01:11 <KGB-1> https://tests.reproducible-builds.org/openwrt/openwrt_x86.html has been updated. (100.0% images and 99.9% packages reproducible in our current test framework.)

08:35 <KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_omap.html has been updated. (11.1% images and 99.8% packages reproducible in our current test framework.)

10:10 <robimarko> rmilecki: Would it be possible to extend the U-Boot NVMEM parser so its able to parse MAC-s from custom variable names as well?

10:11 <robimarko> I am seeing a decent number of devices that have MAC-s stored under non-standard naming so simply matching for "ethaddr" to post-process as its in ASCII is not enough

11:48 <SpectreDev_01> 10.729864] hsl_phy_phydev_get[798]:ERROR:phy_addr 5 phydev is NULL

11:48 <SpectreDev_01> robimarko: I was able to get smem to parser 27 of the partition I gues ubl has to be hardcoded and I labelled ethernet port but only eth0 shows up and none of them seem to work

11:48 <SpectreDev_01> [ 11.779602] hsl_phy_phydev_get[798]:ERROR:phy_addr 5 phydev is NULL

11:48 <SpectreDev_01> [ 12.809616] hsl_phy_phydev_get[798]:ERROR:phy_addr 5 phydev is NULL

11:48 <SpectreDev_01> [ 13.839608] br-lan: port 1(eth0) entered disabled state

11:48 <SpectreDev_01> [ 13.839564] br-lan: port 1(eth0) entered blocking state

11:48 <SpectreDev_01> [ 13.843828] device eth0 entered promiscuous mode

11:50 <SpectreDev_01> Oh wait it's not ubl, it's alt_kernel and alt_rootfs that's missing, I guess that's not needed

11:50 <SpectreDev_01> Oh wait it's not ubl, it's alt_kernel and alt_rootfs that's missing, I guess that's not needed

12:18 <SpectreDev_01> Hmm how do we get WiFi to work

12:39 <SpectreDev_01> Ah damn Linksys use bdwlan.bin based on region

13:12 <robimarko> You need to add the default network config in 02_network

13:12 <robimarko> Its under base-files/etc in qualcomax

13:12 <robimarko> Otherwise they are down by default

13:12 <robimarko> Also, you did not label them if you are seeing eth0

13:16 <SpectreDev_01> robimarko: I used eth0, eth1 etc for labeling like stock does, do you want me to change it to wan, lan1 etc

13:16 <robimarko> Yeah, cause for ethN you dont need to label at all

13:17 <SpectreDev_01> Ah I see

15:41 <SpectreDev_01> robimarko: well ethernet works now all that's left is wlan ig

15:42 <SpectreDev_01> Also web interface is missing, unless it's because I missed adding luci

16:59 <SpectreDev_01> How do I extract the bdwlan.bin to create board-2.bin

17:02 <stintel> https://github.com/qca/qca-swiss-army-knife/tree/master

17:03 <stintel> documentation is meh. I vaguely remember the scripts help options being more helpful.

17:07 <minimal> I just happened to notice that OpenWRT (23.05.0-rc3) is still using MD5 for passwords and am really surprised. I'd assumed it would be using SHA512.

17:09 <SpectreDev_01> stintel: but there's fw extractor for Ath11k from what I can see

17:10 <Znevna> minimal: I'm really surprised you didn't read the forum topics about this subject

17:13 <robimarko> SpectreDev_01: You dont extract it, you wrap it with the ath11k-bdencoder

17:13 <SpectreDev_01> Oh ok

17:14 <minimal> Znevna: haven't search the forum about it yet

17:14 <robimarko> SpectreDev_01: Easiest way being you take one of the BDF-s from the OpenWrt BDF repo

17:14 <robimarko> Unpack it with the -e option, delete the .bin

17:15 <robimarko> And then modify board-2.json for your board

17:15 <robimarko> Then you can just wrap yours with -c

17:24 <minimal> Znevna: are you referring to the "Passwd -a sha256|sha512 less secure than def..." discussion? If so well Alpine also used musl, dropbear, openssh, etc and used sha512 by default so they've managed to address an issues. The only different with OpenWRT would likely be luci-related

17:35 <SpectreDev_01> Guys how do I find the board ID etx

18:10 <SpectreDev_01> Alright I created board-2.bin how do I turn that into .ipq file

18:31 <Lechu> Anybody using Debian 12 for builds? I've stumbled upon strange build error with uqmi:

18:31 <Lechu> https://gist.github.com/Leo-PL/775d02a435fd0ff4cd831f1feb7de479, TL;DR CMake's compiler test fails for some strange reason involving unsupported warning flags

18:34 <Lechu> Back at Debian 11 it seems to have worked. This is using today's main.

19:22 <KanjiMonster> Lechu: are you sure this is related to debian 12? Looks more related to using GCC 11 instead of 12 (the default)

19:54 <SpectreDev_01> How do I find the Gpio pins used for WiFi

21:21 <SpectreDev_01> Got WiFi working

21:47 <Lechu> KanjiMonster: tried on another machine, with freshly installed Debian 12 again, and it worked. Did "make distclean" and updated toolchain according to the wiki, let's see now.

21:50 <Lechu> I wonder why gcc-multilib is among the recommended packages, this caused my mips and armhf cros-toolchains to be removed

22:05 <slh> Lechu: you probably won't need it

22:06 <slh> at least my tpyical package set can do easily without multilb

22:15 <Lechu> I'll restore those cross-toolchains then, I actually use them for some stuff on this machine, unlike gcc-multilib.

22:18 <slh> unless you need nodejs packages, gcc-multilib won't be necessary

22:59 <Lechu> Gaah, indeed, 'twas the case. GCC 11 selected in .config.

23:32 <Lechu> Not sure how those options snuck into my defconfigs for development builds, but now I nuked them, so this should work again.