[PATCH 08/10 V4] workqueue: explicit way to wait for idles workers to finish

From: Lai Jiangshan
Date: Sat Sep 01 2012 - 13:45:51 EST


busy_worker_rebind_fn() can't return until all idle workers are rebound.
This order is ensured by rebind_workers() currently.

We use synchronize_all_idles_rebound() to wait for all idle workers
to be rebound. this is an explicit way and it will ease the pain of
the rebind_workers().

The sleeping synchronize_all_idles_rebound() must be called before
the WORKER_REBIND has been cleared.

It adds a small overhead to the unlikely path.

Signed-off-by: Lai Jiangshan <laijs@xxxxxxxxxxxxxx>
---
kernel/workqueue.c | 16 +++++++++++++++-
1 files changed, 15 insertions(+), 1 deletions(-)

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index c875951..16bcd84 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -1387,9 +1387,23 @@ static void busy_worker_rebind_fn(struct work_struct *work)
{
struct worker *worker = container_of(work, struct worker, rebind_work);
struct global_cwq *gcwq = worker->pool->gcwq;
+ struct idle_rebind *idle_rebind;
+
+ if (worker_maybe_bind_and_lock(worker)) {
+ /* Is idle-rebinding still in progress? */
+ if ((idle_rebind = gcwq->idle_rebind) != NULL) {
+ /* get reference */
+ BUG_ON(idle_rebind->ref_cnt <= 0);
+ idle_rebind->ref_cnt++;
+ spin_unlock_irq(&gcwq->lock);
+
+ synchronize_all_idles_rebound(gcwq, idle_rebind);
+
+ spin_lock_irq(&gcwq->lock);
+ }

- if (worker_maybe_bind_and_lock(worker))
worker_clr_flags(worker, WORKER_REBIND);
+ }

spin_unlock_irq(&gcwq->lock);
}
--
1.7.4.4

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/