In article <98AE79FC-0B84-11D9-81C5-000393DB5366@cs.utk.edu>, Keith Moore wrote:
> 
> maybe iMic's work differently than they used to?

Resurrecting a months-old thread:

With my recently purchased iMic, my 1.6.2_STABLE (2.0 says something
similar) kernel reports it as:

uaudio0 at uhub0 port 2 configuration 1 interface 0: Griffin Technology, Inc iMic USB audio system, rev 1.10/0.06, addr 2
uaudio0: ignored input endpoint of type adaptive
uaudio0: ignored input endpoint of type adaptive
uaudio0: ignored input endpoint of type adaptive
uaudio0: audio rev 1.00
audio1 at uaudio0: full duplex, independent

Perhaps this is a different revision than the version that nathanw@
(among others) reported as working great back in 2003:

  http://article.gmane.org/gmane.os.netbsd.ports.macppc/1798

Or perhaps they were concerned only with playback, and not with
recording.

The comment above the "adaptive" printf in uaudio.c says:

  /* We can't handle endpoints that need a sync pipe yet. */

Though looking at -current, it seems that kent@ is working on support
for this, protected by #ifdef UAUDIO_MULTIPLE_ENDPOINTS.  According to
the rlog,

  "add code to recognize sync-pipe audio stream.
   The code is not enabled yet."

I don't -currently have a -current machine set up to try out the
UAUDIO_MULTIPLE_ENDPOINTS stuff.  Maybe in a few weeks.  :)

  Alan