[PATCH] nvmet_fc: Reduce work_q count

[James Smart]

On 10/1/2017 1:39 AM, Christoph Hellwig wrote:
> On Sat, Sep 30, 2017 at 11:32:39AM -0700, James Smart wrote:
>> Because we're still trying to get cpu affinity on the connection resources
>> that will be shared by the io on the same queue.
> 
> And that's exactly what plain queue_work will give you.  If it doesn't
> do exactly what your open code logic does it has a bug that needs to
> be fixed.
> 
I don't understand how queuing work alone, which can go to any cpu 
(current preferred) helps with avoiding cross-cpu access to the same 
transport/lldd hw queue data structures.

But, as stated. A moot point now. I'll remove this whole IN_ISR, thus 
work_q item, shortly.

-- james