[PATCH V5 0/4] blk-mq: implement queue quiesce via percpu_ref for BLK_MQ_F_BLOCKING

[Ming Lei]

On Fri, Sep 11, 2020 at 01:14:29PM -0600, Jens Axboe wrote:
> On Sep 11, 2020, at 1:12 PM, Sagi Grimberg <sagi at grimberg.me> wrote:
> > 
> > 
> >>> Hi Jens,
> >>> 
> >>> The 1st patch add .mq_quiesce_mutex for serializing quiesce/unquiesce,
> >>> and prepares for replacing srcu with percpu_ref.
> >>> 
> >>> The 2nd patch replaces srcu with percpu_ref.
> >>> 
> >>> The 3rd patch adds tagset quiesce interface.
> >>> 
> >>> The 4th patch applies tagset quiesce interface for NVMe subsystem.
> >> What is this series against?
> > 
> > It didn't apply cleanly to me too until I realized it is
> > on top of v4 of: "percpu_ref & block: reduce memory footprint of percpu_ref in fast path"
> 
> Right, and that’s what has the leak issue you found. I’ll hold off on this one until that’s sorted. 

Actually this patchset doesn't depend on patch of 'percpu_ref & block: reduce memory
footprint of percpu_ref in fast path', and I hold them in one branch, so
causes the conflict.

V6 has been sent out, and just rebase against for-5.10/block directly.

Thanks,
Ming