[patch] sched, vfs: fix scheduling latencies in prune_dcache() and select_parent()

[Ingo Molnar]

the attached patch fixes long scheduling latencies in select_parent()
and prune_dcache(). The prune_dcache() lock-break is easy, but for
select_parent() the only viable solution i found was to break out if
there's a resched necessary - the reordering is not necessary and the
dcache scanning/shrinking will later on do it anyway.

This patch has been in the -VP patchset for weeks.

	Ingo

the attached patch fixes long scheduling latencies in select_parent()
and prune_dcache(). The prune_dcache() lock-break is easy, but for
select_parent() the only viable solution i found was to break out if
there's a resched necessary - the reordering is not necessary and the
dcache scanning/shrinking will later on do it anyway.

This patch has been in the -VP patchset for weeks.

Signed-off-by: Ingo Molnar <mingo@xxxxxxx>

--- linux/fs/dcache.c.orig	
+++ linux/fs/dcache.c	
@@ -381,6 +381,8 @@ static void prune_dcache(int count)
 		struct dentry *dentry;
 		struct list_head *tmp;
 
+		cond_resched_lock(&dcache_lock);
+
 		tmp = dentry_unused.prev;
 		if (tmp == &dentry_unused)
 			break;
@@ -553,6 +555,14 @@ resume:
 		struct dentry *dentry = list_entry(tmp, struct dentry, d_child);
 		next = tmp->next;
 
+		/*
+		 * select_parent() is a performance optimization, it is
+		 * not necessary to complete it. Abort if a reschedule is
+		 * pending:
+		 */
+		if (need_resched())
+			goto out;
+
 		if (!list_empty(&dentry->d_lru)) {
 			dentry_stat.nr_unused--;
 			list_del_init(&dentry->d_lru);
@@ -590,6 +600,7 @@ this_parent->d_parent->d_name.name, this
 #endif
 		goto resume;
 	}
+out:
 	spin_unlock(&dcache_lock);
 	return found;
 }