linux-kernel - Re: [PATCH v2 2/3] lib/find_bit: create find_first_zero_bit

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: <YwYpqRZxCe9NLZyk@yury-laptop>
Date:   Wed, 24 Aug 2022 06:37:45 -0700
From:   Yury Norov <yury.norov@...il.com>
To:     Andy Shevchenko <andy.shevchenko@...il.com>
Cc:     Linus Torvalds <torvalds@...ux-foundation.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Guenter Roeck <linux@...ck-us.net>,
        Dennis Zhou <dennis@...nel.org>,
        Russell King <linux@...linux.org.uk>,
        Catalin Marinas <catalin.marinas@....com>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Rasmus Villemoes <linux@...musvillemoes.dk>,
        Alexey Klimov <aklimov@...hat.com>,
        Kees Cook <keescook@...omium.org>,
        Andy Whitcroft <apw@...onical.com>
Subject: Re: [PATCH v2 2/3] lib/find_bit: create find_first_zero_bit_le()

On Wed, Aug 24, 2022 at 12:22:33PM +0300, Andy Shevchenko wrote:
> On Wed, Aug 24, 2022 at 5:17 AM Yury Norov <yury.norov@...il.com> wrote:
> >
> > find_first_zero_bit_le() is an alias to find_next_zero_bit_le(),
> > despite that 'next' is known to be slower than the 'first' version.
> >
> > Now that we have a common FIND_FIRST_BIT() macro helper, it's trivial
> > to implement find_first_zero_bit_le() as a real function.
> >
> > Moving find_*_le() to a separate file helps to fit the FIND_FIRST_BIT()
> > to the _le needs by wiring word_op to swab.
> >
> > Signed-off-by: Yury Norov <yury.norov@...il.com>
> > ---
> > Like other find_*_le() functions, the new one takes void *addr, instead
> > of unsigned long *. This should be fixed for all in a separate series.
> 
> From this comment it is unclear to me why we can't fix them first and
> then apply this with the correct type?

Because there is a codebase that relies on existing types, mostly in
filesystem code. And those fs fixes would require 5 or 6 patches.

This would triple the length of this series, and is completely
unrelated. That's why I think that:
        > > This should be fixed for all in a separate series.

> ...
> 
> > +#define word_op swab
> > +#include "find_bit.h"
> 
> Looking at this, I would rather always require to define __ffs_word_op
> (or whatever name) in the user and replace #ifndef in the find_bit.h
> with
> #error "The __ffs_word_op must be defined before including find_bit.h!"

This is a local header which is not intended to be included anywhere
except lib/find_bit{,_be}.c. I don't expect someone else would want to
include it, even in lib. So what you suggest is a bit overthinking to
me. But if you insist, I can do that.

Thanks,
Yury