blk-mq: remove blk_mq_put_ctx()

No code that occurs between blk_mq_get_ctx() and blk_mq_put_ctx() depends
on preemption being disabled for its correctness. Since removing the CPU
preemption calls does not measurably affect performance, simplify the
blk-mq code by removing the blk_mq_put_ctx() function and also by not
disabling preemption in blk_mq_get_ctx().

Cc: Hannes Reinecke <hare@suse.com>
Cc: Omar Sandoval <osandov@fb.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
diff --git a/block/blk-mq-sched.c b/block/blk-mq-sched.c
index 956a7aa..c9d183d 100644
--- a/block/blk-mq-sched.c
+++ b/block/blk-mq-sched.c
@@ -330,10 +330,8 @@
 	bool ret = false;
 	enum hctx_type type;
 
-	if (e && e->type->ops.bio_merge) {
-		blk_mq_put_ctx(ctx);
+	if (e && e->type->ops.bio_merge)
 		return e->type->ops.bio_merge(hctx, bio, nr_segs);
-	}
 
 	type = hctx->type;
 	if ((hctx->flags & BLK_MQ_F_SHOULD_MERGE) &&
@@ -344,7 +342,6 @@
 		spin_unlock(&ctx->lock);
 	}
 
-	blk_mq_put_ctx(ctx);
 	return ret;
 }
 
diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c
index 7513c8e..da19f0b 100644
--- a/block/blk-mq-tag.c
+++ b/block/blk-mq-tag.c
@@ -113,7 +113,6 @@
 	struct sbq_wait_state *ws;
 	DEFINE_SBQ_WAIT(wait);
 	unsigned int tag_offset;
-	bool drop_ctx;
 	int tag;
 
 	if (data->flags & BLK_MQ_REQ_RESERVED) {
@@ -136,7 +135,6 @@
 		return BLK_MQ_TAG_FAIL;
 
 	ws = bt_wait_ptr(bt, data->hctx);
-	drop_ctx = data->ctx == NULL;
 	do {
 		struct sbitmap_queue *bt_prev;
 
@@ -161,9 +159,6 @@
 		if (tag != -1)
 			break;
 
-		if (data->ctx)
-			blk_mq_put_ctx(data->ctx);
-
 		bt_prev = bt;
 		io_schedule();
 
@@ -189,9 +184,6 @@
 		ws = bt_wait_ptr(bt, data->hctx);
 	} while (1);
 
-	if (drop_ctx && data->ctx)
-		blk_mq_put_ctx(data->ctx);
-
 	sbitmap_finish_wait(bt, ws, &wait);
 
 found_tag:
diff --git a/block/blk-mq.c b/block/blk-mq.c
index d893838..0cb1b152 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -355,13 +355,13 @@
 	struct elevator_queue *e = q->elevator;
 	struct request *rq;
 	unsigned int tag;
-	bool put_ctx_on_error = false;
+	bool clear_ctx_on_error = false;
 
 	blk_queue_enter_live(q);
 	data->q = q;
 	if (likely(!data->ctx)) {
 		data->ctx = blk_mq_get_ctx(q);
-		put_ctx_on_error = true;
+		clear_ctx_on_error = true;
 	}
 	if (likely(!data->hctx))
 		data->hctx = blk_mq_map_queue(q, data->cmd_flags,
@@ -387,10 +387,8 @@
 
 	tag = blk_mq_get_tag(data);
 	if (tag == BLK_MQ_TAG_FAIL) {
-		if (put_ctx_on_error) {
-			blk_mq_put_ctx(data->ctx);
+		if (clear_ctx_on_error)
 			data->ctx = NULL;
-		}
 		blk_queue_exit(q);
 		return NULL;
 	}
@@ -427,8 +425,6 @@
 	if (!rq)
 		return ERR_PTR(-EWOULDBLOCK);
 
-	blk_mq_put_ctx(alloc_data.ctx);
-
 	rq->__data_len = 0;
 	rq->__sector = (sector_t) -1;
 	rq->bio = rq->biotail = NULL;
@@ -1977,7 +1973,6 @@
 
 	plug = current->plug;
 	if (unlikely(is_flush_fua)) {
-		blk_mq_put_ctx(data.ctx);
 		blk_mq_bio_to_request(rq, bio, nr_segs);
 
 		/* bypass scheduler for flush rq */
@@ -1991,7 +1986,6 @@
 		unsigned int request_count = plug->rq_count;
 		struct request *last = NULL;
 
-		blk_mq_put_ctx(data.ctx);
 		blk_mq_bio_to_request(rq, bio, nr_segs);
 
 		if (!request_count)
@@ -2025,8 +2019,6 @@
 		blk_add_rq_to_plug(plug, rq);
 		trace_block_plug(q);
 
-		blk_mq_put_ctx(data.ctx);
-
 		if (same_queue_rq) {
 			data.hctx = same_queue_rq->mq_hctx;
 			trace_block_unplug(q, 1, true);
@@ -2035,11 +2027,9 @@
 		}
 	} else if ((q->nr_hw_queues > 1 && is_sync) || (!q->elevator &&
 			!data.hctx->dispatch_busy)) {
-		blk_mq_put_ctx(data.ctx);
 		blk_mq_bio_to_request(rq, bio, nr_segs);
 		blk_mq_try_issue_directly(data.hctx, rq, &cookie);
 	} else {
-		blk_mq_put_ctx(data.ctx);
 		blk_mq_bio_to_request(rq, bio, nr_segs);
 		blk_mq_sched_insert_request(rq, false, true, true);
 	}
diff --git a/block/blk-mq.h b/block/blk-mq.h
index 633a5a7..f4bf516 100644
--- a/block/blk-mq.h
+++ b/block/blk-mq.h
@@ -151,12 +151,7 @@
  */
 static inline struct blk_mq_ctx *blk_mq_get_ctx(struct request_queue *q)
 {
-	return __blk_mq_get_ctx(q, get_cpu());
-}
-
-static inline void blk_mq_put_ctx(struct blk_mq_ctx *ctx)
-{
-	put_cpu();
+	return __blk_mq_get_ctx(q, raw_smp_processor_id());
 }
 
 struct blk_mq_alloc_data {
diff --git a/block/kyber-iosched.c b/block/kyber-iosched.c
index 3c26026..34dcea0 100644
--- a/block/kyber-iosched.c
+++ b/block/kyber-iosched.c
@@ -575,7 +575,6 @@
 	spin_lock(&kcq->lock);
 	merged = blk_mq_bio_list_merge(hctx->queue, rq_list, bio, nr_segs);
 	spin_unlock(&kcq->lock);
-	blk_mq_put_ctx(ctx);
 
 	return merged;
 }