[openib-general] RE: [PATCH 3 of 3] mad: large RMPP support

[Michael S. Tsirkin]

> >The hardware fetches two SGEs anyway, since the QP was created with the 
> >number of send s/g entries = 2. 
> 
> It fetches both of the SGEs as part of the WR, but only requires a 
> single fetch for the data, versus two. 

Sean, I wouldnt worry about that.
Hardware can (and does) pipeline these two reads, so the difference
in speed is not measurable. Try testing it on top of ibverbs and see.
It seems obvious to me that any reasonable implementation will work this way.

-- 
What item should I pick to always win in rock, scissors, paper?