Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 24 Jan 2015 14:28:51 -0500 (EST)
From: cve-assign@...re.org
To: oss@...ernot.info
Cc: cve-assign@...re.org, oss-security@...ts.openwall.com
Subject: Re: CVE Request: PHP

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

> Date: Thu, 08 Jan 2015 22:11:09 +1100
> I'm requesting multiple CVE-ID's for multiple vulnerabilities in PHP
> that I found:


> Use after free in 'opcache' component of PHP
> https://bugs.php.net/bug.php?id=68677
> http://git.php.net/?p=php-src.git;a=commit;h=777c39f4042327eac4b63c7ee87dc1c7a09a3115

Use CVE-2015-1351.


(requests 2 and 3 are skipped because of the
http://openwall.com/lists/oss-security/2015/01/08/5 post)


> Null Pointer Dereference in pgsql
> https://bugs.php.net/bug.php?id=68741
> http://git.php.net/?p=php-src.git;a=commit;h=124fb22a13fafa3648e4e15b4f207c7096d8155e

Use CVE-2015-1352 for this issue in which a return value isn't
validated.


> Null Pointer Dereference in ereg(regex)
> https://bugs.php.net/bug.php?id=68740
> http://git.php.net/?p=php-src.git;a=commit;h=124fb22a13fafa3648e4e15b4f207c7096d8155e

Because of an unusual process step on MITRE's end, there was also some
communication about these bugs that was only between MITRE and Joshua
Rogers, without a Cc to oss-security. For Bug #68740, the additional
discussion sent to us was (more or less) was that code in between
lines 140 and 167 wouldn't change g->setbits to a non-NULL value. This
is also essentially implied by the reasoning used in the Description
section of Bug #68740. (We didn't want to send the private e-mail
here, but Joshua Rogers is free to send it if he wants.)

MITRE doesn't have a full code analysis and isn't confident about
whether the "explicit null dereference" exists or not. All we can
offer is that the "wouldn't change g->setbits to a non-NULL value"
seems somewhat implausible because it means that significant intended
functionality of the code wouldn't have worked at all.

As an example, this sequence of function calls seems possible:

  p_str - ordinary - bothcases - p_bracket - allocset

where allocset contains:

  p->g->setbits = (uch *)malloc(nbytes);
  
and a memset (and the code before the line-140 "g->setbits = NULL"
includes a "p->g = g").

We're going to defer a CVE assignment for Bug #68740 until someone
outside MITRE offers additional analysis. It might be worthwhile to
update Bug #68740 so that the "explicit null dereference" term isn't
used. Although maybe a code path with a NULL pointer dereference can
be found, it's apparently not the case that g->setbits is explicitly
guaranteed to be NULL on line 1279.

- -- 
CVE assignment team, MITRE CVE Numbering Authority
M/S M300
202 Burlington Road, Bedford, MA 01730 USA
[ PGP key available through http://cve.mitre.org/cve/request_id.html ]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.14 (SunOS)

iQEcBAEBAgAGBQJUw/F5AAoJEKllVAevmvmsQi0H/A+DkH6FsNqUv6mmXBj7UCbL
rQdAjfZGcMDA43oQBWBbKPqetAae63eBLyxzZOTOPqlRS5vr1U6Ly4s4equlvzsm
govktU8CC7mdg6t5ZRYVh4CQHPsf4VnEf/bAK0ExlDPyl0zSQMXewZ5BJjh9VCXs
Ap6CeWqaN5rS38IDxDOH5MTpqrOAdWP/U5YtSZdUdBvcXR7bla5Aal2aAPXA92kp
HrIU0JXOz5FHCOKeoMvri+RxkrSJe+/8WUQfCI/o4PuUcoq+WHm4YZXHQ3mnck9k
W5SMGA/a+xrTPHXWyqLYo0tY+7VIHQDIpBPTnhw2Hw7+d9vSt5jU9BW4kXNOKrs=
=715+
-----END PGP SIGNATURE-----

Powered by blists - more mailing lists

Your e-mail address:

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.