Skip to contentSkip to conversation

mqueue: sys_mq_open: do not call mnt_drop_write() if read-only

PATCH In review Vladimir Davydov 2 participants 3 messages Mar 19, 2013

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mqueue: sys_mq_open: do not call mnt_drop_write() if read-only

mnt_drop_write() must be called only if mnt_want_write() succeeded, otherwise the mnt_writers counter will diverge.

Cc Doug Ledford <[email protected]>
Cc Andrew Morton <[email protected]>
Cc KOSAKI Motohiro <[email protected]>
Cc "Eric W. Biederman" <[email protected]>
---
 ipc/mqueue.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
ipc/mqueue.c +2 -1
index e5c4f60..3953fda 100644
@@ -840,7 +840,8 @@ out_putfd:
 		fd = error;
 	}
 	mutex_unlock(&root->d_inode->i_mutex);
-	mnt_drop_write(mnt);
+	if (!ro)
+		mnt_drop_write(mnt);
 out_putname:
 	putname(name);
 	return fd;
Signature · 1 line
1.7.1
#1 lore raw
Quoted 8 lines · On Tue, 19 Mar 2013 13:31:18 +0400 Vladimir Davydov <[email protected]>…

mnt_drop_write() must be called only if mnt_want_write() succeeded, otherwise the mnt_writers counter will diverge.

...

ipc/mqueue.c out_putfd: · line 847

Quoted 9 lines
ipc/mqueue.c +2 -1
@@ -840,7 +840,8 @@ out_putfd:
 		fd = error;
 	}
 	mutex_unlock(&root->d_inode->i_mutex);
-	mnt_drop_write(mnt);
+	if (!ro)
+		mnt_drop_write(mnt);
 out_putname:
 	putname(name);
 	return fd;

huh, that's been there for a while. What were the runtime-visible effects of the bug?

#2 lore raw
Quoted 7 lines · Andrew Morton
Quoted 5 lines · On Tue, 19 Mar 2013 13:31:18 +0400 Vladimir Davydov <[email protected]>…

mnt_drop_write() must be called only if mnt_want_write() succeeded, otherwise the mnt_writers counter will diverge.

...

Quoted 15 lines
Quoted 10 lines
ipc/mqueue.c +2 -1
@@ -840,7 +840,8 @@ out_putfd:
 		fd = error;
 	}
 	mutex_unlock(&root->d_inode->i_mutex);
-	mnt_drop_write(mnt);
+	if (!ro)
+		mnt_drop_write(mnt);

out_putname:

Quoted 2 lines
	putname(name);
	return fd;

huh, that's been there for a while. What were the runtime-visible effects of the bug?

mnt_writers counters are used to check if remounting FS as read-only is OK, so after an extra mnt_drop_write() call, it would be impossible to remount mqueue FS as read-only. Besides, on umount a warning would be printed like this one:

[  194.714880] =====================================
[  194.719680] [ BUG: bad unlock balance detected! ]
[  194.724488] 3.9.0-rc3 #5 Not tainted
[  194.728159] -------------------------------------
[  194.732958] a.out/12486 is trying to release lock (sb_writers) at:
[  194.739355] [<ffffffff811b177f>] mnt_drop_write+0x1f/0x30
[  194.744851] but there are no more locks to release!