Rework VRF_DEFAULT and VRF_UNDEFINED

[donaldsharp]

From the static vrf route leaking PR:

So it's unfortunate that VRF_DEFAULT is zero and unknown is 0xffffffff. If we reversed this, then it would be possible to use vrf_id when vrf_id =0=UNKNOWN and it wouldn't be necessary to change all users of zapi. @donaldsharp what do you think it it worth making this change, or does its downsides outweigh the benefit...

I think that there would be benefits from moving this over. At this time I have seen lots of places where VRF_DEFAULT was assumed to be zero and doing this change would cause issues. I don't have time to do this re-org right now so create an issue for someone to pick up in the future

[pguibert6WIND]

why don't you create map the macro to a function ?
Inside, you can do the check that you want.
This is what I propose in #1633

[pguibert6WIND]

some checks on VRF_DEFAULT have been pushed by 4b1e01b

[qlyoung]

@donaldsharp VRF_UNDEFINED no longer exists, I'm just gonna assume we can close this one, if not reopen it