Saturday, 2022-08-20

T42<kreatowastaken> i still couldnt fix it...01:32
T42<kreatowastaken> idk why it suddenly just broke01:32
T42<kreatowastaken> i just did make hybris-hal droidmedia and build_packages.sh01:32
T42<edp_17> @kreatowastaken : regarding that signal: 7 eror. I got this many times. You need to use an older version of TWRP.07:56
T42<kreatowastaken> i dont use TWRP07:57
T42<kreatowastaken> and i searched irc logs07:57
T42<kreatowastaken> the people that got signal 7 error got it because tar OOM or some stuff like that which i dont have that same issue07:57
T42<kreatowastaken> atleast didnt until i decided to rebuild the zip07:57
T42<kreatowastaken> i use lineageos recovery07:58
T42<kreatowastaken> and busybox bzip207:58
T42<kreatowastaken> it says signal 7 when its unmounting the system, not when it extracts the rootfs which what most people had as a issue08:00
T42<kreatowastaken> the oldest version TWRP i have is 3.3.008:01
T42<edp_17> I see. Might worth to try with that twrp...08:05
T42<kreatowastaken> i have no idea how that would help anything but ill try it08:05
voidanix[m]@Thaodan : so, have you tried importing some of the xiaomi changes for sdx55 to see if it fixes the modem?08:06
T42<kreatowastaken> wait08:07
T42<kreatowastaken> it says footer is wrong08:07
T42<kreatowastaken> for some reason its building the zip incorrectly08:07
T42<kreatowastaken> i guess08:08
deathmist1@kreatowastaken so have you tried to extract the zip via adb shell in recovery and run the tar command manually?
T42<kreatowastaken> i didnt08:11
T42<kreatowastaken> ill try that08:11
T42<kreatowastaken> what do i do next?08:14
deathmist1idk, assuming you extracted it successfully to the right dir I guess flash hybris-boot.img and reboot08:14
T42<kreatowastaken> alright08:14
T42<kreatowastaken> deathmist1: that did the trick, thanks08:28
T42<kreatowastaken> droid-hal-init not starting properly, and lipstick having library errors09:01
T42<kreatowastaken> how can i fix these?09:01
T42<kreatowastaken> journalctl | grep droid output09:01
T42<kreatowastaken> droid-hal-init not starting properly, and lipstick having library errors09:01
T42<kreatowastaken> how can i fix these?09:01
T42<kreatowastaken> `journalctl | grep droid` output09:01
T42<kreatowastaken> (edited)09:01
T42<elros34> have you tried following hadk-hot advices? Firs you need to fix droid-hal-init09:06
T42<kreatowastaken> i  checked the lib "" not found error part of it09:08
T42<kreatowastaken> but idk if my device uses /odm or not09:08
T42<kreatowastaken> should i strace droid-hal-init first?09:09
T42<elros34> then check: ls /odm09:09
T42<kreatowastaken> no such file or directory09:09
T42<kreatowastaken> hm09:09
T42<kreatowastaken> but /system_root had it so09:09
T42<kreatowastaken> i thought it may have it09:09
T42<elros34> follow instruction about uevent*rc then get full dmesg/journal and do not filter logs09:10
T42<kreatowastaken> alright09:10
T42<kreatowastaken> do i just build everything thats on make modules?09:14
T42<elros34> of course no only uevent*  and only if it's missing in out directory09:16
T42<kreatowastaken> so stuff like `ltp_uevent01`09:17
T42<kreatowastaken> ok09:17
T42<elros34> no, not really, more like ueventd.qcom09:17
T42<kreatowastaken> well09:17
T42<kreatowastaken> i dont have anything like that09:17
T42<elros34> so that case is also explained, did you bother to read it all?09:18
T42<kreatowastaken> i cant see anything about that09:19
T42<kreatowastaken> i couldnt find anything about that09:21
T42<kreatowastaken> sorry if i misread09:21
T42<kreatowastaken> ok09:25
T42<kreatowastaken> on `./out/target/product/beyond1lte/recovery/root/system/etc/ueventd.rc` theres a ueventd file09:25
T42<kreatowastaken> but thats it09:25
T42<kreatowastaken> dont know if that is enough or not09:25
T42<elros34> then read droid-hal-device.log to figure out whether it's used09:27
T42<kreatowastaken> alright09:27
T42<kreatowastaken> it is09:28
T42<kreatowastaken> theres this warning though09:29
T42<kreatowastaken> `WARNING Rule not converted: firmware_directories /etc/firmware/ /odm/firmware/ /vendor/firmware/ /firmware/image/`09:29
T42<kreatowastaken> should i worry about this?09:29
T42<kreatowastaken> and theres these09:29
T42<kreatowastaken> anyway09:37
T42<kreatowastaken> heres the journalctl output09:37
T42<kreatowastaken> anyway09:37
T42<kreatowastaken> heres the journalctl output09:37
T42<kreatowastaken> @elros34 (edited)09:37
T42<elros34> you waited to long get get journal, initial logs are lost09:56
T42<elros34> what permisions do you have on /dev/kmsg_debug? grep for kmsg_debug in /system_root and /vendor in uevent*files. Is there any entry?10:00
T42<kreatowastaken> `crw--w--w-    1 root     root        1,  11 Aug 20 13:02 /dev/kmsg_debug`10:03
T42<kreatowastaken> permissions10:03
T42<kreatowastaken> `cat /system_root/etc/ueventd.rc | grep kmsg_debug` and `cat /vendor/ueventd.rc | grep kmsg_debug` gave no output10:04
T42<elros34> maybe it's not worth to worry about it now, before you get full logs10:06
T42<kreatowastaken> hope this is full,  i got it as fast as i could after telnet opened10:08
T42<elros34> not really but good enough to see that apex fails. Read 'Generating logs' section if you still didn't. What about dmesg?10:17
T42<kreatowastaken> alright ill read that10:17
T42<elros34> you should read everything that way you will save your time and mine10:18
ThaodanAny idea why my modem should power down from mhi shortly after booting/11:07
Thaodanhm is wifi also connect over mhi through pci on phones?11:09
T42<kreatowastaken> @elros34 i checked out the dmesg, found one bug that was on hadk-hot and that i could fix but couldnt find anything else11:20
T42<kreatowastaken> can i send the dmesg through tg? I cant send it through hastebin because it exceeds file size11:20
T42<kreatowastaken> @elros34 i checked out the dmesg, found one bug that was on hadk-hot and that i could fix but couldnt find anything else11:22
T42<kreatowastaken> can i send the dmesg through tg? I cant send it through hastebin because it exceeds line size (edited)11:22
voidanix[m]afaik no, the soc has a fastconnect subsystem for wifi/bt, only the 5g modem is external11:53
Thaodanvoidanix[m]: Yeah that makes sense, just the mix  cnss = wifi/bt and cnss = rmnet/modem is confusing.11:56
ThaodanA lot of erros are just noise which makes it hard to differenciate11:56
ThaodanI'm not sure about these two:11:57
Thaodan[    0.792851] cnss: Unable to create CNSS MHI IPC log context11:57
Thaodan[    0.792855] cnss: Unable to create CNSS MHICNTRL IPC log context11:57
ThaodanAlso why this [   10.666121] cnss: Setting MHI state: POWER_OFF(4)11:59
ThaodanThe modem shouldn't go off11:59
T42<elros34> @kreatowastaken then use other paste service or gist. BTW what bug?12:07
T42<elros34> gist github12:07
T42<kreatowastaken> alright ill use github gist then12:07
T42<kreatowastaken> it was `   apexd: Failed to LOOP_SET_BLOCK_SIZE: Try again'`12:07
T42<kreatowastaken> i applied the pr and it seems to be gone now so i guess that is fixed12:08
T42<elros34> so apex doesn't fail anymore?12:09
T42<kreatowastaken> seems so12:09
T42<kreatowastaken> ok maybe not but that spesific error is gone12:11
T42<kreatowastaken> heres the gist12:11
T42<TheVancedGamer> what is your kernel cmdline?12:13
T42<TheVancedGamer> do you have `loop.max_parts=7`?12:13
T42<kreatowastaken> i dont12:14
T42<TheVancedGamer> add it12:14
T42<elros34> why did you disable selinux?12:18
T42<kreatowastaken> i did not12:18
T42<elros34> ok its in permissive, I do not remember what is the recommended state for hybris18.12:19
T42<elros34> I wonder because  "Check failed: selinux_status_open" in your logs12:20
T42<kreatowastaken> @TheVancedGamer i did, what next12:21
T42<TheVancedGamer> rebuild kernel and flash12:21
T42<kreatowastaken> i did12:21
T42<TheVancedGamer> difference in apex logs should be obvious now12:22
T42<TheVancedGamer> it should mount all apexes12:22
T42<kreatowastaken> yeah12:22
T42<TheVancedGamer> send logs now12:22
T42<kreatowastaken> okay12:22
T42<elros34> where do you see apex mount failure?12:23
T42<TheVancedGamer> this error is widely regarded as the most confusing one12:24
T42<TheVancedGamer> but it just ran out of loop blocks12:24
T42<elros34> on loop2?:)12:24
T42<TheVancedGamer> well, yes12:25
T42<TheVancedGamer> (I took reference from AOSP docs though)12:25
T42<TheVancedGamer> wait12:25
T42<TheVancedGamer> they're still there12:25
T42<elros34> there is no any kernel cmd line chane12:28
T42<kreatowastaken> i just changed CONFIG_CMDLINE12:29
T42<kreatowastaken> maybe i shouldve rebuilt everything12:29
T42<kreatowastaken> ok forgot a config option, my fault12:30
T42<kreatowastaken> rebuilding and resending now12:30
T42<elros34> Did you see anything in CONFIG_CMDLINE before you change it? Doubt it's used12:31
T42<TheVancedGamer> wait12:31
T42<kreatowastaken> i forgot CONFIG_CMDLINE_EXTEND12:31
T42<TheVancedGamer> it's `loop.max_part=7`, not `max_parts=7`12:31
T42<kreatowastaken> lol guess its the perfect time to correct it then12:31
T42<kreatowastaken> ok now it is on the cmdline12:36
T42<kreatowastaken> ill send dmesg again12:36
T42<TheVancedGamer> I recognize that vndservicemanager panic12:40
T42<TheVancedGamer> why did you disable selinux?12:40
T42<kreatowastaken> i didnt12:40
T42<TheVancedGamer> it fails because of missing selinux shims12:40
T42<TheVancedGamer> oh12:41
T42<TheVancedGamer> watchdog12:41
T42<TheVancedGamer> do you have telnet?12:41
T42<kreatowastaken> i do12:42
T42<TheVancedGamer> `ls -ltha /dev/watchdog`12:42
T42<kreatowastaken> `crw-------    1 root     root       10, 130 Aug 20 15:42 /dev/watchdog`12:42
T42<elros34> error is quiet clear, its something selinux related and not only for watchdog but for servicemanager: selinux_status_open fails12:43
T42<TheVancedGamer> I'm more curious if selinux was even built on android side12:43
T42<TheVancedGamer> do you have ``?12:44
T42<kreatowastaken> on where?12:44
T42<TheVancedGamer> anywhere12:44
T42<kreatowastaken> yeah12:44
T42<TheVancedGamer> where?12:44
T42<TheVancedGamer> did you apply hybris-patches?12:45
T42<kreatowastaken> yes12:45
T42<TheVancedGamer> pastebin the file contents of /system/etc/init/vndservicemanager.rc12:46
T42<kreatowastaken> no such file or directory12:46
T42<TheVancedGamer> find a vndservicemanager.rc12:47
T42<kreatowastaken> theres one on vendor12:47
T42<kreatowastaken> ok12:47
T42<elros34> Did you define android_version_major in droid config?12:47
T42<TheVancedGamer> why is that file in your vendor12:47
T42<kreatowastaken> idk12:47
T42<TheVancedGamer> it's supposed to be always in system12:47
T42<kreatowastaken> droid-hal-init.service                                                    loaded active running   droid-hal-init13:58
T42<kreatowastaken> i got droid-hal-init working lets goo13:58
T42<kreatowastaken> but lipstick still doesnt work14:00
T42<elros34> mind sharing solution with others?14:01
T42<kreatowastaken> following the faq fully xD (re @elros34: mind sharing solutio...)14:01
T42<kreatowastaken> i didnt add the sparse stuff thats on faq14:01
T42<elros34> what sparse stuff you added for hybris18?14:02
T42<kreatowastaken> "Add to sparse: You may need to replace symlinks with the actual files from your device."14:02
T42<kreatowastaken> this14:02
T42<kreatowastaken> i didnt need to replace the symlinks14:03
T42<kreatowastaken> but as i said lipstick doesnt work yet14:03
T42<kreatowastaken> atleast progress though14:03
T42<elros34> so this fixed selinux_status_open failure?14:04
T42<kreatowastaken> yep14:04
T42<kreatowastaken> `user@100000.service` is running14:05
T42<kreatowastaken> but no gui14:05
T42<elros34> did you add android_version_major?14:06
T42<kreatowastaken> i did14:06
T42<kreatowastaken> as 11 though14:06
T42<kreatowastaken> instead of 10 thats on the faq14:07
T42<kreatowastaken> is that ok?14:07
T42<elros34> you use android version so yes14:07
T42<kreatowastaken> should i try to mask user@100000?14:09
T42<elros34> what for, you said it works:)14:10
T42<elros34> find failure error14:10
T42<elros34> in journal or logcat14:10
T42<kreatowastaken> `droid-bootctl.service                                                       loaded failed failed    Droid bootctl`14:13
T42<kreatowastaken> this failed it seems14:13
T42<kreatowastaken> also dev-binderfs.mount and dev-mtp.mount doesnt work but are those even required14:13
T42<elros34> but why lipstick fails14:14
T42<kreatowastaken> `Aug 20 17:15:22 GalaxyS10 droid-hal-init: Could not restart 'surfaceflinger': Cannot find 'surfaceflinger_HYBRIS_DISABLED': No such file or directory`14:15
T42<kreatowastaken> maybe this is related?14:15
T42<elros34> no it's not14:17
T42<elros34> `>_HYBRIS_DISABLED`<14:17
T42<kreatowastaken> okay14:17
malhow does lipstick fail?14:18
T42<kreatowastaken> idk14:19
T42<kreatowastaken> gui just doesnt come out14:19
T42<kreatowastaken> even though stuff seems right14:19
malwhat is in journal log?14:22
T42<kreatowastaken> i couldnt find any relevant errors14:27
T42<elros34> no error, so lipstick doesn't fail?14:32
T42<kreatowastaken> i guess?14:33
T42<kreatowastaken> but gui isnt there14:33
T42<kreatowastaken> just a black screen14:33
T42<elros34> you guess? just check logs14:33
T42<kreatowastaken> i did14:34
T42<kreatowastaken> i couldnt find anything bout lipstick14:34
malis lipstick running?14:39
T42<kreatowastaken> no14:40
T42<kreatowastaken> just checked14:40
malso then something should be in logs telling what happened to it14:41
malhow about you show the full journal log?14:41
T42<kreatowastaken> mal:
T42<kreatowastaken> sorry for the delay14:54
T42<hasajacy> can i use it somehow to build latest sailfish  version?16:34
T42<hasajacy> or can i install order version  and just run  somethink like apt dist-upgrade?16:35
T42<hasajacy> or can i install order version  and just run  somethink like `apt dist-upgrade`? (edited)16:35
T42<hasajacy> or can i install order version  and just run  something like `apt dist-upgrade`? (edited)16:37
mal@kreatowastaken there seems to be a lot of android side processes failing16:47
T42<KernelPanix> hi16:59
T42<KernelPanix> i need help16:59
malwhat is the problem?17:07
mal@KernelPanix edit rpm/dhd/helpers/ like above this comment
T42<KernelPanix> yes this work. thks (re @SailfishFreenodeIRCBridgeBot: <mal>@KernelPanix ed...)18:19
T42<11> hello guys, there is such a problem that after running dsme I get a kernek bug , build based on lineage-14.1 ,21:17
T42<11> 21:17
T42<11> here is the error itself:
T42<11> 21:17
T42<11> here is the line on the git that the error points to:
T42<11> 21:17
T42<11> 21:17
T42<11> 21:17
T42<11> if anyone can help i would be very grateful21:17
malI remember seeing that error on one of my ports, can't remember exactly how I fixed it, which 3.4 kernel do you have?21:19
malalso please do use so many empty lines when posting, it looks terrible on irc side21:20
T42<11> yes if that EleventhLast (re @SailfishFreenodeIRCBridgeBot: <mal>I remember seei...)21:20
T42<11> oh, sorry . okey (re @SailfishFreenodeIRCBridgeBot: <mal>also please do ...)21:20
T42<11> yes, I have a kernel 3.4 msm897421:21
malah, 3.4.11321:22
T42<11> yes21:22
T42<11> It's been 3 weeks and I can't figure it out.21:22
T42<11> deleting the line that calls this function is not an option, then there are a lot of kmem errors21:23
T42<11> mal , so do you remember how you solved this problem?21:26
T42<Herrie1982> We had quite some issues with 3.4 kernels and newer systemd (re @11: mal , so do you reme...)21:28
T42<Herrie1982> We added some patches at our end for LuneOS kernel21:28
T42<11> Can I have links to patches?21:29
T42<Herrie1982> What does dsme do? I don't know SFOS that well, and readme in the repo is not very useful ;)21:29
T42<Herrie1982> Let me check (re @11: Can I have links to ...)21:30
T42<11> I don’t know myself, but as far as I understand it is something like systemd-hostnamed (re @Herrie1982: What does dsme do? I...)21:30
T42<Herrie1982> OK21:31
T42<Herrie1982> Well these are the patches we carry on 3.4 for Hammerhead:
T42<11> If I understand correctly, do I need to add all these patches? (re @Herrie1982: Well these are the p...)21:32
T42<Herrie1982> No, probably not21:32
T42<Herrie1982> I'm not sure what the exact problem is, but we had some issues with some of the namespace stuff for which there's a bunch of patches. But it might be more file system related here.21:32
T42<Herrie1982> So I would start with these:,
mal@11 maybe try backport this
T42<Herrie1982> Or this one from Tenderloin:
T42<11> I'm not sure what the exact problem is, but we had some issues with some of the namespace stuff for which there's a bunch of patches. But it might be more file system related here. (re @SailfishFreenodeIRCBridgeBot: <mal>@11 maybe try b...)21:35
mal@Herrie1982 dsme is a very low level system state handling process, one of the most critical ones, this may explain it
T42<Herrie1982> h OK (re @SailfishFreenodeIRCBridgeBot: <mal>@Herrie1982 dsm...)21:37
mal@Herrie1982 that kernel is 3.4.113 which already has some of those patches, except that revert one21:43
mal@11 so try added the patch I linked and maybe also this
T42<Herrie1982> We kept 3.4 alive as long as we could at our end but seems we're at a dead end now with our new compositor and Qt6 so venturing into mainline for these targets21:44
mal@Herrie1982 yeah, 3.4 is starting have all sorts of issues, I'm doing some mainline porting also21:45
mal@Herrie1982 right now cleaning up msm8974 iommu support code I have locally21:46
malthat would help with hammerhead also :)21:47
T42<11> should I apply this? (re @SailfishFreenodeIRCBridgeBot: <mal>@11 so try adde...)21:48
mal@11 apply these and
T42<Herrie1982> Hammerhead seems to boot OK at our side but some graphic glitches still (re @SailfishFreenodeIRCBridgeBot: <mal>that would help...)21:49
T42<11> I've already used the first one, but I'm not sure about the second one. (re @SailfishFreenodeIRCBridgeBot: <mal>@11 apply these...)21:50
mal@Herrie1982 I assume without gpu? or do you have some patches for it21:50
mal@Herrie1982 or more like without iommu, btw do you use FD_DBG_NOBIN for mesa?21:52
T42<Herrie1982> Well we have UI at our side with Qt6 from what I recall (re @SailfishFreenodeIRCBridgeBot: <mal>@Herrie1982 I a...)21:53
T42<Herrie1982> Let me see21:53
T42<11> the first brocade solved the problem. thank you so much (re @SailfishFreenodeIRCBridgeBot: <mal>@11 apply these...)21:53
T42<Herrie1982> A few patches we carry on top of 5.19.1:
mal@Herrie1982 FD_DBG_NOBIN fixed some graphics issues I had when using iommu, not sure if it helps with other issues, you don't have iommu in use, but that env var did fix some glitches on some other adreno gpus so worth a try if you don't have it21:55
T42<11> by the way, there are a lot of such errors [ 13.617498] systemd-cgroups (4275): undefined instruction: pc=b6861d6621:55
T42<11> [ 13.617506] Code: 8000 ef20 0150 4770 (ec51) 0f1e21:55
malnot sure what those are and if those are critical, maybe check if kernel check script has some cgroup related warnings21:56
T42<Herrie1982> I don't recall we have this, let me check (re @SailfishFreenodeIRCBridgeBot: <mal>@Herrie1982 FD_...)21:57
T42<11> Yes, these errors are very common, but not captured, you can stupidly ignore (re @SailfishFreenodeIRCBridgeBot: <mal>not sure what t...)21:58
mal@Herrie1982 or is it FD_MESA_DEBUG=nobin21:59
T42<Herrie1982> This seems to suggest FD_DBG_NOBIN?
ThaodanPersonally I would like to drop support for anything below linux 4.4.x or even 4.14.x22:00
ThaodanThe best solution is to have ports that upgrade kernel versions22:00
mal@Herrie1982 yes, that was it, I think i forgot which name it has, as an example
mal@Herrie1982 it seems usign that FD_MESA_DEBUG=nobin will set FD_DBG_NOBIN22:01
T42<Herrie1982> Ah OK22:02
T42<Herrie1982> We'll try that at our end and see if it gives improvements22:02
malseveral pmos devices use that22:03
T42<Herrie1982> Our most challenging device seems will be Tenderloin which uses APQ8060/MSM8660 which should be pretty OK, but whole DT needs to be build still mostly22:04
T42<Herrie1982> hammerhead and mako should be in pretty OK shape mostly I guess22:05
malaudio will not be fun, I was thinking if I should try, at least fp2 uses wcd932022:06
T42<Herrie1982> Yes probably not :( Was able to at least get a hold of the schematics for Tenderloin, so that should help a bit (re @SailfishFreenodeIRCBridgeBot: <mal>audio will not ...)22:12

Generated by 2.17.1 by Marius Gedminas - find it at!