All of lore.kernel.org
 help / color / mirror / Atom feed
From: Mike Snitzer <snitzer@redhat.com>
To: dm-devel@redhat.com
Cc: linux-scsi@vger.kernel.org,
	Mike Anderson <andmike@linux.vnet.ibm.com>,
	Mike Christie <michaelc@cs.wisc.edu>
Subject: [PATCH v3] dm mpath: add feature flag to control call to blk_abort_queue
Date: Thu, 18 Nov 2010 10:48:37 -0500	[thread overview]
Message-ID: <1290095317-15684-1-git-send-email-snitzer@redhat.com> (raw)
In-Reply-To: <20101118072032.GA13390@linux.vnet.ibm.com>

Add 'features' member to 'struct multipath' and introduce
MPF_ABORT_QUEUE as the first feature flag.  If "abort_queue_on_fail"
is provided during mpath device configuration the MPF_ABORT_QUEUE
feature flag will be set and blk_abort_queue() will be called during
path failure to allow for lower latency path failures.

But this feature has been disabled by default due to a race (between
blk_abort_queue and scsi_request_fn) that can lead to list corruption,
from: https://www.redhat.com/archives/dm-devel/2010-November/msg00085.html

"the cmd gets blk_abort_queued/timedout run on it and the scsi eh
somehow is able to complete and run scsi_queue_insert while
scsi_request_fn is still trying to process the request."

Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: Mike Anderson <andmike@linux.vnet.ibm.com>
Cc: Mike Christie <michaelc@cs.wisc.edu>
---
 drivers/md/dm-mpath.c |   21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)

Index: linux-2.6/drivers/md/dm-mpath.c
===================================================================
--- linux-2.6.orig/drivers/md/dm-mpath.c
+++ linux-2.6/drivers/md/dm-mpath.c
@@ -56,8 +56,14 @@ struct priority_group {
 	struct list_head pgpaths;
 };
 
+/*
+ * Bits for the m->features
+ */
+#define MPF_ABORT_QUEUE 0
+
 /* Multipath context */
 struct multipath {
+	unsigned long features;
 	struct list_head list;
 	struct dm_target *ti;
 
@@ -146,7 +152,8 @@ static void deactivate_path(struct work_
 	struct pgpath *pgpath =
 		container_of(work, struct pgpath, deactivate_path);
 
-	blk_abort_queue(pgpath->path.dev->bdev->bd_disk->queue);
+	if (test_bit(MPF_ABORT_QUEUE, &pgpath->pg->m->features))
+		blk_abort_queue(pgpath->path.dev->bdev->bd_disk->queue);
 }
 
 static struct priority_group *alloc_priority_group(void)
@@ -813,6 +820,11 @@ static int parse_features(struct arg_set
 			continue;
 		}
 
+		if (!strnicmp(param_name, MESG_STR("abort_queue_on_fail"))) {
+			set_bit(MPF_ABORT_QUEUE, &m->features);
+			continue;
+		}
+
 		if (!strnicmp(param_name, MESG_STR("pg_init_retries")) &&
 		    (argc >= 1)) {
 			r = read_param(_params + 1, shift(as),
@@ -1382,11 +1394,14 @@ static int multipath_status(struct dm_ta
 		DMEMIT("2 %u %u ", m->queue_size, m->pg_init_count);
 	else {
 		DMEMIT("%u ", m->queue_if_no_path +
-			      (m->pg_init_retries > 0) * 2);
+			      (m->pg_init_retries > 0) * 2 +
+			      test_bit(MPF_ABORT_QUEUE, &m->features));
 		if (m->queue_if_no_path)
 			DMEMIT("queue_if_no_path ");
 		if (m->pg_init_retries)
 			DMEMIT("pg_init_retries %u ", m->pg_init_retries);
+		if (test_bit(MPF_ABORT_QUEUE, &m->features))
+			DMEMIT("abort_queue_on_fail ");
 	}
 
 	if (!m->hw_handler_name || type == STATUSTYPE_INFO)
@@ -1655,7 +1670,7 @@ out:
  *---------------------------------------------------------------*/
 static struct target_type multipath_target = {
 	.name = "multipath",
-	.version = {1, 1, 1},
+	.version = {1, 2, 0},
 	.module = THIS_MODULE,
 	.ctr = multipath_ctr,
 	.dtr = multipath_dtr,

  parent reply	other threads:[~2010-11-18 15:48 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-05-12  5:23 block_abort_queue (blk_abort_request) racing with scsi_request_fn Mike Anderson
2010-11-10  7:09 ` [dm-devel] " Mike Christie
2010-11-10  7:30   ` Mike Christie
2010-11-10 16:30     ` Mike Anderson
2010-11-10 21:16       ` Mike Christie
2010-11-12 17:54         ` Mike Anderson
2010-11-16 21:39           ` Mike Snitzer
2010-11-17 17:49             ` [dm-devel] " Mike Anderson
2010-11-17 21:55               ` Mike Snitzer
2010-11-18  4:40                 ` [PATCH v2] dm mpath: add feature flag to control call to blk_abort_queue Mike Snitzer
2010-11-18  7:20                   ` Mike Anderson
2010-11-18 15:48                     ` Mike Snitzer
2010-11-18 15:48                     ` Mike Snitzer [this message]
2010-11-18 19:16                       ` (unknown), Mike Snitzer
2010-11-18 19:21                         ` Mike Snitzer
2010-11-18 19:19                       ` [PATCH v4] dm mpath: avoid call to blk_abort_queue by default Mike Snitzer
2010-11-18 20:07                         ` [PATCH v5] " Mike Snitzer
2010-11-18 20:18                           ` [dm-devel] " Alasdair G Kergon
2010-11-18 20:39                             ` Mike Anderson
2010-11-18 21:48                             ` [PATCH] dm mpath: disable call to blk_abort_queue and related code Mike Snitzer
2010-11-23  1:00                               ` [PATCH v2] dm mpath: revert "dm: Call blk_abort_queue on failed paths" Mike Snitzer

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1290095317-15684-1-git-send-email-snitzer@redhat.com \
    --to=snitzer@redhat.com \
    --cc=andmike@linux.vnet.ibm.com \
    --cc=dm-devel@redhat.com \
    --cc=linux-scsi@vger.kernel.org \
    --cc=michaelc@cs.wisc.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.