svn commit: r331917 - stable/11/sys/dev/mlx5/mlx5_core

Hans Petter Selasky hselasky at FreeBSD.org
Tue Apr 3 09:30:41 UTC 2018


Author: hselasky
Date: Tue Apr  3 09:30:40 2018
New Revision: 331917
URL: https://svnweb.freebsd.org/changeset/base/331917

Log:
  MFC r331824:
  Make sure Giant is locked when allocating bus resources in mlx5core.
  
  During health care IRQ resources will be reallocated.
  Newbus requires that Giant is locked before accessing
  these resources.
  
  Sponsored by:	Mellanox Technologies

Modified:
  stable/11/sys/dev/mlx5/mlx5_core/mlx5_health.c
Directory Properties:
  stable/11/   (props changed)

Modified: stable/11/sys/dev/mlx5/mlx5_core/mlx5_health.c
==============================================================================
--- stable/11/sys/dev/mlx5/mlx5_core/mlx5_health.c	Tue Apr  3 09:29:26 2018	(r331916)
+++ stable/11/sys/dev/mlx5/mlx5_core/mlx5_health.c	Tue Apr  3 09:30:40 2018	(r331917)
@@ -300,6 +300,8 @@ static void health_recover(struct work_struct *work)
 	priv = container_of(health, struct mlx5_priv, health);
 	dev = container_of(priv, struct mlx5_core_dev, priv);
 
+	mtx_lock(&Giant);	/* XXX newbus needs this */
+
 	if (sensor_pci_no_comm(dev)) {
 		dev_err(&dev->pdev->dev, "health recovery flow aborted, PCI reads still not working\n");
 		recover = false;
@@ -322,6 +324,8 @@ static void health_recover(struct work_struct *work)
 		dev_err(&dev->pdev->dev, "starting health recovery flow\n");
 		mlx5_recover_device(dev);
 	}
+
+	mtx_unlock(&Giant);
 }
 
 /* How much time to wait until health resetting the driver (in msecs) */


More information about the svn-src-all mailing list