[openib-general] Re: Re: Re: [PATCH] change Mellanox SDP workaround toa moduleparameter

[Sean Hefty]

Hal Rosenstock wrote:
>>>>>>So the only issue here is the inefficiency in terms of the back and
>>>>>>forth of CM messages to get to the 1K MTU connection. How important is
>>>>>>connection rate for SDP and SRP ? If not, can't we live with how things
>>>>>>are ?
>>>>>
>>>>>SDP implements sockets, and thats a very wide field, so everything is important.
>>>>>AFAIK, connection rate is very important for some socket applications, and does
>>>>>not matter at all for others.
>>>>
>>>>It sounds like better CM handling is important for connection rate.
>>>
>>>Right.
>>
>>I didn't follow what changes are necessary to improve the connection rate.  Can 
>>you clarify?
> 
> 
> Changes in the CM protocol so that a 3 way handshake can occur so that
> for example a lower MTU connection can be accepted with one REQ/REP/RTU
> sequence without REQ/REJ then REQ/REP/RTU.

Okay - this is what I thought.  Personally, I don't think that it makes sense to 
change the architecture to optimize for hardware that happens to perform poorly 
under a given condition.

What would happen if the hardware simply reported that it only supported a 
smaller MTU?

- Sean