On Wed, 2006-10-18 at 20:19 -0700, malahal@xxxxxxxxxx wrote:
> Domain device is freed but the port dev list is not adjusted on some
> discovery errors. Module unload will Oops if this happens.
>
> Signed-off-by: Malahal Naineni <malahal@xxxxxxxxxx>
>
> diff -r 4e1720b3c71b drivers/scsi/libsas/sas_discover.c
> --- a/drivers/scsi/libsas/sas_discover.c Thu Oct 05 10:02:13 2006 -0700
> +++ b/drivers/scsi/libsas/sas_discover.c Fri Oct 13 19:58:54 2006 -0700
> @@ -684,7 +684,23 @@ static void sas_discover_domain(void *da
> }
>
> if (error) {
> - kfree(port->port_dev); /* not kobject_register-ed yet */
> + struct domain_device *dev = port->port_dev;
> +
> + if (dev->rphy) {
> + sas_remove_children(&dev->rphy->dev);
> + sas_rphy_delete(dev->rphy);
> + dev->rphy = NULL;
> + }
> + if (dev->dev_type == EDGE_DEV || dev->dev_type == FANOUT_DEV) {
> + if (dev->ex_dev.ex_phy) {
> + kfree(dev->ex_dev.ex_phy);
> + dev->ex_dev.ex_phy = NULL;
> + }
> + }
> + spin_lock(&port->dev_list_lock);
> + list_del_init(&dev->dev_list_node);
> + spin_unlock(&port->dev_list_lock);
> + kfree(dev); /* not kobject_register-ed yet */
This is a bit of a layering violation
if the various discover functions are going to return an error, I think
it's their job to clean up whatever they did before returning.
James
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
