sched: Pull up the might_sleep() check into cond_resched()
authorFrederic Weisbecker <fweisbec@gmail.com>
Thu, 16 Jul 2009 13:44:29 +0000 (15:44 +0200)
committerIngo Molnar <mingo@elte.hu>
Sat, 18 Jul 2009 13:51:44 +0000 (15:51 +0200)
commit613afbf83298efaead05ebcac23d2285609d7160
tree15fa2025d74cee5f6805ab974c532c6b6a603333
parent6f80bd985fe242c2e6a8b6209ed20b0495d3d63b
sched: Pull up the might_sleep() check into cond_resched()

might_sleep() is called late-ish in cond_resched(), after the
need_resched()/preempt enabled/system running tests are
checked.

It's better to check the sleeps while atomic earlier and not
depend on some environment datas that reduce the chances to
detect a problem.

Also define cond_resched_*() helpers as macros, so that the
FILE/LINE reported in the sleeping while atomic warning
displays the real origin and not sched.h

Changes in v2:

 - Call __might_sleep() directly instead of might_sleep() which
   may call cond_resched()

 - Turn cond_resched() into a macro so that the file:line
   couple reported refers to the caller of cond_resched() and
   not __cond_resched() itself.

Changes in v3:

 - Also propagate this __might_sleep() pull up to
   cond_resched_lock() and cond_resched_softirq()

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1247725694-6082-6-git-send-email-fweisbec@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
fs/dcache.c
include/linux/sched.h
kernel/sched.c