Skip to content

Commit

Permalink
rcu: Allow RCU grace-period cleanup to be preempted
Browse files Browse the repository at this point in the history
RCU grace-period cleanup is currently carried out with interrupts
disabled, which can result in excessive latency spikes on large systems
(many hundreds or thousands of CPUs).  This patch therefore makes the
RCU grace-period cleanup be preemptible, including voluntary preemption
points, which should eliminate those latency spikes.  Similar spikes from
forcing of quiescent states will be dealt with similarly by later patches.

Updated to replace uses of spin_lock_irqsave() with spin_lock_irq(), as
suggested by Peter Zijlstra.

Reported-by: Mike Galbraith <mgalbraith@suse.de>
Reported-by: Dimitri Sivanich <sivanich@sgi.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
  • Loading branch information
Paul E. McKenney committed Sep 23, 2012
1 parent cabc49c commit c856baf
Showing 1 changed file with 5 additions and 6 deletions.
11 changes: 5 additions & 6 deletions kernel/rcutree.c
Original file line number Diff line number Diff line change
Expand Up @@ -1151,7 +1151,7 @@ static int __noreturn rcu_gp_kthread(void *arg)
* completed.
*/
if (*rdp->nxttail[RCU_WAIT_TAIL] == NULL) {
raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */
raw_spin_unlock_irq(&rnp->lock);

/*
* Propagate new ->completed value to rcu_node
Expand All @@ -1160,14 +1160,13 @@ static int __noreturn rcu_gp_kthread(void *arg)
* to process their callbacks.
*/
rcu_for_each_node_breadth_first(rsp, rnp) {
/* irqs already disabled. */
raw_spin_lock(&rnp->lock);
raw_spin_lock_irq(&rnp->lock);
rnp->completed = rsp->gpnum;
/* irqs remain disabled. */
raw_spin_unlock(&rnp->lock);
raw_spin_unlock_irq(&rnp->lock);
cond_resched();
}
rnp = rcu_get_root(rsp);
raw_spin_lock(&rnp->lock); /* irqs already disabled. */
raw_spin_lock_irq(&rnp->lock);
}

rsp->completed = rsp->gpnum; /* Declare grace period done. */
Expand Down

0 comments on commit c856baf

Please sign in to comment.