Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 27 Aug 2023 13:39:41 -0400
From: Demi Marie Obenour <>
Subject: Re: linux-distros list policy and Linux kernel, again

On Sun, Aug 27, 2023 at 09:41:22AM +0200, Eduardo' Vela" <Nava> wrote:
> Hey!
> I'm currently on holidays so sorry for my briefness. I couldn't miss a
> chance to comment on this.
> Our team at Google is working on generating CVEs for Syzkaller findings.
> This is not trivial.

Do you plan to generate CVEs only for fixed issues, or also for unfixed
ones?  The latter might be better, as it would reflect the actual
security of the Linux kernel.

> On Sat, 26 Aug 2023, 23:49 Solar Designer, <> wrote:
> > > If every syzkaller
> > > issue received a CVE automatically, we'd immediately remove the most
> > > noisome posts.
> >
> > Is every syzkaller issue a vulnerability?
> >
> No, they are not. Most (all?) are bugs, so they probably should get fixed,
> but I don't think we can claim them all to be vulnerabilities. Even if we
> did, we probably should help NVD figure out severity for the CVSS or they
> will just have to guess randomly. Figuring out a criteria for what is worth
> a CVE and what is not, as well as deduplicating is probably the main bulk
> of the work here.

The issues also need to actually be fixed.  That will require that
somebody (Google, Red Hat, Oracle, or someone else) hires additional
people to do the work.

> > - Ask Red Hat's CNA to consider setting up an automatic CVE assignment
> > >   process for syzkaller issues. (Red Hat's CNA is now serving as a Root
> > >   CNA for FOSS issues in general, so it feels like a plausible place to
> > >   put this process. Google runs syzkaller and has four CNAs, perhaps
> > >   one of them would be a better fit. Maybe the Linux Foundation could
> > >   run a CNA for this purpose. I'm not picky.)
> >
> > This is an interesting suggestion.  I think we'd first need to determine
> > whether this can be automated at all without ending up with CVEs
> > assigned in cases where they shouldn't have been per MITRE's guidelines
> > (e.g., when no security boundary is crossed in proper documented usage).
> >
> So right now we have been experimenting with this and want to start with a
> basic heuristic to generate OSV identifiers. If it goes well with OSV we
> may start generating CVEs.
> We analyzed crashes and concluded the only ones we are confident on
> generating CVEs automatically are KASAN crashes that aren't null-ptr-deref
> but we will revise this criteria after we have a first version.
> Anyway, as you can imagine, we know generating CVEs automatically can have
> a significant disrupting effect on the industry as a lot of the regulation
> and process depend on it, so we want to minimize the hatemail we'll get.

Much of the problem comes from generating a huge number of reports and
not helping to fix them.  This causes upstream maintainers to
de-prioritize syzbot reports or even outright burn out (as in the case
of Darrick Wong).

> Anyway, for the curious on our progress
>   - has some details
>   -
>     has the output of our current heuristics

Does this include unfixed vulnerabilities?
Demi Marie Obenour (she/her/hers)
Invisible Things Lab

Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)

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.