SiS/Trident 4DWave sound driver oops
I have a very similar OOPS that I reported to linux-kernel just last
week: see full details at
http://uwsg.iu.edu/hypermail/linux/kernel/0110.1/1690.html
Quoted 3 lines
This kernel oops is totally reproducible (on every occasion) in 2.4.9, 2.4.10, and 2.4.12. I have not tried earlier kernels in the 2.4 series
I also have a totally reproduceable OOPS on 2.4.9 through 2.4.12. Kernel 2.4.7 works just fine. If the soundcard is compiled in to the kernel, it dies while booting. If it is compiled as a module, it dies when attempting to load the module (typically when artsd starts).
Quoted 2 lines
The machine in question is a Clevo lp200t SiS630S "all-in-one" machine.
My machine is an ECS K7AMA: ALi 1645 northbridge, ALi Magic 1535D+ southbridge. Sound is onboard the southbridge. I ran ksymoops on my OOPS report, so you can see the trace at the message archive.
Mike Robbins
[email protected]
(Please also cc your reply to me.)
Quoted 4 lines · Michael F. Robbins
I also have a totally reproduceable OOPS on 2.4.9 through 2.4.12. Kernel 2.4.7 works just fine. If the soundcard is compiled in to the kernel, it dies while booting. If it is compiled as a module, it dies when attempting to load the module (typically when artsd starts).
I just tried 2.4.7, and while the module loads, the ac97_codec gives an "unknown id" response, and I get no sound out of the device whatsoever. Any writes to /dev/dsp just hang (not the kernel) and I get nothing out of the device.
I might try compiling the 2.4.7 trident module without symbol versions, and a 2.4.9 kernel the same, and try forcibly inmod'ing the 2.4.7 trident module on 2.4.9 and see what happens.
AMC Enterprises P/L - Stuart Young First Floor - Network and Systems Admin 3 Chesterville Rd - [email protected] Cheltenham Vic 3192 - Ph: (03) 9584-2700
http://www.amc.com.au/ - Fax: (03) 9584-2755
More on this kernel oops with the SiS/Trident 4DWave driver.
Went and compiled 2.4.9 and 2.4.7 without module symbol versions, and tried the trident module from 2.4.7 in 2.4.9. Exactly the same oops that happened with the 2.4.9 module, so it may not explicitly be the driver, but something it relies on, like the ac97_codec.
When I used insmod, it only complained about the kernel revisions, and nothing about symbols, so I'm guessing the interface didn't change much, if at all, so I used the -f option to force a load, producing an oops.
Interestingly enough, with 2.4.9. and 2.4.9 loading the 2.4.7 trident module, lsmod reports that while the trident driver is using ac97_codec, ac97_codec has 0 modules using it.
eg: (hand typed, but accurate)
trident 24448 1 (initializing) soundcore 3812 1 [trident] ac97_codec 8864 0 [trident]
Hope this helps.
AMC Enterprises P/L - Stuart Young First Floor - Network and Systems Admin 3 Chesterville Rd - [email protected] Cheltenham Vic 3192 - Ph: (03) 9584-2700
http://www.amc.com.au/ - Fax: (03) 9584-2755