<mrnuke_>
russell--: do you know how to modify and build realtek-poe from sources?
<russell-->
russell--: yes, do you have a patch?
<mrnuke_>
s/B19200/B115200/ and see what it replies
<russell-->
okay, but first, how do you dump the decoder data in pulse view ... looking at save options, not seeing it.
<mrnuke_>
I suspect one of two: (1) old commands are implemented, and may work, or (2) old commands do nothing, and we have to implement the new protocol
<russell-->
i see export raw binary logic data, but that gives me raw trace, not the decoder afaict
<mrnuke_>
"New View" -> "Binary Decoder Output View"
<mrnuke_>
It's a little icon on the left with a little down arrow next to iy
<russell-->
aha, thank you!
<russell-->
fyi, the pin16 trick causes u-boot to occasionally nuke the first 4k block of one or more of the firmwares
<russell-->
btw, if you don't have a sensepeak pcbite, recommended!
<mrnuke_>
russell--: wait. It's wokring with 115200 baud?
<mrnuke_>
It makes no sense Some replies don't make sense. But it's working ....
<mrnuke_>
reply :20 01 61 ff ff ff ff ff ff ff ff 18 is interpreted as "97 ports" that's nonsense. Why does it work?
<russell-->
when i say it works, i mean "i plugged a meraki mr24 into port 4 and it lights up"
<russell-->
or any port
<mrnuke_>
what does ubus call poe info say?
<\x>
you guys trying to make poe control work on an openwrt'd switch?
<\x>
nice
<mrnuke_>
\x: we have it working with realtek-poe. Issue is, new version of supported switch uses a different PoE controller with similar protocol
<russell-->
s/trying to make/accidentally made/
srslypascal has quit [Remote host closed the connection]
<\x>
hah
srslypascal has joined #openwrt-devel
<mrnuke_>
SO accidental, that one of the authors of realtek-poe is shocket it's working at all
<russell-->
i can't imagine engenius wanted to change the protocol, they would have tried to hide as much of the difference as possible behind the microcontroller
<mrnuke_>
russell--: It's not engenius. It's a different PSE chip altogether
<russell-->
it's engenius's board
<mrnuke_>
It's based on a realtek design
<russell-->
it seems to be their rebranded microcontroller
<mrnuke_>
Those fucking broadcom PSEs are $30 a piece. I suspect realtek just designed their own version
<russell-->
the E == engenius, i'm guessing (of a standard stm32 part)
<mrnuke_>
when half the cost of a device is the damn PSE chips, you're doing it wrong...
* russell--
seems to recall that the license fees for the poe patents used to be around $15 a piece
<russell-->
which is partly what seemed to motivate ubiquiti to use passive poe
<mrnuke_>
russell--: if you compare the vendor dump with that, you'll start seeing that a lot of the commands are the same, just a different command ID
<russell-->
what does PSE stand for anyway, i googled and was lost in a maze of irrelevance
<mrnuke_>
Power Sourcing Equipment
<russell-->
maybe some bits are ingored?
<russell-->
ignored*
<mrnuke_>
russell--: the rtl MCU replies with the same answer to most commands from realtek-poe
<russell-->
mrnuke_: your HW v1.0.0 has a broadcom PSE?
<russell-->
it looks like it might be the LQFP48 version of the same microcontroller on v3 hardware
<mangix>
jesus m1 is fast
<mangix>
my mvebu thing can't keep up
<mrnuke_>
russell--: It's an STM I think. Protocol identifies the MCU, lemme check
<mrnuke_>
ST Micro ST32F100. Anyhow, MCU type is not as relevant as the firmware the MCU is running
<russell-->
mrnuke_: sure, but if it was me, i'd want the SoC to MCU protocol to be the same, since it can translate to whatever the PSE needs. Presumably, the vendor SoC firmware is the same between hardware versions.
<russell-->
"it" being the MCU
<russell-->
the MCU presumably can be programmed to talk serial at arbitrary speeds, so why 115200 instead of 19200? seems like gratuitous complexity.
<mrnuke_>
It's a dfferent PoE solution. The broadcom stuff predates the RTL chips by many years
<mrnuke_>
It might make sense foir things to be consistent... but when you're an engineer looking for a promotion saying you've done something better and faster is probabl bonus points
valku has quit [Remote host closed the connection]
valku has joined #openwrt-devel
Ansuel has joined #openwrt-devel
<Ansuel>
Hi guys, sorry for being silent lately but i'm on holiday and I don't have too much time to check pr and give correct review... I notice some ping so i'm apoligizing for not leaving a message :(
<stintel>
f00b4r0: nope
srslypascal has quit [Ping timeout: 480 seconds]
srslypascal has joined #openwrt-devel
srslypascal is now known as Guest7265
srslypascal has joined #openwrt-devel
Guest7265 has quit [Ping timeout: 480 seconds]
Ansuel has quit [Ping timeout: 480 seconds]
MaxSoniX has quit [Quit: Konversation terminated!]
swalker has quit [Ping timeout: 480 seconds]
srslypascal has quit [Ping timeout: 480 seconds]
srslypascal has joined #openwrt-devel
<mrnuke_>
umh, we don't support IPQ6018 by any chance, do we?
<Habbie>
a quick 'git grep' and ls target/linux smells like no
<mangix>
mrnuke_: too hard cause qualcomm
SlimeyX has quit [Read error: Connection reset by peer]