Running nouveau on PPC64 is a whole experience :-) - Running from git/master crashes X : "(EE) NOUVEAU(0): No known BIOS signature found" - Removing nvidiafb and using only offb was suggested, done, but didn't change a lot of things - Commenting { "PROM", load_vbios_prom }," in nv_bios.c fixed the first issue. - I was then back to the same state I was at a few weeks ago : the screen is full of weird vertical multicolored lines. - Taking a screenshot with ImageMagick's 'import' shows a perfectly good image - Trying to go back to the console (killing X) would show a black screen - Adding "return;" at the start of the run_tmds_table function in nv_bios.c mostly fixes the issu : the screen is now totally readable, even though some artefact appear throughout the screen - Adding `Option "Randr12" "Off"' in your xorg.conf device section eventually totally fixed the issue, letting me use nouveau on a PPC64 machine !!
Curing the artefacts when in randr12 mode is the first step, so please make register dumps from within X (if seeing to type is a problem, do it over SSH), and attach them here -- one with randr12 disabled (and no corruption present), and one with randr12 enabled and the corruption present -- as follows: git clone git://people.freedesktop.org/~stuart/radeontool make ./radeontool regs > logfile Also, adding the xorg log from http://pastebin.com/m443a3b50 (or a similar newer one) to this bug as an attachment will ensure it doesn't get lost.
Created attachment 19575 [details] Xorg output
Created attachment 19576 [details] radeontools reg, without randr12
Created attachment 19577 [details] radeontools reg, with randr12
Added requested attachments. By the way, radeontool's output doesn't seem to change with or without randr
Created attachment 19639 [details] [review] Remove some reg differences shown by radeontool I'm not sure what you mean in the above comment: a diff of the two radeontool logs you attached shows some differences. Please try the attached patch and see if it improves randr12 mode. Note that this patch is made against current git, so you need to update to that to apply it; be sure to apply the "return;" at the start of run_tmds_table, and use offb, as before (the load_vbios_prom issue is now fixed upstream). If it's possible for you to test a monitor with a regular analogue vga input at some stage, that would also be very useful.
No reaction from reporter for two years now. If you still have issues please try with a recent nouveau version and open a new bugreport.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.