Wednesday, 2018-07-18

*** redj_ is now known as redj02:53
*** zama_ is now known as zama05:20
*** ubuntu is now known as Guest4638306:51
*** NeoChapay_ is now known as NeoChapay08:44
NeKitsledges, could you please create xiaomi-hermes target on OBS? (Xiaomi Redmi Note 2)09:13
ohyashin the sb2 installation it shows links for SDK_tooling, SDK_target{armv7hl,i486} but I have set my PORT_ARCH="arm64-v8a"12:47
ohyashdoes that mean I should've stayed with armv7hl in the beginning with PORT_ARCH in ~/.hadk.env?12:47
r0kk3rzport arch should be armv7hl12:49
ohyashdamn. Gonna have to restart from the beginning13:02
r0kk3rzi dont think anything pays attention to it before the sb2 stuff13:04
r0kk3rzso just redo from sb2 target creation13:05
ohyashr0kk3rz, so you mean I dont have to "make" it all again?13:27
r0kk3rzyou shouldnt do13:27
r0kk3rzthat will be built in the proper arch anyway13:27
ohyashcool. Because I hadn't started yet with sb2 target creation.13:29
demonaxshI also have problems with gui - black screen, despite #define MALI_QUIRKS 1. Before, I did not notice it, but now screen for a fraction of a second shows start window (what with the choice of language), after which screen turns black. While I do not know how to fix this, but in the config I found:15:19
demonaxshMTK_GPU_SUPPORT = yes15:19
demonaxshIf I turn off GPU, will QPA have problems? Perhaps software drawing will be very slow, but I can temporarily bypass this problem.15:19
maldemonaxsh: what do you see in logs?15:20
maldmesg and logcat15:21
malprobably also journalctl15:21
demonaxshfor example15:23
demonaxshinteresting from logcat15:25
malso was there a reason why you don't build libminisf?15:28
demonaxshminimer and test_hwcomposer, nothing draws on screen, but there are not segmentation errors15:28
demonaxshbut there are no segmentation errors15:28
demonaxshno, I built, it just did not help15:28
malwell there is no reason to not have it (it will be there later anyway)15:29
demonaxshI built droidmedia on a clean tree, or do I need to build tree with patches?15:30
demonaxshbesides, I thought that droidmedia is necessary for camera, audio, etc15:31
demonaxshfrom interesting15:33
demonaxshifneq ($(FPGA_EARLY_PORTING), yes)15:33
demonaxshMTK_HWC_SUPPORT := yes15:33
demonaxshMTK_HWC_SUPPORT := no15:33
demonaxshbut in ProjectConfig.mk15:33
demonaxshFPGA_EARLY_PORTING = no15:33
malyou seem to have a bad habit of randomly picking parts of logs and not showing everything, the real error might be already a lot before the obvious error15:34
demonaxshwell, now I will show the whole log15:35
demonaxshjournalctl --no-pager15:35
maland when getting logs it's good to get them quite soon after reboot to be sure that even early logs are still available15:35
malfor example logcat can lose the early messages if it's flooded with errors15:36
demonaxshin journald.conf15:37
malthat's how it should be15:37
demonaxshhow can I make logcat not lose message? to set size of the ring buffer?15:39
maljust get the logs early enough15:41
maldemonaxsh: you claim that is all of logcat? where are the message you gave the last time?15:54
demonaxsh i specifically rebooted device. and immediately after gaining access to ssh did  /usr/libexec/droid-hybris/system/bin/logcat > /home/nemo/logcat.log15:56
maltry replacing > with &>15:57
ZuccaCan I do an OTA update straight from to or do I need to first update to and then to My device is LG hammerhead.16:05
malshould be ok to update directly16:06
demonaxsh/usr/libexec/droid-hybris/system/bin/logcat > /home/nemo/logcat2.log 2>&116:06
malwhy did you do that in so complicated way?16:07
malI still find it odd why the logcat has so few messages but maybe it really does16:08
demonaxshbut there is a journalctl, it's same basic logging for systems on systemd,16:10
malthose log different things16:10
NeKitdemonaxsh, which is yours mtk deice?16:11
demonaxshprocessor mt673516:11
demonaxshMTK_GPU_VERSION := mali midgard r7p016:11
NeKitis Android 32-bit?16:11
demonaxshno, it's aarch6416:11
NeKithm, do you see any ioctl error in dmesg?16:12
demonaxshalways thought that dmesg output somehow coincides with journalctl16:13
malNeKit: he has these16:13
malMay 29 15:10:08 Sailfish kernel:  (0)[3353:QSGRenderThread][name:ion&][ION]__ion_alloc buffer is error 0xfffffffffffffff4.16:13
malMay 29 15:10:08 Sailfish kernel:  (0)[3353:QSGRenderThread][name:ion&][ION]ION_IOC_ALLOC handle is invalid. ret = -12.16:13
malMay 29 15:10:08 Sailfish kernel:  (0)[3353:QSGRenderThread][name:compat_ion&][ION]COMPAT_ION_IOC_ALLOC unlocked_ioctl fail! ret = -12.16:13
malso maybe the 32-bit hwcomposer is again broken16:13
NeKiton MTKs it's generally broken on all 64-bit devices16:14
NeKitdemonaxsh, can you try with surfaceflinger plugin?16:14
malthat is probably the easier way to fix that16:14
NeKitbut the errors you mentioned likely come from gralloc, that's weird16:14
NeKitdemonaxsh, also, what happens with test_hwcomposer?16:15
demonaxshI have to built it, from git.merprohect?16:16
demonaxshnothing, just black screen, not crashing - nothing, in minimer it's same16:16
demonaxshnot errors, not sigfaults16:17
NeKitEGL_PLATFORM=hwcomposer test_hwcomposer, right?16:17
maldemonaxsh: if you have new enough rpm/dhd submodule then builds also qt5-qpa-surfaceflinger-plugin16:17
malNeKit: the log has also some other errors, check the google drive link he gave earlier16:21
demonaxshfrom version, but i built for
demonaxshno qt5-qpa-surfaceflinger-plugin not contain in my, ok I will add, or take a new build_packages.sh16:21
maldemonaxsh: update submodules as described in faq linked in channel topic16:23
demonaxshok, i'll try16:23
ZuccaThis should disable all openrepos? ssu lr | awk '/openrepos/ {print $2}' | xargs -n 1 ssu dr16:27
NeKitmal, 05-29 13:18:14.093  1403  1475 E hwcomposer:  ! (0xbf8f8) [COMP] (0) fence 41 didn't signal in 1000 ms16:27
ZuccaI assume dr means "disable repo".16:27
NeKitthat's something I've seen on Gemini, but it was way more flooded with errors16:27
NeKitI don't see ioctl errors in journalctl/dmesg though16:28
malNeKit: the one I gave is there16:35
Zuccaversion --dup fails. :(17:03
ZuccaI fail at ssu-fu. So here's all the enabled repos:
malZucca: you are missing all device repos17:40
malZucca: ah, the second one looks better17:40
malZucca: it's possible OTA updates are broken for that17:41
malZucca: looks like nobody has updated it
ZuccaLet's see...17:42
malZucca: the latest target is too new for it to work17:43
Zuccamal: Too new for hammerhead?17:44
malZucca: too new for the configurations used in those hammerhead repos17:48
ZuccaUh oh.17:48
malwondering why nobody kept maintaining that device17:49
malafaik it was quite popular17:49
ZuccaI only followed this
malthe problem is that it was using the latest target and not the versioned target17:52
ZuccaI see the part "common/sailfish_latest_armv7hl"... This is the problem?17:54
malthis is the problem
malit's using latest which is always the latest sailfish release, but your device configs haven't been updated to match the new needs to it's broken17:55
ZuccaOk. Dang. dhd repo.17:56
wiktorek140_mal:  can u look at this?
malif I had that device I probably would be able to update it very quickly, and if I had access to the repos17:57
wiktorek140_updated repo, hybris-12.1, device surnia17:58
ZuccaBasically I'd need a snapshot of that repo from the time when that part of the wiki was updated.17:59
malZucca: that would be trivial if I had permissions to the OBS project17:59
malsledges: ping17:59
Zuccamal: A read-only branch at one point of time, then I could run the updates?18:05
malZucca: not like that, just adding a proper versioned target there18:07
ohyash$COMPOSITOR_CFGS, how do I know which one of the event# is for touchscreen?18:07
malyou need to check on device18:09
ohyashthere are  8 /dev/input/event#18:10
ohyash0 to 718:10
malso either use evdev_trace from mcetools package or use cat18:10
ohyashthe one that gives continuous weird output in cat is touchscreen?18:12
malthe one that shows something when you touch the screen18:12
malthere might be some sensor device that gives output all the time18:13
malso just find the one that responds the touching the screen18:13
ohyash_this is weird, event0 continuously gives output irrespective of me touching the screen. other event# are pure blank18:19
ohyash_nvm, its event 318:20
Zuccamal: Thanks for your efforts so far! If updating is someday possible can you update the wiki page?18:35
malZucca: but if you want to keep using the device why not continue the port and make it possible to use the latest sailfish in it?18:36
ZuccaI don't think my skills are that high. :)18:39
wiktorek140Zucca: u can always try ;) it cost you only time, what u will learn it will be your ^-^18:51
ohyash_error during PB :
r0kk3rzZucca: it should be fairly straight forward18:53
ohyash_I figured the log file would be required. Here, PB :
wiktorek140ohyash_: use this
ohyash_done, works, thanks19:07
*** Nokius_ is now known as Nokius20:26
steeeveAre there any advantages in building 15.1 over 14.1?21:00
malcurrently only disadvantages21:06
malat the moment 15.1 should be used only if there is nothing else available21:07
steeeveOh yeah i though it, too many changes from 7 to 821:16

Generated by 2.17.1 by Marius Gedminas - find it at!