Re: RFC: squid: Fix free memory corruption and off-by-on error when comparing SNMP OIDs

[Josh Bressers <bressers () redhat com>]

----- "Thomas Biege" <thomas () suse de> wrote:

> Hello people,
> does someone know if this bug has security implications. TIA!
>
> http://www.squid-cache.org/Versions/v3/3.1/changesets/SQUID_3_1_5.html
> http://www.squid-cache.org/Versions/v3/3.1/changesets/squid-3.1-10008.patch

This is really two flaws. The first bit of the patch is an off by one on
the loop that could overflow a heap buffer.

> From looking at the code, I only see this function being called with static
strings for the MIBs. I may be missing something, but it doesn't appear
that arbitrary strings make it into this. I'm not sure if this can be
exploted, or if it's just a bug someone noticed.

The second flaw is an invalid free. I'm not sure if arbitrary data can make
it into this, but with current glibc memory protections, this should be a
DoS only.

Both only seem to affect modern versions of squid. The code seems present
in 3.1.4, but not 2.6.STABLE21 (these are two versions we ship).

Have you mailed upstream at all?

Thanks.

-- 
    JB