Bug 128996

Summary: Sound blaster audigy 2 ZS Notebook CardBus sound card freeze kernel
Product: [openSUSE] SUSE LINUX 10.0 Reporter: Michael Pujos <pujos.michael>
Component: HotplugAssignee: Takashi Iwai <tiwai>
Status: VERIFIED FIXED QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P5 - None CC: tiwai
Version: Final   
Target Milestone: ---   
Hardware: i686   
OS: SuSE Linux 10.0   
Whiteboard:
Found By: Other Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Michael Pujos 2005-10-18 12:35:40 UTC 
This card when inserted in the CardBus slot is known to crash the kernel due 
to a pcmcia bug in the kernel not yet resolved. It is recognized as a SB 
Audigy 2 ZS PCI by hotplug since it use the same ids as this card. 
The problem is if the card is plugged at boot time, OS won't boot due to 
kernel crash. I had to blacklist the yenta_socket module to be able to boot 
with the card. 
 
Here's the relevant bug on the kernel's bugzilla: 
 
 
http://bugzilla.kernel.org/show_bug.cgi?id=5057
Comment 1 Takashi Iwai 2005-10-20 15:55:00 UTC 
The device is known to be buggy.

Christian, any workaround to disable the device?
It's not possible to blacklist on the driver side because it freezes immediately.
Comment 2 Christian Zoz 2005-10-20 19:18:51 UTC 
We can
- disable a device id in two ways: 
  + Add an alias similar to this in /lib/modules/.../modules.alias to 
    modprobe.conf, add a nonexistent module name to it and blacklist this
    module name
  + Add a hwcfg-vpid-* that matches the device with STARTMODE=off
- disble a device location on a machine
  + Add hwcfg-bus-* with STARTMODE=off

I will read http://bugzilla.kernel.org/show_bug.cgi?id=5057 tomorrow.
Comment 3 Takashi Iwai 2005-10-25 13:42:23 UTC 
The problem is that both working and non-working boards have the same PCI ID.
So, we can't add it blindly.
Comment 4 Christian Zoz 2005-10-28 09:40:01 UTC 
Do they have also the same revision?

If we cannot differ working from non working devices, all we could do is to work with a special script hooked in hwup. This script could check if the system crashed and avoid further loading of that module. 
Is it worth the effort?

We can also write SDB articles for such kind of problems.
Comment 5 Jaroslav Kysela 2005-11-01 11:43:52 UTC 
Note that http://bugzilla.kernel.org/show_bug.cgi?id=5057 contains a driver fix now (at least the machine does not hang).
Comment 6 Takashi Iwai 2006-01-19 14:44:19 UTC 
The crash is at least fixed in the latest kernel.
Comment 7 Ihno Krumreich 2007-06-04 16:37:58 UTC 
Closed.