diff options
| author | John Stultz <johnstul@us.ibm.com> | 2010-02-22 18:09:56 -0800 |
|---|---|---|
| committer | Thomas Gleixner <tglx@linutronix.de> | 2010-04-27 17:32:59 +0200 |
| commit | 070976b5b038218900648ea4cc88786d5dfcd58d (patch) | |
| tree | 1c5ffc0d3ba5006981a8e328cee270e96183a584 | |
| parent | 23695896abe0d1c813c2c88f0dd165d435d08689 (diff) | |
| download | lwn-070976b5b038218900648ea4cc88786d5dfcd58d.tar.gz lwn-070976b5b038218900648ea4cc88786d5dfcd58d.zip | |
Fix inc/dec_mnt_count for -rt
With Nick's vfs patches, inc/dec_mnt_count use per-cpu counters, so
this patch makes sure we disable preemption before calling.
Its not a great fix, but works because count_mnt_count() sums all the
percpu values, so each one individually doesn't need to be 0'ed out.
I suspect the better fix for -rt is to revert the mnt_count back to an atomic
counter.
Signed-off-by: John Stultz <johnstul@us.ibm.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
| -rw-r--r-- | fs/namespace.c | 8 |
1 files changed, 7 insertions, 1 deletions
diff --git a/fs/namespace.c b/fs/namespace.c index 7fb7fe3da44e..c1632069b06c 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -757,7 +757,9 @@ void mntput_no_expire(struct vfsmount *mnt) vfsmount_read_unlock(cpu); goto repeat; } + preempt_disable(); dec_mnt_count(mnt); + preempt_enable(); vfsmount_read_unlock(cpu); return; @@ -766,7 +768,9 @@ void mntput_no_expire(struct vfsmount *mnt) repeat: vfsmount_write_lock(); BUG_ON(mnt->mnt_flags & MNT_MOUNTED); + preempt_disable(); dec_mnt_count(mnt); + preempt_enable(); if (count_mnt_count(mnt)) { vfsmount_write_unlock(); return; @@ -819,7 +823,9 @@ void mnt_unpin(struct vfsmount *mnt) { vfsmount_write_lock(); if (mnt->mnt_pinned) { - inc_mnt_count(mnt); + preempt_disable(); + dec_mnt_count(mnt); + preempt_enable(); mnt->mnt_pinned--; } vfsmount_write_unlock(); |