linux-kernel - Re: [PATCH 6/7] cifs: Validate content of native symlink

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <20240930200026.efjs2v6ssqridcqx@pali>
Date: Mon, 30 Sep 2024 22:00:26 +0200
From: Pali Rohár <pali@...nel.org>
To: Steve French <smfrench@...il.com>
Cc: Steve French <sfrench@...ba.org>, Paulo Alcantara <pc@...guebit.com>,
	Ronnie Sahlberg <ronniesahlberg@...il.com>,
	linux-cifs@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 6/7] cifs: Validate content of native symlink

Now I tested it. FSCTL_SET_REPARSE_POINT ioctl call on Windows accepts
also symlink path with zero characters. In attachment I'm sending simple
program which creates Windows relative symlink and accepts binary target
path via \xFF sequence. You can compile it with gcc/mingw option -municode.
So calling "set_reparse_symlink.exe symlink file\x00file" creates new
symlink which points to target path "file<nul>file".

On Monday 30 September 2024 00:19:08 Pali Rohár wrote:
> I think that via pike it could be possible or via windows application
> running locally (to create reparse point manually with prepared buffer
> with such content). I will check it later.
> 
> Just a side note: Windows NT kernel allows for object names any
> characters except backslash. For object names is not used nul-term
> string, but rather string with explicit length. So even a null character
> is a valid in a object name. NT NTFS driver has for file names more
> restrictions and null is not valid. But it does not mean that somebody
> can write own filesystem which allows null bytes in file names...
> And this design of explicit lengths is also in SMB, so NT kernel may
> export nul characters in symlink path buffers...
> 
> On Sunday 29 September 2024 16:48:46 Steve French wrote:
> > Is there any easy way to create such a symlink (with null in it)?
> > 
> > On Sun, Sep 29, 2024 at 1:51 PM Pali Rohár <pali@...nel.org> wrote:
> > >
> > > Check that buffer does not contain UTF-16 null codepoint
> > > because Linux cannot process symlink with null byte.
> > >
> > > Signed-off-by: Pali Rohár <pali@...nel.org>
> > > ---
> > >  fs/smb/client/reparse.c | 10 ++++++++++
> > >  1 file changed, 10 insertions(+)
> > >
> > > diff --git a/fs/smb/client/reparse.c b/fs/smb/client/reparse.c
> > > index 5a738f65b190..ca4f96c43508 100644
> > > --- a/fs/smb/client/reparse.c
> > > +++ b/fs/smb/client/reparse.c
> > > @@ -509,6 +509,16 @@ int smb2_parse_native_symlink(char **target, const char *buf, unsigned int len,
> > >         int rc;
> > >         int i;
> > >
> > > +       /*
> > > +        * Check that buffer does not contain UTF-16 null codepoint
> > > +        * because Linux cannot process symlink with null byte.
> > > +        */
> > > +       if (unicode && UniStrnlen((wchar_t *)buf, len/2) != len/2) {
> > > +               cifs_dbg(VFS, "srv returned null byte in native symlink target location\n");
> > > +               rc = -EIO;
> > > +               goto out;
> > > +       }
> > > +
> > >         smb_target = cifs_strndup_from_utf16(buf, len, unicode, cifs_sb->local_nls);
> > >         if (!smb_target) {
> > >                 rc = -ENOMEM;
> > > --
> > > 2.20.1
> > >
> > >
> > 
> > 
> > -- 
> > Thanks,
> > 
> > Steve

View attachment "set_reparse_symlink.c" of type "text/x-csrc" (3200 bytes)