From: Menyhart Zoltan <Zoltan.Menyhart@bull.net>
To: linux-nfs@vger.kernel.org
Subject: "xprt" reference count drops to 0
Date: Tue, 05 Oct 2010 10:22:30 +0200 [thread overview]
Message-ID: <4CAAE046.5060209@bull.net> (raw)
In-Reply-To: <4C7E4469.70807@duchatelet.net>
[-- Attachment #1: Type: text/plain, Size: 471 bytes --]
Hi,
Due to some race conditions, the reference count can become 0
while "xprt" is still on a "pool":
WARNING: at lib/kref.c:43 kref_get+0x23/0x2d()
[] kref_get+0x23/0x2d
[] svc_xprt_get+0x12/0x14 [sunrpc]
[] svc_recv+0x2db/0x78a [sunrpc]
I think we should increase the reference counter before adding "xprt"
onto any list.
Obviously, we keep increasing the reference count before passing "xprt"
to another thread via "rqstp->rq_xprt".
Thanks,
Zoltan Menyhart
[-- Attachment #2: diff --]
[-- Type: text/plain, Size: 1027 bytes --]
"svc_xprt_get()" should be invoked before adding "xprt" onto "pool" by
"svc_xprt_enqueue()".
Otherwise "xprt->xpt_ref" can drop to 0 due to race conditions.
Signed-off-by: Zoltan Menyhart <Zoltan.Menyhart@bull.net>
--- old/net/sunrpc/svc_xprt.c 2010-10-05 09:47:51.000000000 +0200
+++ new/net/sunrpc/svc_xprt.c 2010-10-05 09:47:20.000000000 +0200
@@ -369,6 +369,7 @@
}
process:
+ svc_xprt_get(xprt);
if (!list_empty(&pool->sp_threads)) {
rqstp = list_entry(pool->sp_threads.next,
struct svc_rqst,
@@ -381,7 +382,6 @@
"svc_xprt_enqueue: server %p, rq_xprt=%p!\n",
rqstp, rqstp->rq_xprt);
rqstp->rq_xprt = xprt;
- svc_xprt_get(xprt);
rqstp->rq_reserved = serv->sv_max_mesg;
atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved);
pool->sp_stats.threads_woken++;
@@ -655,7 +655,6 @@
xprt = svc_xprt_dequeue(pool);
if (xprt) {
rqstp->rq_xprt = xprt;
- svc_xprt_get(xprt);
rqstp->rq_reserved = serv->sv_max_mesg;
atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved);
} else {
next prev parent reply other threads:[~2010-10-05 8:22 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-09-01 12:17 Relocate NFS root FS for maintenance Greg
2010-09-01 17:34 ` J. Bruce Fields
2010-09-01 21:52 ` Tom Haynes
2010-09-02 7:32 ` Greg
2010-09-02 16:06 ` J. Bruce Fields
2010-09-07 6:59 ` Greg
2010-09-02 6:56 ` statfs() gives ESTALE error Menyhart Zoltan
2010-09-07 18:32 ` Trond Myklebust
2010-09-08 13:33 ` Re :statfs() " Menyhart Zoltan
2010-09-08 20:25 ` Trond Myklebust
2010-09-09 8:12 ` Menyhart Zoltan
2010-09-20 12:49 ` Locking question around "...PagePrivate()" Menyhart Zoltan
2010-09-20 13:55 ` Trond Myklebust
2010-10-05 8:22 ` Menyhart Zoltan [this message]
2010-10-21 20:38 ` "xprt" reference count drops to 0 J. Bruce Fields
2010-10-22 15:00 ` Menyhart Zoltan
2010-10-22 21:20 ` J. Bruce Fields
2010-10-22 23:01 ` J. Bruce Fields
2010-10-22 23:21 ` J. Bruce Fields
2010-10-23 3:32 ` J. Bruce Fields
2010-10-25 1:09 ` J. Bruce Fields
2010-10-25 1:21 ` [PATCH 1/4] svcrpc: never clear XPT_BUSY on dead xprt J. Bruce Fields
2010-10-25 1:43 ` Neil Brown
2010-10-25 20:21 ` J. Bruce Fields
2010-10-25 22:58 ` Neil Brown
2010-10-25 23:03 ` J. Bruce Fields
2010-10-25 23:54 ` Neil Brown
2010-10-26 0:11 ` J. Bruce Fields
2010-10-26 0:28 ` J. Bruce Fields
2010-10-26 0:30 ` J. Bruce Fields
2010-10-26 1:28 ` Neil Brown
2010-10-26 12:59 ` J. Bruce Fields
2010-10-26 16:05 ` J. Bruce Fields
2010-11-12 19:00 ` J. Bruce Fields
2010-10-25 1:21 ` [PATCH 2/4] svcrpc: assume svc_delete_xprt() called only once J. Bruce Fields
2010-10-25 1:21 ` [PATCH 3/4] svcrpc: no need for XPT_DEAD check in svc_xprt_enqueue J. Bruce Fields
2010-10-25 1:21 ` [PATCH 4/4] svcrpc: svc_tcp_sendto XTP_DEAD check is redundant J. Bruce Fields
2010-10-25 2:10 ` Neil Brown
2010-10-25 15:03 ` J. Bruce Fields
2010-10-25 17:46 ` J. Bruce Fields
2010-10-25 23:08 ` Neil Brown
2010-10-26 1:33 ` J. Bruce Fields
2010-10-25 23:23 ` Neil Brown
2010-10-26 1:25 ` J. Bruce Fields
2010-10-25 11:56 ` "xprt" reference count drops to 0 Menyhart Zoltan
2010-10-25 14:36 ` J. Bruce Fields
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=4CAAE046.5060209@bull.net \
--to=zoltan.menyhart@bull.net \
--cc=linux-nfs@vger.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 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.