| BUGabundo | eheh | 00:00 |
|---|---|---|
| BUGabundo | so reboot and test? | 00:00 |
| BUGabundo | anything u need me to bring from the _other side: ? | 00:00 |
| Sarvatt | why can't you install nvidia-current? you probably need to install nvidia-common as well btw, looks like you purged it somewhere along the line | 00:03 |
| BUGabundo | reinstalling | 00:04 |
| BUGabundo | Sarvatt: can i install -current from chroot? | 00:04 |
| BUGabundo | or should i boot first? | 00:04 |
| === DanaG1 is now known as DanaG | ||
| Sarvatt | i have no idea, if you do it from the command line you need to set up the xorg.conf manually as well I think | 00:08 |
| BUGabundo | oh ok | 00:08 |
| BUGabundo | though udev took care of all of that now | 00:08 |
| bryceh | scary: http://www2.bryceharrington.org:8080/X/Graphs/totals.svg | 00:09 |
| BUGabundo | Sarvatt: http://paste.ubuntu.com/380034/ | 00:09 |
| bryceh | *sigh* gonna have to redo the scale on that graph again already | 00:09 |
| BUGabundo | bryceh: 150% increase | 00:10 |
| BUGabundo | or better 50% increasw | 00:10 |
| Sarvatt | is the 2.6.32-14 kernel even the main kernel yet or did you manually install it? did you forget headers for it? | 00:11 |
| BUGabundo | WTH | 00:12 |
| BUGabundo | i only had .13 on last boot | 00:12 |
| BUGabundo | 2.6.32-14.19 0 500 http://neacm.fe.up.pt lucid/main Packages 500 http://archive.ubuntu.com lucid/main Packages | 00:13 |
| BUGabundo | its in the archive | 00:13 |
| Sarvatt | yeah but not the linux-meta yet so you dont automatically get the headers and kernel upgrade.. | 00:13 |
| Sarvatt | at least i havent gotten it yet | 00:14 |
| BUGabundo | yeah | 00:14 |
| BUGabundo | ill boot into .13 | 00:14 |
| BUGabundo | its already mirrored | 00:14 |
| Sarvatt | try sudo apt-get install linux-headers-2.6.32-14 linux-headers-2.6.32-14-generic and reinstall nvidia-current | 00:14 |
| Sarvatt | you wont be able to install nvidia-current for -13 either since you have -14 installed with no headers | 00:15 |
| BUGabundo | The following NEW packages will be installed: linux-headers-2.6.32-14 linux-headers-2.6.32-14-generic | 00:15 |
| BUGabundo | run-parts: executing /etc/kernel/header_postinst.d/dkms 2.6.32-14-generic /boot/vmlinuz-2.6.32-14-generic | 00:15 |
| Sarvatt | plus theres no nouveau for -14 yet anyway | 00:16 |
| BUGabundo | yeah | 00:16 |
| BUGabundo | this moves faster then we can keep up | 00:16 |
| BUGabundo | [00:16] <crimsun> linux | 2.6.32-14.20 | lucid | source | 00:16 |
| BUGabundo | LOL | 00:16 |
| Sarvatt | wouldn't be a problem if you waited for it to come through upgrades instead of installing the kernel manually :D | 00:17 |
| BUGabundo | Sarvatt: i DID NOT jump the gun on anything | 00:17 |
| BUGabundo | all i have is from regular updates | 00:17 |
| BUGabundo | via aptitude safe-upgrae | 00:17 |
| Sarvatt | weird.. | 00:17 |
| BUGabundo | specially since i've been running nouveau for 2 weeks, i try to keep it clean | 00:18 |
| BUGabundo | so i can provide as good feedback as possible | 00:18 |
| BUGabundo | and not mess updates | 00:19 |
| Sarvatt | no idea what pulled in just the kernel upgrade | 00:19 |
| Sarvatt | thats strange | 00:19 |
| BUGabundo | :/var/log/apt# pastebinit history.log http://pastebin.com/d3c82e8c6 | 00:20 |
| BUGabundo | ./var/log# pastebinit aptitude http://pastebin.com/d57e34f4f | 00:21 |
| BUGabundo | HTH | 00:21 |
| Sarvatt | none of those even show linux-image-2.6.32-14 getting installed | 00:24 |
| BUGabundo | $ apt-cache policy linux-image-2.6.32-14-generic linux-image-2.6.32-14-generic: Installed: (none) Candidate: 2.6.32-14.19 | 00:26 |
| BUGabundo | its not installed | 00:26 |
| Sarvatt | huh, you pastebinned a long of nvidia-current failing to make the module for 2.6.32-14 | 00:27 |
| Sarvatt | log rather | 00:27 |
| BUGabundo | i know | 00:27 |
| BUGabundo | what can i say, its a mistery | 00:27 |
| Sarvatt | update-initramfs: Generating /boot/initrd.img-2.6.31-14-generic | 00:28 |
| Sarvatt | Cannot find /lib/modules/2.6.31-14-generic | 00:28 |
| Sarvatt | oh .31 | 00:28 |
| Sarvatt | argh lol | 00:28 |
| BUGabundo | # apt-get install linux-image-2.6.32-14 ? | 00:28 |
| BUGabundo | LOLOLOL | 00:28 |
| BUGabundo | E: Couldn't find package linux-image-2.6.32-14 | 00:29 |
| Sarvatt | -generic | 00:29 |
| BUGabundo | so, purge header? | 00:29 |
| Sarvatt | yeah | 00:29 |
| BUGabundo | and boot into -13 ? | 00:29 |
| Sarvatt | it failed because something is screwy, did you manually remove /lib/modules/2.6.31-14-generic or something? just purge linux-image-2.6.31-14-generic if anything | 00:30 |
| BUGabundo | not that i recall | 00:30 |
| BUGabundo | 0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded. | 00:30 |
| Sarvatt | you probably have something left behind in /boot from that kernel then that needs deleting | 00:35 |
| * BUGabundo picks up the HAMMER | 00:36 | |
| BUGabundo | :/boot# ls -l | pastebinit http://pastebin.com/d4cff30eb | 00:37 |
| BUGabundo | Sarvatt: nothing there | 00:37 |
| BUGabundo | # mlocate 2.6.31-14 brings nothing either | 00:38 |
| Sarvatt | you were doing it from a karmic livecd weren't you? lol | 00:38 |
| Sarvatt | when you pasted those logs...... | 00:38 |
| BUGabundo | yes | 00:38 |
| BUGabundo | no lucid livecd at hand | 00:39 |
| BUGabundo | i'm still on it | 00:39 |
| BUGabundo | chrooted to my system | 00:39 |
| BUGabundo | let me guess its picking stuff from /sys ? | 00:39 |
| Sarvatt | uname.. | 00:40 |
| BUGabundo | oh F | 00:40 |
| BUGabundo | sorry for the trouble | 00:40 |
| BUGabundo | rebooting | 00:40 |
| Sarvatt | it installed fine for -13, as long as your /etc/X11/xorg.conf is the nvidia one its fine to reboot | 00:40 |
| BUGabundo | let see whats on the other side | 00:41 |
| BUGabundo | i have no xorg,conf | 00:41 |
| Sarvatt | you need one for nvidia to work | 00:41 |
| BUGabundo | should i run nvidia-soemthing ? | 00:41 |
| Sarvatt | no idea what its called, i just keep the xorg.conf around | 00:41 |
| BUGabundo | # pastebinit xorg.conf-backup-100206200857 http://pastebin.com/d6e9e9950 | 00:42 |
| Sarvatt | nvidia-xconfig? | 00:43 |
| Sarvatt | thats what it says at the top of my xorg.conf | 00:43 |
| BUGabundo | The program 'nvidia-xconfig' can be found in the following packages: | 00:43 |
| Sarvatt | but i probably made it with the blob straight from nvidia | 00:43 |
| BUGabundo | ill try that | 00:44 |
| Sarvatt | yep that backup one is fine | 00:44 |
| BUGabundo | wish me luck | 00:44 |
| Sarvatt | just looked at your paste | 00:44 |
| Sarvatt | yeah lbm-nouveau builds fine straight from linus' tree, too bad i cant merge the nouveau one into that | 00:48 |
| BUGabundo | Sarvatt: thanks. blob and compiz back on | 00:57 |
| BUGabundo | ping me back once nouveau is ready for more testing | 00:57 |
| Sarvatt | if you remove the xorg.conf nouveau should just work | 00:58 |
| Sarvatt | i put a note on the edgers page saying nouveau was broken with it for now | 00:59 |
| Sarvatt | mesa requires the api bump to build nouveau now so i didnt drop it from libdrm but i havent been able to get the lbm-nouveau going | 01:00 |
| Sarvatt | hmm guess i could just remove the list_sort stuff completely, nouveau upstream tree is based on 2.6.32 without it.. | 01:03 |
| Sarvatt | bryceh: yeah all those lbm-* hacks did just extend support for the renaming, didnt replace nouveau in anything so everything should work fine with the backported drm with normal module names. about intel 2.9.1, I'm not sure honestly.. we *could* transition 8xx people onto fbdev on top of KMS instead of -intel if we went 2.10.. 2.9.1 + ppa packages of newer ones in #ubuntu-x would probably be a better idea for a LTS though. Do we have a way we co | 01:33 |
| Sarvatt | uld make 8xx series GPUs default to UMS if we go with 2.9.1? because the KMS overlay support in 2.10 is a pretty vital feature for those cards that cant do textured video | 01:33 |
| Sarvatt | backporting future things will be a heck of alot easier on 2.10 though | 01:36 |
| Sarvatt | i'm trying out that linux-image-2.6.32-14-generic_2.6.32-14.19drmbackportapw2 on all my machines now | 01:37 |
| Sarvatt | wow what the heck, screen is going all garbled for a few seconds when i scroll in xchat | 01:43 |
| Sarvatt | yeep, compile a kernel in the background and then scrolling in xchat does it and doesnt when i'm not compiling, guess its the client side decorations gtk+ change thats making Xorg use craploads of cpu time | 01:49 |
| BUGabundo | Sarvatt: something fishy has been around since the weekend updates | 01:50 |
| BUGabundo | several users compaling of the same slowdown | 01:50 |
| BUGabundo | some point to GTK bug | 01:51 |
| BUGabundo | not sure what | 01:51 |
| BUGabundo | but now I can type full sentnces, way before they show up in my screen | 01:51 |
| Sarvatt | yeah I know what it is, something not ready getting shoved in to make feature freeze and it'll get fixed later :) | 01:51 |
| BUGabundo | eheheh | 01:53 |
| BUGabundo | I know | 01:53 |
| Sarvatt | disabling compiz fixes the screen getting garbled under heavy system load here but metacity is *super* slow minimizing and maximizing with this gtk+ change on this netbook | 01:55 |
| BUGabundo | now I'm using compiz | 01:56 |
| BUGabundo | wasn't all week | 01:56 |
| BUGabundo | it was still super slow | 01:56 |
| BUGabundo | specailly over FreeNX and ADSL | 01:56 |
| Sarvatt | lol wonder how things handle if i launch a gtk app over a ssh connection now that you mention that | 01:57 |
| BUGabundo | eheh | 01:58 |
| BUGabundo | you would go to sleep a awake hopping it had finished | 01:59 |
| Sarvatt | what status should I set bugs to where they are for a feature that wont be included until 10.10 at least? like the backlight OSD under intel KMS bugs | 02:23 |
| Sarvatt | just leave it confirmed? | 02:23 |
| BUGabundo | probably | 02:26 |
| BUGabundo | or in progress | 02:26 |
| BUGabundo | if you have code for it | 02:26 |
| Sarvatt | so many -intel bugs that are 100% kernel bugs, no wonder theres so many open intel bugs :) | 02:29 |
| bryceh | Sarvatt, yeah I've been just moving those to the kernel when I spot them. When you spot ones that you're 100% sure are kernel bugs, retarget them to 'linux' | 06:12 |
| bryceh | I figure it's fine for people to keep filing these against xorg, since that'll ensure they get the proper debugging files before we refile to the proper place | 06:14 |
| bryceh | also if you see ones that you're pretty sure should be fine in lucid, go ahead and close them out; I've been doing similarly as I have time | 06:14 |
| kklimonda | hmm.. xserver-xorg-video-nouveau is trying to pull a -generic kernel when I already have -pae.. | 06:45 |
| tjaalton | kklimonda: because -nouveau depends on it, and there's no linux-backports-modules-nouveau-lucid-generic-pae available which it could depend as an alternative | 07:11 |
| tjaalton | so file a bug against nouveau and linux-meta | 07:11 |
| kklimonda | ach, I see - thanks. should probably wake up before I start asking stupid questions | 07:12 |
| tjaalton | no it's a valid issue that has gone unnoticed | 07:12 |
| RAOF | Hm. It seems that /lib/modules/2.6.32-$ABI/modules.order is there to provide a consistent order of module loading - is that right? Might this be why vga16fb breaks nouveau but not the other KMS drivers? All the other KMS drivers appear prior to vga16fb in modules.order. | 07:57 |
| RAOF | Well, I'll call that hypothesis... partially confirmed. Adding nouveau to modules.order before vga16fb makes boot work for me. | 09:11 |
| RAOF | And trying the reverse to i915 seems to introduce subtle corruption. | 09:14 |
| === BUGabundo is now known as BUGabundo_FEDWS | ||
| === BUGabundo_FEDWS is now known as BUGabundo | ||
| Sarvatt | ok nouveau is all working again with 3d on edgers | 15:53 |
| BUGabundo | YAY | 15:55 |
| BUGabundo | will test tonigh | 15:55 |
| Sarvatt | its even slower now, you might not want to :) | 15:56 |
| BUGabundo | ohhhh ok | 15:57 |
| * BUGabundo changes ideas | 15:57 | |
| BUGabundo | even nvidia blob is slow in lucid | 15:57 |
| BUGabundo | only 180FPS in compiz benchmark plugin | 15:58 |
| BUGabundo | I used to get over 300 in 9.04 | 15:58 |
| jcristau | 180 fps is still 3 times as much as your refresh rate, so.. | 15:59 |
| BUGabundo | jcristau: well nouveu is only at 30 | 16:00 |
| BUGabundo | well it was... if the new one is slower | 16:00 |
| BUGabundo | ..... 15FPS? lol | 16:00 |
| Sarvatt | nvidia-settings -a AccelerateTrapezoids=1 | 16:01 |
| Sarvatt | might want to try that out | 16:01 |
| BUGabundo | ORLY? | 16:02 |
| BUGabundo | ERROR: Error parsing assignment 'AccelerateTrapezoids=1' (Unrecognized attribute name). | 16:03 |
| Sarvatt | ah yeah we have an old nvidia-settings still | 16:04 |
| Sarvatt | ya can add it to xorg.conf still | 16:04 |
| bjsnider | nvidia-settings is still at 180.25? | 16:05 |
| Duke` | wow huge dead code cleanup in xf86-video-intel | 18:21 |
| Sarvatt | anyone currently using karmic on x64 that could tell me if /etc/X11/Xsession.d/65mesa-check-x86-64 exists on their system? | 19:42 |
| hyperair | ls /etc/X11/Xsession.d/65- <-- tab completion doesn't work, so i guess it isn't around. | 19:43 |
| Sarvatt | i cant find that script in any of the karmic packaging but someone is saying it exists, i'm guessing they had it from an old upgrade and didnt get removed | 19:44 |
| hyperair | hmm | 19:44 |
| Sarvatt | https://bugs.edge.launchpad.net/ubuntu/+source/mesa/+bug/460809 | 19:44 |
| ubottu | Ubuntu bug 460809 in mesa "ASM optimizations in mesa are disabled on Intel 64bit CPU" [Undecided,Confirmed] | 19:44 |
| hyperair | i'm using xorg-edgers, if that counts | 19:44 |
| hyperair | ASM optimizations? | 19:44 |
| jcristau | Sarvatt: yeah sounds like the postinst should have removed it on upgrade, and didn't | 19:47 |
| Sarvatt | tormodtjaalton: I still have /etc/X11/Xsession.d/65mesa-check-x86-64 here, is this a conf file that needs to be deleted even if it's not shipped any longer, or is it just me?21:35 | 19:48 |
| Sarvatt | jcristauyes, the postinst should take care of that | 19:48 |
| hyperair | perhaps hookscripts should never go under /etc | 19:48 |
| hyperair | like distribution-shipped hookscripts | 19:48 |
| Sarvatt | 2009/02/10 #ubuntu-x irc log | 19:48 |
| jcristau | hehe | 19:49 |
| Sarvatt | over a year ago :D | 19:49 |
| hyperair | jaunty wasn't even released then, let alone karmic >_> | 19:49 |
| dbro | Hello Xorg experts- I have a problem where xorg memory usage creeps up over time. Eg right now it consumes 40% of my 640MB ram and seems to be causing all kinds of swapping and slowness. Is this the right place to ask for help troubleshooting this? | 20:00 |
| dbro | it appears to match the symptoms in this bug report (but I don't see much guidance there about fixing it) https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/98783 | 20:00 |
| ubottu | Ubuntu bug 98783 in xorg-server "MASTER: memory leak" [Wishlist,Incomplete] | 20:00 |
| jcristau | symptoms don't really matter. what matters is an easy way to reproduce. | 20:02 |
| Sarvatt | I'm guessing you're using the nvidia proprietary drivers | 20:03 |
| Sarvatt | because thats been a problem for a *long* time with those and it doesn't look like something we can fix with those | 20:04 |
| dbro | is there an alternative? something that would play nicely with flash video would be nice | 20:09 |
| Sarvatt | alot of work is being put into the alternative (nouveau) but realistically 10.10 will be the release where nouveau is at a state where it can replace the blob for most people in my opinion | 20:10 |
| Sarvatt | if you dont mind fiddling nouveau is in lucid now, and you can get 3D support from the xorg-edgers PPA | 20:11 |
| dbro | does 10.10 mean ubuntu 10.10 ? | 20:13 |
| Sarvatt | i've already ditched the proprietary drivers on my nvidia system on lucid + xorg-edgers, but the 3D support is slow right now so if you have a high resolution screen I wouldn't bother. if you dont need 3D though the nouveau in lucid handles flash fine | 20:13 |
| Sarvatt | yeah | 20:14 |
| dbro | I dont think I need 3d. I don't play games or use fancy desktop effects | 20:14 |
| dbro | ok | 20:14 |
| Sarvatt | i would try out a lucid livecd once alpha 3 hits in a few days and see how you like it | 20:14 |
| dbro | ok thanks | 20:15 |
| Sarvatt | wish there was some magical fix for it but the memory leaks have been around for over 4 years now and the proprietary drivers are closed source | 20:16 |
| Sarvatt | hmm nouveau is still using a ton more memory than intel for me, 132mb vs 44mb.. the blob would be around 400mb by now though | 20:22 |
| Sarvatt | woohoo http://git.kernel.org/?p=linux/kernel/git/dwmw2/linux-firmware.git;a=commit;h=d9076a54d74e371a11e1206b4a26e2e428045b9e | 20:30 |
| Sarvatt | and we've got the r600+ irq firmware in our linux-firmware now too | 20:30 |
| tjaalton | Sarvatt: bah, so it seems | 22:34 |
| tjaalton | and I didn't do anything the last time | 22:37 |
| tjaalton | go me :) | 22:37 |
| tjaalton | Sarvatt: pushed something to git | 23:13 |
| dbro | does this help wrt to previous request for easy way to reproduce behavior? | 23:18 |
| dbro | switching between different desktop screens also increases memory usage | 23:19 |
| tjaalton | still, if you are using the blob there's nothing we can do | 23:23 |
| dbro | should I try to see if this also happens when using nv ? Is that helpful? | 23:25 |
| tjaalton | not really, since they don't care about that either :) | 23:26 |
| tjaalton | nouveau on the other hand.. | 23:27 |
| dbro | ok, I'm willing to give it a try- this memory leak is driving me mad! | 23:27 |
| tjaalton | ram is cheap, just buy some more :) | 23:28 |
| dbro | that would get me a longer time-to-restart, but not a fix, right? | 23:29 |
| tjaalton | right | 23:34 |
| bjsnider | Sarvatt, the nvidia blob has a memory leak issue? impossible | 23:38 |
| tjaalton | also, it's not like the xserver doesn't have memleaks of it's own, but in this case the driver is most likely to blame | 23:40 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
