[PATCH 0/3] next round of blk-mq updates

[PATCH 0/3] next round of blk-mq updates

[Christoph Hellwig]

A couple more blkmq updates: a trivial cleanup, a consolidation of two very
similar functions into one, and support for partial request completions,
a feature needed by the work in progress scsi blk-mq support.

[PATCH 1/3] blk-mq: remove blk_mq_alloc_rq

[Christoph Hellwig]

[-- Attachment #1: 0001-blk-mq-remove-blk_mq_alloc_rq.patch --]
[-- Type: text/plain, Size: 1231 bytes --]

There's only one caller, which is a straight wrapper and fits the naming
scheme of the related functions a lot better.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-mq.c |   10 ++--------
 1 file changed, 2 insertions(+), 8 deletions(-)

diff --git a/block/blk-mq.c b/block/blk-mq.c
index 1fa9dd1..1e585e3 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -73,8 +73,8 @@ static void blk_mq_hctx_mark_pending(struct blk_mq_hw_ctx *hctx,
 		set_bit(ctx->index_hw, hctx->ctx_map);
 }
 
-static struct request *blk_mq_alloc_rq(struct blk_mq_hw_ctx *hctx, gfp_t gfp,
-				       bool reserved)
+static struct request *__blk_mq_alloc_request(struct blk_mq_hw_ctx *hctx,
+					      gfp_t gfp, bool reserved)
 {
 	struct request *rq;
 	unsigned int tag;
@@ -193,12 +193,6 @@ static void blk_mq_rq_ctx_init(struct request_queue *q, struct blk_mq_ctx *ctx,
 	ctx->rq_dispatched[rw_is_sync(rw_flags)]++;
 }
 
-static struct request *__blk_mq_alloc_request(struct blk_mq_hw_ctx *hctx,
-					      gfp_t gfp, bool reserved)
-{
-	return blk_mq_alloc_rq(hctx, gfp, reserved);
-}
-
 static struct request *blk_mq_alloc_request_pinned(struct request_queue *q,
 						   int rw, gfp_t gfp,
 						   bool reserved)
-- 
1.7.10.4

[PATCH 2/3] blk-mq: merge blk_mq_insert_request and blk_mq_run_request

[Christoph Hellwig]

[-- Attachment #1: 0002-blk-mq-merge-blk_mq_insert_request-and-blk_mq_run_re.patch --]
[-- Type: text/plain, Size: 4784 bytes --]

It's almost identical to blk_mq_insert_request, so fold the two into one
slightly more generic function by making the flush special case a bit
smarted.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-exec.c       |    2 +-
 block/blk-flush.c      |    4 ++--
 block/blk-mq.c         |   53 +++++++++---------------------------------------
 block/blk-mq.h         |    1 -
 include/linux/blk-mq.h |    3 +--
 5 files changed, 14 insertions(+), 49 deletions(-)

diff --git a/block/blk-exec.c b/block/blk-exec.c
index c68613b..dbf4502 100644
--- a/block/blk-exec.c
+++ b/block/blk-exec.c
@@ -65,7 +65,7 @@ void blk_execute_rq_nowait(struct request_queue *q, struct gendisk *bd_disk,
 	 * be resued after dying flag is set
 	 */
 	if (q->mq_ops) {
-		blk_mq_insert_request(q, rq, at_head, true);
+		blk_mq_insert_request(rq, at_head, true, false);
 		return;
 	}
 
diff --git a/block/blk-flush.c b/block/blk-flush.c
index 66e2b69..f598f79 100644
--- a/block/blk-flush.c
+++ b/block/blk-flush.c
@@ -137,7 +137,7 @@ static void mq_flush_run(struct work_struct *work)
 	rq = container_of(work, struct request, mq_flush_work);
 
 	memset(&rq->csd, 0, sizeof(rq->csd));
-	blk_mq_run_request(rq, true, false);
+	blk_mq_insert_request(rq, false, true, false);
 }
 
 static bool blk_flush_queue_rq(struct request *rq)
@@ -411,7 +411,7 @@ void blk_insert_flush(struct request *rq)
 	if ((policy & REQ_FSEQ_DATA) &&
 	    !(policy & (REQ_FSEQ_PREFLUSH | REQ_FSEQ_POSTFLUSH))) {
 		if (q->mq_ops) {
-			blk_mq_run_request(rq, false, true);
+			blk_mq_insert_request(rq, false, false, true);
 		} else
 			list_add_tail(&rq->queuelist, &q->queue_head);
 		return;
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 1e585e3..2af8405 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -724,61 +724,28 @@ static void __blk_mq_insert_request(struct blk_mq_hw_ctx *hctx,
 	blk_mq_add_timer(rq);
 }
 
-void blk_mq_insert_request(struct request_queue *q, struct request *rq,
-			   bool at_head, bool run_queue)
+void blk_mq_insert_request(struct request *rq, bool at_head, bool run_queue,
+		bool async)
 {
+	struct request_queue *q = rq->q;
 	struct blk_mq_hw_ctx *hctx;
-	struct blk_mq_ctx *ctx, *current_ctx;
+	struct blk_mq_ctx *ctx = rq->mq_ctx, *current_ctx;
+
+	current_ctx = blk_mq_get_ctx(q);
+	if (!cpu_online(ctx->cpu))
+		rq->mq_ctx = ctx = current_ctx;
 
-	ctx = rq->mq_ctx;
 	hctx = q->mq_ops->map_queue(q, ctx->cpu);
 
-	if (rq->cmd_flags & (REQ_FLUSH | REQ_FUA)) {
+	if (rq->cmd_flags & (REQ_FLUSH | REQ_FUA) &&
+	    !(rq->cmd_flags & (REQ_FLUSH_SEQ))) {
 		blk_insert_flush(rq);
 	} else {
-		current_ctx = blk_mq_get_ctx(q);
-
-		if (!cpu_online(ctx->cpu)) {
-			ctx = current_ctx;
-			hctx = q->mq_ops->map_queue(q, ctx->cpu);
-			rq->mq_ctx = ctx;
-		}
 		spin_lock(&ctx->lock);
 		__blk_mq_insert_request(hctx, rq, at_head);
 		spin_unlock(&ctx->lock);
-
-		blk_mq_put_ctx(current_ctx);
 	}
 
-	if (run_queue)
-		__blk_mq_run_hw_queue(hctx);
-}
-EXPORT_SYMBOL(blk_mq_insert_request);
-
-/*
- * This is a special version of blk_mq_insert_request to bypass FLUSH request
- * check. Should only be used internally.
- */
-void blk_mq_run_request(struct request *rq, bool run_queue, bool async)
-{
-	struct request_queue *q = rq->q;
-	struct blk_mq_hw_ctx *hctx;
-	struct blk_mq_ctx *ctx, *current_ctx;
-
-	current_ctx = blk_mq_get_ctx(q);
-
-	ctx = rq->mq_ctx;
-	if (!cpu_online(ctx->cpu)) {
-		ctx = current_ctx;
-		rq->mq_ctx = ctx;
-	}
-	hctx = q->mq_ops->map_queue(q, ctx->cpu);
-
-	/* ctx->cpu might be offline */
-	spin_lock(&ctx->lock);
-	__blk_mq_insert_request(hctx, rq, false);
-	spin_unlock(&ctx->lock);
-
 	blk_mq_put_ctx(current_ctx);
 
 	if (run_queue)
diff --git a/block/blk-mq.h b/block/blk-mq.h
index ed0035c..72beba1 100644
--- a/block/blk-mq.h
+++ b/block/blk-mq.h
@@ -23,7 +23,6 @@ struct blk_mq_ctx {
 };
 
 void __blk_mq_complete_request(struct request *rq);
-void blk_mq_run_request(struct request *rq, bool run_queue, bool async);
 void blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx, bool async);
 void blk_mq_init_flush(struct request_queue *q);
 void blk_mq_drain_queue(struct request_queue *q);
diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h
index 18ba8a6..ff28fe3 100644
--- a/include/linux/blk-mq.h
+++ b/include/linux/blk-mq.h
@@ -121,8 +121,7 @@ void blk_mq_init_commands(struct request_queue *, void (*init)(void *data, struc
 
 void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule);
 
-void blk_mq_insert_request(struct request_queue *, struct request *,
-		bool, bool);
+void blk_mq_insert_request(struct request *, bool, bool, bool);
 void blk_mq_run_queues(struct request_queue *q, bool async);
 void blk_mq_free_request(struct request *rq);
 bool blk_mq_can_queue(struct blk_mq_hw_ctx *);
-- 
1.7.10.4

[PATCH 3/3] blk-mq: support partial I/O completions

[Christoph Hellwig]

[-- Attachment #1: 0003-blk-mq-support-partial-I-O-completions.patch --]
[-- Type: text/plain, Size: 2703 bytes --]

Add a new blk_mq_end_io_partial function to partially complete requests
as needed by the SCSI layer.  We do this by reusing blk_update_request
to advance the bio instead of having a simplified version of it in
the blk-mq code.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-mq.c         |   37 +++++--------------------------------
 include/linux/blk-mq.h |    8 +++++++-
 2 files changed, 12 insertions(+), 33 deletions(-)

diff --git a/block/blk-mq.c b/block/blk-mq.c
index 2af8405..1b8b50d 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -283,38 +283,10 @@ void blk_mq_free_request(struct request *rq)
 	__blk_mq_free_request(hctx, ctx, rq);
 }
 
-static void blk_mq_bio_endio(struct request *rq, struct bio *bio, int error)
+bool blk_mq_end_io_partial(struct request *rq, int error, unsigned int nr_bytes)
 {
-	if (error)
-		clear_bit(BIO_UPTODATE, &bio->bi_flags);
-	else if (!test_bit(BIO_UPTODATE, &bio->bi_flags))
-		error = -EIO;
-
-	if (unlikely(rq->cmd_flags & REQ_QUIET))
-		set_bit(BIO_QUIET, &bio->bi_flags);
-
-	/* don't actually finish bio if it's part of flush sequence */
-	if (!(rq->cmd_flags & REQ_FLUSH_SEQ))
-		bio_endio(bio, error);
-}
-
-void blk_mq_end_io(struct request *rq, int error)
-{
-	struct bio *bio = rq->bio;
-	unsigned int bytes = 0;
-
-	trace_block_rq_complete(rq->q, rq);
-
-	while (bio) {
-		struct bio *next = bio->bi_next;
-
-		bio->bi_next = NULL;
-		bytes += bio->bi_iter.bi_size;
-		blk_mq_bio_endio(rq, bio, error);
-		bio = next;
-	}
-
-	blk_account_io_completion(rq, bytes);
+	if (blk_update_request(rq, error, blk_rq_bytes(rq)))
+		return true;
 
 	blk_account_io_done(rq);
 
@@ -322,8 +294,9 @@ void blk_mq_end_io(struct request *rq, int error)
 		rq->end_io(rq, error);
 	else
 		blk_mq_free_request(rq);
+	return false;
 }
-EXPORT_SYMBOL(blk_mq_end_io);
+EXPORT_SYMBOL(blk_mq_end_io_partial);
 
 static void __blk_mq_complete_request_remote(void *data)
 {
diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h
index ff28fe3..2ff2e8d 100644
--- a/include/linux/blk-mq.h
+++ b/include/linux/blk-mq.h
@@ -133,7 +133,13 @@ struct blk_mq_hw_ctx *blk_mq_map_queue(struct request_queue *, const int ctx_ind
 struct blk_mq_hw_ctx *blk_mq_alloc_single_hw_queue(struct blk_mq_reg *, unsigned int);
 void blk_mq_free_single_hw_queue(struct blk_mq_hw_ctx *, unsigned int);
 
-void blk_mq_end_io(struct request *rq, int error);
+bool blk_mq_end_io_partial(struct request *rq, int error,
+		unsigned int nr_bytes);
+static inline void blk_mq_end_io(struct request *rq, int error)
+{
+	bool done = !blk_mq_end_io_partial(rq, error, blk_rq_bytes(rq));
+	BUG_ON(!done);
+}
 
 void blk_mq_complete_request(struct request *rq);
 
-- 
1.7.10.4

Re: [PATCH 0/3] next round of blk-mq updates

[Jens Axboe]

On Thu, Feb 20 2014, Christoph Hellwig wrote:
> A couple more blkmq updates: a trivial cleanup, a consolidation of two very
> similar functions into one, and support for partial request completions,
> a feature needed by the work in progress scsi blk-mq support.

Thanks, applied all 3. The end_io is a bit fatter than it was before,
which is a shame. But on the plus side, it's consistent...

-- 
Jens Axboe