Letter ballot comment - clarification needed in Close Connection vs Begin Close priority

Craig Stoops Craig.Stoops at synopsys.com
Wed Oct 30 17:51:25 PDT 2013


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1310307_f.htm">HTML-formatted message</a>

Hi John,
Here is the new letter ballot comment I wish to submit. George P has agreed
it needs clarifying language else as written the SSP phy would issue 2 DONE
types, which is bad.
In section 6.18.9.6.2.2 it is stated:
If a Close Connection request was received, then this transition shall
include a Close Connection argument.
If a Begin Close message was received, then this transition shall include a
Full Duplex Close argument.
Both of these could be true at the same time, the PL requesting close and an
RRDY / EXTEND (CLOSE) came in from an expander that set the Begin Close msg.
In this case which takes precedence?
Clarifying language should be added to note that Begin Close takes precedence
over Close Connection in TF1. The effect should be that ONLY the argument for
Full Duplex Close is passed on transition to TF2 much like was already done
for the case where a tx frame msg was also received at the same time as Begin
Close. The desired result should be only 1 DONE (CLOSE) is issued if a Begin
Close msg was received.
Thanks,
Craig Stoops
Synopsys
From: Penokie, George [mailto:George.Penokie at lsi.com]
Sent: Wednesday, October 30, 2013 10:38 AM
To: Penokie, George; Craig Stoops; t10 at t10.org
Subject: RE: Which takes precedence for Close in 6.18.9.6.2.2 ?
Craig,
Let me rephrase this. Yes, there appears there may be a problem, but because
SPL-3 is in letter ballot there is nothing I can do about it. I would advise
you to submit a letter ballot comment on SPL-3 with a solution that you think
will work.
Bye for now,
George Penokie
LSI Corporation
3033 41 St NW
Rochester , MN 55901
507-328-9017
george.penokie at lsi.com
From: Penokie, George [mailto:George.Penokie at lsi.com]
Sent: Wednesday, October 30, 2013 7:57 AM
To: Craig Stoops; t10 at t10.org<mailto:t10 at t10.org>
Subject: RE: Which takes precedence for Close in 6.18.9.6.2.2 ?
Craig,
Yes, and the very next paragraph covers the cases where both are received:
If a Begin Close message and a Tx Frame request were received, then this
transition:
a) shall include a Full Duplex Close argument; and
b) shall not include a Transmit Frame Balance Required argument or a Transmit
Frame Balance Not Required argument.
Bye for now,
George Penokie
LSI Corporation
3033 41 St NW
Rochester , MN 55901
507-328-9017
george.penokie at lsi.com
From: Craig Stoops [mailto:Craig.Stoops at synopsys.com]
Sent: Tuesday, October 29, 2013 5:44 PM
To: Penokie, George; t10 at t10.org<mailto:t10 at t10.org>
Subject: Which takes precidence for Close in 6.18.9.6.2.2 ?
HI George,
In section 6.18.9.6.2.2 it is stated:
If a Close Connection request was received, then this transition shall
include a Close Connection argument.
If a Begin Close message was received, then this transition shall include a
Full Duplex Close argument.
Both of these could be true at the same time, the PL requesting close and an
RRDY / EXTEND (CLOSE) came in from an expander that set the Begin Close msg.
In this case which takes precedence? I would think it should be the Begin
Close as that will result in a full duplex close to the remote end. But the
spec has neither a ordered list nor exclusive language, and the 1st one
listed is not the one that I think should take effect.
Thoughts? Do we need some edit in this area?
Craig



More information about the T10 mailing list