2010-01-25 04:42:12

by Frans Pop

Subject: Warnings about tasks on CPU 1 during STR

With 2.6.33-rc5 during suspend to RAM I'm getting a bunch of warnings
I've not seen with earlier kernel versions:

PM: late suspend of devices complete after 28.784 msecs
ACPI: Preparing to enter system sleep state S3
Disabling non-boot CPUs ...
CPU 1 is now offline
SMP alternatives: switching to UP code
Task events/1 (pid = 10) is on cpu 1 (state = 1, flags = 84208040)
Task khelper (pid = 11) is on cpu 1 (state = 1, flags = 80208040)
Task sync_supers (pid = 77) is on cpu 1 (state = 1, flags = 80208040)
Task bdi-default (pid = 79) is on cpu 1 (state = 2, flags = 80a15040)
Task kblockd/1 (pid = 81) is on cpu 1 (state = 1, flags = 84208040)
[...]
Task pm-suspend (pid = 29782) is on cpu 1 (state = 2, flags = 414000)
Extended CMOS year: 2000
Back to C!

All the "Task X is on cpu 1" messages (76 in total) are new. Bug or
intentional? As it is a warning, I suspect a bug.

Cheers,
FJP

2010-01-25 12:51:10

by Peter Zijlstra

Subject: Re: Warnings about tasks on CPU 1 during STR

On Mon, 2010-01-25 at 05:42 +0100, Frans Pop wrote:
> With 2.6.33-rc5 during suspend to RAM I'm getting a bunch of warnings
> I've not seen with earlier kernel versions:
>
> PM: late suspend of devices complete after 28.784 msecs
> ACPI: Preparing to enter system sleep state S3
> Disabling non-boot CPUs ...
> CPU 1 is now offline
> SMP alternatives: switching to UP code
> Task events/1 (pid = 10) is on cpu 1 (state = 1, flags = 84208040)
> Task khelper (pid = 11) is on cpu 1 (state = 1, flags = 80208040)
> Task sync_supers (pid = 77) is on cpu 1 (state = 1, flags = 80208040)
> Task bdi-default (pid = 79) is on cpu 1 (state = 2, flags = 80a15040)
> Task kblockd/1 (pid = 81) is on cpu 1 (state = 1, flags = 84208040)
> [...]
> Task pm-suspend (pid = 29782) is on cpu 1 (state = 2, flags = 414000)
> Extended CMOS year: 2000
> Back to C!
>
> All the "Task X is on cpu 1" messages (76 in total) are new. Bug or
> intentional? As it is a warning, I suspect a bug.

http://lkml.org/lkml/2010/1/21/175

2010-01-25 13:52:38

by Frans Pop

Subject: Re: Warnings about tasks on CPU 1 during STR

On Monday 25 January 2010, Peter Zijlstra wrote:
> On Mon, 2010-01-25 at 05:42 +0100, Frans Pop wrote:
> > All the "Task X is on cpu 1" messages (76 in total) are new. Bug or
> > intentional? As it is a warning, I suspect a bug.
>
> http://lkml.org/lkml/2010/1/21/175

Thanks Peter, that fixes it.

I've a small cleanup patch for the warning that I'll send separately.

Cheers,
FJP

2010-01-25 13:56:39

by Frans Pop

Subject: [PATCH] sched: correct whitespace in warning from cpu down task check

Due to an incorrect line break the output currently contains tabs.
Also remove trailing space.

Signed-off-by: Frans Pop <elendilplanet.nl>
Cc: Ingo Molnar <[email protected]>
Cc: Peter Zijlstra <[email protected]>
---

The actual output that logcheck sent me looked like this:
Task events/1 (pid = 10) is on cpu 1^I^I^I^I(state = 1, flags = 84208040)

After this patch it becomes:
Task events/1 (pid = 10) is on cpu 1 (state = 1, flags = 84208040)

diff --git a/kernel/cpu.c b/kernel/cpu.c
index 08e54e7..677f253 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -154,10 +154,10 @@ static inline void check_for_tasks(int cpu)
if (task_cpu(p) == cpu && p->state == TASK_RUNNING &&
(!cputime_eq(p->utime, cputime_zero) ||
!cputime_eq(p->stime, cputime_zero)))
- printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d\
- (state = %ld, flags = %x) \n",
- p->comm, task_pid_nr(p), cpu,
- p->state, p->flags);
+ printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d "
+ "(state = %ld, flags = %x)\n",
+ p->comm, task_pid_nr(p), cpu,
+ p->state, p->flags);
}
write_unlock_irq(&tasklist_lock);
}

2010-01-27 13:17:06

by Frans Pop

Subject: [tip:sched/urgent] sched: Correct printk whitespace in warning from cpu down task check

Commit-ID: 8048f77492bc428c6783e5039e759de2f960f47d
Gitweb: http://git.kernel.org/tip/8048f77492bc428c6783e5039e759de2f960f47d
Author: Frans Pop <[email protected]>
AuthorDate: Mon, 25 Jan 2010 14:56:34 +0100
Committer: Ingo Molnar <[email protected]>
CommitDate: Wed, 27 Jan 2010 08:34:40 +0100

sched: Correct printk whitespace in warning from cpu down task check

Due to an incorrect line break the output currently contains tabs.
Also remove trailing space.

The actual output that logcheck sent me looked like this:
Task events/1 (pid = 10) is on cpu 1^I^I^I^I(state = 1, flags = 84208040)

After this patch it becomes:
Task events/1 (pid = 10) is on cpu 1 (state = 1, flags = 84208040)

Signed-off-by: Frans Pop <elendilplanet.nl>
Signed-off-by: Peter Zijlstra <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
---
kernel/cpu.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/kernel/cpu.c b/kernel/cpu.c
index 08e54e7..677f253 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -154,10 +154,10 @@ static inline void check_for_tasks(int cpu)
if (task_cpu(p) == cpu && p->state == TASK_RUNNING &&
(!cputime_eq(p->utime, cputime_zero) ||
!cputime_eq(p->stime, cputime_zero)))
- printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d\
- (state = %ld, flags = %x) \n",
- p->comm, task_pid_nr(p), cpu,
- p->state, p->flags);
+ printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d "
+ "(state = %ld, flags = %x)\n",
+ p->comm, task_pid_nr(p), cpu,
+ p->state, p->flags);
}
write_unlock_irq(&tasklist_lock);
}

2010-01-28 07:13:26

by Frans Pop

Subject: [tip:sched/urgent] sched: Correct printk whitespace in warning from cpu down task check

Commit-ID: 9d3cfc4c1d17c6d3bc1373e3b954c56b92607755
Gitweb: http://git.kernel.org/tip/9d3cfc4c1d17c6d3bc1373e3b954c56b92607755
Author: Frans Pop <[email protected]>
AuthorDate: Mon, 25 Jan 2010 14:56:34 +0100
Committer: Ingo Molnar <[email protected]>
CommitDate: Thu, 28 Jan 2010 06:59:55 +0100

sched: Correct printk whitespace in warning from cpu down task check

Due to an incorrect line break the output currently contains tabs.
Also remove trailing space.

The actual output that logcheck sent me looked like this:
Task events/1 (pid = 10) is on cpu 1^I^I^I^I(state = 1, flags = 84208040)

After this patch it becomes:
Task events/1 (pid = 10) is on cpu 1 (state = 1, flags = 84208040)

Signed-off-by: Frans Pop <elendilplanet.nl>
Signed-off-by: Peter Zijlstra <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
---
kernel/cpu.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/kernel/cpu.c b/kernel/cpu.c
index 08e54e7..677f253 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -154,10 +154,10 @@ static inline void check_for_tasks(int cpu)
if (task_cpu(p) == cpu && p->state == TASK_RUNNING &&
(!cputime_eq(p->utime, cputime_zero) ||
!cputime_eq(p->stime, cputime_zero)))
- printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d\
- (state = %ld, flags = %x) \n",
- p->comm, task_pid_nr(p), cpu,
- p->state, p->flags);
+ printk(KERN_WARNING "Task %s (pid = %d) is on cpu %d "
+ "(state = %ld, flags = %x)\n",
+ p->comm, task_pid_nr(p), cpu,
+ p->state, p->flags);
}
write_unlock_irq(&tasklist_lock);
}