fib_hash.c

《嵌入式系统设计与实例开发实验教材二源码》Linux内核移植与编译实验

		if (dst & ~FZ_MASK(fz))
			return -EINVAL;
		key = fz_key(dst, fz);
	}

	if  ((fi = fib_create_info(r, rta, n, &err)) == NULL)
		return err;

#ifdef CONFIG_IP_ROUTE_LARGE_TABLES
	if (fz->fz_nent > (fz->fz_divisor<<2) &&
	    fz->fz_divisor < FZ_MAX_DIVISOR &&
	    (z==32 || (1<<z) > fz->fz_divisor))
		fn_rehash_zone(fz);
#endif

	fp = fz_chain_p(key, fz);


	/*
	 * Scan list to find the first route with the same destination
	 */
	FIB_SCAN(f, fp) {
		if (fn_key_leq(key,f->fn_key))
			break;
	}

#ifdef CONFIG_IP_ROUTE_TOS
	/*
	 * Find route with the same destination and tos.
	 */
	FIB_SCAN_KEY(f, fp, key) {
		if (f->fn_tos <= tos)
			break;
	}
#endif

	del_fp = NULL;

	if (f && (f->fn_state&FN_S_ZOMBIE) &&
#ifdef CONFIG_IP_ROUTE_TOS
	    f->fn_tos == tos &&
#endif
	    fn_key_eq(f->fn_key, key)) {
		del_fp = fp;
		fp = &f->fn_next;
		f = *fp;
		goto create;
	}

	FIB_SCAN_TOS(f, fp, key, tos) {
		if (fi->fib_priority <= FIB_INFO(f)->fib_priority)
			break;
	}

	/* Now f==*fp points to the first node with the same
	   keys [prefix,tos,priority], if such key already
	   exists or to the node, before which we will insert new one.
	 */

	if (f && 
#ifdef CONFIG_IP_ROUTE_TOS
	    f->fn_tos == tos &&
#endif
	    fn_key_eq(f->fn_key, key) &&
	    fi->fib_priority == FIB_INFO(f)->fib_priority) {
		struct fib_node **ins_fp;

		err = -EEXIST;
		if (n->nlmsg_flags&NLM_F_EXCL)
			goto out;

		if (n->nlmsg_flags&NLM_F_REPLACE) {
			del_fp = fp;
			fp = &f->fn_next;
			f = *fp;
			goto replace;
		}

		ins_fp = fp;
		err = -EEXIST;

		FIB_SCAN_TOS(f, fp, key, tos) {
			if (fi->fib_priority != FIB_INFO(f)->fib_priority)
				break;
			if (f->fn_type == type && f->fn_scope == r->rtm_scope
			    && FIB_INFO(f) == fi)
				goto out;
		}

		if (!(n->nlmsg_flags&NLM_F_APPEND)) {
			fp = ins_fp;
			f = *fp;
		}
	}

create:
	err = -ENOENT;
	if (!(n->nlmsg_flags&NLM_F_CREATE))
		goto out;

replace:
	err = -ENOBUFS;
	new_f = kmem_cache_alloc(fn_hash_kmem, SLAB_KERNEL);
	if (new_f == NULL)
		goto out;

	memset(new_f, 0, sizeof(struct fib_node));

	new_f->fn_key = key;
#ifdef CONFIG_IP_ROUTE_TOS
	new_f->fn_tos = tos;
#endif
	new_f->fn_type = type;
	new_f->fn_scope = r->rtm_scope;
	FIB_INFO(new_f) = fi;

	/*
	 * Insert new entry to the list.
	 */

	new_f->fn_next = f;
	write_lock_bh(&fib_hash_lock);
	*fp = new_f;
	write_unlock_bh(&fib_hash_lock);
	fz->fz_nent++;

	if (del_fp) {
		f = *del_fp;
		/* Unlink replaced node */
		write_lock_bh(&fib_hash_lock);
		*del_fp = f->fn_next;
		write_unlock_bh(&fib_hash_lock);

		if (!(f->fn_state&FN_S_ZOMBIE))
			rtmsg_fib(RTM_DELROUTE, f, z, tb->tb_id, n, req);
		if (f->fn_state&FN_S_ACCESSED)
			rt_cache_flush(-1);
		fn_free_node(f);
		fz->fz_nent--;
	} else {
		rt_cache_flush(-1);
	}
	rtmsg_fib(RTM_NEWROUTE, new_f, z, tb->tb_id, n, req);
	return 0;

out:
	fib_release_info(fi);
	return err;
}


static int
fn_hash_delete(struct fib_table *tb, struct rtmsg *r, struct kern_rta *rta,
		struct nlmsghdr *n, struct netlink_skb_parms *req)
{
	struct fn_hash *table = (struct fn_hash*)tb->tb_data;
	struct fib_node **fp, **del_fp, *f;
	int z = r->rtm_dst_len;
	struct fn_zone *fz;
	fn_key_t key;
	int matched;
#ifdef CONFIG_IP_ROUTE_TOS
	u8 tos = r->rtm_tos;
#endif

FTprint("tb(%d)_delete: %d %08x/%d %d\n", tb->tb_id, r->rtm_type, rta->rta_dst ?
       *(u32*)rta->rta_dst : 0, z, rta->rta_oif ? *rta->rta_oif : -1);
	if (z > 32)
		return -EINVAL;
	if ((fz  = table->fn_zones[z]) == NULL)
		return -ESRCH;

	fz_key_0(key);
	if (rta->rta_dst) {
		u32 dst;
		memcpy(&dst, rta->rta_dst, 4);
		if (dst & ~FZ_MASK(fz))
			return -EINVAL;
		key = fz_key(dst, fz);
	}

	fp = fz_chain_p(key, fz);


	FIB_SCAN(f, fp) {
		if (fn_key_eq(f->fn_key, key))
			break;
		if (fn_key_leq(key, f->fn_key)) {
			return -ESRCH;
		}
	}
#ifdef CONFIG_IP_ROUTE_TOS
	FIB_SCAN_KEY(f, fp, key) {
		if (f->fn_tos == tos)
			break;
	}
#endif

	matched = 0;
	del_fp = NULL;
	FIB_SCAN_TOS(f, fp, key, tos) {
		struct fib_info * fi = FIB_INFO(f);

		if (f->fn_state&FN_S_ZOMBIE) {
			return -ESRCH;
		}
		matched++;

		if (del_fp == NULL &&
		    (!r->rtm_type || f->fn_type == r->rtm_type) &&
		    (r->rtm_scope == RT_SCOPE_NOWHERE || f->fn_scope == r->rtm_scope) &&
		    (!r->rtm_protocol || fi->fib_protocol == r->rtm_protocol) &&
		    fib_nh_match(r, n, rta, fi) == 0)
			del_fp = fp;
	}

	if (del_fp) {
		f = *del_fp;
		rtmsg_fib(RTM_DELROUTE, f, z, tb->tb_id, n, req);

		if (matched != 1) {
			write_lock_bh(&fib_hash_lock);
			*del_fp = f->fn_next;
			write_unlock_bh(&fib_hash_lock);

			if (f->fn_state&FN_S_ACCESSED)
				rt_cache_flush(-1);
			fn_free_node(f);
			fz->fz_nent--;
		} else {
			f->fn_state |= FN_S_ZOMBIE;
			if (f->fn_state&FN_S_ACCESSED) {
				f->fn_state &= ~FN_S_ACCESSED;
				rt_cache_flush(-1);
			}
			if (++fib_hash_zombies > 128)
				fib_flush();
		}

		return 0;
	}
	return -ESRCH;
}

extern __inline__ int
fn_flush_list(struct fib_node ** fp, int z, struct fn_hash *table)
{
	int found = 0;
	struct fib_node *f;

	while ((f = *fp) != NULL) {
		struct fib_info *fi = FIB_INFO(f);

		if (fi && ((f->fn_state&FN_S_ZOMBIE) || (fi->fib_flags&RTNH_F_DEAD))) {
			write_lock_bh(&fib_hash_lock);
			*fp = f->fn_next;
			write_unlock_bh(&fib_hash_lock);

			fn_free_node(f);
			found++;
			continue;
		}
		fp = &f->fn_next;
	}
	return found;
}

static int fn_hash_flush(struct fib_table *tb)
{
	struct fn_hash *table = (struct fn_hash*)tb->tb_data;
	struct fn_zone *fz;
	int found = 0;

	fib_hash_zombies = 0;
	for (fz = table->fn_zone_list; fz; fz = fz->fz_next) {
		int i;
		int tmp = 0;
		for (i=fz->fz_divisor-1; i>=0; i--)
			tmp += fn_flush_list(&fz->fz_hash[i], fz->fz_order, table);
		fz->fz_nent -= tmp;
		found += tmp;
	}
	return found;
}


#ifdef CONFIG_PROC_FS

static int fn_hash_get_info(struct fib_table *tb, char *buffer, int first, int count)
{
	struct fn_hash *table = (struct fn_hash*)tb->tb_data;
	struct fn_zone *fz;
	int pos = 0;
	int n = 0;

	read_lock(&fib_hash_lock);
	for (fz=table->fn_zone_list; fz; fz = fz->fz_next) {
		int i;
		struct fib_node *f;
		int maxslot = fz->fz_divisor;
		struct fib_node **fp = fz->fz_hash;

		if (fz->fz_nent == 0)
			continue;

		if (pos + fz->fz_nent <= first) {
			pos += fz->fz_nent;
			continue;
		}

		for (i=0; i < maxslot; i++, fp++) {
			for (f = *fp; f; f = f->fn_next) {
				if (++pos <= first)
					continue;
				fib_node_get_info(f->fn_type,
						  f->fn_state&FN_S_ZOMBIE,
						  FIB_INFO(f),
						  fz_prefix(f->fn_key, fz),
						  FZ_MASK(fz), buffer);
				buffer += 128;
				if (++n >= count)
					goto out;
			}
		}
	}
out:
	read_unlock(&fib_hash_lock);
  	return n;
}
#endif


static __inline__ int
fn_hash_dump_bucket(struct sk_buff *skb, struct netlink_callback *cb,
		     struct fib_table *tb,
		     struct fn_zone *fz,
		     struct fib_node *f)
{
	int i, s_i;

	s_i = cb->args[3];
	for (i=0; f; i++, f=f->fn_next) {
		if (i < s_i) continue;
		if (f->fn_state&FN_S_ZOMBIE) continue;
		if (fib_dump_info(skb, NETLINK_CB(cb->skb).pid, cb->nlh->nlmsg_seq,
				  RTM_NEWROUTE,
				  tb->tb_id, (f->fn_state&FN_S_ZOMBIE) ? 0 : f->fn_type, f->fn_scope,
				  &f->fn_key, fz->fz_order, f->fn_tos,
				  f->fn_info) < 0) {
			cb->args[3] = i;
			return -1;
		}
	}
	cb->args[3] = i;
	return skb->len;
}

static __inline__ int
fn_hash_dump_zone(struct sk_buff *skb, struct netlink_callback *cb,
		   struct fib_table *tb,
		   struct fn_zone *fz)
{
	int h, s_h;

	s_h = cb->args[2];
	for (h=0; h < fz->fz_divisor; h++) {
		if (h < s_h) continue;
		if (h > s_h)
			memset(&cb->args[3], 0, sizeof(cb->args) - 3*sizeof(cb->args[0]));
		if (fz->fz_hash == NULL || fz->fz_hash[h] == NULL)
			continue;
		if (fn_hash_dump_bucket(skb, cb, tb, fz, fz->fz_hash[h]) < 0) {
			cb->args[2] = h;
			return -1;
		}
	}
	cb->args[2] = h;
	return skb->len;
}

static int fn_hash_dump(struct fib_table *tb, struct sk_buff *skb, struct netlink_callback *cb)
{
	int m, s_m;
	struct fn_zone *fz;
	struct fn_hash *table = (struct fn_hash*)tb->tb_data;

	s_m = cb->args[1];
	read_lock(&fib_hash_lock);
	for (fz = table->fn_zone_list, m=0; fz; fz = fz->fz_next, m++) {
		if (m < s_m) continue;
		if (m > s_m)
			memset(&cb->args[2], 0, sizeof(cb->args) - 2*sizeof(cb->args[0]));
		if (fn_hash_dump_zone(skb, cb, tb, fz) < 0) {
			cb->args[1] = m;
			read_unlock(&fib_hash_lock);
			return -1;
		}
	}
	read_unlock(&fib_hash_lock);
	cb->args[1] = m;
	return skb->len;
}

static void rtmsg_fib(int event, struct fib_node* f, int z, int tb_id,
		      struct nlmsghdr *n, struct netlink_skb_parms *req)
{
	struct sk_buff *skb;
	u32 pid = req ? req->pid : 0;
	int size = NLMSG_SPACE(sizeof(struct rtmsg)+256);

	skb = alloc_skb(size, GFP_KERNEL);
	if (!skb)
		return;

	if (fib_dump_info(skb, pid, n->nlmsg_seq, event, tb_id,
			  f->fn_type, f->fn_scope, &f->fn_key, z, f->fn_tos,
			  FIB_INFO(f)) < 0) {
		kfree_skb(skb);
		return;
	}
	NETLINK_CB(skb).dst_groups = RTMGRP_IPV4_ROUTE;
	if (n->nlmsg_flags&NLM_F_ECHO)
		atomic_inc(&skb->users);
	netlink_broadcast(rtnl, skb, pid, RTMGRP_IPV4_ROUTE, GFP_KERNEL);
	if (n->nlmsg_flags&NLM_F_ECHO)
		netlink_unicast(rtnl, skb, pid, MSG_DONTWAIT);
}

#ifdef CONFIG_IP_MULTIPLE_TABLES
struct fib_table * fib_hash_init(int id)
#else
struct fib_table * __init fib_hash_init(int id)
#endif
{
	struct fib_table *tb;

	if (fn_hash_kmem == NULL)
		fn_hash_kmem = kmem_cache_create("ip_fib_hash",
						 sizeof(struct fib_node),
						 0, SLAB_HWCACHE_ALIGN,
						 NULL, NULL);

	tb = kmalloc(sizeof(struct fib_table) + sizeof(struct fn_hash), GFP_KERNEL);
	if (tb == NULL)
		return NULL;

	tb->tb_id = id;
	tb->tb_lookup = fn_hash_lookup;
	tb->tb_insert = fn_hash_insert;
	tb->tb_delete = fn_hash_delete;
	tb->tb_flush = fn_hash_flush;
	tb->tb_select_default = fn_hash_select_default;
	tb->tb_dump = fn_hash_dump;
#ifdef CONFIG_PROC_FS
	tb->tb_get_info = fn_hash_get_info;
#endif
	memset(tb->tb_data, 0, sizeof(struct fn_hash));
	return tb;
}