Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 15 Jun 2019 15:22:22 -0400
From: Alex Gaynor <>
Subject: Re: Thousands of vulnerabilities, almost no CVEs: OSS-Fuzz

You are completely right that this is not a novel phenomenon, though I
think the scale at which OSS-Fuzz has found vulnerabilities has genuinely
exacerbated this problem.

While it's true, some of the bugs found will not be exploitable, I think we
should not be overly dismissive.
an example of such a script-less exploit.
an example of exploiting use of uninitialized value (one of the most common
bug classes in OSS-Fuzz, probably because so few people test with MSAN).

I think you're quite right that the central challenge here is the mismatch
between how Linux distributions operate and what their claims/people's
expectations are.


PS: I'd be remiss if I didn't at least mention that basically all the bug
classes we're discussing are induced by C/C++'s memory unsafety and better
programming language prevent them outright.

On Sat, Jun 15, 2019 at 3:01 PM Hanno Böck <> wrote:

> Hi Alex,
> I think what you're describing has been going on for a while, even
> before oss-fuzz.
> A combination of compiler sanitizers and better fuzzing techniques has
> scaled up bug finding and fixing to a level we haven't had before.
> For distributions that promise to backport all security fixes that
> creates a situation where it's almost impossible to keep that promise,
> they just don't have the manpower to scale up at the same speed as
> people find bugs.
> Maybe the main takeaway here is to just recognize that, and maybe
> distros should be more honest here and be clear what they can and can't
> do. And if you run a parser in a high risk environment you may not want
> to rely on the outdated version shipping in some LTS distribution.
> But I also think it's good to keep some perspective of the bugs we're
> talking about.
> Many of the bugs oss-fuzz finds are of bug classes where it's quite
> unlikely that they directly lead to a security issue (e.g. out of
> bounds memory reads - which asan controversially calls "overflows").
> Even for the scarier looking vulns like write buffer overflows and use
> after free the situation is that these are usually not straightforward
> to exploit. All modern distributions have a combination of stack
> canaries, ASLR and nonexecutable memory. It's my understanding that
> while it's often possible to bypass those, doing so in non-scripting
> scenarios (e.g. in an image parser) is really hard and often impossible.
> I guess therefore it's still an overall win. While there's a number of
> bugs unfixed with public information, in the long term we'll get more
> robust code and the number of bugs present should be in steep decline.
> --
> Hanno Böck
> mail/jabber:
> GPG: FE73757FA60E4E21B937579FA5880072BBB51E42

All that is necessary for evil to succeed is for good people to do nothing.

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.