This is a note to let you know that I've just added the patch titled drm/amdgpu: Add amdgpu suspend-resume code path under SRIOV to the 5.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: drm-amdgpu-add-amdgpu-suspend-resume-code-path-under-sriov.patch and it can be found in the queue-5.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. >From 3b7329cf5a767c1be38352d43066012e220ad43c Mon Sep 17 00:00:00 2001 From: Bokun Zhang <Bokun.Zhang@xxxxxxx> Date: Wed, 28 Sep 2022 00:30:04 +0800 Subject: drm/amdgpu: Add amdgpu suspend-resume code path under SRIOV From: Bokun Zhang <Bokun.Zhang@xxxxxxx> commit 3b7329cf5a767c1be38352d43066012e220ad43c upstream. - Under SRIOV, we need to send REQ_GPU_FINI to the hypervisor during the suspend time. Furthermore, we cannot request a mode 1 reset under SRIOV as VF. Therefore, we will skip it as it is called in suspend_noirq() function. - In the resume code path, we need to send REQ_GPU_INIT to the hypervisor and also resume PSP IP block under SRIOV. Signed-off-by: Bokun Zhang <Bokun.Zhang@xxxxxxx> Reviewed-by: Alex Deucher <alexander.deucher@xxxxxxx> Signed-off-by: Alex Deucher <alexander.deucher@xxxxxxx> Cc: stable@xxxxxxxxxxxxxxx Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c | 4 ++++ drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 27 ++++++++++++++++++++++++++- 2 files changed, 30 insertions(+), 1 deletion(-) --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c @@ -1056,6 +1056,10 @@ bool amdgpu_acpi_should_gpu_reset(struct { if (adev->flags & AMD_IS_APU) return false; + + if (amdgpu_sriov_vf(adev)) + return false; + return pm_suspend_target_state != PM_SUSPEND_TO_IDLE; } --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3137,7 +3137,8 @@ static int amdgpu_device_ip_resume_phase continue; if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON || adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC || - adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH) { + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH || + (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP && amdgpu_sriov_vf(adev))) { r = adev->ip_blocks[i].version->funcs->resume(adev); if (r) { @@ -4001,12 +4002,20 @@ static void amdgpu_device_evict_resource int amdgpu_device_suspend(struct drm_device *dev, bool fbcon) { struct amdgpu_device *adev = drm_to_adev(dev); + int r = 0; if (dev->switch_power_state == DRM_SWITCH_POWER_OFF) return 0; adev->in_suspend = true; + if (amdgpu_sriov_vf(adev)) { + amdgpu_virt_fini_data_exchange(adev); + r = amdgpu_virt_request_full_gpu(adev, false); + if (r) + return r; + } + if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DEV_D3)) DRM_WARN("smart shift update failed\n"); @@ -4035,6 +4044,9 @@ int amdgpu_device_suspend(struct drm_dev */ amdgpu_device_evict_resources(adev); + if (amdgpu_sriov_vf(adev)) + amdgpu_virt_release_full_gpu(adev, false); + return 0; } @@ -4053,6 +4065,12 @@ int amdgpu_device_resume(struct drm_devi struct amdgpu_device *adev = drm_to_adev(dev); int r = 0; + if (amdgpu_sriov_vf(adev)) { + r = amdgpu_virt_request_full_gpu(adev, true); + if (r) + return r; + } + if (dev->switch_power_state == DRM_SWITCH_POWER_OFF) return 0; @@ -4067,6 +4085,13 @@ int amdgpu_device_resume(struct drm_devi } r = amdgpu_device_ip_resume(adev); + + /* no matter what r is, always need to properly release full GPU */ + if (amdgpu_sriov_vf(adev)) { + amdgpu_virt_init_data_exchange(adev); + amdgpu_virt_release_full_gpu(adev, true); + } + if (r) { dev_err(adev->dev, "amdgpu_device_ip_resume failed (%d).\n", r); return r; Patches currently in stable-queue which might be from Bokun.Zhang@xxxxxxx are queue-5.15/drm-amdgpu-add-amdgpu-suspend-resume-code-path-under-sriov.patch