During sound configuration problems (which turned out to be usual error between keyboard and chair) I updated sound code to the latest one from FreeBSD RELENG_6. It didn't help me (obviously), but I realised it should help a lot of hda users and also contains the new envy24 driver, so I continued work on it.
Only ich is tested briefly by me and hda is tested by josepht from #dragonflybsd. So, I'd say that there is good chance it doesn't blow up your system ;). More tests (especially hda and envy24) are very welcome. Don't pay attention to tags etc, this cleanup will follow before commit. There were two problems I'm not sure though that I handled correctly. * New experimental polling mode using simple callout_* mechanism uses callout_drain() we don't have in df. I used callout_stop() instead, but as I'm not familiar with this mechanism, I'm not sure at all. Polling isn't enabled by default, so it shouldn't bite testers anyway. * AFAICS we don't have uncacheable DMA support, so PCI snooping is used. I'm not sure I handled all related stuff correctly (grep for BUS_DMA_NOCACHE from patch). The rest of it should be quite straightforward. http://hasso.linux.ee/stuff/patches/dragonfly/sound-latest-fbsd6.patch.txt -- Hasso Tepper
