[PATCH net 0/6] rxrpc: Miscellaneous fixes

[PATCH net 0/6] rxrpc: Miscellaneous fixes

[David Howells]

Here are some miscellaneous rxrpc fixes:

 (1) Fix the xdr encoding of the contents read from an rxrpc key.

 (2) Fix a BUG() for a unsupported encoding type.

 (3) Fix missing _bh lock annotations.

 (4) Fix acceptance handling for an incoming call where the incoming call
     is encrypted.

 (5) The server token keyring isn't network namespaced - it belongs to the
     server, so there's no need.  Namespacing it means that request_key()
     fails to find it.

 (6) Fix a leak of the server keyring.

The patches are tagged here:

	git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git
	rxrpc-fixes-20201005.txt

and can also be found on the following branch:

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

David
---
David Howells (5):
      rxrpc: Downgrade the BUG() for unsupported token type in rxrpc_read()
      rxrpc: Fix some missing _bh annotations on locking conn->state_lock
      rxrpc: Fix accept on a connection that need securing
      rxrpc: The server keyring isn't network-namespaced
      rxrpc: Fix server keyring leak

Marc Dionne (1):
      rxrpc: Fix rxkad token xdr encoding


 include/uapi/linux/rxrpc.h |   2 +-
 net/rxrpc/ar-internal.h    |   7 +-
 net/rxrpc/call_accept.c    | 263 ++++++-------------------------------
 net/rxrpc/call_object.c    |   5 +-
 net/rxrpc/conn_event.c     |   8 +-
 net/rxrpc/key.c            |   8 +-
 net/rxrpc/recvmsg.c        |  36 +----
 net/rxrpc/sendmsg.c        |  15 +--
 8 files changed, 56 insertions(+), 288 deletions(-)

[PATCH net 1/6] rxrpc: Fix rxkad token xdr encoding

[David Howells]

From: Marc Dionne <marc.dionne@auristor.com>

The session key should be encoded with just the 8 data bytes and
no length; ENCODE_DATA precedes it with a 4 byte length, which
confuses some existing tools that try to parse this format.

Add an ENCODE_BYTES macro that does not include a length, and use
it for the key.  Also adjust the expected length.

Note that commit 774521f353e1d ("rxrpc: Fix an assertion in
rxrpc_read()") had fixed a BUG by changing the length rather than
fixing the encoding.  The original length was correct.

Fixes: 99455153d067 ("RxRPC: Parse security index 5 keys (Kerberos 5)")
Signed-off-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
---

 net/rxrpc/key.c |   12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/net/rxrpc/key.c b/net/rxrpc/key.c
index 94c3df392651..8f7d7a6187db 100644
--- a/net/rxrpc/key.c
+++ b/net/rxrpc/key.c
@@ -1072,7 +1072,7 @@ static long rxrpc_read(const struct key *key,
 
 		switch (token->security_index) {
 		case RXRPC_SECURITY_RXKAD:
-			toksize += 9 * 4;	/* viceid, kvno, key*2 + len, begin,
+			toksize += 8 * 4;	/* viceid, kvno, key*2, begin,
 						 * end, primary, tktlen */
 			toksize += RND(token->kad->ticket_len);
 			break;
@@ -1138,6 +1138,14 @@ static long rxrpc_read(const struct key *key,
 			memcpy((u8 *)xdr + _l, &zero, 4 - (_l & 3));	\
 		xdr += (_l + 3) >> 2;					\
 	} while(0)
+#define ENCODE_BYTES(l, s)						\
+	do {								\
+		u32 _l = (l);						\
+		memcpy(xdr, (s), _l);					\
+		if (_l & 3)						\
+			memcpy((u8 *)xdr + _l, &zero, 4 - (_l & 3));	\
+		xdr += (_l + 3) >> 2;					\
+	} while(0)
 #define ENCODE64(x)					\
 	do {						\
 		__be64 y = cpu_to_be64(x);		\
@@ -1165,7 +1173,7 @@ static long rxrpc_read(const struct key *key,
 		case RXRPC_SECURITY_RXKAD:
 			ENCODE(token->kad->vice_id);
 			ENCODE(token->kad->kvno);
-			ENCODE_DATA(8, token->kad->session_key);
+			ENCODE_BYTES(8, token->kad->session_key);
 			ENCODE(token->kad->start);
 			ENCODE(token->kad->expiry);
 			ENCODE(token->kad->primary_flag);

[PATCH net 2/6] rxrpc: Downgrade the BUG() for unsupported token type in rxrpc_read()

[David Howells]

If rxrpc_read() (which allows KEYCTL_READ to read a key), sees a token of a
type it doesn't recognise, it can BUG in a couple of places, which is
unnecessary as it can easily get back to userspace.

Fix this to print an error message instead.

Fixes: 99455153d067 ("RxRPC: Parse security index 5 keys (Kerberos 5)")
Signed-off-by: David Howells <dhowells@redhat.com>
---

 net/rxrpc/key.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/rxrpc/key.c b/net/rxrpc/key.c
index 8f7d7a6187db..c668e4b7dbff 100644
--- a/net/rxrpc/key.c
+++ b/net/rxrpc/key.c
@@ -1107,7 +1107,8 @@ static long rxrpc_read(const struct key *key,
 			break;
 
 		default: /* we have a ticket we can't encode */
-			BUG();
+			pr_err("Unsupported key token type (%u)\n",
+			       token->security_index);
 			continue;
 		}
 
@@ -1223,7 +1224,6 @@ static long rxrpc_read(const struct key *key,
 			break;
 
 		default:
-			BUG();
 			break;
 		}
 

[PATCH net 3/6] rxrpc: Fix some missing _bh annotations on locking conn->state_lock

[David Howells]

conn->state_lock may be taken in softirq mode, but a previous patch
replaced an outer lock in the response-packet event handling code, and lost
the _bh from that when doing so.

Fix this by applying the _bh annotation to the state_lock locking.

Fixes: a1399f8bb033 ("rxrpc: Call channels should have separate call number spaces")
Signed-off-by: David Howells <dhowells@redhat.com>
---

 net/rxrpc/conn_event.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/net/rxrpc/conn_event.c b/net/rxrpc/conn_event.c
index 447f55ca6886..6e972b4823ef 100644
--- a/net/rxrpc/conn_event.c
+++ b/net/rxrpc/conn_event.c
@@ -340,18 +340,18 @@ static int rxrpc_process_event(struct rxrpc_connection *conn,
 			return ret;
 
 		spin_lock(&conn->channel_lock);
-		spin_lock(&conn->state_lock);
+		spin_lock_bh(&conn->state_lock);
 
 		if (conn->state == RXRPC_CONN_SERVICE_CHALLENGING) {
 			conn->state = RXRPC_CONN_SERVICE;
-			spin_unlock(&conn->state_lock);
+			spin_unlock_bh(&conn->state_lock);
 			for (loop = 0; loop < RXRPC_MAXCALLS; loop++)
 				rxrpc_call_is_secure(
 					rcu_dereference_protected(
 						conn->channels[loop].call,
 						lockdep_is_held(&conn->channel_lock)));
 		} else {
-			spin_unlock(&conn->state_lock);
+			spin_unlock_bh(&conn->state_lock);
 		}
 
 		spin_unlock(&conn->channel_lock);

[PATCH net 4/6] rxrpc: Fix accept on a connection that need securing

[David Howells]

When a new incoming call arrives at an userspace rxrpc socket on a new
connection that has a security class set, the code currently pushes it onto
the accept queue to hold a ref on it for the socket.  This doesn't work,
however, as recvmsg() pops it off, notices that it's in the SERVER_SECURING
state and discards the ref.  This means that the call runs out of refs too
early and the kernel oopses.

By contrast, a kernel rxrpc socket manually pre-charges the incoming call
pool with calls that already have user call IDs assigned, so they are ref'd
by the call tree on the socket.

Change the mode of operation for userspace rxrpc server sockets to work
like this too.  Although this is a UAPI change, server sockets aren't
currently functional.

Fixes: 248f219cb8bc ("rxrpc: Rewrite the data and ack handling code")
Signed-off-by: David Howells <dhowells@redhat.com>
---

 include/uapi/linux/rxrpc.h |    2 
 net/rxrpc/ar-internal.h    |    7 -
 net/rxrpc/call_accept.c    |  263 ++++++--------------------------------------
 net/rxrpc/call_object.c    |    5 -
 net/rxrpc/conn_event.c     |    2 
 net/rxrpc/recvmsg.c        |   36 ------
 net/rxrpc/sendmsg.c        |   15 +--
 7 files changed, 49 insertions(+), 281 deletions(-)

diff --git a/include/uapi/linux/rxrpc.h b/include/uapi/linux/rxrpc.h
index 4accfa7e266d..8f8dc7a937a4 100644
--- a/include/uapi/linux/rxrpc.h
+++ b/include/uapi/linux/rxrpc.h
@@ -51,11 +51,11 @@ enum rxrpc_cmsg_type {
 	RXRPC_BUSY		= 6,	/* -r: server busy received [terminal] */
 	RXRPC_LOCAL_ERROR	= 7,	/* -r: local error generated [terminal] */
 	RXRPC_NEW_CALL		= 8,	/* -r: [Service] new incoming call notification */
-	RXRPC_ACCEPT		= 9,	/* s-: [Service] accept request */
 	RXRPC_EXCLUSIVE_CALL	= 10,	/* s-: Call should be on exclusive connection */
 	RXRPC_UPGRADE_SERVICE	= 11,	/* s-: Request service upgrade for client call */
 	RXRPC_TX_LENGTH		= 12,	/* s-: Total length of Tx data */
 	RXRPC_SET_CALL_TIMEOUT	= 13,	/* s-: Set one or more call timeouts */
+	RXRPC_CHARGE_ACCEPT	= 14,	/* s-: Charge the accept pool with a user call ID */
 	RXRPC__SUPPORTED
 };
 
diff --git a/net/rxrpc/ar-internal.h b/net/rxrpc/ar-internal.h
index 884cff7bb169..97aebb5d19db 100644
--- a/net/rxrpc/ar-internal.h
+++ b/net/rxrpc/ar-internal.h
@@ -518,7 +518,6 @@ enum rxrpc_call_state {
 	RXRPC_CALL_CLIENT_RECV_REPLY,	/* - client receiving reply phase */
 	RXRPC_CALL_SERVER_PREALLOC,	/* - service preallocation */
 	RXRPC_CALL_SERVER_SECURING,	/* - server securing request connection */
-	RXRPC_CALL_SERVER_ACCEPTING,	/* - server accepting request */
 	RXRPC_CALL_SERVER_RECV_REQUEST,	/* - server receiving request */
 	RXRPC_CALL_SERVER_ACK_REQUEST,	/* - server pending ACK of request */
 	RXRPC_CALL_SERVER_SEND_REPLY,	/* - server sending reply */
@@ -714,8 +713,8 @@ struct rxrpc_ack_summary {
 enum rxrpc_command {
 	RXRPC_CMD_SEND_DATA,		/* send data message */
 	RXRPC_CMD_SEND_ABORT,		/* request abort generation */
-	RXRPC_CMD_ACCEPT,		/* [server] accept incoming call */
 	RXRPC_CMD_REJECT_BUSY,		/* [server] reject a call as busy */
+	RXRPC_CMD_CHARGE_ACCEPT,	/* [server] charge accept preallocation */
 };
 
 struct rxrpc_call_params {
@@ -755,9 +754,7 @@ struct rxrpc_call *rxrpc_new_incoming_call(struct rxrpc_local *,
 					   struct rxrpc_sock *,
 					   struct sk_buff *);
 void rxrpc_accept_incoming_calls(struct rxrpc_local *);
-struct rxrpc_call *rxrpc_accept_call(struct rxrpc_sock *, unsigned long,
-				     rxrpc_notify_rx_t);
-int rxrpc_reject_call(struct rxrpc_sock *);
+int rxrpc_user_charge_accept(struct rxrpc_sock *, unsigned long);
 
 /*
  * call_event.c
diff --git a/net/rxrpc/call_accept.c b/net/rxrpc/call_accept.c
index ef160566aa9a..8df1964db333 100644
--- a/net/rxrpc/call_accept.c
+++ b/net/rxrpc/call_accept.c
@@ -39,8 +39,9 @@ static int rxrpc_service_prealloc_one(struct rxrpc_sock *rx,
 				      unsigned int debug_id)
 {
 	const void *here = __builtin_return_address(0);
-	struct rxrpc_call *call;
+	struct rxrpc_call *call, *xcall;
 	struct rxrpc_net *rxnet = rxrpc_net(sock_net(&rx->sk));
+	struct rb_node *parent, **pp;
 	int max, tmp;
 	unsigned int size = RXRPC_BACKLOG_MAX;
 	unsigned int head, tail, call_head, call_tail;
@@ -94,7 +95,7 @@ static int rxrpc_service_prealloc_one(struct rxrpc_sock *rx,
 	}
 
 	/* Now it gets complicated, because calls get registered with the
-	 * socket here, particularly if a user ID is preassigned by the user.
+	 * socket here, with a user ID preassigned by the user.
 	 */
 	call = rxrpc_alloc_call(rx, gfp, debug_id);
 	if (!call)
@@ -107,34 +108,33 @@ static int rxrpc_service_prealloc_one(struct rxrpc_sock *rx,
 			 here, (const void *)user_call_ID);
 
 	write_lock(&rx->call_lock);
-	if (user_attach_call) {
-		struct rxrpc_call *xcall;
-		struct rb_node *parent, **pp;
-
-		/* Check the user ID isn't already in use */
-		pp = &rx->calls.rb_node;
-		parent = NULL;
-		while (*pp) {
-			parent = *pp;
-			xcall = rb_entry(parent, struct rxrpc_call, sock_node);
-			if (user_call_ID < xcall->user_call_ID)
-				pp = &(*pp)->rb_left;
-			else if (user_call_ID > xcall->user_call_ID)
-				pp = &(*pp)->rb_right;
-			else
-				goto id_in_use;
-		}
 
-		call->user_call_ID = user_call_ID;
-		call->notify_rx = notify_rx;
+	/* Check the user ID isn't already in use */
+	pp = &rx->calls.rb_node;
+	parent = NULL;
+	while (*pp) {
+		parent = *pp;
+		xcall = rb_entry(parent, struct rxrpc_call, sock_node);
+		if (user_call_ID < xcall->user_call_ID)
+			pp = &(*pp)->rb_left;
+		else if (user_call_ID > xcall->user_call_ID)
+			pp = &(*pp)->rb_right;
+		else
+			goto id_in_use;
+	}
+
+	call->user_call_ID = user_call_ID;
+	call->notify_rx = notify_rx;
+	if (user_attach_call) {
 		rxrpc_get_call(call, rxrpc_call_got_kernel);
 		user_attach_call(call, user_call_ID);
-		rxrpc_get_call(call, rxrpc_call_got_userid);
-		rb_link_node(&call->sock_node, parent, pp);
-		rb_insert_color(&call->sock_node, &rx->calls);
-		set_bit(RXRPC_CALL_HAS_USERID, &call->flags);
 	}
 
+	rxrpc_get_call(call, rxrpc_call_got_userid);
+	rb_link_node(&call->sock_node, parent, pp);
+	rb_insert_color(&call->sock_node, &rx->calls);
+	set_bit(RXRPC_CALL_HAS_USERID, &call->flags);
+
 	list_add(&call->sock_link, &rx->sock_calls);
 
 	write_unlock(&rx->call_lock);
@@ -157,11 +157,8 @@ static int rxrpc_service_prealloc_one(struct rxrpc_sock *rx,
 }
 
 /*
- * Preallocate sufficient service connections, calls and peers to cover the
- * entire backlog of a socket.  When a new call comes in, if we don't have
- * sufficient of each available, the call gets rejected as busy or ignored.
- *
- * The backlog is replenished when a connection is accepted or rejected.
+ * Allocate the preallocation buffers for incoming service calls.  These must
+ * be charged manually.
  */
 int rxrpc_service_prealloc(struct rxrpc_sock *rx, gfp_t gfp)
 {
@@ -174,13 +171,6 @@ int rxrpc_service_prealloc(struct rxrpc_sock *rx, gfp_t gfp)
 		rx->backlog = b;
 	}
 
-	if (rx->discard_new_call)
-		return 0;
-
-	while (rxrpc_service_prealloc_one(rx, b, NULL, NULL, 0, gfp,
-					  atomic_inc_return(&rxrpc_debug_id)) == 0)
-		;
-
 	return 0;
 }
 
@@ -333,6 +323,7 @@ static struct rxrpc_call *rxrpc_alloc_incoming_call(struct rxrpc_sock *rx,
 	rxrpc_see_call(call);
 	call->conn = conn;
 	call->security = conn->security;
+	call->security_ix = conn->security_ix;
 	call->peer = rxrpc_get_peer(conn->params.peer);
 	call->cong_cwnd = call->peer->cong_cwnd;
 	return call;
@@ -402,8 +393,6 @@ struct rxrpc_call *rxrpc_new_incoming_call(struct rxrpc_local *local,
 
 	if (rx->notify_new_call)
 		rx->notify_new_call(&rx->sk, call, call->user_call_ID);
-	else
-		sk_acceptq_added(&rx->sk);
 
 	spin_lock(&conn->state_lock);
 	switch (conn->state) {
@@ -415,12 +404,8 @@ struct rxrpc_call *rxrpc_new_incoming_call(struct rxrpc_local *local,
 
 	case RXRPC_CONN_SERVICE:
 		write_lock(&call->state_lock);
-		if (call->state < RXRPC_CALL_COMPLETE) {
-			if (rx->discard_new_call)
-				call->state = RXRPC_CALL_SERVER_RECV_REQUEST;
-			else
-				call->state = RXRPC_CALL_SERVER_ACCEPTING;
-		}
+		if (call->state < RXRPC_CALL_COMPLETE)
+			call->state = RXRPC_CALL_SERVER_RECV_REQUEST;
 		write_unlock(&call->state_lock);
 		break;
 
@@ -440,9 +425,6 @@ struct rxrpc_call *rxrpc_new_incoming_call(struct rxrpc_local *local,
 
 	rxrpc_send_ping(call, skb);
 
-	if (call->state == RXRPC_CALL_SERVER_ACCEPTING)
-		rxrpc_notify_socket(call);
-
 	/* We have to discard the prealloc queue's ref here and rely on a
 	 * combination of the RCU read lock and refs held either by the socket
 	 * (recvmsg queue, to-be-accepted queue or user ID tree) or the kernel
@@ -460,187 +442,18 @@ struct rxrpc_call *rxrpc_new_incoming_call(struct rxrpc_local *local,
 }
 
 /*
- * handle acceptance of a call by userspace
- * - assign the user call ID to the call at the front of the queue
- * - called with the socket locked.
+ * Charge up socket with preallocated calls, attaching user call IDs.
  */
-struct rxrpc_call *rxrpc_accept_call(struct rxrpc_sock *rx,
-				     unsigned long user_call_ID,
-				     rxrpc_notify_rx_t notify_rx)
-	__releases(&rx->sk.sk_lock.slock)
-	__acquires(call->user_mutex)
+int rxrpc_user_charge_accept(struct rxrpc_sock *rx, unsigned long user_call_ID)
 {
-	struct rxrpc_call *call;
-	struct rb_node *parent, **pp;
-	int ret;
-
-	_enter(",%lx", user_call_ID);
-
-	ASSERT(!irqs_disabled());
-
-	write_lock(&rx->call_lock);
-
-	if (list_empty(&rx->to_be_accepted)) {
-		write_unlock(&rx->call_lock);
-		release_sock(&rx->sk);
-		kleave(" = -ENODATA [empty]");
-		return ERR_PTR(-ENODATA);
-	}
-
-	/* check the user ID isn't already in use */
-	pp = &rx->calls.rb_node;
-	parent = NULL;
-	while (*pp) {
-		parent = *pp;
-		call = rb_entry(parent, struct rxrpc_call, sock_node);
-
-		if (user_call_ID < call->user_call_ID)
-			pp = &(*pp)->rb_left;
-		else if (user_call_ID > call->user_call_ID)
-			pp = &(*pp)->rb_right;
-		else
-			goto id_in_use;
-	}
-
-	/* Dequeue the first call and check it's still valid.  We gain
-	 * responsibility for the queue's reference.
-	 */
-	call = list_entry(rx->to_be_accepted.next,
-			  struct rxrpc_call, accept_link);
-	write_unlock(&rx->call_lock);
-
-	/* We need to gain the mutex from the interrupt handler without
-	 * upsetting lockdep, so we have to release it there and take it here.
-	 * We are, however, still holding the socket lock, so other accepts
-	 * must wait for us and no one can add the user ID behind our backs.
-	 */
-	if (mutex_lock_interruptible(&call->user_mutex) < 0) {
-		release_sock(&rx->sk);
-		kleave(" = -ERESTARTSYS");
-		return ERR_PTR(-ERESTARTSYS);
-	}
-
-	write_lock(&rx->call_lock);
-	list_del_init(&call->accept_link);
-	sk_acceptq_removed(&rx->sk);
-	rxrpc_see_call(call);
-
-	/* Find the user ID insertion point. */
-	pp = &rx->calls.rb_node;
-	parent = NULL;
-	while (*pp) {
-		parent = *pp;
-		call = rb_entry(parent, struct rxrpc_call, sock_node);
-
-		if (user_call_ID < call->user_call_ID)
-			pp = &(*pp)->rb_left;
-		else if (user_call_ID > call->user_call_ID)
-			pp = &(*pp)->rb_right;
-		else
-			BUG();
-	}
-
-	write_lock_bh(&call->state_lock);
-	switch (call->state) {
-	case RXRPC_CALL_SERVER_ACCEPTING:
-		call->state = RXRPC_CALL_SERVER_RECV_REQUEST;
-		break;
-	case RXRPC_CALL_COMPLETE:
-		ret = call->error;
-		goto out_release;
-	default:
-		BUG();
-	}
-
-	/* formalise the acceptance */
-	call->notify_rx = notify_rx;
-	call->user_call_ID = user_call_ID;
-	rxrpc_get_call(call, rxrpc_call_got_userid);
-	rb_link_node(&call->sock_node, parent, pp);
-	rb_insert_color(&call->sock_node, &rx->calls);
-	if (test_and_set_bit(RXRPC_CALL_HAS_USERID, &call->flags))
-		BUG();
-
-	write_unlock_bh(&call->state_lock);
-	write_unlock(&rx->call_lock);
-	rxrpc_notify_socket(call);
-	rxrpc_service_prealloc(rx, GFP_KERNEL);
-	release_sock(&rx->sk);
-	_leave(" = %p{%d}", call, call->debug_id);
-	return call;
-
-out_release:
-	_debug("release %p", call);
-	write_unlock_bh(&call->state_lock);
-	write_unlock(&rx->call_lock);
-	rxrpc_release_call(rx, call);
-	rxrpc_put_call(call, rxrpc_call_put);
-	goto out;
-
-id_in_use:
-	ret = -EBADSLT;
-	write_unlock(&rx->call_lock);
-out:
-	rxrpc_service_prealloc(rx, GFP_KERNEL);
-	release_sock(&rx->sk);
-	_leave(" = %d", ret);
-	return ERR_PTR(ret);
-}
-
-/*
- * Handle rejection of a call by userspace
- * - reject the call at the front of the queue
- */
-int rxrpc_reject_call(struct rxrpc_sock *rx)
-{
-	struct rxrpc_call *call;
-	bool abort = false;
-	int ret;
-
-	_enter("");
-
-	ASSERT(!irqs_disabled());
-
-	write_lock(&rx->call_lock);
-
-	if (list_empty(&rx->to_be_accepted)) {
-		write_unlock(&rx->call_lock);
-		return -ENODATA;
-	}
-
-	/* Dequeue the first call and check it's still valid.  We gain
-	 * responsibility for the queue's reference.
-	 */
-	call = list_entry(rx->to_be_accepted.next,
-			  struct rxrpc_call, accept_link);
-	list_del_init(&call->accept_link);
-	sk_acceptq_removed(&rx->sk);
-	rxrpc_see_call(call);
+	struct rxrpc_backlog *b = rx->backlog;
 
-	write_lock_bh(&call->state_lock);
-	switch (call->state) {
-	case RXRPC_CALL_SERVER_ACCEPTING:
-		__rxrpc_abort_call("REJ", call, 1, RX_USER_ABORT, -ECONNABORTED);
-		abort = true;
-		fallthrough;
-	case RXRPC_CALL_COMPLETE:
-		ret = call->error;
-		goto out_discard;
-	default:
-		BUG();
-	}
+	if (rx->sk.sk_state == RXRPC_CLOSE)
+		return -ESHUTDOWN;
 
-out_discard:
-	write_unlock_bh(&call->state_lock);
-	write_unlock(&rx->call_lock);
-	if (abort) {
-		rxrpc_send_abort_packet(call);
-		rxrpc_release_call(rx, call);
-		rxrpc_put_call(call, rxrpc_call_put);
-	}
-	rxrpc_service_prealloc(rx, GFP_KERNEL);
-	_leave(" = %d", ret);
-	return ret;
+	return rxrpc_service_prealloc_one(rx, b, NULL, NULL, user_call_ID,
+					  GFP_KERNEL,
+					  atomic_inc_return(&rxrpc_debug_id));
 }
 
 /*
diff --git a/net/rxrpc/call_object.c b/net/rxrpc/call_object.c
index a40fae013942..ed49769b459d 100644
--- a/net/rxrpc/call_object.c
+++ b/net/rxrpc/call_object.c
@@ -23,7 +23,6 @@ const char *const rxrpc_call_states[NR__RXRPC_CALL_STATES] = {
 	[RXRPC_CALL_CLIENT_RECV_REPLY]		= "ClRcvRpl",
 	[RXRPC_CALL_SERVER_PREALLOC]		= "SvPrealc",
 	[RXRPC_CALL_SERVER_SECURING]		= "SvSecure",
-	[RXRPC_CALL_SERVER_ACCEPTING]		= "SvAccept",
 	[RXRPC_CALL_SERVER_RECV_REQUEST]	= "SvRcvReq",
 	[RXRPC_CALL_SERVER_ACK_REQUEST]		= "SvAckReq",
 	[RXRPC_CALL_SERVER_SEND_REPLY]		= "SvSndRpl",
@@ -352,9 +351,7 @@ void rxrpc_incoming_call(struct rxrpc_sock *rx,
 	call->call_id		= sp->hdr.callNumber;
 	call->service_id	= sp->hdr.serviceId;
 	call->cid		= sp->hdr.cid;
-	call->state		= RXRPC_CALL_SERVER_ACCEPTING;
-	if (sp->hdr.securityIndex > 0)
-		call->state	= RXRPC_CALL_SERVER_SECURING;
+	call->state		= RXRPC_CALL_SERVER_SECURING;
 	call->cong_tstamp	= skb->tstamp;
 
 	/* Set the channel for this call.  We don't get channel_lock as we're
diff --git a/net/rxrpc/conn_event.c b/net/rxrpc/conn_event.c
index 6e972b4823ef..64ace2960ecc 100644
--- a/net/rxrpc/conn_event.c
+++ b/net/rxrpc/conn_event.c
@@ -269,7 +269,7 @@ static void rxrpc_call_is_secure(struct rxrpc_call *call)
 	if (call) {
 		write_lock_bh(&call->state_lock);
 		if (call->state == RXRPC_CALL_SERVER_SECURING) {
-			call->state = RXRPC_CALL_SERVER_ACCEPTING;
+			call->state = RXRPC_CALL_SERVER_RECV_REQUEST;
 			rxrpc_notify_socket(call);
 		}
 		write_unlock_bh(&call->state_lock);
diff --git a/net/rxrpc/recvmsg.c b/net/rxrpc/recvmsg.c
index c4684dde1f16..2c842851d72e 100644
--- a/net/rxrpc/recvmsg.c
+++ b/net/rxrpc/recvmsg.c
@@ -178,37 +178,6 @@ static int rxrpc_recvmsg_term(struct rxrpc_call *call, struct msghdr *msg)
 	return ret;
 }
 
-/*
- * Pass back notification of a new call.  The call is added to the
- * to-be-accepted list.  This means that the next call to be accepted might not
- * be the last call seen awaiting acceptance, but unless we leave this on the
- * front of the queue and block all other messages until someone gives us a
- * user_ID for it, there's not a lot we can do.
- */
-static int rxrpc_recvmsg_new_call(struct rxrpc_sock *rx,
-				  struct rxrpc_call *call,
-				  struct msghdr *msg, int flags)
-{
-	int tmp = 0, ret;
-
-	ret = put_cmsg(msg, SOL_RXRPC, RXRPC_NEW_CALL, 0, &tmp);
-
-	if (ret == 0 && !(flags & MSG_PEEK)) {
-		_debug("to be accepted");
-		write_lock_bh(&rx->recvmsg_lock);
-		list_del_init(&call->recvmsg_link);
-		write_unlock_bh(&rx->recvmsg_lock);
-
-		rxrpc_get_call(call, rxrpc_call_got);
-		write_lock(&rx->call_lock);
-		list_add_tail(&call->accept_link, &rx->to_be_accepted);
-		write_unlock(&rx->call_lock);
-	}
-
-	trace_rxrpc_recvmsg(call, rxrpc_recvmsg_to_be_accepted, 1, 0, 0, ret);
-	return ret;
-}
-
 /*
  * End the packet reception phase.
  */
@@ -630,9 +599,6 @@ int rxrpc_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
 	}
 
 	switch (READ_ONCE(call->state)) {
-	case RXRPC_CALL_SERVER_ACCEPTING:
-		ret = rxrpc_recvmsg_new_call(rx, call, msg, flags);
-		break;
 	case RXRPC_CALL_CLIENT_RECV_REPLY:
 	case RXRPC_CALL_SERVER_RECV_REQUEST:
 	case RXRPC_CALL_SERVER_ACK_REQUEST:
@@ -728,7 +694,7 @@ int rxrpc_kernel_recv_data(struct socket *sock, struct rxrpc_call *call,
 	       call->debug_id, rxrpc_call_states[call->state],
 	       iov_iter_count(iter), want_more);
 
-	ASSERTCMP(call->state, !=, RXRPC_CALL_SERVER_ACCEPTING);
+	ASSERTCMP(call->state, !=, RXRPC_CALL_SERVER_SECURING);
 
 	mutex_lock(&call->user_mutex);
 
diff --git a/net/rxrpc/sendmsg.c b/net/rxrpc/sendmsg.c
index 0824e103d037..d27140c836cc 100644
--- a/net/rxrpc/sendmsg.c
+++ b/net/rxrpc/sendmsg.c
@@ -530,10 +530,10 @@ static int rxrpc_sendmsg_cmsg(struct msghdr *msg, struct rxrpc_send_params *p)
 				return -EINVAL;
 			break;
 
-		case RXRPC_ACCEPT:
+		case RXRPC_CHARGE_ACCEPT:
 			if (p->command != RXRPC_CMD_SEND_DATA)
 				return -EINVAL;
-			p->command = RXRPC_CMD_ACCEPT;
+			p->command = RXRPC_CMD_CHARGE_ACCEPT;
 			if (len != 0)
 				return -EINVAL;
 			break;
@@ -659,16 +659,12 @@ int rxrpc_do_sendmsg(struct rxrpc_sock *rx, struct msghdr *msg, size_t len)
 	if (ret < 0)
 		goto error_release_sock;
 
-	if (p.command == RXRPC_CMD_ACCEPT) {
+	if (p.command == RXRPC_CMD_CHARGE_ACCEPT) {
 		ret = -EINVAL;
 		if (rx->sk.sk_state != RXRPC_SERVER_LISTENING)
 			goto error_release_sock;
-		call = rxrpc_accept_call(rx, p.call.user_call_ID, NULL);
-		/* The socket is now unlocked. */
-		if (IS_ERR(call))
-			return PTR_ERR(call);
-		ret = 0;
-		goto out_put_unlock;
+		ret = rxrpc_user_charge_accept(rx, p.call.user_call_ID);
+		goto error_release_sock;
 	}
 
 	call = rxrpc_find_call_by_user_ID(rx, p.call.user_call_ID);
@@ -690,7 +686,6 @@ int rxrpc_do_sendmsg(struct rxrpc_sock *rx, struct msghdr *msg, size_t len)
 		case RXRPC_CALL_CLIENT_AWAIT_CONN:
 		case RXRPC_CALL_SERVER_PREALLOC:
 		case RXRPC_CALL_SERVER_SECURING:
-		case RXRPC_CALL_SERVER_ACCEPTING:
 			rxrpc_put_call(call, rxrpc_call_put);
 			ret = -EBUSY;
 			goto error_release_sock;

[PATCH net 5/6] rxrpc: The server keyring isn't network-namespaced

[David Howells]

The keyring containing the server's tokens isn't network-namespaced, so it
shouldn't be looked up with a network namespace.  It is expected to be
owned specifically by the server, so namespacing is unnecessary.

Fixes: a58946c158a0 ("keys: Pass the network namespace into request_key mechanism")
Signed-off-by: David Howells <dhowells@redhat.com>
---

 net/rxrpc/key.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/rxrpc/key.c b/net/rxrpc/key.c
index c668e4b7dbff..75e84ed4fa63 100644
--- a/net/rxrpc/key.c
+++ b/net/rxrpc/key.c
@@ -940,7 +940,7 @@ int rxrpc_server_keyring(struct rxrpc_sock *rx, sockptr_t optval, int optlen)
 	if (IS_ERR(description))
 		return PTR_ERR(description);
 
-	key = request_key_net(&key_type_keyring, description, sock_net(&rx->sk), NULL);
+	key = request_key(&key_type_keyring, description, NULL);
 	if (IS_ERR(key)) {
 		kfree(description);
 		_leave(" = %ld", PTR_ERR(key));

[PATCH net 6/6] rxrpc: Fix server keyring leak

[David Howells]

If someone calls setsockopt() twice to set a server key keyring, the first
keyring is leaked.

Fix it to return an error instead if the server key keyring is already set.

Fixes: 17926a79320a ("[AF_RXRPC]: Provide secure RxRPC sockets for use by userspace and kernel both")
Signed-off-by: David Howells <dhowells@redhat.com>
---

 net/rxrpc/key.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/rxrpc/key.c b/net/rxrpc/key.c
index 75e84ed4fa63..2e8bd3b97301 100644
--- a/net/rxrpc/key.c
+++ b/net/rxrpc/key.c
@@ -903,7 +903,7 @@ int rxrpc_request_key(struct rxrpc_sock *rx, sockptr_t optval, int optlen)
 
 	_enter("");
 
-	if (optlen <= 0 || optlen > PAGE_SIZE - 1)
+	if (optlen <= 0 || optlen > PAGE_SIZE - 1 || rx->securities)
 		return -EINVAL;
 
 	description = memdup_sockptr_nul(optval, optlen);

Re: [PATCH net 0/6] rxrpc: Miscellaneous fixes

[David Miller]

From: David Howells <dhowells@redhat.com>
Date: Mon, 05 Oct 2020 17:18:44 +0100

> The patches are tagged here:
> 
> 	git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git
> 	rxrpc-fixes-20201005.txt

I think the ".txt" at the end of the branch name is a mistake.

It's really hard to send a clean working pull request isn't it? :-)

Pulled, thanks David.

Re: [PATCH net 0/6] rxrpc: Miscellaneous fixes

[David Howells]

David Miller <davem@davemloft.net> wrote:

> I think the ".txt" at the end of the branch name is a mistake.

Sigh.  That's the name of the file with the cover message in it (named for the
tag).  I need to adjust my script yet more to check not only that the tag name
is in there, but that it also doesn't have any bits on the end.

David