[t13] 04-136r0.pdf SCSI to ATA Command Translations

Jeff Garzik jgarzik at pobox.com
Fri Jun 4 11:28:03 PDT 2004

This message is from the T13 list server.

Pat LaVarre wrote:
>>3.1.2 SCSI INQUIRY CDB format and supported fields
> I mis/remember MMC requires a zero in the byte 2 Version field of op x12
> "INQUIRY" data ... compatibility there could force us to pay attention
> to how to report version in the standard bytes that follow the
> vendor-specific bytes out past offset x24.

Most ATAPI and USB devices report zero in the version field.

My choice in Linux is to simulate MMC-3 version, and supply auto-sensing 
capability even when the low-level device does not.

>>3.2 Read Capacity (10) Command  (25h)
>>This value is currently set to 512 bytes,
>>which is the standard sector size for disk drives. 
> I see us taking the chance to try to fix the bytes/LBA at 0.5 Ki, like
> MMC fixed the bytes/LBA at 2 Ki, shutting out the (empty set of?) ATA MO
> folk.
> I like that.

It shut out nobody.

IMO the OS request block size _should_ be constant at 512 octets.  It 
makes calculations easier, and representations more normal.

For MO devices with >512 sector sizes are fine, just make sure the OS 
always sends down a multiple of 512-byte sectors as required by the device.

However, any device with a non-power-of-two sector size should be shot 
on sight ;-)  SATA specification mentions such devices...  grumble.

>> For the Read(12) command, if the size
>>TRANSFER LENGTH  field is greater than 16
>>bits, then the command will terminate with a
>>check condition where the  sense key set to
>>illegal request and additional sense code set
>>to invalid field in CDB.
> I like that.  I independently reached that conclusion and shipped that
> solution once, together with a comment something like "Comdex, 1996,
> nothing so permanent as a temporary kluge".  So far as I know, noone
> ever returned to that code to kick up the max LBA's/CDB past xFFFF.

<shrug>  It's required by the underlying ATA device, not much else you 
can do but limit to 0xffff -- or even 0xff for non-lba48 devices.

For lba28 devices, Linux limits you to 0xff (actually a tiny bit less) 
as expected.


More information about the T10 mailing list