To anyone interested:

I have a patch for /sys/arch/sparc/dev/esp.c which will prove helpful to
those on the slow machines (SS1/1+ class).  The premise is that if
reselection somehow fails, it turns off reselection AND sync negotiation,
since (for me, anyway), if either one is enabled (despite reassurances from
PK :-) (Hi and thanks for pointing me to the code, Paul!)), I get the

	esp0: RESELECT: 7 bytes in FIFO!

messages.  So I have a small loop that goes in and turns off both attributes
and prints a message to that effect.  Undoubtedly I have probably chosen the
wrong way to go about this, but it works for me.  If anyone else is interes-
ted, please let me know.  I'm considering sending it in as a PR but probably
won't get to do it before Thursday AM US/Pacific time.

It *might* work on an ELC; it DOES work on a SS1+; I need to test it on my
SLC at home.

[pk, can I run it by you before I PR it?  It's just a loop which goes
thru all the targets on the ESP and clears T_NEGOTIATE and sets T_RSELECTOFF]

				--*greywolf;