The Asus A7V motherboard has a built in sound controller (AC'97). When using sndconfig to try to setup this controller, my machine locked completely. I did not respond to network or terminal actions. Another problem related to this is that after a hard reboot the initscripts still attempt to load the module and lock the machine. Since the audio controller is part of the Via KT133 chipset this bug is probably relevant to any KT133 based motherboard. From /var/log/messages: Aug 7 11:07:54 speak modprobe: /lib/modules/2.2.16-17/misc/sb.o: Aug 7 11:07:55 speak modprobe: Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters Aug 7 11:07:55 speak rc.sysinit: Loading sound module (via82cxxx) failed Aug 7 11:07:55 speak modprobe: init_module: Device or resource busy Aug 7 11:07:55 speak modprobe: /lib/modules/2.2.16-17/misc/sb.o: insmod /lib/modules/2.2.16-17/misc/sb.o failed Aug 7 11:07:55 speak modprobe: /lib/modules/2.2.16-17/misc/sb.o: insmod via82cxxx failed
That doesnt seem rational since it actually failed to load the modules in question. Unless maybe the failure path is buggy. This is a sndconfig bug. The sequence that should work is modprobe sb support=1 modprobe via82cxxx (_audio with 2.4) let me know what that does
The modprobe of the sb module seemed ok, but the system locked on the modprobe via82cxxx. There was no kernel oops or anything. Also, the message segment I attached originally was from a boot sequence where I believe I had to kill rc.sysinit (ctrl-c) so that it wouldn't modprobe the via82cxxx because my machine was hanging when the module was being loaded (had to hard reboot a couple times).
OK, that does sound like a driver problem. Bouncing back to the kernel. :)
We need to not load this driver by default then. Any attempt to fix it will take a fair bit of time and mean backporting the native mode driver from 2.4.
OK, kudzu will have to be changed not to load this driver by default. Back to you, Bill...
Fixed in current (0.67) kudzu; we don't configure the via module by default. Users can still configure it in sndconfig if they want to.
I'm changing the component to sndconfig because that is just as large a problem as kudzu configuring this module.
Unfortunately, that module works for other people. :(
What does `lspci -vn' say?
Ok. Sounds good. The Asus A7V motherboard should probably be listed in the hardware compatiblity lists as tier 2 with a note about the onboard sound. This is a thunderbird motherboard, and since there was a note about the athelon mobos with the 6.2 compat lists, it would go under there.
Created attachment 2314 [details] output from lspci -vn
The MSI K7TPro mainboard, also based on the VIA KT133, has exactly the same problem. I found the problem with RC1 and it also occurs in RH 6.2 with the (update package) 2.2.16 kernel.
lspci -vn of that, please?
Created attachment 3338 [details] MSI K7TPro (MS-6330) lspci -vn
sndconfig now no longer writes modules.conf until it's sure it's succeded (as of 0.60-1); the driver itself is fixed in the currently-in-rawhide 2.4.0 kernel.