X.Org 7.7 Goes Into FreeBSD, KMS Is Ready
Phoronix: X.Org 7.7 Goes Into FreeBSD, KMS Is Ready
While X.Org 7.7 was only released last week, this updated set of X packages have already worked their way into FreeBSD. The FreeBSD developers request your help in testing...
I hope FreeBSD can make X.org run without root privileges.
FreeBSD has any plans for anything like Wayland, or are they happy with X.org?
I think manpower is the main issue - they're having enough trouble keeping up with Xorg changes and updates (even though this is getting better), so I think Wayland is not even on the horizon.
so instead of making server features better they are catering to a niche userbase of freebsd desktop user who want awesome graphics performance. facepalm
At one point, I think Intel dropped support for their non-KMS drivers.
Originally Posted by garegin
I think that's back now, but it won't be long until they drop it again.
Then you won't be able to use any graphics at all on FreeBSD without this code. That's already true on the AMD side, where new hardware is only supported by the gallium drivers and requires KMS code FreeBSD doesn't have yet.
how different is the different bsd's. linux distrobutions are pretty much the same stuff just compiled with slightly different compile options of patches and aybe a few different providers of support packages, but in general once code gets merged upstream all the distros have the same codebase withing a short amount of time. is the different bsd's wildly different codebases or somethign to that effect?
is there generic vesa kms for linux or in the works?
KMS is important for security reasons, since it allows you to run the X server as a non-root user. Furthermore, many Xorg drivers are depricating and dropping UMS support. This is something long overdue (and has nothing to do with performance, FYI, since a mode-set is usually something that happens only once per boot).
I think they're pretty different with little code sharing.
Originally Posted by benjamin545
But that hasn't mattered before because X.org was running separate in userspace without tight integration with the kernel.
A good design would be that drivers declare what versions of libraries they want to use or even have a checksum that can be used.
This way it's really easy to spot what needs to be updated when changing api's.
It would be possible to make translation libraries or spot what drivers need to be changed much faster, easier and reliable.