On Sat, Feb 10, 2001 at 11:31:38AM +0000, Chris Gilbert wrote:

 > Certainly faster, the difference is probably even bigger for the slower 
 > risc-pc's.
 > 
 > The fork/exit loop comes in with:
 > 8.43 real         0.00 user         2.07 sys
 > compared with previous to this patch
 > 9.17 real         0.00 user         2.06 sys

Okay, good that it helped things a little -- I've committed it to -current.

Probably another thing getting you is the fact that page tables are
uncached.  This is going to make page table traversal much more expensive.

What was the reason this was done, again?  The clean before TLB invalidate
was too exensive?  Or, it wasn't working properly?

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