PFA proposal 94-190 rev 2

George Penokie GOP at RCHVMP3.VNET.IBM.COM
Wed Nov 2 11:20:12 PST 1994


This is as good as it gets in ASCII.  A postscript version and a Framemaker
version will be on the SCSI BBS soon.


Date:  Nov. 02,1994

To:  X3T10 Committee (SCSI)

From:  George Penokie (IBM)

Subject:  Exception Handling Selection Mode Page

1 Overview

Many of today SCSI devices have asynchronous events that occur.
These devices use several different methods for reporting asynchronous
events. This proposal allows a method for application clients to select the
method it expects the target to use when reporting asynchronous events.

This proposal also define parameters to control reporting of failures that
are predicted by a target.

2 Informational exceptions control page

The informational exceptions control page defines the methods used by
the target to control the reporting and the actions of specific informational
exception conditions. This page shall only apply to informational
exceptions that report an additional sense code of FAILURE
PREDICTION THRESHOLD EXCEEDED to the application client.

Informational exception conditions occur as the result of vendor specific
events within a target.  The informational exception condition may occur
asynchronous to any commands issued by an application client.



Table 1 - Informational exceptions control page

 Bit

Byte

7

6

5

4

3

2

1

0

0

PS

Reserved

Page code (xxh)

1

Page length (0eh)

2

Reserved

LogErr

3

Reserved

Method of reporting informational
exceptions

4

Reserved

5

perf

Reserved

rpf

Reserved

Informational
Exception control

6

Reserved

7

Reserved

8

MSB

 interval timer

9

10

11

LSB

12

MSB

report count

13

14

15

LSB

The log errors bit (logerr) of zero indicates that the logging of
informational exception conditions within a target is vendor specific. A
logerr bit of one indicates the target shall log informational exception
conditions.

The method of reporting informational exceptions field (see table 2)
indicates the methods that shall be used by the target to report
informational exceptions.



Table 2 - Format of method of reporting informational exceptions field

Code

Description

0h

No indications of informational exception condition

This method
instructs the target to not report information exception conditions. The target
shall preserve the informational exception conditions information. To find out
about information exception conditions the Application Client shall poll the
target by issuing a REQUEST SENSE command.

1h

Asynchronous event reporting

This method instructs the target to report
informational exception conditions by using the rules for asynchronous
event reporting as described in the SCSI-3 Architecture Model and the
relevant Protocol Standard.

2h

Generate unit attention

This method instructs the target to report
informational exception conditions by issuing a unit attention condition on
any command. In response to a REQUEST SENSE command the target
shall return sense data. The sense key shall be set to UNIT ATTENTION
and the additional sense code shall be indicate the cause of the
informational exception condition.

The command that has the CHECK CONDITION shall not be executed
before the informational exception condition is reported.

3h

Generate recovered error:

This method instructs the target to report
informational exception conditions by returning a sense of CHECK
CONDITION on any command. In response to a REQUEST SENSE
command the target shall return sense data. The sense key shall be set to
RECOVERED ERROR and the additional sense code shall be indicate the
cause of the informational exception condition.

The command that has the CHECK CONDITION shall complete without
error before any informational exception condition may be reported.

4h-Bh

Reserved

Ch-Fh

Vendor specific

The interval timer and report count fields shall only be valid if the
method of reporting informational exceptions field contains a 2h or
3h code. The rpf field shall only be valid if the method of reporting
informational exceptions filed contains a 3h code.

The informational exception control field (see table 3) indicates if
information exceptions are reported and/or enabled.


Table 3 - Format of informational exceptions control field

Code

Description

00b

All informational exception operations and reporting shall be enabled.

01b

All informational exception operation are enabled but reporting shall be dis-
abled

10b

All informational exception operations and reporting shall be disabled.

11b

Reserved

A report predicted failures bit (rpf) of zero indicates the reporting of
predicted failures shall be dependent on the value of the per bit of the
error recovery parameters mode page. A rpf bit of one indicates all
predicted failures shall be reported regardless of the value of the per bit
of the error recovery parameters mode page.

A performance bit (perf) of zero indicates that informational exception
operations that cause delays are acceptable. A perf bit of one indicates
the target shall not cause delays while doing informational exception
operations.

The interval timer field indicates the period in 100 millisecond
increments for reporting that a informational exception condition has
occurred. The target shall report the informational exception condition as
soon as possible after the timer interval has elapsed. After the
informational exception condition has been reported the interval timer
shall be restarted. A value of zero in the interval timer field indicates
that the target shall only report the informational exception condition one
time. A value of FFFFh in the interval timer field shall indicate the timer
interval is vendor specific.

The report count field indicates the number of times to report an
informational exception condition to the application client. A value of zero
in the report count field indicates there is no limit on the number of
times the target shall report an informational exception condition.





More information about the T10 mailing list