Reenabled linux-xen and xen-image build

[people/pmueller/ipfire-2.x.git] / src / patches / suse-2.6.27.25 / patches.arch / s390-11-14-zfcp_unit_remove.patch

From: Gerald Schaefer <geraldsc@de.ibm.com>
Subject: zfcp: remove unit will fail if add unit is not finished
References: bnc#484767,LTC#48795

Symptom:     On some hardware it can take some time to add a unit. If
             some remove this unit during this process the remove will
             fail.
Problem:     There exists some SCSI work which needs to be finished
             before removing this unit.
Solution:    Wait until all SCSI work is done for that unit.

Acked-by: John Jolly <jjolly@suse.de>
---
 drivers/s390/scsi/zfcp_sysfs.c |   21 +++++++++++++++------
 1 file changed, 15 insertions(+), 6 deletions(-)

Index: linux-sles11/drivers/s390/scsi/zfcp_sysfs.c
===================================================================
--- linux-sles11.orig/drivers/s390/scsi/zfcp_sysfs.c
+++ linux-sles11/drivers/s390/scsi/zfcp_sysfs.c
@@ -253,12 +253,21 @@ static ssize_t zfcp_sysfs_unit_remove_st
 
 	write_lock_irq(&zfcp_data.config_lock);
 	unit = zfcp_get_unit_by_lun(port, fcp_lun);
-	if (unit && (atomic_read(&unit->refcount) == 0)) {
-		zfcp_unit_get(unit);
-		atomic_set_mask(ZFCP_STATUS_COMMON_REMOVE, &unit->status);
-		list_move(&unit->list, &unit_remove_lh);
-	} else
-		unit = NULL;
+	if (unit) {
+		write_unlock_irq(&zfcp_data.config_lock);
+		/* wait for possible timeout during SCSI probe */
+		flush_work(&unit->scsi_work);
+		write_lock_irq(&zfcp_data.config_lock);
+
+		if (atomic_read(&unit->refcount) == 0) {
+			zfcp_unit_get(unit);
+			atomic_set_mask(ZFCP_STATUS_COMMON_REMOVE,
+					&unit->status);
+			list_move(&unit->list, &unit_remove_lh);
+		} else {
+			unit = NULL;
+		}
+	}
 
 	write_unlock_irq(&zfcp_data.config_lock);