From: Henry Martin <bsdhenrymartin@gmail.com>
To: Konstantin Komarov <almaz.alexandrovich@paragon-software.com>
Cc: ntfs3@lists.linux.dev, linux-fsdevel@vger.kernel.org,
linux-kernel@vger.kernel.org,
Henry Martin <bsdhenrymartin@gmail.com>,
stable@vger.kernel.org
Subject: [PATCH] fs/ntfs3: fix UAF in al_destroy() on stream inode eviction
Date: Tue, 27 Oct 2026 23:16:14 +0800 [thread overview]
Message-ID: <20261027151615.971171-1-bsdhenrymartin@gmail.com> (raw)
al_destroy() unconditionally dereferences ni->base and frees the shared
base attribute list:
void al_destroy(struct ntfs_inode *ni)
{
ni = ni->base;
run_close(&ni->attr_list.run);
kvfree(ni->attr_list.le);
ni->attr_list.le = NULL;
ni->attr_list.size = 0;
ni->attr_list.dirty = false;
}
ni_clear() (fs/ntfs3/frecord.c) calls al_destroy() for every inode
being evicted, including ADS stream inodes whose ->base still points
at the base inode: the base's attr_list.le is freed while the base
inode is still alive (iput(&ni->base->vfs_inode) happens only at the
end of ni_clear()), racing the base's ni_lock-held attr_list consumers
(attr_set_size, al_add_le, ntfs_load_attr_list).
Stream inodes have no attr_list of their own; freeing the base's list
on their eviction also leaves the base inode with a destroyed list
(size=0, le=NULL) even without a race.
Only destroy the attribute list for inodes that own it: a stream
inode's ->base points at another inode, while on a base inode it is
self-referencing (fs/ntfs3/inode.c), so the ownership check is
ni->base == ni.
This issue was discovered by Tencent CodeBuddy Security.
Fixes: be71b5cba2e6 ("fs/ntfs3: Add attrib operations")
Cc: stable@vger.kernel.org
Signed-off-by: Henry Martin <bsdhenrymartin@gmail.com>
---
fs/ntfs3/attrlist.c | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/fs/ntfs3/attrlist.c b/fs/ntfs3/attrlist.c
index 8710560cf2944..6aced486a45c7 100644
--- a/fs/ntfs3/attrlist.c
+++ b/fs/ntfs3/attrlist.c
@@ -29,7 +29,17 @@ static inline bool al_is_valid_le(const struct ntfs_inode *ni,
void al_destroy(struct ntfs_inode *ni)
{
- ni = ni->base;
+ struct ntfs_inode *base = ni->base;
+
+ /* Stream inodes share the base's attribute list; freeing it here
+ * would race the base's ni_lock-held consumers and leave the base
+ * with a destroyed list while it is still alive. Only the owner
+ * frees.
+ */
+ if (base != ni)
+ return;
+
+ ni = base;
run_close(&ni->attr_list.run);
kvfree(ni->attr_list.le);
ni->attr_list.le = NULL;
--
2.43.7
next reply other threads:[~2026-10-27 15:16 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-10-27 15:16 Henry Martin [this message]
2026-10-27 15:16 ` [PATCH] fs/ntfs3: fix NULL deref in ni_std5() on stream inodes Henry Martin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261027151615.971171-1-bsdhenrymartin@gmail.com \
--to=bsdhenrymartin@gmail.com \
--cc=almaz.alexandrovich@paragon-software.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=ntfs3@lists.linux.dev \
--cc=stable@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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®