On Mon, Sep 11, 2000 at 08:25:27PM +0200, Manuel Bouyer wrote:

 > Well, I just tried to reproduce the you got. But now that I'm thinking about
 > it, I was using 'make -j16', and with 16 proc running maybe we run out of
 > TBL entries on i386 anyway ?
 > I'll try again with -j2

Note, I didn't add *any* extra -j options ... just "make".  The extact
process I followed was:

	make depend && make clean
	reboot
	time make
	make clean
	reboot into 4MB kernel
	time make

 > Sure, I guess for such patterns the gain can be big. Maybe also for buffer
 > cache pages ?

Buffer cache pages are problematic -- pagemove().

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