Starting the real time clock on virgin systems

[Eric W. Biederman]

steven james <pyro at linuxlabs.com> writes:

> Greetings,
> 
> if the init is what I think it is ( setting bit 5 in RTC index 0x0a, the
> divider), it should be the same for all RTC. The part that tends to differ
> is the PnP stuff. The index accesses from port 0x70-71 tend to stick to
> standards.

Detecting battery failure also tends to vary to the extent that
on some tyan motherboard via chipset systems the only way I have found
so far is a cmos checksum.  The cmos clear jumper is actually a cmos
corrupt jumper on that board.  On other boards cmos clear tends to set
a magic bit. 

Eric