From mboxrd@z Thu Jan  1 00:00:00 1970
From: boris.brezillon@bootlin.com (Boris Brezillon)
Date: Wed, 18 Apr 2018 10:27:55 +0200
Subject: [PATCH v2 4/6] drm/atmel-hlcdc: support bus-width (12/16/18/24)
 in endpoint nodes
In-Reply-To: <8c5847d0-cde1-73fa-9ed1-4f4a5420afbd@axentia.se>
References: <20180417131052.16336-1-peda@axentia.se>
 <20180417131052.16336-5-peda@axentia.se>
 <20180418092948.6dace10a@bbrezillon>
 <8c5847d0-cde1-73fa-9ed1-4f4a5420afbd@axentia.se>
Message-ID: <20180418102755.0491bc9d@bbrezillon>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On Wed, 18 Apr 2018 09:46:09 +0200
Peter Rosin <peda@axentia.se> wrote:

> On 2018-04-18 09:29, Boris Brezillon wrote:
> > On Tue, 17 Apr 2018 15:10:50 +0200
> > Peter Rosin <peda@axentia.se> wrote:
> >   
> >> This beats the heuristic that the connector is involved in what format
> >> should be output for cases where this fails.
> >>
> >> E.g. if there is a bridge that changes format between the encoder and the
> >> connector, or if some of the RGB pins between the lcd controller and the
> >> encoder are not routed on the PCB.
> >>
> >> This is critical for the devices that have the "conflicting output
> >> formats" issue (SAM9N12, SAM9X5, SAMA5D3), since the most significant
> >> RGB bits move around depending on the selected output mode. For
> >> devices that do not have the "conflicting output formats" issue
> >> (SAMA5D2, SAMA5D4), this is completely irrelevant.
> >>
> >> Signed-off-by: Peter Rosin <peda@axentia.se>
> >> ---
> >>  drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c | 85 ++++++++++++++++++++------
> >>  1 file changed, 65 insertions(+), 20 deletions(-)
> >>
> >> diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c
> >> index d73281095fac..2e718959981e 100644
> >> --- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c
> >> +++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c
> >> @@ -19,12 +19,14 @@
> >>   */
> >>  
> >>  #include <linux/clk.h>
> >> +#include <linux/of_graph.h>
> >>  #include <linux/pm.h>
> >>  #include <linux/pm_runtime.h>
> >>  #include <linux/pinctrl/consumer.h>
> >>  
> >>  #include <drm/drm_crtc.h>
> >>  #include <drm/drm_crtc_helper.h>
> >> +#include <drm/drm_of.h>
> >>  #include <drm/drmP.h>
> >>  
> >>  #include <video/videomode.h>
> >> @@ -226,6 +228,68 @@ static void atmel_hlcdc_crtc_atomic_enable(struct drm_crtc *c,
> >>  #define ATMEL_HLCDC_RGB888_OUTPUT	BIT(3)
> >>  #define ATMEL_HLCDC_OUTPUT_MODE_MASK	GENMASK(3, 0)
> >>  
> >> +static int atmel_hlcdc_connector_output_mode(struct drm_connector_state *state)
> >> +{
> >> +	struct drm_connector *connector = state->connector;
> >> +	struct drm_display_info *info = &connector->display_info;
> >> +	unsigned int supported_fmts = 0;
> >> +	struct device_node *ep;
> >> +	int j;
> >> +
> >> +	/*
> >> +	 * Use the connector index as an approximation of the
> >> +	 * endpoint node index. We know it's true for our case
> >> +	 * depending on the driver implementation.
> >> +	 */
> >> +	ep = of_graph_get_endpoint_by_regs(connector->dev->dev->of_node, 0,
> >> +					   connector->index);
> >> +  
> > 
> > Hm, this sounds a bit fragile. Can't we have a reference to the of_node
> > attached to the connector? Or maybe we can parse this earlier and set a
> > constraint on the accepted modes.
> >   
> >> +	if (ep) {
> >> +		int bus_fmt = drm_of_media_bus_fmt(ep);  
> > 
> > Hm, you're extracting this piece of information from the DT every time
> > an atomic modeset is done. I'd really prefer to have this done once at  
> 
> Yes, not happy about it either. I looked for other sensible places too
> hook the info at probe time, but this was just the simplest. I'll take
> another look...
> 
> > probe time. Since this property is attached to the connector, maybe we
> > should overwrite the info->bus_formats[] array or mark some of its
> > entries as invalid.  
> 
> I find it very wrong to mix the connector format with what you want to
> output. In my mind it's a broken assumption that they are related. It is
> only correct for trivial cases. Also note my comment about the connector
> index and the endpoint index, they are only coincidentally the same
> based on our implementation. If the driver has more than one port or
> initializes endpoints out of order for some reason, this is no longer
> true.
> 
> I think it would be better to store this info somewhere near the encoder,
> since that is what I find closest to what I'm trying to change.

Totally agree with you on that: the connector has nothing to do with
how intermediate encoders/bridges exchange data with each other.

> 
> As I said, I'll take another look and see if I can hook this in at some
> other place.

Okay, cool.