<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns="http://www.w3.org/TR/REC-html40" xmlns:ns0="http://schemas.microsoft.com/office/2004/12/omml"> <head> <meta http-equiv=Content-Type content="text/html; charset=us-ascii"> <meta name=Generator content="Microsoft Word 11 (filtered medium)">  <style>  </style>  </head> <body lang=EN-US link=blue vlink=purple> <div class=Section1> Menny,<o:p></o:p> Yes, I agree there is a problem. Wouldn’t it be simpler if you could handle everything scsi_dh.c.. See my response below.. <o:p></o:p> Thanks<o:p></o:p> Babu<o:p></o:p> <o:p></o:p> <div style='border:none;border-left:solid blue 1.5pt;padding:0in 0in 0in 4.0pt'> <div> <div class=MsoNormal align=center style='text-align:center'> <hr size=2 width="100%" align=center tabindex=-1> </div> From: dm-devel-bounces@redhat.com [mailto:dm-devel-bounces@redhat.com] On Behalf Of Menny_Hamburger@Dell.com Sent: Monday, December 13, 2010 9:35 AM To: dm-devel@redhat.com Subject: [dm-devel] [PATCH 1/1]: missing call to pg_init_done causes I/O to be hung forever<o:p></o:p> </div> <o:p> </o:p> When scsi_dh_activate returns SCSI_DH_NOSYS the H/W handler callback is not called, pg_init_done is not called in <o:p></o:p> the multipath layer and pending I/O is requeued forever; this situation causes all userland processes currently performing I/O<o:p></o:p> on the device to I/O hang. A similar situation occurs when the device has transitioned to SDEV_CANCEL/SDEV_DEL and the device<o:p></o:p> handler data had not yet been deleted.<o:p></o:p> <o:p> </o:p> The easiest way to reproduce this is in an ISCSI environment:<o:p></o:p> dd if=/dev/dm-0 of=/dev/zero bs=8k count=1000000 &<o:p></o:p> /etc/init.d/iscsi stop<o:p></o:p> In this example, dd will I/O hang forever and the only way to release it will be to reboot the machine<o:p></o:p> <o:p> </o:p> This patch calls pg_init_done directly from the mpath code when the scsi_dh_activate returns a non SCSI_DH_OK error.<o:p></o:p> <o:p> </o:p> Note:<o:p></o:p> The patch is over RHEL5.5.<o:p></o:p> When running an upstream kernel, the above scenario may not occur because the request queue is aborted in dm-mpath.c:fail_path.<o:p></o:p> This patch makes sure the problem does not occur at all, rather than handling it when it does. In addition, it seems too risky to apply <o:p></o:p> request queue abort functionality on RHEL5 at this stage.<o:p></o:p> <o:p> </o:p> diff -r -U 2 a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c<o:p></o:p> --- a/drivers/md/dm-mpath.c 2010-12-13 09:16:31.358858000 +0200<o:p></o:p> +++ b/drivers/md/dm-mpath.c 2010-12-13 09:16:31.796998000 +0200<o:p></o:p> @@ -1190,4 +1190,5 @@<o:p></o:p> case SCSI_DH_OK:<o:p></o:p> break;<o:p></o:p> + case SCSI_DH_DEV_OFFLINED:<o:p></o:p> <o:p> </o:p> If you are not doing anything special then I would let default take care of it. No need of this change..<o:p></o:p> <o:p> </o:p> case SCSI_DH_NOSYS:<o:p></o:p> if (!m->hw_handler_name) {<o:p></o:p> @@ -1252,7 +1253,15 @@<o:p></o:p> {<o:p></o:p> struct pgpath *pgpath = (struct pgpath *) data;<o:p></o:p> + int err;<o:p></o:p> <o:p></o:p> - scsi_dh_activate(bdev_get_queue(pgpath->path.dev->bdev),<o:p></o:p> + err = scsi_dh_activate(bdev_get_queue(pgpath->path.dev->bdev),<o:p></o:p> pg_init_done, &pgpath->path);<o:p></o:p> +<o:p></o:p> + /*<o:p></o:p> + * If error is not SCSI_DH_OK, we have not entered the scsi_dh H/W handler and did not call pg_init_done - <o:p></o:p> + * need to call pg_init_done directly.<o:p></o:p> + */<o:p></o:p> + if (err) <o:p></o:p> + pg_init_done(&pgpath->path, err);<o:p></o:p> }<o:p></o:p> You can move this to scsi_dh.c<o:p></o:p> <o:p> </o:p> <o:p></o:p> diff -r -U 2 a/drivers/scsi/device_handler/scsi_dh.c b/drivers/scsi/device_handler/scsi_dh.c<o:p></o:p> --- a/drivers/scsi/device_handler/scsi_dh.c 2010-12-13 09:16:31.616554000 +0200<o:p></o:p> +++ b/drivers/scsi/device_handler/scsi_dh.c 2010-12-13 09:16:31.878170000 +0200<o:p></o:p> @@ -443,4 +443,9 @@<o:p></o:p> spin_unlock_irqrestore(q->queue_lock, flags);<o:p></o:p> <o:p></o:p> + if (sdev->sdev_state == SDEV_CANCEL ||<o:p></o:p> + sdev->sdev_state == SDEV_DEL ||<o:p></o:p> + sdev->sdev_state == SDEV_OFFLINE)<o:p></o:p> + err = SCSI_DH_DEV_OFFLINED;<o:p></o:p> +<o:p></o:p> You can change it something like below..<o:p></o:p> if (err) {<o:p></o:p> if(fn)<o:p></o:p> fn(data, err);<o:p></o:p> return err;<o:p></o:p> }<o:p></o:p> <o:p> </o:p> <o:p> </o:p> </div> </div> </body> </html>