drm/amdgpu: disable ras query and iject during gpu reset

added flag to ras context to indicate if ras query functionality is ready

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c

@@ -4168,6 +4168,8 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 	need_full_reset = job_signaled = false;
 	INIT_LIST_HEAD(&device_list);
 
+	amdgpu_ras_set_error_query_ready(adev, false);
+
 	dev_info(adev->dev, "GPU %s begin!\n",
 		(in_ras_intr && !use_baco) ? "jobs stop":"reset");
 
@@ -4224,6 +4226,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 	/* block all schedulers and reset given job's ring */
 	list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) {
 		if (tmp_adev != adev) {
+			amdgpu_ras_set_error_query_ready(tmp_adev, false);
 			amdgpu_device_lock_adev(tmp_adev, false);
 			if (!amdgpu_sriov_vf(tmp_adev))
 			                amdgpu_amdkfd_pre_reset(tmp_adev);

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c

@@ -80,6 +80,20 @@ atomic_t amdgpu_ras_in_intr = ATOMIC_INIT(0);
 static bool amdgpu_ras_check_bad_page(struct amdgpu_device *adev,
 				uint64_t addr);
 
+void amdgpu_ras_set_error_query_ready(struct amdgpu_device *adev, bool ready)
+{
+	if (adev)
+		amdgpu_ras_get_context(adev)->error_query_ready = ready;
+}
+
+bool amdgpu_ras_get_error_query_ready(struct amdgpu_device *adev)
+{
+	if (adev)
+		return amdgpu_ras_get_context(adev)->error_query_ready;
+
+	return false;
+}
+
 static ssize_t amdgpu_ras_debugfs_read(struct file *f, char __user *buf,
 					size_t size, loff_t *pos)
 {
@@ -281,7 +295,7 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f, const char __user *
 	struct ras_debug_if data;
 	int ret = 0;
 
-	if (amdgpu_ras_intr_triggered()) {
+	if (!amdgpu_ras_get_error_query_ready(adev)) {
 		DRM_WARN("RAS WARN: error injection currently inaccessible\n");
 		return size;
 	}
@@ -399,7 +413,7 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev,
 		.head = obj->head,
 	};
 
-	if (amdgpu_ras_intr_triggered())
+	if (!amdgpu_ras_get_error_query_ready(obj->adev))
 		return snprintf(buf, PAGE_SIZE,
 				"Query currently inaccessible\n");
 
@@ -1886,8 +1900,10 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev,
 	}
 
 	/* in resume phase, no need to create ras fs node */
-	if (adev->in_suspend || adev->in_gpu_reset)
+	if (adev->in_suspend || adev->in_gpu_reset) {
+		amdgpu_ras_set_error_query_ready(adev, true);
 		return 0;
+	}
 
 	if (ih_info->cb) {
 		r = amdgpu_ras_interrupt_add_handler(adev, ih_info);
@@ -1899,6 +1915,8 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev,
 	if (r)
 		goto sysfs;
 
+	amdgpu_ras_set_error_query_ready(adev, true);
+
 	return 0;
 cleanup:
 	amdgpu_ras_sysfs_remove(adev, ras_block);

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h

@@ -334,6 +334,8 @@ struct amdgpu_ras {
 	uint32_t flags;
 	bool reboot;
 	struct amdgpu_ras_eeprom_control eeprom_control;
+
+	bool error_query_ready;
 };
 
 struct ras_fs_data {
@@ -629,4 +631,6 @@ static inline void amdgpu_ras_intr_cleared(void)
 
 void amdgpu_ras_global_ras_isr(struct amdgpu_device *adev);
 
+void amdgpu_ras_set_error_query_ready(struct amdgpu_device *adev, bool ready);
+
 #endif