X3T9.2/90-84
Thomas Wicklund
Ciprico, Inc.
612-559-2034
612-559-8799 (FAX)


Comments on Maxtor questions about ESDI, document X3T9.2/90-074R1:


1.  I agree with this change.  The Diagnostic command parameter
is best considered a "command modifier" and "command subscript". 
This is also consistent with the way the Set Configuration
command is documented.


2 + 3.  The Request Status command does not support subscripts
for command modifiers 0xxxb.  The number of valid vendor unique
status words are determined by the "Number of Vendor Unique
Status Words" configuration response (modifier 1001b) and may
specify 0 through 7 words.  This behaviour of vendor unique
status is documented in section 7.6 (page 35 of the Rev 3A
document):

 - Request Vendor Unique Status: When the Command Modifier of the
   Request Status command is 0001-0111, the drive responds with
   Vendor Unique Status (undefined in this specification). The
   number of words available is specified by configuration data.
   Each word of Vendor Unique Status is requested using a
   different Command Modifier configuration. The Command
   Modifier for the first word is 0001 and subsequent words are
   requested by incrementing the Command Modifier. 

Note that each word must be requested "using a different Command
Modifier".


4.  If subscripted status is implemented, you must define the
number of subscripts for each word of vendor unique status (or
only define subscripts for one of them, e.g. vendor status 1).

The subscript used should be as a response to a request
configuration with modifier 0 (as are all currently defined
subscripted configuration words).  Placing all subscripts under
modifier 0 was done at the request of drive vendors to simplify
their firmware.

I think a further question needs to be asked about whether more
than 7 status words are needed outside a manufacturing diagnostic
environment in which the drive type and characteristics are known
ahead of time and a vendor unique extension could be implemented.


5.  I see no problem clarifying this if necessary.


6.  I think paragraph 4 means that for a given diagnostic number
(command modifier), a diagnostic parameter of 0 performs all
routines of the specified diagnostic number.  The standard
diagnostic (#0) would not necessarily perform all diagnostics.

I don't see a problem leaving the diagnostic parameter completely
vendor specific.  The host must know the characteristics of the
drive in order to use a diagnostic parameter other than 0 as
currently defined, and specific diagnostics will in general be
driven by a program which knows something about the drive beyond
configuration data across the interface.


7.  The first paragraph of section 7.12 may not be clear, but the
intent is that if a diagnostic fails ATTENTION is asserted and
status (standard and vendor specific) is returned to document the
error detected by the diagnostic.