Ironalchemy has joined #openwrt-devel
<Tusker> neggles: couldn't we add support for ubifs into the okli-loader and achieve that ?
Ironalchemy has quit [Quit: IronAlchemy]
<Grommish> My debug Octeon3 image is nearly 90MB
<Grommish> for the initramfs image
<neggles> smallest i've gotten so far is 15MB
<neggles> maybe the octeon 4.14 is smaller...
Ironalchemy has joined #openwrt-devel
<Grommish> but this seemingly ethernet driver related memleak is causing issues :(
<Tusker> "added arch-kexecboot, a kexecboot kernel based on kexecboot v0.6 and linux kernel 3.5
<Tusker> (boots linux kernels from ext2/3/4 and ubifs)"
<Tusker> looks like arch-kexecboot fork of kexecboot works with ubifs
<Grommish> If a target/linux/general/hack-* patch needs to be updated, is that submitted like any other PR?
Tapper has quit [Quit: Tapper]
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<Tusker> so, in the source code of https://github.com/kexecboot/kexecboot it looks like it support ubifs already, and use --enable-textui and --enable-timeout would make it boot up automagically
slh has joined #openwrt-devel
slh64 has joined #openwrt-devel
<neggles> well.
<neggles> Tusker: that might be an option, if it builds with the octeon sdk
<neggles> stintel: do you happen to know what the model number of the AP7060DN's flash chip you dumped is? it appears to be SPI NAND, not SPI NOR, based on the strings i'm seeing
goliath has quit [Quit: SIGSEGV]
<neggles> orrr maybe not. never mind.
<stintel> neggles: iirc I read it with flashrom which doesn't support NAND
<neggles> yeah, it's NOR
<neggles> but they've used the android-style qualcomm MSM partition table O_o
<slh> qcom-smem is quite often used on ipq806x and ipq807x (not always by OpenWrt, but for some devices it is)
<stintel> neggles: the device also has a NAND chip, but the bootloader is on the NOR
<stintel> so my plan is to build u-boot, clamp flash the nor, and then it should be easier to get OpenWrt running :P
<slh> other example, ZyXEL NBG6817, ipq8065 - 4 MB spi-nor with qcom-smem partitioning, holding bootloader and calibration data, huge 4 GB eMMC with GPT partitioning with kernel + rootfs + overlay
<neggles> slh: yep it's qcom-smem
<neggles> stintel: fair enough :P does it show anything fun on the serial console at boot?
<neggles> cmdlines in here suggest 96008n1
<neggles> it would be worth trying stopping autoboot and running 'bootipq debug'
<stintel> password protected iirc
<neggles> boo, hiss
<stintel> although I've been in a u-boot like prompt on the device
<stintel> but it was horribly crippled
<stintel> neggles: Unknown command 'bootipq' - try 'help'
<slh> stintel: btw. https://wikidevi.wi-cat.ru/Huawei_AP7060DN suggests IPQ8074, not IPQ8074A - if correct, ath11k wouldn't support your wireless anyways (yes, I'm in the same boat with my IPQ8074 device as well)...
<slh> ipq8071a, ipq8072a, ipq8074a are all supported/ supportable, but not the earlier non-A variants
<slh> /lib/firmware/IPQ8074/HK1.2/ (non A, not supported by ath11k) vs /lib/firmware/IPQ8074/HK2.0/ (A, supported by ath11k)
<stintel> well that would suck
<slh> I may (and hope to be) wrong, but it looks like another case of qca9880-ar1a vs qca9880-br4a
<stintel> unfortunately the photos I took are not clear enough
<slh> at least drivers/net/wireless/ath/ath11k/core.c only seems to know about IPQ8074/hw2.0
<neggles> stintel: ah, okay, you'd need to change to another user which will no doubt be passworded
<neggles> slh: this DTB is based off the IPQ807x AP.HK01 board, which uses the 8074A
<neggles> but it's unclear on whether they used the HK01.2 or what...
dangole has joined #openwrt-devel
<slh> neggles: HK01 was the first ipq807x devboard, there must have been variants with ipq8074, before thecurrent ipq8074a generation
<neggles> slh: you'd think so, but the compex datasheet lists the HK01.2 as having an 8074A
<neggles> gotta be a way to tell em apart
<slh> I'm pretty sure there were HK01.1 and HK01.0
<neggles> ah that'd do it
<stintel> nice! motion on the Zero W was using all the CPU all the time, on the Zero 2, it's using < 20% so not even a full core
<stintel> yikes, 4:00 again
* stintel hits the bed
<slh> ipq8074a should declare "CPU: unavail, SoC Version: 2.0" and "Soc version is not 1, changing clock offsets" in dmesg
<stintel> ahhh bummer, apparently it was rotate 90 that was causing motion to use a lot of CPU
<stintel> if I enable that on the zero 2 it's 25% all the time
<stintel> guess I'll just fix the camera orientation then :D
dangole has quit [Remote host closed the connection]
victhor has quit [Ping timeout: 480 seconds]
strobo_ has joined #openwrt-devel
strobo has quit [Ping timeout: 480 seconds]
<Namidairo> did huawei not bother releasing that model in the united states given the whole political situation
<Namidairo> does not seem to have a fcc submission
* digitalcircuit had to sidetrack into learning WireGuard 'cause family is moving, so he won't be able to use the OpenWRT VPN as a means to remote-support. WireGuard hub-and-spoke to his VPS as bounce host is now in place and hopefully reasonably secure (had to learn a lot).
<digitalcircuit> For anyone wondering about absence, I'm not dead yet! I'll get back to NBG6817 Kingston eMMC troubleshooting Soon™, hopefully.
<digitalcircuit> Just regaining sanity/catching up, and continuing to assist with family stuff, but I hope to return to debugging within the month.
rua has quit [Quit: Leaving.]
<Grommish> nbd: ping
<slh> digitalcircuit: just to be sure, I know that you worked on 5.15 already, but probably still give https://github.com/openwrt/openwrt/pull/4748 a try
<digitalcircuit> slh: Thanks for the heads up! Even though I saw that, I hadn't even considered that there's a chance it could fix the Kingston eMMC Qualcomm clocking issue.
<digitalcircuit> Definitely worth trying before I attempt any further 5.10 bisecting.
<slh> digitalcircuit: chances are low, but still worth checking
<slh> I haven
<slh> I haven't tried it yet, as I'm already full on DSA builds, so my config would be a little difficult
<slh> but it's on my todo list as well
<digitalcircuit> Yeah! If nothing else, Ansuel made sure it builds, so it *should* be more straightforward as "fetch code, wait for build, flash and test" instead of faffing about with the manual kernel bisect fun :)
clayface has joined #openwrt-devel
clayface_ has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
fda has joined #openwrt-devel
nitroshift has joined #openwrt-devel
nitroshift has quit [Read error: Connection reset by peer]
nitroshift has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
fda- has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
Grommish_ has joined #openwrt-devel
rmilecki has joined #openwrt-devel
Grommish has quit [Ping timeout: 480 seconds]
Grommish has joined #openwrt-devel
Grommish_ has quit [Ping timeout: 480 seconds]
strobo_ has quit [Read error: Connection reset by peer]
strobo has joined #openwrt-devel
Grommish has quit [Ping timeout: 480 seconds]
wulfy23 has joined #openwrt-devel
Tapper has joined #openwrt-devel
<wulfy23> wpa-cli is missing in action
<wulfy23> 67f9245ee57eaf47093275670b227c7911f485d4 ?
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
danitool has joined #openwrt-devel
rua has quit [Quit: Leaving.]
goliath has joined #openwrt-devel
<rsalvaterra> Good GMT morning. :)
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
fda- has joined #openwrt-devel
fda has quit [Ping timeout: 480 seconds]
<nitroshift> rsalvaterra, 0/
<nitroshift> _lore_, ping
<nitroshift> _lore_, i tries to remove the armada-3700 condition in mvneta, irq's still sit only on one cpu core
<nitroshift> *tries / tried
Ironalchemy is now known as Guest5505
Ironalchemy has joined #openwrt-devel
victhor has joined #openwrt-devel
Guest5505 has quit [Ping timeout: 480 seconds]
<jow> nbd: any-tld-matching should work just fine with fnmatch(3)
fda has joined #openwrt-devel
<jow> google.*
<nbd> jow: that doesn't work, because * can match any number of components
<nbd> so google.foo.xxx would match
<jow> ah that you mean, yeah sure
<jow> but even with regexp it is hariy
<jow> you need exceptions for google.co.uk for example
<nbd> right
<jow> google\.[^.]+$ wouldn't solve that either
<jow> anyhow, supporting both is fine is guess
<nbd> yes, and it makes sense to use fnmatch as default
<jow> about unpacking the LLVM tarball in phase2...
<jow> ideally it should be done automatically by the OpenWrt SDK
fda- has quit [Ping timeout: 480 seconds]
<nbd> okay, i'll look into that
<nbd> i worked around the issue for now by making qosify nonshared
<jow> or are users expected to do that manually as well when using the OpenWrt SDK ?
<jow> ideally, in phase1, the target/sdk/compile target should simply bundle the LLVM toolchain like it bundles all other toolchain executables
<jow> this will take care of copying the ELF executables, applying the preload wrapper etc.
<nbd> i intentionally left llvm out of the sdk tarball, since it would bloat it up a lot
<nbd> and not everybody needs it
<jow> by how much?
<jow> I think you wrote that already but I forgot
<nbd> 37 MB
<jow> that's acceptable imho
<jow> if it would be 370MB then I'd say no as well
<nbd> i think we should probably move the llvm-bpf toolchain tarball out of the target folder
<nbd> at least on the download site
<nbd> since it's the same for every target
<jow> SDK is around ~120MB atm
<jow> making it ~160MB is totally fine
<jow> bundling LLVM should be as simple as adding the LLVM toolchain path to the find command
<nbd> no, even easier than that
<jow> and with "move out of the target folder" you mean deduplicating and moving it once to a central location?
<nbd> simply remove the exclude line that i added :)
<nbd> yes
<jow> hmm, is it used as download somewhere so that it ends up in dl/?
<nbd> no
<jow> in this case it'll automatically end up in sources.openwrt.org
<nbd> it's host arch specific
<nbd> and binary
<nbd> so it doesn't make sense to put it in the sources dir
<jow> hm, true
<jow> I forgot; do we build LLVM on the buildbots or do we download a precompiled variant?
<nbd> we build it on the buildbots
<jow> ok
<nbd> and the result currently ends up in the target bin folder
<nbd> when we change the location, i will also add the host arch to the tarball name
pmelange has joined #openwrt-devel
<nbd> and then i want to have a script in scripts/ that downloads it from openwrt.org and unpacks it into the source tree
pmelange has left #openwrt-devel [#openwrt-devel]
<nbd> for people that want to build llvm based packages without wanting to build llvm
<jow> for buildroot users that opt to not build their own llvm?
<nbd> yes
<jow> okay
<nbd> buidling llvm takes more time than a standard openwrt build from scratch including the gcc toolchain
<nbd> fine with me
<jow> .../llvm-whatever.tar.gz
<nbd> we just need to change the phase1 code to put it there and leave it out of the target folder
<jow> the uploading part should go into the phase1 buildbot logic
<jow> it'll simply be yet another esync task
<jow> *rsync
fda- has joined #openwrt-devel
pmelange has joined #openwrt-devel
fda| has joined #openwrt-devel
fda has quit [Ping timeout: 480 seconds]
fda- has quit [Ping timeout: 480 seconds]
pmelange has left #openwrt-devel [#openwrt-devel]
rua has joined #openwrt-devel
Grommish has joined #openwrt-devel
decke has joined #openwrt-devel
rua has quit [Quit: Leaving.]
dangole has joined #openwrt-devel
rejoicetreat has joined #openwrt-devel
rejoicetreat has quit []
rua has joined #openwrt-devel
Tapper has quit [Ping timeout: 480 seconds]
Grommish_ has joined #openwrt-devel
Grommish__ has joined #openwrt-devel
Grommish has quit [Ping timeout: 480 seconds]
rua has quit [Quit: Leaving.]
Grommish_ has quit [Ping timeout: 480 seconds]
fda has joined #openwrt-devel
fda| has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
clayface_ has joined #openwrt-devel
clayface has quit [Ping timeout: 480 seconds]
danitool has quit [Ping timeout: 480 seconds]
Tusker has quit [Quit: Time wasted on IRC: 15 hours 1 minute 17 seconds]
rua has quit [Quit: Leaving.]
<fda> i just wanted to say new kernel 5.10 works for me with avm1200 / ipq40xx
<fda> jow i have installed theme bootstrap and openwrt. after an updated now the default theme switched to boostrap(non dark or light). not a big bug but maybe you want to change
<fda> switching again by (l)uci wroks
decke has quit [Quit: Leaving.]
Grommish__ has quit [Remote host closed the connection]
rua has joined #openwrt-devel
Tapper has joined #openwrt-devel
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
danitool has joined #openwrt-devel
pmelange has joined #openwrt-devel
fda- has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
fda has quit [Ping timeout: 480 seconds]
fda has joined #openwrt-devel
fda- has quit [Ping timeout: 480 seconds]
dangole has quit [Ping timeout: 480 seconds]
fda- has joined #openwrt-devel
fda has quit [Ping timeout: 480 seconds]
netprince has joined #openwrt-devel
netprince_ has joined #openwrt-devel
netprince_ has quit []
netprince_ has joined #openwrt-devel
minimal has joined #openwrt-devel
nitroshift has quit [Quit: Gone that way --->]
wulfy23 has quit [Quit: Page closed]
pmelange1 has joined #openwrt-devel
pmelange has joined #openwrt-devel
valku has quit [Quit: valku]
pmelange1 has quit [Ping timeout: 480 seconds]
valku has joined #openwrt-devel
Acinonyx_ has joined #openwrt-devel
Acinonyx has quit [Ping timeout: 480 seconds]
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
Acinonyx_ has quit [Ping timeout: 480 seconds]
Acinonyx has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
Acinonyx_ has joined #openwrt-devel
Acinonyx has quit [Ping timeout: 480 seconds]
Weissnix4711 has joined #openwrt-devel
Weissnix4711 has quit []
Ironalchemy has quit [Read error: Connection reset by peer]
Ironalchemy has joined #openwrt-devel
Tapper has quit [Ping timeout: 480 seconds]
shibboleth has joined #openwrt-devel
dangole has joined #openwrt-devel
danitool has joined #openwrt-devel
dedeckeh has joined #openwrt-devel
Namidairo_ has joined #openwrt-devel
Namidairo has quit [Ping timeout: 480 seconds]
Namidairo_ is now known as Namidairo
Tapper has joined #openwrt-devel
indy has quit []
indy has joined #openwrt-devel
Borromini has joined #openwrt-devel
clayface has joined #openwrt-devel
clayface_ has quit [Ping timeout: 480 seconds]
netprince_ has quit [Read error: Connection reset by peer]
dedeckeh has quit [Remote host closed the connection]
dedeckeh has joined #openwrt-devel
dangole has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
<hurricos> I found a really interesting thread about Marvell Prestera; describing the firmware-heavy strategy that Marvell takes to run Prestera chips with CPSS; and a comparison to a piece of hardware I wasn't aware of before, the Microchip Sparx5: https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg2472438.html
<hurricos> I'm going to be watching out for Sparx5-based boxes getting their EOL and becoming available in the secondhand market in the future. From the sound of it they're very open, it'd be interesting to see a project like OpenWrt running on them.
<hurricos> At the same time, the discussion makes compelling arguments about how the Prestera SoC's mainlining to the Linux kernel is not the same as Prestera being an open platform.
<hurricos> But at the same time, there's zero chance I'm going to go drop $4K on an evaluation board :^)
goliath has joined #openwrt-devel
Tusker has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
linusw has joined #openwrt-devel
Tapper has quit [Quit: Tapper]
Borromini has quit [Quit: leaving]
dedeckeh has quit [Remote host closed the connection]
rmilecki has quit [Ping timeout: 480 seconds]
Tusker has quit [Remote host closed the connection]
robimarko has joined #openwrt-devel
<robimarko> hurricos: Prestera used a slightly lighter approach then Spectrum, so at least it only handles networking
<robimarko> And not everything like the Spectrum does
<robimarko> But its still FW blob magic only
<robimarko> The upstream driver is really light on features, the DENT version is the one with features
<robimarko> I am unfortunate to have to deal with the DENT Prestera based switches
<robimarko> Sparx5 is really cool, its fully open without any FW
<robimarko> They have all of the register level info here
<robimarko> Its up to individual bits with description
<robimarko> And there is 600+ pages PDF datasheet describing how it actually works
<robimarko> But the issue is HW, there is none you can get
<robimarko> Trust me, we tried getting their switches to get rid of FW blobs and having to port the DENT driver
<robimarko> Only thing avaialable is the dev board you linked thats 4k USD
robimarko has quit [Quit: Page closed]