| OrokuSaki | cool | 00:10 |
|---|---|---|
| OrokuSaki | Mal: sensors are now working! except the axis is off.. should I edit primaryuse.conf? Or.. change the sensors in the kernel? | 01:19 |
| OrokuSaki | yay! "0,1,0,-1,0,0,0,0,1" for accelerometer | 01:25 |
| *** ubuntu is now known as Guest97467 | 01:30 | |
| OrokuSaki | mk-cam-conf 0 /etc/gst-droid/gstdroidcamsrc-0.conf gets Failed to start pipeline.. logcat says E/DroidMediaCamera( 4019): Failed to connect to camera service | 01:55 |
| OrokuSaki | Which... I have to manually start minisf and minimedia in a service I have with systemd.. which.. is odd.. codecs work.. but not the camera.. not sure what to do | 01:56 |
| OrokuSaki | Oh a couple of weeks ago I asked about a weird image.. https://ibb.co/naYTAK looked like 4 surfaces.. or a quad image.. it was the kernel. 565 framebuffer colors... | 03:33 |
| OrokuSaki | Oh and anyone trying to install their roms with TWRP... did tar work for you? It didn't me.. had to use busybox binary and use tar | 04:11 |
| OrokuSaki | Can't WAIT to see sfos 3 | 04:24 |
| OrokuSaki | Especially since my device has native framebuffer. Still wish there was a way to run android apps, even if you had to pay for it | 04:41 |
| OrokuSaki | I mean landscape framebuffer... bourbon kickin in. =) | 04:42 |
| OrokuSaki | sledges: where have you been dude? | 04:43 |
| *** ChanServ sets mode: +v T4 | 05:56 | |
| T4 | l0cal0st was added by: l0cal0st | 06:38 |
| mgrover | is there any reason why adding bootmode=debug wouldn't stop the bootloader at first telent | 07:29 |
| Mister_Magister | mgrover: why should it (also kernel is not bootloader) | 07:32 |
| mgrover | Mister_Magister, because https://github.com/mer-hybris/hybris-boot/blob/master/init-script#L362 | 07:34 |
| mgrover | if ive understood the script correctly | 07:34 |
| Mister_Magister | mgrover: here is what you want https://github.com/mer-hybris/hybris-boot/blob/master/init-script#L407 | 07:35 |
| mgrover | Mister_Magister, I wanna stop at first telnet on port 23 before rootfs is mounted | 07:35 |
| mgrover | i can achieve the same just by disabling one of the kernel flags | 07:36 |
| mgrover | but am curious why the kernel arg doesnt work because it should | 07:36 |
| Mister_Magister | why not this then https://github.com/mer-hybris/hybris-boot/blob/master/init-script#L364 | 07:36 |
| mgrover | Mister_Magister, can I just apply that in recovery? | 07:37 |
| Mister_Magister | who knows | 07:37 |
| Mister_Magister | guess | 07:37 |
| mgrover | well it seems that sailfishos will get bound to /target so lets hope so | 07:37 |
| mgrover | need to wait for this puppy to charge and ill give it a test | 07:38 |
| * Mister_Magister flashes roms with 5% battery | 07:39 | |
| mgrover | brave man :P but this device gets stuck in a bootloop if i dont make sure it has atleast some charge | 07:49 |
| Mister_Magister | mgrover: https://i0.kym-cdn.com/entries/icons/facebook/000/022/017/thumb.jpg | 07:54 |
| *** ambernight is now known as qtwyeuritoiy | 08:06 | |
| guhl | Mister_Magister, as I have to analyze the boot process of the chiron again (it does not seam to use configfs while it should, ...) I documented the way I am using pstore for this here https://thinkthinkdo.com/trac/project1/wiki/SF_Ch_pstore . Maybe you can have a short look and comment. | 08:52 |
| * Mister_Magister just destroyed somebody on reddit with sailfishos thats excited that he chrooted to debian from android which i did years ago | 08:55 | |
| Mister_Magister | guhl: should be in /sys/fs/pstore/pmsg-ramoops-X not console-ramoops? | 08:56 |
| mgrover | guhl, is that why it reboots? | 08:57 |
| mgrover | guhl, also mgrover == ghosalmartin :P | 08:57 |
| Mister_Magister | mgrover: oh really? | 08:57 |
| mgrover | yeahh | 08:58 |
| Mister_Magister | why different nickname then | 08:58 |
| guhl | mgrover, well that's not fair | 08:58 |
| mgrover | my machine switches network while connected, so hexchat instantly signs out my main nic | 08:58 |
| Mister_Magister | guhl: and except that everything is fine untill you dont shutdown it | 08:58 |
| guhl | Mister_Magister, where "should be in /sys/fs/pstore/pmsg-ramoops-X not console-ramoops?" | 08:59 |
| Mister_Magister | guhl: ? | 08:59 |
| Mister_Magister | like usually in 3.10 kernel i have it in /sys/fs/pstore/console-ramoops | 09:00 |
| guhl | where in my documentation? | 09:00 |
| Mister_Magister | documentation is tl;dr im just making sure | 09:00 |
| mgrover | guhl, did you get to flash zip? | 09:00 |
| guhl | mgrover, no | 09:01 |
| Mister_Magister | guhl: whatcha mean | 09:01 |
| guhl | the recovery can not flash the zip because tar runs out of memory when extracting the root fs | 09:01 |
| Mister_Magister | guhl: i mean i have it in console-ramoops is it different for you? | 09:02 |
| guhl | so i manually installed the rootfs while booted to LOS | 09:02 |
| Mister_Magister | guhl: ? | 09:03 |
| guhl | Mister_Magister, the console-ramoops for me contains the kernel messages including the panic | 09:03 |
| Mister_Magister | guhl: isn't that what you want? | 09:04 |
| guhl | what i want is the init.log from the mer boot that i cat to /dev/pmsg0 | 09:04 |
| guhl | and that will than be in pmsg-ramoops | 09:05 |
| Mister_Magister | dunno y u want this but ok | 09:05 |
| Mister_Magister | then whats the problem | 09:05 |
| guhl | nothing | 09:05 |
| guhl | i want this because i want to be able to analyze the init.log if it is not written to the filesystem yet | 09:06 |
| guhl | or how would you analyze the boot process if you can not telnet into it (because usb setup fails) and it does not reach DONE_SWITCH | 09:08 |
| mgrover | guhl, why dont you do it mnaually? | 09:08 |
| mgrover | flash hybris-boot | 09:08 |
| mgrover | its how I flashed your realllyyy old cm14 zip | 09:08 |
| mgrover | then boot in and extract it | 09:08 |
| Mister_Magister | guhl: then what did you want from me | 09:10 |
| mgrover | guhl, also pstore works fine atm? | 09:11 |
| guhl | peer review? | 09:11 |
| *** mgrover is now known as ghosalmartin | 09:11 | |
| Mister_Magister | guhl: ??? | 09:12 |
| guhl | Mister_Magister, I just wanted you to read it and comment if there is something wrong, what you did. Thanks! | 09:14 |
| Mister_Magister | just dont shutdown | 09:15 |
| Mister_Magister | and it should be ok i think | 09:15 |
| Mister_Magister | (i usually want kernel log lol) | 09:16 |
| Thaodan | guhl: sounds interresting, maybe you could add such thing to hybris-boot | 09:16 |
| guhl | Well Thaodan I am using this in hybris-boot to debug it but I would not think it makes sense in the standard hybris-boot | 09:18 |
| ghosalmartin | pstore is in hybris boot if the kernel is configured? | 09:18 |
| guhl | ghosalmartin, yes | 09:18 |
| Thaodan | yes, maybe hybris-initrd too if lvm is used. | 09:19 |
| ghosalmartin | guhl, did you get a chance extract zip? or twrp a blocking issue atm? | 09:19 |
| guhl | ghosalmartin, i do have the rootfs intalled in /data/.stowaways/sailfishos | 09:20 |
| ghosalmartin | guhl, does kernel panic? or can you get to telnet 2323? | 09:21 |
| guhl | did it manually using the tar.bz2 while booted to LOS | 09:21 |
| guhl | ghosalmartin, no the kernel does not panic boot i also can not telnet | 09:22 |
| guhl | usb network interface is up but does not get an ip. probably a dhcpd issue | 09:23 |
| Thaodan | did you booted by fastboot or by flashing? | 09:23 |
| guhl | fastboot | 09:23 |
| Thaodan | hm. you can to boot hybris-recovery | 09:23 |
| ghosalmartin | guhl, as a test can you rename the sailfishos folder to sailfishosbak | 09:23 |
| Thaodan | that boots straight into recovery | 09:23 |
| Thaodan | sledges: can you make a repo for my port? sony-h3213? | 09:24 |
| guhl | ghosalmartin, well that was the state before, then it will enter run_debug_session and i can telnet into it | 09:25 |
| Mister_Magister | Thaodan: i would better ask lbt | 09:26 |
| Thaodan | in irc? | 09:27 |
| Mister_Magister | yeah | 09:27 |
| guhl | ghosalmartin, this is the init.log https://pastebin.com/KW4pJWz0 from a boot where it can not boot into real rootfs | 09:32 |
| guhl | and i don't like the | 09:32 |
| guhl | + write /sys/class/android_usb/android0/enable 0 ... lines | 09:32 |
| Mister_Magister | guhl: but you need to disable android usb dont ya | 09:33 |
| Mister_Magister | sfos should handle usb | 09:33 |
| guhl | because it should not write to that if using configfs | 09:33 |
| Mister_Magister | guhl: DBG_REASON=Refusing to boot. See /diagnosis.log (in initrd only) | 09:34 |
| Mister_Magister | did ya check that file? | 09:34 |
| guhl | Mister_Magister, thanks | 09:36 |
| guhl | yes it contains CONFIG_VT=y not found in /proc/config.gz | 09:36 |
| guhl | which is one of our problems with the kernel atm | 09:37 |
| Mister_Magister | fix et? | 09:37 |
| guhl | nah i don't like CONFIG_VT anyway. should have been removed from the kernel years ago | 09:38 |
| Mister_Magister | but it wont boot otherwise | 09:38 |
| guhl | i'll remove the check :-) | 09:38 |
| Mister_Magister | why cant you just enable it | 09:38 |
| guhl | because the kernel will not boot then | 09:39 |
| Mister_Magister | why | 09:39 |
| T4 | minlexx was removed by: minlexx | 09:39 |
| guhl | but maybe i can get a pstore output from that | 09:39 |
| Mister_Magister | wdym | 09:40 |
| Mister_Magister | it panics when you have it enabled? | 09:40 |
| guhl | if we enable that in kernel config it immediately reboots | 09:40 |
| Mister_Magister | aka panics | 09:41 |
| guhl | yes | 09:41 |
| ghosalmartin | guhl, yeah I cant get to that state as well | 09:41 |
| ghosalmartin | well doesnt panic | 09:42 |
| ghosalmartin | the kernel is corrupt, doesnt even call init | 09:42 |
| ghosalmartin | guhl, after that for me it reboots | 09:42 |
| ghosalmartin | ive gotta go, be back on 20-3mins | 09:42 |
| Mister_Magister | well dunno if its needed but kernel ceck has it as needed | 09:43 |
| ghosalmartin | 30mins | 09:43 |
| guhl | i'll give it another try to enable it and find something. If that does not work i'll disable the check in the init-script | 09:44 |
| Mister_Magister | well i had huawei devs that forwarded 32bit calls to 64bit function and motoorola devs having flag hat changes code to broken code so… | 09:44 |
| Mister_Magister | guhl: kernels are magic | 09:48 |
| Thaodan | Mister_Magister: that's to enable usb gadget mode and enable rndis afterwards | 09:51 |
| Mister_Magister | Thaodan: VT? | 09:53 |
| T4 | GiaSen was removed by: GiaSen | 09:54 |
| ghosalmartin | guhl, i dont think config_vt mattes, because my pixel xl could boot to rootfs without it | 10:34 |
| guhl | well enabling VT panics (or reboots however) without any message anywhere | 10:35 |
| Mister_Magister | guhl: then try throwing that flag away it can be worse from there | 10:35 |
| ghosalmartin | guhl, i mean its not required to boot to rootfs, am curious why your device doesnt reboot | 10:36 |
| ghosalmartin | if you can access telnet at port 23 then telnet at 2323 should work fine? | 10:36 |
| T4 | akaWolf was added by: akaWolf | 10:44 |
| guhl | ghosalmartin, i think that for CONFIG_VT we need to set the kernel command line parameter console | 10:52 |
| ghosalmartin | isnt it already set? | 10:53 |
| guhl | nope, if you check the init log i pasted | 10:53 |
| guhl | https://pastebin.com/KW4pJWz0 | 10:53 |
| guhl | only androidboot.console=ttyMSM0 is set | 10:54 |
| ghosalmartin | guhl, this? console=ttyHSL0,115200,n8 | 10:54 |
| guhl | in my old build i had something like this yes | 10:54 |
| ghosalmartin | i think i remember adding it with no luck | 10:54 |
| ghosalmartin | but ill give it another shot | 10:54 |
| guhl | i had this in the old BoardConfigCommon.mk | 10:55 |
| Mister_Magister | guhl: did u try disabling VT or what are you doing now? | 10:55 |
| ghosalmartin | Mister_Magister, it is disabled atm | 10:55 |
| Mister_Magister | k and? | 10:55 |
| Mister_Magister | whats happenin | 10:55 |
| guhl | BOARD_KERNEL_CMDLINE := console=ttyMSM0,115200,n8 androidboot.console=ttyMSM0 earlycon=msm_serial_dm,0xc1b0000 androidboot.hardware=qcom user_debug=31 msm_rtb.filter=0x237 ehci-hcd.park=3 lpm_levels.sleep_disabled=1 sched_enable_hmp=1 sched_enable_power_aware=1 service_locator.enable=1 swiotlb=2048 androidboot.configfs=true androidboot.usbcontroller=a800000.dwc3 androidboot.selinux=permissive | 10:55 |
| ghosalmartin | cant get into rootfs | 10:55 |
| Mister_Magister | why | 10:56 |
| ghosalmartin | but i also think its a red herring, since my other device marlin gets to rootfs fine with it disabled | 10:56 |
| guhl | starting all over? | 10:56 |
| ghosalmartin | but it doesnt use configfs | 10:56 |
| guhl | ghosalmartin, and either we set androidboot.configfs=true again (and check for this in init-script) | 10:57 |
| guhl | or disable usb_gadget | 10:57 |
| ghosalmartin | guhl, thing is, if we can boot to telnet 23, doesnt that prove telnet is working? | 10:57 |
| ghosalmartin | what does sfos do differently when it goes to rootfs? | 10:57 |
| guhl | ghosalmartin, it is a difference if it works in run_debug_session or in real boot | 10:58 |
| Mister_Magister | guhl: i mean you have seen it? mount: mounting /data//.stowaways/sailfishos on /target failed: No such file or directory | 10:58 |
| ghosalmartin | guhl, actually do you have this <guhl> ghosalmartin, it is a difference if it works in run_debug_session or in real boot | 10:58 |
| ghosalmartin | not that :P | 10:58 |
| ghosalmartin | this https://github.com/mer-hybris/droid-config-sony-nile/commit/a4152781498a83cada0241a4fdb9016d7ddf562a | 10:59 |
| Mister_Magister | oh it mounts later | 10:59 |
| guhl | Mister_Magister, i renamed it on purpose so that it stays in debug session | 10:59 |
| ghosalmartin | Mister_Magister, that because he renamed it to saifishosbak to force it to wait at that stage | 10:59 |
| guhl | (that was just a test that ghosalmartin wanted me to do) | 10:59 |
| ghosalmartin | thats the only way i can get to telnet | 10:59 |
| guhl | i told him what will happen but tested it anyway | 11:00 |
| Mister_Magister | i dont get what you are both doing | 11:00 |
| ghosalmartin | my issue is after I continue, mine reboots | 11:00 |
| ghosalmartin | mal: didnt you have a similar issue that phone wasnt rebooting, but link was up? | 11:00 |
| guhl | i will fist of all fix the configfs_usb_setup vs android_usb_setup issue | 11:01 |
| ghosalmartin | what issue is that? | 11:01 |
| guhl | see my log from line 304 it does usb_setup_configfs | 11:02 |
| guhl | but from line 416 it does usb_setup_android_usb | 11:03 |
| ghosalmartin | guhl, ahhhh. hmmm weird, lemme get ya mine to compare | 11:03 |
| guhl | and i think this is because we have CONFIG_USB_GADGET=y | 11:04 |
| guhl | and in the current init-script the test is | 11:05 |
| guhl | usb_setup() { | 11:05 |
| guhl | if [ -d $ANDROID_USB ]; then | 11:05 |
| guhl | usb_setup_android_usb $1 | 11:05 |
| guhl | elif [ -d $GADGET_DIR ]; then | 11:05 |
| guhl | usb_setup_configfs $1 | 11:05 |
| guhl | fi | 11:05 |
| guhl | } | 11:05 |
| Mister_Magister | guhl: never do that again | 11:05 |
| guhl | sorry | 11:05 |
| Mister_Magister | :) | 11:05 |
| Mister_Magister | 5 lines is always max :P | 11:06 |
| ghosalmartin | guhl, this is my init.log https://pastebin.com/gFRQgPcJ | 11:06 |
| ghosalmartin | guessing same issue | 11:06 |
| guhl | OK then i will paste it 5 lines a time | 11:07 |
| guhl | you have the same /sys/class/android_usb/android0 stuff there | 11:10 |
| Mister_Magister | guhl: did you compare with working phone? | 11:11 |
| Mister_Magister | like… how it should be? | 11:11 |
| guhl | in my version (for LOS 14.1) i did the test based on | 11:11 |
| guhl | if [ ! "$(get_opt androidboot.configfs)" = "true" ]; then | 11:11 |
| guhl | Mister_Magister, where would I have a working phone | 11:12 |
| Mister_Magister | u don't have any sfos working? | 11:12 |
| guhl | with a 4.4 kernel and LOS 15.1 based | 11:12 |
| Mister_Magister | but somebody else can have | 11:13 |
| guhl | so I added some cmd_line paramters back and changed the test in init-script lets see what happens, ... | 11:17 |
| Mister_Magister | guhl: i have even bigger problem lol | 11:22 |
| ghosalmartin | guhl, any luck with the diff cmd_line params? | 11:34 |
| guhl | not yet | 11:35 |
| Thaodan | lbt: can you add a repo for my port sony discovery on mer buildservice? | 11:43 |
| mal | Thaodan: do you really want it names discovery since discovery can be two different devices | 11:47 |
| mal | *named | 11:47 |
| Thaodan | The Codename is now as used in the too, so yes. But better with vendor as a prefix in front | 11:51 |
| sledges | wiktorek140: pong | 11:51 |
| sledges | Thaodan: please first see through the rename we're doing: https://github.com/mer-hybris/droid-hal-sony-nile/pull/3 | 11:51 |
| sledges | and rename accodringly | 11:51 |
| sledges | also your droid-config PR will need some corrections, please stand by | 11:54 |
| sledges | all Sailfish X devices are built according to https://sailfishos.org/wiki/Sailfish_X_Build_and_Flash (not only HADK) | 11:55 |
| mal | sledges: we probably want to make a common .inc file for the different variants | 11:56 |
| sledges | mal: we'll push h3113 out first, cc mkosola | 11:56 |
| mal | ok | 11:57 |
| Thaodan | sledges: yes ill change accoraccording to that | 11:57 |
| Mister_Magister | oh hi sledges very long time no see | 11:58 |
| Thaodan | please comment on my pr about whats wrong | 11:58 |
| sledges | ohai | 11:59 |
| sledges | Thaodan: we'll comment on your PR after pushing h3113 out first | 11:59 |
| sledges | currently it will mess up the ssu | 11:59 |
| mal | the repo will probably look quite like the f5121 repo | 11:59 |
| sledges | the repo on mer obs will be qualcomm:sony-nile | 12:00 |
| Thaodan | about the sailfish x built, i built it accoraccording to it but had to use aosp for /system | 12:00 |
| sledges | you can already see that in https://github.com/mer-hybris/droid-config-sony-nile/tree/master/sparse/usr/share/ssu/board-mappings.d | 12:00 |
| mal | sledges: really? how does that work for non-xa2 devices | 12:00 |
| sledges | mal: non sony-nile family will have to have separate repo | 12:01 |
| mal | sledges: are all nile really so close that they can use the same android side? | 12:01 |
| sledges | still better than hardcoding to f5121 only | 12:01 |
| mal | sledges: i.e. are pioneer and discovery devices close enough | 12:02 |
| sledges | Thaodan: sailfish x built uses aosp | 12:02 |
| sledges | mal: only way to find ouy :P[D[D[Dt | 12:02 |
| sledges | mal: only way to find out :P | 12:02 |
| sledges | you have device repos checked out for all respective devices in one tree | 12:03 |
| * Mister_Magister sfosx instructions are unclear for non official devices imo | 12:03 | |
| sledges | kernel config will be selected correctly as per dhd | 12:03 |
| mal | sledges: we should run a diff on discovery and pioneet device repos and defconfig files | 12:04 |
| sledges | we will have droid-hal-pioneer.spec and droid-hal-discovery.spec | 12:04 |
| sledges | where lunch combos etc will be diversified | 12:05 |
| mal | sledges: so the same OBS repo would be used for two several devices? I'm confused | 12:05 |
| mal | is that even possible | 12:05 |
| sledges | yes, because ssu will point to common repo, and device descriptor will make it distinct | 12:06 |
| mal | still wondering how the build dependencies will work | 12:06 |
| sledges | -$DEVICE | 12:06 |
| mal | but how does middleware know which one to use | 12:06 |
| mal | when building | 12:07 |
| sledges | ask mkosola :D | 12:07 |
| mal | or maybe I'm misunderstanding something | 12:07 |
| mkosola | mal: sledges: We don't know yet either 100% sure if it is working, let see | 12:11 |
| sledges | if it doesn't, it would resort to qualcomm:sony-nile-pioneer/sony-nile-discovery, but definitely not to per-variant (e.g. h1234) | 12:14 |
| mal | yes, we'll see | 12:14 |
| r0kk3rz | sledges o/ | 12:15 |
| Mister_Magister | Mikaela: been there done that | 12:18 |
| Mister_Magister | sorry i mean mal | 12:18 |
| Mister_Magister | had same obs repo for titan and thea untill it got too problematic | 12:18 |
| ghosalmartin | sledges, while your still about, do you know what CONFIG_VT kernel flag is for? :P | 12:27 |
| Mister_Magister | ghosalmartin: check documentation :P | 12:29 |
| ghosalmartin | Mister_Magister, I have done, but other devices have reached rootfs without it. so am curious if its a legacy thing | 12:30 |
| Mister_Magister | well iirc couiple of flags needed in merkernelcheck arent really needed | 12:31 |
| ghosalmartin | so just wanna make sure thats its not a red herring | 12:32 |
| Mister_Magister | sure | 12:33 |
| Mister_Magister | mal: dont know? | 12:33 |
| Mister_Magister | doesn't* | 12:34 |
| mal | not sure what that is used for | 12:35 |
| ghosalmartin | i honestly think its doesnt matter do much, my marlin device got to telnet easily from rootfs. | 12:45 |
| Mister_Magister | and im wondering how to port phone without twrp or lineage | 12:45 |
| mal | it's not really that difficult to make a fastboot flashable sailfish image, I have done it myself for fp2 before | 12:49 |
| Mister_Magister | mal: but i dont have lineage | 12:50 |
| mal | do you have aosp? | 12:50 |
| Mister_Magister | mal: only sources from vendor | 12:50 |
| Mister_Magister | and GSI | 12:50 |
| mal | only kernel or some more? | 12:50 |
| Mister_Magister | will have to check | 12:51 |
| Mister_Magister | about GSI we were talking before that it could probably be possible | 12:51 |
| Mister_Magister | ku | 12:51 |
| Mister_Magister | mal: https://www.asus.com/Phone/ZenFone-5-ZE620KL/HelpDesk_Download/ | 12:53 |
| Mister_Magister | mal: looks like kernel only | 12:55 |
| Mister_Magister | mal: but GSI is like AOSP isn't it | 12:57 |
| mal | no idea, not sure if I have heard of that before | 12:58 |
| Mister_Magister | probably naht | 13:01 |
| Mister_Magister | ill look into GSI | 13:01 |
| Mister_Magister | (Generic System Image) | 13:01 |
| Thaodan | sledges: can we put things like simconfig into a package? like ofono-config-sony-hx113? | 13:23 |
| guhl | ghosalmartin, still no luck but i found out that we do not have a ttyMSM0 probably because CONFIG_SERIAL_MSM is not set | 13:23 |
| guhl | so using this in the console parameter might not help | 13:23 |
| guhl | i enabled it and rebuild or otherwise we could try ttyHS0 that is created by SERIAL_MSM_HS | 13:26 |
| Mister_Magister | mal: fancy looking at this? https://github.com/phhusson/treble_experimentations/wiki/How-to-build-a-GSI%3F | 13:49 |
| Mister_Magister | looks lot like simple aosp building | 13:50 |
| Mister_Magister | maybe i can use that | 13:50 |
| mgrover | guhl, we could but am not sure what ti would yield | 14:13 |
| *** mgrover is now known as ghosalmartin | 14:15 | |
| Thaodan | sledges: are there any docs on sailfish-fpd-slave? | 14:30 |
| Thaodan | I saw its based on vendor-sonys-oss-fingeprint, is there a chance that it goes oss? | 14:31 |
| guhl | ghosalmartin, well i disabled the check for CONFIG_VT in init-script and i changed the test if to use configfs as described before renamed the sailfishos folder again and booted untl "init-debug in real rootfs" | 14:40 |
| ghosalmartin | guhl, can you access telnet? | 14:40 |
| guhl | because i created init_enter_debug2 in / of the real rootfs | 14:41 |
| guhl | yes on 2323 | 14:41 |
| guhl | but if i do - echo "continue" >/init-ctl/stdin | 14:43 |
| ghosalmartin | guhl, ahh did you add the usb-moded.ini config? | 14:44 |
| ghosalmartin | it could be usb-moded freaking out, i had a similar issue with marlin, after adding that file it stopped segfaulting | 14:44 |
| guhl | to startup it crashed (reboots). just now i did it again and got this ksmg | 14:45 |
| guhl | https://pastebin.com/8bjKP7qU | 14:45 |
| ghosalmartin | guhl, mask droid-hal-init? | 14:46 |
| ghosalmartin | ohh it could be the usb import in one of the .rc files | 14:46 |
| guhl | i did not do anything with the inits yet - this is the first try | 14:47 |
| guhl | do you want me to paste my init-script or is it clear how i made it work? | 14:47 |
| ghosalmartin | guhl, please | 14:48 |
| ghosalmartin | also i guess disable this? https://github.com/LineageOS/android_device_xiaomi_msm8998-common/blob/lineage-15.1/rootdir/etc/init.qcom.rc#L28 | 14:48 |
| guhl | ghosalmartin, https://github.com/guhl/hybris-boot_chiron/commit/3bf1ffb3425b64793145db25948d20895c6d8339 | 14:54 |
| guhl | you need to add androidboot.configfs=true to the kernel cmd_line to make this work | 14:55 |
| ghosalmartin | guhl, ahh i see, odd that it works in first telnet and not 2nd then | 14:55 |
| mal | guhl: wouldn't it be easier to just disable the old android gadget driver from kernel and not patch hybris-boot for that | 14:59 |
| guhl | mal, yes it would | 15:00 |
| guhl | it would also be easier if hybris-boot would check the kernel config (as it does anyway for other stuff) instead of checking the existence of files | 15:01 |
| mal | still a bit odd why the kernel has some remnants of the old driver enabled if those are not used | 15:02 |
| mal | guhl: that would also be possible, just wondering what would your device report in that case | 15:02 |
| guhl | well for example it would report CONFIG_USB_CONFIGFS=y and CONFIG_USB_CONFIGFS_RNDIS=y and i think that would be a sufficient test | 15:07 |
| guhl | but that is a kind of academic discussion | 15:09 |
| guhl | now i rather focus on the further init process | 15:10 |
| mal | guhl: earlier the check was the other way and it caused issues on some device | 15:10 |
| ghosalmartin | is the gadget driver CONFIG_USB_GADGET | 15:11 |
| guhl | that is one reason why i like my way. I set the parameter in BoardConfig and can be sure what will happen | 15:11 |
| mal | that would require new porters to first figure out which driver their device uses and not sure if that is good | 15:12 |
| ghosalmartin | a warning wouldnt hurt in the mer kernel spec | 15:13 |
| guhl | this https://pastebin.com/CSnxgL2G is the output of logcat before the reboot | 15:19 |
| mal | guhl: which android base? | 15:20 |
| mal | guhl: show your config repo | 15:20 |
| guhl | it is LOS 15.1 | 15:21 |
| mal | looks like you might be missing some things from config repo but need to see what you have there to be sure | 15:23 |
| * guhl is not sure if he want to show the config repo as he fears mal is going to laugh at him | 15:24 | |
| guhl | https://github.com/guhl/droid-config-chiron/tree/hybris-15.1 | 15:24 |
| mal | these https://github.com/mer-hybris/droid-config-sony-nile/tree/master/sparse/usr/libexec/droid-hybris/system/etc/init | 15:24 |
| mal | yes, missing the ones I linked | 15:24 |
| mal | without those the device can't work | 15:24 |
| guhl | k thanks | 15:26 |
| guhl | mal, btw did you ever see this on a device that the rootfs can not be installed via recovery (TWRP) because tar was killed (out of memory) | 15:34 |
| mal | guhl: several reports from some porters of that issue, downgrading twrp seems to have worked for some, it's because of too strict oom killer or something | 15:37 |
| guhl | k will try | 15:38 |
| guhl | no luck, every available twrp fails -> manual install in LOS | 15:52 |
| mal | ok | 15:53 |
| ghosalmartin | guhl, mine still reboots, possibly just a bad zip or extract? | 15:54 |
| guhl | ghosalmartin, did you touch in /data/.stowaways/sailfishos/init_enter_debug2 to make it stop before init | 15:56 |
| guhl | btw. topic 11.1 Extracting the rootfs via adb could be updated as the rootfs file is not created as gz but bz2 | 15:59 |
| ghosalmartin | guhl, yeahh i did, just reboots, ive a feeling somethings broken from my side | 16:07 |
| ghosalmartin | when I enter /target the rootfs is there though | 16:08 |
| ghosalmartin | guhl, do you know how to clear outpstore | 16:09 |
| guhl | did you check that the kernel really built without CONFIG_VT (Check in out/target/product/chiron/obj/KERNEL_OBJ/.config | 16:12 |
| guhl | i had problems with that too until i completely rebuilt the kernel by deleting out/target/product/chiron/obj/KERNEL_OBJ/ | 16:13 |
| guhl | OK thats better, not rebooting anymore but stuck at LOS animation :-) | 16:25 |
| ghosalmartin | guhl, i havent built with CONFIG_VT in a while | 16:28 |
| ghosalmartin | i get to telnet 1 atm | 16:29 |
| mal | guhl: logs logs logs | 16:29 |
| guhl | mal, I am flooded by logs | 16:31 |
| guhl | mal https://pastebin.com/GMZ0ZXst (dmesg) | 16:34 |
| guhl | journalctl https://pastebin.com/pWPG37zc (where it loops atm) | 16:36 |
| mal | guhl: is that dmesg incomplete | 16:37 |
| mal | guhl: does any mount service fail? | 16:38 |
| ghosalmartin | guhl, try to su in as nemo and start a service | 16:40 |
| ghosalmartin | it might complain of permissions | 16:40 |
| ghosalmartin | then its an issue in libbinder | 16:40 |
| guhl | logcat https://pastebin.com/vb4y0J8Q | 16:41 |
| guhl | mal, yes the dmesg paste is very incomplete | 16:42 |
| mal | guhl: check the mount services first | 16:43 |
| guhl | mal, https://pastebin.com/A4UEWYz7 | 16:45 |
| vknecht | could it be a case of disabling qseecomd ? | 16:50 |
| guhl | this https://pastebin.com/MdELYuNg is a journalctl more from the beginning | 16:51 |
| mal | guhl: a lot of permission issues | 16:53 |
| mal | for example Failed to create generator directory /run/user/100000/systemd/generator: Permission denied | 16:53 |
| T4 | NoMeat (MISSING @USERNAME! telegram.org/faq#usernames-and-t-me) was added by: NoMeat (MISSING @USERNAME! telegram.org/faq#usernames-and-t-me) | 17:02 |
| T4 | dreamofdeath_xl was added by: dreamofdeath_xl | 17:02 |
| guhl | mal, well i have no idea how to handle that | 17:16 |
| guhl | these services keep crashing/restarting | 17:16 |
| guhl | 6588 ? Sl 0:00 /vendor/bin/imsdatadaemon | 17:16 |
| guhl | 6592 ? Sl 0:00 /vendor/bin/ims_rtp_daemon | 17:16 |
| guhl | 6602 ? Sl 0:00 /vendor/bin/imsrcsd | 17:16 |
| guhl | 6635 ? S 0:00 /system/bin/tombstoned | 17:16 |
| guhl | 6641 ? S 0:00 /vendor/bin/qti | 17:16 |
| guhl | (that was only 5 lines :-) | 17:17 |
| T4 | dreamofdeath_xl was removed by: dreamofdeath_xl | 17:17 |
| Mister_Magister | mal: gotta time? | 17:22 |
| mal | still at work | 17:24 |
| mal | I probably should leave already | 17:24 |
| Mister_Magister | mal: will you have energy today to help me? | 17:24 |
| T4 | holmes465 was added by: holmes465 | 17:26 |
| mal | probably not much, I have felt terrible today | 17:26 |
| T4 | holmes465 was removed by: holmes465 | 17:27 |
| mal | I did rebase my local anbox again a couple of days ago but didn't build it yet | 17:27 |
| vknecht | about variant handling : how could it be handled at runtime, eg. depending on properties ? iiuc for SailfishX it's handled at build time (distinct spec files) ? | 17:29 |
| Mister_Magister | mal: ah sad | 17:29 |
| mal | vknecht: depends on what you want to detect at runtime | 17:31 |
| Mister_Magister | mal: i got my zenfone 5z and i puke looking at android. Need to slap sfos on it :P I have to look into GSI stuff | 17:32 |
| mal | vknecht: in case of single-sim vs dual-sim you can even use the same android image for both but then you can't detect it at runtime because the codename might not be correct | 17:32 |
| mal | Mister_Magister: I thought you were looking for anbox help | 17:32 |
| Mister_Magister | yeah thats just offtopic | 17:33 |
| Mister_Magister | we were talking bout that today :P | 17:33 |
| mal | yes, but you ask about anbox so often it's the default topic I assume | 17:33 |
| Mister_Magister | its not an easy task slapping sfos there more longer term so there is no much you can do about it rn other than sending ideas | 17:33 |
| Mister_Magister | mal: yes it is | 17:34 |
| Mister_Magister | im just talking :) | 17:34 |
| Mister_Magister | main is anbox | 17:34 |
| *** ChanServ sets mode: +v T4 | 17:34 | |
| vknecht | basically persist.radio.multisim.config ? or rm.cm.curef which is 6045K for multisim, and 6045Y for EU, other letters for other regions... | 17:35 |
| mal | Mister_Magister: based on the link you gave earlier the build in that is lineage based | 17:35 |
| mal | vknecht: as an example in xa2 the difference in properties is this https://github.com/sonyxperiadev/device-sony-pioneer/blob/master/aosp_h4113.mk#L16 | 17:36 |
| mal | and we can use h4113 image also for h3113 | 17:36 |
| Mister_Magister | mal: but the question is can i build sfos using that? | 17:37 |
| mal | Mister_Magister: maybe | 17:37 |
| Mister_Magister | mal: what do you mean by maybe? | 17:38 |
| vknecht | the source tree is the same, the LOS port uses the properties to handle multisim case | 17:38 |
| Mister_Magister | like you dont know if it will actually work but it is probably possible to use those sources? | 17:38 |
| mal | Mister_Magister: not sure how much changes we need, at least that has some extra changes to android repos so we would need separate branches | 17:40 |
| Mister_Magister | mal: hmmm okay but is it possibly doable? | 17:41 |
| vknecht | https://github.com/vknecht/android_device_alcatel_idol3/blob/2b0f7d0911d6a0b76143aec5439fe6250c2d21f0/init/init_idol3.cpp#L120 | 17:41 |
| T4 | and0kr was added by: and0kr | 17:43 |
| Mister_Magister | mal: like… is it worthy giving a shot? | 17:43 |
| kimmoli | Mister_Magister: just do it | 17:44 |
| kimmoli | then we all know | 17:44 |
| vknecht | XD | 17:44 |
| Mister_Magister | kimmoli: seems like a way to live life xD | 17:44 |
| Mister_Magister | first of all lets get this baby open | 17:45 |
| vknecht | offtopic : would anyone have experience with android CTS ? trying to sort the safetynet mess for "my" LOS base... looking for narrowing down to pertinent test, the full one has been running for 20+ hours now and only 7202/81464 tests in current set... | 17:52 |
| vknecht | (there's #idol3 to spare this channel :) | 17:55 |
| kimmoli | Mister_Magister: https://pm1.narvii.com/6057/feeb55f7005e481c25e65664ff82918f470986c5_hq.jpg | 17:56 |
| Mister_Magister | kimmoli: yes, my master | 17:56 |
| kimmoli | vknecht: why that reminds me always about this? https://www.ido.fi/tuotteet/wct-urinaalit-ja-pesuistuimet/lattialle-asennettavat/ido-seven-d-13-peruskorjaus-wc-kiinnitysreiilla-s-lukko/7201/ | 17:57 |
| * kimmoli stops spamming. | 17:57 | |
| kimmoli | well, one to go https://www.youtube.com/watch?v=kfVsfOSbJY0 specially for sledges | 17:57 |
| vknecht | because once it's in, it cannot get out :) | 17:58 |
| Mister_Magister | kimmoli: are you from finnland? | 18:01 |
| kimmoli | lucky guess sherlock | 18:02 |
| Mister_Magister | kimmoli: android is paska | 18:02 |
| Mister_Magister | u'll get that then | 18:02 |
| kimmoli | das ist mir Wurscht | 18:04 |
| Mister_Magister | i don't get deutsch | 18:06 |
| * Mister_Magister google translates | 18:06 | |
| Mister_Magister | oh | 18:06 |
| Mister_Magister | mal: how does sfos handle Bluetooth 5.0 xD (itz ze joke) | 18:21 |
| guhl | vknecht, regarding disabling qseecomd how would you do that? | 18:33 |
| guhl | i tried it similar to https://github.com/mer-hybris/droid-config-sony-nile/blob/master/sparse/usr/libexec/droid-hybris/system/etc/init/netd.rc | 18:33 |
| guhl | but this leads to | 18:37 |
| guhl | droid-hal-init: /usr/libexec/droid-hybris/system/etc/init/qseecomd.rc: 1: ignored duplicate definition of service 'qseecomd' | 18:37 |
| mal | Mister_Magister: I don't know what features BT 5.0 brings, it should depend on device support and bluez5 | 18:43 |
| vknecht | guhl, just adding "disabled" in the existing service definition, in init.qcom.rc I think | 18:43 |
| Mister_Magister | mal: i was just joking bt doesnt work on newer devices anyways | 18:43 |
| mal | Mister_Magister: bluez5 seems to support 5.0, at least btmon can decode the hci commands | 18:44 |
| Mister_Magister | kk | 18:45 |
| Mister_Magister | but nothing >snap810 has bt working right | 18:45 |
| Mister_Magister | ? | 18:45 |
| mal | not sure if that is the defining factor, some devices could use different bt chip | 18:47 |
| guhl | vknecht, the original is in /vendor/etc/init/hw/init.target.rc | 18:47 |
| Mister_Magister | suure :P | 18:47 |
| guhl | well ghosalmartin, mal I think now we know why we need CONFIG_VT | 18:57 |
| guhl | May 29 14:29:13 Sailfish systemd[3018]: pam_systemd(autologin:session): Failed to create session: Seat has no VTs but VT number not 0 | 18:57 |
| * Mister_Magister feels sad that guhl didnt mention him as he was trying his best helping :( | 18:58 | |
| guhl | Mister_Magister, well I am sorry about that but I was just referring to question from ghosalmartin to mal if and for what we need CONFIG_VT | 19:00 |
| ghosalmartin | guhl, ah...damn | 19:00 |
| Mister_Magister | guhl: i dont know if we need too :P | 19:00 |
| vknecht | guhl, ah, maybe a worthy addition to https://public.etherpad-mozilla.org/p/faq-hadk line 697, it's init.qcom.rc in my 14.1 base | 19:00 |
| Mister_Magister | whatev :P | 19:00 |
| ghosalmartin | vknecht, i wrote that, glad someone found it useful :P | 19:01 |
| vknecht | indeed it was, for calls and maybe cam iirc, thanks ghosalmartin :) | 19:03 |
| ghosalmartin | it stumped me for a while as well. it never used to be an issue on cm13 | 19:03 |
| ghosalmartin | am off for a few pints, night and guhl good lucl | 19:04 |
| Mister_Magister | mal: they did lineage and ressurection remix using GSI | 19:12 |
| Mister_Magister | i think we can make SailfishOS GSI | 19:12 |
| Mister_Magister | and slap sfos on every GSI enabled device xD | 19:12 |
| Mister_Magister | (forget last part) | 19:13 |
| Mister_Magister | mal: what if i just sync hybris-15.1 sources and add this https://github.com/phhusson/treble_manifest | 19:30 |
| Mister_Magister | like… yolo it | 19:31 |
| mal | Mister_Magister: if you read the instructions from earlier link you see that you need to patch hybris branch | 19:42 |
| Mister_Magister | mal: somebody from this project told me that drivers are "treblized" and hybris doesnt support that | 19:47 |
| Mister_Magister | mal: or do you actually mean patching libhybris for treble support | 19:47 |
| mal | we use treble (or hidl whatever is the correct term) already in some parts of android 8 support | 19:51 |
| Mister_Magister | oh but what about drivers? | 19:52 |
| Mister_Magister | and blobs on vendor partition | 19:52 |
| mal | ? | 19:52 |
| Mister_Magister | well i can't really explain it as i dont understand it myself | 19:52 |
| Mister_Magister | just somebody said they dont work with libhybris | 19:52 |
| mal | give me some example | 19:54 |
| Mister_Magister | maybe if i get his nick on this channel he'll explain | 19:54 |
| mal | Mister_Magister: we communicate directly with the android binder using the new APIs from a glib based library | 19:59 |
| mal | with these interfaces https://github.com/LineageOS/android_hardware_interfaces | 20:00 |
| Mister_Magister | mal: how much is sfos ready for 8.0 based port? | 20:00 |
| mal | depends on how much the devices uses the new interfaces and how many old hal API implementations it still has available | 20:01 |
| Mister_Magister | well its 2018 device so you can guess :) | 20:02 |
| Mister_Magister | what if not many? :P | 20:02 |
| mal | almost everything | 20:02 |
| Mister_Magister | i mean if there is not many old hal api | 20:03 |
| mal | not sure how many areas can be affected | 20:06 |
| mal | sensors might be broken for example, not sure if they have moved audio stuff to new interfaces | 20:07 |
| Mister_Magister | we'll see then | 20:07 |
| Mister_Magister | too many variables :C | 20:07 |
| Mister_Magister | oakay first i want to get lineage using treble | 20:07 |
| Mister_Magister | lets focus on that for now | 20:07 |
| Mister_Magister | we'll yolo rest xD | 20:07 |
| b0n0 | It'd be better to stay on 7.1 | 20:09 |
| Mister_Magister | there is no 7.1 in 2018 | 20:09 |
| b0n0 | What? | 20:09 |
| Mister_Magister | device came out with 8.0 | 20:09 |
| Mister_Magister | and has no custom rom | 20:09 |
| b0n0 | Ah | 20:10 |
| T4 | <benzands> also some devices in late 2018 will come with 9.0 :) | 20:10 |
| b0n0 | At least it has treble | 20:10 |
| Mister_Magister | b0n0: thats what this whole talk is about | 20:10 |
| Mister_Magister | if its possible to get sfos on treble | 20:11 |
| b0n0 | I dont know | 20:11 |
| b0n0 | I dont think so | 20:11 |
| b0n0 | The sailfish device port has to be done i think | 20:12 |
| Mister_Magister | you just ignore whole talk lol | 20:12 |
| b0n0 | Yeah :P | 20:12 |
| b0n0 | You can always port android 7.1 | 20:13 |
| b0n0 | But if youre noob it will be difficult because there are many changes from 7.1 to 8.1 | 20:13 |
| b0n0 | Probably the hardest android upgrade to port ever | 20:14 |
| mal | Mister_Magister: HIDL is part of treble, we use the new HAL interfaces directly via binder so I think that was the whole point of treble, not sure if there was any other things in treble than that | 20:22 |
| Mister_Magister | mal: okay gotcha | 20:22 |
| Mister_Magister | now i need to try lineage but i dont have twrp or anything to backup stock system.img | 20:23 |
| Mister_Magister | and without backup its risky | 20:23 |
| mal | but there is only some support implemented so far, ril and gnss, sensors is trivial but I know no devices which use the new API yet | 20:24 |
| b0n0 | Treble needs a special twrp iirc | 20:24 |
| Mister_Magister | b0n0: i have no twrp | 20:24 |
| b0n0 | Install it | 20:25 |
| Mister_Magister | i have no twrp? | 20:25 |
| Mister_Magister | like available lol | 20:25 |
| b0n0 | Port it | 20:25 |
| Mister_Magister | funny | 20:25 |
| b0n0 | Hack it | 20:25 |
| Mister_Magister | you can flash GSI images without twrp | 20:26 |
| b0n0 | I dont know, i dont have treble | 20:33 |
| b0n0 | My device is too hold, i have to port every single rom | 20:33 |
| T4 | Dinolek was added by: Dinolek | 21:19 |
| T4 | Dinolek was removed by: Dinolek | 21:20 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
