[openib-general] [PATCH] Add support for querying port width/speed

[Hal Rosenstock]

On Tue, 2004-12-28 at 01:10, Roland Dreier wrote:
> By, the way does anyone know what IPD is supposed to be used when the
> injection port's rate is not a multiple of the rate for the path?  For
> example, suppose an 8X DDR port is sending to a 12X path -- 12 doesn't
> divide 16 evenly.
> 
> I would guess that the correct thing to do is round up, so 8X DDR
> -> 12X uses an IPD of 1, 8X QDR -> 12X uses an IPD of 2, etc.

Yes, rounding up causes the time to wait before scheduling the next
packet to be larger which slows the injection rate (which is what you
would want). 

There is also the recommendation on the HCA that if the IPD value is not
supported then to use the next larger one supported (IBA 1.2 informative
text in paragraph after C9-226). This might result in a slightly larger
slowdown than required for some cases. To do the optimal thing here, you
might need to determine which optional IPDs are supported. However, I'm
not sure if there is a way to determine the optional IPDs supported by
the HCA.

-- Hal