follow up on "xy: cannot exec first read" problem

[Joseph Moran]

I have gotten a number of replies about the "xy: cannot exec first read"
problem during boot on my Sun-3/160 with Rev 2.7 PROM's, a Xylogics 451,
and 2 CDC Saber drives.  It appears that many people have had this problem
themselves and most had some ideas about what they could do to reduce the
likelyhood of this happening (e.g., shortening or changing cables,
cleaning filters, replacing the interface board on CDC).

However, I also got some messages from some folks at Sun and a Sun
customer that also was having this problem that had worked with someone
from Sun on it.  It appears that this is a PROM problem in Rev's 2.7 -
2.8.3.  This seems like the most logical explanation to me since we had
our CPU board upgraded to a 2.7 PROM shortly before we first started
noticing the problem.

Sun has now released a 2.8.4 PROM that supposedly fixes this problem.
Unfortunately, it appears that you need to get an entire new CPU board to
get the new PROM.  We went through the hassle of doing this and now have a
CPU board with a 2.8.4 PROM installed.  I haven't seen the "xy: cannot
exec first read" problem since we got the 2.8.4 PROM (but we haven't had
much time with the new PROM yet).

However, I have already found a new (more minor) problem with the 2.8.4
PROM.  When we try to boot off of a 1/4" data cartridge which is connect
to the system through an older SCSI-2 host adapter ("sc"), the monitor
prints the following message:

	si: peek failed

Fortunately, the tape boot still works after the (debug?) message is
printed.  I guess the Sun PROM saga continues...  In any case, it appears
that having a Rev 2.7 - 2.8.3 PROM with an "xy" SMD boot disk can result
in the "xy: cannot exec first read" problem.

Joseph Moran
Legato Systems Inc.
260 Sheridan Avenue
Palo Alto, CA  94306
(415) 329-7886
mojo at legato.com or {sun,uunet}!legato!mojo