Thanks for the tips, Ash. I've found the Iwai document. Now there's some light reading. I take it that it's still relevant, even though it hasn't been updated since 2005.
I haven't actually read it since 2005 I'm afraid :) but I don't think anything's changed that drastically. The fundamentals are all there.
Yes, the cs46xx driver. That chip is in a lot of machines (at least ten ThinkPad models), and with this bug nobody running Alsa likes Linux very much.
Yeah, it's a pretty poor show that this bug has existed for so long. Let's hope you can zap it!
Is there some (handy) way to dump the registers from DOS or Windows as a comparison?
Well, that depends. I'm not familiar with the hardware. There's a Windows program called PCItree. It allows you to view the current status of the ioport of a chosen device. Now, that's pretty handy if the regs for the card are mapped directly onto the io port. It's usually the case though, I think, that the registers aren't just mapped on and you'll need to query the card. If they're not mapped onto the io port (and forgive me again - not familiar with this chip) then you could use a debugger such as Numega SoftIce to manually issue queries to find out the contents of the registers. Sounds like a happy fun time. Writing the code to do a reg dump might be fun too. You may also want to use a debugger/disassembler with Windows (or other) drivers in order to find out what those drivers do to the card on resume.
For development: Since the drivers go into the kernel, I should be able to find a page on getting a development environment set up... right? I'm
You don't need to mess with the kernel as ALSA is available as a package. All you really need is the latest copy of alsa-driver and a compiler and you're good to go.
Ash
=