| T42 | <Verevka86> @adampigg hi, did you manage to solve the problem with the microphone when making a phone call? | 06:10 |
|---|---|---|
| T42 | <adampigg> yes, it was the android audio-service running (re @Verevka86: @adampigg hi, did yo...) | 06:38 |
| T42 | <Verevka86> https://github.com/mer-hybris/droid-hal-configs/blob/master/sparse-11/usr/libexec/droid-hybris/system/etc/init/disabled_services.rc#L13 (re @adampigg: yes, it was the andr...) | 06:40 |
| T42 | <Verevka86> But was he stopped? | 06:41 |
| T42 | <adampigg> it was a bug in halium which didnt have that disabled ... newer builds do disable it (re @Verevka86: But was he stopped?) | 06:42 |
| T42 | <Verevka86> Strange, in ubuntu touch did not work for me either, I'll check again later | 06:43 |
| T42 | <adampigg> might be .. it was only fixed approx 2 weeks ago, so just update the halium rootfs to the latest | 06:47 |
| T42 | <adampigg> @eugenio_g7 @NotKit in the latest halium image, i noticed also that a usb service is disabled ... does that prevent /config being populated? | 07:15 |
| lotheac | trying to build for lena since a couple months, but it seems that https://github.com/mer-hybris/android/blob/sony-aosp-11/tagged-localbuild.xml#L18 refers to a commit that doesn't exist in sonyxperiadev/device-sony-common when cloning it. https://github.com/sonyxperiadev/device-sony-common/commit/137bf78caf8a464adcac102bdcb2d70c91c98509 | 07:54 |
| lotheac | i guess sony did some force pushing for that branch maybe? | 07:58 |
| T42 | <NotKit> on Volla it was crashing repeatedly. It should not be needed as config is handled host-side (re @adampigg: @eugenio_g7 @NotKit ...) | 08:09 |
| T42 | <edp_17> @elros34 : Hi, Can I get a little help with bluetooth, please? https://paste.ubuntu.com/p/pYDq7VV4sb/ | 08:13 |
| T42 | <edp_17> It doesn't turn on even when manually start brcm_patchram_plus. | 08:14 |
| T42 | <edp_17> bluetooth-rfkill-event-hciattach and broadcom-bluetooth-bluez5 packages are installed and (I think) all files are in place (/etc/bluez5 , /etc/firmware/*.conf , /etc/sysconfig/bluetooth-rfkill-event) | 08:18 |
| T42 | <adampigg> Yeah, because i removed the /init-debug, i lost the script which populates /config ... doesnt matter though, pinephone has a script to do it and I now have that working on pro1x (re @NotKit: on Volla it was cras...) | 08:19 |
| deathmist | mal: it appears you missed https://github.com/mer-hybris/android/pull/155 yesterday, would also be nice to have it for 18.1 whenever that becomes usable without extra local patches | 10:21 |
| mal | deathmist: like I said, it's not approved | 10:55 |
| deathmist | ah right... | 11:25 |
| mal | deathmist: merged now | 11:50 |
| mal | I need to check 18.1 at some point | 11:50 |
| T42 | <elros34> @edp_17 I do not see much point in fixing it until crash party in your log will stop but anyway at least you should do usual change: https://github.com/edp17/bluetooth-rfkill-event/blob/jfltexx4.4/bluetooth-rfkill-event/bluetooth_rfkill_event.c#L86 or even better make it configurable in config file. | 17:45 |
| T42 | <elros34> wrong line, should be 78 | 17:50 |
| T42 | <elros34> Try to remove privatenetwork from systemd-hostnamed.service too see if it will speed up boot | 18:07 |
| T42 | <edp_17> @elros34 : Thanks. I remember that changed bcm43xx to the correct one. Perhaps, I changed somewhere else. I'll do this + the systemd-hostnamed.service. | 18:13 |
| T42 | <edp_17> @elros34 : I've commented privatenetwork in systemd-hostnamed.service and this change speeded up the boot indeed. The gui started (circle animation) but the the device rebooted. | 19:08 |
| T42 | <elros34> so get journal from that failure | 19:08 |
| T42 | <edp_17> You know what's interesting? If I mask droid-hal-init.service, then let the device start and telnet in, then I can unmask the service and can start with systemctl. And it stays happy and eventually lipstick starts. | 19:10 |
| T42 | <edp_17> It doesn't make sense that a process fails but later can be started fine. | 19:11 |
| T42 | <edp_17> I'll get logs but probably late night or more likely tomorrow. | 19:12 |
| T42 | <elros34> sure, I thinkk journal will not be enough and you will have to also redirect dmesg to file. Race condition in boot sequence are not so uncommon | 19:14 |
| T42 | <elros34> @edp_17 btw as you also use glibc repo, should I create separate version or just rename 4.4.0.58 to 4.4.0.64? | 19:24 |
| T42 | <b100dian> > It doesn't make sense that a process fails but later can be started fine. | 20:23 |
| T42 | <b100dian> That process is an init one. Dozens of processes come out of it. | 20:23 |
| T42 | <edp_17> @elros34 : I use glibc for S2 (i9100) and N7000. I stopped building for the N7000, and 4.4 is currently broken on the S2. (Did a build but stuck on black screen, didn't have time to look into it yet. It could be that it has the same/similar issue to this droid-hal-init :). ) This means, you can rename 4.4.0.58 to 4.4.0.64 as I will do a fresh build anyway. Thank you! | 21:08 |
| T42 | <edp_17> How can I redirect dmesg into a file? | 21:09 |
| T42 | <elros34> need to think. dmesg -w works on your kernel? | 21:13 |
| T42 | <edp_17> It did the same as without the -w. It only gave me the list then got the prompt back. I guess this means, -w doesn't work, right? | 21:18 |
| T42 | <elros34> yes, try this: : https://irc.thaodan.de/.imgstore/9a88947c/dmesg_logger_0_0_1_1_armv7hl.rpm | 21:19 |
| T42 | <elros34> as root | 21:19 |
| T42 | <edp_17> Thanks. | 21:21 |
| T42 | <elros34> sorry, previous is probably broken : https://irc.thaodan.de/.imgstore/225cb4cc/dmesg_logger_0_0_1_1_armv7hl.rpm | 21:26 |
| T42 | <edp_17> Hm, I cannot save it from telegram. | 21:29 |
| T42 | <edp_17> The previous one saves but this last one doesn't. | 21:29 |
| T42 | <elros34> another way to download: https://1drv.ms/u/s!AuDqiTFly4jxgyzUL5Za0YnFthaB?e=wBGdy9 | 21:31 |
| T42 | <edp_17> Thanks, I could download and install. How does it work? | 21:34 |
| T42 | <elros34> just type dmesg_logger, it should work like dmesg -w on older kernel. If you append any argument like -l then it will redirect logs to file | 21:35 |
| T42 | <edp_17> I guess still need to initiate this as early as possible. | 21:36 |
| T42 | <edp_17> Okay, I am trying to fetch some logs. (droid-hal-init.service is unmasked, so just need a reboot.) | 21:37 |
| T42 | <elros34> I think something like: "nohup /usr/bin/dmesg_logger -l &" in /init-debug just before exec $INIT should do the job | 21:37 |
| T42 | <edp_17> Shouldn't I need to define a file name? Or where is it outputting the log? | 21:39 |
| T42 | <elros34> run it once to see whether it works without args and you will see filename | 21:40 |
| T42 | <elros34> with arg -l* | 21:41 |
| T42 | <edp_17> Got it (/var/log/dmesg_logger.log) but it is empty now. | 21:41 |
| T42 | <edp_17> Let me reboot to see whether there is anything and collect journal. | 21:42 |
| T42 | <edp_17> Here they are. These are what I could get before the device rebooted. dmesg: https://paste.ubuntu.com/p/89mTjdHXbW/ | 21:46 |
| T42 | <edp_17> journal: https://paste.ubuntu.com/p/4hVgRq295R/ | 21:46 |
| T42 | <elros34> but -b0 is for current boot, with -b -1 you can get full boot log from previous boot which failed | 21:48 |
| T42 | <edp_17> Oh, okay. | 21:48 |
| T42 | <edp_17> Here it is: https://paste.ubuntu.com/p/yyqGjwTzCV/ | 21:50 |
| T42 | <edp_17> The device keeps rebooting so this journal is not from the same as dmesg. Hope that isn't a problem. Or do you need dmesg + journal from the same boot? | 21:51 |
| T42 | <elros34> hm there is no power off/reboot logged. Have you checked /proc/last_kmsg? | 21:56 |
| T42 | <edp_17> No, I haven't. | 21:56 |
| T42 | <edp_17> Do you want me to get full journal then only upload since the last (let say) two reboots? | 21:57 |
| T42 | <elros34> just check whether it contains something about your reboot issue. I can see host name service starts nicely now | 21:58 |
| T42 | <edp_17> journalctl -f --no-tail: https://paste.ubuntu.com/p/9CmXpnGwJP/ | 22:02 |
| T42 | <edp_17> /proc/last_kmsg: https://paste.ubuntu.com/p/ptr529YZvF/ | 22:02 |
| T42 | <edp_17> /var/log/dmesg_logger.log: https://paste.ubuntu.com/p/Nd7gkr66WS/ | 22:02 |
| T42 | <edp_17> Oh, okay, I check that. | 22:02 |
| T42 | <elros34> so looks like kernel panic | 22:04 |
| T42 | <edp_17> Regarding firejail. | 22:05 |
| T42 | <edp_17> Or not? | 22:05 |
| T42 | <elros34> looks like | 22:07 |
| T42 | <elros34> have you tried disabling CONFIG_USER_NS? I do not have it enabled | 22:10 |
| T42 | <edp_17> I wonder, why there is no issue when I manually start droid-hal-init.service later? | 22:10 |
| T42 | <edp_17> Not yet. | 22:10 |
| T42 | <edp_17> Interesting as .config is different than defconfig. .config has a header while defconfig doesn't. | 22:12 |
| T42 | <elros34> dependencies | 22:12 |
| T42 | <elros34> hmm dsme also used to fail in this place but device didn;'t reboot: https://github.com/edp17/android_kernel_samsung_jf/blob/lineage-17.1/drivers/net/loopback.c#L205 | 22:15 |
| T42 | <edp_17> CONFIG_USER_NS depends on: CONFIG_EXPERIMENTAL & CONFIG_UIDGID_CONVERTED | 22:16 |
| T42 | <edp_17> In .config I only have CONFIG_EXPERIMENTAL. I'll try to unset CONFIG_USER_NS in defconfig to see how it looks like it .config. | 22:17 |
| T42 | <elros34> use xconfig or menuconfig, it will be easier | 22:18 |
| T42 | <elros34> ok get rid of this: https://github.com/edp17/android_kernel_samsung_jf/commit/93d8ad07a4938e3f6437ccd21007030f47cbf81e. fp2 kernel doesn't have it and IIRC mal said it works with 4.4.0 | 22:25 |
| T42 | <edp_17> I have the same line in loopback.c in both treltexx kernels (14.1 and 17.1). That device similarly reboots after a minute or so. | 22:25 |
| T42 | <edp_17> Although, that could be a different issue. | 22:25 |
| T42 | <elros34> could be but at least BUG_ON is added by this particular commit | 22:26 |
| T42 | <edp_17> Okay, I revert that and build hybris-hal. (re @elros34: ok get rid of this: ...) | 22:26 |
| T42 | <elros34> maybe this would be better if you care: https://github.com/mer-hybris/android_kernel_sony_msm/commit/e05e90702b2638a39b5ae9d22740f3a1607c54a0 | 22:34 |
| T42 | <edp_17> In my kernel I didn't have the change in include/net/net_namespace.h | 22:34 |
| T42 | <edp_17> How do you mean, keep the current kernel and only add that "net loopback: Set loopback_dev to NULL when freed"? | 22:35 |
| T42 | <elros34> yes but only if you have too much time, first try that revert | 22:36 |
| T42 | <edp_17> Okay, I've reverted. However, in "include/net/net_namespace.h" I didn't have that like "#define LOOPBACK_IFINDEX 1". This means, the "net: Loopback ifindex is constant now" change was not added correctly. | 22:37 |
| T42 | <elros34> about net_namespace.h: how? This is commit from your kernel | 22:37 |
| T42 | <edp_17> I don't know. | 22:38 |
| T42 | <elros34> ah it was moved by other commit | 22:40 |
| T42 | <edp_17> Yep. | 22:40 |
| T42 | <edp_17> Might be just need to add that one line back? | 22:40 |
| T42 | <elros34> this mean, simple reverting 1 commit might not work/not compile | 22:41 |
| T42 | <edp_17> I'll see. Before try, should I unset CONFIG_USER_NS or don't worry about it for now? | 22:42 |
| T42 | <elros34> if it's already set then it simple disabling it will not work | 22:43 |
| T42 | <elros34> with x/menuconfig you will know immediately whether it's possible | 22:43 |
| T42 | <edp_17> I've tried that xmenuconfig once but messed up the config so stopped using it. (I've used it for building linux kernels about 30 years ago though :) ) | 22:45 |
| T42 | <edp_17> Hybris-hal built (I pushed the change to github if you'd like to check) now I am installing it on device. | 22:53 |
| T42 | <edp_17> It seems to be working! :D | 23:05 |
| mal | ah, I remember that commit about loopback | 23:07 |
| T42 | <elros34> so PrivateNetwork=true now works? | 23:09 |
| T42 | <edp_17> Many thanks! Now I can try fixing other broken stuff (wifi keeps forgetting password, no vibra, BT, LEDs, Camera) :) Probably, I'll be back with some. :) | 23:09 |
| T42 | <elros34> quite a lot:) | 23:10 |
| T42 | <edp_17> I haven't tried that yet. I add them back and report. | 23:10 |
| T42 | <elros34> sure no rush | 23:10 |
| T42 | <edp_17> Yeah, many stuff don't work but I have time. Oh, I forgot to mention ofono which is currently masked. | 23:11 |
| T42 | <elros34> so you should mention what works instead:) | 23:11 |
| T42 | <edp_17> :) | 23:12 |
| T42 | <edp_17> Screen, touch, sound, wifi, gps, buttons (home and back), MTP (I need to manually restart usb-moded) | 23:14 |
| T42 | <edp_17> Oh, should I rebuild droidmedia? | 23:15 |
| T42 | <elros34> not unless you changed droidmedia revision | 23:17 |
| T42 | <edp_17> Nope, I haven't. Thanks for confirming. | 23:17 |
| T42 | <edp_17> It seems, systemd-hostnamed.service and dsme.service do not like PrivateNetwork=true. When I added, the device started to reboot. | 23:19 |
| T42 | <edp_17> It was enough to remove privatenetwork from systemd-hostnamed.service. In dsme, I can keep and the device doesn't reboot constantly. | 23:25 |
| mal | kernel might need some namespace fixes, I had to do some for fp2 | 23:42 |
| T42 | <edp_17> mal : Yeah, I remember you mentioned it. What fixes did you do? | 23:45 |
| T42 | <edp_17> Now bluetooth-rfkill-event.service is happy and brcm_patchram_plus starts/stops when I toggle Bluetooth. However, it doesn't work becuase bluetooth.service crashes. | 23:46 |
| T42 | <edp_17> What is this "/usr/libexec/bluetooth/obexd" and why do I need it? | 23:54 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!