sched: fix broken SCHED_RESET_ON_FORK handling

PATCH In review Mike Galbraith 1 participant 3 messages Jul 27, 2011 – Aug 14, 2011

lore

Share & tools

lore

Mike Galbraith

author

2011-07-27 15:15 UTC #0 lore raw

PATCH 1/1 sched: fix broken SCHED_RESET_ON_FORK handling

sched: fix broken SCHED_RESET_ON_FORK handling

Setting child->prio = current->normal_prio _after_ SCHED_RESET_ON_FORK has been handled for an RT parent gives birth to a deranged mutant child with non-RT policy, but RT prio and sched_class.

Move PI leakage protection up, always set priorities and weight, and if the child is leaving RT class, reset rt_priority to the proper value.

Signed-off-by Mike Galbraith <[email protected]>

Cc <[email protected]>

kernel/sched.c +12 ~~-13~~

index 0281f84..a3e65a0 100644
@@ -2804,19 +2804,23 @@ void sched_fork(struct task_struct *p)
 	p->state = TASK_RUNNING;
 
 	/*
+	 * Make sure we do not leak PI boosting priority to the child.
+	 */
+	p->prio = current->normal_prio;
+
+	/*
 	 * Revert to default priority/policy on fork if requested.
 	 */
 	if (unlikely(p->sched_reset_on_fork)) {
-		if (p->policy == SCHED_FIFO || p->policy == SCHED_RR) {
+		if (task_has_rt_policy(p)) {
 			p->policy = SCHED_NORMAL;
-			p->normal_prio = p->static_prio;
-		}
-
-		if (PRIO_TO_NICE(p->static_prio) < 0) {
 			p->static_prio = NICE_TO_PRIO(0);
-			p->normal_prio = p->static_prio;
-			set_load_weight(p);
-		}
+			p->rt_priority = 0;
+		} else if (PRIO_TO_NICE(p->static_prio) < 0)
+			p->static_prio = NICE_TO_PRIO(0);
+
+		p->prio = p->normal_prio = __normal_prio(p);
+		set_load_weight(p);
 
 		/*
 		 * We don't need the reset flag anymore after the fork. It has

@@ -2825,11 +2829,6 @@ void sched_fork(struct task_struct *p)
 		p->sched_reset_on_fork = 0;
 	}
 
-	/*
-	 * Make sure we do not leak PI boosting priority to the child.
-	 */
-	p->prio = current->normal_prio;
-
 	if (!rt_prio(p->prio))
 		p->sched_class = &fair_sched_class;

Mike Galbraith

author

2011-07-27 16:10 UTC #1 lore raw

I knew that thing had been moved before.. see f83f9ac2.

Quoted 62 lines · Mike Galbraith

sched: fix broken SCHED_RESET_ON_FORK handling

Move PI leakage protection up, always set priorities and weight, and if the child is leaving RT class, reset rt_priority to the proper value.

Signed-off-by Mike Galbraith <[email protected]>

Cc <[email protected]>

kernel/sched.c +12 ~~-13~~

index 0281f84..a3e65a0 100644
@@ -2804,19 +2804,23 @@ void sched_fork(struct task_struct *p)
 	p->state = TASK_RUNNING;
 
 	/*
+	 * Make sure we do not leak PI boosting priority to the child.
+	 */
+	p->prio = current->normal_prio;
+
+	/*
 	 * Revert to default priority/policy on fork if requested.
 	 */
 	if (unlikely(p->sched_reset_on_fork)) {
-		if (p->policy == SCHED_FIFO || p->policy == SCHED_RR) {
+		if (task_has_rt_policy(p)) {
 			p->policy = SCHED_NORMAL;
-			p->normal_prio = p->static_prio;
-		}
-
-		if (PRIO_TO_NICE(p->static_prio) < 0) {
 			p->static_prio = NICE_TO_PRIO(0);
-			p->normal_prio = p->static_prio;
-			set_load_weight(p);
-		}
+			p->rt_priority = 0;
+		} else if (PRIO_TO_NICE(p->static_prio) < 0)
+			p->static_prio = NICE_TO_PRIO(0);
+
+		p->prio = p->normal_prio = __normal_prio(p);
+		set_load_weight(p);
 
 		/*
 		 * We don't need the reset flag anymore after the fork. It has

@@ -2825,11 +2829,6 @@ void sched_fork(struct task_struct *p)
 		p->sched_reset_on_fork = 0;
 	}
 
-	/*
-	 * Make sure we do not leak PI boosting priority to the child.
-	 */
-	p->prio = current->normal_prio;
-
 	if (!rt_prio(p->prio))
 		p->sched_class = &fair_sched_class;

tip-bot for Mike Galbraith

author

2011-08-14 15:59 UTC #2 lore raw

In reply to Mike Galbraith

Commit-ID:  c350a04efd1c89cd256b2abc8f07a21d0d53ff24
Gitweb:     http://git.kernel.org/tip/c350a04efd1c89cd256b2abc8f07a21d0d53ff24
Author:     Mike Galbraith <[email protected]>
AuthorDate: Wed, 27 Jul 2011 17:14:55 +0200
Committer:  Ingo Molnar <[email protected]>
CommitDate: Sun, 14 Aug 2011 12:00:43 +0200

sched: fix broken SCHED_RESET_ON_FORK handling

Move PI leakage protection up, always set priorities and weight, and if the child is leaving RT class, reset rt_priority to the proper value.

Signed-off-by Mike Galbraith <[email protected]>

Signed-off-by Peter Zijlstra <[email protected]>

Link http://lkml.kernel.org/r/[email protected] open here

Signed-off-by Ingo Molnar <[email protected]>

---
 kernel/sched.c |   25 ++++++++++++-------------
 1 files changed, 12 insertions(+), 13 deletions(-)

kernel/sched.c +12 ~~-13~~

index b1e8f0e..cf427bb 100644
@@ -2844,19 +2844,23 @@ void sched_fork(struct task_struct *p)
 	p->state = TASK_RUNNING;
 
 	/*
+	 * Make sure we do not leak PI boosting priority to the child.
+	 */
+	p->prio = current->normal_prio;
+
+	/*
 	 * Revert to default priority/policy on fork if requested.
 	 */
 	if (unlikely(p->sched_reset_on_fork)) {
-		if (p->policy == SCHED_FIFO || p->policy == SCHED_RR) {
+		if (task_has_rt_policy(p)) {
 			p->policy = SCHED_NORMAL;
-			p->normal_prio = p->static_prio;
-		}
-
-		if (PRIO_TO_NICE(p->static_prio) < 0) {
 			p->static_prio = NICE_TO_PRIO(0);
-			p->normal_prio = p->static_prio;
-			set_load_weight(p);
-		}
+			p->rt_priority = 0;
+		} else if (PRIO_TO_NICE(p->static_prio) < 0)
+			p->static_prio = NICE_TO_PRIO(0);
+
+		p->prio = p->normal_prio = __normal_prio(p);
+		set_load_weight(p);
 
 		/*
 		 * We don't need the reset flag anymore after the fork. It has

@@ -2865,11 +2869,6 @@ void sched_fork(struct task_struct *p)
 		p->sched_reset_on_fork = 0;
 	}
 
-	/*
-	 * Make sure we do not leak PI boosting priority to the child.
-	 */
-	p->prio = current->normal_prio;
-
 	if (!rt_prio(p->prio))
 		p->sched_class = &fair_sched_class;