Linux kernel IPv6 routing and forwarding
========================================

We don't want to free the routing cache entry after a packet has been
forwarded: the purpose of the cache entry is to be re-used shortly...

If policy routing is enabled...


We've found the FIB table to search for a an entry i.e. a FIB6 node

static struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table, int oif,
				      struct flowi6 *fl6, int flags)

Once we've found the FIB node we call the function below

static struct rt6_info *rt6_select(struct fib6_node *fn, int oif, int strict)

It returns an entry even if no route is found, this is the
net->ipv6.ip6_null_entry If the previous is returned or an entry that
is already a cache entry, the functions returns. Otherwise a clone of
the macth is created with two possible variants: COW or CLONE


A FIB6 node is a routing table entry, not a routing cache entry: where
is the cache entry created ?!

static inline struct rt6_info *ip6_dst_alloc(struct net *net,
					     struct net_device *dev,
					     int flags,
					     struct fib6_table *table)

struct fib6_node {
	struct fib6_node	*parent;
	struct fib6_node	*left;
	struct fib6_node	*right;
#ifdef CONFIG_IPV6_SUBTREES
	struct fib6_node	*subtree;
#endif
	struct rt6_info		*leaf;

	__u16			fn_bit;		/* bit key */
	__u16			fn_flags;
	__u32			fn_sernum;
	struct rt6_info		*rr_ptr;
};


Looking for who is allocating cache entries

/*
 *	Add routing information to the routing tree.
 *	<destination addr>/<source addr>
 *	with source addr info in sub-trees
 */

int fib6_add(struct fib6_node *root, struct rt6_info *rt, struct nl_info *info)
{