We discussed a second interpretation request in the Interworking meeting, this morning, and came up with another resolution:
	
STANDARD: IEEE Std. 802.1ag-2007

SUBSECTION: 20.31.1 ProcessLBR()

APPLICABLE CONDITIONS:

(Question from Henry Fowler, AT&T, edited for brevity.)

It seems that once any LBR with the incorrect LTID is received, then every LBR received after that will be considered to have an incorrect LTID, until LBIactive becomes false. This is because in action c)2), the received Loopback Transaction Identifier is copied to expectedLBRtransID, but expectedLBRtransID is not incremented after that.

It is possible that that was the intended behavior, that all subsequent LBRs are considered incorrect after one incorrect LBR is received.
However, it is also possible that it was intended for expectedLBRtransID to track the new LTID, so that the first LBR is considered incorrect but subsequent ones are considered correct.

Can you clarify what was intended in ProcessLBR()? Is it as written, or was there a typo in the standard?  (e.g., was intended to say "2) The value from the received Loopback Transaction Identifier field is copied into expecedLBRtransID, then expectedLBRtransID and the number of incorrect LBRs received [item z) in 12.14.7.1.3] is incremented by 1.

RESOLUTION:

The intention of 802.1ag was that if LBRs 1 2 3 5 6 7 8 ... are received, only one error would be counted, for the 3-5 sequence.  If LBRs 1 2 3 5 4 6 7 8 ... are received, then three errors would be counted, one for 3-5, one for 5-4, and one for 4-6. This is not the behavior of the standard as currently written; this will be addressed in a future revision of IEEE Std 802.1Q.