05:34alkisg: imirkin: I tested it in the office a lot, but I didn't have a chance yet to deploy it to a school that I was sure it had the issue
14:43sigod: what's the story with the new nvidia hardware docs that were released, are they very useful?
15:24doobz: Anybody (other than I) tried glmark2 on nv4b?
15:42imirkin: alkisg: but your testing so far indicates that things have improved?
15:42imirkin: doobz: i'm 99% sure i've tried running glmark on a nv4x, not a nv4b specifically
15:42alkisg: imirkin: all problems solved! :)
15:43imirkin: alkisg: awesome. i'll do a release, i think that's important enough to do, and it's not like there's a ton of "natural" development on the nouveau ddx anyways - it's pretty stable in general
15:44alkisg: imirkin: thanks a lot; please ping me with the commit link so that I try to get it backported to Ubuntu... I uploaded a patched xorg, but they release newer versions every 6 months, so it's hard to have to patch them all the time...
16:03imirkin: doobz: what version of mesa are you on now?
16:06imirkin: sigod: they're very useful, but the usefulness tends to be for fairly niche features (like the recent release of crc readback)
16:07imirkin: the important things are all RE'd by the time the docs come out
16:11doobz: I tried to upgrade the deb buster installation to testing, and all was not well. Can try again. Am I able to just upgrade a single or select few packages?
16:15imirkin: doobz: i have zero knowledge of debian ... you mentioned you did something, and it still didn't work. do you know what mesa version the something ended up with?
16:19doobz: I'm guessing it was a 19 something. :) It was super glitchy and hung pretty quickly. I have a fresh buster install on a GF7100 desktop that exhibits the same issues as my laptopn. I can try a newer mesa on that fairly quickly.
16:21imirkin: can you find out the actual version somehow?
16:21imirkin: and glitchy in firefox, or in general? was this with xfce?
16:27doobz: The login screen was kinda flashing in a bad way, the clock was stuck. After it hung when copying a couple of file to the network, I lost interest. It was late so nuked install to try sabayon.
16:28doobz: I guess it was 19.2.1-1
16:32imirkin: what's "sabayon"?
16:33doobz: a gentoo based distro.
16:33imirkin: ah ok
16:33imirkin: next time this happens, have a look at /proc/interrupts
16:34imirkin: and see if the number on the nvkm line increases
16:34imirkin: i think we've had some issues with MSI interrupts on nv4x's
16:34imirkin: iirc it was nv46 though
16:43doobz: Just having a look on this install - and seem to be getting ~2 per second. No idea if that is normal or not
16:44imirkin: that's within the realm of possibility
16:56doobz: Is a valgrind output any good to anyone, specifically glmark2 seg-faulting?
16:59imirkin: the backtrace from the segfault itself would be primarily interesting
16:59imirkin: as well as how you invoked glmark2
17:16doobz: $%^&! It's nearly 27MB. Does pastebin work with such large files?
17:17imirkin: but it's not very interesting to look for bugs in glmark2...
17:17imirkin: if you have a stack trace, that should be enough to get started
17:21doobz: Max paste size .5MB
17:27doobz: Valgrind log here - http://www.filedropper.com/valgrind
17:28doobz: Invoked - valgrind --leak-check=yes --log-file=valgrind.txt glmark2
18:12mattst88: h/win 8