On Thu, Jul 10, 2008 at 02:26:26PM +0400, Dima Veselov wrote: [...] > So, is it true, that vlan driver is pretty smart to take vlan tag from card, > when it is provided or parse whole packet if 802.1q is not supported on > hardware? Not sure I'd call that behaviour smart, but yes. > > Also, is it possible to catch clear IP packets, in their real condition? tcpdump(8) will always show you the packet as received from the hardware. On smart hardware wher VLAN tagging is supported, the device sends 802.1q information in control structures, so we don't get the packet as it was on the wire, and there is no way to do it (that's kind of the point of having the hardware handle VLAN tagging, anyway). [...] > In that configuration I can see incoming tagged packets are properly goes to > vlan(4) interfaces, but outgoing are always fall to switch's native vlan. > > What can be the problem in that case? And how can I debug this? Routes were > flushed (even booted with that configuration) The only scenario I can think of would involve a bug in agr stripping the vlan information tag from the mbufs it passes to its child interfaces. That would happen only when the hw is configured to do vlan tagging itself, as otherwise agr would get a full frame including the 802.1q header to pass to its child interfaces. -- Quentin Garnier - cube%cubidou.net@localhost - cube%NetBSD.org@localhost "See the look on my face from staying too long in one place [...] every time the morning breaks I know I'm closer to falling" KT Tunstall, Saving My Face, Drastic Fantastic, 2007.
Description: PGP signature