question about FCP_DL disk I/O...

Matthew Jacob mjacob at feral.com
Fri Sep 19 15:36:23 PDT 2003


* From the T10 Reflector (t10 at t10.org), posted by:
* Matthew Jacob <mjacob at feral.com>
*


The FCP CMND IU has an FCP_DL field, which FCP2 defines to be the
maximum number of bytes that the command is expected to transfer.

In commands that read or write write data blocks (e.g., READ or WRITE)
to DIRECT ACCESS devices, can can anyone think of a non-error case or
where there would be fewer bytes written than FCP_DL?

A T10 commitee member once told me (long long ago) that it's a fatal
system error for a disk to not accept the number of blocks specified by
a READ or a WRITE command, lacking any reported error condition.

If that's the case, then would it be safe to assume that FCP_DL always
equal always to the product of #Logical Blocks and Transfer Length?

I suspect this is just depending on expected behaviours, or you can
detect an underrun after the fact, but I'm wondering if there's any
architectural foundation to this.

-matt

*
* 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