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


[Xen-devel] Re: [PATCH] Correct handling node with CPU populated but no

To: "Jiang, Yunhong" <yunhong.jiang@xxxxxxxxx>
Subject: [Xen-devel] Re: [PATCH] Correct handling node with CPU populated but no memory populated
From: Keir Fraser <keir.fraser@xxxxxxxxxxxxx>
Date: Sat, 26 Dec 2009 07:57:26 +0000
Cc: "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>
Delivery-date: Fri, 25 Dec 2009 23:58:11 -0800
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <C8EDE645B81E5141A8C6B2F73FD9265107A2AACADE@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
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/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
Thread-index: AcqFTuO2u2OxkVxJRv26PUuSaroeEgAsiyXQ
Thread-topic: [PATCH] Correct handling node with CPU populated but no memory populated
User-agent: Microsoft-Entourage/
On 25/12/2009 10:42, "Jiang, Yunhong" <yunhong.jiang@xxxxxxxxx> wrote:

> BTW, when I was working on the patch, I'm really confused by
> srat_detect_node() and init_cpu_to_node().
> Firstly,  Per my understanding to the code itself, these two functions achieve
> the same purpose, but I can't understanding why srat_detect_node need be
> called after the CPU is up. As this is backported from kernel, so I checked
> linux kernel. In linux kernel, the srat_detect_node is mainly for one quirk
> for AMD platform, but that quirk is not implemented in Xen side. So is the
> quirk is really needed?
> Also, are there any special reason that init_cpu_to_node() can't be called
> right after numa_initmem_init()? I think it will be cleaner if we setup the
> whole numa information in one place. Originally I suspect it is related to
> generic_apic_probe(), but after checking the code, I didn't find anything
> related. Did I miss anything?

I have no direct involvement in the NUMA code that got ported, and I think
the guy that originally did the port has moved on. If you think it can be
cleaned up and improved, please go ahead.

 -- Keir

Xen-devel mailing list

<Prev in Thread] Current Thread [Next in Thread>