RE: dt-bindings: display: xlnx: mixer: Inconsistent pixel format terminology in dt docs
From: Hyun Kwon
Date: Thu Aug 20 2020 - 19:43:51 EST
Hi Kenneth,
> -----Original Message-----
> From: Kenneth Sloat <ksloat@xxxxxxxxxxxxxxxx>
> Sent: Thursday, August 20, 2020 2:18 PM
> To: Hyun Kwon <hyunk@xxxxxxxxxx>; linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
> Cc: Michal Simek <michals@xxxxxxxxxx>; dri-devel@xxxxxxxxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx; laurent.pinchart@xxxxxxxxxxxxxxxx;
> devicetree@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> mark.rutland@xxxxxxx
> Subject: dt-bindings: display: xlnx: mixer: Inconsistent pixel format terminology
> in dt docs
>
> Hello,
>
> The Xilinx Video mixer IP uses the DRM fourcc string as a device tree binding in
> order to describe the format for a specific DRM layer/plane. Currently, in the
> device tree doc in
> "Documentation/devicetree/bindings/dma/xilinx/xilinx_frmbuf.txt" there is a list
> called "VIDEO FORMATS" with a table containing "Mixer IP Format" and
> associated "Driver supported Format String" (fourcc). In general, this table is not
> helpful, because "Mixer IP Format" does not match the Xilinx IP datasheet
> terminology for formats but rather some other commonly used nomenclature.
> This is also inconsistent from other Xilinx dts binding docs which specify the
> actual correct IP name and at times list other format nomenclature where
> helpful (i.e. xilinx_frmbuf.txt)
>
> My suggestion is to remove the current "Mixer IP Format" strings in the doc and
> replace them with the actual IP names in the datasheet (which also match those
> used in Xilinx's Vivado FPGA tool). Otherwise, the programmer must cross
> reference other dts documentation or the code in order to figure these out. I am
> happy to provide a patch for this but wanted to throw out the suggestion first.
I agree. The bindings should be in sync with datasheet and not rely on
software definitions. Since the driver is only in Xilinx vendor tree,
please send directly to git@xxxxxxxxxx.
Thanks!
-hyun
>
> Thanks
>
> Sincerely,
> Ken Sloat