danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
Tapper has joined #openwrt-devel
csharper2005 has joined #openwrt-devel
<csharper2005>
rmilecki: ping6
cbeznea has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
cbeznea has quit [Quit: Leaving.]
cbeznea has joined #openwrt-devel
rua has joined #openwrt-devel
lelux_ has joined #openwrt-devel
lelux has quit [Read error: Network is unreachable]
goliath has joined #openwrt-devel
AtomiclyCursed has quit [Quit: ZNC 1.8.2 - https://znc.in]
Atomicly- has joined #openwrt-devel
csrf has quit [Ping timeout: 480 seconds]
ekathva has joined #openwrt-devel
swegener has quit [Quit: leaving]
mrkiko has joined #openwrt-devel
grift has quit [Quit: Bye]
grift has joined #openwrt-devel
swegener has joined #openwrt-devel
Misanthropos has joined #openwrt-devel
pepe2k has joined #openwrt-devel
danitool has joined #openwrt-devel
robimarko has joined #openwrt-devel
bluew has quit [Ping timeout: 480 seconds]
robimarko has quit [Read error: No route to host]
robimarko has joined #openwrt-devel
c0sm1cSlug_ has joined #openwrt-devel
c0sm1cSlug has quit [Ping timeout: 480 seconds]
ecloud_ has quit [Ping timeout: 480 seconds]
ecloud_ has joined #openwrt-devel
<aparcar[m]>
pepe2k: no meeting planed, I got a bit busy with other things. We should have one next week since rc1 is released
<pepe2k>
aparcar[m]: got it, thanks!
<mrkiko>
pepe2k: hi!!
<pepe2k>
mrkiko: hi :)
T-Bone has joined #openwrt-devel
f00b4r__ has quit [Ping timeout: 480 seconds]
csharper2005 has quit [Remote host closed the connection]
ekathva has quit [Remote host closed the connection]
ekathva has joined #openwrt-devel
ekathva has quit [Remote host closed the connection]
ekathva has joined #openwrt-devel
ecloud_ has quit [Ping timeout: 480 seconds]
Tapper has quit [Ping timeout: 480 seconds]
<aiyion>
I've just been looking at leds in OpenWrt and would like someone to review my findings:
<aiyion>
If a target has no override on leds.sh it makes use of the default one in package/base-files.
<aiyion>
after sourcing that file on a running node, one can query led paths with eg `get_dt_led running` which would then return the path to the led, that was referenced through the alias led-running in the dts file?
<aiyion>
And if that alias is not available there's no fallback to another LED.
<aiyion>
Is that right?
<grift>
i accidently selected luci-ssl-openssl in menuconfig and it failed to build due to conflict with ustream-wolfssl, but now there is no way back. i can't deselect openssl anymore even if i deselect all of luci....
ecloud_ has joined #openwrt-devel
<dwfreed>
rm .config :)
<grift>
.... damnit i spent hours just getting it right ...
<stintel>
try harder, there must be something still depending on openssl
<olmari>
maybe (not proven) look what luci-ssl-openssl is choosing and dig in to unselect em all... could be short or long chain
<stintel>
libopenssl should show them
<stintel>
ehr
<stintel>
/ libopenssl
<stintel>
irc client ate the /
<grift>
thanks i will try
<Piraty>
any chance to filter the device table for x86 devices ?
ekathva_ has joined #openwrt-devel
<dwfreed>
Piraty: you mean you only want to see x86 devices? (but also, they're not likely to all be listed, because x86 support is very generic, as there's no need to specialize)
<Piraty>
yes. i know x86 is generetic in regard to kernel but since there are x86 devices specifically for routing / networking purpose i assumed there might be different levels of (device specific) openwrt integration
<Piraty>
think of broadcom driver crap for instance
<Piraty>
which is not x86 specific but you probably get what i mean
<dwfreed>
I mean, then you just install the necessary drivers
ekathva has quit [Ping timeout: 480 seconds]
<dwfreed>
you can filter by target on the full table of hardware, and there are a few devices listed, but the images are not specialized for those devices in the slightest
Tapper has joined #openwrt-devel
<dwfreed>
the only differentiation is on 32 bit, there's geode and p4 images
<grift>
stintel: thanks that pointed me in the right direction, menu config is kind of intimdating or at least overwhelming to me
<stintel>
grift: just use it more often ;)
<grift>
true
<stintel>
I built my first kernel somewhere early 2000s, I've been using it for a while :P
<grift>
yes ok but the kernel is not openwrt
<Habbie>
stintel, yes, i also remember being young :D
<stintel>
Habbie: :P
<stintel>
but it's the "same" menuconfig afaik
<grift>
true
<stintel>
anyway, glad you found it
<grift>
but the contents is different
<stintel>
I know how much time you can spend on getting a config right
<grift>
but youre still right, i should eventually know all the contents
<stintel>
then having to nuke it ... no!
<grift>
so keep using it until you can sum up every option
<Habbie>
"you like menuconfig? name every option"
<dwfreed>
Spent so much time in menuconfig when I ran Gentoo
<dwfreed>
Made very slim kernels, it was great
<Habbie>
my first slackware offered several kernels - some of which supported my HDD, and some of which supported my CDROM drive
<Habbie>
so i was forced into building a kernel on day two
<grift>
name every openwrt menuconfig option and then translate that to the package name
<grift>
like for example gptfdisk -> gdisk
<schmars[m]>
I have an 802.11s plain/no-encryption interface that hostapd lists as unmanaged. Is that because it's unencrypted oder because it's not ap/station?
<Piraty>
thanks dwfreed for hinting to full toh, didn't find it at first
ekathva_ has quit [Quit: Leaving]
goliath has quit [Quit: SIGSEGV]
<neggles>
wow ubiquiti gave me a gpl tarball in record time
<robimarko>
Have they "accidentally" left out all of the interesting stuff?
<stintel>
meanwhile still waiting on watchguard
<neggles>
grift: try `make nconfig` and `make qconfig` too :D
<stintel>
I'm tempted to let the SFC handle it
<neggles>
stintel: I binpatched the u-boot on my T20 to get into it
<neggles>
changed one cbz to cbnz
<stintel>
:)
<neggles>
the PHY layout is weird.
<stintel>
mattytap__ and my attempts to bring network up on the M200 keep failing
<stintel>
TX works, RX doesn't, both on the real ethernet interfaces as on the switch iirc
<stintel>
spotted that the M200 ran a different FMAN firmware (with different features), replaced it with the firmware with same features as on the M300, same problem :(
<robimarko>
stintel: That screams clock issues to me
<mattytap__>
stintel: I'm just cleaning up github.com/mattytap/m200 this morning and taking a break
<neggles>
so MAC0 goes to PHY1, MAC1 to PHY2, MAC2 to PHY3, MAC3 disabled, MAC4 to PHY4, MAC5 to PHY5... except the ports on the outside are in an entirely different order
<mattytap__>
if and when we (or someone else) cracks with RX issues I will return to the Altivec issue on the 5500
<neggles>
port marked eth0/wan is on enet2...
<stintel>
robimarko: well we're using upstream dtsi as much as possible, and without sources from WatchGuard ... trial and error this kind of thing seems suboptimal :)
<robimarko>
I understand
<stintel>
robimarko: I can send you my M200 if you care :P
<mattytap__>
The dts I'm staging later today has all ports in correct order and is pretty clean considering
<robimarko>
stintel: I am stretched really thin currently
<robimarko>
Just catching up on IPQ40xx DSA again to get that over the finish line
<stintel>
robimarko: no worries :)
<stintel>
ipq is more important
<mattytap__>
I've tried everything I can think of in terms of dts - dpaa, rx-delay, NXP trawled forums
<robimarko>
What ethernet PHY-s does that thing use?
cp-- has joined #openwrt-devel
<mattytap__>
Marvell Alaska 1543 in SGMII for eth0,eth1,eth2
<stintel>
robimarko: Marvell 88E1543 for the first 3 ports
cp- has quit [Ping timeout: 480 seconds]
<stintel>
can't find the switch right now
<mattytap__>
The RX issues also exist at uboot and I hate to say it, I cant get TX to work on the 1543 either
<stintel>
mattytap__: I got perfectly working ethernet in u-boot
<stintel>
yeah ok, that's probably not very useful
<stintel>
pretty sure we tried that already
<mattytap__>
The main thing I concluded this week is the m200 boeard is nearer a T1042D4RDB
<robimarko>
stintel: Ahh, they have a ton of erratas on the 88E1543 even in the A1 silicon
<robimarko>
Decent number related to SGMII/QSGMII
<stintel>
robimarko: but the M300 uses 88E1543 also and works fine
<stintel>
robimarko: I suspect the problem is in the MAC actually
<mattytap__>
One humourous thing, googled 'watchguard m200' one last time yesterday; after discounting the OEM links I discovered a new post in April. Yey - someone may have solved this already! - it was from stintel
<neggles>
same phy mode?
<mattytap__>
Yes
<stintel>
mattytap__: hmmm which post is that? :P
<mattytap__>
MB-M200 => mdio list
<mattytap__>
FSL_MDIO0:
<mattytap__>
2 - Marvell 88E1543 <--> FM1@DTSEC2
<mattytap__>
0 - Marvell 88E1543 <--> FM1@DTSEC3
<mattytap__>
1 - Marvell 88E1543 <--> FM1@DTSEC1
<mattytap__>
4 - Marvell 88E6171 <--> FM1@DTSEC5
<mattytap__>
16 - Marvell 88E6171 <--> FM1@DTSEC4
<stintel>
ah :D
<stintel>
twas on IRC
<neggles>
=> mdio list
<neggles>
FSL_MDIO0:
<neggles>
5 - Generic PHY <--> FM1@DTSEC2
<neggles>
6 - Generic PHY <--> FM1@DTSEC5
<neggles>
7 - Generic PHY <--> FM1@DTSEC6
<neggles>
4 - Generic PHY <--> FM1@DTSEC1
<stintel>
I didn't recall posting somewhere, then again I've been all over the place
<neggles>
FM_TGEC_MDIO:
<neggles>
1 - AR8035 <--> FM1@DTSEC3
<neggles>
what the shit, watchguard?
<mattytap__>
MB-M300 => ms dio list
<mattytap__>
0 - Marvell 88E1543 <--> FM1@DTSEC10
<mattytap__>
FSL_MDIO0:
<mattytap__>
1 - Marvell 88E1543 <--> FM1@DTSEC1
<mattytap__>
2 - Marvell 88E1543 <--> FM1@DTSEC2
<mattytap__>
4 - Marvell 88E6171 <--> FM1@DTSEC4
<mattytap__>
16 - Marvell 88E6171 <--> FM1@DTSEC3
<mattytap__>
stintel: my bad - April 2021.
<neggles>
mattytap__: good to know they're just as nuts on every other device
<neggles>
why is the wan port on dtsec3
<stintel>
rng ;)
<neggles>
autorouter/10
<robimarko>
The number of erratas for the PHY is insane, like 20+ pages
<stintel>
:P
<neggles>
robimarko: they have not left out all of the interesting stuff, no, but they *have* left out the bootloader... so now i get to complain about that
<neggles>
(UCK-G2)
<neggles>
but at least i have kernel and userland sources
<robimarko>
stintel: I suppose they have hacked around one of the erratas, but without the GPL to see whats being written its impossible to know
csharper2005 has joined #openwrt-devel
<robimarko>
neggles: They always leave out the bootloader intentionally, cause otherwise it would be really easy to disable their signature check
<neggles>
the signature check can be disabled from the uart command line
<neggles>
stintel: i would also like to take a moment to appreciate that the T20 has that annoying passworded u-boot, but it's not running secure boot.
<neggles>
nor do they have tfa in there.
<neggles>
so you could just flash a regular ole LS1043A-RDB u-boot image to it (with a few dts tweaks) and it'd work fine.
<mattytap__>
neggles: nit like tom cruise movies then. m300 is great, t30 too
<mattytap__>
**bit**
<neggles>
t20 is LS1023A, t40 is LS1043a, T80 is LS1046A
<neggles>
i would like an T80 but havent found a cheap one yet
<neggles>
got this T20 from a pawn shop, it was unactivated and had a year of license on it
pepe2k has quit [Remote host closed the connection]
<stintel>
:P
<neggles>
WARNING failed to get smmu node: FDT_ERR_NOTFOUND
<neggles>
that's new.
<neggles>
oh, no it isn't, so why have i lost boot console
<mattytap__>
m200 I've just added the kernel log to github.com/mattytap/m200
<ynezz>
if you're referring to the following $$(_TARGET): $$(if $$(KERNEL_INSTALL),$(BIN_DIR)/$$(KERNEL_IMAGE)) then I parse it as: if KERNEL_INSTALL is set to some value then add $(BIN_DIR)/$$(KERNEL_IMAGE) as build prerequisite for $$(_TARGET)
<ynezz>
so $(BIN_DIR)/$$(KERNEL_IMAGE) is going to be build before $$(_TARGET)
csrf has joined #openwrt-devel
kkurbjun has quit [Read error: Connection reset by peer]
<neggles>
i'm a bit confused as to why you wouldn't want that
<neggles>
but thanks
<neggles>
OH! I know! it puts the kernel itb in the bin/target/blah/blah folder with the images
<neggles>
my problem was not related to this - turns out, if you want a ramboot image, you have to tell the buildsystem to generate a ramdisk! who knew? :P
<stintel>
https://git.openwrt.org/?p=openwrt/openwrt.git;a=blob;f=include/image.mk;h=b23e192c5d0c3e678f369bf65ce6562d83586a83;hb=HEAD#l360 --> how can I populate $2 in Device/Init to modify the image name? don't see it
<robimarko>
Thats good, those GZIP only devices with 3MB limit are pretty much dead
dangole has joined #openwrt-devel
<f00b4r0>
the AX820 uses the default build which is lzma too iirc
<stintel>
dangole: if you were to introduce a fit kernel for an existing device, how would you do it? new device entry with -fit in the image name and documenting the requires uboot env changes in the commit message so that people don't accidentally soft-brick, keep the legacy device around for a while before deleting, or better suggestion?
<stintel>
basically no more hastle with loading the dtb to some memory location
<Habbie>
i see it, nice
<stintel>
and it could even include rootfs iirc
<dangole>
stintel: I guess it's a nand based device and there are some benefits from using uImage.FIT instead of legacy image? (or you mean uImage.FIT which includes squashfs?)
<stintel>
dangole: sdcard (qoriq)
<stintel>
but I ran into issues with kernel becoming to big and had to change the memory address for the dtb
<stintel>
and fit is just cleaner
<dangole>
stintel: ok got it. yes, FIT is the way to go. I guess changing U-Boot bootcmd can be part of installation instructions, so you could just drop the old legacy image...
<stintel>
dangole: problem is the device support is already there
<dangole>
stintel: so COMPAT_VERSION maybe with a message hinting to update U-Boot env first?
<stintel>
I'll have to research that, thanks for the suggestion
<stintel>
definitely sounds cleaner than adding a new device profile for that
danitool has quit [Ping timeout: 480 seconds]
<hurricos>
neggles: yeah, unfortunately I did create an account but see no order :\
danitool has joined #openwrt-devel
_4fk has joined #openwrt-devel
<_4fk>
hello
<_4fk>
i wanna contribute regarding developing. is there a todo-list anywhere i can take a look on?
sorinello has quit [Ping timeout: 480 seconds]
sorinello has joined #openwrt-devel
<stintel>
ERROR: new format image overwritten - must RESET the board to recover
<stintel>
sigh this is so annoying for stuff in a rack
<Slimey>
:/
<Slimey>
or when its 100's of miles away
<stintel>
well it's short length in full depth rack and the power switch is in the back, annoying enough ;)
<stintel>
but the rack is on the floor below
<Slimey>
ack
<Slimey>
remote pdu is the best thing ever
<stintel>
yeah
<stintel>
should probably just get it over with and order one
<stintel>
but then I'd want to get one that can measure each port too :P
<stintel>
sigh, main m300 again doesn't come back after sysupgrade
<stintel>
fuck sake
dangole has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
<stintel>
also had a pretty nasty oops during sysupgrade
<stintel>
good thing I documented USB recovery on the wiki, mostly copy-paste :)
ecloud_ has quit [Ping timeout: 480 seconds]
jlsalvador has joined #openwrt-devel
<_4fk>
can someone tell me where i can find a list of bugs regarding OpenWRT or do you prefer to ignore other interested parties which wanna conbtribute?
<schmars[m]>
if there is any device that you run openwrt on, see if there's issues for it, maybe there's something you can verify, fix, or reproduce. most packages are separate from the base repo, here: https://github.com/openwrt/packages
<_4fk>
schmars[m] thank you very much. yes, i will do it. i will study the bugs now.
pepe2k has quit [Remote host closed the connection]
<slh>
robimarko, dwfreed: the ipq806x seem to be ipq806x SOC side, not ath10k related. I've run my nbg6817 with wireless disabled (albeit not unloading ath10k completely) for a couple of months, it still rebooted within 2-14 days - haven't tried setting it to max frequencies yet, as I cheated by going x86_64 for my router...
<slh>
*ipq806x reboot issue
<robimarko>
I was suspecting that as I know that Ansuel has been hunting that for a while
shibboleth has joined #openwrt-devel
torv_ has quit [Remote host closed the connection]
torv_ has joined #openwrt-devel
csrf has quit [Ping timeout: 480 seconds]
grift_ has quit [Read error: No route to host]
MAbeeTT2 has quit [Ping timeout: 480 seconds]
<hurricos>
ffs
<hurricos>
WARNING: could not set bootargs FDT_ERR_NOSPACE
<hurricos>
on AP3825i, AGAIN.
<hurricos>
Even after the padding. So the padding fix Chunkeey suggested only worked temporarily.
<hurricos>
s/temporarily/by random chance/
<hurricos>
this irks me to no end, I swear to God
<hurricos>
Let me verify the OpenWrt build isn't just somehow missing the padding step
<hurricos>
(the workaround isn't hard -- just change the bootcmd to add more fdt resizes -- but it's incredibly annoying that this breaks AGAIN after so much work went into fixing it.)
<hurricos>
the vendor u-boot simply refuses to recognize the free space it's been given as being its own unless it's already resized to use it, and it refuses to resize by any more 4k pages than are needed until it's too late
<hurricos>
vendor u-boot doesn't care that it's had tons of space padded to the dtb, it just. won't. touch. it.
grift has joined #openwrt-devel
MAbeeTT has joined #openwrt-devel
<hurricos>
Oh, hmm.
<hurricos>
Looks like the DTB_SIZE is just not sticking.