[PATCH] Implement IP_UNICAST_IF socket option.

Subsystems: networking [general], networking [ipv4/ipv6], the rest, user datagram protocol (udp)

STALE5334d

6 messages, 3 authors, 2012-02-01 · open the first message on its own page

[PATCH] Implement IP_UNICAST_IF socket option.

From: Erich E. Hoover <hidden>
Date: 2012-01-27 00:53:33

The IP_UNICAST_IF feature is needed by the Wine project.  This patch implements the feature by setting the outgoing interface in the same fashion as that of IP_PKTINFO.

Signed-off-by: Erich E. Hoover <redacted>
---
 include/linux/in.h      |    1 +
 include/net/inet_sock.h |    3 ++
 include/net/ip.h        |    2 +-
 net/ipv4/af_inet.c      |    3 ++
 net/ipv4/ip_sockglue.c  |   78 ++++++++++++++++++++++++++++++++++++++++++++++-
 net/ipv4/ping.c         |    2 +-
 net/ipv4/raw.c          |    2 +-
 net/ipv4/udp.c          |    2 +-
 8 files changed, 88 insertions(+), 5 deletions(-)

diff --git a/include/linux/in.h b/include/linux/in.h
index 01129c0..89f6682 100644
--- a/include/linux/in.h
+++ b/include/linux/in.h

@@ -86,6 +86,7 @@ struct in_addr {
 
 #define IP_MINTTL       21
 #define IP_NODEFRAG     22
+#define IP_UNICAST_IF   23
 
 /* IP_MTU_DISCOVER values */
 #define IP_PMTUDISC_DONT		0	/* Never send DF frames */

diff --git a/include/net/inet_sock.h b/include/net/inet_sock.h
index e3e4051..1a8fe64 100644
--- a/include/net/inet_sock.h
+++ b/include/net/inet_sock.h

@@ -132,6 +132,7 @@ struct rtable;
  * @tos - TOS
  * @mc_ttl - Multicasting TTL
  * @is_icsk - is this an inet_connection_sock?
+ * @outif_index - Outgoing device index
  * @mc_index - Multicast device index
  * @mc_list - Group array
  * @cork - info to build ip hdr on each ip frag while socket is corked

@@ -167,6 +168,8 @@ struct inet_sock {
 				transparent:1,
 				mc_all:1,
 				nodefrag:1;
+	int			outif_index;
+	__be32			outif_addr;
 	int			mc_index;
 	__be32			mc_addr;
 	struct ip_mc_socklist __rcu	*mc_list;

diff --git a/include/net/ip.h b/include/net/ip.h
index 775009f..16f7151 100644
--- a/include/net/ip.h
+++ b/include/net/ip.h

@@ -452,7 +452,7 @@ extern int ip_options_rcv_srr(struct sk_buff *skb);
 
 extern void	ipv4_pktinfo_prepare(struct sk_buff *skb);
 extern void	ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
-extern int	ip_cmsg_send(struct net *net,
+extern int	ip_cmsg_send(struct sock *sk,
 			     struct msghdr *msg, struct ipcm_cookie *ipc);
 extern int	ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval, unsigned int optlen);
 extern int	ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, int __user *optlen);

diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index f7b5670..44b1622 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c

@@ -375,6 +375,9 @@ lookup_protocol:
 	sk->sk_protocol	   = protocol;
 	sk->sk_backlog_rcv = sk->sk_prot->backlog_rcv;
 
+	inet->outif_index  = 0;
+	inet->outif_addr   = 0;
+
 	inet->uc_ttl	= -1;
 	inet->mc_loop	= 1;
 	inet->mc_ttl	= 1;

diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index 8aa87c1..b99da51 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c

@@ -186,11 +186,20 @@ void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb)
 }
 EXPORT_SYMBOL(ip_cmsg_recv);
 
-int ip_cmsg_send(struct net *net, struct msghdr *msg, struct ipcm_cookie *ipc)
+int ip_cmsg_send(struct sock *sk, struct msghdr *msg, struct ipcm_cookie *ipc)
 {
+	struct net *net = sock_net(sk);
+	struct inet_sock *inet = inet_sk(sk);
 	int err;
 	struct cmsghdr *cmsg;
 
+	/* set the outgoing interface from IP_UNICAST_IF */
+	if (inet->outif_index || inet->outif_addr)
+	{
+		ipc->oif = inet->outif_index;
+		ipc->addr = inet->outif_addr;
+	}
+
 	for (cmsg = CMSG_FIRSTHDR(msg); cmsg; cmsg = CMSG_NXTHDR(msg, cmsg)) {
 		if (!CMSG_OK(msg, cmsg))
 			return -EINVAL;

@@ -628,6 +637,60 @@ static int do_ip_setsockopt(struct sock *sk, int level,
 			goto e_inval;
 		inet->mc_loop = !!val;
 		break;
+
+	case IP_UNICAST_IF:
+	{
+		struct ip_mreqn mreq;
+		struct net_device *dev = NULL;
+
+		/*
+		 *	Check the arguments are allowable
+		 */
+
+		if (optlen < sizeof(struct in_addr))
+			goto e_inval;
+
+		err = -EFAULT;
+		if (optlen >= sizeof(struct ip_mreqn)) {
+			if (copy_from_user(&mreq, optval, sizeof(mreq)))
+				break;
+		} else {
+			memset(&mreq, 0, sizeof(mreq));
+			if (optlen >= sizeof(struct in_addr) &&
+			    copy_from_user(&mreq.imr_address, optval,
+					   sizeof(struct in_addr)))
+				break;
+		}
+
+		if (!mreq.imr_ifindex) {
+			if (mreq.imr_address.s_addr == htonl(INADDR_ANY)) {
+				inet->outif_index = 0;
+				inet->outif_addr  = 0;
+				err = 0;
+				break;
+			}
+			dev = ip_dev_find(sock_net(sk), mreq.imr_address.s_addr);
+			if (dev)
+				mreq.imr_ifindex = dev->ifindex;
+		} else
+			dev = dev_get_by_index(sock_net(sk), mreq.imr_ifindex);
+
+
+		err = -EADDRNOTAVAIL;
+		if (!dev)
+			break;
+		dev_put(dev);
+
+		err = -EINVAL;
+		if (sk->sk_bound_dev_if &&
+		    mreq.imr_ifindex != sk->sk_bound_dev_if)
+			break;
+
+		inet->outif_index = mreq.imr_ifindex;
+		inet->outif_addr  = mreq.imr_address.s_addr;
+		err = 0;
+		break;
+	}
 	case IP_MULTICAST_IF:
 	{
 		struct ip_mreqn mreq;

@@ -1178,6 +1241,19 @@ static int do_ip_getsockopt(struct sock *sk, int level, int optname,
 	case IP_MULTICAST_LOOP:
 		val = inet->mc_loop;
 		break;
+	case IP_UNICAST_IF:
+	{
+		struct in_addr addr;
+		len = min_t(unsigned int, len, sizeof(struct in_addr));
+		addr.s_addr = inet->outif_addr;
+		release_sock(sk);
+
+		if (put_user(len, optlen))
+			return -EFAULT;
+		if (copy_to_user(optval, &addr, len))
+			return -EFAULT;
+		return 0;
+	}
 	case IP_MULTICAST_IF:
 	{
 		struct in_addr addr;

diff --git a/net/ipv4/ping.c b/net/ipv4/ping.c
index aea5a19..fffdc44 100644
--- a/net/ipv4/ping.c
+++ b/net/ipv4/ping.c

@@ -517,7 +517,7 @@ static int ping_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
 		return err;
 
 	if (msg->msg_controllen) {
-		err = ip_cmsg_send(sock_net(sk), msg, &ipc);
+		err = ip_cmsg_send(sk, msg, &ipc);
 		if (err)
 			return err;
 		if (ipc.opt)

diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c
index 3ccda5a..0aea6c7 100644
--- a/net/ipv4/raw.c
+++ b/net/ipv4/raw.c

@@ -518,7 +518,7 @@ static int raw_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
 	ipc.oif = sk->sk_bound_dev_if;
 
 	if (msg->msg_controllen) {
-		err = ip_cmsg_send(sock_net(sk), msg, &ipc);
+		err = ip_cmsg_send(sk, msg, &ipc);
 		if (err)
 			goto out;
 		if (ipc.opt)

diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 5d075b5..e95c9d4 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c

@@ -874,7 +874,7 @@ int udp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
 	if (err)
 		return err;
 	if (msg->msg_controllen) {
-		err = ip_cmsg_send(sock_net(sk), msg, &ipc);
+		err = ip_cmsg_send(sk, msg, &ipc);
 		if (err)
 			return err;
 		if (ipc.opt)

-- 
1.7.5.4

Re: [PATCH] Implement IP_UNICAST_IF socket option.

From: Hagen Paul Pfeifer <hidden>
Date: 2012-01-27 09:01:51

On Thu, 26 Jan 2012 16:57:48 -0700, "Erich E. Hoover" wrote:

The IP_UNICAST_IF feature is needed by the Wine project.  This patch

implements the feature by setting the outgoing interface in the same

fashion as that of IP_PKTINFO.



Why duplicate functionality - rewrite your application and use IP_PKTINFO

or bind() your socket. This patch duplicate already existing functionality

(and bloat inet_sock unnecessarily).



Hagen

Re: [PATCH] Implement IP_UNICAST_IF socket option.

From: Erich E. Hoover <hidden>
Date: 2012-01-27 15:30:31

On Fri, Jan 27, 2012 at 2:01 AM, Hagen Paul Pfeifer [off-list ref] wrote:

On Thu, 26 Jan 2012 16:57:48 -0700, "Erich E. Hoover" wrote:

quoted

The IP_UNICAST_IF feature is needed by the Wine project.  This patch
implements the feature by setting the outgoing interface in the same
fashion as that of IP_PKTINFO.

Why duplicate functionality - rewrite your application and use IP_PKTINFO
or bind() your socket. This patch duplicate already existing functionality
(and bloat inet_sock unnecessarily).

The IP_UNICAST_IF feature works on broadcast packets, so bind() won't
do the trick.  Also, native Windows applications have the ability to
use both IP_PKTINFO and IP_UNICAST_IF on the same socket, so using
IP_PKTINFO to implement IP_UNICAST_IF might break some applications.
Even so, I did attempt to do this and Alexandre was unsatisfied with
the approach since it required a significant amount of extra code and
the extra communication between the "server" and "user" components of
Wine significantly slowed down socket performance.  Alexendre
indicated that I need to get this functionality added to the kernel to
implement the feature properly, so here I am.

I'm hoping that I can also use this feature, combined with Linux
Socket Filtering, to allow Windows applications to use bind() and have
the socket still work with broadcast packets.  IP_UNICAST_IF is a
relatively new feature, older Windows applications used bind() instead
(Windows bind() still captures broadcast packets).  I've been working
on different ways to get Wine to handle this capability for a long
time and I'd really like to be able to finally check this bug off of
my list.

Anyway, thank you for taking the time to look at my patch.  I really
appreciate and I hope that I've satisfactorily answered your question.

Erich Hoover
ehoover@mines.edu

Re: [PATCH] Implement IP_UNICAST_IF socket option.

From: David Miller <davem@davemloft.net>
Date: 2012-02-01 19:12:35

From: "Erich E. Hoover" <redacted>
Date: Thu, 26 Jan 2012 16:57:48 -0700

The IP_UNICAST_IF feature is needed by the Wine project.  This patch
implements the feature by setting the outgoing interface in the same
fashion as that of IP_PKTINFO.

Signed-off-by: Erich E. Hoover <redacted>

Hmmm, this is not how IP_UNICAST_IF is implemented in other systems.

The passed in argument is always a 32-bit network byte order word, and
it is always interpreted as an interface index only.  There are no
provisions for specifying an IPV4 address.

I see no value in exporting a different API than what other systems
use, especially if you are adding this interface exactly for providing
support for WINE applications, so please don't do this thing with
argument sizes and ip_mreqn.

Re: [PATCH] Implement IP_UNICAST_IF socket option.

From: Erich E. Hoover <hidden>
Date: 2012-02-01 19:32:56

On Wed, Feb 1, 2012 at 12:12 PM, David Miller [off-list ref] wrote:

...
I see no value in exporting a different API than what other systems
use, especially if you are adding this interface exactly for providing
support for WINE applications, so please don't do this thing with
argument sizes and ip_mreqn.

I didn't intend to export the API differently, I clearly misread the
article on MSDN.  If I revise this patch to require an interface index
is there any chance it will be considered for inclusion?

Erich Hoover
ehoover@mines.edu

Re: [PATCH] Implement IP_UNICAST_IF socket option.

From: David Miller <davem@davemloft.net>
Date: 2012-02-01 19:39:16

From: "Erich E. Hoover" <redacted>
Date: Wed, 1 Feb 2012 12:32:54 -0700

On Wed, Feb 1, 2012 at 12:12 PM, David Miller [off-list ref] wrote:

quoted

...
I see no value in exporting a different API than what other systems
use, especially if you are adding this interface exactly for providing
support for WINE applications, so please don't do this thing with
argument sizes and ip_mreqn.

I didn't intend to export the API differently, I clearly misread the
article on MSDN.  If I revise this patch to require an interface index
is there any chance it will be considered for inclusion?

Yes.

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help