"Perry E. Metzger" wrote:
> However, it should be straightforward to build a /dev/nvram equivalent
> if you wish to -- there isn't much that it needs to do. A lot of the
> code you want is associated with the i386 cmos clock functions we
> already have...

That's true. I've read some doc about it and it seems that accessing
the cmos ram is pretty simple. What I would be more afraid of is to 
write the device driver part (never done this :). But as Takahiro
Kambe already did this all, it is probably not worth re-inventing
the wheel.

Werner