Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1571265 > unrolled thread

[PATCH 1/2] blk-mq: allocate blk_mq_tags and requests in correct node

Back to article view | Back to linux.kernel

Contents

  [PATCH 1/2] blk-mq: allocate blk_mq_tags and requests in correct node Shaohua Li <shli@fb.com> - 2017-02-01 05:30 +0100
    Re: [PATCH 1/2] blk-mq: allocate blk_mq_tags and requests in         correct node Christoph Hellwig <hch@lst.de> - 2017-02-01 10:40 +0100

#1571265 — [PATCH 1/2] blk-mq: allocate blk_mq_tags and requests in correct node

blk_mq_tags/requests of specific hardware queue are mostly used in
specific cpus, which might not be in the same numa node as disk. For
example, a nvme card is in node 0. half hardware queue will be used by
node 0, the other node 1.

Signed-off-by: Shaohua Li <shli@fb.com>
---
 block/blk-mq.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/block/blk-mq.c b/block/blk-mq.c
index c3400b5..90a37ba 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1594,9 +1594,13 @@ static struct blk_mq_tags *blk_mq_init_rq_map(struct blk_mq_tag_set *set,
 	struct blk_mq_tags *tags;
 	unsigned int i, j, entries_per_page, max_order = 4;
 	size_t rq_size, left;
+	int node;
+
+	node = blk_mq_hw_queue_to_node(set->mq_map, hctx_idx);
+	if (node == NUMA_NO_NODE)
+		node = set->numa_node;
 
-	tags = blk_mq_init_tags(set->queue_depth, set->reserved_tags,
-				set->numa_node,
+	tags = blk_mq_init_tags(set->queue_depth, set->reserved_tags, node,
 				BLK_MQ_FLAG_TO_ALLOC_POLICY(set->flags));
 	if (!tags)
 		return NULL;
@@ -1605,7 +1609,7 @@ static struct blk_mq_tags *blk_mq_init_rq_map(struct blk_mq_tag_set *set,
 
 	tags->rqs = kzalloc_node(set->queue_depth * sizeof(struct request *),
 				 GFP_NOIO | __GFP_NOWARN | __GFP_NORETRY,
-				 set->numa_node);
+				 node);
 	if (!tags->rqs) {
 		blk_mq_free_tags(tags);
 		return NULL;
@@ -1629,7 +1633,7 @@ static struct blk_mq_tags *blk_mq_init_rq_map(struct blk_mq_tag_set *set,
 			this_order--;
 
 		do {
-			page = alloc_pages_node(set->numa_node,
+			page = alloc_pages_node(node,
 				GFP_NOIO | __GFP_NOWARN | __GFP_NORETRY | __GFP_ZERO,
 				this_order);
 			if (page)
@@ -1660,7 +1664,7 @@ static struct blk_mq_tags *blk_mq_init_rq_map(struct blk_mq_tag_set *set,
 			if (set->ops->init_request) {
 				if (set->ops->init_request(set->driver_data,
 						tags->rqs[i], hctx_idx, i,
-						set->numa_node)) {
+						node)) {
 					tags->rqs[i] = NULL;
 					goto fail;
 				}
-- 
2.9.3

[toc] | [next] | [standalone]

#1571365 — Re: [PATCH 1/2] blk-mq: allocate blk_mq_tags and requests in correct node

Hi Shaohua,

the code your patching has changed a lot in Jens' tree, so I think
you'll have to respin it.  But the idea looks fine.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web