From: Alexander Aring <aahringo@redhat.com>
To: teigland@redhat.com
Cc: cluster-devel@redhat.com, bmarson@redhat.com, stable@vger.kernel.org
Subject: [Cluster-devel] [PATCH dlm/next] dlm: fix plock lookup when using multiple lockspaces
Date: Thu, 24 Aug 2023 16:51:42 -0400 [thread overview]
Message-ID: <20230824205142.2732984-1-aahringo@redhat.com> (raw)
This patch fixes an issues when concurrent fcntl() syscalls are
executing on two different gfs2 filesystems. Each gfs2 filesystem
creates an DLM lockspace, it seems that VFS only allows fcntl() syscalls
at one time on a per filesystem basis. However if there are two
filesystems and we executing fcntl() syscalls our lookup mechanism on the
global plock op list does not work anymore.
It can be reproduced with two mounted gfs2 filesystems using DLM
locking. Then call stress-ng --fcntl 32 on each mount point. The kernel
log will show several:
WARNING: CPU: 4 PID: 943 at fs/dlm/plock.c:574 dev_write+0x15c/0x590
because we have a sanity check if it's was really the meant original
plock op when dev_write() does a lookup. This patch adds just a
additional check for fsid to find the right plock op which is an
indicator that the recv_list should be on a per lockspace basis and not
globally defined. After this patch the sanity check never warned again
that the wrong plock op was being looked up.
Cc: stable@vger.kernel.org
Reported-by: Barry Marson <bmarson@redhat.com>
Fixes: 57e2c2f2d94c ("fs: dlm: fix mismatch of plock results from userspace")
Signed-off-by: Alexander Aring <aahringo@redhat.com>
---
fs/dlm/plock.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/fs/dlm/plock.c b/fs/dlm/plock.c
index 00e1d802a81c..e6b4c1a21446 100644
--- a/fs/dlm/plock.c
+++ b/fs/dlm/plock.c
@@ -556,7 +556,8 @@ static ssize_t dev_write(struct file *file, const char __user *u, size_t count,
op = plock_lookup_waiter(&info);
} else {
list_for_each_entry(iter, &recv_list, list) {
- if (!iter->info.wait) {
+ if (!iter->info.wait &&
+ iter->info.fsid == info.fsid) {
op = iter;
break;
}
@@ -568,8 +569,7 @@ static ssize_t dev_write(struct file *file, const char __user *u, size_t count,
if (info.wait)
WARN_ON(op->info.optype != DLM_PLOCK_OP_LOCK);
else
- WARN_ON(op->info.fsid != info.fsid ||
- op->info.number != info.number ||
+ WARN_ON(op->info.number != info.number ||
op->info.owner != info.owner ||
op->info.optype != info.optype);
--
2.31.1
next reply other threads:[~2023-08-24 20:51 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-24 20:51 Alexander Aring [this message]
2023-08-24 23:22 ` [Cluster-devel] [PATCH dlm/next] dlm: fix plock lookup when using multiple lockspaces Alexander Aring
2023-08-25 13:28 ` Alexander Aring
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=20230824205142.2732984-1-aahringo@redhat.com \
--to=aahringo@redhat.com \
--cc=bmarson@redhat.com \
--cc=cluster-devel@redhat.com \
--cc=stable@vger.kernel.org \
--cc=teigland@redhat.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 read-only IMAP folder(s) and NNTP newsgroup(s).