lkml.org 
[lkml]   [2015]   [May]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH 5/6] x86: replace cpu_**_mask() with topology_**_cpumask()
From
2015-05-02 8:33 GMT+02:00 Ingo Molnar <mingo@kernel.org>:
>
> * Bartosz Golaszewski <bgolaszewski@baylibre.com> wrote:
>
>> The former duplicate the functionalities of the latter but are neither
>> documented nor arch-independent.
>
>> if (!has_mp) {
>> - cpumask_set_cpu(cpu, cpu_sibling_mask(cpu));
>> + cpumask_set_cpu(cpu, topology_thread_cpumask(cpu));
>> cpumask_set_cpu(cpu, cpu_llc_shared_mask(cpu));
>
> So why does topology.h invent a new name for 'sibling CPUs'?
>
> At least in the scheduling context, 'sibling' is the term we are using
> in most places in the scheduler - try 'git grep sibling kernel/sched/'.
>
> 'thread' is a bad name anyway for a CPU, even if we didn't have an
> existing term for it.

Actually those macros used to be called topology_core_siblings() and
topology_thread_siblings() until commit fbd59a8d1.

> So please rename topology_thread_cpumask to topology_sibling_cpumask
> to not replace one inconsistency for another one ...

I'll prepare a new series.

Best regards,
Bartosz Golaszewski


\
 
 \ /
  Last update: 2015-05-04 10:41    [W:0.075 / U:0.148 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site