Re: [PATCH v1 1/4] fbtft: Unorphan the driver

[Helge Deller]

On 1/26/22 11:31, Daniel Vetter wrote:
> On Wed, Jan 26, 2022 at 9:31 AM Greg Kroah-Hartman
> <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>>
>> On Tue, Jan 25, 2022 at 10:21:14PM +0200, Andy Shevchenko wrote:
>>> Let's maintain occasional fixes to the fbtft driver.
>>>
>>> Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
>>> ---
>>>  MAINTAINERS | 4 +++-
>>>  1 file changed, 3 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/MAINTAINERS b/MAINTAINERS
>>> index ea3e6c914384..16e614606ac1 100644
>>> --- a/MAINTAINERS
>>> +++ b/MAINTAINERS
>>> @@ -7372,9 +7372,11 @@ F:     Documentation/fault-injection/
>>>  F:   lib/fault-inject.c
>>>
>>>  FBTFT Framebuffer drivers
>>> +M:   Andy Shevchenko <andy@xxxxxxxxxx>
>>>  L:   dri-devel@xxxxxxxxxxxxxxxxxxxxx
>>>  L:   linux-fbdev@xxxxxxxxxxxxxxx
>>> -S:   Orphan
>>> +S:   Maintained
>>> +T:   git git://git.kernel.org/pub/scm/linux/kernel/git/andy/linux-fbtft.git
>>
>> I'm ok with the files moving if the dri developers agree with it.  It's
>> up to them, not me.
>
> On one hand I'm happy anytime someone volunteers to help out.
>
> On the other hand ... why does it have to be resurrecting fbdev?
> There's an entire community of people who really know graphics and
> display and spent considerable amount of effort on creating useful and
> documented helpers for pretty much anything you might ever want to do.
> And somehow we have to go back to typing out things the hard way, with
> full verbosity, for an uapi that distros are abandoning (e.g. even for
> sdl the direction is to run it on top of drm with a compat layer,
> afaiui fedora is completely ditching any userspace that still even
> uses /dev/fb/0). And yes I know there's still some gaps in drm,
> largely for display features which were really en vogue about 20 years
> ago. And we're happy to add that support, if someone who still has
> such hardware can put in the little bit of work needed ...
>
> I don't get this.

You are describing a transitioning over to DRM - which is Ok.
But on that way there is no need to ignore, deny or even kill usage scenarios
which are different compared to your usage scenarios (e.g. embedded devices,
old platforms, slow devices, slow busses, no 3D hardware features,
low-color devices, ...).

Helge