net: introduce include/net/rps.h

Move RPS related structures and helpers from include/linux/netdevice.h
and include/net/sock.h to a new include file.

Signed-off-by: Eric Dumazet <[email protected]>
Acked-by: Soheil Hassas Yeganeh <[email protected]>
Reviewed-by: David Ahern <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>

Author: edumazet

drivers/net/ethernet/intel/ice/ice_arfs.c

@@ -2,6 +2,7 @@
 /* Copyright (C) 2018-2020, Intel Corporation. */
 
 #include "ice.h"
+#include <net/rps.h>
 
 /**
  * ice_is_arfs_active - helper to check is aRFS is active

drivers/net/ethernet/mellanox/mlx4/en_netdev.c

@@ -42,6 +42,7 @@
 #include <net/ip.h>
 #include <net/vxlan.h>
 #include <net/devlink.h>
+#include <net/rps.h>
 
 #include <linux/mlx4/driver.h>
 #include <linux/mlx4/device.h>

drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c

@@ -34,6 +34,7 @@
 #include <linux/mlx5/fs.h>
 #include <linux/ip.h>
 #include <linux/ipv6.h>
+#include <net/rps.h>
 #include "en.h"
 
 #define ARFS_HASH_SHIFT BITS_PER_BYTE

drivers/net/ethernet/sfc/rx_common.c

@@ -11,6 +11,7 @@
 #include "net_driver.h"
 #include <linux/module.h>
 #include <linux/iommu.h>
+#include <net/rps.h>
 #include "efx.h"
 #include "nic.h"
 #include "rx_common.h"

drivers/net/ethernet/sfc/siena/rx_common.c

@@ -11,6 +11,7 @@
 #include "net_driver.h"
 #include <linux/module.h>
 #include <linux/iommu.h>
+#include <net/rps.h>
 #include "efx.h"
 #include "nic.h"
 #include "rx_common.h"

drivers/net/tun.c

@@ -78,6 +78,7 @@
 #include <net/ax25.h>
 #include <net/rose.h>
 #include <net/6lowpan.h>
+#include <net/rps.h>
 
 #include <linux/uaccess.h>
 #include <linux/proc_fs.h>

include/linux/netdevice.h

@@ -225,12 +225,6 @@ struct net_device_core_stats {
 #include <linux/cache.h>
 #include <linux/skbuff.h>
 
-#ifdef CONFIG_RPS
-#include <linux/static_key.h>
-extern struct static_key_false rps_needed;
-extern struct static_key_false rfs_needed;
-#endif
-
 struct neighbour;
 struct neigh_parms;
 struct sk_buff;
@@ -730,86 +724,10 @@ static inline void netdev_queue_numa_node_write(struct netdev_queue *q, int node
 #endif
 }
 
-#ifdef CONFIG_RPS
-/*
- * This structure holds an RPS map which can be of variable length.  The
- * map is an array of CPUs.
- */
-struct rps_map {
-	unsigned int len;
-	struct rcu_head rcu;
-	u16 cpus[];
-};
-#define RPS_MAP_SIZE(_num) (sizeof(struct rps_map) + ((_num) * sizeof(u16)))
-
-/*
- * The rps_dev_flow structure contains the mapping of a flow to a CPU, the
- * tail pointer for that CPU's input queue at the time of last enqueue, and
- * a hardware filter index.
- */
-struct rps_dev_flow {
-	u16 cpu;
-	u16 filter;
-	unsigned int last_qtail;
-};
-#define RPS_NO_FILTER 0xffff
-
-/*
- * The rps_dev_flow_table structure contains a table of flow mappings.
- */
-struct rps_dev_flow_table {
-	unsigned int mask;
-	struct rcu_head rcu;
-	struct rps_dev_flow flows[];
-};
-#define RPS_DEV_FLOW_TABLE_SIZE(_num) (sizeof(struct rps_dev_flow_table) + \
-    ((_num) * sizeof(struct rps_dev_flow)))
-
-/*
- * The rps_sock_flow_table contains mappings of flows to the last CPU
- * on which they were processed by the application (set in recvmsg).
- * Each entry is a 32bit value. Upper part is the high-order bits
- * of flow hash, lower part is CPU number.
- * rps_cpu_mask is used to partition the space, depending on number of
- * possible CPUs : rps_cpu_mask = roundup_pow_of_two(nr_cpu_ids) - 1
- * For example, if 64 CPUs are possible, rps_cpu_mask = 0x3f,
- * meaning we use 32-6=26 bits for the hash.
- */
-struct rps_sock_flow_table {
-	u32	mask;
-
-	u32	ents[] ____cacheline_aligned_in_smp;
-};
-#define	RPS_SOCK_FLOW_TABLE_SIZE(_num) (offsetof(struct rps_sock_flow_table, ents[_num]))
-
-#define RPS_NO_CPU 0xffff
-
-extern u32 rps_cpu_mask;
-extern struct rps_sock_flow_table __rcu *rps_sock_flow_table;
-
-static inline void rps_record_sock_flow(struct rps_sock_flow_table *table,
-					u32 hash)
-{
-	if (table && hash) {
-		unsigned int index = hash & table->mask;
-		u32 val = hash & ~rps_cpu_mask;
-
-		/* We only give a hint, preemption can change CPU under us */
-		val |= raw_smp_processor_id();
-
-		/* The following WRITE_ONCE() is paired with the READ_ONCE()
-		 * here, and another one in get_rps_cpu().
-		 */
-		if (READ_ONCE(table->ents[index]) != val)
-			WRITE_ONCE(table->ents[index], val);
-	}
-}
-
 #ifdef CONFIG_RFS_ACCEL
 bool rps_may_expire_flow(struct net_device *dev, u16 rxq_index, u32 flow_id,
 			 u16 filter_id);
 #endif
-#endif /* CONFIG_RPS */
 
 /* XPS map type and offset of the xps map within net_device->xps_maps[]. */
 enum xps_map_type {

include/net/rps.h

@@ -0,0 +1,127 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+#ifndef _NET_RPS_H
+#define _NET_RPS_H
+
+#include <linux/types.h>
+#include <linux/static_key.h>
+#include <net/sock.h>
+
+#ifdef CONFIG_RPS
+
+extern struct static_key_false rps_needed;
+extern struct static_key_false rfs_needed;
+
+/*
+ * This structure holds an RPS map which can be of variable length.  The
+ * map is an array of CPUs.
+ */
+struct rps_map {
+	unsigned int	len;
+	struct rcu_head	rcu;
+	u16		cpus[];
+};
+#define RPS_MAP_SIZE(_num) (sizeof(struct rps_map) + ((_num) * sizeof(u16)))
+
+/*
+ * The rps_dev_flow structure contains the mapping of a flow to a CPU, the
+ * tail pointer for that CPU's input queue at the time of last enqueue, and
+ * a hardware filter index.
+ */
+struct rps_dev_flow {
+	u16		cpu;
+	u16		filter;
+	unsigned int	last_qtail;
+};
+#define RPS_NO_FILTER 0xffff
+
+/*
+ * The rps_dev_flow_table structure contains a table of flow mappings.
+ */
+struct rps_dev_flow_table {
+	unsigned int		mask;
+	struct rcu_head		rcu;
+	struct rps_dev_flow	flows[];
+};
+#define RPS_DEV_FLOW_TABLE_SIZE(_num) (sizeof(struct rps_dev_flow_table) + \
+    ((_num) * sizeof(struct rps_dev_flow)))
+
+/*
+ * The rps_sock_flow_table contains mappings of flows to the last CPU
+ * on which they were processed by the application (set in recvmsg).
+ * Each entry is a 32bit value. Upper part is the high-order bits
+ * of flow hash, lower part is CPU number.
+ * rps_cpu_mask is used to partition the space, depending on number of
+ * possible CPUs : rps_cpu_mask = roundup_pow_of_two(nr_cpu_ids) - 1
+ * For example, if 64 CPUs are possible, rps_cpu_mask = 0x3f,
+ * meaning we use 32-6=26 bits for the hash.
+ */
+struct rps_sock_flow_table {
+	u32	mask;
+
+	u32	ents[] ____cacheline_aligned_in_smp;
+};
+#define	RPS_SOCK_FLOW_TABLE_SIZE(_num) (offsetof(struct rps_sock_flow_table, ents[_num]))
+
+#define RPS_NO_CPU 0xffff
+
+extern u32 rps_cpu_mask;
+extern struct rps_sock_flow_table __rcu *rps_sock_flow_table;
+
+static inline void rps_record_sock_flow(struct rps_sock_flow_table *table,
+					u32 hash)
+{
+	unsigned int index = hash & table->mask;
+	u32 val = hash & ~rps_cpu_mask;
+
+	/* We only give a hint, preemption can change CPU under us */
+	val |= raw_smp_processor_id();
+
+	/* The following WRITE_ONCE() is paired with the READ_ONCE()
+	 * here, and another one in get_rps_cpu().
+	 */
+	if (READ_ONCE(table->ents[index]) != val)
+		WRITE_ONCE(table->ents[index], val);
+}
+
+#endif /* CONFIG_RPS */
+
+static inline void sock_rps_record_flow_hash(__u32 hash)
+{
+#ifdef CONFIG_RPS
+	struct rps_sock_flow_table *sock_flow_table;
+
+	if (!hash)
+		return;
+	rcu_read_lock();
+	sock_flow_table = rcu_dereference(rps_sock_flow_table);
+	if (sock_flow_table)
+		rps_record_sock_flow(sock_flow_table, hash);
+	rcu_read_unlock();
+#endif
+}
+
+static inline void sock_rps_record_flow(const struct sock *sk)
+{
+#ifdef CONFIG_RPS
+	if (static_branch_unlikely(&rfs_needed)) {
+		/* Reading sk->sk_rxhash might incur an expensive cache line
+		 * miss.
+		 *
+		 * TCP_ESTABLISHED does cover almost all states where RFS
+		 * might be useful, and is cheaper [1] than testing :
+		 *	IPv4: inet_sk(sk)->inet_daddr
+		 * 	IPv6: ipv6_addr_any(&sk->sk_v6_daddr)
+		 * OR	an additional socket flag
+		 * [1] : sk_state and sk_prot are in the same cache line.
+		 */
+		if (sk->sk_state == TCP_ESTABLISHED) {
+			/* This READ_ONCE() is paired with the WRITE_ONCE()
+			 * from sock_rps_save_rxhash() and sock_rps_reset_rxhash().
+			 */
+			sock_rps_record_flow_hash(READ_ONCE(sk->sk_rxhash));
+		}
+	}
+#endif
+}
+
+#endif /* _NET_RPS_H */

include/net/sock.h

@@ -1117,41 +1117,6 @@ static inline void sk_incoming_cpu_update(struct sock *sk)
 		WRITE_ONCE(sk->sk_incoming_cpu, cpu);
 }
 
-static inline void sock_rps_record_flow_hash(__u32 hash)
-{
-#ifdef CONFIG_RPS
-	struct rps_sock_flow_table *sock_flow_table;
-
-	rcu_read_lock();
-	sock_flow_table = rcu_dereference(rps_sock_flow_table);
-	rps_record_sock_flow(sock_flow_table, hash);
-	rcu_read_unlock();
-#endif
-}
-
-static inline void sock_rps_record_flow(const struct sock *sk)
-{
-#ifdef CONFIG_RPS
-	if (static_branch_unlikely(&rfs_needed)) {
-		/* Reading sk->sk_rxhash might incur an expensive cache line
-		 * miss.
-		 *
-		 * TCP_ESTABLISHED does cover almost all states where RFS
-		 * might be useful, and is cheaper [1] than testing :
-		 *	IPv4: inet_sk(sk)->inet_daddr
-		 * 	IPv6: ipv6_addr_any(&sk->sk_v6_daddr)
-		 * OR	an additional socket flag
-		 * [1] : sk_state and sk_prot are in the same cache line.
-		 */
-		if (sk->sk_state == TCP_ESTABLISHED) {
-			/* This READ_ONCE() is paired with the WRITE_ONCE()
-			 * from sock_rps_save_rxhash() and sock_rps_reset_rxhash().
-			 */
-			sock_rps_record_flow_hash(READ_ONCE(sk->sk_rxhash));
-		}
-	}
-#endif
-}
 
 static inline void sock_rps_save_rxhash(struct sock *sk,
 					const struct sk_buff *skb)

net/core/dev.c

@@ -155,6 +155,7 @@
 #include <net/netdev_rx_queue.h>
 #include <net/page_pool/types.h>
 #include <net/page_pool/helpers.h>
+#include <net/rps.h>
 
 #include "dev.h"
 #include "net-sysfs.h"

net/core/net-sysfs.c

@@ -24,6 +24,7 @@
 #include <linux/of_net.h>
 #include <linux/cpu.h>
 #include <net/netdev_rx_queue.h>
+#include <net/rps.h>
 
 #include "dev.h"
 #include "net-sysfs.h"

net/core/sysctl_net_core.c

@@ -24,6 +24,7 @@
 #include <net/busy_poll.h>
 #include <net/pkt_sched.h>
 #include <net/hotdata.h>
+#include <net/rps.h>
 
 #include "dev.h"
 

net/ipv4/af_inet.c

@@ -119,6 +119,7 @@
 #endif
 #include <net/l3mdev.h>
 #include <net/compat.h>
+#include <net/rps.h>
 
 #include <trace/events/sock.h>
 

net/ipv4/tcp.c

@@ -279,6 +279,7 @@
 #include <linux/uaccess.h>
 #include <asm/ioctls.h>
 #include <net/busy_poll.h>
+#include <net/rps.h>
 
 /* Track pending CMSGs. */
 enum {

net/ipv6/af_inet6.c

@@ -64,6 +64,7 @@
 #include <net/xfrm.h>
 #include <net/ioam6.h>
 #include <net/rawv6.h>
+#include <net/rps.h>
 
 #include <linux/uaccess.h>
 #include <linux/mroute6.h>

net/sctp/socket.c

@@ -67,6 +67,7 @@
 #include <net/sctp/sctp.h>
 #include <net/sctp/sm.h>
 #include <net/sctp/stream_sched.h>
+#include <net/rps.h>
 
 /* Forward declarations for internal helper functions. */
 static bool sctp_writeable(const struct sock *sk);