summaryrefslogtreecommitdiff
path: root/Documentation/networking
diff options
context:
space:
mode:
authorDavid Howells <dhowells@redhat.com>2017-06-05 16:30:49 +0300
committerDavid Howells <dhowells@redhat.com>2017-06-05 16:30:49 +0300
commit4722974d90e06d0164ca1b73a6b34cec6bdb64ad (patch)
treef8f839ed9e2b63691c8122c6e94dc0d4fea9d98f /Documentation/networking
parent28036f44851e2515aa91b547b45cefddcac52ff6 (diff)
downloadlinux-4722974d90e06d0164ca1b73a6b34cec6bdb64ad.tar.xz
rxrpc: Implement service upgrade
Implement AuriStor's service upgrade facility. There are three problems that this is meant to deal with: (1) Various of the standard AFS RPC calls have IPv4 addresses in their requests and/or replies - but there's no room for including IPv6 addresses. (2) Definition of IPv6-specific RPC operations in the standard operation sets has not yet been achieved. (3) One could envision the creation a new service on the same port that as the original service. The new service could implement improved operations - and the client could try this first, falling back to the original service if it's not there. Unfortunately, certain servers ignore packets addressed to a service they don't implement and don't respond in any way - not even with an ABORT. This means that the client must then wait for the call timeout to occur. What service upgrade does is to see if the connection is marked as being 'upgradeable' and if so, change the service ID in the server and thus the request and reply formats. Note that the upgrade isn't mandatory - a server that supports only the original call set will ignore the upgrade request. In the protocol, the procedure is then as follows: (1) To request an upgrade, the first DATA packet in a new connection must have the userStatus set to 1 (this is normally 0). The userStatus value is normally ignored by the server. (2) If the server doesn't support upgrading, the reply packets will contain the same service ID as for the first request packet. (3) If the server does support upgrading, all future reply packets on that connection will contain the new service ID and the new service ID will be applied to *all* further calls on that connection as well. (4) The RPC op used to probe the upgrade must take the same request data as the shadow call in the upgrade set (but may return a different reply). GetCapability RPC ops were added to all standard sets for just this purpose. Ops where the request formats differ cannot be used for probing. (5) The client must wait for completion of the probe before sending any further RPC ops to the same destination. It should then use the service ID that recvmsg() reported back in all future calls. (6) The shadow service must have call definitions for all the operation IDs defined by the original service. To support service upgrading, a server should: (1) Call bind() twice on its AF_RXRPC socket before calling listen(). Each bind() should supply a different service ID, but the transport addresses must be the same. This allows the server to receive requests with either service ID. (2) Enable automatic upgrading by calling setsockopt(), specifying RXRPC_UPGRADEABLE_SERVICE and passing in a two-member array of unsigned shorts as the argument: unsigned short optval[2]; This specifies a pair of service IDs. They must be different and must match the service IDs bound to the socket. Member 0 is the service ID to upgrade from and member 1 is the service ID to upgrade to. Signed-off-by: David Howells <dhowells@redhat.com>
Diffstat (limited to 'Documentation/networking')
-rw-r--r--Documentation/networking/rxrpc.txt34
1 files changed, 26 insertions, 8 deletions
diff --git a/Documentation/networking/rxrpc.txt b/Documentation/networking/rxrpc.txt
index b7115ec55e04..2a1662760450 100644
--- a/Documentation/networking/rxrpc.txt
+++ b/Documentation/networking/rxrpc.txt
@@ -433,6 +433,13 @@ AF_RXRPC sockets support a few socket options at the SOL_RXRPC level:
Encrypted checksum plus entire packet padded and encrypted, including
actual packet length.
+ (*) RXRPC_UPGRADEABLE_SERVICE
+
+ This is used to indicate that a service socket with two bindings may
+ upgrade one bound service to the other if requested by the client. optval
+ must point to an array of two unsigned short ints. The first is the
+ service ID to upgrade from and the second the service ID to upgrade to.
+
========
SECURITY
@@ -588,7 +595,7 @@ A server would be set up to accept operations in the following manner:
The keyring can be manipulated after it has been given to the socket. This
permits the server to add more keys, replace keys, etc. whilst it is live.
- (2) A local address must then be bound:
+ (3) A local address must then be bound:
struct sockaddr_rxrpc srx = {
.srx_family = AF_RXRPC,
@@ -604,11 +611,22 @@ A server would be set up to accept operations in the following manner:
parameters are the same. The limit is currently two. To do this, bind()
should be called twice.
- (3) The server is then set to listen out for incoming calls:
+ (4) If service upgrading is required, first two service IDs must have been
+ bound and then the following option must be set:
+
+ unsigned short service_ids[2] = { from_ID, to_ID };
+ setsockopt(server, SOL_RXRPC, RXRPC_UPGRADEABLE_SERVICE,
+ service_ids, sizeof(service_ids));
+
+ This will automatically upgrade connections on service from_ID to service
+ to_ID if they request it. This will be reflected in msg_name obtained
+ through recvmsg() when the request data is delivered to userspace.
+
+ (5) The server is then set to listen out for incoming calls:
listen(server, 100);
- (4) The kernel notifies the server of pending incoming connections by sending
+ (6) The kernel notifies the server of pending incoming connections by sending
it a message for each. This is received with recvmsg() on the server
socket. It has no data, and has a single dataless control message
attached:
@@ -620,13 +638,13 @@ A server would be set up to accept operations in the following manner:
the time it is accepted - in which case the first call still on the queue
will be accepted.
- (5) The server then accepts the new call by issuing a sendmsg() with two
+ (7) The server then accepts the new call by issuing a sendmsg() with two
pieces of control data and no actual data:
RXRPC_ACCEPT - indicate connection acceptance
RXRPC_USER_CALL_ID - specify user ID for this call
- (6) The first request data packet will then be posted to the server socket for
+ (8) The first request data packet will then be posted to the server socket for
recvmsg() to pick up. At that point, the RxRPC address for the call can
be read from the address fields in the msghdr struct.
@@ -638,7 +656,7 @@ A server would be set up to accept operations in the following manner:
RXRPC_USER_CALL_ID - specifies the user ID for this call
- (8) The reply data should then be posted to the server socket using a series
+ (9) The reply data should then be posted to the server socket using a series
of sendmsg() calls, each with the following control messages attached:
RXRPC_USER_CALL_ID - specifies the user ID for this call
@@ -646,7 +664,7 @@ A server would be set up to accept operations in the following manner:
MSG_MORE should be set in msghdr::msg_flags on all but the last message
for a particular call.
- (9) The final ACK from the client will be posted for retrieval by recvmsg()
+(10) The final ACK from the client will be posted for retrieval by recvmsg()
when it is received. It will take the form of a dataless message with two
control messages attached:
@@ -656,7 +674,7 @@ A server would be set up to accept operations in the following manner:
MSG_EOR will be flagged to indicate that this is the final message for
this call.
-(10) Up to the point the final packet of reply data is sent, the call can be
+(11) Up to the point the final packet of reply data is sent, the call can be
aborted by calling sendmsg() with a dataless message with the following
control messages attached: