<Ansuel> (" are not delimiter they are part of the variable)
csrf1 has quit [Ping timeout: 481 seconds]
<neggles> i guess that's to stop the ole shellscript parameter splitting problem from happening
<neggles> imagine if sh just assumed every ${variable_name} invocation had double quotes around it unless you put like... anti-quotes around it
<neggles> no idea what you'd use as an anti-quote
<Ansuel> well normally you just escape them to treat them litteral
<Ansuel> \"
<neggles> yeah but that's how you end up with "${var} ="'"'"${another_one}"'"'" && more text"
<Ansuel> well perl still have a meme with string and variable concatenation mixed with hash
<neggles> the "'"'" and '"'"' patterns hurt me on a metaphysical level u kno
<Ansuel> print("This variable is $hash{"key"}.\n")
<Ansuel> this will result in compilation error
<Ansuel> print('This variable is $hash{"key"}.\n') won't exapnd the $hash
<Ansuel> solution ?
<Ansuel> print(" ".join("This variable", $hash("key"), "\n"));
csrf1 has joined #openwrt-devel
<Ansuel> :D
<neggles> (don't run that)
<Ansuel> O_O
<neggles> i believe it forms a 'sudo rn
<owrt-snap-builds> Build [#680](https://buildbot.openwrt.org/master/images/#builders/10/builds/680) of `bcm63xx/smp` completed successfully.
<neggles> er, sudo rm -fr / &
<Ansuel> i'm curious to translate that mess
<neggles> I tried and didn't get very far, I am not a Perl person
<neggles> did consider spinning up a vm just so I could trace it
<Ansuel> first workflow with the tools container
<Ansuel> will it work ?
minimalist has quit []
<Ansuel> an arpa ip O_O
<Ansuel> aaaand works nice!
csrf1 has quit [Remote host closed the connection]
csrf1 has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
<dwfreed> Ansuel: probably shouldn't delete the preferred tool from the tool list
<Ansuel> dwfreed the idea is that it will be already tested and if that section has failed we can skip rechecking if the tool exist in the system
<dwfreed> except that's in the preferred_tool branch
<dwfreed> in the second run of download_cmd, it won't be in the list anymore, and line 102 will fail
<Ansuel> you are right! think i should make the tool variable global and move the select tool thing before the while at the end of the script
<Ansuel> the tool should be selected only one time doesn't make sense to make the scan everytime
<dwfreed> aye
csrf1 has quit [Ping timeout: 480 seconds]
<Ansuel> thanks for the review! totally missed the fact that download function is run more than one time for wget and curl
hanetzer1 has joined #openwrt-devel
aiyion has joined #openwrt-devel
hanetzer has quit [Ping timeout: 480 seconds]
aiyion_ has quit [Ping timeout: 480 seconds]
Ansuel has quit [Quit: Probably my PC decided to sleep or I decided to sleep.]
Gaspare has joined #openwrt-devel
aiyion has quit [Read error: Connection reset by peer]
aiyion has joined #openwrt-devel
Gaspare has quit [Quit: Gaspare]
csrf1 has joined #openwrt-devel
danitool has quit [Ping timeout: 480 seconds]
csrf1 has quit [Ping timeout: 480 seconds]
<[Pokey]> neggles: that would make sense but what doesn't is the fact it fails signature checks for official repos too
<[Pokey]> I'll be testing later today
csrf1 has joined #openwrt-devel
soxrok2212 has joined #openwrt-devel
valku has quit [Quit: valku]
soxrok2212 has quit [Quit: Leaving...]
f00b4r0 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
csrf1 has quit [Ping timeout: 480 seconds]
<\x> mrnuke: https://i.imgur.com/zy1rRx7.png ahemm, no direct flashing here aaaaa
rua has quit [Quit: Leaving.]
cbeznea has joined #openwrt-devel
cbeznea has quit [Quit: Leaving.]
rua has joined #openwrt-devel
danitool has joined #openwrt-devel
<neggles> [Pokey]: well
<neggles> depending on what the specific error is
<[Pokey]> jow has told me to switch to the snapshot repo so I am rebuilding at the moment
<neggles> i mean you don't need to rebuild for that really but go off i guess (you can just edit /etc/opkg/distfeeds.conf by hand)
<[Pokey]> neggles: Was too late anyway. I did a distclean and redownloaded the config.buildinfo. So I just adjusted the repo in the menuconfig
<neggles> *bruh*
<neggles> you only need to distclean if you're changing branches/targets or something is Horribly Horribly Broken
<[Pokey]> Generally, I make it horribly broken
<neggles> if `make -j$(( $(nproc)+1 )) world` succeeds, it's not horribly broken :P
<[Pokey]> Yea, it doesn't
<[Pokey]> Dunno why but it often just doesn't
<neggles> complains about missing keys?
<[Pokey]> nah just compilation errors on random shizzles
<neggles> did you do `./scripts/feeds update -a && ./scripts/feeds install -a && ./scripts/feeds install -a`
<neggles> (thanku, past neggles, for putting so many of these into shell aliases)
<[Pokey]> I've done the first two, why do you have a second install?
<neggles> because 9 times out of 10, something fails to install the first time around because it requires something that's further down the alphabet to be installed first
<[Pokey]> hm I haven't had that
<[Pokey]> Ah well. My build literally just completed right now so lets see if I did the repo change correctly or not
<neggles> q
<neggles> in your buildsystem root folder do five? key files exist
<neggles> i don't remember what they're called and i just did a distclean myself so oops
<[Pokey]> 4
<[Pokey]> key-build, key-build.pub, key-build.ucert, key-build.ucert.revoke
danitool has quit [Read error: Connection reset by peer]
<neggles> cool so your packages be signed
<neggles> which means your opkg is going to reject packages signed with any other key
<neggles> (e.g. the ones from the public repos)
<[Pokey]> my opkg rejects packages anyway lol
<[Pokey]> it rejects my packages
<[Pokey]> it rejects official packages
danitool has joined #openwrt-devel
<[Pokey]> ha
<[Pokey]> I broke LuCI
<neggles> [Pokey]: or did it just not install because it's not enabled by default for snapshot?
<neggles> \x: got your linksys then? :P
<\x> there we go neggles
<\x> yup
<\x> just now
<neggles> straight for the console
<neggles> nice
<\x> not really
<\x> i just opened it to check
<\x> the meme on chinese community is that the one with jtag headers is 60x8, the one without is 6000
<\x> so i checked
<\x> then i just used the 1.8GHz build
<\x> seems it doesnt hit the expected coremark score though, I expect 6000 coremarks at 1.8GHz
<neggles> they're probably identical save for the laser etching and a few efuses, knowing qca
<\x> neggles: no console needed really to flash it with lean's modded qsdk
<neggles> oh
<\x> you can flash right from webui
<neggles> that's even more boring
<neggles> it's not fun when they don't fight >:D
<\x> its linksys
<\x> they didnt really make it hard to flash stuff to it
<neggles> i should build openwrt for these puma prototypes
<neggles> so i can poke around and work out what the DT is supposed to look like
PaulFertser has quit [Ping timeout: 480 seconds]
<\x> for now ill be using this qsdk, then once mrnuke or robi figures out mainline for this then ill flash
<[Pokey]> neggles: I'm using the 22.03.0 buildinfo with the snapshots repo configured
<[Pokey]> should be enabled
<[Pokey]> A webserver is online just missing LuCI
<neggles> fwiw
<neggles> you do not need to use config.buildinfo for userspace package compatibility
<neggles> and honestly using it as a base at all is a bad idea
<[Pokey]> The only reason I want to use that as a base is so that I get the exact setup from an official build just with the addition of the latest commits + my commit
<neggles> it's snapshot, therefore, "exact setup from an official build" is meaningless
<neggles> however
<neggles> give me 2 tics
<[Pokey]> Read: exact package configuration from official build
<\x> iw phy
<neggles> it doesn't matter which packages you build or don't build? and snapshot does not represent what release builds look like
<neggles> userspace doesn't care
<neggles> but
<neggles> i know what you're actually trying to achieve which is 'make sure i didn't break anything and am able to download and install userspace packages from the snapshot repo'
<neggles> (and in theory kmods)
<neggles> what config symbols has it generated for your device
<[Pokey]> Argh yea all I wanna achieve is the latest master building with my device support in the exact configuration it would be in with preinstalled kmods and userspace packages as if it were a prod official release
<[Pokey]> neggles: You're gonna have to point me in the right direction for that one I'm afraid
<neggles> [Pokey]: top half a dozen lines of your .config
<neggles> and, my point is that snapshot does not *have* all that much in the way of preinstalled kmods or userspace packages, in order to make it build faster; and by definition, it is not an official release, it's a snapshot of the current tree and not guaranteed to even boot
<[Pokey]> I don't know what I am looking for between the comments so here's probably easier for you: https://termbin.com/3hfy
robimarko has joined #openwrt-devel
<robimarko> \x: I see you mentioned flashing from web gui on MR7350?
<\x> robimarko yes
<robimarko> I am gonna need some details
<\x> if you cant find it, look on "CA" on lower right
<\x> on the webui
<neggles> [Pokey]: the only uncommented one ;) okay, so. to get as close to what a theoretical next-release build would look like as is practical and reasonable, wipe your .config, replace it with this fragment https://paste.neggles.dev/Qs9j7 then run make menuconfig and exit out of it
<neggles> maybe add CONFIG_BUILD_LOG=y
<neggles> do a `make -j4 download` and a `make -j25 world V=s` and ur done
<\x> robimarko: found it?
<robimarko> \x: Yeah, now it allows manual upload
<\x> tried it on 1.1.7 and 1.1.6 firmware
<\x> works
<robimarko> But, fwupd still checks header and signature
<[Pokey]> neggles: Okay, running that right now. How did you know/produce that config though? Just through long term trial?
<\x> well i dont know the specifics
<\x> they should have the recipe there
<neggles> [Pokey]: take the snapshot config.buildinfo, remove everything starting with CONFIG_TARGET, CONFIG_IB, CONFIG_SDK, CONFIG_AUTOREMOVE, CONFIG_BUILDBOT, and add the CONFIG_PACKAGE entries from 22.03 release config.buildinfo
<neggles> then add the three lines for your specific device
<[Pokey]> neggles: damn okay. Prerequisite experience required to know to do that then
<neggles> i mean you can do it manually
<\x> as for switching between the two firmwares, yeah its easy
<\x> you just fail boot two times or so
<neggles> it's mostly "remove the bits that make it build absolutely every single device, as well as the things that are mostly buildbot-specific"
<\x> so you just press the power switch
<\x> once you see the blue light
<\x> do it twice or so
<[Pokey]> Yea you gotta know what those are :P
<robimarko> \x: Yeah, it has a boot counter
<robimarko> So if you interrupt it before it finishes booting 3 times it will switch
<[Pokey]> At this point my motivation is going downhill. All I want is a working production style build so I can do my final usb work and play with my device until the next official release which includes my device support comes out
<[Pokey]> This feels like it's becoming an in-the-know kinda task for which I am largely lacking critical knowledge
<neggles> [Pokey]: you are 100% overthinking this my friend
<neggles> if your device hasn't been merged yet, you can build and play with whatever, all you have to do is get /etc/opkg/distfeeds pointing to the right place
<neggles> s/distfeeds/distfeeds.conf
<neggles> the only really important thing is to have CONFIG_ALL_KMODS
<neggles> * and CONFIG_ALL_NONSHARED
MaxSoniX has joined #openwrt-devel
<[Pokey]> neggles: I feel like my downfall here really is trying to get .config right, and getting the distfeeds pointing right seems to be a pain
<\x> i cant find a way to do vlans on this qsdk ahaha
<[Pokey]> CONFIG_ALL_KMODS seemed to be set anyway with the previous .config I have been using but I haven't known about/touched CONFIG_ALL_NONSHARED
<\x> I guess i need to gre again for a while if I want this as my gateway
<neggles> [Pokey]: config_all_nonshared turns on config_all_kmods which is nice; and you'll probably have better luck using menuconfig, where you can press ? to find out a little bit more about what an option does (usually)
<[Pokey]> neggles: Ah so NONSHARED means build unique things like for example the kmods
<neggles> [Pokey]: it means build anything the target/targets would like to have baked into their image yes
<[Pokey]> Right
<neggles> anything that's not shared between targets
<[Pokey]> Why would you not want to do that :/
<neggles> because you're building a single monolithic image for a device with 8mb of flash and you couldn't install packages into it even if you tried
<[Pokey]> almost never then :P
<neggles> preeeeetty often, actually.
<neggles> i have a couple of quite useful devices which fall into the "oh god why did they put such a tiny flash chip in here" category
<[Pokey]> Fair!
<[Pokey]> I do question though how new users are supposed to know all these little tricks :P
<[Pokey]> I can only assume they go down a similar path to myself bugging the more experienced users
<neggles> well you see, you're not a new user, you're a new developer
<[Pokey]> dev, yea :P
<[Pokey]> s/user/dev
<[Pokey]> That build just finished using your config by the way
<neggles> and uh... ask hurricos or robimarko or stintel or basically any of the other "real" OpenWrt devs how much time I've spent in here asking questions
<neggles> and so far i've been responsible for one device getting added, though indirectly for a few others
<neggles> curse that mips64 bug
<[Pokey]> I would like to document-all-teh-things but the mountain is growing
<[Pokey]> I love making guides which go off on massive tangents to explain why everything is so and what changing it means but OpenWRT. Freaking massive. I'm lost in all the things I have been told already
<neggles> it's easier if you already have some idea of how makefiles and whatnot work
<neggles> like we're starting out under the assumption that you've built a linux kernel before
<[Pokey]> You see, this is the thing. I do not know make whatsoever. I know it can run subcommands and I know the menuconfig is handy. I've built the linux kernel before, yes (see my blog https://alexhorner.cc/linux/building-barebones-linux/) but that doesn't mean I did much more than scrape together lots of other guides. I do have a fundemental chunk of knowledge missing
<[Pokey]> RIGHT
<[Pokey]> neggles: Your config has yeilded LuCI and opkg configured for snapshots
<[Pokey]> I fear hitting the opkg update button though
<[Pokey]> should be okay, right?
<\x> vlan seems to work, my bad, dsa style
<[Pokey]> signature pass signature pass signature pass. Damn, seems like it worked
hanetzer2 has joined #openwrt-devel
hanetzer1 has quit [Ping timeout: 480 seconds]
<neggles> [Pokey]: oh neat
<neggles> did not expect that to work
<[Pokey]> neggles: kmods not playing ball unfortunately
<neggles> kmods will not
<neggles> because you have edited the kernel to add support for your device ;)
<[Pokey]> They worked the first few days
<neggles> gotta pull those from your own package repo
<neggles> ah well in *that* case
<neggles> you need to be building with that ^ hash
<[Pokey]> I just added my repo
<[Pokey]> looks like opkg still prefers the official one over mine
<neggles> where'd you put it
<[Pokey]> custom feeds
<neggles> should be fine
<neggles> it doesn't care where the package came from if there's only one with a matching version
<neggles> `opkg update` doesn't complain?
<[Pokey]> Op update is happy. Signature check pass
<[Pokey]> pkg_hash_check_unresolved: cannot find dependency kernel (= 5.10.144-1-aa517221f18445071c815d295417387e) for kmod-usb-serial
<[Pokey]> + many others
<[Pokey]> wait a sec
<[Pokey]> pkg_hash_fetch_best_installation_candidate: Packages for kmod-usb-serial found, but incompatible with the architectures configured
<[Pokey]> thats new
cbeznea has joined #openwrt-devel
<neggles> what are the exact contents of distfeeds/customfeeds
* neggles sighs and boots up another sophos AP
<[Pokey]> 3 comments and
<neggles> hm
<neggles> ...do i not have a built tree left somewhere...
<[Pokey]> ouchie
<[Pokey]> Thats exactly what I have, for the removal of any doubt
<neggles> looks fine to me 🤔
<neggles> um
<[Pokey]> I'll factory reset my device again and add my feed again and see.... again
<neggles> nono
<neggles> comment out the first line in distfeeds
<[Pokey]> oh okay
<neggles> but also did you sysupgrade this without wiping config?
<[Pokey]> Nope this is a full flash
<neggles> ok good
<[Pokey]> straight from recovery
<neggles> no point factory resetting it then ;)
<[Pokey]> wow okay
<[Pokey]> commenting that out fixed it
* neggles bows
* [Pokey] applauds
<[Pokey]> but how
<[Pokey]> why did that not work yesterday xD
<neggles> remember i said CONFIG_ALL_NONSHARED builds the contents of https://downloads.openwrt.org/snapshots/targets/ramips/mt7621/packages/
* [Pokey] ripping hair out
<[Pokey]> yea?
<neggles> and you just commented out `src/gz openwrt_core https://downloads.openwrt.org/snapshots/targets/ramips/mt7621/packages`
<[Pokey]> yea
<neggles> the contents of your `http://192.168.1.2:8181/targets/ramips/mt7621/packages` directory is....
<neggles> the result of CONFIG_ALL_NONSHARED doing its thing
<KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_sunxi.html has been updated. (0% images and 99.6% packages reproducible in our current test framework.)
<neggles> opkg got confused by version clashes between your repo and the public one.
<[Pokey]> the same, yea. What I did yesterday was turn off openwrt_kmods and add my one
<[Pokey]> This time I notice there is no such feed as openwrt_kmods
<neggles> I have not seen an openwrt_kmods before, might be a 22.03/release thing
<[Pokey]> I guess so
<[Pokey]> Well I still have a fundemental lack of understanding. I couldn't document how I got here and why
<neggles> also on the makefile front, it may help to think of it as basically just being CPP
<[Pokey]> But thank you :D
<neggles> but CPP on steroids
<[Pokey]> I'm very very not CPP capable. Best I got is Arduino
<robimarko> \x: My MR7350 is using IPQ6018, confirmed by reading the SMEM ID
<neggles> ah sorry, c preprocessor
<neggles> the thing that handles turning #DEFINEs/macros into things
<[Pokey]> then I have no idea
<neggles> make is an angry regex engine that became too powerful
<[Pokey]> eek
<neggles> (the secret is that all compilers are just regex engines)
<neggles> compilers/interpreters
<[Pokey]> tokenisers, yea
<neggles> robimarko: it'd be interesting if \x's isnt :P
<neggles> but I take it IPQ6018 is the new IPQ4019
<neggles> and it is likely to be as pervasive as same
<robimarko> Well, their specs say 1.2GHz thus IPQ6000, FCC pictures confirmed the same
<neggles> are they pin-compatible?
<robimarko> Cannot be certain, but I would bet they are
<robimarko> Its probably just binning
<neggles> never could wrap my head around the differences between all the ipq40xx variants
<robimarko> IPQ40xx is simple, you had 4018 which is the base one that has fewer GPIO-s, no RGMII etc
<robimarko> 4028 is the industrial version, same thing just has higher temperature rating
<robimarko> 4019 and 4029 are the top of the range with more GPIO-s, can support more RAM and have RGMII
<neggles> i've seen ipq4010 on spec sheets here and there but not in fcc pics i don't think
<neggles> is that a thing
<neggles> [Pokey]: tbqh i've spent more time than i care to admit digging through makefiles, reading make docs, and running my own tiny little test makefiles to work out exactly wtf goes on during various stages of the build process
<[Pokey]> neggles: So i'm trying to work this one out now. What did you fundementally change compared to the buildinfo? Did you just turn on ALL_KMODS and ALL_NONSHARED?
<neggles> no that's on in those
<neggles> the main thing I did was turn *OFF* CONFIG_BUILDBOT and CONFIG_AUTOREMOVE
<[Pokey]> are those documented?
<neggles> yup
<neggles> well sorta. it doesn't actually tell you what config_buildbot does, not all of ithttps://lounge.neggl.es/uploads/a5ac3b0a12a7f768/image.png
<neggles> oops, missed a space. that particular bit is in <root>/config/Config-build.in as you can see
<[Pokey]> ah yea found it in my menuconfig
<[Pokey]> would be good if it explained it fully :P
<robimarko> neggles: I havent seen IPQ4010 anywhere
<[Pokey]> because if I am reading that correctly, if all it does is those two things it should have no impact
<robimarko> BTW, you didnt happen to find the NSS tarball?
<neggles> I knew i was forgetting something
<neggles> i have the place i think i saw it open, lets see
<neggles> [Pokey]: if you take https://downloads.openwrt.org/snapshots/targets/ramips/mt7621/config.buildinfo and remove everything starting with CONFIG_TARGET or CONFIG_PACKAGE, then diff that with the chunk i gave you (doing the same to it)
<neggles> you'll see what i flipped off
<[Pokey]> neggles: Is that why you asked me to run menuconfig? To poipulate the ones for my device afterwards?
<neggles> yes
<neggles> ...wait have you been running without doing that
<[Pokey]> No I always menuconfig and then defconfig
<neggles> ok good :P
<neggles> but also i asked you for the symbol that it had chosen for your device remember
<neggles> hence the 3 lines at the top, picking target / subtarget / device
<neggles> btw, defconfig is redundant if you already ran menuconfig, if anything do them the other way around
<[Pokey]> fair enough
<[Pokey]> neggles: Oh by sumbol you meant the device name
<[Pokey]> My mind went to debug symbols
<neggles> the bit before =
<[Pokey]> ye
<[Pokey]> not a debug symbol :P
<neggles> config symbol :)
<[Pokey]> I guess by taking 22.03.o's buildinfo I really did overcomplicate things...
<[Pokey]> your one is tiny
<neggles> it's a config fragment, the buildsystem handles expanding it out
<neggles> if you'd like to generate one from your config after changing things in menuconfig i believe it's `make savedefconfig`?
<[Pokey]> oh neat
<neggles> also do yourself a favor and run `./scripts/env` and read the help message
<neggles> mmmmmmm, versioned configs
<[Pokey]> oh yea I read lightly about env
<[Pokey]> I didn't see the point when I could just retain multiple copies of the repo :P
<[Pokey]> I think my next task is to pull another copy and do what you did from fresh as a learning activity
<\x> keks robimarko
<\x> this linksys thing overwrites the other partition when doing sysupgrade
<\x> tahaha
<\x> i now have both QSDK on both partitions
<\x> good thing i backed it up
<robimarko> Thats just great
<\x> so when upgrading best thing to do is, kick yourself to stock linksys, factory flash there, then itll kick you to the flashed partition then sysupgrade from there
Tapper has joined #openwrt-devel
<\x> but if you do sysupgrade from the sysupgrade it flashes the next partition
<\x> what a meme
<\x> hella weird
<robimarko> \x: Well, that is kind of how dual-fw is supposed to work
<robimarko> In the chinese QSDK amalgamation you can find the sysupgrade logic
<\x> i wonder if a modded bootloader is out there
<\x> where it offers a webui, kinda miss it
<robimarko> Not really, maybe GL-inet hacked in the ancient HTTP stack in there
PaulFertser has joined #openwrt-devel
<\x> well, im not used to it yet robimarko
<neggles> [Pokey]: pulling up another copy of the repo is entirely unnecessary, just a waste of bandwidth :P
<[Pokey]> :P
<neggles> `rm -fr build_dir staging_dir tmp bin` and toolchain too if you really must
<neggles> then delete .config
<[Pokey]> neggles: WARNING: your configuration is out of sync. Please run make menuconfig, oldconfig or defconfig!
<[Pokey]> I am a cursed individual
<neggles> well yeah the idea is you drop your fragment into .config
<neggles> then run make menuconfig and just save and quit
<[Pokey]> I got that running menuconfig, chose my device, then make -j25 download clean world
<neggles> did you remember feeds/packages
<[Pokey]> yep
<[Pokey]> I updated -a and did two install -a before running menuconfig
<neggles> i mean... I literally did exactly that just now
<neggles> for these cursed facebookcubes
<[Pokey]> Like I said I am a cursed individual
<\x> robimarko: to switch, it needs 4 boot failures
<[Pokey]> I will distclean, do a feed update and two installs, download the .config, remove all device targets and packages, run menuconfig and choose my device then run that make command
<\x> you watch the blue light
<\x> its blue, then becomes a white blue
<\x> flick the switch
<\x> do this 4 times
<\x> its not just two
<\x> robimarko: i tried to hookup on uart earlier, I was using a CH340G
<\x> I cant get any output
<\x> wtf
<\x> hella weird
<\x> I followed this one
<neggles> [Pokey]: you have to put the three config lines for your target into the .config before menuconfig
<neggles> or the processing logic won't work right
<neggles> it needs to know what it's building for, ordering in a .config matters
<neggles> especially when it's doing the initial defconfig expansion
<neggles> robimarko: no dice, i can't find where i think i saw it
<neggles> i do remember bringing it up at the time and i'm pretty sure you told me it was nothing of any real value
<neggles> maybe the next british teen to set their sights on a big tech company will decide to break into qualcomm :P
<robimarko> mrnuke documented it, and it works for me
srslypascal has quit [Remote host closed the connection]
srslypascal has joined #openwrt-devel
<[Pokey]> neggles: I'm right in the middle of doing a build for it. The config I ended up creating is https://pastebin.com/TqecX2jQ
<neggles> [Pokey]: how'd it go
<\x> ill find another ttl usb, this one is so old and i havent used it for a while
cbeznea has quit [Ping timeout: 480 seconds]
<[Pokey]> neggles: not good lol
<[Pokey]> I have no webserver
<[Pokey]> I went and ate lunch after that. Now I'm back
<neggles> [Pokey]: did you add luci-ssl
<[Pokey]> neggles: Not specifically no. I merely did exactly what I did before with your config just... took the snapshot buildinfo, stripped out all devices and package lines and ran menuconfig to select my devices. ALL_KMODS and ALL_UNSHARED were already set. Did you make additional changes to your .config that I neglected to check on?
<neggles> well i mean
<neggles> packages
<neggles> you should be able to just `opkg install luci-ssl` to get most of it
<[Pokey]> oh yay
<[Pokey]> need to tweak my kmods feed again one sec
<neggles> fwiw this is what i start with myself https://paste.neggles.dev/0T2Rj
<[Pokey]> woah holdup
<[Pokey]> I sent you my .config didn't I?
<neggles> yes
<[Pokey]> It has produced an image where distfeeds contains openwrt_kmods which is the one you said you'd never seen
<neggles> h-uh
<neggles> oh interesting
<neggles> that's new
<[Pokey]> but when I took your config
<[Pokey]> it did not
<neggles> how curious
<neggles> this is where env comes in handy :)
<neggles> it makes a git repo and shoves .config into it so you can diff 'em :P
<[Pokey]> I mean we can diff them without that easy enough
<neggles> actually
<neggles> what's in feeds.conf in buildsystem
<[Pokey]> I don't have a feeds.conf
<[Pokey]> Only a feeds.conf.default
<[Pokey]> ALso, with this build, my local opkg feed fails signature checks
<[Pokey]> This is so frustrating because I kinda thought this would be the easy bit. I thought how to literally build whatever the config you have me previously builds would be like, the standard, documented starting point for new devs
<neggles> heh
<neggles> your C# is showing :P
* [Pokey] hides his C#
<neggles> welcome to the land of low-level systems development
<[Pokey]> neggles: I compare your barebones config to my barebones config. You have a ton of packages selected
<neggles> [Pokey]: i told you I added all the packages from the release config.buildinfo
<[Pokey]> oh balls
<neggles> 'cause you said you wanted a release-like build
<[Pokey]> right
<[Pokey]> thats what I buggered up then
<[Pokey]> okay Pokey, relax, edit your config, rebuild, cry
<neggles> if make succeeds more than it fails
<neggles> you're doing something wrong
<neggles> "ah f#@5" *make menuconfig* *make -jN world* is the standard procedure :P
<[Pokey]> :P
<neggles> believe it or not
<[Pokey]> Your patience with me is amazing, thank you :P
<neggles> OpenWrt's build system is an absolute *dream* compared to some of the others
<neggles> if you really want to cry go look at yocto.
<[Pokey]> Nah I believe it
<[Pokey]> I can see that it is pretty well refined
<neggles> or... *shudders* android...
<[Pokey]> It just isn't C# build refined, ya know
* neggles brandishes a cross
<neggles> an android/chromeos build environment is anywhere between 50 and 250GB
<neggles> yocto usually works out in the same range
<neggles> and, well, of course not; you can't get away with .net down here
<neggles> OpenWrt builds entire system images that are smaller than a single standalone C# .Net binary
<[Pokey]> Right I am comparing my config to yours. I forgot to turn off buildbot. I forgot to turn off autoremove. I also jave a few other options set that you don't
<neggles> ahhhh buildbot will be the problem
<[Pokey]> CONFIG_SDK_LLVM_BPF=y
<[Pokey]> CONFIG_SDK=y
<neggles> ah yes,
<neggles> useless (for local development) things
<[Pokey]> and 4 more: CONFIG_TARGET_MULTI_PROFILE=y and CONFIG_IB=y and CONFIG_KERNEL_BUILD_DOMAIN="buildhost" and CONFIG_KERNEL_BUILD_USER="builder"
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<neggles> SDK builds the SDK .tar.gz, IB is imagebuilder .tar.gz, the build domain and user just change what shows up in `uname -a`
<neggles> autoremove just slows down rebuilds cause it deletes the build directories after it's done with them instead of leaving them for future reuse, buildbot will be why you got the opkg failures and the extra repo
<neggles> bruh i gave you instructions
<neggles> 19:19:40 <neggles> [Pokey]: take the snapshot config.buildinfo, remove everything starting with CONFIG_TARGET, CONFIG_IB, CONFIG_SDK, CONFIG_AUTOREMOVE, CONFIG_BUILDBOT, and add the CONFIG_PACKAGE entries from 22.03 release config.buildinfo
<[Pokey]> Hey look I did give you prewarning that I am a walking talking cursed individual
<[Pokey]> I have read so many IRC messages I'm a mind o mush :P
<[Pokey]> It is no excuse though, let me do another build
<neggles> and to be clear when i say "remove everything starting with" I mean everything starting with *any* of those prefixes
<[Pokey]> even CONFIG_TARGET_ALL_PROFILES=y
<neggles> probably keep that
<[Pokey]> :P
<neggles> but i imagine your target only has one profile
<[Pokey]> Uhh
<neggles> so it's a no-op anyway
<neggles> but, if you remove CONFIG_SDK but leave CONFIG_SDK_LLVM_BPF, CONFIG_SDK_LLVM_BPF will turn CONFIG_SDK back on
<[Pokey]> What counts as a profile? I have 2 DTSes linked to 1 DTSI
<neggles> if you only added one block to the target makefile
<neggles> you only have one profile
<[Pokey]> now i'm worried I messed up
<[Pokey]> I added two sections to m7621.mk
<neggles> oh
<neggles> you do have two profiles
<neggles> neat
<[Pokey]> I only need to build 1 really
<[Pokey]> the 32m variant
<neggles> CONFIG_TARGET_ramips_mt7621_DEVICE_zbtlink_zbt-wg1602-v04-32m=y
<neggles> covers that
<neggles> full disclosure i'm not sure that's actually what profile refers to. but.
<[Pokey]> ye
<[Pokey]> I probably don't need CONFIG_TARGET_ALL_PROFILES=y then
<neggles> ¯\_(ツ)_/¯ YMMV
<[Pokey]> Well its going and if it doesn't work I'll do another build with it :P
<neggles> that's the spirit
<[Pokey]> I am curious about CONFIG_TARGET_PER_DEVICE_ROOTFS=y
<neggles> assembles rootfs images independently per selected target when in multi-target mode
<neggles> with only the packages that device needs
<neggles> rather than building one shared rootfs with all the packages any of them need
<[Pokey]> oh
<[Pokey]> yes that doesn't make a difference to me then but I could see why that would probably want to be y
<neggles> it can save a bit of time if you are, for example, building an image for the Sophos AP55, AP55C, AP100, and AP100C
<neggles> which are the. damn. same.
<[Pokey]> Ahh
<neggles> i still can't find that repo which had the tp-link i was going to gank caldata out of...
<neggles> grr
<[Pokey]> If I was doing both profiles of my device then I should have that as n to save a smol amount of time
<[Pokey]> Caldata refers to the wifi calibration data?
<neggles> yeah
<neggles> so uh
<neggles> the AP55/55C/100/100C are a now-EOL line of wifi APs made by sophos; the AP55 is 2x2 dual band, the AP100 is 3x3 dual band, and the C versions are the indoor-sits-on-ceiling type
<neggles> except
<neggles> the AP55C and AP100C are *hardware identical*
<neggles> the AP55 and AP100 are almost identical, but they have external antennas for 5GHz, so the AP100 has one extra antenna pigtail and antenna. otherwise it's also identical. all four made by edimax.
<neggles> a handful of missing caldata bytes are the only thing stopping the 55 models from being 3x3
<[Pokey]> neggles: Wow okay so they just intentionally crippled it
<[Pokey]> Thats something I am also curious about. Do you need special equipment to *properly* calibrate these things?
<[Pokey]> (I'm also running another build btw)
<neggles> in theory
<neggles> in reality the missing bit of data isn't hugely device specific.
<neggles> I found a repo at one point which was full of various routers' caldata/firmware and i found a tp-link with the same chip that had all 3 chains
<neggles> compared the dumps, and it was literally +1 or +2 dB on each chain compared to the sophos one
<neggles> but had all 3 of them
<neggles> i meant to dump and modify the sophos data to match, then see what happened, but i didn't get around to it and now the repo is lost somewhere in my pins on github
<[Pokey]> by chain is that some sort of configuration bytes for the output radio driving circuitry? The output stage chain?
<neggles> chain being NxN
<neggles> transmit/receive chain
<[Pokey]> Ahh I see
<[Pokey]> Momento of truth
<[Pokey]> Did I config it right
<\x> so heres the tftp details am I right robimarko https://pastebin.com/raw/Hi5p1TmP
<\x> after looking on that print env thing
<[Pokey]> neggles: We have success!
<neggles> yay!
<[Pokey]> neggles: Now the question is... do enough people do what I tried to do that it would be worth documenting this whole thing, and if so, where?
<mrkiko> wiki?
<[Pokey]> mrkiko: Wiki yes lol. Where on the wiki is what I mean. Does it wanna be integrated with the existing build guide, or a separate guide, or somewhere else completely is what I'm thinking more
<[Pokey]> Or if it is even worth it :P
<mrkiko> [Pokey]: sorry, didn't understand then. :D
<[Pokey]> Its ok :P
<neggles> prolly worth it
<neggles> should probably be a part of the getting started guide
<neggles> "a sane and reasonable place to start" - but probably you don't want config_all_nonshared/all_kmods by default
<neggles> takes foreeeeeeever
<[Pokey]> neggles: Hmm, I would have thought both of those would be desirable outputs?
<[Pokey]> maybe for snapshots only?
<neggles> not when you're just trying to add support for a device and you might need to rebuild 30 times before it works right
<neggles> and it can easily mask problems
<neggles> e.g. not having a kmod in the required list for your device
<neggles> but mostly it just makes a rebuild with a tiny little kernel code change take 15 minutes instead of 2
<[Pokey]> Thats a point neggles, so maybe you'd only want that at the end as a final build?
<[Pokey]> I'm trying to think how I would word it to make the most sense
<[Pokey]> To give examples
<neggles> and thus you discover why the documentation is a bit sparse
<neggles> "if you're doing a final build just to make sure you didn't somehow break something, or you're building on a heavy-duty workstation, do this. otherwise, probably don't"
<neggles> to quote the linux kernel kconfig system
<neggles> If unsure, say Y
<[Pokey]> its a pain I can't safe drafts on the wiki that I can see
<neggles> copy the page, put it under your own user path
<neggles> edit it there
<[Pokey]> Can you have multiple user pages?
<neggles> yup
<[Pokey]> D:
<[Pokey]> thats gonna make things a bit easier
<neggles> as far as i know you can put a page at any path you like
<[Pokey]> random screenshot and an advert on a router
Behnam has joined #openwrt-devel
<[Pokey]> I just realised its hard to find pages if you create a random page
Behnam has quit [Remote host closed the connection]
Behnam_ has joined #openwrt-devel
<Behnam_> Hi
<[Pokey]> neggles: nice one!
cbeznea has joined #openwrt-devel
<neggles> Behnam_: 'sup
cbeznea has quit []
Behnam_ has quit [Quit: Page closed]
<Slimey> hm i have a wide range of devices i can submit cal data and bootloaders for
cbeznea has joined #openwrt-devel
Tapper has quit [Ping timeout: 480 seconds]
<KGB-0> https://tests.reproducible-builds.org/openwrt/openwrt_omap.html has been updated. (11.1% images and 99.6% packages reproducible in our current test framework.)
<[Pokey]> neggles: It needs a lot of work and a lot of explanation but here's the start of it https://openwrt.org/user/ahorner/build-guide
<mrnuke> hurricos: Jus found this on my engenius 2910: https://paste.debian.net/1255619/
<neggles> ahahahahahahaahah
<neggles> THEORY PROVEN
<neggles> I have a 3x3 AP55
<[Pokey]> Woop woop congrats
<[Pokey]> I love it when theres hardware capabilities to be unlocked on devices
<mrnuke> hurricos: Was a little nervous at first, until I realized the error was handled as intened :)
<neggles> now i just gotta work out how this stupid eeprom is formatted so i can make the power levels correct
<neggles> accursed hex editing
<[Pokey]> ew
<neggles> ugh checksum
<neggles> big ew
<[Pokey]> EW
<neggles> i think qca-swiss-army-knife should be able to do it
<neggles> idk about you but that sure looks like it's got six radios to me
<neggles> well two radios with three chains each
<neggles> and shockingly enough, it does!
<[Pokey]> neggles: Do you need to solder a dew more connectors though?
<neggles> nope
<neggles> PCB antennas and it has all six
<[Pokey]> oooo
<[Pokey]> how $h177y of them to just disable part of the hardware
<[Pokey]> makes me mad
<[Pokey]> I shouldn't have to pay extra for hardware I already own
<neggles> *shrug* cost saving measure on their part
<[Pokey]> but but
<[Pokey]> it isn't
<neggles> yes it is
<[Pokey]> how does that save money
<neggles> it's cheaper to build one variant than two
<neggles> quite significantly cheaper
<[Pokey]> so just build one variant AA
<neggles> yeah, build one board, then deactivate chain 3 on the cheaper ones
<[Pokey]> crappy companies
<neggles> in this one they omitted one u.fl pigtail and antenna
<neggles> connector's still there tho
<[Pokey]> If they *really* wanted to do that they should make 1 variant same model number and make it a paid one time upgrade to switch it on
<[Pokey]> eFuse it
<neggles> lol that's not how they roll
<[Pokey]> crappy
<neggles> besides unlike the APX series, these were made when sophos was still fairly new to the wifi game, so they're an Edimax board that's been lightly customized
<neggles> e.g. there's a piezo beeper on the C models, which does not work. absolutely everything that's necessary to drive it is installed on the board, *except* the resistor that connects the transistor to a GPIO
<neggles> and it has quite obviously been removed by hand
<[Pokey]> eh
<[Pokey]> why
<neggles> saves on setup costs, don't have to reprogram the pick & place or make a new solder stencil
<neggles> this same board is at least half a dozen different models from different vendors
<[Pokey]> nono why would they go out of their way to take the resistor off
<neggles> so they don't have to put the software in
<[Pokey]> they don't anyway?
<neggles> ah but they do
<neggles> if you don't drive the pin, random electrical noise inside the chip might make it start buzzing
<neggles> it's direct to a gpio, so it needs a pwm signal, and it'll just spew out noise if it's un-driven
<neggles> with the pin disconnected, the pull-down resistor stops that from happening
<[Pokey]> Don't leave the line floating then, have a single command to initialise it output, low
<neggles> but that still requires doing that
<neggles> and this is pre-devicetree
<[Pokey]> ugh it makes me so angry
<neggles> it was most likely a case of edimax having a number of prebuilt boards on hand, sophos ordering their units with no beeper, so they disabled the beeper
<[Pokey]> ;w;
<[Pokey]> you re-enabling it?
<[Pokey]> also M5Stack FTW
<neggles> nah, cbf
<[Pokey]> hah
<neggles> don't want or need a beeper and i've no idea what gpio pin it's on
<neggles> there's a micro-USB OTG port on the Cs as well, but that's missing slightly more than just the socket
<[Pokey]> I was just gonna say that
<neggles> it's also missing a tiny switchmode reg for VBUS and a few resistors
<[Pokey]> ye
<neggles> the edimax has it
<[Pokey]> If I could solder things that small I would add it but I can't
<neggles> but theres also no hole for it in the case
<neggles> and i just, do not even *begin* to care
<[Pokey]> Also... I might get my BT Home Hub 3A out again at some point and see if I can get the WAN interface working...
<neggles> i have five AP55Cs and three AP55s and that's after I gave seven 55Cs away
<[Pokey]> nobody seems to have cared about it for yeats
* [Pokey] is jealous he is not in the vicinity for freebies :P
<neggles> lantiq
<neggles> don't even try
<[Pokey]> that bad? :(
<neggles> z e r o support
<neggles> also
<neggles> its been tried
<neggles> and it's just, not even *close* to being worth the effort
<[Pokey]> bleh
<[Pokey]> Wonder if I can do an old Sky box then...
<neggles> three 100mbit and two 1gbps, 2x2 2.4GHz 802.11n, not even 5ghz
<neggles> you can literally get a better router for 15 quid off amazon
<[Pokey]> I have old kit lying around
<[Pokey]> may as well mess with it
<neggles> not a lot of point for a chip that's never getting upstream support, probably would be rejected for OpenWrt support because nobody wants to maintain it, and is effectively worthless
<neggles> won't even make a good paperweight, doesn't weigh anything :P
<[Pokey]> bleh broadcom
<[Pokey]> I wonder if I have *anything* worth modding
<neggles> > 4x100M ports and 802.11n
<neggles> it's also broadcom mips and already supported :P
<[Pokey]> supported but a complete bastard
<[Pokey]> that page points to google drive links fgs too
<\x> i hate this QSDK shit
<\x> doesnt even work with dumb AP mode
<\x> works as a router though but then SQM doesnt work
<\x> and wifi performance on that QSDK amalgamation mish mash is shit
<\x> only the cpu cores are good so far, its quite nice
<\x> works fine as a switch, vlans work well
<\x> maybe i bought into ipq60xx too early heh
<\x> and the 1.8Ghz strap for cpu doesnt seem to work, it scales up to 1.6GHz so far
<neggles> \x: just leave it in router mode, turn off the dhcp server, and plug into a lan port *shrug*
<\x> then ill just use stock
<neggles> if you want performance out of a recent qca chipset
<neggles> it's gonna be stock or robimarko's nss-enabled tree shenanigans
<\x> on stock, performance is around 700Mbps over 5Ghz with an AX200 client
<neggles> lame
<Slimey> :/
* neggles laughs in 20x20 IPQ8074
* neggles looks at bank account and stops laughing
<\x> the dual firmware shit is nice though once you understand it, just remember to keep backups of mtd14 to mtd17 2x kernel, 2x rootfs
<\x> routing on the qsdk build is hella weird, I guess no packets at all passes through the cpu
<\x> literally 0% cpu usage
<\x> no interrupts even
<\x> neggles: heres how the AP shows up on iw scan, it offers AC wifi on 2.4 too heh https://pastebin.com/raw/M1Ff0bb5
<\x> then here openssl and coremark bench compared to ipq40xx https://i.imgur.com/7tbQblL.png
<\x> I dont think its bad for 40$ though, good hardware but real shit software currently
<neggles> \x: wait it's not even AX on 2.4?
<\x> neggles: it is
<\x> but it also offers WiFi 5
<\x> thats not standard
<neggles> i mean, that's, how backward compatibility works
<neggles> no?
<\x> nope, standard is offer HT then HE on 2.4GHz
<\x> no VHT
<\x> neggles: youre right with that dumb AP mode, I removed the wan port on the bridge, it worked.
<Slimey> i have one of the hawkeye based boards from adtran to play with https://paste.fuckingjuniper.com/?5b2ba629e2932c2f#+S2E7MKrA/yzYTPkWIWAR9Ay5VmSvs7lexJwseugnB8=
<\x> can i bother you Slimey
<\x> coremark on that 8074 pls?
minimal has joined #openwrt-devel
robimarko has quit [Remote host closed the connection]
robimarko has joined #openwrt-devel
goliath has joined #openwrt-devel
Tapper has joined #openwrt-devel
* [Pokey] thinks
<[Pokey]> I wonder if anyone has got OpenWRT running on a TiVO box
<[Pokey]> silly, I know
<[Pokey]> hahahahaha
<[Pokey]> neggles: I just went to set up WiFi and theres no Wireless menu on this buil
<[Pokey]> d
shibboleth has joined #openwrt-devel
<[Pokey]> Drivers were not installed for the radios. Wonder why they weren't on this build but they were when I was using the v22.03.0 buildinfo... I literally copied the package list from that buildinfo and I do not see any other .config options related to that
danitool has joined #openwrt-devel
nezbednik has joined #openwrt-devel
<nezbednik> hey; trying my best to add a entry to the luci web interface: added a file in /usr/share/luci/menu.d and /usr/share/rpcd/acl.d but it is not showing up. any ideas?
<hurricos> mrnuke: nice :o
<hurricos> neggles: < real openwrt devs > hahahaha, very funny
<[Pokey]> hurricos: Realer than me :P
<hurricos> [Pokey]: RE: kernel vermagic, the best way I've found to make sure your version magic is right is to use the config.buildinfo from OpenWrt's build output -- https://downloads.openwrt.org/snapshots/targets/mpc85xx/p1020/config.buildinfo e.g.
<hurricos> ... and then modify it to add your board, of course.
<hurricos> oh, you were doing it right. Didn't even scroll up far enough to find it.
<[Pokey]> hurricos: https://openwrt.org/user/ahorner/build-guide Here is my current build process for my device at the moment, massively based on neggles' work (basically, fully neggles's work)
<robimarko> \x: coremark on 8074a https://pastebin.com/zahZZvQt
<[Pokey]> Just gotta work out why my WiFi drivers are not in by default with that now ehen they are with the 22.03.0 buildinfo
<hurricos> [Pokey]: What's your HEAD at?
<hurricos> Since it sounds like you're building from source with some patches
<hurricos> *sounds* like
<[Pokey]> 918e77465868fc4878ec27826dfc11ba0272f75e + 1 patch
<hurricos> What's the patch? One for your board? :p
<hurricos> iirc packages requested for inclusion by your image Makefile definition don't come with the initramfs
<[Pokey]> Should do with sysupgrade though right?
<hurricos> Should do!
<[Pokey]> didn't
<\x> robimarko: thanks
<\x> also, "auto_recovery=yes" to "auto_recovery=no" on fw saveenv will disable that switching thing?
<[Pokey]> hurricos: I'm rebuilding anyways because I just rebased to what I just told you, though I do not think i'll make any difference. If it does, whoopie!
<\x> and to set that back again if I crash I just need to get into uart then uboot right?
<hurricos> There's also some config stickiness you will have to avoid
<[Pokey]> What do you mean by that sorry?
<hurricos> if you wget the config and then make defconfig you may find you still don't get the packages you want built-in
<robimarko> \x: no idea on that, I have not tried that
<\x> https://i.imgur.com/muvUCTE.png here the ipq60xx
<\x> clearly 1.8GHz strap isnt scaling
<\x> no idea why
<hurricos> When I do board ports, I set up the definitions in my image-file and build board-specifically
<hurricos> that is, just include three lines -- target, subtarget, board-def -- in the diffconfig that you expand out with make defconfig
<[Pokey]> hurricos: why is that so?
<hurricos> Why is config so sticky? Your answer is somewhere in `find . -type f -name 'Makefile' -o -name '*.mk'`
<hurricos> Other advice: I highly recommend using `V=sc` as one of your Make arguments
<hurricos> e.g., just `nohup make -j8 V=sc &`
<[Pokey]> Ye, if it all goes horribly wrong I do usually haha
<[Pokey]> If the config "sticks" and fails to include the packages listed for my board in the makefile, what is the correct way to fix that? This is actually the first time this has happened for me
<\x> also, it seems this MR7350 has bluetooth, after going back to stovk I tried to explore that linksys thing
<\x> it asked my phone to enable bluetooth so it can find the AP
<hurricos> Is it also the first time using `make defconfig` atop a `config.buildinfo`?
<robimarko> \x: Yeah, it has bluetooth over HSUART
<[Pokey]> hurricos: it is not no
<\x> the leds are also fun to play with
<\x> sometime ill try to write a shell script for it where it changes color with the traffic on wan port
<[Pokey]> The menuconfig should update the conf with the required stuff too shouldn't it?
<\x> but thats just literally playing hehe
<hurricos> menuconfig may expand, but you should absolutely try to avoid it
<hurricos> What I'd do is, I'd start with a diffconfig which ONLY builds the one board; confirm that that works, and then do a sorted diff between the config it expands to and the all-boards config you're targeting
<hurricos> Actually, no, I wouldn't bother with `config.buildinfo`, I'd just do a single-board config and test based on that :^)
<hurricos> I've never had the OpenWrt official builds be broken for lack of testing with OpenWrt's config.buildinfo
<hurricos> also, when switching between different configs, you never need a make distclean, but you always need a make clean
<hurricos> The build documentation could be better :(
<[Pokey]> hurricos: just a heads up, not sure if it changes your recommendation at all. My guide is a bit vague as to what I do in menuconfig. I'm actually selecting my target board in there
shibboleth has quit [Quit: shibboleth]
<\x> hi [Pokey] my router is now a gaming router
<KGB-0> https://tests.reproducible-builds.org/openwrt/openwrt_lantiq.html has been updated. (96.2% images and 99.6% packages reproducible in our current test framework.)
<[Pokey]> \x: But can it run crysis?
<[Pokey]> What model is it? And i bet if you could somehow get video output it could run DOOM
<Habbie> it has video output, the last url clearly shows that!
srslypascal is now known as Guest1940
srslypascal has joined #openwrt-devel
<[Pokey]> Habbie: xD
srslypascal has quit []
srslypascal has joined #openwrt-devel
Guest1940 has quit [Ping timeout: 480 seconds]
minimal has quit [Read error: Connection reset by peer]
minimal has joined #openwrt-devel
Tapper has quit [Quit: Tapper]
<[Pokey]> Yes really not sure whats going on here
<[Pokey]> I can see the option set in .config to include my wifi drivers but even after a targetclean, then a make defconfig download clean world I am not getting it included in the image
cmonroe has quit [Read error: Connection reset by peer]
rua has quit [Remote host closed the connection]
rua has joined #openwrt-devel
cmonroe has joined #openwrt-devel
<jow> nezbednik: rm /tmp/luci-indexcache*
* [Pokey] spies jow
<[Pokey]> Signature issues gone
<jow> nezbednik: and clear your local session storage in the browser (log out & in in luci should be enough)
<jow> [Pokey]: ah, good to know
<nbd> jow: i have a draft of the phy renaming stuff in my staging tree
<nbd> still need to test DBDC cards
<[Pokey]> jow: I don't understand it really but from what I can tell it was just using a wrong certificate or something somewhere included in the build
<jow> nbd: great. is there any minipcie dbdc cards around these days? thinking about upgrading my aging apu
* [Pokey] ripping hair out
<[Pokey]> Why is running menuconfig suddenly removing all the package entries I have in my .config every time?
<jow> [Pokey]: unsatisfied dependencies?
<jow> menuconfig will purge all "unreachable" symbols
<jow> nbd: is your current staging tree the latest code?
Borromini has joined #openwrt-devel
<[Pokey]> jow: Thats not very helpful of it, because I do not know what deps I need to satisfy and it doesn't tell me. I'm also using the package list from 22.03.0 so I am uncertain why there would be an unmet dep
<jow> [Pokey]: make menuconfig, type slash (/), enter one of the package names that get removed in search dialog, press enter
<jow> it will display a full depdendency dump including the state (=y/=m etc.)
<[Pokey]> No matches found for CONFIG_PACKAGE_luci
<[Pokey]> wait
<[Pokey]> of course there isnt
<[Pokey]> sorry jow I wasted your time.
<[Pokey]> I fifn't update my feeds.
<nbd> jow: what do you mean by "the latest code"? it's my latest version, and i rebased it to master today
<nbd> jow: not sure what DBDC minipcie cards you can buy
<nbd> maybe you can find some 7915D cards on aliexpress or something
<jow> nbd: I overlooked the "in my staging tree" part of your initial message
csrf1 has joined #openwrt-devel
<nbd> ah :)
<[Pokey]> Yep sorry jow, completely my fault, working now. My next job is to work on that USB switching stuff you told me how to do like a week ago now
<[Pokey]> Yay finally a complete clean and rebuild has made my wifi drivers get in by default again
minimal has quit [Quit: Leaving]
csrf1 has quit [Remote host closed the connection]
csrf1 has joined #openwrt-devel
robimarko has quit [Quit: Leaving]
csrf1 has quit [Remote host closed the connection]
csrf1 has joined #openwrt-devel
Borromini has quit [Quit: Lost terminal]
goliath has quit [Quit: SIGSEGV]
floof58 is now known as Guest1948
floof58 has joined #openwrt-devel
Guest1948 has quit [Ping timeout: 480 seconds]
Tapper has joined #openwrt-devel
<neggles> [Pokey]: sounds like you're missing some stuff in your default packages list in the target makefile if there's no wifi still :P
<[Pokey]> neggles: is fixed!
<[Pokey]> Complete clean and rebuild and were all good
<[Pokey]> I have a working setup
<neggles> nbd: asiaRF has MT7915DAN and MT7916 (2x2 2.4GHz + 3x3:2 5/6GHz) mini-PCIe cards for reasonable-ish prices
<neggles> also [Pokey] all you need for luci is CONFIG_PACKAGE_luci-ssl :P it's a bundle
<neggles> but cool :)
floof58 is now known as Guest1950
floof58 has joined #openwrt-devel
Guest1950 has quit [Ping timeout: 480 seconds]
csrf1 has quit [Remote host closed the connection]
csrf1 has joined #openwrt-devel