<owrt-snap-builds> Build [#474](https://buildbot.openwrt.org/master/images/#builders/27/builds/474) of `mpc85xx/p1020` completed successfully.
mangix has quit [Server closed connection]
mangix has joined #openwrt-devel
Ansuel has joined #openwrt-devel
Ansuel has quit []
rua has quit [Ping timeout: 480 seconds]
ecloud has quit [Ping timeout: 480 seconds]
ecloud has joined #openwrt-devel
rua has joined #openwrt-devel
danitool has quit [Ping timeout: 480 seconds]
ZorPrime has joined #openwrt-devel
<ZorPrime> hello. Trying to compile the tools and tool chain for arm_cortex-a7_neon. Going by the "Build system usage" wiki page I did make menuconfig and set my target arch and platform. The tools build fails on libelf complaining 'checking whether overlapping arrays are copied correctly... yes
<ZorPrime> ugh sorry
<ZorPrime> one min - have this in a pastebin
<KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_kirkwood.html has been updated. (100.0% images and 98.3% packages reproducible in our current test framework.)
<ZorPrime> https://pastebin.com/NezqNaki --- not sure what else I'm missing on specifying I'm trying to build for arm and not x64_86
minimal has quit [Quit: Leaving]
<slh> you need to configure your target (make menuconfig) first, ideally also make a small full build for your target/ device (no need to flash it, just to have the basics built)
<ZorPrime> slh: yeh I did make menuconfig first, and saved the .config file. I didn't do the full build. Will give that a go
<ZorPrime> hmmm something funky with the make script somewhere. Turned up verbosity (-j1 V=s) and when it gets to libelf the target has not been changed ----- bash ./configure --target=x86_64-pc-linux-musl --host=x86_64-pc-linux-musl --build=x86_64-pc-linux-musl
Slimey has quit [Server closed connection]
Slimey has joined #openwrt-devel
<Slimey> ew ID 13b1:003a Linksys AE2500 802.11abgn Wireless Adapter [Broadcom BCM43236]
<owrt-snap-builds> Build [#475](https://buildbot.openwrt.org/master/images/#builders/15/builds/475) of `armvirt/64` failed.
Acinonyx_ has quit [Quit: Peer reset by connector]
philipp64 has quit [Server closed connection]
noahm has quit [Server closed connection]
noahm has joined #openwrt-devel
shibboleth has quit [Quit: shibboleth]
yolo has quit [Server closed connection]
yolo has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<mangix> stintel: as long as nbd uses macOS, support will never die
<mangix> aparcar[m]: https://git.openwrt.org/?p=project/firmware-utils.git;a=commit;h=497726b45263ba6f652724213d14084e89e4f970 broke it
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<owrt-snap-builds> Build [#464](https://buildbot.openwrt.org/master/images/#builders/35/builds/464) of `mvebu/cortexa72` completed successfully.
GNUmoon has quit [Ping timeout: 480 seconds]
philipp64 has joined #openwrt-devel
Misanthropos has quit [Ping timeout: 480 seconds]
<mangix> ynezz: do you run firmware-utils with sanitizers? I remember you doing that with some other packages
<ynezz> mangix: that would be next step, first is usually fix static analyzers warnings https://openwrt.gitlab.io/-/project/firmware-utils/-/jobs/2140706116/artifacts/build/scan/2022-02-26-233926-89-1/index.html
<ynezz> and this is a mine field, so before starting fixing those, there should be some unit tests in order to assure same output
<ynezz> in other words, a lot of work
<mangix> ynezz: I downloaded it as it seems to be failing on macOS. Running gcc's -fanalyzer is interesting. It mostly found missing malloc/fopen NULL checks
<mangix> not sure if worth fixing.
<ynezz> it is always worth fixing
<ynezz> I've tried gcc's static analyzer as well, but it has produced way more false positives then clang's
<mangix> hmm OK I'll submit
<ynezz> mangix: https://git.openwrt.org/?p=openwrt/staging/ynezz.git;a=shortlog;h=refs/heads/wip/firmware-utils-fixes
<ynezz> ldir has some more as well
Misanthropos has joined #openwrt-devel
<ynezz> regarding macOS we should probably consider one buildworker for that purpose
<mangix> ynezz: weeeelll, since firmware-utils was migrated to an external git repo, we can't use the github actions in openwrt.
<owrt-snap-builds> Build [#471](https://buildbot.openwrt.org/master/images/#builders/25/builds/471) of `rockchip/armv8` completed successfully.
goliath has joined #openwrt-devel
<ynezz> mangix: repository location doesn't matter, you can always mirror it to some other place and do whatever you want
<ynezz> mangix: there was 2 reasons I'm aware of: 1. make it stand alone so it's reusable in other projects 2. make the QA easier
<mangix> ynezz: I'm talking about the actions in openwrt's github. I *could* create a temporary PR just to test compilation.
<ynezz> I think, that if it was already decided, that there should be everything on GH, then we should make git.openwrt.org mirror, and make GH master
<ynezz> then move all projects to GH, add CI and start using PR workflow
<ynezz> so that CI builds/check shoulw run in project/firmware-utils.git, not in openwrt.git
<ynezz> is it only myself who finds it insane to build complete toolchain on every push to the PR?
goliath has quit [Quit: SIGSEGV]
<ynezz> ok, that's not the case in firmware-utils, since it's using host build, but you get the picture
<mangix> ynezz: sort of. currently the CI builds ALL tools, not just the changed ones. aparcar needs to fix that.
<mangix> I agree with your mirror idea.
felix has quit []
felix has joined #openwrt-devel
rmilecki has quit [Remote host closed the connection]
aep has joined #openwrt-devel
<aep> any idea if kill_remaining from sysupgrade can be removed safely? this board needs a background script for an external watchdog that gets killed
<mangix> aparcar[m]: CI running: https://github.com/openwrt/openwrt/pull/9356
<ynezz> aep: removing it is likely going to cause issues in a later sysupgrade steps
kenny has quit []
<ynezz> moving it into proper kernel watchdog driver should be less PITA
<aep> i couldnt get the kernel driver to work. it resets anyway.
<aep> it's a toggle with a specific timing :/
<ynezz> it's a gpio toggle?
<ynezz> like some pattern or such?
<aep> yeah, gpio. its just 1/0 toggle, but the timing is important
<aep> couldnt figure out how to make the kernel do that
<ynezz> there is gpio_wdt
<aep> yup, but you can't set the timing
<ynezz> you've source code, you can do everything
<aep> right :)
<ynezz> question is if upstream is going to accept that
<aep> i'm not upstreaming any fixes anymore since i invested a month into a good patch and it's still sitting there ignored years later
<aep> so *shrug*
<aep> oh you mean to lkml. yeah maybe. not sure
<ynezz> whatever `./scripts/get_maintainer.pl drivers/watchdog/gpio_wdt.c` outputs
<aep> might be worth a try
<ynezz> anyway, if it's just that delay value, it might be just a config option in DTS
<aep> doesnt seem to be documented, and from a first glance at the driver i couldnt even figure out where the delay is
rmilecki has joined #openwrt-devel
<aep> nah i dont think thats it
<aep> pulse is different from toggle
<ynezz> what exactly do you need then?
<aep> it looks like userspace sets the timing?
<ynezz> no
<aep> the thing needs to be toggled exactly every 1second
<aep> from the driver, i can't seem to find where it sets the timing of the toggle mode at all
<rsalvaterra> philipp64: I'm not really comfortable for reviewing it, but it looks good to me, so ack. :)
dwfreed has quit [Quit: ZNC - http://znc.in]
dwfreed has joined #openwrt-devel
<aep> yeah not sure what "hw_margin_ms" does. it just forwarded to wdd.max_hw_heartbeat_ms
<aep> oh looks like thats exactly what i need?
Acinonyx has joined #openwrt-devel
<ynezz> those structs are described in include/linux/watchdog.h
<ynezz> there is bunch of min/max values, so if it's not enough, you could simply add support for more in gpio_wdt
<aep> i think its the correct parameter. thanks
Acinonyx has quit [Quit: Peer reset by connector]
goliath has joined #openwrt-devel
Acinonyx has joined #openwrt-devel
danitool has joined #openwrt-devel
<aep> hm can't get the kernel driver in anymore. it's just make kernel_menuconfig and then select the thing and make the image again, right?
f00b4r0 has joined #openwrt-devel
noltari has quit [Remote host closed the connection]
bluew has quit [Quit: Leaving]
noltari has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<aep> ynezz: got it working, thanks! off to the ignore pile :) https://github.com/openwrt/openwrt/pull/9358
Acinonyx has quit [Ping timeout: 480 seconds]
Acinonyx has joined #openwrt-devel
Acinonyx has quit [Ping timeout: 480 seconds]
Acinonyx has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
* f00b4r0 hijacks #3818
<aparcar[m]> jow: do you have a more or less working version of ucode on macOS?
rua has quit [Ping timeout: 480 seconds]
<f00b4r0> odd. Why does mt76 gets built on ath79?
<f00b4r0> ha. rtl8812 too.
<f00b4r0> pkg_hash_check_unresolved: cannot find dependency luci-lib-base for luci-base
rua has joined #openwrt-devel
GNUmoon has joined #openwrt-devel
<jow> aparcar[m]: yes, I have one that fails the testsuite but can run basic code
<karlp> f00b4r0: because it's used in usb wifi dongles?
<f00b4r0> karlp: ah, i see
aep has quit [Quit: WeeChat 3.4]
shibboleth has joined #openwrt-devel
torv has joined #openwrt-devel
shibboleth has quit [Quit: shibboleth]
<owrt-snap-builds> Build [#550](https://buildbot.openwrt.org/master/images/#builders/2/builds/550) of `layerscape/armv7` completed successfully.
<koukoutsi> hi all, when testing a PR what the best way to have you custom build/config done and then all packages as modules just in case? is is just having two .config files (one with the extra "build all modules etc" settings?)?
<koukoutsi> or maybe there is a quick way to do it in one go? (buildbot maybe??)
<koukoutsi> (buildbot locally I mean)
<ynezz> you can start with config as used on the buildbots https://downloads.openwrt.org/snapshots/targets/x86/64/config.buildinfo
<koukoutsi> so manage two configs. one for your image and a simple one with just the "build all modules" option eh?
<ynezz> you can try with one config and run `make CONFIG_BUILDBOT=y`
<koukoutsi> my personal image has a lot of options build inside
<koukoutsi> aha!
<koukoutsi> ynezz: so "make CONFIG_BUILDBOT=y" with build the default "menucofig" options or all the modules too?
user95 has quit [Ping timeout: 480 seconds]
<ynezz> well, ideally you should've some diffconfig produced by `./scripts/diffconfig > myconfig`; then you should just `cp myconfig .config; printf 'CONFIG_BUILDBOT=y\n' >> .config; make defconfig` to populate it for buildbot check
<koukoutsi> ok thanks!
user95 has joined #openwrt-devel
<f00b4r0> wasn't there a copy of the QCA9533 datasheet publicly available somewhere? I'm trying to lookup the gpio multiplexer config
mirko has quit [Ping timeout: 480 seconds]
<f00b4r0> awesome, thanks!
aiyion has quit [Remote host closed the connection]
aiyion has joined #openwrt-devel
<koukoutsi> is there documentation of the builtroot vs openwrt differences or specific things added ?
<rmilecki> koukoutsi: no, it's rather messy
<rmilecki> koukoutsi: I think OpenWrt is based on old Linksys's fork of buildroot
<rmilecki> and meanwhile we picked some recent upstream buildroot changes
<rmilecki> and introduced our changes
Misanthropos has quit [Ping timeout: 480 seconds]
<jow> current openwrt buildroot and buildroot.net only share the name
<jow> by now OpenWrt buildroot is more or less proprietary
<f00b4r0> ok, rb mapl-2nd ath79 support done. time to open PR
<koukoutsi> rmilecki, jow : thank got it! I did check the offical buildroot but looked a lot different so I asked
romany has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<stintel> can I convince the gh CI to build with CONFIG_DWARVES=y ?
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
srslypascal is now known as Guest784
srslypascal has joined #openwrt-devel
Guest784 has quit [Ping timeout: 480 seconds]
<ZorPrime> hmmm... would anyone else might happen to know why the libelf automake/autoreconf insists on building for x86_64-pc-linux-musl instead of the target I specified in 'make menuconfig'?
<ZorPrime> I can only assume I missed something in the directions, but haven't found any other gotchas
<stintel> that's probably HostBuild ?
<ZorPrime> this is my first shot at doing cross-compiling so not sure.
<ZorPrime> wish my CS classes did more embedded stuff :P
<ZorPrime> stintel: I'm following the directions at the 'Build a single package' wiki page
<stintel> to be able to build a single package, you still need to build the toolchain and a bunch of other stuff first
<stintel> and tools etc
<ZorPrime> yep I get that, doing make tools/install goes swimmingly well until libelf
<stintel> ok, make tools/install builds tools for the host, not for the target, so if you are building on x86 that's why you are seeing x86_64-pc-linux-musl there
<stintel> it's probably best you pastebin the full output of your failure
<ZorPrime> one min please
<stintel> build with V=sc for verbose output
<stintel> or BUILD_LOG=1 and get the log from logs/... then
<ynezz> stintel: just modify that tools.yml and add the commands you need
goliath has quit [Quit: SIGSEGV]
<ZorPrime> ok yeh so I am building on a laptop loaded with alpine, targeting ipq40xx generic. I'll have that pastebin in another min or so
<stintel> ynezz: I booted an ubuntu vm to repro what ansuel reported wrt to tools/dwarves
<stintel> turns out I cannot use HOST_BUILD_DEPENDS PKG_BUILD_DEPENDS in tools/
<stintel> no idea how to fix this chicken or egg problem - dwarves requires elfutils/host
<stintel> so the only thing I can think of is move dwarves to package/utils/... but that feels wrong if it's only going to build host stuff / tools
<ynezz> don't you just need to add proper dependency chain to tools/Makefile?
<stintel> how
<stintel> there is nothing in tools/ that has dependencies at all
<ynezz> line 40 # builddir dependencies
<stintel> ynezz: from a quick glance these all appear to depend on things within tools/
<ynezz> then I probably don't follow
<stintel> so I can just add there $(curdir)/dwarves/compile := package/elfutils/host ?
<ZorPrime> stintel: https://pastebin.com/5FB8wG8v --- thanks for your time sir :)
<ynezz> stintel: no, that doesn't seem right
cbeznea has left #openwrt-devel [#openwrt-devel]
hanetzer has quit [Quit: WeeChat 3.4]
cbeznea has joined #openwrt-devel
<stintel> ZorPrime: which version of OpenWrt are you trying to build?
<ynezz> stintel: why it needs to go into tools?
<cbeznea> is there a way to fetch build status from buildbot?
<ZorPrime> v21.02.2 --- focusing on the cjdns package only, don't care about the rest
<cbeznea> or to register somewhere in order to be notified about it?
<stintel> ynezz: where else? it's required at kernel build time when KERNEL_DEBUG_INFO_BTF is enabled - it's not going to be a package for target
<stintel> ZorPrime: looks like 21.02 might not like building on musl systems :(
cbeznea has quit [Read error: Connection reset by peer]
<ZorPrime> stintel: bummer... should I use the docker image for dev work then?
<ZorPrime> I've got docker installed
<stintel> you could give that a try
cbeznea has joined #openwrt-devel
<stintel> I don't even recall if we support building on musl hosts
<stintel> and don't really have time to look into that now
<ZorPrime> stintel: you've done more than enough thanks... just needed a hint
<stintel> awesome :)
<ZorPrime> I've got laptops sitting around I can just throw centos or something and run. Or do docker
<ZorPrime> I'll look through the wiki further for the "preferred" build env
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<stintel> I think most people use debian or ubuntu
Borromini has joined #openwrt-devel
bluetail has joined #openwrt-devel
<bluetail> @stintel that has a good reason. Debian is capable of secureboot. And that is a nightmare to setup yourself and cost a bit of money. ArchLinux is not by default to name one.
<PaulFertser> SecureBoot doesn't cost anything if you take the shim verbatim.
<PaulFertser> And I do not think doing it the way ubuntu or debian does by default is anywhere near complicated. You just put shim to the efi system partition and change efi boot vars to use it.
minimal has joined #openwrt-devel
<bluetail> @PaulFertser: Both Ubuntu or Debian support SecureBoot by default. You just can go ahead and use Rufus or Balena Etcher. With Arch? You have to sign the Kernel with your own enrolled Key through MOK manager and you are betting for problems
<PaulFertser> bluetail: /or/ you can just do the same thing Debian does and boot the shim instead of signing anything of your own.
<bluetail> How would you do that? "Warning: Replacing the platform keys with your own can end up bricking hardware on some machines, including laptops, making it impossible to get into the UEFI/BIOS settings to rectify the situation. This is due to the fact that some device (e.g GPU) firmware (OpROMs), that get executed during boot, are signed using Microsoft's key."
<PaulFertser> bluetail: the shim is signed my MS keys, and shim boots unsigned grub.
<PaulFertser> s/my/by/
<bluetail> Is there a guide to read for using the verbatim shim to boot arch linux on systems with secureboot and uefi enabled?
hanetzer has joined #openwrt-devel
<PaulFertser> bluetail: not sure about the guide but can you try just copying it and booting instead of grub?
<bluetail> 'it' meaning EFI bootfiles? Which ? I know how to mount the EFI.
<PaulFertser> I understand you describe using https://www.rodsbooks.com/efi-bootloaders/secureboot.html with signing the kernel etc. But I think there's an easier way.
rsalvaterra has quit []
rsalvaterra has joined #openwrt-devel
<bluetail> I have been using Ventoy with MokManager to boot some *.iso files that wouldnt boot otherwise. it fails far often as the Kernel isn't loaded. I enrolled a key and then I can boot for instance manjaro. But I cannot proceed for some reason - the installer aborts during installation. My solution? Use debian and create a live usb disk with persistent storage using rufus.
shibboleth has joined #openwrt-devel
<bluetail> To clarify: installer breaks when partitioning. If I dare to install it though a virtualmachine on my physical disk in uefi mode, it wouldnt boot after, too.
<bluetail> though/through
<PaulFertser> bluetail: I have to admit I might be wrong about that regular shim being able to boot unsigned grub.
<bluetail> I still don't know how - I felt lost in that guide earlier. But trying REFind boot manager and clover - both cant boot the bootloader - if tried, it says that I should load the kernel first.
<bluetail> I don't have any access to the BIOS for exporting Keys, if that matters.
goliath has joined #openwrt-devel
Misanthropos has joined #openwrt-devel
danitool has joined #openwrt-devel
GNUmoon2 has joined #openwrt-devel
GNUmoon has quit [Read error: Connection reset by peer]
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
GNUmoon2 has quit [Remote host closed the connection]
GNUmoon2 has joined #openwrt-devel
Acinonyx has quit [Quit: Peer reset by connector]
rua has quit [Remote host closed the connection]
rua has joined #openwrt-devel
Luke-Jr has quit [Ping timeout: 480 seconds]
<mangix> stintel: all tools dependencies must be present within tools. Which means elftools must be under tools and dependency added under tools/Makefile
* stintel forgets about getting these changes in before the freeze
jordbaer has joined #openwrt-devel
Luke-Jr has joined #openwrt-devel
<ZorPrime> stintel: Got that docker build env going, got past the point where I was stuck. Host is ubuntu 20.04 based. Thanks again for the once over
<owrt-snap-builds> Build [#476](https://buildbot.openwrt.org/master/images/#builders/54/builds/476) of `ramips/mt7620` failed.
minimalist has joined #openwrt-devel
minimalist has quit []
ZorPrime has left #openwrt-devel [#openwrt-devel]
minimal has quit [Ping timeout: 480 seconds]
ZorPrime has joined #openwrt-devel
minimal has joined #openwrt-devel
<stintel> ZorPrime: you're welcome :)
<ynezz> I've just fixed that mt7620 build failure, missing kernel config symbol
<torv> ZorPrime: Would you instruct me how to get that docker env up/how to use it?
<ZorPrime> torv: sure... I'll PM it over in a few
PaulFertser has quit [Ping timeout: 480 seconds]
GNUmoon2 has quit [Ping timeout: 480 seconds]
PaulFertser has joined #openwrt-devel
jordbaer has quit [Quit: jordbaer]
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
Luke-Jr has quit [Ping timeout: 480 seconds]
GNUmoon2 has joined #openwrt-devel
GNUmoon2 has quit []
GNUmoon2 has joined #openwrt-devel
<Borromini> well seems MT7613 on master suffers from way less timeouts...
Luke-Jr has joined #openwrt-devel
<svanheule> Borromini: baby steps :)
<svanheule> stintel: thanks for your work on the EAP615! Good to see it merged
<Borromini> svanheule: judging by the amount of stalls/timeouts I saw on 21.02 for the past week... looks like giant's steps xD
<svanheule> :P
Borromini has quit [Quit: leaving]
shibboleth has quit [Quit: shibboleth]
danitool_ has joined #openwrt-devel
danitool has quit [Ping timeout: 480 seconds]
<stintel> svanheule: thanks for the help and review
<owrt-snap-builds> Build [#476](https://buildbot.openwrt.org/master/images/#builders/15/builds/476) of `armvirt/64` completed successfully.
c0sm1cSlug has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
c0sm1cSlug has joined #openwrt-devel
<philipp64> can anyone review (and hopefully merge) my xtables-addons version bump?
goliath has quit [Quit: SIGSEGV]