This message is from the T13 list server.
Can you say specifically how incorrect/incomplete the
signature you are seeing is?
I remember hearing in particular that Phoenix Bios
often writes x90 ExecuteDeviceDiagnostic to the x1F7
Command register, so I wonder how tolerant those hosts
would be of an Atapi device that did not sign at least
x1F5:1F4 Cylinder = xEB:14.
I'd say the host "should" try and zero the x1F6 & x10
DEV bit beforehand, at least if x3F6 & x88 BSY and DRQ
were clear. Otherwise, by design, the host and the
one or two attached devices race to decide when
everyone will agree that DEV has been implicitly zeroed.
When BSY|DRQ are set, how the host "should" prepare to
write x90 is less clear, since writing x1F6 then is
also rude, so then we have to balance rudenesses.
Pat LaVarre
-----Original Message-----
From: Mark Overby [mailto:MOverby@;nvidia.com]
Sent: Thu 10/24/2002 4:12 PM
To: 'Harlan Andrews'; '[EMAIL PROTECTED]'
Cc:
Subject: RE: [t13] Execute diagnostics
This message is from the T13 list server.
Master device, only device on channel.
-----Original Message-----
From: Harlan Andrews [mailto:handrews@;apple.com]
Sent: Thursday, October 24, 2002 12:55 PM
To: Mark Overby; '[EMAIL PROTECTED]'
Subject: Re: [t13] Execute diagnostics
Are you asking about the Master device, or a "Shadowed" slave device ?
>Is it common that ATAPI devices do not put the correct signature into the
>registers after completing to indicate they are ATAPI devices?
>
>
