8 files changed, 59 insertions, 18 deletions

diff --git a/drivers/accel/ivpu/ivpu_debugfs.c b/drivers/accel/ivpu/ivpu_debugfs.c
index f0dad0c9ce33..cd24ccd20ba6 100644
--- a/drivers/accel/ivpu/ivpu_debugfs.c
+++ b/drivers/accel/ivpu/ivpu_debugfs.c
@@ -455,7 +455,7 @@ priority_bands_fops_write(struct file *file, const char __user *user_buf, size_t
 	if (ret < 0)
 		return ret;
 
-	buf[size] = '\0';
+	buf[ret] = '\0';
 	ret = sscanf(buf, "%u %u %u %u", &band, &grace_period, &process_grace_period,
 		     &process_quantum);
 	if (ret != 4)
diff --git a/drivers/accel/ivpu/ivpu_drv.c b/drivers/accel/ivpu/ivpu_drv.c
index eff1d3ca075f..0e7748c5e117 100644
--- a/drivers/accel/ivpu/ivpu_drv.c
+++ b/drivers/accel/ivpu/ivpu_drv.c
@@ -374,6 +374,9 @@ int ivpu_boot(struct ivpu_device *vdev)
 {
 	int ret;
 
+	drm_WARN_ON(&vdev->drm, atomic_read(&vdev->job_timeout_counter));
+	drm_WARN_ON(&vdev->drm, !xa_empty(&vdev->submitted_jobs_xa));
+
 	/* Update boot params located at first 4KB of FW memory */
 	ivpu_fw_boot_params_setup(vdev, ivpu_bo_vaddr(vdev->fw->mem));
 
@@ -573,6 +576,7 @@ static int ivpu_dev_init(struct ivpu_device *vdev)
 	vdev->context_xa_limit.min = IVPU_USER_CONTEXT_MIN_SSID;
 	vdev->context_xa_limit.max = IVPU_USER_CONTEXT_MAX_SSID;
 	atomic64_set(&vdev->unique_id_counter, 0);
+	atomic_set(&vdev->job_timeout_counter, 0);
 	xa_init_flags(&vdev->context_xa, XA_FLAGS_ALLOC | XA_FLAGS_LOCK_IRQ);
 	xa_init_flags(&vdev->submitted_jobs_xa, XA_FLAGS_ALLOC1);
 	xa_init_flags(&vdev->db_xa, XA_FLAGS_ALLOC1);
diff --git a/drivers/accel/ivpu/ivpu_drv.h b/drivers/accel/ivpu/ivpu_drv.h
index 92753effb1c9..5497e7030e91 100644
--- a/drivers/accel/ivpu/ivpu_drv.h
+++ b/drivers/accel/ivpu/ivpu_drv.h
@@ -154,6 +154,7 @@ struct ivpu_device {
 	struct mutex submitted_jobs_lock; /* Protects submitted_jobs */
 	struct xarray submitted_jobs_xa;
 	struct ivpu_ipc_consumer job_done_consumer;
+	atomic_t job_timeout_counter;
 
 	atomic64_t unique_id_counter;
 
diff --git a/drivers/accel/ivpu/ivpu_fw.h b/drivers/accel/ivpu/ivpu_fw.h
index 1d0b2bd9d65c..9a3935be1c05 100644
--- a/drivers/accel/ivpu/ivpu_fw.h
+++ b/drivers/accel/ivpu/ivpu_fw.h
@@ -39,6 +39,7 @@ struct ivpu_fw_info {
 	u64 read_only_addr;
 	u32 read_only_size;
 	u32 sched_mode;
+	u64 last_heartbeat;
 };
 
 int ivpu_fw_init(struct ivpu_device *vdev);
diff --git a/drivers/accel/ivpu/ivpu_gem.c b/drivers/accel/ivpu/ivpu_gem.c
index 8741c73b92ce..e0d242d9f3e5 100644
--- a/drivers/accel/ivpu/ivpu_gem.c
+++ b/drivers/accel/ivpu/ivpu_gem.c
@@ -30,7 +30,7 @@ static inline void ivpu_dbg_bo(struct ivpu_device *vdev, struct ivpu_bo *bo, con
 		 "%6s: bo %8p vpu_addr %9llx size %8zu ctx %d has_pages %d dma_mapped %d mmu_mapped %d wc %d imported %d\n",
 		 action, bo, bo->vpu_addr, ivpu_bo_size(bo), bo->ctx ? bo->ctx->id : 0,
 		 (bool)bo->base.pages, (bool)bo->base.sgt, bo->mmu_mapped, bo->base.map_wc,
-		 (bool)bo->base.base.import_attach);
+		 (bool)drm_gem_is_imported(&bo->base.base));
 }
 
 /*
@@ -122,7 +122,7 @@ static void ivpu_bo_unbind_locked(struct ivpu_bo *bo)
 		bo->ctx = NULL;
 	}
 
-	if (bo->base.base.import_attach)
+	if (drm_gem_is_imported(&bo->base.base))
 		return;
 
 	dma_resv_lock(bo->base.base.resv, NULL);
@@ -282,7 +282,7 @@ static void ivpu_gem_bo_free(struct drm_gem_object *obj)
 	ivpu_bo_unbind_locked(bo);
 	mutex_destroy(&bo->lock);
 
-	drm_WARN_ON(obj->dev, bo->base.pages_use_count > 1);
+	drm_WARN_ON(obj->dev, refcount_read(&bo->base.pages_use_count) > 1);
 	drm_gem_shmem_free(&bo->base);
 }
 
@@ -362,7 +362,7 @@ ivpu_bo_create(struct ivpu_device *vdev, struct ivpu_mmu_context *ctx,
 
 	if (flags & DRM_IVPU_BO_MAPPABLE) {
 		dma_resv_lock(bo->base.base.resv, NULL);
-		ret = drm_gem_shmem_vmap(&bo->base, &map);
+		ret = drm_gem_shmem_vmap_locked(&bo->base, &map);
 		dma_resv_unlock(bo->base.base.resv);
 
 		if (ret)
@@ -387,7 +387,7 @@ void ivpu_bo_free(struct ivpu_bo *bo)
 
 	if (bo->flags & DRM_IVPU_BO_MAPPABLE) {
 		dma_resv_lock(bo->base.base.resv, NULL);
-		drm_gem_shmem_vunmap(&bo->base, &map);
+		drm_gem_shmem_vunmap_locked(&bo->base, &map);
 		dma_resv_unlock(bo->base.base.resv);
 	}
 
@@ -461,7 +461,7 @@ static void ivpu_bo_print_info(struct ivpu_bo *bo, struct drm_printer *p)
 	if (bo->mmu_mapped)
 		drm_printf(p, " mmu_mapped");
 
-	if (bo->base.base.import_attach)
+	if (drm_gem_is_imported(&bo->base.base))
 		drm_printf(p, " imported");
 
 	drm_printf(p, "\n");
diff --git a/drivers/accel/ivpu/ivpu_hw.c b/drivers/accel/ivpu/ivpu_hw.c
index ec9a3629da3a..633160470c93 100644
--- a/drivers/accel/ivpu/ivpu_hw.c
+++ b/drivers/accel/ivpu/ivpu_hw.c
@@ -119,7 +119,7 @@ static void timeouts_init(struct ivpu_device *vdev)
 		else
 			vdev->timeout.autosuspend = 100;
 		vdev->timeout.d0i3_entry_msg = 5;
-		vdev->timeout.state_dump_msg = 10;
+		vdev->timeout.state_dump_msg = 100;
 	}
 }
 
diff --git a/drivers/accel/ivpu/ivpu_job.c b/drivers/accel/ivpu/ivpu_job.c
index 863e3cd6ace5..b28da35c30b6 100644
--- a/drivers/accel/ivpu/ivpu_job.c
+++ b/drivers/accel/ivpu/ivpu_job.c
@@ -681,8 +681,8 @@ static int ivpu_job_submit(struct ivpu_job *job, u8 priority, u32 cmdq_id)
 err_erase_xa:
 	xa_erase(&vdev->submitted_jobs_xa, job->job_id);
 err_unlock:
-	mutex_unlock(&vdev->submitted_jobs_lock);
 	mutex_unlock(&file_priv->lock);
+	mutex_unlock(&vdev->submitted_jobs_lock);
 	ivpu_rpm_put(vdev);
 	return ret;
 }
@@ -874,15 +874,21 @@ int ivpu_cmdq_submit_ioctl(struct drm_device *dev, void *data, struct drm_file *
 int ivpu_cmdq_create_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 {
 	struct ivpu_file_priv *file_priv = file->driver_priv;
+	struct ivpu_device *vdev = file_priv->vdev;
 	struct drm_ivpu_cmdq_create *args = data;
 	struct ivpu_cmdq *cmdq;
+	int ret;
 
-	if (!ivpu_is_capable(file_priv->vdev, DRM_IVPU_CAP_MANAGE_CMDQ))
+	if (!ivpu_is_capable(vdev, DRM_IVPU_CAP_MANAGE_CMDQ))
 		return -ENODEV;
 
 	if (args->priority > DRM_IVPU_JOB_PRIORITY_REALTIME)
 		return -EINVAL;
 
+	ret = ivpu_rpm_get(vdev);
+	if (ret < 0)
+		return ret;
+
 	mutex_lock(&file_priv->lock);
 
 	cmdq = ivpu_cmdq_create(file_priv, ivpu_job_to_jsm_priority(args->priority), false);
@@ -891,6 +897,8 @@ int ivpu_cmdq_create_ioctl(struct drm_device *dev, void *data, struct drm_file *
 
 	mutex_unlock(&file_priv->lock);
 
+	ivpu_rpm_put(vdev);
+
 	return cmdq ? 0 : -ENOMEM;
 }
 
@@ -900,28 +908,35 @@ int ivpu_cmdq_destroy_ioctl(struct drm_device *dev, void *data, struct drm_file
 	struct ivpu_device *vdev = file_priv->vdev;
 	struct drm_ivpu_cmdq_destroy *args = data;
 	struct ivpu_cmdq *cmdq;
-	u32 cmdq_id;
+	u32 cmdq_id = 0;
 	int ret;
 
 	if (!ivpu_is_capable(vdev, DRM_IVPU_CAP_MANAGE_CMDQ))
 		return -ENODEV;
 
+	ret = ivpu_rpm_get(vdev);
+	if (ret < 0)
+		return ret;
+
 	mutex_lock(&file_priv->lock);
 
 	cmdq = xa_load(&file_priv->cmdq_xa, args->cmdq_id);
 	if (!cmdq || cmdq->is_legacy) {
 		ret = -ENOENT;
-		goto err_unlock;
+	} else {
+		cmdq_id = cmdq->id;
+		ivpu_cmdq_destroy(file_priv, cmdq);
+		ret = 0;
 	}
 
-	cmdq_id = cmdq->id;
-	ivpu_cmdq_destroy(file_priv, cmdq);
 	mutex_unlock(&file_priv->lock);
-	ivpu_cmdq_abort_all_jobs(vdev, file_priv->ctx.id, cmdq_id);
-	return 0;
 
-err_unlock:
-	mutex_unlock(&file_priv->lock);
+	/* Abort any pending jobs only if cmdq was destroyed */
+	if (!ret)
+		ivpu_cmdq_abort_all_jobs(vdev, file_priv->ctx.id, cmdq_id);
+
+	ivpu_rpm_put(vdev);
+
 	return ret;
 }
 
diff --git a/drivers/accel/ivpu/ivpu_pm.c b/drivers/accel/ivpu/ivpu_pm.c
index ac0e22454596..ea30db181cd7 100644
--- a/drivers/accel/ivpu/ivpu_pm.c
+++ b/drivers/accel/ivpu/ivpu_pm.c
@@ -34,6 +34,7 @@ module_param_named(tdr_timeout_ms, ivpu_tdr_timeout_ms, ulong, 0644);
 MODULE_PARM_DESC(tdr_timeout_ms, "Timeout for device hang detection, in milliseconds, 0 - default");
 
 #define PM_RESCHEDULE_LIMIT     5
+#define PM_TDR_HEARTBEAT_LIMIT  30
 
 static void ivpu_pm_prepare_cold_boot(struct ivpu_device *vdev)
 {
@@ -44,6 +45,7 @@ static void ivpu_pm_prepare_cold_boot(struct ivpu_device *vdev)
 	ivpu_fw_log_reset(vdev);
 	ivpu_fw_load(vdev);
 	fw->entry_point = fw->cold_boot_entry_point;
+	fw->last_heartbeat = 0;
 }
 
 static void ivpu_pm_prepare_warm_boot(struct ivpu_device *vdev)
@@ -189,7 +191,24 @@ static void ivpu_job_timeout_work(struct work_struct *work)
 {
 	struct ivpu_pm_info *pm = container_of(work, struct ivpu_pm_info, job_timeout_work.work);
 	struct ivpu_device *vdev = pm->vdev;
+	u64 heartbeat;
 
+	if (ivpu_jsm_get_heartbeat(vdev, 0, &heartbeat) || heartbeat <= vdev->fw->last_heartbeat) {
+		ivpu_err(vdev, "Job timeout detected, heartbeat not progressed\n");
+		goto recovery;
+	}
+
+	if (atomic_fetch_inc(&vdev->job_timeout_counter) > PM_TDR_HEARTBEAT_LIMIT) {
+		ivpu_err(vdev, "Job timeout detected, heartbeat limit exceeded\n");
+		goto recovery;
+	}
+
+	vdev->fw->last_heartbeat = heartbeat;
+	ivpu_start_job_timeout_detection(vdev);
+	return;
+
+recovery:
+	atomic_set(&vdev->job_timeout_counter, 0);
 	ivpu_pm_trigger_recovery(vdev, "TDR");
 }
 
@@ -204,6 +223,7 @@ void ivpu_start_job_timeout_detection(struct ivpu_device *vdev)
 void ivpu_stop_job_timeout_detection(struct ivpu_device *vdev)
 {
 	cancel_delayed_work_sync(&vdev->pm->job_timeout_work);
+	atomic_set(&vdev->job_timeout_counter, 0);
 }
 
 int ivpu_pm_suspend_cb(struct device *dev)