Intent of statement
2007-05-07 22:51:58 GMT
Page 36 of draft 07 or RDMAP states that:
HdrCt: 3 bits.
Header control bits:
* M: bit 16. DDP Segment Length valid. See Figure 10 for when this bit SHOULD be set.
* D: bit 17. DDP Header Included. See Figure 10 for when this bit SHOULD be set.
* R: bit 18. RDMAP Header Included. See Figure 10 for when this bit SHOULD be set.
The SHOULD statements in this text indicate that these items are optional. However, in reading the text it appears that the intent may have been that these statements are just incorrectly capitalized “should” statements and best understood as:
HdrCt: 3 bits.
Header control bits:
* M: bit 16. DDP Segment Length valid. See Figure 10 for when to set this bit.
* D: bit 17. DDP Header Included. See Figure 10 for when to set this bit.
* R: bit 18. RDMAP Header Included. See Figure 10 for when to set this bit.
The implication of this would be, for instance, that an implementation MUST set the R bit and return the received RDMA header when generating a terminate for a read request that generated a remote protection error due to an invalid Stag.
A literal reading of the current wording (with SHOULD) leaves a somewhat conflicted state in which figure 10 states that the RDMA header is required but that it is not necessary to set the R bit.
Unless instructed otherwise I am going to assume that the SHOULD in HdrCt is not really normative.
Hope someone is still reading this reflector….
Barry Reinhold
(603) 868-8411
_______________________________________________ rddp mailing list rddp <at> ietf.org https://www1.ietf.org/mailman/listinfo/rddp
RSS Feed