PD690xx PoE daemon - Githubissues

halmartin commented 4 years ago

The init script S03gpio currently auto-configures the PD690xx for PoE. This allows for "plug-and-play" PoE functionality on the switch (at least on the MS220-8P; I don't have other Meraki PoE enabled hardware to test).

However, more advanced functionality which is present in the original Meraki firmware is not possible, this includes:

querying the PoE power budget
disabling PoE on a port (e.g. to power cycle a PoE device)
configuring 802.3at/af on a port

And probably more things I don't know about since I don't have much experience with PoE.

Resources:

halmartin commented 4 years ago

Here is a super basic program to manage PoE. Tested on an MS220-8P.

You can disable PoE on a port
You can enable PoE on a port
You can query for the total power delivered via PoE

I welcome PRs by people who aren't terrible at C code. This is some A+ copy pasta.

halmartin commented 4 years ago

Okay, there are more features now:

Get the PoE power per port (-p PORT)
Get the temperature of the PoE controller with -t (should be correct, the formula differs depending on which page of the datasheet you consult)

Updated the README and now it prints some pretty usage message if you don't specify any arguments (or -h) 🥳

halmartin commented 4 years ago

More new features!

Summary of all PoE port statuses:

$ pd690xx -l
Port    Status          Type    Priority        Power
1       Disabled        at      Critical        0.0
2       Enabled         af      Critical        1.8
3       Enabled         at      Critical        0.0
4       Disabled        at      Critical        0.0
5       Forced          at      Critical        15.4
6       Enabled         at      Critical        0.0
7       Disabled        at      Critical        0.0
8       Enabled         at      Critical        0.0
9       Enabled         at      Critical        0.0
10      Enabled         at      Critical        0.0
11      Enabled         at      Critical        0.0
12      Enabled         at      Critical        0.0

There is no way (that I can see) to query for how many ports exist, so an MS220-8P will display 12 ports (despite having only 8 PoE capable ports) because that's what the PD69012 supports, and apparently the PD69008 has the same port registers ¯_(ツ)_/¯

You can also "force" a port on, though I don't know what that means in practice (maybe don't test this with equipment you care about).

Now you can also query the pd690xx for the PoE voltage within the switch:

$ pd690xx -s
56.2 V

halmartin commented 4 years ago

The only items I consider still to-do are:

Change the port priority so lower priority ports are shed if the PoE power budget is exceeded
Figure out if the MS220-24/48P have different I2C addresses. I suspect the -24P uses two pd690xx on the same I2C bus, and the -48LP/FP use multiple chips on each I2C bus (/dev/i2c-1 and /dev/i2c-2), but lack hardware to confirm this
Implement some checking if /dev/i2c-2 exists and transparently support up to 24 ports if it does
Allow the user to configure a port as 802.3af instead of 802.3at (although I'm not positive there's a use case here)

ben-oxley commented 4 years ago

I've got a MS220-48LP if I can do anything to help confirm your second bullet point above. I've taken a binary dump from the device now which I can share if it helps?

halmartin commented 4 years ago

@ben-oxley thank you for the offer! The Meraki firmware is common to all the MS22/42/220/320 series. Current builds of postmerkOS are also common to all models.

I have decompiled libpoecore from the Meraki firmware and used some information from that in my design of pd690xx

What I am still missing is the I2C addresses of the PoE controllers that manage ports 13-48 (each controller manages 12 ports; there is one controller in the MS220-8P, two in the -24, and four in the -48 models). I would greatly appreciate some assistance from someone with 24 or 48 port MS220/MS320 switch, however you'll need to flash postmerkOS first.

Once you have postmerkOS flashed and booting, you can enable I2C tracing in the kernel and use poe_cli from the Meraki firmware to query ports 23, 35, and 47 (for example). The I2C trace output will provide the I2C address of the remaining controllers and then I can add support for these and pd690xx will be usable on larger switch models too.

ben-oxley commented 4 years ago

We're up and running on postmerkOS (sort of). Seems to have some issues on boot.

Bootlog below:

LinuxLoader built Nov 12 2020 18:01:50
init_pll ok
init_spi ok
init_memctl ok
wait_memctl ok
Training DRAM ok
init_irq ok
init_dram_uncached ok
init_icache ok
init_dcache ok
enable_caches ok
init_pi ok
init_board ok
Low level initialization complete, exiting boot mode
[    0.000000] Linux version 3.18.123-meraki-elemental (hmartin@alp) (gcc version 5.4.0 (GCC) ) #10 Sat Jul 18 17:06:37 UTC 2020
[    0.000000] CPU0 revision is: 01019654 (MIPS 24KEc)
[    0.000000] Determined physical RAM map:
[    0.000000]  memory: 00477000 @ 00100000 (usable)
[    0.000000]  memory: 00049000 @ 00577000 (usable after init)
[    0.000000] User-defined physical RAM map:
[    0.000000]  memory: 07ff0000 @ 00000000 (usable)
[    0.000000] Initmem setup node 0 [mem 0x00000000-0x07feffff]
[    0.000000] Reserving 0MB of memory at 0MB for crashkernel
[    0.000000] Kernel command line: console=ttyS0,115200 mtdparts=m25p80:0x40000(redboot),0x2c0000(kernel),0x800000(squashfs),0x500000(overlay) root=/dev/mtdblock3 ubi.mtd=gen_nand.0 mem=134152192
[    0.000000] PID hash table entries: 512 (order: -1, 2048 bytes)
[    0.000000] Dentry cache hash table entries: 16384 (order: 4, 65536 bytes)
[    0.000000] Inode-cache hash table entries: 8192 (order: 3, 32768 bytes)
[    0.000000] Writing ErrCtl register=8000a004
[    0.000000] Readback ErrCtl register=8000a004
[    0.000000] Cache parity protection enabled
[    0.000000] SLUB: HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
[    0.000000] NR_IRQS:66
[    0.000000] sched_clock: 32 bits at 1kHz, resolution 1000000ns, wraps every 2147483648000000ns
[    0.002000] VCore-III slave device ID 7460 found
[    0.002000] Calibrating delay loop... 275.45 BogoMIPS (lpj=137728)
[    0.011000] pid_max: default: 32768 minimum: 301
[    0.011000] Mount-cache hash table entries: 1024 (order: 0, 4096 bytes)
[    0.011000] Mountpoint-cache hash table entries: 1024 (order: 0, 4096 bytes)
[    0.013000] ftrace: allocating 11994 entries in 24 pages
[    0.041000] Performance counters: mips/24K PMU enabled, 2 32-bit counters available to each CPU, irq -1 (share with timer interrupt)
[    0.048000] devtmpfs: initialized
[    0.055000] NET: Registered protocol family 16
[    0.094000] Switched to clocksource MIPS
[    0.124000] NET: Registered protocol family 2
[    0.127000] TCP established hash table entries: 1024 (order: 0, 4096 bytes)
[    0.128000] TCP bind hash table entries: 1024 (order: 0, 4096 bytes)
[    0.128000] TCP: Hash tables configured (established 1024 bind 1024)
[    0.128000] TCP: reno registered
[    0.128000] UDP hash table entries: 256 (order: 0, 4096 bytes)
[    0.128000] UDP-Lite hash table entries: 256 (order: 0, 4096 bytes)
[    0.129000] NET: Registered protocol family 1
[    0.133000] VCORE-III Watchdog Timer enabled (30 seconds).  Prev boot was not caused by WDT reset.
[    0.158000] futex hash table entries: 256 (order: -1, 3072 bytes)
[    0.196000] squashfs: version 4.0 (2009/01/31) Phillip Lougher
[    0.196000] jffs2: version 2.2. (NAND) © 2001-2006 Red Hat, Inc.
[    0.199000] msgmni has been set to 241
[    0.21=000] io scheduler noop registered
[    0.219000] io scheduler deadline registered (default)
[    0.221000] Serial: 8250/16550 driver, 1 ports, IRQ sharing disabled
[    0.223000] console [ttyS0] disabled
[    0.224000] serial8250.0: ttyS0 at MMIO 0x70100000 (irq = 14, base_baud = 13020833) is a 16550A
[    0.609000] console [ttyS0] enabled
[    0.619000] nand: device found, Manufacturer ID: 0x2c, Chip ID: 0xf1
[    0.625000] nand: Micron MT29F1G08ABADAWP
[    0.630000] nand: 128MiB, SLC, page size: 2048, OOB size: 64
[    0.641000] Scanning device for bad blocks
[    0.758000] m25p80 spi0.1: found mx25l12805d, expected m25p80
[    0.764000] m25p80 spi0.1: mx25l12805d (16384 Kbytes)
[    0.769000] 4 cmdlinepart partitions found on MTD device m25p80
[    0.775000] Creating 4 MTD partitions on "m25p80":
[    0.780000] 0x000000000000-0x000000040000 : "redboot"
[    0.802000] 0x000000040000-0x000000300000 : "kernel"
[    0.812000] 0x000000300000-0x000000b00000 : "squashfs"
[    0.827000] 0x000000b00000-0x000001000000 : "overlay"
[    0.841000] tun: Universal TUN/TAP device driver, 1.6
[    0.846000] tun: (C) 1999-2004 Max Krasnyansky <maxk@qualcomm.com>
[    0.856000] i2c /dev entries driver
[    0.863000] TCP: cubic registered
[    0.867000] Initializing XFRM netlink socket
[    0.874000] NET: Registered protocol family 10
[    0.883000] NET: Registered protocol family 17
[    0.888000] NET: Registered protocol family 15
[    0.893000] 8021q: 802.1Q VLAN Support v1.8
[    0.903000] i2c-gpio i2c-gpio.1: using pins 15 (SDA) and 14 (SCL)
[    0.910000] at24 1-0051: 1024 byte 24c08 EEPROM, writable, 8 bytes/write
[    0.925000] meraki-config meraki-config: Meraki config device loaded
[    1.016000] i2c-gpio-scl i2c-gpio-scl.2: using pins 53 (SDA) and 52 (SCL)
[    1.024000] i2c i2c-2: Added multiplexed i2c bus 3
[    1.031000] i2c i2c-2: Added multiplexed i2c bus 4
[    1.037000] i2c i2c-2: Added multiplexed i2c bus 5
[    1.045000] i2c i2c-2: Added multiplexed i2c bus 6
[    1.062000] tmp401 8-004c: Detected TI TMP411 chip
[    1.082000] i2c-gpio i2c-gpio.8: using pins 47 (SDA) and 46 (SCL)
[    1.090000] at24 3-0050: 256 byte 24c02 EEPROM, read-only, 0 bytes/write
[    1.097000] at24 4-0050: 256 byte 24c02 EEPROM, read-only, 0 bytes/write
[    1.104000] at24 5-0050: 256 byte 24c02 EEPROM, read-only, 0 bytes/write
[    1.112000] at24 6-0050: 256 byte 24c02 EEPROM, read-only, 0 bytes/write
[    1.129000] UBI: attaching mtd0 to ubi0
[    1.933000] UBI: scanning is finished
[    1.975000] UBI: attached mtd0 (name "gen_nand.0", size 128 MiB) to ubi0
[    1.982000] UBI: PEB size: 131072 bytes (128 KiB), LEB size: 129024 bytes
[    1.989000] UBI: min./max. I/O unit sizes: 2048/2048, sub-page size 512
[    1.995000] UBI: VID header offset: 512 (aligned 512), data offset: 2048
[    2.002000] UBI: good PEBs: 1024, bad PEBs: 0, corrupted PEBs: 0
[    2.008000] UBI: user volume: 12, internal volumes: 1, max. volumes count: 128
[    2.016000] UBI: max/mean erase counter: 11/1, WL threshold: 4096, image sequence number: 1363641321
[    2.025000] UBI: available PEBs: 462, total reserved PEBs: 562, PEBs reserved for bad PEB handling: 20
[    2.034000] UBI: background thread "ubi_bgt0d" started, PID 285
[    2.158000] devtmpfs: mounted
[    2.202000] VFS: Mounted root (squashfs filesystem) readonly on device 31:3.
[    2.241000] devtmpfs: mounted
[    2.250000] Freeing unused kernel memory: 292K
[    6.244000] devpts: called with bogus options
Starting syslogd: OK
Starting klogd: OK
Running sysctl: OK
/overlay is already mounted
[    8.701000] random: nonblocking pool is initialized
board-config is /dev/mtd5
insmod: ERROR: could not load module /lib/modules//vtss_core.ko: No such file or directory
Loaded 'proclikefs.ko'
[   31.300000] click: starting router thread pid 458 (87a97e00)
Loaded 'merakiclick.ko'
[   66.332000] elts_meraki: module license 'unspecified' taints kernel.
[   66.339000] Disabling lock debugging due to kernel taint
Loaded 'elts_meraki.ko'
insmod: ERROR: could not load module /lib/modules//vc_click.ko: No such file or directory
[   69.410000] unsatisfied requirement 'vc_click'
[   69.416000] click-config/switch/io-real/vitesse.click:26: unknown element class 'ToVitesse'
[   69.425000] click-config/switch/io-real/vitesse.click:31: unknown element class 'FromVitesse'
[   69.435000] click-config/switch/io-real/vitesse.click:59: unknown element class 'VitesseController'
[   69.854000] click-config/switch.template:74: While configuring 'switch_intf_table :: SwitchIntfTable':
[   69.863000]   SWITCH_HARDWARE: element type mismatch, expected SwitchHardwareIntf
[   69.874000] click-config/switch.template:1045: While configuring 'switch_port_table :: SwitchPortTable':
[   69.884000]   SWITCH_HARDWARE_INTERFACE: element type mismatch, expected SwitchHardwareIntf
[   69.907000] click-config/switch.template:58: While configuring 'l3_update_hardware :: UpdateHardwareTable':
[   69.918000]   SWITCH_HARDWARE_INTERFACE: element type mismatch, expected SwitchHardwareIntf
[   69.936000] click-config/switch.template:383: While configuring 'switch_rpc_server :: SwitchRPCServer':
[   69.945000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   69.955000] click-config/switch.template:566: While configuring 'CheckStackMaster@187 :: CheckStackMaster':
[   69.966000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   69.974000] click-config/switch.template:602: While configuring 'CheckStackMaster@235 :: CheckStackMaster':
[   69.984000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   69.993000] click-config/switch.template:633: While configuring 'switch_mrt :: MulticastRoutingTable':
[   70.002000]   SWITCH_HARDWARE_INTERFACE: element type mismatch, expected SwitchHardwareIntf
[   70.013000] click-config/switch.template:655: While configuring 'ospf_hello_inspector :: OspfHelloInspector':
[   70.024000]   SWITCH_HARDWARE_INTERFACE: element type mismatch, expected SwitchHardwareIntf
[   70.056000] click-config/switch.template:875: While configuring 'msstp_encap_inst :: MSSTPEncap':
[   70.065000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.077000] click-config/switch.template:930: While configuring 'IntraStackResolver@527 :: IntraStackResolver':
[   70.087000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.099000] click-config/switch.template:1072: While configuring 'igmp_table :: IGMPSnoopTable':
[   70.108000]   SWITCH_HARDWARE_INTERFACE: element type mismatch, expected SwitchHardwareIntf
[   70.118000] click-config/switch.template:1093: While configuring 'mld_table :: MLDSnoopTable':
[   70.126000]   SWITCH_HARDWARE_INTERFACE: element type mismatch, expected SwitchHardwareIntf
[   70.137000] click-config/switch.template:1125: While configuring 'msstp_decap :: MSSTPDecap':
[   70.146000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.159000] click-config/switch.template:1187: While configuring 'CheckStackMaster@664 :: CheckStackMaster':
[   70.169000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.176000] click-config/switch.template:1189: While configuring 'CheckStackMaster@669 :: CheckStackMaster':
[   70.186000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.194000] click-config/switch.template:1199: While configuring 'CheckStackMaster@680 :: CheckStackMaster':
[   70.204000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.212000] click-config/switch.template:1212: While configuring 'CheckStackMaster@693 :: CheckStackMaster':
[   70.222000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.231000] click-config/switch.template:1238: While configuring 'CheckStackMaster@726 :: CheckStackMaster':
[   70.241000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.253000] click-config/switch.template:1307: While configuring 'stp_check_master :: CheckStackMaster':
[   70.263000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.270000] click-config/switch.template:1320: While configuring 'lacp_check_master :: CheckStackMaster':
[   70.280000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.289000] click-config/switch.template:1409: While configuring 'udld :: UDLD':
[   70.297000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.306000] click-config/switch.template:1459: While configuring 'CheckStackMaster@861 :: CheckStackMaster':
[   70.316000]   HW_INTF: element type mismatch, expected SwitchHardwareIntf
[   70.405000] Router could not be initialized!
Default port count: 10
net.ipv4.ip_local_reserved_ports = 50000-50127
vc_click isn't loaded; aborting!
/etc/init.d/S11leds: line 8: can't create /click/sw0_ctrl/power_led_orange: nonexistent directory
/etc/init.d/S11leds: line 9: can't create /click/sw0_ctrl/power_led_green: nonexistent directory
Generating dropbear ECDSA host key
Generating 521 bit ecdsa key, this may take a while...
Public key portion is:
ecdsa-sha2-nistp521 AAAAE2VjZHNhLXNoYTItbmlzdHA1MjEAAAAIbmlzdHA1MjEAAACFBACyoNXi8VllFxatkvlFWh6fUgwjfwU9XYpO9ieo+mt892lxCq0pe1bQu8CAa7le/ArQ6j6pncVl8iZ0p3kRzFFgiQFlLOdZBYc202LGorSDtuybFnDpsEnrRBv2QLafafMgXjADZCl7xavDvpcBjCwAwWNYnZyR4Stbvwdpoa3euvDogw== root@m881544301590
Fingerprint: sha1!! 0a:c1:43:72:66:67:ea:ef:6c:b3:41:be:f7:1a:2b:80:d6:d2:56:53
Saving random seed: SKIP (read-only file system detected)
Starting network: OK
Starting chrony: Could not open configuration file /etc/chrony.conf : No such file or directory
FAIL
Starting dropbear sshd: OK
Starting lighttpd: OK

/ #

halmartin commented 4 years ago

@ben-oxley It looks like the board detection in S08kmods doesn't quite work for the MS220-48LP. Could you please provide the output of the following?

cat /proc/cpuinfo

cat /proc/mtd

I'd also really appreciate the strings output of the board-config mtd part:

BCFG=$(cat /proc/mtd | grep board-config | awk '{print $1}' | grep -oE '[0-9]')
dd if=/dev/mtdblock${BCFG} | strings

With these details I should be able to fix the logic in S08kmods. Thanks!

ben-oxley commented 4 years ago

No problem, I'll get this to you asap, probably tomorrow now though!

halmartin commented 4 years ago

Please exclude the last line of board-config as it contains your unit serial number!

ben-oxley commented 4 years ago

/ # cat /proc/cpuinfo system type : Vitesse VCore-III VSC7434 Dual machine : Unknown processor : 0 cpu model : MIPS 24KEc V5.4 BogoMIPS : 275.45 wait instruction : yes microsecond timers : yes tlb_entries : 16 extra interrupt vector : yes hardware watchpoint : yes, count: 4, address/irw mask: [0x0ffc, 0x0ffc, 0x0ffb, 0x0ffb] isa : mips1 mips2 mips32r1 mips32r2 ASEs implemented : mips16 dsp shadow register sets : 1 kscratch registers : 0 package : 0 core : 0 VCED exceptions : not available VCEI exceptions : not available

/ # cat /proc/mtd dev: size erasesize name mtd0: 08000000 00020000 "gen_nand.0" mtd1: 00040000 00010000 "redboot" mtd2: 002c0000 00010000 "kernel" mtd3: 00800000 00010000 "squashfs" mtd4: 00500000 00010000 "overlay" �td5: 0001f800 0001f800 "board-config" mtd6: 00086000 0001f800 "bootroot" mtd7: 0140e800 0001f800 "part1" mtd8: 0140e800 0001f800 "part2" mtd9: 0081f000 0001f800 "storage" mtd10: 0020bdb0 0001f800 "SMBStaX-24" mtd11: 00292240 0001f800 "SMBStaX-48" mtd12: 00230af0 0001f800 "SMBStaX-MS220-8" mtd13: 0022be00 0001f800 "SMBStaX-MS220-24" mtd14: 0029bc88 0001f800 "SMBStaX-MS220-48" mtd15: 0029e7c0 0001f800 "SMBStaX-MS320-24" mtd16: 0029ca88 0001f800 "SMBStaX-MS320-48" / # BCFG=$(cat /proc/mtd | grep board-config | awk '{print $1}' | grep -oE '[0-9 ]') / # echo ${BCFG} 5 / # dd if=/dev/mtdblock${BCFG} | strings 5311 meraki_elemental 000000000 Q2XXXXXXXXXV 252+0 records in 252+0 records out

ben-oxley commented 4 years ago

Failed to initialize PoE server. a=030 a=031 a=033 a=035

Is this the magic you're looking for?

halmartin commented 4 years ago

Thanks to @ben-oxley for help in debugging the MS220-48P and testing, pd690xx has been updated to work with larger switch models such as the MS220-48.

mterron commented 4 years ago

Hi @halmartin there's a whole poe driver infrastructure in the Meraki dump at linux-3.18/drivers/poe/. It's probably worth it to take a look as it is the official Meraki PoE controller infrastructure.

halmartin commented 4 years ago

@mterron Oh man, I should really grep for strings before I write user-space daemons.

Nice to see I wasn't the only person confounded by the poorly written temperature formula in the datasheet.

For some reason, Meraki does not enable POE_PD690XX in the defconfig they supply in their GPL archive. The string pd690 also doesn't appear in the kernel used by their firmware I extracted from my MS220-8P.

I'm not sure why they would go to the effort of writing this driver and then not use it.

mterron commented 4 years ago

Can't comment on that, sorry. It is nicely integrated with thermal monitoring and other goodies so it doesn't look like a POC.

$ grep -ri meraki switch-11-22-ms220/linux-3.18 points to lot's of useful information though.

halmartin commented 2 years ago

There was a serious bug in pd690xx that affects the MS220-24P and MS220-48LP/FP: #27

Users should update their switch firmware to 20221115 to resolve the issue.

halmartin commented 11 months ago

Closing as the pd690xx utility is usable and contains no major bugs.

halmartin / meraki-builder

PD690xx PoE daemon #1