EdgeRouter PoE octeon has 3 Ethernet hardware ports (it is the very same
platform for PoE and Lite). In the case of the PoE unit:
* Two first ports are connected to a PHY device (so you can connect an
actual UTP/FTP cable).
* Third port is connected to an embedded hardware switch rather than a PHY
(so you get no cable for your cnmac2).
So the OpenBSD kernel output seems reasonable as long as you suppose that
nobody has taken the job of writting the driver to enable the embedded
switch. Managing PoE is closely related (as this kind of hardware level
configuration should require its very own driver).
Regards!
2017-04-25 3:14 GMT+02:00 Doggie <dog...@privilege.pl>:
> Hello,
>
> OpenBSD has been my system of choice for router / firewall / access point
> purposes since 2003 (v3.3). And naturally it's been doing great :) Up until
> this year though, I would always use rather old i386 hardware (15-20 years
> old PC's are still in operation), equipped with a bunch of slow NIC's,
> while still dreaming about something neat, efficient, small, silent and
> low-energy. Then, all of a sudden, a few months ago I discovered Ubiquiti
> Networks' EdgeRouter PoE, followed by OpenBSD/octeon port. Didn't need much
> time to decide to buy it, remove its original USB flashdrive, deploy
> OpenBSD to a new one and finally give it a try. 6.0 Release is now
> installed, patched and configured with all the needed packages. Below I
> include dmesg outputs ("sysctl hw.sensors" produces no information):
>
> Copyright (c) 1982, 1986, 1989, 1991, 1993
> The Regents of the University of California. All rights reserved.
> Copyright (c) 1995-2016 OpenBSD. All rights reserved.
> http://www.OpenBSD.org
> OpenBSD 6.0 (GENERIC) #10: Fri Jul 29 04:45:17 UTC 2016
> visa@octeon:/usr/src/sys/arch/octeon/compile/GENERIC
> real mem = 536870912 (512MB)
> avail mem = 524386304 (500MB)
> warning: no entropy supplied by boot loader
> mainbus0 at root
> cpu0 at mainbus0: Cavium OCTEON CPU rev 0.1 500 MHz, Software FP emulation
> cpu0: cache L1-I 32KB 4 way D 8KB 64 way, L2 128KB 8 way
> clock0 at mainbus0: int 5
> iobus0 at mainbus0
> dwctwo0 at iobus0 base 0x1180068000000 irq 56
> usb0 at dwctwo0: USB revision 2.0
> uhub0 at usb0 "Octeon DWC2 root hub" rev 2.00/1.00 addr 1
> octrng0 at iobus0 base 0x1400000000000 irq 0
> cn30xxgmx0 at iobus0 base 0x1180008000000
> cnmac0 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> atphy0 at cnmac0 phy 7: AR8035 10/100/1000 PHY, rev. 2
> cnmac1 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> atphy1 at cnmac1 phy 6: AR8035 10/100/1000 PHY, rev. 2
> cnmac2 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> uartbus0 at mainbus0
> com0 at uartbus0 base 0x1180000000800 irq 34: ns16550a, 64 byte fifo
> com0: console
> com1 at uartbus0 base 0x1180000000c00 irq 35: ns16550a, 64 byte fifo
> /dev/ksyms: Symbol table not valid.
> umass0 at uhub0 port 1 configuration 1 interface 0 "JetFlash Mass Storage
> Device" rev 2.10/11.00 addr 2
> umass0: using SCSI over Bulk-Only
> scsibus0 at umass0: 2 targets, initiator 0
> sd0 at scsibus0 targ 1 lun 0: <JetFlash, Transcend 32GB, 1100> SCSI4
> 0/direct removable serial.********************
> sd0: 30128MB, 512 bytes/sector, 61702144 sectors
> vscsi0 at root
> scsibus1 at vscsi0: 256 targets
> softraid0 at root
> scsibus2 at softraid0: 256 targets
> boot device: sd0
> root on sd0a (****************.a) swap on sd0b dump on sd0b
> WARNING: No TOD clock, believing file system.
> WARNING: CHECK AND RESET THE DATE!
>
> Copyright (c) 1982, 1986, 1989, 1991, 1993
> The Regents of the University of California. All rights reserved.
> Copyright (c) 1995-2016 OpenBSD. All rights reserved.
> http://www.OpenBSD.org
> OpenBSD 6.0 (GENERIC.MP) #0: Thu Apr 6 00:45:05 CEST 2017
>
> root@************************:/usr/src/sys/arch/octeon/compile/GENERIC.MP
> real mem = 536870912 (512MB)
> avail mem = 524353536 (500MB)
> warning: no entropy supplied by boot loader
> mainbus0 at root
> cpu0 at mainbus0: Cavium OCTEON CPU rev 0.1 500 MHz, Software FP emulation
> cpu0: cache L1-I 32KB 4 way D 8KB 64 way, L2 128KB 8 way
> cpu1 at mainbus0: Cavium OCTEON CPU rev 0.1 500 MHz, Software FP emulation
> cpu1: cache L1-I 32KB 4 way D 8KB 64 way, L2 128KB 8 way
> clock0 at mainbus0: int 5
> iobus0 at mainbus0
> dwctwo0 at iobus0 base 0x1180068000000 irq 56
> usb0 at dwctwo0: USB revision 2.0
> uhub0 at usb0 "Octeon DWC2 root hub" rev 2.00/1.00 addr 1
> octrng0 at iobus0 base 0x1400000000000 irq 0
> cn30xxgmx0 at iobus0 base 0x1180008000000
> cnmac0 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> atphy0 at cnmac0 phy 7: AR8035 10/100/1000 PHY, rev. 2
> cnmac1 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> atphy1 at cnmac1 phy 6: AR8035 10/100/1000 PHY, rev. 2
> cnmac2 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> uartbus0 at mainbus0
> com0 at uartbus0 base 0x1180000000800 irq 34: ns16550a, 64 byte fifo
> com0: console
> com1 at uartbus0 base 0x1180000000c00 irq 35: ns16550a, 64 byte fifo
> /dev/ksyms: Symbol table not valid.
> umass0 at uhub0 port 1 configuration 1 interface 0 "JetFlash Mass Storage
> Device" rev 2.10/11.00 addr 2
> umass0: using SCSI over Bulk-Only
> scsibus0 at umass0: 2 targets, initiator 0
> sd0 at scsibus0 targ 1 lun 0: <JetFlash, Transcend 32GB, 1100> SCSI4
> 0/direct removable serial.********************
> sd0: 30128MB, 512 bytes/sector, 61702144 sectors
> vscsi0 at root
> scsibus1 at vscsi0: 256 targets
> softraid0 at root
> scsibus2 at softraid0: 256 targets
> boot device: sd0
> root on sd0a (****************.a) swap on sd0b dump on sd0b
> WARNING: No TOD clock, believing file system.
> WARNING: CHECK AND RESET THE DATE!
> cpu1 launched
>
> Copyright (c) 1982, 1986, 1989, 1991, 1993
> The Regents of the University of California. All rights reserved.
> Copyright (c) 1995-2017 OpenBSD. All rights reserved.
> https://www.OpenBSD.org
> OpenBSD 6.1-current (RAMDISK) #0: Mon Apr 17 09:56:55 UTC 2017
> visa@octeon:/usr/src/sys/arch/octeon/compile/RAMDISK
> real mem = 536870912 (512MB)
> avail mem = 521076736 (496MB)
> mainbus0 at root
> cpu0 at mainbus0: CN50xx CPU rev 0.1 500 MHz, Software FP emulation
> cpu0: cache L1-I 32KB 4 way D 8KB 64 way, L2 128KB 8 way
> clock0 at mainbus0: int 5
> iobus0 at mainbus0
> dwctwo0 at iobus0 base 0x1180068000000 irq 56
> usb0 at dwctwo0: USB revision 2.0
> uhub0 at usb0 configuration 1 interface 0 "Octeon DWC2 root hub" rev
> 2.00/1.00 addr 1
> cn30xxgmx0 at iobus0 base 0x1180008000000
> cnmac0 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> atphy0 at cnmac0 phy 7: AR8035 10/100/1000 PHY, rev. 2
> cnmac1 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> atphy1 at cnmac1 phy 6: AR8035 10/100/1000 PHY, rev. 2
> cnmac2 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> uartbus0 at mainbus0
> com0 at uartbus0 base 0x1180000000800 irq 34: ns16550a, 64 byte fifo
> com0: console
> com1 at uartbus0 base 0x1180000000c00 irq 35: ns16550a, 64 byte fifo
> umass0 at uhub0 port 1 configuration 1 interface 0 "JetFlash Mass Storage
> Device" rev 2.10/11.00 addr 2
> umass0: using SCSI over Bulk-Only
> scsibus0 at umass0: 2 targets, initiator 0
> sd0 at scsibus0 targ 1 lun 0: <JetFlash, Transcend 32GB, 1100> SCSI4
> 0/direct removable serial.********************
> sd0: 30128MB, 512 bytes/sector, 61702144 sectors
> root on rd0a swap on rd0b dump on rd0b
> WARNING: No TOD clock, believing file system.
> WARNING: CHECK AND RESET THE DATE!
>
> ***
>
> The only, yet quite significant, problem I have is with cnmac2 NIC. As I
> understand by checking its MAC address, this is how "switched" group of
> ports (i.e. 3rd, 4th and 5th NIC) is seen by the kernel. And here are the
> details of what's happening:
>
> - looks like cnmac2 is missing its atphy* counterpart - atphy2:
> "cnmac0 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> atphy0 at cnmac0 phy 7: AR8035 10/100/1000 PHY, rev. 2
> cnmac1 at cn30xxgmx0: RGMII, address **:**:**:**:**:**
> atphy1 at cnmac1 phy 6: AR8035 10/100/1000 PHY, rev. 2
> cnmac2 at cn30xxgmx0: RGMII, address **:**:**:**:**:**"
>
> - "ifconfig cnmac2 media" shows rather strange output:
> "cnmac2: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
> lladdr **:**:**:**:**:**
> index 3 priority 0 llprio 3
> media: Ethernet manual (none)
> supported media:
> media manual"
>
> - despite setting IP address and the rest for cnmac2, no external
> connections are working (both inbound and outbound) - I can ping this
> address only from the router itself
>
> According to Ubiquiti documentation, 3rd, 4th and 5th NIC's can either be
> turned into a switch, or work separately. And having 5 independent network
> interfaces of course would be something cool! However, for now I cannot use
> these ports either way.
>
> I did test them using the original operating system and they work fine.
> Also, played with turning them into a switch and back - to check if that
> would somehow affect how bsd.rd sees them after reboot - but it made no
> difference.
>
> ***
>
> And here is a couple of suggestions after my thorough reading of
> "INSTALL.octeon" doc and experiments performed afterwards:
>
> - replacing all occurrences of "rootdev=sd0" with "rootdev=/dev/sd0",
> which fixes the issue with boot process stopping in the middle and asking
> for "root device:"
>
> - replacing all occurrences of "$loadaddr" with "${loadaddr}" - just a
> cosmetic change
>
> - putting an information that a kernel meant to be used, should actually
> be copied to MSDOS partition, so that U-Boot can load it upon boot, e.g. by
> doing this:
> mount /dev/sd0i /mnt && cp /bsd /mnt/bsd && umount /mnt
>
> ***
>
> Currently I use my new Ubiquiti access point with its original PoE
> Injector. But certainly it would be great to be able to turn on 24V / 48V
> on chosen EdgeRouter network ports - like in the original operating system.
> Although maybe this part of router software is closed source, so it may not
> be that easy to implement in OpenBSD.
>
> ***
>
> Another subject - crossing my fingers for successful implementation of
> single user mode option when system starts. Unless it's already possible
> somehow.
>
> ***
>
> That's pretty much it about my current experience with OpenBSD/octeon on
> EdgeRouter PoE. Let me know how I can help further.
>
> And last but not least:
> Many thanks to OpenBSD developers and other folks involved - for all the
> good work across all these years. And best wishes on the 20th anniversary!
>
> --
> Cheers,
> Pawel Waga
>
>
