This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


Re: [Xen-devel] [PATCH 0/6] xen,xend,tools: Add NUMA support to Xen

To: xen-devel@xxxxxxxxxxxxxxxxxxx
Subject: Re: [Xen-devel] [PATCH 0/6] xen,xend,tools: Add NUMA support to Xen
From: Tristan Gingold <Tristan.Gingold@xxxxxxxx>
Date: Wed, 2 Aug 2006 08:14:43 +0200
Delivery-date: Tue, 01 Aug 2006 23:10:31 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
In-reply-to: <20060731190958.GI1694@xxxxxxxxxx>
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
References: <20060731190958.GI1694@xxxxxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: KMail/1.5
Le Lundi 31 Juillet 2006 21:09, Ryan Harper a écrit :
> I've respun the NUMA patches against 10874 and I'm re-submitting them
> with the optimizations mentioned [1]previously on the list.  There was a
> request to see the overhead on non-numa/single-node machines.  I've
> re-run those benchmarks (ballooning up from small mem to multi-gig) as
> well as timing the initially domain increase_reservation time to gauge
> the overhead when allocating from the heap.

two more points.

For DOM0_PHYSINFO, sockets_per_node is wrong:
      pi->sockets_per_node = 
            num_online_cpus() / cpus_weight(cpu_core_map[0]);
[On ia64, I return the max of sockets per node].

The node_to_cpu field of struct dom0_physinfo is:
    XEN_GUEST_HANDLE(u64) node_to_cpu;
This means the max cpus is 64.
I know cpumap_t is also 64 bits.
I think you'd better to use cpumap_t.

Here we have NUMA machines with 128 cpus.  The question is how to upgrade Xen 
to handle more than 64 cpus.  We can either incease cpumap_t size to 256 
(until we reach a new limit) or use an unlimited cpu structure.


Xen-devel mailing list