diff mbox series

[net-next,v2] mctp: Implement extended addressing

Message ID 20211014023739.2032160-1-jk@codeconstruct.com.au (mailing list archive)
State Superseded
Delegated to: Netdev Maintainers
Headers show
Series [net-next,v2] mctp: Implement extended addressing | expand

Checks

Context Check Description
netdev/cover_letter success Single patches do not need cover letters
netdev/fixes_present success Fixes tag not required for -next series
netdev/patch_count success Link
netdev/tree_selection success Clearly marked for net-next
netdev/subject_prefix success Link
netdev/cc_maintainers success CCed 5 of 5 maintainers
netdev/source_inline success Was 0 now: 0
netdev/verify_signedoff success Signed-off-by tag matches author and committer
netdev/module_param success Was 0 now: 0
netdev/build_32bit success Errors and warnings before: 2 this patch: 2
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/verify_fixes success No Fixes tag
netdev/checkpatch success total: 0 errors, 0 warnings, 0 checks, 385 lines checked
netdev/build_allmodconfig_warn success Errors and warnings before: 2 this patch: 2
netdev/header_inline success No static functions without inline keyword in header files

Commit Message

Jeremy Kerr Oct. 14, 2021, 2:37 a.m. UTC 
This change allows an extended address struct - struct sockaddr_mctp_ext
- to be passed to sendmsg/recvmsg. This allows userspace to specify
output ifindex and physical address information (for sendmsg) or receive
the input ifindex/physaddr for incoming messages (for recvmsg). This is
typically used by userspace for MCTP address discovery and assignment
operations.

The extended addressing facility is conditional on a new sockopt:
MCTP_OPT_ADDR_EXT; userspace must explicitly enable addressing before
the kernel will consume/populate the extended address data.

Signed-off-by: Jeremy Kerr <jk@codeconstruct.com.au>

---
v2:
 - non-RFC
 - learn to spell "typically"

RFC: this patch adds a bit of an new ABI, in that the struct sockaddr
for MCTP has been extended, with extra addressing data being available
for applications after setting a sockopt.

Using something like IP_PKTINFO might be suitable instead, but that
requires sendmsg/recvmsg, and control message setup, whereas this is a
a simpler interface, and also allows extended addressing info in
sendto/recvfrom too. Comments most welcome!
---
 include/net/mctp.h        | 16 +++++--
 include/uapi/linux/mctp.h | 13 ++++++
 net/mctp/af_mctp.c        | 86 ++++++++++++++++++++++++++++++----
 net/mctp/route.c          | 97 +++++++++++++++++++++++++++++----------
 4 files changed, 172 insertions(+), 40 deletions(-)

Comments

kernel test robot Oct. 14, 2021, 8:25 a.m. UTC | #1 
Hi Jeremy,

I love your patch! Perhaps something to improve:

[auto build test WARNING on net-next/master]

url:    https://github.com/0day-ci/linux/commits/Jeremy-Kerr/mctp-Implement-extended-addressing/20211014-103942
base:   https://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git 39e222bfd7f37e7a98069869375b903d7096c113
config: hexagon-randconfig-r041-20211014 (attached as .config)
compiler: clang version 14.0.0 (https://github.com/llvm/llvm-project 6c76d0101193aa4eb891a6954ff047eda2f9cf71)
reproduce (this is a W=1 build):
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # https://github.com/0day-ci/linux/commit/2fed3ce68e4fc1a6f8faaa838ce84829cd030f6b
        git remote add linux-review https://github.com/0day-ci/linux
        git fetch --no-tags linux-review Jeremy-Kerr/mctp-Implement-extended-addressing/20211014-103942
        git checkout 2fed3ce68e4fc1a6f8faaa838ce84829cd030f6b
        # save the attached .config to linux build tree
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 ARCH=hexagon 

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>

All warnings (new ones prefixed by >>):

>> net/mctp/route.c:752:7: warning: variable 'rc' is used uninitialized whenever 'if' condition is true [-Wsometimes-uninitialized]
                   if (WARN_ON(!rt->dev))
                       ^~~~~~~~~~~~~~~~~
   include/asm-generic/bug.h:120:28: note: expanded from macro 'WARN_ON'
   #define WARN_ON(condition) ({                                           \
                              ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   net/mctp/route.c:838:9: note: uninitialized use occurs here
           return rc;
                  ^~
   net/mctp/route.c:752:3: note: remove the 'if' if its condition is always false
                   if (WARN_ON(!rt->dev))
                   ^~~~~~~~~~~~~~~~~~~~~~
   net/mctp/route.c:745:8: note: initialize the variable 'rc' to silence this warning
           int rc;
                 ^
                  = 0
   1 warning generated.


vim +752 net/mctp/route.c

4a992bbd365094 Jeremy Kerr 2021-07-29  732  
889b7da23abf92 Jeremy Kerr 2021-07-29  733  int mctp_local_output(struct sock *sk, struct mctp_route *rt,
889b7da23abf92 Jeremy Kerr 2021-07-29  734  		      struct sk_buff *skb, mctp_eid_t daddr, u8 req_tag)
889b7da23abf92 Jeremy Kerr 2021-07-29  735  {
833ef3b91de692 Jeremy Kerr 2021-07-29  736  	struct mctp_sock *msk = container_of(sk, struct mctp_sock, sk);
889b7da23abf92 Jeremy Kerr 2021-07-29  737  	struct mctp_skb_cb *cb = mctp_cb(skb);
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  738  	struct mctp_route tmp_rt;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  739  	struct net_device *dev;
889b7da23abf92 Jeremy Kerr 2021-07-29  740  	struct mctp_hdr *hdr;
889b7da23abf92 Jeremy Kerr 2021-07-29  741  	unsigned long flags;
4a992bbd365094 Jeremy Kerr 2021-07-29  742  	unsigned int mtu;
889b7da23abf92 Jeremy Kerr 2021-07-29  743  	mctp_eid_t saddr;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  744  	bool ext_rt;
889b7da23abf92 Jeremy Kerr 2021-07-29  745  	int rc;
833ef3b91de692 Jeremy Kerr 2021-07-29  746  	u8 tag;
889b7da23abf92 Jeremy Kerr 2021-07-29  747  
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  748  	if (rt) {
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  749  		ext_rt = false;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  750  		dev = NULL;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  751  
889b7da23abf92 Jeremy Kerr 2021-07-29 @752  		if (WARN_ON(!rt->dev))
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  753  			goto out_release;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  754  
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  755  	} else if (cb->ifindex) {
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  756  		ext_rt = true;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  757  		rt = &tmp_rt;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  758  
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  759  		rc = -ENODEV;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  760  		rcu_read_lock();
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  761  		dev = dev_get_by_index_rcu(sock_net(sk), cb->ifindex);
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  762  		if (!dev) {
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  763  			rcu_read_unlock();
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  764  			return rc;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  765  		}
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  766  
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  767  		rt->dev = __mctp_dev_get(dev);
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  768  		rcu_read_unlock();
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  769  
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  770  		if (!rt->dev)
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  771  			goto out_release;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  772  
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  773  		/* establish temporary route - we set up enough to keep
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  774  		 * mctp_route_output happy
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  775  		 */
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  776  		rt->output = mctp_route_output;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  777  		rt->mtu = 0;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  778  
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  779  	} else {
889b7da23abf92 Jeremy Kerr 2021-07-29  780  		return -EINVAL;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  781  	}
889b7da23abf92 Jeremy Kerr 2021-07-29  782  
889b7da23abf92 Jeremy Kerr 2021-07-29  783  	spin_lock_irqsave(&rt->dev->addrs_lock, flags);
889b7da23abf92 Jeremy Kerr 2021-07-29  784  	if (rt->dev->num_addrs == 0) {
889b7da23abf92 Jeremy Kerr 2021-07-29  785  		rc = -EHOSTUNREACH;
889b7da23abf92 Jeremy Kerr 2021-07-29  786  	} else {
889b7da23abf92 Jeremy Kerr 2021-07-29  787  		/* use the outbound interface's first address as our source */
889b7da23abf92 Jeremy Kerr 2021-07-29  788  		saddr = rt->dev->addrs[0];
889b7da23abf92 Jeremy Kerr 2021-07-29  789  		rc = 0;
889b7da23abf92 Jeremy Kerr 2021-07-29  790  	}
889b7da23abf92 Jeremy Kerr 2021-07-29  791  	spin_unlock_irqrestore(&rt->dev->addrs_lock, flags);
889b7da23abf92 Jeremy Kerr 2021-07-29  792  
889b7da23abf92 Jeremy Kerr 2021-07-29  793  	if (rc)
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  794  		goto out_release;
889b7da23abf92 Jeremy Kerr 2021-07-29  795  
833ef3b91de692 Jeremy Kerr 2021-07-29  796  	if (req_tag & MCTP_HDR_FLAG_TO) {
833ef3b91de692 Jeremy Kerr 2021-07-29  797  		rc = mctp_alloc_local_tag(msk, saddr, daddr, &tag);
833ef3b91de692 Jeremy Kerr 2021-07-29  798  		if (rc)
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  799  			goto out_release;
833ef3b91de692 Jeremy Kerr 2021-07-29  800  		tag |= MCTP_HDR_FLAG_TO;
833ef3b91de692 Jeremy Kerr 2021-07-29  801  	} else {
833ef3b91de692 Jeremy Kerr 2021-07-29  802  		tag = req_tag;
833ef3b91de692 Jeremy Kerr 2021-07-29  803  	}
833ef3b91de692 Jeremy Kerr 2021-07-29  804  
4a992bbd365094 Jeremy Kerr 2021-07-29  805  	skb->protocol = htons(ETH_P_MCTP);
4a992bbd365094 Jeremy Kerr 2021-07-29  806  	skb->priority = 0;
889b7da23abf92 Jeremy Kerr 2021-07-29  807  	skb_reset_transport_header(skb);
889b7da23abf92 Jeremy Kerr 2021-07-29  808  	skb_push(skb, sizeof(struct mctp_hdr));
889b7da23abf92 Jeremy Kerr 2021-07-29  809  	skb_reset_network_header(skb);
4a992bbd365094 Jeremy Kerr 2021-07-29  810  	skb->dev = rt->dev->dev;
4a992bbd365094 Jeremy Kerr 2021-07-29  811  
4a992bbd365094 Jeremy Kerr 2021-07-29  812  	/* cb->net will have been set on initial ingress */
4a992bbd365094 Jeremy Kerr 2021-07-29  813  	cb->src = saddr;
4a992bbd365094 Jeremy Kerr 2021-07-29  814  
4a992bbd365094 Jeremy Kerr 2021-07-29  815  	/* set up common header fields */
889b7da23abf92 Jeremy Kerr 2021-07-29  816  	hdr = mctp_hdr(skb);
889b7da23abf92 Jeremy Kerr 2021-07-29  817  	hdr->ver = 1;
889b7da23abf92 Jeremy Kerr 2021-07-29  818  	hdr->dest = daddr;
889b7da23abf92 Jeremy Kerr 2021-07-29  819  	hdr->src = saddr;
889b7da23abf92 Jeremy Kerr 2021-07-29  820  
4a992bbd365094 Jeremy Kerr 2021-07-29  821  	mtu = mctp_route_mtu(rt);
889b7da23abf92 Jeremy Kerr 2021-07-29  822  
4a992bbd365094 Jeremy Kerr 2021-07-29  823  	if (skb->len + sizeof(struct mctp_hdr) <= mtu) {
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  824  		hdr->flags_seq_tag = MCTP_HDR_FLAG_SOM |
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  825  			MCTP_HDR_FLAG_EOM | tag;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  826  		rc = rt->output(rt, skb);
4a992bbd365094 Jeremy Kerr 2021-07-29  827  	} else {
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  828  		rc = mctp_do_fragment_route(rt, skb, mtu, tag);
4a992bbd365094 Jeremy Kerr 2021-07-29  829  	}
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  830  
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  831  out_release:
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  832  	if (!ext_rt)
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  833  		mctp_route_release(rt);
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  834  
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  835  	if (dev)
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  836  		dev_put(dev);
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  837  
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  838  	return rc;
2fed3ce68e4fc1 Jeremy Kerr 2021-10-14  839  

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
diff mbox series

Patch

diff --git a/include/net/mctp.h b/include/net/mctp.h
index b9ed62a63c24..716405e0294c 100644
--- a/include/net/mctp.h
+++ b/include/net/mctp.h
@@ -58,6 +58,9 @@  struct mctp_sock {
 	mctp_eid_t	bind_addr;
 	__u8		bind_type;
 
+	/* sendmsg()/recvmsg() uses struct sockaddr_mctp_ext */
+	bool		addr_ext;
+
 	/* list of mctp_sk_key, for incoming tag lookup. updates protected
 	 * by sk->net->keys_lock
 	 */
@@ -152,10 +155,16 @@  struct mctp_sk_key {
 
 struct mctp_skb_cb {
 	unsigned int	magic;
-	unsigned int	net;
+	int		net;
+	int		ifindex; /* extended/direct addressing if set */
+	unsigned char	halen;
 	mctp_eid_t	src;
+	unsigned char	haddr[];
 };
 
+#define MCTP_SKB_CB_HADDR_MAXLEN (sizeof((((struct sk_buff *)(NULL))->cb)) \
+				  - offsetof(struct mctp_skb_cb, haddr))
+
 /* skb control-block accessors with a little extra debugging for initial
  * development.
  *
@@ -189,8 +198,7 @@  static inline struct mctp_skb_cb *mctp_cb(struct sk_buff *skb)
  *
  * Updates to the route table are performed under rtnl; all reads under RCU,
  * so routes cannot be referenced over a RCU grace period. Specifically: A
- * caller cannot block between mctp_route_lookup and passing the route to
- * mctp_do_route.
+ * caller cannot block between mctp_route_lookup and mctp_route_release()
  */
 struct mctp_route {
 	mctp_eid_t		min, max;
@@ -210,8 +218,6 @@  struct mctp_route {
 struct mctp_route *mctp_route_lookup(struct net *net, unsigned int dnet,
 				     mctp_eid_t daddr);
 
-int mctp_do_route(struct mctp_route *rt, struct sk_buff *skb);
-
 int mctp_local_output(struct sock *sk, struct mctp_route *rt,
 		      struct sk_buff *skb, mctp_eid_t daddr, u8 req_tag);
 
diff --git a/include/uapi/linux/mctp.h b/include/uapi/linux/mctp.h
index 52b54d13f385..896432ce3c71 100644
--- a/include/uapi/linux/mctp.h
+++ b/include/uapi/linux/mctp.h
@@ -10,6 +10,7 @@ 
 #define __UAPI_MCTP_H
 
 #include <linux/types.h>
+#include <linux/netdevice.h>
 
 typedef __u8			mctp_eid_t;
 
@@ -25,6 +26,13 @@  struct sockaddr_mctp {
 	__u8			smctp_tag;
 };
 
+struct sockaddr_mctp_ext {
+	struct sockaddr_mctp	smctp_base;
+	int			smctp_ifindex;
+	unsigned char		smctp_halen;
+	unsigned char		smctp_haddr[MAX_ADDR_LEN];
+};
+
 #define MCTP_NET_ANY		0x0
 
 #define MCTP_ADDR_NULL		0x00
@@ -33,4 +41,9 @@  struct sockaddr_mctp {
 #define MCTP_TAG_MASK		0x07
 #define MCTP_TAG_OWNER		0x08
 
+/* setsockopt(2) level & options */
+#define SOL_MCTP		0
+
+#define MCTP_OPT_ADDR_EXT	1
+
 #endif /* __UAPI_MCTP_H */
diff --git a/net/mctp/af_mctp.c b/net/mctp/af_mctp.c
index 66a411d60b6c..5eae06aaf65c 100644
--- a/net/mctp/af_mctp.c
+++ b/net/mctp/af_mctp.c
@@ -77,6 +77,7 @@  static int mctp_sendmsg(struct socket *sock, struct msghdr *msg, size_t len)
 	const int hlen = MCTP_HEADER_MAXLEN + sizeof(struct mctp_hdr);
 	int rc, addrlen = msg->msg_namelen;
 	struct sock *sk = sock->sk;
+	struct mctp_sock *msk = container_of(sk, struct mctp_sock, sk);
 	struct mctp_skb_cb *cb;
 	struct mctp_route *rt;
 	struct sk_buff *skb;
@@ -100,11 +101,6 @@  static int mctp_sendmsg(struct socket *sock, struct msghdr *msg, size_t len)
 	if (addr->smctp_network == MCTP_NET_ANY)
 		addr->smctp_network = mctp_default_net(sock_net(sk));
 
-	rt = mctp_route_lookup(sock_net(sk), addr->smctp_network,
-			       addr->smctp_addr.s_addr);
-	if (!rt)
-		return -EHOSTUNREACH;
-
 	skb = sock_alloc_send_skb(sk, hlen + 1 + len,
 				  msg->msg_flags & MSG_DONTWAIT, &rc);
 	if (!skb)
@@ -116,19 +112,45 @@  static int mctp_sendmsg(struct socket *sock, struct msghdr *msg, size_t len)
 	*(u8 *)skb_put(skb, 1) = addr->smctp_type;
 
 	rc = memcpy_from_msg((void *)skb_put(skb, len), msg, len);
-	if (rc < 0) {
-		kfree_skb(skb);
-		return rc;
-	}
+	if (rc < 0)
+		goto err_free;
 
 	/* set up cb */
 	cb = __mctp_cb(skb);
 	cb->net = addr->smctp_network;
 
+	/* direct addressing */
+	if (msk->addr_ext && addrlen >= sizeof(struct sockaddr_mctp_ext)) {
+		DECLARE_SOCKADDR(struct sockaddr_mctp_ext *,
+				 extaddr, msg->msg_name);
+
+		if (extaddr->smctp_halen > MCTP_SKB_CB_HADDR_MAXLEN) {
+			rc = -EINVAL;
+			goto err_free;
+		}
+
+		cb->ifindex = extaddr->smctp_ifindex;
+		cb->halen = extaddr->smctp_halen;
+		memcpy(cb->haddr, extaddr->smctp_haddr, cb->halen);
+
+		rt = NULL;
+	} else {
+		rt = mctp_route_lookup(sock_net(sk), addr->smctp_network,
+				       addr->smctp_addr.s_addr);
+		if (!rt) {
+			rc = -EHOSTUNREACH;
+			goto err_free;
+		}
+	}
+
 	rc = mctp_local_output(sk, rt, skb, addr->smctp_addr.s_addr,
 			       addr->smctp_tag);
 
 	return rc ? : len;
+
+err_free:
+	kfree_skb(skb);
+	return rc;
 }
 
 static int mctp_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
@@ -136,6 +158,7 @@  static int mctp_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
 {
 	DECLARE_SOCKADDR(struct sockaddr_mctp *, addr, msg->msg_name);
 	struct sock *sk = sock->sk;
+	struct mctp_sock *msk = container_of(sk, struct mctp_sock, sk);
 	struct sk_buff *skb;
 	size_t msglen;
 	u8 type;
@@ -181,6 +204,16 @@  static int mctp_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
 		addr->smctp_tag = hdr->flags_seq_tag &
 					(MCTP_HDR_TAG_MASK | MCTP_HDR_FLAG_TO);
 		msg->msg_namelen = sizeof(*addr);
+
+		if (msk->addr_ext) {
+			DECLARE_SOCKADDR(struct sockaddr_mctp_ext *, ae,
+					 msg->msg_name);
+			msg->msg_namelen = sizeof(*ae);
+			ae->smctp_ifindex = cb->ifindex;
+			ae->smctp_halen = cb->halen;
+			memset(ae->smctp_haddr, 0x0, sizeof(ae->smctp_haddr));
+			memcpy(ae->smctp_haddr, cb->haddr, cb->halen);
+		}
 	}
 
 	rc = len;
@@ -196,12 +229,45 @@  static int mctp_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
 static int mctp_setsockopt(struct socket *sock, int level, int optname,
 			   sockptr_t optval, unsigned int optlen)
 {
-	return -EINVAL;
+	struct mctp_sock *msk = container_of(sock->sk, struct mctp_sock, sk);
+	int val;
+
+	if (level != SOL_MCTP)
+		return -EINVAL;
+
+	if (optname == MCTP_OPT_ADDR_EXT) {
+		if (optlen != sizeof(int))
+			return -EINVAL;
+		if (copy_from_sockptr(&val, optval, sizeof(int)))
+			return -EFAULT;
+		msk->addr_ext = val;
+		return 0;
+	}
+
+	return -ENOPROTOOPT;
 }
 
 static int mctp_getsockopt(struct socket *sock, int level, int optname,
 			   char __user *optval, int __user *optlen)
 {
+	struct mctp_sock *msk = container_of(sock->sk, struct mctp_sock, sk);
+	int len, val;
+
+	if (level != SOL_MCTP)
+		return -EINVAL;
+
+	if (get_user(len, optlen))
+		return -EFAULT;
+
+	if (optname == MCTP_OPT_ADDR_EXT) {
+		if (len != sizeof(int))
+			return -EINVAL;
+		val = !!msk->addr_ext;
+		if (copy_to_user(optval, &val, len))
+			return -EFAULT;
+		return 0;
+	}
+
 	return -EINVAL;
 }
 
diff --git a/net/mctp/route.c b/net/mctp/route.c
index 04781459b2be..3d77bb5e34d2 100644
--- a/net/mctp/route.c
+++ b/net/mctp/route.c
@@ -433,6 +433,7 @@  static unsigned int mctp_route_mtu(struct mctp_route *rt)
 
 static int mctp_route_output(struct mctp_route *route, struct sk_buff *skb)
 {
+	struct mctp_skb_cb *cb = mctp_cb(skb);
 	struct mctp_hdr *hdr = mctp_hdr(skb);
 	char daddr_buf[MAX_ADDR_LEN];
 	char *daddr = NULL;
@@ -447,9 +448,14 @@  static int mctp_route_output(struct mctp_route *route, struct sk_buff *skb)
 		return -EMSGSIZE;
 	}
 
-	/* If lookup fails let the device handle daddr==NULL */
-	if (mctp_neigh_lookup(route->dev, hdr->dest, daddr_buf) == 0)
-		daddr = daddr_buf;
+	if (cb->ifindex) {
+		/* direct route; use the hwaddr we stashed in sendmsg */
+		daddr = cb->haddr;
+	} else {
+		/* If lookup fails let the device handle daddr==NULL */
+		if (mctp_neigh_lookup(route->dev, hdr->dest, daddr_buf) == 0)
+			daddr = daddr_buf;
+	}
 
 	rc = dev_hard_header(skb, skb->dev, ntohs(skb->protocol),
 			     daddr, skb->dev->dev_addr, skb->len);
@@ -645,16 +651,6 @@  static struct mctp_route *mctp_route_lookup_null(struct net *net,
 	return NULL;
 }
 
-/* sends a skb to rt and releases the route. */
-int mctp_do_route(struct mctp_route *rt, struct sk_buff *skb)
-{
-	int rc;
-
-	rc = rt->output(rt, skb);
-	mctp_route_release(rt);
-	return rc;
-}
-
 static int mctp_do_fragment_route(struct mctp_route *rt, struct sk_buff *skb,
 				  unsigned int mtu, u8 tag)
 {
@@ -721,7 +717,7 @@  static int mctp_do_fragment_route(struct mctp_route *rt, struct sk_buff *skb,
 		/* copy message payload */
 		skb_copy_bits(skb, pos, skb_transport_header(skb2), size);
 
-		/* do route, but don't drop the rt reference */
+		/* do route */
 		rc = rt->output(rt, skb2);
 		if (rc)
 			break;
@@ -730,7 +726,6 @@  static int mctp_do_fragment_route(struct mctp_route *rt, struct sk_buff *skb,
 		pos += size;
 	}
 
-	mctp_route_release(rt);
 	consume_skb(skb);
 	return rc;
 }
@@ -740,15 +735,50 @@  int mctp_local_output(struct sock *sk, struct mctp_route *rt,
 {
 	struct mctp_sock *msk = container_of(sk, struct mctp_sock, sk);
 	struct mctp_skb_cb *cb = mctp_cb(skb);
+	struct mctp_route tmp_rt;
+	struct net_device *dev;
 	struct mctp_hdr *hdr;
 	unsigned long flags;
 	unsigned int mtu;
 	mctp_eid_t saddr;
+	bool ext_rt;
 	int rc;
 	u8 tag;
 
-	if (WARN_ON(!rt->dev))
+	if (rt) {
+		ext_rt = false;
+		dev = NULL;
+
+		if (WARN_ON(!rt->dev))
+			goto out_release;
+
+	} else if (cb->ifindex) {
+		ext_rt = true;
+		rt = &tmp_rt;
+
+		rc = -ENODEV;
+		rcu_read_lock();
+		dev = dev_get_by_index_rcu(sock_net(sk), cb->ifindex);
+		if (!dev) {
+			rcu_read_unlock();
+			return rc;
+		}
+
+		rt->dev = __mctp_dev_get(dev);
+		rcu_read_unlock();
+
+		if (!rt->dev)
+			goto out_release;
+
+		/* establish temporary route - we set up enough to keep
+		 * mctp_route_output happy
+		 */
+		rt->output = mctp_route_output;
+		rt->mtu = 0;
+
+	} else {
 		return -EINVAL;
+	}
 
 	spin_lock_irqsave(&rt->dev->addrs_lock, flags);
 	if (rt->dev->num_addrs == 0) {
@@ -761,18 +791,17 @@  int mctp_local_output(struct sock *sk, struct mctp_route *rt,
 	spin_unlock_irqrestore(&rt->dev->addrs_lock, flags);
 
 	if (rc)
-		return rc;
+		goto out_release;
 
 	if (req_tag & MCTP_HDR_FLAG_TO) {
 		rc = mctp_alloc_local_tag(msk, saddr, daddr, &tag);
 		if (rc)
-			return rc;
+			goto out_release;
 		tag |= MCTP_HDR_FLAG_TO;
 	} else {
 		tag = req_tag;
 	}
 
-
 	skb->protocol = htons(ETH_P_MCTP);
 	skb->priority = 0;
 	skb_reset_transport_header(skb);
@@ -792,12 +821,22 @@  int mctp_local_output(struct sock *sk, struct mctp_route *rt,
 	mtu = mctp_route_mtu(rt);
 
 	if (skb->len + sizeof(struct mctp_hdr) <= mtu) {
-		hdr->flags_seq_tag = MCTP_HDR_FLAG_SOM | MCTP_HDR_FLAG_EOM |
-			tag;
-		return mctp_do_route(rt, skb);
+		hdr->flags_seq_tag = MCTP_HDR_FLAG_SOM |
+			MCTP_HDR_FLAG_EOM | tag;
+		rc = rt->output(rt, skb);
 	} else {
-		return mctp_do_fragment_route(rt, skb, mtu, tag);
+		rc = mctp_do_fragment_route(rt, skb, mtu, tag);
 	}
+
+out_release:
+	if (!ext_rt)
+		mctp_route_release(rt);
+
+	if (dev)
+		dev_put(dev);
+
+	return rc;
+
 }
 
 /* route management */
@@ -938,8 +977,15 @@  static int mctp_pkttype_receive(struct sk_buff *skb, struct net_device *dev,
 	if (mh->ver < MCTP_VER_MIN || mh->ver > MCTP_VER_MAX)
 		goto err_drop;
 
-	cb = __mctp_cb(skb);
+	/* MCTP drivers must populate halen/haddr */
+	if (dev->type == ARPHRD_MCTP) {
+		cb = mctp_cb(skb);
+	} else {
+		cb = __mctp_cb(skb);
+		cb->halen = 0;
+	}
 	cb->net = READ_ONCE(mdev->net);
+	cb->ifindex = dev->ifindex;
 
 	rt = mctp_route_lookup(net, cb->net, mh->dest);
 
@@ -950,7 +996,8 @@  static int mctp_pkttype_receive(struct sk_buff *skb, struct net_device *dev,
 	if (!rt)
 		goto err_drop;
 
-	mctp_do_route(rt, skb);
+	rt->output(rt, skb);
+	mctp_route_release(rt);
 
 	return NET_RX_SUCCESS;