[1766] in Release_7.7_team

home help back first fref pref prev next nref lref last post

RESOLVED! Case 138969: SGI Indy R5000 - DS1386-8K-150 Ramified Timekeeper

daemon@ATHENA.MIT.EDU (Bill Cattey)
Mon Apr 26 17:58:39 1999

Message-Id: <or9C4XgGgE6e00_cs0@mit.edu>
Date: Mon, 26 Apr 1999 17:56:51 -0400 (EDT)
From: Bill Cattey <wdc@MIT.EDU>
To: y2k-help@MIT.EDU, baker@fablab.mit.edu, wdc@MIT.EDU, dhall@MIT.EDU,
        hotline@MIT.EDU, miked@MIT.EDU, owls@MIT.EDU, release-team@MIT.EDU
Cc: mbarker@MIT.EDU, rbasch@MIT.EDU, rferrara@MIT.EDU,
        modica@remington.csd.sgi.com (Steve Modica), ccomanor@csd.sgi.com,
        decastro@boston.sgi.com

I have received information back from SGI which convinces me that they
have taken adequate remediation of the 2 digit year in the DALLAS chip
in the Indy:

Summary:

Back in 1994, they made the OS offest the 2 digit year at 1940, instead
of 1900.  After 1/1/2000, the date command in the PROM monitor will
display the incorrect year, but that is purely cosmetic.  Essentially,
noone uses the PROM date command to view the date.  Everything else is
keyed to operate off the 40 year offset.  Testing for power down, and
reboot for critical dates has been done.

Detail:

    Date: Mon, 26 Apr 1999 13:57:11 -0700 (PDT)
    From: modica@remington.csd.sgi.com (Steve Modica)
    To: wdc@MIT.EDU
    Subject: case 1097556 indy y2k issues, patch issues

    Anyhow, custom engineering went back and examined the module
    supporting the dallas chip.  The last relevant changed were made in
    1994.  "At that time, rev 1.131 changed the time base for the RTC to
    1940 allowing the OS software to properly figure out what the year
    is.  The system has also been extensively tested for Y2K compliance
    in our labs as part of the IRIX 6.5 release.  There is a cosmetic
    bug in the PROM date command that incorrectly displays the year
    after 1/1/2000, but this is harmless and the system boots normally
    with the correct year."

    Additionally, several power down reboot tests were run showing that
    no problems occurred with power down reboots at midnight on these
    dates:

    12/31/99
    2/28/2000
    2/29/2000
    3/31/2000
    4/30/2000
    12/31/2000
    12/31/2009

Special thanks to Steve Modica for resolving the issue, and to Dayton
Wohlman and Kristine Furrer for staying with the case to get it resolved.

-wdc

home help back first fref pref prev next nref lref last post