00:00 skeggsb: oh, i know they work, we used to use them :P
00:00 skeggsb: and sure, i don't see any reason not to use them
00:00 skeggsb: you can clear with ce too btw, though, not all dispnv50 devices have it, so twod is probably the right choice
00:01 Lyude: skeggsb: ahh-I had been curious about that, I'm assuming you just don't set a src and set the data transfer type to NONE + use remapping to set constant pixel values?
00:01 skeggsb: yep :)
00:01 Lyude: cool
00:02 Lyude: anyway-tbh if CE can do it I might try that first since we'll be making that work on fermi at some point anyway, we'll see
00:02 Lyude: wait-no, i forgot we'd still need this for tesla either way oops
00:02 Lyude: twod it is then
00:03 skeggsb: you can use ce on fermi from the kernel now if you want, just don't do it from userspace because i want to change the class :P
00:09 Lyude: skeggsb: and guessing there's no ce on tesla?
00:21 Lyude: skeggsb: also, do you have any idea if the nvidia driver actually powers the eDP panel on/off (e.g. through the GPIO that you use in nvkm_dp_init())?
00:24 skeggsb: Lyude: only from gt21x oneward
00:25 skeggsb: i *believe* so, in some circumstances anyway, i'm not too clear on what it does exactly
00:25 Lyude: cool
00:26 Lyude: skeggsb: so I'm -pretty- sure it's required if we want the panel to not do weird things sometimes, since otherwise the panel never gets shut of
00:26 skeggsb: it *may* just flick the gpio on though, if it's not
00:26 Lyude: hm, let's hope not since then we'll probably need to figure out if nvidia has a vbios script for this or not
00:28 skeggsb: i believe the power control is mostly done from the various Off/OnIntX scripts
00:28 skeggsb: there are random stuff that RM does itself though
00:28 skeggsb: is*
00:28 Lyude: i'll ask andy when i send the email about the T1-3/T12 delays
00:36 Lyude: oh yay, I -think- i'm at least close to getting the panel power working. Initial power on _seems_ to work, and link training passes reliably but the panel looks like it's stuck transmitting black video
01:04 Lyude: bingo!
01:06 Lyude: theory about us needing to power the panel on/off == completely correct (panel on/off issue no longer occurs), actual edp delays still ???, and need some minor changes, but eDP backlight support on nouveau is a success!
01:07 Lyude: that was much faster to solve than anticipated, I hope greg comes back earlier then 2 weeks
07:47 Andriamanitra: i just experienced a hang (display frozen, no reaction to mouse or keyboard) and managed to dig this up from the logs https://bpa.st/raw/4FYA
07:48 Andriamanitra: i didn't realize i was even using nouveau (i use proprietary drivers) but i guess firefox uses it?
07:49 HdkR: Oops you're using Nouveau
07:49 HdkR: Guess you're not using the proprietary driver ;)
07:50 HdkR: It's very easy for a OS update to come in and break the proprietary stack without you noticing
07:51 Andriamanitra: the only display controller that shows up as installed in manjaro settings manager is video-nvidia-450xx
07:52 Andriamanitra: but not sure if that's even relevant; do those logs mean anything to anyone here? should i be reporting a bug, digging up more logs somehow?
07:54 HdkR: Looks like the graphics engine crashed and Nouveau caught it
07:57 HdkR: `glxinfo | grep "OpenGL renderer"` will very much tell you that you're not using the proprietary blob though
08:02 Andriamanitra: hmm i guess i could look into that and maybe try to actually use those proprietary drivers that i have installed, that might resolve the issue
08:39 AndrewR: https://github.com/CRVI/OpenCLIPP - see what I found ... it makes nv50_ir_emit_nv50.cpp sad :}
19:02 AndrewR: karolherbst, I see nv_nir_compile branch, thanks :}
19:28 karolherbst: AndrewR: mind testing it and report back on the MR? thanks
19:37 AndrewR: karolherbst, rebuilding mesa with patch (in Rosa chroot)
20:19 AndrewR: karolherbst, you fixed it :}
23:01 Lyude: hooray, found more issues with this X1 extreme g2's panel and nouveau
23:01 Lyude: fun.
23:02 HdkR: :)
23:17 Lyude: honestly i'm thinking I might just try to find a different laptop, get that working then figure out what the heck we're doing wrong with this thing
23:17 Lyude: tbh the fact I've seen it fail to bring up the display in bios is definitely not encouraging, lol
23:20 Lyude: what's really bizarre though is it only seems like it happens on initial boot. if I get the lcd to come up when booting up, I seem to be able to toggle it on and off as many times as I want
23:20 Lyude: but sometimes I'll see it flicker on for a brief second and then just go black, and I don't see any kind of link training failures and the panel seems to be pretty convinced it's receiving a valid video feed
23:24 Lyude: tommorrow I'll see what i915 does, that'll probably give me some answers