Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [802.3_MAINT] summary of MAC interface issues



 
Brad:

I guess we are talking past each other.  

If I understand the ISO definition of service interface primitives
correctly, there can at most be one MA_DATA.request outstanding and
another can't be issued until the frame resulting from the first request
is transmitted, transmission started and transmission process "aborted"
(e.g., half duplex transmission attempts exhausted).  If this is a
correct understanding of service primitives, then I don't think the
problem between the state diagram and Pascal can exist.  

That said, I'm still inclined to support a "fix" anyway.  I've found
little consensus on how primitives are allowed to work, and consequently
no consensus.  We can avoid a future repeat of this kind of discussion
if our MAC specification implements the abstract characteristics of
primitives.  To do that we should (again) consult service interface
experts in 802.1 as Pat suggested earlier.  

If my understanding of service primitives is right, any change to the
Pascal or state diagram that somehow delays looking for another
MA_DATA.request until the frame from the previous MA_DATA.request is
"transmitted" would I think eliminate the need to understand all the
subtleties of service interface primitives to correctly implement 802.3.


-- Bob

[thread snipped]