jonpackard | Hello. I'm having problems with X in a qemu emulated Ubuntu Jaunty Armel system. I can load gdm or run startx to get to a desktop, but once X loads, I cannot use the keyboard or mouse. Here's my Xorg.0.log http://ubuntu.pastebin.com/f5a153034 -- Could anyone please offer any suggestions? | 03:09 |
---|---|---|
tjaalton | just what we needed.. a poll about zapping | 07:06 |
Alexia_Death | tjaalton: it turns out that the buttons fix for tablets is not in rc1. It was cherrypiced a bit after tagging... | 08:38 |
Alexia_Death | tjaalton: https://bugs.freedesktop.org/show_bug.cgi?id=19282 the patch is here. | 08:41 |
ubottu | Freedesktop bug 19282 in Input/Core "Events handled wrong when master button map differs from the originating device..." [Normal,Resolved: fixed] | 08:41 |
tjaalton | Alexia_Death: is it here? http://wiki.x.org/wiki/Server16Branch | 09:03 |
tjaalton | seems to be | 09:04 |
tjaalton | "mi: ensure chained button mappings from SD -> MD (#19282)" | 09:04 |
Alexia_Death | yes | 09:04 |
Alexia_Death | the hash thingie matches nice. | 09:05 |
tjaalton | I'll add it later today | 09:06 |
tjaalton | and will check the list for others too | 09:06 |
Alexia_Death | I havent tested that pach myself tho... Great... then theres a little catch with the wacom driver and the hotplug should work:D | 09:06 |
tjaalton | a link to the patch and it'll be included as well | 09:07 |
tjaalton | in the driver.. | 09:07 |
tjaalton | afk, cleanup -> | 09:07 |
Alexia_Death | hmm.. that will be a bit tricky... theres a hack and something thats supposed to be a solution... the hack works, the solution does not. the hack however is not nice on anything else than 1.6 because it causes a memory leak. in 1.6 double free kills the server. | 09:08 |
tjaalton | Alexia_Death: ok, I'll wait for a bit then | 10:31 |
ogra | tjaalton, have you seen the request from jonpackard above ? i see the same here | 10:34 |
ogra | (log is identical to his at http://ubuntu.pastebin.com/f5a153034) | 10:35 |
ogra | hal and dbus are running fine | 10:35 |
tjaalton | ogra: not in detail.. will check | 10:37 |
ogra | (# | 10:37 |
ogra | II) Cannot locate a core pointer device. | 10:37 |
ogra | # | 10:37 |
ogra | (II) Cannot locate a core keyboard device. | 10:37 |
ogra | # | 10:37 |
ogra | (II) The server relies on HAL to provide the list of input devices. | 10:37 |
ogra | # | 10:37 |
ogra | If no devices become available, reconfigure HAL or disable AllowEmptyInput. | 10:37 |
ogra | thats the issue here i thikn | 10:37 |
tjaalton | no, the last line is more important | 10:37 |
ogra | [config/dbus] couldn't take over org.x.config: org.freedesktop.DBus.Error.AccessDenied (Connection ":1.6" is not allowed to own the service "org.x.config.display0" due to security policies in the configuration file) | 10:37 |
ogra | that one ? | 10:38 |
tjaalton | yes | 10:38 |
tjaalton | the first ones are just informational | 10:38 |
ogra | note that system is a debootstrapped ARM rootfs | 10:38 |
tjaalton | oh.. | 10:38 |
ogra | with xfce4, xorg and gdm installed on top | 10:38 |
tjaalton | maybe enabling dbus had something to do with it | 10:38 |
ogra | so might be a dependency missing | 10:38 |
tjaalton | can you build the xserver wihtout --enable-dbus? | 10:39 |
ogra | ah, thats builtin now ? | 10:39 |
tjaalton | sorry, --enable-config-dbus | 10:39 |
tjaalton | yes | 10:39 |
tjaalton | just delete the line from rules | 10:39 |
tjaalton | it was enabled for wacom hotplub | 10:40 |
ogra | i can, will take a while, the arm build machine i hav is a bit slower than x86 | 10:40 |
tjaalton | uh, -plug | 10:40 |
ogra | ah | 10:40 |
ogra | googling only found me gentoo probs :) so it might not be enabled in other distros | 10:40 |
tjaalton | Alexia_Death: you've used the dbus setup before.. normal hal hotplug still works? | 10:43 |
* ogra wonders if xfce4 starts up a session bus | 10:43 | |
Alexia_Death | tjaalton: yes. Its designed so HAL can do its thing, and just stylus works. then when the daemon runs, it removes the HAL added device and adds fully configured devices. | 10:45 |
tjaalton | actually, I have the same error | 10:47 |
tjaalton | but then hal kicks in and does it's business | 10:47 |
jcristau | is config/xorg-server.conf installed anywhere? | 10:47 |
tjaalton | good catch | 10:48 |
tjaalton | it isn't :) | 10:48 |
Alexia_Death | tjaalton: did you install it? I had to modify the package to install it for intrepid | 10:48 |
Alexia_Death | :) | 10:48 |
Alexia_Death | the 1.5 packages in my ppa do in an hackish nature... | 10:48 |
tjaalton | the destination is /etc/dbus-1/system.d? | 10:53 |
tjaalton | bah, have to go out -> | 11:07 |
ogra | so copying the .conf file in place made the error in the log go away ... but still neither mouse or kbd | 11:15 |
jcristau | ogra: it would have been unexpected if that made a difference :) | 11:17 |
ogra | well, it fixed the noise in the log :) | 11:17 |
jcristau | heh. i meant other than that | 11:18 |
ogra | yeah :) | 11:18 |
ogra | shuldnt there also be a .service file ? | 11:19 |
jcristau | lshal lists the devices correctly? | 11:19 |
ogra | no idea, i cant get to the console :) | 11:20 |
* ogra boots with the text kernel option | 11:20 | |
ogra | well, withough X running lshal doesnt list any x11 options | 11:22 |
ogra | nor any evdev driven devices | 11:22 |
jcristau | but they show up in /proc/bus/input/devices? | 11:24 |
ogra | there is an AT kbd and a ps/2 mouse | 11:24 |
ogra | argh | 11:25 |
ogra | ah, i thought there was no sysfs mounted, but its all there | 11:25 |
ogra | (note this is qemu ARM, but X used to work in former setups) | 11:26 |
ogra | well, s/X/input/ | 11:27 |
tjaalton | ogra: used to work with 8.10? | 12:12 |
tjaalton | or previous xserver? | 12:12 |
ogra | yep | 12:12 |
ogra | build is running ... | 12:12 |
ogra | lets see | 12:12 |
tjaalton | yep to which question?-) | 12:13 |
ogra | previous server | 12:20 |
ogra | we didnt have armel with 8.10 :) | 12:20 |
tjaalton | right | 12:20 |
jcristau | if hal doesn't have the input.x11_driver properties, that's not X's fault, though :) | 12:26 |
tjaalton | true.. and downgrading the server should confirm that | 12:31 |
tjaalton | bah, still no mesa | 16:12 |
jcristau | tjaalton: what if you use drm_i915_sarea_t instead of struct drm_i915_sarea? | 16:20 |
jcristau | (since the kernel seems to call it struct _drm_i915_sarea instead) | 16:21 |
jcristau | using the typedef should work afaict | 16:21 |
tjaalton | jcristau: where is that? | 16:24 |
tjaalton | I actually googled the error message and found out what it means | 16:25 |
jcristau | in mesa. intelWindowMoved() | 16:25 |
tjaalton | ah.. | 16:25 |
tjaalton | well, changing it to _drm_i915_sare worked | 16:28 |
tjaalton | _sarea | 16:28 |
jcristau | it'd break for people using the libdrm headers. hence the suggestion to use the typedef | 16:29 |
tjaalton | well, it didn't work :) | 16:29 |
jcristau | s/struct drm_i915_sarea/drm_i915_sarea_t/ should have worked just the same, unless i'm missing something | 16:30 |
tjaalton | aah | 16:30 |
tjaalton | I left struct there | 16:30 |
tjaalton | bingo | 16:30 |
tjaalton | (my c sucks) | 16:31 |
tjaalton | ok, will build just to be sure it doesn't explode somewhere else | 16:43 |
tjaalton | I mean if there are other bugs left | 16:43 |
tjaalton | yep | 16:50 |
tjaalton | intel_context.c does the same | 16:54 |
tjaalton | and _screen.c | 16:55 |
tjaalton | .. and _screen.h, bah | 17:37 |
* jcristau hands git grep to tjaalton | 17:39 | |
tjaalton | <blush> | 17:45 |
tjaalton | man, that's fast | 17:45 |
tjaalton | ok, so i915_drm.h from 2.6.28 is missing *bo_handle | 19:18 |
tjaalton | sigh | 19:18 |
tjaalton | there is a commit for it | 19:18 |
tjaalton | so, I got mesa to build, but it'll need a kernel update until it's ready to be published | 21:45 |
tjaalton | phew, compiz works again | 21:55 |
tjaalton | and x11perf -aa10text is now up to ~200k chars | 21:55 |
tjaalton | with EXA, not UXA | 21:55 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!