| Wizzup | uvos: there is now leste-config-mz617, hildon-meta-mz617 and the mapphone kexecboot config pkg for mz617 | 01:32 |
|---|---|---|
| Wizzup | image builder will be slightly more tricky since /boot is on a separate partition so this needs proper fstab and we don't want partition, we just want a tar | 01:33 |
| Wizzup | tmlind: the maemo experimental kernel (some 6.6) doesn't seem to have the ts issues as often (where the probe results in no ts) | 02:02 |
| Wizzup | whether it is completely gone I cannot say yet | 02:03 |
| Wizzup | there is still the audio/modem issue: | 02:03 |
| Wizzup | [ 41.857635] mot-mdm6600-codec 4806a000.serial:modem:audio-codec@2: ASoC: error at snd_soc_component_probe on 4806a000.serial:modem:audio-codec@2: -110 | 02:03 |
| Wizzup | [ 41.901000] asoc-audio-graph-card soundcard: ASoC: failed to instantiate card -110 | 02:03 |
| Wizzup | [ 41.916503] asoc-audio-graph-card soundcard: error -ETIMEDOUT: parse error | 02:03 |
| Wizzup | [ 41.931182] asoc-audio-graph-card: probe of soundcard failed with error -110 | 02:03 |
| Wizzup | arno11: I'll try to get your various changes in a leste-config pkg update for -devel today | 13:03 |
| uvos | tmlind: did you ever figure out why the command mode displays are no longer refreshed since 6.4 or so? | 18:25 |
| Wizzup | uvos: which devices that we support are command line? | 20:23 |
| Wizzup | is that n9 or something? | 20:23 |
| Wizzup | command mode* | 20:23 |
| sicelo | d4 too, iirc | 20:26 |
| Wizzup | arno11: building new experimental kernel with ir stuff | 20:26 |
| Wizzup | uvos: for mz617 it would be nice to have a way to raise keyboard without a button | 20:33 |
| Wizzup | wonder if we can eventually look at some gesture stuff | 20:33 |
| arno11 | Wizzup: (for leste-config and kernel) great ! | 21:04 |
| Wizzup | the kernel is building now | 21:05 |
| Wizzup | did you have a suggested change for the blur change | 21:05 |
| arno11 | ok | 21:05 |
| arno11 | what do you mean exactly for blur ? | 21:06 |
| Wizzup | well, you wrote that disabling the blur in the transitions is helpful | 21:07 |
| Wizzup | oh, just the radius? | 21:07 |
| arno11 | yep so you just need to set radius and radius_more to zero | 21:08 |
| arno11 | that's enough to block the blur effect | 21:08 |
| arno11 | but few other modifications are really cool | 21:09 |
| Wizzup | ok, but under what entry do you set this | 21:09 |
| Wizzup | there's a few different places that can accept radius= I think | 21:09 |
| arno11 | in transitions.ini, (home) | 21:09 |
| Wizzup | ok, [home] | 21:09 |
| arno11 | yep | 21:10 |
| Wizzup | how did you come to these sysctl settings? | 21:11 |
| arno11 | well, a bit long to explain, that's a combination of things found in TMO and in debian/ubuntu known bugs | 21:14 |
| arno11 | mainly, Leste original sysctl settings are too 'generic' for only 256MB of RAM and N900 limited SD bus speed imo | 21:17 |
| Wizzup | ok | 21:18 |
| arno11 | but that's not miraculous :P | 21:18 |
| Wizzup | the nr_requests I will have to make an initscript for :( | 21:18 |
| arno11 | arghhh | 21:18 |
| uvos | Wizzup: all mapphones | 21:18 |
| uvos | except mz6xx | 21:19 |
| Wizzup | ok, so why does it still work for us, or what am I missing? | 21:19 |
| uvos | it dosent work for you | 21:19 |
| uvos | xorg happens to work because it manually calls for a display refesh | 21:19 |
| Wizzup | ok | 21:19 |
| uvos | everything outside xorg is broken | 21:19 |
| uvos | also [ 9.915008] cpcap-core spi0.0: Failed to create device link (0x180) with 40126000.mcbsp | 21:20 |
| uvos | looks like mcbsp is stil broken in 6.6 | 21:21 |
| uvos | which is pretty critcal when combined with the display not refeshing | 21:21 |
| uvos | (mcbsp is omap serial port phi) | 21:21 |
| Wizzup | ok, we can try to figure this out, but let's please also get 6.6 as up to date as we can, with our patches, then at least we have a good base to work from | 21:23 |
| Wizzup | and I do think I remember indeed seeing no serial console output until X started | 21:23 |
| Wizzup | so you're right, that's probably the bug you're describing | 21:23 |
| uvos | sure i have rebased it | 21:23 |
| uvos | the hp paches dont work | 21:23 |
| uvos | we also have some other new issues | 21:23 |
| uvos | [ 9.678344] cpcap_battery cpcap_battery.0: failed to register power supply | 21:23 |
| uvos | [ 10.135711] iio iio:device2: unsupported IRQ trigger specified (3), enforce rising edge. | 21:24 |
| uvos | with the ils being broken | 21:24 |
| uvos | that one is probubly a easy fix | 21:24 |
| uvos | but im looking at the hp first | 21:24 |
| Wizzup | ty :) | 21:24 |
| uvos | the modem seams to work | 21:25 |
| uvos | but havent tried modem audio yet | 21:25 |
| uvos | but no dmesg messages suggesting anything wrong with that | 21:25 |
| Wizzup | well, let me know if you tried it and it works for you | 21:25 |
| Wizzup | arno11: I'm still thinking on the best way to set these values | 21:27 |
| Wizzup | I need to figure out how to add an init script to leste-config-n900 I guess | 21:27 |
| arno11 | ok | 21:27 |
| uvos | these values probubly help mapphones just as mutch or? | 21:27 |
| uvos | not sure they would be harmfull anywhere since all our devices have lowish ram | 21:28 |
| Wizzup | I am not sure, but I figured just having it set for n900 first would help | 21:28 |
| uvos | ok | 21:28 |
| Wizzup | are you talking about sysctl or the block queue nr requests? | 21:28 |
| uvos | the vm changes | 21:28 |
| uvos | since i dont quite know what the nr_requests dose | 21:29 |
| Wizzup | right | 21:29 |
| Wizzup | it will be easy enough to move those over, I will just make sysctl.d file from leste-config-n900 | 21:29 |
| Wizzup | (we can make that a thing in leste-config-common or leste-config-mapphone as well) | 21:29 |
| uvos | hmm? that sentace dosent quite add up | 21:29 |
| Wizzup | also, yeah, I'm sure the default values are optimised for servers :P | 21:30 |
| Wizzup | uvos: what doesn' | 21:30 |
| Wizzup | t? | 21:30 |
| uvos | "I will just make sysctl.d file from leste-config-n900" should be "I will just take sysctl.d file from leste-config-n900" or "I will just make sysctl.d file for leste-config-n900" i think | 21:30 |
| arno11 | bbl (dinner time + kids) | 21:31 |
| Wizzup | the latter | 21:31 |
| uvos | ok | 21:31 |
| uvos | check | 21:31 |
| uvos | https://github.com/maemo-leste/droid4-linux/commits/maemo-6.6-buildeb/ | 21:53 |
| uvos | Wizzup: so what should get into the kernel from that, you have a commit in there DO NOT USE: make dtbs build for now | 21:53 |
| Wizzup | uvos: pull everything from there | 21:53 |
| Wizzup | I just mark some commits as not like, upstream ready or something | 21:54 |
| Wizzup | arno11: the build failed with the ir patches: https://phoenix.maemo.org/job/droid4-linux-binaries/architecture=armhf,label=armhf/192/consoleText | 21:54 |
| Wizzup | I will try to look at this in a bit | 21:55 |
| Wizzup | but jfyi | 21:55 |
| uvos | ok no ir patches for now | 22:08 |
| Wizzup | *nod* | 22:12 |
| Wizzup | btw, mz617 with leste is quite usable now already, just no audio / pm (RET) yet, but those are related | 22:14 |
| uvos | i need figure out how to revive my mz617 | 22:15 |
| Wizzup | What happened to it? | 22:16 |
| uvos | its soft-bricked | 22:16 |
| uvos | i was working on the kernel | 22:16 |
| uvos | and installed one that dosent boot at all | 22:16 |
| uvos | and i had overriden a part of android that makes that not work | 22:16 |
| uvos | and then the battery died while i tried to figure out how to get it back | 22:16 |
| arno11 | Wizzup: ok | 22:17 |
| uvos | its in a tricky state since you cant fix some of the android partitions without allow-mbm.... mbm | 22:18 |
| uvos | but kexecboot still works with its busybox shell | 22:18 |
| uvos | so its possible to recover it | 22:18 |
| uvos | somehow :P | 22:18 |
| Wizzup | you can flash a working leste image to the data partition? | 22:19 |
| uvos | after i take it apart and charge the battery some other way anyhow | 22:19 |
| uvos | maybe yeah | 22:19 |
| Wizzup | assuming 'and installed one that dosent boot at all' is for post-kexecboot kernel | 22:19 |
| uvos | but for now it wont turn on at all | 22:19 |
| Wizzup | ok | 22:19 |
| Wizzup | well I think with image builder I should be able to make a tar tonight or tomorrow that you could extract on the data partition | 22:19 |
| uvos | yeah but how do i get the tar onto the deivce | 22:20 |
| Wizzup | well if you have the busybox shell I think you can probably do it somehow | 22:20 |
| uvos | using ftp and wifi in kexeboot shell i gues | 22:20 |
| Wizzup | either that or base64 encoded over serial :P | 22:20 |
| uvos | uff | 22:20 |
| Wizzup | (that was a joke) | 22:20 |
| Wizzup | I could give you my smaller image that is btrfs I suppose, it's only 700MB | 22:23 |
| Wizzup | but I don't think you can flash it with fastboot to the cdrom partition or something | 22:23 |
| Wizzup | uvos: I suppose usb otg doesn't work in kexecboot? | 22:23 |
| uvos | no | 22:23 |
| Wizzup | ok | 22:24 |
| uvos | but a small enough image you could flash to cache | 22:24 |
| uvos | i put kexecboot somewhere else | 22:24 |
| uvos | and that partition you can flash from fastboot | 22:24 |
| uvos | i dont remeber off hand how big it is | 22:24 |
| uvos | anyhow its recoverable | 22:25 |
| Wizzup | I have this: BTRFS_IMAGE_WORKS_FOR_CACHE_NO_LOCALES.img | 22:25 |
| uvos | its just annoying | 22:25 |
| uvos | i see | 22:25 |
| uvos | sounds good :P | 22:25 |
| Wizzup | I could upload this somewhere | 22:25 |
| Wizzup | ok | 22:25 |
| uvos | yes please | 22:25 |
| Wizzup | per my notes I think I had to dd it from android to cache, but if you're sure you can flash it with fastboot the nyou should be good | 22:27 |
| Wizzup | eta ~11 mins or so | 22:27 |
| Wizzup | I don't know if you have another mz61x, but you could swap batteries if one has android | 22:28 |
| uvos | i dont, but charging the battery is no issue | 22:28 |
| Wizzup | I figured :p | 22:28 |
| uvos | Sending 'cache' (921600 KB) OKAY [ 35.833s] | 23:03 |
| uvos | Writing 'cache' OKAY [ 78.537s] | 23:03 |
| uvos | not sure why your mz617's mbm would be unhappy about that | 23:03 |
| Wizzup | weird | 23:03 |
| arno11 | Wizzup: uvos: btw what's your opinion about gnome-tracker 2 on a phone ? imo, totally useless: once deactivated on leste, no change and the 'find' funtion still works fine. | 23:21 |
| arno11 | and lot of cpu and ram saved | 23:22 |
| arno11 | *function | 23:22 |
| Wizzup | tracker is a pretty integral part for OMP and other things | 23:22 |
| Wizzup | I don't think we should get rid of it, and it doesn't really cause any issues on my d4 | 23:23 |
| uvos | i dont like it either - even on d4 its petty often far up the powertop list | 23:32 |
| uvos | but ok | 23:32 |
| uvos | i think nerfing it is enough rather than removeing it | 23:32 |
| arno11 | ok thx guys, so maybe it should be useful to prevent it to scan the entire home folder ? | 23:36 |
| arno11 | or maybe it scans the entire sd card lol | 23:36 |
| Wizzup | yeah so this is a single gsetting change | 23:38 |
| Wizzup | maybe see if it makes a difference for you | 23:39 |
| arno11 | iirc we already tried | 23:39 |
| arno11 | let me check | 23:41 |
| arno11 | ah i failed to remove the $HOME directory | 23:43 |
| Wizzup | maybe reboot just for good measure | 23:44 |
| Wizzup | sicelo: btw I fixed the image builder again for now | 23:44 |
| Wizzup | (disk filled up...) | 23:45 |
| uvos | can we have a bigger disk or a script that deletes old releases or something | 23:45 |
| uvos | since that happens alot | 23:45 |
| Wizzup | what happens here is that there is still some edge case in arm-sdk where it does not umount | 23:46 |
| Wizzup | and then jenkins cannot deal with that | 23:46 |
| Wizzup | and it doesn't clean | 23:46 |
| uvos | ok | 23:46 |
| Wizzup | the arm-sdk stuff is all zsh and it's not my strong suit but I did try to fix it | 23:46 |
| uvos | ok | 23:46 |
| Wizzup | honestly it's like a 1 minute manual cleanup but yeah, I forget to check... | 23:46 |
| uvos | mz617 tar soon? | 23:47 |
| uvos | cron job hack? | 23:47 |
| Wizzup | it will take me a bit more to make the tar, but I am trying to make the .img (which isn't as useful...) | 23:47 |
| uvos | thats finem i can extract | 23:48 |
| Wizzup | I need to figure out how to make arm-sdk also produce tar files like it used to | 23:48 |
| uvos | i just want something to install to userdata | 23:48 |
| Wizzup | did you manage to boot the minimal image? | 23:48 |
| uvos | yes | 23:48 |
| Wizzup | ah, in that case just take the droid 4 image, losetup, and rsync it to the data partition | 23:48 |
| Wizzup | and then add devel and remove hildon-meta-droid4 and leste-config-droid4 and the mapphone kernel config (or so) and install the mz617 versions of them | 23:48 |
| Wizzup | but yes, I'm working on the image but realistically it's at least a few hours off (it tried to build some 4.14 kernel for some reason...) | 23:49 |
| uvos | ok | 23:49 |
| uvos | 6.6 kernel looks good now (besides mentioned issue with msub and cm pannels) | 23:49 |
| Wizzup | did the call audio work for you? | 23:50 |
| uvos | havent tested | 23:50 |
| Wizzup | ok | 23:50 |
| Wizzup | for completeness sake: hildon-meta-mz617 leste-config-mz617 mz617-kexecboot-config | 23:50 |
| uvos | im installing chimaera on the bionic rn | 23:50 |
| uvos | to try it there | 23:50 |
| Wizzup | ok | 23:50 |
| Wizzup | you probably need to mount the android system partition on /boot/ for the mz617-kexecboot-config to work | 23:51 |
| Wizzup | but you can also just skip mz617-kexecboot-config , just make sure to remove the droid4 version as well | 23:51 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!