09-063r3 clarification regarding BREAK's received during SL_CC8: PS_REQUEST

richard.bohn at seagate.com richard.bohn at seagate.com
Thu Jun 4 12:43:48 PDT 2009

* From the T10 Reflector (t10 at t10.org), posted by:
* richard.bohn at seagate.com
Hi Mark,
Can you confirm my interpretation of the SAS Link Layer changes relating to
BREAKs received during the new PS_REQUEST state?
The case I am curious about is the appropriate behavior of a receiving
device when the BREAK_REPLY method of responding to BREAK sequences is
disabled. I believe the specification indicates that the device should not
respond with a BREAK primitive in this state, but rather transition to the
IDLE state without indicating to the host that the BREAK has been received.
Thus, the host would eventually hit the millisecond BREAK timeout.
Am I correct in thinking that this behavior, which I believe matches the
behavior of the IDLE state, is specified to remain backwords-compatible in
spirit with the SAS 1 specification, which does not appear to contemplate
receiving BREAKs outside of connections (of course, SAS 1 devices shouldn't
be involved with the power management states, which is why I say "in
Richard Bohn
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org

More information about the T10 mailing list