Bug 13094

Summary: Thinkpad X40 randomly hangs while booting.
Product: ACPI Reporter: Daniel Vetter (daniel)
Component: Power-ProcessorAssignee: ykzhao (yakui.zhao)
Status: CLOSED DUPLICATE    
Severity: normal CC: akpm, alex.shi, lenb, rui.zhang, venki, yakui.zhao
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 2.6.29-rc2 Subsystem:
Regression: Yes Bisected commit-id:
Attachments: my .config

Description Daniel Vetter 2009-04-15 21:25:46 UTC 
Created attachment 21003 [details]
my .config

Hardware: Thinkpad X40 Laptop, Pentium M 1.2 GHz, Bios v2.07
Software: Debian unstable x86_32

When booting up 2.6.29-rc2, the system hangs hard somewhat randomly (usually when loading kernel modules). No reaction to sysrq. I used bisecting to track down the problem to the following commit

commit ff69f2bba67bd45514923aaedbf40fe351787c59
Author: alex.shi <alex.shi@intel.com>
Date:   Wed Mar 4 11:55:26 2009 -0800

    acpi: fix of pmtimer overflow that make Cx states time incorrect

Reverting this on top of 2.6.29-rc2 gets rid of the hangs. I'll attach my config. 2.6.29 is not affected, so this is a post-2.6.29 regression.
Comment 1 Andrew Morton 2009-04-15 21:47:27 UTC 
2.6.29-rc2 is a funny kernel version to be testing: -rc2 is always full
of regressions.

Is the bug also present in 2.6.29 final?

Thanks.
Comment 2 Alex Shi 2009-04-16 01:36:09 UTC 
it is duplicated with bug 13087, but I did not found some ways set this as previous bugzilla.  :(
Comment 3 ykzhao 2009-04-16 02:17:10 UTC 
Hi, Alex
    What you said is right. It is the duplicate bug of 13087.
    
    Will you please try the boot option of "clocksource=acpi_pm" and see whether the box still hangs?
    Thanks.
Comment 4 ykzhao 2009-04-16 02:19:48 UTC 
Hi, Daniel
    Will you please try the boot option of "clocksource=acpi_pm" and see whether the box still hangs?
    
    Do you mean that the system can be booted normally if the following commit is reverted on the latest git-tree?
    >ff69f2bba67bd45514923aaedbf40fe351787c59


    thanks.
Comment 5 Zhang Rui 2009-04-16 03:23:17 UTC 

*** This bug has been marked as a duplicate of bug 13087 ***
Comment 6 Daniel Vetter 2009-04-16 08:41:26 UTC 
On Thu, Apr 16, 2009 at 02:19:48AM +0000, bugzilla-daemon@bugzilla.kernel.org wrote:
> --- Comment #4 from ykzhao <yakui.zhao@intel.com>  2009-04-16 02:19:48 ---
> Hi, Daniel
>     Will you please try the boot option of "clocksource=acpi_pm" and see
> whether the box still hangs?

I've just tested and yes, it works with on 2.6.30-rc2 (sorry for yet again
screwing up the version numbers - when bisecting I'm always confused by
the fact that the git versions starts with with the old version number
(2.6.29) before rc1).

>     Do you mean that the system can be booted normally if the following
>     commit
> is reverted on the latest git-tree?
>     >ff69f2bba67bd45514923aaedbf40fe351787c59

Yes, reverting ff69.. on top of 2.6.30-rc2 fixes the hangs, too.

Thanks, Daniel