From: ZhangPeng <zhangpeng362@huawei.com>
Date: Wed, 7 Dec 2022 09:46:10 +0000
Subject: [PATCH] fs/ntfs3: Fix OOB read in indx_insert_into_buffer
References: bsc#1012628
Patch-mainline: 6.3.2
Git-commit: b8c44949044e5f7f864525fdffe8e95135ce9ce5
[ Upstream commit b8c44949044e5f7f864525fdffe8e95135ce9ce5 ]
Syzbot reported a OOB read bug:
BUG: KASAN: slab-out-of-bounds in indx_insert_into_buffer+0xaa3/0x13b0
fs/ntfs3/index.c:1755
Read of size 17168 at addr ffff8880255e06c0 by task syz-executor308/3630
Call Trace:
<TASK>
memmove+0x25/0x60 mm/kasan/shadow.c:54
indx_insert_into_buffer+0xaa3/0x13b0 fs/ntfs3/index.c:1755
indx_insert_entry+0x446/0x6b0 fs/ntfs3/index.c:1863
ntfs_create_inode+0x1d3f/0x35c0 fs/ntfs3/inode.c:1548
ntfs_create+0x3e/0x60 fs/ntfs3/namei.c:100
lookup_open fs/namei.c:3413 [inline]
If the member struct INDEX_BUFFER *index of struct indx_node is
incorrect, that is, the value of __le32 used is greater than the value
of __le32 total in struct INDEX_HDR. Therefore, OOB read occurs when
memmove is called in indx_insert_into_buffer().
Fix this by adding a check in hdr_find_e().
Fixes: 82cae269cfa9 ("fs/ntfs3: Add initialization of super block")
Reported-by: syzbot+d882d57193079e379309@syzkaller.appspotmail.com
Signed-off-by: ZhangPeng <zhangpeng362@huawei.com>
Signed-off-by: Konstantin Komarov <almaz.alexandrovich@paragon-software.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
---
fs/ntfs3/index.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/fs/ntfs3/index.c b/fs/ntfs3/index.c
index 51ab7595..ae9616be 100644
--- a/fs/ntfs3/index.c
+++ b/fs/ntfs3/index.c
@@ -725,9 +725,13 @@ static struct NTFS_DE *hdr_find_e(const struct ntfs_index *indx,
u32 e_size, e_key_len;
u32 end = le32_to_cpu(hdr->used);
u32 off = le32_to_cpu(hdr->de_off);
+ u32 total = le32_to_cpu(hdr->total);
u16 offs[128];
fill_table:
+ if (end > total)
+ return NULL;
+
if (off + sizeof(struct NTFS_DE) > end)
return NULL;
--
2.35.3