> The 60/64 limit is a low-level ethernet thing, and VLANs really shouldn't > matter at that level. What about this scenario: Cisco Switch: Port1: dot1q vlan trunking, vlan1 native Port2: no encap, just vlan1 Port3: no encap, just vlan2 If the cisco receives on port 1 a tagged packet for a device on vlan 2 on port 3, it would remove the header and send the packet out that port. Now if the packet it received was 64 bytes total, when it removes the header it suddenly has a 60 byte packet, too short to send out. I'm not sure if this is really a problem though... maybe the switch is supposed to add the padding itself... > I think the cisco bug is that it does > think you need +4 for the minimum size for VLAN packets. This is sounding more and more likely, except that the two other machines using different network adapters don't exhibit the problem. Maybe they also have a bug that means they always add 4 bytes. The only thing about that is that they are e100 adapters, which if I read the source correctly, aren't vlan aware in hardware and do the padding themselves, and there is nothing in there that seems to say 'if vlan frame then minimum size += 4' > Back (oh, 7 years ago??) when I read the VLAN spec, I don't > remember any requirement to bump the minimum packet size. I just had a brief glance at it, and I'm not sure if I read it correctly, but there does appear to be a provision for the switch doing the padding as required... Thanks James
