Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <cover.1561491964.git.rodrigosiqueiramelo@gmail.com>
 <ea7e3a0daa4ee502d8ec67a010120d53f88fa06b.1561491964.git.rodrigosiqueiramelo@gmail.com>
 <20190711082105.GI15868@phenom.ffwll.local> <hG3hgN80Bt03njzCaW7h3xaog7ppTTBzmsShC0L5LdCbr5dFkHMJHHxizeYa_IYP7uCwMG-vPJOWMhueq2LirNKFhulkkni2KFf3XA24bb8=@emersion.fr>
In-Reply-To: <hG3hgN80Bt03njzCaW7h3xaog7ppTTBzmsShC0L5LdCbr5dFkHMJHHxizeYa_IYP7uCwMG-vPJOWMhueq2LirNKFhulkkni2KFf3XA24bb8=@emersion.fr>
From:   Daniel Vetter <daniel@ffwll.ch>
Date:   Thu, 11 Jul 2019 11:00:56 +0200
Message-ID: <CAKMK7uHfnOHJS8sQ0fpysSTB-SiJ29ACH+rKQ45rhT+7rBH5pQ@mail.gmail.com>
Subject: Re: [PATCH V3 4/5] drm/vkms: Compute CRC without change input data
To:     Simon Ser <contact@emersion.fr>
Cc:     Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com>,
        Brian Starkey <brian.starkey@arm.com>,
        Liviu Dudau <liviu.dudau@arm.com>,
        Haneen Mohammed <hamohammed.sa@gmail.com>,
        "dri-devel@lists.freedesktop.org" <dri-devel@lists.freedesktop.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Jul 11, 2019 at 10:28 AM Simon Ser <contact@emersion.fr> wrote:
>
> On Thursday, July 11, 2019 11:21 AM, Daniel Vetter <daniel@ffwll.ch> wrote:
>
> > On Tue, Jun 25, 2019 at 10:38:31PM -0300, Rodrigo Siqueira wrote:
> >
> > > The compute_crc() function is responsible for calculating the
> > > framebuffer CRC value; due to the XRGB format, this function has to
> > > ignore the alpha channel during the CRC computation. Therefore,
> > > compute_crc() set zero to the alpha channel directly in the input
> > > framebuffer, which is not a problem since this function receives a copy
> > > of the original buffer. However, if we want to use this function in a
> > > context without a buffer copy, it will change the initial value. This
> > > patch makes compute_crc() calculate the CRC value without modifying the
> > > input framebuffer.
> >
> > Uh why? For writeback we're writing the output too, so we can write
> > whatever we want to into the alpha channel. For writeback we should never
> > accept a pixel format where alpha actually matters, that doesn't make
> > sense. You can't see through a real screen either, they are all opaque :-)
>
> I'm not sure about that. See e.g.
> https://en.wikipedia.org/wiki/See-through_display

They have variable opaqueness, independent of the color value?

> Many drivers already accept FBs with alpha channels for the primary
> plane.
> https://drmdb.emersion.fr/formats?plane=1

If you have a background color (we assume it to be black) that makes
sense. Still doesn't mean we render transparent output, we don't.

> Just making sure we aren't painting ourselves into a corner. :P

You can add ARGB to your writeback format support list, there is no
corner here at all to get into (at least in the abstract sense).
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch