linux-kernel - Re: [PATCH] x86/split_lock: Restore warn mode (and add a new one) to avoid userspace regression

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

Date:   Thu, 29 Sep 2022 09:26:38 -0700
From:   Dave Hansen <dave.hansen@...el.com>
To:     "Guilherme G. Piccoli" <gpiccoli@...lia.com>, tony.luck@...el.com,
        tglx@...utronix.de, linux-kernel@...r.kernel.org, x86@...nel.org
Cc:     mingo@...hat.com, bp@...en8.de, dave.hansen@...ux.intel.com,
        hpa@...or.com, luto@...nel.org, kernel-dev@...lia.com,
        kernel@...ccoli.net, Fenghua Yu <fenghua.yu@...el.com>,
        Joshua Ashton <joshua@...ggi.es>,
        Paul Gofman <pgofman@...eweavers.com>,
        Pavel Machek <pavel@...x.de>,
        Pierre-Loup Griffais <pgriffais@...vesoftware.com>,
        Melissa Wen <mwen@...lia.com>
Subject: Re: [PATCH] x86/split_lock: Restore warn mode (and add a new one) to
 avoid userspace regression

On 9/29/22 08:30, Guilherme G. Piccoli wrote:
>> How about we give it a few weeks and see if the current behavior impacts
>> anyone else?  Maybe the best route will be more clear then.
> ...I disagree in just letting it fly for weeks with all players of God
> of War 2 running modern Intel chips unable to play in 5.19+ because of
> this change.

Let's be precise here, though.  It isn't that folks can't play.  It's
that we *intentionally* put something in place that kept them from
playing.  They can play just fine after disabling split lock detection.

> Certainly we have more games/applications that are impacted, I just 
> don't think we should wait on having 3 userspace breakages reported,
> for example, to take an action - why should gamers live with this for
> an arbitrary amount of time, until others report more issues?
They don't have to live with it.  They can turn it off.  That's why the
command-line disable is there.

The real question in my head is whether the misery is intentional or
not.  Is breaking games what folks _intended_ with
split_lock_detect=warn?  Or, is this a more severe penalty than we
expected and maybe we should back off for the default?