On Thu, 26 Feb 2009, solarg wrote:
> Randy Fishel wrote:
> > The Ultra 20 has a header allowing a serial port to be connected and
> > provide extra debug output. It would be far more useful to see output from
> > the serial port before believing that a BIOS update will solve the problem.
> >
> >
> "header allowing a serial port..." What does it mean? do you have some
> pictures to illustrate that?
What it means is that there really is a serial port on the U20, but
to use it requires that the owner do a little bit of work themselves.
> do you mean that nobody at sun has never tested suspend-resume on
> sun ultra20?
No, the U20 was tested, and in fact was the first supported machine.
However, that doesn't mean that some later driver/kernel update didn't
cause it to stop working. Recent runs on later machines have shown
that there is some inconsistancy in USB and ATA drivers, so you may
well be seeing this. Having the serial port connected to another
machine and tracing the output will help confirm or reject this
hypothesis.
History has taught me to not trust the BIOS (as you seem to have
found with the fans), but the BIOS on your machine wouldn't have
changed between a point when it worked to a point when it didn't, so
that implies to me that the problem is more likely with an updated
driver or kernel, than the BIOS (certainly is my claim with our
internal machines showing regressions).
Of course, this message doesn't help solve your immediate problem,
but if you run a text console (disable cde-login:default/gdm:default,
or use the cde-login "command line" login), use:
# mdb -kw
> cpr_debug/W0xff
> <ctrl-d>
# uadmin 3 22
what sort of output do you get?
---- Randy
