/srv/irclogs.ubuntu.com/2010/01/19/#ubuntu-x.txt

Sarvattseb128: fatalerror is segfaulting because of patch 100 in xserver, can see it by just running Xorg from gnome-terminal and seeing the trash after ddxSigGiveUp. Disabling patch 100 fixes the segfaults and gdm handles the fatalerror fine by stopping and starting a new x but then we lose apport, I haven't been able to fix up patch 100 yet but it's really over my head anyway, not sure if we want to drop the xserver apport hooks or what. The error s01:28
Sarvatteems to be coming from the initctl call in the session scripts not being in the exported path anymore (it's in /sbin and adding the path/hardcoding the call to /sbin/initctl fixes it) in the newer GDM's, can fix it in gdm but something does need to be done with that script making it segfault01:28
SarvattI rebuilt xserver without patch 100 and it works fine, also changing the initctl calls to /sbin/initctl and adding /sbin to the paths in the 2 gdm scripts fixes it01:29
Sarvatt/etc/gdm/Init/Default and /etc/gdm/PreSession/Default01:30
Sarvattjust dropping the patch will work but I dont know if you need that initctl to actually work and that needs the gdm change, can see that failing in /var/log/gdm/:0-slave.log when you try to use a guest session01:33
Sarvatterr in the second message I meant patch making it segfault not script01:39
=== jcristau_ is now known as jcristau
faganAny x people awake yet?08:04
tjaaltonalways08:05
RAOFIsh?08:05
faganIm upgrade testing hardy to lucid 08:05
faganand x is giving hassle 08:05
faganhere is one of my many xorg logs http://pastebin.ubuntu.com/358912/08:06
faganI tried doing a Xorg -configure but it segfaults 08:07
faganSo any clue whats causing my computer to start in low graphics mode08:09
tjaaltonwhat's your xorg.conf like?08:09
* fagan pastebins08:09
tjaaltonthough the real reason is that the nv driver doesn't support GF820008:10
* bryyce waves08:11
tjaaltonhi bryyce08:11
bryycehi tjaalton08:11
faganhttp://pastebin.ubuntu.com/358913/08:12
faganHmmm 08:12
mvofagan: did you had the binary nvidia driver in hardy? and on upgrade you got nv that not supports your card?08:12
tjaaltonmvo: not according to the xorg.conf08:12
faganmvo: nope fresh install of hardy08:12
tjaaltonfagan: so you always had vesa08:12
mvooh, so it went vesa -> nv08:13
* mvo nods08:13
faganThats weird08:13
faganWas vesa the default in hardy?08:14
tjaaltonif you had a logfile from hardy..08:14
tjaaltonno, nv has never supported 820008:14
faganThe binary driver supports the 820008:14
tjaaltonthe autoconfig probably didn't work the same way08:14
tjaaltonbut nv doesn't08:14
faganAh08:15
tjaaltonnouveau likely does08:15
tjaaltonbut, meh..08:15
faganSo its just that -nv doesnt support my driver 08:15
tjaaltoncard, yes08:15
faganOh so then its fine then 08:16
* fagan always installs the binary anyway08:16
tjaaltonok so in that case the upgrader removed nvidia from the config?08:16
faganYep you have my logs 08:16
faganIt seems so08:17
tjaaltonprobably would be the best to move the config away, then you'd get vesa straight away and not via a crash08:17
tjaaltonnothing valuable in that config anyway08:18
fagantjaalton: will do 08:18
tjaalton..that would be used08:18
tjaaltonmvo: ^^08:18
tjaaltonI don't know where that should be done though08:18
mvotjaalton: thanks, it should only comment out "nvidia" if it find that and no nvidia driver binary08:21
mvoand the xorg.conf in the pastebin has no "nv" line 08:22
mvoor will xorg still behave differently if there is no xorg.conf or a xorg.conf with no explicit driver in it?08:22
tjaaltonmvo: it'll use the default driver, but if there's no xorg.conf it'll build a list of fallback drivers to try08:25
tjaaltonwith the default one on top08:25
mvonice08:26
faganI have a bug report of the crash that I got earlier https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/50951908:47
ubottuError: This bug is private08:47
faganlol I love that the default is private always 08:48
faganIts public now08:48
seb128_fagan, could be bug #50651008:50
ubottuLaunchpad bug 506510 in xorg-server "Xorg crashed with SIGSEGV in FatalError() when using guest session" [High,Fix released] https://launchpad.net/bugs/50651008:50
seb128_did you try to open a new session or guest session?08:50
faganNew session08:50
seb128_k, likely the same issue then08:51
faganseb128_: Its from an upgrade from hardy to lucid08:51
bryyceshould be fixed with xorg-server (2:1.7.3.902-1ubuntu9)08:51
bryyceupgrade to my latest version and retest08:51
faganCool08:52
seb128_bryyce, I will when it's published08:52
seb128_bryyce, do you plan to bring back apport later? or do you want to catch crashes by some other way?08:52
faganwin 508:54
bryyceseb128_, I don't know; I already sank a lot more time into this patch than I'd like.  signal handling code is tough to debug08:56
bryycebut it's silly for most of our crash bug reports to be due to the patch that enables crash bug reports08:56
seb128_right08:56
seb128_let's see if that uncover an another crash or if that purely due to that one08:57
faganYay the update fixed the segfault 09:20
faganOne more question is the nvidia binary driver still braking x?09:20
faganBug #494166 ?09:27
ubottuLaunchpad bug 494166 in nvidia-graphics-drivers-180 "[lucid] nvidia-glx can't work with new xorg 7.5" [Medium,In progress] https://launchpad.net/bugs/49416609:27
bryycefagan, thanks for testing; please close any bugs you filed that are about this issue09:48
tseliotfagan: I guess not09:55
bryyceheya tseliot09:56
tseliothi bryyce09:56
LLStarksis there anything similar to this for linux? http://www.realtech-vr.com/glview/10:28
akaihola_The page https://wiki.ubuntu.com/X/Troubleshooting/Freeze contradicts itself. It claims that a backtrace is a non-symptom, but lists "[mi] EQ overflowing" and X freeze as a relevant problem. A backtrace is included with such log messages.11:38
akaihola_Alternatively, what is a good place besides IRC to discuss Wiki pages, if I don't want to join the mailing lists? File a bug in Launchpad?11:39
jcristau"what's a good place to discuss this if i don't want to use the right place"?11:39
tseliotakaihola_: I think you can talk to bryyce about this. He lives in the US, therefore you might want to send him an email if you can't chat to him directly on IRC (because of time zones)11:42
akaihola_jcristau: Ok. It's just a hassle to join and unsubscribe a list just to send a quick note.11:53
akaihola_tseliot: Thanks!11:53
jcristauakaihola_: yeah subscriber-only lists are a pain i agree11:55
akaihola_Just e-mailed bryyce. Thanks guys, moving on...12:00
=== yofel_ is now known as yofel
tjaaltonwhee, I got a X61 tablet for a spin to play with the wacom on it14:44
coz_tjaalton,  :)15:04
tjaaltonbah, it's a serial wacom inside the X61 tablet..17:05
tseliot:-/17:05
=== maxb_ is now known as maxb
seb128hum tseliot left17:47
seb128bryyce, the crash on user switch doesn't happen when booting without the splash17:47
seb128ie not using plymonth make it work17:47
Sarvattseb128: did you get any of my messages about the crash happening because the gdm session scripts are trying to call initctl -q emit desktop-session-start DISPLAY_MANAGER=gdm without /sbin in the path? does adding /sbin: to the PATH and changing it to /sbin/initctl -q emit desktop-session-start DISPLAY_MANAGER=gdm in /etc/gdm/PreSession/Default (or /etc/gdm/Init/Default) fix it for you like it does for 18:49
Sarvattme?18:49
seb128Sarvatt, to what path do you add those?18:50
Sarvatt/var/log/gdm/:0-slave.log shows the error when you try to use a guest session18:50
seb128what error?18:50
Sarvattin /etc/gdm/PreSession/Default18:50
Sarvattthat it cant find initctl18:51
seb128let me try that18:51
Sarvattthere were changes so it did it as a user and it cant find /sbin stuff anymore18:51
seb128chrisccoulson, ^18:52
seb128could interest you as well18:52
Sarvattugh spotty connection -- <Sarvatt> dropping patch 100 in xserver fixed it so FatalError wouldnt segfault which fixed it for non-plymouth but I guess with plymouth it relies on what its emitting19:01
seb128Sarvatt, doesn't seem to be enough there19:03
seb128I copied initctl in /bin19:03
=== stgraber_ is now known as stgraber
seb128which should be equivalent to your change19:03
Sarvatt /bin wouldn't be in the path either would it? try /usr/bin?19:04
Sarvattimagine /sbin would be there if /bin was19:05
Sarvattgotta run19:05
seb128good comment19:05
seb128I will try that19:05
seb128thanks19:05
seb128Sarvatt, no, that doesn't fix my issue19:10
chrisccoulsoni'm using /sbin/initctl, and it seems to have fixed my issue19:11
chrisccoulsoni've got Xorg on different VT's now19:11
seb128ok, I will upload a fixed gdm19:15
seb128I've some other changes to do anyway19:15
=== pgraner` is now known as pgraner
tjaaltonI don't understand the tablet.. the only device in the udev db related to wacom is under acpi subsystem, not pnp22:04
superm1is plymouth supposed to be working on non KMS hardware right now?22:59
sebnersuperm1: /here not working23:00
superm1sebner, okay same here then23:00
sebnerseb128: is gdm to be supposed restartable with upstart (like it worked with init)?23:00
superm1got a bug open already?23:01
seb128sebner, I don't know, I guess so, sudo restart gdm?23:01
sebnerseb128: sudo service restart gdm?23:02
sebnersuperm1: dunno23:02
seb128sebner, dunno, don't ask me about upstart23:02
seb128I've no clue about it23:02
sebnerheh kk23:03
cwillusebner, sudo restart gdm is the proper incantation23:37
sebnercwillu: cool. thx for the hint23:38

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