danitool has quit [Remote host closed the connection]
danitool has joined #openwrt-devel
plntyk has quit [Read error: Connection reset by peer]
dwmw2_gone has quit [Ping timeout: 480 seconds]
dorf has quit [Remote host closed the connection]
dorf has joined #openwrt-devel
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
bobthebuilder998 has joined #openwrt-devel
bobthebuilder999 has quit [Ping timeout: 480 seconds]
dwmw2_gone has joined #openwrt-devel
dwmw2_gone is now known as dwmw2
mangix has quit [Read error: Connection reset by peer]
mangix has joined #openwrt-devel
<mangix>
Pepes: with your ax devices, any luck with DFS? I'm getting hostapd: Frequency 5560 (primary) not allowed for AP mode, flags: 0xf0797b NO-IR RADAR on my ax router.
<digitalcircuit>
Those following here might have already seen this, not sure. I've still got to test actual fixes now, but at least it's gone from the convoluted "Deja Dup backup over SFTP to USB HDD on router" to "run this Ash shell script on the router, or optionally use the computer helper shell script to track and organize results."
<digitalcircuit>
(I'll still verify any fixes with the Deja Dup SFTP to USB HDD test, but the shell script is a lot less destructive to backups for testing until I get to that stage.)
bobthebuilder998 has quit [Ping timeout: 480 seconds]
cp- has joined #openwrt-devel
guerby_ has quit [Read error: Connection reset by peer]
<slh>
digitalcircuit: btw. my nbg6817 just crashed after 14 days and 7 hours of uptime, during rather light background traffic - the g10 is still going (roughly the same uptime), let's see how it goes
<digitalcircuit>
slh: Noted! Which is.. huh. I'm not sure. At least that sounds like 19.07 levels of crash instead of what I was experiencing with 21.02.
<digitalcircuit>
Optionally, if you have the time and capacity, would you give the QA script a try, see if that causes a quick (within 10 hours) crash?
<slh>
what worries me, is that it's most of the time bang on around 2 weeks (let's say 11-14 days) - that just looks weird (and yes, I had similar/ worse issues with lantiq and ath79 before, so I'm still not sold on the idea that it's the fault of ipq806x)
<slh>
in 2017/ 2018, I had around ~11 months of uptime on the nbg6817 (before a blown fuse cut it short)
<digitalcircuit>
Ah, good point! Maybe it's an uptime-related crash, possibly entirely unrelated to what I'm experiencing. I guess I need to fix this crash for myself before I can even approach a week of uptime (given weekly backups)...
ashkan has joined #openwrt-devel
Tusker has joined #openwrt-devel
greearb__ has quit [Remote host closed the connection]
greearb__ has joined #openwrt-devel
goliath has joined #openwrt-devel
<mangix>
has anyone modified their router to use external antennas?
<mangix>
seems entirely doable
<mangix>
I've never made holes in plastic though
<Tusker>
drill... dremel... soldering iron...
<mangix>
Tusker: I don't know if it will crack
<mangix>
how big do SMA connector holes have to be?
<Tusker>
i suppose it depends on the SMA connector you purchased... use a micrometer and measure it
<Tusker>
or get something like a bulkhead one with a screw thread
<brpr>
it won
<brpr>
It won't crack if you start with a small bit and go upwards
<Tusker>
i find that actually drilling backwards is the safest method, the drill bit can catch sometimes even if you drill a starter hole
<digitalcircuit>
Drilling backwards - would that be drilling from the inside of case towards the outside?
<Tusker>
i meant set the drill direction in reverse, so it is spinning in a reverse fashion
<Tusker>
much slower to etch away at the plastic, but won't "grab"
<brpr>
Listen to Tusker lol, he probably has way more experience than me
<digitalcircuit>
Oh! I didn't even know that was a thing you could do. Noted for the future :D (Thankfully I hadn't cracked my router's case making a hole for the serial console cable, but I was probably just lucky.)
<Tusker>
especially important when you are using clear acrylic... it is so easy to make it crack... a 1m x 1m sheet can look pretty awful easily
nitroshift has joined #openwrt-devel
dorf has quit [Remote host closed the connection]
<brpr>
Tusker, by the way wink wink do you know how to help me get the PHYs? vendor bootlog is quite vague wink wink
dorf has joined #openwrt-devel
<Tusker>
you don't know what the PHYs are, or don't know how to connect them up to the ethX ?
Tapper has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
<aparcar[m]>
Any 8 switch recommendations?
<brpr>
Tusker, both :)
<Tusker>
ok, did you build mdio-tools and kmod-mdio-netlink in your custom build from master ?
<brpr>
nope
<Tusker>
if you do that, you can see more information on what is connected the mdio bus
<brpr>
yeah, compiling now
<brpr>
Tusker, first I need to clean up my home partition, it's completely full
lmore377 has quit [Quit: No Ping reply in 180 seconds.]
norris has quit [Ping timeout: 480 seconds]
<Tusker>
don't you hate when you look at your share holdings, and one of the shares looks like it made $50k profit... but it turns out you are an idiot and entered the holdings manually and it was simply fiction
lmore377 has joined #openwrt-devel
norris has joined #openwrt-devel
<Tusker>
no wonder they didn't accept my sell request
<robin_>
what is the requirement for mdio-tools and kmod-mdio-netlink? I don't seem to find any of them in "make menuconfig" on current master...
<brpr>
in makefile for your device\
<Tusker>
there is a kernel patch already in master that is needed for 5.4, 5.10 works normally
<Tusker>
perhaps you need to ./scripts/feeds update -a && ./scripts/feeds install -a
decke has joined #openwrt-devel
<robin_>
thanks, i'll have a look.
lmore377 has quit [Quit: No Ping reply in 180 seconds.]
lmore377 has joined #openwrt-devel
<brpr>
Tusker, it actually compiled lol
<Tusker>
once you've pushed it to the device, you should be able type 'mdio bus'
rejoicetreat has joined #openwrt-devel
<brpr>
Tusker, sorry for my inactivity but I got into an agreement with the guy that sold me the faulty USB-serial converter that he'll send me one based on a (supposedly) better chip
<Tusker>
ask him what chip the new one is
<brpr>
FT232RL
<Tusker>
yeah, that one is much better than CP2102
<brpr>
and I get it for free. this seller is actually very good
<Tusker>
FT232RL is more ex than CP2102
<brpr>
Tusker, I know why it compiled right! I was accidentally building for x86 :P I've been messing around with x86 builds lately
<Tusker>
:)
mrkiko has joined #openwrt-devel
<brpr>
yeah, now compiling for the right target
<mrkiko>
Hello folks!!
<brpr>
mrkiko, hi there!
<mrkiko>
brpr: hi! How are things going with your porting?
<brpr>
mrkiko, it's going great! I'm getting a FT232RL based adapter from the seller that sold me the 3.9V unit and I'm building the first build for today with mdio-tools to hopefully find out more about PHYs
Thermi_ has joined #openwrt-devel
Thermi has quit [Quit: Meet your opposition - Profane and disciplined - Take back your pride - With a pounding hammer]
danitool has joined #openwrt-devel
<brpr>
Tusker, mdio not found :(
aleasto has joined #openwrt-devel
<Tusker>
are you sure you used =y rather than =m ?
<Tusker>
in your .config
<brpr>
i'm sorry Tusker but I didn't mess with any .configs
<Tusker>
grep mdio .config
<Tusker>
what does it show ?
<brpr>
CONFIG_DEFAULT_kmod-mdio-netlink=y
<brpr>
CONFIG_DEFAULT_mdio-tools=y
<brpr>
CONFIG_PACKAGE_kmod-mdio-gpio=m
<brpr>
CONFIG_PACKAGE_kmod-of-mdio=m
<brpr>
CONFIG_PACKAGE_kmod-switch-bcm53xx-mdio=m
<brpr>
CONFIG_PACKAGE_kmod-mdio=m
<Tusker>
and you flashed that? perhaps just upload the package files and install manually with opkg
<brpr>
yeah I flashed that one
<brpr>
so install mdio-tools?
<Tusker>
yeah
<brpr>
hmm... i can't find it
<brpr>
oh and great - front usb 3 doesn't work
goliath has joined #openwrt-devel
<Tusker>
brpr: do you see any bin/packages/*/packages/mdio-tools*ipk ?
<Tusker>
ie, from the openwrt git directory, you can type ls bin/packages/*/packages/mdio-tools*ipk
<brpr>
bash-5.1$ ls bin/packages/*/packages/mdio-tools*ipk
<brpr>
ls: cannot access 'bin/packages/*/packages/mdio-tools*ipk': No such file or directory
<brpr>
the local repo is a clone from the git repo
<Tusker>
that's a bit weird...
nitroshift has quit [Remote host closed the connection]
<Tusker>
and you only use make menuconfig to select things in .config, and then do a make after that, and you see the mdio-tools in the list of things being compiled ?
nitroshift has joined #openwrt-devel
<brpr>
yes
<Tusker>
what about something like ./staging_dir/target-arm_cortex-a9+vfpv3-d16_musl_eabi/root-mvebu/usr/bin/mdio ?
<Tusker>
do you see that in the build tree ?
swiftgeek has joined #openwrt-devel
<brpr>
nope
<Tusker>
it must not have compiled for some reason
<Tusker>
ie, ls staging_dir/target*/*/usr/bin/mdio
<Tusker>
root@wrt1900:~# ls /sys/module/mdio_netlink/
<Tusker>
seems to me, yes
<PaulFertser>
brpr: probably you eth1 doesn't need any mdio control and just fixed-link 1000 full-duplex?
<PaulFertser>
brpr: without any &phy reference etc. Have you tried it like that?
<brpr>
i'm gonna try it
<brpr>
PaulFertser, what about qca,ar8327-initvals? is it needed
<PaulFertser>
brpr: I'd try with defaults first.
<PaulFertser>
brpr: so without extra configuration for eth1
<brpr>
PaulFertser, so status ok and fixed link speed and full duplex?
<PaulFertser>
brpr: yep
AbiHalib has joined #openwrt-devel
<brpr>
PaulFertser, how do I check if it's working? sorry, I forgot :)
<brpr>
PaulFertser, now eth1 is completely dead. 0 bytes
brpr has quit [Quit: Taking a long break.]
lynxis has quit [Ping timeout: 480 seconds]
<AbiHalib>
Hello everyone.
<AbiHalib>
I have a question, how do I add new feed (subdirectory with my code files) to feeds?
<Tusker>
need to edit the feeds.conf file
<AbiHalib>
Yeah, but nobody adds packages to feeds like that - I mean, we have svn links here, but all packages are already here in feeds and dl folder
<Tusker>
so, what I do, is create a new directory, say, /home/tusker/dev/my-openwrt-packages/, and then add src-link tusker-packages /home/tusker/dev/my-openwrt-packages
<Tusker>
then ./feeds update will index it
<AbiHalib>
Ok, thanks!
AbiHalib has quit [Quit: Page closed]
lynxis has joined #openwrt-devel
<mrkiko>
PaulFertser: so it's pretty ommon to faith with switches and eth interfaces
<mrkiko>
PaulFertser: is this especially true on some targets or is it general?
<PaulFertser>
mrkiko: to faith?
<mrkiko>
i am compiling lots of drivers/net/phy/rtk/rtl* code, even tough I'm compiling for mt7622 in a hardware with no realtek components as far as I know
<mrkiko>
PaulFertser: sorry, I meant - fight
<karlp>
just let the extra compiles wash over you, bathe in their light.
<PaulFertser>
mrkiko: I think it's relatively uncommon but apparently some MACs peripherals require specific PLL config or something.
<mrkiko>
karlp: :D
cp- has quit [Ping timeout: 480 seconds]
dorf has quit [Remote host closed the connection]
dorf has joined #openwrt-devel
mrkiko has quit [Ping timeout: 480 seconds]
rejoicetreat has quit [Ping timeout: 480 seconds]
mrkiko has joined #openwrt-devel
dorf has quit [Remote host closed the connection]
dedeckeh has joined #openwrt-devel
dorf has joined #openwrt-devel
Tapper has quit [Ping timeout: 480 seconds]
<mrkiko>
I upgraded to openwrt git HEAD - why do my lan wi-fi clients gets like "isolated"? They can talk to router but not to each other?
<mrkiko>
isolate doesn't seem to be true by default
<PaulFertser>
isolate is true by default but the bridging should work on linux software bridge level
<mrkiko>
PaulFertser: well, I did a sysupgrade -n
<mrkiko>
PaulFertser: so I started fresh; and done my changes, but I'm pretty sure they shouldn't cause this :)
<mrkiko>
but observing tcpdump, I see the packets for another host in my lan don't get delivered at all
<mrkiko>
A brave thing I did is to use mt76 git head - may this be the cause?
<mrkiko>
you're right, ap_isolate=1
ldir has quit [Quit: +++Divide by cucumber error+++]
ldir has joined #openwrt-devel
decke has quit [Remote host closed the connection]
decke has joined #openwrt-devel
Tapper has joined #openwrt-devel
ldir has quit [Quit: +++Divide by cucumber error+++]
<karlp>
you just want a comment that you should check that the path you use is persistent?
<dorf>
better commented out defaults I think would be better. it's not a safe assumption that people know /var isn't persistent.
<dorf>
additional comments suggesting a persistent location wouldn't hurt, either, but the most important thing in my view is to not point the commented out defaults at /var/
f12 has quit []
f12 has joined #openwrt-devel
Tusker has quit [Quit: Time wasted on IRC: 8 hours 21 minutes 7 seconds]
f12 has quit []
f12 has joined #openwrt-devel
<zimbabwe>
how do I add package to feeds without modifiyng the feeds.conf file?
<mrkiko>
zimbabwe: I don't know if it's possible - unless you put it in some places like package/network or whatever
brpr has joined #openwrt-devel
<brpr>
PaulFertser, hi again. Sorry that I'm probably really annoying you but now eth1 is dead. 0 bytes in rx_bytes.
zimbabwe has quit [Ping timeout: 481 seconds]
nitroshift has quit [Quit: Gone that way --->]
jbowen has quit [Quit: leaving]
<PaulFertser>
brpr: how is eth1 usually defined with other devices using the same SoC? Have you checked them all?
<brpr>
yes. not much different from eth0, just phy2 instead of phy1 and different pll data.
<brpr>
PaulFertser, the weird thing is that eth0 controls all ports right now, all ports do work.
<PaulFertser>
brpr: so probably you need pll data too.
<PaulFertser>
brpr: what do you mean by "controls"?
<brpr>
because all ports are mapped to eth0
zimbabwe has joined #openwrt-devel
<PaulFertser>
brpr: still do not understand you, what means mapped?
<brpr>
PaulFertser, sorry for the confusion then. If I plug my laptop in to any port it's on eth0. eth1 doesn't transmit any bytes.
decke has quit [Quit: Leaving.]
<PaulFertser>
brpr: well, your eth0 is connected to the switch to port 0. So if you configure the switch to bridge all the ports together of course you'll see there all the traffic.
<PaulFertser>
brpr: I assume eth1 is connected to port 6 there, but probably pll is wrong so eth1 is not seeing the same data eth0 sees.
<brpr>
right. so eth1 is supposed to be a "clone" of eth0? sorry, i'm not too educated on linux networking, especially on routers
<brpr>
PaulFertser, that'd explain why it's always UP
<PaulFertser>
brpr: no, not clone, just another interface that's connected to a separate switch port.
<brpr>
I kinda don't understand. So eth1 would be the WAN port?
goliath has quit [Quit: SIGSEGV]
<brpr>
PaulFertser, if so then that'd bring us to the next part - how to get the PHYs? I can't seem to get mdio-tools to compile
ashkan has quit [Ping timeout: 480 seconds]
<PaulFertser>
brpr: the vendor bootlog suggests eth1 doesn't have its own mdio connection to phy.
<brpr>
PaulFertser, so then what's it for?
<PaulFertser>
brpr: one is supposed to configure the switch to bridge the "wan" port with port connected to eth1 and all the other ports to port connected to eth0. That way you won't be limited by a single switch-SoC link.
<brpr>
Right. Now what do we do next?
Tapper has quit [Ping timeout: 480 seconds]
<brpr>
PaulFertser, I'm not getting mdio-tools compiled :(
<PaulFertser>
brpr: I do not think you need them based on vendor bootlog.
<brpr>
Ah, alright! So now - could you please point me to a line that says the PHYs? Only thing I can find that's close is eth1: Atheros AG71xx at 0xba000000, irq 5, MAC 30:5A:3A:5F:EA:E4
<PaulFertser>
brpr: yes, so no phy configuration (mdio) is performed. While for eth0 there is.
<brpr>
PaulFertser, is "0xb9000000" the PHY? eth0: Atheros AG71xx at 0xb9000000, irq 4, MAC 30:5A:3A:5F:EA:E0
<PaulFertser>
brpr: no, that's the peripheral address of eth0 for MMIO control of the SoC integrted peripheral.
<PaulFertser>
brpr: there's another line about eth0 and phy in your log.
<brpr>
ah, right. that caught me off-guard
<brpr>
uid=004dd036 ?
<PaulFertser>
brpr: that line, yes
<brpr>
ok, now what can we get from that?
<PaulFertser>
brpr: I would assume eth1 is not using mdio bus to control any phy and it's just a fixed link to port 6 of the switch.
<brpr>
PaulFertser, so what's next?
<PaulFertser>
brpr: I would try looking at different pll settings
dorf has quit [Remote host closed the connection]
<brpr>
PaulFertser, you mean phy-handles?
<brpr>
oops, pll data
<brpr>
PaulFertser, just making wild guesses? I'm sorry but I don't know what to mess with as I only see addresses that y'know - I don't know what to change and why to change
<PaulFertser>
brpr: unfortunately I have no idea about pll data for those boards too.
<brpr>
oh no, that sucks
zimbabwe has quit [Ping timeout: 480 seconds]
Larhzu has joined #openwrt-devel
rejoicetreat has quit [Remote host closed the connection]
rejoicetreat has joined #openwrt-devel
AdrianFL has joined #openwrt-devel
goliath has joined #openwrt-devel
ashkan has joined #openwrt-devel
AdrianFL has quit [Remote host closed the connection]
Borromini has joined #openwrt-devel
ashkan_ has joined #openwrt-devel
ashkan has quit [Ping timeout: 480 seconds]
dedeckeh has quit [Remote host closed the connection]
Borromini has quit [Quit: leaving]
Borromini has joined #openwrt-devel
LPaci has joined #openwrt-devel
ashkan has joined #openwrt-devel
ashkan_ has quit [Ping timeout: 480 seconds]
<slh>
aparcar[m]: ZyXEL gs1900-8 tends to sell quite cheap, flash is a bit limited, but flashing is easy and you get a populated 3.3V UART header (2.54mm spacing) and three standard philips heads screws to open the case. used, you can regularly find them around 15-20 EUR
<Borromini>
someone else on the Realtek/ZyXEL boat huh? :^)
<slh>
hehe, we'll see ;) he was asking for an 8 port switch
<Borromini>
;)
<Borromini>
i gave 5.10 a quick shot on my GS1900-8HP v1, and it seems to run fairly well
Borromini has quit [Quit: leaving]
zimbabwe has joined #openwrt-devel
zimbabwe has quit [Ping timeout: 480 seconds]
<stintel>
anyone have a clue how to test mtd / spi-nor quad read ?
brpr has quit [Ping timeout: 480 seconds]
aleasto has quit [Remote host closed the connection]
zimbabwe has joined #openwrt-devel
John has joined #openwrt-devel
LPaci has quit [Remote host closed the connection]
Tapper has joined #openwrt-devel
John has quit [Remote host closed the connection]
zimbabwe has quit [Ping timeout: 480 seconds]
zimbabwe has joined #openwrt-devel
zimbabwe has quit [Ping timeout: 480 seconds]
Random has joined #openwrt-devel
<Random>
How to find out the ram and flash size of my router if it's not mentioned in the website?
<slh>
wikidevi.wi-cat.ru ?
<Random>
I don't have the information about my router's flash and ram size
Random has quit []
<stintel>
aaand it's a brick
Random has joined #openwrt-devel
Random has quit [Remote host closed the connection]
Explorer has joined #openwrt-devel
<Explorer>
Finding out the ram and flash memory of my router?
<stintel>
you try wikidevi replacement(s), fcc, google, or open the device and check the chip labels
<Explorer>
Tried everything except opening the device...
Explorer has quit [Remote host closed the connection]
bobthebuilder999 has joined #openwrt-devel
<bobthebuilder999>
Anyone here a bit more knowledgeable about flashing tp-link devices?
rsalvaterra has joined #openwrt-devel
<bobthebuilder999>
I can load a ramfs image via tftpboot and execute that from RAM just fine. But trying to actually flash the factory.bin results in "00166: Check rsa error."
<PaulFertser>
bobthebuilder999: you can likely flash with sysupgrade from that initramfs image. But will bootloader be happy booting it?
<bobthebuilder999>
My problem is making this accessible to the average joe who isn't going to open the device or solder anything to it.
<PaulFertser>
bobthebuilder999: if the bootloader is so uncooperative then not much you can do I'm afraid.
<PaulFertser>
bobthebuilder999: attaching serial often doesn't require any soldering
<PaulFertser>
bobthebuilder999: but if it's just the userspace not accepting your factory.bin probably you can find a software exploit in some other part of web-interface. Like it was found for xiaomi devices.
<stintel>
bobthebuilder999: sounds similar to TP-Link OC200
<bobthebuilder999>
PaulFertser: I'm already an idiot with a screwdriver who had to tinker with this for the past two weeks to get anything working. I doubt that I would be able to find any exploits here.
<bobthebuilder999>
And the Deco M4R doesn't really have a weg gui. Setup is done with an app.
<PaulFertser>
bobthebuilder999: I think many people will be happy enough using a method that requires just a regular 3.3 V UART adapter.
<bobthebuilder999>
Probably, yes.
<bobthebuilder999>
You still need to solder a bit because UART is on two small testpoints TP1 and TP2 in the middle of the board.
<PaulFertser>
bobthebuilder999: when I was porting to mir3g-v2 I thought nobody would be flashing it by directly attaching to flash chip. I was pleasantly surprised.
<PaulFertser>
bobthebuilder999: or you can carefully hold the wires while someone else's typing commands
<PaulFertser>
And it's not much to type I guess
<bobthebuilder999>
You could probably do it blind and actually only use RX and GND on the board. But holding the wires without soldering needs a surgeon.
<PaulFertser>
If you can solder to a testpoint you can keep a needle touching it too.
<bobthebuilder999>
Let's first see if stintel's link has the solution I need.
<stintel>
no
<stintel>
only ideas
dangole has joined #openwrt-devel
<stintel>
in fact I just bricked one of my OC200 while testing QSPI mode so that I could finally upstream support for the SPI NOR flash chip
<bobthebuilder999>
Hm, yeah, overwriting keys in flash doesn't sound like fun for a layman like me.
<stintel>
and flashrom refuses to detect the chip on the only 1.8v device I have (odroid xu4)
<stintel>
bobthebuilder999: look at the alternative a bit higher :)
<PaulFertser>
stintel: I think "TUMPA" can go down to 1.8 V and it's useful device in general, cheap too.
<PaulFertser>
stintel: I haven't personally tried it with 1.8 V but I see no reason why it shouldn't work, I checked the buffer IC datasheet.
<stintel>
why does all this shit have to come from outside of eu
<stintel>
since first of july, customs is a nightmare in some EU countries
<bobthebuilder999>
I think even the other method is a bit too much for me.
<PaulFertser>
stintel: are you sure it's not the target SoC trying to talk to the flash interfering with your flashrom attempts?
<stintel>
PaulFertser: the chip is removed from the soc\
<PaulFertser>
Ah
<rsalvaterra>
stintel: I've seen a lot of stuff on AliExpress being available from Spain, ever since the EU VAT changes.
<stintel>
PaulFertser: I was suspecting that might be the problem when I tried the previous 7 times on the other oc200 that I have here
<stintel>
apparently it's not
<stintel>
now I tried the "loopback" test on the odroid xu4
<stintel>
and when I connect MISO and MOSI pins, output of spidev_test changes
<stintel>
so it must be the right controller
<stintel>
and it must be working
<stintel>
otherwise there would be no change
<stintel>
so I'm out of ideas
<PaulFertser>
It might be not driving /CS
<stintel>
I can't imaging making a mistake in every of the 10+ attempts I made
<PaulFertser>
I mean SPI driver is probably trying to use hardware /CS control and it's not doing the right thing.
<stintel>
and how would one solve that ?
<PaulFertser>
Hard to tell unless you attach a logic analyser. And even a $10 saleae 8 clone could show that.
<stintel>
I have that, but I'm stupid, never figured out how those things work
<PaulFertser>
Should be reasonably straightforward with SPI talking to flash. Basically you start Pulseview, add SPI decoder, tell it which channel is connected to what signal, then attach m25p80 (or even just spi eeprom) decoder on top. Start capture, run flashrom, and you see which signals do what. But hm, 1.8 V might make that unreliable with a regular cheap LA unfortunately.
bobthebuilder998 has joined #openwrt-devel
<PaulFertser>
In this case I would use an oscilloscope to see if /CS is doing the right thing, if clock is visible on SCK etc.
ashkan has quit [Ping timeout: 480 seconds]
<stintel>
yeah unfortunately I don't have a scope
bobthebuilder999 has quit [Ping timeout: 480 seconds]
<rsalvaterra>
stintel: Let me know if you find a cheap 20+ MHz one. :P
bobthebuilder999 has joined #openwrt-devel
bobthebuilder998 has quit [Ping timeout: 480 seconds]
<bobthebuilder999>
okay, I just tried flashing the factory.bin from the web interface of the ramfs.bin and that just made the device reboot and load the old firmware.
<bobthebuilder999>
I thought I was done after successfully running the ramfs.bin on the device, but I guess this odyssey isn't over yet.
<stintel>
of to a great start
<stintel>
[501377.871274] usb 2-12.2.1.3: device not accepting address 45, error -71
<stintel>
so the pos is constantly disconnecting reconneting
<stintel>
sigh
<stintel>
man people that sell this kind of broken shit should be shot
<stintel>
PulseView (Not Responding)
<stintel>
feels like I'm on fucking windows
<stintel>
goddamnit
dangole has quit [Remote host closed the connection]
dangole has joined #openwrt-devel
<mangix>
stintel: what?
<mangix>
stuff just works on windows
<stintel>
you're funny
<mangix>
wonder if I should mod this router's antennas. People on forums say there's no point
<stintel>
should I try the same with the LA and the SPI NOR chip attached ?
<stintel>
also, what kind of monster monitor do you need to properly view this :/
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<mangix>
at least 3 other people are testing/running musl 1.2 on OpenWrt. cool.
<stintel>
why did you close the pr
<rsalvaterra>
Yeah, I'd like to know too… I mean, musl 1.2.2 is working wonders for me.
<karlp>
are you sure you have mosi/miso the right way around?
<mangix>
stintel: no activity
<karlp>
I'm not really sure what the goal is, but that looks like decent spi at least? I would have expected to see more decoding frrom the flash decoder tough
<karlp>
like, "this is attempting to identify a flash chip, and failing..:"
<rsalvaterra>
mangix: I'm pretty sure one of the first things to be merged after 21.02 is out is your musl bump… :/
<rsalvaterra>
And let's also forget gcc 10 and dial it all the way up to 11. :)
<stintel>
karlp: i'd like to figure out why the odroid xu4 doesn't detected the GD25LQ80C
<stintel>
rsalvaterra: feel free to take it :)
<stintel>
musl 1.2.2 should have been in master already
<rsalvaterra>
stintel: As soon as I can access my dev machine… :$
<stintel>
I mean even on gentoo where this is considered experimental, musl 1.2.2 is considered stable
* rsalvaterra
yolo-upgraded the kernel 290 km away and it didn't go well after reboot…
<stintel>
:D
<stintel>
only 290km
<stintel>
boooooo
<stintel>
weak!
<rsalvaterra>
xD
<karlp>
stintel: isn't there enough tools in teh spi flash chip kernel driver to say why it's failing? I've never looked, but I thought you could get it to tell you about mismatched jedec codes, and what it did detect and so on.
<stintel>
karlp: the flash chip is nuked so the device doesn't boot
<stintel>
karlp: it works fine on the device istels
<stintel>
itself*
<stintel>
but now it doesn't because no OS
<stintel>
nor u-boot
<stintel>
so I need to restore u-boot
<stintel>
with an external flasher. the odroid xu4 being the only 1.8v device I have, so the only option (no level shifters of that kind)
<rsalvaterra>
stintel: Oh, well… I had actually built it for my laptop, my dad's laptop, and both my servers at home. No issues at all. My dev machine didn't like, I have no idea why.
<stintel>
karlp: I nuked the NOR while testing Quad SPI mode, which the datasheet claims it supports, and the mtd/spi maintainer demanded I test
<stintel>
I test, I brick, I curse, yadi yadi
<mangix>
rsalvaterra: the plan was for it to be merged for 21.02
<mangix>
that didn't happen
<stintel>
no, that never was the plan
<rsalvaterra>
Oh, my.
* rsalvaterra
fetches the popcorn
<mangix>
in any case, multiple people carry the patches locally.
<mangix>
it'll get merged eventually
<rsalvaterra>
In all seriousness, there's the possibility of some package failing to build with musl 1.2.2…
<rsalvaterra>
… but I don't build *all* packages.
<mangix>
rsalvaterra: I've fixed 99% of them
<mangix>
I build with CONFIG_ALL
<rsalvaterra>
How long does that take? o_O
<mangix>
~4 hours
<rsalvaterra>
Does CONFIG_ALL include all the feeds too?
<mangix>
yes
<mangix>
well, except the non default ones like video
goliath has quit [Quit: SIGSEGV]
<rsalvaterra>
Oh, ffs… let's merge it, then. :/
<stintel>
there were some comments from other core devs
<stintel>
recently even
<rsalvaterra>
I believe the plan is to merge it after 21.02 is out, but I could be wrong.
<rsalvaterra>
In any case, 21.02 should be tagged Real Soon Now™.
<stintel>
irclogs/2021/08/OFTC/#openwrt-devel.log:20|23:03:12< hauke> mangix: could you split the gdb patch into multiple patches, one for the version update, one for the musl fixes and one ofr the arc supprot
<mangix>
stintel: that was merged
<stintel>
good, then I'm not up2date
<stintel>
I just want to avoid merging things that have outstanding comments/requested changes
<rsalvaterra>
I think my musl-122 branch is down to 3 commits, let me see…
<mangix>
stintel: not subscribed to the mailing list
<stintel>
mangix: then subscribe
<stintel>
core = ML first
<stintel>
get over it
<mangix>
my experience is different
<karlp>
its more that some developers are only in one, and some are only in the other...
<karlp>
so depending on what you want to get merged.... :)
<rsalvaterra>
stintel: Mostly, but I think people generally prefer kernel bumps as GitHub pulls. :)
<stintel>
nbd had comments on the "remove several gnu headers" patch
<stintel>
are those addressed ?
<mangix>
karlp: I've had the same patch go stale in both, so ¯\_(ツ)_/¯
<stintel>
also, bpftools patch should go *before* musl bump
<mangix>
what comments?
<stintel>
so that it never breaks
<stintel>
you guys really need to learn some basic git principles
<rsalvaterra>
stintel: git rebase -i :)
<stintel>
ending up on a commit that doesn't build during git bisect is really fucked up
<mangix>
found it
<mangix>
I responded to his comment
<mangix>
1 year ago...
<rsalvaterra>
stintel: For kernel development, that's a *basic* requirement. No intermediate commit in a series is ever allowed to break the build, precisely not to break bisectability.
<stintel>
rsalvaterra: exactly. and we try to follow that
<stintel>
so I really shouldn't be bitching about it
<stintel>
and I would appreciate it if the same mindset would be followed while maintaining packages feed
<rsalvaterra>
Heh. The packages feed feels a bit like the wild west, in that regard. :)
<stintel>
a bit?
dangole_ has quit [Remote host closed the connection]
<mangix>
stintel: too hard
dangole_ has joined #openwrt-devel
<mangix>
packages is maintained on github, not on the mailing list
<stintel>
that doesn't mean it should be maintained with lower standards
<mangix>
sure it does
<stintel>
heh
<mangix>
oh no
<mangix>
wireguard wrecked my phone
<rsalvaterra>
mangix: How? o_O
<mangix>
battery
<mangix>
I left it on
<Grommish>
rsalvaterra: i think I figured out the arm issue:)
<rsalvaterra>
mangix: Go backend? Or kernel module?
<mangix>
the former
<mangix>
I don't run custom kernels
<rsalvaterra>
mangix: That's unfortunate. The kernel module works wonderfully on my Le Max 2 (LineageOS 18.1).
<Grommish>
Samsung is killing the cameras if you crack Knox
<rsalvaterra>
mangix: My phone doesn't have NFC. :P
<rsalvaterra>
Grommish: I don't know, it feels a bit brittle, but if there's no other way…
<rsalvaterra>
Have you tried targets other than ARM?
<Grommish>
With Rust? Yeah. mips, mipsel, mips64, aarch64, x86_64, ppc still needs testing
<Grommish>
The proper way to fix arm would be to fix ./include/target.mk
<rsalvaterra>
stintel can test the brand spanking new ppc64 target. :)
<Grommish>
but, that would require changes across all the arm targets
<Grommish>
So, I don't see that happening
<rsalvaterra>
stintel: Come to think of it, my Power Mac G5 is ppc64… And has two Broadcom GbE interfaces… Hm.
<Grommish>
The good news is.. I found out it works because it didn't.. I added the new armv7 targets and kept having errors. realized I had made the host bins for rust befor the target changes *sigh*
<stintel>
I tried rust already on pcc64
<stintel>
weird things happened
<stintel>
(during build)
<stintel>
didn't work \
<Grommish>
I probably miscreated the toolchain defines
<Grommish>
If you ever want to try again, let me know so I can do it proper
<stintel>
iirc there was some weird dependency leakage from somewhere
<stintel>
after qoriq target lands in openwrt I'd be happy to
<mangix>
new gerbera version
<stintel>
for now, that's one of my top priorities
<stintel>
I used a breadbord to connect odroid xu4, GD25LQ80C and LA
<Grommish>
stintel: I'm having to manually define the toolchains depending on the target.. I'm using the gnu defines as a base. The arm defines imploded, so I"m sure the ppc is the same
<stintel>
Grommish: ping me every now and then about that \
<mangix>
hmmm
<Grommish>
I'll look at it when I get done with arm.. Esp since rsal might be able to help with the defines
* mangix
is somewhat irritated mt76 doesn't support DFS
<stintel>
DFS is a joke anyway
<stintel>
get 1 radar event, end up on channel 149 with 20MHz width and -10 dBm tx power forever
<stintel>
we need 6E
<mangix>
the issue is I want channel 112 :)
<mangix>
only way to get it is to fake the country code I guess
<stintel>
which you shouldn't
<stintel>
nor should you do 40MHz channel width in the 2.4 band
<rsalvaterra>
mangix: Or have a custom regulatory.db… O:)
* rsalvaterra
hides
<Grommish>
Channel 14?
<stintel>
and people call me antisocial
<rsalvaterra>
:P
<karlp>
stintel: spi layer looks ok, but I don't know enough about spi flash to say much more.
<rsalvaterra>
stintel: I think we desperately need a *lower* frequency band, not higher, like 6E. Lower frequencies, maybe narrower channels, lower throughputs, but much higher range and penetrability.
<mangix>
ehhhh
<mangix>
The router I have performs worse on 2.4ghz
<rsalvaterra>
How crowded is the 2.4 GHz band there?
<mangix>
14 APs visible on 2.4ghz
<rsalvaterra>
That's not good…
<mangix>
5ghz channels 100-144 are currently empty
<mangix>
and I can't use them as mt76 doesn't do DFS with mt7915
<blocktrron>
mangix: which mt76 chip does not support DFS?
<rsalvaterra>
mangix: If you know there are no weather radars around you, why not override it?
<blocktrron>
IIRC mt7613 should be the only one and that one looks to be in a fucked up stage from the beginning
<mangix>
rsalvaterra: I think there are. I've seen DFS throw me off the channels before
<rsalvaterra>
:(
<mangix>
blocktrron: hostapd: Frequency 5560 (primary) not allowed for AP mode, flags: 0xf0797b NO-IR RADAR
<mangix>
is the error I get
<blocktrron>
with which chip?
<mangix>
mt7915
<blocktrron>
Country for radio set?
<mangix>
US
<blocktrron>
can you try a EU regdom of your choice for good measure?
<blocktrron>
I have a U6 Lite running on chan 100
<mangix>
blocktrron: country was set to driver default. I just set it to US.
<blocktrron>
from the source, mt7915 should support FCC, ETSI as well as JP DFS
<blocktrron>
mangix: so does it work now?
<rsalvaterra>
I'm off to bed guys. Cheers! :)
rsalvaterra has quit [Quit: Leaving]
<mangix>
blocktrron: yeah it does
<mangix>
Interesting that the country is not set by default...
<blocktrron>
improved error message would be helpful indeed
<Grommish>
The JP region allows for channel 14 in 2.4ghz range which is empty in the US
<mangix>
there's a link to the upstream binutils bug
<Grommish>
Ok, bt I still build with your gold-ld patches for mips64. I'm running it on the device 2 foot from me
<blocktrron>
mangix: country is always not set when installing
<blocktrron>
"always not" - definitely on point today
<mangix>
blocktrron: isn't it normally specified in EEPROM?
<blocktrron>
not sure for mediatek - ath9k allowed to set one or more per-regdom powerlevels
<blocktrron>
but mac80211 / OpenWrt never read that from the eeprom
<mangix>
also, yeah. CyberTAN's using ubootenv to specify the country code
<Grommish>
mangix: So, either I have no concept for how the build system and git works (which is possible) or.. they fixed it.. either way, I need to figure out how.. Suggestions?