04:37millerthegorilla: Hi, I've just installed the kernel from kernel.org of 4.1-rc6 and am using a gm107 (nvidia gtx 750ti). The xserver is working but I am getting a problem with text blinking and missing. Does anyone know if there is a fix?
04:49millerthegorilla: although I'll need to restart to be sure I think I solved the problem. I turned off sub-pixel order in the font settings.
08:53Prince781: what is nouveau PGR?
08:54Prince781: as in, what does it stand for...
08:54Prince781: if anyone knows
08:56hakzsam: it's probably PGRAPH which is the main engine for 2D/3D
08:58flipmess: Prince781 << 17:56 <hakzsam> it's probably PGRAPH which is the main engine for 2D/3D
08:58Prince781: alright, so what does it mean if it fails to init? (from dmesg) nouveau E[ PGR][0000:02:00.0] init failed, -16
09:00hakzsam: Nouveau fails to init the card which is somehow bad... what card do you have?
09:00Prince781: If it's the main engine for 2D/3D, I guess that explains why I'm getting (in Xorg.log) NOUVEAU(G0): Failed to initialise context object: 2D_NVC0 (0)
09:01Prince781: And I have a geforce gt 750m
09:01Prince781: lspci: NVIDIA Corporation GK107M [GeForce GT 750M] (rev a1)
09:01Prince781: I'm on the Dell XPS 15z.
09:01Prince781: It's setup with an integrated (intel) and discrete (nvidia)
09:01hakzsam: could you please paste full dmesg logs?
09:02Prince781: One sec.
09:03hakzsam: oh it's optimus, I'm not enough comfortable with that, so I'm not going to help a you a lot probably :)
09:04Prince781: hmm..I enabled tracing, so the file is large (ignore the VBIOS tracing), so it's too large to upload to pastebin
09:04hakzsam: disable tracing then
09:07Prince781: should I enable debugging?
09:07Prince781: for nouveau
09:07Prince781: or no
09:08hakzsam: no, disable all tracing in a first time
09:11Prince781: dmesg: http://pastebin.com/8VjFSTCN
09:11Prince781: also this is in /etc/modprobe.d/modules.conf:
09:12Prince781: options i915 modeset=1
09:12Prince781: options nouveau pstate=1 config=NvGrUseFw=1
09:14hakzsam: okay, your issue has already been reported here https://bugs.freedesktop.org/show_bug.cgi?id=70354
09:14hakzsam: unfortunately, this doesn't seem to be fixed
09:14Prince781: so what is the cause of the problem, generally?
09:15hakzsam: no idea, but you could try to apply the patch attached to the comment 21
09:15tobijk: Prince781: it looks to me that the GDDR5 memory is causing this somehow
09:17Prince781: Well, does this mean that the GT 750m is not supported by nouveau, then?
09:17tobijk: Prince781: we dont grant any "support", it may work or not ;-)
09:18Prince781: ah, I see
09:18Prince781: Well, I'll try the patch, then.
09:26Prince781: Hmm...the file it patches (nvkm/engine/graph/nve4.c) no longer exists?
09:27Prince781: I've downloaded linux-4.0.5
09:39tobijk: Prince781: its nvkm/engine/gr/gk104.c
20:30airlied: skeggsb: since imirkin isn't around, can you review the mesa patch I sent :)
20:31airlied:has modesetting/modesetting combo on intel/nouveau sort of doing reverse prime with glamor
20:59skeggsb: airlied: it seems sane to me, i do wonder why we didn't do that in the first place though
21:08airlied: skeggsb: been like that since calim introduced that code 3 years ago
21:09airlied: probably nobody noticed until now
21:10airlied: skeggsb: going to translate "seems sane to me" into "Acked-by" :)
21:11mlankhorst: and the wonder why to reviewed-by? :p
21:12airlied: mlankhorst: oh I meant to ask, are cross-driver fences and i915 ever going to become a thing? :)
21:13mlankhorst: cross driver fences are already sort of supported by nouveau and radeon, but I thought the tendency was towards moving to explicit sync
21:14airlied: mlankhorst: yeah I suppose it is, just wondering for making the prime sharing not suck :)
21:15mlankhorst: skeggsb seems to have dropped my patch that removed the wait during command submission on other-driver fence completion
21:15mlankhorst: so it'll be explicit waiting anyway..
21:25skeggsb: mlankhorst: which patch?