From 1d67e6501b8dba54ef8dcabebe2ad049b8ad0d67 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 22 Sep 2005 03:27:56 -0300 Subject: [LLC]: Make llc_frame_alloc take a net_device as an argument So as to set the newly created sk_buff ->dev member with it, that way we stop using dev_base->next, that is the wrong thing to do, as there may well be several interfaces being used with LLC. This was not such a big problem after all as most of the users of llc_alloc_frame were setting the correct dev, but this way code is reduced. This also fixes another bug in llc_station_ac_send_null_dsap_xid_c, that was not setting the skb->dev field. Signed-off-by: Jochen Friedrich Signed-off-by: Arnaldo Carvalho de Melo --- include/net/llc_sap.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'include/net') diff --git a/include/net/llc_sap.h b/include/net/llc_sap.h index 353baaa627f..6dd4cdccca4 100644 --- a/include/net/llc_sap.h +++ b/include/net/llc_sap.h @@ -13,10 +13,11 @@ */ struct llc_sap; struct sk_buff; +struct net_device; extern void llc_sap_rtn_pdu(struct llc_sap *sap, struct sk_buff *skb); extern void llc_save_primitive(struct sk_buff* skb, unsigned char prim); -extern struct sk_buff *llc_alloc_frame(void); +extern struct sk_buff *llc_alloc_frame(struct net_device *dev); extern void llc_build_and_send_test_pkt(struct llc_sap *sap, struct sk_buff *skb, -- cgit v1.2.3-70-g09d2 From 590232a7150674b2036291eaefce085f3f9659c8 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 22 Sep 2005 04:30:44 -0300 Subject: [LLC]: Add sysctl support for the LLC timeouts Signed-off-by: Jochen Friedrich Signed-off-by: Arnaldo Carvalho de Melo --- include/linux/sysctl.h | 26 ++++++++- include/net/llc.h | 7 +++ include/net/llc_conn.h | 10 ++-- net/llc/Makefile | 1 + net/llc/af_llc.c | 47 +++++++++++----- net/llc/llc_c_ac.c | 12 ++--- net/llc/llc_conn.c | 13 +++-- net/llc/llc_station.c | 11 ++-- net/llc/sysctl_net_llc.c | 136 +++++++++++++++++++++++++++++++++++++++++++++++ 9 files changed, 232 insertions(+), 31 deletions(-) create mode 100644 net/llc/sysctl_net_llc.c (limited to 'include/net') diff --git a/include/linux/sysctl.h b/include/linux/sysctl.h index 3a29a9f9b45..fc8e367f671 100644 --- a/include/linux/sysctl.h +++ b/include/linux/sysctl.h @@ -202,7 +202,8 @@ enum NET_TR=14, NET_DECNET=15, NET_ECONET=16, - NET_SCTP=17, + NET_SCTP=17, + NET_LLC=18, }; /* /proc/sys/kernel/random */ @@ -522,6 +523,29 @@ enum { NET_IPX_FORWARDING=2 }; +/* /proc/sys/net/llc */ +enum { + NET_LLC2=1, + NET_LLC_STATION=2, +}; + +/* /proc/sys/net/llc/llc2 */ +enum { + NET_LLC2_TIMEOUT=1, +}; + +/* /proc/sys/net/llc/station */ +enum { + NET_LLC_STATION_ACK_TIMEOUT=1, +}; + +/* /proc/sys/net/llc/llc2/timeout */ +enum { + NET_LLC2_ACK_TIMEOUT=1, + NET_LLC2_P_TIMEOUT=2, + NET_LLC2_REJ_TIMEOUT=3, + NET_LLC2_BUSY_TIMEOUT=4, +}; /* /proc/sys/net/appletalk */ enum { diff --git a/include/net/llc.h b/include/net/llc.h index 71769a5aeef..8b8e2be289b 100644 --- a/include/net/llc.h +++ b/include/net/llc.h @@ -98,4 +98,11 @@ extern void llc_proc_exit(void); #define llc_proc_init() (0) #define llc_proc_exit() do { } while(0) #endif /* CONFIG_PROC_FS */ +#ifdef CONFIG_SYSCTL +extern int llc_sysctl_init(void); +extern void llc_sysctl_exit(void); +#else +#define llc_sysctl_init() (0) +#define llc_sysctl_exit() do { } while(0) +#endif /* CONFIG_SYSCTL */ #endif /* LLC_H */ diff --git a/include/net/llc_conn.h b/include/net/llc_conn.h index 8ad3bc2c23d..8a8ff481013 100644 --- a/include/net/llc_conn.h +++ b/include/net/llc_conn.h @@ -19,14 +19,14 @@ #define LLC_EVENT 1 #define LLC_PACKET 2 -#define LLC_P_TIME 2 -#define LLC_ACK_TIME 1 -#define LLC_REJ_TIME 3 -#define LLC_BUSY_TIME 3 +#define LLC2_P_TIME 2 +#define LLC2_ACK_TIME 1 +#define LLC2_REJ_TIME 3 +#define LLC2_BUSY_TIME 3 struct llc_timer { struct timer_list timer; - u16 expire; /* timer expire time */ + unsigned long expire; /* timer expire time */ }; struct llc_sock { diff --git a/net/llc/Makefile b/net/llc/Makefile index 5ebd4ed2bd4..4e260cff3c5 100644 --- a/net/llc/Makefile +++ b/net/llc/Makefile @@ -22,3 +22,4 @@ llc2-y := llc_if.o llc_c_ev.o llc_c_ac.o llc_conn.o llc_c_st.o llc_pdu.o \ llc_sap.o llc_s_ac.o llc_s_ev.o llc_s_st.o af_llc.o llc_station.o llc2-$(CONFIG_PROC_FS) += llc_proc.o +llc2-$(CONFIG_SYSCTL) += sysctl_net_llc.o diff --git a/net/llc/af_llc.c b/net/llc/af_llc.c index 95444f22751..ef125345a2d 100644 --- a/net/llc/af_llc.c +++ b/net/llc/af_llc.c @@ -877,22 +877,22 @@ static int llc_ui_setsockopt(struct socket *sock, int level, int optname, case LLC_OPT_ACK_TMR_EXP: if (opt > LLC_OPT_MAX_ACK_TMR_EXP) goto out; - llc->ack_timer.expire = opt; + llc->ack_timer.expire = opt * HZ; break; case LLC_OPT_P_TMR_EXP: if (opt > LLC_OPT_MAX_P_TMR_EXP) goto out; - llc->pf_cycle_timer.expire = opt; + llc->pf_cycle_timer.expire = opt * HZ; break; case LLC_OPT_REJ_TMR_EXP: if (opt > LLC_OPT_MAX_REJ_TMR_EXP) goto out; - llc->rej_sent_timer.expire = opt; + llc->rej_sent_timer.expire = opt * HZ; break; case LLC_OPT_BUSY_TMR_EXP: if (opt > LLC_OPT_MAX_BUSY_TMR_EXP) goto out; - llc->busy_state_timer.expire = opt; + llc->busy_state_timer.expire = opt * HZ; break; case LLC_OPT_TX_WIN: if (opt > LLC_OPT_MAX_WIN) @@ -942,17 +942,17 @@ static int llc_ui_getsockopt(struct socket *sock, int level, int optname, goto out; switch (optname) { case LLC_OPT_RETRY: - val = llc->n2; break; + val = llc->n2; break; case LLC_OPT_SIZE: - val = llc->n1; break; + val = llc->n1; break; case LLC_OPT_ACK_TMR_EXP: - val = llc->ack_timer.expire; break; + val = llc->ack_timer.expire / HZ; break; case LLC_OPT_P_TMR_EXP: - val = llc->pf_cycle_timer.expire; break; + val = llc->pf_cycle_timer.expire / HZ; break; case LLC_OPT_REJ_TMR_EXP: - val = llc->rej_sent_timer.expire; break; + val = llc->rej_sent_timer.expire / HZ; break; case LLC_OPT_BUSY_TMR_EXP: - val = llc->busy_state_timer.expire; break; + val = llc->busy_state_timer.expire / HZ; break; case LLC_OPT_TX_WIN: val = llc->k; break; case LLC_OPT_RX_WIN: @@ -999,6 +999,13 @@ static struct proto_ops llc_ui_ops = { extern void llc_sap_handler(struct llc_sap *sap, struct sk_buff *skb); extern void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb); +static char llc_proc_err_msg[] __initdata = + KERN_CRIT "LLC: Unable to register the proc_fs entries\n"; +static char llc_sysctl_err_msg[] __initdata = + KERN_CRIT "LLC: Unable to register the sysctl entries\n"; +static char llc_sock_err_msg[] __initdata = + KERN_CRIT "LLC: Unable to register the network family\n"; + static int __init llc2_init(void) { int rc = proto_register(&llc_proto, 0); @@ -1010,13 +1017,28 @@ static int __init llc2_init(void) llc_station_init(); llc_ui_sap_last_autoport = LLC_SAP_DYN_START; rc = llc_proc_init(); - if (rc != 0) + if (rc != 0) { + printk(llc_proc_err_msg); goto out_unregister_llc_proto; - sock_register(&llc_ui_family_ops); + } + rc = llc_sysctl_init(); + if (rc) { + printk(llc_sysctl_err_msg); + goto out_proc; + } + rc = sock_register(&llc_ui_family_ops); + if (rc) { + printk(llc_sock_err_msg); + goto out_sysctl; + } llc_add_pack(LLC_DEST_SAP, llc_sap_handler); llc_add_pack(LLC_DEST_CONN, llc_conn_handler); out: return rc; +out_sysctl: + llc_sysctl_exit(); +out_proc: + llc_proc_exit(); out_unregister_llc_proto: proto_unregister(&llc_proto); goto out; @@ -1029,6 +1051,7 @@ static void __exit llc2_exit(void) llc_remove_pack(LLC_DEST_CONN); sock_unregister(PF_LLC); llc_proc_exit(); + llc_sysctl_exit(); proto_unregister(&llc_proto); } diff --git a/net/llc/llc_c_ac.c b/net/llc/llc_c_ac.c index 50c827e13a9..a4daa91003d 100644 --- a/net/llc/llc_c_ac.c +++ b/net/llc/llc_c_ac.c @@ -620,7 +620,7 @@ int llc_conn_ac_set_remote_busy(struct sock *sk, struct sk_buff *skb) if (!llc->remote_busy_flag) { llc->remote_busy_flag = 1; mod_timer(&llc->busy_state_timer.timer, - jiffies + llc->busy_state_timer.expire * HZ); + jiffies + llc->busy_state_timer.expire); } return 0; } @@ -853,7 +853,7 @@ int llc_conn_ac_start_p_timer(struct sock *sk, struct sk_buff *skb) llc_conn_set_p_flag(sk, 1); mod_timer(&llc->pf_cycle_timer.timer, - jiffies + llc->pf_cycle_timer.expire * HZ); + jiffies + llc->pf_cycle_timer.expire); return 0; } @@ -1131,7 +1131,7 @@ int llc_conn_ac_start_ack_timer(struct sock *sk, struct sk_buff *skb) { struct llc_sock *llc = llc_sk(sk); - mod_timer(&llc->ack_timer.timer, jiffies + llc->ack_timer.expire * HZ); + mod_timer(&llc->ack_timer.timer, jiffies + llc->ack_timer.expire); return 0; } @@ -1140,7 +1140,7 @@ int llc_conn_ac_start_rej_timer(struct sock *sk, struct sk_buff *skb) struct llc_sock *llc = llc_sk(sk); mod_timer(&llc->rej_sent_timer.timer, - jiffies + llc->rej_sent_timer.expire * HZ); + jiffies + llc->rej_sent_timer.expire); return 0; } @@ -1151,7 +1151,7 @@ int llc_conn_ac_start_ack_tmr_if_not_running(struct sock *sk, if (!timer_pending(&llc->ack_timer.timer)) mod_timer(&llc->ack_timer.timer, - jiffies + llc->ack_timer.expire * HZ); + jiffies + llc->ack_timer.expire); return 0; } @@ -1199,7 +1199,7 @@ int llc_conn_ac_upd_nr_received(struct sock *sk, struct sk_buff *skb) } if (unacked) mod_timer(&llc->ack_timer.timer, - jiffies + llc->ack_timer.expire * HZ); + jiffies + llc->ack_timer.expire); } else if (llc->failed_data_req) { u8 f_bit; diff --git a/net/llc/llc_conn.c b/net/llc/llc_conn.c index ce7b893ed1a..d3783f8ee48 100644 --- a/net/llc/llc_conn.c +++ b/net/llc/llc_conn.c @@ -40,6 +40,11 @@ static struct llc_conn_state_trans *llc_qualify_conn_ev(struct sock *sk, /* Offset table on connection states transition diagram */ static int llc_offset_table[NBR_CONN_STATES][NBR_CONN_EV]; +int sysctl_llc2_ack_timeout = LLC2_ACK_TIME * HZ; +int sysctl_llc2_p_timeout = LLC2_P_TIME * HZ; +int sysctl_llc2_rej_timeout = LLC2_REJ_TIME * HZ; +int sysctl_llc2_busy_timeout = LLC2_BUSY_TIME * HZ; + /** * llc_conn_state_process - sends event to connection state machine * @sk: connection @@ -799,22 +804,22 @@ static void llc_sk_init(struct sock* sk) llc->dec_step = llc->connect_step = 1; init_timer(&llc->ack_timer.timer); - llc->ack_timer.expire = LLC_ACK_TIME; + llc->ack_timer.expire = sysctl_llc2_ack_timeout; llc->ack_timer.timer.data = (unsigned long)sk; llc->ack_timer.timer.function = llc_conn_ack_tmr_cb; init_timer(&llc->pf_cycle_timer.timer); - llc->pf_cycle_timer.expire = LLC_P_TIME; + llc->pf_cycle_timer.expire = sysctl_llc2_p_timeout; llc->pf_cycle_timer.timer.data = (unsigned long)sk; llc->pf_cycle_timer.timer.function = llc_conn_pf_cycle_tmr_cb; init_timer(&llc->rej_sent_timer.timer); - llc->rej_sent_timer.expire = LLC_REJ_TIME; + llc->rej_sent_timer.expire = sysctl_llc2_rej_timeout; llc->rej_sent_timer.timer.data = (unsigned long)sk; llc->rej_sent_timer.timer.function = llc_conn_rej_tmr_cb; init_timer(&llc->busy_state_timer.timer); - llc->busy_state_timer.expire = LLC_BUSY_TIME; + llc->busy_state_timer.expire = sysctl_llc2_busy_timeout; llc->busy_state_timer.timer.data = (unsigned long)sk; llc->busy_state_timer.timer.function = llc_conn_busy_tmr_cb; diff --git a/net/llc/llc_station.c b/net/llc/llc_station.c index 85a7ac27614..2d764b0382c 100644 --- a/net/llc/llc_station.c +++ b/net/llc/llc_station.c @@ -50,6 +50,10 @@ struct llc_station { struct sk_buff_head mac_pdu_q; }; +#define LLC_STATION_ACK_TIME (3 * HZ) + +int sysctl_llc_station_ack_timeout = LLC_STATION_ACK_TIME; + /* Types of events (possible values in 'ev->type') */ #define LLC_STATION_EV_TYPE_SIMPLE 1 #define LLC_STATION_EV_TYPE_CONDITION 2 @@ -218,7 +222,8 @@ static void llc_station_send_pdu(struct sk_buff *skb) static int llc_station_ac_start_ack_timer(struct sk_buff *skb) { - mod_timer(&llc_main_station.ack_timer, jiffies + LLC_ACK_TIME * HZ); + mod_timer(&llc_main_station.ack_timer, + jiffies + sysctl_llc_station_ack_timeout); return 0; } @@ -687,7 +692,8 @@ int __init llc_station_init(void) init_timer(&llc_main_station.ack_timer); llc_main_station.ack_timer.data = (unsigned long)&llc_main_station; llc_main_station.ack_timer.function = llc_station_ack_tmr_cb; - + llc_main_station.ack_timer.expires = jiffies + + sysctl_llc_station_ack_timeout; skb = alloc_skb(0, GFP_ATOMIC); if (!skb) goto out; @@ -695,7 +701,6 @@ int __init llc_station_init(void) llc_set_station_handler(llc_station_rcv); ev = llc_station_ev(skb); memset(ev, 0, sizeof(*ev)); - llc_main_station.ack_timer.expires = jiffies + 3 * HZ; llc_main_station.maximum_retry = 1; llc_main_station.state = LLC_STATION_STATE_DOWN; ev->type = LLC_STATION_EV_TYPE_SIMPLE; diff --git a/net/llc/sysctl_net_llc.c b/net/llc/sysctl_net_llc.c new file mode 100644 index 00000000000..4d99d2f27a2 --- /dev/null +++ b/net/llc/sysctl_net_llc.c @@ -0,0 +1,136 @@ +/* + * sysctl_net_llc.c: sysctl interface to LLC net subsystem. + * + * Arnaldo Carvalho de Melo + */ + +#include +#include +#include +#include + +#ifndef CONFIG_SYSCTL +#error This file should not be compiled without CONFIG_SYSCTL defined +#endif + +extern int sysctl_llc2_ack_timeout; +extern int sysctl_llc2_busy_timeout; +extern int sysctl_llc2_p_timeout; +extern int sysctl_llc2_rej_timeout; +extern int sysctl_llc_station_ack_timeout; + +static struct ctl_table llc2_timeout_table[] = { + { + .ctl_name = NET_LLC2_ACK_TIMEOUT, + .procname = "ack", + .data = &sysctl_llc2_ack_timeout, + .maxlen = sizeof(long), + .mode = 0644, + .proc_handler = &proc_dointvec_jiffies, + .strategy = &sysctl_jiffies, + }, + { + .ctl_name = NET_LLC2_BUSY_TIMEOUT, + .procname = "busy", + .data = &sysctl_llc2_busy_timeout, + .maxlen = sizeof(long), + .mode = 0644, + .proc_handler = &proc_dointvec_jiffies, + .strategy = &sysctl_jiffies, + }, + { + .ctl_name = NET_LLC2_P_TIMEOUT, + .procname = "p", + .data = &sysctl_llc2_p_timeout, + .maxlen = sizeof(long), + .mode = 0644, + .proc_handler = &proc_dointvec_jiffies, + .strategy = &sysctl_jiffies, + }, + { + .ctl_name = NET_LLC2_REJ_TIMEOUT, + .procname = "rej", + .data = &sysctl_llc2_rej_timeout, + .maxlen = sizeof(long), + .mode = 0644, + .proc_handler = &proc_dointvec_jiffies, + .strategy = &sysctl_jiffies, + }, + { 0 }, +}; + +static struct ctl_table llc_station_table[] = { + { + .ctl_name = NET_LLC_STATION_ACK_TIMEOUT, + .procname = "ack_timeout", + .data = &sysctl_llc_station_ack_timeout, + .maxlen = sizeof(long), + .mode = 0644, + .proc_handler = &proc_dointvec_jiffies, + .strategy = &sysctl_jiffies, + }, + { 0 }, +}; + +static struct ctl_table llc2_dir_timeout_table[] = { + { + .ctl_name = NET_LLC2, + .procname = "timeout", + .mode = 0555, + .child = llc2_timeout_table, + }, + { 0 }, +}; + +static struct ctl_table llc_table[] = { + { + .ctl_name = NET_LLC2, + .procname = "llc2", + .mode = 0555, + .child = llc2_dir_timeout_table, + }, + { + .ctl_name = NET_LLC_STATION, + .procname = "station", + .mode = 0555, + .child = llc_station_table, + }, + { 0 }, +}; + +static struct ctl_table llc_dir_table[] = { + { + .ctl_name = NET_LLC, + .procname = "llc", + .mode = 0555, + .child = llc_table, + }, + { 0 }, +}; + +static struct ctl_table llc_root_table[] = { + { + .ctl_name = CTL_NET, + .procname = "net", + .mode = 0555, + .child = llc_dir_table, + }, + { 0 }, +}; + +static struct ctl_table_header *llc_table_header; + +int __init llc_sysctl_init(void) +{ + llc_table_header = register_sysctl_table(llc_root_table, 1); + + return llc_table_header ? 0 : -ENOMEM; +} + +void llc_sysctl_exit(void) +{ + if (llc_table_header) { + unregister_sysctl_table(llc_table_header); + llc_table_header = NULL; + } +} -- cgit v1.2.3-70-g09d2 From 04e4223f44b89e50f275cb6b95a58ebe2c4909be Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 22 Sep 2005 04:40:59 -0300 Subject: [LLC]: Do better struct sock accounting on skbs Signed-off-by: Jochen Friedrich Signed-off-by: Arnaldo Carvalho de Melo --- include/net/llc_sap.h | 6 ++++-- net/llc/af_llc.c | 2 -- net/llc/llc_c_ac.c | 2 +- net/llc/llc_conn.c | 14 +++++++++----- net/llc/llc_sap.c | 8 ++++---- 5 files changed, 18 insertions(+), 14 deletions(-) (limited to 'include/net') diff --git a/include/net/llc_sap.h b/include/net/llc_sap.h index 6dd4cdccca4..9dcfcf4a4ab 100644 --- a/include/net/llc_sap.h +++ b/include/net/llc_sap.h @@ -12,11 +12,13 @@ * See the GNU General Public License for more details. */ struct llc_sap; -struct sk_buff; struct net_device; +struct sk_buff; +struct sock; extern void llc_sap_rtn_pdu(struct llc_sap *sap, struct sk_buff *skb); -extern void llc_save_primitive(struct sk_buff* skb, unsigned char prim); +extern void llc_save_primitive(struct sock *sk, struct sk_buff* skb, + unsigned char prim); extern struct sk_buff *llc_alloc_frame(struct net_device *dev); extern void llc_build_and_send_test_pkt(struct llc_sap *sap, diff --git a/net/llc/af_llc.c b/net/llc/af_llc.c index f536369cdb5..ad9aad807aa 100644 --- a/net/llc/af_llc.c +++ b/net/llc/af_llc.c @@ -628,7 +628,6 @@ static int llc_ui_accept(struct socket *sock, struct socket *newsock, int flags) /* put original socket back into a clean listen state. */ sk->sk_state = TCP_LISTEN; sk->sk_ack_backlog--; - skb->sk = NULL; dprintk("%s: ok success on %02X, client on %02X\n", __FUNCTION__, llc_sk(sk)->addr.sllc_sap, newllc->daddr.lsap); frees: @@ -740,7 +739,6 @@ static int llc_ui_sendmsg(struct kiocb *iocb, struct socket *sock, lock_sock(sk); if (!skb) goto release; - skb->sk = sk; skb->dev = llc->dev; skb->protocol = llc_proto_type(addr->sllc_arphrd); skb_reserve(skb, hdrlen); diff --git a/net/llc/llc_c_ac.c b/net/llc/llc_c_ac.c index 9d9b6c8aeb4..c1e75103189 100644 --- a/net/llc/llc_c_ac.c +++ b/net/llc/llc_c_ac.c @@ -1335,7 +1335,7 @@ static void llc_conn_tmr_common_cb(unsigned long timeout_data, u8 type) if (skb) { struct llc_conn_state_ev *ev = llc_conn_ev(skb); - skb->sk = sk; + skb_set_owner_r(skb, sk); ev->type = type; llc_process_tmr_ev(sk, skb); } diff --git a/net/llc/llc_conn.c b/net/llc/llc_conn.c index d3783f8ee48..9f0e10e0978 100644 --- a/net/llc/llc_conn.c +++ b/net/llc/llc_conn.c @@ -90,8 +90,8 @@ int llc_conn_state_process(struct sock *sk, struct sk_buff *skb) switch (ev->ind_prim) { case LLC_DATA_PRIM: - llc_save_primitive(skb, LLC_DATA_PRIM); - if (sock_queue_rcv_skb(sk, skb)) { + llc_save_primitive(sk, skb, LLC_DATA_PRIM); + if (unlikely(sock_queue_rcv_skb(sk, skb))) { /* * shouldn't happen */ @@ -103,6 +103,11 @@ int llc_conn_state_process(struct sock *sk, struct sk_buff *skb) case LLC_CONN_PRIM: { struct sock *parent = skb->sk; + skb_orphan(skb); + /* + * Set the skb->sk to the new struct sock, so that at accept + * type the upper layer can get the newly created struct sock. + */ skb->sk = sk; skb_queue_tail(&parent->sk_receive_queue, skb); sk->sk_state_change(parent); @@ -702,10 +707,9 @@ void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb) memcpy(&llc->daddr, &saddr, sizeof(llc->daddr)); llc_sap_add_socket(sap, sk); sock_hold(sk); + skb_set_owner_r(skb, parent); sock_put(parent); - skb->sk = parent; - } else - skb->sk = sk; + } bh_lock_sock(sk); if (!sock_owned_by_user(sk)) llc_conn_rcv(sk, skb); diff --git a/net/llc/llc_sap.c b/net/llc/llc_sap.c index 0adaa289bf0..9f064b3a433 100644 --- a/net/llc/llc_sap.c +++ b/net/llc/llc_sap.c @@ -45,12 +45,12 @@ struct sk_buff *llc_alloc_frame(struct net_device *dev) return skb; } -void llc_save_primitive(struct sk_buff* skb, u8 prim) +void llc_save_primitive(struct sock *sk, struct sk_buff* skb, u8 prim) { struct sockaddr_llc *addr = llc_ui_skb_cb(skb); /* save primitive for use by the user. */ - addr->sllc_family = skb->sk->sk_family; + addr->sllc_family = sk->sk_family; addr->sllc_arphrd = skb->dev->type; addr->sllc_test = prim == LLC_TEST_PRIM; addr->sllc_xid = prim == LLC_XID_PRIM; @@ -190,7 +190,7 @@ static void llc_sap_state_process(struct llc_sap *sap, struct sk_buff *skb) if (skb->sk->sk_state == TCP_LISTEN) kfree_skb(skb); else { - llc_save_primitive(skb, ev->prim); + llc_save_primitive(skb->sk, skb, ev->prim); /* queue skb to the user. */ if (sock_queue_rcv_skb(skb->sk, skb)) @@ -309,7 +309,7 @@ void llc_sap_handler(struct llc_sap *sap, struct sk_buff *skb) sk = llc_lookup_dgram(sap, &laddr); if (sk) { - skb->sk = sk; + skb_set_owner_r(skb, sk); llc_sap_rcv(sap, skb); sock_put(sk); } else -- cgit v1.2.3-70-g09d2 From 6e2144b76840be09924de1626e2dcd7b315f75b3 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 22 Sep 2005 04:43:05 -0300 Subject: [LLC]: Use refcounting with struct llc_sap Signed-off-by: Jochen Friedrich Signed-off-by: Arnaldo Carvalho de Melo --- include/net/llc.h | 16 +++++++++++++++- include/net/llc_conn.h | 1 - net/802/p8022.c | 2 +- net/802/psnap.c | 2 +- net/llc/af_llc.c | 11 ++++++----- net/llc/llc_c_ac.c | 1 + net/llc/llc_conn.c | 28 ++-------------------------- net/llc/llc_core.c | 34 +++++++++++++++++++++++----------- net/llc/llc_input.c | 9 +++++++-- 9 files changed, 56 insertions(+), 48 deletions(-) (limited to 'include/net') diff --git a/include/net/llc.h b/include/net/llc.h index 8b8e2be289b..93e5b443a9a 100644 --- a/include/net/llc.h +++ b/include/net/llc.h @@ -17,6 +17,8 @@ #include #include +#include + struct net_device; struct packet_type; struct sk_buff; @@ -44,6 +46,7 @@ struct llc_sap { unsigned char state; unsigned char p_bit; unsigned char f_bit; + atomic_t refcnt; int (*rcv_func)(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, @@ -81,7 +84,18 @@ extern struct llc_sap *llc_sap_open(unsigned char lsap, struct net_device *dev, struct packet_type *pt, struct net_device *orig_dev)); -extern void llc_sap_close(struct llc_sap *sap); +static inline void llc_sap_hold(struct llc_sap *sap) +{ + atomic_inc(&sap->refcnt); +} + +static inline void llc_sap_put(struct llc_sap *sap) +{ + extern void llc_sap_close(struct llc_sap *sap); + + if (atomic_dec_and_test(&sap->refcnt)) + llc_sap_close(sap); +} extern struct llc_sap *llc_sap_find(unsigned char sap_value); diff --git a/include/net/llc_conn.h b/include/net/llc_conn.h index 8a8ff481013..b2889218c76 100644 --- a/include/net/llc_conn.h +++ b/include/net/llc_conn.h @@ -115,5 +115,4 @@ extern void llc_sap_remove_socket(struct llc_sap *sap, struct sock *sk); extern u8 llc_data_accept_state(u8 state); extern void llc_build_offset_table(void); -extern int llc_release_sockets(struct llc_sap *sap); #endif /* LLC_CONN_H */ diff --git a/net/802/p8022.c b/net/802/p8022.c index b24817c63ca..2530f35241c 100644 --- a/net/802/p8022.c +++ b/net/802/p8022.c @@ -56,7 +56,7 @@ struct datalink_proto *register_8022_client(unsigned char type, void unregister_8022_client(struct datalink_proto *proto) { - llc_sap_close(proto->sap); + llc_sap_put(proto->sap); kfree(proto); } diff --git a/net/802/psnap.c b/net/802/psnap.c index ab80b1fab53..4d638944d93 100644 --- a/net/802/psnap.c +++ b/net/802/psnap.c @@ -106,7 +106,7 @@ module_init(snap_init); static void __exit snap_exit(void) { - llc_sap_close(snap_sap); + llc_sap_put(snap_sap); } module_exit(snap_exit); diff --git a/net/llc/af_llc.c b/net/llc/af_llc.c index ad9aad807aa..a75b8f2aab1 100644 --- a/net/llc/af_llc.c +++ b/net/llc/af_llc.c @@ -21,6 +21,7 @@ * See the GNU General Public License for more details. */ #include +#include #include #include #include @@ -188,10 +189,6 @@ static int llc_ui_release(struct socket *sock) if (!sock_flag(sk, SOCK_ZAPPED)) llc_sap_remove_socket(llc->sap, sk); release_sock(sk); - if (llc->sap && hlist_empty(&llc->sap->sk_list.list)) { - llc_release_sockets(llc->sap); - llc_sap_close(llc->sap); - } if (llc->dev) dev_put(llc->dev); sock_put(sk); @@ -220,6 +217,7 @@ static int llc_ui_autoport(void) llc_ui_sap_last_autoport = i + 2; goto out; } + llc_sap_put(sap); } llc_ui_sap_last_autoport = LLC_SAP_DYN_START; tries++; @@ -310,6 +308,7 @@ static int llc_ui_bind(struct socket *sock, struct sockaddr *uaddr, int addrlen) rc = -EBUSY; /* some other network layer is using the sap */ if (!sap) goto out; + llc_sap_hold(sap); } else { struct llc_addr laddr, daddr; struct sock *ask; @@ -326,7 +325,7 @@ static int llc_ui_bind(struct socket *sock, struct sockaddr *uaddr, int addrlen) ask = llc_lookup_established(sap, &daddr, &laddr); if (ask) { sock_put(ask); - goto out; + goto out_put; } } llc->laddr.lsap = addr->sllc_sap; @@ -336,6 +335,8 @@ static int llc_ui_bind(struct socket *sock, struct sockaddr *uaddr, int addrlen) llc_sap_add_socket(sap, sk); sock_reset_flag(sk, SOCK_ZAPPED); rc = 0; +out_put: + llc_sap_put(sap); out: return rc; } diff --git a/net/llc/llc_c_ac.c b/net/llc/llc_c_ac.c index c1e75103189..05236c2cbb9 100644 --- a/net/llc/llc_c_ac.c +++ b/net/llc/llc_c_ac.c @@ -75,6 +75,7 @@ int llc_conn_ac_conn_ind(struct sock *sk, struct sk_buff *skb) llc->dev = skb->dev; ev->ind_prim = LLC_CONN_PRIM; rc = 0; + llc_sap_put(sap); } return rc; } diff --git a/net/llc/llc_conn.c b/net/llc/llc_conn.c index 9f0e10e0978..ab9e6d7e287 100644 --- a/net/llc/llc_conn.c +++ b/net/llc/llc_conn.c @@ -633,6 +633,7 @@ static int llc_find_offset(int state, int ev_type) */ void llc_sap_add_socket(struct llc_sap *sap, struct sock *sk) { + llc_sap_hold(sap); write_lock_bh(&sap->sk_list.lock); llc_sk(sk)->sap = sap; sk_add_node(sk, &sap->sk_list.list); @@ -652,6 +653,7 @@ void llc_sap_remove_socket(struct llc_sap *sap, struct sock *sk) write_lock_bh(&sap->sk_list.lock); sk_del_node_init(sk); write_unlock_bh(&sap->sk_list.lock); + llc_sap_put(sap); } /** @@ -730,32 +732,6 @@ drop: static atomic_t llc_sock_nr; #endif -/** - * llc_release_sockets - releases all sockets in a sap - * @sap: sap to release its sockets - * - * Releases all connections of a sap. Returns 0 if all actions complete - * successfully, nonzero otherwise - */ -int llc_release_sockets(struct llc_sap *sap) -{ - int rc = 0; - struct sock *sk; - struct hlist_node *node; - - write_lock_bh(&sap->sk_list.lock); - - sk_for_each(sk, node, &sap->sk_list.list) { - llc_sk(sk)->state = LLC_CONN_STATE_TEMP; - - if (llc_send_disc(sk)) - rc = 1; - } - - write_unlock_bh(&sap->sk_list.lock); - return rc; -} - /** * llc_backlog_rcv - Processes rx frames and expired timers. * @sk: LLC sock (p8022 connection) diff --git a/net/llc/llc_core.c b/net/llc/llc_core.c index 9727455bf0e..9ccff1266b2 100644 --- a/net/llc/llc_core.c +++ b/net/llc/llc_core.c @@ -40,6 +40,7 @@ static struct llc_sap *llc_sap_alloc(void) sap->state = LLC_SAP_STATE_ACTIVE; memcpy(sap->laddr.mac, llc_station_mac_sa, ETH_ALEN); rwlock_init(&sap->sk_list.lock); + atomic_set(&sap->refcnt, 1); } return sap; } @@ -52,9 +53,7 @@ static struct llc_sap *llc_sap_alloc(void) */ static void llc_add_sap(struct llc_sap *sap) { - write_lock_bh(&llc_sap_list_lock); list_add_tail(&sap->node, &llc_sap_list); - write_unlock_bh(&llc_sap_list_lock); } /** @@ -70,11 +69,25 @@ static void llc_del_sap(struct llc_sap *sap) write_unlock_bh(&llc_sap_list_lock); } +struct llc_sap *__llc_sap_find(unsigned char sap_value) +{ + struct llc_sap* sap; + + list_for_each_entry(sap, &llc_sap_list, node) + if (sap->laddr.lsap == sap_value) + goto out; + sap = NULL; +out: + return sap; +} + /** * llc_sap_find - searchs a SAP in station * @sap_value: sap to be found * * Searchs for a sap in the sap list of the LLC's station upon the sap ID. + * If the sap is found it will be refcounted and the user will have to do + * a llc_sap_put after use. * Returns the sap or %NULL if not found. */ struct llc_sap *llc_sap_find(unsigned char sap_value) @@ -82,11 +95,9 @@ struct llc_sap *llc_sap_find(unsigned char sap_value) struct llc_sap* sap; read_lock_bh(&llc_sap_list_lock); - list_for_each_entry(sap, &llc_sap_list, node) - if (sap->laddr.lsap == sap_value) - goto out; - sap = NULL; -out: + sap = __llc_sap_find(sap_value); + if (sap) + llc_sap_hold(sap); read_unlock_bh(&llc_sap_list_lock); return sap; } @@ -106,19 +117,20 @@ struct llc_sap *llc_sap_open(unsigned char lsap, struct packet_type *pt, struct net_device *orig_dev)) { - struct llc_sap *sap = llc_sap_find(lsap); + struct llc_sap *sap = NULL; - if (sap) { /* SAP already exists */ - sap = NULL; + write_lock_bh(&llc_sap_list_lock); + if (__llc_sap_find(lsap)) /* SAP already exists */ goto out; - } sap = llc_sap_alloc(); if (!sap) goto out; sap->laddr.lsap = lsap; sap->rcv_func = func; + llc_sap_hold(sap); llc_add_sap(sap); out: + write_unlock_bh(&llc_sap_list_lock); return sap; } diff --git a/net/llc/llc_input.c b/net/llc/llc_input.c index 60c1acac7c9..789eec42645 100644 --- a/net/llc/llc_input.c +++ b/net/llc/llc_input.c @@ -166,17 +166,22 @@ int llc_rcv(struct sk_buff *skb, struct net_device *dev, */ if (sap->rcv_func) { sap->rcv_func(skb, dev, pt, orig_dev); - goto out; + goto out_put; } dest = llc_pdu_type(skb); if (unlikely(!dest || !llc_type_handlers[dest - 1])) - goto drop; + goto drop_put; llc_type_handlers[dest - 1](sap, skb); +out_put: + llc_sap_put(sap); out: return 0; drop: kfree_skb(skb); goto out; +drop_put: + kfree_skb(skb); + goto out_put; handle_station: if (!llc_station_handler) goto drop; -- cgit v1.2.3-70-g09d2 From 2928c19e1086e2f1e90d05931437ab6f1e4cfdc8 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 22 Sep 2005 05:14:33 -0300 Subject: [LLC]: Fix sparse warnings Signed-off-by: Arnaldo Carvalho de Melo --- include/net/llc.h | 13 +++++++++++-- include/net/llc_conn.h | 3 ++- net/llc/af_llc.c | 5 +---- net/llc/llc_c_ac.c | 2 +- net/llc/llc_conn.c | 3 ++- net/llc/llc_core.c | 2 +- net/llc/sysctl_net_llc.c | 7 +------ 7 files changed, 19 insertions(+), 16 deletions(-) (limited to 'include/net') diff --git a/include/net/llc.h b/include/net/llc.h index 93e5b443a9a..1adb2ef3f6f 100644 --- a/include/net/llc.h +++ b/include/net/llc.h @@ -89,10 +89,10 @@ static inline void llc_sap_hold(struct llc_sap *sap) atomic_inc(&sap->refcnt); } +extern void llc_sap_close(struct llc_sap *sap); + static inline void llc_sap_put(struct llc_sap *sap) { - extern void llc_sap_close(struct llc_sap *sap); - if (atomic_dec_and_test(&sap->refcnt)) llc_sap_close(sap); } @@ -102,6 +102,9 @@ extern struct llc_sap *llc_sap_find(unsigned char sap_value); extern int llc_build_and_send_ui_pkt(struct llc_sap *sap, struct sk_buff *skb, unsigned char *dmac, unsigned char dsap); +extern void llc_sap_handler(struct llc_sap *sap, struct sk_buff *skb); +extern void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb); + extern int llc_station_init(void); extern void llc_station_exit(void); @@ -115,6 +118,12 @@ extern void llc_proc_exit(void); #ifdef CONFIG_SYSCTL extern int llc_sysctl_init(void); extern void llc_sysctl_exit(void); + +extern int sysctl_llc2_ack_timeout; +extern int sysctl_llc2_busy_timeout; +extern int sysctl_llc2_p_timeout; +extern int sysctl_llc2_rej_timeout; +extern int sysctl_llc_station_ack_timeout; #else #define llc_sysctl_init() (0) #define llc_sysctl_exit() do { } while(0) diff --git a/include/net/llc_conn.h b/include/net/llc_conn.h index b2889218c76..e44f494bbef 100644 --- a/include/net/llc_conn.h +++ b/include/net/llc_conn.h @@ -92,7 +92,8 @@ static __inline__ char llc_backlog_type(struct sk_buff *skb) return skb->cb[sizeof(skb->cb) - 1]; } -extern struct sock *llc_sk_alloc(int family, int priority, struct proto *prot); +extern struct sock *llc_sk_alloc(int family, unsigned int __nocast priority, + struct proto *prot); extern void llc_sk_free(struct sock *sk); extern void llc_sk_reset(struct sock *sk); diff --git a/net/llc/af_llc.c b/net/llc/af_llc.c index 3361ae900e2..7aa51eb79b1 100644 --- a/net/llc/af_llc.c +++ b/net/llc/af_llc.c @@ -555,7 +555,7 @@ static int llc_ui_wait_for_busy_core(struct sock *sk, long timeout) return rc; } -int llc_wait_data(struct sock *sk, long timeo) +static int llc_wait_data(struct sock *sk, long timeo) { int rc; @@ -1003,9 +1003,6 @@ static struct proto_ops llc_ui_ops = { .sendpage = sock_no_sendpage, }; -extern void llc_sap_handler(struct llc_sap *sap, struct sk_buff *skb); -extern void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb); - static char llc_proc_err_msg[] __initdata = KERN_CRIT "LLC: Unable to register the proc_fs entries\n"; static char llc_sysctl_err_msg[] __initdata = diff --git a/net/llc/llc_c_ac.c b/net/llc/llc_c_ac.c index 05236c2cbb9..8f7b46d2063 100644 --- a/net/llc/llc_c_ac.c +++ b/net/llc/llc_c_ac.c @@ -1321,7 +1321,7 @@ int llc_conn_ac_set_vs_nr(struct sock *sk, struct sk_buff *skb) return 0; } -int llc_conn_ac_inc_vs_by_1(struct sock *sk, struct sk_buff *skb) +static int llc_conn_ac_inc_vs_by_1(struct sock *sk, struct sk_buff *skb) { llc_sk(sk)->vS = (llc_sk(sk)->vS + 1) % 128; return 0; diff --git a/net/llc/llc_conn.c b/net/llc/llc_conn.c index ab9e6d7e287..76f94e0d840 100644 --- a/net/llc/llc_conn.c +++ b/net/llc/llc_conn.c @@ -819,7 +819,8 @@ static void llc_sk_init(struct sock* sk) * Allocates a LLC sock and initializes it. Returns the new LLC sock * or %NULL if there's no memory available for one */ -struct sock *llc_sk_alloc(int family, int priority, struct proto *prot) +struct sock *llc_sk_alloc(int family, unsigned int __nocast priority, + struct proto *prot) { struct sock *sk = sk_alloc(family, priority, prot, 1); diff --git a/net/llc/llc_core.c b/net/llc/llc_core.c index 9ccff1266b2..ab0fcd32fd8 100644 --- a/net/llc/llc_core.c +++ b/net/llc/llc_core.c @@ -69,7 +69,7 @@ static void llc_del_sap(struct llc_sap *sap) write_unlock_bh(&llc_sap_list_lock); } -struct llc_sap *__llc_sap_find(unsigned char sap_value) +static struct llc_sap *__llc_sap_find(unsigned char sap_value) { struct llc_sap* sap; diff --git a/net/llc/sysctl_net_llc.c b/net/llc/sysctl_net_llc.c index 4d99d2f27a2..d1eaddb1363 100644 --- a/net/llc/sysctl_net_llc.c +++ b/net/llc/sysctl_net_llc.c @@ -8,17 +8,12 @@ #include #include #include +#include #ifndef CONFIG_SYSCTL #error This file should not be compiled without CONFIG_SYSCTL defined #endif -extern int sysctl_llc2_ack_timeout; -extern int sysctl_llc2_busy_timeout; -extern int sysctl_llc2_p_timeout; -extern int sysctl_llc2_rej_timeout; -extern int sysctl_llc_station_ack_timeout; - static struct ctl_table llc2_timeout_table[] = { { .ctl_name = NET_LLC2_ACK_TIMEOUT, -- cgit v1.2.3-70-g09d2 From d389424e00f9097cd24b3df4ca0ab7221f140eeb Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 22 Sep 2005 07:57:21 -0300 Subject: [LLC]: Fix the accept path Borrowing the structure of TCP/IP for this. On the receive of new connections I was bh_lock_socking the _new_ sock, not the listening one, duh, now it survives the ssh connections storm I've been using to test this specific bug. Also fixes send side skb sock accounting. Signed-off-by: Arnaldo Carvalho de Melo --- include/net/llc_sap.h | 3 +- net/llc/llc_c_ac.c | 64 +++++++++-------------- net/llc/llc_conn.c | 142 +++++++++++++++++++++++++++++++++----------------- net/llc/llc_if.c | 2 + net/llc/llc_s_ac.c | 4 +- net/llc/llc_sap.c | 4 +- net/llc/llc_station.c | 6 +-- 7 files changed, 132 insertions(+), 93 deletions(-) (limited to 'include/net') diff --git a/include/net/llc_sap.h b/include/net/llc_sap.h index 9dcfcf4a4ab..2c56dbece72 100644 --- a/include/net/llc_sap.h +++ b/include/net/llc_sap.h @@ -19,7 +19,8 @@ struct sock; extern void llc_sap_rtn_pdu(struct llc_sap *sap, struct sk_buff *skb); extern void llc_save_primitive(struct sock *sk, struct sk_buff* skb, unsigned char prim); -extern struct sk_buff *llc_alloc_frame(struct net_device *dev); +extern struct sk_buff *llc_alloc_frame(struct sock *sk, + struct net_device *dev); extern void llc_build_and_send_test_pkt(struct llc_sap *sap, struct sk_buff *skb, diff --git a/net/llc/llc_c_ac.c b/net/llc/llc_c_ac.c index 8f7b46d2063..b0bcfb1f12d 100644 --- a/net/llc/llc_c_ac.c +++ b/net/llc/llc_c_ac.c @@ -60,24 +60,10 @@ int llc_conn_ac_clear_remote_busy(struct sock *sk, struct sk_buff *skb) int llc_conn_ac_conn_ind(struct sock *sk, struct sk_buff *skb) { - int rc = -ENOTCONN; - u8 dsap; - struct llc_sap *sap; - - llc_pdu_decode_dsap(skb, &dsap); - sap = llc_sap_find(dsap); - if (sap) { - struct llc_conn_state_ev *ev = llc_conn_ev(skb); - struct llc_sock *llc = llc_sk(sk); + struct llc_conn_state_ev *ev = llc_conn_ev(skb); - llc_pdu_decode_sa(skb, llc->daddr.mac); - llc_pdu_decode_da(skb, llc->laddr.mac); - llc->dev = skb->dev; - ev->ind_prim = LLC_CONN_PRIM; - rc = 0; - llc_sap_put(sap); - } - return rc; + ev->ind_prim = LLC_CONN_PRIM; + return 0; } int llc_conn_ac_conn_confirm(struct sock *sk, struct sk_buff *skb) @@ -213,7 +199,7 @@ int llc_conn_ac_send_disc_cmd_p_set_x(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -238,7 +224,7 @@ int llc_conn_ac_send_dm_rsp_f_set_p(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -264,7 +250,7 @@ int llc_conn_ac_send_dm_rsp_f_set_1(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -297,7 +283,7 @@ int llc_conn_ac_send_frmr_rsp_f_set_x(struct sock *sk, struct sk_buff *skb) llc_pdu_decode_pf_bit(skb, &f_bit); else f_bit = 0; - nskb = llc_alloc_frame(llc->dev); + nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -321,7 +307,7 @@ int llc_conn_ac_resend_frmr_rsp_f_set_0(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -351,7 +337,7 @@ int llc_conn_ac_resend_frmr_rsp_f_set_p(struct sock *sk, struct sk_buff *skb) struct llc_sock *llc = llc_sk(sk); llc_pdu_decode_pf_bit(skb, &f_bit); - nskb = llc_alloc_frame(llc->dev); + nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb); @@ -439,7 +425,7 @@ int llc_conn_ac_resend_i_xxx_x_set_0_or_send_rr(struct sock *sk, struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb); int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -474,7 +460,7 @@ int llc_conn_ac_send_rej_cmd_p_set_1(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -498,7 +484,7 @@ int llc_conn_ac_send_rej_rsp_f_set_1(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -522,7 +508,7 @@ int llc_conn_ac_send_rej_xxx_x_set_0(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -546,7 +532,7 @@ int llc_conn_ac_send_rnr_cmd_p_set_1(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -570,7 +556,7 @@ int llc_conn_ac_send_rnr_rsp_f_set_1(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -594,7 +580,7 @@ int llc_conn_ac_send_rnr_xxx_x_set_0(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -630,7 +616,7 @@ int llc_conn_ac_opt_send_rnr_xxx_x_set_0(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -654,7 +640,7 @@ int llc_conn_ac_send_rr_cmd_p_set_1(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -678,7 +664,7 @@ int llc_conn_ac_send_rr_rsp_f_set_1(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -703,7 +689,7 @@ int llc_conn_ac_send_ack_rsp_f_set_1(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -727,7 +713,7 @@ int llc_conn_ac_send_rr_xxx_x_set_0(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -751,7 +737,7 @@ int llc_conn_ac_send_ack_xxx_x_set_0(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -785,7 +771,7 @@ int llc_conn_ac_send_sabme_cmd_p_set_x(struct sock *sk, struct sk_buff *skb) { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; @@ -814,7 +800,7 @@ int llc_conn_ac_send_ua_rsp_f_set_p(struct sock *sk, struct sk_buff *skb) u8 f_bit; int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); llc_pdu_decode_pf_bit(skb, &f_bit); if (nskb) { @@ -970,7 +956,7 @@ static int llc_conn_ac_send_rr_rsp_f_set_ackpf(struct sock *sk, { int rc = -ENOBUFS; struct llc_sock *llc = llc_sk(sk); - struct sk_buff *nskb = llc_alloc_frame(llc->dev); + struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev); if (nskb) { struct llc_sap *sap = llc->sap; diff --git a/net/llc/llc_conn.c b/net/llc/llc_conn.c index 76f94e0d840..e10ce5adb10 100644 --- a/net/llc/llc_conn.c +++ b/net/llc/llc_conn.c @@ -58,7 +58,7 @@ int sysctl_llc2_busy_timeout = LLC2_BUSY_TIME * HZ; int llc_conn_state_process(struct sock *sk, struct sk_buff *skb) { int rc; - struct llc_sock *llc = llc_sk(sk); + struct llc_sock *llc = llc_sk(skb->sk); struct llc_conn_state_ev *ev = llc_conn_ev(skb); /* @@ -68,7 +68,10 @@ int llc_conn_state_process(struct sock *sk, struct sk_buff *skb) */ skb_get(skb); ev->ind_prim = ev->cfm_prim = 0; - rc = llc_conn_service(sk, skb); /* sending event to state machine */ + /* + * Send event to state machine + */ + rc = llc_conn_service(skb->sk, skb); if (unlikely(rc != 0)) { printk(KERN_ERR "%s: llc_conn_service failed\n", __FUNCTION__); goto out_kfree_skb; @@ -100,18 +103,14 @@ int llc_conn_state_process(struct sock *sk, struct sk_buff *skb) kfree_skb(skb); } break; - case LLC_CONN_PRIM: { - struct sock *parent = skb->sk; - - skb_orphan(skb); + case LLC_CONN_PRIM: /* - * Set the skb->sk to the new struct sock, so that at accept - * type the upper layer can get the newly created struct sock. + * Can't be sock_queue_rcv_skb, because we have to leave the + * skb->sk pointing to the newly created struct sock in + * llc_conn_handler. -acme */ - skb->sk = sk; - skb_queue_tail(&parent->sk_receive_queue, skb); - sk->sk_state_change(parent); - } + skb_queue_tail(&sk->sk_receive_queue, skb); + sk->sk_state_change(sk); break; case LLC_DISC_PRIM: sock_hold(sk); @@ -475,7 +474,7 @@ static int llc_exec_conn_trans_actions(struct sock *sk, } /** - * llc_lookup_established - Finds connection for the remote/local sap/mac + * __llc_lookup_established - Finds connection for the remote/local sap/mac * @sap: SAP * @daddr: address of remote LLC (MAC + SAP) * @laddr: address of local LLC (MAC + SAP) @@ -483,14 +482,16 @@ static int llc_exec_conn_trans_actions(struct sock *sk, * Search connection list of the SAP and finds connection using the remote * mac, remote sap, local mac, and local sap. Returns pointer for * connection found, %NULL otherwise. + * Caller has to make sure local_bh is disabled. */ -struct sock *llc_lookup_established(struct llc_sap *sap, struct llc_addr *daddr, - struct llc_addr *laddr) +static struct sock *__llc_lookup_established(struct llc_sap *sap, + struct llc_addr *daddr, + struct llc_addr *laddr) { struct sock *rc; struct hlist_node *node; - read_lock_bh(&sap->sk_list.lock); + read_lock(&sap->sk_list.lock); sk_for_each(rc, node, &sap->sk_list.list) { struct llc_sock *llc = llc_sk(rc); @@ -504,10 +505,22 @@ struct sock *llc_lookup_established(struct llc_sap *sap, struct llc_addr *daddr, } rc = NULL; found: - read_unlock_bh(&sap->sk_list.lock); + read_unlock(&sap->sk_list.lock); return rc; } +struct sock *llc_lookup_established(struct llc_sap *sap, + struct llc_addr *daddr, + struct llc_addr *laddr) +{ + struct sock *sk; + + local_bh_disable(); + sk = __llc_lookup_established(sap, daddr, laddr); + local_bh_enable(); + return sk; +} + /** * llc_lookup_listener - Finds listener for local MAC + SAP * @sap: SAP @@ -516,6 +529,7 @@ found: * Search connection list of the SAP and finds connection listening on * local mac, and local sap. Returns pointer for parent socket found, * %NULL otherwise. + * Caller has to make sure local_bh is disabled. */ static struct sock *llc_lookup_listener(struct llc_sap *sap, struct llc_addr *laddr) @@ -523,7 +537,7 @@ static struct sock *llc_lookup_listener(struct llc_sap *sap, struct sock *rc; struct hlist_node *node; - read_lock_bh(&sap->sk_list.lock); + read_lock(&sap->sk_list.lock); sk_for_each(rc, node, &sap->sk_list.list) { struct llc_sock *llc = llc_sk(rc); @@ -537,10 +551,19 @@ static struct sock *llc_lookup_listener(struct llc_sap *sap, } rc = NULL; found: - read_unlock_bh(&sap->sk_list.lock); + read_unlock(&sap->sk_list.lock); return rc; } +static struct sock *__llc_lookup(struct llc_sap *sap, + struct llc_addr *daddr, + struct llc_addr *laddr) +{ + struct sock *sk = __llc_lookup_established(sap, daddr, laddr); + + return sk ? : llc_lookup_listener(sap, laddr); +} + /** * llc_data_accept_state - designates if in this state data can be sent. * @state: state of connection. @@ -666,15 +689,34 @@ void llc_sap_remove_socket(struct llc_sap *sap, struct sock *sk) static int llc_conn_rcv(struct sock* sk, struct sk_buff *skb) { struct llc_conn_state_ev *ev = llc_conn_ev(skb); - struct llc_sock *llc = llc_sk(sk); - if (!llc->dev) - llc->dev = skb->dev; ev->type = LLC_CONN_EV_TYPE_PDU; ev->reason = 0; return llc_conn_state_process(sk, skb); } +static struct sock *llc_create_incoming_sock(struct sock *sk, + struct net_device *dev, + struct llc_addr *saddr, + struct llc_addr *daddr) +{ + struct sock *newsk = llc_sk_alloc(sk->sk_family, GFP_ATOMIC, + sk->sk_prot); + struct llc_sock *newllc, *llc = llc_sk(sk); + + if (!newsk) + goto out; + newllc = llc_sk(newsk); + memcpy(&newllc->laddr, daddr, sizeof(newllc->laddr)); + memcpy(&newllc->daddr, saddr, sizeof(newllc->daddr)); + newllc->dev = dev; + dev_hold(dev); + llc_sap_add_socket(llc->sap, newsk); + llc_sap_hold(llc->sap); +out: + return newsk; +} + void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb) { struct llc_addr saddr, daddr; @@ -685,34 +727,35 @@ void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb) llc_pdu_decode_da(skb, daddr.mac); llc_pdu_decode_dsap(skb, &daddr.lsap); - sk = llc_lookup_established(sap, &saddr, &daddr); - if (!sk) { + sk = __llc_lookup(sap, &saddr, &daddr); + if (!sk) + goto drop; + + bh_lock_sock(sk); + /* + * This has to be done here and not at the upper layer ->accept + * method because of the way the PROCOM state machine works: + * it needs to set several state variables (see, for instance, + * llc_adm_actions_2 in net/llc/llc_c_st.c) and send a packet to + * the originator of the new connection, and this state has to be + * in the newly created struct sock private area. -acme + */ + if (unlikely(sk->sk_state == TCP_LISTEN)) { + struct sock *newsk = llc_create_incoming_sock(sk, skb->dev, + &saddr, &daddr); + if (!newsk) + goto drop_unlock; + skb_set_owner_r(skb, newsk); + } else { /* - * Didn't find an active connection; verify if there - * is a listening socket for this llc addr + * Can't be skb_set_owner_r, this will be done at the + * llc_conn_state_process function, later on, when we will use + * skb_queue_rcv_skb to send it to upper layers, this is + * another trick required to cope with how the PROCOM state + * machine works. -acme */ - struct llc_sock *llc; - struct sock *parent = llc_lookup_listener(sap, &daddr); - - if (!parent) { - dprintk("llc_lookup_listener failed!\n"); - goto drop; - } - - sk = llc_sk_alloc(parent->sk_family, GFP_ATOMIC, parent->sk_prot); - if (!sk) { - sock_put(parent); - goto drop; - } - llc = llc_sk(sk); - memcpy(&llc->laddr, &daddr, sizeof(llc->laddr)); - memcpy(&llc->daddr, &saddr, sizeof(llc->daddr)); - llc_sap_add_socket(sap, sk); - sock_hold(sk); - skb_set_owner_r(skb, parent); - sock_put(parent); + skb->sk = sk; } - bh_lock_sock(sk); if (!sock_owned_by_user(sk)) llc_conn_rcv(sk, skb); else { @@ -720,11 +763,16 @@ void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb) llc_set_backlog_type(skb, LLC_PACKET); sk_add_backlog(sk, skb); } +out: bh_unlock_sock(sk); sock_put(sk); return; drop: kfree_skb(skb); + return; +drop_unlock: + kfree_skb(skb); + goto out; } #undef LLC_REFCNT_DEBUG diff --git a/net/llc/llc_if.c b/net/llc/llc_if.c index 764dbd70405..ba90f7f0801 100644 --- a/net/llc/llc_if.c +++ b/net/llc/llc_if.c @@ -107,6 +107,7 @@ int llc_establish_connection(struct sock *sk, u8 *lmac, u8 *dmac, u8 dsap) ev->type = LLC_CONN_EV_TYPE_PRIM; ev->prim = LLC_CONN_PRIM; ev->prim_type = LLC_PRIM_TYPE_REQ; + skb_set_owner_w(skb, sk); rc = llc_conn_state_process(sk, skb); } out_put: @@ -141,6 +142,7 @@ int llc_send_disc(struct sock *sk) skb = alloc_skb(0, GFP_ATOMIC); if (!skb) goto out; + skb_set_owner_w(skb, sk); sk->sk_state = TCP_CLOSING; ev = llc_conn_ev(skb); ev->type = LLC_CONN_EV_TYPE_PRIM; diff --git a/net/llc/llc_s_ac.c b/net/llc/llc_s_ac.c index 6acdebf03b2..bb3580fb8cf 100644 --- a/net/llc/llc_s_ac.c +++ b/net/llc/llc_s_ac.c @@ -103,7 +103,7 @@ int llc_sap_action_send_xid_r(struct llc_sap *sap, struct sk_buff *skb) llc_pdu_decode_sa(skb, mac_da); llc_pdu_decode_da(skb, mac_sa); llc_pdu_decode_ssap(skb, &dsap); - nskb = llc_alloc_frame(skb->dev); + nskb = llc_alloc_frame(NULL, skb->dev); if (!nskb) goto out; llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap, dsap, @@ -148,7 +148,7 @@ int llc_sap_action_send_test_r(struct llc_sap *sap, struct sk_buff *skb) llc_pdu_decode_sa(skb, mac_da); llc_pdu_decode_da(skb, mac_sa); llc_pdu_decode_ssap(skb, &dsap); - nskb = llc_alloc_frame(skb->dev); + nskb = llc_alloc_frame(NULL, skb->dev); if (!nskb) goto out; llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap, dsap, diff --git a/net/llc/llc_sap.c b/net/llc/llc_sap.c index 9f064b3a433..e6d538937f9 100644 --- a/net/llc/llc_sap.c +++ b/net/llc/llc_sap.c @@ -31,7 +31,7 @@ * Allocates an sk_buff for frame and initializes sk_buff fields. * Returns allocated skb or %NULL when out of memory. */ -struct sk_buff *llc_alloc_frame(struct net_device *dev) +struct sk_buff *llc_alloc_frame(struct sock *sk, struct net_device *dev) { struct sk_buff *skb = alloc_skb(128, GFP_ATOMIC); @@ -41,6 +41,8 @@ struct sk_buff *llc_alloc_frame(struct net_device *dev) skb->protocol = htons(ETH_P_802_2); skb->dev = dev; skb->mac.raw = skb->head; + if (sk != NULL) + skb_set_owner_w(skb, sk); } return skb; } diff --git a/net/llc/llc_station.c b/net/llc/llc_station.c index ec4693fe312..f37dbf8ef12 100644 --- a/net/llc/llc_station.c +++ b/net/llc/llc_station.c @@ -254,7 +254,7 @@ static int llc_station_ac_inc_xid_r_cnt_by_1(struct sk_buff *skb) static int llc_station_ac_send_null_dsap_xid_c(struct sk_buff *skb) { int rc = 1; - struct sk_buff *nskb = llc_alloc_frame(skb->dev); + struct sk_buff *nskb = llc_alloc_frame(NULL, skb->dev); if (!nskb) goto out; @@ -275,7 +275,7 @@ static int llc_station_ac_send_xid_r(struct sk_buff *skb) { u8 mac_da[ETH_ALEN], dsap; int rc = 1; - struct sk_buff* nskb = llc_alloc_frame(skb->dev); + struct sk_buff* nskb = llc_alloc_frame(NULL, skb->dev); if (!nskb) goto out; @@ -299,7 +299,7 @@ static int llc_station_ac_send_test_r(struct sk_buff *skb) { u8 mac_da[ETH_ALEN], dsap; int rc = 1; - struct sk_buff *nskb = llc_alloc_frame(skb->dev); + struct sk_buff *nskb = llc_alloc_frame(NULL, skb->dev); if (!nskb) goto out; -- cgit v1.2.3-70-g09d2 From 8420e1b541fe92aee1d8d4d25d9e33eaca756a7b Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 22 Sep 2005 08:29:08 -0300 Subject: [LLC]: fix llc_ui_recvmsg, making it behave like tcp_recvmsg In fact it is an exact copy of the parts that makes sense to LLC :-) Signed-off-by: Arnaldo Carvalho de Melo --- include/net/llc_conn.h | 1 + net/llc/af_llc.c | 180 ++++++++++++++++++++++++++++++++++++++++--------- net/llc/llc_conn.c | 2 +- net/llc/llc_proc.c | 2 +- net/llc/llc_sap.c | 5 +- 5 files changed, 154 insertions(+), 36 deletions(-) (limited to 'include/net') diff --git a/include/net/llc_conn.h b/include/net/llc_conn.h index e44f494bbef..54852ff6033 100644 --- a/include/net/llc_conn.h +++ b/include/net/llc_conn.h @@ -38,6 +38,7 @@ struct llc_sock { struct llc_addr laddr; /* lsap/mac pair */ struct llc_addr daddr; /* dsap/mac pair */ struct net_device *dev; /* device to send to remote */ + u32 copied_seq; /* head of yet unread data */ u8 retry_count; /* number of retries */ u8 ack_must_be_send; u8 first_pdu_Ns; diff --git a/net/llc/af_llc.c b/net/llc/af_llc.c index 7aa51eb79b1..59d02cbbeb9 100644 --- a/net/llc/af_llc.c +++ b/net/llc/af_llc.c @@ -648,53 +648,167 @@ out: * llc_ui_recvmsg - copy received data to the socket user. * @sock: Socket to copy data from. * @msg: Various user space related information. - * @size: Size of user buffer. + * @len: Size of user buffer. * @flags: User specified flags. * * Copy received data to the socket user. * Returns non-negative upon success, negative otherwise. */ static int llc_ui_recvmsg(struct kiocb *iocb, struct socket *sock, - struct msghdr *msg, size_t size, int flags) + struct msghdr *msg, size_t len, int flags) { - struct sock *sk = sock->sk; struct sockaddr_llc *uaddr = (struct sockaddr_llc *)msg->msg_name; - struct sk_buff *skb; + const int nonblock = flags & MSG_DONTWAIT; + struct sk_buff *skb = NULL; + struct sock *sk = sock->sk; + struct llc_sock *llc = llc_sk(sk); size_t copied = 0; - int rc = -ENOMEM; - int noblock = flags & MSG_DONTWAIT; + u32 peek_seq = 0; + u32 *seq; + unsigned long used; + int target; /* Read at least this many bytes */ + long timeo; - dprintk("%s: receiving in %02X from %02X\n", __FUNCTION__, - llc_sk(sk)->laddr.lsap, llc_sk(sk)->daddr.lsap); lock_sock(sk); - if (skb_queue_empty(&sk->sk_receive_queue)) { - rc = llc_wait_data(sk, sock_rcvtimeo(sk, noblock)); - if (rc) - goto out; - } - skb = skb_dequeue(&sk->sk_receive_queue); - if (!skb) /* shutdown */ + copied = -ENOTCONN; + if (sk->sk_state == TCP_LISTEN) goto out; - copied = skb->len; - if (copied > size) - copied = size; - rc = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied); - if (rc) - goto dgram_free; - if (skb->len > copied) { - skb_pull(skb, copied); - skb_queue_head(&sk->sk_receive_queue, skb); - } - if (uaddr) - memcpy(uaddr, llc_ui_skb_cb(skb), sizeof(*uaddr)); - msg->msg_namelen = sizeof(*uaddr); - if (!skb->next) { -dgram_free: - kfree_skb(skb); - } + + timeo = sock_rcvtimeo(sk, nonblock); + + seq = &llc->copied_seq; + if (flags & MSG_PEEK) { + peek_seq = llc->copied_seq; + seq = &peek_seq; + } + + target = sock_rcvlowat(sk, flags & MSG_WAITALL, len); + copied = 0; + + do { + u32 offset; + + /* + * We need to check signals first, to get correct SIGURG + * handling. FIXME: Need to check this doesn't impact 1003.1g + * and move it down to the bottom of the loop + */ + if (signal_pending(current)) { + if (copied) + break; + copied = timeo ? sock_intr_errno(timeo) : -EAGAIN; + break; + } + + /* Next get a buffer. */ + + skb = skb_peek(&sk->sk_receive_queue); + if (skb) { + offset = *seq; + goto found_ok_skb; + } + /* Well, if we have backlog, try to process it now yet. */ + + if (copied >= target && !sk->sk_backlog.tail) + break; + + if (copied) { + if (sk->sk_err || + sk->sk_state == TCP_CLOSE || + (sk->sk_shutdown & RCV_SHUTDOWN) || + !timeo || + (flags & MSG_PEEK)) + break; + } else { + if (sock_flag(sk, SOCK_DONE)) + break; + + if (sk->sk_err) { + copied = sock_error(sk); + break; + } + if (sk->sk_shutdown & RCV_SHUTDOWN) + break; + + if (sk->sk_state == TCP_CLOSE) { + if (!sock_flag(sk, SOCK_DONE)) { + /* + * This occurs when user tries to read + * from never connected socket. + */ + copied = -ENOTCONN; + break; + } + break; + } + if (!timeo) { + copied = -EAGAIN; + break; + } + } + + if (copied >= target) { /* Do not sleep, just process backlog. */ + release_sock(sk); + lock_sock(sk); + } else + sk_wait_data(sk, &timeo); + + if ((flags & MSG_PEEK) && peek_seq != llc->copied_seq) { + if (net_ratelimit()) + printk(KERN_DEBUG "LLC(%s:%d): Application " + "bug, race in MSG_PEEK.\n", + current->comm, current->pid); + peek_seq = llc->copied_seq; + } + continue; + found_ok_skb: + /* Ok so how much can we use? */ + used = skb->len - offset; + if (len < used) + used = len; + + if (!(flags & MSG_TRUNC)) { + int rc = skb_copy_datagram_iovec(skb, offset, + msg->msg_iov, used); + if (rc) { + /* Exception. Bailout! */ + if (!copied) + copied = -EFAULT; + break; + } + } + + *seq += used; + copied += used; + len -= used; + + if (used + offset < skb->len) + continue; + + if (!(flags & MSG_PEEK)) { + sk_eat_skb(sk, skb); + *seq = 0; + } + } while (len > 0); + + /* + * According to UNIX98, msg_name/msg_namelen are ignored + * on connected socket. -ANK + * But... af_llc still doesn't have separate sets of methods for + * SOCK_DGRAM and SOCK_STREAM :-( So we have to do this test, will + * eventually fix this tho :-) -acme + */ + if (sk->sk_type == SOCK_DGRAM) + goto copy_uaddr; out: release_sock(sk); - return rc ? : copied; + return copied; +copy_uaddr: + if (uaddr != NULL && skb != NULL) { + memcpy(uaddr, llc_ui_skb_cb(skb), sizeof(*uaddr)); + msg->msg_namelen = sizeof(*uaddr); + } + goto out; } /** diff --git a/net/llc/llc_conn.c b/net/llc/llc_conn.c index e10ce5adb10..042b24a8ca4 100644 --- a/net/llc/llc_conn.c +++ b/net/llc/llc_conn.c @@ -120,8 +120,8 @@ int llc_conn_state_process(struct sock *sk, struct sk_buff *skb) sk->sk_socket->state = SS_UNCONNECTED; sk->sk_state = TCP_CLOSE; if (!sock_flag(sk, SOCK_DEAD)) { - sk->sk_state_change(sk); sock_set_flag(sk, SOCK_DEAD); + sk->sk_state_change(sk); } } kfree_skb(skb); diff --git a/net/llc/llc_proc.c b/net/llc/llc_proc.c index 36e8db3fa1a..bd531cb235a 100644 --- a/net/llc/llc_proc.c +++ b/net/llc/llc_proc.c @@ -134,7 +134,7 @@ static int llc_seq_socket_show(struct seq_file *seq, void *v) llc_ui_format_mac(seq, llc->daddr.mac); seq_printf(seq, "@%02X %8d %8d %2d %3d %4d\n", llc->daddr.lsap, atomic_read(&sk->sk_wmem_alloc), - atomic_read(&sk->sk_rmem_alloc), + atomic_read(&sk->sk_rmem_alloc) - llc->copied_seq, sk->sk_state, sk->sk_socket ? SOCK_INODE(sk->sk_socket)->i_uid : -1, llc->link); diff --git a/net/llc/llc_sap.c b/net/llc/llc_sap.c index e6d538937f9..4029ceee9b9 100644 --- a/net/llc/llc_sap.c +++ b/net/llc/llc_sap.c @@ -49,9 +49,12 @@ struct sk_buff *llc_alloc_frame(struct sock *sk, struct net_device *dev) void llc_save_primitive(struct sock *sk, struct sk_buff* skb, u8 prim) { - struct sockaddr_llc *addr = llc_ui_skb_cb(skb); + struct sockaddr_llc *addr; + if (skb->sk->sk_type == SOCK_STREAM) /* See UNIX98 */ + return; /* save primitive for use by the user. */ + addr = llc_ui_skb_cb(skb); addr->sllc_family = sk->sk_family; addr->sllc_arphrd = skb->dev->type; addr->sllc_test = prim == LLC_TEST_PRIM; -- cgit v1.2.3-70-g09d2 From 81c3d5470ecc70564eb9209946730fe2be93ad06 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 3 Oct 2005 14:13:38 -0700 Subject: [INET]: speedup inet (tcp/dccp) lookups Arnaldo and I agreed it could be applied now, because I have other pending patches depending on this one (Thank you Arnaldo) (The other important patch moves skc_refcnt in a separate cache line, so that the SMP/NUMA performance doesnt suffer from cache line ping pongs) 1) First some performance data : -------------------------------- tcp_v4_rcv() wastes a *lot* of time in __inet_lookup_established() The most time critical code is : sk_for_each(sk, node, &head->chain) { if (INET_MATCH(sk, acookie, saddr, daddr, ports, dif)) goto hit; /* You sunk my battleship! */ } The sk_for_each() does use prefetch() hints but only the begining of "struct sock" is prefetched. As INET_MATCH first comparison uses inet_sk(__sk)->daddr, wich is far away from the begining of "struct sock", it has to bring into CPU cache cold cache line. Each iteration has to use at least 2 cache lines. This can be problematic if some chains are very long. 2) The goal ----------- The idea I had is to change things so that INET_MATCH() may return FALSE in 99% of cases only using the data already in the CPU cache, using one cache line per iteration. 3) Description of the patch --------------------------- Adds a new 'unsigned int skc_hash' field in 'struct sock_common', filling a 32 bits hole on 64 bits platform. struct sock_common { unsigned short skc_family; volatile unsigned char skc_state; unsigned char skc_reuse; int skc_bound_dev_if; struct hlist_node skc_node; struct hlist_node skc_bind_node; atomic_t skc_refcnt; + unsigned int skc_hash; struct proto *skc_prot; }; Store in this 32 bits field the full hash, not masked by (ehash_size - 1) Using this full hash as the first comparison done in INET_MATCH permits us immediatly skip the element without touching a second cache line in case of a miss. Suppress the sk_hashent/tw_hashent fields since skc_hash (aliased to sk_hash and tw_hash) already contains the slot number if we mask with (ehash_size - 1) File include/net/inet_hashtables.h 64 bits platforms : #define INET_MATCH(__sk, __hash, __cookie, __saddr, __daddr, __ports, __dif)\ (((__sk)->sk_hash == (__hash)) ((*((__u64 *)&(inet_sk(__sk)->daddr)))== (__cookie)) && \ ((*((__u32 *)&(inet_sk(__sk)->dport))) == (__ports)) && \ (!((__sk)->sk_bound_dev_if) || ((__sk)->sk_bound_dev_if == (__dif)))) 32bits platforms: #define TCP_IPV4_MATCH(__sk, __hash, __cookie, __saddr, __daddr, __ports, __dif)\ (((__sk)->sk_hash == (__hash)) && \ (inet_sk(__sk)->daddr == (__saddr)) && \ (inet_sk(__sk)->rcv_saddr == (__daddr)) && \ (!((__sk)->sk_bound_dev_if) || ((__sk)->sk_bound_dev_if == (__dif)))) - Adds a prefetch(head->chain.first) in __inet_lookup_established()/__tcp_v4_check_established() and __inet6_lookup_established()/__tcp_v6_check_established() and __dccp_v4_check_established() to bring into cache the first element of the list, before the {read|write}_lock(&head->lock); Signed-off-by: Eric Dumazet Acked-by: Arnaldo Carvalho de Melo Signed-off-by: David S. Miller --- include/linux/ipv6.h | 5 +-- include/linux/tc_ematch/tc_em_meta.h | 2 +- include/net/inet6_hashtables.h | 21 ++++++------ include/net/inet_hashtables.h | 64 +++++++++++++++++++++--------------- include/net/inet_timewait_sock.h | 2 +- include/net/sock.h | 5 +-- net/atm/common.c | 2 +- net/dccp/ipv4.c | 12 +++---- net/ipv4/inet_timewait_sock.c | 6 ++-- net/ipv4/tcp_ipv4.c | 11 ++++--- net/ipv6/tcp_ipv6.c | 18 +++++----- net/sched/em_meta.c | 6 ++-- 12 files changed, 85 insertions(+), 69 deletions(-) (limited to 'include/net') diff --git a/include/linux/ipv6.h b/include/linux/ipv6.h index bb6f88e1406..e0b922785d9 100644 --- a/include/linux/ipv6.h +++ b/include/linux/ipv6.h @@ -372,8 +372,9 @@ static inline struct raw6_sock *raw6_sk(const struct sock *sk) #define inet_v6_ipv6only(__sk) 0 #endif /* defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) */ -#define INET6_MATCH(__sk, __saddr, __daddr, __ports, __dif) \ - (((*((__u32 *)&(inet_sk(__sk)->dport))) == (__ports)) && \ +#define INET6_MATCH(__sk, __hash, __saddr, __daddr, __ports, __dif)\ + (((__sk)->sk_hash == (__hash)) && \ + ((*((__u32 *)&(inet_sk(__sk)->dport))) == (__ports)) && \ ((__sk)->sk_family == AF_INET6) && \ ipv6_addr_equal(&inet6_sk(__sk)->daddr, (__saddr)) && \ ipv6_addr_equal(&inet6_sk(__sk)->rcv_saddr, (__daddr)) && \ diff --git a/include/linux/tc_ematch/tc_em_meta.h b/include/linux/tc_ematch/tc_em_meta.h index 081b1ee8516..e21937cf91d 100644 --- a/include/linux/tc_ematch/tc_em_meta.h +++ b/include/linux/tc_ematch/tc_em_meta.h @@ -71,7 +71,7 @@ enum TCF_META_ID_SK_SNDBUF, TCF_META_ID_SK_ALLOCS, TCF_META_ID_SK_ROUTE_CAPS, - TCF_META_ID_SK_HASHENT, + TCF_META_ID_SK_HASH, TCF_META_ID_SK_LINGERTIME, TCF_META_ID_SK_ACK_BACKLOG, TCF_META_ID_SK_MAX_ACK_BACKLOG, diff --git a/include/net/inet6_hashtables.h b/include/net/inet6_hashtables.h index 03df3b15796..5a2beed5a77 100644 --- a/include/net/inet6_hashtables.h +++ b/include/net/inet6_hashtables.h @@ -26,19 +26,18 @@ struct inet_hashinfo; /* I have no idea if this is a good hash for v6 or not. -DaveM */ -static inline int inet6_ehashfn(const struct in6_addr *laddr, const u16 lport, - const struct in6_addr *faddr, const u16 fport, - const int ehash_size) +static inline unsigned int inet6_ehashfn(const struct in6_addr *laddr, const u16 lport, + const struct in6_addr *faddr, const u16 fport) { - int hashent = (lport ^ fport); + unsigned int hashent = (lport ^ fport); hashent ^= (laddr->s6_addr32[3] ^ faddr->s6_addr32[3]); hashent ^= hashent >> 16; hashent ^= hashent >> 8; - return (hashent & (ehash_size - 1)); + return hashent; } -static inline int inet6_sk_ehashfn(const struct sock *sk, const int ehash_size) +static inline int inet6_sk_ehashfn(const struct sock *sk) { const struct inet_sock *inet = inet_sk(sk); const struct ipv6_pinfo *np = inet6_sk(sk); @@ -46,7 +45,7 @@ static inline int inet6_sk_ehashfn(const struct sock *sk, const int ehash_size) const struct in6_addr *faddr = &np->daddr; const __u16 lport = inet->num; const __u16 fport = inet->dport; - return inet6_ehashfn(laddr, lport, faddr, fport, ehash_size); + return inet6_ehashfn(laddr, lport, faddr, fport); } /* @@ -69,14 +68,14 @@ static inline struct sock * /* Optimize here for direct hit, only listening connections can * have wildcards anyways. */ - const int hash = inet6_ehashfn(daddr, hnum, saddr, sport, - hashinfo->ehash_size); - struct inet_ehash_bucket *head = &hashinfo->ehash[hash]; + unsigned int hash = inet6_ehashfn(daddr, hnum, saddr, sport); + struct inet_ehash_bucket *head = inet_ehash_bucket(hashinfo, hash); + prefetch(head->chain.first); read_lock(&head->lock); sk_for_each(sk, node, &head->chain) { /* For IPV6 do the cheaper port and family tests first. */ - if (INET6_MATCH(sk, saddr, daddr, ports, dif)) + if (INET6_MATCH(sk, hash, saddr, daddr, ports, dif)) goto hit; /* You sunk my battleship! */ } /* Must check for a TIME_WAIT'er before going to listener hash. */ diff --git a/include/net/inet_hashtables.h b/include/net/inet_hashtables.h index 646b6ea7fe2..35f49e65e29 100644 --- a/include/net/inet_hashtables.h +++ b/include/net/inet_hashtables.h @@ -108,7 +108,7 @@ struct inet_hashinfo { struct inet_bind_hashbucket *bhash; int bhash_size; - int ehash_size; + unsigned int ehash_size; /* All sockets in TCP_LISTEN state will be in here. This is the only * table where wildcard'd TCP sockets can exist. Hash function here @@ -130,17 +130,16 @@ struct inet_hashinfo { int port_rover; }; -static inline int inet_ehashfn(const __u32 laddr, const __u16 lport, - const __u32 faddr, const __u16 fport, - const int ehash_size) +static inline unsigned int inet_ehashfn(const __u32 laddr, const __u16 lport, + const __u32 faddr, const __u16 fport) { - int h = (laddr ^ lport) ^ (faddr ^ fport); + unsigned int h = (laddr ^ lport) ^ (faddr ^ fport); h ^= h >> 16; h ^= h >> 8; - return h & (ehash_size - 1); + return h; } -static inline int inet_sk_ehashfn(const struct sock *sk, const int ehash_size) +static inline int inet_sk_ehashfn(const struct sock *sk) { const struct inet_sock *inet = inet_sk(sk); const __u32 laddr = inet->rcv_saddr; @@ -148,7 +147,14 @@ static inline int inet_sk_ehashfn(const struct sock *sk, const int ehash_size) const __u32 faddr = inet->daddr; const __u16 fport = inet->dport; - return inet_ehashfn(laddr, lport, faddr, fport, ehash_size); + return inet_ehashfn(laddr, lport, faddr, fport); +} + +static inline struct inet_ehash_bucket *inet_ehash_bucket( + struct inet_hashinfo *hashinfo, + unsigned int hash) +{ + return &hashinfo->ehash[hash & (hashinfo->ehash_size - 1)]; } extern struct inet_bind_bucket * @@ -235,9 +241,11 @@ static inline void __inet_hash(struct inet_hashinfo *hashinfo, lock = &hashinfo->lhash_lock; inet_listen_wlock(hashinfo); } else { - sk->sk_hashent = inet_sk_ehashfn(sk, hashinfo->ehash_size); - list = &hashinfo->ehash[sk->sk_hashent].chain; - lock = &hashinfo->ehash[sk->sk_hashent].lock; + struct inet_ehash_bucket *head; + sk->sk_hash = inet_sk_ehashfn(sk); + head = inet_ehash_bucket(hashinfo, sk->sk_hash); + list = &head->chain; + lock = &head->lock; write_lock(lock); } __sk_add_node(sk, list); @@ -268,9 +276,8 @@ static inline void inet_unhash(struct inet_hashinfo *hashinfo, struct sock *sk) inet_listen_wlock(hashinfo); lock = &hashinfo->lhash_lock; } else { - struct inet_ehash_bucket *head = &hashinfo->ehash[sk->sk_hashent]; - lock = &head->lock; - write_lock_bh(&head->lock); + lock = &inet_ehash_bucket(hashinfo, sk->sk_hash)->lock; + write_lock_bh(lock); } if (__sk_del_node_init(sk)) @@ -337,23 +344,27 @@ sherry_cache: #define INET_ADDR_COOKIE(__name, __saddr, __daddr) \ const __u64 __name = (((__u64)(__daddr)) << 32) | ((__u64)(__saddr)); #endif /* __BIG_ENDIAN */ -#define INET_MATCH(__sk, __cookie, __saddr, __daddr, __ports, __dif)\ - (((*((__u64 *)&(inet_sk(__sk)->daddr))) == (__cookie)) && \ +#define INET_MATCH(__sk, __hash, __cookie, __saddr, __daddr, __ports, __dif)\ + (((__sk)->sk_hash == (__hash)) && \ + ((*((__u64 *)&(inet_sk(__sk)->daddr))) == (__cookie)) && \ ((*((__u32 *)&(inet_sk(__sk)->dport))) == (__ports)) && \ (!((__sk)->sk_bound_dev_if) || ((__sk)->sk_bound_dev_if == (__dif)))) -#define INET_TW_MATCH(__sk, __cookie, __saddr, __daddr, __ports, __dif)\ - (((*((__u64 *)&(inet_twsk(__sk)->tw_daddr))) == (__cookie)) && \ +#define INET_TW_MATCH(__sk, __hash, __cookie, __saddr, __daddr, __ports, __dif)\ + (((__sk)->sk_hash == (__hash)) && \ + ((*((__u64 *)&(inet_twsk(__sk)->tw_daddr))) == (__cookie)) && \ ((*((__u32 *)&(inet_twsk(__sk)->tw_dport))) == (__ports)) && \ (!((__sk)->sk_bound_dev_if) || ((__sk)->sk_bound_dev_if == (__dif)))) #else /* 32-bit arch */ #define INET_ADDR_COOKIE(__name, __saddr, __daddr) -#define INET_MATCH(__sk, __cookie, __saddr, __daddr, __ports, __dif) \ - ((inet_sk(__sk)->daddr == (__saddr)) && \ +#define INET_MATCH(__sk, __hash, __cookie, __saddr, __daddr, __ports, __dif) \ + (((__sk)->sk_hash == (__hash)) && \ + (inet_sk(__sk)->daddr == (__saddr)) && \ (inet_sk(__sk)->rcv_saddr == (__daddr)) && \ ((*((__u32 *)&(inet_sk(__sk)->dport))) == (__ports)) && \ (!((__sk)->sk_bound_dev_if) || ((__sk)->sk_bound_dev_if == (__dif)))) -#define INET_TW_MATCH(__sk, __cookie, __saddr, __daddr, __ports, __dif) \ - ((inet_twsk(__sk)->tw_daddr == (__saddr)) && \ +#define INET_TW_MATCH(__sk, __hash,__cookie, __saddr, __daddr, __ports, __dif) \ + (((__sk)->sk_hash == (__hash)) && \ + (inet_twsk(__sk)->tw_daddr == (__saddr)) && \ (inet_twsk(__sk)->tw_rcv_saddr == (__daddr)) && \ ((*((__u32 *)&(inet_twsk(__sk)->tw_dport))) == (__ports)) && \ (!((__sk)->sk_bound_dev_if) || ((__sk)->sk_bound_dev_if == (__dif)))) @@ -378,18 +389,19 @@ static inline struct sock * /* Optimize here for direct hit, only listening connections can * have wildcards anyways. */ - const int hash = inet_ehashfn(daddr, hnum, saddr, sport, hashinfo->ehash_size); - struct inet_ehash_bucket *head = &hashinfo->ehash[hash]; + unsigned int hash = inet_ehashfn(daddr, hnum, saddr, sport); + struct inet_ehash_bucket *head = inet_ehash_bucket(hashinfo, hash); + prefetch(head->chain.first); read_lock(&head->lock); sk_for_each(sk, node, &head->chain) { - if (INET_MATCH(sk, acookie, saddr, daddr, ports, dif)) + if (INET_MATCH(sk, hash, acookie, saddr, daddr, ports, dif)) goto hit; /* You sunk my battleship! */ } /* Must check for a TIME_WAIT'er before going to listener hash. */ sk_for_each(sk, node, &(head + hashinfo->ehash_size)->chain) { - if (INET_TW_MATCH(sk, acookie, saddr, daddr, ports, dif)) + if (INET_TW_MATCH(sk, hash, acookie, saddr, daddr, ports, dif)) goto hit; } sk = NULL; diff --git a/include/net/inet_timewait_sock.h b/include/net/inet_timewait_sock.h index 3b070352e86..4ade56ef3a4 100644 --- a/include/net/inet_timewait_sock.h +++ b/include/net/inet_timewait_sock.h @@ -112,6 +112,7 @@ struct inet_timewait_sock { #define tw_node __tw_common.skc_node #define tw_bind_node __tw_common.skc_bind_node #define tw_refcnt __tw_common.skc_refcnt +#define tw_hash __tw_common.skc_hash #define tw_prot __tw_common.skc_prot volatile unsigned char tw_substate; /* 3 bits hole, try to pack */ @@ -126,7 +127,6 @@ struct inet_timewait_sock { /* And these are ours. */ __u8 tw_ipv6only:1; /* 31 bits hole, try to pack */ - int tw_hashent; int tw_timeout; unsigned long tw_ttd; struct inet_bind_bucket *tw_tb; diff --git a/include/net/sock.h b/include/net/sock.h index 8c48fbecb7c..b6440805c42 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -99,6 +99,7 @@ struct proto; * @skc_node: main hash linkage for various protocol lookup tables * @skc_bind_node: bind hash linkage for various protocol lookup tables * @skc_refcnt: reference count + * @skc_hash: hash value used with various protocol lookup tables * @skc_prot: protocol handlers inside a network family * * This is the minimal network layer representation of sockets, the header @@ -112,6 +113,7 @@ struct sock_common { struct hlist_node skc_node; struct hlist_node skc_bind_node; atomic_t skc_refcnt; + unsigned int skc_hash; struct proto *skc_prot; }; @@ -139,7 +141,6 @@ struct sock_common { * @sk_no_check: %SO_NO_CHECK setting, wether or not checkup packets * @sk_route_caps: route capabilities (e.g. %NETIF_F_TSO) * @sk_lingertime: %SO_LINGER l_linger setting - * @sk_hashent: hash entry in several tables (e.g. inet_hashinfo.ehash) * @sk_backlog: always used with the per-socket spinlock held * @sk_callback_lock: used with the callbacks in the end of this struct * @sk_error_queue: rarely used @@ -186,6 +187,7 @@ struct sock { #define sk_node __sk_common.skc_node #define sk_bind_node __sk_common.skc_bind_node #define sk_refcnt __sk_common.skc_refcnt +#define sk_hash __sk_common.skc_hash #define sk_prot __sk_common.skc_prot unsigned char sk_shutdown : 2, sk_no_check : 2, @@ -208,7 +210,6 @@ struct sock { unsigned int sk_allocation; int sk_sndbuf; int sk_route_caps; - int sk_hashent; unsigned long sk_flags; unsigned long sk_lingertime; /* diff --git a/net/atm/common.c b/net/atm/common.c index 801a5813ec6..63feea49fb1 100644 --- a/net/atm/common.c +++ b/net/atm/common.c @@ -46,7 +46,7 @@ static void __vcc_insert_socket(struct sock *sk) struct atm_vcc *vcc = atm_sk(sk); struct hlist_head *head = &vcc_hash[vcc->vci & (VCC_HTABLE_SIZE - 1)]; - sk->sk_hashent = vcc->vci & (VCC_HTABLE_SIZE - 1); + sk->sk_hash = vcc->vci & (VCC_HTABLE_SIZE - 1); sk_add_node(sk, head); } diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index 40fe6afacde..ae088d1347a 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -62,27 +62,27 @@ static int __dccp_v4_check_established(struct sock *sk, const __u16 lport, const int dif = sk->sk_bound_dev_if; INET_ADDR_COOKIE(acookie, saddr, daddr) const __u32 ports = INET_COMBINED_PORTS(inet->dport, lport); - const int hash = inet_ehashfn(daddr, lport, saddr, inet->dport, - dccp_hashinfo.ehash_size); - struct inet_ehash_bucket *head = &dccp_hashinfo.ehash[hash]; + unsigned int hash = inet_ehashfn(daddr, lport, saddr, inet->dport); + struct inet_ehash_bucket *head = inet_ehash_bucket(&dccp_hashinfo, hash); const struct sock *sk2; const struct hlist_node *node; struct inet_timewait_sock *tw; + prefetch(head->chain.first); write_lock(&head->lock); /* Check TIME-WAIT sockets first. */ sk_for_each(sk2, node, &(head + dccp_hashinfo.ehash_size)->chain) { tw = inet_twsk(sk2); - if (INET_TW_MATCH(sk2, acookie, saddr, daddr, ports, dif)) + if (INET_TW_MATCH(sk2, hash, acookie, saddr, daddr, ports, dif)) goto not_unique; } tw = NULL; /* And established part... */ sk_for_each(sk2, node, &head->chain) { - if (INET_MATCH(sk2, acookie, saddr, daddr, ports, dif)) + if (INET_MATCH(sk2, hash, acookie, saddr, daddr, ports, dif)) goto not_unique; } @@ -90,7 +90,7 @@ static int __dccp_v4_check_established(struct sock *sk, const __u16 lport, * in hash table socket with a funny identity. */ inet->num = lport; inet->sport = htons(lport); - sk->sk_hashent = hash; + sk->sk_hash = hash; BUG_TRAP(sk_unhashed(sk)); __sk_add_node(sk, &head->chain); sock_prot_inc_use(sk->sk_prot); diff --git a/net/ipv4/inet_timewait_sock.c b/net/ipv4/inet_timewait_sock.c index 4d1502a4985..f9076ef3a1a 100644 --- a/net/ipv4/inet_timewait_sock.c +++ b/net/ipv4/inet_timewait_sock.c @@ -20,7 +20,7 @@ void __inet_twsk_kill(struct inet_timewait_sock *tw, struct inet_hashinfo *hashi struct inet_bind_hashbucket *bhead; struct inet_bind_bucket *tb; /* Unlink from established hashes. */ - struct inet_ehash_bucket *ehead = &hashinfo->ehash[tw->tw_hashent]; + struct inet_ehash_bucket *ehead = inet_ehash_bucket(hashinfo, tw->tw_hash); write_lock(&ehead->lock); if (hlist_unhashed(&tw->tw_node)) { @@ -60,7 +60,7 @@ void __inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk, { const struct inet_sock *inet = inet_sk(sk); const struct inet_connection_sock *icsk = inet_csk(sk); - struct inet_ehash_bucket *ehead = &hashinfo->ehash[sk->sk_hashent]; + struct inet_ehash_bucket *ehead = inet_ehash_bucket(hashinfo, sk->sk_hash); struct inet_bind_hashbucket *bhead; /* Step 1: Put TW into bind hash. Original socket stays there too. Note, that any socket with inet->num != 0 MUST be bound in @@ -106,7 +106,7 @@ struct inet_timewait_sock *inet_twsk_alloc(const struct sock *sk, const int stat tw->tw_dport = inet->dport; tw->tw_family = sk->sk_family; tw->tw_reuse = sk->sk_reuse; - tw->tw_hashent = sk->sk_hashent; + tw->tw_hash = sk->sk_hash; tw->tw_ipv6only = 0; tw->tw_prot = sk->sk_prot_creator; atomic_set(&tw->tw_refcnt, 1); diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 13dfb391cdf..c85819d8474 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -130,19 +130,20 @@ static int __tcp_v4_check_established(struct sock *sk, __u16 lport, int dif = sk->sk_bound_dev_if; INET_ADDR_COOKIE(acookie, saddr, daddr) const __u32 ports = INET_COMBINED_PORTS(inet->dport, lport); - const int hash = inet_ehashfn(daddr, lport, saddr, inet->dport, tcp_hashinfo.ehash_size); - struct inet_ehash_bucket *head = &tcp_hashinfo.ehash[hash]; + unsigned int hash = inet_ehashfn(daddr, lport, saddr, inet->dport); + struct inet_ehash_bucket *head = inet_ehash_bucket(&tcp_hashinfo, hash); struct sock *sk2; const struct hlist_node *node; struct inet_timewait_sock *tw; + prefetch(head->chain.first); write_lock(&head->lock); /* Check TIME-WAIT sockets first. */ sk_for_each(sk2, node, &(head + tcp_hashinfo.ehash_size)->chain) { tw = inet_twsk(sk2); - if (INET_TW_MATCH(sk2, acookie, saddr, daddr, ports, dif)) { + if (INET_TW_MATCH(sk2, hash, acookie, saddr, daddr, ports, dif)) { const struct tcp_timewait_sock *tcptw = tcp_twsk(sk2); struct tcp_sock *tp = tcp_sk(sk); @@ -179,7 +180,7 @@ static int __tcp_v4_check_established(struct sock *sk, __u16 lport, /* And established part... */ sk_for_each(sk2, node, &head->chain) { - if (INET_MATCH(sk2, acookie, saddr, daddr, ports, dif)) + if (INET_MATCH(sk2, hash, acookie, saddr, daddr, ports, dif)) goto not_unique; } @@ -188,7 +189,7 @@ unique: * in hash table socket with a funny identity. */ inet->num = lport; inet->sport = htons(lport); - sk->sk_hashent = hash; + sk->sk_hash = hash; BUG_TRAP(sk_unhashed(sk)); __sk_add_node(sk, &head->chain); sock_prot_inc_use(sk->sk_prot); diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 80643e6b346..d693cb988b7 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -209,9 +209,11 @@ static __inline__ void __tcp_v6_hash(struct sock *sk) lock = &tcp_hashinfo.lhash_lock; inet_listen_wlock(&tcp_hashinfo); } else { - sk->sk_hashent = inet6_sk_ehashfn(sk, tcp_hashinfo.ehash_size); - list = &tcp_hashinfo.ehash[sk->sk_hashent].chain; - lock = &tcp_hashinfo.ehash[sk->sk_hashent].lock; + unsigned int hash; + sk->sk_hash = hash = inet6_sk_ehashfn(sk); + hash &= (tcp_hashinfo.ehash_size - 1); + list = &tcp_hashinfo.ehash[hash].chain; + lock = &tcp_hashinfo.ehash[hash].lock; write_lock(lock); } @@ -322,13 +324,13 @@ static int __tcp_v6_check_established(struct sock *sk, const __u16 lport, const struct in6_addr *saddr = &np->daddr; const int dif = sk->sk_bound_dev_if; const u32 ports = INET_COMBINED_PORTS(inet->dport, lport); - const int hash = inet6_ehashfn(daddr, inet->num, saddr, inet->dport, - tcp_hashinfo.ehash_size); - struct inet_ehash_bucket *head = &tcp_hashinfo.ehash[hash]; + unsigned int hash = inet6_ehashfn(daddr, inet->num, saddr, inet->dport); + struct inet_ehash_bucket *head = inet_ehash_bucket(&tcp_hashinfo, hash); struct sock *sk2; const struct hlist_node *node; struct inet_timewait_sock *tw; + prefetch(head->chain.first); write_lock(&head->lock); /* Check TIME-WAIT sockets first. */ @@ -365,14 +367,14 @@ static int __tcp_v6_check_established(struct sock *sk, const __u16 lport, /* And established part... */ sk_for_each(sk2, node, &head->chain) { - if (INET6_MATCH(sk2, saddr, daddr, ports, dif)) + if (INET6_MATCH(sk2, hash, saddr, daddr, ports, dif)) goto not_unique; } unique: BUG_TRAP(sk_unhashed(sk)); __sk_add_node(sk, &head->chain); - sk->sk_hashent = hash; + sk->sk_hash = hash; sock_prot_inc_use(sk->sk_prot); write_unlock(&head->lock); diff --git a/net/sched/em_meta.c b/net/sched/em_meta.c index 00eae5f9a01..cf68a59fdc5 100644 --- a/net/sched/em_meta.c +++ b/net/sched/em_meta.c @@ -393,10 +393,10 @@ META_COLLECTOR(int_sk_route_caps) dst->value = skb->sk->sk_route_caps; } -META_COLLECTOR(int_sk_hashent) +META_COLLECTOR(int_sk_hash) { SKIP_NONLOCAL(skb); - dst->value = skb->sk->sk_hashent; + dst->value = skb->sk->sk_hash; } META_COLLECTOR(int_sk_lingertime) @@ -515,7 +515,7 @@ static struct meta_ops __meta_ops[TCF_META_TYPE_MAX+1][TCF_META_ID_MAX+1] = { [META_ID(SK_FORWARD_ALLOCS)] = META_FUNC(int_sk_fwd_alloc), [META_ID(SK_ALLOCS)] = META_FUNC(int_sk_alloc), [META_ID(SK_ROUTE_CAPS)] = META_FUNC(int_sk_route_caps), - [META_ID(SK_HASHENT)] = META_FUNC(int_sk_hashent), + [META_ID(SK_HASH)] = META_FUNC(int_sk_hash), [META_ID(SK_LINGERTIME)] = META_FUNC(int_sk_lingertime), [META_ID(SK_ACK_BACKLOG)] = META_FUNC(int_sk_ack_bl), [META_ID(SK_MAX_ACK_BACKLOG)] = META_FUNC(int_sk_max_ack_bl), -- cgit v1.2.3-70-g09d2