[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20190507105126.4be3a6da.cohuck@redhat.com>
Date: Tue, 7 May 2019 10:51:26 +0200
From: Cornelia Huck <cohuck@...hat.com>
To: Yan Zhao <yan.y.zhao@...el.com>
Cc: "cjia@...dia.com" <cjia@...dia.com>,
"kvm@...r.kernel.org" <kvm@...r.kernel.org>,
"aik@...abs.ru" <aik@...abs.ru>,
"Zhengxiao.zx@...baba-inc.com" <Zhengxiao.zx@...baba-inc.com>,
"shuangtai.tst@...baba-inc.com" <shuangtai.tst@...baba-inc.com>,
"qemu-devel@...gnu.org" <qemu-devel@...gnu.org>,
"kwankhede@...dia.com" <kwankhede@...dia.com>,
"eauger@...hat.com" <eauger@...hat.com>,
"Liu, Yi L" <yi.l.liu@...el.com>,
"eskultet@...hat.com" <eskultet@...hat.com>,
"Yang, Ziye" <ziye.yang@...el.com>,
"mlevitsk@...hat.com" <mlevitsk@...hat.com>,
"pasic@...ux.ibm.com" <pasic@...ux.ibm.com>,
"libvir-list@...hat.com" <libvir-list@...hat.com>,
"arei.gonglei@...wei.com" <arei.gonglei@...wei.com>,
"felipe@...anix.com" <felipe@...anix.com>,
"Ken.Xue@....com" <Ken.Xue@....com>,
"Tian, Kevin" <kevin.tian@...el.com>,
"dgilbert@...hat.com" <dgilbert@...hat.com>,
"zhenyuw@...ux.intel.com" <zhenyuw@...ux.intel.com>,
"alex.williamson@...hat.com" <alex.williamson@...hat.com>,
"intel-gvt-dev@...ts.freedesktop.org"
<intel-gvt-dev@...ts.freedesktop.org>,
"Liu, Changpeng" <changpeng.liu@...el.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"Wang, Zhi A" <zhi.a.wang@...el.com>,
"jonathan.davies@...anix.com" <jonathan.davies@...anix.com>,
"He, Shaopeng" <shaopeng.he@...el.com>
Subject: Re: [PATCH 1/2] vfio/mdev: add version field as mandatory attribute
for mdev device
On Tue, 7 May 2019 01:39:13 -0400
Yan Zhao <yan.y.zhao@...el.com> wrote:
> On Tue, Apr 30, 2019 at 11:29:08PM +0800, Cornelia Huck wrote:
> > If I followed the discussion correctly, I think you plan to drop this
> > format, don't you? I'd be happy if a vendor driver can use a simple
> > number without any prefixes if it so chooses.
> >
> > I also like the idea of renaming this "migration_version" so that it is
> > clear we're dealing with versioning of the migration capability (and
> > not a version of the device or so).
> hi Cornelia,
> sorry I just saw this mail after sending v2 of this patch set...
> yes, I dropped the common part and vendor driver now can define whatever it
> wishes to identify a device version.
Ok, I'll look at v2.
> However, I don't agree to rename it to "migration_version", as it still may
> bring some kind of confusing with the migration version a vendor driver is
> using, e.g. vendor driver changes migration code and increases that migration
> version.
> In fact, what info we want to get from this attribute is whether this mdev
> device is compatible with another mdev device, which is tied to device, and not
> necessarily bound to migration.
>
> do you think so?
I'm not 100% convinced; but we can continue the discussion on v2.
Powered by blists - more mailing lists