This message is from the T13 list server.

On Thu, 13 Dec 2001 13:34:18 -0700, Pat LaVarre wrote:

>On the device-side there is a community of people who use more
>or less of the ability of some AtapiPio engines to move
>x0001..FFFF bytes per INTRQ DRQ, rather living by the Ansi rules
>forbidding transfers of odd bytes before the end and xFFFF bytes
>always.

THIS IS NOT A REQUIEMENT ADDED BY T13. THIS WAS IN SFF-8020!!!

>Some of these people fell into this by way of coming from 8-bit
>Scsi, where such abilities can be taken for granted.

Then they are building devices that do not conform to the ATA/ATAPI-x
standards and probably do not even confrom with SFF-8020! These are
BAD DEVICES!!!

>On the host-side, I know Win95B couldn't transfer odd bytes in
>response to any INTRQ DRQ:  to work there the device has to move
>even counts always.

BAD HOST!

>Also on the host-side I can easily imagine hosts that think they
>know how many bytes to move in response to an Atapi INTRQ DRQ ...
>can't you?

No I can't imagine this! This would be in violation of SFF-8020 and
ATA/ATAPI-x. Such hosts would be VERY BAD HOSTS!!!

>I take as confirming evidence that now here I've
>heard someone actually speak up and say that the state diagram
>doesn't tell them any different, as indeed it doesn't.

I'll look at the state diagrams later today...


***  Hale Landis  *** [EMAIL PROTECTED] ***
*** Niwot, CO USA ***   www.ata-atapi.com   ***


Subscribe/Unsubscribe instructions can be found at www.t13.org.

Reply via email to