14:02ignapk: so I used gdb to get exact files and line numbers from the call trace https://paste.rs/eOu
14:06RSpliet: karolherbst: Fedora 33, all updates installed, NVE6 gives me this misrender
14:07RSpliet: Actually... seeing the same on my Intel IGP
14:07RSpliet: That border is wrong. Could be a gnome bug, could be a mesa bug
14:13RSpliet: ignapk: read_pll() from gk104.c should not be executed on Fermi
14:15ignapk: I'm not sure if the trace is correct about that one due to the leading question mark
14:17RSpliet: Probably not then. Same for the other question mark lines
14:18RSpliet: Got no time now, but wondering if your PMU is working at all. Did anything that may prevent firmware uploads?
14:19ignapk: I just double checked and gf100_ram_calc does call nvkm_memx_init in 976 line in ramgf100.c
14:19ignapk: hmm I'll try to find out
18:10Lyude: skeggsb: looking at nv50_dac_disable() and nv50_pior_disable(), and nv50_sor_disable(), and I noticed that they all seem to check that there's a CRTC on the given encoder despite the fact i'm pretty sure that's impossible in atomic. Do you know if that was just leftover from pre-atomic modesetting?
18:18Lyude: (for now I will assume it is and just get rid of it while I'm fixing up the help->atomic_enable/disable stuff