ogra_Fatal server error:11:07
ogra_[    62.328] Inconsistent depth 24 pixmap format.  Exiting11:07
ogra_that seems to happen on all arm test installs for beta2 ... did anything in xorg change recently that makes 24bit being considered bad ?11:08
tjaalton_doubt it11:19
=== tjaalton_ is now known as tjaalton
ogra_tjaalton, well, the point is that neither the kernel, nor the binary driver nor the xorg driver changed on the pandaboard ... its all just a copy from quantal (not even recompiled afaik)11:37
ogra_so the only part failing can be xorg itself11:37
seb128bah, hate you intel driver, compiz frozen again while doing a dnd between screens and I can't restart it11:54
seb128"intel_do_flush_locked failed: Input/output error"11:54
seb128Apr  4 13:53:06 localhost kernel: [16594.405422] [drm:i915_hangcheck_hung] *ERROR* Hangcheck timer elapsed... GPU hung11:55
seb128Apr  4 13:53:06 localhost kernel: [16594.405430] [drm] capturing error event; look for more information in /debug/dri/0/i915_error_state11:55
* seb128 reboots11:55
tjaaltonseb128: maybe try 3.9rc kernel?12:06
seb128tjaalton, it's not like I was having the issue every time, I could try those, but even if that works that doesn't me a fix for raring, just a workaround for me12:07
seb128or is there a specific fix in there that you are considering asking for backport to the raring kernel?12:07
tjaaltonseb128: no, but it would give some working baseline to bisect from12:10
tjaaltonif it's fixed12:10
seb128tjaalton, I will try to see if I can easily reproduce and then test the new kernel if it makes a difference12:19
seb128I wonder if that's the same gpu hang that dholbach was asking about yesterday12:19
seb128seems quite frequent, I hit an hang a day on raring on my i5 laptop :-(12:19
tjaaltonthat's too bad.. I only get the compiz hang after user-switch on my ivybridge desktop12:20
tjaaltonseb128: hey, you could test new mesa too :) https://launchpad.net/~ubuntu-x-swat/+archive/x-staging12:21
tjaaltonthere's a ffe for it12:21
tjaaltonso it's not completely crackful12:21
seb128do you have a dual screen setup? ;-)12:21
tjaaltonit makes a difference I bet12:21
seb128k, no luck then12:21
seb128well, the hang I just hit seems to happen when dnd between the 2 screens12:22
tjaalton3.9 got some pageflip rewrite so it might help12:22
seb128I hit it several times this way this week12:22
tjaaltonhmm, wonder if it was you or someone else who said the same12:22
tjaaltonisn't the dpi hardcoded to 96 somewhere?12:25
seb128tjaalton, in GNOME?12:28
tjaaltoni think12:28
seb128yes, it is12:28
tjaaltonbug 116437912:28
ubottubug 1164379 in xorg (Ubuntu) "DPI not detected" [Undecided,New] https://launchpad.net/bugs/116437912:28
tjaaltonwas about to close it as wontfix12:29
seb128tjaalton, https://git.gnome.org/browse/gnome-settings-daemon/tree/plugins/xsettings/gsd-xsettings-manager.c#n6712:30
=== kees_ is now known as kees
marvin24tjaalton, mlankhorst: xserver segfaults on some ARM in xf86platformBus.c21:17
marvin24the problem is that on my machine, the pci busid is NULL around line 22621:17
marvin24and strncmp doesn't like this21:18
marvin24this is on raring btw21:18
marvin24adding a check in the if expression fixes it for me21:19
tjaaltonmarvin24: bug 116198121:21
ubottubug 1161981 in xorg-server (Ubuntu) "Boot stalls after Ubuntu Raring desktop ARM (Panda board) install" [High,Confirmed] https://launchpad.net/bugs/116198121:21
ogra_hmm, i just had marked that as duplicate 21:22
ogra_i thought ...21:23
marvin24tjaalton: no, xserver crashes very early21:23
marvin24before loading any module21:24
ogra_tjaalton, thats not on a panda 21:24
ogra_marvin24, thats raring, right ?21:24
marvin24ogra_: yes21:25
ogra_intresting is that it works with a 3.1 kernel 21:25
ogra_but not with 3.821:25
ogra_i wonder what changed 21:26
marvin24ogra_: related to pci21:26
marvin24seems pci is detected but no busid21:26
marvin24it worked a few weeks ago21:26
marvin24I also wonder what changed21:27
marvin24(it worked with 3.8 until xserver was upgraded)21:27
tjaaltonit's the same bug21:27
ogra_well, then it is probably the same but manifests differently 21:28
marvin24tjaalton: at so different places in the start process?21:28
ogra_ah, to slow :)21:28
tjaaltonmarvin24: yes21:28
marvin24ah, just read comment #2121:29
marvin24so seems you are right21:29
ogra_awesome 21:30
mlankhorstmarvin24: oh on arm we could ignore the platform string, I tried to be careful and require pci, but I guess I messed up21:30
ogra_well, there are arm boards with PCI buses21:30
mlankhorstmy point is that it shouldn't have crashed to begin with21:30
ogra_i havent seen one that had a graphics card in there ever, but they exist21:31
mlankhorstthat's all21:31
ogra_ah, yeah21:31
mlankhorstand it should be a straightforward fix but im asleep21:31
mlankhorstxorg-server only seems to care about busid for pci devices21:31
mlankhorsti think21:32
marvin24well, busid should be checked against NULL in any case21:32
mlankhorstoh it's loading modesetting too?21:33
marvin24well, kind of21:33
marvin24which is modesetting + some stuff21:33
mlankhorstI mean from the bug report21:33
mlankhorstworst case, I guess it has to determine another way to check if those devices are identical21:34
* marvin24 is also tired 21:35
marvin24I'm off21:35

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