xback has quit [Remote host closed the connection]
xback has joined #openwrt-devel
rua has quit [Remote host closed the connection]
rua has joined #openwrt-devel
<xback>
ynezz: I was going to try bump imx to 6.1 .. but noticed they completely overhauled some dts's :-o (apalis)
Ansuel has quit [Quit: ZNC 1.9.x-git-253-750eac08 - https://znc.in]
Ansuel has joined #openwrt-devel
<lu_zero>
interesting
<lu_zero>
$(qstrip $(CONFIG_EXTERNAL_CPIO)) with CONFIG_EXTERNAL_CPIO=/path/to/my.cpio
<lu_zero>
results in ""
<dwfreed>
probably because it's supposed to be $(call qstrip,$(CONFIG_EXTERNAL_CPIO))
<lu_zero>
let me see if it is just that typo
<lu_zero>
do you remember how to call opkg from the makefile?
<lu_zero>
Ideally I could change the
<lu_zero>
find . | LC_ALL=C sort
<lu_zero>
used below
<lu_zero>
to exclude the contents of the packages I do not want in the initramfs
Ansuel has quit [Quit: ZNC 1.9.x-git-253-750eac08 - https://znc.in]
Ansuel has joined #openwrt-devel
filimonic has quit [Read error: Connection reset by peer]
filimonic has joined #openwrt-devel
<lu_zero>
dwfreed: meanwhile that was the problem ^^;
jeff___m has joined #openwrt-devel
jeff___m has quit [Ping timeout: 480 seconds]
cp- has quit [Read error: Connection reset by peer]
<xback>
nbd: do you recall I once pinged you regarding altered hostapd config upstream? :-) https://git.openwrt.org/?p=openwrt/staging/xback.git;a=commit;h=3eda05c6d70ab7dd5bb35a282299f967cae3ae6c
<xback>
nbd: doing the same for wpa_supplicant now, and then check the Makefile which alters these symbols ..
<nbd>
thx
<Ansuel>
xback news for the wifi test?
* f00b4r0
can't get recent ucode to build on 22.03, sighs
<xback>
Ansuel: No, didnt get to it yet
<xback>
was playing around with WDS and all the sudden cleaning up hostapd config files :-p
* f00b4r0
gets bitten by the wireless iface renaming in 23.05, curses politely
<rmilecki>
huh, all those targets have CONFIG_SLUB_DEBUG=y enabled: gemini layerscape oxnas qoriq rockchip sifiveu
<rmilecki>
why?1
<rmilecki>
linusw: any idea on gemini?
<linusw>
rmilecki: I think it was left on by cargo cult mistakes, probablt as a side effect of using KASAN which requires SLUB_DEBUG
<linusw>
I'm pretty sure noone would react if you just delete that (or "# FOO_CONFIG is not set") from all defconfigs
<rmilecki>
i was thinking about that
<rmilecki>
interesting how many different surprises are there in defconfigs :)
<robimarko>
A lot of stuff probably slipped by
<Ansuel>
rmilecki actually.... all that debug stuff should be disabled on production fw and should be enabled only by tweaking config (aka manualy enabling them)
<Ansuel>
so yes they probably slipped in...
<Ansuel>
the increased overhead is expected with those debug feature enabled
<hitech95>
Ansuel, dumb question, its been a while since last reply I got from dangowrt. who else is skilled with filologic stuff?
<hitech95>
I've seen that you merged some related PRs. I'm facing some issues on porting a device and I'm not able to find a solution after weeks of works. (sysupgrade not working)
<jow>
JiiPee: what do you mean with "private network traffic" ?
<JiiPee>
jow: example I have lan 10.10.10.0/24 and my 5G modem is connected to router wan port in bridge mode. Modem does have management IP from network 192.168.1.0/24
<JiiPee>
OpenWRT lets traffic go to that management IP
<JiiPee>
previouslty I had to specially create virtual interface for that
<jow>
JiiPee: why not? openwrt could be daisy chained behind another router using rfc1918 range
<jow>
Ansuel: pong
<Ansuel>
any chance for lua 5.4 support? I'm having some difficulties in understanding the logic for the shared lib patch
jeff___m has joined #openwrt-devel
<Ansuel>
can you give some hint? (5.4 is needed for the next snmp version since they switched to lua 5.4
<jow>
Ansuel: no capacity or interest from my side
<jow>
maybe simply drop the patches
<jow>
and just ship a vanilla 5.4
<JiiPee>
if I want to daisychain I would do it through lan
<Ansuel>
sure just wanted to know if you had any idea how that worked... what is the logic? think it's needed for having more version of lua in one system
<JiiPee>
why would I want to do it through wan?
<jow>
JiiPee: for example if you have an upstream network using rfc1918 ranges without the ability to configure reverse routes you would need double nat
<JiiPee>
Now bridge device on wan port is open for all kind of malware scripts sniffing unsecure modems or what ever media converters
<jow>
Ansuel: can you be more specific wrt. "the shared lib patch"
<JiiPee>
if user is stupid enough to click all kind of links
<jow>
JiiPee: as I said, it's not the downstream routers task to protect the upstream network
<JiiPee>
so openwrt is going less security is better?
<JiiPee>
before you had to open it up as explained in that link I posted
<jow>
JiiPee: nothing was "opened up", what was done is adding an ip alias to wan and a nat excemption, so that requests toward that management interface are sent with the proper source ip
<jow>
depending on how the mgmt iface on the upstream device is imlemented it may or may not work without that config
<JiiPee>
it didnt work with 22.03
<jow>
then please share your current /etc/config/network, /etc/config/firewall in pastebin
<jow>
as well as the ouput of iptables-save and "nft list ruleset"
<jow>
and "ip -4 route"
<jow>
so we can see whether there's a bug or not
<JiiPee>
dont think 23.05 use iptables?
<jow>
JiiPee: not by default, unless user scripts stage rules
<jow>
Ansuel: seems someone forcibly exported some symbols which are otherwise hidden in liblua.so with that patch
<jow>
Ansuel: I can only guess that there's been some users of those symbols
<JiiPee>
you want me to install that iptables-save?
<jow>
Ansuel: I suppose the author of the patch kept exporting/un-hiding symbols until the "lua" and "luac" executables successfully linked against liblua.so
<Ansuel>
oh! i like bruteforce and no brain way
<jow>
Ansuel: looks like all the methods changed to LUA_API are needed by either lua or luac
<Ansuel>
jow are you still here? I have a question about netifd
<Ansuel>
(i'm adding a way to add toggable gro and wonder if I should also add it with ubus...)
<Ansuel>
(from what i understood adding the entry cause the thing to be configured directly in the network config)
jeff___m has joined #openwrt-devel
<JiiPee>
jow: anything interesting there?
jeff___m has quit [Ping timeout: 480 seconds]
<jow>
JiiPee: no looks standard
<jow>
JiiPee: my cable modem behaves the same here
<JiiPee>
right
<jow>
I am actually wondering why that modem access stuff was needed in the first place on older releases
<JiiPee>
well I would prefer the way 22.03 worked
<jow>
seems like an accidential feature
<jow>
there's no way for the router to know if a random destination IP is just an upstream network host or some kind of protected management
<jow>
... interface
<JiiPee>
then I could just eneable/disable that "modem" interface when I need it/dont need it
<jow>
so dunno what made 22.03 work the way it did but it was certainly not the firewall or nat setup
<jow>
maybe different arp or rpath filter behavior
xback has quit [Remote host closed the connection]