Username or EmailPassword
If you are an OBSD desktop/laptop user with Intel based graphics you are going to appreciate their work when 5.4 is released. I always use -current and having KMS is a great new feature. Works flawlessly on a Dell Optiplex 9010.
It's a different, sadder, story if you have a newer AMD GPU. AMD removed proper support for UMS (Userland Mode Setting) in their recent GPUs, including the HD6310 I have, and OpenBSD doesn't have KMS for AMD even in -current, so you get all kinds of shitty GPU bugs like having to reboot to get back into a text console if you need to leave X.
How would they "remove support" in their GPUs, were the devs using some hardware feature that is no longer in the GPUs?
And if you are talking about the drivers i thought the whole point of AMD opening up their specs is so they wouldn't NEED to maintain a driver for FOSS OSes, the kernel devs could support themselves with the FOSS drivers.
But you have to realize there is a huge change taking place at AMD when it comes to GPUs and APUs and with their limited resources they just can't afford to make drivers for a half a dozen OSes, hence the opening of the code and specs to the communities.
For those that don't know the VLIW design that has been used since the dawn of ATI has been phased out for vector based GCN, this is a fundamental change to the way that AMD GPUs and APUs work as the new design allows the GPUs to do things that before now were only possible on CPUs such as parking unused cores so as to greatly lower power when not doing graphics intensive tasks and to allow anyone writing OpenCL code to treat the CPU and GPU as a single entity, with the GPU able to intelligently split up the load to make the best use of its hardware.
So if they dropped some piece of hardware in the chip that the BSD guys were counting on I'm sorry to hear that but when you are talking about such a giant change in how GPUs work and are treated you can see why there is gonna be some growing pains.