[PATCH 55/91] Make scsi_free_queue() kill pending SCSI commands

[Willy Tarreau]

2.6.27-longterm review patch.  If anyone has any objections, please let us know.

------------------

commit 3308511c93e6ad0d3c58984ecd6e5e57f96b12c8 upstream.

Make sure that SCSI device removal via scsi_remove_host() does finish
all pending SCSI commands. Currently that's not the case and hence
removal of a SCSI host during I/O can cause a deadlock. See also
"blkdev_issue_discard() hangs forever if underlying storage device is
removed" (http://bugzilla.kernel.org/show_bug.cgi?id=40472). See also
http://lkml.org/lkml/2011/8/27/6.

Signed-off-by: Bart Van Assche <bvanassche@xxxxxxx>
Signed-off-by: James Bottomley <JBottomley@xxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxx>
---
 drivers/scsi/hosts.c    |    9 ++++++---
 drivers/scsi/scsi_lib.c |    9 +++++++++
 2 files changed, 15 insertions(+), 3 deletions(-)

Index: longterm-2.6.27/drivers/scsi/hosts.c
===================================================================
--- longterm-2.6.27.orig/drivers/scsi/hosts.c	2012-02-05 22:34:33.360914737 +0100
+++ longterm-2.6.27/drivers/scsi/hosts.c	2012-02-05 22:34:42.692915134 +0100
@@ -269,14 +269,17 @@
 {
 	struct Scsi_Host *shost = dev_to_shost(dev);
 	struct device *parent = dev->parent;
+	struct request_queue *q;
 
 	if (shost->ehandler)
 		kthread_stop(shost->ehandler);
 	if (shost->work_q)
 		destroy_workqueue(shost->work_q);
-	if (shost->uspace_req_q) {
-		kfree(shost->uspace_req_q->queuedata);
-		scsi_free_queue(shost->uspace_req_q);
+	q = shost->uspace_req_q;
+	if (q) {
+		kfree(q->queuedata);
+		q->queuedata = NULL;
+		scsi_free_queue(q);
 	}
 
 	scsi_destroy_command_freelist(shost);
Index: longterm-2.6.27/drivers/scsi/scsi_lib.c
===================================================================
--- longterm-2.6.27.orig/drivers/scsi/scsi_lib.c	2012-02-05 22:34:33.364914767 +0100
+++ longterm-2.6.27/drivers/scsi/scsi_lib.c	2012-02-05 22:34:42.697916475 +0100
@@ -1680,6 +1680,15 @@
 
 void scsi_free_queue(struct request_queue *q)
 {
+	unsigned long flags;
+
+	WARN_ON(q->queuedata);
+
+	/* cause scsi_request_fn() to kill all non-finished requests */
+	spin_lock_irqsave(q->queue_lock, flags);
+	q->request_fn(q);
+	spin_unlock_irqrestore(q->queue_lock, flags);
+
 	blk_cleanup_queue(q);
 }
 


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/