drm/amdgpu: refine codes to avoid reentering GPU recovery

if other threads have holden the reset lock, recovery will
fail to try_lock. Therefore we introduce atomic hive->in_reset
and adev->in_gpu_reset, to avoid reentering GPU recovery.

v2:
drop "? true : false" in the definition of amdgpu_in_reset

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

28 files changed, 127 insertions, 122 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 54666eea1863..72f2a9aeaf0e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -949,7 +949,7 @@ struct amdgpu_device {
 	bool                            in_suspend;
 	bool				in_hibernate;
 
-	bool                            in_gpu_reset;
+	atomic_t 			in_gpu_reset;
 	enum pp_mp1_state               mp1_state;
 	struct mutex  lock_reset;
 	struct amdgpu_doorbell_index doorbell_index;
@@ -1267,4 +1267,8 @@ static inline bool amdgpu_is_tmz(struct amdgpu_device *adev)
        return adev->gmc.tmz_enabled;
 }
 
+static inline int amdgpu_in_reset(struct amdgpu_device *adev)
+{
+	return atomic_read(&adev->in_gpu_reset);
+}
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
index bf927f432506..b0dcc800251e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
@@ -542,7 +542,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	uint32_t temp;
 	struct v10_compute_mqd *m = get_mqd(mqd);
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EIO;
 
 #if 0
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
index 744366c7ee85..275f20399373 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
@@ -423,7 +423,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	unsigned long flags, end_jiffies;
 	int retry;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EIO;
 
 	acquire_queue(kgd, pipe_id, queue_id);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
index feab4cc6e836..4997189d8b36 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
@@ -419,7 +419,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	int retry;
 	struct vi_mqd *m = get_mqd(mqd);
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EIO;
 
 	acquire_queue(kgd, pipe_id, queue_id);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
index a755f63514e1..d08839d9491f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
@@ -552,7 +552,7 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	uint32_t temp;
 	struct v9_mqd *m = get_mqd(mqd);
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EIO;
 
 	acquire_queue(kgd, pipe_id, queue_id);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index 35fed75a4397..79b397800cbc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -127,7 +127,7 @@ static unsigned int amdgpu_debugfs_autodump_poll(struct file *file, struct poll_
 
 	poll_wait(file, &adev->autodump.gpu_hang, poll_table);
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return POLLIN | POLLRDNORM | POLLWRNORM;
 
 	return 0;
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);
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index a819360a4b6a..ebea3f655826 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -719,7 +719,7 @@ uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg)
 	 *
 	 * also don't wait anymore for IRQ context
 	 * */
-	if (r < 1 && (adev->in_gpu_reset || in_interrupt()))
+	if (r < 1 && (amdgpu_in_reset(adev) || in_interrupt()))
 		goto failed_kiq_read;
 
 	might_sleep();
@@ -777,7 +777,7 @@ void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v)
 	 *
 	 * also don't wait anymore for IRQ context
 	 * */
-	if (r < 1 && (adev->in_gpu_reset || in_interrupt()))
+	if (r < 1 && (amdgpu_in_reset(adev) || in_interrupt()))
 		goto failed_kiq_write;
 
 	might_sleep();
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index aa1e77c60c0a..116a89990f39 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -1869,7 +1869,7 @@ static int psp_load_smu_fw(struct psp_context *psp)
 		return 0;
 
 
-	if (adev->in_gpu_reset && ras && ras->supported) {
+	if (amdgpu_in_reset(adev) && ras && ras->supported) {
 		ret = amdgpu_dpm_set_mp1_state(adev, PP_MP1_STATE_UNLOAD);
 		if (ret) {
 			DRM_WARN("Failed to set MP1 state prepare for reload\n");
@@ -1984,7 +1984,7 @@ static int psp_load_fw(struct amdgpu_device *adev)
 	int ret;
 	struct psp_context *psp = &adev->psp;
 
-	if (amdgpu_sriov_vf(adev) && adev->in_gpu_reset) {
+	if (amdgpu_sriov_vf(adev) && amdgpu_in_reset(adev)) {
 		psp_ring_stop(psp, PSP_RING_TYPE__KM); /* should not destroy ring, only stop */
 		goto skip_memalloc;
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 866373c95cae..fa2c28ae9785 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -2081,7 +2081,7 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev,
 			amdgpu_ras_request_reset_on_boot(adev,
 					ras_block->block);
 			return 0;
-		} else if (adev->in_suspend || adev->in_gpu_reset) {
+		} else if (adev->in_suspend || amdgpu_in_reset(adev)) {
 			/* in resume phase, if fail to enable ras,
 			 * clean up all ras fs nodes, and disable ras */
 			goto cleanup;
@@ -2090,7 +2090,7 @@ 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 || amdgpu_in_reset(adev))
 		return 0;
 
 	if (ih_info->cb) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 3d2712a4cf4d..97449ab24c20 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -2098,7 +2098,7 @@ void amdgpu_ttm_set_buffer_funcs_status(struct amdgpu_device *adev, bool enable)
 	uint64_t size;
 	int r;
 
-	if (!adev->mman.initialized || adev->in_gpu_reset ||
+	if (!adev->mman.initialized || amdgpu_in_reset(adev) ||
 	    adev->mman.buffer_funcs_enabled == enable)
 		return;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
index 183743c5fb7b..5f851d17eef4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
@@ -628,7 +628,7 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
 	struct amdgpu_firmware_info *ucode = NULL;
 
  /* for baremetal, the ucode is allocated in gtt, so don't need to fill the bo when reset/suspend */
-	if (!amdgpu_sriov_vf(adev) && (adev->in_gpu_reset || adev->in_suspend))
+	if (!amdgpu_sriov_vf(adev) && (amdgpu_in_reset(adev) || adev->in_suspend))
 		return 0;
 	/*
 	 * if SMU loaded firmware, it needn't add SMC, UVD, and VCE
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
index f826945989c7..b2046c3a404d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
@@ -325,9 +325,9 @@ static inline bool is_virtual_machine(void)
 #define amdgpu_sriov_is_pp_one_vf(adev) \
 	((adev)->virt.gim_feature & AMDGIM_FEATURE_PP_ONE_VF)
 #define amdgpu_sriov_is_debug(adev) \
-	((!adev->in_gpu_reset) && adev->virt.tdr_debug)
+	((!amdgpu_in_reset(adev)) && adev->virt.tdr_debug)
 #define amdgpu_sriov_is_normal(adev) \
-	((!adev->in_gpu_reset) && (!adev->virt.tdr_debug))
+	((!amdgpu_in_reset(adev)) && (!adev->virt.tdr_debug))
 
 bool amdgpu_virt_mmio_blocked(struct amdgpu_device *adev);
 void amdgpu_virt_init_setting(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
index cd6e6eb7d966..7a61dc6738eb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
@@ -372,7 +372,7 @@ struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev, int lo
 	tmp->hive_id = adev->gmc.xgmi.hive_id;
 	INIT_LIST_HEAD(&tmp->device_list);
 	mutex_init(&tmp->hive_lock);
-	mutex_init(&tmp->reset_lock);
+	atomic_set(&tmp->in_reset, 0);
 	task_barrier_init(&tmp->tb);
 
 	if (lock)
@@ -594,7 +594,6 @@ int amdgpu_xgmi_remove_device(struct amdgpu_device *adev)
 	if(!(--hive->number_devices)){
 		amdgpu_xgmi_sysfs_destroy(adev, hive);
 		mutex_destroy(&hive->hive_lock);
-		mutex_destroy(&hive->reset_lock);
 	}
 
 	return psp_xgmi_terminate(&adev->psp);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h
index 6999eab16a72..453336ca9675 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h
@@ -30,10 +30,11 @@ struct amdgpu_hive_info {
 	uint64_t		hive_id;
 	struct list_head	device_list;
 	int number_devices;
-	struct mutex hive_lock, reset_lock;
+	struct mutex hive_lock;
 	struct kobject *kobj;
 	struct device_attribute dev_attr;
 	struct amdgpu_device *adev;
+	atomic_t in_reset;
 	int hi_req_count;
 	struct amdgpu_device *hi_req_gpu;
 	struct task_barrier tb;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index e527be22a3d5..d851fe80eaf4 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -6198,7 +6198,7 @@ static int gfx_v10_0_gfx_init_queue(struct amdgpu_ring *ring)
 	struct v10_gfx_mqd *mqd = ring->mqd_ptr;
 	int mqd_idx = ring - &adev->gfx.gfx_ring[0];
 
-	if (!adev->in_gpu_reset && !adev->in_suspend) {
+	if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
 		memset((void *)mqd, 0, sizeof(*mqd));
 		mutex_lock(&adev->srbm_mutex);
 		nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
@@ -6210,7 +6210,7 @@ static int gfx_v10_0_gfx_init_queue(struct amdgpu_ring *ring)
 		mutex_unlock(&adev->srbm_mutex);
 		if (adev->gfx.me.mqd_backup[mqd_idx])
 			memcpy(adev->gfx.me.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
-	} else if (adev->in_gpu_reset) {
+	} else if (amdgpu_in_reset(adev)) {
 		/* reset mqd with the backup copy */
 		if (adev->gfx.me.mqd_backup[mqd_idx])
 			memcpy(mqd, adev->gfx.me.mqd_backup[mqd_idx], sizeof(*mqd));
@@ -6563,7 +6563,7 @@ static int gfx_v10_0_kiq_init_queue(struct amdgpu_ring *ring)
 
 	gfx_v10_0_kiq_setting(ring);
 
-	if (adev->in_gpu_reset) { /* for GPU_RESET case */
+	if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
 		/* reset MQD to a clean status */
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd));
@@ -6599,7 +6599,7 @@ static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring)
 	struct v10_compute_mqd *mqd = ring->mqd_ptr;
 	int mqd_idx = ring - &adev->gfx.compute_ring[0];
 
-	if (!adev->in_gpu_reset && !adev->in_suspend) {
+	if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
 		memset((void *)mqd, 0, sizeof(*mqd));
 		mutex_lock(&adev->srbm_mutex);
 		nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
@@ -6609,7 +6609,7 @@ static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring)
 
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd));
-	} else if (adev->in_gpu_reset) { /* for GPU_RESET case */
+	} else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
 		/* reset MQD to a clean status */
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd));
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 3b7127202420..d5843e8a6e17 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -4633,7 +4633,7 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)
 
 	gfx_v8_0_kiq_setting(ring);
 
-	if (adev->in_gpu_reset) { /* for GPU_RESET case */
+	if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
 		/* reset MQD to a clean status */
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation));
@@ -4670,7 +4670,7 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)
 	struct vi_mqd *mqd = ring->mqd_ptr;
 	int mqd_idx = ring - &adev->gfx.compute_ring[0];
 
-	if (!adev->in_gpu_reset && !adev->in_suspend) {
+	if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
 		memset((void *)mqd, 0, sizeof(struct vi_mqd_allocation));
 		((struct vi_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
 		((struct vi_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
@@ -4682,7 +4682,7 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)
 
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct vi_mqd_allocation));
-	} else if (adev->in_gpu_reset) { /* for GPU_RESET case */
+	} else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
 		/* reset MQD to a clean status */
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation));
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 2c5bb282cc01..93c63ff3b35e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -3686,7 +3686,7 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring)
 
 	gfx_v9_0_kiq_setting(ring);
 
-	if (adev->in_gpu_reset) { /* for GPU_RESET case */
+	if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
 		/* reset MQD to a clean status */
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation));
@@ -3724,7 +3724,7 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring)
 	struct v9_mqd *mqd = ring->mqd_ptr;
 	int mqd_idx = ring - &adev->gfx.compute_ring[0];
 
-	if (!adev->in_gpu_reset && !adev->in_suspend) {
+	if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
 		memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation));
 		((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
 		((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
@@ -3736,7 +3736,7 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring)
 
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation));
-	} else if (adev->in_gpu_reset) { /* for GPU_RESET case */
+	} else if (amdgpu_in_reset(adev)) { /* for GPU_RESET case */
 		/* reset MQD to a clean status */
 		if (adev->gfx.mec.mqd_backup[mqd_idx])
 			memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation));
@@ -3930,7 +3930,7 @@ static int gfx_v9_0_hw_fini(void *handle)
 	/* Use deinitialize sequence from CAIL when unbinding device from driver,
 	 * otherwise KIQ is hanging when binding back
 	 */
-	if (!adev->in_gpu_reset && !adev->in_suspend) {
+	if (!amdgpu_in_reset(adev) && !adev->in_suspend) {
 		mutex_lock(&adev->srbm_mutex);
 		soc15_grbm_select(adev, adev->gfx.kiq.ring.me,
 				adev->gfx.kiq.ring.pipe,
@@ -4088,7 +4088,7 @@ static uint64_t gfx_v9_0_kiq_read_clock(struct amdgpu_device *adev)
 	 *
 	 * also don't wait anymore for IRQ context
 	 * */
-	if (r < 1 && (adev->in_gpu_reset || in_interrupt()))
+	if (r < 1 && (amdgpu_in_reset(adev) || in_interrupt()))
 		goto failed_kiq_read;
 
 	might_sleep();
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index ec8c0af39553..9d3b1245a339 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -287,7 +287,7 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 	 */
 	if (adev->gfx.kiq.ring.sched.ready &&
 	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev)) &&
-	    !adev->in_gpu_reset) {
+	    !amdgpu_in_reset(adev)) {
 
 		struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
 		const unsigned eng = 17;
@@ -312,7 +312,7 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 
 	if (!adev->mman.buffer_funcs_enabled ||
 	    !adev->ib_pool_ready ||
-	    adev->in_gpu_reset ||
+	    amdgpu_in_reset(adev) ||
 	    ring->sched.ready == false) {
 		gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_GFXHUB_0, 0);
 		mutex_unlock(&adev->mman.gtt_window_lock);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
index 3ce5c1d2fdf2..80c146df338a 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
@@ -434,7 +434,7 @@ static int gmc_v7_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
 	int vmid;
 	unsigned int tmp;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EIO;
 
 	for (vmid = 1; vmid < 16; vmid++) {
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
index 3e6615f9d39c..9ab65ca7df77 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
@@ -635,7 +635,7 @@ static int gmc_v8_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
 	int vmid;
 	unsigned int tmp;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EIO;
 
 	for (vmid = 1; vmid < 16; vmid++) {
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 6a780b674018..86f20426c4af 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -501,7 +501,7 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 	 */
 	if (adev->gfx.kiq.ring.sched.ready &&
 			(amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev)) &&
-			!adev->in_gpu_reset) {
+			!amdgpu_in_reset(adev)) {
 		uint32_t req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
 		uint32_t ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
 
@@ -596,7 +596,7 @@ static int gmc_v9_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
 	struct amdgpu_ring *ring = &adev->gfx.kiq.ring;
 	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EIO;
 
 	if (ring->sched.ready) {
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
index 475ff5df8c87..f27d83f2de78 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
@@ -250,7 +250,7 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work)
 	 */
 	locked = mutex_trylock(&adev->lock_reset);
 	if (locked)
-		adev->in_gpu_reset = true;
+		atomic_set(&adev->in_gpu_reset, 1);
 
 	do {
 		if (xgpu_ai_mailbox_peek_msg(adev) == IDH_FLR_NOTIFICATION_CMPL)
@@ -262,7 +262,7 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work)
 
 flr_done:
 	if (locked) {
-		adev->in_gpu_reset = false;
+		atomic_set(&adev->in_gpu_reset, 0);
 		mutex_unlock(&adev->lock_reset);
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
index 9cf695c05db3..3cb10ab943a6 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
@@ -271,7 +271,7 @@ static void xgpu_nv_mailbox_flr_work(struct work_struct *work)
 	 */
 	locked = mutex_trylock(&adev->lock_reset);
 	if (locked)
-		adev->in_gpu_reset = true;
+		atomic_set(&adev->in_gpu_reset, 1);
 
 	do {
 		if (xgpu_nv_mailbox_peek_msg(adev) == IDH_FLR_NOTIFICATION_CMPL)
@@ -283,7 +283,7 @@ static void xgpu_nv_mailbox_flr_work(struct work_struct *work)
 
 flr_done:
 	if (locked) {
-		adev->in_gpu_reset = false;
+		atomic_set(&adev->in_gpu_reset, 0);
 		mutex_unlock(&adev->lock_reset);
 	}
 
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 0a1e53fda08a..b8f36cbcff7b 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -1658,7 +1658,7 @@ static int dm_suspend(void *handle)
 	struct amdgpu_display_manager *dm = &adev->dm;
 	int ret = 0;
 
-	if (adev->in_gpu_reset) {
+	if (amdgpu_in_reset(adev)) {
 		mutex_lock(&dm->dc_lock);
 		dm->cached_dc_state = dc_copy_state(dm->dc->current_state);
 
@@ -1844,7 +1844,7 @@ static int dm_resume(void *handle)
 	struct dc_state *dc_state;
 	int i, r, j;
 
-	if (adev->in_gpu_reset) {
+	if (amdgpu_in_reset(adev)) {
 		dc_state = dm->cached_dc_state;
 
 		r = dm_dmub_hw_init(adev);
diff --git a/drivers/gpu/drm/amd/pm/amdgpu_pm.c b/drivers/gpu/drm/amd/pm/amdgpu_pm.c
index a77f7347fdfc..e469de005451 100644
--- a/drivers/gpu/drm/amd/pm/amdgpu_pm.c
+++ b/drivers/gpu/drm/amd/pm/amdgpu_pm.c
@@ -121,7 +121,7 @@ static ssize_t amdgpu_get_power_dpm_state(struct device *dev,
 	enum amd_pm_state_type pm;
 	int ret;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EPERM;
 
 	ret = pm_runtime_get_sync(ddev->dev);
@@ -159,7 +159,7 @@ static ssize_t amdgpu_set_power_dpm_state(struct device *dev,
 	enum amd_pm_state_type  state;
 	int ret;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EPERM;
 
 	if (strncmp("battery", buf, strlen("battery")) == 0)
@@ -265,7 +265,7 @@ static ssize_t amdgpu_get_power_dpm_force_performance_level(struct device *dev,
 	enum amd_dpm_forced_level level = 0xff;
 	int ret;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EPERM;
 
 	ret = pm_runtime_get_sync(ddev->dev);
@@ -307,7 +307,7 @@ static ssize_t amdgpu_set_power_dpm_force_performance_level(struct device *dev,
 	enum amd_dpm_forced_level current_level = 0xff;
 	int ret = 0;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EPERM;
 
 	if (strncmp("low", buf, strlen("low")) == 0) {
@@ -411,7 +411,7 @@ static ssize_t amdgpu_get_pp_num_states(struct device *dev,
 	struct pp_states_info data;
 	int i, buf_len, ret;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EPERM;
 
 	ret = pm_runtime_get_sync(ddev->dev);
@@ -455,7 +455,7 @@ static ssize_t amdgpu_get_pp_cur_state(struct device *dev,
 	enum amd_pm_state_type pm = 0;
 	int i = 0, ret = 0;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EPERM;
 
 	ret = pm_runtime_get_sync(ddev->dev);
@@ -496,7 +496,7 @@ static ssize_t amdgpu_get_pp_force_state(struct device *dev,
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = ddev->dev_private;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EPERM;
 
 	if (adev->pp_force_state_enabled)
@@ -516,7 +516,7 @@ static ssize_t amdgpu_set_pp_force_state(struct device *dev,
 	unsigned long idx;
 	int ret;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EPERM;
 
 	if (strlen(buf) == 1)
@@ -576,7 +576,7 @@ static ssize_t amdgpu_get_pp_table(struct device *dev,
 	char *table = NULL;
 	int size, ret;
 
-	if (adev->in_gpu_reset)
+	if (amdgpu_in_reset(adev))
 		return -EPERM;
 
 	ret = pm_runtime_get_sync(ddev->dev);
@@ -620,7 +620,7 @@ static ssize_t amdgpu_set_pp_table(struct device *dev,
 	struct amdgpu_device *adev = ddev->dev_private;
 	int ret = 0;