How to handle if identical GET_CAPABILITY received for two or more times in retry?

[aalokshah-git]

According to point number 890 in 1.2.1. It goes like below,

"A Requester can retry messages. The retries shall be identical to the first, excluding transport variances. However, if the Responder sees two or more non-identical GET_CAPABILITIES or NEGOTIATE_ALGORITHMS , the Responder shall return an ERROR message with ErrorCode=UnexpectedRequest or silently discard non-identical GET_CAPABILITIES or NEGOTIATE_ALGORITHMS requests. Because a retried message is identical to the first, a retried message shall not be used in transcript hash calculations."

It seems libSpdm is not taking care of it properly, and to match this requirement it requires change in Connection State Machine, transcript handling, and also to cache last request explicitly so that it can be compared to next message coming.

Is my understanding correct??

[steven-bellock]

See #2943. But yes, we should probably document libspdm's retry policy.