openvswitch: Mega flow implementation

Add wildcarded flow support in kernel datapath.

Wildcarded flow can improve OVS flow set up performance by avoid sending
matching new flows to the user space program. The exact performance boost
will largely dependent on wildcarded flow hit rate.

In case all new flows hits wildcard flows, the flow set up rate is
within 5% of that of linux bridge module.

Pravin has made significant contributions to this patch. Including API
clean ups and bug fixes.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: Andy Zhou <azhou@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>

Documentation/networking/openvswitch.txt

@@ -91,6 +91,46 @@ Often we ellipsize arguments not important to the discussion, e.g.:
     in_port(1), eth(...), eth_type(0x0800), ipv4(...), tcp(...)
 
 
+Wildcarded flow key format
+--------------------------
+
+A wildcarded flow is described with two sequences of Netlink attributes
+passed over the Netlink socket. A flow key, exactly as described above, and an
+optional corresponding flow mask.
+
+A wildcarded flow can represent a group of exact match flows. Each '1' bit
+in the mask specifies a exact match with the corresponding bit in the flow key.
+A '0' bit specifies a don't care bit, which will match either a '1' or '0' bit
+of a incoming packet. Using wildcarded flow can improve the flow set up rate
+by reduce the number of new flows need to be processed by the user space program.
+
+Support for the mask Netlink attribute is optional for both the kernel and user
+space program. The kernel can ignore the mask attribute, installing an exact
+match flow, or reduce the number of don't care bits in the kernel to less than
+what was specified by the user space program. In this case, variations in bits
+that the kernel does not implement will simply result in additional flow setups.
+The kernel module will also work with user space programs that neither support
+nor supply flow mask attributes.
+
+Since the kernel may ignore or modify wildcard bits, it can be difficult for
+the userspace program to know exactly what matches are installed. There are
+two possible approaches: reactively install flows as they miss the kernel
+flow table (and therefore not attempt to determine wildcard changes at all)
+or use the kernel's response messages to determine the installed wildcards.
+
+When interacting with userspace, the kernel should maintain the match portion
+of the key exactly as originally installed. This will provides a handle to
+identify the flow for all future operations. However, when reporting the
+mask of an installed flow, the mask should include any restrictions imposed
+by the kernel.
+
+The behavior when using overlapping wildcarded flows is undefined. It is the
+responsibility of the user space program to ensure that any incoming packet
+can match at most one flow, wildcarded or not. The current implementation
+performs best-effort detection of overlapping wildcarded flows and may reject
+some but not all of them. However, this behavior may change in future versions.
+
+
 Basic rule for evolving flow keys
 ---------------------------------
 

include/uapi/linux/openvswitch.h

@@ -1,6 +1,6 @@
 
 /*
- * Copyright (c) 2007-2011 Nicira Networks.
+ * Copyright (c) 2007-2013 Nicira, Inc.
  *
  * This program is free software; you can redistribute it and/or
  * modify it under the terms of version 2 of the GNU General Public
@@ -379,6 +379,12 @@ struct ovs_key_nd {
  * @OVS_FLOW_ATTR_CLEAR: If present in a %OVS_FLOW_CMD_SET request, clears the
  * last-used time, accumulated TCP flags, and statistics for this flow.
  * Otherwise ignored in requests.  Never present in notifications.
+ * @OVS_FLOW_ATTR_MASK: Nested %OVS_KEY_ATTR_* attributes specifying the
+ * mask bits for wildcarded flow match. Mask bit value '1' specifies exact
+ * match with corresponding flow key bit, while mask bit value '0' specifies
+ * a wildcarded match. Omitting attribute is treated as wildcarding all
+ * corresponding fields. Optional for all requests. If not present,
+ * all flow key bits are exact match bits.
  *
  * These attributes follow the &struct ovs_header within the Generic Netlink
  * payload for %OVS_FLOW_* commands.
@@ -391,6 +397,7 @@ enum ovs_flow_attr {
 	OVS_FLOW_ATTR_TCP_FLAGS, /* 8-bit OR'd TCP flags. */
 	OVS_FLOW_ATTR_USED,      /* u64 msecs last used in monotonic time. */
 	OVS_FLOW_ATTR_CLEAR,     /* Flag to clear stats, tcp_flags, used. */
+	OVS_FLOW_ATTR_MASK,      /* Sequence of OVS_KEY_ATTR_* attributes. */
 	__OVS_FLOW_ATTR_MAX
 };
 

net/openvswitch/actions.c

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2007-2012 Nicira, Inc.
+ * Copyright (c) 2007-2013 Nicira, Inc.
  *
  * This program is free software; you can redistribute it and/or
  * modify it under the terms of version 2 of the GNU General Public
@@ -376,8 +376,10 @@ static int output_userspace(struct datapath *dp, struct sk_buff *skb,
 	const struct nlattr *a;
 	int rem;
 
+	BUG_ON(!OVS_CB(skb)->pkt_key);
+
 	upcall.cmd = OVS_PACKET_CMD_ACTION;
-	upcall.key = &OVS_CB(skb)->flow->key;
+	upcall.key = OVS_CB(skb)->pkt_key;
 	upcall.userdata = NULL;
 	upcall.portid = 0;