accel/habanalabs: use a mutex rather than a spinlock

There are two reasons why mutex is better here:
1. There's a critical section relatively long, where in
certain scenarios (e.g., multiple VM allocations) taking a spinlock
might cause noticeable performance degradation.
2. It will remove the incorrect usage of mutex under
spin_lock (where preemption is disabled).
Reported-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Koby Elbaz <kelbaz@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>

drivers/accel/habanalabs/common/debugfs.c

@@ -258,7 +258,7 @@ static int vm_show(struct seq_file *s, void *data)
 	if (!dev_entry->hdev->mmu_enable)
 		return 0;
 
-	spin_lock(&dev_entry->ctx_mem_hash_spinlock);
+	mutex_lock(&dev_entry->ctx_mem_hash_mutex);
 
 	list_for_each_entry(ctx, &dev_entry->ctx_mem_hash_list, debugfs_list) {
 		once = false;
@@ -329,7 +329,7 @@ static int vm_show(struct seq_file *s, void *data)
 
 	}
 
-	spin_unlock(&dev_entry->ctx_mem_hash_spinlock);
+	mutex_unlock(&dev_entry->ctx_mem_hash_mutex);
 
 	ctx = hl_get_compute_ctx(dev_entry->hdev);
 	if (ctx) {
@@ -1785,7 +1785,7 @@ void hl_debugfs_add_device(struct hl_device *hdev)
 	spin_lock_init(&dev_entry->cs_spinlock);
 	spin_lock_init(&dev_entry->cs_job_spinlock);
 	spin_lock_init(&dev_entry->userptr_spinlock);
-	spin_lock_init(&dev_entry->ctx_mem_hash_spinlock);
+	mutex_init(&dev_entry->ctx_mem_hash_mutex);
 
 	dev_entry->root = debugfs_create_dir(dev_name(hdev->dev),
 						hl_debug_root);
@@ -1802,6 +1802,7 @@ void hl_debugfs_remove_device(struct hl_device *hdev)
 
 	debugfs_remove_recursive(entry->root);
 
+	mutex_destroy(&entry->ctx_mem_hash_mutex);
 	mutex_destroy(&entry->file_mutex);
 
 	vfree(entry->data_dma_blob_desc.data);
@@ -1908,18 +1909,18 @@ void hl_debugfs_add_ctx_mem_hash(struct hl_device *hdev, struct hl_ctx *ctx)
 {
 	struct hl_dbg_device_entry *dev_entry = &hdev->hl_debugfs;
 
-	spin_lock(&dev_entry->ctx_mem_hash_spinlock);
+	mutex_lock(&dev_entry->ctx_mem_hash_mutex);
 	list_add(&ctx->debugfs_list, &dev_entry->ctx_mem_hash_list);
-	spin_unlock(&dev_entry->ctx_mem_hash_spinlock);
+	mutex_unlock(&dev_entry->ctx_mem_hash_mutex);
 }
 
 void hl_debugfs_remove_ctx_mem_hash(struct hl_device *hdev, struct hl_ctx *ctx)
 {
 	struct hl_dbg_device_entry *dev_entry = &hdev->hl_debugfs;
 
-	spin_lock(&dev_entry->ctx_mem_hash_spinlock);
+	mutex_lock(&dev_entry->ctx_mem_hash_mutex);
 	list_del(&ctx->debugfs_list);
-	spin_unlock(&dev_entry->ctx_mem_hash_spinlock);
+	mutex_unlock(&dev_entry->ctx_mem_hash_mutex);
 }
 
 /**

drivers/accel/habanalabs/common/habanalabs.h

@@ -2320,7 +2320,7 @@ struct hl_debugfs_entry {
  * @userptr_list: list of available userptrs (virtual memory chunk descriptor).
  * @userptr_spinlock: protects userptr_list.
  * @ctx_mem_hash_list: list of available contexts with MMU mappings.
- * @ctx_mem_hash_spinlock: protects cb_list.
+ * @ctx_mem_hash_mutex: protects list of available contexts with MMU mappings.
  * @data_dma_blob_desc: data DMA descriptor of blob.
  * @mon_dump_blob_desc: monitor dump descriptor of blob.
  * @state_dump: data of the system states in case of a bad cs.
@@ -2351,7 +2351,7 @@ struct hl_dbg_device_entry {
 	struct list_head		userptr_list;
 	spinlock_t			userptr_spinlock;
 	struct list_head		ctx_mem_hash_list;
-	spinlock_t			ctx_mem_hash_spinlock;
+	struct mutex			ctx_mem_hash_mutex;
 	struct debugfs_blob_wrapper	data_dma_blob_desc;
 	struct debugfs_blob_wrapper	mon_dump_blob_desc;
 	char				*state_dump[HL_STATE_DUMP_HIST_LEN];