Wizzup | I think it does, if you enable it | 00:00 |
---|---|---|
Wizzup | it does on my laptop with kms | 00:00 |
Wizzup | in any case it looks like kexecboot gets stuck on a non existent file? | 00:00 |
uvos | nah its part of the fbdev code | 00:00 |
uvos | if it shows your useing fbdev for at least part of the boot | 00:00 |
uvos | Wizzup: i gues | 00:00 |
uvos | Wizzup: wierd | 00:00 |
Wizzup | fbdev emulation is on | 00:01 |
Wizzup | but that's about it | 00:01 |
uvos | hmm | 00:01 |
Wizzup | maybe it's time I dig up your specific bionic config | 00:06 |
Wizzup | (it boots to leste with dmse disabled just fine fwiw) | 00:06 |
uvos | /dev/ttyO2 dosent exist on d4 either | 00:07 |
uvos | so | 00:07 |
uvos | its wierd | 00:07 |
Wizzup | maybe it's there on serial too then | 00:07 |
Wizzup | and something else is blocking it from continuing | 00:07 |
uvos | yeah | 00:07 |
uvos | let me check | 00:07 |
uvos | nope | 00:11 |
uvos | dosent complain on bionic | 00:11 |
uvos | and /dev/ttyO2 only exists on the android kernel | 00:11 |
uvos | so im puzzeled | 00:11 |
uvos | ah | 00:15 |
uvos | Wizzup: | 00:15 |
uvos | if you exit kexecboot | 00:16 |
uvos | can't open /dev/ttyO2: No such file or directory repeats | 00:16 |
uvos | so kexecboot (the binary) is crashing | 00:16 |
uvos | this is tmlinds script that enables wifi in this case | 00:16 |
uvos | https://github.com/tmlind/buildroot/blob/9dd76697ccd2bac65a78fd7687899e1c9ca14465/configs/pandaboard_defconfig | 00:19 |
uvos | if you change that you should get a shell | 00:19 |
uvos | to ttyS2 that is | 00:20 |
uvos | password is cat /sys/board_properties/soc/die_id | sha256sum | head -c12 | 00:26 |
Wizzup | I did not enable wifi | 00:33 |
Wizzup | let me get a log | 00:33 |
uvos | Wizzup: kexecboot trys to enable wifi | 00:33 |
uvos | and getty on /dev/ttyO2 | 00:33 |
uvos | these things are android kernel specific | 00:33 |
uvos | but it dose that only after kexecboot exits | 00:33 |
uvos | which it should never doo | 00:33 |
uvos | unless you click exit on the menu | 00:34 |
uvos | or it crashes | 00:34 |
uvos | (otherwise this ends in kexec ofc) | 00:34 |
uvos | https://github.com/tmlind/buildroot/blob/droid4-kexecboot-2017.11/configs/droid4_kexecboot_defconfig | 00:35 |
uvos | the /dev/ttyO2 comes from there | 00:35 |
uvos | you can chang ttyO2 is the serial port on android | 00:35 |
Wizzup | hmm | 00:35 |
uvos | and then you get a shell | 00:35 |
uvos | if you change it to ttyS2 | 00:35 |
uvos | since thats the serial port on mainline | 00:36 |
uvos | ttyO2 is the port on android | 00:36 |
Wizzup | let me get a log | 00:36 |
Wizzup | https://dpaste.com/FKT9GTPNQ | 00:38 |
Wizzup | it doesn't mention drm at all (the log) | 00:40 |
uvos | Wizzup: you forgot init=/sbin/preinit.sh | 00:40 |
uvos | you silly person :P | 00:40 |
Wizzup | oh | 00:41 |
Wizzup | ok, kexecboot works | 00:45 |
uvos | yay :) | 00:45 |
uvos | great work | 00:45 |
Wizzup | at least the menu and everything | 00:46 |
uvos | fbcon=rotate:1 should help | 00:46 |
Wizzup | not sure if it can actually load the other leste kernel | 00:46 |
uvos | did you enable kexec in the mainline kernel? | 00:46 |
uvos | as built in | 00:46 |
Wizzup | it did load something, it just didn't get far | 00:47 |
uvos | ok | 00:47 |
uvos | gn8 | 00:47 |
uvos | progress :) | 00:47 |
Wizzup | https://dpaste.com/BF5UFF5XN | 00:48 |
Wizzup | gn | 00:48 |
uvos | thats it? | 00:48 |
uvos | hmm | 00:48 |
uvos | ok | 00:48 |
Wizzup | [ 0.000000] OF: fdt: Machine model: Motorola Droid Bionic XT875 | 00:49 |
Wizzup | that could be a problem :) | 00:49 |
Wizzup | probably need to check the dts in /boot | 00:49 |
Wizzup | # cat boot.cfg.leste | 00:50 |
Wizzup | LABEL=Maemo Leste | 00:50 |
Wizzup | PRIORITY=8 | 00:50 |
Wizzup | DTB=/boot/omap4-droid-bionic-xt875.dtb | 00:50 |
Wizzup | yeah | 00:50 |
Wizzup | yay | 00:52 |
Wizzup | ok, tomorrow the upstreaming will begin | 00:53 |
Wizzup | thanks for all the help | 00:54 |
Wizzup | uvos: for the record my kernel config is omap2plus + m -> y + olddefconfig, and then some dvb stuff removed | 00:56 |
sicelo | (yeah the dvb stuff is annoying) | 01:06 |
Wizzup | uvos: hehehe the d3 doesn't have the green light always on when it's powered (different led), so the mce patterns actually work properly | 11:19 |
Wizzup | I have a nice orange light now | 11:19 |
uvos | yeah i know it has a white led next to the usb port like every droid < 4 | 11:21 |
Wizzup | yeah, but it's nice :p | 11:21 |
uvos | this is supierior to the n900/d4 anyways becasue if you see a light on the device its allways a notification | 11:21 |
Wizzup | ok so what I need to do is check why the brightness adjustment doesn't work and the modem doesn't (seem) to work, and then it's looking pretty good | 11:21 |
uvos | ok | 11:21 |
bencoh | right, I wanted to ask ... that green light on d4 is fully hw? | 11:21 |
Wizzup | bencoh: it can disabled in cpcap | 11:22 |
bencoh | ah, alright | 11:22 |
Wizzup | like with the n900, which has a gpio for it | 11:22 |
uvos | its a seperate led controled by cpcap | 11:22 |
Wizzup | we disable that in our n900 init script | 11:22 |
uvos | maybe the fw | 11:22 |
uvos | so might not be totaly hw | 11:22 |
uvos | the brightness is likely on a different channel | 11:23 |
Wizzup | yeah, I didn't look at the dts part of that yet | 11:23 |
Wizzup | going to test a bit more, but it really -looks- like the freezes are gone | 11:24 |
Wizzup | which makes me happy | 11:24 |
uvos | that would be wierd | 11:24 |
uvos | why would the double kexec change that | 11:24 |
Wizzup | kexecboot might set up the clocks just right or something | 11:24 |
Wizzup | all I could think of | 11:24 |
Wizzup | ah yeah and there's accelerometer stuff I think | 11:27 |
uvos | i2cdetect -r 1 0x38 0x38 | 11:28 |
uvos | what dose that show on d3 | 11:29 |
uvos | i2cdetect -r 1 0x38 0x38 | 11:29 |
uvos | *i2cdetect -r 0 0x38 0x38 | 11:29 |
Wizzup | do you know what package that is part of? | 11:30 |
Wizzup | (searching) | 11:30 |
tmlind | i2c-tools | 11:30 |
Wizzup | ty | 11:30 |
Wizzup | https://pastebin.com/raw/aZw1W8rs | 11:31 |
uvos | ok | 11:31 |
uvos | the led chip is the same | 11:31 |
uvos | and present | 11:31 |
uvos | different channel then | 11:31 |
uvos | i2cdetect -r 3 0x18 0x18? | 11:32 |
Wizzup | https://dpaste.com/4Y2RPLYP4 | 11:32 |
tmlind | freemangordon: fyi, still slowly setting up pvr-drv with reg read/write access for ocp regs, and interrupt handler for the socs that use it. still need to move the interrupt handler(s).. | 11:33 |
tmlind | i guess the next logical step would be to move the mmu code to a new file pvr-mmu.c or something | 11:33 |
uvos | Wizzup: ok no accel | 11:33 |
uvos | i2cdetect -r 3 0x0f 0x0f shows UU probubly | 11:33 |
uvos | ie it has the same accel as bionic | 11:34 |
* tmlind bbl | 11:34 | |
Wizzup | uvos: yes | 11:34 |
Wizzup | so I'll take that from the bionic dts? | 11:35 |
uvos | yeah | 11:35 |
Wizzup | how about the led controller? | 11:35 |
uvos | https://www.ti.com/lit/ds/symlink/lm3532.pdf | 11:35 |
uvos | use i2cset to mess with the channels until you find what is connected to the backlight | 11:36 |
uvos | ie 0x16 0x18 or 0x1A | 11:37 |
uvos | carefull tho | 11:37 |
uvos | one might be the flash led | 11:37 |
uvos | you can kill that fast by having it on at full power to long | 11:37 |
Wizzup | maybe it's easier to check android? | 11:38 |
uvos | thats boring :P | 11:38 |
uvos | not sure about easier but possible yeah | 11:38 |
Wizzup | ok | 11:43 |
Wizzup | I will try in a bit, but def don't want to damage the led ;) | 11:43 |
sicelo | Which reminds me that I was going to work on the N900 flash LED | 11:46 |
sicelo | Everytime i see the rgb led on my n900 it's due to a notification. If that's not always the case for you, something's borked on your device | 11:49 |
uvos | ? the n900 uses the leds for lots of stuff that isent a notfication, like the silly power down animation and the charging animation etc | 11:51 |
uvos | the old motorola devices have a dedicated led for system state and one for only comunication notifications | 11:51 |
uvos | making its more clear whats happening | 11:51 |
Wizzup | sicelo: what do you mean, due to a notification? | 11:53 |
bencoh | uvos: actually the fact that they both sit in the same area defeats the point of having two :) | 11:57 |
uvos | bencoh: on d3 and lower they dont | 11:57 |
Wizzup | uvos: hm wait, I already had the bionic entry in my dts | 11:57 |
uvos | bencoh: one is where the led on d4 is and one is a big white one next to the usb port | 11:57 |
uvos | on d4 its defeated yeah | 11:57 |
uvos | Wizzup: well dose it work? | 11:58 |
Wizzup | it doens't do anything when I rotate the device | 11:58 |
Wizzup | otherwise I would not have asked ;) | 11:58 |
uvos | check the number in sysfs | 11:58 |
uvos | it might just be oriented differently | 11:59 |
uvos | maybe it rotates when you place the device face down or something :P | 11:59 |
Wizzup | ok | 11:59 |
uvos | what device image is your rootfs based on? | 11:59 |
Wizzup | holding it upside down makes it rotate | 12:00 |
Wizzup | upside down in portrait | 12:00 |
uvos | ok | 12:00 |
uvos | thats easily fixable | 12:00 |
Wizzup | but then holding it normally keeps it there | 12:00 |
Wizzup | rootfs is bionic | 12:00 |
uvos | Wizzup: ofc it keeps it there | 12:01 |
Wizzup | mhm | 12:01 |
uvos | you need to rotate it landscape upside down | 12:01 |
Wizzup | yeah, lol | 12:01 |
uvos | https://github.com/maemo-leste/leste-config/blob/master/leste-config-bionic/lib/udev/rules.d/50-iio-sensors.rules.leste | 12:01 |
uvos | edit this | 12:01 |
Wizzup | ok | 12:01 |
uvos | you need to rotate the matrix 180deg | 12:01 |
uvos | around z | 12:01 |
uvos | https://en.wikipedia.org/wiki/Rotation_matrix | 12:02 |
* Wizzup did linalg at uni | 12:02 | |
uvos | great | 12:02 |
bencoh | linalg <3 | 12:03 |
Wizzup | :) | 12:03 |
Wizzup | I'll look at this in a bit, I will also need to make the leste-config thing then | 12:03 |
Wizzup | uvos: where shall I push the 5.10.y kernel for kexecboot, droid4-linux ? | 12:04 |
uvos | id make a sperate clownboot kernel repo | 12:04 |
uvos | maybe we should gather all of this in one git project | 12:04 |
uvos | (icl the android kernel kexec modules | 12:05 |
uvos | ) | 12:05 |
uvos | *github project | 12:05 |
Wizzup | atm those will go into a branch of the bionic clown boot, but yeah | 12:05 |
uvos | i mean the sources for those | 12:05 |
Wizzup | right | 12:05 |
Wizzup | yeah I suppose we could do that, an extra orga is maybe a bit overkill but we could | 12:05 |
Wizzup | no real opinion(s) | 12:05 |
Wizzup | what do you prefer? | 12:09 |
uvos | idk just make a repo with the kernel that incls the needed patches and defconfig | 12:09 |
uvos | we can pull it into a orga later if need be | 12:09 |
Wizzup | ok | 12:10 |
Wizzup | brb | 12:11 |
lel | MerlijnWajer created a repository: https://github.com/maemo-leste/clown-boot-kernel | 12:39 |
lel | MerlijnWajer created a repository: https://github.com/maemo-leste/clown-boot-kexec | 12:40 |
lel | MerlijnWajer created a repository: https://github.com/maemo-leste/clown-boot | 12:40 |
Wizzup | uvos: ^ something like this? | 12:40 |
Wizzup | clown-boot-kernel for linux stable + patches, clown-boot-kexec for android kexec modules and other android specific stuff, clown-boot for the clown boot stuff | 12:41 |
Wizzup | also, I'd like to host the android images somewhere, I guess we can do maedevu | 12:41 |
Wizzup | it's silly for people to have to search for the android image they want to flash only to load clown boot using it | 12:41 |
Wizzup | uvos: yeah ok freeze is still there | 12:46 |
Wizzup | to be continued | 12:46 |
freemangordon | tmlind: nice | 13:29 |
freemangordon | tmlind: do you have any idea which fd shall I use for mmap so PVRMMap() to be called? | 13:30 |
freemangordon | I tried BO fd and /dev/dric/card1 fd, ni dice | 13:30 |
freemangordon | *no | 13:30 |
buZz | someone is hinting that maemo leste should promote compatibility with https://www.engadget.com/fisher-price-working-chatter-telephone-040026739.html | 13:35 |
buZz | as a meme/commercial/funnonsense i think it could get more eyes on us | 13:35 |
buZz | its just a BT headset so should work totally fine i guess | 13:36 |
dreamer | I think parazyd can guess who it was :P | 13:37 |
buZz | :) | 13:37 |
tmlind | freemangordon: not sure.. i think /dev/dri/card0 currently needs to be used at least for all allocations.. but i'm lost there how it's working | 13:48 |
freemangordon | this is omapdrm, right? | 13:48 |
freemangordon | or pvr | 13:48 |
buZz | parazyd: you wanna meme that fisherprice thing up? :D | 14:10 |
tmlind | freemangordon: at least weston-simple-dmabuf-egl --drm-render-node=/dev/dri/card0 works | 14:23 |
* tmlind on gsm connection with d4 | 14:24 | |
tmlind | yeah that's the omapdrm device, you can use drm_info to see what's available | 14:28 |
freemangordon | tmlind: try glmark-es2-drm | 14:50 |
freemangordon | the fuck? now works | 14:50 |
freemangordon | ah, on the second try device hung | 14:52 |
tmlind | can't try until next week or so | 14:55 |
freemangordon | ok | 14:55 |
uvos | Wizzup: sure idk if it belongs into the leste ogra but ok | 18:53 |
uvos | Wizzup: please fork the clownboot repo then | 18:54 |
uvos | and give me ownership | 18:54 |
Wizzup | uvos: ok | 19:02 |
lel | MerlijnWajer deleted a repository: https://github.com/maemo-leste/clown-boot | 19:03 |
lel | MerlijnWajer created a repository: https://github.com/maemo-leste/bionic-clown-boot | 19:04 |
lel | MerlijnWajer renamed a repository: https://github.com/maemo-leste/clown-boot | 19:04 |
lel | MerlijnWajer closed a pull request: https://github.com/maemo-leste/maemo-input-sounds/pull/1 (remove gconf usage by adding support for the touchscreen.vibration.enabled profiled key) | 19:53 |
Wizzup | uvos: I think all are built now | 20:21 |
Wizzup | well m-i-s is still ongoing | 20:22 |
uvos | Wizzup: in order to now change ts vibration | 20:27 |
uvos | you need to also do a release of profilesx | 20:27 |
uvos | or wait until i do it | 20:27 |
Wizzup | I can wait for you to do it | 20:28 |
uvos | ok | 20:28 |
uvos | https://github.com/maemo-leste/sphone/commit/1f50ad1092499c937dfed6bdafba5574f954e4ca sphone modularization :) | 20:28 |
Wizzup | neat! | 20:29 |
uvos | once you get around to geting ohmd to work writing a module like this should allow switching between ohmd/ucm only at runtime https://github.com/maemo-leste/sphone/blob/master/src/modules/route-pulseaudio.c | 20:29 |
Wizzup | yeah, ohm is next. I will do the clown boot thing you suggested, then the news post, then the audio | 20:31 |
uvos | Wizzup: you dident merge into beowulf-devel for mis | 22:05 |
uvos | Wizzup: and other i think | 22:05 |
uvos | Wizzup: so you just rebuilt the same pacakge | 22:06 |
Wizzup | apart from mis, where did I do that? | 22:09 |
Wizzup | rerunning mis | 22:10 |
uvos | idk but phoenix-bot not showing changes for imple-brightness-applet-repos is suspicous | 22:11 |
uvos | same with profiled | 22:12 |
uvos | profiled is fine tho | 22:12 |
uvos | no idea why phoneix-bot dident show changes | 22:13 |
Wizzup | phoenix-bot is not to be trusted | 22:18 |
Wizzup | it randomly doesn't output | 22:18 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!