On Mon, May 29, 2000 at 10:51:48AM +1000, Simon Burge wrote:

 > The cp_time array is currently longs - I propose changing it to
 > u_int64_t for two reasons:
 > 
 >  + On a mostly idle 256HZ 32-bit machine (eg pmax), the CP_IDLE
 >    counter will overflow in approx 100 days.
 >  + In a compat32 situation on a 64 bit kernel the results of a
 >    userland program reading cp_time will be bogus.
 > 
 > Any reasons not to do this?

Seems okay to me.

That reminds me... need to make cp_time[] a per-CPU variable :-)

-- 
        -- Jason R. Thorpe <thorpej@zembu.com>