ST_ITS (initiator transport server) state machine question

George Penokie george.penokie at avagotech.com
Mon Dec 1 09:37:00 PST 2014


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1412012_f.htm">HTML-formatted message</a>

Ralph,
You don't seem to be missing anything. The task management functions and
the command requests are completely separate. All commands and TMFs that
have an active ST_ITS state machine have their own initiator port transfer
tag so they are processed as separate operations.
The handling of errors in the link layer is described in 8.2.4.
There the case you describe that can get you to step 5 but the behavior of
not sending any data at that point seems to be a problem in the
implementation. The initiator should either send data or if it is confused
it should abort the command and try the whole thing over again.
If nothing happens the device server can timeout the operation if that is
enabled.
Bye for now,
George Penokie
Avago Technologies
Attn: George Penokie
4109 Manor View Dr NW
Rochester , MN 55901
952-921-2495
george.penokie at avagotech.com
On Fri, Nov 28, 2014 at 5:14 PM, Ralph Weber <Ralph.Weber at wdc.com> wrote:
>  I am receiving reports of what seems like a quirky behavior vis a vis
> the SPL-4 working draft.
>
> The observed events are as follows.
>
>    1. A write command is sent to the device server.
>    2. When the device server responses to host with an XFER_RDY IU, one
>    normally expects the application client to send DATA IUs.
>    3. Instead, the task manager receives a TASK IU with the TASK
>    MANAGEMENT FUNCTION field set to 80h (i.e., QUERY TASK).
>    4. The task manager responds to the TASK IU with the Service Response
>    set to FUNCTION SUCCEEDED, but ...
>    5. The device server never receives the DATA IUs expected as a result
>    of the XFER_RDY IU.
>
> SPL-4 r01 appears to say that there is one ST_ITS state machine for each
> possible command and task management function, which suggests to me that
> there should be no interactions between the XFER_RDY IU and the TASK IU
> (unless the TASK IU causes the command to be aborted, as specified
> elsewhere in SPL-4).
>
> What am I missing?
>
> All the best,
>
> .Ralph
>



More information about the T10 mailing list