itimers: remove the per-cpu-ish-ness
authorPeter Zijlstra <peterz@infradead.org>
Mon, 24 Nov 2008 16:06:57 +0000 (17:06 +0100)
committerIngo Molnar <mingo@elte.hu>
Wed, 7 Jan 2009 17:52:44 +0000 (18:52 +0100)
commit490dea45d00f01847ebebd007685d564aaf2cd98
treea1f559fd497b10c21479b378ffb262d517cb627b
parentede6f5aea054d3fb67c78857f7abdee602302043
itimers: remove the per-cpu-ish-ness

Either we bounce once cacheline per cpu per tick, yielding n^2 bounces
or we just bounce a single..

Also, using per-cpu allocations for the thread-groups complicates the
per-cpu allocator in that its currently aimed to be a fixed sized
allocator and the only possible extention to that would be vmap based,
which is seriously constrained on 32 bit archs.

So making the per-cpu memory requirement depend on the number of
processes is an issue.

Lastly, it didn't deal with cpu-hotplug, although admittedly that might
be fixable.

Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
include/linux/init_task.h
include/linux/sched.h
kernel/fork.c
kernel/posix-cpu-timers.c
kernel/sched_stats.h