From: Jeff Skirvin <jeffrey.d.skirvin@intel.com>
Date: Wed, 16 Mar 2011 16:41:59 +0000 (-0700)
Subject: isci: don't hold scic_lock over calls to sas_task_abort()
X-Git-Tag: v3.0-rc6~1^2~1^2~206
X-Git-Url: https://openfabrics.org/gitweb/?a=commitdiff_plain;h=c4b9e24c4be67aeed44cd46ef5ea92948d02a426;p=~emulex%2Finfiniband.git

isci: don't hold scic_lock over calls to sas_task_abort()

In the case where submitted I/Os fail with the status code
SCI_FAILURE_REMOTE_DEVICE_RESET_REQUIRED, the execute function now waits
until scic_lock is cleared before calling the helper function
"isci_request_signal_device_reset" which sets the flag for the pending
reset condition on the I/O.

Signed-off-by: Jeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---

diff --git a/drivers/scsi/isci/request.c b/drivers/scsi/isci/request.c
index 01564310431..eba8e0b3c87 100644
--- a/drivers/scsi/isci/request.c
+++ b/drivers/scsi/isci/request.c
@@ -449,26 +449,11 @@ int isci_request_execute(
 			list_add(&request->dev_node,
 				 &isci_device->reqs_in_process);
 
-			if (status ==
-				SCI_FAILURE_REMOTE_DEVICE_RESET_REQUIRED) {
-				/* Signal libsas that we need the SCSI error
-				 * handler thread to work on this I/O and that
-				 * we want a device reset.
-				 */
-				isci_request_signal_device_reset(request);
-
-				/* Change the status, since we are holding
-				 * the I/O until it is managed by the SCSI
-				 * error handler.
-				 */
-				status = SCI_SUCCESS;
-			}
-			else
+			if (status == SCI_SUCCESS) {
 				/* Save the tag for possible task mgmt later. */
 				request->io_tag = scic_io_request_get_io_tag(
 						     request->sci_request_handle);
-
-
+			}
 		} else
 			dev_warn(&isci_host->pdev->dev,
 				 "%s: failed request start\n",
@@ -476,6 +461,21 @@ int isci_request_execute(
 
 		spin_unlock_irqrestore(&isci_host->scic_lock, flags);
 
+		if (status ==
+		    SCI_FAILURE_REMOTE_DEVICE_RESET_REQUIRED) {
+			/* Signal libsas that we need the SCSI error
+			* handler thread to work on this I/O and that
+			* we want a device reset.
+			*/
+			isci_request_signal_device_reset(request);
+
+			/* Change the status, since we are holding
+			* the I/O until it is managed by the SCSI
+			* error handler.
+			*/
+			status = SCI_SUCCESS;
+		}
+
 	} else
 		dev_warn(&isci_host->pdev->dev,
 			 "%s: request_construct failed - status = 0x%x\n",