Re: [rohc] STATIC-NACK and mode transition

[Klaus Warnke <klaus.warnke at acticom.de>]

> Hi, Klaus

Hi Caiwei,

> I am confused with initiating mode transitions of ROHC. As we all
> known, An ACK or NACK feedback packet carrying a CRC could initiate
> mode transitions.

> Why STATIC-NACK couldn't initiate a mode transition?

You refer to the pictures, which are illustrating ACK(O)/NACK(O) to
initiate the mode transfer, but not STATICK-NACK?

I'm not fully sure, but a (3095, p. 44):

  STATIC-NACK : Indicates that the static context of the decompressor
                is not valid or has not been established.

In this case, the compressor could only send a IR packet, because
IR-DYN does not contain the static chain with the static values. But
the figures are allowing IR-DYN packets too. This is maybe a
explanation for that limit.

On the other hand, I think a de-compressor should not mix mode
transition and de-compression error indications, even it is allowed.

From my point of view, a de-compressor should initiate a mode switch,
if the context is ok.

Another hint: Only the dynamic part of the RTP header (p. 102 3095)
contains a mode field. So it is only with the RTP profile possible for
the compressor to use IR/IR-DYN to acknowledge a mode request. All
other profiles must use UOR-2 packets. This design flaw is well known.

> Thank you very much!

br
Klaus