missMyN900 | Wizzup, sixwheeledbeast: https://paste.debian.net/1242621/ on PP running the 05/29 image (no apps open, charging) | 00:02 |
---|---|---|
missMyN900 | I have installed the WiFi switch applet this time too, so I will trying that. At least that should work. | 00:05 |
missMyN900 | norayr: never mind, I see it is already available! | 00:06 |
missMyN900 | going devel now | 00:09 |
missMyN900 | status menu volume applet has been installed automatically | 00:33 |
missMyN900 | going to test it | 00:33 |
missMyN900 | doesn't work | 00:35 |
missMyN900 | updated sinks.ini; will try again | 00:38 |
missMyN900 | the stock minimal brightness on the Pinephone is way too high. After some trial and error I finally managed to lower the brightness with redshift (which I use on my computers to reduce blue light in the evening) | 06:25 |
missMyN900 | I mean the brightness with status area applet, of course | 06:26 |
missMyN900 | smplayer froze multiple times for some reason during opus playback | 06:28 |
missMyN900 | (with mplayer as backend) | 06:28 |
missMyN900 | using mplayer from the terminal app I did not experience this problem with the Opus files | 06:28 |
missMyN900 | I think I also experienced some general Hildon/X freezes | 06:29 |
missMyN900 | the audio experience still stucks on the PP | 06:30 |
missMyN900 | status area volume applet still does not work | 06:31 |
missMyN900 | with devel enabled (after rebooting) | 06:31 |
missMyN900 | and updating sinks.ini (after rebooting, again) | 06:31 |
missMyN900 | pavucontrol-qt is useless because you can't scroll to access the lower output sink, which happens to be the relevant one... | 06:31 |
missMyN900 | only pavucontrol (GTK3) works | 06:32 |
missMyN900 | however, I discovered that it is also not a good idea to leave pavucontrol running because it causes high CPU usage | 06:32 |
missMyN900 | the application itself uses 22% but pulseaudio usage for some reason jumps from 8 to 20% as well | 06:33 |
missMyN900 | and Xorg also uses a lot of CPU when pavucontrol is running | 06:33 |
missMyN900 | hmm | 06:35 |
missMyN900 | according to pactl info: Default Sink: alsa_output.0.HiFi__hw_PinePhone_0__sink | 06:35 |
missMyN900 | looks like it was not a good idea to change sinks.ini | 06:35 |
missMyN900 | and this is with a fresh install (05-29 image with devel enabled) | 06:36 |
Wizzup | 23:57 < missMyN900> norayr: I have an excellent porting suggestion for you: desktop command execution widget! | 10:02 |
Wizzup | didn't sicelo do thta? | 10:02 |
Wizzup | missMyN900: that's a lot of interrupts | 10:03 |
_uvos_ | intterupts whilest on ac power dosent tell us mutch | 15:18 |
_uvos_ | some drivers ajust behavior | 15:18 |
_uvos_ | crust allows suspend & reaction to modem, not the same kind of pm as omap afaik | 15:19 |
_uvos_ | that widget is in the repo allready | 15:19 |
_uvos_ | try QT_SCALE_FACTOR=1.5 pavucontrol-qt -platform xcb -style fusion | 15:19 |
Wizzup | right, omap pm is different | 15:20 |
_uvos_ | i gues you could try implementing some kind of ip push notifications on the crust uc | 15:22 |
_uvos_ | to further paper over the pm gap on pp | 15:23 |
humpelstilzchen[ | iirc wifi can not wakeup the PP from crust. Only the modem, rtc and a few others | 15:24 |
_uvos_ | that (or mp3 decoding) would also be really neet on omap4's cortex m4 cpus | 15:24 |
_uvos_ | which you can use for anything | 15:24 |
_uvos_ | humpelsilzchen: ok not sure what crust has access to in hw | 15:25 |
_uvos_ | on omap4 the equivalent can access any bus iirc | 15:25 |
_uvos_ | and could directly talk to the wifi chip | 15:26 |
_uvos_ | this would involve the linux driver somehow handing hw resources over to the uc firmware.. diffiult | 15:27 |
_uvos_ | but possible | 15:27 |
Wizzup | are there hidden services for devuan repos? | 15:42 |
Wizzup | i.e. like https://onion.debian.org/ | 15:42 |
Wizzup | ah they do | 15:45 |
Wizzup | ok, I'm able to update using only tor for devuan and maemo repos | 15:47 |
bencoh | nice | 15:49 |
bencoh | :) | 15:49 |
missMyN900 | alright, so looks like I finally got the status menu volume applet to work | 20:56 |
missMyN900 | changed the sink to the "default sink" in the pactl info output | 20:56 |
missMyN900 | and rebooted and now I can see the slider in pavucontrol responding to the slider in the status area | 20:56 |
missMyN900 | but the sink suggested in that GH issue with "platform-sound" definitely does not work | 20:57 |
missMyN900 | maybe it worked on older images or maybe it does on older (distro edition?) 2/16 GB PPs but it certainly doesn't with my setup | 20:57 |
missMyN900 | _uvos_: I doubt crust depends on the modem. Remember, the A64 is a tablet SoC (that is also used in SBCs such as the Pine A64 and laptops such as the original Pinebook) | 20:59 |
missMyN900 | the majority of A64 devices does not have the Quectel/Qualcomm modem | 21:00 |
missMyN900 | the same applies for the other SoCs supported by Crust, which are settop box SoCs that are used in SBCs (except for A83T) | 21:01 |
humpelstilzchen[ | missMyN900: (slider) cool, what exactly did you write in the file and maybe you should add that to the ticket | 21:03 |
missMyN900 | humpelstilzchen[: [9:35:34 pm] <missMyN900> according to pactl info: Default Sink: alsa_output.0.HiFi__hw_PinePhone_0__sink | 21:03 |
missMyN900 | I put that sink in sinks.ini | 21:03 |
missMyN900 | I did not have time to test it yesterday, so I just powered off the PP for the night | 21:04 |
missMyN900 | and booted it up today and now it seems to be working | 21:04 |
missMyN900 | still need to test with (s)mplayer though | 21:04 |
Wizzup | missMyN900: check @ getting it to work | 21:25 |
Wizzup | missMyN900: I think crust depends on the modem to wake it up is the point | 21:25 |
missMyN900 | it is working with mplayer! | 21:27 |
missMyN900 | this is a big improvement | 21:28 |
missMyN900 | Wizzup: most A64 devices do not even have a modem. Crust runs on the Allwinner AR100 (an OpenRISC microcontroller) of the Allwinner SoC. | 21:32 |
missMyN900 | "Similarly, Crust is the lowest-level firmware component that runs on $FRUIT (Banana, Orange, Lichee) Pi single-board computers and other Allwinner-based devices, such as the Pine64 Pinebook and PinePhone." | 21:33 |
missMyN900 | "Note that Crust only provides the mechanism for deep sleep. It does not dictate any system sleep policy. Specifically, Crust does not decide when to go to sleep; the Linux kernel or userspace does that. And with one exception (listening for IR remote control key presses), Crust does not decide when to wake the system up, either; the hardware, as programmed by Linux, does that. Crust is designed to be a mostly-invisible | 21:34 |
missMyN900 | implementation detail of the Linux power management interface." | 21:34 |
humpelstilzchen[ | check /sys/class/wakeup | 21:34 |
humpelstilzchen[ | there are a lot wakeup sources | 21:34 |
missMyN900 | humpelstilzchen[: ok but my point is crust does not depend on the modem. And I don't think that ML should rely on the modem since it would be nice if the pm transferred to the Pinetab, which does not have a modem | 21:37 |
missMyN900 | furthermore, it would be nice if the pm still worked properly with the modem kill switch enabled (or the modem otherwise turned off) on the PP | 21:38 |
humpelstilzchen[ | missMyN900: if in doubt just use the rtc to wakeup | 21:38 |
missMyN900 | humpelstilzchen[: are you saying there is a manual way I can put it into deep sleep? | 21:39 |
humpelstilzchen[ | you might want to keep the modem wakeup source enabled for e.g. wakeup when there is a call. | 21:40 |
humpelstilzchen[ | missMyN900: manual way see https://wiki.pine64.org/wiki/Crust | 21:40 |
missMyN900 | it uses systemd though :/ | 21:41 |
missMyN900 | oh there is non-systemd option | 21:41 |
missMyN900 | so basically I can just experiment with this? | 21:42 |
humpelstilzchen[ | sure | 21:43 |
missMyN900 | interesting :) | 21:43 |
missMyN900 | I am also interested in the unofficial modem "firmware" | 21:43 |
missMyN900 | it can lower the Quectel A7 frequency to 100 MHz from 800/400 MHz | 21:44 |
humpelstilzchen[ | missMyN900: you can also spy on e.g. sxmo how they are using it: https://git.sr.ht/~mil/sxmo-utils/tree/d2d06ef65ac63af41516d334cd8a5446a0f20cc3/item/programs/sxmo_screenlock.c | 21:44 |
missMyN900 | so they are doing all this stuff from userspace? | 21:44 |
missMyN900 | I thought I would have to become a kernel hacker to do this myself | 21:44 |
humpelstilzchen[ | the kernel part is mostly done it seems | 21:45 |
humpelstilzchen[ | firmware: You are talking about https://github.com/Biktorgj/pinephone_modem_sdk I guess. | 21:46 |
Wizzup | missMyN900: the point is that crust without any way to wake it up on action doesn't help | 21:46 |
missMyN900 | yes | 21:46 |
missMyN900 | is there a way to use the power button to wake it up? | 21:47 |
humpelstilzchen[ | I can't say anything about it, since I don't have a sim in the PP, but I've installed the open firmware. GPS still works | 21:47 |
humpelstilzchen[ | power button is already active wakeup source | 21:47 |
missMyN900 | so it is already going to deep sleep? | 21:47 |
missMyN900 | when the screen is locked? | 21:47 |
missMyN900 | battery life does not seem to reflect that... | 21:47 |
humpelstilzchen[ | no, leste does not do that yet. You still need to manually start the crust sleep | 21:48 |
humpelstilzchen[ | but you can that with the simple echo mem > /sys/power/state | 21:48 |
missMyN900 | oh but I can just put it in suspend and it will already automatically wake up when pressing the power button? | 21:49 |
sicelo | Yes | 21:50 |
missMyN900 | why didn't anyone tell me this before? :D | 21:50 |
missMyN900 | it is working! | 21:51 |
missMyN900 | I did it over ssh and then the command appears to "hang" (since it has already gone into deep sleep) | 21:51 |
missMyN900 | and when I press the power button the command "returns" | 21:51 |
humpelstilzchen[ | wakeup crashes my pp from time to time. Had no chance to debug that yet. Also I still have not soldered myself a pp debug cable yet | 21:53 |
humpelstilzchen[ | also you might want to disable some wakeup sources in /sys/class/wakeup/ or the PP will just randomly wakeup | 21:54 |
missMyN900 | the entire device or just X/Hildon? | 21:54 |
humpelstilzchen[ | ? | 21:55 |
missMyN900 | can you still access it over SSH? | 21:55 |
missMyN900 | when that crash occurs? | 21:55 |
humpelstilzchen[ | no network, no x, nothing. | 21:55 |
missMyN900 | so kernel panic? | 21:56 |
humpelstilzchen[ | I guess so | 21:56 |
missMyN900 | maybe you could add some logging code that executes on wakeup? | 21:57 |
humpelstilzchen[ | the notification led is white, everything else is dark. The only thing I can do is long press the power button until it turns off | 21:57 |
missMyN900 | hmm noti LED is always white when it turns off | 21:57 |
missMyN900 | it always takes a little while for the LED to turn off | 21:57 |
humpelstilzchen[ | does not help since a flush is needed to write the disk. I will solder myself a serial debug cable one day | 21:58 |
missMyN900 | hopefully it will get fixed upstream | 21:58 |
missMyN900 | there may be people experimenting with this with SBCs (such as Pine A64) | 21:59 |
missMyN900 | if anyone wants to get a quick list of the PP wakeup sources, this is the easy way to get it: "cat /sys/class/wakeup/*/name" | 22:06 |
missMyN900 | humpelstilzchen[: one question that I have been wanting to ask (somewhere) but forgot to is, does the alternative modem firmware work with ofono too? | 22:46 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!