fagan | hey im trying to test out the nouveu driver and xorg is saying its not configured properly, I ran Xorg -configure but it looks like its not doing anything | 00:20 |
---|---|---|
fagan | Im on lucid by the way | 00:20 |
fagan | I think it may have something to do with residual config from the nvidia driver | 00:21 |
fagan | Anyone have any ideas? | 00:21 |
RAOF | fagan: How are you trying nouveau? | 00:23 |
RAOF | Because the version in the archives currently isn't expected to work, pending some kernel-team decision about how to get nouveau.ko in our kernel build. | 00:24 |
fagan | RAOF: Im using it from the nouveau ppa | 00:24 |
fagan | The one from the archive works though | 00:25 |
fagan | on my machine | 00:25 |
RAOF | Really? It doesn't build against 2.6.32 kernels, though. | 00:25 |
fagan | hmmmm I just installed it rebooted and it worked | 00:25 |
fagan | dont ask me how | 00:25 |
fagan | its definitly not the nv driver | 00:26 |
fagan | That sucked on my hardware | 00:26 |
RAOF | Its possible that you've got nouveau-kernel-source from the nouveau ppa and the DDX from lucid, I guess. | 00:26 |
fagan | I did install the -source package so thats it :) | 00:27 |
fagan | RAOF: so any ideas why im running safe graphics mode? | 00:27 |
fagan | It hung when I did a -configire | 00:27 |
RAOF | No idea, no. | 00:27 |
fagan | -configure | 00:27 |
fagan | hmmm | 00:28 |
fagan | any other things I can try? | 00:28 |
RAOF | What's in your xorg.conf? | 00:28 |
fagan | the nvidia config | 00:29 |
fagan | which caused the problem I think | 00:29 |
RAOF | Well, yes. Can you pastebin your xorg.conf? | 00:29 |
fagan | damn I just rm'd it | 00:29 |
fagan | sorry | 00:30 |
RAOF | That's OK. | 00:30 |
RAOF | All it wants is "Section "driver" Identifier "nouveau" Driver "nouveau" EndSection" in it. | 00:30 |
fagan | oh ok then ill give it a go | 00:30 |
fagan | be back in 5 | 00:31 |
notlistening | A bit later and there is no difference in the cpu usage on that machine with all the extras that i installed so there must be something different between the hardware more than just CPU and GPU | 00:35 |
fagan | that didnt work anyway | 01:00 |
fagan | so im going to go back to the nvidia driver | 01:01 |
fagan | the conflict is http://pastebin.ubuntu.com/337674/ | 01:02 |
fagan | The xorg update conflicts with nvidia's driver | 01:02 |
RAOF | Yes it does. | 01:02 |
fagan | so what do I do? | 01:03 |
RAOF | Not use the nvidia driver in Lucid. | 01:03 |
fagan | That isnt a solution because my computer looks very crap without it | 01:04 |
RAOF | Beacuse it won't work. IIUC, we need a newer nvidia driver, one that will actually work against xserver 1.7 | 01:04 |
RAOF | fagan: Not use Lucid? The proprietary drivers are frequently, and often long-term broken in development. | 01:05 |
fagan | I tried the 190 driver with no luck | 01:05 |
fagan | hmmmmm | 01:05 |
* fagan thinks maybe rolling back to an older version of xorg will help him | 01:05 | |
RAOF | Yes, it might. If you can :) | 01:06 |
fagan | Or maybe a newer one | 01:08 |
RAOF | Well, that won't work, because the nvidia driver won't work against it. | 01:08 |
fagan | Ah ill see what I can do | 01:09 |
fagan | Dont ask me how but the ppa for the xorg team worked | 01:10 |
fagan | no conflict no nothing https://launchpad.net/~xorg-edgers/+archive/ppa | 01:11 |
Sarvatt | 185.18.36 worked for you with that? | 01:11 |
fagan | 190 | 01:11 |
Sarvatt | oh ok | 01:11 |
fagan | Awesomeness | 01:12 |
Sarvatt | i couldnt find a 190 in a ppa that wasnt really borked when i looked at the stuff in debian/ and one screwed up being able to downgrade back, ended up just manually installing it from nvidia for now | 01:15 |
fagan | Nope spoke too soon got killed at the end of installation | 01:16 |
fagan | damn it | 01:17 |
Sarvatt | i had a ton of dangling alternatives from the bad packages | 01:17 |
fagan | im just going to get the one from nvidia themselves | 01:17 |
Sarvatt | 195.22 manually installed is working great with edgers here | 01:18 |
fagan | got to exit the session to install it brb | 01:19 |
fagan | that worked well | 01:27 |
fagan | x and nvidia now play nice | 01:28 |
fagan | why are the packages so bad then? | 01:28 |
fagan | If the nvidia driver works why dont we just package that up that? | 01:29 |
Sarvatt | tseliot is working on it | 01:32 |
fagan | thank god | 01:32 |
notlistening | Can i come back to my original question, aobut the amount of CPU i am using for a video, can the frame buffer in the bios make a big difference or the amount of RAM a system has to use? | 01:54 |
Sarvatt | nvidia gpu acceleration requires either 128 or 256mb last i checked | 01:55 |
Sarvatt | cant remember which, i think its 256 | 01:56 |
fagan | mine has 512 | 01:56 |
Sarvatt | if you are using xbmc, i know i saw a good guide on using gpu acceleration on an ion board in ubuntu on their forums | 01:58 |
Sarvatt | http://blog.xbmc.org/forum/tags.php?tag=ion | 01:59 |
notlistening | the bios limits me to 256 at the moment as i only have a 1 gb a the moment | 01:59 |
notlistening | Thanks Sarvatt | 02:01 |
Sarvatt | gotta be a vdpau thread out there somewhere that'll explain how to make it work, you would know if its working because you'd be getting closer to 10% cpu usage than the 80% | 02:04 |
Sarvatt | ewwww http://lists.x.org/archives/xorg-devel/2009-December/004029.html | 02:49 |
LLStarks | whois keybuk | 06:40 |
LLStarks | damn | 06:40 |
RAOF | Hm. X seems unusually slow, and seems to be spending a lot of time in i915_gem_throttle_ioctl. Is this something others are seeing? | 07:38 |
LLStarks | how would i know? | 07:45 |
LLStarks | jeez, that sounded rude. | 07:45 |
RAOF | You'd know if X was being extremely laggy for you, you had an intel card, and ran gnome-system-monitor and saw X waiting for i915_gem_throttle_ioctl :) | 07:50 |
tjaalton | feels quite normal here | 07:56 |
RAOF | Oh, wow. ubuntu-desktop's new libgtk kills emacs! | 07:59 |
tjaalton | heh, funny.. we ship a cirrus driver with patches to support qemu, and then in xserver a patch that uses vesa for the hw id | 08:59 |
tjaalton | nevermind, it depends on what type kvm is using | 09:22 |
tjaalton | "std" want's mesa | 09:22 |
tjaalton | urgh, vesa | 09:22 |
tjaalton | s/'// | 09:22 |
tjaalton | that should be upstreamed | 09:24 |
tjaalton | mvo: does update-manager always run dist-upgrade, ie. remove packages when there are conflicts during the devel release? | 09:34 |
mvo | tjaalton: sort of, by default it does not do this, but it will prompt for "partial upgrade" mode if it has to remove stuff and tries to be clever | 09:36 |
tjaalton | mvo: ok, thanks | 09:37 |
mvo | why? anything that it should do to make transitions better or something? | 09:38 |
tjaalton | nah, there was just one user who claims that lucid->lucid upgrade broke everything without prompting | 09:39 |
tjaalton | and I don't believe him :) | 09:39 |
tjaalton | the result was that it removed all the drivers | 09:39 |
tjaalton | so the system didn't get X | 09:39 |
mvo | not without prompting :) | 09:40 |
mvo | if he used update-manager we should have logs too | 09:41 |
tseliot | tjaalton: what's the status of input drivers in lucid? An AMD engineer reported that xserver-xorg-core is not installable. Was it just because of synaptics? | 09:41 |
tjaalton | yes, but I won't bother.. told him to install video-all | 09:41 |
tjaalton | tseliot: no, check the mirror | 09:41 |
tjaalton | at least the images build | 09:41 |
tjaalton | tseliot: also, the current synaptics doesn't have your patch in it, had to upload it to make it installable | 09:42 |
tjaalton | but I'll add it now | 09:42 |
tseliot | ok, thanks | 09:42 |
* tseliot forgot about having a lucid chroot already | 09:45 | |
tseliot | xserver-xorg-core was installed correctly here | 09:45 |
tjaalton | if he's using fglrx, it'll get removed | 09:46 |
tjaalton | only the OSS drivers work | 09:46 |
tseliot | ok | 10:01 |
hyperair | does anyone here using the nvidia-96 driver have issues with screen flickering? | 10:03 |
=== seb128_ is now known as seb128 | ||
=== paran_ is now known as paran | ||
=== mac_v_ is now known as mac_v | ||
tseliot | federico1: hey, do you have a few minutes to discuss new features (one of which is more of a fix) for g-s-d and g-d? | 17:00 |
tseliot | otherwise I can file 2 bug reports and we'll discuss there | 17:02 |
federico1 | tseliot: sure, what's up? | 17:21 |
tseliot | federico1: the 1st feature is support for transformations | 17:22 |
federico1 | tseliot: BTW, I'm going to start integrating some of the Moblin patches for RANDR stuff | 17:22 |
tseliot | mainly scaling | 17:22 |
federico1 | ah! | 17:22 |
federico1 | there *is* a patch for that :) | 17:22 |
tseliot | federico1: really? | 17:22 |
tseliot | I'm working on Moblin too | 17:22 |
federico1 | I just haven't looked at it closely | 17:23 |
tseliot | federico1: can you give me the link, please? | 17:23 |
tseliot | federico1: I can't find it in the Moblin 2.1 sources | 17:28 |
tseliot | federico1: anyway the 2nd feature is, put simply, that we check GL_MAX_TEXTURE_SIZE in addition to the max framebuffer size | 17:29 |
tseliot | i.e. that we check the 3D limitations in addition to the 2D limitations | 17:30 |
tseliot | for example on i945 chips you can have 2D up to 4096x4096 and 3D up to 2048x2048 | 17:31 |
tseliot | and if you go beyond the 3D limit when using a compositing manager you get a black screen | 17:32 |
tseliot | therefore we should not validate modes that go beyond that when using, say, mutter (in Moblin) | 17:34 |
federico1 | tseliot: makes sense | 17:46 |
federico1 | tseliot: so, I don't know much about GL. Is the texture size limit just for textures, or does it involve the whole screen? I.e. could a smarter compositing manager simply create two textures for too-wide windows (or something like that?) | 17:47 |
tseliot | federico1: it's the area in which you get 3d acceleration | 17:48 |
tseliot | I have a bug report about it | 17:49 |
tseliot | https://bugs.freedesktop.org/show_bug.cgi?id=23718 | 17:49 |
ubottu | Freedesktop bug 23718 in Driver/intel "[945GME] attaching external monitor causes black screen, with Compiz Fusion in Karmic" [Critical,New] | 17:49 |
tseliot | I've written a patch to access max texture size but I haven't tested it yet. I'll file a bug report and attach it there soon | 17:51 |
* tseliot -> dinner | 17:51 | |
=== yofel_ is now known as yofel | ||
Amaranth | so, my mouse keeps jumping to the center of the screen now... :/ | 23:00 |
Amaranth | federico1: btw, while a compositing manager could split the screen into smaller chunks to work around the texture size limit the only way to do so is to not use GLX_EXT_texture_from_pixmap at all | 23:04 |
Amaranth | the other way of doing OpenGL compositing is...slow | 23:05 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!