IRC logs for #automotive for Wednesday, 2015-09-09

*** bbranch has quit IRC00:10
*** fernandod has quit IRC01:42
*** fernandod has joined #automotive01:42
*** Luming has quit IRC02:26
*** Luming has joined #automotive02:26
*** fernandod has quit IRC03:42
*** fernandod has joined #automotive03:42
*** Luming has quit IRC04:22
*** AlisonChaiken has joined #automotive05:54
*** mpaolino has joined #automotive06:31
*** jacobo has joined #automotive07:32
*** Luming has joined #automotive07:32
*** CTtpollard has joined #automotive07:46
*** toscalix__ has joined #automotive08:04
*** apinheiro has joined #automotive09:00
*** wschaller has joined #automotive09:06
*** wschaller has quit IRC09:21
*** wschaller has joined #automotive09:24
*** CTtpollard has quit IRC09:51
*** wschaller has quit IRC09:56
*** wschaller has joined #automotive09:59
*** CTtpollard has joined #automotive10:37
CTtpollardhello10:43
rjekHello.10:46
*** mpaolino has quit IRC10:47
*** apinheiro has quit IRC11:34
*** wschaller has quit IRC11:59
*** mpaolino has joined #automotive12:10
*** wschaller has joined #automotive12:11
*** jacobo has quit IRC12:35
CTtpollardhas anyone managed to get wayland_debug to work in the weston.service?12:48
*** fredcadete has joined #automotive13:29
CTtpollardhello fredcadete13:31
fredcadetehello13:31
fredcadetegood afternoon all13:31
CTtpollardI'm currently investigating https://bugs.genivi.org/show_bug.cgi?id=349#c2213:32
CTtpollardI'm trying to get weston.service to log to journalctl with the wayland debug variable but have not been successful so far13:33
fredcadeteno success with the log or with setting the variable?13:34
fredcadeteweston prints some stuff by default, you should be able to see that in journalctl even without enabling WAYLAND_DEBUG13:34
CTtpollardyeh, its last print is for the unix log in from root13:35
CTtpollardI've tried adding Environment=WAYLAND_DEBUG=1  to the systemd unit for weston under service13:35
fredcadetethat looks like the way to do it13:36
fredcadeteare you running journalctl as user or root?13:37
CTtpollardI thought I would have to sent stderr to a log file, but apparently jorunalctl syslog should handle it13:37
CTtpollardroot13:37
fredcadetewaaaaait13:38
fredcadetelet me have a look here13:38
fredcadetewhere are the GDP sources these days?13:39
fredcadeteI am having a vague memory of a vague suspicion, that when you use weston-launch it cleans the environment before exec'ing the child process (including WAYLAND_DEBUG=1)13:40
fredcadeteI don't remember if the GDP is using weston-launch or not13:40
fredcadetebut the easiest way to confirm it is to cat /proc/`pidof weston`/env13:40
fredcadetelet me know how it goes13:41
CTtpollardmeta-genivi-demo specifically is here http://git.projects.genivi.org/?p=meta-genivi-demo.git;a=summary13:41
fredcadeteyep, weston-launch13:41
fredcadetethere you go13:42
fredcadetehttp://cgit.freedesktop.org/wayland/weston/tree/src/weston-launch.c#n58113:42
fredcadeteit does not propagate WAYLAND_DEBUG13:42
CTtpollardgaah! thanks for spotting that13:43
fredcadeteI have it in my backlog to propose it upstream13:43
fredcadetebut my backlog is /dev/null13:43
fredcadetesorry about that13:43
CTtpollardI tried simply starting weston manually with the debug flag, but stopping the weston service starts a drm issue with the kernel13:45
CTtpollardor killing weston pid directly13:47
fredcadetea few months ago I was able to do systemctl restart weston (but we were not using weston-launch then)13:47
fredcadetewhen gdb'ing weston I did have those drm issues13:48
*** apinheiro has joined #automotive13:52
CTtpollardhmm, I'm not sure how to go about it if weston-launch resets the env :/14:06
*** jacobo has joined #automotive14:08
CTtpollardyeh restart kills it14:09
fredcadeteyou want to see weston traces and weston-launch is removing the control that enables those traces14:10
fredcadetethe least invasive change would be to patch weston-launch to propagate WAYLAND_DEBUG14:10
fredcadeteor patch wayland to always enable server-side logs14:11
fredcadetegoing into launching weston manually is a more invasive change; because you have to fix the drm issues when you stop, AND you will also have issues with opening the drm device being non-root14:12
fredcadeteoh oh wait14:12
fredcadeteif you really want to hack:14:12
fredcadete1. disable weston.service14:12
fredcadete2. go into a root shell14:13
fredcadete3. login xuser14:13
fredcadete4. su -c weston14:13
fredcadetemoral of the story: there are plenty of options and they are all somewhat ugly14:13
CTtpollardI'm just looking to see the logs so I can grasp the transmissions over the protocol, it's not that important to me that debug needs to work for the platform I don't think14:20
CTtpollardfredcadete: thanks for the suggestions :)14:20
CTtpollardfredcadete: do you think it would be sensible to add a debug flag to weston launch?14:29
fredcadeteyes14:33
*** AlisonChaiken has quit IRC14:39
*** jlrmagnus has joined #automotive14:46
*** wschaller has quit IRC14:55
*** wschaller has joined #automotive14:59
CTtpollardhmm, disabling the weston service doesn't seem to be stopping it launching after reboot15:23
CTtpollardgah15:32
fredcadetesystemctl list-dependencies?15:34
*** jlrmagnus has quit IRC15:39
*** toscalix__ is now known as toscalix15:41
CTtpollardfredcadete: sorry for the delay that just garbaged  my screen session15:47
CTtpollardI have it listing for weston15:50
fredcadetesystemctl list-dependencies is how I would try to guess why systemd launched weston15:50
CTtpollardwith the reverse flag maybe?15:56
fredcadeteyou can see it either way15:57
fredcadetejust see who is the parent of weston15:57
CTtpollardno parents I don't think15:58
fredcadeteif it's none then it was probably launched directly by someone15:58
fredcadetethis is starting to be a matrioska of bugs15:58
CTtpollardit's in the first stage of children after default.targe15:59
CTtpollard:D15:59
fredcadetethen it's launched by default.target15:59
fredcadeteis that on the user session or the system session?15:59
fredcadetedid you get any output when you disabled weston15:59
CTtpollardfredcadete: nope16:00
CTtpollardno prints16:00
fredcadeteif the disable/enable command does have effect, it prints a line saying it's removing a symlink or adding one16:01
CTtpollardok if I enable, and then disable it shows the add and removal of the symlink16:02
fredcadetethat's not the issue then16:03
*** jlrmagnus has joined #automotive16:10
CTtpollardfredcadete: well I've definitely got more things to look at for tomorrow on it, mainly working with weston launch, thanks :)16:19
* CTtpollard drives home16:19
*** wschaller has quit IRC17:03
*** apinheiro has quit IRC17:07
*** jacobo has quit IRC17:09
*** mpaolino has quit IRC17:51
*** toscalix has quit IRC17:53
*** alex_h has quit IRC18:41
*** alex_h has joined #automotive18:42
*** jlrmagnus has quit IRC20:41
*** jlrmagnus has joined #automotive20:58
*** jlrmagnus has quit IRC21:29
*** jlrmagnus has joined #automotive21:46
*** bbranch has joined #automotive22:26
*** bbranch has left #automotive22:26

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!