Credit Advance Question

[Penokie, George]

Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1404231_f.htm">HTML-formatted message</a>

Judy,
It has been pointed out that the proposal that was accepted for Credit
Advance was not properly incorporated. Proposal 13-138r7 stated the
following:
6.16.4.3.4 Transition SL_CC1:ArbSel to SL_CC3:Connected
This transition shall occur after sending:
a) a Connection Opened confirmation.
If the SAS PROTOCOL field in the transmitted OPEN address frame was set to:
a) STP, then this transition shall include an Open STP Connection argument;
b) SSP, then this transition shall include an Open SSP Connection argument;
or
c) SMP, then this transition shall include an Open SMP Connection argument.
If persistent connections are supported (see 4.1.13), and:
a) the SEND EXTEND bit in the transmitted OPEN address frame was set to one
and the INITIATOR PORT bit was set to one, then this transition shall include
an Extend Connection (Wait) argument;
b) the SEND EXTEND bit in the transmitted OPEN address frame was set to zero,
then this transition shall include an Extend Connection (Off) argument. or
c) the INITIATOR PORT bit was set to zero, then this transition shall include
an Extend Connection (Off) argument.
If credit advance is implemented (see 4.1.14), and:
a) the CREDIT ADVANCE bit is set to one in the transmitted OPEN address
frame; and
b) the SAS PROTOCOL field is set to 001b (i.e., SSP) in the transmitted OPEN
address frame,
then this transition shall include an Advance Credit argument.
Based on this editing error, SPL-3 I will correct this error in the next
revision of SPL-3.
Bye for now,
George Penokie
LSI Corporation
3033 41 St NW
Rochester , MN 55901
507-328-9017
george.penokie at lsi.com
From: owner-t10 at t10.org [mailto:owner-t10 at t10.org] On Behalf Of Penokie,
George
Sent: Monday, April 21, 2014 9:59 AM
To: Judy Westby; Tim.Symons at pmcs.com; Gerry Houlder
Cc: Voorhees, Bill; t10 at t10.org
Subject: RE: Credit Advance Question
Judy,
It looks to me like there is an error in the description in section
6.16.4.3.4 Transition SL_CC1:ArbSel to SL_CC3:Connected. This transition only
occurs when arbitration is won by the phy. When arbitration is won and open
address frames cross the information in the losing open address frame is
discarded. So the following statement in section 6.16.4.3.4 Transition
SL_CC1:ArbSel to SL_CC3:Connected is in error:
If credit advance is implemented (see 4.1.14) and:
a) the CREDIT ADVANCE bit is set to one in the received OPEN address frame;
and
b) the SAS PROTOCOL field is set to 001b (i.e., SSP) in the received OPEN
address frame,
then this transition shall include an Advance Credit argument.
It should be deleted or restated as follows:
If credit advance is implemented (see 4.1.14) and:
a) the CREDIT ADVANCE bit is set to one in the << transmitted >>  OPEN
address frame; and
b) the SAS PROTOCOL field is set to 001b (i.e., SSP) in the << transmitted >>
OPEN address frame,
then this transition shall include an Advance Credit argument.
Tim,
What was your intent? Did you want the advance credit on both sends of the
connection or did you want the advance credit only on the side that accepted
the connection.
Bye for now,
George Penokie
LSI Corporation
3033 41 St NW
Rochester , MN 55901
507-328-9017
george.penokie at lsi.com
From: Judy Westby [mailto:judy.westby at seagate.com]
Sent: Friday, April 18, 2014 4:31 PM
To: Penokie, George; Tim.Symons at pmcs.com; Gerry
Houlder
Subject: Credit Advance Question
We were looking at the credit advance function and wanted to clarify
something.
The following is included in the Transition SL_CC0:Idle to SL_CC2:Selected,
Transition SL_CC1:ArbSel to SL_CC3:Connected, and Transition SL_CC1:ArbSel to
SL_CC3:Connected:
If credit advance is implemented (see 4.1.14) and:
a) the CREDIT ADVANCE bit is set to one in the received OPEN address frame;
and
b) the SAS PROTOCOL field is set to 001b (i.e., SSP) in the received OPEN
address frame, then this transition shall include an Advance Credit argument.
When Open Address frames are sent by both sides at about the same time, one
will win and receive OPEN_ACCEPT and one will lose and send OPEN_ACCEPT.  The
Advance Credit argument may be set for both at the same time.  Both
 could then transmit a frame before receiving the first RRY.
We didn't see any issues with this but 13-138 mentioned the OPEN_ACCEPT so
just wanted to confirm that the intent wasn't that it is dependent on which
port wins.
--
Judy Westby
Seagate Technology
952-402-2719
Seagate Confidential Information:
This e-mail message and any files transmitted with it are for the sole use of
the intended recipient(s) and may contain confidential and privileged
information. Any unauthorized review, use, disclosure, or distribution is
prohibited. If you are not the intended recipient, please contact the sender
by reply e-mail and destroy all copies of the original message.