openvswitch: Remove 5-tuple optimization.
The 5-tuple optimization becomes unnecessary with a later per-NUMA node stats patch. Remove it first to make the changes easier to grasp. Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com> Signed-off-by: Jesse Gross <jesse@nicira.com>
This commit is contained in:
committed by
Jesse Gross
parent
8c63ff09bd
commit
23dabf88ab
@@ -155,14 +155,6 @@ struct flow_stats {
|
||||
__be16 tcp_flags; /* Union of seen TCP flags. */
|
||||
};
|
||||
|
||||
struct sw_flow_stats {
|
||||
bool is_percpu;
|
||||
union {
|
||||
struct flow_stats *stat;
|
||||
struct flow_stats __percpu *cpu_stats;
|
||||
};
|
||||
};
|
||||
|
||||
struct sw_flow {
|
||||
struct rcu_head rcu;
|
||||
struct hlist_node hash_node[2];
|
||||
@@ -172,7 +164,7 @@ struct sw_flow {
|
||||
struct sw_flow_key unmasked_key;
|
||||
struct sw_flow_mask *mask;
|
||||
struct sw_flow_actions __rcu *sf_acts;
|
||||
struct sw_flow_stats stats;
|
||||
struct flow_stats __percpu *stats;
|
||||
};
|
||||
|
||||
struct arp_eth_header {
|
||||
|
Reference in New Issue
Block a user