Summary
OSSServer is a fancy sound server that can create /dev/dsp like device-files so one can easily use any OSS-only Applications (Skype, TeamSpeak, ET, Quake III, ...) simultaneously.
Rationale
We want to be able to use more than one OSS-only application without hardware-mixing sound cards.
Use cases
- GuyA wants to talk to his mom while listening to Deep Purple.
GuyB wants to be the l33t3st gamer on earth so he needs TeamSpeak and Wine Sound simultaneously
Implementation
I dont know if this is possible, if it is, we also need a small script that sees which fake-dsps are in use already and returns a free one or requests a new one from OSSServer
This is definitely possible. oss2jack uses a kernel driver (fusd) to create OSS devices that are redirected to the jack audio server. oss2jack could be modified to work with alsa (and thus dmix) rather than using the jack audio server.
Then lets get this up for Google SoC :-)!! - MartinSchmeisser
I tried oss2jack and it works quite good. It requires fusd, a userspace device-creation-library which works with our 2.6.15 kernel with Kors hack. I just sent a patch to Kor that makes multiple instances of oss2jack work at the same time. I have no idea how the alsa api works but i will have a look whether its possible to hack oss2alsa together. It'll be hard to use this concurrently with esd: jack doesn't play well with dmix. However if we can get oss2alsa going we dont need to care about that. Right now i have esd-oss run via oss2jack over the jack server. 