| T42 | <b100dian> hang on, rpm -e --no-deps qt5-qtfeedback | 00:14 |
|---|---|---|
| T42 | <b100dian> (as Elros said:) | 00:14 |
| T42 | <b100dian> mpf.. broken launcher:) I need to modify droid-version too | 00:14 |
| T42 | <b100dian> gonna try again tommorow, gn | 00:20 |
| kquote03[m] | Hello. I've fixed my $DEVICE.xml, it seems to go further, but I'm now getting a sed error. https://pastebin.com/wvd1hmUE | 08:21 |
| kquote03[m] | Here's my fixup-mountpoints (which i suspect to be the culprit, but when comparing to a vanilla fixup-mountpoints, nothing seems wrong) https://pastebin.com/NT43pePZ (my device can be found at the bottom, before last entry) | 08:21 |
| kquote03[m] | verbose.log https://pastebin.com/m8VUrXYX | 08:22 |
| kquote03[m] | build.trace https://pastebin.com/9WfbiFnT | 08:23 |
| kquote03[m] | * last entry) (a5y17lte) | 08:24 |
| T42 | <edp_17> @elros34 : Can I ask a liitle help please? | 11:13 |
| T42 | <elros34> sure you can ask:) | 11:16 |
| T42 | <elros34> kquote03[m] see you mix up path, you have partname/by-name instead by-name/partname | 11:17 |
| T42 | <elros34> also check your fstab again because that path looks strange | 11:19 |
| T42 | <edp_17> @elros34 : Thanks. Actually I have a few questions. 😊 | 11:24 |
| T42 | <edp_17> So, I am trying to use a new kernel for the i9100. This is the kernel that is used for building a los17.1 android. (I need this kernel for waydroid, otherwise I would need to backport a lot). So, I downloaded the kernel source and applied all the sfos related change into it, built hybris-hal and installed the kernel on the device. It stuck on the boot logo and not accessible with telnet. | 11:24 |
| T42 | <edp_17> Question 1: | 11:24 |
| T42 | <edp_17> If I use a new kernel, should I use the relevant device tree too (android_device_samsung_i9100 and android_device_samsung_galaxys2-common)? | 11:24 |
| T42 | <edp_17> and vendor (proprietary_vendor_samsung)? | 11:24 |
| kquote03[m] | <T42> "<elros34> also check your..." <- alright, I'll look into it later today, thanks! | 11:34 |
| T42 | <elros34> @edp_17 hard to answer but generally yes you should use matched device sources. For getting telnet that is irrelevant but everything else could change especially from cm12 or 14? to los17 transition | 11:44 |
| T42 | <edp_17> Yeah, from los14.1 to 17.1. | 11:45 |
| T42 | <edp_17> I've done similarly for the treltexx (from los14.1 to 17.1) and magically the new kernel just works. (I guess got lucky with that.) | 11:45 |
| T42 | <elros34> look at sailfish os changes, they switched to binder everywhere for hybris-17 compared to older bases. | 11:46 |
| T42 | <elros34> Anyway fixing telnet shoudn't be hard so you can always continue your work | 11:46 |
| T42 | <edp_17> Although, on the treltexx there is a little issue because when I shut down the device it always gets rebooted. I guess this is because I didn't match the device tree and I didn't re-build any sfos parts either. | 11:47 |
| T42 | <edp_17> Speaking about fixing the telnet. I've tried to bring up the usb network manually but didn't work. Can you please suggest how to get in to the device? | 11:48 |
| T42 | <elros34> 2323? | 11:48 |
| T42 | <edp_17> Tried that. 😃 | 11:49 |
| T42 | <elros34> I mean do you see that 23 or 2323 port is set up in host dmesg | 11:49 |
| T42 | <edp_17> I don't see any port numbers in there. | 11:52 |
| T42 | <elros34> so what do you see on first boot after flashing? If you do not see mer boatloader messages then telnet will not work | 11:53 |
| T42 | <edp_17> The 'dmesg -w | grep mer' doesn't bring up anything while the device is booting. | 12:02 |
| T42 | <edp_17> and connected | 12:02 |
| T42 | <elros34> but something shows up? also I suggest grep -i mer | 12:04 |
| T42 | <edp_17> Oh well. SImply the 'dmesg -w' doesn't show anything. (I kept it running in a console and in another one I flashed the kernel with heimdall. This way I saw everything that happened since the device got the new kernel and restarted as it was connected to the PC all the time.) | 12:09 |
| T42 | <edp_17> So, it seems bad luck. 😃 | 12:09 |
| T42 | <TheVancedGamer> pull pstore (re @edp_17: Oh well. SImply the ...) | 12:09 |
| T42 | <edp_17> @TheVancedGamer : pstore? What's that? | 12:10 |
| T42 | <TheVancedGamer> pull this directory in recovery: /sys/fs/pstore (re @edp_17: @TheVancedGamer : ps...) | 12:11 |
| T42 | <TheVancedGamer> it will have logs for the bootup process | 12:11 |
| T42 | <TheVancedGamer> from there you can diagnose a little bit | 12:11 |
| T42 | <edp_17> I cannot boot into recovery either. 😞 | 12:11 |
| T42 | <TheVancedGamer> fastboot? wait no nvm it's a samsung (re @edp_17: I cannot boot into r...) | 12:11 |
| T42 | <edp_17> Yeah, it is samsung. | 12:11 |
| T42 | <elros34> @edp_17 if nothing then obviously telnet is not up. Check whether you have initramfs in hybris-boot.img | 12:12 |
| T42 | <edp_17> @elros34 : I don't flash hybris-boot.img to this device. I flash the kernel file directly onto the kernel partition. | 12:14 |
| T42 | <TheVancedGamer> huh? (re @edp_17: @elros34 : I don't f...) | 12:14 |
| T42 | <TheVancedGamer> kernel takes boot.img! | 12:14 |
| T42 | <elros34> ah I need to remember finally that i9100 do not use abootimg | 12:14 |
| T42 | <TheVancedGamer> wait (re @elros34: ah I need to remembe...) | 12:14 |
| T42 | <TheVancedGamer> is this i9100 we are talking about? | 12:14 |
| T42 | <edp_17> I remembered that needed to extract the hybris-boot.img-ramdisk from hybris-boot.img and then copy to the $ANDROID_ROOT. When I used los14.1 kernel. | 12:15 |
| T42 | <edp_17> Yes. (re @TheVancedGamer: is this i9100 we are...) | 12:15 |
| T42 | <elros34> so make sure your hybris-boot.img-ramdisk have 0.5MB size | 12:16 |
| T42 | <edp_17> As I just replaced the kernel folder, I kept hybris-boot.img-ramdisk in $ANDROID_ROOT and changed usr/galaxys2_initramfs.list to point to that hybris-boot.img-ramdisk. Just like for the other kernel. | 12:17 |
| T42 | <edp_17> The size of hybris-boot.img-ramdisk is 776,192 B | 12:17 |
| T42 | <edp_17> It looks okay. | 12:17 |
| T42 | <edp_17> However, this extracted a hybris-boot.img that was created for the other kernel 14.1. Is that a problem? | 12:18 |
| T42 | <elros34> no | 12:18 |
| T42 | <edp_17> Huh. 😊 | 12:18 |
| T42 | <elros34> so devices reboots after some time like 60s? | 12:19 |
| T42 | <edp_17> Nope. It just stays on the nice and shiny Samsung logo. | 12:25 |
| T42 | <elros34> then most likely it doesn't even reach initramfs stage | 12:26 |
| T42 | <edp_17> Looks like it. | 12:29 |
| T42 | <edp_17> If you are interested I pushed all changes to github: https://github.com/edp17/android_kernel_samsung_smdk4412/tree/lineage-17.1 | 12:30 |
| T42 | <elros34> no sure how it works for your device but have you tried with do_skip_initramfs=0 ini initramfs.c? | 12:31 |
| T42 | <edp_17> No but will try now. | 12:31 |
| T42 | <edp_17> Are you referring this line? https://github.com/edp17/android_kernel_samsung_smdk4412/blob/lineage-17.1/init/initramfs.c#L588 | 12:33 |
| T42 | <TheVancedGamer> Yes (re @edp_17: Are you referring th...) | 12:34 |
| T42 | <elros34> yes there is only 1 place where 1 is written | 12:34 |
| T42 | <TheVancedGamer> return 1 exists (re @elros34: yes there is only 1 ...) | 12:34 |
| T42 | <edp_17> Thanks. It is building now. Will let you know the result. 😊 | 12:39 |
| T42 | <edp_17> Nothing changed. Stuck on the boot logo and dmesg -w doesn't show the device. | 12:47 |
| T42 | <elros34> have you tried too boot to recovery also? I can see some changes related to that in you initramfs | 12:49 |
| T42 | <edp_17> How do you mean? I pressed the volume up+home+power which should take me to recovery but didn't do that. I also flahsed the isorec-ramdrive.cpio.lzo (extracted from twrp2.8.7.0.img) to the recovery partition but didn't boot into recovery either. | 13:00 |
| T42 | <elros34> you have added load_image=/stage1/recovery.cpio but I do not see place where /stage1/recovery.cpio is created. Is /dev/block/mmcblk0p6 your recovery on lineage-17? | 13:02 |
| T42 | <edp_17> It didn't boot with the original init file either. Although, I wasn't sure when I added that. | 13:04 |
| T42 | <edp_17> I think worth a try to revert tha change and see whether the do_skip_initramfs=0 has any effect. | 13:05 |
| T42 | <edp_17> If not, I think give up on this kernel and will use the cm14.1 one and try to backport all the missing bits and bobs that requires for waydroid. | 13:06 |
| T42 | <elros34> I think you should revert all changes, maybe except skip_initramfs and try to get recovery working. If recovery works then you are sure that initramfs works and you can try to replace original /stage1/boot.cpio with hybris one | 13:07 |
| T42 | <edp_17> Okay. Thanks. | 13:08 |
| T42 | <elros34> btw is it even possible to get waydroid on arm kernel? I assume nobody cares about arm | 13:10 |
| T42 | <TheVancedGamer> Yeah (re @elros34: btw is it even possi...) | 13:13 |
| T42 | <TheVancedGamer> WayDroid works on arm | 13:13 |
| T42 | <elros34> oh nice, so only one issue: ancient old kernel release | 13:14 |
| T42 | <edp_17> It is working nice on my treltexx port. 😊 (re @elros34: btw is it even possi...) | 13:14 |
| T42 | <edp_17> I reverted those changes except the skip_initramfs and the result is the same. No recovery either. | 13:19 |
| T42 | <edp_17> Thanks for your time, I'll give on this now. 😊 | 13:19 |
| T42 | <elros34> so do you have some known working recovery image for that kernel? | 13:19 |
| T42 | <elros34> did you revert also /stage1/boot.cpio? Not it should boot to android | 13:21 |
| Thaodan | Check if the resulting boot.img is big enough. The skip init ramfs should be easy, I guess setting up the usb network is the issue. You can check from the Android recovery if the way its done in initramfs is working. | 13:36 |
| T42 | <elros34> I think if use network would fail then device would reboot after 60s but that is not the case here | 13:40 |
| T42 | <edp_17> Yes, I do. Although, haven't tried to install it on device yet. (re @elros34: so do you have some ...) | 13:44 |
| mal | did you get a warning about cpio during build? | 13:44 |
| mal | which android base? | 13:44 |
| T42 | <edp_17> Are you referring to this line "load_image=/stage1/boot.cpio"? I didnt add that, it was there, so I kept. (re @elros34: did you revert also ...) | 13:44 |
| T42 | <elros34> no, in usr/galaxys2_initramfs.list also be aware that this https://github.com/edp17/android_kernel_samsung_smdk4412/blob/1278a0a2ccc5b6c8a0b1e988223075f3a33278b7/usr/galaxys2_initramfs_files/init#L42 breaks recovery for sure | 13:45 |
| T42 | <edp_17> The boot.img size is 4,150,172 B so it looks big enough. The issue is that I cannot boot into recovery either. (re @SailfishFreenodeIRCBridgeBot: <Thaodan>Check if th...) | 13:46 |
| T42 | <edp_17> mal: No, there is no warning about cpio during build. It is Android base Los17.1 | 13:47 |
| T42 | <edp_17> Should I remove that line then? (re @elros34: no, in usr/galaxys2...) | 13:50 |
| T42 | <elros34> not remove line, revert whole commit https://github.com/edp17/android_kernel_samsung_smdk4412/commit/1278a0a2ccc5b6c8a0b1e988223075f3a33278b7 | 13:51 |
| T42 | <edp_17> That's what I done. | 13:54 |
| T42 | <edp_17> I put back the original usr/galaxys2_initramfs_files/init file that I got with this kernel. | 13:54 |
| Thaodan | Than check in Android recovery if setting up usb network works | 13:56 |
| Thaodan | rinigus: Care to add https://github.com/sailfishos-sony-tama/miniaudiopolicy/blob/main/miniaudiopolicy.cpp to droidmedia? | 19:46 |
| rinigus | Thaodan: sure, I can look into it. Late right now, I'll try to remember in the next few days. Feel free to remind me | 19:54 |
| Thaodan | rinigus: Np just asking, I feel it belongs to droidmedia. | 19:55 |
| Thaodan | I would need it for Xperia 1 to, I wonder why its not needed with Xperia 10 II | 19:55 |
| rinigus | You know better - I don't have an overview in this respect. Great to hear that you work on Xperia 1! | 19:56 |
| Thaodan | rinigus: Is there anything different in your miniaudiopolicy than the AudioPolicyService::instantiate(); pushed to the end? | 21:45 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!