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

[Ming Lei]

On Mon, Jul 27, 2020 at 02:00:15PM -0700, Sagi Grimberg wrote:
> 
> > > > > > > +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.
> > > > > 
> > > > > That's true, but I want a single interface for both. v2 had exactly
> > > > > that, but I decided that this approach is better.
> > > > 
> > > > Not sure one new interface is needed, and one simple way is to:
> > > > 
> > > > 1) call blk_mq_quiesce_queue_nowait() for each request queue
> > > > 
> > > > 2) wait in driver specific way
> > > > 
> > > > Or just wondering why nvme doesn't use set->tag_list to retrieve NS,
> > > > then you may add per-tagset APIs for the waiting.
> > > 
> > > Because it puts assumptions on how quiesce works, which is something
> > > I'd like to avoid because I think its cleaner, what do others think?
> > > Jens? Christoph?
> > 
> > I'd prefer to have it in a helper, and just have blk_mq_quiesce_queue()
> > call that.
> 
> I agree with this approach as well.
> 
> Jens, this mean that we use the call_rcu mechanism also for non-blocking
> hctxs, because the caller  will call it for multiple request queues (see
> patch 2) and we don't want to call synchronize_rcu for every request
> queue serially, we want it to happen in parallel.
> 
> Which leaves us with the patchset as it is, just to convert the
> rcu_synchronize structure to be dynamically allocated on the heap
> rather than keeping it statically allocated in the hctx.
> 
> This is how it looks:
> --
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index abcf590f6238..d913924117d2 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -209,6 +209,52 @@ 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;
> +       int rcu = false;
> +
> +       blk_mq_quiesce_queue_nowait(q);
> +
> +       queue_for_each_hw_ctx(q, hctx, i) {
> +               hctx->rcu_sync = kmalloc(sizeof(*hctx->rcu_sync),
> GFP_KERNEL);
> +               if (!hctx->rcu_sync) {
> +                       /* fallback to serial rcu sync */
> +                       if (hctx->flags & BLK_MQ_F_BLOCKING)
> +                               synchronize_srcu(hctx->srcu);
> +                       else
> +                               rcu = true;
> +               } else {
> +                       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);
> +               }
> +       }
> +       if (rcu)
> +               synchronize_rcu();
> +}
> +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) {
> +               if (!hctx->rcu_sync)
> +                       continue;
> +               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..7213ce56bb31 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 pointer needn't to be added to blk_mq_hw_ctx, and it can be
allocated on heap and passed to the waiting helper.


Thanks,
Ming