From: David M. <da...@da...> - 2015-01-05 04:43:26
|
From: Thomas Graf <tg...@su...> Date: Sun, 4 Jan 2015 09:53:20 +0000 > As a further optimization, you could add a new atomic function > rhashtable_lookup_and_insert() which holds the per bucket lock during > lookup and use that instead. This would allow you to get rid of the > mutex alltogether. This would definitely be a better implementation in my opinion. |