I wrote

> pk found a race condition and committed a patch that should fix it.
> I haven't reported back on it, though, because I can't reproduce it
> for site-specific reasons.  ([brief explanation])

I got some downtime on the SS5 and test-booted a couple of kernels.  I
am happy to report that kbd.c 1.28 (ie, the version with pk's fix)
cures the problem as far as I can tell.  (I dropped my "fake a reset if
we don't get one" code into 1.28 and didn't see it trip, on the very
same hardware that does exhibit the bug with my tweak in 1.27.)

And thanks to the person (you know who you are, I don't want to ID you
without your consent) who offered a guinea-pig machine, though there's
now no need.

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B