From: Peng Tao <bergwolf@gmail.com>
To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linux-kernel@vger.kernel.org,
Andriy Skulysh <Andriy_Skulysh@xyratex.com>,
Peng Tao <tao.peng@emc.com>,
Andreas Dilger <andreas.dilger@intel.com>
Subject: [PATCH 31/48] staging/lustre/ldlm: Fix flock deadlock detection race
Date: Tue, 23 Jul 2013 00:06:52 +0800 [thread overview]
Message-ID: <1374509230-3324-32-git-send-email-bergwolf@gmail.com> (raw)
In-Reply-To: <1374509230-3324-1-git-send-email-bergwolf@gmail.com>
From: Andriy Skulysh <Andriy_Skulysh@xyratex.com>
Deadlock isn't detected if 2 threads are trying to
grant 2 locks which deadlock on each other.
They call ldlm_flock_deadlock() simultaneously
and deadlock ins't detected.
The soulition is to add lock to blocking list before
calling ldlm_flock_deadlock()
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1602
Lustre-change: http://review.whamcloud.com/3277
Signed-off-by: Andriy Skulysh <Andriy_Skulysh@xyratex.com>
Reviewed-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Reviewed-by: Bruce Korb <bruce_korb@xyratex.com>
Reviewed-by: Keith Mannthey <keith.mannthey@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Peng Tao <tao.peng@emc.com>
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
---
drivers/staging/lustre/lustre/ldlm/ldlm_flock.c | 16 ++++++++++------
1 file changed, 10 insertions(+), 6 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
index aca1073..39973cb 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
@@ -198,6 +198,7 @@ ldlm_flock_deadlock(struct ldlm_lock *req, struct ldlm_lock *bl_lock)
if (lock == NULL)
break;
+ LASSERT(req != lock);
flock = &lock->l_policy_data.l_flock;
LASSERT(flock->owner == bl_owner);
bl_owner = flock->blocking_owner;
@@ -329,18 +330,21 @@ reprocess:
RETURN(LDLM_ITER_STOP);
}
- if (ldlm_flock_deadlock(req, lock)) {
- ldlm_flock_destroy(req, mode, *flags);
- *err = -EDEADLK;
- RETURN(LDLM_ITER_STOP);
- }
-
+ /* add lock to blocking list before deadlock
+ * check to prevent race */
rc = ldlm_flock_blocking_link(req, lock);
if (rc) {
ldlm_flock_destroy(req, mode, *flags);
*err = rc;
RETURN(LDLM_ITER_STOP);
}
+ if (ldlm_flock_deadlock(req, lock)) {
+ ldlm_flock_blocking_unlink(req);
+ ldlm_flock_destroy(req, mode, *flags);
+ *err = -EDEADLK;
+ RETURN(LDLM_ITER_STOP);
+ }
+
ldlm_resource_add_lock(res, &res->lr_waiting, req);
*flags |= LDLM_FL_BLOCK_GRANTED;
RETURN(LDLM_ITER_STOP);
--
1.7.9.5
next prev parent reply other threads:[~2013-07-22 16:12 UTC|newest]
Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-07-22 16:06 [PATCH 00/48] staging/lustre: minor cleanup and Intel Lustre sync Peng Tao
2013-07-22 16:06 ` [PATCH 01/48] staging/lustre: remove bogus ifndef EXPORT_SYMBOL Peng Tao
2013-07-22 16:06 ` [PATCH 02/48] staging/lustre/obdclass: restore lu_ref Peng Tao
2013-07-23 19:37 ` Greg Kroah-Hartman
2013-07-23 20:31 ` Dilger, Andreas
2013-07-23 20:52 ` Greg Kroah-Hartman
2013-07-24 6:06 ` Peng Tao
2013-07-22 16:06 ` [PATCH 03/48] staging/lustre/llite: use READ, WRITE around ll_rw_stats_tally() Peng Tao
2013-07-22 16:06 ` [PATCH 04/48] staging/lustre/llite: missing last bit in ll_have_md_lock Peng Tao
2013-07-22 16:06 ` [PATCH 05/48] staging/lustre: fix 'program hangs' errors Peng Tao
2013-07-22 16:06 ` [PATCH 06/48] staging/lustre/llite: check alloc in ll_file_data_get, ll_dir_ioctl Peng Tao
2013-07-22 16:06 ` [PATCH 07/48] staging/lustre/llite: A not locked mutex can be unlocked Peng Tao
2013-07-22 16:06 ` [PATCH 08/48] staging/lustre/llite: check ll_prep_md_op_data() using IS_ERR() Peng Tao
2013-07-22 16:06 ` [PATCH 09/48] staging/lustre/ldlm: print FID in lvbo_init(), lvbo_update Peng Tao
2013-07-22 16:06 ` [PATCH 10/48] staging/lustre/ptlrpc: race in pinger (use-after-free situation) Peng Tao
2013-07-22 16:06 ` [PATCH 11/48] staging/lustre/ptlrpc: Translate between host and network errnos Peng Tao
2013-07-22 16:29 ` Paul Bolle
2013-07-22 16:36 ` Peng Tao
2013-07-22 16:06 ` [PATCH 12/48] staging/lustre/mdc: layout lock rpc must not take rpc_lock Peng Tao
2013-07-22 16:06 ` [PATCH 13/48] staging/lustre/ldlm: split client namespaces into active and inactive Peng Tao
2013-07-22 16:06 ` [PATCH 14/48] staging/lustre: Only wake up ldlm_poold as frequently as the check interval Peng Tao
2013-07-22 16:06 ` [PATCH 15/48] staging/lustre: Make quota namespace refcounting consistent Peng Tao
2013-07-22 16:06 ` [PATCH 16/48] staging/lustre/dlmlock: compress out unused space Peng Tao
2013-07-22 16:06 ` [PATCH 17/48] staging/lustre/md: fix lu_ucred.c boilerplate Peng Tao
2013-07-22 16:06 ` [PATCH 18/48] staging/lustre/layout: introduce new layout for released files Peng Tao
2013-07-22 16:06 ` [PATCH 19/48] staging/lustre/mdt: add macros for fid string len Peng Tao
2013-07-22 16:06 ` [PATCH 20/48] staging/lustre/llapi: add user space method for lov_user_md Peng Tao
2013-07-22 16:06 ` [PATCH 21/48] staging/lustre/obdclass: use a dummy structure for lu_ref_link Peng Tao
2013-07-23 20:35 ` Greg Kroah-Hartman
2013-07-24 6:10 ` Peng Tao
2013-07-22 16:06 ` [PATCH 22/48] staging/lustre: fix 'code maintainability' errors Peng Tao
2013-07-22 16:06 ` [PATCH 23/48] staging/lustre/fid: prepare FID module for client server split Peng Tao
2013-07-22 16:06 ` [PATCH 24/48] staging/lustre/llite: force lvb_data update after layout change Peng Tao
2013-07-22 16:06 ` [PATCH 25/48] staging/lustre/lfsck: LFSCK 1.5 technical debts (3) Peng Tao
2013-07-22 16:06 ` [PATCH 26/48] staging/lustre/osc: Check return code for lu_kmem_init Peng Tao
2013-07-22 16:06 ` [PATCH 27/48] staging/lustre/ptlrpc: Race between start and stop service threads Peng Tao
2013-07-22 16:06 ` [PATCH 28/48] staging/lustre/crypto: add crc32c module loading to libcfs Peng Tao
2013-07-22 16:06 ` [PATCH 29/48] staging/lustre/mdt: duplicate link names in directory Peng Tao
2013-07-22 16:06 ` [PATCH 30/48] staging/lustre/llite: call simple_setattr() from ll_md_setattr() Peng Tao
2013-07-22 16:06 ` Peng Tao [this message]
2013-07-22 16:06 ` [PATCH 32/48] staging/lustre/lnet: remove empty file lnet/lnet/api-errno.c Peng Tao
2013-07-22 16:06 ` [PATCH 33/48] staging/lustre/fld: prepare FLD module for client server split Peng Tao
2013-07-22 16:06 ` [PATCH 34/48] staging/lustre/lmv: support DNE with HSM Peng Tao
2013-07-22 16:06 ` [PATCH 35/48] " Peng Tao
2013-07-23 20:39 ` Greg Kroah-Hartman
2013-07-22 16:06 ` [PATCH 36/48] staging/lustre/obdclass: add obd_target.h Peng Tao
2013-07-23 20:41 ` Greg Kroah-Hartman
2013-07-22 16:06 ` [PATCH 37/48] staging/lustre/procfs: return -ENOMEM from lprocfs_register() Peng Tao
2013-07-22 16:06 ` [PATCH 38/48] staging/lustre/lmv: fix duplicate directory entries Peng Tao
2013-07-22 16:07 ` [PATCH 39/48] staging/lustre/obdclass: be more careful processing server name Peng Tao
2013-07-22 16:07 ` [PATCH 40/48] staging/lustre/llite: return valid fsid for statfs Peng Tao
2013-07-22 16:07 ` [PATCH 41/48] staging/lustre/llite: error of listxattr when buffer is small Peng Tao
2013-07-22 16:07 ` [PATCH 42/48] staging/lustre/llite: Anonymous dentry incorrectly identified as root Peng Tao
2013-07-22 16:07 ` [PATCH 43/48] staging/lustre/build: fix 'data race condition' issues Peng Tao
2013-07-22 16:07 ` [PATCH 44/48] " Peng Tao
2013-07-23 20:44 ` Greg Kroah-Hartman
2013-07-22 16:07 ` [PATCH 45/48] staging/lustre/style: removes obsolete EXPORT_SYMTAB macros v2 Peng Tao
2013-07-22 16:07 ` [PATCH 46/48] staging/lustre/mdt: HSM on disk actions record Peng Tao
2013-07-22 16:07 ` [PATCH 47/48] staging/lustre/scrub: purge inconsistenct objects after OI scrub Peng Tao
2013-07-22 16:07 ` [PATCH 48/48] staging/lustre/mdc: Keep resend FLocks Peng Tao
2013-07-23 20:51 ` [PATCH 00/48] staging/lustre: minor cleanup and Intel Lustre sync Greg Kroah-Hartman
2013-07-24 6:05 ` Peng Tao
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=1374509230-3324-32-git-send-email-bergwolf@gmail.com \
--to=bergwolf@gmail.com \
--cc=Andriy_Skulysh@xyratex.com \
--cc=andreas.dilger@intel.com \
--cc=gregkh@linuxfoundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=tao.peng@emc.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