linux-kernel - Re: WARNING: can't access registers at asm_common

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:   Wed, 11 Nov 2020 21:42:28 +0100
From:   Peter Zijlstra <peterz@...radead.org>
To:     Josh Poimboeuf <jpoimboe@...hat.com>
Cc:     Shinichiro Kawasaki <shinichiro.kawasaki@....com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Nicholas Piggin <npiggin@...il.com>,
        Damien Le Moal <Damien.LeMoal@....com>,
        andrew.cooper3@...rix.com, jgross@...e.com
Subject: Re: WARNING: can't access registers at asm_common_interrupt

On Wed, Nov 11, 2020 at 12:13:28PM -0600, Josh Poimboeuf wrote:
> On Wed, Nov 11, 2020 at 06:47:36PM +0100, Peter Zijlstra wrote:

> > Yeah, so it's all a giant can of worms that; also see:
> > 
> >   https://lkml.kernel.org/r/20200821084738.508092956@infradead.org
> > 
> > The basic idea is to only trace edges, ie. when the hardware state
> > actually changes. Sadly this means doing a pushf/pop before the cli.
> > Ideally CLI would store the old IF in CF or something like that, but
> > alas.
> 
> Right, that makes sense for save/restore, but is the disabled check
> really needed for local_irq_disable()?  Wouldn't that always be an edge?

IIRC there is code that does local_irq_disable() even though IRQs are
already disabled. This is 'harmless'.

> And anyway I don't see a similar check for local_irq_enable().

I know there is code that does local_irq_enable() with IRQs already
enabled, I'm not exactly sure why this is different. I'll have to put it
on the todo list :/