Bug 12757

Summary: [BISECT]kernels later than 2.6.26.8 die at cardbus init on laptop
Product: Platform Specific/Hardware Reporter: Rafael J. Wysocki (rjw)
Component: i386Assignee: ykzhao (yakui.zhao)
Status: CLOSED CODE_FIX    
Severity: normal CC: acpi-bugzilla, alan, dtpflug, florian, mingo
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 2.6.28 Subsystem:
Regression: Yes Bisected commit-id:
Bug Depends on:    
Bug Blocks: 11167    

Description Rafael J. Wysocki 2009-02-23 14:32:21 UTC 
Subject    : [BUG] >2.6.27 dies at cardbus init on laptop
Submitter  : David Pflug <dtpflug@gmail.com>
Date       : 2009-02-17 19:28
References : http://marc.info/?l=linux-kernel&m=123489904218306&w=4

This entry is being used for tracking a regression from 2.6.27.  Please don't
close it until the problem is fixed in the mainline.
Comment 1 ykzhao 2009-02-23 17:37:59 UTC 
Hi, David
    Will you please attach the output of acpidump, lspci -vxxx?
    It will be great if you can attach the output of dmesg when the following boot option is added:
    a. pci=noacpi
    b. noapic
    c. acpi=off
    
    Will you please confirm whether the box can be booted normally when using the previous kernel(2.6.26.x)? If so, it will be a regression.
   Will you please use git-bisect to identify the commit which causes the regression?
   thanks.
Comment 2 David Pflug 2009-02-26 23:12:00 UTC 
*** Bug 12784 has been marked as a duplicate of this bug. ***
Comment 3 David Pflug 2009-02-26 23:44:24 UTC 
noapic freezes at the same place. That was one of the first things I tried.

pci=noacpi:
http://pastebin.com/f1ee0c148

acpi=off:
http://pastebin.com/fc2898c4

lspci -vxxx:
http://pastebin.com/f738f61dd

acpidump:
http://pastebin.com/f56c9ed77

2.6.26 and the git-bisect will have to wait til the morn. I'm about to fall
asleep on the keys.
Comment 4 David Pflug 2009-02-27 07:37:58 UTC 
2.6.26.8 boots fine. I'll get started on the git-bisect.
Comment 5 Rafael J. Wysocki 2009-03-02 14:42:44 UTC 
Handled-By : Zhao Yakui <yakui.zhao@intel.com>

It turns out to be a regression from 2.6.26.8 (ie. before 2.6.27).  Moving to the other list.
Comment 6 ykzhao 2009-03-02 17:42:48 UTC 
Hi, David
    Can the box be booted if the boot option of "noapic" is added?
    It will be great if you can use the git-bisect to identify the commit which causes the regression.
    Thanks.
Comment 7 David Pflug 2009-03-02 18:08:49 UTC 
noapic was one of the first things I tried. It doesn't help.

git bisect says 5b34653963de7a6d0d8c783527457d68fddc60fb

I have no idea how those changes could be causing a problem, but there you go. I can post the full bisect log if you like.
Comment 8 David Pflug 2009-03-19 22:18:47 UTC 
Is there anything else I can/should do for you?
Comment 9 David Pflug 2010-04-03 06:49:19 UTC 
I don't know what changed, and doing a git bisect to figure out what's fixed would probably be a waste of time.

But it works with 2.6.33. :) Thanks!
Comment 10 Florian Mickler 2010-04-16 06:48:27 UTC 
cool. so this bug can be closed.
Comment 11 Florian Mickler 2010-04-16 06:53:18 UTC 
cool. so this bug can be closed.  

if you are up to an git-bisect for the fix, it could be backported to the stable trees. but don't feel obligated to it.