[SCSI] iscsi: fix writepsace race
We can race and misset the suspend bit if iscsi_write_space is
called then iscsi_send returns with a failure indicating
there is no space.
To handle this this patch returns a error upwards allowing xmitworker
to decide if we need to try and transmit again. For the no
write space case xmitworker will not retry, and instead
let iscsi_write_space queue it back up if needed (this relies
on the work queue code to properly requeue us if needed).
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
diff --git a/include/scsi/scsi_transport_iscsi.h b/include/scsi/scsi_transport_iscsi.h
index 92129b9..b684426 100644
--- a/include/scsi/scsi_transport_iscsi.h
+++ b/include/scsi/scsi_transport_iscsi.h
@@ -57,8 +57,12 @@
* Called from queuecommand with session lock held.
* @init_mgmt_task: Initialize a iscsi_mgmt_task and any internal structs.
* Called from iscsi_conn_send_generic with xmitmutex.
- * @xmit_cmd_task: requests LLD to transfer cmd task
- * @xmit_mgmt_task: requests LLD to transfer mgmt task
+ * @xmit_cmd_task: Requests LLD to transfer cmd task. Returns 0 or the
+ * the number of bytes transferred on success, and -Exyz
+ * value on error.
+ * @xmit_mgmt_task: Requests LLD to transfer mgmt task. Returns 0 or the
+ * the number of bytes transferred on success, and -Exyz
+ * value on error.
* @cleanup_cmd_task: requests LLD to fail cmd task. Called with xmitmutex
* and session->lock after the connection has been
* suspended and terminated during recovery. If called