X-Git-Url: https://git.openpandora.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=net%2Fipv6%2Freassembly.c;h=0a1bf43bd489e8fca1ffe91f8698a7b0ddff68c3;hb=fd9e63544cac30a34c951f0ec958038f0529e244;hp=de795c04e34c1862723eb8100a8f376a5b375ebb;hpb=d6454706c382ab74e2ecad7803c434cc6bd30343;p=pandora-kernel.git diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c index de795c04e34c..0a1bf43bd489 100644 --- a/net/ipv6/reassembly.c +++ b/net/ipv6/reassembly.c @@ -42,6 +42,7 @@ #include #include #include +#include #include #include @@ -53,11 +54,7 @@ #include #include #include - -int sysctl_ip6frag_high_thresh __read_mostly = 256*1024; -int sysctl_ip6frag_low_thresh __read_mostly = 192*1024; - -int sysctl_ip6frag_time __read_mostly = IPV6_FRAG_TIMEOUT; +#include struct ip6frag_skb_cb { @@ -74,53 +71,39 @@ struct ip6frag_skb_cb struct frag_queue { - struct hlist_node list; - struct list_head lru_list; /* lru list member */ + struct inet_frag_queue q; __be32 id; /* fragment id */ struct in6_addr saddr; struct in6_addr daddr; - spinlock_t lock; - atomic_t refcnt; - struct timer_list timer; /* expire timer */ - struct sk_buff *fragments; - int len; - int meat; int iif; - ktime_t stamp; unsigned int csum; - __u8 last_in; /* has first/last segment arrived? */ -#define COMPLETE 4 -#define FIRST_IN 2 -#define LAST_IN 1 __u16 nhoffset; }; -/* Hash table. */ - -#define IP6Q_HASHSZ 64 +struct inet_frags_ctl ip6_frags_ctl __read_mostly = { + .high_thresh = 256 * 1024, + .low_thresh = 192 * 1024, + .timeout = IPV6_FRAG_TIMEOUT, + .secret_interval = 10 * 60 * HZ, +}; -static struct hlist_head ip6_frag_hash[IP6Q_HASHSZ]; -static DEFINE_RWLOCK(ip6_frag_lock); -static u32 ip6_frag_hash_rnd; -static LIST_HEAD(ip6_frag_lru_list); -int ip6_frag_nqueues = 0; +static struct inet_frags ip6_frags; -static __inline__ void __fq_unlink(struct frag_queue *fq) +int ip6_frag_nqueues(void) { - hlist_del(&fq->list); - list_del(&fq->lru_list); - ip6_frag_nqueues--; + return ip6_frags.nqueues; } -static __inline__ void fq_unlink(struct frag_queue *fq) +int ip6_frag_mem(void) { - write_lock(&ip6_frag_lock); - __fq_unlink(fq); - write_unlock(&ip6_frag_lock); + return atomic_read(&ip6_frags.mem); } +static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff *prev, + struct net_device *dev); + /* * callers should be careful not to use the hash value outside the ipfrag_lock * as doing so could race with ipfrag_hash_rnd being recalculated. @@ -136,7 +119,7 @@ static unsigned int ip6qhashfn(__be32 id, struct in6_addr *saddr, a += JHASH_GOLDEN_RATIO; b += JHASH_GOLDEN_RATIO; - c += ip6_frag_hash_rnd; + c += ip6_frags.rnd; __jhash_mix(a, b, c); a += (__force u32)saddr->s6_addr32[3]; @@ -149,60 +132,29 @@ static unsigned int ip6qhashfn(__be32 id, struct in6_addr *saddr, c += (__force u32)id; __jhash_mix(a, b, c); - return c & (IP6Q_HASHSZ - 1); + return c & (INETFRAGS_HASHSZ - 1); } -static struct timer_list ip6_frag_secret_timer; -int sysctl_ip6frag_secret_interval __read_mostly = 10 * 60 * HZ; - -static void ip6_frag_secret_rebuild(unsigned long dummy) +static unsigned int ip6_hashfn(struct inet_frag_queue *q) { - unsigned long now = jiffies; - int i; - - write_lock(&ip6_frag_lock); - get_random_bytes(&ip6_frag_hash_rnd, sizeof(u32)); - for (i = 0; i < IP6Q_HASHSZ; i++) { - struct frag_queue *q; - struct hlist_node *p, *n; - - hlist_for_each_entry_safe(q, p, n, &ip6_frag_hash[i], list) { - unsigned int hval = ip6qhashfn(q->id, - &q->saddr, - &q->daddr); - - if (hval != i) { - hlist_del(&q->list); - - /* Relink to new hash chain. */ - hlist_add_head(&q->list, - &ip6_frag_hash[hval]); - - } - } - } - write_unlock(&ip6_frag_lock); + struct frag_queue *fq; - mod_timer(&ip6_frag_secret_timer, now + sysctl_ip6frag_secret_interval); + fq = container_of(q, struct frag_queue, q); + return ip6qhashfn(fq->id, &fq->saddr, &fq->daddr); } -atomic_t ip6_frag_mem = ATOMIC_INIT(0); - /* Memory Tracking Functions. */ static inline void frag_kfree_skb(struct sk_buff *skb, int *work) { if (work) *work -= skb->truesize; - atomic_sub(skb->truesize, &ip6_frag_mem); + atomic_sub(skb->truesize, &ip6_frags.mem); kfree_skb(skb); } -static inline void frag_free_queue(struct frag_queue *fq, int *work) +static void ip6_frag_free(struct inet_frag_queue *fq) { - if (work) - *work -= sizeof(struct frag_queue); - atomic_sub(sizeof(struct frag_queue), &ip6_frag_mem); - kfree(fq); + kfree(container_of(fq, struct frag_queue, q)); } static inline struct frag_queue *frag_alloc_queue(void) @@ -211,36 +163,15 @@ static inline struct frag_queue *frag_alloc_queue(void) if(!fq) return NULL; - atomic_add(sizeof(struct frag_queue), &ip6_frag_mem); + atomic_add(sizeof(struct frag_queue), &ip6_frags.mem); return fq; } /* Destruction primitives. */ -/* Complete destruction of fq. */ -static void ip6_frag_destroy(struct frag_queue *fq, int *work) -{ - struct sk_buff *fp; - - BUG_TRAP(fq->last_in&COMPLETE); - BUG_TRAP(del_timer(&fq->timer) == 0); - - /* Release all fragment data. */ - fp = fq->fragments; - while (fp) { - struct sk_buff *xp = fp->next; - - frag_kfree_skb(fp, work); - fp = xp; - } - - frag_free_queue(fq, work); -} - -static __inline__ void fq_put(struct frag_queue *fq, int *work) +static __inline__ void fq_put(struct frag_queue *fq) { - if (atomic_dec_and_test(&fq->refcnt)) - ip6_frag_destroy(fq, work); + inet_frag_put(&fq->q, &ip6_frags); } /* Kill fq entry. It is not destroyed immediately, @@ -248,45 +179,16 @@ static __inline__ void fq_put(struct frag_queue *fq, int *work) */ static __inline__ void fq_kill(struct frag_queue *fq) { - if (del_timer(&fq->timer)) - atomic_dec(&fq->refcnt); - - if (!(fq->last_in & COMPLETE)) { - fq_unlink(fq); - atomic_dec(&fq->refcnt); - fq->last_in |= COMPLETE; - } + inet_frag_kill(&fq->q, &ip6_frags); } static void ip6_evictor(struct inet6_dev *idev) { - struct frag_queue *fq; - struct list_head *tmp; - int work; - - work = atomic_read(&ip6_frag_mem) - sysctl_ip6frag_low_thresh; - if (work <= 0) - return; - - while(work > 0) { - read_lock(&ip6_frag_lock); - if (list_empty(&ip6_frag_lru_list)) { - read_unlock(&ip6_frag_lock); - return; - } - tmp = ip6_frag_lru_list.next; - fq = list_entry(tmp, struct frag_queue, lru_list); - atomic_inc(&fq->refcnt); - read_unlock(&ip6_frag_lock); - - spin_lock(&fq->lock); - if (!(fq->last_in&COMPLETE)) - fq_kill(fq); - spin_unlock(&fq->lock); - - fq_put(fq, &work); - IP6_INC_STATS_BH(idev, IPSTATS_MIB_REASMFAILS); - } + int evicted; + + evicted = inet_frag_evictor(&ip6_frags); + if (evicted) + IP6_ADD_STATS_BH(idev, IPSTATS_MIB_REASMFAILS, evicted); } static void ip6_frag_expire(unsigned long data) @@ -294,14 +196,14 @@ static void ip6_frag_expire(unsigned long data) struct frag_queue *fq = (struct frag_queue *) data; struct net_device *dev = NULL; - spin_lock(&fq->lock); + spin_lock(&fq->q.lock); - if (fq->last_in & COMPLETE) + if (fq->q.last_in & COMPLETE) goto out; fq_kill(fq); - dev = dev_get_by_index(fq->iif); + dev = dev_get_by_index(&init_net, fq->iif); if (!dev) goto out; @@ -311,7 +213,7 @@ static void ip6_frag_expire(unsigned long data) rcu_read_unlock(); /* Don't send error if the first segment did not arrive. */ - if (!(fq->last_in&FIRST_IN) || !fq->fragments) + if (!(fq->q.last_in&FIRST_IN) || !fq->q.fragments) goto out; /* @@ -319,59 +221,58 @@ static void ip6_frag_expire(unsigned long data) segment was received. And do not use fq->dev pointer directly, device might already disappeared. */ - fq->fragments->dev = dev; - icmpv6_send(fq->fragments, ICMPV6_TIME_EXCEED, ICMPV6_EXC_FRAGTIME, 0, dev); + fq->q.fragments->dev = dev; + icmpv6_send(fq->q.fragments, ICMPV6_TIME_EXCEED, ICMPV6_EXC_FRAGTIME, 0, dev); out: if (dev) dev_put(dev); - spin_unlock(&fq->lock); - fq_put(fq, NULL); + spin_unlock(&fq->q.lock); + fq_put(fq); } /* Creation primitives. */ -static struct frag_queue *ip6_frag_intern(struct frag_queue *fq_in) +static struct frag_queue *ip6_frag_intern(struct frag_queue *fq_in, + unsigned int hash) { struct frag_queue *fq; - unsigned int hash; #ifdef CONFIG_SMP struct hlist_node *n; #endif - write_lock(&ip6_frag_lock); - hash = ip6qhashfn(fq_in->id, &fq_in->saddr, &fq_in->daddr); + write_lock(&ip6_frags.lock); #ifdef CONFIG_SMP - hlist_for_each_entry(fq, n, &ip6_frag_hash[hash], list) { + hlist_for_each_entry(fq, n, &ip6_frags.hash[hash], q.list) { if (fq->id == fq_in->id && ipv6_addr_equal(&fq_in->saddr, &fq->saddr) && ipv6_addr_equal(&fq_in->daddr, &fq->daddr)) { - atomic_inc(&fq->refcnt); - write_unlock(&ip6_frag_lock); - fq_in->last_in |= COMPLETE; - fq_put(fq_in, NULL); + atomic_inc(&fq->q.refcnt); + write_unlock(&ip6_frags.lock); + fq_in->q.last_in |= COMPLETE; + fq_put(fq_in); return fq; } } #endif fq = fq_in; - if (!mod_timer(&fq->timer, jiffies + sysctl_ip6frag_time)) - atomic_inc(&fq->refcnt); + if (!mod_timer(&fq->q.timer, jiffies + ip6_frags_ctl.timeout)) + atomic_inc(&fq->q.refcnt); - atomic_inc(&fq->refcnt); - hlist_add_head(&fq->list, &ip6_frag_hash[hash]); - INIT_LIST_HEAD(&fq->lru_list); - list_add_tail(&fq->lru_list, &ip6_frag_lru_list); - ip6_frag_nqueues++; - write_unlock(&ip6_frag_lock); + atomic_inc(&fq->q.refcnt); + hlist_add_head(&fq->q.list, &ip6_frags.hash[hash]); + INIT_LIST_HEAD(&fq->q.lru_list); + list_add_tail(&fq->q.lru_list, &ip6_frags.lru_list); + ip6_frags.nqueues++; + write_unlock(&ip6_frags.lock); return fq; } static struct frag_queue * ip6_frag_create(__be32 id, struct in6_addr *src, struct in6_addr *dst, - struct inet6_dev *idev) + struct inet6_dev *idev, unsigned int hash) { struct frag_queue *fq; @@ -382,13 +283,13 @@ ip6_frag_create(__be32 id, struct in6_addr *src, struct in6_addr *dst, ipv6_addr_copy(&fq->saddr, src); ipv6_addr_copy(&fq->daddr, dst); - init_timer(&fq->timer); - fq->timer.function = ip6_frag_expire; - fq->timer.data = (long) fq; - spin_lock_init(&fq->lock); - atomic_set(&fq->refcnt, 1); + init_timer(&fq->q.timer); + fq->q.timer.function = ip6_frag_expire; + fq->q.timer.data = (long) fq; + spin_lock_init(&fq->q.lock); + atomic_set(&fq->q.refcnt, 1); - return ip6_frag_intern(fq); + return ip6_frag_intern(fq, hash); oom: IP6_INC_STATS_BH(idev, IPSTATS_MIB_REASMFAILS); @@ -403,30 +304,31 @@ fq_find(__be32 id, struct in6_addr *src, struct in6_addr *dst, struct hlist_node *n; unsigned int hash; - read_lock(&ip6_frag_lock); + read_lock(&ip6_frags.lock); hash = ip6qhashfn(id, src, dst); - hlist_for_each_entry(fq, n, &ip6_frag_hash[hash], list) { + hlist_for_each_entry(fq, n, &ip6_frags.hash[hash], q.list) { if (fq->id == id && ipv6_addr_equal(src, &fq->saddr) && ipv6_addr_equal(dst, &fq->daddr)) { - atomic_inc(&fq->refcnt); - read_unlock(&ip6_frag_lock); + atomic_inc(&fq->q.refcnt); + read_unlock(&ip6_frags.lock); return fq; } } - read_unlock(&ip6_frag_lock); + read_unlock(&ip6_frags.lock); - return ip6_frag_create(id, src, dst, idev); + return ip6_frag_create(id, src, dst, idev, hash); } -static void ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb, +static int ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb, struct frag_hdr *fhdr, int nhoff) { struct sk_buff *prev, *next; + struct net_device *dev; int offset, end; - if (fq->last_in & COMPLETE) + if (fq->q.last_in & COMPLETE) goto err; offset = ntohs(fhdr->frag_off) & ~0x7; @@ -439,7 +341,7 @@ static void ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb, icmpv6_param_prob(skb, ICMPV6_HDR_FIELD, ((u8 *)&fhdr->frag_off - skb_network_header(skb))); - return; + return -1; } if (skb->ip_summed == CHECKSUM_COMPLETE) { @@ -454,11 +356,11 @@ static void ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb, /* If we already have some bits beyond end * or have different end, the segment is corrupted. */ - if (end < fq->len || - ((fq->last_in & LAST_IN) && end != fq->len)) + if (end < fq->q.len || + ((fq->q.last_in & LAST_IN) && end != fq->q.len)) goto err; - fq->last_in |= LAST_IN; - fq->len = end; + fq->q.last_in |= LAST_IN; + fq->q.len = end; } else { /* Check if the fragment is rounded to 8 bytes. * Required by the RFC. @@ -471,13 +373,13 @@ static void ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb, IPSTATS_MIB_INHDRERRORS); icmpv6_param_prob(skb, ICMPV6_HDR_FIELD, offsetof(struct ipv6hdr, payload_len)); - return; + return -1; } - if (end > fq->len) { + if (end > fq->q.len) { /* Some bits beyond end -> corruption. */ - if (fq->last_in & LAST_IN) + if (fq->q.last_in & LAST_IN) goto err; - fq->len = end; + fq->q.len = end; } } @@ -496,7 +398,7 @@ static void ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb, * this fragment, right? */ prev = NULL; - for(next = fq->fragments; next != NULL; next = next->next) { + for(next = fq->q.fragments; next != NULL; next = next->next) { if (FRAG6_CB(next)->offset >= offset) break; /* bingo! */ prev = next; @@ -533,7 +435,7 @@ static void ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb, if (!pskb_pull(next, i)) goto err; FRAG6_CB(next)->offset += i; /* next fragment */ - fq->meat -= i; + fq->q.meat -= i; if (next->ip_summed != CHECKSUM_UNNECESSARY) next->ip_summed = CHECKSUM_NONE; break; @@ -548,9 +450,9 @@ static void ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb, if (prev) prev->next = next; else - fq->fragments = next; + fq->q.fragments = next; - fq->meat -= free_it->len; + fq->q.meat -= free_it->len; frag_kfree_skb(free_it, NULL); } } @@ -562,30 +464,37 @@ static void ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb, if (prev) prev->next = skb; else - fq->fragments = skb; + fq->q.fragments = skb; - if (skb->dev) - fq->iif = skb->dev->ifindex; - skb->dev = NULL; - fq->stamp = skb->tstamp; - fq->meat += skb->len; - atomic_add(skb->truesize, &ip6_frag_mem); + dev = skb->dev; + if (dev) { + fq->iif = dev->ifindex; + skb->dev = NULL; + } + fq->q.stamp = skb->tstamp; + fq->q.meat += skb->len; + atomic_add(skb->truesize, &ip6_frags.mem); /* The first fragment. * nhoffset is obtained from the first fragment, of course. */ if (offset == 0) { fq->nhoffset = nhoff; - fq->last_in |= FIRST_IN; + fq->q.last_in |= FIRST_IN; } - write_lock(&ip6_frag_lock); - list_move_tail(&fq->lru_list, &ip6_frag_lru_list); - write_unlock(&ip6_frag_lock); - return; + + if (fq->q.last_in == (FIRST_IN | LAST_IN) && fq->q.meat == fq->q.len) + return ip6_frag_reasm(fq, prev, dev); + + write_lock(&ip6_frags.lock); + list_move_tail(&fq->q.lru_list, &ip6_frags.lru_list); + write_unlock(&ip6_frags.lock); + return -1; err: IP6_INC_STATS(ip6_dst_idev(skb->dst), IPSTATS_MIB_REASMFAILS); kfree_skb(skb); + return -1; } /* @@ -597,21 +506,39 @@ err: * queue is eligible for reassembly i.e. it is not COMPLETE, * the last and the first frames arrived and all the bits are here. */ -static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff **skb_in, +static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff *prev, struct net_device *dev) { - struct sk_buff *fp, *head = fq->fragments; + struct sk_buff *fp, *head = fq->q.fragments; int payload_len; unsigned int nhoff; fq_kill(fq); + /* Make the one we just received the head. */ + if (prev) { + head = prev->next; + fp = skb_clone(head, GFP_ATOMIC); + + if (!fp) + goto out_oom; + + fp->next = head->next; + prev->next = fp; + + skb_morph(head, fq->q.fragments); + head->next = fq->q.fragments->next; + + kfree_skb(fq->q.fragments); + fq->q.fragments = head; + } + BUG_TRAP(head != NULL); BUG_TRAP(FRAG6_CB(head)->offset == 0); /* Unfragmented part is taken from the first segment. */ payload_len = ((head->data - skb_network_header(head)) - - sizeof(struct ipv6hdr) + fq->len - + sizeof(struct ipv6hdr) + fq->q.len - sizeof(struct frag_hdr)); if (payload_len > IPV6_MAXPLEN) goto out_oversize; @@ -640,7 +567,7 @@ static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff **skb_in, head->len -= clone->len; clone->csum = 0; clone->ip_summed = head->ip_summed; - atomic_add(clone->truesize, &ip6_frag_mem); + atomic_add(clone->truesize, &ip6_frags.mem); } /* We have to remove fragment header from datagram and to relocate @@ -655,7 +582,7 @@ static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff **skb_in, skb_shinfo(head)->frag_list = head->next; skb_reset_transport_header(head); skb_push(head, head->data - skb_network_header(head)); - atomic_sub(head->truesize, &ip6_frag_mem); + atomic_sub(head->truesize, &ip6_frags.mem); for (fp=head->next; fp; fp = fp->next) { head->data_len += fp->len; @@ -665,17 +592,15 @@ static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff **skb_in, else if (head->ip_summed == CHECKSUM_COMPLETE) head->csum = csum_add(head->csum, fp->csum); head->truesize += fp->truesize; - atomic_sub(fp->truesize, &ip6_frag_mem); + atomic_sub(fp->truesize, &ip6_frags.mem); } head->next = NULL; head->dev = dev; - head->tstamp = fq->stamp; + head->tstamp = fq->q.stamp; ipv6_hdr(head)->payload_len = htons(payload_len); IP6CB(head)->nhoff = nhoff; - *skb_in = head; - /* Yes, and fold redundant checksum back. 8) */ if (head->ip_summed == CHECKSUM_COMPLETE) head->csum = csum_partial(skb_network_header(head), @@ -685,7 +610,7 @@ static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff **skb_in, rcu_read_lock(); IP6_INC_STATS_BH(__in6_dev_get(dev), IPSTATS_MIB_REASMOKS); rcu_read_unlock(); - fq->fragments = NULL; + fq->q.fragments = NULL; return 1; out_oversize: @@ -702,10 +627,8 @@ out_fail: return -1; } -static int ipv6_frag_rcv(struct sk_buff **skbp) +static int ipv6_frag_rcv(struct sk_buff *skb) { - struct sk_buff *skb = *skbp; - struct net_device *dev = skb->dev; struct frag_hdr *fhdr; struct frag_queue *fq; struct ipv6hdr *hdr = ipv6_hdr(skb); @@ -739,23 +662,19 @@ static int ipv6_frag_rcv(struct sk_buff **skbp) return 1; } - if (atomic_read(&ip6_frag_mem) > sysctl_ip6frag_high_thresh) + if (atomic_read(&ip6_frags.mem) > ip6_frags_ctl.high_thresh) ip6_evictor(ip6_dst_idev(skb->dst)); if ((fq = fq_find(fhdr->identification, &hdr->saddr, &hdr->daddr, ip6_dst_idev(skb->dst))) != NULL) { - int ret = -1; - - spin_lock(&fq->lock); + int ret; - ip6_frag_queue(fq, skb, fhdr, IP6CB(skb)->nhoff); + spin_lock(&fq->q.lock); - if (fq->last_in == (FIRST_IN|LAST_IN) && - fq->meat == fq->len) - ret = ip6_frag_reasm(fq, skbp, dev); + ret = ip6_frag_queue(fq, skb, fhdr, IP6CB(skb)->nhoff); - spin_unlock(&fq->lock); - fq_put(fq, NULL); + spin_unlock(&fq->q.lock); + fq_put(fq); return ret; } @@ -775,11 +694,10 @@ void __init ipv6_frag_init(void) if (inet6_add_protocol(&frag_protocol, IPPROTO_FRAGMENT) < 0) printk(KERN_ERR "ipv6_frag_init: Could not register protocol\n"); - ip6_frag_hash_rnd = (u32) ((num_physpages ^ (num_physpages>>7)) ^ - (jiffies ^ (jiffies >> 6))); - - init_timer(&ip6_frag_secret_timer); - ip6_frag_secret_timer.function = ip6_frag_secret_rebuild; - ip6_frag_secret_timer.expires = jiffies + sysctl_ip6frag_secret_interval; - add_timer(&ip6_frag_secret_timer); + ip6_frags.ctl = &ip6_frags_ctl; + ip6_frags.hashfn = ip6_hashfn; + ip6_frags.destructor = ip6_frag_free; + ip6_frags.skb_free = NULL; + ip6_frags.qsize = sizeof(struct frag_queue); + inet_frags_init(&ip6_frags); }