linux-kernel - Re: x86/csum: Remove unnecessary odd handling

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, 28 Jun 2023 13:34:24 -0500
From:   Noah Goldstein <goldstein.w.n@...il.com>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     Borislav Petkov <bp@...en8.de>, x86@...nel.org,
        edumazet@...gle.com, tglx@...utronix.de, mingo@...hat.com,
        dave.hansen@...ux.intel.com, hpa@...or.com,
        lkml <linux-kernel@...r.kernel.org>
Subject: Re: x86/csum: Remove unnecessary odd handling

On Wed, Jun 28, 2023 at 12:44 PM Linus Torvalds
<torvalds@...ux-foundation.org> wrote:
>
> On Wed, 28 Jun 2023 at 08:32, Noah Goldstein <goldstein.w.n@...il.com> wrote:
> >
> > Linus, if you're planning a patch and want to just integrate the codes
> > here I'm happy drop this patch
>
> No, that patch looks good to me.
>
> In fact, I wasn't planning on integrating my patch at all. I literally
> did it as a "I would have done it this way instead" exercise.
>
> And while I am currently running with my patch in the kernel, I don't
> even really know if it works and does the right thing. Maybe my use
> doesn't even trigger csum_partial() at all. I did not do any testing
> that "yes, I get the same checksum as a result".
>

There is a reasonably robust kunit for csum_partial: lib/checksum_kunit.c
so if you happened to run the kunit testsuite with your patch, it's
probably correct.

> So
>
>  (a) removing the pointless one-byte alignment looks good to me.
>
>  (b) I'd actually hope that somebody who _cares_ about this path and
> has put some real work into it (as opposed to my "superficial
> dabbling") would look at my patch and either go "yeah, not worth it",
> or "looks good, I'll take it".
>
> and I'm including that final patch of mine here again in case there
> was any confusion with the earlier versions (there were at least two
> known-broken versions I posted).
>
> *If* somebody likes it, and verifies that the checksum result is
> correct, feel free to do anything with that patch, including adding my
> signed-off-by for it (or taking the credit all for yourself -
> Mwahahahahaahaa!)
>
>               Linus