linux-kernel - Re: [PATCH V3fix ghak120] audit: initialize context values in case of mandatory events

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] [day] [month] [year] [list]

Message-ID: <CAHC9VhSf6eSjTvUA4FfUP+qBv_GDufBPfs=t3+BPPdFcTCD_4w@mail.gmail.com>
Date:   Wed, 29 Jul 2020 10:33:03 -0400
From:   Paul Moore <paul@...l-moore.com>
To:     Richard Guy Briggs <rgb@...hat.com>
Cc:     Eric Paris <eparis@...isplace.org>,
        Linux Security Module list 
        <linux-security-module@...r.kernel.org>,
        Linux-Audit Mailing List <linux-audit@...hat.com>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH V3fix ghak120] audit: initialize context values in case of
 mandatory events

On Tue, Jul 28, 2020 at 10:01 PM Richard Guy Briggs <rgb@...hat.com> wrote:
>
> On 2020-07-28 14:47, Paul Moore wrote:
> > On Tue, Jul 28, 2020 at 12:27 PM Richard Guy Briggs <rgb@...hat.com> wrote:
> > > I know you like only really minimal fixes this late, but this seemed
> > > pretty minimal to me...
> >
> > Minimal is a one (two?) line NULL check in audit_log_name(), this
> > patch is not that.
>
> I didn't try and test that since I'm not sure that would have worked
> because there appeared to be a low non-NULL value in it.  brauer1's trace had
> 0x60 and mine had 0xd0.  Or am I missing something obvious?

Well, you mentioned the obvious already: both 0x60 and 0xd0 are not
NULL.  We already have a NULL check for context->pwd elsewhere so
there is precedence for this solving a similar problem, although
without going through the git log I'm not sure what problem that
solved, or if it was precautionary.

I agree the low value looks suspicious, it almost looks like an offset
to me, ideally it would be good to understand how/why that value is
"off'.  It could be that the audit_context is not being properly
initialized, reset, or something unrelated is clobbering the value;
all things that would be nice to know.

> The patch provided the information rather than ignoring the problem ...

I disagree.

--
paul moore
www.paul-moore.com