<Tusker>
mangix: yeah, this is the first time I've seen it too... I wonder if there was a system package update that decided to be more fussy with executables
<Tusker>
glibc updated to 2.31-13+deb11u2 updated this month, which may be related
musashino has quit [Read error: Connection reset by peer]
Andi_ has quit [Read error: Connection reset by peer]
Andi_ has joined #openwrt-devel
<nbd>
no idea about luci
<nbd>
it shouldn't require much configuration aside from enabling the interface and putting in your bandwidth
rua has joined #openwrt-devel
ashkan has joined #openwrt-devel
dangole has joined #openwrt-devel
<dangole>
Hauke: you are the hero of the month (or even year?). john, stintel and me were stairing at what I already thought would be stack corruption for hours and hours and now you spotted it.
<dangole>
Hauke: and as usual, once you see it, it's very obvious and of course needs to be 64-bit offset on Elf64...
jbowen has quit [Ping timeout: 480 seconds]
rua has quit [Quit: Leaving.]
rua has joined #openwrt-devel
felix has quit [Ping timeout: 480 seconds]
<hauke>
dangole: I think gcc can warn about such problems
jbowen has joined #openwrt-devel
<xdarklight>
nbd: should qosify (and/or bpf-headers) be hidden behind CONFIG_DEVEL because CONFIG_BPF_TOOLCHAIN also depends on CONFIG_DEVEL ? on a Linux host I can select quosify with CONFIG_DEVEL=n which results in a build failure
<dangole>
hauke: i added a note to the commit message of the 2nd commit which is 100% revert of 33b799b ("ujail: elf: work around GCC bug on MIPS64"), stating that. i'd test-run on ARMv7 and Aarch64, then push. ok?
pmelange has joined #openwrt-devel
felix has joined #openwrt-devel
fda has joined #openwrt-devel
felix_ has joined #openwrt-devel
felix__ has joined #openwrt-devel
felix has quit [Ping timeout: 480 seconds]
felix has joined #openwrt-devel
<hauke>
dangole: ok
felix_ has quit [Ping timeout: 480 seconds]
<nbd>
xdarklight: i'll look into it. that would be a bug then
felix__ has quit [Ping timeout: 480 seconds]
<xdarklight>
nbd: thanks. I am using Arch Linux and I don't have clang installed there. I am aware that CONFIG_BPF_TOOLCHAIN_BUILD_LLVM exists, but I was curious as to why the build fails using standard settings
<hauke>
dangole: jow was siggesting to use off_t or size_t instead of long
<hauke>
*suggesting
felix_ has joined #openwrt-devel
felix has quit [Ping timeout: 480 seconds]
<nbd>
xdarklight: fixed now
<xdarklight>
nbd: confirmed - thank you!
dangole has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
jbowen has quit [Ping timeout: 480 seconds]
fda- has joined #openwrt-devel
fda has quit [Ping timeout: 480 seconds]
felix_ has quit [Read error: No route to host]
felix has joined #openwrt-devel
jbowen has joined #openwrt-devel
felix has quit [Quit: No Ping reply in 180 seconds.]
felix has joined #openwrt-devel
fda has joined #openwrt-devel
fda- has quit [Ping timeout: 480 seconds]
<Tapper>
Where is qosify in make menuconfig?
<hauke>
when I activate -Wconversion in libubox I get a lot of warnings
<hauke>
we are strong milli seconds in int and sizes in int
<hauke>
would it be accaptable to fix this even with ABI chanegs?
<jow>
in master yes
ashkan has quit [Ping timeout: 480 seconds]
<Tapper>
nbd is qosify in make menuconfig?
<Tapper>
nbd I did a search and no hits
danitool has joined #openwrt-devel
<nbd>
Tapper: you need to select "advanced configuration options" and change the setting under "BPF Toolchain"
<nbd>
depending on whether you have llvm+clang installed on your host
<nbd>
if you don't, you'd need the build system to build it
<nbd>
but it will take a lot of time
<Tapper>
nbd OK
<Tapper>
Thanks
minimal has quit []
<Tapper>
nbd so after that qosify will show up in make menuconfig?
fda has quit [Ping timeout: 480 seconds]
dorf has joined #openwrt-devel
<nbd>
Tapper: yes
<mangix>
ldir: github PR for ldconfig issue uses /bin/true
<nbd>
Tapper: i also pushed a fix just now that puts it in the right submenu in make menuconfig
<mangix>
so qosify is an SQM replacement I take it?
<ldir>
mangix: unfortunately that didn't permit libluajit to compile
<mangix>
ldir: since you're on a mac, does glib2 compile?
<nbd>
mangix: replacement for sqm-scripts, yes
<Tapper>
nbd OK build failed but did git pull and starting again.
<ldir>
mangix: apols - am on-air at the mo, no real time for openwrt
<Tapper>
mangix O Shit. It's not overclocked or anything like that.
<Tapper>
I am using windows 11 WSL
<Tapper>
Ubuntu
<Tapper>
The same thing hapends 3 times in a row.
<mangix>
Tapper: wsl or wsl2?
<Tapper>
WSL2
<mangix>
that's a VMM
<mangix>
*VM
<mangix>
not enough RAM is assigned to it
<mangix>
one way to work around this is to add PKG_BUILD_PARALLEL:=0 to the Makefile
<Tapper>
I have 16 GB installed and it never gets over 13 GB.
<Tapper>
I have used to build OpenWrt a 1000 times
<mangix>
doesn't matter. LLVM is quite heavy.
<Habbie>
when i make package/dnsdist/compile, i end up building openssl, host-boost, and 30 other things
<Habbie>
i presume there's no big tarball i can get somewhere that has most of those things built already?
<Habbie>
for a given target
<nick[m]1234>
has someone a quick link to a github repro to build a target?
<nick[m]1234>
sorry I mean github workflow, I looked at aparcars one but they build everything and uploads it to some aws repro
schwicht has joined #openwrt-devel
shibboleth has joined #openwrt-devel
<jow>
ldir: any reason to not pick https://git.openwrt.org/?p=openwrt/staging/ldir.git;a=commitdiff;h=4781d38a6e6e145f5d9bbcda0853d31057c0eb80;hp=2a4555aac96544efbae4823eb9584b2460662961 ?
<ldir>
jow: not that I can think of - lack of confidence on my part - don't what to break the firewall for everyone
jbowen has quit [Ping timeout: 480 seconds]
Grommish_ has joined #openwrt-devel
dorf has quit [Remote host closed the connection]
dorf has joined #openwrt-devel
fda has joined #openwrt-devel
Grommish has quit [Ping timeout: 480 seconds]
Borromini has joined #openwrt-devel
<Tapper>
mangix can I use llvm in ubuntu insted of building it?
dangole has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
<jow>
ldir: understood. It looks fine to me
fda- has joined #openwrt-devel
<mangix>
Tapper: haha no
fda has quit [Ping timeout: 480 seconds]
<mangix>
Habbie: ccache is your friend
<Tapper>
Yes you can I have it working
<Habbie>
ack, i love ccache for my other builds, openwrt probably needs a nudge to also use it
<Tapper>
You have to add /usr/bin/llvm- as the path in make menuconfig
<Habbie>
oh, just CONFIG_CCACHE?
<Habbie>
no, not that simple
dorf has quit [Remote host closed the connection]
dorf has joined #openwrt-devel
Borromini has left #openwrt-devel [#openwrt-devel]
Tapper has quit [Read error: Connection reset by peer]
Tapper has joined #openwrt-devel
<Tapper>
nbd Mon Nov 1 21:42:13 2021 daemon.info procd: Instance qosify::instance1 s in a crash loop 6 crashes, 0 seconds since last crash
Andi_ has quit [Read error: Connection reset by peer]
Andi_ has joined #openwrt-devel
<Grommish_>
stintel: ping
Grommish_ is now known as Grommish
<xdarklight>
hauke: which mediatek device do you have that uses NAND flash? I'll try making the upstream NAND driver work on MT7621 in the next days. I don't have any of the already supported SoCs (for example MT7622) to cross check that I didn't break anything
astronautap has joined #openwrt-devel
isinyaaa has joined #openwrt-devel
kenny has quit [Ping timeout: 480 seconds]
<Tusker>
ooh... just got the GPL source code from WatchGuard
pmelange has left #openwrt-devel [#openwrt-devel]
astronautap has quit [Quit: WeeChat 3.3]
jlsalvador has quit [Quit: jlsalvador]
<stintel>
Grommish: pong
<stintel>
Tusker: o_O
<stintel>
that reminds me I should still file my request
<Tusker>
xtm330 and t30-w bundles
<Tusker>
xtm330 I still can't get the switch to function properly, so I am pretty interested in the contents
<Grommish>
stintel: I found a memleak issue under Octeon and 5.10, as well. I left a comment on the PR. I'm not sure what to do now though
<Grommish>
It appears to be a swapper/0 and softirq that's popping up
<stintel>
I arrived back home yesterday, no travel plans for the first 6 weeks, so I should be able t o work on some things
<stintel>
number 1 on my list is getting qoriq in master
<Grommish>
If I can help, let me know. I've got a dev device for octeon/Octeon3 and source build.. I just don't know how to debug/trace things like this
<Grommish>
but at least it wasn't Adblock :D
<stintel>
I have some cn6xxx cards
<Grommish>
cn7020
<stintel>
the ERLs are no longer with me
<stintel>
but the cn6xxx should allow me to try and reproduce
<Grommish>
*nod*. The target still uses octeon+, so it should be fine
<Grommish>
If I can help out, hit me up.. I've got time and some knowledge :D
<stintel>
as for next step, you could try reporting the kmemleak findings upstream
<Grommish>
Upstream as in.. To where? I've done some opensource stuff before, but only here and Android
<stintel>
when I see kmem_cache_* ... I know shit is over my head
kenny has joined #openwrt-devel
<Grommish>
You did better than me.. I stopped and bailed at unreferenced object :
<Grommish>
:D
<stintel>
find cvm_oct_mem_fill in kernel.git, run scripts/find-maintainer.pl path/to/source/containing/that/function
<stintel>
and mail those addresses
<Grommish>
Ok
<stintel>
cc me, too
<Grommish>
stintel: DM me an email?
jlsalvador has joined #openwrt-devel
<stintel>
Grommish: the script is scripts/get_maintainer.pkl
<stintel>
without the k
<stintel>
so in the kernel clone: ./scripts/get_maintainer.pl drivers/staging/octeon/
<stintel>
wait, I assumed the ethernet driver
<stintel>
ok, cvm_oct_mem_fill_fpa is defined there, so that's correct
goliath has quit [Quit: SIGSEGV]
<Grommish>
Yeah, I sent the listed maintainer an email an CC'd you onn it
<stintel>
excellent
<Grommish>
If it's the ethernet driver, it makes sense
<Grommish>
Adblock and dnsmasq would cause huge churn with it