Bernardo | hi | 05:50 |
---|---|---|
RAOF | Howdie | 06:03 |
RAOF | tseliot: Do you want to merge synaptics, or shall I? | 09:37 |
RAOF | Hm. r700+mesa 7.8.1 really doesn't like compiz blur. Unless you like smoothly shaded black/red/yellow/green cornered rectangles for windows, of course. | 09:38 |
tseliot | RAOF: I'm quite busy hacking on radeon's drm code (for different reasons) right now. I'd appreciate if you could do the merge for me. Of course if a patch doesn't apply I'm available to adapt it | 11:11 |
RAOF | tseliot: Cool. Thanks. | 11:12 |
Sarvatt | RAOF: I already merged synaptics locally, did you already do it or do you want me to push? | 12:52 |
RAOF | Sarvatt: I've done some preliminaries, but not fully. Go for a push. | 12:52 |
Sarvatt | ok will do in a minute, just updating the changelog | 13:01 |
Sarvatt | evdev can just be synced, we dont need that patch anymore | 13:10 |
jcristau | did xserver get updated in m? | 13:10 |
jcristau | many drivers have a build dep on 1.7.6.901 which wasn't in lucid | 13:10 |
Sarvatt | not yet, just preparing for when it is | 13:11 |
jcristau | k | 13:12 |
RAOF | Sarvatt: There's an evdev sync bug already. | 13:12 |
Sarvatt | and yeah thats caused no end of headaches in xorg-edgers, been having to use experimental and hoping it doesnt get updated :) | 13:12 |
Sarvatt | RAOF: why universe for this? [raof] Fork a pre-2.10 -intel driver and package it in universe. | 13:18 |
RAOF | Because I 'ain't supportin in. | 13:19 |
RAOF | s/in/it. | 13:19 |
Sarvatt | having it installed side by side with intel would be nice, so it automatically falls back to it with no KMS | 13:19 |
Sarvatt | ah | 13:19 |
Sarvatt | btw SUSE has already done this if you haven't seen | 13:19 |
RAOF | I have not, no. | 13:19 |
Sarvatt | darn lost my bookmark to it last crash | 13:20 |
RAOF | Eh, finding the SuSE packaging can't be _that_ hard :) | 13:20 |
Sarvatt | https://build.opensuse.org/package/show?package=xorg-x11-driver-video-intel-legacy&project=openSUSE:Factory | 13:21 |
Sarvatt | i've been poking at 2.7.1 for newer servers since that was the last one with EXA, 2.8.0 was easy but doesn't seem to be an improvement for anyone | 13:26 |
RAOF | That's probably a better idea; my understanding of the i845 & i855 problems is in the buffer manager, which would disappear without UXA/DRI2 | 13:29 |
=== ara_ is now known as ara | ||
hyperair | Sarvatt: weird. my Xorg.0.log says that page flipping was forcibly disabled, yet i'm seeing the same kinds of hangs that page flipping brought about for you =\ | 14:28 |
hyperair | image frozen on screen, and everything else stalled. | 14:28 |
* Sarvatt makes sure he actually forcibly disabled it.. | 14:28 | |
Sarvatt | wouldn't put it past me to have just updated the message and not actually disabled it :) | 14:29 |
hyperair | heh | 14:29 |
Sarvatt | same patch | 14:30 |
Sarvatt | just the message updated, should be working still.. | 14:31 |
Sarvatt | anything change in your setup? did you install the gallium package or anything? | 14:35 |
RAOF | Why does everyone else see the crazy bugs? I've been using xserver 1.8.1 and -intel 2.11, and *I* don't get these hangs :) | 14:35 |
Sarvatt | on your primary machine? | 14:36 |
Sarvatt | i think thats the big difference :) | 14:36 |
Sarvatt | i dont tend to do the things that really break things on other machines | 14:38 |
RAOF | On (one of) my primary machine(s) | 14:40 |
Sarvatt | ricotz: the gallium dri2 tfp problem fixed in mesa and its build in edgers already btw | 14:56 |
ricotz | Sarvatt, nice! so i can upgrade again ;-) | 14:56 |
Sarvatt | yep :) thinking about uploading compiz without the parts in the ubuntu patch forcing it to not start with swrast now that that has tfp support too | 14:59 |
Sarvatt | hyperair: you crashed twice in an hour earlier? or were you just switching drivers? | 15:13 |
=== Kangarooo1 is now known as Kangarooo | ||
Kangarooo | unbelievable. ssh -x user@ip rocks. i opened other comps google chrome with all its tabs in this comp | 15:28 |
Sarvatt | and then you load a page with flash by mistake? :) | 15:38 |
* Sarvatt is lusting over this sharp netwalker PC-Z1 | 15:39 | |
Sarvatt | eww, a whole bunch of poulsbo based tablet announcements today.. :( | 15:49 |
Sarvatt | which means people are going to want to run ubuntu on them and have it work. great :) | 15:50 |
=== \vish is now known as vish | ||
hyperair | Sarvatt: no, i went offline due to the rain, then it crashed. | 16:07 |
hyperair | Sarvatt: right now i'm trying to coax my btrfs balancing thing to actually complete, since it crashed twice in 24h, interrupting the balancing preocedure twice. | 16:08 |
hyperair | so i'm not starting X until it completes =p | 16:08 |
Kangarooo | Sarvatt: opening ssh -X in browser page in flash does something bad? | 16:17 |
Kangarooo | Sarvatt: im now on pc1 opened vnc pc2 and in there opened ssh -X to pc1 and google-chrome from it :) | 16:17 |
Sarvatt | its just really slow | 16:27 |
Sarvatt | you may want to -XC | 16:28 |
Sarvatt | does gnome-shell really need to depend on xserver-xephyr? does anyone actually use gnome-shell in xephyr? (hint: it doesnt work with glx 1.4..) | 16:36 |
jcristau | depend on Xephyr? what kind of silliness is that? | 16:39 |
Sarvatt | a looong time ago that was the recommended way of testing it | 16:41 |
Sarvatt | like a year before it was packaged | 16:42 |
Sarvatt | ok exaggeration, more like 6 months or so | 16:43 |
Sarvatt | ahhh Recommends: xserver-xephyr | 16:44 |
Kangarooo | if i have FF crashing Xorg when opened a lot gif tabs in FF then can i also replicate it using ssh -X or -XC ? -XC means faster yes? right now i have opened crashed pcs FF in last week debugged pc :) and its not crashin. if ill open otherwise will then it crash? | 16:45 |
Sarvatt | wonder where xserver 1.8.2 (or 1.8.1.901 even) is, has had quite a lot of important fixes in it | 16:57 |
Kangarooo | what is SysRq in Alt+SysRq+1 ? | 17:21 |
Sarvatt | the sysrq key :) | 17:25 |
Sarvatt | ripps: --target=i586-linux in your configure flags for mplayer on maverick? | 17:28 |
Sarvatt | maverick's i686 default now | 17:29 |
ripps | Sarvatt: --cpu='i686' | 17:29 |
ripps | been using it for a while | 17:29 |
Sarvatt | thats in ffmpeg options not mplayer? | 17:30 |
ripps | Sarvatt: yeah, --target=i586-linux is in my i386 arch rules | 17:31 |
Kangarooo | i was making xorg dbg and when in dbg i entered pid nr then all computer froze. mouse and keyb not moving. all screen stopped | 17:33 |
Kangarooo | i can thrue ssh do something | 17:34 |
Sarvatt | attach gdb over ssh and type continue when it happens | 17:36 |
Sarvatt | handle SIGUSR1 nostop (in gdb) | 17:37 |
Sarvatt | darn, I wish nouveau/xf86-video-nouveau was on the xorg-commit list, i always miss updates | 17:41 |
Kangarooo | ah sick of this comp not working. ill sell it. but before if u want i can make it clean and vnc acces to it and if u want u can make all crash and reporting | 17:48 |
Kangarooo | wiki.ubuntu.com/X/Backtracing and wiki.ubuntu.com/DebuggingXorg is the same info | 17:50 |
Kangarooo | (gdb) attach 764 Attaching to program: /usr/bin/Xorg, process 764 ptrace: Operation not permitted. | 17:53 |
Kangarooo | (gdb) cont The program is not being run. (maybe becouse i cant move mouse and keyb on it and screen is showing but not moving anything) | 17:54 |
Bernardo | hi | 18:09 |
Kangarooo | Sarvatt: when i do whats in https://wiki.ubuntu.com/X/Backtracing then my comp crashes. even if i do it thrue ssh. on command (gdb) attach (pidnr) it crashes and nothing moves anymore | 18:20 |
Sarvatt | type cont | 18:21 |
Kangarooo | Sarvatt: so i do again sudo gdb /usr/bin/Xorg 2>&1 | tee gdb-Xorg.txt and then dont do attach but just cont? | 18:37 |
Sarvatt | ahh I give up, nouveau needs newer xserver now too. guess I'll just build xserver from origin/ubuntu instead of using the old one and rebuild the world again in the PPA | 18:37 |
Sarvatt | Kangarooo: no do whats set in the ssh section further down | 18:37 |
Sarvatt | it really says to do that? | 18:37 |
Sarvatt | that'd never work unless you were handling signals in .gdbinit.. | 18:38 |
Kangarooo | ah yes. it was not so easy couse both computers are not in same room. i was with crashing pc vnc to working pc and in there ssh so when in ssh i do that attach that comp crashed. now im at working pc :) | 18:39 |
Sarvatt | you could just gdb -p $(pidof X) -batch -ex 'handle all nostop' -ex 'handle all pass' -ex 'handle 11 stop' -ex 'cont' -ex 'bt full' -ex 'cont' from a VT if you just wanted the backtrace from a segfault, not sure what you're doing | 18:40 |
Kangarooo | Sarvatt: im just following https://wiki.ubuntu.com/X/Backtracing and its the same as https://wiki.ubuntu.com/DebuggingXorg | 18:40 |
Sarvatt | i dont know why you're doing it is what I meant, sorry | 18:41 |
Sarvatt | i'm having a hard time wrapping my head around what you just said you were doing :) | 18:42 |
Kangarooo | ok so now i was able to do next step (gdb) cont but now since comp i want to inwestigate why FF makes it crash Xorg cant be backtraced since it crashed on attach (pid) | 18:43 |
Sarvatt | Kangarooo: are you on radeon or nouveau? | 18:43 |
Sarvatt | because if so i can tell you why its crashing | 18:43 |
Kangarooo | Xorg crashing computer has nvidia | 18:43 |
Kangarooo | geforce fx5500 and i couldnt find gdb for nvidia | 18:44 |
Sarvatt | using the binary nvidia drivers? | 18:44 |
Sarvatt | installed through hardware drivers? | 18:44 |
Sarvatt | or just the default stuff? | 18:44 |
Sarvatt | because if its the default stuff it's the same exa bug | 18:44 |
Kangarooo | system->hardware drivers | 18:45 |
Kangarooo | nvidia somthing airc 173 | 18:45 |
Sarvatt | which ones does that use? nvidia-96? have you filed a bug on launchpad? | 18:45 |
Sarvatt | ah | 18:45 |
Kangarooo | airc = as i recall corretly :D | 18:46 |
bjsnider | the 5500 should be using the nvidia-173 driver | 18:46 |
Kangarooo | i was able to choose somekind 96 and 173 and 173 was working faster | 18:47 |
Sarvatt | did you do anything like upgrade to a 2.6.34 kernel by any chance? | 18:47 |
Sarvatt | because nvidia-96/nvidia-173 dont work on those, just ruling things out | 18:47 |
Kangarooo | i have latest kernel there witch can be updated from default ubuntu universes | 18:48 |
Kangarooo | w8 ill thrue ssh check | 18:48 |
Sarvatt | alrighty | 18:48 |
Sarvatt | its ok you'd know if you installed it | 18:48 |
Kangarooo | 2.6.32-22-generic | 18:48 |
Sarvatt | have you filed a bug on launchpad about it so I can look through your logs? | 18:49 |
Kangarooo | i think i installed that what u named in this comp im at now. couse at what im now has i845 and at bug report i wrote that that solution with kernel .34 didnt work even starting pc. but the realy latest witch 4 days ago was rc7 that worked a little better (loading logo was better resolution) but crash still happened. | 18:50 |
Kangarooo | Sarvatt: w8 3 min.. ill chek ihave a lot bug :) | 18:50 |
Sarvatt | i've never even been to the nvida-graphics-drivers-173 bug list | 18:51 |
* Sarvatt hides | 18:51 | |
Kangarooo | bug 587710 this is bug i made ubuntu-bug and to add more info i was checking https://wiki.ubuntu.com/X/Debugging but since on (gdb) attach (pid) my system crashes i cant backtrace this reported bug. | 18:53 |
ubot4 | Launchpad bug 587710 in nvidia-graphics-drivers (Ubuntu) "if FF wit lot Gif & overloading CPU then X-boom to login (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/587710 | 18:53 |
Sarvatt | thanks, you have a backtrace there in one of the logs too :) | 18:57 |
Kangarooo | oh i dont know how to know :) | 18:58 |
Sarvatt | the xorglogold has it, i put it in the bug description | 18:58 |
Sarvatt | there's another problem, the -dbg packages for xserver-xorg-core that you need in lucid are useless | 18:59 |
Sarvatt | what sucks is I remember the nvidia 170 series having that same problem like 2 years ago and they fixed it in the 180 series | 19:01 |
Sarvatt | not sure if I should move that over to xorg-server or leave it there, the backtrace is all in X | 19:02 |
Kangarooo | ok so a new bug asking about nvidia 180 put to universe is needed? or it has some other bugs? | 19:03 |
Sarvatt | no you cant use it on that GPU, they dropped support for the older cards | 19:04 |
Kangarooo | so now this card is also one of restriced ? | 19:04 |
Sarvatt | does this only happen with compiz enabled? | 19:05 |
Kangarooo | when doing compiz check i got error that this card is restricted. but actually i dont want compiz and dont use it. just tryd that check. | 19:05 |
Kangarooo | ah but actually i tryd compiz installing on xubuntu 1 week ago and it worked. very fast | 19:05 |
Kangarooo | ouh wait. sorry no my card for comp witch that bug is was not compiz check showing as restriceted. compiz check showed ok. | 19:06 |
Kangarooo | so compiz check showed ok for all checks. | 19:07 |
Kangarooo | for other comp witch has i845 compiz check showd restricted card/driver | 19:08 |
Sarvatt | so its enabled then because it defaults to enabled | 19:08 |
Sarvatt | can you try disabling visual effects in system - preferences - appearance to see if you can crash it that way too? | 19:08 |
Kangarooo | Sarvatt: no compiz is removed now. | 19:08 |
Sarvatt | oh the nvidia machine? | 19:08 |
Sarvatt | err on | 19:08 |
Kangarooo | ye :) | 19:08 |
Kangarooo | i put compiz on nvidia 5500 machine and it workd and i dont need fx so i removed. and nvidia 173 is activated | 19:09 |
Sarvatt | oh you're on xubuntu ok | 19:10 |
Kangarooo | ye | 19:10 |
Sarvatt | using any kind of compositing? | 19:10 |
Sarvatt | xcompmgr or anything, no idea what xubuntu uses | 19:11 |
Kangarooo | searching google whats that. so theres Xfce | 19:12 |
Sarvatt | it'd be great if you could get a core dump i could poke around on :) | 19:13 |
Kangarooo | default xubuntu installation nothing then programms added. fx not added. w8 ill ask on #xubuntu | 19:13 |
Kangarooo | Sarvatt: ok thrue ssh ill get it where only? | 19:14 |
Sarvatt | wonder why http://cgit.freedesktop.org/xorg/xserver/commit/?id=4151a13c80f3afa43f88afcf19a7aeb16dace93a didn't make it to 1.7 branch | 19:18 |
Kangarooo | Sarvatt: compositing in xubuntu not by default. so i dont have it also. where is that core dump? ill cp and post to that LP bug | 19:19 |
Sarvatt | have you enabled apport? anything in /var/crash/? | 19:21 |
Kangarooo | Sarvatt: yes apport ive enabled on all comps. about that bug when FF with a lot gifs crashes Xorg apport didnt gave crash. thats why i made ubuntu-bug to post that bug and to get more info i wanted to do today that backtracing | 19:23 |
Kangarooo | now theres noting new in /var/cras | 19:23 |
Kangarooo | theres xfce4-systemload crash and update-apt-xapian crash . when i post crash i delete them so i dont get confused couse i dont post them immidiatly but later all. | 19:25 |
Kangarooo | thouse has nothing to do with nvidia and i got them when FF Xorg crash wanst happening | 19:26 |
Sarvatt | you have to manually enable apport for it to catch things now, but apport is only catching a small number of X crashes anyway | 19:29 |
Sarvatt | change the enabled=0 line to enabled=1 in /etc/default/apport then sudo service apport start and reproduce the crash, hopefully it shows up in /var/crash/ | 19:30 |
Kangarooo | Sarvatt: yes ive done that to all comps | 19:34 |
Kangarooo | no it doens show | 19:35 |
Kangarooo | Sarvatt: interesting news. im now back at crashing comp and now all is moving. GDB still active | 19:36 |
Kangarooo | i went to other comp couse this crashed. and im in pidgin in working comp true vnc from crashing comp | 19:37 |
Kangarooo | so ill now make it make backtace? | 19:38 |
Kangarooo | um reproduce bug while backtracing gdb is working thrue ssh ok? | 19:38 |
Kangarooo | and what about that core dump? where i can get it? | 19:39 |
Kangarooo | ah core dump is a crash file? | 19:40 |
Kangarooo | so apport is started and made to start on comp start enabled=1 | 19:40 |
Kangarooo | it was already 3 weeks ago | 19:40 |
Sarvatt | Kangarooo: I'm building an xserver package in a PPA for you that might fix it, but even if it doesn't fix it it will have a correct -dbg package that'll actually work unlike the one in the archives | 19:41 |
Sarvatt | https://edge.launchpad.net/~sarvatt/+archive/bugsbugsbugs | 19:41 |
Kangarooo | so manually also i have to make? from working comp thrue ssh on crashing comp? | 19:42 |
Sarvatt | make what? | 19:42 |
Sarvatt | oh that means apport isnt catching it, yeah | 19:42 |
Kangarooo | another apport ? | 19:42 |
Sarvatt | sorry was packaging it up and not reading it | 19:42 |
Sarvatt | no | 19:42 |
Kangarooo | yea isnt catching | 19:42 |
Sarvatt | nevermind about that, didnt know you already turned it on | 19:42 |
Sarvatt | yeah have to start X with -core | 19:43 |
Sarvatt | and set the ulimit right first | 19:43 |
Kangarooo | apport i made 3 weeks ago to be starting on comp start | 19:43 |
Kangarooo | w8. im now at comp thats crashing. i went to other comp couse this one stopped again working when put command attach (pid) now im back at it and keyb is working and gdb is working | 19:44 |
Kangarooo | so i can make that backtrace reproducing bug finally if needed | 19:45 |
Kangarooo | ill make backtrace now and then lets try next steps/options | 19:46 |
Kangarooo | ill still be in chanell since im in vnc and chanel is open in working pc | 19:47 |
Sarvatt | try this PPA when it finishes building, it might fix your problem - https://edge.launchpad.net/~sarvatt/+archive/bugsbugsbugs | 19:47 |
Sarvatt | okie, what i was saying is the backtrace in gdb will probably not be very useful because the -dbg packages for xserver-xorg-core in lucid don't work | 19:48 |
Sarvatt | the ones in the PPA will though so you'll be able to get a good backtrace if the patch I added doesn't fix it | 19:49 |
Sarvatt | be sure to get a bt full and not just a bt, you have the bt output already | 19:49 |
Sarvatt | now I see you have another crash in one of the other logs too - *** glibc detected *** /usr/bin/X: corrupted double-linked list: 0x09048e90 *** | 19:58 |
Sarvatt | http://cgit.freedesktop.org/xorg/xserver/commit/?id=4151a13c80f3afa43f88afcf19a7aeb16dace93a is the commit I added to lucid's xserver in that PPA, I think it might fix it | 19:59 |
Sarvatt | its done building | 20:00 |
Kangarooo | this comp was restarted while all finally was working :) | 20:04 |
Kangarooo | ill try again | 20:04 |
Sarvatt | Kangarooo: try with the PPA so even if it doesnt fix things the backtrace you get will actually be useful | 20:05 |
Kangarooo | ah Sarvatt i found out that when i write in (dgb) cont then the crashing comp unfreezes to continue Xorg backtracing | 20:24 |
Sarvatt | yeah | 20:24 |
Sarvatt | Kangarooo: run this then it wont do that anymore - | 20:24 |
Sarvatt | handle all nostop | 20:24 |
Sarvatt | handle all pass | 20:25 |
Sarvatt | handle 11 stop | 20:25 |
Sarvatt | then continue | 20:25 |
Sarvatt | and you wont have to run between rooms anymore until that crashes :) | 20:25 |
Kangarooo | even now? i did cont and now im wainting for crash to happen. or for future do thouse commands? | 20:27 |
Sarvatt | you wont have to cont anymore, it'll keep going until it crashes | 20:37 |
Sarvatt | if it freezes that means it crashed | 20:37 |
Sarvatt | then you can get your bt full | 20:37 |
Kangarooo | no now crash i want to reproduce is making all closed and login open. so no freeze will now happen | 20:44 |
Kangarooo | wow 100 tabs open 600mb ram used swap 0 still not crashing | 20:47 |
bjsnider | if this was windows, i could cause a crash in less than 5 clicks | 20:50 |
Sarvatt | is that with the PPA xserver Kangarooo? | 20:53 |
Kangarooo | no still the same way i was tryng to do today. ill do all backtraces. first this one ill complete | 20:54 |
Sarvatt | ah, ok | 20:54 |
Sarvatt | did you do anything different? you said you enabled compiz now? | 20:55 |
Kangarooo | now already 850mb ram used swap now only first 2 mb. no nothing different. no compiz not installed. this crash was able to reproduce without compiz. i installed compiz 4 days ago only for 10 min and purged it. | 20:56 |
Kangarooo | ill close all tabs and start again opening all gifs | 20:56 |
Kangarooo | Sarvatt: All again froze and in terminal from ssh Program received signal SIGSEGV, Segmentation fault. 0x08087041 in dixChangeGC () . should i cont again? and all other commands u wrote so no stopping happens yes? | 20:59 |
Sarvatt | nope cant continue,thats your crash | 20:59 |
Sarvatt | bt full now | 20:59 |
Kangarooo | no this crash is not the one i wanted to bacrtrace. crash i wanted to reproduce make login screen popup | 21:00 |
Sarvatt | Kangarooo: thats the same crash, it stopped there because you are in GDB | 21:00 |
Sarvatt | set logging file ~/xcrash.txt | 21:00 |
Kangarooo | ok so when ill go out of gdb then crashing comp will be in login promt? | 21:00 |
Sarvatt | set logging on | 21:01 |
Sarvatt | bt full | 21:01 |
Kangarooo | loggin is with | tee filename | 21:01 |
Kangarooo | ok i did bt full. now do i have to quit gdb somehow? | 21:02 |
Sarvatt | yep, just quit | 21:03 |
Kangarooo | and then that comp will respond? | 21:04 |
Kangarooo | quit with quit ? | 21:04 |
Kangarooo | (gdb) quit A debugging session is active. Inferior 1 [process 2562] will be detached. | 21:05 |
Kangarooo | Sarvatt: ? | 21:06 |
Sarvatt | yep | 21:10 |
Sarvatt | you're done, you have your backtrace.. | 21:10 |
Kangarooo | yeah login poped up. where can u see that? so ill upload backtrace to the same bug? | 21:16 |
Kangarooo1 | Sarvatt: so i make sudo add-apt-repository ppa:sarvatt/bugsbugsbugs/ubuntu ? and then install that package ? | 21:29 |
Sarvatt | yep! | 21:30 |
* hyperair is quite sure it doesn't need an extra /ubuntu at the back | 21:31 | |
Sarvatt | i dont use that so i assumed he just pasted it from the ppa page and said yeah lol | 21:35 |
Kangarooo | ok after realod i got 3 package updates. xserver-xorg-core-dgb xserver-common xserver-xorg-core . all install Sarvatt ? and then restart pc? and this crash will not happen? or then make test reproducing? | 21:40 |
Sarvatt | sudo apt-get update && sudo apt-get dist-upgrade, it'll grab everything | 21:40 |
Sarvatt | dont know what you have installed to tell you what needs to be upgraded | 21:41 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!