diff options
Diffstat (limited to 'drivers/gpu/drm/radeon/r100.c')
| -rw-r--r-- | drivers/gpu/drm/radeon/r100.c | 2865 |
1 files changed, 1748 insertions, 1117 deletions
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 71727460968..1544efcf1c3 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -26,18 +26,20 @@ * Jerome Glisse */ #include <linux/seq_file.h> -#include "drmP.h" -#include "drm.h" -#include "radeon_drm.h" +#include <linux/slab.h> +#include <drm/drmP.h> +#include <drm/radeon_drm.h> #include "radeon_reg.h" #include "radeon.h" +#include "radeon_asic.h" #include "r100d.h" #include "rs100d.h" #include "rv200d.h" #include "rv250d.h" +#include "atom.h" #include <linux/firmware.h> -#include <linux/platform_device.h> +#include <linux/module.h> #include "r100_reg_safe.h" #include "rn50_reg_safe.h" @@ -63,9 +65,462 @@ MODULE_FIRMWARE(FIRMWARE_R520); /* This files gather functions specifics to: * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280 + * and others in some cases. */ +static bool r100_is_in_vblank(struct radeon_device *rdev, int crtc) +{ + if (crtc == 0) { + if (RREG32(RADEON_CRTC_STATUS) & RADEON_CRTC_VBLANK_CUR) + return true; + else + return false; + } else { + if (RREG32(RADEON_CRTC2_STATUS) & RADEON_CRTC2_VBLANK_CUR) + return true; + else + return false; + } +} + +static bool r100_is_counter_moving(struct radeon_device *rdev, int crtc) +{ + u32 vline1, vline2; + + if (crtc == 0) { + vline1 = (RREG32(RADEON_CRTC_VLINE_CRNT_VLINE) >> 16) & RADEON_CRTC_V_TOTAL; + vline2 = (RREG32(RADEON_CRTC_VLINE_CRNT_VLINE) >> 16) & RADEON_CRTC_V_TOTAL; + } else { + vline1 = (RREG32(RADEON_CRTC2_VLINE_CRNT_VLINE) >> 16) & RADEON_CRTC_V_TOTAL; + vline2 = (RREG32(RADEON_CRTC2_VLINE_CRNT_VLINE) >> 16) & RADEON_CRTC_V_TOTAL; + } + if (vline1 != vline2) + return true; + else + return false; +} + +/** + * r100_wait_for_vblank - vblank wait asic callback. + * + * @rdev: radeon_device pointer + * @crtc: crtc to wait for vblank on + * + * Wait for vblank on the requested crtc (r1xx-r4xx). + */ +void r100_wait_for_vblank(struct radeon_device *rdev, int crtc) +{ + unsigned i = 0; + + if (crtc >= rdev->num_crtc) + return; + + if (crtc == 0) { + if (!(RREG32(RADEON_CRTC_GEN_CNTL) & RADEON_CRTC_EN)) + return; + } else { + if (!(RREG32(RADEON_CRTC2_GEN_CNTL) & RADEON_CRTC2_EN)) + return; + } + + /* depending on when we hit vblank, we may be close to active; if so, + * wait for another frame. + */ + while (r100_is_in_vblank(rdev, crtc)) { + if (i++ % 100 == 0) { + if (!r100_is_counter_moving(rdev, crtc)) + break; + } + } + + while (!r100_is_in_vblank(rdev, crtc)) { + if (i++ % 100 == 0) { + if (!r100_is_counter_moving(rdev, crtc)) + break; + } + } +} + +/** + * r100_page_flip - pageflip callback. + * + * @rdev: radeon_device pointer + * @crtc_id: crtc to cleanup pageflip on + * @crtc_base: new address of the crtc (GPU MC address) + * + * Does the actual pageflip (r1xx-r4xx). + * 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. + */ +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; + int i; + + /* Lock the graphics update lock */ + /* update the scanout addresses */ + WREG32(RADEON_CRTC_OFFSET + radeon_crtc->crtc_offset, tmp); + + /* Wait for update_pending to go high. */ + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(RADEON_CRTC_OFFSET + radeon_crtc->crtc_offset) & RADEON_CRTC_OFFSET__GUI_TRIG_OFFSET) + break; + udelay(1); + } + DRM_DEBUG("Update pending now high. Unlocking vupdate_lock.\n"); + + /* Unlock the lock, so double-buffering can take place inside vblank */ + 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); +} + +/** + * r100_pm_get_dynpm_state - look up dynpm power state callback. + * + * @rdev: radeon_device pointer + * + * Look up the optimal power state based on the + * current state of the GPU (r1xx-r5xx). + * Used for dynpm only. + */ +void r100_pm_get_dynpm_state(struct radeon_device *rdev) +{ + int i; + rdev->pm.dynpm_can_upclock = true; + rdev->pm.dynpm_can_downclock = true; + + switch (rdev->pm.dynpm_planned_action) { + case DYNPM_ACTION_MINIMUM: + rdev->pm.requested_power_state_index = 0; + rdev->pm.dynpm_can_downclock = false; + break; + case DYNPM_ACTION_DOWNCLOCK: + if (rdev->pm.current_power_state_index == 0) { + rdev->pm.requested_power_state_index = rdev->pm.current_power_state_index; + rdev->pm.dynpm_can_downclock = false; + } else { + if (rdev->pm.active_crtc_count > 1) { + for (i = 0; i < rdev->pm.num_power_states; i++) { + if (rdev->pm.power_state[i].flags & RADEON_PM_STATE_SINGLE_DISPLAY_ONLY) + continue; + else if (i >= rdev->pm.current_power_state_index) { + rdev->pm.requested_power_state_index = rdev->pm.current_power_state_index; + break; + } else { + rdev->pm.requested_power_state_index = i; + break; + } + } + } else + rdev->pm.requested_power_state_index = + rdev->pm.current_power_state_index - 1; + } + /* don't use the power state if crtcs are active and no display flag is set */ + if ((rdev->pm.active_crtc_count > 0) && + (rdev->pm.power_state[rdev->pm.requested_power_state_index].clock_info[0].flags & + RADEON_PM_MODE_NO_DISPLAY)) { + rdev->pm.requested_power_state_index++; + } + break; + case DYNPM_ACTION_UPCLOCK: + if (rdev->pm.current_power_state_index == (rdev->pm.num_power_states - 1)) { + rdev->pm.requested_power_state_index = rdev->pm.current_power_state_index; + rdev->pm.dynpm_can_upclock = false; + } else { + if (rdev->pm.active_crtc_count > 1) { + for (i = (rdev->pm.num_power_states - 1); i >= 0; i--) { + if (rdev->pm.power_state[i].flags & RADEON_PM_STATE_SINGLE_DISPLAY_ONLY) + continue; + else if (i <= rdev->pm.current_power_state_index) { + rdev->pm.requested_power_state_index = rdev->pm.current_power_state_index; + break; + } else { + rdev->pm.requested_power_state_index = i; + break; + } + } + } else + rdev->pm.requested_power_state_index = + rdev->pm.current_power_state_index + 1; + } + break; + case DYNPM_ACTION_DEFAULT: + rdev->pm.requested_power_state_index = rdev->pm.default_power_state_index; + rdev->pm.dynpm_can_upclock = false; + break; + case DYNPM_ACTION_NONE: + default: + DRM_ERROR("Requested mode for not defined action\n"); + return; + } + /* only one clock mode per power state */ + rdev->pm.requested_clock_mode_index = 0; + + DRM_DEBUG_DRIVER("Requested: e: %d m: %d p: %d\n", + rdev->pm.power_state[rdev->pm.requested_power_state_index]. + clock_info[rdev->pm.requested_clock_mode_index].sclk, + rdev->pm.power_state[rdev->pm.requested_power_state_index]. + clock_info[rdev->pm.requested_clock_mode_index].mclk, + rdev->pm.power_state[rdev->pm.requested_power_state_index]. + pcie_lanes); +} + +/** + * r100_pm_init_profile - Initialize power profiles callback. + * + * @rdev: radeon_device pointer + * + * Initialize the power states used in profile mode + * (r1xx-r3xx). + * Used for profile mode only. + */ +void r100_pm_init_profile(struct radeon_device *rdev) +{ + /* default */ + rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_off_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_on_cm_idx = 0; + /* low sh */ + rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0; + /* mid sh */ + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 0; + /* high sh */ + rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_cm_idx = 0; + /* low mh */ + rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0; + /* mid mh */ + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 0; + /* high mh */ + rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = 0; + rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index; + rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_cm_idx = 0; + rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_cm_idx = 0; +} + +/** + * r100_pm_misc - set additional pm hw parameters callback. + * + * @rdev: radeon_device pointer + * + * Set non-clock parameters associated with a power state + * (voltage, pcie lanes, etc.) (r1xx-r4xx). + */ +void r100_pm_misc(struct radeon_device *rdev) +{ + int requested_index = rdev->pm.requested_power_state_index; + struct radeon_power_state *ps = &rdev->pm.power_state[requested_index]; + struct radeon_voltage *voltage = &ps->clock_info[0].voltage; + u32 tmp, sclk_cntl, sclk_cntl2, sclk_more_cntl; + + if ((voltage->type == VOLTAGE_GPIO) && (voltage->gpio.valid)) { + if (ps->misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_SUPPORT) { + tmp = RREG32(voltage->gpio.reg); + if (voltage->active_high) + tmp |= voltage->gpio.mask; + else + tmp &= ~(voltage->gpio.mask); + WREG32(voltage->gpio.reg, tmp); + if (voltage->delay) + udelay(voltage->delay); + } else { + tmp = RREG32(voltage->gpio.reg); + if (voltage->active_high) + tmp &= ~voltage->gpio.mask; + else + tmp |= voltage->gpio.mask; + WREG32(voltage->gpio.reg, tmp); + if (voltage->delay) + udelay(voltage->delay); + } + } + + sclk_cntl = RREG32_PLL(SCLK_CNTL); + sclk_cntl2 = RREG32_PLL(SCLK_CNTL2); + sclk_cntl2 &= ~REDUCED_SPEED_SCLK_SEL(3); + sclk_more_cntl = RREG32_PLL(SCLK_MORE_CNTL); + sclk_more_cntl &= ~VOLTAGE_DELAY_SEL(3); + if (ps->misc & ATOM_PM_MISCINFO_ASIC_REDUCED_SPEED_SCLK_EN) { + sclk_more_cntl |= REDUCED_SPEED_SCLK_EN; + if (ps->misc & ATOM_PM_MISCINFO_DYN_CLK_3D_IDLE) + sclk_cntl2 |= REDUCED_SPEED_SCLK_MODE; + else + sclk_cntl2 &= ~REDUCED_SPEED_SCLK_MODE; + if (ps->misc & ATOM_PM_MISCINFO_DYNAMIC_CLOCK_DIVIDER_BY_2) + sclk_cntl2 |= REDUCED_SPEED_SCLK_SEL(0); + else if (ps->misc & ATOM_PM_MISCINFO_DYNAMIC_CLOCK_DIVIDER_BY_4) + sclk_cntl2 |= REDUCED_SPEED_SCLK_SEL(2); + } else + sclk_more_cntl &= ~REDUCED_SPEED_SCLK_EN; + + if (ps->misc & ATOM_PM_MISCINFO_ASIC_DYNAMIC_VOLTAGE_EN) { + sclk_more_cntl |= IO_CG_VOLTAGE_DROP; + if (voltage->delay) { + sclk_more_cntl |= VOLTAGE_DROP_SYNC; + switch (voltage->delay) { + case 33: + sclk_more_cntl |= VOLTAGE_DELAY_SEL(0); + break; + case 66: + sclk_more_cntl |= VOLTAGE_DELAY_SEL(1); + break; + case 99: + sclk_more_cntl |= VOLTAGE_DELAY_SEL(2); + break; + case 132: + sclk_more_cntl |= VOLTAGE_DELAY_SEL(3); + break; + } + } else + sclk_more_cntl &= ~VOLTAGE_DROP_SYNC; + } else + sclk_more_cntl &= ~IO_CG_VOLTAGE_DROP; + + if (ps->misc & ATOM_PM_MISCINFO_DYNAMIC_HDP_BLOCK_EN) + sclk_cntl &= ~FORCE_HDP; + else + sclk_cntl |= FORCE_HDP; + + WREG32_PLL(SCLK_CNTL, sclk_cntl); + WREG32_PLL(SCLK_CNTL2, sclk_cntl2); + WREG32_PLL(SCLK_MORE_CNTL, sclk_more_cntl); + + /* set pcie lanes */ + if ((rdev->flags & RADEON_IS_PCIE) && + !(rdev->flags & RADEON_IS_IGP) && + rdev->asic->pm.set_pcie_lanes && + (ps->pcie_lanes != + rdev->pm.power_state[rdev->pm.current_power_state_index].pcie_lanes)) { + radeon_set_pcie_lanes(rdev, + ps->pcie_lanes); + DRM_DEBUG_DRIVER("Setting: p: %d\n", ps->pcie_lanes); + } +} + +/** + * r100_pm_prepare - pre-power state change callback. + * + * @rdev: radeon_device pointer + * + * Prepare for a power state change (r1xx-r4xx). + */ +void r100_pm_prepare(struct radeon_device *rdev) +{ + struct drm_device *ddev = rdev->ddev; + struct drm_crtc *crtc; + struct radeon_crtc *radeon_crtc; + u32 tmp; + + /* disable any active CRTCs */ + list_for_each_entry(crtc, &ddev->mode_config.crtc_list, head) { + radeon_crtc = to_radeon_crtc(crtc); + if (radeon_crtc->enabled) { + if (radeon_crtc->crtc_id) { + tmp = RREG32(RADEON_CRTC2_GEN_CNTL); + tmp |= RADEON_CRTC2_DISP_REQ_EN_B; + WREG32(RADEON_CRTC2_GEN_CNTL, tmp); + } else { + tmp = RREG32(RADEON_CRTC_GEN_CNTL); + tmp |= RADEON_CRTC_DISP_REQ_EN_B; + WREG32(RADEON_CRTC_GEN_CNTL, tmp); + } + } + } +} + +/** + * r100_pm_finish - post-power state change callback. + * + * @rdev: radeon_device pointer + * + * Clean up after a power state change (r1xx-r4xx). + */ +void r100_pm_finish(struct radeon_device *rdev) +{ + struct drm_device *ddev = rdev->ddev; + struct drm_crtc *crtc; + struct radeon_crtc *radeon_crtc; + u32 tmp; + + /* enable any active CRTCs */ + list_for_each_entry(crtc, &ddev->mode_config.crtc_list, head) { + radeon_crtc = to_radeon_crtc(crtc); + if (radeon_crtc->enabled) { + if (radeon_crtc->crtc_id) { + tmp = RREG32(RADEON_CRTC2_GEN_CNTL); + tmp &= ~RADEON_CRTC2_DISP_REQ_EN_B; + WREG32(RADEON_CRTC2_GEN_CNTL, tmp); + } else { + tmp = RREG32(RADEON_CRTC_GEN_CNTL); + tmp &= ~RADEON_CRTC_DISP_REQ_EN_B; + WREG32(RADEON_CRTC_GEN_CNTL, tmp); + } + } + } +} + +/** + * r100_gui_idle - gui idle callback. + * + * @rdev: radeon_device pointer + * + * Check of the GUI (2D/3D engines) are idle (r1xx-r5xx). + * Returns true if idle, false if not. + */ +bool r100_gui_idle(struct radeon_device *rdev) +{ + if (RREG32(RADEON_RBBM_STATUS) & RADEON_RBBM_ACTIVE) + return false; + else + return true; +} + /* hpd for digital panel detect/disconnect */ +/** + * r100_hpd_sense - hpd sense callback. + * + * @rdev: radeon_device pointer + * @hpd: hpd (hotplug detect) pin + * + * Checks if a digital monitor is connected (r1xx-r4xx). + * Returns true if connected, false if not connected. + */ bool r100_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd) { bool connected = false; @@ -85,6 +540,14 @@ bool r100_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd) return connected; } +/** + * r100_hpd_set_polarity - hpd set polarity callback. + * + * @rdev: radeon_device pointer + * @hpd: hpd (hotplug detect) pin + * + * Set the polarity of the hpd pin (r1xx-r4xx). + */ void r100_hpd_set_polarity(struct radeon_device *rdev, enum radeon_hpd_id hpd) { @@ -113,45 +576,47 @@ void r100_hpd_set_polarity(struct radeon_device *rdev, } } +/** + * r100_hpd_init - hpd setup callback. + * + * @rdev: radeon_device pointer + * + * Setup the hpd pins used by the card (r1xx-r4xx). + * Set the polarity, and enable the hpd interrupts. + */ void r100_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); - switch (radeon_connector->hpd.hpd) { - case RADEON_HPD_1: - rdev->irq.hpd[0] = true; - break; - case RADEON_HPD_2: - rdev->irq.hpd[1] = true; - break; - default: - break; - } + enable |= 1 << radeon_connector->hpd.hpd; + radeon_hpd_set_polarity(rdev, radeon_connector->hpd.hpd); } - r100_irq_set(rdev); + radeon_irq_kms_enable_hpd(rdev, enable); } +/** + * r100_hpd_fini - hpd tear down callback. + * + * @rdev: radeon_device pointer + * + * Tear down the hpd pins used by the card (r1xx-r4xx). + * Disable the hpd interrupts. + */ void r100_hpd_fini(struct radeon_device *rdev) { struct drm_device *dev = rdev->ddev; struct drm_connector *connector; + unsigned disable = 0; list_for_each_entry(connector, &dev->mode_config.connector_list, head) { struct radeon_connector *radeon_connector = to_radeon_connector(connector); - switch (radeon_connector->hpd.hpd) { - case RADEON_HPD_1: - rdev->irq.hpd[0] = false; - break; - case RADEON_HPD_2: - rdev->irq.hpd[1] = false; - break; - default: - break; - } + disable |= 1 << radeon_connector->hpd.hpd; } + radeon_irq_kms_disable_hpd(rdev, disable); } /* @@ -169,8 +634,8 @@ int r100_pci_gart_init(struct radeon_device *rdev) { int r; - if (rdev->gart.table.ram.ptr) { - WARN(1, "R100 PCI GART already initialized.\n"); + if (rdev->gart.ptr) { + WARN(1, "R100 PCI GART already initialized\n"); return 0; } /* Initialize common gart structure */ @@ -178,36 +643,30 @@ int r100_pci_gart_init(struct radeon_device *rdev) if (r) return r; rdev->gart.table_size = rdev->gart.num_gpu_pages * 4; - rdev->asic->gart_tlb_flush = &r100_pci_gart_tlb_flush; - rdev->asic->gart_set_page = &r100_pci_gart_set_page; + rdev->asic->gart.tlb_flush = &r100_pci_gart_tlb_flush; + rdev->asic->gart.set_page = &r100_pci_gart_set_page; return radeon_gart_table_ram_alloc(rdev); } -/* required on r1xx, r2xx, r300, r(v)350, r420/r481, rs400/rs480 */ -void r100_enable_bm(struct radeon_device *rdev) -{ - uint32_t tmp; - /* Enable bus mastering */ - tmp = RREG32(RADEON_BUS_CNTL) & ~RADEON_BUS_MASTER_DIS; - WREG32(RADEON_BUS_CNTL, tmp); -} - int r100_pci_gart_enable(struct radeon_device *rdev) { uint32_t tmp; + radeon_gart_restore(rdev); /* discard memory request outside of configured range */ tmp = RREG32(RADEON_AIC_CNTL) | RADEON_DIS_OUT_OF_PCI_GART_ACCESS; WREG32(RADEON_AIC_CNTL, tmp); /* set address range for PCI address translate */ - WREG32(RADEON_AIC_LO_ADDR, rdev->mc.gtt_location); - tmp = rdev->mc.gtt_location + rdev->mc.gtt_size - 1; - WREG32(RADEON_AIC_HI_ADDR, tmp); + WREG32(RADEON_AIC_LO_ADDR, rdev->mc.gtt_start); + WREG32(RADEON_AIC_HI_ADDR, rdev->mc.gtt_end); /* set PCI GART page-table base address */ WREG32(RADEON_AIC_PT_BASE, rdev->gart.table_addr); tmp = RREG32(RADEON_AIC_CNTL) | RADEON_PCIGART_TRANSLATE_EN; WREG32(RADEON_AIC_CNTL, tmp); r100_pci_gart_tlb_flush(rdev); + DRM_INFO("PCI GART of %uM enabled (table at 0x%016llX).\n", + (unsigned)(rdev->mc.gtt_size >> 20), + (unsigned long long)rdev->gart.table_addr); rdev->gart.ready = true; return 0; } @@ -223,33 +682,38 @@ 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) { - if (i < 0 || i > rdev->gart.num_gpu_pages) { - return -EINVAL; - } - rdev->gart.table.ram.ptr[i] = cpu_to_le32(lower_32_bits(addr)); - return 0; + u32 *gtt = rdev->gart.ptr; + gtt[i] = cpu_to_le32(lower_32_bits(addr)); } void r100_pci_gart_fini(struct radeon_device *rdev) { + radeon_gart_fini(rdev); r100_pci_gart_disable(rdev); radeon_gart_table_ram_free(rdev); - radeon_gart_fini(rdev); } int r100_irq_set(struct radeon_device *rdev) { uint32_t tmp = 0; - if (rdev->irq.sw_int) { + if (!rdev->irq.installed) { + WARN(1, "Can't enable IRQ/MSI because no handler is installed\n"); + WREG32(R_000040_GEN_INT_CNTL, 0); + return -EINVAL; + } + if (atomic_read(&rdev->irq.ring_int[RADEON_RING_TYPE_GFX_INDEX])) { tmp |= RADEON_SW_INT_ENABLE; } - if (rdev->irq.crtc_vblank_int[0]) { + if (rdev->irq.crtc_vblank_int[0] || + atomic_read(&rdev->irq.pflip[0])) { tmp |= RADEON_CRTC_VBLANK_MASK; } - if (rdev->irq.crtc_vblank_int[1]) { + if (rdev->irq.crtc_vblank_int[1] || + atomic_read(&rdev->irq.pflip[1])) { tmp |= RADEON_CRTC2_VBLANK_MASK; } if (rdev->irq.hpd[0]) { @@ -273,7 +737,7 @@ void r100_irq_disable(struct radeon_device *rdev) WREG32(R_000044_GEN_INT_STATUS, tmp); } -static inline uint32_t r100_irq_ack(struct radeon_device *rdev) +static uint32_t r100_irq_ack(struct radeon_device *rdev) { uint32_t irqs = RREG32(RADEON_GEN_INT_STATUS); uint32_t irq_mask = RADEON_SW_INT_TEST | @@ -301,14 +765,26 @@ int r100_irq_process(struct radeon_device *rdev) while (status) { /* SW interrupt */ if (status & RADEON_SW_INT_TEST) { - radeon_fence_process(rdev); + radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); } /* Vertical blank interrupts */ if (status & RADEON_CRTC_VBLANK_STAT) { - drm_handle_vblank(rdev->ddev, 0); + if (rdev->irq.crtc_vblank_int[0]) { + drm_handle_vblank(rdev->ddev, 0); + rdev->pm.vblank_sync = true; + wake_up(&rdev->irq.vblank_queue); + } + if (atomic_read(&rdev->irq.pflip[0])) + radeon_crtc_handle_vblank(rdev, 0); } if (status & RADEON_CRTC2_VBLANK_STAT) { - drm_handle_vblank(rdev->ddev, 1); + if (rdev->irq.crtc_vblank_int[1]) { + drm_handle_vblank(rdev->ddev, 1); + rdev->pm.vblank_sync = true; + wake_up(&rdev->irq.vblank_queue); + } + if (atomic_read(&rdev->irq.pflip[1])) + radeon_crtc_handle_vblank(rdev, 1); } if (status & RADEON_FP_DETECT_STAT) { queue_hotplug = true; @@ -321,7 +797,7 @@ int r100_irq_process(struct radeon_device *rdev) status = r100_irq_ack(rdev); } if (queue_hotplug) - queue_work(rdev->wq, &rdev->hotplug_work); + schedule_work(&rdev->hotplug_work); if (rdev->msi_enabled) { switch (rdev->family) { case CHIP_RS400: @@ -331,9 +807,7 @@ int r100_irq_process(struct radeon_device *rdev) WREG32(RADEON_AIC_CNTL, msi_rearm | RS400_MSI_REARM); break; default: - msi_rearm = RREG32(RADEON_MSI_REARM_EN) & ~RV370_MSI_REARM_EN; - WREG32(RADEON_MSI_REARM_EN, msi_rearm); - WREG32(RADEON_MSI_REARM_EN, msi_rearm | RV370_MSI_REARM_EN); + WREG32(RADEON_MSI_REARM_EN, RV370_MSI_REARM_EN); break; } } @@ -348,90 +822,53 @@ u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc) return RREG32(RADEON_CRTC2_CRNT_FRAME); } +/* Who ever call radeon_fence_emit should call ring_lock and ask + * for enough space (today caller are ib schedule and buffer move) */ void r100_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence) { - /* Who ever call radeon_fence_emit should call ring_lock and ask - * for enough space (today caller are ib schedule and buffer move) */ + struct radeon_ring *ring = &rdev->ring[fence->ring]; + + /* We have to make sure that caches are flushed before + * CPU might read something from VRAM. */ + radeon_ring_write(ring, PACKET0(RADEON_RB3D_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, RADEON_RB3D_DC_FLUSH_ALL); + radeon_ring_write(ring, PACKET0(RADEON_RB3D_ZCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, RADEON_RB3D_ZC_FLUSH_ALL); /* Wait until IDLE & CLEAN */ - radeon_ring_write(rdev, PACKET0(0x1720, 0)); - radeon_ring_write(rdev, (1 << 16) | (1 << 17)); + radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(ring, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN); + radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0)); + radeon_ring_write(ring, rdev->config.r100.hdp_cntl | + RADEON_HDP_READ_BUFFER_INVALIDATE); + radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0)); + radeon_ring_write(ring, rdev->config.r100.hdp_cntl); /* Emit fence sequence & fire IRQ */ - radeon_ring_write(rdev, PACKET0(rdev->fence_drv.scratch_reg, 0)); - radeon_ring_write(rdev, fence->seq); - radeon_ring_write(rdev, PACKET0(RADEON_GEN_INT_STATUS, 0)); - radeon_ring_write(rdev, RADEON_SW_INT_FIRE); -} - -int r100_wb_init(struct radeon_device *rdev) -{ - int r; - - if (rdev->wb.wb_obj == NULL) { - r = radeon_bo_create(rdev, NULL, RADEON_GPU_PAGE_SIZE, true, - RADEON_GEM_DOMAIN_GTT, - &rdev->wb.wb_obj); - if (r) { - dev_err(rdev->dev, "(%d) create WB buffer failed\n", r); - return r; - } - r = radeon_bo_reserve(rdev->wb.wb_obj, false); - if (unlikely(r != 0)) - return r; - r = radeon_bo_pin(rdev->wb.wb_obj, RADEON_GEM_DOMAIN_GTT, - &rdev->wb.gpu_addr); - if (r) { - dev_err(rdev->dev, "(%d) pin WB buffer failed\n", r); - radeon_bo_unreserve(rdev->wb.wb_obj); - return r; - } - r = radeon_bo_kmap(rdev->wb.wb_obj, (void **)&rdev->wb.wb); - radeon_bo_unreserve(rdev->wb.wb_obj); - if (r) { - dev_err(rdev->dev, "(%d) map WB buffer failed\n", r); - return r; - } - } - WREG32(R_000774_SCRATCH_ADDR, rdev->wb.gpu_addr); - WREG32(R_00070C_CP_RB_RPTR_ADDR, - S_00070C_RB_RPTR_ADDR((rdev->wb.gpu_addr + 1024) >> 2)); - WREG32(R_000770_SCRATCH_UMSK, 0xff); - return 0; + radeon_ring_write(ring, PACKET0(rdev->fence_drv[fence->ring].scratch_reg, 0)); + radeon_ring_write(ring, fence->seq); + radeon_ring_write(ring, PACKET0(RADEON_GEN_INT_STATUS, 0)); + radeon_ring_write(ring, RADEON_SW_INT_FIRE); } -void r100_wb_disable(struct radeon_device *rdev) +bool r100_semaphore_ring_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait) { - WREG32(R_000770_SCRATCH_UMSK, 0); -} - -void r100_wb_fini(struct radeon_device *rdev) -{ - int r; - - r100_wb_disable(rdev); - if (rdev->wb.wb_obj) { - r = radeon_bo_reserve(rdev->wb.wb_obj, false); - if (unlikely(r != 0)) { - dev_err(rdev->dev, "(%d) can't finish WB\n", r); - return; - } - radeon_bo_kunmap(rdev->wb.wb_obj); - radeon_bo_unpin(rdev->wb.wb_obj); - radeon_bo_unreserve(rdev->wb.wb_obj); - radeon_bo_unref(&rdev->wb.wb_obj); - rdev->wb.wb = NULL; - rdev->wb.wb_obj = NULL; - } + /* Unused on older asics, since we don't have semaphores or multiple rings */ + BUG(); + return false; } int r100_copy_blit(struct radeon_device *rdev, uint64_t src_offset, uint64_t dst_offset, - unsigned num_pages, - struct radeon_fence *fence) + unsigned num_gpu_pages, + struct radeon_fence **fence) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; uint32_t cur_pages; - uint32_t stride_bytes = PAGE_SIZE; + uint32_t stride_bytes = RADEON_GPU_PAGE_SIZE; uint32_t pitch; uint32_t stride_pixels; unsigned ndw; @@ -443,26 +880,26 @@ int r100_copy_blit(struct radeon_device *rdev, /* radeon pitch is /64 */ pitch = stride_bytes / 64; stride_pixels = stride_bytes / 4; - num_loops = DIV_ROUND_UP(num_pages, 8191); + num_loops = DIV_ROUND_UP(num_gpu_pages, 8191); /* Ask for enough room for blit + flush + fence */ ndw = 64 + (10 * num_loops); - r = radeon_ring_lock(rdev, ndw); + r = radeon_ring_lock(rdev, ring, ndw); if (r) { DRM_ERROR("radeon: moving bo (%d) asking for %u dw.\n", r, ndw); return -EINVAL; } - while (num_pages > 0) { - cur_pages = num_pages; + while (num_gpu_pages > 0) { + cur_pages = num_gpu_pages; if (cur_pages > 8191) { cur_pages = 8191; } - num_pages -= cur_pages; + num_gpu_pages -= cur_pages; /* pages are in Y direction - height page width in X direction - width */ - radeon_ring_write(rdev, PACKET3(PACKET3_BITBLT_MULTI, 8)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET3(PACKET3_BITBLT_MULTI, 8)); + radeon_ring_write(ring, RADEON_GMC_SRC_PITCH_OFFSET_CNTL | RADEON_GMC_DST_PITCH_OFFSET_CNTL | RADEON_GMC_SRC_CLIPPING | @@ -474,26 +911,26 @@ int r100_copy_blit(struct radeon_device *rdev, RADEON_DP_SRC_SOURCE_MEMORY | RADEON_GMC_CLR_CMP_CNTL_DIS | RADEON_GMC_WR_MSK_DIS); - radeon_ring_write(rdev, (pitch << 22) | (src_offset >> 10)); - radeon_ring_write(rdev, (pitch << 22) | (dst_offset >> 10)); - radeon_ring_write(rdev, (0x1fff) | (0x1fff << 16)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, (0x1fff) | (0x1fff << 16)); - radeon_ring_write(rdev, num_pages); - radeon_ring_write(rdev, num_pages); - radeon_ring_write(rdev, cur_pages | (stride_pixels << 16)); - } - radeon_ring_write(rdev, PACKET0(RADEON_DSTCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, RADEON_RB2D_DC_FLUSH_ALL); - radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, (pitch << 22) | (src_offset >> 10)); + radeon_ring_write(ring, (pitch << 22) | (dst_offset >> 10)); + radeon_ring_write(ring, (0x1fff) | (0x1fff << 16)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, (0x1fff) | (0x1fff << 16)); + radeon_ring_write(ring, num_gpu_pages); + radeon_ring_write(ring, num_gpu_pages); + radeon_ring_write(ring, cur_pages | (stride_pixels << 16)); + } + radeon_ring_write(ring, PACKET0(RADEON_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, RADEON_RB2D_DC_FLUSH_ALL); + radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(ring, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_HOST_IDLECLEAN | RADEON_WAIT_DMA_GUI_IDLE); if (fence) { - r = radeon_fence_emit(rdev, fence); + r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX); } - radeon_ring_unlock_commit(rdev); + radeon_ring_unlock_commit(rdev, ring); return r; } @@ -512,39 +949,32 @@ static int r100_cp_wait_for_idle(struct radeon_device *rdev) return -1; } -void r100_ring_start(struct radeon_device *rdev) +void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring) { int r; - r = radeon_ring_lock(rdev, 2); + r = radeon_ring_lock(rdev, ring, 2); if (r) { return; } - radeon_ring_write(rdev, PACKET0(RADEON_ISYNC_CNTL, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(RADEON_ISYNC_CNTL, 0)); + radeon_ring_write(ring, RADEON_ISYNC_ANY2D_IDLE3D | RADEON_ISYNC_ANY3D_IDLE2D | RADEON_ISYNC_WAIT_IDLEGUI | RADEON_ISYNC_CPSCRATCH_IDLEGUI); - radeon_ring_unlock_commit(rdev); + radeon_ring_unlock_commit(rdev, 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("\n"); + 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)) { @@ -586,8 +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); @@ -602,6 +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; @@ -627,6 +1086,7 @@ static void r100_cp_load_microcode(struct radeon_device *rdev) int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; unsigned rb_bufsz; unsigned rb_blksz; unsigned max_fetch; @@ -640,26 +1100,6 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) if (r100_debugfs_cp_init(rdev)) { DRM_ERROR("Failed to register debugfs file for CP !\n"); } - /* Reset CP */ - tmp = RREG32(RADEON_CP_CSQ_STAT); - if ((tmp & (1 << 31))) { - DRM_INFO("radeon: cp busy (0x%08X) resetting\n", tmp); - WREG32(RADEON_CP_CSQ_MODE, 0); - WREG32(RADEON_CP_CSQ_CNTL, 0); - WREG32(RADEON_RBBM_SOFT_RESET, RADEON_SOFT_RESET_CP); - tmp = RREG32(RADEON_RBBM_SOFT_RESET); - mdelay(2); - WREG32(RADEON_RBBM_SOFT_RESET, 0); - tmp = RREG32(RADEON_RBBM_SOFT_RESET); - mdelay(2); - tmp = RREG32(RADEON_CP_CSQ_STAT); - if ((tmp & (1 << 31))) { - DRM_INFO("radeon: cp reset failed (0x%08X)\n", tmp); - } - } else { - DRM_INFO("radeon: cp idle (0x%08X)\n", tmp); - } - if (!rdev->me_fw) { r = r100_cp_init_microcode(rdev); if (r) { @@ -669,10 +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_size); + r = radeon_ring_init(rdev, ring, ring_size, RADEON_WB_CP_RPTR_OFFSET, + RADEON_CP_PACKET2); if (r) { return r; } @@ -681,7 +1122,7 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) rb_blksz = 9; /* cp will read 128bytes at a time (4 dwords) */ max_fetch = 1; - rdev->cp.align_mask = 16 - 1; + ring->align_mask = 16 - 1; /* Write to CP_RB_WPTR will be delayed for pre_write_timer clocks */ pre_write_timer = 64; /* Force CP_RB_WPTR write if written more than one time before the @@ -704,38 +1145,63 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) WREG32(0x718, pre_write_timer | (pre_write_limit << 28)); tmp = (REG_SET(RADEON_RB_BUFSZ, rb_bufsz) | REG_SET(RADEON_RB_BLKSZ, rb_blksz) | - REG_SET(RADEON_MAX_FETCH, max_fetch) | - RADEON_RB_NO_UPDATE); + REG_SET(RADEON_MAX_FETCH, max_fetch)); #ifdef __BIG_ENDIAN tmp |= RADEON_BUF_SWAP_32BIT; #endif - WREG32(RADEON_CP_RB_CNTL, tmp); + WREG32(RADEON_CP_RB_CNTL, tmp | RADEON_RB_NO_UPDATE); /* Set ring address */ - DRM_INFO("radeon: ring at 0x%016lX\n", (unsigned long)rdev->cp.gpu_addr); - WREG32(RADEON_CP_RB_BASE, rdev->cp.gpu_addr); + DRM_INFO("radeon: ring at 0x%016lX\n", (unsigned long)ring->gpu_addr); + WREG32(RADEON_CP_RB_BASE, ring->gpu_addr); /* Force read & write ptr to 0 */ - WREG32(RADEON_CP_RB_CNTL, tmp | RADEON_RB_RPTR_WR_ENA); + WREG32(RADEON_CP_RB_CNTL, tmp | RADEON_RB_RPTR_WR_ENA | RADEON_RB_NO_UPDATE); WREG32(RADEON_CP_RB_RPTR_WR, 0); - WREG32(RADEON_CP_RB_WPTR, 0); + ring->wptr = 0; + WREG32(RADEON_CP_RB_WPTR, ring->wptr); + + /* set the wb address whether it's enabled or not */ + WREG32(R_00070C_CP_RB_RPTR_ADDR, + S_00070C_RB_RPTR_ADDR((rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) >> 2)); + WREG32(R_000774_SCRATCH_ADDR, rdev->wb.gpu_addr + RADEON_WB_SCRATCH_OFFSET); + + if (rdev->wb.enabled) + WREG32(R_000770_SCRATCH_UMSK, 0xff); + else { + tmp |= RADEON_RB_NO_UPDATE; + WREG32(R_000770_SCRATCH_UMSK, 0); + } + WREG32(RADEON_CP_RB_CNTL, tmp); udelay(10); - rdev->cp.rptr = RREG32(RADEON_CP_RB_RPTR); - rdev->cp.wptr = RREG32(RADEON_CP_RB_WPTR); /* Set cp mode to bus mastering & enable cp*/ WREG32(RADEON_CP_CSQ_MODE, REG_SET(RADEON_INDIRECT2_START, indirect2_start) | REG_SET(RADEON_INDIRECT1_START, indirect1_start)); - WREG32(0x718, 0); - WREG32(0x744, 0x00004D4D); + WREG32(RADEON_CP_RB_WPTR_DELAY, 0); + WREG32(RADEON_CP_CSQ_MODE, 0x00004D4D); WREG32(RADEON_CP_CSQ_CNTL, RADEON_CSQ_PRIBM_INDBM); - radeon_ring_start(rdev); - r = radeon_ring_test(rdev); + + /* at this point everything should be setup correctly to enable master */ + pci_set_master(rdev->pdev); + + radeon_ring_start(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); + r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring); if (r) { DRM_ERROR("radeon: cp isn't working (%d).\n", r); return r; } - rdev->cp.ready = true; + ring->ready = true; + radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); + + if (!ring->rptr_save_reg /* not resuming from suspend */ + && 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; + } + } return 0; } @@ -746,65 +1212,134 @@ void r100_cp_fini(struct radeon_device *rdev) } /* Disable ring */ r100_cp_disable(rdev); - radeon_ring_fini(rdev); + radeon_scratch_free(rdev, rdev->ring[RADEON_RING_TYPE_GFX_INDEX].rptr_save_reg); + radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); DRM_INFO("radeon: cp finalized\n"); } void r100_cp_disable(struct radeon_device *rdev) { /* Disable ring */ - rdev->cp.ready = false; + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; WREG32(RADEON_CP_CSQ_MODE, 0); WREG32(RADEON_CP_CSQ_CNTL, 0); + WREG32(R_000770_SCRATCH_UMSK, 0); if (r100_gui_wait_for_idle(rdev)) { printk(KERN_WARNING "Failed to wait GUI idle while " "programming pipes. Bad things might happen.\n"); } } -int r100_cp_reset(struct radeon_device *rdev) +/* + * CS functions + */ +int r100_reloc_pitch_offset(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt, + unsigned idx, + unsigned reg) { - uint32_t tmp; - bool reinit_cp; - int i; + int r; + u32 tile_flags = 0; + u32 tmp; + struct radeon_cs_reloc *reloc; + u32 value; - reinit_cp = rdev->cp.ready; - rdev->cp.ready = false; - WREG32(RADEON_CP_CSQ_MODE, 0); - WREG32(RADEON_CP_CSQ_CNTL, 0); - WREG32(RADEON_RBBM_SOFT_RESET, RADEON_SOFT_RESET_CP); - (void)RREG32(RADEON_RBBM_SOFT_RESET); - udelay(200); - WREG32(RADEON_RBBM_SOFT_RESET, 0); - /* Wait to prevent race in RBBM_STATUS */ - mdelay(1); - for (i = 0; i < rdev->usec_timeout; i++) { - tmp = RREG32(RADEON_RBBM_STATUS); - if (!(tmp & (1 << 16))) { - DRM_INFO("CP reset succeed (RBBM_STATUS=0x%08X)\n", - tmp); - if (reinit_cp) { - return r100_cp_init(rdev, rdev->cp.ring_size); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); + if (r) { + DRM_ERROR("No reloc for ib[%d]=0x%04X\n", + idx, reg); + radeon_cs_dump_packet(p, pkt); + return r; + } + + value = radeon_get_ib_value(p, idx); + tmp = value & 0x003fffff; + tmp += (((u32)reloc->gpu_offset) >> 10); + + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) + tile_flags |= RADEON_DST_TILE_MACRO; + 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); + return -EINVAL; } - return 0; + tile_flags |= RADEON_DST_TILE_MICRO; } - DRM_UDELAY(1); - } - tmp = RREG32(RADEON_RBBM_STATUS); - DRM_ERROR("Failed to reset CP (RBBM_STATUS=0x%08X)!\n", tmp); - return -1; + + tmp |= tile_flags; + p->ib.ptr[idx] = (value & 0x3fc00000) | tmp; + } else + p->ib.ptr[idx] = (value & 0xffc00000) | tmp; + return 0; } -void r100_cp_commit(struct radeon_device *rdev) +int r100_packet3_load_vbpntr(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt, + int idx) { - WREG32(RADEON_CP_RB_WPTR, rdev->cp.wptr); - (void)RREG32(RADEON_CP_RB_WPTR); -} + unsigned c, i; + struct radeon_cs_reloc *reloc; + struct r100_cs_track *track; + int r = 0; + volatile uint32_t *ib; + u32 idx_value; + ib = p->ib.ptr; + track = (struct r100_cs_track *)p->track; + c = radeon_get_ib_value(p, idx++) & 0x1F; + if (c > 16) { + DRM_ERROR("Only 16 vertex buffers are allowed %d\n", + pkt->opcode); + radeon_cs_dump_packet(p, pkt); + return -EINVAL; + } + track->num_arrays = c; + for (i = 0; i < (c - 1); i+=2, idx+=3) { + r = radeon_cs_packet_next_reloc(p, &reloc, 0); + if (r) { + DRM_ERROR("No reloc for packet3 %d\n", + pkt->opcode); + radeon_cs_dump_packet(p, pkt); + return r; + } + idx_value = radeon_get_ib_value(p, idx); + 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; + track->arrays[i + 0].esize &= 0x7F; + r = radeon_cs_packet_next_reloc(p, &reloc, 0); + if (r) { + DRM_ERROR("No reloc for packet3 %d\n", + pkt->opcode); + radeon_cs_dump_packet(p, pkt); + return r; + } + 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; + } + if (c & 1) { + r = radeon_cs_packet_next_reloc(p, &reloc, 0); + if (r) { + DRM_ERROR("No reloc for packet3 %d\n", + pkt->opcode); + radeon_cs_dump_packet(p, pkt); + return r; + } + idx_value = radeon_get_ib_value(p, idx); + 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; + } + return r; +} -/* - * CS functions - */ int r100_cs_parse_packet0(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt, const unsigned *auth, unsigned n, @@ -850,67 +1385,6 @@ int r100_cs_parse_packet0(struct radeon_cs_parser *p, return 0; } -void r100_cs_dump_packet(struct radeon_cs_parser *p, - struct radeon_cs_packet *pkt) -{ - volatile uint32_t *ib; - unsigned i; - unsigned idx; - - ib = p->ib->ptr; - idx = pkt->idx; - for (i = 0; i <= (pkt->count + 1); i++, idx++) { - DRM_INFO("ib[%d]=0x%08X\n", idx, ib[idx]); - } -} - -/** - * r100_cs_packet_parse() - parse cp packet and point ib index to next packet - * @parser: parser structure holding parsing context. - * @pkt: where to store packet informations - * - * Assume that chunk_ib_index is properly set. Will return -EINVAL - * if packet is bigger than remaining ib size. or if packets is unknown. - **/ -int r100_cs_packet_parse(struct radeon_cs_parser *p, - struct radeon_cs_packet *pkt, - unsigned idx) -{ - struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx]; - uint32_t header; - - if (idx >= ib_chunk->length_dw) { - DRM_ERROR("Can not parse packet at %d after CS end %d !\n", - idx, ib_chunk->length_dw); - return -EINVAL; - } - header = radeon_get_ib_value(p, idx); - pkt->idx = idx; - pkt->type = CP_PACKET_GET_TYPE(header); - pkt->count = CP_PACKET_GET_COUNT(header); - switch (pkt->type) { - case PACKET_TYPE0: - pkt->reg = CP_PACKET0_GET_REG(header); - pkt->one_reg_wr = CP_PACKET0_GET_ONE_REG_WR(header); - break; - case PACKET_TYPE3: - pkt->opcode = CP_PACKET3_GET_OPCODE(header); - break; - case PACKET_TYPE2: - pkt->count = -1; - break; - default: - DRM_ERROR("Unknown packet type %d at %d !\n", pkt->type, idx); - return -EINVAL; - } - if ((pkt->count + 1 + pkt->idx) >= ib_chunk->length_dw) { - DRM_ERROR("Packet (%d:%d:%d) end after CS buffer (%d) !\n", - pkt->idx, pkt->type, pkt->count, ib_chunk->length_dw); - return -EINVAL; - } - return 0; -} - /** * r100_cs_packet_next_vline() - parse userspace VLINE packet * @parser: parser structure holding parsing context. @@ -936,10 +1410,10 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p) uint32_t header, h_idx, reg; volatile uint32_t *ib; - ib = p->ib->ptr; + ib = p->ib.ptr; /* parse the wait until */ - r = r100_cs_packet_parse(p, &waitreloc, p->idx); + r = radeon_cs_packet_parse(p, &waitreloc, p->idx); if (r) return r; @@ -947,18 +1421,16 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p) if (waitreloc.reg != RADEON_WAIT_UNTIL || waitreloc.count != 0) { DRM_ERROR("vline wait had illegal wait until segment\n"); - r = -EINVAL; - return r; + return -EINVAL; } if (radeon_get_ib_value(p, waitreloc.idx + 1) != RADEON_WAIT_CRTC_VLINE) { DRM_ERROR("vline wait had illegal wait until\n"); - r = -EINVAL; - return r; + return -EINVAL; } /* jump over the NOP */ - r = r100_cs_packet_parse(p, &p3reloc, p->idx + waitreloc.count + 2); + r = radeon_cs_packet_parse(p, &p3reloc, p->idx + waitreloc.count + 2); if (r) return r; @@ -968,13 +1440,11 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p) header = radeon_get_ib_value(p, h_idx); crtc_id = radeon_get_ib_value(p, h_idx + 5); - reg = CP_PACKET0_GET_REG(header); - mutex_lock(&p->rdev->ddev->mode_config.mutex); + reg = R100_CP_PACKET0_GET_REG(header); 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); - r = -EINVAL; - goto out; + return -ENOENT; } crtc = obj_to_crtc(obj); radeon_crtc = to_radeon_crtc(crtc); @@ -996,62 +1466,12 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p) break; default: DRM_ERROR("unknown crtc reloc\n"); - r = -EINVAL; - goto out; + return -EINVAL; } ib[h_idx] = header; ib[h_idx + 3] |= RADEON_ENG_DISPLAY_SELECT_CRTC1; } -out: - mutex_unlock(&p->rdev->ddev->mode_config.mutex); - return r; -} -/** - * r100_cs_packet_next_reloc() - parse next packet which should be reloc packet3 - * @parser: parser structure holding parsing context. - * @data: pointer to relocation data - * @offset_start: starting offset - * @offset_mask: offset mask (to align start offset on) - * @reloc: reloc informations - * - * Check next packet is relocation packet3, do bo validation and compute - * GPU offset using the provided start. - **/ -int r100_cs_packet_next_reloc(struct radeon_cs_parser *p, - struct radeon_cs_reloc **cs_reloc) -{ - struct radeon_cs_chunk *relocs_chunk; - struct radeon_cs_packet p3reloc; - unsigned idx; - int r; - - if (p->chunk_relocs_idx == -1) { - DRM_ERROR("No relocation chunk !\n"); - return -EINVAL; - } - *cs_reloc = NULL; - relocs_chunk = &p->chunks[p->chunk_relocs_idx]; - r = r100_cs_packet_parse(p, &p3reloc, p->idx); - if (r) { - return r; - } - p->idx += p3reloc.count + 2; - if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) { - DRM_ERROR("No packet3 for relocation for packet at %d.\n", - p3reloc.idx); - r100_cs_dump_packet(p, &p3reloc); - return -EINVAL; - } - idx = radeon_get_ib_value(p, p3reloc.idx + 1); - if (idx >= relocs_chunk->length_dw) { - DRM_ERROR("Relocs at %d after relocations chunk end %d !\n", - idx, relocs_chunk->length_dw); - r100_cs_dump_packet(p, &p3reloc); - return -EINVAL; - } - /* FIXME: we assume reloc size is 4 dwords */ - *cs_reloc = p->relocs_ptr[(idx / 4)]; return 0; } @@ -1121,7 +1541,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, u32 tile_flags = 0; u32 idx_value; - ib = p->ib->ptr; + ib = p->ib.ptr; track = (struct r100_cs_track *)p->track; idx_value = radeon_get_ib_value(p, idx); @@ -1132,7 +1552,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", idx, reg); - r100_cs_dump_packet(p, pkt); + radeon_cs_dump_packet(p, pkt); return r; } break; @@ -1145,42 +1565,55 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; break; case RADEON_RB3D_DEPTHOFFSET: - r = r100_cs_packet_next_reloc(p, &reloc); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", idx, reg); - r100_cs_dump_packet(p, pkt); + radeon_cs_dump_packet(p, pkt); return r; } track->zb.robj = reloc->robj; track->zb.offset = idx_value; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + track->zb_dirty = true; + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_RB3D_COLOROFFSET: - r = r100_cs_packet_next_reloc(p, &reloc); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", idx, reg); - r100_cs_dump_packet(p, pkt); + radeon_cs_dump_packet(p, pkt); return r; } track->cb[0].robj = reloc->robj; track->cb[0].offset = idx_value; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + track->cb_dirty = true; + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_PP_TXOFFSET_0: case RADEON_PP_TXOFFSET_1: case RADEON_PP_TXOFFSET_2: i = (reg - RADEON_PP_TXOFFSET_0) / 24; - r = r100_cs_packet_next_reloc(p, &reloc); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", idx, reg); - r100_cs_dump_packet(p, pkt); + radeon_cs_dump_packet(p, pkt); return r; } - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) + tile_flags |= RADEON_TXO_MACRO_TILE; + 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->gpu_offset); + } else + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[i].robj = reloc->robj; + track->tex_dirty = true; break; case RADEON_PP_CUBIC_OFFSET_T0_0: case RADEON_PP_CUBIC_OFFSET_T0_1: @@ -1188,16 +1621,17 @@ static int r100_packet0_check(struct radeon_cs_parser *p, case RADEON_PP_CUBIC_OFFSET_T0_3: case RADEON_PP_CUBIC_OFFSET_T0_4: i = (reg - RADEON_PP_CUBIC_OFFSET_T0_0) / 4; - r = r100_cs_packet_next_reloc(p, &reloc); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", idx, reg); - r100_cs_dump_packet(p, pkt); + radeon_cs_dump_packet(p, pkt); 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; case RADEON_PP_CUBIC_OFFSET_T1_0: case RADEON_PP_CUBIC_OFFSET_T1_1: @@ -1205,16 +1639,17 @@ static int r100_packet0_check(struct radeon_cs_parser *p, case RADEON_PP_CUBIC_OFFSET_T1_3: case RADEON_PP_CUBIC_OFFSET_T1_4: i = (reg - RADEON_PP_CUBIC_OFFSET_T1_0) / 4; - r = r100_cs_packet_next_reloc(p, &reloc); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", idx, reg); - r100_cs_dump_packet(p, pkt); + radeon_cs_dump_packet(p, pkt); 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; case RADEON_PP_CUBIC_OFFSET_T2_0: case RADEON_PP_CUBIC_OFFSET_T2_1: @@ -1222,42 +1657,49 @@ static int r100_packet0_check(struct radeon_cs_parser *p, case RADEON_PP_CUBIC_OFFSET_T2_3: case RADEON_PP_CUBIC_OFFSET_T2_4: i = (reg - RADEON_PP_CUBIC_OFFSET_T2_0) / 4; - r = r100_cs_packet_next_reloc(p, &reloc); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", idx, reg); - r100_cs_dump_packet(p, pkt); + radeon_cs_dump_packet(p, pkt); 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; case RADEON_RE_WIDTH_HEIGHT: track->maxy = ((idx_value >> 16) & 0x7FF); + track->cb_dirty = true; + track->zb_dirty = true; break; case RADEON_RB3D_COLORPITCH: - r = r100_cs_packet_next_reloc(p, &reloc); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", idx, reg); - r100_cs_dump_packet(p, pkt); + radeon_cs_dump_packet(p, pkt); return r; } - - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) - tile_flags |= RADEON_COLOR_TILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) - tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; - - tmp = idx_value & ~(0x7 << 16); - tmp |= tile_flags; - ib[idx] = tmp; + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) + tile_flags |= RADEON_COLOR_TILE_ENABLE; + if (reloc->tiling_flags & RADEON_TILING_MICRO) + tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; + + tmp = idx_value & ~(0x7 << 16); + tmp |= tile_flags; + ib[idx] = tmp; + } else + ib[idx] = idx_value; track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK; + track->cb_dirty = true; break; case RADEON_RB3D_DEPTHPITCH: track->zb.pitch = idx_value & RADEON_DEPTHPITCH_MASK; + track->zb_dirty = true; break; case RADEON_RB3D_CNTL: switch ((idx_value >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f) { @@ -1282,6 +1724,8 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return -EINVAL; } track->z_enabled = !!(idx_value & RADEON_Z_ENABLE); + track->cb_dirty = true; + track->zb_dirty = true; break; case RADEON_RB3D_ZSTENCILCNTL: switch (idx_value & 0xf) { @@ -1299,22 +1743,24 @@ static int r100_packet0_check(struct radeon_cs_parser *p, default: break; } + track->zb_dirty = true; break; case RADEON_RB3D_ZPASS_ADDR: - r = r100_cs_packet_next_reloc(p, &reloc); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", idx, reg); - r100_cs_dump_packet(p, pkt); + 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: { uint32_t temp = idx_value >> 4; for (i = 0; i < track->num_texture; i++) track->textures[i].enabled = !!(temp & (1 << i)); + track->tex_dirty = true; } break; case RADEON_SE_VF_CNTL: @@ -1329,12 +1775,14 @@ static int r100_packet0_check(struct radeon_cs_parser *p, i = (reg - RADEON_PP_TEX_SIZE_0) / 8; track->textures[i].width = (idx_value & RADEON_TEX_USIZE_MASK) + 1; track->textures[i].height = ((idx_value & RADEON_TEX_VSIZE_MASK) >> RADEON_TEX_VSIZE_SHIFT) + 1; + track->tex_dirty = true; break; case RADEON_PP_TEX_PITCH_0: case RADEON_PP_TEX_PITCH_1: case RADEON_PP_TEX_PITCH_2: i = (reg - RADEON_PP_TEX_PITCH_0) / 8; track->textures[i].pitch = idx_value + 32; + track->tex_dirty = true; break; case RADEON_PP_TXFILTER_0: case RADEON_PP_TXFILTER_1: @@ -1348,6 +1796,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, tmp = (idx_value >> 27) & 0x7; if (tmp == 2 || tmp == 6) track->textures[i].roundup_h = false; + track->tex_dirty = true; break; case RADEON_PP_TXFORMAT_0: case RADEON_PP_TXFORMAT_1: @@ -1367,6 +1816,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, case RADEON_TXFORMAT_RGB332: case RADEON_TXFORMAT_Y8: track->textures[i].cpp = 1; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case RADEON_TXFORMAT_AI88: case RADEON_TXFORMAT_ARGB1555: @@ -1378,12 +1828,14 @@ static int r100_packet0_check(struct radeon_cs_parser *p, case RADEON_TXFORMAT_LDUDV655: case RADEON_TXFORMAT_DUDV88: track->textures[i].cpp = 2; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case RADEON_TXFORMAT_ARGB8888: case RADEON_TXFORMAT_RGBA8888: case RADEON_TXFORMAT_SHADOW32: case RADEON_TXFORMAT_LDUDUV8888: track->textures[i].cpp = 4; + track->textures[i].compress_format = R100_TRACK_COMP_NONE; break; case RADEON_TXFORMAT_DXT1: track->textures[i].cpp = 1; @@ -1397,6 +1849,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } track->textures[i].cube_info[4].width = 1 << ((idx_value >> 16) & 0xf); track->textures[i].cube_info[4].height = 1 << ((idx_value >> 20) & 0xf); + track->tex_dirty = true; break; case RADEON_PP_CUBIC_FACES_0: case RADEON_PP_CUBIC_FACES_1: @@ -1407,6 +1860,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->textures[i].cube_info[face].width = 1 << ((tmp >> (face * 8)) & 0xf); track->textures[i].cube_info[face].height = 1 << ((tmp >> ((face * 8) + 4)) & 0xf); } + track->tex_dirty = true; break; default: printk(KERN_ERR "Forbidden register 0x%04X in cs at %d\n", @@ -1443,7 +1897,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p, volatile uint32_t *ib; int r; - ib = p->ib->ptr; + ib = p->ib.ptr; idx = pkt->idx + 1; track = (struct r100_cs_track *)p->track; switch (pkt->opcode) { @@ -1453,13 +1907,13 @@ static int r100_packet3_check(struct radeon_cs_parser *p, return r; break; case PACKET3_INDX_BUFFER: - r = r100_cs_packet_next_reloc(p, &reloc); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); if (r) { DRM_ERROR("No reloc for packet3 %d\n", pkt->opcode); - r100_cs_dump_packet(p, pkt); + 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; @@ -1467,13 +1921,13 @@ static int r100_packet3_check(struct radeon_cs_parser *p, break; case 0x23: /* 3D_RNDR_GEN_INDX_PRIM on r100/r200 */ - r = r100_cs_packet_next_reloc(p, &reloc); + r = radeon_cs_packet_next_reloc(p, &reloc, 0); if (r) { DRM_ERROR("No reloc for packet3 %d\n", pkt->opcode); - r100_cs_dump_packet(p, pkt); + 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)); @@ -1493,6 +1947,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("PRIM_WALK must be 3 for IMMD draw\n"); return -EINVAL; } + track->vtx_size = r100_get_vtx_size(radeon_get_ib_value(p, idx + 0)); track->vap_vf_cntl = radeon_get_ib_value(p, idx + 1); track->immd_dwords = pkt->count - 1; r = r100_cs_track_check(p->rdev, track); @@ -1540,6 +1995,11 @@ static int r100_packet3_check(struct radeon_cs_parser *p, return r; break; /* triggers drawing using indices to vertex buffer */ + case PACKET3_3D_CLEAR_HIZ: + case PACKET3_3D_CLEAR_ZMASK: + if (p->rdev->hyperz_filp != p->filp) + return -EINVAL; + break; case PACKET3_NOP: break; default: @@ -1556,108 +2016,438 @@ int r100_cs_parse(struct radeon_cs_parser *p) int r; track = kzalloc(sizeof(*track), GFP_KERNEL); + if (!track) + return -ENOMEM; r100_cs_track_clear(p->rdev, track); p->track = track; do { - r = r100_cs_packet_parse(p, &pkt, p->idx); + r = radeon_cs_packet_parse(p, &pkt, p->idx); if (r) { return r; } p->idx += pkt.count + 2; switch (pkt.type) { - case PACKET_TYPE0: - if (p->rdev->family >= CHIP_R200) - r = r100_cs_parse_packet0(p, &pkt, - p->rdev->config.r100.reg_safe_bm, - p->rdev->config.r100.reg_safe_bm_size, - &r200_packet0_check); - else - r = r100_cs_parse_packet0(p, &pkt, - p->rdev->config.r100.reg_safe_bm, - p->rdev->config.r100.reg_safe_bm_size, - &r100_packet0_check); - break; - case PACKET_TYPE2: - break; - case PACKET_TYPE3: - r = r100_packet3_check(p, &pkt); - break; - default: - DRM_ERROR("Unknown packet type %d !\n", - pkt.type); - return -EINVAL; + case RADEON_PACKET_TYPE0: + if (p->rdev->family >= CHIP_R200) + r = r100_cs_parse_packet0(p, &pkt, + p->rdev->config.r100.reg_safe_bm, + p->rdev->config.r100.reg_safe_bm_size, + &r200_packet0_check); + else + r = r100_cs_parse_packet0(p, &pkt, + p->rdev->config.r100.reg_safe_bm, + p->rdev->config.r100.reg_safe_bm_size, + &r100_packet0_check); + break; + case RADEON_PACKET_TYPE2: + break; + case RADEON_PACKET_TYPE3: + r = r100_packet3_check(p, &pkt); + break; + default: + DRM_ERROR("Unknown packet type %d !\n", + pkt.type); + return -EINVAL; } - if (r) { + if (r) return r; - } } while (p->idx < p->chunks[p->chunk_ib_idx].length_dw); return 0; } +static void r100_cs_track_texture_print(struct r100_cs_track_texture *t) +{ + DRM_ERROR("pitch %d\n", t->pitch); + DRM_ERROR("use_pitch %d\n", t->use_pitch); + DRM_ERROR("width %d\n", t->width); + DRM_ERROR("width_11 %d\n", t->width_11); + DRM_ERROR("height %d\n", t->height); + DRM_ERROR("height_11 %d\n", t->height_11); + DRM_ERROR("num levels %d\n", t->num_levels); + DRM_ERROR("depth %d\n", t->txdepth); + DRM_ERROR("bpp %d\n", t->cpp); + DRM_ERROR("coordinate type %d\n", t->tex_coord_type); + DRM_ERROR("width round to power of 2 %d\n", t->roundup_w); + DRM_ERROR("height round to power of 2 %d\n", t->roundup_h); + DRM_ERROR("compress format %d\n", t->compress_format); +} -/* - * Global GPU functions - */ -void r100_errata(struct radeon_device *rdev) +static int r100_track_compress_size(int compress_format, int w, int h) { - rdev->pll_errata = 0; + int block_width, block_height, block_bytes; + int wblocks, hblocks; + int min_wblocks; + int sz; - if (rdev->family == CHIP_RV200 || rdev->family == CHIP_RS200) { - rdev->pll_errata |= CHIP_ERRATA_PLL_DUMMYREADS; + block_width = 4; + block_height = 4; + + switch (compress_format) { + case R100_TRACK_COMP_DXT1: + block_bytes = 8; + min_wblocks = 4; + break; + default: + case R100_TRACK_COMP_DXT35: + block_bytes = 16; + min_wblocks = 2; + break; } - if (rdev->family == CHIP_RV100 || - rdev->family == CHIP_RS100 || - rdev->family == CHIP_RS200) { - rdev->pll_errata |= CHIP_ERRATA_PLL_DELAY; + hblocks = (h + block_height - 1) / block_height; + wblocks = (w + block_width - 1) / block_width; + if (wblocks < min_wblocks) + wblocks = min_wblocks; + sz = wblocks * hblocks * block_bytes; + return sz; +} + +static int r100_cs_track_cube(struct radeon_device *rdev, + struct r100_cs_track *track, unsigned idx) +{ + unsigned face, w, h; + struct radeon_bo *cube_robj; + unsigned long size; + unsigned compress_format = track->textures[idx].compress_format; + + for (face = 0; face < 5; face++) { + cube_robj = track->textures[idx].cube_info[face].robj; + w = track->textures[idx].cube_info[face].width; + h = track->textures[idx].cube_info[face].height; + + if (compress_format) { + size = r100_track_compress_size(compress_format, w, h); + } else + size = w * h; + size *= track->textures[idx].cpp; + + size += track->textures[idx].cube_info[face].offset; + + if (size > radeon_bo_size(cube_robj)) { + DRM_ERROR("Cube texture offset greater than object size %lu %lu\n", + size, radeon_bo_size(cube_robj)); + r100_cs_track_texture_print(&track->textures[idx]); + return -1; + } } + return 0; } -/* Wait for vertical sync on primary CRTC */ -void r100_gpu_wait_for_vsync(struct radeon_device *rdev) +static int r100_cs_track_texture_check(struct radeon_device *rdev, + struct r100_cs_track *track) { - uint32_t crtc_gen_cntl, tmp; - int i; + struct radeon_bo *robj; + unsigned long size; + unsigned u, i, w, h, d; + int ret; - crtc_gen_cntl = RREG32(RADEON_CRTC_GEN_CNTL); - if ((crtc_gen_cntl & RADEON_CRTC_DISP_REQ_EN_B) || - !(crtc_gen_cntl & RADEON_CRTC_EN)) { - return; + for (u = 0; u < track->num_texture; u++) { + if (!track->textures[u].enabled) + continue; + if (track->textures[u].lookup_disable) + continue; + robj = track->textures[u].robj; + if (robj == NULL) { + DRM_ERROR("No texture bound to unit %u\n", u); + return -EINVAL; + } + size = 0; + for (i = 0; i <= track->textures[u].num_levels; i++) { + if (track->textures[u].use_pitch) { + if (rdev->family < CHIP_R300) + w = (track->textures[u].pitch / track->textures[u].cpp) / (1 << i); + else + w = track->textures[u].pitch / (1 << i); + } else { + w = track->textures[u].width; + if (rdev->family >= CHIP_RV515) + w |= track->textures[u].width_11; + w = w / (1 << i); + if (track->textures[u].roundup_w) + w = roundup_pow_of_two(w); + } + h = track->textures[u].height; + if (rdev->family >= CHIP_RV515) + h |= track->textures[u].height_11; + h = h / (1 << i); + if (track->textures[u].roundup_h) + h = roundup_pow_of_two(h); + if (track->textures[u].tex_coord_type == 1) { + d = (1 << track->textures[u].txdepth) / (1 << i); + if (!d) + d = 1; + } else { + d = 1; + } + if (track->textures[u].compress_format) { + + size += r100_track_compress_size(track->textures[u].compress_format, w, h) * d; + /* compressed textures are block based */ + } else + size += w * h * d; + } + size *= track->textures[u].cpp; + + switch (track->textures[u].tex_coord_type) { + case 0: + case 1: + break; + case 2: + if (track->separate_cube) { + ret = r100_cs_track_cube(rdev, track, u); + if (ret) + return ret; + } else + size *= 6; + break; + default: + DRM_ERROR("Invalid texture coordinate type %u for unit " + "%u\n", track->textures[u].tex_coord_type, u); + return -EINVAL; + } + if (size > radeon_bo_size(robj)) { + DRM_ERROR("Texture of unit %u needs %lu bytes but is " + "%lu\n", u, size, radeon_bo_size(robj)); + r100_cs_track_texture_print(&track->textures[u]); + return -EINVAL; + } } - /* Clear the CRTC_VBLANK_SAVE bit */ - WREG32(RADEON_CRTC_STATUS, RADEON_CRTC_VBLANK_SAVE_CLEAR); - for (i = 0; i < rdev->usec_timeout; i++) { - tmp = RREG32(RADEON_CRTC_STATUS); - if (tmp & RADEON_CRTC_VBLANK_SAVE) { - return; + return 0; +} + +int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) +{ + unsigned i; + unsigned long size; + unsigned prim_walk; + unsigned nverts; + unsigned num_cb = track->cb_dirty ? track->num_cb : 0; + + if (num_cb && !track->zb_cb_clear && !track->color_channel_mask && + !track->blend_read_enable) + num_cb = 0; + + for (i = 0; i < num_cb; i++) { + if (track->cb[i].robj == NULL) { + DRM_ERROR("[drm] No buffer for color buffer %d !\n", i); + return -EINVAL; + } + size = track->cb[i].pitch * track->cb[i].cpp * track->maxy; + size += track->cb[i].offset; + if (size > radeon_bo_size(track->cb[i].robj)) { + DRM_ERROR("[drm] Buffer too small for color buffer %d " + "(need %lu have %lu) !\n", i, size, + radeon_bo_size(track->cb[i].robj)); + DRM_ERROR("[drm] color buffer %d (%u %u %u %u)\n", + i, track->cb[i].pitch, track->cb[i].cpp, + track->cb[i].offset, track->maxy); + return -EINVAL; + } + } + track->cb_dirty = false; + + if (track->zb_dirty && track->z_enabled) { + if (track->zb.robj == NULL) { + DRM_ERROR("[drm] No buffer for z buffer !\n"); + return -EINVAL; + } + size = track->zb.pitch * track->zb.cpp * track->maxy; + size += track->zb.offset; + if (size > radeon_bo_size(track->zb.robj)) { + DRM_ERROR("[drm] Buffer too small for z buffer " + "(need %lu have %lu) !\n", size, + radeon_bo_size(track->zb.robj)); + DRM_ERROR("[drm] zbuffer (%u %u %u %u)\n", + track->zb.pitch, track->zb.cpp, + track->zb.offset, track->maxy); + return -EINVAL; + } + } + track->zb_dirty = false; + + if (track->aa_dirty && track->aaresolve) { + if (track->aa.robj == NULL) { + DRM_ERROR("[drm] No buffer for AA resolve buffer %d !\n", i); + return -EINVAL; + } + /* I believe the format comes from colorbuffer0. */ + size = track->aa.pitch * track->cb[0].cpp * track->maxy; + size += track->aa.offset; + if (size > radeon_bo_size(track->aa.robj)) { + DRM_ERROR("[drm] Buffer too small for AA resolve buffer %d " + "(need %lu have %lu) !\n", i, size, + radeon_bo_size(track->aa.robj)); + DRM_ERROR("[drm] AA resolve buffer %d (%u %u %u %u)\n", + i, track->aa.pitch, track->cb[0].cpp, + track->aa.offset, track->maxy); + return -EINVAL; } - DRM_UDELAY(1); } + track->aa_dirty = false; + + prim_walk = (track->vap_vf_cntl >> 4) & 0x3; + if (track->vap_vf_cntl & (1 << 14)) { + nverts = track->vap_alt_nverts; + } else { + nverts = (track->vap_vf_cntl >> 16) & 0xFFFF; + } + switch (prim_walk) { + case 1: + for (i = 0; i < track->num_arrays; i++) { + size = track->arrays[i].esize * track->max_indx * 4; + if (track->arrays[i].robj == NULL) { + DRM_ERROR("(PW %u) Vertex array %u no buffer " + "bound\n", prim_walk, i); + return -EINVAL; + } + if (size > radeon_bo_size(track->arrays[i].robj)) { + dev_err(rdev->dev, "(PW %u) Vertex array %u " + "need %lu dwords have %lu dwords\n", + prim_walk, i, size >> 2, + radeon_bo_size(track->arrays[i].robj) + >> 2); + DRM_ERROR("Max indices %u\n", track->max_indx); + return -EINVAL; + } + } + break; + case 2: + for (i = 0; i < track->num_arrays; i++) { + size = track->arrays[i].esize * (nverts - 1) * 4; + if (track->arrays[i].robj == NULL) { + DRM_ERROR("(PW %u) Vertex array %u no buffer " + "bound\n", prim_walk, i); + return -EINVAL; + } + if (size > radeon_bo_size(track->arrays[i].robj)) { + dev_err(rdev->dev, "(PW %u) Vertex array %u " + "need %lu dwords have %lu dwords\n", + prim_walk, i, size >> 2, + radeon_bo_size(track->arrays[i].robj) + >> 2); + return -EINVAL; + } + } + break; + case 3: + size = track->vtx_size * nverts; + if (size != track->immd_dwords) { + DRM_ERROR("IMMD draw %u dwors but needs %lu dwords\n", + track->immd_dwords, size); + DRM_ERROR("VAP_VF_CNTL.NUM_VERTICES %u, VTX_SIZE %u\n", + nverts, track->vtx_size); + return -EINVAL; + } + break; + default: + DRM_ERROR("[drm] Invalid primitive walk %d for VAP_VF_CNTL\n", + prim_walk); + return -EINVAL; + } + + if (track->tex_dirty) { + track->tex_dirty = false; + return r100_cs_track_texture_check(rdev, track); + } + return 0; } -/* Wait for vertical sync on secondary CRTC */ -void r100_gpu_wait_for_vsync2(struct radeon_device *rdev) +void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track) { - uint32_t crtc2_gen_cntl, tmp; - int i; + unsigned i, face; - crtc2_gen_cntl = RREG32(RADEON_CRTC2_GEN_CNTL); - if ((crtc2_gen_cntl & RADEON_CRTC2_DISP_REQ_EN_B) || - !(crtc2_gen_cntl & RADEON_CRTC2_EN)) - return; + track->cb_dirty = true; + track->zb_dirty = true; + track->tex_dirty = true; + track->aa_dirty = true; - /* Clear the CRTC_VBLANK_SAVE bit */ - WREG32(RADEON_CRTC2_STATUS, RADEON_CRTC2_VBLANK_SAVE_CLEAR); - for (i = 0; i < rdev->usec_timeout; i++) { - tmp = RREG32(RADEON_CRTC2_STATUS); - if (tmp & RADEON_CRTC2_VBLANK_SAVE) { - return; + if (rdev->family < CHIP_R300) { + track->num_cb = 1; + if (rdev->family <= CHIP_RS200) + track->num_texture = 3; + else + track->num_texture = 6; + track->maxy = 2048; + track->separate_cube = 1; + } else { + track->num_cb = 4; + track->num_texture = 16; + track->maxy = 4096; + track->separate_cube = 0; + track->aaresolve = false; + track->aa.robj = NULL; + } + + for (i = 0; i < track->num_cb; i++) { + track->cb[i].robj = NULL; + track->cb[i].pitch = 8192; + track->cb[i].cpp = 16; + track->cb[i].offset = 0; + } + track->z_enabled = true; + track->zb.robj = NULL; + track->zb.pitch = 8192; + track->zb.cpp = 4; + track->zb.offset = 0; + track->vtx_size = 0x7F; + track->immd_dwords = 0xFFFFFFFFUL; + track->num_arrays = 11; + track->max_indx = 0x00FFFFFFUL; + for (i = 0; i < track->num_arrays; i++) { + track->arrays[i].robj = NULL; + track->arrays[i].esize = 0x7F; + } + for (i = 0; i < track->num_texture; i++) { + track->textures[i].compress_format = R100_TRACK_COMP_NONE; + track->textures[i].pitch = 16536; + track->textures[i].width = 16536; + track->textures[i].height = 16536; + track->textures[i].width_11 = 1 << 11; + track->textures[i].height_11 = 1 << 11; + track->textures[i].num_levels = 12; + if (rdev->family <= CHIP_RS200) { + track->textures[i].tex_coord_type = 0; + track->textures[i].txdepth = 0; + } else { + track->textures[i].txdepth = 16; + track->textures[i].tex_coord_type = 1; } - DRM_UDELAY(1); + track->textures[i].cpp = 64; + track->textures[i].robj = NULL; + /* CS IB emission code makes sure texture unit are disabled */ + track->textures[i].enabled = false; + track->textures[i].lookup_disable = false; + track->textures[i].roundup_w = true; + track->textures[i].roundup_h = true; + if (track->separate_cube) + for (face = 0; face < 5; face++) { + track->textures[i].cube_info[face].robj = NULL; + track->textures[i].cube_info[face].width = 16536; + track->textures[i].cube_info[face].height = 16536; + track->textures[i].cube_info[face].offset = 0; + } + } +} + +/* + * Global GPU functions + */ +static void r100_errata(struct radeon_device *rdev) +{ + rdev->pll_errata = 0; + + if (rdev->family == CHIP_RV200 || rdev->family == CHIP_RS200) { + rdev->pll_errata |= CHIP_ERRATA_PLL_DUMMYREADS; + } + + if (rdev->family == CHIP_RV100 || + rdev->family == CHIP_RS100 || + rdev->family == CHIP_RS200) { + rdev->pll_errata |= CHIP_ERRATA_PLL_DELAY; } } -int r100_rbbm_fifo_wait_for_entry(struct radeon_device *rdev, unsigned n) +static int r100_rbbm_fifo_wait_for_entry(struct radeon_device *rdev, unsigned n) { unsigned i; uint32_t tmp; @@ -1683,7 +2473,7 @@ int r100_gui_wait_for_idle(struct radeon_device *rdev) } for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(RADEON_RBBM_STATUS); - if (!(tmp & (1 << 31))) { + if (!(tmp & RADEON_RBBM_ACTIVE)) { return 0; } DRM_UDELAY(1); @@ -1698,8 +2488,8 @@ int r100_mc_wait_for_idle(struct radeon_device *rdev) for (i = 0; i < rdev->usec_timeout; i++) { /* read MC_STATUS */ - tmp = RREG32(0x0150); - if (tmp & (1 << 2)) { + tmp = RREG32(RADEON_MC_STATUS); + if (tmp & RADEON_MC_IDLE) { return 0; } DRM_UDELAY(1); @@ -1707,89 +2497,106 @@ int r100_mc_wait_for_idle(struct radeon_device *rdev) return -1; } -void r100_gpu_init(struct radeon_device *rdev) +bool r100_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) { - /* TODO: anythings to do here ? pipes ? */ - r100_hdp_reset(rdev); -} + u32 rbbm_status; -void r100_hdp_flush(struct radeon_device *rdev) -{ - u32 tmp; - tmp = RREG32(RADEON_HOST_PATH_CNTL); - tmp |= RADEON_HDP_READ_BUFFER_INVALIDATE; - WREG32(RADEON_HOST_PATH_CNTL, tmp); + rbbm_status = RREG32(R_000E40_RBBM_STATUS); + if (!G_000E40_GUI_ACTIVE(rbbm_status)) { + radeon_ring_lockup_update(rdev, ring); + return false; + } + return radeon_ring_test_lockup(rdev, ring); } -void r100_hdp_reset(struct radeon_device *rdev) +/* required on r1xx, r2xx, r300, r(v)350, r420/r481, rs400/rs480 */ +void r100_enable_bm(struct radeon_device *rdev) { uint32_t tmp; - - tmp = RREG32(RADEON_HOST_PATH_CNTL) & RADEON_HDP_APER_CNTL; - tmp |= (7 << 28); - WREG32(RADEON_HOST_PATH_CNTL, tmp | RADEON_HDP_SOFT_RESET | RADEON_HDP_READ_BUFFER_INVALIDATE); - (void)RREG32(RADEON_HOST_PATH_CNTL); - udelay(200); - WREG32(RADEON_RBBM_SOFT_RESET, 0); - WREG32(RADEON_HOST_PATH_CNTL, tmp); - (void)RREG32(RADEON_HOST_PATH_CNTL); + /* Enable bus mastering */ + tmp = RREG32(RADEON_BUS_CNTL) & ~RADEON_BUS_MASTER_DIS; + WREG32(RADEON_BUS_CNTL, tmp); } -int r100_rb2d_reset(struct radeon_device *rdev) +void r100_bm_disable(struct radeon_device *rdev) { - uint32_t tmp; - int i; + u32 tmp; - WREG32(RADEON_RBBM_SOFT_RESET, RADEON_SOFT_RESET_E2); - (void)RREG32(RADEON_RBBM_SOFT_RESET); - udelay(200); - WREG32(RADEON_RBBM_SOFT_RESET, 0); - /* Wait to prevent race in RBBM_STATUS */ + /* disable bus mastering */ + tmp = RREG32(R_000030_BUS_CNTL); + WREG32(R_000030_BUS_CNTL, (tmp & 0xFFFFFFFF) | 0x00000044); + mdelay(1); + WREG32(R_000030_BUS_CNTL, (tmp & 0xFFFFFFFF) | 0x00000042); + mdelay(1); + WREG32(R_000030_BUS_CNTL, (tmp & 0xFFFFFFFF) | 0x00000040); + tmp = RREG32(RADEON_BUS_CNTL); + mdelay(1); + pci_clear_master(rdev->pdev); mdelay(1); - for (i = 0; i < rdev->usec_timeout; i++) { - tmp = RREG32(RADEON_RBBM_STATUS); - if (!(tmp & (1 << 26))) { - DRM_INFO("RB2D reset succeed (RBBM_STATUS=0x%08X)\n", - tmp); - return 0; - } - DRM_UDELAY(1); - } - tmp = RREG32(RADEON_RBBM_STATUS); - DRM_ERROR("Failed to reset RB2D (RBBM_STATUS=0x%08X)!\n", tmp); - return -1; } -int r100_gpu_reset(struct radeon_device *rdev) +int r100_asic_reset(struct radeon_device *rdev) { - uint32_t status; + struct r100_mc_save save; + u32 status, tmp; + int ret = 0; - /* reset order likely matter */ - status = RREG32(RADEON_RBBM_STATUS); - /* reset HDP */ - r100_hdp_reset(rdev); - /* reset rb2d */ - if (status & ((1 << 17) | (1 << 18) | (1 << 27))) { - r100_rb2d_reset(rdev); + status = RREG32(R_000E40_RBBM_STATUS); + if (!G_000E40_GUI_ACTIVE(status)) { + return 0; } - /* TODO: reset 3D engine */ + r100_mc_stop(rdev, &save); + status = RREG32(R_000E40_RBBM_STATUS); + dev_info(rdev->dev, "(%s:%d) RBBM_STATUS=0x%08X\n", __func__, __LINE__, status); + /* stop CP */ + WREG32(RADEON_CP_CSQ_CNTL, 0); + tmp = RREG32(RADEON_CP_RB_CNTL); + WREG32(RADEON_CP_RB_CNTL, tmp | RADEON_RB_RPTR_WR_ENA); + WREG32(RADEON_CP_RB_RPTR_WR, 0); + WREG32(RADEON_CP_RB_WPTR, 0); + WREG32(RADEON_CP_RB_CNTL, tmp); + /* save PCI state */ + pci_save_state(rdev->pdev); + /* disable bus mastering */ + r100_bm_disable(rdev); + WREG32(R_0000F0_RBBM_SOFT_RESET, S_0000F0_SOFT_RESET_SE(1) | + S_0000F0_SOFT_RESET_RE(1) | + S_0000F0_SOFT_RESET_PP(1) | + S_0000F0_SOFT_RESET_RB(1)); + RREG32(R_0000F0_RBBM_SOFT_RESET); + mdelay(500); + WREG32(R_0000F0_RBBM_SOFT_RESET, 0); + mdelay(1); + status = RREG32(R_000E40_RBBM_STATUS); + dev_info(rdev->dev, "(%s:%d) RBBM_STATUS=0x%08X\n", __func__, __LINE__, status); /* reset CP */ - status = RREG32(RADEON_RBBM_STATUS); - if (status & (1 << 16)) { - r100_cp_reset(rdev); - } + WREG32(R_0000F0_RBBM_SOFT_RESET, S_0000F0_SOFT_RESET_CP(1)); + RREG32(R_0000F0_RBBM_SOFT_RESET); + mdelay(500); + WREG32(R_0000F0_RBBM_SOFT_RESET, 0); + mdelay(1); + status = RREG32(R_000E40_RBBM_STATUS); + dev_info(rdev->dev, "(%s:%d) RBBM_STATUS=0x%08X\n", __func__, __LINE__, status); + /* restore PCI & busmastering */ + pci_restore_state(rdev->pdev); + r100_enable_bm(rdev); /* Check if GPU is idle */ - status = RREG32(RADEON_RBBM_STATUS); - if (status & (1 << 31)) { - DRM_ERROR("Failed to reset GPU (RBBM_STATUS=0x%08X)\n", status); - return -1; - } - DRM_INFO("GPU reset succeed (RBBM_STATUS=0x%08X)\n", status); - return 0; + if (G_000E40_SE_BUSY(status) || G_000E40_RE_BUSY(status) || + G_000E40_TAM_BUSY(status) || G_000E40_PB_BUSY(status)) { + dev_err(rdev->dev, "failed to reset GPU\n"); + ret = -1; + } else + dev_info(rdev->dev, "GPU reset succeed\n"); + r100_mc_resume(rdev, &save); + return ret; } void r100_set_common_regs(struct radeon_device *rdev) { + struct drm_device *dev = rdev->ddev; + bool force_dac2 = false; + u32 tmp; + /* set these so they don't interfere with anything */ WREG32(RADEON_OV0_SCALE_CNTL, 0); WREG32(RADEON_SUBPIC_CNTL, 0); @@ -1798,6 +2605,74 @@ void r100_set_common_regs(struct radeon_device *rdev) WREG32(RADEON_DVI_I2C_CNTL_1, 0); WREG32(RADEON_CAP0_TRIG_CNTL, 0); WREG32(RADEON_CAP1_TRIG_CNTL, 0); + + /* always set up dac2 on rn50 and some rv100 as lots + * of servers seem to wire it up to a VGA port but + * don't report it in the bios connector + * table. + */ + switch (dev->pdev->device) { + /* RN50 */ + case 0x515e: + case 0x5969: + force_dac2 = true; + break; + /* RV100*/ + case 0x5159: + case 0x515a: + /* DELL triple head servers */ + if ((dev->pdev->subsystem_vendor == 0x1028 /* DELL */) && + ((dev->pdev->subsystem_device == 0x016c) || + (dev->pdev->subsystem_device == 0x016d) || + (dev->pdev->subsystem_device == 0x016e) || + (dev->pdev->subsystem_device == 0x016f) || + (dev->pdev->subsystem_device == 0x0170) || + (dev->pdev->subsystem_device == 0x017d) || + (dev->pdev->subsystem_device == 0x017e) || + (dev->pdev->subsystem_device == 0x0183) || + (dev->pdev->subsystem_device == 0x018a) || + (dev->pdev->subsystem_device == 0x019a))) + force_dac2 = true; + break; + } + + if (force_dac2) { + u32 disp_hw_debug = RREG32(RADEON_DISP_HW_DEBUG); + u32 tv_dac_cntl = RREG32(RADEON_TV_DAC_CNTL); + u32 dac2_cntl = RREG32(RADEON_DAC_CNTL2); + + /* For CRT on DAC2, don't turn it on if BIOS didn't + enable it, even it's detected. + */ + + /* force it to crtc0 */ + dac2_cntl &= ~RADEON_DAC2_DAC_CLK_SEL; + dac2_cntl |= RADEON_DAC2_DAC2_CLK_SEL; + disp_hw_debug |= RADEON_CRT2_DISP1_SEL; + + /* set up the TV DAC */ + tv_dac_cntl &= ~(RADEON_TV_DAC_PEDESTAL | + RADEON_TV_DAC_STD_MASK | + RADEON_TV_DAC_RDACPD | + RADEON_TV_DAC_GDACPD | + RADEON_TV_DAC_BDACPD | + RADEON_TV_DAC_BGADJ_MASK | + RADEON_TV_DAC_DACADJ_MASK); + tv_dac_cntl |= (RADEON_TV_DAC_NBLANK | + RADEON_TV_DAC_NHOLD | + RADEON_TV_DAC_STD_PS2 | + (0x58 << 16)); + + WREG32(RADEON_TV_DAC_CNTL, tv_dac_cntl); + WREG32(RADEON_DISP_HW_DEBUG, disp_hw_debug); + WREG32(RADEON_DAC_CNTL2, dac2_cntl); + } + + /* switch PM block to ACPI mode */ + tmp = RREG32_PLL(RADEON_PLL_PWRMGT_CNTL); + tmp &= ~RADEON_PM_MODE_SEL; + WREG32_PLL(RADEON_PLL_PWRMGT_CNTL, tmp); + } /* @@ -1879,17 +2754,20 @@ static u32 r100_get_accessible_vram(struct radeon_device *rdev) void r100_vram_init_sizes(struct radeon_device *rdev) { u64 config_aper_size; - u32 accessible; + /* work out accessible VRAM */ + rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0); + rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0); + rdev->mc.visible_vram_size = r100_get_accessible_vram(rdev); + /* FIXME we don't use the second aperture yet when we could use it */ + if (rdev->mc.visible_vram_size > rdev->mc.aper_size) + rdev->mc.visible_vram_size = rdev->mc.aper_size; config_aper_size = RREG32(RADEON_CONFIG_APER_SIZE); - if (rdev->flags & RADEON_IS_IGP) { uint32_t tom; /* read NB_TOM to get the amount of ram stolen for the GPU */ tom = RREG32(RADEON_NB_TOM); rdev->mc.real_vram_size = (((tom >> 16) - (tom & 0xffff) + 1) << 16); - /* for IGPs we need to keep VRAM where it was put by the BIOS */ - rdev->mc.vram_location = (tom & 0xffff) << 16; WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size); rdev->mc.mc_vram_size = rdev->mc.real_vram_size; } else { @@ -1901,30 +2779,17 @@ void r100_vram_init_sizes(struct radeon_device *rdev) rdev->mc.real_vram_size = 8192 * 1024; WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size); } - /* let driver place VRAM */ - rdev->mc.vram_location = 0xFFFFFFFFUL; - /* Fix for RN50, M6, M7 with 8/16/32(??) MBs of VRAM - - * Novell bug 204882 + along with lots of ubuntu ones */ + /* Fix for RN50, M6, M7 with 8/16/32(??) MBs of VRAM - + * Novell bug 204882 + along with lots of ubuntu ones + */ + if (rdev->mc.aper_size > config_aper_size) + config_aper_size = rdev->mc.aper_size; + if (config_aper_size > rdev->mc.real_vram_size) rdev->mc.mc_vram_size = config_aper_size; else rdev->mc.mc_vram_size = rdev->mc.real_vram_size; } - - /* work out accessible VRAM */ - accessible = r100_get_accessible_vram(rdev); - - rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0); - rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0); - - if (accessible > rdev->mc.aper_size) - accessible = rdev->mc.aper_size; - - if (rdev->mc.mc_vram_size > rdev->mc.aper_size) - rdev->mc.mc_vram_size = rdev->mc.aper_size; - - if (rdev->mc.real_vram_size > rdev->mc.aper_size) - rdev->mc.real_vram_size = rdev->mc.aper_size; } void r100_vga_set_state(struct radeon_device *rdev, bool state) @@ -1933,19 +2798,28 @@ void r100_vga_set_state(struct radeon_device *rdev, bool state) temp = RREG32(RADEON_CONFIG_CNTL); if (state == false) { - temp &= ~(1<<8); - temp |= (1<<9); + temp &= ~RADEON_CFG_VGA_RAM_EN; + temp |= RADEON_CFG_VGA_IO_DIS; } else { - temp &= ~(1<<9); + temp &= ~RADEON_CFG_VGA_IO_DIS; } WREG32(RADEON_CONFIG_CNTL, temp); } -void r100_vram_info(struct radeon_device *rdev) +static void r100_mc_init(struct radeon_device *rdev) { - r100_vram_get_type(rdev); + u64 base; + r100_vram_get_type(rdev); r100_vram_init_sizes(rdev); + base = rdev->mc.aper_base; + if (rdev->flags & RADEON_IS_IGP) + base = (RREG32(RADEON_NB_TOM) & 0xffff) << 16; + radeon_vram_location(rdev, &rdev->mc, base); + rdev->mc.gtt_base_align = 0; + if (!(rdev->flags & RADEON_IS_AGP)) + radeon_gtt_location(rdev, &rdev->mc); + radeon_update_bandwidth_info(rdev); } @@ -1954,11 +2828,10 @@ void r100_vram_info(struct radeon_device *rdev) */ void r100_pll_errata_after_index(struct radeon_device *rdev) { - if (!(rdev->pll_errata & CHIP_ERRATA_PLL_DUMMYREADS)) { - return; + if (rdev->pll_errata & CHIP_ERRATA_PLL_DUMMYREADS) { + (void)RREG32(RADEON_CLOCK_CNTL_DATA); + (void)RREG32(RADEON_CRTC_GEN_CNTL); } - (void)RREG32(RADEON_CLOCK_CNTL_DATA); - (void)RREG32(RADEON_CRTC_GEN_CNTL); } static void r100_pll_errata_after_data(struct radeon_device *rdev) @@ -1967,7 +2840,7 @@ static void r100_pll_errata_after_data(struct radeon_device *rdev) * or the chip could hang on a subsequent access */ if (rdev->pll_errata & CHIP_ERRATA_PLL_DELAY) { - udelay(5000); + mdelay(5); } /* This function is required to workaround a hardware bug in some (all?) @@ -1988,24 +2861,31 @@ 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); } -void r100_set_safe_registers(struct radeon_device *rdev) +static void r100_set_safe_registers(struct radeon_device *rdev) { if (ASIC_IS_RN50(rdev)) { rdev->config.r100.reg_safe_bm = rn50_reg_safe_bm; @@ -2048,21 +2928,24 @@ static int r100_debugfs_cp_ring_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 radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; uint32_t rdp, wdp; unsigned count, i, j; - radeon_ring_free_size(rdev); + radeon_ring_free_size(rdev, ring); rdp = RREG32(RADEON_CP_RB_RPTR); wdp = RREG32(RADEON_CP_RB_WPTR); - count = (rdp + rdev->cp.ring_size - wdp) & rdev->cp.ptr_mask; + count = (rdp + ring->ring_size - wdp) & ring->ptr_mask; seq_printf(m, "CP_STAT 0x%08x\n", RREG32(RADEON_CP_STAT)); seq_printf(m, "CP_RB_WPTR 0x%08x\n", wdp); seq_printf(m, "CP_RB_RPTR 0x%08x\n", rdp); - seq_printf(m, "%u free dwords in ring\n", rdev->cp.ring_free_dw); + 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) & rdev->cp.ptr_mask; - seq_printf(m, "r[%04d]=0x%08x\n", i, rdev->cp.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; } @@ -2196,18 +3079,16 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg, int surf_index = reg * 16; int flags = 0; - /* r100/r200 divide by 16 */ - if (rdev->family < CHIP_R300) - flags = pitch / 16; - else - flags = pitch / 8; - if (rdev->family <= CHIP_RS200) { if ((tiling_flags & (RADEON_TILING_MACRO|RADEON_TILING_MICRO)) == (RADEON_TILING_MACRO|RADEON_TILING_MICRO)) flags |= RADEON_SURF_TILE_COLOR_BOTH; if (tiling_flags & RADEON_TILING_MACRO) flags |= RADEON_SURF_TILE_COLOR_MACRO; + /* setting pitch to 0 disables tiling */ + if ((tiling_flags & (RADEON_TILING_MACRO|RADEON_TILING_MICRO)) + == 0) + pitch = 0; } else if (rdev->family <= CHIP_RV280) { if (tiling_flags & (RADEON_TILING_MACRO)) flags |= R200_SURF_TILE_COLOR_MACRO; @@ -2225,7 +3106,14 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg, if (tiling_flags & RADEON_TILING_SWAP_32BIT) flags |= RADEON_SURF_AP0_SWP_32BPP | RADEON_SURF_AP1_SWP_32BPP; - DRM_DEBUG("writing surface %d %d %x %x\n", reg, flags, offset, offset+obj_size-1); + /* r100/r200 divide by 16 */ + if (rdev->family < CHIP_R300) + flags |= pitch / 16; + else + flags |= pitch / 8; + + + DRM_DEBUG_KMS("writing surface %d %d %x %x\n", reg, flags, offset, offset+obj_size-1); WREG32(RADEON_SURFACE0_INFO + surf_index, flags); WREG32(RADEON_SURFACE0_LOWER_BOUND + surf_index, offset); WREG32(RADEON_SURFACE0_UPPER_BOUND + surf_index, offset + obj_size - 1); @@ -2245,53 +3133,53 @@ void r100_bandwidth_update(struct radeon_device *rdev) fixed20_12 peak_disp_bw, mem_bw, pix_clk, pix_clk2, temp_ff, crit_point_ff; uint32_t temp, data, mem_trcd, mem_trp, mem_tras; fixed20_12 memtcas_ff[8] = { - fixed_init(1), - fixed_init(2), - fixed_init(3), - fixed_init(0), - fixed_init_half(1), - fixed_init_half(2), - fixed_init(0), + dfixed_init(1), + dfixed_init(2), + dfixed_init(3), + dfixed_init(0), + dfixed_init_half(1), + dfixed_init_half(2), + dfixed_init(0), }; fixed20_12 memtcas_rs480_ff[8] = { - fixed_init(0), - fixed_init(1), - fixed_init(2), - fixed_init(3), - fixed_init(0), - fixed_init_half(1), - fixed_init_half(2), - fixed_init_half(3), + dfixed_init(0), + dfixed_init(1), + dfixed_init(2), + dfixed_init(3), + dfixed_init(0), + dfixed_init_half(1), + dfixed_init_half(2), + dfixed_init_half(3), }; fixed20_12 memtcas2_ff[8] = { - fixed_init(0), - fixed_init(1), - fixed_init(2), - fixed_init(3), - fixed_init(4), - fixed_init(5), - fixed_init(6), - fixed_init(7), + dfixed_init(0), + dfixed_init(1), + dfixed_init(2), + dfixed_init(3), + dfixed_init(4), + dfixed_init(5), + dfixed_init(6), + dfixed_init(7), }; fixed20_12 memtrbs[8] = { - fixed_init(1), - fixed_init_half(1), - fixed_init(2), - fixed_init_half(2), - fixed_init(3), - fixed_init_half(3), - fixed_init(4), - fixed_init_half(4) + dfixed_init(1), + dfixed_init_half(1), + dfixed_init(2), + dfixed_init_half(2), + dfixed_init(3), + dfixed_init_half(3), + dfixed_init(4), + dfixed_init_half(4) }; fixed20_12 memtrbs_r4xx[8] = { - fixed_init(4), - fixed_init(5), - fixed_init(6), - fixed_init(7), - fixed_init(8), - fixed_init(9), - fixed_init(10), - fixed_init(11) + dfixed_init(4), + dfixed_init(5), + dfixed_init(6), + dfixed_init(7), + dfixed_init(8), + dfixed_init(9), + dfixed_init(10), + dfixed_init(11) }; fixed20_12 min_mem_eff; fixed20_12 mc_latency_sclk, mc_latency_mclk, k1; @@ -2309,18 +3197,20 @@ void r100_bandwidth_update(struct radeon_device *rdev) uint32_t pixel_bytes1 = 0; uint32_t pixel_bytes2 = 0; + radeon_update_display_priority(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; } } - min_mem_eff.full = rfixed_const_8(0); + min_mem_eff.full = dfixed_const_8(0); /* get modes */ if ((rdev->disp_priority == 2) && ASIC_IS_R300(rdev)) { uint32_t mc_init_misc_lat_timer = RREG32(R300_MC_INIT_MISC_LAT_TIMER); @@ -2337,35 +3227,32 @@ void r100_bandwidth_update(struct radeon_device *rdev) /* * determine is there is enough bw for current mode */ - mclk_ff.full = rfixed_const(rdev->clock.default_mclk); - temp_ff.full = rfixed_const(100); - mclk_ff.full = rfixed_div(mclk_ff, temp_ff); - sclk_ff.full = rfixed_const(rdev->clock.default_sclk); - sclk_ff.full = rfixed_div(sclk_ff, temp_ff); + sclk_ff = rdev->pm.sclk; + mclk_ff = rdev->pm.mclk; temp = (rdev->mc.vram_width / 8) * (rdev->mc.vram_is_ddr ? 2 : 1); - temp_ff.full = rfixed_const(temp); - mem_bw.full = rfixed_mul(mclk_ff, temp_ff); + temp_ff.full = dfixed_const(temp); + mem_bw.full = dfixed_mul(mclk_ff, temp_ff); pix_clk.full = 0; pix_clk2.full = 0; peak_disp_bw.full = 0; if (mode1) { - temp_ff.full = rfixed_const(1000); - pix_clk.full = rfixed_const(mode1->clock); /* convert to fixed point */ - pix_clk.full = rfixed_div(pix_clk, temp_ff); - temp_ff.full = rfixed_const(pixel_bytes1); - peak_disp_bw.full += rfixed_mul(pix_clk, temp_ff); + temp_ff.full = dfixed_const(1000); + pix_clk.full = dfixed_const(mode1->clock); /* convert to fixed point */ + pix_clk.full = dfixed_div(pix_clk, temp_ff); + temp_ff.full = dfixed_const(pixel_bytes1); + peak_disp_bw.full += dfixed_mul(pix_clk, temp_ff); } if (mode2) { - temp_ff.full = rfixed_const(1000); - pix_clk2.full = rfixed_const(mode2->clock); /* convert to fixed point */ - pix_clk2.full = rfixed_div(pix_clk2, temp_ff); - temp_ff.full = rfixed_const(pixel_bytes2); - peak_disp_bw.full += rfixed_mul(pix_clk2, temp_ff); + temp_ff.full = dfixed_const(1000); + pix_clk2.full = dfixed_const(mode2->clock); /* convert to fixed point */ + pix_clk2.full = dfixed_div(pix_clk2, temp_ff); + temp_ff.full = dfixed_const(pixel_bytes2); + peak_disp_bw.full += dfixed_mul(pix_clk2, temp_ff); } - mem_bw.full = rfixed_mul(mem_bw, min_mem_eff); + mem_bw.full = dfixed_mul(mem_bw, min_mem_eff); if (peak_disp_bw.full >= mem_bw.full) { DRM_ERROR("You may not have enough display bandwidth for current mode\n" "If you have flickering problem, try to lower resolution, refresh rate, or color depth\n"); @@ -2407,9 +3294,9 @@ void r100_bandwidth_update(struct radeon_device *rdev) mem_tras = ((temp >> 12) & 0xf) + 4; } /* convert to FF */ - trcd_ff.full = rfixed_const(mem_trcd); - trp_ff.full = rfixed_const(mem_trp); - tras_ff.full = rfixed_const(mem_tras); + trcd_ff.full = dfixed_const(mem_trcd); + trp_ff.full = dfixed_const(mem_trp); + tras_ff.full = dfixed_const(mem_tras); /* Get values from the MEM_SDRAM_MODE_REG register...converting its */ temp = RREG32(RADEON_MEM_SDRAM_MODE_REG); @@ -2427,7 +3314,7 @@ void r100_bandwidth_update(struct radeon_device *rdev) /* extra cas latency stored in bits 23-25 0-4 clocks */ data = (temp >> 23) & 0x7; if (data < 5) - tcas_ff.full += rfixed_const(data); + tcas_ff.full += dfixed_const(data); } if (ASIC_IS_R300(rdev) && !(rdev->flags & RADEON_IS_IGP)) { @@ -2464,72 +3351,72 @@ void r100_bandwidth_update(struct radeon_device *rdev) if (rdev->flags & RADEON_IS_AGP) { fixed20_12 agpmode_ff; - agpmode_ff.full = rfixed_const(radeon_agpmode); - temp_ff.full = rfixed_const_666(16); - sclk_eff_ff.full -= rfixed_mul(agpmode_ff, temp_ff); + agpmode_ff.full = dfixed_const(radeon_agpmode); + temp_ff.full = dfixed_const_666(16); + sclk_eff_ff.full -= dfixed_mul(agpmode_ff, temp_ff); } /* TODO PCIE lanes may affect this - agpmode == 16?? */ if (ASIC_IS_R300(rdev)) { - sclk_delay_ff.full = rfixed_const(250); + sclk_delay_ff.full = dfixed_const(250); } else { if ((rdev->family == CHIP_RV100) || rdev->flags & RADEON_IS_IGP) { if (rdev->mc.vram_is_ddr) - sclk_delay_ff.full = rfixed_const(41); + sclk_delay_ff.full = dfixed_const(41); else - sclk_delay_ff.full = rfixed_const(33); + sclk_delay_ff.full = dfixed_const(33); } else { if (rdev->mc.vram_width == 128) - sclk_delay_ff.full = rfixed_const(57); + sclk_delay_ff.full = dfixed_const(57); else - sclk_delay_ff.full = rfixed_const(41); + sclk_delay_ff.full = dfixed_const(41); } } - mc_latency_sclk.full = rfixed_div(sclk_delay_ff, sclk_eff_ff); + mc_latency_sclk.full = dfixed_div(sclk_delay_ff, sclk_eff_ff); if (rdev->mc.vram_is_ddr) { if (rdev->mc.vram_width == 32) { - k1.full = rfixed_const(40); + k1.full = dfixed_const(40); c = 3; } else { - k1.full = rfixed_const(20); + k1.full = dfixed_const(20); c = 1; } } else { - k1.full = rfixed_const(40); + k1.full = dfixed_const(40); c = 3; } - temp_ff.full = rfixed_const(2); - mc_latency_mclk.full = rfixed_mul(trcd_ff, temp_ff); - temp_ff.full = rfixed_const(c); - mc_latency_mclk.full += rfixed_mul(tcas_ff, temp_ff); - temp_ff.full = rfixed_const(4); - mc_latency_mclk.full += rfixed_mul(tras_ff, temp_ff); - mc_latency_mclk.full += rfixed_mul(trp_ff, temp_ff); + temp_ff.full = dfixed_const(2); + mc_latency_mclk.full = dfixed_mul(trcd_ff, temp_ff); + temp_ff.full = dfixed_const(c); + mc_latency_mclk.full += dfixed_mul(tcas_ff, temp_ff); + temp_ff.full = dfixed_const(4); + mc_latency_mclk.full += dfixed_mul(tras_ff, temp_ff); + mc_latency_mclk.full += dfixed_mul(trp_ff, temp_ff); mc_latency_mclk.full += k1.full; - mc_latency_mclk.full = rfixed_div(mc_latency_mclk, mclk_ff); - mc_latency_mclk.full += rfixed_div(temp_ff, sclk_eff_ff); + mc_latency_mclk.full = dfixed_div(mc_latency_mclk, mclk_ff); + mc_latency_mclk.full += dfixed_div(temp_ff, sclk_eff_ff); /* HW cursor time assuming worst case of full size colour cursor. */ - temp_ff.full = rfixed_const((2 * (cur_size - (rdev->mc.vram_is_ddr + 1)))); + temp_ff.full = dfixed_const((2 * (cur_size - (rdev->mc.vram_is_ddr + 1)))); temp_ff.full += trcd_ff.full; if (temp_ff.full < tras_ff.full) temp_ff.full = tras_ff.full; - cur_latency_mclk.full = rfixed_div(temp_ff, mclk_ff); + cur_latency_mclk.full = dfixed_div(temp_ff, mclk_ff); - temp_ff.full = rfixed_const(cur_size); - cur_latency_sclk.full = rfixed_div(temp_ff, sclk_eff_ff); + temp_ff.full = dfixed_const(cur_size); + cur_latency_sclk.full = dfixed_div(temp_ff, sclk_eff_ff); /* Find the total latency for the display data. */ - disp_latency_overhead.full = rfixed_const(8); - disp_latency_overhead.full = rfixed_div(disp_latency_overhead, sclk_ff); + disp_latency_overhead.full = dfixed_const(8); + disp_latency_overhead.full = dfixed_div(disp_latency_overhead, sclk_ff); mc_latency_mclk.full += disp_latency_overhead.full + cur_latency_mclk.full; mc_latency_sclk.full += disp_latency_overhead.full + cur_latency_sclk.full; @@ -2557,16 +3444,16 @@ void r100_bandwidth_update(struct radeon_device *rdev) /* Find the drain rate of the display buffer. */ - temp_ff.full = rfixed_const((16/pixel_bytes1)); - disp_drain_rate.full = rfixed_div(pix_clk, temp_ff); + temp_ff.full = dfixed_const((16/pixel_bytes1)); + disp_drain_rate.full = dfixed_div(pix_clk, temp_ff); /* Find the critical point of the display buffer. */ - crit_point_ff.full = rfixed_mul(disp_drain_rate, disp_latency); - crit_point_ff.full += rfixed_const_half(0); + crit_point_ff.full = dfixed_mul(disp_drain_rate, disp_latency); + crit_point_ff.full += dfixed_const_half(0); - critical_point = rfixed_trunc(crit_point_ff); + critical_point = dfixed_trunc(crit_point_ff); if (rdev->disp_priority == 2) { critical_point = 0; @@ -2622,7 +3509,7 @@ void r100_bandwidth_update(struct radeon_device *rdev) } #endif - DRM_DEBUG("GRPH_BUFFER_CNTL from to %x\n", + DRM_DEBUG_KMS("GRPH_BUFFER_CNTL from to %x\n", /* (unsigned int)info->SavedReg->grph_buffer_cntl, */ (unsigned int)RREG32(RADEON_GRPH_BUFFER_CNTL)); } @@ -2637,8 +3524,8 @@ void r100_bandwidth_update(struct radeon_device *rdev) /* Find the drain rate of the display buffer. */ - temp_ff.full = rfixed_const((16/pixel_bytes2)); - disp_drain_rate2.full = rfixed_div(pix_clk2, temp_ff); + temp_ff.full = dfixed_const((16/pixel_bytes2)); + disp_drain_rate2.full = dfixed_div(pix_clk2, temp_ff); grph2_cntl = RREG32(RADEON_GRPH2_BUFFER_CNTL); grph2_cntl &= ~(RADEON_GRPH_STOP_REQ_MASK); @@ -2659,8 +3546,8 @@ void r100_bandwidth_update(struct radeon_device *rdev) critical_point2 = 0; else { temp = (rdev->mc.vram_width * rdev->mc.vram_is_ddr + 1)/128; - temp_ff.full = rfixed_const(temp); - temp_ff.full = rfixed_mul(mclk_ff, temp_ff); + temp_ff.full = dfixed_const(temp); + temp_ff.full = dfixed_mul(mclk_ff, temp_ff); if (sclk_ff.full < temp_ff.full) temp_ff.full = sclk_ff.full; @@ -2668,15 +3555,15 @@ void r100_bandwidth_update(struct radeon_device *rdev) if (mode1) { temp_ff.full = read_return_rate.full - disp_drain_rate.full; - time_disp1_drop_priority.full = rfixed_div(crit_point_ff, temp_ff); + time_disp1_drop_priority.full = dfixed_div(crit_point_ff, temp_ff); } else { time_disp1_drop_priority.full = 0; } crit_point_ff.full = disp_latency.full + time_disp1_drop_priority.full + disp_latency.full; - crit_point_ff.full = rfixed_mul(crit_point_ff, disp_drain_rate2); - crit_point_ff.full += rfixed_const_half(0); + crit_point_ff.full = dfixed_mul(crit_point_ff, disp_drain_rate2); + crit_point_ff.full += dfixed_const_half(0); - critical_point2 = rfixed_trunc(crit_point_ff); + critical_point2 = dfixed_trunc(crit_point_ff); if (rdev->disp_priority == 2) { critical_point2 = 0; @@ -2718,333 +3605,12 @@ void r100_bandwidth_update(struct radeon_device *rdev) WREG32(RS400_DISP1_REQ_CNTL1, 0x28FBC3AC); } - DRM_DEBUG("GRPH2_BUFFER_CNTL from to %x\n", + DRM_DEBUG_KMS("GRPH2_BUFFER_CNTL from to %x\n", (unsigned int)RREG32(RADEON_GRPH2_BUFFER_CNTL)); } } -static inline void r100_cs_track_texture_print(struct r100_cs_track_texture *t) -{ - DRM_ERROR("pitch %d\n", t->pitch); - DRM_ERROR("use_pitch %d\n", t->use_pitch); - DRM_ERROR("width %d\n", t->width); - DRM_ERROR("width_11 %d\n", t->width_11); - DRM_ERROR("height %d\n", t->height); - DRM_ERROR("height_11 %d\n", t->height_11); - DRM_ERROR("num levels %d\n", t->num_levels); - DRM_ERROR("depth %d\n", t->txdepth); - DRM_ERROR("bpp %d\n", t->cpp); - DRM_ERROR("coordinate type %d\n", t->tex_coord_type); - DRM_ERROR("width round to power of 2 %d\n", t->roundup_w); - DRM_ERROR("height round to power of 2 %d\n", t->roundup_h); - DRM_ERROR("compress format %d\n", t->compress_format); -} - -static int r100_cs_track_cube(struct radeon_device *rdev, - struct r100_cs_track *track, unsigned idx) -{ - unsigned face, w, h; - struct radeon_bo *cube_robj; - unsigned long size; - - for (face = 0; face < 5; face++) { - cube_robj = track->textures[idx].cube_info[face].robj; - w = track->textures[idx].cube_info[face].width; - h = track->textures[idx].cube_info[face].height; - - size = w * h; - size *= track->textures[idx].cpp; - - size += track->textures[idx].cube_info[face].offset; - - if (size > radeon_bo_size(cube_robj)) { - DRM_ERROR("Cube texture offset greater than object size %lu %lu\n", - size, radeon_bo_size(cube_robj)); - r100_cs_track_texture_print(&track->textures[idx]); - return -1; - } - } - return 0; -} - -static int r100_track_compress_size(int compress_format, int w, int h) -{ - int block_width, block_height, block_bytes; - int wblocks, hblocks; - int min_wblocks; - int sz; - - block_width = 4; - block_height = 4; - - switch (compress_format) { - case R100_TRACK_COMP_DXT1: - block_bytes = 8; - min_wblocks = 4; - break; - default: - case R100_TRACK_COMP_DXT35: - block_bytes = 16; - min_wblocks = 2; - break; - } - - hblocks = (h + block_height - 1) / block_height; - wblocks = (w + block_width - 1) / block_width; - if (wblocks < min_wblocks) - wblocks = min_wblocks; - sz = wblocks * hblocks * block_bytes; - return sz; -} - -static int r100_cs_track_texture_check(struct radeon_device *rdev, - struct r100_cs_track *track) -{ - struct radeon_bo *robj; - unsigned long size; - unsigned u, i, w, h; - int ret; - - for (u = 0; u < track->num_texture; u++) { - if (!track->textures[u].enabled) - continue; - robj = track->textures[u].robj; - if (robj == NULL) { - DRM_ERROR("No texture bound to unit %u\n", u); - return -EINVAL; - } - size = 0; - for (i = 0; i <= track->textures[u].num_levels; i++) { - if (track->textures[u].use_pitch) { - if (rdev->family < CHIP_R300) - w = (track->textures[u].pitch / track->textures[u].cpp) / (1 << i); - else - w = track->textures[u].pitch / (1 << i); - } else { - w = track->textures[u].width; - if (rdev->family >= CHIP_RV515) - w |= track->textures[u].width_11; - w = w / (1 << i); - if (track->textures[u].roundup_w) - w = roundup_pow_of_two(w); - } - h = track->textures[u].height; - if (rdev->family >= CHIP_RV515) - h |= track->textures[u].height_11; - h = h / (1 << i); - if (track->textures[u].roundup_h) - h = roundup_pow_of_two(h); - if (track->textures[u].compress_format) { - - size += r100_track_compress_size(track->textures[u].compress_format, w, h); - /* compressed textures are block based */ - } else - size += w * h; - } - size *= track->textures[u].cpp; - - switch (track->textures[u].tex_coord_type) { - case 0: - break; - case 1: - size *= (1 << track->textures[u].txdepth); - break; - case 2: - if (track->separate_cube) { - ret = r100_cs_track_cube(rdev, track, u); - if (ret) - return ret; - } else - size *= 6; - break; - default: - DRM_ERROR("Invalid texture coordinate type %u for unit " - "%u\n", track->textures[u].tex_coord_type, u); - return -EINVAL; - } - if (size > radeon_bo_size(robj)) { - DRM_ERROR("Texture of unit %u needs %lu bytes but is " - "%lu\n", u, size, radeon_bo_size(robj)); - r100_cs_track_texture_print(&track->textures[u]); - return -EINVAL; - } - } - return 0; -} - -int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) -{ - unsigned i; - unsigned long size; - unsigned prim_walk; - unsigned nverts; - - for (i = 0; i < track->num_cb; i++) { - if (track->cb[i].robj == NULL) { - if (!(track->fastfill || track->color_channel_mask || - track->blend_read_enable)) { - continue; - } - DRM_ERROR("[drm] No buffer for color buffer %d !\n", i); - return -EINVAL; - } - size = track->cb[i].pitch * track->cb[i].cpp * track->maxy; - size += track->cb[i].offset; - if (size > radeon_bo_size(track->cb[i].robj)) { - DRM_ERROR("[drm] Buffer too small for color buffer %d " - "(need %lu have %lu) !\n", i, size, - radeon_bo_size(track->cb[i].robj)); - DRM_ERROR("[drm] color buffer %d (%u %u %u %u)\n", - i, track->cb[i].pitch, track->cb[i].cpp, - track->cb[i].offset, track->maxy); - return -EINVAL; - } - } - if (track->z_enabled) { - if (track->zb.robj == NULL) { - DRM_ERROR("[drm] No buffer for z buffer !\n"); - return -EINVAL; - } - size = track->zb.pitch * track->zb.cpp * track->maxy; - size += track->zb.offset; - if (size > radeon_bo_size(track->zb.robj)) { - DRM_ERROR("[drm] Buffer too small for z buffer " - "(need %lu have %lu) !\n", size, - radeon_bo_size(track->zb.robj)); - DRM_ERROR("[drm] zbuffer (%u %u %u %u)\n", - track->zb.pitch, track->zb.cpp, - track->zb.offset, track->maxy); - return -EINVAL; - } - } - prim_walk = (track->vap_vf_cntl >> 4) & 0x3; - nverts = (track->vap_vf_cntl >> 16) & 0xFFFF; - switch (prim_walk) { - case 1: - for (i = 0; i < track->num_arrays; i++) { - size = track->arrays[i].esize * track->max_indx * 4; - if (track->arrays[i].robj == NULL) { - DRM_ERROR("(PW %u) Vertex array %u no buffer " - "bound\n", prim_walk, i); - return -EINVAL; - } - if (size > radeon_bo_size(track->arrays[i].robj)) { - dev_err(rdev->dev, "(PW %u) Vertex array %u " - "need %lu dwords have %lu dwords\n", - prim_walk, i, size >> 2, - radeon_bo_size(track->arrays[i].robj) - >> 2); - DRM_ERROR("Max indices %u\n", track->max_indx); - return -EINVAL; - } - } - break; - case 2: - for (i = 0; i < track->num_arrays; i++) { - size = track->arrays[i].esize * (nverts - 1) * 4; - if (track->arrays[i].robj == NULL) { - DRM_ERROR("(PW %u) Vertex array %u no buffer " - "bound\n", prim_walk, i); - return -EINVAL; - } - if (size > radeon_bo_size(track->arrays[i].robj)) { - dev_err(rdev->dev, "(PW %u) Vertex array %u " - "need %lu dwords have %lu dwords\n", - prim_walk, i, size >> 2, - radeon_bo_size(track->arrays[i].robj) - >> 2); - return -EINVAL; - } - } - break; - case 3: - size = track->vtx_size * nverts; - if (size != track->immd_dwords) { - DRM_ERROR("IMMD draw %u dwors but needs %lu dwords\n", - track->immd_dwords, size); - DRM_ERROR("VAP_VF_CNTL.NUM_VERTICES %u, VTX_SIZE %u\n", - nverts, track->vtx_size); - return -EINVAL; - } - break; - default: - DRM_ERROR("[drm] Invalid primitive walk %d for VAP_VF_CNTL\n", - prim_walk); - return -EINVAL; - } - return r100_cs_track_texture_check(rdev, track); -} - -void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track) -{ - unsigned i, face; - - if (rdev->family < CHIP_R300) { - track->num_cb = 1; - if (rdev->family <= CHIP_RS200) - track->num_texture = 3; - else - track->num_texture = 6; - track->maxy = 2048; - track->separate_cube = 1; - } else { - track->num_cb = 4; - track->num_texture = 16; - track->maxy = 4096; - track->separate_cube = 0; - } - - for (i = 0; i < track->num_cb; i++) { - track->cb[i].robj = NULL; - track->cb[i].pitch = 8192; - track->cb[i].cpp = 16; - track->cb[i].offset = 0; - } - track->z_enabled = true; - track->zb.robj = NULL; - track->zb.pitch = 8192; - track->zb.cpp = 4; - track->zb.offset = 0; - track->vtx_size = 0x7F; - track->immd_dwords = 0xFFFFFFFFUL; - track->num_arrays = 11; - track->max_indx = 0x00FFFFFFUL; - for (i = 0; i < track->num_arrays; i++) { - track->arrays[i].robj = NULL; - track->arrays[i].esize = 0x7F; - } - for (i = 0; i < track->num_texture; i++) { - track->textures[i].compress_format = R100_TRACK_COMP_NONE; - track->textures[i].pitch = 16536; - track->textures[i].width = 16536; - track->textures[i].height = 16536; - track->textures[i].width_11 = 1 << 11; - track->textures[i].height_11 = 1 << 11; - track->textures[i].num_levels = 12; - if (rdev->family <= CHIP_RS200) { - track->textures[i].tex_coord_type = 0; - track->textures[i].txdepth = 0; - } else { - track->textures[i].txdepth = 16; - track->textures[i].tex_coord_type = 1; - } - track->textures[i].cpp = 64; - track->textures[i].robj = NULL; - /* CS IB emission code makes sure texture unit are disabled */ - track->textures[i].enabled = false; - track->textures[i].roundup_w = true; - track->textures[i].roundup_h = true; - if (track->separate_cube) - for (face = 0; face < 5; face++) { - track->textures[i].cube_info[face].robj = NULL; - track->textures[i].cube_info[face].width = 16536; - track->textures[i].cube_info[face].height = 16536; - track->textures[i].cube_info[face].offset = 0; - } - } -} - -int r100_ring_test(struct radeon_device *rdev) +int r100_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) { uint32_t scratch; uint32_t tmp = 0; @@ -3057,15 +3623,15 @@ int r100_ring_test(struct radeon_device *rdev) return r; } WREG32(scratch, 0xCAFEDEAD); - r = radeon_ring_lock(rdev, 2); + r = radeon_ring_lock(rdev, ring, 2); if (r) { DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); radeon_scratch_free(rdev, scratch); return r; } - radeon_ring_write(rdev, PACKET0(scratch, 0)); - radeon_ring_write(rdev, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, PACKET0(scratch, 0)); + radeon_ring_write(ring, 0xDEADBEEF); + radeon_ring_unlock_commit(rdev, ring); for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) { @@ -3076,7 +3642,7 @@ int r100_ring_test(struct radeon_device *rdev) if (i < rdev->usec_timeout) { DRM_INFO("ring test succeeded in %d usecs\n", i); } else { - DRM_ERROR("radeon: ring test failed (sracth(0x%04X)=0x%08X)\n", + DRM_ERROR("radeon: ring test failed (scratch(0x%04X)=0x%08X)\n", scratch, tmp); r = -EINVAL; } @@ -3086,14 +3652,22 @@ int r100_ring_test(struct radeon_device *rdev) void r100_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) { - radeon_ring_write(rdev, PACKET0(RADEON_CP_IB_BASE, 1)); - radeon_ring_write(rdev, ib->gpu_addr); - radeon_ring_write(rdev, ib->length_dw); + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + + if (ring->rptr_save_reg) { + u32 next_rptr = ring->wptr + 2 + 3; + radeon_ring_write(ring, PACKET0(ring->rptr_save_reg, 0)); + radeon_ring_write(ring, next_rptr); + } + + radeon_ring_write(ring, PACKET0(RADEON_CP_IB_BASE, 1)); + radeon_ring_write(ring, ib->gpu_addr); + radeon_ring_write(ring, ib->length_dw); } -int r100_ib_test(struct radeon_device *rdev) +int r100_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; @@ -3105,28 +3679,29 @@ int r100_ib_test(struct radeon_device *rdev) return r; } WREG32(scratch, 0xCAFEDEAD); - r = radeon_ib_get(rdev, &ib); + r = radeon_ib_get(rdev, RADEON_RING_TYPE_GFX_INDEX, &ib, NULL, 256); if (r) { - return r; - } - ib->ptr[0] = PACKET0(scratch, 0); - ib->ptr[1] = 0xDEADBEEF; - ib->ptr[2] = PACKET2(0); - 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->length_dw = 8; - r = radeon_ib_schedule(rdev, ib); + DRM_ERROR("radeon: failed to get ib (%d).\n", r); + goto free_scratch; + } + ib.ptr[0] = PACKET0(scratch, 0); + ib.ptr[1] = 0xDEADBEEF; + ib.ptr[2] = PACKET2(0); + 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.length_dw = 8; + r = radeon_ib_schedule(rdev, &ib, NULL); if (r) { - radeon_scratch_free(rdev, scratch); - radeon_ib_free(rdev, &ib); - return r; + DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); + goto free_ib; } - r = radeon_fence_wait(ib->fence, false); + r = radeon_fence_wait(ib.fence, false); if (r) { - return r; + DRM_ERROR("radeon: fence wait failed (%d).\n", r); + goto free_ib; } for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); @@ -3138,45 +3713,23 @@ int r100_ib_test(struct radeon_device *rdev) if (i < rdev->usec_timeout) { DRM_INFO("ib test succeeded in %u usecs\n", i); } else { - DRM_ERROR("radeon: ib test failed (sracth(0x%04X)=0x%08X)\n", + 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; } -void r100_ib_fini(struct radeon_device *rdev) -{ - radeon_ib_pool_fini(rdev); -} - -int r100_ib_init(struct radeon_device *rdev) -{ - int r; - - r = radeon_ib_pool_init(rdev); - if (r) { - dev_err(rdev->dev, "failled initializing IB pool (%d).\n", r); - r100_ib_fini(rdev); - return r; - } - r = r100_ib_test(rdev); - if (r) { - dev_err(rdev->dev, "failled testing IB (%d).\n", r); - r100_ib_fini(rdev); - return r; - } - return 0; -} - void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save) { /* Shutdown CP we shouldn't need to do that but better be safe than * sorry */ - rdev->cp.ready = false; + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; WREG32(R_000740_CP_CSQ_CNTL, 0); /* Save few CRTC registers */ @@ -3216,10 +3769,9 @@ void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save) void r100_mc_resume(struct radeon_device *rdev, struct r100_mc_save *save) { /* Update base address for crtc */ - WREG32(R_00023C_DISPLAY_BASE_ADDR, rdev->mc.vram_location); + WREG32(R_00023C_DISPLAY_BASE_ADDR, rdev->mc.vram_start); if (!(rdev->flags & RADEON_SINGLE_CRTC)) { - WREG32(R_00033C_CRTC2_DISPLAY_BASE_ADDR, - rdev->mc.vram_location); + WREG32(R_00033C_CRTC2_DISPLAY_BASE_ADDR, rdev->mc.vram_start); } /* Restore CRTC registers */ WREG8(R_0003C2_GENMO_WT, save->GENMO_WT); @@ -3277,7 +3829,7 @@ static void r100_mc_program(struct radeon_device *rdev) r100_mc_resume(rdev, &save); } -void r100_clock_startup(struct radeon_device *rdev) +static void r100_clock_startup(struct radeon_device *rdev) { u32 tmp; @@ -3301,8 +3853,6 @@ static int r100_startup(struct radeon_device *rdev) r100_mc_program(rdev); /* Resume clock */ r100_clock_startup(rdev); - /* Initialize GPU configuration (# pipes, ...) */ - r100_gpu_init(rdev); /* Initialize GART (initialize after TTM so we can allocate * memory through TTM but finalize after TTM) */ r100_enable_bm(rdev); @@ -3311,34 +3861,54 @@ static int r100_startup(struct radeon_device *rdev) if (r) return r; } + + /* allocate wb buffer */ + r = radeon_wb_init(rdev); + if (r) + return r; + + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ + if (!rdev->irq.installed) { + r = radeon_irq_kms_init(rdev); + if (r) + return r; + } + r100_irq_set(rdev); + rdev->config.r100.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL); /* 1M ring buffer */ r = r100_cp_init(rdev, 1024 * 1024); if (r) { - dev_err(rdev->dev, "failled initializing CP (%d).\n", r); + dev_err(rdev->dev, "failed initializing CP (%d).\n", r); return r; } - r = r100_wb_init(rdev); - if (r) - dev_err(rdev->dev, "failled initializing WB (%d).\n", r); - r = r100_ib_init(rdev); + + r = radeon_ib_pool_init(rdev); if (r) { - dev_err(rdev->dev, "failled initializing IB (%d).\n", r); + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); return r; } + return 0; } int r100_resume(struct radeon_device *rdev) { + int r; + /* Make sur GART are not working */ if (rdev->flags & RADEON_IS_PCI) r100_pci_gart_disable(rdev); /* Resume clock before doing reset */ r100_clock_startup(rdev); /* Reset gpu before posting otherwise ATOM will enter infinite loop */ - if (radeon_gpu_reset(rdev)) { + if (radeon_asic_reset(rdev)) { dev_warn(rdev->dev, "GPU reset failed ! (0xE40=0x%08X, 0x7C0=0x%08X)\n", RREG32(R_000E40_RBBM_STATUS), RREG32(R_0007C0_CP_STAT)); @@ -3349,13 +3919,20 @@ int r100_resume(struct radeon_device *rdev) r100_clock_startup(rdev); /* Initialize surface registers */ radeon_surface_init(rdev); - return r100_startup(rdev); + + rdev->accel_working = true; + r = r100_startup(rdev); + if (r) { + rdev->accel_working = false; + } + return r; } int r100_suspend(struct radeon_device *rdev) { + radeon_pm_suspend(rdev); r100_cp_disable(rdev); - r100_wb_disable(rdev); + radeon_wb_disable(rdev); r100_irq_disable(rdev); if (rdev->flags & RADEON_IS_PCI) r100_pci_gart_disable(rdev); @@ -3364,13 +3941,14 @@ int r100_suspend(struct radeon_device *rdev) void r100_fini(struct radeon_device *rdev) { - r100_suspend(rdev); + radeon_pm_fini(rdev); r100_cp_fini(rdev); - r100_wb_fini(rdev); - r100_ib_fini(rdev); + radeon_wb_fini(rdev); + radeon_ib_pool_fini(rdev); radeon_gem_fini(rdev); if (rdev->flags & RADEON_IS_PCI) r100_pci_gart_fini(rdev); + radeon_agp_fini(rdev); radeon_irq_kms_fini(rdev); radeon_fence_driver_fini(rdev); radeon_bo_fini(rdev); @@ -3379,32 +3957,29 @@ void r100_fini(struct radeon_device *rdev) rdev->bios = NULL; } -int r100_mc_init(struct radeon_device *rdev) +/* + * Due to how kexec works, it can leave the hw fully initialised when it + * boots the new kernel. However doing our init sequence with the CP and + * WB stuff setup causes GPU hangs on the RN50 at least. So at startup + * do some quick sanity checks and restore sane values to avoid this + * problem. + */ +void r100_restore_sanity(struct radeon_device *rdev) { - int r; u32 tmp; - /* Setup GPU memory space */ - rdev->mc.vram_location = 0xFFFFFFFFUL; - rdev->mc.gtt_location = 0xFFFFFFFFUL; - if (rdev->flags & RADEON_IS_IGP) { - tmp = G_00015C_MC_FB_START(RREG32(R_00015C_NB_TOM)); - rdev->mc.vram_location = tmp << 16; + tmp = RREG32(RADEON_CP_CSQ_CNTL); + if (tmp) { + WREG32(RADEON_CP_CSQ_CNTL, 0); } - if (rdev->flags & RADEON_IS_AGP) { - r = radeon_agp_init(rdev); - if (r) { - printk(KERN_WARNING "[drm] Disabling AGP\n"); - rdev->flags &= ~RADEON_IS_AGP; - rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024; - } else { - rdev->mc.gtt_location = rdev->mc.agp_base; - } + tmp = RREG32(RADEON_CP_RB_CNTL); + if (tmp) { + WREG32(RADEON_CP_RB_CNTL, 0); + } + tmp = RREG32(RADEON_SCRATCH_UMSK); + if (tmp) { + WREG32(RADEON_SCRATCH_UMSK, 0); } - r = radeon_mc_setup(rdev); - if (r) - return r; - return 0; } int r100_init(struct radeon_device *rdev) @@ -3419,6 +3994,8 @@ int r100_init(struct radeon_device *rdev) radeon_scratch_init(rdev); /* Initialize surface registers */ radeon_surface_init(rdev); + /* sanity check some register to avoid hangs like after kexec */ + r100_restore_sanity(rdev); /* TODO: disable VGA need to use VGA request */ /* BIOS*/ if (!radeon_get_bios(rdev)) { @@ -3434,7 +4011,7 @@ int r100_init(struct radeon_device *rdev) return r; } /* Reset gpu before posting otherwise ATOM will enter infinite loop */ - if (radeon_gpu_reset(rdev)) { + if (radeon_asic_reset(rdev)) { dev_warn(rdev->dev, "GPU reset failed ! (0xE40=0x%08X, 0x7C0=0x%08X)\n", RREG32(R_000E40_RBBM_STATUS), @@ -3447,21 +4024,19 @@ int r100_init(struct radeon_device *rdev) r100_errata(rdev); /* Initialize clocks */ radeon_get_clock_info(rdev->ddev); - /* Initialize power management */ - radeon_pm_init(rdev); - /* Get vram informations */ - r100_vram_info(rdev); - /* Initialize memory controller (also test AGP) */ - r = r100_mc_init(rdev); - if (r) - return r; + /* initialize AGP */ + if (rdev->flags & RADEON_IS_AGP) { + r = radeon_agp_init(rdev); + if (r) { + radeon_agp_disable(rdev); + } + } + /* initialize VRAM */ + r100_mc_init(rdev); /* Fence driver */ r = radeon_fence_driver_init(rdev); if (r) return r; - r = radeon_irq_kms_init(rdev); - if (r) - return r; /* Memory manager */ r = radeon_bo_init(rdev); if (r) @@ -3472,19 +4047,75 @@ int r100_init(struct radeon_device *rdev) return r; } r100_set_safe_registers(rdev); + + /* Initialize power management */ + radeon_pm_init(rdev); + rdev->accel_working = true; r = r100_startup(rdev); if (r) { /* Somethings want wront with the accel init stop accel */ dev_err(rdev->dev, "Disabling GPU acceleration\n"); - r100_suspend(rdev); r100_cp_fini(rdev); - r100_wb_fini(rdev); - r100_ib_fini(rdev); + radeon_wb_fini(rdev); + radeon_ib_pool_fini(rdev); + radeon_irq_kms_fini(rdev); if (rdev->flags & RADEON_IS_PCI) r100_pci_gart_fini(rdev); - radeon_irq_kms_fini(rdev); rdev->accel_working = false; } return 0; } + +uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg, + bool always_indirect) +{ + if (reg < rdev->rmmio_size && !always_indirect) + return readl(((void __iomem *)rdev->rmmio) + reg); + else { + unsigned long flags; + uint32_t ret; + + spin_lock_irqsave(&rdev->mmio_idx_lock, flags); + writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX); + ret = readl(((void __iomem *)rdev->rmmio) + RADEON_MM_DATA); + spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags); + + return ret; + } +} + +void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v, + bool always_indirect) +{ + if (reg < rdev->rmmio_size && !always_indirect) + writel(v, ((void __iomem *)rdev->rmmio) + reg); + else { + unsigned long flags; + + spin_lock_irqsave(&rdev->mmio_idx_lock, flags); + writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX); + writel(v, ((void __iomem *)rdev->rmmio) + RADEON_MM_DATA); + spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags); + } +} + +u32 r100_io_rreg(struct radeon_device *rdev, u32 reg) +{ + if (reg < rdev->rio_mem_size) + return ioread32(rdev->rio_mem + reg); + else { + iowrite32(reg, rdev->rio_mem + RADEON_MM_INDEX); + return ioread32(rdev->rio_mem + RADEON_MM_DATA); + } +} + +void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v) +{ + if (reg < rdev->rio_mem_size) + iowrite32(v, rdev->rio_mem + reg); + else { + iowrite32(reg, rdev->rio_mem + RADEON_MM_INDEX); + iowrite32(v, rdev->rio_mem + RADEON_MM_DATA); + } +} |