linux-kernel - Re: [PATCH] crypto: xts_crypt() return if walk.nbytes is 0

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:   Sun, 22 Aug 2021 15:23:52 +0200
From:   Ard Biesheuvel <ardb@...nel.org>
To:     Herbert Xu <herbert@...dor.apana.org.au>
Cc:     Shreyansh Chouhan <chouhan.shreyansh630@...il.com>,
        "David S. Miller" <davem@...emloft.net>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        X86 ML <x86@...nel.org>, "H. Peter Anvin" <hpa@...or.com>,
        Linux Crypto Mailing List <linux-crypto@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        syzbot+20191dc583eff8602d2d@...kaller.appspotmail.com
Subject: Re: [PATCH] crypto: xts_crypt() return if walk.nbytes is 0

On Fri, 20 Aug 2021 at 14:53, Herbert Xu <herbert@...dor.apana.org.au> wrote:
>
> On Fri, Aug 20, 2021 at 01:14:52PM +0200, Ard Biesheuvel wrote:
> >
> > In spite of that, I have a slight preference for this version, given
> > that it makes it obvious that we bail on two separate conditions:
> > - an error has occurred
> > - no error has occurred but the resulting walk is empty
> >
> > Testing walk.nbytes only needlessly obfuscates the code, as we need to
> > return 'err' in the end anyway.
>
> I disagree, this is how most skcipher walkers are structured, they
> never explicitly test on err and only terminate the loop when
> walk->nbytes hits zero, in which case err is returned as is.
>
> I don't see why this particular skcipher walker should deviate
> from that paradigm.  In fact it is exactly that deviation that
> caused the bug in the first instance.
>

Fair enough.