linux-kernel - Re: [PATCH v3 1/1] x86/split_lock: check split lock support on initialization

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]

Message-ID: <87zhbojoii.fsf@nanos.tec.linutronix.de>
Date:   Mon, 06 Apr 2020 23:24:53 +0200
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Benjamin Lamowski <benjamin.lamowski@...nkonzept.com>,
        xiaoyao.li@...el.com
Cc:     bp@...en8.de, fenghua.yu@...el.com, hpa@...or.com,
        linux-kernel@...r.kernel.org, luto@...nel.org, mingo@...hat.com,
        nivedita@...m.mit.edu, pbonzini@...hat.com, peterz@...radead.org,
        philipp.eppelt@...nkonzept.com, sean.j.christopherson@...el.com,
        tony.luck@...el.com, x86@...nel.org,
        Benjamin Lamowski <benjamin.lamowski@...nkonzept.com>
Subject: Re: [PATCH v3 1/1] x86/split_lock: check split lock support on initialization

Benjamin Lamowski <benjamin.lamowski@...nkonzept.com> writes:

> While the sld setup code is run only if the TEST_CTRL MSR is available,
> the current sld initialization code unconditionally resets it even on
> systems where this architectural MSR is not available.
>
> This commit introduces a new default sld state sld_unsupported, which is
> changed in split_lock_setup() only if sld is available; and checks for
> split lock detect support before initializing it.

What for? You explain what the patch is doing, but not WHY. See
Documentation/process and while at it please search for 'This patch' as
well.

> Fixes: dbaba47085b0c ("x86/split_lock: Rework the initialization flow
> of split lock detection")

That's blantantly wrong. It fixes your hypervisor which fails to raise a
#GP when an unknown MSR is accessed. 

Are we going to see that kind of 'fixes' everytime we decide to probe an
MSR for simplicity from now on?

Please fix the root cause and not the symptom.

Thanks,

        tglx