Thursday, 2019-01-31

*** doniks_ is now known as doniks00:34
*** ChanServ sets mode: +v T402:47
kimmolirinigus: ok. i'll take a look after c0ffee03:00
kimmolirinigus: we have a green - i wait merbot to shutup about the patterns and lock, and trigger CI03:57
kimmoliaaand it's in the pipeline
riniguskimmoli: thank you very much, will test tonight05:35
rinigusOTA though...05:35
riniguskimmoli: you may want to trigger ci again, obs was still building whole ci was working. Not sure why05:38
kimmoliyes, i did build obs again with merged config05:48
kimmolirinigus: rebuilding CI05:54
kimmolirinigus: i'm doing ota as soon as my test-device gets some charge05:54
dr_gogeta86good morning06:31
kimmolirinigus: brief testing looks good06:42
T4<adampigg> Gm dr_gogeta8606:58
riniguskimmoli: great to hear :)07:05
*** ChanServ sets mode: +v T408:56
HelloPotter_hey there i have image magick installed, but when building it's saying no image magick is installed10:00
HelloPotter_please help10:10
riniguskimmoli: I've updated, rebooted as usual. while idle, I had already 2-3 random reboots. just enabled persistent journal, let's see if it can catch the cause.10:39
kimmolirinigus: i have uptime of >4 hours10:39
riniguskimmoli: 21 min over here so far. but haven't checked it before.10:41
kimmolii meant to say no random reboots10:41
rinigusunderstood :)10:42
malHelloPotter_: when doing what?10:45
HelloPotter_@mal when building hybris-hal, by the way i just solve it , one thing mal, if you dont mind  can you please give the solution again of libsf_compat_layer', needed by 'hybris-hal', missing and no known rule to make it, because i just forget how to fix that10:53
malHelloPotter_: which android base?11:27
malif 15.1, then clone libhybris to some suitable place, either external/ or hybris/mw/? git clone --recurse-submodules -b android8-initial11:27
HelloPotter_@mal yes it 15.1, thanks11:29
riniguskimmoli: rebooted again. uptime about 1h 20min. seems to occure during email sync if I assume the same pattern happens every 5 minutes or so. journal section at . let's see if the same will happen in 1-2 hours11:44
kimmolirinigus: ok. i had no emails active, now activated gmail.11:47
riniguskimmoli: cannot exclude failing hardware either. I had a strange crash (before updates). hasn't happened earlier, but maybe related with the current reboots. although, I would expect (hope) that these reboots are something else. let's see if you can reproduce them with active email11:51
kimmolirinigus: ok. my other onyx has worked mostly without hickups, last time i had to reboot was due missing alarms11:52
kimmoli.e. didn't wakeup from sleep on alarm11:52
T4<elros34> rinigus: check ramoops, few kernels have problem with new firewall rules which cause kernel panic in tcp_mt at xt_tcpudp.c11:55
HelloPotter_@mal hey there i am littlebit confuse about the section 5.4.1(Kernel Config) on sailfish os hardware adaptation developent kit,  i think that i need to verfy the defconfig file which is used to make boot image, so i go to the device/motorola/potter/ file and checked that the file name is potter_defconfig, whic source is kernel/motorola/msm8953/, but i see that there is no file called potter_defconfig12:10
malHelloPotter_: it should be in some subfolder arch/arm64/configs probably12:11
malassuming you have correct kernel in use12:11
HelloPotter_ok im checking12:12
HelloPotter_@mal yes it's there12:13
HelloPotter_@mal by following the official guide when i trying to do make -c kernel/motorola/msm8953/arch/arm64/configs/potter_defconfig it's saying unknown option -c12:16
malHelloPotter_: you don't do that12:19
malHelloPotter_: only make hybris-hal12:19
HelloPotter_@mal i already make hybris-hal12:20
malthat all you need for now12:21
malit will include building kernel12:22
HelloPotter_@mal one more thing when i use mer-kernel-check utility it's showing me some errors like this ERROR: CONFIG_DUMMY is invalid12:24
HelloPotter_ERROR: CONFIG_VT is invalid12:24
HelloPotter_ERROR: CONFIG_DEVTMPFS is invalid12:24
HelloPotter_ERROR: CONFIG_DEVTMPFS_MOUNT is invalid12:24
HelloPotter_ERROR: CONFIG_NLS_UTF8 is invalid12:24
HelloPotter_ERROR: CONFIG_FHANDLE is invalid12:24
HelloPotter_ERROR: CONFIG_SYSVIPC is invalid12:24
maldid you define those in defconfig?12:25
HelloPotter_did i need to deine them?12:25
malHelloPotter_: obviously if those are claimed as errors then you define those as instructed by kernel checker12:33
HelloPotter_@mal like this one ERROR: CONFIG_SYSVIPC is invalid12:39
HelloPotter_It is unset12:39
HelloPotter_Allowed values : y12:39
HelloPotter_Comment says: Inter Process Communication option is required to run Mer12:39
HelloPotter_@mal did i need set it's value as y12:40
HelloPotter_@mal after editing those should i again need to compile the hybris-hal12:43
malHelloPotter_: yes12:45
HelloPotter_ok i will edit the out/target/product/$DEVICE/obj/KERNEL_OBJ/.config file and recompile the hybris-hal12:47
HelloPotter_no no sorry i need to edit the potter_defconfig file12:48
riniguskimmoli: n=2, exactly the same line in the log, just after carddav-client message12:57
kimmoliok, my is still running12:57
rinigusbtw, where are carddav accounts defined. I suspect its using an account that I deleted some time ago and is not anymore in gui either.12:57
riniguskimmoli: what is the time interval between the checks for your email? mine is 5 min - could be related to # times its called12:59
rinigusok, then its probably not that. let's see for a bit longer...13:00
malHelloPotter_: yes, you edit defconfig, you only check .config that the changes were correctly included in the build, you never edit .config13:05
kimmolirinigus: ill check my journals13:06
malrinigus: kimmoli just curious, hopefully you don't get the network crash issue in onyx when using, fp2 doesn't seem to have it13:07
kimmolii have only LTE active13:08
rinigusmal: it could be network related. it happens while checking email or carddav. I'm on LTE as well. how can I test for that network crash condition?13:08
malrinigus: the reports so far have said that heavy usage of network can cause kernel panic, there is bad workaround if that is the case13:10
T4<elros34> check dmesg13:10
malI have so far used on fp2 for several without a single crash, I have used maps and other online services during that time13:11
rinigusmal: haven't used network excessively, phone is mainly just sleeping. as for dmesg (if its for me) what should I check?13:11
mal*several days13:11
kimmoliyeah we have it13:12
kimmolijust rebooted when opening youtube and instagram on tabs13:12
malyes, web browsing can trigger it easily13:12
kimmolistock browser, just two random sites which are bloated13:12
T4<elros34> rinigus: check ramoops, few kernels have problem with new firewall rules which cause kernel panic in tcp_mt at xt_tcpudp.c13:13
malI tried to play youtube videos for quite a while and no crashes13:13
T4<elros34> its ipv6 related13:13
rinigusjust managed to force crash via youtube as well.13:14
rinigusand the browser.13:15
rinigusso, what's the workaround?13:15
T4<elros34> remove /etc/connman/firewall*13:15
malrinigus: the "proper" workaround is to make config repo provide the package that provides those configs and only add the main.conf of that package to config sparse13:16
malafaik that should work13:17
HelloPotter_@mal i just fixed those errors,  but i have now only two errors CONFIG_DEVTMPFS_MOUNT and CONFIG_DUMMY, which both of them are not exist in the defconfig file, similar are exist in the defconfig are  CONFIG_DUMMY_IRQ and CONFIG_DEVTMPFS13:17
riniguselros34: reading irc log - how to check for ramoops?13:17
HelloPotter_@mal what should i do at this point13:17
malHelloPotter_: like I said, if something required is missing you add the flag to defconfig with correct value13:18
T4<elros34> depends on kernel: /proc/last_kmsg ir something in /sys/pstore/13:18
rinigusmal: you mean that /etc/connman/firewall* should be in a separate package that we don't pull in onyx?13:18
HelloPotter_@mal mean i can just add CONFIG_DEVTMPFS_MOUNT=y  in the defconfig13:19
malrinigus: you add Provides: connman-configs-sailfish to droid-config spec and then add etc/connman/main.conf from that package to config sparse13:20
malrinigus: that should do it, hopefully13:21
HelloPotter_thank u , and more thing  did i just ignore all the warnings, or just carefully read the comments and set the flag properly13:23
riniguselros34: indeed, in last_kmsg there is a lot of bitterness about ip6 in kernel panic message.13:24
rinigusI will try to remove filters and see if it works. probably it does and then we can fix it properly via droid-config13:24
T4<elros34> does it crash in tcp_mt function?13:24
malHelloPotter_: make sure you keep CONFIG_AUDIT=y and use CONFIG_SECURITY_SELINUX_BOOTPARAM=y (and set selinux=0 to kernel commandline in device repo
HelloPotter_ok i will check them13:27
riniguselros34: here is the log - . doesn't look like tcp_mt13:31
malrinigus: but ipv6 related anyway13:34
T4<elros34> doesn't reach tcp_mt but otherwise looks similar, anyway offset value is incorrect in ip6t_do_table so it doesn't matter.13:35
riniguselros34: thanks for the insight!13:35
malrinigus: maybe check fp2 kernel (fp2-m-sibon branch) to see if you can see anything related difference there13:36
T4<elros34> this is kernel workarnound I tested since yesterday: Looking at fp2 kernel proper fix its probably here
rinigusmal and elros34: could check what's the state in onyx kernel over there.13:43
riniguslater tonight, though.13:43
malrinigus: that commit is in onyx already13:45
malrinigus: fp2 has two other commits in netfilter though13:50
rinigusmal: thanks!14:01
mal@elros34 rinigus if you are interested, try building a patched connman (checkout the upgrade-3.0.1 branch from mer-core and apply this patch
malif that is not enough also try to patch iptables
T4<elros34> thx I will test it latter14:04
HelloPotter_@mal hey there you said that i need to set selinux=0 to the kernel commandline, u mean in this flag BOARD_KERNEL_CMDLINE14:41
malyes, like this
malthe exact file depends on the device but in lineage it's usually BoardConfig.mk14:45
HelloPotter_ok i got that14:47
malyou should find similar lines in some file14:50
HelloPotter_@mal in boardconfig file  my whole kernel section BOARD_KERNEL_CMDLINE := console=ttyHSL0,115200,n8 androidboot.console=ttyHSL0 androidboot.hardware=qcom user_debug=30 msm_rtb.filter=0x23714:55
HelloPotter_BOARD_KERNEL_CMDLINE += ehci-hcd.park=3 androidboot.bootdevice=7824900.sdhci lpm_levels.sleep_disabled=1 vmalloc=350M14:55
HelloPotter_BOARD_KERNEL_CMDLINE += selinux=014:55
HelloPotter_BOARD_KERNEL_BASE := 0x8000000014:55
HelloPotter_BOARD_KERNEL_IMAGE_NAME := Image.gz14:55
HelloPotter_BOARD_KERNEL_PAGESIZE := 204814:55
HelloPotter_BOARD_KERNEL_SEPARATED_DT := true14:55
HelloPotter_TARGET_CUSTOM_DTBTOOL := dtbTool_custom14:55
HelloPotter_BOARD_DTBTOOL_ARGS := --force-v3 --motorola 114:55
HelloPotter_BOARD_MKBOOTIMG_ARGS := --ramdisk_offset 0x01000000 --tags_offset 0x0000010014:56
HelloPotter_TARGET_KERNEL_ARCH := arm6414:56
HelloPotter_TARGET_KERNEL_HEADER_ARCH := arm6414:56
HelloPotter_TARGET_KERNEL_CONFIG := potter_defconfig14:56
HelloPotter_TARGET_KERNEL_SOURCE := kernel/motorola/msm895314:56
HelloPotter_TARGET_KERNEL_CLANG_COMPILE := true14:56
HelloPotter_TARGET_USE_SDCLANG := true14:56
HelloPotter_is this correct14:56
HelloPotter_i just copied BOARD_KERNEL_CMDLINE += selinux=0 the line from your given link and paste it after the second line14:58
malHelloPotter_: never paste more than 3 lines to the channel14:59
maluse pastebin for large amounts of text14:59
*** ChanServ sets mode: +v T415:00
HelloPotter_@mal ok sorry for that15:01
malanyway that seems fine15:02
HelloPotter_@mal now should i look into those warnings or just ignore them15:03
malyou can skip those for now15:04
HelloPotter_@mal hey there i am getting confused about installing scratchbox2,  beacause my arch is arm6415:38
malPORT_ARCH is always armv7hl15:41
malfor arm devices15:41
HelloPotter_is this will run on arm6415:45
r0kk3rzyeah, its like a 32bit mode of aarch64 devices15:47
malHelloPotter_: just use armv7hl like I said15:47
malalso make sure it's defined in the .hadk.env file you created15:48
HelloPotter_@hey there i previously maked a mistake in .hadk.env file i set the PORT_ARCH="arm64", now i made the the changes, after that i deleted the whole output directory, but  when i am doing breakfast potter, it's saying TARGET_ARCH=arm64, is there any way to fix this15:58
T4<elros34> try source ~/.hadk.env or quit sdk and enter to it again16:04
malHelloPotter_: do not confuse android side build with .hadk.env16:07
malHelloPotter_: .hadk.env is for sailfish side of things16:07
HelloPotter_i tried all of them, but didn't work, can i change the achitechture BoardConfig.mk17:08
malHelloPotter_: NO!17:28
malHelloPotter_: what are doing now? I don't understand the problem17:29
malwe told you to change the PORT_ARCH correctly, then exit sdk and go back, that's it17:29
HelloPotter_i exited the sdk , changed the PORT_ARCH int the .hadk.env file, then re entered the sdk my arch in breakfast result still not changed17:31
r0kk3rzthats fine17:32
HelloPotter_i dont understand if i make the hybris-hal in TRAGET_ARCH=arm64 and then use the tooling and sdk of armv7hl, there is no problem?17:34
r0kk3rzyeah, its kinda mixed mode17:37
malHelloPotter_: TARGET_ARCH is the android side arch, not sailfish arch17:39
malthose are not always the same17:39
HelloPotter_ohh, now i understand, sorry i get too much confused17:41
HelloPotter_hey there when i trying to do sb2 -t $VENDOR-$DEVICE-$PORT_ARCH gcc main.c -o test its saying sb2: Error: Invalid target specified, aborting.18:10
maldid you read the warning box in hadk pdf before creating the target?18:13
malthe instructions in wiki use different naming and the warning box tells what to use instead18:14
HelloPotter_so i need to change the tooling name18:16
HelloPotter_and target name must be motorola-potter-armv7hl18:17
HelloPotter_i did this time correctly but his time no out in the console, but there are no errors also18:28
HelloPotter_sorry i got the output18:30
riniguskimmoli: so far, so good. as expected, stability can be gained through dropping of filters. I can look (probably on the weekend) if commits similar to fp2 kernel ones can help. although, on the first look, they seem to be different in lineageos onyx kernel19:12
malrinigus: did you try the connman patch?19:12
malrinigus: (and in case that not enough also the iptables patch)19:13
rinigusmal: no, I didn't. haven't had a chance to work on it. do you think its a better way than to patch the kernel?19:13
HelloPotter_@mal   rpm/dhd/helpers/ it's saying launch this script from the /home/err0r/Desktop/New_Oss/Sailfish_Os directory19:21
HelloPotter_@mal but i am in PlatformSDK err0r@err0r:~/Desktop/New_Oss/Sailfish_Os$19:21
HelloPotter_it is in the section of creating repositories for new devices19:23
piggzmal: seems quite dated ...requires bluez4!19:28
piggzi feel a fork is needed19:28
piggzmal: i dont even think it needs bluez to build ?19:32
malpiggz: it seems the  brcm_patchram_plus_usb.c might need bluez19:35
malbut see how badly build breaks with bluez519:36
piggzwhy are there 2 versions of the code? it looks like the spec builds the older one19:36
malpiggz: other is upstream, same thing19:37
piggzmal: there is a difference between the 2 dirs..19:39
piggzmainly around forking19:39
piggzmal: well, it builds fine19:44
piggzand i wont be testing the usb part19:45
malpiggz: ok, then it should be fine to use19:48
malpiggz: just fix the dependencies in spec if needed, or add another spec19:48
piggzmal: it seems to work too, even though my version was newer...20:09
piggzmy version seems to just add support for rfkill20:10
*** ChanServ sets mode: +v T420:21
piggz@eugenio_g7 pushed config changes for gst and BT to my repo20:27
piggzone day we will have to better combine and work from a common repo!20:27
HelloPotter_@mal should i configure middleware now20:31

Generated by 2.17.1 by Marius Gedminas - find it at!