Re: [PATCH v2 07/19] media: sun6i-csi: Add support for MIPI CSI-2 bridge input

From: Maxime Ripard <maxime@cerno.tech>
To: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
Cc: linux-media@vger.kernel.org, devicetree@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, linux-doc@vger.kernel.org,
	devel@driverdev.osuosl.org, linux-sunxi@googlegroups.com,
	Yong Deng <yong.deng@magewell.com>,
	Mauro Carvalho Chehab <mchehab@kernel.org>,
	Rob Herring <robh+dt@kernel.org>, Chen-Yu Tsai <wens@csie.org>,
	Jernej Skrabec <jernej.skrabec@siol.net>,
	Jonathan Corbet <corbet@lwn.net>,
	Kishon Vijay Abraham I <kishon@ti.com>,
	Vinod Koul <vkoul@kernel.org>,
	Helen Koike <helen.koike@collabora.com>,
	Dafna Hirschfeld <dafna.hirschfeld@collabora.com>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Philipp Zabel <p.zabel@pengutronix.de>,
	Sakari Ailus <sakari.ailus@linux.intel.com>,
	Hans Verkuil <hans.verkuil@cisco.com>,
	Thomas Petazzoni <thomas.petazzoni@bootlin.com>,
	kevin.lhopital@hotmail.com
Subject: Re: [PATCH v2 07/19] media: sun6i-csi: Add support for MIPI CSI-2 bridge input
Date: Wed, 2 Dec 2020 16:40:53 +0100	[thread overview]
Message-ID: <20201202154053.3fcxiift2uyqnjvp@gilmour> (raw)
In-Reply-To: <X8eiXxYw1iHKbdDV@aptenodytes>

[-- Attachment #1: Type: text/plain, Size: 2710 bytes --]

On Wed, Dec 02, 2020 at 03:19:11PM +0100, Paul Kocialkowski wrote:
> Hi,
> 
> On Tue 01 Dec 20, 13:12, Maxime Ripard wrote:
> > Hi,
> > 
> > On Sat, Nov 28, 2020 at 03:28:27PM +0100, Paul Kocialkowski wrote:
> > > The A31 CSI controller supports a MIPI CSI-2 bridge input, which has
> > > its own dedicated port in the fwnode graph.
> > > 
> > > Support for this input is added with this change:
> > > - two pads are defined for the media entity instead of one
> > >   and only one needs to be connected at a time;
> > > - the pads currently match the fwnode graph representation;
> > > - links are created between our pads and the subdevs for each
> > >   interface and are no longer immutable so that userspace can select
> > >   which interface to use in case both are bound to a subdev;
> > > - fwnode endpoints are parsed and stored for each interface;
> > > - the active subdev (and fwnode endpoint) is retrieved when validating
> > >   the media link at stream on time and cleared at stream off;
> > > - an error is raised if both links are active at the same time;
> > > - the MIPI interface bit is set if the MIPI CSI-2 bridge endpoint is
> > >   active.
> > > 
> > > In the future, the media entity representation might evolve to:
> > > - distinguish the internal parallel bridge and data formatter;
> > > - represent each of the 4 internal channels that can exist between
> > >   the parallel bridge (for BT656 time-multiplex) and MIPI CSI-2
> > >   (internal channels can be mapped to virtual channels);
> > > - connect the controller's output to the ISP instead of its
> > >   DMA engine.
> > > 
> > > Finally note that the MIPI CSI-2 bridges should not be linked in
> > > the fwnode graph unless they have a sensor subdev attached.
> > 
> > I'll leave most of the review to Laurent and Sakari, but I'm not quite
> > sure what you meant in the last paragraph. Did you mean that the
> > MIPI-CSI controller in the Allwinner SoC should only be linked if it has
> > a sensor attached, or did you mean that any MIPI-CSI2 bridge cannot be
> > attached to the controller?
> 
> So the use of plural was a mistake and your first understanding is the correct
> one: if the bridge is linked to the CSI controller in the OF graph but the
> bridge doesn't have a sensor attached, the CSI controller driver will fail
> to probe, as far as I could see.

I'm not sure it's reasonable to not link it in the DTSI then, we'll want
to reduce as much the boilerplate from the board DTS as possible, and
the MIPI-CSI controller is always there anyway. However, we should
definitely have it disabled if there's no sensor, which should solve
your probe issue

Maxime

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]