[PATCH net-next 0/9] rxrpc: Development

[PATCH net-next 0/9] rxrpc: Development

[David Howells]

Here are some development patches for AF_RXRPC.  The most significant points
are:

 (1) Change the tracepoint that indicates a packet has been transmitted
     into one that indicates a packet is about to be transmitted.  Without
     this, the response tracepoint may occur first if the round trip is
     fast enough.

 (2) Sort out AFS address list handling to better enforce maximum capacity
     to use helper functions to fill them and to do an insertion sort to
     order them.  This is here to make (3) easier.

 (3) Keep AF_INET addresses as AF_INET addresses rather than converting
     them to AF_INET6 in both AF_RXRPC and kAFS.  I hadn't realised that a
     UDP6 socket would just call down into UDP4 if given an AF_INET
     address.

 (4) Allow the timestamp on the first DATA packet of a reply to be
     retrieved by a kernel service.  This will give the kAFS a more
     accurate base from which to calculate the callback promise expiration.

 (5) Allow the rxrpc protocol epoch value to be retrieved from an incoming
     call.  This will allow kAFS to determine if the fileserver restarted
     and if two addresses apparently assigned to the same fileserver
     actually are different boxes.

The patches are tagged here:

	git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git
	rxrpc-next-20181004

and can also be found on this branch:

	http://git.kernel.org/cgit/linux/kernel/git/dhowells/linux-fs.git/log/?h=rxrpc-next

David
---
David Howells (9):
      rxrpc: Use rxrpc_free_skb() rather than rxrpc_lose_skb()
      rxrpc: Emit the data Tx trace line before transmitting
      afs: Do better max capacity handling on address lists
      afs: Always build address lists using the helper functions
      afs: Sort address lists so that they are in logical ascending order
      rxrpc: Use IPv4 addresses throught the IPv6
      rxrpc: Drop the local endpoint arg from rxrpc_extract_addr_from_skb()
      rxrpc: Allow the reply time to be obtained on a client call
      rxrpc: Allow the reply time to be obtained on a client call


 Documentation/networking/rxrpc.txt |   25 +++++++++
 fs/afs/addr_list.c                 |  101 ++++++++++++++++++------------------
 fs/afs/internal.h                  |    8 ++-
 include/net/af_rxrpc.h             |    4 +
 net/rxrpc/af_rxrpc.c               |   17 ++++++
 net/rxrpc/ar-internal.h            |    4 -
 net/rxrpc/call_accept.c            |    2 -
 net/rxrpc/conn_object.c            |    7 +-
 net/rxrpc/input.c                  |    2 -
 net/rxrpc/local_event.c            |    2 -
 net/rxrpc/output.c                 |   10 ++--
 net/rxrpc/peer_event.c             |   12 ++--
 net/rxrpc/recvmsg.c                |   43 +++++++++++++++
 net/rxrpc/skbuff.c                 |   15 -----
 net/rxrpc/utils.c                  |   23 ++------
 15 files changed, 169 insertions(+), 106 deletions(-)

[PATCH net-next 1/9] rxrpc: Use rxrpc_free_skb() rather than rxrpc_lose_skb()

[David Howells]

rxrpc_lose_skb() is now exactly the same as rxrpc_free_skb(), so remove it
and use the latter instead.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 net/rxrpc/ar-internal.h |    1 -
 net/rxrpc/input.c       |    2 +-
 net/rxrpc/skbuff.c      |   15 ---------------
 3 files changed, 1 insertion(+), 17 deletions(-)

diff --git a/net/rxrpc/ar-internal.h b/net/rxrpc/ar-internal.h
index ef9554131434..9ba87e4d15c7 100644
--- a/net/rxrpc/ar-internal.h
+++ b/net/rxrpc/ar-internal.h
@@ -1095,7 +1095,6 @@ void rxrpc_new_skb(struct sk_buff *, enum rxrpc_skb_trace);
 void rxrpc_see_skb(struct sk_buff *, enum rxrpc_skb_trace);
 void rxrpc_get_skb(struct sk_buff *, enum rxrpc_skb_trace);
 void rxrpc_free_skb(struct sk_buff *, enum rxrpc_skb_trace);
-void rxrpc_lose_skb(struct sk_buff *, enum rxrpc_skb_trace);
 void rxrpc_purge_queue(struct sk_buff_head *);
 
 /*
diff --git a/net/rxrpc/input.c b/net/rxrpc/input.c
index 4b86b6be6c1f..5b2626929822 100644
--- a/net/rxrpc/input.c
+++ b/net/rxrpc/input.c
@@ -1176,7 +1176,7 @@ void rxrpc_data_ready(struct sock *udp_sk)
 		static int lose;
 		if ((lose++ & 7) == 7) {
 			trace_rxrpc_rx_lose(sp);
-			rxrpc_lose_skb(skb, rxrpc_skb_rx_lost);
+			rxrpc_free_skb(skb, rxrpc_skb_rx_lost);
 			return;
 		}
 	}
diff --git a/net/rxrpc/skbuff.c b/net/rxrpc/skbuff.c
index b8985d01876a..913dca65cc65 100644
--- a/net/rxrpc/skbuff.c
+++ b/net/rxrpc/skbuff.c
@@ -68,21 +68,6 @@ void rxrpc_free_skb(struct sk_buff *skb, enum rxrpc_skb_trace op)
 	}
 }
 
-/*
- * Note the injected loss of a socket buffer.
- */
-void rxrpc_lose_skb(struct sk_buff *skb, enum rxrpc_skb_trace op)
-{
-	const void *here = __builtin_return_address(0);
-	if (skb) {
-		int n;
-		CHECK_SLAB_OKAY(&skb->users);
-		n = atomic_dec_return(select_skb_count(op));
-		trace_rxrpc_skb(skb, op, refcount_read(&skb->users), n, here);
-		kfree_skb(skb);
-	}
-}
-
 /*
  * Clear a queue of socket buffers.
  */

[PATCH net-next 2/9] rxrpc: Emit the data Tx trace line before transmitting

[David Howells]

Print the data Tx trace line before transmitting so that it appears before
the trace lines indicating success or failure of the transmission.  This
makes the trace log less confusing.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 net/rxrpc/output.c |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/net/rxrpc/output.c b/net/rxrpc/output.c
index e8fb8922bca8..993d4cd247f9 100644
--- a/net/rxrpc/output.c
+++ b/net/rxrpc/output.c
@@ -378,11 +378,13 @@ int rxrpc_send_data_packet(struct rxrpc_call *call, struct sk_buff *skb,
 		if ((lose++ & 7) == 7) {
 			ret = 0;
 			lost = true;
-			goto done;
 		}
 	}
 
-	_proto("Tx DATA %%%u { #%u }", serial, sp->hdr.seq);
+	trace_rxrpc_tx_data(call, sp->hdr.seq, serial, whdr.flags,
+			    retrans, lost);
+	if (lost)
+		goto done;
 
 	/* send the packet with the don't fragment bit set if we currently
 	 * think it's small enough */
@@ -415,8 +417,6 @@ int rxrpc_send_data_packet(struct rxrpc_call *call, struct sk_buff *skb,
 		goto send_fragmentable;
 
 done:
-	trace_rxrpc_tx_data(call, sp->hdr.seq, serial, whdr.flags,
-			    retrans, lost);
 	if (ret >= 0) {
 		if (whdr.flags & RXRPC_REQUEST_ACK) {
 			call->peer->rtt_last_req = skb->tstamp;

[PATCH net-next 3/9] afs: Do better max capacity handling on address lists

[David Howells]

Note the maximum allocated capacity in an afs_addr_list struct and discard
addresses that would exceed it in afs_merge_fs_addr{4,6}().

Also, since the current maximum capacity is less than 255, reduce the
relevant members to bytes.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 fs/afs/addr_list.c |   19 +++++++++++--------
 fs/afs/internal.h  |    8 +++++---
 2 files changed, 16 insertions(+), 11 deletions(-)

diff --git a/fs/afs/addr_list.c b/fs/afs/addr_list.c
index 025a9a5e1c32..4dbb8af54668 100644
--- a/fs/afs/addr_list.c
+++ b/fs/afs/addr_list.c
@@ -17,11 +17,6 @@
 #include "internal.h"
 #include "afs_fs.h"
 
-//#define AFS_MAX_ADDRESSES
-//	((unsigned int)((PAGE_SIZE - sizeof(struct afs_addr_list)) /
-//			sizeof(struct sockaddr_rxrpc)))
-#define AFS_MAX_ADDRESSES ((unsigned int)(sizeof(unsigned long) * 8))
-
 /*
  * Release an address list.
  */
@@ -43,11 +38,15 @@ struct afs_addr_list *afs_alloc_addrlist(unsigned int nr,
 
 	_enter("%u,%u,%u", nr, service, port);
 
+	if (nr > AFS_MAX_ADDRESSES)
+		nr = AFS_MAX_ADDRESSES;
+
 	alist = kzalloc(struct_size(alist, addrs, nr), GFP_KERNEL);
 	if (!alist)
 		return NULL;
 
 	refcount_set(&alist->usage, 1);
+	alist->max_addrs = nr;
 
 	for (i = 0; i < nr; i++) {
 		struct sockaddr_rxrpc *srx = &alist->addrs[i];
@@ -109,8 +108,6 @@ struct afs_addr_list *afs_parse_text_addrs(const char *text, size_t len,
 	} while (p < end);
 
 	_debug("%u/%u addresses", nr, AFS_MAX_ADDRESSES);
-	if (nr > AFS_MAX_ADDRESSES)
-		nr = AFS_MAX_ADDRESSES;
 
 	alist = afs_alloc_addrlist(nr, service, port);
 	if (!alist)
@@ -180,7 +177,7 @@ struct afs_addr_list *afs_parse_text_addrs(const char *text, size_t len,
 		}
 
 		alist->nr_addrs++;
-	} while (p < end && alist->nr_addrs < AFS_MAX_ADDRESSES);
+	} while (p < end && alist->nr_addrs < alist->max_addrs);
 
 	_leave(" = [nr %u]", alist->nr_addrs);
 	return alist;
@@ -241,6 +238,9 @@ void afs_merge_fs_addr4(struct afs_addr_list *alist, __be32 xdr, u16 port)
 	__be16 xport = htons(port);
 	int i;
 
+	if (alist->nr_addrs >= alist->max_addrs)
+		return;
+
 	for (i = 0; i < alist->nr_ipv4; i++) {
 		a = &alist->addrs[i].transport.sin6;
 		if (xdr == a->sin6_addr.s6_addr32[3] &&
@@ -277,6 +277,9 @@ void afs_merge_fs_addr6(struct afs_addr_list *alist, __be32 *xdr, u16 port)
 	__be16 xport = htons(port);
 	int i, diff;
 
+	if (alist->nr_addrs >= alist->max_addrs)
+		return;
+
 	for (i = alist->nr_ipv4; i < alist->nr_addrs; i++) {
 		a = &alist->addrs[i].transport.sin6;
 		diff = memcmp(xdr, &a->sin6_addr, 16);
diff --git a/fs/afs/internal.h b/fs/afs/internal.h
index 871a228d7f37..8ae4e2ebb99a 100644
--- a/fs/afs/internal.h
+++ b/fs/afs/internal.h
@@ -73,12 +73,14 @@ struct afs_addr_list {
 	struct rcu_head		rcu;		/* Must be first */
 	refcount_t		usage;
 	u32			version;	/* Version */
-	unsigned short		nr_addrs;
-	unsigned short		index;		/* Address currently in use */
-	unsigned short		nr_ipv4;	/* Number of IPv4 addresses */
+	unsigned char		max_addrs;
+	unsigned char		nr_addrs;
+	unsigned char		index;		/* Address currently in use */
+	unsigned char		nr_ipv4;	/* Number of IPv4 addresses */
 	unsigned long		probed;		/* Mask of servers that have been probed */
 	unsigned long		yfs;		/* Mask of servers that are YFS */
 	struct sockaddr_rxrpc	addrs[];
+#define AFS_MAX_ADDRESSES ((unsigned int)(sizeof(unsigned long) * 8))
 };
 
 /*

[PATCH net-next 4/9] afs: Always build address lists using the helper functions

[David Howells]

Make the address list string parser use the helper functions for adding
addresses to an address list so that they end up appropriately sorted.
This will better handles overruns and make them easier to compare.

It also reduces the number of places that addresses are handled, making it
easier to fix the handling.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 fs/afs/addr_list.c |   32 +++++++++++++++-----------------
 1 file changed, 15 insertions(+), 17 deletions(-)

diff --git a/fs/afs/addr_list.c b/fs/afs/addr_list.c
index 4dbb8af54668..00e87f859b9f 100644
--- a/fs/afs/addr_list.c
+++ b/fs/afs/addr_list.c
@@ -116,8 +116,10 @@ struct afs_addr_list *afs_parse_text_addrs(const char *text, size_t len,
 	/* Extract the addresses */
 	p = text;
 	do {
-		struct sockaddr_rxrpc *srx = &alist->addrs[alist->nr_addrs];
 		const char *q, *stop;
+		unsigned int xport = port;
+		__be32 x[4];
+		int family;
 
 		if (*p == delim) {
 			p++;
@@ -133,19 +135,12 @@ struct afs_addr_list *afs_parse_text_addrs(const char *text, size_t len,
 					break;
 		}
 
-		if (in4_pton(p, q - p,
-			     (u8 *)&srx->transport.sin6.sin6_addr.s6_addr32[3],
-			     -1, &stop)) {
-			srx->transport.sin6.sin6_addr.s6_addr32[0] = 0;
-			srx->transport.sin6.sin6_addr.s6_addr32[1] = 0;
-			srx->transport.sin6.sin6_addr.s6_addr32[2] = htonl(0xffff);
-		} else if (in6_pton(p, q - p,
-				    srx->transport.sin6.sin6_addr.s6_addr,
-				    -1, &stop)) {
-			/* Nothing to do */
-		} else {
+		if (in4_pton(p, q - p, (u8 *)&x[0], -1, &stop))
+			family = AF_INET;
+		else if (in6_pton(p, q - p, (u8 *)x, -1, &stop))
+			family = AF_INET6;
+		else
 			goto bad_address;
-		}
 
 		if (stop != q)
 			goto bad_address;
@@ -157,7 +152,7 @@ struct afs_addr_list *afs_parse_text_addrs(const char *text, size_t len,
 		if (p < end) {
 			if (*p == '+') {
 				/* Port number specification "+1234" */
-				unsigned int xport = 0;
+				xport = 0;
 				p++;
 				if (p >= end || !isdigit(*p))
 					goto bad_address;
@@ -168,7 +163,6 @@ struct afs_addr_list *afs_parse_text_addrs(const char *text, size_t len,
 						goto bad_address;
 					p++;
 				} while (p < end && isdigit(*p));
-				srx->transport.sin6.sin6_port = htons(xport);
 			} else if (*p == delim) {
 				p++;
 			} else {
@@ -176,8 +170,12 @@ struct afs_addr_list *afs_parse_text_addrs(const char *text, size_t len,
 			}
 		}
 
-		alist->nr_addrs++;
-	} while (p < end && alist->nr_addrs < alist->max_addrs);
+		if (family == AF_INET)
+			afs_merge_fs_addr4(alist, x[0], xport);
+		else
+			afs_merge_fs_addr6(alist, x, xport);
+
+	} while (p < end);
 
 	_leave(" = [nr %u]", alist->nr_addrs);
 	return alist;

[PATCH net-next 5/9] afs: Sort address lists so that they are in logical ascending order

[David Howells]

Sort address lists so that they are in logical ascending order rather than
being partially in ascending order of the BE representations of those
values.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 fs/afs/addr_list.c |   51 ++++++++++++++++++++++++---------------------------
 1 file changed, 24 insertions(+), 27 deletions(-)

diff --git a/fs/afs/addr_list.c b/fs/afs/addr_list.c
index 00e87f859b9f..89374de4785c 100644
--- a/fs/afs/addr_list.c
+++ b/fs/afs/addr_list.c
@@ -232,22 +232,23 @@ struct afs_addr_list *afs_dns_query(struct afs_cell *cell, time64_t *_expiry)
  */
 void afs_merge_fs_addr4(struct afs_addr_list *alist, __be32 xdr, u16 port)
 {
-	struct sockaddr_in6 *a;
-	__be16 xport = htons(port);
+	struct sockaddr_in6 *p;
+	u32 addr = ntohl(xdr);
 	int i;
 
 	if (alist->nr_addrs >= alist->max_addrs)
 		return;
 
 	for (i = 0; i < alist->nr_ipv4; i++) {
-		a = &alist->addrs[i].transport.sin6;
-		if (xdr == a->sin6_addr.s6_addr32[3] &&
-		    xport == a->sin6_port)
+		struct sockaddr_in6 *a = &alist->addrs[i].transport.sin6;
+		u32 a_addr = ntohl(a->sin6_addr.s6_addr32[3]);
+		u16 a_port = ntohs(a->sin6_port);
+
+		if (addr == a_addr && port == a_port)
 			return;
-		if (xdr == a->sin6_addr.s6_addr32[3] &&
-		    (u16 __force)xport < (u16 __force)a->sin6_port)
+		if (addr == a_addr && port < a_port)
 			break;
-		if ((u32 __force)xdr < (u32 __force)a->sin6_addr.s6_addr32[3])
+		if (addr < a_addr)
 			break;
 	}
 
@@ -256,12 +257,12 @@ void afs_merge_fs_addr4(struct afs_addr_list *alist, __be32 xdr, u16 port)
 			alist->addrs + i,
 			sizeof(alist->addrs[0]) * (alist->nr_addrs - i));
 
-	a = &alist->addrs[i].transport.sin6;
-	a->sin6_port		  = xport;
-	a->sin6_addr.s6_addr32[0] = 0;
-	a->sin6_addr.s6_addr32[1] = 0;
-	a->sin6_addr.s6_addr32[2] = htonl(0xffff);
-	a->sin6_addr.s6_addr32[3] = xdr;
+	p = &alist->addrs[i].transport.sin6;
+	p->sin6_port		  = htons(port);
+	p->sin6_addr.s6_addr32[0] = 0;
+	p->sin6_addr.s6_addr32[1] = 0;
+	p->sin6_addr.s6_addr32[2] = htonl(0xffff);
+	p->sin6_addr.s6_addr32[3] = xdr;
 	alist->nr_ipv4++;
 	alist->nr_addrs++;
 }
@@ -271,21 +272,20 @@ void afs_merge_fs_addr4(struct afs_addr_list *alist, __be32 xdr, u16 port)
  */
 void afs_merge_fs_addr6(struct afs_addr_list *alist, __be32 *xdr, u16 port)
 {
-	struct sockaddr_in6 *a;
-	__be16 xport = htons(port);
+	struct sockaddr_in6 *p;
 	int i, diff;
 
 	if (alist->nr_addrs >= alist->max_addrs)
 		return;
 
 	for (i = alist->nr_ipv4; i < alist->nr_addrs; i++) {
-		a = &alist->addrs[i].transport.sin6;
+		struct sockaddr_in6 *a = &alist->addrs[i].transport.sin6;
+		u16 a_port = ntohs(a->sin6_port);
+
 		diff = memcmp(xdr, &a->sin6_addr, 16);
-		if (diff == 0 &&
-		    xport == a->sin6_port)
+		if (diff == 0 && port == a_port)
 			return;
-		if (diff == 0 &&
-		    (u16 __force)xport < (u16 __force)a->sin6_port)
+		if (diff == 0 && port < a_port)
 			break;
 		if (diff < 0)
 			break;
@@ -296,12 +296,9 @@ void afs_merge_fs_addr6(struct afs_addr_list *alist, __be32 *xdr, u16 port)
 			alist->addrs + i,
 			sizeof(alist->addrs[0]) * (alist->nr_addrs - i));
 
-	a = &alist->addrs[i].transport.sin6;
-	a->sin6_port		  = xport;
-	a->sin6_addr.s6_addr32[0] = xdr[0];
-	a->sin6_addr.s6_addr32[1] = xdr[1];
-	a->sin6_addr.s6_addr32[2] = xdr[2];
-	a->sin6_addr.s6_addr32[3] = xdr[3];
+	p = &alist->addrs[i].transport.sin6;
+	p->sin6_port = htons(port);
+	memcpy(&p->sin6_addr, xdr, 16);
 	alist->nr_addrs++;
 }
 

[PATCH net-next 6/9] rxrpc: Use IPv4 addresses throught the IPv6

[David Howells]

AF_RXRPC opens an IPv6 socket through which to send and receive network
packets, both IPv6 and IPv4.  It currently turns AF_INET addresses into
AF_INET-as-AF_INET6 addresses based on an assumption that this was
necessary; on further inspection of the code, however, it turns out that
the IPv6 code just farms packets aimed at AF_INET addresses out to the IPv4
code.

Fix AF_RXRPC to use AF_INET addresses directly when given them.

Fixes: 7b674e390e51 ("rxrpc: Fix IPv6 support")
Signed-off-by: David Howells <dhowells@redhat.com>
---

 fs/afs/addr_list.c      |   29 +++++++++++++++--------------
 net/rxrpc/af_rxrpc.c    |    3 ++-
 net/rxrpc/conn_object.c |    5 +++--
 net/rxrpc/peer_event.c  |   12 +++++++-----
 net/rxrpc/utils.c       |   19 +++++--------------
 5 files changed, 32 insertions(+), 36 deletions(-)

diff --git a/fs/afs/addr_list.c b/fs/afs/addr_list.c
index 89374de4785c..55a756c60746 100644
--- a/fs/afs/addr_list.c
+++ b/fs/afs/addr_list.c
@@ -232,7 +232,7 @@ struct afs_addr_list *afs_dns_query(struct afs_cell *cell, time64_t *_expiry)
  */
 void afs_merge_fs_addr4(struct afs_addr_list *alist, __be32 xdr, u16 port)
 {
-	struct sockaddr_in6 *p;
+	struct sockaddr_rxrpc *srx;
 	u32 addr = ntohl(xdr);
 	int i;
 
@@ -240,9 +240,9 @@ void afs_merge_fs_addr4(struct afs_addr_list *alist, __be32 xdr, u16 port)
 		return;
 
 	for (i = 0; i < alist->nr_ipv4; i++) {
-		struct sockaddr_in6 *a = &alist->addrs[i].transport.sin6;
-		u32 a_addr = ntohl(a->sin6_addr.s6_addr32[3]);
-		u16 a_port = ntohs(a->sin6_port);
+		struct sockaddr_in *a = &alist->addrs[i].transport.sin;
+		u32 a_addr = ntohl(a->sin_addr.s_addr);
+		u16 a_port = ntohs(a->sin_port);
 
 		if (addr == a_addr && port == a_port)
 			return;
@@ -257,12 +257,11 @@ void afs_merge_fs_addr4(struct afs_addr_list *alist, __be32 xdr, u16 port)
 			alist->addrs + i,
 			sizeof(alist->addrs[0]) * (alist->nr_addrs - i));
 
-	p = &alist->addrs[i].transport.sin6;
-	p->sin6_port		  = htons(port);
-	p->sin6_addr.s6_addr32[0] = 0;
-	p->sin6_addr.s6_addr32[1] = 0;
-	p->sin6_addr.s6_addr32[2] = htonl(0xffff);
-	p->sin6_addr.s6_addr32[3] = xdr;
+	srx = &alist->addrs[i];
+	srx->transport_len = sizeof(srx->transport.sin);
+	srx->transport.sin.sin_family = AF_INET;
+	srx->transport.sin.sin_port = htons(port);
+	srx->transport.sin.sin_addr.s_addr = xdr;
 	alist->nr_ipv4++;
 	alist->nr_addrs++;
 }
@@ -272,7 +271,7 @@ void afs_merge_fs_addr4(struct afs_addr_list *alist, __be32 xdr, u16 port)
  */
 void afs_merge_fs_addr6(struct afs_addr_list *alist, __be32 *xdr, u16 port)
 {
-	struct sockaddr_in6 *p;
+	struct sockaddr_rxrpc *srx;
 	int i, diff;
 
 	if (alist->nr_addrs >= alist->max_addrs)
@@ -296,9 +295,11 @@ void afs_merge_fs_addr6(struct afs_addr_list *alist, __be32 *xdr, u16 port)
 			alist->addrs + i,
 			sizeof(alist->addrs[0]) * (alist->nr_addrs - i));
 
-	p = &alist->addrs[i].transport.sin6;
-	p->sin6_port = htons(port);
-	memcpy(&p->sin6_addr, xdr, 16);
+	srx = &alist->addrs[i];
+	srx->transport_len = sizeof(srx->transport.sin6);
+	srx->transport.sin6.sin6_family = AF_INET6;
+	srx->transport.sin6.sin6_port = htons(port);
+	memcpy(&srx->transport.sin6.sin6_addr, xdr, 16);
 	alist->nr_addrs++;
 }
 
diff --git a/net/rxrpc/af_rxrpc.c b/net/rxrpc/af_rxrpc.c
index ac44d8afffb1..2fdd276f6842 100644
--- a/net/rxrpc/af_rxrpc.c
+++ b/net/rxrpc/af_rxrpc.c
@@ -97,7 +97,8 @@ static int rxrpc_validate_address(struct rxrpc_sock *rx,
 	    srx->transport_len > len)
 		return -EINVAL;
 
-	if (srx->transport.family != rx->family)
+	if (srx->transport.family != rx->family &&
+	    srx->transport.family == AF_INET && rx->family != AF_INET6)
 		return -EAFNOSUPPORT;
 
 	switch (srx->transport.family) {
diff --git a/net/rxrpc/conn_object.c b/net/rxrpc/conn_object.c
index 885dae829f4a..fb856a1ccda8 100644
--- a/net/rxrpc/conn_object.c
+++ b/net/rxrpc/conn_object.c
@@ -89,8 +89,9 @@ struct rxrpc_connection *rxrpc_find_connection_rcu(struct rxrpc_local *local,
 	if (rxrpc_extract_addr_from_skb(local, &srx, skb) < 0)
 		goto not_found;
 
-	/* We may have to handle mixing IPv4 and IPv6 */
-	if (srx.transport.family != local->srx.transport.family) {
+	if (srx.transport.family != local->srx.transport.family &&
+	    (srx.transport.family == AF_INET &&
+	     local->srx.transport.family != AF_INET6)) {
 		pr_warn_ratelimited("AF_RXRPC: Protocol mismatch %u not %u\n",
 				    srx.transport.family,
 				    local->srx.transport.family);
diff --git a/net/rxrpc/peer_event.c b/net/rxrpc/peer_event.c
index f3e6fc670da2..81a7869325a6 100644
--- a/net/rxrpc/peer_event.c
+++ b/net/rxrpc/peer_event.c
@@ -47,6 +47,8 @@ static struct rxrpc_peer *rxrpc_lookup_peer_icmp_rcu(struct rxrpc_local *local,
 	 */
 	switch (srx->transport.family) {
 	case AF_INET:
+		srx->transport_len = sizeof(srx->transport.sin);
+		srx->transport.family = AF_INET;
 		srx->transport.sin.sin_port = serr->port;
 		switch (serr->ee.ee_origin) {
 		case SO_EE_ORIGIN_ICMP:
@@ -70,20 +72,20 @@ static struct rxrpc_peer *rxrpc_lookup_peer_icmp_rcu(struct rxrpc_local *local,
 
 #ifdef CONFIG_AF_RXRPC_IPV6
 	case AF_INET6:
-		srx->transport.sin6.sin6_port = serr->port;
 		switch (serr->ee.ee_origin) {
 		case SO_EE_ORIGIN_ICMP6:
 			_net("Rx ICMP6");
+			srx->transport.sin6.sin6_port = serr->port;
 			memcpy(&srx->transport.sin6.sin6_addr,
 			       skb_network_header(skb) + serr->addr_offset,
 			       sizeof(struct in6_addr));
 			break;
 		case SO_EE_ORIGIN_ICMP:
 			_net("Rx ICMP on v6 sock");
-			srx->transport.sin6.sin6_addr.s6_addr32[0] = 0;
-			srx->transport.sin6.sin6_addr.s6_addr32[1] = 0;
-			srx->transport.sin6.sin6_addr.s6_addr32[2] = htonl(0xffff);
-			memcpy(srx->transport.sin6.sin6_addr.s6_addr + 12,
+			srx->transport_len = sizeof(srx->transport.sin);
+			srx->transport.family = AF_INET;
+			srx->transport.sin.sin_port = serr->port;
+			memcpy(&srx->transport.sin.sin_addr,
 			       skb_network_header(skb) + serr->addr_offset,
 			       sizeof(struct in_addr));
 			break;
diff --git a/net/rxrpc/utils.c b/net/rxrpc/utils.c
index e801171fa351..017adaa54e90 100644
--- a/net/rxrpc/utils.c
+++ b/net/rxrpc/utils.c
@@ -25,20 +25,11 @@ int rxrpc_extract_addr_from_skb(struct rxrpc_local *local,
 
 	switch (ntohs(skb->protocol)) {
 	case ETH_P_IP:
-		if (local->srx.transport.family == AF_INET6) {
-			srx->transport_type = SOCK_DGRAM;
-			srx->transport_len = sizeof(srx->transport.sin6);
-			srx->transport.sin6.sin6_family = AF_INET6;
-			srx->transport.sin6.sin6_port = udp_hdr(skb)->source;
-			srx->transport.sin6.sin6_addr.s6_addr32[2] = htonl(0xffff);
-			srx->transport.sin6.sin6_addr.s6_addr32[3] = ip_hdr(skb)->saddr;
-		} else {
-			srx->transport_type = SOCK_DGRAM;
-			srx->transport_len = sizeof(srx->transport.sin);
-			srx->transport.sin.sin_family = AF_INET;
-			srx->transport.sin.sin_port = udp_hdr(skb)->source;
-			srx->transport.sin.sin_addr.s_addr = ip_hdr(skb)->saddr;
-		}
+		srx->transport_type = SOCK_DGRAM;
+		srx->transport_len = sizeof(srx->transport.sin);
+		srx->transport.sin.sin_family = AF_INET;
+		srx->transport.sin.sin_port = udp_hdr(skb)->source;
+		srx->transport.sin.sin_addr.s_addr = ip_hdr(skb)->saddr;
 		return 0;
 
 #ifdef CONFIG_AF_RXRPC_IPV6

[PATCH net-next 7/9] rxrpc: Drop the local endpoint arg from rxrpc_extract_addr_from_skb()

[David Howells]

rxrpc_extract_addr_from_skb() doesn't use the argument that points to the
local endpoint, so remove the argument.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 net/rxrpc/ar-internal.h |    3 +--
 net/rxrpc/call_accept.c |    2 +-
 net/rxrpc/conn_object.c |    2 +-
 net/rxrpc/local_event.c |    2 +-
 net/rxrpc/output.c      |    2 +-
 net/rxrpc/utils.c       |    4 +---
 6 files changed, 6 insertions(+), 9 deletions(-)

diff --git a/net/rxrpc/ar-internal.h b/net/rxrpc/ar-internal.h
index 9ba87e4d15c7..76569c178915 100644
--- a/net/rxrpc/ar-internal.h
+++ b/net/rxrpc/ar-internal.h
@@ -1111,8 +1111,7 @@ static inline void rxrpc_sysctl_exit(void) {}
 /*
  * utils.c
  */
-int rxrpc_extract_addr_from_skb(struct rxrpc_local *, struct sockaddr_rxrpc *,
-				struct sk_buff *);
+int rxrpc_extract_addr_from_skb(struct sockaddr_rxrpc *, struct sk_buff *);
 
 static inline bool before(u32 seq1, u32 seq2)
 {
diff --git a/net/rxrpc/call_accept.c b/net/rxrpc/call_accept.c
index 9c7f26d06a52..8354cadbb839 100644
--- a/net/rxrpc/call_accept.c
+++ b/net/rxrpc/call_accept.c
@@ -280,7 +280,7 @@ static struct rxrpc_call *rxrpc_alloc_incoming_call(struct rxrpc_sock *rx,
 			peer = NULL;
 		if (!peer) {
 			peer = b->peer_backlog[peer_tail];
-			if (rxrpc_extract_addr_from_skb(local, &peer->srx, skb) < 0)
+			if (rxrpc_extract_addr_from_skb(&peer->srx, skb) < 0)
 				return NULL;
 			b->peer_backlog[peer_tail] = NULL;
 			smp_store_release(&b->peer_backlog_tail,
diff --git a/net/rxrpc/conn_object.c b/net/rxrpc/conn_object.c
index fb856a1ccda8..c332722820c2 100644
--- a/net/rxrpc/conn_object.c
+++ b/net/rxrpc/conn_object.c
@@ -86,7 +86,7 @@ struct rxrpc_connection *rxrpc_find_connection_rcu(struct rxrpc_local *local,
 
 	_enter(",%x", sp->hdr.cid & RXRPC_CIDMASK);
 
-	if (rxrpc_extract_addr_from_skb(local, &srx, skb) < 0)
+	if (rxrpc_extract_addr_from_skb(&srx, skb) < 0)
 		goto not_found;
 
 	if (srx.transport.family != local->srx.transport.family &&
diff --git a/net/rxrpc/local_event.c b/net/rxrpc/local_event.c
index 13bd8a4dfac7..927ead43df42 100644
--- a/net/rxrpc/local_event.c
+++ b/net/rxrpc/local_event.c
@@ -39,7 +39,7 @@ static void rxrpc_send_version_request(struct rxrpc_local *local,
 
 	_enter("");
 
-	if (rxrpc_extract_addr_from_skb(local, &srx, skb) < 0)
+	if (rxrpc_extract_addr_from_skb(&srx, skb) < 0)
 		return;
 
 	msg.msg_name	= &srx.transport;
diff --git a/net/rxrpc/output.c b/net/rxrpc/output.c
index 993d4cd247f9..0f0b499d1202 100644
--- a/net/rxrpc/output.c
+++ b/net/rxrpc/output.c
@@ -561,7 +561,7 @@ void rxrpc_reject_packets(struct rxrpc_local *local)
 			continue;
 		}
 
-		if (rxrpc_extract_addr_from_skb(local, &srx, skb) == 0) {
+		if (rxrpc_extract_addr_from_skb(&srx, skb) == 0) {
 			msg.msg_namelen = srx.transport_len;
 
 			whdr.epoch	= htonl(sp->hdr.epoch);
diff --git a/net/rxrpc/utils.c b/net/rxrpc/utils.c
index 017adaa54e90..ff7af71c4b49 100644
--- a/net/rxrpc/utils.c
+++ b/net/rxrpc/utils.c
@@ -17,9 +17,7 @@
 /*
  * Fill out a peer address from a socket buffer containing a packet.
  */
-int rxrpc_extract_addr_from_skb(struct rxrpc_local *local,
-				struct sockaddr_rxrpc *srx,
-				struct sk_buff *skb)
+int rxrpc_extract_addr_from_skb(struct sockaddr_rxrpc *srx, struct sk_buff *skb)
 {
 	memset(srx, 0, sizeof(*srx));
 

[PATCH net-next 8/9] rxrpc: Allow the reply time to be obtained on a client call

[David Howells]

Allow the timestamp on the sk_buff holding the first DATA packet of a reply
to be queried.  This can then be used as a base for the expiry time
calculation on the callback promise duration indicated by an operation
result.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 Documentation/networking/rxrpc.txt |   11 +++++++++
 include/net/af_rxrpc.h             |    3 +++
 net/rxrpc/recvmsg.c                |   43 ++++++++++++++++++++++++++++++++++++
 3 files changed, 57 insertions(+)

diff --git a/Documentation/networking/rxrpc.txt b/Documentation/networking/rxrpc.txt
index b5407163d53b..67879992b4c2 100644
--- a/Documentation/networking/rxrpc.txt
+++ b/Documentation/networking/rxrpc.txt
@@ -1069,6 +1069,17 @@ The kernel interface functions are as follows:
 
      This function may transmit a PING ACK.
 
+ (*) Get reply timestamp.
+
+	bool rxrpc_kernel_get_reply_time(struct socket *sock,
+					 struct rxrpc_call *call,
+					 ktime_t *_ts)
+
+     This allows the timestamp on the first DATA packet of the reply of a
+     client call to be queried, provided that it is still in the Rx ring.  If
+     successful, the timestamp will be stored into *_ts and true will be
+     returned; false will be returned otherwise.
+
 
 =======================
 CONFIGURABLE PARAMETERS
diff --git a/include/net/af_rxrpc.h b/include/net/af_rxrpc.h
index f53edb3754bc..c4c912554dee 100644
--- a/include/net/af_rxrpc.h
+++ b/include/net/af_rxrpc.h
@@ -13,6 +13,7 @@
 #define _NET_RXRPC_H
 
 #include <linux/rxrpc.h>
+#include <linux/ktime.h>
 
 struct key;
 struct sock;
@@ -77,5 +78,7 @@ int rxrpc_kernel_retry_call(struct socket *, struct rxrpc_call *,
 int rxrpc_kernel_check_call(struct socket *, struct rxrpc_call *,
 			    enum rxrpc_call_completion *, u32 *);
 u32 rxrpc_kernel_check_life(struct socket *, struct rxrpc_call *);
+bool rxrpc_kernel_get_reply_time(struct socket *, struct rxrpc_call *,
+				 ktime_t *);
 
 #endif /* _NET_RXRPC_H */
diff --git a/net/rxrpc/recvmsg.c b/net/rxrpc/recvmsg.c
index 816b19a78809..eaf19ebaa964 100644
--- a/net/rxrpc/recvmsg.c
+++ b/net/rxrpc/recvmsg.c
@@ -715,3 +715,46 @@ int rxrpc_kernel_recv_data(struct socket *sock, struct rxrpc_call *call,
 	goto out;
 }
 EXPORT_SYMBOL(rxrpc_kernel_recv_data);
+
+/**
+ * rxrpc_kernel_get_reply_time - Get timestamp on first reply packet
+ * @sock: The socket that the call exists on
+ * @call: The call to query
+ * @_ts: Where to put the timestamp
+ *
+ * Retrieve the timestamp from the first DATA packet of the reply if it is
+ * in the ring.  Returns true if successful, false if not.
+ */
+bool rxrpc_kernel_get_reply_time(struct socket *sock, struct rxrpc_call *call,
+				 ktime_t *_ts)
+{
+	struct sk_buff *skb;
+	rxrpc_seq_t hard_ack, top, seq;
+	bool success = false;
+
+	mutex_lock(&call->user_mutex);
+
+	if (READ_ONCE(call->state) != RXRPC_CALL_CLIENT_RECV_REPLY)
+		goto out;
+
+	hard_ack = call->rx_hard_ack;
+	if (hard_ack != 0)
+		goto out;
+
+	seq = hard_ack + 1;
+	top = smp_load_acquire(&call->rx_top);
+	if (after(seq, top))
+		goto out;
+
+	skb = call->rxtx_buffer[seq & RXRPC_RXTX_BUFF_MASK];
+	if (!skb)
+		goto out;
+
+	*_ts = skb_get_ktime(skb);
+	success = true;
+
+out:
+	mutex_unlock(&call->user_mutex);
+	return success;
+}
+EXPORT_SYMBOL(rxrpc_kernel_get_reply_time);

[PATCH net-next 9/9] rxrpc: Allow the reply time to be obtained on a client call

[David Howells]

Allow the epoch value to be queried on a server connection.  This is in the
rxrpc header of every packet for use in routing and is derived from the
client's state.  It's also not supposed to change unless the client gets
restarted.

AFS can make use of this information to deduce whether a fileserver has
been restarted because the fileserver makes client calls to the filesystem
driver's cache manager to send notifications (ie. callback breaks) about
conflicting changes from other clients.  These convey the fileserver's own
epoch value back to the filesystem.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 Documentation/networking/rxrpc.txt |   14 ++++++++++++++
 include/net/af_rxrpc.h             |    1 +
 net/rxrpc/af_rxrpc.c               |   14 ++++++++++++++
 3 files changed, 29 insertions(+)

diff --git a/Documentation/networking/rxrpc.txt b/Documentation/networking/rxrpc.txt
index 67879992b4c2..605e00cdd6be 100644
--- a/Documentation/networking/rxrpc.txt
+++ b/Documentation/networking/rxrpc.txt
@@ -1080,6 +1080,20 @@ The kernel interface functions are as follows:
      successful, the timestamp will be stored into *_ts and true will be
      returned; false will be returned otherwise.
 
+ (*) Get remote client epoch.
+
+	u32 rxrpc_kernel_get_epoch(struct socket *sock,
+				   struct rxrpc_call *call)
+
+     This allows the epoch that's contained in packets of an incoming client
+     call to be queried.  This value is returned.  The function always
+     successful if the call is still in progress.  It shouldn't be called once
+     the call has expired.  Note that calling this on a local client call only
+     returns the local epoch.
+
+     This value can be used to determine if the remote client has been
+     restarted as it shouldn't change otherwise.
+
 
 =======================
 CONFIGURABLE PARAMETERS
diff --git a/include/net/af_rxrpc.h b/include/net/af_rxrpc.h
index c4c912554dee..de587948042a 100644
--- a/include/net/af_rxrpc.h
+++ b/include/net/af_rxrpc.h
@@ -78,6 +78,7 @@ int rxrpc_kernel_retry_call(struct socket *, struct rxrpc_call *,
 int rxrpc_kernel_check_call(struct socket *, struct rxrpc_call *,
 			    enum rxrpc_call_completion *, u32 *);
 u32 rxrpc_kernel_check_life(struct socket *, struct rxrpc_call *);
+u32 rxrpc_kernel_get_epoch(struct socket *, struct rxrpc_call *);
 bool rxrpc_kernel_get_reply_time(struct socket *, struct rxrpc_call *,
 				 ktime_t *);
 
diff --git a/net/rxrpc/af_rxrpc.c b/net/rxrpc/af_rxrpc.c
index 2fdd276f6842..013dbcb052e5 100644
--- a/net/rxrpc/af_rxrpc.c
+++ b/net/rxrpc/af_rxrpc.c
@@ -385,6 +385,20 @@ u32 rxrpc_kernel_check_life(struct socket *sock, struct rxrpc_call *call)
 }
 EXPORT_SYMBOL(rxrpc_kernel_check_life);
 
+/**
+ * rxrpc_kernel_get_epoch - Retrieve the epoch value from a call.
+ * @sock: The socket the call is on
+ * @call: The call to query
+ *
+ * Allow a kernel service to retrieve the epoch value from a service call to
+ * see if the client at the other end rebooted.
+ */
+u32 rxrpc_kernel_get_epoch(struct socket *sock, struct rxrpc_call *call)
+{
+	return call->conn->proto.epoch;
+}
+EXPORT_SYMBOL(rxrpc_kernel_get_epoch);
+
 /**
  * rxrpc_kernel_check_call - Check a call's state
  * @sock: The socket the call is on

Re: [PATCH net-next 0/9] rxrpc: Development

[David Miller]

From: David Howells <dhowells@redhat.com>
Date: Thu, 04 Oct 2018 14:50:31 +0100

> Here are some development patches for AF_RXRPC.  The most significant points
> are:
 ...
> The patches are tagged here:
> 
> 	git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git
> 	rxrpc-next-20181004

Pulled, thanks David.