[PATCH] fs/namespace: fix mntput of ERR_PTR in fsmount error path

Deepanshu Kartikey posted 1 patch 3 days, 17 hours ago
fs/namespace.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
[PATCH] fs/namespace: fix mntput of ERR_PTR in fsmount error path
Posted by Deepanshu Kartikey 3 days, 17 hours ago
When vfs_create_mount() fails and returns an error pointer, the
__free(path_put) cleanup attribute causes path_put() to be called
on the error pointer, which then calls mntput() on an invalid
pointer value (e.g., -ENOENT = 0xfffffffffffffff4).

This results in a general protection fault in mntput() when KASAN
tries to check the shadow memory for the near-null address computed
from the error pointer offset.

Fix this by clearing newmount.mnt to NULL after extracting the error
code, preventing the path_put cleanup from operating on the error
pointer.

Reported-by: syzbot+94048264da5715c251f9@syzkaller.appspotmail.com
Tested-by: syzbot+94048264da5715c251f9@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=94048264da5715c251f9
Fixes: 67c68da01266 ("namespace: convert fsmount() to FD_PREPARE()")
Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
---
 fs/namespace.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/fs/namespace.c b/fs/namespace.c
index e5240df614de..236482fd503f 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -4343,8 +4343,11 @@ SYSCALL_DEFINE3(fsmount, int, fs_fd, unsigned int, flags,
 		warn_mandlock();
 
 	newmount.mnt = vfs_create_mount(fc);
-	if (IS_ERR(newmount.mnt))
-		return PTR_ERR(newmount.mnt);
+	if (IS_ERR(newmount.mnt)) {
+		ret = PTR_ERR(newmount.mnt);
+		newmount.mnt = NULL;
+		return ret;
+	}
 	newmount.dentry = dget(fc->root);
 	newmount.mnt->mnt_flags = mnt_flags;
 
-- 
2.43.0
Re: [PATCH] fs/namespace: fix mntput of ERR_PTR in fsmount error path
Posted by Christian Brauner 3 days, 11 hours ago
On Fri, 28 Nov 2025 09:21:48 +0530, Deepanshu Kartikey wrote:
> When vfs_create_mount() fails and returns an error pointer, the
> __free(path_put) cleanup attribute causes path_put() to be called
> on the error pointer, which then calls mntput() on an invalid
> pointer value (e.g., -ENOENT = 0xfffffffffffffff4).
> 
> This results in a general protection fault in mntput() when KASAN
> tries to check the shadow memory for the near-null address computed
> from the error pointer offset.
> 
> [...]

Thanks! Heavily rewritten but I retained your authorship.

---

Applied to the vfs-6.19.fd_prepare branch of the vfs/vfs.git tree.
Patches in the vfs-6.19.fd_prepare branch should appear in linux-next soon.

Please report any outstanding bugs that were missed during review in a
new review to the original patch series allowing us to drop it.

It's encouraged to provide Acked-bys and Reviewed-bys even though the
patch has now been applied. If possible patch trailers will be updated.

Note that commit hashes shown below are subject to change due to rebase,
trailer updates or similar. If in doubt, please check the listed branch.

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
branch: vfs-6.19.fd_prepare

[1/1] fs/namespace: fix mntput of ERR_PTR in fsmount error path
      https://git.kernel.org/vfs/vfs/c/721ebc9c4a95