05:55octet: Journaelctl -xe with Nvidia Turing 1660ti: Nouveau 0000:01:00.0: unknown chipset (168000a1) Nouveau: probe of 0000:01:00.0 failed with error -12
05:55octet: I read that nouveau has Turing support as of kernel 5 which I am running, so I am curious why I get this error message and flashing cursor
10:12karolherbst: octet: ohh, I guess there is no support for the 1660 ti yet
10:12karolherbst: skeggsb: ?
17:31letterrip: hi all, any suggestions on how to debug 'mouse pointer corruption'? The mouse pointer is displayed as a roughly 1 inch square of red on my screen instead of an arrow
17:31letterrip: with the 'tip' apparently the upper left hand corner of the square
17:32letterrip: the mouse displays normally with the VESA (nomodeset)
17:32letterrip: so I suspect it might be nouveau related
17:39diogenes_: letterrip, wayland?
17:39letterrip: x11 I'm pretty sure
17:40letterrip: whatever default is for disco
17:40diogenes_: what if you change the size of the pointer and also the theme?
17:42letterrip: i think the theme doesn't matter, I'm unaware of how to change the pointer size
17:43letterrip: i'll see if i can change it in the xfce wm settings
17:44diogenes_: yes there is setting to change pointer size and pointer theme in xfce settings
17:44diogenes_: settings > mouse and touchpad > theme
17:45letterrip: no change, tried all themes, all mouse pointers, and a variety of sizes
17:46diogenes_: weird, never came across such a bug
17:46diogenes_: is compositor enabled?
17:46letterrip: i'm a special snowflake bug.. fun
17:46letterrip: nope compositor disabled
17:46letterrip: it causes crashes if enabled
17:47diogenes_: hmm, that's special case indeed XD
17:47diogenes_: maybe you just need a newer kernel
17:47letterrip: 5.0 kernel
17:48letterrip: latest mesa
17:48diogenes_: or it's too new
17:48letterrip: (well not currently latest mesa, but with latest mesa also)
17:48letterrip: have tried 4.x kernels
17:48letterrip: i think this is present in old distros i've tried also
17:49diogenes_: that;s why i like having the optimus systems so i run all the basic stuff with intel graphics and the heavy weight one with nouveau
17:50letterrip: this has in an intel 945gm but it has same memory location and apparently isn't detected - might be disabled in bios?
17:50letterrip: not sure
17:50diogenes_: look in bios
17:50letterrip: the bios is extremely simple - no options for anything but boot order
17:50diogenes_: there shold be the otion which graphics to use by default, onboard or discrete
17:50letterrip: so would have to disasseble the bios and fiddle with bits
17:52letterrip: not sure if i'm hardcore enough to try and do a custom rom :)
17:53letterrip: er custom bios .. not rom
17:54letterrip: i may be misunderstanding my dmesg
17:55letterrip: is there a way to try and force it to use the 945gm as a kernel parameter and disable the noveau?
17:55diogenes_: letterrip, lspci -nnk | grep VGA -A3
17:57diogenes_: and: lsmod | grep i915
17:58diogenes_: yeah, no intel onboard graphics
17:59letterrip: so is 945gm something different then?
17:59diogenes_: could be
17:59diogenes_: we can try a trick to boost the performance though
17:59diogenes_: and maybe it will have some positive impact
18:00diogenes_: edit: /etc/default/grub
18:00diogenes_: line: GRUB_CMDLINE_LINUX_DEFAULT="nouveau.config=NvClkMode=15...."
18:01diogenes_: save, close, update-grub, reboot
18:01letterrip: will give it a try
18:02diogenes_: ok, now i need to go, if something positive improvement happens, hit me up later
18:02letterrip: will do
18:02diogenes_: also with this option, run: sensors to notice if your gpu doesn't get overheated
18:02diogenes_: if it does then remove it
19:02Hijiri: got a crash again, this is what the output of dmesg looks like https://pastebin.com/NZEXYk1a
19:04Hijiri: it happened after playing a game for awhile
19:20karolherbst: Hijiri: uff, essentially that means out of VRAM :/
19:32Hijiri: it was a much shorter play than usual though, and only things else open were web browsers
19:32Hijiri: could chromium consume all my vram?
19:33Hijiri: the game itself is also pretty old and shouldn't be that demanding
19:33karolherbst: might be
19:33Hijiri: is it possible that VRAM might not be reclaimed from previous programs I had open
19:33karolherbst: could also be that too many commands are submitted to the GPU
19:34karolherbst: not quite sure which of the two issues that was
19:34karolherbst: Hijiri: pretty low I think, normally all the memory should get released
19:34Hijiri: I don't really feel like it's VRAM, since I've played minecraft for extended time without crashing
19:34karolherbst: but we also don't really track the current VRAM usage
19:34karolherbst: we should probably start doing that
19:34Hijiri: which would use more VRAM probably
19:36Hijiri: this game supports software rendering for some parts so I could turn that on and hope it doesn't crash
19:36Hijiri: I've just played it without crashing for a long time since the last crash I reported (which had different error messages)
19:36Hijiri: and that other crash went away I think when I stopped using the turn vsync off mesa env variable
19:38Hijiri: forgot why I brought that up, it's because software rendering didn't prevent the other crash, but this crash seems different
19:43Hijiri: so there isn't a way currently to check VRAM usage on nouveau?