From: Andreas Gruenbacher <[email protected]>
[ Upstream commit 4e58543e7da4859c4ba61d15493e3522b6ad71fd ]
It turns out that the .freeze_super and .thaw_super operations require
the filesystem to manage the superblock refcount itself. We are using
the freeze_super() and thaw_super() helpers to mostly take care of that
for us, but this means that the superblock may no longer be around by
when thaw_super() returns, and gfs2_thaw_super() will then access freed
memory. Take an extra superblock reference in gfs2_thaw_super() to fix
that.
Signed-off-by: Andreas Gruenbacher <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
fs/gfs2/super.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c
index 29157f7d9663..2993598d18a4 100644
--- a/fs/gfs2/super.c
+++ b/fs/gfs2/super.c
@@ -1013,6 +1013,7 @@ static int gfs2_freeze(struct super_block *sb)
goto out;
}
+ atomic_inc(&sb->s_active);
for (;;) {
error = gfs2_lock_fs_check_clean(sdp, &sdp->sd_freeze_gh);
if (!error)
@@ -1034,6 +1035,7 @@ static int gfs2_freeze(struct super_block *sb)
error = 0;
out:
mutex_unlock(&sdp->sd_freeze_mutex);
+ deactivate_super(sb);
return error;
}
--
2.43.0
On Tue, Jan 16, 2024 at 9:53 PM Pavel Machek <[email protected]> wrote:
> Hi!
>
> > From: Andreas Gruenbacher <[email protected]>
> >
> > [ Upstream commit 4e58543e7da4859c4ba61d15493e3522b6ad71fd ]
> >
> > It turns out that the .freeze_super and .thaw_super operations require
> > the filesystem to manage the superblock refcount itself. We are using
> > the freeze_super() and thaw_super() helpers to mostly take care of that
> > for us, but this means that the superblock may no longer be around by
> > when thaw_super() returns, and gfs2_thaw_super() will then access freed
> > memory. Take an extra superblock reference in gfs2_thaw_super() to fix
> > that.
>
> Patch was broken during backport.
>
> > +++ b/fs/gfs2/super.c
> > @@ -1013,6 +1013,7 @@ static int gfs2_freeze(struct super_block *sb)
> > goto out;
> > }
> >
> > + atomic_inc(&sb->s_active);
> > for (;;) {
> > error = gfs2_lock_fs_check_clean(sdp, &sdp->sd_freeze_gh);
> > if (!error)
> > @@ -1034,6 +1035,7 @@ static int gfs2_freeze(struct super_block *sb)
> > error = 0;
> > out:
> > mutex_unlock(&sdp->sd_freeze_mutex);
> > + deactivate_super(sb);
> > return error;
> > }
>
> Notice the goto out? That now jumps around the atomic_inc, but we
> still do decrease. This will break 4.19, please fix or drop.
Thanks, Pavel.
Sasha, you don't want that fix without "gfs2: Rework freeze / thaw
logic" and the follow-up fixes, and backporting that probably isn't
going to be worth it.
Thanks,
Andreas
On Thu, Jan 18, 2024 at 12:50:37PM +0100, Andreas Gruenbacher wrote:
>On Tue, Jan 16, 2024 at 9:53 PM Pavel Machek <[email protected]> wrote:
>> Hi!
>>
>> > From: Andreas Gruenbacher <[email protected]>
>> >
>> > [ Upstream commit 4e58543e7da4859c4ba61d15493e3522b6ad71fd ]
>> >
>> > It turns out that the .freeze_super and .thaw_super operations require
>> > the filesystem to manage the superblock refcount itself. We are using
>> > the freeze_super() and thaw_super() helpers to mostly take care of that
>> > for us, but this means that the superblock may no longer be around by
>> > when thaw_super() returns, and gfs2_thaw_super() will then access freed
>> > memory. Take an extra superblock reference in gfs2_thaw_super() to fix
>> > that.
>>
>> Patch was broken during backport.
>>
>> > +++ b/fs/gfs2/super.c
>> > @@ -1013,6 +1013,7 @@ static int gfs2_freeze(struct super_block *sb)
>> > goto out;
>> > }
>> >
>> > + atomic_inc(&sb->s_active);
>> > for (;;) {
>> > error = gfs2_lock_fs_check_clean(sdp, &sdp->sd_freeze_gh);
>> > if (!error)
>> > @@ -1034,6 +1035,7 @@ static int gfs2_freeze(struct super_block *sb)
>> > error = 0;
>> > out:
>> > mutex_unlock(&sdp->sd_freeze_mutex);
>> > + deactivate_super(sb);
>> > return error;
>> > }
>>
>> Notice the goto out? That now jumps around the atomic_inc, but we
>> still do decrease. This will break 4.19, please fix or drop.
>
>Thanks, Pavel.
>
>Sasha, you don't want that fix without "gfs2: Rework freeze / thaw
>logic" and the follow-up fixes, and backporting that probably isn't
>going to be worth it.
I'll drop it, thanks!
--
Thanks,
Sasha