Re: [PATCH v5 2/6] block: wire-up support for passthrough plugging

[Anuj Gupta <anuj20.g@samsung.com>]

[-- Attachment #1: Type: text/plain, Size: 7976 bytes --]

On Thu, May 12, 2022 at 01:25:24PM +0800, Ming Lei wrote:
> Hello,
> 
> On Wed, May 11, 2022 at 11:17:46AM +0530, Kanchan Joshi wrote:
> > From: Jens Axboe <axboe@kernel.dk>
> > 
> > Add support for plugging in passthrough path. When plugging is enabled, the
> > requests are added to a plug instead of getting dispatched to the driver.
> > And when the plug is finished, the whole batch gets dispatched via
> > ->queue_rqs which turns out to be more efficient. Otherwise dispatching
> > used to happen via ->queue_rq, one request at a time.
> > 
> > Signed-off-by: Jens Axboe <axboe@kernel.dk>
> > Reviewed-by: Christoph Hellwig <hch@lst.de>
> > ---
> >  block/blk-mq.c | 73 +++++++++++++++++++++++++++-----------------------
> >  1 file changed, 39 insertions(+), 34 deletions(-)
> > 
> > diff --git a/block/blk-mq.c b/block/blk-mq.c
> > index 84d749511f55..2cf011b57cf9 100644
> > --- a/block/blk-mq.c
> > +++ b/block/blk-mq.c
> > @@ -2340,6 +2340,40 @@ void __blk_mq_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
> >  	blk_mq_hctx_mark_pending(hctx, ctx);
> >  }
> >  
> > +/*
> > + * Allow 2x BLK_MAX_REQUEST_COUNT requests on plug queue for multiple
> > + * queues. This is important for md arrays to benefit from merging
> > + * requests.
> > + */
> > +static inline unsigned short blk_plug_max_rq_count(struct blk_plug *plug)
> > +{
> > +	if (plug->multiple_queues)
> > +		return BLK_MAX_REQUEST_COUNT * 2;
> > +	return BLK_MAX_REQUEST_COUNT;
> > +}
> > +
> > +static void blk_add_rq_to_plug(struct blk_plug *plug, struct request *rq)
> > +{
> > +	struct request *last = rq_list_peek(&plug->mq_list);
> > +
> > +	if (!plug->rq_count) {
> > +		trace_block_plug(rq->q);
> > +	} else if (plug->rq_count >= blk_plug_max_rq_count(plug) ||
> > +		   (!blk_queue_nomerges(rq->q) &&
> > +		    blk_rq_bytes(last) >= BLK_PLUG_FLUSH_SIZE)) {
> > +		blk_mq_flush_plug_list(plug, false);
> > +		trace_block_plug(rq->q);
> > +	}
> > +
> > +	if (!plug->multiple_queues && last && last->q != rq->q)
> > +		plug->multiple_queues = true;
> > +	if (!plug->has_elevator && (rq->rq_flags & RQF_ELV))
> > +		plug->has_elevator = true;
> > +	rq->rq_next = NULL;
> > +	rq_list_add(&plug->mq_list, rq);
> > +	plug->rq_count++;
> > +}
> > +
> >  /**
> >   * blk_mq_request_bypass_insert - Insert a request at dispatch list.
> >   * @rq: Pointer to request to be inserted.
> > @@ -2353,7 +2387,12 @@ void blk_mq_request_bypass_insert(struct request *rq, bool at_head,
> >  				  bool run_queue)
> >  {
> >  	struct blk_mq_hw_ctx *hctx = rq->mq_hctx;
> > +	struct blk_plug *plug = current->plug;
> >  
> > +	if (plug) {
> > +		blk_add_rq_to_plug(plug, rq);
> > +		return;
> > +	}
> 
> This way may cause nested plugging, and breaks xfstests generic/131.
> Also may cause io hang since request can't be polled before flushing
> plug in blk_execute_rq().
>
Hi Ming,
Could you please share your test setup.
I tried test 131 with xfs and it passed.

I followed these steps:
1) mkfs.xfs -f /dev/nvme0n1
2) mount /dev/nvme0n1 /mnt/test
3) ./check tests/generic/131

Tried the same with ext4 and it passed as well.

Thanks,
Anuj

> I'd suggest to apply the plug in blk_execute_rq_nowait(), such as:
> 
> 
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index 2cf011b57cf9..60c29c0229d5 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -1169,6 +1169,62 @@ static void blk_end_sync_rq(struct request *rq, blk_status_t error)
>  	complete(waiting);
>  }
>  
> +/*
> + * Allow 2x BLK_MAX_REQUEST_COUNT requests on plug queue for multiple
> + * queues. This is important for md arrays to benefit from merging
> + * requests.
> + */
> +static inline unsigned short blk_plug_max_rq_count(struct blk_plug *plug)
> +{
> +	if (plug->multiple_queues)
> +		return BLK_MAX_REQUEST_COUNT * 2;
> +	return BLK_MAX_REQUEST_COUNT;
> +}
> +
> +static void blk_add_rq_to_plug(struct blk_plug *plug, struct request *rq)
> +{
> +	struct request *last = rq_list_peek(&plug->mq_list);
> +
> +	if (!plug->rq_count) {
> +		trace_block_plug(rq->q);
> +	} else if (plug->rq_count >= blk_plug_max_rq_count(plug) ||
> +		   (!blk_queue_nomerges(rq->q) &&
> +		    blk_rq_bytes(last) >= BLK_PLUG_FLUSH_SIZE)) {
> +		blk_mq_flush_plug_list(plug, false);
> +		trace_block_plug(rq->q);
> +	}
> +
> +	if (!plug->multiple_queues && last && last->q != rq->q)
> +		plug->multiple_queues = true;
> +	if (!plug->has_elevator && (rq->rq_flags & RQF_ELV))
> +		plug->has_elevator = true;
> +	rq->rq_next = NULL;
> +	rq_list_add(&plug->mq_list, rq);
> +	plug->rq_count++;
> +}
> +
> +static void __blk_execute_rq_nowait(struct request *rq, bool at_head,
> +		rq_end_io_fn *done, bool use_plug)
> +{
> +	WARN_ON(irqs_disabled());
> +	WARN_ON(!blk_rq_is_passthrough(rq));
> +
> +	rq->end_io = done;
> +
> +	blk_account_io_start(rq);
> +
> +	if (use_plug && current->plug) {
> +		blk_add_rq_to_plug(current->plug, rq);
> +		return;
> +	}
> +	/*
> +	 * don't check dying flag for MQ because the request won't
> +	 * be reused after dying flag is set
> +	 */
> +	blk_mq_sched_insert_request(rq, at_head, true, false);
> +}
> +
> +
>  /**
>   * blk_execute_rq_nowait - insert a request to I/O scheduler for execution
>   * @rq:		request to insert
> @@ -1184,18 +1240,8 @@ static void blk_end_sync_rq(struct request *rq, blk_status_t error)
>   */
>  void blk_execute_rq_nowait(struct request *rq, bool at_head, rq_end_io_fn *done)
>  {
> -	WARN_ON(irqs_disabled());
> -	WARN_ON(!blk_rq_is_passthrough(rq));
> -
> -	rq->end_io = done;
> -
> -	blk_account_io_start(rq);
> +	__blk_execute_rq_nowait(rq, at_head, done, true);
>  
> -	/*
> -	 * don't check dying flag for MQ because the request won't
> -	 * be reused after dying flag is set
> -	 */
> -	blk_mq_sched_insert_request(rq, at_head, true, false);
>  }
>  EXPORT_SYMBOL_GPL(blk_execute_rq_nowait);
>  
> @@ -1234,7 +1280,7 @@ blk_status_t blk_execute_rq(struct request *rq, bool at_head)
>  	unsigned long hang_check;
>  
>  	rq->end_io_data = &wait;
> -	blk_execute_rq_nowait(rq, at_head, blk_end_sync_rq);
> +	__blk_execute_rq_nowait(rq, at_head, blk_end_sync_rq, false);
>  
>  	/* Prevent hang_check timer from firing at us during very long I/O */
>  	hang_check = sysctl_hung_task_timeout_secs;
> @@ -2340,40 +2386,6 @@ void __blk_mq_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
>  	blk_mq_hctx_mark_pending(hctx, ctx);
>  }
>  
> -/*
> - * Allow 2x BLK_MAX_REQUEST_COUNT requests on plug queue for multiple
> - * queues. This is important for md arrays to benefit from merging
> - * requests.
> - */
> -static inline unsigned short blk_plug_max_rq_count(struct blk_plug *plug)
> -{
> -	if (plug->multiple_queues)
> -		return BLK_MAX_REQUEST_COUNT * 2;
> -	return BLK_MAX_REQUEST_COUNT;
> -}
> -
> -static void blk_add_rq_to_plug(struct blk_plug *plug, struct request *rq)
> -{
> -	struct request *last = rq_list_peek(&plug->mq_list);
> -
> -	if (!plug->rq_count) {
> -		trace_block_plug(rq->q);
> -	} else if (plug->rq_count >= blk_plug_max_rq_count(plug) ||
> -		   (!blk_queue_nomerges(rq->q) &&
> -		    blk_rq_bytes(last) >= BLK_PLUG_FLUSH_SIZE)) {
> -		blk_mq_flush_plug_list(plug, false);
> -		trace_block_plug(rq->q);
> -	}
> -
> -	if (!plug->multiple_queues && last && last->q != rq->q)
> -		plug->multiple_queues = true;
> -	if (!plug->has_elevator && (rq->rq_flags & RQF_ELV))
> -		plug->has_elevator = true;
> -	rq->rq_next = NULL;
> -	rq_list_add(&plug->mq_list, rq);
> -	plug->rq_count++;
> -}
> -
>  /**
>   * blk_mq_request_bypass_insert - Insert a request at dispatch list.
>   * @rq: Pointer to request to be inserted.
> @@ -2387,12 +2399,7 @@ void blk_mq_request_bypass_insert(struct request *rq, bool at_head,
>  				  bool run_queue)
>  {
>  	struct blk_mq_hw_ctx *hctx = rq->mq_hctx;
> -	struct blk_plug *plug = current->plug;
>  
> -	if (plug) {
> -		blk_add_rq_to_plug(plug, rq);
> -		return;
> -	}
>  	spin_lock(&hctx->lock);
>  	if (at_head)
>  		list_add(&rq->queuelist, &hctx->dispatch);
> 
> 
> Thanks,
> Ming
> 
> 

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]