linux-kernel - Re: Potentially Broken Address Dependency via test

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20211027123429.GE174730@worktop.programming.kicks-ass.net>
Date:   Wed, 27 Oct 2021 14:34:29 +0200
From:   Peter Zijlstra <peterz@...radead.org>
To:     Paul Heidekrüger <paul.heidekrueger@...tum.de>
Cc:     paulmck@...nel.org, will@...nel.org, boqun.feng@...il.com,
        stern@...land.harvard.edu, parri.andrea@...il.com,
        linux-kernel@...r.kernel.org, llvm@...ts.linux.dev,
        elver@...gle.com, charalampos.mainas@...il.com,
        pramod.bhatotia@...tum.de
Subject: Re: Potentially Broken Address Dependency via test_bit() When
 Compiling With Clang

On Wed, Oct 27, 2021 at 02:17:48PM +0200, Peter Zijlstra wrote:
> On Wed, Oct 27, 2021 at 12:19:48PM +0200, Paul Heidekrüger wrote:
> I would personally not consider this a dependend load. The result
> depends on two loads, but there is no actual ordering between them.
> 
>   r1 = *x
>   r2 = *y
>   b = 1 & (r1 >> r2);
> 
> (more or less)

melver pointed out on IRC that I missed the whole BIT_WORD(nr) thing.
And with that restored this should indeed be an address dependency.

Still, I wasn't actually expecting test_bit() to be one. Nice find.