net/tls: reorganize struct tls_context

struct tls_context is slightly badly laid out.  If we reorder things
right we can save 16 bytes (320 -> 304) but also make all fast path
data fit into two cache lines (one read only and one read/write,
down from four cache lines).
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

include/net/tls.h

@@ -236,34 +236,32 @@ struct tls_prot_info {
 };
 
 struct tls_context {
+	/* read-only cache line */
 	struct tls_prot_info prot_info;
 
-	union tls_crypto_context crypto_send;
-	union tls_crypto_context crypto_recv;
+	u8 tx_conf:3;
+	u8 rx_conf:3;
 
-	struct list_head list;
-	struct net_device *netdev;
-	refcount_t refcount;
+	int (*push_pending_record)(struct sock *sk, int flags);
+	void (*sk_write_space)(struct sock *sk);
 
 	void *priv_ctx_tx;
 	void *priv_ctx_rx;
 
-	u8 tx_conf:3;
-	u8 rx_conf:3;
+	struct net_device *netdev;
 
+	/* rw cache line */
 	struct cipher_context tx;
 	struct cipher_context rx;
 
 	struct scatterlist *partially_sent_record;
 	u16 partially_sent_offset;
 
-	unsigned long flags;
 	bool in_tcp_sendpages;
 	bool pending_open_record_frags;
+	unsigned long flags;
 
-	int (*push_pending_record)(struct sock *sk, int flags);
-
-	void (*sk_write_space)(struct sock *sk);
+	/* cache cold stuff */
 	void (*sk_destruct)(struct sock *sk);
 	void (*sk_proto_close)(struct sock *sk, long timeout);
 
@@ -275,6 +273,12 @@ struct tls_context {
 			   int __user *optlen);
 	int  (*hash)(struct sock *sk);
 	void (*unhash)(struct sock *sk);
+
+	union tls_crypto_context crypto_send;
+	union tls_crypto_context crypto_recv;
+
+	struct list_head list;
+	refcount_t refcount;
 };
 
 enum tls_offload_ctx_dir {