diff options
Diffstat (limited to 'drivers/gpu')
62 files changed, 584 insertions, 390 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 260165bbe373..b16cce7c22c3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -213,19 +213,35 @@ int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, spin_lock(&kfd_mem_limit.mem_limit_lock); if (kfd_mem_limit.system_mem_used + system_mem_needed > - kfd_mem_limit.max_system_mem_limit) + kfd_mem_limit.max_system_mem_limit) { pr_debug("Set no_system_mem_limit=1 if using shared memory\n"); + if (!no_system_mem_limit) { + ret = -ENOMEM; + goto release; + } + } - if ((kfd_mem_limit.system_mem_used + system_mem_needed > - kfd_mem_limit.max_system_mem_limit && !no_system_mem_limit) || - (kfd_mem_limit.ttm_mem_used + ttm_mem_needed > - kfd_mem_limit.max_ttm_mem_limit) || - (adev && xcp_id >= 0 && adev->kfd.vram_used[xcp_id] + vram_needed > - vram_size - reserved_for_pt - reserved_for_ras - atomic64_read(&adev->vram_pin_size))) { + if (kfd_mem_limit.ttm_mem_used + ttm_mem_needed > + kfd_mem_limit.max_ttm_mem_limit) { ret = -ENOMEM; goto release; } + /*if is_app_apu is false and apu_prefer_gtt is true, it is an APU with + * carve out < gtt. In that case, VRAM allocation will go to gtt domain, skip + * VRAM check since ttm_mem_limit check already cover this allocation + */ + + if (adev && xcp_id >= 0 && (!adev->apu_prefer_gtt || adev->gmc.is_app_apu)) { + uint64_t vram_available = + vram_size - reserved_for_pt - reserved_for_ras - + atomic64_read(&adev->vram_pin_size); + if (adev->kfd.vram_used[xcp_id] + vram_needed > vram_available) { + ret = -ENOMEM; + goto release; + } + } + /* Update memory accounting by decreasing available system * memory, TTM memory and GPU memory as computed above */ @@ -1626,11 +1642,15 @@ size_t amdgpu_amdkfd_get_available_memory(struct amdgpu_device *adev, uint64_t vram_available, system_mem_available, ttm_mem_available; spin_lock(&kfd_mem_limit.mem_limit_lock); - vram_available = KFD_XCP_MEMORY_SIZE(adev, xcp_id) - - adev->kfd.vram_used_aligned[xcp_id] - - atomic64_read(&adev->vram_pin_size) - - reserved_for_pt - - reserved_for_ras; + if (adev->apu_prefer_gtt && !adev->gmc.is_app_apu) + vram_available = KFD_XCP_MEMORY_SIZE(adev, xcp_id) + - adev->kfd.vram_used_aligned[xcp_id]; + else + vram_available = KFD_XCP_MEMORY_SIZE(adev, xcp_id) + - adev->kfd.vram_used_aligned[xcp_id] + - atomic64_read(&adev->vram_pin_size) + - reserved_for_pt + - reserved_for_ras; if (adev->apu_prefer_gtt) { system_mem_available = no_system_mem_limit ? diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index 23484317a5fa..693357caa9a8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -448,7 +448,7 @@ static int psp_sw_init(struct amdgpu_ip_block *ip_block) psp->cmd = kzalloc(sizeof(struct psp_gfx_cmd_resp), GFP_KERNEL); if (!psp->cmd) { dev_err(adev->dev, "Failed to allocate memory to command buffer!\n"); - ret = -ENOMEM; + return -ENOMEM; } adev->psp.xgmi_context.supports_extended_data = diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index 6379bb25bf5c..486c3646710c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -421,8 +421,6 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring) dma_fence_put(ring->vmid_wait); ring->vmid_wait = NULL; ring->me = 0; - - ring->adev->rings[ring->idx] = NULL; } /** diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c index bf7c22f81cda..ba73518f5cdf 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c @@ -1462,17 +1462,12 @@ static int dce_v10_0_audio_init(struct amdgpu_device *adev) static void dce_v10_0_audio_fini(struct amdgpu_device *adev) { - int i; - if (!amdgpu_audio) return; if (!adev->mode_info.audio.enabled) return; - for (i = 0; i < adev->mode_info.audio.num_pins; i++) - dce_v10_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false); - adev->mode_info.audio.enabled = false; } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c index 47e05783c4a0..b01d88d078fa 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c @@ -1511,17 +1511,12 @@ static int dce_v11_0_audio_init(struct amdgpu_device *adev) static void dce_v11_0_audio_fini(struct amdgpu_device *adev) { - int i; - if (!amdgpu_audio) return; if (!adev->mode_info.audio.enabled) return; - for (i = 0; i < adev->mode_info.audio.num_pins; i++) - dce_v11_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false); - adev->mode_info.audio.enabled = false; } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c index 276c025c4c03..81760a26f2ff 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c @@ -1451,17 +1451,12 @@ static int dce_v6_0_audio_init(struct amdgpu_device *adev) static void dce_v6_0_audio_fini(struct amdgpu_device *adev) { - int i; - if (!amdgpu_audio) return; if (!adev->mode_info.audio.enabled) return; - for (i = 0; i < adev->mode_info.audio.num_pins; i++) - dce_v6_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false); - adev->mode_info.audio.enabled = false; } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c index e62ccf9eb73d..19a265bd4d19 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c @@ -1443,17 +1443,12 @@ static int dce_v8_0_audio_init(struct amdgpu_device *adev) static void dce_v8_0_audio_fini(struct amdgpu_device *adev) { - int i; - if (!amdgpu_audio) return; if (!adev->mode_info.audio.enabled) return; - for (i = 0; i < adev->mode_info.audio.num_pins; i++) - dce_v8_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false); - adev->mode_info.audio.enabled = false; } diff --git a/drivers/gpu/drm/amd/amdgpu/isp_v4_1_1.c b/drivers/gpu/drm/amd/amdgpu/isp_v4_1_1.c index a887df520414..4258d3e0b706 100644 --- a/drivers/gpu/drm/amd/amdgpu/isp_v4_1_1.c +++ b/drivers/gpu/drm/amd/amdgpu/isp_v4_1_1.c @@ -29,6 +29,8 @@ #include "amdgpu.h" #include "isp_v4_1_1.h" +MODULE_FIRMWARE("amdgpu/isp_4_1_1.bin"); + #define ISP_PERFORMANCE_STATE_LOW 0 #define ISP_PERFORMANCE_STATE_HIGH 1 diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c index 28eb846280dd..3f6a828cad8a 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c @@ -641,8 +641,9 @@ static int mes_v11_0_misc_op(struct amdgpu_mes *mes, break; case MES_MISC_OP_CHANGE_CONFIG: if ((mes->adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) < 0x63) { - dev_err(mes->adev->dev, "MES FW version must be larger than 0x63 to support limit single process feature.\n"); - return -EINVAL; + dev_warn_once(mes->adev->dev, + "MES FW version must be larger than 0x63 to support limit single process feature.\n"); + return 0; } misc_pkt.opcode = MESAPI_MISC__CHANGE_CONFIG; misc_pkt.change_config.opcode = diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c index 6cc05d36e359..64b240b51f1a 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c @@ -149,12 +149,12 @@ static int psp_v11_0_wait_for_bootloader(struct psp_context *psp) int ret; int retry_loop; - for (retry_loop = 0; retry_loop < 10; retry_loop++) { + for (retry_loop = 0; retry_loop < 20; retry_loop++) { /* Wait for bootloader to signify that is ready having bit 31 of C2PMSG_35 set to 1 */ ret = psp_wait_for( psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35), - 0x80000000, 0x80000000, PSP_WAITREG_NOVERBOSE); + 0x80000000, 0x8000FFFF, PSP_WAITREG_NOVERBOSE); if (ret == 0) return 0; @@ -397,18 +397,6 @@ static int psp_v11_0_mode1_reset(struct psp_context *psp) msleep(500); - offset = SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_33); - - ret = psp_wait_for(psp, offset, MBOX_TOS_RESP_FLAG, MBOX_TOS_RESP_MASK, - 0); - - if (ret) { - DRM_INFO("psp mode 1 reset failed!\n"); - return -EINVAL; - } - - DRM_INFO("psp mode1 reset succeed \n"); - return 0; } @@ -665,7 +653,8 @@ static const struct psp_funcs psp_v11_0_funcs = { .ring_get_wptr = psp_v11_0_ring_get_wptr, .ring_set_wptr = psp_v11_0_ring_set_wptr, .load_usbc_pd_fw = psp_v11_0_load_usbc_pd_fw, - .read_usbc_pd_fw = psp_v11_0_read_usbc_pd_fw + .read_usbc_pd_fw = psp_v11_0_read_usbc_pd_fw, + .wait_for_bootloader = psp_v11_0_wait_for_bootloader }; void psp_v11_0_set_psp_funcs(struct psp_context *psp) diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c index e6d8eddda2bf..db6e41967f12 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v6_0.c @@ -1377,7 +1377,7 @@ static int sdma_v6_0_sw_init(struct amdgpu_ip_block *ip_block) switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) { case IP_VERSION(6, 0, 0): - if ((adev->sdma.instance[0].fw_version >= 24) && !adev->sdma.disable_uq) + if ((adev->sdma.instance[0].fw_version >= 27) && !adev->sdma.disable_uq) adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs; break; case IP_VERSION(6, 0, 1): @@ -1385,11 +1385,11 @@ static int sdma_v6_0_sw_init(struct amdgpu_ip_block *ip_block) adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs; break; case IP_VERSION(6, 0, 2): - if ((adev->sdma.instance[0].fw_version >= 21) && !adev->sdma.disable_uq) + if ((adev->sdma.instance[0].fw_version >= 23) && !adev->sdma.disable_uq) adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs; break; case IP_VERSION(6, 0, 3): - if ((adev->sdma.instance[0].fw_version >= 25) && !adev->sdma.disable_uq) + if ((adev->sdma.instance[0].fw_version >= 27) && !adev->sdma.disable_uq) adev->userq_funcs[AMDGPU_HW_IP_DMA] = &userq_mes_funcs; break; case IP_VERSION(6, 1, 0): diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index 4b8f4407047f..2811226b0ea5 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -1888,15 +1888,19 @@ static int vcn_v3_0_limit_sched(struct amdgpu_cs_parser *p, struct amdgpu_job *job) { struct drm_gpu_scheduler **scheds; - - /* The create msg must be in the first IB submitted */ - if (atomic_read(&job->base.entity->fence_seq)) - return -EINVAL; + struct dma_fence *fence; /* if VCN0 is harvested, we can't support AV1 */ if (p->adev->vcn.harvest_config & AMDGPU_VCN_HARVEST_VCN0) return -EINVAL; + /* wait for all jobs to finish before switching to instance 0 */ + fence = amdgpu_ctx_get_fence(p->ctx, job->base.entity, ~0ull); + if (fence) { + dma_fence_wait(fence, false); + dma_fence_put(fence); + } + scheds = p->adev->gpu_sched[AMDGPU_HW_IP_VCN_DEC] [AMDGPU_RING_PRIO_DEFAULT].sched; drm_sched_entity_modify_sched(job->base.entity, scheds, 1); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index 1924e075b66f..706f3b2f484f 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -1808,15 +1808,19 @@ static int vcn_v4_0_limit_sched(struct amdgpu_cs_parser *p, struct amdgpu_job *job) { struct drm_gpu_scheduler **scheds; - - /* The create msg must be in the first IB submitted */ - if (atomic_read(&job->base.entity->fence_seq)) - return -EINVAL; + struct dma_fence *fence; /* if VCN0 is harvested, we can't support AV1 */ if (p->adev->vcn.harvest_config & AMDGPU_VCN_HARVEST_VCN0) return -EINVAL; + /* wait for all jobs to finish before switching to instance 0 */ + fence = amdgpu_ctx_get_fence(p->ctx, job->base.entity, ~0ull); + if (fence) { + dma_fence_wait(fence, false); + dma_fence_put(fence); + } + scheds = p->adev->gpu_sched[AMDGPU_HW_IP_VCN_ENC] [AMDGPU_RING_PRIO_0].sched; drm_sched_entity_modify_sched(job->base.entity, scheds, 1); @@ -1907,22 +1911,16 @@ out: #define RADEON_VCN_ENGINE_TYPE_ENCODE (0x00000002) #define RADEON_VCN_ENGINE_TYPE_DECODE (0x00000003) - #define RADEON_VCN_ENGINE_INFO (0x30000001) -#define RADEON_VCN_ENGINE_INFO_MAX_OFFSET 16 - #define RENCODE_ENCODE_STANDARD_AV1 2 #define RENCODE_IB_PARAM_SESSION_INIT 0x00000003 -#define RENCODE_IB_PARAM_SESSION_INIT_MAX_OFFSET 64 -/* return the offset in ib if id is found, -1 otherwise - * to speed up the searching we only search upto max_offset - */ -static int vcn_v4_0_enc_find_ib_param(struct amdgpu_ib *ib, uint32_t id, int max_offset) +/* return the offset in ib if id is found, -1 otherwise */ +static int vcn_v4_0_enc_find_ib_param(struct amdgpu_ib *ib, uint32_t id, int start) { int i; - for (i = 0; i < ib->length_dw && i < max_offset && ib->ptr[i] >= 8; i += ib->ptr[i]/4) { + for (i = start; i < ib->length_dw && ib->ptr[i] >= 8; i += ib->ptr[i] / 4) { if (ib->ptr[i + 1] == id) return i; } @@ -1937,33 +1935,29 @@ static int vcn_v4_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p, struct amdgpu_vcn_decode_buffer *decode_buffer; uint64_t addr; uint32_t val; - int idx; + int idx = 0, sidx; /* The first instance can decode anything */ if (!ring->me) return 0; - /* RADEON_VCN_ENGINE_INFO is at the top of ib block */ - idx = vcn_v4_0_enc_find_ib_param(ib, RADEON_VCN_ENGINE_INFO, - RADEON_VCN_ENGINE_INFO_MAX_OFFSET); - if (idx < 0) /* engine info is missing */ - return 0; - - val = amdgpu_ib_get_value(ib, idx + 2); /* RADEON_VCN_ENGINE_TYPE */ - if (val == RADEON_VCN_ENGINE_TYPE_DECODE) { - decode_buffer = (struct amdgpu_vcn_decode_buffer *)&ib->ptr[idx + 6]; - - if (!(decode_buffer->valid_buf_flag & 0x1)) - return 0; - - addr = ((u64)decode_buffer->msg_buffer_address_hi) << 32 | - decode_buffer->msg_buffer_address_lo; - return vcn_v4_0_dec_msg(p, job, addr); - } else if (val == RADEON_VCN_ENGINE_TYPE_ENCODE) { - idx = vcn_v4_0_enc_find_ib_param(ib, RENCODE_IB_PARAM_SESSION_INIT, - RENCODE_IB_PARAM_SESSION_INIT_MAX_OFFSET); - if (idx >= 0 && ib->ptr[idx + 2] == RENCODE_ENCODE_STANDARD_AV1) - return vcn_v4_0_limit_sched(p, job); + while ((idx = vcn_v4_0_enc_find_ib_param(ib, RADEON_VCN_ENGINE_INFO, idx)) >= 0) { + val = amdgpu_ib_get_value(ib, idx + 2); /* RADEON_VCN_ENGINE_TYPE */ + if (val == RADEON_VCN_ENGINE_TYPE_DECODE) { + decode_buffer = (struct amdgpu_vcn_decode_buffer *)&ib->ptr[idx + 6]; + + if (!(decode_buffer->valid_buf_flag & 0x1)) + return 0; + + addr = ((u64)decode_buffer->msg_buffer_address_hi) << 32 | + decode_buffer->msg_buffer_address_lo; + return vcn_v4_0_dec_msg(p, job, addr); + } else if (val == RADEON_VCN_ENGINE_TYPE_ENCODE) { + sidx = vcn_v4_0_enc_find_ib_param(ib, RENCODE_IB_PARAM_SESSION_INIT, idx); + if (sidx >= 0 && ib->ptr[sidx + 2] == RENCODE_ENCODE_STANDARD_AV1) + return vcn_v4_0_limit_sched(p, job); + } + idx += ib->ptr[idx] / 4; } return 0; } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c index 4ec73f33535e..720b20e842ba 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c @@ -1587,7 +1587,8 @@ static int kfd_dev_create_p2p_links(void) break; if (!dev->gpu || !dev->gpu->adev || (dev->gpu->kfd->hive_id && - dev->gpu->kfd->hive_id == new_dev->gpu->kfd->hive_id)) + dev->gpu->kfd->hive_id == new_dev->gpu->kfd->hive_id && + amdgpu_xgmi_get_is_sharing_enabled(dev->gpu->adev, new_dev->gpu->adev))) goto next; /* check if node(s) is/are peer accessible in one direction or bi-direction */ diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index a0ca3b2c6bd8..4e86370ae705 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -2913,6 +2913,17 @@ static int dm_oem_i2c_hw_init(struct amdgpu_device *adev) return 0; } +static void dm_oem_i2c_hw_fini(struct amdgpu_device *adev) +{ + struct amdgpu_display_manager *dm = &adev->dm; + + if (dm->oem_i2c) { + i2c_del_adapter(&dm->oem_i2c->base); + kfree(dm->oem_i2c); + dm->oem_i2c = NULL; + } +} + /** * dm_hw_init() - Initialize DC device * @ip_block: Pointer to the amdgpu_ip_block for this hw instance. @@ -2963,7 +2974,7 @@ static int dm_hw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - kfree(adev->dm.oem_i2c); + dm_oem_i2c_hw_fini(adev); amdgpu_dm_hpd_fini(adev); @@ -3127,25 +3138,6 @@ static void dm_destroy_cached_state(struct amdgpu_device *adev) dm->cached_state = NULL; } -static void dm_complete(struct amdgpu_ip_block *ip_block) -{ - struct amdgpu_device *adev = ip_block->adev; - - dm_destroy_cached_state(adev); -} - -static int dm_prepare_suspend(struct amdgpu_ip_block *ip_block) -{ - struct amdgpu_device *adev = ip_block->adev; - - if (amdgpu_in_reset(adev)) - return 0; - - WARN_ON(adev->dm.cached_state); - - return dm_cache_state(adev); -} - static int dm_suspend(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; @@ -3571,10 +3563,8 @@ static const struct amd_ip_funcs amdgpu_dm_funcs = { .early_fini = amdgpu_dm_early_fini, .hw_init = dm_hw_init, .hw_fini = dm_hw_fini, - .prepare_suspend = dm_prepare_suspend, .suspend = dm_suspend, .resume = dm_resume, - .complete = dm_complete, .is_idle = dm_is_idle, .wait_for_idle = dm_wait_for_idle, .check_soft_reset = dm_check_soft_reset, @@ -8381,8 +8371,7 @@ static int amdgpu_dm_connector_get_modes(struct drm_connector *connector) drm_add_modes_noedid(connector, 1920, 1080); } else { amdgpu_dm_connector_ddc_get_modes(connector, drm_edid); - if (encoder && (connector->connector_type != DRM_MODE_CONNECTOR_eDP) && - (connector->connector_type != DRM_MODE_CONNECTOR_LVDS)) + if (encoder) amdgpu_dm_connector_add_common_modes(encoder, connector); amdgpu_dm_connector_add_freesync_modes(connector, drm_edid); } diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c index 7187d5aedf0a..77a9d2c7d318 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c @@ -809,6 +809,7 @@ void amdgpu_dm_initialize_dp_connector(struct amdgpu_display_manager *dm, drm_dp_aux_init(&aconnector->dm_dp_aux.aux); drm_dp_cec_register_connector(&aconnector->dm_dp_aux.aux, &aconnector->base); + drm_dp_dpcd_set_probe(&aconnector->dm_dp_aux.aux, false); if (aconnector->base.connector_type == DRM_MODE_CONNECTOR_eDP) return; diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index 59c07756130d..f24e1da68269 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -1145,6 +1145,7 @@ struct dc_debug_options { bool enable_hblank_borrow; bool force_subvp_df_throttle; uint32_t acpi_transition_bitmasks[MAX_PIPES]; + bool enable_pg_cntl_debug_logs; }; diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c index 58c84f555c0f..0ce9489ac6b7 100644 --- a/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c +++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c @@ -133,30 +133,34 @@ enum dsc_clk_source { }; -static void dccg35_set_dsc_clk_rcg(struct dccg *dccg, int inst, bool enable) +static void dccg35_set_dsc_clk_rcg(struct dccg *dccg, int inst, bool allow_rcg) { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dsc && enable) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dsc && allow_rcg) return; switch (inst) { case 0: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, enable ? 0 : 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1); break; case 1: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, enable ? 0 : 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1); break; case 2: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, enable ? 0 : 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1); break; case 3: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, enable ? 0 : 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1); break; default: BREAK_TO_DEBUGGER(); return; } + + /* Wait for clock to ramp */ + if (!allow_rcg) + udelay(10); } static void dccg35_set_symclk32_se_rcg( @@ -385,35 +389,34 @@ static void dccg35_set_dtbclk_p_rcg(struct dccg *dccg, int inst, bool enable) } } -static void dccg35_set_dppclk_rcg(struct dccg *dccg, - int inst, bool enable) +static void dccg35_set_dppclk_rcg(struct dccg *dccg, int inst, bool allow_rcg) { - struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp && enable) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp && allow_rcg) return; switch (inst) { case 0: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK0_ROOT_GATE_DISABLE, enable ? 0 : 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK0_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1); break; case 1: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK1_ROOT_GATE_DISABLE, enable ? 0 : 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK1_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1); break; case 2: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK2_ROOT_GATE_DISABLE, enable ? 0 : 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK2_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1); break; case 3: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK3_ROOT_GATE_DISABLE, enable ? 0 : 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK3_ROOT_GATE_DISABLE, allow_rcg ? 0 : 1); break; default: BREAK_TO_DEBUGGER(); break; } - //DC_LOG_DEBUG("%s: inst(%d) DPPCLK rcg_disable: %d\n", __func__, inst, enable ? 0 : 1); + /* Wait for clock to ramp */ + if (!allow_rcg) + udelay(10); } static void dccg35_set_dpstreamclk_rcg( @@ -1177,32 +1180,34 @@ static void dccg35_update_dpp_dto(struct dccg *dccg, int dpp_inst, } static void dccg35_set_dppclk_root_clock_gating(struct dccg *dccg, - uint32_t dpp_inst, uint32_t enable) + uint32_t dpp_inst, uint32_t disallow_rcg) { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp && !disallow_rcg) return; switch (dpp_inst) { case 0: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK0_ROOT_GATE_DISABLE, enable); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK0_ROOT_GATE_DISABLE, disallow_rcg); break; case 1: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK1_ROOT_GATE_DISABLE, enable); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK1_ROOT_GATE_DISABLE, disallow_rcg); break; case 2: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK2_ROOT_GATE_DISABLE, enable); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK2_ROOT_GATE_DISABLE, disallow_rcg); break; case 3: - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK3_ROOT_GATE_DISABLE, enable); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DPPCLK3_ROOT_GATE_DISABLE, disallow_rcg); break; default: break; } - //DC_LOG_DEBUG("%s: dpp_inst(%d) rcg: %d\n", __func__, dpp_inst, enable); + /* Wait for clock to ramp */ + if (disallow_rcg) + udelay(10); } static void dccg35_get_pixel_rate_div( @@ -1782,8 +1787,7 @@ static void dccg35_enable_dscclk(struct dccg *dccg, int inst) //Disable DTO switch (inst) { case 0: - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK0_ROOT_GATE_DISABLE, 1); REG_UPDATE_2(DSCCLK0_DTO_PARAM, DSCCLK0_DTO_PHASE, 0, @@ -1791,8 +1795,7 @@ static void dccg35_enable_dscclk(struct dccg *dccg, int inst) REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK0_EN, 1); break; case 1: - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK1_ROOT_GATE_DISABLE, 1); REG_UPDATE_2(DSCCLK1_DTO_PARAM, DSCCLK1_DTO_PHASE, 0, @@ -1800,8 +1803,7 @@ static void dccg35_enable_dscclk(struct dccg *dccg, int inst) REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK1_EN, 1); break; case 2: - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK2_ROOT_GATE_DISABLE, 1); REG_UPDATE_2(DSCCLK2_DTO_PARAM, DSCCLK2_DTO_PHASE, 0, @@ -1809,8 +1811,7 @@ static void dccg35_enable_dscclk(struct dccg *dccg, int inst) REG_UPDATE(DSCCLK_DTO_CTRL, DSCCLK2_EN, 1); break; case 3: - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) - REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, 1); + REG_UPDATE(DCCG_GATE_DISABLE_CNTL6, DSCCLK3_ROOT_GATE_DISABLE, 1); REG_UPDATE_2(DSCCLK3_DTO_PARAM, DSCCLK3_DTO_PHASE, 0, @@ -1821,6 +1822,9 @@ static void dccg35_enable_dscclk(struct dccg *dccg, int inst) BREAK_TO_DEBUGGER(); return; } + + /* Wait for clock to ramp */ + udelay(10); } static void dccg35_disable_dscclk(struct dccg *dccg, @@ -1864,6 +1868,9 @@ static void dccg35_disable_dscclk(struct dccg *dccg, default: return; } + + /* Wait for clock ramp */ + udelay(10); } static void dccg35_enable_symclk_se(struct dccg *dccg, uint32_t stream_enc_inst, uint32_t link_enc_inst) @@ -2349,10 +2356,7 @@ static void dccg35_disable_symclk_se_cb( void dccg35_root_gate_disable_control(struct dccg *dccg, uint32_t pipe_idx, uint32_t disable_clock_gating) { - - if (dccg->ctx->dc->debug.root_clock_optimization.bits.dpp) { - dccg35_set_dppclk_root_clock_gating(dccg, pipe_idx, disable_clock_gating); - } + dccg35_set_dppclk_root_clock_gating(dccg, pipe_idx, disable_clock_gating); } static const struct dccg_funcs dccg35_funcs_new = { diff --git a/drivers/gpu/drm/amd/display/dc/dpp/dcn10/dcn10_dpp.c b/drivers/gpu/drm/amd/display/dc/dpp/dcn10/dcn10_dpp.c index 75fb77bca83b..01480a04f85e 100644 --- a/drivers/gpu/drm/amd/display/dc/dpp/dcn10/dcn10_dpp.c +++ b/drivers/gpu/drm/amd/display/dc/dpp/dcn10/dcn10_dpp.c @@ -520,6 +520,15 @@ void dpp1_dppclk_control( REG_UPDATE(DPP_CONTROL, DPP_CLOCK_ENABLE, 0); } +void dpp_force_disable_cursor(struct dpp *dpp_base) +{ + struct dcn10_dpp *dpp = TO_DCN10_DPP(dpp_base); + + /* Force disable cursor */ + REG_UPDATE(CURSOR0_CONTROL, CUR0_ENABLE, 0); + dpp_base->pos.cur0_ctl.bits.cur0_enable = 0; +} + static const struct dpp_funcs dcn10_dpp_funcs = { .dpp_read_state = dpp_read_state, .dpp_reset = dpp_reset, diff --git a/drivers/gpu/drm/amd/display/dc/dpp/dcn10/dcn10_dpp.h b/drivers/gpu/drm/amd/display/dc/dpp/dcn10/dcn10_dpp.h index c48139bed11f..f466182963f7 100644 --- a/drivers/gpu/drm/amd/display/dc/dpp/dcn10/dcn10_dpp.h +++ b/drivers/gpu/drm/amd/display/dc/dpp/dcn10/dcn10_dpp.h @@ -1525,4 +1525,6 @@ void dpp1_construct(struct dcn10_dpp *dpp1, void dpp1_cm_get_gamut_remap(struct dpp *dpp_base, struct dpp_grph_csc_adjustment *adjust); +void dpp_force_disable_cursor(struct dpp *dpp_base); + #endif diff --git a/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c b/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c index 2d70586cef40..09be2a90cc79 100644 --- a/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c +++ b/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c @@ -1494,6 +1494,7 @@ static struct dpp_funcs dcn30_dpp_funcs = { .dpp_dppclk_control = dpp1_dppclk_control, .dpp_set_hdr_multiplier = dpp3_set_hdr_multiplier, .dpp_get_gamut_remap = dpp3_cm_get_gamut_remap, + .dpp_force_disable_cursor = dpp_force_disable_cursor, }; diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c index 3207addbd4eb..b7c2d3095b25 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c @@ -955,7 +955,7 @@ enum dc_status dcn20_enable_stream_timing( return DC_ERROR_UNEXPECTED; } - fsleep(stream->timing.v_total * (stream->timing.h_total * 10000u / stream->timing.pix_clk_100hz)); + udelay(stream->timing.v_total * (stream->timing.h_total * 10000u / stream->timing.pix_clk_100hz)); params.vertical_total_min = stream->adjust.v_total_min; params.vertical_total_max = stream->adjust.v_total_max; diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c index e68f21fd5f0f..560984533950 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c @@ -528,3 +528,75 @@ void dcn314_disable_link_output(struct dc_link *link, apply_symclk_on_tx_off_wa(link); } + +/** + * dcn314_dpp_pg_control - DPP power gate control. + * + * @hws: dce_hwseq reference. + * @dpp_inst: DPP instance reference. + * @power_on: true if we want to enable power gate, false otherwise. + * + * Enable or disable power gate in the specific DPP instance. + * If power gating is disabled, will force disable cursor in the DPP instance. + */ +void dcn314_dpp_pg_control( + struct dce_hwseq *hws, + unsigned int dpp_inst, + bool power_on) +{ + uint32_t power_gate = power_on ? 0 : 1; + uint32_t pwr_status = power_on ? 0 : 2; + + + if (hws->ctx->dc->debug.disable_dpp_power_gate) { + /* Workaround for DCN314 with disabled power gating */ + if (!power_on) { + + /* Force disable cursor if power gating is disabled */ + struct dpp *dpp = hws->ctx->dc->res_pool->dpps[dpp_inst]; + if (dpp && dpp->funcs->dpp_force_disable_cursor) + dpp->funcs->dpp_force_disable_cursor(dpp); + } + return; + } + if (REG(DOMAIN1_PG_CONFIG) == 0) + return; + + switch (dpp_inst) { + case 0: /* DPP0 */ + REG_UPDATE(DOMAIN1_PG_CONFIG, + DOMAIN1_POWER_GATE, power_gate); + + REG_WAIT(DOMAIN1_PG_STATUS, + DOMAIN1_PGFSM_PWR_STATUS, pwr_status, + 1, 1000); + break; + case 1: /* DPP1 */ + REG_UPDATE(DOMAIN3_PG_CONFIG, + DOMAIN3_POWER_GATE, power_gate); + + REG_WAIT(DOMAIN3_PG_STATUS, + DOMAIN3_PGFSM_PWR_STATUS, pwr_status, + 1, 1000); + break; + case 2: /* DPP2 */ + REG_UPDATE(DOMAIN5_PG_CONFIG, + DOMAIN5_POWER_GATE, power_gate); + + REG_WAIT(DOMAIN5_PG_STATUS, + DOMAIN5_PGFSM_PWR_STATUS, pwr_status, + 1, 1000); + break; + case 3: /* DPP3 */ + REG_UPDATE(DOMAIN7_PG_CONFIG, + DOMAIN7_POWER_GATE, power_gate); + + REG_WAIT(DOMAIN7_PG_STATUS, + DOMAIN7_PGFSM_PWR_STATUS, pwr_status, + 1, 1000); + break; + default: + BREAK_TO_DEBUGGER(); + break; + } +} diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.h b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.h index 2305ad282f21..6c072d0274ea 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.h +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.h @@ -47,4 +47,6 @@ void dcn314_dpp_root_clock_control(struct dce_hwseq *hws, unsigned int dpp_inst, void dcn314_disable_link_output(struct dc_link *link, const struct link_resource *link_res, enum signal_type signal); +void dcn314_dpp_pg_control(struct dce_hwseq *hws, unsigned int dpp_inst, bool power_on); + #endif /* __DC_HWSS_DCN314_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_init.c index f5112742edf9..9f454fa90e65 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_init.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_init.c @@ -141,6 +141,7 @@ static const struct hwseq_private_funcs dcn314_private_funcs = { .enable_power_gating_plane = dcn314_enable_power_gating_plane, .dpp_root_clock_control = dcn314_dpp_root_clock_control, .hubp_pg_control = dcn31_hubp_pg_control, + .dpp_pg_control = dcn314_dpp_pg_control, .program_all_writeback_pipes_in_tree = dcn30_program_all_writeback_pipes_in_tree, .update_odm = dcn314_update_odm, .dsc_pg_control = dcn314_dsc_pg_control, diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c index a267f574b619..764eff6a4ec6 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c @@ -113,6 +113,14 @@ static void enable_memory_low_power(struct dc *dc) } #endif +static void print_pg_status(struct dc *dc, const char *debug_func, const char *debug_log) +{ + if (dc->debug.enable_pg_cntl_debug_logs && dc->res_pool->pg_cntl) { + if (dc->res_pool->pg_cntl->funcs->print_pg_status) + dc->res_pool->pg_cntl->funcs->print_pg_status(dc->res_pool->pg_cntl, debug_func, debug_log); + } +} + void dcn35_set_dmu_fgcg(struct dce_hwseq *hws, bool enable) { REG_UPDATE_3(DMU_CLK_CNTL, @@ -137,6 +145,8 @@ void dcn35_init_hw(struct dc *dc) uint32_t user_level = MAX_BACKLIGHT_LEVEL; int i; + print_pg_status(dc, __func__, ": start"); + if (dc->clk_mgr && dc->clk_mgr->funcs->init_clocks) dc->clk_mgr->funcs->init_clocks(dc->clk_mgr); @@ -200,10 +210,7 @@ void dcn35_init_hw(struct dc *dc) /* we want to turn off all dp displays before doing detection */ dc->link_srv->blank_all_dp_displays(dc); -/* - if (hws->funcs.enable_power_gating_plane) - hws->funcs.enable_power_gating_plane(dc->hwseq, true); -*/ + if (res_pool->hubbub && res_pool->hubbub->funcs->dchubbub_init) res_pool->hubbub->funcs->dchubbub_init(dc->res_pool->hubbub); /* If taking control over from VBIOS, we may want to optimize our first @@ -236,6 +243,8 @@ void dcn35_init_hw(struct dc *dc) } hws->funcs.init_pipes(dc, dc->current_state); + print_pg_status(dc, __func__, ": after init_pipes"); + if (dc->res_pool->hubbub->funcs->allow_self_refresh_control && !dc->res_pool->hubbub->ctx->dc->debug.disable_stutter) dc->res_pool->hubbub->funcs->allow_self_refresh_control(dc->res_pool->hubbub, @@ -312,6 +321,7 @@ void dcn35_init_hw(struct dc *dc) if (dc->res_pool->pg_cntl->funcs->init_pg_status) dc->res_pool->pg_cntl->funcs->init_pg_status(dc->res_pool->pg_cntl); } + print_pg_status(dc, __func__, ": after init_pg_status"); } static void update_dsc_on_stream(struct pipe_ctx *pipe_ctx, bool enable) @@ -500,97 +510,6 @@ void dcn35_physymclk_root_clock_control(struct dce_hwseq *hws, unsigned int phy_ } } -void dcn35_dsc_pg_control( - struct dce_hwseq *hws, - unsigned int dsc_inst, - bool power_on) -{ - uint32_t power_gate = power_on ? 0 : 1; - uint32_t pwr_status = power_on ? 0 : 2; - uint32_t org_ip_request_cntl = 0; - - if (hws->ctx->dc->debug.disable_dsc_power_gate) - return; - if (hws->ctx->dc->debug.ignore_pg) - return; - REG_GET(DC_IP_REQUEST_CNTL, IP_REQUEST_EN, &org_ip_request_cntl); - if (org_ip_request_cntl == 0) - REG_SET(DC_IP_REQUEST_CNTL, 0, IP_REQUEST_EN, 1); - - switch (dsc_inst) { - case 0: /* DSC0 */ - REG_UPDATE(DOMAIN16_PG_CONFIG, - DOMAIN_POWER_GATE, power_gate); - - REG_WAIT(DOMAIN16_PG_STATUS, - DOMAIN_PGFSM_PWR_STATUS, pwr_status, - 1, 1000); - break; - case 1: /* DSC1 */ - REG_UPDATE(DOMAIN17_PG_CONFIG, - DOMAIN_POWER_GATE, power_gate); - - REG_WAIT(DOMAIN17_PG_STATUS, - DOMAIN_PGFSM_PWR_STATUS, pwr_status, - 1, 1000); - break; - case 2: /* DSC2 */ - REG_UPDATE(DOMAIN18_PG_CONFIG, - DOMAIN_POWER_GATE, power_gate); - - REG_WAIT(DOMAIN18_PG_STATUS, - DOMAIN_PGFSM_PWR_STATUS, pwr_status, - 1, 1000); - break; - case 3: /* DSC3 */ - REG_UPDATE(DOMAIN19_PG_CONFIG, - DOMAIN_POWER_GATE, power_gate); - - REG_WAIT(DOMAIN19_PG_STATUS, - DOMAIN_PGFSM_PWR_STATUS, pwr_status, - 1, 1000); - break; - default: - BREAK_TO_DEBUGGER(); - break; - } - - if (org_ip_request_cntl == 0) - REG_SET(DC_IP_REQUEST_CNTL, 0, IP_REQUEST_EN, 0); -} - -void dcn35_enable_power_gating_plane(struct dce_hwseq *hws, bool enable) -{ - bool force_on = true; /* disable power gating */ - uint32_t org_ip_request_cntl = 0; - - if (hws->ctx->dc->debug.disable_hubp_power_gate) - return; - if (hws->ctx->dc->debug.ignore_pg) - return; - REG_GET(DC_IP_REQUEST_CNTL, IP_REQUEST_EN, &org_ip_request_cntl); - if (org_ip_request_cntl == 0) - REG_SET(DC_IP_REQUEST_CNTL, 0, IP_REQUEST_EN, 1); - /* DCHUBP0/1/2/3/4/5 */ - REG_UPDATE(DOMAIN0_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on); - REG_UPDATE(DOMAIN2_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on); - /* DPP0/1/2/3/4/5 */ - REG_UPDATE(DOMAIN1_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on); - REG_UPDATE(DOMAIN3_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on); - - force_on = true; /* disable power gating */ - if (enable && !hws->ctx->dc->debug.disable_dsc_power_gate) - force_on = false; - - /* DCS0/1/2/3/4 */ - REG_UPDATE(DOMAIN16_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on); - REG_UPDATE(DOMAIN17_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on); - REG_UPDATE(DOMAIN18_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on); - REG_UPDATE(DOMAIN19_PG_CONFIG, DOMAIN_POWER_FORCEON, force_on); - - -} - /* In headless boot cases, DIG may be turned * on which causes HW/SW discrepancies. * To avoid this, power down hardware on boot @@ -1453,6 +1372,8 @@ void dcn35_prepare_bandwidth( } dcn20_prepare_bandwidth(dc, context); + + print_pg_status(dc, __func__, ": after rcg and power up"); } void dcn35_optimize_bandwidth( @@ -1461,6 +1382,8 @@ void dcn35_optimize_bandwidth( { struct pg_block_update pg_update_state; + print_pg_status(dc, __func__, ": before rcg and power up"); + dcn20_optimize_bandwidth(dc, context); if (dc->hwss.calc_blocks_to_gate) { @@ -1472,6 +1395,8 @@ void dcn35_optimize_bandwidth( if (dc->hwss.root_clock_control) dc->hwss.root_clock_control(dc, &pg_update_state, false); } + + print_pg_status(dc, __func__, ": after rcg and power up"); } void dcn35_set_drr(struct pipe_ctx **pipe_ctx, diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c index a3ccf805bd16..aefb7c473741 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c @@ -115,7 +115,6 @@ static const struct hw_sequencer_funcs dcn35_funcs = { .exit_optimized_pwr_state = dcn21_exit_optimized_pwr_state, .update_visual_confirm_color = dcn10_update_visual_confirm_color, .apply_idle_power_optimizations = dcn35_apply_idle_power_optimizations, - .update_dsc_pg = dcn32_update_dsc_pg, .calc_blocks_to_gate = dcn35_calc_blocks_to_gate, .calc_blocks_to_ungate = dcn35_calc_blocks_to_ungate, .hw_block_power_up = dcn35_hw_block_power_up, @@ -150,7 +149,6 @@ static const struct hwseq_private_funcs dcn35_private_funcs = { .plane_atomic_disable = dcn35_plane_atomic_disable, //.plane_atomic_disable = dcn20_plane_atomic_disable,/*todo*/ //.hubp_pg_control = dcn35_hubp_pg_control, - .enable_power_gating_plane = dcn35_enable_power_gating_plane, .dpp_root_clock_control = dcn35_dpp_root_clock_control, .dpstream_root_clock_control = dcn35_dpstream_root_clock_control, .physymclk_root_clock_control = dcn35_physymclk_root_clock_control, @@ -165,7 +163,6 @@ static const struct hwseq_private_funcs dcn35_private_funcs = { .calculate_dccg_k1_k2_values = dcn32_calculate_dccg_k1_k2_values, .resync_fifo_dccg_dio = dcn314_resync_fifo_dccg_dio, .is_dp_dig_pixel_rate_div_policy = dcn35_is_dp_dig_pixel_rate_div_policy, - .dsc_pg_control = dcn35_dsc_pg_control, .dsc_pg_status = dcn32_dsc_pg_status, .enable_plane = dcn35_enable_plane, .wait_for_pipe_update_if_needed = dcn10_wait_for_pipe_update_if_needed, diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c index 58f2be2a326b..a580a55695c3 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c @@ -114,7 +114,6 @@ static const struct hw_sequencer_funcs dcn351_funcs = { .exit_optimized_pwr_state = dcn21_exit_optimized_pwr_state, .update_visual_confirm_color = dcn10_update_visual_confirm_color, .apply_idle_power_optimizations = dcn35_apply_idle_power_optimizations, - .update_dsc_pg = dcn32_update_dsc_pg, .calc_blocks_to_gate = dcn351_calc_blocks_to_gate, .calc_blocks_to_ungate = dcn351_calc_blocks_to_ungate, .hw_block_power_up = dcn351_hw_block_power_up, @@ -145,7 +144,6 @@ static const struct hwseq_private_funcs dcn351_private_funcs = { .plane_atomic_disable = dcn35_plane_atomic_disable, //.plane_atomic_disable = dcn20_plane_atomic_disable,/*todo*/ //.hubp_pg_control = dcn35_hubp_pg_control, - .enable_power_gating_plane = dcn35_enable_power_gating_plane, .dpp_root_clock_control = dcn35_dpp_root_clock_control, .dpstream_root_clock_control = dcn35_dpstream_root_clock_control, .physymclk_root_clock_control = dcn35_physymclk_root_clock_control, @@ -159,7 +157,6 @@ static const struct hwseq_private_funcs dcn351_private_funcs = { .setup_hpo_hw_control = dcn35_setup_hpo_hw_control, .calculate_dccg_k1_k2_values = dcn32_calculate_dccg_k1_k2_values, .is_dp_dig_pixel_rate_div_policy = dcn35_is_dp_dig_pixel_rate_div_policy, - .dsc_pg_control = dcn35_dsc_pg_control, .dsc_pg_status = dcn32_dsc_pg_status, .enable_plane = dcn35_enable_plane, .wait_for_pipe_update_if_needed = dcn10_wait_for_pipe_update_if_needed, diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h b/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h index 0c5675d1c593..1b7c085dc2cc 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h @@ -349,6 +349,9 @@ struct dpp_funcs { struct dpp *dpp_base, enum dc_color_space color_space, struct dc_csc_transform cursor_csc_color_matrix); + + void (*dpp_force_disable_cursor)(struct dpp *dpp_base); + }; diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/pg_cntl.h b/drivers/gpu/drm/amd/display/dc/inc/hw/pg_cntl.h index 44f86cc2d1d6..227e3f8d7e5f 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/pg_cntl.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/pg_cntl.h @@ -49,6 +49,7 @@ struct pg_cntl_funcs { void (*mem_pg_control)(struct pg_cntl *pg_cntl, bool power_on); void (*dio_pg_control)(struct pg_cntl *pg_cntl, bool power_on); void (*init_pg_status)(struct pg_cntl *pg_cntl); + void (*print_pg_status)(struct pg_cntl *pg_cntl, const char *debug_func, const char *debug_log); }; #endif //__DC_PG_CNTL_H__ diff --git a/drivers/gpu/drm/amd/display/dc/pg/dcn35/dcn35_pg_cntl.c b/drivers/gpu/drm/amd/display/dc/pg/dcn35/dcn35_pg_cntl.c index af21c0a27f86..72bd43f9bbe2 100644 --- a/drivers/gpu/drm/amd/display/dc/pg/dcn35/dcn35_pg_cntl.c +++ b/drivers/gpu/drm/amd/display/dc/pg/dcn35/dcn35_pg_cntl.c @@ -79,16 +79,12 @@ void pg_cntl35_dsc_pg_control(struct pg_cntl *pg_cntl, unsigned int dsc_inst, bo uint32_t power_gate = power_on ? 0 : 1; uint32_t pwr_status = power_on ? 0 : 2; uint32_t org_ip_request_cntl = 0; - bool block_enabled; - - /*need to enable dscclk regardless DSC_PG*/ - if (pg_cntl->ctx->dc->res_pool->dccg->funcs->enable_dsc && power_on) - pg_cntl->ctx->dc->res_pool->dccg->funcs->enable_dsc( - pg_cntl->ctx->dc->res_pool->dccg, dsc_inst); + bool block_enabled = false; + bool skip_pg = pg_cntl->ctx->dc->debug.ignore_pg || + pg_cntl->ctx->dc->debug.disable_dsc_power_gate || + pg_cntl->ctx->dc->idle_optimizations_allowed; - if (pg_cntl->ctx->dc->debug.ignore_pg || - pg_cntl->ctx->dc->debug.disable_dsc_power_gate || - pg_cntl->ctx->dc->idle_optimizations_allowed) + if (skip_pg && !power_on) return; block_enabled = pg_cntl35_dsc_pg_status(pg_cntl, dsc_inst); @@ -111,7 +107,7 @@ void pg_cntl35_dsc_pg_control(struct pg_cntl *pg_cntl, unsigned int dsc_inst, bo REG_WAIT(DOMAIN16_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, - 1, 1000); + 1, 10000); break; case 1: /* DSC1 */ REG_UPDATE(DOMAIN17_PG_CONFIG, @@ -119,7 +115,7 @@ void pg_cntl35_dsc_pg_control(struct pg_cntl *pg_cntl, unsigned int dsc_inst, bo REG_WAIT(DOMAIN17_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, - 1, 1000); + 1, 10000); break; case 2: /* DSC2 */ REG_UPDATE(DOMAIN18_PG_CONFIG, @@ -127,7 +123,7 @@ void pg_cntl35_dsc_pg_control(struct pg_cntl *pg_cntl, unsigned int dsc_inst, bo REG_WAIT(DOMAIN18_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, - 1, 1000); + 1, 10000); break; case 3: /* DSC3 */ REG_UPDATE(DOMAIN19_PG_CONFIG, @@ -135,7 +131,7 @@ void pg_cntl35_dsc_pg_control(struct pg_cntl *pg_cntl, unsigned int dsc_inst, bo REG_WAIT(DOMAIN19_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, - 1, 1000); + 1, 10000); break; default: BREAK_TO_DEBUGGER(); @@ -144,12 +140,6 @@ void pg_cntl35_dsc_pg_control(struct pg_cntl *pg_cntl, unsigned int dsc_inst, bo if (dsc_inst < MAX_PIPES) pg_cntl->pg_pipe_res_enable[PG_DSC][dsc_inst] = power_on; - - if (pg_cntl->ctx->dc->res_pool->dccg->funcs->disable_dsc && !power_on) { - /*this is to disable dscclk*/ - pg_cntl->ctx->dc->res_pool->dccg->funcs->disable_dsc( - pg_cntl->ctx->dc->res_pool->dccg, dsc_inst); - } } static bool pg_cntl35_hubp_dpp_pg_status(struct pg_cntl *pg_cntl, unsigned int hubp_dpp_inst) @@ -189,11 +179,12 @@ void pg_cntl35_hubp_dpp_pg_control(struct pg_cntl *pg_cntl, unsigned int hubp_dp uint32_t pwr_status = power_on ? 0 : 2; uint32_t org_ip_request_cntl; bool block_enabled; + bool skip_pg = pg_cntl->ctx->dc->debug.ignore_pg || + pg_cntl->ctx->dc->debug.disable_hubp_power_gate || + pg_cntl->ctx->dc->debug.disable_dpp_power_gate || + pg_cntl->ctx->dc->idle_optimizations_allowed; - if (pg_cntl->ctx->dc->debug.ignore_pg || - pg_cntl->ctx->dc->debug.disable_hubp_power_gate || - pg_cntl->ctx->dc->debug.disable_dpp_power_gate || - pg_cntl->ctx->dc->idle_optimizations_allowed) + if (skip_pg && !power_on) return; block_enabled = pg_cntl35_hubp_dpp_pg_status(pg_cntl, hubp_dpp_inst); @@ -213,22 +204,22 @@ void pg_cntl35_hubp_dpp_pg_control(struct pg_cntl *pg_cntl, unsigned int hubp_dp case 0: /* DPP0 & HUBP0 */ REG_UPDATE(DOMAIN0_PG_CONFIG, DOMAIN_POWER_GATE, power_gate); - REG_WAIT(DOMAIN0_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 1000); + REG_WAIT(DOMAIN0_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 10000); break; case 1: /* DPP1 & HUBP1 */ REG_UPDATE(DOMAIN1_PG_CONFIG, DOMAIN_POWER_GATE, power_gate); - REG_WAIT(DOMAIN1_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 1000); + REG_WAIT(DOMAIN1_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 10000); break; case 2: /* DPP2 & HUBP2 */ REG_UPDATE(DOMAIN2_PG_CONFIG, DOMAIN_POWER_GATE, power_gate); - REG_WAIT(DOMAIN2_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 1000); + REG_WAIT(DOMAIN2_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 10000); break; case 3: /* DPP3 & HUBP3 */ REG_UPDATE(DOMAIN3_PG_CONFIG, DOMAIN_POWER_GATE, power_gate); - REG_WAIT(DOMAIN3_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 1000); + REG_WAIT(DOMAIN3_PG_STATUS, DOMAIN_PGFSM_PWR_STATUS, pwr_status, 1, 10000); break; default: BREAK_TO_DEBUGGER(); @@ -501,6 +492,36 @@ void pg_cntl35_init_pg_status(struct pg_cntl *pg_cntl) pg_cntl->pg_res_enable[PG_DWB] = block_enabled; } +static void pg_cntl35_print_pg_status(struct pg_cntl *pg_cntl, const char *debug_func, const char *debug_log) +{ + int i = 0; + bool block_enabled = false; + + DC_LOG_DEBUG("%s: %s", debug_func, debug_log); + + DC_LOG_DEBUG("PG_CNTL status:\n"); + + block_enabled = pg_cntl35_io_clk_status(pg_cntl); + DC_LOG_DEBUG("ONO0=%d (DCCG, DIO, DCIO)\n", block_enabled ? 1 : 0); + + block_enabled = pg_cntl35_mem_status(pg_cntl); + DC_LOG_DEBUG("ONO1=%d (DCHUBBUB, DCHVM, DCHUBBUBMEM)\n", block_enabled ? 1 : 0); + + block_enabled = pg_cntl35_plane_otg_status(pg_cntl); + DC_LOG_DEBUG("ONO2=%d (MPC, OPP, OPTC, DWB)\n", block_enabled ? 1 : 0); + + block_enabled = pg_cntl35_hpo_pg_status(pg_cntl); + DC_LOG_DEBUG("ONO3=%d (HPO)\n", block_enabled ? 1 : 0); + + for (i = 0; i < pg_cntl->ctx->dc->res_pool->pipe_count; i++) { + block_enabled = pg_cntl35_hubp_dpp_pg_status(pg_cntl, i); + DC_LOG_DEBUG("ONO%d=%d (DCHUBP%d, DPP%d)\n", 4 + i * 2, block_enabled ? 1 : 0, i, i); + + block_enabled = pg_cntl35_dsc_pg_status(pg_cntl, i); + DC_LOG_DEBUG("ONO%d=%d (DSC%d)\n", 5 + i * 2, block_enabled ? 1 : 0, i); + } +} + static const struct pg_cntl_funcs pg_cntl35_funcs = { .init_pg_status = pg_cntl35_init_pg_status, .dsc_pg_control = pg_cntl35_dsc_pg_control, @@ -511,7 +532,8 @@ static const struct pg_cntl_funcs pg_cntl35_funcs = { .mpcc_pg_control = pg_cntl35_mpcc_pg_control, .opp_pg_control = pg_cntl35_opp_pg_control, .optc_pg_control = pg_cntl35_optc_pg_control, - .dwb_pg_control = pg_cntl35_dwb_pg_control + .dwb_pg_control = pg_cntl35_dwb_pg_control, + .print_pg_status = pg_cntl35_print_pg_status }; struct pg_cntl *pg_cntl35_create( diff --git a/drivers/gpu/drm/bridge/ti-sn65dsi86.c b/drivers/gpu/drm/bridge/ti-sn65dsi86.c index 464390372b34..ae0d08e5e960 100644 --- a/drivers/gpu/drm/bridge/ti-sn65dsi86.c +++ b/drivers/gpu/drm/bridge/ti-sn65dsi86.c @@ -393,6 +393,17 @@ static int __maybe_unused ti_sn65dsi86_resume(struct device *dev) gpiod_set_value_cansleep(pdata->enable_gpio, 1); /* + * After EN is deasserted and an external clock is detected, the bridge + * will sample GPIO3:1 to determine its frequency. The driver will + * overwrite this setting in ti_sn_bridge_set_refclk_freq(). But this is + * racy. Thus we have to wait a couple of us. According to the datasheet + * the GPIO lines has to be stable at least 5 us (td5) but it seems that + * is not enough and the refclk frequency value is still lost or + * overwritten by the bridge itself. Waiting for 20us seems to work. + */ + usleep_range(20, 30); + + /* * If we have a reference clock we can enable communication w/ the * panel (including the aux channel) w/out any need for an input clock * so we can do it in resume which lets us read the EDID before diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c index 273054c22325..c92f3e736228 100644 --- a/drivers/gpu/drm/i915/display/intel_display_power.c +++ b/drivers/gpu/drm/i915/display/intel_display_power.c @@ -1172,7 +1172,7 @@ static void icl_mbus_init(struct intel_display *display) if (DISPLAY_VER(display) == 12) abox_regs |= BIT(0); - for_each_set_bit(i, &abox_regs, sizeof(abox_regs)) + for_each_set_bit(i, &abox_regs, BITS_PER_TYPE(abox_regs)) intel_de_rmw(display, MBUS_ABOX_CTL(i), mask, val); } @@ -1629,11 +1629,11 @@ static void tgl_bw_buddy_init(struct intel_display *display) if (table[config].page_mask == 0) { drm_dbg_kms(display->drm, "Unknown memory configuration; disabling address buddy logic.\n"); - for_each_set_bit(i, &abox_mask, sizeof(abox_mask)) + for_each_set_bit(i, &abox_mask, BITS_PER_TYPE(abox_mask)) intel_de_write(display, BW_BUDDY_CTL(i), BW_BUDDY_DISABLE); } else { - for_each_set_bit(i, &abox_mask, sizeof(abox_mask)) { + for_each_set_bit(i, &abox_mask, BITS_PER_TYPE(abox_mask)) { intel_de_write(display, BW_BUDDY_PAGE_MASK(i), table[config].page_mask); diff --git a/drivers/gpu/drm/mediatek/mtk_drm_drv.c b/drivers/gpu/drm/mediatek/mtk_drm_drv.c index f8a817689e16..eb5537f0ac90 100644 --- a/drivers/gpu/drm/mediatek/mtk_drm_drv.c +++ b/drivers/gpu/drm/mediatek/mtk_drm_drv.c @@ -387,19 +387,21 @@ static bool mtk_drm_get_all_drm_priv(struct device *dev) of_id = of_match_node(mtk_drm_of_ids, node); if (!of_id) - goto next_put_node; + continue; pdev = of_find_device_by_node(node); if (!pdev) - goto next_put_node; + continue; drm_dev = device_find_child(&pdev->dev, NULL, mtk_drm_match); + put_device(&pdev->dev); if (!drm_dev) - goto next_put_device_pdev_dev; + continue; temp_drm_priv = dev_get_drvdata(drm_dev); + put_device(drm_dev); if (!temp_drm_priv) - goto next_put_device_drm_dev; + continue; if (temp_drm_priv->data->main_len) all_drm_priv[CRTC_MAIN] = temp_drm_priv; @@ -411,17 +413,10 @@ static bool mtk_drm_get_all_drm_priv(struct device *dev) if (temp_drm_priv->mtk_drm_bound) cnt++; -next_put_device_drm_dev: - put_device(drm_dev); - -next_put_device_pdev_dev: - put_device(&pdev->dev); - -next_put_node: - of_node_put(node); - - if (cnt == MAX_CRTC) + if (cnt == MAX_CRTC) { + of_node_put(node); break; + } } if (drm_priv->data->mmsys_dev_num == cnt) { diff --git a/drivers/gpu/drm/nouveau/gv100_fence.c b/drivers/gpu/drm/nouveau/gv100_fence.c index cccdeca72002..317e516c4ec7 100644 --- a/drivers/gpu/drm/nouveau/gv100_fence.c +++ b/drivers/gpu/drm/nouveau/gv100_fence.c @@ -18,7 +18,7 @@ gv100_fence_emit32(struct nouveau_channel *chan, u64 virtual, u32 sequence) struct nvif_push *push = &chan->chan.push; int ret; - ret = PUSH_WAIT(push, 8); + ret = PUSH_WAIT(push, 13); if (ret) return ret; @@ -32,6 +32,11 @@ gv100_fence_emit32(struct nouveau_channel *chan, u64 virtual, u32 sequence) NVDEF(NVC36F, SEM_EXECUTE, PAYLOAD_SIZE, 32BIT) | NVDEF(NVC36F, SEM_EXECUTE, RELEASE_TIMESTAMP, DIS)); + PUSH_MTHD(push, NVC36F, MEM_OP_A, 0, + MEM_OP_B, 0, + MEM_OP_C, NVDEF(NVC36F, MEM_OP_C, MEMBAR_TYPE, SYS_MEMBAR), + MEM_OP_D, NVDEF(NVC36F, MEM_OP_D, OPERATION, MEMBAR)); + PUSH_MTHD(push, NVC36F, NON_STALL_INTERRUPT, 0); PUSH_KICK(push); diff --git a/drivers/gpu/drm/nouveau/include/nvhw/class/clc36f.h b/drivers/gpu/drm/nouveau/include/nvhw/class/clc36f.h index 8735dda4c8a7..338f74b9f501 100644 --- a/drivers/gpu/drm/nouveau/include/nvhw/class/clc36f.h +++ b/drivers/gpu/drm/nouveau/include/nvhw/class/clc36f.h @@ -7,6 +7,91 @@ #define NVC36F_NON_STALL_INTERRUPT (0x00000020) #define NVC36F_NON_STALL_INTERRUPT_HANDLE 31:0 +// NOTE - MEM_OP_A and MEM_OP_B have been replaced in gp100 with methods for +// specifying the page address for a targeted TLB invalidate and the uTLB for +// a targeted REPLAY_CANCEL for UVM. +// The previous MEM_OP_A/B functionality is in MEM_OP_C/D, with slightly +// rearranged fields. +#define NVC36F_MEM_OP_A (0x00000028) +#define NVC36F_MEM_OP_A_TLB_INVALIDATE_CANCEL_TARGET_CLIENT_UNIT_ID 5:0 // only relevant for REPLAY_CANCEL_TARGETED +#define NVC36F_MEM_OP_A_TLB_INVALIDATE_INVALIDATION_SIZE 5:0 // Used to specify size of invalidate, used for invalidates which are not of the REPLAY_CANCEL_TARGETED type +#define NVC36F_MEM_OP_A_TLB_INVALIDATE_CANCEL_TARGET_GPC_ID 10:6 // only relevant for REPLAY_CANCEL_TARGETED +#define NVC36F_MEM_OP_A_TLB_INVALIDATE_CANCEL_MMU_ENGINE_ID 6:0 // only relevant for REPLAY_CANCEL_VA_GLOBAL +#define NVC36F_MEM_OP_A_TLB_INVALIDATE_SYSMEMBAR 11:11 +#define NVC36F_MEM_OP_A_TLB_INVALIDATE_SYSMEMBAR_EN 0x00000001 +#define NVC36F_MEM_OP_A_TLB_INVALIDATE_SYSMEMBAR_DIS 0x00000000 +#define NVC36F_MEM_OP_A_TLB_INVALIDATE_TARGET_ADDR_LO 31:12 +#define NVC36F_MEM_OP_B (0x0000002c) +#define NVC36F_MEM_OP_B_TLB_INVALIDATE_TARGET_ADDR_HI 31:0 +#define NVC36F_MEM_OP_C (0x00000030) +#define NVC36F_MEM_OP_C_MEMBAR_TYPE 2:0 +#define NVC36F_MEM_OP_C_MEMBAR_TYPE_SYS_MEMBAR 0x00000000 +#define NVC36F_MEM_OP_C_MEMBAR_TYPE_MEMBAR 0x00000001 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PDB 0:0 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PDB_ONE 0x00000000 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PDB_ALL 0x00000001 // Probably nonsensical for MMU_TLB_INVALIDATE_TARGETED +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_GPC 1:1 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_GPC_ENABLE 0x00000000 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_GPC_DISABLE 0x00000001 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_REPLAY 4:2 // only relevant if GPC ENABLE +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_REPLAY_NONE 0x00000000 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_REPLAY_START 0x00000001 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_REPLAY_START_ACK_ALL 0x00000002 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_REPLAY_CANCEL_TARGETED 0x00000003 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_REPLAY_CANCEL_GLOBAL 0x00000004 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_REPLAY_CANCEL_VA_GLOBAL 0x00000005 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACK_TYPE 6:5 // only relevant if GPC ENABLE +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACK_TYPE_NONE 0x00000000 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACK_TYPE_GLOBALLY 0x00000001 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACK_TYPE_INTRANODE 0x00000002 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACCESS_TYPE 9:7 //only relevant for REPLAY_CANCEL_VA_GLOBAL +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACCESS_TYPE_VIRT_READ 0 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACCESS_TYPE_VIRT_WRITE 1 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACCESS_TYPE_VIRT_ATOMIC_STRONG 2 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACCESS_TYPE_VIRT_RSVRVD 3 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACCESS_TYPE_VIRT_ATOMIC_WEAK 4 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACCESS_TYPE_VIRT_ATOMIC_ALL 5 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACCESS_TYPE_VIRT_WRITE_AND_ATOMIC 6 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_ACCESS_TYPE_VIRT_ALL 7 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PAGE_TABLE_LEVEL 9:7 // Invalidate affects this level and all below +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PAGE_TABLE_LEVEL_ALL 0x00000000 // Invalidate tlb caches at all levels of the page table +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PAGE_TABLE_LEVEL_PTE_ONLY 0x00000001 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PAGE_TABLE_LEVEL_UP_TO_PDE0 0x00000002 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PAGE_TABLE_LEVEL_UP_TO_PDE1 0x00000003 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PAGE_TABLE_LEVEL_UP_TO_PDE2 0x00000004 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PAGE_TABLE_LEVEL_UP_TO_PDE3 0x00000005 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PAGE_TABLE_LEVEL_UP_TO_PDE4 0x00000006 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PAGE_TABLE_LEVEL_UP_TO_PDE5 0x00000007 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PDB_APERTURE 11:10 // only relevant if PDB_ONE +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PDB_APERTURE_VID_MEM 0x00000000 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PDB_APERTURE_SYS_MEM_COHERENT 0x00000002 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PDB_APERTURE_SYS_MEM_NONCOHERENT 0x00000003 +#define NVC36F_MEM_OP_C_TLB_INVALIDATE_PDB_ADDR_LO 31:12 // only relevant if PDB_ONE +#define NVC36F_MEM_OP_C_ACCESS_COUNTER_CLR_TARGETED_NOTIFY_TAG 19:0 +// MEM_OP_D MUST be preceded by MEM_OPs A-C. +#define NVC36F_MEM_OP_D (0x00000034) +#define NVC36F_MEM_OP_D_TLB_INVALIDATE_PDB_ADDR_HI 26:0 // only relevant if PDB_ONE +#define NVC36F_MEM_OP_D_OPERATION 31:27 +#define NVC36F_MEM_OP_D_OPERATION_MEMBAR 0x00000005 +#define NVC36F_MEM_OP_D_OPERATION_MMU_TLB_INVALIDATE 0x00000009 +#define NVC36F_MEM_OP_D_OPERATION_MMU_TLB_INVALIDATE_TARGETED 0x0000000a +#define NVC36F_MEM_OP_D_OPERATION_L2_PEERMEM_INVALIDATE 0x0000000d +#define NVC36F_MEM_OP_D_OPERATION_L2_SYSMEM_INVALIDATE 0x0000000e +// CLEAN_LINES is an alias for Tegra/GPU IP usage +#define NVC36F_MEM_OP_B_OPERATION_L2_INVALIDATE_CLEAN_LINES 0x0000000e +#define NVC36F_MEM_OP_D_OPERATION_L2_CLEAN_COMPTAGS 0x0000000f +#define NVC36F_MEM_OP_D_OPERATION_L2_FLUSH_DIRTY 0x00000010 +#define NVC36F_MEM_OP_D_OPERATION_L2_WAIT_FOR_SYS_PENDING_READS 0x00000015 +#define NVC36F_MEM_OP_D_OPERATION_ACCESS_COUNTER_CLR 0x00000016 +#define NVC36F_MEM_OP_D_ACCESS_COUNTER_CLR_TYPE 1:0 +#define NVC36F_MEM_OP_D_ACCESS_COUNTER_CLR_TYPE_MIMC 0x00000000 +#define NVC36F_MEM_OP_D_ACCESS_COUNTER_CLR_TYPE_MOMC 0x00000001 +#define NVC36F_MEM_OP_D_ACCESS_COUNTER_CLR_TYPE_ALL 0x00000002 +#define NVC36F_MEM_OP_D_ACCESS_COUNTER_CLR_TYPE_TARGETED 0x00000003 +#define NVC36F_MEM_OP_D_ACCESS_COUNTER_CLR_TARGETED_TYPE 2:2 +#define NVC36F_MEM_OP_D_ACCESS_COUNTER_CLR_TARGETED_TYPE_MIMC 0x00000000 +#define NVC36F_MEM_OP_D_ACCESS_COUNTER_CLR_TARGETED_TYPE_MOMC 0x00000001 +#define NVC36F_MEM_OP_D_ACCESS_COUNTER_CLR_TARGETED_BANK 6:3 #define NVC36F_SEM_ADDR_LO (0x0000005c) #define NVC36F_SEM_ADDR_LO_OFFSET 31:2 #define NVC36F_SEM_ADDR_HI (0x00000060) diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c index 9f345a008717..869d4335c0f4 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fence.c +++ b/drivers/gpu/drm/nouveau/nouveau_fence.c @@ -240,21 +240,6 @@ nouveau_fence_emit(struct nouveau_fence *fence) return ret; } -void -nouveau_fence_cancel(struct nouveau_fence *fence) -{ - struct nouveau_fence_chan *fctx = nouveau_fctx(fence); - unsigned long flags; - - spin_lock_irqsave(&fctx->lock, flags); - if (!dma_fence_is_signaled_locked(&fence->base)) { - dma_fence_set_error(&fence->base, -ECANCELED); - if (nouveau_fence_signal(fence)) - nvif_event_block(&fctx->event); - } - spin_unlock_irqrestore(&fctx->lock, flags); -} - bool nouveau_fence_done(struct nouveau_fence *fence) { diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h index 9957a919bd38..183dd43ecfff 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fence.h +++ b/drivers/gpu/drm/nouveau/nouveau_fence.h @@ -29,7 +29,6 @@ void nouveau_fence_unref(struct nouveau_fence **); int nouveau_fence_emit(struct nouveau_fence *); bool nouveau_fence_done(struct nouveau_fence *); -void nouveau_fence_cancel(struct nouveau_fence *fence); int nouveau_fence_wait(struct nouveau_fence *, bool lazy, bool intr); int nouveau_fence_sync(struct nouveau_bo *, struct nouveau_channel *, bool exclusive, bool intr); diff --git a/drivers/gpu/drm/nouveau/nouveau_sched.c b/drivers/gpu/drm/nouveau/nouveau_sched.c index 0cc0bc9f9952..e60f7892f5ce 100644 --- a/drivers/gpu/drm/nouveau/nouveau_sched.c +++ b/drivers/gpu/drm/nouveau/nouveau_sched.c @@ -11,7 +11,6 @@ #include "nouveau_exec.h" #include "nouveau_abi16.h" #include "nouveau_sched.h" -#include "nouveau_chan.h" #define NOUVEAU_SCHED_JOB_TIMEOUT_MS 10000 @@ -122,9 +121,11 @@ nouveau_job_done(struct nouveau_job *job) { struct nouveau_sched *sched = job->sched; - spin_lock(&sched->job_list.lock); + spin_lock(&sched->job.list.lock); list_del(&job->entry); - spin_unlock(&sched->job_list.lock); + spin_unlock(&sched->job.list.lock); + + wake_up(&sched->job.wq); } void @@ -305,9 +306,9 @@ nouveau_job_submit(struct nouveau_job *job) } /* Submit was successful; add the job to the schedulers job list. */ - spin_lock(&sched->job_list.lock); - list_add(&job->entry, &sched->job_list.head); - spin_unlock(&sched->job_list.lock); + spin_lock(&sched->job.list.lock); + list_add(&job->entry, &sched->job.list.head); + spin_unlock(&sched->job.list.lock); drm_sched_job_arm(&job->base); job->done_fence = dma_fence_get(&job->base.s_fence->finished); @@ -392,23 +393,10 @@ nouveau_sched_free_job(struct drm_sched_job *sched_job) nouveau_job_fini(job); } -static void -nouveau_sched_cancel_job(struct drm_sched_job *sched_job) -{ - struct nouveau_fence *fence; - struct nouveau_job *job; - - job = to_nouveau_job(sched_job); - fence = to_nouveau_fence(job->done_fence); - - nouveau_fence_cancel(fence); -} - static const struct drm_sched_backend_ops nouveau_sched_ops = { .run_job = nouveau_sched_run_job, .timedout_job = nouveau_sched_timedout_job, .free_job = nouveau_sched_free_job, - .cancel_job = nouveau_sched_cancel_job, }; static int @@ -458,8 +446,9 @@ nouveau_sched_init(struct nouveau_sched *sched, struct nouveau_drm *drm, goto fail_sched; mutex_init(&sched->mutex); - spin_lock_init(&sched->job_list.lock); - INIT_LIST_HEAD(&sched->job_list.head); + spin_lock_init(&sched->job.list.lock); + INIT_LIST_HEAD(&sched->job.list.head); + init_waitqueue_head(&sched->job.wq); return 0; @@ -493,12 +482,16 @@ nouveau_sched_create(struct nouveau_sched **psched, struct nouveau_drm *drm, return 0; } + static void nouveau_sched_fini(struct nouveau_sched *sched) { struct drm_gpu_scheduler *drm_sched = &sched->base; struct drm_sched_entity *entity = &sched->entity; + rmb(); /* for list_empty to work without lock */ + wait_event(sched->job.wq, list_empty(&sched->job.list.head)); + drm_sched_entity_fini(entity); drm_sched_fini(drm_sched); diff --git a/drivers/gpu/drm/nouveau/nouveau_sched.h b/drivers/gpu/drm/nouveau/nouveau_sched.h index b98c3f0bef30..20cd1da8db73 100644 --- a/drivers/gpu/drm/nouveau/nouveau_sched.h +++ b/drivers/gpu/drm/nouveau/nouveau_sched.h @@ -103,9 +103,12 @@ struct nouveau_sched { struct mutex mutex; struct { - struct list_head head; - spinlock_t lock; - } job_list; + struct { + struct list_head head; + spinlock_t lock; + } list; + struct wait_queue_head wq; + } job; }; int nouveau_sched_create(struct nouveau_sched **psched, struct nouveau_drm *drm, diff --git a/drivers/gpu/drm/nouveau/nouveau_uvmm.c b/drivers/gpu/drm/nouveau/nouveau_uvmm.c index ddfc46bc1b3e..48f105239f42 100644 --- a/drivers/gpu/drm/nouveau/nouveau_uvmm.c +++ b/drivers/gpu/drm/nouveau/nouveau_uvmm.c @@ -1019,8 +1019,8 @@ bind_validate_map_sparse(struct nouveau_job *job, u64 addr, u64 range) u64 end = addr + range; again: - spin_lock(&sched->job_list.lock); - list_for_each_entry(__job, &sched->job_list.head, entry) { + spin_lock(&sched->job.list.lock); + list_for_each_entry(__job, &sched->job.list.head, entry) { struct nouveau_uvmm_bind_job *bind_job = to_uvmm_bind_job(__job); list_for_each_op(op, &bind_job->ops) { @@ -1030,7 +1030,7 @@ again: if (!(end <= op_addr || addr >= op_end)) { nouveau_uvmm_bind_job_get(bind_job); - spin_unlock(&sched->job_list.lock); + spin_unlock(&sched->job.list.lock); wait_for_completion(&bind_job->complete); nouveau_uvmm_bind_job_put(bind_job); goto again; @@ -1038,7 +1038,7 @@ again: } } } - spin_unlock(&sched->job_list.lock); + spin_unlock(&sched->job.list.lock); } static int diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c index fdffa0391b31..6fd4e60634fb 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c @@ -350,6 +350,8 @@ nvkm_fifo_dtor(struct nvkm_engine *engine) nvkm_chid_unref(&fifo->chid); nvkm_event_fini(&fifo->nonstall.event); + if (fifo->func->nonstall_dtor) + fifo->func->nonstall_dtor(fifo); mutex_destroy(&fifo->mutex); if (fifo->func->dtor) diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/ga100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/ga100.c index e74493a4569e..6848a56f20c0 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/ga100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/ga100.c @@ -517,19 +517,11 @@ ga100_fifo_nonstall_intr(struct nvkm_inth *inth) static void ga100_fifo_nonstall_block(struct nvkm_event *event, int type, int index) { - struct nvkm_fifo *fifo = container_of(event, typeof(*fifo), nonstall.event); - struct nvkm_runl *runl = nvkm_runl_get(fifo, index, 0); - - nvkm_inth_block(&runl->nonstall.inth); } static void ga100_fifo_nonstall_allow(struct nvkm_event *event, int type, int index) { - struct nvkm_fifo *fifo = container_of(event, typeof(*fifo), nonstall.event); - struct nvkm_runl *runl = nvkm_runl_get(fifo, index, 0); - - nvkm_inth_allow(&runl->nonstall.inth); } const struct nvkm_event_func @@ -564,12 +556,26 @@ ga100_fifo_nonstall_ctor(struct nvkm_fifo *fifo) if (ret) return ret; + nvkm_inth_allow(&runl->nonstall.inth); + nr = max(nr, runl->id + 1); } return nr; } +void +ga100_fifo_nonstall_dtor(struct nvkm_fifo *fifo) +{ + struct nvkm_runl *runl; + + nvkm_runl_foreach(runl, fifo) { + if (runl->nonstall.vector < 0) + continue; + nvkm_inth_block(&runl->nonstall.inth); + } +} + int ga100_fifo_runl_ctor(struct nvkm_fifo *fifo) { @@ -599,6 +605,7 @@ ga100_fifo = { .runl_ctor = ga100_fifo_runl_ctor, .mmu_fault = &tu102_fifo_mmu_fault, .nonstall_ctor = ga100_fifo_nonstall_ctor, + .nonstall_dtor = ga100_fifo_nonstall_dtor, .nonstall = &ga100_fifo_nonstall, .runl = &ga100_runl, .runq = &ga100_runq, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/ga102.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/ga102.c index 755235f55b3a..18a0b1f4eab7 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/ga102.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/ga102.c @@ -30,6 +30,7 @@ ga102_fifo = { .runl_ctor = ga100_fifo_runl_ctor, .mmu_fault = &tu102_fifo_mmu_fault, .nonstall_ctor = ga100_fifo_nonstall_ctor, + .nonstall_dtor = ga100_fifo_nonstall_dtor, .nonstall = &ga100_fifo_nonstall, .runl = &ga100_runl, .runq = &ga100_runq, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h index 5e81ae195329..fff1428ef267 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/priv.h @@ -41,6 +41,7 @@ struct nvkm_fifo_func { void (*start)(struct nvkm_fifo *, unsigned long *); int (*nonstall_ctor)(struct nvkm_fifo *); + void (*nonstall_dtor)(struct nvkm_fifo *); const struct nvkm_event_func *nonstall; const struct nvkm_runl_func *runl; @@ -200,6 +201,7 @@ u32 tu102_chan_doorbell_handle(struct nvkm_chan *); int ga100_fifo_runl_ctor(struct nvkm_fifo *); int ga100_fifo_nonstall_ctor(struct nvkm_fifo *); +void ga100_fifo_nonstall_dtor(struct nvkm_fifo *); extern const struct nvkm_event_func ga100_fifo_nonstall; extern const struct nvkm_runl_func ga100_runl; extern const struct nvkm_runq_func ga100_runq; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/fifo.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/fifo.c index 1ac5628c5140..4ed54b386a60 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/fifo.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/rm/r535/fifo.c @@ -601,6 +601,7 @@ r535_fifo_new(const struct nvkm_fifo_func *hw, struct nvkm_device *device, rm->chan.func = &r535_chan; rm->nonstall = &ga100_fifo_nonstall; rm->nonstall_ctor = ga100_fifo_nonstall_ctor; + rm->nonstall_dtor = ga100_fifo_nonstall_dtor; return nvkm_fifo_new_(rm, device, type, inst, pfifo); } diff --git a/drivers/gpu/drm/panthor/panthor_drv.c b/drivers/gpu/drm/panthor/panthor_drv.c index 1116f2d2826e..4d8e9b34702a 100644 --- a/drivers/gpu/drm/panthor/panthor_drv.c +++ b/drivers/gpu/drm/panthor/panthor_drv.c @@ -1094,7 +1094,7 @@ static int panthor_ioctl_group_create(struct drm_device *ddev, void *data, struct drm_panthor_queue_create *queue_args; int ret; - if (!args->queues.count) + if (!args->queues.count || args->queues.count > MAX_CS_PER_CSG) return -EINVAL; ret = PANTHOR_UOBJ_GET_ARRAY(queue_args, &args->queues); diff --git a/drivers/gpu/drm/scheduler/sched_entity.c b/drivers/gpu/drm/scheduler/sched_entity.c index 8867b95ab089..3d06f72531ba 100644 --- a/drivers/gpu/drm/scheduler/sched_entity.c +++ b/drivers/gpu/drm/scheduler/sched_entity.c @@ -391,7 +391,8 @@ EXPORT_SYMBOL(drm_sched_entity_set_priority); * Add a callback to the current dependency of the entity to wake up the * scheduler when the entity becomes available. */ -static bool drm_sched_entity_add_dependency_cb(struct drm_sched_entity *entity) +static bool drm_sched_entity_add_dependency_cb(struct drm_sched_entity *entity, + struct drm_sched_job *sched_job) { struct drm_gpu_scheduler *sched = entity->rq->sched; struct dma_fence *fence = entity->dependency; @@ -421,6 +422,10 @@ static bool drm_sched_entity_add_dependency_cb(struct drm_sched_entity *entity) entity->dependency = fence; } + if (trace_drm_sched_job_unschedulable_enabled() && + !test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &entity->dependency->flags)) + trace_drm_sched_job_unschedulable(sched_job, entity->dependency); + if (!dma_fence_add_callback(entity->dependency, &entity->cb, drm_sched_entity_wakeup)) return true; @@ -461,10 +466,8 @@ struct drm_sched_job *drm_sched_entity_pop_job(struct drm_sched_entity *entity) while ((entity->dependency = drm_sched_job_dependency(sched_job, entity))) { - if (drm_sched_entity_add_dependency_cb(entity)) { - trace_drm_sched_job_unschedulable(sched_job, entity->dependency); + if (drm_sched_entity_add_dependency_cb(entity, sched_job)) return NULL; - } } /* skip jobs from entity that marked guilty */ diff --git a/drivers/gpu/drm/xe/tests/xe_bo.c b/drivers/gpu/drm/xe/tests/xe_bo.c index bb469096d072..7b40cc8be1c9 100644 --- a/drivers/gpu/drm/xe/tests/xe_bo.c +++ b/drivers/gpu/drm/xe/tests/xe_bo.c @@ -236,7 +236,7 @@ static int evict_test_run_tile(struct xe_device *xe, struct xe_tile *tile, struc } xe_bo_lock(external, false); - err = xe_bo_pin_external(external); + err = xe_bo_pin_external(external, false); xe_bo_unlock(external); if (err) { KUNIT_FAIL(test, "external bo pin err=%pe\n", diff --git a/drivers/gpu/drm/xe/tests/xe_dma_buf.c b/drivers/gpu/drm/xe/tests/xe_dma_buf.c index c53f67ce4b0a..121f17c112ec 100644 --- a/drivers/gpu/drm/xe/tests/xe_dma_buf.c +++ b/drivers/gpu/drm/xe/tests/xe_dma_buf.c @@ -89,15 +89,7 @@ static void check_residency(struct kunit *test, struct xe_bo *exported, return; } - /* - * If on different devices, the exporter is kept in system if - * possible, saving a migration step as the transfer is just - * likely as fast from system memory. - */ - if (params->mem_mask & XE_BO_FLAG_SYSTEM) - KUNIT_EXPECT_TRUE(test, xe_bo_is_mem_type(exported, XE_PL_TT)); - else - KUNIT_EXPECT_TRUE(test, xe_bo_is_mem_type(exported, mem_type)); + KUNIT_EXPECT_TRUE(test, xe_bo_is_mem_type(exported, mem_type)); if (params->force_different_devices) KUNIT_EXPECT_TRUE(test, xe_bo_is_mem_type(imported, XE_PL_TT)); diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c index 1be2415966df..bae7ff2e5927 100644 --- a/drivers/gpu/drm/xe/xe_bo.c +++ b/drivers/gpu/drm/xe/xe_bo.c @@ -186,6 +186,8 @@ static void try_add_system(struct xe_device *xe, struct xe_bo *bo, bo->placements[*c] = (struct ttm_place) { .mem_type = XE_PL_TT, + .flags = (bo_flags & XE_BO_FLAG_VRAM_MASK) ? + TTM_PL_FLAG_FALLBACK : 0, }; *c += 1; } @@ -819,8 +821,7 @@ static int xe_bo_move(struct ttm_buffer_object *ttm_bo, bool evict, return ret; } - tt_has_data = ttm && (ttm_tt_is_populated(ttm) || - (ttm->page_flags & TTM_TT_FLAG_SWAPPED)); + tt_has_data = ttm && (ttm_tt_is_populated(ttm) || ttm_tt_is_swapped(ttm)); move_lacks_source = !old_mem || (handle_system_ccs ? (!bo->ccs_cleared) : (!mem_type_is_vram(old_mem_type) && !tt_has_data)); @@ -2270,6 +2271,7 @@ uint64_t vram_region_gpu_offset(struct ttm_resource *res) /** * xe_bo_pin_external - pin an external BO * @bo: buffer object to be pinned + * @in_place: Pin in current placement, don't attempt to migrate. * * Pin an external (not tied to a VM, can be exported via dma-buf / prime FD) * BO. Unique call compared to xe_bo_pin as this function has it own set of @@ -2277,7 +2279,7 @@ uint64_t vram_region_gpu_offset(struct ttm_resource *res) * * Returns 0 for success, negative error code otherwise. */ -int xe_bo_pin_external(struct xe_bo *bo) +int xe_bo_pin_external(struct xe_bo *bo, bool in_place) { struct xe_device *xe = xe_bo_device(bo); int err; @@ -2286,9 +2288,11 @@ int xe_bo_pin_external(struct xe_bo *bo) xe_assert(xe, xe_bo_is_user(bo)); if (!xe_bo_is_pinned(bo)) { - err = xe_bo_validate(bo, NULL, false); - if (err) - return err; + if (!in_place) { + err = xe_bo_validate(bo, NULL, false); + if (err) + return err; + } spin_lock(&xe->pinned.lock); list_add_tail(&bo->pinned_link, &xe->pinned.late.external); @@ -2441,6 +2445,9 @@ int xe_bo_validate(struct xe_bo *bo, struct xe_vm *vm, bool allow_res_evict) }; int ret; + if (xe_bo_is_pinned(bo)) + return 0; + if (vm) { lockdep_assert_held(&vm->lock); xe_vm_assert_held(vm); diff --git a/drivers/gpu/drm/xe/xe_bo.h b/drivers/gpu/drm/xe/xe_bo.h index 02e8cde4c6b2..9ce94d252015 100644 --- a/drivers/gpu/drm/xe/xe_bo.h +++ b/drivers/gpu/drm/xe/xe_bo.h @@ -198,7 +198,7 @@ static inline void xe_bo_unlock_vm_held(struct xe_bo *bo) } } -int xe_bo_pin_external(struct xe_bo *bo); +int xe_bo_pin_external(struct xe_bo *bo, bool in_place); int xe_bo_pin(struct xe_bo *bo); void xe_bo_unpin_external(struct xe_bo *bo); void xe_bo_unpin(struct xe_bo *bo); diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h index d4d2c6854790..7ceb0c90f391 100644 --- a/drivers/gpu/drm/xe/xe_device_types.h +++ b/drivers/gpu/drm/xe/xe_device_types.h @@ -553,6 +553,12 @@ struct xe_device { /** @pm_notifier: Our PM notifier to perform actions in response to various PM events. */ struct notifier_block pm_notifier; + /** @pm_block: Completion to block validating tasks on suspend / hibernate prepare */ + struct completion pm_block; + /** @rebind_resume_list: List of wq items to kick on resume. */ + struct list_head rebind_resume_list; + /** @rebind_resume_lock: Lock to protect the rebind_resume_list */ + struct mutex rebind_resume_lock; /** @pmt: Support the PMT driver callback interface */ struct { diff --git a/drivers/gpu/drm/xe/xe_dma_buf.c b/drivers/gpu/drm/xe/xe_dma_buf.c index 346f857f3837..af64baf872ef 100644 --- a/drivers/gpu/drm/xe/xe_dma_buf.c +++ b/drivers/gpu/drm/xe/xe_dma_buf.c @@ -72,7 +72,7 @@ static int xe_dma_buf_pin(struct dma_buf_attachment *attach) return ret; } - ret = xe_bo_pin_external(bo); + ret = xe_bo_pin_external(bo, true); xe_assert(xe, !ret); return 0; diff --git a/drivers/gpu/drm/xe/xe_exec.c b/drivers/gpu/drm/xe/xe_exec.c index 44364c042ad7..374c831e691b 100644 --- a/drivers/gpu/drm/xe/xe_exec.c +++ b/drivers/gpu/drm/xe/xe_exec.c @@ -237,6 +237,15 @@ retry: goto err_unlock_list; } + /* + * It's OK to block interruptible here with the vm lock held, since + * on task freezing during suspend / hibernate, the call will + * return -ERESTARTSYS and the IOCTL will be rerun. + */ + err = wait_for_completion_interruptible(&xe->pm_block); + if (err) + goto err_unlock_list; + vm_exec.vm = &vm->gpuvm; vm_exec.flags = DRM_EXEC_INTERRUPTIBLE_WAIT; if (xe_vm_in_lr_mode(vm)) { diff --git a/drivers/gpu/drm/xe/xe_pm.c b/drivers/gpu/drm/xe/xe_pm.c index e279b47ba03b..bb9b6ecad2af 100644 --- a/drivers/gpu/drm/xe/xe_pm.c +++ b/drivers/gpu/drm/xe/xe_pm.c @@ -24,6 +24,7 @@ #include "xe_pcode.h" #include "xe_pxp.h" #include "xe_trace.h" +#include "xe_vm.h" #include "xe_wa.h" /** @@ -290,6 +291,19 @@ static u32 vram_threshold_value(struct xe_device *xe) return DEFAULT_VRAM_THRESHOLD; } +static void xe_pm_wake_rebind_workers(struct xe_device *xe) +{ + struct xe_vm *vm, *next; + + mutex_lock(&xe->rebind_resume_lock); + list_for_each_entry_safe(vm, next, &xe->rebind_resume_list, + preempt.pm_activate_link) { + list_del_init(&vm->preempt.pm_activate_link); + xe_vm_resume_rebind_worker(vm); + } + mutex_unlock(&xe->rebind_resume_lock); +} + static int xe_pm_notifier_callback(struct notifier_block *nb, unsigned long action, void *data) { @@ -299,30 +313,30 @@ static int xe_pm_notifier_callback(struct notifier_block *nb, switch (action) { case PM_HIBERNATION_PREPARE: case PM_SUSPEND_PREPARE: + reinit_completion(&xe->pm_block); xe_pm_runtime_get(xe); err = xe_bo_evict_all_user(xe); - if (err) { + if (err) drm_dbg(&xe->drm, "Notifier evict user failed (%d)\n", err); - xe_pm_runtime_put(xe); - break; - } err = xe_bo_notifier_prepare_all_pinned(xe); - if (err) { + if (err) drm_dbg(&xe->drm, "Notifier prepare pin failed (%d)\n", err); - xe_pm_runtime_put(xe); - } + /* + * Keep the runtime pm reference until post hibernation / post suspend to + * avoid a runtime suspend interfering with evicted objects or backup + * allocations. + */ break; case PM_POST_HIBERNATION: case PM_POST_SUSPEND: + complete_all(&xe->pm_block); + xe_pm_wake_rebind_workers(xe); xe_bo_notifier_unprepare_all_pinned(xe); xe_pm_runtime_put(xe); break; } - if (err) - return NOTIFY_BAD; - return NOTIFY_DONE; } @@ -344,6 +358,14 @@ int xe_pm_init(struct xe_device *xe) if (err) return err; + err = drmm_mutex_init(&xe->drm, &xe->rebind_resume_lock); + if (err) + goto err_unregister; + + init_completion(&xe->pm_block); + complete_all(&xe->pm_block); + INIT_LIST_HEAD(&xe->rebind_resume_list); + /* For now suspend/resume is only allowed with GuC */ if (!xe_device_uc_enabled(xe)) return 0; diff --git a/drivers/gpu/drm/xe/xe_survivability_mode.c b/drivers/gpu/drm/xe/xe_survivability_mode.c index 41705f5d52e3..8f7b0add2364 100644 --- a/drivers/gpu/drm/xe/xe_survivability_mode.c +++ b/drivers/gpu/drm/xe/xe_survivability_mode.c @@ -41,6 +41,8 @@ * * # echo 1 > /sys/kernel/config/xe/0000:03:00.0/survivability_mode * + * It is the responsibility of the user to clear the mode once firmware flash is complete. + * * Refer :ref:`xe_configfs` for more details on how to use configfs * * Survivability mode is indicated by the below admin-only readable sysfs which provides additional @@ -147,7 +149,6 @@ static void xe_survivability_mode_fini(void *arg) struct pci_dev *pdev = to_pci_dev(xe->drm.dev); struct device *dev = &pdev->dev; - xe_configfs_clear_survivability_mode(pdev); sysfs_remove_file(&dev->kobj, &dev_attr_survivability_mode.attr); } diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index d60c4b115304..dc4f61e56579 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -393,6 +393,9 @@ static int xe_gpuvm_validate(struct drm_gpuvm_bo *vm_bo, struct drm_exec *exec) list_move_tail(&gpuva_to_vma(gpuva)->combined_links.rebind, &vm->rebind_list); + if (!try_wait_for_completion(&vm->xe->pm_block)) + return -EAGAIN; + ret = xe_bo_validate(gem_to_xe_bo(vm_bo->obj), vm, false); if (ret) return ret; @@ -479,6 +482,33 @@ static int xe_preempt_work_begin(struct drm_exec *exec, struct xe_vm *vm, return xe_vm_validate_rebind(vm, exec, vm->preempt.num_exec_queues); } +static bool vm_suspend_rebind_worker(struct xe_vm *vm) +{ + struct xe_device *xe = vm->xe; + bool ret = false; + + mutex_lock(&xe->rebind_resume_lock); + if (!try_wait_for_completion(&vm->xe->pm_block)) { + ret = true; + list_move_tail(&vm->preempt.pm_activate_link, &xe->rebind_resume_list); + } + mutex_unlock(&xe->rebind_resume_lock); + + return ret; +} + +/** + * xe_vm_resume_rebind_worker() - Resume the rebind worker. + * @vm: The vm whose preempt worker to resume. + * + * Resume a preempt worker that was previously suspended by + * vm_suspend_rebind_worker(). + */ +void xe_vm_resume_rebind_worker(struct xe_vm *vm) +{ + queue_work(vm->xe->ordered_wq, &vm->preempt.rebind_work); +} + static void preempt_rebind_work_func(struct work_struct *w) { struct xe_vm *vm = container_of(w, struct xe_vm, preempt.rebind_work); @@ -502,6 +532,11 @@ static void preempt_rebind_work_func(struct work_struct *w) } retry: + if (!try_wait_for_completion(&vm->xe->pm_block) && vm_suspend_rebind_worker(vm)) { + up_write(&vm->lock); + return; + } + if (xe_vm_userptr_check_repin(vm)) { err = xe_vm_userptr_pin(vm); if (err) @@ -1714,6 +1749,7 @@ struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags, struct xe_file *xef) if (flags & XE_VM_FLAG_LR_MODE) { INIT_WORK(&vm->preempt.rebind_work, preempt_rebind_work_func); xe_pm_runtime_get_noresume(xe); + INIT_LIST_HEAD(&vm->preempt.pm_activate_link); } if (flags & XE_VM_FLAG_FAULT_MODE) { @@ -1895,8 +1931,12 @@ void xe_vm_close_and_put(struct xe_vm *vm) xe_assert(xe, !vm->preempt.num_exec_queues); xe_vm_close(vm); - if (xe_vm_in_preempt_fence_mode(vm)) + if (xe_vm_in_preempt_fence_mode(vm)) { + mutex_lock(&xe->rebind_resume_lock); + list_del_init(&vm->preempt.pm_activate_link); + mutex_unlock(&xe->rebind_resume_lock); flush_work(&vm->preempt.rebind_work); + } if (xe_vm_in_fault_mode(vm)) xe_svm_close(vm); diff --git a/drivers/gpu/drm/xe/xe_vm.h b/drivers/gpu/drm/xe/xe_vm.h index 2ecb417c19a2..82b112795807 100644 --- a/drivers/gpu/drm/xe/xe_vm.h +++ b/drivers/gpu/drm/xe/xe_vm.h @@ -273,6 +273,8 @@ struct dma_fence *xe_vm_bind_kernel_bo(struct xe_vm *vm, struct xe_bo *bo, struct xe_exec_queue *q, u64 addr, enum xe_cache_level cache_lvl); +void xe_vm_resume_rebind_worker(struct xe_vm *vm); + /** * xe_vm_resv() - Return's the vm's reservation object * @vm: The vm diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h index 8a07feef503b..6058cf739388 100644 --- a/drivers/gpu/drm/xe/xe_vm_types.h +++ b/drivers/gpu/drm/xe/xe_vm_types.h @@ -293,6 +293,11 @@ struct xe_vm { * BOs */ struct work_struct rebind_work; + /** + * @preempt.pm_activate_link: Link to list of rebind workers to be + * kicked on resume. + */ + struct list_head pm_activate_link; } preempt; /** @um: unified memory state */ diff --git a/drivers/gpu/drm/xe/xe_wa_oob.rules b/drivers/gpu/drm/xe/xe_wa_oob.rules index e990f20eccfe..710f4423726c 100644 --- a/drivers/gpu/drm/xe/xe_wa_oob.rules +++ b/drivers/gpu/drm/xe/xe_wa_oob.rules @@ -30,7 +30,8 @@ 16022287689 GRAPHICS_VERSION(2001) GRAPHICS_VERSION(2004) 13011645652 GRAPHICS_VERSION(2004) - GRAPHICS_VERSION(3001) + GRAPHICS_VERSION_RANGE(3000, 3001) + GRAPHICS_VERSION(3003) 14022293748 GRAPHICS_VERSION_RANGE(2001, 2002) GRAPHICS_VERSION(2004) GRAPHICS_VERSION_RANGE(3000, 3001) diff --git a/drivers/gpu/nova-core/Kconfig b/drivers/gpu/nova-core/Kconfig index 8726d80d6ba4..20d3e6d0d796 100644 --- a/drivers/gpu/nova-core/Kconfig +++ b/drivers/gpu/nova-core/Kconfig @@ -1,5 +1,6 @@ config NOVA_CORE tristate "Nova Core GPU driver" + depends on 64BIT depends on PCI depends on RUST depends on RUST_FW_LOADER_ABSTRACTIONS |