| T42 | <edp_17> It makes me wonder what changed since sfos 4.3 that causes this issue on galaxy devices. | 07:46 |
|---|---|---|
| T42 | <edp_17> Galaxy note 4 (treltexx) with base los14.1 sfos 4.3 and 4.4 are impacted, same device but los17.1 base sfos 4.3 is impacted. Galaxe S4 (jfltexx) with los17.1 base sfos 4.4 is impacted. | 07:49 |
| T42 | <edp_17> However, the note4 with los14.1 sfos 4.2 and before versions are fine. | 07:50 |
| T42 | <edp_17> I also have a galaxy s2 with cm12.1 base and sfos 4.3 works fine but 4.4 stuck on a black screen (haven't yet checked whether that has the same issue). | 07:52 |
| T42 | <elros34> Forced sailjail, libhbyris update?. About selinux permission: did you add to sparse actual selinux files not symlinks? | 07:54 |
| T42 | <edp_17> I am thinking whether worth trying to build sfos 4.2 for this device (jfltexx). Just to see whether that works okay. | 07:55 |
| T42 | <edp_17> Forced sailjail, libhybris update? How do you mean? | 07:55 |
| T42 | <edp_17> About selinux permission: no I only used symlinks from hadk-faq. | 07:56 |
| T42 | <elros34> issue is at android side, android services fails so I do not think older version will help. 4.4.0 enforce sailjail which could affect your devices. Recently there was info on channel about latest libhybris/qpa-hwcomposer-issue | 07:57 |
| T42 | <edp_17> Wait, no. On treltexx (los17.1 - sfos4.4) I copied those files from device. Didn't help | 07:57 |
| T42 | <elros34> re: selinux I am talking about jfltexx, copy files according to faq not create symlinks | 07:59 |
| T42 | <edp_17> Will try that, thanks.. | 08:00 |
| Thaodan | Sailjail doesn't affect any system services only apps started within Sailjail. | 14:03 |
| T42 | <LSolrac> Hello. I've a question. Has anyone Mer/Hybris with Halium before? | 18:40 |
| mal | there have been sailfish ports which used halium | 19:02 |
| xmn | for manjaro phosh, how do you know when to say yes or no, when prompted for using one over the other? | 20:34 |
| xmn | sorry wrong window :) | 20:35 |
| T42 | <edp_17> @elros34 : I've removed the namespace.default.permitted.paths from both "/system/etc/ld.config.txt" and "/system/apex/com.android.runtime.release/etc/ld.config.txt" and copied both files from vendor to /etc/selinux/... | 21:28 |
| T42 | <edp_17> Oh, and I also renamed "/usr/libexec/droid-hybris/system/lib/libcutils.so" | 21:29 |
| T42 | <edp_17> The result: the libcutils.so related error gone from logcat and I got the gui. | 21:30 |
| T42 | <edp_17> However, unfortunately, the device still reboots after a few minutes. :( | 21:30 |
| T42 | <edp_17> I let logcat and journal running until the device rebooted. They are here: logcat: https://paste.ubuntu.com/p/jVkRYbDv8x/ | 21:38 |
| T42 | <edp_17> journalctl -b0: https://paste.ubuntu.com/p/Ct6HMwsKfH/ | 21:38 |
| T42 | <edp_17> dmesg: https://paste.ubuntu.com/p/jsnvvtxC6S/ | 21:38 |
| mal | dsme is failing, that is one of the critical low level sailfish processes | 21:53 |
| mal | it segfaulted, not sure why, there is no clear error in journal | 21:54 |
| T42 | <edp_17> mal : How can I get clearer log? | 21:55 |
| mal | you probably need to try to run that process manually with gdb to see you can get any backtrace | 21:57 |
| T42 | <edp_17> mal : can you give me more details please? Never done that? How should I prevent rebooting? touch /init_enter_debug2 ? | 22:01 |
| mal | that probably prevents it from rebooting but doesn't fix anything | 22:02 |
| T42 | <edp_17> I thought should stop reboot first to be able to run dsme. Okay, I telnet in and run dsme manually. Can you give me the command, please? | 22:04 |
| mal | is your build arm or aarch64? | 22:05 |
| T42 | <edp_17> arm | 22:05 |
| mal | assuming you have gdb installed: "gdb --args /usr/sbin/dsme -- -p /usr/lib/dsme/startup.so" | 22:05 |
| T42 | <edp_17> This gives me a > where the cursore is blinking. | 22:06 |
| mal | are you sure you didn't keep those " accidentally? | 22:07 |
| T42 | <edp_17> Ops, only the last one. Didn't reaise. https://paste.ubuntu.com/p/YrjTnKtrqn/ | 22:08 |
| mal | hmm, it didn't like the parameters | 22:09 |
| mal | did you try again with correct command? | 22:10 |
| mal | the output you gave looks wrong | 22:10 |
| T42 | <edp_17> I did and got this: https://paste.ubuntu.com/p/82vcp7Q6Wc/ | 22:10 |
| T42 | <edp_17> It's not possible to redo anything quick because the device gets rebooted and I need to wait until I have telnet session again. And that telnet session lasts about a minute before the next reboot. | 22:11 |
| T42 | <edp_17> That's why I asked whether I should stop reboot to have more time. | 22:12 |
| mal | you need to the type "run" and it should start the process | 22:13 |
| T42 | <edp_17> Okay, I got this: https://paste.ubuntu.com/p/CrjJ7Gx5ZG/ | 22:15 |
| mal | then type "thread apply all bt" | 22:16 |
| T42 | <edp_17> Got this: https://paste.ubuntu.com/p/dgwx4twztc/ | 22:20 |
| mal | ok, it seems gdb doesn't like running dsme inside it | 22:46 |
| T42 | <edp_17> I've masked dsme but since I don't have telnet and the device gets rebooted quicker (before I can get in) | 22:48 |
| mal | maybe in gdb try typing this "handle SIGHUP nostop" before you type "run" | 22:50 |
| T42 | <edp_17> Will do but now I need to find out how to unmask dsme while I cannot get into the device. Do you have idea? | 22:50 |
| mal | systemctl unmask dsme | 22:51 |
| T42 | <edp_17> There is no symlink on /etc/systemd/system/. | 22:51 |
| mal | how did you mask it? | 22:51 |
| T42 | <edp_17> systemctl mask dsme.service | 22:51 |
| T42 | <edp_17> With 'systemctl unmask dsme' I cannot do it because I cannot get into the device before it reboots. | 22:54 |
| mal | then how do you plan in doing anything? | 22:55 |
| T42 | <edp_17> I think try to put /init_enter_debug2 file then try to get in and unmask dsme. | 22:56 |
| T42 | <edp_17> in twrp | 22:57 |
| T42 | <elros34> or do it directly in twrp, chroot /data/.stowaways/sailfishos/ /bin/bash -l | 22:58 |
| T42 | <edp_17> Thanks @elros34 . I'll make a note about this one but already done it on the other way. :) | 23:03 |
| T42 | <edp_17> mal : I got this: https://paste.ubuntu.com/p/vHQ7XMFtsN/ | 23:03 |
| T42 | <edp_17> Let me try again, I have an idea. | 23:04 |
| T42 | <edp_17> Nope, that idea didn't work. :( | 23:10 |
| mal | so it doesn't crash now? | 23:15 |
| T42 | <edp_17> mal : it is still crahsing | 23:16 |
| mal | ok, then try typing "set follow-fork-mode child" before typing "run" | 23:18 |
| T42 | <edp_17> Okay, give me a minute. | 23:20 |
| T42 | <edp_17> mal : Here it is: https://paste.ubuntu.com/p/95rrJkwmcm/ | 23:22 |
| T42 | <edp_17> @elros34 : How can I permanently avoid /usr/libexec/droid-hybris/system/lib/libcutils.so to get into the build? | 23:24 |
| T42 | <elros34> maybe remove it from out before building droid-hal | 23:25 |
| T42 | <edp_17> @elros34 : From where exactly? There are a few ones : https://paste.ubuntu.com/p/X62ZS6B99s/ | 23:28 |
| T42 | <edp_17> The last three? | 23:28 |
| mal | @edp_17 how does it behave if you didn't rename that lib? | 23:29 |
| T42 | <elros34> latest should be fine I think | 23:29 |
| T42 | <edp_17> Thanks @elros34 . | 23:29 |
| mal | are we really sure that is not needed? | 23:29 |
| mal | but what could cause the sigill in dsme | 23:30 |
| T42 | <edp_17> mal : Since I removed that the error disappeared from logcat. | 23:30 |
| T42 | <edp_17> I am not sure it is not needed. | 23:30 |
| mal | hmm, libcrypto, that sounds familiar | 23:30 |
| T42 | <edp_17> mal : When I had that "/usr/libexec/droid-hybris/system/lib/libcutils.so" file I got this in logcat: https://paste.ubuntu.com/p/PYJyNPWnyY/ | 23:31 |
| T42 | <edp_17> mal : I have this on device: https://paste.ubuntu.com/p/zB2Y2spZs7/ | 23:41 |
| mal | yeah, those are normal, just wondering why it fails, which device is that? | 23:51 |
| mal | and which kernel version? | 23:51 |
| T42 | <edp_17> Galaxy S4 (I9505) kernel 3.4 | 23:52 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!