Investigate feasibility of supporting RTL8238B-based protocol

mrnuke commented 2 years ago

Introduction

The first PoE solution for RTL83xx switches was based on the BCM59111 family of chips. A later solution was implemented by Realtek, using the RTL8238B. realtek-poe speaks the dialect used on the former. IT does not appear to be compatible with the latter.

Instigate the feasibility of implementing the realtek dialect in realtek-poe.

Nomenclature

broadcom dialect - for BCM59111 family, 19200n8
realtek dialect - for RTL8238B family, 115200n8

Techical details

Frame format

Both dialects implement a 12-byte frame format, originally documented by @svanheule [1] :

[ cmd_id, seq_num, {data-[0-8]}, checksum ]

Command IDs

The realtek dialect appears to implement completely new command IDs, simply using IDs not implemented in the briadcom dialect. Some commands map between dialects, albeit with a different ID

##### Example of an equivalent command: Broadcom dialect: ``` TX -> 20 __ ff ff ff ff ff ff ff ff ff __ RX <- 20 __ 00 08 00 e1 11 16 00 03 0a __ ``` Realtek dialect: ``` TX -> 40 __ ff ff ff ff ff ff ff ff ff __ RX <- 40 __ ff 08 01 01 38 10 01 04 11 __ ``` #### Command Equivalency tables * **Setup commands:** * | ID | Number of occurences | Broadcom equivalent ID | Meaning | |----|----------------------|------------------------|---------| | 00 | 1 | 06 | global enable | 01 | 8 | 00 | set port enable | 02 | 1 | 09 | system reset | 04 | 1 | 18 | set global power budget | 08 | 8 | N/A | set auto_powerup | 0c | 8 | 1c | set port powerup mode | 10 | 1 | 17 | set power management mode | 11 | 1 | | | 12 | 8 | 15 | set power limit type | 13 | 8 | 16 | set port power budget | 15 | 8 | 1a | set port priority | 40 | 2 | 20 (?) | get system info | 47 | 1 | | | 49 | 8 | 26 | get extended port config * **Polling commands:** | ID | Command meaning | Broadcom equivalent ID | |----|-----------------|------------------------| | 41 | get power statistics (?) | 23 (consumed_power incorrect) | | 42 | get meaning of the universe and get port status | 21 | | 44 | get port measurements | 30 | | 50 | get port overview | 2a (with caveats) | Command `50` (`2a` in BCM) appears to be specific to eother EnGenius or 8-port switches. Other switches do not support it and return error 0x61. ### References [1] https://svanheule.net/switches/software/broadcom_poe_control_protocol#serial_control_protocol [2] https://svanheule.net/switches/gpl_source_drops

svanheule commented 2 years ago

Anything useful to be found in the GS1900-24EP code archive?

mrnuke commented 2 years ago

Anything useful to be found in the GS1900-24EP code archive?

Thank you! I found some headers and a kernel module with no corresponding source code:

./turnkey/plat/linux/board/poe/realtek/rtl8238b/poe_rtl8238b.h: *      poe_rtl8238b_portPriority_set
./turnkey/plat/linux/board/poe/realtek/rtl8238b/poe_rtl8238b.h:extern int32 poe_rtl8238b_portPriority_set(sys_logic_port_t port, sys_poe_priority_t pri);
grep: ./turnkey/install/board_poe.ko: binary file matches

RussellSenior commented 2 years ago

Fwiw, I requested GPL compliance from Engenius.

mrnuke commented 1 year ago

@RussellSenior, I've started a branch to test out the new dialect:

https://github.com/Hurricos/realtek-poe/commits/mrnuke-what-did-realtek-do-now

Feel free to give it a try (though don't expect it to work just yet). To test individual commands, you may use

ubus call poe sendframe '{"frame" : "99 00 00"}'

I only recommend sending commands listed in https://github.com/Hurricos/realtek-poe/issues/25#issue-1330777591. The BCM dialect has some "erase firmware" commands, and we don't know if the new one does too.

jwh commented 7 months ago

I have a bunch of GS1900 10HP v2s which appear to use this, happy to give access or test anything if we can get it working!

POE delivers power, MCU info and consumption for individual ports sometimes works but otherwise nothing works including status and enabling/disabling poe:


{
        "firmware": "v0.0",
        "budget": 77.000000,
        "consumption": 0.000000,
        "ports": {
                "lan1": {
                        "priority": 2,
                        "status": "unknown"
                },
                "lan2": {
                        "priority": 2,
                        "status": "unknown"
                },
                "lan3": {
                        "priority": 2,
                        "status": "unknown"
                },
                "lan4": {
                        "priority": 2,
                        "status": "unknown"
                },
                "lan5": {
                        "priority": 2,
                        "status": "unknown"
                },
                "lan6": {
                        "priority": 2,
                        "status": "unknown"
                },
                "lan7": {
                        "priority": 2,
                        "status": "unknown"
                },
                "lan8": {
                        "priority": 2,
                        "status": "unknown"
                }
        }
}

``

Wed Dec 27 01:47:37 2023 daemon.warn realtek-poe: Invalid port status packet (port=97)

is spammed a couple of times per second-ish to syslog (using the mrnuke-what-did-realtek-do-now branch)

mrnuke commented 3 months ago

@jwh , @RussellSenior, please see the artifacts in #35, and report with results. No need for complete debug output, just I tried on device "x": "y" and "z" worked, but "w", "a", and "b" failed in "this" manner.