At 08:26 AM 6/8/00 -0700, Ben Greear wrote:
>> At 09:30 PM 6/7/00 -0700, Ben Greear wrote:
>> >Dennis wrote:
>> >They can be called whatever you want, within reason, so long as they are
>> >no more than 15 characters long. They should probably NOT be called eth*,
>> >as that will be confusing, as you mention.
>> >There are some folks that already have bridging working too in the 2.3
>> >kernel I think....
>> We've had our own bridging code for years....the virtual bridge groups are
>> needed for cleanly tying bridged serial interfaces (such as Frame Relay
>> DLCIs, used for DSL bridging) into normal operations without having to hack
>> the arp or dev code in the kernel. The model works quite nicely in freebsd
>> and is working in linux also....I just wanted to be aware of any roadblocks
>> I might have hit up the road. The model allows unlimited bridge groups (on
>> a T3 DSL frame line you may have 900+ necessary, although it seems unlikely
>> you'd have that many ethernet ports in a system :-) ).....
>You could very easily have that many VLAN devices though :)
>Some of us have been thinking about giving a hashed or tree based
>lookup for devices by name or dev_id, instead of the current linear
>walk of the devices list. That should help you as well.
Yes, its a bit sluggish with 1000 devices. Our FR structures are hashed,
but not the devices.
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to firstname.lastname@example.org
This archive was generated by hypermail 2b29 : Thu Jun 15 2000 - 21:00:40 EST