gzed | any how-to to activate gallium nouveau on maverick? xorg-edgers ppa doesn't contain the libgl1-dri-gallium package for maverick... | 12:57 |
---|---|---|
tjaalton | gzed: aiui it's in maverick already, libgl1-mesa-dri-experimental package | 13:06 |
gzed | ok, gonna try it out, thanks | 13:07 |
gzed | hmmm | 13:09 |
gzed | ls /usr/lib/dri/gallium/ | 13:09 |
gzed | i915_dri.so r600_dri.so swrast_dri.so | 13:09 |
gzed | no nouveau in there | 13:09 |
tjaalton | it's in /usr/lib/dri | 13:13 |
gzed | aren't those the classic mesa drivers? | 13:14 |
tjaalton | not necessarily | 13:14 |
tjaalton | hmm | 13:14 |
gzed | I think you're wrong | 13:14 |
tjaalton | maybe | 13:14 |
gzed | I used to have a /usr/lib/dri-gallium directory under lucid | 13:15 |
tjaalton | the packaging has probably changed since | 13:15 |
gzed | not until yesterday | 13:15 |
gzed | I upgraded to maverick this morning... | 13:15 |
tjaalton | well maverick has been in development since may | 13:16 |
tjaalton | checking the packaging now | 13:17 |
tjaalton | ..which is slow since a lot has changed since I touched mesa last :) | 13:20 |
gzed | :) | 13:20 |
gzed | never mind, I think I'll have to build from git to get to play with it again | 13:21 |
tjaalton | it is the gallium driver btw | 13:23 |
tjaalton | just installed in /usr/lib/dri so it's actually used by default | 13:24 |
gzed | running the closed blob right now... :( | 13:24 |
knittl | hey. i don't see any drivers in jockey | 17:30 |
knittl | can i help troubleshoot? | 17:31 |
tjaalton | you don't have nvidia-*-modaliases installed? | 17:45 |
tjaalton | or the hw is too old | 17:46 |
knittl | tjaalton: modaliases are installed, hw is 2 years max | 18:08 |
knittl | now i can see drivers | 18:08 |
knittl | but i bet it fails | 18:08 |
knittl | like everytime | 18:08 |
knittl | :( | 18:08 |
knittl | i will paste the logfile, gimme a minute | 18:08 |
knittl | tjaalton: http://paste2.org/p/1016479 | 18:10 |
knittl | i emptied it before logging in and starting jockey | 18:11 |
tjaalton | fails how? | 18:13 |
tjaalton | it got installed | 18:13 |
knittl | what's the last error then? | 18:13 |
knittl | it also says now in jockey: a different version of this driver is in use | 18:14 |
tjaalton | dunno | 18:14 |
knittl | i try to restart, brb | 18:15 |
knittl | there's no xorg.conf though. is that ok? | 18:15 |
tjaalton | there's 256.53-0ubuntu3 available btw | 18:15 |
tjaalton | yes | 18:15 |
tjaalton | well, no | 18:15 |
tjaalton | not with nvidia | 18:15 |
knittl | so why doesn't jockey create it? | 18:15 |
tjaalton | no idea | 18:15 |
knittl | so maybe it is failing? | 18:16 |
tjaalton | the error is because it's trying to create an alternative link for a file which only the 64bit package includes | 18:17 |
tjaalton | which itself is a bug in the package, but shouldn't be fatal | 18:17 |
knittl | also | 18:17 |
knittl | does not currently work with xserver 1.9 | 18:17 |
knittl | is an interesting error | 18:17 |
tjaalton | but why doesn't your mirror have the latest files? | 18:17 |
tjaalton | no | 18:18 |
knittl | line 10 – 16 | 18:18 |
tjaalton | thats for the old driveres | 18:18 |
tjaalton | -e | 18:18 |
tjaalton | -96 and -173, which don't support the new abi | 18:18 |
knittl | installing nvdia-current by hand, then nvidia-xconfig works | 18:18 |
knittl | so i only have problems getting jockey to work | 18:19 |
tjaalton | xconfig should work just the same if it was installed via jockey | 18:21 |
knittl | and since jockey is the recommended way it's sad :( | 18:21 |
tjaalton | then file a bug | 18:21 |
tjaalton | maybe the error is fatal for jockey | 18:22 |
knittl | maybe, i don't know | 18:22 |
tjaalton | so it doesn't create the conffile when installing on a 32bit system | 18:22 |
knittl | but then it would be fatal on every non-64bit system | 18:22 |
tjaalton | file a bug against nvidia-current.. | 18:22 |
tjaalton | yes | 18:22 |
knittl | jockey even says it failed installing, i should see the log | 18:23 |
tjaalton | this wasn't the logfile? | 18:23 |
knittl | the paste? it was /var/log/jockey.log | 18:24 |
marcog | hey, i just upgraded my netbook to maverick and got the exact same problem as this: http://ubuntuforums.org/showthread.php?t=1587153 | 18:24 |
marcog | any ideas on how to fix? | 18:24 |
knittl | also, standby does not work on my system with nouveau | 18:24 |
knittl | i'd use it instead of nvidia, but no standby is a no-go | 18:24 |
marcog | anyone? to summarise, i get a blank white screen after logging in | 18:29 |
knittl | using gitk, vlc, blender, ding (i guess xlib) applications will crash my xserver, when using xinerama … | 18:36 |
knittl | any pointers? | 18:36 |
tjaalton | buggy nvidia? | 18:39 |
knittl | probably | 18:52 |
bjsnider | knittl, the errors listed on lines 779 and 780 are supposed to happen on i386. there is no way around it because dpkg doesn't support multi-arch yet | 18:54 |
knittl | hmmm | 18:55 |
knittl | well, it still fails while installing | 18:56 |
knittl | and i cannot remove it using jockey | 18:57 |
knittl | using nouveau now | 18:57 |
knittl | no standby <<< xserver crashes | 18:57 |
tjaalton | bjsnider: it's still a bug that it tries to add an alternative for a file that doesn't exist | 18:57 |
tjaalton | and seems like it breaks jockey | 18:57 |
bjsnider | the messages in that log specificaly say that the nvidia alternatives are being used and the kernel module is installed | 18:58 |
bjsnider | so it worked | 18:58 |
tjaalton | the install yes, but jockey still fails | 18:58 |
bjsnider | i don't understand the distinction | 18:59 |
bjsnider | if hte user rebooted it would work fine | 18:59 |
tjaalton | it's not the nvidia install that creates xorg.conf | 18:59 |
bjsnider | i mean maybe it didn't create the correct xorg.conf file? | 18:59 |
tjaalton | but jockey | 19:00 |
tjaalton | exactly | 19:00 |
bjsnider | well then it's been broken the same way going back to lucid | 19:00 |
tjaalton | maybe, or maybe jockey is being more picky in maverick | 19:00 |
knittl | jockey says it failde | 19:01 |
knittl | bjsnider: no it doesnt | 19:01 |
knittl | rebooting will not work properly, because no xorg.conf is generated | 19:01 |
bjsnider | that should be listed as a jockey bug if anything | 19:05 |
tjaalton | it's a packaging bug in nvidia.. | 19:06 |
tjaalton | at least | 19:06 |
bjsnider | how would you suggest it be ffixed? | 19:06 |
tjaalton | not add --slave #VDPAUDIR32#/libvdpau_nvidia.so.1 libvdpau_nvidia.so.1_lib32 #PKGVDPAUDIR32#/libvdpau_nvidia.so.1 \ | 19:07 |
tjaalton | ..for the 32bit package | 19:07 |
bjsnider | the nvidia-current postinst script specifically says that error message will occur on i386 and it is inconsequential | 19:08 |
bjsnider | so it's not like it's a suprise that it's happening | 19:08 |
tjaalton | well, maybe mvo will decide which one is wrong ;) | 19:08 |
tjaalton | I don't see why including known(!) errors in a package is desired | 19:09 |
bjsnider | it may be impossible to get around | 19:09 |
tjaalton | no it's not | 19:09 |
bjsnider | i don't see why you can't just do i arch=i386 then | 19:10 |
bjsnider | you can always test for the arch | 19:10 |
tjaalton | that's the .in file, with a lot of variables that are replaced with real values, so there's no reason it couldn't check the build architecture too | 19:10 |
bjsnider | but it's been in there going back to lucid and it wasn't stopping i386 installs. so why is it a showstopper now? | 19:11 |
tjaalton | like I said, jockey probably got changed | 19:14 |
knittl | it didn't work in lucid for me either | 19:20 |
bjsnider | for the same reason? | 19:22 |
knittl | i guess. i don't know the reason | 19:22 |
knittl | that was when nvidia-current was introduced | 19:23 |
bjsnider | were you trying to use it on i386 then as well? | 19:23 |
knittl | yes | 19:23 |
knittl | it's the same system, only upgraded | 19:24 |
=== yofel_ is now known as yfoel |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!