Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 5 Apr 2011 09:40:13 -0600
From: Vincent Danen <vdanen@...hat.com>
To: oss-security@...ts.openwall.com
Subject: Re: Closed list

* [2011-04-05 08:43:29 -0400] Josh Bressers wrote:

>
>----- Original Message -----
>> On Tue, Apr 05, 2011 at 07:19:08AM -0400, Josh Bressers wrote:
>> > Not adding Apple to any coordination list would be plain silly. They
>> > were far more active than most of the distributions.
>>
>> Yes. But why do they need to be aware, say, of glibc vulnerabilities
>> (ones that are in fact believed to be glibc-specific)?
>
>This is an excellent point. It's a hard problem to solve honestly. I guess
>the question really comes down to this. Do the disadvantages of one list
>outweigh the benefits? I'm not sure what the answer is. There probably
>isn't an "answer" though, just lots of opinions.

Just throwing this out there (I've read the thread but haven't
contributed at all yet).

A lot of userland stuff is shared between BSD and Linux, and probably
some other operating systems.  About the only things that differ between
a lot of these are the Linux kernel, and the *libc.  There is a lot of
cross-over with other stuff, which means there will likely be a lot of
cc'ing going on (which I imagine might be complicated due to encryption
requirements).

Where does the line get drawn?  If vendor A ships with exim, and another
with postfix, which one belongs on the "Linux list"?  Obviously
discussions of exim don't matter to the postfix shipper, and vice versa.
Pick any other software that has a competing open source alternative.
Does Debian not get on the list because they don't technically ship
firefox?

I think if the disqualifier to Apple is that they don't ship a Linux
kernel and glibc, then we're doing them (and ourselves) a disservice.
Apple contributed a lot to vendor-sec (and I'm not going all pro-Apple
here, just stating a fact).

I think it would be reasonable to s/Linux list/open source vendor list/,
like vendor-sec used to be.  Yes, Apple will see some glibc stuff and
some Linux kernel stuff.  So what?  They may also see some exim stuff
that doesn't apply to them, and (hopefully!) the Linux vendors may seem
some stuff that isn't applicable to them, but it is useful for the *BSD
vendors who would (hopefully!) be on the list and feel welcome enough to
use it.  Who knows, it might even be beneficial to have a glibc issue
and someone from Apple or FreeBSD or whatever pipes up and indicates
that the BSD libc once had a similar problem and tells us how they fixed
it.

I think getting hung up on "Linux vendors only" and "BSD vendors can
have their own list" and we end up cross-posting 90% of the issues is
going to be an exercise in frustration.

Either that, or we start to work more closely with a *CERT and deal with
their process for passing along information to other vendors for
userland things that are shared; no offence to oCERT or anyone else, but
that seems like more of a headache than just letting
Apple/FreeBSD/OpenBSD/etc. have a seat at our table.

Just my $0.02.

-- 
Vincent Danen / Red Hat Security Response Team 

Powered by blists - more mailing lists

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.