diff options
Diffstat (limited to 'drivers/gpu/drm/radeon/r600.c')
| -rw-r--r-- | drivers/gpu/drm/radeon/r600.c | 1686 |
1 files changed, 1117 insertions, 569 deletions
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c index 4f08e5e6ee9..3c69f58e46e 100644 --- a/drivers/gpu/drm/radeon/r600.c +++ b/drivers/gpu/drm/radeon/r600.c @@ -28,27 +28,16 @@ #include <linux/slab.h> #include <linux/seq_file.h> #include <linux/firmware.h> -#include <linux/platform_device.h> #include <linux/module.h> -#include "drmP.h" -#include "radeon_drm.h" +#include <drm/drmP.h> +#include <drm/radeon_drm.h> #include "radeon.h" #include "radeon_asic.h" #include "radeon_mode.h" #include "r600d.h" #include "atom.h" #include "avivod.h" - -#define PFP_UCODE_SIZE 576 -#define PM4_UCODE_SIZE 1792 -#define RLC_UCODE_SIZE 768 -#define R700_PFP_UCODE_SIZE 848 -#define R700_PM4_UCODE_SIZE 1360 -#define R700_RLC_UCODE_SIZE 1024 -#define EVERGREEN_PFP_UCODE_SIZE 1120 -#define EVERGREEN_PM4_UCODE_SIZE 1376 -#define EVERGREEN_RLC_UCODE_SIZE 768 -#define CAYMAN_RLC_UCODE_SIZE 1024 +#include "radeon_ucode.h" /* Firmware Names */ MODULE_FIRMWARE("radeon/R600_pfp.bin"); @@ -67,24 +56,32 @@ MODULE_FIRMWARE("radeon/RS780_pfp.bin"); MODULE_FIRMWARE("radeon/RS780_me.bin"); MODULE_FIRMWARE("radeon/RV770_pfp.bin"); MODULE_FIRMWARE("radeon/RV770_me.bin"); +MODULE_FIRMWARE("radeon/RV770_smc.bin"); MODULE_FIRMWARE("radeon/RV730_pfp.bin"); MODULE_FIRMWARE("radeon/RV730_me.bin"); +MODULE_FIRMWARE("radeon/RV730_smc.bin"); +MODULE_FIRMWARE("radeon/RV740_smc.bin"); MODULE_FIRMWARE("radeon/RV710_pfp.bin"); MODULE_FIRMWARE("radeon/RV710_me.bin"); +MODULE_FIRMWARE("radeon/RV710_smc.bin"); MODULE_FIRMWARE("radeon/R600_rlc.bin"); MODULE_FIRMWARE("radeon/R700_rlc.bin"); MODULE_FIRMWARE("radeon/CEDAR_pfp.bin"); MODULE_FIRMWARE("radeon/CEDAR_me.bin"); MODULE_FIRMWARE("radeon/CEDAR_rlc.bin"); +MODULE_FIRMWARE("radeon/CEDAR_smc.bin"); MODULE_FIRMWARE("radeon/REDWOOD_pfp.bin"); MODULE_FIRMWARE("radeon/REDWOOD_me.bin"); MODULE_FIRMWARE("radeon/REDWOOD_rlc.bin"); +MODULE_FIRMWARE("radeon/REDWOOD_smc.bin"); MODULE_FIRMWARE("radeon/JUNIPER_pfp.bin"); MODULE_FIRMWARE("radeon/JUNIPER_me.bin"); MODULE_FIRMWARE("radeon/JUNIPER_rlc.bin"); +MODULE_FIRMWARE("radeon/JUNIPER_smc.bin"); MODULE_FIRMWARE("radeon/CYPRESS_pfp.bin"); MODULE_FIRMWARE("radeon/CYPRESS_me.bin"); MODULE_FIRMWARE("radeon/CYPRESS_rlc.bin"); +MODULE_FIRMWARE("radeon/CYPRESS_smc.bin"); MODULE_FIRMWARE("radeon/PALM_pfp.bin"); MODULE_FIRMWARE("radeon/PALM_me.bin"); MODULE_FIRMWARE("radeon/SUMO_rlc.bin"); @@ -93,14 +90,93 @@ MODULE_FIRMWARE("radeon/SUMO_me.bin"); MODULE_FIRMWARE("radeon/SUMO2_pfp.bin"); MODULE_FIRMWARE("radeon/SUMO2_me.bin"); +static const u32 crtc_offsets[2] = +{ + 0, + AVIVO_D2CRTC_H_TOTAL - AVIVO_D1CRTC_H_TOTAL +}; + int r600_debugfs_mc_info_init(struct radeon_device *rdev); /* r600,rv610,rv630,rv620,rv635,rv670 */ int r600_mc_wait_for_idle(struct radeon_device *rdev); -void r600_gpu_init(struct radeon_device *rdev); +static void r600_gpu_init(struct radeon_device *rdev); 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 + * + * @rdev: radeon_device pointer + * + * Returns the reference clock used by the gfx engine + * (r6xx, IGPs, APUs). + */ +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) @@ -708,10 +784,19 @@ void r600_hpd_init(struct radeon_device *rdev) { struct drm_device *dev = rdev->ddev; struct drm_connector *connector; + unsigned enable = 0; list_for_each_entry(connector, &dev->mode_config.connector_list, head) { struct radeon_connector *radeon_connector = to_radeon_connector(connector); + if (connector->connector_type == DRM_MODE_CONNECTOR_eDP || + connector->connector_type == DRM_MODE_CONNECTOR_LVDS) { + /* don't try to enable hpd on eDP or LVDS avoid breaking the + * aux dp channel on imac and help (but not completely fix) + * https://bugzilla.redhat.com/show_bug.cgi?id=726143 + */ + continue; + } if (ASIC_IS_DCE3(rdev)) { u32 tmp = DC_HPDx_CONNECTION_TIMER(0x9c4) | DC_HPDx_RX_INT_TIMER(0xfa); if (ASIC_IS_DCE32(rdev)) @@ -720,28 +805,22 @@ void r600_hpd_init(struct radeon_device *rdev) switch (radeon_connector->hpd.hpd) { case RADEON_HPD_1: WREG32(DC_HPD1_CONTROL, tmp); - rdev->irq.hpd[0] = true; break; case RADEON_HPD_2: WREG32(DC_HPD2_CONTROL, tmp); - rdev->irq.hpd[1] = true; break; case RADEON_HPD_3: WREG32(DC_HPD3_CONTROL, tmp); - rdev->irq.hpd[2] = true; break; case RADEON_HPD_4: WREG32(DC_HPD4_CONTROL, tmp); - rdev->irq.hpd[3] = true; break; /* DCE 3.2 */ case RADEON_HPD_5: WREG32(DC_HPD5_CONTROL, tmp); - rdev->irq.hpd[4] = true; break; case RADEON_HPD_6: WREG32(DC_HPD6_CONTROL, tmp); - rdev->irq.hpd[5] = true; break; default: break; @@ -750,85 +829,73 @@ void r600_hpd_init(struct radeon_device *rdev) switch (radeon_connector->hpd.hpd) { case RADEON_HPD_1: WREG32(DC_HOT_PLUG_DETECT1_CONTROL, DC_HOT_PLUG_DETECTx_EN); - rdev->irq.hpd[0] = true; break; case RADEON_HPD_2: WREG32(DC_HOT_PLUG_DETECT2_CONTROL, DC_HOT_PLUG_DETECTx_EN); - rdev->irq.hpd[1] = true; break; case RADEON_HPD_3: WREG32(DC_HOT_PLUG_DETECT3_CONTROL, DC_HOT_PLUG_DETECTx_EN); - rdev->irq.hpd[2] = true; break; default: break; } } + enable |= 1 << radeon_connector->hpd.hpd; radeon_hpd_set_polarity(rdev, radeon_connector->hpd.hpd); } - if (rdev->irq.installed) - r600_irq_set(rdev); + radeon_irq_kms_enable_hpd(rdev, enable); } void r600_hpd_fini(struct radeon_device *rdev) { struct drm_device *dev = rdev->ddev; struct drm_connector *connector; + unsigned disable = 0; - if (ASIC_IS_DCE3(rdev)) { - list_for_each_entry(connector, &dev->mode_config.connector_list, head) { - struct radeon_connector *radeon_connector = to_radeon_connector(connector); + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + struct radeon_connector *radeon_connector = to_radeon_connector(connector); + if (ASIC_IS_DCE3(rdev)) { switch (radeon_connector->hpd.hpd) { case RADEON_HPD_1: WREG32(DC_HPD1_CONTROL, 0); - rdev->irq.hpd[0] = false; break; case RADEON_HPD_2: WREG32(DC_HPD2_CONTROL, 0); - rdev->irq.hpd[1] = false; break; case RADEON_HPD_3: WREG32(DC_HPD3_CONTROL, 0); - rdev->irq.hpd[2] = false; break; case RADEON_HPD_4: WREG32(DC_HPD4_CONTROL, 0); - rdev->irq.hpd[3] = false; break; /* DCE 3.2 */ case RADEON_HPD_5: WREG32(DC_HPD5_CONTROL, 0); - rdev->irq.hpd[4] = false; break; case RADEON_HPD_6: WREG32(DC_HPD6_CONTROL, 0); - rdev->irq.hpd[5] = false; break; default: break; } - } - } else { - list_for_each_entry(connector, &dev->mode_config.connector_list, head) { - struct radeon_connector *radeon_connector = to_radeon_connector(connector); + } else { switch (radeon_connector->hpd.hpd) { case RADEON_HPD_1: WREG32(DC_HOT_PLUG_DETECT1_CONTROL, 0); - rdev->irq.hpd[0] = false; break; case RADEON_HPD_2: WREG32(DC_HOT_PLUG_DETECT2_CONTROL, 0); - rdev->irq.hpd[1] = false; break; case RADEON_HPD_3: WREG32(DC_HOT_PLUG_DETECT3_CONTROL, 0); - rdev->irq.hpd[2] = false; break; default: break; } } + disable |= 1 << radeon_connector->hpd.hpd; } + radeon_irq_kms_disable_hpd(rdev, disable); } /* @@ -889,7 +956,7 @@ int r600_pcie_gart_init(struct radeon_device *rdev) return radeon_gart_table_vram_alloc(rdev); } -int r600_pcie_gart_enable(struct radeon_device *rdev) +static int r600_pcie_gart_enable(struct radeon_device *rdev) { u32 tmp; int r, i; @@ -946,7 +1013,7 @@ int r600_pcie_gart_enable(struct radeon_device *rdev) return 0; } -void r600_pcie_gart_disable(struct radeon_device *rdev) +static void r600_pcie_gart_disable(struct radeon_device *rdev) { u32 tmp; int i; @@ -979,14 +1046,14 @@ void r600_pcie_gart_disable(struct radeon_device *rdev) radeon_gart_table_vram_unpin(rdev); } -void r600_pcie_gart_fini(struct radeon_device *rdev) +static void r600_pcie_gart_fini(struct radeon_device *rdev) { radeon_gart_fini(rdev); r600_pcie_gart_disable(rdev); radeon_gart_table_vram_free(rdev); } -void r600_agp_enable(struct radeon_device *rdev) +static void r600_agp_enable(struct radeon_device *rdev) { u32 tmp; int i; @@ -1035,6 +1102,31 @@ int r600_mc_wait_for_idle(struct radeon_device *rdev) return -1; } +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) { struct rv515_mc_save save; @@ -1134,7 +1226,7 @@ static void r600_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc } if (rdev->flags & RADEON_IS_AGP) { size_bf = mc->gtt_start; - size_af = 0xFFFFFFFF - mc->gtt_end + 1; + size_af = mc->mc_mask - mc->gtt_end; if (size_bf > size_af) { if (mc->mc_vram_size > size_bf) { dev_warn(rdev->dev, "limiting VRAM\n"); @@ -1148,7 +1240,7 @@ static void r600_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc mc->real_vram_size = size_af; mc->mc_vram_size = size_af; } - mc->vram_start = mc->gtt_end; + mc->vram_start = mc->gtt_end + 1; } mc->vram_end = mc->vram_start + mc->mc_vram_size - 1; dev_info(rdev->dev, "VRAM: %lluM 0x%08llX - 0x%08llX (%lluM used)\n", @@ -1166,10 +1258,12 @@ static void r600_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc } } -int r600_mc_init(struct radeon_device *rdev) +static int r600_mc_init(struct radeon_device *rdev) { u32 tmp; int chansize, numchan; + uint32_t h_addr, l_addr; + unsigned long long k8_addr; /* Get VRAM informations */ rdev->mc.vram_is_ddr = true; @@ -1210,7 +1304,30 @@ int r600_mc_init(struct radeon_device *rdev) if (rdev->flags & RADEON_IS_IGP) { rs690_pm_info(rdev); rdev->mc.igp_sideport_enabled = radeon_atombios_sideport_present(rdev); + + if (rdev->family == CHIP_RS780 || rdev->family == CHIP_RS880) { + /* Use K8 direct mapping for fast fb access. */ + rdev->fastfb_working = false; + h_addr = G_000012_K8_ADDR_EXT(RREG32_MC(R_000012_MC_MISC_UMA_CNTL)); + l_addr = RREG32_MC(R_000011_K8_FB_LOCATION); + k8_addr = ((unsigned long long)h_addr) << 32 | l_addr; +#if defined(CONFIG_X86_32) && !defined(CONFIG_X86_PAE) + if (k8_addr + rdev->mc.visible_vram_size < 0x100000000ULL) +#endif + { + /* FastFB shall be used with UMA memory. Here it is simply disabled when sideport + * memory is present. + */ + if (rdev->mc.igp_sideport_enabled == false && radeon_fastfb == 1) { + DRM_INFO("Direct mapping: aper base at 0x%llx, replaced by direct mapping base 0x%llx.\n", + (unsigned long long)rdev->mc.aper_base, k8_addr); + rdev->mc.aper_base = (resource_size_t)k8_addr; + rdev->fastfb_working = true; + } + } + } } + radeon_update_bandwidth_info(rdev); return 0; } @@ -1222,7 +1339,7 @@ int r600_vram_scratch_init(struct radeon_device *rdev) if (rdev->vram_scratch.robj == NULL) { r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM, - &rdev->vram_scratch.robj); + NULL, &rdev->vram_scratch.robj); if (r) { return r; } @@ -1262,250 +1379,440 @@ void r600_vram_scratch_fini(struct radeon_device *rdev) radeon_bo_unref(&rdev->vram_scratch.robj); } -/* We doesn't check that the GPU really needs a reset we simply do the - * reset, it's up to the caller to determine if the GPU needs one. We - * might add an helper function to check that. - */ -int r600_gpu_soft_reset(struct radeon_device *rdev) +void r600_set_bios_scratch_engine_hung(struct radeon_device *rdev, bool hung) +{ + u32 tmp = RREG32(R600_BIOS_3_SCRATCH); + + if (hung) + tmp |= ATOM_S3_ASIC_GUI_ENGINE_HUNG; + else + tmp &= ~ATOM_S3_ASIC_GUI_ENGINE_HUNG; + + WREG32(R600_BIOS_3_SCRATCH, tmp); +} + +static void r600_print_gpu_status_regs(struct radeon_device *rdev) +{ + dev_info(rdev->dev, " R_008010_GRBM_STATUS = 0x%08X\n", + RREG32(R_008010_GRBM_STATUS)); + dev_info(rdev->dev, " R_008014_GRBM_STATUS2 = 0x%08X\n", + RREG32(R_008014_GRBM_STATUS2)); + dev_info(rdev->dev, " R_000E50_SRBM_STATUS = 0x%08X\n", + RREG32(R_000E50_SRBM_STATUS)); + dev_info(rdev->dev, " R_008674_CP_STALLED_STAT1 = 0x%08X\n", + RREG32(CP_STALLED_STAT1)); + dev_info(rdev->dev, " R_008678_CP_STALLED_STAT2 = 0x%08X\n", + RREG32(CP_STALLED_STAT2)); + dev_info(rdev->dev, " R_00867C_CP_BUSY_STAT = 0x%08X\n", + RREG32(CP_BUSY_STAT)); + dev_info(rdev->dev, " R_008680_CP_STAT = 0x%08X\n", + RREG32(CP_STAT)); + dev_info(rdev->dev, " R_00D034_DMA_STATUS_REG = 0x%08X\n", + RREG32(DMA_STATUS_REG)); +} + +static bool r600_is_display_hung(struct radeon_device *rdev) +{ + u32 crtc_hung = 0; + u32 crtc_status[2]; + u32 i, j, tmp; + + for (i = 0; i < rdev->num_crtc; i++) { + if (RREG32(AVIVO_D1CRTC_CONTROL + crtc_offsets[i]) & AVIVO_CRTC_EN) { + crtc_status[i] = RREG32(AVIVO_D1CRTC_STATUS_HV_COUNT + crtc_offsets[i]); + crtc_hung |= (1 << i); + } + } + + for (j = 0; j < 10; j++) { + for (i = 0; i < rdev->num_crtc; i++) { + if (crtc_hung & (1 << i)) { + tmp = RREG32(AVIVO_D1CRTC_STATUS_HV_COUNT + crtc_offsets[i]); + if (tmp != crtc_status[i]) + crtc_hung &= ~(1 << i); + } + } + if (crtc_hung == 0) + return false; + udelay(100); + } + + return true; +} + +u32 r600_gpu_check_soft_reset(struct radeon_device *rdev) +{ + u32 reset_mask = 0; + u32 tmp; + + /* GRBM_STATUS */ + tmp = RREG32(R_008010_GRBM_STATUS); + if (rdev->family >= CHIP_RV770) { + if (G_008010_PA_BUSY(tmp) | G_008010_SC_BUSY(tmp) | + G_008010_SH_BUSY(tmp) | G_008010_SX_BUSY(tmp) | + G_008010_TA_BUSY(tmp) | G_008010_VGT_BUSY(tmp) | + G_008010_DB03_BUSY(tmp) | G_008010_CB03_BUSY(tmp) | + G_008010_SPI03_BUSY(tmp) | G_008010_VGT_BUSY_NO_DMA(tmp)) + reset_mask |= RADEON_RESET_GFX; + } else { + if (G_008010_PA_BUSY(tmp) | G_008010_SC_BUSY(tmp) | + G_008010_SH_BUSY(tmp) | G_008010_SX_BUSY(tmp) | + G_008010_TA03_BUSY(tmp) | G_008010_VGT_BUSY(tmp) | + G_008010_DB03_BUSY(tmp) | G_008010_CB03_BUSY(tmp) | + G_008010_SPI03_BUSY(tmp) | G_008010_VGT_BUSY_NO_DMA(tmp)) + reset_mask |= RADEON_RESET_GFX; + } + + if (G_008010_CF_RQ_PENDING(tmp) | G_008010_PF_RQ_PENDING(tmp) | + G_008010_CP_BUSY(tmp) | G_008010_CP_COHERENCY_BUSY(tmp)) + reset_mask |= RADEON_RESET_CP; + + if (G_008010_GRBM_EE_BUSY(tmp)) + reset_mask |= RADEON_RESET_GRBM | RADEON_RESET_GFX | RADEON_RESET_CP; + + /* DMA_STATUS_REG */ + tmp = RREG32(DMA_STATUS_REG); + if (!(tmp & DMA_IDLE)) + reset_mask |= RADEON_RESET_DMA; + + /* SRBM_STATUS */ + tmp = RREG32(R_000E50_SRBM_STATUS); + if (G_000E50_RLC_RQ_PENDING(tmp) | G_000E50_RLC_BUSY(tmp)) + reset_mask |= RADEON_RESET_RLC; + + if (G_000E50_IH_BUSY(tmp)) + reset_mask |= RADEON_RESET_IH; + + if (G_000E50_SEM_BUSY(tmp)) + reset_mask |= RADEON_RESET_SEM; + + if (G_000E50_GRBM_RQ_PENDING(tmp)) + reset_mask |= RADEON_RESET_GRBM; + + if (G_000E50_VMC_BUSY(tmp)) + reset_mask |= RADEON_RESET_VMC; + + if (G_000E50_MCB_BUSY(tmp) | G_000E50_MCDZ_BUSY(tmp) | + G_000E50_MCDY_BUSY(tmp) | G_000E50_MCDX_BUSY(tmp) | + G_000E50_MCDW_BUSY(tmp)) + reset_mask |= RADEON_RESET_MC; + + if (r600_is_display_hung(rdev)) + reset_mask |= RADEON_RESET_DISPLAY; + + /* Skip MC reset as it's mostly likely not hung, just busy */ + if (reset_mask & RADEON_RESET_MC) { + DRM_DEBUG("MC busy: 0x%08X, clearing.\n", reset_mask); + reset_mask &= ~RADEON_RESET_MC; + } + + return reset_mask; +} + +static void r600_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask) { struct rv515_mc_save save; - u32 grbm_busy_mask = S_008010_VC_BUSY(1) | S_008010_VGT_BUSY_NO_DMA(1) | - S_008010_VGT_BUSY(1) | S_008010_TA03_BUSY(1) | - S_008010_TC_BUSY(1) | S_008010_SX_BUSY(1) | - S_008010_SH_BUSY(1) | S_008010_SPI03_BUSY(1) | - S_008010_SMX_BUSY(1) | S_008010_SC_BUSY(1) | - S_008010_PA_BUSY(1) | S_008010_DB03_BUSY(1) | - S_008010_CR_BUSY(1) | S_008010_CB03_BUSY(1) | - S_008010_GUI_ACTIVE(1); - u32 grbm2_busy_mask = S_008014_SPI0_BUSY(1) | S_008014_SPI1_BUSY(1) | - S_008014_SPI2_BUSY(1) | S_008014_SPI3_BUSY(1) | - S_008014_TA0_BUSY(1) | S_008014_TA1_BUSY(1) | - S_008014_TA2_BUSY(1) | S_008014_TA3_BUSY(1) | - S_008014_DB0_BUSY(1) | S_008014_DB1_BUSY(1) | - S_008014_DB2_BUSY(1) | S_008014_DB3_BUSY(1) | - S_008014_CB0_BUSY(1) | S_008014_CB1_BUSY(1) | - S_008014_CB2_BUSY(1) | S_008014_CB3_BUSY(1); + u32 grbm_soft_reset = 0, srbm_soft_reset = 0; u32 tmp; - if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE)) - return 0; + if (reset_mask == 0) + return; + + dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask); + + r600_print_gpu_status_regs(rdev); + + /* 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); + + if (reset_mask & RADEON_RESET_DMA) { + /* Disable DMA */ + tmp = RREG32(DMA_RB_CNTL); + tmp &= ~DMA_RB_ENABLE; + WREG32(DMA_RB_CNTL, tmp); + } + + mdelay(50); - dev_info(rdev->dev, "GPU softreset \n"); - dev_info(rdev->dev, " R_008010_GRBM_STATUS=0x%08X\n", - RREG32(R_008010_GRBM_STATUS)); - dev_info(rdev->dev, " R_008014_GRBM_STATUS2=0x%08X\n", - RREG32(R_008014_GRBM_STATUS2)); - dev_info(rdev->dev, " R_000E50_SRBM_STATUS=0x%08X\n", - RREG32(R_000E50_SRBM_STATUS)); rv515_mc_stop(rdev, &save); if (r600_mc_wait_for_idle(rdev)) { dev_warn(rdev->dev, "Wait for MC idle timedout !\n"); } - /* Disable CP parsing/prefetching */ - WREG32(R_0086D8_CP_ME_CNTL, S_0086D8_CP_ME_HALT(1)); - /* Check if any of the rendering block is busy and reset it */ - if ((RREG32(R_008010_GRBM_STATUS) & grbm_busy_mask) || - (RREG32(R_008014_GRBM_STATUS2) & grbm2_busy_mask)) { - tmp = S_008020_SOFT_RESET_CR(1) | - S_008020_SOFT_RESET_DB(1) | - S_008020_SOFT_RESET_CB(1) | - S_008020_SOFT_RESET_PA(1) | - S_008020_SOFT_RESET_SC(1) | - S_008020_SOFT_RESET_SMX(1) | - S_008020_SOFT_RESET_SPI(1) | - S_008020_SOFT_RESET_SX(1) | - S_008020_SOFT_RESET_SH(1) | - S_008020_SOFT_RESET_TC(1) | - S_008020_SOFT_RESET_TA(1) | - S_008020_SOFT_RESET_VC(1) | + + if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE)) { + if (rdev->family >= CHIP_RV770) + grbm_soft_reset |= S_008020_SOFT_RESET_DB(1) | + S_008020_SOFT_RESET_CB(1) | + S_008020_SOFT_RESET_PA(1) | + S_008020_SOFT_RESET_SC(1) | + S_008020_SOFT_RESET_SPI(1) | + S_008020_SOFT_RESET_SX(1) | + S_008020_SOFT_RESET_SH(1) | + S_008020_SOFT_RESET_TC(1) | + S_008020_SOFT_RESET_TA(1) | + S_008020_SOFT_RESET_VC(1) | + S_008020_SOFT_RESET_VGT(1); + else + grbm_soft_reset |= S_008020_SOFT_RESET_CR(1) | + S_008020_SOFT_RESET_DB(1) | + S_008020_SOFT_RESET_CB(1) | + S_008020_SOFT_RESET_PA(1) | + S_008020_SOFT_RESET_SC(1) | + S_008020_SOFT_RESET_SMX(1) | + S_008020_SOFT_RESET_SPI(1) | + S_008020_SOFT_RESET_SX(1) | + S_008020_SOFT_RESET_SH(1) | + S_008020_SOFT_RESET_TC(1) | + S_008020_SOFT_RESET_TA(1) | + S_008020_SOFT_RESET_VC(1) | + S_008020_SOFT_RESET_VGT(1); + } + + if (reset_mask & RADEON_RESET_CP) { + grbm_soft_reset |= S_008020_SOFT_RESET_CP(1) | S_008020_SOFT_RESET_VGT(1); - dev_info(rdev->dev, " R_008020_GRBM_SOFT_RESET=0x%08X\n", tmp); + + srbm_soft_reset |= S_000E60_SOFT_RESET_GRBM(1); + } + + if (reset_mask & RADEON_RESET_DMA) { + if (rdev->family >= CHIP_RV770) + srbm_soft_reset |= RV770_SOFT_RESET_DMA; + else + srbm_soft_reset |= SOFT_RESET_DMA; + } + + if (reset_mask & RADEON_RESET_RLC) + srbm_soft_reset |= S_000E60_SOFT_RESET_RLC(1); + + if (reset_mask & RADEON_RESET_SEM) + srbm_soft_reset |= S_000E60_SOFT_RESET_SEM(1); + + if (reset_mask & RADEON_RESET_IH) + srbm_soft_reset |= S_000E60_SOFT_RESET_IH(1); + + if (reset_mask & RADEON_RESET_GRBM) + srbm_soft_reset |= S_000E60_SOFT_RESET_GRBM(1); + + if (!(rdev->flags & RADEON_IS_IGP)) { + if (reset_mask & RADEON_RESET_MC) + srbm_soft_reset |= S_000E60_SOFT_RESET_MC(1); + } + + if (reset_mask & RADEON_RESET_VMC) + srbm_soft_reset |= S_000E60_SOFT_RESET_VMC(1); + + if (grbm_soft_reset) { + tmp = RREG32(R_008020_GRBM_SOFT_RESET); + tmp |= grbm_soft_reset; + dev_info(rdev->dev, "R_008020_GRBM_SOFT_RESET=0x%08X\n", tmp); WREG32(R_008020_GRBM_SOFT_RESET, tmp); - RREG32(R_008020_GRBM_SOFT_RESET); - mdelay(15); - WREG32(R_008020_GRBM_SOFT_RESET, 0); + tmp = RREG32(R_008020_GRBM_SOFT_RESET); + + udelay(50); + + tmp &= ~grbm_soft_reset; + WREG32(R_008020_GRBM_SOFT_RESET, tmp); + tmp = RREG32(R_008020_GRBM_SOFT_RESET); } - /* Reset CP (we always reset CP) */ - tmp = S_008020_SOFT_RESET_CP(1); - dev_info(rdev->dev, "R_008020_GRBM_SOFT_RESET=0x%08X\n", tmp); - WREG32(R_008020_GRBM_SOFT_RESET, tmp); - RREG32(R_008020_GRBM_SOFT_RESET); - mdelay(15); - WREG32(R_008020_GRBM_SOFT_RESET, 0); + + if (srbm_soft_reset) { + tmp = RREG32(SRBM_SOFT_RESET); + tmp |= srbm_soft_reset; + dev_info(rdev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp); + WREG32(SRBM_SOFT_RESET, tmp); + tmp = RREG32(SRBM_SOFT_RESET); + + udelay(50); + + tmp &= ~srbm_soft_reset; + WREG32(SRBM_SOFT_RESET, tmp); + tmp = RREG32(SRBM_SOFT_RESET); + } + /* Wait a little for things to settle down */ mdelay(1); - dev_info(rdev->dev, " R_008010_GRBM_STATUS=0x%08X\n", - RREG32(R_008010_GRBM_STATUS)); - dev_info(rdev->dev, " R_008014_GRBM_STATUS2=0x%08X\n", - RREG32(R_008014_GRBM_STATUS2)); - dev_info(rdev->dev, " R_000E50_SRBM_STATUS=0x%08X\n", - RREG32(R_000E50_SRBM_STATUS)); + rv515_mc_resume(rdev, &save); - return 0; + udelay(50); + + r600_print_gpu_status_regs(rdev); } -bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) +static void r600_gpu_pci_config_reset(struct radeon_device *rdev) { - u32 srbm_status; - u32 grbm_status; - u32 grbm_status2; - struct r100_gpu_lockup *lockup; - int r; + 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) - lockup = &rdev->config.rv770.lockup; + WREG32(R_0086D8_CP_ME_CNTL, S_0086D8_CP_ME_HALT(1) | S_0086D8_CP_PFP_HALT(1)); else - lockup = &rdev->config.r600.lockup; + WREG32(R_0086D8_CP_ME_CNTL, S_0086D8_CP_ME_HALT(1)); - srbm_status = RREG32(R_000E50_SRBM_STATUS); - grbm_status = RREG32(R_008010_GRBM_STATUS); - grbm_status2 = RREG32(R_008014_GRBM_STATUS2); - if (!G_008010_GUI_ACTIVE(grbm_status)) { - r100_gpu_lockup_update(lockup, ring); - return false; + /* 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"); } - /* force CP activities */ - r = radeon_ring_lock(rdev, ring, 2); - if (!r) { - /* PACKET2 NOP */ - radeon_ring_write(ring, 0x80000000); - radeon_ring_write(ring, 0x80000000); - radeon_ring_unlock_commit(rdev, ring); + + /* 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); } - ring->rptr = RREG32(ring->rptr_reg); - return r100_gpu_cp_is_lockup(rdev, lockup, ring); } int r600_asic_reset(struct radeon_device *rdev) { - return r600_gpu_soft_reset(rdev); -} + u32 reset_mask; -static u32 r600_get_tile_pipe_to_backend_map(u32 num_tile_pipes, - u32 num_backends, - u32 backend_disable_mask) -{ - u32 backend_map = 0; - u32 enabled_backends_mask; - u32 enabled_backends_count; - u32 cur_pipe; - u32 swizzle_pipe[R6XX_MAX_PIPES]; - u32 cur_backend; - u32 i; + reset_mask = r600_gpu_check_soft_reset(rdev); - if (num_tile_pipes > R6XX_MAX_PIPES) - num_tile_pipes = R6XX_MAX_PIPES; - if (num_tile_pipes < 1) - num_tile_pipes = 1; - if (num_backends > R6XX_MAX_BACKENDS) - num_backends = R6XX_MAX_BACKENDS; - if (num_backends < 1) - num_backends = 1; - - enabled_backends_mask = 0; - enabled_backends_count = 0; - for (i = 0; i < R6XX_MAX_BACKENDS; ++i) { - if (((backend_disable_mask >> i) & 1) == 0) { - enabled_backends_mask |= (1 << i); - ++enabled_backends_count; - } - if (enabled_backends_count == num_backends) - break; - } + if (reset_mask) + r600_set_bios_scratch_engine_hung(rdev, true); - if (enabled_backends_count == 0) { - enabled_backends_mask = 1; - enabled_backends_count = 1; - } + /* try soft reset */ + r600_gpu_soft_reset(rdev, reset_mask); - if (enabled_backends_count != num_backends) - num_backends = enabled_backends_count; + reset_mask = r600_gpu_check_soft_reset(rdev); - memset((uint8_t *)&swizzle_pipe[0], 0, sizeof(u32) * R6XX_MAX_PIPES); - switch (num_tile_pipes) { - case 1: - swizzle_pipe[0] = 0; - break; - case 2: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 1; - break; - case 3: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 1; - swizzle_pipe[2] = 2; - break; - case 4: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 1; - swizzle_pipe[2] = 2; - swizzle_pipe[3] = 3; - break; - case 5: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 1; - swizzle_pipe[2] = 2; - swizzle_pipe[3] = 3; - swizzle_pipe[4] = 4; - break; - case 6: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 5; - swizzle_pipe[4] = 1; - swizzle_pipe[5] = 3; - break; - case 7: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 6; - swizzle_pipe[4] = 1; - swizzle_pipe[5] = 3; - swizzle_pipe[6] = 5; - break; - case 8: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 6; - swizzle_pipe[4] = 1; - swizzle_pipe[5] = 3; - swizzle_pipe[6] = 5; - swizzle_pipe[7] = 7; - break; - } + /* try pci config reset */ + if (reset_mask && radeon_hard_reset) + r600_gpu_pci_config_reset(rdev); - cur_backend = 0; - for (cur_pipe = 0; cur_pipe < num_tile_pipes; ++cur_pipe) { - while (((1 << cur_backend) & enabled_backends_mask) == 0) - cur_backend = (cur_backend + 1) % R6XX_MAX_BACKENDS; + reset_mask = r600_gpu_check_soft_reset(rdev); - backend_map |= (u32)(((cur_backend & 3) << (swizzle_pipe[cur_pipe] * 2))); + if (!reset_mask) + r600_set_bios_scratch_engine_hung(rdev, false); - cur_backend = (cur_backend + 1) % R6XX_MAX_BACKENDS; - } + return 0; +} - return backend_map; +/** + * r600_gfx_is_lockup - Check if the GFX engine is locked up + * + * @rdev: radeon_device pointer + * @ring: radeon_ring structure holding ring information + * + * Check if the GFX engine is locked up. + * Returns true if the engine appears to be locked up, false if not. + */ +bool r600_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) +{ + u32 reset_mask = r600_gpu_check_soft_reset(rdev); + + if (!(reset_mask & (RADEON_RESET_GFX | + RADEON_RESET_COMPUTE | + RADEON_RESET_CP))) { + radeon_ring_lockup_update(rdev, ring); + return false; + } + return radeon_ring_test_lockup(rdev, ring); } -int r600_count_pipe_bits(uint32_t val) +u32 r6xx_remap_render_backend(struct radeon_device *rdev, + u32 tiling_pipe_num, + u32 max_rb_num, + u32 total_max_rb_num, + u32 disabled_rb_mask) { - int i, ret = 0; + u32 rendering_pipe_num, rb_num_width, req_rb_num; + u32 pipe_rb_ratio, pipe_rb_remain, tmp; + u32 data = 0, mask = 1 << (max_rb_num - 1); + unsigned i, j; + + /* mask out the RBs that don't exist on that asic */ + tmp = disabled_rb_mask | ((0xff << max_rb_num) & 0xff); + /* make sure at least one RB is available */ + if ((tmp & 0xff) != 0xff) + disabled_rb_mask = tmp; + + rendering_pipe_num = 1 << tiling_pipe_num; + req_rb_num = total_max_rb_num - r600_count_pipe_bits(disabled_rb_mask); + BUG_ON(rendering_pipe_num < req_rb_num); + + pipe_rb_ratio = rendering_pipe_num / req_rb_num; + pipe_rb_remain = rendering_pipe_num - pipe_rb_ratio * req_rb_num; + + if (rdev->family <= CHIP_RV740) { + /* r6xx/r7xx */ + rb_num_width = 2; + } else { + /* eg+ */ + rb_num_width = 4; + } - for (i = 0; i < 32; i++) { - ret += val & 1; - val >>= 1; + for (i = 0; i < max_rb_num; i++) { + if (!(mask & disabled_rb_mask)) { + for (j = 0; j < pipe_rb_ratio; j++) { + data <<= rb_num_width; + data |= max_rb_num - i - 1; + } + if (pipe_rb_remain) { + data <<= rb_num_width; + data |= max_rb_num - i - 1; + pipe_rb_remain--; + } + } + mask >>= 1; } - return ret; + + return data; } -void r600_gpu_init(struct radeon_device *rdev) +int r600_count_pipe_bits(uint32_t val) +{ + return hweight32(val); +} + +static void r600_gpu_init(struct radeon_device *rdev) { u32 tiling_config; u32 ramcfg; - u32 backend_map; u32 cc_rb_backend_disable; u32 cc_gc_shader_pipe_config; u32 tmp; @@ -1516,8 +1823,9 @@ void r600_gpu_init(struct radeon_device *rdev) u32 sq_thread_resource_mgmt = 0; u32 sq_stack_resource_mgmt_1 = 0; u32 sq_stack_resource_mgmt_2 = 0; + u32 disabled_rb_mask; - /* FIXME: implement */ + rdev->config.r600.tiling_group_size = 256; switch (rdev->family) { case CHIP_R600: rdev->config.r600.max_pipes = 4; @@ -1621,10 +1929,7 @@ void r600_gpu_init(struct radeon_device *rdev) rdev->config.r600.tiling_nbanks = 4 << ((ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT); tiling_config |= BANK_TILING((ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT); tiling_config |= GROUP_SIZE((ramcfg & BURSTLENGTH_MASK) >> BURSTLENGTH_SHIFT); - if ((ramcfg & BURSTLENGTH_MASK) >> BURSTLENGTH_SHIFT) - rdev->config.r600.tiling_group_size = 512; - else - rdev->config.r600.tiling_group_size = 256; + tmp = (ramcfg & NOOFROWS_MASK) >> NOOFROWS_SHIFT; if (tmp > 3) { tiling_config |= ROW_TILING(3); @@ -1636,31 +1941,39 @@ void r600_gpu_init(struct radeon_device *rdev) tiling_config |= BANK_SWAPS(1); cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE) & 0x00ff0000; - cc_rb_backend_disable |= - BACKEND_DISABLE((R6XX_MAX_BACKENDS_MASK << rdev->config.r600.max_backends) & R6XX_MAX_BACKENDS_MASK); - - cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0xffffff00; - cc_gc_shader_pipe_config |= - INACTIVE_QD_PIPES((R6XX_MAX_PIPES_MASK << rdev->config.r600.max_pipes) & R6XX_MAX_PIPES_MASK); - cc_gc_shader_pipe_config |= - INACTIVE_SIMDS((R6XX_MAX_SIMDS_MASK << rdev->config.r600.max_simds) & R6XX_MAX_SIMDS_MASK); - - backend_map = r600_get_tile_pipe_to_backend_map(rdev->config.r600.max_tile_pipes, - (R6XX_MAX_BACKENDS - - r600_count_pipe_bits((cc_rb_backend_disable & - R6XX_MAX_BACKENDS_MASK) >> 16)), - (cc_rb_backend_disable >> 16)); + tmp = R6XX_MAX_BACKENDS - + r600_count_pipe_bits((cc_rb_backend_disable >> 16) & R6XX_MAX_BACKENDS_MASK); + if (tmp < rdev->config.r600.max_backends) { + rdev->config.r600.max_backends = tmp; + } + + cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0x00ffff00; + tmp = R6XX_MAX_PIPES - + r600_count_pipe_bits((cc_gc_shader_pipe_config >> 8) & R6XX_MAX_PIPES_MASK); + if (tmp < rdev->config.r600.max_pipes) { + rdev->config.r600.max_pipes = tmp; + } + tmp = R6XX_MAX_SIMDS - + r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R6XX_MAX_SIMDS_MASK); + 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; + tmp = r6xx_remap_render_backend(rdev, tmp, rdev->config.r600.max_backends, + R6XX_MAX_BACKENDS, disabled_rb_mask); + tiling_config |= tmp << 16; + rdev->config.r600.backend_map = tmp; + rdev->config.r600.tile_config = tiling_config; - rdev->config.r600.backend_map = backend_map; - tiling_config |= BACKEND_MAP(backend_map); WREG32(GB_TILING_CONFIG, tiling_config); WREG32(DCP_TILING_CONFIG, tiling_config & 0xffff); WREG32(HDP_TILING_CONFIG, tiling_config & 0xffff); - - /* Setup pipes */ - WREG32(CC_RB_BACKEND_DISABLE, cc_rb_backend_disable); - WREG32(CC_GC_SHADER_PIPE_CONFIG, cc_gc_shader_pipe_config); - WREG32(GC_USER_SHADER_PIPE_CONFIG, cc_gc_shader_pipe_config); + WREG32(DMA_TILING_CONFIG, tiling_config & 0xffff); tmp = R6XX_MAX_PIPES - r600_count_pipe_bits((cc_gc_shader_pipe_config & INACTIVE_QD_PIPES_MASK) >> 8); WREG32(VGT_OUT_DEALLOC_CNTL, (tmp * 4) & DEALLOC_DIST_MASK); @@ -1905,6 +2218,7 @@ void r600_gpu_init(struct radeon_device *rdev) WREG32(PA_CL_ENHANCE, (CLIP_VTX_REORDER_ENA | NUM_CLIP_SEQ(3))); WREG32(PA_SC_ENHANCE, FORCE_EOV_MAX_CLK_CNT(4095)); + WREG32(VC_ENHANCE, 0); } @@ -1913,20 +2227,27 @@ 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); } /* @@ -1934,29 +2255,24 @@ 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; } int r600_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; + const char *smc_chip_name = "RV770"; + size_t pfp_req_size, me_req_size, rlc_req_size, smc_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_R600: chip_name = "R600"; @@ -1990,32 +2306,51 @@ int r600_init_microcode(struct radeon_device *rdev) case CHIP_RV770: chip_name = "RV770"; rlc_chip_name = "R700"; + smc_chip_name = "RV770"; + smc_req_size = ALIGN(RV770_SMC_UCODE_SIZE, 4); break; case CHIP_RV730: - case CHIP_RV740: chip_name = "RV730"; rlc_chip_name = "R700"; + smc_chip_name = "RV730"; + smc_req_size = ALIGN(RV730_SMC_UCODE_SIZE, 4); break; case CHIP_RV710: chip_name = "RV710"; rlc_chip_name = "R700"; + smc_chip_name = "RV710"; + smc_req_size = ALIGN(RV710_SMC_UCODE_SIZE, 4); + break; + case CHIP_RV740: + chip_name = "RV730"; + rlc_chip_name = "R700"; + smc_chip_name = "RV740"; + smc_req_size = ALIGN(RV740_SMC_UCODE_SIZE, 4); break; case CHIP_CEDAR: chip_name = "CEDAR"; rlc_chip_name = "CEDAR"; + smc_chip_name = "CEDAR"; + smc_req_size = ALIGN(CEDAR_SMC_UCODE_SIZE, 4); break; case CHIP_REDWOOD: chip_name = "REDWOOD"; rlc_chip_name = "REDWOOD"; + smc_chip_name = "REDWOOD"; + smc_req_size = ALIGN(REDWOOD_SMC_UCODE_SIZE, 4); break; case CHIP_JUNIPER: chip_name = "JUNIPER"; rlc_chip_name = "JUNIPER"; + smc_chip_name = "JUNIPER"; + smc_req_size = ALIGN(JUNIPER_SMC_UCODE_SIZE, 4); break; case CHIP_CYPRESS: case CHIP_HEMLOCK: chip_name = "CYPRESS"; rlc_chip_name = "CYPRESS"; + smc_chip_name = "CYPRESS"; + smc_req_size = ALIGN(CYPRESS_SMC_UCODE_SIZE, 4); break; case CHIP_PALM: chip_name = "PALM"; @@ -2041,15 +2376,15 @@ int r600_init_microcode(struct radeon_device *rdev) me_req_size = R700_PM4_UCODE_SIZE * 4; rlc_req_size = R700_RLC_UCODE_SIZE * 4; } else { - pfp_req_size = PFP_UCODE_SIZE * 4; - me_req_size = PM4_UCODE_SIZE * 12; - rlc_req_size = RLC_UCODE_SIZE * 4; + pfp_req_size = R600_PFP_UCODE_SIZE * 4; + me_req_size = R600_PM4_UCODE_SIZE * 12; + rlc_req_size = R600_RLC_UCODE_SIZE * 4; } 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) { @@ -2061,7 +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) { @@ -2072,7 +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) { @@ -2082,9 +2417,25 @@ int r600_init_microcode(struct radeon_device *rdev) err = -EINVAL; } -out: - platform_device_unregister(pdev); + 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, 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); + err = -EINVAL; + } + } +out: if (err) { if (err != -EINVAL) printk(KERN_ERR @@ -2096,10 +2447,42 @@ out: rdev->me_fw = NULL; release_firmware(rdev->rlc_fw); rdev->rlc_fw = NULL; + release_firmware(rdev->smc_fw); + rdev->smc_fw = NULL; } 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; @@ -2126,13 +2509,13 @@ static int r600_cp_load_microcode(struct radeon_device *rdev) fw_data = (const __be32 *)rdev->me_fw->data; WREG32(CP_ME_RAM_WADDR, 0); - for (i = 0; i < PM4_UCODE_SIZE * 3; i++) + for (i = 0; i < R600_PM4_UCODE_SIZE * 3; i++) WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++)); fw_data = (const __be32 *)rdev->pfp_fw->data; WREG32(CP_PFP_UCODE_ADDR, 0); - for (i = 0; i < PFP_UCODE_SIZE; i++) + for (i = 0; i < R600_PFP_UCODE_SIZE; i++) WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++)); @@ -2186,8 +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 @@ -2222,36 +2605,48 @@ 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, ring); + r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring); if (r) { 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; } void r600_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsigned ring_size) { u32 rb_bufsz; + 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; + + if (radeon_ring_supports_scratch_reg(rdev, ring)) { + r = radeon_scratch_get(rdev, &ring->rptr_save_reg); + if (r) { + DRM_ERROR("failed to get scratch reg for rptr save (%d).\n", r); + ring->rptr_save_reg = 0; + } + } } void r600_cp_fini(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; r600_cp_stop(rdev); - radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); + radeon_ring_fini(rdev, ring); + radeon_scratch_free(rdev, ring->rptr_save_reg); } - /* * GPU scratch registers helpers function. */ @@ -2271,7 +2666,7 @@ int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) { uint32_t scratch; uint32_t tmp = 0; - unsigned i, ridx = radeon_ring_index(rdev, ring); + unsigned i; int r; r = radeon_scratch_get(rdev, &scratch); @@ -2282,7 +2677,7 @@ int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) WREG32(scratch, 0xCAFEDEAD); r = radeon_ring_lock(rdev, ring, 3); if (r) { - DRM_ERROR("radeon: cp failed to lock ring %d (%d).\n", ridx, r); + DRM_ERROR("radeon: cp failed to lock ring %d (%d).\n", ring->idx, r); radeon_scratch_free(rdev, scratch); return r; } @@ -2297,44 +2692,49 @@ int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) DRM_UDELAY(1); } if (i < rdev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", ridx, i); + DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("radeon: ring %d test failed (scratch(0x%04X)=0x%08X)\n", - ridx, scratch, tmp); + ring->idx, scratch, tmp); r = -EINVAL; } radeon_scratch_free(rdev, scratch); return r; } +/* + * CP fences/semaphores + */ + 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 */ @@ -2354,7 +2754,7 @@ void r600_fence_ring_emit(struct radeon_device *rdev, } } -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) @@ -2362,46 +2762,94 @@ void r600_semaphore_ring_emit(struct radeon_device *rdev, uint64_t addr = semaphore->gpu_addr; unsigned sel = emit_wait ? PACKET3_SEM_SEL_WAIT : PACKET3_SEM_SEL_SIGNAL; + if (rdev->family < CHIP_CAYMAN) + 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); + + return true; } -int r600_copy_blit(struct radeon_device *rdev, - uint64_t src_offset, - uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence *fence) +/** + * r600_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 (r6xx+). + * Used by the radeon ttm implementation to move pages if + * registered as the asic copy callback. + */ +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 r; + 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, tmp; + 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; + } - mutex_lock(&rdev->r600_blit.mutex); - rdev->r600_blit.vb_ib = NULL; - r = r600_blit_prepare_copy(rdev, num_gpu_pages); + 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) { - if (rdev->r600_blit.vb_ib) - radeon_ib_free(rdev, &rdev->r600_blit.vb_ib); - mutex_unlock(&rdev->r600_blit.mutex); + DRM_ERROR("radeon: moving bo (%d).\n", r); + radeon_semaphore_free(rdev, &sem, NULL); return r; } - r600_kms_blit_copy(rdev, src_offset, dst_offset, num_gpu_pages); - r600_blit_done_copy(rdev, fence); - mutex_unlock(&rdev->r600_blit.mutex); - return 0; -} -void r600_blit_suspend(struct radeon_device *rdev) -{ - int r; + radeon_semaphore_sync_to(sem, *fence); + radeon_semaphore_sync_rings(rdev, sem, ring->idx); - /* unpin shaders bo */ - if (rdev->r600_blit.shader_obj) { - 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_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_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; + } + + radeon_ring_unlock_commit(rdev, ring); + radeon_semaphore_free(rdev, &sem, *fence); + + return r; } int r600_set_surface_reg(struct radeon_device *rdev, int reg, @@ -2417,27 +2865,21 @@ void r600_clear_surface_reg(struct radeon_device *rdev, int reg) /* FIXME: implement */ } -int r600_startup(struct radeon_device *rdev) +static int r600_startup(struct radeon_device *rdev) { - struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + struct radeon_ring *ring; int r; /* 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 { @@ -2446,12 +2888,6 @@ 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 = NULL; - dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); - } /* allocate wb buffer */ r = radeon_wb_init(rdev); @@ -2465,6 +2901,12 @@ int r600_startup(struct radeon_device *rdev) } /* Enable IRQ */ + if (!rdev->irq.installed) { + r = radeon_irq_kms_init(rdev); + if (r) + return r; + } + r = r600_irq_init(rdev); if (r) { DRM_ERROR("radeon: IH init failed (%d).\n", r); @@ -2473,12 +2915,12 @@ int r600_startup(struct radeon_device *rdev) } r600_irq_set(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; + r = r600_cp_load_microcode(rdev); if (r) return r; @@ -2486,14 +2928,15 @@ int r600_startup(struct radeon_device *rdev) if (r) return r; - r = radeon_ib_pool_start(rdev); - if (r) + r = radeon_ib_pool_init(rdev); + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); return r; + } - r = r600_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX); + r = r600_audio_init(rdev); if (r) { - DRM_ERROR("radeon: failed testing IB (%d).\n", r); - rdev->accel_working = false; + DRM_ERROR("radeon: audio init failed\n"); return r; } @@ -2525,16 +2968,14 @@ 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) { DRM_ERROR("r600 startup failed on resume\n"); - return r; - } - - r = r600_audio_init(rdev); - if (r) { - DRM_ERROR("radeon: audio resume failed\n"); + rdev->accel_working = false; return r; } @@ -2543,12 +2984,9 @@ int r600_resume(struct radeon_device *rdev) int r600_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r600_audio_fini(rdev); - radeon_ib_pool_suspend(rdev); - r600_blit_suspend(rdev); - /* FIXME: we should wait for ring to be empty */ r600_cp_stop(rdev); - rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; r600_irq_suspend(rdev); radeon_wb_disable(rdev); r600_pcie_gart_disable(rdev); @@ -2569,10 +3007,6 @@ int r600_init(struct radeon_device *rdev) if (r600_debugfs_mc_info_init(rdev)) { DRM_ERROR("Failed to register debugfs file for mc !\n"); } - /* This don't do much */ - r = radeon_gem_init(rdev); - if (r) - return r; /* Read BIOS */ if (!radeon_get_bios(rdev)) { if (ASIC_IS_AVIVO(rdev)) @@ -2618,9 +3052,16 @@ int r600_init(struct radeon_device *rdev) if (r) return r; - r = radeon_irq_kms_init(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); @@ -2632,45 +3073,35 @@ int r600_init(struct radeon_device *rdev) if (r) return r; - r = radeon_ib_pool_init(rdev); rdev->accel_working = true; - if (r) { - dev_err(rdev->dev, "IB initialization failed (%d).\n", r); - rdev->accel_working = false; - } - r = r600_startup(rdev); if (r) { dev_err(rdev->dev, "disabling GPU acceleration\n"); r600_cp_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); - r100_ib_fini(rdev); + radeon_ib_pool_fini(rdev); radeon_irq_kms_fini(rdev); r600_pcie_gart_fini(rdev); rdev->accel_working = false; } - r = r600_audio_init(rdev); - if (r) - return r; /* TODO error handling */ return 0; } void r600_fini(struct radeon_device *rdev) { + radeon_pm_fini(rdev); r600_audio_fini(rdev); - r600_blit_fini(rdev); r600_cp_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); - r100_ib_fini(rdev); + radeon_ib_pool_fini(rdev); radeon_irq_kms_fini(rdev); r600_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_agp_fini(rdev); radeon_gem_fini(rdev); - radeon_semaphore_driver_fini(rdev); radeon_fence_driver_fini(rdev); radeon_bo_fini(rdev); radeon_atombios_fini(rdev); @@ -2684,9 +3115,24 @@ void r600_fini(struct radeon_device *rdev) */ void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) { - struct radeon_ring *ring = &rdev->ring[ib->fence->ring]; + struct radeon_ring *ring = &rdev->ring[ib->ring]; + u32 next_rptr; + + if (ring->rptr_save_reg) { + next_rptr = ring->wptr + 3 + 4; + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, ((ring->rptr_save_reg - + PACKET3_SET_CONFIG_REG_OFFSET) >> 2)); + radeon_ring_write(ring, next_rptr); + } else if (rdev->wb.enabled) { + next_rptr = ring->wptr + 5 + 4; + radeon_ring_write(ring, PACKET3(PACKET3_MEM_WRITE, 3)); + radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); + radeon_ring_write(ring, (upper_32_bits(ring->next_rptr_gpu_addr) & 0xff) | (1 << 18)); + radeon_ring_write(ring, next_rptr); + radeon_ring_write(ring, 0); + } - /* FIXME: implement */ radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); radeon_ring_write(ring, #ifdef __BIG_ENDIAN @@ -2697,9 +3143,9 @@ void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) radeon_ring_write(ring, ib->length_dw); } -int r600_ib_test(struct radeon_device *rdev, int ring) +int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) { - struct radeon_ib *ib; + struct radeon_ib ib; uint32_t scratch; uint32_t tmp = 0; unsigned i; @@ -2711,39 +3157,24 @@ int r600_ib_test(struct radeon_device *rdev, int ring) return r; } WREG32(scratch, 0xCAFEDEAD); - r = radeon_ib_get(rdev, ring, &ib, 256); + r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256); if (r) { DRM_ERROR("radeon: failed to get ib (%d).\n", r); - return r; + goto free_scratch; } - ib->ptr[0] = PACKET3(PACKET3_SET_CONFIG_REG, 1); - ib->ptr[1] = ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); - ib->ptr[2] = 0xDEADBEEF; - ib->ptr[3] = PACKET2(0); - ib->ptr[4] = PACKET2(0); - ib->ptr[5] = PACKET2(0); - ib->ptr[6] = PACKET2(0); - ib->ptr[7] = PACKET2(0); - ib->ptr[8] = PACKET2(0); - ib->ptr[9] = PACKET2(0); - ib->ptr[10] = PACKET2(0); - ib->ptr[11] = PACKET2(0); - ib->ptr[12] = PACKET2(0); - ib->ptr[13] = PACKET2(0); - ib->ptr[14] = PACKET2(0); - ib->ptr[15] = PACKET2(0); - ib->length_dw = 16; - r = radeon_ib_schedule(rdev, ib); + ib.ptr[0] = PACKET3(PACKET3_SET_CONFIG_REG, 1); + ib.ptr[1] = ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); + ib.ptr[2] = 0xDEADBEEF; + ib.length_dw = 3; + r = radeon_ib_schedule(rdev, &ib, NULL); if (r) { - radeon_scratch_free(rdev, scratch); - radeon_ib_free(rdev, &ib); DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); - return r; + goto free_ib; } - r = radeon_fence_wait(ib->fence, false); + r = radeon_fence_wait(ib.fence, false); if (r) { DRM_ERROR("radeon: fence wait failed (%d).\n", r); - return r; + goto free_ib; } for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); @@ -2752,14 +3183,16 @@ int r600_ib_test(struct radeon_device *rdev, int ring) DRM_UDELAY(1); } if (i < rdev->usec_timeout) { - DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib->fence->ring, i); + DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i); } else { DRM_ERROR("radeon: ib test failed (scratch(0x%04X)=0x%08X)\n", scratch, tmp); r = -EINVAL; } - radeon_scratch_free(rdev, scratch); +free_ib: radeon_ib_free(rdev, &ib); +free_scratch: + radeon_scratch_free(rdev, scratch); return r; } @@ -2779,14 +3212,14 @@ 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; rdev->ih.rptr = 0; } -static int r600_ih_ring_alloc(struct radeon_device *rdev) +int r600_ih_ring_alloc(struct radeon_device *rdev) { int r; @@ -2795,7 +3228,7 @@ static int r600_ih_ring_alloc(struct radeon_device *rdev) r = radeon_bo_create(rdev, rdev->ih.ring_size, PAGE_SIZE, true, RADEON_GEM_DOMAIN_GTT, - &rdev->ih.ring_obj); + NULL, &rdev->ih.ring_obj); if (r) { DRM_ERROR("radeon: failed to create ih ring buffer (%d).\n", r); return r; @@ -2822,7 +3255,7 @@ static int r600_ih_ring_alloc(struct radeon_device *rdev) return 0; } -static void r600_ih_ring_fini(struct radeon_device *rdev) +void r600_ih_ring_fini(struct radeon_device *rdev) { int r; if (rdev->ih.ring_obj) { @@ -2846,7 +3279,7 @@ void r600_rlc_stop(struct radeon_device *rdev) /* r7xx asics need to soft reset RLC before halting */ WREG32(SRBM_SOFT_RESET, SOFT_RESET_RLC); RREG32(SRBM_SOFT_RESET); - udelay(15000); + mdelay(15); WREG32(SRBM_SOFT_RESET, 0); RREG32(SRBM_SOFT_RESET); } @@ -2859,7 +3292,7 @@ static void r600_rlc_start(struct radeon_device *rdev) WREG32(RLC_CNTL, RLC_ENABLE); } -static int r600_rlc_init(struct radeon_device *rdev) +static int r600_rlc_resume(struct radeon_device *rdev) { u32 i; const __be32 *fw_data; @@ -2869,35 +3302,24 @@ static int r600_rlc_init(struct radeon_device *rdev) r600_rlc_stop(rdev); - WREG32(RLC_HB_BASE, 0); WREG32(RLC_HB_CNTL, 0); + + WREG32(RLC_HB_BASE, 0); WREG32(RLC_HB_RPTR, 0); WREG32(RLC_HB_WPTR, 0); - if (rdev->family <= CHIP_CAICOS) { - WREG32(RLC_HB_WPTR_LSB_ADDR, 0); - WREG32(RLC_HB_WPTR_MSB_ADDR, 0); - } + WREG32(RLC_HB_WPTR_LSB_ADDR, 0); + WREG32(RLC_HB_WPTR_MSB_ADDR, 0); WREG32(RLC_MC_CNTL, 0); WREG32(RLC_UCODE_CNTL, 0); fw_data = (const __be32 *)rdev->rlc_fw->data; - if (rdev->family >= CHIP_CAYMAN) { - for (i = 0; i < CAYMAN_RLC_UCODE_SIZE; i++) { - WREG32(RLC_UCODE_ADDR, i); - WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++)); - } - } else if (rdev->family >= CHIP_CEDAR) { - for (i = 0; i < EVERGREEN_RLC_UCODE_SIZE; i++) { - WREG32(RLC_UCODE_ADDR, i); - WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++)); - } - } else if (rdev->family >= CHIP_RV770) { + if (rdev->family >= CHIP_RV770) { for (i = 0; i < R700_RLC_UCODE_SIZE; i++) { WREG32(RLC_UCODE_ADDR, i); WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++)); } } else { - for (i = 0; i < RLC_UCODE_SIZE; i++) { + for (i = 0; i < R600_RLC_UCODE_SIZE; i++) { WREG32(RLC_UCODE_ADDR, i); WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++)); } @@ -2934,7 +3356,6 @@ void r600_disable_interrupts(struct radeon_device *rdev) WREG32(IH_RB_RPTR, 0); WREG32(IH_RB_WPTR, 0); rdev->ih.enabled = false; - rdev->ih.wptr = 0; rdev->ih.rptr = 0; } @@ -2943,6 +3364,8 @@ static void r600_disable_interrupt_state(struct radeon_device *rdev) u32 tmp; WREG32(CP_INT_CNTL, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + tmp = RREG32(DMA_CNTL) & ~TRAP_ENABLE; + WREG32(DMA_CNTL, tmp); WREG32(GRBM_INT_CNTL, 0); WREG32(DxMODE_INT_MASK, 0); WREG32(D1GRPH_INTERRUPT_CONTROL, 0); @@ -2963,6 +3386,15 @@ static void r600_disable_interrupt_state(struct radeon_device *rdev) WREG32(DC_HPD5_INT_CONTROL, tmp); tmp = RREG32(DC_HPD6_INT_CONTROL) & DC_HPDx_INT_POLARITY; WREG32(DC_HPD6_INT_CONTROL, tmp); + tmp = RREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET0) & ~HDMI0_AZ_FORMAT_WTRIG_MASK; + WREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET0, tmp); + tmp = RREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET1) & ~HDMI0_AZ_FORMAT_WTRIG_MASK; + WREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET1, tmp); + } else { + tmp = RREG32(HDMI0_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK; + WREG32(HDMI0_AUDIO_PACKET_CONTROL, tmp); + tmp = RREG32(DCE3_HDMI1_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK; + WREG32(DCE3_HDMI1_AUDIO_PACKET_CONTROL, tmp); } } else { WREG32(DACA_AUTODETECT_INT_CONTROL, 0); @@ -2973,6 +3405,10 @@ static void r600_disable_interrupt_state(struct radeon_device *rdev) WREG32(DC_HOT_PLUG_DETECT2_INT_CONTROL, tmp); tmp = RREG32(DC_HOT_PLUG_DETECT3_INT_CONTROL) & DC_HOT_PLUG_DETECTx_INT_POLARITY; WREG32(DC_HOT_PLUG_DETECT3_INT_CONTROL, tmp); + tmp = RREG32(HDMI0_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK; + WREG32(HDMI0_AUDIO_PACKET_CONTROL, tmp); + tmp = RREG32(HDMI1_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK; + WREG32(HDMI1_AUDIO_PACKET_CONTROL, tmp); } } @@ -2991,7 +3427,10 @@ int r600_irq_init(struct radeon_device *rdev) r600_disable_interrupts(rdev); /* init rlc */ - ret = r600_rlc_init(rdev); + if (rdev->family >= CHIP_CEDAR) + ret = evergreen_rlc_resume(rdev); + else + ret = r600_rlc_resume(rdev); if (ret) { r600_ih_ring_fini(rdev); return ret; @@ -3010,7 +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 | @@ -3042,6 +3481,9 @@ int r600_irq_init(struct radeon_device *rdev) else r600_disable_interrupt_state(rdev); + /* at this point everything should be setup correctly to enable master */ + pci_set_master(rdev->pdev); + /* enable irqs */ r600_enable_interrupts(rdev); @@ -3066,8 +3508,9 @@ int r600_irq_set(struct radeon_device *rdev) u32 mode_int = 0; u32 hpd1, hpd2, hpd3, hpd4 = 0, hpd5 = 0, hpd6 = 0; u32 grbm_int_cntl = 0; - u32 hdmi1, hdmi2; - u32 d1grph = 0, d2grph = 0; + u32 hdmi0, hdmi1; + u32 dma_cntl; + u32 thermal_int = 0; if (!rdev->irq.installed) { WARN(1, "Can't enable IRQ/MSI because no handler is installed\n"); @@ -3081,9 +3524,7 @@ int r600_irq_set(struct radeon_device *rdev) return 0; } - hdmi1 = RREG32(R600_HDMI_BLOCK1 + R600_HDMI_CNTL) & ~R600_HDMI_INT_EN; if (ASIC_IS_DCE3(rdev)) { - hdmi2 = RREG32(R600_HDMI_BLOCK3 + R600_HDMI_CNTL) & ~R600_HDMI_INT_EN; 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; @@ -3091,26 +3532,52 @@ int r600_irq_set(struct radeon_device *rdev) if (ASIC_IS_DCE32(rdev)) { hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN; hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN; + hdmi0 = RREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET0) & ~AFMT_AZ_FORMAT_WTRIG_MASK; + hdmi1 = RREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET1) & ~AFMT_AZ_FORMAT_WTRIG_MASK; + } else { + hdmi0 = RREG32(HDMI0_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK; + hdmi1 = RREG32(DCE3_HDMI1_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK; } } else { - hdmi2 = RREG32(R600_HDMI_BLOCK2 + R600_HDMI_CNTL) & ~R600_HDMI_INT_EN; hpd1 = RREG32(DC_HOT_PLUG_DETECT1_INT_CONTROL) & ~DC_HPDx_INT_EN; hpd2 = RREG32(DC_HOT_PLUG_DETECT2_INT_CONTROL) & ~DC_HPDx_INT_EN; hpd3 = RREG32(DC_HOT_PLUG_DETECT3_INT_CONTROL) & ~DC_HPDx_INT_EN; + hdmi0 = RREG32(HDMI0_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK; + hdmi1 = RREG32(HDMI1_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK; + } + + dma_cntl = RREG32(DMA_CNTL) & ~TRAP_ENABLE; + + if ((rdev->family > CHIP_R600) && (rdev->family < CHIP_RV770)) { + thermal_int = RREG32(CG_THERMAL_INT) & + ~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW); + } else if (rdev->family >= CHIP_RV770) { + thermal_int = RREG32(RV770_CG_THERMAL_INT) & + ~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW); + } + if (rdev->irq.dpm_thermal) { + DRM_DEBUG("dpm thermal\n"); + thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW; } - if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) { + if (atomic_read(&rdev->irq.ring_int[RADEON_RING_TYPE_GFX_INDEX])) { DRM_DEBUG("r600_irq_set: sw int\n"); cp_int_cntl |= RB_INT_ENABLE; cp_int_cntl |= TIME_STAMP_INT_ENABLE; } + + if (atomic_read(&rdev->irq.ring_int[R600_RING_TYPE_DMA_INDEX])) { + DRM_DEBUG("r600_irq_set: sw int dma\n"); + dma_cntl |= TRAP_ENABLE; + } + if (rdev->irq.crtc_vblank_int[0] || - rdev->irq.pflip[0]) { + atomic_read(&rdev->irq.pflip[0])) { DRM_DEBUG("r600_irq_set: vblank 0\n"); mode_int |= D1MODE_VBLANK_INT_MASK; } if (rdev->irq.crtc_vblank_int[1] || - rdev->irq.pflip[1]) { + atomic_read(&rdev->irq.pflip[1])) { DRM_DEBUG("r600_irq_set: vblank 1\n"); mode_int |= D2MODE_VBLANK_INT_MASK; } @@ -3138,27 +3605,22 @@ int r600_irq_set(struct radeon_device *rdev) DRM_DEBUG("r600_irq_set: hpd 6\n"); hpd6 |= DC_HPDx_INT_EN; } - if (rdev->irq.hdmi[0]) { - DRM_DEBUG("r600_irq_set: hdmi 1\n"); - hdmi1 |= R600_HDMI_INT_EN; - } - if (rdev->irq.hdmi[1]) { - DRM_DEBUG("r600_irq_set: hdmi 2\n"); - hdmi2 |= R600_HDMI_INT_EN; + if (rdev->irq.afmt[0]) { + DRM_DEBUG("r600_irq_set: hdmi 0\n"); + hdmi0 |= HDMI0_AZ_FORMAT_WTRIG_MASK; } - if (rdev->irq.gui_idle) { - DRM_DEBUG("gui idle\n"); - grbm_int_cntl |= GUI_IDLE_INT_ENABLE; + if (rdev->irq.afmt[1]) { + DRM_DEBUG("r600_irq_set: hdmi 0\n"); + hdmi1 |= HDMI0_AZ_FORMAT_WTRIG_MASK; } 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); - WREG32(R600_HDMI_BLOCK1 + R600_HDMI_CNTL, hdmi1); if (ASIC_IS_DCE3(rdev)) { - WREG32(R600_HDMI_BLOCK3 + R600_HDMI_CNTL, hdmi2); WREG32(DC_HPD1_INT_CONTROL, hpd1); WREG32(DC_HPD2_INT_CONTROL, hpd2); WREG32(DC_HPD3_INT_CONTROL, hpd3); @@ -3166,12 +3628,23 @@ int r600_irq_set(struct radeon_device *rdev) if (ASIC_IS_DCE32(rdev)) { WREG32(DC_HPD5_INT_CONTROL, hpd5); WREG32(DC_HPD6_INT_CONTROL, hpd6); + WREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET0, hdmi0); + WREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET1, hdmi1); + } else { + WREG32(HDMI0_AUDIO_PACKET_CONTROL, hdmi0); + WREG32(DCE3_HDMI1_AUDIO_PACKET_CONTROL, hdmi1); } } else { - WREG32(R600_HDMI_BLOCK2 + R600_HDMI_CNTL, hdmi2); WREG32(DC_HOT_PLUG_DETECT1_INT_CONTROL, hpd1); WREG32(DC_HOT_PLUG_DETECT2_INT_CONTROL, hpd2); WREG32(DC_HOT_PLUG_DETECT3_INT_CONTROL, hpd3); + WREG32(HDMI0_AUDIO_PACKET_CONTROL, hdmi0); + WREG32(HDMI1_AUDIO_PACKET_CONTROL, hdmi1); + } + if ((rdev->family > CHIP_R600) && (rdev->family < CHIP_RV770)) { + WREG32(CG_THERMAL_INT, thermal_int); + } else if (rdev->family >= CHIP_RV770) { + WREG32(RV770_CG_THERMAL_INT, thermal_int); } return 0; @@ -3185,10 +3658,19 @@ static void r600_irq_ack(struct radeon_device *rdev) rdev->irq.stat_regs.r600.disp_int = RREG32(DCE3_DISP_INTERRUPT_STATUS); rdev->irq.stat_regs.r600.disp_int_cont = RREG32(DCE3_DISP_INTERRUPT_STATUS_CONTINUE); rdev->irq.stat_regs.r600.disp_int_cont2 = RREG32(DCE3_DISP_INTERRUPT_STATUS_CONTINUE2); + if (ASIC_IS_DCE32(rdev)) { + rdev->irq.stat_regs.r600.hdmi0_status = RREG32(AFMT_STATUS + DCE3_HDMI_OFFSET0); + rdev->irq.stat_regs.r600.hdmi1_status = RREG32(AFMT_STATUS + DCE3_HDMI_OFFSET1); + } else { + rdev->irq.stat_regs.r600.hdmi0_status = RREG32(HDMI0_STATUS); + rdev->irq.stat_regs.r600.hdmi1_status = RREG32(DCE3_HDMI1_STATUS); + } } else { rdev->irq.stat_regs.r600.disp_int = RREG32(DISP_INTERRUPT_STATUS); rdev->irq.stat_regs.r600.disp_int_cont = RREG32(DISP_INTERRUPT_STATUS_CONTINUE); rdev->irq.stat_regs.r600.disp_int_cont2 = 0; + rdev->irq.stat_regs.r600.hdmi0_status = RREG32(HDMI0_STATUS); + rdev->irq.stat_regs.r600.hdmi1_status = RREG32(HDMI1_STATUS); } rdev->irq.stat_regs.r600.d1grph_int = RREG32(D1GRPH_INTERRUPT_STATUS); rdev->irq.stat_regs.r600.d2grph_int = RREG32(D2GRPH_INTERRUPT_STATUS); @@ -3254,17 +3736,32 @@ static void r600_irq_ack(struct radeon_device *rdev) tmp |= DC_HPDx_INT_ACK; WREG32(DC_HPD6_INT_CONTROL, tmp); } - } - if (RREG32(R600_HDMI_BLOCK1 + R600_HDMI_STATUS) & R600_HDMI_INT_PENDING) { - WREG32_P(R600_HDMI_BLOCK1 + R600_HDMI_CNTL, R600_HDMI_INT_ACK, ~R600_HDMI_INT_ACK); - } - if (ASIC_IS_DCE3(rdev)) { - if (RREG32(R600_HDMI_BLOCK3 + R600_HDMI_STATUS) & R600_HDMI_INT_PENDING) { - WREG32_P(R600_HDMI_BLOCK3 + R600_HDMI_CNTL, R600_HDMI_INT_ACK, ~R600_HDMI_INT_ACK); + if (rdev->irq.stat_regs.r600.hdmi0_status & AFMT_AZ_FORMAT_WTRIG) { + tmp = RREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET0); + tmp |= AFMT_AZ_FORMAT_WTRIG_ACK; + WREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET0, tmp); + } + if (rdev->irq.stat_regs.r600.hdmi1_status & AFMT_AZ_FORMAT_WTRIG) { + tmp = RREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET1); + tmp |= AFMT_AZ_FORMAT_WTRIG_ACK; + WREG32(AFMT_AUDIO_PACKET_CONTROL + DCE3_HDMI_OFFSET1, tmp); } } else { - if (RREG32(R600_HDMI_BLOCK2 + R600_HDMI_STATUS) & R600_HDMI_INT_PENDING) { - WREG32_P(R600_HDMI_BLOCK2 + R600_HDMI_CNTL, R600_HDMI_INT_ACK, ~R600_HDMI_INT_ACK); + if (rdev->irq.stat_regs.r600.hdmi0_status & HDMI0_AZ_FORMAT_WTRIG) { + tmp = RREG32(HDMI0_AUDIO_PACKET_CONTROL); + tmp |= HDMI0_AZ_FORMAT_WTRIG_ACK; + WREG32(HDMI0_AUDIO_PACKET_CONTROL, tmp); + } + if (rdev->irq.stat_regs.r600.hdmi1_status & HDMI0_AZ_FORMAT_WTRIG) { + if (ASIC_IS_DCE3(rdev)) { + tmp = RREG32(DCE3_HDMI1_AUDIO_PACKET_CONTROL); + tmp |= HDMI0_AZ_FORMAT_WTRIG_ACK; + WREG32(DCE3_HDMI1_AUDIO_PACKET_CONTROL, tmp); + } else { + tmp = RREG32(HDMI1_AUDIO_PACKET_CONTROL); + tmp |= HDMI0_AZ_FORMAT_WTRIG_ACK; + WREG32(HDMI1_AUDIO_PACKET_CONTROL, tmp); + } } } } @@ -3298,6 +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); } @@ -3338,8 +3836,9 @@ int r600_irq_process(struct radeon_device *rdev) u32 rptr; u32 src_id, src_data; u32 ring_index; - unsigned long flags; bool queue_hotplug = false; + bool queue_hdmi = false; + bool queue_thermal = false; if (!rdev->ih.enabled || rdev->shutdown) return IRQ_NONE; @@ -3349,24 +3848,21 @@ int r600_irq_process(struct radeon_device *rdev) RREG32(IH_RB_WPTR); wptr = r600_get_ih_wptr(rdev); - rptr = rdev->ih.rptr; - DRM_DEBUG("r600_irq_process start: rptr %d, wptr %d\n", rptr, wptr); - - spin_lock_irqsave(&rdev->ih.lock, flags); - if (rptr == wptr) { - spin_unlock_irqrestore(&rdev->ih.lock, flags); +restart_ih: + /* is somebody else already processing irqs? */ + if (atomic_xchg(&rdev->ih.lock, 1)) return IRQ_NONE; - } -restart_ih: + rptr = rdev->ih.rptr; + DRM_DEBUG("r600_irq_process start: rptr %d, wptr %d\n", rptr, wptr); + /* Order reading of wptr vs. reading of IH ring data */ rmb(); /* display interrupts */ r600_irq_ack(rdev); - rdev->ih.wptr = wptr; while (rptr != wptr) { /* wptr/rptr are in bytes! */ ring_index = rptr / 4; @@ -3383,8 +3879,8 @@ restart_ih: rdev->pm.vblank_sync = true; wake_up(&rdev->irq.vblank_queue); } - if (rdev->irq.pflip[0]) - radeon_crtc_handle_flip(rdev, 0); + if (atomic_read(&rdev->irq.pflip[0])) + radeon_crtc_handle_vblank(rdev, 0); rdev->irq.stat_regs.r600.disp_int &= ~LB_D1_VBLANK_INTERRUPT; DRM_DEBUG("IH: D1 vblank\n"); } @@ -3409,8 +3905,8 @@ restart_ih: rdev->pm.vblank_sync = true; wake_up(&rdev->irq.vblank_queue); } - if (rdev->irq.pflip[1]) - radeon_crtc_handle_flip(rdev, 1); + if (atomic_read(&rdev->irq.pflip[1])) + radeon_crtc_handle_vblank(rdev, 1); rdev->irq.stat_regs.r600.disp_int &= ~LB_D2_VBLANK_INTERRUPT; DRM_DEBUG("IH: D2 vblank\n"); } @@ -3426,6 +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: @@ -3475,9 +3979,30 @@ restart_ih: break; } break; - case 21: /* HDMI */ - DRM_DEBUG("IH: HDMI: 0x%x\n", src_data); - r600_audio_schedule_polling(rdev); + case 21: /* hdmi */ + switch (src_data) { + case 4: + if (rdev->irq.stat_regs.r600.hdmi0_status & HDMI0_AZ_FORMAT_WTRIG) { + rdev->irq.stat_regs.r600.hdmi0_status &= ~HDMI0_AZ_FORMAT_WTRIG; + queue_hdmi = true; + DRM_DEBUG("IH: HDMI0\n"); + } + break; + case 5: + if (rdev->irq.stat_regs.r600.hdmi1_status & HDMI0_AZ_FORMAT_WTRIG) { + rdev->irq.stat_regs.r600.hdmi1_status &= ~HDMI0_AZ_FORMAT_WTRIG; + queue_hdmi = true; + DRM_DEBUG("IH: HDMI1\n"); + } + break; + default: + DRM_ERROR("Unhandled interrupt: %d %d\n", src_id, src_data); + 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 */ @@ -3489,10 +4014,22 @@ restart_ih: DRM_DEBUG("IH: CP EOP\n"); radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); break; + case 224: /* DMA trap event */ + DRM_DEBUG("IH: DMA trap\n"); + radeon_fence_process(rdev, R600_RING_TYPE_DMA_INDEX); + 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"); - rdev->pm.gui_idle = true; - wake_up(&rdev->irq.idle_queue); break; default: DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); @@ -3503,15 +4040,21 @@ restart_ih: rptr += 16; rptr &= rdev->ih.ptr_mask; } - /* make sure wptr hasn't changed while processing */ - wptr = r600_get_ih_wptr(rdev); - if (wptr != rdev->ih.wptr) - goto restart_ih; if (queue_hotplug) schedule_work(&rdev->hotplug_work); + if (queue_hdmi) + schedule_work(&rdev->audio_work); + if (queue_thermal && rdev->pm.dpm_enabled) + schedule_work(&rdev->pm.dpm.thermal.work); rdev->ih.rptr = rptr; WREG32(IH_RB_RPTR, rdev->ih.rptr); - spin_unlock_irqrestore(&rdev->ih.lock, flags); + atomic_set(&rdev->ih.lock, 0); + + /* make sure wptr hasn't changed while processing */ + wptr = r600_get_ih_wptr(rdev); + if (wptr != rptr) + goto restart_ih; + return IRQ_HANDLED; } @@ -3575,7 +4118,7 @@ void r600_ioctl_wait_idle(struct radeon_device *rdev, struct radeon_bo *bo) void r600_set_pcie_lanes(struct radeon_device *rdev, int lanes) { - u32 link_width_cntl, mask, target_reg; + u32 link_width_cntl, mask; if (rdev->flags & RADEON_IS_IGP) return; @@ -3587,7 +4130,7 @@ void r600_set_pcie_lanes(struct radeon_device *rdev, int lanes) if (ASIC_IS_X2(rdev)) return; - /* FIXME wait for idle */ + radeon_gui_idle(rdev); switch (lanes) { case 0: @@ -3606,53 +4149,24 @@ void r600_set_pcie_lanes(struct radeon_device *rdev, int lanes) mask = RADEON_PCIE_LC_LINK_WIDTH_X8; break; case 12: + /* not actually supported */ mask = RADEON_PCIE_LC_LINK_WIDTH_X12; break; case 16: - default: mask = RADEON_PCIE_LC_LINK_WIDTH_X16; break; - } - - link_width_cntl = RREG32_PCIE_P(RADEON_PCIE_LC_LINK_WIDTH_CNTL); - - if ((link_width_cntl & RADEON_PCIE_LC_LINK_WIDTH_RD_MASK) == - (mask << RADEON_PCIE_LC_LINK_WIDTH_RD_SHIFT)) - return; - - if (link_width_cntl & R600_PCIE_LC_UPCONFIGURE_DIS) + default: + DRM_ERROR("invalid pcie lane request: %d\n", lanes); return; + } - link_width_cntl &= ~(RADEON_PCIE_LC_LINK_WIDTH_MASK | - RADEON_PCIE_LC_RECONFIG_NOW | - R600_PCIE_LC_RENEGOTIATE_EN | - R600_PCIE_LC_RECONFIG_ARC_MISSING_ESCAPE); - link_width_cntl |= mask; - - WREG32_PCIE_P(RADEON_PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); - - /* some northbridges can renegotiate the link rather than requiring - * a complete re-config. - * e.g., AMD 780/790 northbridges (pci ids: 0x5956, 0x5957, 0x5958, etc.) - */ - if (link_width_cntl & R600_PCIE_LC_RENEGOTIATION_SUPPORT) - link_width_cntl |= R600_PCIE_LC_RENEGOTIATE_EN | R600_PCIE_LC_UPCONFIGURE_SUPPORT; - else - link_width_cntl |= R600_PCIE_LC_RECONFIG_ARC_MISSING_ESCAPE; - - WREG32_PCIE_P(RADEON_PCIE_LC_LINK_WIDTH_CNTL, (link_width_cntl | - RADEON_PCIE_LC_RECONFIG_NOW)); - - if (rdev->family >= CHIP_RV770) - target_reg = R700_TARGET_AND_CURRENT_PROFILE_INDEX; - else - target_reg = R600_TARGET_AND_CURRENT_PROFILE_INDEX; - - /* wait for lane set to complete */ - link_width_cntl = RREG32(target_reg); - while (link_width_cntl == 0xffffffff) - link_width_cntl = RREG32(target_reg); + link_width_cntl = RREG32_PCIE_PORT(RADEON_PCIE_LC_LINK_WIDTH_CNTL); + link_width_cntl &= ~RADEON_PCIE_LC_LINK_WIDTH_MASK; + link_width_cntl |= mask << RADEON_PCIE_LC_LINK_WIDTH_SHIFT; + link_width_cntl |= (RADEON_PCIE_LC_RECONFIG_NOW | + R600_PCIE_LC_RECONFIG_ARC_MISSING_ESCAPE); + WREG32_PCIE_PORT(RADEON_PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); } int r600_get_pcie_lanes(struct radeon_device *rdev) @@ -3669,13 +4183,11 @@ int r600_get_pcie_lanes(struct radeon_device *rdev) if (ASIC_IS_X2(rdev)) return 0; - /* FIXME wait for idle */ + radeon_gui_idle(rdev); - link_width_cntl = RREG32_PCIE_P(RADEON_PCIE_LC_LINK_WIDTH_CNTL); + link_width_cntl = RREG32_PCIE_PORT(RADEON_PCIE_LC_LINK_WIDTH_CNTL); switch ((link_width_cntl & RADEON_PCIE_LC_LINK_WIDTH_RD_MASK) >> RADEON_PCIE_LC_LINK_WIDTH_RD_SHIFT) { - case RADEON_PCIE_LC_LINK_WIDTH_X0: - return 0; case RADEON_PCIE_LC_LINK_WIDTH_X1: return 1; case RADEON_PCIE_LC_LINK_WIDTH_X2: @@ -3684,6 +4196,10 @@ int r600_get_pcie_lanes(struct radeon_device *rdev) 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; @@ -3712,28 +4228,40 @@ static void r600_pcie_gen2_enable(struct radeon_device *rdev) if (rdev->family <= CHIP_R600) return; + if ((rdev->pdev->bus->max_bus_speed != PCIE_SPEED_5_0GT) && + (rdev->pdev->bus->max_bus_speed != PCIE_SPEED_8_0GT)) + return; + + speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); + if (speed_cntl & LC_CURRENT_DATA_RATE) { + 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"); + /* 55 nm r6xx asics */ if ((rdev->family == CHIP_RV670) || (rdev->family == CHIP_RV620) || (rdev->family == CHIP_RV635)) { /* advertise upconfig capability */ - link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL); + link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL); link_width_cntl &= ~LC_UPCONFIGURE_DIS; - WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); - link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL); + WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); + link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL); if (link_width_cntl & LC_RENEGOTIATION_SUPPORT) { lanes = (link_width_cntl & LC_LINK_WIDTH_RD_MASK) >> LC_LINK_WIDTH_RD_SHIFT; link_width_cntl &= ~(LC_LINK_WIDTH_MASK | LC_RECONFIG_ARC_MISSING_ESCAPE); link_width_cntl |= lanes | LC_RECONFIG_NOW | LC_RENEGOTIATE_EN; - WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); + WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); } else { link_width_cntl |= LC_UPCONFIGURE_DIS; - WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); + WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); } } - speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL); + speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); if ((speed_cntl & LC_OTHER_SIDE_EVER_SENT_GEN2) && (speed_cntl & LC_OTHER_SIDE_SUPPORTS_GEN2)) { @@ -3754,7 +4282,7 @@ static void r600_pcie_gen2_enable(struct radeon_device *rdev) speed_cntl &= ~LC_VOLTAGE_TIMER_SEL_MASK; speed_cntl &= ~LC_FORCE_DIS_HW_SPEED_CHANGE; speed_cntl |= LC_FORCE_EN_HW_SPEED_CHANGE; - WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl); + WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl); tmp = RREG32(0x541c); WREG32(0x541c, tmp | 0x8); @@ -3768,26 +4296,46 @@ static void r600_pcie_gen2_enable(struct radeon_device *rdev) if ((rdev->family == CHIP_RV670) || (rdev->family == CHIP_RV620) || (rdev->family == CHIP_RV635)) { - training_cntl = RREG32_PCIE_P(PCIE_LC_TRAINING_CNTL); + training_cntl = RREG32_PCIE_PORT(PCIE_LC_TRAINING_CNTL); training_cntl &= ~LC_POINT_7_PLUS_EN; - WREG32_PCIE_P(PCIE_LC_TRAINING_CNTL, training_cntl); + WREG32_PCIE_PORT(PCIE_LC_TRAINING_CNTL, training_cntl); } else { - speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL); + speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); speed_cntl &= ~LC_TARGET_LINK_SPEED_OVERRIDE_EN; - WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl); + WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl); } - speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL); + speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL); speed_cntl |= LC_GEN2_EN_STRAP; - WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl); + WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl); } else { - link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL); + link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL); /* XXX: only disable it if gen1 bridge vendor == 0x111d or 0x1106 */ if (1) link_width_cntl |= LC_UPCONFIGURE_DIS; else link_width_cntl &= ~LC_UPCONFIGURE_DIS; - WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); + WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); } } + +/** + * r600_get_gpu_clock_counter - return GPU clock counter snapshot + * + * @rdev: radeon_device pointer + * + * Fetches a GPU clock counter snapshot (R6xx-cayman). + * Returns the 64 bit clock counter snapshot. + */ +uint64_t r600_get_gpu_clock_counter(struct radeon_device *rdev) +{ + uint64_t clock; + + mutex_lock(&rdev->gpu_clock_mutex); + WREG32(RLC_CAPTURE_GPU_CLOCK_COUNT, 1); + clock = (uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_LSB) | + ((uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_MSB) << 32ULL); + mutex_unlock(&rdev->gpu_clock_mutex); + return clock; +} |