12:13cosurgi: Pie_Mage: so only sweat is allowed here, huh? ;)
12:15imirkin_: nah, blood and tears are good here too
12:15imirkin_: just a good idea to get them checked out
18:16andres: imirkin_: Interesting. New HDMI cable finally arrived, and it does recognize the external monitor with full resolution. So there indeed was something odd with previous one (whatever weird active component one would want to add to a hdmi<->hdmi cable for whatever reason.)
18:18andres: (still has the same issues I had with a usb-c -> hdmi cable, around cursor updates not being successful, and some other timeouts, but that's not related to the resolution problem, I think)
18:18imirkin_: gremlins in the cable
18:18imirkin_: unfortunately i don't know too much about physical cable issues with hdmi, or what causes them
18:19andres: Not worth investigating much, I'd say.
18:20imirkin_: you have the mobile TU117 right?
18:41imirkin_: i think that should be getting accel soonish
18:42imirkin_: at least the basic copy accel needed for reverse prime
18:42imirkin_: commits for it here https://github.com/skeggsb/nouveau/commits/master
18:42imirkin_: still pending firmware
18:42imirkin_: skeggsb: although you said that copy engine should be able to work without firmware?
18:53andres: I'm currently just booting it without the intel card enabled, which, as I understand, ought to make modesetting work on the external card too. But I'm getting the errors I'd mentioned a few days ago (https://gist.github.com/anarazel/c27a2ad9cbfec336a1298096917a0e11).
18:54imirkin_: andres: did you try the patches skeggsb mentioned?
18:54imirkin_: https://github.com/skeggsb/nouveau/commit/0764baeeb737771e71b204630fca57f6bd3974e4 + https://github.com/skeggsb/nouveau/commit/e259bf65680933c80e25231b7a09223a68b11029
18:55andres: Yea, that's with them applied.
18:56andres: IIRC I didn't get through boot with a usable internal display without them.
20:22Lyude: imirkin_: i'm about to come out with some more turing patches jfyi
20:23Lyude: so if you guys have any error codes from the disp core I can see if it's something I've fixed yet
20:23Lyude: assuming ben doesn't beat me to it
20:23imirkin_: have a look at andres's log
20:23imirkin_: code 0x2d
20:44Lyude: andres: what kind of laptop is this?
20:46Lyude: oh hey, ok, I have this laptop :)
20:46Lyude: and it is literally next on my todo list
20:46Lyude: andres: can you get a backtrace of this with drm.debug=0x16 debug=outp=trace
20:50Lyude: oh-ok, I think I know what the problem is her
20:50Lyude: something something lut programming gone wrong
20:51imirkin_: didn't you have a fix for that?
20:51Lyude: imirkin_: yeah but that was for resetting the lut appropriately when we clear it in the state, this one basically just amounts to us giving it an invalid lut
20:51imirkin_: we shouldn't do that =]
20:52imirkin_: hopefully not due to my 10-bit lut efforts
20:52Lyude: might be :P, we
20:52Lyude: *we'll see
21:44andres: Lyude: Heh. It's an x1 extreme gen2. Do you still want the debug output? I can get it (different monitor, but same resolution, so it should be the same result).
21:45Lyude: andres: yeah, I can get it myself but that would save me some time since I've got a lot on my plate :P
21:45andres: What might be different from "your" case is that I booted with the intel card disabled in bios.
21:45andres: Will do in a few (just have to survive this meeting first).
21:56Manoa: my xorg log says nouveau unknown chipset: NVf1 ? this is a 780 Ti he is suposed to be suported :x nouveau have no any problems in dmesg and this the latest nouveau xf86 driver 0.16
22:14andres: Lyude: https://gist.github.com/anarazel/a3bcf2736e21e6839b86b329f4bce911 - that's with 5.6 shortly before rc1, with skeggsb's linux-5.6 changes merged in
22:15andres: internal and external display showed the same picture, despite xrandr thinking they're different, and despite having different resolutions.
22:36imirkin_: Manoa: nvf1 should be very well supported
22:36imirkin_: nouveau 0.16 is from like 2008
22:36imirkin_: i think 1.0.16 would be better.
22:42imirkin_: looks like 0.0.16 was bumped to 1.0.0 in 2012.
22:44Manoa: I used wrong version ? xD
22:44Manoa: it was 1.0.16 yhe
22:45Manoa: I look again ...
22:45Manoa: thank :)
22:46imirkin_: pastebin xorg log then? coz it should be fine...
23:12Lyude: ok, interlace fixes done, xlut setmask fix done, now the laptops
23:26imirkin_: Lyude: 6-hex-digit for register id's
23:26imirkin_: Lyude: also, is it safe to read 0x640144 on nv50?
23:27Lyude: imirkin_: apparently not according to skeggsb, but we've got another solution we're looking at
23:27imirkin_: ah ok
23:27Manoa: this it's ilia: https://pastebin.com/u0wgFkJs
23:28imirkin_: Manoa: [ 129.330] (II) Module nouveau: vendor="X.Org Foundation"
23:28imirkin_: [ 129.330] compiled for 1.17.4, module version = 1.0.7
23:28imirkin_: 1.0.7 is before nvf0/nvf1 support was added
23:29Manoa: ok I delete and rebuild him
23:29imirkin_: and before i upper-cased the chipset, leading to the stupid NVf1 message instead of NVF1 :)
23:29imirkin_:definitely doesn't have OCD
23:30imirkin_: 2013. very good year.
23:33imirkin_: i was still young.
23:33imirkin_: my first contribution to upstream nouveau.
23:33imirkin_: (i had previously done stuff with renouveau back in the bad old days)
23:35imirkin_: nouveau still had a chance then.
23:35imirkin_: ooo welll
23:35Manoa: lel xD
23:35Manoa: still have chanse :)
23:36Tashtari: A chance?
23:36Manoa: whay not, I think he will work soon :)
23:37Lyude: a lot of it depends on if we ever get firmware from nvidia so we can do reclocking
23:37imirkin_: which they haven't released in like 5 years, so ... i'm not holding my breath
23:37Lyude: because until then it's not really possible to complete with the performance of the closed source driver
23:37imirkin_: every year someone comes along saying "hold your breath, it's coming real soon now"
23:37Lyude: yeah, take that for what you will
23:51Manoa: mybe the intel cards give effect on nvidea to give the firmware ?
23:52Manoa: intel make they own distribution in linux, I think intel can make a strong presence in linux
23:54Manoa: it amazing, less than year now for RTX cards and you can drive 2000 serie alredy :)
23:54Manoa: only one year to programming driver for new card I think it grate