[PATCH 01/14] nvmet: for pt I/O cmds annotate req->sg_cnt likely

[Chaitanya Kulkarni]

On 8/10/20 15:56, Logan Gunthorpe wrote:
>> They are present in the target code. I don't see any problem having it
>> in a passthru code unless it shows severe performance degradation.
> Isn't that backwards? Don't we only add performance optimizations (which
> likely/unlikely are) when they show demonstrable increases in performance?
> 
> Logan
> 

We can always do micro-benchmark, in this case code, frequency and 
conditions are pretty clear and straight forward.