[PATCH] sched: return 1 if this cpu is proper for balancing inshould_we_balance()

[Joonsoo Kim]

Commit 23f0d20('sched: Factor out code to should_we_balance()') introduces
should_we_balance() function. This function should return 1 if this cpu is
appropriate for balancing. But current code doesn't do so. When this
happens, it returns 0, instead of 1.

This introduces performance regression which Dave Chinner reports.

                        v4 filesystem           v5 filesystem
3.11+xfsdev:            220k files/s            225k files/s
3.12-git                180k files/s            185k files/s
3.12-git-revert         245k files/s            247k files/s

You can find more detailed information in below link.
https://lkml.org/lkml/2013/9/10/1

This patch corrects the return value of should_we_balance() function
as orignally intended.

With this patch, Dave Chinner said that regression are gone.

                        v4 filesystem           v5 filesystem
3.11+xfsdev:            220k files/s            225k files/s
3.12-git                180k files/s            185k files/s
3.12-git-revert         245k files/s            247k files/s
3.12-git-fix            249k files/s            248k files/s

Reported-by: Dave Chinner <dchinner@xxxxxxxxxx>
Tested-by: Dave Chinner <dchinner@xxxxxxxxxx>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 7f0a5e6..9b3fe1c 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -5151,7 +5151,7 @@ static int should_we_balance(struct lb_env *env)
 	 * First idle cpu or the first cpu(busiest) in this sched group
 	 * is eligible for doing load balancing at this and above domains.
 	 */
-	return balance_cpu != env->dst_cpu;
+	return balance_cpu == env->dst_cpu;
 }
 
 /*
-- 
1.7.9.5

--
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/