On Mon, 25 Sep 2000, Bill Sommerfeld wrote:

> > 	Currently we default to 10% of the first 2MB and 5% of the remaining
> > 	memory for bufpages. On a machine with 4MB or less of RAM this is
> > 	probably more than can be afforded.
> 
> I don't think so.  Buffers getting 10% of kernel memory was the
> default back in the days where 4MB of memory was the norm...

	But kernels were much smaller then.

	I'm trying to help the low memory case, (For example 4MB arm26
	boxes where the default ~260k of buffers is much more than can
	be afforded). Install kernels can manually tune down the bufpages,
	though it would obvous help there also.

	Currently we calculate bufpages based on physmem. Is uvmexp.free
	valid at that point - if so would it make more sense to calculate
	buffers based on that?

	(I'm assuming UBC will change all the variables here, but for
	now...)

                David/absolute
			       -- www.netbsd.org: A pmap for every occasion --