READ/WRITE LONG

Jim McGrath jmcgrath at mail.qntm.com
Wed Oct 5 07:48:36 PDT 1994


                      Subject:                              Time:  7:45 AM
  OFFICE MEMO         READ/WRITE LONG                       Date:  10/5/94


It seems what Ralph wants is an ability to associate more bytes with a given
block of data (for control information), but not for all blocks on the disk. 
Sort of like mixing 512 byte and 524 byte blocks on the same device.

As long as people understood that EC protection was given up (and thus error
rates were REAL bad), and people did not do READs of data written with WRITE
LONGs, that would work.  But in that case I would sugest substantial
rewording of the section to justify the usage Ralph wants and warn people
that ECC testing is not appropriate (for the reasons I cited earlier).

Jim






More information about the T10 mailing list