Jasen Betts wrote in a message to Roy J. Tellason:

 RJT> What,  battery dead?  Ok,  but that should make any difference if
 RJT> I don't shut it off,  right?  Wrong.  I'd go through and put all
 RJT> of the settings in,  and then hit "save and exit" and I'd get the
 RJT> _same_ message!  Go back into setup, and there's nothing there
 RJT> but the default settings

 RJT> Checked the battery,  it's a common coin cell,  CR2032,  and
 RJT> measures 0.5v! Never saw one get that low...

 JB> They often don't power the cmos from the motherboard, and only rely
 JB> on the coin cell.

Oh really?  As far as I can tell it's not a separate chip but built into
one of the other ones,  so that'd be a little hard to do.

 JB> linux only needs the cmos to boot.

*Anything* needs cmos right to boot...

 RJT> other reasons (keyboard connector's in a strange place),  and
 RJT> checked that -- 2.8v -- but it gives me the same result

 JB> 2.8 should be enough to hold the data, maybe it's not enough for
 JB> the writes to the cmos to "take"  OTOH the CMOS could be dead

It's dead.  Tried another one,  measuring 3.12V (like new) and that still
didn't work.

 RJT> Ok,  time to start swapping hardware...

 RJT> Giving me "assertion flubbed" (?!) and "segment
violation" and I
 RJT> see things loading that weren't supposed to be loading according
 RJT> to _my_ configs.  Takes an awful long time to get to a login
 RJT> prompt,  and then when it does before I can try anything I see
 RJT> another segment violation,  kernel PANIC

 JB> unfamiliar hardware - time for a rescue floppy.

I dunno,  I think that MB had some serious problems,  it's in the pile now.
 Not that the replacement turned out to be _that_ much better.

 RJT> Or maybe not,  I feel lousy,  probaby have that bug that's been
 RJT> going around, and won't be going anywhere today...  :-(

 JB> or maybe it's from staying up all night fighting baulky hardware.
 JB> :-( 

Nah,  it was something going around.  I'm (mostly) over it.  Still fighting
the hardware,  though.

---