Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 40399 - iBook G4/2.6.1: i2c_keywest hangs if therm_adt7467 is loaded
Summary: iBook G4/2.6.1: i2c_keywest hangs if therm_adt7467 is loaded
Status: RESOLVED FIXED
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: Current packages (show other bugs)
Hardware: PPC Linux
: High critical (vote)
Assignee: PPC Porters
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2004-02-04 12:15 UTC by Amanda Yilmaz
Modified: 2006-02-04 06:04 UTC (History)
0 users

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Amanda Yilmaz 2004-02-04 12:15:46 UTC
I'm running kernel 2.6.1-benh1 (built from ppc-development-sources) on a 12" iBook G4. My kernel configuration has the iBook G4 thermal management option (CONFIG_THERM_IBOOKG4, therm_adt7467) and all the I2C and sound options, both OSS and ALSA, enabled as modules to the extent possible.

With this configuration, attempting to load the i2c_keywest module hangs the computer if the therm_adt7467 module was loaded first.  This wreaks particular havoc with the Mac sound modules, OSS or ALSA, since all of them depend on the i2c_keywest module and will attempt to load it if necessary.  If therm_adt7467 was configured to be autoloaded (by listing it in /etc/modules.autoload.d/kernel-2.6), and the alsasound script was added to the default runlevel (via 'rc-update add alsasound default'), the computer will hang during boot, since therm_adt7467 will have been loaded first.  The same thing happens if you're attempting to use OSS sound and dmasound_pmac is listed after therm_adt7467 in /etc/modules.autoload.d/kernel-2.6.

It is easily possible to work around this for OSS by listing therm_adt7467 after dmasound_pmac in /etc/modules.autoload.d/kernel-2.6.  But for ALSA, since its modules are normally loaded later in the boot process via the alsasound init script, it is necessary to *not* list therm_adt7467 in /etc/modules.autoload.d/kernel-2.6.  Instead, one has to add a "modprobe therm_adt7467" line to /etc/conf.d/local.start, which will be executed after everything else.  This works, but it's clunky.

I don't know what would happen if any of these items were compiled into the kernel rather than as modules. I suspect the same thing would happen, only without the benefit of being able to change everything to load in the proper order.  A kernel which has therm_adt7467 compiled in but has the I2C stuff compiled as modules might not be able to load sound modules at all without hanging.

I did a Google search in an attempt to find more information about this, and I found one post to the debian-powerpc list that looks like it's another example of the same problem:

http://lists.debian.org/debian-powerpc/2004/debian-powerpc-200401/mail3.html

One other observation that may or may not be related to this: although i2c_keywest is loaded by the dmasound_pmac module, this doesn't appear as a dependency in the 'lsmod' list (i2c_keywest is not listed as being used by dmasound_pmac, even though it actually is).
Comment 1 Amanda Yilmaz 2004-02-04 12:32:37 UTC
I got the URL wrong for that related post on the debian-powerpc list... the correct URL is:

http://lists.debian.org/debian-powerpc/2004/debian-powerpc-200401/msg00047.html

Sorry about that!
Comment 2 Bartosch Pixa (RETIRED) gentoo-dev 2004-02-04 12:47:45 UTC
alsasound should be added to the boot runlevel, not default
Comment 3 Amanda Yilmaz 2004-02-04 14:02:32 UTC
Oops, my mistake.  This is definitely not my day.  Nevertheless, I'm more concerned about the potential bug in the kernel code.
Comment 4 Bartosch Pixa (RETIRED) gentoo-dev 2004-02-04 14:06:37 UTC
try is ppc-development-sources 2.6.2-r1 fixes your problem
Comment 5 Amanda Yilmaz 2004-02-04 16:54:00 UTC
Yes, 2.6.2-r1 fixes the problem... and as a plus, it resolves a couple of other module problems I was having.  Thanks.  (ALSA still isn't emitting sound, but I'm assuming that's a problem with my own configuration.)

I am a Gentoo newbie, and this is the very first bug I have ever reported using Bugzilla.  As a matter of Bugzilla protocol, is it now my responsibility to resolve the bug as FIXED?  Or is it up to the developer to do that?  (I don't want to step on anyone's toes!)
Comment 6 Bartosch Pixa (RETIRED) gentoo-dev 2004-02-04 17:13:22 UTC
doesn't matter either way, though if there were others who are in CC on this bug who have this problem too leave it open toll there are more success reports.

closing