sysctl: Nuke sched_boost feature

This is perhaps the worst performance-related feature ever, and is thoroughly abused by userspace in all the worst ways possible. Triggering a massive task migration via sched_boost during latency-critical operations such as app launches is a very bad idea, and leads to noticeably-horrible stuttering. Signed-off-by: Sultan Alsawaf <[email protected]> Signed-off-by: Danny Lin <[email protected]> Signed-off-by: Tashfin Shakeer Rhythm <[email protected]>
mvaisakh · Apr 24, 2023 · f483f1e · f483f1e
1 parent 21c7cb7
commit f483f1e
Showing 1 changed file with 2 additions and 0 deletions.
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
@@ -429,6 +429,7 @@ static struct ctl_table kern_table[] = {
 		.extra1		= SYSCTL_ZERO,
 		.extra2		= &sysctl_sched_group_upmigrate_pct,
 	},
+#if 0
 	{
 		.procname	= "sched_boost",
 		.data		= &sysctl_sched_boost,
@@ -438,6 +439,7 @@ static struct ctl_table kern_table[] = {
 		.extra1		= &neg_three,
 		.extra2		= &three,
 	},
+#endif
 	{
 		.procname	= "sched_conservative_pl",
 		.data		= &sysctl_sched_conservative_pl,