On Apr 24, 2014, at 6:15 PM, Ethan Jackson <[email protected]> wrote:
> In the cache_push_back function, you might consider using the
> x2nrealloc() function.
>
I did not know this existed, thanks!
> Does this actually help? I've found we spend most of our time getting
> the memory for the rule, not the subtable itself.
>
After this series I see cache misses only on fetching the ‘next’ rule. Before,
I saw also misses for the subtable (being fetched via a next pointer as well)
and/or it’s mask. So this helps :-)
Jarno
_______________________________________________
dev mailing list
[email protected]
http://openvswitch.org/mailman/listinfo/dev
