00:07:20 Quit: andyrtr (Ping timeout: 250 seconds) 00:14:19 andyrtr joined the channel 00:21:35 Quit: jreznik_ (Ping timeout: 246 seconds) 00:34:23 anYc | not yet, I didn't have much time today to work on it. First, I want to avoid that I made a stupid mistake :) 01:00:03 anYc | kido: out of curiosity, why do you disable CPU cores on tetra in the android-init.rc? (https://github.com/AsteroidOS/meta-tetra-hybris/commit/5716fe0680671002087a30a546efe5ed828ab111) 01:07:02 kido | battery life 01:07:10 kido | android wear does the same thing 01:07:42 anYc | okay, so, are they enabled on demand or do they stay off? 01:37:15 kido | Stay off 01:38:35 anYc | interesting, thanks! 01:40:58 anYc | btw, I noticed that libhwcomposer.so isn't loaded in the lipstick process although "EGL_PLATFORM=hwcomposer" and "-platform hwcomposer" are set. I guess it should though? I will investigate this tomorrow. good night and thanks again! 01:47:13 no joined the channel 01:47:18 Quit: no (Client Quit) 02:02:30 maker2222 joined the channel 02:13:27 SailorMoon joined the channel 02:55:19 Quit: andyrtr (Ping timeout: 258 seconds) 03:03:02 andyrtr joined the channel 04:01:25 Quit: SailorMoon (Read error: Connection reset by peer) 04:01:44 SailorMoon joined the channel 04:08:21 Quit: andyrtr (Ping timeout: 266 seconds) 04:09:11 Quit: ahoneybun (Excess Flood) 04:11:19 ahoneybun joined the channel 04:11:20 Quit: ahoneybun (Changing host) 04:11:20 ahoneybun joined the channel 04:12:57 lizzie_ joined the channel 04:13:07 Nokius_ joined the channel 04:13:19 Quit: Nokius (Write error: Broken pipe) 04:13:23 Quit: lizzie (Remote host closed the connection) 04:14:18 andyrtr joined the channel 04:30:10 Quit: andyrtr (Ping timeout: 256 seconds) 04:35:50 Logiczero joined the channel 04:36:06 Quit: Logiczero (Client Quit) 04:37:01 andyrtr joined the channel 04:48:24 Quit: andyrtr (Ping timeout: 258 seconds) 04:54:23 andyrtr joined the channel 05:04:32 Quit: SailorMoon (Read error: Connection reset by peer) 05:19:27 SailorMoon joined the channel 05:28:19 pbryan joined the channel 07:13:11 Quit: andyrtr (Ping timeout: 246 seconds) 07:19:21 andyrtr joined the channel 07:28:19 Quit: andyrtr (Ping timeout: 250 seconds) 07:34:20 andyrtr joined the channel 07:51:55 kido` joined the channel 07:53:12 lizzie joined the channel 07:53:23 Quit: kido (Ping timeout: 252 seconds) 07:53:23 Quit: anYc (Ping timeout: 252 seconds) 07:53:24 Quit: lizzie_ (Ping timeout: 252 seconds) 07:53:24 Quit: OsakaFoo (Ping timeout: 252 seconds) 07:53:24 Quit: TheAppleMan (Read error: Connection reset by peer) 07:53:28 Quit: fnukis (Ping timeout: 250 seconds) 07:53:36 lizzie is now known as Guest59534 07:54:15 OsakaFoo joined the channel 07:54:17 fnukis joined the channel 07:59:05 anYc joined the channel 09:39:25 Quit: SailorMoon (Quit: Leaving) 10:16:56 lepage joined the channel 10:37:37 Nokius_ is now known as Nokius 11:16:50 Quit: maker2222 (Ping timeout: 264 seconds) 11:33:09 reererer joined the channel 11:33:09 Quit: reererer (Client Quit) 11:42:22 Quit: Guest59534 (Changing host) 11:42:22 Guest59534 joined the channel 11:42:26 Guest59534 is now known as lizzie 11:58:10 Mister_Magister joined the channel 12:02:12 Quit: Mister_Magister (Remote host closed the connection) 12:02:35 Mister_Magister joined the channel 12:05:14 Quit: pashik (Read error: Connection reset by peer) 12:06:07 pashik joined the channel 13:31:26 jreznik_ joined the channel 13:40:53 TheAppleMan joined the channel 14:06:51 bencord0_ joined the channel 14:09:33 Quit: andyrtr (Ping timeout: 260 seconds) 14:10:06 Quit: lepage (Quit: Konversation terminated!) 14:11:03 bencord0_ | Merry christmas everyone 14:11:47 bencord0_ | So, errm.... I may have a booting kernel for sturgeon (the Huawei watch) 14:12:28 bencord0_ | but I'm stuck in the initramfs 14:13:38 bencord0_ | with a bit of hacking I've managed to get adbd running, `adb devices` works, but I can't `adb shell`. 14:14:20 andyrtr joined the channel 14:14:32 bencord0_ | At somepoint today, I'll upload my diffs to github somewhere and share 14:14:44 kido` | hi bencord0_ 14:15:15 kido` | merry christmas, that is very cool 14:15:49 kido` | have you double checked how you setup your functionfs? 14:17:07 TheAppleMan | hey kido`, i've got the fossil q founder but it has a wireless charger not a cradle :( 14:17:50 kido` | it seems that santa brought cool toys 14:18:01 TheAppleMan | XD 14:18:17 TheAppleMan | it's lucky we didn't buy two of the same watch 14:18:41 bencord0_ | I'm still reading what's in the initramfs, but I thought I'd report back progress 14:20:11 kido` | TheAppleMan: meh indeed it doesn't seem to have pins out 14:20:34 kido` | on the long term this is gonna become a problem 14:20:42 kido` is now known as kido 14:20:48 Quit: kido (Changing host) 14:20:48 kido joined the channel 14:24:36 kido | maybe we can find pins inside the watch but I'm afraid no one will want to disassemble watches (understandably) 14:25:31 TheAppleMan | this one is pretty solid 14:26:37 TheAppleMan | apparently you can do adb over bluetooth but i don't think that'll work for fastboot 14:26:55 kido | yes, that's why it's not an issue for android developers but an issue for us 14:29:28 kido | meh I'm thinking about some kind of chainloading from the android's userspace to an asteroid boot.img but it would still require a custom kernel on the boot partition which would need fastboot at some point 14:29:49 bencord0_ | kexec? 14:30:07 bencord0_ | It maybe enough to do it from a root shell 14:30:09 kido | I doubt it's enabled by default on watches :) 14:31:01 bencord0_ | can you write to boot.img or the recovery from android wear? 14:31:29 kido | yes you can dd the partition but if you fail you're screwed 14:33:30 kido | mhhh 14:34:00 kido | maybe you're right, just rewriting the recovery partition and booting into it with reboot recovery could do the trick 14:34:21 kido | easily revertable, no need for fastboot etc 14:34:44 kido | but it is still gonna be a pain to make a port without a shell 14:35:01 kido | porters would need some kind of uart access to enable make lipstick and bluetooth work 14:35:14 kido | or wifi, then people could use ssh 14:36:39 bencord0_ | Yeah, I think that's the step that I'm blocked on right now. figuring out the mmcblk layout 14:37:12 bencord0_ | I'm preparing a init.sh that prints output of ls to a file, so that I can `adb pull` it 14:37:24 kido | usually there is a folder with symlinks having nice names and pointing to the corresponding mmc device 14:37:44 kido | you should be able to check that from android 14:38:59 kido | http://www.epiccm.org/2012/07/kexec-and-proof-of-concept-source-built.html this is interesting 14:41:12 bencord0_ | That looks like what we've just been talking about :) 14:41:17 kido | yep 14:42:15 bencord0_ | jolla does something interesting, I think that their "fastboot" mode is done in software 14:43:04 bencord0_ | or it has a very dynamic nature to it, it does things that I don't understand how a firmware only implementation would know 14:44:00 bencord0_ | e.g. pin entry access control to determin if you're even allowed to boot other kernels 14:44:24 kido | interesting, I've never checked it 14:48:17 bencord0_ | If I edit `recipes-core/initrdscripts/initramfs-boot-android/init.sh`, is there a quick way to regen that into the kernel image? 14:48:31 bencord0_ | I'm using `bitbake linux-sturgeon -c bundle_initramfs` 14:48:48 kido | you could re-run mkboot/mkbootimg/abootimg yourself 14:49:32 kido | I usualy keep a folder with the original boot.img packed and extracted and the new boot.img packed and extracted with instructions on how to re-pack modification 14:51:11 bencord0_ | I think the bitbake commands are fine, it's just a bit hard to tell how much of sstate it invalidates based on timestamps 14:51:30 bencord0_ | I'm used to gentoo's `ebuild` command, and these `tasks` are cool, but new to me 14:52:45 kido | ok 14:55:11 bencord0_ | yeah, TIL: `-c bundle_initramfs` then `-c build` works. 14:55:50 TheAppleMan | bencord0_: what are you changing, just the initramfs init? no kernel changes? 14:56:58 bencord0_ | yep 14:57:40 bencord0_ | working around a lack of `adb shell`, so I'm feeding in commands to the init.sh, then `adb pull`ing the output 14:58:03 bencord0_ | takes 1min 3secs to cycle it 15:15:30 Quit: jreznik_ (Ping timeout: 246 seconds) 15:57:19 Quit: M4GNV5 (Read error: Connection reset by peer) 16:20:05 bencord0_ | https://twitter.com/bencord0/status/813403933632765956 16:22:57 kido | yay :D 16:24:38 bencord0_ | FYI: sdcard is p26, system is p24 16:24:50 kido | still no adb though? 16:25:08 bencord0_ | I have `adb shell` in asteroidos now 16:25:20 kido | perfect 16:25:54 bencord0_ | systemd has loaded, and is reporting ok 16:27:54 kido | yep, now you're good to debug lipstick with strace and gdb 16:37:58 bencord0_ | Do you have any docs about how systemd hands over to the gui? 16:44:24 kido | I don't 16:44:59 kido | but this hack is of interest https://github.com/AsteroidOS/meta-asteroid/blob/master/recipes-core/systemd/systemd_%25.bbappend#L13 16:45:07 kido | especialy l16 16:45:20 kido | and this https://github.com/AsteroidOS/meta-asteroid/blob/master/recipes-asteroid/asteroid-launcher/asteroid-launcher/asteroid-launcher.service 16:51:24 bencord0_ | http://pastebin.com/RuWaXvBp 16:51:39 bencord0_ | :( 16:52:42 bencord0_ | I think that's as far as I'm going to get today 16:52:44 kido | how did you compile your /system directory? try to bitbake strace, put in to the fatch 16:52:46 kido | watch* 16:52:54 kido | and run lipstick inside strace 16:53:17 bencord0_ | yeah, that will be my next steps. maybe even ldd it to make sure all the bits are there 16:53:59 kido | usualy there are stupid things to fix, like it could be trying to load a non-existing gralloc module in /system/lib/hw/ 17:12:38 TheAppleMan | bencord0_: what about systemctl --user as ceres? 17:16:42 bencord0 | Segv 17:17:13 TheAppleMan | could be why... 17:17:27 bencord0_ | same reason, lipstick is crashing out 17:17:44 bencord0_ | I'm building some tools to debug it 17:18:24 bencord0_ | I think the stuff in asteroid-image-dev should be enough 17:18:24 TheAppleMan | doesn't sound like the firstboot issue 17:18:56 bencord0_ | asteroid os is up, systemd is happy. just having some problems starting asteroid-launcer.service 17:19:12 kido | asteroid-image-dbg should be lighter than dev 17:19:17 kido | -dev isn't really useful 17:19:47 bencord0_ | isn't it a superset? 17:20:05 kido | yep but it sometimes too large for the watch's storage 17:20:12 bencord0_ | righto 17:20:30 kido | -dbg with strace should be all that you need to debug the lipstick's segfault 17:20:35 TheAppleMan | oh yeah, we had to increase the rootfs already 17:20:37 kido | but don't worry about it, that's the main part of porting 17:20:39 bencord0_ | I'll keep an eye on it. Already needed to bump the IMAGE_ROOTFS_SIZE 17:21:15 bencord0_ | rootfs is 170M ish on my laptop (zfs) but almost 300M in ext2 17:21:44 bencord0_ | anyway, gdb is compiling, so I'm going to grab some food 17:22:19 kido | :) 17:28:38 zefzefz joined the channel 17:29:04 Quit: zefzefz (Client Quit) 17:53:43 jreznik_ joined the channel 18:38:11 bencord0_ | http://pastebin.com/2eW2hxSd 18:38:51 bencord0_ | Yeah, those -dev and -dbg images are big. Ended up just using opkg. 18:45:43 bencord0_ | looks like lipstick is defaulting to XCB, which isn't going to work since we don't use X11 18:46:23 kido | according to this http://pastebin.com/RuWaXvBp it's not using the XCB QPA 18:54:45 bencord0_ | I think it's just cycling over all of the QPA platforms 19:33:30 Nokius | bencord0_: nice \o/ 19:41:14 SylvieLorxu joined the channel 20:39:38 Quit: andyrtr (Ping timeout: 264 seconds) 20:44:49 andyrtr joined the channel 20:53:10 anYc | bencord0_: what kernel version do you use? I'm currently working on the zenwatch3 which uses 3.18 20:53:13 anYc | +mm 20:56:10 Quit: andyrtr (Ping timeout: 256 seconds) 20:56:30 bencord0_ | I'm using https://android.googlesource.com/kernel/msm/+/android-msm-sturgeon-3.10-marshmallow-mr1-wear-release-1 20:56:52 anYc | ah ok, but also marshmallow 20:57:00 bencord0_ | yes 20:58:53 anYc | I was also wondering yesterday if it doesn't load the hwcomposer qpa on my watch but I didn't had the time to further investigate it 21:01:13 bencord0_ | I can see it opening and reading the qpa into memory. Then it decides to sigabrt. 21:01:25 andyrtr joined the channel 21:01:37 bencord0_ | I think it's still in qt's startup code 21:03:21 anYc | on my watch he segfaults in the qt constructor indeed 21:08:36 anYc | btw, I switched to mount the rootfs over NFS from my PC. Makes the debugging a bit easier 21:12:17 bencord0_ | I haven't even started looking at the network yet 21:14:39 Quit: SylvieLorxu (Quit: ZNC - http://znc.in) 21:32:06 Quit: Mister_Magister (Ping timeout: 246 seconds) 21:38:46 Mister_Magister joined the channel 21:57:39 bencord0_ | http://pastebin.com/7GJ1EFN6 21:57:55 bencord0_ | `connect(3, {sa_family=AF_UNIX, sun_path="/run/user/1000/wayland-0"}, 27) = -1 ENOENT (No such file or directory)` 21:58:05 bencord0_ | any idea how to get that socket? 22:54:56 Nokius_ joined the channel 22:57:57 Quit: Nokius (Ping timeout: 248 seconds) 23:24:31 Nokius_ is now known as Nokius 23:39:30 Quit: jreznik_ (Ping timeout: 246 seconds)