Dave,
Thanks for taking the time to help me here!
On Wed, Feb 22, 2017 at 4:47 PM, Dave Barach (dbarach) <dbar...@cisco.com>
wrote:
> You would certainly have to rewrite vnet_create_loopback_interface(...) a
> little bit. That’s not necessarily a bad thing.
>
Yeah, I figured as much on that front... :-) I'm not afraid of
giving it a swing.
> Please brain-police the instance number: flunk attempts to create specific
> loopback instances (e.g. loop0) more than once, and so forth. Bitmap, plus
> some upper limit (maybe 16k?) on the loopback instance number.
>
Can do. Two questions. If a loopback interface is deleted,
would that free-up the instance number for re-use? Or is
there lingering reason to never reuse an instance number?
And second, is there a bit-vector implementation around?
I would assume so and I just need to go looking...?
I still need to resolve the name expectations; that is, expecting
"Loopback" to be the base, and not "loop". Does one of these
options make more sense here?
1) Have the API create call supply the desired name directly,
defaulting to "loop"-names if none are specified.
2) Have a loopback-module-global-ish config variable stating
to be in "loop" or "Loopback" mode. Mode can be set via
new loopback API config call, or via VPP startup config setting.
3) Tell our users to get used to the New "loop" World Order? :-)
HTH… Dave
>
Thanks,
jdl
> PS -- At the risk of foretelling the future, bridge names are going to
have this same naming issue, right? 'Cuz those are next up!
_______________________________________________
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev
