*** Piece_Maker <Piece_Maker!~Acou_Bass@host-78-144-157-133.as13285.net> has joined #sfdroid | 02:31 | |
*** Acou_Bass <Acou_Bass!~Acou_Bass@host-89-241-242-228.as13285.net> has quit IRC (Ping timeout: 255 seconds) | 02:34 | |
*** Piece_Maker is now known as Acou_Bass | 02:34 | |
*** pseudodev <pseudodev!uid205973@gateway/web/irccloud.com/x-kihrpfdjdxrbqgcz> has joined #sfdroid | 02:46 | |
*** piggz_ <piggz_!~piggz@95.147.209.146> has joined #sfdroid | 06:25 | |
*** piggz_ <piggz_!~piggz@95.147.209.146> has quit IRC (Ping timeout: 268 seconds) | 06:31 | |
*** piggz_ <piggz_!~piggz@95.147.209.146> has joined #sfdroid | 07:06 | |
*** piggz_ <piggz_!~piggz@95.147.209.146> has quit IRC (Ping timeout: 272 seconds) | 07:13 | |
locusf | yay! | 07:52 |
---|---|---|
locusf | now just gotta figure out the correct kernel | 08:00 |
mal | locusf: I can build it, unless you are in a hurry to get it right now, the changes should be quite small afaik | 09:00 |
locusf | mal: nope | 09:03 |
locusf | not in a hurry | 09:03 |
mal | I just rebase hybris-14.1 to lineage yesterday, need to test it soon | 09:05 |
*** nh1402 <nh1402!~nh1402@81.133.137.69> has joined #sfdroid | 10:15 | |
*** Mister_Magister_ <Mister_Magister_!~foidbgen@user-94-254-146-124.play-internet.pl> has joined #sfdroid | 10:35 | |
saidinesh5 | locusf: how do i use a more recent gcc for building this? https://build.merproject.org/package/live_build_log/home:saidinesh5:anbox/android-tools/sailfish_latest_armv7hl/armv8el | 10:46 |
saidinesh5 | nh1402: https://build.merproject.org/package/show/home:saidinesh5:anbox/anbox | 10:47 |
*** Mister_Magister_ <Mister_Magister_!~foidbgen@user-94-254-146-124.play-internet.pl> has quit IRC (Ping timeout: 272 seconds) | 10:47 | |
*** Mister_Magister_ <Mister_Magister_!~foidbgen@user-94-254-146-124.play-internet.pl> has joined #sfdroid | 10:51 | |
*** Mister_Magister_ is now known as Mister_Magister | 12:02 | |
locusf | saidinesh5: don't know if its possible | 12:17 |
*** TheKit <TheKit!~nekit@178.45.144.138> has quit IRC (Remote host closed the connection) | 12:18 | |
saidinesh5 | oh.. hmm.. | 12:18 |
*** NeKit <NeKit!~nekit@178.45.144.138> has joined #sfdroid | 12:35 | |
*** mp107 <mp107!~mp107___@PC-109-125-232-42.wegrow.domtel.com.pl> has joined #sfdroid | 14:35 | |
*** mp107 <mp107!~mp107___@PC-109-125-232-42.wegrow.domtel.com.pl> has quit IRC (Quit: Leaving) | 14:48 | |
*** nh1402 <nh1402!~nh1402@81.133.137.69> has quit IRC (Quit: Leaving) | 15:38 | |
krnlyng__ | saidinesh5, neat | 16:36 |
krnlyng__ | saidinesh5, now we just need to test it if we've broken anything :D | 16:36 |
krnlyng__ | saidinesh5, what was the issue? | 16:36 |
krnlyng__ | ah i see the link | 16:36 |
saidinesh5 | krnlyng__: dunno.. something related to file name length i think | 16:37 |
saidinesh5 | krnlyng__: btw. i think we have to split up that prepare-anbox script .. it was even doing bind mounts.. | 16:39 |
krnlyng__ | saidinesh5, why split? i thought everything that belongs together should be in one place | 16:40 |
saidinesh5 | krnlyng__: the copying of files happens just once.. at the time of install | 16:40 |
saidinesh5 | the bind mounts happen everytime anbox is started.. | 16:40 |
krnlyng__ | saidinesh5, yes | 16:40 |
krnlyng__ | but | 16:40 |
krnlyng__ | saidinesh5, this script only sets up a config file | 16:40 |
krnlyng__ | saidinesh5, which is loaded everytime by anbox | 16:40 |
saidinesh5 | Ahhh.. now i see the umount $ROOTFS | 16:41 |
saidinesh5 | i just saw the mount and jumped at that conclusion .. nevermind :p | 16:41 |
saidinesh5 | krnlyng__: another little nitpick.. could you rename the other project to anbox-sailfishos-image? | 16:42 |
saidinesh5 | if this project is anbox-sailfishos , the other being anbox-sfdroid-image feels odd.. | 16:42 |
krnlyng__ | alright | 16:42 |
saidinesh5 | i am looking at that headless mode bit meanwhile though.. havent yet gotten to test anbox on device.. need to free up some space for the image to install i think | 16:43 |
saidinesh5 | willl do it tonight.. | 16:44 |
saidinesh5 | my next phone will definitely be a 32GB+ model.. | 16:46 |
*** Mister_Magister <Mister_Magister!~foidbgen@user-94-254-146-124.play-internet.pl> has quit IRC (Remote host closed the connection) | 16:53 | |
*** Mister_Magister <Mister_Magister!~foidbgen@user-94-254-146-124.play-internet.pl> has joined #sfdroid | 16:57 | |
*** Mister_Magister_ <Mister_Magister_!~foidbgen@user-94-254-146-124.play-internet.pl> has joined #sfdroid | 17:11 | |
*** Mister_Magister <Mister_Magister!~foidbgen@user-94-254-146-124.play-internet.pl> has quit IRC (Ping timeout: 268 seconds) | 17:11 | |
mal | did anyone have any issues applying the binder namespaces patches? https://lkml.org/lkml/diff/2013/12/22/16/1 https://lkml.org/lkml/diff/2013/12/22/15/1 | 17:45 |
krnlyng__ | mal, i think everyone had | 17:57 |
krnlyng__ | so far | 17:58 |
krnlyng__ | mal, for me it was minor issues, had to fix 2 rej's | 17:58 |
mal | krnlyng__: I have 3 rejs for one and more than 11 for the other patch | 17:59 |
nh1402[m] | i tried saidinesh5 patch from a few weeks ago now, that worked fine for me | 17:59 |
mal | nh1402[m]: where are those? I saw only the links I gave just a moment ago | 18:00 |
krnlyng__ | mal, which kernel version do you have? | 18:01 |
krnlyng__ | mal, there are patches for more recent kernel versions in the anbox repo | 18:01 |
mal | 3.4 | 18:01 |
nh1402[m] | https://bpaste.net/show/adeda1721303 | 18:01 |
nh1402[m] | mal: ^ | 18:01 |
mal | nh1402[m]: is that both of them or just one? | 18:02 |
nh1402[m] | I'm not sure | 18:02 |
krnlyng__ | mal, thats both | 18:02 |
saidinesh5 | mal: both | 18:02 |
mal | ok, trying | 18:02 |
saidinesh5 | yeah had to clean up this patch for 3.4/Android .. it wasnt applying otherwise | 18:02 |
mal | only one fail | 18:03 |
saidinesh5 | Ahh.. for me there were a bunch.. | 18:03 |
mal | saidinesh5: the original patches had a lot but that seems almost ok | 18:04 |
saidinesh5 | yeah .. had to apply the original patches manually using a text editor.. otherwise it was full of rejects for me | 18:05 |
saidinesh5 | mal: 3 changes basically: https://github.com/CancroSailors/android_kernel_xiaomi_cancro/commits/anbox | 18:07 |
saidinesh5 | support for overlayfs, so we can reuse the android binary blobs and overlay them into the container's namespace, binder namespaces (so RIL wont break) , and the third one to make lxc-checkconfig happy | 18:08 |
mal | saidinesh5: ah, more defconfig changes than I assumed first | 18:12 |
saidinesh5 | yeah lxc2 wanted a lot more.. | 18:12 |
saidinesh5 | mal: the def config changes, make sure you verify them with lxc-checkconfig | 18:12 |
mal | saidinesh5: ok, need to find that | 18:15 |
saidinesh5 | it is installed when you install lxc .. you can install lxc from that repo | 18:15 |
saidinesh5 | ssu ar anbox-obs http://repo.merproject.org/obs/home:/saidinesh5:/anbox/sailfish_latest_armv7hl/ | 18:15 |
saidinesh5 | zypper ref anbox-obs && zypper in anbox-sailfishos | 18:16 |
mal | saidinesh5: ok, I thought it would be some offline script | 18:16 |
saidinesh5 | havent bothered separating that script | 18:16 |
saidinesh5 | cp /proc/config.gz /tmp/ && gunzip /tmp/config.gz | 18:16 |
saidinesh5 | CONFIG=/tmp/config lxc-checkconfig | 18:16 |
locusf | isn't it in any repo? | 18:17 |
saidinesh5 | it should be in lxc repo | 18:17 |
locusf | like git | 18:17 |
krnlyng__ | saidinesh5 has built the correct version of lxc in obs | 18:17 |
saidinesh5 | https://github.com/lxc/lxc/blob/master/src/lxc/tools/lxc-checkconfig.in | 18:17 |
mal | saidinesh5: nothing provides anbox-sailfishos-image ? | 18:17 |
saidinesh5 | oh.. odd.. it is here: https://build.merproject.org/package/show/home:saidinesh5:anbox/anbox-sailfishos-image | 18:18 |
saidinesh5 | krnlyng__: did yu rebuild the rpm? or rename the old rpm? | 18:19 |
krnlyng__ | i rebuilt it | 18:19 |
krnlyng__ | https://bpaste.net/show/6e1b1bba618d | 18:19 |
saidinesh5 | hmm.. odd it isnt in the repo: http://repo.merproject.org/obs/home:/saidinesh5:/anbox/sailfish_latest_armv7hl/armv7hl/ | 18:20 |
krnlyng__ | publish flag? | 18:20 |
krnlyng__ | i added it and triggered a rebuild | 18:22 |
krnlyng__ | time to test the rpms myself | 18:22 |
saidinesh5 | Ahh | 18:22 |
saidinesh5 | now we see it: http://repo.merproject.org/obs/home:/saidinesh5:/anbox/sailfish_latest_armv7hl/armv7hl/ | 18:22 |
krnlyng__ | but the wayland code isn't doing itself either xD | 18:22 |
krnlyng__ | oh so rebuild wasn't necessary | 18:23 |
saidinesh5 | yeah odd.. i think it was just a publish flag thing | 18:23 |
saidinesh5 | oh well.. at least the repo is functional now.. | 18:23 |
saidinesh5 | almost.. there's still the wonky adb | 18:24 |
saidinesh5 | adb/android-tools | 18:24 |
krnlyng__ | yeah | 18:24 |
krnlyng__ | we should probably just build an older version | 18:24 |
mal | saidinesh5: maybe the rpm has some issues | 18:24 |
locusf | it takes time to copy 270MB package | 18:24 |
krnlyng__ | because we don't have gcc-5/6 | 18:24 |
saidinesh5 | krnlyng__: how did you build it on your device? it should have the same gcc compiler right? | 18:25 |
krnlyng__ | i used an older version | 18:25 |
saidinesh5 | Ah so lets build the same version! | 18:25 |
saidinesh5 | just make sure to rename the binary to anbox-adb though | 18:25 |
krnlyng__ | yep | 18:26 |
krnlyng__ | just trying to complete the initial wayland code inside android first | 18:26 |
krnlyng__ | and we also should try and fix the boost mess at some point | 18:27 |
saidinesh5 | yeah, the boost mess will be the last thing for now though | 18:28 |
krnlyng__ | yeah | 18:28 |
locusf | good job guys | 18:28 |
saidinesh5 | locusf: O_O you ran it? | 18:28 |
locusf | not yet | 18:29 |
saidinesh5 | Ahh | 18:29 |
locusf | just giving some encouragement :) | 18:29 |
saidinesh5 | hehh i see.. thank you.. | 18:29 |
locusf | the boost mess is solved with some massaging of mer:core with a proper MR | 18:29 |
locusf | I suppose the versions arent that different? | 18:30 |
saidinesh5 | locusf: they already have boost 1.61 in the repos | 18:30 |
saidinesh5 | its just not in 2.1 .. | 18:30 |
locusf | ok | 18:30 |
saidinesh5 | 2.1 has 1.51 , and boost-1.51 doesnt have the modules anbox guys are using. | 18:31 |
saidinesh5 | i manually added boost-log to 1.51 and it compiled.. but 2 days ago they had to pull in boost-endian which pulls in some other boost module | 18:31 |
locusf | heh | 18:32 |
krnlyng__ | the boost mess will be dropped by writing our own lxc container start stop functionality xD | 18:33 |
krnlyng__ | if and when we actually do that | 18:33 |
mal | saidinesh5: now it finds the package | 18:36 |
saidinesh5 | yeah krnlyng__ had to enable the publish flags and rebuild it | 18:36 |
krnlyng__ | saidinesh5, seems the prepare-anbox script has an issue | 18:36 |
krnlyng__ | :( | 18:36 |
mal | kernel changes added, now build time | 18:36 |
krnlyng__ | ln -sf should be used instead of ln -s | 18:37 |
krnlyng__ | sorry | 18:37 |
saidinesh5 | Ahh | 18:37 |
krnlyng__ | or wait | 18:37 |
krnlyng__ | sorry | 18:37 |
krnlyng__ | i had the package installed already | 18:37 |
saidinesh5 | I see | 18:37 |
krnlyng__ | or rather the symlinks | 18:37 |
krnlyng__ | it's fine | 18:37 |
saidinesh5 | ln -sf should be the thing either way | 18:37 |
saidinesh5 | or else what'd happen to updates? | 18:37 |
krnlyng__ | yup | 18:38 |
locusf | mal: \o/ | 18:38 |
krnlyng__ | it's crashing lol | 18:39 |
saidinesh5 | krnlyng__: unsurprising :p | 18:39 |
saidinesh5 | where is it crashing though? | 18:39 |
krnlyng__ | after starting the container definately | 18:40 |
krnlyng__ | not sure where yet | 18:40 |
krnlyng__ | gdb | 18:40 |
saidinesh5 | maybe some file was missing and it was trying to read from it.. | 18:40 |
krnlyng__ | or wait no | 18:41 |
krnlyng__ | it fails when starting the container it seems | 18:41 |
saidinesh5 | oh.. | 18:41 |
saidinesh5 | error? | 18:41 |
krnlyng__ | same as always | 18:42 |
krnlyng__ | tripple Failed to start container | 18:42 |
saidinesh5 | hmm.. strace mayeb? | 18:43 |
krnlyng__ | already at it | 18:43 |
krnlyng__ | it's complaining about already existing files in /sys and cgroups | 18:48 |
*** nh1402 <nh1402!~nh1402@212.182.159.143.dyn.plus.net> has joined #sfdroid | 18:51 | |
saidinesh5 | odd but sounds familiar | 18:51 |
mal | saidinesh5: not all of the lxc flags are support in 3.4 kernel | 18:52 |
krnlyng__ | we're not already starting container manager after the rpm installation? | 18:52 |
saidinesh5 | mal: yup. although the ones missing dont seem to be mandatory | 18:52 |
saidinesh5 | so will have to add those patches again later on.. | 18:52 |
mal | saidinesh5: ok | 18:53 |
saidinesh5 | krnlyng__: we're not but what if it was to do with the container from your compiled anbox? | 18:53 |
saidinesh5 | krnlyng__: if you rebooted the device after an installation, it probably is started by systemd.. | 18:54 |
saidinesh5 | ps aux | grep anbox ? | 18:54 |
krnlyng__ | nothing | 19:00 |
saidinesh5 | krnlyng__ what flags did you compile lxc with? | 19:03 |
krnlyng__ | i think just the default ones | 19:03 |
krnlyng__ | and --prefix=/usr | 19:03 |
saidinesh5 | Ah hmm.. | 19:03 |
*** piggz_ <piggz_!~piggz@95.147.209.146> has joined #sfdroid | 19:04 | |
saidinesh5 | %{?_with_runtime_path} --disable-rpath .. | 19:05 |
mal | saidinesh5: ok, now patched kernel and anbox are installed but not sure how to debug this, nothing happens | 19:05 |
saidinesh5 | yeah there seems to be an error starting the container.. | 19:05 |
krnlyng__ | mal, as root: anbox container-manager --daemon | 19:06 |
krnlyng__ | mal, then as nemo anbox session-manager --single-window | 19:06 |
mal | Failed to mount Android rootfs | 19:06 |
krnlyng__ | mal, do you have CONFIG_SQUASHFS_FS or what it's called? | 19:07 |
saidinesh5 | CONFIG_OVERLAYFS_FS i think | 19:07 |
saidinesh5 | okay even SQUASHFS... | 19:07 |
krnlyng__ | yeah both are required | 19:07 |
mal | saidinesh5: hmm, your defconfig doesn't have SQUASHFS | 19:09 |
saidinesh5 | Ahh.. didnt push tht one | 19:10 |
krnlyng__ | ok | 19:10 |
krnlyng__ | container starts with the systemd services | 19:10 |
krnlyng__ | systemctl start anbox-container-manager.service | 19:10 |
krnlyng__ | as root | 19:11 |
krnlyng__ | systemctl start --user anbox-session-manager.service | 19:11 |
mal | saidinesh5: ok, adding that also | 19:11 |
saidinesh5 | CONFIG_SQUASHFS=y | 19:11 |
saidinesh5 | krnlyng__: oh.. so its working after a reboot? | 19:12 |
krnlyng__ | nope had to start them manually | 19:12 |
krnlyng__ | you commented it out in the rpm spec | 19:12 |
krnlyng__ | still no app icons though | 19:13 |
saidinesh5 | Ah .. | 19:13 |
krnlyng__ | not sure if it is still booting | 19:13 |
krnlyng__ | it's doing *something* | 19:13 |
krnlyng__ | ahh i forgot | 19:15 |
saidinesh5 | hmm? | 19:15 |
krnlyng__ | anbox container-manager --privileged | 19:15 |
saidinesh5 | could you start that via. systemd? | 19:16 |
krnlyng__ | it think that command should work for manual too | 19:16 |
saidinesh5 | systemctl start anbox-container-manager | 19:16 |
krnlyng__ | yes i used that | 19:16 |
krnlyng__ | so segfault now... | 19:16 |
saidinesh5 | so that process should be running right? | 19:16 |
saidinesh5 | Ahh | 19:16 |
saidinesh5 | now gdb. | 19:16 |
krnlyng__ | yup | 19:16 |
saidinesh5 | i think that segfault prolly creates those files in sys and doesnt clear them? | 19:17 |
krnlyng__ | ok i know the fix | 19:18 |
saidinesh5 | Ah.. what's the issue? | 19:19 |
krnlyng__ | it's due to libhybris issue | 19:19 |
saidinesh5 | oh? | 19:19 |
krnlyng__ | or rather my mistake | 19:19 |
saidinesh5 | Ahh is this from a recent commit? | 19:19 |
krnlyng__ | yes | 19:19 |
saidinesh5 | Ahh | 19:19 |
*** piggz_ <piggz_!~piggz@95.147.209.146> has quit IRC (Ping timeout: 240 seconds) | 19:21 | |
krnlyng__ | rebuilding | 19:21 |
krnlyng__ | should work nwo | 19:21 |
saidinesh5 | Aye | 19:22 |
mal | still fails to mount | 19:22 |
mal | saidinesh5: push the updated commit to github so I can check if something else is missing | 19:22 |
krnlyng__ | mal, do you have the overlayfs patches? | 19:22 |
mal | yes | 19:23 |
krnlyng__ | hmm wait there was somehting | 19:23 |
saidinesh5 | mal: link? | 19:23 |
krnlyng__ | mal, you need to enable squashfs compression | 19:23 |
krnlyng__ | decompression* | 19:23 |
mal | saidinesh5: I mean you said you forgot to push something to your github | 19:23 |
mal | krnlyng__: ok | 19:24 |
saidinesh5 | Ahh .. | 19:24 |
krnlyng__ | mal, while youre at it, lxc-checkconfig should at least look like this: | 19:24 |
saidinesh5 | just this: CONFIG_SQUASHFS=y | 19:24 |
saidinesh5 | i didnt push it as i havent verified that working on my device yet | 19:24 |
krnlyng__ | https://bpaste.net/show/552138c91727 | 19:24 |
krnlyng__ | so veth should be enabled | 19:24 |
mal | krnlyng__: which one of the compression methods? | 19:24 |
krnlyng__ | if i remember correctly xz | 19:25 |
krnlyng__ | let me check | 19:25 |
krnlyng__ | mal, yes XZ | 19:25 |
saidinesh5 | krnlyng__: CONFIG_HAVE_KERNEL_XZ=y ? | 19:25 |
krnlyng__ | +CONFIG_SQUASHFS_XZ=y | 19:25 |
krnlyng__ | -+ | 19:26 |
saidinesh5 | Ah i have that one already | 19:26 |
krnlyng__ | we should definatley make a list afterwards | 19:26 |
saidinesh5 | yeah hence made all my kernel patches so far in a separate branch | 19:26 |
mal | krnlyng__: https://pastebin.com/S761Rseu | 19:26 |
krnlyng__ | mal, CONFIG_NF_NAT_IPV4 might be necessary not 100% sure | 19:27 |
krnlyng__ | otherwise looks good to me | 19:27 |
mal | krnlyng__: that is not possible, I added all of those and that came out | 19:28 |
krnlyng__ | ah ok | 19:28 |
mal | so maybe some are not supported | 19:28 |
krnlyng__ | well we'll see if it works :) | 19:28 |
krnlyng__ | in worst case you don't have networking | 19:28 |
mal | so do I need both CONFIG_HAVE_KERNEL_XZ=y and CONFIG_SQUASHFS_XZ=y? | 19:29 |
mal | krnlyng__: then I will patch the kernel | 19:29 |
krnlyng__ | mal, you definately need CONFIG_SQUASFS_XZ. CONFIG_HAVE_KERNEL_XZ reading docs | 19:30 |
krnlyng__ | doesn't seem it's necessary at first look | 19:31 |
mal | I think latter is not related, that is for kernel compression | 19:32 |
krnlyng__ | yeah | 19:32 |
krnlyng__ | it's still segfaulting :( | 19:34 |
saidinesh5 | krnlyng__: now where? | 19:35 |
krnlyng__ | same | 19:35 |
saidinesh5 | cleared space on the phone | 19:38 |
saidinesh5 | installing now | 19:38 |
mal | hmm, now Failed to mount Android overlay | 19:39 |
mal | so something is still wrong | 19:39 |
saidinesh5 | strace maybe? | 19:41 |
mal | looks like I forgot to enable overlayfs in defconfig | 19:44 |
saidinesh5 | Ahh | 19:44 |
krnlyng__ | https://bpaste.net/show/a2e9ac8a846b | 19:48 |
krnlyng__ | is there anything wrong with that other than that it is an ugly hack? | 19:48 |
saidinesh5 | not sure.. | 19:48 |
krnlyng__ | $7 = 0x18a9a8 "/usr/lib/libhybris/eglplatform_wayland.so: undefined symbol: hybris_egl_display_get_mapping" | 19:50 |
krnlyng__ | hmm | 19:50 |
krnlyng__ | damn | 19:51 |
krnlyng__ | it's in libEGL.so | 19:51 |
krnlyng__ | it will work when built :D | 19:53 |
*** piggz_ <piggz_!~piggz@95.147.209.146> has joined #sfdroid | 19:53 | |
saidinesh5 | lol | 19:54 |
saidinesh5 | fixing the icon issue too | 19:54 |
mal | krnlyng__: ok, no more mount errors, should be --daemon put it to background, it doesn't | 19:54 |
krnlyng__ | mal, daemon only skips a warning apparently | 19:56 |
krnlyng__ | mal, and btw wait till the new rpm is built | 19:57 |
mal | krnlyng__: ok | 19:58 |
mal | krnlyng__: how about those systemd services, are those the proper way? | 19:58 |
*** Mister_Magister <Mister_Magister!~foidbgen@user-94-254-147-32.play-internet.pl> has joined #sfdroid | 19:58 | |
saidinesh5 | mal: got them from arch | 19:58 |
saidinesh5 | the systemd serices | 19:58 |
*** Mister_Magister_ <Mister_Magister_!~foidbgen@user-94-254-146-124.play-internet.pl> has quit IRC (Ping timeout: 246 seconds) | 19:59 | |
mal | let's see what happens with the new package once it's built | 20:00 |
krnlyng__ | mal, which android base do you have? aosp? and version? | 20:00 |
mal | krnlyng__: aosp5 | 20:01 |
krnlyng__ | ok | 20:01 |
krnlyng__ | mal, well i am not 100% sure but it might fail due to blob incompatibilities. these are usually fixable but a bit of a bummer. we need to figure out a good way to handle those. | 20:01 |
mal | ok | 20:02 |
krnlyng__ | i thought about modifying the linker when undefined references are detected in libEGL.so and libGLES...so and automatically hook them to abort. | 20:03 |
krnlyng__ | they are usually not called ;) | 20:03 |
mal | saidinesh5: hopefully the version number won't cause issues | 20:03 |
saidinesh5 | yeah vendor blobs usually can work across aosp versions too | 20:04 |
saidinesh5 | unless they fail.. | 20:04 |
saidinesh5 | like missing symbols and what not | 20:04 |
krnlyng__ | i've fixed the errors that were there for jolla c and it has 5 too? | 20:04 |
krnlyng__ | i received a frame :P | 20:05 |
saidinesh5 | jolla C has 4.4 right? | 20:05 |
krnlyng__ | now waiting for more | 20:05 |
krnlyng__ | (takes a while) | 20:05 |
saidinesh5 | frame for? | 20:05 |
krnlyng__ | a window update* buffer* black image* | 20:05 |
saidinesh5 | Ahh | 20:05 |
krnlyng__ | the initial one is always black | 20:05 |
nh1402[m] | 4.4 errors? | 20:05 |
krnlyng__ | and i have UI :D | 20:05 |
saidinesh5 | yeah | 20:05 |
saidinesh5 | wow | 20:05 |
mal | OBS is still building | 20:06 |
krnlyng__ | oh | 20:06 |
krnlyng__ | :( | 20:06 |
krnlyng__ | i think we lost touch input somewhere along the way | 20:06 |
saidinesh5 | Ahh... | 20:06 |
krnlyng__ | but in any case it works | 20:06 |
* saidinesh5 checks | 20:06 | |
krnlyng__ | now heading for proper rendering... | 20:06 |
saidinesh5 | krnlyng__: i think the touch maybe broken inside the image.. | 20:07 |
saidinesh5 | there is case SDL_FINGERUP: .. in the code.. | 20:08 |
krnlyng__ | hmm | 20:08 |
*** piggz_ <piggz_!~piggz@95.147.209.146> has quit IRC (Ping timeout: 240 seconds) | 20:08 | |
krnlyng__ | let me try again | 20:08 |
krnlyng__ | in the image i did not change the input code | 20:08 |
krnlyng__ | i mean at all | 20:08 |
krnlyng__ | it's still plain anbox | 20:08 |
saidinesh5 | oh.. | 20:09 |
krnlyng__ | i will try later | 20:10 |
krnlyng__ | want this to compile :D | 20:10 |
saidinesh5 | yup | 20:10 |
mal | krnlyng__: just curious, how did anbox behave before you added the latest fix? what crashed? | 20:11 |
saidinesh5 | anbox session manager probably.. | 20:12 |
mal | saidinesh5: krnlyng__ OBS build failed | 20:12 |
saidinesh5 | Bah | 20:12 |
saidinesh5 | yeah | 20:12 |
saidinesh5 | my bad | 20:12 |
mal | error: File not found: /home/abuild/rpmbuild/BUILDROOT/anbox-sailfishos-1.0+anbox.20170515195631.3.gab990fa-10.35.1.jolla.arm/usr/share/pixmaps/anbox.png | 20:12 |
saidinesh5 | yup | 20:13 |
saidinesh5 | fixed | 20:13 |
saidinesh5 | mal: package built | 20:25 |
saidinesh5 | okay: root 7245 0.1 0.2 55228 4692 ? Ssl 01:56 0:00 /usr/bin/anbox container-manager --privileged | 20:27 |
saidinesh5 | May 16 01:58:06 Sailfish systemd[993]: anbox-session-manager.service: main process exited, code=killed, status=11/SEGV | 20:28 |
saidinesh5 | hmm... | 20:28 |
krnlyng__ | oh | 20:29 |
saidinesh5 | lets see | 20:29 |
saidinesh5 | put it through gdb | 20:29 |
*** piggz_ <piggz_!~piggz@95.147.209.146> has joined #sfdroid | 20:30 | |
saidinesh5 | krnlyng__: https://bpaste.net/show/abd52438df1d ? | 20:31 |
saidinesh5 | familiar? | 20:31 |
krnlyng__ | i've seen that | 20:32 |
krnlyng__ | how do you start session-manager? | 20:32 |
saidinesh5 | from nemo | 20:32 |
saidinesh5 | gdb anbox | 20:32 |
saidinesh5 | r session-manager | 20:32 |
krnlyng__ | does it print anything before that? | 20:32 |
krnlyng__ | i think that error happens after the actual error | 20:33 |
saidinesh5 | Ahh right. [EE 2017-05-15 20:33:26] [client.cpp:49@start] Failed to start container: Failed to start container: Failed to start container | 20:33 |
krnlyng__ | yup | 20:33 |
krnlyng__ | lxc-checkconfig? | 20:33 |
saidinesh5 | yeah 1 sec | 20:34 |
saidinesh5 | havent rebooted the phone after applying a patch | 20:34 |
saidinesh5 | doing that now | 20:34 |
krnlyng__ | ah | 20:34 |
saidinesh5 | https://bpaste.net/show/dcd6aef20747 | 20:35 |
mal | doesn't work | 20:36 |
mal | no both container-manager and session-manager start ok | 20:37 |
mal | -no | 20:37 |
krnlyng__ | mal, did you put --single-window to session-manager? | 20:38 |
mal | no, I used the systemd-service | 20:38 |
krnlyng__ | it's required for now, until we change the rendering method | 20:38 |
krnlyng__ | i think it's not yet in the service | 20:39 |
mal | krnlyng__: where are the logs from anbox? | 20:40 |
saidinesh5 | lets see .. running it via. strace | 20:40 |
krnlyng__ | mal, /var/lib/anbox/ | 20:41 |
mal | krnlyng__: only container.log | 20:42 |
krnlyng__ | mal, for logcat: adb shell | 20:42 |
mal | krnlyng__: how to use that? | 20:43 |
krnlyng__ | mal, you need to run adb shell on the device itself | 20:43 |
krnlyng__ | for that you need adb, we need a package for that though | 20:44 |
saidinesh5 | krnlyng__: lxc 20170515204110.946 ERROR lxc_utils - utils.c:mkdir_p:253 - Invalid argument - failed to create directory '/sys/fs/cgroup/net_cls//lxc/default-999' | 20:44 |
krnlyng__ | /system/bin/adb doesn't seem to work | 20:44 |
krnlyng__ | at least for me | 20:44 |
krnlyng__ | saidinesh5, container-manager is started with --privileged? | 20:44 |
saidinesh5 | yups: root 4046 0.0 0.2 57276 4960 ? Ssl 02:09 0:00 /usr/bin/anbox container-manager --privileged | 20:44 |
*** piggz_ <piggz_!~piggz@95.147.209.146> has quit IRC (Ping timeout: 272 seconds) | 20:50 | |
saidinesh5 | krnlyng__: what does your ls -l /sys/fs/cgroup/net_cls look like? | 20:51 |
*** piggz_ <piggz_!~piggz@95.147.209.146> has joined #sfdroid | 20:54 | |
*** nh1402 <nh1402!~nh1402@212.182.159.143.dyn.plus.net> has quit IRC (Quit: Leaving) | 20:54 | |
krnlyng__ | lrwxrwxrwx 1 root root 16 Jan 4 1970 /sys/fs/cgroup/net_cls -> net_cls,net_prio | 20:55 |
krnlyng__ | container currently not running | 20:56 |
saidinesh5 | and what about /sys/fs/cgroup/net_cls/lxc ? | 20:56 |
mal | maybe I'll continue some other day, don't know what to try next | 20:56 |
saidinesh5 | Aye.. maybe by next time we'll have an adb too? | 20:57 |
saidinesh5 | drwxr-xr-x 2 root root 0 May 16 02:09 lxc | 20:57 |
saidinesh5 | ? | 20:57 |
mal | krnlyng__: what did you mean by the blobs you mentioned earlier? are there some blobs somewhere? | 21:01 |
krnlyng__ | mal, in order to render hw accelerated we use vendor blobs from the host system | 21:03 |
krnlyng__ | mal, thats the difference between anbox and anbox+sfdroid | 21:03 |
krnlyng__ | mal, because pure anbox would be too slow on a mobile device | 21:03 |
mal | ok | 21:03 |
krnlyng__ | mal, maybe not too slow but slower | 21:03 |
saidinesh5 | mal: /var/lib/anbox/overlay/ that contains copies fo libraries form your system | 21:03 |
krnlyng__ | ok libwayland is built | 21:04 |
mal | ok, but the rendered is not built againsta those? | 21:04 |
krnlyng__ | properly now i hope | 21:04 |
krnlyng__ | android-wlegl and surfaceflinger in progress | 21:04 |
saidinesh5 | and it is overlaid on top of the anbox's rootfs and the surfaceflinger then makes use of those libraries to render things | 21:04 |
krnlyng__ | mal, no and thats kind of a hack. for some devices we need an additional layer maybe | 21:04 |
mal | ok, maybe that could be a problem then on fp2 | 21:05 |
krnlyng__ | mal, especially for gralloc, other libs are ok unless they have undefined symbols | 21:05 |
krnlyng__ | mal, yes | 21:05 |
krnlyng__ | mal, we will build adb soon | 21:05 |
mal | I get no errors anywhere | 21:05 |
krnlyng__ | mal, then you can check with logcat | 21:05 |
mal | ok | 21:06 |
krnlyng__ | mal, but it's cool that the container starts | 21:06 |
krnlyng__ | mal, thats the most important thing | 21:06 |
krnlyng__ | mal, the other problems can be fixed | 21:06 |
mal | ok, the UI side does nothing, from UI it just has the waiting indicator for a while and then goes away, from commandline the same command starts but nothing happens | 21:07 |
krnlyng__ | mal, remind me tomorrow to build adb then we can check | 21:07 |
krnlyng__ | mal, then i will either implement a "fix" or a linker workaround | 21:08 |
krnlyng__ | depending on whats easier | 21:08 |
krnlyng__ | but i have to sleep now | 21:08 |
krnlyng__ | good night:) | 21:08 |
saidinesh5 | Ah nighters krnlyng__ | 21:08 |
mal | I won't be available tomorrow | 21:10 |
saidinesh5 | oh dont worry, im here to remind :p | 21:11 |
krnlyng__ | i also noted it down in sfdroid todo | 21:13 |
saidinesh5 | krnlyng__: what''s the boot mess? | 21:15 |
saidinesh5 | Ahh boost mess | 21:15 |
saidinesh5 | nvm | 21:15 |
saidinesh5 | i need sleep now | 21:15 |
saidinesh5 | that's already in the todo :p | 21:16 |
saidinesh5 | hmm.. odd.. mkdir: cannot create directory `/sys/fs/cgroup/net_cls//lxc/default': Invalid argument | 21:24 |
*** piggz_ <piggz_!~piggz@95.147.209.146> has quit IRC (Ping timeout: 260 seconds) | 21:52 | |
*** pseudodev <pseudodev!uid205973@gateway/web/irccloud.com/x-kihrpfdjdxrbqgcz> has quit IRC (Quit: Connection closed for inactivity) | 21:55 | |
*** Mister_Magister <Mister_Magister!~foidbgen@user-94-254-147-32.play-internet.pl> has quit IRC (Ping timeout: 258 seconds) | 22:35 | |
*** TheKit <TheKit!~nekit@178-45-4-8.saransk.ru> has joined #sfdroid | 22:41 | |
*** NeKit <NeKit!~nekit@178.45.144.138> has quit IRC (Ping timeout: 240 seconds) | 22:44 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!