SAS DONE timers

Day, Brian Brian.Day at lsil.com
Wed Jul 26 15:18:40 PDT 2006


* From the T10 Reflector (t10 at t10.org), posted by:
* "Day, Brian" <Brian.Day at lsil.com>
*
Rich...
Whether it is "clear" is relative <grin>, but I believe that behavior is
covered in the port layer state machines.
In the PL_PM3:Connected state it says:
"If this state receives a DONE Received (ACK/NAK Timeout) or DONE
Transmitted confirmation, then this
state shall send a Disable Tx Frames message to the PL_OC state
machine."
And the PL_OC2:Overall_Control state machine says:
"If this state receives a Disable Tx Frames message from a PL_PM state
machine, then this state should send
no more Tx Frame messages to that state machine until a new connection
is established."
At least that is the intention.
Brian Day
LSI Logic
-----Original Message-----
From: owner-t10 at t10.org [mailto:owner-t10 at t10.org] On Behalf Of Richard
Deglin
Sent: Wednesday, July 26, 2006 2:53 PM
To: t10 at t10.org
Subject: RE: SAS DONE timers
* From the T10 Reflector (t10 at t10.org), posted by:
* "Richard Deglin" <rdeglin at vitesse.com>
*
Rob, in the DONE (ACK/NAK TIMEOUT) scenario, it seems to me that the
following points are true:
1. The sender of DONE (ACK/NAK TIMEOUT) will no longer reinitialize and
restart his DONE timer upon receipt of a frame from the other end.
2. This implies that the recipient of DONE (ACK/NAK TIMEOUT) must stop
transmitting frames within 1 ms of receipt of the primitive. Is there a
clause in the standard that makes this clear?
Thanks
Rich
-----Original Message-----
From: owner-t10 at t10.org [mailto:owner-t10 at t10.org] On Behalf Of Elliott,
Robert (Server Storage)
Sent: Tuesday, July 18, 2006 11:45 AM
To: t10 at t10.org
Subject: RE: SAS DONE timers
> -----Original Message-----
> From: owner-t10 at t10.org [mailto:owner-t10 at t10.org] On Behalf 
> Of Richard Deglin
> Sent: Monday, July 17, 2006 1:36 PM
> To: t10 at t10.org
> Subject: SAS DONE timers
> 
> I have been studying section 7.16.8.5 of the SAS-1.1 standard 
> and I have a few questions.
> 
> 1. What exactly does paragraph 6 mean, when it says "the DONE Timeout
> timer shall be reinitialized" ?
Set it to its initial value (1 ms, per table 119 in sas2r04a).
> 2. Why does the ACK/NAK timeout situation treat incoming EOF 
> differently
> than the other states? Just curious about the reasoning for this
> behavior.
When a phy sends DONE (any reason), it starts the 1 ms timer.
If it sent DONE (NORMAL) or DONE (CREDIT TIMEOUT), it will
keep reinitializing the DONE timeout timer if the other side
continues to send frames; useful work is still being performed.
If it sent DONE (ACK/NAK TIMEOUT), it does not keep reinitializing 
the timer, since this is an error condition and it needs to 
close the connection.
> 
> Also, I cannot find any specific definitions of the terms used when
> describing timer operations, such as initialize, reinitialize, start,
> stop, etc. Did I miss something?
Those are intended to be common engineering terms.  A timer is
intialized to a certain value.	Once started, it counts down.  Once
stopped, it stops counting.  When it reaches zero, it is considered
to be expired.
> 
> Thanks
> 
> Rich Deglin
> Principal Software Engineer
> Storage Products Division
> Vitesse Semiconductor
> Milpitas, CA
--
Rob Elliott, elliott at hp.com
Hewlett-Packard Industry Standard Server Storage Advanced Technology
https://ecardfile.com/id/RobElliott
*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org
*
* 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