mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] gfs2: Fix NULL pointer dereference in gfs2_seek_data/gfs2_seek_hole
@ 2026-11-16 18:00 William Chuang
  2026-11-16 22:08 ` Andreas Gruenbacher
  0 siblings, 1 reply; 3+ messages in thread
From: William Chuang @ 2026-11-16 18:00 UTC (permalink / raw)
  To: Andreas Gruenbacher
  Cc: gfs2, linux-kernel, William Chuang, syzbot+5a5f492ccae698fd7434, stable

When the filesystem has been withdrawn, gfs2_glock_nq() returns -EIO
without queuing the holder, and gfs2_glock_nq_init() then calls
gfs2_holder_uninit(), which sets gh.gh_gl to NULL.

gfs2_seek_data() and gfs2_seek_hole() call gfs2_glock_dq_uninit()
unconditionally, so in that case gfs2_glock_dq() dereferences a NULL
glock when it takes gl->gl_lockref.lock, causing a general protection
fault.

Only release the holder when it was successfully acquired, matching the
pattern already used elsewhere in this file (such as gfs2_fiemap() and
gfs2_getattr()).

Fixes: 3a27411cb4bc ("gfs2: Implement SEEK_HOLE / SEEK_DATA via iomap")
Reported-by: syzbot+5a5f492ccae698fd7434@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=5a5f492ccae698fd7434
Cc: stable@vger.kernel.org
Signed-off-by: William Chuang <founder@logarcheon.com>
---
 fs/gfs2/inode.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 75fb1a0c91425..4110b2d0b549c 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -2255,7 +2255,8 @@ loff_t gfs2_seek_data(struct file *file, loff_t offset)
 	ret = gfs2_glock_nq_init(gl, LM_ST_SHARED, 0, &gh);
 	if (!ret)
 		ret = iomap_seek_data(inode, offset, gfs2_iomap_next);
-	gfs2_glock_dq_uninit(&gh);
+	if (gfs2_holder_initialized(&gh))
+		gfs2_glock_dq_uninit(&gh);
 	inode_unlock_shared(inode);
 
 	if (ret < 0)
@@ -2274,7 +2275,8 @@ loff_t gfs2_seek_hole(struct file *file, loff_t offset)
 	ret = gfs2_glock_nq_init(gl, LM_ST_SHARED, 0, &gh);
 	if (!ret)
 		ret = iomap_seek_hole(inode, offset, gfs2_iomap_next);
-	gfs2_glock_dq_uninit(&gh);
+	if (gfs2_holder_initialized(&gh))
+		gfs2_glock_dq_uninit(&gh);
 	inode_unlock_shared(inode);
 
 	if (ret < 0)
-- 
2.43.0



^ permalink raw reply	[flat|nested] 3+ messages in thread

* [PATCH] gfs2: Fix NULL pointer dereference in gfs2_seek_data/gfs2_seek_hole
  2026-11-16 18:00 [PATCH] gfs2: Fix NULL pointer dereference in gfs2_seek_data/gfs2_seek_hole William Chuang
@ 2026-11-16 22:08 ` Andreas Gruenbacher
  0 siblings, 0 replies; 3+ messages in thread
From: Andreas Gruenbacher @ 2026-11-16 22:08 UTC (permalink / raw)
  To: William Chuang
  Cc: Andreas Gruenbacher, gfs2, linux-kernel,
	syzbot+5a5f492ccae698fd7434, stable

On Mon, Nov 16, 2026 at 7:01 PM William Chuang <founder@logarcheon.com> wrote:
> When the filesystem has been withdrawn, gfs2_glock_nq() returns -EIO
> without queuing the holder, and gfs2_glock_nq_init() then calls
> gfs2_holder_uninit(), which sets gh.gh_gl to NULL.
>
> gfs2_seek_data() and gfs2_seek_hole() call gfs2_glock_dq_uninit()
> unconditionally, so in that case gfs2_glock_dq() dereferences a NULL
> glock when it takes gl->gl_lockref.lock, causing a general protection
> fault.
>
> Only release the holder when it was successfully acquired, matching the
> pattern already used elsewhere in this file (such as gfs2_fiemap() and
> gfs2_getattr()).
>
> Fixes: 3a27411cb4bc ("gfs2: Implement SEEK_HOLE / SEEK_DATA via iomap")
> Reported-by: syzbot+5a5f492ccae698fd7434@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=5a5f492ccae698fd7434
> Cc: stable@vger.kernel.org
> Signed-off-by: William Chuang <founder@logarcheon.com>

Thanks.

Could you please replace this with the following fix and resend?

Andreas

--

gfs2_seek_data() and gfs2_seek_hole() call gfs2_glock_nq_init(), and
independent of the outcome, they call gfs2_glock_dq_uninit().  When the
gfs2_glock_nq() call in gfs2_glock_nq_init() fails, gfs2_glock_nq_init()
returns with an error code and gh.gh_gl is set NULL.  This causes a NULL
pointer dereference in gfs2_glock_dq().

Fix that by only calling gfs2_glock_dq_uninit() when
gfs2_glock_nq_init() succeeds.

Fixes: 3a27411cb4bc ("gfs2: Implement SEEK_HOLE / SEEK_DATA via iomap")
Reported-by: syzbot+5a5f492ccae698fd7434@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=5a5f492ccae698fd7434
Cc: stable@vger.kernel.org
Signed-off-by: William Chuang <founder@logarcheon.com>

diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 2c2dc459c037..d4678e4832b3 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -2253,9 +2253,11 @@ loff_t gfs2_seek_data(struct file *file, loff_t offset)
 
 	inode_lock_shared(inode);
 	ret = gfs2_glock_nq_init(gl, LM_ST_SHARED, 0, &gh);
-	if (!ret)
-		ret = iomap_seek_data(inode, offset, &gfs2_iomap_ops);
+	if (ret)
+		goto out;
+	ret = iomap_seek_data(inode, offset, &gfs2_iomap_ops);
 	gfs2_glock_dq_uninit(&gh);
+out:
 	inode_unlock_shared(inode);
 
 	if (ret < 0)
@@ -2272,9 +2274,11 @@ loff_t gfs2_seek_hole(struct file *file, loff_t offset)
 
 	inode_lock_shared(inode);
 	ret = gfs2_glock_nq_init(gl, LM_ST_SHARED, 0, &gh);
-	if (!ret)
-		ret = iomap_seek_hole(inode, offset, &gfs2_iomap_ops);
+	if (ret)
+		goto out;
+	ret = iomap_seek_hole(inode, offset, &gfs2_iomap_ops);
 	gfs2_glock_dq_uninit(&gh);
+out:
 	inode_unlock_shared(inode);
 
 	if (ret < 0)


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] gfs2: Fix NULL pointer dereference in gfs2_seek_data/gfs2_seek_hole
       [not found] <20261116175346.44221-1-founder@logarcheon.com>
@ 2026-11-16 17:54 ` syzbot
  0 siblings, 0 replies; 3+ messages in thread
From: syzbot @ 2026-11-16 17:54 UTC (permalink / raw)
  To: founder; +Cc: founder, syzkaller-bugs, linux-kernel

> #syz test

Patch testing via #syz test is temporarily disabled.
See https://github.com/google/syzkaller/issues/8122 for details.

>
> When the filesystem has been withdrawn, gfs2_glock_nq() returns -EIO
> without queuing the holder, and gfs2_glock_nq_init() then calls
> gfs2_holder_uninit(), which sets gh.gh_gl to NULL.
>
> gfs2_seek_data() and gfs2_seek_hole() call gfs2_glock_dq_uninit()
> unconditionally, so in that case gfs2_glock_dq() dereferences a NULL
> glock when it takes gl->gl_lockref.lock, causing a general protection
> fault.
>
> Only release the holder when it was successfully acquired, matching the
> pattern already used elsewhere in this file (such as gfs2_fiemap() and
> gfs2_getattr()).
>
> Reported-by: syzbot+5a5f492ccae698fd7434@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=5a5f492ccae698fd7434
> Signed-off-by: William Chuang <founder@logarcheon.com>
> ---
>  fs/gfs2/inode.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
> index 75fb1a0..4110b2d 100644
> --- a/fs/gfs2/inode.c
> +++ b/fs/gfs2/inode.c
> @@ -2255,7 +2255,8 @@ loff_t gfs2_seek_data(struct file *file, loff_t offset)
>  	ret = gfs2_glock_nq_init(gl, LM_ST_SHARED, 0, &gh);
>  	if (!ret)
>  		ret = iomap_seek_data(inode, offset, gfs2_iomap_next);
> -	gfs2_glock_dq_uninit(&gh);
> +	if (gfs2_holder_initialized(&gh))
> +		gfs2_glock_dq_uninit(&gh);
>  	inode_unlock_shared(inode);
>  
>  	if (ret < 0)
> @@ -2274,7 +2275,8 @@ loff_t gfs2_seek_hole(struct file *file, loff_t offset)
>  	ret = gfs2_glock_nq_init(gl, LM_ST_SHARED, 0, &gh);
>  	if (!ret)
>  		ret = iomap_seek_hole(inode, offset, gfs2_iomap_next);
> -	gfs2_glock_dq_uninit(&gh);
> +	if (gfs2_holder_initialized(&gh))
> +		gfs2_glock_dq_uninit(&gh);
>  	inode_unlock_shared(inode);
>  
>  	if (ret < 0)
> -- 
> 2.43.0
>
>

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2026-11-16 22:08 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-16 18:00 [PATCH] gfs2: Fix NULL pointer dereference in gfs2_seek_data/gfs2_seek_hole William Chuang
2026-11-16 22:08 ` Andreas Gruenbacher
     [not found] <20261116175346.44221-1-founder@logarcheon.com>
2026-11-16 17:54 ` syzbot

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®