[Info-vax] Changed Boot behavior on XP1000 since 2013

[Hans Vlems]

On 29 jan, 14:34, Stephen Hoffman <seaoh... at hoffmanlabs.invalid>
wrote:
> On 2013-01-29 11:07:50 +0000, Hans Vlems said:
>
> > The next question is : can we fix the error?
> > Very likely we won't get any support from HP, presumably there are very
> > few XP1000's or ALpha Server 1200's (let alone Digital Server 5305's)
> > covered by a
> > hardware maintenance contract.
>
> AlphaStation XP1000 is listed as a supported system for OpenVMS Alpha
> V8.4 <http://h71000.www7.hp.com/doc/spdalphaintegrity84.pdf>, so HP
> will either be changing the SPD here, or changing the firmware, or
> (likely) changing the OpenVMS Alpha code.
>
> > So is it possible to patch or edit the binary firmware file?
>
> Yes, but not easily.
>
> The relevant OpenVMS kernel code is easier to get at, it's in the
> source listings, and it's easier to patch.
>
> Pending any comments from HP and pending any patches, the easiest fix
> is probably whacking the toy value directly from the SRM nvram script.
>
> Something like the following (untested!) should work, if edited into
> the nvram script
>
> 10 deposit -b toy:9 0d
>
> The line number (10, in the example) must be unique within the script.
>  See the SRM edit command and the nvram script documentation for
> details.
>
> Given I see no obvious way to do math on the values examined by
> previous SRM console commands, this nvram hack script will have to be
> changed once a year.  Set a calendar alarm for December 2013, with some
> notes to either look for a patch, or to tweak the nvram script.
>
> --
> Pure Personal Opinion | HoffmanLabs LLC

There is one very odd side effect. VMS is not supported on white box
alpha's.
To make it boot you needed two commands in nvram:
  10 set boot_reset on
  20 set srm_boot on
I noticed that the nvram file was empty on this machine and VMS still
booted alright.
Strange. But the deposit command worked its magic.

Hans