Re: [PATCH 1/4] ARM/div: Drop __div64_fls()

[Michal Orzel <michal.orzel@xxxxxxx>]

On 02/09/2024 12:03, Andrew Cooper wrote:
> 
> 
> Following the improvements to Xen's bitops, fls() does constant propagation in
> all cases.  Use it, and drop the local opencoded helper.
> 
> No functional change.
> 
> Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
Reviewed-by: Michal Orzel <michal.orzel@xxxxxxx>

> ---
> CC: Stefano Stabellini <sstabellini@xxxxxxxxxx>
> CC: Julien Grall <julien@xxxxxxx>
> CC: Volodymyr Babchuk <Volodymyr_Babchuk@xxxxxxxx>
> CC: Bertrand Marquis <bertrand.marquis@xxxxxxx>
> CC: Michal Orzel <michal.orzel@xxxxxxx>
> 
> ARM32 gets a very minor code generation improvement:
> 
>   xen.git/xen$ ../scripts/bloat-o-meter xen-syms-arm32-{before,after}
>   add/remove: 0/0 grow/shrink: 0/6 up/down: 0/-48 (-48)
>   Function                                     old     new   delta
>   wallclock_time                               288     280      -8
>   printk_start_of_line                         560     552      -8
>   domain_vtimer_init                           472     464      -8
>   do_settime                                   376     368      -8
>   burn_credits                                 760     752      -8
>   __printk_ratelimit                           424     416      -8
> 
> But it's just a couple of operations improvement and no real change in code
> structure.  I expect that the constant propagation being done through
> __builtin_clz(), rather than pure C, is giving the optimiser a bit more
> information to work with.
> 
> This file also has an __GNUC__ < 4 case which seems ripe for removing...
Agree and noted.

~Michal