> > > ... Zip 250s die if you send a command to a LUN other than 0.

> > I'm inclined to think that absolute any device response is
> > compliant - but that provoking such a response makes the host
> > in question non-compliant.

> Yes, any response would be good.  However, as far as I can
> tell, the device crashes -- no commands are accepted, no
> status is transferred.

A controlled flight into terrain is an allowed response
to a non-compliant host.

No?  What am I missing here?  I've heard that what happens
is STALL on both bulk pipes until the host apologises for
its non-compliant rudeness by way of a BBB 5.3.4 Reset Recovery?

> > I'd quote BBB 6.2.2 ... CBW meaningful when ... LUN ...
> > 6.4 "The response of a device to a CBW that is not
> > meaningful is not specified."

> Yes.  However, a subsequent command to a valid LUN
> should work according to this, and they don't.

Please walk me through this thinking more slowly?  To say that
a response is not specified is as broad a gaping hole as one
can write into a text?  No?

Clamp down until host indicates resynch via Reset Recovery
IS an allowed response?

> > I've been told that idProduct x0001 Zip 100 respond to
> > nvalid LUN with some obsolete bCSWStatus value.  I've been
> > told that Zip 250 and later Zip 100 respond to invalid LUN
> > as if a phase error had occurred i.e. require a BBB 5.3.4
> > Reset Recovery to continue.
> 
> Hrm, I didn't see a CSW returned at all.

Whoops, let me add precision to my choice of words ...

The bridges that throw a tantrum in response to an invalid
LUN throw this tantrum AFTER returning a CSW to report a
phase error.  They throw this trantrum immediately after
receiving a meaningless CBW, without first returning a CSW.

The bridges that are more tolerant of invalid LUN return
a CSW with an obsolete bCSWStatus (x04 I think, x03 being
the obsolete value reserved for the "persistent failure"
concept of command queueing).


x4402 Pat LaVarre of iOmega   [EMAIL PROTECTED]   http://members.aol.com/ppaatt/


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to