> However if you actually try the above on a test machine after removing
> the dupedkey walk from ire_find_best_route(), I would expect it to fail
> because of bug thiru-27 which needs to be fixed as a pre-requisite.

Turns out thiru-27 was a pre-requisite, and the problems I was seeing
earlier without the rn_dupedkey match were caused by not having thiru-27.

> And if you really need that logic in ire_find_best_route() it is bad,
> because all the radix tree implementation logic is hidden from IP and is
> (should be) buried completely in radix.c.

fortunately that was not the case.

Also, you correctly pointed out that my example with the /32 node in
its own branch can never happen unless the key itself had different bits.
Sorry for that confusion.

So, in summary, thiru-27 and thiru-22 are both ACCEPTed.

Thanks,
--Sowmini

_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to