From: Yu Kuai <yukuai3@huawei.com>
To: <josef@toxicpanda.com>, <axboe@kernel.dk>, <ming.lei@redhat.com>,
<hch@infradead.org>
Cc: <linux-block@vger.kernel.org>, <nbd@other.debian.org>,
<linux-kernel@vger.kernel.org>, <yukuai3@huawei.com>,
<yi.zhang@huawei.com>
Subject: [patch v8 2/7] nbd: make sure request completion won't concurrent
Date: Thu, 16 Sep 2021 17:33:45 +0800 [thread overview]
Message-ID: <20210916093350.1410403-3-yukuai3@huawei.com> (raw)
In-Reply-To: <20210916093350.1410403-1-yukuai3@huawei.com>
commit cddce0116058 ("nbd: Aovid double completion of a request")
try to fix that nbd_clear_que() and recv_work() can complete a
request concurrently. However, the problem still exists:
t1 t2 t3
nbd_disconnect_and_put
flush_workqueue
recv_work
blk_mq_complete_request
blk_mq_complete_request_remote -> this is true
WRITE_ONCE(rq->state, MQ_RQ_COMPLETE)
blk_mq_raise_softirq
blk_done_softirq
blk_complete_reqs
nbd_complete_rq
blk_mq_end_request
blk_mq_free_request
WRITE_ONCE(rq->state, MQ_RQ_IDLE)
nbd_clear_que
blk_mq_tagset_busy_iter
nbd_clear_req
__blk_mq_free_request
blk_mq_put_tag
blk_mq_complete_request -> complete again
There are three places where request can be completed in nbd:
recv_work(), nbd_clear_que() and nbd_xmit_timeout(). Since they
all hold cmd->lock before completing the request, it's easy to
avoid the problem by setting and checking a cmd flag.
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
---
drivers/block/nbd.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
index 23ded569e8d3..614c6ab2b8fe 100644
--- a/drivers/block/nbd.c
+++ b/drivers/block/nbd.c
@@ -406,7 +406,11 @@ static enum blk_eh_timer_return nbd_xmit_timeout(struct request *req,
if (!mutex_trylock(&cmd->lock))
return BLK_EH_RESET_TIMER;
- __clear_bit(NBD_CMD_INFLIGHT, &cmd->flags);
+ if (!__test_and_clear_bit(NBD_CMD_INFLIGHT, &cmd->flags)) {
+ mutex_unlock(&cmd->lock);
+ return BLK_EH_DONE;
+ }
+
if (!refcount_inc_not_zero(&nbd->config_refs)) {
cmd->status = BLK_STS_TIMEOUT;
mutex_unlock(&cmd->lock);
@@ -841,7 +845,10 @@ static bool nbd_clear_req(struct request *req, void *data, bool reserved)
return true;
mutex_lock(&cmd->lock);
- __clear_bit(NBD_CMD_INFLIGHT, &cmd->flags);
+ if (!__test_and_clear_bit(NBD_CMD_INFLIGHT, &cmd->flags)) {
+ mutex_unlock(&cmd->lock);
+ return true;
+ }
cmd->status = BLK_STS_IOERR;
mutex_unlock(&cmd->lock);
--
2.31.1
next prev parent reply other threads:[~2021-09-16 9:24 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-16 9:33 [patch v8 0/7] handle unexpected message from server Yu Kuai
2021-09-16 9:33 ` [patch v8 1/7] nbd: don't handle response without a corresponding request message Yu Kuai
2021-09-16 9:33 ` Yu Kuai [this message]
2021-09-16 9:33 ` [patch v8 3/7] nbd: check sock index in nbd_read_stat() Yu Kuai
2021-09-19 10:34 ` yukuai (C)
2021-09-22 9:22 ` Ming Lei
2021-09-22 12:12 ` Eric Blake
2021-09-22 12:21 ` yukuai (C)
2021-09-22 15:56 ` Wouter Verhelst
2021-09-23 0:29 ` Ming Lei
2021-09-16 9:33 ` [patch v8 4/7] nbd: don't start request if nbd_queue_rq() failed Yu Kuai
2021-09-16 12:48 ` Ming Lei
2021-09-16 9:33 ` [patch v8 5/7] nbd: clean up return value checking of sock_xmit() Yu Kuai
2021-09-16 12:49 ` Ming Lei
2021-09-16 9:33 ` [patch v8 6/7] nbd: partition nbd_read_stat() into nbd_read_reply() and nbd_handle_reply() Yu Kuai
2021-09-16 12:53 ` Ming Lei
2021-09-16 9:33 ` [patch v8 7/7] nbd: fix uaf in nbd_handle_reply() Yu Kuai
2021-09-16 12:58 ` Ming Lei
2021-09-16 13:10 ` yukuai (C)
2021-09-16 13:55 ` Ming Lei
2021-09-16 14:05 ` yukuai (C)
2021-09-16 14:18 ` [PATCH v9] " Yu Kuai
2021-09-17 3:33 ` Ming Lei
2021-10-17 13:09 ` Jens Axboe
2021-09-23 13:33 ` [patch v8 0/7] handle unexpected message from server yukuai (C)
2021-09-29 12:54 ` yukuai (C)
2021-10-08 7:17 ` yukuai (C)
2021-10-15 8:19 ` yukuai (C)
2021-10-15 16:08 ` Josef Bacik
2021-10-17 13:09 ` 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=20210916093350.1410403-3-yukuai3@huawei.com \
--to=yukuai3@huawei.com \
--cc=axboe@kernel.dk \
--cc=hch@infradead.org \
--cc=josef@toxicpanda.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=ming.lei@redhat.com \
--cc=nbd@other.debian.org \
--cc=yi.zhang@huawei.com \
/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).