From: Ming Lei <ming.lei@redhat.com>
To: Jens Axboe <axboe@kernel.dk>
Cc: linux-block@vger.kernel.org, Ming Lei <ming.lei@redhat.com>,
Alan Stern <stern@rowland.harvard.edu>,
Christoph Hellwig <hch@lst.de>,
Bart Van Assche <bart.vanassche@wdc.com>,
Jianchao Wang <jianchao.w.wang@oracle.com>,
Hannes Reinecke <hare@suse.de>,
Johannes Thumshirn <jthumshirn@suse.de>,
Adrian Hunter <adrian.hunter@intel.com>,
"James E.J. Bottomley" <jejb@linux.vnet.ibm.com>,
"Martin K. Petersen" <martin.petersen@oracle.com>,
linux-scsi@vger.kernel.org
Subject: [RFC PATCH V2 15/17] SCSI: use admin queue to implement queue QUIESCE
Date: Sat, 11 Aug 2018 15:12:18 +0800 [thread overview]
Message-ID: <20180811071220.357-16-ming.lei@redhat.com> (raw)
In-Reply-To: <20180811071220.357-1-ming.lei@redhat.com>
All admin commands are sent via per-host admin queue, so we can simply
freeze the IO queue for quiescing scsi device.
Also the current SCSI stack guarantees that any request originated from
admin queue won't be called back to block layer via the associated IO queue,
and it is always dealt with by the admin queue.
So it is safe to submit admin request via admin queue when the associated IO
queue is frozen, and this way matches the PREEMPT flag perfectly.
Finally, we can remove the preempt_only approach for supporting SCSI
quiesce, then the code in block fast path is simplified a lot.
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Jianchao Wang <jianchao.w.wang@oracle.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: "James E.J. Bottomley" <jejb@linux.vnet.ibm.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: linux-scsi@vger.kernel.org
Signed-off-by: Ming Lei <ming.lei@redhat.com>
---
block/blk-core.c | 44 ++------------------------------------------
block/blk-mq-debugfs.c | 1 -
drivers/scsi/scsi_lib.c | 29 +++--------------------------
include/linux/blkdev.h | 6 ------
include/scsi/scsi_device.h | 1 -
5 files changed, 5 insertions(+), 76 deletions(-)
diff --git a/block/blk-core.c b/block/blk-core.c
index 12550340418d..67d34a43359f 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -421,26 +421,6 @@ void blk_sync_queue(struct request_queue *q)
EXPORT_SYMBOL(blk_sync_queue);
/**
- * blk_set_preempt_only - set QUEUE_FLAG_PREEMPT_ONLY
- * @q: request queue pointer
- *
- * Returns the previous value of the PREEMPT_ONLY flag - 0 if the flag was not
- * set and 1 if the flag was already set.
- */
-int blk_set_preempt_only(struct request_queue *q)
-{
- return blk_queue_flag_test_and_set(QUEUE_FLAG_PREEMPT_ONLY, q);
-}
-EXPORT_SYMBOL_GPL(blk_set_preempt_only);
-
-void blk_clear_preempt_only(struct request_queue *q)
-{
- blk_queue_flag_clear(QUEUE_FLAG_PREEMPT_ONLY, q);
- wake_up_all(&q->mq_freeze_wq);
-}
-EXPORT_SYMBOL_GPL(blk_clear_preempt_only);
-
-/**
* __blk_run_queue_uncond - run a queue whether or not it has been stopped
* @q: The queue to run
*
@@ -917,27 +897,8 @@ EXPORT_SYMBOL(blk_alloc_queue);
*/
int blk_queue_enter(struct request_queue *q, blk_mq_req_flags_t flags)
{
- const bool preempt = flags & BLK_MQ_REQ_PREEMPT;
-
while (true) {
- bool success = false;
-
- rcu_read_lock();
- if (percpu_ref_tryget_live(&q->q_usage_counter)) {
- /*
- * The code that sets the PREEMPT_ONLY flag is
- * responsible for ensuring that that flag is globally
- * visible before the queue is unfrozen.
- */
- if (preempt || !blk_queue_preempt_only(q)) {
- success = true;
- } else {
- percpu_ref_put(&q->q_usage_counter);
- }
- }
- rcu_read_unlock();
-
- if (success)
+ if (percpu_ref_tryget_live(&q->q_usage_counter))
return 0;
if (flags & BLK_MQ_REQ_NOWAIT)
@@ -953,8 +914,7 @@ int blk_queue_enter(struct request_queue *q, blk_mq_req_flags_t flags)
smp_rmb();
wait_event(q->mq_freeze_wq,
- (atomic_read(&q->mq_freeze_depth) == 0 &&
- (preempt || !blk_queue_preempt_only(q))) ||
+ atomic_read(&q->mq_freeze_depth) == 0 ||
blk_queue_dying(q));
if (blk_queue_dying(q))
return -ENODEV;
diff --git a/block/blk-mq-debugfs.c b/block/blk-mq-debugfs.c
index 8df013e9f242..82df43ec322f 100644
--- a/block/blk-mq-debugfs.c
+++ b/block/blk-mq-debugfs.c
@@ -132,7 +132,6 @@ static const char *const blk_queue_flag_name[] = {
QUEUE_FLAG_NAME(REGISTERED),
QUEUE_FLAG_NAME(SCSI_PASSTHROUGH),
QUEUE_FLAG_NAME(QUIESCED),
- QUEUE_FLAG_NAME(PREEMPT_ONLY),
QUEUE_FLAG_NAME(ADMIN),
};
#undef QUEUE_FLAG_NAME
diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index ebd36ae549e8..965781e2879c 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -3130,34 +3130,12 @@ static void scsi_wait_for_queuecommand(struct scsi_device *sdev)
int
scsi_device_quiesce(struct scsi_device *sdev)
{
- struct request_queue *q = sdev->request_queue;
int err;
- /*
- * It is allowed to call scsi_device_quiesce() multiple times from
- * the same context but concurrent scsi_device_quiesce() calls are
- * not allowed.
- */
- WARN_ON_ONCE(sdev->quiesced_by && sdev->quiesced_by != current);
-
- blk_set_preempt_only(q);
-
- blk_mq_freeze_queue(q);
- /*
- * Ensure that the effect of blk_set_preempt_only() will be visible
- * for percpu_ref_tryget() callers that occur after the queue
- * unfreeze even if the queue was already frozen before this function
- * was called. See also https://lwn.net/Articles/573497/.
- */
- synchronize_rcu();
- blk_mq_unfreeze_queue(q);
+ blk_mq_freeze_queue(sdev->request_queue);
mutex_lock(&sdev->state_mutex);
err = scsi_device_set_state(sdev, SDEV_QUIESCE);
- if (err == 0)
- sdev->quiesced_by = current;
- else
- blk_clear_preempt_only(q);
mutex_unlock(&sdev->state_mutex);
return err;
@@ -3180,12 +3158,11 @@ void scsi_device_resume(struct scsi_device *sdev)
* device deleted during suspend)
*/
mutex_lock(&sdev->state_mutex);
- WARN_ON_ONCE(!sdev->quiesced_by);
- sdev->quiesced_by = NULL;
- blk_clear_preempt_only(sdev->request_queue);
if (sdev->sdev_state == SDEV_QUIESCE)
scsi_device_set_state(sdev, SDEV_RUNNING);
mutex_unlock(&sdev->state_mutex);
+
+ blk_mq_unfreeze_queue(sdev->request_queue);
}
EXPORT_SYMBOL(scsi_device_resume);
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 2dbc7524a169..1bd4f02d11c0 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -698,7 +698,6 @@ struct request_queue {
#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_QUIESCED 28 /* queue has been quiesced */
-#define QUEUE_FLAG_PREEMPT_ONLY 29 /* only process REQ_PREEMPT requests */
#define QUEUE_FLAG_ADMIN 30 /* admin queue */
#define QUEUE_FLAG_DEFAULT ((1 << QUEUE_FLAG_IO_STAT) | \
@@ -740,14 +739,9 @@ bool blk_queue_flag_test_and_clear(unsigned int flag, struct request_queue *q);
((rq)->cmd_flags & (REQ_FAILFAST_DEV|REQ_FAILFAST_TRANSPORT| \
REQ_FAILFAST_DRIVER))
#define blk_queue_quiesced(q) test_bit(QUEUE_FLAG_QUIESCED, &(q)->queue_flags)
-#define blk_queue_preempt_only(q) \
- test_bit(QUEUE_FLAG_PREEMPT_ONLY, &(q)->queue_flags)
#define blk_queue_fua(q) test_bit(QUEUE_FLAG_FUA, &(q)->queue_flags)
#define blk_queue_admin(q) test_bit(QUEUE_FLAG_ADMIN, &(q)->queue_flags)
-extern int blk_set_preempt_only(struct request_queue *q);
-extern void blk_clear_preempt_only(struct request_queue *q);
-
static inline int queue_in_flight(struct request_queue *q)
{
return q->in_flight[0] + q->in_flight[1];
diff --git a/include/scsi/scsi_device.h b/include/scsi/scsi_device.h
index f6820da1dc37..666b58799cec 100644
--- a/include/scsi/scsi_device.h
+++ b/include/scsi/scsi_device.h
@@ -226,7 +226,6 @@ struct scsi_device {
unsigned char access_state;
struct mutex state_mutex;
enum scsi_device_state sdev_state;
- struct task_struct *quiesced_by;
atomic_t nr_admin_pending;
wait_queue_head_t admin_wq;
--
2.9.5
next prev parent reply other threads:[~2018-08-11 7:12 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-08-11 7:12 [RFC PATCH V2 00/17] SCSI: introduce per-host admin queue & enable runtime PM Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 01/17] blk-mq: allow to pass default queue flags for creating & initializing queue Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 02/17] blk-mq: convert BLK_MQ_F_NO_SCHED into per-queue flag Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 03/17] block: rename QUEUE_FLAG_NO_SCHED as QUEUE_FLAG_ADMIN Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 04/17] blk-mq: don't reserve tags for admin queue Ming Lei
2018-08-13 10:02 ` jianchao.wang
2018-08-13 10:48 ` Ming Lei
2018-08-14 1:29 ` jianchao.wang
2018-08-14 2:10 ` Ming Lei
2018-08-14 2:47 ` jianchao.wang
2018-08-14 3:06 ` Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 05/17] SCSI: try to retrieve request_queue via 'scsi_cmnd' if possible Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 06/17] SCSI: pass 'scsi_device' instance from 'scsi_request' Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 07/17] SCSI: prepare for introducing admin queue for legacy path Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 08/17] SCSI: pass scsi_device to scsi_mq_prep_fn Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 09/17] SCSI: don't set .queuedata in scsi_mq_alloc_queue() Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 10/17] SCSI: deal with admin queue busy Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 11/17] SCSI: track pending admin commands Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 12/17] SCSI: create admin queue for each host Ming Lei
2018-08-14 5:56 ` jianchao.wang
2018-08-14 6:03 ` jianchao.wang
2018-08-14 11:34 ` Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 13/17] SCSI: use the dedicated admin queue to send admin commands Ming Lei
2018-08-11 7:12 ` [RFC PATCH V2 14/17] SCSI: transport_spi: resume a quiesced device Ming Lei
2018-08-11 7:12 ` Ming Lei [this message]
2018-08-11 7:12 ` [RFC PATCH V2 16/17] block: simplify runtime PM support Ming Lei
2018-08-15 6:39 ` jianchao.wang
2018-08-15 8:28 ` Ming Lei
2018-08-15 9:47 ` jianchao.wang
2018-08-15 11:23 ` Ming Lei
2018-08-16 8:26 ` jianchao.wang
2018-08-11 7:12 ` [RFC PATCH V2 17/17] block: enable runtime PM for blk-mq Ming Lei
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180811071220.357-16-ming.lei@redhat.com \
--to=ming.lei@redhat.com \
--cc=adrian.hunter@intel.com \
--cc=axboe@kernel.dk \
--cc=bart.vanassche@wdc.com \
--cc=hare@suse.de \
--cc=hch@lst.de \
--cc=jejb@linux.vnet.ibm.com \
--cc=jianchao.w.wang@oracle.com \
--cc=jthumshirn@suse.de \
--cc=linux-block@vger.kernel.org \
--cc=linux-scsi@vger.kernel.org \
--cc=martin.petersen@oracle.com \
--cc=stern@rowland.harvard.edu \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).