[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20211109075100.GA27284@lst.de>
Date: Tue, 9 Nov 2021 08:51:00 +0100
From: Christoph Hellwig <hch@....de>
To: Jani Nikula <jani.nikula@...ux.intel.com>
Cc: Zhi Wang <zhi.wang.linux@...il.com>,
joonas.lahtinen@...ux.intel.com, rodrigo.vivi@...el.com,
zhenyuw@...ux.intel.com, zhi.a.wang@...el.com, jgg@...dia.com,
intel-gfx@...ts.freedesktop.org,
intel-gvt-dev@...ts.freedesktop.org,
dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
hch@....de
Subject: Re: [PATCH 1/3] i915/gvt: seperate tracked MMIO table from
handlers.c
On Tue, Nov 09, 2021 at 09:00:39AM +0200, Jani Nikula wrote:
> On Mon, 08 Nov 2021, Zhi Wang <zhi.wang.linux@...il.com> wrote:
> > From: Zhi Wang <zhi.wang.linux@...il.com>
> >
> > To support the new mdev interfaces and the re-factor patches from
> > Christoph, which moves the GVT-g code into a dedicated module, the GVT-g
> > MMIO snapshot still needs to be saved in i915 so that the inital clean HW
> > state can be used for the further vGPU. Seperate the tracked MMIO table
> > from GVT-g, so that GVT-g and i915 can both use it.
>
> Do you really have to both put code in a header and then include that in
> multiple places?
>
> I think you may need to rethink the whole approach, maybe make them
> actual tables instead of code.
Without understanding this code too well: an approach that makes in
actual table and uses an accessor seems more useful to me as well.
Powered by blists - more mailing lists