diff options
Diffstat (limited to 'drivers/gpu/drm/radeon')
34 files changed, 263 insertions, 156 deletions
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile index 0013ad0db9e..f77b7135ee4 100644 --- a/drivers/gpu/drm/radeon/Makefile +++ b/drivers/gpu/drm/radeon/Makefile @@ -76,7 +76,7 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \ evergreen.o evergreen_cs.o evergreen_blit_shaders.o \ evergreen_hdmi.o radeon_trace_points.o ni.o cayman_blit_shaders.o \ atombios_encoders.o radeon_semaphore.o radeon_sa.o atombios_i2c.o si.o \ - si_blit_shaders.o radeon_prime.o radeon_uvd.o cik.o cik_blit_shaders.o \ + si_blit_shaders.o radeon_prime.o cik.o cik_blit_shaders.o \ r600_dpm.o rs780_dpm.o rv6xx_dpm.o rv770_dpm.o rv730_dpm.o rv740_dpm.o \ rv770_smc.o cypress_dpm.o btc_dpm.o sumo_dpm.o sumo_smc.o trinity_dpm.o \ trinity_smc.o ni_dpm.o si_smc.o si_dpm.o kv_smc.o kv_dpm.o ci_smc.o \ diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c index 022561e2870..d416bb2ff48 100644 --- a/drivers/gpu/drm/radeon/ci_dpm.c +++ b/drivers/gpu/drm/radeon/ci_dpm.c @@ -869,6 +869,9 @@ static int ci_set_thermal_temperature_range(struct radeon_device *rdev, WREG32_SMC(CG_THERMAL_CTRL, tmp); #endif + rdev->pm.dpm.thermal.min_temp = low_temp; + rdev->pm.dpm.thermal.max_temp = high_temp; + return 0; } diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index b625646bf3e..fa9565957f9 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -3483,7 +3483,7 @@ static void cik_gpu_init(struct radeon_device *rdev) u32 mc_shared_chmap, mc_arb_ramcfg; u32 hdp_host_path_cntl; u32 tmp; - int i, j, k; + int i, j; switch (rdev->family) { case CHIP_BONAIRE: @@ -3544,6 +3544,7 @@ static void cik_gpu_init(struct radeon_device *rdev) (rdev->pdev->device == 0x130B) || (rdev->pdev->device == 0x130E) || (rdev->pdev->device == 0x1315) || + (rdev->pdev->device == 0x1318) || (rdev->pdev->device == 0x131B)) { rdev->config.cik.max_cu_per_sh = 4; rdev->config.cik.max_backends_per_se = 1; @@ -3672,12 +3673,11 @@ static void cik_gpu_init(struct radeon_device *rdev) rdev->config.cik.max_sh_per_se, rdev->config.cik.max_backends_per_se); + rdev->config.cik.active_cus = 0; for (i = 0; i < rdev->config.cik.max_shader_engines; i++) { for (j = 0; j < rdev->config.cik.max_sh_per_se; j++) { - for (k = 0; k < rdev->config.cik.max_cu_per_sh; k++) { - rdev->config.cik.active_cus += - hweight32(cik_get_cu_active_bitmap(rdev, i, j)); - } + rdev->config.cik.active_cus += + hweight32(cik_get_cu_active_bitmap(rdev, i, j)); } } @@ -3801,7 +3801,7 @@ int cik_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1)); radeon_ring_write(ring, ((scratch - PACKET3_SET_UCONFIG_REG_START) >> 2)); radeon_ring_write(ring, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); @@ -3920,6 +3920,17 @@ void cik_fence_compute_ring_emit(struct radeon_device *rdev, radeon_ring_write(ring, 0); } +/** + * cik_semaphore_ring_emit - emit a semaphore on the CP ring + * + * @rdev: radeon_device pointer + * @ring: radeon ring buffer object + * @semaphore: radeon semaphore object + * @emit_wait: Is this a sempahore wait? + * + * Emits a semaphore signal/wait packet to the CP ring and prevents the PFP + * from running ahead of semaphore waits. + */ bool cik_semaphore_ring_emit(struct radeon_device *rdev, struct radeon_ring *ring, struct radeon_semaphore *semaphore, @@ -3932,6 +3943,12 @@ bool cik_semaphore_ring_emit(struct radeon_device *rdev, radeon_ring_write(ring, lower_32_bits(addr)); radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | sel); + if (emit_wait && ring->idx == RADEON_RING_TYPE_GFX_INDEX) { + /* Prevent the PFP from running ahead of the semaphore wait */ + radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0)); + radeon_ring_write(ring, 0x0); + } + return true; } @@ -4004,7 +4021,7 @@ int cik_copy_cpdma(struct radeon_device *rdev, return r; } - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); radeon_semaphore_free(rdev, &sem, *fence); return r; @@ -4103,7 +4120,7 @@ int cik_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) ib.ptr[1] = ((scratch - PACKET3_SET_UCONFIG_REG_START) >> 2); ib.ptr[2] = 0xDEADBEEF; ib.length_dw = 3; - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) { radeon_scratch_free(rdev, scratch); radeon_ib_free(rdev, &ib); @@ -4324,7 +4341,7 @@ static int cik_cp_gfx_start(struct radeon_device *rdev) radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */ - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); return 0; } @@ -5732,20 +5749,17 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev) WREG32(0x15D8, 0); WREG32(0x15DC, 0); - /* empty context1-15 */ - /* FIXME start with 4G, once using 2 level pt switch to full - * vm size space - */ + /* restore context1-15 */ /* set vm size, must be a multiple of 4 */ WREG32(VM_CONTEXT1_PAGE_TABLE_START_ADDR, 0); WREG32(VM_CONTEXT1_PAGE_TABLE_END_ADDR, rdev->vm_manager.max_pfn); for (i = 1; i < 16; i++) { if (i < 8) WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2), - rdev->gart.table_addr >> 12); + rdev->vm_manager.saved_table_addr[i]); else WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2), - rdev->gart.table_addr >> 12); + rdev->vm_manager.saved_table_addr[i]); } /* enable context1-15 */ @@ -5810,6 +5824,17 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev) */ static void cik_pcie_gart_disable(struct radeon_device *rdev) { + unsigned i; + + for (i = 1; i < 16; ++i) { + uint32_t reg; + if (i < 8) + reg = VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2); + else + reg = VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2); + rdev->vm_manager.saved_table_addr[i] = RREG32(reg); + } + /* Disable all tables */ WREG32(VM_CONTEXT0_CNTL, 0); WREG32(VM_CONTEXT1_CNTL, 0); @@ -5958,14 +5983,14 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) /* update SH_MEM_* regs */ radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) | WRITE_DATA_DST_SEL(0))); radeon_ring_write(ring, SRBM_GFX_CNTL >> 2); radeon_ring_write(ring, 0); radeon_ring_write(ring, VMID(vm->id)); radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 6)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) | WRITE_DATA_DST_SEL(0))); radeon_ring_write(ring, SH_MEM_BASES >> 2); radeon_ring_write(ring, 0); @@ -5976,7 +6001,7 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) radeon_ring_write(ring, 0); /* SH_MEM_APE1_LIMIT */ radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) | WRITE_DATA_DST_SEL(0))); radeon_ring_write(ring, SRBM_GFX_CNTL >> 2); radeon_ring_write(ring, 0); @@ -5987,7 +6012,7 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) /* bits 0-15 are the VM contexts0-15 */ radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) | WRITE_DATA_DST_SEL(0))); radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2); radeon_ring_write(ring, 0); @@ -9538,6 +9563,9 @@ static void cik_pcie_gen3_enable(struct radeon_device *rdev) int ret, i; u16 tmp16; + if (pci_is_root_bus(rdev->pdev->bus)) + return; + if (radeon_pcie_gen2 == 0) return; @@ -9764,7 +9792,8 @@ static void cik_program_aspm(struct radeon_device *rdev) if (orig != data) WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, data); - if (!disable_clkreq) { + if (!disable_clkreq && + !pci_is_root_bus(rdev->pdev->bus)) { struct pci_dev *root = rdev->pdev->bus->self; u32 lnkcap; diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c index bcf480510ac..192278bc993 100644 --- a/drivers/gpu/drm/radeon/cik_sdma.c +++ b/drivers/gpu/drm/radeon/cik_sdma.c @@ -596,7 +596,7 @@ int cik_copy_dma(struct radeon_device *rdev, return r; } - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); radeon_semaphore_free(rdev, &sem, *fence); return r; @@ -638,7 +638,7 @@ int cik_sdma_ring_test(struct radeon_device *rdev, radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr)); radeon_ring_write(ring, 1); /* number of DWs to follow */ radeon_ring_write(ring, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); for (i = 0; i < rdev->usec_timeout; i++) { tmp = readl(ptr); @@ -695,7 +695,7 @@ int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) ib.ptr[4] = 0xDEADBEEF; ib.length_dw = 5; - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) { radeon_ib_free(rdev, &ib); DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c index 4fedd14e670..dbca60c7d09 100644 --- a/drivers/gpu/drm/radeon/evergreen.c +++ b/drivers/gpu/drm/radeon/evergreen.c @@ -2869,7 +2869,7 @@ static int evergreen_cp_start(struct radeon_device *rdev) radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1)); radeon_ring_write(ring, 0); radeon_ring_write(ring, 0); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); cp_me = 0xff; WREG32(CP_ME_CNTL, cp_me); @@ -2912,7 +2912,7 @@ static int evergreen_cp_start(struct radeon_device *rdev) radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ radeon_ring_write(ring, 0x00000010); /* */ - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); return 0; } diff --git a/drivers/gpu/drm/radeon/evergreen_dma.c b/drivers/gpu/drm/radeon/evergreen_dma.c index 478caefe0fe..afaba388c36 100644 --- a/drivers/gpu/drm/radeon/evergreen_dma.c +++ b/drivers/gpu/drm/radeon/evergreen_dma.c @@ -155,7 +155,7 @@ int evergreen_copy_dma(struct radeon_device *rdev, return r; } - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); radeon_semaphore_free(rdev, &sem, *fence); return r; diff --git a/drivers/gpu/drm/radeon/kv_dpm.c b/drivers/gpu/drm/radeon/kv_dpm.c index 9ef8c38f2d6..8b58e11b64f 100644 --- a/drivers/gpu/drm/radeon/kv_dpm.c +++ b/drivers/gpu/drm/radeon/kv_dpm.c @@ -1438,14 +1438,14 @@ static int kv_update_uvd_dpm(struct radeon_device *rdev, bool gate) return kv_enable_uvd_dpm(rdev, !gate); } -static u8 kv_get_vce_boot_level(struct radeon_device *rdev) +static u8 kv_get_vce_boot_level(struct radeon_device *rdev, u32 evclk) { u8 i; struct radeon_vce_clock_voltage_dependency_table *table = &rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table; for (i = 0; i < table->count; i++) { - if (table->entries[i].evclk >= 0) /* XXX */ + if (table->entries[i].evclk >= evclk) break; } @@ -1468,7 +1468,7 @@ static int kv_update_vce_dpm(struct radeon_device *rdev, if (pi->caps_stable_p_state) pi->vce_boot_level = table->count - 1; else - pi->vce_boot_level = kv_get_vce_boot_level(rdev); + pi->vce_boot_level = kv_get_vce_boot_level(rdev, radeon_new_state->evclk); ret = kv_copy_bytes_to_smc(rdev, pi->dpm_table_start + @@ -2726,7 +2726,10 @@ int kv_dpm_init(struct radeon_device *rdev) pi->caps_sclk_ds = true; pi->enable_auto_thermal_throttling = true; pi->disable_nb_ps3_in_battery = false; - pi->bapm_enable = true; + if (radeon_bapm == 0) + pi->bapm_enable = false; + else + pi->bapm_enable = true; pi->voltage_drop_t = 0; pi->caps_sclk_throttle_low_notification = false; pi->caps_fps = false; /* true? */ diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c index 327b85f7fd0..3faee58946d 100644 --- a/drivers/gpu/drm/radeon/ni.c +++ b/drivers/gpu/drm/radeon/ni.c @@ -1271,7 +1271,7 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev) WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (i << 2), 0); WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (i << 2), rdev->vm_manager.max_pfn); WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2), - rdev->gart.table_addr >> 12); + rdev->vm_manager.saved_table_addr[i]); } /* enable context1-7 */ @@ -1303,6 +1303,13 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev) static void cayman_pcie_gart_disable(struct radeon_device *rdev) { + unsigned i; + + for (i = 1; i < 8; ++i) { + rdev->vm_manager.saved_table_addr[i] = RREG32( + VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2)); + } + /* Disable all tables */ WREG32(VM_CONTEXT0_CNTL, 0); WREG32(VM_CONTEXT1_CNTL, 0); @@ -1505,7 +1512,7 @@ static int cayman_cp_start(struct radeon_device *rdev) radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1)); radeon_ring_write(ring, 0); radeon_ring_write(ring, 0); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); cayman_cp_enable(rdev, true); @@ -1547,7 +1554,7 @@ static int cayman_cp_start(struct radeon_device *rdev) radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ radeon_ring_write(ring, 0x00000010); /* */ - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); /* XXX init other rings */ diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 04b5940b892..4c5ec44ff32 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -925,7 +925,7 @@ int r100_copy_blit(struct radeon_device *rdev, if (fence) { r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX); } - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); return r; } @@ -958,7 +958,7 @@ void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring) RADEON_ISYNC_ANY3D_IDLE2D | RADEON_ISYNC_WAIT_IDLEGUI | RADEON_ISYNC_CPSCRATCH_IDLEGUI); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); } @@ -3638,7 +3638,7 @@ int r100_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) } radeon_ring_write(ring, PACKET0(scratch, 0)); radeon_ring_write(ring, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) { @@ -3700,7 +3700,7 @@ int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) ib.ptr[6] = PACKET2(0); ib.ptr[7] = PACKET2(0); ib.length_dw = 8; - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) { DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); goto free_ib; diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c index 58f0473aa73..67780374a65 100644 --- a/drivers/gpu/drm/radeon/r200.c +++ b/drivers/gpu/drm/radeon/r200.c @@ -121,7 +121,7 @@ int r200_copy_dma(struct radeon_device *rdev, if (fence) { r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX); } - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); return r; } diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index 75b30338c22..1bc4704034c 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -295,7 +295,7 @@ void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_write(ring, R300_GEOMETRY_ROUND_NEAREST | R300_COLOR_ROUND_NEAREST); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); } static void r300_errata(struct radeon_device *rdev) diff --git a/drivers/gpu/drm/radeon/r420.c b/drivers/gpu/drm/radeon/r420.c index 802b19220a2..2828605aef3 100644 --- a/drivers/gpu/drm/radeon/r420.c +++ b/drivers/gpu/drm/radeon/r420.c @@ -219,7 +219,7 @@ static void r420_cp_errata_init(struct radeon_device *rdev) radeon_ring_write(ring, PACKET0(R300_CP_RESYNC_ADDR, 1)); radeon_ring_write(ring, rdev->config.r300.resync_scratch); radeon_ring_write(ring, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); } static void r420_cp_errata_fini(struct radeon_device *rdev) @@ -232,7 +232,7 @@ static void r420_cp_errata_fini(struct radeon_device *rdev) radeon_ring_lock(rdev, ring, 8); radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0)); radeon_ring_write(ring, R300_RB3D_DC_FINISH); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); radeon_scratch_free(rdev, rdev->config.r300.resync_scratch); } diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c index c70a504d96a..e616eb5f6e7 100644 --- a/drivers/gpu/drm/radeon/r600.c +++ b/drivers/gpu/drm/radeon/r600.c @@ -1812,7 +1812,6 @@ static void r600_gpu_init(struct radeon_device *rdev) { u32 tiling_config; u32 ramcfg; - u32 cc_rb_backend_disable; u32 cc_gc_shader_pipe_config; u32 tmp; int i, j; @@ -1939,29 +1938,20 @@ static void r600_gpu_init(struct radeon_device *rdev) } tiling_config |= BANK_SWAPS(1); - cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE) & 0x00ff0000; - tmp = R6XX_MAX_BACKENDS - - r600_count_pipe_bits((cc_rb_backend_disable >> 16) & R6XX_MAX_BACKENDS_MASK); - if (tmp < rdev->config.r600.max_backends) { - rdev->config.r600.max_backends = tmp; - } - cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0x00ffff00; - tmp = R6XX_MAX_PIPES - - r600_count_pipe_bits((cc_gc_shader_pipe_config >> 8) & R6XX_MAX_PIPES_MASK); - if (tmp < rdev->config.r600.max_pipes) { - rdev->config.r600.max_pipes = tmp; - } - tmp = R6XX_MAX_SIMDS - - r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R6XX_MAX_SIMDS_MASK); - if (tmp < rdev->config.r600.max_simds) { - rdev->config.r600.max_simds = tmp; - } tmp = rdev->config.r600.max_simds - r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R6XX_MAX_SIMDS_MASK); rdev->config.r600.active_simds = tmp; disabled_rb_mask = (RREG32(CC_RB_BACKEND_DISABLE) >> 16) & R6XX_MAX_BACKENDS_MASK; + tmp = 0; + for (i = 0; i < rdev->config.r600.max_backends; i++) + tmp |= (1 << i); + /* if all the backends are disabled, fix it up here */ + if ((disabled_rb_mask & tmp) == tmp) { + for (i = 0; i < rdev->config.r600.max_backends; i++) + disabled_rb_mask &= ~(1 << i); + } tmp = (tiling_config & PIPE_TILING__MASK) >> PIPE_TILING__SHIFT; tmp = r6xx_remap_render_backend(rdev, tmp, rdev->config.r600.max_backends, R6XX_MAX_BACKENDS, disabled_rb_mask); @@ -2547,7 +2537,7 @@ int r600_cp_start(struct radeon_device *rdev) radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1)); radeon_ring_write(ring, 0); radeon_ring_write(ring, 0); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); cp_me = 0xff; WREG32(R_0086D8_CP_ME_CNTL, cp_me); @@ -2683,7 +2673,7 @@ int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); radeon_ring_write(ring, ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2)); radeon_ring_write(ring, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) @@ -2753,6 +2743,17 @@ void r600_fence_ring_emit(struct radeon_device *rdev, } } +/** + * r600_semaphore_ring_emit - emit a semaphore on the CP ring + * + * @rdev: radeon_device pointer + * @ring: radeon ring buffer object + * @semaphore: radeon semaphore object + * @emit_wait: Is this a sempahore wait? + * + * Emits a semaphore signal/wait packet to the CP ring and prevents the PFP + * from running ahead of semaphore waits. + */ bool r600_semaphore_ring_emit(struct radeon_device *rdev, struct radeon_ring *ring, struct radeon_semaphore *semaphore, @@ -2768,6 +2769,13 @@ bool r600_semaphore_ring_emit(struct radeon_device *rdev, radeon_ring_write(ring, lower_32_bits(addr)); radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | sel); + /* PFP_SYNC_ME packet only exists on 7xx+ */ + if (emit_wait && (rdev->family >= CHIP_RV770)) { + /* Prevent the PFP from running ahead of the semaphore wait */ + radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0)); + radeon_ring_write(ring, 0x0); + } + return true; } @@ -2845,7 +2853,7 @@ int r600_copy_cpdma(struct radeon_device *rdev, return r; } - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); radeon_semaphore_free(rdev, &sem, *fence); return r; @@ -3165,7 +3173,7 @@ int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) ib.ptr[1] = ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); ib.ptr[2] = 0xDEADBEEF; ib.length_dw = 3; - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) { DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); goto free_ib; diff --git a/drivers/gpu/drm/radeon/r600_dma.c b/drivers/gpu/drm/radeon/r600_dma.c index 4969cef44a1..51fd98553ea 100644 --- a/drivers/gpu/drm/radeon/r600_dma.c +++ b/drivers/gpu/drm/radeon/r600_dma.c @@ -261,7 +261,7 @@ int r600_dma_ring_test(struct radeon_device *rdev, radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc); radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xff); radeon_ring_write(ring, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); for (i = 0; i < rdev->usec_timeout; i++) { tmp = readl(ptr); @@ -368,7 +368,7 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) ib.ptr[3] = 0xDEADBEEF; ib.length_dw = 4; - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) { radeon_ib_free(rdev, &ib); DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); @@ -493,7 +493,7 @@ int r600_copy_dma(struct radeon_device *rdev, return r; } - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); radeon_semaphore_free(rdev, &sem, *fence); return r; diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h index f94e7a9afe7..0c4a7d8d93e 100644 --- a/drivers/gpu/drm/radeon/r600d.h +++ b/drivers/gpu/drm/radeon/r600d.h @@ -1597,6 +1597,7 @@ */ # define PACKET3_CP_DMA_CMD_SAIC (1 << 28) # define PACKET3_CP_DMA_CMD_DAIC (1 << 29) +#define PACKET3_PFP_SYNC_ME 0x42 /* r7xx+ only */ #define PACKET3_SURFACE_SYNC 0x43 # define PACKET3_CB0_DEST_BASE_ENA (1 << 6) # define PACKET3_FULL_CACHE_ENA (1 << 20) /* r7xx+ only */ diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 9e1732eb402..5f05b4c8433 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -105,6 +105,7 @@ extern int radeon_vm_size; extern int radeon_vm_block_size; extern int radeon_deep_color; extern int radeon_use_pflipirq; +extern int radeon_bapm; /* * Copy from radeon_drv.h so we don't have to include both and have conflicting @@ -914,6 +915,8 @@ struct radeon_vm_manager { u64 vram_base_offset; /* is vm enabled? */ bool enabled; + /* for hw to save the PD addr on suspend/resume */ + uint32_t saved_table_addr[RADEON_NUM_VM]; }; /* @@ -967,7 +970,7 @@ int radeon_ib_get(struct radeon_device *rdev, int ring, unsigned size); void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib); int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, - struct radeon_ib *const_ib); + struct radeon_ib *const_ib, bool hdp_flush); int radeon_ib_pool_init(struct radeon_device *rdev); void radeon_ib_pool_fini(struct radeon_device *rdev); int radeon_ib_ring_tests(struct radeon_device *rdev); @@ -977,8 +980,10 @@ bool radeon_ring_supports_scratch_reg(struct radeon_device *rdev, void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *cp); int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ndw); int radeon_ring_lock(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ndw); -void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *cp); -void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *cp); +void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *cp, + bool hdp_flush); +void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *cp, + bool hdp_flush); void radeon_ring_undo(struct radeon_ring *ring); void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *cp); int radeon_ring_test(struct radeon_device *rdev, struct radeon_ring *cp); diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index ee712c199b2..83f382e8e40 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -132,7 +132,8 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) * the buffers used for read only, which doubles the range * to 0 to 31. 32 is reserved for the kernel driver. */ - priority = (r->flags & 0xf) * 2 + !!r->write_domain; + priority = (r->flags & RADEON_RELOC_PRIO_MASK) * 2 + + !!r->write_domain; /* the first reloc of an UVD job is the msg and that must be in VRAM, also but everything into VRAM on AGP cards to avoid @@ -450,7 +451,7 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev, radeon_vce_note_usage(rdev); radeon_cs_sync_rings(parser); - r = radeon_ib_schedule(rdev, &parser->ib, NULL); + r = radeon_ib_schedule(rdev, &parser->ib, NULL, true); if (r) { DRM_ERROR("Failed to schedule IB !\n"); } @@ -541,9 +542,9 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, if ((rdev->family >= CHIP_TAHITI) && (parser->chunk_const_ib_idx != -1)) { - r = radeon_ib_schedule(rdev, &parser->ib, &parser->const_ib); + r = radeon_ib_schedule(rdev, &parser->ib, &parser->const_ib, true); } else { - r = radeon_ib_schedule(rdev, &parser->ib, NULL); + r = radeon_ib_schedule(rdev, &parser->ib, NULL, true); } out: diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index c8ea050c8fa..6a219bcee66 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -1680,8 +1680,8 @@ int radeon_gpu_reset(struct radeon_device *rdev) radeon_save_bios_scratch_regs(rdev); /* block TTM */ resched = ttm_bo_lock_delayed_workqueue(&rdev->mman.bdev); - radeon_pm_suspend(rdev); radeon_suspend(rdev); + radeon_hpd_fini(rdev); for (i = 0; i < RADEON_NUM_RINGS; ++i) { ring_sizes[i] = radeon_ring_backup(rdev, &rdev->ring[i], @@ -1726,9 +1726,39 @@ retry: } } - radeon_pm_resume(rdev); + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { + /* do dpm late init */ + r = radeon_pm_late_init(rdev); + if (r) { + rdev->pm.dpm_enabled = false; + DRM_ERROR("radeon_pm_late_init failed, disabling dpm\n"); + } + } else { + /* resume old pm late */ + radeon_pm_resume(rdev); + } + + /* init dig PHYs, disp eng pll */ + if (rdev->is_atom_bios) { + radeon_atom_encoder_init(rdev); + radeon_atom_disp_eng_pll_init(rdev); + /* turn on the BL */ + if (rdev->mode_info.bl_encoder) { + u8 bl_level = radeon_get_backlight_level(rdev, + rdev->mode_info.bl_encoder); + radeon_set_backlight_level(rdev, rdev->mode_info.bl_encoder, + bl_level); + } + } + /* reset hpd state */ + radeon_hpd_init(rdev); + drm_helper_resume_force_mode(rdev->ddev); + /* set the power state here in case we are a PX system or headless */ + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) + radeon_pm_compute_clocks(rdev); + ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched); if (r) { /* bad news, how to tell it to userspace ? */ diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index 092d067f93e..8df88890883 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -180,6 +180,7 @@ int radeon_vm_size = 8; int radeon_vm_block_size = -1; int radeon_deep_color = 0; int radeon_use_pflipirq = 2; +int radeon_bapm = -1; MODULE_PARM_DESC(no_wb, "Disable AGP writeback for scratch registers"); module_param_named(no_wb, radeon_no_wb, int, 0444); @@ -259,6 +260,9 @@ module_param_named(deep_color, radeon_deep_color, int, 0444); MODULE_PARM_DESC(use_pflipirq, "Pflip irqs for pageflip completion (0 = disable, 1 = as fallback, 2 = exclusive (default))"); module_param_named(use_pflipirq, radeon_use_pflipirq, int, 0444); +MODULE_PARM_DESC(bapm, "BAPM support (1 = enable, 0 = disable, -1 = auto)"); +module_param_named(bapm, radeon_bapm, int, 0444); + static struct pci_device_id pciidlist[] = { radeon_PCI_IDS }; diff --git a/drivers/gpu/drm/radeon/radeon_ib.c b/drivers/gpu/drm/radeon/radeon_ib.c index 65b0c213488..5bf2c0a0582 100644 --- a/drivers/gpu/drm/radeon/radeon_ib.c +++ b/drivers/gpu/drm/radeon/radeon_ib.c @@ -107,6 +107,7 @@ void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib) * @rdev: radeon_device pointer * @ib: IB object to schedule * @const_ib: Const IB to schedule (SI only) + * @hdp_flush: Whether or not to perform an HDP cache flush * * Schedule an IB on the associated ring (all asics). * Returns 0 on success, error on failure. @@ -122,7 +123,7 @@ void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib) * to SI there was just a DE IB. */ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, - struct radeon_ib *const_ib) + struct radeon_ib *const_ib, bool hdp_flush) { struct radeon_ring *ring = &rdev->ring[ib->ring]; int r = 0; @@ -176,7 +177,7 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, if (ib->vm) radeon_vm_fence(rdev, ib->vm, ib->fence); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, hdp_flush); return 0; } diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c index 23314be4948..164898b0010 100644 --- a/drivers/gpu/drm/radeon/radeon_pm.c +++ b/drivers/gpu/drm/radeon/radeon_pm.c @@ -460,10 +460,6 @@ static ssize_t radeon_get_dpm_state(struct device *dev, struct radeon_device *rdev = ddev->dev_private; enum radeon_pm_state_type pm = rdev->pm.dpm.user_state; - if ((rdev->flags & RADEON_IS_PX) && - (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) - return snprintf(buf, PAGE_SIZE, "off\n"); - return snprintf(buf, PAGE_SIZE, "%s\n", (pm == POWER_STATE_TYPE_BATTERY) ? "battery" : (pm == POWER_STATE_TYPE_BALANCED) ? "balanced" : "performance"); @@ -477,11 +473,6 @@ static ssize_t radeon_set_dpm_state(struct device *dev, struct drm_device *ddev = dev_get_drvdata(dev); struct radeon_device *rdev = ddev->dev_private; - /* Can't set dpm state when the card is off */ - if ((rdev->flags & RADEON_IS_PX) && - (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) - return -EINVAL; - mutex_lock(&rdev->pm.mutex); if (strncmp("battery", buf, strlen("battery")) == 0) rdev->pm.dpm.user_state = POWER_STATE_TYPE_BATTERY; @@ -495,7 +486,12 @@ static ssize_t radeon_set_dpm_state(struct device *dev, goto fail; } mutex_unlock(&rdev->pm.mutex); - radeon_pm_compute_clocks(rdev); + + /* Can't set dpm state when the card is off */ + if (!(rdev->flags & RADEON_IS_PX) || + (ddev->switch_power_state == DRM_SWITCH_POWER_ON)) + radeon_pm_compute_clocks(rdev); + fail: return count; } diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 5b4e0cf231a..d6560790253 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -177,16 +177,18 @@ int radeon_ring_lock(struct radeon_device *rdev, struct radeon_ring *ring, unsig * * @rdev: radeon_device pointer * @ring: radeon_ring structure holding ring information + * @hdp_flush: Whether or not to perform an HDP cache flush * * Update the wptr (write pointer) to tell the GPU to * execute new commands on the ring buffer (all asics). */ -void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring) +void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring, + bool hdp_flush) { /* If we are emitting the HDP flush via the ring buffer, we need to * do it before padding. */ - if (rdev->asic->ring[ring->idx]->hdp_flush) + if (hdp_flush && rdev->asic->ring[ring->idx]->hdp_flush) rdev->asic->ring[ring->idx]->hdp_flush(rdev, ring); /* We pad to match fetch size */ while (ring->wptr & ring->align_mask) { @@ -196,7 +198,7 @@ void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring) /* If we are emitting the HDP flush via MMIO, we need to do it after * all CPU writes to VRAM finished. */ - if (rdev->asic->mmio_hdp_flush) + if (hdp_flush && rdev->asic->mmio_hdp_flush) rdev->asic->mmio_hdp_flush(rdev); radeon_ring_set_wptr(rdev, ring); } @@ -207,12 +209,14 @@ void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring) * * @rdev: radeon_device pointer * @ring: radeon_ring structure holding ring information + * @hdp_flush: Whether or not to perform an HDP cache flush * * Call radeon_ring_commit() then unlock the ring (all asics). */ -void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *ring) +void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *ring, + bool hdp_flush) { - radeon_ring_commit(rdev, ring); + radeon_ring_commit(rdev, ring, hdp_flush); mutex_unlock(&rdev->ring_lock); } @@ -372,7 +376,7 @@ int radeon_ring_restore(struct radeon_device *rdev, struct radeon_ring *ring, radeon_ring_write(ring, data[i]); } - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); kfree(data); return 0; } @@ -400,9 +404,7 @@ int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsig /* Allocate ring buffer */ if (ring->ring_obj == NULL) { r = radeon_bo_create(rdev, ring->ring_size, PAGE_SIZE, true, - RADEON_GEM_DOMAIN_GTT, - (rdev->flags & RADEON_IS_PCIE) ? - RADEON_GEM_GTT_WC : 0, + RADEON_GEM_DOMAIN_GTT, 0, NULL, &ring->ring_obj); if (r) { dev_err(rdev->dev, "(%d) ring create failed\n", r); diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c index dbd6bcde92d..56d9fd66d8a 100644 --- a/drivers/gpu/drm/radeon/radeon_semaphore.c +++ b/drivers/gpu/drm/radeon/radeon_semaphore.c @@ -179,7 +179,7 @@ int radeon_semaphore_sync_rings(struct radeon_device *rdev, continue; } - radeon_ring_commit(rdev, &rdev->ring[i]); + radeon_ring_commit(rdev, &rdev->ring[i], false); radeon_fence_note_sync(fence, ring); semaphore->gpu_addr += 8; diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c index 5adf4207453..17bc3dced9f 100644 --- a/drivers/gpu/drm/radeon/radeon_test.c +++ b/drivers/gpu/drm/radeon/radeon_test.c @@ -288,7 +288,7 @@ static int radeon_test_create_and_emit_fence(struct radeon_device *rdev, return r; } radeon_fence_emit(rdev, fence, ring->idx); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); } return 0; } @@ -313,7 +313,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev, goto out_cleanup; } radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore); - radeon_ring_unlock_commit(rdev, ringA); + radeon_ring_unlock_commit(rdev, ringA, false); r = radeon_test_create_and_emit_fence(rdev, ringA, &fence1); if (r) @@ -325,7 +325,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev, goto out_cleanup; } radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore); - radeon_ring_unlock_commit(rdev, ringA); + radeon_ring_unlock_commit(rdev, ringA, false); r = radeon_test_create_and_emit_fence(rdev, ringA, &fence2); if (r) @@ -344,7 +344,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev, goto out_cleanup; } radeon_semaphore_emit_signal(rdev, ringB->idx, semaphore); - radeon_ring_unlock_commit(rdev, ringB); + radeon_ring_unlock_commit(rdev, ringB, false); r = radeon_fence_wait(fence1, false); if (r) { @@ -365,7 +365,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev, goto out_cleanup; } radeon_semaphore_emit_signal(rdev, ringB->idx, semaphore); - radeon_ring_unlock_commit(rdev, ringB); + radeon_ring_unlock_commit(rdev, ringB, false); r = radeon_fence_wait(fence2, false); if (r) { @@ -408,7 +408,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev, goto out_cleanup; } radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore); - radeon_ring_unlock_commit(rdev, ringA); + radeon_ring_unlock_commit(rdev, ringA, false); r = radeon_test_create_and_emit_fence(rdev, ringA, &fenceA); if (r) @@ -420,7 +420,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev, goto out_cleanup; } radeon_semaphore_emit_wait(rdev, ringB->idx, semaphore); - radeon_ring_unlock_commit(rdev, ringB); + radeon_ring_unlock_commit(rdev, ringB, false); r = radeon_test_create_and_emit_fence(rdev, ringB, &fenceB); if (r) goto out_cleanup; @@ -442,7 +442,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev, goto out_cleanup; } radeon_semaphore_emit_signal(rdev, ringC->idx, semaphore); - radeon_ring_unlock_commit(rdev, ringC); + radeon_ring_unlock_commit(rdev, ringC, false); for (i = 0; i < 30; ++i) { mdelay(100); @@ -468,7 +468,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev, goto out_cleanup; } radeon_semaphore_emit_signal(rdev, ringC->idx, semaphore); - radeon_ring_unlock_commit(rdev, ringC); + radeon_ring_unlock_commit(rdev, ringC, false); mdelay(1000); diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c index 6bf55ec85b6..341848a1437 100644 --- a/drivers/gpu/drm/radeon/radeon_uvd.c +++ b/drivers/gpu/drm/radeon/radeon_uvd.c @@ -646,7 +646,7 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev, ib.ptr[i] = PACKET2(0); ib.length_dw = 16; - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) goto err; ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence); diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c index f9b70a43aa5..c7190aadbd8 100644 --- a/drivers/gpu/drm/radeon/radeon_vce.c +++ b/drivers/gpu/drm/radeon/radeon_vce.c @@ -368,7 +368,7 @@ int radeon_vce_get_create_msg(struct radeon_device *rdev, int ring, for (i = ib.length_dw; i < ib_size_dw; ++i) ib.ptr[i] = 0x0; - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) { DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); } @@ -425,7 +425,7 @@ int radeon_vce_get_destroy_msg(struct radeon_device *rdev, int ring, for (i = ib.length_dw; i < ib_size_dw; ++i) ib.ptr[i] = 0x0; - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) { DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); } @@ -715,7 +715,7 @@ int radeon_vce_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) return r; } radeon_ring_write(ring, VCE_CMD_END); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); for (i = 0; i < rdev->usec_timeout; i++) { if (vce_v1_0_get_rptr(rdev, ring) != rptr) diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c index ccae4d9dc3d..088ffdc2f57 100644 --- a/drivers/gpu/drm/radeon/radeon_vm.c +++ b/drivers/gpu/drm/radeon/radeon_vm.c @@ -420,7 +420,7 @@ static int radeon_vm_clear_bo(struct radeon_device *rdev, radeon_asic_vm_pad_ib(rdev, &ib); WARN_ON(ib.length_dw > 64); - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) goto error; @@ -483,6 +483,10 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev, /* add a clone of the bo_va to clear the old address */ struct radeon_bo_va *tmp; tmp = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL); + if (!tmp) { + mutex_unlock(&vm->mutex); + return -ENOMEM; + } tmp->it.start = bo_va->it.start; tmp->it.last = bo_va->it.last; tmp->vm = vm; @@ -693,7 +697,7 @@ int radeon_vm_update_page_directory(struct radeon_device *rdev, radeon_semaphore_sync_to(ib.semaphore, pd->tbo.sync_obj); radeon_semaphore_sync_to(ib.semaphore, vm->last_id_use); WARN_ON(ib.length_dw > ndw); - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) { radeon_ib_free(rdev, &ib); return r; @@ -957,7 +961,7 @@ int radeon_vm_bo_update(struct radeon_device *rdev, WARN_ON(ib.length_dw > ndw); radeon_semaphore_sync_to(ib.semaphore, vm->fence); - r = radeon_ib_schedule(rdev, &ib, NULL); + r = radeon_ib_schedule(rdev, &ib, NULL, false); if (r) { radeon_ib_free(rdev, &ib); return r; diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c index 3e21e869015..8a477bf1fdb 100644 --- a/drivers/gpu/drm/radeon/rv515.c +++ b/drivers/gpu/drm/radeon/rv515.c @@ -124,7 +124,7 @@ void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_write(ring, GEOMETRY_ROUND_NEAREST | COLOR_ROUND_NEAREST); radeon_ring_write(ring, PACKET0(0x20C8, 0)); radeon_ring_write(ring, 0); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); } int rv515_mc_wait_for_idle(struct radeon_device *rdev) diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c index 2983f17ea1b..d9f5ce715c9 100644 --- a/drivers/gpu/drm/radeon/rv770.c +++ b/drivers/gpu/drm/radeon/rv770.c @@ -1177,7 +1177,6 @@ static void rv770_gpu_init(struct radeon_device *rdev) u32 hdp_host_path_cntl; u32 sq_dyn_gpr_size_simd_ab_0; u32 gb_tiling_config = 0; - u32 cc_rb_backend_disable = 0; u32 cc_gc_shader_pipe_config = 0; u32 mc_arb_ramcfg; u32 db_debug4, tmp; @@ -1311,21 +1310,7 @@ static void rv770_gpu_init(struct radeon_device *rdev) WREG32(SPI_CONFIG_CNTL, 0); } - cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE) & 0x00ff0000; - tmp = R7XX_MAX_BACKENDS - r600_count_pipe_bits(cc_rb_backend_disable >> 16); - if (tmp < rdev->config.rv770.max_backends) { - rdev->config.rv770.max_backends = tmp; - } - cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0xffffff00; - tmp = R7XX_MAX_PIPES - r600_count_pipe_bits((cc_gc_shader_pipe_config >> 8) & R7XX_MAX_PIPES_MASK); - if (tmp < rdev->config.rv770.max_pipes) { - rdev->config.rv770.max_pipes = tmp; - } - tmp = R7XX_MAX_SIMDS - r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R7XX_MAX_SIMDS_MASK); - if (tmp < rdev->config.rv770.max_simds) { - rdev->config.rv770.max_simds = tmp; - } tmp = rdev->config.rv770.max_simds - r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R7XX_MAX_SIMDS_MASK); rdev->config.rv770.active_simds = tmp; @@ -1348,6 +1333,14 @@ static void rv770_gpu_init(struct radeon_device *rdev) rdev->config.rv770.tiling_npipes = rdev->config.rv770.max_tile_pipes; disabled_rb_mask = (RREG32(CC_RB_BACKEND_DISABLE) >> 16) & R7XX_MAX_BACKENDS_MASK; + tmp = 0; + for (i = 0; i < rdev->config.rv770.max_backends; i++) + tmp |= (1 << i); + /* if all the backends are disabled, fix it up here */ + if ((disabled_rb_mask & tmp) == tmp) { + for (i = 0; i < rdev->config.rv770.max_backends; i++) + disabled_rb_mask &= ~(1 << i); + } tmp = (gb_tiling_config & PIPE_TILING__MASK) >> PIPE_TILING__SHIFT; tmp = r6xx_remap_render_backend(rdev, tmp, rdev->config.rv770.max_backends, R7XX_MAX_BACKENDS, disabled_rb_mask); diff --git a/drivers/gpu/drm/radeon/rv770_dma.c b/drivers/gpu/drm/radeon/rv770_dma.c index bbf2e076ee4..74426ac2bb5 100644 --- a/drivers/gpu/drm/radeon/rv770_dma.c +++ b/drivers/gpu/drm/radeon/rv770_dma.c @@ -90,7 +90,7 @@ int rv770_copy_dma(struct radeon_device *rdev, return r; } - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); radeon_semaphore_free(rdev, &sem, *fence); return r; diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index 011779bd2b3..6bce4084775 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -3057,7 +3057,7 @@ static void si_gpu_init(struct radeon_device *rdev) u32 sx_debug_1; u32 hdp_host_path_cntl; u32 tmp; - int i, j, k; + int i, j; switch (rdev->family) { case CHIP_TAHITI: @@ -3255,12 +3255,11 @@ static void si_gpu_init(struct radeon_device *rdev) rdev->config.si.max_sh_per_se, rdev->config.si.max_cu_per_sh); + rdev->config.si.active_cus = 0; for (i = 0; i < rdev->config.si.max_shader_engines; i++) { for (j = 0; j < rdev->config.si.max_sh_per_se; j++) { - for (k = 0; k < rdev->config.si.max_cu_per_sh; k++) { - rdev->config.si.active_cus += - hweight32(si_get_cu_active_bitmap(rdev, i, j)); - } + rdev->config.si.active_cus += + hweight32(si_get_cu_active_bitmap(rdev, i, j)); } } @@ -3541,7 +3540,7 @@ static int si_cp_start(struct radeon_device *rdev) radeon_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE)); radeon_ring_write(ring, 0xc000); radeon_ring_write(ring, 0xe000); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); si_cp_enable(rdev, true); @@ -3570,7 +3569,7 @@ static int si_cp_start(struct radeon_device *rdev) radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */ - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); for (i = RADEON_RING_TYPE_GFX_INDEX; i <= CAYMAN_RING_TYPE_CP2_INDEX; ++i) { ring = &rdev->ring[i]; @@ -3580,7 +3579,7 @@ static int si_cp_start(struct radeon_device *rdev) radeon_ring_write(ring, PACKET3_COMPUTE(PACKET3_CLEAR_STATE, 0)); radeon_ring_write(ring, 0); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); } return 0; @@ -4291,10 +4290,10 @@ static int si_pcie_gart_enable(struct radeon_device *rdev) for (i = 1; i < 16; i++) { if (i < 8) WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2), - rdev->gart.table_addr >> 12); + rdev->vm_manager.saved_table_addr[i]); else WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2), - rdev->gart.table_addr >> 12); + rdev->vm_manager.saved_table_addr[i]); } /* enable context1-15 */ @@ -4326,6 +4325,17 @@ static int si_pcie_gart_enable(struct radeon_device *rdev) static void si_pcie_gart_disable(struct radeon_device *rdev) { + unsigned i; + + for (i = 1; i < 16; ++i) { + uint32_t reg; + if (i < 8) + reg = VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2); + else + reg = VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2); + rdev->vm_manager.saved_table_addr[i] = RREG32(reg); + } + /* Disable all tables */ WREG32(VM_CONTEXT0_CNTL, 0); WREG32(VM_CONTEXT1_CNTL, 0); @@ -5028,7 +5038,7 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) /* flush hdp cache */ radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) | WRITE_DATA_DST_SEL(0))); radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2); radeon_ring_write(ring, 0); @@ -5036,7 +5046,7 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm) /* bits 0-15 are the VM contexts0-15 */ radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) | WRITE_DATA_DST_SEL(0))); radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2); radeon_ring_write(ring, 0); @@ -7178,6 +7188,9 @@ static void si_pcie_gen3_enable(struct radeon_device *rdev) int ret, i; u16 tmp16; + if (pci_is_root_bus(rdev->pdev->bus)) + return; + if (radeon_pcie_gen2 == 0) return; @@ -7455,7 +7468,8 @@ static void si_program_aspm(struct radeon_device *rdev) if (orig != data) WREG32_PIF_PHY1(PB1_PIF_CNTL, data); - if (!disable_clkreq) { + if (!disable_clkreq && + !pci_is_root_bus(rdev->pdev->bus)) { struct pci_dev *root = rdev->pdev->bus->self; u32 lnkcap; diff --git a/drivers/gpu/drm/radeon/si_dma.c b/drivers/gpu/drm/radeon/si_dma.c index 71650512945..7c22baaf94d 100644 --- a/drivers/gpu/drm/radeon/si_dma.c +++ b/drivers/gpu/drm/radeon/si_dma.c @@ -275,7 +275,7 @@ int si_copy_dma(struct radeon_device *rdev, return r; } - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); radeon_semaphore_free(rdev, &sem, *fence); return r; diff --git a/drivers/gpu/drm/radeon/trinity_dpm.c b/drivers/gpu/drm/radeon/trinity_dpm.c index 32e50be9c4a..57f780053b3 100644 --- a/drivers/gpu/drm/radeon/trinity_dpm.c +++ b/drivers/gpu/drm/radeon/trinity_dpm.c @@ -1874,16 +1874,22 @@ int trinity_dpm_init(struct radeon_device *rdev) for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++) pi->at[i] = TRINITY_AT_DFLT; - /* There are stability issues reported on with - * bapm enabled when switching between AC and battery - * power. At the same time, some MSI boards hang - * if it's not enabled and dpm is enabled. Just enable - * it for MSI boards right now. - */ - if (rdev->pdev->subsystem_vendor == 0x1462) - pi->enable_bapm = true; - else + if (radeon_bapm == -1) { + /* There are stability issues reported on with + * bapm enabled when switching between AC and battery + * power. At the same time, some MSI boards hang + * if it's not enabled and dpm is enabled. Just enable + * it for MSI boards right now. + */ + if (rdev->pdev->subsystem_vendor == 0x1462) + pi->enable_bapm = true; + else + pi->enable_bapm = false; + } else if (radeon_bapm == 0) { pi->enable_bapm = false; + } else { + pi->enable_bapm = true; + } pi->enable_nbps_policy = true; pi->enable_sclk_ds = true; pi->enable_gfx_power_gating = true; diff --git a/drivers/gpu/drm/radeon/uvd_v1_0.c b/drivers/gpu/drm/radeon/uvd_v1_0.c index be42c812520..cda39134728 100644 --- a/drivers/gpu/drm/radeon/uvd_v1_0.c +++ b/drivers/gpu/drm/radeon/uvd_v1_0.c @@ -124,7 +124,7 @@ int uvd_v1_0_init(struct radeon_device *rdev) radeon_ring_write(ring, PACKET0(UVD_SEMA_CNTL, 0)); radeon_ring_write(ring, 3); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); done: /* lower clocks again */ @@ -331,7 +331,7 @@ int uvd_v1_0_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) } radeon_ring_write(ring, PACKET0(UVD_CONTEXT_ID, 0)); radeon_ring_write(ring, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev, ring); + radeon_ring_unlock_commit(rdev, ring, false); for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(UVD_CONTEXT_ID); if (tmp == 0xDEADBEEF) |