linux-kernel - Re: [PATCH] stop_machine: avoid potential race behaviour

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:   Mon, 14 Oct 2019 11:09:45 +0100
From:   Mark Rutland <mark.rutland@....com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     linux-kernel@...r.kernel.org, Marco Elver <elver@...gle.com>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH] stop_machine: avoid potential race behaviour

On Tue, Oct 08, 2019 at 10:36:37AM +0200, Peter Zijlstra wrote:
> On Mon, Oct 07, 2019 at 11:45:36AM +0100, Mark Rutland wrote:
> > Both multi_cpu_stop() and set_state() access multi_stop_data::state
> > racily using plain accesses. These are subject to compiler
> > transformations which could break the intended behaviour of the code,
> > and this situation is detected by KCSAN on both arm64 and x86 (splats
> > below).
> 
> I really don't think there is anything the compiler can do wrong here.
> 
> That is, I'm thinking I'd like to get this called out as a false-positive.

I agree that in practice, it's very unlikely this would go wrong.

There are some things the compiler could do, e.g. with re-ordering of
volatile and plain reads of the same variable:

  https://lore.kernel.org/lkml/20191003161233.GB38140@lakrids.cambridge.arm.com/

... and while I agree that's vanishingly unlikely to happen here, I
couldn't say how to rule that out without ruling out cases that would
actually blow up in practice.

> That said, the patch looks obviously fine and will help with the
> validation effort so no real objection there.

Great! Can I take that as an Acked-by?

I assume this should go via the tip tree.

Thanks,
Mark.