danitool has quit [Remote host closed the connection]
minimal has quit [Quit: Leaving]
tSYS has quit [Quit: *squeak*]
tSYS has joined #openwrt-devel
Mangix has joined #openwrt-devel
<dwfreed>
Kufat: pastebin the contents of /home/kufat/gits/openwrt/build_dir/target-aarch64_cortex-a53_musl/root-mediatek/usr/lib/opkg/info/acme-common.postinst-pkg ?
<dwfreed>
I can tell you part of it, though
<dwfreed>
it is not performing actions on the correct path
fda has joined #openwrt-devel
<dwfreed>
it needs to be taking into account ${IPKG_INSTROOT} (and since the postinst is written in the makefile, the $ needs to be doubled there)
<Kufat>
hmm
<dwfreed>
I'm not sure how you're ending up with /etc/init.d/acme called, though
<dwfreed>
it has a uci-defaults file, so it should just skip the postinst if IPKG_INSTROOT is set
<Kufat>
so yeah, exactly right. thanks for putting me on the path to figuring out what was happening.
<fda>
ubus/libubus.so.20220615 can not be compiled for mips-fritzbox7320 anymore: "lto1: internal compiler error: compiler does not support ZSTD LTO compression". You need a big log ?
<Kufat>
...how did that ever work successfully even before the recent commit?
<Kufat>
(talking about the link I pasted, not the issue fda mentioned)
<slh>
fda: if you've been building from an unclean checkout/ used it to build OpenWrt before, you may have to clean the toolchain
<fda>
slh: it was a new directory 1 h ago
<slh>
o.k., there goes my idea ;)
<fda>
currently in another directory a build for rpi2 runs, lets see if it happens with that too...
<fda>
i dont understand what to do: "Please submit a full bug report, with preprocessed source (by using -freport-bug)."
dangole_ has quit [Remote host closed the connection]
dangole_ has joined #openwrt-devel
<dwfreed>
in that instance I'm not sure that would be useful
<dwfreed>
you'd need to have the build log of binutils, I think
<dwfreed>
figure out how it ended up with broken zstd LTO
<slh>
also run ./scripts/diffconfig.sh maybe you notice a weird config combination
fakuivan has quit [Ping timeout: 480 seconds]
dangole_ has quit [Ping timeout: 480 seconds]
hanetzer has quit [Read error: Connection reset by peer]
hanetzer has joined #openwrt-devel
fakuivan has joined #openwrt-devel
sasa has joined #openwrt-devel
sasa has left #openwrt-devel [#openwrt-devel]
<dhewg>
fda: rm -rf .ccache
cbeznea has joined #openwrt-devel
SlimeyX has quit [Read error: Connection reset by peer]
SlimeyX has joined #openwrt-devel
Tapper has joined #openwrt-devel
csrf has joined #openwrt-devel
csrf has quit [Remote host closed the connection]
valku has quit [Remote host closed the connection]
<f00b4r0>
djfe: that's ok. They should be autodetected
goliath has joined #openwrt-devel
cbeznea has quit [Quit: Leaving.]
cbeznea has joined #openwrt-devel
<djfe>
f00b4r0: the only thing I thought about was, if they maybe used something like sercomm and moved partitions to counter bad blocks. but I'm not sure whether that even makes sense on NOR flash
<djfe>
and if there was maybe a partition that could act as a partition map.
<Znevna>
morning
<djfe>
anyways if damex encounters any issues again he might want to compare the partitions to a backup to figure out if they start at the same offsets and if some bytes are broken or whatever. Apart from that I have no clue 😅
<djfe>
morning :)
<Znevna>
what device was he having issues with?
<djfe>
mikrotik hex-s
<djfe>
he ran dmesg over ssh and the input froze
<damex>
Znevna: mikrotik hex-s (nor) and bananapi r3 (emmc) - same issue with something going 'corrupt' after plugging out and plugging in sfp adapter
<damex>
djfe: yeah, it froze also when i did cat /etc/config/network
<damex>
less /etc/config/network woked though
<djfe>
the device has nor flash but still somehow openwrt got corrupted (he says), reinstall fixed it in both cases
<djfe>
interesting, but I don't have a mikrotik and don't have much knowledge so I can't really help
<djfe>
hopfeully it doesn't happen again
<djfe>
if it does happen again, then console output via serial could be interesting
<damex>
it is happing (right now) on last snapshot installed on mikrotik hap ac2.
<djfe>
was it always happening in snapshot (on kernel 5.15)?
<damex>
djfe: banana pi r3 (emmc) is on last snapshot. mikrotik hex-s is on last 22.03 release
<damex>
only reason i use snapshot on hap ac2 is dsa ;)
<damex>
and banana pi r3 is not supported on any actual release
<damex>
any should i do before i am reflashing hap ac2 back to stock -> openwrt ?
<djfe>
backing up partitions is probably not possible either on the hap ac2, is it?
<djfe>
it probably freezes when you run "dd" or later when you try to transfer via scp
<djfe>
you could try hexdumping the first x bytes though
<djfe>
about sfp: openwrt doesn't support that, yet. Or does it?
<ukleinek>
djfe: I'm unsure if reading can have strange effect, soo.
<ukleinek>
s/soo/too/
<djfe>
damex: write that down maybe/or save the full partitions and somehow compare if they look the same after a reinstall
<ukleinek>
if you don't hit bad blocks (which is given as it's NOR IIUC) it should be fine.
<damex>
ha, i just saved mtd partitions to /tmp and transfer freeze over ssh
<djfe>
dang
<djfe>
ukleinek: thx for pointing this stuff out.
<djfe>
I still have a lot to learn here
<ukleinek>
djfe: if the mtd isn't idle it might matter. The kernel might switch the NOR into another mode (e.g. to write something) and then reading from mtd might result in strange things
<damex>
so some flash partition/offsets become super slow to read
<damex>
beyond usable
<damex>
is there something like fstrim ?
noltari has quit [Quit: Bye ~ Happy Hacking!]
noltari has joined #openwrt-devel
<mrkiko>
damex: probably not; whatk ind of flash is that?
goliath has quit [Quit: SIGSEGV]
<damex>
mrkiko: W25Q128JVSM - nor 16mb
<mrkiko>
damex: definitely no fstrim here, this is raw flash
<mrkiko>
damex: if possible, you should not be writing to it very often
<damex>
mrkiko: i don't specifically write to that flash. install openwrt, configure and sometime go over for adjustments. no extra services or anything this time
<damex>
not even snmpd ;)
<damex>
actually it didn't work for hap ac2 - i reinstalled openwrt and it still freeze ._.
<Znevna>
my hEX has an uptime of 76 days
<Znevna>
no issues
<Znevna>
kernel 5.15.82
<f00b4r0>
damex: define freeze? I don't see the relation between mtd and having ssh hang when issuing "dmesg"
<damex>
f00b4r0: ot constantly lose i/o capability
<damex>
freezong on fopen
<damex>
'freezing' on fopen
<Znevna>
a bad chip isn't unheard of :P
<f00b4r0>
although rare on SPI NOR
<f00b4r0>
unless you *really* hammered it
<damex>
f00b4r0: i didn't write much on it over last two years owning device
<damex>
it was a dumb ap with the same config between upgrades
<damex>
changed addressing few times but that's it
<f00b4r0>
damex: have you tried booting initramfs and checking if you experience the same problem (this would rule out mtd)
<damex>
f00b4r0: sure, it does not seem to be happening on initramfs
<f00b4r0>
k. Last thing you could try is flashing say 22.03. If it happens there too then odds of your flash being toast are high
<f00b4r0>
use standard image from downloads.openwrt.org
<f00b4r0>
just to be sure you're using something "known working"
<damex>
f00b4r0: sure, i just redid router 6.49.7 (last 6.x release) revert -> boot 23.03 initramfs -> seem to work sell and does not freeze
<f00b4r0>
ok. Try flashing 22.03 and check if it works from the flashed image
<damex>
well, some configuration is needed to make it works or not but sure
<f00b4r0>
what do you mean "some configuration is needed"?
<damex>
f00b4r0: i notice that it is repeatable on naked openwrt image but sometime it is not when i lets say change ip address and reboot host
<f00b4r0>
ah. That does hint at flash problem.
<f00b4r0>
what I'm curious is whether you can reproduce in 22.03. Otherwise it could be a kernel problem in current snapshot
<damex>
jffs2_scan_eraseblock(): End of filesystem marker found at 0x0 jffs2_build_filesystem(): erasing all blocks after the end marker...
<damex>
is that actually do maintenance similar to what fstrim does?
<f00b4r0>
no that's a one shot at first boot after flash
<damex>
f00b4r0: repeated on 23.03
<damex>
flashing naked image is fine. adjust ip address -> reboot -> ssh after reboot and run dmesg -> input freeze
<damex>
if i try the same multiple times - sometime dmesg runs fine
<fda>
i've now build additionla images for e8450, repeater1200 and raspbery
<fda>
all failed to build dropbear :)
<fda>
bad thing is that by defaul no logs where saved :(
<fda>
for freetz i've done this: when building quiet, the logs go to a temp file. the build of every package flushes it. but when there is an error, this file is shown. maybe an idea for openwrt...
<Znevna>
f00b4r0: it just looks like a dead wan port
<f00b4r0>
Znevna: I doubt a dead wan port would occur immediately after boot and go away after the next one though
<Znevna>
dead as in software dead
<f00b4r0>
same remark :)
<Znevna>
>.>
<Znevna>
same device?
dangole_ has joined #openwrt-devel
<f00b4r0>
no
minimal has joined #openwrt-devel
<Znevna>
I never encountered something similar
<PaulFertser>
I once had an ethernet port integrated in a desktop motherboard suddenly not working at all, no link etc. Rebooting didn't help. Power-cycling did. Never again. A latch-up event probably?
hanetzer2 has joined #openwrt-devel
hanetzer1 has quit [Ping timeout: 480 seconds]
floof58 is now known as Guest6492
floof58 has joined #openwrt-devel
Guest6492 has quit [Ping timeout: 480 seconds]
dansan has quit []
dansan has joined #openwrt-devel
<damex>
mikrotik devices have capability to 'overclock' - is it applicable on bootloader or it is routeros thing? let's say i enable >716Mhz on mikrotik hap ac2 - would it persist on openwrt too?
<robimarko>
You can overclock it via OpenWrt as well by adding an OPP
Mangix has quit [Ping timeout: 480 seconds]
<robimarko>
Choosing anything in RouterBoot wont apply on OpenWrt as Linux only scales on the OPP table
Borromini has joined #openwrt-devel
rua1 has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
bluew has joined #openwrt-devel
Borromini has quit [Quit: Lost terminal]
djfe_ is now known as djfe
Ansuel has joined #openwrt-devel
<Ansuel>
guys i need a hint for the ath bdf directory...
<Ansuel>
was thinking ath-bdf ?
<karlp>
what's bdf?
<stintel>
one of the many qca idiocies :)
<Ansuel>
board data file
<Ansuel>
think i will create repo in /firmware/ath
<Ansuel>
seems patching
<Ansuel>
matching*
<Ansuel>
(they are calibration bin that ath10k and ath11k loads to correctly int on the board)