Sunday, 2018-12-02

*** ChanServ sets mode: +v T402:09
*** ChanServ sets mode: +v T408:34
piggzmal: on mido, udevadm infor -e works as exptected, so need to figure out whats wrong with that rule10:47
T4<adampigg> mal: i wonder if it becuase i dont have a /dev/block11:33
mal@adampigg so the device never has /dev/block ? even in android?11:40
T4<adampigg> mal: i dont know for sure, but i would assume it does have it in android becuase thats what is used in the fstab file13:09
T4<adampigg> mal:
*** ChanServ sets mode: +v T414:02
piggzmal: would android init be responsible for creating /dev/block?
malpiggz: no, it's done in systemd, init shouldn't do it in sailfish14:51
malpiggz: or actually in udev14:52
piggzyes, but i think my udev isnt working14:52
piggzit has no devices :D14:52
malpiggz: maybe something wrong with the default udev rules14:53
piggzmal:  well, there is the rule it complains about, but i dont understand why14:53
piggzmal: this....15:03
piggzDEVLINKS=="* platform/pci0000:00/80860F14:00/by-name/android_persistent *|platform/pci0000:00/80860F14:00/by-name/android_persistent *|* platform/pci0000:00/80860F14:00/by-name/android_persistent|platform/pci0000:00/80860F14:00/by-name/android_persistent"15:03
piggzmy /sys doesnt contain any by_name15:03
piggzbut, i dont really know what im looking for atm15:04
piggzmal: i commented out that line .. now there are no errors reported by udev, but still my udev DB is empty15:08
piggzmal: should uevent files in /sys be mode rw, or w ?15:42
malpiggz: those are 644 on my device15:44
piggzmal: doing a strace on udeavedmin trigger shows loats of failures opening them fro read15:46
malpiggz: maybe you should get some logs from udev behavior15:48
piggzmal: so, i have a bunch of files that are 644, and others which are 2 ... so maybe it is correct15:57
piggzmal: is this correct...15:59
piggzsh-3.2# file ./sbin/ueventd15:59
piggz./sbin/ueventd: symbolic link to `./droid-hal-init'15:59
piggzi guerss it is16:01
piggzkrnlyng: hi, any thoughts about why udev database on my device might be emtpy, or wht may cause it?18:04
piggzor just places to look18:04
krnlyngare there any warnings during the generation?18:05
krnlyngor errors18:05
krnlyngwhich db are you refering too? the generated rules files?18:05
piggzkrnlyng: there is one error in a rule, and with that one rule commented out, it is still mepty....18:05
piggzkrnlyng: udevedm info -e18:06
piggzlists all devices18:06
piggzon mido, lists lots, on latte, nothing18:06
piggzsystemd is unable to do any mounting until udev says devices are there18:06
krnlynghm ok18:09
krnlyngany errors from udev in systemd?18:09
krnlyngpiggz, hmm maybe it is somehow stuck?18:15
krnlyngpiggz, can you run /sbin/ueventd and see what happens?18:16
krnlyngnot sure if that shows us anything..18:16
piggzkrnlyng: well, that may be rleated....that is a symlink to droid-hal-init isnt it?18:16
piggzand it aborts18:16
krnlyngit should be a symlink to droid-hal-init yes18:18
krnlyngpiggz, are there devices in /dev before running it, and after?18:19
krnlynghm ok18:19
piggzkrnlyng: yes, i have devices in /dev, but nothing like /dev/block18:19
krnlyngcan you rm /dev/.coldboot_done18:20
krnlyngand run it agian?18:20
krnlyngwith strace18:20
* krnlyng is just guessing and hoping something reveals something18:20
malpiggz: stack corruption error is visible in that18:22
malso maybe ueventd crashes18:23
krnlyngpiggz, maybe the block rules are broken18:24
piggzit does, as does droid-hal-init18:24
piggzkrnlyng: one moment....18:24
krnlyngpiggz, which android version?18:24
piggzkrnlyng: 6.0.1, LOS 1318:24
krnlyngi see18:26
krnlyngcould be tls breakage again18:26
malkrnlyng: but that is i48618:26
krnlyngthe dhi crash*18:26
krnlyngi486 device?18:27
piggzkrnlyng: yes18:27
krnlyngthen it almost certainly is tls breakage?18:27
krnlyngnot sure if we applied the patch to 618:27
piggzkrnlyng: new territory ;)18:27
piggzkrnlyng: udev doesnt like this rul18:27
piggzSUBSYSTEM=="block", DEVLINKS=="* platform/pci0000:00/80860F14:00/by-name/android_persistent *|platform/pci0000:00/80860F14:00/by-name/android_persistent *|* platform/pci0000:00/80860F14:00/by-name/android_persistent|platform/pci0000:00/80860F14:00/by-name/android_persistent", MODE="0660", GROUP="system", OWNER="system"18:27
piggzDec 02 18:27:39 Sailfish systemd-udevd[1058]: unknown key 'DEVLINKS' in /lib/udev/rules.d/999-android-system.rules:26818:28
piggzDec 02 18:27:39 Sailfish systemd-udevd[1058]: invalid rule '/lib/udev/rules.d/999-android-system.rules:268'18:28
krnlyngpiggz, yeah i am aware of that one18:30
krnlyngbut it shouldn't cause this18:30
krnlyngbut the other rules for the block devices18:30
piggzis there a dhi patch i need?18:32
malat least hybris-13.0 and 14.1 have different patches in bionic18:33
krnlyngsomething like this one18:33
TheKit<piggz> krnlyng: hi, any thoughts about why udev database on my device might be emtpy, or wht may cause it?18:55
TheKitit is not resolved yet?18:55
piggznot yet...18:55
TheKiton mine, there was a change in kernel by Intel that sysfs symlinks18:56
piggzTheKit: and the fix was?19:01
TheKitpiggz, [PATCH 269/429] Skip creating subsystem/ in sysfs nodes19:03
TheKittook me a while to find, but check if it is applied: - if yes, unapply19:03
piggzTheKit: yup, unapplied19:11
piggzTheKit: thanks, i have a udev database now! mal, krnlyng^^19:14
piggzand system is mounted19:15
piggzand /vendor19:15
piggzso,that is something new to me anyway, silly kernel patches breaking udev!19:17
malpiggz: nice, now you can continue debugging other issues19:26
malpiggz: do you see the crashes still?19:27
piggzmal: dhi works now with the bionic patch19:29
piggztho, logcat not working19:30
piggzsh-3.2# /usr/libexec/droid-hybris/system/bin/logcat19:30
piggzlogcat read failuresh-3.2#19:30
piggzmal: and after some time, i loose my usb net connection, not sure if its usb_moded or DHI19:32
malpiggz: check that usb stuff is disabled in init rc files19:32
piggzmal: yes, just need to be connect long enought19:33
steeeveGuess whos back20:17
piggzmal: even comnent lines like
piggzall the way to 49920:36
*** ChanServ sets mode: +v T420:38
T4<adampigg> @eugenio_g7 looking at init.latte.rc, the wifi mac address seems to be loaded from /config (which i havnt got mounted yet)20:48
piggzmal: yes, commenting htose seems to have helped20:58
piggzmal: this is how droid-hal-init looks now
piggzand now got logcat21:01
T4<morfeokmg> how can start to install sailfish to one tablet itab 421:03
Mister_Magisterthis english is beautiful (i wasn't better tho)21:04
malpiggz: does that mean dhi exists?21:29
piggzmal: i think so, i have a second-stage dhi21:30
piggzmal: surfaceflinger worked, hwcomposer not yet21:37
TheKitpiggz, you mean you have graphics on surfaceflinger?21:37
piggzTheKit: yes, i got the LOS logo21:38
TheKitI thought Mer one21:38
TheKit(with qmlscene)21:38
piggzno, just tested the surfaceflinger executable21:38
eugeniojoining the fun :)21:39
piggzeugenio: so, small kernel fix needed to fix udev21:39
piggzand comment out all the usb stuff, which i assume youve done21:39
piggzwondering now if i shoould install all your packages21:39
piggzeugenio: oh, and small bionic fix also needed21:40
eugenioyup I do have the bionic fixes, but still no dice on libhybris (on cm12.1 though)21:40
eugeniogoing to (un)apply the kernel fix21:42
piggzeugenio: is your extremely slow to interact with over telnet/usb ?21:50
eugenioafter starting systemd? Try masking usb-moded21:50
piggzeugenio: no, all the time21:51
piggzits fsking painful!21:51
eugenioworks smoothly here...21:54
eugeniodo you use NetworkManager in your host computer?21:54
piggzit doesnt even refresh the text properly, have to hit enter to arrow keys to refresh output21:57
eugeniotry disabling NM and bringing up the rndis interface manually21:58
eugeniohuh, finally got the sensor modules to load23:48
eugenioI guess I should now recompile sensorfwd with the iio adaptor23:48

Generated by 2.17.1 by Marius Gedminas - find it at!