SCSI-3 Reseved Fields and Codes

[Roy Shenfield x2473]

While I agree that the checking of reserved fields is a cumbersome task, and it would
be nice not to have to do so, based on experience, there are some practical concerns
that should be taken into consideration.

Consider the senario where drive company A (who chose not to do reserved field checking),
is competing against drive company B (who implements this checking and reports CHECK
CONDITION if a reserved bit is not set to zero), for a large volume contract at systems
company C.

Company C will run both competing drives through an array of tests, including specific drive
testing, and system testing.  At some point in these tests the fact that drive A does not 
(and that drive B does) report CHECK CONDITION when a reserved bit is non zero will be
noticed and reported.  Guess which drive company's management gets told there is a problem,
and which drive company's engineers will be working weekends until drive A operates just likedrive B?

Now maybe drive company A could argue that it's drive's performance will be better than
drive B due to not having to check reserved bits, but maybe some of C's customers also do
qualification testing that would notice this difference, so that this argument will not
satisfy C.

In summary, it seems to me that making the setting of reserved bits equal to zero mandatory, but the checking of the same reserved bits optional could lead to some potential problems.

-------------------------------------------------
Roy Shenfield
Senior Staff Engineer
Adaptec
email roys at adaptec.com

These opinions are mine only and do not reflect any company position.