Hi Philipp,

On 28 April 2017 at 09:53, Philipp Tomsich
<philipp.toms...@theobroma-systems.com> wrote:
> This commit enables RK3399 support for HDMI through the following
> changes:
> - adds a driverdata structure to mirror some subtle version
>   differences between the RK3399 VOPs and those in the RK3288
>   (e.g. the pin-polarity configuration)
> - configures the VOP to output 32bpp for HDMI
> - handles whether a VOP can output 10BIT data or not (i.e. RGBaaa vs. RGB888)
>   using the driverdata structure
>
> And as we touch this file anyway, we also increase the size of the
> framebuffer to a slightly overzealous 4K2K@32bpp.
>
> Signed-off-by: Philipp Tomsich <philipp.toms...@theobroma-systems.com>
> ---
>
>  arch/arm/include/asm/arch-rockchip/vop_rk3288.h |  11 ++
>  drivers/video/rockchip/rk_vop.c                 | 180 
> ++++++++++++++++++++----
>  2 files changed, 161 insertions(+), 30 deletions(-)

I'd like to somehow keep the SoC-specific code out of this driver.
You've done a nice job separating it out, but I wonder if we can go a
bit further.

I'm thinking of perhaps having two vop drivers, one for rk3288 and one
for rk3399. They can share most of the operations (e.g. bind()) which
can stay in the existing rk_vop.c file. For probe() you can rename the
existing probe() function to something like rockchip_vop_probe(), and
pass it the device and a rkvop_driverdata *.

Does that make sense? Then when we add more chips we'll have a small
extra file with the SoC-specific functionality.

Also please put the plat->size change in its own patch.

Regards,
Simon
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
https://lists.denx.de/listinfo/u-boot

Reply via email to