Re: hash table sizes

[Matt Mackall <mpm@selenic.com>]

On Tue, Nov 25, 2003 at 01:24:39PM -0800, Andrew Morton wrote:
> jbarnes@sgi.com (Jesse Barnes) wrote:
> >
> > On Tue, Nov 25, 2003 at 01:07:41PM -0800, Andrew Morton wrote:
> > > the size of these tables dependent upon the number of dentries/inodes/etc
> > > which the system is likely to support.  And that does depend upon the
> > > amount of direct-addressible memory.
> > > 
> > > 
> > > So hum.  As a starting point, what happens if we do:
> > > 
> > > -	vfs_caches_init(num_physpages);
> > > +	vfs_caches_init(min(num_physpages, pages_in_ZONE_NORMAL));
> > > 
> > > ?
> > 
> > Something like that might be ok, but on our system, all memory is in
> > ZONE_DMA...
> > 
> 
> Well yes, we'd want
> 
> 	vfs_caches_init(min(num_physpages, some_platform_limit()));
> 
> which on ia32 would evaluate to nr_free_buffer_pages() and on ia64 would
> evaluate to the size of one of those zones.

I actually just added this to the tree I'm working on:

+       vfs_caches_init(min(1000, num_physpages-16000));

Caches are too expensive on the low end of the scale as well, when the
kernel is taking up most of RAM.

-- 
Matt Mackall : http://www.selenic.com : Linux development and consulting