[PATCH 3/4] blk-mq: deal with shared queue mapping reliably

[Ming Lei <ming.lei@redhat.com>]

This patch sets map->nr_queues as zero explictly if there is zero
queues for such queue type, then blk_mq_map_swqueue() can become
more robust to deal with shared mappings.

Cc: Jeff Moyer <jmoyer@redhat.com>
Cc: Mike Snitzer <snitzer@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
---
 block/blk-mq.c          |  3 +++
 drivers/nvme/host/pci.c | 37 ++++++++++++++++++++-----------------
 2 files changed, 23 insertions(+), 17 deletions(-)

diff --git a/block/blk-mq.c b/block/blk-mq.c
index a4a0895dae65..a737d912c46b 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -2435,6 +2435,9 @@ static void blk_mq_map_swqueue(struct request_queue *q)
 		for (j = 0; j < set->nr_maps; j++) {
 			hctx = blk_mq_map_queue_type(q, j, i);
 
+			if (!set->map[j].nr_queues)
+					continue;
+
 			/*
 			 * If the CPU is already set in the mask, then we've
 			 * mapped this one already. This can happen if
diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
index 95bd68be2078..43074c54279e 100644
--- a/drivers/nvme/host/pci.c
+++ b/drivers/nvme/host/pci.c
@@ -492,29 +492,32 @@ static int nvme_pci_map_queues(struct blk_mq_tag_set *set)
 	offset = queue_irq_offset(dev);
 	for (i = 0, qoff = 0; i < set->nr_maps; i++) {
 		struct blk_mq_queue_map *map = &set->map[i];
+		unsigned nr_queues;
 
-		map->nr_queues = dev->io_queues[i];
-		if (!map->nr_queues) {
+		nr_queues = map->nr_queues = dev->io_queues[i];
+		if (!nr_queues) {
 			BUG_ON(i == HCTX_TYPE_DEFAULT);
 
-			/* shared set, resuse read set parameters */
-			map->nr_queues = dev->io_queues[HCTX_TYPE_DEFAULT];
+			/* shared set, resuse default set parameters and table */
+			nr_queues = dev->io_queues[HCTX_TYPE_DEFAULT];
 			qoff = 0;
 			offset = queue_irq_offset(dev);
-		}
 
-		/*
-		 * The poll queue(s) doesn't have an IRQ (and hence IRQ
-		 * affinity), so use the regular blk-mq cpu mapping if
-		 * poll queue(s) don't share mapping with TYPE_DEFAULT.
-		 */
-		map->queue_offset = qoff;
-		if (i != HCTX_TYPE_POLL || !qoff)
-			blk_mq_pci_map_queues(map, to_pci_dev(dev->dev), offset);
-		else
-			blk_mq_map_queues(map);
-		qoff += map->nr_queues;
-		offset += map->nr_queues;
+			memcpy(map->mq_map, set->map[HCTX_TYPE_DEFAULT].mq_map,
+				   nr_cpu_ids * sizeof(map->mq_map[0]));
+		} else {
+			/*
+			 * The poll queue(s) doesn't have an IRQ (and hence IRQ
+			 * affinity), so use the regular blk-mq cpu mapping.
+			 */
+			map->queue_offset = qoff;
+			if (i != HCTX_TYPE_POLL)
+				blk_mq_pci_map_queues(map, to_pci_dev(dev->dev), offset);
+			else
+				blk_mq_map_queues(map);
+		}
+		qoff += nr_queues;
+		offset += nr_queues;
 	}
 
 	return 0;
-- 
2.9.5