/srv/irclogs.ubuntu.com/2011/02/14/#ubuntu-x.txt

tjaaltonRAOF: whee, you found the commit to fix the mesa assert?08:58
tjaaltonI was told that the assert itself was unnecessary, but that there probably was something to it08:58
tjaaltonnouveau dri hasn't crashed on me during the weekend.. helps if the window count is kept low09:06
tjaaltonbut can't understand why links from TB don't work on my desktop, laptop is fine09:07
tjaalton(ffox opens them as 'www.%u.com')09:07
RAOFtjaalton: You're after the nv50 hunk of 7401590de for the mesa assert.09:14
RAOFWhich, piglit pending, should totally be nominated for 7.1009:15
hyperairhi. where can i find xrandr's C functions documentation?09:21
hyperairthat monolithic manpage isn't bringing me anywhere09:22
jcristauread the protocol spec09:22
jcristauthe libXrandr api is a thin wrapper around that.09:22
hyperairhmm okay09:22
hyperairbut what about the strange datatypes that it returns?09:22
hyperairall i'm trying to do is get the current resolution09:23
jcristaulike what?09:23
hyperairof individual outputs09:23
hyperairXRRScreenConfiguration*09:23
hyperairXRRScreenSize*09:23
hyperairyeah, stuff like that09:23
jcristauwhat's strange about a struct?09:24
hyperairoh wait a sec, the structs are documented!09:24
hyperair=O09:24
* hyperair facepalms09:24
tjaaltonRAOF: ah, cool.. will check09:27
RAOFhyperair: :)09:27
hyperairRAOF: are the stuff mentioned inside XRandr(3) all the functions i can use? the protocol document seems to mention some other stuff..09:43
hyperairlike RRGetOutputInfo and the like09:43
jcristauno, the manpage wasn't updated after 1.109:43
RAOFhyperair: I'd be looking at the protocol doc in x11proto-randr-dev09:44
hyperairoh okay.09:44
jcristauand <X11/extensions/Xrandr.h>09:44
hyperairthanks.09:44
klattimerDoes anyone have any ideas about this bug; https://bugs.launchpad.net/compiz/+bug/71866311:07
ubot4Launchpad bug 718663 in xserver-xorg-driver-vesa (and 2 other projects) "Blank unity desktop (affects: 1) (heat: 8)" [Undecided,New]11:07
jcristaunot enough info there.11:09
tjaaltonyep, best to use 'ubuntu-bug' when filing a new bug11:11
tjaaltonthere are three "upstreams" marked for that bug11:11
tjaaltonor projects, none of which is ubuntu11:12
klattimertjaalton: ah :/ 11:18
klattimerwhat info do you need?11:18
tjaaltonklattimer: please run 'apport-collect 718663'11:24
tjaaltonit'll attach relevant information to the bug11:25
klattimertjaalton: done11:25
klattimerthis'll take a minute11:26
tjaaltonhmm does it have a poulsbo in it?11:27
tjaaltonguess we'll see11:27
tjaaltonah no11:27
klattimertjaalton: sorry I'm having trouble with apport now :/11:28
klattimerrunning it as user works but fails to collect log files, running it as root tells me gtk couldn't be initialised 11:29
klattimer:/11:29
tjaaltonit should ask for your password11:29
tjaaltonto collect the logs11:29
klattimertjaalton: that doesn't happen :/11:30
tjaaltonwait, you're on 10.10?11:30
tjaaltonaka maverick11:31
klattimertjaalton: nope natty11:31
klattimerah, it appears gdm was causing problems 11:31
tjaaltonthen you don't have all the updates11:32
klattimerhmm11:32
tjaaltonsince xserver-xorg should be 1:7.6~3ubuntu411:32
klattimerI just updated this morning 11:32
tjaaltonfrom maverick?11:32
klattimertjaalton: no on natty11:32
klattimerthis is a natty install 11:32
klattimerright through 11:32
klattimerI'm working on indicator-datetime (contracted) 11:32
klattimerso I need to use natty 11:32
tjaaltonok so what does 'apt-cache policy xserver-xorg' show as the installed version?11:33
klattimerinstalled 7.5, candidate 7.611:33
tjaaltonsee11:33
klattimerhmm, maybe something was a little wrong during the update11:34
klattimerthanks tjaalton11:34
tjaaltonhow did you upgrade?11:34
klattimertjaalton: apt-get update, apt-get upgrade 11:35
klattimerbut it seems all of the xorg stuff was either half installed, or not installed 11:36
klattimermaybe something went wrong during the upgrade11:36
tjaaltonit refuses to remove packages11:36
tjaaltonso doesn't always work during the devel phase11:36
tjaaltonuse dist-upgrade or update-manager..11:38
klattimertjaalton: thanks ;)11:39
gordboyso why are x-swat/x-updates offering a beta nvidia driver (that doesn't work properly) when their mission statement says they offer *stable* upsteam ? it would be very nice if you offered 260.19 nvidia driver, instead13:50
tseliotgordboy: maybe because they're supposed to work with the new Xserver with the IgnoreABI option. I don't have a real answer for you though14:00
gordboytseliot: interesting. if that were the case, then why are the xorg-edgers (the bleeding cutting edge, break your system weekly ppa) offering the stable driver ? seems to me there is a bit of confusion ...14:03
tseliotgordboy: I wouldn't know, maybe 270.18 was really meant to be uploaded to xorg-edgers instead14:04
ricotzbjsnider, ^^^ ;)14:12
tjaaltonklattimer: so do things work better after the upgrade?14:20
tjaaltonklattimer: all that the logs reveal is that compiz had segfaulted three times14:21
tjaaltonthe x log was fine14:21
tjaaltonhmm, I take that back, the xserver had crashed as well14:27
bjsnidergordboy, what distros is this?14:32
gordboybjsnider: lucid14:35
gordboyi know maverick had trouble with 260 and the NVRM: os_raise_smp_barrier(), invalid context! thing. but lucid doesn't14:35
bjsniderwell, i uploaded the 270 blob because nvidia doesn't generally break things retroactive to previous x-servers. it works fine here on maverick14:36
klattimertjaalton: yeah thanks14:37
klattimeryou can close the bug14:37
tjaaltonklattimer: cool14:38
gordboybjsnider: ok. well just so you know, the 270 on lucid gives freezes and full log files. the 260 is fine. tested on 8500gt, 9500gt and ge21014:40
bjsnideralright, well if i were you i'd go back to using the standard lucid driver14:40
gordboybjsnider: the problem with that, is that flash 10.2 crashes immediately with 195, and libvdpau1. i know this is an inexact science ...14:42
bjsnideryeah but thi is above our pay grade14:42
bjsniderwe can't fix closed-source flash and a binary graphics driver14:43
bjsniderat some point we have to throw up our hands and bow to the absurd14:43
gordboypoint taken. but the 260 works well with lucid, and is the current stable offering. just saying14:43
jcristauthen use that?14:43
gordboythe only place to get it is from xorg-edgers :)14:44
jcristauor nvidia.14:45
bjsnidergordboy, it works well for _you_ in lucid. i could probably scare up a dozen bug reports from others who say it causes problems for them14:46
gordboyi'll trade my 3 definites for your dozen maybes, any day :)14:48
bjsnidertseliot, i submitted a bug about some packaging issues with nvidia-current, as you requested15:15
bjsnider70460715:16
tseliotbug 70460715:16
ubot4tseliot: Bug 704607 on http://launchpad.net/bugs/704607 is private15:16
bjsniderwhat?15:16
tseliotbjsnider: thanks for reporting. I'm working on nvidia today so I should fix it now in my git branch15:17
bjsnidervery trivial changes. easy to implement15:18
bjsnidertseliot, i also created a bug about moving the nvidia preinstall script to jockey-common from nvidia-common. can you tell pitti about it? i don't think he was alerted it's bug 70459715:19
ubot4bjsnider: Bug 704597 on http://launchpad.net/bugs/704597 is private15:19
tseliotbjsnider: oh, right the hook for the installer. Thanks for reminding me about it. I'll talk to pitti15:21
bjsnidergood, because i don't know what channels he's in15:22
thesheff17RAOF: ping.....I finally believe I got the backtrace  from the crash through gdb. http://pastebin.com/Hc19Q0fg not sure if that helps at all but let me know if you have any ideas to fix it.15:43
jcristauwell, fglrx.15:45
thesheff17jcristau: yea I know...I'm really looking for a 4 port video card PCI-express that supports up to 4 monitors without breaking the bank. I got this ATI FirePro 2450 512MB PCI-e x16 Graphics Video Card for about 150 off ebay :-/  15:50
kklimondahey, does nouveau in natty support power managment?17:13
Sarvattkklimonda: http://nouveau.freedesktop.org/wiki/PowerManagement17:17
kklimondaSarvatt: btw, I've had a problem with Firefox causing X to use 100% of cpu (pretty much halting everything) when I browsed google images, or other pages with a lot of images using nouveau. Any idea what would cause it? That was the reason I went back to nvidia17:20
kklimondaso, from the matrix is seems that there is some power managment for nv50, as long as I don't have a fan connected to it? :)17:23
Sarvattit could be a lot of things :( have you been able to narrow it down any? only happen with compiz or something like that?17:25
Sarvattkklimonda: no fan speed adjustment control which isn't a big deal since your laptop doesn't have a separate fan17:25
kklimondaSarvatt: no, but it did happen every time.17:25
kklimondaSarvatt: I think I didn't run compiz17:25
kklimondaI could check it - it should be simple test :)17:26
ricotzSarvatt, hi :)17:26
Sarvatthiya! whats up ricotz?17:27
ricotzi wanted to play a bit with wayland17:27
ricotzSarvatt, do you think pixman 0.21.4 would be considered for a sync?17:27
Sarvattoh yeah we need a cairo-gl in xorg-edgers17:27
ricotzSarvatt, is it? the current cairo package builds its gl extension?17:28
Sarvattricotz: really I've been torn on that, I dont think 0.21.x is a good idea for natty since 0.22 wont be out and its the development branch but 0.21 already went into unstable.. was hoping to put the stable 0.20.x in unstable once squeeze released17:28
jcristauyeah a development snapshot in natty doesn't seem like a good plan17:29
ricotzSarvatt, there was a 0.20 release in exp once, perhaps this one can be pulled17:29
Sarvattit didn't build on our toolchain until that -O2 bug got fixed in our GCC a few weeks ago so we couldn't sync it then17:30
ricotzok, but isnt this package saved somewhere?17:31
Sarvattricotz: also waiting for notify-osd to get updated to not be screwed up with the newer pixman, I pinged MacSlow about it17:31
Sarvattricotz: well it's already up to 0.20.2 upstream and that never had a debian release17:32
Sarvatthttps://bugs.launchpad.net/ubuntu/+source/notify-osd/+bug/65492117:32
ricotzso making an ubuntu only version of it, is no option?17:32
ubot4Launchpad bug 654921 in notify-osd (Ubuntu) "Black border in the notifications when effects are turned off (affects: 10) (dups: 1) (heat: 68)" [Undecided,Confirmed]17:32
Sarvattricotz: thats pretty much our only option besides staying at 0.18.4 and losing out on all the arm speedups :D17:34
ricotzso should be worth it ;)17:35
ricotzSarvatt, i will do it17:39
Sarvattricotz: awesome, thank you for doing that!17:39
jcristaui'm wondering if i should get pixman 0.20 into testing17:41
ricotzSarvatt, http://people.ubuntu.com/~ricotz/pixman/17:54
ricotzjcristau, are you going to do an upload to testing?18:02
jcristaui don't know yet.18:03
ricotznext debian release can probably wait for 0.22 18:04
Sarvattkklimonda: firefox directly from mozilla might be worth testing, they bundle their own pixman in that one and nouveau+pixman 0.18.x have had problems specifically with google image search in the past18:16
kklimondaSarvatt: ok, I will test it out tomorrow18:17
Sarvatthttps://bugs.launchpad.net/ubuntu/+source/pixman/+bug/608613 i'm looking at you18:18
ubot4Launchpad bug 608613 in pixman (Ubuntu) "Nouveau tiling problem with firefox (affects: 2) (heat: 17)" [Undecided,Fix released]18:18
Sarvattthey use their own heavily patched cairo too18:19
ricotzSarvatt, do you if natty will ship the needed firmware for nouveau 3d support of nvc0?18:32
Sarvattricotz: very much doubt it18:32
ricotzSarvatt, hmm :(18:32
SarvattI dont even know how to get it at the moment, its actual code from the blob so I doubt we could ever ship it :(18:32
SarvattI believe*18:33
brycehthesheff17, what bug # is this for?18:54
Sarvattricotz: wonder if its in here - https://github.com/pathscale/envytools/18:56
ricotzSarvatt, yeah, i found that, there are the tools to get ;)19:00
thesheff17bryceh: I was working with RAOF to troubleshoot this problem...I haven't filed a bug yet.  It is completely random about every 48 hours with that ATI FirePro 2450 card.  Do you want me to file a bug?19:06
* penguin42 isn't too sure how to report a bug he's currently seeing - when windows appear they have junk in and then quickly get rendered a fraction of a second later; that's KDE on Radeon HD4350 open driver - any ideas for nailing it down?20:38
jcristaugoes away when you turn off compositing?20:40
penguin42let me try - there appears to be a 'suspend desktop effects' button20:41
penguin42jcristau: No, harder to spot though because stuff is faster, but you can still see it20:42
RAOFthesheff17: Hm.  That looks like it's probably a fglrx problem.  The backtrace suggests that it's occurring when something is changing the cursor image, though, so that might help you reproduce it.21:48
thesheff17RAOF: ah ok...I will have to see if I can pin point it.  I wonder if it is only happening when it changes over a certain app running...maybe chrome22:02
RAOFSomething like that would seem likely.22:03
RAOFI think the time has come to file a bug on fglrx with that backtrace attached.22:03
RAOFThe backtrace makes it look like it really is a fglrx bug, so it's not likely to be fixable by us, but I understand that the amd guys check launchpad every now and then.22:04
thesheff17RAOF: sure will do...it isn't horrible since it only happens every 48 hours or so. should I use these steps to report a bug? https://wiki.ubuntu.com/X/Reporting22:05
RAOFthesheff17: Yes, please.  For you, the main thing to attach will be that backtrace.22:07
thesheff17RAOF: ok sounds good.22:07
thesheff17RAOF: Thanks for the help.22:08
RAOFNo problem!22:12
thesheff17RAOF: not that this would have anything to do with the bug but so I have 2x2 monitors.  The top two monitors I cannot drag applications through the top two monitors.  but I can drag an app down to the bottom monitor and then over to the other one and then up to the other monitor. Does that make sense?22:16
RAOFThat does make sense, but it might be a fglrx quirk?  There's no X-y reason why you shouldn't be able to have a rectangular array of outputs.22:18
thesheff17RAOF: yea not a huge deal just another strange thing I ran into.22:18
afiestasHey, my X's (or maybe the kernel, I'm not sure) are crashing a looooot in Natty23:23
afiestashow can I provide more debug info?23:23
afiestasit happens randomly when I use Xv (via mplayer/vlc/whatever)23:24
penguin42afiestas: What graphics card?23:24
afiestaspenguin42: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (pr23:25
penguin42ok, so it starts up, just sometimes dies?23:25
afiestasit crashes when I use Xv (the X video accelerator api)23:26
afiestasvia Kamoso (Webcam app) Vlc mplayer etc23:26
penguin42how does it crash? Does the whole thing lock up? Does it log out? Does it draw aliens?23:26
afiestasit stop refreshing, and after a few seconds everything turns black23:27
afiestasI can't switch to any tty, but if for example I'm watching a film I still can hear the sound23:27
penguin42can you ping it?23:28
afiestasmm no idea, but I can try the next time23:28
afiestasalso I can leave sshd open and get info if needed23:28
penguin42yeh, so look for any errors at the bottom of dmesg, and also look for the errors at the end of /var/log/Xorg.0.log23:29
afiestasookiz, I'm going to do some patche for Kamoso now (webcam application) so I will reproduce it hopefully23:30
RAOFafiestas: Hm.  Does the apport hook trigger?23:33
afiestasRAOF: mm as far as I can tell, nope23:34
afiestascan I check it someway?23:34
RAOFYou could ssh in and grab the error state.23:35
afiestaswhat is error state exactly? what I have to grab?23:36
RAOFThe output of intel_gpu_dump and /sys/kernel/debug/dri/0/i915_error_state are good.23:38
RAOFIf it's Xv related I could also throw up a testing package for you; there's a possible problem that may be what you're hitting.23:39
afiestasokz23:40
afiestaswell I'm quite sure that it is Xv related23:40
afiestasit crashes only when I'm doing some video stuff23:40
brycehafiestas, grab `dmesg` as well23:42
afiestasookz23:42
RAOFAha!  We have a candidate commit for the i965 gpu hangs.23:47
brycehRAOF, if you mean ae8877e3, I'm already packaging it :-)23:49
RAOFHm.  I means 23f9b123:50
RAOFs/means/meant23:50
brycehoh, then we have fixes for two gpu hangs, lovely23:52
brycehah wait, same patch23:52
brycehgah I'm on crack23:53
brycehRAOF, yes that's the commit I meant.  I'm also pulling in ae8877e323:54
RAOF:)23:54
brycehRAOF, why do you put up with me23:54
afiestasue ue, package package! xD23:55
afiestasright now work on Video stuff in Natty is deadly xd23:56
brycehRAOF, btw what are you working on at the moment?  Would you have time to look at recent commits to mesa?  jcristau said there were some cherrypicks worth including in ubuntu, but looks like I'm not going to get to them today23:56
brycehRAOF, also we're a couple revs behind debian with mesa, although mostly looks like fixes for tangential stuff; still, might be nice to pull in anyway so we're up to date23:57
RAOFI'm currently looking at mesa; specifically unbreaking Unity on nv5x+.  I'll check out the cherry-picks.23:57
brycehbtw, I uploaded a fix to mesa earlier this morning23:58
RAOFThanks.  I'll pull that in.23:58
RAOFDid jcristau happen to mention what cherrypicks were worthwhile, or just following the 7.10 branch?23:59

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