parity handling

Jim McGrath jmcgrath at QNTM.COM
Tue Feb 6 14:39:54 PST 1996

* From the SCSI Reflector, posted by:
* "Jim McGrath" <jmcgrath at>
        Reply to:   RE>parity handling

To echo a previous comment, all host designers should always assume
that writes that are not successfully completed for ANY REASON AT ALL
left the data in the LBA range specified in an indeterminate state.  Any
other assumption will almost certainly get you in trouble some of the
time.  Since there is no value in making any other assumption (errors
occur so rarely that there is not a performance issue), and a lot of
complexity and a nightmare of qualification if you do not, I see making
any other assumption a carreer limiting move.


More information about the T10 mailing list