Tuesday, 2023-03-21

*** Danct12 is now known as Guest837602:03
*** amccarthy is now known as Guest839406:05
*** amccarthy_ is now known as amccarthy06:06
*** amccarthy is now known as Guest839807:27
*** amccarthy_ is now known as amccarthy07:27
T42<k1gen> do I have to flash twrp or I can flash sfos with fastboot?07:34
T42<edp_17> @k1gen : You can flash sfos without twrp just don't forget to flash the android base first. Also, I do recommend a recovery (I use twrp) because you can do backup/restore without a computer.08:29
T42<k1gen> thanks08:29
T42<k1gen> I finally build lineage-18.1, was debugging it whole day. trying to flash sfos zip on top of it got Updater process ended with signal: 7 error: http://ix.io/4rsD19:40
T42<k1gen> any clues?19:50
maldo you have any validation enabled for the zip when you try to install it?20:04
T42<k1gen> no, I skip zip verification in twrp20:12
T42<k1gen> no verifications whatsoever, twrp just fails without any info20:18
T42<k1gen> I unzipped it manually, where should I untar the tar.bz2?20:23
T42<k1gen> tar --numeric-owner -xvjf sailfishos-flame-release- -C /data/.stowaways/sailfishos, then flash hybris-boot.img to boot?20:25
T42<k1gen> I'm sorry, just saw this in hadk pdf20:26
malyes, that is the manual way20:27
T42<k1gen> how to copy static busybox from build environment and use it to unpack image in adb shell?20:28
mal@k1gen it might be easier to just flash kernel in uncompress roots in telnet at port 2320:30
mal*and uncompress20:30
T42<k1gen> ok20:30
malthat is how I usually do it if I need to manually extract rootfs for some reason20:30
T42<k1gen> nothing mer-related in dmesg/lsusb, no telnet, just google logo20:36
T42<k1gen> also rebooting after a minute20:37
T42<elros34> mal btw IIRC you were supposed to talk with somebody about changing sailfish image compression format to something better supported by TWRP20:38
T42<k1gen> dmesg recognises device eather as AOSP on flame or as Pixel 420:39
T42<b100dian> Do you know, if my init.qcom.usb.rc says e.g. `write /config/usb_gadget/g2/configs/b.1/strings/0x409/configuration "rndis"` is it safe to assume that I should change c.1 for b.1 in init shell script20:50
mal@elros34 I think I did mention it internally but nothing happened yet, I will bring it up again20:51
mal@k1gen if init-script can't setup usb interface it will reboot the device in 60 seconds20:53
mal@k1gen try adding these 3 usb option to kernel defconfig https://github.com/mer-hybris/android_kernel_sony_msm/blob/hybris-sony-aosp/LA.UM.9.12.r1/arch/arm64/configs/aosp_lena_pdx213_defconfig#L93120:55
T42<k1gen> mal: thank you21:02
T42<elros34> @b100dian if you have working kernel image now then flash hybris-boot.img instead recovery and then you can test such a change in /init-debug in sfos rootfs21:05
T42<b100dian> Thanks @elros34. Wouldn't that require telnet access? I have a working "kernel packaging" that reboots at different times when I change sleep times in init21:10
T42<elros34> no, hybris-boot do not stop in initramfs unless you have wrong kernel config or data partition then it switch to rootfs and execute same script again. You can use twrp or even lineage to edit script in /data/.stowaways/sailfishos21:12
T42<elros34> additionally you will have some debug messages in init.log21:13
T42<b100dian> I see now what you mean, thanks! it will take some time since I've only built the android bits till now.21:15
T42<elros34> btw do you have these usb configs mentioned 20minutes ago?^21:17
T42<b100dian> I don't have RMNET_BAM but I didn't on my 4.14 kernel in a previous port either. I'll add it in the next test if c.1->b.1 fails21:22
T42<k1gen> mal: now I have Mer Boat Loader in lsusb with iSerial Mer, how do I know the ip for telnet?22:01
T42<k1gen> I see one in ip a, but telnet says Connection refused22:03
T42<k1gen> neither port 23 nor 2323 work22:05
T42<k1gen> so I can't even check /diagnosis.log22:06
T42<k1gen> also in dmesg & lsusb there is Product: Refusing22:09
piggzmal: like you said in the logs, uninstalling gecko-camera-droid-plugin causes browser to use gmp-droid, and that works better here (no green garbage)22:10
T42<elros34> @k1gen you have either wrong /data partition in fixup-mountpoints or wrong kernel config, see hybris-boot/init-script for required configs22:12
T42<k1gen> thanks for suggesting, I'll debug this in the morning22:14
piggzhmm, GECKO_CAMERA_DROID_NO_MEDIA_BUFFER fixes it :)22:19
piggzso, does this mean gmp-droid is no longer needed to be installed?22:20
T42<b100dian> piggz: what green garbage? Is this non-accel vs accelerated or just decoding errors?22:21
T42<k1gen> @elros34 I don't have data partition in fixup-mountpoints because I don't have /data symlink in /dev/block/platform/soc/1d84000.ufshc/by-name22:21
T42<k1gen> what should I do?22:21
T42<elros34> what do you mean? you don't have userdata partition?22:22
T42<adampigg> when using gmp-droid, i had a patched version that allowed to not use droidConvert function for i420, as that caused decoding issues (re @b100dian: piggz: what green ga...)22:22
T42<adampigg> with the new browser plugin, i also had garbage, but that env var fixed it22:23
T42<b100dian> oh, I should know, I use the same patch. Thanks, I wasn't sure what you meant22:23
T42<b100dian> But - I have not updated to 4.5 yet :) Interesting that NO_MEDIA_BUFFER would fix the color conversion22:24
T42<k1gen> I'm sorry, I was looking for data. I do have userdata, and it is pointed to the right block device in fixup-mountpoints (re @elros34: what do you mean? yo...)22:24
T42<elros34> with correct by-name path?22:25
T42<k1gen> haven't been asleep for a long time (for me) I guess22:25
T42<k1gen> yeah, block/platform/soc/1d84000.ufshc/by-name/userdata sda15 (re @elros34: with correct by-name...)22:25
T42<k1gen> userdata -> /dev/block/sda15 in adb shell22:26
T42<elros34> and you compared this by-name path with your android's fstab?22:26
T42<k1gen> well, it's /dev/block/bootdevice/by-name/userdata in device repo fstab, not block/platform/soc/1d84000.ufshc/by-name/userdata22:27
T42<elros34> so you didn't. I wonder how to write instruction so people will really follow it. It's explicitly mentioned in hadk-hot to not guess by-name path but in vain. Maybe you with fresh porting experience have some ideas to improve that documentation22:29
T42<k1gen> I'm sorry for not being smart/experienced enough to make right guesses as I go, it's my first time after all. maybe after this journey is complete I will have enough experience and willing to improve documentation with you guys, who helped me all along. I think you noticed it's my third try with this device, and this time I've got much further then in last two tries, I really want to have daily-driveable linux device when I finish22:34
T42<k1gen> goodnight @elros34, I hope you will be online when I wake up :)22:35
T42<elros34> don't be sorry you are not the first one who have problem with understanding this. It's notorious problem so I am only inviting you (with fresh mind) to improve documention after you figure out everything so it can be easier to understand for new porters22:39
T42<b100dian> I think this is a great idea, fresh eyes to  give feedback on how steps can be improved22:39
malit's difficult to sometimes see how different ways something can be misunderstood22:57
malif there are some ideas for improving instructions for fixup creation those can still get into next hadk pdf, I can still make changes before new version of that pdf is released22:58
T42<elros34> I think this paragraph may be misleading: 'To get that mapping, you should flash and boot and image of your Android base and execute adb shell on your host and this:*'23:22
T42<elros34> I would change it to something like: To get that mapping, you must find android's fstab in device repository or in device itself and get by-name path like: block/bootdevice/by-name/userdata or block/platform/msm_sdcc.1/by-name/userdata from it. Then to get corresponding block devices you should flash and boot image of your Android base or recovery and execute adb shell on your host and this: ls -l /dev/block/platform/*/by-name/ on your device23:24
T42<k1gen> @elros34 you forgot the a/b partitions, that's where I left my first attempt. I understand that I need to pick one slot and use it, but have no clue at all of what to do after I get my build booting past telnet23:28
T42<k1gen> damn, I really should be sleeping rn23:28
T42<elros34> yeah:) but I am not good at writing docs that is why I put some random tips in hadk-hot which of course can be resued23:29

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