From: Bart Van Assche <bvanassche@acm.org>
Cc: linux-scsi <linux-scsi@vger.kernel.org>,
James Bottomley <jbottomley@parallels.com>,
Mike Christie <michaelc@cs.wisc.edu>,
Jens Axboe <axboe@kernel.dk>, Tejun Heo <tj@kernel.org>,
Chanho Min <chanho.min@lge.com>, Hannes Reinecke <hare@suse.de>
Subject: [PATCH v6 05/13] block: Make blk_cleanup_queue() wait until request_fn finished
Date: Wed, 28 Nov 2012 13:46:45 +0100 [thread overview]
Message-ID: <50B607B5.7020808@acm.org> (raw)
In-Reply-To: <50B60619.4080406@acm.org>
Some request_fn implementations, e.g. scsi_request_fn(), unlock
the queue lock internally. This may result in multiple threads
executing request_fn for the same queue simultaneously. Keep
track of the number of active request_fn calls and make sure that
blk_cleanup_queue() waits until all active request_fn invocations
have finished. A block driver may start cleaning up resources
needed by its request_fn as soon as blk_cleanup_queue() finished,
so blk_cleanup_queue() must wait for all outstanding request_fn
invocations to finish.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reported-by: Chanho Min <chanho.min@lge.com>
Cc: James Bottomley <JBottomley@Parallels.com>
Cc: Mike Christie <michaelc@cs.wisc.edu>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Tejun Heo <tj@kernel.org>
---
block/blk-core.c | 10 ++++++++++
include/linux/blkdev.h | 6 ++++++
2 files changed, 16 insertions(+)
diff --git a/block/blk-core.c b/block/blk-core.c
index 12deca9..4d09af4 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -309,7 +309,16 @@ inline void __blk_run_queue_uncond(struct request_queue *q)
if (unlikely(blk_queue_dead(q)))
return;
+ /*
+ * Some request_fn implementations, e.g. scsi_request_fn(), unlock
+ * the queue lock internally. As a result multiple threads may be
+ * running such a request function concurrently. Keep track of the
+ * number of active request_fn invocations such that blk_drain_queue()
+ * can wait until all these request_fn calls have finished.
+ */
+ q->request_fn_active++;
q->request_fn(q);
+ q->request_fn_active--;
}
/**
@@ -408,6 +417,7 @@ static void __blk_drain_queue(struct request_queue *q, bool drain_all)
__blk_run_queue(q);
drain |= q->nr_rqs_elvpriv;
+ drain |= q->request_fn_active;
/*
* Unfortunately, requests are queued at and tracked from
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 8bc46c2..c9d233e 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -378,6 +378,12 @@ struct request_queue {
unsigned int nr_sorted;
unsigned int in_flight[2];
+ /*
+ * Number of active block driver functions for which blk_drain_queue()
+ * must wait. Must be incremented around functions that unlock the
+ * queue_lock internally, e.g. scsi_request_fn().
+ */
+ unsigned int request_fn_active;
unsigned int rq_timeout;
struct timer_list timeout;
--
1.7.10.4
next prev parent reply other threads:[~2012-11-28 12:46 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-11-28 12:39 [PATCH 0/13 v6] More device removal fixes Bart Van Assche
2012-11-28 12:42 ` [PATCH v6 01/13] block: Rename queue dead flag Bart Van Assche
2012-11-28 12:43 ` [PATCH v6 02/13] block: Let blk_drain_queue() caller obtain the queue lock Bart Van Assche
2012-11-28 12:44 ` [PATCH v6 03/13] block: Avoid that request_fn is invoked on a dead queue Bart Van Assche
2012-12-02 13:23 ` Tejun Heo
2012-12-02 13:35 ` Bart Van Assche
2012-11-28 12:45 ` [PATCH v6 04/13] block: Avoid scheduling delayed work " Bart Van Assche
2012-12-02 13:26 ` Tejun Heo
2012-12-02 13:41 ` Bart Van Assche
2012-12-02 13:59 ` Tejun Heo
2012-11-28 12:46 ` Bart Van Assche [this message]
2012-12-02 13:28 ` [PATCH v6 05/13] block: Make blk_cleanup_queue() wait until request_fn finished Tejun Heo
2012-11-28 12:47 ` [PATCH v6 06/13] bsg: Remove unused function bsg_goose_queue() Bart Van Assche
2012-12-02 13:29 ` Tejun Heo
2012-11-28 12:48 ` [PATCH v6 07/13] Fix race between starved list processing and device removal Bart Van Assche
2012-12-02 13:32 ` Tejun Heo
2012-11-28 12:48 ` [PATCH v6 08/13] Remove get_device() / put_device() pair from scsi_request_fn() Bart Van Assche
2012-12-02 13:34 ` Tejun Heo
2012-11-28 12:50 ` [PATCH v6 09/13] Avoid saving/restoring interrupt state inside scsi_remove_host() Bart Van Assche
2012-12-02 13:35 ` Tejun Heo
2012-11-28 12:51 ` [PATCH v6 10/13] Make scsi_remove_host() wait for device removal Bart Van Assche
2012-12-02 13:45 ` Tejun Heo
2012-12-02 13:48 ` Tejun Heo
2012-12-03 8:23 ` Bart Van Assche
2012-12-03 16:15 ` Tejun Heo
2012-12-03 16:38 ` Bart Van Assche
2012-12-03 16:42 ` Tejun Heo
2012-12-07 7:41 ` Bart Van Assche
2012-11-28 12:52 ` [PATCH v6 11/13] Make scsi_remove_host() wait until error handling finished Bart Van Assche
2012-12-02 13:51 ` Tejun Heo
2012-11-28 12:53 ` [PATCH v6 12/13] Avoid that scsi_device_set_state() triggers a race Bart Van Assche
2012-12-02 13:53 ` Tejun Heo
2012-11-28 12:53 ` [PATCH v6 13/13] Do not queue new I/O after scsi_remove_host() started Bart Van Assche
2012-12-02 13:58 ` Tejun Heo
2012-12-02 14:02 ` [PATCH 0/13 v6] More device removal fixes Tejun Heo
2012-12-06 13:33 ` Jens Axboe
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=50B607B5.7020808@acm.org \
--to=bvanassche@acm.org \
--cc=axboe@kernel.dk \
--cc=chanho.min@lge.com \
--cc=hare@suse.de \
--cc=jbottomley@parallels.com \
--cc=linux-scsi@vger.kernel.org \
--cc=michaelc@cs.wisc.edu \
--cc=tj@kernel.org \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).