Re: [v6ops] IPv6 link-local and URLs @ IETF119

[Brian E Carpenter <[e-mail]>]

On 19-Mar-24 20:55, Lorenzo Colitti wrote:
> 
> 
> On Tue, 19 Mar 2024, 17:30 Brian E Carpenter, <[e-mail] <mailto:[e-mail]>> wrote:
> 
>      >> There's a simple solution to this problem: if the user types in a link-local address the browser should open the connection on the default network interface,
> 
>     There is no such thing in some operating systems, and again, this is not the behaviour required by the use cases in the document that passed WGLC more than a year ago and which is not on this week's agenda either.
> 
> 
> How can there be no such thing? If an app connects to a global address, the stack must pick an interface. So such a concept must exist.

RFC4007 says that an implementation MAY have a default interface for link locals when the user doesn't supply one via fe80::1%name. Linux doesn't do that. Since link-locals aren't routeable, you just get 'invalid argument' from sendmsg()  if you don't supply a valid interface number.

> 
> As for the documents... I don't know what the current state is, but it sounds like there's a very real possibility that those will never go anywhere. If so, we might want to explore other options to satisfy at least some of the use cases.

That is the entire purpose of my 6man draft and David Schinazi's https draft.

    Brian
>