On Mon, Jan 14, 2008 at 11:00:37AM +0100, Marco d'Itri wrote:
> On Dec 20, Lionel Elie Mamane <[EMAIL PROTECTED]> wrote:
>> No, not exactly either, I have reproduced the problem on i386
>> machines, too. Here are the results:
> This works for everybody else, so it's reasonable to assume that you
> broke your system some way. Check the generated rules.
I'm not sure what you are talking about with "generated rules". Is
that /etc/udev/rules.d/z25_persistent-net.rules ? They contain the
usual udev-automatically generated rules, with at most the "NAME"
field changed, like for example (the i386 machine with udev 0.105-4):
SUBSYSTEM=="net", DRIVERS=="?*", ATTRS{address}=="00:0e:0c:c0:8d:b9",
NAME="eth_jumbo"
SUBSYSTEM=="net", DRIVERS=="?*", ATTRS{address}=="00:17:31:90:24:db",
NAME="eth_std"
SUBSYSTEM=="net", DRIVERS=="?*", ATTRS{address}=="00:17:31:90:7d:46",
NAME="eth_modem"
another machine (amd64 with udev 0.114-2) contains:
ACTION=="add", SUBSYSTEM=="net", SYSFS{address}=="00:e0:81:54:ab:f8",
NAME="eth0"
ACTION=="add", SUBSYSTEM=="net", SYSFS{address}=="00:e0:81:54:ab:f9",
NAME="eth1"
ACTION=="add", SUBSYSTEM=="net", SYSFS{address}=="00:e0:81:00:00:23:73:96",
NAME="eth2"
I can imagine that these lines match the vlan interfaces, too, and
this is the problem because then udev tries to name the vlan interface
thus but an interface with the said name already exists? But then I
don't understand why it works with some udev/kernel
combinations. Anyway, if that is the problem, I think we should change
the automatically generated lines to fix the problem, or the vlan
package should add rules before these to preempt them.
--
Lionel
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
