Re: [pfSense-discussion] Intel 82559ER switch
Are details on the Broadcom switch controllers openly available? I did not find anything on their web site. Nor did I, but the header file indicates it came from Broadcom: https://svn.openwrt.org/openwrt/tags/kamikaze_7.06/package/switch/src/etc53xx.h Relevant comment: /* * Broadcom Home Gateway Reference Design * BCM53xx Register definitions * * Copyright 2004, Broadcom Corporation * All Rights Reserved. * * THIS SOFTWARE IS OFFERED "AS IS", AND BROADCOM GRANTS NO WARRANTIES OF ANY * KIND, EXPRESS OR IMPLIED, BY STATUTE, COMMUNICATION OR OTHERWISE. BROADCOM * SPECIFICALLY DISCLAIMS ANY IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS * FOR A SPECIFIC PURPOSE OR NONINFRINGEMENT CONCERNING THIS SOFTWARE. * $Id: etc53xx.h,v 1.1 2005/05/14 13:15:46 nbd Exp $ */
Re: [pfSense-discussion] Intel 82559ER switch
On Mon, 18 Jun 2007 21:32:35 -0500, RB <[EMAIL PROTECTED]> wrote: >FWIW, OpenWRT has some utils built-in to deal with these chipsets, I >just don't have it working yet - robocfg. They've deprecated it in >favor of a kernel driver, but all it does is uses a header file from >Broadcom and twiddles some ioctls on the associated ethernet port >according to the magic values therein. Are details on the Broadcom switch controllers openly available? I did not find anything on their web site.
Re: [pfSense-discussion] Intel 82559ER switch
FWIW, OpenWRT has some utils built-in to deal with these chipsets, I just don't have it working yet - robocfg. They've deprecated it in favor of a kernel driver, but all it does is uses a header file from Broadcom and twiddles some ioctls on the associated ethernet port according to the magic values therein. RB
Re: [pfSense-discussion] Intel 82559ER switch
You can probably get the IC part number for the switch by visual inspection and possibly a data sheet through Google. The controllers I am familiar with even have configurable limited VLAN support. Seems it's a Broadcom BCM5325; since it had an adhered t-wing, I was unwilling to disturb the glue. So be it. There are two 8255ERs on-board, one with it's own dedicated port, the other connected to the BCM chip. Now to pick it apart. RB
Re: [pfSense-discussion] Intel 82559ER switch
On Sun, 17 Jun 2007 11:23:29 -0500, RB <[EMAIL PROTECTED]> wrote: >> > As things stand, devices on separate ports are capable of communicating >> > with each other, but the traffic is unseen on fxp0. >> Well, this is the normal behaviour of a switch... > >Agreed - I was just more hoping it was implemented as a bridge instead >of as a simple port on the switch. The last OS on the system had >port-level control over the switch, so there's got to be some control >channel, be it out-of-band over the link, a PCI ioctl, or something. >Guess I'll install Linux on it (my comfort zone) and poke around. >Still open for suggestions, though. The 5 port embedded switch is completely separate from the 82559ER and is probably linked without ethernet level conversion. Usually there are a small number of I/O lines or a serial interface for control. You can probably get the IC part number for the switch by visual inspection and possibly a data sheet through Google. The controllers I am familiar with even have configurable limited VLAN support. The Realtek RTL8305SC is typical: http://www.realtek.com.tw/products/productsView.aspx?Langid=1&PNid=20&PFid=20&Level=5&Conn=4&ProdID=31
Re: [pfSense-discussion] Intel 82559ER switch
On 6/17/07, Nick Buraglio <[EMAIL PROTECTED]> wrote: > What was the OS that had the support that you expected? A crufty blend of only the finest proprietary software, based on VxWorks. x86 architecture. RB
Re: [pfSense-discussion] Intel 82559ER switch
/suggested/expected/ On 6/17/07, Nick Buraglio <[EMAIL PROTECTED]> wrote: What was the OS that had the support that you suggested? nb On 6/17/07, RB <[EMAIL PROTECTED]> wrote: > > > As things stand, devices on separate ports are capable of communicating with each other, but the traffic is unseen on fxp0. > > Well, this is the normal behaviour of a switch... > > Agreed - I was just more hoping it was implemented as a bridge instead > of as a simple port on the switch. The last OS on the system had > port-level control over the switch, so there's got to be some control > channel, be it out-of-band over the link, a PCI ioctl, or something. > Guess I'll install Linux on it (my comfort zone) and poke around. > Still open for suggestions, though. > > > RB >
Re: [pfSense-discussion] Intel 82559ER switch
What was the OS that had the support that you suggested? nb On 6/17/07, RB <[EMAIL PROTECTED]> wrote: > > As things stand, devices on separate ports are capable of communicating with each other, but the traffic is unseen on fxp0. > Well, this is the normal behaviour of a switch... Agreed - I was just more hoping it was implemented as a bridge instead of as a simple port on the switch. The last OS on the system had port-level control over the switch, so there's got to be some control channel, be it out-of-band over the link, a PCI ioctl, or something. Guess I'll install Linux on it (my comfort zone) and poke around. Still open for suggestions, though. RB
Re: [pfSense-discussion] Intel 82559ER switch
> As things stand, devices on separate ports are capable of communicating with each other, but the traffic is unseen on fxp0. Well, this is the normal behaviour of a switch... Agreed - I was just more hoping it was implemented as a bridge instead of as a simple port on the switch. The last OS on the system had port-level control over the switch, so there's got to be some control channel, be it out-of-band over the link, a PCI ioctl, or something. Guess I'll install Linux on it (my comfort zone) and poke around. Still open for suggestions, though. RB
Re: [pfSense-discussion] Intel 82559ER switch
RB wrote: it sees the entire switch as a single interface (fxp0, in this case). This is correct as you are connected to a switch port (the internal). As things stand, devices on separate ports are capable of communicating with each other, but the traffic is unseen on fxp0. Well, this is the normal behaviour of a switch... If you would see the traffic between devices it would be a hub. I don't know in detail the Intel chip you are using but I guess you will not be able to control anything on it. Is like connecting a "dumb" external 5 ports switch. regards - Daniele Guazzoni Senior Network Engineer, CCNP, CCNA Linux and AMD-x86_64 or do you still with Windows and Intel ? -- This message has been scanned for viruses and dangerous content by MailGate, and is believed to be clean.
[pfSense-discussion] Intel 82559ER switch
Anyone have any experience with a 4-port switch using the Intel 82559ER chipset? I've got 1.2-BETA-1-embedded up and running on a piece of hardware with one of these embedded on the motherboard, and it sees the entire switch as a single interface (fxp0, in this case). Related dmesg follows: fxp0: port 0xee00-0xee3f mem 0xc010-0xc0100fff,0xc012-0xc013 irq 9 at device 0.0 on pci1 miibus0: on fxp0 fxp0: Ethernet address: xx:xx:xx:fe:09:98 fxp0: link state changed to UP fxp0: device timeout It's also noteworthy that to get this and it's sister single-port (fxp1, dmesg below) to work, I have to disable tcp-offload. During boot, fxp1 comes up with .. as it's MAC, but then shows up as 02:ea:fa:34:36:2e fxp1: port 0xea00-0xea3f mem 0xc014-0xc0140fff,0xc016-0xc017 irq 12 at device 1.0 on pci1 fxp1: Disabling dynamic standby mode in EEPROM fxp1: New EEPROM ID: 0xfffd fxp1: EEPROM checksum @ 0xff: 0x -> 0xbbb9 miibus1: on fxp1 fxp1: Ethernet address: ff:ff:ff:ff:ff:ff fxp1: link state changed to UP I'm mostly interested in controlling the individual ports, but don't know how possible that is. As things stand, devices on separate ports are capable of communicating with each other, but the traffic is unseen on fxp0. Any priors? fxp0fxp1 | | | | | | | | | PHY PHY PHY PHY PHY RB