1 files changed, 19 insertions, 18 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 6573e1112462..78fd2c9a7b7d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -1940,7 +1940,7 @@ static int amdgpu_device_fw_loading(struct amdgpu_device *adev)
 			if (adev->ip_blocks[i].status.hw == true)
 				break;
 
-			if (adev->in_gpu_reset || adev->in_suspend) {
+			if (amdgpu_in_reset(adev) || adev->in_suspend) {
 				r = adev->ip_blocks[i].version->funcs->resume(adev);
 				if (r) {
 					DRM_ERROR("resume of IP block <%s> failed %d\n",
@@ -2117,7 +2117,7 @@ static bool amdgpu_device_check_vram_lost(struct amdgpu_device *adev)
 			AMDGPU_RESET_MAGIC_NUM))
 		return true;
 
-	if (!adev->in_gpu_reset)
+	if (!amdgpu_in_reset(adev))
 		return false;
 
 	/*
@@ -3053,6 +3053,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	mutex_init(&adev->mn_lock);
 	mutex_init(&adev->virt.vf_errors.lock);
 	hash_init(adev->mn_hash);
+	atomic_set(&adev->in_gpu_reset, 0);
 	mutex_init(&adev->lock_reset);
 	mutex_init(&adev->psp.mutex);
 	mutex_init(&adev->notifier_lock);
@@ -4200,16 +4201,14 @@ end:
 	return r;
 }
 
-static bool amdgpu_device_lock_adev(struct amdgpu_device *adev, bool trylock)
+static bool amdgpu_device_lock_adev(struct amdgpu_device *adev)
 {
-	if (trylock) {
-		if (!mutex_trylock(&adev->lock_reset))
-			return false;
-	} else
-		mutex_lock(&adev->lock_reset);
+	if (atomic_cmpxchg(&adev->in_gpu_reset, 0, 1) != 0)
+		return false;
+
+	mutex_lock(&adev->lock_reset);
 
 	atomic_inc(&adev->gpu_reset_counter);
-	adev->in_gpu_reset = true;
 	switch (amdgpu_asic_reset_method(adev)) {
 	case AMD_RESET_METHOD_MODE1:
 		adev->mp1_state = PP_MP1_STATE_SHUTDOWN;
@@ -4229,7 +4228,7 @@ static void amdgpu_device_unlock_adev(struct amdgpu_device *adev)
 {
 	amdgpu_vf_error_trans_all(adev);
 	adev->mp1_state = PP_MP1_STATE_NONE;
-	adev->in_gpu_reset = false;
+	atomic_set(&adev->in_gpu_reset, 0);
 	mutex_unlock(&adev->lock_reset);
 }
 
@@ -4340,12 +4339,14 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 	 * We always reset all schedulers for device and all devices for XGMI
 	 * hive so that should take care of them too.
 	 */
-	hive = amdgpu_get_xgmi_hive(adev, true);
-	if (hive && !mutex_trylock(&hive->reset_lock)) {
-		DRM_INFO("Bailing on TDR for s_job:%llx, hive: %llx as another already in progress",
-			  job ? job->base.id : -1, hive->hive_id);
-		mutex_unlock(&hive->hive_lock);
-		return 0;
+	hive = amdgpu_get_xgmi_hive(adev, false);
+	if (hive) {
+		if (atomic_cmpxchg(&hive->in_reset, 0, 1) != 0) {
+			DRM_INFO("Bailing on TDR for s_job:%llx, hive: %llx as another already in progress",
+				job ? job->base.id : -1, hive->hive_id);
+			return 0;
+		}
+		mutex_lock(&hive->hive_lock);
 	}
 
 	/*
@@ -4367,7 +4368,7 @@ 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 (!amdgpu_device_lock_adev(tmp_adev, !hive)) {
+		if (!amdgpu_device_lock_adev(tmp_adev)) {
 			DRM_INFO("Bailing on TDR for s_job:%llx, as another already in progress",
 				  job ? job->base.id : -1);
 			mutex_unlock(&hive->hive_lock);
@@ -4505,7 +4506,7 @@ skip_sched_resume:
 	}
 
 	if (hive) {
-		mutex_unlock(&hive->reset_lock);
+		atomic_set(&hive->in_reset, 0);
 		mutex_unlock(&hive->hive_lock);
 	}