Saturday, 2021-11-13

T42<TheVancedGamer> enable CONFIG_DEVTMPFS and DEVTMPFS_MOUNT (re @hacker12455: the problem now is t...)11:23
T42<hacker12455> it's enabled11:25
T42<TheVancedGamer> uh11:25
T42<TheVancedGamer> it should work then11:25
T42<hacker12455> will try again in a bit11:25
T42<elros34> @hacker12455 I guess you have incorrect fixup-mountpoints but you need to telnet to confirm. Have you followed porting instruction about accessing device or what to do when telnet do not work?13:40
T42<hacker12455> hm that was weird14:03
T42<hacker12455> so I got to the init part14:03
T42<hacker12455> then after a moment my screen started turning itself on and off a couple of times14:04
T42<hacker12455> then it turned off14:04
T42<hacker12455> the phone is off14:04
T42<hacker12455> ah I just realized14:07
T42<hacker12455> the phone has no IP assigned14:07
T42<hacker12455> it's like it is trying to boot but dies14:07
T42<hacker12455> hmm14:08
T42<hacker12455> if I press the power button the phone turns the display on14:08
T42<hacker12455> then it gets thrown into charging mode14:08
T42<hacker12455> where I see the mer debug telnet again14:08
T42<hacker12455> though it unregisters itself again14:09
T42<hacker12455> let me try to see if it made a log14:09
T42<elros34> "Product: init-debug in real rootfs" so you fixed previous error, I suggest you to read notes, it has information how to stop rebooting, telnet if inet address is not assigned. It's really all there, just need to read it14:10
T42<hacker12455> yeah I know the info how to stop rebooting is there14:11
T42<hacker12455> I assume the reboot steps would work if I chrooted into the rootfs from recovery?14:13
T42<elros34> not sure what do you mean but you do not need chroot to anything from recovery14:14
T42<hacker12455> how do I access the rootfs to run the commands then if the interface disappears faster than I could assign the IP?14:15
T42<elros34> create init_enter_debug214:16
T42<hacker12455> in /14:17
T42<hacker12455> ??14:17
T42<hacker12455> ? (edited)14:17
T42<elros34> no, in /data/.stowaways/sailfishos/init_enter_debug2. Read notes14:18
T42<elros34> or if by / you meant sailfish / then yes:)14:18
T42<hacker12455> now to set the IP14:21
T42<hacker12455> in telnet now14:23
T42<hacker12455> masking droid-hal-init14:25
T42<hacker12455> /data appears to be mounted correctly14:26
T42<hacker12455> the system.mount points to14:31
T42<hacker12455> What=/dev/block/platform/soc/7824900.sdhci/by-name/system14:31
T42<elros34> most likely result of wrong fixup-mountpoints14:32
T42<hacker12455> could I try changing that manually just to test?14:33
T42<elros34> sure but do not forget about other mount nits, reinstalling whole image just to test one thing is waste of time14:34
T42<hacker12455> so put in /dev/block/mmcblk paths there?14:34
T42<elros34> without block/ part14:36
T42<hacker12455> unmask droid-hal-init?14:37
T42<elros34> if everything else (dmesg) looks sane then sure14:38
T42<hacker12455> reboot?14:38
T42<hacker12455> or continue init14:38
T42<hacker12455> @elros3414:41
T42<elros34> try and you will see:) I depend of what you have tried so far, if you already started droid-hal-init then you must reboot first14:44
T42<hacker12455> well got a blinking led14:46
T42<hacker12455> have to go for now14:48
Mister_Magistermal: don't you think that jolla's repos should have higher priority in zypper so that when users upgrade their system they don't succ packages from openrepos instead of jolla repos?17:38
piggzlbt: ping Certificate Verification Error: Peer certificate subjectAltName does not match host, expected, got DNS:bugs.sailfishos.org17:42
T42<hacker12455> @elros34 so update18:15
T42<hacker12455> right now I get a fast blinking LED18:16
T42<hacker12455> display is dark18:16
T42<elros34> strange but next step is always the same: read journalctl, dmesg, logcat18:17
T42<hacker12455> right now gonna wait a bit to see if it reboots or something18:18
T42<hacker12455> well I don't see any obvious errors from pstore18:27
T42<elros34> why would you look into pstore? device reboots?18:27
T42<hacker12455> doesn't reboot (re @elros34: why would you look i...)18:27
T42<elros34> so read journalctl18:28
T42<hacker12455> weird18:32
T42<hacker12455> can't execute '/sbin/droid-hal-init': Too many levels of symbolic links18:32
T42<hacker12455> so droid-hal-init can't start19:04
T42<hacker12455> should I send the journal output?19:06
T42<hacker12455> I found that when I hold the power button to shut it down19:09
T42<hacker12455> it lights up the backlight then it shuts down normally19:09
T42<elros34> I guess you have /system entry in fstab instead / for system partition so mount units are incorrect. See if /system/ looks like root partition19:15
T42<elros34> ls /system/19:15
T42<hacker12455> from telnet or recovery (re @elros34: ls /system/)19:15
T42<hacker12455> let me mask droid-hal-init for now19:16
malpiggz: when does that happen?19:18
T42<elros34> @hacker12455 no need to mask anything, just run that ls on device19:19
T42<hacker12455> from initramfs telnet? (re @elros34: @hacker12455 no need...)19:19
T42<hacker12455> because I don't have a way to ssh/telnet into the rootfs19:19
T42<elros34> sure this is what we debug now19:19
T42<elros34> why you cant telnet?19:20
T42<hacker12455> it doesn't give me a network device once it goes out of the initramfs (re @elros34: why you cant telnet?)19:21
T42<elros34> have you tried halium instruction linked in hadk-hot?19:22
T42<hacker12455> except in my case it's not a matter of not having an ip assigned19:24
T42<hacker12455> it doesn't expose a device19:24
T42<elros34> so you have nothing in host dmesg?19:25
T42<hacker12455> exactly19:26
T42<hacker12455> nothing in ip a as well19:26
T42<elros34> even with init_enter_debug2? ANDROID0 detected later by usb-moded so this should also shows something19:27
T42<hacker12455> init_enter_debug2 gives me a device19:27
T42<hacker12455> let me check what happens if I continue boot19:27
T42<Itzxdatl> I get tihs error :
T42<hacker12455> nope (re @hacker12455: let me check what ha...)19:29
T42<hacker12455> just unregisters the device19:29
T42<elros34> so mount system partition manualy while you have telnet acces and check, you do not need droid-hal-init for it19:30
T42<hacker12455> though19:36
T42<hacker12455> I found that the telnet connection weirdly gets disconnected after some time19:37
T42<elros34> so it's system-as-root, you need to rectify your fstab /system -> / then rebuid droid-hal19:38
T42<hacker12455> do I need to rebuild the rootfs itself afterwards? (re @elros34: so it's system-as-ro...)19:43
T42<elros34> you can just copy and install droid-hal rpm19:43
T42<elros34> make sure your fixup-mountpoints are correct now and fstab changes are both in device/ and out/19:44
T42<b100dian> anyone else building 4.3 on obs and packaging gmp-droid? I still get unresolvable "droidmedia" even thought the droidmedia package here "provides"
T42<hacker12455> should I change fstab.qcom?19:45
T42<elros34> that is what I am talking about19:45
T42<elros34> @b100dian you have some leftover:
T42<hacker12455> like that? (re @elros34: that is what I am ta...)19:46
T42<hacker12455> /dev/block/platform/soc/7824900.sdhci/by-name/system         /      ext419:46
T42<elros34> yes19:46
lbtpiggz: isn't used anymore but it's still in DNS so you get a cert mismatch19:46
T42<b100dian> @elros34 that was on the next page 🤦 thank you19:47
T42<hacker12455> fixup-mountpoints should be correct (re @elros34: make sure your fixup...)19:51
T42<hacker12455> so19:52
T42<hacker12455> anything else?19:52
T42<hacker12455> built droid-hal19:53
T42<hacker12455> didn't see any changes that I needed to do in fixup-mountpoints19:53
T42<hacker12455> btw, how do I install that package on the device19:53
piggzlbt: what correct method to use obs/osc?19:58
piggzmal: ^^19:59
malit should be build.sailfishos.org19:59
T42<hacker12455> hm do I need to rebuild droid-hal after that fstab change as well?20:01
piggzmal: i just keep getting a 40120:03
piggzseems i cant log in at all on the site either20:04
piggzi shouldnt have logged out!20:05
malpiggz: there seems to some issue, lbt is looking at it, I can't login either, login to bugzilla works fine but not to obs20:05
mal@hacker12455 at least you need to run make hybris-hal and then repackage droid-hal, the reason is that droid-hal packaking creates mount point and uses fstab files in out/ folder of hybris-hal build, of course you could hack the fstab somewhere under out/ and only repackage but not really much difference because running hybris-hal build again should not take long if you have the build already done20:08
T42<hacker12455> weird20:14
T42<hacker12455> so i changed the fstab.qcom in msm8937-common/rootdir/etc20:14
T42<hacker12455> rebuilt hybris-hal20:14
T42<hacker12455> but the changed fstab isn't in out20:14
malif the fstab in out at all?20:15
T42<hacker12455> yes20:15
T42<hacker12455> in vendor/etc20:15
T42<hacker12455> hm what if I tried to delete it?20:16
T42<hacker12455> 🤔20:19
T42<hacker12455> so it just gets replaced with the unedited one20:20
T42<hacker12455> where is the one I need to edit though20:20
T42<hacker12455> unless it grabs the fstab_legacy ?20:21
T42<hacker12455> let me edit that20:21
T42<hacker12455> now it works20:21
T42<hacker12455> now to install that package on the devic20:22
T42<hacker12455> now to install that package on the device (edited)20:22
malcheck if there is fstab in some other folder like intermediates or something in out/20:23
malah, so you edited wrong file20:23
T42<hacker12455> looks like it (re @SailfishFreenodeIRCBridgeBot: <mal>ah, so you edit...)20:24
T42<hacker12455> every instance of fstab.qcom now points system to /20:24
T42<hacker12455> how do I install the package now?20:25
T42<hacker12455> chroot into the rootfs?20:27
lbtSo it seems I created the Mer ldap cert 10 years ago today....20:28
lbtI'll need to do some CA rebuilding so that may take a while20:28
T42<hacker12455> so I installed the droid-hal-cedric.rpm20:36
T42<hacker12455> that's how the system.mount should look like?20:37
T42<hacker12455> What=/system_root/system20:37
T42<hacker12455> system_root.mount is also in /usr/lib/systemd/system20:39
T42<hacker12455> ah it has that same thing20:40
T42<hacker12455> What=/dev/block/platform/soc/7824900.sdhci/by-name/system20:40
T42<hacker12455> so now after changing the mounts to proper values20:45
T42<hacker12455> it just flashed the backlight at me three times20:45
T42<hacker12455> no led this time20:45
malis telnet working?20:46
T42<elros34> @hacker12455 when you create fixup-mountpoints always check fstab, do not assume all entries use same by-name path20:48
T42<hacker12455> alright20:48
T42<hacker12455> btw20:49
T42<hacker12455> grabbed journal20:49
T42<hacker12455> seems like droid-hal-init starts now20:49
T42<hacker12455> that's all I got from system.journal20:49
T42<hacker12455> let me remove init_enter_debug20:50
T42<hacker12455> so that didn't give me either boot or any network device20:52
T42<hacker12455> just kept flashing the backlight20:52
T42<hacker12455> let me reboot and grab the journal20:52
T42<hacker12455> seems like it got cut off shortly after starting droid-hal-init20:57
T42<hacker12455> unless I need to let it keep booting for longer?20:59
malis logcat working?21:00
mali.e. /usr/libexec/droid-hybris/system/bin/logcat21:01
T42<hacker12455> I'd check if I could ssh into the rootfs21:01
T42<hacker12455> it still does not expose a network device for me21:01
malwhich android base?21:01
T42<hacker12455> 17.121:01
maldid you properly setup selinux?21:01
T42<hacker12455> yep21:01
T42<hacker12455> followed hadk-faq21:02
malso you have it enabled in kernel and then you copied the files (no symlinks)21:02
T42<hacker12455> yes21:02
malcan you show the config repo content21:02
T42<hacker12455> actually21:02
T42<hacker12455> wait21:02
T42<hacker12455> they are symlinks21:04
T42<elros34> figure out when you loose telnet: when usb-moded starts or droid-hal-init, it will be much easier to debug it with some access to device21:04
T42<hacker12455> let me try masking droid-hal-init and see what happens21:05
malif you used symlinks for those two selinux files it will break and probably causes reboot21:07
T42<hacker12455> except I'd probably see if it rebooted21:07
T42<hacker12455> it doesn't reboot21:07
malcould also just cause usb to break, difficult to say21:08
T42<hacker12455> you know21:10
T42<hacker12455> let me replace them with the files from my device21:10
T42<hacker12455> then I will update21:10
T42<hacker12455> hm21:12
T42<hacker12455> should it be named as the symlinks or do I grab the same names that are in /vendor?21:12
T42<hacker12455> so for example the github repo has 30.policy that symlinks to precompiled_sepolicy21:14
T42<hacker12455> do I change the name of precompiled_sepolicy to 30.policy when I copy it or not21:14
T42<hacker12455> ok I changed the names of the files to match the symlinks21:21
T42<hacker12455> hmmm21:22
T42<hacker12455> so now it's just sitting there with the backlight on21:22
T42<hacker12455> no usb device detected21:22
malthe names need to match the symlinks, so rename the files21:24
T42<hacker12455> that's what I did now21:24
T42<hacker12455> currently as I said I have the device sitting with the backlight on21:25
maldoes masking droid-hal-init do anything?21:25
T42<hacker12455> did not test yet21:25
T42<hacker12455> mask user@100000 as well?21:29
T42<hacker12455> so with only droid-hal-init masked I get only a fast blinking LED21:33
T42<hacker12455> no usb devices21:33
T42<hacker12455> uh21:47
T42<hacker12455> found that in journal21:47
malprobably not a problem21:53
T42<hacker12455> so any idea how to proceed?21:54
T42<hacker12455> since masking droid-hal-init does not make a difference21:55
T42<hacker12455> just gonna drop that in here21:55
T42<elros34> have you checked usb-moded?21:55
T42<elros34> and I told you to add audit=0, looks at these logs it's pain to read them21:56
T42<hacker12455> how can I check it? (re @elros34: have you checked usb...)21:57
T42<elros34> mask, same as you did for droid-hal-init21:58
T42<hacker12455> so unmask droid-hal-init21:58
T42<hacker12455> and mask usb-moded?21:58
T42<elros34> whatever will help21:59
T42<hacker12455> after masking usb-moded I can get to the telnet22:02
T42<hacker12455> SerialNumber: Mer Debug telnet on port 2323 on rndis0 - also running udhcpd22:02
T42<hacker12455> logcat says it can't find libandroidicu.so22:07
maldo not try to guess what is relevant, and there is simple way to fix issue22:08
malbut that is not the critical part now22:08
malyou could try symlinks like these
malnot that you need to create the odm and needed subfolders22:09
T42<hacker12455> yeah I see22:09
malthat etc is a symlink also22:09
T42<hacker12455> would that help with getting graphics?22:10
mallike I said do try to guess the relvant parts in logs just show the whole log22:11
mal*do not try22:12
T42<hacker12455> logcat or journal22:12
T42<hacker12455> logcat:
malopen /dev/mdss_rotator failed err = 13 errstr = Permission denied22:19
maltry first manually changing permissions of that device node and try starting graphics again22:19
T42<hacker12455> so boot the device22:20
T42<hacker12455> change the permissions and restart graphics?22:20
malwell if you want to reboot the device then you need to manually change udev rule in /lib/udev/rules.d/999-android-system.rules22:22
T42<hacker12455> I had the device in twrp22:22
T42<hacker12455> to grab the logs out22:22
T42<hacker12455> now I booted it to sailfish22:22
T42<hacker12455> what permissions do I give it?22:23
malchange group to graphics22:23
T42<hacker12455> currently it has crw-rw-r--22:23
T42<hacker12455> what owner?22:23
T42<hacker12455> system?22:24
malsystem is fine as user22:24
mal I changed it in the device repo and from there it gets via hybris-hal and droid-hal to the udev rules22:24
T42<hacker12455> how do I restart graphics?22:25
malsystemctl restart user@10000022:25
T42<hacker12455> got graphics22:26
malwhich device is that?22:27
T42<hacker12455> Moto G5 (cedric)22:29
T42<hacker12455> well22:34
T42<hacker12455> I got past the initial setup22:34
T42<hacker12455> but I can't unlock the device22:34
T42<hacker12455> because once I try it it crashes22:34
T42<hacker12455> unless I need minimedia for that?22:35
T42<hacker12455> hm22:36
T42<hacker12455> Oct 27 22:26:23 MotoG5 encsfa[2762]: Device /dev/sailfish/home does not exist or access denied.22:36
mallast one is not an issue22:41
malyou can add the symlinks I mentioned if you want22:42
T42<hacker12455> did just that (re @SailfishFreenodeIRCBridgeBot: <mal>you can add the...)22:42
malcheck logs again22:43
malwhat crashes, whole device or just some process?22:43
T42<hacker12455> some process22:44
T42<hacker12455> lipstick restarts22:44
T42<hacker12455> hm22:45
T42<hacker12455> damn I really need to disable the audit...22:45
T42<hacker12455> it's spammy as hell22:45
T42<hacker12455> did I screw something up here?22:48
T42<hacker12455> also that vendor.time_daemon thing is still crashing22:49
T42<elros34> yeah you did22:49
T42<hacker12455> let me guess22:49
T42<hacker12455> that last line should be on the end of the 3rd one22:50
T42<elros34> time_daemon can be disabled22:50
malyes, also no need to hardcode event device22:51
T42<hacker12455> perfect now it works22:52
T42<rodrisola> i have this error22:56
T42<hacker12455> so now to get networking running22:57
T42<hacker12455> ofono just says Unknown RIL transport: binder22:58
T42<rodrisola> error in fetch clang , is normal?23:01
mal@hacker12455 do you have needed things installed for ofono?23:11
T42<hacker12455> seems like I'm missing ofono-ril-binder-plugin23:12
mal@rodrisola which android base?23:13
T42<rodrisola> Hybris18. 1 (re @SailfishFreenodeIRCBridgeBot: <mal>@rodrisola whic...)23:15
T42<rodrisola> Android 11 (re @SailfishFreenodeIRCBridgeBot: <mal>@rodrisola whic...)23:15
mal@rodrisola that is not yet fully tested, but you could try changing the pybind11 repo in manifest to use master branch instead of the normal android 11 branch23:16
T42<hacker12455> do I have to build ofono-ril-binder-plugin?23:23
T42<rodrisola> Ok (re @SailfishFreenodeIRCBridgeBot: <mal>@rodrisola that...)23:27
mal@hacker12455 it's available in jolla repos, no need to build23:29
T42<hacker12455> how can I download it?23:29
T42<hacker12455> hm do I need to add the repos or something?23:32
T42<hacker12455> because I ran zypper se ofono*23:32
T42<hacker12455> and I don't see the package23:32
maldid you run zypper ref first23:32
T42<hacker12455> yes23:32
malis network working23:33
T42<hacker12455> I'm doing that from PlatformSDK23:33
T42<hacker12455> do I need to do that from the device23:33
malyes, if you want to install it there, otherwise just add it to patterns so it will be installed when you build new image23:34
T42<hacker12455> could I download it and move it to the device?23:35
T42<rodrisola> i have this error23:39
T42<rodrisola> FAILED: out/soong/build.ninja23:39
mal@rodrisola how much memory do you have on your build machine?23:40
T42<rodrisola> 1623:42
malok, it should be enough23:42
T42<hacker12455> well the problem is that zypper does not see it (re @SailfishFreenodeIRCBridgeBot: <mal>yes, if you wan...)23:43
T42<hacker12455> from platform sdk23:43
T42<hacker12455> No provider of 'ofono-ril-binder-plugin' found.23:43
mal@hacker12455 it's only visible in target, not in sdk, not sure what command you used to check23:43
T42<hacker12455> ah23:43

Generated by 2.17.1 by Marius Gedminas - find it at!