On Jul 25, 2023, at 8:51 AM, Stefan Paetow (OpenSource) <oss@xxxxxxxx> wrote: > > > So, I'm thinking of adding it either as an extension (i.e. extending > syntax from s, d, and x, to s, d, x, and v) or as a separate switch > (-V) altogether. Given that the functionality would simply take the > value portion to dig out the vendor enterprise number, its attr_id, > syntax and the value and reuse most of the existing stuff, I'm > wondering which Jouni and the people at large in this group would > prefer. RFC 6929 defines an OID syntax for attributes. https://www.rfc-editor.org/rfc/rfc6929.html#section-2.7.2 > For example, send the VSA of 'Cisco-AVPair="foo=bar"'. > > Existing approach: > > -N 26:x:000000090109666f6f3d626172 I would suggest: -N 26.9.1:s:foo=bar There needs to be special handling for 26, that the next field is a 32-bit vendor ID. But after that, the code could assume that "-N ...n.m..." is just a TLV "n" with sub-TLV "m", in 8-bit form. That would cover the bulk of the VSAs. It should arguably still allow -N 26:x:... Alan DeKok. _______________________________________________ Hostap mailing list Hostap@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/hostap
