From 084e2640e51626f413f85663e3ba7e32d4272477 Mon Sep 17 00:00:00 2001 From: Luben Tuikov Date: Thu, 11 Mar 2021 19:11:01 -0500 Subject: drm/amdgpu: Fix check for RAS support Use positive logic to check for RAS support. Rename the function to actually indicate what it is testing for. Essentially, make the function a predicate with the correct name. Cc: Stanley Yang Cc: Alexander Deucher Signed-off-by: Luben Tuikov Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 0e16683876aa..17652972fd49 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -1933,15 +1933,12 @@ int amdgpu_ras_request_reset_on_boot(struct amdgpu_device *adev, return 0; } -static int amdgpu_ras_check_asic_type(struct amdgpu_device *adev) +static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev) { - if (adev->asic_type != CHIP_VEGA10 && - adev->asic_type != CHIP_VEGA20 && - adev->asic_type != CHIP_ARCTURUS && - adev->asic_type != CHIP_SIENNA_CICHLID) - return 1; - else - return 0; + return adev->asic_type == CHIP_VEGA10 || + adev->asic_type == CHIP_VEGA20 || + adev->asic_type == CHIP_ARCTURUS || + adev->asic_type == CHIP_SIENNA_CICHLID; } /* @@ -1960,7 +1957,7 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev, *supported = 0; if (amdgpu_sriov_vf(adev) || !adev->is_atom_fw || - amdgpu_ras_check_asic_type(adev)) + !amdgpu_ras_asic_supported(adev)) return; if (amdgpu_atomfirmware_mem_ecc_supported(adev)) { -- cgit From 36000c7a51080840902d79e1558851076ecb7a96 Mon Sep 17 00:00:00 2001 From: Tian Tao Date: Wed, 24 Mar 2021 17:17:40 +0800 Subject: drm/amdgpu: Convert sysfs sprintf/snprintf family to sysfs_emit Fix the following coccicheck warning: drivers/gpu//drm/amd/amdgpu/amdgpu_ras.c:434:9-17: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_xgmi.c:220:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_xgmi.c:249:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/df_v3_6.c:208:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_psp.c:2973:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_vram_mgr.c:75:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_vram_mgr.c:112:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_vram_mgr.c:58:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_vram_mgr.c:93:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_vram_mgr.c:125:9-17: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_gtt_mgr.c:52:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_gtt_mgr.c:71:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_device.c:140:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_device.c:164:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_device.c:186:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_device.c:208:8-16: WARNING: use scnprintf or sprintf drivers/gpu//drm/amd/amdgpu/amdgpu_atombios.c:1916:8-16: WARNING: use scnprintf or sprintf Signed-off-by: Tian Tao Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 17652972fd49..c5a73b5c26d9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -431,15 +431,13 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev, }; if (!amdgpu_ras_get_error_query_ready(obj->adev)) - return snprintf(buf, PAGE_SIZE, - "Query currently inaccessible\n"); + return sysfs_emit(buf, "Query currently inaccessible\n"); if (amdgpu_ras_query_error_status(obj->adev, &info)) return -EINVAL; - return snprintf(buf, PAGE_SIZE, "%s: %lu\n%s: %lu\n", - "ue", info.ue_count, - "ce", info.ce_count); + return sysfs_emit(buf, "%s: %lu\n%s: %lu\n", "ue", info.ue_count, + "ce", info.ce_count); } /* obj begin */ -- cgit From 5a4345270474c886dceee48c5b54b5c2c07aa877 Mon Sep 17 00:00:00 2001 From: "Stanley.Yang" Date: Thu, 1 Apr 2021 19:10:54 +0800 Subject: drm/amdgpu: support sdma error injection Signed-off-by: Stanley.Yang Reivewed-by: Dennis Li Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index c5a73b5c26d9..32be6dd1fcd7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -925,6 +925,7 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev, ret = -EINVAL; break; case AMDGPU_RAS_BLOCK__UMC: + case AMDGPU_RAS_BLOCK__SDMA: case AMDGPU_RAS_BLOCK__MMHUB: case AMDGPU_RAS_BLOCK__PCIE_BIF: ret = psp_ras_trigger_error(&adev->psp, &block_info); -- cgit From f08726868c7543e0754212dcadae0d6911f721bd Mon Sep 17 00:00:00 2001 From: Bernard Zhao Date: Wed, 31 Mar 2021 06:12:03 -0700 Subject: drm/amd: cleanup coding style a bit Fix patch check warning: WARNING: suspect code indent for conditional statements (8, 17) + if (obj && obj->use < 0) { + DRM_ERROR("RAS ERROR: Unbalance obj(%s) use\n", obj->head.name); WARNING: braces {} are not necessary for single statement blocks + if (obj && obj->use < 0) { + DRM_ERROR("RAS ERROR: Unbalance obj(%s) use\n", obj->head.name); + } Signed-off-by: Bernard Zhao Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 32be6dd1fcd7..26458946145c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -447,11 +447,10 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev, static inline void put_obj(struct ras_manager *obj) { - if (obj && --obj->use == 0) + if (obj && (--obj->use == 0)) list_del(&obj->node); - if (obj && obj->use < 0) { - DRM_ERROR("RAS ERROR: Unbalance obj(%s) use\n", obj->head.name); - } + if (obj && (obj->use < 0)) + DRM_ERROR("RAS ERROR: Unbalance obj(%s) use\n", obj->head.name); } /* make one obj and return it. */ -- cgit From 75f06251c921baf99c003662c529c25ba9937b2d Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Mon, 8 Mar 2021 16:40:07 +0800 Subject: drm/amdgpu: initialze ras caps per paltform config Driver only manages GFX/SDMA/MMHUB RAS in platforms that gpu node is connected to cpu through XGMI, other than that, it queries VBIOS for RAS capabilities. Signed-off-by: Hawking Zhang Acked-by: Alex Deucher Reviewed-by: John Clements Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 35 ++++++++++++++++++++++----------- 1 file changed, 23 insertions(+), 12 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 26458946145c..1708045e2a0d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -1936,6 +1936,7 @@ static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev) return adev->asic_type == CHIP_VEGA10 || adev->asic_type == CHIP_VEGA20 || adev->asic_type == CHIP_ARCTURUS || + adev->asic_type == CHIP_ALDEBARAN || adev->asic_type == CHIP_SIENNA_CICHLID; } @@ -1958,19 +1959,29 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev, !amdgpu_ras_asic_supported(adev)) return; - if (amdgpu_atomfirmware_mem_ecc_supported(adev)) { - dev_info(adev->dev, "MEM ECC is active.\n"); - *hw_supported |= (1 << AMDGPU_RAS_BLOCK__UMC | - 1 << AMDGPU_RAS_BLOCK__DF); - } else - dev_info(adev->dev, "MEM ECC is not presented.\n"); + if (!adev->gmc.xgmi.connected_to_cpu) { + if (amdgpu_atomfirmware_mem_ecc_supported(adev)) { + dev_info(adev->dev, "MEM ECC is active.\n"); + *hw_supported |= (1 << AMDGPU_RAS_BLOCK__UMC | + 1 << AMDGPU_RAS_BLOCK__DF); + } else { + dev_info(adev->dev, "MEM ECC is not presented.\n"); + } - if (amdgpu_atomfirmware_sram_ecc_supported(adev)) { - dev_info(adev->dev, "SRAM ECC is active.\n"); - *hw_supported |= ~(1 << AMDGPU_RAS_BLOCK__UMC | - 1 << AMDGPU_RAS_BLOCK__DF); - } else - dev_info(adev->dev, "SRAM ECC is not presented.\n"); + if (amdgpu_atomfirmware_sram_ecc_supported(adev)) { + dev_info(adev->dev, "SRAM ECC is active.\n"); + *hw_supported |= ~(1 << AMDGPU_RAS_BLOCK__UMC | + 1 << AMDGPU_RAS_BLOCK__DF); + } else { + dev_info(adev->dev, "SRAM ECC is not presented.\n"); + } + } else { + /* driver only manages a few IP blocks RAS feature + * when GPU is connected cpu through XGMI */ + *hw_supported |= (1 << AMDGPU_RAS_BLOCK__GFX | + 1 << AMDGPU_RAS_BLOCK__SDMA | + 1 << AMDGPU_RAS_BLOCK__MMHUB); + } /* hw_supported needs to be aligned with RAS block mask. */ *hw_supported &= AMDGPU_RAS_BLOCK_MASK; -- cgit From 6e36f23193cc870856a41e87281f62fb2b04bd1f Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Fri, 2 Apr 2021 14:39:36 +0800 Subject: drm/amdgpu: split nbio callbacks into ras and non-ras ones nbio ras is not managed by gpu driver when gpu is connected to cpu through xgmi. split nbio callbacks into ras and non-ras ones so gpu driver only initializes nbio ras callbacks when it manages nbio ras. Signed-off-by: Hawking Zhang Reviewed-by: Dennis Li Reviewed-by: John Clements Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 30 ++++++++++++++++++++++++------ 1 file changed, 24 insertions(+), 6 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 1708045e2a0d..ac3f4c3266bc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -804,8 +804,9 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, adev->mmhub.funcs->query_ras_error_status(adev); break; case AMDGPU_RAS_BLOCK__PCIE_BIF: - if (adev->nbio.funcs->query_ras_error_count) - adev->nbio.funcs->query_ras_error_count(adev, &err_data); + if (adev->nbio.ras_funcs && + adev->nbio.ras_funcs->query_ras_error_count) + adev->nbio.ras_funcs->query_ras_error_count(adev, &err_data); break; case AMDGPU_RAS_BLOCK__XGMI_WAFL: amdgpu_xgmi_query_ras_error_count(adev, &err_data); @@ -2030,14 +2031,31 @@ int amdgpu_ras_init(struct amdgpu_device *adev) /* Might need get this flag from vbios. */ con->flags = RAS_DEFAULT_FLAGS; - if (adev->nbio.funcs->init_ras_controller_interrupt) { - r = adev->nbio.funcs->init_ras_controller_interrupt(adev); + /* initialize nbio ras function ahead of any other + * ras functions so hardware fatal error interrupt + * can be enabled as early as possible */ + switch (adev->asic_type) { + case CHIP_VEGA20: + case CHIP_ARCTURUS: + case CHIP_ALDEBARAN: + if (!adev->gmc.xgmi.connected_to_cpu) + adev->nbio.ras_funcs = &nbio_v7_4_ras_funcs; + break; + default: + /* nbio ras is not available */ + break; + } + + if (adev->nbio.ras_funcs && + adev->nbio.ras_funcs->init_ras_controller_interrupt) { + r = adev->nbio.ras_funcs->init_ras_controller_interrupt(adev); if (r) goto release_con; } - if (adev->nbio.funcs->init_ras_err_event_athub_interrupt) { - r = adev->nbio.funcs->init_ras_err_event_athub_interrupt(adev); + if (adev->nbio.ras_funcs && + adev->nbio.ras_funcs->init_ras_err_event_athub_interrupt) { + r = adev->nbio.ras_funcs->init_ras_err_event_athub_interrupt(adev); if (r) goto release_con; } -- cgit From 52137ca8526cdeceab7651c314ce68ac49963512 Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Thu, 18 Mar 2021 20:18:19 +0800 Subject: drm/amdgpu: move xgmi ras functions to xgmi_ras_funcs xgmi ras is not managed by gpu driver when gpu is connected to cpu through xgmi. move all xgmi ras functions to xgmi_ras_funcs so gpu driver only initializes xgmi ras functions when it manages xgmi ras. Signed-off-by: Hawking Zhang Reviewed-by: Dennis Li Reviewed-by: John Clements Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index ac3f4c3266bc..172738cc99db 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -809,7 +809,9 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, adev->nbio.ras_funcs->query_ras_error_count(adev, &err_data); break; case AMDGPU_RAS_BLOCK__XGMI_WAFL: - amdgpu_xgmi_query_ras_error_count(adev, &err_data); + if (adev->gmc.xgmi.ras_funcs && + adev->gmc.xgmi.ras_funcs->query_ras_error_count) + adev->gmc.xgmi.ras_funcs->query_ras_error_count(adev, &err_data); break; default: break; -- cgit From 49070c4ea3d97b76c5666466efb35dcc42c6c8fd Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Wed, 17 Mar 2021 19:17:52 +0800 Subject: drm/amdgpu: split umc callbacks to ras and non-ras ones umc ras is not managed by gpu driver when gpu is connected to cpu through xgmi. split umc callbacks into ras and non-ras ones so gpu driver only initializes umc ras callbacks when it manages umc ras. Signed-off-by: Hawking Zhang Reviewed-by: Dennis Li Reviewed-by: John Clements Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 172738cc99db..459a470744f4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -774,13 +774,15 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, switch (info->head.block) { case AMDGPU_RAS_BLOCK__UMC: - if (adev->umc.funcs->query_ras_error_count) - adev->umc.funcs->query_ras_error_count(adev, &err_data); + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->query_ras_error_count) + adev->umc.ras_funcs->query_ras_error_count(adev, &err_data); /* umc query_ras_error_address is also responsible for clearing * error status */ - if (adev->umc.funcs->query_ras_error_address) - adev->umc.funcs->query_ras_error_address(adev, &err_data); + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->query_ras_error_address) + adev->umc.ras_funcs->query_ras_error_address(adev, &err_data); break; case AMDGPU_RAS_BLOCK__SDMA: if (adev->sdma.funcs->query_ras_error_count) { -- cgit From 8bc7b360ad4b0a090380d7548dbf24a627f0b035 Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Fri, 19 Mar 2021 15:50:14 +0800 Subject: drm/amdgpu: split mmhub callbacks into ras and non-ras ones mmhub ras is only avaiable in cerntain mmhub ip generation. Signed-off-by: Hawking Zhang Reviewed-by: Dennis Li Reviewed-by: John Clements Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 459a470744f4..b55f470eb747 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -799,11 +799,13 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, adev->gfx.funcs->query_ras_error_status(adev); break; case AMDGPU_RAS_BLOCK__MMHUB: - if (adev->mmhub.funcs->query_ras_error_count) - adev->mmhub.funcs->query_ras_error_count(adev, &err_data); + if (adev->mmhub.ras_funcs && + adev->mmhub.ras_funcs->query_ras_error_count) + adev->mmhub.ras_funcs->query_ras_error_count(adev, &err_data); - if (adev->mmhub.funcs->query_ras_error_status) - adev->mmhub.funcs->query_ras_error_status(adev); + if (adev->mmhub.ras_funcs && + adev->mmhub.ras_funcs->query_ras_error_status) + adev->mmhub.ras_funcs->query_ras_error_status(adev); break; case AMDGPU_RAS_BLOCK__PCIE_BIF: if (adev->nbio.ras_funcs && @@ -857,8 +859,9 @@ int amdgpu_ras_reset_error_status(struct amdgpu_device *adev, adev->gfx.funcs->reset_ras_error_status(adev); break; case AMDGPU_RAS_BLOCK__MMHUB: - if (adev->mmhub.funcs->reset_ras_error_count) - adev->mmhub.funcs->reset_ras_error_count(adev); + if (adev->mmhub.ras_funcs && + adev->mmhub.ras_funcs->reset_ras_error_count) + adev->mmhub.ras_funcs->reset_ras_error_count(adev); break; case AMDGPU_RAS_BLOCK__SDMA: if (adev->sdma.funcs->reset_ras_error_count) @@ -1515,8 +1518,9 @@ static void amdgpu_ras_error_status_query(struct amdgpu_device *adev, adev->gfx.funcs->query_ras_error_status(adev); break; case AMDGPU_RAS_BLOCK__MMHUB: - if (adev->mmhub.funcs->query_ras_error_status) - adev->mmhub.funcs->query_ras_error_status(adev); + if (adev->mmhub.ras_funcs && + adev->mmhub.ras_funcs->query_ras_error_status) + adev->mmhub.ras_funcs->query_ras_error_status(adev); break; default: break; -- cgit From 719a9b332305b8c4b91805c4bedee27ce82ee916 Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Fri, 19 Mar 2021 16:59:09 +0800 Subject: drm/amdgpu: split gfx callbacks into ras and non-ras ones gfx ras is only available in cerntain ip generations. Signed-off-by: Hawking Zhang Reviewed-by: Dennis Li Reviewed-by: John Clements Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 30 ++++++++++++++++++------------ 1 file changed, 18 insertions(+), 12 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index b55f470eb747..1d905bcbc1ac 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -792,11 +792,13 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, } break; case AMDGPU_RAS_BLOCK__GFX: - if (adev->gfx.funcs->query_ras_error_count) - adev->gfx.funcs->query_ras_error_count(adev, &err_data); + if (adev->gfx.ras_funcs && + adev->gfx.ras_funcs->query_ras_error_count) + adev->gfx.ras_funcs->query_ras_error_count(adev, &err_data); - if (adev->gfx.funcs->query_ras_error_status) - adev->gfx.funcs->query_ras_error_status(adev); + if (adev->gfx.ras_funcs && + adev->gfx.ras_funcs->query_ras_error_status) + adev->gfx.ras_funcs->query_ras_error_status(adev); break; case AMDGPU_RAS_BLOCK__MMHUB: if (adev->mmhub.ras_funcs && @@ -852,11 +854,13 @@ int amdgpu_ras_reset_error_status(struct amdgpu_device *adev, switch (block) { case AMDGPU_RAS_BLOCK__GFX: - if (adev->gfx.funcs->reset_ras_error_count) - adev->gfx.funcs->reset_ras_error_count(adev); + if (adev->gfx.ras_funcs && + adev->gfx.ras_funcs->reset_ras_error_count) + adev->gfx.ras_funcs->reset_ras_error_count(adev); - if (adev->gfx.funcs->reset_ras_error_status) - adev->gfx.funcs->reset_ras_error_status(adev); + if (adev->gfx.ras_funcs && + adev->gfx.ras_funcs->reset_ras_error_status) + adev->gfx.ras_funcs->reset_ras_error_status(adev); break; case AMDGPU_RAS_BLOCK__MMHUB: if (adev->mmhub.ras_funcs && @@ -926,8 +930,9 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev, switch (info->head.block) { case AMDGPU_RAS_BLOCK__GFX: - if (adev->gfx.funcs->ras_error_inject) - ret = adev->gfx.funcs->ras_error_inject(adev, info); + if (adev->gfx.ras_funcs && + adev->gfx.ras_funcs->ras_error_inject) + ret = adev->gfx.ras_funcs->ras_error_inject(adev, info); else ret = -EINVAL; break; @@ -1514,8 +1519,9 @@ static void amdgpu_ras_error_status_query(struct amdgpu_device *adev, */ switch (info->head.block) { case AMDGPU_RAS_BLOCK__GFX: - if (adev->gfx.funcs->query_ras_error_status) - adev->gfx.funcs->query_ras_error_status(adev); + if (adev->gfx.ras_funcs && + adev->gfx.ras_funcs->query_ras_error_status) + adev->gfx.ras_funcs->query_ras_error_status(adev); break; case AMDGPU_RAS_BLOCK__MMHUB: if (adev->mmhub.ras_funcs && -- cgit From 134d16d50f0948f00e7172b509e869b6eaecf437 Mon Sep 17 00:00:00 2001 From: John Clements Date: Thu, 25 Mar 2021 17:10:10 +0800 Subject: drm/amdgpu: RAS harvest on driver load In event of RAS UE + warm reset, error counters shall be harvested and cleared on driver load Reviewed-by: Hawking Zhang Signed-off-by: John Clements Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 29 +++++++++++++++++++++++++++++ 1 file changed, 29 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 1d905bcbc1ac..b0fe5885e4c6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -2090,6 +2090,32 @@ release_con: return r; } +static int amdgpu_persistent_edc_harvesting_supported(struct amdgpu_device *adev) +{ + if (adev->gmc.xgmi.connected_to_cpu) + return 1; + return 0; +} + +static int amdgpu_persistent_edc_harvesting(struct amdgpu_device *adev, + struct ras_common_if *ras_block) +{ + struct ras_query_if info = { + .head = *ras_block, + }; + + if (!amdgpu_persistent_edc_harvesting_supported(adev)) + return 0; + + if (amdgpu_ras_query_error_status(adev, &info) != 0) + DRM_WARN("RAS init harvest failure"); + + if (amdgpu_ras_reset_error_status(adev, ras_block->block) != 0) + DRM_WARN("RAS init harvest reset failure"); + + return 0; +} + /* helper function to handle common stuff in ip late init phase */ int amdgpu_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block, @@ -2119,6 +2145,9 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev, return r; } + /* check for errors on warm reset edc persisant supported ASIC */ + amdgpu_persistent_edc_harvesting(adev, ras_block); + /* in resume phase, no need to create ras fs node */ if (adev->in_suspend || amdgpu_in_reset(adev)) return 0; -- cgit From cbb8f989d5a07cb3e39e9c149a6f89d6c83432aa Mon Sep 17 00:00:00 2001 From: John Clements Date: Fri, 9 Apr 2021 17:25:29 +0800 Subject: drm/amdgpu: page retire over debugfs mechanism added support in RAS debugfs to add bad page for isolated page retirement testing Reviewed-by: Hawking Zhang Signed-off-by: John Clements Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 67 +++++++++++++++++++++++++++++++++ 1 file changed, 67 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index b0fe5885e4c6..0541196ae1ed 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -99,6 +99,49 @@ static bool amdgpu_ras_get_error_query_ready(struct amdgpu_device *adev) return false; } +static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t address) +{ + struct ras_err_data err_data = {0, 0, 0, NULL}; + struct eeprom_table_record err_rec; + + if ((address >= adev->gmc.mc_vram_size) || + (address >= RAS_UMC_INJECT_ADDR_LIMIT)) { + dev_warn(adev->dev, + "RAS WARN: input address 0x%llx is invalid.\n", + address); + return -EINVAL; + } + + if (amdgpu_ras_check_bad_page(adev, address)) { + dev_warn(adev->dev, + "RAS WARN: 0x%llx has been marked as bad page!\n", + address); + return 0; + } + + memset(&err_rec, 0x0, sizeof(struct eeprom_table_record)); + + err_rec.address = address; + err_rec.retired_page = address >> AMDGPU_GPU_PAGE_SHIFT; + err_rec.ts = (uint64_t)ktime_get_real_seconds(); + err_rec.err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE; + + err_data.err_addr = &err_rec; + err_data.err_addr_cnt = 1; + + if (amdgpu_bad_page_threshold != 0) { + amdgpu_ras_add_bad_pages(adev, err_data.err_addr, + err_data.err_addr_cnt); + amdgpu_ras_save_bad_pages(adev); + } + + dev_warn(adev->dev, "WARNING: THIS IS ONLY FOR TEST PURPOSES AND WILL CORRUPT RAS EEPROM\n"); + dev_warn(adev->dev, "Clear EEPROM:\n"); + dev_warn(adev->dev, " echo 1 > /sys/kernel/debug/dri/0/ras/ras_eeprom_reset\n"); + + return 0; +} + static ssize_t amdgpu_ras_debugfs_read(struct file *f, char __user *buf, size_t size, loff_t *pos) { @@ -178,11 +221,25 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, op = 1; else if (sscanf(str, "inject %32s %8s", block_name, err) == 2) op = 2; + else if (sscanf(str, "retire_page") == 0) + op = 3; else if (str[0] && str[1] && str[2] && str[3]) /* ascii string, but commands are not matched. */ return -EINVAL; if (op != -1) { + + if (op == 3) { + if (sscanf(str, "%*s %llu", &address) != 1) + if (sscanf(str, "%*s 0x%llx", &address) != 1) + return -EINVAL; + + data->op = op; + data->inject.address = address; + + return 0; + } + if (amdgpu_ras_find_block_id_by_name(block_name, &block_id)) return -EINVAL; @@ -310,6 +367,16 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f, const char __user * if (ret) return -EINVAL; + if (data.op == 3) + { + ret = amdgpu_reserve_page_direct(adev, data.inject.address); + + if (ret) + return size; + else + return ret; + } + if (!amdgpu_ras_is_supported(adev, data.head.block)) return -EINVAL; -- cgit From 6cb7a1d40acb2425b9dabb99cdaf522bf47295cb Mon Sep 17 00:00:00 2001 From: Luben Tuikov Date: Tue, 13 Apr 2021 07:49:11 -0400 Subject: drm/amdgpu: Fix a bug for input with double sscanf Remove double-sscanf to scan for %llu and 0x%llx, as that is not going to work! The %llu will consume the "0" in "0x" of your input, and the hex value you think you're entering will always be 0. That is, a valid hex value can never be consumed. On the other hand, just entering a hex number without leading 0x will either be scanned as a string and not match, for instance FAB123, or the leading decimal portion is scanned as the %llu, for instance 123FAB will be scanned as 123, which is not correct. Thus remove the first %llu scan and leave only the %llx scan, removing the leading 0x since %llx can scan either. Addresses are usually always hex values, so this suffices. Cc: Alexander Deucher Cc: Xinhui Pan Cc: Hawking Zhang Signed-off-by: Luben Tuikov Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 0541196ae1ed..9041453465f1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -230,9 +230,8 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, if (op != -1) { if (op == 3) { - if (sscanf(str, "%*s %llu", &address) != 1) - if (sscanf(str, "%*s 0x%llx", &address) != 1) - return -EINVAL; + if (sscanf(str, "%*s %llx", &address) != 1) + return -EINVAL; data->op = op; data->inject.address = address; @@ -255,11 +254,9 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, data->op = op; if (op == 2) { - if (sscanf(str, "%*s %*s %*s %u %llu %llu", - &sub_block, &address, &value) != 3) - if (sscanf(str, "%*s %*s %*s 0x%x 0x%llx 0x%llx", - &sub_block, &address, &value) != 3) - return -EINVAL; + if (sscanf(str, "%*s %*s %*s %x %llx %llx", + &sub_block, &address, &value) != 3) + return -EINVAL; data->head.sub_block_index = sub_block; data->inject.address = address; data->inject.value = value; -- cgit From 80b0cd0fb97ec3c02b0f2c4dc26fe3e8be94dc01 Mon Sep 17 00:00:00 2001 From: Luben Tuikov Date: Tue, 13 Apr 2021 08:31:59 -0400 Subject: drm/amdgpu: Fix a bug in checking the result of reserve page Fix if (ret) --> if (!ret), a bug, for "retire_page", which caused the kernel to recall the method with *pos == end of file, and that bounced back with error. On the first run, we advanced *pos, but returned 0 back to fs layer, also a bug. Fix the logic of the check of the result of amdgpu_reserve_page_direct()--it is 0 on success, and non-zero on error, not the other way around. This patch fixes this bug. Cc: Alexander Deucher Cc: John Clements Cc: Hawking Zhang Signed-off-by: Luben Tuikov Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 9041453465f1..3ca6b51f0c9c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -114,7 +114,7 @@ static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t addre if (amdgpu_ras_check_bad_page(adev, address)) { dev_warn(adev->dev, - "RAS WARN: 0x%llx has been marked as bad page!\n", + "RAS WARN: 0x%llx has already been marked as bad page!\n", address); return 0; } @@ -228,7 +228,6 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, return -EINVAL; if (op != -1) { - if (op == 3) { if (sscanf(str, "%*s %llx", &address) != 1) return -EINVAL; @@ -364,11 +363,9 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f, const char __user * if (ret) return -EINVAL; - if (data.op == 3) - { + if (data.op == 3) { ret = amdgpu_reserve_page_direct(adev, data.inject.address); - - if (ret) + if (!ret) return size; else return ret; -- cgit From 7fb6407145479dad41deecb5919d248539e38667 Mon Sep 17 00:00:00 2001 From: Luben Tuikov Date: Tue, 13 Apr 2021 08:40:18 -0400 Subject: drm/amdgpu: Add bad_page_cnt_threshold to debugfs Add bad_page_cnt_threshold to debugfs, an optional file system used for debugging, for reporting purposes only--it usually matches the size of EEPROM but may be different depending on the "bad_page_threshold" kernel module option. The "bad_page_cnt_threshold" is a dynamically computed value. It depends on three things: the VRAM size; the size of the EEPROM (or the size allocated to the RAS table therein); and the "bad_page_threshold" module parameter. It is a dynamically computed value, when the amdgpu module is run, on which further parameters and logic depend, and as such it is helpful to see the dynamically computed value in debugfs. Cc: Alexander Deucher Cc: Hawking Zhang Signed-off-by: Luben Tuikov Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 3ca6b51f0c9c..30cda4b8a461 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -1263,6 +1263,8 @@ static struct dentry *amdgpu_ras_debugfs_create_ctrl_node(struct amdgpu_device * &amdgpu_ras_debugfs_ctrl_ops); debugfs_create_file("ras_eeprom_reset", S_IWUGO | S_IRUGO, dir, adev, &amdgpu_ras_debugfs_eeprom_ops); + debugfs_create_u32("bad_page_cnt_threshold", 0444, dir, + &con->bad_page_cnt_threshold); /* * After one uncorrectable error happens, usually GPU recovery will -- cgit From 737c375b88eb33102164ab38fae7d22a501a47d8 Mon Sep 17 00:00:00 2001 From: Luben Tuikov Date: Tue, 13 Apr 2021 08:48:59 -0400 Subject: drm/amdgpu: Fix kernel-doc for the RAS sysfs interface Imporve the kernel-doc for the RAS sysfs interface. Fix the grammar, fix the context. Cc: Alexander Deucher Cc: Hawking Zhang Signed-off-by: Luben Tuikov Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 47 +++++++++++++++++---------------- 1 file changed, 24 insertions(+), 23 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 30cda4b8a461..44dfb3613e37 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -274,7 +274,7 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, /** * DOC: AMDGPU RAS debugfs control interface * - * It accepts struct ras_debug_if who has two members. + * The control interface accepts struct ras_debug_if which has two members. * * First member: ras_debug_if::head or ras_debug_if::inject. * @@ -299,32 +299,33 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, * * How to use the interface? * - * Programs + * In a program * - * Copy the struct ras_debug_if in your codes and initialize it. - * Write the struct to the control node. + * Copy the struct ras_debug_if in your code and initialize it. + * Write the struct to the control interface. * - * Shells + * From shell * * .. code-block:: bash * - * echo op block [error [sub_block address value]] > .../ras/ras_ctrl + * echo "disable " > /sys/kernel/debug/dri//ras/ras_ctrl + * echo "enable " > /sys/kernel/debug/dri//ras/ras_ctrl + * echo "inject
> /sys/kernel/debug/dri//ras/ras_ctrl * - * Parameters: + * Where N, is the card which you want to affect. * - * op: disable, enable, inject - * disable: only block is needed - * enable: block and error are needed - * inject: error, address, value are needed - * block: umc, sdma, gfx, ......... + * "disable" requires only the block. + * "enable" requires the block and error type. + * "inject" requires the block, error type, address, and value. + * The block is one of: umc, sdma, gfx, etc. * see ras_block_string[] for details - * error: ue, ce - * ue: multi_uncorrectable - * ce: single_correctable - * sub_block: - * sub block index, pass 0 if there is no sub block + * The error type is one of: ue, ce, where, + * ue is multi-uncorrectable + * ce is single-correctable + * The sub-block is a the sub-block index, pass 0 if there is no sub-block. + * The address and value are hexadecimal numbers, leading 0x is optional. * - * here are some examples for bash commands: + * For instance, * * .. code-block:: bash * @@ -332,17 +333,17 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, * echo inject umc ce 0 0 0 > /sys/kernel/debug/dri/0/ras/ras_ctrl * echo disable umc > /sys/kernel/debug/dri/0/ras/ras_ctrl * - * How to check the result? + * How to check the result of the operation? * - * For disable/enable, please check ras features at + * To check disable/enable, see "ras" features at, * /sys/class/drm/card[0/1/2...]/device/ras/features * - * For inject, please check corresponding err count at - * /sys/class/drm/card[0/1/2...]/device/ras/[gfx/sdma/...]_err_count + * To check inject, see the corresponding error count at, + * /sys/class/drm/card[0/1/2...]/device/ras/[gfx|sdma|umc|...]_err_count * * .. note:: * Operations are only allowed on blocks which are supported. - * Please check ras mask at /sys/module/amdgpu/parameters/ras_mask + * Check the "ras" mask at /sys/module/amdgpu/parameters/ras_mask * to see which blocks support RAS on a particular asic. * */ -- cgit From 546aa546b05aef582956f44ad917c0eee2cb8027 Mon Sep 17 00:00:00 2001 From: Luben Tuikov Date: Wed, 14 Apr 2021 11:17:01 -0400 Subject: drm/amdgpu: Add double-sscanf but invert Add back the double-sscanf so that both decimal and hexadecimal values could be read in, but this time invert the scan so that hexadecimal format with a leading 0x is tried first, and if that fails, then try decimal format. Also use a logical-AND instead of nesting double if-conditional. See commit "drm/amdgpu: Fix a bug for input with double sscanf" Cc: Alexander Deucher Cc: Hawking Zhang Signed-off-by: Luben Tuikov Reviewed-by: John Clements Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 44dfb3613e37..38a691a3b600 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -229,7 +229,8 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, if (op != -1) { if (op == 3) { - if (sscanf(str, "%*s %llx", &address) != 1) + if (sscanf(str, "%*s 0x%llx", &address) != 1 && + sscanf(str, "%*s %llu", &address) != 1) return -EINVAL; data->op = op; @@ -253,7 +254,9 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, data->op = op; if (op == 2) { - if (sscanf(str, "%*s %*s %*s %x %llx %llx", + if (sscanf(str, "%*s %*s %*s 0x%x 0x%llx 0x%llx", + &sub_block, &address, &value) != 3 && + sscanf(str, "%*s %*s %*s %u %llu %llu", &sub_block, &address, &value) != 3) return -EINVAL; data->head.sub_block_index = sub_block; -- cgit From 6df23f4c5cef5b6d8879720ef7c842adffc565c9 Mon Sep 17 00:00:00 2001 From: Dennis Li Date: Fri, 16 Apr 2021 22:41:11 +0800 Subject: drm/amdgpu: fix a error injection failed issue because "sscanf(str, "retire_page")" always return 0, if application use the raw data for error injection, it always wrongly falls into "op == 3". Change to use strstr instead. Signed-off-by: Dennis Li Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 38a691a3b600..7438d4e84776 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -221,7 +221,7 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f, op = 1; else if (sscanf(str, "inject %32s %8s", block_name, err) == 2) op = 2; - else if (sscanf(str, "retire_page") == 0) + else if (strstr(str, "retire_page") != NULL) op = 3; else if (str[0] && str[1] && str[2] && str[3]) /* ascii string, but commands are not matched. */ -- cgit From 1f0d8e3781f40c6cae7cb68a4cccfc54dd4ad3a1 Mon Sep 17 00:00:00 2001 From: Mukul Joshi Date: Wed, 24 Mar 2021 11:36:33 -0400 Subject: drm/amdgpu: Reset RAS error count and status regs Reset the RAS error count and error status registers after reading to prevent over reporting error counts on Aldebaran. Signed-off-by: Mukul Joshi Reviewed-By: John Clements Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 7438d4e84776..b0d2fc9454ca 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -501,6 +501,12 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev, if (amdgpu_ras_query_error_status(obj->adev, &info)) return -EINVAL; + + if (obj->adev->asic_type == CHIP_ALDEBARAN) { + if (amdgpu_ras_reset_error_status(obj->adev, info.head.block)) + DRM_WARN("Failed to reset error counter and error status"); + } + return sysfs_emit(buf, "%s: %lu\n%s: %lu\n", "ue", info.ue_count, "ce", info.ce_count); } -- cgit