This message is from the T13 list server.

On Thu, 30 Oct 2003 00:53:33 -0800, Mark Overby wrote:
>This message is from the T13 list server.
>I'll throw into the hat:
>Defining a session (host) layer if the commands change away from task-file
>based to FIS-based.

Yea, me too... Merging SATA into ATA/ATAPI-7 was a *MAJOR* mistake.
SATA is not compatible with PATA. Trying to produce a single
ATA/ATAPI standard for will always be a major problem. More and more
every day it is clear that SATA has a number of problems that make it
unique and very much unlike PATA. The fact that SATA even comes close
to looking like (but not executing like) PATA is all in the host
controller for which there is no specification or standard. This mess
is growing larger every day with more and more incompatibilities
being found.

I strongly encourge the move to a "native" SATA host controller
interface ASAP (call it FIS based or whatever). Lets get the world to
move away from this stupid "SATA is just like PATA" idea ASAP.

I think ATA/ATAPI-7 is and will continue to be a huge mistake.
ATA/ATAPI-8 should go back to just being PATA and a entirely new
standard should be created for SATA that *INCLUDES* a new host
controller interface specification (maybe that is AHCI?).

As things are it will be virtually impossible to produce a single
standard for PATA and SATA in one document (like ATA/ATAPI-7) - there
are just too many conflicting things going on in these two
interfaces. Plus even today (when SATA is promoted as a ANSI T13
project) most of the SATA technical issues are still being resolved
by "secret society" meetings (giving many people the false idea that
SATA is now an "open" standard - it is not).

I say all of this because I continue to have problems when I test
SATA devices - I generally can't run tests for more than a few
minutes with getting timeout problems, hung controller/devices, data
compare problems and other strange errors. All of this tells me that
1) SATA implementations have lots of problems (and some of those are
spec issues that have been discussed here) and 2) some are the simple
fact that SATA is not really compatible with PATA.

PATA is still a healthy interface and still the primary interface
being used. Lets don't mess it up in some misguided attempt to merge
the SATA and PATA standards - that has a very low chance of actually
working - as we see today.

Hale



*** Hale Landis *** www.ata-atapi.com ***



Reply via email to