[openib-general] tavor quirks etc (opensm compliance etc)

[Michael S. Tsirkin]

> I am still digesting your response where you have addressed my 
> claims/concerns.

Thatnks for raising this issue, I'll continue to think about this. In
particular, the opensm issue that you raise needs to be addressed by the opensm
guys.

> Anyway what is your response to my suggestion of applying just one 
> trivial patch at the rdma cm?

I think this would work too but I somewhat dislike using an MTU that SM
did not give us - this looks like a spec violation to me. No?
For example, it seems this assumes that any path supports 1/2 MTU but is
that required by spec? Further, might SM make an intelligent decision in selecting
a path if we tell it what MTU we actually want to use?

-- 
MST