From mboxrd@z Thu Jan  1 00:00:00 1970
From: "Michael S. Tsirkin" <mst@redhat.com>
Subject: Re: [Qemu-devel] [PATCH v2 06/11] nvdimm acpi: initialize the
 resource used by NVDIMM ACPI
Date: Thu, 18 Feb 2016 12:20:43 +0200
Message-ID: <20160218120158-mutt-send-email-mst@redhat.com>
References: <20160215114742.382c951e@nial.brq.redhat.com>
 <20160215133722-mutt-send-email-mst@redhat.com>
 <20160215143234.29320a5f@nial.brq.redhat.com>
 <56C1F469.2040602@linux.intel.com>
 <20160215182404.0878474f@nial.brq.redhat.com>
 <56C21A7D.5040902@linux.intel.com>
 <20160216120047.5a50eccf@nial.brq.redhat.com>
 <56C3D522.6090401@linux.intel.com>
 <20160217192356-mutt-send-email-mst@redhat.com>
 <56C54298.3000904@linux.intel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: Igor Mammedov <imammedo@redhat.com>, ehabkost@redhat.com,
	kvm@vger.kernel.org, gleb@kernel.org, mtosatti@redhat.com,
	qemu-devel@nongnu.org, stefanha@redhat.com, pbonzini@redhat.com,
	dan.j.williams@intel.com, rth@twiddle.net
To: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mx1.redhat.com ([209.132.183.28]:49925 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1426118AbcBRKUt (ORCPT <rfc822;kvm@vger.kernel.org>);
	Thu, 18 Feb 2016 05:20:49 -0500
Content-Disposition: inline
In-Reply-To: <56C54298.3000904@linux.intel.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On Thu, Feb 18, 2016 at 12:03:36PM +0800, Xiao Guangrong wrote:
>=20
>=20
> On 02/18/2016 01:26 AM, Michael S. Tsirkin wrote:
> >On Wed, Feb 17, 2016 at 10:04:18AM +0800, Xiao Guangrong wrote:
> >>>>>As for the rest could that commands go via MMIO that we usually
> >>>>>use for control path?
> >>>>
> >>>>So both input data and output data go through single MMIO, we nee=
d to
> >>>>introduce a protocol to pass these data, that is complex?
> >>>>
> >>>>And is any MMIO we can reuse (more complexer=EF=BC=9F) or we shou=
ld allocate this
> >>>>MMIO page =EF=BC=88the old question - where to allocated?=EF=BC=89=
?
> >>>Maybe you could reuse/extend memhotplug IO interface,
> >>>or alternatively as Michael suggested add a vendor specific PCI_Co=
nfig,
> >>>I'd suggest PM device for that (hw/acpi/[piix4.c|ihc9.c])
> >>>which I like even better since you won't need to care about which =
ports
> >>>to allocate at all.
> >>
> >>Well, if Michael does not object, i will do it in the next version.=
 :)
> >
> >Sorry, the thread's so long by now that I'm no longer sure what does=
 "it" refer to.
>=20
> Never mind i saw you were busy on other loops.
>=20
> "It" means the suggestion of Igor that "map each label area right aft=
er each
> NVDIMM's data memory" so we do not emulate it in QEMU and is good for=
 the performance
> of label these are the points i like. However it also brings complexi=
ty/limitation for
> later DSM commands supports since both dsm input & output need to go =
through single MMIO.
>=20
> Your idea?

I think mapping right after data is problematic since it might
use 1G of address space if alignment is used (and alignment is
good for performance, so we typically do want people to use it).
Given you will add more DSM commands anyway,
I don't see a reason not to pass label data this way too.

I don't care much how are commands passed exactly.
Igor, do you have a preference or if it's not MMIO beyong DIMM
then you don't care?

--=20
MST