hi again... poked into it a little further...
* Sebastian Henschel <[EMAIL PROTECTED]> [2004-05-19 15:25 +0200]: > > * [EMAIL PROTECTED] [2004-05-19 14:50 +0200]: > > > On Wed, May 19, 2004 at 02:20:43PM +0200, Sebastian Henschel wrote: > > > howdy folks... > > > > > > SOUND: > > > > > > when trying the old dmasound module i get this when loading: > > > > > > FATAL: Error inserting dmasound_pmac > > > (/lib/modules/2.6.6/kernel/sound/dmasound_pmac.ko): No such device > > > dmasound_pmac: couldn't find a Codec we can handle > > > > > > ALSA can be loaded, but is mute. > > > - trying to play an .au file gives "aplay: set_params:880: Unable to > > > install hw params: ..." > > > - .wav and .mp3 play fine, but i cannot hear a thing. > > > - the alsamixer shows card/chip to be "PowerMac AWACS", although it is > > > "PowerMac Snapper" on my 5,2. there are also fewer items to control: > > > only master, headphone detection, mic boost, pc speak, auto-mute; > > > plus passive items: cd, line, mic. On my 5,2 i can control: > > > master, headphone, headphone detection, bass, treble, pcm, pcm1, pc > > > speak, auto-mute, drc, drc rang, monitor mix. > > > > Same happened on my 17"-PB (1st generation) > > Work-Around: > > dmsaound_pmac works fine when not compiling module but static into kernel. > > hm, not so here. compiling statically does not help. i do not even see a > note about it in dmesg. what should be seen there? found that note in the meantime, telling me the same as the module... > also "cat /dev/[audio|dsp|mixer]" yields: no such device; aumix does not > work, too. > are there any OSS options needed? might the alsa modules interfere? switching back to ALSA... i looked into sound/ppc/pmac.c and commented out the lines 926+932, which leads to the situation that snd-powermac starts in "snapper"-mode. now it seems to work! the question is, why "device_is_compatible" does not return successfully for "snapper". so far, i found out, that this function is located in arch/ppc/syslib/prom.c. what is responsible for filling the property "compatible"? looking back at my cpufreq problem, do you think that the parsing of the openfirmware device tree somewhat fails or misbehaves? i feel like one important item has changed unexpectedly in the OF and subsequent property calls result in failures. i attached the file tree of /proc/device-tree if that could be useful. cheers, sebastian -- ::: .O. ::: ..O ::: OOO ::: lynx -source http://www.kodeaffe.de/shensche.pub | gpg --import
tree.bz2
Description: Binary data
signature.asc
Description: Digital signature
