* [PATCH 0/2] Get rid of "multi-hh_cache" capability.
@ 2011-07-13 8:52 David Miller
0 siblings, 0 replies; only message in thread
From: David Miller @ 2011-07-13 8:52 UTC (permalink / raw)
To: netdev
We support having multiple hh_cache entries in a list hung off
of a neighbour entry, one per unique ->hh_type protocol value.
Problem is, nobody uses this.
Neighbour entries are for one protocol type only (which determines the
type of the key), whose protocol is "dst_ops->protocol".
Kill all of this noise, which dates back to before Alexey Kuznetsov
added the neighbour layer. He just blinded copied over this
capability into the generic layer.
The next step will be to embed "struct hh_cache" directly into "struct
neighbour". Validity will be indicated by a non-zero hh_len.
Then we can get rid of all of this extra memory allocations, funky
refcounting, and SMP memory barriers.
And finally, move to refcount-less neigh entry usage in ip_finish_output2()
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2011-07-13 8:52 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-07-13 8:52 [PATCH 0/2] Get rid of "multi-hh_cache" capability David Miller
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).