habanalabs: add new return code to device fd open

In order to be more informative during device open, we are adding a
new return code -EAGAIN that indicates device is still going through
resource reclaiming and hence it cannot be used yet.
Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers/misc/habanalabs/common/device.c

@@ -107,6 +107,8 @@ static void hpriv_release(struct kref *ref)
 	hdev->is_compute_ctx_active = false;
 	mutex_unlock(&hdev->fpriv_list_lock);
 
+	hdev->compute_ctx_in_release = 0;
+
 	kfree(hpriv);
 }
 
@@ -150,6 +152,8 @@ static int hl_device_release(struct inode *inode, struct file *filp)
 	hl_ts_mgr_fini(hpriv->hdev, &hpriv->ts_mem_mgr);
 	hl_ctx_mgr_fini(hdev, &hpriv->ctx_mgr);
 
+	hdev->compute_ctx_in_release = 1;
+
 	if (!hl_hpriv_put(hpriv))
 		dev_notice(hdev->dev,
 			"User process closed FD but device still in use\n");

drivers/misc/habanalabs/common/habanalabs.h

@@ -2710,6 +2710,7 @@ struct hl_reset_info {
  *                        cases where Linux was not loaded to device CPU
  * @supports_wait_for_multi_cs: true if wait for multi CS is supported
  * @is_compute_ctx_active: Whether there is an active compute context executing.
+ * @compute_ctx_in_release: true if the current compute context is being released.
  */
 struct hl_device {
 	struct pci_dev			*pdev;
@@ -2828,6 +2829,7 @@ struct hl_device {
 	u8				supports_wait_for_multi_cs;
 	u8				stream_master_qid_arr_size;
 	u8				is_compute_ctx_active;
+	u8				compute_ctx_in_release;
 
 	/* Parameters for bring-up */
 	u64				nic_ports_mask;

drivers/misc/habanalabs/common/habanalabs_drv.c

@@ -150,7 +150,20 @@ int hl_device_open(struct inode *inode, struct file *filp)
 		dev_err_ratelimited(hdev->dev,
 			"Can't open %s because it is %s\n",
 			dev_name(hdev->dev), hdev->status[status]);
-		rc = -EPERM;
+
+		if (status == HL_DEVICE_STATUS_IN_RESET)
+			rc = -EAGAIN;
+		else
+			rc = -EPERM;
+
+		goto out_err;
+	}
+
+	if (hdev->compute_ctx_in_release) {
+		dev_dbg_ratelimited(hdev->dev,
+			"Can't open %s because another user is still releasing it\n",
+			dev_name(hdev->dev));
+		rc = -EAGAIN;
 		goto out_err;
 	}