FIBRE CHANNEL - T11

Copies of this document may be purchased from: August 7, 1998 

Global Engineering, 15 Inverness Way East, NCITS TR-xx/Project 1315-DT/Rev 1.07 

Englewood, CO 80112-5704 

Phone: (800) 854-7179 or (303) 792-2181 Fax: (303) 792-2192 

FIBRE CHANNEL 

TAPE (FC-TAPE) 

T11/98-124v8 

REV 1.07 

NCITS working draft proposed 

Technical Report 

August 7, 1998 

Secretariat: 

Information Technology Industry Council 

ABSTRACT: 

NOTE: 

This is a draft proposed Technical Report of Accredited Standards Committee NCITS. As such, this 

is not a completed report. The T11 Technical Committee may modify this document as a result of 

comments received during public review and its approval as a technical report. 

POINTS OF CONTACT: 

Rich Taborek (T11.3 Chairman) 

G2 Networks, Inc. 

16780 Lark Avenue 

Los Gatos, CA 95032-7646 

Phone: (408) 399-3867 

Fax: (408) 399-3899 

EMail: rtaborek@g2networks.com 

Jeffrey Stai (T11.3 Vice Chairman) 

Brocade Communications Systems, Inc. 

15707 Rockfield Boulevard, Suite 215 

Irvine, CA 92618 

Phone: (714) 455-2908 

Fax: (714) 455-9287 

E-Mail: stai@brocadecomm.com 

Dale LaFollette 

(Fibre Channel Tape Facilitator) 

StorageTek 

2270 South 88th Street 

Louisville, CO 80028-4223 

Phone: (303) 673-8791 

Fax: (303) 673-2568 

E-Mail: dale_lafollette@stortek.com 

David A. Peterson 

(Fibre Channel Tape Technical Editor) 

StorageTek Network Systems Group 

7600 Boone Avenue North 

Brooklyn Park, MN 55428 

Phone: (612) 391-1008 

Fax: (612) 391-1095 

E-mail: dap@network.com

Editor’s Note: Release 1.00 Initial Document Differences from PLDA 

– Addresses only streaming and media changer device’s 

– Data Overlay Allowed (for sequence retransmission) 

– Error Detection and Recovery Procedure with new ELS’s added (REC/SRR) 

– FCP_CONF implementation and usage 

Editor’s Note: Release 1.01 Changes/Additions 

– Update of Annex H 

– Added Media Changer SCSI command table 

– Added diagrams to Annex G 


– Added FCP_TOV 

– Incorporated changes discussed in March-98 FC-Tape working group. 

– Added section describing Class 2-Public Loop operation 


– Updated section 8 (SCSI FCP Feature Set) 

– Added changes from comments received 

– Updated PLOGI and FLOGI tables 

Editor’s Note: Release 1.04a Changes/Additions 

– Added changes from comments received 

– Added FCP Command Reference Number text 

– Added SRR to Class 2 error recovery 

– Moved Class 2 diagrams to annex 

– Collapsed FCP Error Detection and Recovery (Class 2 and Class 3) 

– Added Fibre Channel Mode Page 0x19 to annex 


– Removed clause 10. 

– Combined error detection and recovery procedures into one annex (Class 2 and Class 3). 

– Updated Public and Private loop behavior. 

– Updated clause 9. 

– Removed annex D (RR_TOV proposal). 

Editor’s Note: Release 1.06b Changes/Additions 

– Changed error detection and recovery diagrams. 

– Added notes and device reservations to Table 23.


– Moved document towards Public loop operation. 

– Changed SRR ladder diagram. 

– Added ABTS changes text (to be expanded on).

ANSI ® 

NCITS TR-xx 

draft proposed NCITS Technical Report 

Fibre Channel — 

Tape (FC-TAPE) 

Secretariat 

Information Technology Industry Council 

Approved ,199 

American National Standards Institute, Inc. 

Abstract 

This profile selects and restricts logical options from the Fibre Channel Physical and Signalling, Fibre 

Channel Protocol for SCSI, Fibre Channel Arbitrated Loop, and Small Computer Systems Interface standards, 

such that any device complying with the profile should interoperate. 

iv

NCITS 

T11 

Technical 

Report 

Series 

This Technical Report is one in a series produced by the American National 

Standards Committee, NCITS, Information Technology. The secretariat for NCITS 

is held by the Computer and Business Equipment Manufacturers Association 

(CBEMA), 1250 Eye Street, NW Suite 200, Washington DC 20005. 

As a by-product of the standards development process and the resources of 

knowledge devoted to it, NCITS from time to time produces Technical Reports. 

Such Technical Reports are not standards, nor are they intended to be used as 

such. 

Patent 

Statement 

NCITS Technical Reports are produced in some cases to disseminate the 

technical and logical concepts reflected in standards already published or under 

development. In other cases, they derive from studies in areas where it is found 

premature to develop a standard due to a still changing technology, or 

inappropriate to develop a rigorous standard due to the existence of a number of 

viable options, the choice of which depends on the user's particular requirements. 

These Technical Reports, thus, provide guidelines, the use of which can result in 

greater consistency and coherence of information processing systems. 

When the draft Technical Report is completed, the Technical Committee approval 

process is the same as for a draft standard. Processing by NCITS is also similar to 

that for a draft standard. 

CAUTION: The developers of this Technical Report have requested that holder's 

of patents that may be required for the implementation of the Technical Report, 

disclose such patents to the publisher. However, neither the developers nor the 

publisher have undertaken a patent search in order to identify which, if any, patents 

may apply to this Technical Report. 

As of the date of publication of this Technical Report and following calls for the 

identification of patents that may be required for the implementation of the 

Technical Report, no such claims have been made. No further patent search is 

conducted by the developer or the publisher in respect to any Technical Report it 

processes. No representation is made or implied that licenses are not required to 

avoid infringement in the use of this Technical Report. 

Published by 

American National Standards Institute 

11 W. 42nd Street, New York, New York 10036 

Copyright © 199x by American National Standards Institute 

All rights reserved 

No part of this publication may be reproduced in any 

form, in an electronic retrieval system or otherwise, 

without prior written permission of the publisher. 

Printed in the United States of America 

v

NCITS TR-xx Fibre Channel Tape Rev 1.07 August 7, 1998 

Table of Contents 

1 Introduction and scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 

2 Normative references . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 

2.1 Approved references . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 

2.2 References under development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 

2.3 Other references . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 

3 Definitions and conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 

3.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 

3.2 Editorial conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 

3.2.1 Binary notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 

3.2.2 Hexadecimal notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 

3.3 Abbreviations, acronyms, and symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 

3.3.1 Acronyms and abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 

3.4 Applicability and use of this document . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 

3.5 Definitions and abbreviations used in Feature Set tables . . . . . . . . . . . . . . . . . . . . . . 7 

4 Structure and Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 

4.1 Public and Private Loop Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 

5 FC-2 feature sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 

5.1 Node and port naming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 

5.2 FLOGI and PLOGI features and parameters for NL_Ports . . . . . . . . . . . . . . . . . . . . 13 

5.3 NL_Port and N_Port Common Service Parameters (PLOGI) . . . . . . . . . . . . . . . . . . 14 

5.4 NL_Port Common Service Parameters (FLOGI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 

5.5 NL_Port and N_Port Class 2 Service Parameters (PLOGI) . . . . . . . . . . . . . . . . . . . . 16 

5.6 NL_Port Class 2 Service Parameters (FLOGI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 

5.7 NL_Port and N_Port Class 3 Service Parameters (PLOGI) . . . . . . . . . . . . . . . . . . . . 18 

5.8 NL_Port Class 3 Service Parameters (FLOGI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 

5.9 Other FC-2 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 

vii


5.10 Basic Link Service commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 

5.11 Extended Link Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 

5.12 Responses to Link Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 

5.13 Exchange and Sequence management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 

5.13.1 Exchange Originator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 

5.13.2 Exchange Responder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 

5.13.3 Sequence management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 

5.13.4 Sequence errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 

6 FC-AL feature set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 

6.1 FC-AL features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 

6.2 Alternate Login_BB_Credit management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 

6.3 Loop Initialization features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 

6.3.1 Initializing LIP(F7,F7) and LIP(F7,AL_PS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 

6.3.2 Selective Hard Reset LIP(AL_PD,AL_PS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 

6.3.3 Loop Failure LIP(F8,AL_PS) and LIP(F8,F7) . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 

6.3.4 Failure to obtain an AL_PA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 

7 Timers on Public and Private Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 

7.1 Arbitrated Loop Time (AL_TIME) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 

7.2 Loop Initialization Sequence Hold Time (LIS_HOLD_TIME) . . . . . . . . . . . . . . . . . . . 31 

7.3 Receiver_Transmitter Timeout (R_T_TOV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 

7.4 Error_Detect Timeout (E_D_TOV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 

7.5 Resource Allocation Timeout (R_A_TOV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 

7.6 Resource Recovery Timer (RR_TOV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 

7.7 Fibre Channel Protocol Timeout Value (FCP_TOV) . . . . . . . . . . . . . . . . . . . . . . . . . 33 

7.8 Upper Level Protocol Timeout (ULP_TOV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 

8 SCSI-FCP Feature Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 

8.1 Process Login parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 

viii


8.2 FCP Information Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 

8.2.1 FCP_CMND IU (T1/T3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 

8.2.2 FCP_XFER_RDY IU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 

8.2.3 FCP_DATA IU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 

8.2.4 FCP_RSP (I4/I5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 

8.2.5 FCP_CONF (T12) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 

8.3 Task Management Flags and Information Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 

8.4 FCP Task Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 

8.5 Other FCP features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 

8.6 FCP Sequence delivery confirmation in a Class 3 environment . . . . . . . . . . . . . . . . . 41 

8.7 FCP Sequence delivery confirmation in a Class 2 environment . . . . . . . . . . . . . . . . . 42 

9 FCP Error Detection and Recovery Procedure(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 

9.1 FCP Error Detection for Class 3 Delivery Service . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 

9.2 FCP Error Detection for Class 2 Delivery Service . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 

9.3 FCP Error Detection and Recovery (Initiator, Class 2 and Class 3) . . . . . . . . . . . . . . 44 

9.3.1 Error Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 

9.3.2 FCP_CMND Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 

9.3.3 FCP_XFER_RDY Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 

9.3.4 FCP_RSP Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 

9.3.5 FCP_DATA Recovery - Write . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 

9.3.6 FCP_DATA Recovery - Read . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 

9.3.7 Additional error detection by SCSI Initiator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 

9.3.8 Error detection by SCSI Target . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 

9.4 FCP Error Recovery (Target, Class 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 

9.5 Aborting an Exchange Using ABTS Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 

9.5.1 SCSI Initiator ABTS of Exchange behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 

9.5.2 SCSI Target ABTS of Exchange behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 

ix


9.5.3 Second-level error recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 

9.5.4 Abort Sequence (ABTS) Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 

9.5.5 Basic Accept (BA_ACC) Frame to ABTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 

9.5.6 Basic Reject (BA_RJT) Frame to ABTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 

9.5.7 Reinstate Recovery Qualifier (RRQ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 

9.6 SCSI Target error behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 

9.7 Task Management and Multiple-Initiator SCSI Targets . . . . . . . . . . . . . . . . . . . . . . . 51 

9.8 SCSI Target Exchange origination capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 

9.9 Responses to FCP-level Frames before PLOGI or PRLI . . . . . . . . . . . . . . . . . . . . . . 51 

10 SCSI features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 

10.1 Auto Contingent Allegiance (ACA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 

10.2 SCSI Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 

10.3 SCSI Target Discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 

10.4 Exchange Authentication following LIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 

10.4.1 SCSI Initiator Exchange Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 

10.4.2 SCSI Target Exchange Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 

10.5 Clearing effects of ULP, FCP, FC-PH, and FC-AL actions . . . . . . . . . . . . . . . . . . . 56 

11 SCSI Stream Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 

11.1 Applicable Classes of Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 

11.2 Asynchronous Event Notification (AEN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 

11.3 Command Linking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 

11.4 Sequential device commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 

12 SCSI-3 Media Changer Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 

12.1 Applicable Classes of Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 

12.2 Asynchronous Event Notification (AEN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 

12.3 Command Linking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 

12.4 Media Changer Device Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 

x


Annex A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 

A.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 

A.2 Sequence Retransmission Request (SRR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 

A.3 Read Exchange Concise (REC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 

A.4 ABTS Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 

Annex B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 

Annex C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 

C.1 FCP_CONF Information Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 

C.2 FCP_CONF Process Login (PRLI) Parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 

C.2.1 Word 3, Bit 7: FCP_CONF Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 

C.3 FCP_CONF Request Returned in FCP_RSP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 

C.4 FCP_CONF Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 

Annex D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 

D.1 Command Reference Number Process Login (PRLI) Parameter . . . . . . . . . . . . . . . 87 

D.1.1 Word 3, Bit 8: Command Reference Number Support (CRNS) . . . . . . . . . . . . . . 87 

D.2 FC Mode Page (0x19) Parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 

D.2.1 Byte 4, Bit 0: Enable Command Reference Number (ECRN) . . . . . . . . . . . . . . . 87 

D.3 Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 

Annex E . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 

xi


xii


List of Figures 

1. Public and Private NL_Port Coexistence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 

2. FCP Read/Write IU Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 

3. FCP 8-byte LUN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 

4. ABTS Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 

5. BA_ACC Frame to ABTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 

6. BA_RJT Frame to ABTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 

7. Reinstate Recovery Qualifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 

8. Lengthy FCP_CMND or Lost ACK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 

9. FCP_CMND Lost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 

10. FCP_XFER_RDY Lost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 

11. FCP_XFER_RDY Received, ACK Lost (Class 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 

12. FCP_RSPLost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 

13. FCP_RSP Received, ACK Lost (Class 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 

14. Lost Write Data, Last Frame of Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 

15. Lost Write Data, Not Last Frame of Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 

16. Lost Read Data, Last Frame of Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 

17. Lost Read Data, Not Last Frame of Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 

18. ACK Lost on Read (Class 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 

19. REC or REC Response Lost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 

20. SRR or SRR Response Lost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 

21. FCP_CONF Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 

22. FCP_CONF Lost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 

xiii


xiv


List of Tables 

1. Public NL_Port Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 

2. Private NL_Port Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 

3. PDISC, ADISC, ACC Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 

4. FLOGI and PLOGI features and parameters for NL_Ports . . . . . . . . . . . . . . . . . . . . . . . 13 

5. NL_Port and N_Port Common Service Parameters (PLOGI) . . . . . . . . . . . . . . . . . . . . . 14 

6. NL_Port and N_Port Common Service Parameters (FLOGI) . . . . . . . . . . . . . . . . . . . . . 15 

7. Class 2 Service Parameters (PLOGI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 

8. Class 2 Service Parameters (FLOGI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 

9. Class 3 Service Parameters (PLOGI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 

10. Class 3 Service Parameters (FLOGI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 

11. Other FC-2 Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 

12. Basic Link Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 

13. Extended Link Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 

14. Responses to Selected Link Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 

15. FC-AL Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 

16. Timer Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 

17. PRLI Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 

18. FCP_RSP Payload . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 

19. Task Management Function RSP_CODES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 

20. FCP Task Management Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 

21. FCP Task Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 

22. Other FCP Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 

23. Clearing Effects of SCSI Initiator Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 

24. SCSI Tape Device Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 

25. SCSI Independent Media Changer Device Commands . . . . . . . . . . . . . . . . . . . . . . . . 62 

26. SCSI Attached Media Changer Device Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 

xv


27. ABTS Parameter Field Bit Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 

28. FCP information units sent to targets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 

29. FCP_STATUS field format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 

30. Fibre Channel Control Page (19h). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 

xvi

draft proposed NCITS Technical Report 

for Information Technology— 

Fibre Channel — 

Tape (FC-Tape) 

TR-xx-1998 Fibre Channel Tape Rev 1.07 August 7, 1998 

1 Introduction and scope 

This Technical Report specifies Fibre Channel, SCSI-3 Fibre Channel Protocol for SCSI (FCP), and 

SCSI-3 command set options required for operation of streaming devices and medium changers in a 

public arbitrated loop environment. Devices that operate in a public arbitrated loop environment are 

by definition capable of operating in a private loop environment. This Technical Report is intended to 

serve as a guide whose primary objective is to maximize the likelihood of interoperability between 

conforming implementations. This guide prohibits and requires features that are optional as well as 

prohibiting the use of some mandatory features in the referenced ANSI standards. 

A second objective of this guide is to simplify implementations and their associated documentation, 

testing, and support requirements. This means that there will be some optional features which are not 

mutually exclusive, but are still prohibited or required solely for the purpose of this simplification. 

This guide is based on SCSI-3 command sets mapped to FCP using the FC-AL topology. Internal 

characteristics of conformant implementations are not defined by this document. This document incorporates 

features from the standards identified in clause 2. Where needed, changes have been 

proposed to the appropriate NCITS standards to ensure this document remains a strict subset of 

ANSI standards. 

1


2 Normative references 

The following standards contain provisions which, through reference in the text, constitute provisions 

of this document. At the time of publication, the editions indicated were valid. All standards are subject 

to revision, and parties to agreements based on this document are encouraged to investigate the 

possibility of applying the most recent editions of the standards listed below. 

Copies of the following documents can be obtained from ANSI: Approved ANSI standards, approved 

and draft international standards (ISO, IEC, CEN/CENELEC), and approved foreign standards (including 

BSI, JIS, and DIN). For further information, contact ANSI Customer Service Department at 

212-642-4900 (phone), 212-302-1286 (fax). 

Additional availability contact information is provided below as needed. 

2.1 Approved references 

[1] ANSI X3.230-1994, Information technology Fibre channel physical and signaling interface (FC- 

PH) including ANSI X3.230-1994/AM-1-1996 ammendment. 

[2] ANSI X3.269-1996, Information technology Fibre channel protocol for SCSI (FCP) 

[3] ANSI X3.272-1996, Information technology Fibre channel arbitrated loop (FC-AL) 

[4] ANSI X3.270-1996, Information technology SCSI-3 architecture model (SAM) 

[5] ANSI X3.297-1996, Fibre Channel - Physical and Signalling Interface-2 (FC-PH-2), X3T11/Project 

901D/Rev 7.4 

2.2 References under development 

At the time of publication, the following referenced standards were still under development. For information 

about obtaining copies of the following documents or for more information on the current status 

of these documents, contact NCITS at http://www.ncits.org or 202-626-5738. 

[6] ANSI X3.xxx-199x, SCSI-3 Primary Commands (SPC), X3T10/995D/Rev 11a 

[7] ANSI X3.xxx-199x, SCSI-3 Block Commands (SBC), X3T10/996D/Rev 8b 

[8] ANSI X3.xxx-199x, Fibre Channel - Physical and Signalling Interface-3 (FC-PH-3), 

X3T11/Project1119D/Rev 9.0 

[9] ANSI X3.xxx-199x, SCSI-3 Stream Device Command Set (SSC), X3T10/Project 997D/Rev x 

[10] ANSI X3.xxx-199x, SCSI-3 Media Changer Command Set (SMC), X3T10/Project 999D/Rev 10a 

[11] ANSI X3.xxx-199x, SCSI Fibre Channel Protocol-2, X3T10/Project 1144D/Rev x 

[12] ANSI X3.xxx-199x, Fibre Channel - Private Loop Direct Attach (FC-PLDA), T11/Project 1162- 

DT/Rev 2.1 

[13] ANSI TR.xx-199x, Fibre Channel - Fabric Loop Attach (FC-FLA), T11/Project xxxx-xx/Rev 2.7 

[14] ANSI X3.xxx-199x, SCSI Architecture Model - 2 (SAM-2), T10/Project 1157-D Rev x 

2


[15] ANSI X3.272-199x, Fibre Channel - Arbitrated Loop (FC-AL-2), T11/Project 1133D Rev x.x 

2.3 Other references 

[16] FCSI Common FC-PH Feature Sets Used in Multiple Profiles, Rev 3.1 

[17] FCSI SCSI Profile, Rev 2.2 

[18] SFF-8045 Specification for 40-pin SCA-2 Connector w/Parallel Selection, Rev 3.3 

3


3 Definitions and conventions 

The following definitions, conventions, abbreviations, acronyms, and symbols apply to this document. 

3.1 Definitions 

3.1.1 Available BB_Credit: A variable used by a sequence initiator to determine permission to 

transmit Class 2 and Class 3 frames. The transmitter may transmit frames when the Available 

BB_Credit is greater than zero. The rules for modifying Available BB_Credit are defined by FC-AL. 

3.1.2 Available receive buffers: The number of buffers in a receiving port which are available for 

receiving frames at link rate. 

3.1.3 Byte: A group of eight bits. 

3.1.4 Hard Address: The AL_PA which an NL_Port attempts to acquire in the LIHA Loop 

Initialization Sequence. 

3.1.5 Login_BB_Credit: On FC-AL, equal to the number of receive buffers that a receiving 

NL_Port must have available when a loop circuit is established. Login_BB_Credit is established 

using PLOGI and may be discovered by using the PDISC Extended Link Service. 

3.1.6 Loop Circuit: A bidirectional path that allows communication between two L_Ports. 

3.1.7 Loop_ID: Loop_IDs are 7-bit values numbered contiguously from 0 to 126 and represent the 

127 valid AL_PAs on a loop. Loop_IDs correspond to the 7-bit SEL word in SFF-8045 [3] used for 

designating a Hard Addresses that the NL_Port attempts to acquire during the LIHA Loop 

Initialization Sequence. See FC-AL, Annex K (Assigned Loop Identifier) for a complete mapping. 

3.1.8 Native Address Identifier: A 24-bit value divided into three 8-bit fields used to identify an 

NL_Port. The order of bit significance being: Domain (bits 23-16), Area (bits 15-8), Port (bits 7-0). 

3.1.9 Node: An entity containing one or more N_Ports or NL_Ports controlled by a level above FC- 

2. 

3.1.10 OPN Originator: The L_Port on an Arbitrated Loop that enters the OPEN state. 

3.1.11 OPN Recipient: The L_Port on an Arbitrated Loop that enters the OPENED state. 

3.1.12 Previously Acquired Address: During loop initialization, this is the AL_PA value which an 

L_Port attempts to acquire during the LIPA sequence. 

3.1.13 Private NL_Port: An NL_Port which is observing the rules of private loop behavior (see 

reference). 

3.1.14 Public NL_Port: An NL_Port which attempts a fabric login and is permitted to open 

AL_PA=’00’h. A Public NL_Port can observe the rules of either public or private loop behavior (see 

reference). 

3.1.15 ULP process: A function executing within an FC node that conforms to Upper Level 

Protocol (ULP) defined protocols when interacting with ULP processes residing in other FC nodes. 

4


3.2 Editorial conventions 

A number of conditions, mechanisms, sequences, parameters, events, states, or similar terms are 

printed with the following conventions. Each of these terms are explicitly defined in the glossary, in 

the appropriate text, or in the indicated reference and has a special meaning within this document. 

– the first letter of each word in uppercase and the rest lowercase (e.g., Exchange, Class, etc.). 

– a term consisting of multiple words, with the first letter of each word in uppercase and the rest 

lowercase, and each word separated from the other by an underscore (_) character. A word 

may consist of an acronym or abbreviation, which would be printed in uppercase. (e.g., 

NL_Port, Transfer_Length, etc.). 

– a term consisting of multiple words with all letters lowercase and each word separated form the 

other by a dash (-) character. A word may also consist of an acronym or abbreviation, which 

would be printed in uppercase. (e.g., buffer-to-buffer, etc.). 

All terms and words not conforming to the conventions noted above have the normal English meanings. 

Numbered items in this document do not represent any priority. Any priority is explicitly described. 

In the figures, tables, and text, the most significant bit of a binary quantity is shown on the left side. 

Exceptions to this convention are indicated in the appropriate clauses. 

The term ‘shall’ is used to indicate a mandatory rule that must be followed to comply with this profile. 

The fields or control bits which are not applicable shall be set by the sending NL_Port and checked 

by the receiving NL_Port as required by the appropriate standard. 

If a field or a control bit in a frame is specified as not meaningful, the entity which receives the frame 

shall not check that field or control bit. 

In several tables within this document, there is a column on the right side of the table labelled 

“Notes”. These notes are Normative and are mandatory requirements of this document. 

In the event of conflicts between the text, tables, and figures in this document, the following precedence 

shall be used: text, tables, figures. 

3.2.1 Binary notation 

Binary notation may be used to represent some fields. Single bit fields are represented using the binary 

values 0 and 1. For multiple bit fields, the binary value is enclosed in single quotation marks followed 

by the letter b. For example, a four-byte Process_Associator field containing a binary value 

may be represented as ‘00000000 11111111 10011000 11111010’b. 

3.2.2 Hexadecimal notation 

Hexadecimal notation may be used to represent some fields. When this is done, the value is enclosed 

in single quotation marks preceded by the word hex. For example, a four-byte 

Process_Associator field containing a binary value of ‘00000000 11111111 10011000 11111010’b is 

shown in hexadecimal format as hex ‘00 FF 98 FA’. 

5


3.3 Abbreviations, acronyms, and symbols 

Abbreviations and acronyms applicable to this International Standard are listed. Definitions of several 

of these items are also found in 3.1. 

3.3.1 Acronyms and abbreviations 

BB 

Buffer-to-buffer 

ELS 

Extended Link Service 

FC 

Fibre Channel 

FC-PH 

ANSI X3.230-1994, Fibre Channel Physical and Signaling Interface (FC-PH). 

See clause 2.1 for a list of approved references. 

FC-PH-2 

ANSI X3.297-1996, Fibre Channel Physical and Signaling Interface-2 (FC-PH- 

2). See clause 2.2 for a list of references under development. 

FC-PH-3 

ANSI X3.xxx-199x, Fibre Channel Physical and Signaling Interface-3 (FC-PH- 

3). See clause 2.2 for a list of references under development. 

FCP 

Fibre Channel Protocol (see clause 2.1 for a list of approved references) 

FCP-2 

Fibre Channel Protocol-2 

IP 

Internet Protocol 

IU 

Information Unit 

FRU 

Field Replaceable Unit 

LAN 

Local Area Network 

LLC 

Logical Link Control 

MAC 

Media Access Control 

NFS 

Network File System or Network File Server 

SAN 

Storage Area Network 

SBC ANSI X3.xxx-199x, SCSI-3 Block Commands (SBC), X3T10/996D/Rev 4 

SCC 

ANSI X3.xxx-199x, SCSI-3 Controller Commands (SCC), X3T10/1047D Rev 6a 

SMC ANSI X3.xxx-199x, SCSI-3 Media Changer Command Set (SMC), 

X3T10/Project 999D/Rev 5 

SPC 

ANSI X3.xxx-199x, SCSI-3 Primary Commands (SPC), X3T10/995D/Rev 9b 

SSC 

ANSI X3.xxx-199x, SCSI-3 Stream Device Command Set (SSC), X3T10/Project 

997D/Rev 7 

SCSI 

Small Computer System Interface-3 

SPC 

SCSI-3 Primary Commands 

SSC 

SCSI-3 Streaming Commands 

ULP 

Upper Level Protocol 

WAN 

Wide Area Network 

WWN 

World Wide Name 

3.3.2 Symbols 

Unless indicated otherwise, the following symbols have the listed meaning. 

|| concatenation 

3.4 Applicability and use of this document 

This document specifies which features shall be used (“required”) and which features shall not be 

used (“prohibited”) by interoperating compliant Fibre Channel implementations. Use of some features 

is optional (“allowed”); these features may be used but compliant implementations are not required to 

do so. The only features or functions required or prohibited by this document are those which have 

been determined to affect interoperability. 

6


The relationship between use (specified in this document) and support (implemented by a product) is 

subtle. If this document specifies that a feature must be used, then a compliant adapter must support 

it. In some cases, the specification is asymmetric: to ensure interoperability when an optional feature 

is used, this document mandates support for the infrastructure required to use the feature without 

specifying that the feature actually be used when conforming to this profile. 

The requirements of this document are a proper subset of the various relevant standards. They prohibit 

use of many features and options in these standards; and interoperability is not guaranteed if 

these features are used. This document does not prohibit implementation of features, only their use. 

Functions that are mandatory in the appropriate base standard are assumed to be implemented. Implementations 

may support features whose use is prohibited by this document and such prohibited 

features may be required for compliance with the relevant standards or other Profiles or applications 

not covered by this document. 

In the event of conflicts between this profile and other profile documents, the resolution of those conflicts 

is beyond the scope of this document. 

3.5 Definitions and abbreviations used in Feature Set tables 

The following definitions are used to define usage of reference features or options provided by the 

applicable standards. 

Prohibited: If a feature is Prohibited, it means that it shall not be used between FC-TAPE compliant 

implementations. An implementation may use the feature to communicate with non-compliant implementations. 

This document does not prohibit the implementation of features, only their use between 

FC-TAPE compliant implementations. However, interoperability is not guaranteed if Prohibited features 

are used when communicating with devices conforming to this profile. 

Required: If a feature or parameter value is Required, it means that it shall be used between all FC- 

TAPE compliant implementations. FC-TAPE compliant implementations are required to implement 

the feature. An implementation may use the feature to communicate with non-compliant implementations. 

Interoperability with devices conforming to this profile is not guaranteed if Required features 

are not implemented. 

Allowed: If a feature or parameter value is Allowed, it means that it may be used between FC-TAPE 

compliant implementations. FC-TAPE compliant implementations are not required to implement the 

feature, but if they do, the feature shall be implemented as described in the applicable standard. Typically, 

the potential user of a feature may determine if the potential recipient supports that feature via 

a Required discovery process or a minimal response by the recipient. 

Invokable: If a feature or parameter value is Invokable, it means that it may be used between FC- 

TAPE compliant implementations. FC-TAPE compliant implementations are required to implement 

the feature, and make available the use of the feature. Invokable is different than Allowable or Required 

in that an originator may invoke the feature if needed, but the originator is not required to invoke 

it, and may never need to. The recipient shall support Invokable features or provide a response 

that it is not implemented as defined by the appropriate standard. In each case, this profile shall state 

whether the recipient is Required to support the feature or function. 

Features in this document are summarized in the form of Feature Set tables. These tables indicate 

whether the feature is Required, Prohibited, Invokable, or Allowed for FC-TAPE compliance with this 

report; or whether a parameter is Required to be a particular value for compliance with this document. 

Features or parameters that are not listed do not affect interoperability of Private NL_Ports. 

The following legend is used for table entries in these clauses: 

7


‘P’ the implementation is Prohibited from using the specified feature when communicating with a 

FC-TAPE compliant implementation. 

‘R’ the implementation is Required to support the specified feature when communicating with a FC- 

TAPE compliant implementation. 

‘A’ use of the specified feature is Allowed when communicating with a FC-TAPE compliant implementation. 

‘I’ the implementation may Invoke the specified feature when communicating with a FC-TAPE 

compliant implementation. 

‘X’ 

this parameter has no required value; any value is allowed 

‘-’ this parameter or feature is not meaningful 

A blank entry in a Feature Set table indicates that the feature is not part of that Feature Set. 

For many features, explanatory text is provided in the form of notes following each Feature Set table. 

The “Notes” column in each Table contains the reference number of the note applying to an individual 

feature or group of features. 

8


4 Structure and Concepts 

This clause provides an overview of Public and Private loop topology and associated behavior. 

4.1 Public and Private Loop Behavior 

When an NL_Port is operating in an Arbitrated Loop topology as defined by FC-AL and successfully 

completes a Fabric Login (FLOGI) it is called a Public NL_Port. Public NL_Ports whose Fabric Login 

fails revert to Private NL_Port behavior. 

A port which only exhibits Private Loop behavior is called a Private NL_Port. A port that exhibits Public 

Loop behavior is called a Public NL_Port, even though it may communicate with Private NL_Ports. 

A Public NL_Port may have concurrent open Exchanges with Private NL_Ports on the same loop, 

Public NL_Ports on the same loop, and N_Ports or NL_Ports external to the loop. A Private NL_Port 

may have concurrent open Exchanges with other Private and Public NL_Ports on the same loop. 

Table 1 – Public NL_Port Behavior 

Behavior Public NL_Port Notes 

Domain + Area of device’s NL_Port ID = hex ‘0000’ 

Prohibited 

NL_Port may open AL_PA = hex ‘00’ 

Required 

NL_Port may respond to AL_PA = hex ‘00’ 

(NL_Port operating as an F/NL_Port) 

Prohibited 1 

FL_Port may Open NL_Port 

Required 

F/NL_Port may Open NL_Port Prohibited 1 

Public NL_Port may Open any Local NL_Port Allowed 2 

Public NL_Port may be Opened by any Local NL_Port Allowed 3 

NOTES: 

1 A F/NL_Port does not operate in this environment. 

2 The Public NL_Port shall set the D_ID = hex’0000’||AL_PA to address a Local Private NL_Port. 

3 The Public NL_Port shall respond to a S_ID = hex’0000’||AL_PA. 

Table 2 – Private NL_Port Behavior 

Behavior Private NL_Port Notes 

NL_Port may open AL_PA = hex‘00’ 

Prohibited 

Domain + Area of NL_Port ID = hex‘0000’ 

Required 

FL_Port may Open NL_Port 

Prohibited 

Private NL_Port may Open any Local NL_Port 

Allowed 

Private NL_Port may be Opened by any Local NL_Port Allowed 1 

NOTES: 

1 The Private NL_Port shall set the D_ID = hex’0000’||AL_PA to address a Local PublicNL_Port. 

9

FC-Tape Fibre Channel Tape Rev 1.07 August 7, 1998 

Figure 1 — Public and Private NL_Port Coexistence 

NFS Client 

F 

FABRIC 

Public Loop 

Device 

SCSI Target 

Public Loop 

Device 

SCSI Initiator 

F 

FL 

SCSI Target 

SCSI Target 

IP Host 

F 


Public Loop 

Device 

Private Loop 

Device 

Public Loop 

Device 

IP Host 

Public Loop 

Device 

FL 


(NFS Server) 

Public Loop 

Device 

IP Host 

SCSI Target 

Public Loop 

Device 

For example, a Public NL_Port in Figure 9.3 may be a Network File System (NFS) server that communicates 

with NFS Clients accessible via the fabric using Internet Protocol (IP), and with SCSI Targets 

on the same loop using the SCSI-3 FCP protocol. 

10


5 FC-2 feature sets 

The tables in this clause list features described in FC-PH, FC-PH-2, and FC-PH-3. These tables indicate 

whether the feature is Required, Prohibited, Allowable, or Invokable for compliance with this 

specification. Features that are not listed do not affect interoperability of Private NL_Ports. 

Reserved FC-PH, FC-PH-2, and FC-PH-3 fields are processed in accordance with the applicable 

standard or standards. Validity bits set to zero remove any requirement to check the corresponding 

field for zeroes and make the associated field meaningless (e.g., if F_CTL bit 3=0 in an FCP_DATA 

Sequence at a receiving NL_Port, the Relative Offset in the parameter field of the Frame header shall 

not be used in any way when reassembling the Sequence at the recipient). 

5.1 Node and port naming 

All SCSI Targets and Initiators complying with this profile shall have a unique Node_Name (WWN). 

Each FC-TAPE compliant NL_Port associated with these nodes shall also have a unique Port_Name 

(WWN). To ensure uniqueness of these names, all names shall use one of the registered name formats 

such as the IEEE or IEEE extended formats defined by FC-PH-3. 

All NL_Ports must acquire a valid AL_PA and Native Address Identifier before performing PLOGI with 

other NL_Ports. 

NL_Ports must retain the Node_Name and Port_Name of the other port from each PLOGI and associate 

that information with the Native Address Identifier of that NL_Port. This information shall be retained 

for as long as the PLOGI with the other port is active. 

SCSI Targets and SCSI Initiators shall validate current NL_Port logins following every Loop Initialization 

by comparing the Port Name, Node Name, and N_Port ID received during the PLOGI with those 

reported by the PDISC or ADISC that follows the Loop Initialization. All three identifiers reported during 

the PDISC or ADISC shall match the values reported during PLOGI or a configuration change has 

occurred and LOGO is required and all open Exchanges with that SCSI Initiator or SCSI Target will 

be terminated. 

Consider two examples os a DISC or ADISC request arriving at a SCSI Target, or ACC arriving at a 

SCSI Initiator in response to a PDISC or ADISC. 

Table 3 – PDISC, ADISC, ACC Examples 

Example Identifier PLOGI Data ADISC, PDISC, or ACC 

Data 

1 

2 

N_Port ID (24 bits) 00 00 AA 00 00 AB 

Node Name (64 bits) 1000 ABCD EF12 3456 1000 ABCD EF12 3456 

Port Name (64 bits) 1000 ABCD EF12 3457 1000 ABCD EF12 3457 

N_Port ID (24 bits) 0000AA 0000AA 

Node Name (64 bits) 1000 ABCD EF12 3456 1000 ABCD EF12 3456 

Port Name (64 bits) 1000 ABCD EF12 3456 1000 ABCD EF98 7654 

Response 

LOGO 

LOGO 

Example 1 may occur if a node with no Hard Address has been removed and reinserted into a reconfigured 

loop, and takes a different soft address. 

11


Example 2 could occur if both ports of a node are connected to the same loop with the second port 

non-participating. The first port failed and was bypassed, and the second port assumed the AL_PA of 

the (now bypassed) first port. 

These examples, and other situations, such as loop reconfigurations or the connection of two formerly 

independent loops into a single loop, could result in one or more of the identifiers changing. 

12

5.2 FLOGI and PLOGI features and parameters for NL_Ports 


Table 4 lists features and parameters for FLOGI and PLOGI with usage defined by this document. 

Table 4 – FLOGI and PLOGI features and parameters for NL_Ports 

Feature/Parameter 

NL_Port NL_Port Notes 

Originator Responder 

FLOGI S_ID hex’0000’||AL_PA - 

Class of service for FLOGI 1 

Class 2 I A 

Class 3 I A 

Class of service for PLOGI 1 

Class 2 I A 

Class 3 I A 

Supported Classes of Service 1 

Class 1 P P 

Class 2 I A 

Class 3 I A 

Class 4 and up P P 

NOTES: 

1 Implementations conforming to this profile may support Class 2 and/or Class 3. 

13


5.3 NL_Port and N_Port Common Service Parameters (PLOGI) 

Table 5 lists NL_Port and N_Port Common Service Parameters for PLOGI with usage defined by this 

document. The parameters are valid for both Class 2 and Class 3 delivery service. 

Table 5 – NL_Port and N_Port Common Service Parameters (PLOGI) 

Common Service Parameter 

SCSI SCSI Notes 

Initiator Target 

FC-PH Version 

Highest Version X X 1 

Lowest Version hex’20’ hex’20’ 

Buffer-to-Buffer Credit (min) 0 0 4 

Common Features 

Continuously Increasing Relative Offset 1 1 

Random Relative Offset 0 0 

Valid Vendor Version Level 0 0 2 

N_Port/F_Port 0 0 

Alternate BB_Credit Management 1 1 4 

E_D_TOV Resolution 0 0 4 

Dynamic Half Duplex - DHD A A 

SEQ_CNT A A 

Payload Length 0 0 

Buffer-to-Buffer Receive Data Field Size (min) 256 256 4 

Total Concurrent Sequences (min) 1 1 

Relative Offset by Information Category 

Information Category 1 (Solicited Data) R R 

Information Category 5 (Data Descriptor) R R 

All other Information Categories A A 3 

R_A_TOV - - 

E_D_TOV see 7.4 see 7.4 4 

NOTES: 

1 The FC-PH version shall include a version of FC-PH-2 or FC-PH-3 that includes PDISC, ADISC, TPR- 

LO, and RNC (previously known as RVU). PDISC, ADISC, and TPRLO are described in FC-PH-2, RNC 

is described in FC-PH-3. Hex‘20’ is the version assigned to FC-PH-3. 

2 Profile versions are communicated via the Report Node Capabilities (RNC) ELS. The Valid Vendor Version 

Level shall not be set to 1 for communicating support of this Technical Report. 

3 Relative offset is not required for any other Information Categories. The Sequence Initiator is allowed to 

set the bit but the Sequence Recipient may ignore it. 

4 Field does not apply to a direct Fabric-attached device. 

14

5.4 NL_Port Common Service Parameters (FLOGI) 


Table 5 lists NL_Port Common Service Parameters for FLOGI with usage defined by this document. 

The parameters are valid for both Class 2 and Class 3 delivery service. 

Table 6 – NL_Port and N_Port Common Service Parameters (FLOGI) 

Common Service Parameter 

NL_Port Notes 

Originator 

FC-PH 

Highest Version X 1 

Lowest Version 

hex’20’ 

Buffer-to-Buffer Credit (min) 0 

Common Features 

Continuously Increasing Relative Offset - 

Random Relative Offset - 

Valid Vendor Version Level 0 2 

N_Port/F_Port 0 

Alternate BB_Credit Management 1 

E_D_TOV Resolution - 

Dynamic Half Duplex - DHD 

A 

SEQ_CNT - 

Payload Length 0 

Buffer-to-Buffer Receive Data Field Size (min) 256 

Total Concurrent Sequences (min) - 

Relative Offset by Information Category - 

R_A_TOV - 

E_D_TOV - 

NOTES: 

1 The FC-PH version shall include a version of FC-PH-2 or FC-PH-3 that includes PDISC, ADISC, 

TPRLO, and RNC (previously known as RVU). PDISC, ADISC, and TPRLO are described in FC- 

PH-2, RNC is described in FC-PH-3. Hex‘20’ is the version assigned to FC-PH-3. 

2 Profile versions are communicated via the Report Node Capabilities (RNC) ELS. The Valid Vendor 

Version Level shall not be set to 1 for communicating support of this Technical Report. 

15


5.5 NL_Port and N_Port Class 2 Service Parameters (PLOGI) 

Table 7 lists Class 2 Service Parameters for PLOGI with usage defined by this document. 

Table 7 – Class 2 Service Parameters (PLOGI) 

Class 2 Service Parameter 



Class validity 1 1 

Service Options 

Intermix Mode - - 

Stacked Connect Requests - - 

Sequential Delivery - - 

Dedicated Simplex - - 

Camp-On - - 

Buffered Class 1 - - 

Priority 0 0 

Initiator Control 

X_ID reassignment ‘00’b ‘00’b 

Initial Responder Process_Associator ‘00’b ‘00’b 

ACK_0 capable A A 

ACK_N capable 0 0 

ACK generation assistance A A 

Data compression capable 0 0 

Data compression History buffer size - - 

Data encryption capable 0 0 

Clock synchronization capable A A 

Recipient Control 

ACK_0 capable A A 

ACK_N capable 0 0 

X_ID interlock R A 1 

Error Policy Supported 

Abort, discard multiple Sequences (‘00’b) A A 

Abort, discard a single Sequence (’01’b) R R 

Process policy with infinite buffers (‘10’b) P P 

Discard multiple sequences w/immediate retrans (‘11’b) P P 

Categories per Sequence ‘00’b ‘00’b 



Data decryption capable 0 0 


Reserved - Fabric specific 0 0 

Receive data field size (min) 256 256 

Concurrent Sequences (min) 1 1 

N_Port End-to-end Credit (min) 1 1 

Open Sequences per Exchange (min) 1 1 

NOTES: 

1 The Sequence Initiator shall provide X_ID Interlock if required by the Sequence Recipient. If an RX_ID 

other than hex ‘FFFF” is used by an Exchange Responder, it shall return the RX_ID value in the first Sequence 

transmitted as a Sequence Initiator. 

16

5.6 NL_Port Class 2 Service Parameters (FLOGI) 


Table 8 lists Class 2 Service Parameters for FLOGI with usage defined by this document. 

Table 8 – Class 2 Service Parameters (FLOGI) 


SCSI 

Initiator 

SCSI 

Target 





Sequential Delivery 1 1 


Camp-On - - 


Priority 0 0 


X_ID reassignment - - 

Initial Responder Process_Associator - - 

ACK_0 capable - - 

ACK_N capable - - 

ACK generation assistance - - 

Data compression capable - - 


Data encryption capable - - 





X_ID interlock - - 


Abort, discard multiple Sequences (‘00’b) - - 

Abort, discard a single Sequence (’01’b) - - 

Process policy with infinite buffers (‘10’b) - - 

Discard multiple sequences w/immediate retrans (‘11’b) - - 

Categories per Sequence - - 



Data decryption capable - - 



Receive data field size (min) - - 

Concurrent Sequences (min) - - 

N_Port End-to-end Credit (min) - - 

Open Sequences per Exchange (min) - - 

NOTES: 

Notes 

17


5.7 NL_Port and N_Port Class 3 Service Parameters (PLOGI) 

Table 9 lists Class 3 Service Parameters for PLOGI with usage defined by this document. 

Table 9 – Class 3 Service Parameters (PLOGI) 


SCSI 

Initiator 

SCSI 

Target 





Sequential Delivery - - 


Camp-On - - 


Priority - - 



Initial Responder Process_Associator ‘00’b ‘00’b 






Data encryption capable 0 0 







Abort, discard multiple Sequences (‘00’b) A A 

Abort, discard a single Sequence (’01’b) R R 

Process policy with infinite buffers (‘10’b) P P 

Discard multiple sequences w/immediate retrans (‘11’b) P P 

Categories per Sequence ‘00’b ‘00’b 



Data decryption capable 0 0 



Receive data field size (min) 256 256 

Concurrent Sequences (min) 1 1 

N_Port End-to-end Credit - - 

Open Sequences per Exchange (min) 1 1 

Notes 

18

5.8 NL_Port Class 3 Service Parameters (FLOGI) 


Table 9 lists Class 3 Service Parameters for FLOGI with usage defined by this document. 

Table 10 – Class 3 Service Parameters (FLOGI) 


SCSI 

Initiator 

SCSI 

Target 





Sequential Delivery 1 1 


Camp-On - - 


Priority 0 0 



Initial Responder Process_Associator - - 






Data encryption capable - - 







Abort, discard multiple Sequences (‘00’b) - - 

Abort, discard a single Sequence (’01’b) - - 

Process policy with infinite buffers (‘10’b) - - 

Discard multiple sequences w/immediate retrans (‘11’b) - - 

Categories per Sequence - - 



Data decryption capable - - 



Receive data field size (min) - - 

Concurrent Sequences (min) - - 

N_Port End-to-end Credit - - 

Open Sequences per Exchange (min) - - 

Notes 

19


5.9 Other FC-2 features 

Table 11 lists other FC-2 features with usage defined by this document. 

Table 11 – Other FC-2 Features 

Feature 



R_CTL Routing Bits 

FC-4 Device_Data Frame (‘0000’b) R R 

Extended Link_Data Frame (‘0010’b) R R 

FC-4 Link_Data Frame (‘0011’b) P P 

Video_Data Frame (‘0100’b) P P 

Basic Link_Data Frame (‘1000’b) R R 

Link_Control Frame (‘1100’b) 

Class 2 delivery service R R 1 

Class 3 delivery service P P 1 

Continue Sequence Condition = ‘00’b R R 

Ignore nonzero Continue Sequence values A A 

Sequence Chaining (C_S bit in F_CTL) = 0 R R 

Optional Headers (all) P P 

Node Name Format 

Uses a Registered Format R R 2 

Uses a non-Registered Format P P 

Port Name Format 

Uses a Registered Format R R 2 

Uses a non-Registered Format P P 

Node_Name = N_Port Name P P 3 

Continuously increasing Sequence count during consecutive 

Sequences within an Exchange 

NOTES: 

A 

A 

4 and 

see 

5.13.3 

1 Link Control Frames are required for Class 2. A SCSI Initiator or Target that supports only Class 3 shall 

return a P_RJT using Class 2 if a Link Control Frame is received. 

2 Node Names and N_Port Names shall use a format that contains a worldwide unique identifier issued 

by a recognized registration authority. The IEEE and IEEE Extended name formats are two examples of 

acceptable formats. 

3 The use of unique identifiers for both the Node_Name and Port_Name is required. 

4 The use of a continuously increasing sequence count is strongly recommended to provide enhanced error 

detection. 

20


5.10 Basic Link Service commands 

Table 12 lists Basic Link Service commands with usage defined by this document. NL_Ports may implement 

Prohibited Basic Link Service commands although their usage may be Prohibited by this 

profile. 

Feature 

Table 12 – Basic Link Services 

Orig. By 

SCSI 

Initiator 

Resp. By 

SCSI 

Target 

Orig. By 

SCSI 

Target 

Resp. By 

SCSI 

Initiator 

No Operation (NOP) P - P - 

Abort Sequence (ABTS) I R I R 

Basic Accept (BA_ACC) R R R R 

Basic Reject (BA_RJT) R R R R 

Remove Connection (RMC) (Class 1 only) P - P - 

Dedicated Connection Preempted (PRMT) P - P - 

5.11 Extended Link Services 

Table 13 lists Extended Link Services with usage defined by this document. 

Devices that receive requests for Extended Link Services that are unsupported shall return LS_RJT 

with a reason code “Command not supported.” 

Table 13 – Extended Link Services 


Orig. By 

SCSI 

Initiator 

Resp. By 

SCSI 

Target 

Orig. By 

SCSI 

Target 

Resp. By 

SCSI 

Initiator 

Note 

Accept (ACC) R R R R 

Discover Address (ADISC) I R P 1 

Discover N_Port Parameters (PDISC) I R P 1 

Fabric Login (FLOGI) I I 2,5,6 

Fabric Address Notification (FAN) I I 1,5 

Discover F_Port Parameters (FDISC) I I 5 

Loop Initialize (LINIT) I I 6 

Logout (LOGO) I R I R 

Loop Port Control (LPC) I I 6 

Link Service Reject (LS_RJT) R R R R 

Loop Status (LSTS) I I 6 

N_Port Login (PLOGI) R R P 

21


Table 13 – Extended Link Services 


Orig. By 

SCSI 

Resp. By 

SCSI 

Initiator Target Target 

Process Login (PRLI) R R P 

PRLI Common Service Parameters P P 

Single Service Parameter page per 

PRLI request 

R R P 

Multiple Service Parameter pages per 

PRLI request 

P 

P 

ACC contains only those pages 

specified in PRLI 

- R P 

Accept Response code=“Command 

executed” 

- R P 

Process Logout (PRLO) I R I R 

Read Exchange Concise (REC) R R P 4,7 

Read Link Error Status Block (RLS) I R P 3 

Report Node Capabilities Information 

(RNC) 

I R P 

Reinstate Recovery Qualifier (RRQ) I R P 

Request Sequence Initiative (RSI) A A A A 4 

Read Sequence Status Block (RSS) A A A A 4 

Sequence Retransmission Request 

(SRR) 

I R P 7 

Third Party Process Logout (TPRLO) I R P 

Any unlisted Extended Link Service P P 

NOTES: 

1 See 10.3 for a description of SCSI Target Discovery. Only ADISC or PDISC is required by a SCSI Initiator, 

but not both. SCSI Targets shall support receipt of both, for interoperability with SCSI Initiators that 

support one or the other. ADISC is the preferred method for Private NL_Ports. 

2 Private NL_Ports are prohibited from performing FLOGI. Public NL_Ports whose FLOGI fails (i.e. no 

FL_Port is present on the loop) may revert to Private behavior. 

3 Not all fields in the LESB may be supported (see vendor-specific documentation). RLS can be a diagnostic 

tool for isolating link degradation. Using RLS, diagnostic software can use a combination of 

self-directed test Frames, LPB, and LESB polling to determine which device or link segment is causing 

link errors. 

4 These link services permit the progress in a Sequence or Exchange at the Sequence Responder to be 

made visible for recovery purposes by a Sequence Initiator (e.g., determine whether a Sequence was 

delivered before using ABTS). 

5 Required for Public Loop NL_Ports only. 

6 FAN is the preferred method for Public NL_Ports to authenticate addresses following loop initialization. 

7 Refer to clause 9 and Annex A. 

Orig. By 

SCSI 

Resp. By 

SCSI 

Initiator 

Note 

22


5.12 Responses to Link Services 

Table 14 summarizes the responses that shall be generated as a result of receiving selected Link 

Service requests. 

Table 14 – Responses to Selected Link Services 

Frame Received PLOGI Not Successfully Completed PLOGI Successfully Completed 

ABTS Discard and send LOGO BA_ACC, BA_RJT (note 2) 

ADISC Discard and send LOGO ACC (note 1), LS_RJT 

LOGO ACC ACC 

PDISC Discard and send LOGO ACC (note 1), LS_RJT 

PLOGI ACC, LS_RJT (notes 4,5) ACC 

PRLI Discard and send LOGO ACC 

PRLO Discard and send LOGO ACC, LS_RJT (note 3) 

Unlisted Link 

ACC, LS_RJT (as appropriate and 

Discard and send LOGO 

Service 

defined by the applicable standard) 

NOTES: 

1 All three identifiers: N_Port ID, Port Name, and Node Name shall match a logged-in NL_Port for ACC to 

be returned. If all three identifiers do not match LOGO is returned. If other conditions prevent execution 

of the ADISC or PDISC ELS, LS_RJT may be returned as defined by the applicable standard with the 

appropriate reason code. 

2 BA_ACC or BA_RJT as specified in clause 9.5 

3 If PRLI has not been successfully completed, set the reason code to “image pair does not exist”. 

4 An NL_Port that does not wish to accept a PLOGI may return LS_RJT with a reason code of “Command 

Not Supported”. This could occur if a SCSI Initiator receives a PLOGI and does not wish to support SCSI 

Initiator to SCSI Initiator communication. 

5 An NL_Port that is not ready to accept a PLOGI at this time may return LS_RJT with a reason code of 

“Logical Busy”. 

While not shown in the table, an NL_Port shall return LS_RJT to any Extended Link Service request 

if that request is not properly formed or supported by the recipient NL_Port. 

An NL_Port that has discarded a Frame as shown in Table 14 shall not create an Exchange for the 

discarded request. Upon receipt of the LOGO, the originator of the discarded Frame terminates the 

Exchange associated with the discarded Frame and any other open Exchanges for the NL_Port 

sending the LOGO. The LOGO Sequence is in a separate Exchange with the original recipient as the 

Exchange Originator. 

5.13 Exchange and Sequence management 

This subclause describes required and optional behavior required of devices conforming to this profile. 

Refer to the FC-PH document(s) for rules and guidelines pertaining to class 2 operation. 

23


5.13.1 Exchange Originator 

The Originator of each FCP Exchange shall assign an OX_ID value other than hex ’FFFF’ to that Exchange. 

The Exchange Originator shall maintain OX_ID uniqueness with a given Responder. 

Exchange Originators are not assured of RX_ID uniqueness across multiple Responders, nor that a 

Responder will assign an RX_ID other than hex ’FFFF’. The Exchange Originator shall qualify the 

RX_ID with the Native Address Identifier of the Responder, if used by the Responder. 

Since the environment defined by this document ensures in-order delivery of Frames and requires 

that Frames are processed in the order received, a set of simplifying assumptions can be made regarding 

when an Exchange is open. 

The Exchange Originator shall consider the Exchange open from the time the first Frame of the first 

information unit (Frame with the First_Sequence bit in the F_CTL field set to 1) is sent until one of the 

following occurs: 

• the last Frame of the last information unit (Frame with the Last_Sequence bit in the F_CTL field 

set to 1) is sent or received; 

• the Exchange is aborted using ABTS (see 9.5) and a response to the ABTS is received; 

• a LOGO is sent to, or received from, the Exchange Responder; 

• an LS_RJT is sent in response to an ADISC or PDISC during Target discovery. 

• a PLOGI is sent to the Exchange Responder 

5.13.2 Exchange Responder 

Exchange Responders may optionally assign an RX_ID or use the value of hex ’FFFF’. 

Exchange Responders in a multiple Originator environment are not assured of OX_ID uniqueness 

across multiple Originators and shall qualify the OX_ID with the Native Address Identifier of the Originator. 

Exchange Responders are not required to check for OX_ID uniqueness with a given Originator. 

The Exchange Responder shall consider the Exchange open from the time the first Frame of the first 

information unit (Frame with the First_Sequence bit in the F_CTL field set to 1) is received until one 

of the following occurs: 

• the last Frame of the last information unit (Frame with the Last_Sequence bit in the F_CTL field 

set to 1) is sent or a FCP_CONF is received; 

• the Exchange is aborted using ABTS (see 9.5) and a response to the ABTS is sent; 

• a LOGO is sent to, or received from, the Exchange Originator; 

• an LS_RJT is sent in response to an ADISC or PDISC during Target discovery. 

• a PLOGI is received 

24


5.13.3 Sequence management 

The following subclauses define Sequence management requirements for devices operating in accordance 

with this document. 

5.13.3.1 Sequence open 

The Sequence Initiator shall consider a Sequence open from the time that the first Frame of the Sequence 

(the Frame with the SOFi3 delimiter) is sent until one of the following occurs: 

• the last Frame of the Sequence (the Frame with the EOFt delimiter) is sent and 

R_A_TOV SEQ_QUAL has elapsed; 

• the Sequence is aborted using ABTS and a response to the ABTS is sent or received, as appropriate; 

• a LOGO is sent to, or received from, the Sequence recipient. 

The Sequence Recipient shall consider a Sequence open from the time that the first Frame of the Sequence 

(the Frame with the SOFi3 delimiter) is received until one of the following occurs: 

• the last Frame of the Sequence (the Frame with the EOFt delimiter) is received; 

• the Sequence is aborted using ABTS and a response to the ABTS is sent or received, as appropriate; 

• a LOGO is sent to, or received from, the Sequence Initiator. 

5.13.3.2 SEQ_ID and SEQ_CNT usage 

The following paragraphs summarize the rules governing reuse of SEQ_IDs and SEQ_CNT as applicable 

to devices conforming to this profile. 

For all Sequences, 

a) Consecutive Sequences to the same destination within the same Sequence initiative shall use 

different SEQ_IDs. 

For Sequences that transfer Sequence Initiative, 

a) An NL_Port may reuse a SEQ_ID following confirmation of Sequence delivery (see 8.6 for a list 

of FCP Sequences whose delivery can be determined); 

b) An NL_Port may reuse the SEQ_ID of the last Sequence of an Exchange immediately following 

transmission of the last Frame of that Sequence (the Sequence Initiator may consider the Sequence 

delivered upon completion of transmission). 

For Sequences that do not transfer Sequence Initiative, 

a) Consecutive FCP_DATA Sequences for the same Exchange shall follow the FC-PH, FC-PH-2, 

and FC-PH-3 rules for streamed Sequences (an example of this occurs when multiple 

FCP_DATA Sequences are used during a single read operation). The rules for streamed Sequences 

are summarized in the following list (refer to FC-PH, FC-PH-2, and FC-PH-3 for complete 

streamed Sequence requirements): 

25


1) The first FCP_DATA Sequence after transfer of Sequence Initiative is not a streamed Sequence. 

It may use any eligible SEQ_ID and the SEQ_CNT may be either zero or continuously 

increasing (incremented by one from the last Frame of the previous Sequence of that 

Exchange); 

NOTE – Some existing devices require the SEQ_CNT for the first FCP_DATA Frame following the transfer 

of Sequence Initiative to be zero. 

2) The second, and subsequent, Sequences of a group of consecutive Sequences within the 

same Exchange shall be treated as streamed. The following rules apply (where X in the following 

discussion is the number of Open Sequences per Exchange from PLOGI (see table 

9)): 

NOTES: 

i) X+1 different SEQ_IDs shall be used; 

ii) The X+1 different SEQ_IDs may be assigned by the Sequence Initiator in any order (they 

are not required to be monotonically increasing); 

iii) A SEQ_ID may be reused after X different intervening SEQ_IDs have been used, provided 

that R_A_TOV SEQ_QUAL (as defined in 7.5) has expired for all Frames associated with the 

prior use of that SEQ_ID; 

iv) The Sequence Count (SEQ_CNT) across streamed Sequences shall be continuously increasing. 

That is, the SEQ_CNT of the first Frame of the streamed Sequence shall be incremented 

by one from the SEQ_CNT of the last Frame of the previous Sequence; 

1 FC-PH-3 adds a bit to PLOGI that indicates that a port uses continuously increasing SEQ_CNT on all Sequences 

within an Exchange with the first Sequence of the Exchange beginning at SEQ_CNT zero. Ports 

that set this bit during login are obligated to follow this behavior since the Sequence Recipient may use the 

expected behavior to perform Sequence checking. 

2 Some existing devices are not capable of generating continuously increasing SEQ_CNT across consecutive 

Sequences. In this case, the following rules apply: A minimum of two different SEQ_IDs shall be used 

by the Sequence Initiator and the SEQ_CNT across consecutive Sequences may restart at x‘0000’. 

b) The Target may use one or more Sequences to transfer the data requested by a SCSI readtype 

command. The Target may break the associated transfer up into as many arbitrary sized 

Sequences as the Target chooses (subject to the Sequence boundary requirements specified 

elsewhere in this document). This means that: 

1) When there is additional read data still to transfer and all of the data Frames associated with 

a given Sequence have been sent, or all permissible SEQ_CNT values used, the Target may 

begin a new Sequence in order to continue sending the read data; 

2) The Target may send less than the maximum permissable number of data Frames per 

FCP_DATA Sequence, even if there is additional data to send (for example, if the maximum 

Sequence size is 1000 Frames (as defined by the maximum burst size field of the SCSI disconnect/reconnect 

mode page), the Target may terminate the current Sequence after 600 

Frames and subsequently start a new Sequence to send any remaining Frames). 

c) During a read type command with at least one FCP_DATA Sequence, the FCP_RSP Sequence 

may optionally be considered a streamed Sequence and observe the rules for 

streamed Sequences listed previously. 

26


d) A SCSI Target may reuse a SEQ_ID previously used for a Read FCP_DATA Sequence immediately 

following transmission of the FCP_RSP of that Exchange. 

Examples of SEQ_ID and SEQ_CNT usage during streamed Sequences may be found in Annex B. 

5.13.4 Sequence errors 

The following conditions shall constitute a Sequence error when detected by the Sequence Recipient: 

a) if a Frame with an SOFi3 delimiter is received and the SEQ_CNT is not equal to zero or +1 

from the SEQ_CNT of the last Frame of the previous Sequence of that Exchange; 

b) if the SEQ_CNT of a received Frame with an SOFn3 delimiter is not +1 greater than the previous 

Frame received for that Sequence (i.e., a Frame was lost); 

c) if a Frame with an SOFn3 delimiter is received for a SEQ_ID that is not currently open; 

d) if a Frame with an SOFi3 delimiter is received and the previous Sequence of that Exchange is 

still open (see 5.13.3.1 for a definition of when a Sequence is open); 

e) if the relative offset bit in the F_CTL field is set to ‘1’b and the value contained in the parameter 

field of a received Frame with an SOFn3 delimiter is not equal to the (relative offset + the payload 

size) of the previous Frame received for that Sequence; 

f) the next Frame of a Sequence is not received within E_D_TOV; 

g) if, during the same Sequence Initiative, a Sequence is received that has the same SEQ_ID as 

the immediately prior Sequence of that Exchange. 

Refer to the FC-PH document(s) for a description of Sequence errors pertaining to class 2 operation. 

When a Sequence error is detected by the Sequence Recipient, it shall take the appropriate action as 

described in clause 9. 

27


6 FC-AL feature set 

This clause describes the use of Fibre Channel Arbitrated Loop features and capabilities by devices 

that conform to this document. Devices shall be compliant with FC-AL-2. 

6.1 FC-AL features 

Table 15 lists Fibre Channel Arbitrated Loop features with usage defined by this document. 

Table 15 – FC-AL Features 

Feature 

SCSI 

Initiator 

SCSI 

Target 

Open Full Duplex - OPN(yx) 

Open Originator can send I I 

Open Recipient accepts R R 

Open Half Duplex - OPN(yy) 

Open Originator can send I I 

Open Recipient accepts R R 

Open Multicast/Selective Replicate - OPN(yr), OPN(fr) 

Open Originator can send P P 

Login_BB_Credit 

Advertise Login_BB_Credit >= 0 A A 

Accept Login_BB_Credit >= 0 R R 3 

LPEyx/LPByx/LPEfx 

Origination P P 1 

MRKtx origination P P 4 

Attempt to acquire Hard Address (by devices having a Hard Address 

defined) during LIHA Sequence of loop initialization following a R R 5 

power cycle, power-on reset, or recognition of LIP(AL_PD,AL_PS) 

LILP/LIRP 

Loop Master shall originate R R 2 

Non-loop Master L_Ports accept R R 2 

NOTES: 

Notes 

1 LPEfx is useful for resetting the bypass circuits of NL_Ports that have been bypassed and have lost 

their AL_PAs due to an intervening LIP (and therefore cannot be enabled using an addressed LPEyx). 

SCSI Targets complying with SFF-8045 shall be able to recognize LPE/LPB directed to the appropriate 

AL_PA and toggle the corresponding port bypass enable signal. 

2 The initialization Loop Master shall originate the LIRP/LILP Sequences if the Loop Initialization identifier 

(in the LISA Frame) is hex ‘11050100’. Any L_Port (including the Loop Master) that is not capable of 

supporting LIRP/LILP shall set the Loop Initialization identifier (in the LISA Frame) to hex‘11050000’ to 

indicate that the LIRP/LILP Sequences shall not be originated. 

3 The actual value used shall be between zero and the Login_BB_Credit value received during PLOGI. 

4 An NL_Port that receives a MRK shall attempt to forward the MRK per FC-AL rules. FC-AL states that 

an L_Port may remove a MRK if necessary for clock skew management. 

5 Ports complying with SFF-8045 shall attempt to acquire the Hard Address presented by the SFF-8045 

defined interface. 

28

6.2 Alternate Login_BB_Credit management 


OPN Originators may open full or half duplex, regardless of the value of Login_BB_Credit. See Annex 

A for examples. 

Interoperability with devices that advertise Login_BB_Credit=0 is required. 

After completing the Loop Initialization Procedure, Login_BB_Credit shall be zero with all devices until 

after completion of the SCSI Target Discovery protocol (see 10.3). 

NOTE – This is essential for two reasons: The device at a particular address that had Login_BB_Credit > 0 

may have been replaced by a device with Login_BB_Credit = 0. Until SCSI Target Discovery is complete, 

there is no way to be sure that the port is communicating with the Login_BB_Credit > 0 device. Secondly, a 

device may have been trying to balance credit by not sending CLS when the LIP occurred. The LIP could have 

prematurely forced the device into the MONITORING state at a time when it has insufficient buffers available. 

Requiring SCSI Target Discovery before assuming Login BB_Credit > 0 enables the device to free its buffers 

before attempting to resume normal operations. 

6.3 Loop Initialization features 

NOTE – The information in this subclause is provided for information only and is not intended as a replacement 

for the relevant standard. 

NL_Ports shall attempt to acquire their previously acquired AL_PA before they attempt to acquire 

their Hard Address, unless the NL_Port has experienced a power cycle, power-on reset, recognized 

a LIP(AL_PD,AL_PS) for that port, or any other event that causes the previously acquired AL_PA to 

be lost. 

6.3.1 Initializing LIP(F7,F7) and LIP(F7,AL_PS) 

If there are no address conflicts, these LIPs shall have no effect on existing Exchanges except that 

the Exchange in progress may be disrupted, resulting in a ULP timeout. LIP(F7,F7) is used if the initiating 

NL_Port has no valid AL_PA, and LIP(F7,AL_PS) is used if it does. 

6.3.2 Selective Hard Reset LIP(AL_PD,AL_PS) 

Upon recognition of LIP(AL_PD,AL_PS) the NL_Port designated by AL_PD shall perform the appropriate 

resetting actions specified in table 23. All other L_Ports shall treat this as LIP(F7,AL_PS). SCSI 

Targets are prohibited from issuing this LIP. 

6.3.3 Loop Failure LIP(F8,AL_PS) and LIP(F8,F7) 

A Loop Failure is defined as detection of any of the following: 

a) Loss of Signal 

b) Loss of Synchronization for longer than R_T_TOV 

If a Loop Failure occurs, the L_Port which detects the failure shall issue LIP(F8,AL_PS) if it has a valid 

AL_PA, or LIP(F8,F7) if it doesn’t. 

6.3.4 Failure to obtain an AL_PA 

An L_Port which loses its previously acquired address shall implicitly logout all NL_Ports regardless 

of whether or not that AL_PA was previously-assigned, hard-assigned, or soft-assigned. If no soft ad- 

29


dress is available, it shall go to non-participating mode. An NL_Port which has experienced a power 

cycle, received a power-on reset, or recognized a LIP(AL_PD,AL_PS) is not required to retain a previously 

acquired address to use during the next loop initialization. 

30

7 Timers on Public and Private Loop 


FC-PH requirements for timers are summarized in Annex A. This clause defines the timers and timeout 

values used for Public and Private Loop implementations. 

Table 16 – Timer Summary 

Timer Implemented by Description Value 


AL_TIME R R Arbitrated Loop Timeout Value 15 ms per FC-AL 

LIS_HOLD_TIME R R Loop Initialization Sequence Hold 

1 ms 

Time 

R_T_TOV R R Receiver_Transmitter Timeout 100 ms per FC-PH 

Value 

E_D_TOV R A Error_Detect_Timeout Value 

2 sec 

(note 2) 

R_A_TOV SEQ_QUAL R A 

(note 2) 

Resource_Allocation Timeout Value Private = 0 sec 

Public = 10 sec 

(note 1) 

R_A_TOV ELS R R Private = 2 sec 

Public = 10 sec 

RR_TOV R Resource Recovery Timeout Value 2 sec 

FCP_TOV R R FCP Timeout Value >=E_D_TOV + 1 sec 

(minimum) 

ULP_TOV R Upper Level Protocol Timeout Value >=Operationspecific 

timer + 

4xFCP_TOV 

NOTES: 

1 The division of R_A_TOV usage as specified in this table differs from FC-PH. This is due to the unique 

characteristics associated with the FC-TAPE environment. 

2 SCSI Target devices that support Class 2 delivery service are required to implement this timer. 

7.1 Arbitrated Loop Time (AL_TIME) 

AL_TIME is specified by FC-AL as 15 ms, which represents two times the worst case round-trip latency 

for a very large loop. 

7.2 Loop Initialization Sequence Hold Time (LIS_HOLD_TIME) 

LIS_HOLD_TIME is the maximum amount of time between when a node receives a Loop Initialization 

Sequence until it forwards it to the next node. The value for LIS_HOLD_TIME is equal to 1 ms. 

7.3 Receiver_Transmitter Timeout (R_T_TOV) 

The Receiver_Transmitter timeout value (R_T_TOV) is used by the receiver logic to detect a Loop 

Failure. The value specified by FC-PH for R_T_TOV is 100 ms. 

7.4 Error_Detect Timeout (E_D_TOV) 

E_D_TOV is a general error detect timeout value. It is used to time the following: 

31


a) the maximum time permitted for a Sequence Initiator between the transmission of consecutive 

Data Frames within a single Sequence; 

b) the minimum time that a Sequence Recipient shall wait for the reception of the next Frame 

within a single Sequence before recognizing a Sequence timeout; 

c) the minimum time a NL_Port shall wait to win arbitration before it may assume a malfunction or 

excessive unfairness and optionally transmit a LIP Sequence. 

d) The maximum time a SCSI Target shall wait for the ACK response to a Sequence (Class 2) before 

it shall send an ABTS for the Sequence. 

After completion of FLOGI, Public Loop devices shall use the value of E_D_TOV specified by the 

Fabric in the FLOGI ACC. 

E_D_TOV shall include the time required to gain access to the loop in addition to the actual Frame 

transmission time. The default E_D_TOV value for devices conforming to this profile is 2 seconds. 

SCSI Target devices that support Class 2 are required to implement this timer for the purpose of error 

detection. 

7.5 Resource Allocation Timeout (R_A_TOV) 

R_A_TOV has two separate components, which are labeled R_A_TOV SEQ_QUAL and R_A_TOV ELS . 

R_A_TOV SEQ_QUAL is used to define the minimum amount of time that a Sequence Initiator shall wait 

before reusing the Sequence qualifiers (SEQ_ID and SEQ_CNT). This value is also the minimum 

amount of time that a SCSI Initiator shall wait following receipt of the BA_ACC to ABTS before transmitting 

a Reinstate Recovery Qualifier (RRQ) ELS. R_A_TOV SEQ_QUAL is equal to 0 seconds for Private 

Loop devices and 10 seconds for Public Loop devices. 

NOTE – Using a value of 0 for this time out value assumes that a Sequence Initiator does not transmit any 

Frames for a Sequence after an ABTS is sent for that Sequence. If a design uses a queuing mechanism for 

the transmission of Sequences, the queue for a given Sequence shall be empty before an ABTS for that Sequence 

can be sent, or the act of sending the ABTS purges the queue. 

A value of twice R_A_TOV ELS is used to determine the minimum time that the Originator of an Extended 

Link Service or FC-4 Extended Link Service request shall wait for the response to that request. 

R_A_TOV ELS is equal to 2 seconds for Private Loop devices and 10 seconds for Public Loop 

devices. 

NOTE – SCSI Targets are required to implement R_A_TOV ELS in order to time the expected response 

to a LOGO or PRLO Extended Link Service. 

After completion of FLOGI, Public Loop devices shall use the value of R_A_TOV specified by the 

Fabric in the FLOGI ACC. 

7.6 Resource Recovery Timer (RR_TOV) 

RR_TOV is the minimum time a SCSI Target shall wait for a specific SCSI Initiator to perform Exchange 

Authentication (see 10.4) following the completion of the Loop Initialization Protocol (i.e., the 

receipt of CLS while in the OPEN-INIT state). RR_TOV is also the minimum time a SCSI Target shall 

wait for a SCSI Initiator response following transfer of Sequence Initiative from the SCSI Target to the 

32


SCSI Initiator (e.g., following transmission of the FCP_XFER_RDY during a write command). A suggested 

default value is 2 seconds. 

Upon expiration of RR_TOV, a SCSI Target may implicitly or explicitly perform a LOGO with that 

SCSI Initiator, terminate all open Exchanges for that SCSI Initiator, and reclaim the resources associated 

with those Exchanges. The suggested default value of 2 seconds was felt to be a sufficient 

amount of time for a SCSI Initiator to complete the post-LIP processing and initiate the Exchange Authentication 

process with each SCSI Target on the loop. 

7.7 Fibre Channel Protocol Timeout Value (FCP_TOV) 

FCP_TOV is used to time FCP level information units. For Class 2 delivery service the value shall be 

equal to E_D_TOV. For Class 3 delivery service the minimum value shall be equal to E_D_TOV plus 

one second. 

7.8 Upper Level Protocol Timeout (ULP_TOV) 

ULP_TOV is an operation-specific timer maintained by the Upper Level Protocol. ULP_TOV is used 

to time the completion of Exchanges associated with ULP operations. Since the amount of time required 

varies depending upon the operation, the value assigned for this timer is determined by operation 

being timed. Some operations may require extended periods of time to complete. 

The minimum value shall be the operation-specific timer plus four times the value of FCP_TOV. 

33


8 SCSI-FCP Feature Set 

The following Feature Sets are described in FCP. The tables in this clause indicates whether the feature 

is required, prohibited, or optional for compliance with this specification. 

8.1 Process Login parameters 

Table 17 lists usage of Process Login parameters with usage as defined by this document. It does 

not define the values that must be reported during Process Login. 

Table 17 – PRLI Parameters 

Feature 

Use by 

SCSI 

Initiator 

Use by 

SCSI 

Target 

Originator Process Associator Valid = 1 P P 

Responder Process Associator Valid = 1 P P 

Originator Process Associator P P 

Responder Process Associator P P 

Establish Image Pair (bit 13) = 1 R R 2 

Establish Image Pair (bit 13) = 0 I R 2 

Initiator Function =1 R A 1 

SCSI Target Function = 1 A R 1 

Data Overlay Allowed = 1 R R 3 

Data + Response in same Sequence (Read) = 1 P P 

Command + Data in same Sequence (Write)= 1 P P 

Read XFER_RDY Disabled = 1 R R 

Write XFER_RDY Disabled = 1 P P 

FCP_CONF Support = 1 I R 

Command Reference Number Support (CRNS) = 1 I R 

NOTES: 

Notes 

1 What normally might be thought of as Target devices can assume an Initiator role for performing functions 

such as a Copy command. These devices report both Initiator and Target function during Process 

Login. 

2 This bit shall be set to one in the PRLI payload (requesting the image pair be established) and shall be 

set to a one in the PRLI ACC payload when the image pair has been successfully established before 

sending FCP Frames. The bit may be set to zero for the purpose of checking another NL_Port’s capabilities. 

3 Until such time as a future revision or version of ANSI X3.269-1996, Information technoloby Fibre Channel 

Protocol for SCSI provides a formal definition of the term “Data Overlay” this document shall interpret 

that term as follows: “Data overlay occurs when within an Exchange data is transferred to or from 

the same offset of the SCSI application client buffer more than once.” This definition shall apply to both 

read and write type commands. 

34


8.2 FCP Information Units 

From a SCSI Initiator to a SCSI Target, the T1 (CMD/Task Mgmt with SI transferred), and T6 (Write 

Data with SI transferred) are required and the T12 (Confirm with SI held) is optional. Devices that implement 

linking shall also support the T3 (CMD/Task Mgmt) Information Unit. All other SCSI Initiator 

to SCSI Target IUs are prohibited. 

From a SCSI Target to a SCSI Initiator, the I1 (FCP_XFER_RDY on Write), I3 (Read Data), I4 

(FCP_RSP) are required. Devices that implement SCSI linking shall also support the I5 (FCP_RSP) 

Information Unit. All other SCSI Target to SCSI Initiator IUs are prohibited. SCSI Targets shall respond 

to Task Management functions using I4. Examples of write and read type operations are 

shown in Figure 2. 

WRITE CMD 

READ CMD 

T1 

FCP_CMND 

T1 

FCP_CMND 

FCP_XFER_RDY 

I1 

DATA 

T6 

T6 

Single Sequence 

Single Sequence 

T12 

DATA 

DATA 

DATA 

FCP_XFER_RDY 

DATA 

DATA 

DATA 

FCP_RSP 

FCP_CONF 

(optional) 

I1 

I4 

DATA 

DATA 

DATA 

DATA 

DATA 

FCP_RSP 

FCP_CONF 

(optional) 

Single or Multiple Sequences 

I3 

I3 

I4 

T12 

8.2.1 FCP_CMND IU (T1/T3) 

Figure 2 — FCP Read/Write IU Examples 

The T1/T3 Information Unit shall be a Single-Frame Sequence. 

SCSI Targets that report a peripheral device type of Array Controller (hex ’C’) shall conform to the SCC defined 

LUN addressing mechanism. Other SCSI Targets shall address the LUNs using the “first level addressing” 

field of SCC (reference [9]) with “peripheral device addressing method” (00b) as shown in figure 3, unless 

they report a different LUN addressing mechanism in response to a Report LUN command. For example: 

Figure 3 — FCP 8-byte LUN 

Byte 0 Byte 1 Byte 2 Byte 3 Byte 4 Byte 5 Byte 6 Byte 7 

00 LUN 00 00 00 00 00 00 

35


FCP_DL in the FCP_CMND payload should always be equal to the number of bytes expected to be 

transferred for the command. For SCSI commands that specify the Transfer Length in the CDB in 

fixed blocks, the FCP_DL should be equal to (Transfer_Length x Block_Length). SCSI Targets are 

required to check that the FCP_DL is in fact the (Transfer_Length x Block_Length). 

On read commands, the SCSI Initiator is responsible for ensuring the amount of data returned is 

equal to the amount specified by FCP_DL, even if the Status returned is GOOD. Since there are no 

transfers of Sequence Initiative during read operations once the SCSI Target receives the T1 Information 

Unit, status may be returned as GOOD even though some of the data was not received by the 

SCSI Initiator. This may occur as the result of lost or corrupted Frames in the read data. 

If the amount of data returned does not match FCP_DL the error detection and recovery procedure 

described in clause 9.3.6 shall be invoked. 

NOTE – The manner in which a SCSI Initiator determines that the correct amount of data is returned is implementation 

dependent and may include counting the number of bytes returned, computing the number of bytes 

received by use of the relative offsets, or other means not specified. 

8.2.1.1 FCP Command Reference Number (CRN) 

The FCP Command Reference Number shall be used to ensure proper ordering of Exchange’s (SC- 

SI commands). The Target shall indicate support of CRN by setting the Command Reference Number 

Support (CRNS) bit to one in the PRLI request and enable CRN usage on a per-Lun basis by 

setting the Enable Command Reference Number (ECRN) bit to one in the FC Mode Page (0x19) for 

the Lun. See Annex D for a further description of CRN implementation and usage. 

8.2.2 FCP_XFER_RDY IU 

The FCP_XFER_RDY IU (I2) shall not be used during read type (data in) operations. This shall be indicated 

by setting the ‘READ XFER_RDY DISABLED’ bit during process login. 

During write type (data out) operations, FCP_XFER_RDY (I1) shall be sent prior to each Write Data 

Sequence. The I1 Information Unit shall be a Single-Frame Sequence. 

If the Enable Modify Data Pointer (EMDP) in the SCSI Disconnect-reconnect mode page = ‘1’b, then 

for transfers consisting of multiple Write Data Sequences, the DATA_RO parameter contained in 

consecutive FCP_XFER_RDY Sequences does not have to be continuously increasing. This provides 

the logical equivalent of SCSI Modify Data Pointers. NL_Ports conforming to this profile shall 

allow data overlay that may occur as a result of the values chosen for the DATA_RO parameter. This 

is not the same as the Random Relative Offset N_Port login parameter, that refers to Relative Offset 

on consecutive Frames within a Sequence. 

The BURST_LEN parameter in the FCP_XFER_RDY IU on write type operations shall be less than 

or equal to the amount of data remaining to be transferred for the write task and not exceed the maximum 

burst size in the Disconnect-reconnect mode page. 

8.2.3 FCP_DATA IU 

The FCP_DATA IU is used to transfer the data associated with an operation, if any. This data may be 

logical data to or from the medium, as well as command parameter data (such as Mode Select data) 

or command response data (such as Mode Sense data). FCP does not distinguish between logical 

data and command parameter or response data. 

The T6 and I3 Information units may be either Single-Frame or Multiple-Frame Sequences. 

36


Within an Exchange, only the FCP_DATA Frame containing the largest Relative Offset of the last Sequence 

of the Exchange may contain fill bytes. 

8.2.3.1 Write FCP_DATA IU (T6) 

The Parameter field of the first Frame of a Write data Sequence shall be set to the Relative Offset 

specified by the corresponding FCP_XFER_RDY. On subsequent Frames of the Sequence, continuously 

increasing Relative Offset shall be used. 

8.2.3.2 Read FCP_DATA IU (I3) 

Within Read Data Sequences, the Relative Offset on consecutive Frames within that Sequence shall 

be continuously increasing. If the Enable Modify Data Pointer (EMDP) in the SCSI Disconnect-reconnect 

mode page = ‘1’b, then for read-type commands with multiple data (I3) Sequences, the starting 

Relative Offsets (the Relative Offsets in the first Frame of each Sequence) need not be continuously 

increasing. 

During read-type commands with multiple data (I3) Sequences, it is recommended that the NL_Ports 

treat all data Sequences other than the first as streamed Sequences and follow the associated rules 

for streamed Sequences (see 5.13.3 and Annex B). 

8.2.4 FCP_RSP (I4/I5) 

The I4/I5 Information Unit from a device shall be a Single-Frame Sequence. 

According to FC-PH, the first two bits [31:30] of the first word of a command status Frame payload fall 

into one of the following categories: 00 (successful and complete), 01 (successful but incomplete), 10 

(unsuccessful but complete), 11 (unsuccessful and incomplete). 

Since the first word of FCP_RSP Frames are reserved in FCP, these bits are set to zero, regardless 

of the content of the SCSI Status portion of the payload (which may indicate the command was unsuccessful 

and/or incomplete). SCSI Initiators shall not rely on bits 31:30 of word 0 in FCP_RSP to 

determine success or completion status of a command. 

8.2.4.1 FCP_RSP Residual checking 

SCSI Targets that transfer exactly FCP_DL data bytes during the FCP_DATA IUs shall set the 

FCP_RESID_UNDER to ‘0’b. SCSI Targets that transfer less than FCP_DL data bytes during the 

FCP_DATA IUs shall set the FCP_RESID_UNDER to ‘1’b. 

When FCP_RESID_UNDER is set to ‘0’b, the SCSI Initiator shall determine if all of the expected data 

was transferred by comparing the FCP_DL to the actual number of bytes transferred. If these values 

are not the same invoke error detection and recovery procedure(s) as described in clause 9. 

If the FCP_RESID_UNDER bit is set to ‘1’b, a transfer that did not fill the buffer to the expected displacement 

FCP_DL was performed and the value of FCP_RESID shall be equal to: 

FCP_DL - highest offset of any byte transmitted 

the SCSI Initiator shall determine if all of the expected data was transferred by comparing the 

FCP_RESID to the actual number of bytes transferred. If these values are not the same invoke error 

detection and recovery procedure(s) as described in clause 9. 

37


Failure to transfer FCP_DL bytes does not necessarily indicate an error for some devices and commands. 

If the FCP_RESID_OVER bit is set, the transfer was truncated because the data transfer required by 

the SCSI command extended beyond the displacement value of FCP_DL. Those bytes that could be 

transferred without violating the FCP_DL value may or may not have been transferred. The value of 

FCP_RESID shall be equal to: 

(Transfer length required by command) - FCP_DL 

Use of FCP_RESID_OVER is required for both read and write type commands. 

During commands that do not contain an FCP_DATA IU, FCP_RESID_UNDER and 

FCP_RESID_OVER shall both be set to ‘0’b, and the value of the FCP_RESID is undefined. 

8.2.4.2 FCP_RSP payload 

Table 18 lists those FCP_RSP fields with usage defined by this document. 

Table 18 – FCP_RSP Payload 

Feature 

SCSI 

Initiator 

SCSI 

Target 

Notes 

FCP_SNS_LEN (number of bytes) R


Table 19 – Task Management Function RSP_CODES 

RSP_CODE 

Meaning 

00 No Failure or Task Management function complete 

01 FCP_DATA length different than BURST_LEN 

02 FCP_CMND Fields Invalid 

03 FCP_DATA RO mismatch with FCP_XFER_RDY DATA_RO 

04 Task management function was not performed because the function is unsupported 

05 Task management function was not performed, but the function is supported. 

06-FF Reserved 

8.2.5 FCP_CONF (T12) 

The FCP_CONF IU is used to confirm receipt of a FCP_RSP. Support for FCP_CONF is negotiated 

using FCP_CONF Support bit to one in the PRLI request. A Target device conforming to this profile 

shall request a FCP_CONF IU for each FCP_CMND containing a Task Attribute of SIMPLE_Q, 

HEAD_OF_Q, or ORDERED_Q (i.e. tagged task). 

The FCP_CONF is sent by the Initiator after a FCP_RSP has been received with the 

FCP_CONF_REQ bit set in the FCP_STATUS field. The Initiator shall release Exchange information 

such as the Exchange Status Block (ESB) after the FCP_CONF is sent. The Target shall retain Exchange 

information and associated data until a FCP_CONF is received. See Annex C for a description 

of the FCP_CONF_REQ bit and FCP_CONF usage. 

39


8.3 Task Management Flags and Information Units 

All SCSI Initiators shall send profile supported Task Management functions using T1. All SCSI Targets 

shall return FCP_RSP to Task Management functions using I4. The RSP_CODE in the 

FCP_RSP_INFO field shall indicate the result of the Task Management function. The SCSI Status 

byte and FCP_SNS_INFO shall be ignored for I4 information units sent in response to a Task Management 

function. 

Table 20 lists FCP Task Management functions with usage defined by this document. Only one Task 

Management function shall be selected at a time. 

Table 20 – FCP Task Management Flags 

Feature 

SCSI 

Initiator 

SCSI 

Target 

Notes 

Terminate Task = 1 P P 

Clear ACA = 1 (command queuing) R R 1,3 

Clear ACA = 1 (no command queuing) P P 4 

Target Reset = 1 I R 2 

Clear Task Set = 1 I R 2 

Abort Task Set = 1 I R 2 

NOTES: 

1 Clear ACA is only required by the SCSI Target if it indicates support of ACA via the NormACA bit in the 

INQUIRY data. Clear ACA is required by SCSI Initiators that support ACA (NACA bit in CDB=1) when 

communicating with SCSI Targets that have indicated support of ACA. 

2 See 9.5 for SCSI Initiator requirements following transmission of these Task Management functions. 

3 For tagged command queuing operations the SCSI Target shall set the NormACA bit=1 and the SCSI 

Initiator shall set the NACA bit=1 in the CDB. 

4 For non-tagged command queuing operations setting Clear ACA=1 is prohibited. 

8.4 FCP Task Attributes 

Table 21 lists FCP Task Attributes with usage defined by this document. 

Table 21 – FCP Task Attributes 

Feature 

SCSI 

Initiator 

SCSI 

Target 

Notes 

Untagged R R 

Simple Queue Type I A 2 

Ordered Queue Type I A 2 

Head of Queue Type I A 2 

Auto Contingent Allegiance Type I A 1 

NOTES: 

1 ACA Type is only required by the SCSI Target if it advertises support of ACA via the NormACA 

bit in the INQUIRY data. ACA Type is required by SCSI Initiators that use ACA 

(NACA bit in CDB=1). 

2 The setting of the CmdQue bit in the INQUIRY data is allowed by the Target. If CmdQue 

set is one, the Target shall support tagged tasks (i.e. command queuing). 

40


8.5 Other FCP features 

Table 22 lists Other FCP features with usage defined by this document. 

Table 22 – Other FCP Features 

Feature 



FCP_LUN field of the FCP_CMND information unit R R 

Use of LUN = 0 I R 

INQUIRY of LUN = 0 I R 

INQUIRY of LUN > 0 I R 1 

Auto Contingent Allegiance (ACA) A A 2 

NOTES: 

1 Some SCSI Targets will require multiple LUNs. A SCSI Target is Required to send a response to any 

LUN, even if that response indicates the the LUN is not present. 

2 See clause 10.1. 

8.6 FCP Sequence delivery confirmation in a Class 3 environment 

In a Class 3 environment, there are no Acknowledgments to provide direct confirmation of Sequence 

delivery. In many cases, Sequence delivery can be inferred following transmission of a SCSI Initiator 

to SCSI Target Sequence that transfers Sequence Initiative. The following cases provide implicit confirmation 

of Sequence delivery. 

a) Receipt of the any of the following by a SCSI Initiator confirms that the FCP_CMND IU (T1) 

was delivered: 

1) receipt of a Write FCP_XFER_RDY IU (I1) following transmission of an FCP_CMND IU containing 

a write type command; 

2) receipt of a Read FCP_DATA IU (I3) following transmission of an FCP_CMND IU containing 

a read type command; 

3) receipt of FCP_RSP IU (I4) following transmission of an FCP_CMND IU containing any type 

of command. 

b) Receipt of any of the following by a SCSI Initiator confirms that a write FCP_DATA IU (T6) was 

delivered: 

1) receipt of a Write FCP_XFER_RDY IU (I1); 

2) receipt of FCP_RSP IU (I4). 

c) Receipt of the following by a SCSI Target confirms that a write FCP_XFER_RDY (I1) was delivered: 

1) receipt of a write FCP_DATA IU (T6). 

d) Delivery of the following Information Units sent by a SCSI Target cannot be determined directly 

(delivery failure is detected by the error detection procedure or a timeout condition): 

41


1) read FCP_DATA IU (I3); 

2) FCP_RSP IU (I4). 

e) If enabled, the receipt of a FCP_CONF IU by a SCSI Target confirms all FCP IU’s for the Exchange 

have been delivered. 

If Sequence confirmation is determined by receipt of an upper level response, SEQ_ID management 

may be delegated upward to the ULPs. 

8.7 FCP Sequence delivery confirmation in a Class 2 environment 

In a Class 2 environment acknowledgments are used to provide explicit confirmation of Sequence 

delivery. Implicit delivery may be achieved by issuing the REC extended link service and checking 

the response or the status of the Sequence Initiative bit in the returned Exchange Status Block. For 

Class 2 the Sequence recipient shall not consider Sequence Initiative to have been passed until the 

Sequence which passes the Sequence Initiative has completed successfully and the ACK has been 

transmitted with the Sequence Initiative bit = 1. 

42

9 FCP Error Detection and Recovery Procedure(s) 


This clause describes the procedures to be taken by the SCSI Initiator and SCSI Target to detect an 

error condition and recover (if possible) using Class 2 or Class 3 delivery service in a Public or Private 

loop topology. The unit of error recovery for this clause is a Sequence not an Exchange, although 

the Initiator may abort the Exchange if desired. The protocol defining retransmission of 

Exchanges following an abort of the Exchange as documented in 9.5 is beyond the scope of this document. 

Since this profile specifies in-order frame delivery through a Fabric no ABTS is required for Class 3 

error recovery except after a ULP timeout where it is used to abort the Exchange and for each REC 

or SRR sent that does not receive a response. For Class 2 an ABTS shall be sent from the Initiator if 

an ACK is received with an Abort Sequence indication and from the Target after expiration of 

E_D_TOV or if an ACK is received with an Abort Sequence indication. 

The following error detection and recovery procedure(s) are suitable for devices which may have implicit 

positioning and long ULP timeout values. 

9.1 FCP Error Detection for Class 3 Delivery Service 

The Exchange originator (Initiator) shall initiate error detection and recovery described in 9.3 for the 

following: 

1. after expiration of the timeout period FCP_TOV and no reply Sequence has been received for the 

FCP_CMND IU. 

2. after expiration of the timeout period FCP_TOV following the sending of FCP_DATA IU(s) and no 

FCP_RSP IU has been received for the FCP_CMND IU. 

3. a Sequence error is detected. 

4. a read-type command completed with a data count smaller than the CDB indicated. 

The Exchange responder (Target) shall initiate no error detection or recovery. 

9.2 FCP Error Detection for Class 2 Delivery Service 

The Exchange originator (Initiator) shall initate error detection and recovery described in 9.3 for the 

following: 

1. after expiration of the timeout period E_D_TOV and no ACK has been received for the 

FCP_CMND IU. 

2. after expiration of the timeout period E_D_TOV and no ACK has been received for FCP_DATA 

IU(s). 


4. an ABTS is received. 

The Exchange responder (Target) shall initiate error recovery described in 9.4 for the following: 

1. after expiration of the timeout period E_D_TOV and no ACK has been received for the 

FCP_XFER_RDY IU. 

2. after expiration of the timeout period E_D_TOV and no ACK has been received for FCP_DATA 

IU(s). 

3. after expiration of the timeout period E_D_TOV and no ACK has been received for the FCP_RSP 

IU. 

43



If an ACK is received with an Abort, Perform ABTS indication the Exchange originator (Initiator) or 

Exchange responder (Target) shall send an ABTS for a specified Sequence followed by an RRQ 

when the BA_ACC is received for the ABTS. 

9.3 FCP Error Detection and Recovery (Initiator, Class 2 and Class 3) 

The following error detection and recovery procedures are described for Class 2 and Class 3 delivery 

service. 

9.3.1 Error Detection 

NOTE – If a Sequence error has been detected in Class 2, the Exchange originator (Initiator) shall send an 

ACK Frame with an Abort, Perform ABTS indication before issuing the REC. 

Issue a REC for the Exchange containing the FCP_CMND. The REC shall be issued in a new Exchange. 

If the response is an LS_RJT with a reason code indicating that the function is not supported, treat 

the Target as a device not supporting this error detection procedure and perform error recovery as 

documented in 9.5. 

If a proper ACC, LS_RJT, LOGO, or PRLO is not received from the SCSI Target within E_D_TOV, 

second level error recovery as described in 9.5.3. shall be performed. 

9.3.2 FCP_CMND Recovery 

If the FCP_CMND was not received by the Target (i.e., the Initiator receives an LS_RJT for the REC 

with a reason code indicating the OX_ID is unknown), retransmit the FCP_CMND using a new 

OX_ID. 

The Target shall retain Exchange information until the next FCP_CMND has been received for that 

LUN or until a FCP_CONF IU is received for the Exchange. 

If the ACC for the REC indicates that the FCP_CMND was received by the Target and that no reply 

Sequence has been sent, the command is in process and no recovery is needed at this time. At intervals 

of FCP_TOV the REC shall be retransmitted. This is to ensure that no reply Sequences have 

been lost. If at any time, there is no reply to the REC, perform second-level error recovery as described 

in 9.5.3. 

9.3.3 FCP_XFER_RDY Recovery 

If the ACC for an REC indicates that an FCP_XFER_RDY was sent by the Target (by indicating that 

the Initiator holds Sequence Initiative, and that the Exchange is not complete), but not received by 

the Initiator, issue an SRR Extended Link Service Frame in a new Exchange to request retransmission 

of the FCP_XFER_RDY (R_CTL = data descriptor). The Target retransmits the 

FCP_XFER_RDY in a new Sequence containing the same Relative Offset as the originally transmitted 

FCP_XFER_RDY. When the FCP_XFER_RDY is successfully received, the data is sent, and the 

operation continues normally. No error is reported to the ULP, though the error counters in the LESB 

should be updated. 

If the SRR receives a LS_RJT, perform error recovery as documented in 9.5. 

44


9.3.4 FCP_RSP Recovery 

If an ACC for an REC indicates that an FCP_RSP Sequence was sent by the Target, but not received 

by the Initiator (E_STAT indicates the Exchange is complete), issue an SRR Extended Link Service 

Frame in a new Exchange to request retransmission of the FCP_RSP IU. The Target retransmits the 

FCP_RSP in a new Sequence. The response is delivered to the ULP, the error counters in the LESB 

are updated, and no error is reported. If the SRR receives an LS_RJT, perform error recovery as documented 

in 9.5. 

9.3.5 FCP_DATA Recovery - Write 

If the ACC for an REC indicates that an FCP_DATA Sequence was sent by the Initiator, but not received 

by the Target (i.e. the data received count in the REC response is smaller than what the Initiator 

sent, and the Target indicates it does not hold Sequence Initiative): 

Send a SRR Extended Link Service Frame in a new Exchange to request retransmission of an 

FCP_XFER_RDY to request the missing data. As documented in 9.3.8, the Target discards the Sequence 

in error, but does not initiate any recovery action for Class 3. After transmitting the ACC for 

the SRR, the Target transmits an FCP_XFER_RDY with the appropriate Relative Offset parameter 

and the Initiator responds with the requested data. The FCP_XFER_RDY may specify a Relative Offset 

smaller than the Relative Offset specified in the SRR in order to be aligned on an appropriate 

boundary in the Target. The operation should complete with no error indication to the ULP. 

9.3.6 FCP_DATA Recovery - Read 

If the ACC for the REC indicates that data was sent by the Target but not successfully received by 

the Initiator (by indicating a data sent count greater than the Initiator has successfully received): 

Send a SRR Extended Link Service Frame in a new Exchange to request retransmission of the 

FCP_DATA (R_CTL = Solicited Data) that was not successfully received. The Initiator passes the 

Relative Offset of the next data requested. The Target retransmits the data in a new Sequence, and 

then retransmits the FCP_RSP. The Target may send data beginning at a Relative Offset smaller 

than that specified in the SRR in order to be aligned in an appropriate boundary within the Target. 

The received data is delivered to the ULP and no error is reported. If the Target responds to the SRR 

with an LS_RJT and a reason code indicating that the function could not be performed, the Target 

shall present an FCP_RSP IU with an appropriate error status (e.g., Sense key 4, ASC/ASQ of 48/00 

(Initiator Detected Error message received)). 

It is the responsibility of the Initiator to determine the appropriate action required (retry, allow ULP 

time out, or return status to ULP) based on the information determined by REC and other internal 

state. As stated in 9.3.8, the Target does not initiate error recovery for Class 3. 

9.3.7 Additional error detection by SCSI Initiator 

The SCSI Initiator shall abort the Exchange using the ABTS Protocol (see 9.5) for the following: 

a) ULP_TOV Timeout and Exchange Not Complete (see 5.13 for conditions indicating Exchange 

completion) 

9.3.8 Error detection by SCSI Target 

If a SCSI Target detects a Sequence error, it shall discard the Sequence(s) based on the currently 

active Error Policy. For Class 3 the Target shall initiate no error recovery. 

45


SCSI Targets may implement RR_TOV as described in 7.6 to facilitate recovery of resources allocated 

to a non-respondent SCSI Initiator. If a SCSI Target implements an RR_TOV timer for this purpose, 

it may send a LOGO to the SCSI Initiator and terminate all open Exchanges for that SCSI 

Initiator upon detection of the following: 

46 

a) The SCSI Initiator has failed to perform SCSI Target Authentication within RR_TOV. 

b) The SCSI Initiator has failed to transmit the next expected Sequence of an Exchange for all 

open Exchanges with a specific SCSI Initiator within RR_TOV (e.g., FCP write data in response 

to an FCP_XFER_RDY). 

9.4 FCP Error Recovery (Target, Class 2) 

If no ACK has been received the Target shall send an ABTS for the Sequence. If a Recovery Qualifier 

range is returned in the BA_ACC for the ABTS the Target shall send a RRQ ELS following the expiration 

of R_A_TOV SEQ_QUAL after receipt of the BA_ACC. 

If a Sequence error is detected the Target shall send an ACK Frame with an Abort, Perform ABTS indication. 

Since the Target knows it is processing a Sequence error it shall not Abort the Exchange 

upon receiving the ABTS from the Initiator. If an ABTS is received by the Target and it has not detected 

a Sequence error it shall Abort the Exchange as described in 9.5. 

9.5 Aborting an Exchange Using ABTS Protocol 

The ABTS-Last Sequence of Exchange (ABTS-LS) protocol is required, which uses ABTS to abort 

entire Exchanges. 

9.5.1 SCSI Initiator ABTS of Exchange behavior 

This subclause does not define the protocol by which multiple SCSI Initiators communicate or synchronize 

shared peripherals. 

ABTS may be transmitted even if Sequence Initiative is not held. Following the transmission of ABTS, 

any Device_Data Frames received for this Exchange shall be discarded until the BA_ACC with “Last 

Sequence of Exchange” bit is received from the SCSI Target. 

ABTS may not take effect immediately. For example, if ABTS is sent following transmission of a Read 

command, the SCSI Target may send back all or some of the requested read data before replying 

with the BA_ACC to the ABTS (or the data may already be in flight at the time the ABTS was sent). 

The SCSI Initiator shall be capable of receiving this data and providing BB_Credit in order for the 

SCSI Target to send the BA_ACC. 

There are circumstances where a configuration change event could occur very quickly, which results 

in link errors and a ULP_TOV timeout, but no LIP is generated to reconfigure the loop. A SCSI Initiator 

shall attempt the ABTS protocol following ULP_TOV timeout. 

The ABTS protocol shall be invoked: 

a) for all open Exchanges for all LUNs on a designated SCSI Target: 

1) following receipt of “Function Complete” to SCSI Target Reset; 

2) following receipt of CHECK CONDITION status with Sense Key=Unit Attention and ASC/AS- 

CQ of

i) “Power on, reset, or bus device reset occurred” (29 XX); 

ii) Microcode changed (3F 01); 


b) for all open Exchanges for the specified LUN on a designated SCSI Target: 

1) following receipt of “Function Complete” to Abort Task Set; 

2) following receipt of “Function Complete” to Clear Task Set; 

3) upon receipt of CHECK CONDITION status with Sense Key=Unit Attention and ASC/ASCQ 

of “Command cleared by another Initiator” (2F 00); 

c) for a specific Exchange on a specified LUN on a designated SCSI Target: 

1) upon ULP_TOV timer expiration (no STATUS returned for a Command within an application-specified 

amount of time); 

Following receipt of the BA_ACC in response to an ABTS, and after R_A_TOV SEQ_QUAL has 

elapsed, the SCSI Initiator shall transmit RRQ. 

If a proper BA_ACC, BA_RJT, LOGO, or PRLO is not received from the SCSI Target within 

E_D_TOV, second level error recovery as described in 9.5.3. shall be performed. 

9.5.2 SCSI Target ABTS of Exchange behavior 

When ABTS is received at the SCSI Target, it shall abort the designated Exchange and return one of 

the following responses: 

a) the SCSI Target shall discard the ABTS and return LOGO if the NL_Port issuing the ABTS is 

not currently logged in (no PLOGI); 

b) the SCSI Target shall return BA_RJT with Last Sequence of Exchange bit set to one if the received 

ABTS contains an RX_ID, other than hex ‘FFFF’, that is unknown to the SCSI Target. 

The reason code shall be “Logical Error” with a reason code explanation of “Invalid 

OX_ID/RX_ID combination”; 

c) otherwise, the SCSI Target shall return BA_ACC with Last Sequence of Exchange bit set to 

one. 

Upon transmission of any of the above responses, the SCSI Target may reclaim any resources associated 

with the designated Exchange after R_A_TOV SEQ_QUAL has elapsed or a Reinstate Recover 

Qualifier (RRQ) extended link service request has been received. 

When RRQ is received at the SCSI Target, it shall respond according to the same rules stated for the 

ABTS. If the Exchange resources were not reclaimed upon responding to the ABTS, they shall be reclaimed 

at the time the response to the RRQ is sent. 

SCSI Targets shall qualify ABTS based only upon D_ID || S_ID || OX_ID, not RX_ID, since the RX_ID 

is not guaranteed to be known by a SCSI Initiator. 

47


9.5.3 Second-level error recovery 

9.5.3.1 ABTS 

If a response to an ABTS is not received within E_D_TOV, the SCSI Initiator shall: 

a) send the ABTS again. 

If a response to the second ABTS is not received within E_D_TOV, the SCSI Initiator shall explicitly 

logout (FC-PH N_Port Logout, LOGO) the SCSI Target and may issue the Selective Reset LIP 

(LIP,AL_PD,AL_PS) to reset the SCSI Target. All outstanding Exchanges with that SCSI Target are 

terminated at the SCSI Initiator. 

9.5.3.2 REC 

If a response to an REC is not received within E_D_TOV, the SCSI Intiator shall: 

a) send an ABTS for the REC followed by an RRQ if a BA_ACC is received for the ABTS. 

b) send another REC in a new Exchange. 

The REC shall be retried at a rate not to exceed once per the timeout period for at least 3 times. If 

none of the RECs receive a response, the Initiator shall report an error condition to the ULP. 

9.5.3.3 SRR 

If a response to an SRR is not received within E_D_TOV, the SCSI Initiator shall: 

a) send an ABTS for the SRR followed by an RRQ if a BA_ACC is received for the ABTS. 

b) send another SRR in a new Exchange. 

The SRR shall be retried at a rate not to exceed once per the timeout period for at least 3 times. If 

none of the SRRs receive a response, the Initiator shall report an error condition to the ULP. 

48


9.5.4 Abort Sequence (ABTS) Frame 

Only the SCSI Initiator shall transmit an ABTS Frame. When it does so, the specified fields shall be 

set as shown in figure 4. 

Frame 

Header 

Figure 4 — ABTS Frame 

Field Sub-field Content 

Sequence Initiator (even though the ABTS Initiator may not have 

F_CTL 

Context Sequence Initiative for the Sequence being aborted). 

Sequence Transferred, even if the ABTS Initiator did not hold 

Initiative Sequence Initiative prior to the ABTS. 

SEQ_ID 

If ABTS is sent by the Sequence Initiator and the 

Sequence is still open, the SEQ_ID of the open 

Sequence in the Exchange being aborted is used. 

Otherwise, the SEQ_ID is any SEQ_ID not currently open 

(for any Exchange) between that pair of ports. 

SEQ_CNT 

SEQ_CNT of last Frame transmitted in an Open 

Sequence + 1. If no Sequence is open, then 

SEQ_CNT=zero. 

OX_ID 

OX_ID = same as that assigned by the Exchange 

Originator (SCSI Initiator) for the Exchange being 

aborted. 

RX_ID 

Set to hex ’FFFF’ or the same as that assigned by the 

Exchange Responder (SCSI Target) for the task being 

aborted. 

9.5.5 Basic Accept (BA_ACC) Frame to ABTS 

A SCSI Target may accept ABTS with BA_ACC. When it does so, the BA_ACC shall be as shown in 

figure 5. 

Frame 

Header 

Payload 

Figure 5 — BA_ACC Frame to ABTS 


OX_ID 

OX_ID from ABTS Frame 

RX_ID 

RX_ID from ABTS Frame 

F_CTL 

Last_Sequence 1 

Sequence Context Recipient 

SEQ_ID validity hex ’00’ 

SEQ_ID byte 

invalid (don’t care) 

OX_ID 


RX_ID 


Lowest SEQ_CNT hex ’0000’ 

Highest SEQ_CNT 

hex ’FFFF’ 

49


9.5.6 Basic Reject (BA_RJT) Frame to ABTS 

A SCSI Target may reject ABTS with BA_RJT. When it does so, the BA_RJT shall be as shown in figure 

6. 

Figure 6 — BA_RJT Frame to ABTS 


OX_ID 


Frame RX_ID 


Header 

Last_Sequence 1 

F_CTL 

Sequence Context Recipient 

Payload Byte 0 Reserved hex ’00’ 

Byte 1 Reason Code hex‘03’ (Logical error) 

Byte 2 

Reason 

Explanation 

hex‘03’ (Invalid OX_ID-RX_ID 

combination) 

Byte 3 Vendor Unique hex‘00’ 

9.5.7 Reinstate Recovery Qualifier (RRQ) 

The format of the RRQ is shown in figure 7. 

Frame Header 

Payload 

Figure 7 — Reinstate Recovery Qualifier 

Field 

Content 

OX_ID 

Identifier of a new Exchange 

RX_ID 

hex ’FFFF’ 

Originator S_ID Source_ID of the SCSI Initiator 

OX_ID 

OX_ID of XCHG that was previously aborted with ABTS 

RX_ID 

RX_ID of XCHG that was previously aborted with ABTS 

Following successful completion of the RRQ, the SCSI Target shall respond with ACC. 

9.6 SCSI Target error behavior 

If an error is detected by a SCSI Target while it has Sequence Initiative, the only permissible recovery 

action is the transmission of FCP_RSP with CHECK CONDITION status and an appropriate Sense 

Key/ASC/ASCQ. 

If an error is detected by a SCSI Target while it does not have Sequence Initiative, it shall wait until it 

has been given Sequence Initiative before it can return CHECK CONDITION. In the event that Sequence 

Initiative is not received within RR_TOV, the Target may implicitly terminate the affected Exchange 

(see see 7.6). 

If ULP resources in the SCSI Target are unavailable when a command is issued, the SCSI Target 

shall return TASK SET FULL status in FCP_RSP (i.e., shall not return BUSY). 

If a SCSI Target receives a PLOGI request and it detects a limitation of login resources, 

a) if there are no open tasks with another logged-in SCSI Initiator, that other SCSI Initiator may be 

implicitly logged out so that the PLOGI request can be processed; 

50


b) if all logged-in SCSI Initiators have open tasks, the SCSI Target may respond to the PLOGI 

with LS_RJT and reason code “unable to perform command request” and reason code explanation 

“insufficient resources to support Login”. 

9.7 Task Management and Multiple-Initiator SCSI Targets 

If a SCSI Target Reset or Clear Task Set management function is received by a SCSI Target that has 

multiple SCSI Initiators logged in with it, then the SCSI Target shall: 

a) create a Unit Attention Condition for all other SCSI Initiators (an FCP_RSP may have been 

transmitted but not received by the SCSI Initiator, or the SCSI Initiator may have transmitted a 

command that has not yet been received by the SCSI Target); 

b) clear all resources associated with the cleared Exchanges, per SCSI Architectural Model 

(SAM); 

c) return FCP_RSP upon completion of (a) and (b). The payload shall be zeroes with the exception 

of the FCP_RSP_LEN_VALID bit, FCP_RSP_LEN (which shall be set equal to 8), and the 

FCP_RSP_INFO. 

Upon discovery of the Unit Attention Condition set in a), SCSI Initiators shall issue ABTS for all commands 

that are outstanding for the appropriate LUN or LUNs at that SCSI Target as described in 

9.5.1. From a SCSI Initiator perspective, this is all commands for which FCP_RSP has not been received. 

9.8 SCSI Target Exchange origination capability 

The only time that a SCSI Target is required to have Exchange origination capability is to originate a 

LOGO or PRLO request. All other Exchanges are originated by the SCSI Initiator. 

9.9 Responses to FCP-level Frames before PLOGI or PRLI 

If a SCSI Target receives an FCP_CMND from an NL_Port with which it has not successfully completed 

N_Port Login (PLOGI), it shall discard the FCP_CMND and send LOGO to that NL_Port. No 

Exchange is created in the SCSI Target for the discarded request, and the originator of the discarded 

request terminates the Exchange associated with the discarded request and any other open Exchanges 

for the SCSI Target sending the LOGO. The LOGO is not part of the Exchange associated 

with the discarded request. 

If a SCSI Target receives an FCP_DATA Sequence from an NL_Port with which it has not successfully 

completed N_Port Login (PLOGI), it shall discard all Frames of that Sequence and may send 

LOGO. If LOGO is sent, all open Exchanges for the NL_Port that sent the FCP_DATA Sequence 

shall be terminated. 

If a SCSI Target receives an FCP_CMND from an NL_Port with which it has not successfully completed 

Process Login (PRLI), it shall discard the FCP_CMND and send PRLO to the SCSI Initiator. 

No Exchange is created in the recipient NL_Port for the discarded request, and the originator of the 

discarded request terminates the Exchange associated with the discarded request. 

If a SCSI Target receives an FCP_DATA Sequence from a SCSI Initiator with which it has not successfully 

completed Process Login (PRLI), it shall discard all Frames of that Sequence and may send 

PRLO. 

For the action taken on any other received Frame, see table 14. 

51


10 SCSI features 

This clause describes those SCSI features with usage defined by this document. 

10.1 Auto Contingent Allegiance (ACA) 

During the discovery process, FCP SCSI Initiators shall use INQUIRY to discover whether or not 

Logical Units support ACA. 

ACA shall be used (NACA=1 in CDB) only when both the sending SCSI Initiator and receiving Logical 

Unit support ACA. 

ACA shall not be used (NACA=0 in CDB) when either the sending SCSI Initiator or the receiving 

SCSI Target does not support ACA. If a SCSI Target receives a command with NACA=0 and that 

command fails, autosense data is cleared upon transmission of the FCP_RSP to the faulted SCSI Initiator 

and no ACA condition results. 

10.2 SCSI Status 

SCSI Targets are required to be able to return the following status values. Status values not listed are 

optional (not required to be returned by SCSI Targets, not relied upon by SCSI Initiators). 

– ACA ACTIVE (if ACA is supported by both the SCSI Initiator and SCSI Target); 

– RESERVATION CONFLICT; 

– BUSY; 

– GOOD; 

– CHECK CONDITION; 

– TASK SET FULL (if Tagged Queuing is used). 

52


10.3 SCSI Target Discovery 

When the possibility of a configuration change exists, a SCSI Initiator may want to rediscover the new 

configuration. The SCSI Target Discovery procedure for a SCSI Initiator is: 

For all valid AL_PAs: 

OPN(AL_PA) 

IF OPN successful, THEN 

Send ADISC or PDISC to D_ID = hex ’0000’ || AL_PA 

IF (LOGO returned or Node Name or Port Name changed) THEN 

Send PLOGI to D_ID = hex ’0000’ || AL_PA 

IF PLOGI successful THEN 

IF (no hard address conflicts or application tolerant of hard address conflicts) 

Send PRLI to D_ID = hex ’0000’ || AL_PA 

IF PRLI successful, THEN 

Send FCP_CMND with INQUIRY CDB to D_ID = hex ’0000’ || AL_PA(LUN_0) 

ENDIF 

ENDIF 

ENDIF 

ENDIF 

ENDIF 

NEXT AL_PA 

In order to determine whether or not an OPN was “successful” in the above procedure, an NL_Port 

shall be able to: 

a) detect when an OPN has not been intercepted by the designated AL_PA, or 

b) detect that an R_RDY or CLS has not been received from the AL_PA specified in an OPN within 

E_D_TOV of sending that OPN, or 

c) detect that a CLS was received in response to the OPN. In this case, the Target Discovery procedure 

should be retried at a later time, or 

d) detect that the OPN or Frame Extended Link Service failed through some other means not listed. 

If the SCSI Target Discovery procedure revealed that there was a Hard Address conflict (i.e., some 

NL_Port was unable to acquire its Hard Address), then the application may choose to operate in spite 

of that conflict. If this is the case, then the discovery procedure can continue with the PRLI and subsequent 

SCSI INQUIRY command. If the application is not tolerant of Hard Address conflicts, the 

SCSI Initiator may choose not to use that NL_Port. 

Through the use of the SCSI Target Discovery procedure, the SCSI Initiator has the ability to assemble 

a database of [AL_PA:Node Name:Port Name:N_Port ID] tuples for all responding AL_PAs. 

There are several confirmations a SCSI Initiator can perform on that database to determine which 

SCSI Targets it can continue to communicate with that are not defined by this document. See 10.4 for 

rules regarding completion of open Exchanges. 

NOTE – Not all Initiators may perform the exact steps described in the above algorithm, although a SCSI Initiator 

is required to issue ADISC or PDISC to all SCSI Targets it is logged in with within RR_TOV of receiving 

LIP if it wants to remain logged in with those SCSI Targets. 

53


The ADISC/PDISC procedure is designed to avoid the abnormal termination of all open Exchanges 

when a new device is inserted into the loop, or when a device powers on. 

10.4 Exchange Authentication following LIP 

If the node is preparing to change the login parameters of any NL_Port, the node shall perform explicit 

logout with all NL_Ports currently logged-in with that NL_Port before allowing the new parameters 

to take effect. If there are open and active Sequences with the NL_Ports to be logged out, the 

device shall discard any Frames received in those Sequences and send a LOGO, if one has not already 

been sent. 

10.4.1 SCSI Initiator Exchange Authentication 

Following transmission of the CLS signalling completion of loop initialization, SCSI Initiators are required 

to authenticate each SCSI Target with which they have completed PLOGI. 

54 

a) A SCSI Initiator shall suspend execution of all open tasks with a SCSI Target following loop initialization 

until authentication is complete. Following transmission or forwarding of a LIP, only 

Frames related to SCSI Target authentication (i.e., ADISC, PDISC, or PLOGI) shall be transmitted 

to that SCSI Target before Frames for other Exchanges to that SCSI Target are transmitted. 

b) For each SCSI Target with which it has successfully completed PLOGI, a SCSI Initiator shall 

originate ADISC or PDISC to that SCSI Target such that the ADISC or PDISC request Sequence 

arrives at each SCSI Target within RR_TOV of completing loop initialization. 

c) For every SCSI Target that returns an ACC, if the [N_Port ID:Port Name:Node Name] triplet for 

that SCSI Target: 

1) does not match the triplet of a logged-in SCSI Target, the SCSI Initiator shall transmit LOGO 

to that SCSI Target. 

2) does match the triplet of a logged-in SCSI Target, the SCSI Initiator may resume all tasks 

with that SCSI Target. 

d) SCSI Initiator Exchange authentication is complete when the ACC to an ADISC or PDISC has 

been received. If the SCSI Initiator fails to receive an ACC within R_A_TOV, it shall implicitly logout 

the SCSI Target and terminate all tasks for that SCSI Target. 

e) Following loop initialization, any Frames received (other than ACC to an ADISC, PDISC, or 

PLOGI, or a LOGO request) from a SCSI Target that has not been authenticated shall be discarded. 

This means they shall not be considered part of any valid Exchange, and only R_RDYs 

to maintain BB_Credit shall be sent in response to these Frames. 

f) If a SCSI Initiator receives a LOGO from an NL_Port during this procedure, it shall terminate all 

open Exchanges with that NL_Port and send an ACC. 

10.4.2 SCSI Target Exchange Authentication 

Following completion of loop initialization, SCSI Targets are required to wait for Exchange authentication 

by the SCSI Initiators with which they have completed PLOGI. 

a) For each SCSI Initiator that has open tasks, a SCSI Target shall suspend tasks associated with 

that SCSI Initiator until an ADISC or PDISC is received from that SCSI Initiator. Following


transmission or forwarding of a LIP, only Frames related to SCSI Initiator authentication (i.e., 

ACC to an ADISC or PDISC) shall be transmitted to that SCSI Initiator before Frames for other 

Exchanges to that SCSI Initiator are transmitted. 

b) For every SCSI Initiator that sends an ADISC or PDISC, if the 

[N_Port_ID:Port_Name:Node_Name] triplet for that SCSI Initiator: 

1) does not match the triplet of a logged-in SCSI Initiator, the SCSI Target shall transmit LOGO 

to that SCSI Initiator and terminate all Exchanges associated with that SCSI Initiator. 

2) does match the triplet of a logged-in SCSI Initiator, the SCSI Target shall send the ACC to 

the extended link service. The SCSI Target may then resume all tasks with that SCSI Initiator. 

c) If a SCSI Target does not receive a PDISC or ADISC from each logged-in SCSI Initiator within 

RR_TOV of completing loop initialization, then it may implicitly logout that SCSI Initiator and 

terminate all Exchanges associated with that SCSI Initiator. 

d) SCSI Target authentication is complete when the ACC to an ADISC or PDISC has been transmitted. 

e) Following loop initialization, any Frames received (other than ADISC, PDISC, or PLOGI) from a 

SCSI Initiator that has not been authenticated shall be discarded. Only R_RDYs to maintain 

BB_Credit shall be sent in response to those Frames. 

55


10.5 Clearing effects of ULP, FCP, FC-PH, and FC-AL actions 

Table 23 summarizes which FCP SCSI Target objects are cleared as a result of a listed FCP SCSI 

Initiator action. A ‘Y’ in the corresponding column indicates the object is cleared to its default or power-on 

value within the device upon completion of specified action. An ‘N’ in the corresponding column 

indicates the object is not affected by the specified action. A ‘-’ in the column indicates that the action 

is not applicable. 

Table 23 – Clearing Effects of SCSI Initiator Actions 

FCP SCSI Initiator Action Clears Object 

FCP SCSI Target Object 

PLOGI parameters 

For all logged-in SCSI Initiator ports 

Only for SCSI Initiator port initiating action 

Open Sequences Terminated 

For all SCSI Initiator ports with open Sequences 


Only for Sequences associated with Aborted Exchange 

Login BB_Credit_CNT 

For all Logged-In L_Ports 

For transmitting L_Port only 

Power 

Cycle 

Y Y - N N N N N N 

- - Y 

Hard Address Acquisition Attempted Y 1 Y 1 N N N N N N N 

PRLI parameters cleared 

For all logged-in SCSI Initiator ports 

Only for ports of specified TYPE 

Only for port initiating action 

Y 

- 

- 

Y 

- 

- 

N 

N 

Y 

N 

N 

N 

N 

Y 

Y 

N 

Y 

N 

N 

N 

N 

N 

N 

N 

N 

N 

N 

Open Tasks (Exchanges) Aborted 

All tasks for all SCSI Initiator ports with open tasks 

All tasks, only for port initiating action 

Only specified task for port initiating action 

SCSI Target mode page parameters restored from saved 

pages (when saved pages are supported, or default mode 

page parameters when saved pages are not supported) 

For all SCSI Initiators 


Pre-existing ACA, UA 7 and Deferred error conditions 

cleared 



Device Reservations 



Y 

- 

Y 

- 

- 

Y 

- 

- 

Y 

- 

Y 

- 

Y 

- 

Reset 

LIP(y,x) 3 

Y 

- 

Y 

- 

- 

Y 

- 

- 

Y 

- 

Y 

- 

Y 

- 

LOGO, 

PLOGI 

N 

Y 

N 

Y 

- 

N 

Y 

N 

N 

Y 

N 

Y 

N 

Y 6 

ABTS 

w/Last Seq. 

N 

N 

N 

N 

Y 

N 

N 

Y 

N 

N 

N 

N 

N 

N 

PRLI 8 , 

PRLO 

N 

N 

N 

Y 

- 

N 

Y 

N 

N 

Y 

N 

Y 

N 

Y 

TPRLO 5 

N 

N 

Y 

- 

- 

Y 

- 

- 

Y 

- 

Y 

- 

Y 

- 

SCSI Target 

Reset 

N 

N 

Y 2 

- 

- 

Y 

- 

- 

Y 

- 

Y 

- 

Y 

- 

Clear 

Task Set 4 

N 

N 

N 

Y 2 

- 

Y 

- 

- 

N 

N 

N 

N 

N 

N 

Abort 

Task Set 4 

N 

N 

N 

Y 2 

- 

N 

Y 

N 

N 

N 

N 

N 

N 

N 

56


NOTES: 

Table 23 – Clearing Effects of SCSI Initiator Actions 

FCP SCSI Initiator Action Clears Object 

1 If the NL_Port has an AL_PA different that its hard address and the NL_Port experiences a power cycle or 

recognizes LIP(AL_PD,AL_PS), the NL_Port shall relinquish its current AL_PA and attempt to acquire its 

hard address. 

2 Tasks are cleared internally within the SCSI Target, but open Sequences shall be individually aborted by 

the SCSI Initiator via the ABTS_LS protocol that also has the effect of aborting the associated Exchange. 

3 This is also known as LIP(AL_PD,AL_PS). If the destination recognizes a selective hard reset LIP where 

the AL_PD matches the AL_PA of the receiving NL_Port, the receiving NL_Port shall perform the behavior 

described in this column. 

4 For multiple-LUN SCSI Targets, Clear Task Set, and Abort Task set affect only the addressed LUN, not 

the entire SCSI Target. 

5 Actions listed shall be performed when the Global bit = ‘1’b. If the Global bit = ‘0’b, then the actions listed 

under PRLI/PRLO should be performed for the designated SCSI Initiator. 

6 This includes explicit and implicit LOGO/PLOGI. 

7 After clearing the pre-existing UA the Target shall set initial UA condition(s) to its initial power-on value. 

8 The Target shall clear the object only if “establish image pair” = 1. 

57


11 SCSI Stream Devices 

This clause describes commands and features applicable to SCSI stream devices with usage defined 

by this profile. 

11.1 Applicable Classes of Service 

SCSI-3 stream devices conforming to this document shall use Class 2 or Class 3 delivery service 

with parameters as described in clause 5. 

11.2 Asynchronous Event Notification (AEN) 

The use of AEN by stream devices conforming to this profile is prohibited. 

11.3 Command Linking 

Command Linking is allowed by all stream devices conforming to this profile. Support for command 

linking is identified in the Inquiry data. The Flag bit of the CDB shall be set to zero. 

11.4 Sequential device commands 

Commands, and features within commands, that are not listed are optional. Interoperability between 

SCSI Initiators and SCSI Targets is not guaranteed if optional commands or features are used. If 

support of a field is listed as Required without specifying a value for that field, it is assumed that all 

possible values of the bit or field shall be supported in accordance with the appropriate American National 

Standard. Any unlisted commands/features/settings are implicitly Invokable by the SCSI Initiator, 

and Allowed by the SCSI Target. 

Table 24 – SCSI Tape Device Commands 

Feature Initiator Target Doc Notes 

ERASE I R 

IMMED I R 

LONG I R 

SSC 

SHORT I A 

FORMAT MEDIUM I A SSC 

INQUIRY I R 

Standard INQUIRY data (bytes 0-35) I R 

EVPD=1 I R 

Vital Product Data page codes: 

hex‘00’ (Supported vital product pages) I R SPC 

hex‘80’ (Unit serial number page) I R 

hex‘81’ (Implemented operations definition page) I A 

hex‘82’ (ASCII implemented operations definition page) I A 

hex‘83’ (Device identification page) I R 

LOAD UNLOAD I R 

EOT = 1 I A 

Immed = 1 I R SSC 

Load = 1 I A 

Load = 0 I R 

Reten = 1 I A 

58




LOCATE I R 

BT = 0 I R 

BT = 1 I A 

CP = 0 I R SSC 

CP = 1 I A 

Immed=0 I R 

Immed=1 I R 

LOG SELECT I A 

SP = 0 I R 

SP = 1 I A 

PCR = 1 & PLL = 0 I R 

PC = ‘00’b I A 

SPC 

PC = ‘01’b I R 

PC = ‘10’b I A 

PC = ‘11’b I A 

LOG SENSE I R 

SP = 0 I R 

SP = 1 I A 

PPC = 0 I R 

PPC = 1 I A 

PC = ‘00’b I A 

PC = ‘01’b I R 

PC = ‘10’b I A 

SPC 

PC = ‘11’b I A 

Page Select hex’00’ I R 




Page Select hex’0C’ I R 

MODE SENSE/MODE SELECT (6 and 10) I R SPC 

DBD = 0 I R 

DBD = 1 I R 

PC = ‘00’b I A 

PC = ‘01’b I R 

PC = ‘10’b I A 

PC = ‘11’b I A 

Page hex’01’ I R SSC 

Page hex’02’ I R SPC 

Page hex’0A’ I R SSC 

Page hex’10’ I R SSC 

Pages hex’11-14’ I A SSC 

Page hex’19’ I R 

FCP- 

2 

MOVE MEDIUM ATTACHED I A SMC 1 

PERSISTENT RESERVE IN I A SPC 

PERSISTENT RESERVE OUT I A SPC 

PREVENT/ALLOW MEDIUM REMOVAL I A SSC 

59




READ I R 

Fixed = 1 I R 

Fixed = 0 (Required if Read Block Limits values differ) I A SSC 

SILI = 0 I R 

SILI = 1 (Required if Read Block Limits values differ) I A 

READ BLOCK LIMITS I R SSC 

READ BUFFER I A SPC 

READ ELEMENT STATUS ATTACHED I A 

1 

Vol Tag = 0 I R SMC 1 

Vol Tag = 1 I A 1 

READ POSITION I R 

BT = 0 I R SSC 

BT = 1 I A 

READ REVERSE I A 

Fixed = 1 I R 

Fixed = 0 (Required if Read Block Limits values differ) I A 

SILI = 0 I R SSC 


Byte Ord = 0 I A 

Byte Ord = 1 I A 

RECEIVE DIAGNOSTIC RESULTS I A SPC 

RECOVER BUFFERED DATA I R 

Fixed = 1 I R 


SILI = 0 I R SSC 


FIFO I A 

LIFO I R 

RELEASE(6) I A SPC 

RELEASE(10) I R 

3rd Party = 0 I R SPC 

3rd Party = 1 I A 

REPORT DENSITY SUPPORT I R SSC 

REPORT LUN I A SPC 

RESERVE(6) I A SPC 

RESERVE(10) I R 

3rd Party = 0 I R SPC 

3rd Party = 1 I A 

REWIND I R 

Immed = 0 I R SSC 

Immed = 1 I R 

SEND DIAGNOSTIC I R 

ST = 1 I R SPC 

ST = 0 I A 

SPACE I R 

Forward I R SSC 

Reverse I R 

60




Code = ‘000’b (Blocks) I R 

Code = ‘001’b (Filemarks) I R 

Code = ‘010’b (Sequential Filemarks) I A 

Code = ‘011’b (End-of-Data) I A 

Code = ‘100’b (Setmarks) I A 

Code = ‘101’b (Sequential Setmarks) I A 

TEST UNIT READY I R SPC 

VERIFY I A 

Fixed = 1 I R 


BytCmp = 0 I R 

BytCmp = 1 I A 

SSC 

Immed = 0 I R 

Immed = 1 & BytCmp = 0 I R 

Immed = 1 & BytCmp = 1 I A 

WRITE I R 

Fixed = 1 I R SSC 


WRITE BUFFER I A 

Mode = ‘110’b (Download ucode w/ offsets) I R SPC 

Mode = ‘111’b (Download ucode w/ offsets & save) I R 

WRITE FILEMARK I R 

Count = 0 I R 

Count > 0 I R SSC 

Immed = 0 I R 

Immed = 1 I R 

NOTES: 

1 Medium changer commands and features are disabled when Inquiry data indicates MCHGR = ‘0’b and 

enabled when the MCHGR = ‘1’b. See Table 9.3 for a description of the command(s). 

61


12 SCSI-3 Media Changer Devices 

This clause describes commands and features applicable to SCSI media changer devices with usage 

defined by this profile. 

12.1 Applicable Classes of Service 

SCSI-3 media changer devices conforming to this document shall use Class 2 or Class 3 delivery 

service with parameters as described in clause 5. 

12.2 Asynchronous Event Notification (AEN) 

The use of AEN by media changer devices conforming to this profile is prohibited. 

12.3 Command Linking 

Command Linking is allowed by all media changer devices conforming to this profile. Support for 

command linking is identified in the Inquiry data. The Flag bit of the CDB shall be set to zero. 

12.4 Media Changer Device Commands 

Commands, and features within commands, that are not listed are optional. Interoperability between 

SCSI Initiators and SCSI Targets is not guaranteed if optional commands or features are used. If 

support of a field is listed as Required without specifying a value for that field, it is assumed that all 

possible values of the bit or field shall be supported in accordance with the appropriate American National 

Standard. Any unlisted commands/features/settings are implicitly Invokable by the SCSI Initiator, 

and Allowed by the SCSI Target. 

Table 25 – SCSI Independent Media Changer Device Commands 


CHANGE DEFINITION I A SPC 

EXCHANGE MEDIUM I A SMC 

INITIALIZE ELEMENT STATUS I A SMC 

INQUIRY I R 

Standard INQUIRY data (bytes 0-35) I R 

EVPD = 1 I R 

Vital Product Data page codes: 

hex’00’ (Supported vital product pages) I R SPC 

hex’80’ (Unit serial number page) I R 

hex’81’ (Implemented operations definition page) I A 

hex’82’ (ASCII implemented operations definition page) I A 

hex’83’ (Device identification page) I R 

LOG SELECT I A SPC 

MODE SENSE/MODE SELECT (6 and 10) I R SPC 

DBD = 0 I R 

DBD = 1 I R 

PC = ‘00’b I A 

PC = ‘01’b I R 

PC = ‘10’b I A 

PC = ‘11’b I A 

Page hex’1D’ I R SMC 

Page hex’1E’ I R SMC 

62


Table 25 – SCSI Independent Media Changer Device Commands 


Page hex’1F’ I R SMC 

MOVE MEDIUM I R SMC 

PERSISTENT RESERVE IN I A SPC 

PERSISTENT RESERVE OUT I A SPC 

POSITION TO ELEMENT I A SMC 

PREVENT ALLOW MEDIUM REMOVAL I A SPC 

READ BUFFER I A SPC 

READ ELEMENT STATUS I R SMC 

RECEIVE DIAGNOSTIC RESULTS I A SPC 

RESERVE ELEMENT/RELEASE ELEMENT (6) I A 

Element = 0 I R SMC 

Element = 1 I A 

RESERVE ELEMENT/RELEASE ELEMENT (10) I R 

3rd Party = 0 I R 

3rd Party = 1 I A SMC 

Element = 0 I R 

Element = 1 I A 

REQUEST VOLUME ELEMENT ADDRESS I A SMC 

REQUEST SENSE I A SPC 

SEND DIAGNOSTIC I R 

ST = 0 I A SPC 

ST = 1 I R 

SEND VOLUME TAG I A SMC 

TEST UNIT READY I R SPC 

WRITE BUFFER I A 

Mode = ‘110’b (Download ucode w/ offsets) I R SPC 

Mode = ‘111’b (Download ucode w/ offsets & save) I R 

Table 26 – SCSI Attached Media Changer Device Commands 


MOVE MEDIUM ATTACHED (A7h) I R SMC 

MOVE MEDIUM ATTACHED (A5h) I A SMC 1 

READ ELEMENT STATUS ATTACHED (B4h) I R SMC 

READ ELEMENT STATUS ATTACHED (B8h) I A SMC 1 

NOTES: 

1 Sequential devices (primary device type 1) may also use A5h for 

Move Medium Attached and B8h for Read Element Status Attached 

63


64


Annex A 

(informative) 

Error Detection and Recovery Extended Link Services 

A.1 Introduction 

This annex documents the Extended Link Services required to implement error detection and recovery 

for streaming and media changer devices using Class 3 or Class 2 delivery service. The ELS’s 

will be incorporated into a future version of FC-PH. 

R_CTL = hex ‘22/23’ 

A.2 Sequence Retransmission Request (SRR) 

Encoded Value (Bits 31-24) = hex ‘14’ 

The SRR (Sequence Retransmission Request) Extended Link Service request Sequence requests 

an N_Port to retransmit information for the RX_ID or OX_ID originated by the S_ID originating the request 

Sequence. The specification of OX_ID and RX_ID may be useful or required information for the 

destination N_Port to locate the information requested. A Responder destination N_Port would use 

the RX_ID and ignore the OX_ID, unless the RX_ID was undetermined (i.e., RX_ID = 0xffff). An Originator 

N_Port would use the OX_ID and ignore the RX_ID. This function provides the N_Port transmitting 

the request with the ability to request that information not correctly received by either the 

Exchange Originator or Exchange Recipient to be retransmitted so that the Exchange may be completed 

normally. 

If the destination N_Port of the SRR request determines that the Originator S_ID , OX_ID, or RX_ID 

are inconsistent, then it shall reply with an LS_RJT Sequence with a reason code that it is unable to 

perform the command request. 

The SRR shall be sent in a new Exchange. 

Sequence Initiative shall be transferred to the Exchange Responder to resend the requested Sequence. 

The Sequence count for retransmitted FCP_DATA shall start at zero. 

Protocol: 

Sequence Retransmission request Sequence 

Accept (ACC) reply Sequence 

Format: 

FT_1 

Addressing: 

The S_ID field designates the source N_Port requesting the information retransmission. The D_ID 

field designates the destination N_Port to which the request is being made. 

65


In the event that the Target cannot accept this request, the Target shall present a check condition as 

if it had not responded to an Initiator Detected Error with a Restore Pointers message (i.e., Sense 

Key = 4, ASC/ASQ = 48/00). The Target shall not reject requests for retransmission of 

FCP_XFER_RDY or FCP_RSP frames unless the SRR is not supported. 

Payload: 

The format of the Payload is shown in the following table. The amount of data to transfer is implicitly 

the remainder of that for the Exchange. 

SRR Payload 

Item: 

Size 

(bytes): 

hex ‘14000000’ 4 

OX_ID 2 

RX_ID 2 

Relative Offset 4 

R_CTL for IU 1 

Reserved 3 

The R_CTL for IU field encoding is as described in FC-PH section 18.2, Tables 28, 29; i.e., 05 = Data 

Descriptor (FCP_XFER_RDY), 07 = Command Status (FCP_RSP), 01 = Solicited Data 

(FCP_DATA). 

The Relative Offset parameter is only meaningful if the R_CTL for IU field is set to 01 for Solicited Data. 

Reply Link Service Sequence 

Service Reject (LS_RJT) 

Signifies rejection of the SRR request. 

ACC 

Signifies that the information will be retransmitted. - 

Accept payload: 

66


SRR Accept 

Payload 

Item 

Size - 

Bytes 

Hex ‘02000000’ 4 

The new SRR reject reason code is defined below. 

Encoded Value 

Hex ‘00052A00’ 

LS_RJT Reason Code 

explanation 

Cannot resend requested 

Sequence or data at requested 

offset 

A.3 Read Exchange Concise (REC) 

Encoded Value (Bits 31-24) = hex ‘13’ 

The REC (Read Exchange Concise) Extended Link Service request Sequence requests an N_Port to 

return Exchange information for the RX_ID or OX_ID originated by the S_ID specified in the Payload 

of the request Sequence. The specification of OX_ID and RX_ID may be useful or required information 

for the destination N_Port to locate the status information requested. A Responder destination 

N_Port would use the RX_ID and ignore the OX_ID, unless the RX_ID was undetermined (i.e., 

RX_ID = 0xffff). If the RX_ID is unspecified in the request, the Target must identify the Exchange by 

means of the S_ID and OX_ID. An Originator N_Port would use the OX_ID and ignore the RX_ID. 

This function provides the N_Port transmitting the request with information regarding the current status 

of the Exchange specified. 

If the destination N_Port of the REC request determines that the Originator S_ID , OX_ID, or RX_ID 

are inconsistent, then it shall reply with an LS_RJT Sequence with a reason code that it is unable to 

perform the command request. 

The REC shall be sent in a new Exchange. 

Protocol: 

Read Exchange Concise request Sequence 

Accept (ACC) reply Sequence 

Format: 

FT_1 

Addressing: 

67


The S_ID field designates the source N_Port requesting the Exchange information. The D_ID field 

designates the destination N_Port to which the request is being made. 

Payload: 

The format of the Payload is shown in the following table. 

Reply Link Service Sequence: 

Service Reject (LS_RJT) 

Signifies rejection of the REC request. 

ACC 

Signifies that the N_Port has transmitted the requested data. 

- Accept payload: 

REC Payload 

Item 

Size - 

Bytes 

Hex ‘13000000’ 4 

Reserved 1 

Exchange Originator 

S_ID 

3 

OX_ID 2 

RX_ID 2 

- The format of the Accept Payload is shown in the table below. The format of the Concise Exchange 

Status is specified below. 

REC Accept Payload 

Item 

Size - 

Bytes 

Hex ‘02000000’ 4 

Concise Exchange 

Status 

24 

68


Concise Exchange 

Status 

Item 

Size - 

Bytes 

OX_ID 2 

RX_ID 2 

Originator Address 

Identifier 

4 

Responder Address 

Identifier 

4 

Data Transfer Count 4 

E_STAT 4 

E_STAT is as defined in FC-PH Sec. 24.8.1. for the Exchange Status Block. The bits specifying 

whether the Exchange is complete (Bit 29) and whether the responder holds Sequence Initiative (Bit 

30) must be valid; the setting of other bits is not required. 

The Originator Address Identifier is set to the S_ID with the high-order byte reserved. 

The Responder Address Identifier is set the the D_ID with the high-order byte reserved. 

The Data Transfer Count is the count of bytes delivered to the ULP for a write or the count of bytes 

received from the ULP and transmitted to the Initiator for a read. 

A.4 ABTS Changes 

The Abort Sequence Basic Link Service frame shall be used by the Sequence Initiator or Sequence 

Recipient to request that the ABTS Recipient abort one or more sequences, abort the entire Exchange, 

or check the Exchange status, based on the setting of the bits in the parameter field. 

Table 27 – ABTS Parameter Field Bit Definitions 

Parameter Field 

Bit 0 = 1 

Bit 1 = 1 

Bit 2 = 1 

Meaning 

Abort Exchange 

Abort Sequence 

Check Status 

69


70


Annex B 

(informative) 

Error Detection and Recovery Diagrams 

Figure 8 – Lengthy FCP_CMND or Lost ACK 

Error Detection 

Class 3 

Class 2 

FCP_CMND 

FCP_CMND 

FCP_TOV 

REC 

FCP_TOV 

or 

E_D_TOV 

X 

ACK 

ACC 

REC 

ACC 

Error Recovery 

None: the ACC for the REC indicates the Exchange is open and the 

Target holds Sequence Initiative. 

No error recovery is required. 

71


Figure 9 – FCP_CMND Lost 


Class 3 

Class 2 

FCP_TOV 

FCP_CMND 

REC 

X 

LS_RJT 

FCP_TOV 

or 

E_D_TOV 

FCP_CMND 

REC 

X 

LS_RJT 


The LS_RJT for the REC indicates the 

Exchange is unknown. The FCP_CMND is 

retransmitted in a new Exchange using the 

same CRN. 

FCP_CMND 

ACK 

72


Figure 10 – FCP_XFER_RDY Lost 


Class 3 

Class 2 

FCP_CMND 

FCP_CMND 

FCP_TOV 

X 

REC 

FCP_XFER_RDY 

(RO=0) 

FCP_TOV 

X 

ACK 

FCP_XFER_RDY 

(RO=0) 

E_D_TOV 

ACC 

ABTS 


or after 

sending 

BA_ACC 

BA_ACC 

REC 

R_A_TOV 

The ACC for the REC indicates the Initiator 

holds Sequence Initiative and the Exchange is 

open. The Initiator sends an SRR requesting 

the FCP_XFER_RDY be resent. The Target 

resends the FCP_XFER_RDY using the same 

Relative Offset. 

ACC 

ACC 

RRQ 

SRR 

ACC 

FCP_XFER_RDY (RO=0) 

ACK 

73


Figure 11 – FCP_XFER_RDY Received, ACK Lost (Class 2) 


Class 2 

FCP_CMND 

ACK 

ACK 

FCP_XFER_RDY 

X 

ABTS 

E_D_TOV 

Send 

REC after 

sending 

BA_ACC 

BA_ACC 

REC 

R_A_TOV 

ACC 

ACC 

RRQ 


None: the ACC for the REC indicates the 

Initiator holds Sequence Initiative and the 

Exchange is open. No error recovery is 

required. 

74


Figure 12 – FCP_RSPLost 


Class 3 

Class 2 

FCP_CMND 

FCP_CMND 

FCP_TOV 

X 

REC 

FCP_RSP 

FCP_TOV 

X 

ACK 

FCP_RSP 

E_D_TOV 

ACC 

ABTS 

BA_ACC 


or after 

sending 

BA_ACC 

REC 

ACC 

R_A_TOV 

The ACC for the REC indicates the Initiator 

holds Sequence Initiative and the Exchange is 

complete. The Initiator sends an SRR 

requesting the FCP_RSP be resent. The 

Target resends the FCP_RSP. 

ACC 

RRQ 

SRR 

ACC 

FCP_RSP 

ACK 

75



Figure 13 – FCP_RSP Received, ACK Lost (Class 2) 

Class 2 

FCP_CMND 

ACK 

ACK 

BA_RJT 

FCP_RSP 

X 

ABTS 

E_D_TOV 


None: the BA_RJT for the ABTS indicates the 

Exchange is unknown. No error recovery is 

required. 

76


Figure 14 – Lost Write Data, Last Frame of Sequence 

Class 3 


Class 2 

FCP_CMND 


FCP_DATA 

(seq=1,cnt=0,RO=0) 

FCP_CMND 

ACK 


ACK 

FCP_TOV 

FCP_DATA 

(seq=1,cnt=1) 

REC 

X 

ACC 

FCP_TOV 

or 

E_D_TOV 

FCP_DATA 


FCP_DATA 

(seq=1,cnt=1) 

REC 

X 


ACC 

The ACC for the REC indicates the Target 

does not hold Sequence Initiative and the 

Exchange is open. The Initiator sends an SRR 

requesting a FCP_XFER_RDY be resent using 

the specified Relative Offset. The Target 

resends the FCP_XFER_RDY using the 

specified Relative Offset (or a a Relative Offset 

smaller than the Relative Offset specified in 

the SRR in order to be aligned on an 

appropriate boundary in the Target). 

SRR (RO=0) 

ACC 

ACK 


FCP_DATA (seq=2,cnt=0,RO=0) 

FCP_DATA (seq=2,cnt=1) 

ACK 

77


Figure 15 – Lost Write Data, Not Last Frame of Sequence 


FCP_TOV 

FCP_CMND 

Class 3 


FCP_DATA 


FCP_DATA 

(seq=1,cnt=1) 

REC 


X 

ACC 

The ACC for the REC indicates the Target may 

or may not hold Sequence Initiative and the 

Exchange is open. The Initiator sends an SRR 

requesting a FCP_XFER_RDY be resent using 

the specified Relative Offset. The Target 

resends the FCP_XFER_RDY using the 

specified Relative Offset (or a Relative Offset 

smaller than the Relative Offset specified in 

the SRR in order to be aligned on an 

appropriate boundary in the Target). 

Note: sequence 

error detected 

R_A_TOV 

FCP_CMND 

Class 2 

ACK 


ACK 

FCP_DATA 


REC 

X 

FCP_DATA 

(seq=1,cnt=1) 

ACK (Abort) 

ABTS (Abort Sequence) 

BA_ACC 

RRQ 

ACC 

BA_ACC 

Note: sequence error 

detected 

SRR (RO=0) 

ACC 

ACK 




ACK 

78


Figure 16 – Lost Read Data, Last Frame of Sequence 


Note: 

sequence 


X 

FCP_CMND 

REC 

Class 3 

FCP_DATA 


FCP_DATA 

(seq=1,cnt=1) 

FCP_RSP 

ACC 




Exchange is complete for Class 3 and open for 

Class 2. The Initiator sends an SRR requesting 

the FCP_DATA be resent starting at the 

specified Relative Offset. The Target resends 

the FCP_DATA starting at the specified 


X 

FCP_CMND 

REC 

Class 2 

ACK 

FCP_DATA 


FCP_DATA 

(seq=1,cnt=1) 


BA_ACC 

BA_ACC 

ACC 

RRQ 

E_D_TOV 

R_A_TOV 

SRR (RO=0) 

ACC 


ACK 

ACK 


FCP_RSP 

79


Figure 17 – Lost Read Data, Not Last Frame of Sequence 


Class 3 

Class 2 

FCP_CMND 

FCP_CMND 

Note: 

sequence 


X 

REC 

FCP_DATA 


FCP_DATA 

(seq=1,cnt=1) 

FCP_RSP 

Note: 

sequence 


X 

ACK 

FCP_DATA 


FCP_DATA 

(seq=1,cnt=1) 

ACK (Abort) 


E_D_TOV 

ACC 

BA_ACC 

REC 

R_A_TOV 

ACC 




Exchange is complete for Class 3 and open for 

Class 2. The Initiator sends an SRR requesting 

the FCP_DATA be resent starting at the 

specified Relative Offset. The Target resends 

the FCP_DATA starting at the specified 


BA_ACC 

RRQ 

SRR (RO=0) 

ACC 

ACK 

ACK 



FCP_RSP 

80


Figure 18 – ACK Lost on Read (Class 2) 


Class 2 

FCP_CMND 

ACK 

ACK 

BA_ACC 

FCP_DATA 


X 

ABTS 

FCP_RSP 

RRQ 

E_D_TOV 

R_A_TOV 

Note: The BA_ACC 

indicates the 

FCP_DATA sequence 

was received, the 

Target continues the 

Exchange. 

ACC 


None: the Initiator has received the 

FCP_DATA frame or sequence.No error 

recovery is required. 

81


Figure 19 – REC or REC Response Lost 

If no response (i.e. ACC or LS_RJT) for an 

REC is received within E_D_TOV send an 

ABTS for the REC. Retransmit another REC in 

a new Exchange. 

E_D_TOV 

REC 

REC 

or 

X 

X 

ACC 

ABTS (for REC) 

LS_RJT 

R_A_TOV 

REC 

ACC 

RRQ 

ACC 

82


Figure 20 – SRR or SRR Response Lost 

If no response (i.e. ACC or LS_RJT) for an SRR 

is received within E_D_TOV send an ABTS for 

the SRR. Retransmit another SRR in a new 

Exchange. 

E_D_TOV 

SRR 

SRR 

or 

X 

X 

ACC 

ABTS (for SRR) 

LS_RJT 

R_A_TOV 

SRR 

ACC 

RRQ 

ACC 

83


84


Annex C 

(informative) 

FCP_CONF Implementation and Usage 

This annex describes and diagrams the use of FCP_CONF. 

SCSI Initiator and Target devices conforming to this profile are required to support FCP_CONF IUs. 

C.1 FCP_CONF Information Unit 

Table 28 – FCP information units sent to targets 

IU 

SCSI-3 

Primitive 

CAT 

Data Block 

Content 

F/M/L 

SI 

M/O 

T12 Confirm 3 FCP_CONF 

L H O 

C.2 FCP_CONF Process Login (PRLI) Parameter 

C.2.1 

Word 3, Bit 7: FCP_CONF Support 

When this bit is set to 1, the process defined by the page is indicating that it has the capability of supporting 

FCP_CONF IUs. When this bit is set to 0, the process defined by the page is indicating that it 

is not capable of supporting FCP_CONF IUs. If the responder does not support FCP_CONF IUs it 

shall set the bit to 0 in the PRLI accept page. 

C.3 FCP_CONF Request Returned in FCP_RSP 

A FCP_CONF request may be sent by the Target in the FCP_RSP by setting the FCP_CONF_REQ 

bit in the FCP_STATUS field. 

Table 29 – FCP_STATUS field format 

Bit Definition 

Bit 

Byte 2 

FCP_CONF_REQ 4 

FCP_CONF_REQ indicates the Target has requested a FCP_CONF for the Exchange. 

C.4 FCP_CONF Guidelines 

1. The Initiator shall send a FCP_CONF IU after receiving a FCP_RSP from the Target with the 

FCP_CONF_REQ bit set in the FCP_STATUS field. 

2. The Initiator shall release Exchange information after the FCP_CONF is sent or when a response 

to an Exchange abort is received. 

3. The Target shall retain Exchange information until a FCP_CONF is received or the Exchange is 

aborted. 

4. ABTS shall be sent by the Target if no FCP_CONF IU is received by the Target within FCP_TOV 

85


after a FCP_RSP has been sent. 

5. The use of FCP_CONF is required for Tagged Tasks (i.e. command queuing). 

Figure 21 – FCP_CONF Example 

FCP_TOV 

FCP_CMND 

FCP_RSP (FCP_CONF_REQ=1) 

. 

. 

FCP_CONF 

FCP_TOV 

A FCP_RSP requesting a FCP_CONF is sent by the Target. The FCP_CONF is sent by the Initiator 

and received by the Target within FCP_TOV. 

FCP_CMND 

Figure 22 – FCP_CONF Lost 

FCP_TOV 

FCP_RSP (FCP_CONF_REQ=1) 

FCP_CONF 

X 

FCP_TOV 

ABTS 

BA_RJT (Invalid OX_ID-RX_ID combination) 

If a FCP_CONF is sent by the Initiator and not received at the Target the timer will expire and the Target 

sends an ABTS. The Initiator does not know of the Exchange and returns a BA_RJT with an “Invalid 

OX_ID-RX_ID combination” reason code explanation. The Target knows the FCP_CONF was 

sent and releases the Exchange information. 

If the Initiator has not 

If another FCP_CMND is received with the same OX_ID as an Exchange waiting for a FCP_CONF 

the Target may release the Exchange information. 

86


Annex D 

(informative) 

Command Reference Number (CRN) Implementation and Usage 

This annex describes the implementation and usage of Command Reference Number (CRN). 

The CRN is contained in Byte 0 of the FCP_CNTL field in the FCP_CMND IU and is based on an 

I_T_L nexus. Support for CRN is negotiated based on a I_T nexus via PRLI. The use of CRN is enabled 

based on a I_T_L nexus via FC Mode Page (0x19). 

D.1 Command Reference Number Process Login (PRLI) Parameter 

D.1.1 

Word 3, Bit 8: Command Reference Number Support (CRNS) 

When this bit is set to 1, the process defined by the page is indicating that it is capable of supporting 

CRN. When this bit is set to 0, the process defined by the page is indicating that it is not capable of 

supporting CRN. If the responder does not support CRN it shall set the bit to 0 in the PRLI accept 

page. 

D.2 FC Mode Page (0x19) Parameter 

D.2.1 

Byte 4, Bit 0: Enable Command Reference Number (ECRN) 

When the Enable Command Reference Number (ECRN) bit is one, the Initiator shall use a continuously 

increasing Command Reference Number for each Exchange with the Lun (except Task Management 

Functions). When the ECRN bit is zero, the Initiator shall not use a Command Reference 

Number for each Exchange with the Lun. 

D.3 Guidelines 

1. The CRN shall be equal to 1 for the first Exchange between the Initiator and Target and shall be 

continuously increasing. 

2. The CRN shall wrap from 255 to 1 (i.e. a value of 0 in the CRN field is not valid for an Exchange 

using CRN). 

3. A PRLI, Target Reset Task Management Function, and Lun Reset Task Management Function 

shall reset the CRN to 1. 

4. The Initiator shall not reuse a CRN until delivery has been confirmed via a FCP_XFER_RDY, 

FCP_DATA, FCP_RSP IU or a REC or ACK. 

5. The Target shall not execute out-of-order SCSI commands and shall hold the SCSI command 

until prior CRN(s) have been received. 

6. Task Management Functions shall not use a CRN and shall set the CRN field to 0. 

7. SCSI commands with a CRN value of 0 shall not participate in the ordering and verification procedures 

specified in 5. 

Any SCSI command, including such initialization commands as INQUIRY, TEST UNIT READY, and 

MODE SENSE/SELECT, may aways use a CRN of 0 if the state of the ECRN bit is not known or if execution 

ordering is not required for that command. 

87


88


Annex E 

(informative) 

Fibre Channel Mode Page 19h 

This annex describes and diagrams Fibre Channel Mode Page 19h 

Table 30 – Fibre Channel Control Page (19h). 

Bit 

7 6 5 4 3 2 1 0 

Byte 

0 PS Resvd Page Code (19h) 

1 Page Length (06h) 

2 Reserved 

3 Resvd PLPB DDIS DLM DSA ALWI DTIPE DTOLI 

4 Reserved ECRN 

5 Reserved 

6 Reserved 

7 Reserved 

When Disable Target Originated Loop Initialization (DTOLI) bit is one, the Target does not generate 

the Initializing LIP following insertion into the loop. The Target will respond to an Initializing LIP when 

it is received. The Target shall generate the Loop Failure LIP if it detects loop failure at its input and 

the Initializing LIP when the loop failure is corrected. When DTOLI bit is zero, the Target generates 

the Initializing LIP after it enables a port into a loop. 

When Disable Target Initiated Port Enable (DTIPE) bit is one, the Target waits for an Initiator to send 

the Loop Port Enable primitive before inserting itself into the loop. The Target uses the hard Address 

available in the SCA connector or device address jumpers to determine if primitives are addressed to 

it. A Loop Port Enable primitive with the broadcast address shall also cause the Target to insert itself 

into the loop. When DTIPE bit is zero, the Target enables its port into the loop without waiting for a 

Loop Port Enable primitive. 

Wen Allow Login Without Loop Initialization (ALWLI) bit is one, the Target shall use the hard address 

available in the SCA connector or device address jumpers and accept logins without verifying the address 

with loop initialization. When ALWLI bit is zero, the Target is required to verify its address 

through the loop initialization process before a login is accepted. 

When Disable Soft Address (DSA) bit is one, the Target does not select a soft address if there is a 

conflict for the hard address selection during loop initialization. In this case the Target enters the nonparticipating 

state. If the Target detects loop initialization while in the non-participating state, the Target 

will again attempt to get its hard address. When DSA bit is zero, the Target attempts to obtain a 

soft address during the loop initlization process. 

89


When Disable Loop Master (DLM) bit is one, the Target does not become loop master. The Target 

only repeats LISM frames it receives. This allows the Initiator to be loop master during loop initialization. 

When DLM bit is zero, the Target may become loop master during the loop initialization process. 

When Disable Discovery (DDIS) bit is one, the Target does not require receipt of Address or Port Discovery 

following loop initialization. The Target resumes processing of tasks on completion of loop initialization. 

When DDIS bit is zero, the Target must wait to receive an Address or Port Discovery 

before it resumes processing tasks for that Initiator. 

When the Prevent Loop Port Bypass (PLPB) bit is one, the Target shall ignore any Loop Port Bypass 

(LPB) and Loop Port Enable (LPE) primitive sequences. The loop port shall remain enabled. When 

PLPB bit is zero, the Target allows the Loop Port Bypass (LPB) and Loop Port Enable (PBE) primitive 

sequences to control the port bypass circuit. 

It shall be illegal to set DTIPE to one and PLPB to one. 

When an illegal bit combination is sent by the application client the device server shall return CHECK 

CONDITION status and the sense key shall be set to ILLEGAL REQUEST with the addtional sense 

code set to INVALID FIELD IN THE PARAMETER LIST. 

When the Enable Command Reference Number (ECRN) bit is one, the Initiator shall use a continuously 

increasing Command Reference Number for each Exchange with the Lun (except Task Management 

Functions). When the ECRN bit is zero, the Initiator shall not use a Command Reference 

Number for each Exchange with the Lun. 

90

A 

Abort Task Set 40 

ABTS 21, 46 

Accept 21 

Address 

Previously Acquired 4 

ADISC 53, 54 

AEN See Asynchronous Event Notification 

AL_TIME 31 

Asynchronous Event Notification 58, 62 

Auto Contingent Allegiance 40, 52 

Available BB_Credit 4 

B 

BA_ACC 21 

BA_RJT 21 

Basic Link Service Commands 

ABTS 21 

NOP 21 

RMC 21 

BB_Credit 4, 14, 15 

Alternate management 14, 15 

Available 4 

Login 4, 29 

BURST_LEN 36 

C 

Clear 40 

Clear ACA 40 

Clear Task Set 40 

Continue Sequence Condition 20 

D 

Data Compression 16, 17, 18, 19 

Data Overlay 34 

Data overlay 34 

Discover 21 

E 

E_D_TOV 14, 15, 31, 44, 47 

Exchange 23 

X_ID Invalidation 16, 17, 18, 19 

X_ID Reassignment 16, 17, 18, 19 

Extended Link Services 

Discover Address (ADISC) 21 

Discover N_Port Parameters (PDISC) 21 

FLOGI 21 

Logout (LOGO) 21 

N_Port Login (PLOGI) 21 

PRLI 22 

PRLO 22 

RLS 22 

I-i

RNC 22 

RRQ 22 

RSI 22 

RSS 22 

TPRLO 22 

F 

Fabric 21 

FCP 31, 33, 36, 42 

FCP_CMND 35 

FCP_CONF 39 

FCP_DL 36 

FCP_LUN 41 

FCP_RSP 37 

FCP_XFER_RDY 36 

Fibre 33 

Frame 

Abort Sequence 49 

Basic Accept 49 

Basic Accept (BA_ACC) 49 

Basic Reject (BA_RJT) 50 

Parameter Field 37 

Reinstate Recovery Qualifier (RRQ) 50 

H 

Hard Address 4 

Head of Queue Type 40 

Headers 

Optional 20 

I 

Information Units 

FCP_CMND 35 

FCP_RSP 37 

Initiator 

OPN 29 

L 

Lengthy 71 

LILP/LIRP 28 

Link_Control frame 20 

Linking 58, 62 

LIP 

AL_PD,AL_PS 29 

F7,AL_PS 29 

F7,F7 29 

F8,AL_PS 29 

F8,F7 29 

LIS_HOLD_TIME 31 

Loop 21 

Loop Failure 29 

Loop Initialization 

I-ii

LIP(F7,AL_PS) 29 

LIP(F7,F7) 29 

Loop_ID 4 

LUN assignment 41 

M 

Multicast/Selective Replicate 28 

Multiple Initiators 51 

N 

NACA 52 

Native 4 

Node Name 20 

non-participating mode 30 

NOP 21 

O 

Open Full Duplex 28 

Open Half Duplex 28 

OPN Initiator 29 

Ordered Queue Type 40 

P 

PDISC 53, 54 

Port Name 20 

Private NL_Port 4 

Process Associator 34 

Process Login 34 

Process_Associator 16, 17, 18, 19 

Public NL_Port 4 

Q 

Queue Types 

Head of Queue 40 

Ordered 40 

Simple 40 

R 

R 14, 15 

R_A_TOV 26, 31, 32, 47 

R_T_TOV 29 

Timers 

R_T_TOV 29 

Read 37 

Receive Data Field 

Buffer to Buffer 14, 15 

Class 3 16, 17, 18, 19 

Relative Offset 37 

by Information Category 14, 15 

Continously Increasing 14, 15 

Random 14, 15 

I-iii

Resets 


ABTS 56 


LIP(y,x) 56 

LOGO 56 

PLOGI 56 

Power Cycle 56 

PRLI, PRLO 56 

Selective Hard 29 

Target 56 

TPRLO 56 

RMC 21 

RR_TOV 31, 32, 53, 54, 55 

RSP_CODE 38 

S 

SEQ_ID 25 

Sequence 22 

Sequence Chaining 20 

Sequence Initiative 46 

Sequences 

Class 3 Concurrent 16, 17, 18, 19 

Concurrent 14, 15 

Open per Exchange 16, 17, 18, 19 

SFF-8045 4 

Simple Queue Type 40 

T 

Target authentication 54 

Target Reset 40 

Task Management 


Clear ACA 40 


Target Reset 40 

Terminate Task 40 

Terminate Task 40 

Timers 

AL_TIME 31 

E_D_TOV 14, 15, 31, 44, 47 

LIS_HOLD_TIME 31 

R_A_TOV 26, 31, 32, 47 

R_T_TOV 29 

RR_TOV 31, 32, 53, 54, 55 

ULP_TOV 31, 33, 45, 46, 47 

U 

ULP_TOV 31, 33, 45, 46, 47 

Untagged 40 

Upper 33 

I-iv

W 

Write 37 

X 

X_ID 

Invalidation 16, 17, 18, 19 

Reassignment 16, 17, 18, 19 

XFER_RDY Disabled 34 

I-v

FIBRE CHANNEL - T11

Create successful ePaper yourself

Delete template?

Save as template?