<philipp64>
file-system partition too big (more than 13041664 bytes): Success
<philipp64>
how big can an archer squashfs be?
<mangix>
13041664 bytes?
<mangix>
actually hmm
<mangix>
I think I've done 15MB files before
<mangix>
but that's cutting it close
<mangix>
aka barely any overlayfs space
<philipp64>
and 1MB or so is for Perl, which is required for iptgeoip
ylk has joined #openwrt-devel
<slh>
a default OpenWrt image weighs just over 6 MB at the moment, leaving you a good ~5 MB for custom additions
<slh>
so ~1 MB for perl by itself shouldn't be ab issue - but presumably you somehow include two ssl/ tls providers into one image (wolfssl and openssl), just standardizing on one easily saves you 1 MB
<Mike_S>
to be more specific the "Time restriction of internet access" section of that page is not working at all. There is also some "patch" in the forum involving reordering firewall rules, but that also is not working always
<Mike_S>
It is very impacting since it is impossible to limit access to the network by time, and this was already not working in v19.x and maybeearlier. Thanks and bye
Mike_S has left #openwrt-devel [#openwrt-devel]
dangole has joined #openwrt-devel
pmelange has joined #openwrt-devel
sprucie_ is now known as sprucie
jschwart_ has quit []
jschwart has joined #openwrt-devel
rsalvaterra_ has joined #openwrt-devel
rsalvaterra is now known as Guest930
rsalvaterra_ is now known as rsalvaterra
Guest930 has quit [Ping timeout: 480 seconds]
dangole has quit [Ping timeout: 480 seconds]
dorf_ has quit [Remote host closed the connection]
<rsalvaterra>
mangix: I'm not sold yet… MIT vs GPLv3… and what are the runtime requirements of rust applications?
<mangix>
I...don't think there are any. The binaries are just big.
<mangix>
most stuff is statically linked AFAIK
<rsalvaterra>
No "librust" akin to libc? :P
<mangix>
I don't think so
<mangix>
I mean I'm pretty sure it depends on libc, but so does everything
<rsalvaterra>
I had the (probably completely wrong) idea that rust has transpiled to C or something like that.
<rsalvaterra>
*was
<mangix>
no it compiles to LLVM bytecode
<mangix>
which then gets compiled to native
<rsalvaterra>
Ok, so there's a specific rust compiler frontend.
<mangix>
right. it and clang are both frontends for LLVM
<jow>
I really dislike the new static compilation fad
<jow>
and putting that stuff into containers then to make matters worse
<rsalvaterra>
jow: That makes two of us. :/
<mangix>
jow: OTOH, smaller binaries with LTO
<jow>
because managing dependencies is hard, so jsut build statically, put into containers and orchestrate those
<jow>
bam dependency hell again, just ten layers higher
<jow>
now you need to manage your containers, the os in those containers and the orchestrator
<jow>
and since that is not really possible you just throw it away and start over if needed
<rsalvaterra>
mangix: LTO isn't magic.
<mangix>
sure it is
<jow>
mangix: LTO can't eliminate the fact that each application embeds the same common code if there's no shared runtime
<mangix>
right
<rsalvaterra>
mangix: No, it's interprocedural optimization beyond the object file boundary. Sure, it will help statically linked binaries, but you still have to include parts of the libraries with every single executable.
<mangix>
but take the case that only one executable uses said library
<jow>
chances are high that they all call resolvers, socket functions, higher level filesystem calls, algorithms (sorting), data structures (linked lists, hash tables, avl trees, ...)
<rsalvaterra>
mangix: I didn't say it doesn't have its use cases, but it's not a panacea.
<jow>
that stuff that is higher leven than libc but part of the rust standard library or however they call it
<mangix>
stintel: on a CI run, you can press Re-run all jons
<mangix>
*jobs
<Lynx->
"When implementing a new C standard library (referred to as libc henceforth in this document) starting from scratch, it is unrealistic to expect that we will have the entire library available from day one. In such a scenario, a practical approach is to redirect calls to the unimplemented functions to the same functions from another fully functional libc implementation. "
<Lynx->
hmm :)
rua has quit [Remote host closed the connection]
Lynx-- has joined #openwrt-devel
Lynx- has quit [Read error: Connection reset by peer]
<olmari>
sure, I could go look exact revision of master too, but meh, in general :)
rua has quit [Quit: Leaving.]
rua has joined #openwrt-devel
rua has quit []
Tapper has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<olmari>
okay it seems that 5.4 is stil the default choice and 5.10 is the avialable one in master and wdr4900, I shall try flashing in some 4-8-hours, depending what other folks does
<olmari>
or more of I don't expect device to not to be reachable (bootable), bu tI do have recovery stuff near
<PaulFertser>
olmari: guess with "strace -f" you can see what file descriptor it's talking about
<PaulFertser>
New way matrix shows long messages via this bridge is much nicer.
dannyAAM has quit [Quit: znc.saru.moe : ZNC 1.6.2 - http://znc.in]
dannyAAM has joined #openwrt-devel
dedeckeh has joined #openwrt-devel
<olmari>
How does it look? :)
<olmari>
I'm instantly curious :D
<olmari>
as an matrix user I can't say much about ircs side in regards... tho always good to make things work better for everyone, which is ultimate goal anyway
<olmari>
...build with strace -f running, shall see after doned what where :)
<PaulFertser>
olmari: for long messages it now gives the beginning of the message text (reasonably long) and then the link to matrix server to view it in full.
<grid>
stintel: it showed up a few days ago, i haven't done anything with it yet
Borromini has joined #openwrt-devel
<grid>
stintel: is the front panel uart just cisco-style pinout?
Tapper has quit [Ping timeout: 480 seconds]
dannyAAM has quit [Remote host closed the connection]
<olmari>
hm.... now I wonder how I'd spot the exact place from the logs... apparently id doens't output for example "Generating index" anywhere, od I don't know how to use less
dannyAAM has joined #openwrt-devel
<PaulFertser>
olmari: try searching for Bad file descriptor or the corresponding error code like EBADF ?
<olmari>
bingo :)
<olmari>
PaulFertser: I still kinda have no idea hot wo decipher this :D I mean I see all the text, but all brrr for me, I have toadmit :)
<olmari>
ofcourse I can upload log to somewhere too..
<olmari>
I mean I don't know what to seek and so on, plenty waits and resumses and close(0) and(1) etc, but relations to everything..
dannyAAM has quit [Remote host closed the connection]
<PaulFertser>
olmari: but probably you can see what file name it's trying to open that fails with that error.
dannyAAM has joined #openwrt-devel
<olmari>
I honestly can't be too sure :D I mean even likely it exists, but not in near vincinity of said error
<PaulFertser>
olmari: you can search backwards from that point for 2792031.*open.*= 3
danitool has joined #openwrt-devel
<olmari>
hm.. can't seem to get "open" anywhere related... 2792031 seems to beging with `[pid 2792030] clone(child_stack=NULL, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLDstrace: Process 2792031 attached, child_tidptr=0x7f7dd98c1a10) = 2792031`
<PaulFertser>
olmari: probably not open, might be connect or something. Search for "= 3' alone
<PaulFertser>
olmari: and you can see what it "execve" right after that clone
<olmari>
I mean I do see everything, but somewhow I feel deciphering anything... not even 123 nor ABC is there... :D
Borromini has quit [Ping timeout: 480 seconds]
robimarko has joined #openwrt-devel
sprucie has quit [Ping timeout: 480 seconds]
<robimarko>
Is anybody seeing these init errors?
<robimarko>
[ 3.603693] Run /init as init process cp: can't preserve times of '/new_root/bin/uclient-fetch': Invalid argument cp: can't preserve times of '/new_root/bin': Invalid argument cp: can't preserve times of '/new_root/dev/pts': Invalid argument cp: can't preserve times of '/new_root/dev/urandom': Invalid argument cp: can't preserve times of '/new_root/dev/random': Invalid argument cp: can't preserve times of '/new_root/dev/tty1
<mangix>
nope
<hauke>
robimarko: this is related to the musl upgrade
<hauke>
robimarko: please rebuild your toolchain
<mangix>
ah yes. time_t ABI break
<robimarko>
Ok, I rebased today and it obviously just rebuilt the musl then
<hauke>
it only builds musl, but it will not rebuild evrrything else like all packages and the gcc libs
<hauke>
which is needed
<robimarko>
Ok, will rebuild everything then. Thanks
<olmari>
PaulFertser: Didn't see any errors and echo $? gave also 0
<PaulFertser>
olmari: you can try redirecting the output to > /dev/null and see if anything appears on stderr.
<PaulFertser>
olmari: but that's the command that generates the index that's problematic according to your first log.
<olmari>
no errors, just the generating index stuff
<olmari>
also same error on same place when ran make -j1, so definately not something relating parallerism there (likely)
<olmari>
nonetheseless I don't know what and why, nor should if be how critical and so on
<PaulFertser>
Odd
<olmari>
well... I'll take the device literally down and flash it (among other things I do with networking here), weäll generally se does it work after awhile =)
<olmari>
Well, whatever that "index issue" is, it doesn't seem to affect WDR4900 at all... I mean the image itself etc
<olmari>
brr.. the functionality of said image
<olmari>
not that I really expected it at this point, but whatever :)
<olmari>
obviously sometihng is funnily there, but yeah
philipp64 has joined #openwrt-devel
<philipp64>
mangix: Not building wolflssl fortunately... figured out that lldpd pulls in libnetsnmp which is big...
<philipp64>
What ath79 has AX or AC3000 and a decent amount of flash and DRAM?
<philipp64>
also, can you force the speed on an Ethernet interface to 10mb/s?
<philipp64>
I can't get SQM to properly shape, so that might be the best I can do...
<mangix>
philipp64: no ath79 has ax
<philipp64>
and "basic_rate" and "support_rates" don't seem to work, either...
<mangix>
archer c7 has a replaceable mini pcie card
<philipp64>
[ 1556.692276] ath10k_pci 0000:00:00.0: pdev param 0 not supported by firmware
<philipp64>
I think I have an a7-v5... which is similar, right?
<mangix>
no
<mangix>
v5 does not have a mini pcie slot
<philipp64>
ah.
<mangix>
ax mini pcie cards are hard to come by
<philipp64>
The stepkids (who are 14 and 17) haven't been respecting the house rules about using the home WiFi, so I've forced to put them on their own WAP... trying to limit their traffic to 10mb/s but can't figure out how to do so without setting the port on the household VLAN switch to 10mb/s...
pmelange has quit [Quit: Leaving.]
<mangix>
philipp64: I think hostapd can be configured for max mbps
<mangix>
you could also force 802.11g
<philipp64>
got a sample config?
<mangix>
nope
<mangix>
forcing 802.11g is easy with luci though
<mangix>
AFAIK the max becomes ~30mbps
<mangix>
11a for 5ghz is the same speed
dedeckeh has quit [Remote host closed the connection]
<rsalvaterra>
philipp64: I use cake to limit the bandwidth.
<rsalvaterra>
mangix: Limiting to lower rates is terrible for airtime.
<hurricos>
I have an Extreme Networks AP3825i: it's basically identical to an Aerohive AP 330. Anyone know how I might find the U-boot bootloader password?