On 09/11/2007 09:51 PM, Takashi Iwai wrote:

...

Does this mean it might be better to delay the impending 1.0.15 for 1 more week? You're sort of vital to this ALSA thingy...

I guess the next rc will be delayed anyway since Jaroslav is already in the conference.

Mmm. Good :-)

...

Also, although I don't care deeply, I'd also sort of like to get the new Sound Galaxy driver in there and while all the AZT2316 cards are fully functional, AZT1605 is giving me a little grief. The configuration part of its EEPROM is mapped out and the driver is working well, but I haven't yet found a way to know which one of the two I'm looking at runtime and I need to know that to know what configuration register layout I need to use.

The simplest way might in fact be to introduce sound/isa/galaxy/ and provide seperate snd-azt1605 and snd-azt2316 drivers that just share a lot of code through a common galaxy.c (and then just move over snd-azt2320 to there as well ofcourse).

Maybe it's the only way if there is really no difference in the signature or the beginning of EEPROM...

I haven't yet looked at the storing of volumes in the EEPROM (which azt1605 does, and azt2316 not). Requires some reverse engineering again, but that might provide a method.

When I find something, I'll also need to base my nicely simple and direct setup in the ISA match() method around a bit, since I'll certainly not know at _that_ stage. AZT1605 eeprom/config-register layout difference and it in fact supporting different IRQs for MPU-401 (and SB, but that's not used in these drivers)...

For now I've done a seperate azt1605 though and I sort of expect I'm going to try for just splitting the common code of into a galaxy.c and building the two modules from that. I'll not wait too long with submitting the next version, so you'll still have time to slap me around when you really do desperately want a merged driver (azt1605 is really utterly ancient by now though, and an even older galaxy I have will also need differences).

Rene.