Mode page for negotiated transfer speed / width

Sheldon_Kolansky at avid.com Sheldon_Kolansky at avid.com
Thu Nov 4 09:50:00 PST 1999


* From the T10 Reflector (t10 at t10.org), posted by:
* Sheldon_Kolansky at avid.com
*



The real problem is that I (at the application level) must query the O/S which
currently doesn't tell me what speed things are running.  I have no access to
messages or states of any bus signals.  I am also interested in the speed
negotiated between my adapter and the device.

The main thing I need to know is has anything happened to the bus to make it run
slow.

If I had access to the negotiation speed of devices on the bus I could determine
if the bus is running slower than that device could run.  This would provide
feedback as to a problem.  I can do an Inquiry to determine if there are any
single ended devices on an LVD bus which would force it into non-LVD, but I
can't see the terminator.

Thanks
Sheldon







Richard Moore <r_moore at qlc.com> on 11/04/99 12:27:29 PM
                                                                                
                                                                                
                                                                                


                                                              
                                                              
                                                              
 To:      Sheldon Kolansky/AM/Avid at Avid, "'t10 at t10.org'"      
          <t10 at t10.org>                                       
                                                              
 cc:                                                          
                                                              
                                                              
                                                              
 Subject: RE: Mode page for negotiated transfer speed / width 
                                                              








>* From the T10 Reflector (t10 at t10.org), posted by:
>* Sheldon_Kolansky at avid.com
>*
>
>
>I'm sorry if this has been discussed before and I missed it.
>
>We would like to be able to query each device to determine
>what the speed and
>width has been negotiated to.

By negotiating with the device you should know the outcome automatically.
But if you mean you want to know what a device has negotiated with other
devices, this would be unique for each initiator/target combination on the
bus. I don't think there is currently any way to do this (other than
snooping for the negotiation messages between devices).

>We rely upon the full bandwidth of the bus, and now the max
>bus speed can be
>influenced by things that we cannot determine via software like:
>
>Is there a single ended terminator on my LVD bus?

Single-ended devices and terminators will ground the DIFFSENS line. Each
device (except single-ended devices) must be able to sense the level of
DIFFSENS to make this determination. How you propagate this information
up to some application software, if at all, is an implementation-specific
matter.

>
>Has the top bus speed been degraded due to some other bus problems?
>

The tool for determining this is Domain Validation, also known as Physical
Layer Integrity Checking, described in Annex L of SPI-3 rev 10. Like
negotiation, it is a device-to-device procedure so unless you are capable
of snooping this activity between other pairs of devices you would only
know the capabilities between a specific initiator and each target.

Hope this helps,

 -- Richard Moore
    QLogic Corp.

>This is mainly for use as a debug tool for our support group
>to determine if we
>can't keep up to the video and audio due to some physical
>issue with the bus or
>some other problem.
>
>
>*
>* 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