drm / drm2 removal in 12

[blubee blubeeme]

On Sat, Aug 25, 2018 at 8:40 AM Pete Wright <pete at nomadlogic.org> wrote:

>
>
> On 8/24/18 4:07 PM, blubee blubeeme wrote:
>
> > This project: https://wiki.freebsd.org/Use%20linuxkpi%20in%20DRM
> > Goals
> >
> >     - Move DRM headers to a similar location as Linux
> >     -
> >
> >     Use kmalloc() instead of malloc(9)
> >     - Use kref
> >     -
> >
> >     Use idr and get rid of drm_gem_names.c
> >     - Use PCI API
> >     - Use Linux locking primitives
> >
> > is garbage, if you want to use develop Linux code and use Linux then go
> do
> > that on Linux.
> having a hard time not feeding the troll here...but what specifically is
> garbage.  as in, what implementation of all this work do you have
> available that has been developed independently which also enables
> support for modern desktop and portable systems that you can buy today?
> > Are these guys insane and please avoid the nonsense about you're doing
> this
> > in your spare time.
>
The idea that FreeBSD relax it's standards just so that some devs have an
easier time
bringing up a half baked idea is nonsense.

Let's take power management, after you guys do all this work to get the
graphics card working
how much of devd will you need to implement to get that working properly?

I don't understand why this concept seems so hard to grasp but FreeBSD is
not Linux
why are some people continually trying to turn it into some Frankenstein
thing.

If you guys consider yourself developers then do what developers do and
solve problems
with constraints, if you cannot accomplish that stop pushing these breaking
changes.

None of these kmod guys seems to have put any thought into long
term maintenance of
this project. Look at the mailing list, every few days there's some
breaking changes waiting
for patches because something changed in Linux-land...

If you can't solve the problem in a maintainable way, you will just create
bigger problems for
developers down the line. Until you guys have something that's at least as
stable as what's
available now, keep working on it.

Some people take pride in their work and deliver a working product, they
don't need to twist
peoples arms into getting their way.

speaking as someone who's been working on this from pretty much the day
> of the initial CFT (maybe before?) - i don't know anyone who's getting
> paid for this specific work.  at least when it comes to GPU support.
> but, if you have the means, I'd love to work on this full time and am
> open to any serious offers :)
>
> -pete
>
I'd hope you have something more compelling than [http://nomadlogic.org]
as your calling card.

-- 
Pete Wright
pete at nomadlogic.org
@nomadlogicLA