cvs commit: src/sys/net if_ethersubr.c

[Bruce M. Simpson]

Bruce M Simpson wrote:
>   The handling of CARP needs to be revisited architecturally.

By this I mean: we could benefit from a more general mechanism used to 
deal with link-layer addresses other than the hardware station address, 
perhaps kept in a hash table.

Some cards support more than one station address being set up with a 
perfect hash filter entry; I think fxp supports this. Perhaps Bill Paul 
knows more?

>  The M_PROMISC
>   flag may potentially be demoted to a link-layer flag only as it is in
>   NetBSD, where the idea originated.
>   
Separate issue, trivial change.

BMS