| yifei1 | Wizzup: the easiest way to get an interactive shell on that ppp port image seems to be using the keyboard case to switch to a tty and login there. | 10:15 |
|---|---|---|
| Wizzup | Did you manage? | 10:15 |
| yifei1 | yes, using the actual serial port seems to reset the cpu more often than it should. | 10:16 |
| yifei1 | I'm using Ctrl + Fn + Alt + 3 | 10:17 |
| yifei1 | how do I disable auto suspend from commandline? | 10:21 |
| yifei1 | from the tty you can config a ip address on the usb0 interface and then ssh into it. | 10:30 |
| Wizzup | 10:22 < yifei1> how do I disable auto suspend from commandline? | 10:32 |
| Wizzup | suspend of what | 10:32 |
| Wizzup | I think this might be mce turning off the screen | 10:32 |
| yifei1 | The screen turns off automatically after a timeout | 10:33 |
| Wizzup | that is mce | 10:33 |
| Wizzup | you can try to stop it, but yeah it's annoying | 10:33 |
| yifei1 | mce? | 10:33 |
| Wizzup | it is maemo daemon | 10:33 |
| yifei1 | how do I restart the desktop? I crashed xorg... | 10:34 |
| Wizzup | dsmetool -k '/usr/bin/mce --force-syslog' | 10:34 |
| Wizzup | Xorg crash should reboot the device normally | 10:34 |
| Wizzup | are you sure Xorg is dead? | 10:34 |
| Wizzup | I would try to just disable the wrong output in xrandr, and then pkill hildon-desktop and see if it starts to work | 10:34 |
| yifei1 | Yes, I have a core dump of it. | 10:36 |
| yifei1 | nvm I will restart the device. | 10:37 |
| Wizzup | (sorry on the phone, back in a bit) | 10:44 |
| yifei1 | take your time ;) | 10:47 |
| yifei1 | It's working now with correct resolution. the magic seems to be `xrandr --output DSI-1 --auto` | 10:55 |
| yifei1 | to reproduce: - boot the device with keyboard attached, wait until screen turns off | 10:57 |
| yifei1 | - press power key to turn on screen, switch to a tty with Ctrl + Fn + Alt + 3 and configure an ip for usb0 for ssh | 10:58 |
| yifei1 | - ssh login as user, xrandr --output DSI-1 --auto && pkill hildon-desktop | 10:59 |
| yifei1 | There is only one output listed by xrandr. | 10:59 |
| yifei1 | Interestingly the random SoC resets don't seem to be correlated to serial usage, it just happens. | 11:03 |
| Wizzup | I would suggest to put this fix in an xsession script for now, before hildon-desktop is started | 11:13 |
| Wizzup | I will try this in a bit, what happens on rotation? | 11:13 |
| yifei1 | ah yes I forgot that | 11:15 |
| yifei1 | xrandr --output DSI-1 --rotate right | 11:15 |
| yifei1 | I'm now looking into the random reset problem, the uptime for now is too short to do anything... | 11:15 |
| Wizzup | are you sure it's like a kernel reset, or something else maybe? | 11:29 |
| yifei1 | Wizzup: actually, I'm not sure. | 14:03 |
| yifei1 | Wizzup: how is hildon-desktop started on maemo leste? I'm trying to make sure my fixup runs before it. | 14:04 |
| Wizzup | dpkg -L hildon-desktop - but I think it is xsession | 14:21 |
| Wizzup | so /etc/init.d/Xsession.d or /etc/init.d/Xsession.post/ or so | 14:21 |
| yifei1 | got it. | 14:47 |
| yifei1 | How much of the current maemo leste is from Nokia's code? I still see warnings on the email and app manager clients about Nokia (e.g. 'this app is not provided by nokia' 'your info will be send to nokia for processing...'), should those reference be removed? | 14:49 |
| uvos__ | yes these referances should be removed | 14:54 |
| uvos__ | the portion of leste that is nokia code is tiny | 14:54 |
| Wizzup | I think those are probably just in the translations | 14:55 |
| uvos__ | the vast vast majority is the shared linux ecosystem | 14:55 |
| uvos__ | Wizzup: jup | 14:55 |
| uvos__ | still the need to go | 14:55 |
| uvos__ | the warning about the package not being signed should be disabled entirely | 14:55 |
| uvos__ | since we dont have any infrastructure to sign anything like that | 14:55 |
| Wizzup | is this in the app mgr? | 14:58 |
| uvos__ | yeah | 14:58 |
| yifei1 | yes, try install anything | 14:58 |
| yifei1 | I think it just uses apt underneath, right? | 14:58 |
| Wizzup | yes | 14:59 |
| uvos__ | yes except it deals in some custom nokia extensions too | 14:59 |
| yifei1 | so the packages are still signed like regular debian packages. | 14:59 |
| uvos__ | no | 14:59 |
| yifei1 | oh.. so packages are not validated/verified at all? | 14:59 |
| uvos__ | the app mgr only shows extras packages which are not signed | 15:00 |
| Wizzup | 14:59 < yifei1> so the packages are still signed like regular debian packages. | 15:00 |
| Wizzup | yes they are, by our extras key | 15:00 |
| Wizzup | uvos__: ^^ | 15:00 |
| uvos__ | oh ok sorry | 15:01 |
| yifei1 | well that's good to know ;) | 15:03 |
| uvos__ | regardless they are ofc not signed by nokia | 15:03 |
| Wizzup | but yes that needs to be improved for sure | 15:11 |
| Wizzup | it's quite easy really.. | 15:11 |
| Wizzup | yifei1: does the device boot ok with your fix now? | 19:17 |
| Wizzup | the freeze/resets we can take a stab at then, there is probably also a newer kernel avail | 19:17 |
| Wizzup | uvos: do you think the BACKEND_FLAG_CELLULAR? relates to the audio issues? | 19:38 |
| Wizzup | maybe I suppose | 19:38 |
| Wizzup | uvos: should we try a newer kernel re: cpufreq missing? | 20:00 |
| Wizzup | or should we investigate some more since iirc this is/was lts? | 20:00 |
| uvos__ | Wizzup: BACKEND_FLAG_CELLULAR - no not related | 22:06 |
| uvos__ | Wizzup: newer kernel: yes newer linux stable of 6.6 | 22:06 |
| uvos__ | should be tried | 22:06 |
| Wizzup | ok | 22:52 |
| Wizzup | I still wonder if it's not a config issue somehow | 22:52 |
| Wizzup | it must be, right? :D | 22:53 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!