00:24 <Mangix> I do wonder if it's possible to use git magic to automate this

00:26 <Mangix> eg: check for no longer existing folders one level down and explicitly revert.

00:45 <dwfreed> Habbie: fwiw, I think the comment about PKG_SOURCE was about it not being very unique; PKG_SOURCE is a destination filename (what the URL will be downloaded to), rather than a "source" filename (re https://github.com/openwrt/packages-abandoned/pull/39#issuecomment-1989003717 )

00:47 <dwfreed> it "works" because it can be anything; but when you're downloading all sources, it should be sufficiently unique (by including package name, for example) to definitely not collide with any other package's source

01:10 <Mangix> triied asking ChatGPT how to do this, didn't quite get the proper result

01:25 <damo22> what flag am i missing for lzma-loader? i got "error: Unsupported SoC" make -w -C lzma-loader PKG_BUILD_DIR="/home/damien/git/openwrt/build_dir/target-mipsel_24kc_musl/linux-ramips_mt76x8/tmp/openwrt-ramips-mt76x8-tplink_archer-mr200-v6-initramfs-kernel.bin.src" TARGET_DIR="/home/damien/git/openwrt/build_dir/target-mipsel_24kc_musl/linux-ramips_mt76x8/tmp/" LOADER_NAME="openwrt-ramips-mt76x8-tplink_archer-mr200-v6-initramfs-kerne

01:25 <damo22> l.bin" BOARD="MT76x8 based boards" PLATFORM="ralink" LZMA_TEXT_START=0x81800000 LOADADDR=0x80000000 SUBTARGET=mt76x8 LOADER_DATA="/home/damien/git/openwrt/build_dir/target-mipsel_24kc_musl/linux-ramips_mt76x8/tmp/openwrt-ramips-mt76x8-tplink_archer-mr200-v6-initramfs-kernel.bin" compile loader.bin

01:31 <damo22> looks like lzma-loader does not support mt76x8 ?

03:16 <Mangix> wonder if https://forum.openwrt.org/t/openwrt-18-06-build-system-port-to-cygwin/38436 should be modetnized

03:17 <Mangix> it seems there's already some support already

05:51 <damo22> is it safe to compile u-boot and flash it over the vendor uboot?

05:51 <damo22> im having trouble with lzma

05:52 <damo22> i ported lzma-loader to mt76x8

05:53 <damo22> i still get LZMA ERROR 1

05:56 <damo22> isnt the vmlinux already a xz image, and the initramfs is lzma compressed, so what is the point of doing kernel-bin | append-dtb | lzma ?

06:38 <damo22> ok, i have a working initramfs-kernel version

06:38 <damo22> how do i test completing the port?

06:38 <damo22> ive backed up the full flash

06:47 <damo22> all buttons and leds work

06:48 <damo22> and the wwan/wlan are detected in the kernel log

06:55 <damo22> https://zamaudio.com/mbox2/booted.txt

08:31 <PaulFertser> damo22: when everything works in initrams the next logical step is to try to "sysupgrade" using the sysupgrade image from within it.

08:31 <PaulFertser> damo22: congrats on the progress!

08:33 <PaulFertser> damo22: so how did you workaround the lzma error?

08:34 <PaulFertser> damo22: vmlinu_x_ isn't compressed, vmlinu_z_ is, and initramfs there is part of vmlinux, linked in.

08:36 <damo22> i removed the lzma compression so my initramfs-kernel image is 9MB instead of 5MB

08:38 <damo22> i tried with gzip instead but it failed again

08:40 <damo22> i think the sysupgrade may fail because i know the vendor header isnt supported by mktpinkfw2 tool

08:40 <PaulFertser> damo22: can you check how the vendor image (that starts at firmware) differs from what you generate?

08:41 <PaulFertser> damo22: you need to carefully compare what you generate with what's already in firmware partition.

08:41 <damo22> https://github.com/openwrt/openwrt/pull/15610 i sent this so others can collaborate

08:41 <PaulFertser> The part up to the rootfs is essential.

08:42 <PaulFertser> damo22: ok, you're on the right path

08:42 <damo22> the tplink header is at the top of the firmware

08:43 <PaulFertser> Is uImage following? What type of compression does it use?

08:43 <damo22> not sure, i think its uncompressed

08:44 <PaulFertser> You can check by extracting it with binwalk -e and then using mkimage -l

08:45 <damo22> 512 0x200 LZMA compressed data, properties: 0x5D, dictionary size: 8388608 bytes, uncompressed size: 3738100 bytes

08:45 <damo22> 2097152 0x200000 Squashfs filesystem, little endian, version 4.0, compression:xz, size: 8759472 bytes, 689 inodes, blocksize: 65536 bytes, created: 2023-05-16 09:58:34

08:45 <PaulFertser> To figure out the header if it's really different from what's already supported you might want to grab several different versions of this firmware from vendor updates, unpack and see what bytes are different.

08:46 <PaulFertser> damo22: and no uImage header?

08:46 <damo22> thats starting at 0x20000

08:46 <damo22> i dont know what a uImage header looks like

08:46 <PaulFertser> I wonder if your lzma compressed image also has properties: 0x5D

08:47 <damo22> i could recompile one and see

08:48 <PaulFertser> damo22: did you add full log (from startup) from vendor booting somewhere?

08:49 <PaulFertser> damo22: you can create uImage header yourself with "mkimage" tool, if vendor's u-boot isn't talking about FIT then it's probably a "legacy uImage".

08:49 <PaulFertser> But binwalk should have detected it.

08:49 <PaulFertser> damo22: have you downloaded GPL sources for u-boot on this board?

08:51 <damo22> i cant find GPL sources for this version of the board

08:51 <damo22> 82720 0x14320 U-Boot version string, "U-Boot 1.1.3 (May 16 2023 - 17:33:33)"

08:55 <damo22> https://zamaudio.com/mbox2/mr200-vendor.txt

09:16 <Habbie> dwfreed, right, i figured it was something like that, but i don't do guessing games and this person only had guessing games for me :)

09:19 <PaulFertser> damo22: you should probably try requesting it from them, they must provide it

09:24 <PaulFertser> damo22: it's quite possible tplink is doing some odd thing and using their own custom header instead of uImage or FIT header there. So to guess what might be important in that new header type you might need few examples to compare.

21:20 <cation> there are some devices that are being switched to use `compatible = "u-boot,env";` in their DTS files to set which partition has the u-boot environment parameters... some devices have second u-boot-env partitions (searching "u-boot-env2" in the openwrt repo finds these), is it possible to configure these second partitions in the same way (with something like `compatible = "u-boot,sys";`)?

21:22 <cation> or would you have to keep using "ubootenv_add_uci_sys_config" in package/boot/uboot-envtools/files/*?

22:27 <owrt-images-builds> Build [#241](https://buildbot.openwrt.org/images/#/builders/220/builds/241) of `master_mpc85xx/p2020` failed.

22:39 <Mangix> hmm minidlna is super dead

22:39 <Mangix> unfortunate