Discussion
Discussion
- Subject: Discussion
- From: Andy <email@hidden>
- Date: Fri, 20 Sep 2002 21:43:41 +0100
I was talking to another developer today and it turns out that some of
the problems we see while writing drivers for core-midi are very
similar.
For example (I'm being deliberately vague in the hope that discussion
will follow) :
- USB/MIDI devices will suddenly and mysteriously go offline and hot
re-plugging doesn't always bring them back.
(this seems worse in 10.2 than 10.1)
- Sometimes, even restarting the MIDI server won't bring USB devices
back to life, only rebooting the computer will do it.
- During MIDI server initialization, USB can lock up. I had a very real
conflict with the MOTU driver. ReadPipe never returned and
ReadPipeAsync timed out. I solved it by doing my device initialization
in a separate thread and delaying that thread to allow the MOTU device
to complete its initialization before I started mine (a hack I'm not
really proud of) !
Is it possible that there are race conditions inside core-midi and usb
iokit that are being caused by all devices being simultaneously
initialized ?
Should core-midi initialization be altered to stagger the notifications
to drivers so that were not all trying to grab usb at the same time?
Are some drivers sleeping the MIDI server main and IO threads
[sleep() usleep() etc] ?
Do we actually need/want a discussion or am I being paranoid ?
It does appear that something may be awry somewhere in the bowels of
the MIDI server. Or that driver writers, (like me) are not quite
aware of the whole story.
_______________________________________________
coreaudio-api mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/coreaudio-api
Do not post admin requests to the list. They will be ignored.