[PATCH v4 05/10] block: introduce blk_flush_queue to drive flush machinery
From: Ming Lei
Date: Mon Sep 15 2014 - 09:11:46 EST
This patch introduces 'struct blk_flush_queue' and puts all
flush machinery related fields into this structure, so that
- flush implementation details aren't exposed to driver
- it is easy to convert to per dispatch-queue flush machinery
This patch is basically a mechanical replacement.
Reviewed-by: Christoph Hellwig <hch@xxxxxx>
Signed-off-by: Ming Lei <ming.lei@xxxxxxxxxxxxx>
---
block/blk-core.c | 3 +-
block/blk-flush.c | 107 +++++++++++++++++++++++++++++-------------------
block/blk-mq.c | 10 +++--
block/blk.h | 22 +++++++++-
include/linux/blkdev.h | 10 +----
5 files changed, 95 insertions(+), 57 deletions(-)
diff --git a/block/blk-core.c b/block/blk-core.c
index 222fe84..d278a30 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -390,11 +390,12 @@ static void __blk_drain_queue(struct request_queue *q, bool drain_all)
* be drained. Check all the queues and counters.
*/
if (drain_all) {
+ struct blk_flush_queue *fq = blk_get_flush_queue(q);
drain |= !list_empty(&q->queue_head);
for (i = 0; i < 2; i++) {
drain |= q->nr_rqs[i];
drain |= q->in_flight[i];
- drain |= !list_empty(&q->flush_queue[i]);
+ drain |= !list_empty(&fq->flush_queue[i]);
}
}
diff --git a/block/blk-flush.c b/block/blk-flush.c
index a59dd1a..db269d4 100644
--- a/block/blk-flush.c
+++ b/block/blk-flush.c
@@ -28,7 +28,7 @@
*
* The actual execution of flush is double buffered. Whenever a request
* needs to execute PRE or POSTFLUSH, it queues at
- * q->flush_queue[q->flush_pending_idx]. Once certain criteria are met, a
+ * fq->flush_queue[fq->flush_pending_idx]. Once certain criteria are met, a
* flush is issued and the pending_idx is toggled. When the flush
* completes, all the requests which were pending are proceeded to the next
* step. This allows arbitrary merging of different types of FLUSH/FUA
@@ -157,7 +157,7 @@ static bool blk_flush_queue_rq(struct request *rq, bool add_front)
* completion and trigger the next step.
*
* CONTEXT:
- * spin_lock_irq(q->queue_lock or q->mq_flush_lock)
+ * spin_lock_irq(q->queue_lock or fq->mq_flush_lock)
*
* RETURNS:
* %true if requests were added to the dispatch queue, %false otherwise.
@@ -166,7 +166,8 @@ static bool blk_flush_complete_seq(struct request *rq, unsigned int seq,
int error)
{
struct request_queue *q = rq->q;
- struct list_head *pending = &q->flush_queue[q->flush_pending_idx];
+ struct blk_flush_queue *fq = blk_get_flush_queue(q);
+ struct list_head *pending = &fq->flush_queue[fq->flush_pending_idx];
bool queued = false, kicked;
BUG_ON(rq->flush.seq & seq);
@@ -182,12 +183,12 @@ static bool blk_flush_complete_seq(struct request *rq, unsigned int seq,
case REQ_FSEQ_POSTFLUSH:
/* queue for flush */
if (list_empty(pending))
- q->flush_pending_since = jiffies;
+ fq->flush_pending_since = jiffies;
list_move_tail(&rq->flush.list, pending);
break;
case REQ_FSEQ_DATA:
- list_move_tail(&rq->flush.list, &q->flush_data_in_flight);
+ list_move_tail(&rq->flush.list, &fq->flush_data_in_flight);
queued = blk_flush_queue_rq(rq, true);
break;
@@ -222,17 +223,18 @@ static void flush_end_io(struct request *flush_rq, int error)
bool queued = false;
struct request *rq, *n;
unsigned long flags = 0;
+ struct blk_flush_queue *fq = blk_get_flush_queue(q);
if (q->mq_ops) {
- spin_lock_irqsave(&q->mq_flush_lock, flags);
+ spin_lock_irqsave(&fq->mq_flush_lock, flags);
flush_rq->tag = -1;
}
- running = &q->flush_queue[q->flush_running_idx];
- BUG_ON(q->flush_pending_idx == q->flush_running_idx);
+ running = &fq->flush_queue[fq->flush_running_idx];
+ BUG_ON(fq->flush_pending_idx == fq->flush_running_idx);
/* account completion of the flush request */
- q->flush_running_idx ^= 1;
+ fq->flush_running_idx ^= 1;
if (!q->mq_ops)
elv_completed_request(q, flush_rq);
@@ -256,13 +258,13 @@ static void flush_end_io(struct request *flush_rq, int error)
* directly into request_fn may confuse the driver. Always use
* kblockd.
*/
- if (queued || q->flush_queue_delayed) {
+ if (queued || fq->flush_queue_delayed) {
WARN_ON(q->mq_ops);
blk_run_queue_async(q);
}
- q->flush_queue_delayed = 0;
+ fq->flush_queue_delayed = 0;
if (q->mq_ops)
- spin_unlock_irqrestore(&q->mq_flush_lock, flags);
+ spin_unlock_irqrestore(&fq->mq_flush_lock, flags);
}
/**
@@ -273,33 +275,34 @@ static void flush_end_io(struct request *flush_rq, int error)
* Please read the comment at the top of this file for more info.
*
* CONTEXT:
- * spin_lock_irq(q->queue_lock or q->mq_flush_lock)
+ * spin_lock_irq(q->queue_lock or fq->mq_flush_lock)
*
* RETURNS:
* %true if flush was issued, %false otherwise.
*/
static bool blk_kick_flush(struct request_queue *q)
{
- struct list_head *pending = &q->flush_queue[q->flush_pending_idx];
+ struct blk_flush_queue *fq = blk_get_flush_queue(q);
+ struct list_head *pending = &fq->flush_queue[fq->flush_pending_idx];
struct request *first_rq =
list_first_entry(pending, struct request, flush.list);
- struct request *flush_rq = q->flush_rq;
+ struct request *flush_rq = fq->flush_rq;
/* C1 described at the top of this file */
- if (q->flush_pending_idx != q->flush_running_idx || list_empty(pending))
+ if (fq->flush_pending_idx != fq->flush_running_idx || list_empty(pending))
return false;
/* C2 and C3 */
- if (!list_empty(&q->flush_data_in_flight) &&
+ if (!list_empty(&fq->flush_data_in_flight) &&
time_before(jiffies,
- q->flush_pending_since + FLUSH_PENDING_TIMEOUT))
+ fq->flush_pending_since + FLUSH_PENDING_TIMEOUT))
return false;
/*
* Issue flush and toggle pending_idx. This makes pending_idx
* different from running_idx, which means flush is in flight.
*/
- q->flush_pending_idx ^= 1;
+ fq->flush_pending_idx ^= 1;
blk_rq_init(q, flush_rq);
if (q->mq_ops)
@@ -331,6 +334,7 @@ static void mq_flush_data_end_io(struct request *rq, int error)
struct blk_mq_hw_ctx *hctx;
struct blk_mq_ctx *ctx;
unsigned long flags;
+ struct blk_flush_queue *fq = blk_get_flush_queue(q);
ctx = rq->mq_ctx;
hctx = q->mq_ops->map_queue(q, ctx->cpu);
@@ -339,10 +343,10 @@ static void mq_flush_data_end_io(struct request *rq, int error)
* After populating an empty queue, kick it to avoid stall. Read
* the comment in flush_end_io().
*/
- spin_lock_irqsave(&q->mq_flush_lock, flags);
+ spin_lock_irqsave(&fq->mq_flush_lock, flags);
if (blk_flush_complete_seq(rq, REQ_FSEQ_DATA, error))
blk_mq_run_hw_queue(hctx, true);
- spin_unlock_irqrestore(&q->mq_flush_lock, flags);
+ spin_unlock_irqrestore(&fq->mq_flush_lock, flags);
}
/**
@@ -410,11 +414,13 @@ void blk_insert_flush(struct request *rq)
rq->cmd_flags |= REQ_FLUSH_SEQ;
rq->flush.saved_end_io = rq->end_io; /* Usually NULL */
if (q->mq_ops) {
+ struct blk_flush_queue *fq = blk_get_flush_queue(q);
+
rq->end_io = mq_flush_data_end_io;
- spin_lock_irq(&q->mq_flush_lock);
+ spin_lock_irq(&fq->mq_flush_lock);
blk_flush_complete_seq(rq, REQ_FSEQ_ACTIONS & ~policy, 0);
- spin_unlock_irq(&q->mq_flush_lock);
+ spin_unlock_irq(&fq->mq_flush_lock);
return;
}
rq->end_io = flush_data_end_io;
@@ -475,31 +481,48 @@ int blkdev_issue_flush(struct block_device *bdev, gfp_t gfp_mask,
}
EXPORT_SYMBOL(blkdev_issue_flush);
-static int blk_mq_init_flush(struct request_queue *q)
+static struct blk_flush_queue *blk_alloc_flush_queue(
+ struct request_queue *q)
{
- struct blk_mq_tag_set *set = q->tag_set;
+ struct blk_flush_queue *fq;
+ int rq_sz = sizeof(struct request);
- spin_lock_init(&q->mq_flush_lock);
+ fq = kzalloc(sizeof(*fq), GFP_KERNEL);
+ if (!fq)
+ goto fail;
- q->flush_rq = kzalloc(round_up(sizeof(struct request) +
- set->cmd_size, cache_line_size()),
- GFP_KERNEL);
- if (!q->flush_rq)
- return -ENOMEM;
- return 0;
+ if (q->mq_ops) {
+ spin_lock_init(&fq->mq_flush_lock);
+ rq_sz = round_up(rq_sz + q->tag_set->cmd_size,
+ cache_line_size());
+ }
+
+ fq->flush_rq = kzalloc(rq_sz, GFP_KERNEL);
+ if (!fq->flush_rq)
+ goto fail_rq;
+
+ INIT_LIST_HEAD(&fq->flush_queue[0]);
+ INIT_LIST_HEAD(&fq->flush_queue[1]);
+ INIT_LIST_HEAD(&fq->flush_data_in_flight);
+
+ return fq;
+
+ fail_rq:
+ kfree(fq);
+ fail:
+ return NULL;
}
-int blk_init_flush(struct request_queue *q)
+static void blk_free_flush_queue(struct blk_flush_queue *fq)
{
- INIT_LIST_HEAD(&q->flush_queue[0]);
- INIT_LIST_HEAD(&q->flush_queue[1]);
- INIT_LIST_HEAD(&q->flush_data_in_flight);
-
- if (q->mq_ops)
- return blk_mq_init_flush(q);
+ kfree(fq->flush_rq);
+ kfree(fq);
+}
- q->flush_rq = kzalloc(sizeof(struct request), GFP_KERNEL);
- if (!q->flush_rq)
+int blk_init_flush(struct request_queue *q)
+{
+ q->fq = blk_alloc_flush_queue(q);
+ if (!q->fq)
return -ENOMEM;
return 0;
@@ -507,5 +530,5 @@ int blk_init_flush(struct request_queue *q)
void blk_exit_flush(struct request_queue *q)
{
- kfree(q->flush_rq);
+ blk_free_flush_queue(q->fq);
}
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 467b1d8..a819af4 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -508,20 +508,22 @@ void blk_mq_kick_requeue_list(struct request_queue *q)
}
EXPORT_SYMBOL(blk_mq_kick_requeue_list);
-static inline bool is_flush_request(struct request *rq, unsigned int tag)
+static inline bool is_flush_request(struct request *rq,
+ struct blk_flush_queue *fq, unsigned int tag)
{
return ((rq->cmd_flags & REQ_FLUSH_SEQ) &&
- rq->q->flush_rq->tag == tag);
+ fq->flush_rq->tag == tag);
}
struct request *blk_mq_tag_to_rq(struct blk_mq_tags *tags, unsigned int tag)
{
struct request *rq = tags->rqs[tag];
+ struct blk_flush_queue *fq = blk_get_flush_queue(rq->q);
- if (!is_flush_request(rq, tag))
+ if (!is_flush_request(rq, fq, tag))
return rq;
- return rq->q->flush_rq;
+ return fq->flush_rq;
}
EXPORT_SYMBOL(blk_mq_tag_to_rq);
diff --git a/block/blk.h b/block/blk.h
index 261f734..2637349 100644
--- a/block/blk.h
+++ b/block/blk.h
@@ -12,11 +12,28 @@
/* Max future timer expiry for timeouts */
#define BLK_MAX_TIMEOUT (5 * HZ)
+struct blk_flush_queue {
+ unsigned int flush_queue_delayed:1;
+ unsigned int flush_pending_idx:1;
+ unsigned int flush_running_idx:1;
+ unsigned long flush_pending_since;
+ struct list_head flush_queue[2];
+ struct list_head flush_data_in_flight;
+ struct request *flush_rq;
+ spinlock_t mq_flush_lock;
+};
+
extern struct kmem_cache *blk_requestq_cachep;
extern struct kmem_cache *request_cachep;
extern struct kobj_type blk_queue_ktype;
extern struct ida blk_queue_ida;
+static inline struct blk_flush_queue *blk_get_flush_queue(
+ struct request_queue *q)
+{
+ return q->fq;
+}
+
static inline void __blk_get_queue(struct request_queue *q)
{
kobject_get(&q->kobj);
@@ -91,6 +108,7 @@ void blk_insert_flush(struct request *rq);
static inline struct request *__elv_next_request(struct request_queue *q)
{
struct request *rq;
+ struct blk_flush_queue *fq = blk_get_flush_queue(q);
while (1) {
if (!list_empty(&q->queue_head)) {
@@ -113,9 +131,9 @@ static inline struct request *__elv_next_request(struct request_queue *q)
* should be restarted later. Please see flush_end_io() for
* details.
*/
- if (q->flush_pending_idx != q->flush_running_idx &&
+ if (fq->flush_pending_idx != fq->flush_running_idx &&
!queue_flush_queueable(q)) {
- q->flush_queue_delayed = 1;
+ fq->flush_queue_delayed = 1;
return NULL;
}
if (unlikely(blk_queue_bypass(q)) ||
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index e267bf0..49f3461 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -36,6 +36,7 @@ struct request;
struct sg_io_hdr;
struct bsg_job;
struct blkcg_gq;
+struct blk_flush_queue;
#define BLKDEV_MIN_RQ 4
#define BLKDEV_MAX_RQ 128 /* Default maximum */
@@ -455,14 +456,7 @@ struct request_queue {
*/
unsigned int flush_flags;
unsigned int flush_not_queueable:1;
- unsigned int flush_queue_delayed:1;
- unsigned int flush_pending_idx:1;
- unsigned int flush_running_idx:1;
- unsigned long flush_pending_since;
- struct list_head flush_queue[2];
- struct list_head flush_data_in_flight;
- struct request *flush_rq;
- spinlock_t mq_flush_lock;
+ struct blk_flush_queue *fq;
struct list_head requeue_list;
spinlock_t requeue_lock;
--
1.7.9.5
--
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/