PC386 Y2K compliance?

[joel at OARcorp.com (joel at OARcorp.com)]

On Fri, 30 Jul 1999, erik.ivanenko wrote:

> In rtc.c of clock/pc386, there is a #define RTC_CENTURY 32.   This is
> supposed to be the century byte in the CMOS setup.  rtc.c claims that
> this byte requires modification on Dec 99.  How?

The powerpc/score603e BSP has a trick in it to avoid needing a century
bit.  RTEMS epoch is 1988.  So 88-99 --> 1988-1999, 00-87 --> 2000->2087.
Unless I missed something in the leap year math, this works perfectly for
another 88 years. FWIW I do not personally expect to be supporting this in
2087. :)

Remember POSIX time_t breaks in 2037 unless it is changed to 64 bits.

And the RTEMS internal time math runs out around 2125 unless we go to 64
bits before then.