Unknown command 0x23

Pat LaVarre p.lavarre at IEEE.org
Mon Oct 20 09:04:28 PDT 2003


* From the T10 Reflector (t10 at t10.org), posted by:
* Pat LaVarre <p.lavarre at ieee.org>
*
[ BC leguanolog at seznam.cz ]

> Subject: Re: Unknow command 0x23
> From: Ralph Weber <roweber at ieee.org>
> ...
> Have you somehow caused Windows XP
> to think that your device is a CD/DVD-ROM?

Bus traces tell me you have told Windows you have a DVD/CD, perhaps
unwittingly, if into Win 9X/ME 2K/XP you have plugged in an ATAPI
device.

> Operation code 23h is marked Vendor Specific.
>
> Note: Annex C of SPC-3 contains a numeric list of
> operation code values and assigned uses.
> (http://www.t10.org/ftp/t10/drafts/spc3/spc3r15.pdf)

Yes, an incorrect list.

> Subject: Re: Unknown command 0x23
> From: Elliott, Robert ...
> ...
> What is your peripheral device type
> reported in the Standard INQUIRY data?

A key factor, I agree.

> For these device types:
> * direct-access (00h)
> * optical memory (07h)
> * write-once (04h)
> SCSI-2, SPC-n, and SBC-n all agree
> that opcode 23h is vendor-specific,
> not reserved.

Yes, but in the world ruled by Windows that's misleading fiction. 
Windows occasionally behaves as if any ATAPI device reported PDT x05
DVD/CD.

> For the CD-ROM (05h) device type,
> the standards have changed
> in how they document opcode 23h:
> SCSI-2: vendor-specific
> SPC: vendor-specific
> MMC-1: vendor-specific
> SPC-2/3: optional READ FORMAT CAPACITIES command
> MMC-2/3/4: optional READ FORMAT CAPACITIES command
> (see MMC-4 for the current definition)

Yes.

> (presumably all vendor in the industry agreed
> to that change; vendor-specific values usually
> cannot be standardized after they've been
> labeled as such)

No.  I'm sorry to say I personally know of specific products that
shipped with vendor-specific definitions of op x23 that respond
unpleasantly when abused this way by Microsoft.  I'm pleased to say I do
Not know of any products that shipped that way fully assembled in
volume.

> If you're seeing READ FORMAT CAPACITIES being
> sent to anything other than device type 05h,
> then the software is confused.

Yes, naturally, Windows ships in binary code only.

> If you're designing a device of type 05h, you
> should be referencing MMC-4, not SBC-2.

Also I recommend MMC and not SPC.  MMC competes with SPC to define such
ops as op x12 Inquiry and op x5A Mode Sense 10 in more or less binary
incompatible ways.

Pat LaVarre


*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org




More information about the T10 mailing list