[PATCH 0/2] netlink patches

[PATCH 0/2] netlink patches

[Stephen Hemminger]

[PATCH 1/2] netlink: af_netlink cleanup (v2)

[Stephen Hemminger]

[-- Attachment #1: netlink-deinline.patch --]
[-- Type: text/plain, Size: 4401 bytes --]

Don't inline functions that cover several lines, and do inline
the trivial ones. Also make some arguments const.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
---
v2 - cleanup indentation of functions after change

--- a/net/netlink/af_netlink.c	2011-12-21 15:03:32.740915909 -0800
+++ b/net/netlink/af_netlink.c	2011-12-21 15:07:44.807155562 -0800
@@ -98,7 +98,7 @@ static inline struct netlink_sock *nlk_s
 	return container_of(sk, struct netlink_sock, sk);
 }
 
-static inline int netlink_is_kernel(struct sock *sk)
+static inline int netlink_is_kernel(const struct sock *sk)
 {
 	return nlk_sk(sk)->flags & NETLINK_KERNEL_SOCKET;
 }
@@ -139,12 +139,12 @@ static atomic_t nl_table_users = ATOMIC_
 
 static ATOMIC_NOTIFIER_HEAD(netlink_chain);
 
-static u32 netlink_group_mask(u32 group)
+static inline u32 netlink_group_mask(u32 group)
 {
 	return group ? 1 << (group - 1) : 0;
 }
 
-static struct hlist_head *nl_pid_hashfn(struct nl_pid_hash *hash, u32 pid)
+static inline struct hlist_head *nl_pid_hashfn(struct nl_pid_hash *hash, u32 pid)
 {
 	return &hash->table[jhash_1word(pid, hash->rnd) & hash->mask];
 }
@@ -226,8 +226,7 @@ netlink_unlock_table(void)
 		wake_up(&nl_table_wait);
 }
 
-static inline struct sock *netlink_lookup(struct net *net, int protocol,
-					  u32 pid)
+static struct sock *netlink_lookup(struct net *net, int protocol, u32 pid)
 {
 	struct nl_pid_hash *hash = &nl_table[protocol].hash;
 	struct hlist_head *head;
@@ -248,7 +247,7 @@ found:
 	return sk;
 }
 
-static inline struct hlist_head *nl_pid_hash_zalloc(size_t size)
+static struct hlist_head *nl_pid_hash_zalloc(size_t size)
 {
 	if (size <= PAGE_SIZE)
 		return kzalloc(size, GFP_ATOMIC);
@@ -258,7 +257,7 @@ static inline struct hlist_head *nl_pid_
 					 get_order(size));
 }
 
-static inline void nl_pid_hash_free(struct hlist_head *table, size_t size)
+static void nl_pid_hash_free(struct hlist_head *table, size_t size)
 {
 	if (size <= PAGE_SIZE)
 		kfree(table);
@@ -578,7 +577,7 @@ retry:
 	return err;
 }
 
-static inline int netlink_capable(struct socket *sock, unsigned int flag)
+static inline int netlink_capable(const struct socket *sock, unsigned int flag)
 {
 	return (nl_table[sock->sk->sk_protocol].nl_nonroot & flag) ||
 	       capable(CAP_NET_ADMIN);
@@ -846,8 +845,7 @@ void netlink_detachskb(struct sock *sk,
 	sock_put(sk);
 }
 
-static inline struct sk_buff *netlink_trim(struct sk_buff *skb,
-					   gfp_t allocation)
+static struct sk_buff *netlink_trim(struct sk_buff *skb, gfp_t allocation)
 {
 	int delta;
 
@@ -871,7 +869,7 @@ static inline struct sk_buff *netlink_tr
 	return skb;
 }
 
-static inline void netlink_rcv_wake(struct sock *sk)
+static void netlink_rcv_wake(struct sock *sk)
 {
 	struct netlink_sock *nlk = nlk_sk(sk);
 
@@ -881,7 +879,7 @@ static inline void netlink_rcv_wake(stru
 		wake_up_interruptible(&nlk->wait);
 }
 
-static inline int netlink_unicast_kernel(struct sock *sk, struct sk_buff *skb)
+static int netlink_unicast_kernel(struct sock *sk, struct sk_buff *skb)
 {
 	int ret;
 	struct netlink_sock *nlk = nlk_sk(sk);
@@ -952,8 +950,7 @@ int netlink_has_listeners(struct sock *s
 }
 EXPORT_SYMBOL_GPL(netlink_has_listeners);
 
-static inline int netlink_broadcast_deliver(struct sock *sk,
-					    struct sk_buff *skb)
+static int netlink_broadcast_deliver(struct sock *sk, struct sk_buff *skb)
 {
 	struct netlink_sock *nlk = nlk_sk(sk);
 
@@ -982,7 +979,7 @@ struct netlink_broadcast_data {
 	void *tx_data;
 };
 
-static inline int do_one_broadcast(struct sock *sk,
+static int do_one_broadcast(struct sock *sk,
 				   struct netlink_broadcast_data *p)
 {
 	struct netlink_sock *nlk = nlk_sk(sk);
@@ -1110,8 +1107,7 @@ struct netlink_set_err_data {
 	int code;
 };
 
-static inline int do_one_set_err(struct sock *sk,
-				 struct netlink_set_err_data *p)
+static int do_one_set_err(struct sock *sk, struct netlink_set_err_data *p)
 {
 	struct netlink_sock *nlk = nlk_sk(sk);
 	int ret = 0;
--- a/net/netlink/genetlink.c	2011-12-21 15:03:32.724915740 -0800
+++ b/net/netlink/genetlink.c	2011-12-21 15:03:34.992939390 -0800
@@ -106,7 +106,7 @@ static struct genl_ops *genl_get_cmd(u8
 /* Of course we are going to have problems once we hit
  * 2^16 alive types, but that can only happen by year 2K
 */
-static inline u16 genl_generate_id(void)
+static u16 genl_generate_id(void)
 {
 	static u16 id_gen_idx = GENL_MIN_ID;
 	int i;

Re: [PATCH 1/2] netlink: af_netlink cleanup (v2)

[David Miller]

From: Stephen Hemminger <shemminger@vyatta.com>
Date: Thu, 22 Dec 2011 10:52:02 -0800

> Don't inline functions that cover several lines, and do inline
> the trivial ones. Also make some arguments const.
> 
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>

Applied.

[PATCH 2/2] netlink: wake up netlink listeners sooner (v2)

[Stephen Hemminger]

[-- Attachment #1: netlink-congest.patch --]
[-- Type: text/plain, Size: 712 bytes --]

This patch changes it to yield sooner at halfway instead. Still not a cure-all
for listener overrun if listner is slow, but works much reliably.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>

---
v2 - use shift to avoid expensive integer division

--- a/net/netlink/af_netlink.c	2011-12-21 15:15:09.446373421 -0800
+++ b/net/netlink/af_netlink.c	2011-12-21 15:22:35.632062577 -0800
@@ -959,7 +959,7 @@ static int netlink_broadcast_deliver(str
 		skb_set_owner_r(skb, sk);
 		skb_queue_tail(&sk->sk_receive_queue, skb);
 		sk->sk_data_ready(sk, skb->len);
-		return atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf;
+		return atomic_read(&sk->sk_rmem_alloc) > (sk->sk_rcvbuf >> 1);
 	}
 	return -1;
 }

Re: [PATCH 2/2] netlink: wake up netlink listeners sooner (v2)

[David Miller]

From: Stephen Hemminger <shemminger@vyatta.com>
Date: Thu, 22 Dec 2011 10:52:03 -0800

> This patch changes it to yield sooner at halfway instead. Still not a cure-all
> for listener overrun if listner is slow, but works much reliably.
> 
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>

Applied.