05:10 blackpenguin: hi. I experience a problem: the GL parts of xscreensave doesn't show loaded images, but black frames instead. I wonder if this a known problem and if there is an xorg.conf option to fix it
05:11 berz3rk: hello
05:11 berz3rk: is GT 240 supported?
05:15 poisson_rouge: Hi!
05:16 poisson_rouge: I have a problem with my graphics somewhere, sometimes the whole thing kind of crashes and weird colored squares invade my whole screen, I end up having to ctrl alt del to get it back to normal... would anyone know how to fix it?
05:17 poisson_rouge: Or even just have an idea of where it comes from ^^
05:55 blackpenguin: poisson_rouge: if your system still reacts on ctrl-alt-del it seems to run. maybe you get some more information from kernel log or dmesg
05:55 poisson_rouge: oh yes it runs, I can do stuff, i just can't really see what I'm doing anymore
05:55 poisson_rouge: and ok, I'll check that
05:56 blackpenguin: you could ssh-login from a second machine to examine logs/dmesg
05:56 blackpenguin: or even type blind and write output to a file to examine later
05:57 poisson_rouge: if all it takes is typing commands in the terminal, that should be no problem, i can see enough for opening one
05:57 poisson_rouge: ok!
06:07 poisson_rouge: ty :)
09:07 imirkin: berz3rk: yeah, GT240 should be supported just fine. if you have the GDDR5 version there will be a bunch of issues though...
09:07 imirkin: blackpenguin: pastebin logs... dmesg + xorg
09:12 berz3rk: imirkin: I think i have the gddr5 version
09:12 imirkin: berz3rk: until recently that card would hang immediately on load
09:12 imirkin: but now it kinda sorta limps along
09:32 RSpliet: imirkin: is it still bad?
09:33 imirkin: RSpliet: it's as bad as it was after ben's initial fix, i don't think there has been subsequent work
09:35 RSpliet: imirkin: I haven't been able to test
09:35 RSpliet: since... Ben has my card
09:35 imirkin: hehehe
12:01 imirkin: i assume everyone just glossed over my nvc0 double support patches... if that was not the case and you want more time to review, let me know. otherwise i'm pushing tonight.
12:02 imirkin: (i found a couple of minor bugs in the tgsi -> nv50 ir adapter... missed some of the new opcodes, and also src mask computation needed fixing for a few opcodes)
13:01 mupuf: imirkin: did you test that on reator? Sorry for not being responsive, I was pretty sick, stuck in bed and quite delirious so I didn't plug the tesla when I said I would
13:01 imirkin: mupuf: nope, i didn't. don't worry about it
13:01 mupuf: ack
13:01 mupuf: it is plugged though
13:01 imirkin: mupuf: whompy tested that nv50 change for me, already pushed
13:01 mupuf: nv92 and d9
13:01 mupuf: wonderful
13:01 imirkin: mupuf: the double stuff i've been testing on my GF108
13:02 imirkin: hope you feel better soon!
13:02 mupuf: :)
13:03 mupuf: thanks, I feel quite good already, just feel really tired
13:03 imirkin: that's what happens when you're above the arctic circle :p
13:09 imirkin: i will need to do a big tracing session, the d9 will be good for that. hopefully tonight or tomorrow.
14:05 blackpenguin: http://pastebin.com/Nc4d8tyA < the logs to the problem with loaded images
14:07 imirkin: blackpenguin: what issue do you have again?
14:08 blackpenguin: images loaded by xscreensaver do not show. they are just black (only with the GL savers like glslideshow)
14:08 imirkin: so other gl screensavers work fine?
14:08 blackpenguin: yes
14:09 imirkin: i suspect it's because of the 16-bit thing
14:09 imirkin: there's a limitation on the card
14:09 imirkin: in that if you render to a 16-bit color buffer, you must use 16-bit depth
14:09 imirkin: the way i've "fixed" this is i just ignore the depth buffer if a conflicting situation happens
14:10 imirkin: but that doesn't necessarily cause correct rendering :)
14:10 imirkin: but... it avoids gpu faults
14:11 imirkin: blackpenguin: if you're interested, you could try fixing it for real... shouldn't be too hard if you're familiar with C and so on
14:14 blackpenguin: Going to try if the problem disappears with 24bit (24 bit flickers wild and is not very useful, but at least we know if it's really a color depth problem)
14:16 blackpenguin: okay, works fine in 24bit. (and for some reason the flickering disappeared for now. maybe I've set an option in xorg.conf while trying to fix the black images that now fixed the flickering?)
14:17 imirkin: but don't worry -- other things won't work with 24-bit :)
14:17 imirkin: coz if you have a 32-bit color format, you have to have a 32-bit zeta buffer too, not 16-bit
14:19 blackpenguin: as long as the few thing I use work I'm fine with it. I do not use GL very much on that machine.
14:21 blackpenguin: just wonder which option I changed that the flickering disappeared (I changed all of then at least once ;)
14:21 blackpenguin: however... I am quite happy now :) thank you.
14:22 imirkin: enjoy :)
19:16 whompy: So I'm dinking around with nouveau.pstate=1 on my nv50... alas, i'm getting http://ix.io/gsS when I try to write one of the values to /sys/class/drm/card0/device/pstate
19:16 imirkin: whompy: dmesg
19:17 whompy: http://ix.io/gsT
19:17 imirkin: you need 3.19 i think
19:17 whompy: Ah, for some reason i thought it was available in 3.17+
19:18 imirkin: not for GT21x
19:18 whompy: Kk. Thanks again, good sir.
20:09 whompy: Ok. After some quick toying around, I am quite impressed with the reclocking! I was afraid I'd end up going back to the prop driver, but this should be fine for my daily needs thus far. I appreciate the work and help in getting stuff up and running!
20:10 imirkin: enjoy
22:47 imirkin: mwk: does a "join" instruction modifier on nvc0 actually jump to the joinat location? i didn't think it did, but just checking...
22:48 imirkin: i thought it was just a "end of divergent control flow" marker