This patch moves the EADDRINUSE check in SCO sockets from bind() to
listen(), as described in the commit message.
Note that the address matching to detect collisions is very
conservative: BDADDR_ANY will match any other address. This seems to
make sense but differs from the original approach in bind().
This second version includes the review from Johan: the helper function
has been renamed and bool is used instead of int.
Another approach, as proposed by Luiz, would be to use a helper function
that, instead of returning a bool, returns the socket that is using the
matching address. This might be useful in the future, but I personally
don't think it is very likely.
The code could be as follows:
static struct sock *__sco_get_sock_by_addr(int state, bdaddr_t *ba)
{
struct sock *sk = NULL, *sk1 = NULL;
struct hlist_node *node;
sk_for_each(sk, node, &sco_sk_list.head) {
if (sk->sk_state != state)
continue;
/* Exact match. */
if (!bacmp(&bt_sk(sk)->src, src))
break;
/* Closest match */
if (!bacmp(&bt_sk(sk)->src, BDADDR_ANY))
sk1 = sk;
/* Match any if given src is BDADDR_ANY */
if (!bacmp(src, BDADDR_ANY))
sk1 = sk;
}
return node ? sk : sk1;
}
This function would be quite similar from sco_get_sock_listen() with the
exception of the last BDADDR_ANY match, and the lock handling.
Cheers,
Mikel
Mikel Astiz (1):
Bluetooth: Fix EADDRINUSE check in SCO sockets
net/bluetooth/sco.c | 49 ++++++++++++++++++++++++++++++-------------------
1 files changed, 30 insertions(+), 19 deletions(-)
--
1.7.7.6
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
