linux-kernel - Re: [PATCH V2] ocfs2: pass u64 to ocfs2_truncate

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <40f48906-09fd-4695-aef0-7647de46ea42@linux.alibaba.com>
Date: Wed, 16 Oct 2024 10:08:09 +0800
From: Joseph Qi <joseph.qi@...ux.alibaba.com>
To: Edward Adam Davis <eadavis@...com>
Cc: jlbec@...lplan.org, linux-kernel@...r.kernel.org, mark@...heh.com,
 ocfs2-devel@...ts.linux.dev,
 syzbot+81092778aac03460d6b7@...kaller.appspotmail.com,
 syzkaller-bugs@...glegroups.com, Su Yue <l@...enly.org>
Subject: Re: [PATCH V2] ocfs2: pass u64 to ocfs2_truncate_inline maybe
 overflow



On 10/11/24 10:01 AM, Joseph Qi wrote:
> 
> 
> On 10/11/24 9:07 AM, Su Yue wrote:
>>
>> On Thu 10 Oct 2024 at 22:31, Edward Adam Davis <eadavis@...com> wrote:
>>
>>> Syzbot reported a kernel BUG in ocfs2_truncate_inline.
>>> There are two reasons for this: first, the parameter value passed is greater
>>> than UINT_MAX, second, the start and end parameters of ocfs2_truncate_inline
>>> are "unsigned int".
>>>
>>> So, we need to add a sanity check for byte_start and byte_len right before
>>> ocfs2_truncate_inline() in ocfs2_remove_inode_range(), if they are greater
>>> than UINT_MAX return -EFBIG.
>>>
>>> Reported-by: syzbot+81092778aac03460d6b7@...kaller.appspotmail.com
>>> Closes: https://syzkaller.appspot.com/bug?extid=81092778aac03460d6b7
>>> Signed-off-by: Edward Adam Davis <eadavis@...com>
>>> ---
>>> V1 -> V2: move sanity check to ocfs2_remove_inode_range
>>>
>>>  fs/ocfs2/file.c | 5 +++++
>>>  1 file changed, 5 insertions(+)
>>>
>>> diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
>>> index ad131a2fc58e..05d6a8acfcda 100644
>>> --- a/fs/ocfs2/file.c
>>> +++ b/fs/ocfs2/file.c
>>> @@ -1784,6 +1784,11 @@ int ocfs2_remove_inode_range(struct inode *inode,
>>>          return 0;
>>>
>>>      if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
>>> +        if (byte_start > UINT_MAX || byte_start + byte_len > UINT_MAX) {
>>>
>> Why not use ocfs2_max_inline_data_with_xattr() here? Yes, UINT_MAX indeed
>> solves overflow problem Syzbot reported but you can find much lowerer
>> limit if once looked into inline data structures.
> 
> Right, since it is inline data, so the offset can't exceeds block size
> at least. You can refer bad inline data check in
> ocfs2_read_inline_data().
> 
Could you please update the check condition and send a new version?

Thanks,
Joseph