From ec733b15a3ef0b5759141a177f8044a2f40c41e7 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 18 Mar 2010 20:36:06 +0000 Subject: net: snmp mib cleanup There is no point to align or pad mibs to cache lines, they are per cpu allocated with a 8 bytes alignment anyway. This wastes space for no gain. This patch removes __SNMP_MIB_ALIGN__ Since SNMP mibs contain "unsigned long" fields only, we can relax the allocation alignment from "unsigned long long" to "unsigned long" Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/dccp/dccp.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/dccp') diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h index 5ef32c2f0d6a..53f8e12d0c10 100644 --- a/net/dccp/dccp.h +++ b/net/dccp/dccp.h @@ -189,7 +189,7 @@ enum { #define DCCP_MIB_MAX __DCCP_MIB_MAX struct dccp_mib { unsigned long mibs[DCCP_MIB_MAX]; -} __SNMP_MIB_ALIGN__; +}; DECLARE_SNMP_STAT(struct dccp_mib, dccp_statistics); #define DCCP_INC_STATS(field) SNMP_INC_STATS(dccp_statistics, field) -- cgit v1.2.1 From b138338056fc423c61a583d45f8aa64cfad87131 Mon Sep 17 00:00:00 2001 From: Frans Pop Date: Wed, 24 Mar 2010 07:57:28 +0000 Subject: net: remove trailing space in messages Signed-off-by: Frans Pop Signed-off-by: David S. Miller --- net/dccp/ccids/ccid3.c | 2 +- net/dccp/input.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'net/dccp') diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c index bcd7632299f5..d3235899c7e3 100644 --- a/net/dccp/ccids/ccid3.c +++ b/net/dccp/ccids/ccid3.c @@ -208,7 +208,7 @@ static void ccid3_hc_tx_no_feedback_timer(unsigned long data) goto restart_timer; } - ccid3_pr_debug("%s(%p, state=%s) - entry \n", dccp_role(sk), sk, + ccid3_pr_debug("%s(%p, state=%s) - entry\n", dccp_role(sk), sk, ccid3_tx_state_name(hc->tx_state)); if (hc->tx_state == TFRC_SSTATE_FBACK) diff --git a/net/dccp/input.c b/net/dccp/input.c index 7648f316310f..5daa4bdfdb10 100644 --- a/net/dccp/input.c +++ b/net/dccp/input.c @@ -414,7 +414,7 @@ static int dccp_rcv_request_sent_state_process(struct sock *sk, if (!between48(DCCP_SKB_CB(skb)->dccpd_ack_seq, dp->dccps_awl, dp->dccps_awh)) { dccp_pr_debug("invalid ackno: S.AWL=%llu, " - "P.ackno=%llu, S.AWH=%llu \n", + "P.ackno=%llu, S.AWH=%llu\n", (unsigned long long)dp->dccps_awl, (unsigned long long)DCCP_SKB_CB(skb)->dccpd_ack_seq, (unsigned long long)dp->dccps_awh); -- cgit v1.2.1 From bb29624614c2afe2873ee8ee97cf09df42701694 Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Sun, 11 Apr 2010 02:15:55 +0000 Subject: inet: Remove unused send_check length argument inet: Remove unused send_check length argument This patch removes the unused length argument from the send_check function in struct inet_connection_sock_af_ops. Signed-off-by: Herbert Xu Tested-by: Yinghai Signed-off-by: David S. Miller --- net/dccp/dccp.h | 2 +- net/dccp/ipv4.c | 2 +- net/dccp/ipv6.c | 3 +-- net/dccp/output.c | 2 +- 4 files changed, 4 insertions(+), 5 deletions(-) (limited to 'net/dccp') diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h index 53f8e12d0c10..a10a61a1ded2 100644 --- a/net/dccp/dccp.h +++ b/net/dccp/dccp.h @@ -223,7 +223,7 @@ static inline void dccp_csum_outgoing(struct sk_buff *skb) skb->csum = skb_checksum(skb, 0, (cov > skb->len)? skb->len : cov, 0); } -extern void dccp_v4_send_check(struct sock *sk, int len, struct sk_buff *skb); +extern void dccp_v4_send_check(struct sock *sk, struct sk_buff *skb); extern int dccp_retransmit_skb(struct sock *sk); diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index 52ffa1cde15a..d9b11ef8694c 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -349,7 +349,7 @@ static inline __sum16 dccp_v4_csum_finish(struct sk_buff *skb, return csum_tcpudp_magic(src, dst, skb->len, IPPROTO_DCCP, skb->csum); } -void dccp_v4_send_check(struct sock *sk, int unused, struct sk_buff *skb) +void dccp_v4_send_check(struct sock *sk, struct sk_buff *skb) { const struct inet_sock *inet = inet_sk(sk); struct dccp_hdr *dh = dccp_hdr(skb); diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c index 3b11e41a2929..ab1ab95946df 100644 --- a/net/dccp/ipv6.c +++ b/net/dccp/ipv6.c @@ -60,8 +60,7 @@ static inline __sum16 dccp_v6_csum_finish(struct sk_buff *skb, return csum_ipv6_magic(saddr, daddr, skb->len, IPPROTO_DCCP, skb->csum); } -static inline void dccp_v6_send_check(struct sock *sk, int unused_value, - struct sk_buff *skb) +static inline void dccp_v6_send_check(struct sock *sk, struct sk_buff *skb) { struct ipv6_pinfo *np = inet6_sk(sk); struct dccp_hdr *dh = dccp_hdr(skb); diff --git a/net/dccp/output.c b/net/dccp/output.c index fc3f436440b4..b8d98e3c052a 100644 --- a/net/dccp/output.c +++ b/net/dccp/output.c @@ -129,7 +129,7 @@ static int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb) break; } - icsk->icsk_af_ops->send_check(sk, 0, skb); + icsk->icsk_af_ops->send_check(sk, skb); if (set_ack) dccp_event_ack_sent(sk); -- cgit v1.2.1 From b6c6712a42ca3f9fa7f4a3d7c40e3a9dd1fd9e03 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 8 Apr 2010 23:03:29 +0000 Subject: net: sk_dst_cache RCUification With latest CONFIG_PROVE_RCU stuff, I felt more comfortable to make this work. sk->sk_dst_cache is currently protected by a rwlock (sk_dst_lock) This rwlock is readlocked for a very small amount of time, and dst entries are already freed after RCU grace period. This calls for RCU again :) This patch converts sk_dst_lock to a spinlock, and use RCU for readers. __sk_dst_get() is supposed to be called with rcu_read_lock() or if socket locked by user, so use appropriate rcu_dereference_check() condition (rcu_read_lock_held() || sock_owned_by_user(sk)) This patch avoids two atomic ops per tx packet on UDP connected sockets, for example, and permits sk_dst_lock to be much less dirtied. Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/dccp/timer.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net/dccp') diff --git a/net/dccp/timer.c b/net/dccp/timer.c index bbfeb5eae46a..1a9aa05d4dc4 100644 --- a/net/dccp/timer.c +++ b/net/dccp/timer.c @@ -38,7 +38,7 @@ static int dccp_write_timeout(struct sock *sk) if (sk->sk_state == DCCP_REQUESTING || sk->sk_state == DCCP_PARTOPEN) { if (icsk->icsk_retransmits != 0) - dst_negative_advice(&sk->sk_dst_cache, sk); + dst_negative_advice(sk); retry_until = icsk->icsk_syn_retries ? : sysctl_dccp_request_retries; } else { @@ -63,7 +63,7 @@ static int dccp_write_timeout(struct sock *sk) Golden words :-). */ - dst_negative_advice(&sk->sk_dst_cache, sk); + dst_negative_advice(sk); } retry_until = sysctl_dccp_retries2; -- cgit v1.2.1 From 4e15ed4d930297c127d280ca1d0c785be870def4 Mon Sep 17 00:00:00 2001 From: Shan Wei Date: Thu, 15 Apr 2010 16:43:08 +0000 Subject: net: replace ipfragok with skb->local_df As Herbert Xu said: we should be able to simply replace ipfragok with skb->local_df. commit f88037(sctp: Drop ipfargok in sctp_xmit function) has droped ipfragok and set local_df value properly. The patch kills the ipfragok parameter of .queue_xmit(). Signed-off-by: Shan Wei Signed-off-by: David S. Miller --- net/dccp/ipv6.c | 4 ++-- net/dccp/output.c | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'net/dccp') diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c index ab1ab95946df..091698899594 100644 --- a/net/dccp/ipv6.c +++ b/net/dccp/ipv6.c @@ -292,7 +292,7 @@ static int dccp_v6_send_response(struct sock *sk, struct request_sock *req, &ireq6->loc_addr, &ireq6->rmt_addr); ipv6_addr_copy(&fl.fl6_dst, &ireq6->rmt_addr); - err = ip6_xmit(sk, skb, &fl, opt, 0); + err = ip6_xmit(sk, skb, &fl, opt); err = net_xmit_eval(err); } @@ -347,7 +347,7 @@ static void dccp_v6_ctl_send_reset(struct sock *sk, struct sk_buff *rxskb) if (!ip6_dst_lookup(ctl_sk, &dst, &fl)) { if (xfrm_lookup(net, &dst, &fl, NULL, 0) >= 0) { skb_dst_set(skb, dst); - ip6_xmit(ctl_sk, skb, &fl, NULL, 0); + ip6_xmit(ctl_sk, skb, &fl, NULL); DCCP_INC_STATS_BH(DCCP_MIB_OUTSEGS); DCCP_INC_STATS_BH(DCCP_MIB_OUTRSTS); return; diff --git a/net/dccp/output.c b/net/dccp/output.c index b8d98e3c052a..e98b65e9569f 100644 --- a/net/dccp/output.c +++ b/net/dccp/output.c @@ -136,7 +136,7 @@ static int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb) DCCP_INC_STATS(DCCP_MIB_OUTSEGS); - err = icsk->icsk_af_ops->queue_xmit(skb, 0); + err = icsk->icsk_af_ops->queue_xmit(skb); return net_xmit_eval(err); } return -ENOBUFS; -- cgit v1.2.1 From aa395145165cb06a0d0885221bbe0ce4a564391d Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 20 Apr 2010 13:03:51 +0000 Subject: net: sk_sleep() helper Define a new function to return the waitqueue of a "struct sock". static inline wait_queue_head_t *sk_sleep(struct sock *sk) { return sk->sk_sleep; } Change all read occurrences of sk_sleep by a call to this function. Needed for a future RCU conversion. sk_sleep wont be a field directly available. Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/dccp/output.c | 6 +++--- net/dccp/proto.c | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'net/dccp') diff --git a/net/dccp/output.c b/net/dccp/output.c index e98b65e9569f..2d3dcb39851f 100644 --- a/net/dccp/output.c +++ b/net/dccp/output.c @@ -198,7 +198,7 @@ void dccp_write_space(struct sock *sk) read_lock(&sk->sk_callback_lock); if (sk_has_sleeper(sk)) - wake_up_interruptible(sk->sk_sleep); + wake_up_interruptible(sk_sleep(sk)); /* Should agree with poll, otherwise some programs break */ if (sock_writeable(sk)) sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT); @@ -225,7 +225,7 @@ static int dccp_wait_for_ccid(struct sock *sk, struct sk_buff *skb, int delay) dccp_pr_debug("delayed send by %d msec\n", delay); jiffdelay = msecs_to_jiffies(delay); - prepare_to_wait(sk->sk_sleep, &wait, TASK_INTERRUPTIBLE); + prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE); sk->sk_write_pending++; release_sock(sk); @@ -241,7 +241,7 @@ static int dccp_wait_for_ccid(struct sock *sk, struct sk_buff *skb, int delay) rc = ccid_hc_tx_send_packet(dp->dccps_hc_tx_ccid, sk, skb); } while ((delay = rc) > 0); out: - finish_wait(sk->sk_sleep, &wait); + finish_wait(sk_sleep(sk), &wait); return rc; do_error: diff --git a/net/dccp/proto.c b/net/dccp/proto.c index a0e38d8018f5..b03ecf6b2bb0 100644 --- a/net/dccp/proto.c +++ b/net/dccp/proto.c @@ -312,7 +312,7 @@ unsigned int dccp_poll(struct file *file, struct socket *sock, unsigned int mask; struct sock *sk = sock->sk; - sock_poll_wait(file, sk->sk_sleep, wait); + sock_poll_wait(file, sk_sleep(sk), wait); if (sk->sk_state == DCCP_LISTEN) return inet_csk_listen_poll(sk); -- cgit v1.2.1 From 43815482370c510c569fd18edb57afcb0fa8cab6 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 29 Apr 2010 11:01:49 +0000 Subject: net: sock_def_readable() and friends RCU conversion sk_callback_lock rwlock actually protects sk->sk_sleep pointer, so we need two atomic operations (and associated dirtying) per incoming packet. RCU conversion is pretty much needed : 1) Add a new structure, called "struct socket_wq" to hold all fields that will need rcu_read_lock() protection (currently: a wait_queue_head_t and a struct fasync_struct pointer). [Future patch will add a list anchor for wakeup coalescing] 2) Attach one of such structure to each "struct socket" created in sock_alloc_inode(). 3) Respect RCU grace period when freeing a "struct socket_wq" 4) Change sk_sleep pointer in "struct sock" by sk_wq, pointer to "struct socket_wq" 5) Change sk_sleep() function to use new sk->sk_wq instead of sk->sk_sleep 6) Change sk_has_sleeper() to wq_has_sleeper() that must be used inside a rcu_read_lock() section. 7) Change all sk_has_sleeper() callers to : - Use rcu_read_lock() instead of read_lock(&sk->sk_callback_lock) - Use wq_has_sleeper() to eventually wakeup tasks. - Use rcu_read_unlock() instead of read_unlock(&sk->sk_callback_lock) 8) sock_wake_async() is modified to use rcu protection as well. 9) Exceptions : macvtap, drivers/net/tun.c, af_unix use integrated "struct socket_wq" instead of dynamically allocated ones. They dont need rcu freeing. Some cleanups or followups are probably needed, (possible sk_callback_lock conversion to a spinlock for example...). Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/dccp/output.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'net/dccp') diff --git a/net/dccp/output.c b/net/dccp/output.c index 2d3dcb39851f..aadbdb58758b 100644 --- a/net/dccp/output.c +++ b/net/dccp/output.c @@ -195,15 +195,17 @@ EXPORT_SYMBOL_GPL(dccp_sync_mss); void dccp_write_space(struct sock *sk) { - read_lock(&sk->sk_callback_lock); + struct socket_wq *wq; - if (sk_has_sleeper(sk)) - wake_up_interruptible(sk_sleep(sk)); + rcu_read_lock(); + wq = rcu_dereference(sk->sk_wq); + if (wq_has_sleeper(wq)) + wake_up_interruptible(&wq->wait); /* Should agree with poll, otherwise some programs break */ if (sock_writeable(sk)) sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT); - read_unlock(&sk->sk_callback_lock); + rcu_read_unlock(); } /** -- cgit v1.2.1