<robimarko>
All I wanted is missing prefix to resolve the formal issue
<f00b4r0>
robimarko: *shrug* is the correct reaction here ;)
<robimarko>
I dont want to drive anybody away, but I have no clue what I did wrong there
<f00b4r0>
you did nothing wrong.
<aparcar>
robimarko: if think if people have such a hard time understanding things you can also just add the missing formal bits...
<robimarko>
aparcar: I mean, if they cannot do that why should I/anybody else try and understand why they are making that commit?
<f00b4r0>
and then it becomes a habit "look you did it before, why give me a hard time?" and all contributions decline in quality.
<robimarko>
CI was so clear in what needs to be done that it cannot be easier
<aparcar>
sure your choice
<KanjiMonster>
https://openwrt.org/submitting-patches says "commit subject ... must be less than 50 characters long" and "The “summary phrase” in the email's Subject ... must be no more than 70-75 characters"
skynet2 has joined #openwrt-devel
mrkiko has quit [Quit: leaving]
mrkiko has joined #openwrt-devel
fakuivan has quit [Remote host closed the connection]
fakuivan has joined #openwrt-devel
nixuser has quit [Remote host closed the connection]
nixuser has joined #openwrt-devel
<jow>
robimarko: the wrong thing you did was not picking up a random drive-by contribution
<jow>
because it's just a simple change, can't bo se hard to fix it up, right?
<jow>
and it's github after all, so just hit the green button
<Habbie>
yeah robimarko, how hard is it to just push the merge button
<Habbie>
:>
<Habbie>
2024-04-08T12:21:58.0673485Z /builder/staging_dir/toolchain-powerpc_8548_gcc-13.2.0_musl/bin/powerpc-openwrt-linux-musl-ld: cannot find arch/powerpc/lib/crtsavres.o: No such file or directory
<Habbie>
known
<Habbie>
?
<robimarko>
Oh I saw that in some tests yesterday
<robimarko>
But havent had the time to actually look into
<jow>
to be fair, brada's contributions are usually worthwhile (He's also submitting various improvements to the firewall) but his commits are usually very hard to decipher
<robimarko>
That is what threw me off, cause he is ultra active in the issues
<robimarko>
And usually rather helpfull
<Habbie>
jia tan vibes
<Habbie>
"somebody stole his account for a day"
<jow>
no, this guy is helpful and obviously deep into various topics, also adinging in debugging and various low level things
<jow>
but he's not really good at doing proper PRs
<jow>
or commits
<jow>
unfortunately we all lack the manpower to pick up unstructured contributions and steamlining them
<Habbie>
right
<jow>
we can't even handle all good contributions in a timely manner
<Habbie>
that's a common problem, yes
<KanjiMonster>
Habbie: 5.15? a recent update removed a related patch deemed as "upstreamed" (or upstream change assumed making it obsolete) https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=1c28058710fd524054f364deb2b5efc754d695e4
<Habbie>
i just managed to delegate some "fell of the wagon" powerdns PRs to a coworker
<Habbie>
because we also don't like asking people two years later "can you rebase this please"
<Habbie>
KanjiMonster, all i know is I'm PRing to a fork of openwrt/packages
nwf has quit [Ping timeout: 480 seconds]
Znevna has quit [Quit: ZNC 1.8.1+deb1~bpo10+1 - https://znc.in]
<robimarko>
Hm, that powerpc crtsaves is a weird one
<robimarko>
Cause upstream clerly switched building it to always-y
Znevna has joined #openwrt-devel
<KanjiMonster>
robimarko: only PPC64 did, mps53xx is PPC32. PPC32 uses obj-y, and neither extra-y nor always-y
<KanjiMonster>
*mpc85xx
<robimarko>
Ok, let me try and reproduce that locally
Znevna has quit [Ping timeout: 480 seconds]
<jow>
hm, anyone here familar with the dnsmasq multi instance support?
<jow>
are the config host / config cname etc. sections shared among instances?
<robimarko>
KanjiMonster: Well, crazy thing is that using apm821xx which fails in telephony CI dahdi-linux compiles just fine
<robimarko>
Ahah, but using SDK throws that error
<robimarko>
So its missing probably in SDK only
grid has quit [Quit: grid]
<f00b4r0>
so I'm confused, I thought that with "recent" kernels, software flow offloading no longer broke traffic shaping. Yet I have a setup where it breaks *upload* shaping. Is this expected?
grid has joined #openwrt-devel
schwicht has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
* f00b4r0
realizes that download shaping goes through an ifb iface and that probably breaks offloading
Znevna has joined #openwrt-devel
nitroshift has joined #openwrt-devel
<robimarko>
Yeah, classic builds work because arch/powerpc/lib/crtsavres.o is present
<robimarko>
But SDK doesnt have arch/powerpc/lib at all
<robimarko>
I think its filtered out since its just .o
xback has joined #openwrt-devel
<aparcar>
I'm building for Ramips and see " DM audit events (DM_AUDIT) [N/y/?] (NEW)", anyone keen to fix this?
minimal has joined #openwrt-devel
schwicht has quit [Read error: Connection reset by peer]
noltari has quit [Ping timeout: 480 seconds]
<robimarko>
ynezz: Any chance you can push the buildbot tar.xz relaxation so that move to ZSTD for SDK, IB and rest of tarball targets can be done?
<stintel>
ugh wtf, I removed the ubi volumes from the eap683-lr dts yet the kernel is still attaching them at boot
<stintel>
and I verified the image, the volumes nodes are not in the dtb
<robimarko>
cmdline?
xdarklight_ has joined #openwrt-devel
<stintel>
cmdline is empty when booting the initramfs, also not set in uboot env, I suspect the "mtkboardboot" command sets things
lynxis has quit [Read error: Connection reset by peer]
gladiac has quit [Read error: Connection reset by peer]
bookworm has quit [Read error: Connection reset by peer]
takimata has quit [Write error: connection closed]
clandmeter has quit [Write error: connection closed]
svanheule has quit [Read error: Connection reset by peer]
xdarklight has quit [Remote host closed the connection]
bookworm_ has joined #openwrt-devel
xback has quit [Remote host closed the connection]
bookworm_ is now known as bookworm
rmilecki has quit [Remote host closed the connection]
<aparcar>
Ansuel: afaik there is no fallback for the missing subtargets. We could add a default like SUBTARGETS:=generic in case it's not defined at all, however right now it requires the subfolder with target.mk plus BOARDNAME=generic
<Ansuel>
i would error if SUBTARGET is not defined
<Ansuel>
just to enforce it
nitroshift has quit [Quit: Leaving that way ---->]
<robimarko>
Hm, but it depends on TOOLCHAINOPTS so how is it even visible
<robimarko>
Even menuconfig shows Visible if: TOOLCHAINOPTS [=n]
<stintel>
ah, I forgot to remove compatible = "linux,ubi"
Mangix has quit [Read error: Connection reset by peer]
<stintel>
./readelf.c:372:3: error: unknown type name 'off64_t'; did you mean 'off_t'?
<stintel>
ugh
ILEoo has joined #openwrt-devel
philipp64 has quit [Quit: philipp64]
svanheule has joined #openwrt-devel
<mirko>
svanheule: ping
Mangix has joined #openwrt-devel
<Mangix>
stintel: _LARGEFILE64_SOURCE or w/e
Znevna has quit [Ping timeout: 480 seconds]
<stintel>
Mangix: this just appeared after git pull, I had to make dirclean
rmilecki has joined #openwrt-devel
minimal has quit [Quit: Leaving]
goliath has joined #openwrt-devel
<stintel>
christ that caldata extraction is a fucking mess
minimal has joined #openwrt-devel
<mrnuke>
stintel: welcome to QCA!
<stintel>
mrnuke: this is mtk
<stintel>
but caldata_dd() uses count variable in bs= when calling dd
<stintel>
and sets count=1 hardcoded
<stintel>
super confusing
<stintel>
and debugging is impossible, because I don't see those echos in caldata.sh anywhere
<svanheule>
mirko: pong
<mrnuke>
stintel: I just love how one of the arguments to dd is named "bs"
<stintel>
and then the various functions take count and offset in different orders
<stintel>
fucking inconsistent pile of shit
<mrnuke>
stintel: that would be one of the ingredients in "bs", yeah
<mrnuke>
stintel: does it not stick to the "partition, offset, size" order of arguments? I thought that was the convention
<mirko>
svanheule: are you telling me that non of the realtek boards can (currently) actually handle untagged VLAN configs on whatever port(s)?
<svanheule>
apparently they _need_ to be in VLAN 1, although the PVID doesn't have to be 1
<svanheule>
the one switch I'm using runs untagged traffic (default config)
<mirko>
svanheule: ok, then i now bought and committed support for a switch which is useless to me :p
<svanheule>
sorry about that :(
<svanheule>
the good news is that it should be fixable ;-)
<mirko>
svanheule: i just wonder how that didn't pop up more prominently, given the amount of "supported" realtek switches by now
<svanheule>
I've started to look into it, but it's probably going to be a rabbit hole of verifying how the SDK handles a few register blocks and switch control tables
<svanheule>
I guess either a lot of people just don't use VLANs, or use VLAN 1 in on every port in their config (either tagged or untagged)
<robimarko>
Now, this GDB popping up in root level of menuconfig is getting weirder, as long as the newly added config GCC_USE_DEFAULT_VERSION entry exists its "broken"
<robimarko>
As soon as that is removed it magically gets fixed
<stintel>
svanheule: why would you buy a managed switch if you dont use vlans ;)
<svanheule>
I'm not the one to judge people choices :^)
<svanheule>
* hides a cabinet full of "development" switches
<stintel>
:D
<robimarko>
Hehe, I also have more "development" gear than what is actually used
<svanheule>
got to have back-up devices for testing build, right?
<robimarko>
Ok, so it seems that as long as config GCC_USE_DEFAULT_VERSION is put after GCC_VERSION then it doesnt break stuff
<robimarko>
And still works
<stintel>
here it's probably close to 50/50
<stintel>
but once I get the keys to my house I'll be using most of my "development" gear also
<stintel>
fun times ahead, looking to buy a wall chaser soon
<stintel>
can't be having bare UTPs running through the house :P
<robimarko>
Where is the fun in having neat cabling
<robimarko>
Its been only like 3 years since I bought and swore to terminate my UTP-s in a patch panel
SlimeyX has quit [Remote host closed the connection]
goliath has quit [Quit: SIGSEGV]
<stintel>
[ 1.020447] ubiblock: device ubiblock0_1 (rootfs) set to be root filesystem
<stintel>
[ 3.897441] Waiting for root device /dev/ubiblock0_0...
<stintel>
great, more inconsistency
<stintel>
and this is with the bootargs in dts: bootargs = "console=ttyS0,115200n1 ubi.mtd=rootfs root=/dev/ubiblock0_1 rootfstype=squashfs rootwait loglevel=8 earlycon=uart8250,mmio32,0x11002000";