linux-kernel - Re: [PATCH 0/2] Allow assembly code to use BIT(), GENMASK(), etc. and clean-up arm64 header

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: <CAK7LNAQJPMsRtNRYUH+dib0ZMAPqOe5HO0UcAW7zRdjyWWyQWQ@mail.gmail.com>
Date:   Wed, 5 Jun 2019 18:01:10 +0900
From:   Masahiro Yamada <yamada.masahiro@...ionext.com>
To:     Catalin Marinas <catalin.marinas@....com>
Cc:     linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
        Will Deacon <will.deacon@....com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 0/2] Allow assembly code to use BIT(), GENMASK(), etc. and
 clean-up arm64 header

On Wed, Jun 5, 2019 at 4:36 PM Catalin Marinas <catalin.marinas@....com> wrote:
>
> On Mon, May 27, 2019 at 05:34:10PM +0900, Masahiro Yamada wrote:
> > Some in-kernel headers use _BITUL() instead of BIT().
> >
> >  arch/arm64/include/asm/sysreg.h
> >  arch/s390/include/asm/*.h
> >
> > I think the reason is because BIT() is currently not available
> > in assembly. It hard-codes 1UL, which is not available in assembly.
> [...]
> > Masahiro Yamada (2):
> >   linux/bits.h: make BIT(), GENMASK(), and friends available in assembly
> >   arm64: replace _BITUL() with BIT()
> >
> >  arch/arm64/include/asm/sysreg.h | 82 ++++++++++++++++-----------------
> >  include/linux/bits.h            | 17 ++++---
>
> I'm not sure it's worth the hassle. It's nice to have the same BIT macro
> but a quick grep shows arc, arm64, s390 and x86 using _BITUL. Maybe a
> tree-wide clean-up would be more appropriate.


I am happy to clean-up the others
in the next development cycle
once 1/2 lands in the mainline.


Since there is no subsystem that
takes care of include/linux/bits.h,
I just asked Will to pick up both.
I planed per-arch patch submission
to reduce the possibility of merge conflict.


If you guys are not willing to pick up them,
is it better to send treewide conversion to Andrew?


-- 
Best Regards
Masahiro Yamada