Re: [PATCH 0/3] ntfs: three fixes for crafted images, found by fuzzing
From: Namjae Jeon
Date: Fri Nov 06 2026 - 07:31:07 EST
On Fri, Nov 6, 2026 at 6:14 PM Wahid Khan <wahidzk0091@xxxxxxxxx> wrote:
>
> I fuzz the new fs/ntfs driver with mutated images: an image from mkntfs
> populated with ntfs-3g (directories, streams, sparse files, long names), a
> few bytes changed, then mount, walk the tree and write, under KASAN.
> Three problems from that are still present in ntfs-next (e62b384ca).
> Each patch has its own trace and the reproduction steps.
>
> 1. ntfs: do not read an index block outside $INDEX_ALLOCATION
> A directory without a usable $INDEX_ALLOCATION hangs on its first
> lookup. The task waits for an mft record lock that it already holds.
> It needs one changed byte in a crafted image.
>
> 2. ntfs: open the index allocation before reading or writing an index block
> NULL pointer dereference in ntfs_ib_write() on link() when the lookup
> ended in the index root. This is not the path fixed by 9365c4a1357e.
> It needs one changed byte in a crafted image.
>
> 3. ntfs: open a named $DATA stream by name when making room in a record
> WARN_ON() in ntfs_attr_make_non_resident(). This one does not need a
> damaged volume: a file with an alternate data stream and a full mft
> record is enough.
>
> All three need an image to be mounted, so a crafted image, not a remote
> attacker. I can send the images (about 380 KB each, compressed) if they
> are useful.
>
>
> Wahid Khan (3):
> ntfs: do not read an index block outside $INDEX_ALLOCATION
> ntfs: open the index allocation before reading or writing an index
> block
> ntfs: open a named $DATA stream by name when making room in a record
Please check Sashiko's review comments:
https://sashiko.dev/#/patchset/20261106091420.2461683-1-wahidzk0091%40gmail.com
Thanks.