clayface has joined #openwrt-devel
<hurricos> I was surprised because those chips seem to be connected to SPI, right?
<hurricos> wait! no, that device tree example you linked is clearly not SPI ... I don't think
<hurricos> A device spec sheet shows it being driven by a clock; so to manually test I could bit-bang the clock potentially
<blocktrron1> hurricos: You can use bitbanged SPI, so they can be used on regular GPIO pins too
<blocktrron1> Have a look on the board and check whether you need to drive a GPIO for the CLR input also
<hurricos> OK, I think I have to drive the clock
<hurricos> I'm randomly bit-banging a set of GPIO endpoints
<hurricos> using 'grep -om 1 -e [01] /dev/urandom' as the input source
<hurricos> and now the LEDs are lighting up randomly (yes!)
<hurricos> Now I think I just need to reduce which endpoints I use
<hurricos> based on the sheet I have, I am almost certain CLR is pin 482
<hurricos> as when I send 0 to that all the LEDs go dark forever more
<hurricos> go dark until I hit pin 494*
clayface has quit [Read error: Connection reset by peer]
<hurricos> if I remove 493 from the mix, none of the pins ever go dark, only stay light.
<neggles> hurricos: sounds like it's a shift register
<hurricos> it is :D
<neggles> so you'll have DAT, CLK, CLEAR
<neggles> and that's absolute cake to control :D
<hurricos> Yeah, looks like it so far. 482 is CLR, 493 is ?, 494 is ?
<neggles> oh there should be LATCH also
<hurricos> The spec sheet only has CLR, CLK, A and B
<hurricos> no LATCH
<neggles> ah ok
<hurricos> it does look like B might be "latch" tohugh
<neggles> what's the chip again
<hurricos> LV164A
<neggles> oh LV164A
<neggles> ah yeah hello chip design that's literally 50 years old
clayface has joined #openwrt-devel
<neggles> oh that's interesting, it doesn't have a LATCH/output enable, but there's a NAND on the input
<hurricos> 494 is CLK
<hurricos> for sure
<hurricos> 0/1/0/1'ing it causes things to "move"
<neggles> guess they're just relying on shifting all the bits into it fast enough that nobody notices, which is fine for a LED controller
<neggles> nice
<neggles> 493 is probably A then and B is probably tied high somewhere...
<neggles> or there's another GPIO that freezes it even if you toggle CLK?
<blocktrron1> hurricos: just hook up a probe to the inputs of the chip and check which GPIO is connected to them
<hurricos> blocktrron1: oh that'd be easier probably
<hurricos> None of the other input GPIOs care
<hurricos> so I think B is held high
<neggles> makes sense
<neggles> i mean, why they used that particular chip is beyond me
<neggles> but it works
<hurricos> hmm
<hurricos> OK, I'm not finding realstically correct device tree bindings anywhere
<hurricos> https://mjmwired.net/kernel/Documentation/devicetree/bindings/gpio/fairchild,74hc595.yaml is not actually correct -- I am not connecting to SPI
<hurricos> that said, it claims this is a "generic" shift control register ... so
snh_ has joined #openwrt-devel
snh has quit [Ping timeout: 480 seconds]
GNUmoon has quit [Ping timeout: 480 seconds]
snh has joined #openwrt-devel
snh_ has quit [Remote host closed the connection]
GNUmoon has joined #openwrt-devel
<neggles> hurricos: usually you'd use spi to drive a shift register but there should be a module for doing it via gpio
<blocktrron1> hurricos: when your chip is not connected to a SPI bus of your SoC, you can just use the bitbanged spi-gpio driver and add the binding there
<neggles> yeah i was about to say
<blocktrron1> see the archer c58 / fritzbox 4020 as an example
victhor has quit [Remote host closed the connection]
<neggles> fairchild,74hc595.yaml is not present in 5.10.87 hmm
<blocktrron1> nick[m]1234: can you provide a full bootlog of your nanostation m5?
<blocktrron1> just to be sure it is the correct chip
<blocktrron1> the locking pattern is a bit odd, BP3 looks to be a TB bit on crack, i can not figure out the pattern locking s applied in case it is set
<Namidairo> i wish to hand avm the award for ugliest hardware
<blocktrron1> neggles: pre dt-schema
<blocktrron1> Namidairo: :(
<neggles> ah
<neggles> ahhhh right yes the generic 74x164
minimal has quit []
dangole has quit [Remote host closed the connection]
philipp64 has joined #openwrt-devel
philipp64 has quit [Quit: philipp64]
srslypascal is now known as Guest9330
srslypascal has joined #openwrt-devel
Guest9330 has quit [Ping timeout: 480 seconds]
danitool has quit [Ping timeout: 480 seconds]
<hurricos> blocktrron1: Thank you very much for those examples. I was totally unaware you could do such madness. This is very fortunate :^)
<hurricos> I'll try it immediatley
slh has quit [Ping timeout: 480 seconds]
slh64 has quit [Ping timeout: 480 seconds]
Tapper has joined #openwrt-devel
philipp64 has joined #openwrt-devel
philipp64 has quit []
<neggles> I wonder. can you set up GPIO SPI on top of a GPIO expander that is itself connected to GPIO SPI?
<neggles> if so, how deep does the rabbit hole go
<hurricos> it's like USB
<hurricos> i.e.
<hurricos> horrifying
<neggles> I can see no reason why you wouldn't be able to go several levels deep with an SoC with fast enough GPIO pins
<neggles> limiter would probably be minimum spi clock of the gpio spi driver or final device...
rua has quit [Quit: Leaving.]
fda- has quit [Quit: ZNC - https://znc.in]
fda has joined #openwrt-devel
<Namidairo> oh boy they're cutting vcc traces on the board now
<nick[m]1234> the line: "[ 0.437808] spi-nor spi0.0: mx25l6405d (8192 Kbytes)"
<Namidairo> nick[m]1234: oh yeah, did you crack your ax6s open yet, I was wondering if anyone else got different spi-nand
<nick[m]1234> Namidairo: yep. sadly I cracked it because I did not realize that there are screws under the sticker on the bottom ^^
<nick[m]1234> 1 second, I can send you a photo
<nick[m]1234> I have to find my handy first
<Namidairo> no need, just tell me if it's not the ESMT part
<Namidairo> I almost snapped the top of mine before noticing the screws on the bottom
<Namidairo> I should probably note that down on the wiki page that I found yesterday
<nick[m]1234> It is to dark in the room xD I guess it is something like this ESMT FSQL 1641L8
<nick[m]1234> but maybe also somethine else
<Namidairo> F50L1G41LB, probably.
<nick[m]1234> probably
<Namidairo> i'll probably just put the pr in for the 1gbit parts that aren't in the driver yet
slh has joined #openwrt-devel
<nick[m]1234> Namidairo: cool thanks. I have to solder the pins like described in the last forum post...
<nick[m]1234> seems not that trivial :/
slh64 has joined #openwrt-devel
<Namidairo> I assume the other 3.3 rail stuff is drawing too much for piddly usb adapters to keep up
<Namidairo> worst case scenario I kick down the door with a vpn
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
castiel652 has joined #openwrt-devel
danitool has joined #openwrt-devel
mangix has quit [Read error: Connection reset by peer]
mangix has joined #openwrt-devel
onemarcfifty has joined #openwrt-devel
pmelange1 has joined #openwrt-devel
marc__ has joined #openwrt-devel
onemarcfifty has quit [Ping timeout: 480 seconds]
pmelange1 has left #openwrt-devel [#openwrt-devel]
onemarcfifty has joined #openwrt-devel
marc__ has quit [Ping timeout: 480 seconds]
onemarcfifty has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
fda has quit [Quit: ZNC - https://znc.in]
onemarcfifty has joined #openwrt-devel
castiel652 has quit [Ping timeout: 480 seconds]
onemarcfifty has quit [Remote host closed the connection]
onemarcfifty has joined #openwrt-devel
marc__ has joined #openwrt-devel
fda has joined #openwrt-devel
onemarcfifty has quit [Ping timeout: 480 seconds]
marc__ has quit [Remote host closed the connection]
marc__ has joined #openwrt-devel
Borromini has joined #openwrt-devel
marc__ has quit [Ping timeout: 480 seconds]
<neggles> hurricos: you got a GPL dump from extreme for these APs, iirc? feel like sharing? they haven't even activated my support portal account yet >:(
rua1 has joined #openwrt-devel
nlowe has joined #openwrt-devel
hitech95 has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
nlowe has quit []
nlowe has joined #openwrt-devel
<nick[m]1234> blocktrron1: I think it was not the lock, it was just missing flash... Sorry. Wrote just a new email via mailinglist
<blocktrron1> nick[m]1234: i doubt that
nlowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<blocktrron1> you have over 2MB of free overlay space
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
victhor has joined #openwrt-devel
<nick[m]1234> blocktrron1: you could be right. I tftpflashed the factory image. Probably because of that I did not see the jaffs2 errors
<JuniorJPDJ> hey, what is .uc file? I'd like to play a bit with fw3 to add one feature
<blocktrron1> nick[m]1234: https://paste.debian.net/1224795/
<blocktrron1> Can you build with this patch? just threw this together, gives us an idea if kernel is locking by itself and if not, some assertions down the line in the unlock routine fail
nlowe has joined #openwrt-devel
<blocktrron1> If you look at the datasheet of you flash chip, the locking behavior is in line with the codes expectation for all states where BP3 is unset
<blocktrron1> Anyways, this patch always clears BP{0-3}
dangole has joined #openwrt-devel
castiel652 has joined #openwrt-devel
<nick[m]1234> blocktrron1: Thanks I will test it! :)
pmelange has joined #openwrt-devel
<nick[m]1234> blocktrron1: it did not help. still have those jaffs2 errors
onemarcfifty has joined #openwrt-devel
nlowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
rua1 has quit [Quit: Leaving.]
nlowe has joined #openwrt-devel
rua has joined #openwrt-devel
nlowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<JuniorJPDJ> how can I test local firewall3 sources? i mean how to prepare env to build and run it
nlowe has joined #openwrt-devel
snh_ has joined #openwrt-devel
castiel652 has quit [Ping timeout: 480 seconds]
snh has quit [Remote host closed the connection]
onemarcfifty has quit [Read error: Connection reset by peer]
snh_ has quit [Read error: Connection reset by peer]
snh has joined #openwrt-devel
nlowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
castiel652 has joined #openwrt-devel
nlowe has joined #openwrt-devel
onemarcfifty has joined #openwrt-devel
danitool has joined #openwrt-devel
nlowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
KGB-1 has joined #openwrt-devel
Borromini has quit [Quit: leaving]
dangole has quit [Ping timeout: 480 seconds]
nlowe has joined #openwrt-devel
nlowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zenox has joined #openwrt-devel
<blocktrron1> nick[m]1234: please try https://paste.debian.net/1224805/ and post the output from dmesg
<hauke> JuniorJPDJ: You can ue SRC_TREE_OVERRIDE and point to a local checkout of fw3
<blocktrron1> you'll have to flash the resulting image using TFTP from the bootloader, as i suspect the current firmware will not be able to write to the flash
zenox has quit [Remote host closed the connection]
_0x4a6f has quit [Read error: Connection reset by peer]
_0x4a6f has joined #openwrt-devel
minimal has joined #openwrt-devel
<nick[m]1234> blocktrron1: the patch does not apply but I modified it a bit
<nick[m]1234> now
onemarcfifty has quit [Read error: No route to host]
onemarcfifty has joined #openwrt-devel
<nick[m]1234> there is no debug output in dmesg?
Tapper has quit [Ping timeout: 480 seconds]
<hurricos> also, let me create a .torrent file
<hurricos> neggles: Again, you're talking about Extreme Networks? you may want to reach out to nlowe, he is pretty much the Extreme Networks SME here.
zenox has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
rmilecki has joined #openwrt-devel
castiel652 has quit [Read error: Connection reset by peer]
<JuniorJPDJ> hauke: thanks, will play with it
<swalker> updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html
<hurricos> if LEDs are driven by a shift register ...
<hurricos> can I possibly expect to be able to light up one LED without lighting the others?
<hurricos> except of course the first one.
<hurricos> I'm worrying that the WS-AP3825i having its LEDs driven by a shift register means the model is completely wrong and I'm doing the device tree the wrong way
<hurricos> you can't call the outputs of a shift register a GPIO
<hurricos> so why can you register LEDs against them?
<hurricos> I am missing a latch bit
<hurricos> I think
Borromini has joined #openwrt-devel
clayface has quit [Ping timeout: 480 seconds]
<hurricos> the LV164A doesn't have a latch bit input :sweating: https://www.ti.com/lit/ds/symlink/sn74lv164a.pdf
<svanheule> hurricos: it depends on how you use the outputs on a shift register, but they are pretty generic...
<svanheule> hurricos: normally the shift happens fast enough, so you don't notice the new values racing across the LEDs
<hurricos> eugh OK
<hurricos> So that's to say
<hurricos> there's kernel code somewhere for driving input to /sys/class/leds into the right register
clayface has joined #openwrt-devel
<hurricos> this seems to tell me that I need to either expect the kernel knows about how my 8-bit shift register works
<hurricos> or tell it the dimensions of the register
<svanheule> that would be ngpios, I think
<svanheule> although for this one, I think it assumes there are 8 outputs; binding doesn't specify ngpios
<hurricos> So I already have something like this set up: https://www.kernel.org/doc/Documentation/devicetree/bindings/spi/spi-gpio.txt
<hurricos> I had the 74hc595 as a client within this node
<svanheule> ok, good
<hurricos> my confusion then is .....
<hurricos> I guess, the 74hc595-compatible should show up as a separate GPIO, yes?
<hurricos> Under /sys/class/gpios
<hurricos> err
<hurricos> /sys/class/gpio
<svanheule> as a gpiochip, yes
<hurricos> OK cool
<hurricos> I'll fiddle around until I get that
<hurricos> thank you
<hurricos> in this case I think registers-number is ... oh
<hurricos> I see! I was assuming 'daisy-chained' meant different shift register chips
<svanheule> it does, but you use the output of the first one to drive the input of the second one, etc.
<hurricos> but now that I have realized this fundamental problem this is definitely referring to the 8
<hurricos> err I mean different discrete chips
<hurricos> but yes, right, there are 8 registers
<hurricos> literally :P
<svanheule> or rather one 8-bit register
<svanheule> unless you really have eight chips, for a total of 64 outputs, but then registers-number would be <8>
<hurricos> ah
<hurricos> OK
<svanheule> each chip has 8 flip-flops, to build one register
<hurricos> And fwiw the documentation is pretty clear about this being a generic for 8-bit shift registers.
<svanheule> yes, should be able to handle anything that has CLK and DATA, and no output latching
<svanheule> and if it isn't exactly 8-bit in size, you could just round up to the next multiple of 8 to get the equivalent number of registers.
<hurricos> OK, so from the last non-working state the only thing I've added was a reasonable spi-max-frequency to the parent gpio bit-banger
<hurricos> wow
<hurricos> Sorry, another mistake. I left registers-number at 8.
fda has quit [Quit: ZNC - https://znc.in]
<svanheule> :P
<svanheule> I also don't know if it defines the highest gpio to be the first one shifted out, or the last one
<hurricos> well, I can always reverse
<hurricos> the next qualifying step is just getting it to show a new node in /sys/class/gpio
<hurricos> mosi vs miso ... I suspect I am pointing at mosi, because the bus driver is my GPIO?
<hurricos> bus master*
<hurricos> I guess SPI is bi-directional
<hurricos> and that's why this example has both?
<svanheule> mosi is Master In - Slave Out; miso is the opposite
<hurricos> err
<hurricos> m i s o is mosi?
<svanheule> right, sorry :P
<hurricos> OK
<svanheule> those are the words, I'll leave assembling them into an acronym as an exercise to the reader :^)
<hurricos> So by giving spi.mosi-gpios = <&gpio0 $input_pin 0>, knowing 0 is low, I'm doing the right thing, I think
<hurricos> for spi.compatible="spi-gpio", I should say
<hurricos> though the device has no other spi
<svanheule> if 0 means "active high" (would make sense, to have that as default)
<hurricos> in reflection that the state of the $input_pin GPIO is the state that gets carried to the next shift register bit at every cycle of spi.sck-gpio <&gpio0 $clock_pin 0>
<hurricos> alright, let's just see what happens when I compile and boot
<hurricos> dmesg not giving me anything when it comes to debugging isn't helping me much :( I might have to start inserting printk to the spi-gpio.c device driver lmao
goliath has joined #openwrt-devel
<svanheule> do you see the spi-master in sysfs?
<hurricos> oh
<hurricos> let me check in the bad currently-booted version
<hurricos> err, ./class/spi_master/spi0 ?
<hurricos> fwiw I also have spi-nor on the board
<hurricos> so not sure if that's from that
<hurricos> I can boot a pre-LED-attempt version to find out
<hurricos> oh no
<hurricos> this isn't that
<hurricos> this is the spi device at address 0x7000
<hurricos> so I should expect to see an spi_master e.g. "led_spi" (maybe)
<svanheule> not sure what it will be called
<hurricos> I do have /sys/firmware/devicetree/base/led_spi
<svanheule> do you have the the driver for spi-gpio enabled?
<hurricos> oh
<hurricos> you know what
<hurricos> god damn it
<hurricos> I knew it
<hurricos> it wouldn't be a default / builtin
<hurricos> I was checking the ath79 build tree
<hurricos> fwiw
<hurricos> for other devices that bit-bang
<hurricos> for other devices with shift registers*
<hurricos> the one thing I didn't realize was the possibility that either 1) they didn't use bit banging or 2) they have spi-gpio on by default
<hurricos> s/on/enabled/
<svanheule> I would think it's pretty common, but wouldn't hurt to check to make sure you have all the required pieces :)
<hurricos> yeah, I can also check System.map I think
<hurricos> $ cat nohup.out | grep -i spi.*gpio | wc -l
<hurricos> 0
<svanheule> hurricos: target/linux/ath79/config-5.10:CONFIG_SPI_GPIO=y
<hurricos> (nohup.out is detailed build. $(CC).*spi-gpio.c never happened)
<svanheule> should be there, on ath79
<hurricos> ahhhhhhHH RATS
<hurricos> not on mpc85xx I bet
<svanheule> nope :P
<hurricos> I can't just add a depends for one of the symbols can I
<hurricos> oh wait
<hurricos> I could
<hurricos> e.g. for board-specific symbols that mpc85xx gets
<hurricos> because mpc85xx still has a platform file per board
<hurricos> if I added a dependency of CONFIG_WS_AP3825I on CONFIG_SPI_GPIO it'd be fine
<svanheule> there should be de spi-gpio kmod package too
<hurricos> because when building the p1020 kernel the build system selects all targtes
<svanheule> s/de/the
<hurricos> yeah, but I don't have BUILD_ALL_KMODS enabled :P
<hurricos> but if I just add CONFIG_SPI_GPIO in the config-5.10 I risk someone not porting it forward
<hurricos> let me see if adding it to the depends works OK
<hurricos> s/depends/selects/
zenox has quit [Remote host closed the connection]
<svanheule> using it as a kmod does mean you will only be able to use it after it starts loading rootfs
<svanheule> so you wouldn't have failsafe indication, for example
floof58_ has joined #openwrt-devel
<hurricos> right, so I'll just enable it for all p1020 by forcing a select from the CONFIG_WS_AP3825I symbol. Assuming that compiles. Let's find out
<hurricos> config-$kver is a diffconfig, I'm assuming here
<svanheule> yes, there's also the default config
<hurricos> if it's not and it's desirable to show every symbol that WOULD BE selected, this'll screw that up
<hurricos> yeah, all good
floof58 has quit [Ping timeout: 480 seconds]
floof58_ has quit []
floof58 has joined #openwrt-devel
<hurricos> CONFIG_SPI_GPIO gets selected :D
<hurricos> just compiling the initramfs now ....
<hurricos> still nothing under /sys/class/spi_master
<Pepes> Guys, would it be possible to merge this pull request: https://github.com/openwrt/openwrt/pull/4829 ? It is sitting there for almost 20 days. :(
<svanheule> hurricos: :(
<blocktrron1> nick[m]1234: did you apply it with the patch I've sent you prior?
<nick[m]1234> blocktrron1: no. just this
<nick[m]1234> okay I will add it again
<nick[m]1234> ;)
<blocktrron1> nick[m]1234: thanks
<Borromini> Grommish_: ping
<Grommish_> Borromini: pong
<Borromini> which apparently fixes a small memleak but probably unrelated to the ethernet driver
<Grommish_> Borromini: Oo.. So that actually puts frees in place?
<Borromini> i haven't got the slightest idea unfortunately :P
<Borromini> but i'll throw it in my tree here to test drive on my EdgeRouter Lite
<Grommish_> :D Well, I should try it at some point.. I'm still banging head on rust and llvm :/
dangole has joined #openwrt-devel
<Borromini> :)
<Grommish_> Borromini: I can't get LLVM to generate a native Mips64 library. It's what holding up native support on-device :(
<nick[m]1234> blocktrron1: regardless of the flash issue, what do you think of putting ubnt-xm to tiny because of low ressources?
<dangole> ping stintel
<blocktrron1> nick[m]1234: it has sufficient flash, so i see no point in doing so
<nick[m]1234> blocktrron1: it is more about the 32MB RAM, that has also issues with luci and uhttpd
<blocktrron1> i don't like it
<blocktrron1> As then we would have to move all 32M devices
<blocktrron1> and lets face it - you'll need to do a custom build nowadays anyway
<nick[m]1234> okay. ;)
<nick[m]1234> I can close the PR again, but I would be happy if this finds its way into openwrt https://github.com/openwrt/openwrt/pull/4879/commits/7f6b0fdf7fc708a521db12ba9e5494c0cecf5e8f
<nick[m]1234> (just compiling nanostation xm image with patches ... needs some time until I can report new dmesg)
<blocktrron1> nick[m]1234: open a seperate PR
<blocktrron1> or ping me when I'm on a laptop again
<nick[m]1234> I don't see anything useful. :S
<blocktrron1> Do you have your tree online?
<blocktrron1> On GitHub?
<nick[m]1234> whups, sry I added the wrong patch
<nick[m]1234> again...
<nick[m]1234> thats what I compile
<blocktrron1> okay
<blocktrron1> and you flash the resulting image using the bootloader?
Grommish_ is now known as Grommish
<hurricos> uhh
<hurricos> one would expect to see built-in modules in /proc/modules
<hurricos> right?
<hurricos> actually, in lsmod, too, no?
<nick[m]1234> blocktrron1: yes, but still no dmesg output https://gist.github.com/PolynomialDivision/0b5f6c956bf7f2b1cb3f2e8a90c69593
<nick[m]1234> should I flash from openwrt21.02 to this image?
<nick[m]1234> via sysupgrade
<hurricos> > in lsmod too < no, but they should be under /sys/module/. And the modules I'm looking for are not, despite CONFIG_SPI_GPIO being built
<hurricos> So, either: 1) the module has no options and not having options prevents presence in /sys/module/, or 2) I've still fucked up and not actually added the spi_gpio module
<hurricos> hahaha it was the latter =_=
<hurricos> nope, it's there
af has joined #openwrt-devel
af has quit []
<blocktrron1> nick[m]1234: your tree is omitting the patch which adds the ..._HAS_LOCK flag on the CPI chip in macronix.c
<blocktrron1> So the lock functions are never called
<blocktrron1> at that one on top
zenox has joined #openwrt-devel
Borromini has quit [Quit: leaving]
hitech95 has quit [Remote host closed the connection]
minimal has quit []
philipp64 has joined #openwrt-devel
<nick[m]1234> is SPI_NOR_4BIT_BP incorrect?
<nick[m]1234> I don't know, I don't get any debug output
<nick[m]1234> I even changed to dev_err
<nick[m]1234> but nothin
<blocktrron1> nick[m]1234: 4bitbp should not matter at this stage
<blocktrron1> Try to determine, where the code does not follow the locking path
<blocktrron1> e.g. check if spi_nor_unlock_all, spi_nor_unlock and spi_nor_sr_unlock get called at all
<blocktrron1> Somewhere it seems it does not finish this call tree
<blocktrron1> warn laglevel is the minimum compiled & printed by default afair
<nick[m]1234> I will switch to printk ;)
<dangole> blocktron1: i got XMDIO working on the unifi 6 lr but I only got 1000M PoE switch. so if you feel like playing with this AQR112C on 2500Base-T link (which will probably need more work)...
<dangole> * blocktrron1
<blocktrron1> dangole: do you have a tree somewhere?
<blocktrron1> And are you in posession of the required firmware for the PHY?
<dangole> blocktrron1: just pushed to master
<blocktrron1> I can have a look
<blocktrron1> AFAIR there was an SPI chip next to the PHY which might contain the firmware
<blocktrron1> I don't have the device at my place ATM (and do not have a 2.5 NBase-T card)
<dangole> blocktrron1: aq-fw-download works with AQR-G4_v5.4.B-AQR_CIG_WIFI_ID44715_VER1673.cld, i use kmod-mdio-netlink and mdio-tools to trigger the final blessing after firmware upload
eduardo010174 has joined #openwrt-devel
eduardo010174 has quit []
<dangole> blocktrron1: ah interesting. i used a generic binary the winds had whispered.
valku has joined #openwrt-devel
<dangole> blocktrron1: i got an 10GE/5GE/2.5GE/1GE/100M/10M intel nic, but no capable PoE injector, only 1000M PoE switch...
<blocktrron1> PoE injector should not matter all to much someone told me
<blocktrron1> I have no idea about that works, so i just repeat what i was told
<blocktrron1> Just getting a generic microsemi one from ebay might be enough
<blocktrron1> i cant give you an ETA on when I'll have access to the U6-LR again, as its 120 km northwards from me currently.
<blocktrron1> But I''ll come back for the firmware
<nick[m]1234> blocktrron1: spi_nor_unlock and spi_nor_sr_unlock are definitelycalled
<nick[m]1234> spi_nor_try_unlock_all is not called
<blocktrron1> nick[m]1234 you mean "spi_nor_unlock_all" ?
<nick[m]1234> no spi_nor_try_unlock_all, the other function does not exist?
<blocktrron1> there is no try_unlock_all in 5.10
<nick[m]1234> I don't know if something got backported? Where can I find the 5.10 branch from the linux kernel?
<blocktrron1> give me a sec
<nick[m]1234> openwrt is at 5.10.87
<blocktrron1> I've notices i'm on the v5.10 tag, not stable branch
<blocktrron1> ahh
<blocktrron1> This looks much better
<rsalvaterra> dangole: 2.5 GbE on the 6 LR, at last? Now there's a nice Christmas present. :)
<blocktrron1> nick[m]1234: do you have a branch with your patches to determine what is called and what not?
<blocktrron1> spi_nor_try_unlock_all should always be called
<nick[m]1234> I guess I could try something like this: https://lore.kernel.org/all/20201203162959.29589-7-michael@walle.cc/
<nick[m]1234> blocktrron1: give me 1 sec, I will rework that printk
<blocktrron1> nick[m]1234: yes, something like this would be the "ultimate workaround"
<blocktrron1> but hacking to clear the 4 bits regardless of the requested size should at least show some effect
<nick[m]1234> blocktrron1: lol the default is spi_nor_write_16bit_sr_and_check
<nick[m]1234> that is why your debug output was never called xD
<nick[m]1234> lets disable that
rmilecki has quit [Quit: Konversation terminated!]
rmilecki has joined #openwrt-devel
<nick[m]1234> I think disabling the 16bit fixed it https://gist.github.com/PolynomialDivision/8dd8d738c251e3f91c67d68c53d22ccf
<nick[m]1234> I will write a fixup that disables 16 bit. I hope this fixes it completely
<hurricos> ok, this is really weird
<hurricos> openwrt claims I have kmod-spi-gpio installed
<hurricos> and yet I am unable to probe the module
<hurricos> and the kmod-spi-gpio .ipk file only contains the kmodloader
<hurricos> the /etc/modules.d/ entry for it
<hurricos> what have I fucked up? o_e
<hurricos> is kmod-spi-gpio not a real module?
<hurricos> I just installed it on another machine
<hurricos> it's real
<hurricos> so somehow openwrt is generating an empty package for me (?????)
rmilecki has quit [Ping timeout: 480 seconds]
<hurricos> maybe OpenWrt thinks the module is installed in my kernel