===================================================================
RCS file: /MkLinux/floppy-darwin/NOTES,v
retrieving revision 1.3
retrieving revision 1.4
diff -u -r1.3 -r1.4
--- floppy-darwin/NOTES	2002/03/21 23:28:05	1.3
+++ floppy-darwin/NOTES	2002/03/22 07:35:42	1.4
@@ -1,16 +1,24 @@
 NOTES:
 
-Debugged enough to verify that the semaphore timeouts were working, 
-since I just documented that, and I'd need to quickly remove that
-from the docs had it been broken.  So that works.  Now quickly recording
-what's broken.
-
-1.  Somewhere, we're getting stuck in a loop waiting 4000 ms for some event
-    that never happens.  Find and fix.
-2.  On unload, get a 0x400 (instruction access).  It's actually a callback,
-    probably caused by a timeout that wasn't cancelled, probably waiting on
-    a lock, semaphore, or any number of waits.  This is gross.  Need to make
-    sure all timer events are cancelled cleanly in the unload code.  :-)
+Okay, here's what's still broken:
+
+1.  On unload, get a 0x400 (instruction access) if we unload while
+doing a semaphore_timedwait.  Somehow, I need to cancel any thread
+timers that might be outstanding against any user threads that happen
+to be in the driver's code, as well as any thread timers outstanding
+against the polling thread.  I have NO idea how....  Hopefully the
+driver won't be wedging anymore, though, which would mean that this
+wouldn't happen in the future....  :-|
+
+Ah.  We may need to tear down the driver in a sane fashion that prevents
+anything from entering the driver while it is being torn down.  Still
+pondering that one.  Maybe the I/O Kit takes care fo that for us.  I
+hope so.  That would be really ugly.
+
+2.  Need to verify fix for the hangs.
+
+3.  Missing read/write support (both synchronous and asynchronous).
+
 
 
 WARNINGS (that might be mistakes):