**** BEGIN LOGGING AT Fri Dec 17 02:59:57 2021
Dec 17 06:36:01 <eriki73[m]> <JaMa> "eriki73: Herrie: Tofe: the..." <- Fetching is fine now but I am out of memory. Is there a way to reduce threads in make?
Dec 17 06:36:13 <HerrieTP_> JaMa: OK thnx for the info. Good to see that stuff seems to be getting up again
Dec 17 06:41:13 <HerrieTP> Only thing that gets out of memory is qtwebengine usually
Dec 17 06:41:49 <HerrieTP> You could try to just build qtwebengine itself with DEVICE=sargo bb qtwebengine
Dec 17 06:43:32 <eriki73[m]> HerrieTP: I tried running bitbake with only one thread, but make still runs with 16 threads and causes my machine to go out of mem.
Dec 17 07:10:43 <Herrie> You can add swap space or change the threads, see: https://webos-ports.org/wiki/Build_for_Tenderloin#Adding_Swap_Space
Dec 17 09:00:08 <Tofe> eriki73[m]: you can tweak that in local.conf, iirc
Dec 17 09:00:50 <Tofe> I have 16Gb ram, and limit it to 4 threads x 4 bitbake processes
Dec 17 09:03:29 <Tofe> should be PARALLEL_MAKE variable in your case
Dec 17 09:36:00 <Tofe> Herrie: you have.. files in the wic image ? is that right ? not partitions + content of images ?
Dec 17 09:53:05 <Herrie> Tofe: I have 7 .img files inside the WIC, corresponding to the partitions I guess. Then inside the .img files I have a regular file and directory structure
Dec 17 10:03:54 <Herrie> Seeing it boots uBoot, with recent kernel and ATF 2.6 that part is all OK I would say. It's somewhere in the config for the initramfs and the switch to it where it fails
Dec 17 10:11:45 <Tofe> ok
Dec 17 10:16:17 <HerrieTP> The issue could be in my "fitImage" which could/should contain kernel, devicetree and initramfs from what I understood
Dec 17 10:16:53 <HerrieTP> I've never really dealt with fitImage before to be honest, just read a bit about it
Dec 17 10:16:58 <Tofe> Herrie: ok, I didn't understand that Yocto can use these many variables :) the boot.txt is kind of unnecessary, with all this [except if we want to boot recovery when detecting VolDown+Power, but we'll see later]
Dec 17 10:17:11 <HerrieTP> I'll see if I can find some more info
Dec 17 10:18:24 <Herrie> Tofe: Yeah you can do quite a bit in Yocto :)
Dec 17 10:18:27 <Tofe> Herrie: or you can UBOOT_EXTLINUX_FDT + UBOOT_EXTLINUX_INITRD + UBOOT_EXTLINUX_KERNEL_IMAGE , should work too
Dec 17 10:19:20 <Tofe> kernel image can be a zImage also
Dec 17 10:19:30 <Tofe> seems it takes KERNEL_IMAGETYPE
Dec 17 10:20:11 <Herrie> Tofe: Yeah well separate images is always an option, but it seems fitImage was designed to get rid of it... But seems there's something wrong with mine I guess: https://bpa.st/LERQ
Dec 17 10:20:43 <Herrie> If I see this: https://lists.denx.de/pipermail/u-boot/2017-March/283284.html there should be a ramdisk listed as well
Dec 17 10:21:19 <Herrie> https://www.marcusfolkesson.se/blog/fit-vs-legacy-image-format/ some more info about fitImage
Dec 17 10:27:53 <Herrie> Seems I simply might need to add INITRAMFS_IMAGE_BUNDLE=1 to my .conf ;)
Dec 17 10:32:00 <Herrie> OK I'm getting a failure now that it cannot find hte file, so seems the flag is doing something
Dec 17 10:41:21 <Herrie> JaMa: Morning!
Dec 17 10:47:44 <Tofe> Herrie: I guess you have to specify the file name somewhere
Dec 17 10:50:51 <Herrie> Tofe: Yeah, just getting circular dependency stuff now
Dec 17 10:50:54 <Herrie> So trying to work it out
Dec 17 10:51:04 <Herrie> Seems I'm getting there
Dec 17 10:51:23 <Tofe> looks like you're on the right path yes
Dec 17 10:54:14 <Herrie> mkimage doesn't list my initramfs but my kernel turned 24MB, so I guess it's in there
Dec 17 10:54:49 <Tofe> looks about right
Dec 17 10:54:57 <Herrie> OK let's download and flash again
Dec 17 10:55:12 <Herrie> After it boots need to rework stuff a bit so both PP and PPP will work
Dec 17 10:55:21 <Herrie> We might  be able to use fitImage on PP as well BTW
Dec 17 10:55:45 <Tofe> Probably yes
Dec 17 10:56:31 <Tofe> And if we can reduce the gap between the two variants of the phone, it's always better
Dec 17 10:57:08 <Tofe> I wonder if my recovery UI stuff works as-is on PPP
Dec 17 10:57:38 <Tofe> some paths were a bit hardcoded :/
Dec 17 10:57:51 <Herrie> Tofe: Well some of the meta-rockchip stuff seems nicer vs current PP bits, so yes
Dec 17 10:59:15 <Tofe> https://github.com/Tofee/luneos-recovery-ui/blob/master/recovery/ui/actions.c#L11 looks like it can be overriden with a -D define the recipe, if needed
Dec 17 10:59:55 <Tofe> (anyway it's far from rocket science C code here :p )
Dec 17 11:05:07 <Herrie> Tofe: Yeah
Dec 17 11:05:17 <Herrie> Still the same, but seems I changed init in my wks
Dec 17 11:05:19 <Herrie> So I'll change it back
Dec 17 11:21:58 <Herrie> And now utter failure :D
Dec 17 11:22:14 <Herrie> Seems it now just reboots to bootloader all the time
Dec 17 11:31:43 <HerrieTP> Could be I need some more defconfig options, will check
Dec 17 11:38:18 <Tofe> Herrie: you don't even get logs?
Dec 17 11:38:45 <HerrieTP> No just the BL31 on repeat
Dec 17 11:39:31 <HerrieTP> So log, but no more uBoot kernel
Dec 17 11:39:59 <HerrieTP> Seems there's a few defconfig flags, also for size etc that I need to set
Dec 17 11:44:49 <Tofe> but it used to work with the same config, right? could it be something didn't update properly?
Dec 17 12:47:17 <Herrie> Tofe: Yeah it worked before, just now I changed the init line in wks
Dec 17 12:47:19 <Herrie> Last line
Dec 17 12:48:47 <Herrie> So could be bad flash somehow
Dec 17 12:58:22 <Herrie> Well there's things like CONFIG_FIT=y and CONFIG_SPL_LOAD_FIT=y that might be needed
Dec 17 13:07:33 <Herrie> Testing those now
Dec 17 13:11:46 <Herrie> Seems the same... Tofe: I shouldn't need to deploy the initramfs anymore right? SInce this is in the fitImage, or I misunderstood?
Dec 17 13:29:42 <Tofe> maybe misunderstood, in the sense that the fitimage will pick it from the deploy folder
Dec 17 13:29:50 <Tofe> At least I think so
Dec 17 13:34:41 <HerrieTP_> Tofe: I mean I shouldn't need to add it to wks separately anymore like you did for PP
Dec 17 13:34:54 <HerrieTP_> Because it's part of fitImage
Dec 17 13:35:11 <HerrieTP_> At least that is my understanding
Dec 17 13:47:57 <Tofe> I agree
Dec 17 16:43:59 <eriki73[m]> <Tofe> "should be PARALLEL_MAKE variable..." <- In local.conf?
Dec 17 16:45:02 <Tofe> yes
Dec 17 16:45:17 <Tofe> with a value lilke "-j 4" iirc
Dec 17 16:45:57 <eriki73[m]> Hmm, do I need to clean something? I did put it with -j4 but it didn't seem to apply.
Dec 17 17:04:47 <Herrie> I think there are 2 local.conf files
Dec 17 17:04:52 <Herrie> I also get confused
Dec 17 17:04:59 <Herrie> You might want to re-run .setup-env
Dec 17 17:13:33 <eriki73[m]> I didn't find any other local.conf under webos-ports and I re-run . setup-env without any effect.
Dec 17 17:17:07 <Herrie> If you go one level up from webos-ports there should be a common folder with a local.conf as well
Dec 17 17:17:23 <Herrie> I have this: OE @luneos ~/LuneOS/hardknott $ grep -r PARALLEL_MAKE
Dec 17 17:17:23 <Herrie> common/conf/local.conf:# PARALLEL_MAKE = "-j 4"
Dec 17 17:17:23 <Herrie> webos-ports/conf/local.conf:# PARALLEL_MAKE = "-j 4"
Dec 17 17:19:41 <eriki73[m]> Ok, I added it there but still no effect.
Dec 17 17:20:02 <Herrie> Hmmz weird
Dec 17 17:20:11 <Herrie> Sometimes helps to close console, re-open one
Dec 17 17:20:20 <Herrie> Then do . setup-env again
Dec 17 17:20:27 <Herrie> Weird but sometimes like that
Dec 17 17:20:46 <eriki73[m]> I tried that with rebooting already.
Dec 17 17:22:00 <eriki73[m]> And my terminal closes every time I run out of RAM. And when monitoring with system monitor, there are 16 cc1plus processes.
Dec 17 17:26:41 <eriki73[m]> Weirdly enough my laptop seems to dectect PARALLEL_MAKE. Its specs are Intel(R) Core(TM) i7-10510U and 16GB of RAM. It is going to take a bit longer though.
Dec 17 17:50:24 <Herrie> That's weird but well
Dec 17 18:11:21 <JaMa> webos-ports/conf/local.conf is the only one read by bitbake
Dec 17 18:11:27 <JaMa> re-running setup-env isn't needed
Dec 17 18:11:53 <JaMa> use bitbake -e qtwebengine | grep -B 10 ^PARALLEL_MAKE
Dec 17 18:12:00 <JaMa> to see where the value got set for you
Dec 17 18:25:17 <Herrie> JaMa: You happen to have any experience with fitImage, initramfs and Rockchip by any chance?
Dec 17 18:25:32 <Herrie> I'm a bit stuck and seems documentation is lacking ;)
Dec 17 18:50:19 <eriki73[m]> JaMa: hmm, it says -j1 which it the one in my local.conf but building still spawns 16 cc1plus processes.
Dec 17 19:38:56 <JaMa> eriki73[m]: are all cc1plus from qtwebengin?
Dec 17 19:39:05 <JaMa> that's very strange
Dec 17 19:40:34 <JaMa> Herrie: some experience yes, but nothing directly applicable to what I've read here about ppp
Dec 17 19:42:07 <JaMa> that reminds me that I haven't read december update from p64 yet
Dec 17 20:41:25 <HerrieTP_> JaMa: OK
Dec 17 20:41:42 <HerrieTP_> trever suggested to try with meta-rockchip layer just
Dec 17 20:41:51 <HerrieTP_> Which isn't a bad idea I guess
Dec 17 20:42:17 <HerrieTP_> And just adopt one of the machines from there and use linux-yocto-dev
Dec 17 20:43:21 <HerrieTP_> Shouldn't be a lot of work and I can re-do it with honister 
Dec 17 20:43:38 <HerrieTP_> Just to exclude some issues in the build at my end
Dec 17 20:57:04 <Herrie> I could even go for kirkstone/master with bleeding edge stuff
Dec 17 20:57:14 <Herrie> Might be closer to my current frankenbuild
Dec 17 23:01:20 <Herrie> Tofe: FWIW I came across this one for sunxi/Allwinner: https://github.com/linux-sunxi/meta-sunxi
**** ENDING LOGGING AT Sat Dec 18 02:59:56 2021