[PATCH v3 5/6] video: seps525: Add seps525 SPI driver

[Simon Glass]

Hi Michal,

On Tue, 15 Dec 2020 at 07:19, Michal Simek <michal.simek at xilinx.com> wrote:
>
> Add support for the WiseChip Semiconductor Inc. (UG-6028GDEBF02) display
> using the SEPS525 (Syncoam) LCD Controller. Syncoam Seps525 PM-Oled is RGB
> 160x128 display. This driver has been tested through zynq-spi driver.
>
> ZynqMP> load mmc 1 100000 rainbow.bmp
> 61562 bytes read in 20 ms (2.9 MiB/s)
> ZynqMP> bmp info 100000
> Image size    : 160 x 128
> Bits per pixel: 24
> Compression   : 0
> ZynqMP> bmp display 100000
> ZynqMP> setenv stdout vidconsole
>
> Signed-off-by: Michal Simek <michal.simek at xilinx.com>
> Reviewed-by: Simon Glass <sjg at chromium.org>
> ---
>
> Changes in v3:
> - use 32bit reg/value parameters
>
> Changes in v2:
> - Fix Kconfig entry and use VIDEO_SEPS525 instead of SEPS525 and fix
>   location
> - Add kernel-doc structure for seps525_priv structure
> - Remove flags variable and use macros directly
>
>  MAINTAINERS             |   1 +
>  drivers/video/Kconfig   |   7 +
>  drivers/video/Makefile  |   1 +
>  drivers/video/seps525.c | 327 ++++++++++++++++++++++++++++++++++++++++
>  4 files changed, 336 insertions(+)
>  create mode 100644 drivers/video/seps525.c
>

Don't worry about changing it again...but I just wanted to mention
that my interest was in using natural types (e.g. int / uint) for
fuunction parameters - so on arm64 this would actually be a 64-bit
value!

I'm pleased to hear that there is no size impact of using u8. Perhaps
the compiler it better, and after all these are static functions so
there is full visibility for the compiler. I did have a project some
years ago full of u8 params and it added a lot to code size.

Regards,
Simon