From 9b608fe94870fe46bd2c41fcda99e74dabd6bbc6 Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Mon, 29 Sep 2025 17:59:26 +0530 Subject: drm/amdgpu: Check swus/ds for switch state save For saving switch state, check if the GPU is having SWUS/DS architecture. Otherwise, skip saving. Reported-by: Roman Elshin Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4602 Fixes: 1dd2fa0e00f1 ("drm/amdgpu: Save and restore switch state") Signed-off-by: Lijo Lazar Acked-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 23 +++++++++++++++-------- 1 file changed, 15 insertions(+), 8 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 a77000c2e0bb..929936c8d87c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -7157,28 +7157,35 @@ void amdgpu_pci_resume(struct pci_dev *pdev) static void amdgpu_device_cache_switch_state(struct amdgpu_device *adev) { - struct pci_dev *parent = pci_upstream_bridge(adev->pdev); + struct pci_dev *swus, *swds; int r; - if (!parent || parent->vendor != PCI_VENDOR_ID_ATI) + swds = pci_upstream_bridge(adev->pdev); + if (!swds || swds->vendor != PCI_VENDOR_ID_ATI || + pci_pcie_type(swds) != PCI_EXP_TYPE_DOWNSTREAM) + return; + swus = pci_upstream_bridge(swds); + if (!swus || + (swus->vendor != PCI_VENDOR_ID_ATI && + swus->vendor != PCI_VENDOR_ID_AMD) || + pci_pcie_type(swus) != PCI_EXP_TYPE_UPSTREAM) return; /* If already saved, return */ if (adev->pcie_reset_ctx.swus) return; /* Upstream bridge is ATI, assume it's SWUS/DS architecture */ - r = pci_save_state(parent); + r = pci_save_state(swds); if (r) return; - adev->pcie_reset_ctx.swds_pcistate = pci_store_saved_state(parent); + adev->pcie_reset_ctx.swds_pcistate = pci_store_saved_state(swds); - parent = pci_upstream_bridge(parent); - r = pci_save_state(parent); + r = pci_save_state(swus); if (r) return; - adev->pcie_reset_ctx.swus_pcistate = pci_store_saved_state(parent); + adev->pcie_reset_ctx.swus_pcistate = pci_store_saved_state(swus); - adev->pcie_reset_ctx.swus = parent; + adev->pcie_reset_ctx.swus = swus; } static void amdgpu_device_load_switch_state(struct amdgpu_device *adev) -- cgit v1.2.3 From 2e97663760e5fb7ee14f399c68e57b894f01e505 Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Mon, 6 Oct 2025 10:39:03 +0530 Subject: drm/amdgpu: Report individual reset error If reinitialization of one of the GPUs fails after reset, it logs failure on all subsequent GPUs eventhough they have resumed successfully. A sample log where only device at 0000:95:00.0 had a failure - amdgpu 0000:15:00.0: amdgpu: GPU reset(19) succeeded! amdgpu 0000:65:00.0: amdgpu: GPU reset(19) succeeded! amdgpu 0000:75:00.0: amdgpu: GPU reset(19) succeeded! amdgpu 0000:85:00.0: amdgpu: GPU reset(19) succeeded! amdgpu 0000:95:00.0: amdgpu: GPU reset(19) failed amdgpu 0000:e5:00.0: amdgpu: GPU reset(19) failed amdgpu 0000:f5:00.0: amdgpu: GPU reset(19) failed amdgpu 0000:05:00.0: amdgpu: GPU reset(19) failed amdgpu 0000:15:00.0: amdgpu: GPU reset end with ret = -5 To avoid confusion, report the error for each device separately and return the first error as the overall result. Signed-off-by: Lijo Lazar Reviewed-by: Asad Kamal Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 25 +++++++++++++++---------- 1 file changed, 15 insertions(+), 10 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 929936c8d87c..7a899fb4de29 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -6389,23 +6389,28 @@ static int amdgpu_device_sched_resume(struct list_head *device_list, if (!drm_drv_uses_atomic_modeset(adev_to_drm(tmp_adev)) && !job_signaled) drm_helper_resume_force_mode(adev_to_drm(tmp_adev)); - if (tmp_adev->asic_reset_res) - r = tmp_adev->asic_reset_res; - - tmp_adev->asic_reset_res = 0; - - if (r) { + if (tmp_adev->asic_reset_res) { /* bad news, how to tell it to userspace ? * for ras error, we should report GPU bad status instead of * reset failure */ if (reset_context->src != AMDGPU_RESET_SRC_RAS || !amdgpu_ras_eeprom_check_err_threshold(tmp_adev)) - dev_info(tmp_adev->dev, "GPU reset(%d) failed\n", - atomic_read(&tmp_adev->gpu_reset_counter)); - amdgpu_vf_error_put(tmp_adev, AMDGIM_ERROR_VF_GPU_RESET_FAIL, 0, r); + dev_info( + tmp_adev->dev, + "GPU reset(%d) failed with error %d \n", + atomic_read( + &tmp_adev->gpu_reset_counter), + tmp_adev->asic_reset_res); + amdgpu_vf_error_put(tmp_adev, + AMDGIM_ERROR_VF_GPU_RESET_FAIL, 0, + tmp_adev->asic_reset_res); + if (!r) + r = tmp_adev->asic_reset_res; + tmp_adev->asic_reset_res = 0; } else { - dev_info(tmp_adev->dev, "GPU reset(%d) succeeded!\n", atomic_read(&tmp_adev->gpu_reset_counter)); + dev_info(tmp_adev->dev, "GPU reset(%d) succeeded!\n", + atomic_read(&tmp_adev->gpu_reset_counter)); if (amdgpu_acpi_smart_shift_update(tmp_adev, AMDGPU_SS_DEV_D0)) dev_warn(tmp_adev->dev, -- cgit v1.2.3