For the cases I have seen, modesetting works better than nouveau, both regarding speed and suspend-resume (i.e with GPU: GT218M [NVS 3100M] in Bug 31695, as well as with GTX750 [GM107] on my workstation) Guiseppe also told he experienced modesettign working better on the one nvidia card tried. However, this is too little data to make a decision, so we think we better gather results and ideas here.
CC: (none) => bugsquad, kernelAssignee: bugsquad => qa-bugs
Created attachment 13814 [details] patch to use modesetting instead of nouveau for automatic detecting I add a patch for using modesetting instead of nouveau for "GeForce 420 to GeForce 630" and as a fallback for "GeForce 630 to GeForce 920" (nvidia470), and "GeForce 635 and later" (nvidia-current).
CC: (none) => ghibomgx
Created attachment 13815 [details] SPEC file with modesetting fix As reminder here is the corresponding SPEC file (final version should have the patch moved to git ldetect-lst upstream). Resulting packages can be tested (manually, sorry) from here: https://copr-be.cloud.fedoraproject.org/results/ghibo/mageia9-bonus/mageia-cauldron-x86_64/05893860-ldetect-lst/
not a qa bug
Assignee: qa-bugs => pkg-bugs
ah yes, I was just thinking qa people test a lot of hardware = experience. We should write somewhere to tip users to try both nouveau and modesetting. Until recently I myself thought modesetting was only a parameter... Release Notes or Errata? --- Len reports on qa list that modesetting do not work well at all with NVIDIA GP102 [GeForce GTX 1080 Ti] :(