Re: [PATCH] 2.5.28: VM strict overcommit

[Alan Cox <alan@redhat.com>]

> 		if (!vm_enough_memory((maxpos - inode->i_size) >> PAGE_SHIFT)) {
> 			err = -ENOMEM;
> 			goto out_nc;
> 		}
> 	}
> 
> tmpfs supports holes.  Looks to me like a small write which creates
> a big hole will be severely over-accounted for?

Intentionally. The base tree doesn't support page cache removal AS callbacks
so cannot support the ideal behaviour. The other rounding bits for the
tmpfs stuff are I think all fixed in 2.4 by Hugh's stuff, but fixing 2.5
tmpfs is an ongoing seperate project.

> vm_enough_memory() looks really slow.  I'll bench this a bit.

On the benches I've run I can't see any difference whatever the
accounting mode I am using. 

> of memory is dirty" is junk.  It really wants to know more
> information about the dynamic state of the system.  So tracking
> all those datums on-the-fly would be handy.  One day.

We need it three years ago not "one day"

Alan