| inky | sicelo, i have heard in an old podcast that alpine wants to discntinue builds for n900 chipset. do you know about it? | 03:08 |
|---|---|---|
| tmlind | freemangordon: sounds like you figured out the vaudio stuff already, grepping the android values from uvos for 0800: shows 0x25 when in use and 0x65 when idle so only CPCAP_BIT_AUDIO_LOW_PWR gets set for idle, probably because of start-up latency causing audio glitches | 04:49 |
| tmlind | so yeah, the default value of 0 for the register is wrong and should be 0x65 instead | 04:50 |
| tmlind | uvos: heh a retro device suitable playback capture file name there? | 04:56 |
| freemangordon | tmlind: I think 0x65 gets set because regulator is put in standby | 06:45 |
| freemangordon | this is bit CPCAP_BIT_AUDIO_LOW_PWR set | 06:47 |
| freemangordon | this never happens with upstream as we don;t hit off mode, iiuc | 06:48 |
| freemangordon | als, I don;t think cpcap soc driver shall touch that register at all | 06:53 |
| freemangordon | it's already controlled by cpcap-regulator | 06:53 |
| sicelo | inky: i don't listen to their podcasts. which one? anyway, there's no such thing. I'm the maintainer for that device on pmOS and not aware of those plans | 06:53 |
| freemangordon | tmlind: I wonder if "abe-clkctrl:0030:0: failed to disable" I see in dmesg occasionally is harmless and not the reason for varying idle power usage | 06:59 |
| uvos__ | sicelo: well alpine wants to drop armhf no? | 13:06 |
| uvos__ | that would mean pmos has to do armhf itself or drop all old devices like n900 | 13:07 |
| sicelo | uvos__: meaning of armhf in context of alpine/pmos is different from meaning in debian | 13:55 |
| sicelo | debian's armhf is armv7 in alpine speak | 13:56 |
| sicelo | and alpine's armhf is something like armel in debian speak | 13:57 |
| arno11 | good to know | 14:00 |
| freemangordon | and how do they call armhf then? | 14:01 |
| Wizzup | yeah afaik armhf is just armhf | 14:13 |
| Wizzup | armhf==armv7a no? | 14:13 |
| Wizzup | ok, strange | 14:13 |
| freemangordon | Wizzup: how do you read https://patchwork.kernel.org/project/alsa-devel/patch/87y2a1vk3a.wl-kuninori.morimoto.gx@renesas.com/ ? | 14:13 |
| freemangordon | "It doesn't have "SWITCH" control yet, thus it start automatically when probed, but can't stop, so far." | 14:14 |
| freemangordon | I can't find a way to turn off modem output patch | 14:14 |
| freemangordon | *path | 14:14 |
| freemangordon | like, when calls work that is | 14:14 |
| Wizzup | but this is 2021 | 14:14 |
| freemangordon | yeah | 14:14 |
| Wizzup | wasn't this updated since then? | 14:14 |
| freemangordon | but is it still valid? | 14:14 |
| freemangordon | no idea | 14:14 |
| Wizzup | maybe do git log on audio-graph-card2 | 14:15 |
| freemangordon | yeah | 14:15 |
| freemangordon | Wizzup: so, I created widgets in DT for the remote modem and created routes like: | 14:18 |
| freemangordon | https://pastebin.com/SXqRwmZB | 14:19 |
| freemangordon | this, along with c2c config makes it work fine | 14:19 |
| freemangordon | but, "Remote Modem Speaker" is on all the time | 14:20 |
| Wizzup | sorry, I cannot check for another 2-3 hours | 14:39 |
| uvos__ | sicelo: ah so they are dropping armv6 not v7 ok | 15:00 |
| uvos__ | that dosent affect n900 then | 15:00 |
| freemangordon | oh, ok, seems to got it working properly | 15:49 |
| tmlind | freemangordon: yes standby makes sense for vaudio, note that probably for off mode it would not be shut off either as it's the pmic and not the soc, and enable latency could be long | 15:59 |
| tmlind | freemangordon: the abe-clkctrl disable message should be harmless timeout if the clock has the autoidle bit in dts, it just means it's in use and can't idle | 16:00 |
| freemangordon | tmlind: thanks! | 16:15 |
| freemangordon | ok, going to force-push 6.6, keep away :) | 16:15 |
| freemangordon | somehow even ringing while playing musing started to work :) | 16:26 |
| freemangordon | Wizzup: ^^^ | 16:26 |
| freemangordon | please, make a new kernel release | 16:27 |
| freemangordon | perhaps disable dynamic debug | 16:27 |
| freemangordon | dsc_: late pong | 16:29 |
| freemangordon | sorry, was deep in asoc s**t | 16:29 |
| dsc_ | freemangordon: wanted to annoy you with some C bug but I already found Wizzup as a victim | 16:39 |
| dsc_ | anyway, working on Matrix integration in conversations.. | 16:40 |
| freemangordon | why it needs special handling? | 16:41 |
| dsc_ | telepathy-tank is not useable at this point | 16:42 |
| dsc_ | its missing some basic features like joining/leaving groupchats | 16:43 |
| freemangordon | ok, but how is conversations involved? | 16:43 |
| dsc_ | ah | 16:44 |
| dsc_ | conversations needed a change to handle "room names" - which benefits both telegram and Matrix | 16:44 |
| freemangordon | ah, ok | 16:44 |
| dsc_ | but yes I'm mostly working on the telepathy side of things | 16:44 |
| freemangordon | that will help me to implement group names for tp-haze | 16:45 |
| dsc_ | its a property | 16:45 |
| freemangordon | good, but I don;t set it :) | 16:45 |
| dsc_ | QString room_name = props.value(QString("%1.Interface.Room2.RoomName").arg(TP_QT_IFACE_CHANNEL)).toString(); | 16:45 |
| dsc_ | right | 16:45 |
| freemangordon | in tp-haze I mean | 16:46 |
| dsc_ | yes | 16:46 |
| freemangordon | the reason was missing support in conversations ;) | 16:46 |
| dsc_ | I dont know if it should be "Room2", but this is what tp-tank does | 16:47 |
| freemangordon | well, there are specs | 16:47 |
| freemangordon | that describe the interfaces | 16:47 |
| freemangordon | https://telepathy.freedesktop.org/doc/book/ | 16:48 |
| freemangordon | https://telepathy.freedesktop.org/spec/index.html | 16:49 |
| freemangordon | https://telepathy.freedesktop.org/spec/Channel_Interface_Room.html | 16:49 |
| dsc_ | yes | 16:50 |
| dsc_ | https://github.com/maemo-leste/rtcom-eventlogger-plugins/blob/49c1843af4e93b9e7b3cb45cef42533857abd4c2/rtcom-eventlogger-plugins/chat.h#L45 | 16:51 |
| dsc_ | i failed to call this function | 16:51 |
| dsc_ | from conversations | 16:51 |
| dsc_ | https://bpa.st/raw/XIGQ | 16:51 |
| dsc_ | after thats fixed, conversations will have support for room names :) | 16:52 |
| dsc_ | https://plak.infrapuin.nl/selif/69e3l46u.png | 16:52 |
| freemangordon | :) | 16:52 |
| freemangordon | hmm, this is a plugin function | 16:53 |
| freemangordon | I am not sure you are allowed to call it | 16:53 |
| dsc_ | :D | 16:54 |
| freemangordon | oh | 16:54 |
| freemangordon | you need rtcom-eventlogger-plugins.pc | 16:54 |
| dsc_ | sure, i have it | 16:54 |
| freemangordon | weird | 16:54 |
| dsc_ | https://github.com/maemo-leste/conversations/commit/ea2de18ff0958d5a62958a57a708fecf64438d5d | 16:54 |
| dsc_ | yeah not sure | 16:54 |
| freemangordon | maybe the order is wrong | 16:55 |
| freemangordon | ld is picky about libs order | 16:55 |
| dsc_ | thing is I can reach other symbols | 16:55 |
| dsc_ | like enum `RTCOM_EL_FLAG_CHAT_GROUP` | 16:55 |
| dsc_ | well ok, its just an enum | 16:55 |
| freemangordon | this is in the header I guess | 16:55 |
| uvos__ | missing extern c? | 16:55 |
| freemangordon | yeah | 16:55 |
| dsc_ | thats via header include, yeah wrong example | 16:55 |
| uvos__ | it might be looking for a c++ mangled version | 16:56 |
| dsc_ | undefined reference to `rtcom_el_plugin_chat_set_group_title(_RTComEl*, char const*, char const*)' | 16:56 |
| dsc_ | but my args are 'gchar' | 16:56 |
| freemangordon | yeah, this is c++ | 16:56 |
| dsc_ | so this error is weird, no? | 16:56 |
| freemangordon | sec | 16:56 |
| freemangordon | uvos__: yeah, headers are missing extern "C" | 16:57 |
| uvos__ | yup | 16:57 |
| freemangordon | dsc_: make #include like "extern "C" { #include <chat.h> } | 16:58 |
| uvos__ | headers should be fixed with the ifdef cxx extern "C" { pattern | 16:59 |
| uvos__ | instead | 16:59 |
| uvos__ | but yeah that fixes it for dsc | 16:59 |
| dsc_ | compiling | 16:59 |
| freemangordon | uvos__: well, that's C library, not sure it must support c++ | 17:00 |
| dsc_ | that fixes it, nice o/ | 17:00 |
| dsc_ | thanks | 17:00 |
| uvos__ | freemangordon: why should it not? | 17:01 |
| freemangordon | well, I am not saying it should not | 17:02 |
| freemangordon | but that it is not a must | 17:02 |
| freemangordon | uvos__: with fixed 6.6 I have on my device, call ringing while music plays works, no idea why | 17:05 |
| freemangordon | the only remaining issue I am aware of with 6.6 is that reboot/powerdown does not work | 17:06 |
| uvos__ | its racy | 17:06 |
| freemangordon | not anymore it seems | 17:06 |
| uvos__ | it works with ofono on 6.1 too | 17:06 |
| uvos__ | it may return | 17:06 |
| uvos__ | dunno | 17:06 |
| uvos__ | or its fixed | 17:06 |
| freemangordon | it shoudl be, as now everything is controlled by the kernel | 17:07 |
| freemangordon | userspace just switches ucm | 17:07 |
| Wizzup | shutdown racy or call audio? | 17:07 |
| freemangordon | call audio I guess | 17:07 |
| Wizzup | 6.6 is great now | 17:07 |
| Wizzup | but yes 6.1 was also racy | 17:07 |
| freemangordon | did you try it? the one I just pushed? | 17:07 |
| uvos__ | hifi audio between modem activation and ucm swith is racy somehow | 17:07 |
| uvos__ | or was at least | 17:08 |
| freemangordon | it can;t be anymore | 17:08 |
| Wizzup | not that one but the last one | 17:08 |
| freemangordon | try the one I pushed | 17:08 |
| Wizzup | bbiab finally home | 17:08 |
| Wizzup | ok will do | 17:08 |
| freemangordon | this *is* the thing | 17:08 |
| Wizzup | I'm glad the issue wasn't in vcm in any case | 17:08 |
| freemangordon | at least sound wise | 17:08 |
| uvos__ | Wizzup: Wizzup tolled me to remind Wizzup about 2 droid4 when he arrives at home | 17:08 |
| Wizzup | 2 minutes too soon :p | 17:08 |
| uvos__ | ok ok | 17:09 |
| freemangordon | Wizzup: oh, and for some reason when I upgraded iphb was not rebuild | 17:09 |
| freemangordon | so there is still something wrong with debian scripts | 17:10 |
| arno11 | weird, it worked for me last time (iphb) | 17:16 |
| freemangordon | maybe because I installed the same version | 17:17 |
| freemangordon | next time will downgrade the kernel first | 17:17 |
| arno11 | ah yes that's what i did | 17:17 |
| arno11 | btw well done @calls | 17:18 |
| arno11 | for playing music while call ringing, you mean with omp or in general ? | 17:21 |
| freemangordon | омп | 17:22 |
| freemangordon | omp | 17:22 |
| freemangordon | lemme try with ytmusic | 17:23 |
| arno11 | ok | 17:23 |
| freemangordon | arno11: as expected, both sounds play :D | 17:28 |
| arno11 | cool :) | 17:29 |
| freemangordon | not really | 17:29 |
| freemangordon | ringing should stope everything else | 17:30 |
| freemangordon | lemme see what will happen if I pick-up | 17:30 |
| freemangordon | ok, at least that works properly, though I am sure chromium does not stop playback | 17:31 |
| freemangordon | unless PA does something behind the scenes to block non-active streams, dunno | 17:32 |
| arno11 | we can never be sure with PA | 17:35 |
| freemangordon | mafw-tracker-source seems to leak alot | 17:48 |
| freemangordon | arno11: that could be the reason for your n900 slowness | 17:48 |
| arno11 | ah indeed but i already blocked it before | 17:49 |
| arno11 | btw there is no slowness anymore | 17:49 |
| freemangordon | blocked what? | 17:49 |
| arno11 | mafw stuff | 17:49 |
| freemangordon | why? | 17:49 |
| arno11 | because of slowness (long time ago) | 17:50 |
| freemangordon | heh | 17:50 |
| freemangordon | ok | 17:50 |
| arno11 | the recent slowness is due to something wrong with emmc swap | 17:50 |
| arno11 | with sd swap, no more issue | 17:50 |
| arno11 | to resume: no blur, no mafw, no tracker in home and everything is fine | 17:51 |
| arno11 | on N900 | 17:51 |
| freemangordon | mafw should not cause any slowness, in theory | 17:51 |
| freemangordon | but, seems there is some memleak | 17:51 |
| freemangordon | I will look at it | 17:52 |
| arno11 | ok great | 17:52 |
| uvos | freemangordon: pa dose indeed block non active streams | 19:08 |
| uvos | this is something you can configure | 19:08 |
| uvos | you can simply make it not accept any samples anymore | 19:08 |
| uvos | unaware applications using blocking io will simply hang in write() untill the stream unblocks, smarter applications will do something smarter | 19:09 |
| uvos | in addition to this we could eventually use the fdo media interface to tell applications to pause | 19:10 |
| uvos | https://gitlab.freedesktop.org/mpris/mpris-spec/ | 19:11 |
| Wizzup | bonus of having mpris is that it works with bluetooth connected devices like cars | 19:56 |
| Wizzup | uvos: I believe PA calls this "cork"ing a stream, or maybe nokia called it this | 19:57 |
| arno11 | Wizzup: btw (if you have time to have a look) i still have a PR pending for leste-config (swappiness and cache pressure too 'agressive') | 21:54 |
| arno11 | *specially for slow sd cards | 21:55 |
| Wizzup | care to link it here? | 21:56 |
| arno11 | yep let me check | 21:56 |
| arno11 | /leste-config/pull/47 | 21:58 |
| fmg_d4 | where is the new 6.6? | 21:58 |
| fmg_d4 | :p | 21:59 |
| arno11 | in experimental repo i suppose | 22:03 |
| Wizzup | freemangordon: hmm @ 6.6? | 23:32 |
| Wizzup | arno11: ty | 23:32 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
