From 0fa4c25db8b791f79bc0d5a0cd58aff9ad85186b Mon Sep 17 00:00:00 2001 From: Tim Huang Date: Fri, 26 Apr 2024 08:43:30 +0800 Subject: drm/amdgpu: fix uninitialized scalar variable warning Clear warning that field bp is uninitialized when calling amdgpu_virt_ras_add_bps. Signed-off-by: Tim Huang Reviewed-by: Yang Wang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index 54ab51a4ada7..b5fc0e1ad435 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -395,6 +395,8 @@ static void amdgpu_virt_add_bad_page(struct amdgpu_device *adev, else vram_usage_va = adev->mman.drv_vram_usage_va; + memset(&bp, 0, sizeof(bp)); + if (bp_block_size) { bp_cnt = bp_block_size / sizeof(uint64_t); for (bp_idx = 0; bp_idx < bp_cnt; bp_idx++) { -- cgit v1.2.3 From 3bcc0ee14768d886cedff65da72d83d375a31a56 Mon Sep 17 00:00:00 2001 From: Zhigang Luo Date: Tue, 16 Apr 2024 16:35:14 -0400 Subject: drm/amdgpu: avoid reading vf2pf info size from FB VF can't access FB when host is doing mode1 reset. Using sizeof to get vf2pf info size, instead of reading it from vf2pf header stored in FB. Signed-off-by: Zhigang Luo Reviewed-by: Hawking Zhang Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index b5fc0e1ad435..f04b89955fea 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -585,7 +585,7 @@ static int amdgpu_virt_write_vf2pf_data(struct amdgpu_device *adev) } vf2pf_info->checksum = amd_sriov_msg_checksum( - vf2pf_info, vf2pf_info->header.size, 0, 0); + vf2pf_info, sizeof(*vf2pf_info), 0, 0); return 0; } -- cgit v1.2.3 From f4322b9f8ad5f9f62add288c785d2e10bb6a5efe Mon Sep 17 00:00:00 2001 From: Yunxiang Li Date: Mon, 22 Apr 2024 14:59:02 -0400 Subject: drm/amdgpu: Fix two reset triggered in a row Some times a hang GPU causes multiple reset sources to schedule resets. The second source will be able to trigger an unnecessary reset if they schedule after we call amdgpu_device_stop_pending_resets. Move amdgpu_device_stop_pending_resets to after the reset is done. Since at this point the GPU is supposedly in a good state, any reset scheduled after this point would be a legitimate reset. Remove unnecessary and incorrect checks for amdgpu_in_reset that was kinda serving this purpose. Signed-off-by: Yunxiang Li Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 19 ++++++++++--------- drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 2 +- drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c | 2 +- drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c | 2 +- drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c | 2 +- 5 files changed, 14 insertions(+), 13 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 861ccff78af9..8befd10bf007 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -5070,8 +5070,6 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev, retry: amdgpu_amdkfd_pre_reset(adev); - amdgpu_device_stop_pending_resets(adev); - if (from_hypervisor) r = amdgpu_virt_request_full_gpu(adev, true); else @@ -5823,13 +5821,6 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ r, adev_to_drm(tmp_adev)->unique); tmp_adev->asic_reset_res = r; } - - if (!amdgpu_sriov_vf(tmp_adev)) - /* - * Drop all pending non scheduler resets. Scheduler resets - * were already dropped during drm_sched_stop - */ - amdgpu_device_stop_pending_resets(tmp_adev); } /* Actual ASIC resets if needed.*/ @@ -5851,6 +5842,16 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ goto retry; } + list_for_each_entry(tmp_adev, device_list_handle, reset_list) { + /* + * Drop any pending non scheduler resets queued before reset is done. + * Any reset scheduled after this point would be valid. Scheduler resets + * were already dropped during drm_sched_stop and no new ones can come + * in before drm_sched_start. + */ + amdgpu_device_stop_pending_resets(tmp_adev); + } + skip_hw_reset: /* Post ASIC reset for all devs .*/ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index f04b89955fea..d98d619fba97 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -599,7 +599,7 @@ static void amdgpu_virt_update_vf2pf_work_item(struct work_struct *work) if (ret) { adev->virt.vf2pf_update_retry_cnt++; if ((adev->virt.vf2pf_update_retry_cnt >= AMDGPU_VF2PF_UPDATE_MAX_RETRY_LIMIT) && - amdgpu_sriov_runtime(adev) && !amdgpu_in_reset(adev)) { + amdgpu_sriov_runtime(adev)) { amdgpu_ras_set_fed(adev, true); if (amdgpu_reset_domain_schedule(adev->reset_domain, &adev->virt.flr_work)) diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c index 0c7275bca8f7..c5ba9c4757a8 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c @@ -319,7 +319,7 @@ static int xgpu_ai_mailbox_rcv_irq(struct amdgpu_device *adev, switch (event) { case IDH_FLR_NOTIFICATION: - if (amdgpu_sriov_runtime(adev) && !amdgpu_in_reset(adev)) + if (amdgpu_sriov_runtime(adev)) WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain, &adev->virt.flr_work), "Failed to queue work! at %s", diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c index 1cb98e6c09c8..fb7cf4214e3a 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c @@ -358,7 +358,7 @@ static int xgpu_nv_mailbox_rcv_irq(struct amdgpu_device *adev, switch (event) { case IDH_FLR_NOTIFICATION: - if (amdgpu_sriov_runtime(adev) && !amdgpu_in_reset(adev)) + if (amdgpu_sriov_runtime(adev)) WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain, &adev->virt.flr_work), "Failed to queue work! at %s", diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c index 59f53c743362..14a065516ae4 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c @@ -560,7 +560,7 @@ static int xgpu_vi_mailbox_rcv_irq(struct amdgpu_device *adev, r = xgpu_vi_mailbox_rcv_msg(adev, IDH_FLR_NOTIFICATION); /* only handle FLR_NOTIFY now */ - if (!r && !amdgpu_in_reset(adev)) + if (!r) WARN_ONCE(!amdgpu_reset_domain_schedule(adev->reset_domain, &adev->virt.flr_work), "Failed to queue work! at %s", -- cgit v1.2.3 From 5434bc03f52de2ec57d6ce684b1853928f508cbc Mon Sep 17 00:00:00 2001 From: Victor Skvortsov Date: Sun, 19 May 2024 10:39:43 -0400 Subject: drm/amdgpu: Queue KFD reset workitem in VF FED The guest recovery sequence is buggy in Fatal Error when both FLR & KFD reset workitems are queued at the same time. In addition, FLR guest recovery sequence is out of order when PF/VF communication breaks due to a GPU fatal error As a temporary work around, perform a KFD style reset (Initiate reset request from the guest) inside the pf2vf thread on FED. Signed-off-by: Victor Skvortsov Reviewed-by: Zhigang Luo Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index d98d619fba97..3d5f58e76f2d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -602,7 +602,7 @@ static void amdgpu_virt_update_vf2pf_work_item(struct work_struct *work) amdgpu_sriov_runtime(adev)) { amdgpu_ras_set_fed(adev, true); if (amdgpu_reset_domain_schedule(adev->reset_domain, - &adev->virt.flr_work)) + &adev->kfd.reset_work)) return; else dev_err(adev->dev, "Failed to queue work! at %s", __func__); -- cgit v1.2.3 From e864180ee49b4d30e640fd1e1d852b86411420c9 Mon Sep 17 00:00:00 2001 From: Victor Skvortsov Date: Mon, 27 May 2024 16:10:43 -0400 Subject: drm/amdgpu: Add lock around VF RLCG interface flush_gpu_tlb may be called from another thread while device_gpu_recover is running. Both of these threads access registers through the VF RLCG interface during VF Full Access. Add a lock around this interface to prevent race conditions between these threads. Signed-off-by: Victor Skvortsov Reviewed-by: Zhigang Luo Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 6 ++++++ drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h | 2 ++ 3 files changed, 9 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index f5168b4c3b03..6711836054f9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -4049,6 +4049,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, mutex_init(&adev->grbm_idx_mutex); mutex_init(&adev->mn_lock); mutex_init(&adev->virt.vf_errors.lock); + mutex_init(&adev->virt.rlcg_reg_lock); hash_init(adev->mn_hash); mutex_init(&adev->psp.mutex); mutex_init(&adev->notifier_lock); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index 3d5f58e76f2d..a72683f83390 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -982,6 +982,9 @@ u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v, u32 f scratch_reg1 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg1; scratch_reg2 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg2; scratch_reg3 = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->scratch_reg3; + + mutex_lock(&adev->virt.rlcg_reg_lock); + if (reg_access_ctrl->spare_int) spare_int = (void __iomem *)adev->rmmio + 4 * reg_access_ctrl->spare_int; @@ -1038,6 +1041,9 @@ u32 amdgpu_virt_rlcg_reg_rw(struct amdgpu_device *adev, u32 offset, u32 v, u32 f } ret = readl(scratch_reg0); + + mutex_unlock(&adev->virt.rlcg_reg_lock); + return ret; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h index 642f1fd287d8..0ec246c74570 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h @@ -272,6 +272,8 @@ struct amdgpu_virt { /* the ucode id to signal the autoload */ uint32_t autoload_ucode_id; + + struct mutex rlcg_reg_lock; }; struct amdgpu_video_codec_info; -- cgit v1.2.3 From c8ad1bbbc2751063c7a5825911e58996ef849628 Mon Sep 17 00:00:00 2001 From: "Lin.Cao" Date: Fri, 31 May 2024 14:02:39 +0800 Subject: drm/amdgpu: fix failure mapping legacy queue when FLR Flag "mes.ring.shced.ready" will be set as true after mes hw init and set as false when mes hw fini to avoid duplicate initialization. But hw fini will not be called when function level reset, which will cause mes hw init be skipped during FLR, which will leads to mapping legacy queue fail. Set this flag as false when post reset will fix this issue. Signed-off-by: Lin.Cao Acked-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index a72683f83390..b56d243c53ab 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -822,6 +822,8 @@ void amdgpu_virt_post_reset(struct amdgpu_device *adev) */ adev->gfx.is_poweron = false; } + + adev->mes.ring.sched.ready = false; } bool amdgpu_virt_fw_load_skip_check(struct amdgpu_device *adev, uint32_t ucode_id) -- cgit v1.2.3