30 Aug
2001
30 Aug
'01
10:13 a.m.
Well you figure for EOL hardware, Sun has no reason to
fix the
problem. It very well may have been a newbie engineer mistake that wasn't
caught in time. The end result is that it's not worth it to them to fix
the error. I would be surprised if they didn't eventually fix it for new
design hardware.
Just for the record, when I found out about the NVRAM in an SS1 (which I
did when my new kernel left the NVRAM writable and over wrote it leaving my
SPARCStation dead) I flamed Andy Bechtolsheim (the guy who designed it,
founder of Sun, etc, etc) pretty toastily (I was workign at Sun at the
time). And his response was simply that it didn't merit being in EEPROM or
even PROM. You see the boot firmware wanted to store some bits in NVRAM and
as long as you have it, and the firmware guys don't need all of it, why go
to the trouble of adding yet another chip, decoder, etc to put more
non-volatile memory on the mainboard? So here we are.
--Chuck