Thursday, 2019-06-13

KALUbEI've fully rebuilt everything from scratch, hopefully fixing a few minor errors but still can't telnet. Here's my init.log
KALUbEporting to eva.01:15
KALUbEIt also seems to automatically disconnect from rndis after about a minute.01:15
*** OhYash1 is now known as ohyash06:19
masha11Hello! What do I need to fix in order to start test_hwcomposer and lipstick successfully? Device reboots after starting test_hwcomposer.07:16
masha11Here are my logs. strace test_hwcomposer: strace lipstick:
masha11dmesg:, journalctl:
T4<birdzhang> @masha11 line 67007:31
T4<birdzhang> oh, wait, are you porting 15.1 ?07:31
sledgesPSA: Sailfish OS community collaboration in the #mer-meeting channel is starting in a half-hour07:32
masha11@birdzhang 14.107:51
T4<birdzhang> okay, ignore me :P07:51
mal@adampigg you seem to have some issue in mesa spec, you cannot define same driver in both dri-drivers and gallium-drivers11:13
T4<adampigg> mal: yeah, its wip, making changes to spec and building on obs....11:23
T4<adampigg> mal, the interesting q is...what drivers to build?11:29
mal@adampigg those that we think might be useful, current list looks ok for now11:43
mal@adampigg check fedora spec for some ideas where each one goes ref the bug you had
T4<adampigg> yup,/thats  my reference anyway12:16
T4<adampigg> will carry on tonight12:16
T4<adampigg> mal: did u see my issue installin12:16
T4<adampigg> target?12:16
mal@adampigg probably your tooling hasn't been updated and is some older version, updating toolings is a bit of a issue12:29
T4EdwardHack was added by: EdwardHack12:37
T4<adampigg> mal, i suspect that....need to see how to update tooling given that it says its installed13:23
mal@adampigg install it with other name?13:27
malthen the new target will use the new one13:27
T4GodcraftLinXkin was added by: GodcraftLinXkin13:53
T4hassaan_24 was added by: hassaan_2414:16
T4ga1lan was added by: ga1lan15:01
KALUbEYo, any ideas on /sys/class permission issues in init-debug? I have set init_enter_debug2, this is my init.log Huawei P9, reboots instantly on telnet/ping16:05
T4<elros34> do you have configfs enabled in kernel?16:12
T4<elros34> you could try to change order of this if so it will use usb_setup_configfs. Then make hybris-hal and change the same in /data/.stowaways/sailfishos/init-debug16:33
KALUbEThat didn't work, it now waits for ~60s then reboots, and nothing shows up in dmesg or lsusb -v16:45
T4<elros34> show you device repo16:46
KALUbEI'm using branch android-7.1.1_r22. Just wondering if that could be an issue vs 7.x branch...16:49
KALUbENo, I don't think it is.16:50
T4<elros34> try to replace every occurence of rndis_bam.rndis with rndis.gs4 or add similar lines to rndis.gs4 in init-script16:50
T4<elros34> I mean similar lines to these which contains rndis_bam.rndis16:51
KALUbEJust replace rndis_bam.rndis with rndis.gs4?16:54
T4<elros34> yes, you device use rndis.g4 but I am not sure it's relevant16:56
KALUbEAh I see. I also have an issue where the init-debug that get's generated still contains "DATA_PARTITION=%DATA_PARTITION%" and doesn't get updates16:57
malKALUbE: note the difference between init-debug and the init-script in kernel ramdisk16:57
malnot the same thing16:57
KALUbEI'm adding the changes to hybris/hybris-boot/init-script, compiling and flashing hybris-boot.img and then modifying init-debug to match16:58
malthe partitions are relevant only in the kernel ramdisk16:58
KALUbEoh ok16:58
T4<elros34> I dont't think it's important. Mounting data partition is done in initramfs DONE_SWITCH=no in script16:58
malthose are generated from same source but only the one going to to kernel ramdisk has partitions replaced16:59
KALUbEAh ok, not an issue then.17:00
KALUbEI've booted into lineage and /sys/class/android_usb/android0 DOES exist, I also don't get anything in dmesg when using usb_setup_configfs17:06
KALUbEBefore DONE_SWITCH, it does write to iSerial that telnet is on port 23, but it seems that after, INIT_DEBUG for some reason can't write to the files.17:08
malKALUbE: does lineage show anything under that /sys/class/android_usb/android0 and does it show configfs related things in /config/usb_gadget/g1/functions/ ?17:09
KALUbEWhilst lineage flashes. I've previously added "ls /sys/class/android_usb/android0" to init-debug (in usb_setup_android_usb) and got the following:
KALUbEOk here's the contents of the two folders on Lineage:
T4<elros34> port_mode is set to 9 in configfs mode here
KALUbEShould that be set in init-debug?17:28
T4<elros34> worth a try17:28
KALUbEIs it possible that is uses configfs in the bootloader but init-debug switches to android_usb for some reason?17:28
KALUbEBecause it's able to set the serialnumber initially, and tells me that telnet is on port 23. But the init.log says it's calling usb_info and setting telnet to port 2323, but it doesn't change on my PC17:29
KALUbEAnd as I just changed hybris init-script to switch configfs and android_usb it failed to do it initially as it now wasn't using configfs initially where it was before?17:30
T4<elros34> you could check logs in initramfs to see what is really set17:31
KALUbEHow do I get the logs from initramfs? Do they get written?17:31
KALUbEDamn, I need to get food. Back shortly.17:31
KALUbEOh I see, it prevents init-debug from overwriting the log?17:39
T4<elros34> not it should prevents switching to sfos rootfs so it will stay in initramfs with telnet  and mass storage enabled17:42
KALUbEAh ok, I just reflashed, left hybris-boot init-script default but modified init-debug and it succesfully updated to tell me telnet was on 2323 this time17:45
KALUbEHowever it still reboots when I try and telnet in17:46
KALUbEThe iSerial value changes to (error) and then it reboots.17:48
KALUbEOk so telnetd is now running in rootfs which it wasn't before, but it still reboots when I try and connect or ping.18:00
T4<SALPHA8506> anyone porting here for zuk z2_plus ?18:08
piggzmal: i little more work to do on the spec
malpiggz: you'll need to add some if x86 etc to meson command19:12
malfor those i9x5 drivers19:12
malpiggz: also same to the dri-drivers package file list19:13
malpiggz: probably should also limit the arm drivers the same way fedora does19:14
malpiggz: near the beginning of fedora spec19:15
malpiggz: not sure if we should use those with_* variables or just use archs everywhere19:16
piggzmal: yeah, ive seen how they do it19:21
malpiggz: just using ifarch directly might be simpler19:28
piggzmal: sure, that will be v1 anyway!19:28
piggzmal: or, could just copy most of the fedora spec?19:29
piggzit is kinda neat19:29
malpiggz: I think it's a bit overkill to have a separate variable for each driver19:29
malwhen most of those are controlled just based on build arch19:30
malbut that's just my opinion19:30
piggzmal: it has the advantage that if we just decide to enbable/disable a driver, it is a 1 or a 019:31
maltrue, you can just copy the stuff for first version19:31
piggzmal: i'll simplify it a bit19:36
malpiggz: a small comment also, move the libgbm file list earlier so the order of package definitions and file lists is the same19:37
malpiggz: looks like a small issue in arm build20:30
piggzmal: ah, probably doesnt work like a bool?20:32
piggzbest to remove the with_tegra 0 completely20:32
malpiggz: no, you need to enable nouveau driver20:32
malif you want to build also tegra20:32
malpiggz: also fix the error from x86 build20:34
malunpackaged file20:34
piggzmal: i586 success....21:09
malyep, looking good21:09
malpiggz: btw, mesa 19.1 was just released couple of days ago, maybe rebase to that later21:12
piggzpiggz@linux-f1uu ~/mer/android/droid.dontbeevil/hybris/mw/mesa/mesa ((HEAD detached at mesa-19.1.0-rc2)) $ git checkout mesa-19.1.021:12
malpiggz: looks like two unpackaged files in arm build so you have to fix it anyway so good time to rebase21:13
piggzmal: minor fix needed, so will update to 19.1 at the same time21:13
piggzmal: i _hope_that works as im going to sleep now! zzzZZZ21:18
piggzwill fix tomorrow if not :)21:18
piggzmal: rats, also need to add a tag to the repo to get a nice version21:20
malpiggz: the commit history also needs some cleanup21:20
piggzk .. gnight :)21:21
nylI installed messwek i don't see any feedback strange21:49
nylCan't find the irc logs when someone configured sensors in hybris21:49
nylvgrade around?21:53
malhow about you search faq for "sensors"21:53
malit has nice instructions what to do to fix one most common problem with sensors21:59

Generated by 2.17.1 by Marius Gedminas - find it at!