<Habbie>
why does a package build for dnsdist say 'Skipping dnsdist due to unsupported architecture' on powerpc_8540 and arc_archs? I see nothing -explicitly- filtering those archs in dnsdist/Makefile
Tapper has quit [Ping timeout: 480 seconds]
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
userx has joined #openwrt-devel
userx has quit [Remote host closed the connection]
uxerx has joined #openwrt-devel
uxerx has quit [Remote host closed the connection]
userx has joined #openwrt-devel
goliath has joined #openwrt-devel
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
<userx>
Hello. Before build a package that has a number of dependencies, I need to run "make menuconfig" and choose this package?
<userx>
As a result, the required dependencies will be packaged during the build process. Without this step, the package will not be built because the build system will not know what to do.
<userx>
I understood correctly?
<userx>
To automate the process, I can use my own version of the config file.
<userx>
By the build process, I mean make package/<package-name>/compile.
<rmilecki>
userx: yes, you need your config to contain dependencides
<rmilecki>
userx: "make menuconfig" will select them for you
<rmilecki>
userx: maybe you could also modify .config manually and try "make oldconfig" (or "make defconfig" to avoid prompting)
<rmilecki>
oldconfig / defconfig should also enable dependencide in your .config
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
<userx>
rmilecki: Thank you very much!
<Habbie>
2021-11-08T09:27:57.5460239Z crtstuff.c:(.text+0x52): R_PPC_PLT16_HA relocation unsupported for bss-plt
<Habbie>
luajit appears to be broken on powerpc_464fp
victhor has joined #openwrt-devel
musashino has quit [Ping timeout: 480 seconds]
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
pmelange1 has joined #openwrt-devel
pmelange1 has left #openwrt-devel [#openwrt-devel]
Andi_ has quit [Ping timeout: 480 seconds]
Andi_ has joined #openwrt-devel
<stintel>
Habbie: probably another problem mentioned earlier in the output
<stintel>
I've seen that too recently
<Habbie>
ok
<Habbie>
i might just ignore it for now then
<Habbie>
the archs i care about today are fine
tohojo has quit [Quit: Bridge terminating on SIGTERM]
tohojo has joined #openwrt-devel
robimarko has joined #openwrt-devel
<robimarko>
Hi, I am working on IPQ40xx DSA and now am converting the LEDs
<robimarko>
But I am having trouble converting some port masks to their port
<robimarko>
For example Fritzbox 4040 and RT58 use 0x1e as the LAN trigger
<stintel>
robimarko: re: getting ipq40xx reviewed is hard: I have an Edge-Core ipq40xx device, I could test some things if that would help
<stintel>
reviewing .. your changes are probably out of my comfort zone for that
<robimarko>
As far as I know swconfig triggers are 1 << port_nr
<robimarko>
But 0x1e doesnt match any of the ports
<robimarko>
Anybody have a clue what am I misunderstanding
<robimarko>
stintel: You have the ECW5211?
<stintel>
I do
<robimarko>
Since thats PSGMII I would say thats been well tested
<robimarko>
RGMII is missing detailed testing, got it working but VLAN-s etc really need to be tested though it should work just fine
<robimarko>
stintel: Any clue on the LED trigger?
<stintel>
sorry, no clue at all
<robimarko>
It looks like it originates way back from the days where IPQ40xx was a subtarget of IPQ806x
<robimarko>
Thanks anyway
musashino has joined #openwrt-devel
<Habbie>
do we have a tool for extracting the name of a package from a .ipk filename?
<Habbie>
splitting at _ seemed obvious but some things in packages already have a _ in their name
<Pepes>
OpenWrt.org website is down?
<Habbie>
looks down to me
<stintel>
it's slow but works
<Habbie>
there it is
<Habbie>
almost two minutes
<stintel>
was ~15s here
<stintel>
but it's the shitty dokuwiki plugin again
<neggles>
that's... whoever wrote the code which generated that query should feel bad
<neggles>
it's not as bad as f#$*ing magento searches but
<stintel>
and iiuc the taskrunner is causing such query to be executed every time somone opens a device page
<neggles>
...but why
<stintel>
that's just, uhm, wtf, ugh, gdiaf
<stintel>
lack of better? lack of time to test if something can do our usecase better?
<neggles>
is that why toh device pages always take forever to load
<neggles>
oh no no i meant "what even is it doing, and why is it doing it like that"
<stintel>
I don't know, I was not involved in setting it up
<stintel>
I cursed a lot when I broke it due to bringing the wiki machine into the present
<neggles>
so many joins...
<neggles>
i am not a DBA but i'm like 80% sure this could be a view that gets updated when a device is edited
<neggles>
there should not need to be a join for every single row in a ToH entry
<neggles>
*ToH entry/device page
<neggles>
buuuut i'm not volunteering to fix it :P
<neggles>
stintel: i'm sure you get asked this more than you'd like to, but would you happen to have made any progress on the SNIC10E cards since april/may-ish? i've got one of the things booting (i think... don't have anything hooked up the UART and the PCIe serial console appears to not work, but it doesn't show as crashed) so that's cool and good
<neggles>
better make up a cable so I can see wtf it's up to
<stintel>
neggles: have not touched it since then
<stintel>
I'm currently doing some firewall4 work needed to be able to make it default in master, hope to land that this week
<stintel>
then I want to finish the qoriq target
<stintel>
and after that I'll probably play around with that again
<stintel>
I'd first have to revive the only machine I can use these cards in
<neggles>
heh, fair enough! just thought i'd ask :)
<neggles>
no point repeating work
<neggles>
might be able to get somewhere, but my god do Marvell not make it easy to work with MIPS Octeon
<stintel>
I've some cards that have NAND now
<stintel>
the ones I got from ebay only have NOR which is not large enough to install OpenWrt on
<stintel>
but hurricos was so kind to send me a bunch with NAND
<neggles>
I got two NAND-equipped ones for ~$30 AUD or so
<neggles>
the two I have, have amazon version strings in their u-boot and the simple executive binary in a NOR JFFS2 partition
<neggles>
I suspect these are first-gen AWS Nitro cards
Acinonyx_ has joined #openwrt-devel
<neggles>
if they are, then there might be some clues in the NAND as to how the pretend-o-ethernet mode works, which would be awesome to get working in owrt - probably more effort than it's worth, but firewall-in-NIC would be great for a colo'd server for example... but now i'm just rambling
<userx>
I am trying to create a Python package, but it require "wheel" module in "setup.py" script.
<userx>
I created "python-wheel" package and added it to PYTHON3_PKG_SETUP_DIR. In output of "make package/<package-name>/compile" command I can see that "python3-wheel" package has been created. But when it comes to "setup.py install" command, I get the error message: "ModuleNotFoundError: No module named 'wheel'"
<userx>
How can I add build dependency to toolchain (staging_dir)?
<nbd>
jow: with regex, i can do things like match google.[^\.]*$
<nbd>
i meant google\.[^\.]*$
<nbd>
so an any-tld match
<nbd>
doesn't work with fnmatch
<nbd>
i guess i could change it to do fnmatch by default, and support regex via /expr/ syntax
<nbd>
dhewg: you're right
<nbd>
will fix
<nbd>
jow: can you help me with downloading and unpacking the llvm tarball into the sdk dir in the phase2 build?
gladiac is now known as Guest5374
gladiac has joined #openwrt-devel
Guest5374 has quit [Ping timeout: 480 seconds]
Acinonyx has joined #openwrt-devel
Acinonyx_ has quit [Ping timeout: 480 seconds]
Tapper has joined #openwrt-devel
<Habbie>
aparcar[m], do i see correctly that the 'Formalities' (signoff) test passed on https://github.com/openwrt/packages/pull/17078 (the merged one) because 'artifacts:' -happens- to match the syntax?
<aparcar[m]>
Habbie: answered in thread
<Habbie>
ah yes, but that then indeed is how the other one got away with it
<Habbie>
and done
kenny has joined #openwrt-devel
Tapper has quit [Ping timeout: 480 seconds]
<nbd>
jow: the dns syntax has two modes now
<nbd>
dns:<fnmatch>
<nbd>
dns:/<regex>
<nbd>
also, installing files to /etc/qosify/ without updating the config is possible now
<aparcar[m]>
nbd: are there docs on qosify?
<nbd>
there's a readme in the repo
<nbd>
not much else yet
<aparcar[m]>
ok thanks I'll have a look
pmelange has joined #openwrt-devel
<hurricos>
Tusker: Probably the right observation on your part. I'll have to set up my test rig at the Lab tonight so I can test without having to set everything up every time ...
<hurricos>
neggles: I spent a lot of time playing with the CN6640; exact same situation as you RE: amazon version strings in their u-boot. PM me if you want me to give you a brain dump on my understanding of how the cards were used and what might be done to get them working either as OpenWrt hosts themselves or as driven by a Linux host attached over PCIe
<hurricos>
... using the liquidio driver. And by CN6640 I mean the pile of SNIC10E-G cards that I got.
<hurricos>
neggles: > pretend-o-ethernet -- The cards you and I have are probably intended to work with with the linux `liquidio` on Amazon's Xen-based baremetal. Mainline Linux liquidio doesn't actually care what you set `fw_type` to, it always tries to identify the card and push a firmware rather than checking the NOR.
<hurricos>
One problem is that what it pushes by default (lio_210sv_nic.bin) doesn't appear to actually know how to identify the vsc8488 PHY. I never went back and checked the NOR for strings to verify the on-NOR version actually did have a 'vsc8488' string.
<mangix>
Habbie: mips64 can be added. PR welcome.
<Habbie>
mangix, ok! i was thinking, different people might have different needs, but if upstream is happy with expansion, those downstreams can do simple tricks to skip a few, so i might poke at that too
<Habbie>
mangix, will PR soon, thanks
<hurricos>
RE: the NAND, IIRC it was basically empty on all my cards except for an extra copy of U-boot. Cavium leaves the NAND on the development boards because it's at worst, +$2 on the BOM on a widely deployed, development-scale card which costs nearly five figures
Andi_ has quit [Read error: Connection reset by peer]
Andi_ has joined #openwrt-devel
noltari has quit [Read error: Connection reset by peer]
<Habbie>
> Please remove the g, I think more people will think it's part of the hash and end up confused.
<Habbie>
russell--, a question, if I may - the LoC is the party that can make these exemptions? 'Librarian' does not sound like a party that does such things
<russell-->
yeah, it is. obviously this is US-centric too
<Habbie>
oh definitely, that's why I know I don't have enough context :)
rmilecki has quit [Ping timeout: 480 seconds]
<Habbie>
good news in any case
<russell-->
"As required under the
<russell-->
statute, the Register of Copyrights, following a public proceeding, submitted a
<russell-->
Recommendation concerning proposed exemptions to the Librarian of Congress."
<russell-->
it's kind of the mechanism that is used to prevent the DMCA from being overturned outright, a mechanism to address a few of the egregious effects. "this bad framework is tolerable because you can apply for it not to apply to you"