This message is from the T13 list server.
On Tue, 12 Mar 2002 09:29:49 -0800, Thomas Colligan wrote: >This message is from the T13 list server. >It is a good way to force the Short and Long tests to fail and report back a >error. WHY? What are you testing? The ERR and UNC bits in the interface registers? Please explain what it is you think you are testing. What if you must corrupt all 512 bytes of a sector (using these commands) in order to get the drive to set ERR=1 and UNC=1? What does that test? How can you use this command to produce all the types (or even a small number of the types) of inputs to the drive's ECC hardware? Do you think that reading "good data" from the media that happens to have a few bits/bytes flipped is testing anything? Today's drives are doing ECC correction of some type on each and every sector they read (that is what PRML is all about!). Do you really think flipping bits using R/W Long is really testing anything? Come on... It is time to "get over" this idea from *50* years ago! Hard disk drives don't work that way anymore! *** Hale Landis *** www.ata-atapi.com ***
