK_CNTL_0 0x80000004 +# define CK_DISABLE (1 << 0) +# define CKEN (1 << 24) + +#define SMC_SYSCON_MISC_CNTL 0x80000010 + +#define SMC_SYSCON_MSG_ARG_0 0x80000068 + +#define SMC_PC_C 0x80000370 + +#define SMC_SCRATCH9 0x80000424 + +#define RCU_UC_EVENTS 0xC0000004 +# define BOOT_SEQ_DONE (1 << 7) + #define GENERAL_PWRMGT 0xC0200000 +# define GLOBAL_PWRMGT_EN (1 << 0) +# define STATIC_PM_EN (1 << 1) +# define THERMAL_PROTECTION_DIS (1 << 2) +# define THERMAL_PROTECTION_TYPE (1 << 3) +# define SW_SMIO_INDEX(x) ((x) << 6) +# define SW_SMIO_INDEX_MASK (1 << 6) +# define SW_SMIO_INDEX_SHIFT 6 +# define VOLT_PWRMGT_EN (1 << 10) # define GPU_COUNTER_CLK (1 << 15) - +# define DYN_SPREAD_SPECTRUM_EN (1 << 23) + +#define CNB_PWRMGT_CNTL 0xC0200004 +# define GNB_SLOW_MODE(x) ((x) << 0) +# define GNB_SLOW_MODE_MASK (3 << 0) +# define GNB_SLOW_MODE_SHIFT 0 +# define GNB_SLOW (1 << 2) +# define FORCE_NB_PS1 (1 << 3) +# define DPM_ENABLED (1 << 4) + +#define SCLK_PWRMGT_CNTL 0xC0200008 +# define SCLK_PWRMGT_OFF (1 << 0) +# define RESET_BUSY_CNT (1 << 4) +# define RESET_SCLK_CNT (1 << 5) +# define DYNAMIC_PM_EN (1 << 21) + +#define TARGET_AND_CURRENT_PROFILE_INDEX 0xC0200014 +# define CURRENT_STATE_MASK (0xf << 4) +# define CURRENT_STATE_SHIFT 4 +# define CURR_MCLK_INDEX_MASK (0xf << 8) +# define CURR_MCLK_INDEX_SHIFT 8 +# define CURR_SCLK_INDEX_MASK (0x1f << 16) +# define CURR_SCLK_INDEX_SHIFT 16 + +#define CG_SSP 0xC0200044 +# define SST(x) ((x) << 0) +# define SST_MASK (0xffff << 0) +# define SSTU(x) ((x) << 16) +# define SSTU_MASK (0xf << 16) + +#define CG_DISPLAY_GAP_CNTL 0xC0200060 +# define DISP_GAP(x) ((x) << 0) +# define DISP_GAP_MASK (3 << 0) +# define VBI_TIMER_COUNT(x) ((x) << 4) +# define VBI_TIMER_COUNT_MASK (0x3fff << 4) +# define VBI_TIMER_UNIT(x) ((x) << 20) +# define VBI_TIMER_UNIT_MASK (7 << 20) +# define DISP_GAP_MCHG(x) ((x) << 24) +# define DISP_GAP_MCHG_MASK (3 << 24) + +#define SMU_VOLTAGE_STATUS 0xC0200094 +# define SMU_VOLTAGE_CURRENT_LEVEL_MASK (0xff << 1) +# define SMU_VOLTAGE_CURRENT_LEVEL_SHIFT 1 + +#define TARGET_AND_CURRENT_PROFILE_INDEX_1 0xC02000F0 +# define CURR_PCIE_INDEX_MASK (0xf << 24) +# define CURR_PCIE_INDEX_SHIFT 24 + +#define CG_ULV_PARAMETER 0xC0200158 + +#define CG_FTV_0 0xC02001A8 +#define CG_FTV_1 0xC02001AC +#define CG_FTV_2 0xC02001B0 +#define CG_FTV_3 0xC02001B4 +#define CG_FTV_4 0xC02001B8 +#define CG_FTV_5 0xC02001BC +#define CG_FTV_6 0xC02001C0 +#define CG_FTV_7 0xC02001C4 + +#define CG_DISPLAY_GAP_CNTL2 0xC0200230 + +#define LCAC_SX0_OVR_SEL 0xC0400D04 +#define LCAC_SX0_OVR_VAL 0xC0400D08 + +#define LCAC_MC0_CNTL 0xC0400D30 +#define LCAC_MC0_OVR_SEL 0xC0400D34 +#define LCAC_MC0_OVR_VAL 0xC0400D38 +#define LCAC_MC1_CNTL 0xC0400D3C +#define LCAC_MC1_OVR_SEL 0xC0400D40 +#define LCAC_MC1_OVR_VAL 0xC0400D44 + +#define LCAC_MC2_OVR_SEL 0xC0400D4C +#define LCAC_MC2_OVR_VAL 0xC0400D50 + +#define LCAC_MC3_OVR_SEL 0xC0400D58 +#define LCAC_MC3_OVR_VAL 0xC0400D5C + +#define LCAC_CPL_CNTL 0xC0400D80 +#define LCAC_CPL_OVR_SEL 0xC0400D84 +#define LCAC_CPL_OVR_VAL 0xC0400D88 + +/* dGPU */ +#define CG_THERMAL_CTRL 0xC0300004 +#define DPM_EVENT_SRC(x) ((x) << 0) +#define DPM_EVENT_SRC_MASK (7 << 0) +#define DIG_THERM_DPM(x) ((x) << 14) +#define DIG_THERM_DPM_MASK 0x003FC000 +#define DIG_THERM_DPM_SHIFT 14 + +#define CG_THERMAL_INT 0xC030000C +#define CI_DIG_THERM_INTH(x) ((x) << 8) +#define CI_DIG_THERM_INTH_MASK 0x0000FF00 +#define CI_DIG_THERM_INTH_SHIFT 8 +#define CI_DIG_THERM_INTL(x) ((x) << 16) +#define CI_DIG_THERM_INTL_MASK 0x00FF0000 +#define CI_DIG_THERM_INTL_SHIFT 16 +#define THERM_INT_MASK_HIGH (1 << 24) +#define THERM_INT_MASK_LOW (1 << 25) + +#define CG_MULT_THERMAL_STATUS 0xC0300014 +#define ASIC_MAX_TEMP(x) ((x) << 0) +#define ASIC_MAX_TEMP_MASK 0x000001ff +#define ASIC_MAX_TEMP_SHIFT 0 +#define CTF_TEMP(x) ((x) << 9) +#define CTF_TEMP_MASK 0x0003fe00 +#define CTF_TEMP_SHIFT 9 + +#define CG_ECLK_CNTL 0xC05000AC +# define ECLK_DIVIDER_MASK 0x7f +# define ECLK_DIR_CNTL_EN (1 << 8) +#define CG_ECLK_STATUS 0xC05000B0 +# define ECLK_STATUS (1 << 0) + +#define CG_SPLL_FUNC_CNTL 0xC0500140 +#define SPLL_RESET (1 << 0) +#define SPLL_PWRON (1 << 1) +#define SPLL_BYPASS_EN (1 << 3) +#define SPLL_REF_DIV(x) ((x) << 5) +#define SPLL_REF_DIV_MASK (0x3f << 5) +#define SPLL_PDIV_A(x) ((x) << 20) +#define SPLL_PDIV_A_MASK (0x7f << 20) +#define SPLL_PDIV_A_SHIFT 20 +#define CG_SPLL_FUNC_CNTL_2 0xC0500144 +#define SCLK_MUX_SEL(x) ((x) << 0) +#define SCLK_MUX_SEL_MASK (0x1ff << 0) +#define CG_SPLL_FUNC_CNTL_3 0xC0500148 +#define SPLL_FB_DIV(x) ((x) << 0) +#define SPLL_FB_DIV_MASK (0x3ffffff << 0) +#define SPLL_FB_DIV_SHIFT 0 +#define SPLL_DITHEN (1 << 28) +#define CG_SPLL_FUNC_CNTL_4 0xC050014C + +#define CG_SPLL_SPREAD_SPECTRUM 0xC0500164 +#define SSEN (1 << 0) +#define CLK_S(x) ((x) << 4) +#define CLK_S_MASK (0xfff << 4) +#define CLK_S_SHIFT 4 +#define CG_SPLL_SPREAD_SPECTRUM_2 0xC0500168 +#define CLK_V(x) ((x) << 0) +#define CLK_V_MASK (0x3ffffff << 0) +#define CLK_V_SHIFT 0 + +#define MPLL_BYPASSCLK_SEL 0xC050019C +# define MPLL_CLKOUT_SEL(x) ((x) << 8) +# define MPLL_CLKOUT_SEL_MASK 0xFF00 #define CG_CLKPIN_CNTL 0xC05001A0 # define XTALIN_DIVIDE (1 << 1) - +# define BCLK_AS_XCLK (1 << 2) +#define CG_CLKPIN_CNTL_2 0xC05001A4 +# define FORCE_BIF_REFCLK_EN (1 << 3) +# define MUX_TCLK_TO_XCLK (1 << 8) +#define THM_CLK_CNTL 0xC05001A8 +# define CMON_CLK_SEL(x) ((x) << 0) +# define CMON_CLK_SEL_MASK 0xFF +# define TMON_CLK_SEL(x) ((x) << 8) +# define TMON_CLK_SEL_MASK 0xFF00 +#define MISC_CLK_CTRL 0xC05001AC +# define DEEP_SLEEP_CLK_SEL(x) ((x) << 0) +# define DEEP_SLEEP_CLK_SEL_MASK 0xFF +# define ZCLK_SEL(x) ((x) << 8) +# define ZCLK_SEL_MASK 0xFF00 + +/* KV/KB */ +#define CG_THERMAL_INT_CTRL 0xC2100028 +#define DIG_THERM_INTH(x) ((x) << 0) +#define DIG_THERM_INTH_MASK 0x000000FF +#define DIG_THERM_INTH_SHIFT 0 +#define DIG_THERM_INTL(x) ((x) << 8) +#define DIG_THERM_INTL_MASK 0x0000FF00 +#define DIG_THERM_INTL_SHIFT 8 +#define THERM_INTH_MASK (1 << 24) +#define THERM_INTL_MASK (1 << 25) + +/* PCIE registers idx/data 0x38/0x3c */ +#define PB0_PIF_PWRDOWN_0 0x1100012 /* PCIE */ +# define PLL_POWER_STATE_IN_TXS2_0(x) ((x) << 7) +# define PLL_POWER_STATE_IN_TXS2_0_MASK (0x7 << 7) +# define PLL_POWER_STATE_IN_TXS2_0_SHIFT 7 +# define PLL_POWER_STATE_IN_OFF_0(x) ((x) << 10) +# define PLL_POWER_STATE_IN_OFF_0_MASK (0x7 << 10) +# define PLL_POWER_STATE_IN_OFF_0_SHIFT 10 +# define PLL_RAMP_UP_TIME_0(x) ((x) << 24) +# define PLL_RAMP_UP_TIME_0_MASK (0x7 << 24) +# define PLL_RAMP_UP_TIME_0_SHIFT 24 +#define PB0_PIF_PWRDOWN_1 0x1100013 /* PCIE */ +# define PLL_POWER_STATE_IN_TXS2_1(x) ((x) << 7) +# define PLL_POWER_STATE_IN_TXS2_1_MASK (0x7 << 7) +# define PLL_POWER_STATE_IN_TXS2_1_SHIFT 7 +# define PLL_POWER_STATE_IN_OFF_1(x) ((x) << 10) +# define PLL_POWER_STATE_IN_OFF_1_MASK (0x7 << 10) +# define PLL_POWER_STATE_IN_OFF_1_SHIFT 10 +# define PLL_RAMP_UP_TIME_1(x) ((x) << 24) +# define PLL_RAMP_UP_TIME_1_MASK (0x7 << 24) +# define PLL_RAMP_UP_TIME_1_SHIFT 24 + +#define PCIE_CNTL2 0x1001001c /* PCIE */ +# define SLV_MEM_LS_EN (1 << 16) +# define SLV_MEM_AGGRESSIVE_LS_EN (1 << 17) +# define MST_MEM_LS_EN (1 << 18) +# define REPLAY_MEM_LS_EN (1 << 19) + +#define PCIE_LC_STATUS1 0x1400028 /* PCIE */ +# define LC_REVERSE_RCVR (1 << 0) +# define LC_REVERSE_XMIT (1 << 1) +# define LC_OPERATING_LINK_WIDTH_MASK (0x7 << 2) +# define LC_OPERATING_LINK_WIDTH_SHIFT 2 +# define LC_DETECTED_LINK_WIDTH_MASK (0x7 << 5) +# define LC_DETECTED_LINK_WIDTH_SHIFT 5 + +#define PCIE_P_CNTL 0x1400040 /* PCIE */ +# define P_IGNORE_EDB_ERR (1 << 6) + +#define PB1_PIF_PWRDOWN_0 0x2100012 /* PCIE */ +#define PB1_PIF_PWRDOWN_1 0x2100013 /* PCIE */ + +#define PCIE_LC_CNTL 0x100100A0 /* PCIE */ +# define LC_L0S_INACTIVITY(x) ((x) << 8) +# define LC_L0S_INACTIVITY_MASK (0xf << 8) +# define LC_L0S_INACTIVITY_SHIFT 8 +# define LC_L1_INACTIVITY(x) ((x) << 12) +# define LC_L1_INACTIVITY_MASK (0xf << 12) +# define LC_L1_INACTIVITY_SHIFT 12 +# define LC_PMI_TO_L1_DIS (1 << 16) +# define LC_ASPM_TO_L1_DIS (1 << 24) + +#define PCIE_LC_LINK_WIDTH_CNTL 0x100100A2 /* PCIE */ +# define LC_LINK_WIDTH_SHIFT 0 +# define LC_LINK_WIDTH_MASK 0x7 +# define LC_LINK_WIDTH_X0 0 +# define LC_LINK_WIDTH_X1 1 +# define LC_LINK_WIDTH_X2 2 +# define LC_LINK_WIDTH_X4 3 +# define LC_LINK_WIDTH_X8 4 +# define LC_LINK_WIDTH_X16 6 +# define LC_LINK_WIDTH_RD_SHIFT 4 +# define LC_LINK_WIDTH_RD_MASK 0x70 +# define LC_RECONFIG_ARC_MISSING_ESCAPE (1 << 7) +# define LC_RECONFIG_NOW (1 << 8) +# define LC_RENEGOTIATION_SUPPORT (1 << 9) +# define LC_RENEGOTIATE_EN (1 << 10) +# define LC_SHORT_RECONFIG_EN (1 << 11) +# define LC_UPCONFIGURE_SUPPORT (1 << 12) +# define LC_UPCONFIGURE_DIS (1 << 13) +# define LC_DYN_LANES_PWR_STATE(x) ((x) << 21) +# define LC_DYN_LANES_PWR_STATE_MASK (0x3 << 21) +# define LC_DYN_LANES_PWR_STATE_SHIFT 21 +#define PCIE_LC_N_FTS_CNTL 0x100100a3 /* PCIE */ +# define LC_XMIT_N_FTS(x) ((x) << 0) +# define LC_XMIT_N_FTS_MASK (0xff << 0) +# define LC_XMIT_N_FTS_SHIFT 0 +# define LC_XMIT_N_FTS_OVERRIDE_EN (1 << 8) +# define LC_N_FTS_MASK (0xff << 24) +#define PCIE_LC_SPEED_CNTL 0x100100A4 /* PCIE */ +# define LC_GEN2_EN_STRAP (1 << 0) +# define LC_GEN3_EN_STRAP (1 << 1) +# define LC_TARGET_LINK_SPEED_OVERRIDE_EN (1 << 2) +# define LC_TARGET_LINK_SPEED_OVERRIDE_MASK (0x3 << 3) +# define LC_TARGET_LINK_SPEED_OVERRIDE_SHIFT 3 +# define LC_FORCE_EN_SW_SPEED_CHANGE (1 << 5) +# define LC_FORCE_DIS_SW_SPEED_CHANGE (1 << 6) +# define LC_FORCE_EN_HW_SPEED_CHANGE (1 << 7) +# define LC_FORCE_DIS_HW_SPEED_CHANGE (1 << 8) +# define LC_INITIATE_LINK_SPEED_CHANGE (1 << 9) +# define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK (0x3 << 10) +# define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT 10 +# define LC_CURRENT_DATA_RATE_MASK (0x3 << 13) /* 0/1/2 = gen1/2/3 */ +# define LC_CURRENT_DATA_RATE_SHIFT 13 +# define LC_CLR_FAILED_SPD_CHANGE_CNT (1 << 16) +# define LC_OTHER_SIDE_EVER_SENT_GEN2 (1 << 18) +# define LC_OTHER_SIDE_SUPPORTS_GEN2 (1 << 19) +# define LC_OTHER_SIDE_EVER_SENT_GEN3 (1 << 20) +# define LC_OTHER_SIDE_SUPPORTS_GEN3 (1 << 21) + +#define PCIE_LC_CNTL2 0x100100B1 /* PCIE */ +# define LC_ALLOW_PDWN_IN_L1 (1 << 17) +# define LC_ALLOW_PDWN_IN_L23 (1 << 18) + +#define PCIE_LC_CNTL3 0x100100B5 /* PCIE */ +# define LC_GO_TO_RECOVERY (1 << 30) +#define PCIE_LC_CNTL4 0x100100B6 /* PCIE */ +# define LC_REDO_EQ (1 << 5) +# define LC_SET_QUIESCE (1 << 13) + +/* direct registers */ #define PCIE_INDEX 0x38 #define PCIE_DATA 0x3C +#define SMC_IND_INDEX_0 0x200 +#define SMC_IND_DATA_0 0x204 + +#define SMC_IND_ACCESS_CNTL 0x240 +#define AUTO_INCREMENT_IND_0 (1 << 0) + +#define SMC_MESSAGE_0 0x250 +#define SMC_MSG_MASK 0xffff +#define SMC_RESP_0 0x254 +#define SMC_RESP_MASK 0xffff + +#define SMC_MSG_ARG_0 0x290 + #define VGA_HDP_CONTROL 0x328 #define VGA_MEMORY_DISABLE (1 << 4) #define DMIF_ADDR_CALC 0xC00 +#define PIPE0_DMIF_BUFFER_CONTROL 0x0ca0 +# define DMIF_BUFFERS_ALLOCATED(x) ((x) << 0) +# define DMIF_BUFFERS_ALLOCATED_COMPLETED (1 << 4) + #define SRBM_GFX_CNTL 0xE44 #define PIPEID(x) ((x) << 0) #define MEID(x) ((x) << 2) @@ -120,6 +482,7 @@ #define READ_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 16) #define WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 18) #define WRITE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 19) +#define PAGE_TABLE_BLOCK_SIZE(x) (((x) & 0xF) << 24) #define VM_CONTEXT1_CNTL 0x1414 #define VM_CONTEXT0_CNTL2 0x1430 #define VM_CONTEXT1_CNTL2 0x1434 @@ -136,6 +499,23 @@ #define VM_INVALIDATE_RESPONSE 0x147c #define VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x14DC +#define PROTECTIONS_MASK (0xf << 0) +#define PROTECTIONS_SHIFT 0 + /* bit 0: range + * bit 1: pde0 + * bit 2: valid + * bit 3: read + * bit 4: write + */ +#define MEMORY_CLIENT_ID_MASK (0xff << 12) +#define HAWAII_MEMORY_CLIENT_ID_MASK (0x1ff << 12) +#define MEMORY_CLIENT_ID_SHIFT 12 +#define MEMORY_CLIENT_RW_MASK (1 << 24) +#define MEMORY_CLIENT_RW_SHIFT 24 +#define FAULT_VMID_MASK (0xf << 25) +#define FAULT_VMID_SHIFT 25 + +#define VM_CONTEXT1_PROTECTION_FAULT_MCCLIENT 0x14E4 #define VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x14FC @@ -156,6 +536,10 @@ #define VM_CONTEXT0_PAGE_TABLE_END_ADDR 0x157C #define VM_CONTEXT1_PAGE_TABLE_END_ADDR 0x1580 +#define VM_L2_CG 0x15c0 +#define MC_CG_ENABLE (1 << 18) +#define MC_LS_ENABLE (1 << 19) + #define MC_SHARED_CHMAP 0x2004 #define NOOFCHAN_SHIFT 12 #define NOOFCHAN_MASK 0x0000f000 @@ -185,6 +569,17 @@ #define MC_SHARED_BLACKOUT_CNTL 0x20ac +#define MC_HUB_MISC_HUB_CG 0x20b8 +#define MC_HUB_MISC_VM_CG 0x20bc + +#define MC_HUB_MISC_SIP_CG 0x20c0 + +#define MC_XPB_CLK_GAT 0x2478 + +#define MC_CITF_MISC_RD_CG 0x2648 +#define MC_CITF_MISC_WR_CG 0x264c +#define MC_CITF_MISC_VM_CG 0x2650 + #define MC_ARB_RAMCFG 0x2760 #define NOOFBANK_SHIFT 0 #define NOOFBANK_MASK 0x00000003 @@ -199,9 +594,37 @@ #define NOOFGROUPS_SHIFT 12 #define NOOFGROUPS_MASK 0x00001000 +#define MC_ARB_DRAM_TIMING 0x2774 +#define MC_ARB_DRAM_TIMING2 0x2778 + +#define MC_ARB_BURST_TIME 0x2808 +#define STATE0(x) ((x) << 0) +#define STATE0_MASK (0x1f << 0) +#define STATE0_SHIFT 0 +#define STATE1(x) ((x) << 5) +#define STATE1_MASK (0x1f << 5) +#define STATE1_SHIFT 5 +#define STATE2(x) ((x) << 10) +#define STATE2_MASK (0x1f << 10) +#define STATE2_SHIFT 10 +#define STATE3(x) ((x) << 15) +#define STATE3_MASK (0x1f << 15) +#define STATE3_SHIFT 15 + +#define MC_SEQ_RAS_TIMING 0x28a0 +#define MC_SEQ_CAS_TIMING 0x28a4 +#define MC_SEQ_MISC_TIMING 0x28a8 +#define MC_SEQ_MISC_TIMING2 0x28ac +#define MC_SEQ_PMG_TIMING 0x28b0 +#define MC_SEQ_RD_CTL_D0 0x28b4 +#define MC_SEQ_RD_CTL_D1 0x28b8 +#define MC_SEQ_WR_CTL_D0 0x28bc +#define MC_SEQ_WR_CTL_D1 0x28c0 + #define MC_SEQ_SUP_CNTL 0x28c8 #define RUN_MASK (1 << 0) #define MC_SEQ_SUP_PGM 0x28cc +#define MC_PMG_AUTO_CMD 0x28d0 #define MC_SEQ_TRAIN_WAKEUP_CNTL 0x28e8 #define TRAIN_DONE_D0 (1 << 30) @@ -210,10 +633,92 @@ #define MC_IO_PAD_CNTL_D0 0x29d0 #define MEM_FALL_OUT_CMD (1 << 8) +#define MC_SEQ_MISC0 0x2a00 +#define MC_SEQ_MISC0_VEN_ID_SHIFT 8 +#define MC_SEQ_MISC0_VEN_ID_MASK 0x00000f00 +#define MC_SEQ_MISC0_VEN_ID_VALUE 3 +#define MC_SEQ_MISC0_REV_ID_SHIFT 12 +#define MC_SEQ_MISC0_REV_ID_MASK 0x0000f000 +#define MC_SEQ_MISC0_REV_ID_VALUE 1 +#define MC_SEQ_MISC0_GDDR5_SHIFT 28 +#define MC_SEQ_MISC0_GDDR5_MASK 0xf0000000 +#define MC_SEQ_MISC0_GDDR5_VALUE 5 +#define MC_SEQ_MISC1 0x2a04 +#define MC_SEQ_RESERVE_M 0x2a08 +#define MC_PMG_CMD_EMRS 0x2a0c + #define MC_SEQ_IO_DEBUG_INDEX 0x2a44 #define MC_SEQ_IO_DEBUG_DATA 0x2a48 +#define MC_SEQ_MISC5 0x2a54 +#define MC_SEQ_MISC6 0x2a58 + +#define MC_SEQ_MISC7 0x2a64 + +#define MC_SEQ_RAS_TIMING_LP 0x2a6c +#define MC_SEQ_CAS_TIMING_LP 0x2a70 +#define MC_SEQ_MISC_TIMING_LP 0x2a74 +#define MC_SEQ_MISC_TIMING2_LP 0x2a78 +#define MC_SEQ_WR_CTL_D0_LP 0x2a7c +#define MC_SEQ_WR_CTL_D1_LP 0x2a80 +#define MC_SEQ_PMG_CMD_EMRS_LP 0x2a84 +#define MC_SEQ_PMG_CMD_MRS_LP 0x2a88 + +#define MC_PMG_CMD_MRS 0x2aac + +#define MC_SEQ_RD_CTL_D0_LP 0x2b1c +#define MC_SEQ_RD_CTL_D1_LP 0x2b20 + +#define MC_PMG_CMD_MRS1 0x2b44 +#define MC_SEQ_PMG_CMD_MRS1_LP 0x2b48 +#define MC_SEQ_PMG_TIMING_LP 0x2b4c + +#define MC_SEQ_WR_CTL_2 0x2b54 +#define MC_SEQ_WR_CTL_2_LP 0x2b58 +#define MC_PMG_CMD_MRS2 0x2b5c +#define MC_SEQ_PMG_CMD_MRS2_LP 0x2b60 + +#define MCLK_PWRMGT_CNTL 0x2ba0 +# define DLL_SPEED(x) ((x) << 0) +# define DLL_SPEED_MASK (0x1f << 0) +# define DLL_READY (1 << 6) +# define MC_INT_CNTL (1 << 7) +# define MRDCK0_PDNB (1 << 8) +# define MRDCK1_PDNB (1 << 9) +# define MRDCK0_RESET (1 << 16) +# define MRDCK1_RESET (1 << 17) +# define DLL_READY_READ (1 << 24) +#define DLL_CNTL 0x2ba4 +# define MRDCK0_BYPASS (1 << 24) +# define MRDCK1_BYPASS (1 << 25) + +#define MPLL_FUNC_CNTL 0x2bb4 +#define BWCTRL(x) ((x) << 20) +#define BWCTRL_MASK (0xff << 20) +#define MPLL_FUNC_CNTL_1 0x2bb8 +#define VCO_MODE(x) ((x) << 0) +#define VCO_MODE_MASK (3 << 0) +#define CLKFRAC(x) ((x) << 4) +#define CLKFRAC_MASK (0xfff << 4) +#define CLKF(x) ((x) << 16) +#define CLKF_MASK (0xfff << 16) +#define MPLL_FUNC_CNTL_2 0x2bbc +#define MPLL_AD_FUNC_CNTL 0x2bc0 +#define YCLK_POST_DIV(x) ((x) << 0) +#define YCLK_POST_DIV_MASK (7 << 0) +#define MPLL_DQ_FUNC_CNTL 0x2bc4 +#define YCLK_SEL(x) ((x) << 4) +#define YCLK_SEL_MASK (1 << 4) + +#define MPLL_SS1 0x2bcc +#define CLKV(x) ((x) << 0) +#define CLKV_MASK (0x3ffffff << 0) +#define MPLL_SS2 0x2bd0 +#define CLKS(x) ((x) << 0) +#define CLKS_MASK (0xfff << 0) + #define HDP_HOST_PATH_CNTL 0x2C00 +#define CLOCK_GATING_DIS (1 << 23) #define HDP_NONSURFACE_BASE 0x2C04 #define HDP_NONSURFACE_INFO 0x2C08 #define HDP_NONSURFACE_SIZE 0x2C0C @@ -221,6 +726,37 @@ #define HDP_ADDR_CONFIG 0x2F48 #define HDP_MISC_CNTL 0x2F4C #define HDP_FLUSH_INVALIDATE_CACHE (1 << 0) +#define HDP_MEM_POWER_LS 0x2F50 +#define HDP_LS_ENABLE (1 << 0) + +#define ATC_MISC_CG 0x3350 + +#define GMCON_RENG_EXECUTE 0x3508 +#define RENG_EXECUTE_ON_PWR_UP (1 << 0) +#define GMCON_MISC 0x350c +#define RENG_EXECUTE_ON_REG_UPDATE (1 << 11) +#define STCTRL_STUTTER_EN (1 << 16) + +#define GMCON_PGFSM_CONFIG 0x3538 +#define GMCON_PGFSM_WRITE 0x353c +#define GMCON_PGFSM_READ 0x3540 +#define GMCON_MISC3 0x3544 + +#define MC_SEQ_CNTL_3 0x3600 +# define CAC_EN (1 << 31) +#define MC_SEQ_G5PDX_CTRL 0x3604 +#define MC_SEQ_G5PDX_CTRL_LP 0x3608 +#define MC_SEQ_G5PDX_CMD0 0x360c +#define MC_SEQ_G5PDX_CMD0_LP 0x3610 +#define MC_SEQ_G5PDX_CMD1 0x3614 +#define MC_SEQ_G5PDX_CMD1_LP 0x3618 + +#define MC_SEQ_PMG_DVS_CTL 0x3628 +#define MC_SEQ_PMG_DVS_CTL_LP 0x362c +#define MC_SEQ_PMG_DVS_CMD 0x3630 +#define MC_SEQ_PMG_DVS_CMD_LP 0x3634 +#define MC_SEQ_DLL_STBY 0x3638 +#define MC_SEQ_DLL_STBY_LP 0x363c #define IH_RB_CNTL 0x3e00 # define IH_RB_ENABLE (1 << 0) @@ -249,6 +785,9 @@ # define MC_WR_CLEAN_CNT(x) ((x) << 20) # define MC_VMID(x) ((x) << 25) +#define BIF_LNCNT_RESET 0x5220 +# define RESET_LNCNT_EN (1 << 0) + #define CONFIG_MEMSIZE 0x5428 #define INTERRUPT_CNTL 0x5468 @@ -350,6 +889,15 @@ # define DC_HPD6_RX_INTERRUPT (1 << 18) #define DISP_INTERRUPT_STATUS_CONTINUE6 0x6780 +/* 0x6858, 0x7458, 0x10058, 0x10c58, 0x11858, 0x12458 */ +#define GRPH_INT_STATUS 0x6858 +# define GRPH_PFLIP_INT_OCCURRED (1 << 0) +# define GRPH_PFLIP_INT_CLEAR (1 << 8) +/* 0x685c, 0x745c, 0x1005c, 0x10c5c, 0x1185c, 0x1245c */ +#define GRPH_INT_CONTROL 0x685c +# define GRPH_PFLIP_INT_MASK (1 << 0) +# define GRPH_PFLIP_INT_TYPE (1 << 8) + #define DAC_AUTODETECT_INT_CONTROL 0x67c8 #define DC_HPD1_INT_STATUS 0x601c @@ -385,6 +933,42 @@ # define DC_HPDx_RX_INT_TIMER(x) ((x) << 16) # define DC_HPDx_EN (1 << 28) +#define DPG_PIPE_STUTTER_CONTROL 0x6cd4 +# define STUTTER_ENABLE (1 << 0) + +/* DCE8 FMT blocks */ +#define FMT_DYNAMIC_EXP_CNTL 0x6fb4 +# define FMT_DYNAMIC_EXP_EN (1 << 0) +# define FMT_DYNAMIC_EXP_MODE (1 << 4) + /* 0 = 10bit -> 12bit, 1 = 8bit -> 12bit */ +#define FMT_CONTROL 0x6fb8 +# define FMT_PIXEL_ENCODING (1 << 16) + /* 0 = RGB 4:4:4 or YCbCr 4:4:4, 1 = YCbCr 4:2:2 */ +#define FMT_BIT_DEPTH_CONTROL 0x6fc8 +# define FMT_TRUNCATE_EN (1 << 0) +# define FMT_TRUNCATE_MODE (1 << 1) +# define FMT_TRUNCATE_DEPTH(x) ((x) << 4) /* 0 - 18bpp, 1 - 24bpp, 2 - 30bpp */ +# define FMT_SPATIAL_DITHER_EN (1 << 8) +# define FMT_SPATIAL_DITHER_MODE(x) ((x) << 9) +# define FMT_SPATIAL_DITHER_DEPTH(x) ((x) << 11) /* 0 - 18bpp, 1 - 24bpp, 2 - 30bpp */ +# define FMT_FRAME_RANDOM_ENABLE (1 << 13) +# define FMT_RGB_RANDOM_ENABLE (1 << 14) +# define FMT_HIGHPASS_RANDOM_ENABLE (1 << 15) +# define FMT_TEMPORAL_DITHER_EN (1 << 16) +# define FMT_TEMPORAL_DITHER_DEPTH(x) ((x) << 17) /* 0 - 18bpp, 1 - 24bpp, 2 - 30bpp */ +# define FMT_TEMPORAL_DITHER_OFFSET(x) ((x) << 21) +# define FMT_TEMPORAL_LEVEL (1 << 24) +# define FMT_TEMPORAL_DITHER_RESET (1 << 25) +# define FMT_25FRC_SEL(x) ((x) << 26) +# define FMT_50FRC_SEL(x) ((x) << 28) +# define FMT_75FRC_SEL(x) ((x) << 30) +#define FMT_CLAMP_CONTROL 0x6fe4 +# define FMT_CLAMP_DATA_EN (1 << 0) +# define FMT_CLAMP_COLOR_FORMAT(x) ((x) << 16) +# define FMT_CLAMP_6BPC 0 +# define FMT_CLAMP_8BPC 1 +# define FMT_CLAMP_10BPC 2 + #define GRBM_CNTL 0x8000 #define GRBM_READ_TIMEOUT(x) ((x) << 0) @@ -488,6 +1072,9 @@ #define CP_RB0_RPTR 0x8700 #define CP_RB_WPTR_DELAY 0x8704 +#define CP_RB_WPTR_POLL_CNTL 0x8708 +#define IDLE_POLL_COUNT(x) ((x) << 16) +#define IDLE_POLL_COUNT_MASK (0xffff << 16) #define CP_MEQ_THRESHOLDS 0x8764 #define MEQ1_START(x) ((x) << 0) @@ -605,6 +1192,8 @@ # define ADDR_SURF_P8_32x32_16x16 12 # define ADDR_SURF_P8_32x32_16x32 13 # define ADDR_SURF_P8_32x64_32x32 14 +# define ADDR_SURF_P16_32x32_8x16 16 +# define ADDR_SURF_P16_32x32_16x16 17 # define TILE_SPLIT(x) ((x) << 11) # define ADDR_SURF_TILE_SPLIT_64B 0 # define ADDR_SURF_TILE_SPLIT_128B 1 @@ -714,6 +1303,9 @@ # define CP_RINGID1_INT_STAT (1 << 30) # define CP_RINGID0_INT_STAT (1 << 31) +#define CP_MEM_SLP_CNTL 0xC1E4 +# define CP_MEM_LS_EN (1 << 0) + #define CP_CPF_DEBUG 0xC200 #define CP_PQ_WPTR_POLL_CNTL 0xC20C @@ -759,14 +1351,20 @@ #define RLC_MC_CNTL 0xC30C +#define RLC_MEM_SLP_CNTL 0xC318 +# define RLC_MEM_LS_EN (1 << 0) + #define RLC_LB_CNTR_MAX 0xC348 #define RLC_LB_CNTL 0xC364 +# define LOAD_BALANCE_ENABLE (1 << 0) #define RLC_LB_CNTR_INIT 0xC36C #define RLC_SAVE_AND_RESTORE_BASE 0xC374 -#define RLC_DRIVER_DMA_STATUS 0xC378 +#define RLC_DRIVER_DMA_STATUS 0xC378 /* dGPU */ +#define RLC_CP_TABLE_RESTORE 0xC378 /* APU */ +#define RLC_PG_DELAY_2 0xC37C #define RLC_GPM_UCODE_ADDR 0xC388 #define RLC_GPM_UCODE_DATA 0xC38C @@ -775,12 +1373,52 @@ #define RLC_CAPTURE_GPU_CLOCK_COUNT 0xC398 #define RLC_UCODE_CNTL 0xC39C +#define RLC_GPM_STAT 0xC400 +# define RLC_GPM_BUSY (1 << 0) +# define GFX_POWER_STATUS (1 << 1) +# define GFX_CLOCK_STATUS (1 << 2) + +#define RLC_PG_CNTL 0xC40C +# define GFX_PG_ENABLE (1 << 0) +# define GFX_PG_SRC (1 << 1) +# define DYN_PER_CU_PG_ENABLE (1 << 2) +# define STATIC_PER_CU_PG_ENABLE (1 << 3) +# define DISABLE_GDS_PG (1 << 13) +# define DISABLE_CP_PG (1 << 15) +# define SMU_CLK_SLOWDOWN_ON_PU_ENABLE (1 << 17) +# define SMU_CLK_SLOWDOWN_ON_PD_ENABLE (1 << 18) + +#define RLC_CGTT_MGCG_OVERRIDE 0xC420 #define RLC_CGCG_CGLS_CTRL 0xC424 +# define CGCG_EN (1 << 0) +# define CGLS_EN (1 << 1) + +#define RLC_PG_DELAY 0xC434 #define RLC_LB_INIT_CU_MASK 0xC43C #define RLC_LB_PARAMS 0xC444 +#define RLC_PG_AO_CU_MASK 0xC44C + +#define RLC_MAX_PG_CU 0xC450 +# define MAX_PU_CU(x) ((x) << 0) +# define MAX_PU_CU_MASK (0xff << 0) +#define RLC_AUTO_PG_CTRL 0xC454 +# define AUTO_PG_EN (1 << 0) +# define GRBM_REG_SGIT(x) ((x) << 3) +# define GRBM_REG_SGIT_MASK (0xffff << 3) + +#define RLC_SERDES_WR_CU_MASTER_MASK 0xC474 +#define RLC_SERDES_WR_NONCU_MASTER_MASK 0xC478 +#define RLC_SERDES_WR_CTRL 0xC47C +#define BPM_ADDR(x) ((x) << 0) +#define BPM_ADDR_MASK (0xff << 0) +#define CGLS_ENABLE (1 << 16) +#define CGCG_OVERRIDE_0 (1 << 20) +#define MGCG_OVERRIDE_0 (1 << 22) +#define MGCG_OVERRIDE_1 (1 << 23) + #define RLC_SERDES_CU_MASTER_BUSY 0xC484 #define RLC_SERDES_NONCU_MASTER_BUSY 0xC488 # define SE_MASTER_BUSY_MASK 0x0000ffff @@ -791,6 +1429,13 @@ #define RLC_GPM_SCRATCH_ADDR 0xC4B0 #define RLC_GPM_SCRATCH_DATA 0xC4B4 +#define RLC_GPR_REG2 0xC4E8 +#define REQ 0x00000001 +#define MESSAGE(x) ((x) << 1) +#define MESSAGE_MASK 0x0000001e +#define MSG_ENTER_RLC_SAFE_MODE 1 +#define MSG_EXIT_RLC_SAFE_MODE 0 + #define CP_HPD_EOP_BASE_ADDR 0xC904 #define CP_HPD_EOP_BASE_ADDR_HI 0xC908 #define CP_HPD_EOP_VMID 0xC90C @@ -835,11 +1480,14 @@ #define MQD_VMID(x) ((x) << 0) #define MQD_VMID_MASK (0xf << 0) +#define DB_RENDER_CONTROL 0x28000 + #define PA_SC_RASTER_CONFIG 0x28350 # define RASTER_CONFIG_RB_MAP_0 0 # define RASTER_CONFIG_RB_MAP_1 1 # define RASTER_CONFIG_RB_MAP_2 2 # define RASTER_CONFIG_RB_MAP_3 3 +#define PKR_MAP(x) ((x) << 8) #define VGT_EVENT_INITIATOR 0x28a90 # define SAMPLE_STREAMOUTSTATS1 (1 << 0) @@ -928,6 +1576,16 @@ #define CP_PERFMON_CNTL 0x36020 +#define CGTS_SM_CTRL_REG 0x3c000 +#define SM_MODE(x) ((x) << 17) +#define SM_MODE_MASK (0x7 << 17) +#define SM_MODE_ENABLE (1 << 20) +#define CGTS_OVERRIDE (1 << 21) +#define CGTS_LS_OVERRIDE (1 << 22) +#define ON_MONITOR_ADD_EN (1 << 23) +#define ON_MONITOR_ADD(x) ((x) << 24) +#define ON_MONITOR_ADD_MASK (0xff << 24) + #define CGTS_TCC_DISABLE 0x3c00c #define CGTS_USER_TCC_DISABLE 0x3c010 #define TCC_DISABLE_MASK 0xFFFF0000 @@ -1094,12 +1752,12 @@ #define EOP_TC_WB_ACTION_EN (1 << 15) /* L2 */ #define EOP_TCL1_ACTION_EN (1 << 16) #define EOP_TC_ACTION_EN (1 << 17) /* L2 */ +#define EOP_TCL2_VOLATILE (1 << 24) #define EOP_CACHE_POLICY(x) ((x) << 25) /* 0 - LRU * 1 - Stream * 2 - Bypass */ -#define EOP_TCL2_VOLATILE (1 << 27) #define DATA_SEL(x) ((x) << 29) /* 0 - discard * 1 - send low 32bit data @@ -1122,6 +1780,68 @@ # define PACKET3_PREAMBLE_BEGIN_CLEAR_STATE (2 << 28) # define PACKET3_PREAMBLE_END_CLEAR_STATE (3 << 28) #define PACKET3_DMA_DATA 0x50 +/* 1. header + * 2. CONTROL + * 3. SRC_ADDR_LO or DATA [31:0] + * 4. SRC_ADDR_HI [31:0] + * 5. DST_ADDR_LO [31:0] + * 6. DST_ADDR_HI [7:0] + * 7. COMMAND [30:21] | BYTE_COUNT [20:0] + */ +/* CONTROL */ +# define PACKET3_DMA_DATA_ENGINE(x) ((x) << 0) + /* 0 - ME + * 1 - PFP + */ +# define PACKET3_DMA_DATA_SRC_CACHE_POLICY(x) ((x) << 13) + /* 0 - LRU + * 1 - Stream + * 2 - Bypass + */ +# define PACKET3_DMA_DATA_SRC_VOLATILE (1 << 15) +# define PACKET3_DMA_DATA_DST_SEL(x) ((x) << 20) + /* 0 - DST_ADDR using DAS + * 1 - GDS + * 3 - DST_ADDR using L2 + */ +# define PACKET3_DMA_DATA_DST_CACHE_POLICY(x) ((x) << 25) + /* 0 - LRU + * 1 - Stream + * 2 - Bypass + */ +# define PACKET3_DMA_DATA_DST_VOLATILE (1 << 27) +# define PACKET3_DMA_DATA_SRC_SEL(x) ((x) << 29) + /* 0 - SRC_ADDR using SAS + * 1 - GDS + * 2 - DATA + * 3 - SRC_ADDR using L2 + */ +# define PACKET3_DMA_DATA_CP_SYNC (1 << 31) +/* COMMAND */ +# define PACKET3_DMA_DATA_DIS_WC (1 << 21) +# define PACKET3_DMA_DATA_CMD_SRC_SWAP(x) ((x) << 22) + /* 0 - none + * 1 - 8 in 16 + * 2 - 8 in 32 + * 3 - 8 in 64 + */ +# define PACKET3_DMA_DATA_CMD_DST_SWAP(x) ((x) << 24) + /* 0 - none + * 1 - 8 in 16 + * 2 - 8 in 32 + * 3 - 8 in 64 + */ +# define PACKET3_DMA_DATA_CMD_SAS (1 << 26) + /* 0 - memory + * 1 - register + */ +# define PACKET3_DMA_DATA_CMD_DAS (1 << 27) + /* 0 - memory + * 1 - register + */ +# define PACKET3_DMA_DATA_CMD_SAIC (1 << 28) +# define PACKET3_DMA_DATA_CMD_DAIC (1 << 29) +# define PACKET3_DMA_DATA_CMD_RAW_WAIT (1 << 30) #define PACKET3_AQUIRE_MEM 0x58 #define PACKET3_REWIND 0x59 #define PACKET3_LOAD_UCONFIG_REG 0x5E @@ -1160,6 +1880,8 @@ #define SDMA0_UCODE_ADDR 0xD000 #define SDMA0_UCODE_DATA 0xD004 +#define SDMA0_POWER_CNTL 0xD008 +#define SDMA0_CLK_CTRL 0xD00C #define SDMA0_CNTL 0xD010 # define TRAP_ENABLE (1 << 0) @@ -1284,6 +2006,13 @@ #define UVD_RBC_RB_RPTR 0xf690 #define UVD_RBC_RB_WPTR 0xf694 +#define UVD_CGC_CTRL 0xF4B0 +# define DCM (1 << 0) +# define CG_DT(x) ((x) << 2) +# define CG_DT_MASK (0xf << 2) +# define CLK_OD(x) ((x) << 6) +# define CLK_OD_MASK (0x1f << 6) + /* UVD clocks */ #define CG_DCLK_CNTL 0xC050009C @@ -1294,4 +2023,50 @@ #define CG_VCLK_CNTL 0xC05000A4 #define CG_VCLK_STATUS 0xC05000A8 +/* UVD CTX indirect */ +#define UVD_CGC_MEM_CTRL 0xC0 + +/* VCE */ + +#define VCE_VCPU_CACHE_OFFSET0 0x20024 +#define VCE_VCPU_CACHE_SIZE0 0x20028 +#define VCE_VCPU_CACHE_OFFSET1 0x2002c +#define VCE_VCPU_CACHE_SIZE1 0x20030 +#define VCE_VCPU_CACHE_OFFSET2 0x20034 +#define VCE_VCPU_CACHE_SIZE2 0x20038 +#define VCE_RB_RPTR2 0x20178 +#define VCE_RB_WPTR2 0x2017c +#define VCE_RB_RPTR 0x2018c +#define VCE_RB_WPTR 0x20190 +#define VCE_CLOCK_GATING_A 0x202f8 +# define CGC_CLK_GATE_DLY_TIMER_MASK (0xf << 0) +# define CGC_CLK_GATE_DLY_TIMER(x) ((x) << 0) +# define CGC_CLK_GATER_OFF_DLY_TIMER_MASK (0xff << 4) +# define CGC_CLK_GATER_OFF_DLY_TIMER(x) ((x) << 4) +# define CGC_UENC_WAIT_AWAKE (1 << 18) +#define VCE_CLOCK_GATING_B 0x202fc +#define VCE_CGTT_CLK_OVERRIDE 0x207a0 +#define VCE_UENC_CLOCK_GATING 0x207bc +# define CLOCK_ON_DELAY_MASK (0xf << 0) +# define CLOCK_ON_DELAY(x) ((x) << 0) +# define CLOCK_OFF_DELAY_MASK (0xff << 4) +# define CLOCK_OFF_DELAY(x) ((x) << 4) +#define VCE_UENC_REG_CLOCK_GATING 0x207c0 +#define VCE_SYS_INT_EN 0x21300 +# define VCE_SYS_INT_TRAP_INTERRUPT_EN (1 << 3) +#define VCE_LMI_CTRL2 0x21474 +#define VCE_LMI_CTRL 0x21498 +#define VCE_LMI_VM_CTRL 0x214a0 +#define VCE_LMI_SWAP_CNTL 0x214b4 +#define VCE_LMI_SWAP_CNTL1 0x214b8 +#define VCE_LMI_CACHE_CTRL 0x214f4 + +#define VCE_CMD_NO_OP 0x00000000 +#define VCE_CMD_END 0x00000001 +#define VCE_CMD_IB 0x00000002 +#define VCE_CMD_FENCE 0x00000003 +#define VCE_CMD_TRAP 0x00000004 +#define VCE_CMD_IB_AUTO 0x00000005 +#define VCE_CMD_SEMAPHORE 0x00000006 + #endif diff --git a/drivers/gpu/drm/radeon/clearstate_cayman.h b/drivers/gpu/drm/radeon/clearstate_cayman.h index c00339440c5..e48a14037b7 100644 --- a/drivers/gpu/drm/radeon/clearstate_cayman.h +++ b/drivers/gpu/drm/radeon/clearstate_cayman.h @@ -1050,7 +1050,7 @@ static const struct cs_extent_def SECT_CONTEXT_defs[] = {SECT_CONTEXT_def_5, 0x0000a29e, 5 }, {SECT_CONTEXT_def_6, 0x0000a2a5, 56 }, {SECT_CONTEXT_def_7, 0x0000a2de, 290 }, - { 0, 0, 0 } + { NULL, 0, 0 } }; static const u32 SECT_CLEAR_def_1[] = { @@ -1061,7 +1061,7 @@ static const u32 SECT_CLEAR_def_1[] = static const struct cs_extent_def SECT_CLEAR_defs[] = { {SECT_CLEAR_def_1, 0x0000ffc0, 3 }, - { 0, 0, 0 } + { NULL, 0, 0 } }; static const u32 SECT_CTRLCONST_def_1[] = { @@ -1071,11 +1071,11 @@ static const u32 SECT_CTRLCONST_def_1[] = static const struct cs_extent_def SECT_CTRLCONST_defs[] = { {SECT_CTRLCONST_def_1, 0x0000f3fc, 2 }, - { 0, 0, 0 } + { NULL, 0, 0 } }; -struct cs_section_def cayman_cs_data[] = { +static const struct cs_section_def cayman_cs_data[] = { { SECT_CONTEXT_defs, SECT_CONTEXT }, { SECT_CLEAR_defs, SECT_CLEAR }, { SECT_CTRLCONST_defs, SECT_CTRLCONST }, - { 0, SECT_NONE } + { NULL, SECT_NONE } }; diff --git a/drivers/gpu/drm/radeon/clearstate_ci.h b/drivers/gpu/drm/radeon/clearstate_ci.h new file mode 100644 index 00000000000..f55d06664e3 --- /dev/null +++ b/drivers/gpu/drm/radeon/clearstate_ci.h @@ -0,0 +1,944 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +static const unsigned int ci_SECT_CONTEXT_def_1[] = +{ + 0x00000000, // DB_RENDER_CONTROL + 0x00000000, // DB_COUNT_CONTROL + 0x00000000, // DB_DEPTH_VIEW + 0x00000000, // DB_RENDER_OVERRIDE + 0x00000000, // DB_RENDER_OVERRIDE2 + 0x00000000, // DB_HTILE_DATA_BASE + 0, // HOLE + 0, // HOLE + 0x00000000, // DB_DEPTH_BOUNDS_MIN + 0x00000000, // DB_DEPTH_BOUNDS_MAX + 0x00000000, // DB_STENCIL_CLEAR + 0x00000000, // DB_DEPTH_CLEAR + 0x00000000, // PA_SC_SCREEN_SCISSOR_TL + 0x40004000, // PA_SC_SCREEN_SCISSOR_BR + 0, // HOLE + 0x00000000, // DB_DEPTH_INFO + 0x00000000, // DB_Z_INFO + 0x00000000, // DB_STENCIL_INFO + 0x00000000, // DB_Z_READ_BASE + 0x00000000, // DB_STENCIL_READ_BASE + 0x00000000, // DB_Z_WRITE_BASE + 0x00000000, // DB_STENCIL_WRITE_BASE + 0x00000000, // DB_DEPTH_SIZE + 0x00000000, // DB_DEPTH_SLICE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0x00000000, // TA_BC_BASE_ADDR + 0x00000000, // TA_BC_BASE_ADDR_HI + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0x00000000, // COHER_DEST_BASE_HI_0 + 0x00000000, // COHER_DEST_BASE_HI_1 + 0x00000000, // COHER_DEST_BASE_HI_2 + 0x00000000, // COHER_DEST_BASE_HI_3 + 0x00000000, // COHER_DEST_BASE_2 + 0x00000000, // COHER_DEST_BASE_3 + 0x00000000, // PA_SC_WINDOW_OFFSET + 0x80000000, // PA_SC_WINDOW_SCISSOR_TL + 0x40004000, // PA_SC_WINDOW_SCISSOR_BR + 0x0000ffff, // PA_SC_CLIPRECT_RULE + 0x00000000, // PA_SC_CLIPRECT_0_TL + 0x40004000, // PA_SC_CLIPRECT_0_BR + 0x00000000, // PA_SC_CLIPRECT_1_TL + 0x40004000, // PA_SC_CLIPRECT_1_BR + 0x00000000, // PA_SC_CLIPRECT_2_TL + 0x40004000, // PA_SC_CLIPRECT_2_BR + 0x00000000, // PA_SC_CLIPRECT_3_TL + 0x40004000, // PA_SC_CLIPRECT_3_BR + 0xaa99aaaa, // PA_SC_EDGERULE + 0x00000000, // PA_SU_HARDWARE_SCREEN_OFFSET + 0xffffffff, // CB_TARGET_MASK + 0xffffffff, // CB_SHADER_MASK + 0x80000000, // PA_SC_GENERIC_SCISSOR_TL + 0x40004000, // PA_SC_GENERIC_SCISSOR_BR + 0x00000000, // COHER_DEST_BASE_0 + 0x00000000, // COHER_DEST_BASE_1 + 0x80000000, // PA_SC_VPORT_SCISSOR_0_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_0_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_1_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_1_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_2_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_2_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_3_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_3_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_4_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_4_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_5_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_5_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_6_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_6_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_7_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_7_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_8_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_8_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_9_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_9_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_10_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_10_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_11_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_11_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_12_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_12_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_13_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_13_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_14_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_14_BR + 0x80000000, // PA_SC_VPORT_SCISSOR_15_TL + 0x40004000, // PA_SC_VPORT_SCISSOR_15_BR + 0x00000000, // PA_SC_VPORT_ZMIN_0 + 0x3f800000, // PA_SC_VPORT_ZMAX_0 + 0x00000000, // PA_SC_VPORT_ZMIN_1 + 0x3f800000, // PA_SC_VPORT_ZMAX_1 + 0x00000000, // PA_SC_VPORT_ZMIN_2 + 0x3f800000, // PA_SC_VPORT_ZMAX_2 + 0x00000000, // PA_SC_VPORT_ZMIN_3 + 0x3f800000, // PA_SC_VPORT_ZMAX_3 + 0x00000000, // PA_SC_VPORT_ZMIN_4 + 0x3f800000, // PA_SC_VPORT_ZMAX_4 + 0x00000000, // PA_SC_VPORT_ZMIN_5 + 0x3f800000, // PA_SC_VPORT_ZMAX_5 + 0x00000000, // PA_SC_VPORT_ZMIN_6 + 0x3f800000, // PA_SC_VPORT_ZMAX_6 + 0x00000000, // PA_SC_VPORT_ZMIN_7 + 0x3f800000, // PA_SC_VPORT_ZMAX_7 + 0x00000000, // PA_SC_VPORT_ZMIN_8 + 0x3f800000, // PA_SC_VPORT_ZMAX_8 + 0x00000000, // PA_SC_VPORT_ZMIN_9 + 0x3f800000, // PA_SC_VPORT_ZMAX_9 + 0x00000000, // PA_SC_VPORT_ZMIN_10 + 0x3f800000, // PA_SC_VPORT_ZMAX_10 + 0x00000000, // PA_SC_VPORT_ZMIN_11 + 0x3f800000, // PA_SC_VPORT_ZMAX_11 + 0x00000000, // PA_SC_VPORT_ZMIN_12 + 0x3f800000, // PA_SC_VPORT_ZMAX_12 + 0x00000000, // PA_SC_VPORT_ZMIN_13 + 0x3f800000, // PA_SC_VPORT_ZMAX_13 + 0x00000000, // PA_SC_VPORT_ZMIN_14 + 0x3f800000, // PA_SC_VPORT_ZMAX_14 + 0x00000000, // PA_SC_VPORT_ZMIN_15 + 0x3f800000, // PA_SC_VPORT_ZMAX_15 +}; +static const unsigned int ci_SECT_CONTEXT_def_2[] = +{ + 0x00000000, // PA_SC_SCREEN_EXTENT_CONTROL + 0, // HOLE + 0x00000000, // CP_PERFMON_CNTX_CNTL + 0x00000000, // CP_RINGID + 0x00000000, // CP_VMID + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0xffffffff, // VGT_MAX_VTX_INDX + 0x00000000, // VGT_MIN_VTX_INDX + 0x00000000, // VGT_INDX_OFFSET + 0x00000000, // VGT_MULTI_PRIM_IB_RESET_INDX + 0, // HOLE + 0x00000000, // CB_BLEND_RED + 0x00000000, // CB_BLEND_GREEN + 0x00000000, // CB_BLEND_BLUE + 0x00000000, // CB_BLEND_ALPHA + 0, // HOLE + 0, // HOLE + 0x00000000, // DB_STENCIL_CONTROL + 0x00000000, // DB_STENCILREFMASK + 0x00000000, // DB_STENCILREFMASK_BF + 0, // HOLE + 0x00000000, // PA_CL_VPORT_XSCALE + 0x00000000, // PA_CL_VPORT_XOFFSET + 0x00000000, // PA_CL_VPORT_YSCALE + 0x00000000, // PA_CL_VPORT_YOFFSET + 0x00000000, // PA_CL_VPORT_ZSCALE + 0x00000000, // PA_CL_VPORT_ZOFFSET + 0x00000000, // PA_CL_VPORT_XSCALE_1 + 0x00000000, // PA_CL_VPORT_XOFFSET_1 + 0x00000000, // PA_CL_VPORT_YSCALE_1 + 0x00000000, // PA_CL_VPORT_YOFFSET_1 + 0x00000000, // PA_CL_VPORT_ZSCALE_1 + 0x00000000, // PA_CL_VPORT_ZOFFSET_1 + 0x00000000, // PA_CL_VPORT_XSCALE_2 + 0x00000000, // PA_CL_VPORT_XOFFSET_2 + 0x00000000, // PA_CL_VPORT_YSCALE_2 + 0x00000000, // PA_CL_VPORT_YOFFSET_2 + 0x00000000, // PA_CL_VPORT_ZSCALE_2 + 0x00000000, // PA_CL_VPORT_ZOFFSET_2 + 0x00000000, // PA_CL_VPORT_XSCALE_3 + 0x00000000, // PA_CL_VPORT_XOFFSET_3 + 0x00000000, // PA_CL_VPORT_YSCALE_3 + 0x00000000, // PA_CL_VPORT_YOFFSET_3 + 0x00000000, // PA_CL_VPORT_ZSCALE_3 + 0x00000000, // PA_CL_VPORT_ZOFFSET_3 + 0x00000000, // PA_CL_VPORT_XSCALE_4 + 0x00000000, // PA_CL_VPORT_XOFFSET_4 + 0x00000000, // PA_CL_VPORT_YSCALE_4 + 0x00000000, // PA_CL_VPORT_YOFFSET_4 + 0x00000000, // PA_CL_VPORT_ZSCALE_4 + 0x00000000, // PA_CL_VPORT_ZOFFSET_4 + 0x00000000, // PA_CL_VPORT_XSCALE_5 + 0x00000000, // PA_CL_VPORT_XOFFSET_5 + 0x00000000, // PA_CL_VPORT_YSCALE_5 + 0x00000000, // PA_CL_VPORT_YOFFSET_5 + 0x00000000, // PA_CL_VPORT_ZSCALE_5 + 0x00000000, // PA_CL_VPORT_ZOFFSET_5 + 0x00000000, // PA_CL_VPORT_XSCALE_6 + 0x00000000, // PA_CL_VPORT_XOFFSET_6 + 0x00000000, // PA_CL_VPORT_YSCALE_6 + 0x00000000, // PA_CL_VPORT_YOFFSET_6 + 0x00000000, // PA_CL_VPORT_ZSCALE_6 + 0x00000000, // PA_CL_VPORT_ZOFFSET_6 + 0x00000000, // PA_CL_VPORT_XSCALE_7 + 0x00000000, // PA_CL_VPORT_XOFFSET_7 + 0x00000000, // PA_CL_VPORT_YSCALE_7 + 0x00000000, // PA_CL_VPORT_YOFFSET_7 + 0x00000000, // PA_CL_VPORT_ZSCALE_7 + 0x00000000, // PA_CL_VPORT_ZOFFSET_7 + 0x00000000, // PA_CL_VPORT_XSCALE_8 + 0x00000000, // PA_CL_VPORT_XOFFSET_8 + 0x00000000, // PA_CL_VPORT_YSCALE_8 + 0x00000000, // PA_CL_VPORT_YOFFSET_8 + 0x00000000, // PA_CL_VPORT_ZSCALE_8 + 0x00000000, // PA_CL_VPORT_ZOFFSET_8 + 0x00000000, // PA_CL_VPORT_XSCALE_9 + 0x00000000, // PA_CL_VPORT_XOFFSET_9 + 0x00000000, // PA_CL_VPORT_YSCALE_9 + 0x00000000, // PA_CL_VPORT_YOFFSET_9 + 0x00000000, // PA_CL_VPORT_ZSCALE_9 + 0x00000000, // PA_CL_VPORT_ZOFFSET_9 + 0x00000000, // PA_CL_VPORT_XSCALE_10 + 0x00000000, // PA_CL_VPORT_XOFFSET_10 + 0x00000000, // PA_CL_VPORT_YSCALE_10 + 0x00000000, // PA_CL_VPORT_YOFFSET_10 + 0x00000000, // PA_CL_VPORT_ZSCALE_10 + 0x00000000, // PA_CL_VPORT_ZOFFSET_10 + 0x00000000, // PA_CL_VPORT_XSCALE_11 + 0x00000000, // PA_CL_VPORT_XOFFSET_11 + 0x00000000, // PA_CL_VPORT_YSCALE_11 + 0x00000000, // PA_CL_VPORT_YOFFSET_11 + 0x00000000, // PA_CL_VPORT_ZSCALE_11 + 0x00000000, // PA_CL_VPORT_ZOFFSET_11 + 0x00000000, // PA_CL_VPORT_XSCALE_12 + 0x00000000, // PA_CL_VPORT_XOFFSET_12 + 0x00000000, // PA_CL_VPORT_YSCALE_12 + 0x00000000, // PA_CL_VPORT_YOFFSET_12 + 0x00000000, // PA_CL_VPORT_ZSCALE_12 + 0x00000000, // PA_CL_VPORT_ZOFFSET_12 + 0x00000000, // PA_CL_VPORT_XSCALE_13 + 0x00000000, // PA_CL_VPORT_XOFFSET_13 + 0x00000000, // PA_CL_VPORT_YSCALE_13 + 0x00000000, // PA_CL_VPORT_YOFFSET_13 + 0x00000000, // PA_CL_VPORT_ZSCALE_13 + 0x00000000, // PA_CL_VPORT_ZOFFSET_13 + 0x00000000, // PA_CL_VPORT_XSCALE_14 + 0x00000000, // PA_CL_VPORT_XOFFSET_14 + 0x00000000, // PA_CL_VPORT_YSCALE_14 + 0x00000000, // PA_CL_VPORT_YOFFSET_14 + 0x00000000, // PA_CL_VPORT_ZSCALE_14 + 0x00000000, // PA_CL_VPORT_ZOFFSET_14 + 0x00000000, // PA_CL_VPORT_XSCALE_15 + 0x00000000, // PA_CL_VPORT_XOFFSET_15 + 0x00000000, // PA_CL_VPORT_YSCALE_15 + 0x00000000, // PA_CL_VPORT_YOFFSET_15 + 0x00000000, // PA_CL_VPORT_ZSCALE_15 + 0x00000000, // PA_CL_VPORT_ZOFFSET_15 + 0x00000000, // PA_CL_UCP_0_X + 0x00000000, // PA_CL_UCP_0_Y + 0x00000000, // PA_CL_UCP_0_Z + 0x00000000, // PA_CL_UCP_0_W + 0x00000000, // PA_CL_UCP_1_X + 0x00000000, // PA_CL_UCP_1_Y + 0x00000000, // PA_CL_UCP_1_Z + 0x00000000, // PA_CL_UCP_1_W + 0x00000000, // PA_CL_UCP_2_X + 0x00000000, // PA_CL_UCP_2_Y + 0x00000000, // PA_CL_UCP_2_Z + 0x00000000, // PA_CL_UCP_2_W + 0x00000000, // PA_CL_UCP_3_X + 0x00000000, // PA_CL_UCP_3_Y + 0x00000000, // PA_CL_UCP_3_Z + 0x00000000, // PA_CL_UCP_3_W + 0x00000000, // PA_CL_UCP_4_X + 0x00000000, // PA_CL_UCP_4_Y + 0x00000000, // PA_CL_UCP_4_Z + 0x00000000, // PA_CL_UCP_4_W + 0x00000000, // PA_CL_UCP_5_X + 0x00000000, // PA_CL_UCP_5_Y + 0x00000000, // PA_CL_UCP_5_Z + 0x00000000, // PA_CL_UCP_5_W + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0x00000000, // SPI_PS_INPUT_CNTL_0 + 0x00000000, // SPI_PS_INPUT_CNTL_1 + 0x00000000, // SPI_PS_INPUT_CNTL_2 + 0x00000000, // SPI_PS_INPUT_CNTL_3 + 0x00000000, // SPI_PS_INPUT_CNTL_4 + 0x00000000, // SPI_PS_INPUT_CNTL_5 + 0x00000000, // SPI_PS_INPUT_CNTL_6 + 0x00000000, // SPI_PS_INPUT_CNTL_7 + 0x00000000, // SPI_PS_INPUT_CNTL_8 + 0x00000000, // SPI_PS_INPUT_CNTL_9 + 0x00000000, // SPI_PS_INPUT_CNTL_10 + 0x00000000, // SPI_PS_INPUT_CNTL_11 + 0x00000000, // SPI_PS_INPUT_CNTL_12 + 0x00000000, // SPI_PS_INPUT_CNTL_13 + 0x00000000, // SPI_PS_INPUT_CNTL_14 + 0x00000000, // SPI_PS_INPUT_CNTL_15 + 0x00000000, // SPI_PS_INPUT_CNTL_16 + 0x00000000, // SPI_PS_INPUT_CNTL_17 + 0x00000000, // SPI_PS_INPUT_CNTL_18 + 0x00000000, // SPI_PS_INPUT_CNTL_19 + 0x00000000, // SPI_PS_INPUT_CNTL_20 + 0x00000000, // SPI_PS_INPUT_CNTL_21 + 0x00000000, // SPI_PS_INPUT_CNTL_22 + 0x00000000, // SPI_PS_INPUT_CNTL_23 + 0x00000000, // SPI_PS_INPUT_CNTL_24 + 0x00000000, // SPI_PS_INPUT_CNTL_25 + 0x00000000, // SPI_PS_INPUT_CNTL_26 + 0x00000000, // SPI_PS_INPUT_CNTL_27 + 0x00000000, // SPI_PS_INPUT_CNTL_28 + 0x00000000, // SPI_PS_INPUT_CNTL_29 + 0x00000000, // SPI_PS_INPUT_CNTL_30 + 0x00000000, // SPI_PS_INPUT_CNTL_31 + 0x00000000, // SPI_VS_OUT_CONFIG + 0, // HOLE + 0x00000000, // SPI_PS_INPUT_ENA + 0x00000000, // SPI_PS_INPUT_ADDR + 0x00000000, // SPI_INTERP_CONTROL_0 + 0x00000002, // SPI_PS_IN_CONTROL + 0, // HOLE + 0x00000000, // SPI_BARYC_CNTL + 0, // HOLE + 0x00000000, // SPI_TMPRING_SIZE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0x00000000, // SPI_SHADER_POS_FORMAT + 0x00000000, // SPI_SHADER_Z_FORMAT + 0x00000000, // SPI_SHADER_COL_FORMAT + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0x00000000, // CB_BLEND0_CONTROL + 0x00000000, // CB_BLEND1_CONTROL + 0x00000000, // CB_BLEND2_CONTROL + 0x00000000, // CB_BLEND3_CONTROL + 0x00000000, // CB_BLEND4_CONTROL + 0x00000000, // CB_BLEND5_CONTROL + 0x00000000, // CB_BLEND6_CONTROL + 0x00000000, // CB_BLEND7_CONTROL +}; +static const unsigned int ci_SECT_CONTEXT_def_3[] = +{ + 0x00000000, // PA_CL_POINT_X_RAD + 0x00000000, // PA_CL_POINT_Y_RAD + 0x00000000, // PA_CL_POINT_SIZE + 0x00000000, // PA_CL_POINT_CULL_RAD + 0x00000000, // VGT_DMA_BASE_HI + 0x00000000, // VGT_DMA_BASE +}; +static const unsigned int ci_SECT_CONTEXT_def_4[] = +{ + 0x00000000, // DB_DEPTH_CONTROL + 0x00000000, // DB_EQAA + 0x00000000, // CB_COLOR_CONTROL + 0x00000000, // DB_SHADER_CONTROL + 0x00090000, // PA_CL_CLIP_CNTL + 0x00000004, // PA_SU_SC_MODE_CNTL + 0x00000000, // PA_CL_VTE_CNTL + 0x00000000, // PA_CL_VS_OUT_CNTL + 0x00000000, // PA_CL_NANINF_CNTL + 0x00000000, // PA_SU_LINE_STIPPLE_CNTL + 0x00000000, // PA_SU_LINE_STIPPLE_SCALE + 0x00000000, // PA_SU_PRIM_FILTER_CNTL + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0x00000000, // PA_SU_POINT_SIZE + 0x00000000, // PA_SU_POINT_MINMAX + 0x00000000, // PA_SU_LINE_CNTL + 0x00000000, // PA_SC_LINE_STIPPLE + 0x00000000, // VGT_OUTPUT_PATH_CNTL + 0x00000000, // VGT_HOS_CNTL + 0x00000000, // VGT_HOS_MAX_TESS_LEVEL + 0x00000000, // VGT_HOS_MIN_TESS_LEVEL + 0x00000000, // VGT_HOS_REUSE_DEPTH + 0x00000000, // VGT_GROUP_PRIM_TYPE + 0x00000000, // VGT_GROUP_FIRST_DECR + 0x00000000, // VGT_GROUP_DECR + 0x00000000, // VGT_GROUP_VECT_0_CNTL + 0x00000000, // VGT_GROUP_VECT_1_CNTL + 0x00000000, // VGT_GROUP_VECT_0_FMT_CNTL + 0x00000000, // VGT_GROUP_VECT_1_FMT_CNTL + 0x00000000, // VGT_GS_MODE + 0x00000000, // VGT_GS_ONCHIP_CNTL + 0x00000000, // PA_SC_MODE_CNTL_0 + 0x00000000, // PA_SC_MODE_CNTL_1 + 0x00000000, // VGT_ENHANCE + 0x00000100, // VGT_GS_PER_ES + 0x00000080, // VGT_ES_PER_GS + 0x00000002, // VGT_GS_PER_VS + 0x00000000, // VGT_GSVS_RING_OFFSET_1 + 0x00000000, // VGT_GSVS_RING_OFFSET_2 + 0x00000000, // VGT_GSVS_RING_OFFSET_3 + 0x00000000, // VGT_GS_OUT_PRIM_TYPE + 0x00000000, // IA_ENHANCE +}; +static const unsigned int ci_SECT_CONTEXT_def_5[] = +{ + 0x00000000, // WD_ENHANCE + 0x00000000, // VGT_PRIMITIVEID_EN +}; +static const unsigned int ci_SECT_CONTEXT_def_6[] = +{ + 0x00000000, // VGT_PRIMITIVEID_RESET +}; +static const unsigned int ci_SECT_CONTEXT_def_7[] = +{ + 0x00000000, // VGT_MULTI_PRIM_IB_RESET_EN + 0, // HOLE + 0, // HOLE + 0x00000000, // VGT_INSTANCE_STEP_RATE_0 + 0x00000000, // VGT_INSTANCE_STEP_RATE_1 + 0x000000ff, // IA_MULTI_VGT_PARAM + 0x00000000, // VGT_ESGS_RING_ITEMSIZE + 0x00000000, // VGT_GSVS_RING_ITEMSIZE + 0x00000000, // VGT_REUSE_OFF + 0x00000000, // VGT_VTX_CNT_EN + 0x00000000, // DB_HTILE_SURFACE + 0x00000000, // DB_SRESULTS_COMPARE_STATE0 + 0x00000000, // DB_SRESULTS_COMPARE_STATE1 + 0x00000000, // DB_PRELOAD_CONTROL + 0, // HOLE + 0x00000000, // VGT_STRMOUT_BUFFER_SIZE_0 + 0x00000000, // VGT_STRMOUT_VTX_STRIDE_0 + 0, // HOLE + 0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_0 + 0x00000000, // VGT_STRMOUT_BUFFER_SIZE_1 + 0x00000000, // VGT_STRMOUT_VTX_STRIDE_1 + 0, // HOLE + 0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_1 + 0x00000000, // VGT_STRMOUT_BUFFER_SIZE_2 + 0x00000000, // VGT_STRMOUT_VTX_STRIDE_2 + 0, // HOLE + 0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_2 + 0x00000000, // VGT_STRMOUT_BUFFER_SIZE_3 + 0x00000000, // VGT_STRMOUT_VTX_STRIDE_3 + 0, // HOLE + 0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_3 + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_OFFSET + 0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE + 0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE + 0, // HOLE + 0x00000000, // VGT_GS_MAX_VERT_OUT + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0x00000000, // VGT_SHADER_STAGES_EN + 0x00000000, // VGT_LS_HS_CONFIG + 0x00000000, // VGT_GS_VERT_ITEMSIZE + 0x00000000, // VGT_GS_VERT_ITEMSIZE_1 + 0x00000000, // VGT_GS_VERT_ITEMSIZE_2 + 0x00000000, // VGT_GS_VERT_ITEMSIZE_3 + 0x00000000, // VGT_TF_PARAM + 0x00000000, // DB_ALPHA_TO_MASK + 0, // HOLE + 0x00000000, // PA_SU_POLY_OFFSET_DB_FMT_CNTL + 0x00000000, // PA_SU_POLY_OFFSET_CLAMP + 0x00000000, // PA_SU_POLY_OFFSET_FRONT_SCALE + 0x00000000, // PA_SU_POLY_OFFSET_FRONT_OFFSET + 0x00000000, // PA_SU_POLY_OFFSET_BACK_SCALE + 0x00000000, // PA_SU_POLY_OFFSET_BACK_OFFSET + 0x00000000, // VGT_GS_INSTANCE_CNT + 0x00000000, // VGT_STRMOUT_CONFIG + 0x00000000, // VGT_STRMOUT_BUFFER_CONFIG + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0x00000000, // PA_SC_CENTROID_PRIORITY_0 + 0x00000000, // PA_SC_CENTROID_PRIORITY_1 + 0x00001000, // PA_SC_LINE_CNTL + 0x00000000, // PA_SC_AA_CONFIG + 0x00000005, // PA_SU_VTX_CNTL + 0x3f800000, // PA_CL_GB_VERT_CLIP_ADJ + 0x3f800000, // PA_CL_GB_VERT_DISC_ADJ + 0x3f800000, // PA_CL_GB_HORZ_CLIP_ADJ + 0x3f800000, // PA_CL_GB_HORZ_DISC_ADJ + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_1 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_2 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_3 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_1 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_2 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_3 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_1 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_2 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_3 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_1 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_2 + 0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_3 + 0xffffffff, // PA_SC_AA_MASK_X0Y0_X1Y0 + 0xffffffff, // PA_SC_AA_MASK_X0Y1_X1Y1 + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0, // HOLE + 0x0000000e, // VGT_VERTEX_REUSE_BLOCK_CNTL + 0x00000010, // VGT_OUT_DEALLOC_CNTL + 0x00000000, // CB_COLOR0_BASE + 0x00000000, // CB_COLOR0_PITCH + 0x00000000, // CB_COLOR0_SLICE + 0x00000000, // CB_COLOR0_VIEW + 0x00000000, // CB_COLOR0_INFO + 0x00000000, // CB_COLOR0_ATTRIB + 0, // HOLE + 0x00000000, // CB_COLOR0_CMASK + 0x00000000, // CB_COLOR0_CMASK_SLICE + 0x00000000, // CB_COLOR0_FMASK + 0x00000000, // CB_COLOR0_FMASK_SLICE + 0x00000000, // CB_COLOR0_CLEAR_WORD0 + 0x00000000, // CB_COLOR0_CLEAR_WORD1 + 0, // HOLE + 0, // HOLE + 0x00000000, // CB_COLOR1_BASE + 0x00000000, // CB_COLOR1_PITCH + 0x00000000, // CB_COLOR1_SLICE + 0x00000000, // CB_COLOR1_VIEW + 0x00000000, // CB_COLOR1_INFO + 0x00000000, // CB_COLOR1_ATTRIB + 0, // HOLE + 0x00000000, // CB_COLOR1_CMASK + 0x00000000, // CB_COLOR1_CMASK_SLICE + 0x00000000, // CB_COLOR1_FMASK + 0x00000000, // CB_COLOR1_FMASK_SLICE + 0x00000000, // CB_COLOR1_CLEAR_WORD0 + 0x00000000, // CB_COLOR1_CLEAR_WORD1 + 0, // HOLE + 0, // HOLE + 0x00000000, // CB_COLOR2_BASE + 0x00000000, // CB_COLOR2_PITCH + 0x00000000, // CB_COLOR2_SLICE + 0x00000000, // CB_COLOR2_VIEW + 0x00000000, // CB_COLOR2_INFO + 0x00000000, // CB_COLOR2_ATTRIB + 0, // HOLE + 0x00000000, // CB_COLOR2_CMASK + 0x00000000, // CB_COLOR2_CMASK_SLICE + 0x00000000, // CB_COLOR2_FMASK + 0x00000000, // CB_COLOR2_FMASK_SLICE + 0x00000000, // CB_COLOR2_CLEAR_WORD0 + 0x00000000, // CB_COLOR2_CLEAR_WORD1 + 0, // HOLE + 0, // HOLE + 0x00000000, // CB_COLOR3_BASE + 0x00000000, // CB_COLOR3_PITCH + 0x00000000, // CB_COLOR3_SLICE + 0x00000000, // CB_COLOR3_VIEW + 0x00000000, // CB_COLOR3_INFO + 0x00000000, // CB_COLOR3_ATTRIB + 0, // HOLE + 0x00000000, // CB_COLOR3_CMASK + 0x00000000, // CB_COLOR3_CMASK_SLICE + 0x00000000, // CB_COLOR3_FMASK + 0x00000000, // CB_COLOR3_FMASK_SLICE + 0x00000000, // CB_COLOR3_CLEAR_WORD0 + 0x00000000, // CB_COLOR3_CLEAR_WORD1 + 0, // HOLE + 0, // HOLE + 0x00000000, // CB_COLOR4_BASE + 0x00000000, // CB_COLOR4_PITCH + 0x00000000, // CB_COLOR4_SLICE + 0x00000000, // CB_COLOR4_VIEW + 0x00000000, // CB_COLOR4_INFO + 0x00000000, // CB_COLOR4_ATTRIB + 0, // HOLE + 0x00000000, // CB_COLOR4_CMASK + 0x00000000, // CB_COLOR4_CMASK_SLICE + 0x00000000, // CB_COLOR4_FMASK + 0x00000000, // CB_COLOR4_FMASK_SLICE + 0x00000000, // CB_COLOR4_CLEAR_WORD0 + 0x00000000, // CB_COLOR4_CLEAR_WORD1 + 0, // HOLE + 0, // HOLE + 0x00000000, // CB_COLOR5_BASE + 0x00000000, // CB_COLOR5_PITCH + 0x00000000, // CB_COLOR5_SLICE + 0x00000000, // CB_COLOR5_VIEW + 0x00000000, // CB_COLOR5_INFO + 0x00000000, // CB_COLOR5_ATTRIB + 0, // HOLE + 0x00000000, // CB_COLOR5_CMASK + 0x00000000, // CB_COLOR5_CMASK_SLICE + 0x00000000, // CB_COLOR5_FMASK + 0x00000000, // CB_COLOR5_FMASK_SLICE + 0x00000000, // CB_COLOR5_CLEAR_WORD0 + 0x00000000, // CB_COLOR5_CLEAR_WORD1 + 0, // HOLE + 0, // HOLE + 0x00000000, // CB_COLOR6_BASE + 0x00000000, // CB_COLOR6_PITCH + 0x00000000, // CB_COLOR6_SLICE + 0x00000000, // CB_COLOR6_VIEW + 0x00000000, // CB_COLOR6_INFO + 0x00000000, // CB_COLOR6_ATTRIB + 0, // HOLE + 0x00000000, // CB_COLOR6_CMASK + 0x00000000, // CB_COLOR6_CMASK_SLICE + 0x00000000, // CB_COLOR6_FMASK + 0x00000000, // CB_COLOR6_FMASK_SLICE + 0x00000000, // CB_COLOR6_CLEAR_WORD0 + 0x00000000, // CB_COLOR6_CLEAR_WORD1 + 0, // HOLE + 0, // HOLE + 0x00000000, // CB_COLOR7_BASE + 0x00000000, // CB_COLOR7_PITCH + 0x00000000, // CB_COLOR7_SLICE + 0x00000000, // CB_COLOR7_VIEW + 0x00000000, // CB_COLOR7_INFO + 0x00000000, // CB_COLOR7_ATTRIB + 0, // HOLE + 0x00000000, // CB_COLOR7_CMASK + 0x00000000, // CB_COLOR7_CMASK_SLICE + 0x00000000, // CB_COLOR7_FMASK + 0x00000000, // CB_COLOR7_FMASK_SLICE + 0x00000000, // CB_COLOR7_CLEAR_WORD0 + 0x00000000, // CB_COLOR7_CLEAR_WORD1 +}; +static const struct cs_extent_def ci_SECT_CONTEXT_defs[] = +{ + {ci_SECT_CONTEXT_def_1, 0x0000a000, 212 }, + {ci_SECT_CONTEXT_def_2, 0x0000a0d6, 274 }, + {ci_SECT_CONTEXT_def_3, 0x0000a1f5, 6 }, + {ci_SECT_CONTEXT_def_4, 0x0000a200, 157 }, + {ci_SECT_CONTEXT_def_5, 0x0000a2a0, 2 }, + {ci_SECT_CONTEXT_def_6, 0x0000a2a3, 1 }, + {ci_SECT_CONTEXT_def_7, 0x0000a2a5, 233 }, + { NULL, 0, 0 } +}; +static const struct cs_section_def ci_cs_data[] = { + { ci_SECT_CONTEXT_defs, SECT_CONTEXT }, + { NULL, SECT_NONE } +}; diff --git a/drivers/gpu/drm/radeon/clearstate_evergreen.h b/drivers/gpu/drm/radeon/clearstate_evergreen.h index 4791d856b7f..63a1ffbb3ce 100644 --- a/drivers/gpu/drm/radeon/clearstate_evergreen.h +++ b/drivers/gpu/drm/radeon/clearstate_evergreen.h @@ -1072,7 +1072,7 @@ static const struct cs_extent_def SECT_CTRLCONST_defs[] = {SECT_CTRLCONST_def_1, 0x0000f3fc, 2 }, { 0, 0, 0 } }; -struct cs_section_def evergreen_cs_data[] = { +static const struct cs_section_def evergreen_cs_data[] = { { SECT_CONTEXT_defs, SECT_CONTEXT }, { SECT_CLEAR_defs, SECT_CLEAR }, { SECT_CTRLCONST_defs, SECT_CTRLCONST }, diff --git a/drivers/gpu/drm/radeon/clearstate_si.h b/drivers/gpu/drm/radeon/clearstate_si.h index b994cb2a35a..66e39cdb5cb 100644 --- a/drivers/gpu/drm/radeon/clearstate_si.h +++ b/drivers/gpu/drm/radeon/clearstate_si.h @@ -933,9 +933,9 @@ static const struct cs_extent_def si_SECT_CONTEXT_defs[] = {si_SECT_CONTEXT_def_5, 0x0000a2a1, 1 }, {si_SECT_CONTEXT_def_6, 0x0000a2a3, 1 }, {si_SECT_CONTEXT_def_7, 0x0000a2a5, 233 }, - { 0, 0, 0 } + { NULL, 0, 0 } }; static const struct cs_section_def si_cs_data[] = { { si_SECT_CONTEXT_defs, SECT_CONTEXT }, - { 0, SECT_NONE } + { NULL, SECT_NONE } }; diff --git a/drivers/gpu/drm/radeon/cypress_dpm.c b/drivers/gpu/drm/radeon/cypress_dpm.c index 9bcdd174780..47d31e91575 100644 --- a/drivers/gpu/drm/radeon/cypress_dpm.c +++ b/drivers/gpu/drm/radeon/cypress_dpm.c @@ -299,7 +299,9 @@ void cypress_program_response_times(struct radeon_device *rdev) static int cypress_pcie_performance_request(struct radeon_device *rdev, u8 perf_req, bool advertise) { +#if defined(CONFIG_ACPI) struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev); +#endif u32 tmp; udelay(10); @@ -1549,7 +1551,7 @@ int cypress_populate_smc_voltage_tables(struct radeon_device *rdev, table->voltageMaskTable.highMask[RV770_SMC_VOLTAGEMASK_VDDCI] = 0; table->voltageMaskTable.lowMask[RV770_SMC_VOLTAGEMASK_VDDCI] = - cpu_to_be32(eg_pi->vddc_voltage_table.mask_low); + cpu_to_be32(eg_pi->vddci_voltage_table.mask_low); } return 0; @@ -1903,21 +1905,6 @@ int cypress_dpm_enable(struct radeon_device *rdev) if (pi->mg_clock_gating) cypress_mg_clock_gating_enable(rdev, true); - if (rdev->irq.installed && - r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { - PPSMC_Result result; - - ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX); - if (ret) - return ret; - rdev->irq.dpm_thermal = true; - radeon_irq_set(rdev); - result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt); - - if (result != PPSMC_Result_OK) - DRM_DEBUG_KMS("Could not enable thermal interrupts.\n"); - } - rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true); return 0; @@ -2014,12 +2001,6 @@ int cypress_dpm_set_power_state(struct radeon_device *rdev) if (eg_pi->pcie_performance_request) cypress_notify_link_speed_change_after_state_change(rdev, new_ps, old_ps); - ret = rv770_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO); - if (ret) { - DRM_ERROR("rv770_dpm_force_performance_level failed\n"); - return ret; - } - return 0; } @@ -2038,9 +2019,6 @@ int cypress_dpm_init(struct radeon_device *rdev) { struct rv7xx_power_info *pi; struct evergreen_power_info *eg_pi; - int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info); - uint16_t data_offset, size; - uint8_t frev, crev; struct atom_clock_dividers dividers; int ret; @@ -2058,6 +2036,10 @@ int cypress_dpm_init(struct radeon_device *rdev) pi->min_vddc_in_table = 0; pi->max_vddc_in_table = 0; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = rv7xx_parse_power_table(rdev); if (ret) return ret; @@ -2092,16 +2074,7 @@ int cypress_dpm_init(struct radeon_device *rdev) eg_pi->vddci_control = radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0); - if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size, - &frev, &crev, &data_offset)) { - pi->sclk_ss = true; - pi->mclk_ss = true; - pi->dynamic_ss = true; - } else { - pi->sclk_ss = false; - pi->mclk_ss = false; - pi->dynamic_ss = true; - } + rv770_get_engine_memory_ss(rdev); pi->asi = RV770_ASI_DFLT; pi->pasi = CYPRESS_HASI_DFLT; @@ -2122,8 +2095,7 @@ int cypress_dpm_init(struct radeon_device *rdev) pi->dynamic_pcie_gen2 = true; - if (pi->gfx_clock_gating && - (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE)) + if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE) pi->thermal_protection = true; else pi->thermal_protection = false; @@ -2179,7 +2151,8 @@ bool cypress_dpm_vblank_too_short(struct radeon_device *rdev) { struct rv7xx_power_info *pi = rv770_get_pi(rdev); u32 vblank_time = r600_dpm_get_vblank_time(rdev); - u32 switch_limit = pi->mem_gddr5 ? 450 : 300; + /* we never hit the non-gddr5 limit so disable it */ + u32 switch_limit = pi->mem_gddr5 ? 450 : 0; if (vblank_time < switch_limit) return true; diff --git a/drivers/gpu/drm/radeon/dce3_1_afmt.c b/drivers/gpu/drm/radeon/dce3_1_afmt.c new file mode 100644 index 00000000000..51800e340a5 --- /dev/null +++ b/drivers/gpu/drm/radeon/dce3_1_afmt.c @@ -0,0 +1,244 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * Copyright 2014 Rafał Miłecki + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ +#include <linux/hdmi.h> +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "r600d.h" + +static void dce3_2_afmt_write_speaker_allocation(struct drm_encoder *encoder) +{ + struct radeon_device *rdev = encoder->dev->dev_private; + struct drm_connector *connector; + struct radeon_connector *radeon_connector = NULL; + u32 tmp; + u8 *sadb; + int sad_count; + + list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) { + if (connector->encoder == encoder) { + radeon_connector = to_radeon_connector(connector); + break; + } + } + + if (!radeon_connector) { + DRM_ERROR("Couldn't find encoder's connector\n"); + return; + } + + sad_count = drm_edid_to_speaker_allocation(radeon_connector->edid, &sadb); + if (sad_count < 0) { + DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count); + return; + } + + /* program the speaker allocation */ + tmp = RREG32(AZ_F0_CODEC_PIN0_CONTROL_CHANNEL_SPEAKER); + tmp &= ~(DP_CONNECTION | SPEAKER_ALLOCATION_MASK); + /* set HDMI mode */ + tmp |= HDMI_CONNECTION; + if (sad_count) + tmp |= SPEAKER_ALLOCATION(sadb[0]); + else + tmp |= SPEAKER_ALLOCATION(5); /* stereo */ + WREG32(AZ_F0_CODEC_PIN0_CONTROL_CHANNEL_SPEAKER, tmp); + + kfree(sadb); +} + +static void dce3_2_afmt_write_sad_regs(struct drm_encoder *encoder) +{ + struct radeon_device *rdev = encoder->dev->dev_private; + struct drm_connector *connector; + struct radeon_connector *radeon_connector = NULL; + struct cea_sad *sads; + int i, sad_count; + + static const u16 eld_reg_to_type[][2] = { + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR0, HDMI_AUDIO_CODING_TYPE_PCM }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR1, HDMI_AUDIO_CODING_TYPE_AC3 }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR2, HDMI_AUDIO_CODING_TYPE_MPEG1 }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR3, HDMI_AUDIO_CODING_TYPE_MP3 }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR4, HDMI_AUDIO_CODING_TYPE_MPEG2 }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR5, HDMI_AUDIO_CODING_TYPE_AAC_LC }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR6, HDMI_AUDIO_CODING_TYPE_DTS }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR7, HDMI_AUDIO_CODING_TYPE_ATRAC }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR9, HDMI_AUDIO_CODING_TYPE_EAC3 }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR10, HDMI_AUDIO_CODING_TYPE_DTS_HD }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR11, HDMI_AUDIO_CODING_TYPE_MLP }, + { AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR13, HDMI_AUDIO_CODING_TYPE_WMA_PRO }, + }; + + list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) { + if (connector->encoder == encoder) { + radeon_connector = to_radeon_connector(connector); + break; + } + } + + if (!radeon_connector) { + DRM_ERROR("Couldn't find encoder's connector\n"); + return; + } + + sad_count = drm_edid_to_sad(radeon_connector->edid, &sads); + if (sad_count < 0) { + DRM_ERROR("Couldn't read SADs: %d\n", sad_count); + return; + } + BUG_ON(!sads); + + for (i = 0; i < ARRAY_SIZE(eld_reg_to_type); i++) { + u32 value = 0; + u8 stereo_freqs = 0; + int max_channels = -1; + int j; + + for (j = 0; j < sad_count; j++) { + struct cea_sad *sad = &sads[j]; + + if (sad->format == eld_reg_to_type[i][1]) { + if (sad->channels > max_channels) { + value = MAX_CHANNELS(sad->channels) | + DESCRIPTOR_BYTE_2(sad->byte2) | + SUPPORTED_FREQUENCIES(sad->freq); + max_channels = sad->channels; + } + + if (sad->format == HDMI_AUDIO_CODING_TYPE_PCM) + stereo_freqs |= sad->freq; + else + break; + } + } + + value |= SUPPORTED_FREQUENCIES_STEREO(stereo_freqs); + + WREG32(eld_reg_to_type[i][0], value); + } + + kfree(sads); +} + +/* + * update the info frames with the data from the current display mode + */ +void dce3_1_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode) +{ + struct drm_device *dev = encoder->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + u8 buffer[HDMI_INFOFRAME_HEADER_SIZE + HDMI_AVI_INFOFRAME_SIZE]; + struct hdmi_avi_infoframe frame; + uint32_t offset; + ssize_t err; + + if (!dig || !dig->afmt) + return; + + /* Silent, r600_hdmi_enable will raise WARN for us */ + if (!dig->afmt->enabled) + return; + offset = dig->afmt->offset; + + /* disable audio prior to setting up hw */ + dig->afmt->pin = r600_audio_get_pin(rdev); + r600_audio_enable(rdev, dig->afmt->pin, false); + + r600_audio_set_dto(encoder, mode->clock); + + WREG32(HDMI0_VBI_PACKET_CONTROL + offset, + HDMI0_NULL_SEND); /* send null packets when required */ + + WREG32(HDMI0_AUDIO_CRC_CONTROL + offset, 0x1000); + + if (ASIC_IS_DCE32(rdev)) { + WREG32(HDMI0_AUDIO_PACKET_CONTROL + offset, + HDMI0_AUDIO_DELAY_EN(1) | /* default audio delay */ + HDMI0_AUDIO_PACKETS_PER_LINE(3)); /* should be suffient for all audio modes and small enough for all hblanks */ + WREG32(AFMT_AUDIO_PACKET_CONTROL + offset, + AFMT_AUDIO_SAMPLE_SEND | /* send audio packets */ + AFMT_60958_CS_UPDATE); /* allow 60958 channel status fields to be updated */ + } else { + WREG32(HDMI0_AUDIO_PACKET_CONTROL + offset, + HDMI0_AUDIO_SAMPLE_SEND | /* send audio packets */ + HDMI0_AUDIO_DELAY_EN(1) | /* default audio delay */ + HDMI0_AUDIO_PACKETS_PER_LINE(3) | /* should be suffient for all audio modes and small enough for all hblanks */ + HDMI0_60958_CS_UPDATE); /* allow 60958 channel status fields to be updated */ + } + + if (ASIC_IS_DCE32(rdev)) { + dce3_2_afmt_write_speaker_allocation(encoder); + dce3_2_afmt_write_sad_regs(encoder); + } + + WREG32(HDMI0_ACR_PACKET_CONTROL + offset, + HDMI0_ACR_SOURCE | /* select SW CTS value - XXX verify that hw CTS works on all families */ + HDMI0_ACR_AUTO_SEND); /* allow hw to sent ACR packets when required */ + + WREG32(HDMI0_VBI_PACKET_CONTROL + offset, + HDMI0_NULL_SEND | /* send null packets when required */ + HDMI0_GC_SEND | /* send general control packets */ + HDMI0_GC_CONT); /* send general control packets every frame */ + + /* TODO: HDMI0_AUDIO_INFO_UPDATE */ + WREG32(HDMI0_INFOFRAME_CONTROL0 + offset, + HDMI0_AVI_INFO_SEND | /* enable AVI info frames */ + HDMI0_AVI_INFO_CONT | /* send AVI info frames every frame/field */ + HDMI0_AUDIO_INFO_SEND | /* enable audio info frames (frames won't be set until audio is enabled) */ + HDMI0_AUDIO_INFO_CONT); /* send audio info frames every frame/field */ + + WREG32(HDMI0_INFOFRAME_CONTROL1 + offset, + HDMI0_AVI_INFO_LINE(2) | /* anything other than 0 */ + HDMI0_AUDIO_INFO_LINE(2)); /* anything other than 0 */ + + WREG32(HDMI0_GC + offset, 0); /* unset HDMI0_GC_AVMUTE */ + + err = drm_hdmi_avi_infoframe_from_display_mode(&frame, mode); + if (err < 0) { + DRM_ERROR("failed to setup AVI infoframe: %zd\n", err); + return; + } + + err = hdmi_avi_infoframe_pack(&frame, buffer, sizeof(buffer)); + if (err < 0) { + DRM_ERROR("failed to pack AVI infoframe: %zd\n", err); + return; + } + + r600_hdmi_update_avi_infoframe(encoder, buffer, sizeof(buffer)); + r600_hdmi_update_ACR(encoder, mode->clock); + + /* it's unknown what these bits do excatly, but it's indeed quite useful for debugging */ + WREG32(HDMI0_RAMP_CONTROL0 + offset, 0x00FFFFFF); + WREG32(HDMI0_RAMP_CONTROL1 + offset, 0x007FFFFF); + WREG32(HDMI0_RAMP_CONTROL2 + offset, 0x00000001); + WREG32(HDMI0_RAMP_CONTROL3 + offset, 0x00000001); + + r600_hdmi_audio_workaround(encoder); + + /* enable audio after to setting up hw */ + r600_audio_enable(rdev, dig->afmt->pin, true); +} diff --git a/drivers/gpu/drm/radeon/dce6_afmt.c b/drivers/gpu/drm/radeon/dce6_afmt.c new file mode 100644 index 00000000000..0a65dc7e93e --- /dev/null +++ b/drivers/gpu/drm/radeon/dce6_afmt.c @@ -0,0 +1,352 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#include <linux/hdmi.h> +#include <drm/drmP.h> +#include "radeon.h" +#include "sid.h" + +static u32 dce6_endpoint_rreg(struct radeon_device *rdev, + u32 block_offset, u32 reg) +{ + unsigned long flags; + u32 r; + + spin_lock_irqsave(&rdev->end_idx_lock, flags); + WREG32(AZ_F0_CODEC_ENDPOINT_INDEX + block_offset, reg); + r = RREG32(AZ_F0_CODEC_ENDPOINT_DATA + block_offset); + spin_unlock_irqrestore(&rdev->end_idx_lock, flags); + + return r; +} + +static void dce6_endpoint_wreg(struct radeon_device *rdev, + u32 block_offset, u32 reg, u32 v) +{ + unsigned long flags; + + spin_lock_irqsave(&rdev->end_idx_lock, flags); + if (ASIC_IS_DCE8(rdev)) + WREG32(AZ_F0_CODEC_ENDPOINT_INDEX + block_offset, reg); + else + WREG32(AZ_F0_CODEC_ENDPOINT_INDEX + block_offset, + AZ_ENDPOINT_REG_WRITE_EN | AZ_ENDPOINT_REG_INDEX(reg)); + WREG32(AZ_F0_CODEC_ENDPOINT_DATA + block_offset, v); + spin_unlock_irqrestore(&rdev->end_idx_lock, flags); +} + +#define RREG32_ENDPOINT(block, reg) dce6_endpoint_rreg(rdev, (block), (reg)) +#define WREG32_ENDPOINT(block, reg, v) dce6_endpoint_wreg(rdev, (block), (reg), (v)) + + +static void dce6_afmt_get_connected_pins(struct radeon_device *rdev) +{ + int i; + u32 offset, tmp; + + for (i = 0; i < rdev->audio.num_pins; i++) { + offset = rdev->audio.pin[i].offset; + tmp = RREG32_ENDPOINT(offset, + AZ_F0_CODEC_PIN_CONTROL_RESPONSE_CONFIGURATION_DEFAULT); + if (((tmp & PORT_CONNECTIVITY_MASK) >> PORT_CONNECTIVITY_SHIFT) == 1) + rdev->audio.pin[i].connected = false; + else + rdev->audio.pin[i].connected = true; + } +} + +struct r600_audio_pin *dce6_audio_get_pin(struct radeon_device *rdev) +{ + int i; + + dce6_afmt_get_connected_pins(rdev); + + for (i = 0; i < rdev->audio.num_pins; i++) { + if (rdev->audio.pin[i].connected) + return &rdev->audio.pin[i]; + } + DRM_ERROR("No connected audio pins found!\n"); + return NULL; +} + +void dce6_afmt_select_pin(struct drm_encoder *encoder) +{ + struct radeon_device *rdev = encoder->dev->dev_private; + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + u32 offset; + + if (!dig || !dig->afmt || !dig->afmt->pin) + return; + + offset = dig->afmt->offset; + + WREG32(AFMT_AUDIO_SRC_CONTROL + offset, + AFMT_AUDIO_SRC_SELECT(dig->afmt->pin->id)); +} + +void dce6_afmt_write_latency_fields(struct drm_encoder *encoder, + struct drm_display_mode *mode) +{ + struct radeon_device *rdev = encoder->dev->dev_private; + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + struct drm_connector *connector; + struct radeon_connector *radeon_connector = NULL; + u32 tmp = 0, offset; + + if (!dig || !dig->afmt || !dig->afmt->pin) + return; + + offset = dig->afmt->pin->offset; + + list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) { + if (connector->encoder == encoder) { + radeon_connector = to_radeon_connector(connector); + break; + } + } + + if (!radeon_connector) { + DRM_ERROR("Couldn't find encoder's connector\n"); + return; + } + + if (mode->flags & DRM_MODE_FLAG_INTERLACE) { + if (connector->latency_present[1]) + tmp = VIDEO_LIPSYNC(connector->video_latency[1]) | + AUDIO_LIPSYNC(connector->audio_latency[1]); + else + tmp = VIDEO_LIPSYNC(255) | AUDIO_LIPSYNC(255); + } else { + if (connector->latency_present[0]) + tmp = VIDEO_LIPSYNC(connector->video_latency[0]) | + AUDIO_LIPSYNC(connector->audio_latency[0]); + else + tmp = VIDEO_LIPSYNC(255) | AUDIO_LIPSYNC(255); + } + WREG32_ENDPOINT(offset, AZ_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC, tmp); +} + +void dce6_afmt_write_speaker_allocation(struct drm_encoder *encoder) +{ + struct radeon_device *rdev = encoder->dev->dev_private; + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + struct drm_connector *connector; + struct radeon_connector *radeon_connector = NULL; + u32 offset, tmp; + u8 *sadb; + int sad_count; + + if (!dig || !dig->afmt || !dig->afmt->pin) + return; + + offset = dig->afmt->pin->offset; + + list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) { + if (connector->encoder == encoder) + radeon_connector = to_radeon_connector(connector); + } + + if (!radeon_connector) { + DRM_ERROR("Couldn't find encoder's connector\n"); + return; + } + + sad_count = drm_edid_to_speaker_allocation(radeon_connector->edid, &sadb); + if (sad_count <= 0) { + DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count); + return; + } + + /* program the speaker allocation */ + tmp = RREG32_ENDPOINT(offset, AZ_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER); + tmp &= ~(DP_CONNECTION | SPEAKER_ALLOCATION_MASK); + /* set HDMI mode */ + tmp |= HDMI_CONNECTION; + if (sad_count) + tmp |= SPEAKER_ALLOCATION(sadb[0]); + else + tmp |= SPEAKER_ALLOCATION(5); /* stereo */ + WREG32_ENDPOINT(offset, AZ_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER, tmp); + + kfree(sadb); +} + +void dce6_afmt_write_sad_regs(struct drm_encoder *encoder) +{ + struct radeon_device *rdev = encoder->dev->dev_private; + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + u32 offset; + struct drm_connector *connector; + struct radeon_connector *radeon_connector = NULL; + struct cea_sad *sads; + int i, sad_count; + + static const u16 eld_reg_to_type[][2] = { + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0, HDMI_AUDIO_CODING_TYPE_PCM }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR1, HDMI_AUDIO_CODING_TYPE_AC3 }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR2, HDMI_AUDIO_CODING_TYPE_MPEG1 }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR3, HDMI_AUDIO_CODING_TYPE_MP3 }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR4, HDMI_AUDIO_CODING_TYPE_MPEG2 }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR5, HDMI_AUDIO_CODING_TYPE_AAC_LC }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR6, HDMI_AUDIO_CODING_TYPE_DTS }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR7, HDMI_AUDIO_CODING_TYPE_ATRAC }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR9, HDMI_AUDIO_CODING_TYPE_EAC3 }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR10, HDMI_AUDIO_CODING_TYPE_DTS_HD }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR11, HDMI_AUDIO_CODING_TYPE_MLP }, + { AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR13, HDMI_AUDIO_CODING_TYPE_WMA_PRO }, + }; + + if (!dig || !dig->afmt || !dig->afmt->pin) + return; + + offset = dig->afmt->pin->offset; + + list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) { + if (connector->encoder == encoder) + radeon_connector = to_radeon_connector(connector); + } + + if (!radeon_connector) { + DRM_ERROR("Couldn't find encoder's connector\n"); + return; + } + + sad_count = drm_edid_to_sad(radeon_connector->edid, &sads); + if (sad_count <= 0) { + DRM_ERROR("Couldn't read SADs: %d\n", sad_count); + return; + } + BUG_ON(!sads); + + for (i = 0; i < ARRAY_SIZE(eld_reg_to_type); i++) { + u32 value = 0; + u8 stereo_freqs = 0; + int max_channels = -1; + int j; + + for (j = 0; j < sad_count; j++) { + struct cea_sad *sad = &sads[j]; + + if (sad->format == eld_reg_to_type[i][1]) { + if (sad->channels > max_channels) { + value = MAX_CHANNELS(sad->channels) | + DESCRIPTOR_BYTE_2(sad->byte2) | + SUPPORTED_FREQUENCIES(sad->freq); + max_channels = sad->channels; + } + + if (sad->format == HDMI_AUDIO_CODING_TYPE_PCM) + stereo_freqs |= sad->freq; + else + break; + } + } + + value |= SUPPORTED_FREQUENCIES_STEREO(stereo_freqs); + + WREG32_ENDPOINT(offset, eld_reg_to_type[i][0], value); + } + + kfree(sads); +} + +static int dce6_audio_chipset_supported(struct radeon_device *rdev) +{ + return !ASIC_IS_NODCE(rdev); +} + +void dce6_audio_enable(struct radeon_device *rdev, + struct r600_audio_pin *pin, + bool enable) +{ + if (!pin) + return; + + WREG32_ENDPOINT(pin->offset, AZ_F0_CODEC_PIN_CONTROL_HOTPLUG_CONTROL, + enable ? AUDIO_ENABLED : 0); +} + +static const u32 pin_offsets[7] = +{ + (0x5e00 - 0x5e00), + (0x5e18 - 0x5e00), + (0x5e30 - 0x5e00), + (0x5e48 - 0x5e00), + (0x5e60 - 0x5e00), + (0x5e78 - 0x5e00), + (0x5e90 - 0x5e00), +}; + +int dce6_audio_init(struct radeon_device *rdev) +{ + int i; + + if (!radeon_audio || !dce6_audio_chipset_supported(rdev)) + return 0; + + rdev->audio.enabled = true; + + if (ASIC_IS_DCE81(rdev)) /* KV: 4 streams, 7 endpoints */ + rdev->audio.num_pins = 7; + else if (ASIC_IS_DCE83(rdev)) /* KB: 2 streams, 3 endpoints */ + rdev->audio.num_pins = 3; + else if (ASIC_IS_DCE8(rdev)) /* BN/HW: 6 streams, 7 endpoints */ + rdev->audio.num_pins = 7; + else if (ASIC_IS_DCE61(rdev)) /* TN: 4 streams, 6 endpoints */ + rdev->audio.num_pins = 6; + else if (ASIC_IS_DCE64(rdev)) /* OL: 2 streams, 2 endpoints */ + rdev->audio.num_pins = 2; + else /* SI: 6 streams, 6 endpoints */ + rdev->audio.num_pins = 6; + + for (i = 0; i < rdev->audio.num_pins; i++) { + rdev->audio.pin[i].channels = -1; + rdev->audio.pin[i].rate = -1; + rdev->audio.pin[i].bits_per_sample = -1; + rdev->audio.pin[i].status_bits = 0; + rdev->audio.pin[i].category_code = 0; + rdev->audio.pin[i].connected = false; + rdev->audio.pin[i].offset = pin_offsets[i]; + rdev->audio.pin[i].id = i; + /* disable audio. it will be set up later */ + dce6_audio_enable(rdev, &rdev->audio.pin[i], false); + } + + return 0; +} + +void dce6_audio_fini(struct radeon_device *rdev) +{ + int i; + + if (!rdev->audio.enabled) + return; + + for (i = 0; i < rdev->audio.num_pins; i++) + dce6_audio_enable(rdev, &rdev->audio.pin[i], false); + + rdev->audio.enabled = false; +} diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c index e49059dc9b8..15e4f28015e 100644 --- a/drivers/gpu/drm/radeon/evergreen.c +++ b/drivers/gpu/drm/radeon/evergreen.c @@ -47,7 +47,7 @@ static const u32 crtc_offsets[6] = #include "clearstate_evergreen.h" -static u32 sumo_rlc_save_restore_register_list[] = +static const u32 sumo_rlc_save_restore_register_list[] = { 0x98fc, 0x9830, @@ -131,7 +131,6 @@ static u32 sumo_rlc_save_restore_register_list[] = 0x9150, 0x802c, }; -static u32 sumo_rlc_save_restore_register_list_size = ARRAY_SIZE(sumo_rlc_save_restore_register_list); static void evergreen_gpu_init(struct radeon_device *rdev); void evergreen_fini(struct radeon_device *rdev); @@ -139,6 +138,15 @@ void evergreen_pcie_gen2_enable(struct radeon_device *rdev); void evergreen_program_aspm(struct radeon_device *rdev); extern void cayman_cp_int_cntl_setup(struct radeon_device *rdev, int ring, u32 cp_int_cntl); +extern void cayman_vm_decode_fault(struct radeon_device *rdev, + u32 status, u32 addr); +void cik_init_cp_pg_table(struct radeon_device *rdev); + +extern u32 si_get_csb_size(struct radeon_device *rdev); +extern void si_get_csb_buffer(struct radeon_device *rdev, volatile u32 *buffer); +extern u32 cik_get_csb_size(struct radeon_device *rdev); +extern void cik_get_csb_buffer(struct radeon_device *rdev, volatile u32 *buffer); +extern void rv770_set_clk_bypass_mode(struct radeon_device *rdev); static const u32 evergreen_golden_registers[] = { @@ -181,7 +189,7 @@ static const u32 evergreen_golden_registers[] = 0x8c1c, 0xffffffff, 0x00001010, 0x28350, 0xffffffff, 0x00000000, 0xa008, 0xffffffff, 0x00010000, - 0x5cc, 0xffffffff, 0x00000001, + 0x5c4, 0xffffffff, 0x00000001, 0x9508, 0xffffffff, 0x00000002, 0x913c, 0x0000000f, 0x0000000a }; @@ -468,7 +476,7 @@ static const u32 cedar_golden_registers[] = 0x8c1c, 0xffffffff, 0x00001010, 0x28350, 0xffffffff, 0x00000000, 0xa008, 0xffffffff, 0x00010000, - 0x5cc, 0xffffffff, 0x00000001, + 0x5c4, 0xffffffff, 0x00000001, 0x9508, 0xffffffff, 0x00000002 }; @@ -627,7 +635,7 @@ static const u32 juniper_mgcg_init[] = static const u32 supersumo_golden_registers[] = { 0x5eb4, 0xffffffff, 0x00000002, - 0x5cc, 0xffffffff, 0x00000001, + 0x5c4, 0xffffffff, 0x00000001, 0x7030, 0xffffffff, 0x00000011, 0x7c30, 0xffffffff, 0x00000011, 0x6104, 0x01000300, 0x00000000, @@ -711,7 +719,7 @@ static const u32 sumo_golden_registers[] = static const u32 wrestler_golden_registers[] = { 0x5eb4, 0xffffffff, 0x00000002, - 0x5cc, 0xffffffff, 0x00000001, + 0x5c4, 0xffffffff, 0x00000001, 0x7030, 0xffffffff, 0x00000011, 0x7c30, 0xffffffff, 0x00000011, 0x6104, 0x01000300, 0x00000000, @@ -1167,23 +1175,72 @@ int evergreen_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk) void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev) { - u16 ctl, v; - int err; - - err = pcie_capability_read_word(rdev->pdev, PCI_EXP_DEVCTL, &ctl); - if (err) - return; - - v = (ctl & PCI_EXP_DEVCTL_READRQ) >> 12; + int readrq; + u16 v; + readrq = pcie_get_readrq(rdev->pdev); + v = ffs(readrq) - 8; /* if bios or OS sets MAX_READ_REQUEST_SIZE to an invalid value, fix it * to avoid hangs or perfomance issues */ - if ((v == 0) || (v == 6) || (v == 7)) { - ctl &= ~PCI_EXP_DEVCTL_READRQ; - ctl |= (2 << 12); - pcie_capability_write_word(rdev->pdev, PCI_EXP_DEVCTL, ctl); + if ((v == 0) || (v == 6) || (v == 7)) + pcie_set_readrq(rdev->pdev, 512); +} + +void dce4_program_fmt(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); + struct drm_connector *connector = radeon_get_connector_for_encoder(encoder); + int bpc = 0; + u32 tmp = 0; + enum radeon_connector_dither dither = RADEON_FMT_DITHER_DISABLE; + + if (connector) { + struct radeon_connector *radeon_connector = to_radeon_connector(connector); + bpc = radeon_get_monitor_bpc(connector); + dither = radeon_connector->dither; } + + /* LVDS/eDP FMT is set up by atom */ + if (radeon_encoder->devices & ATOM_DEVICE_LCD_SUPPORT) + return; + + /* not needed for analog */ + if ((radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1) || + (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2)) + return; + + if (bpc == 0) + return; + + switch (bpc) { + case 6: + if (dither == RADEON_FMT_DITHER_ENABLE) + /* XXX sort out optimal dither settings */ + tmp |= (FMT_FRAME_RANDOM_ENABLE | FMT_HIGHPASS_RANDOM_ENABLE | + FMT_SPATIAL_DITHER_EN); + else + tmp |= FMT_TRUNCATE_EN; + break; + case 8: + if (dither == RADEON_FMT_DITHER_ENABLE) + /* XXX sort out optimal dither settings */ + tmp |= (FMT_FRAME_RANDOM_ENABLE | FMT_HIGHPASS_RANDOM_ENABLE | + FMT_RGB_RANDOM_ENABLE | + FMT_SPATIAL_DITHER_EN | FMT_SPATIAL_DITHER_DEPTH); + else + tmp |= (FMT_TRUNCATE_EN | FMT_TRUNCATE_DEPTH); + break; + case 10: + default: + /* not needed */ + break; + } + + WREG32(FMT_BIT_DEPTH_CONTROL + radeon_crtc->crtc_offset, tmp); } static bool dce4_is_in_vblank(struct radeon_device *rdev, int crtc) @@ -1244,36 +1301,6 @@ void dce4_wait_for_vblank(struct radeon_device *rdev, int crtc) } /** - * radeon_irq_kms_pflip_irq_get - pre-pageflip callback. - * - * @rdev: radeon_device pointer - * @crtc: crtc to prepare for pageflip on - * - * Pre-pageflip callback (evergreen+). - * Enables the pageflip irq (vblank irq). - */ -void evergreen_pre_page_flip(struct radeon_device *rdev, int crtc) -{ - /* enable the pflip int */ - radeon_irq_kms_pflip_irq_get(rdev, crtc); -} - -/** - * evergreen_post_page_flip - pos-pageflip callback. - * - * @rdev: radeon_device pointer - * @crtc: crtc to cleanup pageflip on - * - * Post-pageflip callback (evergreen+). - * Disables the pageflip irq (vblank irq). - */ -void evergreen_post_page_flip(struct radeon_device *rdev, int crtc) -{ - /* disable the pflip int */ - radeon_irq_kms_pflip_irq_put(rdev, crtc); -} - -/** * evergreen_page_flip - pageflip callback. * * @rdev: radeon_device pointer @@ -1286,7 +1313,7 @@ void evergreen_post_page_flip(struct radeon_device *rdev, int crtc) * double buffered update to take place. * Returns the current update pending status. */ -u32 evergreen_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) +void evergreen_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) { struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id]; u32 tmp = RREG32(EVERGREEN_GRPH_UPDATE + radeon_crtc->crtc_offset); @@ -1318,9 +1345,23 @@ u32 evergreen_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) /* Unlock the lock, so double-buffering can take place inside vblank */ tmp &= ~EVERGREEN_GRPH_UPDATE_LOCK; WREG32(EVERGREEN_GRPH_UPDATE + radeon_crtc->crtc_offset, tmp); +} + +/** + * evergreen_page_flip_pending - check if page flip is still pending + * + * @rdev: radeon_device pointer + * @crtc_id: crtc to check + * + * Returns the current update pending status. + */ +bool evergreen_page_flip_pending(struct radeon_device *rdev, int crtc_id) +{ + struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id]; /* Return current update_pending status: */ - return RREG32(EVERGREEN_GRPH_UPDATE + radeon_crtc->crtc_offset) & EVERGREEN_GRPH_SURFACE_UPDATE_PENDING; + return !!(RREG32(EVERGREEN_GRPH_UPDATE + radeon_crtc->crtc_offset) & + EVERGREEN_GRPH_SURFACE_UPDATE_PENDING); } /* get temperature in millidegrees */ @@ -1623,7 +1664,7 @@ bool evergreen_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd) case RADEON_HPD_6: if (RREG32(DC_HPD6_INT_STATUS) & DC_HPDx_SENSE) connected = true; - break; + break; default: break; } @@ -1805,7 +1846,8 @@ static u32 evergreen_line_buffer_adjust(struct radeon_device *rdev, struct drm_display_mode *mode, struct drm_display_mode *other_mode) { - u32 tmp; + u32 tmp, buffer_alloc, i; + u32 pipe_offset = radeon_crtc->crtc_id * 0x20; /* * Line Buffer Setup * There are 3 line buffers, each one shared by 2 display controllers. @@ -1828,18 +1870,34 @@ static u32 evergreen_line_buffer_adjust(struct radeon_device *rdev, * non-linked crtcs for maximum line buffer allocation. */ if (radeon_crtc->base.enabled && mode) { - if (other_mode) + if (other_mode) { tmp = 0; /* 1/2 */ - else + buffer_alloc = 1; + } else { tmp = 2; /* whole */ - } else + buffer_alloc = 2; + } + } else { tmp = 0; + buffer_alloc = 0; + } /* second controller of the pair uses second half of the lb */ if (radeon_crtc->crtc_id % 2) tmp += 4; WREG32(DC_LB_MEMORY_SPLIT + radeon_crtc->crtc_offset, tmp); + if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE5(rdev)) { + WREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset, + DMIF_BUFFERS_ALLOCATED(buffer_alloc)); + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset) & + DMIF_BUFFERS_ALLOCATED_COMPLETED) + break; + udelay(1); + } + } + if (radeon_crtc->base.enabled && mode) { switch (tmp) { case 0: @@ -2584,8 +2642,9 @@ void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *s for (i = 0; i < rdev->num_crtc; i++) { if (save->crtc_enabled[i]) { tmp = RREG32(EVERGREEN_MASTER_UPDATE_MODE + crtc_offsets[i]); - if ((tmp & 0x3) != 0) { - tmp &= ~0x3; + if ((tmp & 0x7) != 3) { + tmp &= ~0x7; + tmp |= 0x3; WREG32(EVERGREEN_MASTER_UPDATE_MODE + crtc_offsets[i], tmp); } tmp = RREG32(EVERGREEN_GRPH_UPDATE + crtc_offsets[i]); @@ -2879,8 +2938,8 @@ static int evergreen_cp_resume(struct radeon_device *rdev) RREG32(GRBM_SOFT_RESET); /* Set ring buffer size */ - rb_bufsz = drm_order(ring->ring_size / 8); - tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; + rb_bufsz = order_base_2(ring->ring_size / 8); + tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif @@ -2916,8 +2975,6 @@ static int evergreen_cp_resume(struct radeon_device *rdev) WREG32(CP_RB_BASE, ring->gpu_addr >> 8); WREG32(CP_DEBUG, (1 << 27) | (1 << 28)); - ring->rptr = RREG32(CP_RB_RPTR); - evergreen_cp_start(rdev); ring->ready = true; r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring); @@ -3107,7 +3164,7 @@ static void evergreen_gpu_init(struct radeon_device *rdev) rdev->config.evergreen.sx_max_export_size = 256; rdev->config.evergreen.sx_max_export_pos_size = 64; rdev->config.evergreen.sx_max_export_smx_size = 192; - rdev->config.evergreen.max_hw_contexts = 8; + rdev->config.evergreen.max_hw_contexts = 4; rdev->config.evergreen.sq_num_cf_insts = 2; rdev->config.evergreen.sc_prim_fifo_size = 0x40; @@ -3281,6 +3338,18 @@ static void evergreen_gpu_init(struct radeon_device *rdev) disabled_rb_mask &= ~(1 << i); } + for (i = 0; i < rdev->config.evergreen.num_ses; i++) { + u32 simd_disable_bitmap; + + WREG32(GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_INDEX(i)); + WREG32(RLC_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_INDEX(i)); + simd_disable_bitmap = (RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0xffff0000) >> 16; + simd_disable_bitmap |= 0xffffffff << rdev->config.evergreen.max_simds; + tmp <<= 16; + tmp |= simd_disable_bitmap; + } + rdev->config.evergreen.active_simds = hweight32(~tmp); + WREG32(GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_BROADCAST_WRITES); WREG32(RLC_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_BROADCAST_WRITES); @@ -3611,7 +3680,7 @@ bool evergreen_is_display_hung(struct radeon_device *rdev) return true; } -static u32 evergreen_gpu_check_soft_reset(struct radeon_device *rdev) +u32 evergreen_gpu_check_soft_reset(struct radeon_device *rdev) { u32 reset_mask = 0; u32 tmp; @@ -3794,6 +3863,48 @@ static void evergreen_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask) evergreen_print_gpu_status_regs(rdev); } +void evergreen_gpu_pci_config_reset(struct radeon_device *rdev) +{ + struct evergreen_mc_save save; + u32 tmp, i; + + dev_info(rdev->dev, "GPU pci config reset\n"); + + /* disable dpm? */ + + /* Disable CP parsing/prefetching */ + WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT); + udelay(50); + /* Disable DMA */ + tmp = RREG32(DMA_RB_CNTL); + tmp &= ~DMA_RB_ENABLE; + WREG32(DMA_RB_CNTL, tmp); + /* XXX other engines? */ + + /* halt the rlc */ + r600_rlc_stop(rdev); + + udelay(50); + + /* set mclk/sclk to bypass */ + rv770_set_clk_bypass_mode(rdev); + /* disable BM */ + pci_clear_master(rdev->pdev); + /* disable mem access */ + evergreen_mc_stop(rdev, &save); + if (evergreen_mc_wait_for_idle(rdev)) { + dev_warn(rdev->dev, "Wait for MC idle timed out !\n"); + } + /* reset */ + radeon_pci_config_reset(rdev); + /* wait for asic to come out of reset */ + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(CONFIG_MEMSIZE) != 0xffffffff) + break; + udelay(1); + } +} + int evergreen_asic_reset(struct radeon_device *rdev) { u32 reset_mask; @@ -3803,10 +3914,17 @@ int evergreen_asic_reset(struct radeon_device *rdev) if (reset_mask) r600_set_bios_scratch_engine_hung(rdev, true); + /* try soft reset */ evergreen_gpu_soft_reset(rdev, reset_mask); reset_mask = evergreen_gpu_check_soft_reset(rdev); + /* try pci config reset */ + if (reset_mask && radeon_hard_reset) + evergreen_gpu_pci_config_reset(rdev); + + reset_mask = evergreen_gpu_check_soft_reset(rdev); + if (!reset_mask) r600_set_bios_scratch_engine_hung(rdev, false); @@ -3829,33 +3947,9 @@ bool evergreen_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *rin if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); - return false; - } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); - return radeon_ring_test_lockup(rdev, ring); -} - -/** - * evergreen_dma_is_lockup - Check if the DMA engine is locked up - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Check if the async DMA engine is locked up. - * Returns true if the engine appears to be locked up, false if not. - */ -bool evergreen_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) -{ - u32 reset_mask = evergreen_gpu_check_soft_reset(rdev); - - if (!(reset_mask & RADEON_RESET_DMA)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } @@ -3892,147 +3986,231 @@ void sumo_rlc_fini(struct radeon_device *rdev) radeon_bo_unref(&rdev->rlc.clear_state_obj); rdev->rlc.clear_state_obj = NULL; } + + /* clear state block */ + if (rdev->rlc.cp_table_obj) { + r = radeon_bo_reserve(rdev->rlc.cp_table_obj, false); + if (unlikely(r != 0)) + dev_warn(rdev->dev, "(%d) reserve RLC cp table bo failed\n", r); + radeon_bo_unpin(rdev->rlc.cp_table_obj); + radeon_bo_unreserve(rdev->rlc.cp_table_obj); + + radeon_bo_unref(&rdev->rlc.cp_table_obj); + rdev->rlc.cp_table_obj = NULL; + } } +#define CP_ME_TABLE_SIZE 96 + int sumo_rlc_init(struct radeon_device *rdev) { - u32 *src_ptr; + const u32 *src_ptr; volatile u32 *dst_ptr; u32 dws, data, i, j, k, reg_num; - u32 reg_list_num, reg_list_hdr_blk_index, reg_list_blk_index; + u32 reg_list_num, reg_list_hdr_blk_index, reg_list_blk_index = 0; u64 reg_list_mc_addr; - struct cs_section_def *cs_data; + const struct cs_section_def *cs_data; int r; src_ptr = rdev->rlc.reg_list; dws = rdev->rlc.reg_list_size; + if (rdev->family >= CHIP_BONAIRE) { + dws += (5 * 16) + 48 + 48 + 64; + } cs_data = rdev->rlc.cs_data; - /* save restore block */ - if (rdev->rlc.save_restore_obj == NULL) { - r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true, - RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.save_restore_obj); + if (src_ptr) { + /* save restore block */ + if (rdev->rlc.save_restore_obj == NULL) { + r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true, + RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.save_restore_obj); + if (r) { + dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r); + return r; + } + } + + r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false); + if (unlikely(r != 0)) { + sumo_rlc_fini(rdev); + return r; + } + r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM, + &rdev->rlc.save_restore_gpu_addr); if (r) { - dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r); + radeon_bo_unreserve(rdev->rlc.save_restore_obj); + dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r); + sumo_rlc_fini(rdev); return r; } - } - r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false); - if (unlikely(r != 0)) { - sumo_rlc_fini(rdev); - return r; - } - r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM, - &rdev->rlc.save_restore_gpu_addr); - if (r) { + r = radeon_bo_kmap(rdev->rlc.save_restore_obj, (void **)&rdev->rlc.sr_ptr); + if (r) { + dev_warn(rdev->dev, "(%d) map RLC sr bo failed\n", r); + sumo_rlc_fini(rdev); + return r; + } + /* write the sr buffer */ + dst_ptr = rdev->rlc.sr_ptr; + if (rdev->family >= CHIP_TAHITI) { + /* SI */ + for (i = 0; i < rdev->rlc.reg_list_size; i++) + dst_ptr[i] = cpu_to_le32(src_ptr[i]); + } else { + /* ON/LN/TN */ + /* format: + * dw0: (reg2 << 16) | reg1 + * dw1: reg1 save space + * dw2: reg2 save space + */ + for (i = 0; i < dws; i++) { + data = src_ptr[i] >> 2; + i++; + if (i < dws) + data |= (src_ptr[i] >> 2) << 16; + j = (((i - 1) * 3) / 2); + dst_ptr[j] = cpu_to_le32(data); + } + j = ((i * 3) / 2); + dst_ptr[j] = cpu_to_le32(RLC_SAVE_RESTORE_LIST_END_MARKER); + } + radeon_bo_kunmap(rdev->rlc.save_restore_obj); radeon_bo_unreserve(rdev->rlc.save_restore_obj); - dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r); - sumo_rlc_fini(rdev); - return r; - } - r = radeon_bo_kmap(rdev->rlc.save_restore_obj, (void **)&rdev->rlc.sr_ptr); - if (r) { - dev_warn(rdev->dev, "(%d) map RLC sr bo failed\n", r); - sumo_rlc_fini(rdev); - return r; } - /* write the sr buffer */ - dst_ptr = rdev->rlc.sr_ptr; - /* format: - * dw0: (reg2 << 16) | reg1 - * dw1: reg1 save space - * dw2: reg2 save space - */ - for (i = 0; i < dws; i++) { - data = src_ptr[i] >> 2; - i++; - if (i < dws) - data |= (src_ptr[i] >> 2) << 16; - j = (((i - 1) * 3) / 2); - dst_ptr[j] = data; - } - j = ((i * 3) / 2); - dst_ptr[j] = RLC_SAVE_RESTORE_LIST_END_MARKER; - - radeon_bo_kunmap(rdev->rlc.save_restore_obj); - radeon_bo_unreserve(rdev->rlc.save_restore_obj); - /* clear state block */ - reg_list_num = 0; - dws = 0; - for (i = 0; cs_data[i].section != NULL; i++) { - for (j = 0; cs_data[i].section[j].extent != NULL; j++) { - reg_list_num++; - dws += cs_data[i].section[j].reg_count; + if (cs_data) { + /* clear state block */ + if (rdev->family >= CHIP_BONAIRE) { + rdev->rlc.clear_state_size = dws = cik_get_csb_size(rdev); + } else if (rdev->family >= CHIP_TAHITI) { + rdev->rlc.clear_state_size = si_get_csb_size(rdev); + dws = rdev->rlc.clear_state_size + (256 / 4); + } else { + reg_list_num = 0; + dws = 0; + for (i = 0; cs_data[i].section != NULL; i++) { + for (j = 0; cs_data[i].section[j].extent != NULL; j++) { + reg_list_num++; + dws += cs_data[i].section[j].reg_count; + } + } + reg_list_blk_index = (3 * reg_list_num + 2); + dws += reg_list_blk_index; + rdev->rlc.clear_state_size = dws; } - } - reg_list_blk_index = (3 * reg_list_num + 2); - dws += reg_list_blk_index; - if (rdev->rlc.clear_state_obj == NULL) { - r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true, - RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj); + if (rdev->rlc.clear_state_obj == NULL) { + r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true, + RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj); + if (r) { + dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r); + sumo_rlc_fini(rdev); + return r; + } + } + r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false); + if (unlikely(r != 0)) { + sumo_rlc_fini(rdev); + return r; + } + r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM, + &rdev->rlc.clear_state_gpu_addr); if (r) { - dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r); + radeon_bo_unreserve(rdev->rlc.clear_state_obj); + dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r); sumo_rlc_fini(rdev); return r; } - } - r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false); - if (unlikely(r != 0)) { - sumo_rlc_fini(rdev); - return r; - } - r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM, - &rdev->rlc.clear_state_gpu_addr); - if (r) { + r = radeon_bo_kmap(rdev->rlc.clear_state_obj, (void **)&rdev->rlc.cs_ptr); + if (r) { + dev_warn(rdev->dev, "(%d) map RLC c bo failed\n", r); + sumo_rlc_fini(rdev); + return r; + } + /* set up the cs buffer */ + dst_ptr = rdev->rlc.cs_ptr; + if (rdev->family >= CHIP_BONAIRE) { + cik_get_csb_buffer(rdev, dst_ptr); + } else if (rdev->family >= CHIP_TAHITI) { + reg_list_mc_addr = rdev->rlc.clear_state_gpu_addr + 256; + dst_ptr[0] = cpu_to_le32(upper_32_bits(reg_list_mc_addr)); + dst_ptr[1] = cpu_to_le32(lower_32_bits(reg_list_mc_addr)); + dst_ptr[2] = cpu_to_le32(rdev->rlc.clear_state_size); + si_get_csb_buffer(rdev, &dst_ptr[(256/4)]); + } else { + reg_list_hdr_blk_index = 0; + reg_list_mc_addr = rdev->rlc.clear_state_gpu_addr + (reg_list_blk_index * 4); + data = upper_32_bits(reg_list_mc_addr); + dst_ptr[reg_list_hdr_blk_index] = cpu_to_le32(data); + reg_list_hdr_blk_index++; + for (i = 0; cs_data[i].section != NULL; i++) { + for (j = 0; cs_data[i].section[j].extent != NULL; j++) { + reg_num = cs_data[i].section[j].reg_count; + data = reg_list_mc_addr & 0xffffffff; + dst_ptr[reg_list_hdr_blk_index] = cpu_to_le32(data); + reg_list_hdr_blk_index++; + + data = (cs_data[i].section[j].reg_index * 4) & 0xffffffff; + dst_ptr[reg_list_hdr_blk_index] = cpu_to_le32(data); + reg_list_hdr_blk_index++; + + data = 0x08000000 | (reg_num * 4); + dst_ptr[reg_list_hdr_blk_index] = cpu_to_le32(data); + reg_list_hdr_blk_index++; + + for (k = 0; k < reg_num; k++) { + data = cs_data[i].section[j].extent[k]; + dst_ptr[reg_list_blk_index + k] = cpu_to_le32(data); + } + reg_list_mc_addr += reg_num * 4; + reg_list_blk_index += reg_num; + } + } + dst_ptr[reg_list_hdr_blk_index] = cpu_to_le32(RLC_CLEAR_STATE_END_MARKER); + } + radeon_bo_kunmap(rdev->rlc.clear_state_obj); radeon_bo_unreserve(rdev->rlc.clear_state_obj); - dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r); - sumo_rlc_fini(rdev); - return r; - } - r = radeon_bo_kmap(rdev->rlc.clear_state_obj, (void **)&rdev->rlc.cs_ptr); - if (r) { - dev_warn(rdev->dev, "(%d) map RLC c bo failed\n", r); - sumo_rlc_fini(rdev); - return r; } - /* set up the cs buffer */ - dst_ptr = rdev->rlc.cs_ptr; - reg_list_hdr_blk_index = 0; - reg_list_mc_addr = rdev->rlc.clear_state_gpu_addr + (reg_list_blk_index * 4); - data = upper_32_bits(reg_list_mc_addr); - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - for (i = 0; cs_data[i].section != NULL; i++) { - for (j = 0; cs_data[i].section[j].extent != NULL; j++) { - reg_num = cs_data[i].section[j].reg_count; - data = reg_list_mc_addr & 0xffffffff; - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - - data = (cs_data[i].section[j].reg_index * 4) & 0xffffffff; - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - data = 0x08000000 | (reg_num * 4); - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - - for (k = 0; k < reg_num; k++) { - data = cs_data[i].section[j].extent[k]; - dst_ptr[reg_list_blk_index + k] = data; + if (rdev->rlc.cp_table_size) { + if (rdev->rlc.cp_table_obj == NULL) { + r = radeon_bo_create(rdev, rdev->rlc.cp_table_size, PAGE_SIZE, true, + RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.cp_table_obj); + if (r) { + dev_warn(rdev->dev, "(%d) create RLC cp table bo failed\n", r); + sumo_rlc_fini(rdev); + return r; } - reg_list_mc_addr += reg_num * 4; - reg_list_blk_index += reg_num; } - } - dst_ptr[reg_list_hdr_blk_index] = RLC_CLEAR_STATE_END_MARKER; - radeon_bo_kunmap(rdev->rlc.clear_state_obj); - radeon_bo_unreserve(rdev->rlc.clear_state_obj); + r = radeon_bo_reserve(rdev->rlc.cp_table_obj, false); + if (unlikely(r != 0)) { + dev_warn(rdev->dev, "(%d) reserve RLC cp table bo failed\n", r); + sumo_rlc_fini(rdev); + return r; + } + r = radeon_bo_pin(rdev->rlc.cp_table_obj, RADEON_GEM_DOMAIN_VRAM, + &rdev->rlc.cp_table_gpu_addr); + if (r) { + radeon_bo_unreserve(rdev->rlc.cp_table_obj); + dev_warn(rdev->dev, "(%d) pin RLC cp_table bo failed\n", r); + sumo_rlc_fini(rdev); + return r; + } + r = radeon_bo_kmap(rdev->rlc.cp_table_obj, (void **)&rdev->rlc.cp_table_ptr); + if (r) { + dev_warn(rdev->dev, "(%d) map RLC cp table bo failed\n", r); + sumo_rlc_fini(rdev); + return r; + } + + cik_init_cp_pg_table(rdev); + + radeon_bo_kunmap(rdev->rlc.cp_table_obj); + radeon_bo_unreserve(rdev->rlc.cp_table_obj); + + } return 0; } @@ -4163,8 +4341,8 @@ void evergreen_disable_interrupt_state(struct radeon_device *rdev) WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); } - /* only one DAC on DCE6 */ - if (!ASIC_IS_DCE6(rdev)) + /* only one DAC on DCE5 */ + if (!ASIC_IS_DCE5(rdev)) WREG32(DACA_AUTODETECT_INT_CONTROL, 0); WREG32(DACB_AUTODETECT_INT_CONTROL, 0); @@ -4190,7 +4368,6 @@ int evergreen_irq_set(struct radeon_device *rdev) u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0; u32 hpd1, hpd2, hpd3, hpd4, hpd5, hpd6; u32 grbm_int_cntl = 0; - u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0; u32 afmt1 = 0, afmt2 = 0, afmt3 = 0, afmt4 = 0, afmt5 = 0, afmt6 = 0; u32 dma_cntl, dma_cntl1 = 0; u32 thermal_int = 0; @@ -4373,15 +4550,21 @@ int evergreen_irq_set(struct radeon_device *rdev) WREG32(INT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, crtc6); } - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, grph1); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, grph2); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); if (rdev->num_crtc >= 4) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, grph3); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, grph4); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); } if (rdev->num_crtc >= 6) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, grph5); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, grph6); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); } WREG32(DC_HPD1_INT_CONTROL, hpd1); @@ -4573,6 +4756,7 @@ static u32 evergreen_get_ih_wptr(struct radeon_device *rdev) tmp = RREG32(IH_RB_CNTL); tmp |= IH_WPTR_OVERFLOW_CLEAR; WREG32(IH_RB_CNTL, tmp); + wptr &= ~RB_OVERFLOW; } return (wptr & rdev->ih.ptr_mask); } @@ -4586,6 +4770,7 @@ int evergreen_irq_process(struct radeon_device *rdev) bool queue_hotplug = false; bool queue_hdmi = false; bool queue_thermal = false; + u32 status, addr; if (!rdev->ih.enabled || rdev->shutdown) return IRQ_NONE; @@ -4623,7 +4808,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[0])) - radeon_crtc_handle_flip(rdev, 0); + radeon_crtc_handle_vblank(rdev, 0); rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VBLANK_INTERRUPT; DRM_DEBUG("IH: D1 vblank\n"); } @@ -4649,7 +4834,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[1])) - radeon_crtc_handle_flip(rdev, 1); + radeon_crtc_handle_vblank(rdev, 1); rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VBLANK_INTERRUPT; DRM_DEBUG("IH: D2 vblank\n"); } @@ -4675,7 +4860,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[2])) - radeon_crtc_handle_flip(rdev, 2); + radeon_crtc_handle_vblank(rdev, 2); rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VBLANK_INTERRUPT; DRM_DEBUG("IH: D3 vblank\n"); } @@ -4701,7 +4886,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[3])) - radeon_crtc_handle_flip(rdev, 3); + radeon_crtc_handle_vblank(rdev, 3); rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VBLANK_INTERRUPT; DRM_DEBUG("IH: D4 vblank\n"); } @@ -4727,7 +4912,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[4])) - radeon_crtc_handle_flip(rdev, 4); + radeon_crtc_handle_vblank(rdev, 4); rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VBLANK_INTERRUPT; DRM_DEBUG("IH: D5 vblank\n"); } @@ -4753,7 +4938,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[5])) - radeon_crtc_handle_flip(rdev, 5); + radeon_crtc_handle_vblank(rdev, 5); rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VBLANK_INTERRUPT; DRM_DEBUG("IH: D6 vblank\n"); } @@ -4769,6 +4954,15 @@ restart_ih: break; } break; + case 8: /* D1 page flip */ + case 10: /* D2 page flip */ + case 12: /* D3 page flip */ + case 14: /* D4 page flip */ + case 16: /* D5 page flip */ + case 18: /* D6 page flip */ + DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1); + radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1); + break; case 42: /* HPD hotplug */ switch (src_data) { case 0: @@ -4872,13 +5066,18 @@ restart_ih: break; case 146: case 147: + addr = RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR); + status = RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS); + /* reset addr and status */ + WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1); + if (addr == 0x0 && status == 0x0) + break; dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data); dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n", - RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR)); + addr); dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", - RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS)); - /* reset addr and status */ - WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1); + status); + cayman_vm_decode_fault(rdev, status, addr); break; case 176: /* CP_INT in ring buffer */ case 177: /* CP_INT in IB1 */ @@ -4953,143 +5152,6 @@ restart_ih: return IRQ_HANDLED; } -/** - * evergreen_dma_fence_ring_emit - emit a fence on the DMA ring - * - * @rdev: radeon_device pointer - * @fence: radeon fence object - * - * Add a DMA fence packet to the ring to write - * the fence seq number and DMA trap packet to generate - * an interrupt if needed (evergreen-SI). - */ -void evergreen_dma_fence_ring_emit(struct radeon_device *rdev, - struct radeon_fence *fence) -{ - struct radeon_ring *ring = &rdev->ring[fence->ring]; - u64 addr = rdev->fence_drv[fence->ring].gpu_addr; - /* write the fence */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_FENCE, 0, 0)); - radeon_ring_write(ring, addr & 0xfffffffc); - radeon_ring_write(ring, (upper_32_bits(addr) & 0xff)); - radeon_ring_write(ring, fence->seq); - /* generate an interrupt */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_TRAP, 0, 0)); - /* flush HDP */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0)); - radeon_ring_write(ring, (0xf << 16) | (HDP_MEM_COHERENCY_FLUSH_CNTL >> 2)); - radeon_ring_write(ring, 1); -} - -/** - * evergreen_dma_ring_ib_execute - schedule an IB on the DMA engine - * - * @rdev: radeon_device pointer - * @ib: IB object to schedule - * - * Schedule an IB in the DMA ring (evergreen). - */ -void evergreen_dma_ring_ib_execute(struct radeon_device *rdev, - struct radeon_ib *ib) -{ - struct radeon_ring *ring = &rdev->ring[ib->ring]; - - if (rdev->wb.enabled) { - u32 next_rptr = ring->wptr + 4; - while ((next_rptr & 7) != 5) - next_rptr++; - next_rptr += 3; - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 1)); - radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xff); - radeon_ring_write(ring, next_rptr); - } - - /* The indirect buffer packet must end on an 8 DW boundary in the DMA ring. - * Pad as necessary with NOPs. - */ - while ((ring->wptr & 7) != 5) - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_NOP, 0, 0)); - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_INDIRECT_BUFFER, 0, 0)); - radeon_ring_write(ring, (ib->gpu_addr & 0xFFFFFFE0)); - radeon_ring_write(ring, (ib->length_dw << 12) | (upper_32_bits(ib->gpu_addr) & 0xFF)); - -} - -/** - * evergreen_copy_dma - copy pages using the DMA engine - * - * @rdev: radeon_device pointer - * @src_offset: src GPU address - * @dst_offset: dst GPU address - * @num_gpu_pages: number of GPU pages to xfer - * @fence: radeon fence object - * - * Copy GPU paging using the DMA engine (evergreen-cayman). - * Used by the radeon ttm implementation to move pages if - * registered as the asic copy callback. - */ -int evergreen_copy_dma(struct radeon_device *rdev, - uint64_t src_offset, uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence **fence) -{ - struct radeon_semaphore *sem = NULL; - int ring_index = rdev->asic->copy.dma_ring_index; - struct radeon_ring *ring = &rdev->ring[ring_index]; - u32 size_in_dw, cur_size_in_dw; - int i, num_loops; - int r = 0; - - r = radeon_semaphore_create(rdev, &sem); - if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - return r; - } - - size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4; - num_loops = DIV_ROUND_UP(size_in_dw, 0xfffff); - r = radeon_ring_lock(rdev, ring, num_loops * 5 + 11); - if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - radeon_semaphore_free(rdev, &sem, NULL); - return r; - } - - if (radeon_fence_need_sync(*fence, ring->idx)) { - radeon_semaphore_sync_rings(rdev, sem, (*fence)->ring, - ring->idx); - radeon_fence_note_sync(*fence, ring->idx); - } else { - radeon_semaphore_free(rdev, &sem, NULL); - } - - for (i = 0; i < num_loops; i++) { - cur_size_in_dw = size_in_dw; - if (cur_size_in_dw > 0xFFFFF) - cur_size_in_dw = 0xFFFFF; - size_in_dw -= cur_size_in_dw; - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 0, cur_size_in_dw)); - radeon_ring_write(ring, dst_offset & 0xfffffffc); - radeon_ring_write(ring, src_offset & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff); - radeon_ring_write(ring, upper_32_bits(src_offset) & 0xff); - src_offset += cur_size_in_dw * 4; - dst_offset += cur_size_in_dw * 4; - } - - r = radeon_fence_emit(rdev, fence, ring->idx); - if (r) { - radeon_ring_unlock_undo(rdev, ring); - return r; - } - - radeon_ring_unlock_commit(rdev, ring); - radeon_semaphore_free(rdev, &sem, *fence); - - return r; -} - static int evergreen_startup(struct radeon_device *rdev) { struct radeon_ring *ring; @@ -5100,34 +5162,21 @@ static int evergreen_startup(struct radeon_device *rdev) /* enable aspm */ evergreen_program_aspm(rdev); - if (ASIC_IS_DCE5(rdev)) { - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) { - r = ni_init_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load firmware!\n"); - return r; - } - } + /* scratch needs to be initialized before MC */ + r = r600_vram_scratch_init(rdev); + if (r) + return r; + + evergreen_mc_program(rdev); + + if (ASIC_IS_DCE5(rdev) && !rdev->pm.dpm_enabled) { r = ni_mc_load_microcode(rdev); if (r) { DRM_ERROR("Failed to load MC firmware!\n"); return r; } - } else { - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) { - r = r600_init_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load firmware!\n"); - return r; - } - } } - r = r600_vram_scratch_init(rdev); - if (r) - return r; - - evergreen_mc_program(rdev); if (rdev->flags & RADEON_IS_AGP) { evergreen_agp_enable(rdev); } else { @@ -5137,17 +5186,11 @@ static int evergreen_startup(struct radeon_device *rdev) } evergreen_gpu_init(rdev); - r = evergreen_blit_init(rdev); - if (r) { - r600_blit_fini(rdev); - rdev->asic->copy.copy = NULL; - dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); - } - /* allocate rlc buffers */ if (rdev->flags & RADEON_IS_IGP) { rdev->rlc.reg_list = sumo_rlc_save_restore_register_list; - rdev->rlc.reg_list_size = sumo_rlc_save_restore_register_list_size; + rdev->rlc.reg_list_size = + (u32)ARRAY_SIZE(sumo_rlc_save_restore_register_list); rdev->rlc.cs_data = evergreen_cs_data; r = sumo_rlc_init(rdev); if (r) { @@ -5173,7 +5216,7 @@ static int evergreen_startup(struct radeon_device *rdev) return r; } - r = rv770_uvd_resume(rdev); + r = uvd_v2_2_resume(rdev); if (!r) { r = radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_UVD_INDEX); @@ -5201,15 +5244,13 @@ static int evergreen_startup(struct radeon_device *rdev) ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, - R600_CP_RB_RPTR, R600_CP_RB_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + RADEON_CP_PACKET2); if (r) return r; ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET, - DMA_RB_RPTR, DMA_RB_WPTR, - 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0)); + DMA_PACKET(DMA_PACKET_NOP, 0, 0)); if (r) return r; @@ -5225,12 +5266,10 @@ static int evergreen_startup(struct radeon_device *rdev) ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; if (ring->ring_size) { - r = radeon_ring_init(rdev, ring, ring->ring_size, - R600_WB_UVD_RPTR_OFFSET, - UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, + RADEON_CP_PACKET2); if (!r) - r = r600_uvd_init(rdev); + r = uvd_v1_0_init(rdev); if (r) DRM_ERROR("radeon: error initializing UVD (%d).\n", r); @@ -5270,6 +5309,9 @@ int evergreen_resume(struct radeon_device *rdev) /* init golden registers */ evergreen_init_golden_registers(rdev); + if (rdev->pm.pm_method == PM_METHOD_DPM) + radeon_pm_resume(rdev); + rdev->accel_working = true; r = evergreen_startup(rdev); if (r) { @@ -5284,11 +5326,12 @@ int evergreen_resume(struct radeon_device *rdev) int evergreen_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r600_audio_fini(rdev); + uvd_v1_0_fini(rdev); radeon_uvd_suspend(rdev); r700_cp_stop(rdev); r600_dma_stop(rdev); - r600_uvd_rbc_stop(rdev); evergreen_irq_suspend(rdev); radeon_wb_disable(rdev); evergreen_pcie_gart_disable(rdev); @@ -5360,6 +5403,27 @@ int evergreen_init(struct radeon_device *rdev) if (r) return r; + if (ASIC_IS_DCE5(rdev)) { + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) { + r = ni_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + } else { + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) { + r = r600_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + } + + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL; r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024); @@ -5412,8 +5476,8 @@ int evergreen_init(struct radeon_device *rdev) void evergreen_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); r600_audio_fini(rdev); - r600_blit_fini(rdev); r700_cp_fini(rdev); r600_dma_fini(rdev); r600_irq_fini(rdev); @@ -5422,8 +5486,9 @@ void evergreen_fini(struct radeon_device *rdev) radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); radeon_irq_kms_fini(rdev); - evergreen_pcie_gart_fini(rdev); + uvd_v1_0_fini(rdev); radeon_uvd_fini(rdev); + evergreen_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_gem_fini(rdev); radeon_fence_driver_fini(rdev); @@ -5509,6 +5574,9 @@ void evergreen_program_aspm(struct radeon_device *rdev) */ bool fusion_platform = false; + if (radeon_aspm == 0) + return; + if (!(rdev->flags & RADEON_IS_PCIE)) return; diff --git a/drivers/gpu/drm/radeon/evergreen_blit_kms.c b/drivers/gpu/drm/radeon/evergreen_blit_kms.c deleted file mode 100644 index 057c87b6515..00000000000 --- a/drivers/gpu/drm/radeon/evergreen_blit_kms.c +++ /dev/null @@ -1,729 +0,0 @@ -/* - * Copyright 2010 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - * - * Authors: - * Alex Deucher <alexander.deucher@amd.com> - */ - -#include <drm/drmP.h> -#include <drm/radeon_drm.h> -#include "radeon.h" - -#include "evergreend.h" -#include "evergreen_blit_shaders.h" -#include "cayman_blit_shaders.h" -#include "radeon_blit_common.h" - -/* emits 17 */ -static void -set_render_target(struct radeon_device *rdev, int format, - int w, int h, u64 gpu_addr) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u32 cb_color_info; - int pitch, slice; - - h = ALIGN(h, 8); - if (h < 8) - h = 8; - - cb_color_info = CB_FORMAT(format) | - CB_SOURCE_FORMAT(CB_SF_EXPORT_NORM) | - CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1); - pitch = (w / 8) - 1; - slice = ((w * h) / 64) - 1; - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 15)); - radeon_ring_write(ring, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(ring, gpu_addr >> 8); - radeon_ring_write(ring, pitch); - radeon_ring_write(ring, slice); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, cb_color_info); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, (w - 1) | ((h - 1) << 16)); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); -} - -/* emits 5dw */ -static void -cp_set_surface_sync(struct radeon_device *rdev, - u32 sync_type, u32 size, - u64 mc_addr) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u32 cp_coher_size; - - if (size == 0xffffffff) - cp_coher_size = 0xffffffff; - else - cp_coher_size = ((size + 255) >> 8); - - if (rdev->family >= CHIP_CAYMAN) { - /* CP_COHER_CNTL2 has to be set manually when submitting a surface_sync - * to the RB directly. For IBs, the CP programs this as part of the - * surface_sync packet. - */ - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(ring, (0x85e8 - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(ring, 0); /* CP_COHER_CNTL2 */ - } - radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(ring, sync_type); - radeon_ring_write(ring, cp_coher_size); - radeon_ring_write(ring, mc_addr >> 8); - radeon_ring_write(ring, 10); /* poll interval */ -} - -/* emits 11dw + 1 surface sync = 16dw */ -static void -set_shaders(struct radeon_device *rdev) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u64 gpu_addr; - - /* VS */ - gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset; - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 3)); - radeon_ring_write(ring, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(ring, gpu_addr >> 8); - radeon_ring_write(ring, 2); - radeon_ring_write(ring, 0); - - /* PS */ - gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.ps_offset; - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 4)); - radeon_ring_write(ring, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(ring, gpu_addr >> 8); - radeon_ring_write(ring, 1); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 2); - - gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset; - cp_set_surface_sync(rdev, PACKET3_SH_ACTION_ENA, 512, gpu_addr); -} - -/* emits 10 + 1 sync (5) = 15 */ -static void -set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u32 sq_vtx_constant_word2, sq_vtx_constant_word3; - - /* high addr, stride */ - sq_vtx_constant_word2 = SQ_VTXC_BASE_ADDR_HI(upper_32_bits(gpu_addr) & 0xff) | - SQ_VTXC_STRIDE(16); -#ifdef __BIG_ENDIAN - sq_vtx_constant_word2 |= SQ_VTXC_ENDIAN_SWAP(SQ_ENDIAN_8IN32); -#endif - /* xyzw swizzles */ - sq_vtx_constant_word3 = SQ_VTCX_SEL_X(SQ_SEL_X) | - SQ_VTCX_SEL_Y(SQ_SEL_Y) | - SQ_VTCX_SEL_Z(SQ_SEL_Z) | - SQ_VTCX_SEL_W(SQ_SEL_W); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 8)); - radeon_ring_write(ring, 0x580); - radeon_ring_write(ring, gpu_addr & 0xffffffff); - radeon_ring_write(ring, 48 - 1); /* size */ - radeon_ring_write(ring, sq_vtx_constant_word2); - radeon_ring_write(ring, sq_vtx_constant_word3); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, S__SQ_CONSTANT_TYPE(SQ_TEX_VTX_VALID_BUFFER)); - - if ((rdev->family == CHIP_CEDAR) || - (rdev->family == CHIP_PALM) || - (rdev->family == CHIP_SUMO) || - (rdev->family == CHIP_SUMO2) || - (rdev->family == CHIP_CAICOS)) - cp_set_surface_sync(rdev, - PACKET3_TC_ACTION_ENA, 48, gpu_addr); - else - cp_set_surface_sync(rdev, - PACKET3_VC_ACTION_ENA, 48, gpu_addr); - -} - -/* emits 10 */ -static void -set_tex_resource(struct radeon_device *rdev, - int format, int w, int h, int pitch, - u64 gpu_addr, u32 size) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u32 sq_tex_resource_word0, sq_tex_resource_word1; - u32 sq_tex_resource_word4, sq_tex_resource_word7; - - if (h < 1) - h = 1; - - sq_tex_resource_word0 = TEX_DIM(SQ_TEX_DIM_2D); - sq_tex_resource_word0 |= ((((pitch >> 3) - 1) << 6) | - ((w - 1) << 18)); - sq_tex_resource_word1 = ((h - 1) << 0) | - TEX_ARRAY_MODE(ARRAY_1D_TILED_THIN1); - /* xyzw swizzles */ - sq_tex_resource_word4 = TEX_DST_SEL_X(SQ_SEL_X) | - TEX_DST_SEL_Y(SQ_SEL_Y) | - TEX_DST_SEL_Z(SQ_SEL_Z) | - TEX_DST_SEL_W(SQ_SEL_W); - - sq_tex_resource_word7 = format | - S__SQ_CONSTANT_TYPE(SQ_TEX_VTX_VALID_TEXTURE); - - cp_set_surface_sync(rdev, - PACKET3_TC_ACTION_ENA, size, gpu_addr); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 8)); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, sq_tex_resource_word0); - radeon_ring_write(ring, sq_tex_resource_word1); - radeon_ring_write(ring, gpu_addr >> 8); - radeon_ring_write(ring, gpu_addr >> 8); - radeon_ring_write(ring, sq_tex_resource_word4); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, sq_tex_resource_word7); -} - -/* emits 12 */ -static void -set_scissors(struct radeon_device *rdev, int x1, int y1, - int x2, int y2) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - /* workaround some hw bugs */ - if (x2 == 0) - x1 = 1; - if (y2 == 0) - y1 = 1; - if (rdev->family >= CHIP_CAYMAN) { - if ((x2 == 1) && (y2 == 1)) - x2 = 2; - } - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(ring, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(ring, (x1 << 0) | (y1 << 16)); - radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(ring, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31)); - radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(ring, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31)); - radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); -} - -/* emits 10 */ -static void -draw_auto(struct radeon_device *rdev) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(ring, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(ring, DI_PT_RECTLIST); - - radeon_ring_write(ring, PACKET3(PACKET3_INDEX_TYPE, 0)); - radeon_ring_write(ring, -#ifdef __BIG_ENDIAN - (2 << 2) | -#endif - DI_INDEX_SIZE_16_BIT); - - radeon_ring_write(ring, PACKET3(PACKET3_NUM_INSTANCES, 0)); - radeon_ring_write(ring, 1); - - radeon_ring_write(ring, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1)); - radeon_ring_write(ring, 3); - radeon_ring_write(ring, DI_SRC_SEL_AUTO_INDEX); - -} - -/* emits 39 */ -static void -set_default_state(struct radeon_device *rdev) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u32 sq_config, sq_gpr_resource_mgmt_1, sq_gpr_resource_mgmt_2, sq_gpr_resource_mgmt_3; - u32 sq_thread_resource_mgmt, sq_thread_resource_mgmt_2; - u32 sq_stack_resource_mgmt_1, sq_stack_resource_mgmt_2, sq_stack_resource_mgmt_3; - int num_ps_gprs, num_vs_gprs, num_temp_gprs; - int num_gs_gprs, num_es_gprs, num_hs_gprs, num_ls_gprs; - int num_ps_threads, num_vs_threads, num_gs_threads, num_es_threads; - int num_hs_threads, num_ls_threads; - int num_ps_stack_entries, num_vs_stack_entries, num_gs_stack_entries, num_es_stack_entries; - int num_hs_stack_entries, num_ls_stack_entries; - u64 gpu_addr; - int dwords; - - /* set clear context state */ - radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0)); - radeon_ring_write(ring, 0); - - if (rdev->family < CHIP_CAYMAN) { - switch (rdev->family) { - case CHIP_CEDAR: - default: - num_ps_gprs = 93; - num_vs_gprs = 46; - num_temp_gprs = 4; - num_gs_gprs = 31; - num_es_gprs = 31; - num_hs_gprs = 23; - num_ls_gprs = 23; - num_ps_threads = 96; - num_vs_threads = 16; - num_gs_threads = 16; - num_es_threads = 16; - num_hs_threads = 16; - num_ls_threads = 16; - num_ps_stack_entries = 42; - num_vs_stack_entries = 42; - num_gs_stack_entries = 42; - num_es_stack_entries = 42; - num_hs_stack_entries = 42; - num_ls_stack_entries = 42; - break; - case CHIP_REDWOOD: - num_ps_gprs = 93; - num_vs_gprs = 46; - num_temp_gprs = 4; - num_gs_gprs = 31; - num_es_gprs = 31; - num_hs_gprs = 23; - num_ls_gprs = 23; - num_ps_threads = 128; - num_vs_threads = 20; - num_gs_threads = 20; - num_es_threads = 20; - num_hs_threads = 20; - num_ls_threads = 20; - num_ps_stack_entries = 42; - num_vs_stack_entries = 42; - num_gs_stack_entries = 42; - num_es_stack_entries = 42; - num_hs_stack_entries = 42; - num_ls_stack_entries = 42; - break; - case CHIP_JUNIPER: - num_ps_gprs = 93; - num_vs_gprs = 46; - num_temp_gprs = 4; - num_gs_gprs = 31; - num_es_gprs = 31; - num_hs_gprs = 23; - num_ls_gprs = 23; - num_ps_threads = 128; - num_vs_threads = 20; - num_gs_threads = 20; - num_es_threads = 20; - num_hs_threads = 20; - num_ls_threads = 20; - num_ps_stack_entries = 85; - num_vs_stack_entries = 85; - num_gs_stack_entries = 85; - num_es_stack_entries = 85; - num_hs_stack_entries = 85; - num_ls_stack_entries = 85; - break; - case CHIP_CYPRESS: - case CHIP_HEMLOCK: - num_ps_gprs = 93; - num_vs_gprs = 46; - num_temp_gprs = 4; - num_gs_gprs = 31; - num_es_gprs = 31; - num_hs_gprs = 23; - num_ls_gprs = 23; - num_ps_threads = 128; - num_vs_threads = 20; - num_gs_threads = 20; - num_es_threads = 20; - num_hs_threads = 20; - num_ls_threads = 20; - num_ps_stack_entries = 85; - num_vs_stack_entries = 85; - num_gs_stack_entries = 85; - num_es_stack_entries = 85; - num_hs_stack_entries = 85; - num_ls_stack_entries = 85; - break; - case CHIP_PALM: - num_ps_gprs = 93; - num_vs_gprs = 46; - num_temp_gprs = 4; - num_gs_gprs = 31; - num_es_gprs = 31; - num_hs_gprs = 23; - num_ls_gprs = 23; - num_ps_threads = 96; - num_vs_threads = 16; - num_gs_threads = 16; - num_es_threads = 16; - num_hs_threads = 16; - num_ls_threads = 16; - num_ps_stack_entries = 42; - num_vs_stack_entries = 42; - num_gs_stack_entries = 42; - num_es_stack_entries = 42; - num_hs_stack_entries = 42; - num_ls_stack_entries = 42; - break; - case CHIP_SUMO: - num_ps_gprs = 93; - num_vs_gprs = 46; - num_temp_gprs = 4; - num_gs_gprs = 31; - num_es_gprs = 31; - num_hs_gprs = 23; - num_ls_gprs = 23; - num_ps_threads = 96; - num_vs_threads = 25; - num_gs_threads = 25; - num_es_threads = 25; - num_hs_threads = 25; - num_ls_threads = 25; - num_ps_stack_entries = 42; - num_vs_stack_entries = 42; - num_gs_stack_entries = 42; - num_es_stack_entries = 42; - num_hs_stack_entries = 42; - num_ls_stack_entries = 42; - break; - case CHIP_SUMO2: - num_ps_gprs = 93; - num_vs_gprs = 46; - num_temp_gprs = 4; - num_gs_gprs = 31; - num_es_gprs = 31; - num_hs_gprs = 23; - num_ls_gprs = 23; - num_ps_threads = 96; - num_vs_threads = 25; - num_gs_threads = 25; - num_es_threads = 25; - num_hs_threads = 25; - num_ls_threads = 25; - num_ps_stack_entries = 85; - num_vs_stack_entries = 85; - num_gs_stack_entries = 85; - num_es_stack_entries = 85; - num_hs_stack_entries = 85; - num_ls_stack_entries = 85; - break; - case CHIP_BARTS: - num_ps_gprs = 93; - num_vs_gprs = 46; - num_temp_gprs = 4; - num_gs_gprs = 31; - num_es_gprs = 31; - num_hs_gprs = 23; - num_ls_gprs = 23; - num_ps_threads = 128; - num_vs_threads = 20; - num_gs_threads = 20; - num_es_threads = 20; - num_hs_threads = 20; - num_ls_threads = 20; - num_ps_stack_entries = 85; - num_vs_stack_entries = 85; - num_gs_stack_entries = 85; - num_es_stack_entries = 85; - num_hs_stack_entries = 85; - num_ls_stack_entries = 85; - break; - case CHIP_TURKS: - num_ps_gprs = 93; - num_vs_gprs = 46; - num_temp_gprs = 4; - num_gs_gprs = 31; - num_es_gprs = 31; - num_hs_gprs = 23; - num_ls_gprs = 23; - num_ps_threads = 128; - num_vs_threads = 20; - num_gs_threads = 20; - num_es_threads = 20; - num_hs_threads = 20; - num_ls_threads = 20; - num_ps_stack_entries = 42; - num_vs_stack_entries = 42; - num_gs_stack_entries = 42; - num_es_stack_entries = 42; - num_hs_stack_entries = 42; - num_ls_stack_entries = 42; - break; - case CHIP_CAICOS: - num_ps_gprs = 93; - num_vs_gprs = 46; - num_temp_gprs = 4; - num_gs_gprs = 31; - num_es_gprs = 31; - num_hs_gprs = 23; - num_ls_gprs = 23; - num_ps_threads = 128; - num_vs_threads = 10; - num_gs_threads = 10; - num_es_threads = 10; - num_hs_threads = 10; - num_ls_threads = 10; - num_ps_stack_entries = 42; - num_vs_stack_entries = 42; - num_gs_stack_entries = 42; - num_es_stack_entries = 42; - num_hs_stack_entries = 42; - num_ls_stack_entries = 42; - break; - } - - if ((rdev->family == CHIP_CEDAR) || - (rdev->family == CHIP_PALM) || - (rdev->family == CHIP_SUMO) || - (rdev->family == CHIP_SUMO2) || - (rdev->family == CHIP_CAICOS)) - sq_config = 0; - else - sq_config = VC_ENABLE; - - sq_config |= (EXPORT_SRC_C | - CS_PRIO(0) | - LS_PRIO(0) | - HS_PRIO(0) | - PS_PRIO(0) | - VS_PRIO(1) | - GS_PRIO(2) | - ES_PRIO(3)); - - sq_gpr_resource_mgmt_1 = (NUM_PS_GPRS(num_ps_gprs) | - NUM_VS_GPRS(num_vs_gprs) | - NUM_CLAUSE_TEMP_GPRS(num_temp_gprs)); - sq_gpr_resource_mgmt_2 = (NUM_GS_GPRS(num_gs_gprs) | - NUM_ES_GPRS(num_es_gprs)); - sq_gpr_resource_mgmt_3 = (NUM_HS_GPRS(num_hs_gprs) | - NUM_LS_GPRS(num_ls_gprs)); - sq_thread_resource_mgmt = (NUM_PS_THREADS(num_ps_threads) | - NUM_VS_THREADS(num_vs_threads) | - NUM_GS_THREADS(num_gs_threads) | - NUM_ES_THREADS(num_es_threads)); - sq_thread_resource_mgmt_2 = (NUM_HS_THREADS(num_hs_threads) | - NUM_LS_THREADS(num_ls_threads)); - sq_stack_resource_mgmt_1 = (NUM_PS_STACK_ENTRIES(num_ps_stack_entries) | - NUM_VS_STACK_ENTRIES(num_vs_stack_entries)); - sq_stack_resource_mgmt_2 = (NUM_GS_STACK_ENTRIES(num_gs_stack_entries) | - NUM_ES_STACK_ENTRIES(num_es_stack_entries)); - sq_stack_resource_mgmt_3 = (NUM_HS_STACK_ENTRIES(num_hs_stack_entries) | - NUM_LS_STACK_ENTRIES(num_ls_stack_entries)); - - /* disable dyn gprs */ - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(ring, (SQ_DYN_GPR_CNTL_PS_FLUSH_REQ - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(ring, 0); - - /* setup LDS */ - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(ring, (SQ_LDS_RESOURCE_MGMT - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(ring, 0x10001000); - - /* SQ config */ - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 11)); - radeon_ring_write(ring, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(ring, sq_config); - radeon_ring_write(ring, sq_gpr_resource_mgmt_1); - radeon_ring_write(ring, sq_gpr_resource_mgmt_2); - radeon_ring_write(ring, sq_gpr_resource_mgmt_3); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, sq_thread_resource_mgmt); - radeon_ring_write(ring, sq_thread_resource_mgmt_2); - radeon_ring_write(ring, sq_stack_resource_mgmt_1); - radeon_ring_write(ring, sq_stack_resource_mgmt_2); - radeon_ring_write(ring, sq_stack_resource_mgmt_3); - } - - /* CONTEXT_CONTROL */ - radeon_ring_write(ring, 0xc0012800); - radeon_ring_write(ring, 0x80000000); - radeon_ring_write(ring, 0x80000000); - - /* SQ_VTX_BASE_VTX_LOC */ - radeon_ring_write(ring, 0xc0026f00); - radeon_ring_write(ring, 0x00000000); - radeon_ring_write(ring, 0x00000000); - radeon_ring_write(ring, 0x00000000); - - /* SET_SAMPLER */ - radeon_ring_write(ring, 0xc0036e00); - radeon_ring_write(ring, 0x00000000); - radeon_ring_write(ring, 0x00000012); - radeon_ring_write(ring, 0x00000000); - radeon_ring_write(ring, 0x00000000); - - /* set to DX10/11 mode */ - radeon_ring_write(ring, PACKET3(PACKET3_MODE_CONTROL, 0)); - radeon_ring_write(ring, 1); - - /* emit an IB pointing at default state */ - dwords = ALIGN(rdev->r600_blit.state_len, 0x10); - gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.state_offset; - radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); - radeon_ring_write(ring, gpu_addr & 0xFFFFFFFC); - radeon_ring_write(ring, upper_32_bits(gpu_addr) & 0xFF); - radeon_ring_write(ring, dwords); - -} - -int evergreen_blit_init(struct radeon_device *rdev) -{ - u32 obj_size; - int i, r, dwords; - void *ptr; - u32 packet2s[16]; - int num_packet2s = 0; - - rdev->r600_blit.primitives.set_render_target = set_render_target; - rdev->r600_blit.primitives.cp_set_surface_sync = cp_set_surface_sync; - rdev->r600_blit.primitives.set_shaders = set_shaders; - rdev->r600_blit.primitives.set_vtx_resource = set_vtx_resource; - rdev->r600_blit.primitives.set_tex_resource = set_tex_resource; - rdev->r600_blit.primitives.set_scissors = set_scissors; - rdev->r600_blit.primitives.draw_auto = draw_auto; - rdev->r600_blit.primitives.set_default_state = set_default_state; - - rdev->r600_blit.ring_size_common = 8; /* sync semaphore */ - rdev->r600_blit.ring_size_common += 55; /* shaders + def state */ - rdev->r600_blit.ring_size_common += 16; /* fence emit for VB IB */ - rdev->r600_blit.ring_size_common += 5; /* done copy */ - rdev->r600_blit.ring_size_common += 16; /* fence emit for done copy */ - - rdev->r600_blit.ring_size_per_loop = 74; - if (rdev->family >= CHIP_CAYMAN) - rdev->r600_blit.ring_size_per_loop += 9; /* additional DWs for surface sync */ - - rdev->r600_blit.max_dim = 16384; - - rdev->r600_blit.state_offset = 0; - - if (rdev->family < CHIP_CAYMAN) - rdev->r600_blit.state_len = evergreen_default_size; - else - rdev->r600_blit.state_len = cayman_default_size; - - dwords = rdev->r600_blit.state_len; - while (dwords & 0xf) { - packet2s[num_packet2s++] = cpu_to_le32(PACKET2(0)); - dwords++; - } - - obj_size = dwords * 4; - obj_size = ALIGN(obj_size, 256); - - rdev->r600_blit.vs_offset = obj_size; - if (rdev->family < CHIP_CAYMAN) - obj_size += evergreen_vs_size * 4; - else - obj_size += cayman_vs_size * 4; - obj_size = ALIGN(obj_size, 256); - - rdev->r600_blit.ps_offset = obj_size; - if (rdev->family < CHIP_CAYMAN) - obj_size += evergreen_ps_size * 4; - else - obj_size += cayman_ps_size * 4; - obj_size = ALIGN(obj_size, 256); - - /* pin copy shader into vram if not already initialized */ - if (!rdev->r600_blit.shader_obj) { - r = radeon_bo_create(rdev, obj_size, PAGE_SIZE, true, - RADEON_GEM_DOMAIN_VRAM, - NULL, &rdev->r600_blit.shader_obj); - if (r) { - DRM_ERROR("evergreen failed to allocate shader\n"); - return r; - } - - r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); - if (unlikely(r != 0)) - return r; - r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM, - &rdev->r600_blit.shader_gpu_addr); - radeon_bo_unreserve(rdev->r600_blit.shader_obj); - if (r) { - dev_err(rdev->dev, "(%d) pin blit object failed\n", r); - return r; - } - } - - DRM_DEBUG("evergreen blit allocated bo %08x vs %08x ps %08x\n", - obj_size, - rdev->r600_blit.vs_offset, rdev->r600_blit.ps_offset); - - r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); - if (unlikely(r != 0)) - return r; - r = radeon_bo_kmap(rdev->r600_blit.shader_obj, &ptr); - if (r) { - DRM_ERROR("failed to map blit object %d\n", r); - return r; - } - - if (rdev->family < CHIP_CAYMAN) { - memcpy_toio(ptr + rdev->r600_blit.state_offset, - evergreen_default_state, rdev->r600_blit.state_len * 4); - - if (num_packet2s) - memcpy_toio(ptr + rdev->r600_blit.state_offset + (rdev->r600_blit.state_len * 4), - packet2s, num_packet2s * 4); - for (i = 0; i < evergreen_vs_size; i++) - *(u32 *)((unsigned long)ptr + rdev->r600_blit.vs_offset + i * 4) = cpu_to_le32(evergreen_vs[i]); - for (i = 0; i < evergreen_ps_size; i++) - *(u32 *)((unsigned long)ptr + rdev->r600_blit.ps_offset + i * 4) = cpu_to_le32(evergreen_ps[i]); - } else { - memcpy_toio(ptr + rdev->r600_blit.state_offset, - cayman_default_state, rdev->r600_blit.state_len * 4); - - if (num_packet2s) - memcpy_toio(ptr + rdev->r600_blit.state_offset + (rdev->r600_blit.state_len * 4), - packet2s, num_packet2s * 4); - for (i = 0; i < cayman_vs_size; i++) - *(u32 *)((unsigned long)ptr + rdev->r600_blit.vs_offset + i * 4) = cpu_to_le32(cayman_vs[i]); - for (i = 0; i < cayman_ps_size; i++) - *(u32 *)((unsigned long)ptr + rdev->r600_blit.ps_offset + i * 4) = cpu_to_le32(cayman_ps[i]); - } - radeon_bo_kunmap(rdev->r600_blit.shader_obj); - radeon_bo_unreserve(rdev->r600_blit.shader_obj); - - radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); - return 0; -} diff --git a/drivers/gpu/drm/radeon/evergreen_blit_shaders.c b/drivers/gpu/drm/radeon/evergreen_blit_shaders.c index f85c0af115b..d43383470cd 100644 --- a/drivers/gpu/drm/radeon/evergreen_blit_shaders.c +++ b/drivers/gpu/drm/radeon/evergreen_blit_shaders.c @@ -300,58 +300,4 @@ const u32 evergreen_default_state[] = 0x00000010, /* */ }; -const u32 evergreen_vs[] = -{ - 0x00000004, - 0x80800400, - 0x0000a03c, - 0x95000688, - 0x00004000, - 0x15200688, - 0x00000000, - 0x00000000, - 0x3c000000, - 0x67961001, -#ifdef __BIG_ENDIAN - 0x000a0000, -#else - 0x00080000, -#endif - 0x00000000, - 0x1c000000, - 0x67961000, -#ifdef __BIG_ENDIAN - 0x00020008, -#else - 0x00000008, -#endif - 0x00000000, -}; - -const u32 evergreen_ps[] = -{ - 0x00000003, - 0xa00c0000, - 0x00000008, - 0x80400000, - 0x00000000, - 0x95200688, - 0x00380400, - 0x00146b10, - 0x00380000, - 0x20146b10, - 0x00380400, - 0x40146b00, - 0x80380000, - 0x60146b00, - 0x00000000, - 0x00000000, - 0x00000010, - 0x000d1000, - 0xb0800000, - 0x00000000, -}; - -const u32 evergreen_ps_size = ARRAY_SIZE(evergreen_ps); -const u32 evergreen_vs_size = ARRAY_SIZE(evergreen_vs); const u32 evergreen_default_size = ARRAY_SIZE(evergreen_default_state); diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c index eb8ac315f92..5c8b358f9fb 100644 --- a/drivers/gpu/drm/radeon/evergreen_cs.c +++ b/drivers/gpu/drm/radeon/evergreen_cs.c @@ -967,7 +967,10 @@ static int evergreen_cs_track_check(struct radeon_cs_parser *p) if (track->cb_dirty) { tmp = track->cb_target_mask; for (i = 0; i < 8; i++) { - if ((tmp >> (i * 4)) & 0xF) { + u32 format = G_028C70_FORMAT(track->cb_color_info[i]); + + if (format != V_028C70_COLOR_INVALID && + (tmp >> (i * 4)) & 0xF) { /* at least one component is enabled */ if (track->cb_color_bo[i] == NULL) { dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n", @@ -1162,7 +1165,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case DB_DEPTH_CONTROL: track->db_depth_control = radeon_get_ib_value(p, idx); @@ -1193,12 +1196,12 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } ib[idx] &= ~Z_ARRAY_MODE(0xf); track->db_z_info &= ~Z_ARRAY_MODE(0xf); - ib[idx] |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); - track->db_z_info |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + ib[idx] |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); + track->db_z_info |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); + if (reloc->tiling_flags & RADEON_TILING_MACRO) { unsigned bankw, bankh, mtaspect, tile_split; - evergreen_tiling_fields(reloc->lobj.tiling_flags, + evergreen_tiling_fields(reloc->tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); ib[idx] |= DB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); @@ -1234,7 +1237,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->db_z_read_offset = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->db_z_read_bo = reloc->robj; track->db_dirty = true; break; @@ -1246,7 +1249,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->db_z_write_offset = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->db_z_write_bo = reloc->robj; track->db_dirty = true; break; @@ -1258,7 +1261,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->db_s_read_offset = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->db_s_read_bo = reloc->robj; track->db_dirty = true; break; @@ -1270,7 +1273,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->db_s_write_offset = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->db_s_write_bo = reloc->robj; track->db_dirty = true; break; @@ -1294,7 +1297,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = (reg - VGT_STRMOUT_BUFFER_BASE_0) / 16; track->vgt_strmout_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->vgt_strmout_bo[tmp] = reloc->robj; track->streamout_dirty = true; break; @@ -1314,7 +1317,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); case CB_TARGET_MASK: track->cb_target_mask = radeon_get_ib_value(p, idx); track->cb_dirty = true; @@ -1378,8 +1381,8 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); - track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); + ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); + track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); } track->cb_dirty = true; break; @@ -1396,8 +1399,8 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); - track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); + ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); + track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); } track->cb_dirty = true; break; @@ -1458,10 +1461,10 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) { unsigned bankw, bankh, mtaspect, tile_split; - evergreen_tiling_fields(reloc->lobj.tiling_flags, + evergreen_tiling_fields(reloc->tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); @@ -1486,10 +1489,10 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) { unsigned bankw, bankh, mtaspect, tile_split; - evergreen_tiling_fields(reloc->lobj.tiling_flags, + evergreen_tiling_fields(reloc->tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); @@ -1517,7 +1520,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->cb_color_fmask_bo[tmp] = reloc->robj; break; case CB_COLOR0_CMASK: @@ -1534,7 +1537,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->cb_color_cmask_bo[tmp] = reloc->robj; break; case CB_COLOR0_FMASK_SLICE: @@ -1575,7 +1578,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = (reg - CB_COLOR0_BASE) / 0x3c; track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->cb_color_bo[tmp] = reloc->robj; track->cb_dirty = true; break; @@ -1591,7 +1594,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = ((reg - CB_COLOR8_BASE) / 0x1c) + 8; track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->cb_color_bo[tmp] = reloc->robj; track->cb_dirty = true; break; @@ -1603,7 +1606,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->htile_offset = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->htile_bo = reloc->robj; track->db_dirty = true; break; @@ -1720,7 +1723,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case SX_MEMORY_EXPORT_BASE: if (p->rdev->family >= CHIP_CAYMAN) { @@ -1734,7 +1737,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case CAYMAN_SX_SCATTER_EXPORT_BASE: if (p->rdev->family < CHIP_CAYMAN) { @@ -1748,7 +1751,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case SX_MISC: track->sx_misc_kill_all_prims = (radeon_get_ib_value(p, idx) & 0x1) != 0; @@ -1833,7 +1836,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (idx_value & 0xfffffff0) + ((u64)(tmp & 0xff) << 32); @@ -1879,7 +1882,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + idx_value + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); @@ -1906,7 +1909,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + idx_value + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); @@ -1934,7 +1937,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + radeon_get_ib_value(p, idx+1) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -2024,7 +2027,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad DISPATCH_INDIRECT\n"); return -EINVAL; } - ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); + ib[idx+0] = idx_value + (u32)(reloc->gpu_offset & 0xffffffff); r = evergreen_cs_track_check(p); if (r) { dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); @@ -2046,7 +2049,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -2103,7 +2106,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, tmp = radeon_get_ib_value(p, idx) + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); - offset = reloc->lobj.gpu_offset + tmp; + offset = reloc->gpu_offset + tmp; if ((tmp + size) > radeon_bo_size(reloc->robj)) { dev_warn(p->dev, "CP DMA src buffer too small (%llu %lu)\n", @@ -2141,7 +2144,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, tmp = radeon_get_ib_value(p, idx+2) + ((u64)(radeon_get_ib_value(p, idx+3) & 0xff) << 32); - offset = reloc->lobj.gpu_offset + tmp; + offset = reloc->gpu_offset + tmp; if ((tmp + size) > radeon_bo_size(reloc->robj)) { dev_warn(p->dev, "CP DMA dst buffer too small (%llu %lu)\n", @@ -2171,7 +2174,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SURFACE_SYNC\n"); return -EINVAL; } - ib[idx+2] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx+2] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); } break; case PACKET3_EVENT_WRITE: @@ -2187,7 +2190,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad EVENT_WRITE\n"); return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffff8) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -2209,7 +2212,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -2231,7 +2234,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -2299,11 +2302,11 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { ib[idx+1+(i*8)+1] |= - TEX_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + TEX_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); + if (reloc->tiling_flags & RADEON_TILING_MACRO) { unsigned bankw, bankh, mtaspect, tile_split; - evergreen_tiling_fields(reloc->lobj.tiling_flags, + evergreen_tiling_fields(reloc->tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); ib[idx+1+(i*8)+6] |= TEX_TILE_SPLIT(tile_split); @@ -2315,7 +2318,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, } } texture = reloc->robj; - toffset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + toffset = (u32)((reloc->gpu_offset >> 8) & 0xffffffff); /* tex mip base */ tex_dim = ib[idx+1+(i*8)+0] & 0x7; @@ -2334,7 +2337,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SET_RESOURCE (tex)\n"); return -EINVAL; } - moffset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + moffset = (u32)((reloc->gpu_offset >> 8) & 0xffffffff); mipmap = reloc->robj; } @@ -2361,7 +2364,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, ib[idx+1+(i*8)+1] = radeon_bo_size(reloc->robj) - offset; } - offset64 = reloc->lobj.gpu_offset + offset; + offset64 = reloc->gpu_offset + offset; ib[idx+1+(i*8)+0] = offset64; ib[idx+1+(i*8)+2] = (ib[idx+1+(i*8)+2] & 0xffffff00) | (upper_32_bits(offset64) & 0xff); @@ -2442,7 +2445,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+1] = offset; ib[idx+2] = upper_32_bits(offset) & 0xff; } @@ -2461,7 +2464,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+3] = offset; ib[idx+4] = upper_32_bits(offset) & 0xff; } @@ -2490,7 +2493,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, offset + 8, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+0] = offset; ib[idx+1] = upper_32_bits(offset) & 0xff; break; @@ -2515,7 +2518,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+1] = offset; ib[idx+2] = upper_32_bits(offset) & 0xff; } else { @@ -2539,7 +2542,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+3] = offset; ib[idx+4] = upper_32_bits(offset) & 0xff; } else { @@ -2714,7 +2717,7 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst_offset = radeon_get_ib_value(p, idx+1); dst_offset <<= 8; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); p->idx += count + 7; break; /* linear */ @@ -2722,8 +2725,8 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst_offset = radeon_get_ib_value(p, idx+1); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; p->idx += count + 3; break; default: @@ -2765,10 +2768,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; - ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; + ib[idx+4] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 5; break; /* Copy L2T/T2L */ @@ -2778,22 +2781,22 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) /* tiled src, linear dst */ src_offset = radeon_get_ib_value(p, idx+1); src_offset <<= 8; - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(src_reloc->gpu_offset >> 8); dst_offset = radeon_get_ib_value(p, idx + 7); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32; - ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; } else { /* linear src, tiled dst */ src_offset = radeon_get_ib_value(p, idx+7); src_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32; - ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(src_reloc->gpu_offset) & 0xff; dst_offset = radeon_get_ib_value(p, idx+1); dst_offset <<= 8; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); } if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) { dev_warn(p->dev, "DMA L2T, src buffer too small (%llu %lu)\n", @@ -2824,10 +2827,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst_offset + count, radeon_bo_size(dst_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+3] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; - ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xffffffff); + ib[idx+2] += (u32)(src_reloc->gpu_offset & 0xffffffff); + ib[idx+3] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; + ib[idx+4] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 5; break; /* Copy L2L, partial */ @@ -2837,10 +2840,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) DRM_ERROR("L2L Partial is cayman only !\n"); return -EINVAL; } - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; - ib[idx+4] += (u32)(dst_reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+5] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(src_reloc->gpu_offset & 0xffffffff); + ib[idx+2] += upper_32_bits(src_reloc->gpu_offset) & 0xff; + ib[idx+4] += (u32)(dst_reloc->gpu_offset & 0xffffffff); + ib[idx+5] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; p->idx += 9; break; @@ -2873,12 +2876,12 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+4] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; - ib[idx+5] += upper_32_bits(dst2_reloc->lobj.gpu_offset) & 0xff; - ib[idx+6] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += (u32)(dst2_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+4] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; + ib[idx+5] += upper_32_bits(dst2_reloc->gpu_offset) & 0xff; + ib[idx+6] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 7; break; /* Copy L2T Frame to Field */ @@ -2913,10 +2916,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); - ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8); - ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); + ib[idx+2] += (u32)(dst2_reloc->gpu_offset >> 8); + ib[idx+8] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+9] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 10; break; /* Copy L2T/T2L, partial */ @@ -2929,16 +2932,16 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) /* detile bit */ if (radeon_get_ib_value(p, idx + 2) & (1 << 31)) { /* tiled src, linear dst */ - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(src_reloc->gpu_offset >> 8); - ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; } else { /* linear src, tiled dst */ - ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(src_reloc->gpu_offset) & 0xff; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); } p->idx += 12; break; @@ -2975,10 +2978,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); - ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8); - ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); + ib[idx+2] += (u32)(dst2_reloc->gpu_offset >> 8); + ib[idx+8] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+9] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 10; break; /* Copy L2T/T2L (tile units) */ @@ -2989,22 +2992,22 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) /* tiled src, linear dst */ src_offset = radeon_get_ib_value(p, idx+1); src_offset <<= 8; - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(src_reloc->gpu_offset >> 8); dst_offset = radeon_get_ib_value(p, idx+7); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32; - ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; } else { /* linear src, tiled dst */ src_offset = radeon_get_ib_value(p, idx+7); src_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32; - ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(src_reloc->gpu_offset) & 0xff; dst_offset = radeon_get_ib_value(p, idx+1); dst_offset <<= 8; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); } if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) { dev_warn(p->dev, "DMA L2T, T2L src buffer too small (%llu %lu)\n", @@ -3025,8 +3028,8 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) DRM_ERROR("L2T, T2L Partial is cayman only !\n"); return -EINVAL; } - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8); - ib[idx+4] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(src_reloc->gpu_offset >> 8); + ib[idx+4] += (u32)(dst_reloc->gpu_offset >> 8); p->idx += 13; break; /* Copy L2T broadcast (tile units) */ @@ -3062,10 +3065,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); - ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8); - ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); + ib[idx+2] += (u32)(dst2_reloc->gpu_offset >> 8); + ib[idx+8] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+9] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 10; break; default: @@ -3086,8 +3089,8 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst_offset, radeon_bo_size(dst_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += (upper_32_bits(dst_reloc->lobj.gpu_offset) << 16) & 0x00ff0000; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += (upper_32_bits(dst_reloc->gpu_offset) << 16) & 0x00ff0000; p->idx += 4; break; case DMA_PACKET_NOP: diff --git a/drivers/gpu/drm/radeon/evergreen_dma.c b/drivers/gpu/drm/radeon/evergreen_dma.c new file mode 100644 index 00000000000..478caefe0fe --- /dev/null +++ b/drivers/gpu/drm/radeon/evergreen_dma.c @@ -0,0 +1,184 @@ +/* + * Copyright 2010 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "evergreend.h" + +u32 evergreen_gpu_check_soft_reset(struct radeon_device *rdev); + +/** + * evergreen_dma_fence_ring_emit - emit a fence on the DMA ring + * + * @rdev: radeon_device pointer + * @fence: radeon fence object + * + * Add a DMA fence packet to the ring to write + * the fence seq number and DMA trap packet to generate + * an interrupt if needed (evergreen-SI). + */ +void evergreen_dma_fence_ring_emit(struct radeon_device *rdev, + struct radeon_fence *fence) +{ + struct radeon_ring *ring = &rdev->ring[fence->ring]; + u64 addr = rdev->fence_drv[fence->ring].gpu_addr; + /* write the fence */ + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_FENCE, 0, 0)); + radeon_ring_write(ring, addr & 0xfffffffc); + radeon_ring_write(ring, (upper_32_bits(addr) & 0xff)); + radeon_ring_write(ring, fence->seq); + /* generate an interrupt */ + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_TRAP, 0, 0)); + /* flush HDP */ + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0)); + radeon_ring_write(ring, (0xf << 16) | (HDP_MEM_COHERENCY_FLUSH_CNTL >> 2)); + radeon_ring_write(ring, 1); +} + +/** + * evergreen_dma_ring_ib_execute - schedule an IB on the DMA engine + * + * @rdev: radeon_device pointer + * @ib: IB object to schedule + * + * Schedule an IB in the DMA ring (evergreen). + */ +void evergreen_dma_ring_ib_execute(struct radeon_device *rdev, + struct radeon_ib *ib) +{ + struct radeon_ring *ring = &rdev->ring[ib->ring]; + + if (rdev->wb.enabled) { + u32 next_rptr = ring->wptr + 4; + while ((next_rptr & 7) != 5) + next_rptr++; + next_rptr += 3; + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 1)); + radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); + radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xff); + radeon_ring_write(ring, next_rptr); + } + + /* The indirect buffer packet must end on an 8 DW boundary in the DMA ring. + * Pad as necessary with NOPs. + */ + while ((ring->wptr & 7) != 5) + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_NOP, 0, 0)); + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_INDIRECT_BUFFER, 0, 0)); + radeon_ring_write(ring, (ib->gpu_addr & 0xFFFFFFE0)); + radeon_ring_write(ring, (ib->length_dw << 12) | (upper_32_bits(ib->gpu_addr) & 0xFF)); + +} + +/** + * evergreen_copy_dma - copy pages using the DMA engine + * + * @rdev: radeon_device pointer + * @src_offset: src GPU address + * @dst_offset: dst GPU address + * @num_gpu_pages: number of GPU pages to xfer + * @fence: radeon fence object + * + * Copy GPU paging using the DMA engine (evergreen-cayman). + * Used by the radeon ttm implementation to move pages if + * registered as the asic copy callback. + */ +int evergreen_copy_dma(struct radeon_device *rdev, + uint64_t src_offset, uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence **fence) +{ + struct radeon_semaphore *sem = NULL; + int ring_index = rdev->asic->copy.dma_ring_index; + struct radeon_ring *ring = &rdev->ring[ring_index]; + u32 size_in_dw, cur_size_in_dw; + int i, num_loops; + int r = 0; + + r = radeon_semaphore_create(rdev, &sem); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + return r; + } + + size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4; + num_loops = DIV_ROUND_UP(size_in_dw, 0xfffff); + r = radeon_ring_lock(rdev, ring, num_loops * 5 + 11); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_semaphore_sync_to(sem, *fence); + radeon_semaphore_sync_rings(rdev, sem, ring->idx); + + for (i = 0; i < num_loops; i++) { + cur_size_in_dw = size_in_dw; + if (cur_size_in_dw > 0xFFFFF) + cur_size_in_dw = 0xFFFFF; + size_in_dw -= cur_size_in_dw; + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 0, cur_size_in_dw)); + radeon_ring_write(ring, dst_offset & 0xfffffffc); + radeon_ring_write(ring, src_offset & 0xfffffffc); + radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff); + radeon_ring_write(ring, upper_32_bits(src_offset) & 0xff); + src_offset += cur_size_in_dw * 4; + dst_offset += cur_size_in_dw * 4; + } + + r = radeon_fence_emit(rdev, fence, ring->idx); + if (r) { + radeon_ring_unlock_undo(rdev, ring); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_ring_unlock_commit(rdev, ring); + radeon_semaphore_free(rdev, &sem, *fence); + + return r; +} + +/** + * evergreen_dma_is_lockup - Check if the DMA engine is locked up + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * + * Check if the async DMA engine is locked up. + * Returns true if the engine appears to be locked up, false if not. + */ +bool evergreen_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) +{ + u32 reset_mask = evergreen_gpu_check_soft_reset(rdev); + + if (!(reset_mask & RADEON_RESET_DMA)) { + radeon_ring_lockup_update(rdev, ring); + return false; + } + return radeon_ring_test_lockup(rdev, ring); +} + + diff --git a/drivers/gpu/drm/radeon/evergreen_hdmi.c b/drivers/gpu/drm/radeon/evergreen_hdmi.c index b9c6f7675e5..1ec0e6e83f9 100644 --- a/drivers/gpu/drm/radeon/evergreen_hdmi.c +++ b/drivers/gpu/drm/radeon/evergreen_hdmi.c @@ -32,6 +32,12 @@ #include "evergreend.h" #include "atom.h" +extern void dce6_afmt_write_speaker_allocation(struct drm_encoder *encoder); +extern void dce6_afmt_write_sad_regs(struct drm_encoder *encoder); +extern void dce6_afmt_select_pin(struct drm_encoder *encoder); +extern void dce6_afmt_write_latency_fields(struct drm_encoder *encoder, + struct drm_display_mode *mode); + /* * update the N and CTS parameters for a given pixel clock rate */ @@ -54,6 +60,83 @@ static void evergreen_hdmi_update_ACR(struct drm_encoder *encoder, uint32_t cloc WREG32(HDMI_ACR_48_1 + offset, acr.n_48khz); } +static void dce4_afmt_write_latency_fields(struct drm_encoder *encoder, + struct drm_display_mode *mode) +{ + struct radeon_device *rdev = encoder->dev->dev_private; + struct drm_connector *connector; + struct radeon_connector *radeon_connector = NULL; + u32 tmp = 0; + + list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) { + if (connector->encoder == encoder) { + radeon_connector = to_radeon_connector(connector); + break; + } + } + + if (!radeon_connector) { + DRM_ERROR("Couldn't find encoder's connector\n"); + return; + } + + if (mode->flags & DRM_MODE_FLAG_INTERLACE) { + if (connector->latency_present[1]) + tmp = VIDEO_LIPSYNC(connector->video_latency[1]) | + AUDIO_LIPSYNC(connector->audio_latency[1]); + else + tmp = VIDEO_LIPSYNC(255) | AUDIO_LIPSYNC(255); + } else { + if (connector->latency_present[0]) + tmp = VIDEO_LIPSYNC(connector->video_latency[0]) | + AUDIO_LIPSYNC(connector->audio_latency[0]); + else + tmp = VIDEO_LIPSYNC(255) | AUDIO_LIPSYNC(255); + } + WREG32(AZ_F0_CODEC_PIN0_CONTROL_RESPONSE_LIPSYNC, tmp); +} + +static void dce4_afmt_write_speaker_allocation(struct drm_encoder *encoder) +{ + struct radeon_device *rdev = encoder->dev->dev_private; + struct drm_connector *connector; + struct radeon_connector *radeon_connector = NULL; + u32 tmp; + u8 *sadb; + int sad_count; + + list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) { + if (connector->encoder == encoder) { + radeon_connector = to_radeon_connector(connector); + break; + } + } + + if (!radeon_connector) { + DRM_ERROR("Couldn't find encoder's connector\n"); + return; + } + + sad_count = drm_edid_to_speaker_allocation(radeon_connector->edid, &sadb); + if (sad_count <= 0) { + DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count); + return; + } + + /* program the speaker allocation */ + tmp = RREG32(AZ_F0_CODEC_PIN0_CONTROL_CHANNEL_SPEAKER); + tmp &= ~(DP_CONNECTION | SPEAKER_ALLOCATION_MASK); + /* set HDMI mode */ + tmp |= HDMI_CONNECTION; + if (sad_count) + tmp |= SPEAKER_ALLOCATION(sadb[0]); + else + tmp |= SPEAKER_ALLOCATION(5); /* stereo */ + WREG32(AZ_F0_CODEC_PIN0_CONTROL_CHANNEL_SPEAKER, tmp); + + kfree(sadb); +} + static void evergreen_hdmi_write_sad_regs(struct drm_encoder *encoder) { struct radeon_device *rdev = encoder->dev->dev_private; @@ -78,8 +161,10 @@ static void evergreen_hdmi_write_sad_regs(struct drm_encoder *encoder) }; list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) { - if (connector->encoder == encoder) + if (connector->encoder == encoder) { radeon_connector = to_radeon_connector(connector); + break; + } } if (!radeon_connector) { @@ -88,7 +173,7 @@ static void evergreen_hdmi_write_sad_regs(struct drm_encoder *encoder) } sad_count = drm_edid_to_sad(radeon_connector->edid, &sads); - if (sad_count < 0) { + if (sad_count <= 0) { DRM_ERROR("Couldn't read SADs: %d\n", sad_count); return; } @@ -96,20 +181,30 @@ static void evergreen_hdmi_write_sad_regs(struct drm_encoder *encoder) for (i = 0; i < ARRAY_SIZE(eld_reg_to_type); i++) { u32 value = 0; + u8 stereo_freqs = 0; + int max_channels = -1; int j; for (j = 0; j < sad_count; j++) { struct cea_sad *sad = &sads[j]; if (sad->format == eld_reg_to_type[i][1]) { - value = MAX_CHANNELS(sad->channels) | - DESCRIPTOR_BYTE_2(sad->byte2) | - SUPPORTED_FREQUENCIES(sad->freq); + if (sad->channels > max_channels) { + value = MAX_CHANNELS(sad->channels) | + DESCRIPTOR_BYTE_2(sad->byte2) | + SUPPORTED_FREQUENCIES(sad->freq); + max_channels = sad->channels; + } + if (sad->format == HDMI_AUDIO_CODING_TYPE_PCM) - value |= SUPPORTED_FREQUENCIES_STEREO(sad->freq); - break; + stereo_freqs |= sad->freq; + else + break; } } + + value |= SUPPORTED_FREQUENCIES_STEREO(stereo_freqs); + WREG32(eld_reg_to_type[i][0], value); } @@ -148,18 +243,44 @@ static void evergreen_audio_set_dto(struct drm_encoder *encoder, u32 clock) struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); u32 base_rate = 24000; + u32 max_ratio = clock / base_rate; + u32 dto_phase; + u32 dto_modulo = clock; + u32 wallclock_ratio; + u32 dto_cntl; if (!dig || !dig->afmt) return; + if (ASIC_IS_DCE6(rdev)) { + dto_phase = 24 * 1000; + } else { + if (max_ratio >= 8) { + dto_phase = 192 * 1000; + wallclock_ratio = 3; + } else if (max_ratio >= 4) { + dto_phase = 96 * 1000; + wallclock_ratio = 2; + } else if (max_ratio >= 2) { + dto_phase = 48 * 1000; + wallclock_ratio = 1; + } else { + dto_phase = 24 * 1000; + wallclock_ratio = 0; + } + dto_cntl = RREG32(DCCG_AUDIO_DTO0_CNTL) & ~DCCG_AUDIO_DTO_WALLCLOCK_RATIO_MASK; + dto_cntl |= DCCG_AUDIO_DTO_WALLCLOCK_RATIO(wallclock_ratio); + WREG32(DCCG_AUDIO_DTO0_CNTL, dto_cntl); + } + /* XXX two dtos; generally use dto0 for hdmi */ /* Express [24MHz / target pixel clock] as an exact rational * number (coefficient of two integer numbers. DCCG_AUDIO_DTOx_PHASE * is the numerator, DCCG_AUDIO_DTOx_MODULE is the denominator */ - WREG32(DCCG_AUDIO_DTO0_PHASE, base_rate * 100); - WREG32(DCCG_AUDIO_DTO0_MODULE, clock * 100); WREG32(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO0_SOURCE_SEL(radeon_crtc->crtc_id)); + WREG32(DCCG_AUDIO_DTO0_PHASE, dto_phase); + WREG32(DCCG_AUDIO_DTO0_MODULE, dto_modulo); } @@ -172,16 +293,37 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode struct radeon_device *rdev = dev->dev_private; struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + struct drm_connector *connector = radeon_get_connector_for_encoder(encoder); u8 buffer[HDMI_INFOFRAME_HEADER_SIZE + HDMI_AVI_INFOFRAME_SIZE]; struct hdmi_avi_infoframe frame; uint32_t offset; ssize_t err; + uint32_t val; + int bpc = 8; + + if (!dig || !dig->afmt) + return; /* Silent, r600_hdmi_enable will raise WARN for us */ if (!dig->afmt->enabled) return; offset = dig->afmt->offset; + /* hdmi deep color mode general control packets setup, if bpc > 8 */ + if (encoder->crtc) { + struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); + bpc = radeon_crtc->bpc; + } + + /* disable audio prior to setting up hw */ + if (ASIC_IS_DCE6(rdev)) { + dig->afmt->pin = dce6_audio_get_pin(rdev); + dce6_audio_enable(rdev, dig->afmt->pin, false); + } else { + dig->afmt->pin = r600_audio_get_pin(rdev); + r600_audio_enable(rdev, dig->afmt->pin, false); + } + evergreen_audio_set_dto(encoder, mode->clock); WREG32(HDMI_VBI_PACKET_CONTROL + offset, @@ -189,6 +331,35 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode WREG32(AFMT_AUDIO_CRC_CONTROL + offset, 0x1000); + val = RREG32(HDMI_CONTROL + offset); + val &= ~HDMI_DEEP_COLOR_ENABLE; + val &= ~HDMI_DEEP_COLOR_DEPTH_MASK; + + switch (bpc) { + case 0: + case 6: + case 8: + case 16: + default: + DRM_DEBUG("%s: Disabling hdmi deep color for %d bpc.\n", + connector->name, bpc); + break; + case 10: + val |= HDMI_DEEP_COLOR_ENABLE; + val |= HDMI_DEEP_COLOR_DEPTH(HDMI_30BIT_DEEP_COLOR); + DRM_DEBUG("%s: Enabling hdmi deep color 30 for 10 bpc.\n", + connector->name); + break; + case 12: + val |= HDMI_DEEP_COLOR_ENABLE; + val |= HDMI_DEEP_COLOR_DEPTH(HDMI_36BIT_DEEP_COLOR); + DRM_DEBUG("%s: Enabling hdmi deep color 36 for 12 bpc.\n", + connector->name); + break; + } + + WREG32(HDMI_CONTROL + offset, val); + WREG32(HDMI_VBI_PACKET_CONTROL + offset, HDMI_NULL_SEND | /* send null packets when required */ HDMI_GC_SEND | /* send general control packets */ @@ -215,9 +386,13 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode /* fglrx clears sth in AFMT_AUDIO_PACKET_CONTROL2 here */ - WREG32(HDMI_ACR_PACKET_CONTROL + offset, - HDMI_ACR_AUTO_SEND | /* allow hw to sent ACR packets when required */ - HDMI_ACR_SOURCE); /* select SW CTS value */ + if (bpc > 8) + WREG32(HDMI_ACR_PACKET_CONTROL + offset, + HDMI_ACR_AUTO_SEND); /* allow hw to sent ACR packets when required */ + else + WREG32(HDMI_ACR_PACKET_CONTROL + offset, + HDMI_ACR_SOURCE | /* select SW CTS value */ + HDMI_ACR_AUTO_SEND); /* allow hw to sent ACR packets when required */ evergreen_hdmi_update_ACR(encoder, mode->clock); @@ -235,13 +410,25 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode AFMT_60958_CS_CHANNEL_NUMBER_6(7) | AFMT_60958_CS_CHANNEL_NUMBER_7(8)); - /* fglrx sets 0x0001005f | (x & 0x00fc0000) in 0x5f78 here */ + if (ASIC_IS_DCE6(rdev)) { + dce6_afmt_write_speaker_allocation(encoder); + } else { + dce4_afmt_write_speaker_allocation(encoder); + } WREG32(AFMT_AUDIO_PACKET_CONTROL2 + offset, AFMT_AUDIO_CHANNEL_ENABLE(0xff)); /* fglrx sets 0x40 in 0x5f80 here */ - evergreen_hdmi_write_sad_regs(encoder); + + if (ASIC_IS_DCE6(rdev)) { + dce6_afmt_select_pin(encoder); + dce6_afmt_write_sad_regs(encoder); + dce6_afmt_write_latency_fields(encoder, mode); + } else { + evergreen_hdmi_write_sad_regs(encoder); + dce4_afmt_write_latency_fields(encoder, mode); + } err = drm_hdmi_avi_infoframe_from_display_mode(&frame, mode); if (err < 0) { @@ -273,6 +460,12 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode WREG32(AFMT_RAMP_CONTROL1 + offset, 0x007FFFFF); WREG32(AFMT_RAMP_CONTROL2 + offset, 0x00000001); WREG32(AFMT_RAMP_CONTROL3 + offset, 0x00000001); + + /* enable audio after to setting up hw */ + if (ASIC_IS_DCE6(rdev)) + dce6_audio_enable(rdev, dig->afmt->pin, true); + else + r600_audio_enable(rdev, dig->afmt->pin, true); } void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable) @@ -280,6 +473,9 @@ void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable) struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + if (!dig || !dig->afmt) + return; + /* Silent, r600_hdmi_enable will raise WARN for us */ if (enable && dig->afmt->enabled) return; diff --git a/drivers/gpu/drm/radeon/evergreen_reg.h b/drivers/gpu/drm/radeon/evergreen_reg.h index 8a4e641f0e3..23bff590fb6 100644 --- a/drivers/gpu/drm/radeon/evergreen_reg.h +++ b/drivers/gpu/drm/radeon/evergreen_reg.h @@ -33,6 +33,7 @@ #define EVERGREEN_PIF_PHY0_DATA 0xc #define EVERGREEN_PIF_PHY1_INDEX 0x10 #define EVERGREEN_PIF_PHY1_DATA 0x14 +#define EVERGREEN_MM_INDEX_HI 0x18 #define EVERGREEN_VGA_MEMORY_BASE_ADDRESS 0x310 #define EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH 0x324 @@ -115,6 +116,8 @@ # define EVERGREEN_GRPH_ARRAY_LINEAR_ALIGNED 1 # define EVERGREEN_GRPH_ARRAY_1D_TILED_THIN1 2 # define EVERGREEN_GRPH_ARRAY_2D_TILED_THIN1 4 +#define EVERGREEN_GRPH_LUT_10BIT_BYPASS_CONTROL 0x6808 +# define EVERGREEN_LUT_10BIT_BYPASS_EN (1 << 8) #define EVERGREEN_GRPH_SWAP_CONTROL 0x680c # define EVERGREEN_GRPH_ENDIAN_SWAP(x) (((x) & 0x3) << 0) # define EVERGREEN_GRPH_ENDIAN_NONE 0 @@ -236,7 +239,6 @@ # define EVERGREEN_CRTC_V_BLANK (1 << 0) #define EVERGREEN_CRTC_STATUS_POSITION 0x6e90 #define EVERGREEN_CRTC_STATUS_HV_COUNT 0x6ea0 -#define EVERGREEN_MASTER_UPDATE_MODE 0x6ef8 #define EVERGREEN_CRTC_UPDATE_LOCK 0x6ed4 #define EVERGREEN_MASTER_UPDATE_LOCK 0x6ef4 #define EVERGREEN_MASTER_UPDATE_MODE 0x6ef8 diff --git a/drivers/gpu/drm/radeon/evergreen_smc.h b/drivers/gpu/drm/radeon/evergreen_smc.h index 76ada8cfe90..3a03ba37d04 100644 --- a/drivers/gpu/drm/radeon/evergreen_smc.h +++ b/drivers/gpu/drm/radeon/evergreen_smc.h @@ -57,7 +57,7 @@ typedef struct SMC_Evergreen_MCRegisters SMC_Evergreen_MCRegisters; #define EVERGREEN_SMC_FIRMWARE_HEADER_LOCATION 0x100 -#define EVERGREEN_SMC_FIRMWARE_HEADER_softRegisters 0x0 +#define EVERGREEN_SMC_FIRMWARE_HEADER_softRegisters 0x8 #define EVERGREEN_SMC_FIRMWARE_HEADER_stateTable 0xC #define EVERGREEN_SMC_FIRMWARE_HEADER_mcRegisterTable 0x20 diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h index a7baf67aef6..b066d6711b8 100644 --- a/drivers/gpu/drm/radeon/evergreend.h +++ b/drivers/gpu/drm/radeon/evergreend.h @@ -82,12 +82,16 @@ #define CG_SPLL_FUNC_CNTL_2 0x604 #define SCLK_MUX_SEL(x) ((x) << 0) #define SCLK_MUX_SEL_MASK (0x1ff << 0) +#define SCLK_MUX_UPDATE (1 << 26) #define CG_SPLL_FUNC_CNTL_3 0x608 #define SPLL_FB_DIV(x) ((x) << 0) #define SPLL_FB_DIV_MASK (0x3ffffff << 0) #define SPLL_DITHEN (1 << 28) +#define CG_SPLL_STATUS 0x60c +#define SPLL_CHG_STATUS (1 << 1) #define MPLL_CNTL_MODE 0x61c +# define MPLL_MCLK_SEL (1 << 11) # define SS_SSEN (1 << 24) # define SS_DSMODE_EN (1 << 25) @@ -497,6 +501,9 @@ #define DCCG_AUDIO_DTO0_MODULE 0x05b4 #define DCCG_AUDIO_DTO0_LOAD 0x05b8 #define DCCG_AUDIO_DTO0_CNTL 0x05bc +# define DCCG_AUDIO_DTO_WALLCLOCK_RATIO(x) (((x) & 7) << 0) +# define DCCG_AUDIO_DTO_WALLCLOCK_RATIO_MASK 7 +# define DCCG_AUDIO_DTO_WALLCLOCK_RATIO_SHIFT 0 #define DCCG_AUDIO_DTO1_PHASE 0x05c0 #define DCCG_AUDIO_DTO1_MODULE 0x05c4 @@ -510,10 +517,11 @@ # define HDMI_ERROR_ACK (1 << 8) # define HDMI_ERROR_MASK (1 << 9) # define HDMI_DEEP_COLOR_ENABLE (1 << 24) -# define HDMI_DEEP_COLOR_DEPTH (((x) & 3) << 28) +# define HDMI_DEEP_COLOR_DEPTH(x) (((x) & 3) << 28) # define HDMI_24BIT_DEEP_COLOR 0 # define HDMI_30BIT_DEEP_COLOR 1 # define HDMI_36BIT_DEEP_COLOR 2 +# define HDMI_DEEP_COLOR_DEPTH_MASK (3 << 28) #define HDMI_STATUS 0x7034 # define HDMI_ACTIVE_AVMUTE (1 << 0) # define HDMI_AUDIO_PACKET_ERROR (1 << 16) @@ -711,6 +719,13 @@ #define AFMT_GENERIC0_7 0x7138 /* DCE4/5 ELD audio interface */ +#define AZ_F0_CODEC_PIN0_CONTROL_CHANNEL_SPEAKER 0x5f78 +#define SPEAKER_ALLOCATION(x) (((x) & 0x7f) << 0) +#define SPEAKER_ALLOCATION_MASK (0x7f << 0) +#define SPEAKER_ALLOCATION_SHIFT 0 +#define HDMI_CONNECTION (1 << 16) +#define DP_CONNECTION (1 << 17) + #define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR0 0x5f84 /* LPCM */ #define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR1 0x5f88 /* AC3 */ #define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR2 0x5f8c /* MPEG1 */ @@ -740,6 +755,44 @@ * bit6 = 192 kHz */ +#define AZ_CHANNEL_COUNT_CONTROL 0x5fe4 +# define HBR_CHANNEL_COUNT(x) (((x) & 0x7) << 0) +# define COMPRESSED_CHANNEL_COUNT(x) (((x) & 0x7) << 4) +/* HBR_CHANNEL_COUNT, COMPRESSED_CHANNEL_COUNT + * 0 = use stream header + * 1-7 = channel count - 1 + */ +#define AZ_F0_CODEC_PIN0_CONTROL_RESPONSE_LIPSYNC 0x5fe8 +# define VIDEO_LIPSYNC(x) (((x) & 0xff) << 0) +# define AUDIO_LIPSYNC(x) (((x) & 0xff) << 8) +/* VIDEO_LIPSYNC, AUDIO_LIPSYNC + * 0 = invalid + * x = legal delay value + * 255 = sync not supported + */ +#define AZ_F0_CODEC_PIN0_CONTROL_RESPONSE_HBR 0x5fec +# define HBR_CAPABLE (1 << 0) /* enabled by default */ + +#define AZ_F0_CODEC_PIN0_CONTROL_RESPONSE_AV_ASSOCIATION0 0x5ff4 +# define DISPLAY0_TYPE(x) (((x) & 0x3) << 0) +# define DISPLAY_TYPE_NONE 0 +# define DISPLAY_TYPE_HDMI 1 +# define DISPLAY_TYPE_DP 2 +# define DISPLAY0_ID(x) (((x) & 0x3f) << 2) +# define DISPLAY1_TYPE(x) (((x) & 0x3) << 8) +# define DISPLAY1_ID(x) (((x) & 0x3f) << 10) +# define DISPLAY2_TYPE(x) (((x) & 0x3) << 16) +# define DISPLAY2_ID(x) (((x) & 0x3f) << 18) +# define DISPLAY3_TYPE(x) (((x) & 0x3) << 24) +# define DISPLAY3_ID(x) (((x) & 0x3f) << 26) +#define AZ_F0_CODEC_PIN0_CONTROL_RESPONSE_AV_ASSOCIATION1 0x5ff8 +# define DISPLAY4_TYPE(x) (((x) & 0x3) << 0) +# define DISPLAY4_ID(x) (((x) & 0x3f) << 2) +# define DISPLAY5_TYPE(x) (((x) & 0x3) << 8) +# define DISPLAY5_ID(x) (((x) & 0x3f) << 10) +#define AZ_F0_CODEC_PIN0_CONTROL_RESPONSE_AV_NUMBER 0x5ffc +# define NUMBER_OF_DISPLAY_ID(x) (((x) & 0x7) << 0) + #define AZ_HOT_PLUG_CONTROL 0x5e78 # define AZ_FORCE_CODEC_WAKE (1 << 0) # define PIN0_JACK_DETECTION_ENABLE (1 << 4) @@ -1150,6 +1203,10 @@ # define LATENCY_LOW_WATERMARK(x) ((x) << 0) # define LATENCY_HIGH_WATERMARK(x) ((x) << 16) +#define PIPE0_DMIF_BUFFER_CONTROL 0x0ca0 +# define DMIF_BUFFERS_ALLOCATED(x) ((x) << 0) +# define DMIF_BUFFERS_ALLOCATED_COMPLETED (1 << 4) + #define IH_RB_CNTL 0x3e00 # define IH_RB_ENABLE (1 << 0) # define IH_IB_SIZE(x) ((x) << 1) /* log2 */ @@ -1298,6 +1355,38 @@ # define DC_HPDx_RX_INT_TIMER(x) ((x) << 16) # define DC_HPDx_EN (1 << 28) +/* DCE4/5/6 FMT blocks */ +#define FMT_DYNAMIC_EXP_CNTL 0x6fb4 +# define FMT_DYNAMIC_EXP_EN (1 << 0) +# define FMT_DYNAMIC_EXP_MODE (1 << 4) + /* 0 = 10bit -> 12bit, 1 = 8bit -> 12bit */ +#define FMT_CONTROL 0x6fb8 +# define FMT_PIXEL_ENCODING (1 << 16) + /* 0 = RGB 4:4:4 or YCbCr 4:4:4, 1 = YCbCr 4:2:2 */ +#define FMT_BIT_DEPTH_CONTROL 0x6fc8 +# define FMT_TRUNCATE_EN (1 << 0) +# define FMT_TRUNCATE_DEPTH (1 << 4) +# define FMT_SPATIAL_DITHER_EN (1 << 8) +# define FMT_SPATIAL_DITHER_MODE(x) ((x) << 9) +# define FMT_SPATIAL_DITHER_DEPTH (1 << 12) +# define FMT_FRAME_RANDOM_ENABLE (1 << 13) +# define FMT_RGB_RANDOM_ENABLE (1 << 14) +# define FMT_HIGHPASS_RANDOM_ENABLE (1 << 15) +# define FMT_TEMPORAL_DITHER_EN (1 << 16) +# define FMT_TEMPORAL_DITHER_DEPTH (1 << 20) +# define FMT_TEMPORAL_DITHER_OFFSET(x) ((x) << 21) +# define FMT_TEMPORAL_LEVEL (1 << 24) +# define FMT_TEMPORAL_DITHER_RESET (1 << 25) +# define FMT_25FRC_SEL(x) ((x) << 26) +# define FMT_50FRC_SEL(x) ((x) << 28) +# define FMT_75FRC_SEL(x) ((x) << 30) +#define FMT_CLAMP_CONTROL 0x6fe4 +# define FMT_CLAMP_DATA_EN (1 << 0) +# define FMT_CLAMP_COLOR_FORMAT(x) ((x) << 16) +# define FMT_CLAMP_6BPC 0 +# define FMT_CLAMP_8BPC 1 +# define FMT_CLAMP_10BPC 2 + /* ASYNC DMA */ #define DMA_RB_RPTR 0xd008 #define DMA_RB_WPTR 0xd00c @@ -1487,7 +1576,7 @@ * 6. COMMAND [29:22] | BYTE_COUNT [20:0] */ # define PACKET3_CP_DMA_DST_SEL(x) ((x) << 20) - /* 0 - SRC_ADDR + /* 0 - DST_ADDR * 1 - GDS */ # define PACKET3_CP_DMA_ENGINE(x) ((x) << 27) @@ -1502,7 +1591,7 @@ # define PACKET3_CP_DMA_CP_SYNC (1 << 31) /* COMMAND */ # define PACKET3_CP_DMA_DIS_WC (1 << 21) -# define PACKET3_CP_DMA_CMD_SRC_SWAP(x) ((x) << 23) +# define PACKET3_CP_DMA_CMD_SRC_SWAP(x) ((x) << 22) /* 0 - none * 1 - 8 in 16 * 2 - 8 in 32 diff --git a/drivers/gpu/drm/radeon/kv_dpm.c b/drivers/gpu/drm/radeon/kv_dpm.c new file mode 100644 index 00000000000..9ef8c38f2d6 --- /dev/null +++ b/drivers/gpu/drm/radeon/kv_dpm.c @@ -0,0 +1,2830 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include "drmP.h" +#include "radeon.h" +#include "cikd.h" +#include "r600_dpm.h" +#include "kv_dpm.h" +#include "radeon_asic.h" +#include <linux/seq_file.h> + +#define KV_MAX_DEEPSLEEP_DIVIDER_ID 5 +#define KV_MINIMUM_ENGINE_CLOCK 800 +#define SMC_RAM_END 0x40000 + +static void kv_init_graphics_levels(struct radeon_device *rdev); +static int kv_calculate_ds_divider(struct radeon_device *rdev); +static int kv_calculate_nbps_level_settings(struct radeon_device *rdev); +static int kv_calculate_dpm_settings(struct radeon_device *rdev); +static void kv_enable_new_levels(struct radeon_device *rdev); +static void kv_program_nbps_index_settings(struct radeon_device *rdev, + struct radeon_ps *new_rps); +static int kv_set_enabled_level(struct radeon_device *rdev, u32 level); +static int kv_set_enabled_levels(struct radeon_device *rdev); +static int kv_force_dpm_highest(struct radeon_device *rdev); +static int kv_force_dpm_lowest(struct radeon_device *rdev); +static void kv_apply_state_adjust_rules(struct radeon_device *rdev, + struct radeon_ps *new_rps, + struct radeon_ps *old_rps); +static int kv_set_thermal_temperature_range(struct radeon_device *rdev, + int min_temp, int max_temp); +static int kv_init_fps_limits(struct radeon_device *rdev); + +void kv_dpm_powergate_uvd(struct radeon_device *rdev, bool gate); +static void kv_dpm_powergate_vce(struct radeon_device *rdev, bool gate); +static void kv_dpm_powergate_samu(struct radeon_device *rdev, bool gate); +static void kv_dpm_powergate_acp(struct radeon_device *rdev, bool gate); + +extern void cik_enter_rlc_safe_mode(struct radeon_device *rdev); +extern void cik_exit_rlc_safe_mode(struct radeon_device *rdev); +extern void cik_update_cg(struct radeon_device *rdev, + u32 block, bool enable); + +static const struct kv_lcac_config_values sx_local_cac_cfg_kv[] = +{ + { 0, 4, 1 }, + { 1, 4, 1 }, + { 2, 5, 1 }, + { 3, 4, 2 }, + { 4, 1, 1 }, + { 5, 5, 2 }, + { 6, 6, 1 }, + { 7, 9, 2 }, + { 0xffffffff } +}; + +static const struct kv_lcac_config_values mc0_local_cac_cfg_kv[] = +{ + { 0, 4, 1 }, + { 0xffffffff } +}; + +static const struct kv_lcac_config_values mc1_local_cac_cfg_kv[] = +{ + { 0, 4, 1 }, + { 0xffffffff } +}; + +static const struct kv_lcac_config_values mc2_local_cac_cfg_kv[] = +{ + { 0, 4, 1 }, + { 0xffffffff } +}; + +static const struct kv_lcac_config_values mc3_local_cac_cfg_kv[] = +{ + { 0, 4, 1 }, + { 0xffffffff } +}; + +static const struct kv_lcac_config_values cpl_local_cac_cfg_kv[] = +{ + { 0, 4, 1 }, + { 1, 4, 1 }, + { 2, 5, 1 }, + { 3, 4, 1 }, + { 4, 1, 1 }, + { 5, 5, 1 }, + { 6, 6, 1 }, + { 7, 9, 1 }, + { 8, 4, 1 }, + { 9, 2, 1 }, + { 10, 3, 1 }, + { 11, 6, 1 }, + { 12, 8, 2 }, + { 13, 1, 1 }, + { 14, 2, 1 }, + { 15, 3, 1 }, + { 16, 1, 1 }, + { 17, 4, 1 }, + { 18, 3, 1 }, + { 19, 1, 1 }, + { 20, 8, 1 }, + { 21, 5, 1 }, + { 22, 1, 1 }, + { 23, 1, 1 }, + { 24, 4, 1 }, + { 27, 6, 1 }, + { 28, 1, 1 }, + { 0xffffffff } +}; + +static const struct kv_lcac_config_reg sx0_cac_config_reg[] = +{ + { 0xc0400d00, 0x003e0000, 17, 0x3fc00000, 22, 0x0001fffe, 1, 0x00000001, 0 } +}; + +static const struct kv_lcac_config_reg mc0_cac_config_reg[] = +{ + { 0xc0400d30, 0x003e0000, 17, 0x3fc00000, 22, 0x0001fffe, 1, 0x00000001, 0 } +}; + +static const struct kv_lcac_config_reg mc1_cac_config_reg[] = +{ + { 0xc0400d3c, 0x003e0000, 17, 0x3fc00000, 22, 0x0001fffe, 1, 0x00000001, 0 } +}; + +static const struct kv_lcac_config_reg mc2_cac_config_reg[] = +{ + { 0xc0400d48, 0x003e0000, 17, 0x3fc00000, 22, 0x0001fffe, 1, 0x00000001, 0 } +}; + +static const struct kv_lcac_config_reg mc3_cac_config_reg[] = +{ + { 0xc0400d54, 0x003e0000, 17, 0x3fc00000, 22, 0x0001fffe, 1, 0x00000001, 0 } +}; + +static const struct kv_lcac_config_reg cpl_cac_config_reg[] = +{ + { 0xc0400d80, 0x003e0000, 17, 0x3fc00000, 22, 0x0001fffe, 1, 0x00000001, 0 } +}; + +static const struct kv_pt_config_reg didt_config_kv[] = +{ + { 0x10, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x10, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x10, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x10, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x11, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x11, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x11, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x11, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x12, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x12, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x12, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x12, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x2, 0x00003fff, 0, 0x4, KV_CONFIGREG_DIDT_IND }, + { 0x2, 0x03ff0000, 16, 0x80, KV_CONFIGREG_DIDT_IND }, + { 0x2, 0x78000000, 27, 0x3, KV_CONFIGREG_DIDT_IND }, + { 0x1, 0x0000ffff, 0, 0x3FFF, KV_CONFIGREG_DIDT_IND }, + { 0x1, 0xffff0000, 16, 0x3FFF, KV_CONFIGREG_DIDT_IND }, + { 0x0, 0x00000001, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x30, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x30, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x30, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x30, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x31, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x31, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x31, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x31, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x32, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x32, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x32, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x32, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x22, 0x00003fff, 0, 0x4, KV_CONFIGREG_DIDT_IND }, + { 0x22, 0x03ff0000, 16, 0x80, KV_CONFIGREG_DIDT_IND }, + { 0x22, 0x78000000, 27, 0x3, KV_CONFIGREG_DIDT_IND }, + { 0x21, 0x0000ffff, 0, 0x3FFF, KV_CONFIGREG_DIDT_IND }, + { 0x21, 0xffff0000, 16, 0x3FFF, KV_CONFIGREG_DIDT_IND }, + { 0x20, 0x00000001, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x50, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x50, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x50, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x50, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x51, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x51, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x51, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x51, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x52, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x52, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x52, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x52, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x42, 0x00003fff, 0, 0x4, KV_CONFIGREG_DIDT_IND }, + { 0x42, 0x03ff0000, 16, 0x80, KV_CONFIGREG_DIDT_IND }, + { 0x42, 0x78000000, 27, 0x3, KV_CONFIGREG_DIDT_IND }, + { 0x41, 0x0000ffff, 0, 0x3FFF, KV_CONFIGREG_DIDT_IND }, + { 0x41, 0xffff0000, 16, 0x3FFF, KV_CONFIGREG_DIDT_IND }, + { 0x40, 0x00000001, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x70, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x70, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x70, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x70, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x71, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x71, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x71, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x71, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x72, 0x000000ff, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x72, 0x0000ff00, 8, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x72, 0x00ff0000, 16, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x72, 0xff000000, 24, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0x62, 0x00003fff, 0, 0x4, KV_CONFIGREG_DIDT_IND }, + { 0x62, 0x03ff0000, 16, 0x80, KV_CONFIGREG_DIDT_IND }, + { 0x62, 0x78000000, 27, 0x3, KV_CONFIGREG_DIDT_IND }, + { 0x61, 0x0000ffff, 0, 0x3FFF, KV_CONFIGREG_DIDT_IND }, + { 0x61, 0xffff0000, 16, 0x3FFF, KV_CONFIGREG_DIDT_IND }, + { 0x60, 0x00000001, 0, 0x0, KV_CONFIGREG_DIDT_IND }, + { 0xFFFFFFFF } +}; + +static struct kv_ps *kv_get_ps(struct radeon_ps *rps) +{ + struct kv_ps *ps = rps->ps_priv; + + return ps; +} + +static struct kv_power_info *kv_get_pi(struct radeon_device *rdev) +{ + struct kv_power_info *pi = rdev->pm.dpm.priv; + + return pi; +} + +#if 0 +static void kv_program_local_cac_table(struct radeon_device *rdev, + const struct kv_lcac_config_values *local_cac_table, + const struct kv_lcac_config_reg *local_cac_reg) +{ + u32 i, count, data; + const struct kv_lcac_config_values *values = local_cac_table; + + while (values->block_id != 0xffffffff) { + count = values->signal_id; + for (i = 0; i < count; i++) { + data = ((values->block_id << local_cac_reg->block_shift) & + local_cac_reg->block_mask); + data |= ((i << local_cac_reg->signal_shift) & + local_cac_reg->signal_mask); + data |= ((values->t << local_cac_reg->t_shift) & + local_cac_reg->t_mask); + data |= ((1 << local_cac_reg->enable_shift) & + local_cac_reg->enable_mask); + WREG32_SMC(local_cac_reg->cntl, data); + } + values++; + } +} +#endif + +static int kv_program_pt_config_registers(struct radeon_device *rdev, + const struct kv_pt_config_reg *cac_config_regs) +{ + const struct kv_pt_config_reg *config_regs = cac_config_regs; + u32 data; + u32 cache = 0; + + if (config_regs == NULL) + return -EINVAL; + + while (config_regs->offset != 0xFFFFFFFF) { + if (config_regs->type == KV_CONFIGREG_CACHE) { + cache |= ((config_regs->value << config_regs->shift) & config_regs->mask); + } else { + switch (config_regs->type) { + case KV_CONFIGREG_SMC_IND: + data = RREG32_SMC(config_regs->offset); + break; + case KV_CONFIGREG_DIDT_IND: + data = RREG32_DIDT(config_regs->offset); + break; + default: + data = RREG32(config_regs->offset << 2); + break; + } + + data &= ~config_regs->mask; + data |= ((config_regs->value << config_regs->shift) & config_regs->mask); + data |= cache; + cache = 0; + + switch (config_regs->type) { + case KV_CONFIGREG_SMC_IND: + WREG32_SMC(config_regs->offset, data); + break; + case KV_CONFIGREG_DIDT_IND: + WREG32_DIDT(config_regs->offset, data); + break; + default: + WREG32(config_regs->offset << 2, data); + break; + } + } + config_regs++; + } + + return 0; +} + +static void kv_do_enable_didt(struct radeon_device *rdev, bool enable) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 data; + + if (pi->caps_sq_ramping) { + data = RREG32_DIDT(DIDT_SQ_CTRL0); + if (enable) + data |= DIDT_CTRL_EN; + else + data &= ~DIDT_CTRL_EN; + WREG32_DIDT(DIDT_SQ_CTRL0, data); + } + + if (pi->caps_db_ramping) { + data = RREG32_DIDT(DIDT_DB_CTRL0); + if (enable) + data |= DIDT_CTRL_EN; + else + data &= ~DIDT_CTRL_EN; + WREG32_DIDT(DIDT_DB_CTRL0, data); + } + + if (pi->caps_td_ramping) { + data = RREG32_DIDT(DIDT_TD_CTRL0); + if (enable) + data |= DIDT_CTRL_EN; + else + data &= ~DIDT_CTRL_EN; + WREG32_DIDT(DIDT_TD_CTRL0, data); + } + + if (pi->caps_tcp_ramping) { + data = RREG32_DIDT(DIDT_TCP_CTRL0); + if (enable) + data |= DIDT_CTRL_EN; + else + data &= ~DIDT_CTRL_EN; + WREG32_DIDT(DIDT_TCP_CTRL0, data); + } +} + +static int kv_enable_didt(struct radeon_device *rdev, bool enable) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret; + + if (pi->caps_sq_ramping || + pi->caps_db_ramping || + pi->caps_td_ramping || + pi->caps_tcp_ramping) { + cik_enter_rlc_safe_mode(rdev); + + if (enable) { + ret = kv_program_pt_config_registers(rdev, didt_config_kv); + if (ret) { + cik_exit_rlc_safe_mode(rdev); + return ret; + } + } + + kv_do_enable_didt(rdev, enable); + + cik_exit_rlc_safe_mode(rdev); + } + + return 0; +} + +#if 0 +static void kv_initialize_hardware_cac_manager(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + if (pi->caps_cac) { + WREG32_SMC(LCAC_SX0_OVR_SEL, 0); + WREG32_SMC(LCAC_SX0_OVR_VAL, 0); + kv_program_local_cac_table(rdev, sx_local_cac_cfg_kv, sx0_cac_config_reg); + + WREG32_SMC(LCAC_MC0_OVR_SEL, 0); + WREG32_SMC(LCAC_MC0_OVR_VAL, 0); + kv_program_local_cac_table(rdev, mc0_local_cac_cfg_kv, mc0_cac_config_reg); + + WREG32_SMC(LCAC_MC1_OVR_SEL, 0); + WREG32_SMC(LCAC_MC1_OVR_VAL, 0); + kv_program_local_cac_table(rdev, mc1_local_cac_cfg_kv, mc1_cac_config_reg); + + WREG32_SMC(LCAC_MC2_OVR_SEL, 0); + WREG32_SMC(LCAC_MC2_OVR_VAL, 0); + kv_program_local_cac_table(rdev, mc2_local_cac_cfg_kv, mc2_cac_config_reg); + + WREG32_SMC(LCAC_MC3_OVR_SEL, 0); + WREG32_SMC(LCAC_MC3_OVR_VAL, 0); + kv_program_local_cac_table(rdev, mc3_local_cac_cfg_kv, mc3_cac_config_reg); + + WREG32_SMC(LCAC_CPL_OVR_SEL, 0); + WREG32_SMC(LCAC_CPL_OVR_VAL, 0); + kv_program_local_cac_table(rdev, cpl_local_cac_cfg_kv, cpl_cac_config_reg); + } +} +#endif + +static int kv_enable_smc_cac(struct radeon_device *rdev, bool enable) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret = 0; + + if (pi->caps_cac) { + if (enable) { + ret = kv_notify_message_to_smu(rdev, PPSMC_MSG_EnableCac); + if (ret) + pi->cac_enabled = false; + else + pi->cac_enabled = true; + } else if (pi->cac_enabled) { + kv_notify_message_to_smu(rdev, PPSMC_MSG_DisableCac); + pi->cac_enabled = false; + } + } + + return ret; +} + +static int kv_process_firmware_header(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 tmp; + int ret; + + ret = kv_read_smc_sram_dword(rdev, SMU7_FIRMWARE_HEADER_LOCATION + + offsetof(SMU7_Firmware_Header, DpmTable), + &tmp, pi->sram_end); + + if (ret == 0) + pi->dpm_table_start = tmp; + + ret = kv_read_smc_sram_dword(rdev, SMU7_FIRMWARE_HEADER_LOCATION + + offsetof(SMU7_Firmware_Header, SoftRegisters), + &tmp, pi->sram_end); + + if (ret == 0) + pi->soft_regs_start = tmp; + + return ret; +} + +static int kv_enable_dpm_voltage_scaling(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret; + + pi->graphics_voltage_change_enable = 1; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, GraphicsVoltageChangeEnable), + &pi->graphics_voltage_change_enable, + sizeof(u8), pi->sram_end); + + return ret; +} + +static int kv_set_dpm_interval(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret; + + pi->graphics_interval = 1; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, GraphicsInterval), + &pi->graphics_interval, + sizeof(u8), pi->sram_end); + + return ret; +} + +static int kv_set_dpm_boot_state(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, GraphicsBootLevel), + &pi->graphics_boot_level, + sizeof(u8), pi->sram_end); + + return ret; +} + +static void kv_program_vc(struct radeon_device *rdev) +{ + WREG32_SMC(CG_FTV_0, 0x3FFFC100); +} + +static void kv_clear_vc(struct radeon_device *rdev) +{ + WREG32_SMC(CG_FTV_0, 0); +} + +static int kv_set_divider_value(struct radeon_device *rdev, + u32 index, u32 sclk) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct atom_clock_dividers dividers; + int ret; + + ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM, + sclk, false, ÷rs); + if (ret) + return ret; + + pi->graphics_level[index].SclkDid = (u8)dividers.post_div; + pi->graphics_level[index].SclkFrequency = cpu_to_be32(sclk); + + return 0; +} + +static u32 kv_convert_vid2_to_vid7(struct radeon_device *rdev, + struct sumo_vid_mapping_table *vid_mapping_table, + u32 vid_2bit) +{ + struct radeon_clock_voltage_dependency_table *vddc_sclk_table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + u32 i; + + if (vddc_sclk_table && vddc_sclk_table->count) { + if (vid_2bit < vddc_sclk_table->count) + return vddc_sclk_table->entries[vid_2bit].v; + else + return vddc_sclk_table->entries[vddc_sclk_table->count - 1].v; + } else { + for (i = 0; i < vid_mapping_table->num_entries; i++) { + if (vid_mapping_table->entries[i].vid_2bit == vid_2bit) + return vid_mapping_table->entries[i].vid_7bit; + } + return vid_mapping_table->entries[vid_mapping_table->num_entries - 1].vid_7bit; + } +} + +static u32 kv_convert_vid7_to_vid2(struct radeon_device *rdev, + struct sumo_vid_mapping_table *vid_mapping_table, + u32 vid_7bit) +{ + struct radeon_clock_voltage_dependency_table *vddc_sclk_table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + u32 i; + + if (vddc_sclk_table && vddc_sclk_table->count) { + for (i = 0; i < vddc_sclk_table->count; i++) { + if (vddc_sclk_table->entries[i].v == vid_7bit) + return i; + } + return vddc_sclk_table->count - 1; + } else { + for (i = 0; i < vid_mapping_table->num_entries; i++) { + if (vid_mapping_table->entries[i].vid_7bit == vid_7bit) + return vid_mapping_table->entries[i].vid_2bit; + } + + return vid_mapping_table->entries[vid_mapping_table->num_entries - 1].vid_2bit; + } +} + +static u16 kv_convert_8bit_index_to_voltage(struct radeon_device *rdev, + u16 voltage) +{ + return 6200 - (voltage * 25); +} + +static u16 kv_convert_2bit_index_to_voltage(struct radeon_device *rdev, + u32 vid_2bit) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 vid_8bit = kv_convert_vid2_to_vid7(rdev, + &pi->sys_info.vid_mapping_table, + vid_2bit); + + return kv_convert_8bit_index_to_voltage(rdev, (u16)vid_8bit); +} + + +static int kv_set_vid(struct radeon_device *rdev, u32 index, u32 vid) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + pi->graphics_level[index].VoltageDownH = (u8)pi->voltage_drop_t; + pi->graphics_level[index].MinVddNb = + cpu_to_be32(kv_convert_2bit_index_to_voltage(rdev, vid)); + + return 0; +} + +static int kv_set_at(struct radeon_device *rdev, u32 index, u32 at) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + pi->graphics_level[index].AT = cpu_to_be16((u16)at); + + return 0; +} + +static void kv_dpm_power_level_enable(struct radeon_device *rdev, + u32 index, bool enable) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + pi->graphics_level[index].EnabledForActivity = enable ? 1 : 0; +} + +static void kv_start_dpm(struct radeon_device *rdev) +{ + u32 tmp = RREG32_SMC(GENERAL_PWRMGT); + + tmp |= GLOBAL_PWRMGT_EN; + WREG32_SMC(GENERAL_PWRMGT, tmp); + + kv_smc_dpm_enable(rdev, true); +} + +static void kv_stop_dpm(struct radeon_device *rdev) +{ + kv_smc_dpm_enable(rdev, false); +} + +static void kv_start_am(struct radeon_device *rdev) +{ + u32 sclk_pwrmgt_cntl = RREG32_SMC(SCLK_PWRMGT_CNTL); + + sclk_pwrmgt_cntl &= ~(RESET_SCLK_CNT | RESET_BUSY_CNT); + sclk_pwrmgt_cntl |= DYNAMIC_PM_EN; + + WREG32_SMC(SCLK_PWRMGT_CNTL, sclk_pwrmgt_cntl); +} + +static void kv_reset_am(struct radeon_device *rdev) +{ + u32 sclk_pwrmgt_cntl = RREG32_SMC(SCLK_PWRMGT_CNTL); + + sclk_pwrmgt_cntl |= (RESET_SCLK_CNT | RESET_BUSY_CNT); + + WREG32_SMC(SCLK_PWRMGT_CNTL, sclk_pwrmgt_cntl); +} + +static int kv_freeze_sclk_dpm(struct radeon_device *rdev, bool freeze) +{ + return kv_notify_message_to_smu(rdev, freeze ? + PPSMC_MSG_SCLKDPM_FreezeLevel : PPSMC_MSG_SCLKDPM_UnfreezeLevel); +} + +static int kv_force_lowest_valid(struct radeon_device *rdev) +{ + return kv_force_dpm_lowest(rdev); +} + +static int kv_unforce_levels(struct radeon_device *rdev) +{ + if (rdev->family == CHIP_KABINI || rdev->family == CHIP_MULLINS) + return kv_notify_message_to_smu(rdev, PPSMC_MSG_NoForcedLevel); + else + return kv_set_enabled_levels(rdev); +} + +static int kv_update_sclk_t(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 low_sclk_interrupt_t = 0; + int ret = 0; + + if (pi->caps_sclk_throttle_low_notification) { + low_sclk_interrupt_t = cpu_to_be32(pi->low_sclk_interrupt_t); + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, LowSclkInterruptT), + (u8 *)&low_sclk_interrupt_t, + sizeof(u32), pi->sram_end); + } + return ret; +} + +static int kv_program_bootup_state(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 i; + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + + if (table && table->count) { + for (i = pi->graphics_dpm_level_count - 1; i > 0; i--) { + if (table->entries[i].clk == pi->boot_pl.sclk) + break; + } + + pi->graphics_boot_level = (u8)i; + kv_dpm_power_level_enable(rdev, i, true); + } else { + struct sumo_sclk_voltage_mapping_table *table = + &pi->sys_info.sclk_voltage_mapping_table; + + if (table->num_max_dpm_entries == 0) + return -EINVAL; + + for (i = pi->graphics_dpm_level_count - 1; i > 0; i--) { + if (table->entries[i].sclk_frequency == pi->boot_pl.sclk) + break; + } + + pi->graphics_boot_level = (u8)i; + kv_dpm_power_level_enable(rdev, i, true); + } + return 0; +} + +static int kv_enable_auto_thermal_throttling(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret; + + pi->graphics_therm_throttle_enable = 1; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, GraphicsThermThrottleEnable), + &pi->graphics_therm_throttle_enable, + sizeof(u8), pi->sram_end); + + return ret; +} + +static int kv_upload_dpm_settings(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, GraphicsLevel), + (u8 *)&pi->graphics_level, + sizeof(SMU7_Fusion_GraphicsLevel) * SMU7_MAX_LEVELS_GRAPHICS, + pi->sram_end); + + if (ret) + return ret; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, GraphicsDpmLevelCount), + &pi->graphics_dpm_level_count, + sizeof(u8), pi->sram_end); + + return ret; +} + +static u32 kv_get_clock_difference(u32 a, u32 b) +{ + return (a >= b) ? a - b : b - a; +} + +static u32 kv_get_clk_bypass(struct radeon_device *rdev, u32 clk) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 value; + + if (pi->caps_enable_dfs_bypass) { + if (kv_get_clock_difference(clk, 40000) < 200) + value = 3; + else if (kv_get_clock_difference(clk, 30000) < 200) + value = 2; + else if (kv_get_clock_difference(clk, 20000) < 200) + value = 7; + else if (kv_get_clock_difference(clk, 15000) < 200) + value = 6; + else if (kv_get_clock_difference(clk, 10000) < 200) + value = 8; + else + value = 0; + } else { + value = 0; + } + + return value; +} + +static int kv_populate_uvd_table(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_uvd_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table; + struct atom_clock_dividers dividers; + int ret; + u32 i; + + if (table == NULL || table->count == 0) + return 0; + + pi->uvd_level_count = 0; + for (i = 0; i < table->count; i++) { + if (pi->high_voltage_t && + (pi->high_voltage_t < table->entries[i].v)) + break; + + pi->uvd_level[i].VclkFrequency = cpu_to_be32(table->entries[i].vclk); + pi->uvd_level[i].DclkFrequency = cpu_to_be32(table->entries[i].dclk); + pi->uvd_level[i].MinVddNb = cpu_to_be16(table->entries[i].v); + + pi->uvd_level[i].VClkBypassCntl = + (u8)kv_get_clk_bypass(rdev, table->entries[i].vclk); + pi->uvd_level[i].DClkBypassCntl = + (u8)kv_get_clk_bypass(rdev, table->entries[i].dclk); + + ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM, + table->entries[i].vclk, false, ÷rs); + if (ret) + return ret; + pi->uvd_level[i].VclkDivider = (u8)dividers.post_div; + + ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM, + table->entries[i].dclk, false, ÷rs); + if (ret) + return ret; + pi->uvd_level[i].DclkDivider = (u8)dividers.post_div; + + pi->uvd_level_count++; + } + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, UvdLevelCount), + (u8 *)&pi->uvd_level_count, + sizeof(u8), pi->sram_end); + if (ret) + return ret; + + pi->uvd_interval = 1; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, UVDInterval), + &pi->uvd_interval, + sizeof(u8), pi->sram_end); + if (ret) + return ret; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, UvdLevel), + (u8 *)&pi->uvd_level, + sizeof(SMU7_Fusion_UvdLevel) * SMU7_MAX_LEVELS_UVD, + pi->sram_end); + + return ret; + +} + +static int kv_populate_vce_table(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret; + u32 i; + struct radeon_vce_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table; + struct atom_clock_dividers dividers; + + if (table == NULL || table->count == 0) + return 0; + + pi->vce_level_count = 0; + for (i = 0; i < table->count; i++) { + if (pi->high_voltage_t && + pi->high_voltage_t < table->entries[i].v) + break; + + pi->vce_level[i].Frequency = cpu_to_be32(table->entries[i].evclk); + pi->vce_level[i].MinVoltage = cpu_to_be16(table->entries[i].v); + + pi->vce_level[i].ClkBypassCntl = + (u8)kv_get_clk_bypass(rdev, table->entries[i].evclk); + + ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM, + table->entries[i].evclk, false, ÷rs); + if (ret) + return ret; + pi->vce_level[i].Divider = (u8)dividers.post_div; + + pi->vce_level_count++; + } + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, VceLevelCount), + (u8 *)&pi->vce_level_count, + sizeof(u8), + pi->sram_end); + if (ret) + return ret; + + pi->vce_interval = 1; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, VCEInterval), + (u8 *)&pi->vce_interval, + sizeof(u8), + pi->sram_end); + if (ret) + return ret; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, VceLevel), + (u8 *)&pi->vce_level, + sizeof(SMU7_Fusion_ExtClkLevel) * SMU7_MAX_LEVELS_VCE, + pi->sram_end); + + return ret; +} + +static int kv_populate_samu_table(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table; + struct atom_clock_dividers dividers; + int ret; + u32 i; + + if (table == NULL || table->count == 0) + return 0; + + pi->samu_level_count = 0; + for (i = 0; i < table->count; i++) { + if (pi->high_voltage_t && + pi->high_voltage_t < table->entries[i].v) + break; + + pi->samu_level[i].Frequency = cpu_to_be32(table->entries[i].clk); + pi->samu_level[i].MinVoltage = cpu_to_be16(table->entries[i].v); + + pi->samu_level[i].ClkBypassCntl = + (u8)kv_get_clk_bypass(rdev, table->entries[i].clk); + + ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM, + table->entries[i].clk, false, ÷rs); + if (ret) + return ret; + pi->samu_level[i].Divider = (u8)dividers.post_div; + + pi->samu_level_count++; + } + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, SamuLevelCount), + (u8 *)&pi->samu_level_count, + sizeof(u8), + pi->sram_end); + if (ret) + return ret; + + pi->samu_interval = 1; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, SAMUInterval), + (u8 *)&pi->samu_interval, + sizeof(u8), + pi->sram_end); + if (ret) + return ret; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, SamuLevel), + (u8 *)&pi->samu_level, + sizeof(SMU7_Fusion_ExtClkLevel) * SMU7_MAX_LEVELS_SAMU, + pi->sram_end); + if (ret) + return ret; + + return ret; +} + + +static int kv_populate_acp_table(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table; + struct atom_clock_dividers dividers; + int ret; + u32 i; + + if (table == NULL || table->count == 0) + return 0; + + pi->acp_level_count = 0; + for (i = 0; i < table->count; i++) { + pi->acp_level[i].Frequency = cpu_to_be32(table->entries[i].clk); + pi->acp_level[i].MinVoltage = cpu_to_be16(table->entries[i].v); + + ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM, + table->entries[i].clk, false, ÷rs); + if (ret) + return ret; + pi->acp_level[i].Divider = (u8)dividers.post_div; + + pi->acp_level_count++; + } + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, AcpLevelCount), + (u8 *)&pi->acp_level_count, + sizeof(u8), + pi->sram_end); + if (ret) + return ret; + + pi->acp_interval = 1; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, ACPInterval), + (u8 *)&pi->acp_interval, + sizeof(u8), + pi->sram_end); + if (ret) + return ret; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, AcpLevel), + (u8 *)&pi->acp_level, + sizeof(SMU7_Fusion_ExtClkLevel) * SMU7_MAX_LEVELS_ACP, + pi->sram_end); + if (ret) + return ret; + + return ret; +} + +static void kv_calculate_dfs_bypass_settings(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 i; + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + + if (table && table->count) { + for (i = 0; i < pi->graphics_dpm_level_count; i++) { + if (pi->caps_enable_dfs_bypass) { + if (kv_get_clock_difference(table->entries[i].clk, 40000) < 200) + pi->graphics_level[i].ClkBypassCntl = 3; + else if (kv_get_clock_difference(table->entries[i].clk, 30000) < 200) + pi->graphics_level[i].ClkBypassCntl = 2; + else if (kv_get_clock_difference(table->entries[i].clk, 26600) < 200) + pi->graphics_level[i].ClkBypassCntl = 7; + else if (kv_get_clock_difference(table->entries[i].clk , 20000) < 200) + pi->graphics_level[i].ClkBypassCntl = 6; + else if (kv_get_clock_difference(table->entries[i].clk , 10000) < 200) + pi->graphics_level[i].ClkBypassCntl = 8; + else + pi->graphics_level[i].ClkBypassCntl = 0; + } else { + pi->graphics_level[i].ClkBypassCntl = 0; + } + } + } else { + struct sumo_sclk_voltage_mapping_table *table = + &pi->sys_info.sclk_voltage_mapping_table; + for (i = 0; i < pi->graphics_dpm_level_count; i++) { + if (pi->caps_enable_dfs_bypass) { + if (kv_get_clock_difference(table->entries[i].sclk_frequency, 40000) < 200) + pi->graphics_level[i].ClkBypassCntl = 3; + else if (kv_get_clock_difference(table->entries[i].sclk_frequency, 30000) < 200) + pi->graphics_level[i].ClkBypassCntl = 2; + else if (kv_get_clock_difference(table->entries[i].sclk_frequency, 26600) < 200) + pi->graphics_level[i].ClkBypassCntl = 7; + else if (kv_get_clock_difference(table->entries[i].sclk_frequency, 20000) < 200) + pi->graphics_level[i].ClkBypassCntl = 6; + else if (kv_get_clock_difference(table->entries[i].sclk_frequency, 10000) < 200) + pi->graphics_level[i].ClkBypassCntl = 8; + else + pi->graphics_level[i].ClkBypassCntl = 0; + } else { + pi->graphics_level[i].ClkBypassCntl = 0; + } + } + } +} + +static int kv_enable_ulv(struct radeon_device *rdev, bool enable) +{ + return kv_notify_message_to_smu(rdev, enable ? + PPSMC_MSG_EnableULV : PPSMC_MSG_DisableULV); +} + +static void kv_reset_acp_boot_level(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + pi->acp_boot_level = 0xff; +} + +static void kv_update_current_ps(struct radeon_device *rdev, + struct radeon_ps *rps) +{ + struct kv_ps *new_ps = kv_get_ps(rps); + struct kv_power_info *pi = kv_get_pi(rdev); + + pi->current_rps = *rps; + pi->current_ps = *new_ps; + pi->current_rps.ps_priv = &pi->current_ps; +} + +static void kv_update_requested_ps(struct radeon_device *rdev, + struct radeon_ps *rps) +{ + struct kv_ps *new_ps = kv_get_ps(rps); + struct kv_power_info *pi = kv_get_pi(rdev); + + pi->requested_rps = *rps; + pi->requested_ps = *new_ps; + pi->requested_rps.ps_priv = &pi->requested_ps; +} + +void kv_dpm_enable_bapm(struct radeon_device *rdev, bool enable) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret; + + if (pi->bapm_enable) { + ret = kv_smc_bapm_enable(rdev, enable); + if (ret) + DRM_ERROR("kv_smc_bapm_enable failed\n"); + } +} + +int kv_dpm_enable(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret; + + ret = kv_process_firmware_header(rdev); + if (ret) { + DRM_ERROR("kv_process_firmware_header failed\n"); + return ret; + } + kv_init_fps_limits(rdev); + kv_init_graphics_levels(rdev); + ret = kv_program_bootup_state(rdev); + if (ret) { + DRM_ERROR("kv_program_bootup_state failed\n"); + return ret; + } + kv_calculate_dfs_bypass_settings(rdev); + ret = kv_upload_dpm_settings(rdev); + if (ret) { + DRM_ERROR("kv_upload_dpm_settings failed\n"); + return ret; + } + ret = kv_populate_uvd_table(rdev); + if (ret) { + DRM_ERROR("kv_populate_uvd_table failed\n"); + return ret; + } + ret = kv_populate_vce_table(rdev); + if (ret) { + DRM_ERROR("kv_populate_vce_table failed\n"); + return ret; + } + ret = kv_populate_samu_table(rdev); + if (ret) { + DRM_ERROR("kv_populate_samu_table failed\n"); + return ret; + } + ret = kv_populate_acp_table(rdev); + if (ret) { + DRM_ERROR("kv_populate_acp_table failed\n"); + return ret; + } + kv_program_vc(rdev); +#if 0 + kv_initialize_hardware_cac_manager(rdev); +#endif + kv_start_am(rdev); + if (pi->enable_auto_thermal_throttling) { + ret = kv_enable_auto_thermal_throttling(rdev); + if (ret) { + DRM_ERROR("kv_enable_auto_thermal_throttling failed\n"); + return ret; + } + } + ret = kv_enable_dpm_voltage_scaling(rdev); + if (ret) { + DRM_ERROR("kv_enable_dpm_voltage_scaling failed\n"); + return ret; + } + ret = kv_set_dpm_interval(rdev); + if (ret) { + DRM_ERROR("kv_set_dpm_interval failed\n"); + return ret; + } + ret = kv_set_dpm_boot_state(rdev); + if (ret) { + DRM_ERROR("kv_set_dpm_boot_state failed\n"); + return ret; + } + ret = kv_enable_ulv(rdev, true); + if (ret) { + DRM_ERROR("kv_enable_ulv failed\n"); + return ret; + } + kv_start_dpm(rdev); + ret = kv_enable_didt(rdev, true); + if (ret) { + DRM_ERROR("kv_enable_didt failed\n"); + return ret; + } + ret = kv_enable_smc_cac(rdev, true); + if (ret) { + DRM_ERROR("kv_enable_smc_cac failed\n"); + return ret; + } + + kv_reset_acp_boot_level(rdev); + + ret = kv_smc_bapm_enable(rdev, false); + if (ret) { + DRM_ERROR("kv_smc_bapm_enable failed\n"); + return ret; + } + + kv_update_current_ps(rdev, rdev->pm.dpm.boot_ps); + + return ret; +} + +int kv_dpm_late_enable(struct radeon_device *rdev) +{ + int ret = 0; + + if (rdev->irq.installed && + r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { + ret = kv_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX); + if (ret) { + DRM_ERROR("kv_set_thermal_temperature_range failed\n"); + return ret; + } + rdev->irq.dpm_thermal = true; + radeon_irq_set(rdev); + } + + /* powerdown unused blocks for now */ + kv_dpm_powergate_acp(rdev, true); + kv_dpm_powergate_samu(rdev, true); + kv_dpm_powergate_vce(rdev, true); + kv_dpm_powergate_uvd(rdev, true); + + return ret; +} + +void kv_dpm_disable(struct radeon_device *rdev) +{ + kv_smc_bapm_enable(rdev, false); + + /* powerup blocks */ + kv_dpm_powergate_acp(rdev, false); + kv_dpm_powergate_samu(rdev, false); + kv_dpm_powergate_vce(rdev, false); + kv_dpm_powergate_uvd(rdev, false); + + kv_enable_smc_cac(rdev, false); + kv_enable_didt(rdev, false); + kv_clear_vc(rdev); + kv_stop_dpm(rdev); + kv_enable_ulv(rdev, false); + kv_reset_am(rdev); + + kv_update_current_ps(rdev, rdev->pm.dpm.boot_ps); +} + +#if 0 +static int kv_write_smc_soft_register(struct radeon_device *rdev, + u16 reg_offset, u32 value) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + return kv_copy_bytes_to_smc(rdev, pi->soft_regs_start + reg_offset, + (u8 *)&value, sizeof(u16), pi->sram_end); +} + +static int kv_read_smc_soft_register(struct radeon_device *rdev, + u16 reg_offset, u32 *value) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + return kv_read_smc_sram_dword(rdev, pi->soft_regs_start + reg_offset, + value, pi->sram_end); +} +#endif + +static void kv_init_sclk_t(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + pi->low_sclk_interrupt_t = 0; +} + +static int kv_init_fps_limits(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret = 0; + + if (pi->caps_fps) { + u16 tmp; + + tmp = 45; + pi->fps_high_t = cpu_to_be16(tmp); + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, FpsHighT), + (u8 *)&pi->fps_high_t, + sizeof(u16), pi->sram_end); + + tmp = 30; + pi->fps_low_t = cpu_to_be16(tmp); + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, FpsLowT), + (u8 *)&pi->fps_low_t, + sizeof(u16), pi->sram_end); + + } + return ret; +} + +static void kv_init_powergate_state(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + pi->uvd_power_gated = false; + pi->vce_power_gated = false; + pi->samu_power_gated = false; + pi->acp_power_gated = false; + +} + +static int kv_enable_uvd_dpm(struct radeon_device *rdev, bool enable) +{ + return kv_notify_message_to_smu(rdev, enable ? + PPSMC_MSG_UVDDPM_Enable : PPSMC_MSG_UVDDPM_Disable); +} + +static int kv_enable_vce_dpm(struct radeon_device *rdev, bool enable) +{ + return kv_notify_message_to_smu(rdev, enable ? + PPSMC_MSG_VCEDPM_Enable : PPSMC_MSG_VCEDPM_Disable); +} + +static int kv_enable_samu_dpm(struct radeon_device *rdev, bool enable) +{ + return kv_notify_message_to_smu(rdev, enable ? + PPSMC_MSG_SAMUDPM_Enable : PPSMC_MSG_SAMUDPM_Disable); +} + +static int kv_enable_acp_dpm(struct radeon_device *rdev, bool enable) +{ + return kv_notify_message_to_smu(rdev, enable ? + PPSMC_MSG_ACPDPM_Enable : PPSMC_MSG_ACPDPM_Disable); +} + +static int kv_update_uvd_dpm(struct radeon_device *rdev, bool gate) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_uvd_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table; + int ret; + u32 mask; + + if (!gate) { + if (table->count) + pi->uvd_boot_level = table->count - 1; + else + pi->uvd_boot_level = 0; + + if (!pi->caps_uvd_dpm || pi->caps_stable_p_state) { + mask = 1 << pi->uvd_boot_level; + } else { + mask = 0x1f; + } + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, UvdBootLevel), + (uint8_t *)&pi->uvd_boot_level, + sizeof(u8), pi->sram_end); + if (ret) + return ret; + + kv_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_UVDDPM_SetEnabledMask, + mask); + } + + return kv_enable_uvd_dpm(rdev, !gate); +} + +static u8 kv_get_vce_boot_level(struct radeon_device *rdev) +{ + u8 i; + struct radeon_vce_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table; + + for (i = 0; i < table->count; i++) { + if (table->entries[i].evclk >= 0) /* XXX */ + break; + } + + return i; +} + +static int kv_update_vce_dpm(struct radeon_device *rdev, + struct radeon_ps *radeon_new_state, + struct radeon_ps *radeon_current_state) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_vce_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table; + int ret; + + if (radeon_new_state->evclk > 0 && radeon_current_state->evclk == 0) { + kv_dpm_powergate_vce(rdev, false); + /* turn the clocks on when encoding */ + cik_update_cg(rdev, RADEON_CG_BLOCK_VCE, false); + if (pi->caps_stable_p_state) + pi->vce_boot_level = table->count - 1; + else + pi->vce_boot_level = kv_get_vce_boot_level(rdev); + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, VceBootLevel), + (u8 *)&pi->vce_boot_level, + sizeof(u8), + pi->sram_end); + if (ret) + return ret; + + if (pi->caps_stable_p_state) + kv_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_VCEDPM_SetEnabledMask, + (1 << pi->vce_boot_level)); + + kv_enable_vce_dpm(rdev, true); + } else if (radeon_new_state->evclk == 0 && radeon_current_state->evclk > 0) { + kv_enable_vce_dpm(rdev, false); + /* turn the clocks off when not encoding */ + cik_update_cg(rdev, RADEON_CG_BLOCK_VCE, true); + kv_dpm_powergate_vce(rdev, true); + } + + return 0; +} + +static int kv_update_samu_dpm(struct radeon_device *rdev, bool gate) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table; + int ret; + + if (!gate) { + if (pi->caps_stable_p_state) + pi->samu_boot_level = table->count - 1; + else + pi->samu_boot_level = 0; + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, SamuBootLevel), + (u8 *)&pi->samu_boot_level, + sizeof(u8), + pi->sram_end); + if (ret) + return ret; + + if (pi->caps_stable_p_state) + kv_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_SAMUDPM_SetEnabledMask, + (1 << pi->samu_boot_level)); + } + + return kv_enable_samu_dpm(rdev, !gate); +} + +static u8 kv_get_acp_boot_level(struct radeon_device *rdev) +{ + u8 i; + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table; + + for (i = 0; i < table->count; i++) { + if (table->entries[i].clk >= 0) /* XXX */ + break; + } + + if (i >= table->count) + i = table->count - 1; + + return i; +} + +static void kv_update_acp_boot_level(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u8 acp_boot_level; + + if (!pi->caps_stable_p_state) { + acp_boot_level = kv_get_acp_boot_level(rdev); + if (acp_boot_level != pi->acp_boot_level) { + pi->acp_boot_level = acp_boot_level; + kv_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_ACPDPM_SetEnabledMask, + (1 << pi->acp_boot_level)); + } + } +} + +static int kv_update_acp_dpm(struct radeon_device *rdev, bool gate) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table; + int ret; + + if (!gate) { + if (pi->caps_stable_p_state) + pi->acp_boot_level = table->count - 1; + else + pi->acp_boot_level = kv_get_acp_boot_level(rdev); + + ret = kv_copy_bytes_to_smc(rdev, + pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, AcpBootLevel), + (u8 *)&pi->acp_boot_level, + sizeof(u8), + pi->sram_end); + if (ret) + return ret; + + if (pi->caps_stable_p_state) + kv_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_ACPDPM_SetEnabledMask, + (1 << pi->acp_boot_level)); + } + + return kv_enable_acp_dpm(rdev, !gate); +} + +void kv_dpm_powergate_uvd(struct radeon_device *rdev, bool gate) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + if (pi->uvd_power_gated == gate) + return; + + pi->uvd_power_gated = gate; + + if (gate) { + if (pi->caps_uvd_pg) { + uvd_v1_0_stop(rdev); + cik_update_cg(rdev, RADEON_CG_BLOCK_UVD, false); + } + kv_update_uvd_dpm(rdev, gate); + if (pi->caps_uvd_pg) + kv_notify_message_to_smu(rdev, PPSMC_MSG_UVDPowerOFF); + } else { + if (pi->caps_uvd_pg) { + kv_notify_message_to_smu(rdev, PPSMC_MSG_UVDPowerON); + uvd_v4_2_resume(rdev); + uvd_v1_0_start(rdev); + cik_update_cg(rdev, RADEON_CG_BLOCK_UVD, true); + } + kv_update_uvd_dpm(rdev, gate); + } +} + +static void kv_dpm_powergate_vce(struct radeon_device *rdev, bool gate) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + if (pi->vce_power_gated == gate) + return; + + pi->vce_power_gated = gate; + + if (gate) { + if (pi->caps_vce_pg) { + /* XXX do we need a vce_v1_0_stop() ? */ + kv_notify_message_to_smu(rdev, PPSMC_MSG_VCEPowerOFF); + } + } else { + if (pi->caps_vce_pg) { + kv_notify_message_to_smu(rdev, PPSMC_MSG_VCEPowerON); + vce_v2_0_resume(rdev); + vce_v1_0_start(rdev); + } + } +} + +static void kv_dpm_powergate_samu(struct radeon_device *rdev, bool gate) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + if (pi->samu_power_gated == gate) + return; + + pi->samu_power_gated = gate; + + if (gate) { + kv_update_samu_dpm(rdev, true); + if (pi->caps_samu_pg) + kv_notify_message_to_smu(rdev, PPSMC_MSG_SAMPowerOFF); + } else { + if (pi->caps_samu_pg) + kv_notify_message_to_smu(rdev, PPSMC_MSG_SAMPowerON); + kv_update_samu_dpm(rdev, false); + } +} + +static void kv_dpm_powergate_acp(struct radeon_device *rdev, bool gate) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + if (pi->acp_power_gated == gate) + return; + + if (rdev->family == CHIP_KABINI || rdev->family == CHIP_MULLINS) + return; + + pi->acp_power_gated = gate; + + if (gate) { + kv_update_acp_dpm(rdev, true); + if (pi->caps_acp_pg) + kv_notify_message_to_smu(rdev, PPSMC_MSG_ACPPowerOFF); + } else { + if (pi->caps_acp_pg) + kv_notify_message_to_smu(rdev, PPSMC_MSG_ACPPowerON); + kv_update_acp_dpm(rdev, false); + } +} + +static void kv_set_valid_clock_range(struct radeon_device *rdev, + struct radeon_ps *new_rps) +{ + struct kv_ps *new_ps = kv_get_ps(new_rps); + struct kv_power_info *pi = kv_get_pi(rdev); + u32 i; + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + + if (table && table->count) { + for (i = 0; i < pi->graphics_dpm_level_count; i++) { + if ((table->entries[i].clk >= new_ps->levels[0].sclk) || + (i == (pi->graphics_dpm_level_count - 1))) { + pi->lowest_valid = i; + break; + } + } + + for (i = pi->graphics_dpm_level_count - 1; i > 0; i--) { + if (table->entries[i].clk <= new_ps->levels[new_ps->num_levels - 1].sclk) + break; + } + pi->highest_valid = i; + + if (pi->lowest_valid > pi->highest_valid) { + if ((new_ps->levels[0].sclk - table->entries[pi->highest_valid].clk) > + (table->entries[pi->lowest_valid].clk - new_ps->levels[new_ps->num_levels - 1].sclk)) + pi->highest_valid = pi->lowest_valid; + else + pi->lowest_valid = pi->highest_valid; + } + } else { + struct sumo_sclk_voltage_mapping_table *table = + &pi->sys_info.sclk_voltage_mapping_table; + + for (i = 0; i < (int)pi->graphics_dpm_level_count; i++) { + if (table->entries[i].sclk_frequency >= new_ps->levels[0].sclk || + i == (int)(pi->graphics_dpm_level_count - 1)) { + pi->lowest_valid = i; + break; + } + } + + for (i = pi->graphics_dpm_level_count - 1; i > 0; i--) { + if (table->entries[i].sclk_frequency <= + new_ps->levels[new_ps->num_levels - 1].sclk) + break; + } + pi->highest_valid = i; + + if (pi->lowest_valid > pi->highest_valid) { + if ((new_ps->levels[0].sclk - + table->entries[pi->highest_valid].sclk_frequency) > + (table->entries[pi->lowest_valid].sclk_frequency - + new_ps->levels[new_ps->num_levels -1].sclk)) + pi->highest_valid = pi->lowest_valid; + else + pi->lowest_valid = pi->highest_valid; + } + } +} + +static int kv_update_dfs_bypass_settings(struct radeon_device *rdev, + struct radeon_ps *new_rps) +{ + struct kv_ps *new_ps = kv_get_ps(new_rps); + struct kv_power_info *pi = kv_get_pi(rdev); + int ret = 0; + u8 clk_bypass_cntl; + + if (pi->caps_enable_dfs_bypass) { + clk_bypass_cntl = new_ps->need_dfs_bypass ? + pi->graphics_level[pi->graphics_boot_level].ClkBypassCntl : 0; + ret = kv_copy_bytes_to_smc(rdev, + (pi->dpm_table_start + + offsetof(SMU7_Fusion_DpmTable, GraphicsLevel) + + (pi->graphics_boot_level * sizeof(SMU7_Fusion_GraphicsLevel)) + + offsetof(SMU7_Fusion_GraphicsLevel, ClkBypassCntl)), + &clk_bypass_cntl, + sizeof(u8), pi->sram_end); + } + + return ret; +} + +static int kv_enable_nb_dpm(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + int ret = 0; + + if (pi->enable_nb_dpm && !pi->nb_dpm_enabled) { + ret = kv_notify_message_to_smu(rdev, PPSMC_MSG_NBDPM_Enable); + if (ret == 0) + pi->nb_dpm_enabled = true; + } + + return ret; +} + +int kv_dpm_force_performance_level(struct radeon_device *rdev, + enum radeon_dpm_forced_level level) +{ + int ret; + + if (level == RADEON_DPM_FORCED_LEVEL_HIGH) { + ret = kv_force_dpm_highest(rdev); + if (ret) + return ret; + } else if (level == RADEON_DPM_FORCED_LEVEL_LOW) { + ret = kv_force_dpm_lowest(rdev); + if (ret) + return ret; + } else if (level == RADEON_DPM_FORCED_LEVEL_AUTO) { + ret = kv_unforce_levels(rdev); + if (ret) + return ret; + } + + rdev->pm.dpm.forced_level = level; + + return 0; +} + +int kv_dpm_pre_set_power_state(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps; + struct radeon_ps *new_ps = &requested_ps; + + kv_update_requested_ps(rdev, new_ps); + + kv_apply_state_adjust_rules(rdev, + &pi->requested_rps, + &pi->current_rps); + + return 0; +} + +int kv_dpm_set_power_state(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_ps *new_ps = &pi->requested_rps; + struct radeon_ps *old_ps = &pi->current_rps; + int ret; + + if (pi->bapm_enable) { + ret = kv_smc_bapm_enable(rdev, rdev->pm.dpm.ac_power); + if (ret) { + DRM_ERROR("kv_smc_bapm_enable failed\n"); + return ret; + } + } + + if (rdev->family == CHIP_KABINI || rdev->family == CHIP_MULLINS) { + if (pi->enable_dpm) { + kv_set_valid_clock_range(rdev, new_ps); + kv_update_dfs_bypass_settings(rdev, new_ps); + ret = kv_calculate_ds_divider(rdev); + if (ret) { + DRM_ERROR("kv_calculate_ds_divider failed\n"); + return ret; + } + kv_calculate_nbps_level_settings(rdev); + kv_calculate_dpm_settings(rdev); + kv_force_lowest_valid(rdev); + kv_enable_new_levels(rdev); + kv_upload_dpm_settings(rdev); + kv_program_nbps_index_settings(rdev, new_ps); + kv_unforce_levels(rdev); + kv_set_enabled_levels(rdev); + kv_force_lowest_valid(rdev); + kv_unforce_levels(rdev); + + ret = kv_update_vce_dpm(rdev, new_ps, old_ps); + if (ret) { + DRM_ERROR("kv_update_vce_dpm failed\n"); + return ret; + } + kv_update_sclk_t(rdev); + if (rdev->family == CHIP_MULLINS) + kv_enable_nb_dpm(rdev); + } + } else { + if (pi->enable_dpm) { + kv_set_valid_clock_range(rdev, new_ps); + kv_update_dfs_bypass_settings(rdev, new_ps); + ret = kv_calculate_ds_divider(rdev); + if (ret) { + DRM_ERROR("kv_calculate_ds_divider failed\n"); + return ret; + } + kv_calculate_nbps_level_settings(rdev); + kv_calculate_dpm_settings(rdev); + kv_freeze_sclk_dpm(rdev, true); + kv_upload_dpm_settings(rdev); + kv_program_nbps_index_settings(rdev, new_ps); + kv_freeze_sclk_dpm(rdev, false); + kv_set_enabled_levels(rdev); + ret = kv_update_vce_dpm(rdev, new_ps, old_ps); + if (ret) { + DRM_ERROR("kv_update_vce_dpm failed\n"); + return ret; + } + kv_update_acp_boot_level(rdev); + kv_update_sclk_t(rdev); + kv_enable_nb_dpm(rdev); + } + } + + return 0; +} + +void kv_dpm_post_set_power_state(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_ps *new_ps = &pi->requested_rps; + + kv_update_current_ps(rdev, new_ps); +} + +void kv_dpm_setup_asic(struct radeon_device *rdev) +{ + sumo_take_smu_control(rdev, true); + kv_init_powergate_state(rdev); + kv_init_sclk_t(rdev); +} + +void kv_dpm_reset_asic(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + if (rdev->family == CHIP_KABINI || rdev->family == CHIP_MULLINS) { + kv_force_lowest_valid(rdev); + kv_init_graphics_levels(rdev); + kv_program_bootup_state(rdev); + kv_upload_dpm_settings(rdev); + kv_force_lowest_valid(rdev); + kv_unforce_levels(rdev); + } else { + kv_init_graphics_levels(rdev); + kv_program_bootup_state(rdev); + kv_freeze_sclk_dpm(rdev, true); + kv_upload_dpm_settings(rdev); + kv_freeze_sclk_dpm(rdev, false); + kv_set_enabled_level(rdev, pi->graphics_boot_level); + } +} + +//XXX use sumo_dpm_display_configuration_changed + +static void kv_construct_max_power_limits_table(struct radeon_device *rdev, + struct radeon_clock_and_voltage_limits *table) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + if (pi->sys_info.sclk_voltage_mapping_table.num_max_dpm_entries > 0) { + int idx = pi->sys_info.sclk_voltage_mapping_table.num_max_dpm_entries - 1; + table->sclk = + pi->sys_info.sclk_voltage_mapping_table.entries[idx].sclk_frequency; + table->vddc = + kv_convert_2bit_index_to_voltage(rdev, + pi->sys_info.sclk_voltage_mapping_table.entries[idx].vid_2bit); + } + + table->mclk = pi->sys_info.nbp_memory_clock[0]; +} + +static void kv_patch_voltage_values(struct radeon_device *rdev) +{ + int i; + struct radeon_uvd_clock_voltage_dependency_table *uvd_table = + &rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table; + struct radeon_vce_clock_voltage_dependency_table *vce_table = + &rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table; + struct radeon_clock_voltage_dependency_table *samu_table = + &rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table; + struct radeon_clock_voltage_dependency_table *acp_table = + &rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table; + + if (uvd_table->count) { + for (i = 0; i < uvd_table->count; i++) + uvd_table->entries[i].v = + kv_convert_8bit_index_to_voltage(rdev, + uvd_table->entries[i].v); + } + + if (vce_table->count) { + for (i = 0; i < vce_table->count; i++) + vce_table->entries[i].v = + kv_convert_8bit_index_to_voltage(rdev, + vce_table->entries[i].v); + } + + if (samu_table->count) { + for (i = 0; i < samu_table->count; i++) + samu_table->entries[i].v = + kv_convert_8bit_index_to_voltage(rdev, + samu_table->entries[i].v); + } + + if (acp_table->count) { + for (i = 0; i < acp_table->count; i++) + acp_table->entries[i].v = + kv_convert_8bit_index_to_voltage(rdev, + acp_table->entries[i].v); + } + +} + +static void kv_construct_boot_state(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + pi->boot_pl.sclk = pi->sys_info.bootup_sclk; + pi->boot_pl.vddc_index = pi->sys_info.bootup_nb_voltage_index; + pi->boot_pl.ds_divider_index = 0; + pi->boot_pl.ss_divider_index = 0; + pi->boot_pl.allow_gnb_slow = 1; + pi->boot_pl.force_nbp_state = 0; + pi->boot_pl.display_wm = 0; + pi->boot_pl.vce_wm = 0; +} + +static int kv_force_dpm_highest(struct radeon_device *rdev) +{ + int ret; + u32 enable_mask, i; + + ret = kv_dpm_get_enable_mask(rdev, &enable_mask); + if (ret) + return ret; + + for (i = SMU7_MAX_LEVELS_GRAPHICS - 1; i > 0; i--) { + if (enable_mask & (1 << i)) + break; + } + + if (rdev->family == CHIP_KABINI || rdev->family == CHIP_MULLINS) + return kv_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_DPM_ForceState, i); + else + return kv_set_enabled_level(rdev, i); +} + +static int kv_force_dpm_lowest(struct radeon_device *rdev) +{ + int ret; + u32 enable_mask, i; + + ret = kv_dpm_get_enable_mask(rdev, &enable_mask); + if (ret) + return ret; + + for (i = 0; i < SMU7_MAX_LEVELS_GRAPHICS; i++) { + if (enable_mask & (1 << i)) + break; + } + + if (rdev->family == CHIP_KABINI || rdev->family == CHIP_MULLINS) + return kv_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_DPM_ForceState, i); + else + return kv_set_enabled_level(rdev, i); +} + +static u8 kv_get_sleep_divider_id_from_clock(struct radeon_device *rdev, + u32 sclk, u32 min_sclk_in_sr) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 i; + u32 temp; + u32 min = (min_sclk_in_sr > KV_MINIMUM_ENGINE_CLOCK) ? + min_sclk_in_sr : KV_MINIMUM_ENGINE_CLOCK; + + if (sclk < min) + return 0; + + if (!pi->caps_sclk_ds) + return 0; + + for (i = KV_MAX_DEEPSLEEP_DIVIDER_ID; i > 0; i--) { + temp = sclk / sumo_get_sleep_divider_from_id(i); + if (temp >= min) + break; + } + + return (u8)i; +} + +static int kv_get_high_voltage_limit(struct radeon_device *rdev, int *limit) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + int i; + + if (table && table->count) { + for (i = table->count - 1; i >= 0; i--) { + if (pi->high_voltage_t && + (kv_convert_8bit_index_to_voltage(rdev, table->entries[i].v) <= + pi->high_voltage_t)) { + *limit = i; + return 0; + } + } + } else { + struct sumo_sclk_voltage_mapping_table *table = + &pi->sys_info.sclk_voltage_mapping_table; + + for (i = table->num_max_dpm_entries - 1; i >= 0; i--) { + if (pi->high_voltage_t && + (kv_convert_2bit_index_to_voltage(rdev, table->entries[i].vid_2bit) <= + pi->high_voltage_t)) { + *limit = i; + return 0; + } + } + } + + *limit = 0; + return 0; +} + +static void kv_apply_state_adjust_rules(struct radeon_device *rdev, + struct radeon_ps *new_rps, + struct radeon_ps *old_rps) +{ + struct kv_ps *ps = kv_get_ps(new_rps); + struct kv_power_info *pi = kv_get_pi(rdev); + u32 min_sclk = 10000; /* ??? */ + u32 sclk, mclk = 0; + int i, limit; + bool force_high; + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + u32 stable_p_state_sclk = 0; + struct radeon_clock_and_voltage_limits *max_limits = + &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + + if (new_rps->vce_active) { + new_rps->evclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].evclk; + new_rps->ecclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].ecclk; + } else { + new_rps->evclk = 0; + new_rps->ecclk = 0; + } + + mclk = max_limits->mclk; + sclk = min_sclk; + + if (pi->caps_stable_p_state) { + stable_p_state_sclk = (max_limits->sclk * 75) / 100; + + for (i = table->count - 1; i >= 0; i++) { + if (stable_p_state_sclk >= table->entries[i].clk) { + stable_p_state_sclk = table->entries[i].clk; + break; + } + } + + if (i > 0) + stable_p_state_sclk = table->entries[0].clk; + + sclk = stable_p_state_sclk; + } + + if (new_rps->vce_active) { + if (sclk < rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].sclk) + sclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].sclk; + } + + ps->need_dfs_bypass = true; + + for (i = 0; i < ps->num_levels; i++) { + if (ps->levels[i].sclk < sclk) + ps->levels[i].sclk = sclk; + } + + if (table && table->count) { + for (i = 0; i < ps->num_levels; i++) { + if (pi->high_voltage_t && + (pi->high_voltage_t < + kv_convert_8bit_index_to_voltage(rdev, ps->levels[i].vddc_index))) { + kv_get_high_voltage_limit(rdev, &limit); + ps->levels[i].sclk = table->entries[limit].clk; + } + } + } else { + struct sumo_sclk_voltage_mapping_table *table = + &pi->sys_info.sclk_voltage_mapping_table; + + for (i = 0; i < ps->num_levels; i++) { + if (pi->high_voltage_t && + (pi->high_voltage_t < + kv_convert_8bit_index_to_voltage(rdev, ps->levels[i].vddc_index))) { + kv_get_high_voltage_limit(rdev, &limit); + ps->levels[i].sclk = table->entries[limit].sclk_frequency; + } + } + } + + if (pi->caps_stable_p_state) { + for (i = 0; i < ps->num_levels; i++) { + ps->levels[i].sclk = stable_p_state_sclk; + } + } + + pi->video_start = new_rps->dclk || new_rps->vclk || + new_rps->evclk || new_rps->ecclk; + + if ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) == + ATOM_PPLIB_CLASSIFICATION_UI_BATTERY) + pi->battery_state = true; + else + pi->battery_state = false; + + if (rdev->family == CHIP_KABINI || rdev->family == CHIP_MULLINS) { + ps->dpm0_pg_nb_ps_lo = 0x1; + ps->dpm0_pg_nb_ps_hi = 0x0; + ps->dpmx_nb_ps_lo = 0x1; + ps->dpmx_nb_ps_hi = 0x0; + } else { + ps->dpm0_pg_nb_ps_lo = 0x3; + ps->dpm0_pg_nb_ps_hi = 0x0; + ps->dpmx_nb_ps_lo = 0x3; + ps->dpmx_nb_ps_hi = 0x0; + + if (pi->sys_info.nb_dpm_enable) { + force_high = (mclk >= pi->sys_info.nbp_memory_clock[3]) || + pi->video_start || (rdev->pm.dpm.new_active_crtc_count >= 3) || + pi->disable_nb_ps3_in_battery; + ps->dpm0_pg_nb_ps_lo = force_high ? 0x2 : 0x3; + ps->dpm0_pg_nb_ps_hi = 0x2; + ps->dpmx_nb_ps_lo = force_high ? 0x2 : 0x3; + ps->dpmx_nb_ps_hi = 0x2; + } + } +} + +static void kv_dpm_power_level_enabled_for_throttle(struct radeon_device *rdev, + u32 index, bool enable) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + pi->graphics_level[index].EnabledForThrottle = enable ? 1 : 0; +} + +static int kv_calculate_ds_divider(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 sclk_in_sr = 10000; /* ??? */ + u32 i; + + if (pi->lowest_valid > pi->highest_valid) + return -EINVAL; + + for (i = pi->lowest_valid; i <= pi->highest_valid; i++) { + pi->graphics_level[i].DeepSleepDivId = + kv_get_sleep_divider_id_from_clock(rdev, + be32_to_cpu(pi->graphics_level[i].SclkFrequency), + sclk_in_sr); + } + return 0; +} + +static int kv_calculate_nbps_level_settings(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 i; + bool force_high; + struct radeon_clock_and_voltage_limits *max_limits = + &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + u32 mclk = max_limits->mclk; + + if (pi->lowest_valid > pi->highest_valid) + return -EINVAL; + + if (rdev->family == CHIP_KABINI || rdev->family == CHIP_MULLINS) { + for (i = pi->lowest_valid; i <= pi->highest_valid; i++) { + pi->graphics_level[i].GnbSlow = 1; + pi->graphics_level[i].ForceNbPs1 = 0; + pi->graphics_level[i].UpH = 0; + } + + if (!pi->sys_info.nb_dpm_enable) + return 0; + + force_high = ((mclk >= pi->sys_info.nbp_memory_clock[3]) || + (rdev->pm.dpm.new_active_crtc_count >= 3) || pi->video_start); + + if (force_high) { + for (i = pi->lowest_valid; i <= pi->highest_valid; i++) + pi->graphics_level[i].GnbSlow = 0; + } else { + if (pi->battery_state) + pi->graphics_level[0].ForceNbPs1 = 1; + + pi->graphics_level[1].GnbSlow = 0; + pi->graphics_level[2].GnbSlow = 0; + pi->graphics_level[3].GnbSlow = 0; + pi->graphics_level[4].GnbSlow = 0; + } + } else { + for (i = pi->lowest_valid; i <= pi->highest_valid; i++) { + pi->graphics_level[i].GnbSlow = 1; + pi->graphics_level[i].ForceNbPs1 = 0; + pi->graphics_level[i].UpH = 0; + } + + if (pi->sys_info.nb_dpm_enable && pi->battery_state) { + pi->graphics_level[pi->lowest_valid].UpH = 0x28; + pi->graphics_level[pi->lowest_valid].GnbSlow = 0; + if (pi->lowest_valid != pi->highest_valid) + pi->graphics_level[pi->lowest_valid].ForceNbPs1 = 1; + } + } + return 0; +} + +static int kv_calculate_dpm_settings(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 i; + + if (pi->lowest_valid > pi->highest_valid) + return -EINVAL; + + for (i = pi->lowest_valid; i <= pi->highest_valid; i++) + pi->graphics_level[i].DisplayWatermark = (i == pi->highest_valid) ? 1 : 0; + + return 0; +} + +static void kv_init_graphics_levels(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 i; + struct radeon_clock_voltage_dependency_table *table = + &rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk; + + if (table && table->count) { + u32 vid_2bit; + + pi->graphics_dpm_level_count = 0; + for (i = 0; i < table->count; i++) { + if (pi->high_voltage_t && + (pi->high_voltage_t < + kv_convert_8bit_index_to_voltage(rdev, table->entries[i].v))) + break; + + kv_set_divider_value(rdev, i, table->entries[i].clk); + vid_2bit = kv_convert_vid7_to_vid2(rdev, + &pi->sys_info.vid_mapping_table, + table->entries[i].v); + kv_set_vid(rdev, i, vid_2bit); + kv_set_at(rdev, i, pi->at[i]); + kv_dpm_power_level_enabled_for_throttle(rdev, i, true); + pi->graphics_dpm_level_count++; + } + } else { + struct sumo_sclk_voltage_mapping_table *table = + &pi->sys_info.sclk_voltage_mapping_table; + + pi->graphics_dpm_level_count = 0; + for (i = 0; i < table->num_max_dpm_entries; i++) { + if (pi->high_voltage_t && + pi->high_voltage_t < + kv_convert_2bit_index_to_voltage(rdev, table->entries[i].vid_2bit)) + break; + + kv_set_divider_value(rdev, i, table->entries[i].sclk_frequency); + kv_set_vid(rdev, i, table->entries[i].vid_2bit); + kv_set_at(rdev, i, pi->at[i]); + kv_dpm_power_level_enabled_for_throttle(rdev, i, true); + pi->graphics_dpm_level_count++; + } + } + + for (i = 0; i < SMU7_MAX_LEVELS_GRAPHICS; i++) + kv_dpm_power_level_enable(rdev, i, false); +} + +static void kv_enable_new_levels(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 i; + + for (i = 0; i < SMU7_MAX_LEVELS_GRAPHICS; i++) { + if (i >= pi->lowest_valid && i <= pi->highest_valid) + kv_dpm_power_level_enable(rdev, i, true); + } +} + +static int kv_set_enabled_level(struct radeon_device *rdev, u32 level) +{ + u32 new_mask = (1 << level); + + return kv_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_SCLKDPM_SetEnabledMask, + new_mask); +} + +static int kv_set_enabled_levels(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 i, new_mask = 0; + + for (i = pi->lowest_valid; i <= pi->highest_valid; i++) + new_mask |= (1 << i); + + return kv_send_msg_to_smc_with_parameter(rdev, + PPSMC_MSG_SCLKDPM_SetEnabledMask, + new_mask); +} + +static void kv_program_nbps_index_settings(struct radeon_device *rdev, + struct radeon_ps *new_rps) +{ + struct kv_ps *new_ps = kv_get_ps(new_rps); + struct kv_power_info *pi = kv_get_pi(rdev); + u32 nbdpmconfig1; + + if (rdev->family == CHIP_KABINI || rdev->family == CHIP_MULLINS) + return; + + if (pi->sys_info.nb_dpm_enable) { + nbdpmconfig1 = RREG32_SMC(NB_DPM_CONFIG_1); + nbdpmconfig1 &= ~(Dpm0PgNbPsLo_MASK | Dpm0PgNbPsHi_MASK | + DpmXNbPsLo_MASK | DpmXNbPsHi_MASK); + nbdpmconfig1 |= (Dpm0PgNbPsLo(new_ps->dpm0_pg_nb_ps_lo) | + Dpm0PgNbPsHi(new_ps->dpm0_pg_nb_ps_hi) | + DpmXNbPsLo(new_ps->dpmx_nb_ps_lo) | + DpmXNbPsHi(new_ps->dpmx_nb_ps_hi)); + WREG32_SMC(NB_DPM_CONFIG_1, nbdpmconfig1); + } +} + +static int kv_set_thermal_temperature_range(struct radeon_device *rdev, + int min_temp, int max_temp) +{ + int low_temp = 0 * 1000; + int high_temp = 255 * 1000; + u32 tmp; + + if (low_temp < min_temp) + low_temp = min_temp; + if (high_temp > max_temp) + high_temp = max_temp; + if (high_temp < low_temp) { + DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp); + return -EINVAL; + } + + tmp = RREG32_SMC(CG_THERMAL_INT_CTRL); + tmp &= ~(DIG_THERM_INTH_MASK | DIG_THERM_INTL_MASK); + tmp |= (DIG_THERM_INTH(49 + (high_temp / 1000)) | + DIG_THERM_INTL(49 + (low_temp / 1000))); + WREG32_SMC(CG_THERMAL_INT_CTRL, tmp); + + rdev->pm.dpm.thermal.min_temp = low_temp; + rdev->pm.dpm.thermal.max_temp = high_temp; + + return 0; +} + +union igp_info { + struct _ATOM_INTEGRATED_SYSTEM_INFO info; + struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 info_2; + struct _ATOM_INTEGRATED_SYSTEM_INFO_V5 info_5; + struct _ATOM_INTEGRATED_SYSTEM_INFO_V6 info_6; + struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7 info_7; + struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_8 info_8; +}; + +static int kv_parse_sys_info_table(struct radeon_device *rdev) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct radeon_mode_info *mode_info = &rdev->mode_info; + int index = GetIndexIntoMasterTable(DATA, IntegratedSystemInfo); + union igp_info *igp_info; + u8 frev, crev; + u16 data_offset; + int i; + + if (atom_parse_data_header(mode_info->atom_context, index, NULL, + &frev, &crev, &data_offset)) { + igp_info = (union igp_info *)(mode_info->atom_context->bios + + data_offset); + + if (crev != 8) { + DRM_ERROR("Unsupported IGP table: %d %d\n", frev, crev); + return -EINVAL; + } + pi->sys_info.bootup_sclk = le32_to_cpu(igp_info->info_8.ulBootUpEngineClock); + pi->sys_info.bootup_uma_clk = le32_to_cpu(igp_info->info_8.ulBootUpUMAClock); + pi->sys_info.bootup_nb_voltage_index = + le16_to_cpu(igp_info->info_8.usBootUpNBVoltage); + if (igp_info->info_8.ucHtcTmpLmt == 0) + pi->sys_info.htc_tmp_lmt = 203; + else + pi->sys_info.htc_tmp_lmt = igp_info->info_8.ucHtcTmpLmt; + if (igp_info->info_8.ucHtcHystLmt == 0) + pi->sys_info.htc_hyst_lmt = 5; + else + pi->sys_info.htc_hyst_lmt = igp_info->info_8.ucHtcHystLmt; + if (pi->sys_info.htc_tmp_lmt <= pi->sys_info.htc_hyst_lmt) { + DRM_ERROR("The htcTmpLmt should be larger than htcHystLmt.\n"); + } + + if (le32_to_cpu(igp_info->info_8.ulSystemConfig) & (1 << 3)) + pi->sys_info.nb_dpm_enable = true; + else + pi->sys_info.nb_dpm_enable = false; + + for (i = 0; i < KV_NUM_NBPSTATES; i++) { + pi->sys_info.nbp_memory_clock[i] = + le32_to_cpu(igp_info->info_8.ulNbpStateMemclkFreq[i]); + pi->sys_info.nbp_n_clock[i] = + le32_to_cpu(igp_info->info_8.ulNbpStateNClkFreq[i]); + } + if (le32_to_cpu(igp_info->info_8.ulGPUCapInfo) & + SYS_INFO_GPUCAPS__ENABEL_DFS_BYPASS) + pi->caps_enable_dfs_bypass = true; + + sumo_construct_sclk_voltage_mapping_table(rdev, + &pi->sys_info.sclk_voltage_mapping_table, + igp_info->info_8.sAvail_SCLK); + + sumo_construct_vid_mapping_table(rdev, + &pi->sys_info.vid_mapping_table, + igp_info->info_8.sAvail_SCLK); + + kv_construct_max_power_limits_table(rdev, + &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac); + } + return 0; +} + +union power_info { + struct _ATOM_POWERPLAY_INFO info; + struct _ATOM_POWERPLAY_INFO_V2 info_2; + struct _ATOM_POWERPLAY_INFO_V3 info_3; + struct _ATOM_PPLIB_POWERPLAYTABLE pplib; + struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2; + struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3; +}; + +union pplib_clock_info { + struct _ATOM_PPLIB_R600_CLOCK_INFO r600; + struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780; + struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen; + struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo; +}; + +union pplib_power_state { + struct _ATOM_PPLIB_STATE v1; + struct _ATOM_PPLIB_STATE_V2 v2; +}; + +static void kv_patch_boot_state(struct radeon_device *rdev, + struct kv_ps *ps) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + ps->num_levels = 1; + ps->levels[0] = pi->boot_pl; +} + +static void kv_parse_pplib_non_clock_info(struct radeon_device *rdev, + struct radeon_ps *rps, + struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info, + u8 table_rev) +{ + struct kv_ps *ps = kv_get_ps(rps); + + rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings); + rps->class = le16_to_cpu(non_clock_info->usClassification); + rps->class2 = le16_to_cpu(non_clock_info->usClassification2); + + if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) { + rps->vclk = le32_to_cpu(non_clock_info->ulVCLK); + rps->dclk = le32_to_cpu(non_clock_info->ulDCLK); + } else { + rps->vclk = 0; + rps->dclk = 0; + } + + if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) { + rdev->pm.dpm.boot_ps = rps; + kv_patch_boot_state(rdev, ps); + } + if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE) + rdev->pm.dpm.uvd_ps = rps; +} + +static void kv_parse_pplib_clock_info(struct radeon_device *rdev, + struct radeon_ps *rps, int index, + union pplib_clock_info *clock_info) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct kv_ps *ps = kv_get_ps(rps); + struct kv_pl *pl = &ps->levels[index]; + u32 sclk; + + sclk = le16_to_cpu(clock_info->sumo.usEngineClockLow); + sclk |= clock_info->sumo.ucEngineClockHigh << 16; + pl->sclk = sclk; + pl->vddc_index = clock_info->sumo.vddcIndex; + + ps->num_levels = index + 1; + + if (pi->caps_sclk_ds) { + pl->ds_divider_index = 5; + pl->ss_divider_index = 5; + } +} + +static int kv_parse_power_table(struct radeon_device *rdev) +{ + struct radeon_mode_info *mode_info = &rdev->mode_info; + struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info; + union pplib_power_state *power_state; + int i, j, k, non_clock_array_index, clock_array_index; + union pplib_clock_info *clock_info; + struct _StateArray *state_array; + struct _ClockInfoArray *clock_info_array; + struct _NonClockInfoArray *non_clock_info_array; + union power_info *power_info; + int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo); + u16 data_offset; + u8 frev, crev; + u8 *power_state_offset; + struct kv_ps *ps; + + if (!atom_parse_data_header(mode_info->atom_context, index, NULL, + &frev, &crev, &data_offset)) + return -EINVAL; + power_info = (union power_info *)(mode_info->atom_context->bios + data_offset); + + state_array = (struct _StateArray *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(power_info->pplib.usStateArrayOffset)); + clock_info_array = (struct _ClockInfoArray *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(power_info->pplib.usClockInfoArrayOffset)); + non_clock_info_array = (struct _NonClockInfoArray *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset)); + + rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) * + state_array->ucNumEntries, GFP_KERNEL); + if (!rdev->pm.dpm.ps) + return -ENOMEM; + power_state_offset = (u8 *)state_array->states; + for (i = 0; i < state_array->ucNumEntries; i++) { + u8 *idx; + power_state = (union pplib_power_state *)power_state_offset; + non_clock_array_index = power_state->v2.nonClockInfoIndex; + non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *) + &non_clock_info_array->nonClockInfo[non_clock_array_index]; + if (!rdev->pm.power_state[i].clock_info) + return -EINVAL; + ps = kzalloc(sizeof(struct kv_ps), GFP_KERNEL); + if (ps == NULL) { + kfree(rdev->pm.dpm.ps); + return -ENOMEM; + } + rdev->pm.dpm.ps[i].ps_priv = ps; + k = 0; + idx = (u8 *)&power_state->v2.clockInfoIndex[0]; + for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) { + clock_array_index = idx[j]; + if (clock_array_index >= clock_info_array->ucNumEntries) + continue; + if (k >= SUMO_MAX_HARDWARE_POWERLEVELS) + break; + clock_info = (union pplib_clock_info *) + ((u8 *)&clock_info_array->clockInfo[0] + + (clock_array_index * clock_info_array->ucEntrySize)); + kv_parse_pplib_clock_info(rdev, + &rdev->pm.dpm.ps[i], k, + clock_info); + k++; + } + kv_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i], + non_clock_info, + non_clock_info_array->ucEntrySize); + power_state_offset += 2 + power_state->v2.ucNumDPMLevels; + } + rdev->pm.dpm.num_ps = state_array->ucNumEntries; + + /* fill in the vce power states */ + for (i = 0; i < RADEON_MAX_VCE_LEVELS; i++) { + u32 sclk; + clock_array_index = rdev->pm.dpm.vce_states[i].clk_idx; + clock_info = (union pplib_clock_info *) + &clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize]; + sclk = le16_to_cpu(clock_info->sumo.usEngineClockLow); + sclk |= clock_info->sumo.ucEngineClockHigh << 16; + rdev->pm.dpm.vce_states[i].sclk = sclk; + rdev->pm.dpm.vce_states[i].mclk = 0; + } + + return 0; +} + +int kv_dpm_init(struct radeon_device *rdev) +{ + struct kv_power_info *pi; + int ret, i; + + pi = kzalloc(sizeof(struct kv_power_info), GFP_KERNEL); + if (pi == NULL) + return -ENOMEM; + rdev->pm.dpm.priv = pi; + + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + + ret = r600_parse_extended_power_table(rdev); + if (ret) + return ret; + + for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++) + pi->at[i] = TRINITY_AT_DFLT; + + pi->sram_end = SMC_RAM_END; + + pi->enable_nb_dpm = true; + + pi->caps_power_containment = true; + pi->caps_cac = true; + pi->enable_didt = false; + if (pi->enable_didt) { + pi->caps_sq_ramping = true; + pi->caps_db_ramping = true; + pi->caps_td_ramping = true; + pi->caps_tcp_ramping = true; + } + + pi->caps_sclk_ds = true; + pi->enable_auto_thermal_throttling = true; + pi->disable_nb_ps3_in_battery = false; + pi->bapm_enable = true; + pi->voltage_drop_t = 0; + pi->caps_sclk_throttle_low_notification = false; + pi->caps_fps = false; /* true? */ + pi->caps_uvd_pg = true; + pi->caps_uvd_dpm = true; + pi->caps_vce_pg = false; /* XXX true */ + pi->caps_samu_pg = false; + pi->caps_acp_pg = false; + pi->caps_stable_p_state = false; + + ret = kv_parse_sys_info_table(rdev); + if (ret) + return ret; + + kv_patch_voltage_values(rdev); + kv_construct_boot_state(rdev); + + ret = kv_parse_power_table(rdev); + if (ret) + return ret; + + pi->enable_dpm = true; + + return 0; +} + +void kv_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, + struct seq_file *m) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + u32 current_index = + (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX) & CURR_SCLK_INDEX_MASK) >> + CURR_SCLK_INDEX_SHIFT; + u32 sclk, tmp; + u16 vddc; + + if (current_index >= SMU__NUM_SCLK_DPM_STATE) { + seq_printf(m, "invalid dpm profile %d\n", current_index); + } else { + sclk = be32_to_cpu(pi->graphics_level[current_index].SclkFrequency); + tmp = (RREG32_SMC(SMU_VOLTAGE_STATUS) & SMU_VOLTAGE_CURRENT_LEVEL_MASK) >> + SMU_VOLTAGE_CURRENT_LEVEL_SHIFT; + vddc = kv_convert_8bit_index_to_voltage(rdev, (u16)tmp); + seq_printf(m, "power level %d sclk: %u vddc: %u\n", + current_index, sclk, vddc); + } +} + +void kv_dpm_print_power_state(struct radeon_device *rdev, + struct radeon_ps *rps) +{ + int i; + struct kv_ps *ps = kv_get_ps(rps); + + r600_dpm_print_class_info(rps->class, rps->class2); + r600_dpm_print_cap_info(rps->caps); + printk("\tuvd vclk: %d dclk: %d\n", rps->vclk, rps->dclk); + for (i = 0; i < ps->num_levels; i++) { + struct kv_pl *pl = &ps->levels[i]; + printk("\t\tpower level %d sclk: %u vddc: %u\n", + i, pl->sclk, + kv_convert_8bit_index_to_voltage(rdev, pl->vddc_index)); + } + r600_dpm_print_ps_status(rdev, rps); +} + +void kv_dpm_fini(struct radeon_device *rdev) +{ + int i; + + for (i = 0; i < rdev->pm.dpm.num_ps; i++) { + kfree(rdev->pm.dpm.ps[i].ps_priv); + } + kfree(rdev->pm.dpm.ps); + kfree(rdev->pm.dpm.priv); + r600_free_extended_power_table(rdev); +} + +void kv_dpm_display_configuration_changed(struct radeon_device *rdev) +{ + +} + +u32 kv_dpm_get_sclk(struct radeon_device *rdev, bool low) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + struct kv_ps *requested_state = kv_get_ps(&pi->requested_rps); + + if (low) + return requested_state->levels[0].sclk; + else + return requested_state->levels[requested_state->num_levels - 1].sclk; +} + +u32 kv_dpm_get_mclk(struct radeon_device *rdev, bool low) +{ + struct kv_power_info *pi = kv_get_pi(rdev); + + return pi->sys_info.bootup_uma_clk; +} + diff --git a/drivers/gpu/drm/radeon/kv_dpm.h b/drivers/gpu/drm/radeon/kv_dpm.h new file mode 100644 index 00000000000..8cef7525d7a --- /dev/null +++ b/drivers/gpu/drm/radeon/kv_dpm.h @@ -0,0 +1,200 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ +#ifndef __KV_DPM_H__ +#define __KV_DPM_H__ + +#define SMU__NUM_SCLK_DPM_STATE 8 +#define SMU__NUM_MCLK_DPM_LEVELS 4 +#define SMU__NUM_LCLK_DPM_LEVELS 8 +#define SMU__NUM_PCIE_DPM_LEVELS 0 /* ??? */ +#include "smu7_fusion.h" +#include "trinity_dpm.h" +#include "ppsmc.h" + +#define KV_NUM_NBPSTATES 4 + +enum kv_pt_config_reg_type { + KV_CONFIGREG_MMR = 0, + KV_CONFIGREG_SMC_IND, + KV_CONFIGREG_DIDT_IND, + KV_CONFIGREG_CACHE, + KV_CONFIGREG_MAX +}; + +struct kv_pt_config_reg { + u32 offset; + u32 mask; + u32 shift; + u32 value; + enum kv_pt_config_reg_type type; +}; + +struct kv_lcac_config_values { + u32 block_id; + u32 signal_id; + u32 t; +}; + +struct kv_lcac_config_reg { + u32 cntl; + u32 block_mask; + u32 block_shift; + u32 signal_mask; + u32 signal_shift; + u32 t_mask; + u32 t_shift; + u32 enable_mask; + u32 enable_shift; +}; + +struct kv_pl { + u32 sclk; + u8 vddc_index; + u8 ds_divider_index; + u8 ss_divider_index; + u8 allow_gnb_slow; + u8 force_nbp_state; + u8 display_wm; + u8 vce_wm; +}; + +struct kv_ps { + struct kv_pl levels[SUMO_MAX_HARDWARE_POWERLEVELS]; + u32 num_levels; + bool need_dfs_bypass; + u8 dpm0_pg_nb_ps_lo; + u8 dpm0_pg_nb_ps_hi; + u8 dpmx_nb_ps_lo; + u8 dpmx_nb_ps_hi; +}; + +struct kv_sys_info { + u32 bootup_uma_clk; + u32 bootup_sclk; + u32 dentist_vco_freq; + u32 nb_dpm_enable; + u32 nbp_memory_clock[KV_NUM_NBPSTATES]; + u32 nbp_n_clock[KV_NUM_NBPSTATES]; + u16 bootup_nb_voltage_index; + u8 htc_tmp_lmt; + u8 htc_hyst_lmt; + struct sumo_sclk_voltage_mapping_table sclk_voltage_mapping_table; + struct sumo_vid_mapping_table vid_mapping_table; + u32 uma_channel_number; +}; + +struct kv_power_info { + u32 at[SUMO_MAX_HARDWARE_POWERLEVELS]; + u32 voltage_drop_t; + struct kv_sys_info sys_info; + struct kv_pl boot_pl; + bool enable_nb_ps_policy; + bool disable_nb_ps3_in_battery; + bool video_start; + bool battery_state; + u32 lowest_valid; + u32 highest_valid; + u16 high_voltage_t; + bool cac_enabled; + bool bapm_enable; + /* smc offsets */ + u32 sram_end; + u32 dpm_table_start; + u32 soft_regs_start; + /* dpm SMU tables */ + u8 graphics_dpm_level_count; + u8 uvd_level_count; + u8 vce_level_count; + u8 acp_level_count; + u8 samu_level_count; + u16 fps_high_t; + SMU7_Fusion_GraphicsLevel graphics_level[SMU__NUM_SCLK_DPM_STATE]; + SMU7_Fusion_ACPILevel acpi_level; + SMU7_Fusion_UvdLevel uvd_level[SMU7_MAX_LEVELS_UVD]; + SMU7_Fusion_ExtClkLevel vce_level[SMU7_MAX_LEVELS_VCE]; + SMU7_Fusion_ExtClkLevel acp_level[SMU7_MAX_LEVELS_ACP]; + SMU7_Fusion_ExtClkLevel samu_level[SMU7_MAX_LEVELS_SAMU]; + u8 uvd_boot_level; + u8 vce_boot_level; + u8 acp_boot_level; + u8 samu_boot_level; + u8 uvd_interval; + u8 vce_interval; + u8 acp_interval; + u8 samu_interval; + u8 graphics_boot_level; + u8 graphics_interval; + u8 graphics_therm_throttle_enable; + u8 graphics_voltage_change_enable; + u8 graphics_clk_slow_enable; + u8 graphics_clk_slow_divider; + u8 fps_low_t; + u32 low_sclk_interrupt_t; + bool uvd_power_gated; + bool vce_power_gated; + bool acp_power_gated; + bool samu_power_gated; + bool nb_dpm_enabled; + /* flags */ + bool enable_didt; + bool enable_dpm; + bool enable_auto_thermal_throttling; + bool enable_nb_dpm; + /* caps */ + bool caps_cac; + bool caps_power_containment; + bool caps_sq_ramping; + bool caps_db_ramping; + bool caps_td_ramping; + bool caps_tcp_ramping; + bool caps_sclk_throttle_low_notification; + bool caps_fps; + bool caps_uvd_dpm; + bool caps_uvd_pg; + bool caps_vce_pg; + bool caps_samu_pg; + bool caps_acp_pg; + bool caps_stable_p_state; + bool caps_enable_dfs_bypass; + bool caps_sclk_ds; + struct radeon_ps current_rps; + struct kv_ps current_ps; + struct radeon_ps requested_rps; + struct kv_ps requested_ps; +}; + + +/* kv_smc.c */ +int kv_notify_message_to_smu(struct radeon_device *rdev, u32 id); +int kv_dpm_get_enable_mask(struct radeon_device *rdev, u32 *enable_mask); +int kv_send_msg_to_smc_with_parameter(struct radeon_device *rdev, + PPSMC_Msg msg, u32 parameter); +int kv_read_smc_sram_dword(struct radeon_device *rdev, u32 smc_address, + u32 *value, u32 limit); +int kv_smc_dpm_enable(struct radeon_device *rdev, bool enable); +int kv_smc_bapm_enable(struct radeon_device *rdev, bool enable); +int kv_copy_bytes_to_smc(struct radeon_device *rdev, + u32 smc_start_address, + const u8 *src, u32 byte_count, u32 limit); + +#endif diff --git a/drivers/gpu/drm/radeon/kv_smc.c b/drivers/gpu/drm/radeon/kv_smc.c new file mode 100644 index 00000000000..0000b59a6d0 --- /dev/null +++ b/drivers/gpu/drm/radeon/kv_smc.c @@ -0,0 +1,215 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ + +#include "drmP.h" +#include "radeon.h" +#include "cikd.h" +#include "kv_dpm.h" + +int kv_notify_message_to_smu(struct radeon_device *rdev, u32 id) +{ + u32 i; + u32 tmp = 0; + + WREG32(SMC_MESSAGE_0, id & SMC_MSG_MASK); + + for (i = 0; i < rdev->usec_timeout; i++) { + if ((RREG32(SMC_RESP_0) & SMC_RESP_MASK) != 0) + break; + udelay(1); + } + tmp = RREG32(SMC_RESP_0) & SMC_RESP_MASK; + + if (tmp != 1) { + if (tmp == 0xFF) + return -EINVAL; + else if (tmp == 0xFE) + return -EINVAL; + } + + return 0; +} + +int kv_dpm_get_enable_mask(struct radeon_device *rdev, u32 *enable_mask) +{ + int ret; + + ret = kv_notify_message_to_smu(rdev, PPSMC_MSG_SCLKDPM_GetEnabledMask); + + if (ret == 0) + *enable_mask = RREG32_SMC(SMC_SYSCON_MSG_ARG_0); + + return ret; +} + +int kv_send_msg_to_smc_with_parameter(struct radeon_device *rdev, + PPSMC_Msg msg, u32 parameter) +{ + + WREG32(SMC_MSG_ARG_0, parameter); + + return kv_notify_message_to_smu(rdev, msg); +} + +static int kv_set_smc_sram_address(struct radeon_device *rdev, + u32 smc_address, u32 limit) +{ + if (smc_address & 3) + return -EINVAL; + if ((smc_address + 3) > limit) + return -EINVAL; + + WREG32(SMC_IND_INDEX_0, smc_address); + WREG32_P(SMC_IND_ACCESS_CNTL, 0, ~AUTO_INCREMENT_IND_0); + + return 0; +} + +int kv_read_smc_sram_dword(struct radeon_device *rdev, u32 smc_address, + u32 *value, u32 limit) +{ + int ret; + + ret = kv_set_smc_sram_address(rdev, smc_address, limit); + if (ret) + return ret; + + *value = RREG32(SMC_IND_DATA_0); + return 0; +} + +int kv_smc_dpm_enable(struct radeon_device *rdev, bool enable) +{ + if (enable) + return kv_notify_message_to_smu(rdev, PPSMC_MSG_DPM_Enable); + else + return kv_notify_message_to_smu(rdev, PPSMC_MSG_DPM_Disable); +} + +int kv_smc_bapm_enable(struct radeon_device *rdev, bool enable) +{ + if (enable) + return kv_notify_message_to_smu(rdev, PPSMC_MSG_EnableBAPM); + else + return kv_notify_message_to_smu(rdev, PPSMC_MSG_DisableBAPM); +} + +int kv_copy_bytes_to_smc(struct radeon_device *rdev, + u32 smc_start_address, + const u8 *src, u32 byte_count, u32 limit) +{ + int ret; + u32 data, original_data, addr, extra_shift, t_byte, count, mask; + + if ((smc_start_address + byte_count) > limit) + return -EINVAL; + + addr = smc_start_address; + t_byte = addr & 3; + + /* RMW for the initial bytes */ + if (t_byte != 0) { + addr -= t_byte; + + ret = kv_set_smc_sram_address(rdev, addr, limit); + if (ret) + return ret; + + original_data = RREG32(SMC_IND_DATA_0); + + data = 0; + mask = 0; + count = 4; + while (count > 0) { + if (t_byte > 0) { + mask = (mask << 8) | 0xff; + t_byte--; + } else if (byte_count > 0) { + data = (data << 8) + *src++; + byte_count--; + mask <<= 8; + } else { + data <<= 8; + mask = (mask << 8) | 0xff; + } + count--; + } + + data |= original_data & mask; + + ret = kv_set_smc_sram_address(rdev, addr, limit); + if (ret) + return ret; + + WREG32(SMC_IND_DATA_0, data); + + addr += 4; + } + + while (byte_count >= 4) { + /* SMC address space is BE */ + data = (src[0] << 24) + (src[1] << 16) + (src[2] << 8) + src[3]; + + ret = kv_set_smc_sram_address(rdev, addr, limit); + if (ret) + return ret; + + WREG32(SMC_IND_DATA_0, data); + + src += 4; + byte_count -= 4; + addr += 4; + } + + /* RMW for the final bytes */ + if (byte_count > 0) { + data = 0; + + ret = kv_set_smc_sram_address(rdev, addr, limit); + if (ret) + return ret; + + original_data= RREG32(SMC_IND_DATA_0); + + extra_shift = 8 * (4 - byte_count); + + while (byte_count > 0) { + /* SMC address space is BE */ + data = (data << 8) + *src++; + byte_count--; + } + + data <<= extra_shift; + + data |= (original_data & ~((~0UL) << extra_shift)); + + ret = kv_set_smc_sram_address(rdev, addr, limit); + if (ret) + return ret; + + WREG32(SMC_IND_DATA_0, data); + } + return 0; +} + diff --git a/drivers/gpu/drm/radeon/mkregtable.c b/drivers/gpu/drm/radeon/mkregtable.c index af85299f212..4a85bb644e2 100644 --- a/drivers/gpu/drm/radeon/mkregtable.c +++ b/drivers/gpu/drm/radeon/mkregtable.c @@ -655,7 +655,7 @@ static int parser_auth(struct table *t, const char *filename) /* first line will contain the last register * and gpu name */ - sscanf(buf, "%s %s", gpu_name, last_reg_s); + sscanf(buf, "%9s %9s", gpu_name, last_reg_s); t->gpu_prefix = gpu_name; last_reg = strtol(last_reg_s, NULL, 16); diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c index f30127cb30e..5a33ca68186 100644 --- a/drivers/gpu/drm/radeon/ni.c +++ b/drivers/gpu/drm/radeon/ni.c @@ -22,7 +22,6 @@ * Authors: Alex Deucher */ #include <linux/firmware.h> -#include <linux/platform_device.h> #include <linux/slab.h> #include <linux/module.h> #include <drm/drmP.h> @@ -36,7 +35,7 @@ #include "radeon_ucode.h" #include "clearstate_cayman.h" -static u32 tn_rlc_save_restore_register_list[] = +static const u32 tn_rlc_save_restore_register_list[] = { 0x98fc, 0x98f0, @@ -161,7 +160,6 @@ static u32 tn_rlc_save_restore_register_list[] = 0x9830, 0x802c, }; -static u32 tn_rlc_save_restore_register_list_size = ARRAY_SIZE(tn_rlc_save_restore_register_list); extern bool evergreen_is_display_hung(struct radeon_device *rdev); extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev); @@ -176,6 +174,7 @@ extern void evergreen_pcie_gen2_enable(struct radeon_device *rdev); extern void evergreen_program_aspm(struct radeon_device *rdev); extern void sumo_rlc_fini(struct radeon_device *rdev); extern int sumo_rlc_init(struct radeon_device *rdev); +extern void evergreen_gpu_pci_config_reset(struct radeon_device *rdev); /* Firmware Names */ MODULE_FIRMWARE("radeon/BARTS_pfp.bin"); @@ -684,7 +683,6 @@ int ni_mc_load_microcode(struct radeon_device *rdev) int ni_init_microcode(struct radeon_device *rdev) { - struct platform_device *pdev; const char *chip_name; const char *rlc_chip_name; size_t pfp_req_size, me_req_size, rlc_req_size, mc_req_size; @@ -694,13 +692,6 @@ int ni_init_microcode(struct radeon_device *rdev) DRM_DEBUG("\n"); - pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0); - err = IS_ERR(pdev); - if (err) { - printk(KERN_ERR "radeon_cp: Failed to register firmware\n"); - return -EINVAL; - } - switch (rdev->family) { case CHIP_BARTS: chip_name = "BARTS"; @@ -753,7 +744,7 @@ int ni_init_microcode(struct radeon_device *rdev) DRM_INFO("Loading %s Microcode\n", chip_name); snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name); - err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->pfp_fw->size != pfp_req_size) { @@ -765,7 +756,7 @@ int ni_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name); - err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->me_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->me_fw->size != me_req_size) { @@ -776,7 +767,7 @@ int ni_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name); - err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->rlc_fw->size != rlc_req_size) { @@ -789,7 +780,7 @@ int ni_init_microcode(struct radeon_device *rdev) /* no MC ucode on TN */ if (!(rdev->flags & RADEON_IS_IGP)) { snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); - err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->mc_fw->size != mc_req_size) { @@ -802,10 +793,15 @@ int ni_init_microcode(struct radeon_device *rdev) if ((rdev->family >= CHIP_BARTS) && (rdev->family <= CHIP_CAYMAN)) { snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name); - err = request_firmware(&rdev->smc_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->smc_fw->size != smc_req_size) { + err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev); + if (err) { + printk(KERN_ERR + "smc: error loading firmware \"%s\"\n", + fw_name); + release_firmware(rdev->smc_fw); + rdev->smc_fw = NULL; + err = 0; + } else if (rdev->smc_fw->size != smc_req_size) { printk(KERN_ERR "ni_mc: Bogus length %zu in firmware \"%s\"\n", rdev->mc_fw->size, fw_name); @@ -814,8 +810,6 @@ int ni_init_microcode(struct radeon_device *rdev) } out: - platform_device_unregister(pdev); - if (err) { if (err != -EINVAL) printk(KERN_ERR @@ -902,6 +896,10 @@ static void cayman_gpu_init(struct radeon_device *rdev) (rdev->pdev->device == 0x999C)) { rdev->config.cayman.max_simds_per_se = 6; rdev->config.cayman.max_backends_per_se = 2; + rdev->config.cayman.max_hw_contexts = 8; + rdev->config.cayman.sx_max_export_size = 256; + rdev->config.cayman.sx_max_export_pos_size = 64; + rdev->config.cayman.sx_max_export_smx_size = 192; } else if ((rdev->pdev->device == 0x9903) || (rdev->pdev->device == 0x9904) || (rdev->pdev->device == 0x990A) || @@ -912,6 +910,10 @@ static void cayman_gpu_init(struct radeon_device *rdev) (rdev->pdev->device == 0x999D)) { rdev->config.cayman.max_simds_per_se = 4; rdev->config.cayman.max_backends_per_se = 2; + rdev->config.cayman.max_hw_contexts = 8; + rdev->config.cayman.sx_max_export_size = 256; + rdev->config.cayman.sx_max_export_pos_size = 64; + rdev->config.cayman.sx_max_export_smx_size = 192; } else if ((rdev->pdev->device == 0x9919) || (rdev->pdev->device == 0x9990) || (rdev->pdev->device == 0x9991) || @@ -922,9 +924,17 @@ static void cayman_gpu_init(struct radeon_device *rdev) (rdev->pdev->device == 0x99A0)) { rdev->config.cayman.max_simds_per_se = 3; rdev->config.cayman.max_backends_per_se = 1; + rdev->config.cayman.max_hw_contexts = 4; + rdev->config.cayman.sx_max_export_size = 128; + rdev->config.cayman.sx_max_export_pos_size = 32; + rdev->config.cayman.sx_max_export_smx_size = 96; } else { rdev->config.cayman.max_simds_per_se = 2; rdev->config.cayman.max_backends_per_se = 1; + rdev->config.cayman.max_hw_contexts = 4; + rdev->config.cayman.sx_max_export_size = 128; + rdev->config.cayman.sx_max_export_pos_size = 32; + rdev->config.cayman.sx_max_export_smx_size = 96; } rdev->config.cayman.max_texture_channel_caches = 2; rdev->config.cayman.max_gprs = 256; @@ -932,10 +942,6 @@ static void cayman_gpu_init(struct radeon_device *rdev) rdev->config.cayman.max_gs_threads = 32; rdev->config.cayman.max_stack_entries = 512; rdev->config.cayman.sx_num_of_sets = 8; - rdev->config.cayman.sx_max_export_size = 256; - rdev->config.cayman.sx_max_export_pos_size = 64; - rdev->config.cayman.sx_max_export_smx_size = 192; - rdev->config.cayman.max_hw_contexts = 8; rdev->config.cayman.sq_num_cf_insts = 2; rdev->config.cayman.sc_prim_fifo_size = 0x40; @@ -1051,6 +1057,18 @@ static void cayman_gpu_init(struct radeon_device *rdev) disabled_rb_mask &= ~(1 << i); } + for (i = 0; i < rdev->config.cayman.max_shader_engines; i++) { + u32 simd_disable_bitmap; + + WREG32(GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_INDEX(i)); + WREG32(RLC_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_INDEX(i)); + simd_disable_bitmap = (RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0xffff0000) >> 16; + simd_disable_bitmap |= 0xffffffff << rdev->config.cayman.max_simds_per_se; + tmp <<= 16; + tmp |= simd_disable_bitmap; + } + rdev->config.cayman.active_simds = hweight32(~tmp); + WREG32(GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_BROADCAST_WRITES); WREG32(RLC_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_BROADCAST_WRITES); @@ -1222,12 +1240,14 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev) SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU); /* Setup L2 cache */ WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | + ENABLE_L2_FRAGMENT_PROCESSING | ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE | ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE | EFFECTIVE_L2_QUEUE_SIZE(7) | CONTEXT1_IDENTITY_ACCESS_MODE(1)); WREG32(VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS | INVALIDATE_L2_CACHE); WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY | + BANK_SELECT(6) | L2_CACHE_BIGK_FRAGMENT_SIZE(6)); /* setup context0 */ WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12); @@ -1260,6 +1280,7 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev) (u32)(rdev->dummy_page.addr >> 12)); WREG32(VM_CONTEXT1_CNTL2, 4); WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(1) | + PAGE_TABLE_BLOCK_SIZE(radeon_vm_block_size - 9) | RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT | RANGE_PROTECTION_FAULT_ENABLE_DEFAULT | DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT | @@ -1325,20 +1346,19 @@ void cayman_fence_ring_emit(struct radeon_device *rdev, { struct radeon_ring *ring = &rdev->ring[fence->ring]; u64 addr = rdev->fence_drv[fence->ring].gpu_addr; + u32 cp_coher_cntl = PACKET3_FULL_CACHE_ENA | PACKET3_TC_ACTION_ENA | + PACKET3_SH_ACTION_ENA; /* flush read cache over gart for this vmid */ - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(ring, 0); radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(ring, PACKET3_TC_ACTION_ENA | PACKET3_SH_ACTION_ENA); + radeon_ring_write(ring, PACKET3_ENGINE_ME | cp_coher_cntl); radeon_ring_write(ring, 0xFFFFFFFF); radeon_ring_write(ring, 0); radeon_ring_write(ring, 10); /* poll interval */ /* EVENT_WRITE_EOP - flush caches, send int */ radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4)); radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT_TS) | EVENT_INDEX(5)); - radeon_ring_write(ring, addr & 0xffffffff); + radeon_ring_write(ring, lower_32_bits(addr)); radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2)); radeon_ring_write(ring, fence->seq); radeon_ring_write(ring, 0); @@ -1347,6 +1367,8 @@ void cayman_fence_ring_emit(struct radeon_device *rdev, void cayman_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) { struct radeon_ring *ring = &rdev->ring[ib->ring]; + u32 cp_coher_cntl = PACKET3_FULL_CACHE_ENA | PACKET3_TC_ACTION_ENA | + PACKET3_SH_ACTION_ENA; /* set to DX10/11 mode */ radeon_ring_write(ring, PACKET3(PACKET3_MODE_CONTROL, 0)); @@ -1371,31 +1393,11 @@ void cayman_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) (ib->vm ? (ib->vm->id << 24) : 0)); /* flush read cache over gart for this vmid */ - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(ring, ib->vm ? ib->vm->id : 0); radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(ring, PACKET3_TC_ACTION_ENA | PACKET3_SH_ACTION_ENA); + radeon_ring_write(ring, PACKET3_ENGINE_ME | cp_coher_cntl); radeon_ring_write(ring, 0xFFFFFFFF); radeon_ring_write(ring, 0); - radeon_ring_write(ring, 10); /* poll interval */ -} - -void cayman_uvd_semaphore_emit(struct radeon_device *rdev, - struct radeon_ring *ring, - struct radeon_semaphore *semaphore, - bool emit_wait) -{ - uint64_t addr = semaphore->gpu_addr; - - radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_LOW, 0)); - radeon_ring_write(ring, (addr >> 3) & 0x000FFFFF); - - radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_HIGH, 0)); - radeon_ring_write(ring, (addr >> 23) & 0x000FFFFF); - - radeon_ring_write(ring, PACKET0(UVD_SEMA_CMD, 0)); - radeon_ring_write(ring, 0x80 | (emit_wait ? 1 : 0)); + radeon_ring_write(ring, ((ib->vm ? ib->vm->id : 0) << 24) | 10); /* poll interval */ } static void cayman_cp_enable(struct radeon_device *rdev, bool enable) @@ -1403,13 +1405,63 @@ static void cayman_cp_enable(struct radeon_device *rdev, bool enable) if (enable) WREG32(CP_ME_CNTL, 0); else { - radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT)); WREG32(SCRATCH_UMSK, 0); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; } } +u32 cayman_gfx_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 rptr; + + if (rdev->wb.enabled) + rptr = rdev->wb.wb[ring->rptr_offs/4]; + else { + if (ring->idx == RADEON_RING_TYPE_GFX_INDEX) + rptr = RREG32(CP_RB0_RPTR); + else if (ring->idx == CAYMAN_RING_TYPE_CP1_INDEX) + rptr = RREG32(CP_RB1_RPTR); + else + rptr = RREG32(CP_RB2_RPTR); + } + + return rptr; +} + +u32 cayman_gfx_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 wptr; + + if (ring->idx == RADEON_RING_TYPE_GFX_INDEX) + wptr = RREG32(CP_RB0_WPTR); + else if (ring->idx == CAYMAN_RING_TYPE_CP1_INDEX) + wptr = RREG32(CP_RB1_WPTR); + else + wptr = RREG32(CP_RB2_WPTR); + + return wptr; +} + +void cayman_gfx_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + if (ring->idx == RADEON_RING_TYPE_GFX_INDEX) { + WREG32(CP_RB0_WPTR, ring->wptr); + (void)RREG32(CP_RB0_WPTR); + } else if (ring->idx == CAYMAN_RING_TYPE_CP1_INDEX) { + WREG32(CP_RB1_WPTR, ring->wptr); + (void)RREG32(CP_RB1_WPTR); + } else { + WREG32(CP_RB2_WPTR, ring->wptr); + (void)RREG32(CP_RB2_WPTR); + } +} + static int cayman_cp_load_microcode(struct radeon_device *rdev) { const __be32 *fw_data; @@ -1538,6 +1590,16 @@ static int cayman_cp_resume(struct radeon_device *rdev) CP_RB1_BASE, CP_RB2_BASE }; + static const unsigned cp_rb_rptr[] = { + CP_RB0_RPTR, + CP_RB1_RPTR, + CP_RB2_RPTR + }; + static const unsigned cp_rb_wptr[] = { + CP_RB0_WPTR, + CP_RB1_WPTR, + CP_RB2_WPTR + }; struct radeon_ring *ring; int i, r; @@ -1571,8 +1633,8 @@ static int cayman_cp_resume(struct radeon_device *rdev) /* Set ring buffer size */ ring = &rdev->ring[ridx[i]]; - rb_cntl = drm_order(ring->ring_size / 8); - rb_cntl |= drm_order(RADEON_GPU_PAGE_SIZE/8) << 8; + rb_cntl = order_base_2(ring->ring_size / 8); + rb_cntl |= order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8; #ifdef __BIG_ENDIAN rb_cntl |= BUF_SWAP_32BIT; #endif @@ -1595,9 +1657,9 @@ static int cayman_cp_resume(struct radeon_device *rdev) ring = &rdev->ring[ridx[i]]; WREG32_P(cp_rb_cntl[i], RB_RPTR_WR_ENA, ~RB_RPTR_WR_ENA); - ring->rptr = ring->wptr = 0; - WREG32(ring->rptr_reg, ring->rptr); - WREG32(ring->wptr_reg, ring->wptr); + ring->wptr = 0; + WREG32(cp_rb_rptr[i], 0); + WREG32(cp_rb_wptr[i], ring->wptr); mdelay(1); WREG32_P(cp_rb_cntl[i], 0, ~RB_RPTR_WR_ENA); @@ -1617,189 +1679,13 @@ static int cayman_cp_resume(struct radeon_device *rdev) return r; } - return 0; -} - -/* - * DMA - * Starting with R600, the GPU has an asynchronous - * DMA engine. The programming model is very similar - * to the 3D engine (ring buffer, IBs, etc.), but the - * DMA controller has it's own packet format that is - * different form the PM4 format used by the 3D engine. - * It supports copying data, writing embedded data, - * solid fills, and a number of other things. It also - * has support for tiling/detiling of buffers. - * Cayman and newer support two asynchronous DMA engines. - */ -/** - * cayman_dma_ring_ib_execute - Schedule an IB on the DMA engine - * - * @rdev: radeon_device pointer - * @ib: IB object to schedule - * - * Schedule an IB in the DMA ring (cayman-SI). - */ -void cayman_dma_ring_ib_execute(struct radeon_device *rdev, - struct radeon_ib *ib) -{ - struct radeon_ring *ring = &rdev->ring[ib->ring]; - - if (rdev->wb.enabled) { - u32 next_rptr = ring->wptr + 4; - while ((next_rptr & 7) != 5) - next_rptr++; - next_rptr += 3; - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1)); - radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xff); - radeon_ring_write(ring, next_rptr); - } - - /* The indirect buffer packet must end on an 8 DW boundary in the DMA ring. - * Pad as necessary with NOPs. - */ - while ((ring->wptr & 7) != 5) - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); - radeon_ring_write(ring, DMA_IB_PACKET(DMA_PACKET_INDIRECT_BUFFER, ib->vm ? ib->vm->id : 0, 0)); - radeon_ring_write(ring, (ib->gpu_addr & 0xFFFFFFE0)); - radeon_ring_write(ring, (ib->length_dw << 12) | (upper_32_bits(ib->gpu_addr) & 0xFF)); - -} - -/** - * cayman_dma_stop - stop the async dma engines - * - * @rdev: radeon_device pointer - * - * Stop the async dma engines (cayman-SI). - */ -void cayman_dma_stop(struct radeon_device *rdev) -{ - u32 rb_cntl; - - radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); - - /* dma0 */ - rb_cntl = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET); - rb_cntl &= ~DMA_RB_ENABLE; - WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, rb_cntl); - - /* dma1 */ - rb_cntl = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET); - rb_cntl &= ~DMA_RB_ENABLE; - WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, rb_cntl); - - rdev->ring[R600_RING_TYPE_DMA_INDEX].ready = false; - rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX].ready = false; -} - -/** - * cayman_dma_resume - setup and start the async dma engines - * - * @rdev: radeon_device pointer - * - * Set up the DMA ring buffers and enable them. (cayman-SI). - * Returns 0 for success, error for failure. - */ -int cayman_dma_resume(struct radeon_device *rdev) -{ - struct radeon_ring *ring; - u32 rb_cntl, dma_cntl, ib_cntl; - u32 rb_bufsz; - u32 reg_offset, wb_offset; - int i, r; - - /* Reset dma */ - WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA | SOFT_RESET_DMA1); - RREG32(SRBM_SOFT_RESET); - udelay(50); - WREG32(SRBM_SOFT_RESET, 0); - - for (i = 0; i < 2; i++) { - if (i == 0) { - ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; - reg_offset = DMA0_REGISTER_OFFSET; - wb_offset = R600_WB_DMA_RPTR_OFFSET; - } else { - ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]; - reg_offset = DMA1_REGISTER_OFFSET; - wb_offset = CAYMAN_WB_DMA1_RPTR_OFFSET; - } - - WREG32(DMA_SEM_INCOMPLETE_TIMER_CNTL + reg_offset, 0); - WREG32(DMA_SEM_WAIT_FAIL_TIMER_CNTL + reg_offset, 0); - - /* Set ring buffer size in dwords */ - rb_bufsz = drm_order(ring->ring_size / 4); - rb_cntl = rb_bufsz << 1; -#ifdef __BIG_ENDIAN - rb_cntl |= DMA_RB_SWAP_ENABLE | DMA_RPTR_WRITEBACK_SWAP_ENABLE; -#endif - WREG32(DMA_RB_CNTL + reg_offset, rb_cntl); - - /* Initialize the ring buffer's read and write pointers */ - WREG32(DMA_RB_RPTR + reg_offset, 0); - WREG32(DMA_RB_WPTR + reg_offset, 0); - - /* set the wb address whether it's enabled or not */ - WREG32(DMA_RB_RPTR_ADDR_HI + reg_offset, - upper_32_bits(rdev->wb.gpu_addr + wb_offset) & 0xFF); - WREG32(DMA_RB_RPTR_ADDR_LO + reg_offset, - ((rdev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC)); - - if (rdev->wb.enabled) - rb_cntl |= DMA_RPTR_WRITEBACK_ENABLE; - - WREG32(DMA_RB_BASE + reg_offset, ring->gpu_addr >> 8); - - /* enable DMA IBs */ - ib_cntl = DMA_IB_ENABLE | CMD_VMID_FORCE; -#ifdef __BIG_ENDIAN - ib_cntl |= DMA_IB_SWAP_ENABLE; -#endif - WREG32(DMA_IB_CNTL + reg_offset, ib_cntl); - - dma_cntl = RREG32(DMA_CNTL + reg_offset); - dma_cntl &= ~CTXEMPTY_INT_ENABLE; - WREG32(DMA_CNTL + reg_offset, dma_cntl); - - ring->wptr = 0; - WREG32(DMA_RB_WPTR + reg_offset, ring->wptr << 2); - - ring->rptr = RREG32(DMA_RB_RPTR + reg_offset) >> 2; - - WREG32(DMA_RB_CNTL + reg_offset, rb_cntl | DMA_RB_ENABLE); - - ring->ready = true; - - r = radeon_ring_test(rdev, ring->idx, ring); - if (r) { - ring->ready = false; - return r; - } - } - - radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); return 0; } -/** - * cayman_dma_fini - tear down the async dma engines - * - * @rdev: radeon_device pointer - * - * Stop the async dma engines and free the rings (cayman-SI). - */ -void cayman_dma_fini(struct radeon_device *rdev) -{ - cayman_dma_stop(rdev); - radeon_ring_fini(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX]); - radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]); -} - -static u32 cayman_gpu_check_soft_reset(struct radeon_device *rdev) +u32 cayman_gpu_check_soft_reset(struct radeon_device *rdev) { u32 reset_mask = 0; u32 tmp; @@ -2022,8 +1908,10 @@ int cayman_asic_reset(struct radeon_device *rdev) reset_mask = cayman_gpu_check_soft_reset(rdev); - if (!reset_mask) - r600_set_bios_scratch_engine_hung(rdev, false); + if (reset_mask) + evergreen_gpu_pci_config_reset(rdev); + + r600_set_bios_scratch_engine_hung(rdev, false); return 0; } @@ -2044,39 +1932,9 @@ bool cayman_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); - return false; - } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); - return radeon_ring_test_lockup(rdev, ring); -} - -/** - * cayman_dma_is_lockup - Check if the DMA engine is locked up - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Check if the async DMA engine is locked up. - * Returns true if the engine appears to be locked up, false if not. - */ -bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) -{ - u32 reset_mask = cayman_gpu_check_soft_reset(rdev); - u32 mask; - - if (ring->idx == R600_RING_TYPE_DMA_INDEX) - mask = RADEON_RESET_DMA; - else - mask = RADEON_RESET_DMA1; - - if (!(reset_mask & mask)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } @@ -2090,23 +1948,14 @@ static int cayman_startup(struct radeon_device *rdev) /* enable aspm */ evergreen_program_aspm(rdev); - if (rdev->flags & RADEON_IS_IGP) { - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) { - r = ni_init_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load firmware!\n"); - return r; - } - } - } else { - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) { - r = ni_init_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load firmware!\n"); - return r; - } - } + /* scratch needs to be initialized before MC */ + r = r600_vram_scratch_init(rdev); + if (r) + return r; + + evergreen_mc_program(rdev); + if (!(rdev->flags & RADEON_IS_IGP) && !rdev->pm.dpm_enabled) { r = ni_mc_load_microcode(rdev); if (r) { DRM_ERROR("Failed to load MC firmware!\n"); @@ -2114,27 +1963,16 @@ static int cayman_startup(struct radeon_device *rdev) } } - r = r600_vram_scratch_init(rdev); - if (r) - return r; - - evergreen_mc_program(rdev); r = cayman_pcie_gart_enable(rdev); if (r) return r; cayman_gpu_init(rdev); - r = evergreen_blit_init(rdev); - if (r) { - r600_blit_fini(rdev); - rdev->asic->copy.copy = NULL; - dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); - } - /* allocate rlc buffers */ if (rdev->flags & RADEON_IS_IGP) { rdev->rlc.reg_list = tn_rlc_save_restore_register_list; - rdev->rlc.reg_list_size = tn_rlc_save_restore_register_list_size; + rdev->rlc.reg_list_size = + (u32)ARRAY_SIZE(tn_rlc_save_restore_register_list); rdev->rlc.cs_data = cayman_cs_data; r = sumo_rlc_init(rdev); if (r) { @@ -2154,7 +1992,7 @@ static int cayman_startup(struct radeon_device *rdev) return r; } - r = rv770_uvd_resume(rdev); + r = uvd_v2_2_resume(rdev); if (!r) { r = radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_UVD_INDEX); @@ -2204,24 +2042,19 @@ static int cayman_startup(struct radeon_device *rdev) evergreen_irq_set(rdev); r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, - CP_RB0_RPTR, CP_RB0_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + RADEON_CP_PACKET2); if (r) return r; ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET, - DMA_RB_RPTR + DMA0_REGISTER_OFFSET, - DMA_RB_WPTR + DMA0_REGISTER_OFFSET, - 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); + DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); if (r) return r; ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, CAYMAN_WB_DMA1_RPTR_OFFSET, - DMA_RB_RPTR + DMA1_REGISTER_OFFSET, - DMA_RB_WPTR + DMA1_REGISTER_OFFSET, - 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); + DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); if (r) return r; @@ -2238,12 +2071,10 @@ static int cayman_startup(struct radeon_device *rdev) ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; if (ring->ring_size) { - r = radeon_ring_init(rdev, ring, ring->ring_size, - R600_WB_UVD_RPTR_OFFSET, - UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, + RADEON_CP_PACKET2); if (!r) - r = r600_uvd_init(rdev); + r = uvd_v1_0_init(rdev); if (r) DRM_ERROR("radeon: failed initializing UVD (%d).\n", r); } @@ -2260,9 +2091,15 @@ static int cayman_startup(struct radeon_device *rdev) return r; } - r = r600_audio_init(rdev); - if (r) - return r; + if (ASIC_IS_DCE6(rdev)) { + r = dce6_audio_init(rdev); + if (r) + return r; + } else { + r = r600_audio_init(rdev); + if (r) + return r; + } return 0; } @@ -2281,6 +2118,9 @@ int cayman_resume(struct radeon_device *rdev) /* init golden registers */ ni_init_golden_registers(rdev); + if (rdev->pm.pm_method == PM_METHOD_DPM) + radeon_pm_resume(rdev); + rdev->accel_working = true; r = cayman_startup(rdev); if (r) { @@ -2293,11 +2133,15 @@ int cayman_resume(struct radeon_device *rdev) int cayman_suspend(struct radeon_device *rdev) { - r600_audio_fini(rdev); + radeon_pm_suspend(rdev); + if (ASIC_IS_DCE6(rdev)) + dce6_audio_fini(rdev); + else + r600_audio_fini(rdev); radeon_vm_manager_fini(rdev); cayman_cp_enable(rdev, false); cayman_dma_stop(rdev); - r600_uvd_rbc_stop(rdev); + uvd_v1_0_fini(rdev); radeon_uvd_suspend(rdev); evergreen_irq_suspend(rdev); radeon_wb_disable(rdev); @@ -2360,6 +2204,27 @@ int cayman_init(struct radeon_device *rdev) if (r) return r; + if (rdev->flags & RADEON_IS_IGP) { + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) { + r = ni_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + } else { + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) { + r = ni_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + } + + /* Initialize power management */ + radeon_pm_init(rdev); + ring->ring_obj = NULL; r600_ring_init(rdev, ring, 1024 * 1024); @@ -2419,7 +2284,7 @@ int cayman_init(struct radeon_device *rdev) void cayman_fini(struct radeon_device *rdev) { - r600_blit_fini(rdev); + radeon_pm_fini(rdev); cayman_cp_fini(rdev); cayman_dma_fini(rdev); r600_irq_fini(rdev); @@ -2429,6 +2294,7 @@ void cayman_fini(struct radeon_device *rdev) radeon_vm_manager_fini(rdev); radeon_ib_pool_fini(rdev); radeon_irq_kms_fini(rdev); + uvd_v1_0_fini(rdev); radeon_uvd_fini(rdev); cayman_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); @@ -2461,129 +2327,165 @@ void cayman_vm_fini(struct radeon_device *rdev) { } -#define R600_ENTRY_VALID (1 << 0) -#define R600_PTE_SYSTEM (1 << 1) -#define R600_PTE_SNOOPED (1 << 2) -#define R600_PTE_READABLE (1 << 5) -#define R600_PTE_WRITEABLE (1 << 6) - -uint32_t cayman_vm_page_flags(struct radeon_device *rdev, uint32_t flags) -{ - uint32_t r600_flags = 0; - r600_flags |= (flags & RADEON_VM_PAGE_VALID) ? R600_ENTRY_VALID : 0; - r600_flags |= (flags & RADEON_VM_PAGE_READABLE) ? R600_PTE_READABLE : 0; - r600_flags |= (flags & RADEON_VM_PAGE_WRITEABLE) ? R600_PTE_WRITEABLE : 0; - if (flags & RADEON_VM_PAGE_SYSTEM) { - r600_flags |= R600_PTE_SYSTEM; - r600_flags |= (flags & RADEON_VM_PAGE_SNOOPED) ? R600_PTE_SNOOPED : 0; - } - return r600_flags; -} - /** - * cayman_vm_set_page - update the page tables using the CP + * cayman_vm_decode_fault - print human readable fault info * * @rdev: radeon_device pointer - * @ib: indirect buffer to fill with commands - * @pe: addr of the page entry - * @addr: dst addr to write into pe - * @count: number of page entries to update - * @incr: increase next addr by incr bytes - * @flags: access flags + * @status: VM_CONTEXT1_PROTECTION_FAULT_STATUS register value + * @addr: VM_CONTEXT1_PROTECTION_FAULT_ADDR register value * - * Update the page tables using the CP (cayman/TN). + * Print human readable fault information (cayman/TN). */ -void cayman_vm_set_page(struct radeon_device *rdev, - struct radeon_ib *ib, - uint64_t pe, - uint64_t addr, unsigned count, - uint32_t incr, uint32_t flags) +void cayman_vm_decode_fault(struct radeon_device *rdev, + u32 status, u32 addr) { - uint32_t r600_flags = cayman_vm_page_flags(rdev, flags); - uint64_t value; - unsigned ndw; - - if (rdev->asic->vm.pt_ring_index == RADEON_RING_TYPE_GFX_INDEX) { - while (count) { - ndw = 1 + count * 2; - if (ndw > 0x3FFF) - ndw = 0x3FFF; - - ib->ptr[ib->length_dw++] = PACKET3(PACKET3_ME_WRITE, ndw); - ib->ptr[ib->length_dw++] = pe; - ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; - for (; ndw > 1; ndw -= 2, --count, pe += 8) { - if (flags & RADEON_VM_PAGE_SYSTEM) { - value = radeon_vm_map_gart(rdev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & RADEON_VM_PAGE_VALID) { - value = addr; - } else { - value = 0; - } - addr += incr; - value |= r600_flags; - ib->ptr[ib->length_dw++] = value; - ib->ptr[ib->length_dw++] = upper_32_bits(value); - } - } - } else { - if ((flags & RADEON_VM_PAGE_SYSTEM) || - (count == 1)) { - while (count) { - ndw = count * 2; - if (ndw > 0xFFFFE) - ndw = 0xFFFFE; - - /* for non-physically contiguous pages (system) */ - ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, ndw); - ib->ptr[ib->length_dw++] = pe; - ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; - for (; ndw > 0; ndw -= 2, --count, pe += 8) { - if (flags & RADEON_VM_PAGE_SYSTEM) { - value = radeon_vm_map_gart(rdev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & RADEON_VM_PAGE_VALID) { - value = addr; - } else { - value = 0; - } - addr += incr; - value |= r600_flags; - ib->ptr[ib->length_dw++] = value; - ib->ptr[ib->length_dw++] = upper_32_bits(value); - } - } - while (ib->length_dw & 0x7) - ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0); - } else { - while (count) { - ndw = count * 2; - if (ndw > 0xFFFFE) - ndw = 0xFFFFE; - - if (flags & RADEON_VM_PAGE_VALID) - value = addr; - else - value = 0; - /* for physically contiguous pages (vram) */ - ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw); - ib->ptr[ib->length_dw++] = pe; /* dst addr */ - ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; - ib->ptr[ib->length_dw++] = r600_flags; /* mask */ - ib->ptr[ib->length_dw++] = 0; - ib->ptr[ib->length_dw++] = value; /* value */ - ib->ptr[ib->length_dw++] = upper_32_bits(value); - ib->ptr[ib->length_dw++] = incr; /* increment size */ - ib->ptr[ib->length_dw++] = 0; - pe += ndw * 4; - addr += (ndw / 2) * incr; - count -= ndw / 2; - } - } - while (ib->length_dw & 0x7) - ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0); + u32 mc_id = (status & MEMORY_CLIENT_ID_MASK) >> MEMORY_CLIENT_ID_SHIFT; + u32 vmid = (status & FAULT_VMID_MASK) >> FAULT_VMID_SHIFT; + u32 protections = (status & PROTECTIONS_MASK) >> PROTECTIONS_SHIFT; + char *block; + + switch (mc_id) { + case 32: + case 16: + case 96: + case 80: + case 160: + case 144: + case 224: + case 208: + block = "CB"; + break; + case 33: + case 17: + case 97: + case 81: + case 161: + case 145: + case 225: + case 209: + block = "CB_FMASK"; + break; + case 34: + case 18: + case 98: + case 82: + case 162: + case 146: + case 226: + case 210: + block = "CB_CMASK"; + break; + case 35: + case 19: + case 99: + case 83: + case 163: + case 147: + case 227: + case 211: + block = "CB_IMMED"; + break; + case 36: + case 20: + case 100: + case 84: + case 164: + case 148: + case 228: + case 212: + block = "DB"; + break; + case 37: + case 21: + case 101: + case 85: + case 165: + case 149: + case 229: + case 213: + block = "DB_HTILE"; + break; + case 38: + case 22: + case 102: + case 86: + case 166: + case 150: + case 230: + case 214: + block = "SX"; + break; + case 39: + case 23: + case 103: + case 87: + case 167: + case 151: + case 231: + case 215: + block = "DB_STEN"; + break; + case 40: + case 24: + case 104: + case 88: + case 232: + case 216: + case 168: + case 152: + block = "TC_TFETCH"; + break; + case 41: + case 25: + case 105: + case 89: + case 233: + case 217: + case 169: + case 153: + block = "TC_VFETCH"; + break; + case 42: + case 26: + case 106: + case 90: + case 234: + case 218: + case 170: + case 154: + block = "VC"; + break; + case 112: + block = "CP"; + break; + case 113: + case 114: + block = "SH"; + break; + case 115: + block = "VGT"; + break; + case 178: + block = "IH"; + break; + case 51: + block = "RLC"; + break; + case 55: + block = "DMA"; + break; + case 56: + block = "HDP"; + break; + default: + block = "unknown"; + break; } + + printk("VM fault (0x%02x, vmid %d) at page %u, %s from %s (%d)\n", + protections, vmid, addr, + (status & MEMORY_CLIENT_RW_MASK) ? "write" : "read", + block, mc_id); } /** @@ -2616,26 +2518,3 @@ void cayman_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0)); radeon_ring_write(ring, 0x0); } - -void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) -{ - struct radeon_ring *ring = &rdev->ring[ridx]; - - if (vm == NULL) - return; - - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0)); - radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2)); - radeon_ring_write(ring, vm->pd_gpu_addr >> 12); - - /* flush hdp cache */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0)); - radeon_ring_write(ring, (0xf << 16) | (HDP_MEM_COHERENCY_FLUSH_CNTL >> 2)); - radeon_ring_write(ring, 1); - - /* bits 0-7 are the VM contexts0-7 */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0)); - radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2)); - radeon_ring_write(ring, 1 << vm->id); -} - diff --git a/drivers/gpu/drm/radeon/ni_dma.c b/drivers/gpu/drm/radeon/ni_dma.c new file mode 100644 index 00000000000..6378e027669 --- /dev/null +++ b/drivers/gpu/drm/radeon/ni_dma.c @@ -0,0 +1,408 @@ +/* + * Copyright 2010 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "radeon_trace.h" +#include "nid.h" + +u32 cayman_gpu_check_soft_reset(struct radeon_device *rdev); + +/* + * DMA + * Starting with R600, the GPU has an asynchronous + * DMA engine. The programming model is very similar + * to the 3D engine (ring buffer, IBs, etc.), but the + * DMA controller has it's own packet format that is + * different form the PM4 format used by the 3D engine. + * It supports copying data, writing embedded data, + * solid fills, and a number of other things. It also + * has support for tiling/detiling of buffers. + * Cayman and newer support two asynchronous DMA engines. + */ + +/** + * cayman_dma_get_rptr - get the current read pointer + * + * @rdev: radeon_device pointer + * @ring: radeon ring pointer + * + * Get the current rptr from the hardware (cayman+). + */ +uint32_t cayman_dma_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 rptr, reg; + + if (rdev->wb.enabled) { + rptr = rdev->wb.wb[ring->rptr_offs/4]; + } else { + if (ring->idx == R600_RING_TYPE_DMA_INDEX) + reg = DMA_RB_RPTR + DMA0_REGISTER_OFFSET; + else + reg = DMA_RB_RPTR + DMA1_REGISTER_OFFSET; + + rptr = RREG32(reg); + } + + return (rptr & 0x3fffc) >> 2; +} + +/** + * cayman_dma_get_wptr - get the current write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon ring pointer + * + * Get the current wptr from the hardware (cayman+). + */ +uint32_t cayman_dma_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 reg; + + if (ring->idx == R600_RING_TYPE_DMA_INDEX) + reg = DMA_RB_WPTR + DMA0_REGISTER_OFFSET; + else + reg = DMA_RB_WPTR + DMA1_REGISTER_OFFSET; + + return (RREG32(reg) & 0x3fffc) >> 2; +} + +/** + * cayman_dma_set_wptr - commit the write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon ring pointer + * + * Write the wptr back to the hardware (cayman+). + */ +void cayman_dma_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 reg; + + if (ring->idx == R600_RING_TYPE_DMA_INDEX) + reg = DMA_RB_WPTR + DMA0_REGISTER_OFFSET; + else + reg = DMA_RB_WPTR + DMA1_REGISTER_OFFSET; + + WREG32(reg, (ring->wptr << 2) & 0x3fffc); +} + +/** + * cayman_dma_ring_ib_execute - Schedule an IB on the DMA engine + * + * @rdev: radeon_device pointer + * @ib: IB object to schedule + * + * Schedule an IB in the DMA ring (cayman-SI). + */ +void cayman_dma_ring_ib_execute(struct radeon_device *rdev, + struct radeon_ib *ib) +{ + struct radeon_ring *ring = &rdev->ring[ib->ring]; + + if (rdev->wb.enabled) { + u32 next_rptr = ring->wptr + 4; + while ((next_rptr & 7) != 5) + next_rptr++; + next_rptr += 3; + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1)); + radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); + radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xff); + radeon_ring_write(ring, next_rptr); + } + + /* The indirect buffer packet must end on an 8 DW boundary in the DMA ring. + * Pad as necessary with NOPs. + */ + while ((ring->wptr & 7) != 5) + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); + radeon_ring_write(ring, DMA_IB_PACKET(DMA_PACKET_INDIRECT_BUFFER, ib->vm ? ib->vm->id : 0, 0)); + radeon_ring_write(ring, (ib->gpu_addr & 0xFFFFFFE0)); + radeon_ring_write(ring, (ib->length_dw << 12) | (upper_32_bits(ib->gpu_addr) & 0xFF)); + +} + +/** + * cayman_dma_stop - stop the async dma engines + * + * @rdev: radeon_device pointer + * + * Stop the async dma engines (cayman-SI). + */ +void cayman_dma_stop(struct radeon_device *rdev) +{ + u32 rb_cntl; + + if ((rdev->asic->copy.copy_ring_index == R600_RING_TYPE_DMA_INDEX) || + (rdev->asic->copy.copy_ring_index == CAYMAN_RING_TYPE_DMA1_INDEX)) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + + /* dma0 */ + rb_cntl = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET); + rb_cntl &= ~DMA_RB_ENABLE; + WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, rb_cntl); + + /* dma1 */ + rb_cntl = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET); + rb_cntl &= ~DMA_RB_ENABLE; + WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, rb_cntl); + + rdev->ring[R600_RING_TYPE_DMA_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX].ready = false; +} + +/** + * cayman_dma_resume - setup and start the async dma engines + * + * @rdev: radeon_device pointer + * + * Set up the DMA ring buffers and enable them. (cayman-SI). + * Returns 0 for success, error for failure. + */ +int cayman_dma_resume(struct radeon_device *rdev) +{ + struct radeon_ring *ring; + u32 rb_cntl, dma_cntl, ib_cntl; + u32 rb_bufsz; + u32 reg_offset, wb_offset; + int i, r; + + /* Reset dma */ + WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA | SOFT_RESET_DMA1); + RREG32(SRBM_SOFT_RESET); + udelay(50); + WREG32(SRBM_SOFT_RESET, 0); + + for (i = 0; i < 2; i++) { + if (i == 0) { + ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; + reg_offset = DMA0_REGISTER_OFFSET; + wb_offset = R600_WB_DMA_RPTR_OFFSET; + } else { + ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]; + reg_offset = DMA1_REGISTER_OFFSET; + wb_offset = CAYMAN_WB_DMA1_RPTR_OFFSET; + } + + WREG32(DMA_SEM_INCOMPLETE_TIMER_CNTL + reg_offset, 0); + WREG32(DMA_SEM_WAIT_FAIL_TIMER_CNTL + reg_offset, 0); + + /* Set ring buffer size in dwords */ + rb_bufsz = order_base_2(ring->ring_size / 4); + rb_cntl = rb_bufsz << 1; +#ifdef __BIG_ENDIAN + rb_cntl |= DMA_RB_SWAP_ENABLE | DMA_RPTR_WRITEBACK_SWAP_ENABLE; +#endif + WREG32(DMA_RB_CNTL + reg_offset, rb_cntl); + + /* Initialize the ring buffer's read and write pointers */ + WREG32(DMA_RB_RPTR + reg_offset, 0); + WREG32(DMA_RB_WPTR + reg_offset, 0); + + /* set the wb address whether it's enabled or not */ + WREG32(DMA_RB_RPTR_ADDR_HI + reg_offset, + upper_32_bits(rdev->wb.gpu_addr + wb_offset) & 0xFF); + WREG32(DMA_RB_RPTR_ADDR_LO + reg_offset, + ((rdev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC)); + + if (rdev->wb.enabled) + rb_cntl |= DMA_RPTR_WRITEBACK_ENABLE; + + WREG32(DMA_RB_BASE + reg_offset, ring->gpu_addr >> 8); + + /* enable DMA IBs */ + ib_cntl = DMA_IB_ENABLE | CMD_VMID_FORCE; +#ifdef __BIG_ENDIAN + ib_cntl |= DMA_IB_SWAP_ENABLE; +#endif + WREG32(DMA_IB_CNTL + reg_offset, ib_cntl); + + dma_cntl = RREG32(DMA_CNTL + reg_offset); + dma_cntl &= ~CTXEMPTY_INT_ENABLE; + WREG32(DMA_CNTL + reg_offset, dma_cntl); + + ring->wptr = 0; + WREG32(DMA_RB_WPTR + reg_offset, ring->wptr << 2); + + WREG32(DMA_RB_CNTL + reg_offset, rb_cntl | DMA_RB_ENABLE); + + ring->ready = true; + + r = radeon_ring_test(rdev, ring->idx, ring); + if (r) { + ring->ready = false; + return r; + } + } + + if ((rdev->asic->copy.copy_ring_index == R600_RING_TYPE_DMA_INDEX) || + (rdev->asic->copy.copy_ring_index == CAYMAN_RING_TYPE_DMA1_INDEX)) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); + + return 0; +} + +/** + * cayman_dma_fini - tear down the async dma engines + * + * @rdev: radeon_device pointer + * + * Stop the async dma engines and free the rings (cayman-SI). + */ +void cayman_dma_fini(struct radeon_device *rdev) +{ + cayman_dma_stop(rdev); + radeon_ring_fini(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX]); + radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]); +} + +/** + * cayman_dma_is_lockup - Check if the DMA engine is locked up + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * + * Check if the async DMA engine is locked up. + * Returns true if the engine appears to be locked up, false if not. + */ +bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) +{ + u32 reset_mask = cayman_gpu_check_soft_reset(rdev); + u32 mask; + + if (ring->idx == R600_RING_TYPE_DMA_INDEX) + mask = RADEON_RESET_DMA; + else + mask = RADEON_RESET_DMA1; + + if (!(reset_mask & mask)) { + radeon_ring_lockup_update(rdev, ring); + return false; + } + return radeon_ring_test_lockup(rdev, ring); +} + +/** + * cayman_dma_vm_set_page - update the page tables using the DMA + * + * @rdev: radeon_device pointer + * @ib: indirect buffer to fill with commands + * @pe: addr of the page entry + * @addr: dst addr to write into pe + * @count: number of page entries to update + * @incr: increase next addr by incr bytes + * @flags: hw access flags + * + * Update the page tables using the DMA (cayman/TN). + */ +void cayman_dma_vm_set_page(struct radeon_device *rdev, + struct radeon_ib *ib, + uint64_t pe, + uint64_t addr, unsigned count, + uint32_t incr, uint32_t flags) +{ + uint64_t value; + unsigned ndw; + + trace_radeon_vm_set_page(pe, addr, count, incr, flags); + + if ((flags & R600_PTE_SYSTEM) || (count == 1)) { + while (count) { + ndw = count * 2; + if (ndw > 0xFFFFE) + ndw = 0xFFFFE; + + /* for non-physically contiguous pages (system) */ + ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, ndw); + ib->ptr[ib->length_dw++] = pe; + ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; + for (; ndw > 0; ndw -= 2, --count, pe += 8) { + if (flags & R600_PTE_SYSTEM) { + value = radeon_vm_map_gart(rdev, addr); + value &= 0xFFFFFFFFFFFFF000ULL; + } else if (flags & R600_PTE_VALID) { + value = addr; + } else { + value = 0; + } + addr += incr; + value |= flags; + ib->ptr[ib->length_dw++] = value; + ib->ptr[ib->length_dw++] = upper_32_bits(value); + } + } + } else { + while (count) { + ndw = count * 2; + if (ndw > 0xFFFFE) + ndw = 0xFFFFE; + + if (flags & R600_PTE_VALID) + value = addr; + else + value = 0; + /* for physically contiguous pages (vram) */ + ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw); + ib->ptr[ib->length_dw++] = pe; /* dst addr */ + ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; + ib->ptr[ib->length_dw++] = flags; /* mask */ + ib->ptr[ib->length_dw++] = 0; + ib->ptr[ib->length_dw++] = value; /* value */ + ib->ptr[ib->length_dw++] = upper_32_bits(value); + ib->ptr[ib->length_dw++] = incr; /* increment size */ + ib->ptr[ib->length_dw++] = 0; + pe += ndw * 4; + addr += (ndw / 2) * incr; + count -= ndw / 2; + } + } + while (ib->length_dw & 0x7) + ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0); +} + +void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) +{ + struct radeon_ring *ring = &rdev->ring[ridx]; + + if (vm == NULL) + return; + + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0)); + radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2)); + radeon_ring_write(ring, vm->pd_gpu_addr >> 12); + + /* flush hdp cache */ + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0)); + radeon_ring_write(ring, (0xf << 16) | (HDP_MEM_COHERENCY_FLUSH_CNTL >> 2)); + radeon_ring_write(ring, 1); + + /* bits 0-7 are the VM contexts0-7 */ + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0)); + radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2)); + radeon_ring_write(ring, 1 << vm->id); +} + diff --git a/drivers/gpu/drm/radeon/ni_dpm.c b/drivers/gpu/drm/radeon/ni_dpm.c index 559cf24d51a..01fc4888e6f 100644 --- a/drivers/gpu/drm/radeon/ni_dpm.c +++ b/drivers/gpu/drm/radeon/ni_dpm.c @@ -720,6 +720,8 @@ static const u32 cayman_sysls_enable[] = struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev); struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev); +extern int ni_mc_load_microcode(struct radeon_device *rdev); + struct ni_power_info *ni_get_pi(struct radeon_device *rdev) { struct ni_power_info *pi = rdev->pm.dpm.priv; @@ -769,7 +771,8 @@ bool ni_dpm_vblank_too_short(struct radeon_device *rdev) { struct rv7xx_power_info *pi = rv770_get_pi(rdev); u32 vblank_time = r600_dpm_get_vblank_time(rdev); - u32 switch_limit = pi->mem_gddr5 ? 450 : 300; + /* we never hit the non-gddr5 limit so disable it */ + u32 switch_limit = pi->mem_gddr5 ? 450 : 0; if (vblank_time < switch_limit) return true; @@ -784,8 +787,9 @@ static void ni_apply_state_adjust_rules(struct radeon_device *rdev, struct ni_ps *ps = ni_get_ps(rps); struct radeon_clock_and_voltage_limits *max_limits; bool disable_mclk_switching; - u32 mclk, sclk; - u16 vddc, vddci; + u32 mclk; + u16 vddci; + u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc; int i; if ((rdev->pm.dpm.new_active_crtc_count > 1) || @@ -812,26 +816,39 @@ static void ni_apply_state_adjust_rules(struct radeon_device *rdev, } } + /* limit clocks to max supported clocks based on voltage dependency tables */ + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk, + &max_sclk_vddc); + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk, + &max_mclk_vddci); + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk, + &max_mclk_vddc); + + for (i = 0; i < ps->performance_level_count; i++) { + if (max_sclk_vddc) { + if (ps->performance_levels[i].sclk > max_sclk_vddc) + ps->performance_levels[i].sclk = max_sclk_vddc; + } + if (max_mclk_vddci) { + if (ps->performance_levels[i].mclk > max_mclk_vddci) + ps->performance_levels[i].mclk = max_mclk_vddci; + } + if (max_mclk_vddc) { + if (ps->performance_levels[i].mclk > max_mclk_vddc) + ps->performance_levels[i].mclk = max_mclk_vddc; + } + } + /* XXX validate the min clocks required for display */ + /* adjust low state */ if (disable_mclk_switching) { - mclk = ps->performance_levels[ps->performance_level_count - 1].mclk; - sclk = ps->performance_levels[0].sclk; - vddc = ps->performance_levels[0].vddc; - vddci = ps->performance_levels[ps->performance_level_count - 1].vddci; - } else { - sclk = ps->performance_levels[0].sclk; - mclk = ps->performance_levels[0].mclk; - vddc = ps->performance_levels[0].vddc; - vddci = ps->performance_levels[0].vddci; + ps->performance_levels[0].mclk = + ps->performance_levels[ps->performance_level_count - 1].mclk; + ps->performance_levels[0].vddci = + ps->performance_levels[ps->performance_level_count - 1].vddci; } - /* adjusted low state */ - ps->performance_levels[0].sclk = sclk; - ps->performance_levels[0].mclk = mclk; - ps->performance_levels[0].vddc = vddc; - ps->performance_levels[0].vddci = vddci; - btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk, &ps->performance_levels[0].sclk, &ps->performance_levels[0].mclk); @@ -843,11 +860,15 @@ static void ni_apply_state_adjust_rules(struct radeon_device *rdev, ps->performance_levels[i].vddc = ps->performance_levels[i - 1].vddc; } + /* adjust remaining states */ if (disable_mclk_switching) { mclk = ps->performance_levels[0].mclk; + vddci = ps->performance_levels[0].vddci; for (i = 1; i < ps->performance_level_count; i++) { if (mclk < ps->performance_levels[i].mclk) mclk = ps->performance_levels[i].mclk; + if (vddci < ps->performance_levels[i].vddci) + vddci = ps->performance_levels[i].vddci; } for (i = 0; i < ps->performance_level_count; i++) { ps->performance_levels[i].mclk = mclk; @@ -1054,10 +1075,6 @@ static int ni_restrict_performance_levels_before_switch(struct radeon_device *rd int ni_dpm_force_performance_level(struct radeon_device *rdev, enum radeon_dpm_forced_level level) { - struct radeon_ps *rps = rdev->pm.dpm.current_ps; - struct ni_ps *ps = ni_get_ps(rps); - u32 levels; - if (level == RADEON_DPM_FORCED_LEVEL_HIGH) { if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK) return -EINVAL; @@ -1068,8 +1085,7 @@ int ni_dpm_force_performance_level(struct radeon_device *rdev, if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK) return -EINVAL; - levels = ps->performance_level_count - 1; - if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, levels) != PPSMC_Result_OK) + if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 1) != PPSMC_Result_OK) return -EINVAL; } else if (level == RADEON_DPM_FORCED_LEVEL_AUTO) { if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK) @@ -1299,7 +1315,7 @@ static void ni_populate_smc_voltage_tables(struct radeon_device *rdev, table->voltageMaskTable.highMask[NISLANDS_SMC_VOLTAGEMASK_VDDCI] = 0; table->voltageMaskTable.lowMask[NISLANDS_SMC_VOLTAGEMASK_VDDCI] = - cpu_to_be32(eg_pi->vddc_voltage_table.mask_low); + cpu_to_be32(eg_pi->vddci_voltage_table.mask_low); } } @@ -2572,7 +2588,7 @@ static int ni_populate_sq_ramping_values(struct radeon_device *rdev, if (NISLANDS_DPM2_SQ_RAMP_STI_SIZE > (STI_SIZE_MASK >> STI_SIZE_SHIFT)) enable_sq_ramping = false; - if (NISLANDS_DPM2_SQ_RAMP_LTI_RATIO <= (LTI_RATIO_MASK >> LTI_RATIO_SHIFT)) + if (NISLANDS_DPM2_SQ_RAMP_LTI_RATIO > (LTI_RATIO_MASK >> LTI_RATIO_SHIFT)) enable_sq_ramping = false; for (i = 0; i < state->performance_level_count; i++) { @@ -3425,9 +3441,9 @@ static int ni_enable_smc_cac(struct radeon_device *rdev, static int ni_pcie_performance_request(struct radeon_device *rdev, u8 perf_req, bool advertise) { +#if defined(CONFIG_ACPI) struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev); -#if defined(CONFIG_ACPI) if ((perf_req == PCIE_PERF_REQ_PECI_GEN1) || (perf_req == PCIE_PERF_REQ_PECI_GEN2)) { if (eg_pi->pcie_performance_request_registered == false) @@ -3551,7 +3567,11 @@ void ni_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev, void ni_dpm_setup_asic(struct radeon_device *rdev) { struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev); + int r; + r = ni_mc_load_microcode(rdev); + if (r) + DRM_ERROR("Failed to load MC firmware!\n"); ni_read_clock_registers(rdev); btc_read_arb_registers(rdev); rv770_get_memory_type(rdev); @@ -3696,21 +3716,6 @@ int ni_dpm_enable(struct radeon_device *rdev) if (eg_pi->ls_clock_gating) ni_ls_clockgating_enable(rdev, true); - if (rdev->irq.installed && - r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { - PPSMC_Result result; - - ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, 0xff * 1000); - if (ret) - return ret; - rdev->irq.dpm_thermal = true; - radeon_irq_set(rdev); - result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt); - - if (result != PPSMC_Result_OK) - DRM_DEBUG_KMS("Could not enable thermal interrupts.\n"); - } - rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true); ni_update_current_ps(rdev, boot_ps); @@ -3869,12 +3874,6 @@ int ni_dpm_set_power_state(struct radeon_device *rdev) return ret; } - ret = ni_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO); - if (ret) { - DRM_ERROR("ni_dpm_force_performance_level failed\n"); - return ret; - } - return 0; } @@ -3946,7 +3945,6 @@ static void ni_parse_pplib_clock_info(struct radeon_device *rdev, struct rv7xx_power_info *pi = rv770_get_pi(rdev); struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev); struct ni_ps *ps = ni_get_ps(rps); - u16 vddc; struct rv7xx_pl *pl = &ps->performance_levels[index]; ps->performance_level_count = index + 1; @@ -3962,8 +3960,8 @@ static void ni_parse_pplib_clock_info(struct radeon_device *rdev, /* patch up vddc if necessary */ if (pl->vddc == 0xff01) { - if (radeon_atom_get_max_vddc(rdev, 0, 0, &vddc) == 0) - pl->vddc = vddc; + if (pi->max_vddc) + pl->vddc = pi->max_vddc; } if (rps->class & ATOM_PPLIB_CLASSIFICATION_ACPI) { @@ -4027,9 +4025,6 @@ static int ni_parse_power_table(struct radeon_device *rdev) power_info->pplib.ucNumStates, GFP_KERNEL); if (!rdev->pm.dpm.ps) return -ENOMEM; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < power_info->pplib.ucNumStates; i++) { power_state = (union pplib_power_state *) @@ -4042,6 +4037,7 @@ static int ni_parse_power_table(struct radeon_device *rdev) (power_state->v1.ucNonClockStateIndex * power_info->pplib.ucNonClockSize)); if (power_info->pplib.ucStateEntrySize - 1) { + u8 *idx; ps = kzalloc(sizeof(struct ni_ps), GFP_KERNEL); if (ps == NULL) { kfree(rdev->pm.dpm.ps); @@ -4051,12 +4047,12 @@ static int ni_parse_power_table(struct radeon_device *rdev) ni_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i], non_clock_info, power_info->pplib.ucNonClockSize); + idx = (u8 *)&power_state->v1.ucClockStateIndices[0]; for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) { clock_info = (union pplib_clock_info *) (mode_info->atom_context->bios + data_offset + le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) + - (power_state->v1.ucClockStateIndices[j] * - power_info->pplib.ucClockInfoSize)); + (idx[j] * power_info->pplib.ucClockInfoSize)); ni_parse_pplib_clock_info(rdev, &rdev->pm.dpm.ps[i], j, clock_info); @@ -4072,9 +4068,6 @@ int ni_dpm_init(struct radeon_device *rdev) struct rv7xx_power_info *pi; struct evergreen_power_info *eg_pi; struct ni_power_info *ni_pi; - int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info); - u16 data_offset, size; - u8 frev, crev; struct atom_clock_dividers dividers; int ret; @@ -4093,6 +4086,10 @@ int ni_dpm_init(struct radeon_device *rdev) pi->min_vddc_in_table = 0; pi->max_vddc_in_table = 0; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = ni_parse_power_table(rdev); if (ret) return ret; @@ -4167,16 +4164,7 @@ int ni_dpm_init(struct radeon_device *rdev) eg_pi->vddci_control = radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0); - if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size, - &frev, &crev, &data_offset)) { - pi->sclk_ss = true; - pi->mclk_ss = true; - pi->dynamic_ss = true; - } else { - pi->sclk_ss = false; - pi->mclk_ss = false; - pi->dynamic_ss = true; - } + rv770_get_engine_memory_ss(rdev); pi->asi = RV770_ASI_DFLT; pi->pasi = CYPRESS_HASI_DFLT; @@ -4193,8 +4181,7 @@ int ni_dpm_init(struct radeon_device *rdev) pi->dynamic_pcie_gen2 = true; - if (pi->gfx_clock_gating && - (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE)) + if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE) pi->thermal_protection = true; else pi->thermal_protection = false; @@ -4288,6 +4275,12 @@ int ni_dpm_init(struct radeon_device *rdev) ni_pi->use_power_boost_limit = true; + /* make sure dc limits are valid */ + if ((rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.sclk == 0) || + (rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.mclk == 0)) + rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc = + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + return 0; } @@ -4329,7 +4322,8 @@ void ni_dpm_print_power_state(struct radeon_device *rdev, void ni_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, struct seq_file *m) { - struct radeon_ps *rps = rdev->pm.dpm.current_ps; + struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev); + struct radeon_ps *rps = &eg_pi->current_rps; struct ni_ps *ps = ni_get_ps(rps); struct rv7xx_pl *pl; u32 current_index = diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h index fe24a93542e..2e12e4d6925 100644 --- a/drivers/gpu/drm/radeon/nid.h +++ b/drivers/gpu/drm/radeon/nid.h @@ -128,11 +128,28 @@ #define READ_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 16) #define WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 18) #define WRITE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 19) +#define PAGE_TABLE_BLOCK_SIZE(x) (((x) & 0xF) << 24) #define VM_CONTEXT1_CNTL 0x1414 #define VM_CONTEXT0_CNTL2 0x1430 #define VM_CONTEXT1_CNTL2 0x1434 #define VM_INVALIDATE_REQUEST 0x1478 #define VM_INVALIDATE_RESPONSE 0x147c +#define VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x14FC +#define VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x14DC +#define PROTECTIONS_MASK (0xf << 0) +#define PROTECTIONS_SHIFT 0 + /* bit 0: range + * bit 2: pde0 + * bit 3: valid + * bit 4: read + * bit 5: write + */ +#define MEMORY_CLIENT_ID_MASK (0xff << 12) +#define MEMORY_CLIENT_ID_SHIFT 12 +#define MEMORY_CLIENT_RW_MASK (1 << 24) +#define MEMORY_CLIENT_RW_SHIFT 24 +#define FAULT_VMID_MASK (0x7 << 25) +#define FAULT_VMID_SHIFT 25 #define VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR 0x1518 #define VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR 0x151c #define VM_CONTEXT0_PAGE_TABLE_BASE_ADDR 0x153C @@ -1138,6 +1155,7 @@ # define PACKET3_DB_ACTION_ENA (1 << 26) # define PACKET3_SH_ACTION_ENA (1 << 27) # define PACKET3_SX_ACTION_ENA (1 << 28) +# define PACKET3_ENGINE_ME (1 << 31) #define PACKET3_ME_INITIALIZE 0x44 #define PACKET3_ME_INITIALIZE_DEVICE_ID(x) ((x) << 16) #define PACKET3_COND_WRITE 0x45 diff --git a/drivers/gpu/drm/radeon/ppsmc.h b/drivers/gpu/drm/radeon/ppsmc.h index b5564a3645d..5670b829128 100644 --- a/drivers/gpu/drm/radeon/ppsmc.h +++ b/drivers/gpu/drm/radeon/ppsmc.h @@ -99,13 +99,72 @@ typedef uint8_t PPSMC_Result; #define PPSMC_MSG_ThrottleOVRDSCLKDS ((uint8_t)0x96) #define PPSMC_MSG_CancelThrottleOVRDSCLKDS ((uint8_t)0x97) +/* CI/KV/KB */ +#define PPSMC_MSG_UVDDPM_SetEnabledMask ((uint16_t) 0x12D) +#define PPSMC_MSG_VCEDPM_SetEnabledMask ((uint16_t) 0x12E) +#define PPSMC_MSG_ACPDPM_SetEnabledMask ((uint16_t) 0x12F) +#define PPSMC_MSG_SAMUDPM_SetEnabledMask ((uint16_t) 0x130) +#define PPSMC_MSG_MCLKDPM_ForceState ((uint16_t) 0x131) +#define PPSMC_MSG_MCLKDPM_NoForcedLevel ((uint16_t) 0x132) +#define PPSMC_MSG_Voltage_Cntl_Disable ((uint16_t) 0x135) +#define PPSMC_MSG_PCIeDPM_Enable ((uint16_t) 0x136) +#define PPSMC_MSG_PCIeDPM_Disable ((uint16_t) 0x13d) +#define PPSMC_MSG_ACPPowerOFF ((uint16_t) 0x137) +#define PPSMC_MSG_ACPPowerON ((uint16_t) 0x138) +#define PPSMC_MSG_SAMPowerOFF ((uint16_t) 0x139) +#define PPSMC_MSG_SAMPowerON ((uint16_t) 0x13a) +#define PPSMC_MSG_PCIeDPM_Disable ((uint16_t) 0x13d) +#define PPSMC_MSG_NBDPM_Enable ((uint16_t) 0x140) +#define PPSMC_MSG_NBDPM_Disable ((uint16_t) 0x141) +#define PPSMC_MSG_SCLKDPM_SetEnabledMask ((uint16_t) 0x145) +#define PPSMC_MSG_MCLKDPM_SetEnabledMask ((uint16_t) 0x146) +#define PPSMC_MSG_PCIeDPM_ForceLevel ((uint16_t) 0x147) +#define PPSMC_MSG_PCIeDPM_UnForceLevel ((uint16_t) 0x148) +#define PPSMC_MSG_EnableVRHotGPIOInterrupt ((uint16_t) 0x14a) +#define PPSMC_MSG_DPM_Enable ((uint16_t) 0x14e) +#define PPSMC_MSG_DPM_Disable ((uint16_t) 0x14f) +#define PPSMC_MSG_MCLKDPM_Enable ((uint16_t) 0x150) +#define PPSMC_MSG_MCLKDPM_Disable ((uint16_t) 0x151) +#define PPSMC_MSG_UVDDPM_Enable ((uint16_t) 0x154) +#define PPSMC_MSG_UVDDPM_Disable ((uint16_t) 0x155) +#define PPSMC_MSG_SAMUDPM_Enable ((uint16_t) 0x156) +#define PPSMC_MSG_SAMUDPM_Disable ((uint16_t) 0x157) +#define PPSMC_MSG_ACPDPM_Enable ((uint16_t) 0x158) +#define PPSMC_MSG_ACPDPM_Disable ((uint16_t) 0x159) +#define PPSMC_MSG_VCEDPM_Enable ((uint16_t) 0x15a) +#define PPSMC_MSG_VCEDPM_Disable ((uint16_t) 0x15b) +#define PPSMC_MSG_VddC_Request ((uint16_t) 0x15f) +#define PPSMC_MSG_SCLKDPM_GetEnabledMask ((uint16_t) 0x162) +#define PPSMC_MSG_PCIeDPM_SetEnabledMask ((uint16_t) 0x167) +#define PPSMC_MSG_TDCLimitEnable ((uint16_t) 0x169) +#define PPSMC_MSG_TDCLimitDisable ((uint16_t) 0x16a) +#define PPSMC_MSG_PkgPwrLimitEnable ((uint16_t) 0x185) +#define PPSMC_MSG_PkgPwrLimitDisable ((uint16_t) 0x186) +#define PPSMC_MSG_PkgPwrSetLimit ((uint16_t) 0x187) +#define PPSMC_MSG_OverDriveSetTargetTdp ((uint16_t) 0x188) +#define PPSMC_MSG_SCLKDPM_FreezeLevel ((uint16_t) 0x189) +#define PPSMC_MSG_SCLKDPM_UnfreezeLevel ((uint16_t) 0x18A) +#define PPSMC_MSG_MCLKDPM_FreezeLevel ((uint16_t) 0x18B) +#define PPSMC_MSG_MCLKDPM_UnfreezeLevel ((uint16_t) 0x18C) +#define PPSMC_MSG_MASTER_DeepSleep_ON ((uint16_t) 0x18F) +#define PPSMC_MSG_MASTER_DeepSleep_OFF ((uint16_t) 0x190) +#define PPSMC_MSG_Remove_DC_Clamp ((uint16_t) 0x191) + +#define PPSMC_MSG_API_GetSclkFrequency ((uint16_t) 0x200) +#define PPSMC_MSG_API_GetMclkFrequency ((uint16_t) 0x201) + /* TN */ #define PPSMC_MSG_DPM_Config ((uint32_t) 0x102) #define PPSMC_MSG_DPM_ForceState ((uint32_t) 0x104) #define PPSMC_MSG_PG_SIMD_Config ((uint32_t) 0x108) #define PPSMC_MSG_DPM_N_LevelsDisabled ((uint32_t) 0x112) +#define PPSMC_MSG_Voltage_Cntl_Enable ((uint32_t) 0x109) +#define PPSMC_MSG_VCEPowerOFF ((uint32_t) 0x10e) +#define PPSMC_MSG_VCEPowerON ((uint32_t) 0x10f) #define PPSMC_MSG_DCE_RemoveVoltageAdjustment ((uint32_t) 0x11d) #define PPSMC_MSG_DCE_AllowVoltageAdjustment ((uint32_t) 0x11e) +#define PPSMC_MSG_EnableBAPM ((uint32_t) 0x120) +#define PPSMC_MSG_DisableBAPM ((uint32_t) 0x121) #define PPSMC_MSG_UVD_DPM_Config ((uint32_t) 0x124) diff --git a/drivers/gpu/drm/radeon/pptable.h b/drivers/gpu/drm/radeon/pptable.h new file mode 100644 index 00000000000..2d532996c69 --- /dev/null +++ b/drivers/gpu/drm/radeon/pptable.h @@ -0,0 +1,682 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef _PPTABLE_H +#define _PPTABLE_H + +#pragma pack(1) + +typedef struct _ATOM_PPLIB_THERMALCONTROLLER + +{ + UCHAR ucType; // one of ATOM_PP_THERMALCONTROLLER_* + UCHAR ucI2cLine; // as interpreted by DAL I2C + UCHAR ucI2cAddress; + UCHAR ucFanParameters; // Fan Control Parameters. + UCHAR ucFanMinRPM; // Fan Minimum RPM (hundreds) -- for display purposes only. + UCHAR ucFanMaxRPM; // Fan Maximum RPM (hundreds) -- for display purposes only. + UCHAR ucReserved; // ---- + UCHAR ucFlags; // to be defined +} ATOM_PPLIB_THERMALCONTROLLER; + +#define ATOM_PP_FANPARAMETERS_TACHOMETER_PULSES_PER_REVOLUTION_MASK 0x0f +#define ATOM_PP_FANPARAMETERS_NOFAN 0x80 // No fan is connected to this controller. + +#define ATOM_PP_THERMALCONTROLLER_NONE 0 +#define ATOM_PP_THERMALCONTROLLER_LM63 1 // Not used by PPLib +#define ATOM_PP_THERMALCONTROLLER_ADM1032 2 // Not used by PPLib +#define ATOM_PP_THERMALCONTROLLER_ADM1030 3 // Not used by PPLib +#define ATOM_PP_THERMALCONTROLLER_MUA6649 4 // Not used by PPLib +#define ATOM_PP_THERMALCONTROLLER_LM64 5 +#define ATOM_PP_THERMALCONTROLLER_F75375 6 // Not used by PPLib +#define ATOM_PP_THERMALCONTROLLER_RV6xx 7 +#define ATOM_PP_THERMALCONTROLLER_RV770 8 +#define ATOM_PP_THERMALCONTROLLER_ADT7473 9 +#define ATOM_PP_THERMALCONTROLLER_KONG 10 +#define ATOM_PP_THERMALCONTROLLER_EXTERNAL_GPIO 11 +#define ATOM_PP_THERMALCONTROLLER_EVERGREEN 12 +#define ATOM_PP_THERMALCONTROLLER_EMC2103 13 /* 0x0D */ // Only fan control will be implemented, do NOT show this in PPGen. +#define ATOM_PP_THERMALCONTROLLER_SUMO 14 /* 0x0E */ // Sumo type, used internally +#define ATOM_PP_THERMALCONTROLLER_NISLANDS 15 +#define ATOM_PP_THERMALCONTROLLER_SISLANDS 16 +#define ATOM_PP_THERMALCONTROLLER_LM96163 17 +#define ATOM_PP_THERMALCONTROLLER_CISLANDS 18 +#define ATOM_PP_THERMALCONTROLLER_KAVERI 19 + + +// Thermal controller 'combo type' to use an external controller for Fan control and an internal controller for thermal. +// We probably should reserve the bit 0x80 for this use. +// To keep the number of these types low we should also use the same code for all ASICs (i.e. do not distinguish RV6xx and RV7xx Internal here). +// The driver can pick the correct internal controller based on the ASIC. + +#define ATOM_PP_THERMALCONTROLLER_ADT7473_WITH_INTERNAL 0x89 // ADT7473 Fan Control + Internal Thermal Controller +#define ATOM_PP_THERMALCONTROLLER_EMC2103_WITH_INTERNAL 0x8D // EMC2103 Fan Control + Internal Thermal Controller + +typedef struct _ATOM_PPLIB_STATE +{ + UCHAR ucNonClockStateIndex; + UCHAR ucClockStateIndices[1]; // variable-sized +} ATOM_PPLIB_STATE; + + +typedef struct _ATOM_PPLIB_FANTABLE +{ + UCHAR ucFanTableFormat; // Change this if the table format changes or version changes so that the other fields are not the same. + UCHAR ucTHyst; // Temperature hysteresis. Integer. + USHORT usTMin; // The temperature, in 0.01 centigrades, below which we just run at a minimal PWM. + USHORT usTMed; // The middle temperature where we change slopes. + USHORT usTHigh; // The high point above TMed for adjusting the second slope. + USHORT usPWMMin; // The minimum PWM value in percent (0.01% increments). + USHORT usPWMMed; // The PWM value (in percent) at TMed. + USHORT usPWMHigh; // The PWM value at THigh. +} ATOM_PPLIB_FANTABLE; + +typedef struct _ATOM_PPLIB_FANTABLE2 +{ + ATOM_PPLIB_FANTABLE basicTable; + USHORT usTMax; // The max temperature +} ATOM_PPLIB_FANTABLE2; + +typedef struct _ATOM_PPLIB_EXTENDEDHEADER +{ + USHORT usSize; + ULONG ulMaxEngineClock; // For Overdrive. + ULONG ulMaxMemoryClock; // For Overdrive. + // Add extra system parameters here, always adjust size to include all fields. + USHORT usVCETableOffset; //points to ATOM_PPLIB_VCE_Table + USHORT usUVDTableOffset; //points to ATOM_PPLIB_UVD_Table + USHORT usSAMUTableOffset; //points to ATOM_PPLIB_SAMU_Table + USHORT usPPMTableOffset; //points to ATOM_PPLIB_PPM_Table + USHORT usACPTableOffset; //points to ATOM_PPLIB_ACP_Table + USHORT usPowerTuneTableOffset; //points to ATOM_PPLIB_POWERTUNE_Table +} ATOM_PPLIB_EXTENDEDHEADER; + +//// ATOM_PPLIB_POWERPLAYTABLE::ulPlatformCaps +#define ATOM_PP_PLATFORM_CAP_BACKBIAS 1 +#define ATOM_PP_PLATFORM_CAP_POWERPLAY 2 +#define ATOM_PP_PLATFORM_CAP_SBIOSPOWERSOURCE 4 +#define ATOM_PP_PLATFORM_CAP_ASPM_L0s 8 +#define ATOM_PP_PLATFORM_CAP_ASPM_L1 16 +#define ATOM_PP_PLATFORM_CAP_HARDWAREDC 32 +#define ATOM_PP_PLATFORM_CAP_GEMINIPRIMARY 64 +#define ATOM_PP_PLATFORM_CAP_STEPVDDC 128 +#define ATOM_PP_PLATFORM_CAP_VOLTAGECONTROL 256 +#define ATOM_PP_PLATFORM_CAP_SIDEPORTCONTROL 512 +#define ATOM_PP_PLATFORM_CAP_TURNOFFPLL_ASPML1 1024 +#define ATOM_PP_PLATFORM_CAP_HTLINKCONTROL 2048 +#define ATOM_PP_PLATFORM_CAP_MVDDCONTROL 4096 +#define ATOM_PP_PLATFORM_CAP_GOTO_BOOT_ON_ALERT 0x2000 // Go to boot state on alerts, e.g. on an AC->DC transition. +#define ATOM_PP_PLATFORM_CAP_DONT_WAIT_FOR_VBLANK_ON_ALERT 0x4000 // Do NOT wait for VBLANK during an alert (e.g. AC->DC transition). +#define ATOM_PP_PLATFORM_CAP_VDDCI_CONTROL 0x8000 // Does the driver control VDDCI independently from VDDC. +#define ATOM_PP_PLATFORM_CAP_REGULATOR_HOT 0x00010000 // Enable the 'regulator hot' feature. +#define ATOM_PP_PLATFORM_CAP_BACO 0x00020000 // Does the driver supports BACO state. +#define ATOM_PP_PLATFORM_CAP_NEW_CAC_VOLTAGE 0x00040000 // Does the driver supports new CAC voltage table. +#define ATOM_PP_PLATFORM_CAP_REVERT_GPIO5_POLARITY 0x00080000 // Does the driver supports revert GPIO5 polarity. +#define ATOM_PP_PLATFORM_CAP_OUTPUT_THERMAL2GPIO17 0x00100000 // Does the driver supports thermal2GPIO17. +#define ATOM_PP_PLATFORM_CAP_VRHOT_GPIO_CONFIGURABLE 0x00200000 // Does the driver supports VR HOT GPIO Configurable. +#define ATOM_PP_PLATFORM_CAP_TEMP_INVERSION 0x00400000 // Does the driver supports Temp Inversion feature. +#define ATOM_PP_PLATFORM_CAP_EVV 0x00800000 + +typedef struct _ATOM_PPLIB_POWERPLAYTABLE +{ + ATOM_COMMON_TABLE_HEADER sHeader; + + UCHAR ucDataRevision; + + UCHAR ucNumStates; + UCHAR ucStateEntrySize; + UCHAR ucClockInfoSize; + UCHAR ucNonClockSize; + + // offset from start of this table to array of ucNumStates ATOM_PPLIB_STATE structures + USHORT usStateArrayOffset; + + // offset from start of this table to array of ASIC-specific structures, + // currently ATOM_PPLIB_CLOCK_INFO. + USHORT usClockInfoArrayOffset; + + // offset from start of this table to array of ATOM_PPLIB_NONCLOCK_INFO + USHORT usNonClockInfoArrayOffset; + + USHORT usBackbiasTime; // in microseconds + USHORT usVoltageTime; // in microseconds + USHORT usTableSize; //the size of this structure, or the extended structure + + ULONG ulPlatformCaps; // See ATOM_PPLIB_CAPS_* + + ATOM_PPLIB_THERMALCONTROLLER sThermalController; + + USHORT usBootClockInfoOffset; + USHORT usBootNonClockInfoOffset; + +} ATOM_PPLIB_POWERPLAYTABLE; + +typedef struct _ATOM_PPLIB_POWERPLAYTABLE2 +{ + ATOM_PPLIB_POWERPLAYTABLE basicTable; + UCHAR ucNumCustomThermalPolicy; + USHORT usCustomThermalPolicyArrayOffset; +}ATOM_PPLIB_POWERPLAYTABLE2, *LPATOM_PPLIB_POWERPLAYTABLE2; + +typedef struct _ATOM_PPLIB_POWERPLAYTABLE3 +{ + ATOM_PPLIB_POWERPLAYTABLE2 basicTable2; + USHORT usFormatID; // To be used ONLY by PPGen. + USHORT usFanTableOffset; + USHORT usExtendendedHeaderOffset; +} ATOM_PPLIB_POWERPLAYTABLE3, *LPATOM_PPLIB_POWERPLAYTABLE3; + +typedef struct _ATOM_PPLIB_POWERPLAYTABLE4 +{ + ATOM_PPLIB_POWERPLAYTABLE3 basicTable3; + ULONG ulGoldenPPID; // PPGen use only + ULONG ulGoldenRevision; // PPGen use only + USHORT usVddcDependencyOnSCLKOffset; + USHORT usVddciDependencyOnMCLKOffset; + USHORT usVddcDependencyOnMCLKOffset; + USHORT usMaxClockVoltageOnDCOffset; + USHORT usVddcPhaseShedLimitsTableOffset; // Points to ATOM_PPLIB_PhaseSheddingLimits_Table + USHORT usMvddDependencyOnMCLKOffset; +} ATOM_PPLIB_POWERPLAYTABLE4, *LPATOM_PPLIB_POWERPLAYTABLE4; + +typedef struct _ATOM_PPLIB_POWERPLAYTABLE5 +{ + ATOM_PPLIB_POWERPLAYTABLE4 basicTable4; + ULONG ulTDPLimit; + ULONG ulNearTDPLimit; + ULONG ulSQRampingThreshold; + USHORT usCACLeakageTableOffset; // Points to ATOM_PPLIB_CAC_Leakage_Table + ULONG ulCACLeakage; // The iLeakage for driver calculated CAC leakage table + USHORT usTDPODLimit; + USHORT usLoadLineSlope; // in milliOhms * 100 +} ATOM_PPLIB_POWERPLAYTABLE5, *LPATOM_PPLIB_POWERPLAYTABLE5; + +//// ATOM_PPLIB_NONCLOCK_INFO::usClassification +#define ATOM_PPLIB_CLASSIFICATION_UI_MASK 0x0007 +#define ATOM_PPLIB_CLASSIFICATION_UI_SHIFT 0 +#define ATOM_PPLIB_CLASSIFICATION_UI_NONE 0 +#define ATOM_PPLIB_CLASSIFICATION_UI_BATTERY 1 +#define ATOM_PPLIB_CLASSIFICATION_UI_BALANCED 3 +#define ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE 5 +// 2, 4, 6, 7 are reserved + +#define ATOM_PPLIB_CLASSIFICATION_BOOT 0x0008 +#define ATOM_PPLIB_CLASSIFICATION_THERMAL 0x0010 +#define ATOM_PPLIB_CLASSIFICATION_LIMITEDPOWERSOURCE 0x0020 +#define ATOM_PPLIB_CLASSIFICATION_REST 0x0040 +#define ATOM_PPLIB_CLASSIFICATION_FORCED 0x0080 +#define ATOM_PPLIB_CLASSIFICATION_3DPERFORMANCE 0x0100 +#define ATOM_PPLIB_CLASSIFICATION_OVERDRIVETEMPLATE 0x0200 +#define ATOM_PPLIB_CLASSIFICATION_UVDSTATE 0x0400 +#define ATOM_PPLIB_CLASSIFICATION_3DLOW 0x0800 +#define ATOM_PPLIB_CLASSIFICATION_ACPI 0x1000 +#define ATOM_PPLIB_CLASSIFICATION_HD2STATE 0x2000 +#define ATOM_PPLIB_CLASSIFICATION_HDSTATE 0x4000 +#define ATOM_PPLIB_CLASSIFICATION_SDSTATE 0x8000 + +//// ATOM_PPLIB_NONCLOCK_INFO::usClassification2 +#define ATOM_PPLIB_CLASSIFICATION2_LIMITEDPOWERSOURCE_2 0x0001 +#define ATOM_PPLIB_CLASSIFICATION2_ULV 0x0002 +#define ATOM_PPLIB_CLASSIFICATION2_MVC 0x0004 //Multi-View Codec (BD-3D) + +//// ATOM_PPLIB_NONCLOCK_INFO::ulCapsAndSettings +#define ATOM_PPLIB_SINGLE_DISPLAY_ONLY 0x00000001 +#define ATOM_PPLIB_SUPPORTS_VIDEO_PLAYBACK 0x00000002 + +// 0 is 2.5Gb/s, 1 is 5Gb/s +#define ATOM_PPLIB_PCIE_LINK_SPEED_MASK 0x00000004 +#define ATOM_PPLIB_PCIE_LINK_SPEED_SHIFT 2 + +// lanes - 1: 1, 2, 4, 8, 12, 16 permitted by PCIE spec +#define ATOM_PPLIB_PCIE_LINK_WIDTH_MASK 0x000000F8 +#define ATOM_PPLIB_PCIE_LINK_WIDTH_SHIFT 3 + +// lookup into reduced refresh-rate table +#define ATOM_PPLIB_LIMITED_REFRESHRATE_VALUE_MASK 0x00000F00 +#define ATOM_PPLIB_LIMITED_REFRESHRATE_VALUE_SHIFT 8 + +#define ATOM_PPLIB_LIMITED_REFRESHRATE_UNLIMITED 0 +#define ATOM_PPLIB_LIMITED_REFRESHRATE_50HZ 1 +// 2-15 TBD as needed. + +#define ATOM_PPLIB_SOFTWARE_DISABLE_LOADBALANCING 0x00001000 +#define ATOM_PPLIB_SOFTWARE_ENABLE_SLEEP_FOR_TIMESTAMPS 0x00002000 + +#define ATOM_PPLIB_DISALLOW_ON_DC 0x00004000 + +#define ATOM_PPLIB_ENABLE_VARIBRIGHT 0x00008000 + +//memory related flags +#define ATOM_PPLIB_SWSTATE_MEMORY_DLL_OFF 0x000010000 + +//M3 Arb //2bits, current 3 sets of parameters in total +#define ATOM_PPLIB_M3ARB_MASK 0x00060000 +#define ATOM_PPLIB_M3ARB_SHIFT 17 + +#define ATOM_PPLIB_ENABLE_DRR 0x00080000 + +// remaining 16 bits are reserved +typedef struct _ATOM_PPLIB_THERMAL_STATE +{ + UCHAR ucMinTemperature; + UCHAR ucMaxTemperature; + UCHAR ucThermalAction; +}ATOM_PPLIB_THERMAL_STATE, *LPATOM_PPLIB_THERMAL_STATE; + +// Contained in an array starting at the offset +// in ATOM_PPLIB_POWERPLAYTABLE::usNonClockInfoArrayOffset. +// referenced from ATOM_PPLIB_STATE_INFO::ucNonClockStateIndex +#define ATOM_PPLIB_NONCLOCKINFO_VER1 12 +#define ATOM_PPLIB_NONCLOCKINFO_VER2 24 +typedef struct _ATOM_PPLIB_NONCLOCK_INFO +{ + USHORT usClassification; + UCHAR ucMinTemperature; + UCHAR ucMaxTemperature; + ULONG ulCapsAndSettings; + UCHAR ucRequiredPower; + USHORT usClassification2; + ULONG ulVCLK; + ULONG ulDCLK; + UCHAR ucUnused[5]; +} ATOM_PPLIB_NONCLOCK_INFO; + +// Contained in an array starting at the offset +// in ATOM_PPLIB_POWERPLAYTABLE::usClockInfoArrayOffset. +// referenced from ATOM_PPLIB_STATE::ucClockStateIndices +typedef struct _ATOM_PPLIB_R600_CLOCK_INFO +{ + USHORT usEngineClockLow; + UCHAR ucEngineClockHigh; + + USHORT usMemoryClockLow; + UCHAR ucMemoryClockHigh; + + USHORT usVDDC; + USHORT usUnused1; + USHORT usUnused2; + + ULONG ulFlags; // ATOM_PPLIB_R600_FLAGS_* + +} ATOM_PPLIB_R600_CLOCK_INFO; + +// ulFlags in ATOM_PPLIB_R600_CLOCK_INFO +#define ATOM_PPLIB_R600_FLAGS_PCIEGEN2 1 +#define ATOM_PPLIB_R600_FLAGS_UVDSAFE 2 +#define ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE 4 +#define ATOM_PPLIB_R600_FLAGS_MEMORY_ODT_OFF 8 +#define ATOM_PPLIB_R600_FLAGS_MEMORY_DLL_OFF 16 +#define ATOM_PPLIB_R600_FLAGS_LOWPOWER 32 // On the RV770 use 'low power' setting (sequencer S0). + +typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO + +{ + USHORT usLowEngineClockLow; // Low Engine clock in MHz (the same way as on the R600). + UCHAR ucLowEngineClockHigh; + USHORT usHighEngineClockLow; // High Engine clock in MHz. + UCHAR ucHighEngineClockHigh; + USHORT usMemoryClockLow; // For now one of the ATOM_PPLIB_RS780_SPMCLK_XXXX constants. + UCHAR ucMemoryClockHigh; // Currentyl unused. + UCHAR ucPadding; // For proper alignment and size. + USHORT usVDDC; // For the 780, use: None, Low, High, Variable + UCHAR ucMaxHTLinkWidth; // From SBIOS - {2, 4, 8, 16} + UCHAR ucMinHTLinkWidth; // From SBIOS - {2, 4, 8, 16}. Effective only if CDLW enabled. Minimum down stream width could + USHORT usHTLinkFreq; // See definition ATOM_PPLIB_RS780_HTLINKFREQ_xxx or in MHz(>=200). + ULONG ulFlags; +} ATOM_PPLIB_RS780_CLOCK_INFO; + +#define ATOM_PPLIB_RS780_VOLTAGE_NONE 0 +#define ATOM_PPLIB_RS780_VOLTAGE_LOW 1 +#define ATOM_PPLIB_RS780_VOLTAGE_HIGH 2 +#define ATOM_PPLIB_RS780_VOLTAGE_VARIABLE 3 + +#define ATOM_PPLIB_RS780_SPMCLK_NONE 0 // We cannot change the side port memory clock, leave it as it is. +#define ATOM_PPLIB_RS780_SPMCLK_LOW 1 +#define ATOM_PPLIB_RS780_SPMCLK_HIGH 2 + +#define ATOM_PPLIB_RS780_HTLINKFREQ_NONE 0 +#define ATOM_PPLIB_RS780_HTLINKFREQ_LOW 1 +#define ATOM_PPLIB_RS780_HTLINKFREQ_HIGH 2 + +typedef struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO +{ + USHORT usEngineClockLow; + UCHAR ucEngineClockHigh; + + USHORT usMemoryClockLow; + UCHAR ucMemoryClockHigh; + + USHORT usVDDC; + USHORT usVDDCI; + USHORT usUnused; + + ULONG ulFlags; // ATOM_PPLIB_R600_FLAGS_* + +} ATOM_PPLIB_EVERGREEN_CLOCK_INFO; + +typedef struct _ATOM_PPLIB_SI_CLOCK_INFO +{ + USHORT usEngineClockLow; + UCHAR ucEngineClockHigh; + + USHORT usMemoryClockLow; + UCHAR ucMemoryClockHigh; + + USHORT usVDDC; + USHORT usVDDCI; + UCHAR ucPCIEGen; + UCHAR ucUnused1; + + ULONG ulFlags; // ATOM_PPLIB_SI_FLAGS_*, no flag is necessary for now + +} ATOM_PPLIB_SI_CLOCK_INFO; + +typedef struct _ATOM_PPLIB_CI_CLOCK_INFO +{ + USHORT usEngineClockLow; + UCHAR ucEngineClockHigh; + + USHORT usMemoryClockLow; + UCHAR ucMemoryClockHigh; + + UCHAR ucPCIEGen; + USHORT usPCIELane; +} ATOM_PPLIB_CI_CLOCK_INFO; + +typedef struct _ATOM_PPLIB_SUMO_CLOCK_INFO{ + USHORT usEngineClockLow; //clockfrequency & 0xFFFF. The unit is in 10khz + UCHAR ucEngineClockHigh; //clockfrequency >> 16. + UCHAR vddcIndex; //2-bit vddc index; + USHORT tdpLimit; + //please initalize to 0 + USHORT rsv1; + //please initialize to 0s + ULONG rsv2[2]; +}ATOM_PPLIB_SUMO_CLOCK_INFO; + +typedef struct _ATOM_PPLIB_STATE_V2 +{ + //number of valid dpm levels in this state; Driver uses it to calculate the whole + //size of the state: sizeof(ATOM_PPLIB_STATE_V2) + (ucNumDPMLevels - 1) * sizeof(UCHAR) + UCHAR ucNumDPMLevels; + + //a index to the array of nonClockInfos + UCHAR nonClockInfoIndex; + /** + * Driver will read the first ucNumDPMLevels in this array + */ + UCHAR clockInfoIndex[1]; +} ATOM_PPLIB_STATE_V2; + +typedef struct _StateArray{ + //how many states we have + UCHAR ucNumEntries; + + ATOM_PPLIB_STATE_V2 states[1]; +}StateArray; + + +typedef struct _ClockInfoArray{ + //how many clock levels we have + UCHAR ucNumEntries; + + //sizeof(ATOM_PPLIB_CLOCK_INFO) + UCHAR ucEntrySize; + + UCHAR clockInfo[1]; +}ClockInfoArray; + +typedef struct _NonClockInfoArray{ + + //how many non-clock levels we have. normally should be same as number of states + UCHAR ucNumEntries; + //sizeof(ATOM_PPLIB_NONCLOCK_INFO) + UCHAR ucEntrySize; + + ATOM_PPLIB_NONCLOCK_INFO nonClockInfo[1]; +}NonClockInfoArray; + +typedef struct _ATOM_PPLIB_Clock_Voltage_Dependency_Record +{ + USHORT usClockLow; + UCHAR ucClockHigh; + USHORT usVoltage; +}ATOM_PPLIB_Clock_Voltage_Dependency_Record; + +typedef struct _ATOM_PPLIB_Clock_Voltage_Dependency_Table +{ + UCHAR ucNumEntries; // Number of entries. + ATOM_PPLIB_Clock_Voltage_Dependency_Record entries[1]; // Dynamically allocate entries. +}ATOM_PPLIB_Clock_Voltage_Dependency_Table; + +typedef struct _ATOM_PPLIB_Clock_Voltage_Limit_Record +{ + USHORT usSclkLow; + UCHAR ucSclkHigh; + USHORT usMclkLow; + UCHAR ucMclkHigh; + USHORT usVddc; + USHORT usVddci; +}ATOM_PPLIB_Clock_Voltage_Limit_Record; + +typedef struct _ATOM_PPLIB_Clock_Voltage_Limit_Table +{ + UCHAR ucNumEntries; // Number of entries. + ATOM_PPLIB_Clock_Voltage_Limit_Record entries[1]; // Dynamically allocate entries. +}ATOM_PPLIB_Clock_Voltage_Limit_Table; + +union _ATOM_PPLIB_CAC_Leakage_Record +{ + struct + { + USHORT usVddc; // We use this field for the "fake" standardized VDDC for power calculations; For CI and newer, we use this as the real VDDC value. in CI we read it as StdVoltageHiSidd + ULONG ulLeakageValue; // For CI and newer we use this as the "fake" standar VDDC value. in CI we read it as StdVoltageLoSidd + + }; + struct + { + USHORT usVddc1; + USHORT usVddc2; + USHORT usVddc3; + }; +}; + +typedef union _ATOM_PPLIB_CAC_Leakage_Record ATOM_PPLIB_CAC_Leakage_Record; + +typedef struct _ATOM_PPLIB_CAC_Leakage_Table +{ + UCHAR ucNumEntries; // Number of entries. + ATOM_PPLIB_CAC_Leakage_Record entries[1]; // Dynamically allocate entries. +}ATOM_PPLIB_CAC_Leakage_Table; + +typedef struct _ATOM_PPLIB_PhaseSheddingLimits_Record +{ + USHORT usVoltage; + USHORT usSclkLow; + UCHAR ucSclkHigh; + USHORT usMclkLow; + UCHAR ucMclkHigh; +}ATOM_PPLIB_PhaseSheddingLimits_Record; + +typedef struct _ATOM_PPLIB_PhaseSheddingLimits_Table +{ + UCHAR ucNumEntries; // Number of entries. + ATOM_PPLIB_PhaseSheddingLimits_Record entries[1]; // Dynamically allocate entries. +}ATOM_PPLIB_PhaseSheddingLimits_Table; + +typedef struct _VCEClockInfo{ + USHORT usEVClkLow; + UCHAR ucEVClkHigh; + USHORT usECClkLow; + UCHAR ucECClkHigh; +}VCEClockInfo; + +typedef struct _VCEClockInfoArray{ + UCHAR ucNumEntries; + VCEClockInfo entries[1]; +}VCEClockInfoArray; + +typedef struct _ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record +{ + USHORT usVoltage; + UCHAR ucVCEClockInfoIndex; +}ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record; + +typedef struct _ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table +{ + UCHAR numEntries; + ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record entries[1]; +}ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table; + +typedef struct _ATOM_PPLIB_VCE_State_Record +{ + UCHAR ucVCEClockInfoIndex; + UCHAR ucClockInfoIndex; //highest 2 bits indicates memory p-states, lower 6bits indicates index to ClockInfoArrary +}ATOM_PPLIB_VCE_State_Record; + +typedef struct _ATOM_PPLIB_VCE_State_Table +{ + UCHAR numEntries; + ATOM_PPLIB_VCE_State_Record entries[1]; +}ATOM_PPLIB_VCE_State_Table; + + +typedef struct _ATOM_PPLIB_VCE_Table +{ + UCHAR revid; +// VCEClockInfoArray array; +// ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table limits; +// ATOM_PPLIB_VCE_State_Table states; +}ATOM_PPLIB_VCE_Table; + + +typedef struct _UVDClockInfo{ + USHORT usVClkLow; + UCHAR ucVClkHigh; + USHORT usDClkLow; + UCHAR ucDClkHigh; +}UVDClockInfo; + +typedef struct _UVDClockInfoArray{ + UCHAR ucNumEntries; + UVDClockInfo entries[1]; +}UVDClockInfoArray; + +typedef struct _ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record +{ + USHORT usVoltage; + UCHAR ucUVDClockInfoIndex; +}ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record; + +typedef struct _ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table +{ + UCHAR numEntries; + ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record entries[1]; +}ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table; + +typedef struct _ATOM_PPLIB_UVD_Table +{ + UCHAR revid; +// UVDClockInfoArray array; +// ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table limits; +}ATOM_PPLIB_UVD_Table; + +typedef struct _ATOM_PPLIB_SAMClk_Voltage_Limit_Record +{ + USHORT usVoltage; + USHORT usSAMClockLow; + UCHAR ucSAMClockHigh; +}ATOM_PPLIB_SAMClk_Voltage_Limit_Record; + +typedef struct _ATOM_PPLIB_SAMClk_Voltage_Limit_Table{ + UCHAR numEntries; + ATOM_PPLIB_SAMClk_Voltage_Limit_Record entries[1]; +}ATOM_PPLIB_SAMClk_Voltage_Limit_Table; + +typedef struct _ATOM_PPLIB_SAMU_Table +{ + UCHAR revid; + ATOM_PPLIB_SAMClk_Voltage_Limit_Table limits; +}ATOM_PPLIB_SAMU_Table; + +typedef struct _ATOM_PPLIB_ACPClk_Voltage_Limit_Record +{ + USHORT usVoltage; + USHORT usACPClockLow; + UCHAR ucACPClockHigh; +}ATOM_PPLIB_ACPClk_Voltage_Limit_Record; + +typedef struct _ATOM_PPLIB_ACPClk_Voltage_Limit_Table{ + UCHAR numEntries; + ATOM_PPLIB_ACPClk_Voltage_Limit_Record entries[1]; +}ATOM_PPLIB_ACPClk_Voltage_Limit_Table; + +typedef struct _ATOM_PPLIB_ACP_Table +{ + UCHAR revid; + ATOM_PPLIB_ACPClk_Voltage_Limit_Table limits; +}ATOM_PPLIB_ACP_Table; + +typedef struct _ATOM_PowerTune_Table{ + USHORT usTDP; + USHORT usConfigurableTDP; + USHORT usTDC; + USHORT usBatteryPowerLimit; + USHORT usSmallPowerLimit; + USHORT usLowCACLeakage; + USHORT usHighCACLeakage; +}ATOM_PowerTune_Table; + +typedef struct _ATOM_PPLIB_POWERTUNE_Table +{ + UCHAR revid; + ATOM_PowerTune_Table power_tune_table; +}ATOM_PPLIB_POWERTUNE_Table; + +typedef struct _ATOM_PPLIB_POWERTUNE_Table_V1 +{ + UCHAR revid; + ATOM_PowerTune_Table power_tune_table; + USHORT usMaximumPowerDeliveryLimit; + USHORT usReserve[7]; +} ATOM_PPLIB_POWERTUNE_Table_V1; + +#define ATOM_PPM_A_A 1 +#define ATOM_PPM_A_I 2 +typedef struct _ATOM_PPLIB_PPM_Table +{ + UCHAR ucRevId; + UCHAR ucPpmDesign; //A+I or A+A + USHORT usCpuCoreNumber; + ULONG ulPlatformTDP; + ULONG ulSmallACPlatformTDP; + ULONG ulPlatformTDC; + ULONG ulSmallACPlatformTDC; + ULONG ulApuTDP; + ULONG ulDGpuTDP; + ULONG ulDGpuUlvPower; + ULONG ulTjmax; +} ATOM_PPLIB_PPM_Table; + +#pragma pack() + +#endif diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index c9affefd79f..1544efcf1c3 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -39,7 +39,6 @@ #include "atom.h" #include <linux/firmware.h> -#include <linux/platform_device.h> #include <linux/module.h> #include "r100_reg_safe.h" @@ -143,36 +142,6 @@ void r100_wait_for_vblank(struct radeon_device *rdev, int crtc) } /** - * r100_pre_page_flip - pre-pageflip callback. - * - * @rdev: radeon_device pointer - * @crtc: crtc to prepare for pageflip on - * - * Pre-pageflip callback (r1xx-r4xx). - * Enables the pageflip irq (vblank irq). - */ -void r100_pre_page_flip(struct radeon_device *rdev, int crtc) -{ - /* enable the pflip int */ - radeon_irq_kms_pflip_irq_get(rdev, crtc); -} - -/** - * r100_post_page_flip - pos-pageflip callback. - * - * @rdev: radeon_device pointer - * @crtc: crtc to cleanup pageflip on - * - * Post-pageflip callback (r1xx-r4xx). - * Disables the pageflip irq (vblank irq). - */ -void r100_post_page_flip(struct radeon_device *rdev, int crtc) -{ - /* disable the pflip int */ - radeon_irq_kms_pflip_irq_put(rdev, crtc); -} - -/** * r100_page_flip - pageflip callback. * * @rdev: radeon_device pointer @@ -183,9 +152,8 @@ void r100_post_page_flip(struct radeon_device *rdev, int crtc) * During vblank we take the crtc lock and wait for the update_pending * bit to go high, when it does, we release the lock, and allow the * double buffered update to take place. - * Returns the current update pending status. */ -u32 r100_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) +void r100_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) { struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id]; u32 tmp = ((u32)crtc_base) | RADEON_CRTC_OFFSET__OFFSET_LOCK; @@ -207,8 +175,24 @@ u32 r100_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) tmp &= ~RADEON_CRTC_OFFSET__OFFSET_LOCK; WREG32(RADEON_CRTC_OFFSET + radeon_crtc->crtc_offset, tmp); +} + +/** + * r100_page_flip_pending - check if page flip is still pending + * + * @rdev: radeon_device pointer + * @crtc_id: crtc to check + * + * Check if the last pagefilp is still pending (r1xx-r4xx). + * Returns the current update pending status. + */ +bool r100_page_flip_pending(struct radeon_device *rdev, int crtc_id) +{ + struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id]; + /* Return current update_pending status: */ - return RREG32(RADEON_CRTC_OFFSET + radeon_crtc->crtc_offset) & RADEON_CRTC_OFFSET__GUI_TRIG_OFFSET; + return !!(RREG32(RADEON_CRTC_OFFSET + radeon_crtc->crtc_offset) & + RADEON_CRTC_OFFSET__GUI_TRIG_OFFSET); } /** @@ -698,15 +682,11 @@ void r100_pci_gart_disable(struct radeon_device *rdev) WREG32(RADEON_AIC_HI_ADDR, 0); } -int r100_pci_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr) +void r100_pci_gart_set_page(struct radeon_device *rdev, unsigned i, + uint64_t addr) { u32 *gtt = rdev->gart.ptr; - - if (i < 0 || i > rdev->gart.num_gpu_pages) { - return -EINVAL; - } gtt[i] = cpu_to_le32(lower_32_bits(addr)); - return 0; } void r100_pci_gart_fini(struct radeon_device *rdev) @@ -795,7 +775,7 @@ int r100_irq_process(struct radeon_device *rdev) wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[0])) - radeon_crtc_handle_flip(rdev, 0); + radeon_crtc_handle_vblank(rdev, 0); } if (status & RADEON_CRTC2_VBLANK_STAT) { if (rdev->irq.crtc_vblank_int[1]) { @@ -804,7 +784,7 @@ int r100_irq_process(struct radeon_device *rdev) wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[1])) - radeon_crtc_handle_flip(rdev, 1); + radeon_crtc_handle_vblank(rdev, 1); } if (status & RADEON_FP_DETECT_STAT) { queue_hotplug = true; @@ -870,13 +850,14 @@ void r100_fence_ring_emit(struct radeon_device *rdev, radeon_ring_write(ring, RADEON_SW_INT_FIRE); } -void r100_semaphore_ring_emit(struct radeon_device *rdev, +bool r100_semaphore_ring_emit(struct radeon_device *rdev, struct radeon_ring *ring, struct radeon_semaphore *semaphore, bool emit_wait) { /* Unused on older asics, since we don't have semaphores or multiple rings */ BUG(); + return false; } int r100_copy_blit(struct radeon_device *rdev, @@ -989,18 +970,11 @@ void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring) /* Load the microcode for the CP */ static int r100_cp_init_microcode(struct radeon_device *rdev) { - struct platform_device *pdev; const char *fw_name = NULL; int err; DRM_DEBUG_KMS("\n"); - pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0); - err = IS_ERR(pdev); - if (err) { - printk(KERN_ERR "radeon_cp: Failed to register firmware\n"); - return -EINVAL; - } if ((rdev->family == CHIP_R100) || (rdev->family == CHIP_RV100) || (rdev->family == CHIP_RV200) || (rdev->family == CHIP_RS100) || (rdev->family == CHIP_RS200)) { @@ -1042,8 +1016,7 @@ static int r100_cp_init_microcode(struct radeon_device *rdev) fw_name = FIRMWARE_R520; } - err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev); - platform_device_unregister(pdev); + err = request_firmware(&rdev->me_fw, fw_name, rdev->dev); if (err) { printk(KERN_ERR "radeon_cp: Failed to load firmware \"%s\"\n", fw_name); @@ -1058,6 +1031,36 @@ static int r100_cp_init_microcode(struct radeon_device *rdev) return err; } +u32 r100_gfx_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 rptr; + + if (rdev->wb.enabled) + rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]); + else + rptr = RREG32(RADEON_CP_RB_RPTR); + + return rptr; +} + +u32 r100_gfx_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 wptr; + + wptr = RREG32(RADEON_CP_RB_WPTR); + + return wptr; +} + +void r100_gfx_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + WREG32(RADEON_CP_RB_WPTR, ring->wptr); + (void)RREG32(RADEON_CP_RB_WPTR); +} + static void r100_cp_load_microcode(struct radeon_device *rdev) { const __be32 *fw_data; @@ -1106,12 +1109,11 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) } /* Align ring size */ - rb_bufsz = drm_order(ring_size / 8); + rb_bufsz = order_base_2(ring_size / 8); ring_size = (1 << (rb_bufsz + 1)) * 4; r100_cp_load_microcode(rdev); r = radeon_ring_init(rdev, ring, ring_size, RADEON_WB_CP_RPTR_OFFSET, - RADEON_CP_RB_RPTR, RADEON_CP_RB_WPTR, - 0, 0x7fffff, RADEON_CP_PACKET2); + RADEON_CP_PACKET2); if (r) { return r; } @@ -1172,7 +1174,6 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) WREG32(RADEON_CP_RB_CNTL, tmp); udelay(10); - ring->rptr = RREG32(RADEON_CP_RB_RPTR); /* Set cp mode to bus mastering & enable cp*/ WREG32(RADEON_CP_CSQ_MODE, REG_SET(RADEON_INDIRECT2_START, indirect2_start) | @@ -1254,12 +1255,12 @@ int r100_reloc_pitch_offset(struct radeon_cs_parser *p, value = radeon_get_ib_value(p, idx); tmp = value & 0x003fffff; - tmp += (((u32)reloc->lobj.gpu_offset) >> 10); + tmp += (((u32)reloc->gpu_offset) >> 10); if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= RADEON_DST_TILE_MACRO; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) { + if (reloc->tiling_flags & RADEON_TILING_MICRO) { if (reg == RADEON_SRC_PITCH_OFFSET) { DRM_ERROR("Cannot src blit from microtiled surface\n"); radeon_cs_dump_packet(p, pkt); @@ -1305,7 +1306,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p, return r; } idx_value = radeon_get_ib_value(p, idx); - ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset); + ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->gpu_offset); track->arrays[i + 0].esize = idx_value >> 8; track->arrays[i + 0].robj = reloc->robj; @@ -1317,7 +1318,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx+2] = radeon_get_ib_value(p, idx + 2) + ((u32)reloc->lobj.gpu_offset); + ib[idx+2] = radeon_get_ib_value(p, idx + 2) + ((u32)reloc->gpu_offset); track->arrays[i + 1].robj = reloc->robj; track->arrays[i + 1].esize = idx_value >> 24; track->arrays[i + 1].esize &= 0x7F; @@ -1331,7 +1332,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p, return r; } idx_value = radeon_get_ib_value(p, idx); - ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset); + ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->gpu_offset); track->arrays[i + 0].robj = reloc->robj; track->arrays[i + 0].esize = idx_value >> 8; track->arrays[i + 0].esize &= 0x7F; @@ -1443,7 +1444,7 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p) obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC); if (!obj) { DRM_ERROR("cannot find crtc %d\n", crtc_id); - return -EINVAL; + return -ENOENT; } crtc = obj_to_crtc(obj); radeon_crtc = to_radeon_crtc(crtc); @@ -1574,7 +1575,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->zb.robj = reloc->robj; track->zb.offset = idx_value; track->zb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_RB3D_COLOROFFSET: r = radeon_cs_packet_next_reloc(p, &reloc, 0); @@ -1587,7 +1588,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->cb[0].robj = reloc->robj; track->cb[0].offset = idx_value; track->cb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_PP_TXOFFSET_0: case RADEON_PP_TXOFFSET_1: @@ -1601,16 +1602,16 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= RADEON_TXO_MACRO_TILE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= RADEON_TXO_MICRO_TILE_X2; tmp = idx_value & ~(0x7 << 2); tmp |= tile_flags; - ib[idx] = tmp + ((u32)reloc->lobj.gpu_offset); + ib[idx] = tmp + ((u32)reloc->gpu_offset); } else - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -1628,7 +1629,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } track->textures[0].cube_info[i].offset = idx_value; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[0].cube_info[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -1646,7 +1647,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } track->textures[1].cube_info[i].offset = idx_value; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[1].cube_info[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -1664,7 +1665,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } track->textures[2].cube_info[i].offset = idx_value; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[2].cube_info[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -1682,9 +1683,9 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= RADEON_COLOR_TILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; tmp = idx_value & ~(0x7 << 16); @@ -1752,7 +1753,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_PP_CNTL: { @@ -1912,7 +1913,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx+1] = radeon_get_ib_value(p, idx+1) + ((u32)reloc->lobj.gpu_offset); + ib[idx+1] = radeon_get_ib_value(p, idx+1) + ((u32)reloc->gpu_offset); r = r100_cs_track_check_pkt3_indx_buffer(p, pkt, reloc->robj); if (r) { return r; @@ -1926,7 +1927,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx] = radeon_get_ib_value(p, idx) + ((u32)reloc->lobj.gpu_offset); + ib[idx] = radeon_get_ib_value(p, idx) + ((u32)reloc->gpu_offset); track->num_arrays = 1; track->vtx_size = r100_get_vtx_size(radeon_get_ib_value(p, idx + 2)); @@ -2502,11 +2503,9 @@ bool r100_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) rbbm_status = RREG32(R_000E40_RBBM_STATUS); if (!G_000E40_GUI_ACTIVE(rbbm_status)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } @@ -2862,21 +2861,28 @@ static void r100_pll_errata_after_data(struct radeon_device *rdev) uint32_t r100_pll_rreg(struct radeon_device *rdev, uint32_t reg) { + unsigned long flags; uint32_t data; + spin_lock_irqsave(&rdev->pll_idx_lock, flags); WREG8(RADEON_CLOCK_CNTL_INDEX, reg & 0x3f); r100_pll_errata_after_index(rdev); data = RREG32(RADEON_CLOCK_CNTL_DATA); r100_pll_errata_after_data(rdev); + spin_unlock_irqrestore(&rdev->pll_idx_lock, flags); return data; } void r100_pll_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->pll_idx_lock, flags); WREG8(RADEON_CLOCK_CNTL_INDEX, ((reg & 0x3f) | RADEON_PLL_WR_EN)); r100_pll_errata_after_index(rdev); WREG32(RADEON_CLOCK_CNTL_DATA, v); r100_pll_errata_after_data(rdev); + spin_unlock_irqrestore(&rdev->pll_idx_lock, flags); } static void r100_set_safe_registers(struct radeon_device *rdev) @@ -2935,9 +2941,11 @@ static int r100_debugfs_cp_ring_info(struct seq_file *m, void *data) seq_printf(m, "CP_RB_RPTR 0x%08x\n", rdp); seq_printf(m, "%u free dwords in ring\n", ring->ring_free_dw); seq_printf(m, "%u dwords in ring\n", count); - for (j = 0; j <= count; j++) { - i = (rdp + j) & ring->ptr_mask; - seq_printf(m, "r[%04d]=0x%08x\n", i, ring->ring[i]); + if (ring->ready) { + for (j = 0; j <= count; j++) { + i = (rdp + j) & ring->ptr_mask; + seq_printf(m, "r[%04d]=0x%08x\n", i, ring->ring[i]); + } } return 0; } @@ -3193,12 +3201,12 @@ void r100_bandwidth_update(struct radeon_device *rdev) if (rdev->mode_info.crtcs[0]->base.enabled) { mode1 = &rdev->mode_info.crtcs[0]->base.mode; - pixel_bytes1 = rdev->mode_info.crtcs[0]->base.fb->bits_per_pixel / 8; + pixel_bytes1 = rdev->mode_info.crtcs[0]->base.primary->fb->bits_per_pixel / 8; } if (!(rdev->flags & RADEON_SINGLE_CRTC)) { if (rdev->mode_info.crtcs[1]->base.enabled) { mode2 = &rdev->mode_info.crtcs[1]->base.mode; - pixel_bytes2 = rdev->mode_info.crtcs[1]->base.fb->bits_per_pixel / 8; + pixel_bytes2 = rdev->mode_info.crtcs[1]->base.primary->fb->bits_per_pixel / 8; } } @@ -3922,6 +3930,7 @@ int r100_resume(struct radeon_device *rdev) int r100_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); r100_irq_disable(rdev); @@ -3932,6 +3941,7 @@ int r100_suspend(struct radeon_device *rdev) void r100_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); r100_cp_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); @@ -4038,6 +4048,9 @@ int r100_init(struct radeon_device *rdev) } r100_set_safe_registers(rdev); + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->accel_working = true; r = r100_startup(rdev); if (r) { diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c index b3807edb193..58f0473aa73 100644 --- a/drivers/gpu/drm/radeon/r200.c +++ b/drivers/gpu/drm/radeon/r200.c @@ -185,7 +185,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, track->zb.robj = reloc->robj; track->zb.offset = idx_value; track->zb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_RB3D_COLOROFFSET: r = radeon_cs_packet_next_reloc(p, &reloc, 0); @@ -198,7 +198,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, track->cb[0].robj = reloc->robj; track->cb[0].offset = idx_value; track->cb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case R200_PP_TXOFFSET_0: case R200_PP_TXOFFSET_1: @@ -215,16 +215,16 @@ int r200_packet0_check(struct radeon_cs_parser *p, return r; } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= R200_TXO_MACRO_TILE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= R200_TXO_MICRO_TILE; tmp = idx_value & ~(0x7 << 2); tmp |= tile_flags; - ib[idx] = tmp + ((u32)reloc->lobj.gpu_offset); + ib[idx] = tmp + ((u32)reloc->gpu_offset); } else - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -268,7 +268,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, return r; } track->textures[i].cube_info[face - 1].offset = idx_value; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[i].cube_info[face - 1].robj = reloc->robj; track->tex_dirty = true; break; @@ -287,9 +287,9 @@ int r200_packet0_check(struct radeon_cs_parser *p, } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= RADEON_COLOR_TILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; tmp = idx_value & ~(0x7 << 16); @@ -362,7 +362,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_PP_CNTL: { diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index b9b776f1e58..3c21d77a483 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -72,13 +72,11 @@ void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev) #define R300_PTE_WRITEABLE (1 << 2) #define R300_PTE_READABLE (1 << 3) -int rv370_pcie_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr) +void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i, + uint64_t addr) { void __iomem *ptr = rdev->gart.ptr; - if (i < 0 || i > rdev->gart.num_gpu_pages) { - return -EINVAL; - } addr = (lower_32_bits(addr) >> 8) | ((upper_32_bits(addr) & 0xff) << 24) | R300_PTE_WRITEABLE | R300_PTE_READABLE; @@ -86,7 +84,6 @@ int rv370_pcie_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr) * on powerpc without HW swappers, it'll get swapped on way * into VRAM - so no need for cpu_to_le32 on VRAM tables */ writel(addr, ((void __iomem *)ptr) + (i * 4)); - return 0; } int rv370_pcie_gart_init(struct radeon_device *rdev) @@ -640,7 +637,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, track->cb[i].robj = reloc->robj; track->cb[i].offset = idx_value; track->cb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case R300_ZB_DEPTHOFFSET: r = radeon_cs_packet_next_reloc(p, &reloc, 0); @@ -653,7 +650,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, track->zb.robj = reloc->robj; track->zb.offset = idx_value; track->zb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case R300_TX_OFFSET_0: case R300_TX_OFFSET_0+4: @@ -682,16 +679,16 @@ static int r300_packet0_check(struct radeon_cs_parser *p, if (p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) { ib[idx] = (idx_value & 31) | /* keep the 1st 5 bits */ - ((idx_value & ~31) + (u32)reloc->lobj.gpu_offset); + ((idx_value & ~31) + (u32)reloc->gpu_offset); } else { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= R300_TXO_MACRO_TILE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= R300_TXO_MICRO_TILE; - else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE) + else if (reloc->tiling_flags & RADEON_TILING_MICRO_SQUARE) tile_flags |= R300_TXO_MICRO_TILE_SQUARE; - tmp = idx_value + ((u32)reloc->lobj.gpu_offset); + tmp = idx_value + ((u32)reloc->gpu_offset); tmp |= tile_flags; ib[idx] = tmp; } @@ -753,11 +750,11 @@ static int r300_packet0_check(struct radeon_cs_parser *p, return r; } - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= R300_COLOR_TILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= R300_COLOR_MICROTILE_ENABLE; - else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE) + else if (reloc->tiling_flags & RADEON_TILING_MICRO_SQUARE) tile_flags |= R300_COLOR_MICROTILE_SQUARE_ENABLE; tmp = idx_value & ~(0x7 << 16); @@ -838,11 +835,11 @@ static int r300_packet0_check(struct radeon_cs_parser *p, return r; } - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= R300_DEPTHMACROTILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= R300_DEPTHMICROTILE_TILED; - else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE) + else if (reloc->tiling_flags & RADEON_TILING_MICRO_SQUARE) tile_flags |= R300_DEPTHMICROTILE_TILED_SQUARE; tmp = idx_value & ~(0x7 << 16); @@ -1052,7 +1049,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case 0x4e0c: /* RB3D_COLOR_CHANNEL_MASK */ @@ -1097,7 +1094,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, track->aa.robj = reloc->robj; track->aa.offset = idx_value; track->aa_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case R300_RB3D_AARESOLVE_PITCH: track->aa.pitch = idx_value & 0x3FFE; @@ -1162,7 +1159,7 @@ static int r300_packet3_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset); + ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->gpu_offset); r = r100_cs_track_check_pkt3_indx_buffer(p, pkt, reloc->robj); if (r) { return r; @@ -1440,6 +1437,7 @@ int r300_resume(struct radeon_device *rdev) int r300_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); r100_irq_disable(rdev); @@ -1452,6 +1450,7 @@ int r300_suspend(struct radeon_device *rdev) void r300_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); r100_cp_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); @@ -1538,10 +1537,13 @@ int r300_init(struct radeon_device *rdev) } r300_set_reg_safe(rdev); + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->accel_working = true; r = r300_startup(rdev); if (r) { - /* Somethings want wront with the accel init stop accel */ + /* Something went wrong with the accel init, so stop accel */ dev_err(rdev->dev, "Disabling GPU acceleration\n"); r100_cp_fini(rdev); radeon_wb_fini(rdev); diff --git a/drivers/gpu/drm/radeon/r300_cmdbuf.c b/drivers/gpu/drm/radeon/r300_cmdbuf.c index 60170ea5e3a..84b1d5367a1 100644 --- a/drivers/gpu/drm/radeon/r300_cmdbuf.c +++ b/drivers/gpu/drm/radeon/r300_cmdbuf.c @@ -75,7 +75,7 @@ static int r300_emit_cliprects(drm_radeon_private_t *dev_priv, OUT_RING(CP_PACKET0(R300_RE_CLIPRECT_TL_0, nr * 2 - 1)); for (i = 0; i < nr; ++i) { - if (DRM_COPY_FROM_USER + if (copy_from_user (&box, &cmdbuf->boxes[n + i], sizeof(box))) { DRM_ERROR("copy cliprect faulted\n"); return -EFAULT; @@ -928,12 +928,12 @@ static int r300_scratch(drm_radeon_private_t *dev_priv, buf_idx = drm_buffer_pointer_to_dword(cmdbuf->buffer, 0); *buf_idx *= 2; /* 8 bytes per buf */ - if (DRM_COPY_TO_USER(ref_age_base + *buf_idx, + if (copy_to_user(ref_age_base + *buf_idx, &dev_priv->scratch_ages[header.scratch.reg], sizeof(u32))) return -EINVAL; - if (DRM_COPY_FROM_USER(&h_pending, + if (copy_from_user(&h_pending, ref_age_base + *buf_idx + 1, sizeof(u32))) return -EINVAL; @@ -943,7 +943,7 @@ static int r300_scratch(drm_radeon_private_t *dev_priv, h_pending--; - if (DRM_COPY_TO_USER(ref_age_base + *buf_idx + 1, + if (copy_to_user(ref_age_base + *buf_idx + 1, &h_pending, sizeof(u32))) return -EINVAL; diff --git a/drivers/gpu/drm/radeon/r420.c b/drivers/gpu/drm/radeon/r420.c index 4e796ecf9ea..802b19220a2 100644 --- a/drivers/gpu/drm/radeon/r420.c +++ b/drivers/gpu/drm/radeon/r420.c @@ -160,18 +160,25 @@ void r420_pipes_init(struct radeon_device *rdev) u32 r420_mc_rreg(struct radeon_device *rdev, u32 reg) { + unsigned long flags; u32 r; + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(R_0001F8_MC_IND_INDEX, S_0001F8_MC_IND_ADDR(reg)); r = RREG32(R_0001FC_MC_IND_DATA); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); return r; } void r420_mc_wreg(struct radeon_device *rdev, u32 reg, u32 v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(R_0001F8_MC_IND_INDEX, S_0001F8_MC_IND_ADDR(reg) | S_0001F8_MC_IND_WR_EN(1)); WREG32(R_0001FC_MC_IND_DATA, v); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); } static void r420_debugfs(struct radeon_device *rdev) @@ -328,6 +335,7 @@ int r420_resume(struct radeon_device *rdev) int r420_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r420_cp_errata_fini(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); @@ -341,6 +349,7 @@ int r420_suspend(struct radeon_device *rdev) void r420_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); r100_cp_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); @@ -437,6 +446,9 @@ int r420_init(struct radeon_device *rdev) } r420_set_reg_safe(rdev); + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->accel_working = true; r = r420_startup(rdev); if (r) { diff --git a/drivers/gpu/drm/radeon/r500_reg.h b/drivers/gpu/drm/radeon/r500_reg.h index 1dd0d32993d..136b7bc7cd2 100644 --- a/drivers/gpu/drm/radeon/r500_reg.h +++ b/drivers/gpu/drm/radeon/r500_reg.h @@ -402,6 +402,7 @@ * block and vice versa. This applies to GRPH, CUR, etc. */ #define AVIVO_D1GRPH_LUT_SEL 0x6108 +# define AVIVO_LUT_10BIT_BYPASS_EN (1 << 8) #define AVIVO_D1GRPH_PRIMARY_SURFACE_ADDRESS 0x6110 #define R700_D1GRPH_PRIMARY_SURFACE_ADDRESS_HIGH 0x6914 #define R700_D2GRPH_PRIMARY_SURFACE_ADDRESS_HIGH 0x6114 diff --git a/drivers/gpu/drm/radeon/r520.c b/drivers/gpu/drm/radeon/r520.c index e1aece73b37..98d6053c36c 100644 --- a/drivers/gpu/drm/radeon/r520.c +++ b/drivers/gpu/drm/radeon/r520.c @@ -312,6 +312,9 @@ int r520_init(struct radeon_device *rdev) return r; rv515_set_safe_registers(rdev); + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->accel_working = true; r = r520_startup(rdev); if (r) { diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c index 2d3655f7f41..3c69f58e46e 100644 --- a/drivers/gpu/drm/radeon/r600.c +++ b/drivers/gpu/drm/radeon/r600.c @@ -28,7 +28,6 @@ #include <linux/slab.h> #include <linux/seq_file.h> #include <linux/firmware.h> -#include <linux/platform_device.h> #include <linux/module.h> #include <drm/drmP.h> #include <drm/radeon_drm.h> @@ -106,6 +105,7 @@ void r600_fini(struct radeon_device *rdev); void r600_irq_disable(struct radeon_device *rdev); static void r600_pcie_gen2_enable(struct radeon_device *rdev); extern int evergreen_rlc_resume(struct radeon_device *rdev); +extern void rv770_set_clk_bypass_mode(struct radeon_device *rdev); /** * r600_get_xclk - get the xclk @@ -120,6 +120,64 @@ u32 r600_get_xclk(struct radeon_device *rdev) return rdev->clock.spll.reference_freq; } +int r600_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk) +{ + return 0; +} + +void dce3_program_fmt(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); + struct drm_connector *connector = radeon_get_connector_for_encoder(encoder); + int bpc = 0; + u32 tmp = 0; + enum radeon_connector_dither dither = RADEON_FMT_DITHER_DISABLE; + + if (connector) { + struct radeon_connector *radeon_connector = to_radeon_connector(connector); + bpc = radeon_get_monitor_bpc(connector); + dither = radeon_connector->dither; + } + + /* LVDS FMT is set up by atom */ + if (radeon_encoder->devices & ATOM_DEVICE_LCD_SUPPORT) + return; + + /* not needed for analog */ + if ((radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1) || + (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2)) + return; + + if (bpc == 0) + return; + + switch (bpc) { + case 6: + if (dither == RADEON_FMT_DITHER_ENABLE) + /* XXX sort out optimal dither settings */ + tmp |= FMT_SPATIAL_DITHER_EN; + else + tmp |= FMT_TRUNCATE_EN; + break; + case 8: + if (dither == RADEON_FMT_DITHER_ENABLE) + /* XXX sort out optimal dither settings */ + tmp |= (FMT_SPATIAL_DITHER_EN | FMT_SPATIAL_DITHER_DEPTH); + else + tmp |= (FMT_TRUNCATE_EN | FMT_TRUNCATE_DEPTH); + break; + case 10: + default: + /* not needed */ + break; + } + + WREG32(FMT_BIT_DEPTH_CONTROL + radeon_crtc->crtc_offset, tmp); +} + /* get temperature in millidegrees */ int rv6xx_get_temp(struct radeon_device *rdev) { @@ -1046,20 +1104,27 @@ int r600_mc_wait_for_idle(struct radeon_device *rdev) uint32_t rs780_mc_rreg(struct radeon_device *rdev, uint32_t reg) { + unsigned long flags; uint32_t r; + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(R_0028F8_MC_INDEX, S_0028F8_MC_IND_ADDR(reg)); r = RREG32(R_0028FC_MC_DATA); WREG32(R_0028F8_MC_INDEX, ~C_0028F8_MC_IND_ADDR); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); return r; } void rs780_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(R_0028F8_MC_INDEX, S_0028F8_MC_IND_ADDR(reg) | S_0028F8_MC_IND_WR_EN(1)); WREG32(R_0028FC_MC_DATA, v); WREG32(R_0028F8_MC_INDEX, 0x7F); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); } static void r600_mc_program(struct radeon_device *rdev) @@ -1375,7 +1440,7 @@ static bool r600_is_display_hung(struct radeon_device *rdev) return true; } -static u32 r600_gpu_check_soft_reset(struct radeon_device *rdev) +u32 r600_gpu_check_soft_reset(struct radeon_device *rdev) { u32 reset_mask = 0; u32 tmp; @@ -1580,6 +1645,67 @@ static void r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask) r600_print_gpu_status_regs(rdev); } +static void r600_gpu_pci_config_reset(struct radeon_device *rdev) +{ + struct rv515_mc_save save; + u32 tmp, i; + + dev_info(rdev->dev, "GPU pci config reset\n"); + + /* disable dpm? */ + + /* Disable CP parsing/prefetching */ + if (rdev->family >= CHIP_RV770) + WREG32(R_0086D8_CP_ME_CNTL, S_0086D8_CP_ME_HALT(1) | S_0086D8_CP_PFP_HALT(1)); + else + WREG32(R_0086D8_CP_ME_CNTL, S_0086D8_CP_ME_HALT(1)); + + /* disable the RLC */ + WREG32(RLC_CNTL, 0); + + /* Disable DMA */ + tmp = RREG32(DMA_RB_CNTL); + tmp &= ~DMA_RB_ENABLE; + WREG32(DMA_RB_CNTL, tmp); + + mdelay(50); + + /* set mclk/sclk to bypass */ + if (rdev->family >= CHIP_RV770) + rv770_set_clk_bypass_mode(rdev); + /* disable BM */ + pci_clear_master(rdev->pdev); + /* disable mem access */ + rv515_mc_stop(rdev, &save); + if (r600_mc_wait_for_idle(rdev)) { + dev_warn(rdev->dev, "Wait for MC idle timedout !\n"); + } + + /* BIF reset workaround. Not sure if this is needed on 6xx */ + tmp = RREG32(BUS_CNTL); + tmp |= VGA_COHE_SPEC_TIMER_DIS; + WREG32(BUS_CNTL, tmp); + + tmp = RREG32(BIF_SCRATCH0); + + /* reset */ + radeon_pci_config_reset(rdev); + mdelay(1); + + /* BIF reset workaround. Not sure if this is needed on 6xx */ + tmp = SOFT_RESET_BIF; + WREG32(SRBM_SOFT_RESET, tmp); + mdelay(1); + WREG32(SRBM_SOFT_RESET, 0); + + /* wait for asic to come out of reset */ + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(CONFIG_MEMSIZE) != 0xffffffff) + break; + udelay(1); + } +} + int r600_asic_reset(struct radeon_device *rdev) { u32 reset_mask; @@ -1589,10 +1715,17 @@ int r600_asic_reset(struct radeon_device *rdev) if (reset_mask) r600_set_bios_scratch_engine_hung(rdev, true); + /* try soft reset */ r600_gpu_soft_reset(rdev, reset_mask); reset_mask = r600_gpu_check_soft_reset(rdev); + /* try pci config reset */ + if (reset_mask && radeon_hard_reset) + r600_gpu_pci_config_reset(rdev); + + reset_mask = r600_gpu_check_soft_reset(rdev); + if (!reset_mask) r600_set_bios_scratch_engine_hung(rdev, false); @@ -1615,33 +1748,9 @@ bool r600_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); - return radeon_ring_test_lockup(rdev, ring); -} - -/** - * r600_dma_is_lockup - Check if the DMA engine is locked up - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Check if the async DMA engine is locked up. - * Returns true if the engine appears to be locked up, false if not. - */ -bool r600_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) -{ - u32 reset_mask = r600_gpu_check_soft_reset(rdev); - - if (!(reset_mask & RADEON_RESET_DMA)) { - radeon_ring_lockup_update(ring); - return false; - } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } @@ -1849,6 +1958,9 @@ static void r600_gpu_init(struct radeon_device *rdev) if (tmp < rdev->config.r600.max_simds) { rdev->config.r600.max_simds = tmp; } + tmp = rdev->config.r600.max_simds - + r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R6XX_MAX_SIMDS_MASK); + rdev->config.r600.active_simds = tmp; disabled_rb_mask = (RREG32(CC_RB_BACKEND_DISABLE) >> 16) & R6XX_MAX_BACKENDS_MASK; tmp = (tiling_config & PIPE_TILING__MASK) >> PIPE_TILING__SHIFT; @@ -2115,20 +2227,27 @@ static void r600_gpu_init(struct radeon_device *rdev) */ u32 r600_pciep_rreg(struct radeon_device *rdev, u32 reg) { + unsigned long flags; u32 r; + spin_lock_irqsave(&rdev->pciep_idx_lock, flags); WREG32(PCIE_PORT_INDEX, ((reg) & 0xff)); (void)RREG32(PCIE_PORT_INDEX); r = RREG32(PCIE_PORT_DATA); + spin_unlock_irqrestore(&rdev->pciep_idx_lock, flags); return r; } void r600_pciep_wreg(struct radeon_device *rdev, u32 reg, u32 v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->pciep_idx_lock, flags); WREG32(PCIE_PORT_INDEX, ((reg) & 0xff)); (void)RREG32(PCIE_PORT_INDEX); WREG32(PCIE_PORT_DATA, (v)); (void)RREG32(PCIE_PORT_DATA); + spin_unlock_irqrestore(&rdev->pciep_idx_lock, flags); } /* @@ -2136,7 +2255,8 @@ void r600_pciep_wreg(struct radeon_device *rdev, u32 reg, u32 v) */ void r600_cp_stop(struct radeon_device *rdev) { - radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); WREG32(R_0086D8_CP_ME_CNTL, S_0086D8_CP_ME_HALT(1)); WREG32(SCRATCH_UMSK, 0); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; @@ -2144,7 +2264,6 @@ void r600_cp_stop(struct radeon_device *rdev) int r600_init_microcode(struct radeon_device *rdev) { - struct platform_device *pdev; const char *chip_name; const char *rlc_chip_name; const char *smc_chip_name = "RV770"; @@ -2154,13 +2273,6 @@ int r600_init_microcode(struct radeon_device *rdev) DRM_DEBUG("\n"); - pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0); - err = IS_ERR(pdev); - if (err) { - printk(KERN_ERR "radeon_cp: Failed to register firmware\n"); - return -EINVAL; - } - switch (rdev->family) { case CHIP_R600: chip_name = "R600"; @@ -2272,7 +2384,7 @@ int r600_init_microcode(struct radeon_device *rdev) DRM_INFO("Loading %s Microcode\n", chip_name); snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name); - err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->pfp_fw->size != pfp_req_size) { @@ -2284,7 +2396,7 @@ int r600_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name); - err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->me_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->me_fw->size != me_req_size) { @@ -2295,7 +2407,7 @@ int r600_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name); - err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->rlc_fw->size != rlc_req_size) { @@ -2307,10 +2419,15 @@ int r600_init_microcode(struct radeon_device *rdev) if ((rdev->family >= CHIP_RV770) && (rdev->family <= CHIP_HEMLOCK)) { snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", smc_chip_name); - err = request_firmware(&rdev->smc_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->smc_fw->size != smc_req_size) { + err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev); + if (err) { + printk(KERN_ERR + "smc: error loading firmware \"%s\"\n", + fw_name); + release_firmware(rdev->smc_fw); + rdev->smc_fw = NULL; + err = 0; + } else if (rdev->smc_fw->size != smc_req_size) { printk(KERN_ERR "smc: Bogus length %zu in firmware \"%s\"\n", rdev->smc_fw->size, fw_name); @@ -2319,8 +2436,6 @@ int r600_init_microcode(struct radeon_device *rdev) } out: - platform_device_unregister(pdev); - if (err) { if (err != -EINVAL) printk(KERN_ERR @@ -2338,6 +2453,36 @@ out: return err; } +u32 r600_gfx_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 rptr; + + if (rdev->wb.enabled) + rptr = rdev->wb.wb[ring->rptr_offs/4]; + else + rptr = RREG32(R600_CP_RB_RPTR); + + return rptr; +} + +u32 r600_gfx_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 wptr; + + wptr = RREG32(R600_CP_RB_WPTR); + + return wptr; +} + +void r600_gfx_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + WREG32(R600_CP_RB_WPTR, ring->wptr); + (void)RREG32(R600_CP_RB_WPTR); +} + static int r600_cp_load_microcode(struct radeon_device *rdev) { const __be32 *fw_data; @@ -2424,8 +2569,8 @@ int r600_cp_resume(struct radeon_device *rdev) WREG32(GRBM_SOFT_RESET, 0); /* Set ring buffer size */ - rb_bufsz = drm_order(ring->ring_size / 8); - tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; + rb_bufsz = order_base_2(ring->ring_size / 8); + tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif @@ -2460,8 +2605,6 @@ int r600_cp_resume(struct radeon_device *rdev) WREG32(CP_RB_BASE, ring->gpu_addr >> 8); WREG32(CP_DEBUG, (1 << 27) | (1 << 28)); - ring->rptr = RREG32(CP_RB_RPTR); - r600_cp_start(rdev); ring->ready = true; r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring); @@ -2469,6 +2612,10 @@ int r600_cp_resume(struct radeon_device *rdev) ring->ready = false; return r; } + + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); + return 0; } @@ -2478,7 +2625,7 @@ void r600_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsign int r; /* Align ring size */ - rb_bufsz = drm_order(ring_size / 8); + rb_bufsz = order_base_2(ring_size / 8); ring_size = (1 << (rb_bufsz + 1)) * 4; ring->ring_size = ring_size; ring->align_mask = 16 - 1; @@ -2501,327 +2648,6 @@ void r600_cp_fini(struct radeon_device *rdev) } /* - * DMA - * Starting with R600, the GPU has an asynchronous - * DMA engine. The programming model is very similar - * to the 3D engine (ring buffer, IBs, etc.), but the - * DMA controller has it's own packet format that is - * different form the PM4 format used by the 3D engine. - * It supports copying data, writing embedded data, - * solid fills, and a number of other things. It also - * has support for tiling/detiling of buffers. - */ -/** - * r600_dma_stop - stop the async dma engine - * - * @rdev: radeon_device pointer - * - * Stop the async dma engine (r6xx-evergreen). - */ -void r600_dma_stop(struct radeon_device *rdev) -{ - u32 rb_cntl = RREG32(DMA_RB_CNTL); - - radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); - - rb_cntl &= ~DMA_RB_ENABLE; - WREG32(DMA_RB_CNTL, rb_cntl); - - rdev->ring[R600_RING_TYPE_DMA_INDEX].ready = false; -} - -/** - * r600_dma_resume - setup and start the async dma engine - * - * @rdev: radeon_device pointer - * - * Set up the DMA ring buffer and enable it. (r6xx-evergreen). - * Returns 0 for success, error for failure. - */ -int r600_dma_resume(struct radeon_device *rdev) -{ - struct radeon_ring *ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; - u32 rb_cntl, dma_cntl, ib_cntl; - u32 rb_bufsz; - int r; - - /* Reset dma */ - if (rdev->family >= CHIP_RV770) - WREG32(SRBM_SOFT_RESET, RV770_SOFT_RESET_DMA); - else - WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA); - RREG32(SRBM_SOFT_RESET); - udelay(50); - WREG32(SRBM_SOFT_RESET, 0); - - WREG32(DMA_SEM_INCOMPLETE_TIMER_CNTL, 0); - WREG32(DMA_SEM_WAIT_FAIL_TIMER_CNTL, 0); - - /* Set ring buffer size in dwords */ - rb_bufsz = drm_order(ring->ring_size / 4); - rb_cntl = rb_bufsz << 1; -#ifdef __BIG_ENDIAN - rb_cntl |= DMA_RB_SWAP_ENABLE | DMA_RPTR_WRITEBACK_SWAP_ENABLE; -#endif - WREG32(DMA_RB_CNTL, rb_cntl); - - /* Initialize the ring buffer's read and write pointers */ - WREG32(DMA_RB_RPTR, 0); - WREG32(DMA_RB_WPTR, 0); - - /* set the wb address whether it's enabled or not */ - WREG32(DMA_RB_RPTR_ADDR_HI, - upper_32_bits(rdev->wb.gpu_addr + R600_WB_DMA_RPTR_OFFSET) & 0xFF); - WREG32(DMA_RB_RPTR_ADDR_LO, - ((rdev->wb.gpu_addr + R600_WB_DMA_RPTR_OFFSET) & 0xFFFFFFFC)); - - if (rdev->wb.enabled) - rb_cntl |= DMA_RPTR_WRITEBACK_ENABLE; - - WREG32(DMA_RB_BASE, ring->gpu_addr >> 8); - - /* enable DMA IBs */ - ib_cntl = DMA_IB_ENABLE; -#ifdef __BIG_ENDIAN - ib_cntl |= DMA_IB_SWAP_ENABLE; -#endif - WREG32(DMA_IB_CNTL, ib_cntl); - - dma_cntl = RREG32(DMA_CNTL); - dma_cntl &= ~CTXEMPTY_INT_ENABLE; - WREG32(DMA_CNTL, dma_cntl); - - if (rdev->family >= CHIP_RV770) - WREG32(DMA_MODE, 1); - - ring->wptr = 0; - WREG32(DMA_RB_WPTR, ring->wptr << 2); - - ring->rptr = RREG32(DMA_RB_RPTR) >> 2; - - WREG32(DMA_RB_CNTL, rb_cntl | DMA_RB_ENABLE); - - ring->ready = true; - - r = radeon_ring_test(rdev, R600_RING_TYPE_DMA_INDEX, ring); - if (r) { - ring->ready = false; - return r; - } - - radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); - - return 0; -} - -/** - * r600_dma_fini - tear down the async dma engine - * - * @rdev: radeon_device pointer - * - * Stop the async dma engine and free the ring (r6xx-evergreen). - */ -void r600_dma_fini(struct radeon_device *rdev) -{ - r600_dma_stop(rdev); - radeon_ring_fini(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX]); -} - -/* - * UVD - */ -int r600_uvd_rbc_start(struct radeon_device *rdev) -{ - struct radeon_ring *ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; - uint64_t rptr_addr; - uint32_t rb_bufsz, tmp; - int r; - - rptr_addr = rdev->wb.gpu_addr + R600_WB_UVD_RPTR_OFFSET; - - if (upper_32_bits(rptr_addr) != upper_32_bits(ring->gpu_addr)) { - DRM_ERROR("UVD ring and rptr not in the same 4GB segment!\n"); - return -EINVAL; - } - - /* force RBC into idle state */ - WREG32(UVD_RBC_RB_CNTL, 0x11010101); - - /* Set the write pointer delay */ - WREG32(UVD_RBC_RB_WPTR_CNTL, 0); - - /* set the wb address */ - WREG32(UVD_RBC_RB_RPTR_ADDR, rptr_addr >> 2); - - /* programm the 4GB memory segment for rptr and ring buffer */ - WREG32(UVD_LMI_EXT40_ADDR, upper_32_bits(rptr_addr) | - (0x7 << 16) | (0x1 << 31)); - - /* Initialize the ring buffer's read and write pointers */ - WREG32(UVD_RBC_RB_RPTR, 0x0); - - ring->wptr = ring->rptr = RREG32(UVD_RBC_RB_RPTR); - WREG32(UVD_RBC_RB_WPTR, ring->wptr); - - /* set the ring address */ - WREG32(UVD_RBC_RB_BASE, ring->gpu_addr); - - /* Set ring buffer size */ - rb_bufsz = drm_order(ring->ring_size); - rb_bufsz = (0x1 << 8) | rb_bufsz; - WREG32(UVD_RBC_RB_CNTL, rb_bufsz); - - ring->ready = true; - r = radeon_ring_test(rdev, R600_RING_TYPE_UVD_INDEX, ring); - if (r) { - ring->ready = false; - return r; - } - - r = radeon_ring_lock(rdev, ring, 10); - if (r) { - DRM_ERROR("radeon: ring failed to lock UVD ring (%d).\n", r); - return r; - } - - tmp = PACKET0(UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL, 0); - radeon_ring_write(ring, tmp); - radeon_ring_write(ring, 0xFFFFF); - - tmp = PACKET0(UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL, 0); - radeon_ring_write(ring, tmp); - radeon_ring_write(ring, 0xFFFFF); - - tmp = PACKET0(UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL, 0); - radeon_ring_write(ring, tmp); - radeon_ring_write(ring, 0xFFFFF); - - /* Clear timeout status bits */ - radeon_ring_write(ring, PACKET0(UVD_SEMA_TIMEOUT_STATUS, 0)); - radeon_ring_write(ring, 0x8); - - radeon_ring_write(ring, PACKET0(UVD_SEMA_CNTL, 0)); - radeon_ring_write(ring, 3); - - radeon_ring_unlock_commit(rdev, ring); - - return 0; -} - -void r600_uvd_rbc_stop(struct radeon_device *rdev) -{ - struct radeon_ring *ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; - - /* force RBC into idle state */ - WREG32(UVD_RBC_RB_CNTL, 0x11010101); - ring->ready = false; -} - -int r600_uvd_init(struct radeon_device *rdev) -{ - int i, j, r; - /* disable byte swapping */ - u32 lmi_swap_cntl = 0; - u32 mp_swap_cntl = 0; - - /* raise clocks while booting up the VCPU */ - radeon_set_uvd_clocks(rdev, 53300, 40000); - - /* disable clock gating */ - WREG32(UVD_CGC_GATE, 0); - - /* disable interupt */ - WREG32_P(UVD_MASTINT_EN, 0, ~(1 << 1)); - - /* put LMI, VCPU, RBC etc... into reset */ - WREG32(UVD_SOFT_RESET, LMI_SOFT_RESET | VCPU_SOFT_RESET | - LBSI_SOFT_RESET | RBC_SOFT_RESET | CSM_SOFT_RESET | - CXW_SOFT_RESET | TAP_SOFT_RESET | LMI_UMC_SOFT_RESET); - mdelay(5); - - /* take UVD block out of reset */ - WREG32_P(SRBM_SOFT_RESET, 0, ~SOFT_RESET_UVD); - mdelay(5); - - /* initialize UVD memory controller */ - WREG32(UVD_LMI_CTRL, 0x40 | (1 << 8) | (1 << 13) | - (1 << 21) | (1 << 9) | (1 << 20)); - -#ifdef __BIG_ENDIAN - /* swap (8 in 32) RB and IB */ - lmi_swap_cntl = 0xa; - mp_swap_cntl = 0; -#endif - WREG32(UVD_LMI_SWAP_CNTL, lmi_swap_cntl); - WREG32(UVD_MP_SWAP_CNTL, mp_swap_cntl); - - WREG32(UVD_MPC_SET_MUXA0, 0x40c2040); - WREG32(UVD_MPC_SET_MUXA1, 0x0); - WREG32(UVD_MPC_SET_MUXB0, 0x40c2040); - WREG32(UVD_MPC_SET_MUXB1, 0x0); - WREG32(UVD_MPC_SET_ALU, 0); - WREG32(UVD_MPC_SET_MUX, 0x88); - - /* Stall UMC */ - WREG32_P(UVD_LMI_CTRL2, 1 << 8, ~(1 << 8)); - WREG32_P(UVD_RB_ARB_CTRL, 1 << 3, ~(1 << 3)); - - /* take all subblocks out of reset, except VCPU */ - WREG32(UVD_SOFT_RESET, VCPU_SOFT_RESET); - mdelay(5); - - /* enable VCPU clock */ - WREG32(UVD_VCPU_CNTL, 1 << 9); - - /* enable UMC */ - WREG32_P(UVD_LMI_CTRL2, 0, ~(1 << 8)); - - /* boot up the VCPU */ - WREG32(UVD_SOFT_RESET, 0); - mdelay(10); - - WREG32_P(UVD_RB_ARB_CTRL, 0, ~(1 << 3)); - - for (i = 0; i < 10; ++i) { - uint32_t status; - for (j = 0; j < 100; ++j) { - status = RREG32(UVD_STATUS); - if (status & 2) - break; - mdelay(10); - } - r = 0; - if (status & 2) - break; - - DRM_ERROR("UVD not responding, trying to reset the VCPU!!!\n"); - WREG32_P(UVD_SOFT_RESET, VCPU_SOFT_RESET, ~VCPU_SOFT_RESET); - mdelay(10); - WREG32_P(UVD_SOFT_RESET, 0, ~VCPU_SOFT_RESET); - mdelay(10); - r = -1; - } - - if (r) { - DRM_ERROR("UVD not responding, giving up!!!\n"); - radeon_set_uvd_clocks(rdev, 0, 0); - return r; - } - - /* enable interupt */ - WREG32_P(UVD_MASTINT_EN, 3<<1, ~(3 << 1)); - - r = r600_uvd_rbc_start(rdev); - if (!r) - DRM_INFO("UVD initialized successfully.\n"); - - /* lower clocks again */ - radeon_set_uvd_clocks(rdev, 0, 0); - - return r; -} - -/* * GPU scratch registers helpers function. */ void r600_scratch_init(struct radeon_device *rdev) @@ -2876,94 +2702,6 @@ int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) return r; } -/** - * r600_dma_ring_test - simple async dma engine test - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Test the DMA engine by writing using it to write an - * value to memory. (r6xx-SI). - * Returns 0 for success, error for failure. - */ -int r600_dma_ring_test(struct radeon_device *rdev, - struct radeon_ring *ring) -{ - unsigned i; - int r; - void __iomem *ptr = (void *)rdev->vram_scratch.ptr; - u32 tmp; - - if (!ptr) { - DRM_ERROR("invalid vram scratch pointer\n"); - return -EINVAL; - } - - tmp = 0xCAFEDEAD; - writel(tmp, ptr); - - r = radeon_ring_lock(rdev, ring, 4); - if (r) { - DRM_ERROR("radeon: dma failed to lock ring %d (%d).\n", ring->idx, r); - return r; - } - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1)); - radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xff); - radeon_ring_write(ring, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev, ring); - - for (i = 0; i < rdev->usec_timeout; i++) { - tmp = readl(ptr); - if (tmp == 0xDEADBEEF) - break; - DRM_UDELAY(1); - } - - if (i < rdev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); - } else { - DRM_ERROR("radeon: ring %d test failed (0x%08X)\n", - ring->idx, tmp); - r = -EINVAL; - } - return r; -} - -int r600_uvd_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) -{ - uint32_t tmp = 0; - unsigned i; - int r; - - WREG32(UVD_CONTEXT_ID, 0xCAFEDEAD); - r = radeon_ring_lock(rdev, ring, 3); - if (r) { - DRM_ERROR("radeon: cp failed to lock ring %d (%d).\n", - ring->idx, r); - return r; - } - radeon_ring_write(ring, PACKET0(UVD_CONTEXT_ID, 0)); - radeon_ring_write(ring, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev, ring); - for (i = 0; i < rdev->usec_timeout; i++) { - tmp = RREG32(UVD_CONTEXT_ID); - if (tmp == 0xDEADBEEF) - break; - DRM_UDELAY(1); - } - - if (i < rdev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", - ring->idx, i); - } else { - DRM_ERROR("radeon: ring %d test failed (0x%08X)\n", - ring->idx, tmp); - r = -EINVAL; - } - return r; -} - /* * CP fences/semaphores */ @@ -2972,30 +2710,31 @@ void r600_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence) { struct radeon_ring *ring = &rdev->ring[fence->ring]; + u32 cp_coher_cntl = PACKET3_TC_ACTION_ENA | PACKET3_VC_ACTION_ENA | + PACKET3_SH_ACTION_ENA; + + if (rdev->family >= CHIP_RV770) + cp_coher_cntl |= PACKET3_FULL_CACHE_ENA; if (rdev->wb.use_event) { u64 addr = rdev->fence_drv[fence->ring].gpu_addr; /* flush read cache over gart */ radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(ring, PACKET3_TC_ACTION_ENA | - PACKET3_VC_ACTION_ENA | - PACKET3_SH_ACTION_ENA); + radeon_ring_write(ring, cp_coher_cntl); radeon_ring_write(ring, 0xFFFFFFFF); radeon_ring_write(ring, 0); radeon_ring_write(ring, 10); /* poll interval */ /* EVENT_WRITE_EOP - flush caches, send int */ radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4)); radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT_TS) | EVENT_INDEX(5)); - radeon_ring_write(ring, addr & 0xffffffff); + radeon_ring_write(ring, lower_32_bits(addr)); radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2)); radeon_ring_write(ring, fence->seq); radeon_ring_write(ring, 0); } else { /* flush read cache over gart */ radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(ring, PACKET3_TC_ACTION_ENA | - PACKET3_VC_ACTION_ENA | - PACKET3_SH_ACTION_ENA); + radeon_ring_write(ring, cp_coher_cntl); radeon_ring_write(ring, 0xFFFFFFFF); radeon_ring_write(ring, 0); radeon_ring_write(ring, 10); /* poll interval */ @@ -3015,31 +2754,7 @@ void r600_fence_ring_emit(struct radeon_device *rdev, } } -void r600_uvd_fence_emit(struct radeon_device *rdev, - struct radeon_fence *fence) -{ - struct radeon_ring *ring = &rdev->ring[fence->ring]; - uint32_t addr = rdev->fence_drv[fence->ring].gpu_addr; - - radeon_ring_write(ring, PACKET0(UVD_CONTEXT_ID, 0)); - radeon_ring_write(ring, fence->seq); - radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA0, 0)); - radeon_ring_write(ring, addr & 0xffffffff); - radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA1, 0)); - radeon_ring_write(ring, upper_32_bits(addr) & 0xff); - radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_CMD, 0)); - radeon_ring_write(ring, 0); - - radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA0, 0)); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA1, 0)); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_CMD, 0)); - radeon_ring_write(ring, 2); - return; -} - -void r600_semaphore_ring_emit(struct radeon_device *rdev, +bool r600_semaphore_ring_emit(struct radeon_device *rdev, struct radeon_ring *ring, struct radeon_semaphore *semaphore, bool emit_wait) @@ -3051,101 +2766,14 @@ void r600_semaphore_ring_emit(struct radeon_device *rdev, sel |= PACKET3_SEM_WAIT_ON_SIGNAL; radeon_ring_write(ring, PACKET3(PACKET3_MEM_SEMAPHORE, 1)); - radeon_ring_write(ring, addr & 0xffffffff); + radeon_ring_write(ring, lower_32_bits(addr)); radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | sel); -} - -/* - * DMA fences/semaphores - */ - -/** - * r600_dma_fence_ring_emit - emit a fence on the DMA ring - * - * @rdev: radeon_device pointer - * @fence: radeon fence object - * - * Add a DMA fence packet to the ring to write - * the fence seq number and DMA trap packet to generate - * an interrupt if needed (r6xx-r7xx). - */ -void r600_dma_fence_ring_emit(struct radeon_device *rdev, - struct radeon_fence *fence) -{ - struct radeon_ring *ring = &rdev->ring[fence->ring]; - u64 addr = rdev->fence_drv[fence->ring].gpu_addr; - - /* write the fence */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_FENCE, 0, 0, 0)); - radeon_ring_write(ring, addr & 0xfffffffc); - radeon_ring_write(ring, (upper_32_bits(addr) & 0xff)); - radeon_ring_write(ring, lower_32_bits(fence->seq)); - /* generate an interrupt */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_TRAP, 0, 0, 0)); -} -/** - * r600_dma_semaphore_ring_emit - emit a semaphore on the dma ring - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * @semaphore: radeon semaphore object - * @emit_wait: wait or signal semaphore - * - * Add a DMA semaphore packet to the ring wait on or signal - * other rings (r6xx-SI). - */ -void r600_dma_semaphore_ring_emit(struct radeon_device *rdev, - struct radeon_ring *ring, - struct radeon_semaphore *semaphore, - bool emit_wait) -{ - u64 addr = semaphore->gpu_addr; - u32 s = emit_wait ? 0 : 1; - - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SEMAPHORE, 0, s, 0)); - radeon_ring_write(ring, addr & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(addr) & 0xff); -} - -void r600_uvd_semaphore_emit(struct radeon_device *rdev, - struct radeon_ring *ring, - struct radeon_semaphore *semaphore, - bool emit_wait) -{ - uint64_t addr = semaphore->gpu_addr; - - radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_LOW, 0)); - radeon_ring_write(ring, (addr >> 3) & 0x000FFFFF); - - radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_HIGH, 0)); - radeon_ring_write(ring, (addr >> 23) & 0x000FFFFF); - - radeon_ring_write(ring, PACKET0(UVD_SEMA_CMD, 0)); - radeon_ring_write(ring, emit_wait ? 1 : 0); -} - -int r600_copy_blit(struct radeon_device *rdev, - uint64_t src_offset, - uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence **fence) -{ - struct radeon_semaphore *sem = NULL; - struct radeon_sa_bo *vb = NULL; - int r; - - r = r600_blit_prepare_copy(rdev, num_gpu_pages, fence, &vb, &sem); - if (r) { - return r; - } - r600_kms_blit_copy(rdev, src_offset, dst_offset, num_gpu_pages, vb); - r600_blit_done_copy(rdev, fence, vb, sem); - return 0; + return true; } /** - * r600_copy_dma - copy pages using the DMA engine + * r600_copy_cpdma - copy pages using the CP DMA engine * * @rdev: radeon_device pointer * @src_offset: src GPU address @@ -3153,19 +2781,19 @@ int r600_copy_blit(struct radeon_device *rdev, * @num_gpu_pages: number of GPU pages to xfer * @fence: radeon fence object * - * Copy GPU paging using the DMA engine (r6xx). + * Copy GPU paging using the CP DMA engine (r6xx+). * Used by the radeon ttm implementation to move pages if * registered as the asic copy callback. */ -int r600_copy_dma(struct radeon_device *rdev, - uint64_t src_offset, uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence **fence) +int r600_copy_cpdma(struct radeon_device *rdev, + uint64_t src_offset, uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence **fence) { struct radeon_semaphore *sem = NULL; - int ring_index = rdev->asic->copy.dma_ring_index; + int ring_index = rdev->asic->copy.blit_ring_index; struct radeon_ring *ring = &rdev->ring[ring_index]; - u32 size_in_dw, cur_size_in_dw; + u32 size_in_bytes, cur_size_in_bytes, tmp; int i, num_loops; int r = 0; @@ -3175,40 +2803,46 @@ int r600_copy_dma(struct radeon_device *rdev, return r; } - size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4; - num_loops = DIV_ROUND_UP(size_in_dw, 0xFFFE); - r = radeon_ring_lock(rdev, ring, num_loops * 4 + 8); + size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT); + num_loops = DIV_ROUND_UP(size_in_bytes, 0x1fffff); + r = radeon_ring_lock(rdev, ring, num_loops * 6 + 24); if (r) { DRM_ERROR("radeon: moving bo (%d).\n", r); radeon_semaphore_free(rdev, &sem, NULL); return r; } - if (radeon_fence_need_sync(*fence, ring->idx)) { - radeon_semaphore_sync_rings(rdev, sem, (*fence)->ring, - ring->idx); - radeon_fence_note_sync(*fence, ring->idx); - } else { - radeon_semaphore_free(rdev, &sem, NULL); - } + radeon_semaphore_sync_to(sem, *fence); + radeon_semaphore_sync_rings(rdev, sem, ring->idx); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); + radeon_ring_write(ring, WAIT_3D_IDLE_bit); for (i = 0; i < num_loops; i++) { - cur_size_in_dw = size_in_dw; - if (cur_size_in_dw > 0xFFFE) - cur_size_in_dw = 0xFFFE; - size_in_dw -= cur_size_in_dw; - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 0, 0, cur_size_in_dw)); - radeon_ring_write(ring, dst_offset & 0xfffffffc); - radeon_ring_write(ring, src_offset & 0xfffffffc); - radeon_ring_write(ring, (((upper_32_bits(dst_offset) & 0xff) << 16) | - (upper_32_bits(src_offset) & 0xff))); - src_offset += cur_size_in_dw * 4; - dst_offset += cur_size_in_dw * 4; + cur_size_in_bytes = size_in_bytes; + if (cur_size_in_bytes > 0x1fffff) + cur_size_in_bytes = 0x1fffff; + size_in_bytes -= cur_size_in_bytes; + tmp = upper_32_bits(src_offset) & 0xff; + if (size_in_bytes == 0) + tmp |= PACKET3_CP_DMA_CP_SYNC; + radeon_ring_write(ring, PACKET3(PACKET3_CP_DMA, 4)); + radeon_ring_write(ring, lower_32_bits(src_offset)); + radeon_ring_write(ring, tmp); + radeon_ring_write(ring, lower_32_bits(dst_offset)); + radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff); + radeon_ring_write(ring, cur_size_in_bytes); + src_offset += cur_size_in_bytes; + dst_offset += cur_size_in_bytes; } + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); + radeon_ring_write(ring, WAIT_CP_DMA_IDLE_bit); r = radeon_fence_emit(rdev, fence, ring->idx); if (r) { radeon_ring_unlock_undo(rdev, ring); + radeon_semaphore_free(rdev, &sem, NULL); return r; } @@ -3239,19 +2873,13 @@ static int r600_startup(struct radeon_device *rdev) /* enable pcie gen2 link */ r600_pcie_gen2_enable(rdev); - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) { - r = r600_init_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load firmware!\n"); - return r; - } - } - + /* scratch needs to be initialized before MC */ r = r600_vram_scratch_init(rdev); if (r) return r; r600_mc_program(rdev); + if (rdev->flags & RADEON_IS_AGP) { r600_agp_enable(rdev); } else { @@ -3260,12 +2888,6 @@ static int r600_startup(struct radeon_device *rdev) return r; } r600_gpu_init(rdev); - r = r600_blit_init(rdev); - if (r) { - r600_blit_fini(rdev); - rdev->asic->copy.copy = NULL; - dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); - } /* allocate wb buffer */ r = radeon_wb_init(rdev); @@ -3278,12 +2900,6 @@ static int r600_startup(struct radeon_device *rdev) return r; } - r = radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_DMA_INDEX); - if (r) { - dev_err(rdev->dev, "failed initializing DMA fences (%d).\n", r); - return r; - } - /* Enable IRQ */ if (!rdev->irq.installed) { r = radeon_irq_kms_init(rdev); @@ -3301,15 +2917,7 @@ static int r600_startup(struct radeon_device *rdev) ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, - R600_CP_RB_RPTR, R600_CP_RB_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); - if (r) - return r; - - ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; - r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET, - DMA_RB_RPTR, DMA_RB_WPTR, - 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); + RADEON_CP_PACKET2); if (r) return r; @@ -3320,10 +2928,6 @@ static int r600_startup(struct radeon_device *rdev) if (r) return r; - r = r600_dma_resume(rdev); - if (r) - return r; - r = radeon_ib_pool_init(rdev); if (r) { dev_err(rdev->dev, "IB initialization failed (%d).\n", r); @@ -3364,6 +2968,9 @@ int r600_resume(struct radeon_device *rdev) /* post card */ atom_asic_init(rdev->mode_info.atom_context); + if (rdev->pm.pm_method == PM_METHOD_DPM) + radeon_pm_resume(rdev); + rdev->accel_working = true; r = r600_startup(rdev); if (r) { @@ -3377,9 +2984,9 @@ int r600_resume(struct radeon_device *rdev) int r600_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r600_audio_fini(rdev); r600_cp_stop(rdev); - r600_dma_stop(rdev); r600_irq_suspend(rdev); radeon_wb_disable(rdev); r600_pcie_gart_disable(rdev); @@ -3445,12 +3052,20 @@ int r600_init(struct radeon_device *rdev) if (r) return r; + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) { + r = r600_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL; r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024); - rdev->ring[R600_RING_TYPE_DMA_INDEX].ring_obj = NULL; - r600_ring_init(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX], 64 * 1024); - rdev->ih.ring_obj = NULL; r600_ih_ring_init(rdev, 64 * 1024); @@ -3463,7 +3078,6 @@ int r600_init(struct radeon_device *rdev) if (r) { dev_err(rdev->dev, "disabling GPU acceleration\n"); r600_cp_fini(rdev); - r600_dma_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); @@ -3477,10 +3091,9 @@ int r600_init(struct radeon_device *rdev) void r600_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); r600_audio_fini(rdev); - r600_blit_fini(rdev); r600_cp_fini(rdev); - r600_dma_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); @@ -3530,16 +3143,6 @@ void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) radeon_ring_write(ring, ib->length_dw); } -void r600_uvd_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) -{ - struct radeon_ring *ring = &rdev->ring[ib->ring]; - - radeon_ring_write(ring, PACKET0(UVD_RBC_IB_BASE, 0)); - radeon_ring_write(ring, ib->gpu_addr); - radeon_ring_write(ring, PACKET0(UVD_RBC_IB_SIZE, 0)); - radeon_ring_write(ring, ib->length_dw); -} - int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) { struct radeon_ib ib; @@ -3593,139 +3196,6 @@ free_scratch: return r; } -/** - * r600_dma_ib_test - test an IB on the DMA engine - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Test a simple IB in the DMA ring (r6xx-SI). - * Returns 0 on success, error on failure. - */ -int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) -{ - struct radeon_ib ib; - unsigned i; - int r; - void __iomem *ptr = (void *)rdev->vram_scratch.ptr; - u32 tmp = 0; - - if (!ptr) { - DRM_ERROR("invalid vram scratch pointer\n"); - return -EINVAL; - } - - tmp = 0xCAFEDEAD; - writel(tmp, ptr); - - r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256); - if (r) { - DRM_ERROR("radeon: failed to get ib (%d).\n", r); - return r; - } - - ib.ptr[0] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1); - ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc; - ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xff; - ib.ptr[3] = 0xDEADBEEF; - ib.length_dw = 4; - - r = radeon_ib_schedule(rdev, &ib, NULL); - if (r) { - radeon_ib_free(rdev, &ib); - DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); - return r; - } - r = radeon_fence_wait(ib.fence, false); - if (r) { - DRM_ERROR("radeon: fence wait failed (%d).\n", r); - return r; - } - for (i = 0; i < rdev->usec_timeout; i++) { - tmp = readl(ptr); - if (tmp == 0xDEADBEEF) - break; - DRM_UDELAY(1); - } - if (i < rdev->usec_timeout) { - DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i); - } else { - DRM_ERROR("radeon: ib test failed (0x%08X)\n", tmp); - r = -EINVAL; - } - radeon_ib_free(rdev, &ib); - return r; -} - -int r600_uvd_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) -{ - struct radeon_fence *fence = NULL; - int r; - - r = radeon_set_uvd_clocks(rdev, 53300, 40000); - if (r) { - DRM_ERROR("radeon: failed to raise UVD clocks (%d).\n", r); - return r; - } - - r = radeon_uvd_get_create_msg(rdev, ring->idx, 1, NULL); - if (r) { - DRM_ERROR("radeon: failed to get create msg (%d).\n", r); - goto error; - } - - r = radeon_uvd_get_destroy_msg(rdev, ring->idx, 1, &fence); - if (r) { - DRM_ERROR("radeon: failed to get destroy ib (%d).\n", r); - goto error; - } - - r = radeon_fence_wait(fence, false); - if (r) { - DRM_ERROR("radeon: fence wait failed (%d).\n", r); - goto error; - } - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); -error: - radeon_fence_unref(&fence); - radeon_set_uvd_clocks(rdev, 0, 0); - return r; -} - -/** - * r600_dma_ring_ib_execute - Schedule an IB on the DMA engine - * - * @rdev: radeon_device pointer - * @ib: IB object to schedule - * - * Schedule an IB in the DMA ring (r6xx-r7xx). - */ -void r600_dma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) -{ - struct radeon_ring *ring = &rdev->ring[ib->ring]; - - if (rdev->wb.enabled) { - u32 next_rptr = ring->wptr + 4; - while ((next_rptr & 7) != 5) - next_rptr++; - next_rptr += 3; - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1)); - radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xff); - radeon_ring_write(ring, next_rptr); - } - - /* The indirect buffer packet must end on an 8 DW boundary in the DMA ring. - * Pad as necessary with NOPs. - */ - while ((ring->wptr & 7) != 5) - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_INDIRECT_BUFFER, 0, 0, 0)); - radeon_ring_write(ring, (ib->gpu_addr & 0xFFFFFFE0)); - radeon_ring_write(ring, (ib->length_dw << 16) | (upper_32_bits(ib->gpu_addr) & 0xFF)); - -} - /* * Interrupts * @@ -3742,7 +3212,7 @@ void r600_ih_ring_init(struct radeon_device *rdev, unsigned ring_size) u32 rb_bufsz; /* Align ring size */ - rb_bufsz = drm_order(ring_size / 4); + rb_bufsz = order_base_2(ring_size / 4); ring_size = (1 << rb_bufsz) * 4; rdev->ih.ring_size = ring_size; rdev->ih.ptr_mask = rdev->ih.ring_size - 1; @@ -3979,7 +3449,7 @@ int r600_irq_init(struct radeon_device *rdev) WREG32(INTERRUPT_CNTL, interrupt_cntl); WREG32(IH_RB_BASE, rdev->ih.gpu_addr >> 8); - rb_bufsz = drm_order(rdev->ih.ring_size / 4); + rb_bufsz = order_base_2(rdev->ih.ring_size / 4); ih_rb_cntl = (IH_WPTR_OVERFLOW_ENABLE | IH_WPTR_OVERFLOW_CLEAR | @@ -4039,7 +3509,6 @@ int r600_irq_set(struct radeon_device *rdev) u32 hpd1, hpd2, hpd3, hpd4 = 0, hpd5 = 0, hpd6 = 0; u32 grbm_int_cntl = 0; u32 hdmi0, hdmi1; - u32 d1grph = 0, d2grph = 0; u32 dma_cntl; u32 thermal_int = 0; @@ -4148,8 +3617,8 @@ int r600_irq_set(struct radeon_device *rdev) WREG32(CP_INT_CNTL, cp_int_cntl); WREG32(DMA_CNTL, dma_cntl); WREG32(DxMODE_INT_MASK, mode_int); - WREG32(D1GRPH_INTERRUPT_CONTROL, d1grph); - WREG32(D2GRPH_INTERRUPT_CONTROL, d2grph); + WREG32(D1GRPH_INTERRUPT_CONTROL, DxGRPH_PFLIP_INT_MASK); + WREG32(D2GRPH_INTERRUPT_CONTROL, DxGRPH_PFLIP_INT_MASK); WREG32(GRBM_INT_CNTL, grbm_int_cntl); if (ASIC_IS_DCE3(rdev)) { WREG32(DC_HPD1_INT_CONTROL, hpd1); @@ -4326,6 +3795,7 @@ static u32 r600_get_ih_wptr(struct radeon_device *rdev) tmp = RREG32(IH_RB_CNTL); tmp |= IH_WPTR_OVERFLOW_CLEAR; WREG32(IH_RB_CNTL, tmp); + wptr &= ~RB_OVERFLOW; } return (wptr & rdev->ih.ptr_mask); } @@ -4410,7 +3880,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[0])) - radeon_crtc_handle_flip(rdev, 0); + radeon_crtc_handle_vblank(rdev, 0); rdev->irq.stat_regs.r600.disp_int &= ~LB_D1_VBLANK_INTERRUPT; DRM_DEBUG("IH: D1 vblank\n"); } @@ -4436,7 +3906,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[1])) - radeon_crtc_handle_flip(rdev, 1); + radeon_crtc_handle_vblank(rdev, 1); rdev->irq.stat_regs.r600.disp_int &= ~LB_D2_VBLANK_INTERRUPT; DRM_DEBUG("IH: D2 vblank\n"); } @@ -4452,6 +3922,14 @@ restart_ih: break; } break; + case 9: /* D1 pflip */ + DRM_DEBUG("IH: D1 flip\n"); + radeon_crtc_handle_flip(rdev, 0); + break; + case 11: /* D2 pflip */ + DRM_DEBUG("IH: D2 flip\n"); + radeon_crtc_handle_flip(rdev, 1); + break; case 19: /* HPD/DAC hotplug */ switch (src_data) { case 0: @@ -4522,6 +4000,10 @@ restart_ih: break; } break; + case 124: /* UVD */ + DRM_DEBUG("IH: UVD int: 0x%08x\n", src_data); + radeon_fence_process(rdev, R600_RING_TYPE_UVD_INDEX); + break; case 176: /* CP_INT in ring buffer */ case 177: /* CP_INT in IB1 */ case 178: /* CP_INT in IB2 */ diff --git a/drivers/gpu/drm/radeon/r600_audio.c b/drivers/gpu/drm/radeon/r600_audio.c index c92eb86a8e5..bffac10c429 100644 --- a/drivers/gpu/drm/radeon/r600_audio.c +++ b/drivers/gpu/drm/radeon/r600_audio.c @@ -57,12 +57,12 @@ static bool radeon_dig_encoder(struct drm_encoder *encoder) */ static int r600_audio_chipset_supported(struct radeon_device *rdev) { - return ASIC_IS_DCE2(rdev) && !ASIC_IS_DCE6(rdev); + return ASIC_IS_DCE2(rdev) && !ASIC_IS_NODCE(rdev); } -struct r600_audio r600_audio_status(struct radeon_device *rdev) +struct r600_audio_pin r600_audio_status(struct radeon_device *rdev) { - struct r600_audio status; + struct r600_audio_pin status; uint32_t value; value = RREG32(R600_AUDIO_RATE_BPS_CHANNEL); @@ -120,16 +120,16 @@ void r600_audio_update_hdmi(struct work_struct *work) struct radeon_device *rdev = container_of(work, struct radeon_device, audio_work); struct drm_device *dev = rdev->ddev; - struct r600_audio audio_status = r600_audio_status(rdev); + struct r600_audio_pin audio_status = r600_audio_status(rdev); struct drm_encoder *encoder; bool changed = false; - if (rdev->audio_status.channels != audio_status.channels || - rdev->audio_status.rate != audio_status.rate || - rdev->audio_status.bits_per_sample != audio_status.bits_per_sample || - rdev->audio_status.status_bits != audio_status.status_bits || - rdev->audio_status.category_code != audio_status.category_code) { - rdev->audio_status = audio_status; + if (rdev->audio.pin[0].channels != audio_status.channels || + rdev->audio.pin[0].rate != audio_status.rate || + rdev->audio.pin[0].bits_per_sample != audio_status.bits_per_sample || + rdev->audio.pin[0].status_bits != audio_status.status_bits || + rdev->audio.pin[0].category_code != audio_status.category_code) { + rdev->audio.pin[0] = audio_status; changed = true; } @@ -141,13 +141,16 @@ void r600_audio_update_hdmi(struct work_struct *work) } } -/* - * turn on/off audio engine - */ -static void r600_audio_engine_enable(struct radeon_device *rdev, bool enable) +/* enable the audio stream */ +void r600_audio_enable(struct radeon_device *rdev, + struct r600_audio_pin *pin, + bool enable) { u32 value = 0; - DRM_INFO("%s audio support\n", enable ? "Enabling" : "Disabling"); + + if (!pin) + return; + if (ASIC_IS_DCE4(rdev)) { if (enable) { value |= 0x81000000; /* Required to enable audio */ @@ -158,7 +161,6 @@ static void r600_audio_engine_enable(struct radeon_device *rdev, bool enable) WREG32_P(R600_AUDIO_ENABLE, enable ? 0x81000000 : 0x0, ~0x81000000); } - rdev->audio_enabled = enable; } /* @@ -169,13 +171,17 @@ int r600_audio_init(struct radeon_device *rdev) if (!radeon_audio || !r600_audio_chipset_supported(rdev)) return 0; - r600_audio_engine_enable(rdev, true); + rdev->audio.enabled = true; - rdev->audio_status.channels = -1; - rdev->audio_status.rate = -1; - rdev->audio_status.bits_per_sample = -1; - rdev->audio_status.status_bits = 0; - rdev->audio_status.category_code = 0; + rdev->audio.num_pins = 1; + rdev->audio.pin[0].channels = -1; + rdev->audio.pin[0].rate = -1; + rdev->audio.pin[0].bits_per_sample = -1; + rdev->audio.pin[0].status_bits = 0; + rdev->audio.pin[0].category_code = 0; + rdev->audio.pin[0].id = 0; + /* disable audio. it will be set up later */ + r600_audio_enable(rdev, &rdev->audio.pin[0], false); return 0; } @@ -186,8 +192,16 @@ int r600_audio_init(struct radeon_device *rdev) */ void r600_audio_fini(struct radeon_device *rdev) { - if (!rdev->audio_enabled) + if (!rdev->audio.enabled) return; - r600_audio_engine_enable(rdev, false); + r600_audio_enable(rdev, &rdev->audio.pin[0], false); + + rdev->audio.enabled = false; +} + +struct r600_audio_pin *r600_audio_get_pin(struct radeon_device *rdev) +{ + /* only one pin on 6xx-NI */ + return &rdev->audio.pin[0]; } diff --git a/drivers/gpu/drm/radeon/r600_blit.c b/drivers/gpu/drm/radeon/r600_blit.c index f651881eb0a..daf7572be97 100644 --- a/drivers/gpu/drm/radeon/r600_blit.c +++ b/drivers/gpu/drm/radeon/r600_blit.c @@ -31,6 +31,37 @@ #include "r600_blit_shaders.h" +/* 23 bits of float fractional data */ +#define I2F_FRAC_BITS 23 +#define I2F_MASK ((1 << I2F_FRAC_BITS) - 1) + +/* + * Converts unsigned integer into 32-bit IEEE floating point representation. + * Will be exact from 0 to 2^24. Above that, we round towards zero + * as the fractional bits will not fit in a float. (It would be better to + * round towards even as the fpu does, but that is slower.) + */ +static __pure uint32_t int2float(uint32_t x) +{ + uint32_t msb, exponent, fraction; + + /* Zero is special */ + if (!x) return 0; + + /* Get location of the most significant bit */ + msb = __fls(x); + + /* + * Use a rotate instead of a shift because that works both leftwards + * and rightwards due to the mod(32) behaviour. This means we don't + * need to check to see if we are above 2^24 or not. + */ + fraction = ror32(x, (msb - I2F_FRAC_BITS) & 0x1f) & I2F_MASK; + exponent = (127 + msb) << I2F_FRAC_BITS; + + return fraction + exponent; +} + #define DI_PT_RECTLIST 0x11 #define DI_INDEX_SIZE_16_BIT 0x0 #define DI_SRC_SEL_AUTO_INDEX 0x2 diff --git a/drivers/gpu/drm/radeon/r600_blit_kms.c b/drivers/gpu/drm/radeon/r600_blit_kms.c deleted file mode 100644 index 9fb5780a552..00000000000 --- a/drivers/gpu/drm/radeon/r600_blit_kms.c +++ /dev/null @@ -1,785 +0,0 @@ -/* - * Copyright 2009 Advanced Micro Devices, Inc. - * Copyright 2009 Red Hat Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - * - */ - -#include <drm/drmP.h> -#include <drm/radeon_drm.h> -#include "radeon.h" - -#include "r600d.h" -#include "r600_blit_shaders.h" -#include "radeon_blit_common.h" - -/* 23 bits of float fractional data */ -#define I2F_FRAC_BITS 23 -#define I2F_MASK ((1 << I2F_FRAC_BITS) - 1) - -/* - * Converts unsigned integer into 32-bit IEEE floating point representation. - * Will be exact from 0 to 2^24. Above that, we round towards zero - * as the fractional bits will not fit in a float. (It would be better to - * round towards even as the fpu does, but that is slower.) - */ -__pure uint32_t int2float(uint32_t x) -{ - uint32_t msb, exponent, fraction; - - /* Zero is special */ - if (!x) return 0; - - /* Get location of the most significant bit */ - msb = __fls(x); - - /* - * Use a rotate instead of a shift because that works both leftwards - * and rightwards due to the mod(32) behaviour. This means we don't - * need to check to see if we are above 2^24 or not. - */ - fraction = ror32(x, (msb - I2F_FRAC_BITS) & 0x1f) & I2F_MASK; - exponent = (127 + msb) << I2F_FRAC_BITS; - - return fraction + exponent; -} - -/* emits 21 on rv770+, 23 on r600 */ -static void -set_render_target(struct radeon_device *rdev, int format, - int w, int h, u64 gpu_addr) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u32 cb_color_info; - int pitch, slice; - - h = ALIGN(h, 8); - if (h < 8) - h = 8; - - cb_color_info = CB_FORMAT(format) | - CB_SOURCE_FORMAT(CB_SF_EXPORT_NORM) | - CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1); - pitch = (w / 8) - 1; - slice = ((w * h) / 64) - 1; - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, gpu_addr >> 8); - - if (rdev->family > CHIP_R600 && rdev->family < CHIP_RV770) { - radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_BASE_UPDATE, 0)); - radeon_ring_write(ring, 2 << 0); - } - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (CB_COLOR0_SIZE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, (pitch << 0) | (slice << 10)); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (CB_COLOR0_VIEW - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, 0); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (CB_COLOR0_INFO - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, cb_color_info); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (CB_COLOR0_TILE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, 0); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (CB_COLOR0_FRAG - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, 0); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (CB_COLOR0_MASK - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, 0); -} - -/* emits 5dw */ -static void -cp_set_surface_sync(struct radeon_device *rdev, - u32 sync_type, u32 size, - u64 mc_addr) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u32 cp_coher_size; - - if (size == 0xffffffff) - cp_coher_size = 0xffffffff; - else - cp_coher_size = ((size + 255) >> 8); - - radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(ring, sync_type); - radeon_ring_write(ring, cp_coher_size); - radeon_ring_write(ring, mc_addr >> 8); - radeon_ring_write(ring, 10); /* poll interval */ -} - -/* emits 21dw + 1 surface sync = 26dw */ -static void -set_shaders(struct radeon_device *rdev) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u64 gpu_addr; - u32 sq_pgm_resources; - - /* setup shader regs */ - sq_pgm_resources = (1 << 0); - - /* VS */ - gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset; - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, gpu_addr >> 8); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (SQ_PGM_RESOURCES_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, sq_pgm_resources); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (SQ_PGM_CF_OFFSET_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, 0); - - /* PS */ - gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.ps_offset; - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, gpu_addr >> 8); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (SQ_PGM_RESOURCES_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, sq_pgm_resources | (1 << 28)); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (SQ_PGM_EXPORTS_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, 2); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(ring, (SQ_PGM_CF_OFFSET_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, 0); - - gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset; - cp_set_surface_sync(rdev, PACKET3_SH_ACTION_ENA, 512, gpu_addr); -} - -/* emits 9 + 1 sync (5) = 14*/ -static void -set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u32 sq_vtx_constant_word2; - - sq_vtx_constant_word2 = SQ_VTXC_BASE_ADDR_HI(upper_32_bits(gpu_addr) & 0xff) | - SQ_VTXC_STRIDE(16); -#ifdef __BIG_ENDIAN - sq_vtx_constant_word2 |= SQ_VTXC_ENDIAN_SWAP(SQ_ENDIAN_8IN32); -#endif - - radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 7)); - radeon_ring_write(ring, 0x460); - radeon_ring_write(ring, gpu_addr & 0xffffffff); - radeon_ring_write(ring, 48 - 1); - radeon_ring_write(ring, sq_vtx_constant_word2); - radeon_ring_write(ring, 1 << 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, SQ_TEX_VTX_VALID_BUFFER << 30); - - if ((rdev->family == CHIP_RV610) || - (rdev->family == CHIP_RV620) || - (rdev->family == CHIP_RS780) || - (rdev->family == CHIP_RS880) || - (rdev->family == CHIP_RV710)) - cp_set_surface_sync(rdev, - PACKET3_TC_ACTION_ENA, 48, gpu_addr); - else - cp_set_surface_sync(rdev, - PACKET3_VC_ACTION_ENA, 48, gpu_addr); -} - -/* emits 9 */ -static void -set_tex_resource(struct radeon_device *rdev, - int format, int w, int h, int pitch, - u64 gpu_addr, u32 size) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - uint32_t sq_tex_resource_word0, sq_tex_resource_word1, sq_tex_resource_word4; - - if (h < 1) - h = 1; - - sq_tex_resource_word0 = S_038000_DIM(V_038000_SQ_TEX_DIM_2D) | - S_038000_TILE_MODE(V_038000_ARRAY_1D_TILED_THIN1); - sq_tex_resource_word0 |= S_038000_PITCH((pitch >> 3) - 1) | - S_038000_TEX_WIDTH(w - 1); - - sq_tex_resource_word1 = S_038004_DATA_FORMAT(format); - sq_tex_resource_word1 |= S_038004_TEX_HEIGHT(h - 1); - - sq_tex_resource_word4 = S_038010_REQUEST_SIZE(1) | - S_038010_DST_SEL_X(SQ_SEL_X) | - S_038010_DST_SEL_Y(SQ_SEL_Y) | - S_038010_DST_SEL_Z(SQ_SEL_Z) | - S_038010_DST_SEL_W(SQ_SEL_W); - - cp_set_surface_sync(rdev, - PACKET3_TC_ACTION_ENA, size, gpu_addr); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 7)); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, sq_tex_resource_word0); - radeon_ring_write(ring, sq_tex_resource_word1); - radeon_ring_write(ring, gpu_addr >> 8); - radeon_ring_write(ring, gpu_addr >> 8); - radeon_ring_write(ring, sq_tex_resource_word4); - radeon_ring_write(ring, 0); - radeon_ring_write(ring, SQ_TEX_VTX_VALID_TEXTURE << 30); -} - -/* emits 12 */ -static void -set_scissors(struct radeon_device *rdev, int x1, int y1, - int x2, int y2) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(ring, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, (x1 << 0) | (y1 << 16)); - radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(ring, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31)); - radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); - - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(ring, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31)); - radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); -} - -/* emits 10 */ -static void -draw_auto(struct radeon_device *rdev) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(ring, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); - radeon_ring_write(ring, DI_PT_RECTLIST); - - radeon_ring_write(ring, PACKET3(PACKET3_INDEX_TYPE, 0)); - radeon_ring_write(ring, -#ifdef __BIG_ENDIAN - (2 << 2) | -#endif - DI_INDEX_SIZE_16_BIT); - - radeon_ring_write(ring, PACKET3(PACKET3_NUM_INSTANCES, 0)); - radeon_ring_write(ring, 1); - - radeon_ring_write(ring, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1)); - radeon_ring_write(ring, 3); - radeon_ring_write(ring, DI_SRC_SEL_AUTO_INDEX); - -} - -/* emits 14 */ -static void -set_default_state(struct radeon_device *rdev) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - u32 sq_config, sq_gpr_resource_mgmt_1, sq_gpr_resource_mgmt_2; - u32 sq_thread_resource_mgmt, sq_stack_resource_mgmt_1, sq_stack_resource_mgmt_2; - int num_ps_gprs, num_vs_gprs, num_temp_gprs, num_gs_gprs, num_es_gprs; - int num_ps_threads, num_vs_threads, num_gs_threads, num_es_threads; - int num_ps_stack_entries, num_vs_stack_entries, num_gs_stack_entries, num_es_stack_entries; - u64 gpu_addr; - int dwords; - - switch (rdev->family) { - case CHIP_R600: - num_ps_gprs = 192; - num_vs_gprs = 56; - num_temp_gprs = 4; - num_gs_gprs = 0; - num_es_gprs = 0; - num_ps_threads = 136; - num_vs_threads = 48; - num_gs_threads = 4; - num_es_threads = 4; - num_ps_stack_entries = 128; - num_vs_stack_entries = 128; - num_gs_stack_entries = 0; - num_es_stack_entries = 0; - break; - case CHIP_RV630: - case CHIP_RV635: - num_ps_gprs = 84; - num_vs_gprs = 36; - num_temp_gprs = 4; - num_gs_gprs = 0; - num_es_gprs = 0; - num_ps_threads = 144; - num_vs_threads = 40; - num_gs_threads = 4; - num_es_threads = 4; - num_ps_stack_entries = 40; - num_vs_stack_entries = 40; - num_gs_stack_entries = 32; - num_es_stack_entries = 16; - break; - case CHIP_RV610: - case CHIP_RV620: - case CHIP_RS780: - case CHIP_RS880: - default: - num_ps_gprs = 84; - num_vs_gprs = 36; - num_temp_gprs = 4; - num_gs_gprs = 0; - num_es_gprs = 0; - num_ps_threads = 136; - num_vs_threads = 48; - num_gs_threads = 4; - num_es_threads = 4; - num_ps_stack_entries = 40; - num_vs_stack_entries = 40; - num_gs_stack_entries = 32; - num_es_stack_entries = 16; - break; - case CHIP_RV670: - num_ps_gprs = 144; - num_vs_gprs = 40; - num_temp_gprs = 4; - num_gs_gprs = 0; - num_es_gprs = 0; - num_ps_threads = 136; - num_vs_threads = 48; - num_gs_threads = 4; - num_es_threads = 4; - num_ps_stack_entries = 40; - num_vs_stack_entries = 40; - num_gs_stack_entries = 32; - num_es_stack_entries = 16; - break; - case CHIP_RV770: - num_ps_gprs = 192; - num_vs_gprs = 56; - num_temp_gprs = 4; - num_gs_gprs = 0; - num_es_gprs = 0; - num_ps_threads = 188; - num_vs_threads = 60; - num_gs_threads = 0; - num_es_threads = 0; - num_ps_stack_entries = 256; - num_vs_stack_entries = 256; - num_gs_stack_entries = 0; - num_es_stack_entries = 0; - break; - case CHIP_RV730: - case CHIP_RV740: - num_ps_gprs = 84; - num_vs_gprs = 36; - num_temp_gprs = 4; - num_gs_gprs = 0; - num_es_gprs = 0; - num_ps_threads = 188; - num_vs_threads = 60; - num_gs_threads = 0; - num_es_threads = 0; - num_ps_stack_entries = 128; - num_vs_stack_entries = 128; - num_gs_stack_entries = 0; - num_es_stack_entries = 0; - break; - case CHIP_RV710: - num_ps_gprs = 192; - num_vs_gprs = 56; - num_temp_gprs = 4; - num_gs_gprs = 0; - num_es_gprs = 0; - num_ps_threads = 144; - num_vs_threads = 48; - num_gs_threads = 0; - num_es_threads = 0; - num_ps_stack_entries = 128; - num_vs_stack_entries = 128; - num_gs_stack_entries = 0; - num_es_stack_entries = 0; - break; - } - - if ((rdev->family == CHIP_RV610) || - (rdev->family == CHIP_RV620) || - (rdev->family == CHIP_RS780) || - (rdev->family == CHIP_RS880) || - (rdev->family == CHIP_RV710)) - sq_config = 0; - else - sq_config = VC_ENABLE; - - sq_config |= (DX9_CONSTS | - ALU_INST_PREFER_VECTOR | - PS_PRIO(0) | - VS_PRIO(1) | - GS_PRIO(2) | - ES_PRIO(3)); - - sq_gpr_resource_mgmt_1 = (NUM_PS_GPRS(num_ps_gprs) | - NUM_VS_GPRS(num_vs_gprs) | - NUM_CLAUSE_TEMP_GPRS(num_temp_gprs)); - sq_gpr_resource_mgmt_2 = (NUM_GS_GPRS(num_gs_gprs) | - NUM_ES_GPRS(num_es_gprs)); - sq_thread_resource_mgmt = (NUM_PS_THREADS(num_ps_threads) | - NUM_VS_THREADS(num_vs_threads) | - NUM_GS_THREADS(num_gs_threads) | - NUM_ES_THREADS(num_es_threads)); - sq_stack_resource_mgmt_1 = (NUM_PS_STACK_ENTRIES(num_ps_stack_entries) | - NUM_VS_STACK_ENTRIES(num_vs_stack_entries)); - sq_stack_resource_mgmt_2 = (NUM_GS_STACK_ENTRIES(num_gs_stack_entries) | - NUM_ES_STACK_ENTRIES(num_es_stack_entries)); - - /* emit an IB pointing at default state */ - dwords = ALIGN(rdev->r600_blit.state_len, 0x10); - gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.state_offset; - radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); - radeon_ring_write(ring, -#ifdef __BIG_ENDIAN - (2 << 0) | -#endif - (gpu_addr & 0xFFFFFFFC)); - radeon_ring_write(ring, upper_32_bits(gpu_addr) & 0xFF); - radeon_ring_write(ring, dwords); - - /* SQ config */ - radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 6)); - radeon_ring_write(ring, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); - radeon_ring_write(ring, sq_config); - radeon_ring_write(ring, sq_gpr_resource_mgmt_1); - radeon_ring_write(ring, sq_gpr_resource_mgmt_2); - radeon_ring_write(ring, sq_thread_resource_mgmt); - radeon_ring_write(ring, sq_stack_resource_mgmt_1); - radeon_ring_write(ring, sq_stack_resource_mgmt_2); -} - -int r600_blit_init(struct radeon_device *rdev) -{ - u32 obj_size; - int i, r, dwords; - void *ptr; - u32 packet2s[16]; - int num_packet2s = 0; - - rdev->r600_blit.primitives.set_render_target = set_render_target; - rdev->r600_blit.primitives.cp_set_surface_sync = cp_set_surface_sync; - rdev->r600_blit.primitives.set_shaders = set_shaders; - rdev->r600_blit.primitives.set_vtx_resource = set_vtx_resource; - rdev->r600_blit.primitives.set_tex_resource = set_tex_resource; - rdev->r600_blit.primitives.set_scissors = set_scissors; - rdev->r600_blit.primitives.draw_auto = draw_auto; - rdev->r600_blit.primitives.set_default_state = set_default_state; - - rdev->r600_blit.ring_size_common = 8; /* sync semaphore */ - rdev->r600_blit.ring_size_common += 40; /* shaders + def state */ - rdev->r600_blit.ring_size_common += 5; /* done copy */ - rdev->r600_blit.ring_size_common += 16; /* fence emit for done copy */ - - rdev->r600_blit.ring_size_per_loop = 76; - /* set_render_target emits 2 extra dwords on rv6xx */ - if (rdev->family > CHIP_R600 && rdev->family < CHIP_RV770) - rdev->r600_blit.ring_size_per_loop += 2; - - rdev->r600_blit.max_dim = 8192; - - rdev->r600_blit.state_offset = 0; - - if (rdev->family >= CHIP_RV770) - rdev->r600_blit.state_len = r7xx_default_size; - else - rdev->r600_blit.state_len = r6xx_default_size; - - dwords = rdev->r600_blit.state_len; - while (dwords & 0xf) { - packet2s[num_packet2s++] = cpu_to_le32(PACKET2(0)); - dwords++; - } - - obj_size = dwords * 4; - obj_size = ALIGN(obj_size, 256); - - rdev->r600_blit.vs_offset = obj_size; - obj_size += r6xx_vs_size * 4; - obj_size = ALIGN(obj_size, 256); - - rdev->r600_blit.ps_offset = obj_size; - obj_size += r6xx_ps_size * 4; - obj_size = ALIGN(obj_size, 256); - - /* pin copy shader into vram if not already initialized */ - if (rdev->r600_blit.shader_obj == NULL) { - r = radeon_bo_create(rdev, obj_size, PAGE_SIZE, true, - RADEON_GEM_DOMAIN_VRAM, - NULL, &rdev->r600_blit.shader_obj); - if (r) { - DRM_ERROR("r600 failed to allocate shader\n"); - return r; - } - - r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); - if (unlikely(r != 0)) - return r; - r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM, - &rdev->r600_blit.shader_gpu_addr); - radeon_bo_unreserve(rdev->r600_blit.shader_obj); - if (r) { - dev_err(rdev->dev, "(%d) pin blit object failed\n", r); - return r; - } - } - - DRM_DEBUG("r6xx blit allocated bo %08x vs %08x ps %08x\n", - obj_size, - rdev->r600_blit.vs_offset, rdev->r600_blit.ps_offset); - - r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); - if (unlikely(r != 0)) - return r; - r = radeon_bo_kmap(rdev->r600_blit.shader_obj, &ptr); - if (r) { - DRM_ERROR("failed to map blit object %d\n", r); - return r; - } - if (rdev->family >= CHIP_RV770) - memcpy_toio(ptr + rdev->r600_blit.state_offset, - r7xx_default_state, rdev->r600_blit.state_len * 4); - else - memcpy_toio(ptr + rdev->r600_blit.state_offset, - r6xx_default_state, rdev->r600_blit.state_len * 4); - if (num_packet2s) - memcpy_toio(ptr + rdev->r600_blit.state_offset + (rdev->r600_blit.state_len * 4), - packet2s, num_packet2s * 4); - for (i = 0; i < r6xx_vs_size; i++) - *(u32 *)((unsigned long)ptr + rdev->r600_blit.vs_offset + i * 4) = cpu_to_le32(r6xx_vs[i]); - for (i = 0; i < r6xx_ps_size; i++) - *(u32 *)((unsigned long)ptr + rdev->r600_blit.ps_offset + i * 4) = cpu_to_le32(r6xx_ps[i]); - radeon_bo_kunmap(rdev->r600_blit.shader_obj); - radeon_bo_unreserve(rdev->r600_blit.shader_obj); - - radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); - return 0; -} - -void r600_blit_fini(struct radeon_device *rdev) -{ - int r; - - radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); - if (rdev->r600_blit.shader_obj == NULL) - return; - /* If we can't reserve the bo, unref should be enough to destroy - * it when it becomes idle. - */ - r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); - if (!r) { - radeon_bo_unpin(rdev->r600_blit.shader_obj); - radeon_bo_unreserve(rdev->r600_blit.shader_obj); - } - radeon_bo_unref(&rdev->r600_blit.shader_obj); -} - -static unsigned r600_blit_create_rect(unsigned num_gpu_pages, - int *width, int *height, int max_dim) -{ - unsigned max_pages; - unsigned pages = num_gpu_pages; - int w, h; - - if (num_gpu_pages == 0) { - /* not supposed to be called with no pages, but just in case */ - h = 0; - w = 0; - pages = 0; - WARN_ON(1); - } else { - int rect_order = 2; - h = RECT_UNIT_H; - while (num_gpu_pages / rect_order) { - h *= 2; - rect_order *= 4; - if (h >= max_dim) { - h = max_dim; - break; - } - } - max_pages = (max_dim * h) / (RECT_UNIT_W * RECT_UNIT_H); - if (pages > max_pages) - pages = max_pages; - w = (pages * RECT_UNIT_W * RECT_UNIT_H) / h; - w = (w / RECT_UNIT_W) * RECT_UNIT_W; - pages = (w * h) / (RECT_UNIT_W * RECT_UNIT_H); - BUG_ON(pages == 0); - } - - - DRM_DEBUG("blit_rectangle: h=%d, w=%d, pages=%d\n", h, w, pages); - - /* return width and height only of the caller wants it */ - if (height) - *height = h; - if (width) - *width = w; - - return pages; -} - - -int r600_blit_prepare_copy(struct radeon_device *rdev, unsigned num_gpu_pages, - struct radeon_fence **fence, struct radeon_sa_bo **vb, - struct radeon_semaphore **sem) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - int r; - int ring_size; - int num_loops = 0; - int dwords_per_loop = rdev->r600_blit.ring_size_per_loop; - - /* num loops */ - while (num_gpu_pages) { - num_gpu_pages -= - r600_blit_create_rect(num_gpu_pages, NULL, NULL, - rdev->r600_blit.max_dim); - num_loops++; - } - - /* 48 bytes for vertex per loop */ - r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, vb, - (num_loops*48)+256, 256, true); - if (r) { - return r; - } - - r = radeon_semaphore_create(rdev, sem); - if (r) { - radeon_sa_bo_free(rdev, vb, NULL); - return r; - } - - /* calculate number of loops correctly */ - ring_size = num_loops * dwords_per_loop; - ring_size += rdev->r600_blit.ring_size_common; - r = radeon_ring_lock(rdev, ring, ring_size); - if (r) { - radeon_sa_bo_free(rdev, vb, NULL); - radeon_semaphore_free(rdev, sem, NULL); - return r; - } - - if (radeon_fence_need_sync(*fence, RADEON_RING_TYPE_GFX_INDEX)) { - radeon_semaphore_sync_rings(rdev, *sem, (*fence)->ring, - RADEON_RING_TYPE_GFX_INDEX); - radeon_fence_note_sync(*fence, RADEON_RING_TYPE_GFX_INDEX); - } else { - radeon_semaphore_free(rdev, sem, NULL); - } - - rdev->r600_blit.primitives.set_default_state(rdev); - rdev->r600_blit.primitives.set_shaders(rdev); - return 0; -} - -void r600_blit_done_copy(struct radeon_device *rdev, struct radeon_fence **fence, - struct radeon_sa_bo *vb, struct radeon_semaphore *sem) -{ - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - int r; - - r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX); - if (r) { - radeon_ring_unlock_undo(rdev, ring); - return; - } - - radeon_ring_unlock_commit(rdev, ring); - radeon_sa_bo_free(rdev, &vb, *fence); - radeon_semaphore_free(rdev, &sem, *fence); -} - -void r600_kms_blit_copy(struct radeon_device *rdev, - u64 src_gpu_addr, u64 dst_gpu_addr, - unsigned num_gpu_pages, - struct radeon_sa_bo *vb) -{ - u64 vb_gpu_addr; - u32 *vb_cpu_addr; - - DRM_DEBUG("emitting copy %16llx %16llx %d\n", - src_gpu_addr, dst_gpu_addr, num_gpu_pages); - vb_cpu_addr = (u32 *)radeon_sa_bo_cpu_addr(vb); - vb_gpu_addr = radeon_sa_bo_gpu_addr(vb); - - while (num_gpu_pages) { - int w, h; - unsigned size_in_bytes; - unsigned pages_per_loop = - r600_blit_create_rect(num_gpu_pages, &w, &h, - rdev->r600_blit.max_dim); - - size_in_bytes = pages_per_loop * RADEON_GPU_PAGE_SIZE; - DRM_DEBUG("rectangle w=%d h=%d\n", w, h); - - vb_cpu_addr[0] = 0; - vb_cpu_addr[1] = 0; - vb_cpu_addr[2] = 0; - vb_cpu_addr[3] = 0; - - vb_cpu_addr[4] = 0; - vb_cpu_addr[5] = int2float(h); - vb_cpu_addr[6] = 0; - vb_cpu_addr[7] = int2float(h); - - vb_cpu_addr[8] = int2float(w); - vb_cpu_addr[9] = int2float(h); - vb_cpu_addr[10] = int2float(w); - vb_cpu_addr[11] = int2float(h); - - rdev->r600_blit.primitives.set_tex_resource(rdev, FMT_8_8_8_8, - w, h, w, src_gpu_addr, size_in_bytes); - rdev->r600_blit.primitives.set_render_target(rdev, COLOR_8_8_8_8, - w, h, dst_gpu_addr); - rdev->r600_blit.primitives.set_scissors(rdev, 0, 0, w, h); - rdev->r600_blit.primitives.set_vtx_resource(rdev, vb_gpu_addr); - rdev->r600_blit.primitives.draw_auto(rdev); - rdev->r600_blit.primitives.cp_set_surface_sync(rdev, - PACKET3_CB_ACTION_ENA | PACKET3_CB0_DEST_BASE_ENA, - size_in_bytes, dst_gpu_addr); - - vb_cpu_addr += 12; - vb_gpu_addr += 4*12; - src_gpu_addr += size_in_bytes; - dst_gpu_addr += size_in_bytes; - num_gpu_pages -= pages_per_loop; - } -} diff --git a/drivers/gpu/drm/radeon/r600_blit_shaders.h b/drivers/gpu/drm/radeon/r600_blit_shaders.h index 2f3ce7a7597..f437d36dd98 100644 --- a/drivers/gpu/drm/radeon/r600_blit_shaders.h +++ b/drivers/gpu/drm/radeon/r600_blit_shaders.h @@ -35,5 +35,4 @@ extern const u32 r6xx_default_state[]; extern const u32 r6xx_ps_size, r6xx_vs_size; extern const u32 r6xx_default_size, r7xx_default_size; -__pure uint32_t int2float(uint32_t x); #endif diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c index 1c51c08b1fd..8c9b7e26533 100644 --- a/drivers/gpu/drm/radeon/r600_cp.c +++ b/drivers/gpu/drm/radeon/r600_cp.c @@ -2200,13 +2200,13 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init, dev_priv->ring.end = ((u32 *) dev_priv->cp_ring->handle + init->ring_size / sizeof(u32)); dev_priv->ring.size = init->ring_size; - dev_priv->ring.size_l2qw = drm_order(init->ring_size / 8); + dev_priv->ring.size_l2qw = order_base_2(init->ring_size / 8); dev_priv->ring.rptr_update = /* init->rptr_update */ 4096; - dev_priv->ring.rptr_update_l2qw = drm_order(/* init->rptr_update */ 4096 / 8); + dev_priv->ring.rptr_update_l2qw = order_base_2(/* init->rptr_update */ 4096 / 8); dev_priv->ring.fetch_size = /* init->fetch_size */ 32; - dev_priv->ring.fetch_size_l2ow = drm_order(/* init->fetch_size */ 32 / 16); + dev_priv->ring.fetch_size_l2ow = order_base_2(/* init->fetch_size */ 32 / 16); dev_priv->ring.tail_mask = (dev_priv->ring.size / sizeof(u32)) - 1; @@ -2515,7 +2515,7 @@ int r600_cp_dispatch_texture(struct drm_device *dev, buf = radeon_freelist_get(dev); if (!buf) { DRM_DEBUG("EAGAIN\n"); - if (DRM_COPY_TO_USER(tex->image, image, sizeof(*image))) + if (copy_to_user(tex->image, image, sizeof(*image))) return -EFAULT; return -EAGAIN; } @@ -2528,7 +2528,7 @@ int r600_cp_dispatch_texture(struct drm_device *dev, buffer = (u32 *) ((char *)dev->agp_buffer_map->handle + buf->offset); - if (DRM_COPY_FROM_USER(buffer, data, pass_size)) { + if (copy_from_user(buffer, data, pass_size)) { DRM_ERROR("EFAULT on pad, %d bytes\n", pass_size); return -EFAULT; } diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c index 01a3ec83f28..12511bb5fd6 100644 --- a/drivers/gpu/drm/radeon/r600_cs.c +++ b/drivers/gpu/drm/radeon/r600_cs.c @@ -749,7 +749,10 @@ static int r600_cs_track_check(struct radeon_cs_parser *p) } for (i = 0; i < 8; i++) { - if ((tmp >> (i * 4)) & 0xF) { + u32 format = G_0280A0_FORMAT(track->cb_color_info[i]); + + if (format != V_0280A0_COLOR_INVALID && + (tmp >> (i * 4)) & 0xF) { /* at least one component is enabled */ if (track->cb_color_bo[i] == NULL) { dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n", @@ -887,7 +890,7 @@ int r600_cs_common_vline_parse(struct radeon_cs_parser *p, obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC); if (!obj) { DRM_ERROR("cannot find crtc %d\n", crtc_id); - return -EINVAL; + return -ENOENT; } crtc = obj_to_crtc(obj); radeon_crtc = to_radeon_crtc(crtc); @@ -1004,8 +1007,22 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case R_008C64_SQ_VSTMP_RING_SIZE: case R_0288C8_SQ_GS_VERT_ITEMSIZE: /* get value to populate the IB don't remove */ - tmp =radeon_get_ib_value(p, idx); - ib[idx] = 0; + /*tmp =radeon_get_ib_value(p, idx); + ib[idx] = 0;*/ + break; + case SQ_ESGS_RING_BASE: + case SQ_GSVS_RING_BASE: + case SQ_ESTMP_RING_BASE: + case SQ_GSTMP_RING_BASE: + case SQ_PSTMP_RING_BASE: + case SQ_VSTMP_RING_BASE: + r = radeon_cs_packet_next_reloc(p, &reloc, 0); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case SQ_CONFIG: track->sq_config = radeon_get_ib_value(p, idx); @@ -1026,7 +1043,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) track->db_depth_info = radeon_get_ib_value(p, idx); ib[idx] &= C_028010_ARRAY_MODE; track->db_depth_info &= C_028010_ARRAY_MODE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) { ib[idx] |= S_028010_ARRAY_MODE(V_028010_ARRAY_2D_TILED_THIN1); track->db_depth_info |= S_028010_ARRAY_MODE(V_028010_ARRAY_2D_TILED_THIN1); } else { @@ -1067,9 +1084,9 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = (reg - VGT_STRMOUT_BUFFER_BASE_0) / 16; track->vgt_strmout_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->vgt_strmout_bo[tmp] = reloc->robj; - track->vgt_strmout_bo_mc[tmp] = reloc->lobj.gpu_offset; + track->vgt_strmout_bo_mc[tmp] = reloc->gpu_offset; track->streamout_dirty = true; break; case VGT_STRMOUT_BUFFER_SIZE_0: @@ -1088,7 +1105,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case R_028238_CB_TARGET_MASK: track->cb_target_mask = radeon_get_ib_value(p, idx); @@ -1125,10 +1142,10 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = (reg - R_0280A0_CB_COLOR0_INFO) / 4; track->cb_color_info[tmp] = radeon_get_ib_value(p, idx); - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) { ib[idx] |= S_0280A0_ARRAY_MODE(V_0280A0_ARRAY_2D_TILED_THIN1); track->cb_color_info[tmp] |= S_0280A0_ARRAY_MODE(V_0280A0_ARRAY_2D_TILED_THIN1); - } else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) { + } else if (reloc->tiling_flags & RADEON_TILING_MICRO) { ib[idx] |= S_0280A0_ARRAY_MODE(V_0280A0_ARRAY_1D_TILED_THIN1); track->cb_color_info[tmp] |= S_0280A0_ARRAY_MODE(V_0280A0_ARRAY_1D_TILED_THIN1); } @@ -1197,7 +1214,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } track->cb_color_frag_bo[tmp] = reloc->robj; track->cb_color_frag_offset[tmp] = (u64)ib[idx] << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); } if (G_0280A0_TILE_MODE(track->cb_color_info[tmp])) { track->cb_dirty = true; @@ -1228,7 +1245,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } track->cb_color_tile_bo[tmp] = reloc->robj; track->cb_color_tile_offset[tmp] = (u64)ib[idx] << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); } if (G_0280A0_TILE_MODE(track->cb_color_info[tmp])) { track->cb_dirty = true; @@ -1264,10 +1281,10 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = (reg - CB_COLOR0_BASE) / 4; track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->cb_color_base_last[tmp] = ib[idx]; track->cb_color_bo[tmp] = reloc->robj; - track->cb_color_bo_mc[tmp] = reloc->lobj.gpu_offset; + track->cb_color_bo_mc[tmp] = reloc->gpu_offset; track->cb_dirty = true; break; case DB_DEPTH_BASE: @@ -1278,9 +1295,9 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->db_offset = radeon_get_ib_value(p, idx) << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->db_bo = reloc->robj; - track->db_bo_mc = reloc->lobj.gpu_offset; + track->db_bo_mc = reloc->gpu_offset; track->db_dirty = true; break; case DB_HTILE_DATA_BASE: @@ -1291,7 +1308,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->htile_offset = radeon_get_ib_value(p, idx) << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->htile_bo = reloc->robj; track->db_dirty = true; break; @@ -1360,7 +1377,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case SX_MEMORY_EXPORT_BASE: r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm); @@ -1369,7 +1386,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case SX_MISC: track->sx_misc_kill_all_prims = (radeon_get_ib_value(p, idx) & 0x1) != 0; @@ -1655,7 +1672,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (idx_value & 0xfffffff0) + ((u64)(tmp & 0xff) << 32); @@ -1696,7 +1713,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + idx_value + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); @@ -1748,7 +1765,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffff0) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -1788,7 +1805,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, tmp = radeon_get_ib_value(p, idx) + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); - offset = reloc->lobj.gpu_offset + tmp; + offset = reloc->gpu_offset + tmp; if ((tmp + size) > radeon_bo_size(reloc->robj)) { dev_warn(p->dev, "CP DMA src buffer too small (%llu %lu)\n", @@ -1818,7 +1835,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, tmp = radeon_get_ib_value(p, idx+2) + ((u64)(radeon_get_ib_value(p, idx+3) & 0xff) << 32); - offset = reloc->lobj.gpu_offset + tmp; + offset = reloc->gpu_offset + tmp; if ((tmp + size) > radeon_bo_size(reloc->robj)) { dev_warn(p->dev, "CP DMA dst buffer too small (%llu %lu)\n", @@ -1844,7 +1861,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SURFACE_SYNC\n"); return -EINVAL; } - ib[idx+2] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx+2] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); } break; case PACKET3_EVENT_WRITE: @@ -1860,7 +1877,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad EVENT_WRITE\n"); return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffff8) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -1882,7 +1899,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -1947,11 +1964,11 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SET_RESOURCE\n"); return -EINVAL; } - base_offset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + base_offset = (u32)((reloc->gpu_offset >> 8) & 0xffffffff); if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) ib[idx+1+(i*7)+0] |= S_038000_TILE_MODE(V_038000_ARRAY_2D_TILED_THIN1); - else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + else if (reloc->tiling_flags & RADEON_TILING_MICRO) ib[idx+1+(i*7)+0] |= S_038000_TILE_MODE(V_038000_ARRAY_1D_TILED_THIN1); } texture = reloc->robj; @@ -1961,13 +1978,13 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SET_RESOURCE\n"); return -EINVAL; } - mip_offset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + mip_offset = (u32)((reloc->gpu_offset >> 8) & 0xffffffff); mipmap = reloc->robj; r = r600_check_texture_resource(p, idx+(i*7)+1, texture, mipmap, base_offset + radeon_get_ib_value(p, idx+1+(i*7)+2), mip_offset + radeon_get_ib_value(p, idx+1+(i*7)+3), - reloc->lobj.tiling_flags); + reloc->tiling_flags); if (r) return r; ib[idx+1+(i*7)+2] += base_offset; @@ -1991,7 +2008,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, ib[idx+1+(i*7)+1] = radeon_bo_size(reloc->robj) - offset; } - offset64 = reloc->lobj.gpu_offset + offset; + offset64 = reloc->gpu_offset + offset; ib[idx+1+(i*8)+0] = offset64; ib[idx+1+(i*8)+2] = (ib[idx+1+(i*8)+2] & 0xffffff00) | (upper_32_bits(offset64) & 0xff); @@ -2101,7 +2118,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx+1] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); } break; case PACKET3_SURFACE_BASE_UPDATE: @@ -2134,7 +2151,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+1] = offset; ib[idx+2] = upper_32_bits(offset) & 0xff; } @@ -2153,7 +2170,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+3] = offset; ib[idx+4] = upper_32_bits(offset) & 0xff; } @@ -2182,7 +2199,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 8, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+0] = offset; ib[idx+1] = upper_32_bits(offset) & 0xff; break; @@ -2207,7 +2224,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+1] = offset; ib[idx+2] = upper_32_bits(offset) & 0xff; } else { @@ -2231,7 +2248,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+3] = offset; ib[idx+4] = upper_32_bits(offset) & 0xff; } else { @@ -2328,13 +2345,8 @@ static void r600_cs_parser_fini(struct radeon_cs_parser *parser, int error) unsigned i; kfree(parser->relocs); - for (i = 0; i < parser->nchunks; i++) { - kfree(parser->chunks[i].kdata); - if (parser->rdev && (parser->rdev->flags & RADEON_IS_AGP)) { - kfree(parser->chunks[i].kpage[0]); - kfree(parser->chunks[i].kpage[1]); - } - } + for (i = 0; i < parser->nchunks; i++) + drm_free_large(parser->chunks[i].kdata); kfree(parser->chunks); kfree(parser->chunks_array); } @@ -2391,13 +2403,12 @@ int r600_cs_legacy(struct drm_device *dev, void *data, struct drm_file *filp, ib_chunk = &parser.chunks[parser.chunk_ib_idx]; parser.ib.length_dw = ib_chunk->length_dw; *l = parser.ib.length_dw; - r = r600_cs_parse(&parser); - if (r) { - DRM_ERROR("Invalid command stream !\n"); + if (copy_from_user(ib, ib_chunk->user_ptr, ib_chunk->length_dw * 4)) { + r = -EFAULT; r600_cs_parser_fini(&parser, r); return r; } - r = radeon_cs_finish_pages(&parser); + r = r600_cs_parse(&parser); if (r) { DRM_ERROR("Invalid command stream !\n"); r600_cs_parser_fini(&parser, r); @@ -2494,14 +2505,14 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p) dst_offset = radeon_get_ib_value(p, idx+1); dst_offset <<= 8; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); p->idx += count + 5; } else { dst_offset = radeon_get_ib_value(p, idx+1); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; p->idx += count + 3; } if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) { @@ -2528,22 +2539,22 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p) /* tiled src, linear dst */ src_offset = radeon_get_ib_value(p, idx+1); src_offset <<= 8; - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(src_reloc->gpu_offset >> 8); dst_offset = radeon_get_ib_value(p, idx+5); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+6) & 0xff)) << 32; - ib[idx+5] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+6] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+5] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+6] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; } else { /* linear src, tiled dst */ src_offset = radeon_get_ib_value(p, idx+5); src_offset |= ((u64)(radeon_get_ib_value(p, idx+6) & 0xff)) << 32; - ib[idx+5] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+6] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+5] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+6] += upper_32_bits(src_reloc->gpu_offset) & 0xff; dst_offset = radeon_get_ib_value(p, idx+1); dst_offset <<= 8; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); } p->idx += 7; } else { @@ -2553,10 +2564,10 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p) dst_offset = radeon_get_ib_value(p, idx+1); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0xff)) << 32; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; - ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; + ib[idx+4] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 5; } else { src_offset = radeon_get_ib_value(p, idx+2); @@ -2564,10 +2575,10 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p) dst_offset = radeon_get_ib_value(p, idx+1); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0xff0000)) << 16; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; - ib[idx+3] += (upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff) << 16; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += upper_32_bits(src_reloc->gpu_offset) & 0xff; + ib[idx+3] += (upper_32_bits(dst_reloc->gpu_offset) & 0xff) << 16; p->idx += 4; } } @@ -2599,8 +2610,8 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p) dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += (upper_32_bits(dst_reloc->lobj.gpu_offset) << 16) & 0x00ff0000; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += (upper_32_bits(dst_reloc->gpu_offset) << 16) & 0x00ff0000; p->idx += 4; break; case DMA_PACKET_NOP: diff --git a/drivers/gpu/drm/radeon/r600_dma.c b/drivers/gpu/drm/radeon/r600_dma.c new file mode 100644 index 00000000000..4969cef44a1 --- /dev/null +++ b/drivers/gpu/drm/radeon/r600_dma.c @@ -0,0 +1,500 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "r600d.h" + +u32 r600_gpu_check_soft_reset(struct radeon_device *rdev); + +/* + * DMA + * Starting with R600, the GPU has an asynchronous + * DMA engine. The programming model is very similar + * to the 3D engine (ring buffer, IBs, etc.), but the + * DMA controller has it's own packet format that is + * different form the PM4 format used by the 3D engine. + * It supports copying data, writing embedded data, + * solid fills, and a number of other things. It also + * has support for tiling/detiling of buffers. + */ + +/** + * r600_dma_get_rptr - get the current read pointer + * + * @rdev: radeon_device pointer + * @ring: radeon ring pointer + * + * Get the current rptr from the hardware (r6xx+). + */ +uint32_t r600_dma_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + u32 rptr; + + if (rdev->wb.enabled) + rptr = rdev->wb.wb[ring->rptr_offs/4]; + else + rptr = RREG32(DMA_RB_RPTR); + + return (rptr & 0x3fffc) >> 2; +} + +/** + * r600_dma_get_wptr - get the current write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon ring pointer + * + * Get the current wptr from the hardware (r6xx+). + */ +uint32_t r600_dma_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + return (RREG32(DMA_RB_WPTR) & 0x3fffc) >> 2; +} + +/** + * r600_dma_set_wptr - commit the write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon ring pointer + * + * Write the wptr back to the hardware (r6xx+). + */ +void r600_dma_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + WREG32(DMA_RB_WPTR, (ring->wptr << 2) & 0x3fffc); +} + +/** + * r600_dma_stop - stop the async dma engine + * + * @rdev: radeon_device pointer + * + * Stop the async dma engine (r6xx-evergreen). + */ +void r600_dma_stop(struct radeon_device *rdev) +{ + u32 rb_cntl = RREG32(DMA_RB_CNTL); + + if (rdev->asic->copy.copy_ring_index == R600_RING_TYPE_DMA_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + + rb_cntl &= ~DMA_RB_ENABLE; + WREG32(DMA_RB_CNTL, rb_cntl); + + rdev->ring[R600_RING_TYPE_DMA_INDEX].ready = false; +} + +/** + * r600_dma_resume - setup and start the async dma engine + * + * @rdev: radeon_device pointer + * + * Set up the DMA ring buffer and enable it. (r6xx-evergreen). + * Returns 0 for success, error for failure. + */ +int r600_dma_resume(struct radeon_device *rdev) +{ + struct radeon_ring *ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; + u32 rb_cntl, dma_cntl, ib_cntl; + u32 rb_bufsz; + int r; + + /* Reset dma */ + if (rdev->family >= CHIP_RV770) + WREG32(SRBM_SOFT_RESET, RV770_SOFT_RESET_DMA); + else + WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA); + RREG32(SRBM_SOFT_RESET); + udelay(50); + WREG32(SRBM_SOFT_RESET, 0); + + WREG32(DMA_SEM_INCOMPLETE_TIMER_CNTL, 0); + WREG32(DMA_SEM_WAIT_FAIL_TIMER_CNTL, 0); + + /* Set ring buffer size in dwords */ + rb_bufsz = order_base_2(ring->ring_size / 4); + rb_cntl = rb_bufsz << 1; +#ifdef __BIG_ENDIAN + rb_cntl |= DMA_RB_SWAP_ENABLE | DMA_RPTR_WRITEBACK_SWAP_ENABLE; +#endif + WREG32(DMA_RB_CNTL, rb_cntl); + + /* Initialize the ring buffer's read and write pointers */ + WREG32(DMA_RB_RPTR, 0); + WREG32(DMA_RB_WPTR, 0); + + /* set the wb address whether it's enabled or not */ + WREG32(DMA_RB_RPTR_ADDR_HI, + upper_32_bits(rdev->wb.gpu_addr + R600_WB_DMA_RPTR_OFFSET) & 0xFF); + WREG32(DMA_RB_RPTR_ADDR_LO, + ((rdev->wb.gpu_addr + R600_WB_DMA_RPTR_OFFSET) & 0xFFFFFFFC)); + + if (rdev->wb.enabled) + rb_cntl |= DMA_RPTR_WRITEBACK_ENABLE; + + WREG32(DMA_RB_BASE, ring->gpu_addr >> 8); + + /* enable DMA IBs */ + ib_cntl = DMA_IB_ENABLE; +#ifdef __BIG_ENDIAN + ib_cntl |= DMA_IB_SWAP_ENABLE; +#endif + WREG32(DMA_IB_CNTL, ib_cntl); + + dma_cntl = RREG32(DMA_CNTL); + dma_cntl &= ~CTXEMPTY_INT_ENABLE; + WREG32(DMA_CNTL, dma_cntl); + + if (rdev->family >= CHIP_RV770) + WREG32(DMA_MODE, 1); + + ring->wptr = 0; + WREG32(DMA_RB_WPTR, ring->wptr << 2); + + WREG32(DMA_RB_CNTL, rb_cntl | DMA_RB_ENABLE); + + ring->ready = true; + + r = radeon_ring_test(rdev, R600_RING_TYPE_DMA_INDEX, ring); + if (r) { + ring->ready = false; + return r; + } + + if (rdev->asic->copy.copy_ring_index == R600_RING_TYPE_DMA_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); + + return 0; +} + +/** + * r600_dma_fini - tear down the async dma engine + * + * @rdev: radeon_device pointer + * + * Stop the async dma engine and free the ring (r6xx-evergreen). + */ +void r600_dma_fini(struct radeon_device *rdev) +{ + r600_dma_stop(rdev); + radeon_ring_fini(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX]); +} + +/** + * r600_dma_is_lockup - Check if the DMA engine is locked up + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * + * Check if the async DMA engine is locked up. + * Returns true if the engine appears to be locked up, false if not. + */ +bool r600_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) +{ + u32 reset_mask = r600_gpu_check_soft_reset(rdev); + + if (!(reset_mask & RADEON_RESET_DMA)) { + radeon_ring_lockup_update(rdev, ring); + return false; + } + return radeon_ring_test_lockup(rdev, ring); +} + + +/** + * r600_dma_ring_test - simple async dma engine test + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * + * Test the DMA engine by writing using it to write an + * value to memory. (r6xx-SI). + * Returns 0 for success, error for failure. + */ +int r600_dma_ring_test(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + unsigned i; + int r; + void __iomem *ptr = (void *)rdev->vram_scratch.ptr; + u32 tmp; + + if (!ptr) { + DRM_ERROR("invalid vram scratch pointer\n"); + return -EINVAL; + } + + tmp = 0xCAFEDEAD; + writel(tmp, ptr); + + r = radeon_ring_lock(rdev, ring, 4); + if (r) { + DRM_ERROR("radeon: dma failed to lock ring %d (%d).\n", ring->idx, r); + return r; + } + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1)); + radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc); + radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xff); + radeon_ring_write(ring, 0xDEADBEEF); + radeon_ring_unlock_commit(rdev, ring); + + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = readl(ptr); + if (tmp == 0xDEADBEEF) + break; + DRM_UDELAY(1); + } + + if (i < rdev->usec_timeout) { + DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); + } else { + DRM_ERROR("radeon: ring %d test failed (0x%08X)\n", + ring->idx, tmp); + r = -EINVAL; + } + return r; +} + +/** + * r600_dma_fence_ring_emit - emit a fence on the DMA ring + * + * @rdev: radeon_device pointer + * @fence: radeon fence object + * + * Add a DMA fence packet to the ring to write + * the fence seq number and DMA trap packet to generate + * an interrupt if needed (r6xx-r7xx). + */ +void r600_dma_fence_ring_emit(struct radeon_device *rdev, + struct radeon_fence *fence) +{ + struct radeon_ring *ring = &rdev->ring[fence->ring]; + u64 addr = rdev->fence_drv[fence->ring].gpu_addr; + + /* write the fence */ + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_FENCE, 0, 0, 0)); + radeon_ring_write(ring, addr & 0xfffffffc); + radeon_ring_write(ring, (upper_32_bits(addr) & 0xff)); + radeon_ring_write(ring, lower_32_bits(fence->seq)); + /* generate an interrupt */ + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_TRAP, 0, 0, 0)); +} + +/** + * r600_dma_semaphore_ring_emit - emit a semaphore on the dma ring + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * @semaphore: radeon semaphore object + * @emit_wait: wait or signal semaphore + * + * Add a DMA semaphore packet to the ring wait on or signal + * other rings (r6xx-SI). + */ +bool r600_dma_semaphore_ring_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait) +{ + u64 addr = semaphore->gpu_addr; + u32 s = emit_wait ? 0 : 1; + + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SEMAPHORE, 0, s, 0)); + radeon_ring_write(ring, addr & 0xfffffffc); + radeon_ring_write(ring, upper_32_bits(addr) & 0xff); + + return true; +} + +/** + * r600_dma_ib_test - test an IB on the DMA engine + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * + * Test a simple IB in the DMA ring (r6xx-SI). + * Returns 0 on success, error on failure. + */ +int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) +{ + struct radeon_ib ib; + unsigned i; + int r; + void __iomem *ptr = (void *)rdev->vram_scratch.ptr; + u32 tmp = 0; + + if (!ptr) { + DRM_ERROR("invalid vram scratch pointer\n"); + return -EINVAL; + } + + tmp = 0xCAFEDEAD; + writel(tmp, ptr); + + r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256); + if (r) { + DRM_ERROR("radeon: failed to get ib (%d).\n", r); + return r; + } + + ib.ptr[0] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1); + ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc; + ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xff; + ib.ptr[3] = 0xDEADBEEF; + ib.length_dw = 4; + + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + radeon_ib_free(rdev, &ib); + DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); + return r; + } + r = radeon_fence_wait(ib.fence, false); + if (r) { + DRM_ERROR("radeon: fence wait failed (%d).\n", r); + return r; + } + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = readl(ptr); + if (tmp == 0xDEADBEEF) + break; + DRM_UDELAY(1); + } + if (i < rdev->usec_timeout) { + DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i); + } else { + DRM_ERROR("radeon: ib test failed (0x%08X)\n", tmp); + r = -EINVAL; + } + radeon_ib_free(rdev, &ib); + return r; +} + +/** + * r600_dma_ring_ib_execute - Schedule an IB on the DMA engine + * + * @rdev: radeon_device pointer + * @ib: IB object to schedule + * + * Schedule an IB in the DMA ring (r6xx-r7xx). + */ +void r600_dma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) +{ + struct radeon_ring *ring = &rdev->ring[ib->ring]; + + if (rdev->wb.enabled) { + u32 next_rptr = ring->wptr + 4; + while ((next_rptr & 7) != 5) + next_rptr++; + next_rptr += 3; + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1)); + radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); + radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xff); + radeon_ring_write(ring, next_rptr); + } + + /* The indirect buffer packet must end on an 8 DW boundary in the DMA ring. + * Pad as necessary with NOPs. + */ + while ((ring->wptr & 7) != 5) + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_INDIRECT_BUFFER, 0, 0, 0)); + radeon_ring_write(ring, (ib->gpu_addr & 0xFFFFFFE0)); + radeon_ring_write(ring, (ib->length_dw << 16) | (upper_32_bits(ib->gpu_addr) & 0xFF)); + +} + +/** + * r600_copy_dma - copy pages using the DMA engine + * + * @rdev: radeon_device pointer + * @src_offset: src GPU address + * @dst_offset: dst GPU address + * @num_gpu_pages: number of GPU pages to xfer + * @fence: radeon fence object + * + * Copy GPU paging using the DMA engine (r6xx). + * Used by the radeon ttm implementation to move pages if + * registered as the asic copy callback. + */ +int r600_copy_dma(struct radeon_device *rdev, + uint64_t src_offset, uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence **fence) +{ + struct radeon_semaphore *sem = NULL; + int ring_index = rdev->asic->copy.dma_ring_index; + struct radeon_ring *ring = &rdev->ring[ring_index]; + u32 size_in_dw, cur_size_in_dw; + int i, num_loops; + int r = 0; + + r = radeon_semaphore_create(rdev, &sem); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + return r; + } + + size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4; + num_loops = DIV_ROUND_UP(size_in_dw, 0xFFFE); + r = radeon_ring_lock(rdev, ring, num_loops * 4 + 8); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_semaphore_sync_to(sem, *fence); + radeon_semaphore_sync_rings(rdev, sem, ring->idx); + + for (i = 0; i < num_loops; i++) { + cur_size_in_dw = size_in_dw; + if (cur_size_in_dw > 0xFFFE) + cur_size_in_dw = 0xFFFE; + size_in_dw -= cur_size_in_dw; + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 0, 0, cur_size_in_dw)); + radeon_ring_write(ring, dst_offset & 0xfffffffc); + radeon_ring_write(ring, src_offset & 0xfffffffc); + radeon_ring_write(ring, (((upper_32_bits(dst_offset) & 0xff) << 16) | + (upper_32_bits(src_offset) & 0xff))); + src_offset += cur_size_in_dw * 4; + dst_offset += cur_size_in_dw * 4; + } + + r = radeon_fence_emit(rdev, fence, ring->idx); + if (r) { + radeon_ring_unlock_undo(rdev, ring); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_ring_unlock_commit(rdev, ring); + radeon_semaphore_free(rdev, &sem, *fence); + + return r; +} diff --git a/drivers/gpu/drm/radeon/r600_dpm.c b/drivers/gpu/drm/radeon/r600_dpm.c index b88f54b134a..9c61b74ef44 100644 --- a/drivers/gpu/drm/radeon/r600_dpm.c +++ b/drivers/gpu/drm/radeon/r600_dpm.c @@ -158,22 +158,43 @@ u32 r600_dpm_get_vblank_time(struct radeon_device *rdev) u32 line_time_us, vblank_lines; u32 vblank_time_us = 0xffffffff; /* if the displays are off, vblank time is max */ - list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { - radeon_crtc = to_radeon_crtc(crtc); - if (crtc->enabled && radeon_crtc->enabled && radeon_crtc->hw_mode.clock) { - line_time_us = (radeon_crtc->hw_mode.crtc_htotal * 1000) / - radeon_crtc->hw_mode.clock; - vblank_lines = radeon_crtc->hw_mode.crtc_vblank_end - - radeon_crtc->hw_mode.crtc_vdisplay + - (radeon_crtc->v_border * 2); - vblank_time_us = vblank_lines * line_time_us; - break; + if (rdev->num_crtc && rdev->mode_info.mode_config_initialized) { + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { + radeon_crtc = to_radeon_crtc(crtc); + if (crtc->enabled && radeon_crtc->enabled && radeon_crtc->hw_mode.clock) { + line_time_us = (radeon_crtc->hw_mode.crtc_htotal * 1000) / + radeon_crtc->hw_mode.clock; + vblank_lines = radeon_crtc->hw_mode.crtc_vblank_end - + radeon_crtc->hw_mode.crtc_vdisplay + + (radeon_crtc->v_border * 2); + vblank_time_us = vblank_lines * line_time_us; + break; + } } } return vblank_time_us; } +u32 r600_dpm_get_vrefresh(struct radeon_device *rdev) +{ + struct drm_device *dev = rdev->ddev; + struct drm_crtc *crtc; + struct radeon_crtc *radeon_crtc; + u32 vrefresh = 0; + + if (rdev->num_crtc && rdev->mode_info.mode_config_initialized) { + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { + radeon_crtc = to_radeon_crtc(crtc); + if (crtc->enabled && radeon_crtc->enabled && radeon_crtc->hw_mode.clock) { + vrefresh = radeon_crtc->hw_mode.vrefresh; + break; + } + } + } + return vrefresh; +} + void r600_calculate_u_and_p(u32 i, u32 r_c, u32 p_b, u32 *p, u32 *u) { @@ -278,9 +299,9 @@ bool r600_dynamicpm_enabled(struct radeon_device *rdev) void r600_enable_sclk_control(struct radeon_device *rdev, bool enable) { if (enable) - WREG32_P(GENERAL_PWRMGT, 0, ~SCLK_PWRMGT_OFF); + WREG32_P(SCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_OFF); else - WREG32_P(GENERAL_PWRMGT, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF); + WREG32_P(SCLK_PWRMGT_CNTL, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF); } void r600_enable_mclk_control(struct radeon_device *rdev, bool enable) @@ -711,8 +732,8 @@ bool r600_is_uvd_state(u32 class, u32 class2) return false; } -int r600_set_thermal_temperature_range(struct radeon_device *rdev, - int min_temp, int max_temp) +static int r600_set_thermal_temperature_range(struct radeon_device *rdev, + int min_temp, int max_temp) { int low_temp = 0 * 1000; int high_temp = 255 * 1000; @@ -745,6 +766,8 @@ bool r600_is_internal_thermal_sensor(enum radeon_int_thermal_type sensor) case THERMAL_TYPE_SUMO: case THERMAL_TYPE_NI: case THERMAL_TYPE_SI: + case THERMAL_TYPE_CI: + case THERMAL_TYPE_KV: return true; case THERMAL_TYPE_ADT7473_WITH_INTERNAL: case THERMAL_TYPE_EMC2103_WITH_INTERNAL: @@ -757,6 +780,22 @@ bool r600_is_internal_thermal_sensor(enum radeon_int_thermal_type sensor) } } +int r600_dpm_late_enable(struct radeon_device *rdev) +{ + int ret; + + if (rdev->irq.installed && + r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { + ret = r600_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX); + if (ret) + return ret; + rdev->irq.dpm_thermal = true; + radeon_irq_set(rdev); + } + + return 0; +} + union power_info { struct _ATOM_POWERPLAY_INFO info; struct _ATOM_POWERPLAY_INFO_V2 info_2; @@ -779,21 +818,45 @@ static int r600_parse_clk_voltage_dep_table(struct radeon_clock_voltage_dependen u32 size = atom_table->ucNumEntries * sizeof(struct radeon_clock_voltage_dependency_entry); int i; + ATOM_PPLIB_Clock_Voltage_Dependency_Record *entry; radeon_table->entries = kzalloc(size, GFP_KERNEL); if (!radeon_table->entries) return -ENOMEM; + entry = &atom_table->entries[0]; for (i = 0; i < atom_table->ucNumEntries; i++) { - radeon_table->entries[i].clk = le16_to_cpu(atom_table->entries[i].usClockLow) | - (atom_table->entries[i].ucClockHigh << 16); - radeon_table->entries[i].v = le16_to_cpu(atom_table->entries[i].usVoltage); + radeon_table->entries[i].clk = le16_to_cpu(entry->usClockLow) | + (entry->ucClockHigh << 16); + radeon_table->entries[i].v = le16_to_cpu(entry->usVoltage); + entry = (ATOM_PPLIB_Clock_Voltage_Dependency_Record *) + ((u8 *)entry + sizeof(ATOM_PPLIB_Clock_Voltage_Dependency_Record)); } radeon_table->count = atom_table->ucNumEntries; return 0; } +int r600_get_platform_caps(struct radeon_device *rdev) +{ + struct radeon_mode_info *mode_info = &rdev->mode_info; + union power_info *power_info; + int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo); + u16 data_offset; + u8 frev, crev; + + if (!atom_parse_data_header(mode_info->atom_context, index, NULL, + &frev, &crev, &data_offset)) + return -EINVAL; + power_info = (union power_info *)(mode_info->atom_context->bios + data_offset); + + rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); + rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); + rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); + + return 0; +} + /* sizeof(ATOM_PPLIB_EXTENDEDHEADER) */ #define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V2 12 #define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V3 14 @@ -875,6 +938,19 @@ int r600_parse_extended_power_table(struct radeon_device *rdev) return ret; } } + if (power_info->pplib4.usMvddDependencyOnMCLKOffset) { + dep_table = (ATOM_PPLIB_Clock_Voltage_Dependency_Table *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(power_info->pplib4.usMvddDependencyOnMCLKOffset)); + ret = r600_parse_clk_voltage_dep_table(&rdev->pm.dpm.dyn_state.mvdd_dependency_on_mclk, + dep_table); + if (ret) { + kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries); + kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries); + kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries); + return ret; + } + } if (power_info->pplib4.usMaxClockVoltageOnDCOffset) { ATOM_PPLIB_Clock_Voltage_Limit_Table *clk_v = (ATOM_PPLIB_Clock_Voltage_Limit_Table *) @@ -898,27 +974,27 @@ int r600_parse_extended_power_table(struct radeon_device *rdev) (ATOM_PPLIB_PhaseSheddingLimits_Table *) (mode_info->atom_context->bios + data_offset + le16_to_cpu(power_info->pplib4.usVddcPhaseShedLimitsTableOffset)); + ATOM_PPLIB_PhaseSheddingLimits_Record *entry; rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries = kzalloc(psl->ucNumEntries * sizeof(struct radeon_phase_shedding_limits_entry), GFP_KERNEL); if (!rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries) { - kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries); - kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries); - kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries); + r600_free_extended_power_table(rdev); return -ENOMEM; } + entry = &psl->entries[0]; for (i = 0; i < psl->ucNumEntries; i++) { rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries[i].sclk = - le16_to_cpu(psl->entries[i].usSclkLow) | - (psl->entries[i].ucSclkHigh << 16); + le16_to_cpu(entry->usSclkLow) | (entry->ucSclkHigh << 16); rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries[i].mclk = - le16_to_cpu(psl->entries[i].usMclkLow) | - (psl->entries[i].ucMclkHigh << 16); + le16_to_cpu(entry->usMclkLow) | (entry->ucMclkHigh << 16); rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries[i].voltage = - le16_to_cpu(psl->entries[i].usVoltage); + le16_to_cpu(entry->usVoltage); + entry = (ATOM_PPLIB_PhaseSheddingLimits_Record *) + ((u8 *)entry + sizeof(ATOM_PPLIB_PhaseSheddingLimits_Record)); } rdev->pm.dpm.dyn_state.phase_shedding_limits_table.count = psl->ucNumEntries; @@ -945,30 +1021,166 @@ int r600_parse_extended_power_table(struct radeon_device *rdev) (ATOM_PPLIB_CAC_Leakage_Table *) (mode_info->atom_context->bios + data_offset + le16_to_cpu(power_info->pplib5.usCACLeakageTableOffset)); + ATOM_PPLIB_CAC_Leakage_Record *entry; u32 size = cac_table->ucNumEntries * sizeof(struct radeon_cac_leakage_table); rdev->pm.dpm.dyn_state.cac_leakage_table.entries = kzalloc(size, GFP_KERNEL); if (!rdev->pm.dpm.dyn_state.cac_leakage_table.entries) { - kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries); - kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries); - kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries); + r600_free_extended_power_table(rdev); return -ENOMEM; } + entry = &cac_table->entries[0]; for (i = 0; i < cac_table->ucNumEntries; i++) { - rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc = - le16_to_cpu(cac_table->entries[i].usVddc); - rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].leakage = - le32_to_cpu(cac_table->entries[i].ulLeakageValue); + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_EVV) { + rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc1 = + le16_to_cpu(entry->usVddc1); + rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc2 = + le16_to_cpu(entry->usVddc2); + rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc3 = + le16_to_cpu(entry->usVddc3); + } else { + rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc = + le16_to_cpu(entry->usVddc); + rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].leakage = + le32_to_cpu(entry->ulLeakageValue); + } + entry = (ATOM_PPLIB_CAC_Leakage_Record *) + ((u8 *)entry + sizeof(ATOM_PPLIB_CAC_Leakage_Record)); } rdev->pm.dpm.dyn_state.cac_leakage_table.count = cac_table->ucNumEntries; } } - /* ppm table */ + /* ext tables */ if (le16_to_cpu(power_info->pplib.usTableSize) >= sizeof(struct _ATOM_PPLIB_POWERPLAYTABLE3)) { ATOM_PPLIB_EXTENDEDHEADER *ext_hdr = (ATOM_PPLIB_EXTENDEDHEADER *) (mode_info->atom_context->bios + data_offset + le16_to_cpu(power_info->pplib3.usExtendendedHeaderOffset)); + if ((le16_to_cpu(ext_hdr->usSize) >= SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V2) && + ext_hdr->usVCETableOffset) { + VCEClockInfoArray *array = (VCEClockInfoArray *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usVCETableOffset) + 1); + ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table *limits = + (ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usVCETableOffset) + 1 + + 1 + array->ucNumEntries * sizeof(VCEClockInfo)); + ATOM_PPLIB_VCE_State_Table *states = + (ATOM_PPLIB_VCE_State_Table *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usVCETableOffset) + 1 + + 1 + (array->ucNumEntries * sizeof (VCEClockInfo)) + + 1 + (limits->numEntries * sizeof(ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record))); + ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record *entry; + ATOM_PPLIB_VCE_State_Record *state_entry; + VCEClockInfo *vce_clk; + u32 size = limits->numEntries * + sizeof(struct radeon_vce_clock_voltage_dependency_entry); + rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries = + kzalloc(size, GFP_KERNEL); + if (!rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries) { + r600_free_extended_power_table(rdev); + return -ENOMEM; + } + rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.count = + limits->numEntries; + entry = &limits->entries[0]; + state_entry = &states->entries[0]; + for (i = 0; i < limits->numEntries; i++) { + vce_clk = (VCEClockInfo *) + ((u8 *)&array->entries[0] + + (entry->ucVCEClockInfoIndex * sizeof(VCEClockInfo))); + rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries[i].evclk = + le16_to_cpu(vce_clk->usEVClkLow) | (vce_clk->ucEVClkHigh << 16); + rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries[i].ecclk = + le16_to_cpu(vce_clk->usECClkLow) | (vce_clk->ucECClkHigh << 16); + rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries[i].v = + le16_to_cpu(entry->usVoltage); + entry = (ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record *) + ((u8 *)entry + sizeof(ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record)); + } + for (i = 0; i < states->numEntries; i++) { + if (i >= RADEON_MAX_VCE_LEVELS) + break; + vce_clk = (VCEClockInfo *) + ((u8 *)&array->entries[0] + + (state_entry->ucVCEClockInfoIndex * sizeof(VCEClockInfo))); + rdev->pm.dpm.vce_states[i].evclk = + le16_to_cpu(vce_clk->usEVClkLow) | (vce_clk->ucEVClkHigh << 16); + rdev->pm.dpm.vce_states[i].ecclk = + le16_to_cpu(vce_clk->usECClkLow) | (vce_clk->ucECClkHigh << 16); + rdev->pm.dpm.vce_states[i].clk_idx = + state_entry->ucClockInfoIndex & 0x3f; + rdev->pm.dpm.vce_states[i].pstate = + (state_entry->ucClockInfoIndex & 0xc0) >> 6; + state_entry = (ATOM_PPLIB_VCE_State_Record *) + ((u8 *)state_entry + sizeof(ATOM_PPLIB_VCE_State_Record)); + } + } + if ((le16_to_cpu(ext_hdr->usSize) >= SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V3) && + ext_hdr->usUVDTableOffset) { + UVDClockInfoArray *array = (UVDClockInfoArray *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usUVDTableOffset) + 1); + ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table *limits = + (ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usUVDTableOffset) + 1 + + 1 + (array->ucNumEntries * sizeof (UVDClockInfo))); + ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record *entry; + u32 size = limits->numEntries * + sizeof(struct radeon_uvd_clock_voltage_dependency_entry); + rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries = + kzalloc(size, GFP_KERNEL); + if (!rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries) { + r600_free_extended_power_table(rdev); + return -ENOMEM; + } + rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.count = + limits->numEntries; + entry = &limits->entries[0]; + for (i = 0; i < limits->numEntries; i++) { + UVDClockInfo *uvd_clk = (UVDClockInfo *) + ((u8 *)&array->entries[0] + + (entry->ucUVDClockInfoIndex * sizeof(UVDClockInfo))); + rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[i].vclk = + le16_to_cpu(uvd_clk->usVClkLow) | (uvd_clk->ucVClkHigh << 16); + rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[i].dclk = + le16_to_cpu(uvd_clk->usDClkLow) | (uvd_clk->ucDClkHigh << 16); + rdev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[i].v = + le16_to_cpu(entry->usVoltage); + entry = (ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record *) + ((u8 *)entry + sizeof(ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record)); + } + } + if ((le16_to_cpu(ext_hdr->usSize) >= SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V4) && + ext_hdr->usSAMUTableOffset) { + ATOM_PPLIB_SAMClk_Voltage_Limit_Table *limits = + (ATOM_PPLIB_SAMClk_Voltage_Limit_Table *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usSAMUTableOffset) + 1); + ATOM_PPLIB_SAMClk_Voltage_Limit_Record *entry; + u32 size = limits->numEntries * + sizeof(struct radeon_clock_voltage_dependency_entry); + rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.entries = + kzalloc(size, GFP_KERNEL); + if (!rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.entries) { + r600_free_extended_power_table(rdev); + return -ENOMEM; + } + rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.count = + limits->numEntries; + entry = &limits->entries[0]; + for (i = 0; i < limits->numEntries; i++) { + rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.entries[i].clk = + le16_to_cpu(entry->usSAMClockLow) | (entry->ucSAMClockHigh << 16); + rdev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.entries[i].v = + le16_to_cpu(entry->usVoltage); + entry = (ATOM_PPLIB_SAMClk_Voltage_Limit_Record *) + ((u8 *)entry + sizeof(ATOM_PPLIB_SAMClk_Voltage_Limit_Record)); + } + } if ((le16_to_cpu(ext_hdr->usSize) >= SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V5) && ext_hdr->usPPMTableOffset) { ATOM_PPLIB_PPM_Table *ppm = (ATOM_PPLIB_PPM_Table *) @@ -977,10 +1189,7 @@ int r600_parse_extended_power_table(struct radeon_device *rdev) rdev->pm.dpm.dyn_state.ppm_table = kzalloc(sizeof(struct radeon_ppm_table), GFP_KERNEL); if (!rdev->pm.dpm.dyn_state.ppm_table) { - kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries); - kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries); - kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries); - kfree(rdev->pm.dpm.dyn_state.cac_leakage_table.entries); + r600_free_extended_power_table(rdev); return -ENOMEM; } rdev->pm.dpm.dyn_state.ppm_table->ppm_design = ppm->ucPpmDesign; @@ -1003,6 +1212,71 @@ int r600_parse_extended_power_table(struct radeon_device *rdev) rdev->pm.dpm.dyn_state.ppm_table->tj_max = le32_to_cpu(ppm->ulTjmax); } + if ((le16_to_cpu(ext_hdr->usSize) >= SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V6) && + ext_hdr->usACPTableOffset) { + ATOM_PPLIB_ACPClk_Voltage_Limit_Table *limits = + (ATOM_PPLIB_ACPClk_Voltage_Limit_Table *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usACPTableOffset) + 1); + ATOM_PPLIB_ACPClk_Voltage_Limit_Record *entry; + u32 size = limits->numEntries * + sizeof(struct radeon_clock_voltage_dependency_entry); + rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.entries = + kzalloc(size, GFP_KERNEL); + if (!rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.entries) { + r600_free_extended_power_table(rdev); + return -ENOMEM; + } + rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.count = + limits->numEntries; + entry = &limits->entries[0]; + for (i = 0; i < limits->numEntries; i++) { + rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.entries[i].clk = + le16_to_cpu(entry->usACPClockLow) | (entry->ucACPClockHigh << 16); + rdev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.entries[i].v = + le16_to_cpu(entry->usVoltage); + entry = (ATOM_PPLIB_ACPClk_Voltage_Limit_Record *) + ((u8 *)entry + sizeof(ATOM_PPLIB_ACPClk_Voltage_Limit_Record)); + } + } + if ((le16_to_cpu(ext_hdr->usSize) >= SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V7) && + ext_hdr->usPowerTuneTableOffset) { + u8 rev = *(u8 *)(mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usPowerTuneTableOffset)); + ATOM_PowerTune_Table *pt; + rdev->pm.dpm.dyn_state.cac_tdp_table = + kzalloc(sizeof(struct radeon_cac_tdp_table), GFP_KERNEL); + if (!rdev->pm.dpm.dyn_state.cac_tdp_table) { + r600_free_extended_power_table(rdev); + return -ENOMEM; + } + if (rev > 0) { + ATOM_PPLIB_POWERTUNE_Table_V1 *ppt = (ATOM_PPLIB_POWERTUNE_Table_V1 *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usPowerTuneTableOffset)); + rdev->pm.dpm.dyn_state.cac_tdp_table->maximum_power_delivery_limit = + ppt->usMaximumPowerDeliveryLimit; + pt = &ppt->power_tune_table; + } else { + ATOM_PPLIB_POWERTUNE_Table *ppt = (ATOM_PPLIB_POWERTUNE_Table *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usPowerTuneTableOffset)); + rdev->pm.dpm.dyn_state.cac_tdp_table->maximum_power_delivery_limit = 255; + pt = &ppt->power_tune_table; + } + rdev->pm.dpm.dyn_state.cac_tdp_table->tdp = le16_to_cpu(pt->usTDP); + rdev->pm.dpm.dyn_state.cac_tdp_table->configurable_tdp = + le16_to_cpu(pt->usConfigurableTDP); + rdev->pm.dpm.dyn_state.cac_tdp_table->tdc = le16_to_cpu(pt->usTDC); + rdev->pm.dpm.dyn_state.cac_tdp_table->battery_power_limit = + le16_to_cpu(pt->usBatteryPowerLimit); + rdev->pm.dpm.dyn_state.cac_tdp_table->small_power_limit = + le16_to_cpu(pt->usSmallPowerLimit); + rdev->pm.dpm.dyn_state.cac_tdp_table->low_cac_leakage = + le16_to_cpu(pt->usLowCACLeakage); + rdev->pm.dpm.dyn_state.cac_tdp_table->high_cac_leakage = + le16_to_cpu(pt->usHighCACLeakage); + } } return 0; @@ -1010,18 +1284,20 @@ int r600_parse_extended_power_table(struct radeon_device *rdev) void r600_free_extended_power_table(struct radeon_device *rdev) { - if (rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries) - kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries); - if (rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries) - kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries); - if (rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries) - kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries); - if (rdev->pm.dpm.dyn_state.cac_leakage_table.entries) - kfree(rdev->pm.dpm.dyn_state.cac_leakage_table.entries); - if (rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries) - kfree(rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries); - if (rdev->pm.dpm.dyn_state.ppm_table) - kfree(rdev->pm.dpm.dyn_state.ppm_table); + struct radeon_dpm_dynamic_state *dyn_state = &rdev->pm.dpm.dyn_state; + + kfree(dyn_state->vddc_dependency_on_sclk.entries); + kfree(dyn_state->vddci_dependency_on_mclk.entries); + kfree(dyn_state->vddc_dependency_on_mclk.entries); + kfree(dyn_state->mvdd_dependency_on_mclk.entries); + kfree(dyn_state->cac_leakage_table.entries); + kfree(dyn_state->phase_shedding_limits_table.entries); + kfree(dyn_state->ppm_table); + kfree(dyn_state->cac_tdp_table); + kfree(dyn_state->vce_clock_voltage_dependency_table.entries); + kfree(dyn_state->uvd_clock_voltage_dependency_table.entries); + kfree(dyn_state->samu_clock_voltage_dependency_table.entries); + kfree(dyn_state->acp_clock_voltage_dependency_table.entries); } enum radeon_pcie_gen r600_get_pcie_gen_support(struct radeon_device *rdev, @@ -1046,3 +1322,36 @@ enum radeon_pcie_gen r600_get_pcie_gen_support(struct radeon_device *rdev, } return RADEON_PCIE_GEN1; } + +u16 r600_get_pcie_lane_support(struct radeon_device *rdev, + u16 asic_lanes, + u16 default_lanes) +{ + switch (asic_lanes) { + case 0: + default: + return default_lanes; + case 1: + return 1; + case 2: + return 2; + case 4: + return 4; + case 8: + return 8; + case 12: + return 12; + case 16: + return 16; + } +} + +u8 r600_encode_pci_lane_width(u32 lanes) +{ + u8 encoded_lanes[] = { 0, 1, 2, 0, 3, 0, 0, 0, 4, 0, 0, 0, 5, 0, 0, 0, 6 }; + + if (lanes > 16) + return 0; + + return encoded_lanes[lanes]; +} diff --git a/drivers/gpu/drm/radeon/r600_dpm.h b/drivers/gpu/drm/radeon/r600_dpm.h index 7c822d9ae53..46b9d2a0301 100644 --- a/drivers/gpu/drm/radeon/r600_dpm.h +++ b/drivers/gpu/drm/radeon/r600_dpm.h @@ -130,6 +130,7 @@ void r600_dpm_print_cap_info(u32 caps); void r600_dpm_print_ps_status(struct radeon_device *rdev, struct radeon_ps *rps); u32 r600_dpm_get_vblank_time(struct radeon_device *rdev); +u32 r600_dpm_get_vrefresh(struct radeon_device *rdev); bool r600_is_uvd_state(u32 class, u32 class2); void r600_calculate_u_and_p(u32 i, u32 r_c, u32 p_b, u32 *p, u32 *u); @@ -212,10 +213,10 @@ void r600_wait_for_power_level(struct radeon_device *rdev, void r600_start_dpm(struct radeon_device *rdev); void r600_stop_dpm(struct radeon_device *rdev); -int r600_set_thermal_temperature_range(struct radeon_device *rdev, - int min_temp, int max_temp); bool r600_is_internal_thermal_sensor(enum radeon_int_thermal_type sensor); +int r600_get_platform_caps(struct radeon_device *rdev); + int r600_parse_extended_power_table(struct radeon_device *rdev); void r600_free_extended_power_table(struct radeon_device *rdev); @@ -224,4 +225,9 @@ enum radeon_pcie_gen r600_get_pcie_gen_support(struct radeon_device *rdev, enum radeon_pcie_gen asic_gen, enum radeon_pcie_gen default_gen); +u16 r600_get_pcie_lane_support(struct radeon_device *rdev, + u16 asic_lanes, + u16 default_lanes); +u8 r600_encode_pci_lane_width(u32 lanes); + #endif diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c index e73b2a73494..26ef8ced6f8 100644 --- a/drivers/gpu/drm/radeon/r600_hdmi.c +++ b/drivers/gpu/drm/radeon/r600_hdmi.c @@ -24,6 +24,7 @@ * Authors: Christian König */ #include <linux/hdmi.h> +#include <linux/gcd.h> #include <drm/drmP.h> #include <drm/radeon_drm.h> #include "radeon.h" @@ -57,28 +58,57 @@ enum r600_hdmi_iec_status_bits { static const struct radeon_hdmi_acr r600_hdmi_predefined_acr[] = { /* 32kHz 44.1kHz 48kHz */ /* Clock N CTS N CTS N CTS */ - { 25174, 4576, 28125, 7007, 31250, 6864, 28125 }, /* 25,20/1.001 MHz */ + { 25175, 4096, 25175, 28224, 125875, 6144, 25175 }, /* 25,20/1.001 MHz */ { 25200, 4096, 25200, 6272, 28000, 6144, 25200 }, /* 25.20 MHz */ { 27000, 4096, 27000, 6272, 30000, 6144, 27000 }, /* 27.00 MHz */ { 27027, 4096, 27027, 6272, 30030, 6144, 27027 }, /* 27.00*1.001 MHz */ { 54000, 4096, 54000, 6272, 60000, 6144, 54000 }, /* 54.00 MHz */ { 54054, 4096, 54054, 6272, 60060, 6144, 54054 }, /* 54.00*1.001 MHz */ - { 74175, 11648, 210937, 17836, 234375, 11648, 140625 }, /* 74.25/1.001 MHz */ + { 74176, 4096, 74176, 5733, 75335, 6144, 74176 }, /* 74.25/1.001 MHz */ { 74250, 4096, 74250, 6272, 82500, 6144, 74250 }, /* 74.25 MHz */ - { 148351, 11648, 421875, 8918, 234375, 5824, 140625 }, /* 148.50/1.001 MHz */ + { 148352, 4096, 148352, 5733, 150670, 6144, 148352 }, /* 148.50/1.001 MHz */ { 148500, 4096, 148500, 6272, 165000, 6144, 148500 }, /* 148.50 MHz */ - { 0, 4096, 0, 6272, 0, 6144, 0 } /* Other */ }; + /* - * calculate CTS value if it's not found in the table + * calculate CTS and N values if they are not found in the table */ -static void r600_hdmi_calc_cts(uint32_t clock, int *CTS, int N, int freq) +static void r600_hdmi_calc_cts(uint32_t clock, int *CTS, int *N, int freq) { - if (*CTS == 0) - *CTS = clock * N / (128 * freq) * 1000; - DRM_DEBUG("Using ACR timing N=%d CTS=%d for frequency %d\n", - N, *CTS, freq); + int n, cts; + unsigned long div, mul; + + /* Safe, but overly large values */ + n = 128 * freq; + cts = clock * 1000; + + /* Smallest valid fraction */ + div = gcd(n, cts); + + n /= div; + cts /= div; + + /* + * The optimal N is 128*freq/1000. Calculate the closest larger + * value that doesn't truncate any bits. + */ + mul = ((128*freq/1000) + (n-1))/n; + + n *= mul; + cts *= mul; + + /* Check that we are in spec (not always possible) */ + if (n < (128*freq/1500)) + printk(KERN_WARNING "Calculated ACR N value is too small. You may experience audio problems.\n"); + if (n > (128*freq/300)) + printk(KERN_WARNING "Calculated ACR N value is too large. You may experience audio problems.\n"); + + *N = n; + *CTS = cts; + + DRM_DEBUG("Calculated ACR timing N=%d CTS=%d for frequency %d\n", + *N, *CTS, freq); } struct radeon_hdmi_acr r600_hdmi_acr(uint32_t clock) @@ -86,15 +116,16 @@ struct radeon_hdmi_acr r600_hdmi_acr(uint32_t clock) struct radeon_hdmi_acr res; u8 i; - for (i = 0; r600_hdmi_predefined_acr[i].clock != clock && - r600_hdmi_predefined_acr[i].clock != 0; i++) - ; - res = r600_hdmi_predefined_acr[i]; + /* Precalculated values for common clocks */ + for (i = 0; i < ARRAY_SIZE(r600_hdmi_predefined_acr); i++) { + if (r600_hdmi_predefined_acr[i].clock == clock) + return r600_hdmi_predefined_acr[i]; + } - /* In case some CTS are missing */ - r600_hdmi_calc_cts(clock, &res.cts_32khz, res.n_32khz, 32000); - r600_hdmi_calc_cts(clock, &res.cts_44_1khz, res.n_44_1khz, 44100); - r600_hdmi_calc_cts(clock, &res.cts_48khz, res.n_48khz, 48000); + /* And odd clocks get manually calculated */ + r600_hdmi_calc_cts(clock, &res.cts_32khz, &res.n_32khz, 32000); + r600_hdmi_calc_cts(clock, &res.cts_44_1khz, &res.n_44_1khz, 44100); + r600_hdmi_calc_cts(clock, &res.cts_48khz, &res.n_48khz, 48000); return res; } @@ -102,7 +133,7 @@ struct radeon_hdmi_acr r600_hdmi_acr(uint32_t clock) /* * update the N and CTS parameters for a given pixel clock rate */ -static void r600_hdmi_update_ACR(struct drm_encoder *encoder, uint32_t clock) +void r600_hdmi_update_ACR(struct drm_encoder *encoder, uint32_t clock) { struct drm_device *dev = encoder->dev; struct radeon_device *rdev = dev->dev_private; @@ -111,21 +142,33 @@ static void r600_hdmi_update_ACR(struct drm_encoder *encoder, uint32_t clock) struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; uint32_t offset = dig->afmt->offset; - WREG32(HDMI0_ACR_32_0 + offset, HDMI0_ACR_CTS_32(acr.cts_32khz)); - WREG32(HDMI0_ACR_32_1 + offset, acr.n_32khz); - - WREG32(HDMI0_ACR_44_0 + offset, HDMI0_ACR_CTS_44(acr.cts_44_1khz)); - WREG32(HDMI0_ACR_44_1 + offset, acr.n_44_1khz); - - WREG32(HDMI0_ACR_48_0 + offset, HDMI0_ACR_CTS_48(acr.cts_48khz)); - WREG32(HDMI0_ACR_48_1 + offset, acr.n_48khz); + WREG32_P(HDMI0_ACR_32_0 + offset, + HDMI0_ACR_CTS_32(acr.cts_32khz), + ~HDMI0_ACR_CTS_32_MASK); + WREG32_P(HDMI0_ACR_32_1 + offset, + HDMI0_ACR_N_32(acr.n_32khz), + ~HDMI0_ACR_N_32_MASK); + + WREG32_P(HDMI0_ACR_44_0 + offset, + HDMI0_ACR_CTS_44(acr.cts_44_1khz), + ~HDMI0_ACR_CTS_44_MASK); + WREG32_P(HDMI0_ACR_44_1 + offset, + HDMI0_ACR_N_44(acr.n_44_1khz), + ~HDMI0_ACR_N_44_MASK); + + WREG32_P(HDMI0_ACR_48_0 + offset, + HDMI0_ACR_CTS_48(acr.cts_48khz), + ~HDMI0_ACR_CTS_48_MASK); + WREG32_P(HDMI0_ACR_48_1 + offset, + HDMI0_ACR_N_48(acr.n_48khz), + ~HDMI0_ACR_N_48_MASK); } /* * build a HDMI Video Info Frame */ -static void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder, - void *buffer, size_t size) +void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder, void *buffer, + size_t size) { struct drm_device *dev = encoder->dev; struct radeon_device *rdev = dev->dev_private; @@ -200,7 +243,7 @@ int r600_hdmi_buffer_status_changed(struct drm_encoder *encoder) /* * write the audio workaround status to the hardware */ -static void r600_hdmi_audio_workaround(struct drm_encoder *encoder) +void r600_hdmi_audio_workaround(struct drm_encoder *encoder) { struct drm_device *dev = encoder->dev; struct radeon_device *rdev = dev->dev_private; @@ -226,10 +269,29 @@ void r600_audio_set_dto(struct drm_encoder *encoder, u32 clock) struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; u32 base_rate = 24000; + u32 max_ratio = clock / base_rate; + u32 dto_phase; + u32 dto_modulo = clock; + u32 wallclock_ratio; + u32 dto_cntl; if (!dig || !dig->afmt) return; + if (max_ratio >= 8) { + dto_phase = 192 * 1000; + wallclock_ratio = 3; + } else if (max_ratio >= 4) { + dto_phase = 96 * 1000; + wallclock_ratio = 2; + } else if (max_ratio >= 2) { + dto_phase = 48 * 1000; + wallclock_ratio = 1; + } else { + dto_phase = 24 * 1000; + wallclock_ratio = 0; + } + /* there are two DTOs selected by DCCG_AUDIO_DTO_SELECT. * doesn't matter which one you use. Just use the first one. */ @@ -238,17 +300,35 @@ void r600_audio_set_dto(struct drm_encoder *encoder, u32 clock) * number (coefficient of two integer numbers. DCCG_AUDIO_DTOx_PHASE * is the numerator, DCCG_AUDIO_DTOx_MODULE is the denominator */ - if (ASIC_IS_DCE3(rdev)) { + if (ASIC_IS_DCE32(rdev)) { + if (dig->dig_encoder == 0) { + dto_cntl = RREG32(DCCG_AUDIO_DTO0_CNTL) & ~DCCG_AUDIO_DTO_WALLCLOCK_RATIO_MASK; + dto_cntl |= DCCG_AUDIO_DTO_WALLCLOCK_RATIO(wallclock_ratio); + WREG32(DCCG_AUDIO_DTO0_CNTL, dto_cntl); + WREG32(DCCG_AUDIO_DTO0_PHASE, dto_phase); + WREG32(DCCG_AUDIO_DTO0_MODULE, dto_modulo); + WREG32(DCCG_AUDIO_DTO_SELECT, 0); /* select DTO0 */ + } else { + dto_cntl = RREG32(DCCG_AUDIO_DTO1_CNTL) & ~DCCG_AUDIO_DTO_WALLCLOCK_RATIO_MASK; + dto_cntl |= DCCG_AUDIO_DTO_WALLCLOCK_RATIO(wallclock_ratio); + WREG32(DCCG_AUDIO_DTO1_CNTL, dto_cntl); + WREG32(DCCG_AUDIO_DTO1_PHASE, dto_phase); + WREG32(DCCG_AUDIO_DTO1_MODULE, dto_modulo); + WREG32(DCCG_AUDIO_DTO_SELECT, 1); /* select DTO1 */ + } + } else { /* according to the reg specs, this should DCE3.2 only, but in - * practice it seems to cover DCE3.0 as well. + * practice it seems to cover DCE2.0/3.0/3.1 as well. */ - WREG32(DCCG_AUDIO_DTO0_PHASE, base_rate * 100); - WREG32(DCCG_AUDIO_DTO0_MODULE, clock * 100); - WREG32(DCCG_AUDIO_DTO_SELECT, 0); /* select DTO0 */ - } else { - /* according to the reg specs, this should be DCE2.0 and DCE3.0 */ - WREG32(AUDIO_DTO, AUDIO_DTO_PHASE(base_rate / 10) | - AUDIO_DTO_MODULE(clock / 10)); + if (dig->dig_encoder == 0) { + WREG32(DCCG_AUDIO_DTO0_PHASE, base_rate * 100); + WREG32(DCCG_AUDIO_DTO0_MODULE, clock * 100); + WREG32(DCCG_AUDIO_DTO_SELECT, 0); /* select DTO0 */ + } else { + WREG32(DCCG_AUDIO_DTO1_PHASE, base_rate * 100); + WREG32(DCCG_AUDIO_DTO1_MODULE, clock * 100); + WREG32(DCCG_AUDIO_DTO_SELECT, 1); /* select DTO1 */ + } } } @@ -264,56 +344,61 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod u8 buffer[HDMI_INFOFRAME_HEADER_SIZE + HDMI_AVI_INFOFRAME_SIZE]; struct hdmi_avi_infoframe frame; uint32_t offset; + uint32_t acr_ctl; ssize_t err; + if (!dig || !dig->afmt) + return; + /* Silent, r600_hdmi_enable will raise WARN for us */ if (!dig->afmt->enabled) return; offset = dig->afmt->offset; - r600_audio_set_dto(encoder, mode->clock); - - WREG32(HDMI0_VBI_PACKET_CONTROL + offset, - HDMI0_NULL_SEND); /* send null packets when required */ - - WREG32(HDMI0_AUDIO_CRC_CONTROL + offset, 0x1000); - - if (ASIC_IS_DCE32(rdev)) { - WREG32(HDMI0_AUDIO_PACKET_CONTROL + offset, - HDMI0_AUDIO_DELAY_EN(1) | /* default audio delay */ - HDMI0_AUDIO_PACKETS_PER_LINE(3)); /* should be suffient for all audio modes and small enough for all hblanks */ - WREG32(AFMT_AUDIO_PACKET_CONTROL + offset, - AFMT_AUDIO_SAMPLE_SEND | /* send audio packets */ - AFMT_60958_CS_UPDATE); /* allow 60958 channel status fields to be updated */ - } else { - WREG32(HDMI0_AUDIO_PACKET_CONTROL + offset, - HDMI0_AUDIO_SAMPLE_SEND | /* send audio packets */ - HDMI0_AUDIO_DELAY_EN(1) | /* default audio delay */ - HDMI0_AUDIO_PACKETS_PER_LINE(3) | /* should be suffient for all audio modes and small enough for all hblanks */ - HDMI0_60958_CS_UPDATE); /* allow 60958 channel status fields to be updated */ - } - - WREG32(HDMI0_ACR_PACKET_CONTROL + offset, - HDMI0_ACR_AUTO_SEND | /* allow hw to sent ACR packets when required */ - HDMI0_ACR_SOURCE); /* select SW CTS value */ - - WREG32(HDMI0_VBI_PACKET_CONTROL + offset, - HDMI0_NULL_SEND | /* send null packets when required */ - HDMI0_GC_SEND | /* send general control packets */ - HDMI0_GC_CONT); /* send general control packets every frame */ + /* disable audio prior to setting up hw */ + dig->afmt->pin = r600_audio_get_pin(rdev); + r600_audio_enable(rdev, dig->afmt->pin, false); - /* TODO: HDMI0_AUDIO_INFO_UPDATE */ - WREG32(HDMI0_INFOFRAME_CONTROL0 + offset, - HDMI0_AVI_INFO_SEND | /* enable AVI info frames */ - HDMI0_AVI_INFO_CONT | /* send AVI info frames every frame/field */ - HDMI0_AUDIO_INFO_SEND | /* enable audio info frames (frames won't be set until audio is enabled) */ - HDMI0_AUDIO_INFO_CONT); /* send audio info frames every frame/field */ - - WREG32(HDMI0_INFOFRAME_CONTROL1 + offset, - HDMI0_AVI_INFO_LINE(2) | /* anything other than 0 */ - HDMI0_AUDIO_INFO_LINE(2)); /* anything other than 0 */ + r600_audio_set_dto(encoder, mode->clock); - WREG32(HDMI0_GC + offset, 0); /* unset HDMI0_GC_AVMUTE */ + WREG32_P(HDMI0_AUDIO_PACKET_CONTROL + offset, + HDMI0_AUDIO_SAMPLE_SEND | /* send audio packets */ + HDMI0_AUDIO_DELAY_EN(1) | /* default audio delay */ + HDMI0_AUDIO_PACKETS_PER_LINE(3) | /* should be suffient for all audio modes and small enough for all hblanks */ + HDMI0_60958_CS_UPDATE, /* allow 60958 channel status fields to be updated */ + ~(HDMI0_AUDIO_SAMPLE_SEND | + HDMI0_AUDIO_DELAY_EN_MASK | + HDMI0_AUDIO_PACKETS_PER_LINE_MASK | + HDMI0_60958_CS_UPDATE)); + + /* DCE 3.0 uses register that's normally for CRC_CONTROL */ + acr_ctl = ASIC_IS_DCE3(rdev) ? DCE3_HDMI0_ACR_PACKET_CONTROL : + HDMI0_ACR_PACKET_CONTROL; + WREG32_P(acr_ctl + offset, + HDMI0_ACR_SOURCE | /* select SW CTS value - XXX verify that hw CTS works on all families */ + HDMI0_ACR_AUTO_SEND, /* allow hw to sent ACR packets when required */ + ~(HDMI0_ACR_SOURCE | + HDMI0_ACR_AUTO_SEND)); + + WREG32_OR(HDMI0_VBI_PACKET_CONTROL + offset, + HDMI0_NULL_SEND | /* send null packets when required */ + HDMI0_GC_SEND | /* send general control packets */ + HDMI0_GC_CONT); /* send general control packets every frame */ + + WREG32_OR(HDMI0_INFOFRAME_CONTROL0 + offset, + HDMI0_AVI_INFO_SEND | /* enable AVI info frames */ + HDMI0_AVI_INFO_CONT | /* send AVI info frames every frame/field */ + HDMI0_AUDIO_INFO_SEND | /* enable audio info frames (frames won't be set until audio is enabled) */ + HDMI0_AUDIO_INFO_UPDATE); /* required for audio info values to be updated */ + + WREG32_P(HDMI0_INFOFRAME_CONTROL1 + offset, + HDMI0_AVI_INFO_LINE(2) | /* anything other than 0 */ + HDMI0_AUDIO_INFO_LINE(2), /* anything other than 0 */ + ~(HDMI0_AVI_INFO_LINE_MASK | + HDMI0_AUDIO_INFO_LINE_MASK)); + + WREG32_AND(HDMI0_GC + offset, + ~HDMI0_GC_AVMUTE); /* unset HDMI0_GC_AVMUTE */ err = drm_hdmi_avi_infoframe_from_display_mode(&frame, mode); if (err < 0) { @@ -328,19 +413,45 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod } r600_hdmi_update_avi_infoframe(encoder, buffer, sizeof(buffer)); + + /* fglrx duplicates INFOFRAME_CONTROL0 & INFOFRAME_CONTROL1 ops here */ + + WREG32_AND(HDMI0_GENERIC_PACKET_CONTROL + offset, + ~(HDMI0_GENERIC0_SEND | + HDMI0_GENERIC0_CONT | + HDMI0_GENERIC0_UPDATE | + HDMI0_GENERIC1_SEND | + HDMI0_GENERIC1_CONT | + HDMI0_GENERIC0_LINE_MASK | + HDMI0_GENERIC1_LINE_MASK)); + r600_hdmi_update_ACR(encoder, mode->clock); + WREG32_P(HDMI0_60958_0 + offset, + HDMI0_60958_CS_CHANNEL_NUMBER_L(1), + ~(HDMI0_60958_CS_CHANNEL_NUMBER_L_MASK | + HDMI0_60958_CS_CLOCK_ACCURACY_MASK)); + + WREG32_P(HDMI0_60958_1 + offset, + HDMI0_60958_CS_CHANNEL_NUMBER_R(2), + ~HDMI0_60958_CS_CHANNEL_NUMBER_R_MASK); + /* it's unknown what these bits do excatly, but it's indeed quite useful for debugging */ WREG32(HDMI0_RAMP_CONTROL0 + offset, 0x00FFFFFF); WREG32(HDMI0_RAMP_CONTROL1 + offset, 0x007FFFFF); WREG32(HDMI0_RAMP_CONTROL2 + offset, 0x00000001); WREG32(HDMI0_RAMP_CONTROL3 + offset, 0x00000001); - r600_hdmi_audio_workaround(encoder); + /* enable audio after to setting up hw */ + r600_audio_enable(rdev, dig->afmt->pin, true); } -/* - * update settings with current parameters from audio engine +/** + * r600_hdmi_update_audio_settings - Update audio infoframe + * + * @encoder: drm encoder + * + * Gets info about current audio stream and updates audio infoframe. */ void r600_hdmi_update_audio_settings(struct drm_encoder *encoder) { @@ -348,11 +459,11 @@ void r600_hdmi_update_audio_settings(struct drm_encoder *encoder) struct radeon_device *rdev = dev->dev_private; struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; - struct r600_audio audio = r600_audio_status(rdev); + struct r600_audio_pin audio = r600_audio_status(rdev); uint8_t buffer[HDMI_INFOFRAME_HEADER_SIZE + HDMI_AUDIO_INFOFRAME_SIZE]; struct hdmi_audio_infoframe frame; uint32_t offset; - uint32_t iec; + uint32_t value; ssize_t err; if (!dig->afmt || !dig->afmt->enabled) @@ -365,60 +476,6 @@ void r600_hdmi_update_audio_settings(struct drm_encoder *encoder) DRM_DEBUG("0x%02X IEC60958 status bits and 0x%02X category code\n", (int)audio.status_bits, (int)audio.category_code); - iec = 0; - if (audio.status_bits & AUDIO_STATUS_PROFESSIONAL) - iec |= 1 << 0; - if (audio.status_bits & AUDIO_STATUS_NONAUDIO) - iec |= 1 << 1; - if (audio.status_bits & AUDIO_STATUS_COPYRIGHT) - iec |= 1 << 2; - if (audio.status_bits & AUDIO_STATUS_EMPHASIS) - iec |= 1 << 3; - - iec |= HDMI0_60958_CS_CATEGORY_CODE(audio.category_code); - - switch (audio.rate) { - case 32000: - iec |= HDMI0_60958_CS_SAMPLING_FREQUENCY(0x3); - break; - case 44100: - iec |= HDMI0_60958_CS_SAMPLING_FREQUENCY(0x0); - break; - case 48000: - iec |= HDMI0_60958_CS_SAMPLING_FREQUENCY(0x2); - break; - case 88200: - iec |= HDMI0_60958_CS_SAMPLING_FREQUENCY(0x8); - break; - case 96000: - iec |= HDMI0_60958_CS_SAMPLING_FREQUENCY(0xa); - break; - case 176400: - iec |= HDMI0_60958_CS_SAMPLING_FREQUENCY(0xc); - break; - case 192000: - iec |= HDMI0_60958_CS_SAMPLING_FREQUENCY(0xe); - break; - } - - WREG32(HDMI0_60958_0 + offset, iec); - - iec = 0; - switch (audio.bits_per_sample) { - case 16: - iec |= HDMI0_60958_CS_WORD_LENGTH(0x2); - break; - case 20: - iec |= HDMI0_60958_CS_WORD_LENGTH(0x3); - break; - case 24: - iec |= HDMI0_60958_CS_WORD_LENGTH(0xb); - break; - } - if (audio.status_bits & AUDIO_STATUS_V) - iec |= 0x5 << 16; - WREG32_P(HDMI0_60958_1 + offset, iec, ~0x5000f); - err = hdmi_audio_infoframe_init(&frame); if (err < 0) { DRM_ERROR("failed to setup audio infoframe\n"); @@ -433,8 +490,22 @@ void r600_hdmi_update_audio_settings(struct drm_encoder *encoder) return; } + value = RREG32(HDMI0_AUDIO_PACKET_CONTROL + offset); + if (value & HDMI0_AUDIO_TEST_EN) + WREG32(HDMI0_AUDIO_PACKET_CONTROL + offset, + value & ~HDMI0_AUDIO_TEST_EN); + + WREG32_OR(HDMI0_CONTROL + offset, + HDMI0_ERROR_ACK); + + WREG32_AND(HDMI0_INFOFRAME_CONTROL0 + offset, + ~HDMI0_AUDIO_INFO_SOURCE); + r600_hdmi_update_audio_infoframe(encoder, buffer, sizeof(buffer)); - r600_hdmi_audio_workaround(encoder); + + WREG32_OR(HDMI0_INFOFRAME_CONTROL0 + offset, + HDMI0_AUDIO_INFO_CONT | + HDMI0_AUDIO_INFO_UPDATE); } /* @@ -448,6 +519,9 @@ void r600_hdmi_enable(struct drm_encoder *encoder, bool enable) struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; u32 hdmi = HDMI0_ERROR_ACK; + if (!dig || !dig->afmt) + return; + /* Silent, r600_hdmi_enable will raise WARN for us */ if (enable && dig->afmt->enabled) return; diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h index f1b3084d8f5..f94e7a9afe7 100644 --- a/drivers/gpu/drm/radeon/r600d.h +++ b/drivers/gpu/drm/radeon/r600d.h @@ -602,6 +602,7 @@ #define L2_BUSY (1 << 0) #define WAIT_UNTIL 0x8040 +#define WAIT_CP_DMA_IDLE_bit (1 << 8) #define WAIT_2D_IDLE_bit (1 << 14) #define WAIT_3D_IDLE_bit (1 << 15) #define WAIT_2D_IDLECLEAN_bit (1 << 16) @@ -700,11 +701,18 @@ #define RLC_UCODE_DATA 0x3f30 #define SRBM_SOFT_RESET 0xe60 +# define SOFT_RESET_BIF (1 << 1) # define SOFT_RESET_DMA (1 << 12) # define SOFT_RESET_RLC (1 << 13) # define SOFT_RESET_UVD (1 << 18) # define RV770_SOFT_RESET_DMA (1 << 20) +#define BIF_SCRATCH0 0x5438 + +#define BUS_CNTL 0x5420 +# define BIOS_ROM_DIS (1 << 1) +# define VGA_COHE_SPEC_TIMER_DIS (1 << 9) + #define CP_INT_CNTL 0xc124 # define CNTX_BUSY_INT_ENABLE (1 << 19) # define CNTX_EMPTY_INT_ENABLE (1 << 20) @@ -932,6 +940,9 @@ #define DCCG_AUDIO_DTO0_LOAD 0x051c # define DTO_LOAD (1 << 31) #define DCCG_AUDIO_DTO0_CNTL 0x0520 +# define DCCG_AUDIO_DTO_WALLCLOCK_RATIO(x) (((x) & 7) << 0) +# define DCCG_AUDIO_DTO_WALLCLOCK_RATIO_MASK 7 +# define DCCG_AUDIO_DTO_WALLCLOCK_RATIO_SHIFT 0 #define DCCG_AUDIO_DTO1_PHASE 0x0524 #define DCCG_AUDIO_DTO1_MODULE 0x0528 @@ -956,6 +967,42 @@ # define DIG_MODE_SDVO 4 #define DIG1_CNTL 0x79a0 +#define AZ_F0_CODEC_PIN0_CONTROL_CHANNEL_SPEAKER 0x71bc +#define SPEAKER_ALLOCATION(x) (((x) & 0x7f) << 0) +#define SPEAKER_ALLOCATION_MASK (0x7f << 0) +#define SPEAKER_ALLOCATION_SHIFT 0 +#define HDMI_CONNECTION (1 << 16) +#define DP_CONNECTION (1 << 17) + +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR0 0x71c8 /* LPCM */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR1 0x71cc /* AC3 */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR2 0x71d0 /* MPEG1 */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR3 0x71d4 /* MP3 */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR4 0x71d8 /* MPEG2 */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR5 0x71dc /* AAC */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR6 0x71e0 /* DTS */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR7 0x71e4 /* ATRAC */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR8 0x71e8 /* one bit audio - leave at 0 (default) */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR9 0x71ec /* Dolby Digital */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR10 0x71f0 /* DTS-HD */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR11 0x71f4 /* MAT-MLP */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR12 0x71f8 /* DTS */ +#define AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR13 0x71fc /* WMA Pro */ +# define MAX_CHANNELS(x) (((x) & 0x7) << 0) +/* max channels minus one. 7 = 8 channels */ +# define SUPPORTED_FREQUENCIES(x) (((x) & 0xff) << 8) +# define DESCRIPTOR_BYTE_2(x) (((x) & 0xff) << 16) +# define SUPPORTED_FREQUENCIES_STEREO(x) (((x) & 0xff) << 24) /* LPCM only */ +/* SUPPORTED_FREQUENCIES, SUPPORTED_FREQUENCIES_STEREO + * bit0 = 32 kHz + * bit1 = 44.1 kHz + * bit2 = 48 kHz + * bit3 = 88.2 kHz + * bit4 = 96 kHz + * bit5 = 176.4 kHz + * bit6 = 192 kHz + */ + /* rs6xx/rs740 and r6xx share the same HDMI blocks, however, rs6xx has only one * instance of the blocks while r6xx has 2. DCE 3.0 cards are slightly * different due to the new DIG blocks, but also have 2 instances. @@ -982,15 +1029,18 @@ #define HDMI0_AUDIO_PACKET_CONTROL 0x7408 # define HDMI0_AUDIO_SAMPLE_SEND (1 << 0) # define HDMI0_AUDIO_DELAY_EN(x) (((x) & 3) << 4) +# define HDMI0_AUDIO_DELAY_EN_MASK (3 << 4) # define HDMI0_AUDIO_SEND_MAX_PACKETS (1 << 8) # define HDMI0_AUDIO_TEST_EN (1 << 12) # define HDMI0_AUDIO_PACKETS_PER_LINE(x) (((x) & 0x1f) << 16) +# define HDMI0_AUDIO_PACKETS_PER_LINE_MASK (0x1f << 16) # define HDMI0_AUDIO_CHANNEL_SWAP (1 << 24) # define HDMI0_60958_CS_UPDATE (1 << 26) # define HDMI0_AZ_FORMAT_WTRIG_MASK (1 << 28) # define HDMI0_AZ_FORMAT_WTRIG_ACK (1 << 29) #define HDMI0_AUDIO_CRC_CONTROL 0x740c # define HDMI0_AUDIO_CRC_EN (1 << 0) +#define DCE3_HDMI0_ACR_PACKET_CONTROL 0x740c #define HDMI0_VBI_PACKET_CONTROL 0x7410 # define HDMI0_NULL_SEND (1 << 0) # define HDMI0_GC_SEND (1 << 4) @@ -1000,14 +1050,16 @@ # define HDMI0_AVI_INFO_CONT (1 << 1) # define HDMI0_AUDIO_INFO_SEND (1 << 4) # define HDMI0_AUDIO_INFO_CONT (1 << 5) -# define HDMI0_AUDIO_INFO_SOURCE (1 << 6) /* 0 - sound block; 1 - hmdi regs */ +# define HDMI0_AUDIO_INFO_SOURCE (1 << 6) /* 0 - sound block; 1 - hdmi regs */ # define HDMI0_AUDIO_INFO_UPDATE (1 << 7) # define HDMI0_MPEG_INFO_SEND (1 << 8) # define HDMI0_MPEG_INFO_CONT (1 << 9) # define HDMI0_MPEG_INFO_UPDATE (1 << 10) #define HDMI0_INFOFRAME_CONTROL1 0x7418 # define HDMI0_AVI_INFO_LINE(x) (((x) & 0x3f) << 0) +# define HDMI0_AVI_INFO_LINE_MASK (0x3f << 0) # define HDMI0_AUDIO_INFO_LINE(x) (((x) & 0x3f) << 8) +# define HDMI0_AUDIO_INFO_LINE_MASK (0x3f << 8) # define HDMI0_MPEG_INFO_LINE(x) (((x) & 0x3f) << 16) #define HDMI0_GENERIC_PACKET_CONTROL 0x741c # define HDMI0_GENERIC0_SEND (1 << 0) @@ -1016,7 +1068,9 @@ # define HDMI0_GENERIC1_SEND (1 << 4) # define HDMI0_GENERIC1_CONT (1 << 5) # define HDMI0_GENERIC0_LINE(x) (((x) & 0x3f) << 16) +# define HDMI0_GENERIC0_LINE_MASK (0x3f << 16) # define HDMI0_GENERIC1_LINE(x) (((x) & 0x3f) << 24) +# define HDMI0_GENERIC1_LINE_MASK (0x3f << 24) #define HDMI0_GC 0x7428 # define HDMI0_GC_AVMUTE (1 << 0) #define HDMI0_AVI_INFO0 0x7454 @@ -1072,16 +1126,22 @@ #define HDMI0_GENERIC1_6 0x74a8 #define HDMI0_ACR_32_0 0x74ac # define HDMI0_ACR_CTS_32(x) (((x) & 0xfffff) << 12) +# define HDMI0_ACR_CTS_32_MASK (0xfffff << 12) #define HDMI0_ACR_32_1 0x74b0 # define HDMI0_ACR_N_32(x) (((x) & 0xfffff) << 0) +# define HDMI0_ACR_N_32_MASK (0xfffff << 0) #define HDMI0_ACR_44_0 0x74b4 # define HDMI0_ACR_CTS_44(x) (((x) & 0xfffff) << 12) +# define HDMI0_ACR_CTS_44_MASK (0xfffff << 12) #define HDMI0_ACR_44_1 0x74b8 # define HDMI0_ACR_N_44(x) (((x) & 0xfffff) << 0) +# define HDMI0_ACR_N_44_MASK (0xfffff << 0) #define HDMI0_ACR_48_0 0x74bc # define HDMI0_ACR_CTS_48(x) (((x) & 0xfffff) << 12) +# define HDMI0_ACR_CTS_48_MASK (0xfffff << 12) #define HDMI0_ACR_48_1 0x74c0 # define HDMI0_ACR_N_48(x) (((x) & 0xfffff) << 0) +# define HDMI0_ACR_N_48_MASK (0xfffff << 0) #define HDMI0_ACR_STATUS_0 0x74c4 #define HDMI0_ACR_STATUS_1 0x74c8 #define HDMI0_AUDIO_INFO0 0x74cc @@ -1101,14 +1161,17 @@ # define HDMI0_60958_CS_CATEGORY_CODE(x) (((x) & 0xff) << 8) # define HDMI0_60958_CS_SOURCE_NUMBER(x) (((x) & 0xf) << 16) # define HDMI0_60958_CS_CHANNEL_NUMBER_L(x) (((x) & 0xf) << 20) +# define HDMI0_60958_CS_CHANNEL_NUMBER_L_MASK (0xf << 20) # define HDMI0_60958_CS_SAMPLING_FREQUENCY(x) (((x) & 0xf) << 24) # define HDMI0_60958_CS_CLOCK_ACCURACY(x) (((x) & 3) << 28) +# define HDMI0_60958_CS_CLOCK_ACCURACY_MASK (3 << 28) #define HDMI0_60958_1 0x74d8 # define HDMI0_60958_CS_WORD_LENGTH(x) (((x) & 0xf) << 0) # define HDMI0_60958_CS_ORIGINAL_SAMPLING_FREQUENCY(x) (((x) & 0xf) << 4) # define HDMI0_60958_CS_VALID_L(x) (((x) & 1) << 16) # define HDMI0_60958_CS_VALID_R(x) (((x) & 1) << 18) # define HDMI0_60958_CS_CHANNEL_NUMBER_R(x) (((x) & 0xf) << 20) +# define HDMI0_60958_CS_CHANNEL_NUMBER_R_MASK (0xf << 20) #define HDMI0_ACR_PACKET_CONTROL 0x74dc # define HDMI0_ACR_SEND (1 << 0) # define HDMI0_ACR_CONT (1 << 1) @@ -1119,6 +1182,7 @@ # define HDMI0_ACR_48 3 # define HDMI0_ACR_SOURCE (1 << 8) /* 0 - hw; 1 - cts value */ # define HDMI0_ACR_AUTO_SEND (1 << 12) +#define DCE3_HDMI0_AUDIO_CRC_CONTROL 0x74dc #define HDMI0_RAMP_CONTROL0 0x74e0 # define HDMI0_RAMP_MAX_COUNT(x) (((x) & 0xffffff) << 0) #define HDMI0_RAMP_CONTROL1 0x74e4 @@ -1159,6 +1223,34 @@ # define AFMT_AZ_FORMAT_WTRIG_ACK (1 << 29) # define AFMT_AZ_AUDIO_ENABLE_CHG_ACK (1 << 30) +/* DCE3 FMT blocks */ +#define FMT_CONTROL 0x6700 +# define FMT_PIXEL_ENCODING (1 << 16) + /* 0 = RGB 4:4:4 or YCbCr 4:4:4, 1 = YCbCr 4:2:2 */ +#define FMT_BIT_DEPTH_CONTROL 0x6710 +# define FMT_TRUNCATE_EN (1 << 0) +# define FMT_TRUNCATE_DEPTH (1 << 4) +# define FMT_SPATIAL_DITHER_EN (1 << 8) +# define FMT_SPATIAL_DITHER_MODE(x) ((x) << 9) +# define FMT_SPATIAL_DITHER_DEPTH (1 << 12) +# define FMT_FRAME_RANDOM_ENABLE (1 << 13) +# define FMT_RGB_RANDOM_ENABLE (1 << 14) +# define FMT_HIGHPASS_RANDOM_ENABLE (1 << 15) +# define FMT_TEMPORAL_DITHER_EN (1 << 16) +# define FMT_TEMPORAL_DITHER_DEPTH (1 << 20) +# define FMT_TEMPORAL_DITHER_OFFSET(x) ((x) << 21) +# define FMT_TEMPORAL_LEVEL (1 << 24) +# define FMT_TEMPORAL_DITHER_RESET (1 << 25) +# define FMT_25FRC_SEL(x) ((x) << 26) +# define FMT_50FRC_SEL(x) ((x) << 28) +# define FMT_75FRC_SEL(x) ((x) << 30) +#define FMT_CLAMP_CONTROL 0x672c +# define FMT_CLAMP_DATA_EN (1 << 0) +# define FMT_CLAMP_COLOR_FORMAT(x) ((x) << 16) +# define FMT_CLAMP_6BPC 0 +# define FMT_CLAMP_8BPC 1 +# define FMT_CLAMP_10BPC 2 + /* Power management */ #define CG_SPLL_FUNC_CNTL 0x600 # define SPLL_RESET (1 << 0) @@ -1483,7 +1575,7 @@ */ # define PACKET3_CP_DMA_CP_SYNC (1 << 31) /* COMMAND */ -# define PACKET3_CP_DMA_CMD_SRC_SWAP(x) ((x) << 23) +# define PACKET3_CP_DMA_CMD_SRC_SWAP(x) ((x) << 22) /* 0 - none * 1 - 8 in 16 * 2 - 8 in 32 @@ -1507,6 +1599,7 @@ # define PACKET3_CP_DMA_CMD_DAIC (1 << 29) #define PACKET3_SURFACE_SYNC 0x43 # define PACKET3_CB0_DEST_BASE_ENA (1 << 6) +# define PACKET3_FULL_CACHE_ENA (1 << 20) /* r7xx+ only */ # define PACKET3_TC_ACTION_ENA (1 << 23) # define PACKET3_VC_ACTION_ENA (1 << 24) # define PACKET3_CB_ACTION_ENA (1 << 25) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 9b7025d02cd..60c47f82912 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -97,6 +97,12 @@ extern int radeon_msi; extern int radeon_lockup_timeout; extern int radeon_fastfb; extern int radeon_dpm; +extern int radeon_aspm; +extern int radeon_runtime_pm; +extern int radeon_hard_reset; +extern int radeon_vm_size; +extern int radeon_vm_block_size; +extern int radeon_deep_color; /* * Copy from radeon_drv.h so we don't have to include both and have conflicting @@ -110,19 +116,16 @@ extern int radeon_dpm; #define RADEONFB_CONN_LIMIT 4 #define RADEON_BIOS_NUM_SCRATCH 8 -/* max number of rings */ -#define RADEON_NUM_RINGS 6 - /* fence seq are set to this number when signaled */ #define RADEON_FENCE_SIGNALED_SEQ 0LL /* internal ring indices */ /* r1xx+ has gfx CP ring */ -#define RADEON_RING_TYPE_GFX_INDEX 0 +#define RADEON_RING_TYPE_GFX_INDEX 0 /* cayman has 2 compute CP rings */ -#define CAYMAN_RING_TYPE_CP1_INDEX 1 -#define CAYMAN_RING_TYPE_CP2_INDEX 2 +#define CAYMAN_RING_TYPE_CP1_INDEX 1 +#define CAYMAN_RING_TYPE_CP2_INDEX 2 /* R600+ has an async dma ring */ #define R600_RING_TYPE_DMA_INDEX 3 @@ -130,13 +133,29 @@ extern int radeon_dpm; #define CAYMAN_RING_TYPE_DMA1_INDEX 4 /* R600+ */ -#define R600_RING_TYPE_UVD_INDEX 5 +#define R600_RING_TYPE_UVD_INDEX 5 + +/* TN+ */ +#define TN_RING_TYPE_VCE1_INDEX 6 +#define TN_RING_TYPE_VCE2_INDEX 7 + +/* max number of rings */ +#define RADEON_NUM_RINGS 8 + +/* number of hw syncs before falling back on blocking */ +#define RADEON_NUM_SYNCS 4 + +/* number of hw syncs before falling back on blocking */ +#define RADEON_NUM_SYNCS 4 /* hardcode those limit for now */ #define RADEON_VA_IB_OFFSET (1 << 20) #define RADEON_VA_RESERVED_SIZE (8 << 20) #define RADEON_IB_VM_MAX_SIZE (64 << 10) +/* hard reset data */ +#define RADEON_ASIC_RESET_DATA 0x39d5e86b + /* reset flags */ #define RADEON_RESET_GFX (1 << 0) #define RADEON_RESET_COMPUTE (1 << 1) @@ -151,6 +170,47 @@ extern int radeon_dpm; #define RADEON_RESET_MC (1 << 10) #define RADEON_RESET_DISPLAY (1 << 11) +/* CG block flags */ +#define RADEON_CG_BLOCK_GFX (1 << 0) +#define RADEON_CG_BLOCK_MC (1 << 1) +#define RADEON_CG_BLOCK_SDMA (1 << 2) +#define RADEON_CG_BLOCK_UVD (1 << 3) +#define RADEON_CG_BLOCK_VCE (1 << 4) +#define RADEON_CG_BLOCK_HDP (1 << 5) +#define RADEON_CG_BLOCK_BIF (1 << 6) + +/* CG flags */ +#define RADEON_CG_SUPPORT_GFX_MGCG (1 << 0) +#define RADEON_CG_SUPPORT_GFX_MGLS (1 << 1) +#define RADEON_CG_SUPPORT_GFX_CGCG (1 << 2) +#define RADEON_CG_SUPPORT_GFX_CGLS (1 << 3) +#define RADEON_CG_SUPPORT_GFX_CGTS (1 << 4) +#define RADEON_CG_SUPPORT_GFX_CGTS_LS (1 << 5) +#define RADEON_CG_SUPPORT_GFX_CP_LS (1 << 6) +#define RADEON_CG_SUPPORT_GFX_RLC_LS (1 << 7) +#define RADEON_CG_SUPPORT_MC_LS (1 << 8) +#define RADEON_CG_SUPPORT_MC_MGCG (1 << 9) +#define RADEON_CG_SUPPORT_SDMA_LS (1 << 10) +#define RADEON_CG_SUPPORT_SDMA_MGCG (1 << 11) +#define RADEON_CG_SUPPORT_BIF_LS (1 << 12) +#define RADEON_CG_SUPPORT_UVD_MGCG (1 << 13) +#define RADEON_CG_SUPPORT_VCE_MGCG (1 << 14) +#define RADEON_CG_SUPPORT_HDP_LS (1 << 15) +#define RADEON_CG_SUPPORT_HDP_MGCG (1 << 16) + +/* PG flags */ +#define RADEON_PG_SUPPORT_GFX_PG (1 << 0) +#define RADEON_PG_SUPPORT_GFX_SMG (1 << 1) +#define RADEON_PG_SUPPORT_GFX_DMG (1 << 2) +#define RADEON_PG_SUPPORT_UVD (1 << 3) +#define RADEON_PG_SUPPORT_VCE (1 << 4) +#define RADEON_PG_SUPPORT_CP (1 << 5) +#define RADEON_PG_SUPPORT_GDS (1 << 6) +#define RADEON_PG_SUPPORT_RLC_SMU_HS (1 << 7) +#define RADEON_PG_SUPPORT_SDMA (1 << 8) +#define RADEON_PG_SUPPORT_ACP (1 << 9) +#define RADEON_PG_SUPPORT_SAMU (1 << 10) + /* max cursor sizes (in pixels) */ #define CURSOR_WIDTH 64 #define CURSOR_HEIGHT 64 @@ -209,6 +269,7 @@ struct radeon_clock { * Power management */ int radeon_pm_init(struct radeon_device *rdev); +int radeon_pm_late_init(struct radeon_device *rdev); void radeon_pm_fini(struct radeon_device *rdev); void radeon_pm_compute_clocks(struct radeon_device *rdev); void radeon_pm_suspend(struct radeon_device *rdev); @@ -237,6 +298,12 @@ int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type, int radeon_atom_get_leakage_vddc_based_on_leakage_idx(struct radeon_device *rdev, u16 *voltage, u16 leakage_idx); +int radeon_atom_get_leakage_id_from_vbios(struct radeon_device *rdev, + u16 *leakage_id); +int radeon_atom_get_leakage_vddc_based_on_leakage_params(struct radeon_device *rdev, + u16 *vddc, u16 *vddci, + u16 virtual_voltage_id, + u16 vbios_voltage_id); int radeon_atom_round_to_true_voltage(struct radeon_device *rdev, u8 voltage_type, u16 nominal_voltage, @@ -279,7 +346,6 @@ struct radeon_fence_driver { /* sync_seq is protected by ring emission lock */ uint64_t sync_seq[RADEON_NUM_RINGS]; atomic64_t last_seq; - unsigned long last_activity; bool initialized; }; @@ -300,8 +366,8 @@ int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence **fence, i void radeon_fence_process(struct radeon_device *rdev, int ring); bool radeon_fence_signaled(struct radeon_fence *fence); int radeon_fence_wait(struct radeon_fence *fence, bool interruptible); -int radeon_fence_wait_next_locked(struct radeon_device *rdev, int ring); -int radeon_fence_wait_empty_locked(struct radeon_device *rdev, int ring); +int radeon_fence_wait_next(struct radeon_device *rdev, int ring); +int radeon_fence_wait_empty(struct radeon_device *rdev, int ring); int radeon_fence_wait_any(struct radeon_device *rdev, struct radeon_fence **fences, bool intr); @@ -364,6 +430,11 @@ struct radeon_mman { struct ttm_bo_device bdev; bool mem_global_referenced; bool initialized; + +#if defined(CONFIG_DEBUG_FS) + struct dentry *vram; + struct dentry *gtt; +#endif }; /* bo virtual address in a specific vm */ @@ -378,6 +449,7 @@ struct radeon_bo_va { /* protected by vm mutex */ struct list_head vm_list; + struct list_head vm_status; /* constant after initialization */ struct radeon_vm *vm; @@ -388,6 +460,7 @@ struct radeon_bo { /* Protected by gem.mutex */ struct list_head list; /* Protected by tbo.reserved */ + u32 initial_domain; u32 placements[3]; struct ttm_placement placement; struct ttm_buffer_object tbo; @@ -410,16 +483,6 @@ struct radeon_bo { }; #define gem_to_radeon_bo(gobj) container_of((gobj), struct radeon_bo, gem_base) -struct radeon_bo_list { - struct ttm_validate_buffer tv; - struct radeon_bo *bo; - uint64_t gpu_offset; - bool written; - unsigned domain; - unsigned alt_domain; - u32 tiling_flags; -}; - int radeon_gem_debugfs_init(struct radeon_device *rdev); /* sub-allocation manager, it has to be protected by another lock. @@ -455,6 +518,7 @@ struct radeon_sa_manager { uint64_t gpu_addr; void *cpu_ptr; uint32_t domain; + uint32_t align; }; struct radeon_sa_bo; @@ -490,29 +554,28 @@ int radeon_mode_dumb_create(struct drm_file *file_priv, int radeon_mode_dumb_mmap(struct drm_file *filp, struct drm_device *dev, uint32_t handle, uint64_t *offset_p); -int radeon_mode_dumb_destroy(struct drm_file *file_priv, - struct drm_device *dev, - uint32_t handle); /* * Semaphores. */ -/* everything here is constant */ struct radeon_semaphore { struct radeon_sa_bo *sa_bo; signed waiters; uint64_t gpu_addr; + struct radeon_fence *sync_to[RADEON_NUM_RINGS]; }; int radeon_semaphore_create(struct radeon_device *rdev, struct radeon_semaphore **semaphore); -void radeon_semaphore_emit_signal(struct radeon_device *rdev, int ring, +bool radeon_semaphore_emit_signal(struct radeon_device *rdev, int ring, struct radeon_semaphore *semaphore); -void radeon_semaphore_emit_wait(struct radeon_device *rdev, int ring, +bool radeon_semaphore_emit_wait(struct radeon_device *rdev, int ring, struct radeon_semaphore *semaphore); +void radeon_semaphore_sync_to(struct radeon_semaphore *semaphore, + struct radeon_fence *fence); int radeon_semaphore_sync_rings(struct radeon_device *rdev, struct radeon_semaphore *semaphore, - int signaler, int waiter); + int waiting_ring); void radeon_semaphore_free(struct radeon_device *rdev, struct radeon_semaphore **semaphore, struct radeon_fence *fence); @@ -599,13 +662,15 @@ void radeon_scratch_free(struct radeon_device *rdev, uint32_t reg); /* * GPU doorbell structures, functions & helpers */ +#define RADEON_MAX_DOORBELLS 1024 /* Reserve at most 1024 doorbell slots for radeon-owned rings. */ + struct radeon_doorbell { - u32 num_pages; - bool free[1024]; /* doorbell mmio */ - resource_size_t base; - resource_size_t size; - void __iomem *ptr; + resource_size_t base; + resource_size_t size; + u32 __iomem *ptr; + u32 num_doorbells; /* Number of doorbells actually reserved for radeon. */ + unsigned long used[DIV_ROUND_UP(RADEON_MAX_DOORBELLS, BITS_PER_LONG)]; }; int radeon_doorbell_get(struct radeon_device *rdev, u32 *page); @@ -615,14 +680,15 @@ void radeon_doorbell_free(struct radeon_device *rdev, u32 doorbell); * IRQS. */ -struct radeon_unpin_work { - struct work_struct work; - struct radeon_device *rdev; - int crtc_id; - struct radeon_fence *fence; +struct radeon_flip_work { + struct work_struct flip_work; + struct work_struct unpin_work; + struct radeon_device *rdev; + int crtc_id; + uint64_t base; struct drm_pending_vblank_event *event; - struct radeon_bo *old_rbo; - u64 new_crtc_base; + struct radeon_bo *old_rbo; + struct radeon_fence *fence; }; struct r500_irq_stat_regs { @@ -669,6 +735,12 @@ struct cik_irq_stat_regs { u32 disp_int_cont4; u32 disp_int_cont5; u32 disp_int_cont6; + u32 d1grph_int; + u32 d2grph_int; + u32 d3grph_int; + u32 d4grph_int; + u32 d5grph_int; + u32 d6grph_int; }; union radeon_irq_stat_regs { @@ -678,10 +750,6 @@ union radeon_irq_stat_regs { struct cik_irq_stat_regs cik; }; -#define RADEON_MAX_HPD_PINS 6 -#define RADEON_MAX_CRTCS 6 -#define RADEON_MAX_AFMT_BLOCKS 6 - struct radeon_irq { bool installed; spinlock_t lock; @@ -719,33 +787,27 @@ struct radeon_ib { struct radeon_fence *fence; struct radeon_vm *vm; bool is_const_ib; - struct radeon_fence *sync_to[RADEON_NUM_RINGS]; struct radeon_semaphore *semaphore; }; struct radeon_ring { struct radeon_bo *ring_obj; volatile uint32_t *ring; - unsigned rptr; unsigned rptr_offs; - unsigned rptr_reg; unsigned rptr_save_reg; u64 next_rptr_gpu_addr; volatile u32 *next_rptr_cpu_addr; unsigned wptr; unsigned wptr_old; - unsigned wptr_reg; unsigned ring_size; unsigned ring_free_dw; int count_dw; - unsigned long last_activity; - unsigned last_rptr; + atomic_t last_rptr; + atomic64_t last_activity; uint64_t gpu_addr; uint32_t align_mask; uint32_t ptr_mask; bool ready; - u32 ptr_reg_shift; - u32 ptr_reg_mask; u32 nop; u32 idx; u64 last_semaphore_signal_addr; @@ -755,8 +817,7 @@ struct radeon_ring { u32 pipe; u32 queue; struct radeon_bo *mqd_obj; - u32 doorbell_page_num; - u32 doorbell_offset; + u32 doorbell_index; unsigned wptr_offs; }; @@ -775,38 +836,62 @@ struct radeon_mec { /* maximum number of VMIDs */ #define RADEON_NUM_VM 16 -/* defines number of bits in page table versus page directory, - * a page is 4KB so we have 12 bits offset, 9 bits in the page - * table and the remaining 19 bits are in the page directory */ -#define RADEON_VM_BLOCK_SIZE 9 - /* number of entries in page table */ -#define RADEON_VM_PTE_COUNT (1 << RADEON_VM_BLOCK_SIZE) +#define RADEON_VM_PTE_COUNT (1 << radeon_vm_block_size) + +/* PTBs (Page Table Blocks) need to be aligned to 32K */ +#define RADEON_VM_PTB_ALIGN_SIZE 32768 +#define RADEON_VM_PTB_ALIGN_MASK (RADEON_VM_PTB_ALIGN_SIZE - 1) +#define RADEON_VM_PTB_ALIGN(a) (((a) + RADEON_VM_PTB_ALIGN_MASK) & ~RADEON_VM_PTB_ALIGN_MASK) + +#define R600_PTE_VALID (1 << 0) +#define R600_PTE_SYSTEM (1 << 1) +#define R600_PTE_SNOOPED (1 << 2) +#define R600_PTE_READABLE (1 << 5) +#define R600_PTE_WRITEABLE (1 << 6) + +/* PTE (Page Table Entry) fragment field for different page sizes */ +#define R600_PTE_FRAG_4KB (0 << 7) +#define R600_PTE_FRAG_64KB (4 << 7) +#define R600_PTE_FRAG_256KB (6 << 7) + +/* flags used for GART page table entries on R600+ */ +#define R600_PTE_GART ( R600_PTE_VALID | R600_PTE_SYSTEM | R600_PTE_SNOOPED \ + | R600_PTE_READABLE | R600_PTE_WRITEABLE) + +struct radeon_vm_pt { + struct radeon_bo *bo; + uint64_t addr; +}; struct radeon_vm { - struct list_head list; struct list_head va; unsigned id; + /* BOs freed, but not yet updated in the PT */ + struct list_head freed; + /* contains the page directory */ - struct radeon_sa_bo *page_directory; + struct radeon_bo *page_directory; uint64_t pd_gpu_addr; + unsigned max_pde_used; /* array of page tables, one for each page directory entry */ - struct radeon_sa_bo **page_tables; + struct radeon_vm_pt *page_tables; + + struct radeon_bo_va *ib_bo_va; struct mutex mutex; /* last fence for cs using this vm */ struct radeon_fence *fence; /* last flush or NULL if we still need to flush */ struct radeon_fence *last_flush; + /* last use of vmid */ + struct radeon_fence *last_id_use; }; struct radeon_vm_manager { - struct mutex lock; - struct list_head lru_vm; struct radeon_fence *active[RADEON_NUM_VM]; - struct radeon_sa_manager sa_manager; uint32_t max_pfn; /* number of VMIDs */ unsigned nvm; @@ -837,35 +922,6 @@ struct r600_ih { bool enabled; }; -struct r600_blit_cp_primitives { - void (*set_render_target)(struct radeon_device *rdev, int format, - int w, int h, u64 gpu_addr); - void (*cp_set_surface_sync)(struct radeon_device *rdev, - u32 sync_type, u32 size, - u64 mc_addr); - void (*set_shaders)(struct radeon_device *rdev); - void (*set_vtx_resource)(struct radeon_device *rdev, u64 gpu_addr); - void (*set_tex_resource)(struct radeon_device *rdev, - int format, int w, int h, int pitch, - u64 gpu_addr, u32 size); - void (*set_scissors)(struct radeon_device *rdev, int x1, int y1, - int x2, int y2); - void (*draw_auto)(struct radeon_device *rdev); - void (*set_default_state)(struct radeon_device *rdev); -}; - -struct r600_blit { - struct radeon_bo *shader_obj; - struct r600_blit_cp_primitives primitives; - int max_dim; - int ring_size_common; - int ring_size_per_loop; - u64 shader_gpu_addr; - u32 vs_offset, ps_offset; - u32 state_offset; - u32 state_len; -}; - /* * RLC stuff */ @@ -876,20 +932,25 @@ struct radeon_rlc { struct radeon_bo *save_restore_obj; uint64_t save_restore_gpu_addr; volatile uint32_t *sr_ptr; - u32 *reg_list; + const u32 *reg_list; u32 reg_list_size; /* for clear state */ struct radeon_bo *clear_state_obj; uint64_t clear_state_gpu_addr; volatile uint32_t *cs_ptr; - struct cs_section_def *cs_data; + const struct cs_section_def *cs_data; + u32 clear_state_size; + /* for cp tables */ + struct radeon_bo *cp_table_obj; + uint64_t cp_table_gpu_addr; + volatile uint32_t *cp_table_ptr; + u32 cp_table_size; }; int radeon_ib_get(struct radeon_device *rdev, int ring, struct radeon_ib *ib, struct radeon_vm *vm, unsigned size); void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib); -void radeon_ib_sync_to(struct radeon_ib *ib, struct radeon_fence *fence); int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, struct radeon_ib *const_ib); int radeon_ib_pool_init(struct radeon_device *rdev); @@ -906,16 +967,15 @@ void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *c void radeon_ring_undo(struct radeon_ring *ring); void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *cp); int radeon_ring_test(struct radeon_device *rdev, struct radeon_ring *cp); -void radeon_ring_force_activity(struct radeon_device *rdev, struct radeon_ring *ring); -void radeon_ring_lockup_update(struct radeon_ring *ring); +void radeon_ring_lockup_update(struct radeon_device *rdev, + struct radeon_ring *ring); bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *ring); unsigned radeon_ring_backup(struct radeon_device *rdev, struct radeon_ring *ring, uint32_t **data); int radeon_ring_restore(struct radeon_device *rdev, struct radeon_ring *ring, unsigned size, uint32_t *data); int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ring_size, - unsigned rptr_offs, unsigned rptr_reg, unsigned wptr_reg, - u32 ptr_reg_shift, u32 ptr_reg_mask, u32 nop); + unsigned rptr_offs, u32 nop); void radeon_ring_fini(struct radeon_device *rdev, struct radeon_ring *cp); @@ -934,20 +994,19 @@ void cayman_dma_fini(struct radeon_device *rdev); struct radeon_cs_reloc { struct drm_gem_object *gobj; struct radeon_bo *robj; - struct radeon_bo_list lobj; + struct ttm_validate_buffer tv; + uint64_t gpu_offset; + unsigned prefered_domains; + unsigned allowed_domains; + uint32_t tiling_flags; uint32_t handle; - uint32_t flags; }; struct radeon_cs_chunk { uint32_t chunk_id; uint32_t length_dw; - int kpage_idx[2]; - uint32_t *kpage[2]; uint32_t *kdata; void __user *user_ptr; - int last_copied_page; - int last_page_index; }; struct radeon_cs_parser { @@ -964,6 +1023,7 @@ struct radeon_cs_parser { unsigned nrelocs; struct radeon_cs_reloc *relocs; struct radeon_cs_reloc **relocs_ptr; + struct radeon_cs_reloc *vm_bos; struct list_head validated; unsigned dma_reloc_idx; /* indices of various chunks */ @@ -982,8 +1042,15 @@ struct radeon_cs_parser { struct ww_acquire_ctx ticket; }; -extern int radeon_cs_finish_pages(struct radeon_cs_parser *p); -extern u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx); +static inline u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx) +{ + struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx]; + + if (ibc->kdata) + return ibc->kdata[idx]; + return p->ib.ptr[idx]; +} + struct radeon_cs_packet { unsigned idx; @@ -1029,7 +1096,6 @@ struct radeon_wb { #define R600_WB_DMA_RPTR_OFFSET 1792 #define R600_WB_IH_WPTR_OFFSET 2048 #define CAYMAN_WB_DMA1_RPTR_OFFSET 2304 -#define R600_WB_UVD_RPTR_OFFSET 2560 #define R600_WB_EVENT_OFFSET 3072 #define CIK_WB_CP1_WPTR_OFFSET 3328 #define CIK_WB_CP2_WPTR_OFFSET 3584 @@ -1140,6 +1206,7 @@ enum radeon_int_thermal_type { THERMAL_TYPE_SI, THERMAL_TYPE_EMC2103_WITH_INTERNAL, THERMAL_TYPE_CI, + THERMAL_TYPE_KV, }; struct radeon_voltage { @@ -1206,6 +1273,17 @@ enum radeon_dpm_event_src { RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL = 4 }; +#define RADEON_MAX_VCE_LEVELS 6 + +enum radeon_vce_level { + RADEON_VCE_LEVEL_AC_ALL = 0, /* AC, All cases */ + RADEON_VCE_LEVEL_DC_EE = 1, /* DC, entropy encoding */ + RADEON_VCE_LEVEL_DC_LL_LOW = 2, /* DC, low latency queue, res <= 720 */ + RADEON_VCE_LEVEL_DC_LL_HIGH = 3, /* DC, low latency queue, 1080 >= res > 720 */ + RADEON_VCE_LEVEL_DC_GP_LOW = 4, /* DC, general purpose queue, res <= 720 */ + RADEON_VCE_LEVEL_DC_GP_HIGH = 5, /* DC, general purpose queue, 1080 >= res > 720 */ +}; + struct radeon_ps { u32 caps; /* vbios flags */ u32 class; /* vbios flags */ @@ -1213,6 +1291,11 @@ struct radeon_ps { /* UVD clocks */ u32 vclk; u32 dclk; + /* VCE clocks */ + u32 evclk; + u32 ecclk; + bool vce_active; + enum radeon_vce_level vce_level; /* asic priv */ void *ps_priv; }; @@ -1244,8 +1327,8 @@ struct radeon_blacklist_clocks struct radeon_clock_and_voltage_limits { u32 sclk; u32 mclk; - u32 vddc; - u32 vddci; + u16 vddc; + u16 vddci; }; struct radeon_clock_array { @@ -1263,14 +1346,21 @@ struct radeon_clock_voltage_dependency_table { struct radeon_clock_voltage_dependency_entry *entries; }; -struct radeon_cac_leakage_entry { - u16 vddc; - u32 leakage; +union radeon_cac_leakage_entry { + struct { + u16 vddc; + u32 leakage; + }; + struct { + u16 vddc1; + u16 vddc2; + u16 vddc3; + }; }; struct radeon_cac_leakage_table { u32 count; - struct radeon_cac_leakage_entry *entries; + union radeon_cac_leakage_entry *entries; }; struct radeon_phase_shedding_limits_entry { @@ -1284,6 +1374,28 @@ struct radeon_phase_shedding_limits_table { struct radeon_phase_shedding_limits_entry *entries; }; +struct radeon_uvd_clock_voltage_dependency_entry { + u32 vclk; + u32 dclk; + u16 v; +}; + +struct radeon_uvd_clock_voltage_dependency_table { + u8 count; + struct radeon_uvd_clock_voltage_dependency_entry *entries; +}; + +struct radeon_vce_clock_voltage_dependency_entry { + u32 ecclk; + u32 evclk; + u16 v; +}; + +struct radeon_vce_clock_voltage_dependency_table { + u8 count; + struct radeon_vce_clock_voltage_dependency_entry *entries; +}; + struct radeon_ppm_table { u8 ppm_design; u16 cpu_core_number; @@ -1297,11 +1409,27 @@ struct radeon_ppm_table { u32 tj_max; }; +struct radeon_cac_tdp_table { + u16 tdp; + u16 configurable_tdp; + u16 tdc; + u16 battery_power_limit; + u16 small_power_limit; + u16 low_cac_leakage; + u16 high_cac_leakage; + u16 maximum_power_delivery_limit; +}; + struct radeon_dpm_dynamic_state { struct radeon_clock_voltage_dependency_table vddc_dependency_on_sclk; struct radeon_clock_voltage_dependency_table vddci_dependency_on_mclk; struct radeon_clock_voltage_dependency_table vddc_dependency_on_mclk; + struct radeon_clock_voltage_dependency_table mvdd_dependency_on_mclk; struct radeon_clock_voltage_dependency_table vddc_dependency_on_dispclk; + struct radeon_uvd_clock_voltage_dependency_table uvd_clock_voltage_dependency_table; + struct radeon_vce_clock_voltage_dependency_table vce_clock_voltage_dependency_table; + struct radeon_clock_voltage_dependency_table samu_clock_voltage_dependency_table; + struct radeon_clock_voltage_dependency_table acp_clock_voltage_dependency_table; struct radeon_clock_array valid_sclk_values; struct radeon_clock_array valid_mclk_values; struct radeon_clock_and_voltage_limits max_clock_voltage_on_dc; @@ -1313,6 +1441,7 @@ struct radeon_dpm_dynamic_state { struct radeon_cac_leakage_table cac_leakage_table; struct radeon_phase_shedding_limits_table phase_shedding_limits_table; struct radeon_ppm_table *ppm_table; + struct radeon_cac_tdp_table *cac_tdp_table; }; struct radeon_dpm_fan { @@ -1341,6 +1470,17 @@ enum radeon_dpm_forced_level { RADEON_DPM_FORCED_LEVEL_HIGH = 2, }; +struct radeon_vce_state { + /* vce clocks */ + u32 evclk; + u32 ecclk; + /* gpu clocks */ + u32 sclk; + u32 mclk; + u8 clk_idx; + u8 pstate; +}; + struct radeon_dpm { struct radeon_ps *ps; /* number of valid power states */ @@ -1353,6 +1493,9 @@ struct radeon_dpm { struct radeon_ps *boot_ps; /* default uvd power state */ struct radeon_ps *uvd_ps; + /* vce requirements */ + struct radeon_vce_state vce_states[RADEON_MAX_VCE_LEVELS]; + enum radeon_vce_level vce_level; enum radeon_pm_state_type state; enum radeon_pm_state_type user_state; u32 platform_caps; @@ -1378,15 +1521,18 @@ struct radeon_dpm { /* special states active */ bool thermal_active; bool uvd_active; + bool vce_active; /* thermal handling */ struct radeon_dpm_thermal thermal; /* forced levels */ enum radeon_dpm_forced_level forced_level; + /* track UVD streams */ + unsigned sd; + unsigned hd; }; -void radeon_dpm_enable_power_state(struct radeon_device *rdev, - enum radeon_pm_state_type dpm_state); - +void radeon_dpm_enable_uvd(struct radeon_device *rdev, bool enable); +void radeon_dpm_enable_vce(struct radeon_device *rdev, bool enable); struct radeon_pm { struct mutex mutex; @@ -1460,8 +1606,10 @@ struct radeon_uvd { struct radeon_bo *vcpu_bo; void *cpu_addr; uint64_t gpu_addr; + void *saved_bo; atomic_t handles[RADEON_MAX_UVD_HANDLES]; struct drm_file *filp[RADEON_MAX_UVD_HANDLES]; + unsigned img_size[RADEON_MAX_UVD_HANDLES]; struct delayed_work idle_work; }; @@ -1490,12 +1638,61 @@ int radeon_uvd_calc_upll_dividers(struct radeon_device *rdev, int radeon_uvd_send_upll_ctlreq(struct radeon_device *rdev, unsigned cg_upll_func_cntl); -struct r600_audio { +/* + * VCE + */ +#define RADEON_MAX_VCE_HANDLES 16 +#define RADEON_VCE_STACK_SIZE (1024*1024) +#define RADEON_VCE_HEAP_SIZE (4*1024*1024) + +struct radeon_vce { + struct radeon_bo *vcpu_bo; + uint64_t gpu_addr; + unsigned fw_version; + unsigned fb_version; + atomic_t handles[RADEON_MAX_VCE_HANDLES]; + struct drm_file *filp[RADEON_MAX_VCE_HANDLES]; + unsigned img_size[RADEON_MAX_VCE_HANDLES]; + struct delayed_work idle_work; +}; + +int radeon_vce_init(struct radeon_device *rdev); +void radeon_vce_fini(struct radeon_device *rdev); +int radeon_vce_suspend(struct radeon_device *rdev); +int radeon_vce_resume(struct radeon_device *rdev); +int radeon_vce_get_create_msg(struct radeon_device *rdev, int ring, + uint32_t handle, struct radeon_fence **fence); +int radeon_vce_get_destroy_msg(struct radeon_device *rdev, int ring, + uint32_t handle, struct radeon_fence **fence); +void radeon_vce_free_handles(struct radeon_device *rdev, struct drm_file *filp); +void radeon_vce_note_usage(struct radeon_device *rdev); +int radeon_vce_cs_reloc(struct radeon_cs_parser *p, int lo, int hi, unsigned size); +int radeon_vce_cs_parse(struct radeon_cs_parser *p); +bool radeon_vce_semaphore_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait); +void radeon_vce_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); +void radeon_vce_fence_emit(struct radeon_device *rdev, + struct radeon_fence *fence); +int radeon_vce_ring_test(struct radeon_device *rdev, struct radeon_ring *ring); +int radeon_vce_ib_test(struct radeon_device *rdev, struct radeon_ring *ring); + +struct r600_audio_pin { int channels; int rate; int bits_per_sample; u8 status_bits; u8 category_code; + u32 offset; + bool connected; + u32 id; +}; + +struct r600_audio { + bool enabled; + struct r600_audio_pin pin[RADEON_MAX_AFMT_BLOCKS]; + int num_pins; }; /* @@ -1527,6 +1724,34 @@ int radeon_debugfs_add_files(struct radeon_device *rdev, unsigned nfiles); int radeon_debugfs_fence_init(struct radeon_device *rdev); +/* + * ASIC ring specific functions. + */ +struct radeon_asic_ring { + /* ring read/write ptr handling */ + u32 (*get_rptr)(struct radeon_device *rdev, struct radeon_ring *ring); + u32 (*get_wptr)(struct radeon_device *rdev, struct radeon_ring *ring); + void (*set_wptr)(struct radeon_device *rdev, struct radeon_ring *ring); + + /* validating and patching of IBs */ + int (*ib_parse)(struct radeon_device *rdev, struct radeon_ib *ib); + int (*cs_parse)(struct radeon_cs_parser *p); + + /* command emmit functions */ + void (*ib_execute)(struct radeon_device *rdev, struct radeon_ib *ib); + void (*emit_fence)(struct radeon_device *rdev, struct radeon_fence *fence); + bool (*emit_semaphore)(struct radeon_device *rdev, struct radeon_ring *cp, + struct radeon_semaphore *semaphore, bool emit_wait); + void (*vm_flush)(struct radeon_device *rdev, int ridx, struct radeon_vm *vm); + + /* testing functions */ + int (*ring_test)(struct radeon_device *rdev, struct radeon_ring *cp); + int (*ib_test)(struct radeon_device *rdev, struct radeon_ring *cp); + bool (*is_lockup)(struct radeon_device *rdev, struct radeon_ring *cp); + + /* deprecated */ + void (*ring_start)(struct radeon_device *rdev, struct radeon_ring *cp); +}; /* * ASIC specific functions. @@ -1556,13 +1781,12 @@ struct radeon_asic { /* gart */ struct { void (*tlb_flush)(struct radeon_device *rdev); - int (*set_page)(struct radeon_device *rdev, int i, uint64_t addr); + void (*set_page)(struct radeon_device *rdev, unsigned i, + uint64_t addr); } gart; struct { int (*init)(struct radeon_device *rdev); void (*fini)(struct radeon_device *rdev); - - u32 pt_ring_index; void (*set_page)(struct radeon_device *rdev, struct radeon_ib *ib, uint64_t pe, @@ -1570,23 +1794,7 @@ struct radeon_asic { uint32_t incr, uint32_t flags); } vm; /* ring specific callbacks */ - struct { - void (*ib_execute)(struct radeon_device *rdev, struct radeon_ib *ib); - int (*ib_parse)(struct radeon_device *rdev, struct radeon_ib *ib); - void (*emit_fence)(struct radeon_device *rdev, struct radeon_fence *fence); - void (*emit_semaphore)(struct radeon_device *rdev, struct radeon_ring *cp, - struct radeon_semaphore *semaphore, bool emit_wait); - int (*cs_parse)(struct radeon_cs_parser *p); - void (*ring_start)(struct radeon_device *rdev, struct radeon_ring *cp); - int (*ring_test)(struct radeon_device *rdev, struct radeon_ring *cp); - int (*ib_test)(struct radeon_device *rdev, struct radeon_ring *cp); - bool (*is_lockup)(struct radeon_device *rdev, struct radeon_ring *cp); - void (*vm_flush)(struct radeon_device *rdev, int ridx, struct radeon_vm *vm); - - u32 (*get_rptr)(struct radeon_device *rdev, struct radeon_ring *ring); - u32 (*get_wptr)(struct radeon_device *rdev, struct radeon_ring *ring); - void (*set_wptr)(struct radeon_device *rdev, struct radeon_ring *ring); - } ring[RADEON_NUM_RINGS]; + struct radeon_asic_ring *ring[RADEON_NUM_RINGS]; /* irqs */ struct { int (*set)(struct radeon_device *rdev); @@ -1660,6 +1868,7 @@ struct radeon_asic { void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes); void (*set_clock_gating)(struct radeon_device *rdev, int enable); int (*set_uvd_clocks)(struct radeon_device *rdev, u32 vclk, u32 dclk); + int (*set_vce_clocks)(struct radeon_device *rdev, u32 evclk, u32 ecclk); int (*get_temperature)(struct radeon_device *rdev); } pm; /* dynamic power management */ @@ -1667,6 +1876,7 @@ struct radeon_asic { int (*init)(struct radeon_device *rdev); void (*setup_asic)(struct radeon_device *rdev); int (*enable)(struct radeon_device *rdev); + int (*late_enable)(struct radeon_device *rdev); void (*disable)(struct radeon_device *rdev); int (*pre_set_power_state)(struct radeon_device *rdev); int (*set_power_state)(struct radeon_device *rdev); @@ -1679,12 +1889,13 @@ struct radeon_asic { void (*debugfs_print_current_performance_level)(struct radeon_device *rdev, struct seq_file *m); int (*force_performance_level)(struct radeon_device *rdev, enum radeon_dpm_forced_level level); bool (*vblank_too_short)(struct radeon_device *rdev); + void (*powergate_uvd)(struct radeon_device *rdev, bool gate); + void (*enable_bapm)(struct radeon_device *rdev, bool enable); } dpm; /* pageflipping */ struct { - void (*pre_page_flip)(struct radeon_device *rdev, int crtc); - u32 (*page_flip)(struct radeon_device *rdev, int crtc, u64 crtc_base); - void (*post_page_flip)(struct radeon_device *rdev, int crtc); + void (*page_flip)(struct radeon_device *rdev, int crtc, u64 crtc_base); + bool (*page_flip_pending)(struct radeon_device *rdev, int crtc); } pflip; }; @@ -1723,6 +1934,7 @@ struct r600_asic { unsigned tiling_group_size; unsigned tile_config; unsigned backend_map; + unsigned active_simds; }; struct rv770_asic { @@ -1748,6 +1960,7 @@ struct rv770_asic { unsigned tiling_group_size; unsigned tile_config; unsigned backend_map; + unsigned active_simds; }; struct evergreen_asic { @@ -1774,6 +1987,7 @@ struct evergreen_asic { unsigned tiling_group_size; unsigned tile_config; unsigned backend_map; + unsigned active_simds; }; struct cayman_asic { @@ -1812,6 +2026,7 @@ struct cayman_asic { unsigned multi_gpu_tile_size; unsigned tile_config; + unsigned active_simds; }; struct si_asic { @@ -1830,7 +2045,7 @@ struct si_asic { unsigned sc_earlyz_tile_fifo_size; unsigned num_tile_pipes; - unsigned num_backends_per_se; + unsigned backend_enable_mask; unsigned backend_disable_mask_per_asic; unsigned backend_map; unsigned num_texture_channel_caches; @@ -1842,6 +2057,7 @@ struct si_asic { unsigned tile_config; uint32_t tile_mode_array[32]; + uint32_t active_cus; }; struct cik_asic { @@ -1860,7 +2076,7 @@ struct cik_asic { unsigned sc_earlyz_tile_fifo_size; unsigned num_tile_pipes; - unsigned num_backends_per_se; + unsigned backend_enable_mask; unsigned backend_disable_mask_per_asic; unsigned backend_map; unsigned num_texture_channel_caches; @@ -1872,6 +2088,8 @@ struct cik_asic { unsigned tile_config; uint32_t tile_mode_array[32]; + uint32_t macrotile_mode_array[16]; + uint32_t active_cus; }; union radeon_asic_config { @@ -1917,6 +2135,8 @@ int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); int radeon_gem_va_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); +int radeon_gem_op_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp); int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); int radeon_gem_set_tiling_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); @@ -2011,6 +2231,28 @@ struct radeon_device { resource_size_t rmmio_size; /* protects concurrent MM_INDEX/DATA based register access */ spinlock_t mmio_idx_lock; + /* protects concurrent SMC based register access */ + spinlock_t smc_idx_lock; + /* protects concurrent PLL register access */ + spinlock_t pll_idx_lock; + /* protects concurrent MC register access */ + spinlock_t mc_idx_lock; + /* protects concurrent PCIE register access */ + spinlock_t pcie_idx_lock; + /* protects concurrent PCIE_PORT register access */ + spinlock_t pciep_idx_lock; + /* protects concurrent PIF register access */ + spinlock_t pif_idx_lock; + /* protects concurrent CG register access */ + spinlock_t cg_idx_lock; + /* protects concurrent UVD register access */ + spinlock_t uvd_idx_lock; + /* protects concurrent RCU register access */ + spinlock_t rcu_idx_lock; + /* protects concurrent DIDT register access */ + spinlock_t didt_idx_lock; + /* protects concurrent ENDPOINT (audio) register access */ + spinlock_t end_idx_lock; void __iomem *rmmio; radeon_rreg_t mc_rreg; radeon_wreg_t mc_wreg; @@ -2040,6 +2282,7 @@ struct radeon_device { struct radeon_gem gem; struct radeon_pm pm; struct radeon_uvd uvd; + struct radeon_vce vce; uint32_t bios_scratch[RADEON_BIOS_NUM_SCRATCH]; struct radeon_wb wb; struct radeon_dummy_page dummy_page; @@ -2048,17 +2291,18 @@ struct radeon_device { bool need_dma32; bool accel_working; bool fastfb_working; /* IGP feature*/ + bool needs_reset; struct radeon_surface_reg surface_regs[RADEON_GEM_MAX_SURFACES]; const struct firmware *me_fw; /* all family ME firmware */ const struct firmware *pfp_fw; /* r6/700 PFP firmware */ const struct firmware *rlc_fw; /* r6/700 RLC firmware */ const struct firmware *mc_fw; /* NI MC firmware */ const struct firmware *ce_fw; /* SI CE firmware */ - const struct firmware *uvd_fw; /* UVD firmware */ const struct firmware *mec_fw; /* CIK MEC firmware */ const struct firmware *sdma_fw; /* CIK SDMA firmware */ const struct firmware *smc_fw; /* SMC firmware */ - struct r600_blit r600_blit; + const struct firmware *uvd_fw; /* UVD firmware */ + const struct firmware *vce_fw; /* VCE firmware */ struct r600_vram_scratch vram_scratch; int msi_enabled; /* msi enabled */ struct r600_ih ih; /* r6/700 interrupt ring */ @@ -2069,9 +2313,8 @@ struct radeon_device { struct work_struct reset_work; int num_crtc; /* number of crtcs */ struct mutex dc_hw_i2c_mutex; /* display controller hw i2c mutex */ - bool audio_enabled; bool has_uvd; - struct r600_audio audio_status; /* audio stuff */ + struct r600_audio audio; /* audio stuff */ struct notifier_block acpi_nb; /* only one userspace can use Hyperz features or CMASK at a time */ struct drm_file *hyperz_filp; @@ -2084,11 +2327,24 @@ struct radeon_device { /* virtual memory */ struct radeon_vm_manager vm_manager; struct mutex gpu_clock_mutex; + /* memory stats */ + atomic64_t vram_usage; + atomic64_t gtt_usage; + atomic64_t num_bytes_moved; /* ACPI interface */ struct radeon_atif atif; struct radeon_atcs atcs; + /* srbm instance registers */ + struct mutex srbm_mutex; + /* clock, powergating flags */ + u32 cg_flags; + u32 pg_flags; + + struct dev_pm_domain vga_pm_domain; + bool have_disp_power_ref; }; +bool radeon_is_px(struct drm_device *dev); int radeon_device_init(struct radeon_device *rdev, struct drm_device *ddev, struct pci_dev *pdev, @@ -2103,8 +2359,8 @@ void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v, u32 r100_io_rreg(struct radeon_device *rdev, u32 reg); void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v); -u32 cik_mm_rdoorbell(struct radeon_device *rdev, u32 offset); -void cik_mm_wdoorbell(struct radeon_device *rdev, u32 offset, u32 v); +u32 cik_mm_rdoorbell(struct radeon_device *rdev, u32 index); +void cik_mm_wdoorbell(struct radeon_device *rdev, u32 index, u32 v); /* * Cast helper @@ -2145,6 +2401,8 @@ void cik_mm_wdoorbell(struct radeon_device *rdev, u32 offset, u32 v); #define WREG32_PIF_PHY1(reg, v) eg_pif_phy1_wreg(rdev, (reg), (v)) #define RREG32_UVD_CTX(reg) r600_uvd_ctx_rreg(rdev, (reg)) #define WREG32_UVD_CTX(reg, v) r600_uvd_ctx_wreg(rdev, (reg), (v)) +#define RREG32_DIDT(reg) cik_didt_rreg(rdev, (reg)) +#define WREG32_DIDT(reg, v) cik_didt_wreg(rdev, (reg), (v)) #define WREG32_P(reg, val, mask) \ do { \ uint32_t tmp_ = RREG32(reg); \ @@ -2153,7 +2411,7 @@ void cik_mm_wdoorbell(struct radeon_device *rdev, u32 offset, u32 v); WREG32(reg, tmp_); \ } while (0) #define WREG32_AND(reg, and) WREG32_P(reg, 0, and) -#define WREG32_OR(reg, or) WREG32_P(reg, or, ~or) +#define WREG32_OR(reg, or) WREG32_P(reg, or, ~(or)) #define WREG32_PLL_P(reg, val, mask) \ do { \ uint32_t tmp_ = RREG32_PLL(reg); \ @@ -2165,115 +2423,187 @@ void cik_mm_wdoorbell(struct radeon_device *rdev, u32 offset, u32 v); #define RREG32_IO(reg) r100_io_rreg(rdev, (reg)) #define WREG32_IO(reg, v) r100_io_wreg(rdev, (reg), (v)) -#define RDOORBELL32(offset) cik_mm_rdoorbell(rdev, (offset)) -#define WDOORBELL32(offset, v) cik_mm_wdoorbell(rdev, (offset), (v)) +#define RDOORBELL32(index) cik_mm_rdoorbell(rdev, (index)) +#define WDOORBELL32(index, v) cik_mm_wdoorbell(rdev, (index), (v)) /* * Indirect registers accessor */ static inline uint32_t rv370_pcie_rreg(struct radeon_device *rdev, uint32_t reg) { + unsigned long flags; uint32_t r; + spin_lock_irqsave(&rdev->pcie_idx_lock, flags); WREG32(RADEON_PCIE_INDEX, ((reg) & rdev->pcie_reg_mask)); r = RREG32(RADEON_PCIE_DATA); + spin_unlock_irqrestore(&rdev->pcie_idx_lock, flags); return r; } static inline void rv370_pcie_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->pcie_idx_lock, flags); WREG32(RADEON_PCIE_INDEX, ((reg) & rdev->pcie_reg_mask)); WREG32(RADEON_PCIE_DATA, (v)); + spin_unlock_irqrestore(&rdev->pcie_idx_lock, flags); } static inline u32 tn_smc_rreg(struct radeon_device *rdev, u32 reg) { + unsigned long flags; u32 r; + spin_lock_irqsave(&rdev->smc_idx_lock, flags); WREG32(TN_SMC_IND_INDEX_0, (reg)); r = RREG32(TN_SMC_IND_DATA_0); + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); return r; } static inline void tn_smc_wreg(struct radeon_device *rdev, u32 reg, u32 v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->smc_idx_lock, flags); WREG32(TN_SMC_IND_INDEX_0, (reg)); WREG32(TN_SMC_IND_DATA_0, (v)); + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); } static inline u32 r600_rcu_rreg(struct radeon_device *rdev, u32 reg) { + unsigned long flags; u32 r; + spin_lock_irqsave(&rdev->rcu_idx_lock, flags); WREG32(R600_RCU_INDEX, ((reg) & 0x1fff)); r = RREG32(R600_RCU_DATA); + spin_unlock_irqrestore(&rdev->rcu_idx_lock, flags); return r; } static inline void r600_rcu_wreg(struct radeon_device *rdev, u32 reg, u32 v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->rcu_idx_lock, flags); WREG32(R600_RCU_INDEX, ((reg) & 0x1fff)); WREG32(R600_RCU_DATA, (v)); + spin_unlock_irqrestore(&rdev->rcu_idx_lock, flags); } static inline u32 eg_cg_rreg(struct radeon_device *rdev, u32 reg) { + unsigned long flags; u32 r; + spin_lock_irqsave(&rdev->cg_idx_lock, flags); WREG32(EVERGREEN_CG_IND_ADDR, ((reg) & 0xffff)); r = RREG32(EVERGREEN_CG_IND_DATA); + spin_unlock_irqrestore(&rdev->cg_idx_lock, flags); return r; } static inline void eg_cg_wreg(struct radeon_device *rdev, u32 reg, u32 v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->cg_idx_lock, flags); WREG32(EVERGREEN_CG_IND_ADDR, ((reg) & 0xffff)); WREG32(EVERGREEN_CG_IND_DATA, (v)); + spin_unlock_irqrestore(&rdev->cg_idx_lock, flags); } static inline u32 eg_pif_phy0_rreg(struct radeon_device *rdev, u32 reg) { + unsigned long flags; u32 r; + spin_lock_irqsave(&rdev->pif_idx_lock, flags); WREG32(EVERGREEN_PIF_PHY0_INDEX, ((reg) & 0xffff)); r = RREG32(EVERGREEN_PIF_PHY0_DATA); + spin_unlock_irqrestore(&rdev->pif_idx_lock, flags); return r; } static inline void eg_pif_phy0_wreg(struct radeon_device *rdev, u32 reg, u32 v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->pif_idx_lock, flags); WREG32(EVERGREEN_PIF_PHY0_INDEX, ((reg) & 0xffff)); WREG32(EVERGREEN_PIF_PHY0_DATA, (v)); + spin_unlock_irqrestore(&rdev->pif_idx_lock, flags); } static inline u32 eg_pif_phy1_rreg(struct radeon_device *rdev, u32 reg) { + unsigned long flags; u32 r; + spin_lock_irqsave(&rdev->pif_idx_lock, flags); WREG32(EVERGREEN_PIF_PHY1_INDEX, ((reg) & 0xffff)); r = RREG32(EVERGREEN_PIF_PHY1_DATA); + spin_unlock_irqrestore(&rdev->pif_idx_lock, flags); return r; } static inline void eg_pif_phy1_wreg(struct radeon_device *rdev, u32 reg, u32 v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->pif_idx_lock, flags); WREG32(EVERGREEN_PIF_PHY1_INDEX, ((reg) & 0xffff)); WREG32(EVERGREEN_PIF_PHY1_DATA, (v)); + spin_unlock_irqrestore(&rdev->pif_idx_lock, flags); } static inline u32 r600_uvd_ctx_rreg(struct radeon_device *rdev, u32 reg) { + unsigned long flags; u32 r; + spin_lock_irqsave(&rdev->uvd_idx_lock, flags); WREG32(R600_UVD_CTX_INDEX, ((reg) & 0x1ff)); r = RREG32(R600_UVD_CTX_DATA); + spin_unlock_irqrestore(&rdev->uvd_idx_lock, flags); return r; } static inline void r600_uvd_ctx_wreg(struct radeon_device *rdev, u32 reg, u32 v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->uvd_idx_lock, flags); WREG32(R600_UVD_CTX_INDEX, ((reg) & 0x1ff)); WREG32(R600_UVD_CTX_DATA, (v)); + spin_unlock_irqrestore(&rdev->uvd_idx_lock, flags); +} + + +static inline u32 cik_didt_rreg(struct radeon_device *rdev, u32 reg) +{ + unsigned long flags; + u32 r; + + spin_lock_irqsave(&rdev->didt_idx_lock, flags); + WREG32(CIK_DIDT_IND_INDEX, (reg)); + r = RREG32(CIK_DIDT_IND_DATA); + spin_unlock_irqrestore(&rdev->didt_idx_lock, flags); + return r; +} + +static inline void cik_didt_wreg(struct radeon_device *rdev, u32 reg, u32 v) +{ + unsigned long flags; + + spin_lock_irqsave(&rdev->didt_idx_lock, flags); + WREG32(CIK_DIDT_IND_INDEX, (reg)); + WREG32(CIK_DIDT_IND_DATA, (v)); + spin_unlock_irqrestore(&rdev->didt_idx_lock, flags); } void r100_pll_errata_after_index(struct radeon_device *rdev); @@ -2325,6 +2655,10 @@ void r100_pll_errata_after_index(struct radeon_device *rdev); #define ASIC_IS_DCE64(rdev) ((rdev->family == CHIP_OLAND)) #define ASIC_IS_NODCE(rdev) ((rdev->family == CHIP_HAINAN)) #define ASIC_IS_DCE8(rdev) ((rdev->family >= CHIP_BONAIRE)) +#define ASIC_IS_DCE81(rdev) ((rdev->family == CHIP_KAVERI)) +#define ASIC_IS_DCE82(rdev) ((rdev->family == CHIP_BONAIRE)) +#define ASIC_IS_DCE83(rdev) ((rdev->family == CHIP_KABINI) || \ + (rdev->family == CHIP_MULLINS)) #define ASIC_IS_LOMBOK(rdev) ((rdev->ddev->pdev->device == 0x6849) || \ (rdev->ddev->pdev->device == 0x6850) || \ @@ -2371,7 +2705,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v); #define radeon_fini(rdev) (rdev)->asic->fini((rdev)) #define radeon_resume(rdev) (rdev)->asic->resume((rdev)) #define radeon_suspend(rdev) (rdev)->asic->suspend((rdev)) -#define radeon_cs_parse(rdev, r, p) (rdev)->asic->ring[(r)].cs_parse((p)) +#define radeon_cs_parse(rdev, r, p) (rdev)->asic->ring[(r)]->cs_parse((p)) #define radeon_vga_set_state(rdev, state) (rdev)->asic->vga_set_state((rdev), (state)) #define radeon_asic_reset(rdev) (rdev)->asic->asic_reset((rdev)) #define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart.tlb_flush((rdev)) @@ -2379,16 +2713,16 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v); #define radeon_asic_vm_init(rdev) (rdev)->asic->vm.init((rdev)) #define radeon_asic_vm_fini(rdev) (rdev)->asic->vm.fini((rdev)) #define radeon_asic_vm_set_page(rdev, ib, pe, addr, count, incr, flags) ((rdev)->asic->vm.set_page((rdev), (ib), (pe), (addr), (count), (incr), (flags))) -#define radeon_ring_start(rdev, r, cp) (rdev)->asic->ring[(r)].ring_start((rdev), (cp)) -#define radeon_ring_test(rdev, r, cp) (rdev)->asic->ring[(r)].ring_test((rdev), (cp)) -#define radeon_ib_test(rdev, r, cp) (rdev)->asic->ring[(r)].ib_test((rdev), (cp)) -#define radeon_ring_ib_execute(rdev, r, ib) (rdev)->asic->ring[(r)].ib_execute((rdev), (ib)) -#define radeon_ring_ib_parse(rdev, r, ib) (rdev)->asic->ring[(r)].ib_parse((rdev), (ib)) -#define radeon_ring_is_lockup(rdev, r, cp) (rdev)->asic->ring[(r)].is_lockup((rdev), (cp)) -#define radeon_ring_vm_flush(rdev, r, vm) (rdev)->asic->ring[(r)].vm_flush((rdev), (r), (vm)) -#define radeon_ring_get_rptr(rdev, r) (rdev)->asic->ring[(r)->idx].get_rptr((rdev), (r)) -#define radeon_ring_get_wptr(rdev, r) (rdev)->asic->ring[(r)->idx].get_wptr((rdev), (r)) -#define radeon_ring_set_wptr(rdev, r) (rdev)->asic->ring[(r)->idx].set_wptr((rdev), (r)) +#define radeon_ring_start(rdev, r, cp) (rdev)->asic->ring[(r)]->ring_start((rdev), (cp)) +#define radeon_ring_test(rdev, r, cp) (rdev)->asic->ring[(r)]->ring_test((rdev), (cp)) +#define radeon_ib_test(rdev, r, cp) (rdev)->asic->ring[(r)]->ib_test((rdev), (cp)) +#define radeon_ring_ib_execute(rdev, r, ib) (rdev)->asic->ring[(r)]->ib_execute((rdev), (ib)) +#define radeon_ring_ib_parse(rdev, r, ib) (rdev)->asic->ring[(r)]->ib_parse((rdev), (ib)) +#define radeon_ring_is_lockup(rdev, r, cp) (rdev)->asic->ring[(r)]->is_lockup((rdev), (cp)) +#define radeon_ring_vm_flush(rdev, r, vm) (rdev)->asic->ring[(r)]->vm_flush((rdev), (r), (vm)) +#define radeon_ring_get_rptr(rdev, r) (rdev)->asic->ring[(r)->idx]->get_rptr((rdev), (r)) +#define radeon_ring_get_wptr(rdev, r) (rdev)->asic->ring[(r)->idx]->get_wptr((rdev), (r)) +#define radeon_ring_set_wptr(rdev, r) (rdev)->asic->ring[(r)->idx]->set_wptr((rdev), (r)) #define radeon_irq_set(rdev) (rdev)->asic->irq.set((rdev)) #define radeon_irq_process(rdev) (rdev)->asic->irq.process((rdev)) #define radeon_get_vblank_counter(rdev, crtc) (rdev)->asic->display.get_vblank_counter((rdev), (crtc)) @@ -2396,8 +2730,8 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v); #define radeon_get_backlight_level(rdev, e) (rdev)->asic->display.get_backlight_level((e)) #define radeon_hdmi_enable(rdev, e, b) (rdev)->asic->display.hdmi_enable((e), (b)) #define radeon_hdmi_setmode(rdev, e, m) (rdev)->asic->display.hdmi_setmode((e), (m)) -#define radeon_fence_ring_emit(rdev, r, fence) (rdev)->asic->ring[(r)].emit_fence((rdev), (fence)) -#define radeon_semaphore_ring_emit(rdev, r, cp, semaphore, emit_wait) (rdev)->asic->ring[(r)].emit_semaphore((rdev), (cp), (semaphore), (emit_wait)) +#define radeon_fence_ring_emit(rdev, r, fence) (rdev)->asic->ring[(r)]->emit_fence((rdev), (fence)) +#define radeon_semaphore_ring_emit(rdev, r, cp, semaphore, emit_wait) (rdev)->asic->ring[(r)]->emit_semaphore((rdev), (cp), (semaphore), (emit_wait)) #define radeon_copy_blit(rdev, s, d, np, f) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (f)) #define radeon_copy_dma(rdev, s, d, np, f) (rdev)->asic->copy.dma((rdev), (s), (d), (np), (f)) #define radeon_copy(rdev, s, d, np, f) (rdev)->asic->copy.copy((rdev), (s), (d), (np), (f)) @@ -2412,6 +2746,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v); #define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->pm.set_pcie_lanes((rdev), (l)) #define radeon_set_clock_gating(rdev, e) (rdev)->asic->pm.set_clock_gating((rdev), (e)) #define radeon_set_uvd_clocks(rdev, v, d) (rdev)->asic->pm.set_uvd_clocks((rdev), (v), (d)) +#define radeon_set_vce_clocks(rdev, ev, ec) (rdev)->asic->pm.set_vce_clocks((rdev), (ev), (ec)) #define radeon_get_temperature(rdev) (rdev)->asic->pm.get_temperature((rdev)) #define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->surface.set_reg((rdev), (r), (f), (p), (o), (s))) #define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->surface.clear_reg((rdev), (r))) @@ -2426,9 +2761,8 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v); #define radeon_pm_finish(rdev) (rdev)->asic->pm.finish((rdev)) #define radeon_pm_init_profile(rdev) (rdev)->asic->pm.init_profile((rdev)) #define radeon_pm_get_dynpm_state(rdev) (rdev)->asic->pm.get_dynpm_state((rdev)) -#define radeon_pre_page_flip(rdev, crtc) (rdev)->asic->pflip.pre_page_flip((rdev), (crtc)) #define radeon_page_flip(rdev, crtc, base) (rdev)->asic->pflip.page_flip((rdev), (crtc), (base)) -#define radeon_post_page_flip(rdev, crtc) (rdev)->asic->pflip.post_page_flip((rdev), (crtc)) +#define radeon_page_flip_pending(rdev, crtc) (rdev)->asic->pflip.page_flip_pending((rdev), (crtc)) #define radeon_wait_for_vblank(rdev, crtc) (rdev)->asic->display.wait_for_vblank((rdev), (crtc)) #define radeon_mc_wait_for_idle(rdev) (rdev)->asic->mc_wait_for_idle((rdev)) #define radeon_get_xclk(rdev) (rdev)->asic->get_xclk((rdev)) @@ -2436,6 +2770,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v); #define radeon_dpm_init(rdev) rdev->asic->dpm.init((rdev)) #define radeon_dpm_setup_asic(rdev) rdev->asic->dpm.setup_asic((rdev)) #define radeon_dpm_enable(rdev) rdev->asic->dpm.enable((rdev)) +#define radeon_dpm_late_enable(rdev) rdev->asic->dpm.late_enable((rdev)) #define radeon_dpm_disable(rdev) rdev->asic->dpm.disable((rdev)) #define radeon_dpm_pre_set_power_state(rdev) rdev->asic->dpm.pre_set_power_state((rdev)) #define radeon_dpm_set_power_state(rdev) rdev->asic->dpm.set_power_state((rdev)) @@ -2448,10 +2783,13 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v); #define radeon_dpm_debugfs_print_current_performance_level(rdev, m) rdev->asic->dpm.debugfs_print_current_performance_level((rdev), (m)) #define radeon_dpm_force_performance_level(rdev, l) rdev->asic->dpm.force_performance_level((rdev), (l)) #define radeon_dpm_vblank_too_short(rdev) rdev->asic->dpm.vblank_too_short((rdev)) +#define radeon_dpm_powergate_uvd(rdev, g) rdev->asic->dpm.powergate_uvd((rdev), (g)) +#define radeon_dpm_enable_bapm(rdev, e) rdev->asic->dpm.enable_bapm((rdev), (e)) /* Common functions */ /* AGP */ extern int radeon_gpu_reset(struct radeon_device *rdev); +extern void radeon_pci_config_reset(struct radeon_device *rdev); extern void r600_set_bios_scratch_engine_hung(struct radeon_device *rdev, bool hung); extern void radeon_agp_disable(struct radeon_device *rdev); extern int radeon_modeset_init(struct radeon_device *rdev); @@ -2472,8 +2810,8 @@ extern void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain); extern bool radeon_ttm_bo_is_radeon_bo(struct ttm_buffer_object *bo); extern void radeon_vram_location(struct radeon_device *rdev, struct radeon_mc *mc, u64 base); extern void radeon_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc); -extern int radeon_resume_kms(struct drm_device *dev); -extern int radeon_suspend_kms(struct drm_device *dev, pm_message_t state); +extern int radeon_resume_kms(struct drm_device *dev, bool resume, bool fbcon); +extern int radeon_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon); extern void radeon_ttm_set_active_vram_size(struct radeon_device *rdev, u64 size); extern void radeon_program_register_sequence(struct radeon_device *rdev, const u32 *registers, @@ -2484,20 +2822,27 @@ extern void radeon_program_register_sequence(struct radeon_device *rdev, */ int radeon_vm_manager_init(struct radeon_device *rdev); void radeon_vm_manager_fini(struct radeon_device *rdev); -void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm); +int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm); void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm); -int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm); -void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm); +struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev, + struct radeon_vm *vm, + struct list_head *head); struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev, struct radeon_vm *vm, int ring); +void radeon_vm_flush(struct radeon_device *rdev, + struct radeon_vm *vm, + int ring); void radeon_vm_fence(struct radeon_device *rdev, struct radeon_vm *vm, struct radeon_fence *fence); uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr); -int radeon_vm_bo_update_pte(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_bo *bo, - struct ttm_mem_reg *mem); +int radeon_vm_update_page_directory(struct radeon_device *rdev, + struct radeon_vm *vm); +int radeon_vm_clear_freed(struct radeon_device *rdev, + struct radeon_vm *vm); +int radeon_vm_bo_update(struct radeon_device *rdev, + struct radeon_bo_va *bo_va, + struct ttm_mem_reg *mem); void radeon_vm_bo_invalidate(struct radeon_device *rdev, struct radeon_bo *bo); struct radeon_bo_va *radeon_vm_bo_find(struct radeon_vm *vm, @@ -2509,11 +2854,19 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev, struct radeon_bo_va *bo_va, uint64_t offset, uint32_t flags); -int radeon_vm_bo_rmv(struct radeon_device *rdev, - struct radeon_bo_va *bo_va); +void radeon_vm_bo_rmv(struct radeon_device *rdev, + struct radeon_bo_va *bo_va); /* audio */ void r600_audio_update_hdmi(struct work_struct *work); +struct r600_audio_pin *r600_audio_get_pin(struct radeon_device *rdev); +struct r600_audio_pin *dce6_audio_get_pin(struct radeon_device *rdev); +void r600_audio_enable(struct radeon_device *rdev, + struct r600_audio_pin *pin, + bool enable); +void dce6_audio_enable(struct radeon_device *rdev, + struct r600_audio_pin *pin, + bool enable); /* * R600 vram scratch functions diff --git a/drivers/gpu/drm/radeon/radeon_acpi.c b/drivers/gpu/drm/radeon/radeon_acpi.c index 10f98c7742d..77e9d07c55b 100644 --- a/drivers/gpu/drm/radeon/radeon_acpi.c +++ b/drivers/gpu/drm/radeon/radeon_acpi.c @@ -25,18 +25,14 @@ #include <linux/acpi.h> #include <linux/slab.h> #include <linux/power_supply.h> -#include <acpi/acpi_drivers.h> -#include <acpi/acpi_bus.h> +#include <linux/vga_switcheroo.h> #include <acpi/video.h> - #include <drm/drmP.h> #include <drm/drm_crtc_helper.h> #include "radeon.h" #include "radeon_acpi.h" #include "atom.h" -#include <linux/vga_switcheroo.h> - #define ACPI_AC_CLASS "ac_adapter" extern void radeon_pm_acpi_event_handler(struct radeon_device *rdev); @@ -369,7 +365,7 @@ int radeon_atif_handler(struct radeon_device *rdev, return NOTIFY_DONE; /* Check pending SBIOS requests */ - handle = DEVICE_ACPI_HANDLE(&rdev->pdev->dev); + handle = ACPI_HANDLE(&rdev->pdev->dev); count = radeon_atif_get_sbios_requests(handle, &req); if (count <= 0) @@ -556,7 +552,7 @@ int radeon_acpi_pcie_notify_device_ready(struct radeon_device *rdev) struct radeon_atcs *atcs = &rdev->atcs; /* Get the device handle */ - handle = DEVICE_ACPI_HANDLE(&rdev->pdev->dev); + handle = ACPI_HANDLE(&rdev->pdev->dev); if (!handle) return -EINVAL; @@ -596,7 +592,7 @@ int radeon_acpi_pcie_performance_request(struct radeon_device *rdev, u32 retry = 3; /* Get the device handle */ - handle = DEVICE_ACPI_HANDLE(&rdev->pdev->dev); + handle = ACPI_HANDLE(&rdev->pdev->dev); if (!handle) return -EINVAL; @@ -699,7 +695,7 @@ int radeon_acpi_init(struct radeon_device *rdev) int ret; /* Get the device handle */ - handle = DEVICE_ACPI_HANDLE(&rdev->pdev->dev); + handle = ACPI_HANDLE(&rdev->pdev->dev); /* No need to proceed if we're sure that ATIF is not supported */ if (!ASIC_IS_AVIVO(rdev) || !rdev->bios || !handle) diff --git a/drivers/gpu/drm/radeon/radeon_agp.c b/drivers/gpu/drm/radeon/radeon_agp.c index 42433344cb1..a9297b2c352 100644 --- a/drivers/gpu/drm/radeon/radeon_agp.c +++ b/drivers/gpu/drm/radeon/radeon_agp.c @@ -117,9 +117,6 @@ static struct radeon_agpmode_quirk radeon_agpmode_quirk_list[] = { /* ATI Host Bridge / RV280 [M9+] Needs AGPMode 1 (phoronix forum) */ { PCI_VENDOR_ID_ATI, 0xcbb2, PCI_VENDOR_ID_ATI, 0x5c61, PCI_VENDOR_ID_SONY, 0x8175, 1}, - /* HP Host Bridge / R300 [FireGL X1] Needs AGPMode 2 (fdo #7770) */ - { PCI_VENDOR_ID_HP, 0x122e, PCI_VENDOR_ID_ATI, 0x4e47, - PCI_VENDOR_ID_ATI, 0x0152, 2}, { 0, 0, 0, 0, 0, 0, 0 }, }; #endif diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c index 097077499cc..34b9aa9e3c0 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.c +++ b/drivers/gpu/drm/radeon/radeon_asic.c @@ -172,6 +172,21 @@ void radeon_agp_disable(struct radeon_device *rdev) /* * ASIC */ + +static struct radeon_asic_ring r100_gfx_ring = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r100_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r100_cs_parse, + .ring_start = &r100_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + .is_lockup = &r100_gpu_is_lockup, + .get_rptr = &r100_gfx_get_rptr, + .get_wptr = &r100_gfx_get_wptr, + .set_wptr = &r100_gfx_set_wptr, +}; + static struct radeon_asic r100_asic = { .init = &r100_init, .fini = &r100_fini, @@ -187,19 +202,7 @@ static struct radeon_asic r100_asic = { .set_page = &r100_pci_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r100_ring_ib_execute, - .emit_fence = &r100_fence_ring_emit, - .emit_semaphore = &r100_semaphore_ring_emit, - .cs_parse = &r100_cs_parse, - .ring_start = &r100_ring_start, - .ring_test = &r100_ring_test, - .ib_test = &r100_ib_test, - .is_lockup = &r100_gpu_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r100_gfx_ring }, .irq = { .set = &r100_irq_set, @@ -245,9 +248,8 @@ static struct radeon_asic r100_asic = { .set_clock_gating = &radeon_legacy_set_clock_gating, }, .pflip = { - .pre_page_flip = &r100_pre_page_flip, .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .page_flip_pending = &r100_page_flip_pending, }, }; @@ -266,19 +268,7 @@ static struct radeon_asic r200_asic = { .set_page = &r100_pci_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r100_ring_ib_execute, - .emit_fence = &r100_fence_ring_emit, - .emit_semaphore = &r100_semaphore_ring_emit, - .cs_parse = &r100_cs_parse, - .ring_start = &r100_ring_start, - .ring_test = &r100_ring_test, - .ib_test = &r100_ib_test, - .is_lockup = &r100_gpu_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r100_gfx_ring }, .irq = { .set = &r100_irq_set, @@ -324,12 +314,25 @@ static struct radeon_asic r200_asic = { .set_clock_gating = &radeon_legacy_set_clock_gating, }, .pflip = { - .pre_page_flip = &r100_pre_page_flip, .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .page_flip_pending = &r100_page_flip_pending, }, }; +static struct radeon_asic_ring r300_gfx_ring = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r300_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r300_cs_parse, + .ring_start = &r300_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + .is_lockup = &r100_gpu_is_lockup, + .get_rptr = &r100_gfx_get_rptr, + .get_wptr = &r100_gfx_get_wptr, + .set_wptr = &r100_gfx_set_wptr, +}; + static struct radeon_asic r300_asic = { .init = &r300_init, .fini = &r300_fini, @@ -345,19 +348,7 @@ static struct radeon_asic r300_asic = { .set_page = &r100_pci_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r100_ring_ib_execute, - .emit_fence = &r300_fence_ring_emit, - .emit_semaphore = &r100_semaphore_ring_emit, - .cs_parse = &r300_cs_parse, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ib_test = &r100_ib_test, - .is_lockup = &r100_gpu_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r300_gfx_ring }, .irq = { .set = &r100_irq_set, @@ -403,9 +394,8 @@ static struct radeon_asic r300_asic = { .set_clock_gating = &radeon_legacy_set_clock_gating, }, .pflip = { - .pre_page_flip = &r100_pre_page_flip, .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .page_flip_pending = &r100_page_flip_pending, }, }; @@ -424,19 +414,7 @@ static struct radeon_asic r300_asic_pcie = { .set_page = &rv370_pcie_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r100_ring_ib_execute, - .emit_fence = &r300_fence_ring_emit, - .emit_semaphore = &r100_semaphore_ring_emit, - .cs_parse = &r300_cs_parse, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ib_test = &r100_ib_test, - .is_lockup = &r100_gpu_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r300_gfx_ring }, .irq = { .set = &r100_irq_set, @@ -482,9 +460,8 @@ static struct radeon_asic r300_asic_pcie = { .set_clock_gating = &radeon_legacy_set_clock_gating, }, .pflip = { - .pre_page_flip = &r100_pre_page_flip, .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .page_flip_pending = &r100_page_flip_pending, }, }; @@ -503,19 +480,7 @@ static struct radeon_asic r420_asic = { .set_page = &rv370_pcie_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r100_ring_ib_execute, - .emit_fence = &r300_fence_ring_emit, - .emit_semaphore = &r100_semaphore_ring_emit, - .cs_parse = &r300_cs_parse, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ib_test = &r100_ib_test, - .is_lockup = &r100_gpu_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r300_gfx_ring }, .irq = { .set = &r100_irq_set, @@ -561,9 +526,8 @@ static struct radeon_asic r420_asic = { .set_clock_gating = &radeon_atom_set_clock_gating, }, .pflip = { - .pre_page_flip = &r100_pre_page_flip, .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .page_flip_pending = &r100_page_flip_pending, }, }; @@ -582,19 +546,7 @@ static struct radeon_asic rs400_asic = { .set_page = &rs400_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r100_ring_ib_execute, - .emit_fence = &r300_fence_ring_emit, - .emit_semaphore = &r100_semaphore_ring_emit, - .cs_parse = &r300_cs_parse, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ib_test = &r100_ib_test, - .is_lockup = &r100_gpu_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r300_gfx_ring }, .irq = { .set = &r100_irq_set, @@ -640,9 +592,8 @@ static struct radeon_asic rs400_asic = { .set_clock_gating = &radeon_legacy_set_clock_gating, }, .pflip = { - .pre_page_flip = &r100_pre_page_flip, .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .page_flip_pending = &r100_page_flip_pending, }, }; @@ -661,19 +612,7 @@ static struct radeon_asic rs600_asic = { .set_page = &rs600_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r100_ring_ib_execute, - .emit_fence = &r300_fence_ring_emit, - .emit_semaphore = &r100_semaphore_ring_emit, - .cs_parse = &r300_cs_parse, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ib_test = &r100_ib_test, - .is_lockup = &r100_gpu_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r300_gfx_ring }, .irq = { .set = &rs600_irq_set, @@ -721,9 +660,8 @@ static struct radeon_asic rs600_asic = { .set_clock_gating = &radeon_atom_set_clock_gating, }, .pflip = { - .pre_page_flip = &rs600_pre_page_flip, .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .page_flip_pending = &rs600_page_flip_pending, }, }; @@ -742,19 +680,7 @@ static struct radeon_asic rs690_asic = { .set_page = &rs400_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r100_ring_ib_execute, - .emit_fence = &r300_fence_ring_emit, - .emit_semaphore = &r100_semaphore_ring_emit, - .cs_parse = &r300_cs_parse, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ib_test = &r100_ib_test, - .is_lockup = &r100_gpu_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r300_gfx_ring }, .irq = { .set = &rs600_irq_set, @@ -802,9 +728,8 @@ static struct radeon_asic rs690_asic = { .set_clock_gating = &radeon_atom_set_clock_gating, }, .pflip = { - .pre_page_flip = &rs600_pre_page_flip, .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .page_flip_pending = &rs600_page_flip_pending, }, }; @@ -823,19 +748,7 @@ static struct radeon_asic rv515_asic = { .set_page = &rv370_pcie_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r100_ring_ib_execute, - .emit_fence = &r300_fence_ring_emit, - .emit_semaphore = &r100_semaphore_ring_emit, - .cs_parse = &r300_cs_parse, - .ring_start = &rv515_ring_start, - .ring_test = &r100_ring_test, - .ib_test = &r100_ib_test, - .is_lockup = &r100_gpu_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r300_gfx_ring }, .irq = { .set = &rs600_irq_set, @@ -881,9 +794,8 @@ static struct radeon_asic rv515_asic = { .set_clock_gating = &radeon_atom_set_clock_gating, }, .pflip = { - .pre_page_flip = &rs600_pre_page_flip, .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .page_flip_pending = &rs600_page_flip_pending, }, }; @@ -902,19 +814,7 @@ static struct radeon_asic r520_asic = { .set_page = &rv370_pcie_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r100_ring_ib_execute, - .emit_fence = &r300_fence_ring_emit, - .emit_semaphore = &r100_semaphore_ring_emit, - .cs_parse = &r300_cs_parse, - .ring_start = &rv515_ring_start, - .ring_test = &r100_ring_test, - .ib_test = &r100_ib_test, - .is_lockup = &r100_gpu_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r300_gfx_ring }, .irq = { .set = &rs600_irq_set, @@ -960,12 +860,37 @@ static struct radeon_asic r520_asic = { .set_clock_gating = &radeon_atom_set_clock_gating, }, .pflip = { - .pre_page_flip = &rs600_pre_page_flip, .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .page_flip_pending = &rs600_page_flip_pending, }, }; +static struct radeon_asic_ring r600_gfx_ring = { + .ib_execute = &r600_ring_ib_execute, + .emit_fence = &r600_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &r600_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + .is_lockup = &r600_gfx_is_lockup, + .get_rptr = &r600_gfx_get_rptr, + .get_wptr = &r600_gfx_get_wptr, + .set_wptr = &r600_gfx_set_wptr, +}; + +static struct radeon_asic_ring r600_dma_ring = { + .ib_execute = &r600_dma_ring_ib_execute, + .emit_fence = &r600_dma_fence_ring_emit, + .emit_semaphore = &r600_dma_semaphore_ring_emit, + .cs_parse = &r600_dma_cs_parse, + .ring_test = &r600_dma_ring_test, + .ib_test = &r600_dma_ib_test, + .is_lockup = &r600_dma_is_lockup, + .get_rptr = &r600_dma_get_rptr, + .get_wptr = &r600_dma_get_wptr, + .set_wptr = &r600_dma_set_wptr, +}; + static struct radeon_asic r600_asic = { .init = &r600_init, .fini = &r600_fini, @@ -983,30 +908,8 @@ static struct radeon_asic r600_asic = { .set_page = &rs600_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r600_ring_ib_execute, - .emit_fence = &r600_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &r600_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &r600_gfx_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &r600_dma_ring_ib_execute, - .emit_fence = &r600_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &r600_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &r600_dma_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r600_gfx_ring, + [R600_RING_TYPE_DMA_INDEX] = &r600_dma_ring, }, .irq = { .set = &r600_irq_set, @@ -1022,12 +925,12 @@ static struct radeon_asic r600_asic = { .hdmi_setmode = &r600_hdmi_setmode, }, .copy = { - .blit = &r600_copy_blit, + .blit = &r600_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &r600_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, - .copy = &r600_copy_dma, - .copy_ring_index = R600_RING_TYPE_DMA_INDEX, + .copy = &r600_copy_cpdma, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, }, .surface = { .set_reg = r600_set_surface_reg, @@ -1055,9 +958,8 @@ static struct radeon_asic r600_asic = { .get_temperature = &rv6xx_get_temp, }, .pflip = { - .pre_page_flip = &rs600_pre_page_flip, .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .page_flip_pending = &rs600_page_flip_pending, }, }; @@ -1078,30 +980,8 @@ static struct radeon_asic rv6xx_asic = { .set_page = &rs600_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r600_ring_ib_execute, - .emit_fence = &r600_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &r600_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &r600_gfx_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &r600_dma_ring_ib_execute, - .emit_fence = &r600_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &r600_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &r600_dma_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r600_gfx_ring, + [R600_RING_TYPE_DMA_INDEX] = &r600_dma_ring, }, .irq = { .set = &r600_irq_set, @@ -1113,14 +993,16 @@ static struct radeon_asic rv6xx_asic = { .wait_for_vblank = &avivo_wait_for_vblank, .set_backlight_level = &atombios_set_backlight_level, .get_backlight_level = &atombios_get_backlight_level, + .hdmi_enable = &r600_hdmi_enable, + .hdmi_setmode = &r600_hdmi_setmode, }, .copy = { - .blit = &r600_copy_blit, + .blit = &r600_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &r600_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, - .copy = &r600_copy_dma, - .copy_ring_index = R600_RING_TYPE_DMA_INDEX, + .copy = &r600_copy_cpdma, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, }, .surface = { .set_reg = r600_set_surface_reg, @@ -1146,11 +1028,13 @@ static struct radeon_asic rv6xx_asic = { .set_pcie_lanes = &r600_set_pcie_lanes, .set_clock_gating = NULL, .get_temperature = &rv6xx_get_temp, + .set_uvd_clocks = &r600_set_uvd_clocks, }, .dpm = { .init = &rv6xx_dpm_init, .setup_asic = &rv6xx_setup_asic, .enable = &rv6xx_dpm_enable, + .late_enable = &r600_dpm_late_enable, .disable = &rv6xx_dpm_disable, .pre_set_power_state = &r600_dpm_pre_set_power_state, .set_power_state = &rv6xx_dpm_set_power_state, @@ -1161,11 +1045,11 @@ static struct radeon_asic rv6xx_asic = { .get_mclk = &rv6xx_dpm_get_mclk, .print_power_state = &rv6xx_dpm_print_power_state, .debugfs_print_current_performance_level = &rv6xx_dpm_debugfs_print_current_performance_level, + .force_performance_level = &rv6xx_dpm_force_performance_level, }, .pflip = { - .pre_page_flip = &rs600_pre_page_flip, .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .page_flip_pending = &rs600_page_flip_pending, }, }; @@ -1186,30 +1070,8 @@ static struct radeon_asic rs780_asic = { .set_page = &rs600_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r600_ring_ib_execute, - .emit_fence = &r600_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &r600_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &r600_gfx_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &r600_dma_ring_ib_execute, - .emit_fence = &r600_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &r600_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &r600_dma_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r600_gfx_ring, + [R600_RING_TYPE_DMA_INDEX] = &r600_dma_ring, }, .irq = { .set = &r600_irq_set, @@ -1225,12 +1087,12 @@ static struct radeon_asic rs780_asic = { .hdmi_setmode = &r600_hdmi_setmode, }, .copy = { - .blit = &r600_copy_blit, + .blit = &r600_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &r600_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, - .copy = &r600_copy_dma, - .copy_ring_index = R600_RING_TYPE_DMA_INDEX, + .copy = &r600_copy_cpdma, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, }, .surface = { .set_reg = r600_set_surface_reg, @@ -1256,11 +1118,13 @@ static struct radeon_asic rs780_asic = { .set_pcie_lanes = NULL, .set_clock_gating = NULL, .get_temperature = &rv6xx_get_temp, + .set_uvd_clocks = &r600_set_uvd_clocks, }, .dpm = { .init = &rs780_dpm_init, .setup_asic = &rs780_dpm_setup_asic, .enable = &rs780_dpm_enable, + .late_enable = &r600_dpm_late_enable, .disable = &rs780_dpm_disable, .pre_set_power_state = &r600_dpm_pre_set_power_state, .set_power_state = &rs780_dpm_set_power_state, @@ -1270,14 +1134,28 @@ static struct radeon_asic rs780_asic = { .get_sclk = &rs780_dpm_get_sclk, .get_mclk = &rs780_dpm_get_mclk, .print_power_state = &rs780_dpm_print_power_state, + .debugfs_print_current_performance_level = &rs780_dpm_debugfs_print_current_performance_level, + .force_performance_level = &rs780_dpm_force_performance_level, }, .pflip = { - .pre_page_flip = &rs600_pre_page_flip, .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .page_flip_pending = &rs600_page_flip_pending, }, }; +static struct radeon_asic_ring rv770_uvd_ring = { + .ib_execute = &uvd_v1_0_ib_execute, + .emit_fence = &uvd_v2_2_fence_emit, + .emit_semaphore = &uvd_v1_0_semaphore_emit, + .cs_parse = &radeon_uvd_cs_parse, + .ring_test = &uvd_v1_0_ring_test, + .ib_test = &uvd_v1_0_ib_test, + .is_lockup = &radeon_ring_test_lockup, + .get_rptr = &uvd_v1_0_get_rptr, + .get_wptr = &uvd_v1_0_get_wptr, + .set_wptr = &uvd_v1_0_set_wptr, +}; + static struct radeon_asic rv770_asic = { .init = &rv770_init, .fini = &rv770_fini, @@ -1295,42 +1173,9 @@ static struct radeon_asic rv770_asic = { .set_page = &rs600_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &r600_ring_ib_execute, - .emit_fence = &r600_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &r600_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &r600_gfx_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &r600_dma_ring_ib_execute, - .emit_fence = &r600_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &r600_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &r600_dma_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_UVD_INDEX] = { - .ib_execute = &r600_uvd_ib_execute, - .emit_fence = &r600_uvd_fence_emit, - .emit_semaphore = &r600_uvd_semaphore_emit, - .cs_parse = &radeon_uvd_cs_parse, - .ring_test = &r600_uvd_ring_test, - .ib_test = &r600_uvd_ib_test, - .is_lockup = &radeon_ring_test_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &r600_gfx_ring, + [R600_RING_TYPE_DMA_INDEX] = &r600_dma_ring, + [R600_RING_TYPE_UVD_INDEX] = &rv770_uvd_ring, }, .irq = { .set = &r600_irq_set, @@ -1343,10 +1188,10 @@ static struct radeon_asic rv770_asic = { .set_backlight_level = &atombios_set_backlight_level, .get_backlight_level = &atombios_get_backlight_level, .hdmi_enable = &r600_hdmi_enable, - .hdmi_setmode = &r600_hdmi_setmode, + .hdmi_setmode = &dce3_1_hdmi_setmode, }, .copy = { - .blit = &r600_copy_blit, + .blit = &r600_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &rv770_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, @@ -1383,6 +1228,7 @@ static struct radeon_asic rv770_asic = { .init = &rv770_dpm_init, .setup_asic = &rv770_dpm_setup_asic, .enable = &rv770_dpm_enable, + .late_enable = &rv770_dpm_late_enable, .disable = &rv770_dpm_disable, .pre_set_power_state = &r600_dpm_pre_set_power_state, .set_power_state = &rv770_dpm_set_power_state, @@ -1397,12 +1243,37 @@ static struct radeon_asic rv770_asic = { .vblank_too_short = &rv770_dpm_vblank_too_short, }, .pflip = { - .pre_page_flip = &rs600_pre_page_flip, .page_flip = &rv770_page_flip, - .post_page_flip = &rs600_post_page_flip, + .page_flip_pending = &rv770_page_flip_pending, }, }; +static struct radeon_asic_ring evergreen_gfx_ring = { + .ib_execute = &evergreen_ring_ib_execute, + .emit_fence = &r600_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + .is_lockup = &evergreen_gfx_is_lockup, + .get_rptr = &r600_gfx_get_rptr, + .get_wptr = &r600_gfx_get_wptr, + .set_wptr = &r600_gfx_set_wptr, +}; + +static struct radeon_asic_ring evergreen_dma_ring = { + .ib_execute = &evergreen_dma_ring_ib_execute, + .emit_fence = &evergreen_dma_fence_ring_emit, + .emit_semaphore = &r600_dma_semaphore_ring_emit, + .cs_parse = &evergreen_dma_cs_parse, + .ring_test = &r600_dma_ring_test, + .ib_test = &r600_dma_ib_test, + .is_lockup = &evergreen_dma_is_lockup, + .get_rptr = &r600_dma_get_rptr, + .get_wptr = &r600_dma_get_wptr, + .set_wptr = &r600_dma_set_wptr, +}; + static struct radeon_asic evergreen_asic = { .init = &evergreen_init, .fini = &evergreen_fini, @@ -1420,42 +1291,9 @@ static struct radeon_asic evergreen_asic = { .set_page = &rs600_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &evergreen_ring_ib_execute, - .emit_fence = &r600_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &evergreen_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &evergreen_gfx_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &evergreen_dma_ring_ib_execute, - .emit_fence = &evergreen_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &evergreen_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &evergreen_dma_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_UVD_INDEX] = { - .ib_execute = &r600_uvd_ib_execute, - .emit_fence = &r600_uvd_fence_emit, - .emit_semaphore = &r600_uvd_semaphore_emit, - .cs_parse = &radeon_uvd_cs_parse, - .ring_test = &r600_uvd_ring_test, - .ib_test = &r600_uvd_ib_test, - .is_lockup = &radeon_ring_test_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &evergreen_gfx_ring, + [R600_RING_TYPE_DMA_INDEX] = &evergreen_dma_ring, + [R600_RING_TYPE_UVD_INDEX] = &rv770_uvd_ring, }, .irq = { .set = &evergreen_irq_set, @@ -1471,7 +1309,7 @@ static struct radeon_asic evergreen_asic = { .hdmi_setmode = &evergreen_hdmi_setmode, }, .copy = { - .blit = &r600_copy_blit, + .blit = &r600_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &evergreen_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, @@ -1508,6 +1346,7 @@ static struct radeon_asic evergreen_asic = { .init = &cypress_dpm_init, .setup_asic = &cypress_dpm_setup_asic, .enable = &cypress_dpm_enable, + .late_enable = &rv770_dpm_late_enable, .disable = &cypress_dpm_disable, .pre_set_power_state = &r600_dpm_pre_set_power_state, .set_power_state = &cypress_dpm_set_power_state, @@ -1522,9 +1361,8 @@ static struct radeon_asic evergreen_asic = { .vblank_too_short = &cypress_dpm_vblank_too_short, }, .pflip = { - .pre_page_flip = &evergreen_pre_page_flip, .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .page_flip_pending = &evergreen_page_flip_pending, }, }; @@ -1545,42 +1383,9 @@ static struct radeon_asic sumo_asic = { .set_page = &rs600_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &evergreen_ring_ib_execute, - .emit_fence = &r600_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &evergreen_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &evergreen_gfx_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &evergreen_dma_ring_ib_execute, - .emit_fence = &evergreen_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &evergreen_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &evergreen_dma_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_UVD_INDEX] = { - .ib_execute = &r600_uvd_ib_execute, - .emit_fence = &r600_uvd_fence_emit, - .emit_semaphore = &r600_uvd_semaphore_emit, - .cs_parse = &radeon_uvd_cs_parse, - .ring_test = &r600_uvd_ring_test, - .ib_test = &r600_uvd_ib_test, - .is_lockup = &radeon_ring_test_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &evergreen_gfx_ring, + [R600_RING_TYPE_DMA_INDEX] = &evergreen_dma_ring, + [R600_RING_TYPE_UVD_INDEX] = &rv770_uvd_ring, }, .irq = { .set = &evergreen_irq_set, @@ -1596,7 +1401,7 @@ static struct radeon_asic sumo_asic = { .hdmi_setmode = &evergreen_hdmi_setmode, }, .copy = { - .blit = &r600_copy_blit, + .blit = &r600_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &evergreen_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, @@ -1633,6 +1438,7 @@ static struct radeon_asic sumo_asic = { .init = &sumo_dpm_init, .setup_asic = &sumo_dpm_setup_asic, .enable = &sumo_dpm_enable, + .late_enable = &sumo_dpm_late_enable, .disable = &sumo_dpm_disable, .pre_set_power_state = &sumo_dpm_pre_set_power_state, .set_power_state = &sumo_dpm_set_power_state, @@ -1646,9 +1452,8 @@ static struct radeon_asic sumo_asic = { .force_performance_level = &sumo_dpm_force_performance_level, }, .pflip = { - .pre_page_flip = &evergreen_pre_page_flip, .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .page_flip_pending = &evergreen_page_flip_pending, }, }; @@ -1669,42 +1474,9 @@ static struct radeon_asic btc_asic = { .set_page = &rs600_gart_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &evergreen_ring_ib_execute, - .emit_fence = &r600_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &evergreen_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &evergreen_gfx_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &evergreen_dma_ring_ib_execute, - .emit_fence = &evergreen_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &evergreen_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &evergreen_dma_is_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_UVD_INDEX] = { - .ib_execute = &r600_uvd_ib_execute, - .emit_fence = &r600_uvd_fence_emit, - .emit_semaphore = &r600_uvd_semaphore_emit, - .cs_parse = &radeon_uvd_cs_parse, - .ring_test = &r600_uvd_ring_test, - .ib_test = &r600_uvd_ib_test, - .is_lockup = &radeon_ring_test_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &evergreen_gfx_ring, + [R600_RING_TYPE_DMA_INDEX] = &evergreen_dma_ring, + [R600_RING_TYPE_UVD_INDEX] = &rv770_uvd_ring, }, .irq = { .set = &evergreen_irq_set, @@ -1720,7 +1492,7 @@ static struct radeon_asic btc_asic = { .hdmi_setmode = &evergreen_hdmi_setmode, }, .copy = { - .blit = &r600_copy_blit, + .blit = &r600_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &evergreen_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, @@ -1757,6 +1529,7 @@ static struct radeon_asic btc_asic = { .init = &btc_dpm_init, .setup_asic = &btc_dpm_setup_asic, .enable = &btc_dpm_enable, + .late_enable = &rv770_dpm_late_enable, .disable = &btc_dpm_disable, .pre_set_power_state = &btc_dpm_pre_set_power_state, .set_power_state = &btc_dpm_set_power_state, @@ -1766,17 +1539,59 @@ static struct radeon_asic btc_asic = { .get_sclk = &btc_dpm_get_sclk, .get_mclk = &btc_dpm_get_mclk, .print_power_state = &rv770_dpm_print_power_state, - .debugfs_print_current_performance_level = &rv770_dpm_debugfs_print_current_performance_level, + .debugfs_print_current_performance_level = &btc_dpm_debugfs_print_current_performance_level, .force_performance_level = &rv770_dpm_force_performance_level, .vblank_too_short = &btc_dpm_vblank_too_short, }, .pflip = { - .pre_page_flip = &evergreen_pre_page_flip, .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .page_flip_pending = &evergreen_page_flip_pending, }, }; +static struct radeon_asic_ring cayman_gfx_ring = { + .ib_execute = &cayman_ring_ib_execute, + .ib_parse = &evergreen_ib_parse, + .emit_fence = &cayman_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + .is_lockup = &cayman_gfx_is_lockup, + .vm_flush = &cayman_vm_flush, + .get_rptr = &cayman_gfx_get_rptr, + .get_wptr = &cayman_gfx_get_wptr, + .set_wptr = &cayman_gfx_set_wptr, +}; + +static struct radeon_asic_ring cayman_dma_ring = { + .ib_execute = &cayman_dma_ring_ib_execute, + .ib_parse = &evergreen_dma_ib_parse, + .emit_fence = &evergreen_dma_fence_ring_emit, + .emit_semaphore = &r600_dma_semaphore_ring_emit, + .cs_parse = &evergreen_dma_cs_parse, + .ring_test = &r600_dma_ring_test, + .ib_test = &r600_dma_ib_test, + .is_lockup = &cayman_dma_is_lockup, + .vm_flush = &cayman_dma_vm_flush, + .get_rptr = &cayman_dma_get_rptr, + .get_wptr = &cayman_dma_get_wptr, + .set_wptr = &cayman_dma_set_wptr +}; + +static struct radeon_asic_ring cayman_uvd_ring = { + .ib_execute = &uvd_v1_0_ib_execute, + .emit_fence = &uvd_v2_2_fence_emit, + .emit_semaphore = &uvd_v3_1_semaphore_emit, + .cs_parse = &radeon_uvd_cs_parse, + .ring_test = &uvd_v1_0_ring_test, + .ib_test = &uvd_v1_0_ib_test, + .is_lockup = &radeon_ring_test_lockup, + .get_rptr = &uvd_v1_0_get_rptr, + .get_wptr = &uvd_v1_0_get_wptr, + .set_wptr = &uvd_v1_0_set_wptr, +}; + static struct radeon_asic cayman_asic = { .init = &cayman_init, .fini = &cayman_fini, @@ -1796,92 +1611,15 @@ static struct radeon_asic cayman_asic = { .vm = { .init = &cayman_vm_init, .fini = &cayman_vm_fini, - .pt_ring_index = R600_RING_TYPE_DMA_INDEX, - .set_page = &cayman_vm_set_page, + .set_page = &cayman_dma_vm_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &cayman_ring_ib_execute, - .ib_parse = &evergreen_ib_parse, - .emit_fence = &cayman_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &evergreen_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &cayman_gfx_is_lockup, - .vm_flush = &cayman_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_CP1_INDEX] = { - .ib_execute = &cayman_ring_ib_execute, - .ib_parse = &evergreen_ib_parse, - .emit_fence = &cayman_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &evergreen_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &cayman_gfx_is_lockup, - .vm_flush = &cayman_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_CP2_INDEX] = { - .ib_execute = &cayman_ring_ib_execute, - .ib_parse = &evergreen_ib_parse, - .emit_fence = &cayman_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &evergreen_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &cayman_gfx_is_lockup, - .vm_flush = &cayman_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &cayman_dma_ring_ib_execute, - .ib_parse = &evergreen_dma_ib_parse, - .emit_fence = &evergreen_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &evergreen_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &cayman_dma_is_lockup, - .vm_flush = &cayman_dma_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_DMA1_INDEX] = { - .ib_execute = &cayman_dma_ring_ib_execute, - .ib_parse = &evergreen_dma_ib_parse, - .emit_fence = &evergreen_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &evergreen_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &cayman_dma_is_lockup, - .vm_flush = &cayman_dma_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_UVD_INDEX] = { - .ib_execute = &r600_uvd_ib_execute, - .emit_fence = &r600_uvd_fence_emit, - .emit_semaphore = &cayman_uvd_semaphore_emit, - .cs_parse = &radeon_uvd_cs_parse, - .ring_test = &r600_uvd_ring_test, - .ib_test = &r600_uvd_ib_test, - .is_lockup = &radeon_ring_test_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &cayman_gfx_ring, + [CAYMAN_RING_TYPE_CP1_INDEX] = &cayman_gfx_ring, + [CAYMAN_RING_TYPE_CP2_INDEX] = &cayman_gfx_ring, + [R600_RING_TYPE_DMA_INDEX] = &cayman_dma_ring, + [CAYMAN_RING_TYPE_DMA1_INDEX] = &cayman_dma_ring, + [R600_RING_TYPE_UVD_INDEX] = &cayman_uvd_ring, }, .irq = { .set = &evergreen_irq_set, @@ -1897,7 +1635,7 @@ static struct radeon_asic cayman_asic = { .hdmi_setmode = &evergreen_hdmi_setmode, }, .copy = { - .blit = &r600_copy_blit, + .blit = &r600_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &evergreen_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, @@ -1934,6 +1672,7 @@ static struct radeon_asic cayman_asic = { .init = &ni_dpm_init, .setup_asic = &ni_dpm_setup_asic, .enable = &ni_dpm_enable, + .late_enable = &rv770_dpm_late_enable, .disable = &ni_dpm_disable, .pre_set_power_state = &ni_dpm_pre_set_power_state, .set_power_state = &ni_dpm_set_power_state, @@ -1948,9 +1687,8 @@ static struct radeon_asic cayman_asic = { .vblank_too_short = &ni_dpm_vblank_too_short, }, .pflip = { - .pre_page_flip = &evergreen_pre_page_flip, .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .page_flip_pending = &evergreen_page_flip_pending, }, }; @@ -1973,92 +1711,15 @@ static struct radeon_asic trinity_asic = { .vm = { .init = &cayman_vm_init, .fini = &cayman_vm_fini, - .pt_ring_index = R600_RING_TYPE_DMA_INDEX, - .set_page = &cayman_vm_set_page, + .set_page = &cayman_dma_vm_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &cayman_ring_ib_execute, - .ib_parse = &evergreen_ib_parse, - .emit_fence = &cayman_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &evergreen_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &cayman_gfx_is_lockup, - .vm_flush = &cayman_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_CP1_INDEX] = { - .ib_execute = &cayman_ring_ib_execute, - .ib_parse = &evergreen_ib_parse, - .emit_fence = &cayman_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &evergreen_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &cayman_gfx_is_lockup, - .vm_flush = &cayman_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_CP2_INDEX] = { - .ib_execute = &cayman_ring_ib_execute, - .ib_parse = &evergreen_ib_parse, - .emit_fence = &cayman_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = &evergreen_cs_parse, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &cayman_gfx_is_lockup, - .vm_flush = &cayman_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &cayman_dma_ring_ib_execute, - .ib_parse = &evergreen_dma_ib_parse, - .emit_fence = &evergreen_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &evergreen_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &cayman_dma_is_lockup, - .vm_flush = &cayman_dma_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_DMA1_INDEX] = { - .ib_execute = &cayman_dma_ring_ib_execute, - .ib_parse = &evergreen_dma_ib_parse, - .emit_fence = &evergreen_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = &evergreen_dma_cs_parse, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &cayman_dma_is_lockup, - .vm_flush = &cayman_dma_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_UVD_INDEX] = { - .ib_execute = &r600_uvd_ib_execute, - .emit_fence = &r600_uvd_fence_emit, - .emit_semaphore = &cayman_uvd_semaphore_emit, - .cs_parse = &radeon_uvd_cs_parse, - .ring_test = &r600_uvd_ring_test, - .ib_test = &r600_uvd_ib_test, - .is_lockup = &radeon_ring_test_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &cayman_gfx_ring, + [CAYMAN_RING_TYPE_CP1_INDEX] = &cayman_gfx_ring, + [CAYMAN_RING_TYPE_CP2_INDEX] = &cayman_gfx_ring, + [R600_RING_TYPE_DMA_INDEX] = &cayman_dma_ring, + [CAYMAN_RING_TYPE_DMA1_INDEX] = &cayman_dma_ring, + [R600_RING_TYPE_UVD_INDEX] = &cayman_uvd_ring, }, .irq = { .set = &evergreen_irq_set, @@ -2070,9 +1731,11 @@ static struct radeon_asic trinity_asic = { .wait_for_vblank = &dce4_wait_for_vblank, .set_backlight_level = &atombios_set_backlight_level, .get_backlight_level = &atombios_get_backlight_level, + .hdmi_enable = &evergreen_hdmi_enable, + .hdmi_setmode = &evergreen_hdmi_setmode, }, .copy = { - .blit = &r600_copy_blit, + .blit = &r600_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &evergreen_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, @@ -2109,6 +1772,7 @@ static struct radeon_asic trinity_asic = { .init = &trinity_dpm_init, .setup_asic = &trinity_dpm_setup_asic, .enable = &trinity_dpm_enable, + .late_enable = &trinity_dpm_late_enable, .disable = &trinity_dpm_disable, .pre_set_power_state = &trinity_dpm_pre_set_power_state, .set_power_state = &trinity_dpm_set_power_state, @@ -2120,14 +1784,44 @@ static struct radeon_asic trinity_asic = { .print_power_state = &trinity_dpm_print_power_state, .debugfs_print_current_performance_level = &trinity_dpm_debugfs_print_current_performance_level, .force_performance_level = &trinity_dpm_force_performance_level, + .enable_bapm = &trinity_dpm_enable_bapm, }, .pflip = { - .pre_page_flip = &evergreen_pre_page_flip, .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .page_flip_pending = &evergreen_page_flip_pending, }, }; +static struct radeon_asic_ring si_gfx_ring = { + .ib_execute = &si_ring_ib_execute, + .ib_parse = &si_ib_parse, + .emit_fence = &si_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = NULL, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + .is_lockup = &si_gfx_is_lockup, + .vm_flush = &si_vm_flush, + .get_rptr = &cayman_gfx_get_rptr, + .get_wptr = &cayman_gfx_get_wptr, + .set_wptr = &cayman_gfx_set_wptr, +}; + +static struct radeon_asic_ring si_dma_ring = { + .ib_execute = &cayman_dma_ring_ib_execute, + .ib_parse = &evergreen_dma_ib_parse, + .emit_fence = &evergreen_dma_fence_ring_emit, + .emit_semaphore = &r600_dma_semaphore_ring_emit, + .cs_parse = NULL, + .ring_test = &r600_dma_ring_test, + .ib_test = &r600_dma_ib_test, + .is_lockup = &si_dma_is_lockup, + .vm_flush = &si_dma_vm_flush, + .get_rptr = &cayman_dma_get_rptr, + .get_wptr = &cayman_dma_get_wptr, + .set_wptr = &cayman_dma_set_wptr, +}; + static struct radeon_asic si_asic = { .init = &si_init, .fini = &si_fini, @@ -2147,92 +1841,15 @@ static struct radeon_asic si_asic = { .vm = { .init = &si_vm_init, .fini = &si_vm_fini, - .pt_ring_index = R600_RING_TYPE_DMA_INDEX, - .set_page = &si_vm_set_page, + .set_page = &si_dma_vm_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &si_ring_ib_execute, - .ib_parse = &si_ib_parse, - .emit_fence = &si_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &si_gfx_is_lockup, - .vm_flush = &si_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_CP1_INDEX] = { - .ib_execute = &si_ring_ib_execute, - .ib_parse = &si_ib_parse, - .emit_fence = &si_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &si_gfx_is_lockup, - .vm_flush = &si_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_CP2_INDEX] = { - .ib_execute = &si_ring_ib_execute, - .ib_parse = &si_ib_parse, - .emit_fence = &si_fence_ring_emit, - .emit_semaphore = &r600_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &r600_ring_test, - .ib_test = &r600_ib_test, - .is_lockup = &si_gfx_is_lockup, - .vm_flush = &si_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &cayman_dma_ring_ib_execute, - .ib_parse = &evergreen_dma_ib_parse, - .emit_fence = &evergreen_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &si_dma_is_lockup, - .vm_flush = &si_dma_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_DMA1_INDEX] = { - .ib_execute = &cayman_dma_ring_ib_execute, - .ib_parse = &evergreen_dma_ib_parse, - .emit_fence = &evergreen_dma_fence_ring_emit, - .emit_semaphore = &r600_dma_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &r600_dma_ring_test, - .ib_test = &r600_dma_ib_test, - .is_lockup = &si_dma_is_lockup, - .vm_flush = &si_dma_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_UVD_INDEX] = { - .ib_execute = &r600_uvd_ib_execute, - .emit_fence = &r600_uvd_fence_emit, - .emit_semaphore = &cayman_uvd_semaphore_emit, - .cs_parse = &radeon_uvd_cs_parse, - .ring_test = &r600_uvd_ring_test, - .ib_test = &r600_uvd_ib_test, - .is_lockup = &radeon_ring_test_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &si_gfx_ring, + [CAYMAN_RING_TYPE_CP1_INDEX] = &si_gfx_ring, + [CAYMAN_RING_TYPE_CP2_INDEX] = &si_gfx_ring, + [R600_RING_TYPE_DMA_INDEX] = &si_dma_ring, + [CAYMAN_RING_TYPE_DMA1_INDEX] = &si_dma_ring, + [R600_RING_TYPE_UVD_INDEX] = &cayman_uvd_ring, }, .irq = { .set = &si_irq_set, @@ -2244,9 +1861,11 @@ static struct radeon_asic si_asic = { .wait_for_vblank = &dce4_wait_for_vblank, .set_backlight_level = &atombios_set_backlight_level, .get_backlight_level = &atombios_get_backlight_level, + .hdmi_enable = &evergreen_hdmi_enable, + .hdmi_setmode = &evergreen_hdmi_setmode, }, .copy = { - .blit = NULL, + .blit = &r600_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &si_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, @@ -2283,6 +1902,7 @@ static struct radeon_asic si_asic = { .init = &si_dpm_init, .setup_asic = &si_dpm_setup_asic, .enable = &si_dpm_enable, + .late_enable = &si_dpm_late_enable, .disable = &si_dpm_disable, .pre_set_power_state = &si_dpm_pre_set_power_state, .set_power_state = &si_dpm_set_power_state, @@ -2297,12 +1917,69 @@ static struct radeon_asic si_asic = { .vblank_too_short = &ni_dpm_vblank_too_short, }, .pflip = { - .pre_page_flip = &evergreen_pre_page_flip, .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .page_flip_pending = &evergreen_page_flip_pending, }, }; +static struct radeon_asic_ring ci_gfx_ring = { + .ib_execute = &cik_ring_ib_execute, + .ib_parse = &cik_ib_parse, + .emit_fence = &cik_fence_gfx_ring_emit, + .emit_semaphore = &cik_semaphore_ring_emit, + .cs_parse = NULL, + .ring_test = &cik_ring_test, + .ib_test = &cik_ib_test, + .is_lockup = &cik_gfx_is_lockup, + .vm_flush = &cik_vm_flush, + .get_rptr = &cik_gfx_get_rptr, + .get_wptr = &cik_gfx_get_wptr, + .set_wptr = &cik_gfx_set_wptr, +}; + +static struct radeon_asic_ring ci_cp_ring = { + .ib_execute = &cik_ring_ib_execute, + .ib_parse = &cik_ib_parse, + .emit_fence = &cik_fence_compute_ring_emit, + .emit_semaphore = &cik_semaphore_ring_emit, + .cs_parse = NULL, + .ring_test = &cik_ring_test, + .ib_test = &cik_ib_test, + .is_lockup = &cik_gfx_is_lockup, + .vm_flush = &cik_vm_flush, + .get_rptr = &cik_compute_get_rptr, + .get_wptr = &cik_compute_get_wptr, + .set_wptr = &cik_compute_set_wptr, +}; + +static struct radeon_asic_ring ci_dma_ring = { + .ib_execute = &cik_sdma_ring_ib_execute, + .ib_parse = &cik_ib_parse, + .emit_fence = &cik_sdma_fence_ring_emit, + .emit_semaphore = &cik_sdma_semaphore_ring_emit, + .cs_parse = NULL, + .ring_test = &cik_sdma_ring_test, + .ib_test = &cik_sdma_ib_test, + .is_lockup = &cik_sdma_is_lockup, + .vm_flush = &cik_dma_vm_flush, + .get_rptr = &cik_sdma_get_rptr, + .get_wptr = &cik_sdma_get_wptr, + .set_wptr = &cik_sdma_set_wptr, +}; + +static struct radeon_asic_ring ci_vce_ring = { + .ib_execute = &radeon_vce_ib_execute, + .emit_fence = &radeon_vce_fence_emit, + .emit_semaphore = &radeon_vce_semaphore_emit, + .cs_parse = &radeon_vce_cs_parse, + .ring_test = &radeon_vce_ring_test, + .ib_test = &radeon_vce_ib_test, + .is_lockup = &radeon_ring_test_lockup, + .get_rptr = &vce_v1_0_get_rptr, + .get_wptr = &vce_v1_0_get_wptr, + .set_wptr = &vce_v1_0_set_wptr, +}; + static struct radeon_asic ci_asic = { .init = &cik_init, .fini = &cik_fini, @@ -2322,92 +1999,17 @@ static struct radeon_asic ci_asic = { .vm = { .init = &cik_vm_init, .fini = &cik_vm_fini, - .pt_ring_index = R600_RING_TYPE_DMA_INDEX, - .set_page = &cik_vm_set_page, + .set_page = &cik_sdma_vm_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &cik_ring_ib_execute, - .ib_parse = &cik_ib_parse, - .emit_fence = &cik_fence_gfx_ring_emit, - .emit_semaphore = &cik_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &cik_ring_test, - .ib_test = &cik_ib_test, - .is_lockup = &cik_gfx_is_lockup, - .vm_flush = &cik_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_CP1_INDEX] = { - .ib_execute = &cik_ring_ib_execute, - .ib_parse = &cik_ib_parse, - .emit_fence = &cik_fence_compute_ring_emit, - .emit_semaphore = &cik_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &cik_ring_test, - .ib_test = &cik_ib_test, - .is_lockup = &cik_gfx_is_lockup, - .vm_flush = &cik_vm_flush, - .get_rptr = &cik_compute_ring_get_rptr, - .get_wptr = &cik_compute_ring_get_wptr, - .set_wptr = &cik_compute_ring_set_wptr, - }, - [CAYMAN_RING_TYPE_CP2_INDEX] = { - .ib_execute = &cik_ring_ib_execute, - .ib_parse = &cik_ib_parse, - .emit_fence = &cik_fence_compute_ring_emit, - .emit_semaphore = &cik_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &cik_ring_test, - .ib_test = &cik_ib_test, - .is_lockup = &cik_gfx_is_lockup, - .vm_flush = &cik_vm_flush, - .get_rptr = &cik_compute_ring_get_rptr, - .get_wptr = &cik_compute_ring_get_wptr, - .set_wptr = &cik_compute_ring_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &cik_sdma_ring_ib_execute, - .ib_parse = &cik_ib_parse, - .emit_fence = &cik_sdma_fence_ring_emit, - .emit_semaphore = &cik_sdma_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &cik_sdma_ring_test, - .ib_test = &cik_sdma_ib_test, - .is_lockup = &cik_sdma_is_lockup, - .vm_flush = &cik_dma_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_DMA1_INDEX] = { - .ib_execute = &cik_sdma_ring_ib_execute, - .ib_parse = &cik_ib_parse, - .emit_fence = &cik_sdma_fence_ring_emit, - .emit_semaphore = &cik_sdma_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &cik_sdma_ring_test, - .ib_test = &cik_sdma_ib_test, - .is_lockup = &cik_sdma_is_lockup, - .vm_flush = &cik_dma_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_UVD_INDEX] = { - .ib_execute = &r600_uvd_ib_execute, - .emit_fence = &r600_uvd_fence_emit, - .emit_semaphore = &cayman_uvd_semaphore_emit, - .cs_parse = &radeon_uvd_cs_parse, - .ring_test = &r600_uvd_ring_test, - .ib_test = &r600_uvd_ib_test, - .is_lockup = &radeon_ring_test_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &ci_gfx_ring, + [CAYMAN_RING_TYPE_CP1_INDEX] = &ci_cp_ring, + [CAYMAN_RING_TYPE_CP2_INDEX] = &ci_cp_ring, + [R600_RING_TYPE_DMA_INDEX] = &ci_dma_ring, + [CAYMAN_RING_TYPE_DMA1_INDEX] = &ci_dma_ring, + [R600_RING_TYPE_UVD_INDEX] = &cayman_uvd_ring, + [TN_RING_TYPE_VCE1_INDEX] = &ci_vce_ring, + [TN_RING_TYPE_VCE2_INDEX] = &ci_vce_ring, }, .irq = { .set = &cik_irq_set, @@ -2417,9 +2019,13 @@ static struct radeon_asic ci_asic = { .bandwidth_update = &dce8_bandwidth_update, .get_vblank_counter = &evergreen_get_vblank_counter, .wait_for_vblank = &dce4_wait_for_vblank, + .set_backlight_level = &atombios_set_backlight_level, + .get_backlight_level = &atombios_get_backlight_level, + .hdmi_enable = &evergreen_hdmi_enable, + .hdmi_setmode = &evergreen_hdmi_setmode, }, .copy = { - .blit = NULL, + .blit = &cik_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &cik_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, @@ -2450,11 +2056,31 @@ static struct radeon_asic ci_asic = { .set_pcie_lanes = NULL, .set_clock_gating = NULL, .set_uvd_clocks = &cik_set_uvd_clocks, + .set_vce_clocks = &cik_set_vce_clocks, + .get_temperature = &ci_get_temp, + }, + .dpm = { + .init = &ci_dpm_init, + .setup_asic = &ci_dpm_setup_asic, + .enable = &ci_dpm_enable, + .late_enable = &ci_dpm_late_enable, + .disable = &ci_dpm_disable, + .pre_set_power_state = &ci_dpm_pre_set_power_state, + .set_power_state = &ci_dpm_set_power_state, + .post_set_power_state = &ci_dpm_post_set_power_state, + .display_configuration_changed = &ci_dpm_display_configuration_changed, + .fini = &ci_dpm_fini, + .get_sclk = &ci_dpm_get_sclk, + .get_mclk = &ci_dpm_get_mclk, + .print_power_state = &ci_dpm_print_power_state, + .debugfs_print_current_performance_level = &ci_dpm_debugfs_print_current_performance_level, + .force_performance_level = &ci_dpm_force_performance_level, + .vblank_too_short = &ci_dpm_vblank_too_short, + .powergate_uvd = &ci_dpm_powergate_uvd, }, .pflip = { - .pre_page_flip = &evergreen_pre_page_flip, .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .page_flip_pending = &evergreen_page_flip_pending, }, }; @@ -2477,92 +2103,17 @@ static struct radeon_asic kv_asic = { .vm = { .init = &cik_vm_init, .fini = &cik_vm_fini, - .pt_ring_index = R600_RING_TYPE_DMA_INDEX, - .set_page = &cik_vm_set_page, + .set_page = &cik_sdma_vm_set_page, }, .ring = { - [RADEON_RING_TYPE_GFX_INDEX] = { - .ib_execute = &cik_ring_ib_execute, - .ib_parse = &cik_ib_parse, - .emit_fence = &cik_fence_gfx_ring_emit, - .emit_semaphore = &cik_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &cik_ring_test, - .ib_test = &cik_ib_test, - .is_lockup = &cik_gfx_is_lockup, - .vm_flush = &cik_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_CP1_INDEX] = { - .ib_execute = &cik_ring_ib_execute, - .ib_parse = &cik_ib_parse, - .emit_fence = &cik_fence_compute_ring_emit, - .emit_semaphore = &cik_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &cik_ring_test, - .ib_test = &cik_ib_test, - .is_lockup = &cik_gfx_is_lockup, - .vm_flush = &cik_vm_flush, - .get_rptr = &cik_compute_ring_get_rptr, - .get_wptr = &cik_compute_ring_get_wptr, - .set_wptr = &cik_compute_ring_set_wptr, - }, - [CAYMAN_RING_TYPE_CP2_INDEX] = { - .ib_execute = &cik_ring_ib_execute, - .ib_parse = &cik_ib_parse, - .emit_fence = &cik_fence_compute_ring_emit, - .emit_semaphore = &cik_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &cik_ring_test, - .ib_test = &cik_ib_test, - .is_lockup = &cik_gfx_is_lockup, - .vm_flush = &cik_vm_flush, - .get_rptr = &cik_compute_ring_get_rptr, - .get_wptr = &cik_compute_ring_get_wptr, - .set_wptr = &cik_compute_ring_set_wptr, - }, - [R600_RING_TYPE_DMA_INDEX] = { - .ib_execute = &cik_sdma_ring_ib_execute, - .ib_parse = &cik_ib_parse, - .emit_fence = &cik_sdma_fence_ring_emit, - .emit_semaphore = &cik_sdma_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &cik_sdma_ring_test, - .ib_test = &cik_sdma_ib_test, - .is_lockup = &cik_sdma_is_lockup, - .vm_flush = &cik_dma_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [CAYMAN_RING_TYPE_DMA1_INDEX] = { - .ib_execute = &cik_sdma_ring_ib_execute, - .ib_parse = &cik_ib_parse, - .emit_fence = &cik_sdma_fence_ring_emit, - .emit_semaphore = &cik_sdma_semaphore_ring_emit, - .cs_parse = NULL, - .ring_test = &cik_sdma_ring_test, - .ib_test = &cik_sdma_ib_test, - .is_lockup = &cik_sdma_is_lockup, - .vm_flush = &cik_dma_vm_flush, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - }, - [R600_RING_TYPE_UVD_INDEX] = { - .ib_execute = &r600_uvd_ib_execute, - .emit_fence = &r600_uvd_fence_emit, - .emit_semaphore = &cayman_uvd_semaphore_emit, - .cs_parse = &radeon_uvd_cs_parse, - .ring_test = &r600_uvd_ring_test, - .ib_test = &r600_uvd_ib_test, - .is_lockup = &radeon_ring_test_lockup, - .get_rptr = &radeon_ring_generic_get_rptr, - .get_wptr = &radeon_ring_generic_get_wptr, - .set_wptr = &radeon_ring_generic_set_wptr, - } + [RADEON_RING_TYPE_GFX_INDEX] = &ci_gfx_ring, + [CAYMAN_RING_TYPE_CP1_INDEX] = &ci_cp_ring, + [CAYMAN_RING_TYPE_CP2_INDEX] = &ci_cp_ring, + [R600_RING_TYPE_DMA_INDEX] = &ci_dma_ring, + [CAYMAN_RING_TYPE_DMA1_INDEX] = &ci_dma_ring, + [R600_RING_TYPE_UVD_INDEX] = &cayman_uvd_ring, + [TN_RING_TYPE_VCE1_INDEX] = &ci_vce_ring, + [TN_RING_TYPE_VCE2_INDEX] = &ci_vce_ring, }, .irq = { .set = &cik_irq_set, @@ -2572,9 +2123,13 @@ static struct radeon_asic kv_asic = { .bandwidth_update = &dce8_bandwidth_update, .get_vblank_counter = &evergreen_get_vblank_counter, .wait_for_vblank = &dce4_wait_for_vblank, + .set_backlight_level = &atombios_set_backlight_level, + .get_backlight_level = &atombios_get_backlight_level, + .hdmi_enable = &evergreen_hdmi_enable, + .hdmi_setmode = &evergreen_hdmi_setmode, }, .copy = { - .blit = NULL, + .blit = &cik_copy_cpdma, .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, .dma = &cik_copy_dma, .dma_ring_index = R600_RING_TYPE_DMA_INDEX, @@ -2605,11 +2160,31 @@ static struct radeon_asic kv_asic = { .set_pcie_lanes = NULL, .set_clock_gating = NULL, .set_uvd_clocks = &cik_set_uvd_clocks, + .set_vce_clocks = &cik_set_vce_clocks, + .get_temperature = &kv_get_temp, + }, + .dpm = { + .init = &kv_dpm_init, + .setup_asic = &kv_dpm_setup_asic, + .enable = &kv_dpm_enable, + .late_enable = &kv_dpm_late_enable, + .disable = &kv_dpm_disable, + .pre_set_power_state = &kv_dpm_pre_set_power_state, + .set_power_state = &kv_dpm_set_power_state, + .post_set_power_state = &kv_dpm_post_set_power_state, + .display_configuration_changed = &kv_dpm_display_configuration_changed, + .fini = &kv_dpm_fini, + .get_sclk = &kv_dpm_get_sclk, + .get_mclk = &kv_dpm_get_mclk, + .print_power_state = &kv_dpm_print_power_state, + .debugfs_print_current_performance_level = &kv_dpm_debugfs_print_current_performance_level, + .force_performance_level = &kv_dpm_force_performance_level, + .powergate_uvd = &kv_dpm_powergate_uvd, + .enable_bapm = &kv_dpm_enable_bapm, }, .pflip = { - .pre_page_flip = &evergreen_pre_page_flip, .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .page_flip_pending = &evergreen_page_flip_pending, }, }; @@ -2774,19 +2349,210 @@ int radeon_asic_init(struct radeon_device *rdev) rdev->has_uvd = false; else rdev->has_uvd = true; + switch (rdev->family) { + case CHIP_TAHITI: + rdev->cg_flags = + RADEON_CG_SUPPORT_GFX_MGCG | + RADEON_CG_SUPPORT_GFX_MGLS | + /*RADEON_CG_SUPPORT_GFX_CGCG |*/ + RADEON_CG_SUPPORT_GFX_CGLS | + RADEON_CG_SUPPORT_GFX_CGTS | + RADEON_CG_SUPPORT_GFX_CP_LS | + RADEON_CG_SUPPORT_MC_MGCG | + RADEON_CG_SUPPORT_SDMA_MGCG | + RADEON_CG_SUPPORT_BIF_LS | + RADEON_CG_SUPPORT_VCE_MGCG | + RADEON_CG_SUPPORT_UVD_MGCG | + RADEON_CG_SUPPORT_HDP_LS | + RADEON_CG_SUPPORT_HDP_MGCG; + rdev->pg_flags = 0; + break; + case CHIP_PITCAIRN: + rdev->cg_flags = + RADEON_CG_SUPPORT_GFX_MGCG | + RADEON_CG_SUPPORT_GFX_MGLS | + /*RADEON_CG_SUPPORT_GFX_CGCG |*/ + RADEON_CG_SUPPORT_GFX_CGLS | + RADEON_CG_SUPPORT_GFX_CGTS | + RADEON_CG_SUPPORT_GFX_CP_LS | + RADEON_CG_SUPPORT_GFX_RLC_LS | + RADEON_CG_SUPPORT_MC_LS | + RADEON_CG_SUPPORT_MC_MGCG | + RADEON_CG_SUPPORT_SDMA_MGCG | + RADEON_CG_SUPPORT_BIF_LS | + RADEON_CG_SUPPORT_VCE_MGCG | + RADEON_CG_SUPPORT_UVD_MGCG | + RADEON_CG_SUPPORT_HDP_LS | + RADEON_CG_SUPPORT_HDP_MGCG; + rdev->pg_flags = 0; + break; + case CHIP_VERDE: + rdev->cg_flags = + RADEON_CG_SUPPORT_GFX_MGCG | + RADEON_CG_SUPPORT_GFX_MGLS | + /*RADEON_CG_SUPPORT_GFX_CGCG |*/ + RADEON_CG_SUPPORT_GFX_CGLS | + RADEON_CG_SUPPORT_GFX_CGTS | + RADEON_CG_SUPPORT_GFX_CP_LS | + RADEON_CG_SUPPORT_GFX_RLC_LS | + RADEON_CG_SUPPORT_MC_LS | + RADEON_CG_SUPPORT_MC_MGCG | + RADEON_CG_SUPPORT_SDMA_MGCG | + RADEON_CG_SUPPORT_BIF_LS | + RADEON_CG_SUPPORT_VCE_MGCG | + RADEON_CG_SUPPORT_UVD_MGCG | + RADEON_CG_SUPPORT_HDP_LS | + RADEON_CG_SUPPORT_HDP_MGCG; + rdev->pg_flags = 0 | + /*RADEON_PG_SUPPORT_GFX_PG | */ + RADEON_PG_SUPPORT_SDMA; + break; + case CHIP_OLAND: + rdev->cg_flags = + RADEON_CG_SUPPORT_GFX_MGCG | + RADEON_CG_SUPPORT_GFX_MGLS | + /*RADEON_CG_SUPPORT_GFX_CGCG |*/ + RADEON_CG_SUPPORT_GFX_CGLS | + RADEON_CG_SUPPORT_GFX_CGTS | + RADEON_CG_SUPPORT_GFX_CP_LS | + RADEON_CG_SUPPORT_GFX_RLC_LS | + RADEON_CG_SUPPORT_MC_LS | + RADEON_CG_SUPPORT_MC_MGCG | + RADEON_CG_SUPPORT_SDMA_MGCG | + RADEON_CG_SUPPORT_BIF_LS | + RADEON_CG_SUPPORT_UVD_MGCG | + RADEON_CG_SUPPORT_HDP_LS | + RADEON_CG_SUPPORT_HDP_MGCG; + rdev->pg_flags = 0; + break; + case CHIP_HAINAN: + rdev->cg_flags = + RADEON_CG_SUPPORT_GFX_MGCG | + RADEON_CG_SUPPORT_GFX_MGLS | + /*RADEON_CG_SUPPORT_GFX_CGCG |*/ + RADEON_CG_SUPPORT_GFX_CGLS | + RADEON_CG_SUPPORT_GFX_CGTS | + RADEON_CG_SUPPORT_GFX_CP_LS | + RADEON_CG_SUPPORT_GFX_RLC_LS | + RADEON_CG_SUPPORT_MC_LS | + RADEON_CG_SUPPORT_MC_MGCG | + RADEON_CG_SUPPORT_SDMA_MGCG | + RADEON_CG_SUPPORT_BIF_LS | + RADEON_CG_SUPPORT_HDP_LS | + RADEON_CG_SUPPORT_HDP_MGCG; + rdev->pg_flags = 0; + break; + default: + rdev->cg_flags = 0; + rdev->pg_flags = 0; + break; + } break; case CHIP_BONAIRE: + case CHIP_HAWAII: rdev->asic = &ci_asic; rdev->num_crtc = 6; + rdev->has_uvd = true; + if (rdev->family == CHIP_BONAIRE) { + rdev->cg_flags = + RADEON_CG_SUPPORT_GFX_MGCG | + RADEON_CG_SUPPORT_GFX_MGLS | + RADEON_CG_SUPPORT_GFX_CGCG | + RADEON_CG_SUPPORT_GFX_CGLS | + RADEON_CG_SUPPORT_GFX_CGTS | + RADEON_CG_SUPPORT_GFX_CGTS_LS | + RADEON_CG_SUPPORT_GFX_CP_LS | + RADEON_CG_SUPPORT_MC_LS | + RADEON_CG_SUPPORT_MC_MGCG | + RADEON_CG_SUPPORT_SDMA_MGCG | + RADEON_CG_SUPPORT_SDMA_LS | + RADEON_CG_SUPPORT_BIF_LS | + RADEON_CG_SUPPORT_VCE_MGCG | + RADEON_CG_SUPPORT_UVD_MGCG | + RADEON_CG_SUPPORT_HDP_LS | + RADEON_CG_SUPPORT_HDP_MGCG; + rdev->pg_flags = 0; + } else { + rdev->cg_flags = + RADEON_CG_SUPPORT_GFX_MGCG | + RADEON_CG_SUPPORT_GFX_MGLS | + RADEON_CG_SUPPORT_GFX_CGCG | + RADEON_CG_SUPPORT_GFX_CGLS | + RADEON_CG_SUPPORT_GFX_CGTS | + RADEON_CG_SUPPORT_GFX_CP_LS | + RADEON_CG_SUPPORT_MC_LS | + RADEON_CG_SUPPORT_MC_MGCG | + RADEON_CG_SUPPORT_SDMA_MGCG | + RADEON_CG_SUPPORT_SDMA_LS | + RADEON_CG_SUPPORT_BIF_LS | + RADEON_CG_SUPPORT_VCE_MGCG | + RADEON_CG_SUPPORT_UVD_MGCG | + RADEON_CG_SUPPORT_HDP_LS | + RADEON_CG_SUPPORT_HDP_MGCG; + rdev->pg_flags = 0; + } break; case CHIP_KAVERI: case CHIP_KABINI: + case CHIP_MULLINS: rdev->asic = &kv_asic; /* set num crtcs */ - if (rdev->family == CHIP_KAVERI) + if (rdev->family == CHIP_KAVERI) { rdev->num_crtc = 4; - else + rdev->cg_flags = + RADEON_CG_SUPPORT_GFX_MGCG | + RADEON_CG_SUPPORT_GFX_MGLS | + RADEON_CG_SUPPORT_GFX_CGCG | + RADEON_CG_SUPPORT_GFX_CGLS | + RADEON_CG_SUPPORT_GFX_CGTS | + RADEON_CG_SUPPORT_GFX_CGTS_LS | + RADEON_CG_SUPPORT_GFX_CP_LS | + RADEON_CG_SUPPORT_SDMA_MGCG | + RADEON_CG_SUPPORT_SDMA_LS | + RADEON_CG_SUPPORT_BIF_LS | + RADEON_CG_SUPPORT_VCE_MGCG | + RADEON_CG_SUPPORT_UVD_MGCG | + RADEON_CG_SUPPORT_HDP_LS | + RADEON_CG_SUPPORT_HDP_MGCG; + rdev->pg_flags = 0; + /*RADEON_PG_SUPPORT_GFX_PG | + RADEON_PG_SUPPORT_GFX_SMG | + RADEON_PG_SUPPORT_GFX_DMG | + RADEON_PG_SUPPORT_UVD | + RADEON_PG_SUPPORT_VCE | + RADEON_PG_SUPPORT_CP | + RADEON_PG_SUPPORT_GDS | + RADEON_PG_SUPPORT_RLC_SMU_HS | + RADEON_PG_SUPPORT_ACP | + RADEON_PG_SUPPORT_SAMU;*/ + } else { rdev->num_crtc = 2; + rdev->cg_flags = + RADEON_CG_SUPPORT_GFX_MGCG | + RADEON_CG_SUPPORT_GFX_MGLS | + RADEON_CG_SUPPORT_GFX_CGCG | + RADEON_CG_SUPPORT_GFX_CGLS | + RADEON_CG_SUPPORT_GFX_CGTS | + RADEON_CG_SUPPORT_GFX_CGTS_LS | + RADEON_CG_SUPPORT_GFX_CP_LS | + RADEON_CG_SUPPORT_SDMA_MGCG | + RADEON_CG_SUPPORT_SDMA_LS | + RADEON_CG_SUPPORT_BIF_LS | + RADEON_CG_SUPPORT_VCE_MGCG | + RADEON_CG_SUPPORT_UVD_MGCG | + RADEON_CG_SUPPORT_HDP_LS | + RADEON_CG_SUPPORT_HDP_MGCG; + rdev->pg_flags = 0; + /*RADEON_PG_SUPPORT_GFX_PG | + RADEON_PG_SUPPORT_GFX_SMG | + RADEON_PG_SUPPORT_UVD | + RADEON_PG_SUPPORT_VCE | + RADEON_PG_SUPPORT_CP | + RADEON_PG_SUPPORT_GDS | + RADEON_PG_SUPPORT_RLC_SMU_HS | + RADEON_PG_SUPPORT_SAMU;*/ + } + rdev->has_uvd = true; break; default: /* FIXME: not supported yet */ diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h index 45d0693cddd..01e7c0ad8f0 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.h +++ b/drivers/gpu/drm/radeon/radeon_asic.h @@ -47,13 +47,6 @@ u8 atombios_get_backlight_level(struct radeon_encoder *radeon_encoder); void radeon_legacy_set_backlight_level(struct radeon_encoder *radeon_encoder, u8 level); u8 radeon_legacy_get_backlight_level(struct radeon_encoder *radeon_encoder); -u32 radeon_ring_generic_get_rptr(struct radeon_device *rdev, - struct radeon_ring *ring); -u32 radeon_ring_generic_get_wptr(struct radeon_device *rdev, - struct radeon_ring *ring); -void radeon_ring_generic_set_wptr(struct radeon_device *rdev, - struct radeon_ring *ring); - /* * r100,rv100,rs100,rv200,rs200 */ @@ -74,13 +67,14 @@ bool r100_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp); int r100_asic_reset(struct radeon_device *rdev); u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc); void r100_pci_gart_tlb_flush(struct radeon_device *rdev); -int r100_pci_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr); +void r100_pci_gart_set_page(struct radeon_device *rdev, unsigned i, + uint64_t addr); void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring); int r100_irq_set(struct radeon_device *rdev); int r100_irq_process(struct radeon_device *rdev); void r100_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); -void r100_semaphore_ring_emit(struct radeon_device *rdev, +bool r100_semaphore_ring_emit(struct radeon_device *rdev, struct radeon_ring *cp, struct radeon_semaphore *semaphore, bool emit_wait); @@ -142,12 +136,19 @@ extern void r100_pm_prepare(struct radeon_device *rdev); extern void r100_pm_finish(struct radeon_device *rdev); extern void r100_pm_init_profile(struct radeon_device *rdev); extern void r100_pm_get_dynpm_state(struct radeon_device *rdev); -extern void r100_pre_page_flip(struct radeon_device *rdev, int crtc); -extern u32 r100_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base); -extern void r100_post_page_flip(struct radeon_device *rdev, int crtc); +extern void r100_page_flip(struct radeon_device *rdev, int crtc, + u64 crtc_base); +extern bool r100_page_flip_pending(struct radeon_device *rdev, int crtc); extern void r100_wait_for_vblank(struct radeon_device *rdev, int crtc); extern int r100_mc_wait_for_idle(struct radeon_device *rdev); +u32 r100_gfx_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +u32 r100_gfx_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void r100_gfx_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); + /* * r200,rv250,rs300,rv280 */ @@ -171,7 +172,8 @@ extern void r300_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); extern int r300_cs_parse(struct radeon_cs_parser *p); extern void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev); -extern int rv370_pcie_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr); +extern void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i, + uint64_t addr); extern void rv370_set_pcie_lanes(struct radeon_device *rdev, int lanes); extern int rv370_get_pcie_lanes(struct radeon_device *rdev); extern void r300_set_reg_safe(struct radeon_device *rdev); @@ -206,7 +208,8 @@ extern void rs400_fini(struct radeon_device *rdev); extern int rs400_suspend(struct radeon_device *rdev); extern int rs400_resume(struct radeon_device *rdev); void rs400_gart_tlb_flush(struct radeon_device *rdev); -int rs400_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr); +void rs400_gart_set_page(struct radeon_device *rdev, unsigned i, + uint64_t addr); uint32_t rs400_mc_rreg(struct radeon_device *rdev, uint32_t reg); void rs400_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); int rs400_gart_init(struct radeon_device *rdev); @@ -229,7 +232,8 @@ int rs600_irq_process(struct radeon_device *rdev); void rs600_irq_disable(struct radeon_device *rdev); u32 rs600_get_vblank_counter(struct radeon_device *rdev, int crtc); void rs600_gart_tlb_flush(struct radeon_device *rdev); -int rs600_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr); +void rs600_gart_set_page(struct radeon_device *rdev, unsigned i, + uint64_t addr); uint32_t rs600_mc_rreg(struct radeon_device *rdev, uint32_t reg); void rs600_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); void rs600_bandwidth_update(struct radeon_device *rdev); @@ -241,9 +245,9 @@ void rs600_hpd_set_polarity(struct radeon_device *rdev, extern void rs600_pm_misc(struct radeon_device *rdev); extern void rs600_pm_prepare(struct radeon_device *rdev); extern void rs600_pm_finish(struct radeon_device *rdev); -extern void rs600_pre_page_flip(struct radeon_device *rdev, int crtc); -extern u32 rs600_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base); -extern void rs600_post_page_flip(struct radeon_device *rdev, int crtc); +extern void rs600_page_flip(struct radeon_device *rdev, int crtc, + u64 crtc_base); +extern bool rs600_page_flip_pending(struct radeon_device *rdev, int crtc); void rs600_set_safe_registers(struct radeon_device *rdev); extern void avivo_wait_for_vblank(struct radeon_device *rdev, int crtc); extern int rs600_mc_wait_for_idle(struct radeon_device *rdev); @@ -313,13 +317,13 @@ int r600_cs_parse(struct radeon_cs_parser *p); int r600_dma_cs_parse(struct radeon_cs_parser *p); void r600_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); -void r600_semaphore_ring_emit(struct radeon_device *rdev, +bool r600_semaphore_ring_emit(struct radeon_device *rdev, struct radeon_ring *cp, struct radeon_semaphore *semaphore, bool emit_wait); void r600_dma_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); -void r600_dma_semaphore_ring_emit(struct radeon_device *rdev, +bool r600_dma_semaphore_ring_emit(struct radeon_device *rdev, struct radeon_ring *ring, struct radeon_semaphore *semaphore, bool emit_wait); @@ -336,10 +340,9 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring); void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *cp); int r600_dma_ring_test(struct radeon_device *rdev, struct radeon_ring *cp); -int r600_uvd_ring_test(struct radeon_device *rdev, struct radeon_ring *ring); -int r600_copy_blit(struct radeon_device *rdev, - uint64_t src_offset, uint64_t dst_offset, - unsigned num_gpu_pages, struct radeon_fence **fence); +int r600_copy_cpdma(struct radeon_device *rdev, + uint64_t src_offset, uint64_t dst_offset, + unsigned num_gpu_pages, struct radeon_fence **fence); int r600_copy_dma(struct radeon_device *rdev, uint64_t src_offset, uint64_t dst_offset, unsigned num_gpu_pages, struct radeon_fence **fence); @@ -368,9 +371,13 @@ int r600_count_pipe_bits(uint32_t val); int r600_mc_wait_for_idle(struct radeon_device *rdev); int r600_pcie_gart_init(struct radeon_device *rdev); void r600_scratch_init(struct radeon_device *rdev); -int r600_blit_init(struct radeon_device *rdev); -void r600_blit_fini(struct radeon_device *rdev); int r600_init_microcode(struct radeon_device *rdev); +u32 r600_gfx_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +u32 r600_gfx_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void r600_gfx_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); /* r600 irq */ int r600_irq_process(struct radeon_device *rdev); int r600_irq_init(struct radeon_device *rdev); @@ -382,28 +389,32 @@ void r600_disable_interrupts(struct radeon_device *rdev); void r600_rlc_stop(struct radeon_device *rdev); /* r600 audio */ int r600_audio_init(struct radeon_device *rdev); -struct r600_audio r600_audio_status(struct radeon_device *rdev); +struct r600_audio_pin r600_audio_status(struct radeon_device *rdev); void r600_audio_fini(struct radeon_device *rdev); +void r600_audio_set_dto(struct drm_encoder *encoder, u32 clock); +void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder, void *buffer, + size_t size); +void r600_hdmi_update_ACR(struct drm_encoder *encoder, uint32_t clock); +void r600_hdmi_audio_workaround(struct drm_encoder *encoder); int r600_hdmi_buffer_status_changed(struct drm_encoder *encoder); void r600_hdmi_update_audio_settings(struct drm_encoder *encoder); void r600_hdmi_enable(struct drm_encoder *encoder, bool enable); void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode); -/* r600 blit */ -int r600_blit_prepare_copy(struct radeon_device *rdev, unsigned num_gpu_pages, - struct radeon_fence **fence, struct radeon_sa_bo **vb, - struct radeon_semaphore **sem); -void r600_blit_done_copy(struct radeon_device *rdev, struct radeon_fence **fence, - struct radeon_sa_bo *vb, struct radeon_semaphore *sem); -void r600_kms_blit_copy(struct radeon_device *rdev, - u64 src_gpu_addr, u64 dst_gpu_addr, - unsigned num_gpu_pages, - struct radeon_sa_bo *vb); int r600_mc_wait_for_idle(struct radeon_device *rdev); u32 r600_get_xclk(struct radeon_device *rdev); uint64_t r600_get_gpu_clock_counter(struct radeon_device *rdev); int rv6xx_get_temp(struct radeon_device *rdev); +int r600_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk); int r600_dpm_pre_set_power_state(struct radeon_device *rdev); void r600_dpm_post_set_power_state(struct radeon_device *rdev); +int r600_dpm_late_enable(struct radeon_device *rdev); +/* r600 dma */ +uint32_t r600_dma_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +uint32_t r600_dma_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void r600_dma_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); /* rv6xx dpm */ int rv6xx_dpm_init(struct radeon_device *rdev); int rv6xx_dpm_enable(struct radeon_device *rdev); @@ -418,6 +429,8 @@ void rv6xx_dpm_print_power_state(struct radeon_device *rdev, struct radeon_ps *ps); void rv6xx_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, struct seq_file *m); +int rv6xx_dpm_force_performance_level(struct radeon_device *rdev, + enum radeon_dpm_forced_level level); /* rs780 dpm */ int rs780_dpm_init(struct radeon_device *rdev); int rs780_dpm_enable(struct radeon_device *rdev); @@ -430,19 +443,10 @@ u32 rs780_dpm_get_sclk(struct radeon_device *rdev, bool low); u32 rs780_dpm_get_mclk(struct radeon_device *rdev, bool low); void rs780_dpm_print_power_state(struct radeon_device *rdev, struct radeon_ps *ps); - -/* uvd */ -int r600_uvd_init(struct radeon_device *rdev); -int r600_uvd_rbc_start(struct radeon_device *rdev); -void r600_uvd_rbc_stop(struct radeon_device *rdev); -int r600_uvd_ib_test(struct radeon_device *rdev, struct radeon_ring *ring); -void r600_uvd_fence_emit(struct radeon_device *rdev, - struct radeon_fence *fence); -void r600_uvd_semaphore_emit(struct radeon_device *rdev, - struct radeon_ring *ring, - struct radeon_semaphore *semaphore, - bool emit_wait); -void r600_uvd_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); +void rs780_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, + struct seq_file *m); +int rs780_dpm_force_performance_level(struct radeon_device *rdev, + enum radeon_dpm_forced_level level); /* * rv770,rv730,rv710,rv740 @@ -452,7 +456,8 @@ void rv770_fini(struct radeon_device *rdev); int rv770_suspend(struct radeon_device *rdev); int rv770_resume(struct radeon_device *rdev); void rv770_pm_misc(struct radeon_device *rdev); -u32 rv770_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base); +void rv770_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base); +bool rv770_page_flip_pending(struct radeon_device *rdev, int crtc); void r700_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc); void r700_cp_stop(struct radeon_device *rdev); void r700_cp_fini(struct radeon_device *rdev); @@ -461,12 +466,14 @@ int rv770_copy_dma(struct radeon_device *rdev, unsigned num_gpu_pages, struct radeon_fence **fence); u32 rv770_get_xclk(struct radeon_device *rdev); -int rv770_uvd_resume(struct radeon_device *rdev); int rv770_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk); int rv770_get_temp(struct radeon_device *rdev); +/* hdmi */ +void dce3_1_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode); /* rv7xx pm */ int rv770_dpm_init(struct radeon_device *rdev); int rv770_dpm_enable(struct radeon_device *rdev); +int rv770_dpm_late_enable(struct radeon_device *rdev); void rv770_dpm_disable(struct radeon_device *rdev); int rv770_dpm_set_power_state(struct radeon_device *rdev); void rv770_dpm_setup_asic(struct radeon_device *rdev); @@ -518,12 +525,11 @@ extern void sumo_pm_init_profile(struct radeon_device *rdev); extern void btc_pm_init_profile(struct radeon_device *rdev); int sumo_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk); int evergreen_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk); -extern void evergreen_pre_page_flip(struct radeon_device *rdev, int crtc); -extern u32 evergreen_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base); -extern void evergreen_post_page_flip(struct radeon_device *rdev, int crtc); +extern void evergreen_page_flip(struct radeon_device *rdev, int crtc, + u64 crtc_base); +extern bool evergreen_page_flip_pending(struct radeon_device *rdev, int crtc); extern void dce4_wait_for_vblank(struct radeon_device *rdev, int crtc); void evergreen_disable_interrupt_state(struct radeon_device *rdev); -int evergreen_blit_init(struct radeon_device *rdev); int evergreen_mc_wait_for_idle(struct radeon_device *rdev); void evergreen_dma_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); @@ -557,8 +563,11 @@ void btc_dpm_fini(struct radeon_device *rdev); u32 btc_dpm_get_sclk(struct radeon_device *rdev, bool low); u32 btc_dpm_get_mclk(struct radeon_device *rdev, bool low); bool btc_dpm_vblank_too_short(struct radeon_device *rdev); +void btc_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, + struct seq_file *m); int sumo_dpm_init(struct radeon_device *rdev); int sumo_dpm_enable(struct radeon_device *rdev); +int sumo_dpm_late_enable(struct radeon_device *rdev); void sumo_dpm_disable(struct radeon_device *rdev); int sumo_dpm_pre_set_power_state(struct radeon_device *rdev); int sumo_dpm_set_power_state(struct radeon_device *rdev); @@ -580,10 +589,6 @@ int sumo_dpm_force_performance_level(struct radeon_device *rdev, */ void cayman_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); -void cayman_uvd_semaphore_emit(struct radeon_device *rdev, - struct radeon_ring *ring, - struct radeon_semaphore *semaphore, - bool emit_wait); void cayman_pcie_gart_tlb_flush(struct radeon_device *rdev); int cayman_init(struct radeon_device *rdev); void cayman_fini(struct radeon_device *rdev); @@ -595,19 +600,33 @@ int cayman_vm_init(struct radeon_device *rdev); void cayman_vm_fini(struct radeon_device *rdev); void cayman_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm); uint32_t cayman_vm_page_flags(struct radeon_device *rdev, uint32_t flags); -void cayman_vm_set_page(struct radeon_device *rdev, - struct radeon_ib *ib, - uint64_t pe, - uint64_t addr, unsigned count, - uint32_t incr, uint32_t flags); int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib); int evergreen_dma_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib); void cayman_dma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); bool cayman_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring); bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring); +void cayman_dma_vm_set_page(struct radeon_device *rdev, + struct radeon_ib *ib, + uint64_t pe, + uint64_t addr, unsigned count, + uint32_t incr, uint32_t flags); + void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm); +u32 cayman_gfx_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +u32 cayman_gfx_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void cayman_gfx_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +uint32_t cayman_dma_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +uint32_t cayman_dma_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void cayman_dma_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); + int ni_dpm_init(struct radeon_device *rdev); void ni_dpm_setup_asic(struct radeon_device *rdev); int ni_dpm_enable(struct radeon_device *rdev); @@ -627,6 +646,7 @@ int ni_dpm_force_performance_level(struct radeon_device *rdev, bool ni_dpm_vblank_too_short(struct radeon_device *rdev); int trinity_dpm_init(struct radeon_device *rdev); int trinity_dpm_enable(struct radeon_device *rdev); +int trinity_dpm_late_enable(struct radeon_device *rdev); void trinity_dpm_disable(struct radeon_device *rdev); int trinity_dpm_pre_set_power_state(struct radeon_device *rdev); int trinity_dpm_set_power_state(struct radeon_device *rdev); @@ -642,9 +662,12 @@ void trinity_dpm_debugfs_print_current_performance_level(struct radeon_device *r struct seq_file *m); int trinity_dpm_force_performance_level(struct radeon_device *rdev, enum radeon_dpm_forced_level level); +void trinity_dpm_enable_bapm(struct radeon_device *rdev, bool enable); /* DCE6 - SI */ void dce6_bandwidth_update(struct radeon_device *rdev); +int dce6_audio_init(struct radeon_device *rdev); +void dce6_audio_fini(struct radeon_device *rdev); /* * si @@ -664,17 +687,17 @@ int si_irq_set(struct radeon_device *rdev); int si_irq_process(struct radeon_device *rdev); int si_vm_init(struct radeon_device *rdev); void si_vm_fini(struct radeon_device *rdev); -void si_vm_set_page(struct radeon_device *rdev, - struct radeon_ib *ib, - uint64_t pe, - uint64_t addr, unsigned count, - uint32_t incr, uint32_t flags); void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm); int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib); int si_copy_dma(struct radeon_device *rdev, uint64_t src_offset, uint64_t dst_offset, unsigned num_gpu_pages, struct radeon_fence **fence); +void si_dma_vm_set_page(struct radeon_device *rdev, + struct radeon_ib *ib, + uint64_t pe, + uint64_t addr, unsigned count, + uint32_t incr, uint32_t flags); void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm); u32 si_get_xclk(struct radeon_device *rdev); uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev); @@ -683,6 +706,7 @@ int si_get_temp(struct radeon_device *rdev); int si_dpm_init(struct radeon_device *rdev); void si_dpm_setup_asic(struct radeon_device *rdev); int si_dpm_enable(struct radeon_device *rdev); +int si_dpm_late_enable(struct radeon_device *rdev); void si_dpm_disable(struct radeon_device *rdev); int si_dpm_pre_set_power_state(struct radeon_device *rdev); int si_dpm_set_power_state(struct radeon_device *rdev); @@ -705,10 +729,10 @@ u32 cik_get_xclk(struct radeon_device *rdev); uint32_t cik_pciep_rreg(struct radeon_device *rdev, uint32_t reg); void cik_pciep_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); int cik_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk); -int cik_uvd_resume(struct radeon_device *rdev); +int cik_set_vce_clocks(struct radeon_device *rdev, u32 evclk, u32 ecclk); void cik_sdma_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); -void cik_sdma_semaphore_ring_emit(struct radeon_device *rdev, +bool cik_sdma_semaphore_ring_emit(struct radeon_device *rdev, struct radeon_ring *ring, struct radeon_semaphore *semaphore, bool emit_wait); @@ -717,6 +741,10 @@ int cik_copy_dma(struct radeon_device *rdev, uint64_t src_offset, uint64_t dst_offset, unsigned num_gpu_pages, struct radeon_fence **fence); +int cik_copy_cpdma(struct radeon_device *rdev, + uint64_t src_offset, uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence **fence); int cik_sdma_ring_test(struct radeon_device *rdev, struct radeon_ring *ring); int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring); bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring); @@ -724,7 +752,7 @@ void cik_fence_gfx_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); void cik_fence_compute_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); -void cik_semaphore_ring_emit(struct radeon_device *rdev, +bool cik_semaphore_ring_emit(struct radeon_device *rdev, struct radeon_ring *cp, struct radeon_semaphore *semaphore, bool emit_wait); @@ -743,18 +771,122 @@ int cik_irq_process(struct radeon_device *rdev); int cik_vm_init(struct radeon_device *rdev); void cik_vm_fini(struct radeon_device *rdev); void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm); -void cik_vm_set_page(struct radeon_device *rdev, - struct radeon_ib *ib, - uint64_t pe, - uint64_t addr, unsigned count, - uint32_t incr, uint32_t flags); +void cik_sdma_vm_set_page(struct radeon_device *rdev, + struct radeon_ib *ib, + uint64_t pe, + uint64_t addr, unsigned count, + uint32_t incr, uint32_t flags); void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm); int cik_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib); -u32 cik_compute_ring_get_rptr(struct radeon_device *rdev, - struct radeon_ring *ring); -u32 cik_compute_ring_get_wptr(struct radeon_device *rdev, - struct radeon_ring *ring); -void cik_compute_ring_set_wptr(struct radeon_device *rdev, - struct radeon_ring *ring); +u32 cik_gfx_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +u32 cik_gfx_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void cik_gfx_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +u32 cik_compute_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +u32 cik_compute_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void cik_compute_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +u32 cik_sdma_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +u32 cik_sdma_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void cik_sdma_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +int ci_get_temp(struct radeon_device *rdev); +int kv_get_temp(struct radeon_device *rdev); + +int ci_dpm_init(struct radeon_device *rdev); +int ci_dpm_enable(struct radeon_device *rdev); +int ci_dpm_late_enable(struct radeon_device *rdev); +void ci_dpm_disable(struct radeon_device *rdev); +int ci_dpm_pre_set_power_state(struct radeon_device *rdev); +int ci_dpm_set_power_state(struct radeon_device *rdev); +void ci_dpm_post_set_power_state(struct radeon_device *rdev); +void ci_dpm_setup_asic(struct radeon_device *rdev); +void ci_dpm_display_configuration_changed(struct radeon_device *rdev); +void ci_dpm_fini(struct radeon_device *rdev); +u32 ci_dpm_get_sclk(struct radeon_device *rdev, bool low); +u32 ci_dpm_get_mclk(struct radeon_device *rdev, bool low); +void ci_dpm_print_power_state(struct radeon_device *rdev, + struct radeon_ps *ps); +void ci_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, + struct seq_file *m); +int ci_dpm_force_performance_level(struct radeon_device *rdev, + enum radeon_dpm_forced_level level); +bool ci_dpm_vblank_too_short(struct radeon_device *rdev); +void ci_dpm_powergate_uvd(struct radeon_device *rdev, bool gate); + +int kv_dpm_init(struct radeon_device *rdev); +int kv_dpm_enable(struct radeon_device *rdev); +int kv_dpm_late_enable(struct radeon_device *rdev); +void kv_dpm_disable(struct radeon_device *rdev); +int kv_dpm_pre_set_power_state(struct radeon_device *rdev); +int kv_dpm_set_power_state(struct radeon_device *rdev); +void kv_dpm_post_set_power_state(struct radeon_device *rdev); +void kv_dpm_setup_asic(struct radeon_device *rdev); +void kv_dpm_display_configuration_changed(struct radeon_device *rdev); +void kv_dpm_fini(struct radeon_device *rdev); +u32 kv_dpm_get_sclk(struct radeon_device *rdev, bool low); +u32 kv_dpm_get_mclk(struct radeon_device *rdev, bool low); +void kv_dpm_print_power_state(struct radeon_device *rdev, + struct radeon_ps *ps); +void kv_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, + struct seq_file *m); +int kv_dpm_force_performance_level(struct radeon_device *rdev, + enum radeon_dpm_forced_level level); +void kv_dpm_powergate_uvd(struct radeon_device *rdev, bool gate); +void kv_dpm_enable_bapm(struct radeon_device *rdev, bool enable); + +/* uvd v1.0 */ +uint32_t uvd_v1_0_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +uint32_t uvd_v1_0_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void uvd_v1_0_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); + +int uvd_v1_0_init(struct radeon_device *rdev); +void uvd_v1_0_fini(struct radeon_device *rdev); +int uvd_v1_0_start(struct radeon_device *rdev); +void uvd_v1_0_stop(struct radeon_device *rdev); + +int uvd_v1_0_ring_test(struct radeon_device *rdev, struct radeon_ring *ring); +int uvd_v1_0_ib_test(struct radeon_device *rdev, struct radeon_ring *ring); +bool uvd_v1_0_semaphore_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait); +void uvd_v1_0_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); + +/* uvd v2.2 */ +int uvd_v2_2_resume(struct radeon_device *rdev); +void uvd_v2_2_fence_emit(struct radeon_device *rdev, + struct radeon_fence *fence); + +/* uvd v3.1 */ +bool uvd_v3_1_semaphore_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait); + +/* uvd v4.2 */ +int uvd_v4_2_resume(struct radeon_device *rdev); + +/* vce v1.0 */ +uint32_t vce_v1_0_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +uint32_t vce_v1_0_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void vce_v1_0_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +int vce_v1_0_init(struct radeon_device *rdev); +int vce_v1_0_start(struct radeon_device *rdev); + +/* vce v2.0 */ +int vce_v2_0_resume(struct radeon_device *rdev); #endif diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c index fbdaff55556..173f378428a 100644 --- a/drivers/gpu/drm/radeon/radeon_atombios.c +++ b/drivers/gpu/drm/radeon/radeon_atombios.c @@ -30,27 +30,10 @@ #include "atom.h" #include "atom-bits.h" -/* from radeon_encoder.c */ -extern uint32_t -radeon_get_encoder_enum(struct drm_device *dev, uint32_t supported_device, - uint8_t dac); -extern void radeon_link_encoder_connector(struct drm_device *dev); extern void radeon_add_atom_encoder(struct drm_device *dev, uint32_t encoder_enum, uint32_t supported_device, u16 caps); -/* from radeon_connector.c */ -extern void -radeon_add_atom_connector(struct drm_device *dev, - uint32_t connector_id, - uint32_t supported_device, - int connector_type, - struct radeon_i2c_bus_rec *i2c_bus, - uint32_t igp_lane_info, - uint16_t connector_object_id, - struct radeon_hpd *hpd, - struct radeon_router *router); - /* from radeon_legacy_encoder.c */ extern void radeon_add_legacy_encoder(struct drm_device *dev, uint32_t encoder_enum, @@ -163,8 +146,8 @@ static struct radeon_i2c_bus_rec radeon_lookup_i2c_gpio(struct radeon_device *rd num_indices = (size - sizeof(ATOM_COMMON_TABLE_HEADER)) / sizeof(ATOM_GPIO_I2C_ASSIGMENT); + gpio = &i2c_info->asGPIO_Info[0]; for (i = 0; i < num_indices; i++) { - gpio = &i2c_info->asGPIO_Info[i]; radeon_lookup_i2c_gpio_quirks(rdev, gpio, i); @@ -172,6 +155,8 @@ static struct radeon_i2c_bus_rec radeon_lookup_i2c_gpio(struct radeon_device *rd i2c = radeon_get_bus_rec_for_i2c_gpio(gpio); break; } + gpio = (ATOM_GPIO_I2C_ASSIGMENT *) + ((u8 *)gpio + sizeof(ATOM_GPIO_I2C_ASSIGMENT)); } } @@ -195,9 +180,8 @@ void radeon_atombios_i2c_init(struct radeon_device *rdev) num_indices = (size - sizeof(ATOM_COMMON_TABLE_HEADER)) / sizeof(ATOM_GPIO_I2C_ASSIGMENT); + gpio = &i2c_info->asGPIO_Info[0]; for (i = 0; i < num_indices; i++) { - gpio = &i2c_info->asGPIO_Info[i]; - radeon_lookup_i2c_gpio_quirks(rdev, gpio, i); i2c = radeon_get_bus_rec_for_i2c_gpio(gpio); @@ -206,12 +190,14 @@ void radeon_atombios_i2c_init(struct radeon_device *rdev) sprintf(stmp, "0x%x", i2c.i2c_id); rdev->i2c_bus[i] = radeon_i2c_create(rdev->ddev, &i2c, stmp); } + gpio = (ATOM_GPIO_I2C_ASSIGMENT *) + ((u8 *)gpio + sizeof(ATOM_GPIO_I2C_ASSIGMENT)); } } } static struct radeon_gpio_rec radeon_lookup_gpio(struct radeon_device *rdev, - u8 id) + u8 id) { struct atom_context *ctx = rdev->mode_info.atom_context; struct radeon_gpio_rec gpio; @@ -230,8 +216,8 @@ static struct radeon_gpio_rec radeon_lookup_gpio(struct radeon_device *rdev, num_indices = (size - sizeof(ATOM_COMMON_TABLE_HEADER)) / sizeof(ATOM_GPIO_PIN_ASSIGNMENT); + pin = gpio_info->asGPIO_Pin; for (i = 0; i < num_indices; i++) { - pin = &gpio_info->asGPIO_Pin[i]; if (id == pin->ucGPIO_ID) { gpio.id = pin->ucGPIO_ID; gpio.reg = le16_to_cpu(pin->usGpioPin_AIndex) * 4; @@ -239,6 +225,8 @@ static struct radeon_gpio_rec radeon_lookup_gpio(struct radeon_device *rdev, gpio.valid = true; break; } + pin = (ATOM_GPIO_PIN_ASSIGNMENT *) + ((u8 *)pin + sizeof(ATOM_GPIO_PIN_ASSIGNMENT)); } } @@ -711,13 +699,16 @@ bool radeon_get_atom_connector_info_from_object_table(struct drm_device *dev) (ATOM_SRC_DST_TABLE_FOR_ONE_OBJECT *) (ctx->bios + data_offset + le16_to_cpu(router_obj->asObjects[k].usSrcDstTableOffset)); + u8 *num_dst_objs = (u8 *) + ((u8 *)router_src_dst_table + 1 + + (router_src_dst_table->ucNumberOfSrc * 2)); + u16 *dst_objs = (u16 *)(num_dst_objs + 1); int enum_id; router.router_id = router_obj_id; - for (enum_id = 0; enum_id < router_src_dst_table->ucNumberOfDst; - enum_id++) { + for (enum_id = 0; enum_id < (*num_dst_objs); enum_id++) { if (le16_to_cpu(path->usConnObjectId) == - le16_to_cpu(router_src_dst_table->usDstObjectID[enum_id])) + le16_to_cpu(dst_objs[enum_id])) break; } @@ -1236,11 +1227,19 @@ bool radeon_atom_get_clock_info(struct drm_device *dev) rdev->clock.default_dispclk = le32_to_cpu(firmware_info->info_21.ulDefaultDispEngineClkFreq); if (rdev->clock.default_dispclk == 0) { - if (ASIC_IS_DCE5(rdev)) + if (ASIC_IS_DCE6(rdev)) + rdev->clock.default_dispclk = 60000; /* 600 Mhz */ + else if (ASIC_IS_DCE5(rdev)) rdev->clock.default_dispclk = 54000; /* 540 Mhz */ else rdev->clock.default_dispclk = 60000; /* 600 Mhz */ } + /* set a reasonable default for DP */ + if (ASIC_IS_DCE6(rdev) && (rdev->clock.default_dispclk < 53900)) { + DRM_INFO("Changing default dispclk from %dMhz to 600Mhz\n", + rdev->clock.default_dispclk / 100); + rdev->clock.default_dispclk = 60000; + } rdev->clock.dp_extclk = le16_to_cpu(firmware_info->info_21.usUniphyDPModeExtClkFreq); rdev->clock.current_dispclk = rdev->clock.default_dispclk; @@ -1359,6 +1358,7 @@ bool radeon_atombios_get_ppll_ss_info(struct radeon_device *rdev, int index = GetIndexIntoMasterTable(DATA, PPLL_SS_Info); uint16_t data_offset, size; struct _ATOM_SPREAD_SPECTRUM_INFO *ss_info; + struct _ATOM_SPREAD_SPECTRUM_ASSIGNMENT *ss_assign; uint8_t frev, crev; int i, num_indices; @@ -1370,18 +1370,21 @@ bool radeon_atombios_get_ppll_ss_info(struct radeon_device *rdev, num_indices = (size - sizeof(ATOM_COMMON_TABLE_HEADER)) / sizeof(ATOM_SPREAD_SPECTRUM_ASSIGNMENT); - + ss_assign = (struct _ATOM_SPREAD_SPECTRUM_ASSIGNMENT*) + ((u8 *)&ss_info->asSS_Info[0]); for (i = 0; i < num_indices; i++) { - if (ss_info->asSS_Info[i].ucSS_Id == id) { + if (ss_assign->ucSS_Id == id) { ss->percentage = - le16_to_cpu(ss_info->asSS_Info[i].usSpreadSpectrumPercentage); - ss->type = ss_info->asSS_Info[i].ucSpreadSpectrumType; - ss->step = ss_info->asSS_Info[i].ucSS_Step; - ss->delay = ss_info->asSS_Info[i].ucSS_Delay; - ss->range = ss_info->asSS_Info[i].ucSS_Range; - ss->refdiv = ss_info->asSS_Info[i].ucRecommendedRef_Div; + le16_to_cpu(ss_assign->usSpreadSpectrumPercentage); + ss->type = ss_assign->ucSpreadSpectrumType; + ss->step = ss_assign->ucSS_Step; + ss->delay = ss_assign->ucSS_Delay; + ss->range = ss_assign->ucSS_Range; + ss->refdiv = ss_assign->ucRecommendedRef_Div; return true; } + ss_assign = (struct _ATOM_SPREAD_SPECTRUM_ASSIGNMENT*) + ((u8 *)ss_assign + sizeof(struct _ATOM_SPREAD_SPECTRUM_ASSIGNMENT)); } } return false; @@ -1469,6 +1472,12 @@ union asic_ss_info { struct _ATOM_ASIC_INTERNAL_SS_INFO_V3 info_3; }; +union asic_ss_assignment { + struct _ATOM_ASIC_SS_ASSIGNMENT v1; + struct _ATOM_ASIC_SS_ASSIGNMENT_V2 v2; + struct _ATOM_ASIC_SS_ASSIGNMENT_V3 v3; +}; + bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev, struct radeon_atom_ss *ss, int id, u32 clock) @@ -1477,9 +1486,19 @@ bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev, int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info); uint16_t data_offset, size; union asic_ss_info *ss_info; + union asic_ss_assignment *ss_assign; uint8_t frev, crev; int i, num_indices; + if (id == ASIC_INTERNAL_MEMORY_SS) { + if (!(rdev->mode_info.firmware_flags & ATOM_BIOS_INFO_MEMORY_CLOCK_SS_SUPPORT)) + return false; + } + if (id == ASIC_INTERNAL_ENGINE_SS) { + if (!(rdev->mode_info.firmware_flags & ATOM_BIOS_INFO_ENGINE_CLOCK_SS_SUPPORT)) + return false; + } + memset(ss, 0, sizeof(struct radeon_atom_ss)); if (atom_parse_data_header(mode_info->atom_context, index, &size, &frev, &crev, &data_offset)) { @@ -1492,45 +1511,59 @@ bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev, num_indices = (size - sizeof(ATOM_COMMON_TABLE_HEADER)) / sizeof(ATOM_ASIC_SS_ASSIGNMENT); + ss_assign = (union asic_ss_assignment *)((u8 *)&ss_info->info.asSpreadSpectrum[0]); for (i = 0; i < num_indices; i++) { - if ((ss_info->info.asSpreadSpectrum[i].ucClockIndication == id) && - (clock <= le32_to_cpu(ss_info->info.asSpreadSpectrum[i].ulTargetClockRange))) { + if ((ss_assign->v1.ucClockIndication == id) && + (clock <= le32_to_cpu(ss_assign->v1.ulTargetClockRange))) { ss->percentage = - le16_to_cpu(ss_info->info.asSpreadSpectrum[i].usSpreadSpectrumPercentage); - ss->type = ss_info->info.asSpreadSpectrum[i].ucSpreadSpectrumMode; - ss->rate = le16_to_cpu(ss_info->info.asSpreadSpectrum[i].usSpreadRateInKhz); + le16_to_cpu(ss_assign->v1.usSpreadSpectrumPercentage); + ss->type = ss_assign->v1.ucSpreadSpectrumMode; + ss->rate = le16_to_cpu(ss_assign->v1.usSpreadRateInKhz); + ss->percentage_divider = 100; return true; } + ss_assign = (union asic_ss_assignment *) + ((u8 *)ss_assign + sizeof(ATOM_ASIC_SS_ASSIGNMENT)); } break; case 2: num_indices = (size - sizeof(ATOM_COMMON_TABLE_HEADER)) / sizeof(ATOM_ASIC_SS_ASSIGNMENT_V2); + ss_assign = (union asic_ss_assignment *)((u8 *)&ss_info->info_2.asSpreadSpectrum[0]); for (i = 0; i < num_indices; i++) { - if ((ss_info->info_2.asSpreadSpectrum[i].ucClockIndication == id) && - (clock <= le32_to_cpu(ss_info->info_2.asSpreadSpectrum[i].ulTargetClockRange))) { + if ((ss_assign->v2.ucClockIndication == id) && + (clock <= le32_to_cpu(ss_assign->v2.ulTargetClockRange))) { ss->percentage = - le16_to_cpu(ss_info->info_2.asSpreadSpectrum[i].usSpreadSpectrumPercentage); - ss->type = ss_info->info_2.asSpreadSpectrum[i].ucSpreadSpectrumMode; - ss->rate = le16_to_cpu(ss_info->info_2.asSpreadSpectrum[i].usSpreadRateIn10Hz); + le16_to_cpu(ss_assign->v2.usSpreadSpectrumPercentage); + ss->type = ss_assign->v2.ucSpreadSpectrumMode; + ss->rate = le16_to_cpu(ss_assign->v2.usSpreadRateIn10Hz); + ss->percentage_divider = 100; if ((crev == 2) && ((id == ASIC_INTERNAL_ENGINE_SS) || (id == ASIC_INTERNAL_MEMORY_SS))) ss->rate /= 100; return true; } + ss_assign = (union asic_ss_assignment *) + ((u8 *)ss_assign + sizeof(ATOM_ASIC_SS_ASSIGNMENT_V2)); } break; case 3: num_indices = (size - sizeof(ATOM_COMMON_TABLE_HEADER)) / sizeof(ATOM_ASIC_SS_ASSIGNMENT_V3); + ss_assign = (union asic_ss_assignment *)((u8 *)&ss_info->info_3.asSpreadSpectrum[0]); for (i = 0; i < num_indices; i++) { - if ((ss_info->info_3.asSpreadSpectrum[i].ucClockIndication == id) && - (clock <= le32_to_cpu(ss_info->info_3.asSpreadSpectrum[i].ulTargetClockRange))) { + if ((ss_assign->v3.ucClockIndication == id) && + (clock <= le32_to_cpu(ss_assign->v3.ulTargetClockRange))) { ss->percentage = - le16_to_cpu(ss_info->info_3.asSpreadSpectrum[i].usSpreadSpectrumPercentage); - ss->type = ss_info->info_3.asSpreadSpectrum[i].ucSpreadSpectrumMode; - ss->rate = le16_to_cpu(ss_info->info_3.asSpreadSpectrum[i].usSpreadRateIn10Hz); + le16_to_cpu(ss_assign->v3.usSpreadSpectrumPercentage); + ss->type = ss_assign->v3.ucSpreadSpectrumMode; + ss->rate = le16_to_cpu(ss_assign->v3.usSpreadRateIn10Hz); + if (ss_assign->v3.ucSpreadSpectrumMode & + SS_MODE_V3_PERCENTAGE_DIV_BY_1000_MASK) + ss->percentage_divider = 1000; + else + ss->percentage_divider = 100; if ((id == ASIC_INTERNAL_ENGINE_SS) || (id == ASIC_INTERNAL_MEMORY_SS)) ss->rate /= 100; @@ -1538,6 +1571,8 @@ bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev, radeon_atombios_get_igp_ss_overrides(rdev, ss, id); return true; } + ss_assign = (union asic_ss_assignment *) + ((u8 *)ss_assign + sizeof(ATOM_ASIC_SS_ASSIGNMENT_V3)); } break; default: @@ -1672,7 +1707,9 @@ struct radeon_encoder_atom_dig *radeon_atombios_get_lvds_info(struct kfree(edid); } } - record += sizeof(ATOM_FAKE_EDID_PATCH_RECORD); + record += fake_edid_record->ucFakeEDIDLength ? + fake_edid_record->ucFakeEDIDLength + 2 : + sizeof(ATOM_FAKE_EDID_PATCH_RECORD); break; case LCD_PANEL_RESOLUTION_RECORD_TYPE: panel_res_record = (ATOM_PANEL_RESOLUTION_PATCH_RECORD *)record; @@ -1770,7 +1807,8 @@ bool radeon_atom_get_tv_timings(struct radeon_device *rdev, int index, if (misc & ATOM_DOUBLE_CLOCK_MODE) mode->flags |= DRM_MODE_FLAG_DBLSCAN; - mode->clock = le16_to_cpu(tv_info->aModeTimings[index].usPixelClock) * 10; + mode->crtc_clock = mode->clock = + le16_to_cpu(tv_info->aModeTimings[index].usPixelClock) * 10; if (index == 1) { /* PAL timings appear to have wrong values for totals */ @@ -1813,7 +1851,8 @@ bool radeon_atom_get_tv_timings(struct radeon_device *rdev, int index, if (misc & ATOM_DOUBLE_CLOCK_MODE) mode->flags |= DRM_MODE_FLAG_DBLSCAN; - mode->clock = le16_to_cpu(dtd_timings->usPixClk) * 10; + mode->crtc_clock = mode->clock = + le16_to_cpu(dtd_timings->usPixClk) * 10; break; } return true; @@ -2237,6 +2276,11 @@ static void radeon_atombios_add_pplib_thermal_controller(struct radeon_device *r (controller->ucFanParameters & ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with"); rdev->pm.int_thermal_type = THERMAL_TYPE_CI; + } else if (controller->ucType == ATOM_PP_THERMALCONTROLLER_KAVERI) { + DRM_INFO("Internal thermal controller %s fan control\n", + (controller->ucFanParameters & + ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with"); + rdev->pm.int_thermal_type = THERMAL_TYPE_KV; } else if ((controller->ucType == ATOM_PP_THERMALCONTROLLER_EXTERNAL_GPIO) || (controller->ucType == @@ -2782,7 +2826,7 @@ int radeon_atom_get_clock_dividers(struct radeon_device *rdev, ATOM_PLL_CNTL_FLAG_PLL_POST_DIV_EN) ? true : false; dividers->enable_dithen = (args.v3.ucCntlFlag & ATOM_PLL_CNTL_FLAG_FRACTION_DISABLE) ? false : true; - dividers->fb_div = le16_to_cpu(args.v3.ulFbDiv.usFbDiv); + dividers->whole_fb_div = le16_to_cpu(args.v3.ulFbDiv.usFbDiv); dividers->frac_fb_div = le16_to_cpu(args.v3.ulFbDiv.usFbDivFrac); dividers->ref_div = args.v3.ucRefDiv; dividers->vco_mode = (args.v3.ucCntlFlag & @@ -2874,7 +2918,7 @@ int radeon_atom_get_memory_pll_dividers(struct radeon_device *rdev, mpll_param->dll_speed = args.ucDllSpeed; mpll_param->bwcntl = args.ucBWCntl; mpll_param->vco_mode = - (args.ucPllCntlFlag & MPLL_CNTL_FLAG_VCO_MODE_MASK) ? 1 : 0; + (args.ucPllCntlFlag & MPLL_CNTL_FLAG_VCO_MODE_MASK); mpll_param->yclk_sel = (args.ucPllCntlFlag & MPLL_CNTL_FLAG_BYPASS_DQ_PLL) ? 1 : 0; mpll_param->qdr = @@ -3077,6 +3121,121 @@ int radeon_atom_get_leakage_vddc_based_on_leakage_idx(struct radeon_device *rdev return radeon_atom_get_max_vddc(rdev, VOLTAGE_TYPE_VDDC, leakage_idx, voltage); } +int radeon_atom_get_leakage_id_from_vbios(struct radeon_device *rdev, + u16 *leakage_id) +{ + union set_voltage args; + int index = GetIndexIntoMasterTable(COMMAND, SetVoltage); + u8 frev, crev; + + if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev)) + return -EINVAL; + + switch (crev) { + case 3: + case 4: + args.v3.ucVoltageType = 0; + args.v3.ucVoltageMode = ATOM_GET_LEAKAGE_ID; + args.v3.usVoltageLevel = 0; + + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + + *leakage_id = le16_to_cpu(args.v3.usVoltageLevel); + break; + default: + DRM_ERROR("Unknown table version %d, %d\n", frev, crev); + return -EINVAL; + } + + return 0; +} + +int radeon_atom_get_leakage_vddc_based_on_leakage_params(struct radeon_device *rdev, + u16 *vddc, u16 *vddci, + u16 virtual_voltage_id, + u16 vbios_voltage_id) +{ + int index = GetIndexIntoMasterTable(DATA, ASIC_ProfilingInfo); + u8 frev, crev; + u16 data_offset, size; + int i, j; + ATOM_ASIC_PROFILING_INFO_V2_1 *profile; + u16 *leakage_bin, *vddc_id_buf, *vddc_buf, *vddci_id_buf, *vddci_buf; + + *vddc = 0; + *vddci = 0; + + if (!atom_parse_data_header(rdev->mode_info.atom_context, index, &size, + &frev, &crev, &data_offset)) + return -EINVAL; + + profile = (ATOM_ASIC_PROFILING_INFO_V2_1 *) + (rdev->mode_info.atom_context->bios + data_offset); + + switch (frev) { + case 1: + return -EINVAL; + case 2: + switch (crev) { + case 1: + if (size < sizeof(ATOM_ASIC_PROFILING_INFO_V2_1)) + return -EINVAL; + leakage_bin = (u16 *) + (rdev->mode_info.atom_context->bios + data_offset + + le16_to_cpu(profile->usLeakageBinArrayOffset)); + vddc_id_buf = (u16 *) + (rdev->mode_info.atom_context->bios + data_offset + + le16_to_cpu(profile->usElbVDDC_IdArrayOffset)); + vddc_buf = (u16 *) + (rdev->mode_info.atom_context->bios + data_offset + + le16_to_cpu(profile->usElbVDDC_LevelArrayOffset)); + vddci_id_buf = (u16 *) + (rdev->mode_info.atom_context->bios + data_offset + + le16_to_cpu(profile->usElbVDDCI_IdArrayOffset)); + vddci_buf = (u16 *) + (rdev->mode_info.atom_context->bios + data_offset + + le16_to_cpu(profile->usElbVDDCI_LevelArrayOffset)); + + if (profile->ucElbVDDC_Num > 0) { + for (i = 0; i < profile->ucElbVDDC_Num; i++) { + if (vddc_id_buf[i] == virtual_voltage_id) { + for (j = 0; j < profile->ucLeakageBinNum; j++) { + if (vbios_voltage_id <= leakage_bin[j]) { + *vddc = vddc_buf[j * profile->ucElbVDDC_Num + i]; + break; + } + } + break; + } + } + } + if (profile->ucElbVDDCI_Num > 0) { + for (i = 0; i < profile->ucElbVDDCI_Num; i++) { + if (vddci_id_buf[i] == virtual_voltage_id) { + for (j = 0; j < profile->ucLeakageBinNum; j++) { + if (vbios_voltage_id <= leakage_bin[j]) { + *vddci = vddci_buf[j * profile->ucElbVDDCI_Num + i]; + break; + } + } + break; + } + } + } + break; + default: + DRM_ERROR("Unknown table version %d, %d\n", frev, crev); + return -EINVAL; + } + break; + default: + DRM_ERROR("Unknown table version %d, %d\n", frev, crev); + return -EINVAL; + } + + return 0; +} + int radeon_atom_get_voltage_gpio_settings(struct radeon_device *rdev, u16 voltage_level, u8 voltage_type, u32 *gpio_value, u32 *gpio_mask) @@ -3279,10 +3438,11 @@ int radeon_atom_get_max_voltage(struct radeon_device *rdev, ATOM_VOLTAGE_FORMULA_V2 *formula = &voltage_object->v2.asFormula; if (formula->ucNumOfVoltageEntries) { + VOLTAGE_LUT_ENTRY *lut = (VOLTAGE_LUT_ENTRY *) + ((u8 *)&formula->asVIDAdjustEntries[0] + + (sizeof(VOLTAGE_LUT_ENTRY) * (formula->ucNumOfVoltageEntries - 1))); *max_voltage = - le16_to_cpu(formula->asVIDAdjustEntries[ - formula->ucNumOfVoltageEntries - 1 - ].usVoltageValue); + le16_to_cpu(lut->usVoltageValue); return 0; } } @@ -3442,11 +3602,13 @@ int radeon_atom_get_voltage_table(struct radeon_device *rdev, if (voltage_object) { ATOM_VOLTAGE_FORMULA_V2 *formula = &voltage_object->v2.asFormula; + VOLTAGE_LUT_ENTRY *lut; if (formula->ucNumOfVoltageEntries > MAX_VOLTAGE_ENTRIES) return -EINVAL; + lut = &formula->asVIDAdjustEntries[0]; for (i = 0; i < formula->ucNumOfVoltageEntries; i++) { voltage_table->entries[i].value = - le16_to_cpu(formula->asVIDAdjustEntries[i].usVoltageValue); + le16_to_cpu(lut->usVoltageValue); ret = radeon_atom_get_voltage_gpio_settings(rdev, voltage_table->entries[i].value, voltage_type, @@ -3454,6 +3616,8 @@ int radeon_atom_get_voltage_table(struct radeon_device *rdev, &voltage_table->mask_low); if (ret) return ret; + lut = (VOLTAGE_LUT_ENTRY *) + ((u8 *)lut + sizeof(VOLTAGE_LUT_ENTRY)); } voltage_table->count = formula->ucNumOfVoltageEntries; return 0; @@ -3473,13 +3637,17 @@ int radeon_atom_get_voltage_table(struct radeon_device *rdev, if (voltage_object) { ATOM_GPIO_VOLTAGE_OBJECT_V3 *gpio = &voltage_object->v3.asGpioVoltageObj; + VOLTAGE_LUT_ENTRY_V2 *lut; if (gpio->ucGpioEntryNum > MAX_VOLTAGE_ENTRIES) return -EINVAL; + lut = &gpio->asVolGpioLut[0]; for (i = 0; i < gpio->ucGpioEntryNum; i++) { voltage_table->entries[i].value = - le16_to_cpu(gpio->asVolGpioLut[i].usVoltageValue); + le16_to_cpu(lut->usVoltageValue); voltage_table->entries[i].smio_low = - le32_to_cpu(gpio->asVolGpioLut[i].ulVoltageId); + le32_to_cpu(lut->ulVoltageId); + lut = (VOLTAGE_LUT_ENTRY_V2 *) + ((u8 *)lut + sizeof(VOLTAGE_LUT_ENTRY_V2)); } voltage_table->mask_low = le32_to_cpu(gpio->ulGpioMaskVal); voltage_table->count = gpio->ucGpioEntryNum; @@ -3513,7 +3681,6 @@ int radeon_atom_get_memory_info(struct radeon_device *rdev, u8 frev, crev, i; u16 data_offset, size; union vram_info *vram_info; - u8 *p; memset(mem_info, 0, sizeof(struct atom_memory_info)); @@ -3529,13 +3696,12 @@ int radeon_atom_get_memory_info(struct radeon_device *rdev, if (module_index < vram_info->v1_3.ucNumOfVRAMModule) { ATOM_VRAM_MODULE_V3 *vram_module = (ATOM_VRAM_MODULE_V3 *)vram_info->v1_3.aVramInfo; - p = (u8 *)vram_info->v1_3.aVramInfo; for (i = 0; i < module_index; i++) { - vram_module = (ATOM_VRAM_MODULE_V3 *)p; if (le16_to_cpu(vram_module->usSize) == 0) return -EINVAL; - p += le16_to_cpu(vram_module->usSize); + vram_module = (ATOM_VRAM_MODULE_V3 *) + ((u8 *)vram_module + le16_to_cpu(vram_module->usSize)); } mem_info->mem_vendor = vram_module->asMemory.ucMemoryVenderID & 0xf; mem_info->mem_type = vram_module->asMemory.ucMemoryType & 0xf0; @@ -3547,13 +3713,12 @@ int radeon_atom_get_memory_info(struct radeon_device *rdev, if (module_index < vram_info->v1_4.ucNumOfVRAMModule) { ATOM_VRAM_MODULE_V4 *vram_module = (ATOM_VRAM_MODULE_V4 *)vram_info->v1_4.aVramInfo; - p = (u8 *)vram_info->v1_4.aVramInfo; for (i = 0; i < module_index; i++) { - vram_module = (ATOM_VRAM_MODULE_V4 *)p; if (le16_to_cpu(vram_module->usModuleSize) == 0) return -EINVAL; - p += le16_to_cpu(vram_module->usModuleSize); + vram_module = (ATOM_VRAM_MODULE_V4 *) + ((u8 *)vram_module + le16_to_cpu(vram_module->usModuleSize)); } mem_info->mem_vendor = vram_module->ucMemoryVenderID & 0xf; mem_info->mem_type = vram_module->ucMemoryType & 0xf0; @@ -3572,13 +3737,12 @@ int radeon_atom_get_memory_info(struct radeon_device *rdev, if (module_index < vram_info->v2_1.ucNumOfVRAMModule) { ATOM_VRAM_MODULE_V7 *vram_module = (ATOM_VRAM_MODULE_V7 *)vram_info->v2_1.aVramInfo; - p = (u8 *)vram_info->v2_1.aVramInfo; for (i = 0; i < module_index; i++) { - vram_module = (ATOM_VRAM_MODULE_V7 *)p; if (le16_to_cpu(vram_module->usModuleSize) == 0) return -EINVAL; - p += le16_to_cpu(vram_module->usModuleSize); + vram_module = (ATOM_VRAM_MODULE_V7 *) + ((u8 *)vram_module + le16_to_cpu(vram_module->usModuleSize)); } mem_info->mem_vendor = vram_module->ucMemoryVenderID & 0xf; mem_info->mem_type = vram_module->ucMemoryType & 0xf0; @@ -3609,7 +3773,6 @@ int radeon_atom_get_mclk_range_table(struct radeon_device *rdev, union vram_info *vram_info; u32 mem_timing_size = gddr5 ? sizeof(ATOM_MEMORY_TIMING_FORMAT_V2) : sizeof(ATOM_MEMORY_TIMING_FORMAT); - u8 *p; memset(mclk_range_table, 0, sizeof(struct atom_memory_clock_range_table)); @@ -3629,22 +3792,21 @@ int radeon_atom_get_mclk_range_table(struct radeon_device *rdev, ATOM_VRAM_MODULE_V4 *vram_module = (ATOM_VRAM_MODULE_V4 *)vram_info->v1_4.aVramInfo; ATOM_MEMORY_TIMING_FORMAT *format; - p = (u8 *)vram_info->v1_4.aVramInfo; for (i = 0; i < module_index; i++) { - vram_module = (ATOM_VRAM_MODULE_V4 *)p; if (le16_to_cpu(vram_module->usModuleSize) == 0) return -EINVAL; - p += le16_to_cpu(vram_module->usModuleSize); + vram_module = (ATOM_VRAM_MODULE_V4 *) + ((u8 *)vram_module + le16_to_cpu(vram_module->usModuleSize)); } mclk_range_table->num_entries = (u8) - ((vram_module->usModuleSize - offsetof(ATOM_VRAM_MODULE_V4, asMemTiming)) / + ((le16_to_cpu(vram_module->usModuleSize) - offsetof(ATOM_VRAM_MODULE_V4, asMemTiming)) / mem_timing_size); - p = (u8 *)vram_module->asMemTiming; + format = &vram_module->asMemTiming[0]; for (i = 0; i < mclk_range_table->num_entries; i++) { - format = (ATOM_MEMORY_TIMING_FORMAT *)p; mclk_range_table->mclk[i] = le32_to_cpu(format->ulClkRange); - p += mem_timing_size; + format = (ATOM_MEMORY_TIMING_FORMAT *) + ((u8 *)format + mem_timing_size); } } else return -EINVAL; @@ -3705,29 +3867,35 @@ int radeon_atom_init_mc_reg_table(struct radeon_device *rdev, (ATOM_MEMORY_SETTING_DATA_BLOCK *) ((u8 *)reg_block + (2 * sizeof(u16)) + le16_to_cpu(reg_block->usRegIndexTblSize)); + ATOM_INIT_REG_INDEX_FORMAT *format = ®_block->asRegIndexBuf[0]; num_entries = (u8)((le16_to_cpu(reg_block->usRegIndexTblSize)) / sizeof(ATOM_INIT_REG_INDEX_FORMAT)) - 1; if (num_entries > VBIOS_MC_REGISTER_ARRAY_SIZE) return -EINVAL; - while (!(reg_block->asRegIndexBuf[i].ucPreRegDataLength & ACCESS_PLACEHOLDER) && - (i < num_entries)) { + while (i < num_entries) { + if (format->ucPreRegDataLength & ACCESS_PLACEHOLDER) + break; reg_table->mc_reg_address[i].s1 = - (u16)(le16_to_cpu(reg_block->asRegIndexBuf[i].usRegIndex)); + (u16)(le16_to_cpu(format->usRegIndex)); reg_table->mc_reg_address[i].pre_reg_data = - (u8)(reg_block->asRegIndexBuf[i].ucPreRegDataLength); + (u8)(format->ucPreRegDataLength); i++; + format = (ATOM_INIT_REG_INDEX_FORMAT *) + ((u8 *)format + sizeof(ATOM_INIT_REG_INDEX_FORMAT)); } reg_table->last = i; - while ((*(u32 *)reg_data != END_OF_REG_DATA_BLOCK) && + while ((le32_to_cpu(*(u32 *)reg_data) != END_OF_REG_DATA_BLOCK) && (num_ranges < VBIOS_MAX_AC_TIMING_ENTRIES)) { - t_mem_id = (u8)((*(u32 *)reg_data & MEM_ID_MASK) >> MEM_ID_SHIFT); + t_mem_id = (u8)((le32_to_cpu(*(u32 *)reg_data) & MEM_ID_MASK) + >> MEM_ID_SHIFT); if (module_index == t_mem_id) { reg_table->mc_reg_table_entry[num_ranges].mclk_max = - (u32)((*(u32 *)reg_data & CLOCK_RANGE_MASK) >> CLOCK_RANGE_SHIFT); + (u32)((le32_to_cpu(*(u32 *)reg_data) & CLOCK_RANGE_MASK) + >> CLOCK_RANGE_SHIFT); for (i = 0, j = 1; i < reg_table->last; i++) { if ((reg_table->mc_reg_address[i].pre_reg_data & LOW_NIBBLE_MASK) == DATA_FROM_TABLE) { reg_table->mc_reg_table_entry[num_ranges].mc_data[i] = - (u32)*((u32 *)reg_data + j); + (u32)le32_to_cpu(*((u32 *)reg_data + j)); j++; } else if ((reg_table->mc_reg_address[i].pre_reg_data & LOW_NIBBLE_MASK) == DATA_EQU_PREV) { reg_table->mc_reg_table_entry[num_ranges].mc_data[i] = @@ -3739,7 +3907,7 @@ int radeon_atom_init_mc_reg_table(struct radeon_device *rdev, reg_data = (ATOM_MEMORY_SETTING_DATA_BLOCK *) ((u8 *)reg_data + le16_to_cpu(reg_block->usRegDataBlkSize)); } - if (*(u32 *)reg_data != END_OF_REG_DATA_BLOCK) + if (le32_to_cpu(*(u32 *)reg_data) != END_OF_REG_DATA_BLOCK) return -EINVAL; reg_table->num_entries = num_ranges; } else @@ -3778,6 +3946,10 @@ void radeon_atom_initialize_bios_scratch_regs(struct drm_device *dev) /* tell the bios not to handle mode switching */ bios_6_scratch |= ATOM_S6_ACC_BLOCK_DISPLAY_SWITCH; + /* clear the vbios dpms state */ + if (ASIC_IS_DCE4(rdev)) + bios_2_scratch &= ~ATOM_S2_DEVICE_DPMS_STATE; + if (rdev->family >= CHIP_R600) { WREG32(R600_BIOS_2_SCRATCH, bios_2_scratch); WREG32(R600_BIOS_6_SCRATCH, bios_6_scratch); diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c index d96070bf838..a9fb0d016d3 100644 --- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c +++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c @@ -8,8 +8,7 @@ */ #include <linux/vga_switcheroo.h> #include <linux/slab.h> -#include <acpi/acpi.h> -#include <acpi/acpi_bus.h> +#include <linux/acpi.h> #include <linux/pci.h> #include "radeon_acpi.h" @@ -34,6 +33,7 @@ static struct radeon_atpx_priv { bool atpx_detected; /* handle for device - and atpx */ acpi_handle dhandle; + acpi_handle other_handle; struct radeon_atpx atpx; } radeon_atpx_priv; @@ -59,6 +59,10 @@ struct atpx_mux { u16 mux; } __packed; +bool radeon_has_atpx(void) { + return radeon_atpx_priv.atpx_detected; +} + /** * radeon_atpx_call - call an ATPX method * @@ -215,7 +219,8 @@ static int radeon_atpx_verify_interface(struct radeon_atpx *atpx) memcpy(&output, info->buffer.pointer, size); /* TODO: check version? */ - printk("ATPX version %u\n", output.version); + printk("ATPX version %u, functions 0x%08x\n", + output.version, output.function_bits); radeon_atpx_parse_functions(&atpx->functions, output.function_bits); @@ -443,14 +448,15 @@ static bool radeon_atpx_pci_probe_handle(struct pci_dev *pdev) acpi_handle dhandle, atpx_handle; acpi_status status; - dhandle = DEVICE_ACPI_HANDLE(&pdev->dev); + dhandle = ACPI_HANDLE(&pdev->dev); if (!dhandle) return false; status = acpi_get_handle(dhandle, "ATPX", &atpx_handle); - if (ACPI_FAILURE(status)) + if (ACPI_FAILURE(status)) { + radeon_atpx_priv.other_handle = dhandle; return false; - + } radeon_atpx_priv.dhandle = dhandle; radeon_atpx_priv.atpx.handle = atpx_handle; return true; @@ -489,7 +495,7 @@ static int radeon_atpx_init(void) */ static int radeon_atpx_get_client_id(struct pci_dev *pdev) { - if (radeon_atpx_priv.dhandle == DEVICE_ACPI_HANDLE(&pdev->dev)) + if (radeon_atpx_priv.dhandle == ACPI_HANDLE(&pdev->dev)) return VGA_SWITCHEROO_IGD; else return VGA_SWITCHEROO_DIS; @@ -522,11 +528,28 @@ static bool radeon_atpx_detect(void) has_atpx |= (radeon_atpx_pci_probe_handle(pdev) == true); } + /* some newer PX laptops mark the dGPU as a non-VGA display device */ + while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_OTHER << 8, pdev)) != NULL) { + vga_count++; + + has_atpx |= (radeon_atpx_pci_probe_handle(pdev) == true); + } + if (has_atpx && vga_count == 2) { acpi_get_name(radeon_atpx_priv.atpx.handle, ACPI_FULL_PATHNAME, &buffer); printk(KERN_INFO "VGA switcheroo: detected switching method %s handle\n", acpi_method_name); radeon_atpx_priv.atpx_detected = true; + /* + * On some systems hotplug events are generated for the device + * being switched off when ATPX is executed. They cause ACPI + * hotplug to trigger and attempt to remove the device from + * the system, which causes it to break down. Prevent that from + * happening by setting the no_hotplug flag for the involved + * ACPI device objects. + */ + acpi_bus_no_hotplug(radeon_atpx_priv.dhandle); + acpi_bus_no_hotplug(radeon_atpx_priv.other_handle); return true; } return false; diff --git a/drivers/gpu/drm/radeon/radeon_bios.c b/drivers/gpu/drm/radeon/radeon_bios.c index 061b227dae0..6a03624fada 100644 --- a/drivers/gpu/drm/radeon/radeon_bios.c +++ b/drivers/gpu/drm/radeon/radeon_bios.c @@ -185,7 +185,7 @@ static bool radeon_atrm_get_bios(struct radeon_device *rdev) return false; while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, pdev)) != NULL) { - dhandle = DEVICE_ACPI_HANDLE(&pdev->dev); + dhandle = ACPI_HANDLE(&pdev->dev); if (!dhandle) continue; @@ -196,6 +196,20 @@ static bool radeon_atrm_get_bios(struct radeon_device *rdev) } } + if (!found) { + while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_OTHER << 8, pdev)) != NULL) { + dhandle = ACPI_HANDLE(&pdev->dev); + if (!dhandle) + continue; + + status = acpi_get_handle(dhandle, "ATRM", &atrm_handle); + if (!ACPI_FAILURE(status)) { + found = true; + break; + } + } + } + if (!found) return false; @@ -499,7 +513,7 @@ static bool legacy_read_disabled_bios(struct radeon_device *rdev) crtc_ext_cntl = RREG32(RADEON_CRTC_EXT_CNTL); fp2_gen_cntl = 0; - if (rdev->ddev->pci_device == PCI_DEVICE_ID_ATI_RADEON_QY) { + if (rdev->ddev->pdev->device == PCI_DEVICE_ID_ATI_RADEON_QY) { fp2_gen_cntl = RREG32(RADEON_FP2_GEN_CNTL); } @@ -536,7 +550,7 @@ static bool legacy_read_disabled_bios(struct radeon_device *rdev) (RADEON_CRTC_SYNC_TRISTAT | RADEON_CRTC_DISPLAY_DIS))); - if (rdev->ddev->pci_device == PCI_DEVICE_ID_ATI_RADEON_QY) { + if (rdev->ddev->pdev->device == PCI_DEVICE_ID_ATI_RADEON_QY) { WREG32(RADEON_FP2_GEN_CNTL, (fp2_gen_cntl & ~RADEON_FP2_ON)); } @@ -554,7 +568,7 @@ static bool legacy_read_disabled_bios(struct radeon_device *rdev) WREG32(RADEON_CRTC2_GEN_CNTL, crtc2_gen_cntl); } WREG32(RADEON_CRTC_EXT_CNTL, crtc_ext_cntl); - if (rdev->ddev->pci_device == PCI_DEVICE_ID_ATI_RADEON_QY) { + if (rdev->ddev->pdev->device == PCI_DEVICE_ID_ATI_RADEON_QY) { WREG32(RADEON_FP2_GEN_CNTL, fp2_gen_cntl); } return r; @@ -612,7 +626,7 @@ static bool radeon_acpi_vfct_bios(struct radeon_device *rdev) vhdr->DeviceID != rdev->pdev->device) { DRM_INFO("ACPI VFCT table is not for this card\n"); goto out_unmap; - }; + } if (vfct->VBIOSImageOffset + sizeof(VFCT_IMAGE_HEADER) + vhdr->ImageLength > tbl_size) { DRM_ERROR("ACPI VFCT image truncated\n"); diff --git a/drivers/gpu/drm/radeon/radeon_blit_common.h b/drivers/gpu/drm/radeon/radeon_blit_common.h deleted file mode 100644 index 4ecbe72c9d2..00000000000 --- a/drivers/gpu/drm/radeon/radeon_blit_common.h +++ /dev/null @@ -1,44 +0,0 @@ -/* - * Copyright 2009 Advanced Micro Devices, Inc. - * Copyright 2009 Red Hat Inc. - * Copyright 2012 Alcatel-Lucent, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - * - */ - -#ifndef __RADEON_BLIT_COMMON_H__ - -#define DI_PT_RECTLIST 0x11 -#define DI_INDEX_SIZE_16_BIT 0x0 -#define DI_SRC_SEL_AUTO_INDEX 0x2 - -#define FMT_8 0x1 -#define FMT_5_6_5 0x8 -#define FMT_8_8_8_8 0x1a -#define COLOR_8 0x1 -#define COLOR_5_6_5 0x8 -#define COLOR_8_8_8_8 0x1a - -#define RECT_UNIT_H 32 -#define RECT_UNIT_W (RADEON_GPU_PAGE_SIZE / 4 / RECT_UNIT_H) - -#define __RADEON_BLIT_COMMON_H__ -#endif diff --git a/drivers/gpu/drm/radeon/radeon_combios.c b/drivers/gpu/drm/radeon/radeon_combios.c index 78edadc9e86..6651177110f 100644 --- a/drivers/gpu/drm/radeon/radeon_combios.c +++ b/drivers/gpu/drm/radeon/radeon_combios.c @@ -37,22 +37,6 @@ #include <asm/pci-bridge.h> #endif /* CONFIG_PPC_PMAC */ -/* from radeon_encoder.c */ -extern uint32_t -radeon_get_encoder_enum(struct drm_device *dev, uint32_t supported_device, - uint8_t dac); -extern void radeon_link_encoder_connector(struct drm_device *dev); - -/* from radeon_connector.c */ -extern void -radeon_add_legacy_connector(struct drm_device *dev, - uint32_t connector_id, - uint32_t supported_device, - int connector_type, - struct radeon_i2c_bus_rec *i2c_bus, - uint16_t connector_object_id, - struct radeon_hpd *hpd); - /* from radeon_legacy_encoder.c */ extern void radeon_add_legacy_encoder(struct drm_device *dev, uint32_t encoder_enum, @@ -147,7 +131,7 @@ static uint16_t combios_get_table_offset(struct drm_device *dev, enum radeon_combios_table_offset table) { struct radeon_device *rdev = dev->dev_private; - int rev; + int rev, size; uint16_t offset = 0, check_offset; if (!rdev->bios) @@ -156,174 +140,106 @@ static uint16_t combios_get_table_offset(struct drm_device *dev, switch (table) { /* absolute offset tables */ case COMBIOS_ASIC_INIT_1_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0xc); - if (check_offset) - offset = check_offset; + check_offset = 0xc; break; case COMBIOS_BIOS_SUPPORT_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x14); - if (check_offset) - offset = check_offset; + check_offset = 0x14; break; case COMBIOS_DAC_PROGRAMMING_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x2a); - if (check_offset) - offset = check_offset; + check_offset = 0x2a; break; case COMBIOS_MAX_COLOR_DEPTH_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x2c); - if (check_offset) - offset = check_offset; + check_offset = 0x2c; break; case COMBIOS_CRTC_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x2e); - if (check_offset) - offset = check_offset; + check_offset = 0x2e; break; case COMBIOS_PLL_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x30); - if (check_offset) - offset = check_offset; + check_offset = 0x30; break; case COMBIOS_TV_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x32); - if (check_offset) - offset = check_offset; + check_offset = 0x32; break; case COMBIOS_DFP_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x34); - if (check_offset) - offset = check_offset; + check_offset = 0x34; break; case COMBIOS_HW_CONFIG_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x36); - if (check_offset) - offset = check_offset; + check_offset = 0x36; break; case COMBIOS_MULTIMEDIA_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x38); - if (check_offset) - offset = check_offset; + check_offset = 0x38; break; case COMBIOS_TV_STD_PATCH_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x3e); - if (check_offset) - offset = check_offset; + check_offset = 0x3e; break; case COMBIOS_LCD_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x40); - if (check_offset) - offset = check_offset; + check_offset = 0x40; break; case COMBIOS_MOBILE_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x42); - if (check_offset) - offset = check_offset; + check_offset = 0x42; break; case COMBIOS_PLL_INIT_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x46); - if (check_offset) - offset = check_offset; + check_offset = 0x46; break; case COMBIOS_MEM_CONFIG_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x48); - if (check_offset) - offset = check_offset; + check_offset = 0x48; break; case COMBIOS_SAVE_MASK_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x4a); - if (check_offset) - offset = check_offset; + check_offset = 0x4a; break; case COMBIOS_HARDCODED_EDID_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x4c); - if (check_offset) - offset = check_offset; + check_offset = 0x4c; break; case COMBIOS_ASIC_INIT_2_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x4e); - if (check_offset) - offset = check_offset; + check_offset = 0x4e; break; case COMBIOS_CONNECTOR_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x50); - if (check_offset) - offset = check_offset; + check_offset = 0x50; break; case COMBIOS_DYN_CLK_1_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x52); - if (check_offset) - offset = check_offset; + check_offset = 0x52; break; case COMBIOS_RESERVED_MEM_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x54); - if (check_offset) - offset = check_offset; + check_offset = 0x54; break; case COMBIOS_EXT_TMDS_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x58); - if (check_offset) - offset = check_offset; + check_offset = 0x58; break; case COMBIOS_MEM_CLK_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x5a); - if (check_offset) - offset = check_offset; + check_offset = 0x5a; break; case COMBIOS_EXT_DAC_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x5c); - if (check_offset) - offset = check_offset; + check_offset = 0x5c; break; case COMBIOS_MISC_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x5e); - if (check_offset) - offset = check_offset; + check_offset = 0x5e; break; case COMBIOS_CRT_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x60); - if (check_offset) - offset = check_offset; + check_offset = 0x60; break; case COMBIOS_INTEGRATED_SYSTEM_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x62); - if (check_offset) - offset = check_offset; + check_offset = 0x62; break; case COMBIOS_COMPONENT_VIDEO_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x64); - if (check_offset) - offset = check_offset; + check_offset = 0x64; break; case COMBIOS_FAN_SPEED_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x66); - if (check_offset) - offset = check_offset; + check_offset = 0x66; break; case COMBIOS_OVERDRIVE_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x68); - if (check_offset) - offset = check_offset; + check_offset = 0x68; break; case COMBIOS_OEM_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x6a); - if (check_offset) - offset = check_offset; + check_offset = 0x6a; break; case COMBIOS_DYN_CLK_2_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x6c); - if (check_offset) - offset = check_offset; + check_offset = 0x6c; break; case COMBIOS_POWER_CONNECTOR_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x6e); - if (check_offset) - offset = check_offset; + check_offset = 0x6e; break; case COMBIOS_I2C_INFO_TABLE: - check_offset = RBIOS16(rdev->bios_header_start + 0x70); - if (check_offset) - offset = check_offset; + check_offset = 0x70; break; /* relative offset tables */ case COMBIOS_ASIC_INIT_3_TABLE: /* offset from misc info */ @@ -439,11 +355,16 @@ static uint16_t combios_get_table_offset(struct drm_device *dev, } break; default: + check_offset = 0; break; } - return offset; + size = RBIOS8(rdev->bios_header_start + 0x6); + /* check absolute offset tables */ + if (table < COMBIOS_ASIC_INIT_3_TABLE && check_offset && check_offset < size) + offset = RBIOS16(rdev->bios_header_start + check_offset); + return offset; } bool radeon_combios_check_hardcoded_edid(struct radeon_device *rdev) @@ -965,16 +886,22 @@ struct radeon_encoder_primary_dac *radeon_combios_get_primary_dac_info(struct dac = RBIOS8(dac_info + 0x3) & 0xf; p_dac->ps2_pdac_adj = (bg << 8) | (dac); } - /* if the values are all zeros, use the table */ - if (p_dac->ps2_pdac_adj) + /* if the values are zeros, use the table */ + if ((dac == 0) || (bg == 0)) + found = 0; + else found = 1; } /* quirks */ + /* Radeon 7000 (RV100) */ + if (((dev->pdev->device == 0x5159) && + (dev->pdev->subsystem_vendor == 0x174B) && + (dev->pdev->subsystem_device == 0x7c28)) || /* Radeon 9100 (R200) */ - if ((dev->pdev->device == 0x514D) && + ((dev->pdev->device == 0x514D) && (dev->pdev->subsystem_vendor == 0x174B) && - (dev->pdev->subsystem_device == 0x7149)) { + (dev->pdev->subsystem_device == 0x7149))) { /* vbios value is bad, use the default */ found = 0; } diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c index 2399f25ec03..44831197e82 100644 --- a/drivers/gpu/drm/radeon/radeon_connectors.c +++ b/drivers/gpu/drm/radeon/radeon_connectors.c @@ -31,14 +31,7 @@ #include "radeon.h" #include "atom.h" -extern void -radeon_combios_connected_scratch_regs(struct drm_connector *connector, - struct drm_encoder *encoder, - bool connected); -extern void -radeon_atombios_connected_scratch_regs(struct drm_connector *connector, - struct drm_encoder *encoder, - bool connected); +#include <linux/pm_runtime.h> void radeon_connector_hotplug(struct drm_connector *connector) { @@ -55,6 +48,7 @@ void radeon_connector_hotplug(struct drm_connector *connector) radeon_hpd_set_polarity(rdev, radeon_connector->hpd.hpd); /* if the connector is already off, don't turn it back on */ + /* FIXME: This access isn't protected by any locks. */ if (connector->dpms != DRM_MODE_DPMS_ON) return; @@ -96,7 +90,7 @@ static void radeon_property_change_mode(struct drm_encoder *encoder) if (crtc && crtc->enabled) { drm_crtc_helper_set_mode(crtc, &crtc->mode, - crtc->x, crtc->y, crtc->fb); + crtc->x, crtc->y, crtc->primary->fb); } } @@ -107,6 +101,7 @@ int radeon_get_monitor_bpc(struct drm_connector *connector) struct radeon_connector *radeon_connector = to_radeon_connector(connector); struct radeon_connector_atom_dig *dig_connector; int bpc = 8; + int mode_clock, max_tmds_clock; switch (connector->connector_type) { case DRM_MODE_CONNECTOR_DVII: @@ -152,6 +147,64 @@ int radeon_get_monitor_bpc(struct drm_connector *connector) } break; } + + if (drm_detect_hdmi_monitor(radeon_connector->edid)) { + /* hdmi deep color only implemented on DCE4+ */ + if ((bpc > 8) && !ASIC_IS_DCE4(rdev)) { + DRM_DEBUG("%s: HDMI deep color %d bpc unsupported. Using 8 bpc.\n", + connector->name, bpc); + bpc = 8; + } + + /* + * Pre DCE-8 hw can't handle > 12 bpc, and more than 12 bpc doesn't make + * much sense without support for > 12 bpc framebuffers. RGB 4:4:4 at + * 12 bpc is always supported on hdmi deep color sinks, as this is + * required by the HDMI-1.3 spec. Clamp to a safe 12 bpc maximum. + */ + if (bpc > 12) { + DRM_DEBUG("%s: HDMI deep color %d bpc unsupported. Using 12 bpc.\n", + connector->name, bpc); + bpc = 12; + } + + /* Any defined maximum tmds clock limit we must not exceed? */ + if (connector->max_tmds_clock > 0) { + /* mode_clock is clock in kHz for mode to be modeset on this connector */ + mode_clock = radeon_connector->pixelclock_for_modeset; + + /* Maximum allowable input clock in kHz */ + max_tmds_clock = connector->max_tmds_clock * 1000; + + DRM_DEBUG("%s: hdmi mode dotclock %d kHz, max tmds input clock %d kHz.\n", + connector->name, mode_clock, max_tmds_clock); + + /* Check if bpc is within clock limit. Try to degrade gracefully otherwise */ + if ((bpc == 12) && (mode_clock * 3/2 > max_tmds_clock)) { + if ((connector->display_info.edid_hdmi_dc_modes & DRM_EDID_HDMI_DC_30) && + (mode_clock * 5/4 <= max_tmds_clock)) + bpc = 10; + else + bpc = 8; + + DRM_DEBUG("%s: HDMI deep color 12 bpc exceeds max tmds clock. Using %d bpc.\n", + connector->name, bpc); + } + + if ((bpc == 10) && (mode_clock * 5/4 > max_tmds_clock)) { + bpc = 8; + DRM_DEBUG("%s: HDMI deep color 10 bpc exceeds max tmds clock. Using %d bpc.\n", + connector->name, bpc); + } + } + } + + if ((radeon_deep_color == 0) && (bpc > 8)) + bpc = 8; + + DRM_DEBUG("%s: Display bpc=%d, returned bpc=%d\n", + connector->name, connector->display_info.bpc, bpc); + return bpc; } @@ -267,13 +320,17 @@ radeon_connector_analog_encoder_conflict_solve(struct drm_connector *connector, continue; if (priority == true) { - DRM_DEBUG_KMS("1: conflicting encoders switching off %s\n", drm_get_connector_name(conflict)); - DRM_DEBUG_KMS("in favor of %s\n", drm_get_connector_name(connector)); + DRM_DEBUG_KMS("1: conflicting encoders switching off %s\n", + conflict->name); + DRM_DEBUG_KMS("in favor of %s\n", + connector->name); conflict->status = connector_status_disconnected; radeon_connector_update_scratch_regs(conflict, connector_status_disconnected); } else { - DRM_DEBUG_KMS("2: conflicting encoders switching off %s\n", drm_get_connector_name(connector)); - DRM_DEBUG_KMS("in favor of %s\n", drm_get_connector_name(conflict)); + DRM_DEBUG_KMS("2: conflicting encoders switching off %s\n", + connector->name); + DRM_DEBUG_KMS("in favor of %s\n", + conflict->name); current_status = connector_status_disconnected; } break; @@ -396,6 +453,36 @@ static int radeon_connector_set_property(struct drm_connector *connector, struct } } + if (property == rdev->mode_info.audio_property) { + struct radeon_connector *radeon_connector = to_radeon_connector(connector); + /* need to find digital encoder on connector */ + encoder = radeon_find_encoder(connector, DRM_MODE_ENCODER_TMDS); + if (!encoder) + return 0; + + radeon_encoder = to_radeon_encoder(encoder); + + if (radeon_connector->audio != val) { + radeon_connector->audio = val; + radeon_property_change_mode(&radeon_encoder->base); + } + } + + if (property == rdev->mode_info.dither_property) { + struct radeon_connector *radeon_connector = to_radeon_connector(connector); + /* need to find digital encoder on connector */ + encoder = radeon_find_encoder(connector, DRM_MODE_ENCODER_TMDS); + if (!encoder) + return 0; + + radeon_encoder = to_radeon_encoder(encoder); + + if (radeon_connector->dither != val) { + radeon_connector->dither = val; + radeon_property_change_mode(&radeon_encoder->base); + } + } + if (property == rdev->mode_info.underscan_property) { /* need to find digital encoder on connector */ encoder = radeon_find_encoder(connector, DRM_MODE_ENCODER_TMDS); @@ -611,6 +698,11 @@ radeon_lvds_detect(struct drm_connector *connector, bool force) struct radeon_connector *radeon_connector = to_radeon_connector(connector); struct drm_encoder *encoder = radeon_best_single_encoder(connector); enum drm_connector_status ret = connector_status_disconnected; + int r; + + r = pm_runtime_get_sync(connector->dev->dev); + if (r < 0) + return connector_status_disconnected; if (encoder) { struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); @@ -636,6 +728,8 @@ radeon_lvds_detect(struct drm_connector *connector, bool force) /* check acpi lid status ??? */ radeon_connector_update_scratch_regs(connector, ret); + pm_runtime_mark_last_busy(connector->dev->dev); + pm_runtime_put_autosuspend(connector->dev->dev); return ret; } @@ -735,6 +829,11 @@ radeon_vga_detect(struct drm_connector *connector, bool force) struct drm_encoder_helper_funcs *encoder_funcs; bool dret = false; enum drm_connector_status ret = connector_status_disconnected; + int r; + + r = pm_runtime_get_sync(connector->dev->dev); + if (r < 0) + return connector_status_disconnected; encoder = radeon_best_single_encoder(connector); if (!encoder) @@ -752,7 +851,7 @@ radeon_vga_detect(struct drm_connector *connector, bool force) if (!radeon_connector->edid) { DRM_ERROR("%s: probed a monitor but no|invalid EDID\n", - drm_get_connector_name(connector)); + connector->name); ret = connector_status_connected; } else { radeon_connector->use_digital = !!(radeon_connector->edid->input & DRM_EDID_INPUT_DIGITAL); @@ -775,9 +874,8 @@ radeon_vga_detect(struct drm_connector *connector, bool force) * detected a monitor via load. */ if (radeon_connector->detected_by_load) - return connector->status; - else - return ret; + ret = connector->status; + goto out; } if (radeon_connector->dac_load_detect && encoder) { @@ -802,6 +900,11 @@ radeon_vga_detect(struct drm_connector *connector, bool force) } radeon_connector_update_scratch_regs(connector, ret); + +out: + pm_runtime_mark_last_busy(connector->dev->dev); + pm_runtime_put_autosuspend(connector->dev->dev); + return ret; } @@ -858,10 +961,15 @@ radeon_tv_detect(struct drm_connector *connector, bool force) struct drm_encoder_helper_funcs *encoder_funcs; struct radeon_connector *radeon_connector = to_radeon_connector(connector); enum drm_connector_status ret = connector_status_disconnected; + int r; if (!radeon_connector->dac_load_detect) return ret; + r = pm_runtime_get_sync(connector->dev->dev); + if (r < 0) + return connector_status_disconnected; + encoder = radeon_best_single_encoder(connector); if (!encoder) ret = connector_status_disconnected; @@ -872,6 +980,8 @@ radeon_tv_detect(struct drm_connector *connector, bool force) if (ret == connector_status_connected) ret = radeon_connector_analog_encoder_conflict_solve(connector, encoder, ret, false); radeon_connector_update_scratch_regs(connector, ret); + pm_runtime_mark_last_busy(connector->dev->dev); + pm_runtime_put_autosuspend(connector->dev->dev); return ret; } @@ -939,12 +1049,18 @@ radeon_dvi_detect(struct drm_connector *connector, bool force) struct drm_encoder *encoder = NULL; struct drm_encoder_helper_funcs *encoder_funcs; struct drm_mode_object *obj; - int i; + int i, r; enum drm_connector_status ret = connector_status_disconnected; bool dret = false, broken_edid = false; - if (!force && radeon_check_hpd_status_unchanged(connector)) - return connector->status; + r = pm_runtime_get_sync(connector->dev->dev); + if (r < 0) + return connector_status_disconnected; + + if (!force && radeon_check_hpd_status_unchanged(connector)) { + ret = connector->status; + goto exit; + } if (radeon_connector->ddc_bus) dret = radeon_ddc_probe(radeon_connector, false); @@ -958,12 +1074,13 @@ radeon_dvi_detect(struct drm_connector *connector, bool force) if (!radeon_connector->edid) { DRM_ERROR("%s: probed a monitor but no|invalid EDID\n", - drm_get_connector_name(connector)); + connector->name); /* rs690 seems to have a problem with connectors not existing and always * return a block of 0's. If we see this just stop polling on this output */ if ((rdev->family == CHIP_RS690 || rdev->family == CHIP_RS740) && radeon_connector->base.null_edid_counter) { ret = connector_status_disconnected; - DRM_ERROR("%s: detected RS690 floating bus bug, stopping ddc detect\n", drm_get_connector_name(connector)); + DRM_ERROR("%s: detected RS690 floating bus bug, stopping ddc detect\n", + connector->name); radeon_connector->ddc_bus = NULL; } else { ret = connector_status_connected; @@ -1095,6 +1212,11 @@ out: /* updated in get modes as well since we need to know if it's analog or digital */ radeon_connector_update_scratch_regs(connector, ret); + +exit: + pm_runtime_mark_last_busy(connector->dev->dev); + pm_runtime_put_autosuspend(connector->dev->dev); + return ret; } @@ -1169,17 +1291,15 @@ static int radeon_dvi_mode_valid(struct drm_connector *connector, (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D) || (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_B)) return MODE_OK; - else if (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_A) { - if (ASIC_IS_DCE6(rdev)) { - /* HDMI 1.3+ supports max clock of 340 Mhz */ - if (mode->clock > 340000) - return MODE_CLOCK_HIGH; - else - return MODE_OK; - } else + else if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) { + /* HDMI 1.3+ supports max clock of 340 Mhz */ + if (mode->clock > 340000) return MODE_CLOCK_HIGH; - } else + else + return MODE_OK; + } else { return MODE_CLOCK_HIGH; + } } /* check against the max pixel clock */ @@ -1204,21 +1324,6 @@ static const struct drm_connector_funcs radeon_dvi_connector_funcs = { .force = radeon_dvi_force, }; -static void radeon_dp_connector_destroy(struct drm_connector *connector) -{ - struct radeon_connector *radeon_connector = to_radeon_connector(connector); - struct radeon_connector_atom_dig *radeon_dig_connector = radeon_connector->con_priv; - - if (radeon_connector->edid) - kfree(radeon_connector->edid); - if (radeon_dig_connector->dp_i2c_bus) - radeon_i2c_destroy(radeon_dig_connector->dp_i2c_bus); - kfree(radeon_connector->con_priv); - drm_sysfs_connector_remove(connector); - drm_connector_cleanup(connector); - kfree(connector); -} - static int radeon_dp_get_modes(struct drm_connector *connector) { struct radeon_connector *radeon_connector = to_radeon_connector(connector); @@ -1345,7 +1450,7 @@ bool radeon_connector_is_dp12_capable(struct drm_connector *connector) struct radeon_device *rdev = dev->dev_private; if (ASIC_IS_DCE5(rdev) && - (rdev->clock.dp_extclk >= 53900) && + (rdev->clock.default_dispclk >= 53900) && radeon_connector_encoder_is_hbr2(connector)) { return true; } @@ -1362,9 +1467,16 @@ radeon_dp_detect(struct drm_connector *connector, bool force) enum drm_connector_status ret = connector_status_disconnected; struct radeon_connector_atom_dig *radeon_dig_connector = radeon_connector->con_priv; struct drm_encoder *encoder = radeon_best_single_encoder(connector); + int r; - if (!force && radeon_check_hpd_status_unchanged(connector)) - return connector->status; + r = pm_runtime_get_sync(connector->dev->dev); + if (r < 0) + return connector_status_disconnected; + + if (!force && radeon_check_hpd_status_unchanged(connector)) { + ret = connector->status; + goto out; + } if (radeon_connector->edid) { kfree(radeon_connector->edid); @@ -1420,7 +1532,7 @@ radeon_dp_detect(struct drm_connector *connector, bool force) if (radeon_dp_getdpcd(radeon_connector)) ret = connector_status_connected; } else { - /* try non-aux ddc (DP to DVI/HMDI/etc. adapter) */ + /* try non-aux ddc (DP to DVI/HDMI/etc. adapter) */ if (radeon_ddc_probe(radeon_connector, false)) ret = connector_status_connected; } @@ -1428,12 +1540,18 @@ radeon_dp_detect(struct drm_connector *connector, bool force) } radeon_connector_update_scratch_regs(connector, ret); +out: + pm_runtime_mark_last_busy(connector->dev->dev); + pm_runtime_put_autosuspend(connector->dev->dev); + return ret; } static int radeon_dp_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) { + struct drm_device *dev = connector->dev; + struct radeon_device *rdev = dev->dev_private; struct radeon_connector *radeon_connector = to_radeon_connector(connector); struct radeon_connector_atom_dig *radeon_dig_connector = radeon_connector->con_priv; @@ -1464,14 +1582,23 @@ static int radeon_dp_mode_valid(struct drm_connector *connector, return MODE_PANEL; } } - return MODE_OK; } else { if ((radeon_dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) || - (radeon_dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) + (radeon_dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) { return radeon_dp_mode_valid_helper(connector, mode); - else - return MODE_OK; + } else { + if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) { + /* HDMI 1.3+ supports max clock of 340 Mhz */ + if (mode->clock > 340000) + return MODE_CLOCK_HIGH; + } else { + if (mode->clock > 165000) + return MODE_CLOCK_HIGH; + } + } } + + return MODE_OK; } static const struct drm_connector_helper_funcs radeon_dp_connector_helper_funcs = { @@ -1485,7 +1612,25 @@ static const struct drm_connector_funcs radeon_dp_connector_funcs = { .detect = radeon_dp_detect, .fill_modes = drm_helper_probe_single_connector_modes, .set_property = radeon_connector_set_property, - .destroy = radeon_dp_connector_destroy, + .destroy = radeon_connector_destroy, + .force = radeon_dvi_force, +}; + +static const struct drm_connector_funcs radeon_edp_connector_funcs = { + .dpms = drm_helper_connector_dpms, + .detect = radeon_dp_detect, + .fill_modes = drm_helper_probe_single_connector_modes, + .set_property = radeon_lvds_set_property, + .destroy = radeon_connector_destroy, + .force = radeon_dvi_force, +}; + +static const struct drm_connector_funcs radeon_lvds_bridge_connector_funcs = { + .dpms = drm_helper_connector_dpms, + .detect = radeon_dp_detect, + .fill_modes = drm_helper_probe_single_connector_modes, + .set_property = radeon_lvds_set_property, + .destroy = radeon_connector_destroy, .force = radeon_dvi_force, }; @@ -1509,6 +1654,7 @@ radeon_add_atom_connector(struct drm_device *dev, uint32_t subpixel_order = SubPixelNone; bool shared_ddc = false; bool is_dp_bridge = false; + bool has_aux = false; if (connector_type == DRM_MODE_CONNECTOR_Unknown) return; @@ -1580,24 +1726,21 @@ radeon_add_atom_connector(struct drm_device *dev, goto failed; radeon_dig_connector->igp_lane_info = igp_lane_info; radeon_connector->con_priv = radeon_dig_connector; - drm_connector_init(dev, &radeon_connector->base, &radeon_dp_connector_funcs, connector_type); - drm_connector_helper_add(&radeon_connector->base, &radeon_dp_connector_helper_funcs); if (i2c_bus->valid) { - /* add DP i2c bus */ - if (connector_type == DRM_MODE_CONNECTOR_eDP) - radeon_dig_connector->dp_i2c_bus = radeon_i2c_create_dp(dev, i2c_bus, "eDP-auxch"); - else - radeon_dig_connector->dp_i2c_bus = radeon_i2c_create_dp(dev, i2c_bus, "DP-auxch"); - if (!radeon_dig_connector->dp_i2c_bus) - DRM_ERROR("DP: Failed to assign dp ddc bus! Check dmesg for i2c errors.\n"); radeon_connector->ddc_bus = radeon_i2c_lookup(rdev, i2c_bus); - if (!radeon_connector->ddc_bus) + if (radeon_connector->ddc_bus) + has_aux = true; + else DRM_ERROR("DP: Failed to assign ddc bus! Check dmesg for i2c errors.\n"); } switch (connector_type) { case DRM_MODE_CONNECTOR_VGA: case DRM_MODE_CONNECTOR_DVIA: default: + drm_connector_init(dev, &radeon_connector->base, + &radeon_dp_connector_funcs, connector_type); + drm_connector_helper_add(&radeon_connector->base, + &radeon_dp_connector_helper_funcs); connector->interlace_allowed = true; connector->doublescan_allowed = true; radeon_connector->dac_load_detect = true; @@ -1610,6 +1753,10 @@ radeon_add_atom_connector(struct drm_device *dev, case DRM_MODE_CONNECTOR_HDMIA: case DRM_MODE_CONNECTOR_HDMIB: case DRM_MODE_CONNECTOR_DisplayPort: + drm_connector_init(dev, &radeon_connector->base, + &radeon_dp_connector_funcs, connector_type); + drm_connector_helper_add(&radeon_connector->base, + &radeon_dp_connector_helper_funcs); drm_object_attach_property(&radeon_connector->base.base, rdev->mode_info.underscan_property, UNDERSCAN_OFF); @@ -1619,6 +1766,16 @@ radeon_add_atom_connector(struct drm_device *dev, drm_object_attach_property(&radeon_connector->base.base, rdev->mode_info.underscan_vborder_property, 0); + + drm_object_attach_property(&radeon_connector->base.base, + rdev->mode_info.dither_property, + RADEON_FMT_DITHER_DISABLE); + + if (radeon_audio != 0) + drm_object_attach_property(&radeon_connector->base.base, + rdev->mode_info.audio_property, + RADEON_AUDIO_AUTO); + subpixel_order = SubPixelHorizontalRGB; connector->interlace_allowed = true; if (connector_type == DRM_MODE_CONNECTOR_HDMIB) @@ -1634,6 +1791,10 @@ radeon_add_atom_connector(struct drm_device *dev, break; case DRM_MODE_CONNECTOR_LVDS: case DRM_MODE_CONNECTOR_eDP: + drm_connector_init(dev, &radeon_connector->base, + &radeon_lvds_bridge_connector_funcs, connector_type); + drm_connector_helper_add(&radeon_connector->base, + &radeon_dp_connector_helper_funcs); drm_object_attach_property(&radeon_connector->base.base, dev->mode_config.scaling_mode_property, DRM_MODE_SCALE_FULLSCREEN); @@ -1708,6 +1869,16 @@ radeon_add_atom_connector(struct drm_device *dev, rdev->mode_info.underscan_vborder_property, 0); } + if (ASIC_IS_DCE2(rdev) && (radeon_audio != 0)) { + drm_object_attach_property(&radeon_connector->base.base, + rdev->mode_info.audio_property, + RADEON_AUDIO_AUTO); + } + if (ASIC_IS_AVIVO(rdev)) { + drm_object_attach_property(&radeon_connector->base.base, + rdev->mode_info.dither_property, + RADEON_FMT_DITHER_DISABLE); + } if (connector_type == DRM_MODE_CONNECTOR_DVII) { radeon_connector->dac_load_detect = true; drm_object_attach_property(&radeon_connector->base.base, @@ -1748,6 +1919,16 @@ radeon_add_atom_connector(struct drm_device *dev, rdev->mode_info.underscan_vborder_property, 0); } + if (ASIC_IS_DCE2(rdev) && (radeon_audio != 0)) { + drm_object_attach_property(&radeon_connector->base.base, + rdev->mode_info.audio_property, + RADEON_AUDIO_AUTO); + } + if (ASIC_IS_AVIVO(rdev)) { + drm_object_attach_property(&radeon_connector->base.base, + rdev->mode_info.dither_property, + RADEON_FMT_DITHER_DISABLE); + } subpixel_order = SubPixelHorizontalRGB; connector->interlace_allowed = true; if (connector_type == DRM_MODE_CONNECTOR_HDMIB) @@ -1764,12 +1945,10 @@ radeon_add_atom_connector(struct drm_device *dev, drm_connector_init(dev, &radeon_connector->base, &radeon_dp_connector_funcs, connector_type); drm_connector_helper_add(&radeon_connector->base, &radeon_dp_connector_helper_funcs); if (i2c_bus->valid) { - /* add DP i2c bus */ - radeon_dig_connector->dp_i2c_bus = radeon_i2c_create_dp(dev, i2c_bus, "DP-auxch"); - if (!radeon_dig_connector->dp_i2c_bus) - DRM_ERROR("DP: Failed to assign dp ddc bus! Check dmesg for i2c errors.\n"); radeon_connector->ddc_bus = radeon_i2c_lookup(rdev, i2c_bus); - if (!radeon_connector->ddc_bus) + if (radeon_connector->ddc_bus) + has_aux = true; + else DRM_ERROR("DP: Failed to assign ddc bus! Check dmesg for i2c errors.\n"); } subpixel_order = SubPixelHorizontalRGB; @@ -1787,6 +1966,17 @@ radeon_add_atom_connector(struct drm_device *dev, rdev->mode_info.underscan_vborder_property, 0); } + if (ASIC_IS_DCE2(rdev) && (radeon_audio != 0)) { + drm_object_attach_property(&radeon_connector->base.base, + rdev->mode_info.audio_property, + RADEON_AUDIO_AUTO); + } + if (ASIC_IS_AVIVO(rdev)) { + drm_object_attach_property(&radeon_connector->base.base, + rdev->mode_info.dither_property, + RADEON_FMT_DITHER_DISABLE); + + } connector->interlace_allowed = true; /* in theory with a DP to VGA converter... */ connector->doublescan_allowed = false; @@ -1797,15 +1987,13 @@ radeon_add_atom_connector(struct drm_device *dev, goto failed; radeon_dig_connector->igp_lane_info = igp_lane_info; radeon_connector->con_priv = radeon_dig_connector; - drm_connector_init(dev, &radeon_connector->base, &radeon_dp_connector_funcs, connector_type); + drm_connector_init(dev, &radeon_connector->base, &radeon_edp_connector_funcs, connector_type); drm_connector_helper_add(&radeon_connector->base, &radeon_dp_connector_helper_funcs); if (i2c_bus->valid) { - /* add DP i2c bus */ - radeon_dig_connector->dp_i2c_bus = radeon_i2c_create_dp(dev, i2c_bus, "eDP-auxch"); - if (!radeon_dig_connector->dp_i2c_bus) - DRM_ERROR("DP: Failed to assign dp ddc bus! Check dmesg for i2c errors.\n"); radeon_connector->ddc_bus = radeon_i2c_lookup(rdev, i2c_bus); - if (!radeon_connector->ddc_bus) + if (radeon_connector->ddc_bus) + has_aux = true; + else DRM_ERROR("DP: Failed to assign ddc bus! Check dmesg for i2c errors.\n"); } drm_object_attach_property(&radeon_connector->base.base, @@ -1863,6 +2051,10 @@ radeon_add_atom_connector(struct drm_device *dev, connector->display_info.subpixel_order = subpixel_order; drm_sysfs_connector_add(connector); + + if (has_aux) + radeon_dp_aux_init(radeon_connector); + return; failed: diff --git a/drivers/gpu/drm/radeon/radeon_cp.c b/drivers/gpu/drm/radeon/radeon_cp.c index efc4f6441ef..bb0d5c3a831 100644 --- a/drivers/gpu/drm/radeon/radeon_cp.c +++ b/drivers/gpu/drm/radeon/radeon_cp.c @@ -1444,13 +1444,13 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init, dev_priv->ring.end = ((u32 *) dev_priv->cp_ring->handle + init->ring_size / sizeof(u32)); dev_priv->ring.size = init->ring_size; - dev_priv->ring.size_l2qw = drm_order(init->ring_size / 8); + dev_priv->ring.size_l2qw = order_base_2(init->ring_size / 8); dev_priv->ring.rptr_update = /* init->rptr_update */ 4096; - dev_priv->ring.rptr_update_l2qw = drm_order( /* init->rptr_update */ 4096 / 8); + dev_priv->ring.rptr_update_l2qw = order_base_2( /* init->rptr_update */ 4096 / 8); dev_priv->ring.fetch_size = /* init->fetch_size */ 32; - dev_priv->ring.fetch_size_l2ow = drm_order( /* init->fetch_size */ 32 / 16); + dev_priv->ring.fetch_size_l2ow = order_base_2( /* init->fetch_size */ 32 / 16); dev_priv->ring.tail_mask = (dev_priv->ring.size / sizeof(u32)) - 1; dev_priv->ring.high_mark = RADEON_RING_HIGH_MARK; @@ -2020,10 +2020,10 @@ static int radeon_cp_get_buffers(struct drm_device *dev, buf->file_priv = file_priv; - if (DRM_COPY_TO_USER(&d->request_indices[i], &buf->idx, + if (copy_to_user(&d->request_indices[i], &buf->idx, sizeof(buf->idx))) return -EFAULT; - if (DRM_COPY_TO_USER(&d->request_sizes[i], &buf->total, + if (copy_to_user(&d->request_sizes[i], &buf->total, sizeof(buf->total))) return -EFAULT; @@ -2228,7 +2228,7 @@ void radeon_commit_ring(drm_radeon_private_t *dev_priv) dev_priv->ring.tail &= dev_priv->ring.tail_mask; - DRM_MEMORYBARRIER(); + mb(); GET_RING_HEAD( dev_priv ); if ((dev_priv->flags & RADEON_FAMILY_MASK) >= CHIP_R600) { diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 13a130fb351..ae763f60c8a 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -24,15 +24,59 @@ * Authors: * Jerome Glisse <glisse@freedesktop.org> */ +#include <linux/list_sort.h> #include <drm/drmP.h> #include <drm/radeon_drm.h> #include "radeon_reg.h" #include "radeon.h" +#include "radeon_trace.h" + +#define RADEON_CS_MAX_PRIORITY 32u +#define RADEON_CS_NUM_BUCKETS (RADEON_CS_MAX_PRIORITY + 1) + +/* This is based on the bucket sort with O(n) time complexity. + * An item with priority "i" is added to bucket[i]. The lists are then + * concatenated in descending order. + */ +struct radeon_cs_buckets { + struct list_head bucket[RADEON_CS_NUM_BUCKETS]; +}; + +static void radeon_cs_buckets_init(struct radeon_cs_buckets *b) +{ + unsigned i; + + for (i = 0; i < RADEON_CS_NUM_BUCKETS; i++) + INIT_LIST_HEAD(&b->bucket[i]); +} + +static void radeon_cs_buckets_add(struct radeon_cs_buckets *b, + struct list_head *item, unsigned priority) +{ + /* Since buffers which appear sooner in the relocation list are + * likely to be used more often than buffers which appear later + * in the list, the sort mustn't change the ordering of buffers + * with the same priority, i.e. it must be stable. + */ + list_add_tail(item, &b->bucket[min(priority, RADEON_CS_MAX_PRIORITY)]); +} + +static void radeon_cs_buckets_get_list(struct radeon_cs_buckets *b, + struct list_head *out_list) +{ + unsigned i; + + /* Connect the sorted buckets in the output list. */ + for (i = 0; i < RADEON_CS_NUM_BUCKETS; i++) { + list_splice(&b->bucket[i], out_list); + } +} static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) { struct drm_device *ddev = p->rdev->ddev; struct radeon_cs_chunk *chunk; + struct radeon_cs_buckets buckets; unsigned i, j; bool duplicate; @@ -51,8 +95,12 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) if (p->relocs == NULL) { return -ENOMEM; } + + radeon_cs_buckets_init(&buckets); + for (i = 0; i < p->nrelocs; i++) { struct drm_radeon_cs_reloc *r; + unsigned priority; duplicate = false; r = (struct drm_radeon_cs_reloc *)&chunk->kdata[i*4]; @@ -77,36 +125,59 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) } p->relocs_ptr[i] = &p->relocs[i]; p->relocs[i].robj = gem_to_radeon_bo(p->relocs[i].gobj); - p->relocs[i].lobj.bo = p->relocs[i].robj; - p->relocs[i].lobj.written = !!r->write_domain; - /* the first reloc of an UVD job is the - msg and that must be in VRAM */ - if (p->ring == R600_RING_TYPE_UVD_INDEX && i == 0) { + /* The userspace buffer priorities are from 0 to 15. A higher + * number means the buffer is more important. + * Also, the buffers used for write have a higher priority than + * the buffers used for read only, which doubles the range + * to 0 to 31. 32 is reserved for the kernel driver. + */ + priority = (r->flags & 0xf) * 2 + !!r->write_domain; + + /* the first reloc of an UVD job is the msg and that must be in + VRAM, also but everything into VRAM on AGP cards to avoid + image corruptions */ + if (p->ring == R600_RING_TYPE_UVD_INDEX && + (i == 0 || drm_pci_device_is_agp(p->rdev->ddev))) { /* TODO: is this still needed for NI+ ? */ - p->relocs[i].lobj.domain = + p->relocs[i].prefered_domains = RADEON_GEM_DOMAIN_VRAM; - p->relocs[i].lobj.alt_domain = + p->relocs[i].allowed_domains = RADEON_GEM_DOMAIN_VRAM; + /* prioritize this over any other relocation */ + priority = RADEON_CS_MAX_PRIORITY; } else { uint32_t domain = r->write_domain ? r->write_domain : r->read_domains; - p->relocs[i].lobj.domain = domain; + if (domain & RADEON_GEM_DOMAIN_CPU) { + DRM_ERROR("RADEON_GEM_DOMAIN_CPU is not valid " + "for command submission\n"); + return -EINVAL; + } + + p->relocs[i].prefered_domains = domain; if (domain == RADEON_GEM_DOMAIN_VRAM) domain |= RADEON_GEM_DOMAIN_GTT; - p->relocs[i].lobj.alt_domain = domain; + p->relocs[i].allowed_domains = domain; } - p->relocs[i].lobj.tv.bo = &p->relocs[i].robj->tbo; + p->relocs[i].tv.bo = &p->relocs[i].robj->tbo; p->relocs[i].handle = r->handle; - radeon_bo_list_add_object(&p->relocs[i].lobj, - &p->validated); + radeon_cs_buckets_add(&buckets, &p->relocs[i].tv.head, + priority); } - return radeon_bo_list_validate(&p->ticket, &p->validated, p->ring); + + radeon_cs_buckets_get_list(&buckets, &p->validated); + + if (p->cs_flags & RADEON_CS_USE_VM) + p->vm_bos = radeon_vm_get_bos(p->rdev, p->ib.vm, + &p->validated); + + return radeon_bo_list_validate(p->rdev, &p->ticket, &p->validated, p->ring); } static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority) @@ -135,7 +206,7 @@ static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority p->ring = R600_RING_TYPE_DMA_INDEX; else p->ring = CAYMAN_RING_TYPE_DMA1_INDEX; - } else if (p->rdev->family >= CHIP_R600) { + } else if (p->rdev->family >= CHIP_RV770) { p->ring = R600_RING_TYPE_DMA_INDEX; } else { return -EINVAL; @@ -144,6 +215,10 @@ static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority case RADEON_CS_RING_UVD: p->ring = R600_RING_TYPE_UVD_INDEX; break; + case RADEON_CS_RING_VCE: + /* TODO: only use the low priority ring for now */ + p->ring = TN_RING_TYPE_VCE1_INDEX; + break; } return 0; } @@ -156,7 +231,8 @@ static void radeon_cs_sync_rings(struct radeon_cs_parser *p) if (!p->relocs[i].robj) continue; - radeon_ib_sync_to(&p->ib, p->relocs[i].robj->tbo.sync_obj); + radeon_semaphore_sync_to(p->ib.semaphore, + p->relocs[i].robj->tbo.sync_obj); } } @@ -188,7 +264,7 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) return -ENOMEM; } chunk_array_ptr = (uint64_t *)(unsigned long)(cs->chunks); - if (DRM_COPY_FROM_USER(p->chunks_array, chunk_array_ptr, + if (copy_from_user(p->chunks_array, chunk_array_ptr, sizeof(uint64_t)*cs->num_chunks)) { return -EFAULT; } @@ -204,14 +280,12 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) uint32_t __user *cdata; chunk_ptr = (void __user*)(unsigned long)p->chunks_array[i]; - if (DRM_COPY_FROM_USER(&user_chunk, chunk_ptr, + if (copy_from_user(&user_chunk, chunk_ptr, sizeof(struct drm_radeon_cs_chunk))) { return -EFAULT; } p->chunks[i].length_dw = user_chunk.length_dw; - p->chunks[i].kdata = NULL; p->chunks[i].chunk_id = user_chunk.chunk_id; - p->chunks[i].user_ptr = (void __user *)(unsigned long)user_chunk.chunk_data; if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) { p->chunk_relocs_idx = i; } @@ -234,25 +308,31 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) return -EINVAL; } - cdata = (uint32_t *)(unsigned long)user_chunk.chunk_data; - if ((p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) || - (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS)) { - size = p->chunks[i].length_dw * sizeof(uint32_t); - p->chunks[i].kdata = kmalloc(size, GFP_KERNEL); - if (p->chunks[i].kdata == NULL) { - return -ENOMEM; - } - if (DRM_COPY_FROM_USER(p->chunks[i].kdata, - p->chunks[i].user_ptr, size)) { - return -EFAULT; - } - if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) { - p->cs_flags = p->chunks[i].kdata[0]; - if (p->chunks[i].length_dw > 1) - ring = p->chunks[i].kdata[1]; - if (p->chunks[i].length_dw > 2) - priority = (s32)p->chunks[i].kdata[2]; - } + size = p->chunks[i].length_dw; + cdata = (void __user *)(unsigned long)user_chunk.chunk_data; + p->chunks[i].user_ptr = cdata; + if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_CONST_IB) + continue; + + if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_IB) { + if (!p->rdev || !(p->rdev->flags & RADEON_IS_AGP)) + continue; + } + + p->chunks[i].kdata = drm_malloc_ab(size, sizeof(uint32_t)); + size *= sizeof(uint32_t); + if (p->chunks[i].kdata == NULL) { + return -ENOMEM; + } + if (copy_from_user(p->chunks[i].kdata, cdata, size)) { + return -EFAULT; + } + if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) { + p->cs_flags = p->chunks[i].kdata[0]; + if (p->chunks[i].length_dw > 1) + ring = p->chunks[i].kdata[1]; + if (p->chunks[i].length_dw > 2) + priority = (s32)p->chunks[i].kdata[2]; } } @@ -268,44 +348,33 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) return -EINVAL; /* we only support VM on some SI+ rings */ - if ((p->rdev->asic->ring[p->ring].cs_parse == NULL) && - ((p->cs_flags & RADEON_CS_USE_VM) == 0)) { - DRM_ERROR("Ring %d requires VM!\n", p->ring); - return -EINVAL; - } - } - - /* deal with non-vm */ - if ((p->chunk_ib_idx != -1) && - ((p->cs_flags & RADEON_CS_USE_VM) == 0) && - (p->chunks[p->chunk_ib_idx].chunk_id == RADEON_CHUNK_ID_IB)) { - if (p->chunks[p->chunk_ib_idx].length_dw > (16 * 1024)) { - DRM_ERROR("cs IB too big: %d\n", - p->chunks[p->chunk_ib_idx].length_dw); - return -EINVAL; - } - if (p->rdev && (p->rdev->flags & RADEON_IS_AGP)) { - p->chunks[p->chunk_ib_idx].kpage[0] = kmalloc(PAGE_SIZE, GFP_KERNEL); - p->chunks[p->chunk_ib_idx].kpage[1] = kmalloc(PAGE_SIZE, GFP_KERNEL); - if (p->chunks[p->chunk_ib_idx].kpage[0] == NULL || - p->chunks[p->chunk_ib_idx].kpage[1] == NULL) { - kfree(p->chunks[p->chunk_ib_idx].kpage[0]); - kfree(p->chunks[p->chunk_ib_idx].kpage[1]); - p->chunks[p->chunk_ib_idx].kpage[0] = NULL; - p->chunks[p->chunk_ib_idx].kpage[1] = NULL; - return -ENOMEM; + if ((p->cs_flags & RADEON_CS_USE_VM) == 0) { + if (p->rdev->asic->ring[p->ring]->cs_parse == NULL) { + DRM_ERROR("Ring %d requires VM!\n", p->ring); + return -EINVAL; + } + } else { + if (p->rdev->asic->ring[p->ring]->ib_parse == NULL) { + DRM_ERROR("VM not supported on ring %d!\n", + p->ring); + return -EINVAL; } } - p->chunks[p->chunk_ib_idx].kpage_idx[0] = -1; - p->chunks[p->chunk_ib_idx].kpage_idx[1] = -1; - p->chunks[p->chunk_ib_idx].last_copied_page = -1; - p->chunks[p->chunk_ib_idx].last_page_index = - ((p->chunks[p->chunk_ib_idx].length_dw * 4) - 1) / PAGE_SIZE; } return 0; } +static int cmp_size_smaller_first(void *priv, struct list_head *a, + struct list_head *b) +{ + struct radeon_cs_reloc *la = list_entry(a, struct radeon_cs_reloc, tv.head); + struct radeon_cs_reloc *lb = list_entry(b, struct radeon_cs_reloc, tv.head); + + /* Sort A before B if A is smaller. */ + return (int)la->robj->tbo.num_pages - (int)lb->robj->tbo.num_pages; +} + /** * cs_parser_fini() - clean parser states * @parser: parser structure holding parsing context. @@ -319,6 +388,18 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo unsigned i; if (!error) { + /* Sort the buffer list from the smallest to largest buffer, + * which affects the order of buffers in the LRU list. + * This assures that the smallest buffers are added first + * to the LRU list, so they are likely to be later evicted + * first, instead of large buffers whose eviction is more + * expensive. + * + * This slightly lowers the number of bytes moved by TTM + * per frame under memory pressure. + */ + list_sort(NULL, &parser->validated, cmp_size_smaller_first); + ttm_eu_fence_buffer_objects(&parser->ticket, &parser->validated, parser->ib.fence); @@ -336,13 +417,9 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo kfree(parser->track); kfree(parser->relocs); kfree(parser->relocs_ptr); - for (i = 0; i < parser->nchunks; i++) { - kfree(parser->chunks[i].kdata); - if ((parser->rdev->flags & RADEON_IS_AGP)) { - kfree(parser->chunks[i].kpage[0]); - kfree(parser->chunks[i].kpage[1]); - } - } + kfree(parser->vm_bos); + for (i = 0; i < parser->nchunks; i++) + drm_free_large(parser->chunks[i].kdata); kfree(parser->chunks); kfree(parser->chunks_array); radeon_ib_free(parser->rdev, &parser->ib); @@ -352,7 +429,6 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo static int radeon_cs_ib_chunk(struct radeon_device *rdev, struct radeon_cs_parser *parser) { - struct radeon_cs_chunk *ib_chunk; int r; if (parser->chunk_ib_idx == -1) @@ -361,28 +437,18 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev, if (parser->cs_flags & RADEON_CS_USE_VM) return 0; - ib_chunk = &parser->chunks[parser->chunk_ib_idx]; - /* Copy the packet into the IB, the parser will read from the - * input memory (cached) and write to the IB (which can be - * uncached). - */ - r = radeon_ib_get(rdev, parser->ring, &parser->ib, - NULL, ib_chunk->length_dw * 4); - if (r) { - DRM_ERROR("Failed to get ib !\n"); - return r; - } - parser->ib.length_dw = ib_chunk->length_dw; r = radeon_cs_parse(rdev, parser->ring, parser); if (r || parser->parser_error) { DRM_ERROR("Invalid command stream !\n"); return r; } - r = radeon_cs_finish_pages(parser); - if (r) { - DRM_ERROR("Invalid command stream !\n"); - return r; - } + + if (parser->ring == R600_RING_TYPE_UVD_INDEX) + radeon_uvd_note_usage(rdev); + else if ((parser->ring == TN_RING_TYPE_VCE1_INDEX) || + (parser->ring == TN_RING_TYPE_VCE2_INDEX)) + radeon_vce_note_usage(rdev); + radeon_cs_sync_rings(parser); r = radeon_ib_schedule(rdev, &parser->ib, NULL); if (r) { @@ -391,24 +457,48 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev, return r; } -static int radeon_bo_vm_update_pte(struct radeon_cs_parser *parser, +static int radeon_bo_vm_update_pte(struct radeon_cs_parser *p, struct radeon_vm *vm) { - struct radeon_device *rdev = parser->rdev; - struct radeon_bo_list *lobj; - struct radeon_bo *bo; - int r; + struct radeon_device *rdev = p->rdev; + struct radeon_bo_va *bo_va; + int i, r; - r = radeon_vm_bo_update_pte(rdev, vm, rdev->ring_tmp_bo.bo, &rdev->ring_tmp_bo.bo->tbo.mem); - if (r) { + r = radeon_vm_update_page_directory(rdev, vm); + if (r) + return r; + + r = radeon_vm_clear_freed(rdev, vm); + if (r) return r; + + if (vm->ib_bo_va == NULL) { + DRM_ERROR("Tmp BO not in VM!\n"); + return -EINVAL; } - list_for_each_entry(lobj, &parser->validated, tv.head) { - bo = lobj->bo; - r = radeon_vm_bo_update_pte(parser->rdev, vm, bo, &bo->tbo.mem); - if (r) { - return r; + + r = radeon_vm_bo_update(rdev, vm->ib_bo_va, + &rdev->ring_tmp_bo.bo->tbo.mem); + if (r) + return r; + + for (i = 0; i < p->nrelocs; i++) { + struct radeon_bo *bo; + + /* ignore duplicates */ + if (p->relocs_ptr[i] != &p->relocs[i]) + continue; + + bo = p->relocs[i].robj; + bo_va = radeon_vm_bo_find(vm, bo); + if (bo_va == NULL) { + dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm); + return -EINVAL; } + + r = radeon_vm_bo_update(rdev, bo_va, &bo->tbo.mem); + if (r) + return r; } return 0; } @@ -416,7 +506,6 @@ static int radeon_bo_vm_update_pte(struct radeon_cs_parser *parser, static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, struct radeon_cs_parser *parser) { - struct radeon_cs_chunk *ib_chunk; struct radeon_fpriv *fpriv = parser->filp->driver_priv; struct radeon_vm *vm = &fpriv->vm; int r; @@ -426,68 +515,28 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, if ((parser->cs_flags & RADEON_CS_USE_VM) == 0) return 0; - if ((rdev->family >= CHIP_TAHITI) && - (parser->chunk_const_ib_idx != -1)) { - ib_chunk = &parser->chunks[parser->chunk_const_ib_idx]; - if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) { - DRM_ERROR("cs IB CONST too big: %d\n", ib_chunk->length_dw); - return -EINVAL; - } - r = radeon_ib_get(rdev, parser->ring, &parser->const_ib, - vm, ib_chunk->length_dw * 4); - if (r) { - DRM_ERROR("Failed to get const ib !\n"); - return r; - } - parser->const_ib.is_const_ib = true; - parser->const_ib.length_dw = ib_chunk->length_dw; - /* Copy the packet into the IB */ - if (DRM_COPY_FROM_USER(parser->const_ib.ptr, ib_chunk->user_ptr, - ib_chunk->length_dw * 4)) { - return -EFAULT; - } + if (parser->const_ib.length_dw) { r = radeon_ring_ib_parse(rdev, parser->ring, &parser->const_ib); if (r) { return r; } } - ib_chunk = &parser->chunks[parser->chunk_ib_idx]; - if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) { - DRM_ERROR("cs IB too big: %d\n", ib_chunk->length_dw); - return -EINVAL; - } - r = radeon_ib_get(rdev, parser->ring, &parser->ib, - vm, ib_chunk->length_dw * 4); - if (r) { - DRM_ERROR("Failed to get ib !\n"); - return r; - } - parser->ib.length_dw = ib_chunk->length_dw; - /* Copy the packet into the IB */ - if (DRM_COPY_FROM_USER(parser->ib.ptr, ib_chunk->user_ptr, - ib_chunk->length_dw * 4)) { - return -EFAULT; - } r = radeon_ring_ib_parse(rdev, parser->ring, &parser->ib); if (r) { return r; } - mutex_lock(&rdev->vm_manager.lock); + if (parser->ring == R600_RING_TYPE_UVD_INDEX) + radeon_uvd_note_usage(rdev); + mutex_lock(&vm->mutex); - r = radeon_vm_alloc_pt(rdev, vm); - if (r) { - goto out; - } r = radeon_bo_vm_update_pte(parser, vm); if (r) { goto out; } radeon_cs_sync_rings(parser); - radeon_ib_sync_to(&parser->ib, vm->fence); - radeon_ib_sync_to(&parser->ib, radeon_vm_grab_id( - rdev, vm, parser->ring)); + radeon_semaphore_sync_to(parser->ib.semaphore, vm->fence); if ((rdev->family >= CHIP_TAHITI) && (parser->chunk_const_ib_idx != -1)) { @@ -496,14 +545,8 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, r = radeon_ib_schedule(rdev, &parser->ib, NULL); } - if (!r) { - radeon_vm_fence(rdev, vm, parser->ib.fence); - } - out: - radeon_vm_add_to_lru(rdev, vm); mutex_unlock(&vm->mutex); - mutex_unlock(&rdev->vm_manager.lock); return r; } @@ -517,6 +560,62 @@ static int radeon_cs_handle_lockup(struct radeon_device *rdev, int r) return r; } +static int radeon_cs_ib_fill(struct radeon_device *rdev, struct radeon_cs_parser *parser) +{ + struct radeon_cs_chunk *ib_chunk; + struct radeon_vm *vm = NULL; + int r; + + if (parser->chunk_ib_idx == -1) + return 0; + + if (parser->cs_flags & RADEON_CS_USE_VM) { + struct radeon_fpriv *fpriv = parser->filp->driver_priv; + vm = &fpriv->vm; + + if ((rdev->family >= CHIP_TAHITI) && + (parser->chunk_const_ib_idx != -1)) { + ib_chunk = &parser->chunks[parser->chunk_const_ib_idx]; + if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) { + DRM_ERROR("cs IB CONST too big: %d\n", ib_chunk->length_dw); + return -EINVAL; + } + r = radeon_ib_get(rdev, parser->ring, &parser->const_ib, + vm, ib_chunk->length_dw * 4); + if (r) { + DRM_ERROR("Failed to get const ib !\n"); + return r; + } + parser->const_ib.is_const_ib = true; + parser->const_ib.length_dw = ib_chunk->length_dw; + if (copy_from_user(parser->const_ib.ptr, + ib_chunk->user_ptr, + ib_chunk->length_dw * 4)) + return -EFAULT; + } + + ib_chunk = &parser->chunks[parser->chunk_ib_idx]; + if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) { + DRM_ERROR("cs IB too big: %d\n", ib_chunk->length_dw); + return -EINVAL; + } + } + ib_chunk = &parser->chunks[parser->chunk_ib_idx]; + + r = radeon_ib_get(rdev, parser->ring, &parser->ib, + vm, ib_chunk->length_dw * 4); + if (r) { + DRM_ERROR("Failed to get ib !\n"); + return r; + } + parser->ib.length_dw = ib_chunk->length_dw; + if (ib_chunk->kdata) + memcpy(parser->ib.ptr, ib_chunk->kdata, ib_chunk->length_dw * 4); + else if (copy_from_user(parser->ib.ptr, ib_chunk->user_ptr, ib_chunk->length_dw * 4)) + return -EFAULT; + return 0; +} + int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) { struct radeon_device *rdev = dev->dev_private; @@ -542,19 +641,22 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) r = radeon_cs_handle_lockup(rdev, r); return r; } - r = radeon_cs_parser_relocs(&parser); - if (r) { - if (r != -ERESTARTSYS) + + r = radeon_cs_ib_fill(rdev, &parser); + if (!r) { + r = radeon_cs_parser_relocs(&parser); + if (r && r != -ERESTARTSYS) DRM_ERROR("Failed to parse relocation %d!\n", r); + } + + if (r) { radeon_cs_parser_fini(&parser, r, false); up_read(&rdev->exclusive_lock); r = radeon_cs_handle_lockup(rdev, r); return r; } - /* XXX pick SD/HD/MVC */ - if (parser.ring == R600_RING_TYPE_UVD_INDEX) - radeon_uvd_note_usage(rdev); + trace_radeon_cs(&parser); r = radeon_cs_ib_chunk(rdev, &parser); if (r) { @@ -571,97 +673,6 @@ out: return r; } -int radeon_cs_finish_pages(struct radeon_cs_parser *p) -{ - struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx]; - int i; - int size = PAGE_SIZE; - - for (i = ibc->last_copied_page + 1; i <= ibc->last_page_index; i++) { - if (i == ibc->last_page_index) { - size = (ibc->length_dw * 4) % PAGE_SIZE; - if (size == 0) - size = PAGE_SIZE; - } - - if (DRM_COPY_FROM_USER(p->ib.ptr + (i * (PAGE_SIZE/4)), - ibc->user_ptr + (i * PAGE_SIZE), - size)) - return -EFAULT; - } - return 0; -} - -static int radeon_cs_update_pages(struct radeon_cs_parser *p, int pg_idx) -{ - int new_page; - struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx]; - int i; - int size = PAGE_SIZE; - bool copy1 = (p->rdev && (p->rdev->flags & RADEON_IS_AGP)) ? - false : true; - - for (i = ibc->last_copied_page + 1; i < pg_idx; i++) { - if (DRM_COPY_FROM_USER(p->ib.ptr + (i * (PAGE_SIZE/4)), - ibc->user_ptr + (i * PAGE_SIZE), - PAGE_SIZE)) { - p->parser_error = -EFAULT; - return 0; - } - } - - if (pg_idx == ibc->last_page_index) { - size = (ibc->length_dw * 4) % PAGE_SIZE; - if (size == 0) - size = PAGE_SIZE; - } - - new_page = ibc->kpage_idx[0] < ibc->kpage_idx[1] ? 0 : 1; - if (copy1) - ibc->kpage[new_page] = p->ib.ptr + (pg_idx * (PAGE_SIZE / 4)); - - if (DRM_COPY_FROM_USER(ibc->kpage[new_page], - ibc->user_ptr + (pg_idx * PAGE_SIZE), - size)) { - p->parser_error = -EFAULT; - return 0; - } - - /* copy to IB for non single case */ - if (!copy1) - memcpy((void *)(p->ib.ptr+(pg_idx*(PAGE_SIZE/4))), ibc->kpage[new_page], size); - - ibc->last_copied_page = pg_idx; - ibc->kpage_idx[new_page] = pg_idx; - - return new_page; -} - -u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx) -{ - struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx]; - u32 pg_idx, pg_offset; - u32 idx_value = 0; - int new_page; - - pg_idx = (idx * 4) / PAGE_SIZE; - pg_offset = (idx * 4) % PAGE_SIZE; - - if (ibc->kpage_idx[0] == pg_idx) - return ibc->kpage[0][pg_offset/4]; - if (ibc->kpage_idx[1] == pg_idx) - return ibc->kpage[1][pg_offset/4]; - - new_page = radeon_cs_update_pages(p, pg_idx); - if (new_page < 0) { - p->parser_error = new_page; - return 0; - } - - idx_value = ibc->kpage[new_page][pg_offset/4]; - return idx_value; -} - /** * radeon_cs_packet_parse() - parse cp packet and point ib index to next packet * @parser: parser structure holding parsing context. @@ -803,9 +814,9 @@ int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p, /* FIXME: we assume reloc size is 4 dwords */ if (nomm) { *cs_reloc = p->relocs; - (*cs_reloc)->lobj.gpu_offset = + (*cs_reloc)->gpu_offset = (u64)relocs_chunk->kdata[idx + 3] << 32; - (*cs_reloc)->lobj.gpu_offset |= relocs_chunk->kdata[idx + 0]; + (*cs_reloc)->gpu_offset |= relocs_chunk->kdata[idx + 0]; } else *cs_reloc = p->relocs_ptr[(idx / 4)]; return 0; diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index 82335e38ec4..697add2cd4e 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -98,9 +98,20 @@ static const char radeon_family_name[][16] = { "BONAIRE", "KAVERI", "KABINI", + "HAWAII", + "MULLINS", "LAST", }; +bool radeon_is_px(struct drm_device *dev) +{ + struct radeon_device *rdev = dev->dev_private; + + if (rdev->flags & RADEON_IS_PX) + return true; + return false; +} + /** * radeon_program_register_sequence - program an array of registers. * @@ -137,6 +148,11 @@ void radeon_program_register_sequence(struct radeon_device *rdev, } } +void radeon_pci_config_reset(struct radeon_device *rdev) +{ + pci_write_config_dword(rdev->pdev, 0x7c, RADEON_ASIC_RESET_DATA); +} + /** * radeon_surface_init - Clear GPU surface registers. * @@ -242,30 +258,25 @@ void radeon_scratch_free(struct radeon_device *rdev, uint32_t reg) * Init doorbell driver information (CIK) * Returns 0 on success, error on failure. */ -int radeon_doorbell_init(struct radeon_device *rdev) +static int radeon_doorbell_init(struct radeon_device *rdev) { - int i; - /* doorbell bar mapping */ rdev->doorbell.base = pci_resource_start(rdev->pdev, 2); rdev->doorbell.size = pci_resource_len(rdev->pdev, 2); - /* limit to 4 MB for now */ - if (rdev->doorbell.size > (4 * 1024 * 1024)) - rdev->doorbell.size = 4 * 1024 * 1024; + rdev->doorbell.num_doorbells = min_t(u32, rdev->doorbell.size / sizeof(u32), RADEON_MAX_DOORBELLS); + if (rdev->doorbell.num_doorbells == 0) + return -EINVAL; - rdev->doorbell.ptr = ioremap(rdev->doorbell.base, rdev->doorbell.size); + rdev->doorbell.ptr = ioremap(rdev->doorbell.base, rdev->doorbell.num_doorbells * sizeof(u32)); if (rdev->doorbell.ptr == NULL) { return -ENOMEM; } DRM_INFO("doorbell mmio base: 0x%08X\n", (uint32_t)rdev->doorbell.base); DRM_INFO("doorbell mmio size: %u\n", (unsigned)rdev->doorbell.size); - rdev->doorbell.num_pages = rdev->doorbell.size / PAGE_SIZE; + memset(&rdev->doorbell.used, 0, sizeof(rdev->doorbell.used)); - for (i = 0; i < rdev->doorbell.num_pages; i++) { - rdev->doorbell.free[i] = true; - } return 0; } @@ -276,47 +287,45 @@ int radeon_doorbell_init(struct radeon_device *rdev) * * Tear down doorbell driver information (CIK) */ -void radeon_doorbell_fini(struct radeon_device *rdev) +static void radeon_doorbell_fini(struct radeon_device *rdev) { iounmap(rdev->doorbell.ptr); rdev->doorbell.ptr = NULL; } /** - * radeon_doorbell_get - Allocate a doorbell page + * radeon_doorbell_get - Allocate a doorbell entry * * @rdev: radeon_device pointer - * @doorbell: doorbell page number + * @doorbell: doorbell index * - * Allocate a doorbell page for use by the driver (all asics). + * Allocate a doorbell for use by the driver (all asics). * Returns 0 on success or -EINVAL on failure. */ int radeon_doorbell_get(struct radeon_device *rdev, u32 *doorbell) { - int i; - - for (i = 0; i < rdev->doorbell.num_pages; i++) { - if (rdev->doorbell.free[i]) { - rdev->doorbell.free[i] = false; - *doorbell = i; - return 0; - } + unsigned long offset = find_first_zero_bit(rdev->doorbell.used, rdev->doorbell.num_doorbells); + if (offset < rdev->doorbell.num_doorbells) { + __set_bit(offset, rdev->doorbell.used); + *doorbell = offset; + return 0; + } else { + return -EINVAL; } - return -EINVAL; } /** - * radeon_doorbell_free - Free a doorbell page + * radeon_doorbell_free - Free a doorbell entry * * @rdev: radeon_device pointer - * @doorbell: doorbell page number + * @doorbell: doorbell index * - * Free a doorbell page allocated for use by the driver (all asics) + * Free a doorbell allocated for use by the driver (all asics) */ void radeon_doorbell_free(struct radeon_device *rdev, u32 doorbell) { - if (doorbell < rdev->doorbell.num_pages) - rdev->doorbell.free[doorbell] = true; + if (doorbell < rdev->doorbell.num_doorbells) + __clear_bit(doorbell, rdev->doorbell.used); } /* @@ -1003,16 +1012,28 @@ static void radeon_check_arguments(struct radeon_device *rdev) radeon_vram_limit = 0; } + if (radeon_gart_size == -1) { + /* default to a larger gart size on newer asics */ + if (rdev->family >= CHIP_RV770) + radeon_gart_size = 1024; + else + radeon_gart_size = 512; + } /* gtt size must be power of two and greater or equal to 32M */ if (radeon_gart_size < 32) { - dev_warn(rdev->dev, "gart size (%d) too small forcing to 512M\n", + dev_warn(rdev->dev, "gart size (%d) too small\n", radeon_gart_size); - radeon_gart_size = 512; - + if (rdev->family >= CHIP_RV770) + radeon_gart_size = 1024; + else + radeon_gart_size = 512; } else if (!radeon_check_pot_argument(radeon_gart_size)) { dev_warn(rdev->dev, "gart size (%d) must be a power of 2\n", radeon_gart_size); - radeon_gart_size = 512; + if (rdev->family >= CHIP_RV770) + radeon_gart_size = 1024; + else + radeon_gart_size = 512; } rdev->mc.gtt_size = (uint64_t)radeon_gart_size << 20; @@ -1031,6 +1052,43 @@ static void radeon_check_arguments(struct radeon_device *rdev) radeon_agpmode = 0; break; } + + if (!radeon_check_pot_argument(radeon_vm_size)) { + dev_warn(rdev->dev, "VM size (%d) must be a power of 2\n", + radeon_vm_size); + radeon_vm_size = 4; + } + + if (radeon_vm_size < 1) { + dev_warn(rdev->dev, "VM size (%d) to small, min is 1GB\n", + radeon_vm_size); + radeon_vm_size = 4; + } + + /* + * Max GPUVM size for Cayman, SI and CI are 40 bits. + */ + if (radeon_vm_size > 1024) { + dev_warn(rdev->dev, "VM size (%d) too large, max is 1TB\n", + radeon_vm_size); + radeon_vm_size = 4; + } + + /* defines number of bits in page table versus page directory, + * a page is 4KB so we have 12 bits offset, minimum 9 bits in the + * page table and the remaining bits are in the page directory */ + if (radeon_vm_block_size < 9) { + dev_warn(rdev->dev, "VM page table size (%d) too small\n", + radeon_vm_block_size); + radeon_vm_block_size = 9; + } + + if (radeon_vm_block_size > 24 || + (radeon_vm_size * 1024) < (1ull << radeon_vm_block_size)) { + dev_warn(rdev->dev, "VM page table size (%d) too large\n", + radeon_vm_block_size); + radeon_vm_block_size = 9; + } } /** @@ -1064,7 +1122,10 @@ static bool radeon_switcheroo_quirk_long_wakeup(struct pci_dev *pdev) static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_state state) { struct drm_device *dev = pci_get_drvdata(pdev); - pm_message_t pmm = { .event = PM_EVENT_SUSPEND }; + + if (radeon_is_px(dev) && state == VGA_SWITCHEROO_OFF) + return; + if (state == VGA_SWITCHEROO_ON) { unsigned d3_delay = dev->pdev->d3_delay; @@ -1075,7 +1136,7 @@ static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero if (d3_delay < 20 && radeon_switcheroo_quirk_long_wakeup(pdev)) dev->pdev->d3_delay = 20; - radeon_resume_kms(dev); + radeon_resume_kms(dev, true, true); dev->pdev->d3_delay = d3_delay; @@ -1085,7 +1146,7 @@ static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero printk(KERN_INFO "radeon: switched off\n"); drm_kms_helper_poll_disable(dev); dev->switch_power_state = DRM_SWITCH_POWER_CHANGING; - radeon_suspend_kms(dev, pmm); + radeon_suspend_kms(dev, true, true); dev->switch_power_state = DRM_SWITCH_POWER_OFF; } } @@ -1102,12 +1163,13 @@ static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero static bool radeon_switcheroo_can_switch(struct pci_dev *pdev) { struct drm_device *dev = pci_get_drvdata(pdev); - bool can_switch; - spin_lock(&dev->count_lock); - can_switch = (dev->open_count == 0); - spin_unlock(&dev->count_lock); - return can_switch; + /* + * FIXME: open_count is protected by drm_global_mutex but that would lead to + * locking inversion with the driver load path. And the access here is + * completely racy anyway. So don't bother with locking for now. + */ + return dev->open_count == 0; } static const struct vga_switcheroo_client_ops radeon_switcheroo_ops = { @@ -1135,6 +1197,7 @@ int radeon_device_init(struct radeon_device *rdev, { int r, i; int dma_bits; + bool runtime = false; rdev->shutdown = false; rdev->dev = &pdev->dev; @@ -1144,7 +1207,7 @@ int radeon_device_init(struct radeon_device *rdev, rdev->family = flags & RADEON_FAMILY_MASK; rdev->is_atom_bios = false; rdev->usec_timeout = RADEON_MAX_USEC_TIMEOUT; - rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024; + rdev->mc.gtt_size = 512 * 1024 * 1024; rdev->accel_working = false; /* set up ring ids */ for (i = 0; i < RADEON_NUM_RINGS; i++) { @@ -1163,26 +1226,24 @@ int radeon_device_init(struct radeon_device *rdev, mutex_init(&rdev->gem.mutex); mutex_init(&rdev->pm.mutex); mutex_init(&rdev->gpu_clock_mutex); + mutex_init(&rdev->srbm_mutex); init_rwsem(&rdev->pm.mclk_lock); init_rwsem(&rdev->exclusive_lock); init_waitqueue_head(&rdev->irq.vblank_queue); r = radeon_gem_init(rdev); if (r) return r; - /* initialize vm here */ - mutex_init(&rdev->vm_manager.lock); + + radeon_check_arguments(rdev); /* Adjust VM size here. - * Currently set to 4GB ((1 << 20) 4k pages). - * Max GPUVM size for cayman and SI is 40 bits. + * Max GPUVM size for cayman+ is 40 bits. */ - rdev->vm_manager.max_pfn = 1 << 20; - INIT_LIST_HEAD(&rdev->vm_manager.lru_vm); + rdev->vm_manager.max_pfn = radeon_vm_size << 18; /* Set asic functions */ r = radeon_asic_init(rdev); if (r) return r; - radeon_check_arguments(rdev); /* all of the newer IGP chips have an internal gart * However some rs4xx report as AGP, so remove that here. @@ -1236,6 +1297,17 @@ int radeon_device_init(struct radeon_device *rdev, /* Registers mapping */ /* TODO: block userspace mapping of io register */ spin_lock_init(&rdev->mmio_idx_lock); + spin_lock_init(&rdev->smc_idx_lock); + spin_lock_init(&rdev->pll_idx_lock); + spin_lock_init(&rdev->mc_idx_lock); + spin_lock_init(&rdev->pcie_idx_lock); + spin_lock_init(&rdev->pciep_idx_lock); + spin_lock_init(&rdev->pif_idx_lock); + spin_lock_init(&rdev->cg_idx_lock); + spin_lock_init(&rdev->uvd_idx_lock); + spin_lock_init(&rdev->rcu_idx_lock); + spin_lock_init(&rdev->didt_idx_lock); + spin_lock_init(&rdev->end_idx_lock); if (rdev->family >= CHIP_BONAIRE) { rdev->rmmio_base = pci_resource_start(rdev->pdev, 5); rdev->rmmio_size = pci_resource_len(rdev->pdev, 5); @@ -1269,7 +1341,12 @@ int radeon_device_init(struct radeon_device *rdev, /* this will fail for cards that aren't VGA class devices, just * ignore it */ vga_client_register(rdev->pdev, rdev, NULL, radeon_vga_set_decode); - vga_switcheroo_register_client(rdev->pdev, &radeon_switcheroo_ops); + + if (rdev->flags & RADEON_IS_PX) + runtime = true; + vga_switcheroo_register_client(rdev->pdev, &radeon_switcheroo_ops, runtime); + if (runtime) + vga_switcheroo_init_domain_pm_ops(rdev->dev, &rdev->vga_pm_domain); r = radeon_init(rdev); if (r) @@ -1295,14 +1372,24 @@ int radeon_device_init(struct radeon_device *rdev, if (r) return r; } + if ((radeon_testing & 1)) { - radeon_test_moves(rdev); + if (rdev->accel_working) + radeon_test_moves(rdev); + else + DRM_INFO("radeon: acceleration disabled, skipping move tests\n"); } if ((radeon_testing & 2)) { - radeon_test_syncing(rdev); + if (rdev->accel_working) + radeon_test_syncing(rdev); + else + DRM_INFO("radeon: acceleration disabled, skipping sync tests\n"); } if (radeon_benchmarking) { - radeon_benchmark(rdev, radeon_benchmarking); + if (rdev->accel_working) + radeon_benchmark(rdev, radeon_benchmarking); + else + DRM_INFO("radeon: acceleration disabled, skipping benchmarks\n"); } return 0; } @@ -1350,7 +1437,7 @@ void radeon_device_fini(struct radeon_device *rdev) * Returns 0 for success or an error on failure. * Called at driver suspend. */ -int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) +int radeon_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon) { struct radeon_device *rdev; struct drm_crtc *crtc; @@ -1361,9 +1448,7 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) if (dev == NULL || dev->dev_private == NULL) { return -ENODEV; } - if (state.event == PM_EVENT_PRETHAW) { - return 0; - } + rdev = dev->dev_private; if (dev->switch_power_state == DRM_SWITCH_POWER_OFF) @@ -1378,7 +1463,7 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) /* unpin the front buffers */ list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { - struct radeon_framebuffer *rfb = to_radeon_framebuffer(crtc->fb); + struct radeon_framebuffer *rfb = to_radeon_framebuffer(crtc->primary->fb); struct radeon_bo *robj; if (rfb == NULL || rfb->obj == NULL) { @@ -1397,10 +1482,9 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) /* evict vram memory */ radeon_bo_evict_vram(rdev); - mutex_lock(&rdev->ring_lock); /* wait for gpu to finish processing current batch */ for (i = 0; i < RADEON_NUM_RINGS; i++) { - r = radeon_fence_wait_empty_locked(rdev, i); + r = radeon_fence_wait_empty(rdev, i); if (r) { /* delay GPU reset to resume */ force_completion = true; @@ -1409,11 +1493,9 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) if (force_completion) { radeon_fence_driver_force_completion(rdev); } - mutex_unlock(&rdev->ring_lock); radeon_save_bios_scratch_regs(rdev); - radeon_pm_suspend(rdev); radeon_suspend(rdev); radeon_hpd_fini(rdev); /* evict remaining vram memory */ @@ -1422,14 +1504,17 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) radeon_agp_suspend(rdev); pci_save_state(dev->pdev); - if (state.event == PM_EVENT_SUSPEND) { + if (suspend) { /* Shut down the device */ pci_disable_device(dev->pdev); pci_set_power_state(dev->pdev, PCI_D3hot); } - console_lock(); - radeon_fbdev_set_suspend(rdev, 1); - console_unlock(); + + if (fbcon) { + console_lock(); + radeon_fbdev_set_suspend(rdev, 1); + console_unlock(); + } return 0; } @@ -1442,7 +1527,7 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) * Returns 0 for success or an error on failure. * Called at driver resume. */ -int radeon_resume_kms(struct drm_device *dev) +int radeon_resume_kms(struct drm_device *dev, bool resume, bool fbcon) { struct drm_connector *connector; struct radeon_device *rdev = dev->dev_private; @@ -1451,12 +1536,17 @@ int radeon_resume_kms(struct drm_device *dev) if (dev->switch_power_state == DRM_SWITCH_POWER_OFF) return 0; - console_lock(); - pci_set_power_state(dev->pdev, PCI_D0); - pci_restore_state(dev->pdev); - if (pci_enable_device(dev->pdev)) { - console_unlock(); - return -1; + if (fbcon) { + console_lock(); + } + if (resume) { + pci_set_power_state(dev->pdev, PCI_D0); + pci_restore_state(dev->pdev); + if (pci_enable_device(dev->pdev)) { + if (fbcon) + console_unlock(); + return -1; + } } /* resume AGP if in use */ radeon_agp_resume(rdev); @@ -1466,11 +1556,19 @@ int radeon_resume_kms(struct drm_device *dev) if (r) DRM_ERROR("ib ring test failed (%d).\n", r); - radeon_pm_resume(rdev); - radeon_restore_bios_scratch_regs(rdev); + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { + /* do dpm late init */ + r = radeon_pm_late_init(rdev); + if (r) { + rdev->pm.dpm_enabled = false; + DRM_ERROR("radeon_pm_late_init failed, disabling dpm\n"); + } + } else { + /* resume old pm late */ + radeon_pm_resume(rdev); + } - radeon_fbdev_set_suspend(rdev, 0); - console_unlock(); + radeon_restore_bios_scratch_regs(rdev); /* init dig PHYs, disp eng pll */ if (rdev->is_atom_bios) { @@ -1487,13 +1585,25 @@ int radeon_resume_kms(struct drm_device *dev) /* reset hpd state */ radeon_hpd_init(rdev); /* blat the mode back in */ - drm_helper_resume_force_mode(dev); - /* turn on display hw */ - list_for_each_entry(connector, &dev->mode_config.connector_list, head) { - drm_helper_connector_dpms(connector, DRM_MODE_DPMS_ON); + if (fbcon) { + drm_helper_resume_force_mode(dev); + /* turn on display hw */ + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + drm_helper_connector_dpms(connector, DRM_MODE_DPMS_ON); + } } drm_kms_helper_poll_enable(dev); + + /* set the power state here in case we are a PX system or headless */ + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) + radeon_pm_compute_clocks(rdev); + + if (fbcon) { + radeon_fbdev_set_suspend(rdev, 0); + console_unlock(); + } + return 0; } @@ -1516,9 +1626,18 @@ int radeon_gpu_reset(struct radeon_device *rdev) int resched; down_write(&rdev->exclusive_lock); + + if (!rdev->needs_reset) { + up_write(&rdev->exclusive_lock); + return 0; + } + + rdev->needs_reset = false; + radeon_save_bios_scratch_regs(rdev); /* block TTM */ resched = ttm_bo_lock_delayed_workqueue(&rdev->mman.bdev); + radeon_pm_suspend(rdev); radeon_suspend(rdev); for (i = 0; i < RADEON_NUM_RINGS; ++i) { @@ -1564,6 +1683,7 @@ retry: } } + radeon_pm_resume(rdev); drm_helper_resume_force_mode(rdev->ddev); ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched); diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index c2b67b4e1ac..bf25061c8ac 100644 --- a/drivers/gpu/drm/radeon/radeon_display.c +++ b/drivers/gpu/drm/radeon/radeon_display.c @@ -30,9 +30,12 @@ #include "atom.h" #include <asm/div64.h> +#include <linux/pm_runtime.h> #include <drm/drm_crtc_helper.h> #include <drm/drm_edid.h> +#include <linux/gcd.h> + static void avivo_crtc_load_lut(struct drm_crtc *crtc) { struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); @@ -63,7 +66,8 @@ static void avivo_crtc_load_lut(struct drm_crtc *crtc) (radeon_crtc->lut_b[i] << 0)); } - WREG32(AVIVO_D1GRPH_LUT_SEL + radeon_crtc->crtc_offset, radeon_crtc->crtc_id); + /* Only change bit 0 of LUT_SEL, other bits are set elsewhere */ + WREG32_P(AVIVO_D1GRPH_LUT_SEL + radeon_crtc->crtc_offset, radeon_crtc->crtc_id, ~1); } static void dce4_crtc_load_lut(struct drm_crtc *crtc) @@ -246,16 +250,21 @@ static void radeon_crtc_destroy(struct drm_crtc *crtc) struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); drm_crtc_cleanup(crtc); + destroy_workqueue(radeon_crtc->flip_queue); kfree(radeon_crtc); } -/* - * Handle unpin events outside the interrupt handler proper. +/** + * radeon_unpin_work_func - unpin old buffer object + * + * @__work - kernel work item + * + * Unpin the old frame buffer object outside of the interrupt handler */ static void radeon_unpin_work_func(struct work_struct *__work) { - struct radeon_unpin_work *work = - container_of(__work, struct radeon_unpin_work, work); + struct radeon_flip_work *work = + container_of(__work, struct radeon_flip_work, unpin_work); int r; /* unpin of the old buffer */ @@ -273,40 +282,35 @@ static void radeon_unpin_work_func(struct work_struct *__work) kfree(work); } -void radeon_crtc_handle_flip(struct radeon_device *rdev, int crtc_id) +void radeon_crtc_handle_vblank(struct radeon_device *rdev, int crtc_id) { struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id]; - struct radeon_unpin_work *work; unsigned long flags; u32 update_pending; int vpos, hpos; + /* can happen during initialization */ + if (radeon_crtc == NULL) + return; + spin_lock_irqsave(&rdev->ddev->event_lock, flags); - work = radeon_crtc->unpin_work; - if (work == NULL || - (work->fence && !radeon_fence_signaled(work->fence))) { + if (radeon_crtc->flip_status != RADEON_FLIP_SUBMITTED) { + DRM_DEBUG_DRIVER("radeon_crtc->flip_status = %d != " + "RADEON_FLIP_SUBMITTED(%d)\n", + radeon_crtc->flip_status, + RADEON_FLIP_SUBMITTED); spin_unlock_irqrestore(&rdev->ddev->event_lock, flags); return; } - /* New pageflip, or just completion of a previous one? */ - if (!radeon_crtc->deferred_flip_completion) { - /* do the flip (mmio) */ - update_pending = radeon_page_flip(rdev, crtc_id, work->new_crtc_base); - } else { - /* This is just a completion of a flip queued in crtc - * at last invocation. Make sure we go directly to - * completion routine. - */ - update_pending = 0; - radeon_crtc->deferred_flip_completion = 0; - } + + update_pending = radeon_page_flip_pending(rdev, crtc_id); /* Has the pageflip already completed in crtc, or is it certain * to complete in this vblank? */ if (update_pending && - (DRM_SCANOUTPOS_VALID & radeon_get_crtc_scanoutpos(rdev->ddev, crtc_id, - &vpos, &hpos)) && + (DRM_SCANOUTPOS_VALID & radeon_get_crtc_scanoutpos(rdev->ddev, crtc_id, 0, + &vpos, &hpos, NULL, NULL)) && ((vpos >= (99 * rdev->mode_info.crtcs[crtc_id]->base.hwmode.crtc_vdisplay)/100) || (vpos < 0 && !ASIC_IS_AVIVO(rdev)))) { /* crtc didn't flip in this target vblank interval, @@ -317,19 +321,43 @@ void radeon_crtc_handle_flip(struct radeon_device *rdev, int crtc_id) */ update_pending = 0; } - if (update_pending) { - /* crtc didn't flip in this target vblank interval, - * but flip is pending in crtc. It will complete it - * in next vblank interval, so complete the flip at - * next vblank irq. - */ - radeon_crtc->deferred_flip_completion = 1; + spin_unlock_irqrestore(&rdev->ddev->event_lock, flags); + if (!update_pending) + radeon_crtc_handle_flip(rdev, crtc_id); +} + +/** + * radeon_crtc_handle_flip - page flip completed + * + * @rdev: radeon device pointer + * @crtc_id: crtc number this event is for + * + * Called when we are sure that a page flip for this crtc is completed. + */ +void radeon_crtc_handle_flip(struct radeon_device *rdev, int crtc_id) +{ + struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id]; + struct radeon_flip_work *work; + unsigned long flags; + + /* this can happen at init */ + if (radeon_crtc == NULL) + return; + + spin_lock_irqsave(&rdev->ddev->event_lock, flags); + work = radeon_crtc->flip_work; + if (radeon_crtc->flip_status != RADEON_FLIP_SUBMITTED) { + DRM_DEBUG_DRIVER("radeon_crtc->flip_status = %d != " + "RADEON_FLIP_SUBMITTED(%d)\n", + radeon_crtc->flip_status, + RADEON_FLIP_SUBMITTED); spin_unlock_irqrestore(&rdev->ddev->event_lock, flags); return; } - /* Pageflip (will be) certainly completed in this vblank. Clean up. */ - radeon_crtc->unpin_work = NULL; + /* Pageflip completed. Clean up. */ + radeon_crtc->flip_status = RADEON_FLIP_NONE; + radeon_crtc->flip_work = NULL; /* wakeup userspace */ if (work->event) @@ -338,14 +366,65 @@ void radeon_crtc_handle_flip(struct radeon_device *rdev, int crtc_id) spin_unlock_irqrestore(&rdev->ddev->event_lock, flags); drm_vblank_put(rdev->ddev, radeon_crtc->crtc_id); - radeon_fence_unref(&work->fence); - radeon_post_page_flip(work->rdev, work->crtc_id); - schedule_work(&work->work); + radeon_irq_kms_pflip_irq_put(rdev, work->crtc_id); + queue_work(radeon_crtc->flip_queue, &work->unpin_work); +} + +/** + * radeon_flip_work_func - page flip framebuffer + * + * @work - kernel work item + * + * Wait for the buffer object to become idle and do the actual page flip + */ +static void radeon_flip_work_func(struct work_struct *__work) +{ + struct radeon_flip_work *work = + container_of(__work, struct radeon_flip_work, flip_work); + struct radeon_device *rdev = work->rdev; + struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[work->crtc_id]; + + struct drm_crtc *crtc = &radeon_crtc->base; + unsigned long flags; + int r; + + down_read(&rdev->exclusive_lock); + if (work->fence) { + r = radeon_fence_wait(work->fence, false); + if (r == -EDEADLK) { + up_read(&rdev->exclusive_lock); + r = radeon_gpu_reset(rdev); + down_read(&rdev->exclusive_lock); + } + if (r) + DRM_ERROR("failed to wait on page flip fence (%d)!\n", r); + + /* We continue with the page flip even if we failed to wait on + * the fence, otherwise the DRM core and userspace will be + * confused about which BO the CRTC is scanning out + */ + + radeon_fence_unref(&work->fence); + } + + /* We borrow the event spin lock for protecting flip_status */ + spin_lock_irqsave(&crtc->dev->event_lock, flags); + + /* set the proper interrupt */ + radeon_irq_kms_pflip_irq_get(rdev, radeon_crtc->crtc_id); + + /* do the flip (mmio) */ + radeon_page_flip(rdev, radeon_crtc->crtc_id, work->base); + + radeon_crtc->flip_status = RADEON_FLIP_SUBMITTED; + spin_unlock_irqrestore(&crtc->dev->event_lock, flags); + up_read(&rdev->exclusive_lock); } static int radeon_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb, - struct drm_pending_vblank_event *event) + struct drm_pending_vblank_event *event, + uint32_t page_flip_flags) { struct drm_device *dev = crtc->dev; struct radeon_device *rdev = dev->dev_private; @@ -353,69 +432,61 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc, struct radeon_framebuffer *old_radeon_fb; struct radeon_framebuffer *new_radeon_fb; struct drm_gem_object *obj; - struct radeon_bo *rbo; - struct radeon_unpin_work *work; + struct radeon_flip_work *work; + struct radeon_bo *new_rbo; + uint32_t tiling_flags, pitch_pixels; + uint64_t base; unsigned long flags; - u32 tiling_flags, pitch_pixels; - u64 base; int r; work = kzalloc(sizeof *work, GFP_KERNEL); if (work == NULL) return -ENOMEM; - work->event = event; + INIT_WORK(&work->flip_work, radeon_flip_work_func); + INIT_WORK(&work->unpin_work, radeon_unpin_work_func); + work->rdev = rdev; work->crtc_id = radeon_crtc->crtc_id; - old_radeon_fb = to_radeon_framebuffer(crtc->fb); - new_radeon_fb = to_radeon_framebuffer(fb); + work->event = event; + /* schedule unpin of the old buffer */ + old_radeon_fb = to_radeon_framebuffer(crtc->primary->fb); obj = old_radeon_fb->obj; + /* take a reference to the old object */ drm_gem_object_reference(obj); - rbo = gem_to_radeon_bo(obj); - work->old_rbo = rbo; - obj = new_radeon_fb->obj; - rbo = gem_to_radeon_bo(obj); + work->old_rbo = gem_to_radeon_bo(obj); - spin_lock(&rbo->tbo.bdev->fence_lock); - if (rbo->tbo.sync_obj) - work->fence = radeon_fence_ref(rbo->tbo.sync_obj); - spin_unlock(&rbo->tbo.bdev->fence_lock); - - INIT_WORK(&work->work, radeon_unpin_work_func); + new_radeon_fb = to_radeon_framebuffer(fb); + obj = new_radeon_fb->obj; + new_rbo = gem_to_radeon_bo(obj); - /* We borrow the event spin lock for protecting unpin_work */ - spin_lock_irqsave(&dev->event_lock, flags); - if (radeon_crtc->unpin_work) { - DRM_DEBUG_DRIVER("flip queue: crtc already busy\n"); - r = -EBUSY; - goto unlock_free; - } - radeon_crtc->unpin_work = work; - radeon_crtc->deferred_flip_completion = 0; - spin_unlock_irqrestore(&dev->event_lock, flags); + spin_lock(&new_rbo->tbo.bdev->fence_lock); + if (new_rbo->tbo.sync_obj) + work->fence = radeon_fence_ref(new_rbo->tbo.sync_obj); + spin_unlock(&new_rbo->tbo.bdev->fence_lock); /* pin the new buffer */ - DRM_DEBUG_DRIVER("flip-ioctl() cur_fbo = %p, cur_bbo = %p\n", - work->old_rbo, rbo); + DRM_DEBUG_DRIVER("flip-ioctl() cur_rbo = %p, new_rbo = %p\n", + work->old_rbo, new_rbo); - r = radeon_bo_reserve(rbo, false); + r = radeon_bo_reserve(new_rbo, false); if (unlikely(r != 0)) { DRM_ERROR("failed to reserve new rbo buffer before flip\n"); - goto pflip_cleanup; + goto cleanup; } /* Only 27 bit offset for legacy CRTC */ - r = radeon_bo_pin_restricted(rbo, RADEON_GEM_DOMAIN_VRAM, + r = radeon_bo_pin_restricted(new_rbo, RADEON_GEM_DOMAIN_VRAM, ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27, &base); if (unlikely(r != 0)) { - radeon_bo_unreserve(rbo); + radeon_bo_unreserve(new_rbo); r = -EINVAL; DRM_ERROR("failed to pin new rbo buffer before flip\n"); - goto pflip_cleanup; + goto cleanup; } - radeon_bo_get_tiling_flags(rbo, &tiling_flags, NULL); - radeon_bo_unreserve(rbo); + radeon_bo_get_tiling_flags(new_rbo, &tiling_flags, NULL); + radeon_bo_unreserve(new_rbo); if (!ASIC_IS_AVIVO(rdev)) { /* crtc offset is from display base addr not FB location */ @@ -452,52 +523,104 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc, } base &= ~7; } + work->base = base; - spin_lock_irqsave(&dev->event_lock, flags); - work->new_crtc_base = base; - spin_unlock_irqrestore(&dev->event_lock, flags); - - /* update crtc fb */ - crtc->fb = fb; - - r = drm_vblank_get(dev, radeon_crtc->crtc_id); + r = drm_vblank_get(crtc->dev, radeon_crtc->crtc_id); if (r) { DRM_ERROR("failed to get vblank before flip\n"); - goto pflip_cleanup1; + goto pflip_cleanup; } - /* set the proper interrupt */ - radeon_pre_page_flip(rdev, radeon_crtc->crtc_id); + /* We borrow the event spin lock for protecting flip_work */ + spin_lock_irqsave(&crtc->dev->event_lock, flags); + + if (radeon_crtc->flip_status != RADEON_FLIP_NONE) { + DRM_DEBUG_DRIVER("flip queue: crtc already busy\n"); + spin_unlock_irqrestore(&crtc->dev->event_lock, flags); + r = -EBUSY; + goto vblank_cleanup; + } + radeon_crtc->flip_status = RADEON_FLIP_PENDING; + radeon_crtc->flip_work = work; + + /* update crtc fb */ + crtc->primary->fb = fb; + + spin_unlock_irqrestore(&crtc->dev->event_lock, flags); + queue_work(radeon_crtc->flip_queue, &work->flip_work); return 0; -pflip_cleanup1: - if (unlikely(radeon_bo_reserve(rbo, false) != 0)) { +vblank_cleanup: + drm_vblank_put(crtc->dev, radeon_crtc->crtc_id); + +pflip_cleanup: + if (unlikely(radeon_bo_reserve(new_rbo, false) != 0)) { DRM_ERROR("failed to reserve new rbo in error path\n"); - goto pflip_cleanup; + goto cleanup; } - if (unlikely(radeon_bo_unpin(rbo) != 0)) { + if (unlikely(radeon_bo_unpin(new_rbo) != 0)) { DRM_ERROR("failed to unpin new rbo in error path\n"); } - radeon_bo_unreserve(rbo); + radeon_bo_unreserve(new_rbo); -pflip_cleanup: - spin_lock_irqsave(&dev->event_lock, flags); - radeon_crtc->unpin_work = NULL; -unlock_free: - spin_unlock_irqrestore(&dev->event_lock, flags); - drm_gem_object_unreference_unlocked(old_radeon_fb->obj); +cleanup: + drm_gem_object_unreference_unlocked(&work->old_rbo->gem_base); radeon_fence_unref(&work->fence); kfree(work); return r; } +static int +radeon_crtc_set_config(struct drm_mode_set *set) +{ + struct drm_device *dev; + struct radeon_device *rdev; + struct drm_crtc *crtc; + bool active = false; + int ret; + + if (!set || !set->crtc) + return -EINVAL; + + dev = set->crtc->dev; + + ret = pm_runtime_get_sync(dev->dev); + if (ret < 0) + return ret; + + ret = drm_crtc_helper_set_config(set); + + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) + if (crtc->enabled) + active = true; + + pm_runtime_mark_last_busy(dev->dev); + + rdev = dev->dev_private; + /* if we have active crtcs and we don't have a power ref, + take the current one */ + if (active && !rdev->have_disp_power_ref) { + rdev->have_disp_power_ref = true; + return ret; + } + /* if we have no active crtcs, then drop the power ref + we got before */ + if (!active && rdev->have_disp_power_ref) { + pm_runtime_put_autosuspend(dev->dev); + rdev->have_disp_power_ref = false; + } + + /* drop the power reference we got coming in here */ + pm_runtime_put_autosuspend(dev->dev); + return ret; +} static const struct drm_crtc_funcs radeon_crtc_funcs = { .cursor_set = radeon_crtc_cursor_set, .cursor_move = radeon_crtc_cursor_move, .gamma_set = radeon_crtc_gamma_set, - .set_config = drm_crtc_helper_set_config, + .set_config = radeon_crtc_set_config, .destroy = radeon_crtc_destroy, .page_flip = radeon_crtc_page_flip, }; @@ -516,6 +639,7 @@ static void radeon_crtc_init(struct drm_device *dev, int index) drm_mode_crtc_set_gamma_size(&radeon_crtc->base, 256); radeon_crtc->crtc_id = index; + radeon_crtc->flip_queue = create_singlethread_workqueue("radeon-crtc"); rdev->mode_info.crtcs[index] = radeon_crtc; if (rdev->family >= CHIP_BONAIRE) { @@ -525,6 +649,8 @@ static void radeon_crtc_init(struct drm_device *dev, int index) radeon_crtc->max_cursor_width = CURSOR_WIDTH; radeon_crtc->max_cursor_height = CURSOR_HEIGHT; } + dev->mode_config.cursor_width = radeon_crtc->max_cursor_width; + dev->mode_config.cursor_height = radeon_crtc->max_cursor_height; #if 0 radeon_crtc->mode_set.crtc = &radeon_crtc->base; @@ -607,7 +733,7 @@ static void radeon_print_display_setup(struct drm_device *dev) list_for_each_entry(connector, &dev->mode_config.connector_list, head) { radeon_connector = to_radeon_connector(connector); DRM_INFO("Connector %d:\n", i); - DRM_INFO(" %s\n", drm_get_connector_name(connector)); + DRM_INFO(" %s\n", connector->name); if (radeon_connector->hpd.hpd != RADEON_HPD_NONE) DRM_INFO(" %s\n", hpd_names[radeon_connector->hpd.hpd]); if (radeon_connector->ddc_bus) { @@ -703,25 +829,28 @@ int radeon_ddc_get_modes(struct radeon_connector *radeon_connector) struct radeon_device *rdev = dev->dev_private; int ret = 0; + /* don't leak the edid if we already fetched it in detect() */ + if (radeon_connector->edid) + goto got_edid; + /* on hw with routers, select right port */ if (radeon_connector->router.ddc_valid) radeon_router_select_ddc_port(radeon_connector); if (radeon_connector_encoder_get_dp_bridge_encoder_id(&radeon_connector->base) != ENCODER_OBJECT_ID_NONE) { - struct radeon_connector_atom_dig *dig = radeon_connector->con_priv; - - if (dig->dp_i2c_bus) + if (radeon_connector->ddc_bus->has_aux) radeon_connector->edid = drm_get_edid(&radeon_connector->base, - &dig->dp_i2c_bus->adapter); + &radeon_connector->ddc_bus->aux.ddc); } else if ((radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_DisplayPort) || (radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_eDP)) { struct radeon_connector_atom_dig *dig = radeon_connector->con_priv; if ((dig->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT || - dig->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) && dig->dp_i2c_bus) + dig->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) && + radeon_connector->ddc_bus->has_aux) radeon_connector->edid = drm_get_edid(&radeon_connector->base, - &dig->dp_i2c_bus->adapter); + &radeon_connector->ddc_bus->aux.ddc); else if (radeon_connector->ddc_bus && !radeon_connector->edid) radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter); @@ -742,8 +871,10 @@ int radeon_ddc_get_modes(struct radeon_connector *radeon_connector) radeon_connector->edid = radeon_bios_get_hardcoded_edid(rdev); } if (radeon_connector->edid) { +got_edid: drm_mode_connector_update_edid_property(&radeon_connector->base, radeon_connector->edid); ret = drm_add_edid_modes(&radeon_connector->base, radeon_connector->edid); + drm_edid_to_eld(&radeon_connector->base, radeon_connector->edid); return ret; } drm_mode_connector_update_edid_property(&radeon_connector->base, NULL); @@ -751,66 +882,89 @@ int radeon_ddc_get_modes(struct radeon_connector *radeon_connector) } /* avivo */ -static void avivo_get_fb_div(struct radeon_pll *pll, - u32 target_clock, - u32 post_div, - u32 ref_div, - u32 *fb_div, - u32 *frac_fb_div) -{ - u32 tmp = post_div * ref_div; - tmp *= target_clock; - *fb_div = tmp / pll->reference_freq; - *frac_fb_div = tmp % pll->reference_freq; +/** + * avivo_reduce_ratio - fractional number reduction + * + * @nom: nominator + * @den: denominator + * @nom_min: minimum value for nominator + * @den_min: minimum value for denominator + * + * Find the greatest common divisor and apply it on both nominator and + * denominator, but make nominator and denominator are at least as large + * as their minimum values. + */ +static void avivo_reduce_ratio(unsigned *nom, unsigned *den, + unsigned nom_min, unsigned den_min) +{ + unsigned tmp; + + /* reduce the numbers to a simpler ratio */ + tmp = gcd(*nom, *den); + *nom /= tmp; + *den /= tmp; + + /* make sure nominator is large enough */ + if (*nom < nom_min) { + tmp = DIV_ROUND_UP(nom_min, *nom); + *nom *= tmp; + *den *= tmp; + } - if (*fb_div > pll->max_feedback_div) - *fb_div = pll->max_feedback_div; - else if (*fb_div < pll->min_feedback_div) - *fb_div = pll->min_feedback_div; + /* make sure the denominator is large enough */ + if (*den < den_min) { + tmp = DIV_ROUND_UP(den_min, *den); + *nom *= tmp; + *den *= tmp; + } } -static u32 avivo_get_post_div(struct radeon_pll *pll, - u32 target_clock) +/** + * avivo_get_fb_ref_div - feedback and ref divider calculation + * + * @nom: nominator + * @den: denominator + * @post_div: post divider + * @fb_div_max: feedback divider maximum + * @ref_div_max: reference divider maximum + * @fb_div: resulting feedback divider + * @ref_div: resulting reference divider + * + * Calculate feedback and reference divider for a given post divider. Makes + * sure we stay within the limits. + */ +static void avivo_get_fb_ref_div(unsigned nom, unsigned den, unsigned post_div, + unsigned fb_div_max, unsigned ref_div_max, + unsigned *fb_div, unsigned *ref_div) { - u32 vco, post_div, tmp; - - if (pll->flags & RADEON_PLL_USE_POST_DIV) - return pll->post_div; - - if (pll->flags & RADEON_PLL_PREFER_MINM_OVER_MAXP) { - if (pll->flags & RADEON_PLL_IS_LCD) - vco = pll->lcd_pll_out_min; - else - vco = pll->pll_out_min; - } else { - if (pll->flags & RADEON_PLL_IS_LCD) - vco = pll->lcd_pll_out_max; - else - vco = pll->pll_out_max; - } + /* limit reference * post divider to a maximum */ + ref_div_max = max(min(100 / post_div, ref_div_max), 1u); - post_div = vco / target_clock; - tmp = vco % target_clock; + /* get matching reference and feedback divider */ + *ref_div = min(max(DIV_ROUND_CLOSEST(den, post_div), 1u), ref_div_max); + *fb_div = DIV_ROUND_CLOSEST(nom * *ref_div * post_div, den); - if (pll->flags & RADEON_PLL_PREFER_MINM_OVER_MAXP) { - if (tmp) - post_div++; - } else { - if (!tmp) - post_div--; + /* limit fb divider to its maximum */ + if (*fb_div > fb_div_max) { + *ref_div = DIV_ROUND_CLOSEST(*ref_div * fb_div_max, *fb_div); + *fb_div = fb_div_max; } - - if (post_div > pll->max_post_div) - post_div = pll->max_post_div; - else if (post_div < pll->min_post_div) - post_div = pll->min_post_div; - - return post_div; } -#define MAX_TOLERANCE 10 - +/** + * radeon_compute_pll_avivo - compute PLL paramaters + * + * @pll: information about the PLL + * @dot_clock_p: resulting pixel clock + * fb_div_p: resulting feedback divider + * frac_fb_div_p: fractional part of the feedback divider + * ref_div_p: resulting reference divider + * post_div_p: resulting reference divider + * + * Try to calculate the PLL parameters to generate the given frequency: + * dot_clock = (ref_freq * feedback_div) / (ref_div * post_div) + */ void radeon_compute_pll_avivo(struct radeon_pll *pll, u32 freq, u32 *dot_clock_p, @@ -819,53 +973,135 @@ void radeon_compute_pll_avivo(struct radeon_pll *pll, u32 *ref_div_p, u32 *post_div_p) { - u32 target_clock = freq / 10; - u32 post_div = avivo_get_post_div(pll, target_clock); - u32 ref_div = pll->min_ref_div; - u32 fb_div = 0, frac_fb_div = 0, tmp; + unsigned target_clock = pll->flags & RADEON_PLL_USE_FRAC_FB_DIV ? + freq : freq / 10; - if (pll->flags & RADEON_PLL_USE_REF_DIV) - ref_div = pll->reference_div; + unsigned fb_div_min, fb_div_max, fb_div; + unsigned post_div_min, post_div_max, post_div; + unsigned ref_div_min, ref_div_max, ref_div; + unsigned post_div_best, diff_best; + unsigned nom, den; + + /* determine allowed feedback divider range */ + fb_div_min = pll->min_feedback_div; + fb_div_max = pll->max_feedback_div; if (pll->flags & RADEON_PLL_USE_FRAC_FB_DIV) { - avivo_get_fb_div(pll, target_clock, post_div, ref_div, &fb_div, &frac_fb_div); - frac_fb_div = (100 * frac_fb_div) / pll->reference_freq; - if (frac_fb_div >= 5) { - frac_fb_div -= 5; - frac_fb_div = frac_fb_div / 10; - frac_fb_div++; + fb_div_min *= 10; + fb_div_max *= 10; + } + + /* determine allowed ref divider range */ + if (pll->flags & RADEON_PLL_USE_REF_DIV) + ref_div_min = pll->reference_div; + else + ref_div_min = pll->min_ref_div; + + if (pll->flags & RADEON_PLL_USE_FRAC_FB_DIV && + pll->flags & RADEON_PLL_USE_REF_DIV) + ref_div_max = pll->reference_div; + else + ref_div_max = pll->max_ref_div; + + /* determine allowed post divider range */ + if (pll->flags & RADEON_PLL_USE_POST_DIV) { + post_div_min = pll->post_div; + post_div_max = pll->post_div; + } else { + unsigned vco_min, vco_max; + + if (pll->flags & RADEON_PLL_IS_LCD) { + vco_min = pll->lcd_pll_out_min; + vco_max = pll->lcd_pll_out_max; + } else { + vco_min = pll->pll_out_min; + vco_max = pll->pll_out_max; } - if (frac_fb_div >= 10) { - fb_div++; - frac_fb_div = 0; + + if (pll->flags & RADEON_PLL_USE_FRAC_FB_DIV) { + vco_min *= 10; + vco_max *= 10; } - } else { - while (ref_div <= pll->max_ref_div) { - avivo_get_fb_div(pll, target_clock, post_div, ref_div, - &fb_div, &frac_fb_div); - if (frac_fb_div >= (pll->reference_freq / 2)) - fb_div++; - frac_fb_div = 0; - tmp = (pll->reference_freq * fb_div) / (post_div * ref_div); - tmp = (tmp * 10000) / target_clock; - - if (tmp > (10000 + MAX_TOLERANCE)) - ref_div++; - else if (tmp >= (10000 - MAX_TOLERANCE)) - break; - else - ref_div++; + + post_div_min = vco_min / target_clock; + if ((target_clock * post_div_min) < vco_min) + ++post_div_min; + if (post_div_min < pll->min_post_div) + post_div_min = pll->min_post_div; + + post_div_max = vco_max / target_clock; + if ((target_clock * post_div_max) > vco_max) + --post_div_max; + if (post_div_max > pll->max_post_div) + post_div_max = pll->max_post_div; + } + + /* represent the searched ratio as fractional number */ + nom = target_clock; + den = pll->reference_freq; + + /* reduce the numbers to a simpler ratio */ + avivo_reduce_ratio(&nom, &den, fb_div_min, post_div_min); + + /* now search for a post divider */ + if (pll->flags & RADEON_PLL_PREFER_MINM_OVER_MAXP) + post_div_best = post_div_min; + else + post_div_best = post_div_max; + diff_best = ~0; + + for (post_div = post_div_min; post_div <= post_div_max; ++post_div) { + unsigned diff; + avivo_get_fb_ref_div(nom, den, post_div, fb_div_max, + ref_div_max, &fb_div, &ref_div); + diff = abs(target_clock - (pll->reference_freq * fb_div) / + (ref_div * post_div)); + + if (diff < diff_best || (diff == diff_best && + !(pll->flags & RADEON_PLL_PREFER_MINM_OVER_MAXP))) { + + post_div_best = post_div; + diff_best = diff; + } + } + post_div = post_div_best; + + /* get the feedback and reference divider for the optimal value */ + avivo_get_fb_ref_div(nom, den, post_div, fb_div_max, ref_div_max, + &fb_div, &ref_div); + + /* reduce the numbers to a simpler ratio once more */ + /* this also makes sure that the reference divider is large enough */ + avivo_reduce_ratio(&fb_div, &ref_div, fb_div_min, ref_div_min); + + /* avoid high jitter with small fractional dividers */ + if (pll->flags & RADEON_PLL_USE_FRAC_FB_DIV && (fb_div % 10)) { + fb_div_min = max(fb_div_min, (9 - (fb_div % 10)) * 20 + 50); + if (fb_div < fb_div_min) { + unsigned tmp = DIV_ROUND_UP(fb_div_min, fb_div); + fb_div *= tmp; + ref_div *= tmp; } } - *dot_clock_p = ((pll->reference_freq * fb_div * 10) + (pll->reference_freq * frac_fb_div)) / - (ref_div * post_div * 10); - *fb_div_p = fb_div; - *frac_fb_div_p = frac_fb_div; + /* and finally save the result */ + if (pll->flags & RADEON_PLL_USE_FRAC_FB_DIV) { + *fb_div_p = fb_div / 10; + *frac_fb_div_p = fb_div % 10; + } else { + *fb_div_p = fb_div; + *frac_fb_div_p = 0; + } + + *dot_clock_p = ((pll->reference_freq * *fb_div_p * 10) + + (pll->reference_freq * *frac_fb_div_p)) / + (ref_div * post_div * 10); *ref_div_p = ref_div; *post_div_p = post_div; - DRM_DEBUG_KMS("%d, pll dividers - fb: %d.%d ref: %d, post %d\n", - *dot_clock_p, fb_div, frac_fb_div, ref_div, post_div); + + DRM_DEBUG_KMS("%d - %d, pll dividers - fb: %d.%d ref: %d, post %d\n", + freq, *dot_clock_p * 10, *fb_div_p, *frac_fb_div_p, + ref_div, post_div); } /* pre-avivo */ @@ -1171,6 +1407,18 @@ static struct drm_prop_enum_list radeon_underscan_enum_list[] = { UNDERSCAN_AUTO, "auto" }, }; +static struct drm_prop_enum_list radeon_audio_enum_list[] = +{ { RADEON_AUDIO_DISABLE, "off" }, + { RADEON_AUDIO_ENABLE, "on" }, + { RADEON_AUDIO_AUTO, "auto" }, +}; + +/* XXX support different dither options? spatial, temporal, both, etc. */ +static struct drm_prop_enum_list radeon_dither_enum_list[] = +{ { RADEON_FMT_DITHER_DISABLE, "off" }, + { RADEON_FMT_DITHER_ENABLE, "on" }, +}; + static int radeon_modeset_create_props(struct radeon_device *rdev) { int sz; @@ -1221,6 +1469,18 @@ static int radeon_modeset_create_props(struct radeon_device *rdev) if (!rdev->mode_info.underscan_vborder_property) return -ENOMEM; + sz = ARRAY_SIZE(radeon_audio_enum_list); + rdev->mode_info.audio_property = + drm_property_create_enum(rdev->ddev, 0, + "audio", + radeon_audio_enum_list, sz); + + sz = ARRAY_SIZE(radeon_dither_enum_list); + rdev->mode_info.dither_property = + drm_property_create_enum(rdev->ddev, 0, + "dither", + radeon_dither_enum_list, sz); + return 0; } @@ -1254,41 +1514,41 @@ static void radeon_afmt_init(struct radeon_device *rdev) for (i = 0; i < RADEON_MAX_AFMT_BLOCKS; i++) rdev->mode_info.afmt[i] = NULL; - if (ASIC_IS_DCE6(rdev)) { - /* todo */ + if (ASIC_IS_NODCE(rdev)) { + /* nothing to do */ } else if (ASIC_IS_DCE4(rdev)) { + static uint32_t eg_offsets[] = { + EVERGREEN_CRTC0_REGISTER_OFFSET, + EVERGREEN_CRTC1_REGISTER_OFFSET, + EVERGREEN_CRTC2_REGISTER_OFFSET, + EVERGREEN_CRTC3_REGISTER_OFFSET, + EVERGREEN_CRTC4_REGISTER_OFFSET, + EVERGREEN_CRTC5_REGISTER_OFFSET, + 0x13830 - 0x7030, + }; + int num_afmt; + + /* DCE8 has 7 audio blocks tied to DIG encoders */ + /* DCE6 has 6 audio blocks tied to DIG encoders */ /* DCE4/5 has 6 audio blocks tied to DIG encoders */ /* DCE4.1 has 2 audio blocks tied to DIG encoders */ - rdev->mode_info.afmt[0] = kzalloc(sizeof(struct radeon_afmt), GFP_KERNEL); - if (rdev->mode_info.afmt[0]) { - rdev->mode_info.afmt[0]->offset = EVERGREEN_CRTC0_REGISTER_OFFSET; - rdev->mode_info.afmt[0]->id = 0; - } - rdev->mode_info.afmt[1] = kzalloc(sizeof(struct radeon_afmt), GFP_KERNEL); - if (rdev->mode_info.afmt[1]) { - rdev->mode_info.afmt[1]->offset = EVERGREEN_CRTC1_REGISTER_OFFSET; - rdev->mode_info.afmt[1]->id = 1; - } - if (!ASIC_IS_DCE41(rdev)) { - rdev->mode_info.afmt[2] = kzalloc(sizeof(struct radeon_afmt), GFP_KERNEL); - if (rdev->mode_info.afmt[2]) { - rdev->mode_info.afmt[2]->offset = EVERGREEN_CRTC2_REGISTER_OFFSET; - rdev->mode_info.afmt[2]->id = 2; - } - rdev->mode_info.afmt[3] = kzalloc(sizeof(struct radeon_afmt), GFP_KERNEL); - if (rdev->mode_info.afmt[3]) { - rdev->mode_info.afmt[3]->offset = EVERGREEN_CRTC3_REGISTER_OFFSET; - rdev->mode_info.afmt[3]->id = 3; - } - rdev->mode_info.afmt[4] = kzalloc(sizeof(struct radeon_afmt), GFP_KERNEL); - if (rdev->mode_info.afmt[4]) { - rdev->mode_info.afmt[4]->offset = EVERGREEN_CRTC4_REGISTER_OFFSET; - rdev->mode_info.afmt[4]->id = 4; - } - rdev->mode_info.afmt[5] = kzalloc(sizeof(struct radeon_afmt), GFP_KERNEL); - if (rdev->mode_info.afmt[5]) { - rdev->mode_info.afmt[5]->offset = EVERGREEN_CRTC5_REGISTER_OFFSET; - rdev->mode_info.afmt[5]->id = 5; + if (ASIC_IS_DCE8(rdev)) + num_afmt = 7; + else if (ASIC_IS_DCE6(rdev)) + num_afmt = 6; + else if (ASIC_IS_DCE5(rdev)) + num_afmt = 6; + else if (ASIC_IS_DCE41(rdev)) + num_afmt = 2; + else /* DCE4 */ + num_afmt = 6; + + BUG_ON(num_afmt > ARRAY_SIZE(eg_offsets)); + for (i = 0; i < num_afmt; i++) { + rdev->mode_info.afmt[i] = kzalloc(sizeof(struct radeon_afmt), GFP_KERNEL); + if (rdev->mode_info.afmt[i]) { + rdev->mode_info.afmt[i]->offset = eg_offsets[i]; + rdev->mode_info.afmt[i]->id = i; } } } else if (ASIC_IS_DCE3(rdev)) { @@ -1394,12 +1654,22 @@ int radeon_modeset_init(struct radeon_device *rdev) /* setup afmt */ radeon_afmt_init(rdev); - /* Initialize power management */ - radeon_pm_init(rdev); - radeon_fbdev_init(rdev); drm_kms_helper_poll_init(rdev->ddev); + if (rdev->pm.dpm_enabled) { + /* do dpm late init */ + ret = radeon_pm_late_init(rdev); + if (ret) { + rdev->pm.dpm_enabled = false; + DRM_ERROR("radeon_pm_late_init failed, disabling dpm\n"); + } + /* set the dpm state for PX since there won't be + * a modeset to call this. + */ + radeon_pm_compute_clocks(rdev); + } + return 0; } @@ -1407,7 +1677,6 @@ void radeon_modeset_fini(struct radeon_device *rdev) { radeon_fbdev_fini(rdev); kfree(rdev->mode_info.bios_hardcoded_edid); - radeon_pm_fini(rdev); if (rdev->mode_info.mode_config_initialized) { radeon_afmt_fini(rdev); @@ -1526,12 +1795,18 @@ bool radeon_crtc_scaling_mode_fixup(struct drm_crtc *crtc, } /* - * Retrieve current video scanout position of crtc on a given gpu. + * Retrieve current video scanout position of crtc on a given gpu, and + * an optional accurate timestamp of when query happened. * * \param dev Device to query. * \param crtc Crtc to query. + * \param flags Flags from caller (DRM_CALLED_FROM_VBLIRQ or 0). * \param *vpos Location where vertical scanout position should be stored. * \param *hpos Location where horizontal scanout position should go. + * \param *stime Target location for timestamp taken immediately before + * scanout position query. Can be NULL to skip timestamp. + * \param *etime Target location for timestamp taken immediately after + * scanout position query. Can be NULL to skip timestamp. * * Returns vpos as a positive number while in active scanout area. * Returns vpos as a negative number inside vblank, counting the number @@ -1547,7 +1822,8 @@ bool radeon_crtc_scaling_mode_fixup(struct drm_crtc *crtc, * unknown small number of scanlines wrt. real scanout position. * */ -int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, int *vpos, int *hpos) +int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, unsigned int flags, + int *vpos, int *hpos, ktime_t *stime, ktime_t *etime) { u32 stat_crtc = 0, vbl = 0, position = 0; int vbl_start, vbl_end, vtotal, ret = 0; @@ -1555,6 +1831,12 @@ int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, int *vpos, int struct radeon_device *rdev = dev->dev_private; + /* preempt_disable_rt() should go right here in PREEMPT_RT patchset. */ + + /* Get optional system timestamp before query. */ + if (stime) + *stime = ktime_get(); + if (ASIC_IS_DCE4(rdev)) { if (crtc == 0) { vbl = RREG32(EVERGREEN_CRTC_V_BLANK_START_END + @@ -1637,6 +1919,12 @@ int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, int *vpos, int } } + /* Get optional system timestamp after query. */ + if (etime) + *etime = ktime_get(); + + /* preempt_enable_rt() should go right here in PREEMPT_RT patchset. */ + /* Decode into vertical and horizontal scanout position. */ *vpos = position & 0x1fff; *hpos = (position >> 16) & 0x1fff; @@ -1677,5 +1965,27 @@ int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, int *vpos, int if (in_vbl) ret |= DRM_SCANOUTPOS_INVBL; + /* Is vpos outside nominal vblank area, but less than + * 1/100 of a frame height away from start of vblank? + * If so, assume this isn't a massively delayed vblank + * interrupt, but a vblank interrupt that fired a few + * microseconds before true start of vblank. Compensate + * by adding a full frame duration to the final timestamp. + * Happens, e.g., on ATI R500, R600. + * + * We only do this if DRM_CALLED_FROM_VBLIRQ. + */ + if ((flags & DRM_CALLED_FROM_VBLIRQ) && !in_vbl) { + vbl_start = rdev->mode_info.crtcs[crtc]->base.hwmode.crtc_vdisplay; + vtotal = rdev->mode_info.crtcs[crtc]->base.hwmode.crtc_vtotal; + + if (vbl_start - *vpos < vtotal / 100) { + *vpos -= vtotal; + + /* Signal this correction as "applied". */ + ret |= 0x8; + } + } + return ret; } diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index e5419b35017..e9e36108424 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -36,8 +36,9 @@ #include <drm/drm_pciids.h> #include <linux/console.h> #include <linux/module.h> - - +#include <linux/pm_runtime.h> +#include <linux/vga_switcheroo.h> +#include "drm_crtc_helper.h" /* * KMS wrapper. * - 2.0.0 - initial interface @@ -75,21 +76,26 @@ * 2.32.0 - new info request for rings working * 2.33.0 - Add SI tiling mode array query * 2.34.0 - Add CIK tiling mode array query + * 2.35.0 - Add CIK macrotile mode array query + * 2.36.0 - Fix CIK DCE tiling setup + * 2.37.0 - allow GS ring setup on r6xx/r7xx + * 2.38.0 - RADEON_GEM_OP (GET_INITIAL_DOMAIN, SET_INITIAL_DOMAIN), + * CIK: 1D and linear tiling modes contain valid PIPE_CONFIG + * 2.39.0 - Add INFO query for number of active CUs */ #define KMS_DRIVER_MAJOR 2 -#define KMS_DRIVER_MINOR 34 +#define KMS_DRIVER_MINOR 39 #define KMS_DRIVER_PATCHLEVEL 0 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags); int radeon_driver_unload_kms(struct drm_device *dev); -int radeon_driver_firstopen_kms(struct drm_device *dev); void radeon_driver_lastclose_kms(struct drm_device *dev); int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv); void radeon_driver_postclose_kms(struct drm_device *dev, struct drm_file *file_priv); void radeon_driver_preclose_kms(struct drm_device *dev, struct drm_file *file_priv); -int radeon_suspend_kms(struct drm_device *dev, pm_message_t state); -int radeon_resume_kms(struct drm_device *dev); +int radeon_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon); +int radeon_resume_kms(struct drm_device *dev, bool resume, bool fbcon); u32 radeon_get_vblank_counter_kms(struct drm_device *dev, int crtc); int radeon_enable_vblank_kms(struct drm_device *dev, int crtc); void radeon_disable_vblank_kms(struct drm_device *dev, int crtc); @@ -100,18 +106,18 @@ int radeon_get_vblank_timestamp_kms(struct drm_device *dev, int crtc, void radeon_driver_irq_preinstall_kms(struct drm_device *dev); int radeon_driver_irq_postinstall_kms(struct drm_device *dev); void radeon_driver_irq_uninstall_kms(struct drm_device *dev); -irqreturn_t radeon_driver_irq_handler_kms(DRM_IRQ_ARGS); -int radeon_dma_ioctl_kms(struct drm_device *dev, void *data, - struct drm_file *file_priv); -int radeon_gem_object_init(struct drm_gem_object *obj); +irqreturn_t radeon_driver_irq_handler_kms(int irq, void *arg); void radeon_gem_object_free(struct drm_gem_object *obj); int radeon_gem_object_open(struct drm_gem_object *obj, struct drm_file *file_priv); void radeon_gem_object_close(struct drm_gem_object *obj, struct drm_file *file_priv); extern int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, - int *vpos, int *hpos); -extern struct drm_ioctl_desc radeon_ioctls_kms[]; + unsigned int flags, + int *vpos, int *hpos, ktime_t *stime, + ktime_t *etime); +extern bool radeon_is_px(struct drm_device *dev); +extern const struct drm_ioctl_desc radeon_ioctls_kms[]; extern int radeon_max_kms_ioctl; int radeon_mmap(struct file *filp, struct vm_area_struct *vma); int radeon_mode_dumb_mmap(struct drm_file *filp, @@ -120,9 +126,6 @@ int radeon_mode_dumb_mmap(struct drm_file *filp, int radeon_mode_dumb_create(struct drm_file *file_priv, struct drm_device *dev, struct drm_mode_create_dumb *args); -int radeon_mode_dumb_destroy(struct drm_file *file_priv, - struct drm_device *dev, - uint32_t handle); struct sg_table *radeon_gem_prime_get_sg_table(struct drm_gem_object *obj); struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev, size_t size, @@ -154,12 +157,12 @@ int radeon_dynclks = -1; int radeon_r4xx_atom = 0; int radeon_agpmode = 0; int radeon_vram_limit = 0; -int radeon_gart_size = 512; /* default gart size */ +int radeon_gart_size = -1; /* auto */ int radeon_benchmarking = 0; int radeon_testing = 0; int radeon_connector_table = 0; int radeon_tv = 1; -int radeon_audio = 0; +int radeon_audio = -1; int radeon_disp_priority = 0; int radeon_hw_i2c = 0; int radeon_pcie_gen2 = -1; @@ -167,6 +170,12 @@ int radeon_msi = -1; int radeon_lockup_timeout = 10000; int radeon_fastfb = 0; int radeon_dpm = -1; +int radeon_aspm = -1; +int radeon_runtime_pm = -1; +int radeon_hard_reset = 0; +int radeon_vm_size = 4; +int radeon_vm_block_size = 9; +int radeon_deep_color = 0; MODULE_PARM_DESC(no_wb, "Disable AGP writeback for scratch registers"); module_param_named(no_wb, radeon_no_wb, int, 0444); @@ -180,13 +189,13 @@ module_param_named(dynclks, radeon_dynclks, int, 0444); MODULE_PARM_DESC(r4xx_atom, "Enable ATOMBIOS modesetting for R4xx"); module_param_named(r4xx_atom, radeon_r4xx_atom, int, 0444); -MODULE_PARM_DESC(vramlimit, "Restrict VRAM for testing"); +MODULE_PARM_DESC(vramlimit, "Restrict VRAM for testing, in megabytes"); module_param_named(vramlimit, radeon_vram_limit, int, 0600); MODULE_PARM_DESC(agpmode, "AGP Mode (-1 == PCI)"); module_param_named(agpmode, radeon_agpmode, int, 0444); -MODULE_PARM_DESC(gartsize, "Size of PCIE/IGP gart to setup in megabytes (32, 64, etc)"); +MODULE_PARM_DESC(gartsize, "Size of PCIE/IGP gart to setup in megabytes (32, 64, etc., -1 = auto)"); module_param_named(gartsize, radeon_gart_size, int, 0600); MODULE_PARM_DESC(benchmark, "Run benchmark"); @@ -201,7 +210,7 @@ module_param_named(connector_table, radeon_connector_table, int, 0444); MODULE_PARM_DESC(tv, "TV enable (0 = disable)"); module_param_named(tv, radeon_tv, int, 0444); -MODULE_PARM_DESC(audio, "Audio enable (1 = enable)"); +MODULE_PARM_DESC(audio, "Audio enable (-1 = auto, 0 = disable, 1 = enable)"); module_param_named(audio, radeon_audio, int, 0444); MODULE_PARM_DESC(disp_priority, "Display Priority (0 = auto, 1 = normal, 2 = high)"); @@ -225,6 +234,24 @@ module_param_named(fastfb, radeon_fastfb, int, 0444); MODULE_PARM_DESC(dpm, "DPM support (1 = enable, 0 = disable, -1 = auto)"); module_param_named(dpm, radeon_dpm, int, 0444); +MODULE_PARM_DESC(aspm, "ASPM support (1 = enable, 0 = disable, -1 = auto)"); +module_param_named(aspm, radeon_aspm, int, 0444); + +MODULE_PARM_DESC(runpm, "PX runtime pm (1 = force enable, 0 = disable, -1 = PX only default)"); +module_param_named(runpm, radeon_runtime_pm, int, 0444); + +MODULE_PARM_DESC(hard_reset, "PCI config reset (1 = force enable, 0 = disable (default))"); +module_param_named(hard_reset, radeon_hard_reset, int, 0444); + +MODULE_PARM_DESC(vm_size, "VM address space size in gigabytes (default 4GB)"); +module_param_named(vm_size, radeon_vm_size, int, 0444); + +MODULE_PARM_DESC(vm_block_size, "VM page table size in bits (default 9)"); +module_param_named(vm_block_size, radeon_vm_block_size, int, 0444); + +MODULE_PARM_DESC(deep_color, "Deep Color support (1 = enable, 0 = disable (default))"); +module_param_named(deep_color, radeon_deep_color, int, 0444); + static struct pci_device_id pciidlist[] = { radeon_PCI_IDS }; @@ -261,6 +288,7 @@ static int radeon_resume(struct drm_device *dev) return 0; } + static const struct file_operations radeon_driver_old_fops = { .owner = THIS_MODULE, .open = drm_open, @@ -268,7 +296,6 @@ static const struct file_operations radeon_driver_old_fops = { .unlocked_ioctl = drm_ioctl, .mmap = drm_mmap, .poll = drm_poll, - .fasync = drm_fasync, .read = drm_read, #ifdef CONFIG_COMPAT .compat_ioctl = radeon_compat_ioctl, @@ -278,7 +305,7 @@ static const struct file_operations radeon_driver_old_fops = { static struct drm_driver driver_old = { .driver_features = - DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_PCI_DMA | DRIVER_SG | + DRIVER_USE_AGP | DRIVER_PCI_DMA | DRIVER_SG | DRIVER_HAVE_IRQ | DRIVER_HAVE_DMA | DRIVER_IRQ_SHARED, .dev_priv_size = sizeof(drm_radeon_buf_priv_t), .load = radeon_driver_load, @@ -356,28 +383,144 @@ radeon_pci_remove(struct pci_dev *pdev) drm_put_dev(dev); } -static int -radeon_pci_suspend(struct pci_dev *pdev, pm_message_t state) +static int radeon_pmops_suspend(struct device *dev) { - struct drm_device *dev = pci_get_drvdata(pdev); - return radeon_suspend_kms(dev, state); + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + return radeon_suspend_kms(drm_dev, true, true); } -static int -radeon_pci_resume(struct pci_dev *pdev) +static int radeon_pmops_resume(struct device *dev) { - struct drm_device *dev = pci_get_drvdata(pdev); - return radeon_resume_kms(dev); + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + return radeon_resume_kms(drm_dev, true, true); +} + +static int radeon_pmops_freeze(struct device *dev) +{ + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + return radeon_suspend_kms(drm_dev, false, true); } +static int radeon_pmops_thaw(struct device *dev) +{ + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + return radeon_resume_kms(drm_dev, false, true); +} + +static int radeon_pmops_runtime_suspend(struct device *dev) +{ + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + int ret; + + if (!radeon_is_px(drm_dev)) { + pm_runtime_forbid(dev); + return -EBUSY; + } + + drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING; + drm_kms_helper_poll_disable(drm_dev); + vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_OFF); + + ret = radeon_suspend_kms(drm_dev, false, false); + pci_save_state(pdev); + pci_disable_device(pdev); + pci_set_power_state(pdev, PCI_D3cold); + drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF; + + return 0; +} + +static int radeon_pmops_runtime_resume(struct device *dev) +{ + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + int ret; + + if (!radeon_is_px(drm_dev)) + return -EINVAL; + + drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING; + + pci_set_power_state(pdev, PCI_D0); + pci_restore_state(pdev); + ret = pci_enable_device(pdev); + if (ret) + return ret; + pci_set_master(pdev); + + ret = radeon_resume_kms(drm_dev, false, false); + drm_kms_helper_poll_enable(drm_dev); + vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_ON); + drm_dev->switch_power_state = DRM_SWITCH_POWER_ON; + return 0; +} + +static int radeon_pmops_runtime_idle(struct device *dev) +{ + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + struct drm_crtc *crtc; + + if (!radeon_is_px(drm_dev)) { + pm_runtime_forbid(dev); + return -EBUSY; + } + + list_for_each_entry(crtc, &drm_dev->mode_config.crtc_list, head) { + if (crtc->enabled) { + DRM_DEBUG_DRIVER("failing to power off - crtc active\n"); + return -EBUSY; + } + } + + pm_runtime_mark_last_busy(dev); + pm_runtime_autosuspend(dev); + /* we don't want the main rpm_idle to call suspend - we want to autosuspend */ + return 1; +} + +long radeon_drm_ioctl(struct file *filp, + unsigned int cmd, unsigned long arg) +{ + struct drm_file *file_priv = filp->private_data; + struct drm_device *dev; + long ret; + dev = file_priv->minor->dev; + ret = pm_runtime_get_sync(dev->dev); + if (ret < 0) + return ret; + + ret = drm_ioctl(filp, cmd, arg); + + pm_runtime_mark_last_busy(dev->dev); + pm_runtime_put_autosuspend(dev->dev); + return ret; +} + +static const struct dev_pm_ops radeon_pm_ops = { + .suspend = radeon_pmops_suspend, + .resume = radeon_pmops_resume, + .freeze = radeon_pmops_freeze, + .thaw = radeon_pmops_thaw, + .poweroff = radeon_pmops_freeze, + .restore = radeon_pmops_resume, + .runtime_suspend = radeon_pmops_runtime_suspend, + .runtime_resume = radeon_pmops_runtime_resume, + .runtime_idle = radeon_pmops_runtime_idle, +}; + static const struct file_operations radeon_driver_kms_fops = { .owner = THIS_MODULE, .open = drm_open, .release = drm_release, - .unlocked_ioctl = drm_ioctl, + .unlocked_ioctl = radeon_drm_ioctl, .mmap = radeon_mmap, .poll = drm_poll, - .fasync = drm_fasync, .read = drm_read, #ifdef CONFIG_COMPAT .compat_ioctl = radeon_kms_compat_ioctl, @@ -386,19 +529,15 @@ static const struct file_operations radeon_driver_kms_fops = { static struct drm_driver kms_driver = { .driver_features = - DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_PCI_DMA | DRIVER_SG | - DRIVER_HAVE_IRQ | DRIVER_HAVE_DMA | DRIVER_IRQ_SHARED | DRIVER_GEM | - DRIVER_PRIME, - .dev_priv_size = 0, + DRIVER_USE_AGP | + DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED | DRIVER_GEM | + DRIVER_PRIME | DRIVER_RENDER, .load = radeon_driver_load_kms, - .firstopen = radeon_driver_firstopen_kms, .open = radeon_driver_open_kms, .preclose = radeon_driver_preclose_kms, .postclose = radeon_driver_postclose_kms, .lastclose = radeon_driver_lastclose_kms, .unload = radeon_driver_unload_kms, - .suspend = radeon_suspend_kms, - .resume = radeon_resume_kms, .get_vblank_counter = radeon_get_vblank_counter_kms, .enable_vblank = radeon_enable_vblank_kms, .disable_vblank = radeon_disable_vblank_kms, @@ -413,14 +552,12 @@ static struct drm_driver kms_driver = { .irq_uninstall = radeon_driver_irq_uninstall_kms, .irq_handler = radeon_driver_irq_handler_kms, .ioctls = radeon_ioctls_kms, - .gem_init_object = radeon_gem_object_init, .gem_free_object = radeon_gem_object_free, .gem_open_object = radeon_gem_object_open, .gem_close_object = radeon_gem_object_close, - .dma_ioctl = radeon_dma_ioctl_kms, .dumb_create = radeon_mode_dumb_create, .dumb_map_offset = radeon_mode_dumb_mmap, - .dumb_destroy = radeon_mode_dumb_destroy, + .dumb_destroy = drm_gem_dumb_destroy, .fops = &radeon_driver_kms_fops, .prime_handle_to_fd = drm_gem_prime_handle_to_fd, @@ -457,8 +594,7 @@ static struct pci_driver radeon_kms_pci_driver = { .id_table = pciidlist, .probe = radeon_pci_probe, .remove = radeon_pci_remove, - .suspend = radeon_pci_suspend, - .resume = radeon_pci_resume, + .driver.pm = &radeon_pm_ops, }; static int __init radeon_init(void) diff --git a/drivers/gpu/drm/radeon/radeon_drv.h b/drivers/gpu/drm/radeon/radeon_drv.h index b369d42f7de..dafd812e457 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.h +++ b/drivers/gpu/drm/radeon/radeon_drv.h @@ -108,11 +108,15 @@ * 1.31- Add support for num Z pipes from GET_PARAM * 1.32- fixes for rv740 setup * 1.33- Add r6xx/r7xx const buffer support + * 1.34- fix evergreen/cayman GS register */ #define DRIVER_MAJOR 1 -#define DRIVER_MINOR 33 +#define DRIVER_MINOR 34 #define DRIVER_PATCHLEVEL 0 +long radeon_drm_ioctl(struct file *filp, + unsigned int cmd, unsigned long arg); + /* The rest of the file is DEPRECATED! */ #ifdef CONFIG_DRM_RADEON_UMS @@ -401,7 +405,7 @@ extern void radeon_do_release(struct drm_device * dev); extern u32 radeon_get_vblank_counter(struct drm_device *dev, int crtc); extern int radeon_enable_vblank(struct drm_device *dev, int crtc); extern void radeon_disable_vblank(struct drm_device *dev, int crtc); -extern irqreturn_t radeon_driver_irq_handler(DRM_IRQ_ARGS); +extern irqreturn_t radeon_driver_irq_handler(int irq, void *arg); extern void radeon_driver_irq_preinstall(struct drm_device * dev); extern int radeon_driver_irq_postinstall(struct drm_device *dev); extern void radeon_driver_irq_uninstall(struct drm_device * dev); diff --git a/drivers/gpu/drm/radeon/radeon_family.h b/drivers/gpu/drm/radeon/radeon_family.h index 3c8289083f9..4b7b87f71a6 100644 --- a/drivers/gpu/drm/radeon/radeon_family.h +++ b/drivers/gpu/drm/radeon/radeon_family.h @@ -96,6 +96,8 @@ enum radeon_family { CHIP_BONAIRE, CHIP_KAVERI, CHIP_KABINI, + CHIP_HAWAII, + CHIP_MULLINS, CHIP_LAST, }; @@ -114,6 +116,7 @@ enum radeon_chip_flags { RADEON_NEW_MEMMAP = 0x00400000UL, RADEON_IS_PCI = 0x00800000UL, RADEON_IS_IGPGART = 0x01000000UL, + RADEON_IS_PX = 0x02000000UL, }; #endif diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c index ddb8f8e04eb..913787085df 100644 --- a/drivers/gpu/drm/radeon/radeon_fence.c +++ b/drivers/gpu/drm/radeon/radeon_fence.c @@ -121,7 +121,7 @@ int radeon_fence_emit(struct radeon_device *rdev, (*fence)->seq = ++rdev->fence_drv[ring].sync_seq[ring]; (*fence)->ring = ring; radeon_fence_ring_emit(rdev, ring, *fence); - trace_radeon_fence_emit(rdev->ddev, (*fence)->seq); + trace_radeon_fence_emit(rdev->ddev, ring, (*fence)->seq); return 0; } @@ -190,10 +190,8 @@ void radeon_fence_process(struct radeon_device *rdev, int ring) } } while (atomic64_xchg(&rdev->fence_drv[ring].last_seq, seq) > seq); - if (wake) { - rdev->fence_drv[ring].last_activity = jiffies; + if (wake) wake_up_all(&rdev->fence_queue); - } } /** @@ -212,13 +210,13 @@ static void radeon_fence_destroy(struct kref *kref) } /** - * radeon_fence_seq_signaled - check if a fence sequeuce number has signaled + * radeon_fence_seq_signaled - check if a fence sequence number has signaled * * @rdev: radeon device pointer * @seq: sequence number * @ring: ring index the fence is associated with * - * Check if the last singled fence sequnce number is >= the requested + * Check if the last signaled fence sequnce number is >= the requested * sequence number (all asics). * Returns true if the fence has signaled (current fence value * is >= requested value) or false if it has not (current fence @@ -263,113 +261,122 @@ bool radeon_fence_signaled(struct radeon_fence *fence) } /** - * radeon_fence_wait_seq - wait for a specific sequence number + * radeon_fence_any_seq_signaled - check if any sequence number is signaled * * @rdev: radeon device pointer - * @target_seq: sequence number we want to wait for - * @ring: ring index the fence is associated with + * @seq: sequence numbers + * + * Check if the last signaled fence sequnce number is >= the requested + * sequence number (all asics). + * Returns true if any has signaled (current value is >= requested value) + * or false if it has not. Helper function for radeon_fence_wait_seq. + */ +static bool radeon_fence_any_seq_signaled(struct radeon_device *rdev, u64 *seq) +{ + unsigned i; + + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + if (seq[i] && radeon_fence_seq_signaled(rdev, seq[i], i)) + return true; + } + return false; +} + +/** + * radeon_fence_wait_seq - wait for a specific sequence numbers + * + * @rdev: radeon device pointer + * @target_seq: sequence number(s) we want to wait for * @intr: use interruptable sleep - * @lock_ring: whether the ring should be locked or not * - * Wait for the requested sequence number to be written (all asics). + * Wait for the requested sequence number(s) to be written by any ring + * (all asics). Sequnce number array is indexed by ring id. * @intr selects whether to use interruptable (true) or non-interruptable * (false) sleep when waiting for the sequence number. Helper function - * for radeon_fence_wait(), et al. + * for radeon_fence_wait_*(). * Returns 0 if the sequence number has passed, error for all other cases. - * -EDEADLK is returned when a GPU lockup has been detected and the ring is - * marked as not ready so no further jobs get scheduled until a successful - * reset. + * -EDEADLK is returned when a GPU lockup has been detected. */ -static int radeon_fence_wait_seq(struct radeon_device *rdev, u64 target_seq, - unsigned ring, bool intr, bool lock_ring) +static int radeon_fence_wait_seq(struct radeon_device *rdev, u64 *target_seq, + bool intr) { - unsigned long timeout, last_activity; - uint64_t seq; - unsigned i; + uint64_t last_seq[RADEON_NUM_RINGS]; bool signaled; - int r; + int i, r; + + while (!radeon_fence_any_seq_signaled(rdev, target_seq)) { - while (target_seq > atomic64_read(&rdev->fence_drv[ring].last_seq)) { - if (!rdev->ring[ring].ready) { - return -EBUSY; + /* Save current sequence values, used to check for GPU lockups */ + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + if (!target_seq[i]) + continue; + + last_seq[i] = atomic64_read(&rdev->fence_drv[i].last_seq); + trace_radeon_fence_wait_begin(rdev->ddev, i, target_seq[i]); + radeon_irq_kms_sw_irq_get(rdev, i); } - timeout = jiffies - RADEON_FENCE_JIFFIES_TIMEOUT; - if (time_after(rdev->fence_drv[ring].last_activity, timeout)) { - /* the normal case, timeout is somewhere before last_activity */ - timeout = rdev->fence_drv[ring].last_activity - timeout; + if (intr) { + r = wait_event_interruptible_timeout(rdev->fence_queue, ( + (signaled = radeon_fence_any_seq_signaled(rdev, target_seq)) + || rdev->needs_reset), RADEON_FENCE_JIFFIES_TIMEOUT); } else { - /* either jiffies wrapped around, or no fence was signaled in the last 500ms - * anyway we will just wait for the minimum amount and then check for a lockup - */ - timeout = 1; + r = wait_event_timeout(rdev->fence_queue, ( + (signaled = radeon_fence_any_seq_signaled(rdev, target_seq)) + || rdev->needs_reset), RADEON_FENCE_JIFFIES_TIMEOUT); } - seq = atomic64_read(&rdev->fence_drv[ring].last_seq); - /* Save current last activity valuee, used to check for GPU lockups */ - last_activity = rdev->fence_drv[ring].last_activity; - trace_radeon_fence_wait_begin(rdev->ddev, seq); - radeon_irq_kms_sw_irq_get(rdev, ring); - if (intr) { - r = wait_event_interruptible_timeout(rdev->fence_queue, - (signaled = radeon_fence_seq_signaled(rdev, target_seq, ring)), - timeout); - } else { - r = wait_event_timeout(rdev->fence_queue, - (signaled = radeon_fence_seq_signaled(rdev, target_seq, ring)), - timeout); + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + if (!target_seq[i]) + continue; + + radeon_irq_kms_sw_irq_put(rdev, i); + trace_radeon_fence_wait_end(rdev->ddev, i, target_seq[i]); } - radeon_irq_kms_sw_irq_put(rdev, ring); - if (unlikely(r < 0)) { + + if (unlikely(r < 0)) return r; - } - trace_radeon_fence_wait_end(rdev->ddev, seq); if (unlikely(!signaled)) { + if (rdev->needs_reset) + return -EDEADLK; + /* we were interrupted for some reason and fence * isn't signaled yet, resume waiting */ - if (r) { + if (r) continue; - } - /* check if sequence value has changed since last_activity */ - if (seq != atomic64_read(&rdev->fence_drv[ring].last_seq)) { - continue; - } + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + if (!target_seq[i]) + continue; - if (lock_ring) { - mutex_lock(&rdev->ring_lock); + if (last_seq[i] != atomic64_read(&rdev->fence_drv[i].last_seq)) + break; } - /* test if somebody else has already decided that this is a lockup */ - if (last_activity != rdev->fence_drv[ring].last_activity) { - if (lock_ring) { - mutex_unlock(&rdev->ring_lock); - } + if (i != RADEON_NUM_RINGS) continue; + + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + if (!target_seq[i]) + continue; + + if (radeon_ring_is_lockup(rdev, i, &rdev->ring[i])) + break; } - if (radeon_ring_is_lockup(rdev, ring, &rdev->ring[ring])) { + if (i < RADEON_NUM_RINGS) { /* good news we believe it's a lockup */ - dev_warn(rdev->dev, "GPU lockup (waiting for 0x%016llx last fence id 0x%016llx)\n", - target_seq, seq); - - /* change last activity so nobody else think there is a lockup */ - for (i = 0; i < RADEON_NUM_RINGS; ++i) { - rdev->fence_drv[i].last_activity = jiffies; - } - - /* mark the ring as not ready any more */ - rdev->ring[ring].ready = false; - if (lock_ring) { - mutex_unlock(&rdev->ring_lock); - } + dev_warn(rdev->dev, "GPU lockup (waiting for " + "0x%016llx last fence id 0x%016llx on" + " ring %d)\n", + target_seq[i], last_seq[i], i); + + /* remember that we need an reset */ + rdev->needs_reset = true; + wake_up_all(&rdev->fence_queue); return -EDEADLK; } - - if (lock_ring) { - mutex_unlock(&rdev->ring_lock); - } } } return 0; @@ -388,6 +395,7 @@ static int radeon_fence_wait_seq(struct radeon_device *rdev, u64 target_seq, */ int radeon_fence_wait(struct radeon_fence *fence, bool intr) { + uint64_t seq[RADEON_NUM_RINGS] = {}; int r; if (fence == NULL) { @@ -395,147 +403,15 @@ int radeon_fence_wait(struct radeon_fence *fence, bool intr) return -EINVAL; } - r = radeon_fence_wait_seq(fence->rdev, fence->seq, - fence->ring, intr, true); - if (r) { - return r; - } - fence->seq = RADEON_FENCE_SIGNALED_SEQ; - return 0; -} - -static bool radeon_fence_any_seq_signaled(struct radeon_device *rdev, u64 *seq) -{ - unsigned i; - - for (i = 0; i < RADEON_NUM_RINGS; ++i) { - if (seq[i] && radeon_fence_seq_signaled(rdev, seq[i], i)) { - return true; - } - } - return false; -} - -/** - * radeon_fence_wait_any_seq - wait for a sequence number on any ring - * - * @rdev: radeon device pointer - * @target_seq: sequence number(s) we want to wait for - * @intr: use interruptable sleep - * - * Wait for the requested sequence number(s) to be written by any ring - * (all asics). Sequnce number array is indexed by ring id. - * @intr selects whether to use interruptable (true) or non-interruptable - * (false) sleep when waiting for the sequence number. Helper function - * for radeon_fence_wait_any(), et al. - * Returns 0 if the sequence number has passed, error for all other cases. - */ -static int radeon_fence_wait_any_seq(struct radeon_device *rdev, - u64 *target_seq, bool intr) -{ - unsigned long timeout, last_activity, tmp; - unsigned i, ring = RADEON_NUM_RINGS; - bool signaled; - int r; - - for (i = 0, last_activity = 0; i < RADEON_NUM_RINGS; ++i) { - if (!target_seq[i]) { - continue; - } - - /* use the most recent one as indicator */ - if (time_after(rdev->fence_drv[i].last_activity, last_activity)) { - last_activity = rdev->fence_drv[i].last_activity; - } - - /* For lockup detection just pick the lowest ring we are - * actively waiting for - */ - if (i < ring) { - ring = i; - } - } - - /* nothing to wait for ? */ - if (ring == RADEON_NUM_RINGS) { - return -ENOENT; - } - - while (!radeon_fence_any_seq_signaled(rdev, target_seq)) { - timeout = jiffies - RADEON_FENCE_JIFFIES_TIMEOUT; - if (time_after(last_activity, timeout)) { - /* the normal case, timeout is somewhere before last_activity */ - timeout = last_activity - timeout; - } else { - /* either jiffies wrapped around, or no fence was signaled in the last 500ms - * anyway we will just wait for the minimum amount and then check for a lockup - */ - timeout = 1; - } - - trace_radeon_fence_wait_begin(rdev->ddev, target_seq[ring]); - for (i = 0; i < RADEON_NUM_RINGS; ++i) { - if (target_seq[i]) { - radeon_irq_kms_sw_irq_get(rdev, i); - } - } - if (intr) { - r = wait_event_interruptible_timeout(rdev->fence_queue, - (signaled = radeon_fence_any_seq_signaled(rdev, target_seq)), - timeout); - } else { - r = wait_event_timeout(rdev->fence_queue, - (signaled = radeon_fence_any_seq_signaled(rdev, target_seq)), - timeout); - } - for (i = 0; i < RADEON_NUM_RINGS; ++i) { - if (target_seq[i]) { - radeon_irq_kms_sw_irq_put(rdev, i); - } - } - if (unlikely(r < 0)) { - return r; - } - trace_radeon_fence_wait_end(rdev->ddev, target_seq[ring]); + seq[fence->ring] = fence->seq; + if (seq[fence->ring] == RADEON_FENCE_SIGNALED_SEQ) + return 0; - if (unlikely(!signaled)) { - /* we were interrupted for some reason and fence - * isn't signaled yet, resume waiting */ - if (r) { - continue; - } - - mutex_lock(&rdev->ring_lock); - for (i = 0, tmp = 0; i < RADEON_NUM_RINGS; ++i) { - if (time_after(rdev->fence_drv[i].last_activity, tmp)) { - tmp = rdev->fence_drv[i].last_activity; - } - } - /* test if somebody else has already decided that this is a lockup */ - if (last_activity != tmp) { - last_activity = tmp; - mutex_unlock(&rdev->ring_lock); - continue; - } - - if (radeon_ring_is_lockup(rdev, ring, &rdev->ring[ring])) { - /* good news we believe it's a lockup */ - dev_warn(rdev->dev, "GPU lockup (waiting for 0x%016llx)\n", - target_seq[ring]); - - /* change last activity so nobody else think there is a lockup */ - for (i = 0; i < RADEON_NUM_RINGS; ++i) { - rdev->fence_drv[i].last_activity = jiffies; - } + r = radeon_fence_wait_seq(fence->rdev, seq, intr); + if (r) + return r; - /* mark the ring as not ready any more */ - rdev->ring[ring].ready = false; - mutex_unlock(&rdev->ring_lock); - return -EDEADLK; - } - mutex_unlock(&rdev->ring_lock); - } - } + fence->seq = RADEON_FENCE_SIGNALED_SEQ; return 0; } @@ -557,7 +433,7 @@ int radeon_fence_wait_any(struct radeon_device *rdev, bool intr) { uint64_t seq[RADEON_NUM_RINGS]; - unsigned i; + unsigned i, num_rings = 0; int r; for (i = 0; i < RADEON_NUM_RINGS; ++i) { @@ -567,15 +443,19 @@ int radeon_fence_wait_any(struct radeon_device *rdev, continue; } - if (fences[i]->seq == RADEON_FENCE_SIGNALED_SEQ) { - /* something was allready signaled */ - return 0; - } - seq[i] = fences[i]->seq; + ++num_rings; + + /* test if something was allready signaled */ + if (seq[i] == RADEON_FENCE_SIGNALED_SEQ) + return 0; } - r = radeon_fence_wait_any_seq(rdev, seq, intr); + /* nothing to wait for ? */ + if (num_rings == 0) + return -ENOENT; + + r = radeon_fence_wait_seq(rdev, seq, intr); if (r) { return r; } @@ -583,7 +463,7 @@ int radeon_fence_wait_any(struct radeon_device *rdev, } /** - * radeon_fence_wait_next_locked - wait for the next fence to signal + * radeon_fence_wait_next - wait for the next fence to signal * * @rdev: radeon device pointer * @ring: ring index the fence is associated with @@ -592,21 +472,21 @@ int radeon_fence_wait_any(struct radeon_device *rdev, * Returns 0 if the next fence has passed, error for all other cases. * Caller must hold ring lock. */ -int radeon_fence_wait_next_locked(struct radeon_device *rdev, int ring) +int radeon_fence_wait_next(struct radeon_device *rdev, int ring) { - uint64_t seq; + uint64_t seq[RADEON_NUM_RINGS] = {}; - seq = atomic64_read(&rdev->fence_drv[ring].last_seq) + 1ULL; - if (seq >= rdev->fence_drv[ring].sync_seq[ring]) { + seq[ring] = atomic64_read(&rdev->fence_drv[ring].last_seq) + 1ULL; + if (seq[ring] >= rdev->fence_drv[ring].sync_seq[ring]) { /* nothing to wait for, last_seq is already the last emited fence */ return -ENOENT; } - return radeon_fence_wait_seq(rdev, seq, ring, false, false); + return radeon_fence_wait_seq(rdev, seq, false); } /** - * radeon_fence_wait_empty_locked - wait for all fences to signal + * radeon_fence_wait_empty - wait for all fences to signal * * @rdev: radeon device pointer * @ring: ring index the fence is associated with @@ -615,16 +495,20 @@ int radeon_fence_wait_next_locked(struct radeon_device *rdev, int ring) * Returns 0 if the fences have passed, error for all other cases. * Caller must hold ring lock. */ -int radeon_fence_wait_empty_locked(struct radeon_device *rdev, int ring) +int radeon_fence_wait_empty(struct radeon_device *rdev, int ring) { - uint64_t seq = rdev->fence_drv[ring].sync_seq[ring]; + uint64_t seq[RADEON_NUM_RINGS] = {}; int r; - r = radeon_fence_wait_seq(rdev, seq, ring, false, false); + seq[ring] = rdev->fence_drv[ring].sync_seq[ring]; + if (!seq[ring]) + return 0; + + r = radeon_fence_wait_seq(rdev, seq, false); if (r) { - if (r == -EDEADLK) { + if (r == -EDEADLK) return -EDEADLK; - } + dev_err(rdev->dev, "error waiting for ring[%d] to become idle (%d)\n", ring, r); } @@ -826,7 +710,6 @@ static void radeon_fence_driver_init_ring(struct radeon_device *rdev, int ring) for (i = 0; i < RADEON_NUM_RINGS; ++i) rdev->fence_drv[ring].sync_seq[i] = 0; atomic64_set(&rdev->fence_drv[ring].last_seq, 0); - rdev->fence_drv[ring].last_activity = jiffies; rdev->fence_drv[ring].initialized = false; } @@ -872,7 +755,7 @@ void radeon_fence_driver_fini(struct radeon_device *rdev) for (ring = 0; ring < RADEON_NUM_RINGS; ring++) { if (!rdev->fence_drv[ring].initialized) continue; - r = radeon_fence_wait_empty_locked(rdev, ring); + r = radeon_fence_wait_empty(rdev, ring); if (r) { /* no need to trigger GPU reset as we are unloading */ radeon_fence_driver_force_completion(rdev); @@ -919,6 +802,8 @@ static int radeon_debugfs_fence_info(struct seq_file *m, void *data) if (!rdev->fence_drv[i].initialized) continue; + radeon_fence_process(rdev, i); + seq_printf(m, "--- ring %d ---\n", i); seq_printf(m, "Last signaled fence 0x%016llx\n", (unsigned long long)atomic64_read(&rdev->fence_drv[i].last_seq)); @@ -934,15 +819,35 @@ static int radeon_debugfs_fence_info(struct seq_file *m, void *data) return 0; } +/** + * radeon_debugfs_gpu_reset - manually trigger a gpu reset + * + * Manually trigger a gpu reset at the next fence wait. + */ +static int radeon_debugfs_gpu_reset(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct radeon_device *rdev = dev->dev_private; + + down_read(&rdev->exclusive_lock); + seq_printf(m, "%d\n", rdev->needs_reset); + rdev->needs_reset = true; + up_read(&rdev->exclusive_lock); + + return 0; +} + static struct drm_info_list radeon_debugfs_fence_list[] = { {"radeon_fence_info", &radeon_debugfs_fence_info, 0, NULL}, + {"radeon_gpu_reset", &radeon_debugfs_gpu_reset, 0, NULL} }; #endif int radeon_debugfs_fence_init(struct radeon_device *rdev) { #if defined(CONFIG_DEBUG_FS) - return radeon_debugfs_add_files(rdev, radeon_debugfs_fence_list, 1); + return radeon_debugfs_add_files(rdev, radeon_debugfs_fence_list, 2); #else return 0; #endif diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c index 43ec4a401f0..2e723651069 100644 --- a/drivers/gpu/drm/radeon/radeon_gart.c +++ b/drivers/gpu/drm/radeon/radeon_gart.c @@ -28,7 +28,6 @@ #include <drm/drmP.h> #include <drm/radeon_drm.h> #include "radeon.h" -#include "radeon_reg.h" /* * GART @@ -207,7 +206,6 @@ void radeon_gart_table_vram_free(struct radeon_device *rdev) if (rdev->gart.robj == NULL) { return; } - radeon_gart_table_vram_unpin(rdev); radeon_bo_unref(&rdev->gart.robj); } @@ -394,893 +392,3 @@ void radeon_gart_fini(struct radeon_device *rdev) radeon_dummy_page_fini(rdev); } - -/* - * GPUVM - * GPUVM is similar to the legacy gart on older asics, however - * rather than there being a single global gart table - * for the entire GPU, there are multiple VM page tables active - * at any given time. The VM page tables can contain a mix - * vram pages and system memory pages and system memory pages - * can be mapped as snooped (cached system pages) or unsnooped - * (uncached system pages). - * Each VM has an ID associated with it and there is a page table - * associated with each VMID. When execting a command buffer, - * the kernel tells the the ring what VMID to use for that command - * buffer. VMIDs are allocated dynamically as commands are submitted. - * The userspace drivers maintain their own address space and the kernel - * sets up their pages tables accordingly when they submit their - * command buffers and a VMID is assigned. - * Cayman/Trinity support up to 8 active VMs at any given time; - * SI supports 16. - */ - -/* - * vm helpers - * - * TODO bind a default page at vm initialization for default address - */ - -/** - * radeon_vm_num_pde - return the number of page directory entries - * - * @rdev: radeon_device pointer - * - * Calculate the number of page directory entries (cayman+). - */ -static unsigned radeon_vm_num_pdes(struct radeon_device *rdev) -{ - return rdev->vm_manager.max_pfn >> RADEON_VM_BLOCK_SIZE; -} - -/** - * radeon_vm_directory_size - returns the size of the page directory in bytes - * - * @rdev: radeon_device pointer - * - * Calculate the size of the page directory in bytes (cayman+). - */ -static unsigned radeon_vm_directory_size(struct radeon_device *rdev) -{ - return RADEON_GPU_PAGE_ALIGN(radeon_vm_num_pdes(rdev) * 8); -} - -/** - * radeon_vm_manager_init - init the vm manager - * - * @rdev: radeon_device pointer - * - * Init the vm manager (cayman+). - * Returns 0 for success, error for failure. - */ -int radeon_vm_manager_init(struct radeon_device *rdev) -{ - struct radeon_vm *vm; - struct radeon_bo_va *bo_va; - int r; - unsigned size; - - if (!rdev->vm_manager.enabled) { - /* allocate enough for 2 full VM pts */ - size = radeon_vm_directory_size(rdev); - size += rdev->vm_manager.max_pfn * 8; - size *= 2; - r = radeon_sa_bo_manager_init(rdev, &rdev->vm_manager.sa_manager, - RADEON_GPU_PAGE_ALIGN(size), - RADEON_GEM_DOMAIN_VRAM); - if (r) { - dev_err(rdev->dev, "failed to allocate vm bo (%dKB)\n", - (rdev->vm_manager.max_pfn * 8) >> 10); - return r; - } - - r = radeon_asic_vm_init(rdev); - if (r) - return r; - - rdev->vm_manager.enabled = true; - - r = radeon_sa_bo_manager_start(rdev, &rdev->vm_manager.sa_manager); - if (r) - return r; - } - - /* restore page table */ - list_for_each_entry(vm, &rdev->vm_manager.lru_vm, list) { - if (vm->page_directory == NULL) - continue; - - list_for_each_entry(bo_va, &vm->va, vm_list) { - bo_va->valid = false; - } - } - return 0; -} - -/** - * radeon_vm_free_pt - free the page table for a specific vm - * - * @rdev: radeon_device pointer - * @vm: vm to unbind - * - * Free the page table of a specific vm (cayman+). - * - * Global and local mutex must be lock! - */ -static void radeon_vm_free_pt(struct radeon_device *rdev, - struct radeon_vm *vm) -{ - struct radeon_bo_va *bo_va; - int i; - - if (!vm->page_directory) - return; - - list_del_init(&vm->list); - radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); - - list_for_each_entry(bo_va, &vm->va, vm_list) { - bo_va->valid = false; - } - - if (vm->page_tables == NULL) - return; - - for (i = 0; i < radeon_vm_num_pdes(rdev); i++) - radeon_sa_bo_free(rdev, &vm->page_tables[i], vm->fence); - - kfree(vm->page_tables); -} - -/** - * radeon_vm_manager_fini - tear down the vm manager - * - * @rdev: radeon_device pointer - * - * Tear down the VM manager (cayman+). - */ -void radeon_vm_manager_fini(struct radeon_device *rdev) -{ - struct radeon_vm *vm, *tmp; - int i; - - if (!rdev->vm_manager.enabled) - return; - - mutex_lock(&rdev->vm_manager.lock); - /* free all allocated page tables */ - list_for_each_entry_safe(vm, tmp, &rdev->vm_manager.lru_vm, list) { - mutex_lock(&vm->mutex); - radeon_vm_free_pt(rdev, vm); - mutex_unlock(&vm->mutex); - } - for (i = 0; i < RADEON_NUM_VM; ++i) { - radeon_fence_unref(&rdev->vm_manager.active[i]); - } - radeon_asic_vm_fini(rdev); - mutex_unlock(&rdev->vm_manager.lock); - - radeon_sa_bo_manager_suspend(rdev, &rdev->vm_manager.sa_manager); - radeon_sa_bo_manager_fini(rdev, &rdev->vm_manager.sa_manager); - rdev->vm_manager.enabled = false; -} - -/** - * radeon_vm_evict - evict page table to make room for new one - * - * @rdev: radeon_device pointer - * @vm: VM we want to allocate something for - * - * Evict a VM from the lru, making sure that it isn't @vm. (cayman+). - * Returns 0 for success, -ENOMEM for failure. - * - * Global and local mutex must be locked! - */ -static int radeon_vm_evict(struct radeon_device *rdev, struct radeon_vm *vm) -{ - struct radeon_vm *vm_evict; - - if (list_empty(&rdev->vm_manager.lru_vm)) - return -ENOMEM; - - vm_evict = list_first_entry(&rdev->vm_manager.lru_vm, - struct radeon_vm, list); - if (vm_evict == vm) - return -ENOMEM; - - mutex_lock(&vm_evict->mutex); - radeon_vm_free_pt(rdev, vm_evict); - mutex_unlock(&vm_evict->mutex); - return 0; -} - -/** - * radeon_vm_alloc_pt - allocates a page table for a VM - * - * @rdev: radeon_device pointer - * @vm: vm to bind - * - * Allocate a page table for the requested vm (cayman+). - * Returns 0 for success, error for failure. - * - * Global and local mutex must be locked! - */ -int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm) -{ - unsigned pd_size, pts_size; - u64 *pd_addr; - int r; - - if (vm == NULL) { - return -EINVAL; - } - - if (vm->page_directory != NULL) { - return 0; - } - -retry: - pd_size = RADEON_GPU_PAGE_ALIGN(radeon_vm_directory_size(rdev)); - r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, - &vm->page_directory, pd_size, - RADEON_GPU_PAGE_SIZE, false); - if (r == -ENOMEM) { - r = radeon_vm_evict(rdev, vm); - if (r) - return r; - goto retry; - - } else if (r) { - return r; - } - - vm->pd_gpu_addr = radeon_sa_bo_gpu_addr(vm->page_directory); - - /* Initially clear the page directory */ - pd_addr = radeon_sa_bo_cpu_addr(vm->page_directory); - memset(pd_addr, 0, pd_size); - - pts_size = radeon_vm_num_pdes(rdev) * sizeof(struct radeon_sa_bo *); - vm->page_tables = kzalloc(pts_size, GFP_KERNEL); - - if (vm->page_tables == NULL) { - DRM_ERROR("Cannot allocate memory for page table array\n"); - radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); - return -ENOMEM; - } - - return 0; -} - -/** - * radeon_vm_add_to_lru - add VMs page table to LRU list - * - * @rdev: radeon_device pointer - * @vm: vm to add to LRU - * - * Add the allocated page table to the LRU list (cayman+). - * - * Global mutex must be locked! - */ -void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm) -{ - list_del_init(&vm->list); - list_add_tail(&vm->list, &rdev->vm_manager.lru_vm); -} - -/** - * radeon_vm_grab_id - allocate the next free VMID - * - * @rdev: radeon_device pointer - * @vm: vm to allocate id for - * @ring: ring we want to submit job to - * - * Allocate an id for the vm (cayman+). - * Returns the fence we need to sync to (if any). - * - * Global and local mutex must be locked! - */ -struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev, - struct radeon_vm *vm, int ring) -{ - struct radeon_fence *best[RADEON_NUM_RINGS] = {}; - unsigned choices[2] = {}; - unsigned i; - - /* check if the id is still valid */ - if (vm->fence && vm->fence == rdev->vm_manager.active[vm->id]) - return NULL; - - /* we definately need to flush */ - radeon_fence_unref(&vm->last_flush); - - /* skip over VMID 0, since it is the system VM */ - for (i = 1; i < rdev->vm_manager.nvm; ++i) { - struct radeon_fence *fence = rdev->vm_manager.active[i]; - - if (fence == NULL) { - /* found a free one */ - vm->id = i; - return NULL; - } - - if (radeon_fence_is_earlier(fence, best[fence->ring])) { - best[fence->ring] = fence; - choices[fence->ring == ring ? 0 : 1] = i; - } - } - - for (i = 0; i < 2; ++i) { - if (choices[i]) { - vm->id = choices[i]; - return rdev->vm_manager.active[choices[i]]; - } - } - - /* should never happen */ - BUG(); - return NULL; -} - -/** - * radeon_vm_fence - remember fence for vm - * - * @rdev: radeon_device pointer - * @vm: vm we want to fence - * @fence: fence to remember - * - * Fence the vm (cayman+). - * Set the fence used to protect page table and id. - * - * Global and local mutex must be locked! - */ -void radeon_vm_fence(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_fence *fence) -{ - radeon_fence_unref(&rdev->vm_manager.active[vm->id]); - rdev->vm_manager.active[vm->id] = radeon_fence_ref(fence); - - radeon_fence_unref(&vm->fence); - vm->fence = radeon_fence_ref(fence); -} - -/** - * radeon_vm_bo_find - find the bo_va for a specific vm & bo - * - * @vm: requested vm - * @bo: requested buffer object - * - * Find @bo inside the requested vm (cayman+). - * Search inside the @bos vm list for the requested vm - * Returns the found bo_va or NULL if none is found - * - * Object has to be reserved! - */ -struct radeon_bo_va *radeon_vm_bo_find(struct radeon_vm *vm, - struct radeon_bo *bo) -{ - struct radeon_bo_va *bo_va; - - list_for_each_entry(bo_va, &bo->va, bo_list) { - if (bo_va->vm == vm) { - return bo_va; - } - } - return NULL; -} - -/** - * radeon_vm_bo_add - add a bo to a specific vm - * - * @rdev: radeon_device pointer - * @vm: requested vm - * @bo: radeon buffer object - * - * Add @bo into the requested vm (cayman+). - * Add @bo to the list of bos associated with the vm - * Returns newly added bo_va or NULL for failure - * - * Object has to be reserved! - */ -struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_bo *bo) -{ - struct radeon_bo_va *bo_va; - - bo_va = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL); - if (bo_va == NULL) { - return NULL; - } - bo_va->vm = vm; - bo_va->bo = bo; - bo_va->soffset = 0; - bo_va->eoffset = 0; - bo_va->flags = 0; - bo_va->valid = false; - bo_va->ref_count = 1; - INIT_LIST_HEAD(&bo_va->bo_list); - INIT_LIST_HEAD(&bo_va->vm_list); - - mutex_lock(&vm->mutex); - list_add(&bo_va->vm_list, &vm->va); - list_add_tail(&bo_va->bo_list, &bo->va); - mutex_unlock(&vm->mutex); - - return bo_va; -} - -/** - * radeon_vm_bo_set_addr - set bos virtual address inside a vm - * - * @rdev: radeon_device pointer - * @bo_va: bo_va to store the address - * @soffset: requested offset of the buffer in the VM address space - * @flags: attributes of pages (read/write/valid/etc.) - * - * Set offset of @bo_va (cayman+). - * Validate and set the offset requested within the vm address space. - * Returns 0 for success, error for failure. - * - * Object has to be reserved! - */ -int radeon_vm_bo_set_addr(struct radeon_device *rdev, - struct radeon_bo_va *bo_va, - uint64_t soffset, - uint32_t flags) -{ - uint64_t size = radeon_bo_size(bo_va->bo); - uint64_t eoffset, last_offset = 0; - struct radeon_vm *vm = bo_va->vm; - struct radeon_bo_va *tmp; - struct list_head *head; - unsigned last_pfn; - - if (soffset) { - /* make sure object fit at this offset */ - eoffset = soffset + size; - if (soffset >= eoffset) { - return -EINVAL; - } - - last_pfn = eoffset / RADEON_GPU_PAGE_SIZE; - if (last_pfn > rdev->vm_manager.max_pfn) { - dev_err(rdev->dev, "va above limit (0x%08X > 0x%08X)\n", - last_pfn, rdev->vm_manager.max_pfn); - return -EINVAL; - } - - } else { - eoffset = last_pfn = 0; - } - - mutex_lock(&vm->mutex); - head = &vm->va; - last_offset = 0; - list_for_each_entry(tmp, &vm->va, vm_list) { - if (bo_va == tmp) { - /* skip over currently modified bo */ - continue; - } - - if (soffset >= last_offset && eoffset <= tmp->soffset) { - /* bo can be added before this one */ - break; - } - if (eoffset > tmp->soffset && soffset < tmp->eoffset) { - /* bo and tmp overlap, invalid offset */ - dev_err(rdev->dev, "bo %p va 0x%08X conflict with (bo %p 0x%08X 0x%08X)\n", - bo_va->bo, (unsigned)bo_va->soffset, tmp->bo, - (unsigned)tmp->soffset, (unsigned)tmp->eoffset); - mutex_unlock(&vm->mutex); - return -EINVAL; - } - last_offset = tmp->eoffset; - head = &tmp->vm_list; - } - - bo_va->soffset = soffset; - bo_va->eoffset = eoffset; - bo_va->flags = flags; - bo_va->valid = false; - list_move(&bo_va->vm_list, head); - - mutex_unlock(&vm->mutex); - return 0; -} - -/** - * radeon_vm_map_gart - get the physical address of a gart page - * - * @rdev: radeon_device pointer - * @addr: the unmapped addr - * - * Look up the physical address of the page that the pte resolves - * to (cayman+). - * Returns the physical address of the page. - */ -uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr) -{ - uint64_t result; - - /* page table offset */ - result = rdev->gart.pages_addr[addr >> PAGE_SHIFT]; - - /* in case cpu page size != gpu page size*/ - result |= addr & (~PAGE_MASK); - - return result; -} - -/** - * radeon_vm_update_pdes - make sure that page directory is valid - * - * @rdev: radeon_device pointer - * @vm: requested vm - * @start: start of GPU address range - * @end: end of GPU address range - * - * Allocates new page tables if necessary - * and updates the page directory (cayman+). - * Returns 0 for success, error for failure. - * - * Global and local mutex must be locked! - */ -static int radeon_vm_update_pdes(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_ib *ib, - uint64_t start, uint64_t end) -{ - static const uint32_t incr = RADEON_VM_PTE_COUNT * 8; - - uint64_t last_pde = ~0, last_pt = ~0; - unsigned count = 0; - uint64_t pt_idx; - int r; - - start = (start / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE; - end = (end / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE; - - /* walk over the address space and update the page directory */ - for (pt_idx = start; pt_idx <= end; ++pt_idx) { - uint64_t pde, pt; - - if (vm->page_tables[pt_idx]) - continue; - -retry: - r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, - &vm->page_tables[pt_idx], - RADEON_VM_PTE_COUNT * 8, - RADEON_GPU_PAGE_SIZE, false); - - if (r == -ENOMEM) { - r = radeon_vm_evict(rdev, vm); - if (r) - return r; - goto retry; - } else if (r) { - return r; - } - - pde = vm->pd_gpu_addr + pt_idx * 8; - - pt = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]); - - if (((last_pde + 8 * count) != pde) || - ((last_pt + incr * count) != pt)) { - - if (count) { - radeon_asic_vm_set_page(rdev, ib, last_pde, - last_pt, count, incr, - RADEON_VM_PAGE_VALID); - } - - count = 1; - last_pde = pde; - last_pt = pt; - } else { - ++count; - } - } - - if (count) { - radeon_asic_vm_set_page(rdev, ib, last_pde, last_pt, count, - incr, RADEON_VM_PAGE_VALID); - - } - - return 0; -} - -/** - * radeon_vm_update_ptes - make sure that page tables are valid - * - * @rdev: radeon_device pointer - * @vm: requested vm - * @start: start of GPU address range - * @end: end of GPU address range - * @dst: destination address to map to - * @flags: mapping flags - * - * Update the page tables in the range @start - @end (cayman+). - * - * Global and local mutex must be locked! - */ -static void radeon_vm_update_ptes(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_ib *ib, - uint64_t start, uint64_t end, - uint64_t dst, uint32_t flags) -{ - static const uint64_t mask = RADEON_VM_PTE_COUNT - 1; - - uint64_t last_pte = ~0, last_dst = ~0; - unsigned count = 0; - uint64_t addr; - - start = start / RADEON_GPU_PAGE_SIZE; - end = end / RADEON_GPU_PAGE_SIZE; - - /* walk over the address space and update the page tables */ - for (addr = start; addr < end; ) { - uint64_t pt_idx = addr >> RADEON_VM_BLOCK_SIZE; - unsigned nptes; - uint64_t pte; - - if ((addr & ~mask) == (end & ~mask)) - nptes = end - addr; - else - nptes = RADEON_VM_PTE_COUNT - (addr & mask); - - pte = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]); - pte += (addr & mask) * 8; - - if ((last_pte + 8 * count) != pte) { - - if (count) { - radeon_asic_vm_set_page(rdev, ib, last_pte, - last_dst, count, - RADEON_GPU_PAGE_SIZE, - flags); - } - - count = nptes; - last_pte = pte; - last_dst = dst; - } else { - count += nptes; - } - - addr += nptes; - dst += nptes * RADEON_GPU_PAGE_SIZE; - } - - if (count) { - radeon_asic_vm_set_page(rdev, ib, last_pte, - last_dst, count, - RADEON_GPU_PAGE_SIZE, flags); - } -} - -/** - * radeon_vm_bo_update_pte - map a bo into the vm page table - * - * @rdev: radeon_device pointer - * @vm: requested vm - * @bo: radeon buffer object - * @mem: ttm mem - * - * Fill in the page table entries for @bo (cayman+). - * Returns 0 for success, -EINVAL for failure. - * - * Object have to be reserved & global and local mutex must be locked! - */ -int radeon_vm_bo_update_pte(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_bo *bo, - struct ttm_mem_reg *mem) -{ - unsigned ridx = rdev->asic->vm.pt_ring_index; - struct radeon_ib ib; - struct radeon_bo_va *bo_va; - unsigned nptes, npdes, ndw; - uint64_t addr; - int r; - - /* nothing to do if vm isn't bound */ - if (vm->page_directory == NULL) - return 0; - - bo_va = radeon_vm_bo_find(vm, bo); - if (bo_va == NULL) { - dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm); - return -EINVAL; - } - - if (!bo_va->soffset) { - dev_err(rdev->dev, "bo %p don't has a mapping in vm %p\n", - bo, vm); - return -EINVAL; - } - - if ((bo_va->valid && mem) || (!bo_va->valid && mem == NULL)) - return 0; - - bo_va->flags &= ~RADEON_VM_PAGE_VALID; - bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM; - if (mem) { - addr = mem->start << PAGE_SHIFT; - if (mem->mem_type != TTM_PL_SYSTEM) { - bo_va->flags |= RADEON_VM_PAGE_VALID; - bo_va->valid = true; - } - if (mem->mem_type == TTM_PL_TT) { - bo_va->flags |= RADEON_VM_PAGE_SYSTEM; - } else { - addr += rdev->vm_manager.vram_base_offset; - } - } else { - addr = 0; - bo_va->valid = false; - } - - nptes = radeon_bo_ngpu_pages(bo); - - /* assume two extra pdes in case the mapping overlaps the borders */ - npdes = (nptes >> RADEON_VM_BLOCK_SIZE) + 2; - - /* padding, etc. */ - ndw = 64; - - if (RADEON_VM_BLOCK_SIZE > 11) - /* reserve space for one header for every 2k dwords */ - ndw += (nptes >> 11) * 4; - else - /* reserve space for one header for - every (1 << BLOCK_SIZE) entries */ - ndw += (nptes >> RADEON_VM_BLOCK_SIZE) * 4; - - /* reserve space for pte addresses */ - ndw += nptes * 2; - - /* reserve space for one header for every 2k dwords */ - ndw += (npdes >> 11) * 4; - - /* reserve space for pde addresses */ - ndw += npdes * 2; - - /* update too big for an IB */ - if (ndw > 0xfffff) - return -ENOMEM; - - r = radeon_ib_get(rdev, ridx, &ib, NULL, ndw * 4); - ib.length_dw = 0; - - r = radeon_vm_update_pdes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset); - if (r) { - radeon_ib_free(rdev, &ib); - return r; - } - - radeon_vm_update_ptes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset, - addr, bo_va->flags); - - radeon_ib_sync_to(&ib, vm->fence); - r = radeon_ib_schedule(rdev, &ib, NULL); - if (r) { - radeon_ib_free(rdev, &ib); - return r; - } - radeon_fence_unref(&vm->fence); - vm->fence = radeon_fence_ref(ib.fence); - radeon_ib_free(rdev, &ib); - radeon_fence_unref(&vm->last_flush); - - return 0; -} - -/** - * radeon_vm_bo_rmv - remove a bo to a specific vm - * - * @rdev: radeon_device pointer - * @bo_va: requested bo_va - * - * Remove @bo_va->bo from the requested vm (cayman+). - * Remove @bo_va->bo from the list of bos associated with the bo_va->vm and - * remove the ptes for @bo_va in the page table. - * Returns 0 for success. - * - * Object have to be reserved! - */ -int radeon_vm_bo_rmv(struct radeon_device *rdev, - struct radeon_bo_va *bo_va) -{ - int r = 0; - - mutex_lock(&rdev->vm_manager.lock); - mutex_lock(&bo_va->vm->mutex); - if (bo_va->soffset) { - r = radeon_vm_bo_update_pte(rdev, bo_va->vm, bo_va->bo, NULL); - } - mutex_unlock(&rdev->vm_manager.lock); - list_del(&bo_va->vm_list); - mutex_unlock(&bo_va->vm->mutex); - list_del(&bo_va->bo_list); - - kfree(bo_va); - return r; -} - -/** - * radeon_vm_bo_invalidate - mark the bo as invalid - * - * @rdev: radeon_device pointer - * @vm: requested vm - * @bo: radeon buffer object - * - * Mark @bo as invalid (cayman+). - */ -void radeon_vm_bo_invalidate(struct radeon_device *rdev, - struct radeon_bo *bo) -{ - struct radeon_bo_va *bo_va; - - list_for_each_entry(bo_va, &bo->va, bo_list) { - bo_va->valid = false; - } -} - -/** - * radeon_vm_init - initialize a vm instance - * - * @rdev: radeon_device pointer - * @vm: requested vm - * - * Init @vm fields (cayman+). - */ -void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm) -{ - vm->id = 0; - vm->fence = NULL; - mutex_init(&vm->mutex); - INIT_LIST_HEAD(&vm->list); - INIT_LIST_HEAD(&vm->va); -} - -/** - * radeon_vm_fini - tear down a vm instance - * - * @rdev: radeon_device pointer - * @vm: requested vm - * - * Tear down @vm (cayman+). - * Unbind the VM and remove all bos from the vm bo list - */ -void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm) -{ - struct radeon_bo_va *bo_va, *tmp; - int r; - - mutex_lock(&rdev->vm_manager.lock); - mutex_lock(&vm->mutex); - radeon_vm_free_pt(rdev, vm); - mutex_unlock(&rdev->vm_manager.lock); - - if (!list_empty(&vm->va)) { - dev_err(rdev->dev, "still active bo inside vm\n"); - } - list_for_each_entry_safe(bo_va, tmp, &vm->va, vm_list) { - list_del_init(&bo_va->vm_list); - r = radeon_bo_reserve(bo_va->bo, false); - if (!r) { - list_del_init(&bo_va->bo_list); - radeon_bo_unreserve(bo_va->bo); - kfree(bo_va); - } - } - radeon_fence_unref(&vm->fence); - radeon_fence_unref(&vm->last_flush); - mutex_unlock(&vm->mutex); -} diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c index aa796031ab6..d09650c1d72 100644 --- a/drivers/gpu/drm/radeon/radeon_gem.c +++ b/drivers/gpu/drm/radeon/radeon_gem.c @@ -29,13 +29,6 @@ #include <drm/radeon_drm.h> #include "radeon.h" -int radeon_gem_object_init(struct drm_gem_object *obj) -{ - BUG(); - - return 0; -} - void radeon_gem_object_free(struct drm_gem_object *gobj) { struct radeon_bo *robj = gem_to_radeon_bo(gobj); @@ -93,7 +86,7 @@ retry: return 0; } -int radeon_gem_set_domain(struct drm_gem_object *gobj, +static int radeon_gem_set_domain(struct drm_gem_object *gobj, uint32_t rdomain, uint32_t wdomain) { struct radeon_bo *robj; @@ -351,18 +344,7 @@ int radeon_gem_busy_ioctl(struct drm_device *dev, void *data, } robj = gem_to_radeon_bo(gobj); r = radeon_bo_wait(robj, &cur_placement, true); - switch (cur_placement) { - case TTM_PL_VRAM: - args->domain = RADEON_GEM_DOMAIN_VRAM; - break; - case TTM_PL_TT: - args->domain = RADEON_GEM_DOMAIN_GTT; - break; - case TTM_PL_SYSTEM: - args->domain = RADEON_GEM_DOMAIN_CPU; - default: - break; - } + args->domain = radeon_mem_type_to_domain(cur_placement); drm_gem_object_unreference_unlocked(gobj); r = radeon_gem_handle_lockup(rdev, r); return r; @@ -540,6 +522,42 @@ out: return r; } +int radeon_gem_op_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp) +{ + struct drm_radeon_gem_op *args = data; + struct drm_gem_object *gobj; + struct radeon_bo *robj; + int r; + + gobj = drm_gem_object_lookup(dev, filp, args->handle); + if (gobj == NULL) { + return -ENOENT; + } + robj = gem_to_radeon_bo(gobj); + r = radeon_bo_reserve(robj, false); + if (unlikely(r)) + goto out; + + switch (args->op) { + case RADEON_GEM_OP_GET_INITIAL_DOMAIN: + args->value = robj->initial_domain; + break; + case RADEON_GEM_OP_SET_INITIAL_DOMAIN: + robj->initial_domain = args->value & (RADEON_GEM_DOMAIN_VRAM | + RADEON_GEM_DOMAIN_GTT | + RADEON_GEM_DOMAIN_CPU); + break; + default: + r = -EINVAL; + } + + radeon_bo_unreserve(robj); +out: + drm_gem_object_unreference_unlocked(gobj); + return r; +} + int radeon_mode_dumb_create(struct drm_file *file_priv, struct drm_device *dev, struct drm_mode_create_dumb *args) @@ -570,13 +588,6 @@ int radeon_mode_dumb_create(struct drm_file *file_priv, return 0; } -int radeon_mode_dumb_destroy(struct drm_file *file_priv, - struct drm_device *dev, - uint32_t handle) -{ - return drm_gem_handle_delete(file_priv, handle); -} - #if defined(CONFIG_DEBUG_FS) static int radeon_debugfs_gem_info(struct seq_file *m, void *data) { diff --git a/drivers/gpu/drm/radeon/radeon_i2c.c b/drivers/gpu/drm/radeon/radeon_i2c.c index fc60b74ee30..add62200840 100644 --- a/drivers/gpu/drm/radeon/radeon_i2c.c +++ b/drivers/gpu/drm/radeon/radeon_i2c.c @@ -64,8 +64,7 @@ bool radeon_ddc_probe(struct radeon_connector *radeon_connector, bool use_aux) radeon_router_select_ddc_port(radeon_connector); if (use_aux) { - struct radeon_connector_atom_dig *dig = radeon_connector->con_priv; - ret = i2c_transfer(&dig->dp_i2c_bus->adapter, msgs, 2); + ret = i2c_transfer(&radeon_connector->ddc_bus->aux.ddc, msgs, 2); } else { ret = i2c_transfer(&radeon_connector->ddc_bus->adapter, msgs, 2); } @@ -95,6 +94,8 @@ static int pre_xfer(struct i2c_adapter *i2c_adap) struct radeon_i2c_bus_rec *rec = &i2c->rec; uint32_t temp; + mutex_lock(&i2c->mutex); + /* RV410 appears to have a bug where the hw i2c in reset * holds the i2c port in a bad state - switch hw i2c away before * doing DDC - do this for all r200s/r300s/r400s for safety sake @@ -171,6 +172,8 @@ static void post_xfer(struct i2c_adapter *i2c_adap) temp = RREG32(rec->mask_data_reg) & ~rec->mask_data_mask; WREG32(rec->mask_data_reg, temp); temp = RREG32(rec->mask_data_reg); + + mutex_unlock(&i2c->mutex); } static int get_clock(void *i2c_priv) @@ -814,6 +817,8 @@ static int radeon_hw_i2c_xfer(struct i2c_adapter *i2c_adap, struct radeon_i2c_bus_rec *rec = &i2c->rec; int ret = 0; + mutex_lock(&i2c->mutex); + switch (rdev->family) { case CHIP_R100: case CHIP_RV100: @@ -880,6 +885,8 @@ static int radeon_hw_i2c_xfer(struct i2c_adapter *i2c_adap, break; } + mutex_unlock(&i2c->mutex); + return ret; } @@ -920,6 +927,7 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev, i2c->adapter.dev.parent = &dev->pdev->dev; i2c->dev = dev; i2c_set_adapdata(&i2c->adapter, i2c); + mutex_init(&i2c->mutex); if (rec->mm_i2c || (rec->hw_capable && radeon_hw_i2c && @@ -950,16 +958,16 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev, /* set the radeon bit adapter */ snprintf(i2c->adapter.name, sizeof(i2c->adapter.name), "Radeon i2c bit bus %s", name); - i2c->adapter.algo_data = &i2c->algo.bit; - i2c->algo.bit.pre_xfer = pre_xfer; - i2c->algo.bit.post_xfer = post_xfer; - i2c->algo.bit.setsda = set_data; - i2c->algo.bit.setscl = set_clock; - i2c->algo.bit.getsda = get_data; - i2c->algo.bit.getscl = get_clock; - i2c->algo.bit.udelay = 10; - i2c->algo.bit.timeout = usecs_to_jiffies(2200); /* from VESA */ - i2c->algo.bit.data = i2c; + i2c->adapter.algo_data = &i2c->bit; + i2c->bit.pre_xfer = pre_xfer; + i2c->bit.post_xfer = post_xfer; + i2c->bit.setsda = set_data; + i2c->bit.setscl = set_clock; + i2c->bit.getsda = get_data; + i2c->bit.getscl = get_clock; + i2c->bit.udelay = 10; + i2c->bit.timeout = usecs_to_jiffies(2200); /* from VESA */ + i2c->bit.data = i2c; ret = i2c_bit_add_bus(&i2c->adapter); if (ret) { DRM_ERROR("Failed to register bit i2c %s\n", name); @@ -974,52 +982,22 @@ out_free: } -struct radeon_i2c_chan *radeon_i2c_create_dp(struct drm_device *dev, - struct radeon_i2c_bus_rec *rec, - const char *name) -{ - struct radeon_i2c_chan *i2c; - int ret; - - i2c = kzalloc(sizeof(struct radeon_i2c_chan), GFP_KERNEL); - if (i2c == NULL) - return NULL; - - i2c->rec = *rec; - i2c->adapter.owner = THIS_MODULE; - i2c->adapter.class = I2C_CLASS_DDC; - i2c->adapter.dev.parent = &dev->pdev->dev; - i2c->dev = dev; - snprintf(i2c->adapter.name, sizeof(i2c->adapter.name), - "Radeon aux bus %s", name); - i2c_set_adapdata(&i2c->adapter, i2c); - i2c->adapter.algo_data = &i2c->algo.dp; - i2c->algo.dp.aux_ch = radeon_dp_i2c_aux_ch; - i2c->algo.dp.address = 0; - ret = i2c_dp_aux_add_bus(&i2c->adapter); - if (ret) { - DRM_INFO("Failed to register i2c %s\n", name); - goto out_free; - } - - return i2c; -out_free: - kfree(i2c); - return NULL; - -} - void radeon_i2c_destroy(struct radeon_i2c_chan *i2c) { if (!i2c) return; i2c_del_adapter(&i2c->adapter); + if (i2c->has_aux) + drm_dp_aux_unregister(&i2c->aux); kfree(i2c); } /* Add the default buses */ void radeon_i2c_init(struct radeon_device *rdev) { + if (radeon_hw_i2c) + DRM_INFO("hw_i2c forced on, you may experience display detection problems!\n"); + if (rdev->is_atom_bios) radeon_atombios_i2c_init(rdev); else diff --git a/drivers/gpu/drm/radeon/radeon_ioc32.c b/drivers/gpu/drm/radeon/radeon_ioc32.c index c180df8e84d..0b98ea13457 100644 --- a/drivers/gpu/drm/radeon/radeon_ioc32.c +++ b/drivers/gpu/drm/radeon/radeon_ioc32.c @@ -399,7 +399,7 @@ long radeon_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) if (nr < DRM_COMMAND_BASE) return drm_compat_ioctl(filp, cmd, arg); - if (nr < DRM_COMMAND_BASE + DRM_ARRAY_SIZE(radeon_compat_ioctls)) + if (nr < DRM_COMMAND_BASE + ARRAY_SIZE(radeon_compat_ioctls)) fn = radeon_compat_ioctls[nr - DRM_COMMAND_BASE]; if (fn != NULL) @@ -418,7 +418,7 @@ long radeon_kms_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long if (nr < DRM_COMMAND_BASE) return drm_compat_ioctl(filp, cmd, arg); - ret = drm_ioctl(filp, cmd, arg); + ret = radeon_drm_ioctl(filp, cmd, arg); return ret; } diff --git a/drivers/gpu/drm/radeon/radeon_irq.c b/drivers/gpu/drm/radeon/radeon_irq.c index 8d68e972789..244b19bab2e 100644 --- a/drivers/gpu/drm/radeon/radeon_irq.c +++ b/drivers/gpu/drm/radeon/radeon_irq.c @@ -181,7 +181,7 @@ static u32 radeon_acknowledge_irqs(drm_radeon_private_t *dev_priv, u32 *r500_dis * tied to dma at all, this is just a hangover from dri prehistory. */ -irqreturn_t radeon_driver_irq_handler(DRM_IRQ_ARGS) +irqreturn_t radeon_driver_irq_handler(int irq, void *arg) { struct drm_device *dev = (struct drm_device *) arg; drm_radeon_private_t *dev_priv = @@ -203,7 +203,7 @@ irqreturn_t radeon_driver_irq_handler(DRM_IRQ_ARGS) /* SW interrupt */ if (stat & RADEON_SW_INT_TEST) - DRM_WAKEUP(&dev_priv->swi_queue); + wake_up(&dev_priv->swi_queue); /* VBLANK interrupt */ if ((dev_priv->flags & RADEON_FAMILY_MASK) >= CHIP_RS600) { @@ -249,7 +249,7 @@ static int radeon_wait_irq(struct drm_device * dev, int swi_nr) dev_priv->stats.boxes |= RADEON_BOX_WAIT_IDLE; - DRM_WAIT_ON(ret, dev_priv->swi_queue, 3 * DRM_HZ, + DRM_WAIT_ON(ret, dev_priv->swi_queue, 3 * HZ, RADEON_READ(RADEON_LAST_SWI_REG) >= swi_nr); return ret; @@ -302,7 +302,7 @@ int radeon_irq_emit(struct drm_device *dev, void *data, struct drm_file *file_pr result = radeon_emit_irq(dev); - if (DRM_COPY_TO_USER(emit->irq_seq, &result, sizeof(int))) { + if (copy_to_user(emit->irq_seq, &result, sizeof(int))) { DRM_ERROR("copy_to_user\n"); return -EFAULT; } @@ -354,7 +354,7 @@ int radeon_driver_irq_postinstall(struct drm_device *dev) (drm_radeon_private_t *) dev->dev_private; atomic_set(&dev_priv->swi_emitted, 0); - DRM_INIT_WAITQUEUE(&dev_priv->swi_queue); + init_waitqueue_head(&dev_priv->swi_queue); dev->max_vblank_count = 0x001fffff; diff --git a/drivers/gpu/drm/radeon/radeon_irq_kms.c b/drivers/gpu/drm/radeon/radeon_irq_kms.c index bcdefd1dcd4..16807afab36 100644 --- a/drivers/gpu/drm/radeon/radeon_irq_kms.c +++ b/drivers/gpu/drm/radeon/radeon_irq_kms.c @@ -32,23 +32,29 @@ #include "radeon.h" #include "atom.h" +#include <linux/pm_runtime.h> + #define RADEON_WAIT_IDLE_TIMEOUT 200 /** * radeon_driver_irq_handler_kms - irq handler for KMS * - * @DRM_IRQ_ARGS: args + * @int irq, void *arg: args * * This is the irq handler for the radeon KMS driver (all asics). * radeon_irq_process is a macro that points to the per-asic * irq handler callback. */ -irqreturn_t radeon_driver_irq_handler_kms(DRM_IRQ_ARGS) +irqreturn_t radeon_driver_irq_handler_kms(int irq, void *arg) { struct drm_device *dev = (struct drm_device *) arg; struct radeon_device *rdev = dev->dev_private; + irqreturn_t ret; - return radeon_irq_process(rdev); + ret = radeon_irq_process(rdev); + if (ret == IRQ_HANDLED) + pm_runtime_mark_last_busy(dev->dev); + return ret; } /* @@ -260,10 +266,6 @@ int radeon_irq_kms_init(struct radeon_device *rdev) { int r = 0; - INIT_WORK(&rdev->hotplug_work, radeon_hotplug_work_func); - INIT_WORK(&rdev->audio_work, r600_audio_update_hdmi); - INIT_WORK(&rdev->reset_work, radeon_irq_reset_work_func); - spin_lock_init(&rdev->irq.lock); r = drm_vblank_init(rdev->ddev, rdev->num_crtc); if (r) { @@ -279,12 +281,19 @@ int radeon_irq_kms_init(struct radeon_device *rdev) dev_info(rdev->dev, "radeon: using MSI.\n"); } } + + INIT_WORK(&rdev->hotplug_work, radeon_hotplug_work_func); + INIT_WORK(&rdev->audio_work, r600_audio_update_hdmi); + INIT_WORK(&rdev->reset_work, radeon_irq_reset_work_func); + rdev->irq.installed = true; - r = drm_irq_install(rdev->ddev); + r = drm_irq_install(rdev->ddev, rdev->ddev->pdev->irq); if (r) { rdev->irq.installed = false; + flush_work(&rdev->hotplug_work); return r; } + DRM_INFO("radeon: irq initialized.\n"); return 0; } @@ -304,8 +313,8 @@ void radeon_irq_kms_fini(struct radeon_device *rdev) rdev->irq.installed = false; if (rdev->msi_enabled) pci_disable_msi(rdev->pdev); + flush_work(&rdev->hotplug_work); } - flush_work(&rdev->hotplug_work); } /** diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 49ff3d1a610..d25ae6acfd5 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -32,6 +32,13 @@ #include <linux/vga_switcheroo.h> #include <linux/slab.h> +#include <linux/pm_runtime.h> + +#if defined(CONFIG_VGA_SWITCHEROO) +bool radeon_has_atpx(void); +#else +static inline bool radeon_has_atpx(void) { return false; } +#endif /** * radeon_driver_unload_kms - Main unload function for KMS. @@ -50,9 +57,14 @@ int radeon_driver_unload_kms(struct drm_device *dev) if (rdev == NULL) return 0; + if (rdev->rmmio == NULL) goto done_free; + + pm_runtime_get_sync(dev->dev); + radeon_acpi_fini(rdev); + radeon_modeset_fini(rdev); radeon_device_fini(rdev); @@ -95,6 +107,11 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags) flags |= RADEON_IS_PCI; } + if ((radeon_runtime_pm != 0) && + radeon_has_atpx() && + ((flags & RADEON_IS_IGP) == 0)) + flags |= RADEON_IS_PX; + /* radeon_device_init should report only fatal error * like memory allocation failure or iomapping failure, * or memory manager initialization failure, it must @@ -125,9 +142,20 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags) "Error during ACPI methods call\n"); } + if (radeon_is_px(dev)) { + pm_runtime_use_autosuspend(dev->dev); + pm_runtime_set_autosuspend_delay(dev->dev, 5000); + pm_runtime_set_active(dev->dev); + pm_runtime_allow(dev->dev); + pm_runtime_mark_last_busy(dev->dev); + pm_runtime_put_autosuspend(dev->dev); + } + out: if (r) radeon_driver_unload_kms(dev); + + return r; } @@ -175,7 +203,7 @@ static void radeon_set_filp_rights(struct drm_device *dev, * etc. (all asics). * Returns 0 on success, -EINVAL on failure. */ -int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) +static int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) { struct radeon_device *rdev = dev->dev_private; struct drm_radeon_info *info = data; @@ -191,7 +219,7 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) switch (info->request) { case RADEON_INFO_DEVICE_ID: - *value = dev->pci_device; + *value = dev->pdev->device; break; case RADEON_INFO_NUM_GB_PIPES: *value = rdev->num_gb_pipes; @@ -207,7 +235,7 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) *value = rdev->accel_working; break; case RADEON_INFO_CRTC_FROM_ID: - if (DRM_COPY_FROM_USER(value, value_ptr, sizeof(uint32_t))) { + if (copy_from_user(value, value_ptr, sizeof(uint32_t))) { DRM_ERROR("copy_from_user %s:%u\n", __func__, __LINE__); return -EFAULT; } @@ -253,7 +281,7 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) * * When returning, the value is 1 if filp owns hyper-z access, * 0 otherwise. */ - if (DRM_COPY_FROM_USER(value, value_ptr, sizeof(uint32_t))) { + if (copy_from_user(value, value_ptr, sizeof(uint32_t))) { DRM_ERROR("copy_from_user %s:%u\n", __func__, __LINE__); return -EFAULT; } @@ -265,7 +293,7 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) break; case RADEON_INFO_WANT_CMASK: /* The same logic as Hyper-Z. */ - if (DRM_COPY_FROM_USER(value, value_ptr, sizeof(uint32_t))) { + if (copy_from_user(value, value_ptr, sizeof(uint32_t))) { DRM_ERROR("copy_from_user %s:%u\n", __func__, __LINE__); return -EFAULT; } @@ -324,7 +352,7 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) break; case RADEON_INFO_BACKEND_MAP: if (rdev->family >= CHIP_BONAIRE) - return -EINVAL; + *value = rdev->config.cik.backend_map; else if (rdev->family >= CHIP_TAHITI) *value = rdev->config.si.backend_map; else if (rdev->family >= CHIP_CAYMAN) @@ -401,7 +429,7 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) *value = rdev->fastfb_working; break; case RADEON_INFO_RING_WORKING: - if (DRM_COPY_FROM_USER(value, value_ptr, sizeof(uint32_t))) { + if (copy_from_user(value, value_ptr, sizeof(uint32_t))) { DRM_ERROR("copy_from_user %s:%u\n", __func__, __LINE__); return -EFAULT; } @@ -417,6 +445,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) case RADEON_CS_RING_UVD: *value = rdev->ring[R600_RING_TYPE_UVD_INDEX].ready; break; + case RADEON_CS_RING_VCE: + *value = rdev->ring[TN_RING_TYPE_VCE1_INDEX].ready; + break; default: return -EINVAL; } @@ -433,11 +464,76 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) return -EINVAL; } break; + case RADEON_INFO_CIK_MACROTILE_MODE_ARRAY: + if (rdev->family >= CHIP_BONAIRE) { + value = rdev->config.cik.macrotile_mode_array; + value_size = sizeof(uint32_t)*16; + } else { + DRM_DEBUG_KMS("macrotile mode array is cik+ only!\n"); + return -EINVAL; + } + break; + case RADEON_INFO_SI_CP_DMA_COMPUTE: + *value = 1; + break; + case RADEON_INFO_SI_BACKEND_ENABLED_MASK: + if (rdev->family >= CHIP_BONAIRE) { + *value = rdev->config.cik.backend_enable_mask; + } else if (rdev->family >= CHIP_TAHITI) { + *value = rdev->config.si.backend_enable_mask; + } else { + DRM_DEBUG_KMS("BACKEND_ENABLED_MASK is si+ only!\n"); + } + break; + case RADEON_INFO_MAX_SCLK: + if ((rdev->pm.pm_method == PM_METHOD_DPM) && + rdev->pm.dpm_enabled) + *value = rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk * 10; + else + *value = rdev->pm.default_sclk * 10; + break; + case RADEON_INFO_VCE_FW_VERSION: + *value = rdev->vce.fw_version; + break; + case RADEON_INFO_VCE_FB_VERSION: + *value = rdev->vce.fb_version; + break; + case RADEON_INFO_NUM_BYTES_MOVED: + value = (uint32_t*)&value64; + value_size = sizeof(uint64_t); + value64 = atomic64_read(&rdev->num_bytes_moved); + break; + case RADEON_INFO_VRAM_USAGE: + value = (uint32_t*)&value64; + value_size = sizeof(uint64_t); + value64 = atomic64_read(&rdev->vram_usage); + break; + case RADEON_INFO_GTT_USAGE: + value = (uint32_t*)&value64; + value_size = sizeof(uint64_t); + value64 = atomic64_read(&rdev->gtt_usage); + break; + case RADEON_INFO_ACTIVE_CU_COUNT: + if (rdev->family >= CHIP_BONAIRE) + *value = rdev->config.cik.active_cus; + else if (rdev->family >= CHIP_TAHITI) + *value = rdev->config.si.active_cus; + else if (rdev->family >= CHIP_CAYMAN) + *value = rdev->config.cayman.active_simds; + else if (rdev->family >= CHIP_CEDAR) + *value = rdev->config.evergreen.active_simds; + else if (rdev->family >= CHIP_RV770) + *value = rdev->config.rv770.active_simds; + else if (rdev->family >= CHIP_R600) + *value = rdev->config.r600.active_simds; + else + *value = 1; + break; default: DRM_DEBUG_KMS("Invalid request %d\n", info->request); return -EINVAL; } - if (DRM_COPY_TO_USER(value_ptr, (char*)value, value_size)) { + if (copy_to_user(value_ptr, (char*)value, value_size)) { DRM_ERROR("copy_to_user %s:%u\n", __func__, __LINE__); return -EFAULT; } @@ -449,19 +545,6 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) * Outdated mess for old drm with Xorg being in charge (void function now). */ /** - * radeon_driver_firstopen_kms - drm callback for first open - * - * @dev: drm dev pointer - * - * Nothing to be done for KMS (all asics). - * Returns 0 on success. - */ -int radeon_driver_firstopen_kms(struct drm_device *dev) -{ - return 0; -} - -/** * radeon_driver_firstopen_kms - drm callback for last close * * @dev: drm dev pointer @@ -485,13 +568,18 @@ void radeon_driver_lastclose_kms(struct drm_device *dev) int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv) { struct radeon_device *rdev = dev->dev_private; + int r; file_priv->driver_priv = NULL; + r = pm_runtime_get_sync(dev->dev); + if (r < 0) + return r; + /* new gpu have virtual address space support */ if (rdev->family >= CHIP_CAYMAN) { struct radeon_fpriv *fpriv; - struct radeon_bo_va *bo_va; + struct radeon_vm *vm; int r; fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL); @@ -499,23 +587,42 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv) return -ENOMEM; } - radeon_vm_init(rdev, &fpriv->vm); - - /* map the ib pool buffer read only into - * virtual address space */ - bo_va = radeon_vm_bo_add(rdev, &fpriv->vm, - rdev->ring_tmp_bo.bo); - r = radeon_vm_bo_set_addr(rdev, bo_va, RADEON_VA_IB_OFFSET, - RADEON_VM_PAGE_READABLE | - RADEON_VM_PAGE_SNOOPED); + vm = &fpriv->vm; + r = radeon_vm_init(rdev, vm); if (r) { - radeon_vm_fini(rdev, &fpriv->vm); kfree(fpriv); return r; } + if (rdev->accel_working) { + r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false); + if (r) { + radeon_vm_fini(rdev, vm); + kfree(fpriv); + return r; + } + + /* map the ib pool buffer read only into + * virtual address space */ + vm->ib_bo_va = radeon_vm_bo_add(rdev, vm, + rdev->ring_tmp_bo.bo); + r = radeon_vm_bo_set_addr(rdev, vm->ib_bo_va, + RADEON_VA_IB_OFFSET, + RADEON_VM_PAGE_READABLE | + RADEON_VM_PAGE_SNOOPED); + + radeon_bo_unreserve(rdev->ring_tmp_bo.bo); + if (r) { + radeon_vm_fini(rdev, vm); + kfree(fpriv); + return r; + } + } file_priv->driver_priv = fpriv; } + + pm_runtime_mark_last_busy(dev->dev); + pm_runtime_put_autosuspend(dev->dev); return 0; } @@ -535,19 +642,19 @@ void radeon_driver_postclose_kms(struct drm_device *dev, /* new gpu have virtual address space support */ if (rdev->family >= CHIP_CAYMAN && file_priv->driver_priv) { struct radeon_fpriv *fpriv = file_priv->driver_priv; - struct radeon_bo_va *bo_va; + struct radeon_vm *vm = &fpriv->vm; int r; - r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false); - if (!r) { - bo_va = radeon_vm_bo_find(&fpriv->vm, - rdev->ring_tmp_bo.bo); - if (bo_va) - radeon_vm_bo_rmv(rdev, bo_va); - radeon_bo_unreserve(rdev->ring_tmp_bo.bo); + if (rdev->accel_working) { + r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false); + if (!r) { + if (vm->ib_bo_va) + radeon_vm_bo_rmv(rdev, vm->ib_bo_va); + radeon_bo_unreserve(rdev->ring_tmp_bo.bo); + } } - radeon_vm_fini(rdev, &fpriv->vm); + radeon_vm_fini(rdev, vm); kfree(fpriv); file_priv->driver_priv = NULL; } @@ -571,6 +678,7 @@ void radeon_driver_preclose_kms(struct drm_device *dev, if (rdev->cmask_filp == file_priv) rdev->cmask_filp = NULL; radeon_uvd_free_handles(rdev, file_priv); + radeon_vce_free_handles(rdev, file_priv); } /* @@ -680,21 +788,12 @@ int radeon_get_vblank_timestamp_kms(struct drm_device *dev, int crtc, /* Helper routine in DRM core does all the work: */ return drm_calc_vbltimestamp_from_scanoutpos(dev, crtc, max_error, vblank_time, flags, - drmcrtc); -} - -/* - * IOCTL. - */ -int radeon_dma_ioctl_kms(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - /* Not valid in KMS. */ - return -EINVAL; + drmcrtc, &drmcrtc->hwmode); } #define KMS_INVALID_IOCTL(name) \ -int name(struct drm_device *dev, void *data, struct drm_file *file_priv)\ +static int name(struct drm_device *dev, void *data, struct drm_file \ + *file_priv) \ { \ DRM_ERROR("invalid ioctl with kms %s\n", __func__); \ return -EINVAL; \ @@ -732,7 +831,7 @@ KMS_INVALID_IOCTL(radeon_surface_alloc_kms) KMS_INVALID_IOCTL(radeon_surface_free_kms) -struct drm_ioctl_desc radeon_ioctls_kms[] = { +const struct drm_ioctl_desc radeon_ioctls_kms[] = { DRM_IOCTL_DEF_DRV(RADEON_CP_INIT, radeon_cp_init_kms, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), DRM_IOCTL_DEF_DRV(RADEON_CP_START, radeon_cp_start_kms, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), DRM_IOCTL_DEF_DRV(RADEON_CP_STOP, radeon_cp_stop_kms, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), @@ -761,18 +860,19 @@ struct drm_ioctl_desc radeon_ioctls_kms[] = { DRM_IOCTL_DEF_DRV(RADEON_SURF_ALLOC, radeon_surface_alloc_kms, DRM_AUTH), DRM_IOCTL_DEF_DRV(RADEON_SURF_FREE, radeon_surface_free_kms, DRM_AUTH), /* KMS */ - DRM_IOCTL_DEF_DRV(RADEON_GEM_INFO, radeon_gem_info_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF_DRV(RADEON_GEM_CREATE, radeon_gem_create_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF_DRV(RADEON_GEM_MMAP, radeon_gem_mmap_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF_DRV(RADEON_GEM_SET_DOMAIN, radeon_gem_set_domain_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF_DRV(RADEON_GEM_INFO, radeon_gem_info_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_GEM_CREATE, radeon_gem_create_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_GEM_MMAP, radeon_gem_mmap_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_GEM_SET_DOMAIN, radeon_gem_set_domain_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(RADEON_GEM_PREAD, radeon_gem_pread_ioctl, DRM_AUTH|DRM_UNLOCKED), DRM_IOCTL_DEF_DRV(RADEON_GEM_PWRITE, radeon_gem_pwrite_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF_DRV(RADEON_GEM_WAIT_IDLE, radeon_gem_wait_idle_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF_DRV(RADEON_CS, radeon_cs_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF_DRV(RADEON_INFO, radeon_info_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF_DRV(RADEON_GEM_SET_TILING, radeon_gem_set_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF_DRV(RADEON_GEM_GET_TILING, radeon_gem_get_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF_DRV(RADEON_GEM_BUSY, radeon_gem_busy_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF_DRV(RADEON_GEM_VA, radeon_gem_va_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF_DRV(RADEON_GEM_WAIT_IDLE, radeon_gem_wait_idle_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_CS, radeon_cs_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_INFO, radeon_info_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_GEM_SET_TILING, radeon_gem_set_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_GEM_GET_TILING, radeon_gem_get_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_GEM_BUSY, radeon_gem_busy_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_GEM_VA, radeon_gem_va_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_GEM_OP, radeon_gem_op_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), }; -int radeon_max_kms_ioctl = DRM_ARRAY_SIZE(radeon_ioctls_kms); +int radeon_max_kms_ioctl = ARRAY_SIZE(radeon_ioctls_kms); diff --git a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c index 7cb178a34a0..cafb1ccf2ec 100644 --- a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c +++ b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c @@ -385,7 +385,7 @@ int radeon_crtc_do_set_base(struct drm_crtc *crtc, DRM_DEBUG_KMS("\n"); /* no fb bound */ - if (!atomic && !crtc->fb) { + if (!atomic && !crtc->primary->fb) { DRM_DEBUG_KMS("No FB bound\n"); return 0; } @@ -395,8 +395,8 @@ int radeon_crtc_do_set_base(struct drm_crtc *crtc, target_fb = fb; } else { - radeon_fb = to_radeon_framebuffer(crtc->fb); - target_fb = crtc->fb; + radeon_fb = to_radeon_framebuffer(crtc->primary->fb); + target_fb = crtc->primary->fb; } switch (target_fb->bits_per_pixel) { @@ -422,6 +422,7 @@ int radeon_crtc_do_set_base(struct drm_crtc *crtc, /* Pin framebuffer & get tilling informations */ obj = radeon_fb->obj; rbo = gem_to_radeon_bo(obj); +retry: r = radeon_bo_reserve(rbo, false); if (unlikely(r != 0)) return r; @@ -430,6 +431,33 @@ int radeon_crtc_do_set_base(struct drm_crtc *crtc, &base); if (unlikely(r != 0)) { radeon_bo_unreserve(rbo); + + /* On old GPU like RN50 with little vram pining can fails because + * current fb is taking all space needed. So instead of unpining + * the old buffer after pining the new one, first unpin old one + * and then retry pining new one. + * + * As only master can set mode only master can pin and it is + * unlikely the master client will race with itself especialy + * on those old gpu with single crtc. + * + * We don't shutdown the display controller because new buffer + * will end up in same spot. + */ + if (!atomic && fb && fb != crtc->primary->fb) { + struct radeon_bo *old_rbo; + unsigned long nsize, osize; + + old_rbo = gem_to_radeon_bo(to_radeon_framebuffer(fb)->obj); + osize = radeon_bo_size(old_rbo); + nsize = radeon_bo_size(rbo); + if (nsize <= osize && !radeon_bo_reserve(old_rbo, false)) { + radeon_bo_unpin(old_rbo); + radeon_bo_unreserve(old_rbo); + fb = NULL; + goto retry; + } + } return -EINVAL; } radeon_bo_get_tiling_flags(rbo, &tiling_flags, NULL); @@ -527,7 +555,7 @@ int radeon_crtc_do_set_base(struct drm_crtc *crtc, WREG32(RADEON_CRTC_OFFSET + radeon_crtc->crtc_offset, crtc_offset); WREG32(RADEON_CRTC_PITCH + radeon_crtc->crtc_offset, crtc_pitch); - if (!atomic && fb && fb != crtc->fb) { + if (!atomic && fb && fb != crtc->primary->fb) { radeon_fb = to_radeon_framebuffer(fb); rbo = gem_to_radeon_bo(radeon_fb->obj); r = radeon_bo_reserve(rbo, false); @@ -571,7 +599,7 @@ static bool radeon_set_crtc_timing(struct drm_crtc *crtc, struct drm_display_mod } } - switch (crtc->fb->bits_per_pixel) { + switch (crtc->primary->fb->bits_per_pixel) { case 8: format = 2; break; @@ -1056,6 +1084,26 @@ static void radeon_crtc_commit(struct drm_crtc *crtc) } } +static void radeon_crtc_disable(struct drm_crtc *crtc) +{ + radeon_crtc_dpms(crtc, DRM_MODE_DPMS_OFF); + if (crtc->primary->fb) { + int r; + struct radeon_framebuffer *radeon_fb; + struct radeon_bo *rbo; + + radeon_fb = to_radeon_framebuffer(crtc->primary->fb); + rbo = gem_to_radeon_bo(radeon_fb->obj); + r = radeon_bo_reserve(rbo, false); + if (unlikely(r)) + DRM_ERROR("failed to reserve rbo before unpin\n"); + else { + radeon_bo_unpin(rbo); + radeon_bo_unreserve(rbo); + } + } +} + static const struct drm_crtc_helper_funcs legacy_helper_funcs = { .dpms = radeon_crtc_dpms, .mode_fixup = radeon_crtc_mode_fixup, @@ -1065,6 +1113,7 @@ static const struct drm_crtc_helper_funcs legacy_helper_funcs = { .prepare = radeon_crtc_prepare, .commit = radeon_crtc_commit, .load_lut = radeon_crtc_load_lut, + .disable = radeon_crtc_disable }; diff --git a/drivers/gpu/drm/radeon/radeon_legacy_encoders.c b/drivers/gpu/drm/radeon/radeon_legacy_encoders.c index 62cd512f5c8..c89971d904c 100644 --- a/drivers/gpu/drm/radeon/radeon_legacy_encoders.c +++ b/drivers/gpu/drm/radeon/radeon_legacy_encoders.c @@ -392,7 +392,7 @@ void radeon_legacy_backlight_init(struct radeon_encoder *radeon_encoder, props.type = BACKLIGHT_RAW; snprintf(bl_name, sizeof(bl_name), "radeon_bl%d", dev->primary->index); - bd = backlight_device_register(bl_name, &drm_connector->kdev, + bd = backlight_device_register(bl_name, drm_connector->kdev, pdata, &radeon_backlight_ops, &props); if (IS_ERR(bd)) { DRM_ERROR("Backlight registration failed\n"); diff --git a/drivers/gpu/drm/radeon/radeon_mem.c b/drivers/gpu/drm/radeon/radeon_mem.c index d54d2d7c903..146d253f113 100644 --- a/drivers/gpu/drm/radeon/radeon_mem.c +++ b/drivers/gpu/drm/radeon/radeon_mem.c @@ -243,7 +243,7 @@ int radeon_mem_alloc(struct drm_device *dev, void *data, struct drm_file *file_p if (!block) return -ENOMEM; - if (DRM_COPY_TO_USER(alloc->region_offset, &block->start, + if (copy_to_user(alloc->region_offset, &block->start, sizeof(int))) { DRM_ERROR("copy_to_user\n"); return -EFAULT; diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h index 8296632a423..0592ddb0904 100644 --- a/drivers/gpu/drm/radeon/radeon_mode.h +++ b/drivers/gpu/drm/radeon/radeon_mode.h @@ -46,6 +46,10 @@ struct radeon_device; #define to_radeon_encoder(x) container_of(x, struct radeon_encoder, base) #define to_radeon_framebuffer(x) container_of(x, struct radeon_framebuffer, base) +#define RADEON_MAX_HPD_PINS 7 +#define RADEON_MAX_CRTCS 6 +#define RADEON_MAX_AFMT_BLOCKS 7 + enum radeon_rmx_type { RMX_OFF, RMX_FULL, @@ -187,11 +191,11 @@ struct radeon_pll { struct radeon_i2c_chan { struct i2c_adapter adapter; struct drm_device *dev; - union { - struct i2c_algo_bit_data bit; - struct i2c_algo_dp_aux_data dp; - } algo; + struct i2c_algo_bit_data bit; struct radeon_i2c_bus_rec rec; + struct drm_dp_aux aux; + bool has_aux; + struct mutex mutex; }; /* mostly for macs, but really any system without connector tables */ @@ -225,6 +229,7 @@ struct radeon_afmt { int offset; bool last_buffer_filled_status; int id; + struct r600_audio_pin *pin; }; struct radeon_mode_info { @@ -232,8 +237,8 @@ struct radeon_mode_info { struct card_info *atom_card_info; enum radeon_connector_table connector_table; bool mode_config_initialized; - struct radeon_crtc *crtcs[6]; - struct radeon_afmt *afmt[6]; + struct radeon_crtc *crtcs[RADEON_MAX_CRTCS]; + struct radeon_afmt *afmt[RADEON_MAX_AFMT_BLOCKS]; /* DVI-I properties */ struct drm_property *coherent_mode_property; /* DAC enable load detect */ @@ -246,6 +251,10 @@ struct radeon_mode_info { struct drm_property *underscan_property; struct drm_property *underscan_hborder_property; struct drm_property *underscan_vborder_property; + /* audio */ + struct drm_property *audio_property; + /* FMT dithering */ + struct drm_property *dither_property; /* hardcoded DFP edid from BIOS */ struct edid *bios_hardcoded_edid; int bios_hardcoded_edid_size; @@ -286,6 +295,7 @@ struct radeon_tv_regs { struct radeon_atom_ss { uint16_t percentage; + uint16_t percentage_divider; uint8_t type; uint16_t step; uint8_t delay; @@ -296,6 +306,12 @@ struct radeon_atom_ss { uint16_t amount; }; +enum radeon_flip_status { + RADEON_FLIP_NONE, + RADEON_FLIP_PENDING, + RADEON_FLIP_SUBMITTED +}; + struct radeon_crtc { struct drm_crtc base; int crtc_id; @@ -319,8 +335,9 @@ struct radeon_crtc { struct drm_display_mode native_mode; int pll_id; /* page flipping */ - struct radeon_unpin_work *unpin_work; - int deferred_flip_completion; + struct workqueue_struct *flip_queue; + struct radeon_flip_work *flip_work; + enum radeon_flip_status flip_status; /* pll sharing */ struct radeon_atom_ss ss; bool ss_enabled; @@ -433,7 +450,6 @@ struct radeon_encoder { struct radeon_connector_atom_dig { uint32_t igp_lane_info; /* displayport */ - struct radeon_i2c_chan *dp_i2c_bus; u8 dpcd[DP_RECEIVER_CAP_SIZE]; u8 dp_sink_type; int dp_clock; @@ -470,6 +486,17 @@ struct radeon_router { u8 cd_mux_state; }; +enum radeon_connector_audio { + RADEON_AUDIO_DISABLE = 0, + RADEON_AUDIO_ENABLE = 1, + RADEON_AUDIO_AUTO = 2 +}; + +enum radeon_connector_dither { + RADEON_FMT_DITHER_DISABLE = 0, + RADEON_FMT_DITHER_ENABLE = 1, +}; + struct radeon_connector { struct drm_connector base; uint32_t connector_id; @@ -488,6 +515,9 @@ struct radeon_connector { struct radeon_hpd hpd; struct radeon_router router; struct radeon_i2c_chan *router_bus; + enum radeon_connector_audio audio; + enum radeon_connector_dither dither; + int pixelclock_for_modeset; }; struct radeon_framebuffer { @@ -606,6 +636,30 @@ struct atom_voltage_table struct atom_voltage_table_entry entries[MAX_VOLTAGE_ENTRIES]; }; + +extern void +radeon_add_atom_connector(struct drm_device *dev, + uint32_t connector_id, + uint32_t supported_device, + int connector_type, + struct radeon_i2c_bus_rec *i2c_bus, + uint32_t igp_lane_info, + uint16_t connector_object_id, + struct radeon_hpd *hpd, + struct radeon_router *router); +extern void +radeon_add_legacy_connector(struct drm_device *dev, + uint32_t connector_id, + uint32_t supported_device, + int connector_type, + struct radeon_i2c_bus_rec *i2c_bus, + uint16_t connector_object_id, + struct radeon_hpd *hpd); +extern uint32_t +radeon_get_encoder_enum(struct drm_device *dev, uint32_t supported_device, + uint8_t dac); +extern void radeon_link_encoder_connector(struct drm_device *dev); + extern enum radeon_tv_std radeon_combios_get_tv_info(struct radeon_device *rdev); extern enum radeon_tv_std @@ -613,6 +667,15 @@ radeon_atombios_get_tv_info(struct radeon_device *rdev); extern void radeon_atombios_get_default_voltages(struct radeon_device *rdev, u16 *vddc, u16 *vddci, u16 *mvdd); +extern void +radeon_combios_connected_scratch_regs(struct drm_connector *connector, + struct drm_encoder *encoder, + bool connected); +extern void +radeon_atombios_connected_scratch_regs(struct drm_connector *connector, + struct drm_encoder *encoder, + bool connected); + extern struct drm_connector * radeon_get_connector_for_encoder(struct drm_encoder *encoder); extern struct drm_connector * @@ -638,6 +701,9 @@ extern u8 radeon_dp_getsinktype(struct radeon_connector *radeon_connector); extern bool radeon_dp_getdpcd(struct radeon_connector *radeon_connector); extern int radeon_dp_get_panel_mode(struct drm_encoder *encoder, struct drm_connector *connector); +extern void radeon_dp_set_rx_power_state(struct drm_connector *connector, + u8 power_state); +extern void radeon_dp_aux_init(struct radeon_connector *radeon_connector); extern void atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mode); extern void radeon_atom_encoder_init(struct radeon_device *rdev); extern void radeon_atom_disp_eng_pll_init(struct radeon_device *rdev); @@ -646,8 +712,7 @@ extern void atombios_dig_transmitter_setup(struct drm_encoder *encoder, uint8_t lane_set); extern void radeon_atom_ext_encoder_setup_ddc(struct drm_encoder *encoder); extern struct drm_encoder *radeon_get_external_encoder(struct drm_encoder *encoder); -extern int radeon_dp_i2c_aux_ch(struct i2c_adapter *adapter, int mode, - u8 write_byte, u8 *read_byte); +void radeon_atom_copy_swap(u8 *dst, u8 *src, u8 num_bytes, bool to_le); extern void radeon_i2c_init(struct radeon_device *rdev); extern void radeon_i2c_fini(struct radeon_device *rdev); @@ -658,9 +723,6 @@ extern void radeon_i2c_add(struct radeon_device *rdev, const char *name); extern struct radeon_i2c_chan *radeon_i2c_lookup(struct radeon_device *rdev, struct radeon_i2c_bus_rec *i2c_bus); -extern struct radeon_i2c_chan *radeon_i2c_create_dp(struct drm_device *dev, - struct radeon_i2c_bus_rec *rec, - const char *name); extern struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev, struct radeon_i2c_bus_rec *rec, const char *name); @@ -748,7 +810,9 @@ extern int radeon_crtc_cursor_move(struct drm_crtc *crtc, int x, int y); extern int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, - int *vpos, int *hpos); + unsigned int flags, + int *vpos, int *hpos, ktime_t *stime, + ktime_t *etime); extern bool radeon_combios_check_hardcoded_edid(struct radeon_device *rdev); extern struct edid * @@ -840,6 +904,12 @@ void radeon_legacy_tv_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode); +/* fmt blocks */ +void avivo_program_fmt(struct drm_encoder *encoder); +void dce3_program_fmt(struct drm_encoder *encoder); +void dce4_program_fmt(struct drm_encoder *encoder); +void dce8_program_fmt(struct drm_encoder *encoder); + /* fbdev layer */ int radeon_fbdev_init(struct radeon_device *rdev); void radeon_fbdev_fini(struct radeon_device *rdev); @@ -849,6 +919,7 @@ bool radeon_fbdev_robj_is_fb(struct radeon_device *rdev, struct radeon_bo *robj) void radeon_fb_output_poll_changed(struct radeon_device *rdev); +void radeon_crtc_handle_vblank(struct radeon_device *rdev, int crtc_id); void radeon_crtc_handle_flip(struct radeon_device *rdev, int crtc_id); int radeon_align_pitch(struct radeon_device *rdev, int width, int bpp, bool tiled); diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index 0219d263e2d..6c717b257d6 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -46,7 +46,7 @@ static void radeon_bo_clear_surface_reg(struct radeon_bo *bo); * function are calling it. */ -void radeon_bo_clear_va(struct radeon_bo *bo) +static void radeon_bo_clear_va(struct radeon_bo *bo) { struct radeon_bo_va *bo_va, *tmp; @@ -56,11 +56,36 @@ void radeon_bo_clear_va(struct radeon_bo *bo) } } +static void radeon_update_memory_usage(struct radeon_bo *bo, + unsigned mem_type, int sign) +{ + struct radeon_device *rdev = bo->rdev; + u64 size = (u64)bo->tbo.num_pages << PAGE_SHIFT; + + switch (mem_type) { + case TTM_PL_TT: + if (sign > 0) + atomic64_add(size, &rdev->gtt_usage); + else + atomic64_sub(size, &rdev->gtt_usage); + break; + case TTM_PL_VRAM: + if (sign > 0) + atomic64_add(size, &rdev->vram_usage); + else + atomic64_sub(size, &rdev->vram_usage); + break; + } +} + static void radeon_ttm_bo_destroy(struct ttm_buffer_object *tbo) { struct radeon_bo *bo; bo = container_of(tbo, struct radeon_bo, tbo); + + radeon_update_memory_usage(bo, bo->tbo.mem.mem_type, -1); + mutex_lock(&bo->rdev->gem.mutex); list_del_init(&bo->list); mutex_unlock(&bo->rdev->gem.mutex); @@ -79,7 +104,7 @@ bool radeon_ttm_bo_is_radeon_bo(struct ttm_buffer_object *bo) void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain) { - u32 c = 0; + u32 c = 0, i; rbo->placement.fpfn = 0; rbo->placement.lpfn = 0; @@ -106,6 +131,17 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain) rbo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM; rbo->placement.num_placement = c; rbo->placement.num_busy_placement = c; + + /* + * Use two-ended allocation depending on the buffer size to + * improve fragmentation quality. + * 512kb was measured as the most optimal number. + */ + if (rbo->tbo.mem.size > 512 * 1024) { + for (i = 0; i < c; i++) { + rbo->placements[i] |= TTM_PL_FLAG_TOPDOWN; + } + } } int radeon_bo_create(struct radeon_device *rdev, @@ -120,7 +156,6 @@ int radeon_bo_create(struct radeon_device *rdev, size = ALIGN(size, PAGE_SIZE); - rdev->mman.bdev.dev_mapping = rdev->ddev->dev_mapping; if (kernel) { type = ttm_bo_type_kernel; } else if (sg) { @@ -142,10 +177,12 @@ int radeon_bo_create(struct radeon_device *rdev, return r; } bo->rdev = rdev; - bo->gem_base.driver_private = NULL; bo->surface_reg = -1; INIT_LIST_HEAD(&bo->list); INIT_LIST_HEAD(&bo->va); + bo->initial_domain = domain & (RADEON_GEM_DOMAIN_VRAM | + RADEON_GEM_DOMAIN_GTT | + RADEON_GEM_DOMAIN_CPU); radeon_ttm_placement_from_domain(bo, domain); /* Kernel allocation are uninterruptible */ down_read(&rdev->pm.mclk_lock); @@ -339,44 +376,112 @@ void radeon_bo_fini(struct radeon_device *rdev) arch_phys_wc_del(rdev->mc.vram_mtrr); } -void radeon_bo_list_add_object(struct radeon_bo_list *lobj, - struct list_head *head) +/* Returns how many bytes TTM can move per IB. + */ +static u64 radeon_bo_get_threshold_for_moves(struct radeon_device *rdev) { - if (lobj->written) { - list_add(&lobj->tv.head, head); - } else { - list_add_tail(&lobj->tv.head, head); - } + u64 real_vram_size = rdev->mc.real_vram_size; + u64 vram_usage = atomic64_read(&rdev->vram_usage); + + /* This function is based on the current VRAM usage. + * + * - If all of VRAM is free, allow relocating the number of bytes that + * is equal to 1/4 of the size of VRAM for this IB. + + * - If more than one half of VRAM is occupied, only allow relocating + * 1 MB of data for this IB. + * + * - From 0 to one half of used VRAM, the threshold decreases + * linearly. + * __________________ + * 1/4 of -|\ | + * VRAM | \ | + * | \ | + * | \ | + * | \ | + * | \ | + * | \ | + * | \________|1 MB + * |----------------| + * VRAM 0 % 100 % + * used used + * + * Note: It's a threshold, not a limit. The threshold must be crossed + * for buffer relocations to stop, so any buffer of an arbitrary size + * can be moved as long as the threshold isn't crossed before + * the relocation takes place. We don't want to disable buffer + * relocations completely. + * + * The idea is that buffers should be placed in VRAM at creation time + * and TTM should only do a minimum number of relocations during + * command submission. In practice, you need to submit at least + * a dozen IBs to move all buffers to VRAM if they are in GTT. + * + * Also, things can get pretty crazy under memory pressure and actual + * VRAM usage can change a lot, so playing safe even at 50% does + * consistently increase performance. + */ + + u64 half_vram = real_vram_size >> 1; + u64 half_free_vram = vram_usage >= half_vram ? 0 : half_vram - vram_usage; + u64 bytes_moved_threshold = half_free_vram >> 1; + return max(bytes_moved_threshold, 1024*1024ull); } -int radeon_bo_list_validate(struct ww_acquire_ctx *ticket, +int radeon_bo_list_validate(struct radeon_device *rdev, + struct ww_acquire_ctx *ticket, struct list_head *head, int ring) { - struct radeon_bo_list *lobj; + struct radeon_cs_reloc *lobj; struct radeon_bo *bo; - u32 domain; int r; + u64 bytes_moved = 0, initial_bytes_moved; + u64 bytes_moved_threshold = radeon_bo_get_threshold_for_moves(rdev); r = ttm_eu_reserve_buffers(ticket, head); if (unlikely(r != 0)) { return r; } + list_for_each_entry(lobj, head, tv.head) { - bo = lobj->bo; + bo = lobj->robj; if (!bo->pin_count) { - domain = lobj->domain; - + u32 domain = lobj->prefered_domains; + u32 current_domain = + radeon_mem_type_to_domain(bo->tbo.mem.mem_type); + + /* Check if this buffer will be moved and don't move it + * if we have moved too many buffers for this IB already. + * + * Note that this allows moving at least one buffer of + * any size, because it doesn't take the current "bo" + * into account. We don't want to disallow buffer moves + * completely. + */ + if ((lobj->allowed_domains & current_domain) != 0 && + (domain & current_domain) == 0 && /* will be moved */ + bytes_moved > bytes_moved_threshold) { + /* don't move it */ + domain = current_domain; + } + retry: radeon_ttm_placement_from_domain(bo, domain); if (ring == R600_RING_TYPE_UVD_INDEX) radeon_uvd_force_into_uvd_segment(bo); - r = ttm_bo_validate(&bo->tbo, &bo->placement, - true, false); + + initial_bytes_moved = atomic64_read(&rdev->num_bytes_moved); + r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); + bytes_moved += atomic64_read(&rdev->num_bytes_moved) - + initial_bytes_moved; + if (unlikely(r)) { - if (r != -ERESTARTSYS && domain != lobj->alt_domain) { - domain = lobj->alt_domain; + if (r != -ERESTARTSYS && + domain != lobj->allowed_domains) { + domain = lobj->allowed_domains; goto retry; } + ttm_eu_backoff_reservation(ticket, head); return r; } } @@ -564,14 +669,23 @@ int radeon_bo_check_tiling(struct radeon_bo *bo, bool has_moved, } void radeon_bo_move_notify(struct ttm_buffer_object *bo, - struct ttm_mem_reg *mem) + struct ttm_mem_reg *new_mem) { struct radeon_bo *rbo; + if (!radeon_ttm_bo_is_radeon_bo(bo)) return; + rbo = container_of(bo, struct radeon_bo, tbo); radeon_bo_check_tiling(rbo, 0, 1); radeon_vm_bo_invalidate(rbo->rdev, rbo); + + /* update statistics */ + if (!new_mem) + return; + + radeon_update_memory_usage(rbo, bo->mem.mem_type, -1); + radeon_update_memory_usage(rbo, new_mem->mem_type, 1); } int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo) @@ -586,22 +700,30 @@ int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo) rbo = container_of(bo, struct radeon_bo, tbo); radeon_bo_check_tiling(rbo, 0, 0); rdev = rbo->rdev; - if (bo->mem.mem_type == TTM_PL_VRAM) { - size = bo->mem.num_pages << PAGE_SHIFT; - offset = bo->mem.start << PAGE_SHIFT; - if ((offset + size) > rdev->mc.visible_vram_size) { - /* hurrah the memory is not visible ! */ - radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_VRAM); - rbo->placement.lpfn = rdev->mc.visible_vram_size >> PAGE_SHIFT; - r = ttm_bo_validate(bo, &rbo->placement, false, false); - if (unlikely(r != 0)) - return r; - offset = bo->mem.start << PAGE_SHIFT; - /* this should not happen */ - if ((offset + size) > rdev->mc.visible_vram_size) - return -EINVAL; - } + if (bo->mem.mem_type != TTM_PL_VRAM) + return 0; + + size = bo->mem.num_pages << PAGE_SHIFT; + offset = bo->mem.start << PAGE_SHIFT; + if ((offset + size) <= rdev->mc.visible_vram_size) + return 0; + + /* hurrah the memory is not visible ! */ + radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_VRAM); + rbo->placement.lpfn = rdev->mc.visible_vram_size >> PAGE_SHIFT; + r = ttm_bo_validate(bo, &rbo->placement, false, false); + if (unlikely(r == -ENOMEM)) { + radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_GTT); + return ttm_bo_validate(bo, &rbo->placement, false, false); + } else if (unlikely(r != 0)) { + return r; } + + offset = bo->mem.start << PAGE_SHIFT; + /* this should never happen */ + if ((offset + size) > rdev->mc.visible_vram_size) + return -EINVAL; + return 0; } @@ -609,7 +731,7 @@ int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type, bool no_wait) { int r; - r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, 0); + r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, NULL); if (unlikely(r != 0)) return r; spin_lock(&bo->tbo.bdev->fence_lock); diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h index 91519a5622b..5a873f31a17 100644 --- a/drivers/gpu/drm/radeon/radeon_object.h +++ b/drivers/gpu/drm/radeon/radeon_object.h @@ -65,7 +65,7 @@ static inline int radeon_bo_reserve(struct radeon_bo *bo, bool no_intr) { int r; - r = ttm_bo_reserve(&bo->tbo, !no_intr, false, false, 0); + r = ttm_bo_reserve(&bo->tbo, !no_intr, false, false, NULL); if (unlikely(r != 0)) { if (r != -ERESTARTSYS) dev_err(bo->rdev->dev, "%p reserve failed\n", bo); @@ -113,13 +113,10 @@ static inline unsigned radeon_bo_gpu_page_alignment(struct radeon_bo *bo) * @bo: radeon object for which we query the offset * * Returns mmap offset of the object. - * - * Note: addr_space_offset is constant after ttm bo init thus isn't protected - * by any lock. */ static inline u64 radeon_bo_mmap_offset(struct radeon_bo *bo) { - return bo->tbo.addr_space_offset; + return drm_vma_node_offset_addr(&bo->tbo.vma_node); } extern int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type, @@ -141,9 +138,8 @@ extern int radeon_bo_evict_vram(struct radeon_device *rdev); extern void radeon_bo_force_delete(struct radeon_device *rdev); extern int radeon_bo_init(struct radeon_device *rdev); extern void radeon_bo_fini(struct radeon_device *rdev); -extern void radeon_bo_list_add_object(struct radeon_bo_list *lobj, - struct list_head *head); -extern int radeon_bo_list_validate(struct ww_acquire_ctx *ticket, +extern int radeon_bo_list_validate(struct radeon_device *rdev, + struct ww_acquire_ctx *ticket, struct list_head *head, int ring); extern int radeon_bo_fbdev_mmap(struct radeon_bo *bo, struct vm_area_struct *vma); @@ -154,7 +150,7 @@ extern void radeon_bo_get_tiling_flags(struct radeon_bo *bo, extern int radeon_bo_check_tiling(struct radeon_bo *bo, bool has_moved, bool force_drop); extern void radeon_bo_move_notify(struct ttm_buffer_object *bo, - struct ttm_mem_reg *mem); + struct ttm_mem_reg *new_mem); extern int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo); extern int radeon_bo_get_surface_reg(struct radeon_bo *bo); @@ -174,7 +170,7 @@ static inline void * radeon_sa_bo_cpu_addr(struct radeon_sa_bo *sa_bo) extern int radeon_sa_bo_manager_init(struct radeon_device *rdev, struct radeon_sa_manager *sa_manager, - unsigned size, u32 domain); + unsigned size, u32 align, u32 domain); extern void radeon_sa_bo_manager_fini(struct radeon_device *rdev, struct radeon_sa_manager *sa_manager); extern int radeon_sa_bo_manager_start(struct radeon_device *rdev, @@ -184,7 +180,7 @@ extern int radeon_sa_bo_manager_suspend(struct radeon_device *rdev, extern int radeon_sa_bo_new(struct radeon_device *rdev, struct radeon_sa_manager *sa_manager, struct radeon_sa_bo **sa_bo, - unsigned size, unsigned align, bool block); + unsigned size, unsigned align); extern void radeon_sa_bo_free(struct radeon_device *rdev, struct radeon_sa_bo **sa_bo, struct radeon_fence *fence); diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c index f374c467aac..e447e390d09 100644 --- a/drivers/gpu/drm/radeon/radeon_pm.c +++ b/drivers/gpu/drm/radeon/radeon_pm.c @@ -67,7 +67,18 @@ int radeon_pm_get_type_index(struct radeon_device *rdev, void radeon_pm_acpi_event_handler(struct radeon_device *rdev) { - if (rdev->pm.pm_method == PM_METHOD_PROFILE) { + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { + mutex_lock(&rdev->pm.mutex); + if (power_supply_is_system_supplied() > 0) + rdev->pm.dpm.ac_power = true; + else + rdev->pm.dpm.ac_power = false; + if (rdev->family == CHIP_ARUBA) { + if (rdev->asic->dpm.enable_bapm) + radeon_dpm_enable_bapm(rdev, rdev->pm.dpm.ac_power); + } + mutex_unlock(&rdev->pm.mutex); + } else if (rdev->pm.pm_method == PM_METHOD_PROFILE) { if (rdev->pm.profile == PM_PROFILE_AUTO) { mutex_lock(&rdev->pm.mutex); radeon_pm_update_profile(rdev); @@ -251,7 +262,7 @@ static void radeon_pm_set_clocks(struct radeon_device *rdev) if (!ring->ready) { continue; } - r = radeon_fence_wait_empty_locked(rdev, i); + r = radeon_fence_wait_empty(rdev, i); if (r) { /* needs a GPU reset dont reset here */ mutex_unlock(&rdev->ring_lock); @@ -333,7 +344,7 @@ static ssize_t radeon_get_pm_profile(struct device *dev, struct device_attribute *attr, char *buf) { - struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev)); + struct drm_device *ddev = dev_get_drvdata(dev); struct radeon_device *rdev = ddev->dev_private; int cp = rdev->pm.profile; @@ -349,9 +360,14 @@ static ssize_t radeon_set_pm_profile(struct device *dev, const char *buf, size_t count) { - struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev)); + struct drm_device *ddev = dev_get_drvdata(dev); struct radeon_device *rdev = ddev->dev_private; + /* Can't set profile when the card is off */ + if ((rdev->flags & RADEON_IS_PX) && + (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return -EINVAL; + mutex_lock(&rdev->pm.mutex); if (rdev->pm.pm_method == PM_METHOD_PROFILE) { if (strncmp("default", buf, strlen("default")) == 0) @@ -383,7 +399,7 @@ static ssize_t radeon_get_pm_method(struct device *dev, struct device_attribute *attr, char *buf) { - struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev)); + struct drm_device *ddev = dev_get_drvdata(dev); struct radeon_device *rdev = ddev->dev_private; int pm = rdev->pm.pm_method; @@ -397,9 +413,16 @@ static ssize_t radeon_set_pm_method(struct device *dev, const char *buf, size_t count) { - struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev)); + struct drm_device *ddev = dev_get_drvdata(dev); struct radeon_device *rdev = ddev->dev_private; + /* Can't set method when the card is off */ + if ((rdev->flags & RADEON_IS_PX) && + (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) { + count = -EINVAL; + goto fail; + } + /* we don't support the legacy modes with dpm */ if (rdev->pm.pm_method == PM_METHOD_DPM) { count = -EINVAL; @@ -433,10 +456,14 @@ static ssize_t radeon_get_dpm_state(struct device *dev, struct device_attribute *attr, char *buf) { - struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev)); + struct drm_device *ddev = dev_get_drvdata(dev); struct radeon_device *rdev = ddev->dev_private; enum radeon_pm_state_type pm = rdev->pm.dpm.user_state; + if ((rdev->flags & RADEON_IS_PX) && + (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return snprintf(buf, PAGE_SIZE, "off\n"); + return snprintf(buf, PAGE_SIZE, "%s\n", (pm == POWER_STATE_TYPE_BATTERY) ? "battery" : (pm == POWER_STATE_TYPE_BALANCED) ? "balanced" : "performance"); @@ -447,9 +474,14 @@ static ssize_t radeon_set_dpm_state(struct device *dev, const char *buf, size_t count) { - struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev)); + struct drm_device *ddev = dev_get_drvdata(dev); struct radeon_device *rdev = ddev->dev_private; + /* Can't set dpm state when the card is off */ + if ((rdev->flags & RADEON_IS_PX) && + (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return -EINVAL; + mutex_lock(&rdev->pm.mutex); if (strncmp("battery", buf, strlen("battery")) == 0) rdev->pm.dpm.user_state = POWER_STATE_TYPE_BATTERY; @@ -472,10 +504,14 @@ static ssize_t radeon_get_dpm_forced_performance_level(struct device *dev, struct device_attribute *attr, char *buf) { - struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev)); + struct drm_device *ddev = dev_get_drvdata(dev); struct radeon_device *rdev = ddev->dev_private; enum radeon_dpm_forced_level level = rdev->pm.dpm.forced_level; + if ((rdev->flags & RADEON_IS_PX) && + (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return snprintf(buf, PAGE_SIZE, "off\n"); + return snprintf(buf, PAGE_SIZE, "%s\n", (level == RADEON_DPM_FORCED_LEVEL_AUTO) ? "auto" : (level == RADEON_DPM_FORCED_LEVEL_LOW) ? "low" : "high"); @@ -486,11 +522,16 @@ static ssize_t radeon_set_dpm_forced_performance_level(struct device *dev, const char *buf, size_t count) { - struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev)); + struct drm_device *ddev = dev_get_drvdata(dev); struct radeon_device *rdev = ddev->dev_private; enum radeon_dpm_forced_level level; int ret = 0; + /* Can't force performance level when the card is off */ + if ((rdev->flags & RADEON_IS_PX) && + (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return -EINVAL; + mutex_lock(&rdev->pm.mutex); if (strncmp("low", buf, strlen("low")) == 0) { level = RADEON_DPM_FORCED_LEVEL_LOW; @@ -499,17 +540,21 @@ static ssize_t radeon_set_dpm_forced_performance_level(struct device *dev, } else if (strncmp("auto", buf, strlen("auto")) == 0) { level = RADEON_DPM_FORCED_LEVEL_AUTO; } else { - mutex_unlock(&rdev->pm.mutex); count = -EINVAL; goto fail; } if (rdev->asic->dpm.force_performance_level) { + if (rdev->pm.dpm.thermal_active) { + count = -EINVAL; + goto fail; + } ret = radeon_dpm_force_performance_level(rdev, level); if (ret) count = -EINVAL; } - mutex_unlock(&rdev->pm.mutex); fail: + mutex_unlock(&rdev->pm.mutex); + return count; } @@ -524,10 +569,15 @@ static ssize_t radeon_hwmon_show_temp(struct device *dev, struct device_attribute *attr, char *buf) { - struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev)); - struct radeon_device *rdev = ddev->dev_private; + struct radeon_device *rdev = dev_get_drvdata(dev); + struct drm_device *ddev = rdev->ddev; int temp; + /* Can't get temperature when the card is off */ + if ((rdev->flags & RADEON_IS_PX) && + (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return -EINVAL; + if (rdev->asic->pm.get_temperature) temp = radeon_get_temperature(rdev); else @@ -536,32 +586,62 @@ static ssize_t radeon_hwmon_show_temp(struct device *dev, return snprintf(buf, PAGE_SIZE, "%d\n", temp); } -static ssize_t radeon_hwmon_show_name(struct device *dev, - struct device_attribute *attr, - char *buf) +static ssize_t radeon_hwmon_show_temp_thresh(struct device *dev, + struct device_attribute *attr, + char *buf) { - return sprintf(buf, "radeon\n"); + struct radeon_device *rdev = dev_get_drvdata(dev); + int hyst = to_sensor_dev_attr(attr)->index; + int temp; + + if (hyst) + temp = rdev->pm.dpm.thermal.min_temp; + else + temp = rdev->pm.dpm.thermal.max_temp; + + return snprintf(buf, PAGE_SIZE, "%d\n", temp); } static SENSOR_DEVICE_ATTR(temp1_input, S_IRUGO, radeon_hwmon_show_temp, NULL, 0); -static SENSOR_DEVICE_ATTR(name, S_IRUGO, radeon_hwmon_show_name, NULL, 0); +static SENSOR_DEVICE_ATTR(temp1_crit, S_IRUGO, radeon_hwmon_show_temp_thresh, NULL, 0); +static SENSOR_DEVICE_ATTR(temp1_crit_hyst, S_IRUGO, radeon_hwmon_show_temp_thresh, NULL, 1); static struct attribute *hwmon_attributes[] = { &sensor_dev_attr_temp1_input.dev_attr.attr, - &sensor_dev_attr_name.dev_attr.attr, + &sensor_dev_attr_temp1_crit.dev_attr.attr, + &sensor_dev_attr_temp1_crit_hyst.dev_attr.attr, NULL }; +static umode_t hwmon_attributes_visible(struct kobject *kobj, + struct attribute *attr, int index) +{ + struct device *dev = container_of(kobj, struct device, kobj); + struct radeon_device *rdev = dev_get_drvdata(dev); + + /* Skip limit attributes if DPM is not enabled */ + if (rdev->pm.pm_method != PM_METHOD_DPM && + (attr == &sensor_dev_attr_temp1_crit.dev_attr.attr || + attr == &sensor_dev_attr_temp1_crit_hyst.dev_attr.attr)) + return 0; + + return attr->mode; +} + static const struct attribute_group hwmon_attrgroup = { .attrs = hwmon_attributes, + .is_visible = hwmon_attributes_visible, +}; + +static const struct attribute_group *hwmon_groups[] = { + &hwmon_attrgroup, + NULL }; static int radeon_hwmon_init(struct radeon_device *rdev) { int err = 0; - rdev->pm.int_hwmon_dev = NULL; - switch (rdev->pm.int_thermal_type) { case THERMAL_TYPE_RV6XX: case THERMAL_TYPE_RV770: @@ -569,22 +649,17 @@ static int radeon_hwmon_init(struct radeon_device *rdev) case THERMAL_TYPE_NI: case THERMAL_TYPE_SUMO: case THERMAL_TYPE_SI: + case THERMAL_TYPE_CI: + case THERMAL_TYPE_KV: if (rdev->asic->pm.get_temperature == NULL) return err; - rdev->pm.int_hwmon_dev = hwmon_device_register(rdev->dev); + rdev->pm.int_hwmon_dev = hwmon_device_register_with_groups(rdev->dev, + "radeon", rdev, + hwmon_groups); if (IS_ERR(rdev->pm.int_hwmon_dev)) { err = PTR_ERR(rdev->pm.int_hwmon_dev); dev_err(rdev->dev, "Unable to register hwmon device: %d\n", err); - break; - } - dev_set_drvdata(rdev->pm.int_hwmon_dev, rdev->ddev); - err = sysfs_create_group(&rdev->pm.int_hwmon_dev->kobj, - &hwmon_attrgroup); - if (err) { - dev_err(rdev->dev, - "Unable to create hwmon sysfs file: %d\n", err); - hwmon_device_unregister(rdev->dev); } break; default: @@ -596,10 +671,8 @@ static int radeon_hwmon_init(struct radeon_device *rdev) static void radeon_hwmon_fini(struct radeon_device *rdev) { - if (rdev->pm.int_hwmon_dev) { - sysfs_remove_group(&rdev->pm.int_hwmon_dev->kobj, &hwmon_attrgroup); + if (rdev->pm.int_hwmon_dev) hwmon_device_unregister(rdev->pm.int_hwmon_dev); - } } static void radeon_dpm_thermal_work_handler(struct work_struct *work) @@ -624,7 +697,15 @@ static void radeon_dpm_thermal_work_handler(struct work_struct *work) /* switch back the user state */ dpm_state = rdev->pm.dpm.user_state; } - radeon_dpm_enable_power_state(rdev, dpm_state); + mutex_lock(&rdev->pm.mutex); + if (dpm_state == POWER_STATE_TYPE_INTERNAL_THERMAL) + rdev->pm.dpm.thermal_active = true; + else + rdev->pm.dpm.thermal_active = false; + rdev->pm.dpm.state = dpm_state; + mutex_unlock(&rdev->pm.mutex); + + radeon_pm_compute_clocks(rdev); } static struct radeon_ps *radeon_dpm_pick_power_state(struct radeon_device *rdev, @@ -687,7 +768,10 @@ restart_search: break; /* internal states */ case POWER_STATE_TYPE_INTERNAL_UVD: - return rdev->pm.dpm.uvd_ps; + if (rdev->pm.dpm.uvd_ps) + return rdev->pm.dpm.uvd_ps; + else + break; case POWER_STATE_TYPE_INTERNAL_UVD_SD: if (ps->class & ATOM_PPLIB_CLASSIFICATION_SDSTATE) return ps; @@ -729,10 +813,17 @@ restart_search: /* use a fallback state if we didn't match */ switch (dpm_state) { case POWER_STATE_TYPE_INTERNAL_UVD_SD: + dpm_state = POWER_STATE_TYPE_INTERNAL_UVD_HD; + goto restart_search; case POWER_STATE_TYPE_INTERNAL_UVD_HD: case POWER_STATE_TYPE_INTERNAL_UVD_HD2: case POWER_STATE_TYPE_INTERNAL_UVD_MVC: - return rdev->pm.dpm.uvd_ps; + if (rdev->pm.dpm.uvd_ps) { + return rdev->pm.dpm.uvd_ps; + } else { + dpm_state = POWER_STATE_TYPE_PERFORMANCE; + goto restart_search; + } case POWER_STATE_TYPE_INTERNAL_THERMAL: dpm_state = POWER_STATE_TYPE_INTERNAL_ACPI; goto restart_search; @@ -778,6 +869,9 @@ static void radeon_dpm_change_power_state_locked(struct radeon_device *rdev) /* no need to reprogram if nothing changed unless we are on BTC+ */ if (rdev->pm.dpm.current_ps == rdev->pm.dpm.requested_ps) { + /* vce just modifies an existing state so force a change */ + if (ps->vce_active != rdev->pm.dpm.vce_active) + goto force; if ((rdev->family < CHIP_BARTS) || (rdev->flags & RADEON_IS_IGP)) { /* for pre-BTC and APUs if the num crtcs changed but state is the same, * all we need to do is update the display configuration. @@ -814,15 +908,21 @@ static void radeon_dpm_change_power_state_locked(struct radeon_device *rdev) } } - printk("switching from power state:\n"); - radeon_dpm_print_power_state(rdev, rdev->pm.dpm.current_ps); - printk("switching to power state:\n"); - radeon_dpm_print_power_state(rdev, rdev->pm.dpm.requested_ps); +force: + if (radeon_dpm == 1) { + printk("switching from power state:\n"); + radeon_dpm_print_power_state(rdev, rdev->pm.dpm.current_ps); + printk("switching to power state:\n"); + radeon_dpm_print_power_state(rdev, rdev->pm.dpm.requested_ps); + } mutex_lock(&rdev->ddev->struct_mutex); down_write(&rdev->pm.mclk_lock); mutex_lock(&rdev->ring_lock); + /* update whether vce is active */ + ps->vce_active = rdev->pm.dpm.vce_active; + ret = radeon_dpm_pre_set_power_state(rdev); if (ret) goto done; @@ -839,7 +939,7 @@ static void radeon_dpm_change_power_state_locked(struct radeon_device *rdev) for (i = 0; i < RADEON_NUM_RINGS; i++) { struct radeon_ring *ring = &rdev->ring[i]; if (ring->ready) - radeon_fence_wait_empty_locked(rdev, i); + radeon_fence_wait_empty(rdev, i); } /* program the new power state */ @@ -850,37 +950,81 @@ static void radeon_dpm_change_power_state_locked(struct radeon_device *rdev) radeon_dpm_post_set_power_state(rdev); + if (rdev->asic->dpm.force_performance_level) { + if (rdev->pm.dpm.thermal_active) { + enum radeon_dpm_forced_level level = rdev->pm.dpm.forced_level; + /* force low perf level for thermal */ + radeon_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_LOW); + /* save the user's level */ + rdev->pm.dpm.forced_level = level; + } else { + /* otherwise, user selected level */ + radeon_dpm_force_performance_level(rdev, rdev->pm.dpm.forced_level); + } + } + done: mutex_unlock(&rdev->ring_lock); up_write(&rdev->pm.mclk_lock); mutex_unlock(&rdev->ddev->struct_mutex); } -void radeon_dpm_enable_power_state(struct radeon_device *rdev, - enum radeon_pm_state_type dpm_state) +void radeon_dpm_enable_uvd(struct radeon_device *rdev, bool enable) { - if (!rdev->pm.dpm_enabled) - return; + enum radeon_pm_state_type dpm_state; - mutex_lock(&rdev->pm.mutex); - switch (dpm_state) { - case POWER_STATE_TYPE_INTERNAL_THERMAL: - rdev->pm.dpm.thermal_active = true; - break; - case POWER_STATE_TYPE_INTERNAL_UVD: - case POWER_STATE_TYPE_INTERNAL_UVD_SD: - case POWER_STATE_TYPE_INTERNAL_UVD_HD: - case POWER_STATE_TYPE_INTERNAL_UVD_HD2: - case POWER_STATE_TYPE_INTERNAL_UVD_MVC: - rdev->pm.dpm.uvd_active = true; - break; - default: - rdev->pm.dpm.thermal_active = false; - rdev->pm.dpm.uvd_active = false; - break; + if (rdev->asic->dpm.powergate_uvd) { + mutex_lock(&rdev->pm.mutex); + /* don't powergate anything if we + have active but pause streams */ + enable |= rdev->pm.dpm.sd > 0; + enable |= rdev->pm.dpm.hd > 0; + /* enable/disable UVD */ + radeon_dpm_powergate_uvd(rdev, !enable); + mutex_unlock(&rdev->pm.mutex); + } else { + if (enable) { + mutex_lock(&rdev->pm.mutex); + rdev->pm.dpm.uvd_active = true; + /* disable this for now */ +#if 0 + if ((rdev->pm.dpm.sd == 1) && (rdev->pm.dpm.hd == 0)) + dpm_state = POWER_STATE_TYPE_INTERNAL_UVD_SD; + else if ((rdev->pm.dpm.sd == 2) && (rdev->pm.dpm.hd == 0)) + dpm_state = POWER_STATE_TYPE_INTERNAL_UVD_HD; + else if ((rdev->pm.dpm.sd == 0) && (rdev->pm.dpm.hd == 1)) + dpm_state = POWER_STATE_TYPE_INTERNAL_UVD_HD; + else if ((rdev->pm.dpm.sd == 0) && (rdev->pm.dpm.hd == 2)) + dpm_state = POWER_STATE_TYPE_INTERNAL_UVD_HD2; + else +#endif + dpm_state = POWER_STATE_TYPE_INTERNAL_UVD; + rdev->pm.dpm.state = dpm_state; + mutex_unlock(&rdev->pm.mutex); + } else { + mutex_lock(&rdev->pm.mutex); + rdev->pm.dpm.uvd_active = false; + mutex_unlock(&rdev->pm.mutex); + } + + radeon_pm_compute_clocks(rdev); } - rdev->pm.dpm.state = dpm_state; - mutex_unlock(&rdev->pm.mutex); +} + +void radeon_dpm_enable_vce(struct radeon_device *rdev, bool enable) +{ + if (enable) { + mutex_lock(&rdev->pm.mutex); + rdev->pm.dpm.vce_active = true; + /* XXX select vce level based on ring/task */ + rdev->pm.dpm.vce_level = RADEON_VCE_LEVEL_AC_ALL; + mutex_unlock(&rdev->pm.mutex); + } else { + mutex_lock(&rdev->pm.mutex); + rdev->pm.dpm.vce_active = false; + mutex_unlock(&rdev->pm.mutex); + } + radeon_pm_compute_clocks(rdev); } @@ -919,7 +1063,7 @@ static void radeon_pm_resume_old(struct radeon_device *rdev) { /* set up the default clocks if the MC ucode is loaded */ if ((rdev->family >= CHIP_BARTS) && - (rdev->family <= CHIP_HAINAN) && + (rdev->family <= CHIP_CAYMAN) && rdev->mc_fw) { if (rdev->pm.default_vddc) radeon_atom_set_voltage(rdev, rdev->pm.default_vddc, @@ -938,8 +1082,10 @@ static void radeon_pm_resume_old(struct radeon_device *rdev) rdev->pm.current_clock_mode_index = 0; rdev->pm.current_sclk = rdev->pm.default_sclk; rdev->pm.current_mclk = rdev->pm.default_mclk; - rdev->pm.current_vddc = rdev->pm.power_state[rdev->pm.default_power_state_index].clock_info[0].voltage.voltage; - rdev->pm.current_vddci = rdev->pm.power_state[rdev->pm.default_power_state_index].clock_info[0].voltage.vddci; + if (rdev->pm.power_state) { + rdev->pm.current_vddc = rdev->pm.power_state[rdev->pm.default_power_state_index].clock_info[0].voltage.voltage; + rdev->pm.current_vddci = rdev->pm.power_state[rdev->pm.default_power_state_index].clock_info[0].voltage.vddci; + } if (rdev->pm.pm_method == PM_METHOD_DYNPM && rdev->pm.dynpm_state == DYNPM_STATE_SUSPENDED) { rdev->pm.dynpm_state = DYNPM_STATE_ACTIVE; @@ -960,25 +1106,26 @@ static void radeon_pm_resume_dpm(struct radeon_device *rdev) radeon_dpm_setup_asic(rdev); ret = radeon_dpm_enable(rdev); mutex_unlock(&rdev->pm.mutex); - if (ret) { - DRM_ERROR("radeon: dpm resume failed\n"); - if ((rdev->family >= CHIP_BARTS) && - (rdev->family <= CHIP_HAINAN) && - rdev->mc_fw) { - if (rdev->pm.default_vddc) - radeon_atom_set_voltage(rdev, rdev->pm.default_vddc, - SET_VOLTAGE_TYPE_ASIC_VDDC); - if (rdev->pm.default_vddci) - radeon_atom_set_voltage(rdev, rdev->pm.default_vddci, - SET_VOLTAGE_TYPE_ASIC_VDDCI); - if (rdev->pm.default_sclk) - radeon_set_engine_clock(rdev, rdev->pm.default_sclk); - if (rdev->pm.default_mclk) - radeon_set_memory_clock(rdev, rdev->pm.default_mclk); - } - } else { - rdev->pm.dpm_enabled = true; - radeon_pm_compute_clocks(rdev); + if (ret) + goto dpm_resume_fail; + rdev->pm.dpm_enabled = true; + return; + +dpm_resume_fail: + DRM_ERROR("radeon: dpm resume failed\n"); + if ((rdev->family >= CHIP_BARTS) && + (rdev->family <= CHIP_CAYMAN) && + rdev->mc_fw) { + if (rdev->pm.default_vddc) + radeon_atom_set_voltage(rdev, rdev->pm.default_vddc, + SET_VOLTAGE_TYPE_ASIC_VDDC); + if (rdev->pm.default_vddci) + radeon_atom_set_voltage(rdev, rdev->pm.default_vddci, + SET_VOLTAGE_TYPE_ASIC_VDDCI); + if (rdev->pm.default_sclk) + radeon_set_engine_clock(rdev, rdev->pm.default_sclk); + if (rdev->pm.default_mclk) + radeon_set_memory_clock(rdev, rdev->pm.default_mclk); } } @@ -1014,7 +1161,7 @@ static int radeon_pm_init_old(struct radeon_device *rdev) radeon_pm_init_profile(rdev); /* set up the default clocks if the MC ucode is loaded */ if ((rdev->family >= CHIP_BARTS) && - (rdev->family <= CHIP_HAINAN) && + (rdev->family <= CHIP_CAYMAN) && rdev->mc_fw) { if (rdev->pm.default_vddc) radeon_atom_set_voltage(rdev, rdev->pm.default_vddc, @@ -1069,9 +1216,10 @@ static int radeon_pm_init_dpm(struct radeon_device *rdev) { int ret; - /* default to performance state */ + /* default to balanced state */ rdev->pm.dpm.state = POWER_STATE_TYPE_BALANCED; rdev->pm.dpm.user_state = POWER_STATE_TYPE_BALANCED; + rdev->pm.dpm.forced_level = RADEON_DPM_FORCED_LEVEL_AUTO; rdev->pm.default_sclk = rdev->clock.default_sclk; rdev->pm.default_mclk = rdev->clock.default_mclk; rdev->pm.current_sclk = rdev->clock.default_sclk; @@ -1092,55 +1240,55 @@ static int radeon_pm_init_dpm(struct radeon_device *rdev) mutex_lock(&rdev->pm.mutex); radeon_dpm_init(rdev); rdev->pm.dpm.current_ps = rdev->pm.dpm.requested_ps = rdev->pm.dpm.boot_ps; - radeon_dpm_print_power_states(rdev); + if (radeon_dpm == 1) + radeon_dpm_print_power_states(rdev); radeon_dpm_setup_asic(rdev); ret = radeon_dpm_enable(rdev); mutex_unlock(&rdev->pm.mutex); - if (ret) { - rdev->pm.dpm_enabled = false; - if ((rdev->family >= CHIP_BARTS) && - (rdev->family <= CHIP_HAINAN) && - rdev->mc_fw) { - if (rdev->pm.default_vddc) - radeon_atom_set_voltage(rdev, rdev->pm.default_vddc, - SET_VOLTAGE_TYPE_ASIC_VDDC); - if (rdev->pm.default_vddci) - radeon_atom_set_voltage(rdev, rdev->pm.default_vddci, - SET_VOLTAGE_TYPE_ASIC_VDDCI); - if (rdev->pm.default_sclk) - radeon_set_engine_clock(rdev, rdev->pm.default_sclk); - if (rdev->pm.default_mclk) - radeon_set_memory_clock(rdev, rdev->pm.default_mclk); - } - DRM_ERROR("radeon: dpm initialization failed\n"); - return ret; - } + if (ret) + goto dpm_failed; rdev->pm.dpm_enabled = true; - radeon_pm_compute_clocks(rdev); - if (rdev->pm.num_power_states > 1) { - ret = device_create_file(rdev->dev, &dev_attr_power_dpm_state); - if (ret) - DRM_ERROR("failed to create device file for dpm state\n"); - ret = device_create_file(rdev->dev, &dev_attr_power_dpm_force_performance_level); - if (ret) - DRM_ERROR("failed to create device file for dpm state\n"); - /* XXX: these are noops for dpm but are here for backwards compat */ - ret = device_create_file(rdev->dev, &dev_attr_power_profile); - if (ret) - DRM_ERROR("failed to create device file for power profile\n"); - ret = device_create_file(rdev->dev, &dev_attr_power_method); - if (ret) - DRM_ERROR("failed to create device file for power method\n"); - - if (radeon_debugfs_pm_init(rdev)) { - DRM_ERROR("Failed to register debugfs file for dpm!\n"); - } + ret = device_create_file(rdev->dev, &dev_attr_power_dpm_state); + if (ret) + DRM_ERROR("failed to create device file for dpm state\n"); + ret = device_create_file(rdev->dev, &dev_attr_power_dpm_force_performance_level); + if (ret) + DRM_ERROR("failed to create device file for dpm state\n"); + /* XXX: these are noops for dpm but are here for backwards compat */ + ret = device_create_file(rdev->dev, &dev_attr_power_profile); + if (ret) + DRM_ERROR("failed to create device file for power profile\n"); + ret = device_create_file(rdev->dev, &dev_attr_power_method); + if (ret) + DRM_ERROR("failed to create device file for power method\n"); - DRM_INFO("radeon: dpm initialized\n"); + if (radeon_debugfs_pm_init(rdev)) { + DRM_ERROR("Failed to register debugfs file for dpm!\n"); } + DRM_INFO("radeon: dpm initialized\n"); + return 0; + +dpm_failed: + rdev->pm.dpm_enabled = false; + if ((rdev->family >= CHIP_BARTS) && + (rdev->family <= CHIP_CAYMAN) && + rdev->mc_fw) { + if (rdev->pm.default_vddc) + radeon_atom_set_voltage(rdev, rdev->pm.default_vddc, + SET_VOLTAGE_TYPE_ASIC_VDDC); + if (rdev->pm.default_vddci) + radeon_atom_set_voltage(rdev, rdev->pm.default_vddci, + SET_VOLTAGE_TYPE_ASIC_VDDCI); + if (rdev->pm.default_sclk) + radeon_set_engine_clock(rdev, rdev->pm.default_sclk); + if (rdev->pm.default_mclk) + radeon_set_memory_clock(rdev, rdev->pm.default_mclk); + } + DRM_ERROR("radeon: dpm initialization failed\n"); + return ret; } int radeon_pm_init(struct radeon_device *rdev) @@ -1155,6 +1303,22 @@ int radeon_pm_init(struct radeon_device *rdev) case CHIP_RS780: case CHIP_RS880: case CHIP_RV770: + case CHIP_BARTS: + case CHIP_TURKS: + case CHIP_CAICOS: + case CHIP_CAYMAN: + /* DPM requires the RLC, RV770+ dGPU requires SMC */ + if (!rdev->rlc_fw) + rdev->pm.pm_method = PM_METHOD_PROFILE; + else if ((rdev->family >= CHIP_RV770) && + (!(rdev->flags & RADEON_IS_IGP)) && + (!rdev->smc_fw)) + rdev->pm.pm_method = PM_METHOD_PROFILE; + else if (radeon_dpm == 1) + rdev->pm.pm_method = PM_METHOD_DPM; + else + rdev->pm.pm_method = PM_METHOD_PROFILE; + break; case CHIP_RV730: case CHIP_RV710: case CHIP_RV740: @@ -1166,20 +1330,28 @@ int radeon_pm_init(struct radeon_device *rdev) case CHIP_PALM: case CHIP_SUMO: case CHIP_SUMO2: - case CHIP_BARTS: - case CHIP_TURKS: - case CHIP_CAICOS: - case CHIP_CAYMAN: case CHIP_ARUBA: case CHIP_TAHITI: case CHIP_PITCAIRN: case CHIP_VERDE: case CHIP_OLAND: case CHIP_HAINAN: - if (radeon_dpm == 1) - rdev->pm.pm_method = PM_METHOD_DPM; - else + case CHIP_BONAIRE: + case CHIP_KABINI: + case CHIP_KAVERI: + case CHIP_HAWAII: + case CHIP_MULLINS: + /* DPM requires the RLC, RV770+ dGPU requires SMC */ + if (!rdev->rlc_fw) + rdev->pm.pm_method = PM_METHOD_PROFILE; + else if ((rdev->family >= CHIP_RV770) && + (!(rdev->flags & RADEON_IS_IGP)) && + (!rdev->smc_fw)) + rdev->pm.pm_method = PM_METHOD_PROFILE; + else if (radeon_dpm == 0) rdev->pm.pm_method = PM_METHOD_PROFILE; + else + rdev->pm.pm_method = PM_METHOD_DPM; break; default: /* default to profile method */ @@ -1193,6 +1365,18 @@ int radeon_pm_init(struct radeon_device *rdev) return radeon_pm_init_old(rdev); } +int radeon_pm_late_init(struct radeon_device *rdev) +{ + int ret = 0; + + if (rdev->pm.pm_method == PM_METHOD_DPM) { + mutex_lock(&rdev->pm.mutex); + ret = radeon_dpm_late_enable(rdev); + mutex_unlock(&rdev->pm.mutex); + } + return ret; +} + static void radeon_pm_fini_old(struct radeon_device *rdev) { if (rdev->pm.num_power_states > 1) { @@ -1215,10 +1399,10 @@ static void radeon_pm_fini_old(struct radeon_device *rdev) device_remove_file(rdev->dev, &dev_attr_power_method); } + radeon_hwmon_fini(rdev); + if (rdev->pm.power_state) kfree(rdev->pm.power_state); - - radeon_hwmon_fini(rdev); } static void radeon_pm_fini_dpm(struct radeon_device *rdev) @@ -1236,10 +1420,10 @@ static void radeon_pm_fini_dpm(struct radeon_device *rdev) } radeon_dpm_fini(rdev); + radeon_hwmon_fini(rdev); + if (rdev->pm.power_state) kfree(rdev->pm.power_state); - - radeon_hwmon_fini(rdev); } void radeon_pm_fini(struct radeon_device *rdev) @@ -1263,12 +1447,14 @@ static void radeon_pm_compute_clocks_old(struct radeon_device *rdev) rdev->pm.active_crtcs = 0; rdev->pm.active_crtc_count = 0; - list_for_each_entry(crtc, - &ddev->mode_config.crtc_list, head) { - radeon_crtc = to_radeon_crtc(crtc); - if (radeon_crtc->enabled) { - rdev->pm.active_crtcs |= (1 << radeon_crtc->crtc_id); - rdev->pm.active_crtc_count++; + if (rdev->num_crtc && rdev->mode_info.mode_config_initialized) { + list_for_each_entry(crtc, + &ddev->mode_config.crtc_list, head) { + radeon_crtc = to_radeon_crtc(crtc); + if (radeon_crtc->enabled) { + rdev->pm.active_crtcs |= (1 << radeon_crtc->crtc_id); + rdev->pm.active_crtc_count++; + } } } @@ -1327,17 +1513,22 @@ static void radeon_pm_compute_clocks_dpm(struct radeon_device *rdev) struct drm_crtc *crtc; struct radeon_crtc *radeon_crtc; + if (!rdev->pm.dpm_enabled) + return; + mutex_lock(&rdev->pm.mutex); /* update active crtc counts */ rdev->pm.dpm.new_active_crtcs = 0; rdev->pm.dpm.new_active_crtc_count = 0; - list_for_each_entry(crtc, - &ddev->mode_config.crtc_list, head) { - radeon_crtc = to_radeon_crtc(crtc); - if (crtc->enabled) { - rdev->pm.dpm.new_active_crtcs |= (1 << radeon_crtc->crtc_id); - rdev->pm.dpm.new_active_crtc_count++; + if (rdev->num_crtc && rdev->mode_info.mode_config_initialized) { + list_for_each_entry(crtc, + &ddev->mode_config.crtc_list, head) { + radeon_crtc = to_radeon_crtc(crtc); + if (crtc->enabled) { + rdev->pm.dpm.new_active_crtcs |= (1 << radeon_crtc->crtc_id); + rdev->pm.dpm.new_active_crtc_count++; + } } } @@ -1371,7 +1562,7 @@ static bool radeon_pm_in_vbl(struct radeon_device *rdev) */ for (crtc = 0; (crtc < rdev->num_crtc) && in_vbl; crtc++) { if (rdev->pm.active_crtcs & (1 << crtc)) { - vbl_status = radeon_get_crtc_scanoutpos(rdev->ddev, crtc, &vpos, &hpos); + vbl_status = radeon_get_crtc_scanoutpos(rdev->ddev, crtc, 0, &vpos, &hpos, NULL, NULL); if ((vbl_status & DRM_SCANOUTPOS_VALID) && !(vbl_status & DRM_SCANOUTPOS_INVBL)) in_vbl = false; @@ -1463,8 +1654,12 @@ static int radeon_debugfs_pm_info(struct seq_file *m, void *data) struct drm_info_node *node = (struct drm_info_node *) m->private; struct drm_device *dev = node->minor->dev; struct radeon_device *rdev = dev->dev_private; + struct drm_device *ddev = rdev->ddev; - if (rdev->pm.dpm_enabled) { + if ((rdev->flags & RADEON_IS_PX) && + (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) { + seq_printf(m, "PX asic powered off\n"); + } else if (rdev->pm.dpm_enabled) { mutex_lock(&rdev->pm.mutex); if (rdev->asic->dpm.debugfs_print_current_performance_level) radeon_dpm_debugfs_print_current_performance_level(rdev, m); diff --git a/drivers/gpu/drm/radeon/radeon_prime.c b/drivers/gpu/drm/radeon/radeon_prime.c index 65b9eabd5a2..20074560fc2 100644 --- a/drivers/gpu/drm/radeon/radeon_prime.c +++ b/drivers/gpu/drm/radeon/radeon_prime.c @@ -68,7 +68,6 @@ struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev, RADEON_GEM_DOMAIN_GTT, sg, &bo); if (ret) return ERR_PTR(ret); - bo->gem_base.driver_private = bo; mutex_lock(&rdev->gem.mutex); list_add_tail(&bo->list, &rdev->gem.objects); diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 5f1c51a776e..f8050f5429e 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -61,9 +61,9 @@ int radeon_ib_get(struct radeon_device *rdev, int ring, struct radeon_ib *ib, struct radeon_vm *vm, unsigned size) { - int i, r; + int r; - r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &ib->sa_bo, size, 256, true); + r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &ib->sa_bo, size, 256); if (r) { dev_err(rdev->dev, "failed to get a new IB (%d)\n", r); return r; @@ -87,8 +87,6 @@ int radeon_ib_get(struct radeon_device *rdev, int ring, ib->gpu_addr = radeon_sa_bo_gpu_addr(ib->sa_bo); } ib->is_const_ib = false; - for (i = 0; i < RADEON_NUM_RINGS; ++i) - ib->sync_to[i] = NULL; return 0; } @@ -109,25 +107,6 @@ void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib) } /** - * radeon_ib_sync_to - sync to fence before executing the IB - * - * @ib: IB object to add fence to - * @fence: fence to sync to - * - * Sync to the fence before executing the IB - */ -void radeon_ib_sync_to(struct radeon_ib *ib, struct radeon_fence *fence) -{ - struct radeon_fence *other; - - if (!fence) - return; - - other = ib->sync_to[fence->ring]; - ib->sync_to[fence->ring] = radeon_fence_later(fence, other); -} - -/** * radeon_ib_schedule - schedule an IB (Indirect Buffer) on the ring * * @rdev: radeon_device pointer @@ -151,8 +130,7 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, struct radeon_ib *const_ib) { struct radeon_ring *ring = &rdev->ring[ib->ring]; - bool need_sync = false; - int i, r = 0; + int r = 0; if (!ib->length_dw || !ring->ready) { /* TODO: Nothings in the ib we should report. */ @@ -161,29 +139,30 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, } /* 64 dwords should be enough for fence too */ - r = radeon_ring_lock(rdev, ring, 64 + RADEON_NUM_RINGS * 8); + r = radeon_ring_lock(rdev, ring, 64 + RADEON_NUM_SYNCS * 8); if (r) { dev_err(rdev->dev, "scheduling IB failed (%d).\n", r); return r; } - for (i = 0; i < RADEON_NUM_RINGS; ++i) { - struct radeon_fence *fence = ib->sync_to[i]; - if (radeon_fence_need_sync(fence, ib->ring)) { - need_sync = true; - radeon_semaphore_sync_rings(rdev, ib->semaphore, - fence->ring, ib->ring); - radeon_fence_note_sync(fence, ib->ring); - } - } - /* immediately free semaphore when we don't need to sync */ - if (!need_sync) { - radeon_semaphore_free(rdev, &ib->semaphore, NULL); + + /* grab a vm id if necessary */ + if (ib->vm) { + struct radeon_fence *vm_id_fence; + vm_id_fence = radeon_vm_grab_id(rdev, ib->vm, ib->ring); + radeon_semaphore_sync_to(ib->semaphore, vm_id_fence); } - /* if we can't remember our last VM flush then flush now! */ - /* XXX figure out why we have to flush for every IB */ - if (ib->vm /*&& !ib->vm->last_flush*/) { - radeon_ring_vm_flush(rdev, ib->ring, ib->vm); + + /* sync with other rings */ + r = radeon_semaphore_sync_rings(rdev, ib->semaphore, ib->ring); + if (r) { + dev_err(rdev->dev, "failed to sync rings (%d)\n", r); + radeon_ring_unlock_undo(rdev, ring); + return r; } + + if (ib->vm) + radeon_vm_flush(rdev, ib->vm, ib->ring); + if (const_ib) { radeon_ring_ib_execute(rdev, const_ib->ring, const_ib); radeon_semaphore_free(rdev, &const_ib->semaphore, NULL); @@ -198,10 +177,10 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, if (const_ib) { const_ib->fence = radeon_fence_ref(ib->fence); } - /* we just flushed the VM, remember that */ - if (ib->vm && !ib->vm->last_flush) { - ib->vm->last_flush = radeon_fence_ref(ib->fence); - } + + if (ib->vm) + radeon_vm_fence(rdev, ib->vm, ib->fence); + radeon_ring_unlock_commit(rdev, ring); return 0; } @@ -224,6 +203,7 @@ int radeon_ib_pool_init(struct radeon_device *rdev) } r = radeon_sa_bo_manager_init(rdev, &rdev->ring_tmp_bo, RADEON_IB_POOL_SIZE*64*1024, + RADEON_GPU_PAGE_SIZE, RADEON_GEM_DOMAIN_GTT); if (r) { return r; @@ -282,6 +262,7 @@ int radeon_ib_ring_tests(struct radeon_device *rdev) r = radeon_ib_test(rdev, i, ring); if (r) { ring->ready = false; + rdev->needs_reset = false; if (i == RADEON_RING_TYPE_GFX_INDEX) { /* oh, oh, that's really bad */ @@ -357,38 +338,6 @@ bool radeon_ring_supports_scratch_reg(struct radeon_device *rdev, } } -u32 radeon_ring_generic_get_rptr(struct radeon_device *rdev, - struct radeon_ring *ring) -{ - u32 rptr; - - if (rdev->wb.enabled && ring != &rdev->ring[R600_RING_TYPE_UVD_INDEX]) - rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]); - else - rptr = RREG32(ring->rptr_reg); - rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift; - - return rptr; -} - -u32 radeon_ring_generic_get_wptr(struct radeon_device *rdev, - struct radeon_ring *ring) -{ - u32 wptr; - - wptr = RREG32(ring->wptr_reg); - wptr = (wptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift; - - return wptr; -} - -void radeon_ring_generic_set_wptr(struct radeon_device *rdev, - struct radeon_ring *ring) -{ - WREG32(ring->wptr_reg, (ring->wptr << ring->ptr_reg_shift) & ring->ptr_reg_mask); - (void)RREG32(ring->wptr_reg); -} - /** * radeon_ring_free_size - update the free size * @@ -399,13 +348,17 @@ void radeon_ring_generic_set_wptr(struct radeon_device *rdev, */ void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *ring) { - ring->rptr = radeon_ring_get_rptr(rdev, ring); + uint32_t rptr = radeon_ring_get_rptr(rdev, ring); + /* This works because ring_size is a power of 2 */ - ring->ring_free_dw = (ring->rptr + (ring->ring_size / 4)); + ring->ring_free_dw = rptr + (ring->ring_size / 4); ring->ring_free_dw -= ring->wptr; ring->ring_free_dw &= ring->ptr_mask; if (!ring->ring_free_dw) { + /* this is an empty ring */ ring->ring_free_dw = ring->ring_size / 4; + /* update lockup info to avoid false positive */ + radeon_ring_lockup_update(rdev, ring); } } @@ -429,19 +382,13 @@ int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *ring, unsi /* Align requested size with padding so unlock_commit can * pad safely */ radeon_ring_free_size(rdev, ring); - if (ring->ring_free_dw == (ring->ring_size / 4)) { - /* This is an empty ring update lockup info to avoid - * false positive. - */ - radeon_ring_lockup_update(ring); - } ndw = (ndw + ring->align_mask) & ~ring->align_mask; while (ndw > (ring->ring_free_dw - 1)) { radeon_ring_free_size(rdev, ring); if (ndw < ring->ring_free_dw) { break; } - r = radeon_fence_wait_next_locked(rdev, ring->idx); + r = radeon_fence_wait_next(rdev, ring->idx); if (r) return r; } @@ -490,7 +437,7 @@ void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring) while (ring->wptr & ring->align_mask) { radeon_ring_write(ring, ring->nop); } - DRM_MEMORYBARRIER(); + mb(); radeon_ring_set_wptr(rdev, ring); } @@ -535,39 +482,17 @@ void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *rin } /** - * radeon_ring_force_activity - add some nop packets to the ring - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Add some nop packets to the ring to force activity (all asics). - * Used for lockup detection to see if the rptr is advancing. - */ -void radeon_ring_force_activity(struct radeon_device *rdev, struct radeon_ring *ring) -{ - int r; - - radeon_ring_free_size(rdev, ring); - if (ring->rptr == ring->wptr) { - r = radeon_ring_alloc(rdev, ring, 1); - if (!r) { - radeon_ring_write(ring, ring->nop); - radeon_ring_commit(rdev, ring); - } - } -} - -/** * radeon_ring_lockup_update - update lockup variables * * @ring: radeon_ring structure holding ring information * * Update the last rptr value and timestamp (all asics). */ -void radeon_ring_lockup_update(struct radeon_ring *ring) +void radeon_ring_lockup_update(struct radeon_device *rdev, + struct radeon_ring *ring) { - ring->last_rptr = ring->rptr; - ring->last_activity = jiffies; + atomic_set(&ring->last_rptr, radeon_ring_get_rptr(rdev, ring)); + atomic64_set(&ring->last_activity, jiffies_64); } /** @@ -575,40 +500,23 @@ void radeon_ring_lockup_update(struct radeon_ring *ring) * @rdev: radeon device structure * @ring: radeon_ring structure holding ring information * - * We don't need to initialize the lockup tracking information as we will either - * have CP rptr to a different value of jiffies wrap around which will force - * initialization of the lockup tracking informations. - * - * A possible false positivie is if we get call after while and last_cp_rptr == - * the current CP rptr, even if it's unlikely it might happen. To avoid this - * if the elapsed time since last call is bigger than 2 second than we return - * false and update the tracking information. Due to this the caller must call - * radeon_ring_test_lockup several time in less than 2sec for lockup to be reported - * the fencing code should be cautious about that. - * - * Caller should write to the ring to force CP to do something so we don't get - * false positive when CP is just gived nothing to do. - * - **/ + */ bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *ring) { - unsigned long cjiffies, elapsed; + uint32_t rptr = radeon_ring_get_rptr(rdev, ring); + uint64_t last = atomic64_read(&ring->last_activity); + uint64_t elapsed; - cjiffies = jiffies; - if (!time_after(cjiffies, ring->last_activity)) { - /* likely a wrap around */ - radeon_ring_lockup_update(ring); - return false; - } - ring->rptr = radeon_ring_get_rptr(rdev, ring); - if (ring->rptr != ring->last_rptr) { - /* CP is still working no lockup */ - radeon_ring_lockup_update(ring); + if (rptr != atomic_read(&ring->last_rptr)) { + /* ring is still working, no lockup */ + radeon_ring_lockup_update(rdev, ring); return false; } - elapsed = jiffies_to_msecs(cjiffies - ring->last_activity); + + elapsed = jiffies_to_msecs(jiffies_64 - last); if (radeon_lockup_timeout && elapsed >= radeon_lockup_timeout) { - dev_err(rdev->dev, "GPU lockup CP stall for more than %lumsec\n", elapsed); + dev_err(rdev->dev, "ring %d stalled for more than %llumsec\n", + ring->idx, elapsed); return true; } /* give a chance to the GPU ... */ @@ -716,27 +624,18 @@ int radeon_ring_restore(struct radeon_device *rdev, struct radeon_ring *ring, * @ring: radeon_ring structure holding ring information * @ring_size: size of the ring * @rptr_offs: offset of the rptr writeback location in the WB buffer - * @rptr_reg: MMIO offset of the rptr register - * @wptr_reg: MMIO offset of the wptr register - * @ptr_reg_shift: bit offset of the rptr/wptr values - * @ptr_reg_mask: bit mask of the rptr/wptr values * @nop: nop packet for this ring * * Initialize the driver information for the selected ring (all asics). * Returns 0 on success, error on failure. */ int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsigned ring_size, - unsigned rptr_offs, unsigned rptr_reg, unsigned wptr_reg, - u32 ptr_reg_shift, u32 ptr_reg_mask, u32 nop) + unsigned rptr_offs, u32 nop) { int r; ring->ring_size = ring_size; ring->rptr_offs = rptr_offs; - ring->rptr_reg = rptr_reg; - ring->wptr_reg = wptr_reg; - ring->ptr_reg_shift = ptr_reg_shift; - ring->ptr_reg_mask = ptr_reg_mask; ring->nop = nop; /* Allocate ring buffer */ if (ring->ring_obj == NULL) { @@ -775,7 +674,7 @@ int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsig if (radeon_debugfs_ring_init(rdev, ring)) { DRM_ERROR("Failed to register debugfs file for rings !\n"); } - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return 0; } @@ -822,31 +721,51 @@ static int radeon_debugfs_ring_info(struct seq_file *m, void *data) struct radeon_device *rdev = dev->dev_private; int ridx = *(int*)node->info_ent->data; struct radeon_ring *ring = &rdev->ring[ridx]; + + uint32_t rptr, wptr, rptr_next; unsigned count, i, j; - u32 tmp; radeon_ring_free_size(rdev, ring); count = (ring->ring_size / 4) - ring->ring_free_dw; - tmp = radeon_ring_get_wptr(rdev, ring); - seq_printf(m, "wptr(0x%04x): 0x%08x [%5d]\n", ring->wptr_reg, tmp, tmp); - tmp = radeon_ring_get_rptr(rdev, ring); - seq_printf(m, "rptr(0x%04x): 0x%08x [%5d]\n", ring->rptr_reg, tmp, tmp); + + wptr = radeon_ring_get_wptr(rdev, ring); + seq_printf(m, "wptr: 0x%08x [%5d]\n", + wptr, wptr); + + rptr = radeon_ring_get_rptr(rdev, ring); + seq_printf(m, "rptr: 0x%08x [%5d]\n", + rptr, rptr); + if (ring->rptr_save_reg) { - seq_printf(m, "rptr next(0x%04x): 0x%08x\n", ring->rptr_save_reg, - RREG32(ring->rptr_save_reg)); - } - seq_printf(m, "driver's copy of the wptr: 0x%08x [%5d]\n", ring->wptr, ring->wptr); - seq_printf(m, "driver's copy of the rptr: 0x%08x [%5d]\n", ring->rptr, ring->rptr); - seq_printf(m, "last semaphore signal addr : 0x%016llx\n", ring->last_semaphore_signal_addr); - seq_printf(m, "last semaphore wait addr : 0x%016llx\n", ring->last_semaphore_wait_addr); + rptr_next = RREG32(ring->rptr_save_reg); + seq_printf(m, "rptr next(0x%04x): 0x%08x [%5d]\n", + ring->rptr_save_reg, rptr_next, rptr_next); + } else + rptr_next = ~0; + + seq_printf(m, "driver's copy of the wptr: 0x%08x [%5d]\n", + ring->wptr, ring->wptr); + seq_printf(m, "last semaphore signal addr : 0x%016llx\n", + ring->last_semaphore_signal_addr); + seq_printf(m, "last semaphore wait addr : 0x%016llx\n", + ring->last_semaphore_wait_addr); seq_printf(m, "%u free dwords in ring\n", ring->ring_free_dw); seq_printf(m, "%u dwords in ring\n", count); + + if (!ring->ready) + return 0; + /* print 8 dw before current rptr as often it's the last executed * packet that is the root issue */ - i = (ring->rptr + ring->ptr_mask + 1 - 32) & ring->ptr_mask; + i = (rptr + ring->ptr_mask + 1 - 32) & ring->ptr_mask; for (j = 0; j <= (count + 32); j++) { - seq_printf(m, "r[%5d]=0x%08x\n", i, ring->ring[i]); + seq_printf(m, "r[%5d]=0x%08x", i, ring->ring[i]); + if (rptr == i) + seq_puts(m, " *"); + if (rptr_next == i) + seq_puts(m, " #"); + seq_puts(m, "\n"); i = (i + 1) & ring->ptr_mask; } return 0; @@ -858,6 +777,8 @@ static int cayman_cp2_index = CAYMAN_RING_TYPE_CP2_INDEX; static int radeon_dma1_index = R600_RING_TYPE_DMA_INDEX; static int radeon_dma2_index = CAYMAN_RING_TYPE_DMA1_INDEX; static int r600_uvd_index = R600_RING_TYPE_UVD_INDEX; +static int si_vce1_index = TN_RING_TYPE_VCE1_INDEX; +static int si_vce2_index = TN_RING_TYPE_VCE2_INDEX; static struct drm_info_list radeon_debugfs_ring_info_list[] = { {"radeon_ring_gfx", radeon_debugfs_ring_info, 0, &radeon_gfx_index}, @@ -866,6 +787,8 @@ static struct drm_info_list radeon_debugfs_ring_info_list[] = { {"radeon_ring_dma1", radeon_debugfs_ring_info, 0, &radeon_dma1_index}, {"radeon_ring_dma2", radeon_debugfs_ring_info, 0, &radeon_dma2_index}, {"radeon_ring_uvd", radeon_debugfs_ring_info, 0, &r600_uvd_index}, + {"radeon_ring_vce1", radeon_debugfs_ring_info, 0, &si_vce1_index}, + {"radeon_ring_vce2", radeon_debugfs_ring_info, 0, &si_vce2_index}, }; static int radeon_debugfs_sa_info(struct seq_file *m, void *data) diff --git a/drivers/gpu/drm/radeon/radeon_sa.c b/drivers/gpu/drm/radeon/radeon_sa.c index 0abe5a9431b..adcf3e2f07d 100644 --- a/drivers/gpu/drm/radeon/radeon_sa.c +++ b/drivers/gpu/drm/radeon/radeon_sa.c @@ -49,7 +49,7 @@ static void radeon_sa_bo_try_free(struct radeon_sa_manager *sa_manager); int radeon_sa_bo_manager_init(struct radeon_device *rdev, struct radeon_sa_manager *sa_manager, - unsigned size, u32 domain) + unsigned size, u32 align, u32 domain) { int i, r; @@ -57,13 +57,14 @@ int radeon_sa_bo_manager_init(struct radeon_device *rdev, sa_manager->bo = NULL; sa_manager->size = size; sa_manager->domain = domain; + sa_manager->align = align; sa_manager->hole = &sa_manager->olist; INIT_LIST_HEAD(&sa_manager->olist); for (i = 0; i < RADEON_NUM_RINGS; ++i) { INIT_LIST_HEAD(&sa_manager->flist[i]); } - r = radeon_bo_create(rdev, size, RADEON_GPU_PAGE_SIZE, true, + r = radeon_bo_create(rdev, size, align, true, domain, NULL, &sa_manager->bo); if (r) { dev_err(rdev->dev, "(%d) failed to allocate bo for manager\n", r); @@ -311,13 +312,13 @@ static bool radeon_sa_bo_next_hole(struct radeon_sa_manager *sa_manager, int radeon_sa_bo_new(struct radeon_device *rdev, struct radeon_sa_manager *sa_manager, struct radeon_sa_bo **sa_bo, - unsigned size, unsigned align, bool block) + unsigned size, unsigned align) { struct radeon_fence *fences[RADEON_NUM_RINGS]; unsigned tries[RADEON_NUM_RINGS]; int i, r; - BUG_ON(align > RADEON_GPU_PAGE_SIZE); + BUG_ON(align > sa_manager->align); BUG_ON(size > sa_manager->size); *sa_bo = kmalloc(sizeof(struct radeon_sa_bo), GFP_KERNEL); @@ -352,14 +353,11 @@ int radeon_sa_bo_new(struct radeon_device *rdev, r = radeon_fence_wait_any(rdev, fences, false); spin_lock(&sa_manager->wq.lock); /* if we have nothing to wait for block */ - if (r == -ENOENT && block) { + if (r == -ENOENT) { r = wait_event_interruptible_locked( sa_manager->wq, radeon_sa_event(sa_manager, size, align) ); - - } else if (r == -ENOENT) { - r = -ENOMEM; } } while (!r); @@ -401,13 +399,15 @@ void radeon_sa_bo_dump_debug_info(struct radeon_sa_manager *sa_manager, spin_lock(&sa_manager->wq.lock); list_for_each_entry(i, &sa_manager->olist, olist) { + uint64_t soffset = i->soffset + sa_manager->gpu_addr; + uint64_t eoffset = i->eoffset + sa_manager->gpu_addr; if (&i->olist == sa_manager->hole) { seq_printf(m, ">"); } else { seq_printf(m, " "); } - seq_printf(m, "[0x%08x 0x%08x] size %8d", - i->soffset, i->eoffset, i->eoffset - i->soffset); + seq_printf(m, "[0x%010llx 0x%010llx] size %8lld", + soffset, eoffset, eoffset - soffset); if (i->fence) { seq_printf(m, " protected by 0x%016llx on ring %d", i->fence->seq, i->fence->ring); diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c index 8dcc20f53d7..dbd6bcde92d 100644 --- a/drivers/gpu/drm/radeon/radeon_semaphore.c +++ b/drivers/gpu/drm/radeon/radeon_semaphore.c @@ -29,19 +29,20 @@ */ #include <drm/drmP.h> #include "radeon.h" - +#include "radeon_trace.h" int radeon_semaphore_create(struct radeon_device *rdev, struct radeon_semaphore **semaphore) { - int r; + uint32_t *cpu_addr; + int i, r; *semaphore = kmalloc(sizeof(struct radeon_semaphore), GFP_KERNEL); if (*semaphore == NULL) { return -ENOMEM; } - r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, - &(*semaphore)->sa_bo, 8, 8, true); + r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &(*semaphore)->sa_bo, + 8 * RADEON_NUM_SYNCS, 8); if (r) { kfree(*semaphore); *semaphore = NULL; @@ -49,55 +50,140 @@ int radeon_semaphore_create(struct radeon_device *rdev, } (*semaphore)->waiters = 0; (*semaphore)->gpu_addr = radeon_sa_bo_gpu_addr((*semaphore)->sa_bo); - *((uint64_t*)radeon_sa_bo_cpu_addr((*semaphore)->sa_bo)) = 0; + + cpu_addr = radeon_sa_bo_cpu_addr((*semaphore)->sa_bo); + for (i = 0; i < RADEON_NUM_SYNCS; ++i) + cpu_addr[i] = 0; + + for (i = 0; i < RADEON_NUM_RINGS; ++i) + (*semaphore)->sync_to[i] = NULL; + return 0; } -void radeon_semaphore_emit_signal(struct radeon_device *rdev, int ring, +bool radeon_semaphore_emit_signal(struct radeon_device *rdev, int ridx, struct radeon_semaphore *semaphore) { - --semaphore->waiters; - radeon_semaphore_ring_emit(rdev, ring, &rdev->ring[ring], semaphore, false); + struct radeon_ring *ring = &rdev->ring[ridx]; + + trace_radeon_semaphore_signale(ridx, semaphore); + + if (radeon_semaphore_ring_emit(rdev, ridx, ring, semaphore, false)) { + --semaphore->waiters; + + /* for debugging lockup only, used by sysfs debug files */ + ring->last_semaphore_signal_addr = semaphore->gpu_addr; + return true; + } + return false; } -void radeon_semaphore_emit_wait(struct radeon_device *rdev, int ring, +bool radeon_semaphore_emit_wait(struct radeon_device *rdev, int ridx, struct radeon_semaphore *semaphore) { - ++semaphore->waiters; - radeon_semaphore_ring_emit(rdev, ring, &rdev->ring[ring], semaphore, true); + struct radeon_ring *ring = &rdev->ring[ridx]; + + trace_radeon_semaphore_wait(ridx, semaphore); + + if (radeon_semaphore_ring_emit(rdev, ridx, ring, semaphore, true)) { + ++semaphore->waiters; + + /* for debugging lockup only, used by sysfs debug files */ + ring->last_semaphore_wait_addr = semaphore->gpu_addr; + return true; + } + return false; } -/* caller must hold ring lock */ +/** + * radeon_semaphore_sync_to - use the semaphore to sync to a fence + * + * @semaphore: semaphore object to add fence to + * @fence: fence to sync to + * + * Sync to the fence using this semaphore object + */ +void radeon_semaphore_sync_to(struct radeon_semaphore *semaphore, + struct radeon_fence *fence) +{ + struct radeon_fence *other; + + if (!fence) + return; + + other = semaphore->sync_to[fence->ring]; + semaphore->sync_to[fence->ring] = radeon_fence_later(fence, other); +} + +/** + * radeon_semaphore_sync_rings - sync ring to all registered fences + * + * @rdev: radeon_device pointer + * @semaphore: semaphore object to use for sync + * @ring: ring that needs sync + * + * Ensure that all registered fences are signaled before letting + * the ring continue. The caller must hold the ring lock. + */ int radeon_semaphore_sync_rings(struct radeon_device *rdev, struct radeon_semaphore *semaphore, - int signaler, int waiter) + int ring) { - int r; + unsigned count = 0; + int i, r; - /* no need to signal and wait on the same ring */ - if (signaler == waiter) { - return 0; - } + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + struct radeon_fence *fence = semaphore->sync_to[i]; - /* prevent GPU deadlocks */ - if (!rdev->ring[signaler].ready) { - dev_err(rdev->dev, "Trying to sync to a disabled ring!"); - return -EINVAL; - } + /* check if we really need to sync */ + if (!radeon_fence_need_sync(fence, ring)) + continue; - r = radeon_ring_alloc(rdev, &rdev->ring[signaler], 8); - if (r) { - return r; - } - radeon_semaphore_emit_signal(rdev, signaler, semaphore); - radeon_ring_commit(rdev, &rdev->ring[signaler]); + /* prevent GPU deadlocks */ + if (!rdev->ring[i].ready) { + dev_err(rdev->dev, "Syncing to a disabled ring!"); + return -EINVAL; + } + + if (++count > RADEON_NUM_SYNCS) { + /* not enough room, wait manually */ + r = radeon_fence_wait(fence, false); + if (r) + return r; + continue; + } - /* we assume caller has already allocated space on waiters ring */ - radeon_semaphore_emit_wait(rdev, waiter, semaphore); + /* allocate enough space for sync command */ + r = radeon_ring_alloc(rdev, &rdev->ring[i], 16); + if (r) { + return r; + } - /* for debugging lockup only, used by sysfs debug files */ - rdev->ring[signaler].last_semaphore_signal_addr = semaphore->gpu_addr; - rdev->ring[waiter].last_semaphore_wait_addr = semaphore->gpu_addr; + /* emit the signal semaphore */ + if (!radeon_semaphore_emit_signal(rdev, i, semaphore)) { + /* signaling wasn't successful wait manually */ + radeon_ring_undo(&rdev->ring[i]); + r = radeon_fence_wait(fence, false); + if (r) + return r; + continue; + } + + /* we assume caller has already allocated space on waiters ring */ + if (!radeon_semaphore_emit_wait(rdev, ring, semaphore)) { + /* waiting wasn't successful wait manually */ + radeon_ring_undo(&rdev->ring[i]); + r = radeon_fence_wait(fence, false); + if (r) + return r; + continue; + } + + radeon_ring_commit(rdev, &rdev->ring[i]); + radeon_fence_note_sync(fence, ring); + + semaphore->gpu_addr += 8; + } return 0; } diff --git a/drivers/gpu/drm/radeon/radeon_state.c b/drivers/gpu/drm/radeon/radeon_state.c index 4d20910899d..23bb64fd775 100644 --- a/drivers/gpu/drm/radeon/radeon_state.c +++ b/drivers/gpu/drm/radeon/radeon_state.c @@ -1810,7 +1810,7 @@ static int radeon_cp_dispatch_texture(struct drm_device * dev, } if (!buf) { DRM_DEBUG("EAGAIN\n"); - if (DRM_COPY_TO_USER(tex->image, image, sizeof(*image))) + if (copy_to_user(tex->image, image, sizeof(*image))) return -EFAULT; return -EAGAIN; } @@ -1823,7 +1823,7 @@ static int radeon_cp_dispatch_texture(struct drm_device * dev, #define RADEON_COPY_MT(_buf, _data, _width) \ do { \ - if (DRM_COPY_FROM_USER(_buf, _data, (_width))) {\ + if (copy_from_user(_buf, _data, (_width))) {\ DRM_ERROR("EFAULT on pad, %d bytes\n", (_width)); \ return -EFAULT; \ } \ @@ -2168,7 +2168,7 @@ static int radeon_cp_clear(struct drm_device *dev, void *data, struct drm_file * if (sarea_priv->nbox > RADEON_NR_SAREA_CLIPRECTS) sarea_priv->nbox = RADEON_NR_SAREA_CLIPRECTS; - if (DRM_COPY_FROM_USER(&depth_boxes, clear->depth_boxes, + if (copy_from_user(&depth_boxes, clear->depth_boxes, sarea_priv->nbox * sizeof(depth_boxes[0]))) return -EFAULT; @@ -2436,7 +2436,7 @@ static int radeon_cp_texture(struct drm_device *dev, void *data, struct drm_file return -EINVAL; } - if (DRM_COPY_FROM_USER(&image, + if (copy_from_user(&image, (drm_radeon_tex_image_t __user *) tex->image, sizeof(image))) return -EFAULT; @@ -2460,7 +2460,7 @@ static int radeon_cp_stipple(struct drm_device *dev, void *data, struct drm_file LOCK_TEST_WITH_RETURN(dev, file_priv); - if (DRM_COPY_FROM_USER(&mask, stipple->mask, 32 * sizeof(u32))) + if (copy_from_user(&mask, stipple->mask, 32 * sizeof(u32))) return -EFAULT; RING_SPACE_TEST_WITH_RETURN(dev_priv); @@ -2585,13 +2585,13 @@ static int radeon_cp_vertex2(struct drm_device *dev, void *data, struct drm_file drm_radeon_prim_t prim; drm_radeon_tcl_prim_t tclprim; - if (DRM_COPY_FROM_USER(&prim, &vertex->prim[i], sizeof(prim))) + if (copy_from_user(&prim, &vertex->prim[i], sizeof(prim))) return -EFAULT; if (prim.stateidx != laststate) { drm_radeon_state_t state; - if (DRM_COPY_FROM_USER(&state, + if (copy_from_user(&state, &vertex->state[prim.stateidx], sizeof(state))) return -EFAULT; @@ -2799,7 +2799,7 @@ static int radeon_emit_packet3_cliprect(struct drm_device *dev, do { if (i < cmdbuf->nbox) { - if (DRM_COPY_FROM_USER(&box, &boxes[i], sizeof(box))) + if (copy_from_user(&box, &boxes[i], sizeof(box))) return -EFAULT; /* FIXME The second and subsequent times round * this loop, send a WAIT_UNTIL_3D_IDLE before @@ -3054,7 +3054,7 @@ static int radeon_cp_getparam(struct drm_device *dev, void *data, struct drm_fil if ((dev_priv->flags & RADEON_FAMILY_MASK) >= CHIP_R600) value = 0; else - value = drm_dev_to_irq(dev); + value = dev->pdev->irq; break; case RADEON_PARAM_GART_BASE: value = dev_priv->gart_vm_start; @@ -3116,7 +3116,7 @@ static int radeon_cp_getparam(struct drm_device *dev, void *data, struct drm_fil return -EINVAL; } - if (DRM_COPY_TO_USER(param->value, &value, sizeof(int))) { + if (copy_to_user(param->value, &value, sizeof(int))) { DRM_ERROR("copy_to_user\n"); return -EFAULT; } @@ -3258,4 +3258,4 @@ struct drm_ioctl_desc radeon_ioctls[] = { DRM_IOCTL_DEF_DRV(RADEON_CS, r600_cs_legacy_ioctl, DRM_AUTH) }; -int radeon_max_ioctl = DRM_ARRAY_SIZE(radeon_ioctls); +int radeon_max_ioctl = ARRAY_SIZE(radeon_ioctls); diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c index f4d6bcee900..3a13e0d1055 100644 --- a/drivers/gpu/drm/radeon/radeon_test.c +++ b/drivers/gpu/drm/radeon/radeon_test.c @@ -36,8 +36,8 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag) struct radeon_bo *vram_obj = NULL; struct radeon_bo **gtt_obj = NULL; uint64_t gtt_addr, vram_addr; - unsigned i, n, size; - int r, ring; + unsigned n, size; + int i, r, ring; switch (flag) { case RADEON_TEST_COPY_DMA: @@ -257,20 +257,36 @@ static int radeon_test_create_and_emit_fence(struct radeon_device *rdev, struct radeon_ring *ring, struct radeon_fence **fence) { + uint32_t handle = ring->idx ^ 0xdeafbeef; int r; if (ring->idx == R600_RING_TYPE_UVD_INDEX) { - r = radeon_uvd_get_create_msg(rdev, ring->idx, 1, NULL); + r = radeon_uvd_get_create_msg(rdev, ring->idx, handle, NULL); if (r) { DRM_ERROR("Failed to get dummy create msg\n"); return r; } - r = radeon_uvd_get_destroy_msg(rdev, ring->idx, 1, fence); + r = radeon_uvd_get_destroy_msg(rdev, ring->idx, handle, fence); if (r) { DRM_ERROR("Failed to get dummy destroy msg\n"); return r; } + + } else if (ring->idx == TN_RING_TYPE_VCE1_INDEX || + ring->idx == TN_RING_TYPE_VCE2_INDEX) { + r = radeon_vce_get_create_msg(rdev, ring->idx, handle, NULL); + if (r) { + DRM_ERROR("Failed to get dummy create msg\n"); + return r; + } + + r = radeon_vce_get_destroy_msg(rdev, ring->idx, handle, fence); + if (r) { + DRM_ERROR("Failed to get dummy destroy msg\n"); + return r; + } + } else { r = radeon_ring_lock(rdev, ring, 64); if (r) { @@ -486,6 +502,16 @@ out_cleanup: printk(KERN_WARNING "Error while testing ring sync (%d).\n", r); } +static bool radeon_test_sync_possible(struct radeon_ring *ringA, + struct radeon_ring *ringB) +{ + if (ringA->idx == TN_RING_TYPE_VCE2_INDEX && + ringB->idx == TN_RING_TYPE_VCE1_INDEX) + return false; + + return true; +} + void radeon_test_syncing(struct radeon_device *rdev) { int i, j, k; @@ -500,6 +526,9 @@ void radeon_test_syncing(struct radeon_device *rdev) if (!ringB->ready) continue; + if (!radeon_test_sync_possible(ringA, ringB)) + continue; + DRM_INFO("Testing syncing between rings %d and %d...\n", i, j); radeon_test_ring_sync(rdev, ringA, ringB); @@ -511,6 +540,12 @@ void radeon_test_syncing(struct radeon_device *rdev) if (!ringC->ready) continue; + if (!radeon_test_sync_possible(ringA, ringC)) + continue; + + if (!radeon_test_sync_possible(ringB, ringC)) + continue; + DRM_INFO("Testing syncing between rings %d, %d and %d...\n", i, j, k); radeon_test_ring_sync2(rdev, ringA, ringB, ringC); diff --git a/drivers/gpu/drm/radeon/radeon_trace.h b/drivers/gpu/drm/radeon/radeon_trace.h index eafd8160a15..f749f2c3bbd 100644 --- a/drivers/gpu/drm/radeon/radeon_trace.h +++ b/drivers/gpu/drm/radeon/radeon_trace.h @@ -27,51 +27,160 @@ TRACE_EVENT(radeon_bo_create, TP_printk("bo=%p, pages=%u", __entry->bo, __entry->pages) ); +TRACE_EVENT(radeon_cs, + TP_PROTO(struct radeon_cs_parser *p), + TP_ARGS(p), + TP_STRUCT__entry( + __field(u32, ring) + __field(u32, dw) + __field(u32, fences) + ), + + TP_fast_assign( + __entry->ring = p->ring; + __entry->dw = p->chunks[p->chunk_ib_idx].length_dw; + __entry->fences = radeon_fence_count_emitted( + p->rdev, p->ring); + ), + TP_printk("ring=%u, dw=%u, fences=%u", + __entry->ring, __entry->dw, + __entry->fences) +); + +TRACE_EVENT(radeon_vm_grab_id, + TP_PROTO(unsigned vmid, int ring), + TP_ARGS(vmid, ring), + TP_STRUCT__entry( + __field(u32, vmid) + __field(u32, ring) + ), + + TP_fast_assign( + __entry->vmid = vmid; + __entry->ring = ring; + ), + TP_printk("vmid=%u, ring=%u", __entry->vmid, __entry->ring) +); + +TRACE_EVENT(radeon_vm_bo_update, + TP_PROTO(struct radeon_bo_va *bo_va), + TP_ARGS(bo_va), + TP_STRUCT__entry( + __field(u64, soffset) + __field(u64, eoffset) + __field(u32, flags) + ), + + TP_fast_assign( + __entry->soffset = bo_va->soffset; + __entry->eoffset = bo_va->eoffset; + __entry->flags = bo_va->flags; + ), + TP_printk("soffs=%010llx, eoffs=%010llx, flags=%08x", + __entry->soffset, __entry->eoffset, __entry->flags) +); + +TRACE_EVENT(radeon_vm_set_page, + TP_PROTO(uint64_t pe, uint64_t addr, unsigned count, + uint32_t incr, uint32_t flags), + TP_ARGS(pe, addr, count, incr, flags), + TP_STRUCT__entry( + __field(u64, pe) + __field(u64, addr) + __field(u32, count) + __field(u32, incr) + __field(u32, flags) + ), + + TP_fast_assign( + __entry->pe = pe; + __entry->addr = addr; + __entry->count = count; + __entry->incr = incr; + __entry->flags = flags; + ), + TP_printk("pe=%010Lx, addr=%010Lx, incr=%u, flags=%08x, count=%u", + __entry->pe, __entry->addr, __entry->incr, + __entry->flags, __entry->count) +); + DECLARE_EVENT_CLASS(radeon_fence_request, - TP_PROTO(struct drm_device *dev, u32 seqno), + TP_PROTO(struct drm_device *dev, int ring, u32 seqno), - TP_ARGS(dev, seqno), + TP_ARGS(dev, ring, seqno), TP_STRUCT__entry( __field(u32, dev) + __field(int, ring) __field(u32, seqno) ), TP_fast_assign( __entry->dev = dev->primary->index; + __entry->ring = ring; __entry->seqno = seqno; ), - TP_printk("dev=%u, seqno=%u", __entry->dev, __entry->seqno) + TP_printk("dev=%u, ring=%d, seqno=%u", + __entry->dev, __entry->ring, __entry->seqno) ); DEFINE_EVENT(radeon_fence_request, radeon_fence_emit, - TP_PROTO(struct drm_device *dev, u32 seqno), + TP_PROTO(struct drm_device *dev, int ring, u32 seqno), - TP_ARGS(dev, seqno) + TP_ARGS(dev, ring, seqno) ); -DEFINE_EVENT(radeon_fence_request, radeon_fence_retire, +DEFINE_EVENT(radeon_fence_request, radeon_fence_wait_begin, - TP_PROTO(struct drm_device *dev, u32 seqno), + TP_PROTO(struct drm_device *dev, int ring, u32 seqno), - TP_ARGS(dev, seqno) + TP_ARGS(dev, ring, seqno) ); -DEFINE_EVENT(radeon_fence_request, radeon_fence_wait_begin, +DEFINE_EVENT(radeon_fence_request, radeon_fence_wait_end, - TP_PROTO(struct drm_device *dev, u32 seqno), + TP_PROTO(struct drm_device *dev, int ring, u32 seqno), - TP_ARGS(dev, seqno) + TP_ARGS(dev, ring, seqno) ); -DEFINE_EVENT(radeon_fence_request, radeon_fence_wait_end, +DECLARE_EVENT_CLASS(radeon_semaphore_request, + + TP_PROTO(int ring, struct radeon_semaphore *sem), + + TP_ARGS(ring, sem), + + TP_STRUCT__entry( + __field(int, ring) + __field(signed, waiters) + __field(uint64_t, gpu_addr) + ), + + TP_fast_assign( + __entry->ring = ring; + __entry->waiters = sem->waiters; + __entry->gpu_addr = sem->gpu_addr; + ), + + TP_printk("ring=%u, waiters=%d, addr=%010Lx", __entry->ring, + __entry->waiters, __entry->gpu_addr) +); + +DEFINE_EVENT(radeon_semaphore_request, radeon_semaphore_signale, + + TP_PROTO(int ring, struct radeon_semaphore *sem), + + TP_ARGS(ring, sem) +); + +DEFINE_EVENT(radeon_semaphore_request, radeon_semaphore_wait, - TP_PROTO(struct drm_device *dev, u32 seqno), + TP_PROTO(int ring, struct radeon_semaphore *sem), - TP_ARGS(dev, seqno) + TP_ARGS(ring, sem) ); #endif diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c index 6c0ce8915fa..c8a8a5144ec 100644 --- a/drivers/gpu/drm/radeon/radeon_ttm.c +++ b/drivers/gpu/drm/radeon/radeon_ttm.c @@ -39,12 +39,14 @@ #include <linux/seq_file.h> #include <linux/slab.h> #include <linux/swiotlb.h> +#include <linux/debugfs.h> #include "radeon_reg.h" #include "radeon.h" #define DRM_FILE_PAGE_OFFSET (0x100000000ULL >> PAGE_SHIFT) static int radeon_ttm_debugfs_init(struct radeon_device *rdev); +static void radeon_ttm_debugfs_fini(struct radeon_device *rdev); static struct radeon_device *radeon_get_rdev(struct ttm_bo_device *bdev) { @@ -142,7 +144,7 @@ static int radeon_init_mem_type(struct ttm_bo_device *bdev, uint32_t type, man->flags = TTM_MEMTYPE_FLAG_MAPPABLE | TTM_MEMTYPE_FLAG_CMA; #if __OS_HAS_AGP if (rdev->flags & RADEON_IS_AGP) { - if (!(drm_core_has_AGP(rdev->ddev) && rdev->ddev->agp)) { + if (!rdev->ddev->agp) { DRM_ERROR("AGP is not enabled for memory type %u\n", (unsigned)type); return -EINVAL; @@ -203,7 +205,9 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo, static int radeon_verify_access(struct ttm_buffer_object *bo, struct file *filp) { - return 0; + struct radeon_bo *rbo = container_of(bo, struct radeon_bo, tbo); + + return drm_vma_node_verify_access(&rbo->gem_base.vma_node, filp); } static void radeon_move_null(struct ttm_buffer_object *bo, @@ -402,8 +406,14 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, if (r) { memcpy: r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem); + if (r) { + return r; + } } - return r; + + /* update statistics */ + atomic64_add((u64)bo->num_pages << PAGE_SHIFT, &rdev->num_bytes_moved); + return 0; } static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem) @@ -697,7 +707,9 @@ int radeon_ttm_init(struct radeon_device *rdev) /* No others user of address space so set it to 0 */ r = ttm_bo_device_init(&rdev->mman.bdev, rdev->mman.bo_global_ref.ref.object, - &radeon_bo_driver, DRM_FILE_PAGE_OFFSET, + &radeon_bo_driver, + rdev->ddev->anon_inode->i_mapping, + DRM_FILE_PAGE_OFFSET, rdev->need_dma32); if (r) { DRM_ERROR("failed initializing buffer object driver(%d).\n", r); @@ -710,6 +722,9 @@ int radeon_ttm_init(struct radeon_device *rdev) DRM_ERROR("Failed initializing VRAM heap.\n"); return r; } + /* Change the size here instead of the init above so only lpfn is affected */ + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + r = radeon_bo_create(rdev, 256 * 1024, PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->stollen_vga_memory); @@ -735,7 +750,6 @@ int radeon_ttm_init(struct radeon_device *rdev) } DRM_INFO("radeon: %uM of GTT memory ready.\n", (unsigned)(rdev->mc.gtt_size / (1024 * 1024))); - rdev->mman.bdev.dev_mapping = rdev->ddev->dev_mapping; r = radeon_ttm_debugfs_init(rdev); if (r) { @@ -751,6 +765,7 @@ void radeon_ttm_fini(struct radeon_device *rdev) if (!rdev->mman.initialized) return; + radeon_ttm_debugfs_fini(rdev); if (rdev->stollen_vga_memory) { r = radeon_bo_reserve(rdev->stollen_vga_memory, false); if (r == 0) { @@ -830,16 +845,15 @@ int radeon_mmap(struct file *filp, struct vm_area_struct *vma) return 0; } - -#define RADEON_DEBUGFS_MEM_TYPES 2 - #if defined(CONFIG_DEBUG_FS) + static int radeon_mm_dump_table(struct seq_file *m, void *data) { struct drm_info_node *node = (struct drm_info_node *)m->private; - struct drm_mm *mm = (struct drm_mm *)node->info_ent->data; + unsigned ttm_pl = *(int *)node->info_ent->data; struct drm_device *dev = node->minor->dev; struct radeon_device *rdev = dev->dev_private; + struct drm_mm *mm = (struct drm_mm *)rdev->mman.bdev.man[ttm_pl].priv; int ret; struct ttm_bo_global *glob = rdev->mman.bdev.glob; @@ -848,46 +862,169 @@ static int radeon_mm_dump_table(struct seq_file *m, void *data) spin_unlock(&glob->lru_lock); return ret; } + +static int ttm_pl_vram = TTM_PL_VRAM; +static int ttm_pl_tt = TTM_PL_TT; + +static struct drm_info_list radeon_ttm_debugfs_list[] = { + {"radeon_vram_mm", radeon_mm_dump_table, 0, &ttm_pl_vram}, + {"radeon_gtt_mm", radeon_mm_dump_table, 0, &ttm_pl_tt}, + {"ttm_page_pool", ttm_page_alloc_debugfs, 0, NULL}, +#ifdef CONFIG_SWIOTLB + {"ttm_dma_page_pool", ttm_dma_page_alloc_debugfs, 0, NULL} #endif +}; -static int radeon_ttm_debugfs_init(struct radeon_device *rdev) +static int radeon_ttm_vram_open(struct inode *inode, struct file *filep) { -#if defined(CONFIG_DEBUG_FS) - static struct drm_info_list radeon_mem_types_list[RADEON_DEBUGFS_MEM_TYPES+2]; - static char radeon_mem_types_names[RADEON_DEBUGFS_MEM_TYPES+2][32]; - unsigned i; + struct radeon_device *rdev = inode->i_private; + i_size_write(inode, rdev->mc.mc_vram_size); + filep->private_data = inode->i_private; + return 0; +} - for (i = 0; i < RADEON_DEBUGFS_MEM_TYPES; i++) { - if (i == 0) - sprintf(radeon_mem_types_names[i], "radeon_vram_mm"); - else - sprintf(radeon_mem_types_names[i], "radeon_gtt_mm"); - radeon_mem_types_list[i].name = radeon_mem_types_names[i]; - radeon_mem_types_list[i].show = &radeon_mm_dump_table; - radeon_mem_types_list[i].driver_features = 0; - if (i == 0) - radeon_mem_types_list[i].data = rdev->mman.bdev.man[TTM_PL_VRAM].priv; - else - radeon_mem_types_list[i].data = rdev->mman.bdev.man[TTM_PL_TT].priv; +static ssize_t radeon_ttm_vram_read(struct file *f, char __user *buf, + size_t size, loff_t *pos) +{ + struct radeon_device *rdev = f->private_data; + ssize_t result = 0; + int r; + + if (size & 0x3 || *pos & 0x3) + return -EINVAL; + + while (size) { + unsigned long flags; + uint32_t value; + + if (*pos >= rdev->mc.mc_vram_size) + return result; + + spin_lock_irqsave(&rdev->mmio_idx_lock, flags); + WREG32(RADEON_MM_INDEX, ((uint32_t)*pos) | 0x80000000); + if (rdev->family >= CHIP_CEDAR) + WREG32(EVERGREEN_MM_INDEX_HI, *pos >> 31); + value = RREG32(RADEON_MM_DATA); + spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags); + r = put_user(value, (uint32_t *)buf); + if (r) + return r; + + result += 4; + buf += 4; + *pos += 4; + size -= 4; } - /* Add ttm page pool to debugfs */ - sprintf(radeon_mem_types_names[i], "ttm_page_pool"); - radeon_mem_types_list[i].name = radeon_mem_types_names[i]; - radeon_mem_types_list[i].show = &ttm_page_alloc_debugfs; - radeon_mem_types_list[i].driver_features = 0; - radeon_mem_types_list[i++].data = NULL; -#ifdef CONFIG_SWIOTLB - if (swiotlb_nr_tbl()) { - sprintf(radeon_mem_types_names[i], "ttm_dma_page_pool"); - radeon_mem_types_list[i].name = radeon_mem_types_names[i]; - radeon_mem_types_list[i].show = &ttm_dma_page_alloc_debugfs; - radeon_mem_types_list[i].driver_features = 0; - radeon_mem_types_list[i++].data = NULL; + + return result; +} + +static const struct file_operations radeon_ttm_vram_fops = { + .owner = THIS_MODULE, + .open = radeon_ttm_vram_open, + .read = radeon_ttm_vram_read, + .llseek = default_llseek +}; + +static int radeon_ttm_gtt_open(struct inode *inode, struct file *filep) +{ + struct radeon_device *rdev = inode->i_private; + i_size_write(inode, rdev->mc.gtt_size); + filep->private_data = inode->i_private; + return 0; +} + +static ssize_t radeon_ttm_gtt_read(struct file *f, char __user *buf, + size_t size, loff_t *pos) +{ + struct radeon_device *rdev = f->private_data; + ssize_t result = 0; + int r; + + while (size) { + loff_t p = *pos / PAGE_SIZE; + unsigned off = *pos & ~PAGE_MASK; + size_t cur_size = min_t(size_t, size, PAGE_SIZE - off); + struct page *page; + void *ptr; + + if (p >= rdev->gart.num_cpu_pages) + return result; + + page = rdev->gart.pages[p]; + if (page) { + ptr = kmap(page); + ptr += off; + + r = copy_to_user(buf, ptr, cur_size); + kunmap(rdev->gart.pages[p]); + } else + r = clear_user(buf, cur_size); + + if (r) + return -EFAULT; + + result += cur_size; + buf += cur_size; + *pos += cur_size; + size -= cur_size; } + + return result; +} + +static const struct file_operations radeon_ttm_gtt_fops = { + .owner = THIS_MODULE, + .open = radeon_ttm_gtt_open, + .read = radeon_ttm_gtt_read, + .llseek = default_llseek +}; + #endif - return radeon_debugfs_add_files(rdev, radeon_mem_types_list, i); +static int radeon_ttm_debugfs_init(struct radeon_device *rdev) +{ +#if defined(CONFIG_DEBUG_FS) + unsigned count; + + struct drm_minor *minor = rdev->ddev->primary; + struct dentry *ent, *root = minor->debugfs_root; + + ent = debugfs_create_file("radeon_vram", S_IFREG | S_IRUGO, root, + rdev, &radeon_ttm_vram_fops); + if (IS_ERR(ent)) + return PTR_ERR(ent); + rdev->mman.vram = ent; + + ent = debugfs_create_file("radeon_gtt", S_IFREG | S_IRUGO, root, + rdev, &radeon_ttm_gtt_fops); + if (IS_ERR(ent)) + return PTR_ERR(ent); + rdev->mman.gtt = ent; + + count = ARRAY_SIZE(radeon_ttm_debugfs_list); + +#ifdef CONFIG_SWIOTLB + if (!swiotlb_nr_tbl()) + --count; #endif + + return radeon_debugfs_add_files(rdev, radeon_ttm_debugfs_list, count); +#else + return 0; +#endif +} + +static void radeon_ttm_debugfs_fini(struct radeon_device *rdev) +{ +#if defined(CONFIG_DEBUG_FS) + + debugfs_remove(rdev->mman.vram); + rdev->mman.vram = NULL; + + debugfs_remove(rdev->mman.gtt); + rdev->mman.gtt = NULL; +#endif } diff --git a/drivers/gpu/drm/radeon/radeon_ucode.h b/drivers/gpu/drm/radeon/radeon_ucode.h index d8b05f7bcf1..4e7c3269b18 100644 --- a/drivers/gpu/drm/radeon/radeon_ucode.h +++ b/drivers/gpu/drm/radeon/radeon_ucode.h @@ -35,6 +35,12 @@ #define SI_PFP_UCODE_SIZE 2144 #define SI_PM4_UCODE_SIZE 2144 #define SI_CE_UCODE_SIZE 2144 +#define CIK_PFP_UCODE_SIZE 2144 +#define CIK_ME_UCODE_SIZE 2144 +#define CIK_CE_UCODE_SIZE 2144 + +/* MEC */ +#define CIK_MEC_UCODE_SIZE 4192 /* RLC */ #define R600_RLC_UCODE_SIZE 768 @@ -43,12 +49,27 @@ #define CAYMAN_RLC_UCODE_SIZE 1024 #define ARUBA_RLC_UCODE_SIZE 1536 #define SI_RLC_UCODE_SIZE 2048 +#define BONAIRE_RLC_UCODE_SIZE 2048 +#define KB_RLC_UCODE_SIZE 2560 +#define KV_RLC_UCODE_SIZE 2560 +#define ML_RLC_UCODE_SIZE 2560 /* MC */ #define BTC_MC_UCODE_SIZE 6024 #define CAYMAN_MC_UCODE_SIZE 6037 #define SI_MC_UCODE_SIZE 7769 +#define TAHITI_MC_UCODE_SIZE 7808 +#define PITCAIRN_MC_UCODE_SIZE 7775 +#define VERDE_MC_UCODE_SIZE 7875 #define OLAND_MC_UCODE_SIZE 7863 +#define BONAIRE_MC_UCODE_SIZE 7866 +#define BONAIRE_MC2_UCODE_SIZE 7948 +#define HAWAII_MC_UCODE_SIZE 7933 +#define HAWAII_MC2_UCODE_SIZE 8091 + +/* SDMA */ +#define CIK_SDMA_UCODE_SIZE 1050 +#define CIK_SDMA_UCODE_VERSION 64 /* SMC */ #define RV770_SMC_UCODE_START 0x0100 @@ -126,4 +147,10 @@ #define HAINAN_SMC_UCODE_START 0x10000 #define HAINAN_SMC_UCODE_SIZE 0xe67C +#define BONAIRE_SMC_UCODE_START 0x20000 +#define BONAIRE_SMC_UCODE_SIZE 0x1FDEC + +#define HAWAII_SMC_UCODE_START 0x20000 +#define HAWAII_SMC_UCODE_SIZE 0x1FDEC + #endif diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c index 41efcec28cd..a4ad270e826 100644 --- a/drivers/gpu/drm/radeon/radeon_uvd.c +++ b/drivers/gpu/drm/radeon/radeon_uvd.c @@ -56,20 +56,12 @@ static void radeon_uvd_idle_work_handler(struct work_struct *work); int radeon_uvd_init(struct radeon_device *rdev) { - struct platform_device *pdev; unsigned long bo_size; const char *fw_name; int i, r; INIT_DELAYED_WORK(&rdev->uvd.idle_work, radeon_uvd_idle_work_handler); - pdev = platform_device_register_simple("radeon_uvd", 0, NULL, 0); - r = IS_ERR(pdev); - if (r) { - dev_err(rdev->dev, "radeon_uvd: Failed to register firmware\n"); - return -EINVAL; - } - switch (rdev->family) { case CHIP_RV710: case CHIP_RV730: @@ -99,12 +91,15 @@ int radeon_uvd_init(struct radeon_device *rdev) case CHIP_VERDE: case CHIP_PITCAIRN: case CHIP_ARUBA: + case CHIP_OLAND: fw_name = FIRMWARE_TAHITI; break; case CHIP_BONAIRE: case CHIP_KABINI: case CHIP_KAVERI: + case CHIP_HAWAII: + case CHIP_MULLINS: fw_name = FIRMWARE_BONAIRE; break; @@ -112,16 +107,13 @@ int radeon_uvd_init(struct radeon_device *rdev) return -EINVAL; } - r = request_firmware(&rdev->uvd_fw, fw_name, &pdev->dev); + r = request_firmware(&rdev->uvd_fw, fw_name, rdev->dev); if (r) { dev_err(rdev->dev, "radeon_uvd: Can't load firmware \"%s\"\n", fw_name); - platform_device_unregister(pdev); return r; } - platform_device_unregister(pdev); - bo_size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 8) + RADEON_UVD_STACK_SIZE + RADEON_UVD_HEAP_SIZE; r = radeon_bo_create(rdev, bo_size, PAGE_SIZE, true, @@ -131,20 +123,34 @@ int radeon_uvd_init(struct radeon_device *rdev) return r; } - r = radeon_uvd_resume(rdev); - if (r) + r = radeon_bo_reserve(rdev->uvd.vcpu_bo, false); + if (r) { + radeon_bo_unref(&rdev->uvd.vcpu_bo); + dev_err(rdev->dev, "(%d) failed to reserve UVD bo\n", r); return r; + } - memset(rdev->uvd.cpu_addr, 0, bo_size); - memcpy(rdev->uvd.cpu_addr, rdev->uvd_fw->data, rdev->uvd_fw->size); + r = radeon_bo_pin(rdev->uvd.vcpu_bo, RADEON_GEM_DOMAIN_VRAM, + &rdev->uvd.gpu_addr); + if (r) { + radeon_bo_unreserve(rdev->uvd.vcpu_bo); + radeon_bo_unref(&rdev->uvd.vcpu_bo); + dev_err(rdev->dev, "(%d) UVD bo pin failed\n", r); + return r; + } - r = radeon_uvd_suspend(rdev); - if (r) + r = radeon_bo_kmap(rdev->uvd.vcpu_bo, &rdev->uvd.cpu_addr); + if (r) { + dev_err(rdev->dev, "(%d) UVD map failed\n", r); return r; + } + + radeon_bo_unreserve(rdev->uvd.vcpu_bo); for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) { atomic_set(&rdev->uvd.handles[i], 0); rdev->uvd.filp[i] = NULL; + rdev->uvd.img_size[i] = 0; } return 0; @@ -152,70 +158,75 @@ int radeon_uvd_init(struct radeon_device *rdev) void radeon_uvd_fini(struct radeon_device *rdev) { - radeon_uvd_suspend(rdev); - radeon_bo_unref(&rdev->uvd.vcpu_bo); -} - -int radeon_uvd_suspend(struct radeon_device *rdev) -{ int r; if (rdev->uvd.vcpu_bo == NULL) - return 0; + return; r = radeon_bo_reserve(rdev->uvd.vcpu_bo, false); if (!r) { radeon_bo_kunmap(rdev->uvd.vcpu_bo); radeon_bo_unpin(rdev->uvd.vcpu_bo); - rdev->uvd.cpu_addr = NULL; - if (!radeon_bo_pin(rdev->uvd.vcpu_bo, RADEON_GEM_DOMAIN_CPU, NULL)) { - radeon_bo_kmap(rdev->uvd.vcpu_bo, &rdev->uvd.cpu_addr); - } radeon_bo_unreserve(rdev->uvd.vcpu_bo); - - if (rdev->uvd.cpu_addr) { - radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_UVD_INDEX); - } else { - rdev->fence_drv[R600_RING_TYPE_UVD_INDEX].cpu_addr = NULL; - } } - return r; + + radeon_bo_unref(&rdev->uvd.vcpu_bo); + + radeon_ring_fini(rdev, &rdev->ring[R600_RING_TYPE_UVD_INDEX]); + + release_firmware(rdev->uvd_fw); +} + +int radeon_uvd_suspend(struct radeon_device *rdev) +{ + unsigned size; + void *ptr; + int i; + + if (rdev->uvd.vcpu_bo == NULL) + return 0; + + for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) + if (atomic_read(&rdev->uvd.handles[i])) + break; + + if (i == RADEON_MAX_UVD_HANDLES) + return 0; + + size = radeon_bo_size(rdev->uvd.vcpu_bo); + size -= rdev->uvd_fw->size; + + ptr = rdev->uvd.cpu_addr; + ptr += rdev->uvd_fw->size; + + rdev->uvd.saved_bo = kmalloc(size, GFP_KERNEL); + memcpy(rdev->uvd.saved_bo, ptr, size); + + return 0; } int radeon_uvd_resume(struct radeon_device *rdev) { - int r; + unsigned size; + void *ptr; if (rdev->uvd.vcpu_bo == NULL) return -EINVAL; - r = radeon_bo_reserve(rdev->uvd.vcpu_bo, false); - if (r) { - radeon_bo_unref(&rdev->uvd.vcpu_bo); - dev_err(rdev->dev, "(%d) failed to reserve UVD bo\n", r); - return r; - } + memcpy(rdev->uvd.cpu_addr, rdev->uvd_fw->data, rdev->uvd_fw->size); - /* Have been pin in cpu unmap unpin */ - radeon_bo_kunmap(rdev->uvd.vcpu_bo); - radeon_bo_unpin(rdev->uvd.vcpu_bo); + size = radeon_bo_size(rdev->uvd.vcpu_bo); + size -= rdev->uvd_fw->size; - r = radeon_bo_pin(rdev->uvd.vcpu_bo, RADEON_GEM_DOMAIN_VRAM, - &rdev->uvd.gpu_addr); - if (r) { - radeon_bo_unreserve(rdev->uvd.vcpu_bo); - radeon_bo_unref(&rdev->uvd.vcpu_bo); - dev_err(rdev->dev, "(%d) UVD bo pin failed\n", r); - return r; - } + ptr = rdev->uvd.cpu_addr; + ptr += rdev->uvd_fw->size; - r = radeon_bo_kmap(rdev->uvd.vcpu_bo, &rdev->uvd.cpu_addr); - if (r) { - dev_err(rdev->dev, "(%d) UVD map failed\n", r); - return r; - } - - radeon_bo_unreserve(rdev->uvd.vcpu_bo); + if (rdev->uvd.saved_bo != NULL) { + memcpy(ptr, rdev->uvd.saved_bo, size); + kfree(rdev->uvd.saved_bo); + rdev->uvd.saved_bo = NULL; + } else + memset(ptr, 0, size); return 0; } @@ -230,10 +241,12 @@ void radeon_uvd_free_handles(struct radeon_device *rdev, struct drm_file *filp) { int i, r; for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) { - if (rdev->uvd.filp[i] == filp) { - uint32_t handle = atomic_read(&rdev->uvd.handles[i]); + uint32_t handle = atomic_read(&rdev->uvd.handles[i]); + if (handle != 0 && rdev->uvd.filp[i] == filp) { struct radeon_fence *fence; + radeon_uvd_note_usage(rdev); + r = radeon_uvd_get_destroy_msg(rdev, R600_RING_TYPE_UVD_INDEX, handle, &fence); if (r) { @@ -342,6 +355,7 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo, unsigned offset, unsigned buf_sizes[]) { int32_t *msg, msg_type, handle; + unsigned img_size = 0; void *ptr; int i, r; @@ -351,9 +365,19 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo, return -EINVAL; } + if (bo->tbo.sync_obj) { + r = radeon_fence_wait(bo->tbo.sync_obj, false); + if (r) { + DRM_ERROR("Failed waiting for UVD message (%d)!\n", r); + return r; + } + } + r = radeon_bo_kmap(bo, &ptr); - if (r) + if (r) { + DRM_ERROR("Failed mapping the UVD message (%d)!\n", r); return r; + } msg = ptr + offset; @@ -368,6 +392,8 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo, if (msg_type == 1) { /* it's a decode msg, calc buffer sizes */ r = radeon_uvd_cs_msg_decode(msg, buf_sizes); + /* calc image size (width * height) */ + img_size = msg[6] * msg[7]; radeon_bo_kunmap(bo); if (r) return r; @@ -379,8 +405,16 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo, radeon_bo_kunmap(bo); return 0; } else { - /* it's a create msg, no special handling needed */ + /* it's a create msg, calc image size (width * height) */ + img_size = msg[7] * msg[8]; radeon_bo_kunmap(bo); + + if (msg_type != 0) { + DRM_ERROR("Illegal UVD message type (%d)!\n", msg_type); + return -EINVAL; + } + + /* it's a create msg, no special handling needed */ } /* create or decode, validate the handle */ @@ -393,6 +427,7 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo, for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) { if (!atomic_cmpxchg(&p->rdev->uvd.handles[i], 0, handle)) { p->rdev->uvd.filp[i] = p->filp; + p->rdev->uvd.img_size[i] = img_size; return 0; } } @@ -403,7 +438,7 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo, static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p, int data0, int data1, - unsigned buf_sizes[]) + unsigned buf_sizes[], bool *has_msg_cmd) { struct radeon_cs_chunk *relocs_chunk; struct radeon_cs_reloc *reloc; @@ -421,7 +456,7 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p, } reloc = p->relocs_ptr[(idx / 4)]; - start = reloc->lobj.gpu_offset; + start = reloc->gpu_offset; end = start + radeon_bo_size(reloc->robj); start += offset; @@ -431,8 +466,12 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p, cmd = radeon_get_ib_value(p, p->idx) >> 1; if (cmd < 0x4) { + if (end <= start) { + DRM_ERROR("invalid reloc offset %X!\n", offset); + return -EINVAL; + } if ((end - start) < buf_sizes[cmd]) { - DRM_ERROR("buffer to small (%d / %d)!\n", + DRM_ERROR("buffer (%d) to small (%d / %d)!\n", cmd, (unsigned)(end - start), buf_sizes[cmd]); return -EINVAL; } @@ -442,7 +481,7 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p, return -EINVAL; } - if ((start >> 28) != (end >> 28)) { + if ((start >> 28) != ((end - 1) >> 28)) { DRM_ERROR("reloc %LX-%LX crossing 256MB boundary!\n", start, end); return -EINVAL; @@ -457,9 +496,17 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p, } if (cmd == 0) { + if (*has_msg_cmd) { + DRM_ERROR("More than one message in a UVD-IB!\n"); + return -EINVAL; + } + *has_msg_cmd = true; r = radeon_uvd_cs_msg(p, reloc->robj, offset, buf_sizes); if (r) return r; + } else if (!*has_msg_cmd) { + DRM_ERROR("Message needed before other commands are send!\n"); + return -EINVAL; } return 0; @@ -468,7 +515,8 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p, static int radeon_uvd_cs_reg(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt, int *data0, int *data1, - unsigned buf_sizes[]) + unsigned buf_sizes[], + bool *has_msg_cmd) { int i, r; @@ -482,7 +530,8 @@ static int radeon_uvd_cs_reg(struct radeon_cs_parser *p, *data1 = p->idx; break; case UVD_GPCOM_VCPU_CMD: - r = radeon_uvd_cs_reloc(p, *data0, *data1, buf_sizes); + r = radeon_uvd_cs_reloc(p, *data0, *data1, + buf_sizes, has_msg_cmd); if (r) return r; break; @@ -503,6 +552,9 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p) struct radeon_cs_packet pkt; int r, data0 = 0, data1 = 0; + /* does the IB has a msg command */ + bool has_msg_cmd = false; + /* minimum buffer sizes */ unsigned buf_sizes[] = { [0x00000000] = 2048, @@ -529,8 +581,8 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p) return r; switch (pkt.type) { case RADEON_PACKET_TYPE0: - r = radeon_uvd_cs_reg(p, &pkt, &data0, - &data1, buf_sizes); + r = radeon_uvd_cs_reg(p, &pkt, &data0, &data1, + buf_sizes, &has_msg_cmd); if (r) return r; break; @@ -542,6 +594,12 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p) return -EINVAL; } } while (p->idx < p->chunks[p->chunk_ib_idx].length_dw); + + if (!has_msg_cmd) { + DRM_ERROR("UVD-IBs need a msg command!\n"); + return -EINVAL; + } + return 0; } @@ -573,7 +631,7 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev, if (r) goto err; - r = radeon_ib_get(rdev, ring, &ib, NULL, 16); + r = radeon_ib_get(rdev, ring, &ib, NULL, 64); if (r) goto err; @@ -693,6 +751,34 @@ int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring, return radeon_uvd_send_msg(rdev, ring, bo, fence); } +/** + * radeon_uvd_count_handles - count number of open streams + * + * @rdev: radeon_device pointer + * @sd: number of SD streams + * @hd: number of HD streams + * + * Count the number of open SD/HD streams as a hint for power mangement + */ +static void radeon_uvd_count_handles(struct radeon_device *rdev, + unsigned *sd, unsigned *hd) +{ + unsigned i; + + *sd = 0; + *hd = 0; + + for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) { + if (!atomic_read(&rdev->uvd.handles[i])) + continue; + + if (rdev->uvd.img_size[i] >= 720*576) + ++(*hd); + else + ++(*sd); + } +} + static void radeon_uvd_idle_work_handler(struct work_struct *work) { struct radeon_device *rdev = @@ -700,10 +786,9 @@ static void radeon_uvd_idle_work_handler(struct work_struct *work) if (radeon_fence_count_emitted(rdev, R600_RING_TYPE_UVD_INDEX) == 0) { if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { - mutex_lock(&rdev->pm.mutex); - rdev->pm.dpm.uvd_active = false; - mutex_unlock(&rdev->pm.mutex); - radeon_pm_compute_clocks(rdev); + radeon_uvd_count_handles(rdev, &rdev->pm.dpm.sd, + &rdev->pm.dpm.hd); + radeon_dpm_enable_uvd(rdev, false); } else { radeon_set_uvd_clocks(rdev, 0, 0); } @@ -715,13 +800,26 @@ static void radeon_uvd_idle_work_handler(struct work_struct *work) void radeon_uvd_note_usage(struct radeon_device *rdev) { + bool streams_changed = false; bool set_clocks = !cancel_delayed_work_sync(&rdev->uvd.idle_work); set_clocks &= schedule_delayed_work(&rdev->uvd.idle_work, msecs_to_jiffies(UVD_IDLE_TIMEOUT_MS)); - if (set_clocks) { + + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { + unsigned hd = 0, sd = 0; + radeon_uvd_count_handles(rdev, &sd, &hd); + if ((rdev->pm.dpm.sd != sd) || + (rdev->pm.dpm.hd != hd)) { + rdev->pm.dpm.sd = sd; + rdev->pm.dpm.hd = hd; + /* disable this for now */ + /*streams_changed = true;*/ + } + } + + if (set_clocks || streams_changed) { if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { - /* XXX pick SD/HD/MVC */ - radeon_dpm_enable_power_state(rdev, POWER_STATE_TYPE_INTERNAL_UVD); + radeon_dpm_enable_uvd(rdev, true); } else { radeon_set_uvd_clocks(rdev, 53300, 40000); } diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c new file mode 100644 index 00000000000..aa21c31a846 --- /dev/null +++ b/drivers/gpu/drm/radeon/radeon_vce.c @@ -0,0 +1,771 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * Authors: Christian König <christian.koenig@amd.com> + */ + +#include <linux/firmware.h> +#include <linux/module.h> +#include <drm/drmP.h> +#include <drm/drm.h> + +#include "radeon.h" +#include "radeon_asic.h" +#include "sid.h" + +/* 1 second timeout */ +#define VCE_IDLE_TIMEOUT_MS 1000 + +/* Firmware Names */ +#define FIRMWARE_BONAIRE "radeon/BONAIRE_vce.bin" + +MODULE_FIRMWARE(FIRMWARE_BONAIRE); + +static void radeon_vce_idle_work_handler(struct work_struct *work); + +/** + * radeon_vce_init - allocate memory, load vce firmware + * + * @rdev: radeon_device pointer + * + * First step to get VCE online, allocate memory and load the firmware + */ +int radeon_vce_init(struct radeon_device *rdev) +{ + static const char *fw_version = "[ATI LIB=VCEFW,"; + static const char *fb_version = "[ATI LIB=VCEFWSTATS,"; + unsigned long size; + const char *fw_name, *c; + uint8_t start, mid, end; + int i, r; + + INIT_DELAYED_WORK(&rdev->vce.idle_work, radeon_vce_idle_work_handler); + + switch (rdev->family) { + case CHIP_BONAIRE: + case CHIP_KAVERI: + case CHIP_KABINI: + case CHIP_HAWAII: + case CHIP_MULLINS: + fw_name = FIRMWARE_BONAIRE; + break; + + default: + return -EINVAL; + } + + r = request_firmware(&rdev->vce_fw, fw_name, rdev->dev); + if (r) { + dev_err(rdev->dev, "radeon_vce: Can't load firmware \"%s\"\n", + fw_name); + return r; + } + + /* search for firmware version */ + + size = rdev->vce_fw->size - strlen(fw_version) - 9; + c = rdev->vce_fw->data; + for (;size > 0; --size, ++c) + if (strncmp(c, fw_version, strlen(fw_version)) == 0) + break; + + if (size == 0) + return -EINVAL; + + c += strlen(fw_version); + if (sscanf(c, "%2hhd.%2hhd.%2hhd]", &start, &mid, &end) != 3) + return -EINVAL; + + /* search for feedback version */ + + size = rdev->vce_fw->size - strlen(fb_version) - 3; + c = rdev->vce_fw->data; + for (;size > 0; --size, ++c) + if (strncmp(c, fb_version, strlen(fb_version)) == 0) + break; + + if (size == 0) + return -EINVAL; + + c += strlen(fb_version); + if (sscanf(c, "%2u]", &rdev->vce.fb_version) != 1) + return -EINVAL; + + DRM_INFO("Found VCE firmware/feedback version %hhd.%hhd.%hhd / %d!\n", + start, mid, end, rdev->vce.fb_version); + + rdev->vce.fw_version = (start << 24) | (mid << 16) | (end << 8); + + /* we can only work with this fw version for now */ + if (rdev->vce.fw_version != ((40 << 24) | (2 << 16) | (2 << 8))) + return -EINVAL; + + /* allocate firmware, stack and heap BO */ + + size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size) + + RADEON_VCE_STACK_SIZE + RADEON_VCE_HEAP_SIZE; + r = radeon_bo_create(rdev, size, PAGE_SIZE, true, + RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->vce.vcpu_bo); + if (r) { + dev_err(rdev->dev, "(%d) failed to allocate VCE bo\n", r); + return r; + } + + r = radeon_bo_reserve(rdev->vce.vcpu_bo, false); + if (r) { + radeon_bo_unref(&rdev->vce.vcpu_bo); + dev_err(rdev->dev, "(%d) failed to reserve VCE bo\n", r); + return r; + } + + r = radeon_bo_pin(rdev->vce.vcpu_bo, RADEON_GEM_DOMAIN_VRAM, + &rdev->vce.gpu_addr); + radeon_bo_unreserve(rdev->vce.vcpu_bo); + if (r) { + radeon_bo_unref(&rdev->vce.vcpu_bo); + dev_err(rdev->dev, "(%d) VCE bo pin failed\n", r); + return r; + } + + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) { + atomic_set(&rdev->vce.handles[i], 0); + rdev->vce.filp[i] = NULL; + } + + return 0; +} + +/** + * radeon_vce_fini - free memory + * + * @rdev: radeon_device pointer + * + * Last step on VCE teardown, free firmware memory + */ +void radeon_vce_fini(struct radeon_device *rdev) +{ + if (rdev->vce.vcpu_bo == NULL) + return; + + radeon_bo_unref(&rdev->vce.vcpu_bo); + + release_firmware(rdev->vce_fw); +} + +/** + * radeon_vce_suspend - unpin VCE fw memory + * + * @rdev: radeon_device pointer + * + */ +int radeon_vce_suspend(struct radeon_device *rdev) +{ + int i; + + if (rdev->vce.vcpu_bo == NULL) + return 0; + + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) + if (atomic_read(&rdev->vce.handles[i])) + break; + + if (i == RADEON_MAX_VCE_HANDLES) + return 0; + + /* TODO: suspending running encoding sessions isn't supported */ + return -EINVAL; +} + +/** + * radeon_vce_resume - pin VCE fw memory + * + * @rdev: radeon_device pointer + * + */ +int radeon_vce_resume(struct radeon_device *rdev) +{ + void *cpu_addr; + int r; + + if (rdev->vce.vcpu_bo == NULL) + return -EINVAL; + + r = radeon_bo_reserve(rdev->vce.vcpu_bo, false); + if (r) { + dev_err(rdev->dev, "(%d) failed to reserve VCE bo\n", r); + return r; + } + + r = radeon_bo_kmap(rdev->vce.vcpu_bo, &cpu_addr); + if (r) { + radeon_bo_unreserve(rdev->vce.vcpu_bo); + dev_err(rdev->dev, "(%d) VCE map failed\n", r); + return r; + } + + memcpy(cpu_addr, rdev->vce_fw->data, rdev->vce_fw->size); + + radeon_bo_kunmap(rdev->vce.vcpu_bo); + + radeon_bo_unreserve(rdev->vce.vcpu_bo); + + return 0; +} + +/** + * radeon_vce_idle_work_handler - power off VCE + * + * @work: pointer to work structure + * + * power of VCE when it's not used any more + */ +static void radeon_vce_idle_work_handler(struct work_struct *work) +{ + struct radeon_device *rdev = + container_of(work, struct radeon_device, vce.idle_work.work); + + if ((radeon_fence_count_emitted(rdev, TN_RING_TYPE_VCE1_INDEX) == 0) && + (radeon_fence_count_emitted(rdev, TN_RING_TYPE_VCE2_INDEX) == 0)) { + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { + radeon_dpm_enable_vce(rdev, false); + } else { + radeon_set_vce_clocks(rdev, 0, 0); + } + } else { + schedule_delayed_work(&rdev->vce.idle_work, + msecs_to_jiffies(VCE_IDLE_TIMEOUT_MS)); + } +} + +/** + * radeon_vce_note_usage - power up VCE + * + * @rdev: radeon_device pointer + * + * Make sure VCE is powerd up when we want to use it + */ +void radeon_vce_note_usage(struct radeon_device *rdev) +{ + bool streams_changed = false; + bool set_clocks = !cancel_delayed_work_sync(&rdev->vce.idle_work); + set_clocks &= schedule_delayed_work(&rdev->vce.idle_work, + msecs_to_jiffies(VCE_IDLE_TIMEOUT_MS)); + + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { + /* XXX figure out if the streams changed */ + streams_changed = false; + } + + if (set_clocks || streams_changed) { + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { + radeon_dpm_enable_vce(rdev, true); + } else { + radeon_set_vce_clocks(rdev, 53300, 40000); + } + } +} + +/** + * radeon_vce_free_handles - free still open VCE handles + * + * @rdev: radeon_device pointer + * @filp: drm file pointer + * + * Close all VCE handles still open by this file pointer + */ +void radeon_vce_free_handles(struct radeon_device *rdev, struct drm_file *filp) +{ + int i, r; + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) { + uint32_t handle = atomic_read(&rdev->vce.handles[i]); + if (!handle || rdev->vce.filp[i] != filp) + continue; + + radeon_vce_note_usage(rdev); + + r = radeon_vce_get_destroy_msg(rdev, TN_RING_TYPE_VCE1_INDEX, + handle, NULL); + if (r) + DRM_ERROR("Error destroying VCE handle (%d)!\n", r); + + rdev->vce.filp[i] = NULL; + atomic_set(&rdev->vce.handles[i], 0); + } +} + +/** + * radeon_vce_get_create_msg - generate a VCE create msg + * + * @rdev: radeon_device pointer + * @ring: ring we should submit the msg to + * @handle: VCE session handle to use + * @fence: optional fence to return + * + * Open up a stream for HW test + */ +int radeon_vce_get_create_msg(struct radeon_device *rdev, int ring, + uint32_t handle, struct radeon_fence **fence) +{ + const unsigned ib_size_dw = 1024; + struct radeon_ib ib; + uint64_t dummy; + int i, r; + + r = radeon_ib_get(rdev, ring, &ib, NULL, ib_size_dw * 4); + if (r) { + DRM_ERROR("radeon: failed to get ib (%d).\n", r); + return r; + } + + dummy = ib.gpu_addr + 1024; + + /* stitch together an VCE create msg */ + ib.length_dw = 0; + ib.ptr[ib.length_dw++] = 0x0000000c; /* len */ + ib.ptr[ib.length_dw++] = 0x00000001; /* session cmd */ + ib.ptr[ib.length_dw++] = handle; + + ib.ptr[ib.length_dw++] = 0x00000030; /* len */ + ib.ptr[ib.length_dw++] = 0x01000001; /* create cmd */ + ib.ptr[ib.length_dw++] = 0x00000000; + ib.ptr[ib.length_dw++] = 0x00000042; + ib.ptr[ib.length_dw++] = 0x0000000a; + ib.ptr[ib.length_dw++] = 0x00000001; + ib.ptr[ib.length_dw++] = 0x00000080; + ib.ptr[ib.length_dw++] = 0x00000060; + ib.ptr[ib.length_dw++] = 0x00000100; + ib.ptr[ib.length_dw++] = 0x00000100; + ib.ptr[ib.length_dw++] = 0x0000000c; + ib.ptr[ib.length_dw++] = 0x00000000; + + ib.ptr[ib.length_dw++] = 0x00000014; /* len */ + ib.ptr[ib.length_dw++] = 0x05000005; /* feedback buffer */ + ib.ptr[ib.length_dw++] = upper_32_bits(dummy); + ib.ptr[ib.length_dw++] = dummy; + ib.ptr[ib.length_dw++] = 0x00000001; + + for (i = ib.length_dw; i < ib_size_dw; ++i) + ib.ptr[i] = 0x0; + + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); + } + + if (fence) + *fence = radeon_fence_ref(ib.fence); + + radeon_ib_free(rdev, &ib); + + return r; +} + +/** + * radeon_vce_get_destroy_msg - generate a VCE destroy msg + * + * @rdev: radeon_device pointer + * @ring: ring we should submit the msg to + * @handle: VCE session handle to use + * @fence: optional fence to return + * + * Close up a stream for HW test or if userspace failed to do so + */ +int radeon_vce_get_destroy_msg(struct radeon_device *rdev, int ring, + uint32_t handle, struct radeon_fence **fence) +{ + const unsigned ib_size_dw = 1024; + struct radeon_ib ib; + uint64_t dummy; + int i, r; + + r = radeon_ib_get(rdev, ring, &ib, NULL, ib_size_dw * 4); + if (r) { + DRM_ERROR("radeon: failed to get ib (%d).\n", r); + return r; + } + + dummy = ib.gpu_addr + 1024; + + /* stitch together an VCE destroy msg */ + ib.length_dw = 0; + ib.ptr[ib.length_dw++] = 0x0000000c; /* len */ + ib.ptr[ib.length_dw++] = 0x00000001; /* session cmd */ + ib.ptr[ib.length_dw++] = handle; + + ib.ptr[ib.length_dw++] = 0x00000014; /* len */ + ib.ptr[ib.length_dw++] = 0x05000005; /* feedback buffer */ + ib.ptr[ib.length_dw++] = upper_32_bits(dummy); + ib.ptr[ib.length_dw++] = dummy; + ib.ptr[ib.length_dw++] = 0x00000001; + + ib.ptr[ib.length_dw++] = 0x00000008; /* len */ + ib.ptr[ib.length_dw++] = 0x02000001; /* destroy cmd */ + + for (i = ib.length_dw; i < ib_size_dw; ++i) + ib.ptr[i] = 0x0; + + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); + } + + if (fence) + *fence = radeon_fence_ref(ib.fence); + + radeon_ib_free(rdev, &ib); + + return r; +} + +/** + * radeon_vce_cs_reloc - command submission relocation + * + * @p: parser context + * @lo: address of lower dword + * @hi: address of higher dword + * @size: size of checker for relocation buffer + * + * Patch relocation inside command stream with real buffer address + */ +int radeon_vce_cs_reloc(struct radeon_cs_parser *p, int lo, int hi, + unsigned size) +{ + struct radeon_cs_chunk *relocs_chunk; + struct radeon_cs_reloc *reloc; + uint64_t start, end, offset; + unsigned idx; + + relocs_chunk = &p->chunks[p->chunk_relocs_idx]; + offset = radeon_get_ib_value(p, lo); + idx = radeon_get_ib_value(p, hi); + + if (idx >= relocs_chunk->length_dw) { + DRM_ERROR("Relocs at %d after relocations chunk end %d !\n", + idx, relocs_chunk->length_dw); + return -EINVAL; + } + + reloc = p->relocs_ptr[(idx / 4)]; + start = reloc->gpu_offset; + end = start + radeon_bo_size(reloc->robj); + start += offset; + + p->ib.ptr[lo] = start & 0xFFFFFFFF; + p->ib.ptr[hi] = start >> 32; + + if (end <= start) { + DRM_ERROR("invalid reloc offset %llX!\n", offset); + return -EINVAL; + } + if ((end - start) < size) { + DRM_ERROR("buffer to small (%d / %d)!\n", + (unsigned)(end - start), size); + return -EINVAL; + } + + return 0; +} + +/** + * radeon_vce_validate_handle - validate stream handle + * + * @p: parser context + * @handle: handle to validate + * + * Validates the handle and return the found session index or -EINVAL + * we we don't have another free session index. + */ +int radeon_vce_validate_handle(struct radeon_cs_parser *p, uint32_t handle) +{ + unsigned i; + + /* validate the handle */ + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) { + if (atomic_read(&p->rdev->vce.handles[i]) == handle) + return i; + } + + /* handle not found try to alloc a new one */ + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) { + if (!atomic_cmpxchg(&p->rdev->vce.handles[i], 0, handle)) { + p->rdev->vce.filp[i] = p->filp; + p->rdev->vce.img_size[i] = 0; + return i; + } + } + + DRM_ERROR("No more free VCE handles!\n"); + return -EINVAL; +} + +/** + * radeon_vce_cs_parse - parse and validate the command stream + * + * @p: parser context + * + */ +int radeon_vce_cs_parse(struct radeon_cs_parser *p) +{ + int session_idx = -1; + bool destroyed = false; + uint32_t tmp, handle = 0; + uint32_t *size = &tmp; + int i, r; + + while (p->idx < p->chunks[p->chunk_ib_idx].length_dw) { + uint32_t len = radeon_get_ib_value(p, p->idx); + uint32_t cmd = radeon_get_ib_value(p, p->idx + 1); + + if ((len < 8) || (len & 3)) { + DRM_ERROR("invalid VCE command length (%d)!\n", len); + return -EINVAL; + } + + if (destroyed) { + DRM_ERROR("No other command allowed after destroy!\n"); + return -EINVAL; + } + + switch (cmd) { + case 0x00000001: // session + handle = radeon_get_ib_value(p, p->idx + 2); + session_idx = radeon_vce_validate_handle(p, handle); + if (session_idx < 0) + return session_idx; + size = &p->rdev->vce.img_size[session_idx]; + break; + + case 0x00000002: // task info + break; + + case 0x01000001: // create + *size = radeon_get_ib_value(p, p->idx + 8) * + radeon_get_ib_value(p, p->idx + 10) * + 8 * 3 / 2; + break; + + case 0x04000001: // config extension + case 0x04000002: // pic control + case 0x04000005: // rate control + case 0x04000007: // motion estimation + case 0x04000008: // rdo + break; + + case 0x03000001: // encode + r = radeon_vce_cs_reloc(p, p->idx + 10, p->idx + 9, + *size); + if (r) + return r; + + r = radeon_vce_cs_reloc(p, p->idx + 12, p->idx + 11, + *size / 3); + if (r) + return r; + break; + + case 0x02000001: // destroy + destroyed = true; + break; + + case 0x05000001: // context buffer + r = radeon_vce_cs_reloc(p, p->idx + 3, p->idx + 2, + *size * 2); + if (r) + return r; + break; + + case 0x05000004: // video bitstream buffer + tmp = radeon_get_ib_value(p, p->idx + 4); + r = radeon_vce_cs_reloc(p, p->idx + 3, p->idx + 2, + tmp); + if (r) + return r; + break; + + case 0x05000005: // feedback buffer + r = radeon_vce_cs_reloc(p, p->idx + 3, p->idx + 2, + 4096); + if (r) + return r; + break; + + default: + DRM_ERROR("invalid VCE command (0x%x)!\n", cmd); + return -EINVAL; + } + + if (session_idx == -1) { + DRM_ERROR("no session command at start of IB\n"); + return -EINVAL; + } + + p->idx += len / 4; + } + + if (destroyed) { + /* IB contains a destroy msg, free the handle */ + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) + atomic_cmpxchg(&p->rdev->vce.handles[i], handle, 0); + } + + return 0; +} + +/** + * radeon_vce_semaphore_emit - emit a semaphore command + * + * @rdev: radeon_device pointer + * @ring: engine to use + * @semaphore: address of semaphore + * @emit_wait: true=emit wait, false=emit signal + * + */ +bool radeon_vce_semaphore_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait) +{ + uint64_t addr = semaphore->gpu_addr; + + radeon_ring_write(ring, VCE_CMD_SEMAPHORE); + radeon_ring_write(ring, (addr >> 3) & 0x000FFFFF); + radeon_ring_write(ring, (addr >> 23) & 0x000FFFFF); + radeon_ring_write(ring, 0x01003000 | (emit_wait ? 1 : 0)); + if (!emit_wait) + radeon_ring_write(ring, VCE_CMD_END); + + return true; +} + +/** + * radeon_vce_ib_execute - execute indirect buffer + * + * @rdev: radeon_device pointer + * @ib: the IB to execute + * + */ +void radeon_vce_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) +{ + struct radeon_ring *ring = &rdev->ring[ib->ring]; + radeon_ring_write(ring, VCE_CMD_IB); + radeon_ring_write(ring, ib->gpu_addr); + radeon_ring_write(ring, upper_32_bits(ib->gpu_addr)); + radeon_ring_write(ring, ib->length_dw); +} + +/** + * radeon_vce_fence_emit - add a fence command to the ring + * + * @rdev: radeon_device pointer + * @fence: the fence + * + */ +void radeon_vce_fence_emit(struct radeon_device *rdev, + struct radeon_fence *fence) +{ + struct radeon_ring *ring = &rdev->ring[fence->ring]; + uint64_t addr = rdev->fence_drv[fence->ring].gpu_addr; + + radeon_ring_write(ring, VCE_CMD_FENCE); + radeon_ring_write(ring, addr); + radeon_ring_write(ring, upper_32_bits(addr)); + radeon_ring_write(ring, fence->seq); + radeon_ring_write(ring, VCE_CMD_TRAP); + radeon_ring_write(ring, VCE_CMD_END); +} + +/** + * radeon_vce_ring_test - test if VCE ring is working + * + * @rdev: radeon_device pointer + * @ring: the engine to test on + * + */ +int radeon_vce_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) +{ + uint32_t rptr = vce_v1_0_get_rptr(rdev, ring); + unsigned i; + int r; + + r = radeon_ring_lock(rdev, ring, 16); + if (r) { + DRM_ERROR("radeon: vce failed to lock ring %d (%d).\n", + ring->idx, r); + return r; + } + radeon_ring_write(ring, VCE_CMD_END); + radeon_ring_unlock_commit(rdev, ring); + + for (i = 0; i < rdev->usec_timeout; i++) { + if (vce_v1_0_get_rptr(rdev, ring) != rptr) + break; + DRM_UDELAY(1); + } + + if (i < rdev->usec_timeout) { + DRM_INFO("ring test on %d succeeded in %d usecs\n", + ring->idx, i); + } else { + DRM_ERROR("radeon: ring %d test failed\n", + ring->idx); + r = -ETIMEDOUT; + } + + return r; +} + +/** + * radeon_vce_ib_test - test if VCE IBs are working + * + * @rdev: radeon_device pointer + * @ring: the engine to test on + * + */ +int radeon_vce_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) +{ + struct radeon_fence *fence = NULL; + int r; + + r = radeon_vce_get_create_msg(rdev, ring->idx, 1, NULL); + if (r) { + DRM_ERROR("radeon: failed to get create msg (%d).\n", r); + goto error; + } + + r = radeon_vce_get_destroy_msg(rdev, ring->idx, 1, &fence); + if (r) { + DRM_ERROR("radeon: failed to get destroy ib (%d).\n", r); + goto error; + } + + r = radeon_fence_wait(fence, false); + if (r) { + DRM_ERROR("radeon: fence wait failed (%d).\n", r); + } else { + DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + } +error: + radeon_fence_unref(&fence); + return r; +} diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c new file mode 100644 index 00000000000..725d3669014 --- /dev/null +++ b/drivers/gpu/drm/radeon/radeon_vm.c @@ -0,0 +1,1089 @@ +/* + * Copyright 2008 Advanced Micro Devices, Inc. + * Copyright 2008 Red Hat Inc. + * Copyright 2009 Jerome Glisse. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Dave Airlie + * Alex Deucher + * Jerome Glisse + */ +#include <drm/drmP.h> +#include <drm/radeon_drm.h> +#include "radeon.h" +#include "radeon_trace.h" + +/* + * GPUVM + * GPUVM is similar to the legacy gart on older asics, however + * rather than there being a single global gart table + * for the entire GPU, there are multiple VM page tables active + * at any given time. The VM page tables can contain a mix + * vram pages and system memory pages and system memory pages + * can be mapped as snooped (cached system pages) or unsnooped + * (uncached system pages). + * Each VM has an ID associated with it and there is a page table + * associated with each VMID. When execting a command buffer, + * the kernel tells the the ring what VMID to use for that command + * buffer. VMIDs are allocated dynamically as commands are submitted. + * The userspace drivers maintain their own address space and the kernel + * sets up their pages tables accordingly when they submit their + * command buffers and a VMID is assigned. + * Cayman/Trinity support up to 8 active VMs at any given time; + * SI supports 16. + */ + +/** + * radeon_vm_num_pde - return the number of page directory entries + * + * @rdev: radeon_device pointer + * + * Calculate the number of page directory entries (cayman+). + */ +static unsigned radeon_vm_num_pdes(struct radeon_device *rdev) +{ + return rdev->vm_manager.max_pfn >> radeon_vm_block_size; +} + +/** + * radeon_vm_directory_size - returns the size of the page directory in bytes + * + * @rdev: radeon_device pointer + * + * Calculate the size of the page directory in bytes (cayman+). + */ +static unsigned radeon_vm_directory_size(struct radeon_device *rdev) +{ + return RADEON_GPU_PAGE_ALIGN(radeon_vm_num_pdes(rdev) * 8); +} + +/** + * radeon_vm_manager_init - init the vm manager + * + * @rdev: radeon_device pointer + * + * Init the vm manager (cayman+). + * Returns 0 for success, error for failure. + */ +int radeon_vm_manager_init(struct radeon_device *rdev) +{ + int r; + + if (!rdev->vm_manager.enabled) { + r = radeon_asic_vm_init(rdev); + if (r) + return r; + + rdev->vm_manager.enabled = true; + } + return 0; +} + +/** + * radeon_vm_manager_fini - tear down the vm manager + * + * @rdev: radeon_device pointer + * + * Tear down the VM manager (cayman+). + */ +void radeon_vm_manager_fini(struct radeon_device *rdev) +{ + int i; + + if (!rdev->vm_manager.enabled) + return; + + for (i = 0; i < RADEON_NUM_VM; ++i) + radeon_fence_unref(&rdev->vm_manager.active[i]); + radeon_asic_vm_fini(rdev); + rdev->vm_manager.enabled = false; +} + +/** + * radeon_vm_get_bos - add the vm BOs to a validation list + * + * @vm: vm providing the BOs + * @head: head of validation list + * + * Add the page directory to the list of BOs to + * validate for command submission (cayman+). + */ +struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev, + struct radeon_vm *vm, + struct list_head *head) +{ + struct radeon_cs_reloc *list; + unsigned i, idx; + + list = kmalloc_array(vm->max_pde_used + 2, + sizeof(struct radeon_cs_reloc), GFP_KERNEL); + if (!list) + return NULL; + + /* add the vm page table to the list */ + list[0].gobj = NULL; + list[0].robj = vm->page_directory; + list[0].prefered_domains = RADEON_GEM_DOMAIN_VRAM; + list[0].allowed_domains = RADEON_GEM_DOMAIN_VRAM; + list[0].tv.bo = &vm->page_directory->tbo; + list[0].tiling_flags = 0; + list[0].handle = 0; + list_add(&list[0].tv.head, head); + + for (i = 0, idx = 1; i <= vm->max_pde_used; i++) { + if (!vm->page_tables[i].bo) + continue; + + list[idx].gobj = NULL; + list[idx].robj = vm->page_tables[i].bo; + list[idx].prefered_domains = RADEON_GEM_DOMAIN_VRAM; + list[idx].allowed_domains = RADEON_GEM_DOMAIN_VRAM; + list[idx].tv.bo = &list[idx].robj->tbo; + list[idx].tiling_flags = 0; + list[idx].handle = 0; + list_add(&list[idx++].tv.head, head); + } + + return list; +} + +/** + * radeon_vm_grab_id - allocate the next free VMID + * + * @rdev: radeon_device pointer + * @vm: vm to allocate id for + * @ring: ring we want to submit job to + * + * Allocate an id for the vm (cayman+). + * Returns the fence we need to sync to (if any). + * + * Global and local mutex must be locked! + */ +struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev, + struct radeon_vm *vm, int ring) +{ + struct radeon_fence *best[RADEON_NUM_RINGS] = {}; + unsigned choices[2] = {}; + unsigned i; + + /* check if the id is still valid */ + if (vm->last_id_use && vm->last_id_use == rdev->vm_manager.active[vm->id]) + return NULL; + + /* we definately need to flush */ + radeon_fence_unref(&vm->last_flush); + + /* skip over VMID 0, since it is the system VM */ + for (i = 1; i < rdev->vm_manager.nvm; ++i) { + struct radeon_fence *fence = rdev->vm_manager.active[i]; + + if (fence == NULL) { + /* found a free one */ + vm->id = i; + trace_radeon_vm_grab_id(vm->id, ring); + return NULL; + } + + if (radeon_fence_is_earlier(fence, best[fence->ring])) { + best[fence->ring] = fence; + choices[fence->ring == ring ? 0 : 1] = i; + } + } + + for (i = 0; i < 2; ++i) { + if (choices[i]) { + vm->id = choices[i]; + trace_radeon_vm_grab_id(vm->id, ring); + return rdev->vm_manager.active[choices[i]]; + } + } + + /* should never happen */ + BUG(); + return NULL; +} + +/** + * radeon_vm_flush - hardware flush the vm + * + * @rdev: radeon_device pointer + * @vm: vm we want to flush + * @ring: ring to use for flush + * + * Flush the vm (cayman+). + * + * Global and local mutex must be locked! + */ +void radeon_vm_flush(struct radeon_device *rdev, + struct radeon_vm *vm, + int ring) +{ + uint64_t pd_addr = radeon_bo_gpu_offset(vm->page_directory); + + /* if we can't remember our last VM flush then flush now! */ + /* XXX figure out why we have to flush all the time */ + if (!vm->last_flush || true || pd_addr != vm->pd_gpu_addr) { + vm->pd_gpu_addr = pd_addr; + radeon_ring_vm_flush(rdev, ring, vm); + } +} + +/** + * radeon_vm_fence - remember fence for vm + * + * @rdev: radeon_device pointer + * @vm: vm we want to fence + * @fence: fence to remember + * + * Fence the vm (cayman+). + * Set the fence used to protect page table and id. + * + * Global and local mutex must be locked! + */ +void radeon_vm_fence(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_fence *fence) +{ + radeon_fence_unref(&vm->fence); + vm->fence = radeon_fence_ref(fence); + + radeon_fence_unref(&rdev->vm_manager.active[vm->id]); + rdev->vm_manager.active[vm->id] = radeon_fence_ref(fence); + + radeon_fence_unref(&vm->last_id_use); + vm->last_id_use = radeon_fence_ref(fence); + + /* we just flushed the VM, remember that */ + if (!vm->last_flush) + vm->last_flush = radeon_fence_ref(fence); +} + +/** + * radeon_vm_bo_find - find the bo_va for a specific vm & bo + * + * @vm: requested vm + * @bo: requested buffer object + * + * Find @bo inside the requested vm (cayman+). + * Search inside the @bos vm list for the requested vm + * Returns the found bo_va or NULL if none is found + * + * Object has to be reserved! + */ +struct radeon_bo_va *radeon_vm_bo_find(struct radeon_vm *vm, + struct radeon_bo *bo) +{ + struct radeon_bo_va *bo_va; + + list_for_each_entry(bo_va, &bo->va, bo_list) { + if (bo_va->vm == vm) { + return bo_va; + } + } + return NULL; +} + +/** + * radeon_vm_bo_add - add a bo to a specific vm + * + * @rdev: radeon_device pointer + * @vm: requested vm + * @bo: radeon buffer object + * + * Add @bo into the requested vm (cayman+). + * Add @bo to the list of bos associated with the vm + * Returns newly added bo_va or NULL for failure + * + * Object has to be reserved! + */ +struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_bo *bo) +{ + struct radeon_bo_va *bo_va; + + bo_va = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL); + if (bo_va == NULL) { + return NULL; + } + bo_va->vm = vm; + bo_va->bo = bo; + bo_va->soffset = 0; + bo_va->eoffset = 0; + bo_va->flags = 0; + bo_va->valid = false; + bo_va->ref_count = 1; + INIT_LIST_HEAD(&bo_va->bo_list); + INIT_LIST_HEAD(&bo_va->vm_list); + INIT_LIST_HEAD(&bo_va->vm_status); + + mutex_lock(&vm->mutex); + list_add(&bo_va->vm_list, &vm->va); + list_add_tail(&bo_va->bo_list, &bo->va); + mutex_unlock(&vm->mutex); + + return bo_va; +} + +/** + * radeon_vm_clear_bo - initially clear the page dir/table + * + * @rdev: radeon_device pointer + * @bo: bo to clear + */ +static int radeon_vm_clear_bo(struct radeon_device *rdev, + struct radeon_bo *bo) +{ + struct ttm_validate_buffer tv; + struct ww_acquire_ctx ticket; + struct list_head head; + struct radeon_ib ib; + unsigned entries; + uint64_t addr; + int r; + + memset(&tv, 0, sizeof(tv)); + tv.bo = &bo->tbo; + + INIT_LIST_HEAD(&head); + list_add(&tv.head, &head); + + r = ttm_eu_reserve_buffers(&ticket, &head); + if (r) + return r; + + r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); + if (r) + goto error; + + addr = radeon_bo_gpu_offset(bo); + entries = radeon_bo_size(bo) / 8; + + r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, + NULL, entries * 2 + 64); + if (r) + goto error; + + ib.length_dw = 0; + + radeon_asic_vm_set_page(rdev, &ib, addr, 0, entries, 0, 0); + + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) + goto error; + + ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence); + radeon_ib_free(rdev, &ib); + + return 0; + +error: + ttm_eu_backoff_reservation(&ticket, &head); + return r; +} + +/** + * radeon_vm_bo_set_addr - set bos virtual address inside a vm + * + * @rdev: radeon_device pointer + * @bo_va: bo_va to store the address + * @soffset: requested offset of the buffer in the VM address space + * @flags: attributes of pages (read/write/valid/etc.) + * + * Set offset of @bo_va (cayman+). + * Validate and set the offset requested within the vm address space. + * Returns 0 for success, error for failure. + * + * Object has to be reserved! + */ +int radeon_vm_bo_set_addr(struct radeon_device *rdev, + struct radeon_bo_va *bo_va, + uint64_t soffset, + uint32_t flags) +{ + uint64_t size = radeon_bo_size(bo_va->bo); + uint64_t eoffset, last_offset = 0; + struct radeon_vm *vm = bo_va->vm; + struct radeon_bo_va *tmp; + struct list_head *head; + unsigned last_pfn, pt_idx; + int r; + + if (soffset) { + /* make sure object fit at this offset */ + eoffset = soffset + size; + if (soffset >= eoffset) { + return -EINVAL; + } + + last_pfn = eoffset / RADEON_GPU_PAGE_SIZE; + if (last_pfn > rdev->vm_manager.max_pfn) { + dev_err(rdev->dev, "va above limit (0x%08X > 0x%08X)\n", + last_pfn, rdev->vm_manager.max_pfn); + return -EINVAL; + } + + } else { + eoffset = last_pfn = 0; + } + + mutex_lock(&vm->mutex); + head = &vm->va; + last_offset = 0; + list_for_each_entry(tmp, &vm->va, vm_list) { + if (bo_va == tmp) { + /* skip over currently modified bo */ + continue; + } + + if (soffset >= last_offset && eoffset <= tmp->soffset) { + /* bo can be added before this one */ + break; + } + if (eoffset > tmp->soffset && soffset < tmp->eoffset) { + /* bo and tmp overlap, invalid offset */ + dev_err(rdev->dev, "bo %p va 0x%08X conflict with (bo %p 0x%08X 0x%08X)\n", + bo_va->bo, (unsigned)bo_va->soffset, tmp->bo, + (unsigned)tmp->soffset, (unsigned)tmp->eoffset); + mutex_unlock(&vm->mutex); + return -EINVAL; + } + last_offset = tmp->eoffset; + head = &tmp->vm_list; + } + + if (bo_va->soffset) { + /* add a clone of the bo_va to clear the old address */ + tmp = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL); + if (!tmp) { + mutex_unlock(&vm->mutex); + return -ENOMEM; + } + tmp->soffset = bo_va->soffset; + tmp->eoffset = bo_va->eoffset; + tmp->vm = vm; + list_add(&tmp->vm_status, &vm->freed); + } + + bo_va->soffset = soffset; + bo_va->eoffset = eoffset; + bo_va->flags = flags; + bo_va->valid = false; + list_move(&bo_va->vm_list, head); + + soffset = (soffset / RADEON_GPU_PAGE_SIZE) >> radeon_vm_block_size; + eoffset = (eoffset / RADEON_GPU_PAGE_SIZE) >> radeon_vm_block_size; + + BUG_ON(eoffset >= radeon_vm_num_pdes(rdev)); + + if (eoffset > vm->max_pde_used) + vm->max_pde_used = eoffset; + + radeon_bo_unreserve(bo_va->bo); + + /* walk over the address space and allocate the page tables */ + for (pt_idx = soffset; pt_idx <= eoffset; ++pt_idx) { + struct radeon_bo *pt; + + if (vm->page_tables[pt_idx].bo) + continue; + + /* drop mutex to allocate and clear page table */ + mutex_unlock(&vm->mutex); + + r = radeon_bo_create(rdev, RADEON_VM_PTE_COUNT * 8, + RADEON_GPU_PAGE_SIZE, true, + RADEON_GEM_DOMAIN_VRAM, NULL, &pt); + if (r) + return r; + + r = radeon_vm_clear_bo(rdev, pt); + if (r) { + radeon_bo_unref(&pt); + radeon_bo_reserve(bo_va->bo, false); + return r; + } + + /* aquire mutex again */ + mutex_lock(&vm->mutex); + if (vm->page_tables[pt_idx].bo) { + /* someone else allocated the pt in the meantime */ + mutex_unlock(&vm->mutex); + radeon_bo_unref(&pt); + mutex_lock(&vm->mutex); + continue; + } + + vm->page_tables[pt_idx].addr = 0; + vm->page_tables[pt_idx].bo = pt; + } + + mutex_unlock(&vm->mutex); + return radeon_bo_reserve(bo_va->bo, false); +} + +/** + * radeon_vm_map_gart - get the physical address of a gart page + * + * @rdev: radeon_device pointer + * @addr: the unmapped addr + * + * Look up the physical address of the page that the pte resolves + * to (cayman+). + * Returns the physical address of the page. + */ +uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr) +{ + uint64_t result; + + /* page table offset */ + result = rdev->gart.pages_addr[addr >> PAGE_SHIFT]; + + /* in case cpu page size != gpu page size*/ + result |= addr & (~PAGE_MASK); + + return result; +} + +/** + * radeon_vm_page_flags - translate page flags to what the hw uses + * + * @flags: flags comming from userspace + * + * Translate the flags the userspace ABI uses to hw flags. + */ +static uint32_t radeon_vm_page_flags(uint32_t flags) +{ + uint32_t hw_flags = 0; + hw_flags |= (flags & RADEON_VM_PAGE_VALID) ? R600_PTE_VALID : 0; + hw_flags |= (flags & RADEON_VM_PAGE_READABLE) ? R600_PTE_READABLE : 0; + hw_flags |= (flags & RADEON_VM_PAGE_WRITEABLE) ? R600_PTE_WRITEABLE : 0; + if (flags & RADEON_VM_PAGE_SYSTEM) { + hw_flags |= R600_PTE_SYSTEM; + hw_flags |= (flags & RADEON_VM_PAGE_SNOOPED) ? R600_PTE_SNOOPED : 0; + } + return hw_flags; +} + +/** + * radeon_vm_update_pdes - make sure that page directory is valid + * + * @rdev: radeon_device pointer + * @vm: requested vm + * @start: start of GPU address range + * @end: end of GPU address range + * + * Allocates new page tables if necessary + * and updates the page directory (cayman+). + * Returns 0 for success, error for failure. + * + * Global and local mutex must be locked! + */ +int radeon_vm_update_page_directory(struct radeon_device *rdev, + struct radeon_vm *vm) +{ + struct radeon_bo *pd = vm->page_directory; + uint64_t pd_addr = radeon_bo_gpu_offset(pd); + uint32_t incr = RADEON_VM_PTE_COUNT * 8; + uint64_t last_pde = ~0, last_pt = ~0; + unsigned count = 0, pt_idx, ndw; + struct radeon_ib ib; + int r; + + /* padding, etc. */ + ndw = 64; + + /* assume the worst case */ + ndw += vm->max_pde_used * 16; + + /* update too big for an IB */ + if (ndw > 0xfffff) + return -ENOMEM; + + r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, ndw * 4); + if (r) + return r; + ib.length_dw = 0; + + /* walk over the address space and update the page directory */ + for (pt_idx = 0; pt_idx <= vm->max_pde_used; ++pt_idx) { + struct radeon_bo *bo = vm->page_tables[pt_idx].bo; + uint64_t pde, pt; + + if (bo == NULL) + continue; + + pt = radeon_bo_gpu_offset(bo); + if (vm->page_tables[pt_idx].addr == pt) + continue; + vm->page_tables[pt_idx].addr = pt; + + pde = pd_addr + pt_idx * 8; + if (((last_pde + 8 * count) != pde) || + ((last_pt + incr * count) != pt)) { + + if (count) { + radeon_asic_vm_set_page(rdev, &ib, last_pde, + last_pt, count, incr, + R600_PTE_VALID); + } + + count = 1; + last_pde = pde; + last_pt = pt; + } else { + ++count; + } + } + + if (count) + radeon_asic_vm_set_page(rdev, &ib, last_pde, last_pt, count, + incr, R600_PTE_VALID); + + if (ib.length_dw != 0) { + radeon_semaphore_sync_to(ib.semaphore, pd->tbo.sync_obj); + radeon_semaphore_sync_to(ib.semaphore, vm->last_id_use); + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + radeon_ib_free(rdev, &ib); + return r; + } + radeon_fence_unref(&vm->fence); + vm->fence = radeon_fence_ref(ib.fence); + radeon_fence_unref(&vm->last_flush); + } + radeon_ib_free(rdev, &ib); + + return 0; +} + +/** + * radeon_vm_frag_ptes - add fragment information to PTEs + * + * @rdev: radeon_device pointer + * @ib: IB for the update + * @pe_start: first PTE to handle + * @pe_end: last PTE to handle + * @addr: addr those PTEs should point to + * @flags: hw mapping flags + * + * Global and local mutex must be locked! + */ +static void radeon_vm_frag_ptes(struct radeon_device *rdev, + struct radeon_ib *ib, + uint64_t pe_start, uint64_t pe_end, + uint64_t addr, uint32_t flags) +{ + /** + * The MC L1 TLB supports variable sized pages, based on a fragment + * field in the PTE. When this field is set to a non-zero value, page + * granularity is increased from 4KB to (1 << (12 + frag)). The PTE + * flags are considered valid for all PTEs within the fragment range + * and corresponding mappings are assumed to be physically contiguous. + * + * The L1 TLB can store a single PTE for the whole fragment, + * significantly increasing the space available for translation + * caching. This leads to large improvements in throughput when the + * TLB is under pressure. + * + * The L2 TLB distributes small and large fragments into two + * asymmetric partitions. The large fragment cache is significantly + * larger. Thus, we try to use large fragments wherever possible. + * Userspace can support this by aligning virtual base address and + * allocation size to the fragment size. + */ + + /* NI is optimized for 256KB fragments, SI and newer for 64KB */ + uint64_t frag_flags = rdev->family == CHIP_CAYMAN ? + R600_PTE_FRAG_256KB : R600_PTE_FRAG_64KB; + uint64_t frag_align = rdev->family == CHIP_CAYMAN ? 0x200 : 0x80; + + uint64_t frag_start = ALIGN(pe_start, frag_align); + uint64_t frag_end = pe_end & ~(frag_align - 1); + + unsigned count; + + /* system pages are non continuously */ + if ((flags & R600_PTE_SYSTEM) || !(flags & R600_PTE_VALID) || + (frag_start >= frag_end)) { + + count = (pe_end - pe_start) / 8; + radeon_asic_vm_set_page(rdev, ib, pe_start, addr, count, + RADEON_GPU_PAGE_SIZE, flags); + return; + } + + /* handle the 4K area at the beginning */ + if (pe_start != frag_start) { + count = (frag_start - pe_start) / 8; + radeon_asic_vm_set_page(rdev, ib, pe_start, addr, count, + RADEON_GPU_PAGE_SIZE, flags); + addr += RADEON_GPU_PAGE_SIZE * count; + } + + /* handle the area in the middle */ + count = (frag_end - frag_start) / 8; + radeon_asic_vm_set_page(rdev, ib, frag_start, addr, count, + RADEON_GPU_PAGE_SIZE, flags | frag_flags); + + /* handle the 4K area at the end */ + if (frag_end != pe_end) { + addr += RADEON_GPU_PAGE_SIZE * count; + count = (pe_end - frag_end) / 8; + radeon_asic_vm_set_page(rdev, ib, frag_end, addr, count, + RADEON_GPU_PAGE_SIZE, flags); + } +} + +/** + * radeon_vm_update_ptes - make sure that page tables are valid + * + * @rdev: radeon_device pointer + * @vm: requested vm + * @start: start of GPU address range + * @end: end of GPU address range + * @dst: destination address to map to + * @flags: mapping flags + * + * Update the page tables in the range @start - @end (cayman+). + * + * Global and local mutex must be locked! + */ +static void radeon_vm_update_ptes(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_ib *ib, + uint64_t start, uint64_t end, + uint64_t dst, uint32_t flags) +{ + uint64_t mask = RADEON_VM_PTE_COUNT - 1; + uint64_t last_pte = ~0, last_dst = ~0; + unsigned count = 0; + uint64_t addr; + + start = start / RADEON_GPU_PAGE_SIZE; + end = end / RADEON_GPU_PAGE_SIZE; + + /* walk over the address space and update the page tables */ + for (addr = start; addr < end; ) { + uint64_t pt_idx = addr >> radeon_vm_block_size; + struct radeon_bo *pt = vm->page_tables[pt_idx].bo; + unsigned nptes; + uint64_t pte; + + radeon_semaphore_sync_to(ib->semaphore, pt->tbo.sync_obj); + + if ((addr & ~mask) == (end & ~mask)) + nptes = end - addr; + else + nptes = RADEON_VM_PTE_COUNT - (addr & mask); + + pte = radeon_bo_gpu_offset(pt); + pte += (addr & mask) * 8; + + if ((last_pte + 8 * count) != pte) { + + if (count) { + radeon_vm_frag_ptes(rdev, ib, last_pte, + last_pte + 8 * count, + last_dst, flags); + } + + count = nptes; + last_pte = pte; + last_dst = dst; + } else { + count += nptes; + } + + addr += nptes; + dst += nptes * RADEON_GPU_PAGE_SIZE; + } + + if (count) { + radeon_vm_frag_ptes(rdev, ib, last_pte, + last_pte + 8 * count, + last_dst, flags); + } +} + +/** + * radeon_vm_bo_update - map a bo into the vm page table + * + * @rdev: radeon_device pointer + * @vm: requested vm + * @bo: radeon buffer object + * @mem: ttm mem + * + * Fill in the page table entries for @bo (cayman+). + * Returns 0 for success, -EINVAL for failure. + * + * Object have to be reserved and mutex must be locked! + */ +int radeon_vm_bo_update(struct radeon_device *rdev, + struct radeon_bo_va *bo_va, + struct ttm_mem_reg *mem) +{ + struct radeon_vm *vm = bo_va->vm; + struct radeon_ib ib; + unsigned nptes, ndw; + uint64_t addr; + int r; + + + if (!bo_va->soffset) { + dev_err(rdev->dev, "bo %p don't has a mapping in vm %p\n", + bo_va->bo, vm); + return -EINVAL; + } + + if ((bo_va->valid && mem) || (!bo_va->valid && mem == NULL)) + return 0; + + bo_va->flags &= ~RADEON_VM_PAGE_VALID; + bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM; + if (mem) { + addr = mem->start << PAGE_SHIFT; + if (mem->mem_type != TTM_PL_SYSTEM) { + bo_va->flags |= RADEON_VM_PAGE_VALID; + bo_va->valid = true; + } + if (mem->mem_type == TTM_PL_TT) { + bo_va->flags |= RADEON_VM_PAGE_SYSTEM; + } else { + addr += rdev->vm_manager.vram_base_offset; + } + } else { + addr = 0; + bo_va->valid = false; + } + + trace_radeon_vm_bo_update(bo_va); + + nptes = (bo_va->eoffset - bo_va->soffset) / RADEON_GPU_PAGE_SIZE; + + /* padding, etc. */ + ndw = 64; + + if (radeon_vm_block_size > 11) + /* reserve space for one header for every 2k dwords */ + ndw += (nptes >> 11) * 4; + else + /* reserve space for one header for + every (1 << BLOCK_SIZE) entries */ + ndw += (nptes >> radeon_vm_block_size) * 4; + + /* reserve space for pte addresses */ + ndw += nptes * 2; + + /* update too big for an IB */ + if (ndw > 0xfffff) + return -ENOMEM; + + r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, ndw * 4); + if (r) + return r; + ib.length_dw = 0; + + radeon_vm_update_ptes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset, + addr, radeon_vm_page_flags(bo_va->flags)); + + radeon_semaphore_sync_to(ib.semaphore, vm->fence); + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + radeon_ib_free(rdev, &ib); + return r; + } + radeon_fence_unref(&vm->fence); + vm->fence = radeon_fence_ref(ib.fence); + radeon_ib_free(rdev, &ib); + radeon_fence_unref(&vm->last_flush); + + return 0; +} + +/** + * radeon_vm_clear_freed - clear freed BOs in the PT + * + * @rdev: radeon_device pointer + * @vm: requested vm + * + * Make sure all freed BOs are cleared in the PT. + * Returns 0 for success. + * + * PTs have to be reserved and mutex must be locked! + */ +int radeon_vm_clear_freed(struct radeon_device *rdev, + struct radeon_vm *vm) +{ + struct radeon_bo_va *bo_va, *tmp; + int r; + + list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status) { + list_del(&bo_va->vm_status); + r = radeon_vm_bo_update(rdev, bo_va, NULL); + kfree(bo_va); + if (r) + return r; + } + return 0; + +} + +/** + * radeon_vm_bo_rmv - remove a bo to a specific vm + * + * @rdev: radeon_device pointer + * @bo_va: requested bo_va + * + * Remove @bo_va->bo from the requested vm (cayman+). + * + * Object have to be reserved! + */ +void radeon_vm_bo_rmv(struct radeon_device *rdev, + struct radeon_bo_va *bo_va) +{ + struct radeon_vm *vm = bo_va->vm; + + list_del(&bo_va->bo_list); + + mutex_lock(&vm->mutex); + list_del(&bo_va->vm_list); + + if (bo_va->soffset) { + bo_va->bo = NULL; + list_add(&bo_va->vm_status, &vm->freed); + } else { + kfree(bo_va); + } + + mutex_unlock(&vm->mutex); +} + +/** + * radeon_vm_bo_invalidate - mark the bo as invalid + * + * @rdev: radeon_device pointer + * @vm: requested vm + * @bo: radeon buffer object + * + * Mark @bo as invalid (cayman+). + */ +void radeon_vm_bo_invalidate(struct radeon_device *rdev, + struct radeon_bo *bo) +{ + struct radeon_bo_va *bo_va; + + list_for_each_entry(bo_va, &bo->va, bo_list) { + bo_va->valid = false; + } +} + +/** + * radeon_vm_init - initialize a vm instance + * + * @rdev: radeon_device pointer + * @vm: requested vm + * + * Init @vm fields (cayman+). + */ +int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm) +{ + const unsigned align = min(RADEON_VM_PTB_ALIGN_SIZE, + RADEON_VM_PTE_COUNT * 8); + unsigned pd_size, pd_entries, pts_size; + int r; + + vm->id = 0; + vm->ib_bo_va = NULL; + vm->fence = NULL; + vm->last_flush = NULL; + vm->last_id_use = NULL; + mutex_init(&vm->mutex); + INIT_LIST_HEAD(&vm->va); + INIT_LIST_HEAD(&vm->freed); + + pd_size = radeon_vm_directory_size(rdev); + pd_entries = radeon_vm_num_pdes(rdev); + + /* allocate page table array */ + pts_size = pd_entries * sizeof(struct radeon_vm_pt); + vm->page_tables = kzalloc(pts_size, GFP_KERNEL); + if (vm->page_tables == NULL) { + DRM_ERROR("Cannot allocate memory for page table array\n"); + return -ENOMEM; + } + + r = radeon_bo_create(rdev, pd_size, align, true, + RADEON_GEM_DOMAIN_VRAM, NULL, + &vm->page_directory); + if (r) + return r; + + r = radeon_vm_clear_bo(rdev, vm->page_directory); + if (r) { + radeon_bo_unref(&vm->page_directory); + vm->page_directory = NULL; + return r; + } + + return 0; +} + +/** + * radeon_vm_fini - tear down a vm instance + * + * @rdev: radeon_device pointer + * @vm: requested vm + * + * Tear down @vm (cayman+). + * Unbind the VM and remove all bos from the vm bo list + */ +void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm) +{ + struct radeon_bo_va *bo_va, *tmp; + int i, r; + + if (!list_empty(&vm->va)) { + dev_err(rdev->dev, "still active bo inside vm\n"); + } + list_for_each_entry_safe(bo_va, tmp, &vm->va, vm_list) { + list_del_init(&bo_va->vm_list); + r = radeon_bo_reserve(bo_va->bo, false); + if (!r) { + list_del_init(&bo_va->bo_list); + radeon_bo_unreserve(bo_va->bo); + kfree(bo_va); + } + } + list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status) + kfree(bo_va); + + for (i = 0; i < radeon_vm_num_pdes(rdev); i++) + radeon_bo_unref(&vm->page_tables[i].bo); + kfree(vm->page_tables); + + radeon_bo_unref(&vm->page_directory); + + radeon_fence_unref(&vm->fence); + radeon_fence_unref(&vm->last_flush); + radeon_fence_unref(&vm->last_id_use); + + mutex_destroy(&vm->mutex); +} diff --git a/drivers/gpu/drm/radeon/reg_srcs/cayman b/drivers/gpu/drm/radeon/reg_srcs/cayman index a072fa8c46b..d46b58d078a 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/cayman +++ b/drivers/gpu/drm/radeon/reg_srcs/cayman @@ -21,7 +21,7 @@ cayman 0x9400 0x000089AC VGT_COMPUTE_THREAD_GOURP_SIZE 0x000089B0 VGT_HS_OFFCHIP_PARAM 0x00008A14 PA_CL_ENHANCE -0x00008A60 PA_SC_LINE_STIPPLE_VALUE +0x00008A60 PA_SU_LINE_STIPPLE_VALUE 0x00008B10 PA_SC_LINE_STIPPLE_STATE 0x00008BF0 PA_SC_ENHANCE 0x00008D8C SQ_DYN_GPR_CNTL_PS_FLUSH_REQ @@ -532,7 +532,7 @@ cayman 0x9400 0x00028B84 PA_SU_POLY_OFFSET_FRONT_OFFSET 0x00028B88 PA_SU_POLY_OFFSET_BACK_SCALE 0x00028B8C PA_SU_POLY_OFFSET_BACK_OFFSET -0x00028B74 VGT_GS_INSTANCE_CNT +0x00028B90 VGT_GS_INSTANCE_CNT 0x00028BD4 PA_SC_CENTROID_PRIORITY_0 0x00028BD8 PA_SC_CENTROID_PRIORITY_1 0x00028BDC PA_SC_LINE_CNTL diff --git a/drivers/gpu/drm/radeon/reg_srcs/evergreen b/drivers/gpu/drm/radeon/reg_srcs/evergreen index b912a37689b..57745c8761c 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/evergreen +++ b/drivers/gpu/drm/radeon/reg_srcs/evergreen @@ -22,7 +22,7 @@ evergreen 0x9400 0x000089A4 VGT_COMPUTE_START_Z 0x000089AC VGT_COMPUTE_THREAD_GOURP_SIZE 0x00008A14 PA_CL_ENHANCE -0x00008A60 PA_SC_LINE_STIPPLE_VALUE +0x00008A60 PA_SU_LINE_STIPPLE_VALUE 0x00008B10 PA_SC_LINE_STIPPLE_STATE 0x00008BF0 PA_SC_ENHANCE 0x00008D8C SQ_DYN_GPR_CNTL_PS_FLUSH_REQ @@ -545,7 +545,7 @@ evergreen 0x9400 0x00028B84 PA_SU_POLY_OFFSET_FRONT_OFFSET 0x00028B88 PA_SU_POLY_OFFSET_BACK_SCALE 0x00028B8C PA_SU_POLY_OFFSET_BACK_OFFSET -0x00028B74 VGT_GS_INSTANCE_CNT +0x00028B90 VGT_GS_INSTANCE_CNT 0x00028C00 PA_SC_LINE_CNTL 0x00028C08 PA_SU_VTX_CNTL 0x00028C0C PA_CL_GB_VERT_CLIP_ADJ diff --git a/drivers/gpu/drm/radeon/reg_srcs/r600 b/drivers/gpu/drm/radeon/reg_srcs/r600 index 20bfbda7b3f..ec0c6829c1d 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/r600 +++ b/drivers/gpu/drm/radeon/reg_srcs/r600 @@ -18,6 +18,7 @@ r600 0x9400 0x00028A3C VGT_GROUP_VECT_1_FMT_CNTL 0x00028A40 VGT_GS_MODE 0x00028A6C VGT_GS_OUT_PRIM_TYPE +0x00028B38 VGT_GS_MAX_VERT_OUT 0x000088C8 VGT_GS_PER_ES 0x000088E8 VGT_GS_PER_VS 0x000088D4 VGT_GS_VERTEX_REUSE diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c index 233a9b9fa1f..a0f96decece 100644 --- a/drivers/gpu/drm/radeon/rs400.c +++ b/drivers/gpu/drm/radeon/rs400.c @@ -174,10 +174,13 @@ int rs400_gart_enable(struct radeon_device *rdev) /* FIXME: according to doc we should set HIDE_MMCFG_BAR=0, * AGPMODE30=0 & AGP30ENHANCED=0 in NB_CNTL */ if ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740)) { - WREG32_MC(RS480_MC_MISC_CNTL, - (RS480_GART_INDEX_REG_EN | RS690_BLOCK_GFX_D3_EN)); + tmp = RREG32_MC(RS480_MC_MISC_CNTL); + tmp |= RS480_GART_INDEX_REG_EN | RS690_BLOCK_GFX_D3_EN; + WREG32_MC(RS480_MC_MISC_CNTL, tmp); } else { - WREG32_MC(RS480_MC_MISC_CNTL, RS480_GART_INDEX_REG_EN); + tmp = RREG32_MC(RS480_MC_MISC_CNTL); + tmp |= RS480_GART_INDEX_REG_EN; + WREG32_MC(RS480_MC_MISC_CNTL, tmp); } /* Enable gart */ WREG32_MC(RS480_AGP_ADDRESS_SPACE_SIZE, (RS480_GART_EN | size_reg)); @@ -209,21 +212,16 @@ void rs400_gart_fini(struct radeon_device *rdev) #define RS400_PTE_WRITEABLE (1 << 2) #define RS400_PTE_READABLE (1 << 3) -int rs400_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr) +void rs400_gart_set_page(struct radeon_device *rdev, unsigned i, uint64_t addr) { uint32_t entry; u32 *gtt = rdev->gart.ptr; - if (i < 0 || i > rdev->gart.num_gpu_pages) { - return -EINVAL; - } - entry = (lower_32_bits(addr) & PAGE_MASK) | ((upper_32_bits(addr) & 0xff) << 4) | RS400_PTE_WRITEABLE | RS400_PTE_READABLE; entry = cpu_to_le32(entry); gtt[i] = entry; - return 0; } int rs400_mc_wait_for_idle(struct radeon_device *rdev) @@ -271,19 +269,26 @@ static void rs400_mc_init(struct radeon_device *rdev) uint32_t rs400_mc_rreg(struct radeon_device *rdev, uint32_t reg) { + unsigned long flags; uint32_t r; + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(RS480_NB_MC_INDEX, reg & 0xff); r = RREG32(RS480_NB_MC_DATA); WREG32(RS480_NB_MC_INDEX, 0xff); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); return r; } void rs400_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(RS480_NB_MC_INDEX, ((reg) & 0xff) | RS480_NB_MC_IND_WR_EN); WREG32(RS480_NB_MC_DATA, (v)); WREG32(RS480_NB_MC_INDEX, 0xff); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); } #if defined(CONFIG_DEBUG_FS) @@ -474,6 +479,7 @@ int rs400_resume(struct radeon_device *rdev) int rs400_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); r100_irq_disable(rdev); @@ -483,6 +489,7 @@ int rs400_suspend(struct radeon_device *rdev) void rs400_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); r100_cp_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); @@ -550,6 +557,9 @@ int rs400_init(struct radeon_device *rdev) return r; r300_set_reg_safe(rdev); + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->accel_working = true; r = rs400_startup(rdev); if (r) { diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c index 670b555d2ca..d1a35cb1c91 100644 --- a/drivers/gpu/drm/radeon/rs600.c +++ b/drivers/gpu/drm/radeon/rs600.c @@ -109,19 +109,7 @@ void avivo_wait_for_vblank(struct radeon_device *rdev, int crtc) } } -void rs600_pre_page_flip(struct radeon_device *rdev, int crtc) -{ - /* enable the pflip int */ - radeon_irq_kms_pflip_irq_get(rdev, crtc); -} - -void rs600_post_page_flip(struct radeon_device *rdev, int crtc) -{ - /* disable the pflip int */ - radeon_irq_kms_pflip_irq_put(rdev, crtc); -} - -u32 rs600_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) +void rs600_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) { struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id]; u32 tmp = RREG32(AVIVO_D1GRPH_UPDATE + radeon_crtc->crtc_offset); @@ -148,9 +136,79 @@ u32 rs600_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) /* Unlock the lock, so double-buffering can take place inside vblank */ tmp &= ~AVIVO_D1GRPH_UPDATE_LOCK; WREG32(AVIVO_D1GRPH_UPDATE + radeon_crtc->crtc_offset, tmp); +} + +bool rs600_page_flip_pending(struct radeon_device *rdev, int crtc_id) +{ + struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id]; /* Return current update_pending status: */ - return RREG32(AVIVO_D1GRPH_UPDATE + radeon_crtc->crtc_offset) & AVIVO_D1GRPH_SURFACE_UPDATE_PENDING; + return !!(RREG32(AVIVO_D1GRPH_UPDATE + radeon_crtc->crtc_offset) & + AVIVO_D1GRPH_SURFACE_UPDATE_PENDING); +} + +void avivo_program_fmt(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct drm_connector *connector = radeon_get_connector_for_encoder(encoder); + int bpc = 0; + u32 tmp = 0; + enum radeon_connector_dither dither = RADEON_FMT_DITHER_DISABLE; + + if (connector) { + struct radeon_connector *radeon_connector = to_radeon_connector(connector); + bpc = radeon_get_monitor_bpc(connector); + dither = radeon_connector->dither; + } + + /* LVDS FMT is set up by atom */ + if (radeon_encoder->devices & ATOM_DEVICE_LCD_SUPPORT) + return; + + if (bpc == 0) + return; + + switch (bpc) { + case 6: + if (dither == RADEON_FMT_DITHER_ENABLE) + /* XXX sort out optimal dither settings */ + tmp |= AVIVO_TMDS_BIT_DEPTH_CONTROL_SPATIAL_DITHER_EN; + else + tmp |= AVIVO_TMDS_BIT_DEPTH_CONTROL_TRUNCATE_EN; + break; + case 8: + if (dither == RADEON_FMT_DITHER_ENABLE) + /* XXX sort out optimal dither settings */ + tmp |= (AVIVO_TMDS_BIT_DEPTH_CONTROL_SPATIAL_DITHER_EN | + AVIVO_TMDS_BIT_DEPTH_CONTROL_SPATIAL_DITHER_DEPTH); + else + tmp |= (AVIVO_TMDS_BIT_DEPTH_CONTROL_TRUNCATE_EN | + AVIVO_TMDS_BIT_DEPTH_CONTROL_TRUNCATE_DEPTH); + break; + case 10: + default: + /* not needed */ + break; + } + + switch (radeon_encoder->encoder_id) { + case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1: + WREG32(AVIVO_TMDSA_BIT_DEPTH_CONTROL, tmp); + break; + case ENCODER_OBJECT_ID_INTERNAL_LVTM1: + WREG32(AVIVO_LVTMA_BIT_DEPTH_CONTROL, tmp); + break; + case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1: + WREG32(AVIVO_DVOA_BIT_DEPTH_CONTROL, tmp); + break; + case ENCODER_OBJECT_ID_INTERNAL_DDI: + WREG32(AVIVO_DDIA_BIT_DEPTH_CONTROL, tmp); + break; + default: + break; + } } void rs600_pm_misc(struct radeon_device *rdev) @@ -568,24 +626,16 @@ static void rs600_gart_fini(struct radeon_device *rdev) radeon_gart_table_vram_free(rdev); } -#define R600_PTE_VALID (1 << 0) -#define R600_PTE_SYSTEM (1 << 1) -#define R600_PTE_SNOOPED (1 << 2) -#define R600_PTE_READABLE (1 << 5) -#define R600_PTE_WRITEABLE (1 << 6) - -int rs600_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr) +void rs600_gart_set_page(struct radeon_device *rdev, unsigned i, uint64_t addr) { void __iomem *ptr = (void *)rdev->gart.ptr; - if (i < 0 || i > rdev->gart.num_gpu_pages) { - return -EINVAL; - } addr = addr & 0xFFFFFFFFFFFFF000ULL; - addr |= R600_PTE_VALID | R600_PTE_SYSTEM | R600_PTE_SNOOPED; - addr |= R600_PTE_READABLE | R600_PTE_WRITEABLE; + if (addr == rdev->dummy_page.addr) + addr |= R600_PTE_SYSTEM | R600_PTE_SNOOPED; + else + addr |= R600_PTE_GART; writeq(addr, ptr + (i * 8)); - return 0; } int rs600_irq_set(struct radeon_device *rdev) @@ -723,7 +773,7 @@ int rs600_irq_process(struct radeon_device *rdev) wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[0])) - radeon_crtc_handle_flip(rdev, 0); + radeon_crtc_handle_vblank(rdev, 0); } if (G_007EDC_LB_D2_VBLANK_INTERRUPT(rdev->irq.stat_regs.r500.disp_int)) { if (rdev->irq.crtc_vblank_int[1]) { @@ -732,7 +782,7 @@ int rs600_irq_process(struct radeon_device *rdev) wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[1])) - radeon_crtc_handle_flip(rdev, 1); + radeon_crtc_handle_vblank(rdev, 1); } if (G_007EDC_DC_HOT_PLUG_DETECT1_INTERRUPT(rdev->irq.stat_regs.r500.disp_int)) { queue_hotplug = true; @@ -847,16 +897,26 @@ void rs600_bandwidth_update(struct radeon_device *rdev) uint32_t rs600_mc_rreg(struct radeon_device *rdev, uint32_t reg) { + unsigned long flags; + u32 r; + + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(R_000070_MC_IND_INDEX, S_000070_MC_IND_ADDR(reg) | S_000070_MC_IND_CITF_ARB0(1)); - return RREG32(R_000074_MC_IND_DATA); + r = RREG32(R_000074_MC_IND_DATA); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); + return r; } void rs600_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(R_000070_MC_IND_INDEX, S_000070_MC_IND_ADDR(reg) | S_000070_MC_IND_CITF_ARB0(1) | S_000070_MC_IND_WR_EN(1)); WREG32(R_000074_MC_IND_DATA, v); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); } static void rs600_debugfs(struct radeon_device *rdev) @@ -984,6 +1044,7 @@ int rs600_resume(struct radeon_device *rdev) int rs600_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r600_audio_fini(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); @@ -994,6 +1055,7 @@ int rs600_suspend(struct radeon_device *rdev) void rs600_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); r600_audio_fini(rdev); r100_cp_fini(rdev); radeon_wb_fini(rdev); @@ -1062,6 +1124,9 @@ int rs600_init(struct radeon_device *rdev) return r; rs600_set_safe_registers(rdev); + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->accel_working = true; r = rs600_startup(rdev); if (r) { diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c index d8ddfb34545..3462b64369b 100644 --- a/drivers/gpu/drm/radeon/rs690.c +++ b/drivers/gpu/drm/radeon/rs690.c @@ -162,6 +162,16 @@ static void rs690_mc_init(struct radeon_device *rdev) base = RREG32_MC(R_000100_MCCFG_FB_LOCATION); base = G_000100_MC_FB_START(base) << 16; rdev->mc.igp_sideport_enabled = radeon_atombios_sideport_present(rdev); + /* Some boards seem to be configured for 128MB of sideport memory, + * but really only have 64MB. Just skip the sideport and use + * UMA memory. + */ + if (rdev->mc.igp_sideport_enabled && + (rdev->mc.real_vram_size == (384 * 1024 * 1024))) { + base += 128 * 1024 * 1024; + rdev->mc.real_vram_size -= 128 * 1024 * 1024; + rdev->mc.mc_vram_size = rdev->mc.real_vram_size; + } /* Use K8 direct mapping for fast fb access. */ rdev->fastfb_working = false; @@ -345,9 +355,11 @@ static void rs690_crtc_bandwidth_compute(struct radeon_device *rdev, if (max_bandwidth.full > rdev->pm.sideport_bandwidth.full && rdev->pm.sideport_bandwidth.full) max_bandwidth = rdev->pm.sideport_bandwidth; - read_delay_latency.full = dfixed_const(370 * 800 * 1000); - read_delay_latency.full = dfixed_div(read_delay_latency, - rdev->pm.igp_sideport_mclk); + read_delay_latency.full = dfixed_const(370 * 800); + a.full = dfixed_const(1000); + b.full = dfixed_div(rdev->pm.igp_sideport_mclk, a); + read_delay_latency.full = dfixed_div(read_delay_latency, b); + read_delay_latency.full = dfixed_mul(read_delay_latency, a); } else { if (max_bandwidth.full > rdev->pm.k8_bandwidth.full && rdev->pm.k8_bandwidth.full) @@ -488,14 +500,10 @@ static void rs690_compute_mode_priority(struct radeon_device *rdev, } if (wm0->priority_mark.full > priority_mark02.full) priority_mark02.full = wm0->priority_mark.full; - if (dfixed_trunc(priority_mark02) < 0) - priority_mark02.full = 0; if (wm0->priority_mark_max.full > priority_mark02.full) priority_mark02.full = wm0->priority_mark_max.full; if (wm1->priority_mark.full > priority_mark12.full) priority_mark12.full = wm1->priority_mark.full; - if (dfixed_trunc(priority_mark12) < 0) - priority_mark12.full = 0; if (wm1->priority_mark_max.full > priority_mark12.full) priority_mark12.full = wm1->priority_mark_max.full; *d1mode_priority_a_cnt = dfixed_trunc(priority_mark02); @@ -526,8 +534,6 @@ static void rs690_compute_mode_priority(struct radeon_device *rdev, } if (wm0->priority_mark.full > priority_mark02.full) priority_mark02.full = wm0->priority_mark.full; - if (dfixed_trunc(priority_mark02) < 0) - priority_mark02.full = 0; if (wm0->priority_mark_max.full > priority_mark02.full) priority_mark02.full = wm0->priority_mark_max.full; *d1mode_priority_a_cnt = dfixed_trunc(priority_mark02); @@ -555,8 +561,6 @@ static void rs690_compute_mode_priority(struct radeon_device *rdev, } if (wm1->priority_mark.full > priority_mark12.full) priority_mark12.full = wm1->priority_mark.full; - if (dfixed_trunc(priority_mark12) < 0) - priority_mark12.full = 0; if (wm1->priority_mark_max.full > priority_mark12.full) priority_mark12.full = wm1->priority_mark_max.full; *d2mode_priority_a_cnt = dfixed_trunc(priority_mark12); @@ -631,20 +635,27 @@ void rs690_bandwidth_update(struct radeon_device *rdev) uint32_t rs690_mc_rreg(struct radeon_device *rdev, uint32_t reg) { + unsigned long flags; uint32_t r; + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(R_000078_MC_INDEX, S_000078_MC_IND_ADDR(reg)); r = RREG32(R_00007C_MC_DATA); WREG32(R_000078_MC_INDEX, ~C_000078_MC_IND_ADDR); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); return r; } void rs690_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(R_000078_MC_INDEX, S_000078_MC_IND_ADDR(reg) | S_000078_MC_IND_WR_EN(1)); WREG32(R_00007C_MC_DATA, v); WREG32(R_000078_MC_INDEX, 0x7F); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); } static void rs690_mc_program(struct radeon_device *rdev) @@ -755,6 +766,7 @@ int rs690_resume(struct radeon_device *rdev) int rs690_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r600_audio_fini(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); @@ -765,6 +777,7 @@ int rs690_suspend(struct radeon_device *rdev) void rs690_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); r600_audio_fini(rdev); r100_cp_fini(rdev); radeon_wb_fini(rdev); @@ -834,6 +847,9 @@ int rs690_init(struct radeon_device *rdev) return r; rs600_set_safe_registers(rdev); + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->accel_working = true; r = rs690_startup(rdev); if (r) { diff --git a/drivers/gpu/drm/radeon/rs780_dpm.c b/drivers/gpu/drm/radeon/rs780_dpm.c index bef832a62fe..02f7710de47 100644 --- a/drivers/gpu/drm/radeon/rs780_dpm.c +++ b/drivers/gpu/drm/radeon/rs780_dpm.c @@ -28,6 +28,7 @@ #include "r600_dpm.h" #include "rs780_dpm.h" #include "atom.h" +#include <linux/seq_file.h> static struct igp_ps *rs780_get_ps(struct radeon_ps *rps) { @@ -61,9 +62,7 @@ static void rs780_get_pm_mode_parameters(struct radeon_device *rdev) radeon_crtc = to_radeon_crtc(crtc); pi->crtc_id = radeon_crtc->crtc_id; if (crtc->mode.htotal && crtc->mode.vtotal) - pi->refresh_rate = - (crtc->mode.clock * 1000) / - (crtc->mode.htotal * crtc->mode.vtotal); + pi->refresh_rate = drm_mode_vrefresh(&crtc->mode); break; } } @@ -375,9 +374,8 @@ static void rs780_disable_vbios_powersaving(struct radeon_device *rdev) WREG32_P(CG_INTGFX_MISC, 0, ~0xFFF00000); } -static void rs780_force_voltage_to_high(struct radeon_device *rdev) +static void rs780_force_voltage(struct radeon_device *rdev, u16 voltage) { - struct igp_power_info *pi = rs780_get_pi(rdev); struct igp_ps *current_state = rs780_get_ps(rdev->pm.dpm.current_ps); if ((current_state->max_voltage == RS780_VDDC_LEVEL_HIGH) && @@ -389,7 +387,7 @@ static void rs780_force_voltage_to_high(struct radeon_device *rdev) udelay(1); WREG32_P(FVTHROT_PWM_CTRL_REG0, - STARTING_PWM_HIGHTIME(pi->max_voltage), + STARTING_PWM_HIGHTIME(voltage), ~STARTING_PWM_HIGHTIME_MASK); WREG32_P(FVTHROT_PWM_CTRL_REG0, @@ -403,6 +401,26 @@ static void rs780_force_voltage_to_high(struct radeon_device *rdev) WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~SPLL_BYPASS_CNTL); } +static void rs780_force_fbdiv(struct radeon_device *rdev, u32 fb_div) +{ + struct igp_ps *current_state = rs780_get_ps(rdev->pm.dpm.current_ps); + + if (current_state->sclk_low == current_state->sclk_high) + return; + + WREG32_P(GFX_MACRO_BYPASS_CNTL, SPLL_BYPASS_CNTL, ~SPLL_BYPASS_CNTL); + + WREG32_P(FVTHROT_FBDIV_REG2, FORCED_FEEDBACK_DIV(fb_div), + ~FORCED_FEEDBACK_DIV_MASK); + WREG32_P(FVTHROT_FBDIV_REG1, STARTING_FEEDBACK_DIV(fb_div), + ~STARTING_FEEDBACK_DIV_MASK); + WREG32_P(FVTHROT_FBDIV_REG1, FORCE_FEEDBACK_DIV, ~FORCE_FEEDBACK_DIV); + + udelay(100); + + WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~SPLL_BYPASS_CNTL); +} + static int rs780_set_engine_clock_scaling(struct radeon_device *rdev, struct radeon_ps *new_ps, struct radeon_ps *old_ps) @@ -431,17 +449,13 @@ static int rs780_set_engine_clock_scaling(struct radeon_device *rdev, if (ret) return ret; - WREG32_P(GFX_MACRO_BYPASS_CNTL, SPLL_BYPASS_CNTL, ~SPLL_BYPASS_CNTL); - - WREG32_P(FVTHROT_FBDIV_REG2, FORCED_FEEDBACK_DIV(max_dividers.fb_div), - ~FORCED_FEEDBACK_DIV_MASK); - WREG32_P(FVTHROT_FBDIV_REG1, STARTING_FEEDBACK_DIV(max_dividers.fb_div), - ~STARTING_FEEDBACK_DIV_MASK); - WREG32_P(FVTHROT_FBDIV_REG1, FORCE_FEEDBACK_DIV, ~FORCE_FEEDBACK_DIV); - - udelay(100); + if ((min_dividers.ref_div != max_dividers.ref_div) || + (min_dividers.post_div != max_dividers.post_div) || + (max_dividers.ref_div != current_max_dividers.ref_div) || + (max_dividers.post_div != current_max_dividers.post_div)) + return -EINVAL; - WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~SPLL_BYPASS_CNTL); + rs780_force_fbdiv(rdev, max_dividers.fb_div); if (max_dividers.fb_div > min_dividers.fb_div) { WREG32_P(FVTHROT_FBDIV_REG0, @@ -485,6 +499,9 @@ static void rs780_activate_engine_clk_scaling(struct radeon_device *rdev, (new_state->sclk_low == old_state->sclk_low)) return; + if (new_state->sclk_high == new_state->sclk_low) + return; + rs780_clk_scaling_enable(rdev, true); } @@ -606,14 +623,6 @@ int rs780_dpm_enable(struct radeon_device *rdev) if (pi->gfx_clock_gating) r600_gfx_clockgating_enable(rdev, true); - if (rdev->irq.installed && (rdev->pm.int_thermal_type == THERMAL_TYPE_RV6XX)) { - ret = r600_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX); - if (ret) - return ret; - rdev->irq.dpm_thermal = true; - radeon_irq_set(rdev); - } - return 0; } @@ -648,7 +657,7 @@ int rs780_dpm_set_power_state(struct radeon_device *rdev) rs780_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps); if (pi->voltage_control) { - rs780_force_voltage_to_high(rdev); + rs780_force_voltage(rdev, pi->max_voltage); mdelay(5); } @@ -716,14 +725,18 @@ static void rs780_parse_pplib_non_clock_info(struct radeon_device *rdev, if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) { rps->vclk = le32_to_cpu(non_clock_info->ulVCLK); rps->dclk = le32_to_cpu(non_clock_info->ulDCLK); - } else if (r600_is_uvd_state(rps->class, rps->class2)) { - rps->vclk = RS780_DEFAULT_VCLK_FREQ; - rps->dclk = RS780_DEFAULT_DCLK_FREQ; } else { rps->vclk = 0; rps->dclk = 0; } + if (r600_is_uvd_state(rps->class, rps->class2)) { + if ((rps->vclk == 0) || (rps->dclk == 0)) { + rps->vclk = RS780_DEFAULT_VCLK_FREQ; + rps->dclk = RS780_DEFAULT_DCLK_FREQ; + } + } + if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) rdev->pm.dpm.boot_ps = rps; if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE) @@ -794,9 +807,6 @@ static int rs780_parse_power_table(struct radeon_device *rdev) power_info->pplib.ucNumStates, GFP_KERNEL); if (!rdev->pm.dpm.ps) return -ENOMEM; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < power_info->pplib.ucNumStates; i++) { power_state = (union pplib_power_state *) @@ -846,6 +856,10 @@ int rs780_dpm_init(struct radeon_device *rdev) return -ENOMEM; rdev->pm.dpm.priv = pi; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = rs780_parse_power_table(rdev); if (ret) return ret; @@ -961,3 +975,79 @@ u32 rs780_dpm_get_mclk(struct radeon_device *rdev, bool low) return pi->bootup_uma_clk; } + +void rs780_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, + struct seq_file *m) +{ + struct radeon_ps *rps = rdev->pm.dpm.current_ps; + struct igp_ps *ps = rs780_get_ps(rps); + u32 current_fb_div = RREG32(FVTHROT_STATUS_REG0) & CURRENT_FEEDBACK_DIV_MASK; + u32 func_cntl = RREG32(CG_SPLL_FUNC_CNTL); + u32 ref_div = ((func_cntl & SPLL_REF_DIV_MASK) >> SPLL_REF_DIV_SHIFT) + 1; + u32 post_div = ((func_cntl & SPLL_SW_HILEN_MASK) >> SPLL_SW_HILEN_SHIFT) + 1 + + ((func_cntl & SPLL_SW_LOLEN_MASK) >> SPLL_SW_LOLEN_SHIFT) + 1; + u32 sclk = (rdev->clock.spll.reference_freq * current_fb_div) / + (post_div * ref_div); + + seq_printf(m, "uvd vclk: %d dclk: %d\n", rps->vclk, rps->dclk); + + /* guess based on the current sclk */ + if (sclk < (ps->sclk_low + 500)) + seq_printf(m, "power level 0 sclk: %u vddc_index: %d\n", + ps->sclk_low, ps->min_voltage); + else + seq_printf(m, "power level 1 sclk: %u vddc_index: %d\n", + ps->sclk_high, ps->max_voltage); +} + +int rs780_dpm_force_performance_level(struct radeon_device *rdev, + enum radeon_dpm_forced_level level) +{ + struct igp_power_info *pi = rs780_get_pi(rdev); + struct radeon_ps *rps = rdev->pm.dpm.current_ps; + struct igp_ps *ps = rs780_get_ps(rps); + struct atom_clock_dividers dividers; + int ret; + + rs780_clk_scaling_enable(rdev, false); + rs780_voltage_scaling_enable(rdev, false); + + if (level == RADEON_DPM_FORCED_LEVEL_HIGH) { + if (pi->voltage_control) + rs780_force_voltage(rdev, pi->max_voltage); + + ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM, + ps->sclk_high, false, ÷rs); + if (ret) + return ret; + + rs780_force_fbdiv(rdev, dividers.fb_div); + } else if (level == RADEON_DPM_FORCED_LEVEL_LOW) { + ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM, + ps->sclk_low, false, ÷rs); + if (ret) + return ret; + + rs780_force_fbdiv(rdev, dividers.fb_div); + + if (pi->voltage_control) + rs780_force_voltage(rdev, pi->min_voltage); + } else { + if (pi->voltage_control) + rs780_force_voltage(rdev, pi->max_voltage); + + if (ps->sclk_high != ps->sclk_low) { + WREG32_P(FVTHROT_FBDIV_REG1, 0, ~FORCE_FEEDBACK_DIV); + rs780_clk_scaling_enable(rdev, true); + } + + if (pi->voltage_control) { + rs780_voltage_scaling_enable(rdev, true); + rs780_enable_voltage_scaling(rdev, rps); + } + } + + rdev->pm.dpm.forced_level = level; + + return 0; +} diff --git a/drivers/gpu/drm/radeon/rs780d.h b/drivers/gpu/drm/radeon/rs780d.h index b1142ed1c62..cfbe9a43d97 100644 --- a/drivers/gpu/drm/radeon/rs780d.h +++ b/drivers/gpu/drm/radeon/rs780d.h @@ -28,6 +28,7 @@ # define SPLL_SLEEP (1 << 1) # define SPLL_REF_DIV(x) ((x) << 2) # define SPLL_REF_DIV_MASK (7 << 2) +# define SPLL_REF_DIV_SHIFT 2 # define SPLL_FB_DIV(x) ((x) << 5) # define SPLL_FB_DIV_MASK (0xff << 2) # define SPLL_FB_DIV_SHIFT 2 @@ -36,8 +37,10 @@ # define SPLL_PULSENUM_MASK (3 << 14) # define SPLL_SW_HILEN(x) ((x) << 16) # define SPLL_SW_HILEN_MASK (0xf << 16) +# define SPLL_SW_HILEN_SHIFT 16 # define SPLL_SW_LOLEN(x) ((x) << 20) # define SPLL_SW_LOLEN_MASK (0xf << 20) +# define SPLL_SW_LOLEN_SHIFT 20 # define SPLL_DIVEN (1 << 24) # define SPLL_BYPASS_EN (1 << 25) # define SPLL_CHG_STATUS (1 << 29) diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c index 8ea1573ae82..3e21e869015 100644 --- a/drivers/gpu/drm/radeon/rv515.c +++ b/drivers/gpu/drm/radeon/rv515.c @@ -209,19 +209,27 @@ static void rv515_mc_init(struct radeon_device *rdev) uint32_t rv515_mc_rreg(struct radeon_device *rdev, uint32_t reg) { + unsigned long flags; uint32_t r; + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(MC_IND_INDEX, 0x7f0000 | (reg & 0xffff)); r = RREG32(MC_IND_DATA); WREG32(MC_IND_INDEX, 0); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); + return r; } void rv515_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) { + unsigned long flags; + + spin_lock_irqsave(&rdev->mc_idx_lock, flags); WREG32(MC_IND_INDEX, 0xff0000 | ((reg) & 0xffff)); WREG32(MC_IND_DATA, (v)); WREG32(MC_IND_INDEX, 0); + spin_unlock_irqrestore(&rdev->mc_idx_lock, flags); } #if defined(CONFIG_DEBUG_FS) @@ -398,8 +406,9 @@ void rv515_mc_resume(struct radeon_device *rdev, struct rv515_mc_save *save) for (i = 0; i < rdev->num_crtc; i++) { if (save->crtc_enabled[i]) { tmp = RREG32(AVIVO_D1MODE_MASTER_UPDATE_MODE + crtc_offsets[i]); - if ((tmp & 0x3) != 0) { - tmp &= ~0x3; + if ((tmp & 0x7) != 3) { + tmp &= ~0x7; + tmp |= 0x3; WREG32(AVIVO_D1MODE_MASTER_UPDATE_MODE + crtc_offsets[i], tmp); } tmp = RREG32(AVIVO_D1GRPH_UPDATE + crtc_offsets[i]); @@ -588,6 +597,7 @@ int rv515_resume(struct radeon_device *rdev) int rv515_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); rs600_irq_disable(rdev); @@ -604,6 +614,7 @@ void rv515_set_safe_registers(struct radeon_device *rdev) void rv515_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); r100_cp_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); @@ -677,6 +688,9 @@ int rv515_init(struct radeon_device *rdev) return r; rv515_set_safe_registers(rdev); + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->accel_working = true; r = rv515_startup(rdev); if (r) { @@ -1147,14 +1161,10 @@ static void rv515_compute_mode_priority(struct radeon_device *rdev, } if (wm0->priority_mark.full > priority_mark02.full) priority_mark02.full = wm0->priority_mark.full; - if (dfixed_trunc(priority_mark02) < 0) - priority_mark02.full = 0; if (wm0->priority_mark_max.full > priority_mark02.full) priority_mark02.full = wm0->priority_mark_max.full; if (wm1->priority_mark.full > priority_mark12.full) priority_mark12.full = wm1->priority_mark.full; - if (dfixed_trunc(priority_mark12) < 0) - priority_mark12.full = 0; if (wm1->priority_mark_max.full > priority_mark12.full) priority_mark12.full = wm1->priority_mark_max.full; *d1mode_priority_a_cnt = dfixed_trunc(priority_mark02); @@ -1185,8 +1195,6 @@ static void rv515_compute_mode_priority(struct radeon_device *rdev, } if (wm0->priority_mark.full > priority_mark02.full) priority_mark02.full = wm0->priority_mark.full; - if (dfixed_trunc(priority_mark02) < 0) - priority_mark02.full = 0; if (wm0->priority_mark_max.full > priority_mark02.full) priority_mark02.full = wm0->priority_mark_max.full; *d1mode_priority_a_cnt = dfixed_trunc(priority_mark02); @@ -1214,8 +1222,6 @@ static void rv515_compute_mode_priority(struct radeon_device *rdev, } if (wm1->priority_mark.full > priority_mark12.full) priority_mark12.full = wm1->priority_mark.full; - if (dfixed_trunc(priority_mark12) < 0) - priority_mark12.full = 0; if (wm1->priority_mark_max.full > priority_mark12.full) priority_mark12.full = wm1->priority_mark_max.full; *d2mode_priority_a_cnt = dfixed_trunc(priority_mark12); diff --git a/drivers/gpu/drm/radeon/rv6xx_dpm.c b/drivers/gpu/drm/radeon/rv6xx_dpm.c index 8303de267ee..e7045b08571 100644 --- a/drivers/gpu/drm/radeon/rv6xx_dpm.c +++ b/drivers/gpu/drm/radeon/rv6xx_dpm.c @@ -407,9 +407,9 @@ static void rv6xx_enable_engine_feedback_and_reference_sync(struct radeon_device WREG32_P(SPLL_CNTL_MODE, SPLL_DIV_SYNC, ~SPLL_DIV_SYNC); } -static u64 rv6xx_clocks_per_unit(u32 unit) +static u32 rv6xx_clocks_per_unit(u32 unit) { - u64 tmp = 1 << (2 * unit); + u32 tmp = 1 << (2 * unit); return tmp; } @@ -417,7 +417,7 @@ static u64 rv6xx_clocks_per_unit(u32 unit) static u32 rv6xx_scale_count_given_unit(struct radeon_device *rdev, u32 unscaled_count, u32 unit) { - u32 count_per_unit = (u32)rv6xx_clocks_per_unit(unit); + u32 count_per_unit = rv6xx_clocks_per_unit(unit); return (unscaled_count + count_per_unit - 1) / count_per_unit; } @@ -819,7 +819,7 @@ static void rv6xx_program_memory_timing_parameters(struct radeon_device *rdev) POWERMODE1(calculate_memory_refresh_rate(rdev, pi->hw.sclks[R600_POWER_LEVEL_MEDIUM])) | POWERMODE2(calculate_memory_refresh_rate(rdev, - pi->hw.sclks[R600_POWER_LEVEL_MEDIUM])) | + pi->hw.sclks[R600_POWER_LEVEL_HIGH])) | POWERMODE3(calculate_memory_refresh_rate(rdev, pi->hw.sclks[R600_POWER_LEVEL_HIGH]))); WREG32(ARB_RFSH_RATE, arb_refresh_rate); @@ -1182,10 +1182,10 @@ static void rv6xx_program_display_gap(struct radeon_device *rdev) u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL); tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK); - if (RREG32(AVIVO_D1CRTC_CONTROL) & AVIVO_CRTC_EN) { + if (rdev->pm.dpm.new_active_crtcs & 1) { tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK); tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE); - } else if (RREG32(AVIVO_D2CRTC_CONTROL) & AVIVO_CRTC_EN) { + } else if (rdev->pm.dpm.new_active_crtcs & 2) { tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE); tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK); } else { @@ -1546,7 +1546,6 @@ int rv6xx_dpm_enable(struct radeon_device *rdev) { struct rv6xx_power_info *pi = rv6xx_get_pi(rdev); struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps; - int ret; if (r600_dynamicpm_enabled(rdev)) return -EINVAL; @@ -1594,15 +1593,6 @@ int rv6xx_dpm_enable(struct radeon_device *rdev) r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, true); r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, true); - if (rdev->irq.installed && - r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { - ret = r600_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX); - if (ret) - return ret; - rdev->irq.dpm_thermal = true; - radeon_irq_set(rdev); - } - rv6xx_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true); r600_start_dpm(rdev); @@ -1670,6 +1660,8 @@ int rv6xx_dpm_set_power_state(struct radeon_device *rdev) struct radeon_ps *old_ps = rdev->pm.dpm.current_ps; int ret; + pi->restricted_levels = 0; + rv6xx_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps); rv6xx_clear_vc(rdev); @@ -1763,12 +1755,14 @@ void rv6xx_setup_asic(struct radeon_device *rdev) { r600_enable_acpi_pm(rdev); - if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L0s) - rv6xx_enable_l0s(rdev); - if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L1) - rv6xx_enable_l1(rdev); - if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_TURNOFFPLL_ASPML1) - rv6xx_enable_pll_sleep_in_l1(rdev); + if (radeon_aspm != 0) { + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L0s) + rv6xx_enable_l0s(rdev); + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L1) + rv6xx_enable_l1(rdev); + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_TURNOFFPLL_ASPML1) + rv6xx_enable_pll_sleep_in_l1(rdev); + } } void rv6xx_dpm_display_configuration_changed(struct radeon_device *rdev) @@ -1897,9 +1891,6 @@ static int rv6xx_parse_power_table(struct radeon_device *rdev) power_info->pplib.ucNumStates, GFP_KERNEL); if (!rdev->pm.dpm.ps) return -ENOMEM; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < power_info->pplib.ucNumStates; i++) { power_state = (union pplib_power_state *) @@ -1912,6 +1903,7 @@ static int rv6xx_parse_power_table(struct radeon_device *rdev) (power_state->v1.ucNonClockStateIndex * power_info->pplib.ucNonClockSize)); if (power_info->pplib.ucStateEntrySize - 1) { + u8 *idx; ps = kzalloc(sizeof(struct rv6xx_ps), GFP_KERNEL); if (ps == NULL) { kfree(rdev->pm.dpm.ps); @@ -1920,12 +1912,12 @@ static int rv6xx_parse_power_table(struct radeon_device *rdev) rdev->pm.dpm.ps[i].ps_priv = ps; rv6xx_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i], non_clock_info); + idx = (u8 *)&power_state->v1.ucClockStateIndices[0]; for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) { clock_info = (union pplib_clock_info *) (mode_info->atom_context->bios + data_offset + le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) + - (power_state->v1.ucClockStateIndices[j] * - power_info->pplib.ucClockInfoSize)); + (idx[j] * power_info->pplib.ucClockInfoSize)); rv6xx_parse_pplib_clock_info(rdev, &rdev->pm.dpm.ps[i], j, clock_info); @@ -1938,9 +1930,7 @@ static int rv6xx_parse_power_table(struct radeon_device *rdev) int rv6xx_dpm_init(struct radeon_device *rdev) { - int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info); - uint16_t data_offset, size; - uint8_t frev, crev; + struct radeon_atom_ss ss; struct atom_clock_dividers dividers; struct rv6xx_power_info *pi; int ret; @@ -1950,6 +1940,10 @@ int rv6xx_dpm_init(struct radeon_device *rdev) return -ENOMEM; rdev->pm.dpm.priv = pi; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = rv6xx_parse_power_table(rdev); if (ret) return ret; @@ -1983,16 +1977,18 @@ int rv6xx_dpm_init(struct radeon_device *rdev) pi->gfx_clock_gating = true; - if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size, - &frev, &crev, &data_offset)) { - pi->sclk_ss = true; - pi->mclk_ss = true; + pi->sclk_ss = radeon_atombios_get_asic_ss_info(rdev, &ss, + ASIC_INTERNAL_ENGINE_SS, 0); + pi->mclk_ss = radeon_atombios_get_asic_ss_info(rdev, &ss, + ASIC_INTERNAL_MEMORY_SS, 0); + + /* Disable sclk ss, causes hangs on a lot of systems */ + pi->sclk_ss = false; + + if (pi->sclk_ss || pi->mclk_ss) pi->dynamic_ss = true; - } else { - pi->sclk_ss = false; - pi->mclk_ss = false; + else pi->dynamic_ss = false; - } pi->dynamic_pcie_gen2 = true; @@ -2083,3 +2079,34 @@ u32 rv6xx_dpm_get_mclk(struct radeon_device *rdev, bool low) else return requested_state->high.mclk; } + +int rv6xx_dpm_force_performance_level(struct radeon_device *rdev, + enum radeon_dpm_forced_level level) +{ + struct rv6xx_power_info *pi = rv6xx_get_pi(rdev); + + if (level == RADEON_DPM_FORCED_LEVEL_HIGH) { + pi->restricted_levels = 3; + } else if (level == RADEON_DPM_FORCED_LEVEL_LOW) { + pi->restricted_levels = 2; + } else { + pi->restricted_levels = 0; + } + + rv6xx_clear_vc(rdev); + r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true); + r600_set_at(rdev, 0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF); + r600_wait_for_power_level(rdev, R600_POWER_LEVEL_LOW); + r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, false); + r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, false); + rv6xx_enable_medium(rdev); + rv6xx_enable_high(rdev); + if (pi->restricted_levels == 3) + r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, false); + rv6xx_program_vc(rdev); + rv6xx_program_at(rdev); + + rdev->pm.dpm.forced_level = level; + + return 0; +} diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c index 4a62ad2e539..da8703d8d45 100644 --- a/drivers/gpu/drm/radeon/rv770.c +++ b/drivers/gpu/drm/radeon/rv770.c @@ -744,10 +744,10 @@ static void rv770_init_golden_registers(struct radeon_device *rdev) (const u32)ARRAY_SIZE(r7xx_golden_dyn_gpr_registers)); radeon_program_register_sequence(rdev, rv730_golden_registers, - (const u32)ARRAY_SIZE(rv770_golden_registers)); + (const u32)ARRAY_SIZE(rv730_golden_registers)); radeon_program_register_sequence(rdev, rv730_mgcg_init, - (const u32)ARRAY_SIZE(rv770_mgcg_init)); + (const u32)ARRAY_SIZE(rv730_mgcg_init)); break; case CHIP_RV710: radeon_program_register_sequence(rdev, @@ -758,18 +758,18 @@ static void rv770_init_golden_registers(struct radeon_device *rdev) (const u32)ARRAY_SIZE(r7xx_golden_dyn_gpr_registers)); radeon_program_register_sequence(rdev, rv710_golden_registers, - (const u32)ARRAY_SIZE(rv770_golden_registers)); + (const u32)ARRAY_SIZE(rv710_golden_registers)); radeon_program_register_sequence(rdev, rv710_mgcg_init, - (const u32)ARRAY_SIZE(rv770_mgcg_init)); + (const u32)ARRAY_SIZE(rv710_mgcg_init)); break; case CHIP_RV740: radeon_program_register_sequence(rdev, rv740_golden_registers, - (const u32)ARRAY_SIZE(rv770_golden_registers)); + (const u32)ARRAY_SIZE(rv740_golden_registers)); radeon_program_register_sequence(rdev, rv740_mgcg_init, - (const u32)ARRAY_SIZE(rv770_mgcg_init)); + (const u32)ARRAY_SIZE(rv740_mgcg_init)); break; default: break; @@ -801,104 +801,7 @@ u32 rv770_get_xclk(struct radeon_device *rdev) return reference_clock; } -int rv770_uvd_resume(struct radeon_device *rdev) -{ - uint64_t addr; - uint32_t chip_id, size; - int r; - - r = radeon_uvd_resume(rdev); - if (r) - return r; - - /* programm the VCPU memory controller bits 0-27 */ - addr = rdev->uvd.gpu_addr >> 3; - size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 4) >> 3; - WREG32(UVD_VCPU_CACHE_OFFSET0, addr); - WREG32(UVD_VCPU_CACHE_SIZE0, size); - - addr += size; - size = RADEON_UVD_STACK_SIZE >> 3; - WREG32(UVD_VCPU_CACHE_OFFSET1, addr); - WREG32(UVD_VCPU_CACHE_SIZE1, size); - - addr += size; - size = RADEON_UVD_HEAP_SIZE >> 3; - WREG32(UVD_VCPU_CACHE_OFFSET2, addr); - WREG32(UVD_VCPU_CACHE_SIZE2, size); - - /* bits 28-31 */ - addr = (rdev->uvd.gpu_addr >> 28) & 0xF; - WREG32(UVD_LMI_ADDR_EXT, (addr << 12) | (addr << 0)); - - /* bits 32-39 */ - addr = (rdev->uvd.gpu_addr >> 32) & 0xFF; - WREG32(UVD_LMI_EXT40_ADDR, addr | (0x9 << 16) | (0x1 << 31)); - - /* tell firmware which hardware it is running on */ - switch (rdev->family) { - default: - return -EINVAL; - case CHIP_RV710: - chip_id = 0x01000005; - break; - case CHIP_RV730: - chip_id = 0x01000006; - break; - case CHIP_RV740: - chip_id = 0x01000007; - break; - case CHIP_CYPRESS: - case CHIP_HEMLOCK: - chip_id = 0x01000008; - break; - case CHIP_JUNIPER: - chip_id = 0x01000009; - break; - case CHIP_REDWOOD: - chip_id = 0x0100000a; - break; - case CHIP_CEDAR: - chip_id = 0x0100000b; - break; - case CHIP_SUMO: - case CHIP_SUMO2: - chip_id = 0x0100000c; - break; - case CHIP_PALM: - chip_id = 0x0100000e; - break; - case CHIP_CAYMAN: - chip_id = 0x0100000f; - break; - case CHIP_BARTS: - chip_id = 0x01000010; - break; - case CHIP_TURKS: - chip_id = 0x01000011; - break; - case CHIP_CAICOS: - chip_id = 0x01000012; - break; - case CHIP_TAHITI: - chip_id = 0x01000014; - break; - case CHIP_VERDE: - chip_id = 0x01000015; - break; - case CHIP_PITCAIRN: - chip_id = 0x01000016; - break; - case CHIP_ARUBA: - chip_id = 0x01000017; - break; - } - WREG32(UVD_VCPU_CHIP_ID, chip_id); - - return 0; -} - -u32 rv770_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) +void rv770_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) { struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id]; u32 tmp = RREG32(AVIVO_D1GRPH_UPDATE + radeon_crtc->crtc_offset); @@ -932,9 +835,15 @@ u32 rv770_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base) /* Unlock the lock, so double-buffering can take place inside vblank */ tmp &= ~AVIVO_D1GRPH_UPDATE_LOCK; WREG32(AVIVO_D1GRPH_UPDATE + radeon_crtc->crtc_offset, tmp); +} + +bool rv770_page_flip_pending(struct radeon_device *rdev, int crtc_id) +{ + struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id]; /* Return current update_pending status: */ - return RREG32(AVIVO_D1GRPH_UPDATE + radeon_crtc->crtc_offset) & AVIVO_D1GRPH_SURFACE_UPDATE_PENDING; + return !!(RREG32(AVIVO_D1GRPH_UPDATE + radeon_crtc->crtc_offset) & + AVIVO_D1GRPH_SURFACE_UPDATE_PENDING); } /* get temperature in millidegrees */ @@ -1168,7 +1077,8 @@ static void rv770_mc_program(struct radeon_device *rdev) */ void r700_cp_stop(struct radeon_device *rdev) { - radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT)); WREG32(SCRATCH_UMSK, 0); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; @@ -1220,6 +1130,35 @@ void r700_cp_fini(struct radeon_device *rdev) radeon_scratch_free(rdev, ring->rptr_save_reg); } +void rv770_set_clk_bypass_mode(struct radeon_device *rdev) +{ + u32 tmp, i; + + if (rdev->flags & RADEON_IS_IGP) + return; + + tmp = RREG32(CG_SPLL_FUNC_CNTL_2); + tmp &= SCLK_MUX_SEL_MASK; + tmp |= SCLK_MUX_SEL(1) | SCLK_MUX_UPDATE; + WREG32(CG_SPLL_FUNC_CNTL_2, tmp); + + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(CG_SPLL_STATUS) & SPLL_CHG_STATUS) + break; + udelay(1); + } + + tmp &= ~SCLK_MUX_UPDATE; + WREG32(CG_SPLL_FUNC_CNTL_2, tmp); + + tmp = RREG32(MPLL_CNTL_MODE); + if ((rdev->family == CHIP_RV710) || (rdev->family == CHIP_RV730)) + tmp &= ~RV730_MPLL_MCLK_SEL; + else + tmp &= ~MPLL_MCLK_SEL; + WREG32(MPLL_CNTL_MODE, tmp); +} + /* * Core functions */ @@ -1388,6 +1327,9 @@ static void rv770_gpu_init(struct radeon_device *rdev) if (tmp < rdev->config.rv770.max_simds) { rdev->config.rv770.max_simds = tmp; } + tmp = rdev->config.rv770.max_simds - + r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R7XX_MAX_SIMDS_MASK); + rdev->config.rv770.active_simds = tmp; switch (rdev->config.rv770.max_tile_pipes) { case 1: @@ -1747,80 +1689,6 @@ static int rv770_mc_init(struct radeon_device *rdev) return 0; } -/** - * rv770_copy_dma - copy pages using the DMA engine - * - * @rdev: radeon_device pointer - * @src_offset: src GPU address - * @dst_offset: dst GPU address - * @num_gpu_pages: number of GPU pages to xfer - * @fence: radeon fence object - * - * Copy GPU paging using the DMA engine (r7xx). - * Used by the radeon ttm implementation to move pages if - * registered as the asic copy callback. - */ -int rv770_copy_dma(struct radeon_device *rdev, - uint64_t src_offset, uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence **fence) -{ - struct radeon_semaphore *sem = NULL; - int ring_index = rdev->asic->copy.dma_ring_index; - struct radeon_ring *ring = &rdev->ring[ring_index]; - u32 size_in_dw, cur_size_in_dw; - int i, num_loops; - int r = 0; - - r = radeon_semaphore_create(rdev, &sem); - if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - return r; - } - - size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4; - num_loops = DIV_ROUND_UP(size_in_dw, 0xFFFF); - r = radeon_ring_lock(rdev, ring, num_loops * 5 + 8); - if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - radeon_semaphore_free(rdev, &sem, NULL); - return r; - } - - if (radeon_fence_need_sync(*fence, ring->idx)) { - radeon_semaphore_sync_rings(rdev, sem, (*fence)->ring, - ring->idx); - radeon_fence_note_sync(*fence, ring->idx); - } else { - radeon_semaphore_free(rdev, &sem, NULL); - } - - for (i = 0; i < num_loops; i++) { - cur_size_in_dw = size_in_dw; - if (cur_size_in_dw > 0xFFFF) - cur_size_in_dw = 0xFFFF; - size_in_dw -= cur_size_in_dw; - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 0, 0, cur_size_in_dw)); - radeon_ring_write(ring, dst_offset & 0xfffffffc); - radeon_ring_write(ring, src_offset & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff); - radeon_ring_write(ring, upper_32_bits(src_offset) & 0xff); - src_offset += cur_size_in_dw * 4; - dst_offset += cur_size_in_dw * 4; - } - - r = radeon_fence_emit(rdev, fence, ring->idx); - if (r) { - radeon_ring_unlock_undo(rdev, ring); - return r; - } - - radeon_ring_unlock_commit(rdev, ring); - radeon_semaphore_free(rdev, &sem, *fence); - - return r; -} - static int rv770_startup(struct radeon_device *rdev) { struct radeon_ring *ring; @@ -1829,19 +1697,13 @@ static int rv770_startup(struct radeon_device *rdev) /* enable pcie gen2 link */ rv770_pcie_gen2_enable(rdev); - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) { - r = r600_init_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load firmware!\n"); - return r; - } - } - + /* scratch needs to be initialized before MC */ r = r600_vram_scratch_init(rdev); if (r) return r; rv770_mc_program(rdev); + if (rdev->flags & RADEON_IS_AGP) { rv770_agp_enable(rdev); } else { @@ -1851,12 +1713,6 @@ static int rv770_startup(struct radeon_device *rdev) } rv770_gpu_init(rdev); - r = r600_blit_init(rdev); - if (r) { - r600_blit_fini(rdev); - rdev->asic->copy.copy = NULL; - dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); - } /* allocate wb buffer */ r = radeon_wb_init(rdev); @@ -1875,7 +1731,7 @@ static int rv770_startup(struct radeon_device *rdev) return r; } - r = rv770_uvd_resume(rdev); + r = uvd_v2_2_resume(rdev); if (!r) { r = radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_UVD_INDEX); @@ -1903,15 +1759,13 @@ static int rv770_startup(struct radeon_device *rdev) ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, - R600_CP_RB_RPTR, R600_CP_RB_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + RADEON_CP_PACKET2); if (r) return r; ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET, - DMA_RB_RPTR, DMA_RB_WPTR, - 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); + DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0)); if (r) return r; @@ -1928,12 +1782,10 @@ static int rv770_startup(struct radeon_device *rdev) ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; if (ring->ring_size) { - r = radeon_ring_init(rdev, ring, ring->ring_size, - R600_WB_UVD_RPTR_OFFSET, - UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, + RADEON_CP_PACKET2); if (!r) - r = r600_uvd_init(rdev); + r = uvd_v1_0_init(rdev); if (r) DRM_ERROR("radeon: failed initializing UVD (%d).\n", r); @@ -1968,6 +1820,9 @@ int rv770_resume(struct radeon_device *rdev) /* init golden registers */ rv770_init_golden_registers(rdev); + if (rdev->pm.pm_method == PM_METHOD_DPM) + radeon_pm_resume(rdev); + rdev->accel_working = true; r = rv770_startup(rdev); if (r) { @@ -1982,7 +1837,9 @@ int rv770_resume(struct radeon_device *rdev) int rv770_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r600_audio_fini(rdev); + uvd_v1_0_fini(rdev); radeon_uvd_suspend(rdev); r700_cp_stop(rdev); r600_dma_stop(rdev); @@ -2051,6 +1908,17 @@ int rv770_init(struct radeon_device *rdev) if (r) return r; + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) { + r = r600_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL; r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024); @@ -2090,15 +1958,16 @@ int rv770_init(struct radeon_device *rdev) void rv770_fini(struct radeon_device *rdev) { - r600_blit_fini(rdev); + radeon_pm_fini(rdev); r700_cp_fini(rdev); r600_dma_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); radeon_irq_kms_fini(rdev); - rv770_pcie_gart_fini(rdev); + uvd_v1_0_fini(rdev); radeon_uvd_fini(rdev); + rv770_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_gem_fini(rdev); radeon_fence_driver_fini(rdev); diff --git a/drivers/gpu/drm/radeon/rv770_dma.c b/drivers/gpu/drm/radeon/rv770_dma.c new file mode 100644 index 00000000000..bbf2e076ee4 --- /dev/null +++ b/drivers/gpu/drm/radeon/rv770_dma.c @@ -0,0 +1,97 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "rv770d.h" + +/** + * rv770_copy_dma - copy pages using the DMA engine + * + * @rdev: radeon_device pointer + * @src_offset: src GPU address + * @dst_offset: dst GPU address + * @num_gpu_pages: number of GPU pages to xfer + * @fence: radeon fence object + * + * Copy GPU paging using the DMA engine (r7xx). + * Used by the radeon ttm implementation to move pages if + * registered as the asic copy callback. + */ +int rv770_copy_dma(struct radeon_device *rdev, + uint64_t src_offset, uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence **fence) +{ + struct radeon_semaphore *sem = NULL; + int ring_index = rdev->asic->copy.dma_ring_index; + struct radeon_ring *ring = &rdev->ring[ring_index]; + u32 size_in_dw, cur_size_in_dw; + int i, num_loops; + int r = 0; + + r = radeon_semaphore_create(rdev, &sem); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + return r; + } + + size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4; + num_loops = DIV_ROUND_UP(size_in_dw, 0xFFFF); + r = radeon_ring_lock(rdev, ring, num_loops * 5 + 8); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_semaphore_sync_to(sem, *fence); + radeon_semaphore_sync_rings(rdev, sem, ring->idx); + + for (i = 0; i < num_loops; i++) { + cur_size_in_dw = size_in_dw; + if (cur_size_in_dw > 0xFFFF) + cur_size_in_dw = 0xFFFF; + size_in_dw -= cur_size_in_dw; + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 0, 0, cur_size_in_dw)); + radeon_ring_write(ring, dst_offset & 0xfffffffc); + radeon_ring_write(ring, src_offset & 0xfffffffc); + radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff); + radeon_ring_write(ring, upper_32_bits(src_offset) & 0xff); + src_offset += cur_size_in_dw * 4; + dst_offset += cur_size_in_dw * 4; + } + + r = radeon_fence_emit(rdev, fence, ring->idx); + if (r) { + radeon_ring_unlock_undo(rdev, ring); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_ring_unlock_commit(rdev, ring); + radeon_semaphore_free(rdev, &sem, *fence); + + return r; +} diff --git a/drivers/gpu/drm/radeon/rv770_dpm.c b/drivers/gpu/drm/radeon/rv770_dpm.c index d914e04ea39..3c76e1dcdf0 100644 --- a/drivers/gpu/drm/radeon/rv770_dpm.c +++ b/drivers/gpu/drm/radeon/rv770_dpm.c @@ -1863,8 +1863,8 @@ void rv770_enable_auto_throttle_source(struct radeon_device *rdev, } } -int rv770_set_thermal_temperature_range(struct radeon_device *rdev, - int min_temp, int max_temp) +static int rv770_set_thermal_temperature_range(struct radeon_device *rdev, + int min_temp, int max_temp) { int low_temp = 0 * 1000; int high_temp = 255 * 1000; @@ -1966,6 +1966,15 @@ int rv770_dpm_enable(struct radeon_device *rdev) if (pi->mg_clock_gating) rv770_mg_clock_gating_enable(rdev, true); + rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true); + + return 0; +} + +int rv770_dpm_late_enable(struct radeon_device *rdev) +{ + int ret; + if (rdev->irq.installed && r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { PPSMC_Result result; @@ -1981,8 +1990,6 @@ int rv770_dpm_enable(struct radeon_device *rdev) DRM_DEBUG_KMS("Could not enable thermal interrupts.\n"); } - rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true); - return 0; } @@ -2064,12 +2071,6 @@ int rv770_dpm_set_power_state(struct radeon_device *rdev) rv770_program_dcodt_after_state_switch(rdev, new_ps, old_ps); rv770_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps); - ret = rv770_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO); - if (ret) { - DRM_ERROR("rv770_dpm_force_performance_level failed\n"); - return ret; - } - return 0; } @@ -2099,12 +2100,14 @@ void rv770_dpm_setup_asic(struct radeon_device *rdev) rv770_enable_acpi_pm(rdev); - if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L0s) - rv770_enable_l0s(rdev); - if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L1) - rv770_enable_l1(rdev); - if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_TURNOFFPLL_ASPML1) - rv770_enable_pll_sleep_in_l1(rdev); + if (radeon_aspm != 0) { + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L0s) + rv770_enable_l0s(rdev); + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L1) + rv770_enable_l1(rdev); + if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_TURNOFFPLL_ASPML1) + rv770_enable_pll_sleep_in_l1(rdev); + } } void rv770_dpm_display_configuration_changed(struct radeon_device *rdev) @@ -2145,14 +2148,18 @@ static void rv7xx_parse_pplib_non_clock_info(struct radeon_device *rdev, if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) { rps->vclk = le32_to_cpu(non_clock_info->ulVCLK); rps->dclk = le32_to_cpu(non_clock_info->ulDCLK); - } else if (r600_is_uvd_state(rps->class, rps->class2)) { - rps->vclk = RV770_DEFAULT_VCLK_FREQ; - rps->dclk = RV770_DEFAULT_DCLK_FREQ; } else { rps->vclk = 0; rps->dclk = 0; } + if (r600_is_uvd_state(rps->class, rps->class2)) { + if ((rps->vclk == 0) || (rps->dclk == 0)) { + rps->vclk = RV770_DEFAULT_VCLK_FREQ; + rps->dclk = RV770_DEFAULT_DCLK_FREQ; + } + } + if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) rdev->pm.dpm.boot_ps = rps; if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE) @@ -2167,7 +2174,6 @@ static void rv7xx_parse_pplib_clock_info(struct radeon_device *rdev, struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev); struct rv7xx_ps *ps = rv770_get_ps(rps); u32 sclk, mclk; - u16 vddc; struct rv7xx_pl *pl; switch (index) { @@ -2207,8 +2213,8 @@ static void rv7xx_parse_pplib_clock_info(struct radeon_device *rdev, /* patch up vddc if necessary */ if (pl->vddc == 0xff01) { - if (radeon_atom_get_max_vddc(rdev, 0, 0, &vddc) == 0) - pl->vddc = vddc; + if (pi->max_vddc) + pl->vddc = pi->max_vddc; } if (rps->class & ATOM_PPLIB_CLASSIFICATION_ACPI) { @@ -2244,14 +2250,12 @@ static void rv7xx_parse_pplib_clock_info(struct radeon_device *rdev, pl->vddci = vddci; } - if (rdev->family >= CHIP_BARTS) { - if ((rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) == - ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE) { - rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk = pl->sclk; - rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.mclk = pl->mclk; - rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddc = pl->vddc; - rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddci = pl->vddci; - } + if ((rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) == + ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE) { + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk = pl->sclk; + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.mclk = pl->mclk; + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddc = pl->vddc; + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddci = pl->vddci; } } @@ -2277,9 +2281,6 @@ int rv7xx_parse_power_table(struct radeon_device *rdev) power_info->pplib.ucNumStates, GFP_KERNEL); if (!rdev->pm.dpm.ps) return -ENOMEM; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < power_info->pplib.ucNumStates; i++) { power_state = (union pplib_power_state *) @@ -2292,6 +2293,7 @@ int rv7xx_parse_power_table(struct radeon_device *rdev) (power_state->v1.ucNonClockStateIndex * power_info->pplib.ucNonClockSize)); if (power_info->pplib.ucStateEntrySize - 1) { + u8 *idx; ps = kzalloc(sizeof(struct rv7xx_ps), GFP_KERNEL); if (ps == NULL) { kfree(rdev->pm.dpm.ps); @@ -2301,12 +2303,12 @@ int rv7xx_parse_power_table(struct radeon_device *rdev) rv7xx_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i], non_clock_info, power_info->pplib.ucNonClockSize); + idx = (u8 *)&power_state->v1.ucClockStateIndices[0]; for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) { clock_info = (union pplib_clock_info *) (mode_info->atom_context->bios + data_offset + le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) + - (power_state->v1.ucClockStateIndices[j] * - power_info->pplib.ucClockInfoSize)); + (idx[j] * power_info->pplib.ucClockInfoSize)); rv7xx_parse_pplib_clock_info(rdev, &rdev->pm.dpm.ps[i], j, clock_info); @@ -2317,12 +2319,25 @@ int rv7xx_parse_power_table(struct radeon_device *rdev) return 0; } +void rv770_get_engine_memory_ss(struct radeon_device *rdev) +{ + struct rv7xx_power_info *pi = rv770_get_pi(rdev); + struct radeon_atom_ss ss; + + pi->sclk_ss = radeon_atombios_get_asic_ss_info(rdev, &ss, + ASIC_INTERNAL_ENGINE_SS, 0); + pi->mclk_ss = radeon_atombios_get_asic_ss_info(rdev, &ss, + ASIC_INTERNAL_MEMORY_SS, 0); + + if (pi->sclk_ss || pi->mclk_ss) + pi->dynamic_ss = true; + else + pi->dynamic_ss = false; +} + int rv770_dpm_init(struct radeon_device *rdev) { struct rv7xx_power_info *pi; - int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info); - uint16_t data_offset, size; - uint8_t frev, crev; struct atom_clock_dividers dividers; int ret; @@ -2337,6 +2352,10 @@ int rv770_dpm_init(struct radeon_device *rdev) pi->min_vddc_in_table = 0; pi->max_vddc_in_table = 0; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = rv7xx_parse_power_table(rdev); if (ret) return ret; @@ -2367,16 +2386,7 @@ int rv770_dpm_init(struct radeon_device *rdev) pi->mvdd_control = radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, 0); - if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size, - &frev, &crev, &data_offset)) { - pi->sclk_ss = true; - pi->mclk_ss = true; - pi->dynamic_ss = true; - } else { - pi->sclk_ss = false; - pi->mclk_ss = false; - pi->dynamic_ss = false; - } + rv770_get_engine_memory_ss(rdev); pi->asi = RV770_ASI_DFLT; pi->pasi = RV770_HASI_DFLT; @@ -2391,8 +2401,7 @@ int rv770_dpm_init(struct radeon_device *rdev) pi->dynamic_pcie_gen2 = true; - if (pi->gfx_clock_gating && - (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE)) + if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE) pi->thermal_protection = true; else pi->thermal_protection = false; @@ -2512,8 +2521,15 @@ u32 rv770_dpm_get_mclk(struct radeon_device *rdev, bool low) bool rv770_dpm_vblank_too_short(struct radeon_device *rdev) { u32 vblank_time = r600_dpm_get_vblank_time(rdev); + u32 switch_limit = 200; /* 300 */ + + /* RV770 */ + /* mclk switching doesn't seem to work reliably on desktop RV770s */ + if ((rdev->family == CHIP_RV770) && + !(rdev->flags & RADEON_IS_MOBILITY)) + switch_limit = 0xffffffff; /* disable mclk switching */ - if (vblank_time < 300) + if (vblank_time < switch_limit) return true; else return false; diff --git a/drivers/gpu/drm/radeon/rv770_dpm.h b/drivers/gpu/drm/radeon/rv770_dpm.h index 96b1b2a62a8..f776634840c 100644 --- a/drivers/gpu/drm/radeon/rv770_dpm.h +++ b/drivers/gpu/drm/radeon/rv770_dpm.h @@ -275,6 +275,7 @@ void rv770_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev, void rv770_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev, struct radeon_ps *new_ps, struct radeon_ps *old_ps); +void rv770_get_engine_memory_ss(struct radeon_device *rdev); /* smc */ int rv770_read_smc_soft_register(struct radeon_device *rdev, @@ -282,8 +283,4 @@ int rv770_read_smc_soft_register(struct radeon_device *rdev, int rv770_write_smc_soft_register(struct radeon_device *rdev, u16 reg_offset, u32 value); -/* thermal */ -int rv770_set_thermal_temperature_range(struct radeon_device *rdev, - int min_temp, int max_temp); - #endif diff --git a/drivers/gpu/drm/radeon/rv770_smc.c b/drivers/gpu/drm/radeon/rv770_smc.c index ab95da57021..b2a22440736 100644 --- a/drivers/gpu/drm/radeon/rv770_smc.c +++ b/drivers/gpu/drm/radeon/rv770_smc.c @@ -274,8 +274,8 @@ static const u8 cayman_smc_int_vectors[] = 0x08, 0x72, 0x08, 0x72 }; -int rv770_set_smc_sram_address(struct radeon_device *rdev, - u16 smc_address, u16 limit) +static int rv770_set_smc_sram_address(struct radeon_device *rdev, + u16 smc_address, u16 limit) { u32 addr; @@ -296,9 +296,10 @@ int rv770_copy_bytes_to_smc(struct radeon_device *rdev, u16 smc_start_address, const u8 *src, u16 byte_count, u16 limit) { + unsigned long flags; u32 data, original_data, extra_shift; u16 addr; - int ret; + int ret = 0; if (smc_start_address & 3) return -EINVAL; @@ -307,13 +308,14 @@ int rv770_copy_bytes_to_smc(struct radeon_device *rdev, addr = smc_start_address; + spin_lock_irqsave(&rdev->smc_idx_lock, flags); while (byte_count >= 4) { /* SMC address space is BE */ data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3]; ret = rv770_set_smc_sram_address(rdev, addr, limit); if (ret) - return ret; + goto done; WREG32(SMC_SRAM_DATA, data); @@ -328,7 +330,7 @@ int rv770_copy_bytes_to_smc(struct radeon_device *rdev, ret = rv770_set_smc_sram_address(rdev, addr, limit); if (ret) - return ret; + goto done; original_data = RREG32(SMC_SRAM_DATA); @@ -346,12 +348,15 @@ int rv770_copy_bytes_to_smc(struct radeon_device *rdev, ret = rv770_set_smc_sram_address(rdev, addr, limit); if (ret) - return ret; + goto done; WREG32(SMC_SRAM_DATA, data); } - return 0; +done: + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); + + return ret; } static int rv770_program_interrupt_vectors(struct radeon_device *rdev, @@ -461,12 +466,15 @@ PPSMC_Result rv770_wait_for_smc_inactive(struct radeon_device *rdev) static void rv770_clear_smc_sram(struct radeon_device *rdev, u16 limit) { + unsigned long flags; u16 i; + spin_lock_irqsave(&rdev->smc_idx_lock, flags); for (i = 0; i < limit; i += 4) { rv770_set_smc_sram_address(rdev, i, limit); WREG32(SMC_SRAM_DATA, 0); } + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); } int rv770_load_smc_ucode(struct radeon_device *rdev, @@ -595,27 +603,29 @@ int rv770_load_smc_ucode(struct radeon_device *rdev, int rv770_read_smc_sram_dword(struct radeon_device *rdev, u16 smc_address, u32 *value, u16 limit) { + unsigned long flags; int ret; + spin_lock_irqsave(&rdev->smc_idx_lock, flags); ret = rv770_set_smc_sram_address(rdev, smc_address, limit); - if (ret) - return ret; - - *value = RREG32(SMC_SRAM_DATA); + if (ret == 0) + *value = RREG32(SMC_SRAM_DATA); + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); - return 0; + return ret; } int rv770_write_smc_sram_dword(struct radeon_device *rdev, u16 smc_address, u32 value, u16 limit) { + unsigned long flags; int ret; + spin_lock_irqsave(&rdev->smc_idx_lock, flags); ret = rv770_set_smc_sram_address(rdev, smc_address, limit); - if (ret) - return ret; + if (ret == 0) + WREG32(SMC_SRAM_DATA, value); + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); - WREG32(SMC_SRAM_DATA, value); - - return 0; + return ret; } diff --git a/drivers/gpu/drm/radeon/rv770_smc.h b/drivers/gpu/drm/radeon/rv770_smc.h index f78d92a4b32..3b2c963c488 100644 --- a/drivers/gpu/drm/radeon/rv770_smc.h +++ b/drivers/gpu/drm/radeon/rv770_smc.h @@ -187,8 +187,6 @@ typedef struct RV770_SMC_STATETABLE RV770_SMC_STATETABLE; #define RV770_SMC_SOFT_REGISTER_uvd_enabled 0x9C #define RV770_SMC_SOFT_REGISTER_is_asic_lombok 0xA0 -int rv770_set_smc_sram_address(struct radeon_device *rdev, - u16 smc_address, u16 limit); int rv770_copy_bytes_to_smc(struct radeon_device *rdev, u16 smc_start_address, const u8 *src, u16 byte_count, u16 limit); diff --git a/drivers/gpu/drm/radeon/rv770d.h b/drivers/gpu/drm/radeon/rv770d.h index 6bef2b7d601..3cf1e292154 100644 --- a/drivers/gpu/drm/radeon/rv770d.h +++ b/drivers/gpu/drm/radeon/rv770d.h @@ -100,14 +100,21 @@ #define CG_SPLL_FUNC_CNTL_2 0x604 #define SCLK_MUX_SEL(x) ((x) << 0) #define SCLK_MUX_SEL_MASK (0x1ff << 0) +#define SCLK_MUX_UPDATE (1 << 26) #define CG_SPLL_FUNC_CNTL_3 0x608 #define SPLL_FB_DIV(x) ((x) << 0) #define SPLL_FB_DIV_MASK (0x3ffffff << 0) #define SPLL_DITHEN (1 << 28) +#define CG_SPLL_STATUS 0x60c +#define SPLL_CHG_STATUS (1 << 1) #define SPLL_CNTL_MODE 0x610 #define SPLL_DIV_SYNC (1 << 5) +#define MPLL_CNTL_MODE 0x61c +# define MPLL_MCLK_SEL (1 << 11) +# define RV730_MPLL_MCLK_SEL (1 << 25) + #define MPLL_AD_FUNC_CNTL 0x624 #define CLKF(x) ((x) << 0) #define CLKF_MASK (0x7f << 0) @@ -852,7 +859,7 @@ #define AFMT_VBI_PACKET_CONTROL 0x7608 # define AFMT_GENERIC0_UPDATE (1 << 2) #define AFMT_INFOFRAME_CONTROL0 0x760c -# define AFMT_AUDIO_INFO_SOURCE (1 << 6) /* 0 - sound block; 1 - hmdi regs */ +# define AFMT_AUDIO_INFO_SOURCE (1 << 6) /* 0 - sound block; 1 - hdmi regs */ # define AFMT_AUDIO_INFO_UPDATE (1 << 7) # define AFMT_MPEG_INFO_UPDATE (1 << 10) #define AFMT_GENERIC0_7 0x7610 @@ -971,7 +978,21 @@ # define TARGET_LINK_SPEED_MASK (0xf << 0) # define SELECTABLE_DEEMPHASIS (1 << 6) +/* + * PM4 + */ +#define PACKET0(reg, n) ((RADEON_PACKET_TYPE0 << 30) | \ + (((reg) >> 2) & 0xFFFF) | \ + ((n) & 0x3FFF) << 16) +#define PACKET3(op, n) ((RADEON_PACKET_TYPE3 << 30) | \ + (((op) & 0xFF) << 8) | \ + ((n) & 0x3FFF) << 16) + /* UVD */ +#define UVD_GPCOM_VCPU_CMD 0xef0c +#define UVD_GPCOM_VCPU_DATA0 0xef10 +#define UVD_GPCOM_VCPU_DATA1 0xef14 + #define UVD_LMI_EXT40_ADDR 0xf498 #define UVD_VCPU_CHIP_ID 0xf4d4 #define UVD_VCPU_CACHE_OFFSET0 0xf4d8 @@ -985,4 +1006,6 @@ #define UVD_RBC_RB_RPTR 0xf690 #define UVD_RBC_RB_WPTR 0xf694 +#define UVD_CONTEXT_ID 0xf6f4 + #endif diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index 23490670906..9e854fd016d 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -22,7 +22,6 @@ * Authors: Alex Deucher */ #include <linux/firmware.h> -#include <linux/platform_device.h> #include <linux/slab.h> #include <linux/module.h> #include <drm/drmP.h> @@ -40,35 +39,43 @@ MODULE_FIRMWARE("radeon/TAHITI_pfp.bin"); MODULE_FIRMWARE("radeon/TAHITI_me.bin"); MODULE_FIRMWARE("radeon/TAHITI_ce.bin"); MODULE_FIRMWARE("radeon/TAHITI_mc.bin"); +MODULE_FIRMWARE("radeon/TAHITI_mc2.bin"); MODULE_FIRMWARE("radeon/TAHITI_rlc.bin"); MODULE_FIRMWARE("radeon/TAHITI_smc.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_me.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_ce.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_mc.bin"); +MODULE_FIRMWARE("radeon/PITCAIRN_mc2.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_rlc.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_smc.bin"); MODULE_FIRMWARE("radeon/VERDE_pfp.bin"); MODULE_FIRMWARE("radeon/VERDE_me.bin"); MODULE_FIRMWARE("radeon/VERDE_ce.bin"); MODULE_FIRMWARE("radeon/VERDE_mc.bin"); +MODULE_FIRMWARE("radeon/VERDE_mc2.bin"); MODULE_FIRMWARE("radeon/VERDE_rlc.bin"); MODULE_FIRMWARE("radeon/VERDE_smc.bin"); MODULE_FIRMWARE("radeon/OLAND_pfp.bin"); MODULE_FIRMWARE("radeon/OLAND_me.bin"); MODULE_FIRMWARE("radeon/OLAND_ce.bin"); MODULE_FIRMWARE("radeon/OLAND_mc.bin"); +MODULE_FIRMWARE("radeon/OLAND_mc2.bin"); MODULE_FIRMWARE("radeon/OLAND_rlc.bin"); MODULE_FIRMWARE("radeon/OLAND_smc.bin"); MODULE_FIRMWARE("radeon/HAINAN_pfp.bin"); MODULE_FIRMWARE("radeon/HAINAN_me.bin"); MODULE_FIRMWARE("radeon/HAINAN_ce.bin"); MODULE_FIRMWARE("radeon/HAINAN_mc.bin"); +MODULE_FIRMWARE("radeon/HAINAN_mc2.bin"); MODULE_FIRMWARE("radeon/HAINAN_rlc.bin"); MODULE_FIRMWARE("radeon/HAINAN_smc.bin"); +static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh); static void si_pcie_gen3_enable(struct radeon_device *rdev); static void si_program_aspm(struct radeon_device *rdev); +extern void sumo_rlc_fini(struct radeon_device *rdev); +extern int sumo_rlc_init(struct radeon_device *rdev); extern int r600_ih_ring_alloc(struct radeon_device *rdev); extern void r600_ih_ring_fini(struct radeon_device *rdev); extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev); @@ -77,6 +84,13 @@ extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_ extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev); extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev); extern bool evergreen_is_display_hung(struct radeon_device *rdev); +static void si_enable_gui_idle_interrupt(struct radeon_device *rdev, + bool enable); +static void si_init_pg(struct radeon_device *rdev); +static void si_init_cg(struct radeon_device *rdev); +static void si_fini_pg(struct radeon_device *rdev); +static void si_fini_cg(struct radeon_device *rdev); +static void si_rlc_stop(struct radeon_device *rdev); static const u32 verde_rlc_save_restore_register_list[] = { @@ -1454,41 +1468,38 @@ static const u32 hainan_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = { }; /* ucode loading */ -static int si_mc_load_microcode(struct radeon_device *rdev) +int si_mc_load_microcode(struct radeon_device *rdev) { const __be32 *fw_data; u32 running, blackout = 0; u32 *io_mc_regs; - int i, ucode_size, regs_size; + int i, regs_size, ucode_size; if (!rdev->mc_fw) return -EINVAL; + ucode_size = rdev->mc_fw->size / 4; + switch (rdev->family) { case CHIP_TAHITI: io_mc_regs = (u32 *)&tahiti_io_mc_regs; - ucode_size = SI_MC_UCODE_SIZE; regs_size = TAHITI_IO_MC_REGS_SIZE; break; case CHIP_PITCAIRN: io_mc_regs = (u32 *)&pitcairn_io_mc_regs; - ucode_size = SI_MC_UCODE_SIZE; regs_size = TAHITI_IO_MC_REGS_SIZE; break; case CHIP_VERDE: default: io_mc_regs = (u32 *)&verde_io_mc_regs; - ucode_size = SI_MC_UCODE_SIZE; regs_size = TAHITI_IO_MC_REGS_SIZE; break; case CHIP_OLAND: io_mc_regs = (u32 *)&oland_io_mc_regs; - ucode_size = OLAND_MC_UCODE_SIZE; regs_size = TAHITI_IO_MC_REGS_SIZE; break; case CHIP_HAINAN: io_mc_regs = (u32 *)&hainan_io_mc_regs; - ucode_size = OLAND_MC_UCODE_SIZE; regs_size = TAHITI_IO_MC_REGS_SIZE; break; } @@ -1541,23 +1552,15 @@ static int si_mc_load_microcode(struct radeon_device *rdev) static int si_init_microcode(struct radeon_device *rdev) { - struct platform_device *pdev; const char *chip_name; const char *rlc_chip_name; size_t pfp_req_size, me_req_size, ce_req_size, rlc_req_size, mc_req_size; - size_t smc_req_size; + size_t smc_req_size, mc2_req_size; char fw_name[30]; int err; DRM_DEBUG("\n"); - pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0); - err = IS_ERR(pdev); - if (err) { - printk(KERN_ERR "radeon_cp: Failed to register firmware\n"); - return -EINVAL; - } - switch (rdev->family) { case CHIP_TAHITI: chip_name = "TAHITI"; @@ -1567,6 +1570,7 @@ static int si_init_microcode(struct radeon_device *rdev) ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; mc_req_size = SI_MC_UCODE_SIZE * 4; + mc2_req_size = TAHITI_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(TAHITI_SMC_UCODE_SIZE, 4); break; case CHIP_PITCAIRN: @@ -1577,6 +1581,7 @@ static int si_init_microcode(struct radeon_device *rdev) ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; mc_req_size = SI_MC_UCODE_SIZE * 4; + mc2_req_size = PITCAIRN_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(PITCAIRN_SMC_UCODE_SIZE, 4); break; case CHIP_VERDE: @@ -1587,6 +1592,7 @@ static int si_init_microcode(struct radeon_device *rdev) ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; mc_req_size = SI_MC_UCODE_SIZE * 4; + mc2_req_size = VERDE_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(VERDE_SMC_UCODE_SIZE, 4); break; case CHIP_OLAND: @@ -1596,7 +1602,7 @@ static int si_init_microcode(struct radeon_device *rdev) me_req_size = SI_PM4_UCODE_SIZE * 4; ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; - mc_req_size = OLAND_MC_UCODE_SIZE * 4; + mc_req_size = mc2_req_size = OLAND_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(OLAND_SMC_UCODE_SIZE, 4); break; case CHIP_HAINAN: @@ -1606,7 +1612,7 @@ static int si_init_microcode(struct radeon_device *rdev) me_req_size = SI_PM4_UCODE_SIZE * 4; ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; - mc_req_size = OLAND_MC_UCODE_SIZE * 4; + mc_req_size = mc2_req_size = OLAND_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(HAINAN_SMC_UCODE_SIZE, 4); break; default: BUG(); @@ -1615,7 +1621,7 @@ static int si_init_microcode(struct radeon_device *rdev) DRM_INFO("Loading %s Microcode\n", chip_name); snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name); - err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->pfp_fw->size != pfp_req_size) { @@ -1627,7 +1633,7 @@ static int si_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name); - err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->me_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->me_fw->size != me_req_size) { @@ -1638,7 +1644,7 @@ static int si_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name); - err = request_firmware(&rdev->ce_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->ce_fw->size != ce_req_size) { @@ -1649,7 +1655,7 @@ static int si_init_microcode(struct radeon_device *rdev) } snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name); - err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev); + err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev); if (err) goto out; if (rdev->rlc_fw->size != rlc_req_size) { @@ -1659,22 +1665,33 @@ static int si_init_microcode(struct radeon_device *rdev) err = -EINVAL; } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); - err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->mc_fw->size != mc_req_size) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); + if (err) + goto out; + } + if ((rdev->mc_fw->size != mc_req_size) && + (rdev->mc_fw->size != mc2_req_size)) { printk(KERN_ERR "si_mc: Bogus length %zu in firmware \"%s\"\n", rdev->mc_fw->size, fw_name); err = -EINVAL; } + DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size); snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name); - err = request_firmware(&rdev->smc_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->smc_fw->size != smc_req_size) { + err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev); + if (err) { + printk(KERN_ERR + "smc: error loading firmware \"%s\"\n", + fw_name); + release_firmware(rdev->smc_fw); + rdev->smc_fw = NULL; + err = 0; + } else if (rdev->smc_fw->size != smc_req_size) { printk(KERN_ERR "si_smc: Bogus length %zu in firmware \"%s\"\n", rdev->smc_fw->size, fw_name); @@ -1682,8 +1699,6 @@ static int si_init_microcode(struct radeon_device *rdev) } out: - platform_device_unregister(pdev); - if (err) { if (err != -EINVAL) printk(KERN_ERR @@ -1711,7 +1726,8 @@ static u32 dce6_line_buffer_adjust(struct radeon_device *rdev, struct drm_display_mode *mode, struct drm_display_mode *other_mode) { - u32 tmp; + u32 tmp, buffer_alloc, i; + u32 pipe_offset = radeon_crtc->crtc_id * 0x20; /* * Line Buffer Setup * There are 3 line buffers, each one shared by 2 display controllers. @@ -1726,16 +1742,30 @@ static u32 dce6_line_buffer_adjust(struct radeon_device *rdev, * non-linked crtcs for maximum line buffer allocation. */ if (radeon_crtc->base.enabled && mode) { - if (other_mode) + if (other_mode) { tmp = 0; /* 1/2 */ - else + buffer_alloc = 1; + } else { tmp = 2; /* whole */ - } else + buffer_alloc = 2; + } + } else { tmp = 0; + buffer_alloc = 0; + } WREG32(DC_LB_MEMORY_SPLIT + radeon_crtc->crtc_offset, DC_LB_MEMORY_CONFIG(tmp)); + WREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset, + DMIF_BUFFERS_ALLOCATED(buffer_alloc)); + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset) & + DMIF_BUFFERS_ALLOCATED_COMPLETED) + break; + udelay(1); + } + if (radeon_crtc->base.enabled && mode) { switch (tmp) { case 0: @@ -2795,7 +2825,7 @@ static void si_setup_spi(struct radeon_device *rdev, } static u32 si_get_rb_disabled(struct radeon_device *rdev, - u32 max_rb_num, u32 se_num, + u32 max_rb_num_per_se, u32 sh_per_se) { u32 data, mask; @@ -2809,14 +2839,14 @@ static u32 si_get_rb_disabled(struct radeon_device *rdev, data >>= BACKEND_DISABLE_SHIFT; - mask = si_create_bitmask(max_rb_num / se_num / sh_per_se); + mask = si_create_bitmask(max_rb_num_per_se / sh_per_se); return data & mask; } static void si_setup_rb(struct radeon_device *rdev, u32 se_num, u32 sh_per_se, - u32 max_rb_num) + u32 max_rb_num_per_se) { int i, j; u32 data, mask; @@ -2826,19 +2856,21 @@ static void si_setup_rb(struct radeon_device *rdev, for (i = 0; i < se_num; i++) { for (j = 0; j < sh_per_se; j++) { si_select_se_sh(rdev, i, j); - data = si_get_rb_disabled(rdev, max_rb_num, se_num, sh_per_se); + data = si_get_rb_disabled(rdev, max_rb_num_per_se, sh_per_se); disabled_rbs |= data << ((i * sh_per_se + j) * TAHITI_RB_BITMAP_WIDTH_PER_SH); } } si_select_se_sh(rdev, 0xffffffff, 0xffffffff); mask = 1; - for (i = 0; i < max_rb_num; i++) { + for (i = 0; i < max_rb_num_per_se * se_num; i++) { if (!(disabled_rbs & mask)) enabled_rbs |= mask; mask <<= 1; } + rdev->config.si.backend_enable_mask = enabled_rbs; + for (i = 0; i < se_num; i++) { si_select_se_sh(rdev, i, 0xffffffff); data = 0; @@ -2869,7 +2901,7 @@ static void si_gpu_init(struct radeon_device *rdev) u32 sx_debug_1; u32 hdp_host_path_cntl; u32 tmp; - int i, j; + int i, j, k; switch (rdev->family) { case CHIP_TAHITI: @@ -3067,6 +3099,14 @@ static void si_gpu_init(struct radeon_device *rdev) rdev->config.si.max_sh_per_se, rdev->config.si.max_cu_per_sh); + for (i = 0; i < rdev->config.si.max_shader_engines; i++) { + for (j = 0; j < rdev->config.si.max_sh_per_se; j++) { + for (k = 0; k < rdev->config.si.max_cu_per_sh; k++) { + rdev->config.si.active_cus += + hweight32(si_get_cu_active_bitmap(rdev, i, j)); + } + } + } /* set HW defaults for 3D engine */ WREG32(CP_QUEUE_THRESHOLDS, (ROQ_IB1_START(0x16) | @@ -3155,7 +3195,7 @@ void si_fence_ring_emit(struct radeon_device *rdev, /* EVENT_WRITE_EOP - flush caches, send int */ radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4)); radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5)); - radeon_ring_write(ring, addr & 0xffffffff); + radeon_ring_write(ring, lower_32_bits(addr)); radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2)); radeon_ring_write(ring, fence->seq); radeon_ring_write(ring, 0); @@ -3188,7 +3228,7 @@ void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); radeon_ring_write(ring, (1 << 8)); radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff); + radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr)); radeon_ring_write(ring, next_rptr); } @@ -3229,7 +3269,8 @@ static void si_cp_enable(struct radeon_device *rdev, bool enable) if (enable) WREG32(CP_ME_CNTL, 0); else { - radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT)); WREG32(SCRATCH_UMSK, 0); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; @@ -3371,16 +3412,7 @@ static int si_cp_resume(struct radeon_device *rdev) u32 rb_bufsz; int r; - /* Reset cp; if cp is reset, then PA, SH, VGT also need to be reset */ - WREG32(GRBM_SOFT_RESET, (SOFT_RESET_CP | - SOFT_RESET_PA | - SOFT_RESET_VGT | - SOFT_RESET_SPI | - SOFT_RESET_SX)); - RREG32(GRBM_SOFT_RESET); - mdelay(15); - WREG32(GRBM_SOFT_RESET, 0); - RREG32(GRBM_SOFT_RESET); + si_enable_gui_idle_interrupt(rdev, false); WREG32(CP_SEM_WAIT_TIMER, 0x0); WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0); @@ -3394,8 +3426,8 @@ static int si_cp_resume(struct radeon_device *rdev) /* ring 0 - compute and gfx */ /* Set ring buffer size */ ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; - rb_bufsz = drm_order(ring->ring_size / 8); - tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; + rb_bufsz = order_base_2(ring->ring_size / 8); + tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif @@ -3422,13 +3454,11 @@ static int si_cp_resume(struct radeon_device *rdev) WREG32(CP_RB0_BASE, ring->gpu_addr >> 8); - ring->rptr = RREG32(CP_RB0_RPTR); - /* ring1 - compute only */ /* Set ring buffer size */ ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; - rb_bufsz = drm_order(ring->ring_size / 8); - tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; + rb_bufsz = order_base_2(ring->ring_size / 8); + tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif @@ -3448,13 +3478,11 @@ static int si_cp_resume(struct radeon_device *rdev) WREG32(CP_RB1_BASE, ring->gpu_addr >> 8); - ring->rptr = RREG32(CP_RB1_RPTR); - /* ring2 - compute only */ /* Set ring buffer size */ ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; - rb_bufsz = drm_order(ring->ring_size / 8); - tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; + rb_bufsz = order_base_2(ring->ring_size / 8); + tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif @@ -3474,8 +3502,6 @@ static int si_cp_resume(struct radeon_device *rdev) WREG32(CP_RB2_BASE, ring->gpu_addr >> 8); - ring->rptr = RREG32(CP_RB2_RPTR); - /* start the rings */ si_cp_start(rdev); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true; @@ -3497,10 +3523,15 @@ static int si_cp_resume(struct radeon_device *rdev) rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false; } + si_enable_gui_idle_interrupt(rdev, true); + + if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX) + radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); + return 0; } -static u32 si_gpu_check_soft_reset(struct radeon_device *rdev) +u32 si_gpu_check_soft_reset(struct radeon_device *rdev) { u32 reset_mask = 0; u32 tmp; @@ -3598,6 +3629,13 @@ static void si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask) dev_info(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS)); + /* disable PG/CG */ + si_fini_pg(rdev); + si_fini_cg(rdev); + + /* stop the rlc */ + si_rlc_stop(rdev); + /* Disable CP parsing/prefetching */ WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT); @@ -3706,6 +3744,106 @@ static void si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask) evergreen_print_gpu_status_regs(rdev); } +static void si_set_clk_bypass_mode(struct radeon_device *rdev) +{ + u32 tmp, i; + + tmp = RREG32(CG_SPLL_FUNC_CNTL); + tmp |= SPLL_BYPASS_EN; + WREG32(CG_SPLL_FUNC_CNTL, tmp); + + tmp = RREG32(CG_SPLL_FUNC_CNTL_2); + tmp |= SPLL_CTLREQ_CHG; + WREG32(CG_SPLL_FUNC_CNTL_2, tmp); + + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(SPLL_STATUS) & SPLL_CHG_STATUS) + break; + udelay(1); + } + + tmp = RREG32(CG_SPLL_FUNC_CNTL_2); + tmp &= ~(SPLL_CTLREQ_CHG | SCLK_MUX_UPDATE); + WREG32(CG_SPLL_FUNC_CNTL_2, tmp); + + tmp = RREG32(MPLL_CNTL_MODE); + tmp &= ~MPLL_MCLK_SEL; + WREG32(MPLL_CNTL_MODE, tmp); +} + +static void si_spll_powerdown(struct radeon_device *rdev) +{ + u32 tmp; + + tmp = RREG32(SPLL_CNTL_MODE); + tmp |= SPLL_SW_DIR_CONTROL; + WREG32(SPLL_CNTL_MODE, tmp); + + tmp = RREG32(CG_SPLL_FUNC_CNTL); + tmp |= SPLL_RESET; + WREG32(CG_SPLL_FUNC_CNTL, tmp); + + tmp = RREG32(CG_SPLL_FUNC_CNTL); + tmp |= SPLL_SLEEP; + WREG32(CG_SPLL_FUNC_CNTL, tmp); + + tmp = RREG32(SPLL_CNTL_MODE); + tmp &= ~SPLL_SW_DIR_CONTROL; + WREG32(SPLL_CNTL_MODE, tmp); +} + +static void si_gpu_pci_config_reset(struct radeon_device *rdev) +{ + struct evergreen_mc_save save; + u32 tmp, i; + + dev_info(rdev->dev, "GPU pci config reset\n"); + + /* disable dpm? */ + + /* disable cg/pg */ + si_fini_pg(rdev); + si_fini_cg(rdev); + + /* Disable CP parsing/prefetching */ + WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT); + /* dma0 */ + tmp = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET); + tmp &= ~DMA_RB_ENABLE; + WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, tmp); + /* dma1 */ + tmp = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET); + tmp &= ~DMA_RB_ENABLE; + WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, tmp); + /* XXX other engines? */ + + /* halt the rlc, disable cp internal ints */ + si_rlc_stop(rdev); + + udelay(50); + + /* disable mem access */ + evergreen_mc_stop(rdev, &save); + if (evergreen_mc_wait_for_idle(rdev)) { + dev_warn(rdev->dev, "Wait for MC idle timed out !\n"); + } + + /* set mclk/sclk to bypass */ + si_set_clk_bypass_mode(rdev); + /* powerdown spll */ + si_spll_powerdown(rdev); + /* disable BM */ + pci_clear_master(rdev->pdev); + /* reset */ + radeon_pci_config_reset(rdev); + /* wait for asic to come out of reset */ + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(CONFIG_MEMSIZE) != 0xffffffff) + break; + udelay(1); + } +} + int si_asic_reset(struct radeon_device *rdev) { u32 reset_mask; @@ -3715,10 +3853,17 @@ int si_asic_reset(struct radeon_device *rdev) if (reset_mask) r600_set_bios_scratch_engine_hung(rdev, true); + /* try soft reset */ si_gpu_soft_reset(rdev, reset_mask); reset_mask = si_gpu_check_soft_reset(rdev); + /* try pci config reset */ + if (reset_mask && radeon_hard_reset) + si_gpu_pci_config_reset(rdev); + + reset_mask = si_gpu_check_soft_reset(rdev); + if (!reset_mask) r600_set_bios_scratch_engine_hung(rdev, false); @@ -3741,39 +3886,9 @@ bool si_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); - return radeon_ring_test_lockup(rdev, ring); -} - -/** - * si_dma_is_lockup - Check if the DMA engine is locked up - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Check if the async DMA engine is locked up. - * Returns true if the engine appears to be locked up, false if not. - */ -bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) -{ - u32 reset_mask = si_gpu_check_soft_reset(rdev); - u32 mask; - - if (ring->idx == R600_RING_TYPE_DMA_INDEX) - mask = RADEON_RESET_DMA; - else - mask = RADEON_RESET_DMA1; - - if (!(reset_mask & mask)) { - radeon_ring_lockup_update(ring); - return false; - } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } @@ -3894,8 +4009,15 @@ static int si_mc_init(struct radeon_device *rdev) rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0); rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0); /* size in MB on si */ - rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024ULL * 1024ULL; - rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024ULL * 1024ULL; + tmp = RREG32(CONFIG_MEMSIZE); + /* some boards may have garbage in the upper 16 bits */ + if (tmp & 0xffff0000) { + DRM_INFO("Probable bad vram size: 0x%08x\n", tmp); + if (tmp & 0xffff) + tmp &= 0xffff; + } + rdev->mc.mc_vram_size = tmp * 1024ULL * 1024ULL; + rdev->mc.real_vram_size = rdev->mc.mc_vram_size; rdev->mc.visible_vram_size = rdev->mc.aper_size; si_vram_gtt_location(rdev, &rdev->mc); radeon_update_bandwidth_info(rdev); @@ -3931,18 +4053,21 @@ static int si_pcie_gart_enable(struct radeon_device *rdev) WREG32(MC_VM_MX_L1_TLB_CNTL, (0xA << 7) | ENABLE_L1_TLB | + ENABLE_L1_FRAGMENT_PROCESSING | SYSTEM_ACCESS_MODE_NOT_IN_SYS | ENABLE_ADVANCED_DRIVER_MODEL | SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU); /* Setup L2 cache */ WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | + ENABLE_L2_FRAGMENT_PROCESSING | ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE | ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE | EFFECTIVE_L2_QUEUE_SIZE(7) | CONTEXT1_IDENTITY_ACCESS_MODE(1)); WREG32(VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS | INVALIDATE_L2_CACHE); WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY | - L2_CACHE_BIGK_FRAGMENT_SIZE(0)); + BANK_SELECT(4) | + L2_CACHE_BIGK_FRAGMENT_SIZE(4)); /* setup context0 */ WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12); WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR, rdev->mc.gtt_end >> 12); @@ -3979,6 +4104,7 @@ static int si_pcie_gart_enable(struct radeon_device *rdev) (u32)(rdev->dummy_page.addr >> 12)); WREG32(VM_CONTEXT1_CNTL2, 4); WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(1) | + PAGE_TABLE_BLOCK_SIZE(radeon_vm_block_size - 9) | RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT | RANGE_PROTECTION_FAULT_ENABLE_DEFAULT | DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT | @@ -4090,13 +4216,64 @@ static int si_vm_packet3_ce_check(struct radeon_device *rdev, return 0; } +static int si_vm_packet3_cp_dma_check(u32 *ib, u32 idx) +{ + u32 start_reg, reg, i; + u32 command = ib[idx + 4]; + u32 info = ib[idx + 1]; + u32 idx_value = ib[idx]; + if (command & PACKET3_CP_DMA_CMD_SAS) { + /* src address space is register */ + if (((info & 0x60000000) >> 29) == 0) { + start_reg = idx_value << 2; + if (command & PACKET3_CP_DMA_CMD_SAIC) { + reg = start_reg; + if (!si_vm_reg_valid(reg)) { + DRM_ERROR("CP DMA Bad SRC register\n"); + return -EINVAL; + } + } else { + for (i = 0; i < (command & 0x1fffff); i++) { + reg = start_reg + (4 * i); + if (!si_vm_reg_valid(reg)) { + DRM_ERROR("CP DMA Bad SRC register\n"); + return -EINVAL; + } + } + } + } + } + if (command & PACKET3_CP_DMA_CMD_DAS) { + /* dst address space is register */ + if (((info & 0x00300000) >> 20) == 0) { + start_reg = ib[idx + 2]; + if (command & PACKET3_CP_DMA_CMD_DAIC) { + reg = start_reg; + if (!si_vm_reg_valid(reg)) { + DRM_ERROR("CP DMA Bad DST register\n"); + return -EINVAL; + } + } else { + for (i = 0; i < (command & 0x1fffff); i++) { + reg = start_reg + (4 * i); + if (!si_vm_reg_valid(reg)) { + DRM_ERROR("CP DMA Bad DST register\n"); + return -EINVAL; + } + } + } + } + } + return 0; +} + static int si_vm_packet3_gfx_check(struct radeon_device *rdev, u32 *ib, struct radeon_cs_packet *pkt) { + int r; u32 idx = pkt->idx + 1; u32 idx_value = ib[idx]; u32 start_reg, end_reg, reg, i; - u32 command, info; switch (pkt->opcode) { case PACKET3_NOP: @@ -4197,50 +4374,9 @@ static int si_vm_packet3_gfx_check(struct radeon_device *rdev, } break; case PACKET3_CP_DMA: - command = ib[idx + 4]; - info = ib[idx + 1]; - if (command & PACKET3_CP_DMA_CMD_SAS) { - /* src address space is register */ - if (((info & 0x60000000) >> 29) == 0) { - start_reg = idx_value << 2; - if (command & PACKET3_CP_DMA_CMD_SAIC) { - reg = start_reg; - if (!si_vm_reg_valid(reg)) { - DRM_ERROR("CP DMA Bad SRC register\n"); - return -EINVAL; - } - } else { - for (i = 0; i < (command & 0x1fffff); i++) { - reg = start_reg + (4 * i); - if (!si_vm_reg_valid(reg)) { - DRM_ERROR("CP DMA Bad SRC register\n"); - return -EINVAL; - } - } - } - } - } - if (command & PACKET3_CP_DMA_CMD_DAS) { - /* dst address space is register */ - if (((info & 0x00300000) >> 20) == 0) { - start_reg = ib[idx + 2]; - if (command & PACKET3_CP_DMA_CMD_DAIC) { - reg = start_reg; - if (!si_vm_reg_valid(reg)) { - DRM_ERROR("CP DMA Bad DST register\n"); - return -EINVAL; - } - } else { - for (i = 0; i < (command & 0x1fffff); i++) { - reg = start_reg + (4 * i); - if (!si_vm_reg_valid(reg)) { - DRM_ERROR("CP DMA Bad DST register\n"); - return -EINVAL; - } - } - } - } - } + r = si_vm_packet3_cp_dma_check(ib, idx); + if (r) + return r; break; default: DRM_ERROR("Invalid GFX packet3: 0x%x\n", pkt->opcode); @@ -4252,6 +4388,7 @@ static int si_vm_packet3_gfx_check(struct radeon_device *rdev, static int si_vm_packet3_compute_check(struct radeon_device *rdev, u32 *ib, struct radeon_cs_packet *pkt) { + int r; u32 idx = pkt->idx + 1; u32 idx_value = ib[idx]; u32 start_reg, reg, i; @@ -4324,6 +4461,11 @@ static int si_vm_packet3_compute_check(struct radeon_device *rdev, return -EINVAL; } break; + case PACKET3_CP_DMA: + r = si_vm_packet3_cp_dma_check(ib, idx); + if (r) + return r; + break; default: DRM_ERROR("Invalid Compute packet3: 0x%x\n", pkt->opcode); return -EINVAL; @@ -4401,109 +4543,267 @@ void si_vm_fini(struct radeon_device *rdev) } /** - * si_vm_set_page - update the page tables using the CP + * si_vm_decode_fault - print human readable fault info * * @rdev: radeon_device pointer - * @ib: indirect buffer to fill with commands - * @pe: addr of the page entry - * @addr: dst addr to write into pe - * @count: number of page entries to update - * @incr: increase next addr by incr bytes - * @flags: access flags + * @status: VM_CONTEXT1_PROTECTION_FAULT_STATUS register value + * @addr: VM_CONTEXT1_PROTECTION_FAULT_ADDR register value * - * Update the page tables using the CP (SI). + * Print human readable fault information (SI). */ -void si_vm_set_page(struct radeon_device *rdev, - struct radeon_ib *ib, - uint64_t pe, - uint64_t addr, unsigned count, - uint32_t incr, uint32_t flags) -{ - uint32_t r600_flags = cayman_vm_page_flags(rdev, flags); - uint64_t value; - unsigned ndw; - - if (rdev->asic->vm.pt_ring_index == RADEON_RING_TYPE_GFX_INDEX) { - while (count) { - ndw = 2 + count * 2; - if (ndw > 0x3FFE) - ndw = 0x3FFE; - - ib->ptr[ib->length_dw++] = PACKET3(PACKET3_WRITE_DATA, ndw); - ib->ptr[ib->length_dw++] = (WRITE_DATA_ENGINE_SEL(0) | - WRITE_DATA_DST_SEL(1)); - ib->ptr[ib->length_dw++] = pe; - ib->ptr[ib->length_dw++] = upper_32_bits(pe); - for (; ndw > 2; ndw -= 2, --count, pe += 8) { - if (flags & RADEON_VM_PAGE_SYSTEM) { - value = radeon_vm_map_gart(rdev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & RADEON_VM_PAGE_VALID) { - value = addr; - } else { - value = 0; - } - addr += incr; - value |= r600_flags; - ib->ptr[ib->length_dw++] = value; - ib->ptr[ib->length_dw++] = upper_32_bits(value); - } +static void si_vm_decode_fault(struct radeon_device *rdev, + u32 status, u32 addr) +{ + u32 mc_id = (status & MEMORY_CLIENT_ID_MASK) >> MEMORY_CLIENT_ID_SHIFT; + u32 vmid = (status & FAULT_VMID_MASK) >> FAULT_VMID_SHIFT; + u32 protections = (status & PROTECTIONS_MASK) >> PROTECTIONS_SHIFT; + char *block; + + if (rdev->family == CHIP_TAHITI) { + switch (mc_id) { + case 160: + case 144: + case 96: + case 80: + case 224: + case 208: + case 32: + case 16: + block = "CB"; + break; + case 161: + case 145: + case 97: + case 81: + case 225: + case 209: + case 33: + case 17: + block = "CB_FMASK"; + break; + case 162: + case 146: + case 98: + case 82: + case 226: + case 210: + case 34: + case 18: + block = "CB_CMASK"; + break; + case 163: + case 147: + case 99: + case 83: + case 227: + case 211: + case 35: + case 19: + block = "CB_IMMED"; + break; + case 164: + case 148: + case 100: + case 84: + case 228: + case 212: + case 36: + case 20: + block = "DB"; + break; + case 165: + case 149: + case 101: + case 85: + case 229: + case 213: + case 37: + case 21: + block = "DB_HTILE"; + break; + case 167: + case 151: + case 103: + case 87: + case 231: + case 215: + case 39: + case 23: + block = "DB_STEN"; + break; + case 72: + case 68: + case 64: + case 8: + case 4: + case 0: + case 136: + case 132: + case 128: + case 200: + case 196: + case 192: + block = "TC"; + break; + case 112: + case 48: + block = "CP"; + break; + case 49: + case 177: + case 50: + case 178: + block = "SH"; + break; + case 53: + case 190: + block = "VGT"; + break; + case 117: + block = "IH"; + break; + case 51: + case 115: + block = "RLC"; + break; + case 119: + case 183: + block = "DMA0"; + break; + case 61: + block = "DMA1"; + break; + case 248: + case 120: + block = "HDP"; + break; + default: + block = "unknown"; + break; } } else { - /* DMA */ - if (flags & RADEON_VM_PAGE_SYSTEM) { - while (count) { - ndw = count * 2; - if (ndw > 0xFFFFE) - ndw = 0xFFFFE; - - /* for non-physically contiguous pages (system) */ - ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw); - ib->ptr[ib->length_dw++] = pe; - ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; - for (; ndw > 0; ndw -= 2, --count, pe += 8) { - if (flags & RADEON_VM_PAGE_SYSTEM) { - value = radeon_vm_map_gart(rdev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & RADEON_VM_PAGE_VALID) { - value = addr; - } else { - value = 0; - } - addr += incr; - value |= r600_flags; - ib->ptr[ib->length_dw++] = value; - ib->ptr[ib->length_dw++] = upper_32_bits(value); - } - } - } else { - while (count) { - ndw = count * 2; - if (ndw > 0xFFFFE) - ndw = 0xFFFFE; - - if (flags & RADEON_VM_PAGE_VALID) - value = addr; - else - value = 0; - /* for physically contiguous pages (vram) */ - ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw); - ib->ptr[ib->length_dw++] = pe; /* dst addr */ - ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; - ib->ptr[ib->length_dw++] = r600_flags; /* mask */ - ib->ptr[ib->length_dw++] = 0; - ib->ptr[ib->length_dw++] = value; /* value */ - ib->ptr[ib->length_dw++] = upper_32_bits(value); - ib->ptr[ib->length_dw++] = incr; /* increment size */ - ib->ptr[ib->length_dw++] = 0; - pe += ndw * 4; - addr += (ndw / 2) * incr; - count -= ndw / 2; - } + switch (mc_id) { + case 32: + case 16: + case 96: + case 80: + case 160: + case 144: + case 224: + case 208: + block = "CB"; + break; + case 33: + case 17: + case 97: + case 81: + case 161: + case 145: + case 225: + case 209: + block = "CB_FMASK"; + break; + case 34: + case 18: + case 98: + case 82: + case 162: + case 146: + case 226: + case 210: + block = "CB_CMASK"; + break; + case 35: + case 19: + case 99: + case 83: + case 163: + case 147: + case 227: + case 211: + block = "CB_IMMED"; + break; + case 36: + case 20: + case 100: + case 84: + case 164: + case 148: + case 228: + case 212: + block = "DB"; + break; + case 37: + case 21: + case 101: + case 85: + case 165: + case 149: + case 229: + case 213: + block = "DB_HTILE"; + break; + case 39: + case 23: + case 103: + case 87: + case 167: + case 151: + case 231: + case 215: + block = "DB_STEN"; + break; + case 72: + case 68: + case 8: + case 4: + case 136: + case 132: + case 200: + case 196: + block = "TC"; + break; + case 112: + case 48: + block = "CP"; + break; + case 49: + case 177: + case 50: + case 178: + block = "SH"; + break; + case 53: + block = "VGT"; + break; + case 117: + block = "IH"; + break; + case 51: + case 115: + block = "RLC"; + break; + case 119: + case 183: + block = "DMA0"; + break; + case 61: + block = "DMA1"; + break; + case 248: + case 120: + block = "HDP"; + break; + default: + block = "unknown"; + break; } - while (ib->length_dw & 0x7) - ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0); } + + printk("VM fault (0x%02x, vmid %d) at page %u, %s from %s (%d)\n", + protections, vmid, addr, + (status & MEMORY_CLIENT_RW_MASK) ? "write" : "read", + block, mc_id); } void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) @@ -4549,32 +4849,6 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) radeon_ring_write(ring, 0x0); } -void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) -{ - struct radeon_ring *ring = &rdev->ring[ridx]; - - if (vm == NULL) - return; - - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0)); - if (vm->id < 8) { - radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2)); - } else { - radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2)); - } - radeon_ring_write(ring, vm->pd_gpu_addr >> 12); - - /* flush hdp cache */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0)); - radeon_ring_write(ring, (0xf << 16) | (HDP_MEM_COHERENCY_FLUSH_CNTL >> 2)); - radeon_ring_write(ring, 1); - - /* bits 0-7 are the VM contexts0-7 */ - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0)); - radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2)); - radeon_ring_write(ring, 1 << vm->id); -} - /* * Power and clock gating */ @@ -4642,7 +4916,7 @@ static void si_set_uvd_dcm(struct radeon_device *rdev, WREG32_UVD_CTX(UVD_CGC_CTRL2, tmp2); } -static void si_init_uvd_internal_cg(struct radeon_device *rdev) +void si_init_uvd_internal_cg(struct radeon_device *rdev) { bool hw_mode = true; @@ -4685,7 +4959,7 @@ static void si_enable_dma_pg(struct radeon_device *rdev, bool enable) u32 data, orig; orig = data = RREG32(DMA_PG); - if (enable) + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_SDMA)) data |= PG_CNTL_ENABLE; else data &= ~PG_CNTL_ENABLE; @@ -4709,7 +4983,7 @@ static void si_enable_gfx_cgpg(struct radeon_device *rdev, { u32 tmp; - if (enable) { + if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_PG)) { tmp = RLC_PUD(0x10) | RLC_PDD(0x10) | RLC_TTPD(0x10) | RLC_MSD(0x10); WREG32(RLC_TTOP_D, tmp); @@ -4812,9 +5086,9 @@ static void si_enable_cgcg(struct radeon_device *rdev, orig = data = RREG32(RLC_CGCG_CGLS_CTRL); - si_enable_gui_idle_interrupt(rdev, enable); + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_CGCG)) { + si_enable_gui_idle_interrupt(rdev, true); - if (enable) { WREG32(RLC_GCPM_GENERAL_3, 0x00000080); tmp = si_halt_rlc(rdev); @@ -4831,6 +5105,8 @@ static void si_enable_cgcg(struct radeon_device *rdev, data |= CGCG_EN | CGLS_EN; } else { + si_enable_gui_idle_interrupt(rdev, false); + RREG32(CB_CGTT_SCLK_CTRL); RREG32(CB_CGTT_SCLK_CTRL); RREG32(CB_CGTT_SCLK_CTRL); @@ -4848,16 +5124,18 @@ static void si_enable_mgcg(struct radeon_device *rdev, { u32 data, orig, tmp = 0; - if (enable) { + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_MGCG)) { orig = data = RREG32(CGTS_SM_CTRL_REG); data = 0x96940200; if (orig != data) WREG32(CGTS_SM_CTRL_REG, data); - orig = data = RREG32(CP_MEM_SLP_CNTL); - data |= CP_MEM_LS_EN; - if (orig != data) - WREG32(CP_MEM_SLP_CNTL, data); + if (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_CP_LS) { + orig = data = RREG32(CP_MEM_SLP_CNTL); + data |= CP_MEM_LS_EN; + if (orig != data) + WREG32(CP_MEM_SLP_CNTL, data); + } orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE); data &= 0xffffffc0; @@ -4902,7 +5180,7 @@ static void si_enable_uvd_mgcg(struct radeon_device *rdev, { u32 orig, data, tmp; - if (enable) { + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_UVD_MGCG)) { tmp = RREG32_UVD_CTX(UVD_CGC_MEM_CTRL); tmp |= 0x3fff; WREG32_UVD_CTX(UVD_CGC_MEM_CTRL, tmp); @@ -4950,7 +5228,7 @@ static void si_enable_mc_ls(struct radeon_device *rdev, for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) { orig = data = RREG32(mc_cg_registers[i]); - if (enable) + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_MC_LS)) data |= MC_LS_ENABLE; else data &= ~MC_LS_ENABLE; @@ -4959,230 +5237,301 @@ static void si_enable_mc_ls(struct radeon_device *rdev, } } - -static void si_init_cg(struct radeon_device *rdev) +static void si_enable_mc_mgcg(struct radeon_device *rdev, + bool enable) { - bool has_uvd = true; + int i; + u32 orig, data; - si_enable_mgcg(rdev, true); - si_enable_cgcg(rdev, true); - /* disable MC LS on Tahiti */ - if (rdev->family == CHIP_TAHITI) - si_enable_mc_ls(rdev, false); - if (has_uvd) { - si_enable_uvd_mgcg(rdev, true); - si_init_uvd_internal_cg(rdev); + for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) { + orig = data = RREG32(mc_cg_registers[i]); + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_MC_MGCG)) + data |= MC_CG_ENABLE; + else + data &= ~MC_CG_ENABLE; + if (data != orig) + WREG32(mc_cg_registers[i], data); } } -static void si_fini_cg(struct radeon_device *rdev) +static void si_enable_dma_mgcg(struct radeon_device *rdev, + bool enable) { - bool has_uvd = true; + u32 orig, data, offset; + int i; - if (has_uvd) - si_enable_uvd_mgcg(rdev, false); - si_enable_cgcg(rdev, false); - si_enable_mgcg(rdev, false); + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_SDMA_MGCG)) { + for (i = 0; i < 2; i++) { + if (i == 0) + offset = DMA0_REGISTER_OFFSET; + else + offset = DMA1_REGISTER_OFFSET; + orig = data = RREG32(DMA_POWER_CNTL + offset); + data &= ~MEM_POWER_OVERRIDE; + if (data != orig) + WREG32(DMA_POWER_CNTL + offset, data); + WREG32(DMA_CLK_CTRL + offset, 0x00000100); + } + } else { + for (i = 0; i < 2; i++) { + if (i == 0) + offset = DMA0_REGISTER_OFFSET; + else + offset = DMA1_REGISTER_OFFSET; + orig = data = RREG32(DMA_POWER_CNTL + offset); + data |= MEM_POWER_OVERRIDE; + if (data != orig) + WREG32(DMA_POWER_CNTL + offset, data); + + orig = data = RREG32(DMA_CLK_CTRL + offset); + data = 0xff000000; + if (data != orig) + WREG32(DMA_CLK_CTRL + offset, data); + } + } } -static void si_init_pg(struct radeon_device *rdev) +static void si_enable_bif_mgls(struct radeon_device *rdev, + bool enable) { - bool has_pg = false; + u32 orig, data; - /* only cape verde supports PG */ - if (rdev->family == CHIP_VERDE) - has_pg = true; + orig = data = RREG32_PCIE(PCIE_CNTL2); - if (has_pg) { - si_init_ao_cu_mask(rdev); - si_init_dma_pg(rdev); - si_enable_dma_pg(rdev, true); - si_init_gfx_cgpg(rdev); - si_enable_gfx_cgpg(rdev, true); - } else { - WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8); - WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8); - } + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_BIF_LS)) + data |= SLV_MEM_LS_EN | MST_MEM_LS_EN | + REPLAY_MEM_LS_EN | SLV_MEM_AGGRESSIVE_LS_EN; + else + data &= ~(SLV_MEM_LS_EN | MST_MEM_LS_EN | + REPLAY_MEM_LS_EN | SLV_MEM_AGGRESSIVE_LS_EN); + + if (orig != data) + WREG32_PCIE(PCIE_CNTL2, data); } -static void si_fini_pg(struct radeon_device *rdev) +static void si_enable_hdp_mgcg(struct radeon_device *rdev, + bool enable) { - bool has_pg = false; + u32 orig, data; - /* only cape verde supports PG */ - if (rdev->family == CHIP_VERDE) - has_pg = true; + orig = data = RREG32(HDP_HOST_PATH_CNTL); - if (has_pg) { - si_enable_dma_pg(rdev, false); - si_enable_gfx_cgpg(rdev, false); - } + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_HDP_MGCG)) + data &= ~CLOCK_GATING_DIS; + else + data |= CLOCK_GATING_DIS; + + if (orig != data) + WREG32(HDP_HOST_PATH_CNTL, data); } -/* - * RLC - */ -void si_rlc_fini(struct radeon_device *rdev) +static void si_enable_hdp_ls(struct radeon_device *rdev, + bool enable) { - int r; - - /* save restore block */ - if (rdev->rlc.save_restore_obj) { - r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false); - if (unlikely(r != 0)) - dev_warn(rdev->dev, "(%d) reserve RLC sr bo failed\n", r); - radeon_bo_unpin(rdev->rlc.save_restore_obj); - radeon_bo_unreserve(rdev->rlc.save_restore_obj); + u32 orig, data; - radeon_bo_unref(&rdev->rlc.save_restore_obj); - rdev->rlc.save_restore_obj = NULL; - } + orig = data = RREG32(HDP_MEM_POWER_LS); - /* clear state block */ - if (rdev->rlc.clear_state_obj) { - r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false); - if (unlikely(r != 0)) - dev_warn(rdev->dev, "(%d) reserve RLC c bo failed\n", r); - radeon_bo_unpin(rdev->rlc.clear_state_obj); - radeon_bo_unreserve(rdev->rlc.clear_state_obj); + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_HDP_LS)) + data |= HDP_LS_ENABLE; + else + data &= ~HDP_LS_ENABLE; - radeon_bo_unref(&rdev->rlc.clear_state_obj); - rdev->rlc.clear_state_obj = NULL; - } + if (orig != data) + WREG32(HDP_MEM_POWER_LS, data); } -#define RLC_CLEAR_STATE_END_MARKER 0x00000001 - -int si_rlc_init(struct radeon_device *rdev) +static void si_update_cg(struct radeon_device *rdev, + u32 block, bool enable) { - volatile u32 *dst_ptr; - u32 dws, data, i, j, k, reg_num; - u32 reg_list_num, reg_list_hdr_blk_index, reg_list_blk_index; - u64 reg_list_mc_addr; - const struct cs_section_def *cs_data = si_cs_data; - int r; - - /* save restore block */ - if (rdev->rlc.save_restore_obj == NULL) { - r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true, - RADEON_GEM_DOMAIN_VRAM, NULL, - &rdev->rlc.save_restore_obj); - if (r) { - dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r); - return r; + if (block & RADEON_CG_BLOCK_GFX) { + si_enable_gui_idle_interrupt(rdev, false); + /* order matters! */ + if (enable) { + si_enable_mgcg(rdev, true); + si_enable_cgcg(rdev, true); + } else { + si_enable_cgcg(rdev, false); + si_enable_mgcg(rdev, false); } + si_enable_gui_idle_interrupt(rdev, true); } - r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false); - if (unlikely(r != 0)) { - si_rlc_fini(rdev); - return r; + if (block & RADEON_CG_BLOCK_MC) { + si_enable_mc_mgcg(rdev, enable); + si_enable_mc_ls(rdev, enable); } - r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM, - &rdev->rlc.save_restore_gpu_addr); - if (r) { - radeon_bo_unreserve(rdev->rlc.save_restore_obj); - dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r); - si_rlc_fini(rdev); - return r; + + if (block & RADEON_CG_BLOCK_SDMA) { + si_enable_dma_mgcg(rdev, enable); } - if (rdev->family == CHIP_VERDE) { - r = radeon_bo_kmap(rdev->rlc.save_restore_obj, (void **)&rdev->rlc.sr_ptr); - if (r) { - dev_warn(rdev->dev, "(%d) map RLC sr bo failed\n", r); - si_rlc_fini(rdev); - return r; - } - /* write the sr buffer */ - dst_ptr = rdev->rlc.sr_ptr; - for (i = 0; i < ARRAY_SIZE(verde_rlc_save_restore_register_list); i++) { - dst_ptr[i] = verde_rlc_save_restore_register_list[i]; - } - radeon_bo_kunmap(rdev->rlc.save_restore_obj); + if (block & RADEON_CG_BLOCK_BIF) { + si_enable_bif_mgls(rdev, enable); } - radeon_bo_unreserve(rdev->rlc.save_restore_obj); - /* clear state block */ - reg_list_num = 0; - dws = 0; - for (i = 0; cs_data[i].section != NULL; i++) { - for (j = 0; cs_data[i].section[j].extent != NULL; j++) { - reg_list_num++; - dws += cs_data[i].section[j].reg_count; + if (block & RADEON_CG_BLOCK_UVD) { + if (rdev->has_uvd) { + si_enable_uvd_mgcg(rdev, enable); } } - reg_list_blk_index = (3 * reg_list_num + 2); - dws += reg_list_blk_index; - if (rdev->rlc.clear_state_obj == NULL) { - r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true, - RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj); - if (r) { - dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r); - si_rlc_fini(rdev); - return r; - } + if (block & RADEON_CG_BLOCK_HDP) { + si_enable_hdp_mgcg(rdev, enable); + si_enable_hdp_ls(rdev, enable); } - r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false); - if (unlikely(r != 0)) { - si_rlc_fini(rdev); - return r; +} + +static void si_init_cg(struct radeon_device *rdev) +{ + si_update_cg(rdev, (RADEON_CG_BLOCK_GFX | + RADEON_CG_BLOCK_MC | + RADEON_CG_BLOCK_SDMA | + RADEON_CG_BLOCK_BIF | + RADEON_CG_BLOCK_HDP), true); + if (rdev->has_uvd) { + si_update_cg(rdev, RADEON_CG_BLOCK_UVD, true); + si_init_uvd_internal_cg(rdev); } - r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM, - &rdev->rlc.clear_state_gpu_addr); - if (r) { +} - radeon_bo_unreserve(rdev->rlc.clear_state_obj); - dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r); - si_rlc_fini(rdev); - return r; +static void si_fini_cg(struct radeon_device *rdev) +{ + if (rdev->has_uvd) { + si_update_cg(rdev, RADEON_CG_BLOCK_UVD, false); } - r = radeon_bo_kmap(rdev->rlc.clear_state_obj, (void **)&rdev->rlc.cs_ptr); - if (r) { - dev_warn(rdev->dev, "(%d) map RLC c bo failed\n", r); - si_rlc_fini(rdev); - return r; + si_update_cg(rdev, (RADEON_CG_BLOCK_GFX | + RADEON_CG_BLOCK_MC | + RADEON_CG_BLOCK_SDMA | + RADEON_CG_BLOCK_BIF | + RADEON_CG_BLOCK_HDP), false); +} + +u32 si_get_csb_size(struct radeon_device *rdev) +{ + u32 count = 0; + const struct cs_section_def *sect = NULL; + const struct cs_extent_def *ext = NULL; + + if (rdev->rlc.cs_data == NULL) + return 0; + + /* begin clear state */ + count += 2; + /* context control state */ + count += 3; + + for (sect = rdev->rlc.cs_data; sect->section != NULL; ++sect) { + for (ext = sect->section; ext->extent != NULL; ++ext) { + if (sect->id == SECT_CONTEXT) + count += 2 + ext->reg_count; + else + return 0; + } } - /* set up the cs buffer */ - dst_ptr = rdev->rlc.cs_ptr; - reg_list_hdr_blk_index = 0; - reg_list_mc_addr = rdev->rlc.clear_state_gpu_addr + (reg_list_blk_index * 4); - data = upper_32_bits(reg_list_mc_addr); - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - for (i = 0; cs_data[i].section != NULL; i++) { - for (j = 0; cs_data[i].section[j].extent != NULL; j++) { - reg_num = cs_data[i].section[j].reg_count; - data = reg_list_mc_addr & 0xffffffff; - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - - data = (cs_data[i].section[j].reg_index * 4) & 0xffffffff; - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - - data = 0x08000000 | (reg_num * 4); - dst_ptr[reg_list_hdr_blk_index] = data; - reg_list_hdr_blk_index++; - - for (k = 0; k < reg_num; k++) { - data = cs_data[i].section[j].extent[k]; - dst_ptr[reg_list_blk_index + k] = data; + /* pa_sc_raster_config */ + count += 3; + /* end clear state */ + count += 2; + /* clear state */ + count += 2; + + return count; +} + +void si_get_csb_buffer(struct radeon_device *rdev, volatile u32 *buffer) +{ + u32 count = 0, i; + const struct cs_section_def *sect = NULL; + const struct cs_extent_def *ext = NULL; + + if (rdev->rlc.cs_data == NULL) + return; + if (buffer == NULL) + return; + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1)); + buffer[count++] = cpu_to_le32(0x80000000); + buffer[count++] = cpu_to_le32(0x80000000); + + for (sect = rdev->rlc.cs_data; sect->section != NULL; ++sect) { + for (ext = sect->section; ext->extent != NULL; ++ext) { + if (sect->id == SECT_CONTEXT) { + buffer[count++] = + cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count)); + buffer[count++] = cpu_to_le32(ext->reg_index - 0xa000); + for (i = 0; i < ext->reg_count; i++) + buffer[count++] = cpu_to_le32(ext->extent[i]); + } else { + return; } - reg_list_mc_addr += reg_num * 4; - reg_list_blk_index += reg_num; } } - dst_ptr[reg_list_hdr_blk_index] = RLC_CLEAR_STATE_END_MARKER; - radeon_bo_kunmap(rdev->rlc.clear_state_obj); - radeon_bo_unreserve(rdev->rlc.clear_state_obj); + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + buffer[count++] = cpu_to_le32(PA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START); + switch (rdev->family) { + case CHIP_TAHITI: + case CHIP_PITCAIRN: + buffer[count++] = cpu_to_le32(0x2a00126a); + break; + case CHIP_VERDE: + buffer[count++] = cpu_to_le32(0x0000124a); + break; + case CHIP_OLAND: + buffer[count++] = cpu_to_le32(0x00000082); + break; + case CHIP_HAINAN: + buffer[count++] = cpu_to_le32(0x00000000); + break; + default: + buffer[count++] = cpu_to_le32(0x00000000); + break; + } + + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); - return 0; + buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0)); + buffer[count++] = cpu_to_le32(0); +} + +static void si_init_pg(struct radeon_device *rdev) +{ + if (rdev->pg_flags) { + if (rdev->pg_flags & RADEON_PG_SUPPORT_SDMA) { + si_init_dma_pg(rdev); + } + si_init_ao_cu_mask(rdev); + if (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_PG) { + si_init_gfx_cgpg(rdev); + } else { + WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8); + WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8); + } + si_enable_dma_pg(rdev, true); + si_enable_gfx_cgpg(rdev, true); + } else { + WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8); + WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8); + } +} + +static void si_fini_pg(struct radeon_device *rdev) +{ + if (rdev->pg_flags) { + si_enable_dma_pg(rdev, false); + si_enable_gfx_cgpg(rdev, false); + } } -static void si_rlc_reset(struct radeon_device *rdev) +/* + * RLC + */ +void si_rlc_reset(struct radeon_device *rdev) { u32 tmp = RREG32(GRBM_SOFT_RESET); @@ -5312,7 +5661,9 @@ static void si_disable_interrupt_state(struct radeon_device *rdev) { u32 tmp; - WREG32(CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + tmp = RREG32(CP_INT_CNTL_RING0) & + (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + WREG32(CP_INT_CNTL_RING0, tmp); WREG32(CP_INT_CNTL_RING1, 0); WREG32(CP_INT_CNTL_RING2, 0); tmp = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET) & ~TRAP_ENABLE; @@ -5347,7 +5698,7 @@ static void si_disable_interrupt_state(struct radeon_device *rdev) } if (!ASIC_IS_NODCE(rdev)) { - WREG32(DACA_AUTODETECT_INT_CONTROL, 0); + WREG32(DAC_AUTODETECT_INT_CONTROL, 0); tmp = RREG32(DC_HPD1_INT_CONTROL) & DC_HPDx_INT_POLARITY; WREG32(DC_HPD1_INT_CONTROL, tmp); @@ -5398,7 +5749,7 @@ static int si_irq_init(struct radeon_device *rdev) WREG32(INTERRUPT_CNTL, interrupt_cntl); WREG32(IH_RB_BASE, rdev->ih.gpu_addr >> 8); - rb_bufsz = drm_order(rdev->ih.ring_size / 4); + rb_bufsz = order_base_2(rdev->ih.ring_size / 4); ih_rb_cntl = (IH_WPTR_OVERFLOW_ENABLE | IH_WPTR_OVERFLOW_CLEAR | @@ -5437,12 +5788,11 @@ static int si_irq_init(struct radeon_device *rdev) int si_irq_set(struct radeon_device *rdev) { - u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE; + u32 cp_int_cntl; u32 cp_int_cntl1 = 0, cp_int_cntl2 = 0; u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0; u32 hpd1 = 0, hpd2 = 0, hpd3 = 0, hpd4 = 0, hpd5 = 0, hpd6 = 0; u32 grbm_int_cntl = 0; - u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0; u32 dma_cntl, dma_cntl1; u32 thermal_int = 0; @@ -5458,6 +5808,9 @@ int si_irq_set(struct radeon_device *rdev) return 0; } + cp_int_cntl = RREG32(CP_INT_CNTL_RING0) & + (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + if (!ASIC_IS_NODCE(rdev)) { hpd1 = RREG32(DC_HPD1_INT_CONTROL) & ~DC_HPDx_INT_EN; hpd2 = RREG32(DC_HPD2_INT_CONTROL) & ~DC_HPDx_INT_EN; @@ -5578,16 +5931,22 @@ int si_irq_set(struct radeon_device *rdev) } if (rdev->num_crtc >= 2) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, grph1); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, grph2); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); } if (rdev->num_crtc >= 4) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, grph3); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, grph4); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); } if (rdev->num_crtc >= 6) { - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, grph5); - WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, grph6); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, + GRPH_PFLIP_INT_MASK); } if (!ASIC_IS_NODCE(rdev)) { @@ -5744,6 +6103,7 @@ static inline u32 si_get_ih_wptr(struct radeon_device *rdev) tmp = RREG32(IH_RB_CNTL); tmp |= IH_WPTR_OVERFLOW_CLEAR; WREG32(IH_RB_CNTL, tmp); + wptr &= ~RB_OVERFLOW; } return (wptr & rdev->ih.ptr_mask); } @@ -5766,6 +6126,7 @@ int si_irq_process(struct radeon_device *rdev) u32 ring_index; bool queue_hotplug = false; bool queue_thermal = false; + u32 status, addr; if (!rdev->ih.enabled || rdev->shutdown) return IRQ_NONE; @@ -5804,7 +6165,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[0])) - radeon_crtc_handle_flip(rdev, 0); + radeon_crtc_handle_vblank(rdev, 0); rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VBLANK_INTERRUPT; DRM_DEBUG("IH: D1 vblank\n"); } @@ -5830,7 +6191,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[1])) - radeon_crtc_handle_flip(rdev, 1); + radeon_crtc_handle_vblank(rdev, 1); rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VBLANK_INTERRUPT; DRM_DEBUG("IH: D2 vblank\n"); } @@ -5856,7 +6217,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[2])) - radeon_crtc_handle_flip(rdev, 2); + radeon_crtc_handle_vblank(rdev, 2); rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VBLANK_INTERRUPT; DRM_DEBUG("IH: D3 vblank\n"); } @@ -5882,7 +6243,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[3])) - radeon_crtc_handle_flip(rdev, 3); + radeon_crtc_handle_vblank(rdev, 3); rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VBLANK_INTERRUPT; DRM_DEBUG("IH: D4 vblank\n"); } @@ -5908,7 +6269,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[4])) - radeon_crtc_handle_flip(rdev, 4); + radeon_crtc_handle_vblank(rdev, 4); rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VBLANK_INTERRUPT; DRM_DEBUG("IH: D5 vblank\n"); } @@ -5934,7 +6295,7 @@ restart_ih: wake_up(&rdev->irq.vblank_queue); } if (atomic_read(&rdev->irq.pflip[5])) - radeon_crtc_handle_flip(rdev, 5); + radeon_crtc_handle_vblank(rdev, 5); rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VBLANK_INTERRUPT; DRM_DEBUG("IH: D6 vblank\n"); } @@ -5950,6 +6311,15 @@ restart_ih: break; } break; + case 8: /* D1 page flip */ + case 10: /* D2 page flip */ + case 12: /* D3 page flip */ + case 14: /* D4 page flip */ + case 16: /* D5 page flip */ + case 18: /* D6 page flip */ + DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1); + radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1); + break; case 42: /* HPD hotplug */ switch (src_data) { case 0: @@ -5999,15 +6369,24 @@ restart_ih: break; } break; + case 124: /* UVD */ + DRM_DEBUG("IH: UVD int: 0x%08x\n", src_data); + radeon_fence_process(rdev, R600_RING_TYPE_UVD_INDEX); + break; case 146: case 147: + addr = RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR); + status = RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS); + /* reset addr and status */ + WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1); + if (addr == 0x0 && status == 0x0) + break; dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data); dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n", - RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR)); + addr); dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", - RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS)); - /* reset addr and status */ - WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1); + status); + si_vm_decode_fault(rdev, status, addr); break; case 176: /* RINGID0 CP_INT */ radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); @@ -6078,80 +6457,6 @@ restart_ih: return IRQ_HANDLED; } -/** - * si_copy_dma - copy pages using the DMA engine - * - * @rdev: radeon_device pointer - * @src_offset: src GPU address - * @dst_offset: dst GPU address - * @num_gpu_pages: number of GPU pages to xfer - * @fence: radeon fence object - * - * Copy GPU paging using the DMA engine (SI). - * Used by the radeon ttm implementation to move pages if - * registered as the asic copy callback. - */ -int si_copy_dma(struct radeon_device *rdev, - uint64_t src_offset, uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence **fence) -{ - struct radeon_semaphore *sem = NULL; - int ring_index = rdev->asic->copy.dma_ring_index; - struct radeon_ring *ring = &rdev->ring[ring_index]; - u32 size_in_bytes, cur_size_in_bytes; - int i, num_loops; - int r = 0; - - r = radeon_semaphore_create(rdev, &sem); - if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - return r; - } - - size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT); - num_loops = DIV_ROUND_UP(size_in_bytes, 0xfffff); - r = radeon_ring_lock(rdev, ring, num_loops * 5 + 11); - if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - radeon_semaphore_free(rdev, &sem, NULL); - return r; - } - - if (radeon_fence_need_sync(*fence, ring->idx)) { - radeon_semaphore_sync_rings(rdev, sem, (*fence)->ring, - ring->idx); - radeon_fence_note_sync(*fence, ring->idx); - } else { - radeon_semaphore_free(rdev, &sem, NULL); - } - - for (i = 0; i < num_loops; i++) { - cur_size_in_bytes = size_in_bytes; - if (cur_size_in_bytes > 0xFFFFF) - cur_size_in_bytes = 0xFFFFF; - size_in_bytes -= cur_size_in_bytes; - radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 1, 0, 0, cur_size_in_bytes)); - radeon_ring_write(ring, dst_offset & 0xffffffff); - radeon_ring_write(ring, src_offset & 0xffffffff); - radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff); - radeon_ring_write(ring, upper_32_bits(src_offset) & 0xff); - src_offset += cur_size_in_bytes; - dst_offset += cur_size_in_bytes; - } - - r = radeon_fence_emit(rdev, fence, ring->idx); - if (r) { - radeon_ring_unlock_undo(rdev, ring); - return r; - } - - radeon_ring_unlock_commit(rdev, ring); - radeon_semaphore_free(rdev, &sem, *fence); - - return r; -} - /* * startup/shutdown callbacks */ @@ -6165,33 +6470,34 @@ static int si_startup(struct radeon_device *rdev) /* enable aspm */ si_program_aspm(rdev); - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw || - !rdev->rlc_fw || !rdev->mc_fw) { - r = si_init_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load firmware!\n"); - return r; - } - } - - r = si_mc_load_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load MC firmware!\n"); - return r; - } - + /* scratch needs to be initialized before MC */ r = r600_vram_scratch_init(rdev); if (r) return r; si_mc_program(rdev); + + if (!rdev->pm.dpm_enabled) { + r = si_mc_load_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load MC firmware!\n"); + return r; + } + } + r = si_pcie_gart_enable(rdev); if (r) return r; si_gpu_init(rdev); /* allocate rlc buffers */ - r = si_rlc_init(rdev); + if (rdev->family == CHIP_VERDE) { + rdev->rlc.reg_list = verde_rlc_save_restore_register_list; + rdev->rlc.reg_list_size = + (u32)ARRAY_SIZE(verde_rlc_save_restore_register_list); + } + rdev->rlc.cs_data = si_cs_data; + r = sumo_rlc_init(rdev); if (r) { DRM_ERROR("Failed to init rlc BOs!\n"); return r; @@ -6233,7 +6539,7 @@ static int si_startup(struct radeon_device *rdev) } if (rdev->has_uvd) { - r = rv770_uvd_resume(rdev); + r = uvd_v2_2_resume(rdev); if (!r) { r = radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_UVD_INDEX); @@ -6261,38 +6567,31 @@ static int si_startup(struct radeon_device *rdev) ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, - CP_RB0_RPTR, CP_RB0_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + RADEON_CP_PACKET2); if (r) return r; ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP1_RPTR_OFFSET, - CP_RB1_RPTR, CP_RB1_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + RADEON_CP_PACKET2); if (r) return r; ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP2_RPTR_OFFSET, - CP_RB2_RPTR, CP_RB2_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + RADEON_CP_PACKET2); if (r) return r; ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET, - DMA_RB_RPTR + DMA0_REGISTER_OFFSET, - DMA_RB_WPTR + DMA0_REGISTER_OFFSET, - 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0)); + DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0)); if (r) return r; ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]; r = radeon_ring_init(rdev, ring, ring->ring_size, CAYMAN_WB_DMA1_RPTR_OFFSET, - DMA_RB_RPTR + DMA1_REGISTER_OFFSET, - DMA_RB_WPTR + DMA1_REGISTER_OFFSET, - 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0)); + DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0)); if (r) return r; @@ -6310,12 +6609,10 @@ static int si_startup(struct radeon_device *rdev) if (rdev->has_uvd) { ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; if (ring->ring_size) { - r = radeon_ring_init(rdev, ring, ring->ring_size, - R600_WB_UVD_RPTR_OFFSET, - UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR, - 0, 0xfffff, RADEON_CP_PACKET2); + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, + RADEON_CP_PACKET2); if (!r) - r = r600_uvd_init(rdev); + r = uvd_v1_0_init(rdev); if (r) DRM_ERROR("radeon: failed initializing UVD (%d).\n", r); } @@ -6333,6 +6630,10 @@ static int si_startup(struct radeon_device *rdev) return r; } + r = dce6_audio_init(rdev); + if (r) + return r; + return 0; } @@ -6350,6 +6651,9 @@ int si_resume(struct radeon_device *rdev) /* init golden registers */ si_init_golden_registers(rdev); + if (rdev->pm.pm_method == PM_METHOD_DPM) + radeon_pm_resume(rdev); + rdev->accel_working = true; r = si_startup(rdev); if (r) { @@ -6364,13 +6668,17 @@ int si_resume(struct radeon_device *rdev) int si_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); + dce6_audio_fini(rdev); radeon_vm_manager_fini(rdev); si_cp_enable(rdev, false); cayman_dma_stop(rdev); if (rdev->has_uvd) { - r600_uvd_rbc_stop(rdev); + uvd_v1_0_fini(rdev); radeon_uvd_suspend(rdev); } + si_fini_pg(rdev); + si_fini_cg(rdev); si_irq_suspend(rdev); radeon_wb_disable(rdev); si_pcie_gart_disable(rdev); @@ -6434,6 +6742,18 @@ int si_init(struct radeon_device *rdev) if (r) return r; + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw || + !rdev->rlc_fw || !rdev->mc_fw) { + r = si_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + + /* Initialize power management */ + radeon_pm_init(rdev); + ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; ring->ring_obj = NULL; r600_ring_init(rdev, ring, 1024 * 1024); @@ -6477,7 +6797,7 @@ int si_init(struct radeon_device *rdev) si_cp_fini(rdev); cayman_dma_fini(rdev); si_irq_fini(rdev); - si_rlc_fini(rdev); + sumo_rlc_fini(rdev); radeon_wb_fini(rdev); radeon_ib_pool_fini(rdev); radeon_vm_manager_fini(rdev); @@ -6500,18 +6820,21 @@ int si_init(struct radeon_device *rdev) void si_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); si_cp_fini(rdev); cayman_dma_fini(rdev); - si_irq_fini(rdev); - si_rlc_fini(rdev); - si_fini_cg(rdev); si_fini_pg(rdev); + si_fini_cg(rdev); + si_irq_fini(rdev); + sumo_rlc_fini(rdev); radeon_wb_fini(rdev); radeon_vm_manager_fini(rdev); radeon_ib_pool_fini(rdev); radeon_irq_kms_fini(rdev); - if (rdev->has_uvd) + if (rdev->has_uvd) { + uvd_v1_0_fini(rdev); radeon_uvd_fini(rdev); + } si_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_gem_fini(rdev); @@ -6796,6 +7119,9 @@ static void si_program_aspm(struct radeon_device *rdev) bool disable_l0s = false, disable_l1 = false, disable_plloff_in_l1 = false; bool disable_clkreq = false; + if (radeon_aspm == 0) + return; + if (!(rdev->flags & RADEON_IS_PCIE)) return; diff --git a/drivers/gpu/drm/radeon/si_dma.c b/drivers/gpu/drm/radeon/si_dma.c new file mode 100644 index 00000000000..e24c94b6d14 --- /dev/null +++ b/drivers/gpu/drm/radeon/si_dma.c @@ -0,0 +1,243 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "radeon_trace.h" +#include "sid.h" + +u32 si_gpu_check_soft_reset(struct radeon_device *rdev); + +/** + * si_dma_is_lockup - Check if the DMA engine is locked up + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * + * Check if the async DMA engine is locked up. + * Returns true if the engine appears to be locked up, false if not. + */ +bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) +{ + u32 reset_mask = si_gpu_check_soft_reset(rdev); + u32 mask; + + if (ring->idx == R600_RING_TYPE_DMA_INDEX) + mask = RADEON_RESET_DMA; + else + mask = RADEON_RESET_DMA1; + + if (!(reset_mask & mask)) { + radeon_ring_lockup_update(rdev, ring); + return false; + } + return radeon_ring_test_lockup(rdev, ring); +} + +/** + * si_dma_vm_set_page - update the page tables using the DMA + * + * @rdev: radeon_device pointer + * @ib: indirect buffer to fill with commands + * @pe: addr of the page entry + * @addr: dst addr to write into pe + * @count: number of page entries to update + * @incr: increase next addr by incr bytes + * @flags: access flags + * + * Update the page tables using the DMA (SI). + */ +void si_dma_vm_set_page(struct radeon_device *rdev, + struct radeon_ib *ib, + uint64_t pe, + uint64_t addr, unsigned count, + uint32_t incr, uint32_t flags) +{ + uint64_t value; + unsigned ndw; + + trace_radeon_vm_set_page(pe, addr, count, incr, flags); + + if (flags == R600_PTE_GART) { + uint64_t src = rdev->gart.table_addr + (addr >> 12) * 8; + while (count) { + unsigned bytes = count * 8; + if (bytes > 0xFFFF8) + bytes = 0xFFFF8; + + ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY, + 1, 0, 0, bytes); + ib->ptr[ib->length_dw++] = lower_32_bits(pe); + ib->ptr[ib->length_dw++] = lower_32_bits(src); + ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; + ib->ptr[ib->length_dw++] = upper_32_bits(src) & 0xff; + + pe += bytes; + src += bytes; + count -= bytes / 8; + } + } else if (flags & R600_PTE_SYSTEM) { + while (count) { + ndw = count * 2; + if (ndw > 0xFFFFE) + ndw = 0xFFFFE; + + /* for non-physically contiguous pages (system) */ + ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw); + ib->ptr[ib->length_dw++] = pe; + ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; + for (; ndw > 0; ndw -= 2, --count, pe += 8) { + value = radeon_vm_map_gart(rdev, addr); + value &= 0xFFFFFFFFFFFFF000ULL; + addr += incr; + value |= flags; + ib->ptr[ib->length_dw++] = value; + ib->ptr[ib->length_dw++] = upper_32_bits(value); + } + } + } else { + while (count) { + ndw = count * 2; + if (ndw > 0xFFFFE) + ndw = 0xFFFFE; + + if (flags & R600_PTE_VALID) + value = addr; + else + value = 0; + /* for physically contiguous pages (vram) */ + ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw); + ib->ptr[ib->length_dw++] = pe; /* dst addr */ + ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff; + ib->ptr[ib->length_dw++] = flags; /* mask */ + ib->ptr[ib->length_dw++] = 0; + ib->ptr[ib->length_dw++] = value; /* value */ + ib->ptr[ib->length_dw++] = upper_32_bits(value); + ib->ptr[ib->length_dw++] = incr; /* increment size */ + ib->ptr[ib->length_dw++] = 0; + pe += ndw * 4; + addr += (ndw / 2) * incr; + count -= ndw / 2; + } + } + while (ib->length_dw & 0x7) + ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0); +} + +void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) +{ + struct radeon_ring *ring = &rdev->ring[ridx]; + + if (vm == NULL) + return; + + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0)); + if (vm->id < 8) { + radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2)); + } else { + radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2)); + } + radeon_ring_write(ring, vm->pd_gpu_addr >> 12); + + /* flush hdp cache */ + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0)); + radeon_ring_write(ring, (0xf << 16) | (HDP_MEM_COHERENCY_FLUSH_CNTL >> 2)); + radeon_ring_write(ring, 1); + + /* bits 0-7 are the VM contexts0-7 */ + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0)); + radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2)); + radeon_ring_write(ring, 1 << vm->id); +} + +/** + * si_copy_dma - copy pages using the DMA engine + * + * @rdev: radeon_device pointer + * @src_offset: src GPU address + * @dst_offset: dst GPU address + * @num_gpu_pages: number of GPU pages to xfer + * @fence: radeon fence object + * + * Copy GPU paging using the DMA engine (SI). + * Used by the radeon ttm implementation to move pages if + * registered as the asic copy callback. + */ +int si_copy_dma(struct radeon_device *rdev, + uint64_t src_offset, uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence **fence) +{ + struct radeon_semaphore *sem = NULL; + int ring_index = rdev->asic->copy.dma_ring_index; + struct radeon_ring *ring = &rdev->ring[ring_index]; + u32 size_in_bytes, cur_size_in_bytes; + int i, num_loops; + int r = 0; + + r = radeon_semaphore_create(rdev, &sem); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + return r; + } + + size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT); + num_loops = DIV_ROUND_UP(size_in_bytes, 0xfffff); + r = radeon_ring_lock(rdev, ring, num_loops * 5 + 11); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_semaphore_sync_to(sem, *fence); + radeon_semaphore_sync_rings(rdev, sem, ring->idx); + + for (i = 0; i < num_loops; i++) { + cur_size_in_bytes = size_in_bytes; + if (cur_size_in_bytes > 0xFFFFF) + cur_size_in_bytes = 0xFFFFF; + size_in_bytes -= cur_size_in_bytes; + radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 1, 0, 0, cur_size_in_bytes)); + radeon_ring_write(ring, lower_32_bits(dst_offset)); + radeon_ring_write(ring, lower_32_bits(src_offset)); + radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff); + radeon_ring_write(ring, upper_32_bits(src_offset) & 0xff); + src_offset += cur_size_in_bytes; + dst_offset += cur_size_in_bytes; + } + + r = radeon_fence_emit(rdev, fence, ring->idx); + if (r) { + radeon_ring_unlock_undo(rdev, ring); + radeon_semaphore_free(rdev, &sem, NULL); + return r; + } + + radeon_ring_unlock_commit(rdev, ring); + radeon_semaphore_free(rdev, &sem, *fence); + + return r; +} + diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c index 73aaa2e4c31..58918868f89 100644 --- a/drivers/gpu/drm/radeon/si_dpm.c +++ b/drivers/gpu/drm/radeon/si_dpm.c @@ -37,8 +37,6 @@ #define SMC_RAM_END 0x20000 -#define DDR3_DRAM_ROWS 0x2000 - #define SCLK_MIN_DEEPSLEEP_FREQ 1350 static const struct si_cac_config_reg cac_weights_tahiti[] = @@ -1740,6 +1738,8 @@ struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev); struct ni_power_info *ni_get_pi(struct radeon_device *rdev); struct ni_ps *ni_get_ps(struct radeon_ps *rps); +extern int si_mc_load_microcode(struct radeon_device *rdev); + static int si_populate_voltage_value(struct radeon_device *rdev, const struct atom_voltage_table *table, u16 value, SISLANDS_SMC_VOLTAGE_VALUE *voltage); @@ -1767,8 +1767,9 @@ static void si_calculate_leakage_for_v_and_t_formula(const struct ni_leakage_coe { s64 kt, kv, leakage_w, i_leakage, vddc; s64 temperature, t_slope, t_intercept, av, bv, t_ref; + s64 tmp; - i_leakage = drm_int2fixp(ileakage / 100); + i_leakage = div64_s64(drm_int2fixp(ileakage), 100); vddc = div64_s64(drm_int2fixp(v), 1000); temperature = div64_s64(drm_int2fixp(t), 1000); @@ -1778,8 +1779,9 @@ static void si_calculate_leakage_for_v_and_t_formula(const struct ni_leakage_coe bv = div64_s64(drm_int2fixp(coeff->bv), 100000000); t_ref = drm_int2fixp(coeff->t_ref); - kt = drm_fixp_div(drm_fixp_exp(drm_fixp_mul(drm_fixp_mul(t_slope, vddc) + t_intercept, temperature)), - drm_fixp_exp(drm_fixp_mul(drm_fixp_mul(t_slope, vddc) + t_intercept, t_ref))); + tmp = drm_fixp_mul(t_slope, vddc) + t_intercept; + kt = drm_fixp_exp(drm_fixp_mul(tmp, temperature)); + kt = drm_fixp_div(kt, drm_fixp_exp(drm_fixp_mul(tmp, t_ref))); kv = drm_fixp_mul(av, drm_fixp_exp(drm_fixp_mul(bv, vddc))); leakage_w = drm_fixp_mul(drm_fixp_mul(drm_fixp_mul(i_leakage, kt), kv), vddc); @@ -1931,6 +1933,7 @@ static void si_initialize_powertune_defaults(struct radeon_device *rdev) si_pi->cac_override = cac_override_pitcairn; si_pi->powertune_data = &powertune_data_pitcairn; si_pi->dte_data = dte_data_pitcairn; + break; } } else if (rdev->family == CHIP_VERDE) { si_pi->lcac_config = lcac_cape_verde; @@ -1941,9 +1944,14 @@ static void si_initialize_powertune_defaults(struct radeon_device *rdev) case 0x683B: case 0x683F: case 0x6829: + case 0x6835: si_pi->cac_weights = cac_weights_cape_verde_pro; si_pi->dte_data = dte_data_cape_verde; break; + case 0x682C: + si_pi->cac_weights = cac_weights_cape_verde_pro; + si_pi->dte_data = dte_data_sun_xt; + break; case 0x6825: case 0x6827: si_pi->cac_weights = cac_weights_heathrow; @@ -1967,10 +1975,9 @@ static void si_initialize_powertune_defaults(struct radeon_device *rdev) si_pi->dte_data = dte_data_venus_xt; break; case 0x6823: - si_pi->cac_weights = cac_weights_chelsea_pro; - si_pi->dte_data = dte_data_venus_pro; - break; case 0x682B: + case 0x6822: + case 0x682A: si_pi->cac_weights = cac_weights_chelsea_pro; si_pi->dte_data = dte_data_venus_pro; break; @@ -1984,6 +1991,7 @@ static void si_initialize_powertune_defaults(struct radeon_device *rdev) case 0x6601: case 0x6621: case 0x6603: + case 0x6605: si_pi->cac_weights = cac_weights_mars_pro; si_pi->lcac_config = lcac_mars_pro; si_pi->cac_override = cac_override_oland; @@ -1994,6 +2002,7 @@ static void si_initialize_powertune_defaults(struct radeon_device *rdev) case 0x6600: case 0x6606: case 0x6620: + case 0x6604: si_pi->cac_weights = cac_weights_mars_xt; si_pi->lcac_config = lcac_mars_pro; si_pi->cac_override = cac_override_oland; @@ -2002,6 +2011,8 @@ static void si_initialize_powertune_defaults(struct radeon_device *rdev) update_dte_from_pl2 = true; break; case 0x6611: + case 0x6613: + case 0x6608: si_pi->cac_weights = cac_weights_oland_pro; si_pi->lcac_config = lcac_mars_pro; si_pi->cac_override = cac_override_oland; @@ -2391,7 +2402,7 @@ static int si_populate_sq_ramping_values(struct radeon_device *rdev, if (SISLANDS_DPM2_SQ_RAMP_STI_SIZE > (STI_SIZE_MASK >> STI_SIZE_SHIFT)) enable_sq_ramping = false; - if (NISLANDS_DPM2_SQ_RAMP_LTI_RATIO <= (LTI_RATIO_MASK >> LTI_RATIO_SHIFT)) + if (SISLANDS_DPM2_SQ_RAMP_LTI_RATIO > (LTI_RATIO_MASK >> LTI_RATIO_SHIFT)) enable_sq_ramping = false; for (i = 0; i < state->performance_level_count; i++) { @@ -2901,16 +2912,21 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev, { struct ni_ps *ps = ni_get_ps(rps); struct radeon_clock_and_voltage_limits *max_limits; - bool disable_mclk_switching; + bool disable_mclk_switching = false; + bool disable_sclk_switching = false; u32 mclk, sclk; u16 vddc, vddci; + u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc; int i; if ((rdev->pm.dpm.new_active_crtc_count > 1) || ni_dpm_vblank_too_short(rdev)) disable_mclk_switching = true; - else - disable_mclk_switching = false; + + if (rps->vclk || rps->dclk) { + disable_mclk_switching = true; + disable_sclk_switching = true; + } if (rdev->pm.dpm.ac_power) max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; @@ -2934,31 +2950,70 @@ static void si_apply_state_adjust_rules(struct radeon_device *rdev, } } + /* limit clocks to max supported clocks based on voltage dependency tables */ + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk, + &max_sclk_vddc); + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk, + &max_mclk_vddci); + btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk, + &max_mclk_vddc); + + for (i = 0; i < ps->performance_level_count; i++) { + if (max_sclk_vddc) { + if (ps->performance_levels[i].sclk > max_sclk_vddc) + ps->performance_levels[i].sclk = max_sclk_vddc; + } + if (max_mclk_vddci) { + if (ps->performance_levels[i].mclk > max_mclk_vddci) + ps->performance_levels[i].mclk = max_mclk_vddci; + } + if (max_mclk_vddc) { + if (ps->performance_levels[i].mclk > max_mclk_vddc) + ps->performance_levels[i].mclk = max_mclk_vddc; + } + } + /* XXX validate the min clocks required for display */ if (disable_mclk_switching) { mclk = ps->performance_levels[ps->performance_level_count - 1].mclk; - sclk = ps->performance_levels[0].sclk; - vddc = ps->performance_levels[0].vddc; vddci = ps->performance_levels[ps->performance_level_count - 1].vddci; } else { - sclk = ps->performance_levels[0].sclk; mclk = ps->performance_levels[0].mclk; - vddc = ps->performance_levels[0].vddc; vddci = ps->performance_levels[0].vddci; } + if (disable_sclk_switching) { + sclk = ps->performance_levels[ps->performance_level_count - 1].sclk; + vddc = ps->performance_levels[ps->performance_level_count - 1].vddc; + } else { + sclk = ps->performance_levels[0].sclk; + vddc = ps->performance_levels[0].vddc; + } + /* adjusted low state */ ps->performance_levels[0].sclk = sclk; ps->performance_levels[0].mclk = mclk; ps->performance_levels[0].vddc = vddc; ps->performance_levels[0].vddci = vddci; - for (i = 1; i < ps->performance_level_count; i++) { - if (ps->performance_levels[i].sclk < ps->performance_levels[i - 1].sclk) - ps->performance_levels[i].sclk = ps->performance_levels[i - 1].sclk; - if (ps->performance_levels[i].vddc < ps->performance_levels[i - 1].vddc) - ps->performance_levels[i].vddc = ps->performance_levels[i - 1].vddc; + if (disable_sclk_switching) { + sclk = ps->performance_levels[0].sclk; + for (i = 1; i < ps->performance_level_count; i++) { + if (sclk < ps->performance_levels[i].sclk) + sclk = ps->performance_levels[i].sclk; + } + for (i = 0; i < ps->performance_level_count; i++) { + ps->performance_levels[i].sclk = sclk; + ps->performance_levels[i].vddc = vddc; + } + } else { + for (i = 1; i < ps->performance_level_count; i++) { + if (ps->performance_levels[i].sclk < ps->performance_levels[i - 1].sclk) + ps->performance_levels[i].sclk = ps->performance_levels[i - 1].sclk; + if (ps->performance_levels[i].vddc < ps->performance_levels[i - 1].vddc) + ps->performance_levels[i].vddc = ps->performance_levels[i - 1].vddc; + } } if (disable_mclk_switching) { @@ -3237,10 +3292,10 @@ int si_dpm_force_performance_level(struct radeon_device *rdev, { struct radeon_ps *rps = rdev->pm.dpm.current_ps; struct ni_ps *ps = ni_get_ps(rps); - u32 levels; + u32 levels = ps->performance_level_count; if (level == RADEON_DPM_FORCED_LEVEL_HIGH) { - if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK) + if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, levels) != PPSMC_Result_OK) return -EINVAL; if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 1) != PPSMC_Result_OK) @@ -3249,14 +3304,13 @@ int si_dpm_force_performance_level(struct radeon_device *rdev, if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK) return -EINVAL; - levels = ps->performance_level_count - 1; - if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, levels) != PPSMC_Result_OK) + if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 1) != PPSMC_Result_OK) return -EINVAL; } else if (level == RADEON_DPM_FORCED_LEVEL_AUTO) { if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK) return -EINVAL; - if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK) + if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, levels) != PPSMC_Result_OK) return -EINVAL; } @@ -3541,6 +3595,10 @@ static void si_program_display_gap(struct radeon_device *rdev) WREG32(DCCG_DISP_SLOW_SELECT_REG, tmp); } + /* Setting this to false forces the performance state to low if the crtcs are disabled. + * This can be a problem on PowerXpress systems or if you want to use the card + * for offscreen rendering or compute if there are no crtcs enabled. + */ si_notify_smc_display_change(rdev, rdev->pm.dpm.new_active_crtc_count > 0); } @@ -3620,8 +3678,12 @@ static void si_enable_display_gap(struct radeon_device *rdev) { u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL); + tmp &= ~(DISP1_GAP_MASK | DISP2_GAP_MASK); + tmp |= (DISP1_GAP(R600_PM_DISPLAY_GAP_IGNORE) | + DISP2_GAP(R600_PM_DISPLAY_GAP_IGNORE)); + tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK); - tmp |= (DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE) | + tmp |= (DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK) | DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE)); WREG32(CG_DISPLAY_GAP_CNTL, tmp); } @@ -3638,7 +3700,7 @@ static void si_clear_vc(struct radeon_device *rdev) WREG32(CG_FTV, 0); } -static u8 si_get_ddr3_mclk_frequency_ratio(u32 memory_clock) +u8 si_get_ddr3_mclk_frequency_ratio(u32 memory_clock) { u8 mc_para_index; @@ -3651,7 +3713,7 @@ static u8 si_get_ddr3_mclk_frequency_ratio(u32 memory_clock) return mc_para_index; } -static u8 si_get_mclk_frequency_ratio(u32 memory_clock, bool strobe_mode) +u8 si_get_mclk_frequency_ratio(u32 memory_clock, bool strobe_mode) { u8 mc_para_index; @@ -3733,20 +3795,21 @@ static bool si_validate_phase_shedding_tables(struct radeon_device *rdev, return true; } -static void si_trim_voltage_table_to_fit_state_table(struct radeon_device *rdev, - struct atom_voltage_table *voltage_table) +void si_trim_voltage_table_to_fit_state_table(struct radeon_device *rdev, + u32 max_voltage_steps, + struct atom_voltage_table *voltage_table) { unsigned int i, diff; - if (voltage_table->count <= SISLANDS_MAX_NO_VREG_STEPS) + if (voltage_table->count <= max_voltage_steps) return; - diff = voltage_table->count - SISLANDS_MAX_NO_VREG_STEPS; + diff = voltage_table->count - max_voltage_steps; - for (i= 0; i < SISLANDS_MAX_NO_VREG_STEPS; i++) + for (i= 0; i < max_voltage_steps; i++) voltage_table->entries[i] = voltage_table->entries[i + diff]; - voltage_table->count = SISLANDS_MAX_NO_VREG_STEPS; + voltage_table->count = max_voltage_steps; } static int si_construct_voltage_tables(struct radeon_device *rdev) @@ -3762,7 +3825,9 @@ static int si_construct_voltage_tables(struct radeon_device *rdev) return ret; if (eg_pi->vddc_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS) - si_trim_voltage_table_to_fit_state_table(rdev, &eg_pi->vddc_voltage_table); + si_trim_voltage_table_to_fit_state_table(rdev, + SISLANDS_MAX_NO_VREG_STEPS, + &eg_pi->vddc_voltage_table); if (eg_pi->vddci_control) { ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDCI, @@ -3771,7 +3836,9 @@ static int si_construct_voltage_tables(struct radeon_device *rdev) return ret; if (eg_pi->vddci_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS) - si_trim_voltage_table_to_fit_state_table(rdev, &eg_pi->vddci_voltage_table); + si_trim_voltage_table_to_fit_state_table(rdev, + SISLANDS_MAX_NO_VREG_STEPS, + &eg_pi->vddci_voltage_table); } if (pi->mvdd_control) { @@ -3789,7 +3856,9 @@ static int si_construct_voltage_tables(struct radeon_device *rdev) } if (si_pi->mvdd_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS) - si_trim_voltage_table_to_fit_state_table(rdev, &si_pi->mvdd_voltage_table); + si_trim_voltage_table_to_fit_state_table(rdev, + SISLANDS_MAX_NO_VREG_STEPS, + &si_pi->mvdd_voltage_table); } if (si_pi->vddc_phase_shed_control) { @@ -4036,16 +4105,15 @@ static int si_force_switch_to_arb_f0(struct radeon_device *rdev) static u32 si_calculate_memory_refresh_rate(struct radeon_device *rdev, u32 engine_clock) { - struct rv7xx_power_info *pi = rv770_get_pi(rdev); u32 dram_rows; u32 dram_refresh_rate; u32 mc_arb_rfsh_rate; u32 tmp = (RREG32(MC_ARB_RAMCFG) & NOOFROWS_MASK) >> NOOFROWS_SHIFT; - if (pi->mem_gddr5) - dram_rows = 1 << (tmp + 10); + if (tmp >= 4) + dram_rows = 16384; else - dram_rows = DDR3_DRAM_ROWS; + dram_rows = 1 << (tmp + 10); dram_refresh_rate = 1 << ((RREG32(MC_SEQ_MISC0) & 0x3) + 3); mc_arb_rfsh_rate = ((engine_clock * 10) * dram_refresh_rate / dram_rows - 32) / 64; @@ -4495,7 +4563,7 @@ static int si_init_smc_table(struct radeon_device *rdev) table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5; if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REVERT_GPIO5_POLARITY) - table->systemFlags |= PPSMC_EXTRAFLAGS_AC2DC_GPIO5_POLARITY_HIGH; + table->extraFlags |= PPSMC_EXTRAFLAGS_AC2DC_GPIO5_POLARITY_HIGH; if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_VRHOT_GPIO_CONFIGURABLE) { table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT_PROG_GPIO; @@ -5150,7 +5218,7 @@ static int si_set_mc_special_registers(struct radeon_device *rdev, table->mc_reg_table_entry[k].mc_data[j] |= 0x100; } j++; - if (j > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) + if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) return -EINVAL; if (!pi->mem_gddr5) { @@ -5160,7 +5228,7 @@ static int si_set_mc_special_registers(struct radeon_device *rdev, table->mc_reg_table_entry[k].mc_data[j] = (table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16; j++; - if (j > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) + if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) return -EINVAL; } break; @@ -5173,7 +5241,7 @@ static int si_set_mc_special_registers(struct radeon_device *rdev, (temp_reg & 0xffff0000) | (table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff); j++; - if (j > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) + if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) return -EINVAL; break; default: @@ -5351,7 +5419,7 @@ static void si_populate_mc_reg_addresses(struct radeon_device *rdev, for (i = 0, j = 0; j < si_pi->mc_reg_table.last; j++) { if (si_pi->mc_reg_table.valid_flag & (1 << j)) { - if (i >= SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE) + if (i >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) break; mc_reg_table->address[i].s0 = cpu_to_be16(si_pi->mc_reg_table.mc_reg_address[j].s0); @@ -5691,6 +5759,11 @@ static void si_set_pcie_lane_width_in_smc(struct radeon_device *rdev, void si_dpm_setup_asic(struct radeon_device *rdev) { + int r; + + r = si_mc_load_microcode(rdev); + if (r) + DRM_ERROR("Failed to load MC firmware!\n"); rv770_get_memory_type(rdev); si_read_clock_registers(rdev); si_enable_acpi_power_management(rdev); @@ -5830,6 +5903,17 @@ int si_dpm_enable(struct radeon_device *rdev) si_enable_sclk_control(rdev, true); si_start_dpm(rdev); + si_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true); + + ni_update_current_ps(rdev, boot_ps); + + return 0; +} + +int si_dpm_late_enable(struct radeon_device *rdev) +{ + int ret; + if (rdev->irq.installed && r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { PPSMC_Result result; @@ -5845,10 +5929,6 @@ int si_dpm_enable(struct radeon_device *rdev) DRM_DEBUG_KMS("Could not enable thermal interrupts.\n"); } - si_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true); - - ni_update_current_ps(rdev, boot_ps); - return 0; } @@ -6013,17 +6093,6 @@ int si_dpm_set_power_state(struct radeon_device *rdev) return ret; } -#if 0 - /* XXX */ - ret = si_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO); - if (ret) { - DRM_ERROR("si_dpm_force_performance_level failed\n"); - return ret; - } -#else - rdev->pm.dpm.forced_level = RADEON_DPM_FORCED_LEVEL_AUTO; -#endif - return 0; } @@ -6209,10 +6278,8 @@ static int si_parse_power_table(struct radeon_device *rdev) if (!rdev->pm.dpm.ps) return -ENOMEM; power_state_offset = (u8 *)state_array->states; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < state_array->ucNumEntries; i++) { + u8 *idx; power_state = (union pplib_power_state *)power_state_offset; non_clock_array_index = power_state->v2.nonClockInfoIndex; non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *) @@ -6229,14 +6296,16 @@ static int si_parse_power_table(struct radeon_device *rdev) non_clock_info, non_clock_info_array->ucEntrySize); k = 0; + idx = (u8 *)&power_state->v2.clockInfoIndex[0]; for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) { - clock_array_index = power_state->v2.clockInfoIndex[j]; + clock_array_index = idx[j]; if (clock_array_index >= clock_info_array->ucNumEntries) continue; if (k >= SISLANDS_MAX_HARDWARE_POWERLEVELS) break; clock_info = (union pplib_clock_info *) - &clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize]; + ((u8 *)&clock_info_array->clockInfo[0] + + (clock_array_index * clock_info_array->ucEntrySize)); si_parse_pplib_clock_info(rdev, &rdev->pm.dpm.ps[i], k, clock_info); @@ -6254,9 +6323,6 @@ int si_dpm_init(struct radeon_device *rdev) struct evergreen_power_info *eg_pi; struct ni_power_info *ni_pi; struct si_power_info *si_pi; - int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info); - u16 data_offset, size; - u8 frev, crev; struct atom_clock_dividers dividers; int ret; u32 mask; @@ -6288,6 +6354,10 @@ int si_dpm_init(struct radeon_device *rdev) pi->min_vddc_in_table = 0; pi->max_vddc_in_table = 0; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = si_parse_power_table(rdev); if (ret) return ret; @@ -6347,16 +6417,7 @@ int si_dpm_init(struct radeon_device *rdev) si_pi->vddc_phase_shed_control = radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, VOLTAGE_OBJ_PHASE_LUT); - if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size, - &frev, &crev, &data_offset)) { - pi->sclk_ss = true; - pi->mclk_ss = true; - pi->dynamic_ss = true; - } else { - pi->sclk_ss = false; - pi->mclk_ss = false; - pi->dynamic_ss = true; - } + rv770_get_engine_memory_ss(rdev); pi->asi = RV770_ASI_DFLT; pi->pasi = CYPRESS_HASI_DFLT; @@ -6367,8 +6428,7 @@ int si_dpm_init(struct radeon_device *rdev) eg_pi->sclk_deep_sleep = true; si_pi->sclk_deep_sleep_above_low = false; - if (pi->gfx_clock_gating && - (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE)) + if (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE) pi->thermal_protection = true; else pi->thermal_protection = false; @@ -6395,6 +6455,12 @@ int si_dpm_init(struct radeon_device *rdev) si_initialize_powertune_defaults(rdev); + /* make sure dc limits are valid */ + if ((rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.sclk == 0) || + (rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.mclk == 0)) + rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc = + rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + return 0; } @@ -6414,7 +6480,8 @@ void si_dpm_fini(struct radeon_device *rdev) void si_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, struct seq_file *m) { - struct radeon_ps *rps = rdev->pm.dpm.current_ps; + struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev); + struct radeon_ps *rps = &eg_pi->current_rps; struct ni_ps *ps = ni_get_ps(rps); struct rv7xx_pl *pl; u32 current_index = diff --git a/drivers/gpu/drm/radeon/si_smc.c b/drivers/gpu/drm/radeon/si_smc.c index 5f524c0a541..e80efcf0c23 100644 --- a/drivers/gpu/drm/radeon/si_smc.c +++ b/drivers/gpu/drm/radeon/si_smc.c @@ -28,9 +28,10 @@ #include "sid.h" #include "ppsmc.h" #include "radeon_ucode.h" +#include "sislands_smc.h" -int si_set_smc_sram_address(struct radeon_device *rdev, - u32 smc_address, u32 limit) +static int si_set_smc_sram_address(struct radeon_device *rdev, + u32 smc_address, u32 limit) { if (smc_address & 3) return -EINVAL; @@ -47,7 +48,8 @@ int si_copy_bytes_to_smc(struct radeon_device *rdev, u32 smc_start_address, const u8 *src, u32 byte_count, u32 limit) { - int ret; + unsigned long flags; + int ret = 0; u32 data, original_data, addr, extra_shift; if (smc_start_address & 3) @@ -57,13 +59,14 @@ int si_copy_bytes_to_smc(struct radeon_device *rdev, addr = smc_start_address; + spin_lock_irqsave(&rdev->smc_idx_lock, flags); while (byte_count >= 4) { /* SMC address space is BE */ data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3]; ret = si_set_smc_sram_address(rdev, addr, limit); if (ret) - return ret; + goto done; WREG32(SMC_IND_DATA_0, data); @@ -78,7 +81,7 @@ int si_copy_bytes_to_smc(struct radeon_device *rdev, ret = si_set_smc_sram_address(rdev, addr, limit); if (ret) - return ret; + goto done; original_data = RREG32(SMC_IND_DATA_0); @@ -96,11 +99,15 @@ int si_copy_bytes_to_smc(struct radeon_device *rdev, ret = si_set_smc_sram_address(rdev, addr, limit); if (ret) - return ret; + goto done; WREG32(SMC_IND_DATA_0, data); } - return 0; + +done: + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); + + return ret; } void si_start_smc(struct radeon_device *rdev) @@ -203,6 +210,7 @@ PPSMC_Result si_wait_for_smc_inactive(struct radeon_device *rdev) int si_load_smc_ucode(struct radeon_device *rdev, u32 limit) { + unsigned long flags; u32 ucode_start_address; u32 ucode_size; const u8 *src; @@ -241,6 +249,7 @@ int si_load_smc_ucode(struct radeon_device *rdev, u32 limit) return -EINVAL; src = (const u8 *)rdev->smc_fw->data; + spin_lock_irqsave(&rdev->smc_idx_lock, flags); WREG32(SMC_IND_INDEX_0, ucode_start_address); WREG32_P(SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, ~AUTO_INCREMENT_IND_0); while (ucode_size >= 4) { @@ -253,6 +262,7 @@ int si_load_smc_ucode(struct radeon_device *rdev, u32 limit) ucode_size -= 4; } WREG32_P(SMC_IND_ACCESS_CNTL, 0, ~AUTO_INCREMENT_IND_0); + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); return 0; } @@ -260,25 +270,29 @@ int si_load_smc_ucode(struct radeon_device *rdev, u32 limit) int si_read_smc_sram_dword(struct radeon_device *rdev, u32 smc_address, u32 *value, u32 limit) { + unsigned long flags; int ret; + spin_lock_irqsave(&rdev->smc_idx_lock, flags); ret = si_set_smc_sram_address(rdev, smc_address, limit); - if (ret) - return ret; + if (ret == 0) + *value = RREG32(SMC_IND_DATA_0); + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); - *value = RREG32(SMC_IND_DATA_0); - return 0; + return ret; } int si_write_smc_sram_dword(struct radeon_device *rdev, u32 smc_address, u32 value, u32 limit) { + unsigned long flags; int ret; + spin_lock_irqsave(&rdev->smc_idx_lock, flags); ret = si_set_smc_sram_address(rdev, smc_address, limit); - if (ret) - return ret; + if (ret == 0) + WREG32(SMC_IND_DATA_0, value); + spin_unlock_irqrestore(&rdev->smc_idx_lock, flags); - WREG32(SMC_IND_DATA_0, value); - return 0; + return ret; } diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h index 12a20eb77d0..fd414d34d88 100644 --- a/drivers/gpu/drm/radeon/sid.h +++ b/drivers/gpu/drm/radeon/sid.h @@ -94,6 +94,8 @@ #define CG_SPLL_FUNC_CNTL_2 0x604 #define SCLK_MUX_SEL(x) ((x) << 0) #define SCLK_MUX_SEL_MASK (0x1ff << 0) +#define SPLL_CTLREQ_CHG (1 << 23) +#define SCLK_MUX_UPDATE (1 << 26) #define CG_SPLL_FUNC_CNTL_3 0x608 #define SPLL_FB_DIV(x) ((x) << 0) #define SPLL_FB_DIV_MASK (0x3ffffff << 0) @@ -101,9 +103,12 @@ #define SPLL_DITHEN (1 << 28) #define CG_SPLL_FUNC_CNTL_4 0x60c +#define SPLL_STATUS 0x614 +#define SPLL_CHG_STATUS (1 << 1) #define SPLL_CNTL_MODE 0x618 -# define SPLL_REFCLK_SEL(x) ((x) << 8) -# define SPLL_REFCLK_SEL_MASK 0xFF00 +#define SPLL_SW_DIR_CONTROL (1 << 0) +# define SPLL_REFCLK_SEL(x) ((x) << 26) +# define SPLL_REFCLK_SEL_MASK (3 << 26) #define CG_SPLL_SPREAD_SPECTRUM 0x620 #define SSEN (1 << 0) @@ -282,6 +287,10 @@ #define DMIF_ADDR_CALC 0xC00 +#define PIPE0_DMIF_BUFFER_CONTROL 0x0ca0 +# define DMIF_BUFFERS_ALLOCATED(x) ((x) << 0) +# define DMIF_BUFFERS_ALLOCATED_COMPLETED (1 << 4) + #define SRBM_STATUS 0xE50 #define GRBM_RQ_PENDING (1 << 5) #define VMC_BUSY (1 << 8) @@ -353,6 +362,7 @@ #define READ_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 16) #define WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT (1 << 18) #define WRITE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 19) +#define PAGE_TABLE_BLOCK_SIZE(x) (((x) & 0xF) << 24) #define VM_CONTEXT1_CNTL 0x1414 #define VM_CONTEXT0_CNTL2 0x1430 #define VM_CONTEXT1_CNTL2 0x1434 @@ -367,6 +377,20 @@ #define VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x14FC #define VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x14DC +#define PROTECTIONS_MASK (0xf << 0) +#define PROTECTIONS_SHIFT 0 + /* bit 0: range + * bit 1: pde0 + * bit 2: valid + * bit 3: read + * bit 4: write + */ +#define MEMORY_CLIENT_ID_MASK (0xff << 12) +#define MEMORY_CLIENT_ID_SHIFT 12 +#define MEMORY_CLIENT_RW_MASK (1 << 24) +#define MEMORY_CLIENT_RW_SHIFT 24 +#define FAULT_VMID_MASK (0xf << 25) +#define FAULT_VMID_SHIFT 25 #define VM_INVALIDATE_REQUEST 0x1478 #define VM_INVALIDATE_RESPONSE 0x147c @@ -460,7 +484,7 @@ #define STATE3_MASK (0x1f << 15) #define STATE3_SHIFT 15 -#define MC_SEQ_TRAIN_WAKEUP_CNTL 0x2808 +#define MC_SEQ_TRAIN_WAKEUP_CNTL 0x28e8 #define TRAIN_DONE_D0 (1 << 30) #define TRAIN_DONE_D1 (1 << 31) @@ -541,6 +565,8 @@ # define MRDCK0_BYPASS (1 << 24) # define MRDCK1_BYPASS (1 << 25) +#define MPLL_CNTL_MODE 0x2bb0 +# define MPLL_MCLK_SEL (1 << 11) #define MPLL_FUNC_CNTL 0x2bb4 #define BWCTRL(x) ((x) << 20) #define BWCTRL_MASK (0xff << 20) @@ -567,6 +593,7 @@ #define CLKS_MASK (0xfff << 0) #define HDP_HOST_PATH_CNTL 0x2C00 +#define CLOCK_GATING_DIS (1 << 23) #define HDP_NONSURFACE_BASE 0x2C04 #define HDP_NONSURFACE_INFO 0x2C08 #define HDP_NONSURFACE_SIZE 0x2C0C @@ -574,6 +601,8 @@ #define HDP_ADDR_CONFIG 0x2F48 #define HDP_MISC_CNTL 0x2F4C #define HDP_FLUSH_INVALIDATE_CACHE (1 << 0) +#define HDP_MEM_POWER_LS 0x2F50 +#define HDP_LS_ENABLE (1 << 0) #define ATC_MISC_CG 0x3350 @@ -621,6 +650,99 @@ #define HDP_REG_COHERENCY_FLUSH_CNTL 0x54A0 +/* DCE6 ELD audio interface */ +#define AZ_F0_CODEC_ENDPOINT_INDEX 0x5E00 +# define AZ_ENDPOINT_REG_INDEX(x) (((x) & 0xff) << 0) +# define AZ_ENDPOINT_REG_WRITE_EN (1 << 8) +#define AZ_F0_CODEC_ENDPOINT_DATA 0x5E04 + +#define AZ_F0_CODEC_PIN_CONTROL_CHANNEL_SPEAKER 0x25 +#define SPEAKER_ALLOCATION(x) (((x) & 0x7f) << 0) +#define SPEAKER_ALLOCATION_MASK (0x7f << 0) +#define SPEAKER_ALLOCATION_SHIFT 0 +#define HDMI_CONNECTION (1 << 16) +#define DP_CONNECTION (1 << 17) + +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0 0x28 /* LPCM */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR1 0x29 /* AC3 */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR2 0x2A /* MPEG1 */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR3 0x2B /* MP3 */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR4 0x2C /* MPEG2 */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR5 0x2D /* AAC */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR6 0x2E /* DTS */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR7 0x2F /* ATRAC */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR8 0x30 /* one bit audio - leave at 0 (default) */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR9 0x31 /* Dolby Digital */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR10 0x32 /* DTS-HD */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR11 0x33 /* MAT-MLP */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR12 0x34 /* DTS */ +#define AZ_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR13 0x35 /* WMA Pro */ +# define MAX_CHANNELS(x) (((x) & 0x7) << 0) +/* max channels minus one. 7 = 8 channels */ +# define SUPPORTED_FREQUENCIES(x) (((x) & 0xff) << 8) +# define DESCRIPTOR_BYTE_2(x) (((x) & 0xff) << 16) +# define SUPPORTED_FREQUENCIES_STEREO(x) (((x) & 0xff) << 24) /* LPCM only */ +/* SUPPORTED_FREQUENCIES, SUPPORTED_FREQUENCIES_STEREO + * bit0 = 32 kHz + * bit1 = 44.1 kHz + * bit2 = 48 kHz + * bit3 = 88.2 kHz + * bit4 = 96 kHz + * bit5 = 176.4 kHz + * bit6 = 192 kHz + */ + +#define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC 0x37 +# define VIDEO_LIPSYNC(x) (((x) & 0xff) << 0) +# define AUDIO_LIPSYNC(x) (((x) & 0xff) << 8) +/* VIDEO_LIPSYNC, AUDIO_LIPSYNC + * 0 = invalid + * x = legal delay value + * 255 = sync not supported + */ +#define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_HBR 0x38 +# define HBR_CAPABLE (1 << 0) /* enabled by default */ + +#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO0 0x3a +# define MANUFACTURER_ID(x) (((x) & 0xffff) << 0) +# define PRODUCT_ID(x) (((x) & 0xffff) << 16) +#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO1 0x3b +# define SINK_DESCRIPTION_LEN(x) (((x) & 0xff) << 0) +#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO2 0x3c +# define PORT_ID0(x) (((x) & 0xffffffff) << 0) +#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO3 0x3d +# define PORT_ID1(x) (((x) & 0xffffffff) << 0) +#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO4 0x3e +# define DESCRIPTION0(x) (((x) & 0xff) << 0) +# define DESCRIPTION1(x) (((x) & 0xff) << 8) +# define DESCRIPTION2(x) (((x) & 0xff) << 16) +# define DESCRIPTION3(x) (((x) & 0xff) << 24) +#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO5 0x3f +# define DESCRIPTION4(x) (((x) & 0xff) << 0) +# define DESCRIPTION5(x) (((x) & 0xff) << 8) +# define DESCRIPTION6(x) (((x) & 0xff) << 16) +# define DESCRIPTION7(x) (((x) & 0xff) << 24) +#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO6 0x40 +# define DESCRIPTION8(x) (((x) & 0xff) << 0) +# define DESCRIPTION9(x) (((x) & 0xff) << 8) +# define DESCRIPTION10(x) (((x) & 0xff) << 16) +# define DESCRIPTION11(x) (((x) & 0xff) << 24) +#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO7 0x41 +# define DESCRIPTION12(x) (((x) & 0xff) << 0) +# define DESCRIPTION13(x) (((x) & 0xff) << 8) +# define DESCRIPTION14(x) (((x) & 0xff) << 16) +# define DESCRIPTION15(x) (((x) & 0xff) << 24) +#define AZ_F0_CODEC_PIN_CONTROL_SINK_INFO8 0x42 +# define DESCRIPTION16(x) (((x) & 0xff) << 0) +# define DESCRIPTION17(x) (((x) & 0xff) << 8) + +#define AZ_F0_CODEC_PIN_CONTROL_HOTPLUG_CONTROL 0x54 +# define AUDIO_ENABLED (1 << 31) + +#define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_CONFIGURATION_DEFAULT 0x56 +#define PORT_CONNECTIVITY_MASK (3 << 30) +#define PORT_CONNECTIVITY_SHIFT 30 + #define DC_LB_MEMORY_SPLIT 0x6b0c #define DC_LB_MEMORY_CONFIG(x) ((x) << 20) @@ -701,7 +823,7 @@ # define GRPH_PFLIP_INT_MASK (1 << 0) # define GRPH_PFLIP_INT_TYPE (1 << 8) -#define DACA_AUTODETECT_INT_CONTROL 0x66c8 +#define DAC_AUTODETECT_INT_CONTROL 0x67c8 #define DC_HPD1_INT_STATUS 0x601c #define DC_HPD2_INT_STATUS 0x6028 @@ -741,6 +863,17 @@ /* 0x6e98, 0x7a98, 0x10698, 0x11298, 0x11e98, 0x12a98 */ #define CRTC_STATUS_FRAME_COUNT 0x6e98 +#define AFMT_AUDIO_SRC_CONTROL 0x713c +#define AFMT_AUDIO_SRC_SELECT(x) (((x) & 7) << 0) +/* AFMT_AUDIO_SRC_SELECT + * 0 = stream0 + * 1 = stream1 + * 2 = stream2 + * 3 = stream3 + * 4 = stream4 + * 5 = stream5 + */ + #define GRBM_CNTL 0x8000 #define GRBM_READ_TIMEOUT(x) ((x) << 0) @@ -1281,6 +1414,7 @@ /* PCIE registers idx/data 0x30/0x34 */ #define PCIE_CNTL2 0x1c /* PCIE */ # define SLV_MEM_LS_EN (1 << 16) +# define SLV_MEM_AGGRESSIVE_LS_EN (1 << 17) # define MST_MEM_LS_EN (1 << 18) # define REPLAY_MEM_LS_EN (1 << 19) #define PCIE_LC_STATUS1 0x28 /* PCIE */ @@ -1472,7 +1606,7 @@ * 6. COMMAND [30:21] | BYTE_COUNT [20:0] */ # define PACKET3_CP_DMA_DST_SEL(x) ((x) << 20) - /* 0 - SRC_ADDR + /* 0 - DST_ADDR * 1 - GDS */ # define PACKET3_CP_DMA_ENGINE(x) ((x) << 27) @@ -1487,7 +1621,7 @@ # define PACKET3_CP_DMA_CP_SYNC (1 << 31) /* COMMAND */ # define PACKET3_CP_DMA_DIS_WC (1 << 21) -# define PACKET3_CP_DMA_CMD_SRC_SWAP(x) ((x) << 23) +# define PACKET3_CP_DMA_CMD_SRC_SWAP(x) ((x) << 22) /* 0 - none * 1 - 8 in 16 * 2 - 8 in 32 @@ -1630,6 +1764,10 @@ # define DMA_IDLE (1 << 0) #define DMA_TILING_CONFIG 0xd0b8 +#define DMA_POWER_CNTL 0xd0bc +# define MEM_POWER_OVERRIDE (1 << 8) +#define DMA_CLK_CTRL 0xd0c0 + #define DMA_PG 0xd0d4 # define PG_CNTL_ENABLE (1 << 0) #define DMA_PGFSM_CONFIG 0xd0d8 @@ -1661,4 +1799,51 @@ #define DMA_PACKET_CONSTANT_FILL 0xd #define DMA_PACKET_NOP 0xf +#define VCE_STATUS 0x20004 +#define VCE_VCPU_CNTL 0x20014 +#define VCE_CLK_EN (1 << 0) +#define VCE_VCPU_CACHE_OFFSET0 0x20024 +#define VCE_VCPU_CACHE_SIZE0 0x20028 +#define VCE_VCPU_CACHE_OFFSET1 0x2002c +#define VCE_VCPU_CACHE_SIZE1 0x20030 +#define VCE_VCPU_CACHE_OFFSET2 0x20034 +#define VCE_VCPU_CACHE_SIZE2 0x20038 +#define VCE_SOFT_RESET 0x20120 +#define VCE_ECPU_SOFT_RESET (1 << 0) +#define VCE_FME_SOFT_RESET (1 << 2) +#define VCE_RB_BASE_LO2 0x2016c +#define VCE_RB_BASE_HI2 0x20170 +#define VCE_RB_SIZE2 0x20174 +#define VCE_RB_RPTR2 0x20178 +#define VCE_RB_WPTR2 0x2017c +#define VCE_RB_BASE_LO 0x20180 +#define VCE_RB_BASE_HI 0x20184 +#define VCE_RB_SIZE 0x20188 +#define VCE_RB_RPTR 0x2018c +#define VCE_RB_WPTR 0x20190 +#define VCE_CLOCK_GATING_A 0x202f8 +#define VCE_CLOCK_GATING_B 0x202fc +#define VCE_UENC_CLOCK_GATING 0x205bc +#define VCE_UENC_REG_CLOCK_GATING 0x205c0 +#define VCE_FW_REG_STATUS 0x20e10 +# define VCE_FW_REG_STATUS_BUSY (1 << 0) +# define VCE_FW_REG_STATUS_PASS (1 << 3) +# define VCE_FW_REG_STATUS_DONE (1 << 11) +#define VCE_LMI_FW_START_KEYSEL 0x20e18 +#define VCE_LMI_FW_PERIODIC_CTRL 0x20e20 +#define VCE_LMI_CTRL2 0x20e74 +#define VCE_LMI_CTRL 0x20e98 +#define VCE_LMI_VM_CTRL 0x20ea0 +#define VCE_LMI_SWAP_CNTL 0x20eb4 +#define VCE_LMI_SWAP_CNTL1 0x20eb8 +#define VCE_LMI_CACHE_CTRL 0x20ef4 + +#define VCE_CMD_NO_OP 0x00000000 +#define VCE_CMD_END 0x00000001 +#define VCE_CMD_IB 0x00000002 +#define VCE_CMD_FENCE 0x00000003 +#define VCE_CMD_TRAP 0x00000004 +#define VCE_CMD_IB_AUTO 0x00000005 +#define VCE_CMD_SEMAPHORE 0x00000006 + #endif diff --git a/drivers/gpu/drm/radeon/sislands_smc.h b/drivers/gpu/drm/radeon/sislands_smc.h index 5578e983702..10e945a4947 100644 --- a/drivers/gpu/drm/radeon/sislands_smc.h +++ b/drivers/gpu/drm/radeon/sislands_smc.h @@ -374,8 +374,6 @@ typedef struct Smc_SIslands_DTE_Configuration Smc_SIslands_DTE_Configuration; #pragma pack(pop) -int si_set_smc_sram_address(struct radeon_device *rdev, - u32 smc_address, u32 limit); int si_copy_bytes_to_smc(struct radeon_device *rdev, u32 smc_start_address, const u8 *src, u32 byte_count, u32 limit); diff --git a/drivers/gpu/drm/radeon/smu7.h b/drivers/gpu/drm/radeon/smu7.h new file mode 100644 index 00000000000..75a380a1529 --- /dev/null +++ b/drivers/gpu/drm/radeon/smu7.h @@ -0,0 +1,170 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef SMU7_H +#define SMU7_H + +#pragma pack(push, 1) + +#define SMU7_CONTEXT_ID_SMC 1 +#define SMU7_CONTEXT_ID_VBIOS 2 + + +#define SMU7_CONTEXT_ID_SMC 1 +#define SMU7_CONTEXT_ID_VBIOS 2 + +#define SMU7_MAX_LEVELS_VDDC 8 +#define SMU7_MAX_LEVELS_VDDCI 4 +#define SMU7_MAX_LEVELS_MVDD 4 +#define SMU7_MAX_LEVELS_VDDNB 8 + +#define SMU7_MAX_LEVELS_GRAPHICS SMU__NUM_SCLK_DPM_STATE // SCLK + SQ DPM + ULV +#define SMU7_MAX_LEVELS_MEMORY SMU__NUM_MCLK_DPM_LEVELS // MCLK Levels DPM +#define SMU7_MAX_LEVELS_GIO SMU__NUM_LCLK_DPM_LEVELS // LCLK Levels +#define SMU7_MAX_LEVELS_LINK SMU__NUM_PCIE_DPM_LEVELS // PCIe speed and number of lanes. +#define SMU7_MAX_LEVELS_UVD 8 // VCLK/DCLK levels for UVD. +#define SMU7_MAX_LEVELS_VCE 8 // ECLK levels for VCE. +#define SMU7_MAX_LEVELS_ACP 8 // ACLK levels for ACP. +#define SMU7_MAX_LEVELS_SAMU 8 // SAMCLK levels for SAMU. +#define SMU7_MAX_ENTRIES_SMIO 32 // Number of entries in SMIO table. + +#define DPM_NO_LIMIT 0 +#define DPM_NO_UP 1 +#define DPM_GO_DOWN 2 +#define DPM_GO_UP 3 + +#define SMU7_FIRST_DPM_GRAPHICS_LEVEL 0 +#define SMU7_FIRST_DPM_MEMORY_LEVEL 0 + +#define GPIO_CLAMP_MODE_VRHOT 1 +#define GPIO_CLAMP_MODE_THERM 2 +#define GPIO_CLAMP_MODE_DC 4 + +#define SCRATCH_B_TARG_PCIE_INDEX_SHIFT 0 +#define SCRATCH_B_TARG_PCIE_INDEX_MASK (0x7<<SCRATCH_B_TARG_PCIE_INDEX_SHIFT) +#define SCRATCH_B_CURR_PCIE_INDEX_SHIFT 3 +#define SCRATCH_B_CURR_PCIE_INDEX_MASK (0x7<<SCRATCH_B_CURR_PCIE_INDEX_SHIFT) +#define SCRATCH_B_TARG_UVD_INDEX_SHIFT 6 +#define SCRATCH_B_TARG_UVD_INDEX_MASK (0x7<<SCRATCH_B_TARG_UVD_INDEX_SHIFT) +#define SCRATCH_B_CURR_UVD_INDEX_SHIFT 9 +#define SCRATCH_B_CURR_UVD_INDEX_MASK (0x7<<SCRATCH_B_CURR_UVD_INDEX_SHIFT) +#define SCRATCH_B_TARG_VCE_INDEX_SHIFT 12 +#define SCRATCH_B_TARG_VCE_INDEX_MASK (0x7<<SCRATCH_B_TARG_VCE_INDEX_SHIFT) +#define SCRATCH_B_CURR_VCE_INDEX_SHIFT 15 +#define SCRATCH_B_CURR_VCE_INDEX_MASK (0x7<<SCRATCH_B_CURR_VCE_INDEX_SHIFT) +#define SCRATCH_B_TARG_ACP_INDEX_SHIFT 18 +#define SCRATCH_B_TARG_ACP_INDEX_MASK (0x7<<SCRATCH_B_TARG_ACP_INDEX_SHIFT) +#define SCRATCH_B_CURR_ACP_INDEX_SHIFT 21 +#define SCRATCH_B_CURR_ACP_INDEX_MASK (0x7<<SCRATCH_B_CURR_ACP_INDEX_SHIFT) +#define SCRATCH_B_TARG_SAMU_INDEX_SHIFT 24 +#define SCRATCH_B_TARG_SAMU_INDEX_MASK (0x7<<SCRATCH_B_TARG_SAMU_INDEX_SHIFT) +#define SCRATCH_B_CURR_SAMU_INDEX_SHIFT 27 +#define SCRATCH_B_CURR_SAMU_INDEX_MASK (0x7<<SCRATCH_B_CURR_SAMU_INDEX_SHIFT) + + +struct SMU7_PIDController +{ + uint32_t Ki; + int32_t LFWindupUL; + int32_t LFWindupLL; + uint32_t StatePrecision; + uint32_t LfPrecision; + uint32_t LfOffset; + uint32_t MaxState; + uint32_t MaxLfFraction; + uint32_t StateShift; +}; + +typedef struct SMU7_PIDController SMU7_PIDController; + +// ------------------------------------------------------------------------------------------------------------------------- +#define SMU7_MAX_PCIE_LINK_SPEEDS 3 /* 0:Gen1 1:Gen2 2:Gen3 */ + +#define SMU7_SCLK_DPM_CONFIG_MASK 0x01 +#define SMU7_VOLTAGE_CONTROLLER_CONFIG_MASK 0x02 +#define SMU7_THERMAL_CONTROLLER_CONFIG_MASK 0x04 +#define SMU7_MCLK_DPM_CONFIG_MASK 0x08 +#define SMU7_UVD_DPM_CONFIG_MASK 0x10 +#define SMU7_VCE_DPM_CONFIG_MASK 0x20 +#define SMU7_ACP_DPM_CONFIG_MASK 0x40 +#define SMU7_SAMU_DPM_CONFIG_MASK 0x80 +#define SMU7_PCIEGEN_DPM_CONFIG_MASK 0x100 + +#define SMU7_ACP_MCLK_HANDSHAKE_DISABLE 0x00000001 +#define SMU7_ACP_SCLK_HANDSHAKE_DISABLE 0x00000002 +#define SMU7_UVD_MCLK_HANDSHAKE_DISABLE 0x00000100 +#define SMU7_UVD_SCLK_HANDSHAKE_DISABLE 0x00000200 +#define SMU7_VCE_MCLK_HANDSHAKE_DISABLE 0x00010000 +#define SMU7_VCE_SCLK_HANDSHAKE_DISABLE 0x00020000 + +struct SMU7_Firmware_Header +{ + uint32_t Digest[5]; + uint32_t Version; + uint32_t HeaderSize; + uint32_t Flags; + uint32_t EntryPoint; + uint32_t CodeSize; + uint32_t ImageSize; + + uint32_t Rtos; + uint32_t SoftRegisters; + uint32_t DpmTable; + uint32_t FanTable; + uint32_t CacConfigTable; + uint32_t CacStatusTable; + + uint32_t mcRegisterTable; + + uint32_t mcArbDramTimingTable; + + uint32_t PmFuseTable; + uint32_t Globals; + uint32_t Reserved[42]; + uint32_t Signature; +}; + +typedef struct SMU7_Firmware_Header SMU7_Firmware_Header; + +#define SMU7_FIRMWARE_HEADER_LOCATION 0x20000 + +enum DisplayConfig { + PowerDown = 1, + DP54x4, + DP54x2, + DP54x1, + DP27x4, + DP27x2, + DP27x1, + HDMI297, + HDMI162, + LVDS, + DP324x4, + DP324x2, + DP324x1 +}; + +#pragma pack(pop) + +#endif + diff --git a/drivers/gpu/drm/radeon/smu7_discrete.h b/drivers/gpu/drm/radeon/smu7_discrete.h new file mode 100644 index 00000000000..82f70c90a9e --- /dev/null +++ b/drivers/gpu/drm/radeon/smu7_discrete.h @@ -0,0 +1,486 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef SMU7_DISCRETE_H +#define SMU7_DISCRETE_H + +#include "smu7.h" + +#pragma pack(push, 1) + +#define SMU7_DTE_ITERATIONS 5 +#define SMU7_DTE_SOURCES 3 +#define SMU7_DTE_SINKS 1 +#define SMU7_NUM_CPU_TES 0 +#define SMU7_NUM_GPU_TES 1 +#define SMU7_NUM_NON_TES 2 + +struct SMU7_SoftRegisters +{ + uint32_t RefClockFrequency; + uint32_t PmTimerP; + uint32_t FeatureEnables; + uint32_t PreVBlankGap; + uint32_t VBlankTimeout; + uint32_t TrainTimeGap; + + uint32_t MvddSwitchTime; + uint32_t LongestAcpiTrainTime; + uint32_t AcpiDelay; + uint32_t G5TrainTime; + uint32_t DelayMpllPwron; + uint32_t VoltageChangeTimeout; + uint32_t HandshakeDisables; + + uint8_t DisplayPhy1Config; + uint8_t DisplayPhy2Config; + uint8_t DisplayPhy3Config; + uint8_t DisplayPhy4Config; + + uint8_t DisplayPhy5Config; + uint8_t DisplayPhy6Config; + uint8_t DisplayPhy7Config; + uint8_t DisplayPhy8Config; + + uint32_t AverageGraphicsA; + uint32_t AverageMemoryA; + uint32_t AverageGioA; + + uint8_t SClkDpmEnabledLevels; + uint8_t MClkDpmEnabledLevels; + uint8_t LClkDpmEnabledLevels; + uint8_t PCIeDpmEnabledLevels; + + uint8_t UVDDpmEnabledLevels; + uint8_t SAMUDpmEnabledLevels; + uint8_t ACPDpmEnabledLevels; + uint8_t VCEDpmEnabledLevels; + + uint32_t DRAM_LOG_ADDR_H; + uint32_t DRAM_LOG_ADDR_L; + uint32_t DRAM_LOG_PHY_ADDR_H; + uint32_t DRAM_LOG_PHY_ADDR_L; + uint32_t DRAM_LOG_BUFF_SIZE; + uint32_t UlvEnterC; + uint32_t UlvTime; + uint32_t Reserved[3]; + +}; + +typedef struct SMU7_SoftRegisters SMU7_SoftRegisters; + +struct SMU7_Discrete_VoltageLevel +{ + uint16_t Voltage; + uint16_t StdVoltageHiSidd; + uint16_t StdVoltageLoSidd; + uint8_t Smio; + uint8_t padding; +}; + +typedef struct SMU7_Discrete_VoltageLevel SMU7_Discrete_VoltageLevel; + +struct SMU7_Discrete_GraphicsLevel +{ + uint32_t Flags; + uint32_t MinVddc; + uint32_t MinVddcPhases; + + uint32_t SclkFrequency; + + uint8_t padding1[2]; + uint16_t ActivityLevel; + + uint32_t CgSpllFuncCntl3; + uint32_t CgSpllFuncCntl4; + uint32_t SpllSpreadSpectrum; + uint32_t SpllSpreadSpectrum2; + uint32_t CcPwrDynRm; + uint32_t CcPwrDynRm1; + uint8_t SclkDid; + uint8_t DisplayWatermark; + uint8_t EnabledForActivity; + uint8_t EnabledForThrottle; + uint8_t UpH; + uint8_t DownH; + uint8_t VoltageDownH; + uint8_t PowerThrottle; + uint8_t DeepSleepDivId; + uint8_t padding[3]; +}; + +typedef struct SMU7_Discrete_GraphicsLevel SMU7_Discrete_GraphicsLevel; + +struct SMU7_Discrete_ACPILevel +{ + uint32_t Flags; + uint32_t MinVddc; + uint32_t MinVddcPhases; + uint32_t SclkFrequency; + uint8_t SclkDid; + uint8_t DisplayWatermark; + uint8_t DeepSleepDivId; + uint8_t padding; + uint32_t CgSpllFuncCntl; + uint32_t CgSpllFuncCntl2; + uint32_t CgSpllFuncCntl3; + uint32_t CgSpllFuncCntl4; + uint32_t SpllSpreadSpectrum; + uint32_t SpllSpreadSpectrum2; + uint32_t CcPwrDynRm; + uint32_t CcPwrDynRm1; +}; + +typedef struct SMU7_Discrete_ACPILevel SMU7_Discrete_ACPILevel; + +struct SMU7_Discrete_Ulv +{ + uint32_t CcPwrDynRm; + uint32_t CcPwrDynRm1; + uint16_t VddcOffset; + uint8_t VddcOffsetVid; + uint8_t VddcPhase; + uint32_t Reserved; +}; + +typedef struct SMU7_Discrete_Ulv SMU7_Discrete_Ulv; + +struct SMU7_Discrete_MemoryLevel +{ + uint32_t MinVddc; + uint32_t MinVddcPhases; + uint32_t MinVddci; + uint32_t MinMvdd; + + uint32_t MclkFrequency; + + uint8_t EdcReadEnable; + uint8_t EdcWriteEnable; + uint8_t RttEnable; + uint8_t StutterEnable; + + uint8_t StrobeEnable; + uint8_t StrobeRatio; + uint8_t EnabledForThrottle; + uint8_t EnabledForActivity; + + uint8_t UpH; + uint8_t DownH; + uint8_t VoltageDownH; + uint8_t padding; + + uint16_t ActivityLevel; + uint8_t DisplayWatermark; + uint8_t padding1; + + uint32_t MpllFuncCntl; + uint32_t MpllFuncCntl_1; + uint32_t MpllFuncCntl_2; + uint32_t MpllAdFuncCntl; + uint32_t MpllDqFuncCntl; + uint32_t MclkPwrmgtCntl; + uint32_t DllCntl; + uint32_t MpllSs1; + uint32_t MpllSs2; +}; + +typedef struct SMU7_Discrete_MemoryLevel SMU7_Discrete_MemoryLevel; + +struct SMU7_Discrete_LinkLevel +{ + uint8_t PcieGenSpeed; + uint8_t PcieLaneCount; + uint8_t EnabledForActivity; + uint8_t Padding; + uint32_t DownT; + uint32_t UpT; + uint32_t Reserved; +}; + +typedef struct SMU7_Discrete_LinkLevel SMU7_Discrete_LinkLevel; + + +struct SMU7_Discrete_MCArbDramTimingTableEntry +{ + uint32_t McArbDramTiming; + uint32_t McArbDramTiming2; + uint8_t McArbBurstTime; + uint8_t padding[3]; +}; + +typedef struct SMU7_Discrete_MCArbDramTimingTableEntry SMU7_Discrete_MCArbDramTimingTableEntry; + +struct SMU7_Discrete_MCArbDramTimingTable +{ + SMU7_Discrete_MCArbDramTimingTableEntry entries[SMU__NUM_SCLK_DPM_STATE][SMU__NUM_MCLK_DPM_LEVELS]; +}; + +typedef struct SMU7_Discrete_MCArbDramTimingTable SMU7_Discrete_MCArbDramTimingTable; + +struct SMU7_Discrete_UvdLevel +{ + uint32_t VclkFrequency; + uint32_t DclkFrequency; + uint16_t MinVddc; + uint8_t MinVddcPhases; + uint8_t VclkDivider; + uint8_t DclkDivider; + uint8_t padding[3]; +}; + +typedef struct SMU7_Discrete_UvdLevel SMU7_Discrete_UvdLevel; + +struct SMU7_Discrete_ExtClkLevel +{ + uint32_t Frequency; + uint16_t MinVoltage; + uint8_t MinPhases; + uint8_t Divider; +}; + +typedef struct SMU7_Discrete_ExtClkLevel SMU7_Discrete_ExtClkLevel; + +struct SMU7_Discrete_StateInfo +{ + uint32_t SclkFrequency; + uint32_t MclkFrequency; + uint32_t VclkFrequency; + uint32_t DclkFrequency; + uint32_t SamclkFrequency; + uint32_t AclkFrequency; + uint32_t EclkFrequency; + uint16_t MvddVoltage; + uint16_t padding16; + uint8_t DisplayWatermark; + uint8_t McArbIndex; + uint8_t McRegIndex; + uint8_t SeqIndex; + uint8_t SclkDid; + int8_t SclkIndex; + int8_t MclkIndex; + uint8_t PCIeGen; + +}; + +typedef struct SMU7_Discrete_StateInfo SMU7_Discrete_StateInfo; + + +struct SMU7_Discrete_DpmTable +{ + SMU7_PIDController GraphicsPIDController; + SMU7_PIDController MemoryPIDController; + SMU7_PIDController LinkPIDController; + + uint32_t SystemFlags; + + + uint32_t SmioMaskVddcVid; + uint32_t SmioMaskVddcPhase; + uint32_t SmioMaskVddciVid; + uint32_t SmioMaskMvddVid; + + uint32_t VddcLevelCount; + uint32_t VddciLevelCount; + uint32_t MvddLevelCount; + + SMU7_Discrete_VoltageLevel VddcLevel [SMU7_MAX_LEVELS_VDDC]; +// SMU7_Discrete_VoltageLevel VddcStandardReference [SMU7_MAX_LEVELS_VDDC]; + SMU7_Discrete_VoltageLevel VddciLevel [SMU7_MAX_LEVELS_VDDCI]; + SMU7_Discrete_VoltageLevel MvddLevel [SMU7_MAX_LEVELS_MVDD]; + + uint8_t GraphicsDpmLevelCount; + uint8_t MemoryDpmLevelCount; + uint8_t LinkLevelCount; + uint8_t UvdLevelCount; + uint8_t VceLevelCount; + uint8_t AcpLevelCount; + uint8_t SamuLevelCount; + uint8_t MasterDeepSleepControl; + uint32_t Reserved[5]; +// uint32_t SamuDefaultLevel; + + SMU7_Discrete_GraphicsLevel GraphicsLevel [SMU7_MAX_LEVELS_GRAPHICS]; + SMU7_Discrete_MemoryLevel MemoryACPILevel; + SMU7_Discrete_MemoryLevel MemoryLevel [SMU7_MAX_LEVELS_MEMORY]; + SMU7_Discrete_LinkLevel LinkLevel [SMU7_MAX_LEVELS_LINK]; + SMU7_Discrete_ACPILevel ACPILevel; + SMU7_Discrete_UvdLevel UvdLevel [SMU7_MAX_LEVELS_UVD]; + SMU7_Discrete_ExtClkLevel VceLevel [SMU7_MAX_LEVELS_VCE]; + SMU7_Discrete_ExtClkLevel AcpLevel [SMU7_MAX_LEVELS_ACP]; + SMU7_Discrete_ExtClkLevel SamuLevel [SMU7_MAX_LEVELS_SAMU]; + SMU7_Discrete_Ulv Ulv; + + uint32_t SclkStepSize; + uint32_t Smio [SMU7_MAX_ENTRIES_SMIO]; + + uint8_t UvdBootLevel; + uint8_t VceBootLevel; + uint8_t AcpBootLevel; + uint8_t SamuBootLevel; + + uint8_t UVDInterval; + uint8_t VCEInterval; + uint8_t ACPInterval; + uint8_t SAMUInterval; + + uint8_t GraphicsBootLevel; + uint8_t GraphicsVoltageChangeEnable; + uint8_t GraphicsThermThrottleEnable; + uint8_t GraphicsInterval; + + uint8_t VoltageInterval; + uint8_t ThermalInterval; + uint16_t TemperatureLimitHigh; + + uint16_t TemperatureLimitLow; + uint8_t MemoryBootLevel; + uint8_t MemoryVoltageChangeEnable; + + uint8_t MemoryInterval; + uint8_t MemoryThermThrottleEnable; + uint16_t VddcVddciDelta; + + uint16_t VoltageResponseTime; + uint16_t PhaseResponseTime; + + uint8_t PCIeBootLinkLevel; + uint8_t PCIeGenInterval; + uint8_t DTEInterval; + uint8_t DTEMode; + + uint8_t SVI2Enable; + uint8_t VRHotGpio; + uint8_t AcDcGpio; + uint8_t ThermGpio; + + uint16_t PPM_PkgPwrLimit; + uint16_t PPM_TemperatureLimit; + + uint16_t DefaultTdp; + uint16_t TargetTdp; + + uint16_t FpsHighT; + uint16_t FpsLowT; + + uint16_t BAPMTI_R [SMU7_DTE_ITERATIONS][SMU7_DTE_SOURCES][SMU7_DTE_SINKS]; + uint16_t BAPMTI_RC [SMU7_DTE_ITERATIONS][SMU7_DTE_SOURCES][SMU7_DTE_SINKS]; + + uint8_t DTEAmbientTempBase; + uint8_t DTETjOffset; + uint8_t GpuTjMax; + uint8_t GpuTjHyst; + + uint16_t BootVddc; + uint16_t BootVddci; + + uint16_t BootMVdd; + uint16_t padding; + + uint32_t BAPM_TEMP_GRADIENT; + + uint32_t LowSclkInterruptT; +}; + +typedef struct SMU7_Discrete_DpmTable SMU7_Discrete_DpmTable; + +#define SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE 16 +#define SMU7_DISCRETE_MC_REGISTER_ARRAY_SET_COUNT SMU7_MAX_LEVELS_MEMORY + +struct SMU7_Discrete_MCRegisterAddress +{ + uint16_t s0; + uint16_t s1; +}; + +typedef struct SMU7_Discrete_MCRegisterAddress SMU7_Discrete_MCRegisterAddress; + +struct SMU7_Discrete_MCRegisterSet +{ + uint32_t value[SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE]; +}; + +typedef struct SMU7_Discrete_MCRegisterSet SMU7_Discrete_MCRegisterSet; + +struct SMU7_Discrete_MCRegisters +{ + uint8_t last; + uint8_t reserved[3]; + SMU7_Discrete_MCRegisterAddress address[SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE]; + SMU7_Discrete_MCRegisterSet data[SMU7_DISCRETE_MC_REGISTER_ARRAY_SET_COUNT]; +}; + +typedef struct SMU7_Discrete_MCRegisters SMU7_Discrete_MCRegisters; + +struct SMU7_Discrete_PmFuses { + // dw0-dw1 + uint8_t BapmVddCVidHiSidd[8]; + + // dw2-dw3 + uint8_t BapmVddCVidLoSidd[8]; + + // dw4-dw5 + uint8_t VddCVid[8]; + + // dw6 + uint8_t SviLoadLineEn; + uint8_t SviLoadLineVddC; + uint8_t SviLoadLineTrimVddC; + uint8_t SviLoadLineOffsetVddC; + + // dw7 + uint16_t TDC_VDDC_PkgLimit; + uint8_t TDC_VDDC_ThrottleReleaseLimitPerc; + uint8_t TDC_MAWt; + + // dw8 + uint8_t TdcWaterfallCtl; + uint8_t LPMLTemperatureMin; + uint8_t LPMLTemperatureMax; + uint8_t Reserved; + + // dw9-dw10 + uint8_t BapmVddCVidHiSidd2[8]; + + // dw11-dw12 + uint32_t Reserved6[2]; + + // dw13-dw16 + uint8_t GnbLPML[16]; + + // dw17 + uint8_t GnbLPMLMaxVid; + uint8_t GnbLPMLMinVid; + uint8_t Reserved1[2]; + + // dw18 + uint16_t BapmVddCBaseLeakageHiSidd; + uint16_t BapmVddCBaseLeakageLoSidd; +}; + +typedef struct SMU7_Discrete_PmFuses SMU7_Discrete_PmFuses; + + +#pragma pack(pop) + +#endif + diff --git a/drivers/gpu/drm/radeon/smu7_fusion.h b/drivers/gpu/drm/radeon/smu7_fusion.h new file mode 100644 index 00000000000..78ada9ffd50 --- /dev/null +++ b/drivers/gpu/drm/radeon/smu7_fusion.h @@ -0,0 +1,300 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef SMU7_FUSION_H +#define SMU7_FUSION_H + +#include "smu7.h" + +#pragma pack(push, 1) + +#define SMU7_DTE_ITERATIONS 5 +#define SMU7_DTE_SOURCES 5 +#define SMU7_DTE_SINKS 3 +#define SMU7_NUM_CPU_TES 2 +#define SMU7_NUM_GPU_TES 1 +#define SMU7_NUM_NON_TES 2 + +// All 'soft registers' should be uint32_t. +struct SMU7_SoftRegisters +{ + uint32_t RefClockFrequency; + uint32_t PmTimerP; + uint32_t FeatureEnables; + uint32_t HandshakeDisables; + + uint8_t DisplayPhy1Config; + uint8_t DisplayPhy2Config; + uint8_t DisplayPhy3Config; + uint8_t DisplayPhy4Config; + + uint8_t DisplayPhy5Config; + uint8_t DisplayPhy6Config; + uint8_t DisplayPhy7Config; + uint8_t DisplayPhy8Config; + + uint32_t AverageGraphicsA; + uint32_t AverageMemoryA; + uint32_t AverageGioA; + + uint8_t SClkDpmEnabledLevels; + uint8_t MClkDpmEnabledLevels; + uint8_t LClkDpmEnabledLevels; + uint8_t PCIeDpmEnabledLevels; + + uint8_t UVDDpmEnabledLevels; + uint8_t SAMUDpmEnabledLevels; + uint8_t ACPDpmEnabledLevels; + uint8_t VCEDpmEnabledLevels; + + uint32_t DRAM_LOG_ADDR_H; + uint32_t DRAM_LOG_ADDR_L; + uint32_t DRAM_LOG_PHY_ADDR_H; + uint32_t DRAM_LOG_PHY_ADDR_L; + uint32_t DRAM_LOG_BUFF_SIZE; + uint32_t UlvEnterC; + uint32_t UlvTime; + uint32_t Reserved[3]; + +}; + +typedef struct SMU7_SoftRegisters SMU7_SoftRegisters; + +struct SMU7_Fusion_GraphicsLevel +{ + uint32_t MinVddNb; + + uint32_t SclkFrequency; + + uint8_t Vid; + uint8_t VidOffset; + uint16_t AT; + + uint8_t PowerThrottle; + uint8_t GnbSlow; + uint8_t ForceNbPs1; + uint8_t SclkDid; + + uint8_t DisplayWatermark; + uint8_t EnabledForActivity; + uint8_t EnabledForThrottle; + uint8_t UpH; + + uint8_t DownH; + uint8_t VoltageDownH; + uint8_t DeepSleepDivId; + + uint8_t ClkBypassCntl; + + uint32_t reserved; +}; + +typedef struct SMU7_Fusion_GraphicsLevel SMU7_Fusion_GraphicsLevel; + +struct SMU7_Fusion_GIOLevel +{ + uint8_t EnabledForActivity; + uint8_t LclkDid; + uint8_t Vid; + uint8_t VoltageDownH; + + uint32_t MinVddNb; + + uint16_t ResidencyCounter; + uint8_t UpH; + uint8_t DownH; + + uint32_t LclkFrequency; + + uint8_t ActivityLevel; + uint8_t EnabledForThrottle; + + uint8_t ClkBypassCntl; + + uint8_t padding; +}; + +typedef struct SMU7_Fusion_GIOLevel SMU7_Fusion_GIOLevel; + +// UVD VCLK/DCLK state (level) definition. +struct SMU7_Fusion_UvdLevel +{ + uint32_t VclkFrequency; + uint32_t DclkFrequency; + uint16_t MinVddNb; + uint8_t VclkDivider; + uint8_t DclkDivider; + + uint8_t VClkBypassCntl; + uint8_t DClkBypassCntl; + + uint8_t padding[2]; + +}; + +typedef struct SMU7_Fusion_UvdLevel SMU7_Fusion_UvdLevel; + +// Clocks for other external blocks (VCE, ACP, SAMU). +struct SMU7_Fusion_ExtClkLevel +{ + uint32_t Frequency; + uint16_t MinVoltage; + uint8_t Divider; + uint8_t ClkBypassCntl; + + uint32_t Reserved; +}; +typedef struct SMU7_Fusion_ExtClkLevel SMU7_Fusion_ExtClkLevel; + +struct SMU7_Fusion_ACPILevel +{ + uint32_t Flags; + uint32_t MinVddNb; + uint32_t SclkFrequency; + uint8_t SclkDid; + uint8_t GnbSlow; + uint8_t ForceNbPs1; + uint8_t DisplayWatermark; + uint8_t DeepSleepDivId; + uint8_t padding[3]; +}; + +typedef struct SMU7_Fusion_ACPILevel SMU7_Fusion_ACPILevel; + +struct SMU7_Fusion_NbDpm +{ + uint8_t DpmXNbPsHi; + uint8_t DpmXNbPsLo; + uint8_t Dpm0PgNbPsHi; + uint8_t Dpm0PgNbPsLo; + uint8_t EnablePsi1; + uint8_t SkipDPM0; + uint8_t SkipPG; + uint8_t Hysteresis; + uint8_t EnableDpmPstatePoll; + uint8_t padding[3]; +}; + +typedef struct SMU7_Fusion_NbDpm SMU7_Fusion_NbDpm; + +struct SMU7_Fusion_StateInfo +{ + uint32_t SclkFrequency; + uint32_t LclkFrequency; + uint32_t VclkFrequency; + uint32_t DclkFrequency; + uint32_t SamclkFrequency; + uint32_t AclkFrequency; + uint32_t EclkFrequency; + uint8_t DisplayWatermark; + uint8_t McArbIndex; + int8_t SclkIndex; + int8_t MclkIndex; +}; + +typedef struct SMU7_Fusion_StateInfo SMU7_Fusion_StateInfo; + +struct SMU7_Fusion_DpmTable +{ + uint32_t SystemFlags; + + SMU7_PIDController GraphicsPIDController; + SMU7_PIDController GioPIDController; + + uint8_t GraphicsDpmLevelCount; + uint8_t GIOLevelCount; + uint8_t UvdLevelCount; + uint8_t VceLevelCount; + + uint8_t AcpLevelCount; + uint8_t SamuLevelCount; + uint16_t FpsHighT; + + SMU7_Fusion_GraphicsLevel GraphicsLevel [SMU__NUM_SCLK_DPM_STATE]; + SMU7_Fusion_ACPILevel ACPILevel; + SMU7_Fusion_UvdLevel UvdLevel [SMU7_MAX_LEVELS_UVD]; + SMU7_Fusion_ExtClkLevel VceLevel [SMU7_MAX_LEVELS_VCE]; + SMU7_Fusion_ExtClkLevel AcpLevel [SMU7_MAX_LEVELS_ACP]; + SMU7_Fusion_ExtClkLevel SamuLevel [SMU7_MAX_LEVELS_SAMU]; + + uint8_t UvdBootLevel; + uint8_t VceBootLevel; + uint8_t AcpBootLevel; + uint8_t SamuBootLevel; + uint8_t UVDInterval; + uint8_t VCEInterval; + uint8_t ACPInterval; + uint8_t SAMUInterval; + + uint8_t GraphicsBootLevel; + uint8_t GraphicsInterval; + uint8_t GraphicsThermThrottleEnable; + uint8_t GraphicsVoltageChangeEnable; + + uint8_t GraphicsClkSlowEnable; + uint8_t GraphicsClkSlowDivider; + uint16_t FpsLowT; + + uint32_t DisplayCac; + uint32_t LowSclkInterruptT; + + uint32_t DRAM_LOG_ADDR_H; + uint32_t DRAM_LOG_ADDR_L; + uint32_t DRAM_LOG_PHY_ADDR_H; + uint32_t DRAM_LOG_PHY_ADDR_L; + uint32_t DRAM_LOG_BUFF_SIZE; + +}; + +struct SMU7_Fusion_GIODpmTable +{ + + SMU7_Fusion_GIOLevel GIOLevel [SMU7_MAX_LEVELS_GIO]; + + SMU7_PIDController GioPIDController; + + uint32_t GIOLevelCount; + + uint8_t Enable; + uint8_t GIOVoltageChangeEnable; + uint8_t GIOBootLevel; + uint8_t padding; + uint8_t padding1[2]; + uint8_t TargetState; + uint8_t CurrenttState; + uint8_t ThrottleOnHtc; + uint8_t ThermThrottleStatus; + uint8_t ThermThrottleTempSelect; + uint8_t ThermThrottleEnable; + uint16_t TemperatureLimitHigh; + uint16_t TemperatureLimitLow; + +}; + +typedef struct SMU7_Fusion_DpmTable SMU7_Fusion_DpmTable; +typedef struct SMU7_Fusion_GIODpmTable SMU7_Fusion_GIODpmTable; + +#pragma pack(pop) + +#endif + diff --git a/drivers/gpu/drm/radeon/sumo_dpm.c b/drivers/gpu/drm/radeon/sumo_dpm.c index 11b6b9924f1..3f0e8d7b8db 100644 --- a/drivers/gpu/drm/radeon/sumo_dpm.c +++ b/drivers/gpu/drm/radeon/sumo_dpm.c @@ -71,7 +71,7 @@ static const u32 sumo_dtc[SUMO_PM_NUMBER_OF_TC] = SUMO_DTC_DFLT_14, }; -struct sumo_ps *sumo_get_ps(struct radeon_ps *rps) +static struct sumo_ps *sumo_get_ps(struct radeon_ps *rps) { struct sumo_ps *ps = rps->ps_priv; @@ -1202,14 +1202,10 @@ static void sumo_update_requested_ps(struct radeon_device *rdev, int sumo_dpm_enable(struct radeon_device *rdev) { struct sumo_power_info *pi = sumo_get_pi(rdev); - int ret; if (sumo_dpm_enabled(rdev)) return -EINVAL; - ret = sumo_enable_clock_power_gating(rdev); - if (ret) - return ret; sumo_program_bootup_state(rdev); sumo_init_bsp(rdev); sumo_reset_am(rdev); @@ -1233,6 +1229,19 @@ int sumo_dpm_enable(struct radeon_device *rdev) if (pi->enable_boost) sumo_enable_boost_timer(rdev); + sumo_update_current_ps(rdev, rdev->pm.dpm.boot_ps); + + return 0; +} + +int sumo_dpm_late_enable(struct radeon_device *rdev) +{ + int ret; + + ret = sumo_enable_clock_power_gating(rdev); + if (ret) + return ret; + if (rdev->irq.installed && r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { ret = sumo_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX); @@ -1242,8 +1251,6 @@ int sumo_dpm_enable(struct radeon_device *rdev) radeon_irq_set(rdev); } - sumo_update_current_ps(rdev, rdev->pm.dpm.boot_ps); - return 0; } @@ -1319,8 +1326,6 @@ int sumo_dpm_set_power_state(struct radeon_device *rdev) if (pi->enable_dpm) sumo_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps); - rdev->pm.dpm.forced_level = RADEON_DPM_FORCED_LEVEL_AUTO; - return 0; } @@ -1479,10 +1484,8 @@ static int sumo_parse_power_table(struct radeon_device *rdev) if (!rdev->pm.dpm.ps) return -ENOMEM; power_state_offset = (u8 *)state_array->states; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < state_array->ucNumEntries; i++) { + u8 *idx; power_state = (union pplib_power_state *)power_state_offset; non_clock_array_index = power_state->v2.nonClockInfoIndex; non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *) @@ -1496,12 +1499,15 @@ static int sumo_parse_power_table(struct radeon_device *rdev) } rdev->pm.dpm.ps[i].ps_priv = ps; k = 0; + idx = (u8 *)&power_state->v2.clockInfoIndex[0]; for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) { - clock_array_index = power_state->v2.clockInfoIndex[j]; + clock_array_index = idx[j]; if (k >= SUMO_MAX_HARDWARE_POWERLEVELS) break; + clock_info = (union pplib_clock_info *) - &clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize]; + ((u8 *)&clock_info_array->clockInfo[0] + + (clock_array_index * clock_info_array->ucEntrySize)); sumo_parse_pplib_clock_info(rdev, &rdev->pm.dpm.ps[i], k, clock_info); @@ -1530,6 +1536,20 @@ u32 sumo_convert_vid2_to_vid7(struct radeon_device *rdev, return vid_mapping_table->entries[vid_mapping_table->num_entries - 1].vid_7bit; } +u32 sumo_convert_vid7_to_vid2(struct radeon_device *rdev, + struct sumo_vid_mapping_table *vid_mapping_table, + u32 vid_7bit) +{ + u32 i; + + for (i = 0; i < vid_mapping_table->num_entries; i++) { + if (vid_mapping_table->entries[i].vid_7bit == vid_7bit) + return vid_mapping_table->entries[i].vid_2bit; + } + + return vid_mapping_table->entries[vid_mapping_table->num_entries - 1].vid_2bit; +} + static u16 sumo_convert_voltage_index_to_value(struct radeon_device *rdev, u32 vid_2bit) { @@ -1732,7 +1752,13 @@ int sumo_dpm_init(struct radeon_device *rdev) pi->enable_sclk_ds = true; pi->enable_dynamic_m3_arbiter = false; pi->enable_dynamic_patch_ps = true; - pi->enable_gfx_power_gating = true; + /* Some PALM chips don't seem to properly ungate gfx when UVD is in use; + * for now just disable gfx PG. + */ + if (rdev->family == CHIP_PALM) + pi->enable_gfx_power_gating = false; + else + pi->enable_gfx_power_gating = true; pi->enable_gfx_clock_gating = true; pi->enable_mg_clock_gating = true; pi->enable_auto_thermal_throttling = true; @@ -1743,6 +1769,10 @@ int sumo_dpm_init(struct radeon_device *rdev) sumo_construct_boot_and_acpi_state(rdev); + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = sumo_parse_power_table(rdev); if (ret) return ret; @@ -1778,7 +1808,7 @@ void sumo_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev struct seq_file *m) { struct sumo_power_info *pi = sumo_get_pi(rdev); - struct radeon_ps *rps = rdev->pm.dpm.current_ps; + struct radeon_ps *rps = &pi->current_rps; struct sumo_ps *ps = sumo_get_ps(rps); struct sumo_pl *pl; u32 current_index = @@ -1845,6 +1875,8 @@ int sumo_dpm_force_performance_level(struct radeon_device *rdev, return 0; if (level == RADEON_DPM_FORCED_LEVEL_HIGH) { + if (pi->enable_boost) + sumo_enable_boost(rdev, rps, false); sumo_power_level_enable(rdev, ps->num_levels - 1, true); sumo_set_forced_level(rdev, ps->num_levels - 1); sumo_set_forced_mode_enabled(rdev); @@ -1855,6 +1887,8 @@ int sumo_dpm_force_performance_level(struct radeon_device *rdev, sumo_set_forced_mode_enabled(rdev); sumo_set_forced_mode(rdev, false); } else if (level == RADEON_DPM_FORCED_LEVEL_LOW) { + if (pi->enable_boost) + sumo_enable_boost(rdev, rps, false); sumo_power_level_enable(rdev, 0, true); sumo_set_forced_level(rdev, 0); sumo_set_forced_mode_enabled(rdev); @@ -1868,6 +1902,8 @@ int sumo_dpm_force_performance_level(struct radeon_device *rdev, for (i = 0; i < ps->num_levels; i++) { sumo_power_level_enable(rdev, i, true); } + if (pi->enable_boost) + sumo_enable_boost(rdev, rps, true); } rdev->pm.dpm.forced_level = level; diff --git a/drivers/gpu/drm/radeon/sumo_dpm.h b/drivers/gpu/drm/radeon/sumo_dpm.h index 07dda299c78..db1ea32a907 100644 --- a/drivers/gpu/drm/radeon/sumo_dpm.h +++ b/drivers/gpu/drm/radeon/sumo_dpm.h @@ -202,6 +202,9 @@ void sumo_construct_vid_mapping_table(struct radeon_device *rdev, u32 sumo_convert_vid2_to_vid7(struct radeon_device *rdev, struct sumo_vid_mapping_table *vid_mapping_table, u32 vid_2bit); +u32 sumo_convert_vid7_to_vid2(struct radeon_device *rdev, + struct sumo_vid_mapping_table *vid_mapping_table, + u32 vid_7bit); u32 sumo_get_sleep_divider_from_id(u32 id); u32 sumo_get_sleep_divider_id_from_clock(struct radeon_device *rdev, u32 sclk, diff --git a/drivers/gpu/drm/radeon/sumo_smc.c b/drivers/gpu/drm/radeon/sumo_smc.c index 18abba5b581..fb081d2ae37 100644 --- a/drivers/gpu/drm/radeon/sumo_smc.c +++ b/drivers/gpu/drm/radeon/sumo_smc.c @@ -31,7 +31,6 @@ #define SUMO_SMU_SERVICE_ROUTINE_ALTVDDNB_NOTIFY 27 #define SUMO_SMU_SERVICE_ROUTINE_GFX_SRV_ID_20 20 -struct sumo_ps *sumo_get_ps(struct radeon_ps *rps); struct sumo_power_info *sumo_get_pi(struct radeon_device *rdev); static void sumo_send_msg_to_smu(struct radeon_device *rdev, u32 id) diff --git a/drivers/gpu/drm/radeon/trinity_dpm.c b/drivers/gpu/drm/radeon/trinity_dpm.c index a1eb5f59939..32e50be9c4a 100644 --- a/drivers/gpu/drm/radeon/trinity_dpm.c +++ b/drivers/gpu/drm/radeon/trinity_dpm.c @@ -342,14 +342,14 @@ static void trinity_apply_state_adjust_rules(struct radeon_device *rdev, struct radeon_ps *new_rps, struct radeon_ps *old_rps); -struct trinity_ps *trinity_get_ps(struct radeon_ps *rps) +static struct trinity_ps *trinity_get_ps(struct radeon_ps *rps) { struct trinity_ps *ps = rps->ps_priv; return ps; } -struct trinity_power_info *trinity_get_pi(struct radeon_device *rdev) +static struct trinity_power_info *trinity_get_pi(struct radeon_device *rdev) { struct trinity_power_info *pi = rdev->pm.dpm.priv; @@ -1068,10 +1068,20 @@ static void trinity_update_requested_ps(struct radeon_device *rdev, pi->requested_rps.ps_priv = &pi->requested_ps; } +void trinity_dpm_enable_bapm(struct radeon_device *rdev, bool enable) +{ + struct trinity_power_info *pi = trinity_get_pi(rdev); + + if (pi->enable_bapm) { + trinity_acquire_mutex(rdev); + trinity_dpm_bapm_enable(rdev, enable); + trinity_release_mutex(rdev); + } +} + int trinity_dpm_enable(struct radeon_device *rdev) { struct trinity_power_info *pi = trinity_get_pi(rdev); - int ret; trinity_acquire_mutex(rdev); @@ -1080,7 +1090,6 @@ int trinity_dpm_enable(struct radeon_device *rdev) return -EINVAL; } - trinity_enable_clock_power_gating(rdev); trinity_program_bootup_state(rdev); sumo_program_vc(rdev, 0x00C00033); trinity_start_am(rdev); @@ -1091,8 +1100,21 @@ int trinity_dpm_enable(struct radeon_device *rdev) trinity_program_sclk_dpm(rdev); trinity_start_dpm(rdev); trinity_wait_for_dpm_enabled(rdev); + trinity_dpm_bapm_enable(rdev, false); trinity_release_mutex(rdev); + trinity_update_current_ps(rdev, rdev->pm.dpm.boot_ps); + + return 0; +} + +int trinity_dpm_late_enable(struct radeon_device *rdev) +{ + int ret; + + trinity_acquire_mutex(rdev); + trinity_enable_clock_power_gating(rdev); + if (rdev->irq.installed && r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) { ret = trinity_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX); @@ -1103,8 +1125,7 @@ int trinity_dpm_enable(struct radeon_device *rdev) rdev->irq.dpm_thermal = true; radeon_irq_set(rdev); } - - trinity_update_current_ps(rdev, rdev->pm.dpm.boot_ps); + trinity_release_mutex(rdev); return 0; } @@ -1116,6 +1137,7 @@ void trinity_dpm_disable(struct radeon_device *rdev) trinity_release_mutex(rdev); return; } + trinity_dpm_bapm_enable(rdev, false); trinity_disable_clock_power_gating(rdev); sumo_clear_vc(rdev); trinity_wait_for_level_0(rdev); @@ -1212,6 +1234,8 @@ int trinity_dpm_set_power_state(struct radeon_device *rdev) trinity_acquire_mutex(rdev); if (pi->enable_dpm) { + if (pi->enable_bapm) + trinity_dpm_bapm_enable(rdev, rdev->pm.dpm.ac_power); trinity_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps); trinity_enable_power_level_0(rdev); trinity_force_level_0(rdev); @@ -1221,7 +1245,6 @@ int trinity_dpm_set_power_state(struct radeon_device *rdev) trinity_force_level_0(rdev); trinity_unforce_levels(rdev); trinity_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps); - rdev->pm.dpm.forced_level = RADEON_DPM_FORCED_LEVEL_AUTO; } trinity_release_mutex(rdev); @@ -1671,10 +1694,8 @@ static int trinity_parse_power_table(struct radeon_device *rdev) if (!rdev->pm.dpm.ps) return -ENOMEM; power_state_offset = (u8 *)state_array->states; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < state_array->ucNumEntries; i++) { + u8 *idx; power_state = (union pplib_power_state *)power_state_offset; non_clock_array_index = power_state->v2.nonClockInfoIndex; non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *) @@ -1688,14 +1709,16 @@ static int trinity_parse_power_table(struct radeon_device *rdev) } rdev->pm.dpm.ps[i].ps_priv = ps; k = 0; + idx = (u8 *)&power_state->v2.clockInfoIndex[0]; for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) { - clock_array_index = power_state->v2.clockInfoIndex[j]; + clock_array_index = idx[j]; if (clock_array_index >= clock_info_array->ucNumEntries) continue; if (k >= SUMO_MAX_HARDWARE_POWERLEVELS) break; clock_info = (union pplib_clock_info *) - &clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize]; + ((u8 *)&clock_info_array->clockInfo[0] + + (clock_array_index * clock_info_array->ucEntrySize)); trinity_parse_pplib_clock_info(rdev, &rdev->pm.dpm.ps[i], k, clock_info); @@ -1851,13 +1874,23 @@ int trinity_dpm_init(struct radeon_device *rdev) for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++) pi->at[i] = TRINITY_AT_DFLT; + /* There are stability issues reported on with + * bapm enabled when switching between AC and battery + * power. At the same time, some MSI boards hang + * if it's not enabled and dpm is enabled. Just enable + * it for MSI boards right now. + */ + if (rdev->pdev->subsystem_vendor == 0x1462) + pi->enable_bapm = true; + else + pi->enable_bapm = false; pi->enable_nbps_policy = true; pi->enable_sclk_ds = true; pi->enable_gfx_power_gating = true; pi->enable_gfx_clock_gating = true; - pi->enable_mg_clock_gating = true; - pi->enable_gfx_dynamic_mgpg = true; /* ??? */ - pi->override_dynamic_mgpg = true; + pi->enable_mg_clock_gating = false; + pi->enable_gfx_dynamic_mgpg = false; + pi->override_dynamic_mgpg = false; pi->enable_auto_thermal_throttling = true; pi->voltage_drop_in_dce = false; /* need to restructure dpm/modeset interaction */ pi->uvd_dpm = true; /* ??? */ @@ -1868,6 +1901,10 @@ int trinity_dpm_init(struct radeon_device *rdev) trinity_construct_boot_state(rdev); + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = trinity_parse_power_table(rdev); if (ret) return ret; @@ -1899,7 +1936,8 @@ void trinity_dpm_print_power_state(struct radeon_device *rdev, void trinity_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev, struct seq_file *m) { - struct radeon_ps *rps = rdev->pm.dpm.current_ps; + struct trinity_power_info *pi = trinity_get_pi(rdev); + struct radeon_ps *rps = &pi->current_rps; struct trinity_ps *ps = trinity_get_ps(rps); struct trinity_pl *pl; u32 current_index = diff --git a/drivers/gpu/drm/radeon/trinity_dpm.h b/drivers/gpu/drm/radeon/trinity_dpm.h index e82df071f8b..c261657750c 100644 --- a/drivers/gpu/drm/radeon/trinity_dpm.h +++ b/drivers/gpu/drm/radeon/trinity_dpm.h @@ -108,6 +108,7 @@ struct trinity_power_info { bool enable_auto_thermal_throttling; bool enable_dpm; bool enable_sclk_ds; + bool enable_bapm; bool uvd_dpm; struct radeon_ps current_rps; struct trinity_ps current_ps; @@ -118,6 +119,7 @@ struct trinity_power_info { #define TRINITY_AT_DFLT 30 /* trinity_smc.c */ +int trinity_dpm_bapm_enable(struct radeon_device *rdev, bool enable); int trinity_dpm_config(struct radeon_device *rdev, bool enable); int trinity_uvd_dpm_config(struct radeon_device *rdev); int trinity_dpm_force_state(struct radeon_device *rdev, u32 n); diff --git a/drivers/gpu/drm/radeon/trinity_smc.c b/drivers/gpu/drm/radeon/trinity_smc.c index a42d89f1830..99dd0455334 100644 --- a/drivers/gpu/drm/radeon/trinity_smc.c +++ b/drivers/gpu/drm/radeon/trinity_smc.c @@ -27,9 +27,6 @@ #include "trinity_dpm.h" #include "ppsmc.h" -struct trinity_ps *trinity_get_ps(struct radeon_ps *rps); -struct trinity_power_info *trinity_get_pi(struct radeon_device *rdev); - static int trinity_notify_message_to_smu(struct radeon_device *rdev, u32 id) { int i; @@ -56,6 +53,14 @@ static int trinity_notify_message_to_smu(struct radeon_device *rdev, u32 id) return 0; } +int trinity_dpm_bapm_enable(struct radeon_device *rdev, bool enable) +{ + if (enable) + return trinity_notify_message_to_smu(rdev, PPSMC_MSG_EnableBAPM); + else + return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DisableBAPM); +} + int trinity_dpm_config(struct radeon_device *rdev, bool enable) { if (enable) diff --git a/drivers/gpu/drm/radeon/uvd_v1_0.c b/drivers/gpu/drm/radeon/uvd_v1_0.c new file mode 100644 index 00000000000..be42c812520 --- /dev/null +++ b/drivers/gpu/drm/radeon/uvd_v1_0.c @@ -0,0 +1,444 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Christian König <christian.koenig@amd.com> + */ + +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "r600d.h" + +/** + * uvd_v1_0_get_rptr - get read pointer + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Returns the current hardware read pointer + */ +uint32_t uvd_v1_0_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + return RREG32(UVD_RBC_RB_RPTR); +} + +/** + * uvd_v1_0_get_wptr - get write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Returns the current hardware write pointer + */ +uint32_t uvd_v1_0_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + return RREG32(UVD_RBC_RB_WPTR); +} + +/** + * uvd_v1_0_set_wptr - set write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Commits the write pointer to the hardware + */ +void uvd_v1_0_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + WREG32(UVD_RBC_RB_WPTR, ring->wptr); +} + +/** + * uvd_v1_0_init - start and test UVD block + * + * @rdev: radeon_device pointer + * + * Initialize the hardware, boot up the VCPU and do some testing + */ +int uvd_v1_0_init(struct radeon_device *rdev) +{ + struct radeon_ring *ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; + uint32_t tmp; + int r; + + /* raise clocks while booting up the VCPU */ + if (rdev->family < CHIP_RV740) + radeon_set_uvd_clocks(rdev, 10000, 10000); + else + radeon_set_uvd_clocks(rdev, 53300, 40000); + + r = uvd_v1_0_start(rdev); + if (r) + goto done; + + ring->ready = true; + r = radeon_ring_test(rdev, R600_RING_TYPE_UVD_INDEX, ring); + if (r) { + ring->ready = false; + goto done; + } + + r = radeon_ring_lock(rdev, ring, 10); + if (r) { + DRM_ERROR("radeon: ring failed to lock UVD ring (%d).\n", r); + goto done; + } + + tmp = PACKET0(UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL, 0); + radeon_ring_write(ring, tmp); + radeon_ring_write(ring, 0xFFFFF); + + tmp = PACKET0(UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL, 0); + radeon_ring_write(ring, tmp); + radeon_ring_write(ring, 0xFFFFF); + + tmp = PACKET0(UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL, 0); + radeon_ring_write(ring, tmp); + radeon_ring_write(ring, 0xFFFFF); + + /* Clear timeout status bits */ + radeon_ring_write(ring, PACKET0(UVD_SEMA_TIMEOUT_STATUS, 0)); + radeon_ring_write(ring, 0x8); + + radeon_ring_write(ring, PACKET0(UVD_SEMA_CNTL, 0)); + radeon_ring_write(ring, 3); + + radeon_ring_unlock_commit(rdev, ring); + +done: + /* lower clocks again */ + radeon_set_uvd_clocks(rdev, 0, 0); + + if (!r) + DRM_INFO("UVD initialized successfully.\n"); + + return r; +} + +/** + * uvd_v1_0_fini - stop the hardware block + * + * @rdev: radeon_device pointer + * + * Stop the UVD block, mark ring as not ready any more + */ +void uvd_v1_0_fini(struct radeon_device *rdev) +{ + struct radeon_ring *ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; + + uvd_v1_0_stop(rdev); + ring->ready = false; +} + +/** + * uvd_v1_0_start - start UVD block + * + * @rdev: radeon_device pointer + * + * Setup and start the UVD block + */ +int uvd_v1_0_start(struct radeon_device *rdev) +{ + struct radeon_ring *ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX]; + uint32_t rb_bufsz; + int i, j, r; + + /* disable byte swapping */ + u32 lmi_swap_cntl = 0; + u32 mp_swap_cntl = 0; + + /* disable clock gating */ + WREG32(UVD_CGC_GATE, 0); + + /* disable interupt */ + WREG32_P(UVD_MASTINT_EN, 0, ~(1 << 1)); + + /* Stall UMC and register bus before resetting VCPU */ + WREG32_P(UVD_LMI_CTRL2, 1 << 8, ~(1 << 8)); + WREG32_P(UVD_RB_ARB_CTRL, 1 << 3, ~(1 << 3)); + mdelay(1); + + /* put LMI, VCPU, RBC etc... into reset */ + WREG32(UVD_SOFT_RESET, LMI_SOFT_RESET | VCPU_SOFT_RESET | + LBSI_SOFT_RESET | RBC_SOFT_RESET | CSM_SOFT_RESET | + CXW_SOFT_RESET | TAP_SOFT_RESET | LMI_UMC_SOFT_RESET); + mdelay(5); + + /* take UVD block out of reset */ + WREG32_P(SRBM_SOFT_RESET, 0, ~SOFT_RESET_UVD); + mdelay(5); + + /* initialize UVD memory controller */ + WREG32(UVD_LMI_CTRL, 0x40 | (1 << 8) | (1 << 13) | + (1 << 21) | (1 << 9) | (1 << 20)); + +#ifdef __BIG_ENDIAN + /* swap (8 in 32) RB and IB */ + lmi_swap_cntl = 0xa; + mp_swap_cntl = 0; +#endif + WREG32(UVD_LMI_SWAP_CNTL, lmi_swap_cntl); + WREG32(UVD_MP_SWAP_CNTL, mp_swap_cntl); + + WREG32(UVD_MPC_SET_MUXA0, 0x40c2040); + WREG32(UVD_MPC_SET_MUXA1, 0x0); + WREG32(UVD_MPC_SET_MUXB0, 0x40c2040); + WREG32(UVD_MPC_SET_MUXB1, 0x0); + WREG32(UVD_MPC_SET_ALU, 0); + WREG32(UVD_MPC_SET_MUX, 0x88); + + /* take all subblocks out of reset, except VCPU */ + WREG32(UVD_SOFT_RESET, VCPU_SOFT_RESET); + mdelay(5); + + /* enable VCPU clock */ + WREG32(UVD_VCPU_CNTL, 1 << 9); + + /* enable UMC */ + WREG32_P(UVD_LMI_CTRL2, 0, ~(1 << 8)); + + /* boot up the VCPU */ + WREG32(UVD_SOFT_RESET, 0); + mdelay(10); + + WREG32_P(UVD_RB_ARB_CTRL, 0, ~(1 << 3)); + + for (i = 0; i < 10; ++i) { + uint32_t status; + for (j = 0; j < 100; ++j) { + status = RREG32(UVD_STATUS); + if (status & 2) + break; + mdelay(10); + } + r = 0; + if (status & 2) + break; + + DRM_ERROR("UVD not responding, trying to reset the VCPU!!!\n"); + WREG32_P(UVD_SOFT_RESET, VCPU_SOFT_RESET, ~VCPU_SOFT_RESET); + mdelay(10); + WREG32_P(UVD_SOFT_RESET, 0, ~VCPU_SOFT_RESET); + mdelay(10); + r = -1; + } + + if (r) { + DRM_ERROR("UVD not responding, giving up!!!\n"); + return r; + } + + /* enable interupt */ + WREG32_P(UVD_MASTINT_EN, 3<<1, ~(3 << 1)); + + /* force RBC into idle state */ + WREG32(UVD_RBC_RB_CNTL, 0x11010101); + + /* Set the write pointer delay */ + WREG32(UVD_RBC_RB_WPTR_CNTL, 0); + + /* programm the 4GB memory segment for rptr and ring buffer */ + WREG32(UVD_LMI_EXT40_ADDR, upper_32_bits(ring->gpu_addr) | + (0x7 << 16) | (0x1 << 31)); + + /* Initialize the ring buffer's read and write pointers */ + WREG32(UVD_RBC_RB_RPTR, 0x0); + + ring->wptr = RREG32(UVD_RBC_RB_RPTR); + WREG32(UVD_RBC_RB_WPTR, ring->wptr); + + /* set the ring address */ + WREG32(UVD_RBC_RB_BASE, ring->gpu_addr); + + /* Set ring buffer size */ + rb_bufsz = order_base_2(ring->ring_size); + rb_bufsz = (0x1 << 8) | rb_bufsz; + WREG32_P(UVD_RBC_RB_CNTL, rb_bufsz, ~0x11f1f); + + return 0; +} + +/** + * uvd_v1_0_stop - stop UVD block + * + * @rdev: radeon_device pointer + * + * stop the UVD block + */ +void uvd_v1_0_stop(struct radeon_device *rdev) +{ + /* force RBC into idle state */ + WREG32(UVD_RBC_RB_CNTL, 0x11010101); + + /* Stall UMC and register bus before resetting VCPU */ + WREG32_P(UVD_LMI_CTRL2, 1 << 8, ~(1 << 8)); + WREG32_P(UVD_RB_ARB_CTRL, 1 << 3, ~(1 << 3)); + mdelay(1); + + /* put VCPU into reset */ + WREG32(UVD_SOFT_RESET, VCPU_SOFT_RESET); + mdelay(5); + + /* disable VCPU clock */ + WREG32(UVD_VCPU_CNTL, 0x0); + + /* Unstall UMC and register bus */ + WREG32_P(UVD_LMI_CTRL2, 0, ~(1 << 8)); + WREG32_P(UVD_RB_ARB_CTRL, 0, ~(1 << 3)); +} + +/** + * uvd_v1_0_ring_test - register write test + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Test if we can successfully write to the context register + */ +int uvd_v1_0_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) +{ + uint32_t tmp = 0; + unsigned i; + int r; + + WREG32(UVD_CONTEXT_ID, 0xCAFEDEAD); + r = radeon_ring_lock(rdev, ring, 3); + if (r) { + DRM_ERROR("radeon: cp failed to lock ring %d (%d).\n", + ring->idx, r); + return r; + } + radeon_ring_write(ring, PACKET0(UVD_CONTEXT_ID, 0)); + radeon_ring_write(ring, 0xDEADBEEF); + radeon_ring_unlock_commit(rdev, ring); + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = RREG32(UVD_CONTEXT_ID); + if (tmp == 0xDEADBEEF) + break; + DRM_UDELAY(1); + } + + if (i < rdev->usec_timeout) { + DRM_INFO("ring test on %d succeeded in %d usecs\n", + ring->idx, i); + } else { + DRM_ERROR("radeon: ring %d test failed (0x%08X)\n", + ring->idx, tmp); + r = -EINVAL; + } + return r; +} + +/** + * uvd_v1_0_semaphore_emit - emit semaphore command + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * @semaphore: semaphore to emit commands for + * @emit_wait: true if we should emit a wait command + * + * Emit a semaphore command (either wait or signal) to the UVD ring. + */ +bool uvd_v1_0_semaphore_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait) +{ + uint64_t addr = semaphore->gpu_addr; + + radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_LOW, 0)); + radeon_ring_write(ring, (addr >> 3) & 0x000FFFFF); + + radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_HIGH, 0)); + radeon_ring_write(ring, (addr >> 23) & 0x000FFFFF); + + radeon_ring_write(ring, PACKET0(UVD_SEMA_CMD, 0)); + radeon_ring_write(ring, emit_wait ? 1 : 0); + + return true; +} + +/** + * uvd_v1_0_ib_execute - execute indirect buffer + * + * @rdev: radeon_device pointer + * @ib: indirect buffer to execute + * + * Write ring commands to execute the indirect buffer + */ +void uvd_v1_0_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) +{ + struct radeon_ring *ring = &rdev->ring[ib->ring]; + + radeon_ring_write(ring, PACKET0(UVD_RBC_IB_BASE, 0)); + radeon_ring_write(ring, ib->gpu_addr); + radeon_ring_write(ring, PACKET0(UVD_RBC_IB_SIZE, 0)); + radeon_ring_write(ring, ib->length_dw); +} + +/** + * uvd_v1_0_ib_test - test ib execution + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Test if we can successfully execute an IB + */ +int uvd_v1_0_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) +{ + struct radeon_fence *fence = NULL; + int r; + + if (rdev->family < CHIP_RV740) + r = radeon_set_uvd_clocks(rdev, 10000, 10000); + else + r = radeon_set_uvd_clocks(rdev, 53300, 40000); + if (r) { + DRM_ERROR("radeon: failed to raise UVD clocks (%d).\n", r); + return r; + } + + r = radeon_uvd_get_create_msg(rdev, ring->idx, 1, NULL); + if (r) { + DRM_ERROR("radeon: failed to get create msg (%d).\n", r); + goto error; + } + + r = radeon_uvd_get_destroy_msg(rdev, ring->idx, 1, &fence); + if (r) { + DRM_ERROR("radeon: failed to get destroy ib (%d).\n", r); + goto error; + } + + r = radeon_fence_wait(fence, false); + if (r) { + DRM_ERROR("radeon: fence wait failed (%d).\n", r); + goto error; + } + DRM_INFO("ib test on ring %d succeeded\n", ring->idx); +error: + radeon_fence_unref(&fence); + radeon_set_uvd_clocks(rdev, 0, 0); + return r; +} diff --git a/drivers/gpu/drm/radeon/uvd_v2_2.c b/drivers/gpu/drm/radeon/uvd_v2_2.c new file mode 100644 index 00000000000..8bfdadd5659 --- /dev/null +++ b/drivers/gpu/drm/radeon/uvd_v2_2.c @@ -0,0 +1,165 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Christian König <christian.koenig@amd.com> + */ + +#include <linux/firmware.h> +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "rv770d.h" + +/** + * uvd_v2_2_fence_emit - emit an fence & trap command + * + * @rdev: radeon_device pointer + * @fence: fence to emit + * + * Write a fence and a trap command to the ring. + */ +void uvd_v2_2_fence_emit(struct radeon_device *rdev, + struct radeon_fence *fence) +{ + struct radeon_ring *ring = &rdev->ring[fence->ring]; + uint64_t addr = rdev->fence_drv[fence->ring].gpu_addr; + + radeon_ring_write(ring, PACKET0(UVD_CONTEXT_ID, 0)); + radeon_ring_write(ring, fence->seq); + radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA0, 0)); + radeon_ring_write(ring, lower_32_bits(addr)); + radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA1, 0)); + radeon_ring_write(ring, upper_32_bits(addr) & 0xff); + radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_CMD, 0)); + radeon_ring_write(ring, 0); + + radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA0, 0)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA1, 0)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_CMD, 0)); + radeon_ring_write(ring, 2); +} + +/** + * uvd_v2_2_resume - memory controller programming + * + * @rdev: radeon_device pointer + * + * Let the UVD memory controller know it's offsets + */ +int uvd_v2_2_resume(struct radeon_device *rdev) +{ + uint64_t addr; + uint32_t chip_id, size; + int r; + + r = radeon_uvd_resume(rdev); + if (r) + return r; + + /* programm the VCPU memory controller bits 0-27 */ + addr = rdev->uvd.gpu_addr >> 3; + size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 4) >> 3; + WREG32(UVD_VCPU_CACHE_OFFSET0, addr); + WREG32(UVD_VCPU_CACHE_SIZE0, size); + + addr += size; + size = RADEON_UVD_STACK_SIZE >> 3; + WREG32(UVD_VCPU_CACHE_OFFSET1, addr); + WREG32(UVD_VCPU_CACHE_SIZE1, size); + + addr += size; + size = RADEON_UVD_HEAP_SIZE >> 3; + WREG32(UVD_VCPU_CACHE_OFFSET2, addr); + WREG32(UVD_VCPU_CACHE_SIZE2, size); + + /* bits 28-31 */ + addr = (rdev->uvd.gpu_addr >> 28) & 0xF; + WREG32(UVD_LMI_ADDR_EXT, (addr << 12) | (addr << 0)); + + /* bits 32-39 */ + addr = (rdev->uvd.gpu_addr >> 32) & 0xFF; + WREG32(UVD_LMI_EXT40_ADDR, addr | (0x9 << 16) | (0x1 << 31)); + + /* tell firmware which hardware it is running on */ + switch (rdev->family) { + default: + return -EINVAL; + case CHIP_RV710: + chip_id = 0x01000005; + break; + case CHIP_RV730: + chip_id = 0x01000006; + break; + case CHIP_RV740: + chip_id = 0x01000007; + break; + case CHIP_CYPRESS: + case CHIP_HEMLOCK: + chip_id = 0x01000008; + break; + case CHIP_JUNIPER: + chip_id = 0x01000009; + break; + case CHIP_REDWOOD: + chip_id = 0x0100000a; + break; + case CHIP_CEDAR: + chip_id = 0x0100000b; + break; + case CHIP_SUMO: + case CHIP_SUMO2: + chip_id = 0x0100000c; + break; + case CHIP_PALM: + chip_id = 0x0100000e; + break; + case CHIP_CAYMAN: + chip_id = 0x0100000f; + break; + case CHIP_BARTS: + chip_id = 0x01000010; + break; + case CHIP_TURKS: + chip_id = 0x01000011; + break; + case CHIP_CAICOS: + chip_id = 0x01000012; + break; + case CHIP_TAHITI: + chip_id = 0x01000014; + break; + case CHIP_VERDE: + chip_id = 0x01000015; + break; + case CHIP_PITCAIRN: + case CHIP_OLAND: + chip_id = 0x01000016; + break; + case CHIP_ARUBA: + chip_id = 0x01000017; + break; + } + WREG32(UVD_VCPU_CHIP_ID, chip_id); + + return 0; +} diff --git a/drivers/gpu/drm/radeon/uvd_v3_1.c b/drivers/gpu/drm/radeon/uvd_v3_1.c new file mode 100644 index 00000000000..d722db2cf34 --- /dev/null +++ b/drivers/gpu/drm/radeon/uvd_v3_1.c @@ -0,0 +1,57 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Christian König <christian.koenig@amd.com> + */ + +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "nid.h" + +/** + * uvd_v3_1_semaphore_emit - emit semaphore command + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * @semaphore: semaphore to emit commands for + * @emit_wait: true if we should emit a wait command + * + * Emit a semaphore command (either wait or signal) to the UVD ring. + */ +bool uvd_v3_1_semaphore_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait) +{ + uint64_t addr = semaphore->gpu_addr; + + radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_LOW, 0)); + radeon_ring_write(ring, (addr >> 3) & 0x000FFFFF); + + radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_HIGH, 0)); + radeon_ring_write(ring, (addr >> 23) & 0x000FFFFF); + + radeon_ring_write(ring, PACKET0(UVD_SEMA_CMD, 0)); + radeon_ring_write(ring, 0x80 | (emit_wait ? 1 : 0)); + + return true; +} diff --git a/drivers/gpu/drm/radeon/uvd_v4_2.c b/drivers/gpu/drm/radeon/uvd_v4_2.c new file mode 100644 index 00000000000..d04d5073eef --- /dev/null +++ b/drivers/gpu/drm/radeon/uvd_v4_2.c @@ -0,0 +1,68 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Christian König <christian.koenig@amd.com> + */ + +#include <linux/firmware.h> +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "cikd.h" + +/** + * uvd_v4_2_resume - memory controller programming + * + * @rdev: radeon_device pointer + * + * Let the UVD memory controller know it's offsets + */ +int uvd_v4_2_resume(struct radeon_device *rdev) +{ + uint64_t addr; + uint32_t size; + + /* programm the VCPU memory controller bits 0-27 */ + addr = rdev->uvd.gpu_addr >> 3; + size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 4) >> 3; + WREG32(UVD_VCPU_CACHE_OFFSET0, addr); + WREG32(UVD_VCPU_CACHE_SIZE0, size); + + addr += size; + size = RADEON_UVD_STACK_SIZE >> 3; + WREG32(UVD_VCPU_CACHE_OFFSET1, addr); + WREG32(UVD_VCPU_CACHE_SIZE1, size); + + addr += size; + size = RADEON_UVD_HEAP_SIZE >> 3; + WREG32(UVD_VCPU_CACHE_OFFSET2, addr); + WREG32(UVD_VCPU_CACHE_SIZE2, size); + + /* bits 28-31 */ + addr = (rdev->uvd.gpu_addr >> 28) & 0xF; + WREG32(UVD_LMI_ADDR_EXT, (addr << 12) | (addr << 0)); + + /* bits 32-39 */ + addr = (rdev->uvd.gpu_addr >> 32) & 0xFF; + WREG32(UVD_LMI_EXT40_ADDR, addr | (0x9 << 16) | (0x1 << 31)); + + return 0; +} diff --git a/drivers/gpu/drm/radeon/vce_v1_0.c b/drivers/gpu/drm/radeon/vce_v1_0.c new file mode 100644 index 00000000000..b44d9c842f7 --- /dev/null +++ b/drivers/gpu/drm/radeon/vce_v1_0.c @@ -0,0 +1,187 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * Authors: Christian König <christian.koenig@amd.com> + */ + +#include <linux/firmware.h> +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "sid.h" + +/** + * vce_v1_0_get_rptr - get read pointer + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Returns the current hardware read pointer + */ +uint32_t vce_v1_0_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + if (ring->idx == TN_RING_TYPE_VCE1_INDEX) + return RREG32(VCE_RB_RPTR); + else + return RREG32(VCE_RB_RPTR2); +} + +/** + * vce_v1_0_get_wptr - get write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Returns the current hardware write pointer + */ +uint32_t vce_v1_0_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + if (ring->idx == TN_RING_TYPE_VCE1_INDEX) + return RREG32(VCE_RB_WPTR); + else + return RREG32(VCE_RB_WPTR2); +} + +/** + * vce_v1_0_set_wptr - set write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Commits the write pointer to the hardware + */ +void vce_v1_0_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + if (ring->idx == TN_RING_TYPE_VCE1_INDEX) + WREG32(VCE_RB_WPTR, ring->wptr); + else + WREG32(VCE_RB_WPTR2, ring->wptr); +} + +/** + * vce_v1_0_start - start VCE block + * + * @rdev: radeon_device pointer + * + * Setup and start the VCE block + */ +int vce_v1_0_start(struct radeon_device *rdev) +{ + struct radeon_ring *ring; + int i, j, r; + + /* set BUSY flag */ + WREG32_P(VCE_STATUS, 1, ~1); + + ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX]; + WREG32(VCE_RB_RPTR, ring->wptr); + WREG32(VCE_RB_WPTR, ring->wptr); + WREG32(VCE_RB_BASE_LO, ring->gpu_addr); + WREG32(VCE_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); + WREG32(VCE_RB_SIZE, ring->ring_size / 4); + + ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX]; + WREG32(VCE_RB_RPTR2, ring->wptr); + WREG32(VCE_RB_WPTR2, ring->wptr); + WREG32(VCE_RB_BASE_LO2, ring->gpu_addr); + WREG32(VCE_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); + WREG32(VCE_RB_SIZE2, ring->ring_size / 4); + + WREG32_P(VCE_VCPU_CNTL, VCE_CLK_EN, ~VCE_CLK_EN); + + WREG32_P(VCE_SOFT_RESET, + VCE_ECPU_SOFT_RESET | + VCE_FME_SOFT_RESET, ~( + VCE_ECPU_SOFT_RESET | + VCE_FME_SOFT_RESET)); + + mdelay(100); + + WREG32_P(VCE_SOFT_RESET, 0, ~( + VCE_ECPU_SOFT_RESET | + VCE_FME_SOFT_RESET)); + + for (i = 0; i < 10; ++i) { + uint32_t status; + for (j = 0; j < 100; ++j) { + status = RREG32(VCE_STATUS); + if (status & 2) + break; + mdelay(10); + } + r = 0; + if (status & 2) + break; + + DRM_ERROR("VCE not responding, trying to reset the ECPU!!!\n"); + WREG32_P(VCE_SOFT_RESET, VCE_ECPU_SOFT_RESET, ~VCE_ECPU_SOFT_RESET); + mdelay(10); + WREG32_P(VCE_SOFT_RESET, 0, ~VCE_ECPU_SOFT_RESET); + mdelay(10); + r = -1; + } + + /* clear BUSY flag */ + WREG32_P(VCE_STATUS, 0, ~1); + + if (r) { + DRM_ERROR("VCE not responding, giving up!!!\n"); + return r; + } + + return 0; +} + +int vce_v1_0_init(struct radeon_device *rdev) +{ + struct radeon_ring *ring; + int r; + + r = vce_v1_0_start(rdev); + if (r) + return r; + + ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX]; + ring->ready = true; + r = radeon_ring_test(rdev, TN_RING_TYPE_VCE1_INDEX, ring); + if (r) { + ring->ready = false; + return r; + } + + ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX]; + ring->ready = true; + r = radeon_ring_test(rdev, TN_RING_TYPE_VCE2_INDEX, ring); + if (r) { + ring->ready = false; + return r; + } + + DRM_INFO("VCE initialized successfully.\n"); + + return 0; +} diff --git a/drivers/gpu/drm/radeon/vce_v2_0.c b/drivers/gpu/drm/radeon/vce_v2_0.c new file mode 100644 index 00000000000..1ac7bb825a1 --- /dev/null +++ b/drivers/gpu/drm/radeon/vce_v2_0.c @@ -0,0 +1,181 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * Authors: Christian König <christian.koenig@amd.com> + */ + +#include <linux/firmware.h> +#include <drm/drmP.h> +#include "radeon.h" +#include "radeon_asic.h" +#include "cikd.h" + +static void vce_v2_0_set_sw_cg(struct radeon_device *rdev, bool gated) +{ + u32 tmp; + + if (gated) { + tmp = RREG32(VCE_CLOCK_GATING_B); + tmp |= 0xe70000; + WREG32(VCE_CLOCK_GATING_B, tmp); + + tmp = RREG32(VCE_UENC_CLOCK_GATING); + tmp |= 0xff000000; + WREG32(VCE_UENC_CLOCK_GATING, tmp); + + tmp = RREG32(VCE_UENC_REG_CLOCK_GATING); + tmp &= ~0x3fc; + WREG32(VCE_UENC_REG_CLOCK_GATING, tmp); + + WREG32(VCE_CGTT_CLK_OVERRIDE, 0); + } else { + tmp = RREG32(VCE_CLOCK_GATING_B); + tmp |= 0xe7; + tmp &= ~0xe70000; + WREG32(VCE_CLOCK_GATING_B, tmp); + + tmp = RREG32(VCE_UENC_CLOCK_GATING); + tmp |= 0x1fe000; + tmp &= ~0xff000000; + WREG32(VCE_UENC_CLOCK_GATING, tmp); + + tmp = RREG32(VCE_UENC_REG_CLOCK_GATING); + tmp |= 0x3fc; + WREG32(VCE_UENC_REG_CLOCK_GATING, tmp); + } +} + +static void vce_v2_0_set_dyn_cg(struct radeon_device *rdev, bool gated) +{ + u32 orig, tmp; + + tmp = RREG32(VCE_CLOCK_GATING_B); + tmp &= ~0x00060006; + if (gated) { + tmp |= 0xe10000; + } else { + tmp |= 0xe1; + tmp &= ~0xe10000; + } + WREG32(VCE_CLOCK_GATING_B, tmp); + + orig = tmp = RREG32(VCE_UENC_CLOCK_GATING); + tmp &= ~0x1fe000; + tmp &= ~0xff000000; + if (tmp != orig) + WREG32(VCE_UENC_CLOCK_GATING, tmp); + + orig = tmp = RREG32(VCE_UENC_REG_CLOCK_GATING); + tmp &= ~0x3fc; + if (tmp != orig) + WREG32(VCE_UENC_REG_CLOCK_GATING, tmp); + + if (gated) + WREG32(VCE_CGTT_CLK_OVERRIDE, 0); +} + +static void vce_v2_0_disable_cg(struct radeon_device *rdev) +{ + WREG32(VCE_CGTT_CLK_OVERRIDE, 7); +} + +void vce_v2_0_enable_mgcg(struct radeon_device *rdev, bool enable) +{ + bool sw_cg = false; + + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_VCE_MGCG)) { + if (sw_cg) + vce_v2_0_set_sw_cg(rdev, true); + else + vce_v2_0_set_dyn_cg(rdev, true); + } else { + vce_v2_0_disable_cg(rdev); + + if (sw_cg) + vce_v2_0_set_sw_cg(rdev, false); + else + vce_v2_0_set_dyn_cg(rdev, false); + } +} + +static void vce_v2_0_init_cg(struct radeon_device *rdev) +{ + u32 tmp; + + tmp = RREG32(VCE_CLOCK_GATING_A); + tmp &= ~(CGC_CLK_GATE_DLY_TIMER_MASK | CGC_CLK_GATER_OFF_DLY_TIMER_MASK); + tmp |= (CGC_CLK_GATE_DLY_TIMER(0) | CGC_CLK_GATER_OFF_DLY_TIMER(4)); + tmp |= CGC_UENC_WAIT_AWAKE; + WREG32(VCE_CLOCK_GATING_A, tmp); + + tmp = RREG32(VCE_UENC_CLOCK_GATING); + tmp &= ~(CLOCK_ON_DELAY_MASK | CLOCK_OFF_DELAY_MASK); + tmp |= (CLOCK_ON_DELAY(0) | CLOCK_OFF_DELAY(4)); + WREG32(VCE_UENC_CLOCK_GATING, tmp); + + tmp = RREG32(VCE_CLOCK_GATING_B); + tmp |= 0x10; + tmp &= ~0x100000; + WREG32(VCE_CLOCK_GATING_B, tmp); +} + +int vce_v2_0_resume(struct radeon_device *rdev) +{ + uint64_t addr = rdev->vce.gpu_addr; + uint32_t size; + + WREG32_P(VCE_CLOCK_GATING_A, 0, ~(1 << 16)); + WREG32_P(VCE_UENC_CLOCK_GATING, 0x1FF000, ~0xFF9FF000); + WREG32_P(VCE_UENC_REG_CLOCK_GATING, 0x3F, ~0x3F); + WREG32(VCE_CLOCK_GATING_B, 0xf7); + + WREG32(VCE_LMI_CTRL, 0x00398000); + WREG32_P(VCE_LMI_CACHE_CTRL, 0x0, ~0x1); + WREG32(VCE_LMI_SWAP_CNTL, 0); + WREG32(VCE_LMI_SWAP_CNTL1, 0); + WREG32(VCE_LMI_VM_CTRL, 0); + + size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size); + WREG32(VCE_VCPU_CACHE_OFFSET0, addr & 0x7fffffff); + WREG32(VCE_VCPU_CACHE_SIZE0, size); + + addr += size; + size = RADEON_VCE_STACK_SIZE; + WREG32(VCE_VCPU_CACHE_OFFSET1, addr & 0x7fffffff); + WREG32(VCE_VCPU_CACHE_SIZE1, size); + + addr += size; + size = RADEON_VCE_HEAP_SIZE; + WREG32(VCE_VCPU_CACHE_OFFSET2, addr & 0x7fffffff); + WREG32(VCE_VCPU_CACHE_SIZE2, size); + + WREG32_P(VCE_LMI_CTRL2, 0x0, ~0x100); + + WREG32_P(VCE_SYS_INT_EN, VCE_SYS_INT_TRAP_INTERRUPT_EN, + ~VCE_SYS_INT_TRAP_INTERRUPT_EN); + + vce_v2_0_init_cg(rdev); + + return 0; +} |