warn_on regression after 2.6.27-rc5

warn_on regression after 2.6.27-rc5

[Christian Borntraeger]

Hello Thomas,

this commit 

commit 1fb9b7d29d8e85ba3196eaa7ab871bf76fc98d36
Author: Thomas Gleixner <tglx@linutronix.de>
Date:   Wed Sep 3 21:37:14 2008 +0000

    clockevents: prevent endless loop lockup


causes the following warning on my Thinkpad T61p:

CE: __tick_program_event of hpet is stuck 7416220944 7416221ccc
CE: increasing min_delta_ns 5000 to 10000 nsec
------------[ cut here ]------------
WARNING: at kernel/time/tick-oneshot.c:52 tick_dev_program_event+0xd6/0x120()
Modules linked in: kvm_intel kvm fan iptable_nat nf_nat iptable_mangle 
ipt_REJECT nf_conntrack_ipv4 xt_state nf_conntrack iptable_filter ip_tables 
x_tables ctr blowfish rmd160 sha256_generic crypto_null nls_utf8 ntfs fuse 
acpi_cpufreq usbhid pcmcia snd_hda_intel snd_pcm_oss snd_mixer_oss snd_pcm 
snd_timer snd_page_alloc snd_hwdep snd yenta_socket iwlagn rsrc_nonstatic 
pcmcia_core soundcore ehci_hcd uhci_hcd e1000e thermal evdev video battery 
usbcore ac bay processor button
Pid: 0, comm: swapper Not tainted 2.6.27-rc5-00319-g7686ad5 #15
 [<c012e32f>] warn_on_slowpath+0x5f/0xa0
 [<c026f950>] vt_console_print+0x0/0x330
 [<c03f0554>] _spin_lock_irqsave+0x34/0x50
 [<c012eb71>] release_console_sem+0x191/0x1c0
 [<c012f0ce>] vprintk+0x32e/0x3a0
 [<c01487d1>] getnstimeofday+0x41/0xf0
 [<c0114d88>] lapic_next_event+0x18/0x30
 [<c01487d1>] getnstimeofday+0x41/0xf0
 [<c014bd73>] clockevents_program_event+0xa3/0x160
 [<c014cfe6>] tick_dev_program_event+0xd6/0x120
 [<c014c81d>] tick_broadcast_oneshot_control+0x10d/0x120
 [<c014c3a5>] tick_notify+0x295/0x3c0
 [<c0146df6>] notifier_call_chain+0x36/0x70
 [<c0146ee7>] raw_notifier_call_chain+0x17/0x20
 [<c014bbbe>] clockevents_notify+0x1e/0x70
 [<f884f087>] acpi_idle_enter_bm+0x15e/0x2c2 [processor]
 [<c031fed3>] cpuidle_idle_call+0x73/0xd0
 [<c010249d>] cpu_idle+0x6d/0x140
 =======================
---[ end trace c7d6f2472868c172 ]---

While I agree with doubling the time, I am not sure if a WARN_ON is 
appropriate. After all the code is able to fix the timer setting.

Christian

Re: warn_on regression after 2.6.27-rc5

[Christian Borntraeger]

Am Mittwoch, 10. September 2008 schrieb Christian Borntraeger:
> Hello Thomas,
> 
> this commit 
> 
> commit 1fb9b7d29d8e85ba3196eaa7ab871bf76fc98d36
> Author: Thomas Gleixner <tglx@linutronix.de>
> Date:   Wed Sep 3 21:37:14 2008 +0000
> 
>     clockevents: prevent endless loop lockup
> 
> 
> causes the following warning on my Thinkpad T61p:

Ok, forget it. I just found the mail thread and the patch in my last git 
update.

Thanks

Christian

Re: warn_on regression after 2.6.27-rc5

[Thomas Gleixner]

On Wed, 10 Sep 2008, Christian Borntraeger wrote:
> 
> While I agree with doubling the time, I am not sure if a WARN_ON is 
> appropriate. After all the code is able to fix the timer setting.

The WARN_ON is removed already. It was there temporary to gather
information, which worked :)

Thanks,

	tglx