sched: avoid stack var in move_task_off_dead_cpu
authorRusty Russell <rusty@rustcorp.com.au>
Mon, 24 Nov 2008 16:05:11 +0000 (02:35 +1030)
committerIngo Molnar <mingo@elte.hu>
Mon, 24 Nov 2008 16:51:52 +0000 (17:51 +0100)
commite76bd8d9850c2296a7e8e24c9dce9b5e6b55fe2f
treec4f5178e224ece271317c60da569489b5a8704ee
parentf17c860760927c2a8e41a021eab3317e4415e962
sched: avoid stack var in move_task_off_dead_cpu

Impact: stack usage reduction

With some care, we can avoid needing a temporary cpumask (we can't
really allocate here, since we can't fail).

This version calls cpuset_cpus_allowed_locked() with the task_rq_lock
held.  I'm fairly sure this works, but there might be a deadlock
hiding.

And of course, we can't get rid of the last cpumask on stack until we
can use cpumask_of_node instead of node_to_cpumask.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
kernel/sched.c