sicelo | Wizzup: freemangordon: uvos: please review https://github.com/maemo-leste/mce/pull/59 and also https://github.com/maemo-leste-upstream-forks/upower/pull/5 | 05:56 |
---|---|---|
sicelo | with those in place, the battery icon on status stops animating when charge full, and the green led works (at least on N900) | 05:57 |
sicelo | arno11: you can also test :-) | 06:02 |
sicelo | to test the updated packages without rebooting, (1) sudo rc-service mce stop, (2) sudo pkill -f upower, (3) sudo rc-service mce start | 06:17 |
tmlind | freemangordon: seems like that should work though.. at least the wake up enable bit gets set when omap-usb-host is runtime suspended | 08:30 |
freemangordon | tmlind: yes, but no interrupt | 15:03 |
sicelo | i have internet problems today for some reason, so i'm struggling with updating my git repositories for droid4-linux. i have a patch that i sent as an issue. https://github.com/maemo-leste/droid4-linux/issues/7 | 17:26 |
sicelo | please review and merge ... i'll also get it ready for upstream | 17:26 |
freemangordon | sicelo: why not send it to upstream? | 17:35 |
freemangordon | ah, sorry | 17:35 |
sicelo | i would like it to hit Leste in the meantime, since of course upstream will take a while. no idea as well if they'd decide to backport it to the kernel we're on | 17:38 |
sicelo | it fixes that issue with charging animation never stopping while a charger is connected, even when battery is really full | 17:40 |
freemangordon | sicelo: yeah, I understand what it is used for, the point is that we can backport it easily after upstream comments on it. And that usually happens in 1-2 days | 18:03 |
freemangordon | and at least I will not have time soon to test it :( | 18:04 |
sicelo | let's see after i get a chance to refresh my repos | 18:16 |
Wizzup | freemangordon: we often include patches before they go to upstream fwiw | 19:16 |
sicelo | Wizzup: i have a question about https://github.com/maemo-leste/mce/blob/master/src/modules/battery-upower.c#L79-L80 ... do you know any further details about that? i can see it was added by spinal, but you might know something :-) | 19:21 |
sicelo | the thing is ... in status-area-applet-battery, he eventually reverted it | 19:23 |
sicelo | i'll be working on dropping the n900-related blacklist entries, so i was wondering about the one for d4 | 19:24 |
Wizzup | (will check after dinner) | 19:29 |
freemangordon | Wizzup: I know | 19:32 |
freemangordon | I don't object, just saying that perhaps review will happen faster if sent to upstream :) | 19:32 |
tmlind | freemangordon: maybe the kernel driver for qmi does not handle usb suspend so the modem won't know to initiate a wake-up? i'd asssume the pad wake-up event works just fine.. | 19:43 |
tmlind | freemangordon: it could be the kernel driver just tells the qmi modem to shut up while suspended.. | 19:44 |
freemangordon | could be, any idea where to look in? | 19:45 |
freemangordon | this is option driver? | 19:46 |
freemangordon | or cdc-wdm? | 19:47 |
freemangordon | maybe I can connect my ec20 to d4 to see what will happen | 19:47 |
freemangordon | tmlind: ^^^ | 19:48 |
freemangordon | does it make sense to write to the ML? | 19:48 |
tmlind | freemangordon: no idea where to look.. yeah maybe asking the qmi folks would help or try to test with your usb modem on a pc with wakeup and autosuspend enabled | 19:49 |
freemangordon | also, do you know how it behaves with vendor kernel? | 19:49 |
freemangordon | I was not able to make it suspend when connected to my laptop | 19:49 |
tmlind | freemangordon: afaik the android kernel uses the gsmmux for everything except for cell data | 19:50 |
freemangordon | yeah | 19:50 |
freemangordon | do you have any idea why would it not suspend when connected to PC? | 19:50 |
tmlind | testing with a usb mouse with autosuspend and wakeup enabled on some similar device could help.. | 19:50 |
tmlind | freemangordon: did you enable d4 modem autosuspend (and wakeup) on your pc via sysfs? | 19:51 |
freemangordon | autosuspend is enable (200 ms for the modem, 0 ms for the port) | 19:52 |
freemangordon | wakeup is disabled | 19:52 |
freemangordon | *2000 ms | 19:52 |
freemangordon | but runtime_status stays active | 19:52 |
freemangordon | the same happens in leste VM | 19:53 |
freemangordon | maybe because kernel there is older, dunno | 19:53 |
tmlind | so you did echo auto on pc? | 19:53 |
freemangordon | control? | 19:53 |
tmlind | yeah | 19:53 |
freemangordon | oh, wait, it is on | 19:53 |
freemangordon | lemme try | 19:53 |
tmlind | ok | 19:54 |
tmlind | so i think after you enable autosuspend on some usb device like a mouse, the mouse will send the k-state wake when you click on a button | 19:54 |
freemangordon | yeah, ok, ec20 is suspended now | 19:54 |
tmlind | ok | 19:54 |
freemangordon | will try with d4, gimme a minute | 19:54 |
tmlind | so then you need some command to remotely get modem(s) to do something, like send them sms? | 19:55 |
freemangordon | hmm, will take more time | 19:56 |
freemangordon | I need to add my patches in VM ofono | 19:56 |
tmlind | ok | 19:57 |
freemangordon | otherwise it is easy to test it, RSSI gets reported on every change | 19:57 |
freemangordon | ok, will let you know what I test it, most-probably tomorrow as otherwise GF will get angry on me :) | 19:57 |
freemangordon | *when I test | 19:57 |
tmlind | yeah ok, maybe check with a usb mouse too if modems won't work for wake-up | 20:00 |
freemangordon | ubuntu kernel on my laptop refuses to suspend it :( | 20:04 |
freemangordon | tmlind: what happens on d4 seems to happen in the VM when modem gets suspended | 20:17 |
freemangordon | no wakeup happens, despite it is enabled | 20:17 |
freemangordon | I enabled wakeup on both modem and its parent | 20:17 |
freemangordon | to no use | 20:17 |
tmlind | ok, also for your usb qmi modem? | 20:17 |
freemangordon | lemme check with it | 20:18 |
freemangordon | tmlind: yeah, looks like the same | 20:25 |
freemangordon | so most-probably qmi driver does not properly support suspend | 20:25 |
freemangordon | :( | 20:26 |
freemangordon | tmlind: ^^^ | 20:26 |
freemangordon | ttyl | 20:27 |
tmlind | yeah ok ttyl | 20:28 |
tmlind | freemangordon: maybe it's related to the needs_remote_wakeup in usb drivers, at least qmi_wwan_manage_power() uses it | 20:32 |
Wizzup | sicelo: re your questions, I don't really know, maybe we can do without? | 20:36 |
sicelo | sure. i can't think of a reason too. and yes, spinal did revert it in status-area-applet-battery. i'll drop it in an MR i'm yet to submit, while dropping the others. only bq24150 will remain in blacklist | 20:43 |
sicelo | i'll fix the kernel driver for that, and the upower problem will fix itself | 20:44 |
freemangordon | tmlind: https://www.quectel.com/wp-content/uploads/2021/03/Quectel_WCDMALTE_Linux_USB_Driver_User_Guide_V1.7.pdf | 21:22 |
freemangordon | see 4.2 "Enable USB Remote Wakeup" | 21:23 |
freemangordon | I wonder why it is not in upstream | 21:23 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!