[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <59DE3A15.90800@arm.com>
Date: Wed, 11 Oct 2017 16:34:45 +0100
From: James Morse <james.morse@....com>
To: Mark Salyzyn <salyzyn@...roid.com>
CC: linux-kernel@...r.kernel.org, kevin.brodsky@....com,
catalin.marinas@....com, will.deacon@....com, mingo@...nel.org,
borntraeger@...ibm.com, peterz@...radead.org, Dave.Martin@....com,
credmonster@...il.com, zijun_hu@....com, mark.rutland@....com,
jszhang@...vell.com, labbott@...hat.com, john.stultz@...aro.org,
keescook@...omium.org, linux-arm-kernel@...ts.infradead.org,
takahiro.akashi@...aro.org, mmarek@...e.com,
ard.biesheuvel@...aro.org, Russell King <linux@...linux.org.uk>,
Mark Salyzyn <salyzyn@...gle.com>,
Andy Lutomirski <luto@...capital.net>,
Scott Wood <oss@...error.net>,
Dmitry Safonov <dsafonov@...tuozzo.com>
Subject: Re: [PATCH 1/3] arm: vdso: Rewrite gettimeofday to more closely match
arm64 variant
Hi Mark,
On 11/09/17 17:18, Mark Salyzyn wrote:
> Take an effort to recode the arm64 vdso code from assembler to C
> previously submitted by Andrew Pinski <apinski@...ium.com>, rework
> it for use in both arm and arm64, overlapping any optimizations
> for each architecture.
>
> Side effects include renaming some vdso_datapage.h variables, adding
> a few more to enhance features and functionality to match.
There are quite a lot of different things going on here, all of which are
interleaved in this patch. Could you split them out into smaller simpler changes?
The bits I count are:
* Pulling 32bit specific eabi and fallback macro out into compiler.h,
* Adding support for __vdso_clock_getres,
* Flipping tk_is_cntvct's logic to become use_syscall,
* Variable renames to align 32/64 bit,
...are those 'coarse' versions of the timer calls being added, or just moved around?
Thanks,
James
Powered by blists - more mailing lists