linux-ext4 - Re: [PATCH v5 01/12] ovl: Reject mounting over case-insensitive directories

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]

Date: Tue, 30 Jan 2024 21:31:40 -0300
From: Gabriel Krisman Bertazi <krisman@...e.de>
To: Eric Biggers <ebiggers@...nel.org>
Cc: viro@...iv.linux.org.uk,  jaegeuk@...nel.org,  tytso@....edu,
  amir73il@...il.com,  linux-ext4@...r.kernel.org,
  linux-f2fs-devel@...ts.sourceforge.net,  linux-fsdevel@...r.kernel.org
Subject: Re: [PATCH v5 01/12] ovl: Reject mounting over case-insensitive
 directories

Eric Biggers <ebiggers@...nel.org> writes:

> On Mon, Jan 29, 2024 at 05:43:19PM -0300, Gabriel Krisman Bertazi wrote:
>> ovl: Reject mounting over case-insensitive directories
>
> Maybe:
>
>     ovl: Reject mounting over rootdir of case-insensitive capable FS
>
> or:
>
>     ovl: Always reject mounting over case-insensitive directories
>
> ... since as your commit message explains, overlayfs already does reject
> mounting over case-insensitive directories, just not in all cases.
>
>> Since commit bb9cd9106b22 ("fscrypt: Have filesystems handle their
>> d_ops"), we set ->d_op through a hook in ->d_lookup, which
>> means the root dentry won't have them, causing the mount to accidentally
>> succeed.
>
> But this series changes that.  Doesn't that make this overlayfs fix redundant?
> It does improve the error message, which is helpful, but your commit message
> makes it sound like it's an actual fix, not just an error message improvement.

Yes, this series will make it redundant.  But Christian Brauner had
suggested we make this verification explicit instead of relying on d_ops
being set, to avoid future changes breaking this again.  I found the
issue in ovl during testing of v2 and intended to send it separately for
-rc7, but Amir asked it to be sent as part of this series.  And also the
error code is improved.

Anyway, I'll can make this clear in the commit message

-- 
Gabriel Krisman Bertazi