[RFC/PATCH] numa: distinguish associativity domain from node id

Dave Hansen haveblue at us.ibm.com
Thu Apr 7 05:08:31 EST 2005

On Wed, 2005-04-06 at 00:30 -0500, Nathan Lynch wrote:
> The ppc64 numa code makes some possibly invalid assumptions about the
> numbering of "associativity domains" (which may be considered NUMA
> nodes).  As far as I've been able to determine from the architecture
> docs, there is no guarantee about the numbering of associativity
> domains, i.e. the values that are contained in ibm,associativity
> device node properties.  Yet we seem to assume that the numbering of
> the domains begins at zero and that the range is contiguous, and we
> use the domain number for a given resource as its logical node id.
> This strikes me as a problem waiting to happen, and in fact I've been
> seeing some problems in the lab with larger machines violating or at
> least straining these assumptions.

I think I'm responsible for at least some of the bugs that you're
hitting.  This introduces added complexity that I was trying to avoid
when I was touching it, mostly because the power4 systems had much
simpler associativity information that was laid out sequentially.

Your changes look pretty good to me.

One minor nit:

	static int of_node_to_nid(struct device_node *dn)

That sounds to me like a conversion function, but it also does some
setup, like setting the nodes online.  I might separate the conversion
function (the read-only part) from the setup one that writes part of the

-- Dave

More information about the Linuxppc64-dev mailing list