RE: iSCSI: Checking the I bit

[Paul Koning <ni1d@arrl.net>]

>>>>> "Eddy" == Eddy Quicksall <Eddy_Quicksall@ivivity.com> writes:

 Eddy> This has become three issues: 1) The issue of a device checking
 Eddy> to see if the other device is compliant (the "reserved I bit"
 Eddy> happens to fall into this category).  2) The issue of a "1" bit
 Eddy> being called a reserved bit 3) The issue of forcing the "I" bit
 Eddy> to be 1 for all responses

 Eddy> What I am most interested in is issue 1. Because that can be an
 Eddy> unnecessary performance hit (if there are too many things to
 Eddy> check) and a nuisance. This should be an implementation issue.

We had a discussion about this a few weeks ago on the list.  I believe
the outcome was that checking reserved bits is, minimally, not needed,
and in fact either disallowed entirely or at the very least
discouraged.  (I.e., it is either "MUST NOT check" or "SHOULD NOT
check".) 

	 paul