[PATCH 4.15-rc 2/3] nvme-core/loop/rdma: Host delete_work and reset_work on system workqueues

[Sagi Grimberg]

>> From: Roy Shterman <roys at lightbitslabs.com>
>>
>> We need to ensure that delete_work will be hosted on a different
>> workqueue than all the works we flush or cancel from it.
>> Otherwise we may hit a circular dependency warning [1].
>>
>> Also, given that delete_work flushes reset_work, host reset_work
>> on system_wq and delete_work on system_long_wq. In addition,
>> fix the flushing in the individual drivers to flush system_long_wq
>> when draining queued deletes.
> 
> I vaguely remember pointing something like this out :)

You did :)

> I think we really need our own, separate WQ for this, e.g. a
> nvme_reset_wq as reset progress under an swap / paging load is
> essential and we need it isolated from the system.

I agree, this was the thought when introducing nvme_wq.

So you are fine with keeping controller delete in system_wq
(or system_long_wq)?

> Once we have that patch 1 is probably ok, but the patch adding
> nvme_reset_wq should go before the current patch 1.  Also please
> add good comments on the workqueue usage.

OK, I can do that.