Tuesday, 2019-11-19

T42<meierrom> @adampigg: nice pic. Are batteries in separate box?01:26
T42<然捷 %lastname%> (Photo, 1223x675) https://irc.thaodan.de/.imgstore/3aeEgOFBq3.png "精准币圈菠菜棋牌粉" "非小号/mytoken关注惊爆价"1、币圈电报群/微信群炸群推广 2、twitter、facebook关注 3.KYC身份材料出售  4、电报群/微信群/potato拉币圈活人 5、国内外交易所APP实名注册认证 6、电报群客服活聊天  7、批量注册国内外交易所实名账户,,出twitter、facebook、微信带粉账号04:20
T42<dontatmepls> these spam accounts need to stop06:13
r0kk3rzwelcome to smellogram06:38
deathmisthey, if RTC alarms are working are they meant to power on the device (e.g. to actdead mode) when an alarm has been set in SFOS and the device is turned off or am I understanding them incorrectly?06:42
T42<adampigg> @meierrom [@adampigg: nice pic. Are batteries in separate …], Atm...wont be final though06:44
r0kk3rzdeathmist: yeah iirc06:44
deathmistI may know why they aren't working then: I can't set the rtc device as rw. when I do, the boot process stops before it even starts DSME :/  still get LED tho, could try debugging it if only I could telnet in for once06:46
r0kk3rzwhats the go with this error? Error <creator>[11/19 06:49:21] : Command 'modprobe' is not available.06:50
r0kk3rzrunning mic06:50
rinigusr0kk3rz or mal: would you mind to rename nemo:devel:hw:sony:h8216 -> nemo:devel:hw:sony:tama and make new project nemo:testing:hw:sony:tama07:03
T42<birdzhang> r0kk3rz: i asked yesterday, maybe we should ping coderus07:06
malrinigus: not sure if rename is possible, I might need to create the new one07:07
rinigusmal: no problem, let's make new.07:08
rinigusI will let you know when the old one can be removed07:08
malrinigus: https://build.merproject.org/project/show/nemo:devel:hw:sony:tama use osc copypac to copy the old packages07:09
malI can do it if needed07:10
r0kk3rz@birdzhang its maybe a problem with the docker image?07:10
r0kk3rztbh im surprised it even needs modprobe07:10
rinigusmal: will do. would have to learn copypac anyway. what about testing? I'd mainly need it for adding extra-images there07:11
r0kk3rzosc can be a pita to set up sometimes07:12
malrinigus: https://build.merproject.org/project/show/nemo:testing:hw:sony:tama07:13
rinigusmal: thank you very much!07:13
malr0kk3rz: that is why I have commands ready for osc copypac so I can clone a device easily07:13
r0kk3rzyeah its definitely useful07:14
r0kk3rzbut ive also seen it explode with weird python errors too many times :)07:14
T42<adampigg> use osc from your distro, not the sdk07:16
malI always use it from host07:16
rinigusmal: would be great to become maintainer for nemo:devel:hw:sony:tama. if possible :)07:17
malI thought I added you there07:17
malrinigus: you should have permissions to both devel and testing07:19
rinigusmal: now I do, thanks. copypac worked with a small script07:20
deathmistmal: how should I deal with OBS for OP5/5T? maybe create a oneplus:msm8998-common project with most stuff, then :cheeseburger and :dumpling for droid-* packages perhaps with subprojects pointing to the :msm8998-common? hardware is nearly identical between the models07:20
malrinigus: we should add copy script to faq07:21
rinigusmal: maybe. not mine, though - very quick hack :)07:21
maldeathmist: need to think07:22
rinigusohh, surprise, _pattern package after copypac exploded and became broken. ... somebody was telling me _pattern is not really up for it.07:24
rinigusmal: how do I exclude _pattern from building? does it need to be re-patternised?07:25
rinigustama project07:25
malrinigus: remove it, then ask lbt to add patterns for those, and to remove it from the old project07:26
malso we can remove the old project completely07:27
rinigusmal: will do07:28
T42<adampigg> mal, bluetooth-rfkill-event is built from master, so that can be a common package....only need to pr broadcom-bluetooth07:30
mal@adampigg wure, I'll add those at the same time07:31
T42<adampigg> mal, for broadcom.... i should create broadcom-bluetooth-bluez5.spec, and add Provides: line?07:37
T42<adampigg> mal, any easy way to use the new gcc? does it have many deps if i add it toa project?07:43
r0kk3rz@birdzhang so far so good with the docker07:45
T42<birdzhang> r0kk3rz: i see, will rollback to docker image too07:46
r0kk3rzit only runs mic, so it should be ok07:47
T42<birdzhang> yeah07:47
T42<birdzhang> building now ;)07:51
rinigusany idea why initrd-logos is missing in testing and available in devel? at obs. as a result, https://build.merproject.org/package/show/nemo:testing:hw:sony:tama/droid-hal-img-boot-sony-tama-pie is unresolvable08:17
riniguslbt: please patternise nemo:testing:hw:sony:tama and nemo:devel:hw:sony:tama. also, please remove internal patterns for nemo:devel:hw:sony:h8216. we are going to delete this project08:19
rinigusfor now, I copied nemo:devel:hw:common/initrd-logos-sailfish to tama testing. would be great to get that package in nemo:testing:hw:common08:33
mal@adampigg yes, add broadcom-bluetooth-bluez5.spec, you can probably add a Provides there also if we want to have those use the same thing in patterns08:41
rinigusI have been able to get oom killer on xz2 using the following sequence: https://github.com/sailfishos-sony-tama/main/issues/3011:41
riniguson onyx all is fine.11:41
rinigusI wonder whether its reproducible on other newer devices or just my xz2-specific headache11:42
Mister_Magistercamera always used shitton of ram11:56
abransonthat hdr scene mode's hilarious for testing oom. so many buffers.11:57
rinigusdoes it mean you see the same?11:59
rinigusactually, in dmesg I see something like https://dpaste.de/Dbem . from searching online, it seems to be android-feature. should we disable that beauty from our builds?12:10
Mister_Magisterrinigus: lmk is standard part of linux kernel why would you even disable that12:11
Mister_Magisterthat would be madness. If you run out of ram your phone would not get out of it and just freeze12:12
rinigusMister_Magister: that was a question, possible with the wrong assumption. assumption was that its a part of android builds, but not regular linux. from your reply, I deduce that this assumption is, probably, wrong12:13
Mister_Magisterrinigus:  if you enable sysrq on your desktop, hit alt+sysrq+f and you force trigger lmk. you will see same message in dmesg12:14
Mister_Magisterwell similar at least12:15
Mister_Magisterhttps://paste.opensuse.org/59391536 linux can sound scary12:17
rinigusMister_Magister: I am not going to try - as its gonna disturb too much right now. However, we have CONFIG_ANDROID_LOW_MEMORY_KILLER=y and corresponding driver seems to be in drivers/staging/android/lowmemorykiller.c12:18
rinigusnot very desktopy to me12:18
Mister_Magisterye well android can have bit different memory killer that's fair point12:18
Mister_Magisterbut it's still needed12:18
Mister_Magisteror you could use standard memory killer but what's the difference12:19
Mister_Magisteryou setup android one in init files i think12:19
rinigusMister_Magister: I wonder whether my oom with camera is induced by trigger-happy lowmemorykiller. but let's see if others see the same killer at their devices.12:23
Mister_Magisterrinigus: camera eats lots of ram it often kills my apps. you can monitor ram usage in background and see that it just ran out of memory12:28
Mister_Magisterit's like… normal thing to do12:28
abransonyeah there's a special android lmk12:30
riniguswell, its a fast process and will not occur if you do it differently. its just I haven't seen it on onyx and couldn't reproduce it on it. at the same time, onyx has 1gb ram less.12:31
Mister_Magisterabranson: but it should do roughly same job right12:31
rinigus... and in onyx, its commented out12:34
abransonMister_Magister: they score things differently iirc. the android one is quite specific to that process model13:04
Mister_Magisterlet's disable it what worse could happen13:06
Mister_Magisteryeah android one probably chooses between java programs than native ones13:06
abransoni seem to remember it uses thousands for different classes of android process13:19
rinigusits briefly described in https://elinux.org/Android_Kernel_Features under oom handling13:22
Mister_Magisterrinigus: you can try disabling it and switching to linux one if it changes anything13:22
rinigusabranson: android lmk is something additional, as far as I can see https://source.android.com/devices/tech/perf/lmkd13:22
rinigusMister_Magister: yes, I will try13:23
Mister_Magisterdisabling different things is fun ^^13:23
abransonrinigus: I think that daemon communicates with the kernel android lmk13:25
abransonmanages the aggressiveness of the killing depending on how short memory is13:26
rinigusabranson: according to "In Android 9 and higher, userspace lmkd activates if an in-kernel lowmemorykiller driver is not detected", I'd expect they are exclusive13:26
abransonoh really? guess i can't remember then :)13:27
T42<Lukapanio> simStatusChangedInd: radioService[0]->mRadioIndication == NULL13:46
T42<Lukapanio> it seems to be mainproblem with modem13:49
deathmist@Lukapanio your modem or what? mine has that spammed plenty of times, but calls and SMS work, no cellular data tho so it's probably related to that(?)13:51
deathmistyou're just likely missing configs if your modem doesn't work, do you even have UI yet?14:03
deathmistor "ofono" to be precise, not modem, that should mostly work OOB14:04
T42<Lukapanio> yes i have gui16:07
rinigusfun fact: disabling android lowmemorykiller resulted in no oom induced by camera even in the presence of browser and osm scout server + pure maps. and there are no messages regarding low ram either.16:09
rinigusabranson Mister_Magister ^ will monitor whether it has some other side effects16:10
abransonrinigus: make sure you check dmesg to see if the linux oom killer's doing anything16:13
abransonif it's even there...16:13
deathmist@Lukapanio add ofono configs from https://git.io/JeKHH to your device (and possibly modify if needed), afterwards remember to put them in your droid-config repo too16:15
rinigusabranson: I did check, that seems to be clean16:16
rinigusI presume it is there... let me think how to test it16:17
Mister_Magisterrinigus: try to kick looinx oom reaper16:31
Mister_Magistermy moto g2 has # CONFIG_ANDROID_LOW_MEMORY_KILLER is not set16:34
Mister_Magisterbut that's older one let's try newest one16:34
Mister_Magisternewest snapdragon 845 has CONFIG_ANDROID_LOW_MEMORY_KILLER=y16:35
rinigusMister_Magister and abranson : with stress test I managed to reboot phone. not good, will have to see how to get oom killer working16:36
Mister_Magisterthat's what i suspected16:36
abransonmaybe we have to use the android one?16:42
Mister_Magisterfor new devices possibly16:44
rinigusI presume that with android one configured, the classical one should be disabled mostly. well, for now, I can enable it back and postpone hunting for how to setup the classical one.16:45
rinigusand deal with it if it does become an issue in more common workloads16:45
T42<Akatsu %lastname%> Do anyone have idea how wcnss initialization works? I have broken wifi so using logcat i made almost all what it was asking me for but i stopped on wcnss not being able to wcnss_write_cal_data.17:48
riniguslbt: please patternise nemo:testing:hw:sony:tama and nemo:devel:hw:sony:tama. also, please remove internal patterns for nemo:devel:hw:sony:h8216. we are going to delete this project18:21
T42<adampigg> rinigus: save the wait, dont paternize, use meta ;)18:23
rinigusadampigg: I don't mind meta, but then we have to do it properly. namely, generated KS should have already %packages replaced with meta.18:24
rinigusso far, I have found in one of droid-config scripts a call %gen_ks.18:24
rinigusany idea where is that one?18:25
rinigus... defined18:25
piggzmal: https://github.com/mer-hybris/broadcom-bluetooth/pull/219:25
riniguspiggz: me neither.19:32
rinigusmaybe digging through build logs will help. I'll check19:33
piggzrinigus: i think its maybe done in droid-config.inc?19:33
riniguspiggz: grepping it didn't help. but that's where %gen_ks is19:34
riniguspiggz: I wonder, if its this bit - https://git.merproject.org/mer-core/ssu/blob/master/ssuks/ssukickstarter.cpp#L15219:46
piggzrinigus: geez, its a compiled program???19:47
riniguspiggz: yes, called ssuks19:48
piggzrinigus: im sure some sed magic will fix up the file :D19:48
riniguspiggz: yes, that's true. and that sed magick can be added to spec19:49
vknechtgood question, would have expected to find this rpm macro defined in one of https://git.merproject.org/mer-core?utf8=%E2%9C%93&filter=rpm but could find it from a quick look19:55
vknechtbut there must be something somewhere :19:55
vknechtPlatformSDK [vince@dozer ~]$ strings /var/lib/rpm/Packages | grep gen_ks19:55
vknecht- [ks] Add hacky %gen_ks macro for easy kickstart packaging19:55
TheKitdid latest Sailfish update break sailfishos-porters-ci? Error <creator>[11/19 20:23:26] : Command 'modprobe' is not available.20:25
TheKit80 ERROR: Job failed: exit code 120:25
deathmistTheKit: it was discussed earlier today, seems docker image works fine http://www.merproject.org/logs/%23sailfishos-porters/%23sailfishos-porters.2019-11-19.log.html#t2019-11-19T06:50:4920:28
TheKitthank you20:28
deathmistI have an interesting situation on my phone, got both /vendor/firmware/ and a mount /vendor/firmware_mnt/ with the ladder containing an image/ dir, but e.g. only the former having "tfa98xx.cnt" in it, what should I do? I still have some FW issues on 16.020:31
vknechtdeathmist, about rtc & alarm wakeup, maybe you need something like that https://github.com/vknecht/android_kernel_alcatel_msm8916/commit/22eba3aaf141c1b3e868cdfe14074c3cc6133fb820:32
vknechtiirc that's what made powering up device from a sfos alarm work20:33
piggzdeathmist: sounds similar to pro1 mounts20:33
vknecht*work for me20:39
vknechtand probably from los too ; don't be too eager, need like 3 or 4 minutes delay at least iirc20:44
deathmistvknecht: nvm I just re-tested and it seems to work (just boots normally tho, not to actdead but that is perfectly fine imo) :D  that's cool. it just spams my logs like crazy tho in it's current state. primary cause for it seems to be https://git.io/JeKAP20:58
deathmistand a sample of the spam https://pastebin.com/21FvWevT20:58
vknechtactdead is only for battery charging when device is powered off isn't it ? I'd expect a full OS load on alarm wakeup, don't you ? dunno about log spam, should check it here21:02
vknechtgood for you that it works without dts/kernel fiddling, tho ;-)21:03
vknechtrtc write seems like a whole different beast, iirc it crashed when enabled ; I'd have like not to have ~40 years uptime displayed, but let it slide21:07
mcencorahi. now that we know that jolla does not care about the community, is anybody here working on Android 8/9 for Xperia X?22:19
vknechtmcencora, dunno about real Jolla intents, but there's effort about Loire platform toward aosp8 rebasing at https://build.merproject.org/project/show/home:vknecht:hw:qualcomm:f512122:34
vknechtand I'm about to switch to aosp9 base22:34
vknechtafaik, Jolla stance is http://www.merproject.org/logs/%23mer-meeting/%23mer-meeting.2019-11-14.log.html#t2019-11-14T10:47:1022:35
mcencoravknecht: jolla stance was shared today on blog - they won't do aosp8 rebase for xperia x22:35
vknechtoh ?22:36
vknechtdidnt' follow the blog lately22:36
mcencoravknecht: if I have paid Sailfish license, will I be able to install android app support with your port?22:36
vknechtwell, best I can propose is Anbox support, can't say anything about paid-for-dalvik support22:38
vknecht(must say didn't even try Anbox on aosp8 base, lack of time)22:39
mcencorais there any difference w.r.t. to how jolla provides android app support vs anbox?22:40
vknechtthere's probably tons of differences, esp. good knowing of system guts (eg. media/audio routing so it's available to android apps)22:43
vknechtmaybe it's not so much, when you know andbox/lxd innards, but still...22:44
mcencorado you expect to get the bluetooth/usb and higher battery consumption to be fixed with your rebase on top of aosp9?22:46
vknechthmm, just read the blog post and it sounds like a defeat22:48
vknechtwell, about aosp9, I just recently published Loire AOSP 9 builds as baseline for a rebase ; and from what I've seen, it's more on the good side than bad22:50
mcencorayeah, I am really pissed. It basically means that Jolla isn't going to fix Sailfish for Xperia X to finally make it non-beta quality22:51
mcencorabut that seems to be the case for all Xperia's22:51
mcencorathey prefer to offer more beta ports for  another device, then to fix bugs/finish features for existings ports22:52
vknecht*I must say I'm a bit surprised, I'd have expected Jolla could up Alien Dalvik support for X (Loire) a0t least to XA2 (Nile)22:54
vknecht*at least par to XA2 (Nile)22:54
mcencorahmm, your aosp9 port looks interesting. i will most probably try it out22:55
mcencoraone question, do I need to backup my home folder before flashing?22:56
mcencorai.e. will /home/nemo stay untouched?22:56
mcencoraor I need to backup and restory it manually22:56
vknechtit's not started yet, will have to integrate all lessons learned from tama port22:57
vknechtkeep an eye on https://github.com/sailfish-on-loire ;-)22:58
mcencoraaaah, I thought these images are for sailfish with aosp9 base, but this is just plain aosp923:00
mcencoraok, I will23:01
mcencorayou are my only hope vknecht23:01
mcencoraif this port fails, I am going to try ubuntu touch, or switch back to android/lineage os23:01
mcencoraso I keep fingers crossed for your work23:02
vknechtafaik there's no aosp9 image ; aosp8-based ones from my obs still suffer from high battery consuption and no BT23:02
vknecht*aosp9-based image23:03
mcencorado you know the status of upstream Xperia X support in linux kernel?23:05
vknechtso yeah, some patience is needed and we'll see ; next higher hope is mainstream support like https://github.com/kholk/kernel-upstream23:06
vknechtmcencora, I think I'll start on Loire aosp9 base this week-end, maybe make a request to join  sailfish-on-loire GH23:11
vknecht@eugenio_g7 ^23:17
mcencoravknecht: hmm, I think admin must invite me23:22
mcencoraI don't see a way to make request to join23:22
T42<eugenio_g7> Cool stuff, I should finally sync the aosp9 tree :D23:39
T42<eugenio_g7> mcencora: if you remember me tomorrow morning I'll add you to the organization :)23:39

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