🚀 go-pugleaf

Hey all,

Somehow Deja lost this thread, so I'm posting the earlier stuff...

Yes, my end goal is to use this as a gateway machine running IP
Masquerading.

Anyway, I have Linux recognizing the eth0, eth1, and eth2 (yes, I have
two additional cards in), and I've assigned each interface an IP
address, but the only interface that appears to work is the eth0
(motherboard) interface. Why is this? I also tried manually assigning
MAC addresses using

ifconfig eth1 hw ether 08:00:20:0D:9B:AB

(the machine's MAC address is 08:00:20:0D:9B:AA)

among other things. I also "setenv local-mac-address? true" in
the OBP. I don't actually know whether the MAC address of the interface
itself really matters in the end...

Nothing seems to get the non-motherboard Ethernet interfaces to do
anything.

What am I doing wrong? Is there some special trick that I'm still
missing?

In article <2NQc3.28674$eF3.11048@news.rdc1.nj.home.com>,
  Mike Sawicki <fifi@prozac.lojik.net> wrote:
> In comp.sys.sun.hardware max_newell@my-deja.com wrote:
> > All I get from "test net" is the following:
>
> >  Internal loopback test -- succeeded.
> >  External loopback test -- succeeded.
>
> > (If I disconnect the cable from the onboard Ethernet, the external
test
> > fails.)
>
> > I assume that I would be getting two tests if I properly had two
> > Ethernet interfaces installed.
>
> > Here's what I see if I do "cd sbus@1,f8000000" (the sbus device that
> > shows up at /)
>
> > ffec1844 cgsix@3,0
> > ffec17a4 le@2,c00000
> > ffec053c le@0,c00000
> > ffebdddc esp@0,800000
> > ffebdb28 dma@0,400000
>
> Good, you have both le devices.
>
> > Another strange detail: When I installed Linux on this machine,
dmesg
> > reported both eth0 and eth1, both with the same MAC address. ?!?!?
>
> Sparc machines share one hardware address per machine.  Since each
ethernet
> device is assumed to be used on a different physical segment, cloned
MAC
> addresses will not affect operation.  These machines store their MAC
info
> in the NVRAM on the machine.  It can be changed, if need be, but I
don't
> know if there is a way to force them to use separate addresses.  Is
this
> really necessary?  If you are using the machine as some kind of router
> or packet forwarder, then you certainly wouldn't have to worry about
> this.
>
> Are you running Solaris?  or Linux?
>
> If you are still running Linux, then just use the machine.  eth0 and
eth1
> exist and you can set up ipmasq or whatever.  If you are using
solaris,
> create a hostname.le1 in /etc with the ip you want the sbus device to
> use, and upon reboot it should come up.
>
> --
> [ Mike Sawicki < f i f i at HAX d.o.t org >  *DIE SPAM BOTS!* .... ]
> [ UNIX Systems Administrator, Urinal Repairman and Underwear Model ]
> [ HAX Engineering & Administration / Phone:<732.396.8653> ........ ]
> [ url...http://www.HAX.ORG {/fifi for personal info} ............. ]
>
> "Linux is something for windows haters, BSD is something for unix
lovers"
>
>  --- "i am AN hacker." ---
>

Although SUNS do report the MAC located on the clock board by default,
in some architectures ie sun4u setting the OBP param "local-mac-
address? true" will force the local cards to report there own 48 bit
address. If the main interface is on the system board it will still
report the system wide MAC"



Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.