Unfortunately, I didn't apply things in multiple commits, as it was not really possible. Peter's patches wasn't cleanly applicable to the source, so I needed to make same modifications to it. Also, afterwards changed to use current Xorg API, to make it also compile.
It is true, that it contains some code that expects to have some kind of non-standard kernel features, but it's actually used as optional, if you observe ImpactI2FindCflushmode(ImpactPtr pImpact) method. It should also work with vanilla kernel.
No. It's kind of a mess. The patches are originally by Peter Fuerst <post@pfrst.de>
(see https://lists.x.org/archives/xorg-devel/2011-September/024881.html). I haven't confirmed if what is in this MR is exactly what's in those two patches from the mailing list.
The patches from the mailing list don't apply cleanly because they were pasted into an email client and sent that way. I offered to help in both cases, but I don't have hardware to test, the patches apparently have some interaction with kernel changes that never made it upstream, and I don't have sufficient motivation to untangle all this.
If someone cares enough to sort out these things and make an MR with proper patches and authorship and can confirm the patches work on real hardware, I'll be happy to press merge.
Any updates on this?
Pipeline currently fails because driver depends on XAA, which seems to be fixed in !1
Signed-off-by: Alan Coopersmith alan.coopersmith@oracle.com