PF will do page retirement and reset VF in SRIOV.
Signed-off-by: Tao Zhou <tao.zh...@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c | 18 +++++++++++++-----
drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h | 1 +
drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c | 6 ++++++
drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h | 1 +
4 files changed, 21 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
index aad3c8b4c810..b603ab3bd138 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
@@ -123,12 +123,20 @@ int amdgpu_umc_poison_handler(struct amdgpu_device *adev,
};
struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head);
- ret =
- amdgpu_umc_do_page_retirement(adev, ras_error_status, NULL,
reset);
+ if (!amdgpu_sriov_vf(adev)) {
+ ret = amdgpu_umc_do_page_retirement(adev,
+ ras_error_status, NULL, reset);
- if (ret == AMDGPU_RAS_SUCCESS && obj) {
- obj->err_data.ue_count += err_data->ue_count;
- obj->err_data.ce_count += err_data->ce_count;
+ if (ret == AMDGPU_RAS_SUCCESS && obj) {
+ obj->err_data.ue_count += err_data->ue_count;
+ obj->err_data.ce_count += err_data->ce_count;
+ }
+ } else {
+ if (adev->virt.ops && adev->virt.ops->ras_poison_handler)
+ adev->virt.ops->ras_poison_handler(adev);
+ else
+ dev_warn(adev->dev,
+ "No ras_poison_handler interface in SRIOV!\n");
}
return ret;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
index 4534e6f70a4b..41d03ef417d0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
@@ -86,6 +86,7 @@ struct amdgpu_virt_ops {
void (*trans_msg)(struct amdgpu_device *adev, u32 req, u32 data1, u32
data2, u32 data3);
void (*ras_trigger_error)(struct amdgpu_device *adev,
struct ta_ras_trigger_error_input *info);
+ void (*ras_poison_handler)(struct amdgpu_device *adev);
};
/*
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
index 3b4c5162a237..7545d7c0f524 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
@@ -428,6 +428,11 @@ void amdgpu_virt_ras_trigger_error(struct amdgpu_device
*adev,
addr_lo, addr_hi);
}
+void amdgpu_virt_ras_poison_handler(struct amdgpu_device *adev)
+{
+ xgpu_ai_send_access_requests(adev, IDH_RAS_POISON);
+}
+
const struct amdgpu_virt_ops xgpu_ai_virt_ops = {
.req_full_gpu = xgpu_ai_request_full_gpu_access,
.rel_full_gpu = xgpu_ai_release_full_gpu_access,
@@ -436,4 +441,5 @@ const struct amdgpu_virt_ops xgpu_ai_virt_ops = {
.trans_msg = xgpu_ai_mailbox_trans_msg,
.req_init_data = xgpu_ai_request_init_data,
.ras_trigger_error = amdgpu_virt_ras_trigger_error,
+ .ras_poison_handler = amdgpu_virt_ras_poison_handler,
};
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h
b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h
index 0841d6632328..869dff6daddd 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.h
@@ -48,6 +48,7 @@ enum idh_request {
IDH_LOG_VF_ERROR = 200,
IDH_READY_TO_RESET = 201,
IDH_RAS_ERROR_INJECT = 202,
+ IDH_RAS_POISON = 203,
};
enum idh_event {
--
2.35.1
