ATA command pass through (16) opcode

[Pat LaVarre]

* From the T10 Reflector (t10 at t10.org), posted by:
* Pat LaVarre <p.lavarre at ieee.org>
*
> Why would you want to do a SCSI pass through to a device that 
> implements
> a SCSI command set? One could argue that belongs in SAT, not in this
> pass through.

Hi, thanks for your interest.  Three people have responded, you're the 
first online.

Certainly I agree that trying to pass, thru SCSI, the ATA op xA0 Packet 
that initiates ATAPI data protocols, is probably never highly useful, 
because so often we can achieve essentially the same effect by passing 
the packet data as a SCSI command to the ATAPI device.

All the same, I think ATA thru SCSI should accommodate ATAPI as much as 
ATA.  That is, given that we want to pass ATA commands thru SCSI to ATA 
devices, I suspect we also want to pass ATA commands thru SCSI to ATAPI 
devices.

Two specific ATA commands that immediately strike me as useful are the 
op xA1 "IDENTIFY" command that reports slave present, available data 
burst transfer rates, the currently negotiated data burst transfer 
rates, etc.  Also the op xDA Get Media Status originally defined by 
Microsoft.

One correspondent informs me offline that yes people are saying in fact 
MMC does somehow collide with the 12 byte opcode we chose, so that to 
pass ATA thru to ATAPI devices we will have to resort to the 16 byte 
opcode.  ATAPI folk might be well served to check this out in advance, 
because they may find they are pioneering the use of 16 byte CDB's in 
places that choke over anything beyond 12.

Am I yet making any sense?

I'm by now guessing that "A1h for the 12 byte command pass-through" 
means compatible with SBC and RBC only, not MMC.  I'm guessing that 
"85h for the 16 byte command pass-through" is compatible with MMC but 
might not remain so if we fail to find an ATAPI volunteer to get that 
op named there in t10:).

What "09h for the return descriptor" means, I have no idea, I'm not yet 
watching closely enough to know.

Indeed, I'm not now close enough to confidently discriminate SAT SATA 
SATAPI SCSI-over-SATA etc. etc. etc. but in this thread I was hoping we 
were discussing only passing ATA thru SCSI.

Curiously yours,
Pat LaVarre
http://ide-byte-counting.blog-city.com/

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