workqueues: make get_online_cpus() useable for work->func()
authorOleg Nesterov <oleg@tv-sign.ru>
Fri, 25 Jul 2008 08:47:50 +0000 (01:47 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 25 Jul 2008 17:53:40 +0000 (10:53 -0700)
commit3da1c84c00c7e5fa8348336bd8c342f9128b0f14
treeaf960c65a670319fa86846a493a6d97f9fb37cd9
parent8616a89ab761239c963eea3a63be383f127cc7e8
workqueues: make get_online_cpus() useable for work->func()

workqueue_cpu_callback(CPU_DEAD) flushes cwq->thread under
cpu_maps_update_begin().  This means that the multithreaded workqueues
can't use get_online_cpus() due to the possible deadlock, very bad and
very old problem.

Introduce the new state, CPU_POST_DEAD, which is called after
cpu_hotplug_done() but before cpu_maps_update_done().

Change workqueue_cpu_callback() to use CPU_POST_DEAD instead of CPU_DEAD.
This means that create/destroy functions can't rely on get_online_cpus()
any longer and should take cpu_add_remove_lock instead.

[akpm@linux-foundation.org: fix CONFIG_SMP=n]
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Acked-by: Gautham R Shenoy <ego@in.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Max Krasnyansky <maxk@qualcomm.com>
Cc: Paul Jackson <pj@sgi.com>
Cc: Paul Menage <menage@google.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Vegard Nossum <vegard.nossum@gmail.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
include/linux/cpu.h
include/linux/notifier.h
kernel/cpu.c
kernel/workqueue.c