diff mbox series

[v5,2/2] dm: support bio polling

Message ID 20220305020804.54010-3-snitzer@redhat.com (mailing list archive)
State New, archived
Headers show
Series block/dm: support bio polling | expand

Commit Message

Mike Snitzer March 5, 2022, 2:08 a.m. UTC
From: Ming Lei <ming.lei@redhat.com>

Support bio(REQ_POLLED) polling in the following approach:

1) only support io polling on normal READ/WRITE, and other abnormal IOs
still fallback to IRQ mode, so the target io is exactly inside the dm
io.

2) hold one refcnt on io->io_count after submitting this dm bio with
REQ_POLLED

3) support dm native bio splitting, any dm io instance associated with
current bio will be added into one list which head is bio->bi_end_io
which will be recovered before ending this bio

4) implement .poll_bio() callback, call bio_poll() on the single target
bio inside the dm io which is retrieved via bio->bi_bio_drv_data; call
dm_io_dec_pending() after the target io is done in .poll_bio()

5) enable QUEUE_FLAG_POLL if all underlying queues enable QUEUE_FLAG_POLL,
which is based on Jeffle's previous patch.

Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
---
 drivers/md/dm-core.h  |   2 +
 drivers/md/dm-table.c |  27 +++++++++
 drivers/md/dm.c       | 150 +++++++++++++++++++++++++++++++++++++++++++++++++-
 3 files changed, 176 insertions(+), 3 deletions(-)

Comments

Christoph Hellwig March 6, 2022, 9:29 a.m. UTC | #1
> +/*
> + * Reuse ->bi_end_io as hlist head for storing all dm_io instances
> + * associated with this bio, and this bio's bi_end_io has to be
> + * stored in one of 'dm_io' instance first.
> + */
> +static inline struct hlist_head *dm_get_bio_hlist_head(struct bio *bio)
> +{
> +	WARN_ON_ONCE(!(bio->bi_opf & REQ_DM_POLL_LIST));
> +
> +	return (struct hlist_head *)&bio->bi_end_io;
> +}

So this reuse is what I really hated.  I still think we should be able
to find space in the bio by creatively shifting fields around to just
add the hlist there directly, which would remove the need for this
override and more importantly the quite cumbersome saving and restoring
of the end_io handler.
Jens Axboe March 7, 2022, 1:48 a.m. UTC | #2
On 3/6/22 2:29 AM, Christoph Hellwig wrote:
>> +/*
>> + * Reuse ->bi_end_io as hlist head for storing all dm_io instances
>> + * associated with this bio, and this bio's bi_end_io has to be
>> + * stored in one of 'dm_io' instance first.
>> + */
>> +static inline struct hlist_head *dm_get_bio_hlist_head(struct bio *bio)
>> +{
>> +	WARN_ON_ONCE(!(bio->bi_opf & REQ_DM_POLL_LIST));
>> +
>> +	return (struct hlist_head *)&bio->bi_end_io;
>> +}
> 
> So this reuse is what I really hated.  I still think we should be able
> to find space in the bio by creatively shifting fields around to just
> add the hlist there directly, which would remove the need for this
> override and more importantly the quite cumbersome saving and restoring
> of the end_io handler.

If it's possible, then that would be preferable. But I don't think
that's going to be easy to do...
Ming Lei March 7, 2022, 2:20 a.m. UTC | #3
On Sun, Mar 06, 2022 at 06:48:15PM -0700, Jens Axboe wrote:
> On 3/6/22 2:29 AM, Christoph Hellwig wrote:
> >> +/*
> >> + * Reuse ->bi_end_io as hlist head for storing all dm_io instances
> >> + * associated with this bio, and this bio's bi_end_io has to be
> >> + * stored in one of 'dm_io' instance first.
> >> + */
> >> +static inline struct hlist_head *dm_get_bio_hlist_head(struct bio *bio)
> >> +{
> >> +	WARN_ON_ONCE(!(bio->bi_opf & REQ_DM_POLL_LIST));
> >> +
> >> +	return (struct hlist_head *)&bio->bi_end_io;
> >> +}
> > 
> > So this reuse is what I really hated.  I still think we should be able
> > to find space in the bio by creatively shifting fields around to just
> > add the hlist there directly, which would remove the need for this
> > override and more importantly the quite cumbersome saving and restoring
> > of the end_io handler.
> 
> If it's possible, then that would be preferable. But I don't think
> that's going to be easy to do...

I agree, now basically there isn't gap inside bio, so either adding one
new field or reusing one existed field...


Thanks,
Ming
Jens Axboe March 7, 2022, 2:25 a.m. UTC | #4
On 3/6/22 7:20 PM, Ming Lei wrote:
> On Sun, Mar 06, 2022 at 06:48:15PM -0700, Jens Axboe wrote:
>> On 3/6/22 2:29 AM, Christoph Hellwig wrote:
>>>> +/*
>>>> + * Reuse ->bi_end_io as hlist head for storing all dm_io instances
>>>> + * associated with this bio, and this bio's bi_end_io has to be
>>>> + * stored in one of 'dm_io' instance first.
>>>> + */
>>>> +static inline struct hlist_head *dm_get_bio_hlist_head(struct bio *bio)
>>>> +{
>>>> +	WARN_ON_ONCE(!(bio->bi_opf & REQ_DM_POLL_LIST));
>>>> +
>>>> +	return (struct hlist_head *)&bio->bi_end_io;
>>>> +}
>>>
>>> So this reuse is what I really hated.  I still think we should be able
>>> to find space in the bio by creatively shifting fields around to just
>>> add the hlist there directly, which would remove the need for this
>>> override and more importantly the quite cumbersome saving and restoring
>>> of the end_io handler.
>>
>> If it's possible, then that would be preferable. But I don't think
>> that's going to be easy to do...
> 
> I agree, now basically there isn't gap inside bio, so either adding one
> new field or reusing one existed field...

There'd no amount of re-arranging that'll free up 8 bytes, that's just
not happening. I'm not a huge fan of growing struct bio for that, and
the oddity here is mostly (to me) that ->bi_end_io is the one overlayed.
That would usually belong to the owner of the bio.

Maybe some commenting would help? Is bi_next available at this point?
Ming Lei March 7, 2022, 2:41 a.m. UTC | #5
On Sun, Mar 06, 2022 at 07:25:11PM -0700, Jens Axboe wrote:
> On 3/6/22 7:20 PM, Ming Lei wrote:
> > On Sun, Mar 06, 2022 at 06:48:15PM -0700, Jens Axboe wrote:
> >> On 3/6/22 2:29 AM, Christoph Hellwig wrote:
> >>>> +/*
> >>>> + * Reuse ->bi_end_io as hlist head for storing all dm_io instances
> >>>> + * associated with this bio, and this bio's bi_end_io has to be
> >>>> + * stored in one of 'dm_io' instance first.
> >>>> + */
> >>>> +static inline struct hlist_head *dm_get_bio_hlist_head(struct bio *bio)
> >>>> +{
> >>>> +	WARN_ON_ONCE(!(bio->bi_opf & REQ_DM_POLL_LIST));
> >>>> +
> >>>> +	return (struct hlist_head *)&bio->bi_end_io;
> >>>> +}
> >>>
> >>> So this reuse is what I really hated.  I still think we should be able
> >>> to find space in the bio by creatively shifting fields around to just
> >>> add the hlist there directly, which would remove the need for this
> >>> override and more importantly the quite cumbersome saving and restoring
> >>> of the end_io handler.
> >>
> >> If it's possible, then that would be preferable. But I don't think
> >> that's going to be easy to do...
> > 
> > I agree, now basically there isn't gap inside bio, so either adding one
> > new field or reusing one existed field...
> 
> There'd no amount of re-arranging that'll free up 8 bytes, that's just
> not happening. I'm not a huge fan of growing struct bio for that, and
> the oddity here is mostly (to me) that ->bi_end_io is the one overlayed.
> That would usually belong to the owner of the bio.
> 
> Maybe some commenting would help?

OK, ->bi_end_io is safe because it is only called until the bio is
ended, so we can retrieve the list head and recover ->bi_end_io before
polling.

> Is bi_next available at this point?

The same bio can be re-submitted to block layer because of splitting, and
will be linked to current->bio_list[].

BTW, bio splitting can be very often for some dm target, that is why we
don't ignore bio splitting for dm polling.


Thanks,
Ming
Ming Lei March 7, 2022, 3:39 a.m. UTC | #6
On Mon, Mar 07, 2022 at 10:41:31AM +0800, Ming Lei wrote:
> On Sun, Mar 06, 2022 at 07:25:11PM -0700, Jens Axboe wrote:
> > On 3/6/22 7:20 PM, Ming Lei wrote:
> > > On Sun, Mar 06, 2022 at 06:48:15PM -0700, Jens Axboe wrote:
> > >> On 3/6/22 2:29 AM, Christoph Hellwig wrote:
> > >>>> +/*
> > >>>> + * Reuse ->bi_end_io as hlist head for storing all dm_io instances
> > >>>> + * associated with this bio, and this bio's bi_end_io has to be
> > >>>> + * stored in one of 'dm_io' instance first.
> > >>>> + */
> > >>>> +static inline struct hlist_head *dm_get_bio_hlist_head(struct bio *bio)
> > >>>> +{
> > >>>> +	WARN_ON_ONCE(!(bio->bi_opf & REQ_DM_POLL_LIST));
> > >>>> +
> > >>>> +	return (struct hlist_head *)&bio->bi_end_io;
> > >>>> +}
> > >>>
> > >>> So this reuse is what I really hated.  I still think we should be able
> > >>> to find space in the bio by creatively shifting fields around to just
> > >>> add the hlist there directly, which would remove the need for this
> > >>> override and more importantly the quite cumbersome saving and restoring
> > >>> of the end_io handler.
> > >>
> > >> If it's possible, then that would be preferable. But I don't think
> > >> that's going to be easy to do...
> > > 
> > > I agree, now basically there isn't gap inside bio, so either adding one
> > > new field or reusing one existed field...
> > 
> > There'd no amount of re-arranging that'll free up 8 bytes, that's just
> > not happening. I'm not a huge fan of growing struct bio for that, and
> > the oddity here is mostly (to me) that ->bi_end_io is the one overlayed.
> > That would usually belong to the owner of the bio.
> > 
> > Maybe some commenting would help?
> 
> OK, ->bi_end_io is safe because it is only called until the bio is
> ended, so we can retrieve the list head and recover ->bi_end_io before
> polling.

->bi_private can be reused too, is that better?

Yeah, both belong to owner(higher level storage), then block layer can't touch
them inside submit_bio_noacct(), that is also why this trick is safe.

Thanks,
Ming
diff mbox series

Patch

diff --git a/drivers/md/dm-core.h b/drivers/md/dm-core.h
index 8078b6c155ef..b3d1429fba83 100644
--- a/drivers/md/dm-core.h
+++ b/drivers/md/dm-core.h
@@ -235,6 +235,8 @@  struct dm_io {
 	bool start_io_acct:1;
 	int was_accounted;
 	unsigned long start_time;
+	void *saved_bio_end_io;
+	struct hlist_node node;
 	spinlock_t endio_lock;
 	struct dm_stats_aux stats_aux;
 	/* last member of dm_target_io is 'struct bio' */
diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index f4ed756ab391..c0be4f60b427 100644
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -1481,6 +1481,14 @@  struct dm_target *dm_table_find_target(struct dm_table *t, sector_t sector)
 	return &t->targets[(KEYS_PER_NODE * n) + k];
 }
 
+static int device_not_poll_capable(struct dm_target *ti, struct dm_dev *dev,
+				   sector_t start, sector_t len, void *data)
+{
+	struct request_queue *q = bdev_get_queue(dev->bdev);
+
+	return !test_bit(QUEUE_FLAG_POLL, &q->queue_flags);
+}
+
 /*
  * type->iterate_devices() should be called when the sanity check needs to
  * iterate and check all underlying data devices. iterate_devices() will
@@ -1531,6 +1539,11 @@  static int count_device(struct dm_target *ti, struct dm_dev *dev,
 	return 0;
 }
 
+static int dm_table_supports_poll(struct dm_table *t)
+{
+	return !dm_table_any_dev_attr(t, device_not_poll_capable, NULL);
+}
+
 /*
  * Check whether a table has no data devices attached using each
  * target's iterate_devices method.
@@ -2067,6 +2080,20 @@  int dm_table_set_restrictions(struct dm_table *t, struct request_queue *q,
 	dm_update_crypto_profile(q, t);
 	disk_update_readahead(t->md->disk);
 
+	/*
+	 * Check for request-based device is left to
+	 * dm_mq_init_request_queue()->blk_mq_init_allocated_queue().
+	 *
+	 * For bio-based device, only set QUEUE_FLAG_POLL when all
+	 * underlying devices supporting polling.
+	 */
+	if (__table_type_bio_based(t->type)) {
+		if (dm_table_supports_poll(t))
+			blk_queue_flag_set(QUEUE_FLAG_POLL, q);
+		else
+			blk_queue_flag_clear(QUEUE_FLAG_POLL, q);
+	}
+
 	return 0;
 }
 
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 454d39bc7745..c28d453e9930 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -40,6 +40,13 @@ 
 #define DM_COOKIE_ENV_VAR_NAME "DM_COOKIE"
 #define DM_COOKIE_LENGTH 24
 
+/*
+ * For REQ_POLLED fs bio, this flag is set if we link mapped underlying
+ * dm_io into one list, and reuse bio->bi_end_io as the list head. Before
+ * ending this fs bio, we will recover its ->bi_end_io callback.
+ */
+#define REQ_DM_POLL_LIST	REQ_DRV
+
 static const char *_name = DM_NAME;
 
 static unsigned int major = 0;
@@ -73,6 +80,7 @@  struct clone_info {
 	struct dm_io *io;
 	sector_t sector;
 	unsigned sector_count;
+	bool submit_as_polled;
 };
 
 #define DM_TARGET_IO_BIO_OFFSET (offsetof(struct dm_target_io, clone))
@@ -599,6 +607,9 @@  static struct bio *alloc_tio(struct clone_info *ci, struct dm_target *ti,
 		if (!clone)
 			return NULL;
 
+		/* REQ_DM_POLL_LIST shouldn't be inherited */
+		clone->bi_opf &= ~REQ_DM_POLL_LIST;
+
 		tio = clone_to_tio(clone);
 		tio->inside_dm_io = false;
 	}
@@ -888,8 +899,15 @@  void dm_io_dec_pending(struct dm_io *io, blk_status_t error)
 		if (unlikely(wq_has_sleeper(&md->wait)))
 			wake_up(&md->wait);
 
-		if (io_error == BLK_STS_DM_REQUEUE)
+		if (io_error == BLK_STS_DM_REQUEUE) {
+			/*
+			 * Upper layer won't help us poll split bio, io->orig_bio
+			 * may only reflect a subset of the pre-split original,
+			 * so clear REQ_POLLED in case of requeue
+			 */
+			bio->bi_opf &= ~REQ_POLLED;
 			return;
+		}
 
 		if (bio_is_flush_with_data(bio)) {
 			/*
@@ -1440,6 +1458,42 @@  static bool __process_abnormal_io(struct clone_info *ci, struct dm_target *ti,
 	return true;
 }
 
+/*
+ * Reuse ->bi_end_io as hlist head for storing all dm_io instances
+ * associated with this bio, and this bio's bi_end_io has to be
+ * stored in one of 'dm_io' instance first.
+ */
+static inline struct hlist_head *dm_get_bio_hlist_head(struct bio *bio)
+{
+	WARN_ON_ONCE(!(bio->bi_opf & REQ_DM_POLL_LIST));
+
+	return (struct hlist_head *)&bio->bi_end_io;
+}
+
+static void dm_queue_poll_io(struct bio *bio, struct dm_io *io)
+{
+	if (!(bio->bi_opf & REQ_DM_POLL_LIST)) {
+		bio->bi_opf |= REQ_DM_POLL_LIST;
+		/*
+		 * Save .bi_end_io into dm_io, so that we can reuse
+		 * .bi_end_io for storing dm_io list
+		 */
+		io->saved_bio_end_io = bio->bi_end_io;
+
+		INIT_HLIST_HEAD(dm_get_bio_hlist_head(bio));
+
+		/* tell block layer to poll for completion */
+		bio->bi_cookie = ~BLK_QC_T_NONE;
+	} else {
+		/*
+		 * bio recursed due to split, reuse original poll list
+		 */
+		io->saved_bio_end_io = NULL;
+	}
+
+	hlist_add_head(&io->node, dm_get_bio_hlist_head(bio));
+}
+
 /*
  * Select the correct strategy for processing a non-flush bio.
  */
@@ -1457,6 +1511,12 @@  static int __split_and_process_bio(struct clone_info *ci)
 	if (__process_abnormal_io(ci, ti, &r))
 		return r;
 
+	/*
+	 * Only support bio polling for normal IO, and the target io is
+	 * exactly inside the dm_io instance (verified in dm_poll_dm_io)
+	 */
+	ci->submit_as_polled = ci->bio->bi_opf & REQ_POLLED;
+
 	len = min_t(sector_t, max_io_len(ti, ci->sector), ci->sector_count);
 	clone = alloc_tio(ci, ti, 0, &len, GFP_NOIO);
 	__map_bio(clone);
@@ -1473,6 +1533,7 @@  static void init_clone_info(struct clone_info *ci, struct mapped_device *md,
 	ci->map = map;
 	ci->io = alloc_io(md, bio);
 	ci->bio = bio;
+	ci->submit_as_polled = false;
 	ci->sector = bio->bi_iter.bi_sector;
 	ci->sector_count = bio_sectors(bio);
 
@@ -1522,8 +1583,17 @@  static void dm_split_and_process_bio(struct mapped_device *md,
 	if (ci.io->start_io_acct)
 		dm_start_io_acct(ci.io, NULL);
 
-	/* drop the extra reference count */
-	dm_io_dec_pending(ci.io, errno_to_blk_status(error));
+	/*
+	 * Drop the extra reference count for non-POLLED bio, and hold one
+	 * reference for POLLED bio, which will be released in dm_poll_bio
+	 *
+	 * Add every dm_io instance into the hlist_head which is stored in
+	 * bio->bi_end_io, so that dm_poll_bio can poll them all.
+	 */
+	if (error || !ci.submit_as_polled)
+		dm_io_dec_pending(ci.io, errno_to_blk_status(error));
+	else
+		dm_queue_poll_io(bio, ci.io);
 }
 
 static void dm_submit_bio(struct bio *bio)
@@ -1558,6 +1628,79 @@  static void dm_submit_bio(struct bio *bio)
 	dm_put_live_table(md, srcu_idx);
 }
 
+static bool dm_poll_dm_io(struct dm_io *io, struct io_comp_batch *iob,
+			  unsigned int flags)
+{
+	WARN_ON_ONCE(!io->tio.inside_dm_io);
+
+	/* don't poll if the mapped io is done */
+	if (atomic_read(&io->io_count) > 1)
+		bio_poll(&io->tio.clone, iob, flags);
+
+	/* bio_poll holds the last reference */
+	return atomic_read(&io->io_count) == 1;
+}
+
+static int dm_poll_bio(struct bio *bio, struct io_comp_batch *iob,
+		       unsigned int flags)
+{
+	struct hlist_head *head = dm_get_bio_hlist_head(bio);
+	struct hlist_head tmp = HLIST_HEAD_INIT;
+	void *saved_bio_end_io = NULL;
+	struct hlist_node *next;
+	struct dm_io *io;
+
+	/* Only poll normal bio which was marked as REQ_DM_POLL_LIST */
+	if (!(bio->bi_opf & REQ_DM_POLL_LIST))
+		return 0;
+
+	WARN_ON_ONCE(hlist_empty(head));
+
+	hlist_move_list(head, &tmp);
+
+	hlist_for_each_entry(io, &tmp, node) {
+		if (io->saved_bio_end_io) {
+			saved_bio_end_io = io->saved_bio_end_io;
+			break;
+		}
+	}
+
+	/*
+	 * Restore .bi_end_io before possibly completing dm_io.
+	 *
+	 * bio_poll() is only possible once @bio has been completely
+	 * submitted via submit_bio_noacct()'s depth-first submission.
+	 * So there is no dm_queue_poll_io() race associated with
+	 * clearing REQ_DM_POLL_LIST here.
+	 */
+	WARN_ON_ONCE(!saved_bio_end_io);
+	bio->bi_opf &= ~REQ_DM_POLL_LIST;
+	bio->bi_end_io = saved_bio_end_io;
+
+	hlist_for_each_entry_safe(io, next, &tmp, node) {
+		if (dm_poll_dm_io(io, iob, flags)) {
+			hlist_del_init(&io->node);
+			/*
+			 * clone_endio() has already occurred, so passing
+			 * error as 0 here doesn't override io->status
+			 */
+			dm_io_dec_pending(io, 0);
+		}
+	}
+
+	/* Not done? */
+	if (!hlist_empty(&tmp)) {
+		/* Store saved_bio_end_io in a remaining dm_io */
+		io = hlist_entry(tmp.first, struct dm_io, node);
+		io->saved_bio_end_io = saved_bio_end_io;
+		bio->bi_opf |= REQ_DM_POLL_LIST;
+		/* Reset bio->bi_end_io to dm_io list head */
+		hlist_move_list(&tmp, head);
+		return 0;
+	}
+	return 1;
+}
+
 /*-----------------------------------------------------------------
  * An IDR is used to keep track of allocated minor numbers.
  *---------------------------------------------------------------*/
@@ -2983,6 +3126,7 @@  static const struct pr_ops dm_pr_ops = {
 
 static const struct block_device_operations dm_blk_dops = {
 	.submit_bio = dm_submit_bio,
+	.poll_bio = dm_poll_bio,
 	.open = dm_blk_open,
 	.release = dm_blk_close,
 	.ioctl = dm_blk_ioctl,