There are 2 driver versions for current nvidia GPUs. Short lived branch (effectively beta) 387.XX Long lived branch (effectively LTS) 384.XX http://www.nvidia.com/object/unix.html
We used to have both, but it appears 384 has gone missing. 387.XX has issues resulting in blacked out windows (Every DE) and sometimes the entire screen (KDE) while improving performance in some areas (games such as dying light or steambox DE) 384.XX is stable and doesn't have these issues of blacking out windows and screen. (simple test would be runescape-launcher https://bugs.sabayon.org/show_bug.cgi?id=5426 *hint *hint) We also have a few packages that are out of date in our repo that do not match with (or exist within) upstream gentoo libblockdev 0.13 -> 2.14 bintutils 2.25 (is now masked? how do we push people to the unmasked version since we are slotting 2.25 and 2.28.1? any ideas? ) zfs & zfs-kmod 7.2 -> 7.3 (meh, normal) Samba (not really sure what to do here. CVE-2017-15275,14746,11103, list goes on an on) we need to be on 4.5.14, but that doesn't cure all the CVE on samba's page but switching to 4.7.2 is just switching to a list of unknown vulnerabilities. Is it really beneficial to go from 4.5 to 4.7? at the very least we should be 4.5.14, but beyond that I'm not sure how we should proceed or the effort it would take to constantly hop as samba updates at a pretty fast pace.. Looking for some insight on this topic. efivar 0.21 -> 31 efibootmgr 0.12 -> 15 multipath-tools 0.5.0-r99 (are we doing something special with this package that we need the -r99? are there some patches we need specifically for the distro or should we pass them upstream? if not should we update 6.4-r1 (gentoo stable) or 0.7.3 (gentoo unstable)? Thanks, Darksurf
