Sunday, 2020-03-22

*** OhYash1 is now known as ohyash08:56
Nebrassy[m]tried to build in Pie again, same issue I had on oreo, device instantly reboots, console ramoops are of the previous twrp session (I know for a fact this kernel supports ramoops)11:10
Nebrassy[m]I noticed the kernel doesn't include any dtbs so no early mounts11:10
Nebrassy[m]also what's DATA_PARTITION= in the ramdisk supposed to point towards?, I might have wrong mountpoints11:10
T42<elros34> it's should points directly to your data partition so /dev/mmcblk0px or /dev/sdx but unlikley device wil reboot straightaway if it's wrong. What device? Did you revert skip_initramfs support in kernel.11:15
Nebrassy[m]ok I definitely have wrong mountpoints, device is jasmine_sprout, here's all
Nebrassy[m]I also don't see where it mounts system or vendor anywhere11:17
Nebrassy[m]I usually use this to build twrp, don't know if I need the revert for sailfish
Nebrassy[m](also tried adding audit=0 to cmdline)11:19
Nebrassy[m]and, isn't it supposed to be /dev/block/mmcblkXX instead of /dev/mmcblkXX?11:19
T42<elros34> @Nebrassy which branch in kernel is the one you use now?11:32
T42<elros34> disable CONFIG_VT in kernel, flash it and try again11:36
T42<elros34> not very important right now but your droid-hal-device.conf is wrong read again comment you put there "WRITE11:51
T42<elros34> ˓→ALL IN ONE LINE(:"11:51
*** OhYash1 is now known as ohyash12:12
Nebrassy[m]ah didn't notice, thanks12:13
T42<elros34> just to be sure: you know that you can just flash hybris-boot, you dont need to build whole new image12:21
Nebrassy[m]yup, just building hybris-hal12:22
Nebrassy[m]I need to flash manually anyway, my twrp doesn't have busybox/bunzip12:23
Nebrassy[m]it stays live for a bit longer after disabling CONFIG_VT, and got ramoops
T42<elros34> ok so enable CONFIG_VT back and revert: 6c8294064937459a6a0e79b57e01fde0efd554d0 dd96daf41e0ab5c87171b3a55cab2d3ae9e5ee0e Do you have /data/.stowaways/.sailfishos/init.log n12:38
T42ow? How long it stays on now?12:38
Nebrassy[m]used to be like a second, now more like 812:39
Nebrassy[m]and nah no logs in data, pretty sure it isn't reaching mount yet12:40
Nebrassy[m]what exactly should I revert?12:42
Nebrassy[m]nvm didn't see12:42
Nebrassy[m]kernel/xiaomi/sdm660/drivers/tty/tty_port.c:607:34: error: no member named 'worker_thread' in 'struct tty_port'12:59
Nebrassy[m]        return sched_setscheduler(port->worker_thread, policy, &param);12:59
Nebrassy[m]also revert
T42<elros34> probably. The faulty commit is "tty: move tty_port workqueue to be a kthread" but other depends on these changes13:01
Nebrassy[m]it's the same merge so yeah, trying13:02
Nebrassy[m]had to rever too13:25
Nebrassy[m]also unable to open an initial console13:25
T42<elros34> that shouldn't be an issue. So now you have enabled VT?13:27
T42<elros34> pfk initialize for you  but I remember other with similar issue to your have to disable CONFIG_PFK so it's worth to try. Also enable CONFIG_FW_LOADER_USER_HELPER back13:30
Nebrassy[m]hm idk why I thought I didn't disable it, I'll try13:32
T42<elros34> looks like it crash somewhere in init script when mdev is used13:49
Nebrassy[m]revert latest one?13:52
T42<elros34> I would try, you can also compare with some other working kernel (same version)13:53
Nebrassy[m]worked, got telnet
Nebrassy[m]should I keep CONFIG_PFK disabled?14:23
T42<elros34> you can keep it disabled it shouldn't matter14:24
Nebrassy[m]what's the root password for devel-su14:25
T42<elros34> in telnet  you are already root user14:25
Nebrassy[m]right, thanks for the help14:27
Nebrassy[m]so what's next?14:45
T42<elros34> grab full journalctl --no-page14:46
T42<elros34> looks like wrong fixup-mountpoints: compare your changes with your fstab14:53
Nebrassy[m]ah yup, guessing it's time to rebuild everything14:56
Nebrassy[m]already fixed mountpoints just only build hybris-boot after it14:56
T42<elros34> before creating new image you can unpack droid-hal*rpm and check whether /lib/systemd/system/*.mount contains correct partitions (without by-name)14:59
Nebrassy[m]I don't have vendor in fstab, since it's early mounted, should I add it?18:08
T42<elros34> or just create vendor.mount and symlink to enable it and put it to sparse18:19
T42<elros34> by default mount units are enabled by symlink in /etc/systemd/system/ by droid-hal rpm post install script18:25
Nebrassy[m]that saved me a rebuild but I still rather just add to fstab18:39
malsome new devices need a custom vendor mount service because those have also other things mounted under /vendor18:44
malso the mount order has to be correct18:44
T42<elros34> mal: If I understand systemd manual correctly custom vendor shoudn't be needed: "If a mount unit is beneath another mount unit in the file system hierarchy, both a requirement dependency and an ordering dependency between both units are created automatically."18:46
Nebrassy[m]that'd be me then as dsp, bluetooth, persist and modem are mounted in vendor18:46
mal@elros34 hmm, is there some mention about which version has such support18:49
T42<elros34> no, I found it here
mal@elros34 looking at git blame of systemd that should be a very old feature, probably even an original feature18:54
Nebrassy[m]looks like it mounted persist first18:54
T42<elros34> it's about /vendor/dsp and /vendor/bt_firmware not /persist18:56
Nebrassy[m]wait that's /mnt/vendor/persist, why18:58
Nebrassy[m] vendor.mount: Directory /vendor to mount over is not empty, mounting anyway.18:58
Nebrassy[m]so what went wrong here, missing libs?19:05
T42<elros34> what do you mean? mounting over not empty directory is not an error, check order which partitions are mounted19:06
Nebrassy[m]the order seems correct, I mean what's the problem for display19:09
Nebrassy[m]allocator and cas crashes19:13
vknechtyou didn't disable selinux ?19:13
Nebrassy[m]I did19:14
piggzmal: i come bearing bluebinder/bluez logs
Nebrassy[m]do I need to build droidmedia?19:23
malpiggz: you should know to provide full logs :)19:24
T42<elros34> @Nebrassy android service fails for some reason so droidmedia will probably not help. Some firmware fails to load like tas2557_uCDSP.bin or modem.mdt. Do you have such a files anywhere?19:27
Nebrassy[m]those two are meaningless19:28
piggzmal: knock yourself out
malpiggz: also logcat can have bluetooth related things when using bluebinder19:32
T42<elros34> @Nebrassy if you say so. Check if you have uevent* file somewhere in /vendor. Also some device needds custom ldconfig but I don't know the details:
piggzmal: thers are bt related logs at the same time....
Nebrassy[m]I have /vendor/ueventd.rc19:42
T42<elros34> compare it with uevent*rc in out/*/root, maybe rules are missing. These files are used to generate udev rules19:46
piggzmal: i dont know about you, but i cant see anything substantiol in the logcat log19:46
T42<elros34> @Nebrassy I see you also use master dhd submodule. Not sure if it's still required but so far recommended branch is  upgrade-3.2.119:54
T42orboti2p was added by: orboti2p21:00
walidhwhen i ran got this error error: Bad exit status from /var/tmp/rpm-tmp.23FnvD (%build)21:52
vknechtwalidh, impossible to tell with so little information... depending on the package being built, there's probably a related log file22:12
malwalidh: so did you run without the -o switch, I think I said to use that22:16
walidhmal: i try out with cm14.1, got system boot , but hwcomposer not work, no gui22:30
walidhi'm try now with lineageos15.122:31
malwhy do people so often start changing android base after some issues22:39
vknechthope :-)22:42
walidhmal: i have ported LG V20 for ubuntu touch, and work well with cm14.1.22:44
walidhFor SFOS i want at begin to use LOS15.1,22:45
T42<edp_17> Hi all!23:39
T42<edp_17> I have tried to make the MTP work on my S2 port. I used the hammerhead:
T42<edp_17> To port into the S2's kernel:
T42<edp_17> But then I got a few errors at make hybris-hal:
T42<edp_17> Any help is much appreciated. Thanks.23:42
T42<edp_17> I probably missed something from the MTP fix for the hammerhead.23:45

Generated by 2.17.1 by Marius Gedminas - find it at!