Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 18 Sep 1998 13:29:27 +0800
From:      chas <panda@peace.com.my>
To:        freebsd-questions@FreeBSD.ORG
Cc:        dan@dpcsys.com, Clark@open.org, dwhite@resnet.uoregon.edu, SteveFriedrich@Hot-Shot.com
Subject:   Thank you - works OK ! Re: 2 NICs disable all network services. 
Message-ID:  <3.0.32.19980918132927.006b86a0@mail.peace.com.my>

next in thread | raw e-mail | index | archive | help
Thank you to everybody for your help...  all of you were correct.

1. The new NIC took over fxp0 instead of fxp1. 
2. I had the netmask wrong for 10.0.0.1

But this is very strange : last nite I put the UTP into the 
new card and it didn't work. The cure was to do the following :
a) remove the original NIC
b) use only the new NIC which takes fxp0
c) replace the original NIC.
This also had the side effect of changing the IRQ !

The following chronological dmesg outputs explain this in detail :

A) With only 1 card : (the original NIC)
   ------------------
fxp0 <Intel EtherExpress Pro 10/100B Ethernet> rev 2 int a irq 15 on pci0:18:0
fxp0: Ethernet address 00:a0:c9:cf:b8:b2 

   Works OK.

B) With 2 cards : (added second, new NIC) :
   --------------
fxp0 <Intel EtherExpress Pro 10/100B Ethernet> rev 5 int a irq 9 on pci0:16:0
fxp0: Ethernet address 00:a0:c9:e4:c2:ed  
fxp1 <Intel EtherExpress Pro 10/100B Ethernet> rev 2 int a irq 15 on pci0:18:0
fxp1: Ethernet address 00:a0:c9:cf:b8:b2

   This failed.

3) With just the new NIC (removed the original one) :
   ----------------------
fxp0 <Intel EtherExpress Pro 10/100B Ethernet> rev 5 int a irq 9 on pci0:16:0
fxp0: Ethernet address 00:a0:c9:e4:c2:ed  

   This worked.

4) Replaced the original NIC (thus, 2 cards) :
   -------------------------
fxp0 <Intel EtherExpress Pro 10/100B Ethernet> rev 5 int a irq 9 on pci0:16:0
fxp0: Ethernet address 00:a0:c9:e4:c2:ed
fxp1 <Intel EtherExpress Pro 10/100B Ethernet> rev 2 int a irq 11 on pci0:18:0
fxp1: Ethernet address 00:a0:c9:cf:b8:b2

This works :) 

But notice that between B) and D), the IRQ for fxp1 (which
turns out to be the original NIC) has changed ! (I did not
swap the cards between PCI slots) I'm guessing that some
auto-detect magic in the BIOS finally kicked in.

Sincere thanks once again to one and all :-)

chas

Dan Busarow wrote:
>On Fri, 18 Sep 1998, chas wrote:
>> Our FreeBSD box was working fine with one NIC. Upon adding a second
>> NIC, network services didn't work. The extra NIC was detected as dmesg 
>> shows below :
>
>Just a guess.  The new NIC may have taken over as fxp0 leaving your
>networks reversed.  There's no problem running multiple NICs.
>
>Dan
>-- 
> Dan Busarow


Robert Clark wrote :
>Do you have your network cables reversed?
>
>Sometimes PCI bus numbering is counter-intuitive.
>
>IE: The second card you installed, may be fxp0 and not fxp1.
>
>(Been there myself.)
>
>[RC]

"Steve Friedrich" wrote :
>>ifconfig_fxp1="inet 10.0.0.1  netmask 255.255.255.0" 
>
>The netmask for a Class A network, i.e., network 10, is 255.0.0.0
>You may have other problems, as well.  I would reconfigure it as
>original, even going so far as to remove the new NIC, and getting it
>back to where it worked, and then we'll try again.  I've never
>misconfigued a netmask this way, so I've never seen it's effect.  If
>changing it doesn't clear up everything, I'd revert to original
>config...


Doug White wrote:
>On Fri, 18 Sep 1998, chas wrote:
>
>> Added second NIC :
>> ------------------
>> Our FreeBSD box was working fine with one NIC. Upon adding a second
>> NIC, network services didn't work. The extra NIC was detected as dmesg 
>> shows below :
>
>Please explain.  Provide lots of details.  
>
>> fxp0 <Intel EtherExpress Pro 10/100B Ethernet> rev 5 int a irq 9 on
pci0:16:0
>> fxp0: Ethernet address 00:a0:c9:e4:c2:ed
>> fxp1 <Intel EtherExpress Pro 10/100B Ethernet> rev 2 int a irq 15 on
pci0:18:0
>> fxp1: Ethernet address 00:a0:c9:cf:b8:b2
>> 
>> I also added the following to /etc/rc.conf :
>> 
>> network_interfaces="fxp0 fxp1 lo0"
>> ifconfig_fxp0="inet 202.184.153.5  netmask 255.255.255.0"
>> ifconfig_fxp1="inet 10.0.0.1  netmask 255.255.255.0" 
>
>Looks good to me.
>
>> 1) rebuilt the kernel with OPTIONS IPFIREWALL enabled
>> 2) set the following in /etc/rc.conf :
>> 	firewall_enable="YES"
>> 	firewall_type="OPEN"
>> 	gateway_enable="YES"
>> 3) /etc/rc.firewall was left as the default for "OPEN" which, if I 
>>    understand this correctly, should let everything through :
>>    $fwcmd add 100 pass all from any to any via lo0
>>    $fwcmd add 200 deny all from any to 127.0.0.0/8
>>    # Prototype setups.
>>    if [ "${firewall_type}" = "open" -o "${firewall_type}" = "OPEN" ]; then
>>          $fwcmd add 65000 pass all from any to any
>> 
>> This still didn't help. Again, dmesg and ifconfig report the NICs are
there 
>> but IP services are totally dead.
>
>Turn on ipfw logging and see where things are going.

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3.0.32.19980918132927.006b86a0>