[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: ethernet configuration



[ Please turn off the HTML part of your messages (I think the Gmail web
  interface calls it something like "rich format") and please stop
  top-posting (I will fix that for this message). ]

On Fri, Jul 04, 2008 at 11:15:05 -0300, André Timpanaro wrote:
> On Wed, Jul 2, 2008 at 2:01 PM, Florian Kulzer wrote:
> > On Wed, Jul 02, 2008 at 10:43:58 +0300, Andrei Popescu wrote:
> > > On Tue, Jul 01, 2008 at 06:04:17PM -0300, André Timpanaro wrote:
> > > > I've found the /etc/udev/rules.d/z25_persistent-net.rules file:

[...]

> > > I've heard here that some (all?) nVidia chipsets will change MAC on
> > > every boot. That's your problem.

[...]

> > > You can look in the BIOS for a setting to have a persistent MAC or
> > > rewrite the rule to allocate device names based only on the kernel
> > > module (forcedeth?).
> >
> > You could try to use a wildcard in z25_persistent-net.rules. This might
> > work as long as you don't put a second nvidia network card into this
> > machine:
> >
> > # PCI device 0x10de:0x03ef (forcedeth)
> > SUBSYSTEM=="net", DRIVERS=="?*", ATTRS{address}=="00:00:6c:*", NAME="eth0"
>
> I made the change in 'z25_persistent-net.rules' and it stoped creating new
> interfaces.

OK, so far, so good. At least the proliferation of ethX devices should
stop now.

> > To make sure that the rest of your network sees a consistent MAC address
> > for this computer, try to set it explicitly in the eth0 stanza of
> > /etc/network/interfaces:
> >
> > # The primary network interface
> > allow-hotplug eth0
> > iface eth0 inet dhcp
> >    hwaddress ether 00:00:6c:e9:01:77
>
> But specifying the MAC address didn't worked (he configured the
> interface correctly but I still couldn't connect to the internet). Probably
> it wasn't 00:00:6c:e9:01:77 anymore for eth0 (I suspected that would happen
> but decided to try it anyway).

You can check the IP address with "ifconfig eth0". Does your DHCP server
base the DHCP leases on MAC addresses?

> Isn't there a sure way to retrieve the MAC address correctly?

The suggestion above was inspired by this article:

http://www.debian-administration.org/articles/501

I have no idea if it works for all forcedeth-related problems.

> By the way, my BIOS doesn't allow changes in the settings for the net card,
> so I can't make the MAC address persistent.

According to the comments in forcedeth.c, the problem with retrieving
the correct MAC is caused by a buggy BIOS, so a BIOS upgrade might help.

AFAICT, the forcedeth driver does not allow to set a fixed MAC address
via a module parameter. If you compile your own kernel then you can try
to change the code that sets the random address

                dev->dev_addr[0] = 0x00;
                dev->dev_addr[1] = 0x00;
                dev->dev_addr[2] = 0x6c;
                get_random_bytes(&dev->dev_addr[3], 3);

to this

                dev->dev_addr[0] = 0x00;
                dev->dev_addr[1] = 0x00;
                dev->dev_addr[2] = 0x6c;
                dev->dev_addr[3] = 0xe9;
                dev->dev_addr[4] = 0x01;
                dev->dev_addr[5] = 0x77;

I have no way of testing if this actually works. Google finds some
recent posts on LKML about patches for the MAC detection of forcedeth,
so you might want to try the newer code (especially if you are
currently using the version in Etch).

http://lkml.org/lkml/2008/4/8/391
http://marc.info/?t=119215716900001&r=1&w=2

-- 
Regards,            | http://users.icfo.es/Florian.Kulzer
          Florian   |


Reply to: