This message is from the T13 list server.

Hi Pat,

You seem focussed on the "extra clocks" during Ultra DMA burst termination.
Let's try to take this up a level (or down a level, as the case may be).
These are not "extra clocks".  This is explained in several places in the
standard (e.g., item (d) in 9.13.3.2 Host pausing an Ultra DMA data-in
burst), "The additional data words are a result of cable round trip delay
and tRFS timing for the device."  In other words, if the sender is
transmitting data, and the recipient initiates a burst termination, then the
sender shall respond to the burst termination as quickly as possible, but
this response is not instantaneous.  By the time that, a) the termination
signal gets down the wire, b) the sender detects that the recipient has
initiated a termination, and c) the sender can react to the termination by
stopping data transmission, the sender may have already clocked one, two, or
three more times.  This is perfectly good data.  This data will be included
in the CRC calculation.  The only "arbitraryness" is the result of the
inherent asynchronicity of the system and implementation differences.  The
"one, two, or three" is the result of the cable plant and when the
recipient's internal circuitry detects the pause and can react to it.  I
hope this helps.

Please feel free to call or send an email to me with any additional
questions that you may have about this.

Regards,

Mark Evans
Maxtor Corporation
500 McCarthy Boulevard
Milpitas, CA 95035 USA
Tel:  408-894-5310
FAX:  408-324-7432
email:  [EMAIL PROTECTED]
Subscribe/Unsubscribe instructions can be found at www.t13.org.

Reply via email to