04:11 karolherbst: Jayhost: no idea, I kind of just check how the nvidia driver does it's stuff and try to get nouveau do to something similiar
04:46 Umeaboy: Hi! The sched error for GTX cards, has it been resolved yet?
04:46 Umeaboy: The kernel 4.1.15 spews out sched error during boot.
04:47 Umeaboy: Then the computer freezes.
04:50 karolherbst: Umeaboy: what gpu and dmesg
04:51 karolherbst: Umeaboy: also, sched error is more of a effect than the actual cause
04:53 karolherbst: Umeaboy: also before we go into this issue, try out with kernel 4.4
04:56 Umeaboy: karolherbst: The computer in question is an ASUS N550JK (European model) with a GeForce GTX 850M graphics card together with Intel. I managed to make a workaround to make the error from not showing, but my dist Mageia doesn't have 4.4 yet.
04:57 Umeaboy: At least not in the stable version.
04:57 Umeaboy: The development version has it.
05:00 karolherbst: so it is a maxwell one
05:01 karolherbst: you really want to try it out with a 4.4 kernel
05:01 karolherbst: and what workaround do you use?
05:05 Umeaboy: karolherbst: Adding modprobe.blacklist=nouveau to the kernel boot command line and then ran echo options nouveau modeset=0 | sudo tee -a /etc/modprobe.d/nouveau.conf and finally update-initramfs -u
05:06 karolherbst: ahh I see
05:06 karolherbst: so you just disabled nouveau
05:06 karolherbst: okay
05:06 karolherbst: Umeaboy: so nouveau doesn't show up inside lsmod?
05:09 Umeaboy: I think it does. Haven't checked recently.
05:11 Umeaboy: But the problem wasn't about lsmod from the beginning.
05:11 Umeaboy: I could get to any tty to check lsmod.
05:12 karolherbst: yeah, but if nouveau loads at boot time and this causes X to not start, something is really wrong
05:12 Umeaboy: It's the GPU driver according to a thread in the archlinux forum.
05:13 Umeaboy: So therefor I was wondering if it has been updated at all in the 4.4-kernel.
05:13 karolherbst: yeah, I know, but you also have an intel gpu, which drives your display, so nouveau shouldn't affect a thing
05:13 karolherbst: and yeah, nouveau gets updates in each kernel version usually
05:14 karolherbst: what we need is the error output first. It could be that some script messes something up though, and this causes X to not start or something
05:14 karolherbst: but without nowing what gets printed, we can only guess
05:20 Umeaboy: karolherbst: https://bbs.archlinux.org/viewtopic.php?id=199578
05:22 karolherbst: Umeaboy: is this 100% what you get or something which looks kind of what you get?
05:23 karolherbst: if it isn't 100% the same, it's useless
05:23 karolherbst: also this is from half a year ago
05:23 karolherbst: with current kernel maybe it just works, because issues might get fixes
05:23 karolherbst: *fixed
05:25 karolherbst: the "failed to load fuc409c" bothers me though
05:26 karolherbst: Umeaboy: do you have by any chance NvGrUseFW enabled or something?
05:31 Umeaboy: karolherbst: As I said, I'm no expert....... so I wouldn't know as I can't reproduce it until the installation of Mageia 6.
05:32 karolherbst: Umeaboy: can you give me the output of:
05:32 karolherbst: find /etc/modprobe.d/ -exec grep nouveau {} +
05:32 karolherbst: cat /proc/cmdline
05:32 Umeaboy: When I get home.
05:32 karolherbst: ahh okay
05:33 karolherbst: though chances are that it may just work with 4.4 though
05:47 Umeaboy: Yeah. I think Mageia 6 comes in April.
06:10 RSpliet: mwk: does graph/gf100-ctxctl/strand.txt (still) exist somewhere?
06:28 RSpliet: I don't like how HOST_IO_INDEX no longer seems to exist on GK208B
06:45 Umeaboy: karolherbst: I'm home now. Be back in a jiffy.
06:47 karolherbst: Umeaboy: k
06:48 Umeaboy: karolherbst: argument to "-exec" is missing.
06:48 karolherbst: find /etc/modprobe.d/ -exec grep nouveau {} +
06:49 RSpliet: ah, got it - _V5
06:49 RSpliet: *_V4
06:50 Umeaboy: karolherbst: http://pastebin.com/nxvpNw6t
06:50 Umeaboy: What where you looking for?
06:50 Umeaboy: I added the commented line.
06:50 Umeaboy: In case you wonder.
06:50 karolherbst: Umeaboy: remove "options nouveau modeset=0" and if booting still works
06:50 karolherbst: *see
06:51 Umeaboy: OK.
06:51 karolherbst: blacklisting should be enough for now
06:51 Umeaboy: By for now.
06:55 Umeaboy: karolherbst: Still working.
06:55 karolherbst: Umeaboy: very good
06:55 Umeaboy: Now what?
06:55 karolherbst: Umeaboy: now just modprobe nouveau and see what happens
06:56 Umeaboy: No error and a new line.
06:57 karolherbst: new line in dmesg?
06:58 karolherbst: ahh in the console
06:58 karolherbst: I just need dmesg now
06:58 Umeaboy: The sched errors are back!
06:58 karolherbst: yeah
06:58 karolherbst: but X still runs, right?
06:59 karolherbst: but I think I know what causes the issues
06:59 Umeaboy: http://pastebin.com/4VftwHJa
06:59 Umeaboy: Yeah.
06:59 karolherbst: stupid X
07:00 karolherbst: Umeaboy: can you uninstall the nouveau X driver? I have no idea how the package name is for you
07:00 Umeaboy: I think so.
07:00 karolherbst: Umeaboy: but it contains the /usr/lib/xorg/modules/nouveau_drv.so file
07:00 karolherbst: or something like that
07:01 karolherbst: ohh wait
07:01 karolherbst: modesetting is used for maxwell anyway
07:01 karolherbst: forget about that then
07:03 karolherbst: Umeaboy: is the nvidia card at 01:00.0 in lspci?
07:04 karolherbst: Umeaboy: and do you have the dummy xorg driver installed?
07:04 martm: i think i screwd up a bit, i always forgat that on radeon cbranch takes an offset to to jump to as immediate, it looks to be a constant, allthough there should not be any reason why it should be
07:05 martm: so in hw land
07:05 Umeaboy2: karolhoberst: Uninstalling the nouveau driver made my computer extremely slow.
07:05 karolherbst: Umeaboy2: it should have no effect :/
07:06 Umeaboy2: I can't click with my mouse pointer either.
07:06 karolherbst: which package did you remove?
07:06 Umeaboy2: x11-driver-video-nouveau
07:06 karolherbst: I think X is simply upset about the nvidia card :/
07:07 karolherbst: ohh do you know which version it has?
07:08 Umeaboy2: 1.0.11
07:09 karolherbst: ahh
07:09 karolherbst: then the nouveau ddx still was used
07:09 karolherbst: okay
07:09 karolherbst: can you reboot?
07:09 Umeaboy2: Now X won't start.
07:09 karolherbst: aha
07:09 karolherbst: then something is odd with X itself
07:09 karolherbst: k
07:09 karolherbst: can you give me the X log?
07:10 karolherbst: it is in /var/log/Xorg.0.log
07:10 Umeaboy2: How can I paste it from tty?
07:10 Umeaboy2: Only option at the moment.
07:11 karolherbst: ssh?
07:11 karolherbst: otherwise cat /var/log/Xorg.0.log | some_upload_script
07:11 karolherbst: is there a wgetpaste package?
07:11 karolherbst: cat /var/log/Xorg.0.log | wgetpaste
07:11 karolherbst: or any other paste application you might have
07:11 Umeaboy2: Yes.
07:11 RSpliet: how is this even... karolherbst: we're doing something wrong :-P
07:12 karolherbst: RSpliet: totally
07:12 RSpliet: I have here a GT640 and a GTX 650
07:12 RSpliet: there's basically one difference between the two cards. GT640 is DDR3, GTX650 is GDDR5
07:12 karolherbst: which 640?
07:12 RSpliet: NVE7
07:12 karolherbst: yeah, which of them? :D
07:13 RSpliet: the 1782MHz memory clock one
07:13 karolherbst: k
07:13 RSpliet: anyway, 1 GPC, 2 SMs, it's the same chip
07:13 karolherbst: okay
07:13 RSpliet: on the same machine running openarena, I get 225fps for the DDR3 GT640, and 175fps for the GDDR5 one
07:13 RSpliet: both at max clocks
07:13 RSpliet: (which for the GDDR5 card is higher)
07:13 karolherbst: ohh the newer 640 DDR3 is just faster than the OEM 640 DDR3
07:14 karolherbst: yeah, they seem equal
07:14 karolherbst: let me check
07:15 karolherbst: the 650 has gddr5 and slightly higher clocks
07:15 karolherbst: odd
07:15 RSpliet: they are equal, only the DDR3 one has 28.5GiB/s bw, whereas the GDDR5 one has 80GiB/s
07:15 karolherbst: both fully recklocked?
07:15 RSpliet: yep
07:15 pmoreau: Wow!
07:15 karolherbst: and the AC line shows a higher core clock for the 650?
07:15 pmoreau: Interesting
07:16 RSpliet: karolherbst: oh... I was only looking at the mem line, yes you're right, and I should've known
07:16 RSpliet: this is your voltage issue
07:16 karolherbst: :D
07:16 Umeaboy2: karolherbst: http://paste.pound.org/show/g4nQOPSkMDMUp9ecQJ/
07:16 RSpliet: thanks for being a helpdesk :-P
07:17 RSpliet: I'll go and pull in your patches to see how much a difference they make
07:17 karolherbst: RSpliet: well you could try out my branch
07:17 karolherbst: RSpliet: stable_reclocking_kepler_v2
07:17 karolherbst: based on 4.4
07:17 RSpliet: 4.4 is the magic word I was looking for
07:17 karolherbst: and boot with nouveau.config=NvBoost=2
07:17 karolherbst: otherwise the clocks are limited
07:18 RSpliet: (as I also have a hacked-up LITMUS^RT tree and some cruft of my own)
07:19 RSpliet: do I need all 18 patches?
07:19 Umeaboy2: How do you make the text larger in tty?
07:19 karolherbst: I have a lot nwertok issues today, so I might just disappear for a few seconds somtimes :P
07:19 karolherbst: RSpliet: kind of
07:19 RSpliet: okay
07:19 Umeaboy2: I think I wrote a character wrong.
07:20 karolherbst: yep
07:20 karolherbst: b
07:20 karolherbst: I think
07:20 karolherbst: ohh
07:20 karolherbst: no idea
07:20 karolherbst: Umeaboy2: paste.pound.org doesn't exist for me
07:20 RSpliet: karolherbst: oh you don't have those patches in a kernel tree do you>
07:20 karolherbst: RSpliet: no
07:21 Umeaboy2: Sorry. Missed a -python after pound.
07:21 Umeaboy2: Then the link should work.
07:21 karolherbst: "Sorry, but the page you requested was not found on this server."
07:22 Umeaboy2: add an f between the g and 4
07:23 karolherbst: gf4nQOPSkMDMUp9ecQJ
07:23 karolherbst: still won't work
07:25 Umeaboy2: What the heck?
07:29 karolherbst: Umeaboy2: found the error=
07:29 Umeaboy2: Okey.
07:29 Umeaboy2: What was it?
07:29 karolherbst: no, I meant to write ?
07:30 Umeaboy2: Wait...... I'm starting my stationary PC.
07:37 Umeaboy2: http://paste.pound-python.org/show/Xw3IDqDLOT4XpKnLOPMF
07:37 Umeaboy2: I checked that the link works.
07:37 Umeaboy2: karolherbst
07:42 Umeaboy2: karolherbst: Well?
07:42 Umeaboy2: Your thoughts?
07:44 Umeaboy2: Your silence alarms me.
07:50 pmoreau: He might be afk
07:51 RSpliet: Umeaboy2: or karolherbst is just having a bad internet connection
07:52 Tom^: i see only three possible outcomes, he is either eating, comet struck his house or he is focusing on the coding he is doing atm.
07:52 RSpliet: Tom^: if you read up ~25 lines, you'd see
07:52 RSpliet: "< karolherbst> I have a lot nwertok issues today, so I might just disappear for a few seconds somtimes :P"
07:52 Umeaboy2: I'd say #2. :)
07:53 Tom^: then its more aligned towards #2 indeed.
07:53 RSpliet: not *quite* sure what a nwertok, but it's likely networking-related
07:53 karolherbst: yeah, my internet connection really sucks today
07:54 RSpliet: ^ classic
07:54 Tom^: ah excess flood, i forgot the tsunami possibility.
07:55 karolherbst: ..
07:55 RSpliet: karolherbst: tell your network stack not to queue up all your sent messages, that'll get you kicked here :-P
07:56 Tom^: synchronous messages, not async.
07:56 karolherbst: ohh my IRC client is smart enough
07:56 karolherbst: it already notifies me when there is no reply for 5 seconds
07:57 karolherbst: Umeaboy2: you don't have the intel xorg driver isntalled?
07:57 RSpliet: Umeaboy2: is there a reason you have an xorg.conf ?
07:58 karolherbst: X really _can't_ handle the fact, that the nvidia card has no display
07:58 karolherbst: and with _can't_ I mean, it really can't
07:58 karolherbst: :D
08:00 Umeaboy2: RSpliet: Don't know. Mageia have implemented Wayland.
08:00 karolherbst: is somebody messing around with the routing tables or what? :/
08:00 karolherbst: Umeaboy2: k, then give us the content of your /etc/X11/xorg.conf
08:00 RSpliet: Umeaboy2: you don't seem to be using Wayland
08:00 karolherbst: Umeaboy2: also check if you have the intel Xorg driver installed
08:00 karolherbst: x11-driver-video-intel I suppose
08:02 Umeaboy2: No, it isn't installed. I think it got removed when I uninstalled the nouveau driver.
08:03 karolherbst: ....
08:03 martm: Description
08:03 martm: PC = S0.u; S0.u is a byte address of the instruction to jump to
08:03 martm: Description
08:03 martm: PC = PC + signext(SIMM16 * 4) + 4
08:04 Umeaboy2: Checked the requirements for it, but I couldn't find the nouveau driver in the list.
08:04 karolherbst: Umeaboy2: yeah, there is something odd
08:04 martm: actually yeah, it is possible to program a pc for cbranches if correct sequence is used
08:04 karolherbst: Umeaboy2: install x11-driver-video-intel then
08:04 Umeaboy2: Okey.
08:05 karolherbst: Umeaboy2: you won't need x11-driver-video-nouveau anytime soon anyway
08:05 karolherbst: because it can't handly your maxwell card anyway the right way
08:05 Umeaboy2: And then reboot?
08:07 karolherbst: yeah
08:08 Umeaboy2: Now I got X working again after installing the Intel driver.
08:08 Umeaboy2: What next?
08:09 karolherbst: your x config: /etc/X11/xorg.conf
08:11 martm: then i did not screw , a minor modifcation to the pseudo makes it work again
08:14 Umeaboy2: http://paste.pound-python.org/show/dm66Q3wn10ZSZRXrBtSm/
08:14 Umeaboy2: My Xorg.conf
08:14 karolherbst: link doesn't work again
08:15 Umeaboy2: Hmmmmmm.
08:16 Umeaboy2: http://paste.pound-python.org/show/dm66Q3wn1OZSZRXrBtSm/
08:16 Umeaboy2: Made a 0 instead of an O.
08:16 Umeaboy2: Hard to tell sometimes.
08:17 karolherbst: my network is still shitty :/
08:17 karolherbst: around 25% package drop rate...
08:17 karolherbst: actually 40%
08:21 Umeaboy2: Need something more?
08:22 Umeaboy2: Ouch!
08:24 karolherbst: yeah I need stable network...
08:25 karolherbst: Umeaboy2: add this to your xorg.conf: https://gist.github.com/karolherbst/b66678391e305a2ae236
08:28 Umeaboy2: And then?
08:28 Umeaboy2: Reboot again?
08:31 karolherbst: Umeaboy2: yeah, reboot is fine
08:31 Umeaboy2: Okey.
08:32 RSpliet: karolherbst: my machine locks up in the highest perflvl as soon as I try to do any serious rendering
08:32 karolherbst: RSpliet: with my branch?
08:32 Umeaboy2: Rebooted. Now what?
08:32 RSpliet: well, with my branch plus your patches
08:32 karolherbst: Umeaboy2: modprobe nouveau
08:32 karolherbst: Umeaboy2: lsmod | grep nouveau
08:32 karolherbst: RSpliet: mhh
08:32 karolherbst: RSpliet: dmesg
08:32 RSpliet: https://github.com/RSpliet/kernel-nouveau-nv50-pm/commits/ctxswitch_litmus
08:33 RSpliet: dmesg empty
08:34 RSpliet: (pretty sure, as the magic sysrq warnings did end up in my logs)
08:35 karolherbst: RSpliet: fun fact, nvatiming shows much higher ctx clocks on nouveau than the blob for me :/
08:36 karolherbst: RSpliet: odd, so a lockup without anything in dmesg
08:36 Umeaboy2: karolherbst: http://pastebin.com/FjpqLaax
08:38 RSpliet: didn't you hide a cstate debugfs entry somewhere
08:38 RSpliet: karolherbst: didn't you hide a cstate debugfs entry somewhere?
08:39 karolherbst: not anymore
08:39 RSpliet: okay
08:39 karolherbst: RSpliet: only happens with the 650?
08:40 karolherbst: Umeaboy2: k, then dmesg
08:40 RSpliet: it's the only GDDR5 card I have
08:40 karolherbst: Umeaboy2: X doesn't seem to clain the nvidia gpu anymore, so this is out of the way
08:40 RSpliet: oh, but: yes
08:40 karolherbst: RSpliet: k, then after you go to 0xf, pstate and sensors
08:43 karolherbst: RSpliet: did you add the vbios of the gddr5 card into the repository?
08:43 Umeaboy2: karolherbst: http://läste.pound-python.org/show/ApMJyXqaTRo4Gmsf6ha0/
08:43 RSpliet: karolherbst: http://fpaste.org/333078/70234011
08:44 Umeaboy2: paste
08:44 Umeaboy2: Dag nam it!
08:44 Umeaboy2: Stupid Swedish keyboard.
08:45 RSpliet: karolherbst: and no, I didn't add it to the repo, don't have access to that from the machine I work on right now, let me e-mail it to yoy
08:45 RSpliet: *you
08:49 Umeaboy2: Anything else?
08:49 karolherbst: RSpliet: k
08:50 karolherbst: Umeaboy2: there is no nouveau in it for some strange reasons
08:50 karolherbst: Umeaboy2: did you modprobe nouveau before getting that dmesg?
08:50 Umeaboy2: Yes.
08:51 Umeaboy2: But we uninstalled nouveau, remember?
08:51 karolherbst: the X driver
08:51 karolherbst: not the kernel module
08:52 karolherbst: RSpliet: .... nooooo not such a gpu again
08:52 karolherbst: utter pain
08:52 Umeaboy2: I sad now during the uninstallation of nouveau that urpme (the removal/uninstallation command in Mageia) orphaned x11-driver-video-intel as well.
08:52 RSpliet: karolherbst: just a GK107
08:52 Umeaboy2: It was in the list of many.
08:53 karolherbst: RSpliet: no, the vbios is messed up
08:53 karolherbst: big times
08:53 RSpliet: messed up as in the file is damaged?
08:53 karolherbst: RSpliet: no base clocks, 3 cstates in total
08:53 karolherbst: well dynamic recklocking would jump between 3 clocks on this card...
08:54 karolherbst: ahh
08:54 karolherbst: RSpliet: your 650 requires this patch: https://github.com/karolherbst/nouveau/commit/6de3767e479a77fe3d39135f0d16b28bfe23ea94
08:54 karolherbst: that's why stock nouveau can't volt
08:55 RSpliet: I pulled that in
08:55 karolherbst: RSpliet: can you run the blob on the machine with the 650?
08:55 RSpliet: no, but I might be able to take it home over the weekend to do some tests
08:55 karolherbst: I want to know what voltage we read out
08:55 karolherbst: ahh k
08:56 karolherbst: it would be really interessting how nvidia handles that gpu
08:56 karolherbst: but 1.1V sounds right
08:56 karolherbst: Umeaboy2: yeah, but you should still have the kernel module
08:56 Umeaboy2: I'll switch to my computer now that X is working. I also have the stationary running just in case.
08:57 karolherbst: and after you modprobe nouveau
08:57 karolherbst: you should get some nouveau stuff in dmesg
08:57 Umeaboy2: Okey.
09:00 karolherbst: RSpliet: but how does the machine looks up? Just after switching to 0f and add heavy load, the screen freezes?
09:00 karolherbst: but nothing appears in dmesg
09:00 karolherbst: and ssh and stuff still works?
09:00 RSpliet: karolherbst: screen freezes after rendering the first real frame of OpenArena
09:00 RSpliet: consistently, doesn't happen in perflvl a
09:01 karolherbst: mhh
09:01 karolherbst: oddd
09:02 RSpliet: system responds to magic sysrq, so I assume the rest all still works as well
09:03 RSpliet: ah, here's the dmesg messages piling up
09:03 RSpliet: starts with a trap, then GPU lockup, then fifo spits sched errors with reason ctxsw timeout
09:04 RSpliet: in other words, gr is just in over it's head
09:05 RSpliet: karolherbst: system responds to magic sysrq, so I assume the rest all still works as well
09:05 RSpliet: ah, here's the dmesg messages piling up
09:05 RSpliet: starts with a trap, then GPU lockup, then fifo spits sched errors with reason ctxsw timeout
09:05 RSpliet: in other words, gr is just in over it's head
09:05 karolherbst1: mhh
09:05 karolherbst1: maybe I parse the table wrong then
09:05 karolherbst1: would be good to know what nvidia does
09:06 RSpliet: yes I'll see if I can generate some trace for you during the weekend
09:06 karolherbst1: noo, that won't be needed
09:06 karolherbst1: I just want to know what voltage the nouveau code reads out while the blob is on max perf
09:07 RSpliet: oh ok
09:07 RSpliet: still
09:07 RSpliet: weekend
09:08 karolherbst1: k
09:53 Jayhost: karolherbst Havn't crashed yet. Awesome.
10:02 karolherbst: Jayhost: :)
10:02 karolherbst: Jayhost: how is the performance?
10:20 Jayhost: karolherbst. It was great I played RE4 on Dolphin for some half hour
10:24 Yoshimo: nice to hear
10:25 Yoshimo: a lot more stable than karols connection
10:42 karolherbst: ... I just wrote "yeah, my connection is really awesome today"
10:42 karolherbst: Jayhost: but good to know that my stuff also works on maxwell somehow
10:42 karolherbst: Jayhost: how much more perf do you get with memory recklocked?
12:31 karolherbst: RSpliet: nvbios says "Detected ram type: DDR2", does this mean anything or will nouveau detect it right?
12:33 karolherbst: Jayhost: can you also test if the other pstates are stable?
12:33 Jayhost: Yes
12:34 Jayhost: karolherbst was held up at postoffice. Now to install motherboard
12:34 karolherbst: Jayhost: ohh you got a new one?
12:36 Jayhost: Yes, A new used one. It should save me years of time on rebooting.
12:37 karolherbst: RSpliet: I am sure nouveau sets the wrong VIDs on the 650, because the table is so simple, that I can't imagine anything else being wrong. 1.137V is the highest the vmap table allows and 1.1 should be stable enough :/
12:37 karolherbst: Jayhost: :D yeah, bios/uefi implementations usually suck big time
12:43 Jayhost: Eh looks like I need an 8 pin power supply
12:47 karolherbst: for the gpu?
12:54 Jayhost: karolherbst the board I guess. I don't know much. Just read the board has revisions with 4 pin and 8 pin
12:58 Jayhost: Internet says 4 pins is fine so I'll do it after rest
13:06 Jayhost: also if the urgency of the testing is high let me know. I assumed not.
13:25 pmoreau: Where the hell are those extra `ld some_reg c0[]` coming from? I never asked for those!
13:25 pmoreau: `convertToSSA()`: I blame you for those
13:35 pmoreau: Hum no, it's the pre-SSA legalize pass
13:41 pmoreau: Ah! Gotcha, it's the lowering of the store which adds those
14:16 pmoreau: Could anyone explain me what is going [here](https://phabricator.pmoreau.org/diffusion/MESA/browse/spirv_1.0/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp;b25dffd09046fa9a626abaf8a4a68b4e4501cc43$1967)?
14:16 pmoreau: imirkin: You're probably the best person to help me, since you recently wrote that piece of code. :-)
14:22 pmoreau: If I comment that part, hello_world is working. :-)
15:24 hakzsam: pmoreau, yes, he added that piece of code for SSBOs I guess
15:24 hakzsam: what's your problem?
16:05 Mittttens: imirkin, did you ever look into that crash?
17:48 imirkin: pmoreau: yeah.... so the idea there is that global has a file index, which in turn specifies where to find the real base offset. what you're dealing with is a *real* global value, and you don't want any base index. so you need to indicate that somehow to the backend.
17:49 imirkin: Mittttens: nope
17:49 imirkin: Mittttens: tbh i don't really even remember what your issue is. iirc you didn't file a bug.
18:38 Mittttens: imirkin, it was
18:38 Mittttens: java: pushbuf.c:727: nouveau_pushbuf_data: Assertion `kref' failed.
18:40 karolherbst: Mittttens: you should file a bug with all the information (software used, version of lwjgl, mesa, full stacktrace...)
19:14 Mittttens: sounds like a lotof work karolherbst
19:15 karolherbst: Mittttens: yeah, well it is a lot of work for us to guess your issue
19:15 karolherbst: either you help us understanding what your problem is or we can't help you at all
19:15 Mittttens: i told him how to recreate it karolherbst
19:16 Mittttens: i gotta essay due at midnight pal
19:16 karolherbst: and what should we be able to do about that?
19:17 Mittttens: https://ghostbin.com/paste/n6fc4 karolherbst
19:18 karolherbst: ahh okay
19:18 karolherbst: you should still file a bug about that
19:18 karolherbst: devs are prone to sudden memory losses
22:53 danofsatx: greetings, how can I determine which GPUs are supported? In particular, I'm looking at the GTX 970m and 980m. If nouveau supports these, does it support the 4x4k display capability of the card?