diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 106 |
1 files changed, 70 insertions, 36 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index a9c133a09be5..61f589b6fe5e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -877,7 +877,7 @@ static int amdgpu_ras_block_match_default(struct amdgpu_ras_block_object *block_ if (!block_obj) return -EINVAL; - if (block_obj->block == block) + if (block_obj->ras_comm.block == block) return 0; return -EINVAL; @@ -1276,18 +1276,17 @@ static int amdgpu_ras_sysfs_remove_feature_node(struct amdgpu_device *adev) } int amdgpu_ras_sysfs_create(struct amdgpu_device *adev, - struct ras_fs_if *head) + struct ras_common_if *head) { - struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head->head); + struct ras_manager *obj = amdgpu_ras_find_obj(adev, head); if (!obj || obj->attr_inuse) return -EINVAL; get_obj(obj); - memcpy(obj->fs_data.sysfs_name, - head->sysfs_name, - sizeof(obj->fs_data.sysfs_name)); + snprintf(obj->fs_data.sysfs_name, sizeof(obj->fs_data.sysfs_name), + "%s_err_count", head->name); obj->sysfs_attr = (struct device_attribute){ .attr = { @@ -1594,9 +1593,9 @@ int amdgpu_ras_interrupt_dispatch(struct amdgpu_device *adev, } int amdgpu_ras_interrupt_remove_handler(struct amdgpu_device *adev, - struct ras_ih_if *info) + struct ras_common_if *head) { - struct ras_manager *obj = amdgpu_ras_find_obj(adev, &info->head); + struct ras_manager *obj = amdgpu_ras_find_obj(adev, head); struct ras_ih_data *data; if (!obj) @@ -1616,24 +1615,27 @@ int amdgpu_ras_interrupt_remove_handler(struct amdgpu_device *adev, } int amdgpu_ras_interrupt_add_handler(struct amdgpu_device *adev, - struct ras_ih_if *info) + struct ras_common_if *head) { - struct ras_manager *obj = amdgpu_ras_find_obj(adev, &info->head); + struct ras_manager *obj = amdgpu_ras_find_obj(adev, head); struct ras_ih_data *data; + struct amdgpu_ras_block_object *ras_obj; if (!obj) { /* in case we registe the IH before enable ras feature */ - obj = amdgpu_ras_create_obj(adev, &info->head); + obj = amdgpu_ras_create_obj(adev, head); if (!obj) return -EINVAL; } else get_obj(obj); + ras_obj = container_of(head, struct amdgpu_ras_block_object, ras_comm); + data = &obj->ih_data; /* add the callback.etc */ *data = (struct ras_ih_data) { .inuse = 0, - .cb = info->cb, + .cb = ras_obj->ras_cb, .element_size = sizeof(struct amdgpu_iv_entry), .rptr = 0, .wptr = 0, @@ -1662,10 +1664,7 @@ static int amdgpu_ras_interrupt_remove_all(struct amdgpu_device *adev) struct ras_manager *obj, *tmp; list_for_each_entry_safe(obj, tmp, &con->head, node) { - struct ras_ih_if info = { - .head = obj->head, - }; - amdgpu_ras_interrupt_remove_handler(adev, &info); + amdgpu_ras_interrupt_remove_handler(adev, &obj->head); } return 0; @@ -2301,6 +2300,7 @@ int amdgpu_ras_init(struct amdgpu_device *adev) if (!adev->gmc.xgmi.connected_to_cpu) { adev->nbio.ras = &nbio_v7_4_ras; amdgpu_ras_register_ras_block(adev, &adev->nbio.ras->ras_block); + adev->nbio.ras_if = &adev->nbio.ras->ras_block.ras_comm; } break; default: @@ -2397,11 +2397,10 @@ bool amdgpu_ras_is_poison_mode_supported(struct amdgpu_device *adev) } /* 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, - struct ras_fs_if *fs_info, - struct ras_ih_if *ih_info) +int amdgpu_ras_block_late_init(struct amdgpu_device *adev, + struct ras_common_if *ras_block) { + struct amdgpu_ras_block_object *ras_obj; struct amdgpu_ras *con = amdgpu_ras_get_context(adev); unsigned long ue_count, ce_count; int r; @@ -2429,15 +2428,16 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev, if (adev->in_suspend || amdgpu_in_reset(adev)) return 0; - if (ih_info->cb) { - r = amdgpu_ras_interrupt_add_handler(adev, ih_info); + ras_obj = container_of(ras_block, struct amdgpu_ras_block_object, ras_comm); + if (ras_obj->ras_cb) { + r = amdgpu_ras_interrupt_add_handler(adev, ras_block); if (r) - goto interrupt; + goto cleanup; } - r = amdgpu_ras_sysfs_create(adev, fs_info); + r = amdgpu_ras_sysfs_create(adev, ras_block); if (r) - goto sysfs; + goto interrupt; /* Those are the cached values at init. */ @@ -2447,27 +2447,34 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev, } return 0; -cleanup: - amdgpu_ras_sysfs_remove(adev, ras_block); -sysfs: - if (ih_info->cb) - amdgpu_ras_interrupt_remove_handler(adev, ih_info); + interrupt: + if (ras_obj->ras_cb) + amdgpu_ras_interrupt_remove_handler(adev, ras_block); +cleanup: amdgpu_ras_feature_enable(adev, ras_block, 0); return r; } +int amdgpu_ras_block_late_init_default(struct amdgpu_device *adev, + struct ras_common_if *ras_block) +{ + return amdgpu_ras_block_late_init(adev, ras_block); +} + /* helper function to remove ras fs node and interrupt handler */ -void amdgpu_ras_late_fini(struct amdgpu_device *adev, - struct ras_common_if *ras_block, - struct ras_ih_if *ih_info) +void amdgpu_ras_block_late_fini(struct amdgpu_device *adev, + struct ras_common_if *ras_block) { - if (!ras_block || !ih_info) + struct amdgpu_ras_block_object *ras_obj; + if (!ras_block) return; amdgpu_ras_sysfs_remove(adev, ras_block); - if (ih_info->cb) - amdgpu_ras_interrupt_remove_handler(adev, ih_info); + + ras_obj = container_of(ras_block, struct amdgpu_ras_block_object, ras_comm); + if (ras_obj->ras_cb) + amdgpu_ras_interrupt_remove_handler(adev, ras_block); } /* do some init work after IP late init as dependence. @@ -2520,6 +2527,33 @@ void amdgpu_ras_suspend(struct amdgpu_device *adev) amdgpu_ras_disable_all_features(adev, 1); } +int amdgpu_ras_late_init(struct amdgpu_device *adev) +{ + struct amdgpu_ras_block_list *node, *tmp; + struct amdgpu_ras_block_object *obj; + int r; + + list_for_each_entry_safe(node, tmp, &adev->ras_list, node) { + if (!node->ras_obj) { + dev_warn(adev->dev, "Warning: abnormal ras list node.\n"); + continue; + } + + obj = node->ras_obj; + if (obj->ras_late_init) { + r = obj->ras_late_init(adev, &obj->ras_comm); + if (r) { + dev_err(adev->dev, "%s failed to execute ras_late_init! ret:%d\n", + obj->ras_comm.name, r); + return r; + } + } else + amdgpu_ras_block_late_init_default(adev, &obj->ras_comm); + } + + return 0; +} + /* do some fini work before IP fini as dependence */ int amdgpu_ras_pre_fini(struct amdgpu_device *adev) { |