Bug 10813 - 2.6.22 boot hang - 2.6.19 worked - Gericom notebook
Summary: 2.6.22 boot hang - 2.6.19 worked - Gericom notebook
Status: CLOSED CODE_FIX
Alias: None
Product: ACPI
Classification: Unclassified
Component: Other (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: Zhang Rui
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-05-28 05:54 UTC by paolo
Modified: 2008-10-16 22:19 UTC (History)
2 users (show)

See Also:
Kernel Version: 2.6.22.*
Subsystem:
Regression: Yes
Bisected commit-id:


Attachments
acpidump+dmesg(acpi=off)+lscpi of Gericom notebook (184.26 KB, application/gzip)
2008-06-06 14:38 UTC, paolo
Details
full acpidump of Toshiba nb, + screenshot (190.00 KB, application/x-tar)
2008-06-15 11:38 UTC, paolo
Details

Description paolo 2008-05-28 05:54:02 UTC
Latest working kernel version: 2.6.19.7
Earliest failing kernel version: 2.6.22.*
Distribution: Debian
Hardware Environment: notebook
Software Environment:
Problem Description: hangs early on boot

[perhaps it belongs to acpi_bios, nevertheless it might complement/relate to
 #10811]

I failed to make 2.6.22.* boot on at least a couple of notebook (I've made more trials on 2). Booting hangs just after a few 'ACPI ... ' lines are printed, and after (on 2.6.22.19) I get the msg from acpi stack (circa)
'... try with "acpi_osi=!Linux" ...'
which I did but with no success.
I know it's an aged kernel, nevertheless since I'm having troubles with ACPI
in latest version (see #10811), I hope this bit offer some hunch to devels.

I'll try to provide more info on specific request - these are 'travelling' machines and to get such infos I'd need them phisically in my hands.
DSDT and such still easy, though, of course, if needed.


thanks


Steps to reproduce:
Comment 1 ykzhao 2008-05-28 20:54:00 UTC
Will you please boot the system with "acpi=off" and attach the output of dmesg, lspci -vxxx and acpidump?

Will you please capture the screen picture when the system hangs?
thanks.
Comment 2 Len Brown 2008-05-29 14:07:13 UTC
since acpi_osi=!Linux had no effect, I've removed that from the summary.
Can you also provide the full output from dmesg -s64000 from
the latest kernel that did boot -- apparently 2.6.19?
Comment 3 paolo 2008-06-06 14:38:24 UTC
Created attachment 16419 [details]
acpidump+dmesg(acpi=off)+lscpi  of  Gericom notebook

Here's (hopely) the required infos - sorry, didn't have 2.6.19 still aroud, so dumped dmesg from latest Debian's 2.6.18, plus dumps from 2.6.22.19, 2.6.25.4,
acpidump and screenshot of 2.6.22.19 hanging on boot.
Comment 4 Len Brown 2008-06-13 18:23:31 UTC
The last message printed before the hang was:

Initialized 17/17 Regions 10/10 Fields 26/26 Buffers 18/19 Packages (530 nodes)
Initializing Device/Processor/Thermal objects by executing _INI methods:

Please try booting a kernel with CONFIG_ACPI_THERMAL=n
if that doesn't work, try one with CONFIG_ACPI_PROCESSOR=n

Are you sure that 2.6.19 booted properly in ACPI mode?
The dmesg for 2.6.18 was with "acpi=off" (as were the 2.6.22
and 2.6.25 dmesg)

Does 2.6.18 boot properly w/o "acpi=off"?  how about
the FC9 livecd (2.6.25)?
Comment 5 paolo 2008-06-15 11:38:50 UTC
Created attachment 16493 [details]
full acpidump of Toshiba nb, + screenshot 

I could try 2.6.22.19 on the Toshiba, and take a screenshot of last screen
when boot hangs.
Attached also output from acpidump, in case the other referenced to in prev. comment was incomplete.
Comment 6 paolo 2008-06-15 11:47:10 UTC
(In reply to comment #4)
> Please try booting a kernel with CONFIG_ACPI_THERMAL=n
> if that doesn't work, try one with CONFIG_ACPI_PROCESSOR=n

ok, I'll try asap next round
 
> Are you sure that 2.6.19 booted properly in ACPI mode?

yes

> The dmesg for 2.6.18 was with "acpi=off" (as were the 2.6.22
> and 2.6.25 dmesg)
> Does 2.6.18 boot properly w/o "acpi=off"?  how about

yes, I set acpi=off a provided the log as requested. But .18 works fine, 
# cat /proc/acpi/info 
version:                 20060707
 
> the FC9 livecd (2.6.25)?

hm, I'll try to grab just their kernel, but I'd rather test vanilla kernel compiled here, unless there are specific reasons to use (pactched?) FC's 
kernels.
 
Comment 7 Zhang Rui 2008-08-14 00:16:31 UTC
Does the problem still exist in the latest kernel, say 2.6.26?
Comment 8 paolo 2008-08-14 02:19:59 UTC
Sorry, forgot to update...
see http://bugzilla.kernel.org/show_bug.cgi?id=10811#24 :
2.6.26.1 boots on both nb (still another one to test, a 
PackardBell/NEC,if/when possible); but while it seems
fine on the Toshiba, reading bat status hangs on the
Gericom nb.

But the hangs-on-boot problem seems gone, at least on 
these 2 nb, and at least for k as of 2.6.26.1.
Comment 9 Zhang Rui 2008-09-09 19:28:54 UTC
so the problem is fixed in the latest kernel release.

> eading bat status hangs on the Gericom nb
this is another problem. please file a new bug report.

Note You need to log in before you can comment on or make changes to this bug.