Re: too much memory for node

[Ben Dooks <ben-linux@fluff.org>]

On Wed, Jan 20, 2010 at 04:10:53PM +0000, Russell King - ARM Linux wrote:
> On Wed, Jan 20, 2010 at 04:05:25PM +0000, Ben Dooks wrote:
> > On Wed, Jan 20, 2010 at 03:55:10PM +0000, Russell King - ARM Linux wrote:
> > > On Wed, Jan 20, 2010 at 03:34:11PM +0000, Ben Dooks wrote:
> > > > I've been looking at support for a new machine type where the
> > > > current system is using DISCONTIGMEM as the 1G memory map has
> > > > 4x256M RAM ranges which may not all be filled.
> > > 
> > > Firstly, use sparsemem, unless you intentionally want to spend more
> > > cycles in the kernel.
> > 
> > Ok, any idea if that will fix the problem in this case?
> 
> It _may_ do, because it doesn't have the concept of NUMA nodes (which
> is what discontigmem is all about.)
> 
> > Should I send a patch to detect bad DISCONTIGMEM setup? I think that
> > a check of PHYS_TO_PFN((start+end) != PHYS_TO_PFN(start)) should be
> > safe in all cases?
> 
> Just move away from discontigmem completely.  Sparsemem can do everything
> we need on ARM at a lesser cost.  As I say above, discontigmem is really
> for NUMA platforms, and ARM is not a NUMA platform.  No one on ARM should
> be going anywhere near discontigmem.

Thanks, we seem to have working sparsemem support now so I will try and
ensure that is what gets sumitted.

-- 
Ben

Q:      What's a light-year?
A:      One-third less calories than a regular year.