提交 449de6b8 编写于 作者: T Tejun Heo 提交者: Yang Yingliang

blk-mq: add optional request->alloc_time_ns

mainline inclusion
from mainline-5.4-rc1
commit 6f816b4b746c2241540e537682d30d8e9997d674
category: feature
bugzilla: 38688
CVE: NA

---------------------------

There are currently two start time timestamps - start_time_ns and
io_start_time_ns.  The former marks the request allocation and and the
second issue-to-device time.  The planned io.weight controller needs
to measure the total time bios take to execute after it leaves rq_qos
including the time spent waiting for request to become available,
which can easily dominate on saturated devices.

This patch adds request->alloc_time_ns which records when the request
allocation attempt started.  As it isn't used for the usual stats,
make it optional behind CONFIG_BLK_RQ_ALLOC_TIME and
QUEUE_FLAG_RQ_ALLOC_TIME so that it can be compiled out when there are
no users and it's active only on queues which need it even when
compiled in.

v2: s/pre_start_time/alloc_time/ and add CONFIG_BLK_RQ_ALLOC_TIME
    gating as suggested by Jens.
Signed-off-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Conflict:
  include/linux/blkdev.h
  block/Kconfig
  block/blk-mq.c
Signed-off-by: NYu Kuai <yukuai3@huawei.com>
Reviewed-by: NHou Tao <houtao1@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
上级 1f3268e2
...@@ -50,6 +50,9 @@ config LBDAF ...@@ -50,6 +50,9 @@ config LBDAF
If unsure, say Y. If unsure, say Y.
config BLK_RQ_ALLOC_TIME
bool
config BLK_SCSI_REQUEST config BLK_SCSI_REQUEST
bool bool
......
...@@ -275,7 +275,7 @@ bool blk_mq_can_queue(struct blk_mq_hw_ctx *hctx) ...@@ -275,7 +275,7 @@ bool blk_mq_can_queue(struct blk_mq_hw_ctx *hctx)
EXPORT_SYMBOL(blk_mq_can_queue); EXPORT_SYMBOL(blk_mq_can_queue);
static struct request *blk_mq_rq_ctx_init(struct blk_mq_alloc_data *data, static struct request *blk_mq_rq_ctx_init(struct blk_mq_alloc_data *data,
unsigned int tag, unsigned int op) unsigned int tag, unsigned int op, u64 alloc_time_ns)
{ {
struct blk_mq_tags *tags = blk_mq_tags_from_data(data); struct blk_mq_tags *tags = blk_mq_tags_from_data(data);
struct request *rq = tags->static_rqs[tag]; struct request *rq = tags->static_rqs[tag];
...@@ -309,6 +309,9 @@ static struct request *blk_mq_rq_ctx_init(struct blk_mq_alloc_data *data, ...@@ -309,6 +309,9 @@ static struct request *blk_mq_rq_ctx_init(struct blk_mq_alloc_data *data,
RB_CLEAR_NODE(&rq->rb_node); RB_CLEAR_NODE(&rq->rb_node);
rq->rq_disk = NULL; rq->rq_disk = NULL;
rq->part = NULL; rq->part = NULL;
#ifdef CONFIG_BLK_RQ_ALLOC_TIME
rq->alloc_time_ns = alloc_time_ns;
#endif
rq->start_time_ns = ktime_get_ns(); rq->start_time_ns = ktime_get_ns();
rq->io_start_time_ns = 0; rq->io_start_time_ns = 0;
rq->nr_phys_segments = 0; rq->nr_phys_segments = 0;
...@@ -344,8 +347,14 @@ static struct request *blk_mq_get_request(struct request_queue *q, ...@@ -344,8 +347,14 @@ static struct request *blk_mq_get_request(struct request_queue *q,
struct request *rq; struct request *rq;
unsigned int tag; unsigned int tag;
bool put_ctx_on_error = false; bool put_ctx_on_error = false;
u64 alloc_time_ns = 0;
blk_queue_enter_live(q); blk_queue_enter_live(q);
/* alloc_time includes depth and tag waits */
if (blk_queue_rq_alloc_time(q))
alloc_time_ns = ktime_get_ns();
data->q = q; data->q = q;
if (likely(!data->ctx)) { if (likely(!data->ctx)) {
data->ctx = blk_mq_get_ctx(q); data->ctx = blk_mq_get_ctx(q);
...@@ -381,7 +390,7 @@ static struct request *blk_mq_get_request(struct request_queue *q, ...@@ -381,7 +390,7 @@ static struct request *blk_mq_get_request(struct request_queue *q,
return NULL; return NULL;
} }
rq = blk_mq_rq_ctx_init(data, tag, op); rq = blk_mq_rq_ctx_init(data, tag, op, alloc_time_ns);
if (!op_is_flush(op)) { if (!op_is_flush(op)) {
rq->elv.icq = NULL; rq->elv.icq = NULL;
if (e && e->type->ops.mq.prepare_request) { if (e && e->type->ops.mq.prepare_request) {
......
...@@ -212,7 +212,11 @@ struct request { ...@@ -212,7 +212,11 @@ struct request {
struct gendisk *rq_disk; struct gendisk *rq_disk;
struct hd_struct *part; struct hd_struct *part;
/* Time that I/O was submitted to the kernel. */ #ifdef CONFIG_BLK_RQ_ALLOC_TIME
/* Time that the first bio started allocating this request. */
u64 alloc_time_ns;
#endif
/* Time that this request was allocated for this IO. */
u64 start_time_ns; u64 start_time_ns;
/* Time that I/O was submitted to the device. */ /* Time that I/O was submitted to the device. */
u64 io_start_time_ns; u64 io_start_time_ns;
...@@ -720,6 +724,7 @@ struct request_queue { ...@@ -720,6 +724,7 @@ struct request_queue {
#define QUEUE_FLAG_REGISTERED 26 /* queue has been registered to a disk */ #define QUEUE_FLAG_REGISTERED 26 /* queue has been registered to a disk */
#define QUEUE_FLAG_SCSI_PASSTHROUGH 27 /* queue supports SCSI commands */ #define QUEUE_FLAG_SCSI_PASSTHROUGH 27 /* queue supports SCSI commands */
#define QUEUE_FLAG_QUIESCED 28 /* queue has been quiesced */ #define QUEUE_FLAG_QUIESCED 28 /* queue has been quiesced */
#define QUEUE_FLAG_RQ_ALLOC_TIME 27 /* record rq->alloc_time_ns */
#define QUEUE_FLAG_DEFAULT ((1 << QUEUE_FLAG_IO_STAT) | \ #define QUEUE_FLAG_DEFAULT ((1 << QUEUE_FLAG_IO_STAT) | \
(1 << QUEUE_FLAG_SAME_COMP) | \ (1 << QUEUE_FLAG_SAME_COMP) | \
...@@ -753,6 +758,13 @@ bool blk_queue_flag_test_and_clear(unsigned int flag, struct request_queue *q); ...@@ -753,6 +758,13 @@ bool blk_queue_flag_test_and_clear(unsigned int flag, struct request_queue *q);
#define blk_queue_scsi_passthrough(q) \ #define blk_queue_scsi_passthrough(q) \
test_bit(QUEUE_FLAG_SCSI_PASSTHROUGH, &(q)->queue_flags) test_bit(QUEUE_FLAG_SCSI_PASSTHROUGH, &(q)->queue_flags)
#ifdef CONFIG_BLK_RQ_ALLOC_TIME
#define blk_queue_rq_alloc_time(q) \
test_bit(QUEUE_FLAG_RQ_ALLOC_TIME, &(q)->queue_flags)
#else
#define blk_queue_rq_alloc_time(q) false
#endif
#define blk_noretry_request(rq) \ #define blk_noretry_request(rq) \
((rq)->cmd_flags & (REQ_FAILFAST_DEV|REQ_FAILFAST_TRANSPORT| \ ((rq)->cmd_flags & (REQ_FAILFAST_DEV|REQ_FAILFAST_TRANSPORT| \
REQ_FAILFAST_DRIVER)) REQ_FAILFAST_DRIVER))
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册