Re: [PATCH v3 1/2] blk-mq: add async quiesce interface

[Ming Lei <ming.lei@redhat.com>]

Hi Sagi,

On Sat, Jul 25, 2020 at 05:23:00PM -0700, Sagi Grimberg wrote:
> Drivers that may have to quiesce a large amount of request queues at once
> (e.g. controller or adapter reset). These drivers would benefit from an
> async quiesce interface such that the can trigger quiesce asynchronously
> and wait for all in parallel.
> 
> This leaves the synchronization responsibility to the driver, but adds
> a convenient interface to quiesce async and wait in a single pass.
> 
> Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
> ---
>  block/blk-mq.c         | 32 ++++++++++++++++++++++++++++++++
>  include/linux/blk-mq.h |  4 ++++
>  2 files changed, 36 insertions(+)
> 
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index abcf590f6238..60d137265bd9 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -209,6 +209,38 @@ void blk_mq_quiesce_queue_nowait(struct request_queue *q)
>  }
>  EXPORT_SYMBOL_GPL(blk_mq_quiesce_queue_nowait);
>  
> +void blk_mq_quiesce_queue_async(struct request_queue *q)
> +{
> +	struct blk_mq_hw_ctx *hctx;
> +	unsigned int i;
> +
> +	blk_mq_quiesce_queue_nowait(q);
> +
> +	queue_for_each_hw_ctx(q, hctx, i) {
> +		init_completion(&hctx->rcu_sync.completion);
> +		init_rcu_head(&hctx->rcu_sync.head);
> +		if (hctx->flags & BLK_MQ_F_BLOCKING)
> +			call_srcu(hctx->srcu, &hctx->rcu_sync.head,
> +				wakeme_after_rcu);
> +		else
> +			call_rcu(&hctx->rcu_sync.head,
> +				wakeme_after_rcu);
> +	}

Looks not necessary to do anything in case of !BLK_MQ_F_BLOCKING, and single
synchronize_rcu() is OK for all hctx during waiting.

> +}
> +EXPORT_SYMBOL_GPL(blk_mq_quiesce_queue_async);
> +
> +void blk_mq_quiesce_queue_async_wait(struct request_queue *q)
> +{
> +	struct blk_mq_hw_ctx *hctx;
> +	unsigned int i;
> +
> +	queue_for_each_hw_ctx(q, hctx, i) {
> +		wait_for_completion(&hctx->rcu_sync.completion);
> +		destroy_rcu_head(&hctx->rcu_sync.head);
> +	}
> +}
> +EXPORT_SYMBOL_GPL(blk_mq_quiesce_queue_async_wait);
> +
>  /**
>   * blk_mq_quiesce_queue() - wait until all ongoing dispatches have finished
>   * @q: request queue.
> diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h
> index 23230c1d031e..5536e434311a 100644
> --- a/include/linux/blk-mq.h
> +++ b/include/linux/blk-mq.h
> @@ -5,6 +5,7 @@
>  #include <linux/blkdev.h>
>  #include <linux/sbitmap.h>
>  #include <linux/srcu.h>
> +#include <linux/rcupdate_wait.h>
>  
>  struct blk_mq_tags;
>  struct blk_flush_queue;
> @@ -170,6 +171,7 @@ struct blk_mq_hw_ctx {
>  	 */
>  	struct list_head	hctx_list;
>  
> +	struct rcu_synchronize	rcu_sync;
 
The above struct takes at least 5 words, and I'd suggest to avoid it,
and the hctx->srcu should be re-used for waiting BLK_MQ_F_BLOCKING.
Meantime !BLK_MQ_F_BLOCKING doesn't need it.


Thanks,
Ming