Sunday, 2026-07-19

*** n2klb_ is now known as n2klb01:56
rinigusmal: thanks05:43
rinigus@b100dian: those local checkouts and clones can be confusing when syncing with repo later. Had that too once in a while05:46
rinigus@b100dian: re snapshots and HADK - is there anything missing in docs? I should have pointed with url, sorry07:04
T42<b100dian> No, the docs are alright, it just that I went to https://hadk.sailfishos.org/ first to check09:40
rinigus@b100dian:  :)10:19
Mister_Magistermal: while you can't help me figure out why backports cause segfault, would you be so kind as to help me with the fm radio microphone redirection?11:34
T42<b100dian> rinigus: seems like all went well. After redoing almost all the steps, I think i got even newer droid packages because I used —force-sync (when having previous repo errors)11:56
rinigus@b100dian: I guess we can release it then. or do you want to look into esim on boot before it? or it will wait till its ready and get updated during 5.1 cycle11:58
T42<b100dian> So, linkerconfig binary is there, and even camera doesnt lock up on sleep (?), quickly checked calls + bt ->  work, I think we're set11:59
T42<b100dian> One thing I dont understand is why do I have telnet before password entry. Because the update flashed new boot image which is clearly not recoverable11:59
T42<b100dian> *not recovery-boot12:00
T42<b100dian> rinigus: do you have /init_disable_telnet file (or something like that)?12:02
T42<b100dian> rinigus: no need to look at esim at boot, maybe for 5.2 - maybe the gbinder-radio that connects to IRadio@1.3 will be already available12:02
rinigus@b100dian: no /init_disable_telnet in mounted / .12:03
rinigusas for telnet, didn't check12:04
T42<b100dian> So maybe you had telnet just because you were on devel then12:04
T42<b100dian> one thing we could do is to include hybris-recovery in the release zip12:05
T42<b100dian> rinigus: did you test GPS?12:06
rinigusnot really. will do and afk for 1h12:07
riniguswill be away for a bit :)12:07
T42<b100dian> no worries. Seems to work, it just started very slow (0/0) in the first minute12:08
T42<b100dian> I will copypac to 5.112:11
rinigusGps works fine :)12:13
T42<adampigg> mtk is very fast at gps signal12:16
malMister_Magister: logs12:49
Mister_Magisterwhich ones would you like12:49
malMister_Magister: and what do you see in pactl list when you try fm radio12:49
malhow about you just have a look what happens in logs when you try to use fm radio and see if anything interesting looking appears12:50
Mister_Magisteroke12:50
Mister_Magisterfor some reason it stopped loading radio_iris so i gotta figure that one now xd12:55
Mister_Magistermal: oh this is interesting https://paste.opensuse.org/pastes/81096db7760f12:57
Mister_Magisteroh12:58
Mister_Magisterit should be input-fm_rx not fm_tuner12:58
Mister_Magisterit's like it ignores xvars12:59
Mister_Magisterin fmradio.conf there's ports = $input_external@equals:$input_equals->$source_input_fmradio13:00
Mister_Magistersource_input_fmradio not droid_source_input_fmradio13:00
Mister_Magistermal: works13:02
Mister_Magisterso it changed from droid_source_input_fmradio to droid_source_input_fmradio13:02
Mister_Magisterto source_input_fmradio*13:02
Mister_Magisternobody told me :D13:02
Mister_Magistermal: and YOU did it! https://github.com/mer-hybris/droid-hal-configs/commit/12968c5aaf00ccfe4e0281ed9693d78b32fa056713:03
Mister_Magistersmh smh13:03
malMister_Magister: so the usual solution, always check fairphone ports :)13:10
malactually not this time13:11
malsince I use the default value on fp213:11
Mister_Magister:)13:24
Mister_Magisterbut you were right once i tried looking into it i fixed it myself13:25
Mister_Magistershouldn't've bothered you at all13:25
Mister_Magisternow plz fix the segfault in bt backports lmao xd13:27
Mister_Magisternaah i'd rather like it if you fixed the gst-droid :)13:27
Mister_Magisterid fix it myself but the decoding and stuff is way beyond my head i don't get what's going on at all13:28
Mister_Magisterand it's simple segfault to debug13:28
malrinigus: https://github.com/mer-hybris/hybris-patches/commits/hybris-21.0/13:38
malrinigus: also pushed the apexd patch I had, slightly different than your patch but does the same thing13:38
Mister_Magistermal: why are the bt backports built outside of kernel tree? Couldn't i incorporate them into the tree?13:39
malmight be possible to include those but didn't bother trying13:41
malI don't really want to spend too much time on old ports, just to keep them running13:41
malso what was the segfault again?13:43
Mister_Magisterhttps://paste.opensuse.org/pastes/c7f0430a311513:44
Mister_MagisterI've checked conntrack code as well as slub code and it's all fine13:45
Mister_Magisterso it's more like memory corruption13:46
malyeah13:53
malwhen does that happend?13:56
T42<elros34> how about patch in hadk-hot for sysfs_remove_dir crash for  <= 3.10 kernels?14:02
malyeah, that seems like the exact place of that crash14:06
T42<elros34> I remember it happens on some older devices after some sfos update so added it so little strange it happens just now14:07
malcan't remember seeing that on fp2 but will add the patch anyway14:09
malMister_Magister: ^14:12
malhttps://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=26ea12dec0c84133add937455be76d44fe253d85 as shown in hadk-hot14:12
Mister_Magistermal: it happens after device boots to gui14:44
Mister_Magister@elros34 w…what patch? Who the hell figured that out14:45
Mister_Magisteralso hadk-hot is not linked in topic on irc14:45
Mister_Magisterneither it is linked on telegram14:45
Mister_Magisterbut thanks to you both14:46
rinigusmal: thank you very much! we would probably switch to them for the next release14:47
Mister_Magistermal: @elros34 bt works!14:55
Mister_Magistereverything's fixed now14:55
Mister_Magistermal: should I remove those acdb symlinks? https://github.com/VerdandiTeam/droid-config-thea/tree/master/sparse/etc14:56
malnot sure14:56
Mister_MagisterI won't then14:56
T42<b100dian> rinigus: packaging, I don't think we need these anymore https://github.com/sailfishos-sony-nagara/main/compare/main...b100dian-patch-115:29
T42<b100dian> However I admit I still use the no-beautiful-soup version of the script so that's why I am making PR so you can check it yourself :)15:29
Mister_Magisterhmm building zip freezes at installing jolla-developer-mode :/15:41
Mister_Magisterahh nothing that good reboot can't fix15:55
T42<elros34> honestly I do not remember who figure out this patch it was long time ago16:03
Mister_Magistersomeone smart16:03
Mister_Magisteri mean i knew the place, but i didn't think that allowing it to just pass agian would be a solution16:03
Mister_Magisterseemed like a hack16:03
T42<b100dian> rinigus: I have drafted a release where I've uploaded the 5iv build + hybris-recovery just in case16:28
T42<b100dian> let me know if you'd prefer to rebuild with mal's 2 new patches that we kindof already have16:29
Mister_Magistermal: which package provides the default images in gallery? my newer ports don't have it16:35
Mister_Magisterpatterns-sailfish-consumer-generic?16:35
Mister_Magisternah tahts not it16:36
malgood question16:36
Mister_Magisterlike i don't see anything additional in patterns yet my g2 port does come with those images, and the newer ports don't16:37
Mister_Magisterwait i can check with --provides16:37
Mister_Magisterhmmm it returns no results xd16:38
Mister_Magisterso maybe I can't16:38
Mister_Magisterwell that's a mystery for another day i got 6 devices to update still16:43
malsailfish-content-gallery-default but those should get copied atr some point16:43
Mister_Magisterseems like it's installed nevermind then16:43
malhttps://github.com/sailfishos/user-managerd/blob/master/src/sailfishusermanager.cpp#L25016:43
maland similar in home encryption code16:44
malbut why it fails for you I'm not sure16:44
Mister_Magisternah i don't think it fails don't worry about it16:45
Mister_Magistervolume buttons work, everything is back to working order16:45
rinigus@b100dian: great job! no, I don't think we need to rebuild with the same/similar patches coming from different  source. let's release as it is.16:47
Mister_Magisternow time to do this whole ordeal AGAIN on my other 3.4 device16:48
Mister_Magisterpainpeko16:48
Mister_Magisteri probably will do the same, steal latest kernel available16:49
Mister_Magisteri'm kinda done with backporting xd16:49
rinigus@b100dian: I think we can just release it and drop Draft. I've added OTA link16:51
T42<b100dian> Absolutely, if you added the 5iv zip go16:52
T42<b100dian> rinigus: about the OTA, is it a precaution to document that check or did you lack some repo params at the beginning?16:53
T42<b100dian> (I'm talking about the https://github.com/sailfishos-sony-nagara/main/wiki/OTA documentation, to be clear for anybody else)16:53
malMister_Magister: do you have any custom things related to user or home?17:06
Mister_Magisterwhat do you mean?17:06
maljust wondering17:06
Mister_Magisterdon't think so17:06
Mister_Magisterdon't worry about the images it probably works fine17:07
malthere is also code to copy that /etc/skel in create-home script17:07
Mister_Magisterdon't worryyy, worry about gst-droid17:07
rinigus@b100dian: yes, its carried over from tama and partially from onyx. so, reflects some older and occasional issues17:09
T42<b100dian> ringus: alright, so better check than sorry :)17:09
rinigusas for ZIP - do we really need it? I may add one later, would like to get qt6 building in chum:testing first17:10
malrinigus: I'll make sure that apexd patch also goes to other android bases where it might be relevant17:10
T42<b100dian> rinigus: Well, technically creating the zip validates a bit more the packages built on 5.1 subproject.17:11
rinigusmal: nice! it makes sense to distribute these patches as much as possible making sfos better on all ports17:11
T42<b100dian> Also, I didn't test the create-images.sh because I'm using a fork of it 😊17:11
rinigus@b100dian: I'll try to build it soonish - will put into TODO17:12
malrinigus: not sure if that affected battery life on fp5 but with one sim in and otherwise idle the device lost charge about 1% per 4 hours17:12
rinigusmal: I usually keep collectd running. that allows me to check suspend stats over a year see whether patches made a major difference or not. but seems like very few are using it (systemdatascope + collectd) :)17:14
malyeah, maybe would make sense to test that, forgot about that17:14
rinigusits best when it runs always on bg. that way data is always collected17:15
Mister_Magisterdroid-hal-init: critical process 'servicemanager' exited 4 times in 4 minutes; rebooting into recovery mode19:00
Mister_Magisteruh-oh19:00
malwhich android base?19:00
Mister_Magister13 but using 17.1 kernel19:01
Mister_Magistereverything is seemingly segfaulting with the 17.1 kernel19:01
Mister_Magisterugh i'll just go and backport the memfd19:02
Mister_Magisterahh 2 out of 3 patches applied without any conflcts glorious19:07
Mister_Magistermal: it would seem that my backporting was successful! its much easier when you're backporting from same kernel xd20:34
Mister_Magisterand even volume buttons fixed itself20:36
Mister_Magistermal: but dns is still broken20:37
Mister_Magisteri have nameserver 127.0.0.53 but it simply doesn't work20:38
Mister_Magisterwhile loading shared libraries: libdns.so.1110: cannot open shared object file: No such file or directory20:38
Mister_Magisteroh20:38
Mister_Magisterah no that's just one tool nevermind20:40
Mister_Magisternobody is listening at 127.0.0.53 whatever that is20:41
Mister_Magisterthere's /etc/appsupport.conf.d/10-systemd-resolved-dns.conf:DNSServers=127.0.0.53 but idk if that's in use20:43
Mister_Magisterwhoop it seems something is still wrong with seccomp because ssh starts but won't connect21:06
Mister_Magisterapparently in 5.1 sfos switched to systemd-resolved but there's no resolvectl (why would there be)21:23
Mister_Magistermaybe it's also caused by seccomp21:38
Mister_Magisterahh browser straight up segfaults, glorious21:49
Mister_Magisterfailed to seal memfd: Invalid argument21:49
Mister_Magisterbut on fresh flash at least ssh works21:50
Mister_Magisterdns still nyet21:50
Mister_Magisteri wonder what am i missing21:53
Mister_Magisterokay so I went back to g2 and opening browser segfaulted kernel and after reboot it just crashes the browser22:07
Mister_Magisterugh22:07
Mister_Magisteri'd notify people that they also need sealing api but nobody but us has 3.4 kernel xd22:14
Mister_Magisterhopefully sealing api is all i need…22:14
malwhat is that?22:21
Mister_Magisterwish i didn't know22:25
malit seems fp2 kernel has some commit "This adds memfd_create with sealing support as required by Perfetto/traced"22:25
Mister_Magisterhttps://github.com/mlehtima/android_kernel_fairphone_fp2/commit/6591f092f90d6f8d71f39331cfc0f376462e3b8522:25
malseems to be several commits22:26
Mister_Magisterneither my new and shiny g2 kernel nor android 18.0 on x2 has that22:26
Mister_Magisterand seems like browser is asking for it22:26
malok22:26
Mister_Magistermal: does browser work on fp2?22:26
malI'll test it again, I think it worked but I'll verify22:27
malbrowser starts without issues22:28
Mister_Magisterdamn you22:28
Mister_Magisteralso your kernel doesn't have f2fs22:29
Mister_Magistergrrr22:29
Mister_Magisteryou got lucky youknow22:31
malyeah, your kernel hasn't had much backporting done compared to fp222:31
Mister_Magisteryeap22:32
Mister_Magistereven with like android 18 there isn't much22:32
malwe'll see how long we can keep this old device working22:34
malto use latest systemd we would need to revert or patch some things22:35
Mister_Magisterthat's the spirit!22:35
Mister_Magisterg2 is the oldest port and i'd love to have it working as long as possible22:36
Mister_Magisterfor x2 i'm in process of porting mainline so i don't really care22:36
Mister_Magisterbut there isn't really mainline support for 822622:36
Mister_Magisterand its too weak anyway22:36
malbut the systemd version I'll update to still has support for for 3.422:36
Mister_Magisternice22:36
Mister_Magisterthank you for trying to keep g2 alive22:36
mal8226 support is not that bad, I have been part of that22:36
Mister_Magisterxd22:37
Mister_Magisterhttps://wiki.postmarketos.org/wiki/Generic_MSM8226_(qcom-msm8226)22:37
Mister_Magisterpmos says something else22:37
Mister_Magisterhttps://wiki.postmarketos.org/wiki/Motorola_Moto_G_2014_(motorola-titan)22:37
malyeah, many things still to be done22:37
Mister_Magisterso it is bad22:38
Mister_Magisterx2's aka 8974 is much better22:38
Mister_Magisterand all i need is wifi touch and audio anyway22:38
malnot sure how much time people want to spend on such old platform22:38
malaudio is the biggest thing22:38
Mister_Magisterwell x2 is the best device to have ever existed so i'll spend as much time as is needed22:39
Mister_Magisteri won't stop using x222:39
Mister_MagisterI even have backup one in perfect condition :P22:39
maldoes it have working volte?22:39
Mister_Magisterno ofc not22:39
Mister_Magisterits way too old22:39
Mister_Magisterop6 doesn't even have working volte22:40
malyeah, that will be a problem at some point22:40
Mister_Magisterno because i'm using it as wifi only device, and mobile data22:40
Mister_Magisterit was never usable as a phone becaues in-call mic was broken22:40
malok22:40
Mister_Magisterg2 stopped being usable as a phone since it doesn't have volte working though so i'm not sure how much longer people will be using it22:40
Mister_Magisterbut i'd love to have it supported as long as possible simply because it's the oldesst22:41
Mister_Magisterat this point it's statement, not a phone22:41
malI wonder how many users still use fp222:41
Mister_Magisterlike if people use g2 in 2026, a 12 years old phone with almost 11 years old sfos port, that's a statement22:41
Mister_Magisterthat means phones don't get slower, and don't need 16GB ram22:41
malyeah, fp2 is still mostly usable, if volte is not needed, browser is of course a problem because many websites are so heavy they use a lot of memory22:42
Mister_Magisterdo you have 1GB of zram?22:42
malprobably not22:43
Mister_Magisteri recommend22:44
Mister_Magisteressentially makes the phone 2GB ram and makes it way more usable22:44
Mister_Magisteryou can use my scripts if you want22:44
malhow much ram do your phones have?22:45
Mister_Magister1GB22:45
malfp2 has 222:45
Mister_Magisterx2 has 2GB22:45
Mister_Magisterwell zram still good22:45
Mister_Magisterspeaking of mainline, since i got op6 ports which exhibit the best mainline today, would you mind helping me out doing mainline port on them + pipa (xiaomi pad 6 also exhibits insane mainline support)22:45
Mister_Magisternot today ofc just wondering if you'd be interested :D22:46
mala lot of mainline stuff on sailfish is not that difficult22:46
Mister_Magisteryeah but i've never done mainline port so idk how do you even go about it22:47
Mister_Magisterit's not like there's HADK for it22:47
malyeah, maybe I should continue mainline stuff at some point, fairphones have quite nice support22:47
malthe new ones22:48
malI should finally do proper mainline adaptation repo setup22:48
Mister_Magistermaybe you can do it while helping me out making op6 mainline port :)))))22:49
malI can at least show what kind of setup I have for mainline fairphone tests22:50
Mister_Magistersure22:50
Mister_Magisterfor now i have to update g2/x2 so that i can get on with updating op6, op6t, pipa, miami22:50
Mister_Magisteri got lot more ports to go22:50
Mister_Magistertho those have WAY newer kernel so there's no issue22:50
Mister_Magisteroh damn its 1am alraedy22:51
malI still need to verify some things on fp2 before releasing 5.1 for fp2, fp4 and fp522:51
maland new 5.1 based fp6 test image is waiting for testing22:51
Mister_MagisterI realeased g2 already and the browser is broken oops22:51
Mister_Magisteri hope the sealing is the fix if not idk what am i gonna do22:52
Mister_Magistermal: and dns doesn't work!22:52
Mister_Magisterit seems like stuff is blocking systemctl-resolved but not sure22:56
malI don't remember the details of the resolved work22:57
malMister_Magister: any of your newer ports have slow boot?22:58
Mister_Magisterhmmm hard to tell really22:59
malwhat is the newest android base you use?22:59
Mister_Magister22 iirc22:59
malhow long is boot time usually?22:59
Mister_Magisternot sure will check tomorrow22:59
malI saw signicant improvement on some newer bases with some fixing23:00
malsome of those issues are generic, some device specific23:00
Mister_Magisteriirc pipa is 21 miami is 22 or smth like that23:00
malone thing to check is if bluebinder ends up getting timeout and if you see messages about vold in log, latter is a generic issue which can cause about 10 second delay and other issues during boot23:04
malbluebinder issue which is device specific can delay up to a minute23:05
Mister_Magisterdang after adding the sealing it just doesn't boot anymore, great23:32

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!