iscsi_tcp: propogate EAGAIN from sendpage to libiscsi

The net layer might return -EAGAIN because it could not
get space/mem within the sock sndtimeo or becuase the tcp/ip
connection was down. For the latter we do not want to retry
because the conn/session should just be shutdown and restarted.
libiscsi knows the state of the session recovery so propogate
this error to that layer. It will either do iscsi recovery
or have us retry the operation. Right now if we have partially
sent a pdu we would always retry the IO xmit slowing down
recovery.

Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>

1 files changed, 9 insertions, 5 deletions

diff --git a/drivers/scsi/iscsi_tcp.c b/drivers/scsi/iscsi_tcp.c
index b7c092d63bbe..518dbd91df85 100644
--- a/drivers/scsi/iscsi_tcp.c
+++ b/drivers/scsi/iscsi_tcp.c
@@ -253,8 +253,6 @@ static int iscsi_sw_tcp_xmit_segment(struct iscsi_tcp_conn *tcp_conn,
 
 		if (r < 0) {
 			iscsi_tcp_segment_unmap(segment);
-			if (copied || r == -EAGAIN)
-				break;
 			return r;
 		}
 		copied += r;
@@ -275,11 +273,17 @@ static int iscsi_sw_tcp_xmit(struct iscsi_conn *conn)
 
 	while (1) {
 		rc = iscsi_sw_tcp_xmit_segment(tcp_conn, segment);
-		if (rc < 0) {
+		/*
+		 * We may not have been able to send data because the conn
+		 * is getting stopped. libiscsi will know so propogate err
+		 * for it to do the right thing.
+		 */
+		if (rc == -EAGAIN)
+			return rc;
+		else if (rc < 0) {
 			rc = ISCSI_ERR_XMIT_FAILED;
 			goto error;
-		}
-		if (rc == 0)
+		} else if (rc == 0)
 			break;
 
 		consumed += rc;