[00:00] if you mean use the nvidia-installer, forget it [00:00] unless you want to damage your system [00:01] bjsnider: well manually installing the .run from nvidia [00:01] yeah, do not do that thing [00:02] bjsnider: any ideas on how to get the nvidia-installer working then? [00:02] maybe it is working [00:02] it used to be less tolerant of broken edid info than other drivers [00:02] maybe still is [00:04] Well I said above that I installed the driver in 11.10 and upgraded up till now. So it must be something funny going on recently [00:04] it's pictureboxed or windowboxed or letterboxed or something? [00:05] fagan: sudo apt-get install linux-headers-3.5.0-18-generic linux-headers-3.5.0-18 [00:05] its not building because you dont have all the headers, theres a bug open about that [00:05] nice thanks Sarvatt [00:05] or just linux-headers-generic like bjsnider said would work [00:05] (and upgrade automatically) [00:06] i think he was offline when i said that [00:06] I was :) [00:06] yeah was off when i said it too, oops [00:06] that log says: [00:06] Module build for the currently running kernel was skipped since the [00:06] kernel source for this kernel does not seem to be installed. [00:06] i thought this problem was solved already [00:07] nvidia-current should depend on linux-headers-generic shouldn't it? [00:07] Ah I completely looked over that line everything else seemed usual oh except demod didn't run I should have remembered that :) [00:08] bjsnider: some people use -lowlatency or -server, no way to guarantee the right headers are installed :( [00:08] Anyway it looks like it worked going to reboot again thanks :D [00:08] this has been a bug for awhile [00:08] https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-updates/+bug/1068341 [00:08] Launchpad bug 1068341 in software-properties (Ubuntu) "No way to specify correct dependencies for dkms packages (nvidia driver install fails to get matching header)" [High,Triaged] [00:08] lots of people complained baout this in the +1 channel the past few months [00:09] must be because of the switch away from jockey [00:14] Worked! :) [00:14] Sarvatt, hrm, is that bug something we should put on our TODO list? [00:15] of course it worked [00:16] bjsnider: well im excited because I was scratching my head for an hour wondering what was gone wrong :) [00:17] lots of people had th is problem [00:19] Prf_Jakob: is the decimal seperator in your locale a . or a ,? [00:19] because if its , possibly http://steamcommunity.com/app/221410/discussions/1/882965239689221201/ [00:20] Sarvatt: Haha yeah it is. [00:20] utf8-se [00:35] hmm, apt-cache policy shows me fglrx-experimental-9, however jockey-text -l does not [00:43] hmm, lets' try rebooting. [00:57] Sarvatt: thanks! that helped! [01:45] tjaalton: ^ you'll hit the same issue in finland :P [03:20] bryceh, there no nvidia 310 drivers for R :( [04:13] they're in xorg-edgers aren't they? [04:17] bjsnider, oh yes, didnt think to check there [05:40] darkxst: thanks so much for looking into the ppa-purge multiarch thing, i'll run through this tomorrow and see how it works. i didn't even know about dpkg-query! [09:19] Sarvatt, np, I really needed to purge a ppa on my system and turned out it wasnt too hard to fix ;) [17:23] bryceh_: ping¿ [17:24] yep [19:26] tjaalton, looks like 2:1.13.0-0ubuntu7 is unreleased? Is it waiting on anything before going in? === bryceh_ is now known as bryce [19:29] bryce: I think the barrier patch needed a proper review [19:30] mlankhorst: ^? [19:31] ok; I'll just add the poulsbo pci id patch with it then [19:32] or given that it's been a few weeks, maybe we should drop the patch for the time being until it gets reviewed? [19:33] bryce: well the sru queue is moving pretty slow in general too, so don't think it matters that much. [19:33] I was hoping for raof to review it but he lost out on his chance to do so, just keep it in. :P [19:34] well, xserver git targets raring, so no sru queue blockage to worry about there... [19:34] oh sure go for it then [19:34] * bryce shrugs [19:38] first point-release rc coming next week [19:38] 1.13.1rc [19:38] but uploads are cheap [19:38] i'll be back to work starting monday [19:40] sru's are expensive :/ [19:40] enough fixes I do want to push, but not necessarily have a test for [19:40] they need to go to the devel series first [19:40] and let's get that exception for xserver point-releases.. [19:40] mlankhorst, in the changelog you mention 234-composite-borderclip.patch "is wrong". Did you get a better solution for that, or should we drop that for now? [19:41] bryce: drop for now, upstream didn't respond [19:41] ok [19:41] and then the pointer barrier change is the update to 500_pointer_barrier_thresholds.diff. You want to drop that one as well, or is it ok? [19:41] keep [19:44] okie doke [19:46] I'm more amused that games were playable with nouveau, didn't expect that :D [19:50] pretty soon the "all we need is the documentation" crowd will be proven wrong, because they didn't even need that [19:51] bjsnider: well it is a lot faster with documentation.. [19:51] we have no clue how to do memory reclocking right for example [19:52] I'm reclocking my memory with a glue of shell script, c, and assembly for the card to run :-) [19:53] but it's all hardcoded, so it might even damage other cards when you try [19:54] well, i do think it's interesting that a community can build a good graphics driver with no help at all from the manufacturer [19:55] although the manufacturer has to protect other organizations' ip [19:56] wee, crashed mplayer again === yofel_ is now known as yofel [19:58] bjsnider, I get the sense that nouveau is about where radeon was around the time that AMD opened the docs [19:58] oh, i thought it was further along than that [19:58] I recall alex and them mentioning that when they started looking through the docs they discovered all these registers that they'd mis-guessed [19:58] radeon couldn't do any opengl could it? [19:59] I'd have to check but I seem to remember it at least had some experimental support for it [20:00] now we just need a nvidiahd driver then.. [20:01] 17 Apr 2009: AMD releases initial code branches for 3D support on R6xx/R7xx (see more below) [20:01] however I think there was 3D for older devices before that. [20:01] r200 worked just fine [20:02] way before the docs were public. but that was for r5xx anyway [20:05] well this looks even uglier [20:05] :D [20:24] bryce: did you test the xserver patch for the input matrix funkyness? [20:25] tjaalton, on my todo list for today [20:26] bryce: ah :) [20:26] mlankhorst, rtg asked for an r ppa, so I've set up https://launchpad.net/~ubuntu-x-swat/+archive/r-lts-backport [20:26] if it works, add a tested-by on xorg-devel@ :) [20:26] bryce: sure, just need a new mapping file [20:27] I need to add a hack to make it get the prefix from the mapping file, but it ought to work [20:27] I'm also working on getting the changes to mesa upstream, the ones from intel already are [20:28] ones in mesa too, but needs to be backported to the 9.0 branch after release [20:28] mlankhorst, no I mean he was asking for a ppa so he could start uploading his kernels there. :-) [20:28] ah [20:30] mlankhorst, but that's great you're ready to go with our bits. But no worries, take care of higher priority stuff first. [20:30] well it's not a big problem, I tried to keep most of the changes easy enough for forward compatability [20:31] plus i want to upload libdrm 2.4.40 to raring so I can keep playing with mesa master on my weird backport stack :-) [20:35] hi :) [20:36] hmm, who broke linking against mesa? ;) [20:41] sigh mplayer hates -vc , atm (disabling vdpau acceleration). some out of bounds write [21:57] mlankhorst, are you using mplayer or mplayer2? [21:57] uau's mplayer is better than the regular one [22:00] uau? [22:02] he's the lead mplayer2 developer [22:02] ah [22:03] no I probably know why it happened [22:05] ah I was right, silly mistake :D [22:11] bjsnider: actually it works worse with mesa afaict.. [22:12] -vo vdpau -vc ffh264 [22:12] channels = 6 [22:12] vc = ffmpeg12vdpau,ffh264vdpau,ffwmv3vdpau,ffvc1vdpau,ffodivxvdpau, [22:12] vo = vdpau:colorspace=0, [22:12] is what i use atm [22:13] still some glitches though, probably messing up image parameters :-) [22:14] that wouldn't work without the blob [22:14] you'd better have a 6-channel sound system [22:14] that doesn't have a dac [22:14] works fine without the blob ;-) [22:15] ok, so gallium does vdpau? [22:15] well strictly speaking still blobby [22:15] there's a working vdpau state tracker in gallium? [22:15] not upstream, but it does have some vdpau output support, and if you're insane and know nouveau some you can load the decoder blobs into the hardware [22:15] i know they'd talked about it [22:15] yeah [22:15] it's just the drivers that are lacking [22:17] well, i'd use the cpu assuming it's sandybridge or ivybridge [22:17] Where's the fun in that!!! [22:18] 1000 22700 1.0 0.2 542568 39148 pts/1 Sl+ 23:09 0:05 mplayer -framedrop -aid 1 01 The Crystal Empire, Pt. 1 (HD).m4v [22:18] :D [22:18] i'm assuming you want it to work as an htpc or whatever. if you're trying to burn down the house that's a different motive [22:18] i want it working because I can :p [22:18] m4v my butt [22:19] ISO Media, MPEG v4 system, iTunes AVC-LC [22:19] itunes my butt [22:20] anyway, ffmpeg has multithreaded decoding which beats gpu decoding in speed [22:20] although it uses more power [22:20] libav i mean [22:21] Again, where's the fun in that, more fun to hack on nouveau :P [22:21] xserver 1.13.0-0ubuntu7 uploaded [22:25] it would probably be more fun if they'd release the docs [22:27] No, but it would be more productive [22:27] :P [22:27] vdpau's been my puzzle [22:28] wikileaks is going to have to release those docs [22:28] which means they'll need a whistleblower [22:28] I would rather have permission to redistribute the firmware blobs [22:29] bjsnider: does mplayer2 support mvc? [22:29] what does that acronym mean? [22:29] erm h264 multiview codec [22:29] i would suggest going to #mplayer2 and asking uau about it. he will answer [22:30] or whatever it was called [22:30] I know the firmware has support for it, but nvidia doesn't support it [22:30] he's got betetr vdpau support i can tell you that [22:30] aha there's 9.0.1 [22:42] mlankhorst: afair there's a (obsolete?) nouveau-firmware package in multiverse. maybe put the video blobs there? [22:44] jwi: I was thinking of writing an extraction program instead, I don't want to redistribute anything myself..