auanta | hello the wiki has several improvements | 00:27 |
---|---|---|
auanta | question about one of the sections here, are these still applicable?: https://leste.maemo.org/PinePhone#More_Information | 00:28 |
auanta | pertaining to the status/functionality | 00:38 |
Wizzup | I think the virtual keyboard might be on by default | 00:45 |
Wizzup | >Some rendering bugs in portrait mode remain, so the default desktop orientation is landscape for now (xrandr -o right). Please note that if orientation is changed (e.g. with xrandr -o normal) the ui will still be reading original key positions. | 00:45 |
Wizzup | not sure about this one | 00:45 |
Wizzup | >Screen brightness adjustments in UI do not work on the Pinephone yet. There is also a user reported when "brightness" UI setting is set to minimum (no visual change normally), the screen will stay black after reboot with WiFi set to ON with kill switch. However screen is normal after reboot if WiFi set to OFF with kill switch even minimum "brightness" UI setting. You may still change the screen | 00:46 |
Wizzup | brightness with terminal commands: xrandr --output DSI-1 --brightness 0.5 | 00:46 |
Wizzup | this is solved, pretty sure | 00:46 |
auanta | also update on the pinephone eMMC boot, yeah it didn't boot even after a new msdos partition table and dd'ing it to eMMC | 00:58 |
auanta | i'll say it 'booted' but remains a black screen with a yellow light | 00:59 |
auanta | something stopped it in the boot process | 00:59 |
auanta | hmm is there a way i can see the output? | 00:59 |
auanta | during bootup | 00:59 |
auanta | waiting on the others before i edit the wiki to remove the things Wizzup mentioned | 01:00 |
auanta | another power outage XD | 01:28 |
auanta | ! Default user | 02:51 |
auanta | Wizzup looks like the brightness settings still do not work. Also there are indeed some rendering issues in portrait mode, e.g. date and time time zone map. | 02:55 |
auanta | my virtual keyboard is indeed on. ok, checked everything out and just a few minor changes | 03:01 |
auanta | is your pinephone getting hot at all? my modem is so hot it's warping the plastic | 03:18 |
xmn | https://wiki.pine64.org/wiki/PinePhone_Thermal_Tweaks | 03:23 |
auanta | thanks! | 03:24 |
auanta | added page to category while i'm at it.. | 03:51 |
auanta | while unplugged, idle, screen off, the temp is 39292 | 04:25 |
auanta | while plugged in.. let me get a read on that | 04:25 |
auanta | i assume it's because wifi is on (modem kill switch is in the off position, but this didn't fix it) | 04:26 |
auanta | trip point temp 0 is 6000 | 04:27 |
auanta | trip point temp 1 is 8000 | 04:27 |
auanta | trip point temp 2 is 11000 | 04:27 |
auanta | 0 = passive, 1 = hot, 2 = critical | 04:28 |
auanta | ...hmmm | 04:30 |
auanta | missed a few zeros XD | 04:35 |
auanta | 60000, 80000, 110000 | 04:35 |
auanta | always run landscape mode! | 04:35 |
auanta | eh, it's below temp, i'll just keep an eye out.. | 04:37 |
tmlind | freemangordon: sorry fell asleep early again :) sounds like you got close to trying it out, not sure if running tinywl under x produces the same results as running it without h-d though | 06:25 |
freemangordon | tmlind: yeah, I just have to manage to compile mesa on d4 | 08:06 |
tmlind | freemangordon: heh ok that's at least an hour | 08:07 |
freemangordon | that's assuming the charge in battery can hold it :D | 08:09 |
tmlind | freemangordon: you may want to increase the usb charger current limit via sysfs to be higher than the default 500mA :) | 08:17 |
tmlind | echo 700000 > /sys/class/power_supply/usb/input_current_limit or higher depending on the charger | 08:18 |
buZz | this vendor -claims- to have new manufactured EB41 ; https://www.aliexpress.com/item/4001266035225.html | 08:56 |
tmlind | buZz: watch out for the extra capacity scam, they sell you a used eb41 battery relabeled with some bogus capacity.. | 09:20 |
buZz | tmlind: yeah havent hit that one yet, but i noticed some suspicious 'gold' EB41 with >2000mAh :P | 09:22 |
buZz | still a bit sad that capacity in that size didnt increase in the last 10y | 09:23 |
buZz | also, really hard to find a 3.8V lipo in 404565 size :P | 09:23 |
freemangordon | tmlind: already did, does not help | 09:25 |
buZz | whats the issue? powering off during compile? | 09:25 |
buZz | i built the kernel @ droid4 , several times , no issues :) | 09:25 |
freemangordon | ok, I set the limit to 2A, lets see | 09:30 |
freemangordon | charger is 2.4 | 09:30 |
freemangordon | POWER_SUPPLY_CURRENT_NOW=1438000 | 09:31 |
freemangordon | :) | 09:31 |
Wizzup | freemangordon: you can also build using -experimental | 09:36 |
Wizzup | that way it will build in ~10 mins | 09:36 |
Wizzup | jfyi | 09:36 |
buZz | freemangordon: battery not getting very toasty? :P | 09:36 |
buZz | i've had 45C on battery, d4 wasnt happy | 09:36 |
buZz | (but dont think it actually switched off) | 09:36 |
freemangordon | Wizzup: yes, bu I wouldn't be able to make code changes without rebuilding everything | 09:40 |
Wizzup | right | 09:40 |
freemangordon | when on device I can just call ninja | 09:40 |
freemangordon | buZz: batter tem 49 | 09:41 |
freemangordon | *temp | 09:41 |
buZz | :) | 09:41 |
buZz | and is it discharging from the load you're doing on it while charging at 1.4A ? | 09:41 |
freemangordon | tmlind: (or anybody) - do you have access to chromeos device? | 11:20 |
Wizzup | I have one at home I suppose, but not close to me | 11:26 |
freemangordon | ok, I have found what I need | 11:39 |
freemangordon | tmlind: | 12:43 |
freemangordon | 00:00:00.086 [render/egl.c:158] Supported DMA-BUF formats: AB30 AR24 AB24 XR24 XB24 RG16 GR88 R8 AR15 NV12 NV21 YU12 YV12 | 12:43 |
freemangordon | but, then it hits tinywl: types/wlr_linux_dmabuf_v1.c:511: feedback_compile: Assertion `table_len > 0' failed. | 12:43 |
freemangordon | any idae? | 12:43 |
Wizzup | uvos: bionic calls work fine too (wrt audio) | 12:54 |
Wizzup | uvos: I think sphone should also set remote_name if it's provided (for example when I get a SMS from 'takeaway', that is included somehow) | 14:15 |
Wizzup | currently it seems the line_identifier is Takeaway which is a little weird, I'd expect it to be the number if possible, and then have the remote_name set to takeaway | 14:15 |
Wizzup | also, I think it needs to check if the remote already exists | 14:16 |
Wizzup | nevermind the above statement | 14:16 |
Wizzup | in any case setting the remote_name would still be good | 14:16 |
Wizzup | dsc_: pushed a fix to conversations to render sphone-ofono inserted msgs | 14:21 |
Wizzup | uvos: pushed new code to wip-rtcom, this -should- address the things you asked about | 14:23 |
Wizzup | https://github.com/maemo-leste/sphone/compare/master...wip-rtcom | 14:24 |
Wizzup | it also fixes a bunch of typos | 14:24 |
freemangordon | tmlind: https://pastebin.com/MCQeRXZC | 14:54 |
freemangordon | I don't know if this is success or not | 14:54 |
freemangordon | tmlind: how to start that using DRM backend? | 15:06 |
freemangordon | wtf: | 15:10 |
freemangordon | libinput| Run-time dependency gtk4 found: NO (tried pkgconfig and cmake) | 15:10 |
freemangordon | libinput| Run-time dependency gtk+-3.0 found: NO (tried pkgconfig and cmake) | 15:10 |
freemangordon | do we really want that in mce? | 15:10 |
Wizzup | if it doesn't use those, that's fine, right? | 15:12 |
Wizzup | It is the "upstream way" | 15:13 |
freemangordon | well, I don;t know it it uses it, but it requires it to be build | 15:18 |
Wizzup | requires? | 15:18 |
Wizzup | it can't build without it and just skip it? | 15:18 |
freemangordon | yes, it errors out | 15:18 |
freemangordon | this is the current libinput master | 15:18 |
Wizzup | that's surprising since out CI doesn't even have gtk4 | 15:18 |
Wizzup | maybe you don't config it right | 15:19 |
freemangordon | this is not *our* libinput | 15:19 |
freemangordon | but upstream | 15:19 |
freemangordon | our is too old for wlroots libinput backend | 15:19 |
freemangordon | but that's not the point | 15:19 |
Wizzup | what is the point exactly? | 15:21 |
sicelo | freemangordon: start with drm? stop h-d and simply start tinywl in console | 15:35 |
Wizzup | so tinywl is supposed to work on X? when I google for it it says it's a wayland compositor | 15:35 |
sicelo | yes it's a wayland compositor | 15:36 |
Wizzup | why does it mention X then | 15:36 |
sicelo | you can start it under x if you want, which is what fmg did | 15:36 |
Wizzup | aha | 15:37 |
freemangordon | sicelo: does not start | 15:41 |
freemangordon | already tried | 15:41 |
freemangordon | sec, will provide logs | 15:41 |
freemangordon | sicelo: https://pastebin.com/idkuNdTA | 15:41 |
freemangordon | it seems it needs seatd | 15:43 |
sicelo | looks like you need to install and start seatd | 15:43 |
sicelo | yes :-) | 15:43 |
freemangordon | we don;t have such thinkg, afaik | 15:43 |
freemangordon | but ok | 15:43 |
sicelo | i can see it in devuan repos | 15:43 |
freemangordon | hmm | 15:44 |
freemangordon | what i sthe package name? | 15:44 |
sicelo | seatd | 15:44 |
freemangordon | "E: Unable to locate package seatd" | 15:44 |
sicelo | mmm | 15:44 |
freemangordon | not in lester repos, obviouslt | 15:44 |
sicelo | what about elogind? | 15:45 |
* sicelo doesn't have a leste device handy | 15:45 | |
Wizzup | elogind is blacklisted by us | 15:48 |
Wizzup | it does all kinds of crazy things we don't want | 15:48 |
freemangordon | ok, I have it built and started | 15:48 |
sicelo | so yes, fmg can install that temporarily | 15:48 |
sicelo | ah, sweet | 15:49 |
freemangordon | XDG_RUNTIME_DIR is invalid or not set in the environment | 15:49 |
freemangordon | sicelo: ^^^ | 15:49 |
sicelo | look at my paste from last night | 15:50 |
freemangordon | this is not set there | 15:50 |
sicelo | ah, set it to /run/user/$(id) | 15:51 |
freemangordon | literally? | 15:51 |
sicelo | mmm, no :-) | 15:51 |
sicelo | just the uid | 15:52 |
freemangordon | ah | 15:52 |
freemangordon | XDG_RUNTIME_DIR=/run/user/0 sudo ./build/tinywl/tinywl | 15:53 |
freemangordon | results in same error | 15:53 |
Wizzup | does sudo pass the env? | 15:54 |
freemangordon | shoult be | 15:54 |
freemangordon | but well | 15:54 |
sicelo | it's not good idea to start waylabd stuff as root (sudo) | 15:55 |
freemangordon | hmm | 15:55 |
sicelo | start it with regular user and his id | 15:55 |
freemangordon | but seatd does not start if not sude | 15:57 |
freemangordon | *sudo | 15:57 |
sicelo | mmm | 15:57 |
sicelo | seatd should be running in background, and yes, that'll need sudo | 15:57 |
freemangordon | ok, started | 15:57 |
freemangordon | :) | 15:57 |
sicelo | :-) | 15:57 |
freemangordon | well, screnn is black, I guess mce tunbed it off | 15:58 |
sicelo | i think you must ensure no maemo stuff runs ... | 15:58 |
freemangordon | https://pastebin.com/euGbjimP | 15:58 |
freemangordon | Xorg is stopped | 15:58 |
freemangordon | ok, what now :) | 15:59 |
freemangordon | what terminal you said? | 15:59 |
sicelo | looks good to me, but yeah, I'm not sure what you guys are testing, haha | 15:59 |
sicelo | use foot | 15:59 |
freemangordon | where to get that from? | 16:01 |
sicelo | foot? dhould be in repos too | 16:01 |
freemangordon | it is not :) | 16:01 |
sicelo | https://codeberg.org/dnkl/foot to build | 16:02 |
tmlind | freemangordon: i'm pretty sure i just export XDG_RUNTIME_DIR, and then the following | 16:11 |
tmlind | export WLR_BACKENDS=libinput,drm | 16:11 |
tmlind | export WLR_DIRECT_TTY=/dev/tty2 | 16:11 |
tmlind | export WLR_DRM_DEVICES=/dev/dri/card0 | 16:11 |
tmlind | i think that allows starting things like sway and probably tinywl from a ssh connection | 16:12 |
freemangordon | tmlind: I have it running | 16:12 |
freemangordon | I am trying to compule foot now | 16:12 |
tmlind | freemangordon: ok, did you get the errors starting it unless you patch gbm? | 16:13 |
freemangordon | umm... | 16:13 |
tmlind | foot is cool, just lacks touchscreen event support still :) | 16:13 |
freemangordon | I think gbm is already ok in leste repos | 16:14 |
freemangordon | I merged your commit back then | 16:14 |
tmlind | ah ok | 16:14 |
freemangordon | https://github.com/maemo-leste/sgx-ddk-um/commit/ce20a6afdd46d9d24a3588b71b20e395770f7e09 | 16:14 |
freemangordon | tmlind: see https://pastebin.com/euGbjimP | 16:15 |
freemangordon | foot segfaults | 16:16 |
freemangordon | ok, how to test that? | 16:16 |
freemangordon | sicelo: tmlind: ^^^? | 16:16 |
tmlind | hmm works for me | 16:17 |
freemangordon | I mean - what to run? | 16:17 |
freemangordon | tmlind: well, could be, but I justbuild it from source | 16:17 |
freemangordon | it spits "xkbcommon: ERROR: couldn't find a Compose file for locale "en_US.utf-8"" | 16:17 |
freemangordon | and then segfaults | 16:17 |
tmlind | well at least your wlroots log is using dri backend :) | 16:17 |
freemangordon | yes | 16:18 |
tmlind | hmm maybe you just need to export locale? | 16:19 |
freemangordon | LC_ALL=en_US.UTF-8 | 16:20 |
tmlind | ok, i guess no need for LANG=en_US.UTF-8 then | 16:20 |
freemangordon | https://pastebin.com/tbEA2ZTW | 16:21 |
freemangordon | no, adding LC_ALL=en_US.UTF-8 made it start | 16:21 |
tmlind | oh | 16:21 |
tmlind | freemangordon: so presumably if you revert the gbm patch i did, you'll get an error starting tinywl | 16:22 |
freemangordon | could be | 16:22 |
freemangordon | ok, how does foot knows which wl sever to use? | 16:22 |
freemangordon | *server | 16:22 |
tmlind | hmm or was it a mesa patch? i forgot.. | 16:23 |
freemangordon | WAYLAND_DISPLAY=wayland-0 does not help | 16:23 |
freemangordon | tmlind: maybe I can just hand you the mesa patch for you to test | 16:23 |
sicelo | maybe it's wayland-1 :-) | 16:24 |
tmlind | too many branches to patch :) | 16:24 |
freemangordon | sicelo: no, it says "failed to connect to wayland; no compositor running?" | 16:25 |
freemangordon | with wayland1 | 16:25 |
freemangordon | why it just exits? | 16:25 |
tmlind | freemangordon: for reference, so if you revert http://muru.com/linux/d4/0001-egl-dri2-Workaround-for-EGL_EXT_image_dma_buf_import.patch then you should see starting tinywm fail | 16:27 |
sicelo | you're starting it with something like, `./tinywl -s foot` ? | 16:28 |
tmlind | bbl | 16:28 |
freemangordon | sicelo: no | 16:30 |
freemangordon | but I will try | 16:30 |
freemangordon | same thing | 16:30 |
freemangordon | 'goodbye' | 16:30 |
freemangordon | tmlind: we don;t have that in mesa | 16:32 |
freemangordon | see https://github.com/maemo-leste-upstream-forks/mesa/commit/a673ae59b0edd281a20060080554f96331978aa3 | 16:32 |
freemangordon | tmlind: I implementd support for those missing functions | 16:32 |
tmlind | freemangordon: ok sounds like i need to test again then.. what's the commit implementing the missing functions? | 16:43 |
tmlind | food time, bbl | 16:44 |
freemangordon | there is still no commit | 16:46 |
freemangordon | as I want to see if I actually fixed it before doing such commit :) | 16:46 |
freemangordon | es2gears_wayland segfaults in some wayland code | 16:51 |
freemangordon | this is broken all over the place | 16:51 |
freemangordon | glmark2-es2-wayland does not start as well | 16:53 |
freemangordon | PVR:(Error): PVRDisplayBufferCreate: Failed to create display buffer (err=13) [0, ] | 16:53 |
josuah | I am just trying it now on a droid 4, maemo leste feels good! | 17:30 |
josuah | kudos and kudos and kudos guys! | 17:30 |
freemangordon | ok, no idea what has happened, but I have glmark2-es2-wayland running | 18:57 |
freemangordon | it seems that I have some old version of libdbm installed by hand | 18:59 |
freemangordon | reinstalling sgx-ddk-um-ti443x from the repos made it work | 19:00 |
dsc_ | Wizzup: sure thing | 19:02 |
freemangordon | hmm, why the hell glmakr was OOM killed? | 19:03 |
freemangordon | damit, this eats memory like crazy | 19:06 |
sicelo | what does? | 19:07 |
freemangordon | glmark | 19:07 |
freemangordon | hmm, actually I don't know | 19:15 |
freemangordon | looks like kernel allocations | 19:15 |
freemangordon | maybe BOs | 19:15 |
freemangordon | yeah, wlroots or glmark leak BOs | 19:18 |
sicelo | i hope it's not wlroots | 19:27 |
freemangordon | looks like no | 19:35 |
freemangordon | glmark leaks like hell when started with --fullscreen | 19:35 |
Wizzup | josuah: thanks!! | 19:36 |
Wizzup | josuah: there's more to come, telephony and such should be a reality soon | 19:36 |
tmlind | freemangordon: sounds like you're having fun with it | 19:43 |
freemangordon | tmlind: this renders like mad, until gets OOM killed :D | 19:44 |
freemangordon | I have pulled latest glmark2 master, to no user | 19:44 |
freemangordon | *use | 19:44 |
tmlind | ok, i recall the unpatched mesa or gbm had the crashing issue but was about twice as fast | 19:45 |
tmlind | well less patched mesa rather than unpatched | 19:46 |
freemangordon | ok, if I pass -s 538x960 it does not leak | 19:47 |
freemangordon | ok, lets see what score I will get | 19:47 |
freemangordon | tmlind: sicelo: https://pastebin.com/XwC1SV71 | 20:47 |
freemangordon | :D | 20:47 |
sicelo | wow, that's nice. i think score was in the 50s before? | 20:48 |
freemangordon | for x11 is ~87, iirc | 20:48 |
freemangordon | for drm is ~50, but it is vsynced | 20:49 |
freemangordon | now measuring xorg | 20:49 |
freemangordon | hmm, seems gpu-bound tests have higher fps on xorg | 20:51 |
freemangordon | ok, fps is slower, but I it is smoother | 20:52 |
freemangordon | maybe wl just drops lots of frames and xorg displays them | 20:52 |
freemangordon | glmark2 Score: 77 | 20:53 |
freemangordon | ~25% faster | 20:54 |
freemangordon | not bad | 20:54 |
freemangordon | anyway, lets push what I have done so far | 20:54 |
tmlind | nice, but hmm didn't we get the terrain test all the way up to 2 fps from 1 earlier on droid4 at least? | 20:55 |
freemangordon | I don;t think so | 20:55 |
freemangordon | well, maybe we get a bit lower fps because of fences actually work | 20:57 |
freemangordon | they didn;t back then and that could result in higher fps | 20:57 |
tmlind | just ran it here with glmark2-es2-wayland --size 960x540 --benchmark terrain, it produces 2 for me fyi | 21:00 |
freemangordon | 960x540 | 21:09 |
freemangordon | I guess it should be 540x960 | 21:09 |
tmlind | not full screen with 540x960, seems like 960x540 produces full screen | 21:13 |
tmlind | using landscape mode | 21:14 |
freemangordon | ah :) | 21:14 |
freemangordon | how to rotate? | 21:14 |
freemangordon | I am testing in portrait, that may explain the difference | 21:14 |
tmlind | yeah ok, i rotate with swaymsg output DSI-1 transform 90 | 21:15 |
freemangordon | I doubt I have swaymsg | 21:15 |
freemangordon | anyway | 21:15 |
tmlind | yeah.. wlr-randr should do the trick | 21:15 |
tmlind | sleep time here again, ttyl | 21:18 |
freemangordon | Wizzup: do you mind if I release new mesa in -devel? | 21:20 |
freemangordon | tmlind: fixed mease is in -devel, please test | 22:39 |
freemangordon | *mesa | 22:39 |
Wizzup | freemangordon: go for it | 23:39 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!