linux-kernel - Re: fix int_sqrt() for very large numbers

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: <CAHk-=wjwt4M7wGcMDkVLmSCsFP3xj29pAVgsC6-zbc_XPN1yiw@mail.gmail.com>
Date:   Sun, 20 Jan 2019 17:11:14 +1200
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Florian La Roche <florian.laroche@...glemail.com>
Cc:     Will Deacon <will.deacon@....com>,
        Linux List Kernel Mailing <linux-kernel@...r.kernel.org>,
        Crt Mori <cmo@...exis.com>, Joe Perches <joe@...ches.com>,
        Davidlohr Bueso <dave@...olabs.net>,
        Peter Zijlstra <peterz@...radead.org>
Subject: Re: fix int_sqrt() for very large numbers

On Sun, Jan 20, 2019 at 5:03 PM Florian La Roche
<florian.laroche@...glemail.com> wrote:
>
> The real bug is that we compute 1 to 64 for bit 0 to bit 63, whereas
> the algorithm expects 0 to 63 for the value of m.

Florian, you seem to be in denial.

__fls() returns 0-63. Your patch is *wrong* for the __fls() use,
because when you subtract 1, you get -1 to 62, and the -1  now
introduces the very undefined behavior you claim your patch fixes.

So your patch fixes one real case (int_sqrt64(), that has one user
that doesn't care) but it *BREAKS* the other case that is actually
much more widely used (int_sqrt()).

See what Will and I are complainig about?

            Linus