[lkml]   [2015]   [Apr]   [15]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH 1/3] sched, timer: Remove usages of ACCESS_ONCE in the scheduler
On Tue, 2015-04-14 at 22:40 -0400, Steven Rostedt wrote:
> You are correct. Now I'm thinking that the WRITE_ONCE() is not needed,
> and just a:
> p->mm->numa_scan_seq = READ_ONCE(p->numa_scan_seq) + 1;

Just to confirm, is this a typo? Because there really is a numa_scan_seq
in the task_struct itself too :)

p->mm->numa_scan_seq is read in task_numa_placement() with
ACCESS_ONCE(), and so the benefit that I do see with it is that it makes
it consistent by doing the updates with ACCESS_ONCE too (for
documentation purposes).

If that's really the case:

WRITE_ONCE(p->mm->numa_scan_seq, p->mm->numa_scan_seq + 1)

should be enough for that.

 \ /
  Last update: 2015-04-16 04:41    [W:0.214 / U:2.060 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site