SAS XL state machine to XL transmitter messages

Gil Romo gil.romo at qlogic.com
Tue Dec 9 11:04:51 PST 2003


* From the T10 Reflector (t10 at t10.org), posted by:
* Gil Romo <gil.romo at qlogic.com>
*
In "backoff and reverse path", as shown in G.7, the expander phy enters XL2 from 
XL6.

Hence, it is not coming from the XL1 state where the phy typically begins 
transmitting AIP (NORMAL).

Should XL2 begin sending Transmit AIP (Normal) messages to the XL transmitter 
regardless of whether it came from XL1 or XL6?

> 
> * From the T10 Reflector (t10 at t10.org), posted by:
> * "Elliott, Robert (Server Storage)" <elliott at hp.com>
> *
> 
> > * From the T10 Reflector (t10 at t10.org), posted by:
> > * Gil Romo <gil.romo at qlogic.com>
> > *
> > There is a discrepancy between 7.15, XL state machine, and 
> > Annex G, Expander device handling of connections.
> > 
> > In 7.15.5 and 7.15.6, the XL2 and XL3 states repeatedly send 
> > Transmit Idle Dword  messages to the XL transmitter until
> > the first ECR confirmation is received.
> > 
> > However, G.2, G.3, G.5, G.6, G.7, and G.12 show that the 
> > Phy[X] transmitter continuously sends AIP (NORMAL) during 
> > this period.  Note this also occurs with Phy[Y] in G.6 and G.7.
> > 
> > Will someone involved in this area please review?
> > ------------
> > Gilbert Romo
> > Circuits & Integration
> > QLogic Corporation, Aliso Viejo, California
> > Office: 949-389-6266
> > E-mail: gil.romo at qlogic.com
> 
> None of the AIPs are intended to be continuously transferred.  They're
> supposed to be sent once every 128 dwords (see 7.12.5.1) - the fastest
> allowed is 3 out of every 4 dwords and the slowest (what the receiver
> cares about) is 1 every 1 ms (to prevent the Open Timeout timer from
> expiring).
> 
> XL1 and the XL transmitter descriptions should be clearer that repeated
> Transmit AIP messages result in a mix of mostly idle dwords and an
> occasional AIP, not a steady stream of AIPs.
> 
> XL2 should continue repeating the AIP (NORMAL) at that interval, rather
> than just send idle dwords.
> 
> XL3 is already scheduled for cleanup regarding AIP (WAITING ON DEVICE)
> (so only one gets sent by this expander before it starts forwarding
> dwords).  If there is any time gap in XL3 before the first AIP (WAITING
> ON DEVICE) is transmitted, the previous AIP should be repeated every 128
> dwords while the XL state machine waits.
> 
> --
> 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

------------
Gilbert Romo
Circuits & Integration
QLogic Corporation, Aliso Viejo, California
Office: 949-389-6266
E-mail: gil.romo at qlogic.com

*
* 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