> Date: Sun, 02 Mar 1997 12:02:44 -0500
> From: "John F. Woods" <jfw@jfwhome.funhouse.com>

[ serial driver stops processing input ...]

> I don't see that.  Sounds like a bug in the serial driver; maybe the receive
> interrupt clears out some broken state in the transmit path allowing it to
> recover?

Here is a follow up on this problem.  It turns out that when I moved
the clock interrupt handler from locore.s to clock.c, the compiled
code was no longer doing atomic updates to the interrupt register
when it "pulses the clock latch" so we had a race.  The result was
that if a ZS interrupt came in while the clock interrupt was doing
the "pulse" thing, one could lose the "soft" interrupt request.
I changed the code to use Leo's <m68k/asm_single.h> macros.

The fixes should appear in tomorrow's SUP.

Gordon