hibernate causes nobody cared on acpi irq

hibernate causes nobody cared on acpi irq

[Jiri Slaby]

Hi,

with current mmotm I get this while trying to hibernate from a system in
qemu:

PM: Syncing filesystems ... done.
Freezing user space processes ... (elapsed 0.00 seconds) done.
Freezing remaining freezable tasks ... (elapsed 0.03 seconds) done.
PM: Preallocating image memory... done (allocated 7986 pages)
PM: Allocated 31944 kbytes in 0.01 seconds (3194.40 MB/s)
ACPI: Preparing to enter system sleep state S4
PM: Saving platform NVS memory
Disabling non-boot CPUs ...
Cannot set affinity for irq 0
CPU 1 is now offline
SMP alternatives: switching to UP code
CPU1 is down
PM: Creating hibernation image:
PM: Need to copy 7687 pages
PM: Hibernation image created (7687 pages copied)
Enabling non-boot CPUs ...
SMP alternatives: switching to SMP code
Booting processor 1 APIC 0x1 ip 0x6000
Initializing CPU#1
Calibrating delay using timer specific routine.. 5985.37 BogoMIPS
(lpj=11970749)
CPU: L1 I cache: 32K, L1 D cache: 32K
CPU: L2 cache: 2048K
mce: CPU supports 32 MCE banks
CPU1: Intel QEMU Virtual CPU version 0.10.50 stepping 03
checking TSC synchronization [CPU#0 -> CPU#1]: passed.
Switched to high resolution mode on CPU 1
CPU1 is up
ACPI: Waking up from system sleep state S4
irq 9: nobody cared (try booting with the "irqpoll" option)
Pid: 497, comm: sh Tainted: G        W  2.6.31-rc5-mm1_64 #668
Call Trace:
 <IRQ>  [<ffffffff810815c6>] __report_bad_irq+0x26/0xa0
 [<ffffffff810817d0>] note_interrupt+0x190/0x1d0
 [<ffffffff81197e49>] ? acpi_irq+0x11/0x2c
 [<ffffffff81081e17>] handle_fasteoi_irq+0xb7/0xe0
 [<ffffffff8100e80d>] handle_irq+0x1d/0x30
 [<ffffffff8100dd37>] do_IRQ+0x67/0xe0
 [<ffffffff8100c813>] ret_from_intr+0x0/0xa
 [<ffffffff8104543f>] ? __do_softirq+0x5f/0x130
 [<ffffffff8102449d>] ? ack_apic_level+0x7d/0x1e0
 [<ffffffff8100cf9c>] ? call_softirq+0x1c/0x30
 [<ffffffff8100e7bd>] ? do_softirq+0x4d/0x80
 [<ffffffff8104520d>] ? irq_exit+0x7d/0x90
 [<ffffffff8100dd40>] ? do_IRQ+0x70/0xe0
 [<ffffffff8100c813>] ? ret_from_intr+0x0/0xa
 <EOI>  [<ffffffff81429c08>] ? _spin_unlock_irqrestore+0x8/0x10
 [<ffffffff810830a3>] ? resume_device_irqs+0x73/0x90
 [<ffffffff8123fb31>] ? dpm_resume_noirq+0xb1/0xc0
 [<ffffffff8106edaa>] ? hibernation_snapshot+0x12a/0x230
 [<ffffffff8106ef84>] ? hibernate+0xd4/0x1a0
 [<ffffffff8106d88c>] ? state_store+0xec/0x100
 [<ffffffff81168987>] ? kobj_attr_store+0x17/0x20
 [<ffffffff81119190>] ? sysfs_write_file+0xe0/0x160
 [<ffffffff810c1648>] ? vfs_write+0xb8/0x1a0
 [<ffffffff8142c5a5>] ? do_page_fault+0x185/0x350
 [<ffffffff810c1f4c>] ? sys_write+0x4c/0x80
 [<ffffffff810cf1c0>] ? sys_fcntl+0x150/0x410
 [<ffffffff8100beeb>] ? system_call_fastpath+0x16/0x1b
handlers:
[<ffffffff81197e38>] (acpi_irq+0x0/0x2c)
Disabling IRQ #9

I use this hunk to prevent bh enablement btw.:
diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c
index c1f64e6..bc5368d 100644
--- a/kernel/hw_breakpoint.c
+++ b/kernel/hw_breakpoint.c
@@ -80,17 +80,15 @@ void load_debug_registers(void)
        unsigned long flags;
        struct task_struct *tsk = current;

-       spin_lock_bh(&hw_breakpoint_lock);
+       spin_lock_irqsave(&hw_breakpoint_lock, flags);

        /* Prevent IPIs for new kernel breakpoint updates */
-       local_irq_save(flags);
        arch_update_kernel_hw_breakpoint(NULL);
-       local_irq_restore(flags);

        if (test_tsk_thread_flag(tsk, TIF_DEBUG))
                arch_install_thread_hw_breakpoint(tsk);

-       spin_unlock_bh(&hw_breakpoint_lock);
+       spin_unlock_irqrestore(&hw_breakpoint_lock, flags);
 }

 /*

Re: hibernate causes nobody cared on acpi irq

[Rafael J. Wysocki]

On Thursday 13 August 2009, Jiri Slaby wrote:
> Hi,

Hi,

> with current mmotm I get this while trying to hibernate from a system in
> qemu:

Does it trigger with the "core" test:

# echo core > /sys/power/pm_test
# echo disk > /sys/power/state

> PM: Syncing filesystems ... done.
> Freezing user space processes ... (elapsed 0.00 seconds) done.
> Freezing remaining freezable tasks ... (elapsed 0.03 seconds) done.
> PM: Preallocating image memory... done (allocated 7986 pages)
> PM: Allocated 31944 kbytes in 0.01 seconds (3194.40 MB/s)
> ACPI: Preparing to enter system sleep state S4
> PM: Saving platform NVS memory
> Disabling non-boot CPUs ...
> Cannot set affinity for irq 0
> CPU 1 is now offline
> SMP alternatives: switching to UP code
> CPU1 is down
> PM: Creating hibernation image:
> PM: Need to copy 7687 pages
> PM: Hibernation image created (7687 pages copied)
> Enabling non-boot CPUs ...
> SMP alternatives: switching to SMP code
> Booting processor 1 APIC 0x1 ip 0x6000
> Initializing CPU#1
> Calibrating delay using timer specific routine.. 5985.37 BogoMIPS
> (lpj=11970749)
> CPU: L1 I cache: 32K, L1 D cache: 32K
> CPU: L2 cache: 2048K
> mce: CPU supports 32 MCE banks
> CPU1: Intel QEMU Virtual CPU version 0.10.50 stepping 03
> checking TSC synchronization [CPU#0 -> CPU#1]: passed.
> Switched to high resolution mode on CPU 1
> CPU1 is up
> ACPI: Waking up from system sleep state S4
> irq 9: nobody cared (try booting with the "irqpoll" option)
> Pid: 497, comm: sh Tainted: G        W  2.6.31-rc5-mm1_64 #668
> Call Trace:
>  <IRQ>  [<ffffffff810815c6>] __report_bad_irq+0x26/0xa0
>  [<ffffffff810817d0>] note_interrupt+0x190/0x1d0
>  [<ffffffff81197e49>] ? acpi_irq+0x11/0x2c
>  [<ffffffff81081e17>] handle_fasteoi_irq+0xb7/0xe0
>  [<ffffffff8100e80d>] handle_irq+0x1d/0x30
>  [<ffffffff8100dd37>] do_IRQ+0x67/0xe0
>  [<ffffffff8100c813>] ret_from_intr+0x0/0xa
>  [<ffffffff8104543f>] ? __do_softirq+0x5f/0x130
>  [<ffffffff8102449d>] ? ack_apic_level+0x7d/0x1e0
>  [<ffffffff8100cf9c>] ? call_softirq+0x1c/0x30
>  [<ffffffff8100e7bd>] ? do_softirq+0x4d/0x80
>  [<ffffffff8104520d>] ? irq_exit+0x7d/0x90
>  [<ffffffff8100dd40>] ? do_IRQ+0x70/0xe0
>  [<ffffffff8100c813>] ? ret_from_intr+0x0/0xa
>  <EOI>  [<ffffffff81429c08>] ? _spin_unlock_irqrestore+0x8/0x10
>  [<ffffffff810830a3>] ? resume_device_irqs+0x73/0x90
>  [<ffffffff8123fb31>] ? dpm_resume_noirq+0xb1/0xc0
>  [<ffffffff8106edaa>] ? hibernation_snapshot+0x12a/0x230
>  [<ffffffff8106ef84>] ? hibernate+0xd4/0x1a0
>  [<ffffffff8106d88c>] ? state_store+0xec/0x100
>  [<ffffffff81168987>] ? kobj_attr_store+0x17/0x20
>  [<ffffffff81119190>] ? sysfs_write_file+0xe0/0x160
>  [<ffffffff810c1648>] ? vfs_write+0xb8/0x1a0
>  [<ffffffff8142c5a5>] ? do_page_fault+0x185/0x350
>  [<ffffffff810c1f4c>] ? sys_write+0x4c/0x80
>  [<ffffffff810cf1c0>] ? sys_fcntl+0x150/0x410
>  [<ffffffff8100beeb>] ? system_call_fastpath+0x16/0x1b
> handlers:
> [<ffffffff81197e38>] (acpi_irq+0x0/0x2c)
> Disabling IRQ #9
> 
> I use this hunk to prevent bh enablement btw.:
> diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c
> index c1f64e6..bc5368d 100644
> --- a/kernel/hw_breakpoint.c
> +++ b/kernel/hw_breakpoint.c
> @@ -80,17 +80,15 @@ void load_debug_registers(void)
>         unsigned long flags;
>         struct task_struct *tsk = current;
> 
> -       spin_lock_bh(&hw_breakpoint_lock);
> +       spin_lock_irqsave(&hw_breakpoint_lock, flags);
> 
>         /* Prevent IPIs for new kernel breakpoint updates */
> -       local_irq_save(flags);
>         arch_update_kernel_hw_breakpoint(NULL);
> -       local_irq_restore(flags);
> 
>         if (test_tsk_thread_flag(tsk, TIF_DEBUG))
>                 arch_install_thread_hw_breakpoint(tsk);
> 
> -       spin_unlock_bh(&hw_breakpoint_lock);
> +       spin_unlock_irqrestore(&hw_breakpoint_lock, flags);
>  }
> 
>  /*

Rafael

Re: hibernate causes nobody cared on acpi irq

[Jiri Slaby]

On 08/13/2009 12:21 AM, Rafael J. Wysocki wrote:
> On Thursday 13 August 2009, Jiri Slaby wrote:
>> with current mmotm I get this while trying to hibernate from a system in
>> qemu:
> 
> Does it trigger with the "core" test:

Yes, but I found out it's totally unrelated to the susp code. There is
an interrupt storm on IRQ9 after system boot -- 150000 intrs on both
CPUs in the first minute. Then the message emerges.

It's just a coincidence it appears during hibernation (after interrupts
enable phase).

Any idea what it could be?

Thanks.

Re: hibernate causes nobody cared on acpi irq

[Rafael J. Wysocki]

On Thursday 13 August 2009, Jiri Slaby wrote:
> On 08/13/2009 12:21 AM, Rafael J. Wysocki wrote:
> > On Thursday 13 August 2009, Jiri Slaby wrote:
> >> with current mmotm I get this while trying to hibernate from a system in
> >> qemu:
> > 
> > Does it trigger with the "core" test:
> 
> Yes, but I found out it's totally unrelated to the susp code. There is
> an interrupt storm on IRQ9 after system boot -- 150000 intrs on both
> CPUs in the first minute. Then the message emerges.
> 
> It's just a coincidence it appears during hibernation (after interrupts
> enable phase).
> 
> Any idea what it could be?

Well, nothing obvious comes to mind.

Rafael

Re: hibernate causes nobody cared on acpi irq

[Jiri Slaby]

On 08/13/2009 10:20 PM, Rafael J. Wysocki wrote:
> Well, nothing obvious comes to mind.

Hmm, looks like a fresh qemu-kvm bug. noapic fixes that. I never needed
that, but even 2.6.30 doesn't work with it now. Or maybe something on
the host kernel broke recently. -ETOOMANYBUGS. Thanks.

Re: hibernate causes nobody cared on acpi irq

[Jiri Slaby]

On 08/13/2009 10:46 PM, Jiri Slaby wrote:
> On 08/13/2009 10:20 PM, Rafael J. Wysocki wrote:
>> Well, nothing obvious comes to mind.
> 
> Hmm, looks like a fresh qemu-kvm bug.

FWIW downgrading from kvm userspace 88 to 78 fixed the issue.