<PaulFertser>
I think patch-cmdline just changes the whole line specified in DT.
Grommish is now known as Grommish[M]
Grommish[M] is now known as Grommish
<damex>
PaulFertser: CONFIG_MIPS_CMDLINE_FROM_BOOTLOADER this was the one. if i use other option like extend dtb CONFIG_MIPS_CMDLINE_DTB_EXTEND it works
Grommish[M] has joined #openwrt-devel
<damex>
but /proc/cmdline is still console=ttyS0,115200 only
<Grommish>
Morning all
<Grommish>
PaulFertser: Are you any good at package Config.in files for menuconfig selections?
<PaulFertser>
Grommish: I guess about average good :)
<Grommish>
I'm setting up a massive submenu for Rust. Is there a way to use the depends to deselect a CONFIG_RUST_xxx bool if the option is selected?
<PaulFertser>
Grommish: that's above my knowledge but I can suggest you try searching the internet for some keywords like "automatic deselect" and Kconfig because OpenWrt is basically borrowing the Kconfig system from Linux, the kernel.
<Grommish>
Like, if CONFIG_RUST_ENABLE_NINJA is y, then CONFIG_RUST_OPTION_CHECKING would be set to n
<Grommish>
Ok
<Grommish>
Thank you :)
<Grommish>
BTW, I got rust to work inside the buildroot, from source. I'm just getting them to fix cross compile issues :)
<Grommish>
and I even added mips64 soft-float targets successfully
<PaulFertser>
Grommish: massive progress! Does it really take ~100 GB to build it?
<Grommish>
If I did it the way they wanted it.. I compromised
<Grommish>
and it was 100gb per TARGET :(
<Grommish>
so, that was a no go..
<Grommish>
I pull a stage0 precompiled to actually create the stage1 and stage2 compilers and tools
<Grommish>
So i don't have to go up through each rust release to build the next one with the previous
<Grommish>
All this so I can run Suricata because they told me I couldn't
<Grommish>
I don't even know C/C++, let alone rust
gch981213 has quit [Read error: Connection reset by peer]
gch9812130 is now known as gch981213
<Grommish>
Now, if they'd get the issues with rust itself fixed, I'll be able to actually use it hehe. But the don't get many people using it the way we are
<ynezz>
just look at the config/ folder for examples
gch981213 has quit [Read error: Connection reset by peer]
gch9812133 is now known as gch981213
graphine has joined #openwrt-devel
aszeszo has joined #openwrt-devel
samantaz__ has quit [Ping timeout: 272 seconds]
Grommish[M] has joined #openwrt-devel
<damex>
so on stock firmware sfp port could details about insertion of module and its other details. currently all Phy's are detected as 'external' (either with generic phy module or Phy specific - vsc8504) and no extra details
Darkmatter66_ has quit [Quit: ZNC 1.7.5 - https://znc.in]
Darkmatter66 has joined #openwrt-devel
rossi947 has joined #openwrt-devel
<rossi947>
Hello everybody,
<rossi947>
Does anybody here know if you can get a router to work as a vdsl modem + router with sqm and qos using openwrt?
<pkgadd>
yes'ish (lantiq VRX2xx), *BUT*... these lantiq routers are hard pressed to reach 100 MBit/s VDSL2+vectoring, routing and firewalling without SQM at all (~85 MBit/s for dual-core devices, ~55-60 MBit/s for single-core/ FXS supporting devices - only software flow-offloading can get it up to ~115 MBit/s); with SQM enabled, I wouldn't assume it to cope with more than ~25 MBit/s
<pkgadd>
the only reasonable choice is to use the VDSL modem as modem only or *simple* modem-router without SQM/ QoS
<rossi947>
so I should be better of buying a vdsl modem and a ubiquiti ER-4?
<rossi947>
I am trying to adress jitter and bufferbloat
<pkgadd>
in that case you need a VDSL modem and a powerful/ fast OpenWrt supported router. the modem can be running OpenWrt, but doesn't need to (as long as it offers bridge mode). I don't know about the ER-4, so won't make any recommendations in that regard
<rossi947>
Thank you very much, you have helped me a lot
<pkgadd>
sadly lantiq VRX2xx just isn't fast enough to do SQM at 'normal' high speeds
<pkgadd>
otherweise they're cheap and pretty nice
<damex>
rossi947: edgerouter 4 is not in a tree yet so it can't be recommended for purchase (yet?)
<stintel>
depends on how much one wants to tinker :)
<damex>
stintel: btw, made it work with octeon3 target and enabled FPU support. 520 -> 540~550Mbit and bump in ssl benchmark :)
<pkgadd>
in terms of lots of tinkering, xiaomi ax3600 (ipq8071a) might be more interesting, but likewise lots left to do
<stintel>
:P
<pkgadd>
(even though the CN7130 might be faster, hard to guess)
<stintel>
yeah, I have a Huawei AP7060DN in a similar state ;)
<damex>
i wonder how ipq8071a compares to cn7130
<pkgadd>
it seem the ax3600 might be slightly easier than the AP7060DN (it's confirmed that secure-boot isn't enabled on the ax3600 and efforts are in varying stages of basic functionality - if I understand it correctly, the biggest remaining issue seem to be enabling the third 'IoT' (1x1 ath10k) radio, respectively the PCIe bus driving it - aside from the general ipq807x support)
<pkgadd>
(and serial console access seems to be reasonably easy, just 1.8 volts UART and an unpopulated header at the edge of the board)
<stintel>
well once this crisis situation at work settles and I can take my canceled holidays I might look into it again
<pkgadd>
(I haven't bought it, yet(?) but it's tempting, aside from merely 3+1 ethernet ports. no USB and a slightly unclarity if the radios would allow ch52-64/ ch100-149 with ATH_REGD_USER)
rossi947 has quit [Remote host closed the connection]
<Namidairo>
oof that 5ghz channel availability in china
<Namidairo>
I don't even see contiguous 80mhz in there
Nick_Lowe has joined #openwrt-devel
swalker has quit [Ping timeout: 258 seconds]
slh64 has quit [Quit: gone]
slh64 has joined #openwrt-devel
slh64 has quit [Client Quit]
slh64 has joined #openwrt-devel
Borromini has quit [Ping timeout: 246 seconds]
<Namidairo>
looks like they have a cut-down version without the 1x1 radio, also with the 8071a
opal has quit [Ping timeout: 240 seconds]
<pkgadd>
the redmi ax6, yes - but apparently with half the rf power
bookworm has quit [Read error: Connection reset by peer]
bookworm has joined #openwrt-devel
MichaelOF has joined #openwrt-devel
Grommish has quit [Read error: Connection reset by peer]
Grommish has joined #openwrt-devel
opal has joined #openwrt-devel
Borromini has joined #openwrt-devel
<Grommish>
ynezz: ping
<rsalvaterra>
Christ, I've been beating my head against the wall for over two hours… because conditional directives can't be indented with tabs in makefiles.
<rsalvaterra>
I mean, really? In 2020?
<rsalvaterra>
GNU Make is weird.
<Grommish>
haha
<Grommish>
Sorry
<PaulFertser>
It has to behave in a compatible way
<Grommish>
yeah, tab = shell
<rsalvaterra>
Yeah, yeah, I know, but… bah. :P
adrianschmutzler has joined #openwrt-devel
<Grommish>
Ok, because I'm a glutton and bored. I want to start testing 5.8 kernel.. I've added the kernel-version.mk changes for it, but it errors on kconfig.pl.. Where else do I need to change things? I didn't see any other (non-patches) commits
<Grommish>
the kconfig.pl is missing the first arg.. hmm
<rsalvaterra>
Grommish: You'll probably have to forward-port all the 5.4 patches to 5.8, no?
<Grommish>
Probably, but I can't even get into kernel_menuconfig because it's erroring on copying the .config in the kernel build dir
<damex>
seems like 20% difference in favor of openwrt ;)
<karlp>
not for md4... ;)
<karlp>
or hmac, or sha1...?
<Grommish>
karlp: 5.8 is the stable, so i figure I'd have the best chances.. and I think i figured it out :)
decke has quit [Quit: Leaving.]
<Borromini>
Grommish: 5.9 is already released as well by now? October 11th.
<karlp>
or aes128?
<karlp>
damex: which ones are better for openwrt?
<Grommish>
Borromini: Go big or go home, you mean?
<damex>
karlp: higher bytes throughput and sign/verify. check numbers at the end
<graphine>
I once tried to have linux from master branch in openwrt as kernel version, didn't end too well
<damex>
some of them faster, some not though ;p
<karlp>
damex: whcih numbers at the end, only sha512 is faster that I can see?
<Borromini>
Grommish: hehehe :P no i thought maybe you were still assuming it was an rc
<karlp>
ok, your rsa sign/verify is too.
<Borromini>
Grommish: grated
<Borromini>
Grommish: granted, the .0 release is usually a bit fresh, at least i remember from the olden days. lots of people would indeed wait until it had received some patches
<damex>
karlp: hmm... rechecked again - yeah. sha512 is killing it on openwrt.
<lemmi>
damex: tell me what to do :D
<damex>
lemmi: openssl speed -elapsed
<lemmi>
damex: wrt to er-12
<damex>
karlp: sign/verify for rsa is faster ;p
<damex>
lemmi: oh
<damex>
i haven't finished with octeon3 target yet
<lemmi>
damex: btw, was able to snack an er-4 on ebay
<damex>
lemmi: expensive ? :)
<lemmi>
50% off
<damex>
oh shi
<damex>
for 100 bucks it is a steal ;p
<lemmi>
80€
<damex>
95$ :)
<damex>
nice catch ;p
<lemmi>
yeah. that series of routers really fits for what i'm currently building here. so i'd love to see wrt running on there.
feriman has joined #openwrt-devel
<damex>
lemmi: can you also extract readable dts on working system ?:)
adrianschmutzler has quit [Quit: ~ Trillian - www.trillian.im ~]
<lemmi>
damex: /sys/firmware/devicetree/base nor dtc exist on edgeos
<damex>
oh
<damex>
lemmi: maybe try /proc/device-tree ?
<damex>
if it is there
<damex>
oh no dtc ;(
<lemmi>
yep /proc/device-tree exists
<lemmi>
hm. maybe i can statically compile dtc, let me have a look
<damex>
lemmi: there is debian repos available
<damex>
you can install it from that repos
<lemmi>
true
* lemmi
closes 20 tabs
<damex>
any suggestions how to find what support i need to do a monitoring for sfp module? [ 210.359033] SFP module removed from eth3 [ 212.429167] SFP module inserted at eth3
<lemmi>
damex: any wishes what openwrt to boot, or is the generic octeon snapshot from a couple weeks ago fine?
<damex>
lemmi: openwrt will boot. generic openwrt. but you won't be able to flash it without modifying. if you can wait a bit - i will be done with new target pretty soon and you could flash er12 using custom dts (limited to phy ports so no switch for now)
<damex>
ah
<damex>
you mean which openwrt to boot... any of them is fine
<damex>
as long as it boots octeon target
<rsalvaterra>
Guys, just to have a general idea… how emotionally attached are we to RSA in Dropbear? :P
<stintel>
ed25519 or gtfo ;)
<rsalvaterra>
stintel: <3
<jow>
I am emotionally attached to it
<jow>
and practically as well
<rsalvaterra>
Aww…
<rsalvaterra>
Alright, I won't default to Ed25519 only… :P
<rsalvaterra>
… even if it cuts about 8 kiB from the executable…
state has joined #openwrt-devel
<Grommish>
Oof.. This backporting is going to take forever
<Borromini>
:P
<Borromini>
Grommish: how does that octeon compare to say the Armada 88F3720? Any idea?
<lemmi>
damex: gist updated
<damex>
lemmi: thanks
<Borromini>
i see that ER-4 is using an Octeon as well, and you were porting an octeon no?
<Borromini>
was looking at a gl-inet brume as an edge router until now.
luke-jr has quit [Read error: Connection reset by peer]
<stintel>
that brume is a dual-core vs er-4 quad-core afaik
<stintel>
I have a 3720-based device but it has only a 100Mbps so can't do very useful testing
<Borromini>
stintel: yeah, i'm not sure how well networking can be parallelised?
<Borromini>
the brume has openwrt support already but only three ports, the er-4 has four
<stintel>
well if you have multiple real ethernet interfaces you can balance the IRQs over different cores
luke-jr has joined #openwrt-devel
<Borromini>
stintel: you mean as in: not a switch chip but separate nics?
<stintel>
yes
<Borromini>
ok. it looks like both have an actual switch, not separate nics
<stintel>
the er-4 has 2 NICs
<Borromini>
yes sorry. one for the wan and a QCA switch for the lan
<damex>
2x QSGMII but 4 PHY connected to a single QSGMII
<damex>
that is ER4 as far as i understand
<damex>
same for whole lineup of ER* with cavium3
hbug has quit [Remote host closed the connection]
hbug has joined #openwrt-devel
Grommish[M] has quit [Ping timeout: 244 seconds]
Grommish[M] has joined #openwrt-devel
<Borromini>
damex: you're running tests on yours correct?
<lemmi>
damex: have you tried mii-tool -w for monitoring the sfp module?
luke-jr has quit [Excess Flood]
luke-jr has joined #openwrt-devel
luke-jr has quit [Excess Flood]
<lemmi>
damex: ah, that only reports something once you establish a link
<Borromini>
damex: hey. i did leave, but since the channel is logged it's easy to find stuff if someone responds :)
gcontrol has joined #openwrt-devel
<damex>
Borromini: aswering your question - i am testing on my ER4
<Borromini>
i'm curious about your finds if you can run tests similar to those in the topic i linked to. I have no idea about how the octeons compare to the Marvell ARM CPUs
<Borromini>
damex: cool, thanks!
<damex>
wanted to sell it off but managed to a push to get openwrt running on it
<damex>
now it is for keeping. turns out to be a very nice piece of hardware that held back by software :)
<Borromini>
=)
<Borromini>
like a lot of hardware unfortunately.
<damex>
Borromini: do you have that Marvell for testing?
<Borromini>
the look on friends' faces when you say you can turn a router into an AP...
<damex>
well, lots of manufacturers let you do that
<damex>
found where mac is kept and used eeprom for that ;)
<damex>
not really where mac is kept... it is generated from serial
<lemmi>
ah
<damex>
so find serial -> increment them
<damex>
anyway fix is in place for the last error
<lemmi>
already compiling
flokli has quit [Ping timeout: 272 seconds]
<lemmi>
usb is working btw
vidister has quit [Ping timeout: 272 seconds]
brn has left #openwrt-devel [#openwrt-devel]
brn has joined #openwrt-devel
<damex>
lemmi: nice ;p
state has quit [Remote host closed the connection]
Nick_Lowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
vidister has joined #openwrt-devel
state has joined #openwrt-devel
state has quit [Remote host closed the connection]
state has joined #openwrt-devel
Nick_Lowe has joined #openwrt-devel
flokli has joined #openwrt-devel
gch981213 has quit [Quit: Ping timeout (120 seconds)]
gch981213 has joined #openwrt-devel
state has quit [Remote host closed the connection]
state has joined #openwrt-devel
lnslbrty has joined #openwrt-devel
Grommish[M] has quit [Ping timeout: 260 seconds]
<lemmi>
damex: ok... something *very* odd :D but one thing at a time: i can set eth0-eth3 up. but there seems to be something wired wrong. when i plug in a cable at port 8, it says eth0 is connected, but i can't reach it or ping anything from it. when i put in a cable that loops from port 9 to 10, eth1 and eth2 are connected and i can now reach with the link local address of eth2 anyething that's
<lemmi>
connected via eth0
<lemmi>
man i hope that makes any sense
jg has quit [Ping timeout: 260 seconds]
<damex>
lemmi: weird. very weird. let's try to bring all the stuff fup then. no panics at a time is good
<lemmi>
damex: eth0-eth3 are up. eth4-eth7 cause the same warning as before (expected)
<damex>
yeah
<lemmi>
so the phyical port seems to be correctly assigned to what you intended
<damex>
> 119 insertions
<lemmi>
but traffic from eth2 seems go out the physical eth0
<damex>
lemmi: try to pull and build, let me know
<lemmi>
k
<damex>
you're supposed to have all 12 ports after that
<lemmi>
i hope you are right :D
<damex>
you will have wrong switch configuration - it will probably give you some panics
<damex>
and lan/wan configs too
state has quit [Remote host closed the connection]
<rsalvaterra>
Oh, Linux 5.9 now supports REJECT in the PREROUTING chain. Cute.
aszeszo has joined #openwrt-devel
graphine has quit [Ping timeout: 272 seconds]
<lemmi>
damex: so.. still no luck. so i booted current octeon generic. eth2 maps to port 8, eth3 to port 9, eth1 to port 10 (sfp) and eth0 to port 11 (sfp)
<lemmi>
(and networking works on all of those 4 ports
<damex>
lemmi: what ip link and ls /sys/bus/mdio_bus/drivers/* says ?
<damex>
lemmi: added your new reginit that is extracted from openwrt
<damex>
lemmi: please try again :)
<lemmi>
sure
feriman has quit [Quit: WeeChat 2.9]
Ycarus has quit [Quit: Ycarus]
feriman has joined #openwrt-devel
opal has quit [Ping timeout: 240 seconds]
<lemmi>
damex: so.. the interfaces seem to correctly map to the ports, but again no traffic
<damex>
lemmi: sure, lets see... i manage to repeat same thing locally with er4
<lemmi>
k
<damex>
gonna try fixing it
opal has joined #openwrt-devel
swalker has joined #openwrt-devel
awgh has quit [Ping timeout: 256 seconds]
Borromini has quit [Quit: Lost terminal]
feriman has quit [Quit: WeeChat 2.9]
<Grommish>
make sure you have the Octeon ethernet drivers enables
<Grommish>
in Device Drivers/Staging
<Grommish>
Otherwise, it'll look like it's doing something and never leaves the PHY
<Grommish>
Ask me how I know :P
graphine has joined #openwrt-devel
shibboleth has joined #openwrt-devel
<shibboleth>
I've got a device led (wlan) which for some reason gets turned on during boot (even though the radio is disabled). if i enable the radio the leds goes off and then back on and stays off if i then disable the radio
<shibboleth>
i tried option default '0'
Misanthropos has quit [Ping timeout: 260 seconds]
feriman has joined #openwrt-devel
<shibboleth>
any way to debug why it turns on during boot and prevent it?
graphine has quit [Quit: Leaving]
black_ant has quit [Ping timeout: 272 seconds]
<lemmi>
Grommish: just checked and it's selected
<damex>
Grommish: ofcourse octeon driver (octeon_ethernet) is enabled and used :)
<Grommish>
hehe
<ynezz>
Grommish: poing
<aparcar[m]>
ynezz: are you packages.git admin?
<damex>
Grommish: it never leaves phy although phy have proper status
<Grommish>
ynezz: hey hey.. I was going to ask you about the kernel versioning, but i figured it out
<damex>
happens after you feed kernel the same device tree device had and it goes crazy. probably something fishy is happening
<ynezz>
ldir: are you able to somehow reach Simon of dnsmasq fame?
<damex>
only thing generic does not do is not telling kernel that <this board is e300 board>
<damex>
but it does not excuse it for having such behaviour
<damex>
ubnt_prune_device_tree();
<damex>
oh really... that ubnt
<ldir>
ynezz: I checked through my emails and can't find any magic back channel for him. I've sent an email to the address that I (and everyone else) has.
<graphine>
are there any future plans to use nftables in fw3?
jg has quit [Ping timeout: 260 seconds]
Nick_Lowe has joined #openwrt-devel
Darkmatter66_ has quit [Quit: ZNC 1.7.5 - https://znc.in]
Darkmatter66 has joined #openwrt-devel
brn has quit [Remote host closed the connection]
damex has quit [Remote host closed the connection]
damex has joined #openwrt-devel
<ynezz>
graphine: AFAIK yes, in fw4
<graphine>
has work started on fw4?
goliath has quit [Quit: SIGSEGV]
csrf has quit [Read error: Connection reset by peer]
guerby has quit [Read error: Connection reset by peer]
csrf has joined #openwrt-devel
guerby has joined #openwrt-devel
<karlp>
aparcar[m]: do you have any figures on the time saved before/after all these cleanups? it should add up to something meaningful now right? I know each little one is tiny, but with the amount of them, it shoudl be better?
<aparcar[m]>
karlp: No figures sorry. I'm just guessing it's a tiny bit on each build. Mostly done for cleaning
<karlp>
I meant all the tings are "cosmetic" but save time
<aparcar[m]>
oh yea
<aparcar[m]>
I just send the other patchwork patch because it's also imagebuilder/sdk related
<aparcar[m]>
and the patchwork patch is a bugfix
<karlp>
there's been a bunch of this sort of stuff, that's all micro-opts, but they add up, I was just curious if there was any before/after, but I guess with kernel updates over the time as well, it's hard to get a good measure
Nick_Lowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Nick_Lowe has joined #openwrt-devel
graphine1 has joined #openwrt-devel
graphine1 has quit [Client Quit]
graphine has quit [Ping timeout: 272 seconds]
gch981213 has quit [Read error: Connection reset by peer]
gch981213 has joined #openwrt-devel
Nick_Lowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Darkmatter66 has quit [Quit: ZNC 1.7.5 - https://znc.in]
Mister_X has joined #openwrt-devel
Mister_X has quit [Changing host]
Mister_X has joined #openwrt-devel
matteo has quit [Ping timeout: 256 seconds]
Darkmatter66 has joined #openwrt-devel
Nick_Lowe has joined #openwrt-devel
Nick_Lowe has quit [Client Quit]
txt_file has joined #openwrt-devel
linzst has quit [Quit: Leaving]
txt_file is now known as txt-file
txt-file has quit [Quit: Leaving]
aszeszo_ has joined #openwrt-devel
aszeszo has quit [Ping timeout: 260 seconds]
philipp64|work_ has joined #openwrt-devel
philipp64|work has quit [Ping timeout: 240 seconds]
philipp64|work has joined #openwrt-devel
<mirko>
Maybe it's (just) late, but scrolling through libubox I wonder what's its actual license? I mean, all files have proper headers - but if I'd be selling a commercial product with libubox - what's my obligations?
philipp64|work_ has quit [Ping timeout: 258 seconds]
SmugLeaf has quit [Remote host closed the connection]
SmugLeaf has joined #openwrt-devel
th3g1z_ has quit [Quit: aloha]
heffer has quit [Remote host closed the connection]
damex has quit [Ping timeout: 256 seconds]
Ultrasauce has quit [Remote host closed the connection]