Bug 6340

Summary:	Kernel Oops under sil_init() - Toshiba A100
Product:	SCSI Drivers	Reporter:	Carsten Niehaus (cniehaus)
Component:	Other	Assignee:	acpi_other
Status:	CLOSED PATCH_ALREADY_AVAILABLE
Severity:	normal	CC:	acpi-bugzilla
Priority:	P2
Hardware:	i386
OS:	Linux
Kernel Version:	2.6.16	Subsystem:
Regression:	---	Bisected commit-id:
Attachments:	Foto of the oops The output of 'lspci -vvvn' Output of dmidecode

Description Carsten Niehaus 2006-04-06 12:25:38 UTC

Most recent kernel where this bug did not occur:
Distribution: Ubuntu or Knoppix
Hardware Environment: Toshiba Satellite a100-533
Software Environment: Linux 2.6.15 and 2.6.16
Problem Description: Booting the kernel results in the oops I took a foto of

Steps to reproduce: Boot any kernel with acpi enabled

Comment 1 Carsten Niehaus 2006-04-06 12:32:55 UTC

Created attachment 7790 [details]
Foto of the oops

Comment 2 Carsten Niehaus 2006-04-06 12:42:39 UTC

Sorry, I cannot provide you lspci -vvv no as I don't have physical access to 
the PC now and didn't save the info on my usbstick as I thought I did. 

The notebook is a Toshiba Satellite a100-533, the product page is 

http://de.computers.toshiba-europe.com/cgi-bin/ToshibaCSG/product_page.jsp?service=DE&PRODUCT_ID=113743 

It is a Intel Celeron M Prozessor 360 with 512 MB DDR2-RAM. Currently I could 
reproduce the crash on Ubuntu Dapper Flight 5 and 6 (both running 
2.5.15+patches) and Knoppix 5 (running 2.6.16). The foto of the oops was taken 
from Knoppix. Booting in failsafe mode works because failsafe implies that 
acpi=off.

Please tell me what kind of information you need so that I can help you to fix 
this problem. I really want to support you fixing this bug...

Comment 3 Carsten Niehaus 2006-04-08 00:28:03 UTC

Created attachment 7810 [details]
The output of 'lspci -vvvn'

I can also give you the output with '-n' if you need it.

Comment 4 Carsten Niehaus 2006-04-08 00:29:45 UTC

Created attachment 7811 [details]
Output of dmidecode

Comment 5 Shaohua 2006-04-10 19:07:03 UTC

This sounds like a sata bug. 'failsafe' might also enable other workaround. so 
please try boot option 'acpi=off', if it works please change the bug catalogy 
to 'SCSI drivers'

Comment 6 Carsten Niehaus 2006-04-11 00:57:27 UTC

I tried to only pass acpi=off to knoppix5, doesn't work. In Ubuntu Dapper 
"acpi=off nolapic noapic" works. The notebook boots (other things don't work, 
but at least no oops). Well, this is of course conflicting so I don't know if 
you want me to change the category.

If you are interested, I also filed a bug against Ubuntu itself:

https://launchpad.net/distros/ubuntu/+source/kernel-image-2.6.15-18-386-di/+bug/38476

Then I also tested DSL (Damn Small Linux) and get the same oops. So on four 
different distros I get the same oops.

Comment 7 Len Brown 2006-04-19 18:54:40 UTC

> I tried to only pass acpi=off to knoppix5, doesn't work

doesn't work how?  same failure, or different failure?

Has any ACPI-enabled linux ever booted on this box,
or has it failed with everythig you have tried?

one thing you might investigate is to set the BIOS
to present the SATA drive in legacy mode to the OS --
though I would think sata_sil() should be able to handle
this in native mode.

Comment 8 Carsten Niehaus 2006-05-27 11:33:35 UTC

Ok, I updated to dapper release-candidate. That is, if I am not mistaken, 
2.6.15.16 with an updated (backported from 2.6.16) acpi-table. This problem is 
now fixed in Ubuntu. I expect it to be fixed in 2.6.16 as well.