linux-kernel - Re: [PATCH 05/32] locking/lockdep: Prepare valid

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]

Message-ID: <20190213151648.GD8524@lenoir>
Date:   Wed, 13 Feb 2019 16:16:49 +0100
From:   Frederic Weisbecker <frederic@...nel.org>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     LKML <linux-kernel@...r.kernel.org>,
        Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
        Peter Zijlstra <peterz@...radead.org>,
        Mauro Carvalho Chehab <mchehab@...pensource.com>,
        "David S . Miller" <davem@...emloft.net>,
        Thomas Gleixner <tglx@...utronix.de>,
        "Paul E . McKenney" <paulmck@...ux.vnet.ibm.com>,
        Frederic Weisbecker <fweisbec@...il.com>,
        Pavan Kondeti <pkondeti@...eaurora.org>,
        Ingo Molnar <mingo@...nel.org>,
        Joel Fernandes <joel@...lfernandes.org>
Subject: Re: [PATCH 05/32] locking/lockdep: Prepare valid_state() to handle
 plain masks

On Tue, Feb 12, 2019 at 09:45:52AM -0800, Linus Torvalds wrote:
> On Tue, Feb 12, 2019 at 9:14 AM Frederic Weisbecker <frederic@...nel.org> wrote:
> >
> > +
> > +       while (vectors) {
> > +               long fs = __ffs64(vectors) + 1;
> > +
> > +               vectors >>= fs;
> 
> This is wrong.
> 
> If "vectors" only has the high hit set, you end up with "fs" having
> the value "64".
> 
> And then "vectors >>= fs" is undefined and won't actually do anything
> at all on x86.

Oh! ok didn't know that...

> 
> In general, avoid "ffs()", and the stupid pattern of "__ffs(x)+1".
> 
> Bit numbering starts at 0. "ffs()" is wrong. And you never *ever* just
> add one to a bit number in order to shift by one more bit, exactly
> because of overflow issues.
> 
> So it may look inefficient, but the correct thing to do is
> 
>     long bit = __ffs64(vectors);
>     vectors >>= fs;
>     vectors >>= 1;
> 
> because that actually works.

I see, perhaps I should use for_each_set_bit() that should take care about those
details?

Thanks.