From 1bea57ea7544a8058b32d8605075e24aa2814fcc Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Thu, 25 Sep 2025 12:09:56 +0200 Subject: drm/amdgpu: reduce queue timeout to 2 seconds v2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There has been multiple complains that 10 seconds are usually to long. The original requirement for longer timeout came from compute tests on AMDVLK, since that is no longer a topic reduce the timeout back to 2 seconds for all queues. While at it also remove any special handling for compute queues under SRIOV or pass through. v2: fix checkpatch warning. Signed-off-by: Christian König Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 85 ++++++++++++++---------------- 1 file changed, 40 insertions(+), 45 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 3d032c4e2dce..6d228012187c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -4285,58 +4285,53 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev) long timeout; int ret = 0; - /* - * By default timeout for jobs is 10 sec - */ - adev->compute_timeout = adev->gfx_timeout = msecs_to_jiffies(10000); - adev->sdma_timeout = adev->video_timeout = adev->gfx_timeout; + /* By default timeout for all queues is 2 sec */ + adev->gfx_timeout = adev->compute_timeout = adev->sdma_timeout = + adev->video_timeout = msecs_to_jiffies(2000); - if (strnlen(input, AMDGPU_MAX_TIMEOUT_PARAM_LENGTH)) { - while ((timeout_setting = strsep(&input, ",")) && - strnlen(timeout_setting, AMDGPU_MAX_TIMEOUT_PARAM_LENGTH)) { - ret = kstrtol(timeout_setting, 0, &timeout); - if (ret) - return ret; + if (!strnlen(input, AMDGPU_MAX_TIMEOUT_PARAM_LENGTH)) + return 0; - if (timeout == 0) { - index++; - continue; - } else if (timeout < 0) { - timeout = MAX_SCHEDULE_TIMEOUT; - dev_warn(adev->dev, "lockup timeout disabled"); - add_taint(TAINT_SOFTLOCKUP, LOCKDEP_STILL_OK); - } else { - timeout = msecs_to_jiffies(timeout); - } + while ((timeout_setting = strsep(&input, ",")) && + strnlen(timeout_setting, AMDGPU_MAX_TIMEOUT_PARAM_LENGTH)) { + ret = kstrtol(timeout_setting, 0, &timeout); + if (ret) + return ret; - switch (index++) { - case 0: - adev->gfx_timeout = timeout; - break; - case 1: - adev->compute_timeout = timeout; - break; - case 2: - adev->sdma_timeout = timeout; - break; - case 3: - adev->video_timeout = timeout; - break; - default: - break; - } + if (timeout == 0) { + index++; + continue; + } else if (timeout < 0) { + timeout = MAX_SCHEDULE_TIMEOUT; + dev_warn(adev->dev, "lockup timeout disabled"); + add_taint(TAINT_SOFTLOCKUP, LOCKDEP_STILL_OK); + } else { + timeout = msecs_to_jiffies(timeout); } - /* - * There is only one value specified and - * it should apply to all non-compute jobs. - */ - if (index == 1) { - adev->sdma_timeout = adev->video_timeout = adev->gfx_timeout; - if (amdgpu_sriov_vf(adev) || amdgpu_passthrough(adev)) - adev->compute_timeout = adev->gfx_timeout; + + switch (index++) { + case 0: + adev->gfx_timeout = timeout; + break; + case 1: + adev->compute_timeout = timeout; + break; + case 2: + adev->sdma_timeout = timeout; + break; + case 3: + adev->video_timeout = timeout; + break; + default: + break; } } + /* When only one value specified apply it to all queues. */ + if (index == 1) + adev->gfx_timeout = adev->compute_timeout = adev->sdma_timeout = + adev->video_timeout = timeout; + return ret; } -- cgit v1.2.3 From 6062ede68097d21aec896820071257362c965cc2 Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Thu, 2 Oct 2025 12:42:40 -0500 Subject: drm/amd: Stop exporting amdgpu_device_ip_suspend() outside amdgpu_device amdgpu_device_ip_suspend() doesn't have a caller outside of amdgpu_device.c. Make it static. No intended functional changes. Reviewed-by: Alex Deucher Signed-off-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 6d228012187c..edd086213049 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3895,7 +3895,7 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev) * in each IP into a state suitable for suspend. * Returns 0 on success, negative error code on failure. */ -int amdgpu_device_ip_suspend(struct amdgpu_device *adev) +static int amdgpu_device_ip_suspend(struct amdgpu_device *adev) { int r; -- cgit v1.2.3 From f35f2541780b05d0ebc4cdf5a6e9d8aea6573369 Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Thu, 2 Oct 2025 12:42:41 -0500 Subject: drm/amd: Remove comment about handling errors in amdgpu_device_ip_suspend_phase1() Error handling was introduced in commit e095026f0066 ("drm/amdgpu: validate suspend before function call") so the comment about TODO is no longer needed. Fixes: e095026f0066 ("drm/amdgpu: validate suspend before function call") Reviewed-by: Alex Deucher Signed-off-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 - 1 file changed, 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index edd086213049..3c1d784d6f3b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3780,7 +3780,6 @@ static int amdgpu_device_ip_suspend_phase1(struct amdgpu_device *adev) if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_DCE) continue; - /* XXX handle errors */ r = amdgpu_ip_block_suspend(&adev->ip_blocks[i]); if (r) return r; -- cgit v1.2.3 From b7ff2e79240a51e8639ae0c304423bb1cb460bb5 Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Thu, 2 Oct 2025 12:42:42 -0500 Subject: drm/amd: Don't always set IP block HW status to false amdgpu_device_ip_suspend_phase2() calls amdgpu_ip_block_suspend() which already sets HW block status to false when succeeding with IP suspend. Remove the explicit call in amdgpu_device_ip_suspend_phase2() so that the status is accurate. Reviewed-by: Alex Deucher Signed-off-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 - 1 file changed, 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 3c1d784d6f3b..70f5ff4c63e7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3864,7 +3864,6 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev) /* XXX handle errors */ r = amdgpu_ip_block_suspend(&adev->ip_blocks[i]); - adev->ip_blocks[i].status.hw = false; /* handle putting the SMC in the appropriate state */ if (!amdgpu_sriov_vf(adev)) { -- cgit v1.2.3 From 173360fe49c2cd5ec581056985a7a3d4fbd9f1ab Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Thu, 2 Oct 2025 12:42:43 -0500 Subject: drm/amd: Pass IP suspend errors up to callers If IP suspend fails the callers should be notified so that they can potentially react. Reviewed-by: Alex Deucher Signed-off-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 70f5ff4c63e7..a734c05cc00e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3862,8 +3862,9 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev) adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP) continue; - /* XXX handle errors */ r = amdgpu_ip_block_suspend(&adev->ip_blocks[i]); + if (r) + return r; /* handle putting the SMC in the appropriate state */ if (!amdgpu_sriov_vf(adev)) { @@ -5218,7 +5219,9 @@ int amdgpu_device_suspend(struct drm_device *dev, bool notify_clients) amdgpu_ras_suspend(adev); - amdgpu_device_ip_suspend_phase1(adev); + r = amdgpu_device_ip_suspend_phase1(adev); + if (r) + return r; amdgpu_amdkfd_suspend(adev, !amdgpu_sriov_vf(adev) && !adev->in_runpm); amdgpu_userq_suspend(adev); @@ -5231,7 +5234,9 @@ int amdgpu_device_suspend(struct drm_device *dev, bool notify_clients) amdgpu_fence_driver_hw_fini(adev); - amdgpu_device_ip_suspend_phase2(adev); + r = amdgpu_device_ip_suspend_phase2(adev); + if (r) + return r; if (amdgpu_sriov_vf(adev)) amdgpu_virt_release_full_gpu(adev, false); -- cgit v1.2.3 From 1f3cca77943b6c8bfc7342722e2ce7b348484ee3 Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Thu, 2 Oct 2025 12:42:45 -0500 Subject: drm/amd: Pass userq suspend failures up to caller If a userq failed to suspend the rest of the suspend sequence may have problems. Pass the error code up to the caller for a decision on what to do. Reviewed-by: Alex Deucher Signed-off-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index a734c05cc00e..c7c999ae2a28 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -5224,7 +5224,9 @@ int amdgpu_device_suspend(struct drm_device *dev, bool notify_clients) return r; amdgpu_amdkfd_suspend(adev, !amdgpu_sriov_vf(adev) && !adev->in_runpm); - amdgpu_userq_suspend(adev); + r = amdgpu_userq_suspend(adev); + if (r) + return r; r = amdgpu_device_evict_resources(adev); if (r) -- cgit v1.2.3 From db36632ea51e8b02dd637d291c755899b20a5a31 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 27 Aug 2025 11:34:14 -0400 Subject: drm/amdgpu: clean up and unify hw fence handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Decouple the amdgpu fence from the amdgpu_job structure. This lets us clean up the separate fence ops for the embedded fence and other fences. This also allows us to allocate the vm fence up front when we allocate the job. v2: Additional cleanup suggested by Christian v3: Additional cleanups suggested by Christian v4: Additional cleanups suggested by David and vm fence fix v5: cast seqno (David) Cc: David.Wu3@amd.com Cc: christian.koenig@amd.com Tested-by: David (Ming Qiang) Wu Reviewed-by: David (Ming Qiang) Wu Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index c7c999ae2a28..b6d2b9c0bac0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -5809,11 +5809,6 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, if (!amdgpu_ring_sched_ready(ring)) continue; - /* Clear job fence from fence drv to avoid force_completion - * leave NULL and vm flush fence in fence drv - */ - amdgpu_fence_driver_clear_job_fences(ring); - /* after all hw jobs are reset, hw fence is meaningless, so force_completion */ amdgpu_fence_driver_force_completion(ring); } @@ -6542,7 +6537,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, * * job->base holds a reference to parent fence */ - if (job && dma_fence_is_signaled(&job->hw_fence.base)) { + if (job && dma_fence_is_signaled(&job->hw_fence->base)) { job_signaled = true; dev_info(adev->dev, "Guilty job already signaled, skipping HW reset"); goto skip_hw_reset; -- cgit v1.2.3 From aa6674f2da055da0629a62eb8debe774515645df Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Tue, 7 Oct 2025 18:12:04 +0530 Subject: drm/amdgpu: Reorganize sysfs ini/fini calls Aggregate sysfs ini/fini calls into separate functions. No functional change. Signed-off-by: Lijo Lazar Reviewed-by: Asad Kamal Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 88 +++++++++++++++++------------- 1 file changed, 51 insertions(+), 37 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index b6d2b9c0bac0..86818aa9ba13 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -4385,6 +4385,55 @@ static void amdgpu_device_set_mcbp(struct amdgpu_device *adev) dev_info(adev->dev, "MCBP is enabled\n"); } +static int amdgpu_device_sys_interface_init(struct amdgpu_device *adev) +{ + int r; + + r = amdgpu_atombios_sysfs_init(adev); + if (r) + drm_err(&adev->ddev, + "registering atombios sysfs failed (%d).\n", r); + + r = amdgpu_pm_sysfs_init(adev); + if (r) + dev_err(adev->dev, "registering pm sysfs failed (%d).\n", r); + + r = amdgpu_ucode_sysfs_init(adev); + if (r) { + adev->ucode_sysfs_en = false; + dev_err(adev->dev, "Creating firmware sysfs failed (%d).\n", r); + } else + adev->ucode_sysfs_en = true; + + r = amdgpu_device_attr_sysfs_init(adev); + if (r) + dev_err(adev->dev, "Could not create amdgpu device attr\n"); + + r = devm_device_add_group(adev->dev, &amdgpu_board_attrs_group); + if (r) + dev_err(adev->dev, + "Could not create amdgpu board attributes\n"); + + amdgpu_fru_sysfs_init(adev); + amdgpu_reg_state_sysfs_init(adev); + amdgpu_xcp_sysfs_init(adev); + + return r; +} + +static void amdgpu_device_sys_interface_fini(struct amdgpu_device *adev) +{ + if (adev->pm.sysfs_initialized) + amdgpu_pm_sysfs_fini(adev); + if (adev->ucode_sysfs_en) + amdgpu_ucode_sysfs_fini(adev); + amdgpu_device_attr_sysfs_fini(adev); + amdgpu_fru_sysfs_fini(adev); + + amdgpu_reg_state_sysfs_fini(adev); + amdgpu_xcp_sysfs_fini(adev); +} + /** * amdgpu_device_init - initialize the driver * @@ -4813,34 +4862,7 @@ fence_driver_init: * operations performed in `late_init` might affect the sysfs * interfaces creating. */ - r = amdgpu_atombios_sysfs_init(adev); - if (r) - drm_err(&adev->ddev, - "registering atombios sysfs failed (%d).\n", r); - - r = amdgpu_pm_sysfs_init(adev); - if (r) - dev_err(adev->dev, "registering pm sysfs failed (%d).\n", r); - - r = amdgpu_ucode_sysfs_init(adev); - if (r) { - adev->ucode_sysfs_en = false; - dev_err(adev->dev, "Creating firmware sysfs failed (%d).\n", r); - } else - adev->ucode_sysfs_en = true; - - r = amdgpu_device_attr_sysfs_init(adev); - if (r) - dev_err(adev->dev, "Could not create amdgpu device attr\n"); - - r = devm_device_add_group(adev->dev, &amdgpu_board_attrs_group); - if (r) - dev_err(adev->dev, - "Could not create amdgpu board attributes\n"); - - amdgpu_fru_sysfs_init(adev); - amdgpu_reg_state_sysfs_init(adev); - amdgpu_xcp_sysfs_init(adev); + r = amdgpu_device_sys_interface_init(adev); if (IS_ENABLED(CONFIG_PERF_EVENTS)) r = amdgpu_pmu_init(adev); @@ -4962,15 +4984,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev) } amdgpu_fence_driver_hw_fini(adev); - if (adev->pm.sysfs_initialized) - amdgpu_pm_sysfs_fini(adev); - if (adev->ucode_sysfs_en) - amdgpu_ucode_sysfs_fini(adev); - amdgpu_device_attr_sysfs_fini(adev); - amdgpu_fru_sysfs_fini(adev); - - amdgpu_reg_state_sysfs_fini(adev); - amdgpu_xcp_sysfs_fini(adev); + amdgpu_device_sys_interface_fini(adev); /* disable ras feature must before hw fini */ amdgpu_ras_pre_fini(adev); -- cgit v1.2.3 From 9e2096baab9addedde324f406d3ec9166e70d15b Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Fri, 10 Oct 2025 17:12:52 +0530 Subject: drm/amdgpu: Add amdgpu_discovery_info Add amdgpu_discovery_info structure to keep all discovery related information. Signed-off-by: Lijo Lazar Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 86818aa9ba13..50acf2a94492 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -2633,7 +2633,7 @@ static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev) chip_name = "arcturus"; break; case CHIP_NAVI12: - if (adev->mman.discovery_bin) + if (adev->discovery.bin) return 0; chip_name = "navi12"; break; @@ -5059,7 +5059,7 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev) if (IS_ENABLED(CONFIG_PERF_EVENTS)) amdgpu_pmu_fini(adev); - if (adev->mman.discovery_bin) + if (adev->discovery.bin) amdgpu_discovery_fini(adev); amdgpu_reset_put_reset_domain(adev->reset_domain); -- cgit v1.2.3 From 1cbac73d1a15db0f149cb0c6a932301de5f15b05 Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Tue, 7 Oct 2025 18:30:24 +0530 Subject: drm/amdgpu: Move reset-on-init sequence earlier Complete reset-on-init sequence before sysfs interfaces are created. Devices get properly initiaized only after reset, and then only sysfs interfaces should be made available. Signed-off-by: Lijo Lazar Reviewed-by: Asad Kamal Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 50acf2a94492..0d5585bc3b04 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -4857,6 +4857,8 @@ fence_driver_init: flush_delayed_work(&adev->delayed_init_work); } + if (adev->init_lvl->level == AMDGPU_INIT_LEVEL_MINIMAL_XGMI) + amdgpu_xgmi_reset_on_init(adev); /* * Place those sysfs registering after `late_init`. As some of those * operations performed in `late_init` might affect the sysfs @@ -4890,9 +4892,6 @@ fence_driver_init: if (px) vga_switcheroo_init_domain_pm_ops(adev->dev, &adev->vga_pm_domain); - if (adev->init_lvl->level == AMDGPU_INIT_LEVEL_MINIMAL_XGMI) - amdgpu_xgmi_reset_on_init(adev); - amdgpu_device_check_iommu_direct_map(adev); adev->pm_nb.notifier_call = amdgpu_device_pm_notifier; -- cgit v1.2.3 From 152dca4ea77e575b21c5ee96c582fc1083a25f8a Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Tue, 14 Oct 2025 14:30:35 -0500 Subject: drm/amd: Add a helper to tell whether an IP block HW is enabled There is already a helper for telling if a block is valid, but if IP handling wants to check if it's HW is enabled no such helper exists. Reviewed-by: Harry Wentland Signed-off-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 0d5585bc3b04..038804d48341 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -2387,7 +2387,7 @@ int amdgpu_device_ip_wait_for_idle(struct amdgpu_device *adev, } /** - * amdgpu_device_ip_is_valid - is the hardware IP enabled + * amdgpu_device_ip_is_hw - is the hardware IP enabled * * @adev: amdgpu_device pointer * @block_type: Type of hardware IP (SMU, GFX, UVD, etc.) @@ -2395,6 +2395,27 @@ int amdgpu_device_ip_wait_for_idle(struct amdgpu_device *adev, * Check if the hardware IP is enable or not. * Returns true if it the IP is enable, false if not. */ +bool amdgpu_device_ip_is_hw(struct amdgpu_device *adev, + enum amd_ip_block_type block_type) +{ + int i; + + for (i = 0; i < adev->num_ip_blocks; i++) { + if (adev->ip_blocks[i].version->type == block_type) + return adev->ip_blocks[i].status.hw; + } + return false; +} + +/** + * amdgpu_device_ip_is_valid - is the hardware IP valid + * + * @adev: amdgpu_device pointer + * @block_type: Type of hardware IP (SMU, GFX, UVD, etc.) + * + * Check if the hardware IP is valid or not. + * Returns true if it the IP is valid, false if not. + */ bool amdgpu_device_ip_is_valid(struct amdgpu_device *adev, enum amd_ip_block_type block_type) { -- cgit v1.2.3 From 6169b555db1392e79159e114fff105987231e4ce Mon Sep 17 00:00:00 2001 From: Victor Zhao Date: Thu, 9 Oct 2025 10:42:48 +0800 Subject: drm/amdgpu: use GPU_HDP_FLUSH for sriov Currently SRIOV runtime will use kiq to write HDP_MEM_FLUSH_CNTL for hdp flush. This register need to be write from CPU for nbif to aware, otherwise it will not work. Implement amdgpu_kiq_hdp_flush and use kiq to do gpu hdp flush during sriov runtime. v2: - fallback to amdgpu_asic_flush_hdp when amdgpu_kiq_hdp_flush failed - add function amdgpu_mes_hdp_flush v3: - changed returned error Reviewed-by: Lijo Lazar Signed-off-by: Victor Zhao Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 038804d48341..a99185ed0642 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -7315,10 +7315,17 @@ void amdgpu_device_flush_hdp(struct amdgpu_device *adev, if (adev->gmc.xgmi.connected_to_cpu) return; - if (ring && ring->funcs->emit_hdp_flush) + if (ring && ring->funcs->emit_hdp_flush) { amdgpu_ring_emit_hdp_flush(ring); - else - amdgpu_asic_flush_hdp(adev, ring); + return; + } + + if (!ring && amdgpu_sriov_runtime(adev)) { + if (!amdgpu_kiq_hdp_flush(adev)) + return; + } + + amdgpu_asic_flush_hdp(adev, ring); } void amdgpu_device_invalidate_hdp(struct amdgpu_device *adev, -- cgit v1.2.3 From 07009df6494dc9272b233eeddc10d26f545ff09a Mon Sep 17 00:00:00 2001 From: Ellen Pan Date: Tue, 7 Oct 2025 11:00:16 -0500 Subject: drm/amdgpu: Introduce SRIOV critical regions v2 during VF init 1. Introduced amdgpu_virt_init_critical_region during VF init. - VFs use init_data_header_offset and init_data_header_size_kb transmitted via PF2VF mailbox to fetch the offset of critical regions' offsets/sizes in VRAM and save to adev->virt.crit_region_offsets and adev->virt.crit_region_sizes_kb. Signed-off-by: Ellen Pan Reviewed-by: Alex Deucher Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index a99185ed0642..3ffb9bb1ec0b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -2782,6 +2782,10 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) r = amdgpu_virt_request_full_gpu(adev, true); if (r) return r; + + r = amdgpu_virt_init_critical_region(adev); + if (r) + return r; } switch (adev->asic_type) { -- cgit v1.2.3