dangole_ has joined #openwrt-devel
dangole has quit [Remote host closed the connection]
danitool has quit [Ping timeout: 480 seconds]
dangole_ has quit [Ping timeout: 480 seconds]
csrf has joined #openwrt-devel
minimal has quit [Quit: Leaving]
cmonroe has quit [Quit: Textual IRC Client: www.textualapp.com]
guifipedro_ has quit [Quit: No Ping reply in 180 seconds.]
guifipedro_ has joined #openwrt-devel
cmonroe has joined #openwrt-devel
srslypascal has joined #openwrt-devel
ptudor has joined #openwrt-devel
ptudor_ has quit [Ping timeout: 480 seconds]
srslypascal has quit [Ping timeout: 480 seconds]
cmonroe has quit [Read error: No route to host]
cmonroe has joined #openwrt-devel
ekathva has joined #openwrt-devel
srslypascal has joined #openwrt-devel
Grommish_ is now known as Grommish
csrf has quit [Ping timeout: 480 seconds]
ecloud_ has quit [Ping timeout: 480 seconds]
ecloud_ has joined #openwrt-devel
goliath has joined #openwrt-devel
cbeznea has joined #openwrt-devel
<KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_tegra.html has been updated. (100.0% images and 99.8% packages reproducible in our current test framework.)
goliath has quit [Quit: SIGSEGV]
goliath has joined #openwrt-devel
Tapper has joined #openwrt-devel
csharper2005 has joined #openwrt-devel
<csharper2005> rmilecki: ping. I would like some advice on a patch your *.yaml in linux.
grift_ has joined #openwrt-devel
grift has quit [Ping timeout: 480 seconds]
<stintel> neggles: that wasn't the main problem
<stintel> neggles: I mean, after switching to FIT and hitting that problem changing loadaddr is the first thing I did. loadaddr=0x10000000 wasn't enough though so ended up going for loadaddr=0x20000000
danitool has joined #openwrt-devel
<stintel> but my kernel is huge, has BTF enabled etc :)
<neggles> stintel: what's the FIT image's load address (as opposed to u-boot's temp one)
<stintel> no idea
<neggles> m300 yeah?
<stintel> yes
<neggles> 'KERNEL_LOADADDR := 0x00000000' h-uh
<neggles> it shows you when it does the image summary just before booting
<stintel> yes, I don't have that summary anymore :)
<stintel> I'd need to reboot
<stintel> let me replug the console cable to the backup m300 then I can check it
<stintel> really need to get some USB -> RJ45 console cables, then I can leave all network devices in the rack connected
<neggles> um
<neggles> `dumpimage -l` on the .fit? :P
<stintel> GP Header: Size d00dfeed LoadAddr 760e48
<neggles> ah
<neggles> yeah dumpimage eats itself when the file is longer than the header says it should be.
<neggles> h-uh. loadaddr really *is* 0x0
<neggles> how big is the uncompressed kernel?
<stintel> my biggest issue now is my main M300 not coming back after sysupgrade - first boot after sysupgrade u-boot can't even find mmc 0:1
<stintel> if I then reset, it boots, but the image on the sdcard is corrupt
<neggles> that's... odd
<stintel> so no rootfs, panic, boot loop
<stintel> much fun
<neggles> does a normal reboot/poweroff work?
<neggles> without the sysupgrade?
<stintel> -rwxr-xr-x 1 stijn users 350M apr 26 18:55 vmlinux-initramfs.debug
<stintel> -rwxr-xr-x 1 stijn users 178M apr 26 18:55 vmlinux-initramfs.elf
<neggles> would be looking for the size of Image or vmlinux without initramfs
<stintel> oh
<stintel> it's early, skipped coffee
<neggles> hehe no worries
<stintel> -rwxr-xr-x 1 stijn users 22M apr 26 18:50 vmlinux
<stintel> -rwxr-xr-x 1 stijn users 193M apr 26 18:47 vmlinux.debug
<neggles> interesting... you have a 150MB initramfs???
<neggles> what u-boot ver does it run
<stintel> WatchGuard U-Boot 2014.07 - Aug 12 2015 13:55:34
<neggles> *sigh* of course
<neggles> if you're sdbooting i'd say the fix there would be to run without an initramfs entirely
<stintel> I don't use initramfs on the sdcard
<stintel> just squash+f2fs
<neggles> right, so the kernel image for that should be 22mib not 175mib
<neggles> decompressed
<neggles> which checks out
<neggles> but if it can't find the mmc that's problematic
<stintel> that problem is magically solved after "reset"
<neggles> right, so... something during shutdown is putting the mmc controller into a state that u-boot isn't expecting?
<stintel> maybe u-boot uses some memory range somewhere that gets used after X uptime, only run into it on the main because the backup router usually doesn't do anything
<stintel> just did a reboot on the main, comes back fine
<stintel> maybe I can run some memory hogging thing on the backup and see if I can repro the issue on the main
grift has joined #openwrt-devel
<neggles> u-boot completely failing to detect the card but only from a post-sysupgrade reset... yikes
<stintel> and only on one of my 2 M300s
<stintel> that's why I suspect it's memory related
<stintel> maybe some memory region that is reserved in the WG kernel - no hints in oem bootlog though
<neggles> hrm
<stintel> OEM: [ 0.000000] Memory: 3942828K/4194304K available (8360K kernel code, 1316K rwdata, 3096K rodata, 420K init, 2047K bss, 251476K reserved)
grift_ has quit [Ping timeout: 480 seconds]
<stintel> OWRT [ 0.000000] Memory: 3979176K/4194304K available (11176K kernel code, 2284K rwdata, 4632K rodata, 3528K init, 325K bss, 215128K reserved, 0K cma-reserved)
grift has quit []
grift has joined #openwrt-devel
<neggles> hmmm
<neggles> slighty amount of missing reserved maybe
pepe2k has joined #openwrt-devel
<neggles> stintel: do you happen to have the tree you're using published somewhere
<stintel> https://git.openwrt.org/?p=openwrt/staging/stintel.git;a=shortlog;h=refs/heads/qoriq-5.15
torv has quit [Remote host closed the connection]
torv has joined #openwrt-devel
<neggles> ah it was right in front of me
<neggles> typical
<stintel> nah it wasn't there when you asked ;)
<neggles> oh :P well
<stintel> but the problem I'm having also exists with 5.10 kernel, and even without FIT image
<neggles> yeah
<neggles> um stupid question but
<neggles> have you tried a different sd card
<stintel> I haven't, as I can every time just dd the sdcard image to the sd, restore backup and reboot without issues
<stintel> also it would require to remove the thing from the rack
<stintel> short length in full rack, power cable in the back, much annoy
<neggles> I wonder if the SD card is just getting reset/powered off too quickly after having a bunch of stuff written to it
<neggles> is it the same kind as the other one?
<stintel> at some point I suspected the USB stick might be causing the problem but yesterday the USB stick wasn't plugged in
<stintel> both M300 have the same SD, yes
<neggles> does it still happen on a sysupgrade immediately after you've recovered it? i.e. is repeatable?
<stintel> I was about to try that now
<stintel> option conloglevel '8' hmmm
Tapper has quit [Ping timeout: 480 seconds]
<stintel> ** Bad device mmc 0 **
<neggles> so that's a yes then
<neggles> oh
<neggles> if you run `mmc info`
<neggles> does that work?
<stintel> too late :P
<stintel> ran reset already :P
<neggles> nooooooo
<stintel> well can sysupgrade again no worries
<stintel> actually now the SD is not corrupt
<stintel> that must have been due to the oops yesterday
<stintel> seen that gist ?
<neggles> i have not
<neggles> i was here when you linked it i think
<neggles> ah no i was asleep by then
<neggles> did
<neggles> did it oops on drop_caches????
<stintel> that's what it looks like
<neggles> what?????
<stintel> lol and now it boots fine after sysupgrade
<stintel> but I didn't let it get into active router state
<stintel> sigh
<neggles> oh right, that's overlayfs
<neggles> an oops during flushing overlayfs would definitely explain the fs corruption
<neggles> what i was looking at was, i noticed on the T20 the stock boot command runs `mmc info` before doing `ext2load`
<stintel> what's the default value for STOP if it's not set in a procd-based init script ?
<stintel> and does STOP=01 stop before STOP=99 or is that reversed?
<stintel> bird needs to stop before network does otherwise it can't send goodbye packets to peers
<stintel> messing up failover
<neggles> good q
<neggles> it looks like it's a bad idea to not define STOP
<stintel> b580ebb5a8e hints at 01
<neggles> yeah it looks like the 'you must define START and STOP' has been dropped from earlier versions of the doco
<neggles> er, dropped, but was present in earlier
<neggles> isn't ordering done via /etc/rcS.d/S__blah symlink numbering? startup order via rcS and shutdown via rcK?
<neggles> ah right just /etc/rc.d
<stintel> going to make a PR to use STOP=10 in bird2, like was the case in bird1
<mattytap__> stintel: anything specific you need me to test on my m300?
<stintel> is it running in production?
<mattytap__> nope
<stintel> if you like, use this diffconfig: https://gist.github.com/stintel/27cf088e79a4e40818ea42a89090353f run for a while, then try sysupgrade
Borromini has joined #openwrt-devel
<mattytap__> sounds good.
<neggles> maybe even just try `sync; echo 3 > /proc/sys/vm/drop_caches`
<stintel> bleergh now can't even reproduce the bird warnings
<neggles> oh damn i can get an M300 for $100
<stintel> neggles: that was the first time I've seen that really - but the console cable (or screen) isn't always attached
<stintel> ok just repro'd it
<stintel> ** Bad device mmc 0 **
<neggles> mmc info?
<stintel> < 600s uptime
<stintel> but made it into active state
<neggles> now try boot
<stintel> lol works just fine
<neggles> lmfao i knew it
<neggles> add `mmc info;` to your boot command just before ext4ls
<neggles> er ext2load
<stintel> but how do you explain it doesn't happen on the other M300? :/
<neggles> luck
robimarko has joined #openwrt-devel
<neggles> timing perhaps
<neggles> all the freescale default boot commands run mmc info before doing anything with it
<neggles> same on the LS1043, LS1046
<neggles> https://gist.github.com/neg2led/735f4ab5d039615575cc77c4bbcf336d <-- and the T20 watchguard boot commands, too
<neggles> `mmc dev 0` would probably work too
<stintel> bah
<stintel> really strange, never seen this before pushing qoriq to master
<neggles> all the RDB/QDS envs do it, for all the boards with this controller - which to me screams "WONTFIX, here's a workaround"
<stintel> does the GPL state a timeframe within which source has to be supplied ?
<stintel> I really had it with WG
<jow> stintel: iirc no. Within "reasonable" limits
<stintel> > 6 months surely isn't within those limits :)
<stintel> I'm going to forward some mails to SFC
<stintel> fuck it
<jow> the license states no limits at all
<neggles> somehow i feel like no
<neggles> my case is still open too
<neggles> at this point they must just be pulling delay tactics
<jow> GPLv2 section 3.b)
<jow> Accompany it with a written offer, valid for at least three years, to give any third party, for a charge no more than your cost of physically performing source distribution, a complete machine-readable copy of the corresponding source code, to be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange
<jow> not sure if they could claim that the written offer expired and that they're now not eligable to publish anything
<jow> s/eligable/required/
<jow> there's a lot of trolling potential
<neggles> the most recently published software update is from march this year
<neggles> so
<jow> like using a harddisk as "medium customarily used for software interchange"
<neggles> they're pretty SOL on a timeout
<jow> then mail that using snail mail
<jow> and charge USD 2000 for the mail fees and medium costs
<neggles> I don't think a HDD qualifies as a medium customarily used for software interchange, since those are typically fixed, but you could probably get away with "here's three 3.5" magneto-optical disks"
<neggles> in the US, anyway. if you tried to pull that here (AU) the courts would slap you upside the head
<jow> not arguing with that
<jow> but for that to happen it has to reach the court in the first place
<stintel> yeah I'm just going to prepare a mail to the SFC and let them handle it
<jow> being in the right and getting ones rights are sadly not the same things
<stintel> I commented on the WG case that if I don't have the sources by May 6th I'll let SFC handle it, that date is exactly 6 months after my initial request
<stintel> and added a reminder in my calendar to make sure to keep my promise
<Borromini> jow: is there a way to get LuCI to be more verbose? It looks like the wireless page is somehow not registering on one device I have
mattytap_ has joined #openwrt-devel
<Borromini> there's no menu entry under wireless and when I type the path manually I get the above ^^ clearing the caches does not help
<jow> Borromini: that's likely caused by /etc/config/wireless being unparsable (containing syntax errors)
<jow> check "uci show wireless >/dev/null"
<Borromini> ok, let me check :)
<Borromini> yep, you're right
mattytap__ has quit [Ping timeout: 480 seconds]
<Borromini> jow: fixed, thanks a bunch.
<Borromini> the wireless page doesn't show in the submenu yet, but i can call it directly, suppose that will be fixed with a reboot (uhttpd restart doesn't cut it)
* neggles has updated his watchguard ticket with some appropriate levels of snark
<neggles> Borromini: restart uhttpd
<neggles> oh
<neggles> i should've finished reading that message before i replied hey
<jow> Borromini: close and reopen the browser tab
<jow> the menu is cached in session storage
<neggles> or if you're in chrome/edge, open dev tools, right click refresh button, 'empty cache and hard reload' :D
<Borromini> yep, you're right - thanks :)
<Borromini> neggles: i'm on rusty firefox, i suppose a shift+ctrl+r does the same as that though?
<neggles> not quite
<Borromini> ok.
<neggles> open dev tools, network tab, check 'disable cache' box, then ctrl-shift-r
<neggles> i should open a feature request or see if there's an addon or something
<owrt-2102-builds> Build [#8](https://buildbot.openwrt.org/openwrt-21.02/images/#builders/1/builds/8) of `tegra/generic` failed.
<owrt-2102-builds> Build [#8](https://buildbot.openwrt.org/openwrt-21.02/images/#builders/43/builds/8) of `ipq806x/generic` failed.
f00b4r__ has joined #openwrt-devel
T-Bone has quit [Ping timeout: 480 seconds]
Borromini has quit [Quit: leaving]
<owrt-2102-builds> Build [#8](https://buildbot.openwrt.org/openwrt-21.02/images/#builders/65/builds/8) of `lantiq/ase` failed.
dansan has quit [Ping timeout: 480 seconds]
dangole_ has joined #openwrt-devel
<rmilecki> nick[m]1234: why would you want/need bcm53xx initramfs images build by default?
<rmilecki> csharper2005: if you ask a question it may be easier to help you :)
* rmilecki is back after 2 days semi-off
neggles has quit [Quit: bye friends - ZNC - https://znc.in]
neggles has joined #openwrt-devel
<stintel> wb rmilecki
<owrt-2102-builds> Build [#8](https://buildbot.openwrt.org/openwrt-21.02/images/#builders/26/builds/8) of `lantiq/xrx200` failed.
neggles has quit [Quit: bye friends - ZNC - https://znc.in]
neggles has joined #openwrt-devel
neggles has quit []
neggles has joined #openwrt-devel
<owrt-snap-builds> Build [#535](https://buildbot.openwrt.org/master/images/#builders/37/builds/535) of `sunxi/cortexa7` completed successfully.
grift_ has joined #openwrt-devel
dangole_ has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
grift has quit [Ping timeout: 480 seconds]
T-Bone has joined #openwrt-devel
f00b4r__ has quit [Ping timeout: 480 seconds]
<mrkiko> rmilecki: hi!
<mrkiko> rmilecki: nice to have you back! :)
<rsalvaterra> aparcar[m]: WireGuard throughput on the Belkin. :) https://paste.debian.net/1239170/
<mrkiko> rsalvaterra: great!
<rsalvaterra> It's insane. This thing eats my Omnia for breakfast.
<stintel> what Belkin is that ?
<mrkiko> rsalvaterra: it's the RT3200, right?
<rsalvaterra> mrkiko: Yep!
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<mrkiko> rsalvaterra: I guess you're on UBI layout, right? Even because there is no alternative. Are you using it's wi-fi?
<rsalvaterra> mrkiko: Went directly from first boot of OEM firmware to UBI layout. :)
<mrkiko> rsalvaterra: can you report me the output of dmesg|grep -i sanity
<rsalvaterra> And this is wired. I wanted to measure the raw CPU power.
<mrkiko> rsalvaterra: ehehe, I am a big fan of that device
<rsalvaterra> Yeah, I saw the message.
<rsalvaterra> [ 0.066573] CPU features: SANITY CHECK: Unexpected variation in SYS_CNTFRQ_EL0. Boot CPU: 0x00000000bebc20, CPU1: 0x00000000000000
<mrkiko> rsalvaterra: do you think the power supplier of the omnia is compatible with the RT3200? I think so, the jack is compatible, but can you look the voltage / amperage for me?
<rsalvaterra> Yesterday me and dangole fixed a couple of things (the GICv2 detection, for example).
<rsalvaterra> mrkiko: 12 V, 2 A.
<rsalvaterra> 24 W.
<mrkiko> rsalvaterra: so I can safely use omnia power supplier for the RT3200?
<mrkiko> rsalvaterra: for the fixes, amazing! Are they going to improve stability or cpu usage?
<rsalvaterra> Check the polarity first.
<rsalvaterra> I haven't checked. If it's the same, sure, you can use the same PSU.
<dangole> the RT3200 has + on the inside of the barrel and - on the outside. i'd be surprised if it's different for the omnia
<mrkiko> rsalvaterra: I am a blind person so I can't check all of the parameters myself, but for the polarity I know it's correct because the device seems to work fine. Still I needed some assistance to confirm volts and amperage are sufficient to operatecorrectly
<rsalvaterra> mrkiko: Oh, I had no idea, sorry about that.
<rsalvaterra> In any case, like dangole said, they're compatible, yes.
<dangole> 12V is fine, the RT3200 takes about 1.2A peak and comes with 12V 2.0A power supply
<rsalvaterra> dangole: I played with PCIe payload tuning this morning. :)
<mrkiko> dangole: thanks! So I can use both !! The omnia supplier has a more confortable form factor
<rsalvaterra> The bus accepts a packed/request size of 256 bytes, double the default. :D
<mrkiko> rsalvaterra: dangole: sorry for the questions about power supplying, but I didn't want to fray the device long term
<dangole> rsalvaterra: very interesting
<dangole> mrkiko: better safe than sorry ;)
arinc9 has joined #openwrt-devel
shibboleth has joined #openwrt-devel
danitool has joined #openwrt-devel
mattytap has joined #openwrt-devel
mattytap_ has quit [Read error: Connection reset by peer]
dedeckeh has joined #openwrt-devel
arinc9 has quit [Remote host closed the connection]
arinc9 has joined #openwrt-devel
arthur_melo has joined #openwrt-devel
arinc9 has quit [Quit: Page closed]
<csharper2005> rmilecki: hi! Glad to see you here again. I'm working under upstreaming of sercomm-partitions patch (you reviewed my PR a month ago) and guys from linux recommended me to add bindings in linux doc. You're the maintainer of these yaml files. What do you think about these changes? https://github.com/csharper2005/scmap_temp/commit/3be7d5100e71f78d7b5a30b2b4dbf7e1bbee2cfc
robimarko_ has joined #openwrt-devel
robimarko has quit [Ping timeout: 480 seconds]
Borromini has joined #openwrt-devel
<rsalvaterra> mrkiko: For the record, I'm now powering the RT3200 with the Omnia brick. :)
dangole has quit [Ping timeout: 480 seconds]
<rmilecki> csharper2005: i think I was discussing that with Rob, i'll see if your change matches Rob's opinion
<rsalvaterra> Aw, yiss! 5.15.36 is out.
<rmilecki> csharper2005: i found it, it's [PATCH] dt-bindings: mtd: partitions: convert Broadcom's TRX to the json-schema
<rmilecki> csharper2005: so that is different set of YAML files (trx & ns-firmware) but Rob suggested to merge them
<rmilecki> csharper2005: so your solution with extending fixed-partitions.yaml seems correct
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<rsalvaterra> s/out/tagged… the atom feed hasn't been updated yet. :P
<rmilecki> csharper2005: I think you shouldn't do "(for using with sercomm,sc-partitions compatible only)"
<rmilecki> if that property is for specific case, you should define it so
<rmilecki> it makes "resets" property available only for a specific compatible string
<owrt-2203-builds> Build [#30](https://buildbot.openwrt.org/openwrt-22.03/images/#builders/31/builds/30) of `bcm27xx/bcm2710` completed successfully.
pepe2k has quit [Remote host closed the connection]
<mrkiko> rsalvaterra: :D :D
<mrkiko> rmilecki: hi
<mrkiko> rmilecki: I answered your mail - hope the answer was clear enough
<rmilecki> mrkiko: hey, that description needs to go into commit if it's meant to be applied
<rmilecki> so people who look at it later can understand why it was reverted
<rmilecki> and if they try again - they make sure to do it right that time
<rmilecki> mrkiko: but ideally that should be debugged properly, to really understand why that commit broke anything
<rmilecki> maybe it's driver no checking for the "pre-calibration" string? is there a typo there?
<rmilecki> did you maybe check a driver sources for that?
<mrkiko> rmilecki: infact, the intent of the patch wasn't to be committed necessarily, just to help spark out discussion. That said, I would hope gl-b2200 can be supported in 22.x with all three radios working
<mrkiko> rmilecki: no, i did not look at it
cbeznea has quit [Quit: Leaving.]
<mrkiko> rsalvaterra: I am curious to see the fixes land in master
<mrkiko> rmilecki: what actually I checked was the file requested by driver was there, and my impression is it was present. But I may well be wrong
<rsalvaterra> mrkiko: The GICv2 fix will probably come through a stable kernel update. Seems eligible. Otherwise, we'll just apply the patch ourselves. https://patchwork.kernel.org/project/linux-arm-kernel/patch/YmhNSLgp/yg8Vr1F@makrotopia.org/
cbeznea has joined #openwrt-devel
<rmilecki> mrkiko: try grep -r '"calibration"' drivers/net/wireless/ath/ath10k/
<mrkiko> rsalvaterra: thanks!!
<rmilecki> mrkiko: use that to find code responsible for getting cal data
<rmilecki> mrkiko: check if driver also checks for "pre-calibration"
Borromini has quit [Quit: Lost terminal]
<mrkiko> rmilecki: oh, I found only the occurrence: drivers/net/wireless/ath/ath10k/core.c: ret = ath10k_download_cal_nvmem(ar, "calibration");
<rmilecki> rmilecki: check that file for "pre-calibration"
<robimarko_> calibration should only be used for Wawe-1 cards
<robimarko_> pre-calibration is for Wawe-2
<csharper2005> rmilecki: Thanks a lot!
<rmilecki> robimarko_: thanks
<rmilecki> mrkiko: is that Wawe-1 or Wawe-2 card?
<mrkiko> robimarko_: rmilecki: thanks!
<mrkiko> rmilecki: wave-2
<rmilecki> csharper2005: of cours you also need a proper commit subject + body with change description
<rmilecki> mrkiko: so that should work...
<rmilecki> it may mean ath10k incorrectly treats your wave-2 card as wave-1 and looks for "calibration" instead of "pre-calibration"
<mrkiko> rmilecki: at least, this is what I understand from commit message of 80d34d9d593865248bf5a23794e9163895140de7
<rmilecki> mrkiko: you should check for code paths in ath10k/core.c and try to understand why ath10k doesn't use "pre-calibration" for your wave-2 car
<mrkiko> rmilecki: thanks!
<rmilecki> mrkiko: np, i provided you only an idea what to look for
<rmilecki> the real fun starts now with actual debugging ;)
<csharper2005> rmilecki: of course)) link was for illustration purpose only
<rmilecki> csharper2005: great :)
<robimarko_> It should detect it as v2 if the PCI ID in the original commit is correct
<robimarko_> #define QCA9888_2_0_DEVICE_ID(0x0056)
<robimarko_> But yeah, only was it so see what ath10k is requesting for your board
<mrkiko> robimarko_: but from what I can understand looking at the code, the driver tries to load pre-calibration data and calibration only after
ekathva has quit [Quit: Leaving]
<mrkiko> robimarko_: but this doesn't seem like a problem with missing file
<robimarko_> That link is missing the custom board-2.bin
<robimarko_> Probably using the buildbot generated initramfs
<mrkiko> robimarko_: no, was an image I built
<robimarko_> Well, the BDF is there in ipq-wifi
<mrkiko> robimarko_: and since it works with my revert applied, I guess the custom board-2.bin is there
<robimarko_> It makes no sense for QCA9988 to not work with pre-cal
<robimarko_> Whats the error you get with pre-cal?
<rsalvaterra> Oh, my…! No gc-sections fo the arm64 kernel…? o_O
<rsalvaterra> I need to fix that…
<robimarko_> mrkiko: BoardNames[0]: 'bus=pci,vendor=168c,device=0056,subsystem-vendor=0000,subsystem-device=0000,variant=GL-B2200'
<robimarko_> This is not correct in the QCA9888 BDF
<robimarko_> Since its Wawe-2 and pre-cal is used it has BMI ID
<mrkiko> so the problem is in BDF file?
<robimarko_> So it should be: bus=pci,bmi-chip-id=0,bmi-board-id=16,variant=GL-B2200
<robimarko_> mrkiko: Looks like it, whats the error you get?
<robimarko_> The one you linked?
<mrkiko> robimarko_: yes
<robimarko_> Then its incorrect BDF
<mrkiko> robimarko_: how may I fix that?
<robimarko_> That is why you are only seeing it with pre-cal despite that being correct
<robimarko_> Simply unpack the board-glinet_gl-b2200.qca9888
<robimarko_> And repack it with the correct name
<mrkiko> robimarko_: ok, I never did this, so I should understand where to start
<robimarko_> Use -e board-glinet_gl-b2200.qca9888
<robimarko_> That will extract the BDF and provide you with a board-2.json
<robimarko_> And bus=pci,vendor=168c,device=0056,subsystem-vendor=0000,subsystem-device=0000,variant=GL-B2200.bin
minimal has joined #openwrt-devel
<robimarko_> Rename bus=pci,vendor=168c,device=0056,subsystem-vendor=0000,subsystem-device=0000,variant=GL-B2200.bin to bus=pci,bmi-chip-id=0,bmi-board-id=16,variant=GL-B2200.bin as the driver wants
<robimarko_> And update the board-2.json accordingly
<robimarko_> Then you just use -c board-2.json and get the new BDF packaged
<mrkiko> wondering what else is incorrect withok, will come back once I have a working .bin
rua has quit [Ping timeout: 480 seconds]
<arnd> rsalvaterra: I am not currently planning to revisit this, but feel free to pick it up and address the remaining issues
csharper2005 has quit [Quit: Page closed]
<stintel> > Network gateway. If omitted, the gateway from the parent interface is taken if any, otherwise creates a link scope route; if set to 0.0.0.0 no gateway will be specified for the route
<stintel> "the gateway from the parent interface is taken if any" doesn't seem to work with dhcp :/
<rsalvaterra> arnd: Oh… I don't know which sections I should KEEP(), but I'll see what I can do…
<dwfreed> stintel: because the static route is set up before dhcp finishes
<stintel> yeah pretty annoying
<stintel> I have some test devices (rpi4) that I shuffle around all the time, they connect via wifi but I want mgmt network via wired, but also that is shuffled around, so the only workable option is to use DHCP
<stintel> but then with defaultroute 0 on the wired part you can't really access things properly :/
<arnd> rsalvaterra: if you run into specific issues, you can ask me on #armlinux (libera.chat), most of the other folks that might be interested should be there
rua has joined #openwrt-devel
<stintel> I would argue that netifd knows the parent interface uses dhcp, and should wait until it's been set up before adding static routes
<dwfreed> yeah, static route at least shouldn't be set until the interface has an IP
<dwfreed> which for dhcp comes from a proto update
<dwfreed> (which should nominally resolve the issue for dhcp and a few other protocol handlers)
<mrkiko> robimarko_: http://ix.io/3WnU
<robimarko_> mrkiko: Yeah, that should be it
<mrkiko> robimarko_: ok, replacing the binary in the openwrt tree and reflashing without the revert - will keep you up to date
<mrkiko> robimarko_: so am I actually running 'uncalibrated' hardware?
<mrkiko> for the PCI device at least
<robimarko_> Well, kind of
<mrkiko> robirmor, in other words, why does it work without pre-calibration?
<robimarko_> Well, you were feeding it calibration
<robimarko_> And thus it used PCI ID based name to get the BDF
jlsalvador has quit [Read error: Connection reset by peer]
<mrkiko> robimarko_: ok, but if the name was wrong, why does the problem only surface with pre-calibration?
<robimarko_> Instead of using BMI ID-s that are used in Wawe-2
<mrkiko> ok, so it' different method of detecting what data goes to what hw
<robimarko_> Yeah
<robimarko_> Cause BMI was only introduced with Wawe-2 HW
jlsalvador has joined #openwrt-devel
<mrkiko> does the qca4019 data for that device look sensible in your opinion?
<neggles> why does this watchguard's device tree have two maxim T1/E1 transceivers in it
<mrkiko> I guess yes but...
<robimarko_> It probably is, cause I see pre-cal being used and BMI ID-s are there
<robimarko_> You can check the BDF name it requests in the debugfs
Tapper has joined #openwrt-devel
<mrkiko> robimarko_: thanks!
<mrkiko> robimarko_: on a totally unrelated topic - once it's rebased I'l also test DSA on this device
dangole has joined #openwrt-devel
<KGB-1> https://tests.reproducible-builds.org/openwrt/openwrt_mediatek.html has been updated. (100.0% images and 99.8% packages reproducible in our current test framework.)
<mrkiko> robimarko_: ok, now it works
<robimarko_> Great
csrf has joined #openwrt-devel
srslypascal is now known as Guest2980
srslypascal has joined #openwrt-devel
Guest2980 has quit [Ping timeout: 480 seconds]
<mrkiko> robimarko_: should I submit a patch as normal or is there a particular procedure to follow?
romany has quit [Quit: The Lounge - https://thelounge.github.io]
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
romany has joined #openwrt-devel
goliath has quit [Ping timeout: 480 seconds]
mattytap has quit [Ping timeout: 480 seconds]
<mrkiko> robimarko_: any tag(s) you want me to add?
Tapper has quit [Ping timeout: 480 seconds]
goliath has joined #openwrt-devel
<robimarko_> mrkiko: I dont have commit rights, so I dont set any standards
<robimarko_> Just adding fixes tag should be enough along with a proper description
<csrf> weird question somewhat related to dev: has anyone ever wired up an (FTDI) USB-TTL adapter so that it can reset the SoC? Was thinking about doing something like this in order to facilitate image loading (over serial)
<csrf> or, is there a better/easier way to do it? basically, I need to be able build & flash my test images, and reboot the board remotely
<csrf> I can also have someone hook up the test clip to the SPI flash chip, but figured it wouldn't work for what I'm trying to achieve
noltari has quit [Remote host closed the connection]
noltari has joined #openwrt-devel
<robimarko_> csrf: You can probably reuse one of the RS232 signals if that FTDI board exposes them
f00b4r__ has joined #openwrt-devel
<robimarko_> But it will need to be hooked up to the CPU reset
<csrf> my idea was to load images over serial from uboot console, tie the DTR line from the FTDI adapter to the reset line (at the reset button) to GND, and then toggle it programmatically
<robimarko_> Yeah, thats I had in my mind as well
<csrf> *toggle it to GND programmatically
<robimarko_> I think Arduino basically does that for UART programming
<csrf> arduino does it a little bit different; they use a 'bounce' circuit to create a pulse anytime the DTR line toggles
<csrf> I would have to manually toggle the line & bounce it myself (using picocom or whatever)
<robimarko_> Oh yeah, they have a capacitor for that
<csrf> just worried i might draw too much current or something by pulling it to GND, dunno
<csrf> robimarko_, yup exactly
<csrf> though I figure the reset button circuit should be high impedance, right?
<robimarko_> If you tie it to the CPU reset then you shouldnt be able to damage anything
<robimarko_> It has to be High Z anyway
<robimarko_> I have boards that tie CPU reset to a micro push button directly and that works
<csrf> i have to have my 'remote hands' tech look over the reset button circuit then
<csrf> maybe it's not even tied to CPU reset
<csrf> might just be a regular GPIO
<csrf> ughh
<csrf> getting the tech to solder directly to the CPU pin is going to get messy
T-Bone has quit [Ping timeout: 480 seconds]
danitool has joined #openwrt-devel
<KGB-0> https://tests.reproducible-builds.org/openwrt/openwrt_kirkwood.html has been updated. (100.0% images and 99.8% packages reproducible in our current test framework.)
mattytap has joined #openwrt-devel
csrf has quit [Ping timeout: 480 seconds]
<owrt-snap-builds> Build [#534](https://buildbot.openwrt.org/master/images/#builders/25/builds/534) of `rockchip/armv8` completed successfully.
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<owrt-snap-builds> Build [#563](https://buildbot.openwrt.org/master/images/#builders/8/builds/563) of `x86/64` failed.
Borromini has joined #openwrt-devel
dedeckeh has quit [Remote host closed the connection]
<owrt-snap-builds> Build [#539](https://buildbot.openwrt.org/master/images/#builders/27/builds/539) of `mpc85xx/p1020` completed successfully.
Tapper has joined #openwrt-devel
shibboleth has quit [Remote host closed the connection]
shibboleth has joined #openwrt-devel
dansan has joined #openwrt-devel
mattytap_ has joined #openwrt-devel
csrf has joined #openwrt-devel
mattytap has quit [Ping timeout: 480 seconds]
<nick[m]1234> someone mentioned me because of building initramfs default, however, it is not important anymore.
Borromini has quit [Quit: Lost terminal]
cbeznea has quit [Quit: Leaving.]
cbeznea has joined #openwrt-devel
shibboleth has quit [Quit: shibboleth]
csrf1 has joined #openwrt-devel
csrf has quit [Ping timeout: 480 seconds]
robimarko_ has quit [Quit: Leaving]
rua has quit [Ping timeout: 480 seconds]
csrf1 has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
cbeznea has quit [Quit: Leaving.]
c0sm1cSlug_ has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
c0sm1cSlug has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
c0sm1cSlug has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
<jow> hmm.... seems there's some buildbot specific packaging quirks
rua has joined #openwrt-devel
<jow> contains "Package: ucode" ... "Depends: libc, libucode"
c0sm1cSlug has joined #openwrt-devel
<jow> however it should be "Depends: libc, libucode20220322" (including the ABI version)
<dwfreed> only the ucode package is like that
<dwfreed> the others that depend on libucode have the abi version
<dwfreed> {rpcd,uhttpd}-mod-ucode being the others
<jow> yeah
<jow> what's worrying me is that locally it is all fine here
<jow> my locally generates Packages contains "Depends: libc, libucode20220322" also for ucode itself
<jow> so there's some non-deterministic (I guess racy) behaviour
csrf1 has joined #openwrt-devel
T-Bone has joined #openwrt-devel
f00b4r__ has quit [Ping timeout: 480 seconds]
Tapper has quit [Ping timeout: 480 seconds]
<KGB-0> https://tests.reproducible-builds.org/openwrt/openwrt_x86.html has been updated. (100.0% images and 99.8% packages reproducible in our current test framework.)
c0sm1cSlug has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
zatwai_ has joined #openwrt-devel
zatwai has quit [Quit: ZNC 1.8.2+deb2+b1 - https://znc.in]
embargo has quit [Ping timeout: 480 seconds]
rua has quit [Ping timeout: 480 seconds]
xes has quit [Ping timeout: 480 seconds]
embargo has joined #openwrt-devel
rua has joined #openwrt-devel
xes has joined #openwrt-devel
srslypascal has quit [Quit: Leaving]
srslypascal has joined #openwrt-devel
csrf1 has quit [Ping timeout: 480 seconds]
csrf1 has joined #openwrt-devel
dangole has quit [Ping timeout: 480 seconds]
arthur_melo has quit []
minimal has quit [Quit: Leaving]