Skip to contentSkip to conversation

sched/numa: kill the wrong/dead TASK_DEAD check in task_numa_fault()

PATCH In review Oleg Nesterov 4 participants 4 messages Sep 21, 2014 – Sep 24, 2014 Scheduler

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 sched/numa: kill the wrong/dead TASK_DEAD check in task_numa_fault()

current->state == TASK_DEAD means that the task is doing its last schedule(), page fault is obviously impossible at this stage.

---
 kernel/sched/fair.c |    4 ----
 1 files changed, 0 insertions(+), 4 deletions(-)
kernel/sched/fair.c +0 -4
index 82088b2..0090e8c 100644
@@ -1804,10 +1804,6 @@ void task_numa_fault(int last_cpupid, int mem_node, int pages, int flags)
 	if (!p->mm)
 		return;
 
-	/* Do not worry about placement if exiting */
-	if (p->state == TASK_DEAD)
-		return;
-
 	/* Allocate buffer to track faults on a per-node basis */
 	if (unlikely(!p->numa_faults_memory)) {
 		int size = sizeof(*p->numa_faults_memory) *
Signature · 1 line
1.5.5.1
reviewer Acked-by on v1 1/1
#1 lore raw
Quoted 4 lines · Oleg Nesterov

current->state == TASK_DEAD means that the task is doing its last schedule(), page fault is obviously impossible at this stage.

I was going to complain that the check would protect against a spurious allocation during fault but a task in this state should not be faulting any more.

Thanks.

Signature · 2 lines
Mel Gorman
SUSE Labs
Acked-by on v1 1/1
#2 lore raw
In reply to Oleg Nesterov

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 4 lines · Oleg Nesterov

current->state == TASK_DEAD means that the task is doing its last schedule(), page fault is obviously impossible at this stage.

- --
All rights reversed
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iQEcBAEBAgAGBQJUIDVyAAoJEM553pKExN6D50oH/A5RpFUhx5zsDYgFjyq071+n NYgwgy6Nr6DDkmfEGtRePxQY7CmysKCTWdPzAFxmvhaFfOg2zP+P7uh85/lbIYXM 7ATj4yO8u3Loa8zdz+VjwrOR0lJr/Y/nsjQC0q+50oNM0pwFC+HPuSPTrpz4o+HB ReeZ6BKNtFoiC5RfR0XpNHk4ozwT5qW2oCDEIT6bsMQiV0413CXksHFg6Flz7Acb 2EFCzwY2BgfKRawGR6e4Q09UZ0aG1gW9RkLkUuIWlmbOHiMzduLCIKQNCZaXTgVn xVJcsgdANnQk1BGSU5rozqtlDUmRoKq3teXlE00CcJEEfzUrUHMw0HYZ62xbQ+s= =fg3H
-----END PGP SIGNATURE-----

#3 lore raw

[tip:sched/core] sched/numa: Kill the wrong/ dead TASK_DEAD check in task_numa_fault()

In reply to Oleg Nesterov
sched/numa: Kill the wrong/dead TASK_DEAD check in task_numa_fault()
Commit-ID:  be34f0f3e6aed6e828a8059247d169d38da128d7
Gitweb:     http://git.kernel.org/tip/be34f0f3e6aed6e828a8059247d169d38da128d7
Author:     Oleg Nesterov <[email protected]>
AuthorDate: Sun, 21 Sep 2014 21:47:43 +0200
Committer:  Ingo Molnar <[email protected]>
CommitDate: Wed, 24 Sep 2014 09:35:05 +0200

sched/numa: Kill the wrong/dead TASK_DEAD check in task_numa_fault()

current->state == TASK_DEAD means that the task is doing its last schedule(), page fault is obviously impossible at this stage.

---
 kernel/sched/fair.c | 4 ----
 1 file changed, 4 deletions(-)
kernel/sched/fair.c +0 -4
index 2a1e6ac..9ee3d4f 100644
@@ -1817,10 +1817,6 @@ void task_numa_fault(int last_cpupid, int mem_node, int pages, int flags)
 	if (!p->mm)
 		return;
 
-	/* Do not worry about placement if exiting */
-	if (p->state == TASK_DEAD)
-		return;
-
 	/* Allocate buffer to track faults on a per-node basis */
 	if (unlikely(!p->numa_faults_memory)) {
 		int size = sizeof(*p->numa_faults_memory) *