[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190528025127.GI29553@zhen-hp.sh.intel.com>
Date: Tue, 28 May 2019 10:51:27 +0800
From: Zhenyu Wang <zhenyuw@...ux.intel.com>
To: Gerd Hoffmann <kraxel@...hat.com>
Cc: Tina Zhang <tina.zhang@...el.com>,
intel-gvt-dev@...ts.freedesktop.org, kvm@...r.kernel.org,
linux-kernel@...r.kernel.org, alex.williamson@...hat.com,
hang.yuan@...el.com, zhiyuan.lv@...el.com
Subject: Re: [PATCH 1/2] vfio: ABI for setting mdev display flip eventfd
On 2019.05.27 14:22:37 +0200, Gerd Hoffmann wrote:
> On Mon, May 27, 2019 at 05:07:41PM +0800, Zhenyu Wang wrote:
> > On 2019.05.27 16:43:11 +0800, Tina Zhang wrote:
> > > Add VFIO_DEVICE_SET_GFX_FLIP_EVENTFD ioctl command to set eventfd
> > > based signaling mechanism to deliver vGPU framebuffer page flip
> > > event to userspace.
> >
> > Should we add probe to see if driver can support gfx flip event?
>
> Userspace can simply call VFIO_DEVICE_SET_GFX_FLIP_EVENTFD and see if it
> worked. If so -> use the eventfd. Otherwise take the fallback path
> (timer based polling). I can't see any advantage a separate feature
> probe steps adds.
>
Then we need to define error return which means driver doesn't support
e.g -ENOTTY, and driver shouldn't return that for other possible
failure, so user space won't get confused.
I think if we can define this as generic display event notification?
Not necessarily just for flip, just a display change notification to
let user space query current state.
--
Open Source Technology Center, Intel ltd.
$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827
Download attachment "signature.asc" of type "application/pgp-signature" (196 bytes)
Powered by blists - more mailing lists