Merge branch 'flower-offload'

Amir Vadai says:

====================
cls_flower hardware offload support

Please see changes from V2 at the bottom.

This patchset introduces cls_flower hardware offload support over ConnectX-4
driver, more hardware vendors are welcome to use it too.

This patchset is based on John's infrastructure for tc offloading [2] to add
hardware offload support to the flower filter. It also extends the support to
an additional tc action - skbedit mark operation.
NIC driver that was used is ConnectX-4. Feature is off by default and could be
turned on using ethtool.

Some commands to use this code:

export TC=../iproute2/tc/tc
export ETH=ens9

ethtool  -K ens9 hw-tc-offload on

$TC qdisc add dev $ETH ingress

$TC filter add dev $ETH protocol ip prio 20 parent ffff: \
	flower ip_proto 1 \
	dst_mac 7c:fe:90:69:81:62 \
	src_mac 7c:fe:90:69:81:56 \
	dst_ip 11.11.11.11 \
	src_ip 11.11.11.12 \
	indev $ETH \
	action drop

$TC filter add dev $ETH protocol ip prio 30 parent ffff: \
	flower ip_proto 6 \
	indev $ETH \
	action skbedit mark 0x1234

$TC filter add dev $ETH protocol ip prio 10 parent ffff: \
	handle 0x1234 fw action pass

The code was tested and applied on top of commit 3ebeac1 ("Merge branch
'cxgb4-next'")

Changes from V2:
- patch 1/10 ("net/flower: Introduce hardware offload support")
  - Remove unused variable [Dave]
  - Don't fail command when HW can't offload filter [John]
- patch 3/10 ("net/sched: Macro instead of CONFIG_NET_CLS_ACT ifdef")
  - Mention in changelog that struct tc_action is now exposed out of the ifdef.
- patch 4/10 ("net/act_skbedit: Utility functions for mark action")
  - Document clearly that is_tcf_skbedit_mark() is returning true if and only
    if the only action is mark [Dave]
- patch 8/10 ("net/mlx5e: Introduce tc offload support")
  - make mlx5e_tc_add_flow() static

Changes from V1:
- patch 3/10 ("net/sched: Macro instead of CONFIG_NET_CLS_ACT ifdef")
  - fixed return value of tc_no_actions

Changes from V0:
- Use tc_no_actions and tc_for_each_action instead of ifdef CONFIG_NET_CLS_ACT
- Replace ENOTSUPP (and some EINVAL) with EOPNOTSUPP
- Name the flower command enum
- fl_hw_destroy_filter() to return void - nobody uses the return value
- mlx5e_tc_init() and mlx5e_tc_cleanup() to be called from the right places.
- When adding HW rule fails - fail the command
- Rules are added to be processed both by HW and SW unless SKIP_HW is given
- Adding patch 6/10 ("net/mlx5e: Relax ndo_setup_tc handle restriction")

Main changes from the RFC [1]:
- API
  - Using ndo_setup_tc() instead of switchdev
- act_skbedit, act_gact
  - Actions are not serialized to NIC driver, instead using access functions.
- cls_flower
  - prevent double classification by software by not adding
    successfuly offloaded filters to the hashtable
  - Fixed some bugs in original RFC with rule delete
- mlx5
  - Adding flow table to kernel namespace instead of a new namespace
  - s/offload/tc/ in many places
  - no need for a special kconfig since switchdev is not used
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 63 insertions, 1 deletions

diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 95b021243233..25d87666bf1e 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -165,6 +165,51 @@ static void fl_destroy_filter(struct rcu_head *head)
 	kfree(f);
 }
 
+static void fl_hw_destroy_filter(struct tcf_proto *tp, u64 cookie)
+{
+	struct net_device *dev = tp->q->dev_queue->dev;
+	struct tc_cls_flower_offload offload = {0};
+	struct tc_to_netdev tc;
+
+	if (!tc_should_offload(dev, 0))
+		return;
+
+	offload.command = TC_CLSFLOWER_DESTROY;
+	offload.cookie = cookie;
+
+	tc.type = TC_SETUP_CLSFLOWER;
+	tc.cls_flower = &offload;
+
+	dev->netdev_ops->ndo_setup_tc(dev, tp->q->handle, tp->protocol, &tc);
+}
+
+static void fl_hw_replace_filter(struct tcf_proto *tp,
+				 struct flow_dissector *dissector,
+				 struct fl_flow_key *mask,
+				 struct fl_flow_key *key,
+				 struct tcf_exts *actions,
+				 u64 cookie, u32 flags)
+{
+	struct net_device *dev = tp->q->dev_queue->dev;
+	struct tc_cls_flower_offload offload = {0};
+	struct tc_to_netdev tc;
+
+	if (!tc_should_offload(dev, flags))
+		return;
+
+	offload.command = TC_CLSFLOWER_REPLACE;
+	offload.cookie = cookie;
+	offload.dissector = dissector;
+	offload.mask = mask;
+	offload.key = key;
+	offload.exts = actions;
+
+	tc.type = TC_SETUP_CLSFLOWER;
+	tc.cls_flower = &offload;
+
+	dev->netdev_ops->ndo_setup_tc(dev, tp->q->handle, tp->protocol, &tc);
+}
+
 static bool fl_destroy(struct tcf_proto *tp, bool force)
 {
 	struct cls_fl_head *head = rtnl_dereference(tp->root);
@@ -174,6 +219,7 @@ static bool fl_destroy(struct tcf_proto *tp, bool force)
 		return false;
 
 	list_for_each_entry_safe(f, next, &head->filters, list) {
+		fl_hw_destroy_filter(tp, (u64)f);
 		list_del_rcu(&f->list);
 		call_rcu(&f->rcu, fl_destroy_filter);
 	}
@@ -459,6 +505,7 @@ static int fl_change(struct net *net, struct sk_buff *in_skb,
 	struct cls_fl_filter *fnew;
 	struct nlattr *tb[TCA_FLOWER_MAX + 1];
 	struct fl_flow_mask mask = {};
+	u32 flags = 0;
 	int err;
 
 	if (!tca[TCA_OPTIONS])
@@ -486,6 +533,9 @@ static int fl_change(struct net *net, struct sk_buff *in_skb,
 	}
 	fnew->handle = handle;
 
+	if (tb[TCA_FLOWER_FLAGS])
+		flags = nla_get_u32(tb[TCA_FLOWER_FLAGS]);
+
 	err = fl_set_parms(net, tp, fnew, &mask, base, tb, tca[TCA_RATE], ovr);
 	if (err)
 		goto errout;
@@ -498,9 +548,20 @@ static int fl_change(struct net *net, struct sk_buff *in_skb,
 				     head->ht_params);
 	if (err)
 		goto errout;
-	if (fold)
+
+	fl_hw_replace_filter(tp,
+			     &head->dissector,
+			     &mask.key,
+			     &fnew->key,
+			     &fnew->exts,
+			     (u64)fnew,
+			     flags);
+
+	if (fold) {
 		rhashtable_remove_fast(&head->ht, &fold->ht_node,
 				       head->ht_params);
+		fl_hw_destroy_filter(tp, (u64)fold);
+	}
 
 	*arg = (unsigned long) fnew;
 
@@ -527,6 +588,7 @@ static int fl_delete(struct tcf_proto *tp, unsigned long arg)
 	rhashtable_remove_fast(&head->ht, &f->ht_node,
 			       head->ht_params);
 	list_del_rcu(&f->list);
+	fl_hw_destroy_filter(tp, (u64)f);
 	tcf_unbind_filter(tp, &f->res);
 	call_rcu(&f->rcu, fl_destroy_filter);
 	return 0;