On 18/11/2020 11:40, Mark Brown wrote:
On Wed, Nov 18, 2020 at 02:03:41AM +0100, Linus Walleij wrote:
On Mon, Nov 16, 2020 at 10:06 PM Mark Brown <broonie@xxxxxxxxxx> wrote:
I think the main push in the other direction has always been people who
want to not have to write a driver at all and put absolutely everything
into DT which has scaling issues :/
What I can't understand is what gave them that idea.
This thing looks like a dream to these people for example:
https://gist.github.com/Minecrell/56c2b20118ba00a9723f0785301bc5ec#file-dsi_panel_s6e88a0_ams452ef01_qhd_octa_video-dtsi
And it looks like a nightmare to me.
(There is even a tool to convert this description into a proper display
driver now.)
It just seems to be one of those golden hammer things: everything
start to look like nails.
What people think they were sold was the idea that they shouldn't have
to write driver code or upstream things, something with more AML like
capabilities (not realising that AML works partly because ACPI hugely
constrains system design).
And is also untrue. AML only provides an API abstraction for a specific
power management model. All the actual driving of the device still
requires driver code and requires reading devices-specific properties
out of the ACPI node.
g.
