In message <tkr8zrud2rx.fsf@parity-error.sm.sony.co.jp>enami tsugutomo writes
>Jonathan Stone <jonathan@DSG.Stanford.EDU> writes:
>
>I guess wi_intr should ignore interupts while !RUNNING.

Wonderful!  I added a check for sc->sc_enabled == 0 (cf  dp8390_intr()),

and now the wi card is *much* happier.  I had occasionally seen
wi_seek() timeouts on insertions before i hacked the delay() code;
those seem to be fixed also.  Thank you. I'll commit a fix and request
a pullup after I get some sleep--unless someoen beats me to it.


I'm tempted to submit a 1.5 pullup to change the delay() calls to
pccb_delay (basicaly a copy of pcic_delay(), also.  If the tsleep()
with no locks in the pcic code isnt a bug, it wont be a bug for pccbb,
either.  I just dont know if its safe for -current ;-).  Jason?