drm/amdgpu: refactor bad_page_work for corner case handling

When a poison is consumed on the guest before the guest receives the host's poison creation msg, a corner case may occur to have poison_handler complete processing earlier than it should to cause the guest to hang waiting for the req_bad_pages reply during a VF FLR, resulting in the VM becoming inaccessible in stress tests.

To fix this issue, this patch refactored the mailbox sequence by seperating the bad_page_work into two parts req_bad_pages_work and handle_bad_pages_work.
Old sequence:
  1.Stop data exchange work
  2.Guest sends MB_REQ_RAS_BAD_PAGES to host and keep polling for IDH_RAS_BAD_PAGES_READY
  3.If the IDH_RAS_BAD_PAGES_READY arrives within timeout limit, re-init the data exchange region for updated bad page info
    else timeout with error message
New sequence:
req_bad_pages_work:
  1.Stop data exhange work
  2.Guest sends MB_REQ_RAS_BAD_PAGES to host
Once Guest receives IDH_RAS_BAD_PAGES_READY event
handle_bad_pages_work:
  3.re-init the data exchange region for updated bad page info

Signed-off-by: Chenglei Xie <Chenglei.Xie@amd.com>
Reviewed-by: Shravan Kumar Gande <Shravankumar.Gande@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

1 files changed, 28 insertions, 4 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
index 48101a34e049..9a40107a0869 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
@@ -292,14 +292,32 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work)
 	}
 }
 
-static void xgpu_ai_mailbox_bad_pages_work(struct work_struct *work)
+static void xgpu_ai_mailbox_req_bad_pages_work(struct work_struct *work)
 {
-	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, bad_pages_work);
+	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, req_bad_pages_work);
 	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
 
 	if (down_read_trylock(&adev->reset_domain->sem)) {
 		amdgpu_virt_fini_data_exchange(adev);
 		amdgpu_virt_request_bad_pages(adev);
+		up_read(&adev->reset_domain->sem);
+	}
+}
+
+/**
+ * xgpu_ai_mailbox_handle_bad_pages_work - Reinitialize the data exchange region to get fresh bad page information
+ * @work: pointer to the work_struct
+ *
+ * This work handler is triggered when bad pages are ready, and it reinitializes
+ * the data exchange region to retrieve updated bad page information from the host.
+ */
+static void xgpu_ai_mailbox_handle_bad_pages_work(struct work_struct *work)
+{
+	struct amdgpu_virt *virt = container_of(work, struct amdgpu_virt, handle_bad_pages_work);
+	struct amdgpu_device *adev = container_of(virt, struct amdgpu_device, virt);
+
+	if (down_read_trylock(&adev->reset_domain->sem)) {
+		amdgpu_virt_fini_data_exchange(adev);
 		amdgpu_virt_init_data_exchange(adev);
 		up_read(&adev->reset_domain->sem);
 	}
@@ -327,10 +345,15 @@ static int xgpu_ai_mailbox_rcv_irq(struct amdgpu_device *adev,
 	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 
 	switch (event) {
+	case IDH_RAS_BAD_PAGES_READY:
+		xgpu_ai_mailbox_send_ack(adev);
+		if (amdgpu_sriov_runtime(adev))
+			schedule_work(&adev->virt.handle_bad_pages_work);
+		break;
 	case IDH_RAS_BAD_PAGES_NOTIFICATION:
 		xgpu_ai_mailbox_send_ack(adev);
 		if (amdgpu_sriov_runtime(adev))
-			schedule_work(&adev->virt.bad_pages_work);
+			schedule_work(&adev->virt.req_bad_pages_work);
 		break;
 	case IDH_UNRECOV_ERR_NOTIFICATION:
 		xgpu_ai_mailbox_send_ack(adev);
@@ -415,7 +438,8 @@ int xgpu_ai_mailbox_get_irq(struct amdgpu_device *adev)
 	}
 
 	INIT_WORK(&adev->virt.flr_work, xgpu_ai_mailbox_flr_work);
-	INIT_WORK(&adev->virt.bad_pages_work, xgpu_ai_mailbox_bad_pages_work);
+	INIT_WORK(&adev->virt.req_bad_pages_work, xgpu_ai_mailbox_req_bad_pages_work);
+	INIT_WORK(&adev->virt.handle_bad_pages_work, xgpu_ai_mailbox_handle_bad_pages_work);
 
 	return 0;
 }