**** BEGIN LOGGING AT Fri Dec 10 02:59:56 2021 Dec 10 07:23:01 Morning! Dec 10 07:26:27 So, looks like I missed quite a few messages while travelling :p Dec 10 07:31:24 Herrie: yes, if the dependency on the meta-arm recipes becomes more justified, I don't see why we should try to avoid adding it Dec 10 07:32:05 so far I had only the firmware, so it was not worth it, that's all Dec 10 07:32:37 I just hope it doesn't bring any conflict or so Dec 10 07:46:54 Morning! Dec 10 07:47:12 Tofe: Yeah I was traveling yesterday all day too LOL Dec 10 07:47:30 Woke up at 4:30 (because of silly brain thinking I'll oversleep for my 9:30 flight).... Dec 10 07:47:47 Got home at 21:30... So long day flying to and back from Spain ;) Dec 10 07:48:18 But decided to toy a little with PinePhonePro Dec 10 07:48:24 Seems it comes with Android 9 out of the box Dec 10 07:50:17 Some Chinese version without Gapps, but managed to switch it to English Dec 10 07:51:27 December 2019 security patch LOL, kernel 4.4.167 Dec 10 07:52:22 Yeah, sounds like a basic Chinese version Dec 10 07:52:46 I guess any smartphone haadware does, nowadays Dec 10 07:56:42 Yeah, well it seems pretty fluid in general, but our aim is mainline of course Dec 10 08:13:35 Let me first upstream those pmOS support patches for rk3399/pinephonepro to u-Boot. Not sure why that wasn't done Dec 10 13:30:21 OK got kernel to compile now as well, needed some tweaks because of 5.16 but well Dec 10 16:20:33 Last obstacle seems u-boot for now Dec 10 17:14:11 u-boot should be farily similar to what we do for pinephone already, I guess? Dec 10 17:23:51 Tofe: Well it's Rockchip vs Allwinner, seems quite different Dec 10 17:25:56 This is what pmOS does for it: https://gitlab.com/postmarketOS/pmaports/-/merge_requests/2607/diffs#3cd848f3d6227eef9c70aafcde2d650fd43386d1 Dec 10 17:28:55 most of that is done by our u-boot recipe Dec 10 17:29:08 we need to use Martijn's fork I guess Dec 10 17:29:46 and of course, the right defconfig Dec 10 17:29:56 but I don't see anything else, tbh Dec 10 17:32:23 Well it's the various bin files etc as well Dec 10 17:33:24 we already have that, bl31.elf from ATF and the bin output of course Dec 10 17:34:05 Well yes, but somehow I had other issues... Now my do_configure is getting 100 GB somehow for u-boot :s Dec 10 17:34:42 https://github.com/webOS-ports/meta-pine64-luneos/blob/master/recipes-bsp/u-boot/u-boot_%25.bbappend#L22 and we even have the Crust binaries, maybe not needed on the Pro Dec 10 17:34:58 that's a bit too much! :D Dec 10 17:35:44 Can you share your current u-boot recipe? Dec 10 17:35:49 I dropped the crust already Dec 10 17:36:08 Give me a minute to clean it up Dec 10 17:36:21 (not sure I can help much, though, but worth a try) Dec 10 17:37:33 Well it's more the machine.conf it pulls some stuff from where I'm not sure what to do Dec 10 17:38:29 Stuff like: https://github.com/webOS-ports/meta-pine64-luneos/blob/master/conf/machine/pinephone.conf#L10-L14 Dec 10 17:38:58 And https://github.com/webOS-ports/meta-pine64-luneos/blob/master/conf/machine/pinephone.conf#L36-L46 Dec 10 17:40:12 These are more related to the u-boot bb recipe we're extendind Dec 10 17:40:58 IMAGE_BOOT_FILES will be used to build the wic final image file Dec 10 17:41:13 and it's renaming some files on the fly Dec 10 17:42:09 KERNEL_IMAGETYPE is the output of the kernel recipe, will be used as an input for u-boot boot image Dec 10 17:42:19 https://github.com/webOS-ports/meta-pine64-luneos/commit/9cc9d20efd4be0047d06216bd2b94a8a9da55409 Dec 10 17:42:26 That's my u-boot recipe for now Dec 10 17:43:53 well, looks quite ok to me Dec 10 17:44:12 Yeah to me too Dec 10 17:44:24 But it somehow breaks Dec 10 17:45:35 is the fork very patched, compared to upstream? Dec 10 17:46:09 even so, I don't see how you get so much data generated Dec 10 17:46:56 it's generating only one output u-boot image? Dec 10 17:47:19 I'm not really sure there... Manjaro also has a build with 2 more patches: https://gitlab.manjaro.org/manjaro-arm/packages/core/uboot-pinephonepro/-/blob/master/PKGBUILD Dec 10 17:47:31 But they only added boot order Dec 10 17:50:32 Seems they're using older 2021.01-rc3 u-boot Dec 10 17:50:49 I rebased the regular Pinephone patches in my branch on top of latest release Dec 10 17:51:20 ah, you don't use Martijn's fork, like in pmOS? Dec 10 17:51:28 No I didn't Dec 10 17:52:01 ok; note that for pinephone I don't use a particularly recent u-boot, it's also a fork Dec 10 17:52:08 Ah OK Dec 10 17:52:14 Well I could use theirs I guess Dec 10 17:53:03 it's around 2021.07 for pinephone Dec 10 17:53:04 There was 1 recipe where there were a ton of patches, which I was able to drop by using upstream, but that wasn't u-boot I guess Dec 10 17:54:28 I'd say follow pmOS fork until you get a result (i.e. not 100GB :) ) and then tweak it more Dec 10 17:54:55 Tofe: OK. Well I bumped arm-trusted-fimrware because I could drop a ton of patches Dec 10 17:55:33 that should be fine; the recipe does work, isn't it Dec 10 17:56:04 Yeah or was it something else Dec 10 17:56:07 I now cannot find it Dec 10 17:57:14 you should get a bl31.elf binary deployed, mainly Dec 10 17:57:37 OK Dec 10 17:57:41 Well I got it to build Dec 10 17:57:44 That seems to work Dec 10 17:59:07 Ok good Dec 10 18:00:00 you're pulling meta-arm already? Dec 10 18:00:34 Tofe: No not yet Dec 10 18:00:46 Just copied the necessary recipes for now Dec 10 18:00:57 ah, ok, so it's not that which could pull the 100GB of dependencies Dec 10 18:01:25 This heap of output is quite a mystery Dec 10 18:01:33 Tofe: No... This is the log: https://bpa.st/AWIQ Dec 10 18:01:42 Trimmed the end, because it basically keeps repeating that line Dec 10 18:03:14 Seems some issue with "symbol value '' invalid for SYS_LOAD_ADDR" Dec 10 18:09:11 could it be pmOS is feeding the build with this value somehow Dec 10 18:10:41 I found this: https://patchwork.ozlabs.org/project/uboot/patch/20210823142531.5449-3-trini@konsulko.com/ I wonder where the fork is regarding this migration Dec 10 18:13:21 Well the fork is in 2021.01, this was about half a year later Dec 10 18:16:12 CONFIG_SYS_LOAD_ADDR should be defined by this header then: https://git.sr.ht/~martijnbraam/u-boot/tree/13c8c6945856e57958fdf7e4ef338c21e1199f19/item/include/configs/pinephone-pro-rk3399.h Dec 10 18:16:27 maybe it's in the common .h Dec 10 18:16:37 gotta leave for a bit Dec 10 18:17:30 https://git.sr.ht/~martijnbraam/u-boot/tree/13c8c6945856e57958fdf7e4ef338c21e1199f19/item/include/configs/rk3399_common.h#L20 it's here, weird... Dec 10 18:18:00 Let me set the value simply Dec 10 18:18:15 In the defconfig of uBoot Dec 10 18:18:18 or you can add it to the defconfig Dec 10 18:18:19 ah yes Dec 10 18:18:21 In my fork Dec 10 18:18:26 exactly Dec 10 18:19:30 Solved it Dec 10 18:19:36 Thnx for pointer! Dec 10 18:19:55 Do I need this boot.txt file and where do I get thse values from? Dec 10 18:20:50 yes, boot.txt is needed, it's the script executed at boot by u-boot Dec 10 18:21:14 it's compiled as boot.scr by u-boot Dec 10 18:21:25 OK Dec 10 18:21:51 maybe the same .txt as pinephone would work, just check the names it uses for the kernel and initramfs Dec 10 18:22:06 Yeah I'll play around a bit now Dec 10 18:22:14 ok :) Dec 10 19:56:55 Just not really sure what to do with the wks Dec 10 20:00:49 https://github.com/webOS-ports/meta-pine64-luneos/blob/master/wic/pinephone-luneos-bsp-image.wks#L10 Dec 10 20:00:55 That line specifically Dec 10 20:04:15 Might just be u-boot-rockchip.bin going by https://u-boot.readthedocs.io/en/latest/board/rockchip/rockchip.html Dec 10 20:06:40 Yeah that seems to build Dec 10 20:07:24 OK I have an image now it seems Dec 10 20:08:11 Just need to find sdcard adapter somewhere Dec 10 20:15:35 Found it, question is if this sd card still works, that's another good one ;) Dec 10 20:54:29 :) Dec 10 20:55:31 Tofe: Seems I cannot find WIC on Ubuntu? Dec 10 20:58:43 Could I use dd instead? Dec 10 20:59:24 dd is fine; but don't you have a bmap image? Dec 10 21:00:25 because bmap images are sparse, so you'll have to use bmapcopy Dec 10 21:00:52 Well I'm going by your instructions in meta-pine64-luneos README.md Dec 10 21:01:31 ok, then it's fine :) Dec 10 21:03:39 This is what I have in my images folder: https://bpa.st/MYMA Dec 10 21:05:03 oh, ok, so the wic.gz would actually be a wic.bmap.gz? I don't know Dec 10 21:05:44 you'll see soon enough if the sdcard can be read Dec 10 21:07:51 Inside the .wic.gz I have a .wic Dec 10 21:11:00 Well let's see Dec 10 21:11:13 I would be surprised if it boots on first attempt, but well Dec 10 21:11:17 gunzip will just propose the name without .gz as output; but yes, should be a wic... so why is there a bmap sparse index file besides the rest?... Dec 10 21:11:43 do your system see the two partitions on the sdcard? Dec 10 21:11:44 Well it's a small file with layout info Dec 10 21:11:52 I have one, flashing now Dec 10 21:11:57 Will check after Dec 10 21:12:01 ok Dec 10 21:12:34 luneos-boot and luneos-rootfs my UI shows me as being there Dec 10 21:12:36 So I guess yes Dec 10 21:13:02 ok, looks fine then Dec 10 21:13:07 Yup sde1 and sde2 Dec 10 21:13:09 you used dd ? Dec 10 21:13:13 130MB and 1.8GB Dec 10 21:13:26 bmaptool Dec 10 21:13:27 sudo bmaptool copy luneos-dev-image-pinephonepro.wic.gz /dev/sde Dec 10 21:13:51 did the tool indicate that it was a sparse file? Dec 10 21:14:11 like, 1.8GB written from 1.3GB file Dec 10 21:15:10 I'm just surprised by the wic.gz file extension Dec 10 21:16:37 Something like this yes: https://bpa.st/W7MA Dec 10 21:16:49 Boots to Android, let me erase Android like they suggest on Wiki Dec 10 21:17:23 ok yes, same as I have usually, so maybe it changed recently with hardknott Dec 10 21:18:11 Tofe: Yeah could be... Well this was hackidy hacked together based on your regular Pinephone bits, so I'm sure there are mistakes in various places Dec 10 21:18:13 for the boot, do you have a jack-USB cable? Dec 10 21:18:13 https://wiki.pine64.org/wiki/PinePhone_Pro_Developer_Edition#Nuking_the_factory_AOSP_installation Dec 10 21:18:15 FYI Dec 10 21:18:21 Tofe: ehm no.... Dec 10 21:18:29 And it's USD 13 shipping on USD 7 cable Dec 10 21:18:47 ok, might be a bit tricky to debug, we'll see Dec 10 21:19:30 Yeah let's see Dec 10 21:19:38 I could always put something together like this: https://dimitris.cc/kde/2020/05/01/USB_Serial_Pinephone.html Dec 10 21:20:44 if you feel playful, yes :) Dec 10 21:22:19 I should check my cable box downstairs, might have something, but doubt it Dec 10 21:22:29 I think I have USB-> Old Nokia for unlocking Dec 10 21:24:30 The 10 minute wait for erase is about 1/3 Dec 10 21:24:33 So we'll see Dec 10 21:29:16 not sure why they propose to wipe out the whole eMMC, only the first MB would be enough to erase the GPT Dec 10 21:32:50 Well I now got a "brick", Windows sees it as an "unknown" device which has been successfully setup Dec 10 21:32:54 So I might have terminal? Dec 10 21:35:37 Windows sees "USB\VID_2207&PID_330C\5&1DAC11DF&0&3" which seems to be some Rockchip rom mode Dec 10 21:38:17 during boot? Dec 10 21:39:38 Yes Dec 10 21:39:47 Well it just seems stuck in some state Dec 10 21:39:59 Screen shows nothing Dec 10 21:40:03 But it doesn't seem to bootloop Dec 10 21:40:35 might be a mistake in the boot.txt Dec 10 21:41:19 Well I didn't change anything there except for the filename Dec 10 21:41:24 I'm not sure what to put there Dec 10 21:41:43 or, if initrd is started, maybe you have a telnet on... ehmmm... some subnet I forgot :p Dec 10 21:42:27 https://github.com/webOS-ports/meta-pine64-luneos/blob/master/recipes-core/initrdscripts/initramfs-scripts-simple/init.sh#L19 Dec 10 21:43:03 ah, it's just in recovery mode Dec 10 21:43:30 Well stuff like this will of course never work: https://github.com/webOS-ports/meta-pine64-luneos/blob/master/recipes-bsp/u-boot/files/boot.txt#L11-L14 Dec 10 21:43:36 I need to update that Dec 10 21:44:56 what needs changing? Dec 10 21:45:12 sun50i-a64-pinephone.dtb for sure Dec 10 21:45:22 initramfs-uboot-image-pinephone.uboot -> initramfs-uboot-image-pinephonepro.uboot Dec 10 21:45:34 ah yes, that's part of the filenames :) Dec 10 21:45:48 I'm just not sure what dtb name to use Dec 10 21:45:58 Let me search in my build folder Dec 10 21:49:18 you can also tweak IMAGE_BOOT_FILES to get the right filename on the image Dec 10 21:50:15 I'll leave you here for tonight though, it's +1 here and we'll have to go early tomorrow Dec 10 21:51:14 OK thnx for the pointers Dec 10 21:51:18 Also bedtime here Dec 10 21:52:50 Ah this one might help for rk3399: https://github.com/jakogut/uboot-orangepi-rk3399/blob/master/boot.txt **** ENDING LOGGING AT Sat Dec 11 02:59:56 2021