Re: -mm hang while shutdown with printk_delay

[Dave Young]

On Wed, Aug 19, 2009 at 9:47 PM, Dave Young<hidave.darkstar@xxxxxxxxx> wrote:
> On Tue, Aug 18, 2009 at 12:12 PM, Dave Young<hidave.darkstar@xxxxxxxxx> wrote:
>> On Tue, Aug 18, 2009 at 11:30 AM, Andrew
>> Morton<akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
>>> On Tue, 18 Aug 2009 11:25:22 +0800 Dave Young <hidave.darkstar@xxxxxxxxx> wrote:
>>>
>>>> Hi,
>>>>
>>>> Test with printk_delay=200,
>>>>
>>>> /sbin/halt -p
>>>>
>>>> kernel hang after following message:
>>>>
>>>> Disabling non-boot CPUs
>>>>
>>>> There's no such problem with 2.6.31-rc6
>>>
>>> umm, OK. ÂBut -mm contains six patches:
>>>
>>> printk-boot_delay-rename-printk_delay_msec-to-loops_per_msec.patch
>>> printk-boot_delay-rename-printk_delay_msec-to-loops_per_msec-fix.patch
>>> printk-boot_delay-rename-printk_delay_msec-to-loops_per_msec-fix-2.patch
>>> printk-add-printk_delay-to-make-messages-readable-for-some-scenarios.patch
>>> printk-add-printk_delay-to-make-messages-readable-for-some-scenarios-fix.patch
>>> printk-add-printk_delay-to-make-messages-readable-for-some-scenarios-cleanup.patch
>>>
>>> from yourself. ÂAre they the cause?
>>>
>>
>> Actually I tested 2.6.31-rc6 with the above six patches applied,
>> there's no such problem.
>>
>
> I'm manually bisecting the mm patch series, hope find cause.

Again weird result:

linux-futexh-place-kernel-types-behind-__kernel__.patch
#bisect good
#
#
# edac
#
edac-mpc85xx-add-p2020ds-support.patch
#edac-mpc85xx-add-mpc83xx-support.patch: smp_processor_id() bug?
edac-mpc85xx-add-mpc83xx-support.patch
edac-fix-resource-size-calculation.patch
edac-i3200-memory-controller-driver.patch
edac-i3200-memory-controller-driver-fix-offset-of-reg-in-i3200_edac-module.patch
# bisect bad

I put printk_delay patches to the head of queue. test five times, same results.

bash-3.1$ grep EDAC .config
CONFIG_EDAC=y
# CONFIG_EDAC_DEBUG is not set
# CONFIG_EDAC_MM_EDAC is not set

There's no difference between bad/good kernel indeed because the edac
code is not compiled.

hardware is dell e5400
distribution is slackware 12.2

BTW, there's lockdep warnings while umounting a reiserfs partition,
but I think it is not relevant.

Confused...

-- 
Regards
dave
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/