| *** slate has quit IRC | 03:16 | |
| *** slate has joined #sailfishos-porters | 03:18 | |
| *** slate has quit IRC | 03:38 | |
| *** slate has joined #sailfishos-porters | 03:54 | |
| *** slate has quit IRC | 04:00 | |
| *** slate has joined #sailfishos-porters | 04:27 | |
| *** slate has quit IRC | 04:36 | |
| *** slate has joined #sailfishos-porters | 04:41 | |
| *** Vgrade-j_ has joined #sailfishos-porters | 10:02 | |
| Vgrade-j_ | morning | 10:03 |
|---|---|---|
| situ | Morning | 10:07 |
| Vgrade-j_ | at bowling this morning doing some trsting with u7 | 10:09 |
| Vgrade-j_ | there is a definite timing issue with things running before /system is there | 10:10 |
| Vgrade-j_ | if i place some execstartpre=sleeps then for example language selector comes up. without sleep no language selectot | 10:12 |
| Vgrade-j_ | looking at systemd configs to try and undetstand | 10:13 |
| situ | Vgrade-j_: It comes up every time with sleep ? | 10:17 |
| netchip | TODAY IT'S INTERNATIONAL CAPSLOCK DAY :P | 10:18 |
| netchip | hehe | 10:18 |
| netchip | I wonder which idiot came up with that idea | 10:19 |
| Vgrade-j_ | situ tried | 10:20 |
| Vgrade-j_ | twice . with and without | 10:22 |
| situ | but systemctl start lipstick.service is not succesful everytime. | 10:24 |
| Vgrade-j_ | it seems to be for me | 10:32 |
| Vgrade-j_ | on our ks at least | 10:32 |
| situ | ok, I am using other image. | 10:34 |
| Vgrade-j_ | if it does not come sutonatically on boot i can akways systenctl start lipstick.service | 10:34 |
| Vgrade-j_ | ok. i dint have access to that here | 10:34 |
| situ | I moved /usr/bin/lipstick to /usr/bin/lipstick-original, and created /usr/bin/lipstick shell script which adds a 5 second delay before kicking /usr/bin/lipstick-original | 10:35 |
| netchip | hrm | 10:36 |
| netchip | I feel like coding, and expanding the code I was working on, MSameer | 10:36 |
| netchip | and sledges, gastboot boot <cm boot.img> didn't work for me | 10:36 |
| netchip | fastboot* | 10:36 |
| situ | Vgrade-j_: : We don't have droid-hal-hammerhead-kernel-0.0.6-201406220757.armv7hl package installed in automatically generated .ks, could that make any difference ? | 10:40 |
| Vgrade-j_ | possibly if there are kernel modules but i dont think we have any on hammerhead | 10:42 |
| Vgrade-j_ | lsmod returns nothing on our ks based imagr | 10:43 |
| situ | Vgrade-j_: oops... I booted with image generaed with our .ks and getprop doesn't list all the properties http://pastie.org/9334259 | 10:56 |
| situ | journal logs http://pastebin.com/41ZpA4VH | 10:57 |
| Vgrade-j_ | and i think thats because /sytem has not beem mounted in time | 10:59 |
| situ | Vgrade-j_: Is there anything in the logs to prove it ? | 11:00 |
| situ | Vgrade-j_: Output of mount and getprop just before lipstick is launched http://pastie.org/9334287 | 11:06 |
| situ | Vgrade-j_: I think you are right, I printed output of mount command just before executing 'droid-hal-init' and /system is not mounted. | 11:21 |
| Vgrade-j_ | right | 11:22 |
| situ | This invalidates following line in droid-hal-init.service : | 11:38 |
| situ | After=local-fs.target systemd-udev-settle.service | 11:38 |
| situ | In system.mout prbably replace 'WantedBy=local-fs.target' with 'RequiredBy=local-fs.target' | 11:39 |
| situ | seems to work after 2 reboots, let me try a few more | 11:42 |
| situ | it didn't work on my last try :( | 11:49 |
| situ | Now trying this : | 11:55 |
| situ | Requires=local-fs.target swap.target | 11:55 |
| situ | in sysinit.target | 11:55 |
| *** Vgrade-j_ has quit IRC | 11:58 | |
| *** slate has quit IRC | 12:09 | |
| *** slate has joined #sailfishos-porters | 12:17 | |
| vgrade_ | situ: back home now | 12:18 |
| * Stskeeps is home for a while, too | 12:18 | |
| situ | vgrade_: So we agree that mouting /system is inconsistent. | 12:18 |
| vgrade_ | situ: well I think its when it gets mouted is the important thing | 12:20 |
| vgrade_ | hi Stskeeps | 12:21 |
| situ | vgrade_: Yeah, that only :) | 12:21 |
| Stskeeps | moo | 12:21 |
| *** slate has quit IRC | 12:21 | |
| Stskeeps | what does getprop output in general say when it's broken | 12:22 |
| vgrade_ | the last pastie from situ 14:06 | 12:23 |
| Stskeeps | ysess | 12:24 |
| Stskeeps | that looks broken | 12:24 |
| situ | Stskeeps: Right, I actually tried printing output of getprop before launching 'droid-hal-init' and /system is not mounted sometimes. | 12:25 |
| situ | My /usr/bin/droid/droid-hal-startup.sh http://pastie.org/9334546 | 12:26 |
| Stskeeps | hmm | 12:26 |
| Stskeeps | well getprop won't work before droid-hal-init is up | 12:26 |
| Stskeeps | won't work properly, at least | 12:27 |
| Stskeeps | droid-hal-init needs /system to be up though, afaik | 12:27 |
| situ | Stskeeps: sorry, I meant output of mount. | 12:27 |
| *** phdeswer has quit IRC | 12:30 | |
| *** slate has joined #sailfishos-porters | 12:31 | |
| situ | Is it a bug in systemd ? | 12:44 |
| lbt | "Note that requirement dependencies do not influence the order in which services are started or stopped" Before=, After= .... "Configures ordering dependencies between units. If a unit foo.service contains a setting Before=bar.service and both units are being started, bar.service's start-up is delayed until foo.service is started up." | 13:05 |
| lbt | http://www.freedesktop.org/software/systemd/man/systemd.unit.html | 13:05 |
| lbt | vgrade_: situ: so add an After=local-fs.target and that should be reliable | 13:06 |
| situ | lbt: for droid-hal-init.service ? | 13:06 |
| lbt | yes | 13:07 |
| lbt | hmm | 13:08 |
| lbt | it seems to have one | 13:08 |
| lbt | but oddly doesn not Require it | 13:09 |
| situ | After=local-fs.target systemd-udev-settle.service | 13:09 |
| situ | Yes, it's there. | 13:09 |
| lbt | I think I should have added some Requires= in there too | 13:10 |
| lbt | Requires=local-fs.target | 13:11 |
| situ | lbt: will check that | 13:13 |
| *** slate has quit IRC | 13:16 | |
| situ | lbt: I added that line and it didn't mount /system before starting droid-hal-init in third reboot | 13:23 |
| lbt | hmm | 13:24 |
| lbt | the WantedBy change you made | 13:24 |
| lbt | that only takes effect when a unit is installed | 13:24 |
| lbt | so it's possible that local-fs is failing to mount /system but it's not Required at that point | 13:25 |
| lbt | so local-fs succeeds | 13:25 |
| *** slate has joined #sailfishos-porters | 13:26 | |
| situ | but why would it fail to mount ? | 13:29 |
| *** slate has quit IRC | 13:30 | |
| lbt | you need to do a systemctl enable system.mount I think | 13:30 |
| lbt | erm | 13:31 |
| lbt | use reenable | 13:31 |
| lbt | systemctl reenable system.mount | 13:31 |
| lbt | "Reenable one or more unit files, as specified on the command line. This is a combination of disable and enable and is useful to reset the symlinks a unit is enabled with to the defaults configured in the "[Install]" section of the unit file." | 13:31 |
| *** slate has joined #sailfishos-porters | 13:32 | |
| situ | lbt: I had added that line in droid-hal-init.service | 13:32 |
| lbt | [28/06/2014 12:39] <situ> In system.mout prbably replace 'WantedBy=local-fs.target' with 'RequiredBy=local-fs.target' | 13:32 |
| lbt | ? | 13:32 |
| situ | lbt: Ohh right, I have done that before. | 13:32 |
| lbt | but did you do a reenable ? | 13:33 |
| situ | I didn't | 13:33 |
| lbt | the [Install] section is only read by enable/disable and it is used to create symlinks | 13:33 |
| situ | ok | 13:33 |
| lbt | systemd is very very logical but can be a bit confusing :) | 13:34 |
| lbt | eg Requires & Before/After are not tied together but you may casually expect them to be | 13:34 |
| *** slate has quit IRC | 13:38 | |
| situ | lbt: It worked for several boots, but got stuck twice at this point http://pastie.org/9334840 | 13:50 |
| situ | hmm | 13:51 |
| situ | Jul 06 16:45:06 Jolla mount[385]: mount: none is already mounted or /acct busy | 13:51 |
| lbt | sounds better | 13:59 |
| lbt | it feels to me like local-fs is not defined tightly enough | 14:00 |
| lbt | I wonder | 14:00 |
| lbt | logically local-fs could signal complete when all 'Required' fs are mounted and continue to try and mount "Wants" | 14:00 |
| lbt | since that's a softer dependency | 14:01 |
| lbt | I dunno if systemd actually does that | 14:01 |
| netchip | hrm | 14:02 |
| netchip | anyone knows CAN bus? | 14:02 |
| netchip | it's a communication bus | 14:02 |
| netchip | dunno how it works... | 14:02 |
| lbt | know of, yes - know details .. google :D | 14:04 |
| netchip | ah, found it: http://en.wikipedia.org/wiki/CAN_bus | 14:05 |
| netchip | my school has joined a 'karting' competition... Electrical motors for the win :P Wanted to look into reading the information the motor outputs | 14:06 |
| netchip | engines* | 14:07 |
| netchip | I think, not motors | 14:07 |
| netchip | or maybe it are motors | 14:07 |
| *** slate has joined #sailfishos-porters | 14:17 | |
| situ | lbt: But it just gets stuck at this error and nothing proceeds. | 15:30 |
| *** slate has quit IRC | 20:57 | |
| *** slate has joined #sailfishos-porters | 21:07 | |
| *** slate has quit IRC | 21:13 | |
| *** slate has joined #sailfishos-porters | 21:21 | |
| situ | lbt: I think all those 'WantedBy" should be replaced with "RequiredBy" in systemd files. | 21:21 |
| situ | I see lipstick crashing several times after this function call : | 21:29 |
| situ | clock_gettime(CLOCK_MONOTONIC, {223, 931506376}) = 0 | 21:29 |
| *** slate has quit IRC | 22:05 | |
| *** slate has joined #sailfishos-porters | 22:09 | |
| *** slate has quit IRC | 22:49 | |
| *** slate has joined #sailfishos-porters | 22:52 | |
| *** slate has quit IRC | 23:29 | |
| *** slate has joined #sailfishos-porters | 23:33 | |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!