| *** olesalscheider_ is now known as olesalscheider | 05:41 | |
| *** zama_ is now known as zama | 05:50 | |
| *** ahjolinna_ is now known as ahjolinna | 09:03 | |
| T4 | <adampigg> @eugenio_g7 if youget a chance to look over the bt/fw PR, i will build anothr test image | 09:41 |
|---|---|---|
| T4 | <eugenio_g7> @adampigg merged, and I just opened another PR to merge the dhv changes: https://github.com/sailfish-on-latte/droid-hal-version-latte/pull/1 | 11:17 |
| T4 | <adampigg> Cool, will look.shortly | 11:25 |
| piggz | @eugenio_g7 need to make a wee fix to d-h-v | 12:12 |
| heroic_1 | where can I find the sources for this build? http://images.devaamo.fi/sfe/fp2-sibon/beta4/ I think "mal" built it | 12:20 |
| vknecht | heroic_1, https://build.merproject.org/project/show/nemo:testing:hw:fairphone:fp2-sibon I guess | 12:26 |
| heroic_1 | ah, I was only checking the mer-hybris github repos... thanks you! | 12:26 |
| mal | piggz: nitpicking again, the commit message format in the dhv PR is not following the example set by the previous commits :) | 12:43 |
| piggz | !"£$^£$$£$%^%$£$££$ | 12:44 |
| merbot | piggz: Error: No closing quotation | 12:44 |
| mal | piggz: :D | 12:44 |
| mal | piggz: sorry | 12:44 |
| piggz | im doing too much at one .. d-h-v, d-h-v-l, and sorting out my daughetrs car insurance! | 12:45 |
| T4 | <eugenio_g7> :D | 13:01 |
| piggz | mal: happy with adding [dhv] at the start? https://github.com/mer-hybris/droid-hal-version/pull/14 | 13:18 |
| piggz | @eugenio_g7 https://github.com/sailfish-on-latte/droid-hal-version-latte/pull/2 | 13:19 |
| piggz | make sure it is ok, i had to merge etc to get it to commit | 13:20 |
| mal | piggz: much better thanks :) | 13:27 |
| *** kernling is now known as krnlyng | 14:19 | |
| T4 | <eugenio_g7> piggz: although I generally avoid doing that on the master branches, I think that we could rewrite the git history in this specific case rather that further polluting it by merging this PR | 16:19 |
| T4 | <adampigg> @eugenio_g7 feel free! | 16:20 |
| T4 | <eugenio_g7> great :) | 16:20 |
| T4 | <eugenio_g7> @adampigg: https://github.com/sailfish-on-latte/droid-hal-version-latte/commit/f16016fa3ccd9b5f7d5d8362aa221664db1255fc also fixed up the dhv commit reference as it didn't exist anymore | 16:29 |
| piggz | @eugenio_g7 hmm, somehting is still wrong... | 16:37 |
| piggz | its looking for the vibrator packages.. | 16:37 |
| piggz | the droid one... | 16:38 |
| piggz | but, afaict, becuase have_vibrator is 0, it shouldnt? | 16:38 |
| mal | piggz: what is looking those vibra packages? | 16:40 |
| piggz | mal: droid-hal-version-latte | 16:41 |
| mal | piggz: try removing the whole have_vibrator define from your spec | 16:42 |
| piggz | mal: yes, that works, guess that makes the whole 0/1 thing a bit irrelevant! | 16:50 |
| piggz | mal: actually, it passed the unresolvable stage, into failing stage https://build.merproject.org/package/live_build_log/nemo:devel:hw:xiaomi:latte/droid-hal-version-latte/sailfish_latest_i486/i586 | 16:53 |
| piggz | mal: hang on, i thnk i see it | 17:06 |
| piggz | mal: @eugenio_g7: yes, thats worked | 17:09 |
| piggz | @eugenio_g7 feel free to flash https://gitlab.com/sailfishos-porters-ci/latte-ci/-/jobs !! | 17:38 |
| heroic_1 | kagura looking good: https://sx.ix5.org/logs/smooth-sailing.jpg | 18:48 |
| heroic_1 | now to get touch up... | 18:48 |
| heroic_1 | is there a way to ssh in? I have no adb sadly | 18:49 |
| piggz | huh, the redmi note 5 is the same soc as the note 4 | 19:18 |
| TheKit | heroic_1, what about telnet according to HADK? | 19:23 |
| TheKit | adb is not used with Sailfish | 19:23 |
| heroic_1 | yup, that looks good to me, but how do I set up the init file? | 19:23 |
| TheKit | for what? | 19:24 |
| TheKit | it should be already enabled when building community port according to HADK | 19:24 |
| heroic_1 | "The hybris-boot repository contains such an initrd with convenient USB networking" | 19:24 |
| heroic_1 | AHH nvm | 19:24 |
| heroic_1 | but do I only have it in recovery or also in regular boot? | 19:25 |
| heroic_1 | no, apparently not. I can't telnet into the device, only when in recovery | 19:27 |
| TheKit | after OS is booted, it should be setup by /init script in rootfs | 19:31 |
| TheKit | does ssh ip work? | 19:31 |
| TheKit | and make sure you try telnet 192.168.2.15 2323, not 23 after switching to roofs | 19:32 |
| TheKit | *rootfs | 19:32 |
| heroic_1 | seems I don't have telnet, this is what dmesg tells me: https://del.dog/amuguzicoc.cs | 19:36 |
| heroic_1 | do I need to copy "hybris-boot/init-script" to somewhere? | 19:37 |
| TheKit | heroic_1, you mean you do not have network device? | 20:09 |
| TheKit | usb0 or similar on host | 20:09 |
| heroic_1 | I see no new network device | 20:11 |
| heroic_1 | do you mean I could have ssh up? | 20:12 |
| heroic_1 | Ideally I'd like to have telnet up at least like in the recovery | 20:14 |
| heroic_1 | Just noticed I have "qt5-plugin-generic-evdev" in my "jolla-hw-adaptation-kagura.yaml", but somehow it didn't get installed to my target. Odd. Let's see if I can get touch up now | 20:14 |
| TheKit | missing touch is probably due to wrong /dev/input/eventx device specified in environment file | 20:16 |
| heroic_1 | should I hard-code that? | 20:19 |
| heroic_1 | Because the nile devices don't do it. And for me, the input devices seem to get shuffled at every boot | 20:19 |
| mal | heroic_1: is that android 8 based port? | 20:19 |
| heroic_1 | yes, android 8 based | 20:19 |
| heroic_1 | Right now I'm pulling the systemd logs upon returning to recovery, but that doesn't make for a very brisk development pace | 20:20 |
| mal | heroic_1: some android 8 devices have issues with usb network, try setting CONFIG_RNDIS_IPA=n, CONFIG_USB_CONFIGFS_QCRNDIS=n, CONFIG_USB_CONFIGFS_RNDIS=y and CONFIG_USB_CONFIGFS_RMNET_BAM=n in kernel defconfig and rebuild kernel using make hybris-boot and then fastboot flash the new hybris-boot.img from out/ | 20:22 |
| heroic_1 | I do have telnet in recovery, are you sure that that's the issue? | 20:22 |
| mal | heroic_1: if you don't have telnet in normal boot then yes | 20:22 |
| mal | I have had that | 20:23 |
| heroic_1 | let's see https://github.com/marina-bay/android_kernel_sony_msm/tree/sfos-3.0.1.14-4.4 | 20:23 |
| heroic_1 | yeah, the RNDIS stuff is enabled | 20:24 |
| heroic_1 | ok, will try without that stuff | 20:25 |
| heroic_1 | TheKit: do I need to set "evdevtouch:/dev/touchscreen" in LIPSTICK_OPTIONS ? | 20:26 |
| heroic_1 | nile seems to be doing it without: "LIPSTICK_OPTIONS=-plugin evdevtouch -plugin evdevkeyboard:keymap=/usr/share/qt5/keymaps/droid.qmap" | 20:26 |
| TheKit | earlier it used to be /dev/input/eventx (where x is 0..n), I suppose /dev/touchscreen symlink should be created by udev rule | 20:28 |
| heroic_1 | I see there is a package that does that https://github.com/mer-hybris/evcap | 20:28 |
| heroic_1 | But it hasn't been updated since 2015, and I don't see jolla using it anywhere | 20:28 |
| heroic_1 | hmm, now I get "connection refused" at least | 20:30 |
| heroic_1 | ugh, and my display is dead.. how did I manage that? -_- | 20:31 |
| heroic_1 | https://del.dog/udazotitag.coffeescript some progress | 20:32 |
| heroic_1 | it's stuck on the "setting up" stage I guess: "iSerial 3 Mer Debug telnet on port 23 on usb0 192.168.2.15 - also running udhcpd" | 20:33 |
| *** OhYash1 is now known as ohyash | 20:41 | |
| heroic_1 | Hmm, I think I know why it fails to boot. I made the kernel with "make hybris-boot" instead of going through "droid-hal-kagura-img-boot" | 20:49 |
| T4 | <elros34> don't add :/dev/touchscreen so touchscreen node will be auto detected | 20:57 |
| T4 | <elros34> or /dev/input/event* | 20:58 |
| heroic_1 | alright, that's what I have now. But no touch | 20:58 |
| heroic_1 | Alright, I get the issue now. When I compile with "droid-hal-img-boot-kagura", I get the "jolla-init" script as init https://github.com/mer-hybris/hybris-initrd/blob/master/jolla-init | 21:50 |
| heroic_1 | But when I compile with "make droid-boot", I get the debug "init-script" from hybris-boot: https://github.com/mer-hybris/hybris-boot/blob/master/init-script | 21:51 |
| heroic_1 | that's why it seemed like I could get telnet up when compiling from within the android tree, but not when compiling from the SDK | 21:51 |
| mal | I didn't know you are not using the normal hadk way to build | 21:56 |
| heroic_1 | Yeah, it's all a bit wild with sony stuff :) | 22:02 |
| heroic_1 | But I think I got it now | 22:02 |
| heroic_1 | Added the gadgetfs functions and mangled a bit with the "jolla-init" file, will see if I get telnet now | 22:03 |
| heroic_1 | btw, the helpers from https://github.com/mer-hybris/droid-hal-img-boot-sony-nile and https://github.com/mer-hybris/hybris-initrd/ do make a lot of sense | 22:04 |
| r0kk3rz | tbh it would be easier to do normal hadk way first, then switch to productised way after | 22:12 |
| mal | nothing prevents using sonys with hadk way either | 22:12 |
| r0kk3rz | no thats whe way i did my tablet | 22:19 |
| heroic_1 | hadk way doesn't boot for me | 22:24 |
| heroic_1 | or at least with the current config, would be quite a lot of work to switch now, only to switch back later | 22:24 |
| r0kk3rz | its fine if you know what you're doing | 22:27 |
| heroic_1 | let's say I'm getting there... | 22:28 |
| r0kk3rz | heh, yes its a learning curve | 22:28 |
| r0kk3rz | but if you do things the usual way then we can give you the usual answers | 22:28 |
| heroic_1 | heads up: old init script uses /bin/busybox-static, but the jolla initramfs has /sbin/busybox-static | 22:29 |
| heroic_1 | r0kk3rz: sure, but I figure it would be easier to start with a working system | 22:29 |
| r0kk3rz | ish | 22:30 |
| r0kk3rz | if it was working that well, you wouldnt be talking to us :P | 22:31 |
| r0kk3rz | anyway, sounds like you've gotten fairly far already | 22:32 |
| heroic_1 | "-ish", hehe yeah :) | 22:38 |
| heroic_1 | this is my current boot log: https://del.dog/anigetivor.sql | 23:13 |
| heroic_1 | and this is the modified init I am using: https://del.dog/ubumifosiy.sh | 23:14 |
| heroic_1 | It boots fine, but I can see the rndis interfaces disappearing in dmesg once it has reached the "switch-root" stage | 23:14 |
| mal | heroic_1: so the kernel change didn't help? | 23:15 |
| heroic_1 | It seems to me there is no way to preserve telnet functionality throughout the root switch, only to set it up again afterwards | 23:15 |
| heroic_1 | I changed the kernel, but it did not help | 23:15 |
| heroic_1 | I see that the rootfs has an "init-debug" file, is that ever executed? Looks just like the recovery telnet setup | 23:16 |
| mal | just curious try masking usb-moded and see what happens | 23:17 |
| mal | hmm, actually yes, the change won't work all the way, usb-moded has a limitations currently in configfs support | 23:18 |
| heroic_1 | Oh I just noticed something stupid. I think I am missing sailfish-porter-tools | 23:18 |
| mal | does your rootfs have a file /init_disable_telnet? porter tools does provide busybox-static it seems | 23:20 |
| heroic_1 | rootfs as in the ramdisk? then no | 23:21 |
| mal | no, sfos rootfs | 23:22 |
| heroic_1 | sfos rootfs as well | 23:22 |
| mal | heroic_1: how far does the device boot now? | 23:22 |
| heroic_1 | all the way | 23:22 |
| heroic_1 | to the welcome screen that is | 23:22 |
| mal | ok, did you fix touchscreen already? | 23:22 |
| heroic_1 | no, sadly not | 23:23 |
| heroic_1 | I am hunting down what it could be. I thought maybe some weirdness because nile uses SDE/sdm and tone/kagura is still stuck on msmfb, but then I remembered that nile only got sde on pie | 23:23 |
| heroic_1 | the "qt5-plugin-generic-evdev" is supposed to auto-detect the right touchscreen device, right? | 23:24 |
| mal | how did you try to fix touch? | 23:24 |
| heroic_1 | I don't think it's kernel since my aosp builds work just fine | 23:24 |
| heroic_1 | Currently trying to telnet in, *in order to* start exploring /dev/input | 23:25 |
| heroic_1 | what common pitfalls are there with touchscreen? the hadk-faq recommends to try using evdev instead of evdevtouch, is that a good idea? | 23:26 |
| mal | how does your /var/lib/environment/compositor/droid-hal-*.conf file look like | 23:26 |
| heroic_1 | https://github.com/marina-bay/droid-config-sony-tone/blob/sfos-3.0.1.14-android-8.1/sparse/var/lib/environment/compositor/droid-hal-device-kagura.conf | 23:28 |
| heroic_1 | kanged straight from nile. "physical" properties are only for dpi scaling, right? | 23:28 |
| mal | ok, not sure which sfos version brings the touchscreen autodetect | 23:29 |
| heroic_1 | I am porting 3.0.1.11(latest) | 23:29 |
| mal | https://github.com/mlehtima/droid-config-fp2-sibon/blob/master/sparse/var/lib/environment/compositor/droid-hal-device.conf#L9 try defining the device manually, it's usually 0, 1 or 2 | 23:30 |
| heroic_1 | I have saved kernel logs from 4.4 on aosp, and it seems the device jumps per-boot | 23:31 |
| heroic_1 | but I will try manually, maybe in sfos it stays stable | 23:32 |
| mal | hmm, well then there is the udev rule way | 23:32 |
| mal | https://github.com/mer-hybris/droid-config-sony-nile/commit/25fbfb35f9b76418420f3e8f0a1a78d522625f1c | 23:33 |
| heroic_1 | just revert that one? | 23:33 |
| mal | that was how it was dropped from nile so add something similar back | 23:33 |
| mal | depends on what the name of the touchscreen is on your device | 23:34 |
| mal | I think you should be able to see that in recovery | 23:34 |
| heroic_1 | how? now I'm curious | 23:34 |
| heroic_1 | wut? apparently my sfos rootfs doesn't have /var/lib/udev at all... | 23:35 |
| mal | are you sure you are checking correct place? | 23:36 |
| mal | and for the name check /sys/class/input/input0/name etc | 23:36 |
| mal | depending on which event device is correct | 23:37 |
| heroic_1 | lil heart attack :) was checking /var/lib instead of /lib/udev | 23:38 |
| heroic_1 | WE HAVE TOUCH!!!! | 23:43 |
| heroic_1 | Thanks a million mal! | 23:43 |
| heroic_1 | wifi as well... | 23:45 |
| heroic_1 | sound too | 23:45 |
| heroic_1 | wifi won't accept the password though. will fix later | 23:46 |
| mal | nice | 23:47 |
| heroic_1 | heh, the tutorial is kinda cute | 23:48 |
| heroic_1 | haptics up as well | 23:49 |
| mal | it can be skipped by touching screen corners clockwise starting from top left corner, before starting the tutorial | 23:49 |
| mal | just in case you need to reinstall at some poit | 23:50 |
| mal | *point | 23:50 |
| heroic_1 | nah, I'm new to it, I really dig it :) but thx | 23:50 |
| mal | heroic_1: once you have reinstalled many times it becomes annoying :) | 23:50 |
| heroic_1 | CAMERA up | 23:50 |
| heroic_1 | let's check ril | 23:51 |
| heroic_1 | torch up | 23:51 |
| mal | maybe check that the camera configs are correct, there could be different resolutions | 23:51 |
| mal | installing droid-camres from here can be used https://build.merproject.org/project/show/nemo:devel:hw:common | 23:52 |
| heroic_1 | it's stuttery as well | 23:52 |
| heroic_1 | but we gotta fix that on our side, our camera hal is garbage anyway | 23:53 |
| heroic_1 | our side meaning sony open devices | 23:53 |
| heroic_1 | then we can check which resolutions we enable for sailfish | 23:53 |
| heroic_1 | high framerate stuff is only in testing on pie as of now | 23:53 |
| mal | which of the tone devices are you testing on? | 23:54 |
| heroic_1 | can already see the thermals need some love | 23:54 |
| heroic_1 | for tone, only kagura | 23:54 |
| heroic_1 | I have some friends with dora, and we had someone with the elusive keyaki | 23:54 |
| mal | ok, good progress anyway | 23:54 |
| heroic_1 | how do I change camera resolution in the ui? i don't see any pulley | 23:55 |
| mal | in settings->apps->camera | 23:55 |
| mal | but it only has one resolution for each aspect ratio | 23:56 |
| mal | which are defined in https://github.com/marina-bay/droid-config-sony-tone/blob/sfos-3.0.1.14-android-8.1/sparse/etc/dconf/db/vendor.d/jolla-camera-hw.txt | 23:56 |
| heroic_1 | no effing way... we got working ril and calls with audio | 23:57 |
| mal | :) | 23:58 |
| heroic_1 | camera resolutions are 20mp and 13mp | 23:58 |
| heroic_1 | front with 4 and 8, selfie cam confirmed working | 23:58 |
| heroic_1 | hmm, seems no mobile data though | 23:59 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
