From e7219858ac1f98213a4714d0e24e7a003e1bf6a2 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki / 吉藤英明 Date: Sun, 13 Jan 2013 05:02:01 +0000 Subject: ipv6: Use ipv6_get_dsfield() instead of ipv6_tclass(). Commit 7a3198a8 ("ipv6: helper function to get tclass") introduced ipv6_tclass(), but similar function is already available as ipv6_get_dsfield(). We might be able to call ipv6_tclass() from ipv6_get_dsfield(), but it is confusing to have two versions. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- include/linux/ipv6.h | 5 ----- 1 file changed, 5 deletions(-) (limited to 'include/linux/ipv6.h') diff --git a/include/linux/ipv6.h b/include/linux/ipv6.h index faed1e357dd6..304a9f46b578 100644 --- a/include/linux/ipv6.h +++ b/include/linux/ipv6.h @@ -77,11 +77,6 @@ static inline struct ipv6hdr *ipipv6_hdr(const struct sk_buff *skb) return (struct ipv6hdr *)skb_transport_header(skb); } -static inline __u8 ipv6_tclass(const struct ipv6hdr *iph) -{ - return (ntohl(*(__be32 *)iph) >> 20) & 0xff; -} - /* This structure contains results of exthdrs parsing as offsets from skb->nh. -- cgit v1.2.3 From dd3332bfcb2223458f553f341d3388cb84040e6a Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki / 吉藤英明 Date: Sun, 13 Jan 2013 05:02:45 +0000 Subject: ipv6: Store Router Alert option in IP6CB directly. Router Alert option is very small and we can store the value itself in the skb. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- include/linux/ipv6.h | 3 ++- include/uapi/linux/ipv6.h | 2 ++ net/ipv6/exthdrs.c | 3 ++- net/ipv6/ip6_input.c | 5 ++--- 4 files changed, 8 insertions(+), 5 deletions(-) (limited to 'include/linux/ipv6.h') diff --git a/include/linux/ipv6.h b/include/linux/ipv6.h index 304a9f46b578..e971e3742172 100644 --- a/include/linux/ipv6.h +++ b/include/linux/ipv6.h @@ -84,7 +84,7 @@ static inline struct ipv6hdr *ipipv6_hdr(const struct sk_buff *skb) struct inet6_skb_parm { int iif; - __u16 ra; + __be16 ra; __u16 hop; __u16 dst0; __u16 srcrt; @@ -100,6 +100,7 @@ struct inet6_skb_parm { #define IP6SKB_XFRM_TRANSFORMED 1 #define IP6SKB_FORWARDED 2 #define IP6SKB_REROUTED 4 +#define IP6SKB_ROUTERALERT 8 }; #define IP6CB(skb) ((struct inet6_skb_parm*)((skb)->cb)) diff --git a/include/uapi/linux/ipv6.h b/include/uapi/linux/ipv6.h index 5a2991cf0251..4bda4cf5b0f5 100644 --- a/include/uapi/linux/ipv6.h +++ b/include/uapi/linux/ipv6.h @@ -63,6 +63,8 @@ struct ipv6_opt_hdr { #define ipv6_destopt_hdr ipv6_opt_hdr #define ipv6_hopopt_hdr ipv6_opt_hdr +/* Router Alert option values (RFC2711) */ +#define IPV6_OPT_ROUTERALERT_MLD 0x0000 /* MLD(RFC2710) */ /* * routing header type 0 (used in cmsghdr struct) diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c index 473f628f9f20..07a7d65a7cb6 100644 --- a/net/ipv6/exthdrs.c +++ b/net/ipv6/exthdrs.c @@ -553,7 +553,8 @@ static bool ipv6_hop_ra(struct sk_buff *skb, int optoff) const unsigned char *nh = skb_network_header(skb); if (nh[optoff + 1] == 2) { - IP6CB(skb)->ra = optoff; + IP6CB(skb)->flags |= IP6SKB_ROUTERALERT; + memcpy(&IP6CB(skb)->ra, nh + optoff + 2, sizeof(IP6CB(skb)->ra)); return true; } LIMIT_NETDEBUG(KERN_DEBUG "ipv6_hop_ra: wrong RA length %d\n", diff --git a/net/ipv6/ip6_input.c b/net/ipv6/ip6_input.c index 2ccd35ec3628..4ac5bf30e16a 100644 --- a/net/ipv6/ip6_input.c +++ b/net/ipv6/ip6_input.c @@ -280,9 +280,8 @@ int ip6_mc_input(struct sk_buff *skb) struct inet6_skb_parm *opt = IP6CB(skb); /* Check for MLD */ - if (unlikely(opt->ra)) { + if (unlikely(opt->flags & IP6SKB_ROUTERALERT)) { /* Check if this is a mld message */ - u8 *ptr = skb_network_header(skb) + opt->ra; u8 nexthdr = hdr->nexthdr; __be16 frag_off; int offset; @@ -290,7 +289,7 @@ int ip6_mc_input(struct sk_buff *skb) /* Check if the value of Router Alert * is for MLD (0x0000). */ - if ((ptr[2] | ptr[3]) == 0) { + if (opt->ra == htons(IPV6_OPT_ROUTERALERT_MLD)) { deliver = false; if (!ipv6_ext_hdr(nexthdr)) { -- cgit v1.2.3 From 18367681a10bd29c3f2305e6b7b984de5b33d548 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki / 吉藤英明 Date: Wed, 30 Jan 2013 09:27:52 +0000 Subject: ipv6 flowlabel: Convert np->ipv6_fl_list to RCU. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- include/linux/ipv6.h | 2 +- include/net/ipv6.h | 1 + net/ipv6/ip6_flowlabel.c | 72 +++++++++++++++++++++++++++--------------------- 3 files changed, 42 insertions(+), 33 deletions(-) (limited to 'include/linux/ipv6.h') diff --git a/include/linux/ipv6.h b/include/linux/ipv6.h index e971e3742172..850e95bc766c 100644 --- a/include/linux/ipv6.h +++ b/include/linux/ipv6.h @@ -214,7 +214,7 @@ struct ipv6_pinfo { struct ipv6_mc_socklist __rcu *ipv6_mc_list; struct ipv6_ac_socklist *ipv6_ac_list; - struct ipv6_fl_socklist *ipv6_fl_list; + struct ipv6_fl_socklist __rcu *ipv6_fl_list; struct ipv6_txoptions *opt; struct sk_buff *pktoptions; diff --git a/include/net/ipv6.h b/include/net/ipv6.h index 1d457161def2..851d5412a299 100644 --- a/include/net/ipv6.h +++ b/include/net/ipv6.h @@ -239,6 +239,7 @@ struct ip6_flowlabel { struct ipv6_fl_socklist { struct ipv6_fl_socklist *next; struct ip6_flowlabel *fl; + struct rcu_head rcu; }; extern struct ip6_flowlabel *fl6_sock_lookup(struct sock *sk, __be32 label); diff --git a/net/ipv6/ip6_flowlabel.c b/net/ipv6/ip6_flowlabel.c index da156015d827..22494afd981c 100644 --- a/net/ipv6/ip6_flowlabel.c +++ b/net/ipv6/ip6_flowlabel.c @@ -62,7 +62,7 @@ static DEFINE_SPINLOCK(ip6_fl_lock); /* Big socket sock */ -static DEFINE_RWLOCK(ip6_sk_fl_lock); +static DEFINE_SPINLOCK(ip6_sk_fl_lock); #define for_each_fl_rcu(hash, fl) \ for (fl = rcu_dereference(fl_ht[(hash)]); \ @@ -73,6 +73,11 @@ static DEFINE_RWLOCK(ip6_sk_fl_lock); fl != NULL; \ fl = rcu_dereference(fl->next)) +#define for_each_sk_fl_rcu(np, sfl) \ + for (sfl = rcu_dereference_bh(np->ipv6_fl_list); \ + sfl != NULL; \ + sfl = rcu_dereference_bh(sfl->next)) + static inline struct ip6_flowlabel *__fl_lookup(struct net *net, __be32 label) { struct ip6_flowlabel *fl; @@ -244,17 +249,17 @@ struct ip6_flowlabel * fl6_sock_lookup(struct sock *sk, __be32 label) label &= IPV6_FLOWLABEL_MASK; - read_lock_bh(&ip6_sk_fl_lock); - for (sfl=np->ipv6_fl_list; sfl; sfl = sfl->next) { + rcu_read_lock_bh(); + for_each_sk_fl_rcu(np, sfl) { struct ip6_flowlabel *fl = sfl->fl; if (fl->label == label) { fl->lastuse = jiffies; atomic_inc(&fl->users); - read_unlock_bh(&ip6_sk_fl_lock); + rcu_read_unlock_bh(); return fl; } } - read_unlock_bh(&ip6_sk_fl_lock); + rcu_read_unlock_bh(); return NULL; } @@ -265,20 +270,21 @@ void fl6_free_socklist(struct sock *sk) struct ipv6_pinfo *np = inet6_sk(sk); struct ipv6_fl_socklist *sfl; - if (!np->ipv6_fl_list) + if (!rcu_access_pointer(np->ipv6_fl_list)) return; - write_lock_bh(&ipv6_sk_fl_lock); - sfl = np->ipv6_fl_list; - np->ipv6_fl_list = NULL; - write_unlock_bh(&ipv6_sk_fl_lock); + spin_lock_bh(&ip6_sk_fl_lock); + while ((sfl = rcu_dereference_protected(np->ipv6_fl_list, + lockdep_is_held(&ip6_sk_fl_lock))) != NULL) { + np->ipv6_fl_list = sfl->next; + spin_unlock_bh(&ip6_sk_fl_lock); - while (sfl) { - struct ipv6_fl_socklist *next = sfl->next; fl_release(sfl->fl); - kfree(sfl); - sfl = next; + kfree_rcu(sfl, rcu); + + spin_lock_bh(&ip6_sk_fl_lock); } + spin_unlock_bh(&ip6_sk_fl_lock); } /* Service routines */ @@ -443,7 +449,7 @@ static int mem_check(struct sock *sk) if (room > FL_MAX_SIZE - FL_MAX_PER_SOCK) return 0; - for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next) + for_each_sk_fl_rcu(np, sfl) count++; if (room <= 0 || @@ -486,11 +492,11 @@ static bool ipv6_opt_cmp(struct ipv6_txoptions *o1, struct ipv6_txoptions *o2) static inline void fl_link(struct ipv6_pinfo *np, struct ipv6_fl_socklist *sfl, struct ip6_flowlabel *fl) { - write_lock_bh(&ip6_sk_fl_lock); + spin_lock_bh(&ip6_sk_fl_lock); sfl->fl = fl; sfl->next = np->ipv6_fl_list; - np->ipv6_fl_list = sfl; - write_unlock_bh(&ip6_sk_fl_lock); + rcu_assign_pointer(np->ipv6_fl_list, sfl); + spin_unlock_bh(&ip6_sk_fl_lock); } int ipv6_flowlabel_opt(struct sock *sk, char __user *optval, int optlen) @@ -512,31 +518,33 @@ int ipv6_flowlabel_opt(struct sock *sk, char __user *optval, int optlen) switch (freq.flr_action) { case IPV6_FL_A_PUT: - write_lock_bh(&ip6_sk_fl_lock); - for (sflp = &np->ipv6_fl_list; (sfl=*sflp)!=NULL; sflp = &sfl->next) { + spin_lock_bh(&ip6_sk_fl_lock); + for (sflp = &np->ipv6_fl_list; + (sfl = rcu_dereference(*sflp))!=NULL; + sflp = &sfl->next) { if (sfl->fl->label == freq.flr_label) { if (freq.flr_label == (np->flow_label&IPV6_FLOWLABEL_MASK)) np->flow_label &= ~IPV6_FLOWLABEL_MASK; - *sflp = sfl->next; - write_unlock_bh(&ip6_sk_fl_lock); + *sflp = rcu_dereference(sfl->next); + spin_unlock_bh(&ip6_sk_fl_lock); fl_release(sfl->fl); - kfree(sfl); + kfree_rcu(sfl, rcu); return 0; } } - write_unlock_bh(&ip6_sk_fl_lock); + spin_unlock_bh(&ip6_sk_fl_lock); return -ESRCH; case IPV6_FL_A_RENEW: - read_lock_bh(&ip6_sk_fl_lock); - for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next) { + rcu_read_lock_bh(); + for_each_sk_fl_rcu(np, sfl) { if (sfl->fl->label == freq.flr_label) { err = fl6_renew(sfl->fl, freq.flr_linger, freq.flr_expires); - read_unlock_bh(&ip6_sk_fl_lock); + rcu_read_unlock_bh(); return err; } } - read_unlock_bh(&ip6_sk_fl_lock); + rcu_read_unlock_bh(); if (freq.flr_share == IPV6_FL_S_NONE && ns_capable(net->user_ns, CAP_NET_ADMIN)) { @@ -560,11 +568,11 @@ int ipv6_flowlabel_opt(struct sock *sk, char __user *optval, int optlen) if (freq.flr_label) { err = -EEXIST; - read_lock_bh(&ip6_sk_fl_lock); - for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next) { + rcu_read_lock_bh(); + for_each_sk_fl_rcu(np, sfl) { if (sfl->fl->label == freq.flr_label) { if (freq.flr_flags&IPV6_FL_F_EXCL) { - read_unlock_bh(&ip6_sk_fl_lock); + rcu_read_unlock_bh(); goto done; } fl1 = sfl->fl; @@ -572,7 +580,7 @@ int ipv6_flowlabel_opt(struct sock *sk, char __user *optval, int optlen) break; } } - read_unlock_bh(&ip6_sk_fl_lock); + rcu_read_unlock_bh(); if (fl1 == NULL) fl1 = fl_lookup(net, freq.flr_label); -- cgit v1.2.3