Subject: Re: 'set foo' not sticking (PC164)
To: Peter Seebach <>
From: Robert Elz <kre@munnari.OZ.AU>
List: port-alpha
Date: 08/17/1999 07:08:26
    Date:        Mon, 16 Aug 1999 14:03:59 -0500
    From: (Peter Seebach)
    Message-ID:  <>

  | Am I missing
  | something obvious, like a "save" command?

I don't think so.   Unless the PC164 SRM is different from those on
other Alphas, they ought to just be remembered.

I see the same thing - I have considered sticking in a new battery,
but the tech types here tell me the battery is just fine, and doing
so would be a waste of time (but then again, batteries are cheap, so
I may do so anyway).

My PC164 also tells me that the isacfg table is corrupt, which it attempts
to fix by resetting it back to the defaults (which it seems happy enough to
do), but then it is always corrupt again, any time it is referenced.
("isacfg -init" doesn't help, though after complaining about the table
being corrupt, it seems to work).   Of course, I have no ISA cards installed
to configure, so it is always possible that the isacfg table can't be
correctly formatted until there is one, though that seems unlikely.

I am considering switching the 5.4 SRM I have in there now to the 5.3
version and seeing if that makes any difference.

While I'm here, and assuming it is relevant at all, can someone point me
at the NetBSD source location where the mapping between isacfg handles and
the kernel is kept?   That is, if I did, for some reason, want to install
an ISA card, how would I discover what handle to tell isacfg, so the
kernel would later recognise the card.  Or is NetBSD pretending that the
alpha is an i386, ignoring the isacfg table, and just probing the bus
hunting for whatever might be out there?