linux-kernel - Re: [PATCH 3.16 208/245] namei: allow restricted O

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

Date:   Fri, 24 Apr 2020 19:38:28 +0200
From:   Solar Designer <solar@...nwall.com>
To:     Ben Hutchings <ben@...adent.org.uk>
Cc:     linux-kernel@...r.kernel.org, stable@...r.kernel.org,
        akpm@...ux-foundation.org, Denis Kirjanov <kda@...ux-powerpc.org>,
        Dan Carpenter <dan.carpenter@...cle.com>,
        Al Viro <viro@...iv.linux.org.uk>,
        Kees Cook <keescook@...omium.org>,
        Salvatore Mesoraca <s.mesoraca16@...il.com>,
        Linus Torvalds <torvalds@...ux-foundation.org>
Subject: Re: [PATCH 3.16 208/245] namei: allow restricted O_CREAT of FIFOs and regular files

On Fri, Apr 24, 2020 at 04:13:22PM +0100, Ben Hutchings wrote:
> On Fri, 2020-04-24 at 15:52 +0200, Solar Designer wrote:
> > On Fri, Apr 24, 2020 at 12:07:15AM +0100, Ben Hutchings wrote:
> > > 3.16.83-rc1 review patch.  If anyone has any objections, please let me know.
> > 
> > I do.  This patch is currently known-buggy, see this thread:
> > 
> > https://www.openwall.com/lists/oss-security/2020/01/28/2
> > 
> > It is (partially) fixed with these newer commits in 5.5 and 5.5.2:
> > 
> > commit d0cb50185ae942b03c4327be322055d622dc79f6
> > Author: Al Viro <viro@...iv.linux.org.uk>
> > Date:   Sun Jan 26 09:29:34 2020 -0500
> > 
> >     do_last(): fetch directory ->i_mode and ->i_uid before it's too late
> >     
> >     may_create_in_sticky() call is done when we already have dropped the
> >     reference to dir.
> >     
> >     Fixes: 30aba6656f61e (namei: allow restricted O_CREAT of FIFOs and regular files)
> >     Signed-off-by: Al Viro <viro@...iv.linux.org.uk>
> > 
> > commit d76341d93dedbcf6ed5a08dfc8bce82d3e9a772b
> > Author: Al Viro <viro@...iv.linux.org.uk>
> > Date:   Sat Feb 1 16:26:45 2020 +0000
> > 
> >     vfs: fix do_last() regression
> >     
> >     commit 6404674acd596de41fd3ad5f267b4525494a891a upstream.
> [...]
> > At least inclusion of the above fixes is mandatory for any backports.
> 
> I know, and those are the next 2 patches in the series.

Ah, then no objections from me.

> > Also, I think no one has fixed the logic of may_create_in_sticky() so
> > that it wouldn't unintentionally apply the "protection" when the file
> > is neither a FIFO nor a regular file (something I found and mentioned in
> > the oss-security posting above).
> [...]
> > I think the implementation of may_create_in_sticky() should be rewritten
> > such that it'd directly correspond to the textual description in the
> > comment above.  As we've seen, trying to write the code "more optimally"
> > resulted in its logic actually being different from the description.
> > 
> > Meanwhile, I think backporting known-so-buggy code is a bad idea.
> 
> I can see that it's not quite right, but does it matter in practice? 
> Directories and symlinks are handled separately; sockets can't be
> opened anyway; block and character devices wonn't normally appear in a
> sticky directory.

Clearly, it doesn't matter all that much in practice - I'm not aware of
anyone having complained about it causing issues on their system.

I think it primarily mattered as an attack vector on the issue fixed
with Al's commits above.

I think we should nevertheless fix the code to match its intent and the
comment, but meanwhile this isn't a blocker for the backport.

Alexander