If that is not enough, you could probably argue to add some more ioctls there

is it enough for fbink like controls?

if fbink cant manage that, we should just rust library it, should be managable

maybe for basic stuff

then we need more ioctls

¯\_(ツ)_/¯

you have the standard drm interface for reporting dirty stuff, so you are at the level you have with pmOS + kobos. I have not analyzed the custom stuff yet

Fbink is pretty firmly stuck in fbdev land, can't do much with drm 😉

as I thought

1. Where do I see the actual exposed API from the screen 2. why the hell is this file 2.5k, cant they like split it

thats definitely not all https://github.com/m-weigand/linux/blob/11a849851a8d31dd8f0eec003d45763dd3212c77/drivers/gpu/drm/rockchip/rockchip_ebc.c#L406-L413

well, ricoh619-power.c in kobo vendor kernel has > 7k lines... that is even worse. Partial refresh is done via DRM_IOCTL_MODE_DIRTYFB. All that was three files once, they took the layout from my kobo drm driver at first

so it supports bare DRM ioctls

and DRM supports so specific eink refreshes?

well i hoped linux kernel enforces small files

well, that is not upstream

1. Where is the documentation for it? As I understand, just source files, like this is everything we have to work with: https://elixir.bootlin.com/linux/v4.6/source/include/uapi/drm/drm.h#L686-L792 2. DRM_IOCTL_MODE_DIRTYFB marks it as dirty, so 2 questions to it: - How do I replace the buffer in the first place? There is a probably a IOCTL for that, somewhere - How do I make it not redraw it in the first place?, Another ioctl somewhere I suppose? - And the real question, how do I actually pass with what waveform mode the refresh is actually done, I doubt there is a Regular DRM ioctl for that, the driver doesn't expose much ioctls - I think this thing is not implemented at all

The more I look at it the less I know

Help needed

The guy doing high FPS video in the pine discord might know

the ioctl triggers the redraw directly

probably you should go through libdrm

Who

and apparently the driver seems not to support manual selection of waveform

Yea that sucks

Like, kobo drivers gived more control

if he makes the display faster, maybe we won't need to do anything to optimise it further