00:21airlied[d]: karolherbst[d]: I assume it's just a ctrl call
00:21karolherbst[d]: I wonder.. because there are commands on the subchans that could be tweaked with more knowledge from the application
00:22karolherbst[d]: like especially as it mentions SCG as well
12:41karolherbst: lyudess, dakr: I went through the entire nouveau ML moderation queue, so if a loooot of emails show up, don't be surprised
12:41karolherbst: not sure if the original sent date is shown tho...
12:42karolherbst: there are also a few bug reports here and there among them
13:28karolherbst: anyway, mailing list also got migrated to mailman3
13:28karolherbst: https://lists.freedesktop.org/postorius/lists/nouveau.lists.freedesktop.org/
13:29karolherbst: well.. the archive is here: https://lore.freedesktop.org/nouveau/
13:31karolherbst: ohh, lore has a git clone thing.. interesting
16:55phomes_[d]: mhenning[d]: I can borrow a turing card later tonight and test the values from "nak: Use the hardware's max warps_per_sm value" with it if that helps
19:14mhenning[d]: I'm not especially concerned about the values being wrong, since they're reasonably well documented.
19:55phomes_[d]: Ok. I might do only a very quick check then. I just wanted to verify in case that could help land it
20:51mhenning[d]: Okay. It may possibly improve perf in some situations on affected cards but I don't know if you'll see that in practice
20:55karolherbst: ohh I also noticed that the nouveau ML is configured to never send out a mail about the bounce...