2.6.28-rc6: hpet: object is on stack, but not annotated

2.6.28-rc6: hpet: object is on stack, but not annotated

[Alexander Beregalov]

Hi

It is X86_64, SMP
It appears during boot process.

CONFIG_HPET_TIMER=y
CONFIG_HPET_EMULATE_RTC=y
CONFIG_HPET=y
CONFIG_HPET_MMAP=y
CONFIG_DEBUG_OBJECTS=y
# CONFIG_DEBUG_OBJECTS_SELFTEST is not set
CONFIG_DEBUG_OBJECTS_FREE=y
CONFIG_DEBUG_OBJECTS_TIMERS=y

hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0
hpet0: 3 comparators, 64-bit 14.318180 MHz counter
ODEBUG: object is on stack, but not annotated
------------[ cut here ]------------
WARNING: at lib/debugobjects.c:251 __debug_object_init+0x2a4/0x352()
Modules linked in:
Pid: 1, comm: swapper Not tainted 2.6.28-rc6-00007-ged31348 #5
Call Trace:
 [<ffffffff8023661b>] warn_on_slowpath+0x58/0x7d
 [<ffffffff804de4b6>] ? printk+0x67/0x69
 [<ffffffff8037fd54>] ? __debug_object_init+0x185/0x352
 [<ffffffff8037fe73>] __debug_object_init+0x2a4/0x352
 [<ffffffff8037ff4e>] debug_object_init+0x14/0x16
 [<ffffffff8023f681>] init_timer+0x18/0x5b
 [<ffffffff80220739>] hpet_cpuhp_notify+0x93/0x105
 [<ffffffff80220872>] ? hpet_work+0x0/0x206
 [<ffffffff803d0301>] ? hpet_alloc+0x32d/0x389
 [<ffffffff80256db2>] ? trace_hardirqs_on_caller+0x10d/0x138
 [<ffffffff80256dea>] ? trace_hardirqs_on+0xd/0xf
 [<ffffffff806c45b7>] ? hpet_late_init+0x0/0x19e
 [<ffffffff806c45b7>] ? hpet_late_init+0x0/0x19e
 [<ffffffff806c4721>] hpet_late_init+0x16a/0x19e
 [<ffffffff806c27a9>] ? print_all_ICs+0x0/0x540
 [<ffffffff80209058>] _stext+0x58/0x138
 [<ffffffff804e0ecb>] ? _spin_unlock+0x4a/0x57
 [<ffffffff802db669>] ? proc_register+0x17f/0x193
 [<ffffffff802db7a5>] ? create_proc_entry+0x7e/0x94
 [<ffffffff80267d15>] ? register_irq_proc+0xb0/0xcc
 [<ffffffff802d0000>] ? elf_core_dump+0x662/0xc71
 [<ffffffff806b9625>] kernel_init+0x125/0x179
 [<ffffffff804e087f>] ? trace_hardirqs_on_thunk+0x3a/0x3f
 [<ffffffff8020c849>] child_rip+0xa/0x11
 [<ffffffff8020bd38>] ? restore_args+0x0/0x30
 [<ffffffff806b9500>] ? kernel_init+0x0/0x179
 [<ffffffff8020c83f>] ? child_rip+0x0/0x11
---[ end trace 4eaa2a86a8e2da22 ]---
Switched to high resolution mode on CPU 0
Switched to high resolution mode on CPU 3
Switched to high resolution mode on CPU 2
Switched to high resolution mode on CPU 1

Re: 2.6.28-rc6: hpet: object is on stack, but not annotated

[Alexander Beregalov]

2008/11/24 Alexander Beregalov <a.beregalov-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>:
> hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0
> hpet0: 3 comparators, 64-bit 14.318180 MHz counter
> ODEBUG: object is on stack, but not annotated
> ------------[ cut here ]------------
> WARNING: at lib/debugobjects.c:251 __debug_object_init+0x2a4/0x352()

Bisected down to 26afe5f2fbf06ea0765aaa316640c4dd472310c0
(x86: HPET_MSI Initialise per-cpu HPET timers)

Revert helps

Re: 2.6.28-rc6: hpet: object is on stack, but not annotated

[Thomas Gleixner]

On Mon, 24 Nov 2008, Alexander Beregalov wrote:

> 2008/11/24 Alexander Beregalov <a.beregalov-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>:
> > hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0
> > hpet0: 3 comparators, 64-bit 14.318180 MHz counter
> > ODEBUG: object is on stack, but not annotated
> > ------------[ cut here ]------------
> > WARNING: at lib/debugobjects.c:251 __debug_object_init+0x2a4/0x352()
> 
> Bisected down to 26afe5f2fbf06ea0765aaa316640c4dd472310c0
> (x86: HPET_MSI Initialise per-cpu HPET timers)
> 
> Revert helps

That's just a warning about a missing annotation. The commit is fine.

Thanks,

	tglx

Re: 2.6.28-rc6: hpet: object is on stack, but not annotated

[Pavel Machek]

On Mon 2008-11-24 19:17:44, Thomas Gleixner wrote:
> On Mon, 24 Nov 2008, Alexander Beregalov wrote:
> 
> > 2008/11/24 Alexander Beregalov <a.beregalov-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>:
> > > hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0
> > > hpet0: 3 comparators, 64-bit 14.318180 MHz counter
> > > ODEBUG: object is on stack, but not annotated
> > > ------------[ cut here ]------------
> > > WARNING: at lib/debugobjects.c:251 __debug_object_init+0x2a4/0x352()
> > 
> > Bisected down to 26afe5f2fbf06ea0765aaa316640c4dd472310c0
> > (x86: HPET_MSI Initialise per-cpu HPET timers)
> > 
> > Revert helps
> 
> That's just a warning about a missing annotation. The commit is fine.

You must be using very different metric of what "fine" means than rest
of the world.

(No, introducing WARN_ON()s is not fine; it spams kerneloops.org at least...)

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: 2.6.28-rc6: hpet: object is on stack, but not annotated

[Alexander Beregalov]

2008/12/4 Pavel Machek <pavel-AlSwsSmVLrQ@public.gmane.org>:
> On Mon 2008-11-24 19:17:44, Thomas Gleixner wrote:
>> On Mon, 24 Nov 2008, Alexander Beregalov wrote:
>>
>> > 2008/11/24 Alexander Beregalov <a.beregalov-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>:
>> > > hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0
>> > > hpet0: 3 comparators, 64-bit 14.318180 MHz counter
>> > > ODEBUG: object is on stack, but not annotated
>> > > ------------[ cut here ]------------
>> > > WARNING: at lib/debugobjects.c:251 __debug_object_init+0x2a4/0x352()
>> >
>> > Bisected down to 26afe5f2fbf06ea0765aaa316640c4dd472310c0
>> > (x86: HPET_MSI Initialise per-cpu HPET timers)
>> >
>> > Revert helps
>>
>> That's just a warning about a missing annotation. The commit is fine.
>
> You must be using very different metric of what "fine" means than rest
> of the world.
>
> (No, introducing WARN_ON()s is not fine; it spams kerneloops.org at least...)

The warning is still here on 2.6.29-rc1.
It taints the kernel.

Re: 2.6.28-rc6: hpet: object is on stack, but not annotated

[Peter Zijlstra]

On Mon, 2009-01-12 at 13:40 +0300, Alexander Beregalov wrote:
> 2008/12/4 Pavel Machek <pavel@suse.cz>:
> > On Mon 2008-11-24 19:17:44, Thomas Gleixner wrote:
> >> On Mon, 24 Nov 2008, Alexander Beregalov wrote:
> >>
> >> > 2008/11/24 Alexander Beregalov <a.beregalov@gmail.com>:
> >> > > hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0
> >> > > hpet0: 3 comparators, 64-bit 14.318180 MHz counter
> >> > > ODEBUG: object is on stack, but not annotated
> >> > > ------------[ cut here ]------------
> >> > > WARNING: at lib/debugobjects.c:251 __debug_object_init+0x2a4/0x352()
> >> >
> >> > Bisected down to 26afe5f2fbf06ea0765aaa316640c4dd472310c0
> >> > (x86: HPET_MSI Initialise per-cpu HPET timers)
> >> >
> >> > Revert helps
> >>
> >> That's just a warning about a missing annotation. The commit is fine.
> >
> > You must be using very different metric of what "fine" means than rest
> > of the world.
> >
> > (No, introducing WARN_ON()s is not fine; it spams kerneloops.org at least...)
> 
> The warning is still here on 2.6.29-rc1.
> It taints the kernel.

Does this fix it?

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
---
 arch/x86/kernel/hpet.c    |    2 +-
 include/linux/workqueue.h |    6 ++++++
 2 files changed, 7 insertions(+), 1 deletions(-)

diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
index cd759ad..bb2e0f0 100644
--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -628,7 +628,7 @@ static int hpet_cpuhp_notify(struct notifier_block *n,
 
 	switch (action & 0xf) {
 	case CPU_ONLINE:
-		INIT_DELAYED_WORK(&work.work, hpet_work);
+		INIT_DELAYED_WORK_ON_STACK(&work.work, hpet_work);
 		init_completion(&work.complete);
 		/* FIXME: add schedule_work_on() */
 		schedule_delayed_work_on(cpu, &work.work, 0);
diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h
index b362911..47151c8 100644
--- a/include/linux/workqueue.h
+++ b/include/linux/workqueue.h
@@ -118,6 +118,12 @@ struct execute_work {
 		init_timer(&(_work)->timer);			\
 	} while (0)
 
+#define INIT_DELAYED_WORK_ON_STACK(_work, _func)		\
+	do {							\
+		INIT_WORK(&(_work)->work, (_func));		\
+		init_timer_on_stack(&(_work)->timer);		\
+	} while (0)
+
 #define INIT_DELAYED_WORK_DEFERRABLE(_work, _func)			\
 	do {							\
 		INIT_WORK(&(_work)->work, (_func));		\

Re: 2.6.28-rc6: hpet: object is on stack, but not annotated

[Ingo Molnar]

* Peter Zijlstra <peterz-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org> wrote:

> > > (No, introducing WARN_ON()s is not fine; it spams kerneloops.org at 
> > > least...)
> > 
> > The warning is still here on 2.6.29-rc1. It taints the kernel.
> 
> Does this fix it?

i've applied it to tip/timers/urgent and will send it to Linus if it 
passes Alexander's testing too. Below is the prettified commit.

Thanks guys,

	Ingo
-------------->
From b1f1006971b6f8d330c6e8ecf240f51c9eb67f4d Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <peterz-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
Date: Mon, 12 Jan 2009 12:52:23 +0100
Subject: [PATCH] locking, hpet: annotate false positive warning

Alexander Beregalov reported that this warning is caused by the HPET code:

> hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0
> hpet0: 3 comparators, 64-bit 14.318180 MHz counter
> ODEBUG: object is on stack, but not annotated
> ------------[ cut here ]------------
> WARNING: at lib/debugobjects.c:251 __debug_object_init+0x2a4/0x352()

> Bisected down to 26afe5f2fbf06ea0765aaa316640c4dd472310c0
> (x86: HPET_MSI Initialise per-cpu HPET timers)

The commit is fine - but the on-stack workqueue entry needs annotation.

Reported-and-bisected-by: Alexander Beregalov <a.beregalov-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Signed-off-by: Peter Zijlstra <a.p.zijlstra-/NLkJaSkS4VmR6Xm/wNWPw@public.gmane.org>
Signed-off-by: Ingo Molnar <mingo-X9Un+BFzKDI@public.gmane.org>
---
 arch/x86/kernel/hpet.c    |    2 +-
 include/linux/workqueue.h |    6 ++++++
 2 files changed, 7 insertions(+), 1 deletions(-)

diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
index cd759ad..bb2e0f0 100644
--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -628,7 +628,7 @@ static int hpet_cpuhp_notify(struct notifier_block *n,
 
 	switch (action & 0xf) {
 	case CPU_ONLINE:
-		INIT_DELAYED_WORK(&work.work, hpet_work);
+		INIT_DELAYED_WORK_ON_STACK(&work.work, hpet_work);
 		init_completion(&work.complete);
 		/* FIXME: add schedule_work_on() */
 		schedule_delayed_work_on(cpu, &work.work, 0);
diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h
index b362911..47151c8 100644
--- a/include/linux/workqueue.h
+++ b/include/linux/workqueue.h
@@ -118,6 +118,12 @@ struct execute_work {
 		init_timer(&(_work)->timer);			\
 	} while (0)
 
+#define INIT_DELAYED_WORK_ON_STACK(_work, _func)		\
+	do {							\
+		INIT_WORK(&(_work)->work, (_func));		\
+		init_timer_on_stack(&(_work)->timer);		\
+	} while (0)
+
 #define INIT_DELAYED_WORK_DEFERRABLE(_work, _func)			\
 	do {							\
 		INIT_WORK(&(_work)->work, (_func));		\

Re: 2.6.28-rc6: hpet: object is on stack, but not annotated

[Alexander Beregalov]

2009/1/12 Ingo Molnar <mingo-X9Un+BFzKDI@public.gmane.org>:
>
> * Peter Zijlstra <peterz-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org> wrote:
>
>> > > (No, introducing WARN_ON()s is not fine; it spams kerneloops.org at
>> > > least...)
>> >
>> > The warning is still here on 2.6.29-rc1. It taints the kernel.
>>
>> Does this fix it?
Yes, it does. Thanks Peter.
>
> i've applied it to tip/timers/urgent and will send it to Linus if it
> passes Alexander's testing too. Below is the prettified commit.
Yes, please.
>
> Thanks guys,
>
>        Ingo
> -------------->
> From b1f1006971b6f8d330c6e8ecf240f51c9eb67f4d Mon Sep 17 00:00:00 2001
> From: Peter Zijlstra <peterz-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
> Date: Mon, 12 Jan 2009 12:52:23 +0100
> Subject: [PATCH] locking, hpet: annotate false positive warning
>
> Alexander Beregalov reported that this warning is caused by the HPET code:
>
>> hpet0: at MMIO 0xfed00000, IRQs 2, 8, 0
>> hpet0: 3 comparators, 64-bit 14.318180 MHz counter
>> ODEBUG: object is on stack, but not annotated
>> ------------[ cut here ]------------
>> WARNING: at lib/debugobjects.c:251 __debug_object_init+0x2a4/0x352()
>
>> Bisected down to 26afe5f2fbf06ea0765aaa316640c4dd472310c0
>> (x86: HPET_MSI Initialise per-cpu HPET timers)
>
> The commit is fine - but the on-stack workqueue entry needs annotation.
>
> Reported-and-bisected-by: Alexander Beregalov <a.beregalov-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> Signed-off-by: Peter Zijlstra <a.p.zijlstra-/NLkJaSkS4VmR6Xm/wNWPw@public.gmane.org>
> Signed-off-by: Ingo Molnar <mingo-X9Un+BFzKDI@public.gmane.org>
> ---
>  arch/x86/kernel/hpet.c    |    2 +-
>  include/linux/workqueue.h |    6 ++++++
>  2 files changed, 7 insertions(+), 1 deletions(-)
>
> diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
> index cd759ad..bb2e0f0 100644
> --- a/arch/x86/kernel/hpet.c
> +++ b/arch/x86/kernel/hpet.c
> @@ -628,7 +628,7 @@ static int hpet_cpuhp_notify(struct notifier_block *n,
>
>        switch (action & 0xf) {
>        case CPU_ONLINE:
> -               INIT_DELAYED_WORK(&work.work, hpet_work);
> +               INIT_DELAYED_WORK_ON_STACK(&work.work, hpet_work);
>                init_completion(&work.complete);
>                /* FIXME: add schedule_work_on() */
>                schedule_delayed_work_on(cpu, &work.work, 0);
> diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h
> index b362911..47151c8 100644
> --- a/include/linux/workqueue.h
> +++ b/include/linux/workqueue.h
> @@ -118,6 +118,12 @@ struct execute_work {
>                init_timer(&(_work)->timer);                    \
>        } while (0)
>
> +#define INIT_DELAYED_WORK_ON_STACK(_work, _func)               \
> +       do {                                                    \
> +               INIT_WORK(&(_work)->work, (_func));             \
> +               init_timer_on_stack(&(_work)->timer);           \
> +       } while (0)
> +
>  #define INIT_DELAYED_WORK_DEFERRABLE(_work, _func)                     \
>        do {                                                    \
>                INIT_WORK(&(_work)->work, (_func));             \
>

Re: 2.6.28-rc6: hpet: object is on stack, but not annotated

[Ingo Molnar]

* Alexander Beregalov <a.beregalov-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> 2009/1/12 Ingo Molnar <mingo-X9Un+BFzKDI@public.gmane.org>:
> >
> > * Peter Zijlstra <peterz-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org> wrote:
> >
> >> > > (No, introducing WARN_ON()s is not fine; it spams kerneloops.org at
> >> > > least...)
> >> >
> >> > The warning is still here on 2.6.29-rc1. It taints the kernel.
> >>
> >> Does this fix it?
> Yes, it does. Thanks Peter.

thanks - added your Tested-by tag to the commit too.

	Ingo