Sarvattlet me know if xserver-xorg-dev, libpixman-1-dev, x11proto-core-dev, is enough for your pbuilder build RAOF, thanks for the heads up00:02
Sarvatti'll try building without the xorg state tracker to see if it goes00:02
* bryceh waves00:03
jcristauxserver-xorg-dev should be enough, it'll pull in the rest00:03
Sarvattdon't know if --with-state-trackers=dri is enough, i always built it with --with-state-trackers=dri,glx and i used --with-state-trackers=dri,xorg,glx that time because thats what fedora uses :D00:05
Sarvattheyo bryceh!00:05
SarvattRAOF: do you have edgers set up in your pbuilder? because you need the newer protos00:06
RAOFI do have edgers set up in the pbuilder.00:09
RAOFOne of the few nice side-effects of pbuilder00:09
RAOFOne of the few nice side-effects of pbuilder-dist always passing --override-config is that you can add --othermirror very easily00:10
Sarvattjcristau: xserver-xorg-dev pulls in libpixman-1-dev?00:13
RAOFSarvatt: Looks like everything's building nicely00:13
Sarvattalready uploaded with the new deps, hopefully it wont get mozilla-ed00:13
jcristauSarvatt: yes00:13
Sarvattsweet, didn't know that00:14
Sarvattoh yeah it does now that i look at the control00:14
RAOFMmm.  pbuilder-on-tmpfs is wonderfully fast until you start to hit swap :)00:15
kklimondaheh, or (in case you have no swap) your builds start to fail ;)00:15
Sarvatti ran out of hdd space building mesa last time, thats why it took so long to see if the rule changes worked :D00:16
Sarvattused up around 2 gigs building00:16
RAOFWooop!  Ran out of swap space!00:17
kklimondabtw, what is the status of hibernation with nouveau drivers? have devs started working on that?00:19
RAOFI don't know; does it not work now?00:20
Sarvatti'd blame other userspace stuff for suspend/hibernate not working, haven't looked through pm-utils fully yet but i'm pretty sure there are some spots where it checks for the quirks to do based on the normal module names00:20
RAOFAFAIK it should work, same as suspend.00:20
kklimondaRAOF, it seems to hibernate just fine (i.e. computer shuts down) but after I start it again I get a gdm login prompt00:22
kklimondano idea how to debug hibernation on Linux as it has never worked for me00:22
kklimondawhich isn't really an excuse now that I think about it.. :D00:23
Sarvattguessing you mean it doesnt restore the old gdm session after you login again too?00:23
* Sarvatt isnt sure if a gdm login prompt after hibernate is the expected behavior because he's never hibernated00:23
Sarvattno swap partition here00:24
Sarvattfinally mesa shows up on edgers, building now00:26
Sarvatt/dev/dri/card0 didnt have the right permissions on nouveau without the udev change in the xorg-edgers/nouveau udev package00:27
kklimondawhat are the right permissions? 0660 root:video ?00:29
Sarvattlooks like its fine with --with-state-trackers=dri,glx, i'll do tomorrows update with that00:39
* johanbr reboots to try again00:46
Sarvattits not done building, what are you trying?00:46
Sarvattyep works fine without the xorg state tracker00:49
Sarvattalmost done building on edgers, its packaging now00:50
RAOFWoop!  Woop!00:52
Sarvattjohanbr: its not ready until you get mesa 7.8.0~git20100207.1a45c2bc-0ubuntu0sarvatt3 offered, you need both xorg-edgers/ppa and xorg-edgers/nouveau in your sources too incase you missed that part, then you just install xserver-xorg-video-nouveau and reboot/restart gdm00:55
Sarvattcan someone pass that along when he comes back? i have to run again00:55
RAOFSarvatt: Certainly.00:56
johanbrhmm... nouveau seems to have trouble opening the DRI device: http://pastebin.com/f1cb1a31301:02
johanbrand the permissions on /dev/dri/card0 are 660 root:video01:03
johanbrshould the kernel modules with or without lbm_ be loaded?01:08
Sarvattwhen's the last time you did a dist-upgrade? Current version of pixman: 0.16.2??01:09
Sarvatt<Sarvatt> johanbr: its not ready until you get mesa 7.8.0~git20100207.1a45c2bc-0ubuntu0sarvatt3 offered, you need both xorg-edgers/ppa and xorg-edgers/nouveau in your sources too incase you missed that part, then you just install xserver-xorg-video-nouveau and reboot/restart gdm01:09
Sarvattdoes dmesg | grep efifb show anything?01:10
kklimondaSarvatt, works fine for me01:11
johanbrI built the mesa package locally, maybe that's the problem...01:12
johanbrwe'll see how it goes after a reboot01:14
Sarvatthow did he manage to have a pixman version over a month old is what i'm wondering01:15
Sarvattkklimonda: compiz work? tried any 3D apps yet? I'm doing it all remotely so beyond compiz and glxgears I haven't seen if anything works01:17
Sarvatti dont know if theres any visual corruption in compiz either, just that it didnt give any errors01:18
kklimondaSarvatt, well - compiz works just fine. the quality of scaled windows (in alt+tab, super+w) isn't too good but it may be something I can tweak01:18
RAOFYou'll probably find gnome-shell exposes a little bit of corruption.01:19
kklimondaSarvatt, my cats are alive so at least it doesn't kill wicked ones ;)01:19
Sarvatttheres probably something you can tweak in driconf to fix that01:19
Sarvattkitty halftime show on the puppy bowl is over :(01:20
Sarvattthe alt-tab display looks like crap on intel here too01:21
johanbrthere we go... problem disappeared after I uninstalled plymouth01:22
Sarvattyeah not surprised there, i have to do all kinds of junk to get plymouth working right on every machine i have01:22
Sarvattits segfaulting every 3rd boot or so on intel too, and getting hung displaying the splash every time ureadahead profiles01:23
Sarvatthow did you get in a state where you have pixman 16.2 though?01:26
johanbrI'm a bit leery of running full dist-upgrades, so I tend to upgrade packages by hand every now and then01:27
johanbrguess I mixed pixman01:27
Sarvatti was a bit worried that you might have missed a grub fix that was causing the same error you were getting on warm boots here01:27
johanbrthat is also possible, and that removing plymouth just papered over the symptoms01:29
johanbranyway, time to head home... thanks for the help!01:30
* RAOF checks the -rc7 gitlog to discover whether there's any nouveau updates that could be shoved into l-b-m-nouveau.01:39
BUGabundoRAOF: bug I mention in cube with nouveau gallium01:54
hyperairoh hey compiz works on nouveau =O01:55
BUGabundospend the evening testing it01:56
BUGabundowant my log from +1 ?01:56
hyperairno need =)01:56
hyperairi'll test it myself when i get home =p01:56
BUGabundo(01:46:17 AM) freenode: conclusion: nouveau gallium 3D support works ok, but is *very* slow01:56
hyperairoh heh01:57
BUGabundo(01:45:15 AM) freenode: TOTALLY FORGET RAIN EFFECTS01:57
BUGabundo(01:43:47 AM) freenode: windows previews FAIL01:57
hyperairmy card can't support rain anyway >_>01:57
kklimondaBUGabundo, how do the fail?01:57
BUGabundo(01:38:32 AM) freenode: if I go from Compiz to metacity, I get a freezes background01:57
BUGabundo(01:43:59 AM) freenode: got the windown, but nothing in there01:57
BUGabundo(01:43:10 AM) freenode: selective zoom works01:57
BUGabundo(01:43:25 AM) freenode: all apps switcher seem to work01:57
BUGabundo(01:42:52 AM) freenode: both apps expose and Desktop expose work01:58
BUGabundo(01:42:57 AM) freenode: fire writing works01:58
kklimondanow the next linux graphics gizmo I'm interested in is wayland..02:00
BUGabundoRAOF: is it normal for collors to seem off with this driver??02:00
BUGabundosome yellows are no longer as yellow02:00
* BUGabundo wonders if that is sleep deprivation 02:00
kklimondait may be02:00
RAOFNo; it should be at least as good as nv.  If you happen to have an 18bit LDVS panel (and your laptop almost certainly does), it should actually look better.02:00
RAOFBecause nouveau sets the dither mode properly.02:01
BUGabundoI noticed it looks smoother02:01
BUGabundobut collors are off02:01
BUGabundoeven greys 02:01
hyperairor colors.02:01
BUGabundogramar nazy02:01
BUGabundothanks !02:01
BUGabundoRAOF: any way I can test this better?02:03
BUGabundoor make sure I got one of those 18bit LDVS panel  ?02:03
RAOFWell, if you didn't pay lots and lots of money to get a 24bit panel, you've got an 18bit panel.02:03
BUGabundogot it02:03
BUGabundothe screen look brighter too02:04
RAOFAnd just general testing; suspend, resume, hibernate.  Use it for a while; does it die screaming after 5 hours?02:04
RAOFThat sort of thing.02:04
BUGabundothen again I'm tired 02:04
BUGabundosuspend test coming up02:04
RAOFJust use it as your driver; if it's not obnoxious, great.02:05
=== BUGabundo1 is now known as BUGabundo
kklimondaBUGabundo says that suspend has failed for him - "I had sound a net, but no x or tty acess"02:10
Sarvatthow are you getting gallium 3D support BUGabundo? building mesa yourself? i only added it to edgers a few hours ago, sure you aren't just using swrast?02:49
RAOFCompiz would bomb out with swrast.02:52
kklimondaSarvatt, he uses xorg-edgers ppas02:53
Sarvatti just dont know how he could have spent the evening testing it when it was only up for under an hour when he said that and i only mentioned it in here, sounded like he might have been using swrast :)02:57
kklimondaSarvatt, either I or RAOF has mentioned it to him on #ubuntu+102:58
Sarvattno problems here so far though outside of a 5 second stall closing the glxgears window once under compiz, openarena extreme tux racer and neverball are all very playable at 1280x80002:58
Sarvattthe alt-tab graphics look as nasty under nouveau as intel, comparing them side by side :D02:59
Sarvattreally ugly text rendering scaled down like that03:00
kklimondaI'm suprised how good 3D support is03:01
kklimondaI was expecting it to bork my system completely03:01
johanbryes, me too03:02
johanbra few minor rendering artifacts, but overall it works very well03:02
RAOFI get a very fetching all-white screen.03:02
Sarvattdoing what?03:03
Sarvattfont shadows look a little funky03:03
RAOFStarting compiz.03:03
Sarvattguessing everyone else testing is using nv50 :D03:04
RAOFAh.  Because I've got bad permissions somewhere.03:05
Sarvattya get the udev upgrade?03:07
johanbrchanging the window translucency looks a bit funky...03:08
Sarvattstandard main compiz profile RAOF? or do you have any extra plugins enabled?03:08
RAOFNo, I don't think I do have the udev upgrade.  Where is it?03:09
Sarvattshould be 1:149-6~nouveau03:09
Sarvattits in xorg-edgers/nouveau03:09
RAOFOh.  I *can* make compiz SIGFPE in nouveau_dri.so :)03:10
Sarvattwonder if its related to not having ctxprogs from the blob on your nv4003:12
RAOFPossibly.  I'll restart using the blob's ctxprogs.03:12
RAOFI think it's more likely to be AIGLX being broken.03:12
Sarvattprobably, seems to work decently on nv50 at least. going to test the patched lbm-nouveau that doesnt need firmware for nv50 again in a bit03:15
johanbrdid anyone try gnome-shell?03:15
RAOFWell, that wasn't a ringing success; it seems setting ctxfw=1 is a good way of making nouveau just not come up.03:16
Sarvattoh it doesnt add the firmware to the initrd03:16
RAOFIt doesn't?03:16
Sarvattnoticed that when i built the one that doesnt need firmware, it only installed the firmware for the other models that need it03:17
Sarvattmodinfo lbm-nouveau lists the stuff it adds03:17
kklimondajohanbr, not really - we all expect it to fail though ;)03:18
johanbrhmm, sounds like a challenge ;)03:19
Sarvatti dont, mutter works fine :)03:19
Sarvatti did test that out, just not the full gnome-shell03:20
johanbrstarting gnome-shell causes a hard lockup03:23
johanbrand no DRI after a cold boot03:24
RAOFSarvatt: Have you actually uploaded udev 149-6~nouveau?  I can't see it.03:26
RAOFNo, it should be there...03:26
RAOFWhy can't I see it?03:27
Sarvattwhat version do you have now?03:28
Sarvattyou can just makes these changes to the 3 files in /lib/udev/rules.d/ http://sarvatt.com/downloads/patches/udev-lbm-nouveau.patch03:29
Sarvattyou have xorg-edgers/nouveau enabled right?03:30
Sarvattthat'd be the only reason i can think of03:31
RAOFI do, yes.03:31
RAOFOh.  It seems my apt-proxy was being stupid.03:36
RAOFWith bonus corrupted VTs!03:47
BUGabundoguud afternoon14:10
BUGabundoFYI suspend to RAM is broken with nouveau gallium on nvidia 8400gm14:39
BUGabundofunny view of Cube bugs14:43
kklimondaBUGabundo, what's weird is that it works on my quadro 140m just fine (and quadro is based on 8400something)14:49
BUGabundomaybe an option14:50
BUGabundodon't know14:50
BUGabundomaybe Reflexion plugin14:50
kklimondaright, you should try with default plugins probably14:51
BUGabundo3D Windows BUG14:59
BUGabundo"Paint Windows Like Backgroud" seems to be the cause of my Windows Preview bug15:02
=== yofel_ is now known as yofel
BUGabundowindows shifting may be slower16:36
BUGabundobut browsers and some ups, scrolling is *much* smoother16:36
BUGabundothere's no comparition in Firefox16:37
hyperairno comparition?16:38
BUGabundomuch nicer16:38
BUGabundoyea, ytpo16:38
hyperairwell i've heard nouveau does well in the 2D acceleration field16:39
hyperairstuff like banshee's smooth scrolling works better on nouveau too.16:39
hyperairalthough it's very jerky with intel16:39
BUGabundoit needs a bit work here : http://dl.dropbox.com/u/112892/ubuntu-x.png16:40
BUGabundoother then that, its very nice16:40
hyperairthat looks bad16:41
BUGabundoeverything from tooltips, to other  overlays is like that16:42
BUGabundohyperair: I'm almost sure it's a bug in 3D Windows plugin16:43
BUGabundoturning that off it does improve a lot16:43
hyperairi see.16:43
BUGabundokklimonda: as in stuff that goes overing other16:43
* hyperair doe more16:43
hyperairstupid lagging X16:43
hyperairit just drops my keyboard events damnit >_>16:43
hyperairi meant i don't use the cube any more16:44
BUGabundoits _pretty_ 16:45
hyperairyeah, but not practical.16:45
hyperairi have three rows of workspaces16:45
BUGabundoexpose and zoom are nicer and usefull16:45
hyperairit's a 3x3 grid.16:45
BUGabundousually have 4 desktops filled16:45
hyperairso with a maximum of 2 desktop switches, i can switch to any workspace in that 3x3 grid.16:45
hyperairand have 9 workspaces.16:45
hyperairthe cube can't do that =\16:45
hyperairif the cube supported multiple rows of workspaces, then sure16:46
BUGabundowindows expose FTW16:46
hyperairwindows expose?16:46
hyperairyou mean scale?16:46
* hyperair uses that intensively16:46
hyperaireven for closing windows16:46
BUGabundogot to have a BIG screen16:47
BUGabundoto hit those X16:47
hyperairwhat X?16:47
BUGabundoclose wind16:47
BUGabundothe cross 16:47
hyperairum you close the X in expo mode?16:47
hyperairthat's not scale16:48
hyperairthat's expo!16:48
hyperairi'm talking about scale. middle clicking on a window closes it16:48
BUGabundothere used to be this 3rd party compiz plugin that would stick scale on a secondary Monitor16:48
BUGabundoI loved it16:48
hyperairnever used it16:48
BUGabundohttp://dl.dropbox.com/u/112892/expose2.png this ?16:49
hyperairyeah tht16:49
hyperairmiddle click on a window and it closes16:49
BUGabundonice to find a wind16:49
* BUGabundo tries16:49
hyperairand rightclick on it and it zooms in to normal size16:50
hyperairif you've got the text filter thing you can even filter the windows based on title16:50
BUGabundomine does ZOOM16:50
hyperairconfigure it in scale addons16:51
BUGabundoits already enabled16:53
BUGabundomaybe it conflicts with zoom16:53
BUGabundowow that theme is scary16:53
hyperairand no it doesn't conflict with zoom16:54
BUGabundothen I'm doing it wrong16:54
hyperairi think compiz by default has it enabled16:54
hyperairer has it enabled by default16:54
hyperairwhat's scary about my theme anyway? O_o16:55
hyperairor am i missing something16:55
BUGabundobright blue16:55
hyperairis bright blue scary?16:56
BUGabundofor a DE: YES16:56
hyperairbut why?16:56
BUGabundolets see: super + w, get it scaled, then without letting go, mouse middle click?16:56
BUGabundodoesn't work16:57
hyperairoh, i see.16:57
hyperairyou're using the keyboard!16:57
hyperairit doesn't work that way16:57
hyperairyou need to trigger the scale using one of the hot-edges16:57
kklimondait works from keyboard here16:57
hyperairkklimonda: the scale addons thing?16:57
BUGabundoconfilts with Super-mouse2 for zoom box16:58
hyperairBUGabundo: you're doing it wrong >_>16:58
BUGabundoI don't use mouse!!!16:58
BUGabundoif I can avoid it16:58
BUGabundoso corners is a NO GO16:58
hyperairokay, let's think..16:58
hyperairthere are keyboard shortcuts available for setting in scale addons16:58
kklimondahyperair, yes16:58
hyperairset something there16:58
hyperairBUGabundo: and set your scale plugin's behaviour to not require you to hold down your key16:59
hyperairBUGabundo: like when you release your scale's keybinding, it should stay in scale mode16:59
hyperair"Key  bindings toggle scale mode"17:00
BUGabundonever mind17:00
BUGabundodon't use it either17:00
BUGabundowould eventually forget about it 17:01
=== johanbr_ is now known as johanbr
Sarvatthmm, should this be fixed released or wontfix? his problem is r600_dri.so doesnt exist in karmics ia32-libs, its fixed in lucid but wont be fixed in karmic https://bugs.edge.launchpad.net/ubuntu/+source/mesa/+bug/51593318:52
ubottuUbuntu bug 515933 in mesa "[xorg-edgers] after rebuilding, Wine doesn't get same results with DRI R600 as glxinfo" [Undecided,New]18:52
Sarvattfix released it is, i can't set wontfix anyway :)18:55
Sarvattdarn.. ya got me hoping the fix in your bug would fix the flickering post-resume alot of us are getting on 945 jcristau :D http://bugs.freedesktop.org/show_bug.cgi?id=2537119:06
ubottuFreedesktop bug 25371 in DRM/Intel "[945gm] lvds panel blinking from time to time" [Normal,New]19:06
Sarvattstill getting the flickering and lockup post resume here with drm-intel-next though19:07
Sarvatt(with powersave=1)19:07
Sarvatti still cant suspend if i ever mount a sd card on the lastest .33, thats a nasty one19:08
knittlEVIOCGNAME <- could this be related to failing wakeup from standby?19:19
Sarvattnah thats normal right now with the udevification and harmless afaik19:22
Sarvattlooks like its loading evdev for every device and when something else like synaptics takes it over it gives that error and unloads evdev for that device unless I'm mistaken19:23
Sarvatt(EE) ioctl EVIOCGNAME failed: Inappropriate ioctl for device <- that?19:24
Sarvattjbarnes: are you around by any chance? I have a question about this commit http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=7121413f2accf14cf05b38539fb7a8be7754337019:25
Sarvattaspire ones have a DMI product name of AOA110 or AOA150 and not Aspire one so I'm not sure it ever gets used?19:26
jbarnesSarvatt: hm19:26
Sarvattmine is quirky and always reports the lid closed to devicekit-power, I cant boot with the lid closed19:26
jbarnesworked on my machine19:26
jbarnesit had an old bios though iirc19:27
Sarvattwould something like this potentially work? I'm no kernel hacker, not sure if its worth fixing kernel-package to build a new kernel to try it :D http://sarvatt.com/downloads/patches/0001-Make-drm-i915-blacklist-Acer-AspireOne-lid-status-ac.patch19:28
Sarvattam I supposed to be able to boot with the lid closed though? not sure if thats intended or if that would fix that19:28
jbarneswell you should probably leave the existing one alone19:29
jbarnessince I tested that on my aspire one19:29
jbarnesthe others look ok, though I'm not sure about the * matching19:29
Sarvatti'm using a gateway bios on my acer aspire one, its the same thing as the acer bios except with a gateway vendor name and it has more brightness levels exposed, acer blocked out the lower brightness levels because it causes backlight flickering but i like the lower power :D19:29
jbarnesif you've tested and confirmed against those machines and it makes things work, definitely send a patch to intel-gfx and cc eric@anholt.net19:31
Sarvattok I'll make another leaving the Aspire one match alone and add AOA1* seperately, I'm positive AOA1* only matches AOA110 (ssd model) and AOA150 (hdd model) aspire ones. I just wasn't sure if that commit was intended to let you boot with the lid closed because I dont have a problem booting with the lid open and didnt before that commit and I'm pretty sure that commit isn't doing anything here because Aspire one isnt anywhere in the dmi 19:35
Sarvattwill try it out, thanks :)19:36
Sarvattdevkit-power always reports lid-is-closed:   yes on this thing19:36
Sarvattnouveau has a module option to ignore acpi lid status that I need to boot if the lid is closed, guess its intentional that things dont start up right with a closed lid19:39
Sarvattwas hoping [PATCH] drm/i915: handle FBC and self-refresh better would fix the blinking and eventual hangs with intel_gpu_top reporting 100% use in framebuffer compression that only happens after resume here on this netbook but no luck :(19:41
Sarvatthmm powertop saying Xorg is writing to /var/log/Xorg.0.log preventing the hdd from sleeping but its not writing anything to it, wonder what thats about19:48
johanbrSarvatt, I've noticed that powertop "writing to..." line can be pretty inaccurate19:50
johanbrsometimes it claims that user programs are writing to /usr19:51
Sarvattjbarnes: just curious, but have you tried using intel 2.10 on your aspire one? I'm not aware of *anyone* able to use 2.10+ without random batchbuffer I/O error hangs on 915-94519:51
Sarvattsorry to bug you so much :)19:51
Sarvattsome change in libdrm between 11-25 and 12-03 made 915-945 completely unstable with intel 2.10 on any kernel19:53
Sarvattthats blocking us going to 2.10 in lucid :(19:54
jbarnesSarvatt: I haven't tried it yet19:54
Sarvattheyo BUGabundo 19:54
jbarnesbut yeah I've heard similar reports19:54
jbarnesa bisect might help19:54
Sarvatti can't bisect libdrm to find out what did it because almost all the commits between are unstable on their own, I've been trying for months19:55
Sarvattbut I know a libdrm git checkout from 11-25 is stable with intel 2.1019:56
Duke`During the past days, I suffered less often of the Execbuf Error19:56
jbarnesSarvatt: just libdrm huh?19:56
jbarnesI figured it must be an xf86-video-intel problem19:57
jbarnesbut if it's libdrm that could be even more serious19:57
Sarvattyeah some interaction between libdrm and the new things in the ddx there, libdrm git is 100% stable on its own with intel 2.9.119:57
Sarvattand intel from 11-11 is fine with the newer libdrm as well19:58
jbarnesis there new execbuf2 stuff in there possibly?19:58
Sarvattit happens on 2.6.30 and 2.6.31 kernels as well19:58
Sarvattintel git was broken between 11-11 and 12-06 or so for me as well so I can't bisect the driver to see what change in there messed it up20:00
brycehheya jbarnes20:01
jbarnesbryceh: hi20:01
BUGabundo 1211   0.80s   0.07s     0K     0K     0K     0K  --   - S  44% Xorg20:10
BUGabundoI'm seeing much more CPU usage since I changed to nouveau gallium20:10
Sarvattseeing alot of  invalid BO alignment 4096 crashes under 2.6.31 as well, at least those dont happen under .32+ but people with the karmic kernel are having problems using edgers20:10
SarvattBUGabundo: just dont expect nouveau gallium to work right, the call for testing was for the xorg-edgers/nouveau PPA because that's what's going to be in lucid :D20:12
BUGabundoSarvatt: that's what I started using last night20:12
Sarvatti'm just enabling 3D support in edgers for people that want to break things, they dont even want upstream 3D related bug reports yet because they know its broken :)20:12
BUGabundoits terribly slow20:12
BUGabundoI WANT 3D !  I WANT 3D !  I WANT 3D !  I WANT 3D !  20:13
BUGabundoThat's the only reason I installed it20:13
BUGabundo3D is NOT broken20:13
BUGabundoit works fine here20:13
Sarvattthat will change on a day to day basis for sure, it breaks all the time :D20:13
kklimondaSarvatt, so you were just lucky when grabbing git snapshot? :)20:14
Sarvattwell unlucky for some people, at least we're lucky its working well on nv50 :) its completely usable here, i already removed the blob20:15
* BUGabundo stabs Sarvatt20:15
BUGabundodon't go kill it for me20:15
BUGabundoI need it on its BEST mode for Saturday FLOSS presentation20:15
BUGabundogonna try getting a few more users to Ubuntu 20:15
BUGabundoif Compiz is broken, then Ill lose points!!!20:16
BUGabundoyou don't want that on your shoulders do you ? 20:16
SarvattBUGabundo: not sure if you got the message but its expected to use both xorg-edgers/ppa and xorg-edgers/nouveau at the same time if you use nouveau20:16
BUGabundothat's what I have20:17
Sarvattokie, just wanted to be sure since i said it in here when you werent around :)20:17
BUGabundoI followed RAOF pointers20:17
Sarvatttheres udev and plymouth updates to work for nouveau in xorg-edgers/nouveau that i didnt put in xorg-edgers/ppa yet20:17
BUGabundo 1211   0.98s   0.18s     0K     0K     0K     0K  --   - D  57% Xorg20:17
BUGabundostill this is TOO MUCH cpu20:17
Sarvattthats using compiz?20:17
BUGabundonv blob was much nicer20:18
BUGabundo$ ls20:18
BUGabundototal 12K20:18
BUGabundo-rw-r--r-- 1 root root 67 2010-02-07 23:13 xorg-edgers-nouveau-lucid.list20:18
BUGabundo-rw-r--r-- 1 root root 63 2010-02-07 23:13 xorg-edgers-ppa-lucid.list20:18
BUGabundo-rw-r--r-- 1 root root 63 2010-02-07 23:13 xorg-edgers-ppa-lucid.list.save20:18
Sarvattpastebin your xorg.0.log and dmesg? just curious what they look like on your system with that setup to see if i spot any more errors :D20:19
Sarvattbtw if you want to go back to nouveau without 3D support just sudo ppa-purge xorg-edgers, it'll just remove xorg-edgers/ppa and leave xorg-edgers/nouveau20:20
kklimondaSarvatt, have I already thanked you for the ppa-purge script?20:20
kklimondaSarvatt, it's the best thing since sliced bread ;)20:20
kklimondaok, I got weird font colors in xchat :D20:20
Sarvattthe -p argument to ppa-purge is broken after some changes tormod did so you cant remove xorg-edgers/nouveau yet :(20:20
Sarvattno worries kklimonda glad it helped, surprised there wasn't anything similar out there already20:21
BUGabundoSarvatt: $ pastebinit Xorg.0.log http://paste.ubuntu.com/372001/20:21
BUGabundo$ dmesg | pastebinit http://paste.ubuntu.com/37200320:21
BUGabundoDescription: Disables a PPA and reverts to official packages20:22
BUGabundo This package provides a script capable of automatically downgrading all20:22
BUGabundo packages from a given PPA back to the official Ubuntu versions.20:22
BUGabundohey that's nice20:22
* BUGabundo is sad20:32
BUGabundothis drivers suck for video!!!!20:32
BUGabundocan't watch a single video .... it just shows horizontal strikes20:32
kklimondawhoa, I've just all my fonts.. they have all disappeared20:33
brycehBUGabundo, please keep comments constructive20:35
BUGabundokklimonda: also you no longer know how to speak english20:35
BUGabundobryceh: I know20:35
BUGabundojust frustrated20:35
brycehBUGabundo, perhaps take a break and come back to it later when you're more calm?20:36
BUGabundobryceh: no worry my friend20:36
BUGabundoall cool20:36
BUGabundoI wouldn't be testing this, if I wasn't capable of having it breaking on me, would I ?20:37
BUGabundojust that sometimes, one of you guys can have some tips that might make this better for me (and everyone else) or maybe I provide some useful report20:37
brycehBUGabundo, keep in mind people on this channel are busy with development work so requests for help on bugs often interrupts other important work, so it is a good idea to avoid impoliteness when asking for help20:39
BUGabundookidoki. go back to your work... I'll be in +1 :D20:40
BUGabundohelping others too20:40
=== BUGabundo is now known as BUGabundo_dinner
Sarvattit being slow is fully expected unfortunately, i'm worried about bugs like kklimonda's fonts disappearing.. is that with compiz disabled or enabled?20:56
kklimondaSarvatt, it was with compiz enabled but replacing it with metacity haven't fixed that20:57
kklimondaSarvatt, only after I've switched to VT and back again I got fonts back20:57
Sarvattdo you have Force Synchornization between X and GLX enabled in compiz?21:00
Sarvattlooks like you're hitting https://bugs.launchpad.net/bugs/517807 too btw BUGabundo_dinner 21:04
ubottuUbuntu bug 517807 in linux "iwlagn: Failed to allocate SKB buffer with GFP_ATOMIC. Only 0 free buffers remaining." [Undecided,New]21:04
=== BUGabundo_dinner is now known as BUGabundo
BUGabundoSarvatt: for exactly what symptom ?21:08
Sarvatti just saw the [22076.107618] swapper: page allocation failure. order:2, mode:0x4020 in your dmesg21:09
BUGabundoah thanks21:09
BUGabundoSarvatt: my X fallback to no WM, I $DISPLAY=:0 compiz --replace &, but now I lost my top bar (gnome)21:11
coz_just noticed nvidia drivers showing up in hardware drivers... is it safe to install and reboot??22:15
tseliotyes, it should be22:16
tseliotif you're referring to lucid22:17
coz_tseliot,  ok let me try it now    22:17
* coz_ crosses fingers :)22:17
coz_tseliot,  yes I have another system with lucid on it  so its installing now 22:17
coz_tseliot,  do you know what the issue was this week ??  plymouth?22:18
coz_nope  nvidia still not installing 22:23
coz_let me reboot this again22:24
coz_doesnt even go to text console22:24
coz_pure black lol22:25
coz_ah another reinstall I suppose lol22:25
bjsnideris that problem with nvidia-current being blacklisted fixed?22:29
coz_not sure   there were no drivers even listed in hardware drivers  all week22:30
coz_manual install  didnt work last time I tried22:30
coz_system freeze with manual install22:30
brycehRAOF, Sarvatt, question on https://edge.launchpad.net/~xorg-edgers/+archive/nouveau - is this all stuff staged to be uploaded for Lucid?  Any tasks needing done before I start looking at uploading these packages?22:59
brycehAlso, do you suggest a particular order to upload them?22:59
RAOFI think it's all ready to go, modulo the obvious versioning.23:02
kklimondabryceh, right now you can't install lbm -generic and -generic-pae packages at the same time - that should probably be fixed (even if they aren't to be installed side-by-side a Conflicts: field have to be added).23:02
RAOFkklimonda: Thanks for reminding me.  Yes.23:02
brycehdoes lbm need uploaded before the X bits?23:03
RAOFThat should be fixed; l-b-m-n ships an initramfs hook.  That should be moved to the existing framebuffer copyfest.23:03
RAOFYes.  The X bits need lbm before they'll work at all.23:03
Sarvattthe linux-backports is the most important one to get in, ROAF would be the one to ask though because I'm not sure whats going on with the libdrm/nouveau packages in there. looks like he has the libdrm updates queued in git though. it would be nice to have a metapackage in lbm that the ddx can depend on so its not broken every abi bump before going in as well23:03
brycehok, so  0. nouveau-firmware, 1. lbm, 2. libdrm, 3. -nouveau, 4. xorg-server23:04
brycehwhat about the plymouth and udev bits?23:05
brycehguess those should follow lbm?23:05
RAOFI think they want to precede lbm, actually.23:05
Sarvattthose dont hurt anything if they are uploaded early23:06
brycehok, so  0. nouveau-firmware + udev + plymouth, 1. lbm, 2. libdrm, 3. -nouveau, 4. xorg-server23:07
Sarvattit shouldnt hurt anything if xserver is updated too before the rest too, it'll just fail to load nouveau and move on with no xorg.conf. it was just back when we had a default xorg.conf in early karmic with 1 device section and no named driver in there that we couldn't use those23:08
* bryceh nods23:09
brycehI put it last though since nouveau is not yet in main, so having it load nv is going to still be appropriate until we're closer to that point23:09
bryceh(but perhaps I'm wrong?)23:10
Sarvattyeah, if anything it'll just be opt-in for people who choose it in an xorg.conf without that patch23:10
Sarvatta libdrm version bump would be nice right about now.. :)23:11
RAOFWe'd need to pull in a new DDX too, though.23:11
Sarvattcould we sneak nouveau-firmware in the lbm package with the other firmware in it? :D23:12
brycehis nouveau-firmware going to be a Ubuntu-X package, or the kernel team's?23:13
Sarvattit'll go away lucid+1 almost guaranteed23:14
bryceh(I've been assuming the latter)23:14
brycehoh btw I talked to them about this a bit last week, and it sounded like they'd be keen on backporting the upstream stuff that makes it go away, so that it never becomes an issue for lucid23:14
brycehso if you have recommendations for what to backport, we should just let them know23:15
Sarvatti've been using that all this time with no problems but i'm sure theres some reason it's not upstream yet :D23:16
Sarvatti got nouveau gallium set up on xorg-edgers/ppa to play with too if you didn't see, compiz is working great here and the only game I've been had problems with is nexuiz (who doesnt have problems with that one)23:20
RAOFIt doesn't work for me, though, which is not great :`(23:21
Sarvattoh yeah about that - glxinfo show gallium being used? was it just compiz starting that killed things for you? tried any 3D games or anything without compiz?23:24
SarvattI'll break the kernel API and bump to 0.0.16 real soon now anyway, so23:27
SarvattI'll fold the changes in when I update libdrm for it.23:27
RAOFCompiz killed things; texture_from_pixmap is broken.23:28
SarvattLIBGL_ALWAYS_INDIRECT=1 compiz --replace work?23:28
RAOFOh, it probably would, wouldn't it.  Let me check.23:30
Sarvattah compiz --indirect-rendering does the same thing23:36
jcristauoh, nice, tormod got a newer radeontool into debian23:37
brycehSarvatt, RAOF, I added a small "how to revert nouveau" section at https://wiki.ubuntu.com/X/Nouveau - could you take a minute to look at that and let me know if I missed anything we'd need to do?23:53

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!