<digitalcircuit>
slh: I've built with your commit base (NOT including the PR, "openwrt-snapshot-r17525-a46fa5c3a7-ipq806x-generic-zyxel_nbg6817-squashfs-sysupgrade.bin") and the feeds.default.conf that you shared, and.. same MMC error.
<digitalcircuit>
I'm wondering if the PR is required for snapshot to boot successfully, or if there's something I'm overlooking?
<slh>
that's really weird... do you have a second for a query?
<digitalcircuit>
slh: Sure! I'm waiting for a phone call so I might get pulled away (hence delay in rebooting to test), but I'm happy to talk via query and respond when I can.
<digitalcircuit>
Hrmm. I've tried slh's build setup and still run into issues. I'll try reflashing the currently working build to see if there's something about the partitions that's messed up.
<digitalcircuit>
Reflashing the current build I'm running works just fine. Might be leftovers not getting cleared off the MMC partitions, maybe..?
<digitalcircuit>
(And by reflashing the current build, it toggled the dualboot flag, so I'm booting from the side that had been problematic)
danitool has quit [Ping timeout: 480 seconds]
goliath has quit [Quit: SIGSEGV]
shibboleth has quit [Quit: shibboleth]
<digitalcircuit>
After discussing with slh, I'm trying git bisecting from my current commit (that works) and slh's build base commit (which didn't work), ignoring package feeds since this is a lower-level "kernel can't mount root" problem. I'm trying re-building my current commit without package feeds to make sure that works properly, and if so, I'll git bisect until I find the problematic commit.
<digitalcircuit>
In case others are following along, zeroing out the alternative eMMC partitions before flashing didn't change anything, and flashing the currently working build is fine (both sets of partitions function correctly).
<digitalcircuit>
slh: The minimal build from the same commit I'm currently running as a full build has worked! Time to git bisect until I find the faulty commit (with pauses for food/sleep/etc, so.. probably starting this next build and pausing).
<Lynx->
danitool I have 3x RT3200's and one of them 5Ghz doesn't always come up on reboot and I see: "Fri Oct 1 22:03:55 2021 kern.err kernel: [ 29.325310] mt7915e 0000:01:00.0: Message 000021ed (seq 2) timeout" <--- any idea what I can do?
<Lynx->
(or anyone for that matter)
<danitool>
Lynx-: I have no idea
<Lynx->
faulty hardware?
<Lynx->
it happens about 1 in 3 reboots then I see 5Ghz hasn't come up and those timeout errors
<Lynx->
anything else I can look for?
<Lynx->
or try
noltari has joined #openwrt-devel
Lynx-- has joined #openwrt-devel
Lynx- has quit [Ping timeout: 480 seconds]
Lynx-- is now known as Lynx-
clayface_ has joined #openwrt-devel
Lynx- has quit [Ping timeout: 480 seconds]
clayface has quit [Ping timeout: 480 seconds]
goliath has joined #openwrt-devel
Tapper has quit [Ping timeout: 480 seconds]
rua has quit [Quit: Leaving.]
Tapper has joined #openwrt-devel
hanetzer1 is now known as hanetzer
victhor has joined #openwrt-devel
Lynx- has joined #openwrt-devel
<Lynx->
is it necessary/advisable to install wpad-mesh-openssl for 802.11s?
<Lynx->
(and replace wpad-openssl)
<hanetzer>
Lynx-: does not wpad's full version (forget if its wpad and wpad-full or wpad-mini and wpad, been a while) do .11s?\
<Lynx->
I don't know
<Lynx->
I'm trying to get to the bottom of problem. I have x3 RT3200's set up in a 802.11s mesh. On one of them, sometimes when I boot it up the 5ghz radio won't come up
<Lynx->
and it doesn't work until I power off then power on
chwba has quit [Remote host closed the connection]
rua has joined #openwrt-devel
PtitGNU has quit [Quit: Quassel terminated!]
Tapper has quit [Ping timeout: 480 seconds]
PtitGNU has joined #openwrt-devel
ephemer0l has quit [Remote host closed the connection]
chwba has joined #openwrt-devel
guidosarducci has quit [Remote host closed the connection]
guidosarducci has joined #openwrt-devel
pmelange has quit [Quit: Leaving.]
victhor has quit [Ping timeout: 480 seconds]
<ldir>
I can't build vnstat2 since the bump to libwebp https://git.openwrt.org/?p=feed/packages.git;a=commit;h=17126e81749e03aa4735fc169b32404203760328
Lynx- has quit [Quit: Going offline, see ya! (www.adiirc.com)]
victhor has joined #openwrt-devel
<mangix>
ldit: there's a patch for it
shibboleth has joined #openwrt-devel
schwicht_ has joined #openwrt-devel
schwicht has quit [Read error: Connection reset by peer]
<shibboleth>
has dsa been merged current master?
schwicht_ has quit []
shibboleth has quit [Quit: shibboleth]
Tapper has joined #openwrt-devel
pmelange has joined #openwrt-devel
cp- has joined #openwrt-devel
* Slimey
hits ldir with a small interstellar spaceship
cp- has quit [Quit: Disappeared in a puff of smoke]
cp- has joined #openwrt-devel
cp- has quit []
<ldir>
Slimey: what why?
<Slimey>
:P
<Slimey>
i got slap happy
<ldir>
fair enough :-D
* Slimey
pours a tuborg super light beer down the throat on Slimey
<rsalvaterra>
Oh, beer o'clock again, nice!
<Slimey>
if you dont know what beer is non alcohol
<Slimey>
:P gross
cp- has joined #openwrt-devel
shibboleth has joined #openwrt-devel
cp- has quit []
<shibboleth>
ah, and now the images i've built for ath79 also bootloop in addition to ipq806x
<Slimey>
:o
<shibboleth>
lemme guess, the decision was made to merge DSA and make bloody sure there would be no warning when sysupgrading?
<rsalvaterra>
shibboleth: You guessed… poorly.
<shibboleth>
ok, so not at all DSA-related?
<rsalvaterra>
Disclaimer: My ath79 devices are running with DSA.
<shibboleth>
or is that you take issue with my use of adjective?
<rsalvaterra>
Definitely not DSA related.
<rsalvaterra>
It's still under development, and only for qca8k, at the moment.
<shibboleth>
well, i find it hard to believe master is broken for both ipq806x and ath79 and no one but me has noticed
<rsalvaterra>
shibboleth: It's broken since when, do you have any idea? It's working for me (built an image for my Archer C6 v2 yesterday).
<shibboleth>
since yesterday 1800 GMT when i first tried
<rsalvaterra>
And for which subtargets?
<shibboleth>
c2600, ad7200, archerc7 v2, v4
<shibboleth>
boot proceeds to when the switch LEDs come on, then reboot
<shibboleth>
failsafe works
<rsalvaterra>
shibboleth: Are you doing your own builds?
<shibboleth>
yes
<rsalvaterra>
Have you dircleaned since the musl merge?
cp- has joined #openwrt-devel
<shibboleth>
fortunately failsafe works, onlike the last two times in the last three months someone merged code that "forgot" to connect the cpu to eth0/eth1
<rsalvaterra>
I find it hard to believe the C7 v2 is broken, since mangix uses it all the time.
<shibboleth>
i've noticed new toolchain dirs
<shibboleth>
which... dirs should i clean?
<rsalvaterra>
make dirclean
<shibboleth>
i usually pull master and then rsync -avtP --delete to the buildbox
<rsalvaterra>
Since the musl merge, you need to rebuild *everything*.
cp- has quit []
<shibboleth>
ok, i assumed --delete would take care of that. i do get to keep stuff in buildroot/dl (kernel, firmware, etc), right?
<rsalvaterra>
The move to 64-bit time_t is a complete ABI breakage.
cp- has joined #openwrt-devel
<shibboleth>
well, that at least gives me hope that we're not talking gremlins on stuka-pills here
<ldir>
mangix: ah, yes, found the PR - will test & merge if ok
<shibboleth>
ok, i ran dirclean, but there are still huge toolchain dirs in stagin/build_dir?
<shibboleth>
i'll nuke them manually
<rsalvaterra>
shibboleth: It only does a complete clean for the current target/subtarget.
<rsalvaterra>
You also have make distclean, which will clean everything, including your .config.
<mangix>
shibboleth: 2 port DSA is totally broken. Remove that patchset.
<shibboleth>
is it currently in main master?
<mangix>
No
<mangix>
Neither is 5ht ath79 DSA conversion
<mangix>
*the
<ldir>
bloody hell - macbook panic'd mid build - what a pain in the arse
<ldir>
make clean & start again.
<mangix>
ldir: since you use macOS, some guy has an issue building the tool chain.
<shibboleth>
ok, distclean, now building for archerc7,v2
<shibboleth>
and now for the dramatic element of... suspense
<shibboleth>
meanwhile, kinda offtopic but since there are a few ppl present: i've been asked to look at some donated dell servers with broken idracs
<shibboleth>
seems the on-mobo emmc is stuck in some weird state where it's read-only
<shibboleth>
if i attach to the debug uart and don't interrupt the idrac uboot both read and write fails, if i interrupt uboot read works just fine, but erase/write won't
<shibboleth>
anyone come across anything similar?
<shibboleth>
somewhere down the line i'll have to attempt to build openwrt for the idrac and tftpboot it, the embedded recovery environment is pretty limited in terms of being able to debug the issue
<shibboleth>
shorter version: anyone have tips/tricks with re to apparently corrupted emmc?
<mangix>
shibboleth: short of throwing in trash?
<shibboleth>
well, thing is, it's not corrupted
<shibboleth>
i've been able to write the emmc contents to a regular sd-card, nothing wrong
<mangix>
Besides random Google searches, I can't really help.
<PaulFertser>
shibboleth: with eMMC issues I had on my ac100 netbook I think this can help: 1. disconnect all the power for a minute; 2. Write bs=4096 from /dev/zero to faulty block on eMMC; 3. run blkdiscard on whole device. But if it's already detected as R/O right from the bootup I wouldn't expect any of that to help.
<shibboleth>
yeah, indeed
<shibboleth>
i was hoping for some kernel flag-fu to instruct the emmc to reset/clear itself
<mangix>
this is somewhat worrying. I'm removing this driver but this driver is also used for the builtin switch for various ath79 devices
<mangix>
I wonder if getting rid of the compatible entry is enough...
schwicht has joined #openwrt-devel
<mangix>
Ansuel: what's this expensive polling mode?
<Ansuel>
dsa make multiple call every sec to check port status... i remember swconfig didn't do it but i could be wrong...
<ldir>
mangix: the build env on macos requires a number of components not supplied with the OS - in my case they are provided by 'homebrew' - based on the relatively 'early' failure to build a toolchain my guess is that some pre-requisite component is missing.
<mangix>
Ansuel: could this be related to the faster speed of qca8k?
<Ansuel>
also i think the driver can be improved by finding the right wait values but currently the main priority is stability... if you tried the driver without all my patch you would get port dropping every 30 minutes and having disconnected port detected as connected LOL
<mangix>
30 minutes...
<Ansuel>
luckely they quickly merged the fixes upstream and we now have an usable driver...
<mangix>
ldir: sure. I have no idea what this guy's setup is.
<shibboleth>
rsalvaterra, yeah, distclean/nuking the toolchains worked
<mangix>
gotta love C ABI breakage
<shibboleth>
also freed up about 40gigs :P
<ldir>
for reasons I can't remember I also have a /usr/local/bin/python3 wrapper script that literally does exec /usr/bin/python3 "$@" - something about some package somewhere managed to find python2 even though it was looking for python3!
pmelange has left #openwrt-devel [#openwrt-devel]
schwicht has quit [Read error: Connection reset by peer]
Luke-Jr has quit [Ping timeout: 480 seconds]
Luke-Jr has joined #openwrt-devel
<karlp>
rsalvaterra:n you're very welcome, thanks for reminding me, I needed to convert an old apacche config to modern nginx and add a domain to a cert, very long overdue :)