ceph: drop src address(es) from message header [new protocol feature]

The CEPH_FEATURE_NOSRCADDR protocol feature avoids putting the full source
address in each message header (twice).  This patch switches the client to
the new scheme, and _requires_ this feature on the server.  The server
will support both the old and new schemes.  That means an old client will
work with a new server, but a new client will not work with an old server.

Signed-off-by: Sage Weil <sage@newdream.net>

3 files changed, 36 insertions, 11 deletions

diff --git a/fs/ceph/ceph_fs.h b/fs/ceph/ceph_fs.h
index 0c2241e..8b194c6 100644
--- a/fs/ceph/ceph_fs.h
+++ b/fs/ceph/ceph_fs.h
@@ -53,8 +53,17 @@
 /*
  * feature bits
  */
-#define CEPH_FEATURE_SUPPORTED  0
-#define CEPH_FEATURE_REQUIRED   0
+#define CEPH_FEATURE_UID        1
+#define CEPH_FEATURE_NOSRCADDR  2
+
+#define CEPH_FEATURE_SUPPORTED_MON  CEPH_FEATURE_UID|CEPH_FEATURE_NOSRCADDR
+#define CEPH_FEATURE_REQUIRED_MON   CEPH_FEATURE_UID
+#define CEPH_FEATURE_SUPPORTED_MDS  CEPH_FEATURE_UID|CEPH_FEATURE_NOSRCADDR
+#define CEPH_FEATURE_REQUIRED_MDS   CEPH_FEATURE_UID
+#define CEPH_FEATURE_SUPPORTED_OSD  CEPH_FEATURE_UID|CEPH_FEATURE_NOSRCADDR
+#define CEPH_FEATURE_REQUIRED_OSD   CEPH_FEATURE_UID
+#define CEPH_FEATURE_SUPPORTED_CLIENT CEPH_FEATURE_NOSRCADDR
+#define CEPH_FEATURE_REQUIRED_CLIENT CEPH_FEATURE_NOSRCADDR
 
 
 /*
diff --git a/fs/ceph/messenger.c b/fs/ceph/messenger.c
index bb16edb..fe75ec7 100644
--- a/fs/ceph/messenger.c
+++ b/fs/ceph/messenger.c
@@ -663,7 +663,7 @@ static void prepare_write_connect(struct ceph_messenger *msgr,
 	dout("prepare_write_connect %p cseq=%d gseq=%d proto=%d\n", con,
 	     con->connect_seq, global_seq, proto);
 
-	con->out_connect.features = CEPH_FEATURE_SUPPORTED;
+	con->out_connect.features = CEPH_FEATURE_SUPPORTED_CLIENT;
 	con->out_connect.host_type = cpu_to_le32(CEPH_ENTITY_TYPE_CLIENT);
 	con->out_connect.connect_seq = cpu_to_le32(con->connect_seq);
 	con->out_connect.global_seq = cpu_to_le32(global_seq);
@@ -1126,8 +1126,8 @@ static void fail_protocol(struct ceph_connection *con)
 
 static int process_connect(struct ceph_connection *con)
 {
-	u64 sup_feat = CEPH_FEATURE_SUPPORTED;
-	u64 req_feat = CEPH_FEATURE_REQUIRED;
+	u64 sup_feat = CEPH_FEATURE_SUPPORTED_CLIENT;
+	u64 req_feat = CEPH_FEATURE_REQUIRED_CLIENT;
 	u64 server_feat = le64_to_cpu(con->in_reply.features);
 
 	dout("process_connect on %p tag %d\n", con, (int)con->in_tag);
@@ -1514,14 +1514,14 @@ static void process_message(struct ceph_connection *con)
 
 	/* if first message, set peer_name */
 	if (con->peer_name.type == 0)
-		con->peer_name = msg->hdr.src.name;
+		con->peer_name = msg->hdr.src;
 
 	con->in_seq++;
 	mutex_unlock(&con->mutex);
 
 	dout("===== %p %llu from %s%lld %d=%s len %d+%d (%u %u %u) =====\n",
 	     msg, le64_to_cpu(msg->hdr.seq),
-	     ENTITY_NAME(msg->hdr.src.name),
+	     ENTITY_NAME(msg->hdr.src),
 	     le16_to_cpu(msg->hdr.type),
 	     ceph_msg_type_name(le16_to_cpu(msg->hdr.type)),
 	     le32_to_cpu(msg->hdr.front_len),
@@ -1987,9 +1987,7 @@ void ceph_con_send(struct ceph_connection *con, struct ceph_msg *msg)
 	}
 
 	/* set src+dst */
-	msg->hdr.src.name = con->msgr->inst.name;
-	msg->hdr.src.addr = con->msgr->my_enc_addr;
-	msg->hdr.orig_src = msg->hdr.src;
+	msg->hdr.src = con->msgr->inst.name;
 
 	BUG_ON(msg->front.iov_len != le32_to_cpu(msg->hdr.front_len));
 
diff --git a/fs/ceph/msgr.h b/fs/ceph/msgr.h
index 8aaab41..6baa8e4f 100644
--- a/fs/ceph/msgr.h
+++ b/fs/ceph/msgr.h
@@ -120,7 +120,7 @@ struct ceph_msg_connect_reply {
 /*
  * message header
  */
-struct ceph_msg_header {
+struct ceph_msg_header_old {
 	__le64 seq;       /* message seq# for this session */
 	__le64 tid;       /* transaction id */
 	__le16 type;      /* message type */
@@ -138,6 +138,24 @@ struct ceph_msg_header {
 	__le32 crc;       /* header crc32c */
 } __attribute__ ((packed));
 
+struct ceph_msg_header {
+	__le64 seq;       /* message seq# for this session */
+	__le64 tid;       /* transaction id */
+	__le16 type;      /* message type */
+	__le16 priority;  /* priority.  higher value == higher priority */
+	__le16 version;   /* version of message encoding */
+
+	__le32 front_len; /* bytes in main payload */
+	__le32 middle_len;/* bytes in middle payload */
+	__le32 data_len;  /* bytes of data payload */
+	__le16 data_off;  /* sender: include full offset;
+			     receiver: mask against ~PAGE_MASK */
+
+	struct ceph_entity_name src;
+	__le32 reserved;
+	__le32 crc;       /* header crc32c */
+} __attribute__ ((packed));
+
 #define CEPH_MSG_PRIO_LOW     64
 #define CEPH_MSG_PRIO_DEFAULT 127
 #define CEPH_MSG_PRIO_HIGH    196