[t13] SCSI ATA f/w revision bytes how many

Jeff Garzik jgarzik at pobox.com
Tue May 25 13:49:31 PDT 2004

* From the T10 Reflector (t10 at t10.org), posted by:
* Jeff Garzik <jgarzik at pobox.com>
Pat LaVarre wrote:
>> The only thing I care about is the 16-byte CDB limit currently in 
>> Linux...
> If we take advantage of the extra four bytes there, we will not design a 
> solution usable by the max-12-CDB-byte connections, in particular normal 
> PATAPI, that also wish to give ATA pass thru to hosts that think they 
> support only SCSI pass thru.  Those people will therefore push to 
> revisit this issue, yet again.

No.  I think you are missing some context...

This opcode is used to tunnel through a _fake_ SCSI layer to reach a 
real ATA device behind it.

ATAPI does not come into play here.  This "ata passthru" opcode is only 
needed for ATA.  For ATAPI, one does not need to pass through a fake 
SCSI layer, but rather avoid the hardware device's SCSI layer and talk 
directly to ATA.  Subtle difference... but substantial.


* 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