Acked-by: Justin Pettit
On Sep 3, 2013, at 5:40 PM, Ethan Jackson wrote:
> xlate_normal() held the mac_learning lock while calling
> output_normal(). When running with patch ports, this could cause
> xlate_actions() to be called again, possibly attempting to take a
> write lock on the same l
xlate_normal() held the mac_learning lock while calling
output_normal(). When running with patch ports, this could cause
xlate_actions() to be called again, possibly attempting to take a
write lock on the same learning table causing a deadlock. This patch
solves the problem by holding the lock fo
