Re: [PATCH 1/5] scsi: rescan VPD attributes

2015-07-25 Thread Hannes Reinecke
On 07/24/2015 04:43 PM, Christoph Hellwig wrote: On Fri, Jul 24, 2015 at 04:40:42PM +0200, Hannes Reinecke wrote: Yeah, possibly. After all, the variable isn't expected to change under rcu_read_lock(). Actually it can and will change, that's the point. But if you use a local variable you

Re: [PATCH 1/5] scsi: rescan VPD attributes

2015-07-24 Thread Christoph Hellwig
On Fri, Jul 24, 2015 at 04:40:42PM +0200, Hannes Reinecke wrote: Yeah, possibly. After all, the variable isn't expected to change under rcu_read_lock(). Actually it can and will change, that's the point. But if you use a local variable you keep a single version of it, which won't be freed

Re: [PATCH 1/5] scsi: rescan VPD attributes

2015-07-24 Thread Christoph Hellwig
On Wed, Jul 08, 2015 at 01:09:44PM +0200, Hannes Reinecke wrote: +++ b/drivers/scsi/device_handler/scsi_dh_alua.c @@ -264,8 +264,11 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h) int device_id_size, device_id_type = 0; struct alua_port_group

[PATCH 1/5] scsi: rescan VPD attributes

2015-07-08 Thread Hannes Reinecke
This patch implements a VPD page rescan if the 'rescan' sysfs attribute is triggered. Signed-off-by: Hannes Reinecke h...@suse.de --- drivers/scsi/device_handler/scsi_dh_alua.c | 11 +++ drivers/scsi/scsi.c| 20 +--- drivers/scsi/scsi_scan.c