01:48 martm: yeah i know this airlied is trying to make a big screw with this med stuff, i will mention that in autobiagoraphy, so he would have a broken head every time his in public places
01:49 martm: they with some irish self made gangstas, looked like were trying to bother me overseas, half a military is under this self-made heros control
01:50 martm: here we have lot of those similar male quasimodos, who never made it
01:51 martm: i call them abortion leftovers, they always tend to think that their birth faults is my mistake
01:54 martm: airlied: it seems like you can not really access stuff with your brain anyways, what dynamic IP address, crazy human being
01:55 martm: this dave airlie is a lunatic, this motherfucker sinks like the last ones when they get such a dose to him
01:56 martm: they mostly think that going against a bigger name with theyr meds stuff, is that i get some more kilos and i am not atractive to their women
01:56 martm: it's very highly documented approach of quasimodos
01:59 martm: those are very horror guys from red hat, this ajax_ and airlied i have not seen cockier neither dummier bastards in life
01:59 mooch3: dummier is not a word
01:59 martm: brittish have some good stuff generally, but i think all know
01:59 mooch3: *britihs
01:59 mooch3: *british
01:59 martm: that those guys are not the ones
02:00 mooch: can someone ban martm please?
02:00 imirkin: mooch: just use /ignore
02:01 imirkin: but really just use ignore. it avoids annoyance.
02:03 mooch: oh, sorry
02:09 airlied: imirkin: ignore only works for you
02:09 imirkin: and everyone else who has ignored ;)
02:10 airlied: yes but it means poor newbies end up seeing him
02:10 airlied: and interacting with him before we notice
02:10 airlied: we should probably just ban all of estonia
02:10 imirkin: seems a bit drastic. but iirc that's what intel-gfx did
02:10 imirkin: along with banning all irc proxies
02:10 imirkin: etc
02:10 imirkin: imho it's not worth it for just one annoying person
02:11 airlied: he produces quite a lot of noise
02:11 imirkin: apparently hates you quite a bit (just reading the logs...)
02:11 airlied: oh he pms me for days on end as well
02:14 imirkin: hakzsam: so it looks like commenting out the whole predication thing makes things a little better for elemental, but ... not much much
05:44 hakzsam: which predication thing? the image one?
08:24 xexaxo: hakzsam: yes
08:25 hakzsam: cool
12:03 RSpliet: what's the state of nouveau for drm-next?
12:37 pmoreau: RSpliet: You mean, what will be merged for 4.7, or can you build Ben’s tree against drm-next (to which the answer is no: Ben is missing some ttm and vgaswitcheroo changes)?
14:44 RSpliet: pmoreau: the former :-)
15:46 mmu_man: plop
15:47 Leftmost: The envytools docs are incomplete for Kepler+, but I assume that I'd have to do some digging with the RE tools to get a sense of what needs to change for Maxwell tess.
15:47 mmu_man: I recall that nvidia at some point finally released "some" specs, anyone has a list around?
15:47 mmu_man: IIRC they first published some docs about a table in the BIOS or something, then some Tegra specs, which I honestly couldn't care less. Is there anything else I missed?
15:48 mmu_man: or did they finally come to their senses and published everything else? :D
15:48 hakzsam: mmu_man, search for "nvidia open gpu doc"
15:48 hakzsam: Leftmost, yep, you probably have to RE the thing before
15:48 hakzsam: Leftmost, did you already use valgrind-mmt?
15:48 mmu_man: thx hakzsam
15:49 mmu_man: ah yes DCB that was the first one
15:49 Leftmost: I did not. This is entirely new territory for me, so I'm still getting a sense of how I should get started.
15:51 Leftmost: Basically as far as I've gotten in my thinking is, "Write a simple test with a basic tess program, run against proprietary with RE tools". That's as sophisticated as it's gotten. :-P
15:51 hakzsam: Leftmost, exactly :)
15:51 hakzsam: but we already have a bunch of tests for tess in piglit
15:52 hakzsam: so, I would say, the first thing is to grab both piglit and valgrind-mmt
15:52 hakzsam: find a test which passes on blob, make a trace, study :)
15:52 hakzsam: here's how to use valgrind-mmt https://nouveau.freedesktop.org/wiki/Valgrind-mmt/
15:52 Leftmost: Great, thank you.
15:52 hakzsam: and ask questions :)
15:53 Leftmost: Oh, don't worry on that count. I'll have lots of questions, and I will ask them. :)
15:53 hakzsam: cool
15:53 Leftmost: Thanks for the advice on getting started.
15:54 hakzsam: np
15:54 mmu_man: btw, what was that bug already which crashed nouveau 3 times a week and made me switch to the proprietary driver on this board...
15:54 hakzsam: Leftmost, I would suggest to play with nop.shader-test (in piglit/tests/spec/arb_tessellation_shader) or something at beginning
15:55 hakzsam: Leftmost, this one should be "simple"
15:59 mmu_man: https://bugs.freedesktop.org/show_bug.cgi?id=26980
15:59 mmu_man: oh, RESOLVED FIXED \o/
16:00 mmu_man: maybe I should try again :)
16:02 imirkin_: Leftmost: https://people.freedesktop.org/~imirkin/traces/gm107/
16:02 imirkin_: Leftmost: already some traces in there that correspond to piglits
16:02 imirkin_: sanity = sanity.shader_test, quads = quads.shader_test
16:06 Leftmost: Cheers.
16:23 pascalgp104: http://www.geforce.com/hardware/10series/geforce-gtx-1070 Pascal GTX 1070
18:29 kloofy: listen guys i am sorry for being angry again, but today i did not bother to search for keplers specs for branching but there are fermi tesla ones, we talked with karolherbst, that i am able to find from the net , how both of those archs with branching
18:29 kloofy: http://people.engr.ncsu.edu/hzhou/ispass_15-poster.pdf
18:30 kloofy: if you inspect it then there is seen that there special stack handling stuff
18:35 kloofy: https://www.youtube.com/watch?v=PQqUTigWKHY
19:49 pmoreau: RSpliet: No idea then… Probably a good chunk of those fixes https://github.com/skeggsb/nouveau/compare/linux-4.6...master, but otherwise :-/
19:57 i64: Hello, I am trying to run nouveau with a GTX960, using linux 4.6, mesa 11.2.2 and nouveau 1.0.12 (archlinux packages), but somehow I still get "(EE) Unknown chipset: NV126"
19:57 i64: Am I doing something wrong ?
19:58 imirkin_: i64: no, that's fine
19:58 imirkin_: i64: you should end up with the modesetting X driver, which is right
19:59 i64: oh thanks, so that's not the issue
19:59 imirkin_: if that message bothers you, you can uninstall xf86-video-nouveau
20:00 i64: but my X session won't run, next error is "next error is [drm] Failed to open DRM device for pci ... : -13", no idea where it's comming from
20:01 imirkin_: pastebin dmesg
20:01 i64: thank you, give me a few minutes
20:04 i64: ln
20:05 i64: oh I think I found the problem myself
20:05 i64: let me check
20:06 i64: I had an old xorg conf file specifying nouveau as driver, that was what was preventing X to start
20:06 i64: sorry about the waste of time
20:26 pmoreau: imirkin_: Would it make sense to have `BuildUtil::getScratch()` return an LValue of `size = max(given_size, 4)` if file == FILE_GPR? Or should I do it on my side each before I call getScratch?
20:26 pmoreau: s/each/each time