This message is from the T13 list server.
This is a t10.org issue, for which we would might more complete/ accurate answers talking @ t10.org, but I will pipe up here to disagree with something already said: > Once you determine that a device supports the PACKET command set, it is > safe to assume the device supports the SCSI command op x12 INQUIRY as > defined in SPC. This leap is courtesy of the following from SPC-2 > (T10/1236-D Revision 20 pg. 19): > > This standard defines three commands that all SCSI device servers > shall implement - INQUIRY, REQUEST > SENSE, and TEST UNIT READY. These commands are used to configure the > system, to test devices, and to > return important information concerning errors and exception > conditions. > > The INQUIRY response data reflects the command sets supported by the > device. That leap to SPC from SCSI-over-whatever is written there, aye. But also we see MMC-4 then goes and redefines op x12 Inquiry. We say SPC is the common root from which all SCSI talk flows for any PDT whatsoever yes ... provided we then remember that PDT x05 MMC is the exception which prints its own root. That the MMC root is redundant with SPC is clear. How often that redundancy actually rather than merely potentially leads to binary incompatibility, answering that question accurately takes study. In so far as t10.org MMC is an adoption of the standard pioneered by SFF, getting an answer that corresponds to the reality of a gazillion shipped DVD/ CD drives may involve study of SFF precedent, not just ANSI t10.org precedent. I mis/remember that although MMC before MMC-4 didn't admit to rewriting op x12 Inquiry, SFF did, so by rewriting Inquiry MMC-4 has more accurately reflected the reality of the SFF departure. May also be worth remembering that PDT x00 SBC, PDT x0E RBC, PDT x05 MMC etc., all from t10.org, give binary incompatible definitions for the same commands. For example, PDT x00 SBC pretends that blocks may contain other than 0.5 KiB, but PDT x05 MMC does not, instead it explicitly requires 2 KiB. So the USB folk defining bootable SCSI (which is a SCSI flavour developed outside of t10.org) may have erred by requiring op x25 Read Capacity. A more real definition might be to require op x25 Read Capacity unless PDT x05 DVD/CD or unless assuming the 0.5 KiB/block of PDT x00 HDD that excludes certain obscure MO etc. Pat LaVarre
