[PATCH] sched_fair: test list head instead of list entry in throttle_cfs_rq()

[PATCH] sched_fair: test list head instead of list entry in throttle_cfs_rq()

[Cong Wang]

According to the comments, we need to test if this is
the first throttled task, however, list_empty() tests on
the entry cfs_rq->throttled_list, not the head, this is wrong.

This is a bug because we don't re-init the list entry after
removing it from the list, so list_empty() could return false
even if the list is really empty.

Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Cong Wang <cwang@twopensource.com>
---
 kernel/sched/fair.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 40a7fcb..ee91917 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3683,7 +3683,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
 	cfs_rq->throttled = 1;
 	cfs_rq->throttled_clock = rq_clock(rq);
 	raw_spin_lock(&cfs_b->lock);
-	empty = list_empty(&cfs_rq->throttled_list);
+	empty = list_empty(&cfs_b->throttled_cfs_rq);
 
 	/*
 	 * Add to the _head_ of the list, so that an already-started
-- 
1.8.3.1

Re: [PATCH] sched_fair: test list head instead of list entry in throttle_cfs_rq()

[bsegall]

Cong Wang <xiyou.wangcong@gmail.com> writes:

> According to the comments, we need to test if this is
> the first throttled task, however, list_empty() tests on
> the entry cfs_rq->throttled_list, not the head, this is wrong.
>
> This is a bug because we don't re-init the list entry after
> removing it from the list, so list_empty() could return false
> even if the list is really empty.
>
> Cc: Ingo Molnar <mingo@redhat.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
> Signed-off-by: Cong Wang <cwang@twopensource.com>

Agreed.
Reviewed-by: Ben Segall <bsegall@google.com>

> ---
>  kernel/sched/fair.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 40a7fcb..ee91917 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -3683,7 +3683,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
>  	cfs_rq->throttled = 1;
>  	cfs_rq->throttled_clock = rq_clock(rq);
>  	raw_spin_lock(&cfs_b->lock);
> -	empty = list_empty(&cfs_rq->throttled_list);
> +	empty = list_empty(&cfs_b->throttled_cfs_rq);
>  
>  	/*
>  	 * Add to the _head_ of the list, so that an already-started

Re: [PATCH] sched_fair: test list head instead of list entry in throttle_cfs_rq()

[Cong Wang]

On Wed, Jun 24, 2015 at 12:41 PM, Cong Wang <xiyou.wangcong@gmail.com> wrote:
> According to the comments, we need to test if this is
> the first throttled task, however, list_empty() tests on
> the entry cfs_rq->throttled_list, not the head, this is wrong.
>
> This is a bug because we don't re-init the list entry after
> removing it from the list, so list_empty() could return false
> even if the list is really empty.
>
> Cc: Ingo Molnar <mingo@redhat.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
> Signed-off-by: Cong Wang <cwang@twopensource.com>

ping.

What's the status of this patch?

Re: [PATCH] sched_fair: test list head instead of list entry in throttle_cfs_rq()

[Peter Zijlstra]

On Fri, Jul 03, 2015 at 09:09:12AM -0700, Cong Wang wrote:
> On Wed, Jun 24, 2015 at 12:41 PM, Cong Wang <xiyou.wangcong@gmail.com> wrote:
> > According to the comments, we need to test if this is
> > the first throttled task, however, list_empty() tests on
> > the entry cfs_rq->throttled_list, not the head, this is wrong.
> >
> > This is a bug because we don't re-init the list entry after
> > removing it from the list, so list_empty() could return false
> > even if the list is really empty.
> >
> > Cc: Ingo Molnar <mingo@redhat.com>
> > Cc: Peter Zijlstra <peterz@infradead.org>
> > Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
> > Signed-off-by: Cong Wang <cwang@twopensource.com>
> 
> ping.
> 
> What's the status of this patch?

Queued, will likely hit tip after the merge window etc has settled.

Re: [PATCH] sched_fair: test list head instead of list entry in throttle_cfs_rq()

[Cong Wang]

On Fri, Jul 3, 2015 at 9:10 AM, Peter Zijlstra <peterz@infradead.org> wrote:
>
> Queued, will likely hit tip after the merge window etc has settled.

Great.

Thanks!

[tip:sched/core] sched/fair: Test list head instead of list entry in throttle_cfs_rq()

[tip-bot for Cong Wang]

Commit-ID:  d49db342f0e276b354383b3281c5668b6b80f5c2
Gitweb:     http://git.kernel.org/tip/d49db342f0e276b354383b3281c5668b6b80f5c2
Author:     Cong Wang <xiyou.wangcong@gmail.com>
AuthorDate: Wed, 24 Jun 2015 12:41:47 -0700
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Mon, 6 Jul 2015 14:15:08 +0200

sched/fair: Test list head instead of list entry in throttle_cfs_rq()

According to the comments, we need to test if this is
the first throttled task, however, list_empty() tests on
the entry cfs_rq->throttled_list, not the head, this is wrong.

This is a bug because we don't re-init the list entry after
removing it from the list, so list_empty() could return false
even if the list is really empty.

Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Cong Wang <cwang@twopensource.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ben Segall <bsegall@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1435174907-432-1-git-send-email-xiyou.wangcong@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/sched/fair.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 65c8f3e..d113c3b 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3683,7 +3683,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
 	cfs_rq->throttled = 1;
 	cfs_rq->throttled_clock = rq_clock(rq);
 	raw_spin_lock(&cfs_b->lock);
-	empty = list_empty(&cfs_rq->throttled_list);
+	empty = list_empty(&cfs_b->throttled_cfs_rq);
 
 	/*
 	 * Add to the _head_ of the list, so that an already-started