T42 | <oMinimalist> https://pastebin.com/U3zmBHMS | 01:00 |
---|---|---|
rinigus | Thaodan I wasn't aware we have any major issues with Bluetooth. | 04:38 |
T42 | <Mad_on_Pubg> /notes | 05:08 |
Thaodan | rinigus: Well there is a native driver so I tried out to see what's left that is needed to use it. | 08:04 |
Thaodan | From what I see its' device tree nodes. | 08:04 |
rinigus | Thaodan: so, did this native driver work for you? so, if it works fine, we can drop bluebinder, right? | 08:13 |
Thaodan | rinigus: I need to figure which uart to attach to, the device tree misses that information. | 08:16 |
Thaodan | For mainline there is sdm45 working with that. I took the device tree node from there and it did not work but maybe does for you. | 08:18 |
rinigus | Thaodan: thanks! will have to keep that in mind when I get to kernel updates. but I must say that I am reluctant to touch something which works just fine as it is :) | 08:25 |
kquote03 | Hello | 09:35 |
kquote03 | I've done a little hack I'm not proud of, but did get me further, I added chmod 666 /dev/*binder*; to the top of droid-hal-early-init.sh | 09:36 |
kquote03 | it can access binder devices fine now | 09:37 |
kquote03 | but it's complaining about something else | 09:37 |
kquote03 | https://susepaste.org/67876881 here's logcat | 09:37 |
kquote03 | I've already symlinked /apex/..../lib64 to /odm/lib64 | 09:38 |
kquote03 | launching lipstick now segfaults | 09:39 |
kquote03 | https://susepaste.org/92951763 | 09:39 |
T42 | <TheVancedGamer> oh it's so linked to lineage (re @SailfishFreenodeIRCBridgeBot: <kquote03>https://su...) | 09:39 |
T42 | <TheVancedGamer> is vendor from lineage? | 09:39 |
kquote03 | can you clarify? my phone doesnt have a /vendor partition | 09:40 |
kquote03 | so i symlinked /system/vendor to /vendor | 09:40 |
T42 | <TheVancedGamer> ah | 09:40 |
T42 | <TheVancedGamer> it wants a fast charge lib | 09:40 |
kquote03 | I saw that but i shrugged it off, doesnt seem like something that would cause it fail? At least I guess? | 09:41 |
T42 | <TheVancedGamer> Strace test_hwcomposer first IG | 09:44 |
T42 | <elros34> I told you use minimer instead lipstick, it's easier to debug and strace. Also if you would disable audit via cmdline that would clean up your logs | 09:48 |
kquote03 | Alright! | 09:49 |
T42 | <elros34> libandroidicu.so symlinks is also missing | 09:50 |
kquote03 | elros34: https://susepaste.org/24299281 it's not tho? | 09:52 |
T42 | <elros34> looks goo but it still complain | 09:54 |
kquote03 | yea :/ | 09:54 |
T42 | <elros34> you would have to strace -f -e file /usr/libexec/droid-hybris/system/bin/minimediaservice | 09:56 |
kquote03 | Alright, will update you once kernel is recompiled. Thanks for your help!! | 09:57 |
kquote03 | Alright, I ran the strace command for libandroidicu.so https://susepaste.org/45026537 | 10:13 |
kquote03 | i'm getting strace output for minimer | 10:14 |
T42 | <elros34> is your minimediaservice 32bit? | 10:14 |
kquote03 | i dont think so? | 10:15 |
T42 | <elros34> then check? | 10:15 |
kquote03 | gimme a min | 10:15 |
kquote03 | it is 32 bit | 10:15 |
T42 | <elros34> you can clearly see that is search for lib in /odm/lib | 10:15 |
T42 | <elros34> so add another symlink | 10:16 |
kquote03 | will do | 10:16 |
kquote03 | alright, running the strace command you gave no longer crashes | 10:16 |
kquote03 | again, thanks a lot! | 10:16 |
*** Daaanct12 is now known as Danct12 | 10:21 | |
kquote03 | minimer strace https://susepaste.org/1326789 (idk which package provices qt5/accessible, i have to leave now sadly) | 10:26 |
T42 | <elros34> /usr/lib64/qt5/bin/accessible doesn't looks like anything important | 10:28 |
kquote03 | turns out i had 5 more minutes so i took strace of hwcomposer as well just in case https://susepaste.org/80325912 | 10:39 |
kquote03 | notice it crashes right after accessing /dev/binder ? | 10:39 |
T42 | <elros34> there must be something more useful in logcat or journal/dmesg | 10:41 |
kquote03 | alright gimme a sec | 10:41 |
T42 | <elros34> grep for mali-hist-dump in your /vendor/lib64/egl/libGLES_mali.so. If it's found then make sure you build libhybris with mali quirks | 10:42 |
kquote03 | https://susepaste.org/14492343. it is found | 10:44 |
kquote03 | but i gotta leave | 10:44 |
T42 | <elros34> @ kquote03: I think mali-hist-dump should be also in /usr/lib64/libhybris-common.so.1.0.0 | 10:51 |
kquote03[m] | Hey I don’t have access to my equipment now, but for when I return , when I apply mali quirks, I only rebuild droid-Hal correct? | 11:39 |
T42 | <elros34> nope, IIRC droid-hal only adds quirks to header file which is used when building libhybris | 11:43 |
kquote03 | elros34: Hey, if i ask why the second command on https://github.com/mer-hybris/hadk-faq#43devices-with-mali-gpu is a mount command? | 15:28 |
kquote03 | mind if i ask* | 15:28 |
T42 | <elros34> probably mistake when faq was moved form one place to another | 15:43 |
T42 | <elros34> from* | 15:43 |
T42 | <elros34> that mount is from different place: https://wiki.merproject.org/wiki/Adaptations/faq-hadk | 15:45 |
kquote03 | Ah, thanks! | 15:45 |
T42 | <oMinimalist> Anyone has porting sfos on any pi single board computer? | 15:48 |
kquote03 | I rebuilt the packages and copied them over, but some of the packages it's asking me to install as new (they were not installed from before) https://susepaste.org/9777703 maybe that was the problem?? | 16:06 |
kquote03 | minimer still segfaults tho | 16:11 |
kquote03 | https://susepaste.org/52714004 strace of minimer | 16:15 |
T42 | <elros34> you should not install packages which were not installed already, especially *-devel packages> have you checked whether mali-hist-dump is now defined in lib I asked? | 16:23 |
kquote03 | i have not, how do you do that? | 16:24 |
kquote03 | ill search it | 16:25 |
T42 | <elros34> grep or strings should be enough | 16:25 |
kquote03[m] | so i cat the file and grep it? | 16:25 |
T42 | <elros34> why cat, its not text file just grep "pattern" /path/to/file | 16:26 |
kquote03[m] | i dont use grep efficiently :( | 16:27 |
kquote03 | yep it's there https://susepaste.org/7724742 | 16:28 |
kquote03 | anyways for now i will reinstall rootfs becuase i messed with the packages | 16:34 |
kquote03 | then i will give logcat | 16:35 |
T42 | <elros34> waste of time if device still boot up:) | 16:39 |
kquote03 | thank god my cat decided to jump on my laptop, causing me to waste time initially, but causing me to gain more time because I put off reinstalling the whole system | 16:40 |
kquote03 | current logcat https://susepaste.org/35251305 | 16:44 |
kquote03 | currentl logcat after executing minimer https://susepaste.org/68565226 | 16:47 |
T42 | <elros34> this is disaster, have you tried mention audit=0? | 16:49 |
kquote03 | yes | 16:50 |
kquote03 | https://susepaste.org/34852195 | 16:50 |
T42 | <elros34> are you sure about syntax? cat /proc/cmdline use spaces instead comma | 16:52 |
kquote03 | .... | 16:52 |
kquote03 | oh no | 16:52 |
T42 | <elros34> do you even see values you set in /proc/cmdline? | 16:54 |
kquote03 | gimme a sec | 16:54 |
kquote03 | yes i do | 16:59 |
kquote03 | but they're separated by commas XD | 16:59 |
kquote03 | yea time to recompile | 16:59 |
kquote03 | (everything is separated by spaces) | 16:59 |
T42 | <elros34> after that get full dmesg/journal/logcat logs after reboot | 17:02 |
kquote03 | alright | 17:02 |
kquote03 | logcat https://susepaste.org/120979 | 17:08 |
kquote03 | (after running minimer once) | 17:08 |
T42 | <elros34> either lipstick or mask user@100000 and then play with minimer/test_hwcomposer/surfaceflinger | 17:10 |
kquote03 | i mask lipstick? | 17:11 |
kquote03 | i did anyway | 17:11 |
kquote03 | but wait ill put the rest of the pasted | 17:11 |
kquote03 | dmesg https://susepaste.org/50242132 | 17:12 |
kquote03 | journalctl https://susepaste.org/99135890 | 17:13 |
kquote03 | I maked both. here's strace of minimer https://susepaste.org/76579862 (it turned the screem black) | 17:20 |
kquote03 | (did restart) | 17:20 |
T42 | <elros34> it crash same way so stracing it will not give us more information | 17:20 |
kquote03 | i see | 17:21 |
T42 | <elros34> as you can see "KNOX: The mark is out of range" is everywhere so logs are lost so I suggest to get rid of it or/and enable journal persistent logging. /usr/bin/surfaceflinger is next to test | 17:21 |
kquote03 | If you want I can just delete all KNOX lines before uploading | 17:21 |
kquote03 | takes 2 secs in emacs | 17:22 |
T42 | <elros34> that will not bring lost messages | 17:22 |
kquote03 | but ultimately I will find a way to disable knox because it is not necessary | 17:22 |
kquote03 | elros34: can you clarify? | 17:22 |
T42 | <elros34> journalctl should have all messages since boot | 17:23 |
kquote03 | but doesnt journalctl have the KNOX spam as well? | 17:23 |
T42 | <elros34> yes it has that is why early boot messages are lost | 17:24 |
kquote03 | Ah | 17:24 |
kquote03 | elros34: I enabled persistent logging. The file is 4.6 megabytes in size. No paste service accepts it. And I wouldn't expect a mere mortal to sift through 5 MB of pure log hell | 20:20 |
kquote03 | Also I do not have /usr/bin/surfaceflinger | 20:26 |
kquote03 | nvm its in /system/usr/bin | 20:29 |
kquote03 | strace of surfaceflinger https://susepaste.org/99846940 (it ends like that) | 20:48 |
kquote03 | tho idk how helpful that is | 20:48 |
kquote03 | is it normal that lsmod does not print anything? | 20:51 |
kquote03 | dmesg where I quickly ran surfaceflinger and removed all KNOX spam https://susepaste.org/30814129 | 21:05 |
kquote03 | will try again but with journalctl | 21:05 |
kquote03 | same but with journalctl https://susepaste.org/28976908 | 21:14 |
T42 | <elros34> about lsmod: it depends on whether you have any modules enabled in kernel. I wonder whether you need to remove bootanim from disabled_services.rc to get surfaceflinger working | 21:46 |
kquote03 | ill do that anyway | 21:47 |
kquote03 | comments for the rc file are by # right? | 21:48 |
T42 | <elros34> yes, according to /init*rc | 21:49 |
kquote03 | alright, ill reboot now | 21:49 |
kquote03 | no dice | 21:50 |
kquote03 | will post strace in a bit | 21:50 |
kquote03 | https://susepaste.org/57984266 looks the same tho | 21:53 |
T42 | <elros34> at least it not crash like other. Check other logs also, bootanim should be started. | 21:58 |
kquote03 | alright, will reboot tho because of exynos spam | 21:59 |
T42 | <elros34> why don't you just comment it out from kernel sources? You can try one more thing, some devices needs to run surfaceflinger for a while, then stop it and run minimer | 22:00 |
kquote03 | i setting the knox kernel config to n but that changed nothing | 22:01 |
kquote03 | for a while then i stop it manually? | 22:02 |
T42 | <elros34> yes | 22:03 |
kquote03 | alright | 22:03 |
T42 | <elros34> grep for that message in kernel source and comment it out. There is no any ifdef for this knox spam, somebody already had same issue | 22:04 |
kquote03 | you mean with rgrep? | 22:04 |
kquote03 | or can normal grep do that/ | 22:04 |
T42 | <elros34> grep -rn | 22:04 |
kquote03 | honestly im learning a lot. i did not expect that | 22:05 |
kquote03 | thanks | 22:05 |
T42 | <elros34> you know you could just use gnome-search-tool it's very useful also | 22:06 |
kquote03 | i will try it, thanks for the recommendation | 22:07 |
kquote03 | is now a good time to stop surfaceflinger? | 22:09 |
T42 | <oMinimalist> Has anyone ported sfos on any pi single board computer? | 22:15 |
kquote03 | not that im aware of? (only answering because you asked this question twice now) | 22:17 |
kquote03 | (otherwise probably not the guy most qualified to give an answer) | 22:17 |
T42 | <elros34> @oMinimalist I think there was something long time ago. Quick search shows https://techcresendo.com/sailpi-sailfish-os-on-raspberry-pi/ | 22:22 |
kquote03 | alright i stopped surfaceflinger, minimer sigfaults like usual | 22:22 |
T42 | <elros34> no good | 22:23 |
kquote03 | ill reboot into the new kernel with the knox message commented out now btw | 22:24 |
T42 | <oMinimalist> Thanks. That’s too old. (re @elros34: @oMinimalist I think...) | 22:24 |
kquote03 | I mean the raspberrypi can run android, even builds of LOS exist, doesnt seem out of the question to port hybris/hal to it ? | 22:25 |
kquote03 | dmesg after runnnig minimer https://susepaste.org/16778178 | 22:34 |
kquote03 | i do not like the look of that 'cut here' error | 22:36 |
kquote03 | https://susepaste.org/7611268 surfaceflinger. same cut here irq error :/ | 22:46 |
Thaodan | kquote03: Why would you so that when you can have native port? | 22:56 |
Thaodan | You can see the pinephone port as an example | 22:57 |
kquote03 | because big dumb (tbf it's 2 am here. i am sleep deprived) | 22:57 |
Thaodan | np maybe you want to have hybris port :D | 22:58 |
kquote03 | you mean on the pi? | 22:59 |
kquote03 | sorry my reading comprehension is just broken today | 22:59 |
Thaodan | yes | 22:59 |
kquote03 | btw shall I try porting using an older version of LOS / hybris? Would that change anything? | 23:01 |
T42 | <oMinimalist> https://t.me/sailfishos_porters/128450 | 23:05 |
T42 | <kquote03> I see, will avoid building on older version. Thanks for the heads up | 23:07 |
T42 | <elros34> it was about old sfos release not hybris/lineage base | 23:14 |
T42 | <elros34> @kquote03 I would advice you to try qpa-surfaceflinger-plugin instead qpa-hwcomposer-plugin but unfortunately it's abandoned and I am not sure there is good intruction for it | 23:21 |
T42 | <kquote03> Ah (re @elros34: it was about old sfo...) | 23:22 |
T42 | <kquote03> Any particular reason to try it? (re @elros34: @kquote03 I would ad...) | 23:22 |
T42 | <elros34> only one: hwcomposer fails | 23:23 |
T42 | <kquote03> Sorry just read the rules won’t use reply | 23:23 |
T42 | <kquote03> I see, so minimer and current surfaceflinger depend on hwcomposer then | 23:24 |
mal | how does it fail? | 23:34 |
T42 | <kquote03> mal: it segfaults | 23:48 |
T42 | <kquote03> Here’s dmesg after running minimer https://susepaste.org/16778178 and here’s strace https://susepaste.org/76579862 | 23:50 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!