net/sched: cls_flower: Try to offload only if skip_hw flag isn't set
authorHadar Hen Zion <hadarh@mellanox.com>
Thu, 1 Dec 2016 12:06:34 +0000 (14:06 +0200)
committerDavid S. Miller <davem@davemloft.net>
Fri, 2 Dec 2016 18:28:36 +0000 (13:28 -0500)
Check skip_hw flag isn't set before calling
fl_hw_{replace/destroy}_filter and fl_hw_update_stats functions.

Replace the call to tc_should_offload with tc_can_offload.
tc_can_offload only checks if the device supports offloading, the check for
skip_hw flag is done earlier in the flow.

Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/sched/cls_flower.c

index e8dd09af0d0ce60fccaced8bbd90f9af5434cc27..5e70f65442d393d9469999136cb9e8c660c8fa16 100644 (file)
@@ -207,7 +207,7 @@ static void fl_hw_destroy_filter(struct tcf_proto *tp, unsigned long cookie)
        struct tc_cls_flower_offload offload = {0};
        struct tc_to_netdev tc;
 
-       if (!tc_should_offload(dev, tp, 0))
+       if (!tc_can_offload(dev, tp))
                return;
 
        offload.command = TC_CLSFLOWER_DESTROY;
@@ -231,7 +231,7 @@ static int fl_hw_replace_filter(struct tcf_proto *tp,
        struct tc_to_netdev tc;
        int err;
 
-       if (!tc_should_offload(dev, tp, flags))
+       if (!tc_can_offload(dev, tp))
                return tc_skip_sw(flags) ? -EINVAL : 0;
 
        offload.command = TC_CLSFLOWER_REPLACE;
@@ -259,7 +259,7 @@ static void fl_hw_update_stats(struct tcf_proto *tp, struct cls_fl_filter *f)
        struct tc_cls_flower_offload offload = {0};
        struct tc_to_netdev tc;
 
-       if (!tc_should_offload(dev, tp, 0))
+       if (!tc_can_offload(dev, tp))
                return;
 
        offload.command = TC_CLSFLOWER_STATS;
@@ -275,7 +275,8 @@ static void fl_hw_update_stats(struct tcf_proto *tp, struct cls_fl_filter *f)
 static void __fl_delete(struct tcf_proto *tp, struct cls_fl_filter *f)
 {
        list_del_rcu(&f->list);
-       fl_hw_destroy_filter(tp, (unsigned long)f);
+       if (!tc_skip_hw(f->flags))
+               fl_hw_destroy_filter(tp, (unsigned long)f);
        tcf_unbind_filter(tp, &f->res);
        call_rcu(&f->rcu, fl_destroy_filter);
 }
@@ -743,20 +744,23 @@ static int fl_change(struct net *net, struct sk_buff *in_skb,
                        goto errout;
        }
 
-       err = fl_hw_replace_filter(tp,
-                                  &head->dissector,
-                                  &mask.key,
-                                  &fnew->key,
-                                  &fnew->exts,
-                                  (unsigned long)fnew,
-                                  fnew->flags);
-       if (err)
-               goto errout;
+       if (!tc_skip_hw(fnew->flags)) {
+               err = fl_hw_replace_filter(tp,
+                                          &head->dissector,
+                                          &mask.key,
+                                          &fnew->key,
+                                          &fnew->exts,
+                                          (unsigned long)fnew,
+                                          fnew->flags);
+               if (err)
+                       goto errout;
+       }
 
        if (fold) {
                rhashtable_remove_fast(&head->ht, &fold->ht_node,
                                       head->ht_params);
-               fl_hw_destroy_filter(tp, (unsigned long)fold);
+               if (!tc_skip_hw(fold->flags))
+                       fl_hw_destroy_filter(tp, (unsigned long)fold);
        }
 
        *arg = (unsigned long) fnew;
@@ -879,7 +883,8 @@ static int fl_dump(struct net *net, struct tcf_proto *tp, unsigned long fh,
                        goto nla_put_failure;
        }
 
-       fl_hw_update_stats(tp, f);
+       if (!tc_skip_hw(f->flags))
+               fl_hw_update_stats(tp, f);
 
        if (fl_dump_key_val(skb, key->eth.dst, TCA_FLOWER_KEY_ETH_DST,
                            mask->eth.dst, TCA_FLOWER_KEY_ETH_DST_MASK,