linux-kernel - Re: [PATCH] drm/etnaviv: only reject timeouts with tv

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Tue, 28 Jan 2020 14:07:51 +0100
From:   Arnd Bergmann <arnd@...db.de>
To:     Guido Günther <agx@...xcpu.org>
Cc:     Russell King - ARM Linux admin <linux@...linux.org.uk>,
        Lucas Stach <l.stach@...gutronix.de>,
        Christian Gmeiner <christian.gmeiner@...il.com>,
        David Airlie <airlied@...ux.ie>,
        Daniel Vetter <daniel@...ll.ch>,
        Philipp Zabel <p.zabel@...gutronix.de>,
        Sam Ravnborg <sam@...nborg.org>, Rob Herring <robh@...nel.org>,
        Emil Velikov <emil.velikov@...labora.com>,
        The etnaviv authors <etnaviv@...ts.freedesktop.org>,
        dri-devel <dri-devel@...ts.freedesktop.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] drm/etnaviv: only reject timeouts with tv_nsec >= 2 seconds

On Fri, Jan 24, 2020 at 9:56 AM Guido Günther <agx@...xcpu.org> wrote:
> On Wed, Jan 22, 2020 at 10:35:53AM +0000, Russell King - ARM Linux admin wrote:
> > On Wed, Jan 22, 2020 at 11:30:34AM +0100, Guido Günther wrote:

> > I think it would probably be better for the kernel to print a
> > warning once when noticing over-large nsec values, suggesting a
> > userspace upgrade is in order, but continue the existing behaviour.
>
> That makes sense to me. This also makes sure we don't break other (non
> mesa using) stuff accidentally. We have
>
>   https://gitlab.freedesktop.org/mesa/mesa/commit/d817f2c69615cf37b78f484a25b7831ebe9dbe6f
>
> and
>
>   https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3534
>
> to normalize nsec to [0..999999999] now.
>

I have reverted my patch that adds the range check now, so I can send the rest
of the series for inclusion.

I've played around with ways to add a ratelimited warning message and to
make sure that a malicious application cannot cause a long stall, but haven't
managed to get a version I'm actually happy with.

I'll follow up once the series is merged, and then we can add a better
workaround
later through the drm tree.

      Arnd