From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43)
	id 1LCjOb-0004Vi-IN
	for qemu-devel@nongnu.org; Tue, 16 Dec 2008 18:32:09 -0500
Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43)
	id 1LCjOY-0004VF-SO
	for qemu-devel@nongnu.org; Tue, 16 Dec 2008 18:32:09 -0500
Received: from [199.232.76.173] (port=56936 helo=monty-python.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1LCjOY-0004VC-O0
	for qemu-devel@nongnu.org; Tue, 16 Dec 2008 18:32:06 -0500
Received: from outbound-sin.frontbridge.com ([207.46.51.80]:41402
	helo=SG2EHSOBE001.bigfish.com)
	by monty-python.gnu.org with esmtps (TLS-1.0:RSA_ARCFOUR_MD5:16)
	(Exim 4.60) (envelope-from <Andre.Przywara@amd.com>)
	id 1LCjOY-0000Sz-AI
	for qemu-devel@nongnu.org; Tue, 16 Dec 2008 18:32:06 -0500
Message-ID: <49483A64.3040607@amd.com>
Date: Wed, 17 Dec 2008 00:31:48 +0100
From: Andre Przywara <andre.przywara@amd.com>
MIME-Version: 1.0
References: <4947B784.8020300@amd.com> <49481A15.6080402@codemonkey.ws>
In-Reply-To: <49481A15.6080402@codemonkey.ws>
Content-Type: text/plain; charset="ISO-8859-1"; format=flowed
Content-Transfer-Encoding: 7bit
Subject: [Qemu-devel] Re: [PATCH 1/8] v2: introduce -numa command line option
Reply-To: qemu-devel@nongnu.org
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.gnu.org/pipermail/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Anthony Liguori <anthony@codemonkey.ws>
Cc: qemu-devel@nongnu.org, Avi Kivity <avi@redhat.com>

Anthony Liguori wrote:
>> +        if (i == numnumanodes) {
>> +            for (i = 0; i < smp_cpus; i++)
>> +                node_to_cpus[i % numnumanodes] |= 1<<i;
>>   
> 
> The way CPUs are allocate here seems strange?  Each CPU is assigned 
> round robin?  Should you have node 0 contain 1..X, node 1 contain X..Y, 
> node 2 contain Y..smp_cpus?
This is because an algorithm to do it this way is more complicated than 
this round-robin scheme (imagine distributing 8 CPUs over 3 nodes). 
Actually I know machines which do this natively (probably for the same 
reason, code simplicity in the BIOS). OSes can cope with my variant too, 
there is no guarantee that sequentially numbered CPUs belong to one node.
But if you insist, I fix it.

Regards,
Andre.

-- 
Andre Przywara
AMD-OSRC (Dresden)
Tel: x84917