Date: Sat, 27 Mar 2004 08:48:37 -0800 From: "Jacob S. Barrett" <jbarrett@amduat.net> To: Ruslan Ermilov <ru@freebsd.org> Cc: freebsd-net@freebsd.org Subject: Re: Disabling VLAN_HWTAGGING Message-ID: <200403270848.37996.jbarrett@amduat.net> In-Reply-To: <200403270753.53476.jbarrett@amduat.net> References: <200403251118.40718.jbarrett@amduat.net> <20040327074205.GA32984@ip.net.ua> <200403270753.53476.jbarrett@amduat.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Saturday 27 March 2004 07:53 am, Jacob S. Barrett wrote: > Well with a new "correct" MAC that pings go back and forth just fine now. > I will back out all my changes and see if they still work with the hardware > tagging/detagging enabled. OK, with the hardware support re-enabled the frame now enters the driver. It is detected as VLAN frame and sent to VLAN_INPUT_TAG. The frame is then delivered via lower hook to the ng_vlan where it doesn't match the vlan tag so it goes out the nomatch hook. I guess with the VLAN tag stripped from the frame that ng_vlan can't match it. Is this the expected behavior with ng_vlan? I can just comment out the VLAN stripping line in the driver if it is the expected behavior. That isn't a big deal really. I haven't done a whole lot with the network interface drivers other than a few minor fixes here and there, but would it be hard to add some sort of flag to enable/disable the tag stripping via ifconfig? I was thinking that could be down through a "linkx" flag right? If the driver got the say "link1" it would down the interface, set clear the stripping config bits, and then re-up the interface. Or would this be better handled by a sysctl option? Does that sound safe and do-able? -- Jacob S. Barrett jbarrett@amduat.net www.amduat.net "I don't suffer from insanity, I enjoy every minute of it."
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200403270848.37996.jbarrett>