[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-Id: <20080617.225427.66863666.davem@davemloft.net>
Date: Tue, 17 Jun 2008 22:54:27 -0700 (PDT)
From: David Miller <davem@...emloft.net>
To: acme@...hat.com
Cc: johnpol@....mipt.ru, netdev@...r.kernel.org
Subject: Re: sk->sk_socket
From: Arnaldo Carvalho de Melo <acme@...hat.com>
Date: Tue, 17 Jun 2008 11:14:39 -0300
> Em Tue, Jun 17, 2008 at 04:32:55PM +0400, Evgeniy Polyakov escreveu:
> > On Tue, Jun 17, 2008 at 03:25:11AM -0700, David Miller (davem@...emloft.net) wrote:
> > > For the curious, here are the only assignments to sk->sk_socket
> > > remaining in the net-next-2.6 tree right now:
> > >
> > > include/net/sock.h: sk->sk_socket = NULL;
> > > include/net/sock.h: sk->sk_socket = parent;
>
> Isn't this sock_graft and sock_orphan already? :)
>
> > > net/core/sock.c: newsk->sk_socket = NULL;
> > > net/core/sock.c: sk->sk_socket = sock;
>
> I see, perhaps we can have __sock_orphan() in sk_clone() and
> __sock_graft() in sock_init_data() 8)
Even though Evgeniy was only half-serious I think adding such
a helper makes sense. I'm going to have to slap an
rcu_assign_pointer() there anyways if I do the RCU socket
freeing changes.
Applied and pushed to net-next-2.6:
net: Add sk_set_socket() helper.
In order to more easily grep for all things that set
sk->sk_socket, add sk_set_socket() helper inline function.
Suggested (although only half-seriously) by Evgeniy Polyakov.
Signed-off-by: David S. Miller <davem@...emloft.net>
---
include/net/sock.h | 9 +++++++--
net/core/sock.c | 4 ++--
2 files changed, 9 insertions(+), 4 deletions(-)
diff --git a/include/net/sock.h b/include/net/sock.h
index a7c3041..3f4897a 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h
@@ -990,6 +990,11 @@ static inline void sock_put(struct sock *sk)
extern int sk_receive_skb(struct sock *sk, struct sk_buff *skb,
const int nested);
+static inline void sk_set_socket(struct sock *sk, struct socket *sock)
+{
+ sk->sk_socket = sock;
+}
+
/* Detach socket from process context.
* Announce socket dead, detach it from wait queue and inode.
* Note that parent inode held reference count on this struct sock,
@@ -1001,7 +1006,7 @@ static inline void sock_orphan(struct sock *sk)
{
write_lock_bh(&sk->sk_callback_lock);
sock_set_flag(sk, SOCK_DEAD);
- sk->sk_socket = NULL;
+ sk_set_socket(sk, NULL);
sk->sk_sleep = NULL;
write_unlock_bh(&sk->sk_callback_lock);
}
@@ -1011,7 +1016,7 @@ static inline void sock_graft(struct sock *sk, struct socket *parent)
write_lock_bh(&sk->sk_callback_lock);
sk->sk_sleep = &parent->wait;
parent->sk = sk;
- sk->sk_socket = parent;
+ sk_set_socket(sk, parent);
security_sock_graft(sk, parent);
write_unlock_bh(&sk->sk_callback_lock);
}
diff --git a/net/core/sock.c b/net/core/sock.c
index 3879bf6..2c0ba52 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -1066,7 +1066,7 @@ struct sock *sk_clone(const struct sock *sk, const gfp_t priority)
* to be taken into account in all callers. -acme
*/
sk_refcnt_debug_inc(newsk);
- newsk->sk_socket = NULL;
+ sk_set_socket(newsk, NULL);
newsk->sk_sleep = NULL;
if (newsk->sk_prot->sockets_allocated)
@@ -1702,7 +1702,7 @@ void sock_init_data(struct socket *sock, struct sock *sk)
sk->sk_rcvbuf = sysctl_rmem_default;
sk->sk_sndbuf = sysctl_wmem_default;
sk->sk_state = TCP_CLOSE;
- sk->sk_socket = sock;
+ sk_set_socket(sk, sock);
sock_set_flag(sk, SOCK_ZAPPED);
--
1.5.5.1.308.g1fbb5
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists