Fujitsu MBA3147RC, MBA3073RC, MBA3300RC, MBC2036RC, MBC2073RC User Manual

Download

C141-C013-01EN

MBA3073RC, MBA3147RC, MBA3300RC SERIES,

MBB2073RC, MBB2147RC SERIES,

MBC2036RC, MBC2073RC SERIES

DISK DRIVES

SERIAL ATTACHED SCSI INTERFACE

SPECIFICATIONS

FOR SAFE OPERATION

Handling of This Manual

This manual contains important information for using this product. Read thoroughly before using the product. Use this product only after thoroughly reading and understanding especially the section "Important Alert Items" in this manual. Keep this manual handy, and keep it carefully.

FUJITSU makes every effort to prevent users and bystanders from being injured or from suffering damage to their property. Use the product according to this manual.

This product is designed and manufactured for use in standard applications such as office work, personal devices and household appliances. This product is not intended for special uses (atomic controls, aeronautic or space systems, mass transport vehicle operating controls, medical devices for life support, or weapons firing controls) where particularly high reliability requirements exist, where the pertinent levels of safety are not guaranteed, or where a failure or operational error could threaten a life or cause a physical injury (hereafter referred to as "mission-critical" use). Customers considering the use of these products for mission-critical applications must have safety-assurance measures in place beforehand. Moreover, they are requested to consult our sales representative before embarking on such specialized use.

First Edition February 2007

The contents of this manual may be revised without prior notice.

The contents of this manual shall not be disclosed in any way or reproduced in any media without

the express written permission of Fujitsu Limited.

C141-C013

Revision History

(1/1)

Edition Date

01 2007.02.28 — —

Revised section (*1)

(Added/Deleted/Altered)

Details

*1 Section(s) with asterisk (*) refer to the previous edition when those were deleted.

C141-C013

This page is intentionally left blank.

Preface

This manual explains concerning the hard disk drives with internal Serial Attached SCSI (SAS) controller.

The purpose of this manual is to provide the specifications and functions of SAS for use of these magnetic disk drives incorporated into user systems, and to present the information necessary for creating host system software. This manual is written for users who have a basic knowledge of hard disk drives and their use in computer systems.

The composition of manuals related to these disk drives and the range of subjects covered in this manual are shown in "Manual Organization," provided on a subsequent page. Please use these other manuals along with this manual as necessary.

The organization of this manual, related reference manual and conventions for alert messages follow.

Overview of Manual

This manual consists of the following six chapters, glossary, abbreviation, and index:

Chapter 1 SAS Interface

This chapter describes the topology, physical and electrical requirements, interface protocol, and other operations of the interface.

Chapter 2 Command Processing

This chapter describes the basic logical specifications related to command processing.

Chapter 3 Data Buffer Management

This chapter describes the data buffer configuration, data transfer processing functions and cache operations.

Chapter 4 Command Specifications

This chapter describes detailed command specifications and how to use them.

Chapter 5 Parameter Data Format

This chapter describes the parameter data formats provided by the disk drives and how to use them.

C141-C013 1

Preface

Chapter 6 Sense Data and Error Recovery Methods

This chapter describes the configuration and contents of sense data which report to the host system when an error occurs, etc., key information necessary for error recovery, recommended procedures for error recovery to be executed through host system software and retry processing.

Chapter 7 Disk Media Management

This chapter describes the procedure for initializing the disk media, methods of treating media defects and data recovery methods.

Glossary

The glossary explains technical terms which are necessary to the reader's understanding when reading this manual.

Acronyms and Abbreviations

This list shows the full spelling of abbreviations used in this manual.

Index

2 C141-C013

Preface

CONVENTIONS USED IN THIS MANUAL

The model names of the disk drives covered by this manual differ depending on their device types and capacity (*1). In addition, these disk drives are called

Hard Disk Drive (HDD), "drive" or "device" in this manual.

Note: Model Name

M BA 3 147 RC

Interface type

Formatted capacity [One gigabyte (GB) = one billion bytes; accessible capacity will be less and actual capacity depends on the operating environment and formatting.]

Disk size 3: 3.5 inch 2: 2.5 inch

Type BA: 3.5-inch, 15,000rpm BB: 2.5-inch, 10,025rpm BC: 2.5-inch, 15,000rpm

Decimal numbers are represented normally. Hexadecimal numbers are represented as shown in the following examples:

X'17B9', 17B9h, 17B9H, or 17B9H. Binary number is represented as "010", 010b. An X is used to represent mode parameters that are ignored by the MODE

SELECT and MODE SELECT EXTENDED commands. An X is also used to represent mode parameters reported by the MODE SELECT and MODE SELECT EXTENDED commands and that vary depending on conditions at the time.

C141-C013 3

Preface

Conventions for Alert Messages

This manual uses the following conventions to show the alert messages. An alert message consists of an alert signal and alert statements. The alert signal consists of an alert symbol and a signal word or just a signal word.

The following are the alert signals and their meanings:

This indicates a hazardous situation likely to result in serious personal injury if the user does not perform the procedure correctly.

This indicates a hazardous situation could result in serious personal injury if the user does not perform the procedure correctly.

This indicates a hazardous situation could result in minor or moderate personal injury if the user does not perform the procedure correctly. This alert signal also indicates that damages to the product or other property, may occur if the user does not perform the product correctly.

Attention

This indicates information that could help the user use the product more efficiently.

In the text, the alert signal is centered, followed below by the indented message. A wider line space precedes and follows the alert message to show where the alert message begins and ends. The following is an example:

(Example)

It is possible to use bit 7 and bit 6 of the control byte as an inherent control field in future product specifications. It is recommended that the INIT specify zero in this field.

Please forward any comments you may have regarding this manual.

To make this manual easier for users to understand, opinions from readers are needed. Please write your opinions or requests on the Comment at the back of this manual and forward it to the address described in the sheet.

4 C141-C013

Preface

MANUAL ORGANIZATION

Product Maintenance

Manual

Interface Specifications

(This Manual)

1. General Description

2. Specifications

3. Data Format

4. Installation Requirements

5. Installation

6. Diagnostics and Maintenance

7. Error Analysis

1. Serial Attached SCSI (SAS) Interface

2. Command Processing

3. Data Buffer Management

4. Command Specifications

5. Parameter Data Formats

6. Sense Data and Error Recovery

7. Disk Media Management

C141-C013 5

Preface

REFERENCED STANDARDS

The product specifications and functions described in this manual conform to the following standards:

Specification

(document) number

T10/1562-D Revision 05

T10/1601-D Revision 10

T10/1236D Revision 20

T10/996D Revision 8C

T10/1157D Revision 24

T10/1561 Revision 14

Name

Serial Attached SCSI (SAS) American national

Serial Attached SCSI-1.1 (SAS-1.1) American national

Information technology SCSI Primary Commands-2 (SPC-2)

Information Technology SCSI-3 Block Commands (SBC)

Information technology SCSI-3 Architecture Model (SAM-2)

Information technology SCSI-3 Architecture Model (SAM-3)

Concerned

organization

Standards Institute (ANSI)

American national Standards Institute (ANSI)

6 C141-C013

CHAPTER 1 SAS Interface .............................................................................21

1.1 Topologies in SAS Interface....................................................................22

1.1.1 SAS Layering........................................................................................23

1.1.2 Physical links and phys .........................................................................24

1.1.3 Ports (narrow ports and wide ports) ......................................................24

1.1.4 SAS devices...........................................................................................26

1.1.5 Pathways................................................................................................27

1.1.6 Connections...........................................................................................28

1.2 Names and identifiers...............................................................................29

1.2.1 SAS addresses .......................................................................................29

1.2.2 Hashed SAS address..............................................................................30

1.3 Phy layer ..................................................................................................31

1.3.1 8b10b coding.........................................................................................31

1.3.2 Link reset sequence ...............................................................................32

1.3.3 Start conditions of the link reset sequence ............................................33

1.3.4 Out of band (OOB) signals....................................................................33

1.3.5 SAS OOB sequency ..............................................................................36

1.3.5.1 Exception handling in the OOB sequence.............................................38

1.3.6 SAS speed negotiation sequence...........................................................38

1.3.6.1 Phy reset sequence after devices are attached.......................................42

1.3.6.2 When the speed negotiation sequence is successful..............................43

1.3.6.3 Exception handling in the speed negotiation sequence.........................43

1.3.7 Phy layer dword synchronization (DWS)..............................................43

1.4 Link layer.................................................................................................44

1.4.1 Primitives...............................................................................................44

1.4.2 Primitive sequences...............................................................................47

1.4.3 Primitives not specific to type of connections.......................................47

1.4.3.1 AIP (Arbitration in progress).................................................................47

1.4.3.2 ALIGN...................................................................................................47

C141-C013 7

Contents

1.4.3.3 BREAK................................................................................................. 48

1.4.3.4 BROADCAST...................................................................................... 48

1.4.3.5 CLOSE.................................................................................................. 49

1.4.3.6 EOAF (End of address frame).............................................................. 49

1.4.3.7 ERROR................................................................................................. 49

1.4.3.8 HARD_RESET..................................................................................... 49

1.4.3.9 NOTIFY................................................................................................ 50

1.4.3.10 OPEN_ACCEPT................................................................................ 50

1.4.3.11 OPEN_REJECT................................................................................. 50

1.4.3.12 SOAF (Start of address frame)........................................................... 53

1.4.4 Primitives used only inside SSP and SMP connections ....................... 54

1.4.4.1 ACK (acknowledge)............................................................................. 54

1.4.4.2 CREDIT_BLOCKED........................................................................... 54

1.4.4.3 DONE................................................................................................... 54

1.4.4.4 EOF (End of frame).............................................................................. 56

1.4.4.5 NAK (negative acknowledgement) ...................................................... 56

1.4.4.6 RRDY................................................................................................... 56

1.4.4.7 SOF (Start of frame)............................................................................. 56

1.4.5 Clock skew management...................................................................... 57

1.4.6 Idle physical link .................................................................................. 57

1.4.7 Scrambling............................................................................................58

1.5 Address frames........................................................................................ 59

1.5.1 Address frames overview...................................................................... 59

1.5.2 IDENTIFY address frame..................................................................... 60

1.5.3 OPEN address frame............................................................................. 62

1.5.4 Identification and hard reset sequence.................................................. 65

1.5.5 Connections.......................................................................................... 66

1.5.5.1 Connections overview .......................................................................... 66

1.5.5.2 Connection request............................................................................... 66

1.5.5.3 Connection responses........................................................................... 67

1.5.5.4 Arbitration fairness............................................................................... 67

1.5.5.5 Aborting a connection request.............................................................. 69

1.5.5.6 Closing a connection ............................................................................ 70

1.5.5.7 Breaking a connection.......................................................................... 71

1.5.5.8 Rate matching....................................................................................... 71

8 C141-C013

Contents

1.5.6 SSP link layer........................................................................................72

1.5.6.1 SSP frame transmission and reception..................................................73

1.5.6.2 SSP flow control....................................................................................73

1.5.6.3 Interlocked frames.................................................................................73

1.5.6.4 Closing an SSP connection....................................................................76

1.6 Transport layer.........................................................................................78

1.6.1 SSP frame format ..................................................................................78

1.6.2 Information units...................................................................................81

1.6.2.1 COMMAND information unit...............................................................81

1.6.2.2 TASK information unit..........................................................................84

1.6.2.3 XFER_RDY information unit...............................................................86

1.6.2.4 DATA information unit.........................................................................89

1.6.2.5 RESPONSE information unit................................................................92

1.6.3 Sequences of SSP frames ......................................................................96

1.6.4 Exceptional event processing of a drive................................................98

CHAPTER 2 Command Processing.............................................................103

2.1 Command Format ..................................................................................103

2.2 Status Byte .............................................................................................108

2.3 Command Queuing Function.................................................................110

2.4 UNIT ATTENTION Condition .............................................................111

2.4.1 Generation of the UNIT ATTENTION condition...............................111

2.4.2 Response and release condition at UNIT ATTENTION condition

hol d s t ate...............................................................................................112

2.4.3 UNIT ATTENTION condition multiple hold .....................................113

2.5 Sense Data Hold State............................................................................113

2.6 Power Condition ....................................................................................114

2.7 LED Display ..........................................................................................116

2.8 Command Processing Exceptions..........................................................116

2.8.1 Overlapped tag ....................................................................................116

2.8.2 Illegal LUN specification....................................................................117

C141-C013 9

Contents

2.8.3 Reserved operation code.....................................................................118

2.8.4 Error recovery processing...................................................................118

2.8.5 Abort processing................................................................................. 119

2.8.6 Fatal hardware errors.......................................................................... 122

2.9 Data Block Addressing.......................................................................... 122

2.9.1 Definition of data space...................................................................... 122

2.9.2 Logical block addressing.................................................................... 125

CHAPTER 3 Data Buffer Management........................................................ 127

3.1 Data Buffer............................................................................................ 127

3.1.1 Data buffer configuration and basic operation ................................... 127

3.2 Look-Ahead Cache Feature................................................................... 130

3.2.1 Caching operation............................................................................... 130

3.2.2 Caching parameters ............................................................................ 133

3.2.3 Look-Ahead operation, Look-Ahead volume..................................... 133

CHAPTER 4 Command Specifications........................................................ 135

4.1 Control/Sense Commands..................................................................... 135

4.1.1 TEST UNIT READY (00)..................................................................135

4.1.2 INQUIRY (12).................................................................................... 136

4.1.3 READ CAPACITY (25).....................................................................151

4.1.4 MODE SELECT (15)......................................................................... 153

4.1.5 MODE SELECT EXTENDED (55)................................................... 163

4.1.6 MODE SENSE (1A)........................................................................... 166

4.1.7 MODE SENSE EXTENDED (5A)..................................................... 174

4.1.8 REZERO UNIT (01)........................................................................... 177

4.1.9 START/STOP UNIT (1B).................................................................. 178

4.1.10 RESERVE (16)................................................................................... 180

4.1.11 RESERVE EXTENDED (56)............................................................. 182

4.1.12 RELEASE (17)................................................................................... 183

4.1.13 RELEASE EXTENDED (57)............................................................. 184

4.1.14 REQUEST SENSE (03) ..................................................................... 185

4.1.15 LOG SELECT (4C)............................................................................ 187

10 C141-C013

Contents

4.1.16 LOG SENSE (4D)...............................................................................192

4.1.17 PERSISTENT RESERVE IN (5E)......................................................194

4.1.18 PERSISTENT RESERVE OUT (5F)..................................................201

4.1.19 REPORT LUNS (A0)..........................................................................207

4.1.20 REPORT DEVICE IDENTIFIER (A3)...............................................209

4.1.21 SET DEVICE IDENTIFIER (A4).......................................................211

4.2 Data Access Commands.........................................................................213

4.2.1 READ (08) ..........................................................................................213

4.2.2 READ EXTENDED (28)....................................................................215

4.2.3 WRITE (0A)........................................................................................216

4.2.4 WRITE EXTENDED (2A)..................................................................218

4.2.5 WRITE AND VERIFY (2E) ...............................................................219

4.2.6 VERIFY (2F).......................................................................................220

4.2.7 SEEK (0B)...........................................................................................221

4.2.8 SEEK EXTENDED (2B) ....................................................................222

4.2.9 SYNCHRONIZE CACHE (35)...........................................................223

4.3 Format Commands.................................................................................224

4.3.1 FORMAT UNIT (04)..........................................................................224

4.3.2 REASSIGN BLOCKS (07).................................................................235

4.3.3 READ DEFECT DATA (37) ..............................................................239

4.3.4 READ DEFECT DATA (B7)..............................................................244

4.4 Maintenance, Diagnostic Commands.....................................................246

4.4.1 SEND DIAGNOSTIC (1D).................................................................246

4.4.2 RECEIVE DIAGNOSTIC RESULTS (1C) ........................................258

4.4.3 WRITE BUFFER (3B)........................................................................263

4.4.4 READ BUFFER (3C)..........................................................................270

4.4.5 READ LONG (3E)..............................................................................275

4.4.6 WRITE LONG (3F) ............................................................................277

4.4.7 WRITE SAME (41).............................................................................279

CHAPTER 5 Parameter Data Format ...........................................................281

5.1 Mode Parameters....................................................................................281

5.1.1 Read/Write Error Recovery Parameters (Page Code = 1) ...................282

5.1.2 Disconnect/Reconnect Parameters (Page Code = 2) ...........................287

C141-C013 11

Contents

5.1.3 Format Parameters (Page Code = 3)................................................... 289

5.1.4 Drive Parameters (Page Code = 4) ..................................................... 294

5.1.5 Verify Error Recovery Parameters (Page Code = 7) .......................... 296

5.1.6 Caching Parameters (Page Code = 8) ................................................. 298

5.1.7 Control Mode Parameters (Page Code = 0A)..................................... 304

5.1.8 Notch Parameters (Page Code = 0C).................................................. 308

5.1.9 Port Control Parameter (Page Code = 19).......................................... 310

5.1.10 Power Condition Parameter (Page Code = 1A).................................. 316

5.1.11 Informational Exceptions Control Page (Page Code = 1C)................ 317

5.1.12 Background Control Mode Parameter (Page Code =

1C/subpage = 01)................................................................................322

5.1.13 Additional Error Recovery Parameters (Page Code = 21).................. 325

5.2 Log Parameters...................................................................................... 327

5.2.1 Support Log Page (X'00') ................................................................... 329

5.2.2 Buffer Overrun/Underrun Page (X'01') .............................................. 330

5.2.3 Write Error Count Page (X'02') .......................................................... 331

5.2.4 Read Error Count Page (X'03')........................................................... 335

5.2.5 Verify Error Count Page (X'05')......................................................... 338

5.2.6 Non-Medium Error Count Page (X'06').............................................. 342

5.2.7 Temperature Page (X'0D') .................................................................. 342

5.2.8 Start-Stop Cycle Counter Page (X'0E')............................................... 344

5.2.9 Application Client Page (X'0F')..........................................................347

5.2.10 Self-Test Result Page (X'10').............................................................. 348

5.2.11 Background Medium Scan Page (X'15')............................................. 350

5.2.12 Protocol Specific Port Log Page (X'18').............................................353

5.2.13 SMART Status Page (X'2F')...............................................................357

5.2.14 SMART Data Page (X'38')................................................................. 357

CHAPTER 6 Sense Data and Error Recovery Methods............................. 359

6.1 Sense Data............................................................................................ 359

6.1.1 Sense data format................................................................................ 359

6.1.2 Sense data basic information .............................................................. 361

6.1.3 Sense data additional information ...................................................... 371

6.2 INIT Error Recovery Methods (Recommended).................................. 372

12 C141-C013

Contents

6.2.1 Termination status analysis and error recovery methods ....................372

6.2.2 Sense data analysis and error recovery methods.................................374

6.2.3 Error logging .......................................................................................382

6.3 Disk Drive Error Recovery Processing.................................................382

6.3.1 Error states and retry processing procedures.......................................382

6.3.2 Auto alternate block allocation processing..........................................384

6.3.3 Error recovery processing control.......................................................386

CHAPTER 7 Disk Media Management.........................................................389

7.1 Defect Management...............................................................................389

7.2 Disk Media Initialization ....................................................................... 392

7.2.1 Initialization during installation ..........................................................392

7.2.2 Re-initialization...................................................................................393

7.3 Alternate Block Allocation Processing.................................................. 395

7.4 Background Media Scan (BMS)............................................................396

7.4.1 Overview.............................................................................................396

7.4.2 Background Media Scan......................................................................396

7.4.2.1 Background Media Scan (BMS) mode................................................396

7.4.2.2 Pre-Scan mode.....................................................................................396

7.4.2.3 Write command operation during a Pre-Scan......................................397

7.4.2.4 Conditions for operation......................................................................397

7.4.3 Mode Page...........................................................................................398

7.4.4 Background Media Scan Log Page .....................................................400

7.5 Drive Self-Test (DST)...........................................................................404

7.5.1 Overview.............................................................................................404

7.5.2 Self-test modes ....................................................................................405

7.5.2.1 Foreground mode ................................................................................405

7.5.2.2 Background mode................................................................................406

7.5.3 Matters that are common to both modes .............................................407

7.5.4 Short self-test.......................................................................................407

7.5.5 Extended self-test ................................................................................407

7.5.6 Test segments......................................................................................407

C141-C013 13

Contents

7.6 SMART (Self-Monitoring Analysis and Reporting Technology)......... 410

7.6.1 Overview ............................................................................................ 410

7.6.2 Data analysis....................................................................................... 411

7.6.3 Failure prediction method................................................................... 412

7.6.4 Reporting function.............................................................................. 416

Glossary ............................................................................................................ 417

Acronyms and Abbreviations.......................................................................... 419

Index ................................................................................................................. 421

14 C141-C013

Contents

Illustrations

Figures

Figure 1.1 SAS drive connection patterns............................................................. 22

Figure 1.2 SAS control layers................................................................................ 23

Figure 1.3 Physical links and phys........................................................................ 24

Figure 1.4 Ports (narrow ports and wide ports)..................................................... 25

Figure 1.5 SAS devices ......................................................................................... 26

Figure 1.6 Example of potential pathways............................................................ 27

Figure 1.7 Reset-related terminology.................................................................... 32

Figure 1.8 OOB signal transmission...................................................................... 34

Figure 1.9 OOB signal detection........................................................................... 36

Figure 1.10 SAS to SAS OOB sequence................................................................. 37

Figure 1.11 SAS speed negotiation window............................................................ 39

Figure 1.12 SAS speed negotiation sequence (Example 1)..................................... 41

Figure 1.13 SAS speed negotiation sequence (Example 2)..................................... 41

Figure 1.14 Phy reset sequence (Example).............................................................. 42

Figure 1.15 Connection request timeout example................................................... 69

Figure 1.16 Closing a connection example.............................................................. 70

Figure 1.17 Interlocked frames................................................................................75

Figure 1.18 Non-interlocked frames with the same tag........................................... 76

Figure 1.19 Non-interlocked frames with different tags.......................................... 76

Figure 1.20 Closing an SSP connection example.................................................... 77

Figure 1.21 Example of XFER_RDY frames.......................................................... 88

Figure 1.22 Example of TASK frame...................................................................... 96

Figure 1.23 Example of write command ................................................................. 97

Figure 1.24 Example of read command................................................................... 97

Figure 1.25 Example of the processing sequence for an exceptional event ............ 99

Figure 2.1 Data space configuration.................................................................... 124

Figure 3.1 Data buffer configuration (in the case of 8 cache segments) ............. 128

Figure 4.1 MODE SELECT parameter structure................................................. 155

Figure 4.2 Correction of the defect descriptor.....................................................238

Figure 6.1 Analysis of the termination status...................................................... 372

Tables

Table 1.1 SAS address format..............................................................................29

Table 1.2 Hashed SAS address code parameter................................................... 30

Table 1.3 Usage of special characters .................................................................. 31

Table 1.4 OOB signal timing specifications ........................................................ 33

C141-C013 15

Contents

Table 1.5 OOB signal transmitter requirements...................................................34

Table 1.6 OOB signal receiver burst time detection requirements....................... 35

Table 1.7 OOB signal receiver idle time detection requirements......................... 35

Table 1.8 OOB signal receiver negation time detection requirements.................35

Table 1.9 SAS speed negotiation sequence timing specifications........................ 39

Table 1.10 Primitives not specific to type of connection .......................................44

Table 1.11 Primitives used only inside SSP and SMP connections .......................46

Table 1.12 Primitive sequences..............................................................................47

Table 1.13 OPEN_REJECT abandon primitives....................................................51

Table 1.14 OPEN_REJECT retry primitives..........................................................52

Table 1.15 DONE primitives..................................................................................55

Table 1.16 Clock skew management ALIGN insertion requirement .....................57

Table 1.17 Scrambling for different data dword types...........................................58

Table 1.18 Address frame format........................................................................... 59

Table 1.19 IDENTIFY address frame format.........................................................60

Table 1.20 DEVICE TYPE field............................................................................ 60

Table 1.21 OPEN address frame format.................................................................62

Table 1.22 PROTOCOL field................................................................................. 62

Table 1.23 CONNECTION RATE field ................................................................ 63

Table 1.24 ARBITRATION WAIT TIME field..................................................... 64

Table 1.25 Connection responses........................................................................... 67

Table 1.26 Arbitration priority for OPEN address frames passing on a

physical link ......................................................................................... 68

Table 1.27 Abort connection responses.................................................................. 69

Table 1.28 Close connection responses..................................................................70

Table 1.29 Break connection responses ................................................................. 71

Table 1.30 Rate matching ALIGN and/or NOTIFY insertion requirements..........72

Table 1.31 SSP frame interlock requirements........................................................ 74

Table 1.32 SSP frame format ................................................................................. 78

Table 1.33 FRAME TYPE field.............................................................................79

Table 1.34 COMMAND information unit .............................................................. 81

Table 1.35 TASK ATTRIBUTE field....................................................................82

Table 1.36 TASK information unit......................................................................... 84

Table 1.37 TASK MANAGEMENT FUNCTION field ........................................ 85

Table 1.38 XFER_RDY information unit .............................................................. 86

Table 1.39 An example of requested offset............................................................88

Table 1.40 DATA information unit........................................................................89

Table 1.41 RESPONSE information unit...............................................................92

Table 1.42 DATAPRES field................................................................................. 92

Table 1.43 RESPONSE DATA field......................................................................94

Table 1.44 RESPONSE CODE field......................................................................94

Table 1.45 Exceptional event processing of a drive.............................................100

Table 2.1 6-Byte CDB basic format................................................................... 104

Table 2.2 10-Byte CDB basic format................................................................. 104

Table 2.3 12-Byte CDB basic format................................................................. 105

16 C141-C013

Contents

Table 2.4 Operation code................................................................................... 105

Table 2.5 Control byte........................................................................................ 107

Table 2.6 Status.................................................................................................. 108

Table 2.7 LED display ....................................................................................... 116

Table 2.8 Outline of disk drive error recovery processing................................. 118

Table 2.9 Comparison between SAS and SCSI about definition....................... 119

Table 2.10 Reset processing during write ............................................................ 121

Table 4.1 Standard INQUIRY data.................................................................... 138

Table 4.2 VERSION field.................................................................................. 139

Table 4.3 Command queuing............................................................................. 141

Table 4.4 Version descriptor.............................................................................. 142

Table 4.5 Command support data....................................................................... 143

Table 4.6 Support............................................................................................... 144

Table 4.7 VPD information................................................................................ 145

Table 4.8 VPD information: VPD identifier list ............................................... 145

Table 4.9 VPD information: device serial No................................................... 146

Table 4.10 VPD information: device unique information....................................147

Table 4.11 READ CAPACITY data .................................................................... 152

Table 4.12 MODE SELECT command (Group 0) parameter configuration ....... 157

Table 4.13 MODE SELECT parameters.............................................................. 161

Table 4.14 MODE SELECT EXTENDED command (group 2) parameter

configuration...................................................................................... 164

Table 4.15 Mode page.......................................................................................... 167

Table 4.16 MODE SENSE data type specifications ............................................ 168

Table 4.17 MODE SENSE command (group 0) parameter configuration...........170

Table 4.18 MODE SENSE EXTENDED command (group 2) parameter

configuration...................................................................................... 175

Table 4.19 PC (page control) ............................................................................... 188

Table 4.20 LOG SELECT command parameter configuration............................ 188

Table 4.21 Page code ........................................................................................... 189

Table 4.22 Log parameter .................................................................................... 190

Table 4.23 "Page Code" assignment for the log pages......................................... 193

Table 4.24 PERSISTENT RESERVE IN service actions.................................... 195

Table 4.25 PERSISTENT RESERVE IN parameter data for READ KEYS ....... 196

Table 4.26 PERSISTENT RESERVE IN parameter data for READ

RESERVATIONS.............................................................................. 197

Table 4.27 Format of reservation descriptors.......................................................198

Table 4.28 Persistent reservations scope.............................................................. 199

Table 4.29 Persistent reservations type codes...................................................... 200

Table 4.30 PERSISTENT RESERVE OUT service action codes........................ 203

Table 4.31 PERSISTENT RESERVE OUT parameter list.................................. 204

Table 4.32 PERSISTENT RESERVE OUT service action and valid

parameters.......................................................................................... 206

Table 4.33 REPORT LUNS parameter data......................................................... 208

Table 4.34 REPORT DEVICE IDENTIFIER parameter data ............................. 210

Table 4.35 SET DEVICE IDENTIFIER parameter data...................................... 212

C141-C013 17

Contents

Table 4.36 Defect list format................................................................................226

Table 4.37 FORMAT UNIT command parameter list configuration...................227

Table 4.38 Defect descriptor: byte distance from index format ..........................230

Table 4.39 Defect descriptor: physical sector address format............................. 231

Table 4.40 FORMAT UNIT command defect processing ................................... 233

Table 4.41 REASSIGN BLOCK command: defect data list configuration.........236

Table 4.42 Defect data type..................................................................................239

Table 4.43 Defect data format.............................................................................. 240

Table 4.44 READ DEFECT DATA command: defect data configuration ......... 240

Table 4.45 Defect data conditions........................................................................ 242

Table 4.46 READ DEFECT DATA command (B7): defect data

configuration ...................................................................................... 245

Table 4.47 Self-diagnosis test............................................................................... 247

Table 4.48 Error recovery control flags during the self-diagnosis test................. 248

Table 4.49 SEND DIAGNOSTIC command: parameter list configuration........250

Table 4.50 Page code............................................................................................250

Table 4.51 SEND DIAGNOSTIC parameters: page code list.............................251

Table 4.52 SEND DIAGNOSTIC parameters: PHY Test function..................... 252

Table 4.53 SEND DIAGNOSTIC parameters: logical/physical address

conversion .......................................................................................... 255

Table 4.54 Specifying address format.................................................................. 255

Table 4.55 SELF-TEST........................................................................................257

Table 4.56 RECEIVE DIAGNOSTIC RESULTS command: response data

configuration ...................................................................................... 259

Table 4.57 RECEIVE DIAGNOSTIC RESULTS response data: page

code list...............................................................................................260

Table 4.58 RECEIVE DIAGNOSTIC RESULTS response data:

logical/physical address conversion ...................................................261

Table 4.59 Address format ...................................................................................262

Table 4.60 WRITE BUFFER transfer mode ........................................................264

Table 4.61 WRITE BUFFER command: buffer data (mode = 000, 001) ........... 265

Table 4.62 READ BUFFER transfer mode.......................................................... 270

Table 4.63 READ BUFFER command: buffer data (mode = 0000, 0001)......... 271

Table 4.64 READ BUFFER command: buffer descriptor.................................... 273

Table 4.65 READ BUFFER command: echo buffer descriptor.......................... 274

Table 5.1 MODE SELECT parameters: read/write error recovery

parameters...........................................................................................282

Table 5.2 Combinations of error recovery flags.................................................286

Table 5.3 MODE SELECT parameters: disconnect/reconnect parameters........287

Table 5.4 MODE SELECT parameters: format parameters..............................289

Table 5.5 Details of parameters on MODE SELECT parameter: pages 3,

4, and C...............................................................................................293

Table 5.6 MODE SELECT parameters: drive parameters ................................294

Table 5.7 MODE SELECT parameters: verify error recovery parameters .......296

Table 5.8 MODE SELECT parameters: caching parameters............................ 298

Table 5.9 MODE SELECT parameters: control mode parameters....................304

18 C141-C013

Contents

Table 5.10 MODE SELECT parameters: notch parameters................................ 308

Table 5.11 Port control parameter: Page 0 Format (Short Page Format)............. 311

Table 5.12 Port control parameter: Sub Page Format (Long Format).................. 312

Table 5.13 SAS phy mode descriptor format....................................................... 313

Table 5.14 Power condition parameter: Page 0 Format (Short Page

Format)............................................................................................... 316

Table 5.15 MODE SELECT parameters: informational exception control

page.................................................................................................... 317

Table 5.16 MRIE.................................................................................................. 320

Table 5.17 Interval timer...................................................................................... 321

Table 5.18 MODE SELECT parameters: background control mode

parameter............................................................................................322

Table 5.19 MODE SELECT parameters: additional error recovery

parameters.......................................................................................... 325

Table 5.20 Mode parameter default values .......................................................... 326

Table 5.21 Log parameter format.........................................................................327

Table 5.22 Support log page (X'00') .................................................................... 329

Table 5.23 Buffer overrun/underrun page (X'01')................................................ 330

Table 5.24 Write error count page (X'02')............................................................ 331

Table 5.25 Write errors recovered without delays (page 02, code 0000)............ 332

Table 5.26 Write errors recovered with possible delays (page 02, code

0001) .................................................................................................. 332

Table 5.27 Total write errors posted (page 02, code 0002)..................................333

Table 5.28 Total recoverable write errors posted to INIT (page 02, code

0003) .................................................................................................. 333

Table 5.29 Total write bytes processed (page 02, code 0005) ............................. 334

Table 5.30 Total unrecoverable write errors posted to INIT (page 02, code

0006) .................................................................................................. 334

Table 5.31 Read error count page (X'03') ............................................................ 335

Table 5.32 Read errors recovered without delays (page 03, code 0000) ............. 335

Table 5.33 Read errors recovered with possible delays (page 03, code

0001) .................................................................................................. 336

Table 5.34 Total read errors posted (page 03, code 0002)................................... 336

Table 5.35 Total recoverable read errors posted to INIT (page 03, code

0003) .................................................................................................. 337

Table 5.36 Total read bytes processed (page 03, code 0005)............................... 337

Table 5.37 Total unrecoverable read errors posted to INIT (page 03, code

0006) .................................................................................................. 338

Table 5.38 Verify error count page (X'05') .......................................................... 338

Table 5.39 Verify errors recovered without delays (page 05, code 0000) ........... 339

Table 5.40 Verify errors recovered with possible delays (page 05, code

0001) .................................................................................................. 339

Table 5.41 Total verify errors posted (page 05, code 0002) ................................ 340

Table 5.42 Total recoverable verify errors posted to INIT (page 05, code

0003) .................................................................................................. 340

Table 5.43 Total verify bytes processed (page 05, code 0005)............................ 341

C141-C013 19

Contents

Table 5.44 Total unrecoverable verify errors posted to INIT (page 05,

code 0006).......................................................................................... 341

Table 5.45 Non-medium error count page (X'06')................................................ 342

Table 5.46 Temperature page (X'0D') ..................................................................342

Table 5.47 Temperature (page 0D, code 0000).................................................... 343

Table 5.48 Reference temperature (page 0D, code 0001) ....................................343

Table 5.49 Start-stop cycle counter page (X'0E') .................................................344

Table 5.50 Date of manufacture (page 0E, code 0001)........................................ 344

Table 5.51 Accounting date (page 0E, code 0002)............................................... 345

Table 5.52 Specified cycle count over device lifetime (page 0E, code

0003)................................................................................................... 345

Table 5.53 Start-stop cycle counter (page 0E, code 0004)...................................346

Table 5.54 Application client page (X'0F')........................................................... 347

Table 5.55 General usage application client parameter data (page 0F, code

0000-003F).........................................................................................347

Table 5.56 Self-test result page (X'10') ................................................................ 348

Table 5.57 Self-test result parameter data (page 10, code 0001-0014) ................348

Table 5.58 Self-test results values........................................................................ 349

Table 5.59 Background medium scan page (X'15')..............................................350

Table 5.60 Background medium scan status parameter ....................................... 350

Table 5.61 BMS status ......................................................................................... 351

Table 5.62 Background medium scan parameter ................................................. 352

Table 5.63 Reassign status ...................................................................................352

Table 5.64 Protocol specific port log page format ............................................... 353

Table 5.65 Log parameter format.........................................................................354

Table 5.66 SMART status page (X'2F') ............................................................... 357

Table 5.67 SMART data page (X'38').................................................................. 357

Table 6.1 Sense data format ...............................................................................360

Table 6.2 Sense key inherent information..........................................................363

Table 6.3 Sense key............................................................................................364

Table 6.4 Additional Sense Code and Additional Sense Code Qualifier ...........365

Table 6.5 Sense data error classification ............................................................375

Table 6.6 Error recovery processing procedures................................................378

Table 6.7 Disk drive errors and number of retries.............................................. 387

Table 7.1 Mode Page 0x1C SubPage 0x01 (Background Control Mode

Page)...................................................................................................398

Table 7.2 Background Media Scan Log Page format......................................... 400

Table 7.3 BMS Status parameter format............................................................401

Table 7.4 Medium Scan parameter format......................................................... 402

Table 7.5 SMART ASC/ASCQ.......................................................................... 412

Table 7.6 SMART thresholds.............................................................................415

20 C141-C013

CHAPTER 1 SAS Interface

1.1 Topologies in SAS Interface

1.2 Names and identifiers

1.3 Phy layer

1.4 Link layer

1.5 Address frames

1.6 Transport layer

This chapter describes the topology, interface protocol, and operation of the SAS interface.

C141-C013 21

SAS Interface

1.1 Topologies in SAS Interface

As shown in Figure 1.1, SAS drives have two connection patterns: point-topoint connection and expander connection.

Figure 1.1 SAS drive connection patterns

SAS supports the three protocols listed below. Among these protocols, SAS drives support only SSP.

• Serial SCSI Protocol (SSP)

• Serial ATA Tunneled Protocol (STP)

• Serial Management Protocol (SMP)

22 C141-C013

1.1 Topologies in SAS Interface

1.1.1 SAS Layering

As shown in Figure 1.2, for SAS, the following six control layers are defined:

• Physical layer: Electric properties related to cables, connectors,

and signals

• Phy (transceiver) layer: 8B/10B code, OOB, and speed negotiation

• Link layer: Primitives, address frames, and connection control

• Port layer: Wide port control

• Transport layer: Frame control

• Application layer: SCSI commands, mode pages, and log pages

Application layer

Transport layer

Port layer

Link layer

Phy layer

Physical layer

Figure 1.2 SAS control layers

C141-C013 23

SAS Interface

1.1.2 Physical links and phys

A physical link is a set of four wires used as two differential signal pairs. One differential signal transmits in one direction while the other differential signal transmits in the opposite direction. Data may be transmitted in both directions simultaneously.

A physical phy contains a transceiver which electrically interfaces to a physical link, which attaches to another physical phy.

Phys are contained in ports. Phys interface to the service delivery subsystem. Figure 1.3 shows two phys attached with a physical link.

An attached phy is the phy to which a phy is attached over a physical link. A device may contain one or more phys. Each phy has a phy identifier which is unique within the device.

Figure 1.3 Physical links and phys

1.1.3 Ports (narrow ports and wide ports)

A port contains one or more phys. Ports in a device are associated with physical phys based on the identification sequence. A port is created from a set of physical phys if one or more physical phys contained within a device:

a) transmit the same SAS address during the identification sequence; and b) receive the same SAS address during the identification sequence (i.e., the

corresponding attached phy or phys transmit the same SAS address).

24 C141-C013

1.1 Topologies in SAS Interface

A wide port is created if there is more than one phy in the port. A narrow port is a port with only one phy.

A wide link is the set of physical links that attach a wide port to another wide port. A narrow link is the physical link that attaches a narrow port to another narrow port.

Figure 1.4 shows examples of narrow ports and wide ports, with a representation of the SAS address transmitted during the identification sequence. Although several phys on the left transmit SAS addresses of B, only phys attached to the same SAS addresses become part of the same ports. The set of phys with SAS address B attached to the set of phys with SAS address D become one port, while the set of phys with SAS address B attached to the set of phys with SAS address E become another port.

Figure 1.4 Ports (narrow ports and wide ports)

C141-C013 25

SAS Interface

1.1.4 SAS devices

A SAS device contains one or more SAS ports, each containing one or more phys (i.e., a SAS port may be a narrow port or a wide port).

Each single HDD unit is a separate SAS device. Usually, it is a single port or dual port device, but does not use a wide port.

Figure 1.5 shows examples of SAS devices with different port and phy configurations.

Figure 1.5 SAS devices

26 C141-C013

1.1 Topologies in SAS Interface

1.1.5 Pathways

A potential pathway is a set of physical links between a SAS INIT phy and a SAS TARG phy. When a SAS INIT phy is directly attached to a SAS TARG phy, there is one potential pathway. When there are expander devices between a SAS INIT phy and a SAS TARG phy, there are multiple potential pathways, each consisting of a set of physical links between the SAS INIT phy and the SAS TARG phy. The physical links may or may not be using the same physical link rate.

A pathway is a set of physical links between a SAS INIT phy and a SAS TARG phy being used by a connection.

Figure 1.6 shows examples of potential pathways.

Figure 1.6 Example of potential pathways

A partial pathway is the set of physical links participating in a connection request that has not reached the destination phy (e.g., the OPEN address frame has been transmitted by the source phy but the OPEN address frame has not yet reached the destination phy).

A partial pathway is blocked when path resources it requires are held by another partial pathway.

C141-C013 27

SAS Interface

1.1.6 Connections

A connection is a temporary association between a SAS INIT port and a SAS TARG port. During a connection all dwords from the SAS INIT port are forwarded to the SAS TARG port, and all dwords from the SAS TARG port are forwarded to the SAS INIT port.

A connection is pending when an OPEN address frame has been delivered along a completed pathway to the destination phy but the destination phy has not yet responded to the connection request. A connection is established when an OPEN_ACCEPT is returned to the source phy.

A connection enables communication for one protocol: SSP, STP, or SMP. For SSP and STP, connections may be opened and closed multiple times during the processing of a command.

The connection rate is the effective rate of dwords through the pathway between a SAS INIT phy and a SAS TARG phy, established through the connection request. Every phy shall support a 1,5 Gbps connection rate regardless of its physical link rate.

One connection may be active on a physical link at a time. If the connection is an SSP or SMP connection and there are no dwords to transmit associated with that connection, idle dwords are transmitted.

The number of connections established by a SAS port shall not exceed the number of SAS phys within the SAS port (i.e., only one connection per SAS phy is allowed). There shall be a separate connection on each physical link.

If multiple potential pathways exist between the SAS INIT port(s) and the SAS TARG port(s), multiple connections may be established by a SAS port between the following:

a) one SAS INIT port to multiple SAS TARG ports; b) one SAS TARG port to multiple SAS INIT ports; or c) one SAS INIT port to one SAS TARG port. Once a connection is established, the pathway used for that connection shall not

be changed (i.e., all the physical links that make up the pathway remain dedicated to the connection until it is closed).

28 C141-C013

1.2 Names and identifiers

Device names are worldwide unique names for devices within a transport protocol. Port names are worldwide unique names for ports within a transport protocol. Port identifiers are the values by which ports are identified within a domain, and are used as SAS addresses. Phy identifiers are unique within a device.

1.2.1 SAS addresses

Table 1.1 defines the SAS address format. SAS addresses shall be compatible with the NAA (Name Address Authority) IEEE Registered format identification descriptor defined in SPC-3. The SAS address shall be worldwide unique. A SAS address of 00000000_00000000h indicates an invalid SAS address.

Table 1.1 SAS address format

Bit

Byte

0 NAA (5h) (MSB) 1 2 3 (LSB) (MSB) 4 5 6 7 (LSB)

7 6 5 4 3 2 1 0

IEEE COMPANY ID

SAS addresses represent any types of WWNs: a device SAS address representing a device address, a port SAS address representing Port-A, or a port SAS address representing Port-B.

− device SAS address: 5CCC CCCX XXXX XXX yy00

− port SAS address (Port-A): 5CCC CCCX XXXX XXX yy10

− port SAS address (Port-B): 5CCC CCCX XXXX XXX yy11

C = 4 bits: Company ID X = 4 bits: Serial number for WWN y = 1 bit: Reserved

C141-C013 29

SAS Interface

1.2.2 Hashed SAS address

SSP frames include a hashed version of the SAS address to provide an additional level of verification of proper frame routing.

The code used for the hashing algorithm is a cyclic binary Bose, Chaudhuri, and Hocquenghem (BCH) (63, 39, 9) codes. Table 1.2 lists the parameters for the code.

Table 1.2 Hashed SAS address code parameter

Parameter Value

Number of bits per codeword 63 Number of data bits 39 Number of redundant bits 24 Minimum distance of the code 9

The generator polynomial for this code is:

G(x) = (x 6 + x + 1) (x 6 + x 4 + x 2 + x + 1) (x 6 + x 5 + x 2 + x + 1) (x 6 + x 3 + 1)

After multiplication of the factors, the generator polynomial is:

G(x) = x 24 + x 23 + x 22 + x 20 + x 19 + x 17 + x 16 + x 13 + x 10 + x 9 + x 8 + x 6 + x 5 + x 4 + x 2 + x + 1

30 C141-C013

1.3 Phy layer

In the phy layer, 8b10b coding and the link reset sequence are defined.

1.3.1 8b10b coding

All information transferred in SAS is encoded into 10-bit characters using 8b10b encoding. Information includes data bytes representing data in a frame and control characters used for frame delimiters.

Ten-bit characters are transferred in units of dwords, where a dword contains four ten-bit characters.

A primitive is a dword that consists of a control character at the beginning and the other three data characters.

In the SAS specifications, a primitive that starts with the K28.5 control character and another one that starts with the K28.6 control character are defined. Table 1.3 shows the usage of special characters.

Table 1.3 Usage of special characters

Beginning character Usage in SAS Usage in SATA

K28.3

K28.5

K28.6

Dxx.y Data Data

A single data dword is a dword that starts with a data character.

Running disparity shall be maintained separately on each physical link. During a connection, expander devices shall convert incoming 10-bit characters to 8-bit bytes and generate the 10-bit character with correct disparity for the output physical link. Phys within a device may or may not begin operation with the same disparity after the reset sequence.

Information to be transmitted across a physical link shall be encoded eight bits at a time into a 10-bit character and then transmitted serially bit-by-bit across the physical link. Information received over the physical link shall be collected ten bits at a time, and those characters that are used for data, called data characters, shall be decoded into the correct 8-bit data bytes. The 10-bit characters support all 256 8-bit combinations. 8b10b coding ensures that sufficient transitions are present in the serial bit stream to make clock recovery possible at the receiver. Such encoding also greatly increases the likelihood of detecting any single or

Primitive to be used only within STP connection

ALIGN and most primitives that are defined in SAS

SATA_ERROR (used on the physical link layer of SATA)

All the primitives except ALIGN

ALIGN

Not used.

C141-C013 31

SAS Interface

multiple bit errors that may occur during transmission and reception of information. In addition, some of the control characters of the transmission code contain a distinct and easily recognizable bit pattern called a comma pattern which assists a receiver in achieving character and dword alignment on the incoming bit stream.

1.3.2 Link reset sequence

The link reset sequence consists of the phy reset sequence and the identification sequence. The phy reset sequence consists of the out of band (OOB) sequence and speed negotiation sequence. The hard reset sequence is performed only when the HARD_RESET primitive is detected after the completion of the speed negotiation. The HARD_RESET primitive detected at any time later than the identification sequence is invalid. The hard reset sequence corresponds to the reset condition of pSCSI.

Figure 1.7 Reset-related terminology

32 C141-C013

1.3 Phy layer

1.3.3 Start conditions of the link reset sequence

Drives start the link reset sequence when they detect any of the following conditions:

• The power is turned on.

• A loss of signal is detected (OOB from the INIT).

• A loss of sync is detected.

• A HARD_RESET primitive is received during the link reset sequence.

• A hot-plug timeout (500 ms) is detected during the link reset sequence.

• The IDENTIFY address frame could not be received.

1.3.4 Out of band (OOB) signals

Out of band (OOB) signals are low-speed signal patterns detected by the phy that do not appear in normal data streams. They consist of defined amounts of idle time followed by defined amounts of burst time. During the idle time, D.C. idle is transmitted. During the burst time, ALIGN (0) primitives are transmitted repeatedly. The signals are differentiated by the length of idle time between the burst times.

Table 1.4 OOB signal timing specifications

Parameter Minimum Nominal Maximum Comments

OOB interval (OOBI) *1

666,600 ps 666,666 ps 666,733 ps

The time basis for burst times and idle times used to create OOB signals. Based on 1,5 Gbps clock tolerance.

The minimum time a receiver shall COMSAS detect timeout

13,65 ìs

− −

allow to detect COMSAS after

transmitting COMSAS.

Derived from: OOBI × 512 × 40

*1 The OOBI is different from the UI (OOB) defined in SATA (for example,

stricter clock tolerance applies to SAS). This is because the OOBI is a fixed value that is equal to the UI value of G1, and that does not depend on actual transfer speed used to set up the burst time.

Table 1.5 describes the OOB signal transmitter requirements for the burst time, idle time, and negation times that comprise each OOB signal.

C141-C013 33

SAS Interface

COMINIT

idle

COMRESET/COMINIT

COMSAS

2 3 4 5 6

1 2 3 4 5 6

Table 1.5 OOB signal transmitter requirements

Signal Burst time Idle time Negation time

COMINIT/RESET 160 OOBI 480 OOBI 800 OOBI COMSAS 160 OOBI 1440 OOBI 2400 OOBI

To transmit an OOB signal, a transmitter shall repeat these steps six times:

1) transmit D.C. idle for an idle time; and

2) transmit an ALIGN burst for a burst time. It shall then transmit D.C. idle for an OOB signal negation time.

The ALIGNs used in OOB signals should be at generation 1 (G1) physical link rates (i.e., 1,5 Gbps). The ALIGNs are only required to generate an envelope for the detection circuitry, as required for any signaling that may be A.C. coupled. If G2 ALIGNs are used, the number of ALIGNs doubles compared with G1 ALIGNs.

A SAS transmitter should transmit ALIGNs at the G1 physical link rate to create the burst portion of the OOB signal, but may transmit ALIGNs at its lowest supported physical link rate if it is not able to transmit at the G1 physical link rate and shall not transmit them at a physical link rate faster than its lowest supported physical link rate.

Figure 1.8 describes OOB signal transmission by the SP transmitter.

ALIGN burst

COMINIT Transmitt

COMINIT

negation

Transmitted

COMSAS

COMSAS idle

negation

Figure 1.8 OOB signal transmission

34 C141-C013

1.3 Phy layer

Table 1.6 describes the OOB signal receiver requirements for detecting burst times, assuming Tburst is the length of the detected burst time. The burst time is not used to distinguish between signals.

Table 1.6 OOB signal receiver burst time detection requirements

Signal

COMINIT/COMRESET COMSAS

Table 1.7 describes the OOB signal receiver requirements for detecting idle times, assuming Tidle is the length of the detected idle time.

Table 1.7 OOB signal receiver idle time detection requirements

Signal

COMINIT/COMRESET COMSAS

Table 1.8 describes the OOB signal receiver requirements for detecting negation times, assuming Tidle is the length of the detected idle time.

Detection requirements

may detect shall detect

Tburst ≤ 100 ns Tburst ≤ 100 ns

Tburst > 100 ns Tburst > 100 ns

Detection requirements

may detect shall detect

175 ns ≤ Tidle < 525 ns 304 ns ≤ Tidle < 336 ns

525 ns ≤ Tidle < 1575 ns 911,7 ns ≤ Tidle < 1008 ns

Table 1.8 OOB signal receiver negation time detection requirements

Signal Detection requirements COMINIT/COMRESET Tidle > 525 ns COMSAS Tidle > 1575 ns

A receiver shall detect an OOB signal after receiving four consecutive idle time/burst time pairs. It is not an error to receive more than four idle time/burst time pairs. A receiver shall not detect the same OOB signal again until it has detected the corresponding negation time (i.e., a COMINIT negation time for a COMINIT) or has detected a different OOB signal.

A SAS receiver shall detect OOB signals comprised of ALIGNs transmitted at any rate up to its highest supported physical link rate. This includes physical link rates below its lowest supported physical link rate (e.g., a SAS receiver supporting only 3,0 Gbps detects 1,5 Gbps based ALIGNs, providing interoperability with a SAS transmitter supporting both 1,5 Gbps and 3,0 Gbps).

C141-C013 35

SAS Interface

1 2 3 4 n

COMRESET/COMINIT

COMSAS

idle

ALIGN burst

idle/Burst pair

COMSAS

Figure 1.9 describes SAS OOB signal detection by the SP receiver.

Any transitions

COMINIT detected

nth idle/Burst pair

COMINIT negation

negation

COMSAS detected

Figure 1.9 OOB signal detection

1.3.5 SAS OOB sequency

During the OOB sequence, the INIT and the drive first send the COMINIT signal to each other, then the COMSAS signal. When they both have received the COMSAS signal from each other, the OOB sequence is recognized as having been performed successfully. Depending on the power-on timing, reception of the COMINIT signal from the other side may not be detected. In such a case, however, the OOB sequence is recognized as having been performed successfully if the COMINIT and COMSAS signals have been sent by the time the COMSAS signal from the other side is received.

36 C141-C013

When the OOB sequence is successful, the SAS speed negotiation sequence starts.

Figure 1.10 shows several different SAS OOB sequences between phy A and phy B, with phy A starting the SAS OOB sequence at the same time as phy B, before phy B, and before phy B powers on.

1.3 Phy layer

Phy B Rx

Phy B Tx

COMINIT

COMSAS

Phy A Tx/

Phy B Rx

Phy A Rx/

Phy B Tx

Time=0

Time=z

COMINIT

COMSAS

Scenario 1: Both SAS phys start SAS OOB sequence at same time

COMINIT

Phy A Tx/ Phy B Rx

Phy A Rx/ Phy B Tx

Time=0 Time=z

COMINIT COMSAS

Scenario 2: SAS phy A starts SAS OOB sequence

COMINIT

Phy A Tx/

Time=0 Time=z

COMSAS

Phy A Rx/

COMINIT

Scenario 3: SAS phy B misses SAS phy A's COMINIT

A: SAS phy A power on B: SAS phy B power on Time 0 : SAS phy reset sequence begins Time z : SAS speed negotiation sequence begins

COMSAS

Figure 1.10 SAS to SAS OOB sequence

C141-C013 37

SAS Interface

1.3.5.1 Exception handling in the OOB sequence

• When the COMINIT signal cannot be detected

When a drive sends the COMINIT signal but the COMINIT signal from the other side cannot be detected, the drive waits for the reception of the COMINIT signal from the INIT or expander until the hot-plug timeout time (500 ms) elapses. If the hot-plug timeout time elapses before the COMINIT signal from the other side has been received, the drive sends the COMINIT signal again to start a new link reset sequence. If the drive receives the COMSAS signal while waiting for the COMINIT signal, it sends the COMSAS signal and takes the OOB sequence as having been performed successfully in spite of the fact that it has not received the COMINIT signal.

• When the COMSAS signal cannot be detected

When a drive sends the COMSAS signal but the COMSAS signal from the other side cannot be detected, the drive waits for the reception of the COMSAS signal from the INIT or expander until the hot-plug timeout time (500 ms) elapses. If the hot-plug timeout time elapses before the COMSAS signal from the other side has been received, the drive sends the COMSAS signal again to start a new link reset sequence.

1.3.6 SAS speed negotiation sequence

The SAS speed negotiation sequence is a peer-to-peer negotiation technique that does not assume initiator and target (i.e., host and device) roles. The sequence consists of a set of speed negotiation windows for each physical link rate, starting with 1,5 Gbps, then 3,0 Gbps, then the next rate. The length of the speed negotiation sequence is determined by the number of physical link rates supported by the phys.

Figure 1.11 defines the speed negotiation window, including:

a) speed negotiation window time; b) rate change delay time (RCDT); c) speed negotiation transmit time (SNTT); and d) speed negotiation lock time (SNLT).

38 C141-C013

1.3 Phy layer

ALIGN(0)s

ALIGN(1)s

Speed negotiation lock time(SNLT)

Speed negotiation transmit time(SNTT)

Rate change delay time

(RCDT)

Speed negotiation window

Figure 1.11 SAS speed negotiation window

Table 1.9 defines the timing specifications for the SAS speed negotiation sequence.

Table 1.9 SAS speed negotiation sequence timing specifications

Parameter Time Comments

Rate change delay time (RCDT)

750 000 OOBI

The time the transmitter shall transmit D.C. idle between rates during speed negotiation.

The time during which ALIGN(0) or Speed negotiation transmit time (SNTT)

163 840 OOBI

ALIGN(1) is transmitted at each physical link

rate during the speed negotiation sequence.

Derived from: OOBI x 4096 x 40.

The maximum time during the speed Speed negotiation lock time (SNLT)

153 600 OOBI

negotiation window for a transmitter to reply

with ALIGN(1). Derived from:

OOBI x 3840 x 40 Speed negotiation

window time

913 840 OOBI

The duration of a speed negotiation window.

Derived from: RCDT + SNTT.

The speed negotiation sequence shall consist of the following transmission sequence:

1) a transmission of D.C. idle for an RCDT; and

C141-C013 39

2) if the phy supports the physical link rate, a transmission of ALIGNs at that physical link rate for the remainder of the entire speed negotiation window. If the phy does not support the physical link rate, transmission of D.C. idle for the remainder of the entire speed negotiation window.

SAS Interface

If the phy supports the physical link rate, it shall attempt to synchronize on an incoming series of dwords at that rate for the SNLT. The received dwords may be ALIGN (0) or ALIGN (1) primitives. If the phy achieves dword synchronization within the SNLT, it shall change from transmitting ALIGN (0) primitives to transmitting ALIGN (1) primitives for the remainder of the SNTT (i.e., the remainder of the speed negotiation window). If the phy does not achieve dword synchronization within the SNLT, it shall continue transmitting ALIGN (0)s for the remainder of the SNTT (i.e., the remainder of the speed negotiation window).

At the end of the SNTT, if a phy is both transmitting and receiving ALIGN (1) primitives, it shall consider that physical link rate valid. The phy shall then proceed to the next speed negotiation window. A phy shall participate in all speed negotiation windows:

a) up to its highest supported physical link rate plus one; or b) until it runs a speed negotiation window that does not detect a valid physical

link rate after having detected a valid physical link rate in a previous speed negotiation window.

If the phy has detected a valid physical link rate in the previous speed negotiation window, it shall enter the final speed negotiation window using the highest previously successful link rate.

Figure 1.12 shows speed negotiation between a phy A that supports G1 through G3 link rates and a phy B that only supports the G2 link rate. Both phys run:

1) the G1 speed negotiation window, supported by phy A but not by phy B;

2) the G2 speed negotiation window, supported by both phys; and

3) the G3 speed negotiation window, supported by phy A but not by phy B. Both phys then select G2 for the final speed negotiation window to establish the

negotiated physical link rate.

40 C141-C013

1.3 Phy layer

RCD

SNTT

RCD

no dword

sync

dword

sync

dword

sync

SNTT

dword

sync

dword

sync

Phy B Tx

Phy B Rx

no dword

sync

by phy B

G1 rate

G2 rate

G3 rate

G2 rate

(negotiated rate)

ALIGN(0)s

ALIGN(1)s

Time

RCD

Hot Plug

Timeout delay

Phy A Tx/

Phy B Rx

Phy A Rx/

Phy B Tx SNTT

SNTT

G2 rate

COMINIT

Phy A Tx/

Phy A Rx/

Not supported

Time z: SAS speed negotiation sequence begins

Not supported

Figure 1.12 SAS speed negotiation sequence (Example 1)

If the phy does not obtain dword synchronization during the final speed negotiation window, the SAS speed negotiation sequence fails.

Figure 1.13 shows the same speed negotiation sequence as in Figure 1.12 when phy B does not obtain dword synchronization during the final speed negotiation window. If this occurs, the handshake is not complete and the OOB sequence shall be retried starting with COMINIT, forcing the phy to retry the whole reset sequence.

RCD

dword

sync

RCD

dword

sync

RCD

no dword

sync

dword

sync

dword

sync

G2 rate

ALIGN(1)s

no dword

RCD

G3 rate

Time z: SAS speed negotiation sequence begins

sync

RCD

(negotiated rate)

Time

RCD

G1 rate

ALIGN(0)s

dword

sync

RCD

Figure 1.13 SAS speed negotiation sequence (Example 2)

C141-C013 41

SAS Interface

COMINIT

COMSAS

Timeout

Incomplete

COMINIT

Timeout

1.3.6.1 Phy reset sequence after devices are attached

Since SAS signal cable connector does not include power lines, it is not possible to detect the physical insertion of the signal cable connector onto a plug. Noncabled environments may similarly not have a way to detect physical insertion of a device. As a result, every time a phy reset sequence is originated:

Figure 1.14 shows how two phys complete the phy reset sequence if the phys are not attached at power on. In this example, phy A and phy B are attached some time before phy B's second hot-plug timeout occurs. Phy B's OOB detection circuitry detects a COMINIT after the attachment, and therefore phy B transmits COMSAS, since it has both transmitted and received a COMINIT. Upon receiving COMSAS, phy A transmits its own COMSAS. The SAS speed negotiation sequence follows.

Phy A Tx

Phy A Rx

Phy B Tx

Phy B Rx

Time=0

Hot-Plug

Time=y

Hot-Plug

Time=y

Hot-Plug

COMINIT

COMSAS

Time=z

COMSAS

A: SAS phy A power on B: SAS phy B power on Time y : SAS phy A attached to SAS phy B Time z: SAS phy A and SAS phy B start the SAS speed negotiation sequence

Figure 1.14 Phy reset sequence (Example)

42 C141-C013

1.3 Phy layer

1.3.6.2 When the speed negotiation sequence is successful

When the speed negotiation sequence is successful, the identification sequence (sending and receiving the IDENTIFY frame) is performed subsequently. However, the INIT may have started the hard reset sequence at the time when the drive starts the identification sequence. In this case, the drive will start the link reset sequence, because it receives the HARD_RESET.

1.3.6.3 Exception handling in the speed negotiation sequence

• When synchronization cannot be established at the G1 or G2 rate

• When synchronization cannot be established at the last negotiated rate

In both of the cases referred to above, the speed negotiation sequence is considered to have failed. The drive waits for the reception of the COMINIT signal from the INIT or expander until the hot-plug timeout time (500 ms) elapses. When the hot-plug timeout time elapses before the signal has been received from the other side, the drive sends the COMINIT signal again to start a new link reset sequence. At the same time, the value in the PHY RESET PROBLEM field of the Protocol-Specific log page (page 18) is incremented when synchronization cannot be established at the last negotiated rate.

1.3.7 Phy layer dword synchronization (DWS)

The dword synchronization monitoring performed by the drive after the completion of the speed negotiation sequence has the following characteristics:

1) When four invalid dwords are detected in a frame (address frame or SSP frame), the dword synchronization is considered to have been lost. At this time, the drive starts the link reset sequence without monitoring whether any valid primitive is received (and without trying to establish resynchronization).

2) If an invalid dword is detected in the idle state (the state where random data is being sent and/or received without transfer of frames or primitives), the DWS Reset timeout timer (1 ms) is started. If at least one valid primitive is detected within 1 ms, the monitoring assumes that the drive is in the SynAcquired state (the state that the dword synchronization is established). Otherwise, the drive starts the link reset sequence.

C141-C013 43

SAS Interface

1.4 Link layer

The link layer defines primitives, address frames, and connections.

1.4.1 Primitives

Table 1.10 defines the primitives not specific to the type of connection.

Table 1.10 Primitives not specific to type of connection (1/2)

From To Sequence

Primitive USE (*4)

I E T I E T Type

AIP (NORMAL) NoConn - E - I E T single supported supported AIP (RESERVED 0) NoConn - - - I E T single

AIP (RESERVED 1) NoConn - - - I E T single

AIP (RESERVED 2) NoConn - - - I E T single

AIP (WAITING ON CONNECTION) NoConn - E - I E T single supported supported AIP (WAITING ON DEVICE) NoConn - E - I E T single supported supported AIP (WAITING ON PARTIAL) NoConn - E - I E T single supported supported AIP (RESERVED WAITING ON PARTIAL) NoConn - - - I E T single

ALIGN (0) ALL I E T I E T single supported supported ALIGN (1) ALL I E T I E T single supported supported ALIGN (2) ALL I E T I E T single supported supported ALIGN (3) ALL I E T I E T single supported supported BREAK ALL I E T I E T redundant supported supported

SAS 1.1

rev. 10

supported prim

↑ ↑ ↑

↑

FJ supported

prim

Supported (*1)

BROADCAST (CHANGE) NoConn I E - I E T redundant supported ignored BROADCAST (RESERVED CHANGE 0) NoConn - - - I E T redundant BROADCAST (RESERVED CHANGE 1) NoConn - - - I E T redundant BROADCAST (SES) NoConn - - T I E T redundant supported ignored BROADCAST (RESERVED 1) NoConn - - - I E T redundant reserved ignored BROADCAST (RESERVED 2) NoConn - - - I E T redundant reserved ignored BROADCAST (RESERVED 3) NoConn - - - I E T redundant reserved ignored BROADCAST (RESERVED 4) NoConn - - - I E T redundant reserved ignored CLOSE (NORMAL) Conn I - T I - T triple supported supported CLOSE (RESERVED 0) Conn - - - I - T triple CLOSE (RESERVED 1) Conn - - - I - T triple CLOSE (CLEAR AFFILIATION) STP I - - - - T triple (Not STP)

↑ ↑

↑ ↑ ↑ ↑

ignored ignored

↑

44 C141-C013

1.4 Link layer

Table 1.10 Primitives not specific to type of connection (2/2)

From To Sequence

Primitive USE (*4)

I E T I E T Type

EOAF NoConn I E T I E T single supported supported ERROR ALL - E - I E T single supported supported HARD RESET NoConn I E - I E T redundant supported supported NOTIFY (ENABLE SPINUP) ALL I E - I E T single supported supported NOTIFY (RESERVED 0) ALL - - - I E T single reserved ignored NOTIFY (RESERVED 1) ALL - - - I E T single reserved ignored NOTIFY (RESERVED 2) ALL - - - I E T single reserved ignored OPEN_ACCEPT NoConn I - T I - T single supported supported OPEN_REJECT (RETRY) NoConn I - T I - T single supported supported OPEN_REJECT (RESERVED CONTINUE 0) NoConn - - - I - T single OPEN_REJECT (RESERVED CONTINUE 1) NoConn - - - I - T single OPEN_REJECT (BAD DESTINATION) NoConn - E - I - T single supported supported OPEN_REJECT (CONNECTION RATE NOT

SUPPORTED) OPEN_REJECT (NO DESTINATION) NoConn - E - I - T single supported supported

NoConn I E T I - T single supported supported

SAS 1.1

rev. 10

supported prim

↑ ↑ ↑ ↑

FJ supported

prim

OPEN_REJECT (RESERVED INITIALIZE 0) NoConn - - - I - T single OPEN_REJECT (RESERVED INITIALIZE 1) NoConn - - - I - T single OPEN_REJECT (PATHWAY BLOCKED) NoConn - E - I - T single supported supported OPEN_REJECT (RESERVED STOP 0) NoConn - - - I - T single OPEN_REJECT (RESERVED STOP 1) NoConn - - - I - T single OPEN_REJECT (PROTOCOL NOT

SUPPORTED) OPEN_REJECT (STP RESOURCES BUSY) NoConn - E T I - - single supported supported (*2) OPEN_REJECT (WRONG DESTINATON) NoConn I - T I - T single supported supported OPNE_REJECT (RESERVED ABANDON 0) NoConn - - - I - T single OPNE_REJECT (RESERVED ABANDON 1) NoConn - - - I - T single OPNE_REJECT (RESERVED ABANDON 2) NoConn - - - I - T single OPNE_REJECT (RESERVED ABANDON 3) NoConn - - - I - T single SOAF NoConn I E T I E T single supported supported

NoConn I - T I - T single supported supported

↑ ↑ ↑ ↑

↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑

Table 1.11 defines the primitives used only inside SSP and SMP connections.

C141-C013 45

SAS Interface

Table 1.11 Primitives used only inside SSP and SMP connections

From To Sequence

Primitive USE (*4)

I E T I E T Type

ACK Conn I - T I - T single supported supported CREDIT_BLOCKED Conn I - T I - T single supported supported DONE (NORMAL) Conn I - T I - T single supported supported DONE (RESERVED 0)

DONE (RESERVED 1)

DONE (ACK/NAK TIMEOUT) Conn I - T I - T single supported supported DONE (RESERVED TIMEOUT 0)

DONE (RESERVED TIMEOUT 1)

DONE (CREDIT TIMEOUT) Conn I - T I - T single supported supported EOF Conn I - T I - T single supported supported NAK (CRC ERROR) Conn I - T I - T single supported supported NAK (RESERVED 0) Conn - - - I - T single NAK (RESERVED 1) Conn - - - I - T single NAK (RESERVED 2) Conn - - - I - T single RRDY (NORMAL) Conn I - T I - T single supported supported

Conn - - - I - T single

SAS 1.1

rev. 10

supported prim

↑

↑ ↑ ↑ ↑ ↑ ↑

FJ supported

prim

supported (*3)

RRDY (RESERVED 0) Conn - - - I - T single RRDY (RESERVED 1) Conn - - - I - T single SOF Conn I - T I - T single supported supported

↑ ↑ ↑ ↑

(*1) These primitives are received as AIP primitives.

(*2) When this primitive is received, the connection request is canceled and the

command is aborted.

(*3) These primitives are received as DONE primitives.

(*4) When a primitive falling under "NoConn" is received during the inside

connection, this primitive is discarded. When a primitive falling under "NoConn" is received during the outside connection, this primitive is also discarded. As for the HARD_RESET primitive, when it is received at any time except during the period between the completion of the speed negotiation and the reception of the first valid address frame, the received HARD_RESET primitive is discarded as well.

46 C141-C013

1.4 Link layer

1.4.2 Primitive sequences

Table 1.12 summarizes the types of primitive sequences.

Table 1.12 Primitive sequences

Primitive

sequence type

Number of times the transmitter

transmits the primitive to transmit

the primitive sequence

Number of times the receiver

receives the primitive to detect the

Single 1 1 Triple 3 3

Redundant 6 3

Any number of ALIGNs and NOTIFYs may be sent inside primitive sequences without affecting the count or breaking the consecutiveness requirements. Rate matching ALIGNs and NOTIFYs shall be sent inside primitive sequences inside of connections if rate matching is enabled.

1.4.3 Primitives not specific to type of connections

1.4.3.1 AIP (Arbitration in progress)

AIP is sent by an expander device after a connection request to specify that the connection request is being processed and specify the status of the connection request.

A drive performs the following processing when an AIP primitive is received:

primitive sequence

1) When the drive receives an AIP primitive after sending the OPEN frame, the

2) When the drive receives an AIP primitive before sending the OPEN frame,

3) In a case of OPEN cross (a case where both the drive and initiator send the

1.4.3.2 ALIGN

ALIGNs are used for: a) OOB signals;

b) character and dword alignment during the speed negotiation sequence; c) clock skew management after the phy reset sequence; d) rate matching during connections; and

drive initializes the open timer held in the drive, and restarts it.

the received AIP primitive is discarded.

OPEN frame), when the drive receives an AIP primitive after sending the OPEN frame and, then, receives the OPEN frame from the expander, the drive does not perform the arbitration fairness procedure. Instead, the attempt from the drive is treated as unsuccessful, and the drive sends the OPEN_ACCEPT or an OPEN_REJECT primitive.

C141-C013 47

SAS Interface

Phys shall use ALIGN (0) to construct OOB signals as described in 6.6. Phys shall use ALIGN (0) and ALIGN (1) during the speed negotiation sequence as described in 6.7.4.2. Phys shall rotate through ALIGN (0), ALIGN (1), ALIGN (2), and ALIGN (3) for all ALIGNs sent after the phy reset sequence.

1.4.3.3 BREAK

The BREAK primitive is used to cancel the connection request, or to cancel the connection. The BREAK primitive is sent in cases 1) to 4) shown below.

1) When the connection request is to be canceled after sending an OPEN frame

2) When neither OPEN_ACCEPT nor OPEN_REJECT primitive is received within 1 ms of sending an OPEN frame

3) When neither a DONE primitive nor an SSP frame is received within 1 ms of sending a DONE primitive

4) When a CLOSE primitive is not received within 1 ms of sending a CLOSE primitive

When the BREAK primitive is sent in any of the cases above, the BREAK timer (1 ms) is started in anticipation of the reception of the BREAK primitive within 1 ms. If the BREAK primitive is not received within 1 ms, the connection is regarded as having been closed.

After the BREAK primitive is sent, when any primitives and/or frames other than the BREAK primitive are received, all of these (except NOTIFY primitives) are discarded.

When the BREAK primitive is not received within 1 ms of sending the BREAK primitive, a drive may start the link reset sequence.

1.4.3.4 BROADCAST

BROADCASTs are used to notify all SAS ports in a domain of an event.

When an expander port receives a BROADCAST it shall transmit the same BROADCAST on at least one phy in all other expander ports. BROADCAST shall only be sent outside of connections after the phy reset sequence has completed.

BROADCAST (CHANGE) is sent by an expander device to notify SAS INIT ports and other expander devices that a configuration change has occurred. BROADCAST (CHANGE) may also be transmitted by SAS INIT ports. BROADCAST (CHANGE) shall be ignored by SAS TARG ports.

BROADCAST (SES) is sent by a SAS TARG port to notify SAS INIT ports that an asynchronous event has occurred in an enclosure, and SSP INIT ports should poll all the SCSI enclosure services logical units in the SAS domain. BROADCAST (SES) shall be ignored by SAS TARG ports.

The drive ignores all received BROADCAST primitives.

48 C141-C013

1.4 Link layer

1.4.3.5 CLOSE

A CLOSE primitive is used to close a connection. When a CLOSE primitive is sent, the close timer (1 ms) is started. If a CLOSE primitive is not received within 1 ms of sending CLOSE primitive, the BREAK primitive is sent. The recipient must send a CLOSE primitive within 1 ms of receiving a CLOSE primitive.

Drives support only the CLOSE (NORMAL) primitive as a CLOSE primitive, while the CLOSE (CLEAR AFFILIATION), CLOSE (RESERVED 0), and CLOSE (RESERVED 1) primitives are ignored.

1.4.3.6 EOAF (End of address frame)

EOAF specifies the end of an address frame.

1.4.3.7 ERROR

ERROR should be sent by an expander device when it is forwarding dwords from a SAS physical link or SATA physical link to a SAS physical link and it receives an invalid dword or an ERROR.

When a drive receives an ERROR primitive, the drive discards it. When a received address frame contains the ERROR primitive, the whole address frame is discarded even if CRC is correct.

When a drive receives an SSP frame containing the ERROR primitive, the drive discards the whole SSP frame even if CRC is correct, and sends the NAK primitive.

1.4.3.8 HARD_RESET

HARD_RESET is used to force a phy to generate a hard reset to its port. This primitive is only valid after the phy reset sequence without an intervening identification sequence and shall be ignored at other times.

Upon receipt of the HARD_RESET primitive, a drive clears all the commands it is currently holding.

In a dual port drive, the HARD_RESET primitive received at one port (e.g., at port A) affects the other port (e.g., at port B).

Meanwhile, as can be expected, a link reset sequence that occurs at one port (e.g., port A) does not affect the other port (e.g., port B).

C141-C013 49

SAS Interface

1.4.3.9 NOTIFY

The NOTIFY (ENABLE SPINUP) primitive is sent by a SAS INIT device or expander device. When receiving the NOTIFY (ENABLE SPINUP) primitive in the active wait state of SCSI power conditions, a drive enters the active state.

A NOTIFY primitive is sometime used as a substitute for an ALIGN primitive, and it is sent by the initiator or expander at regular intervals.

A SAS TARG device with multiple SAS TARG ports treats the NOTIFY (ENABLE SPINUP) primitives from all the SAS TARG ports equally (for example, the NOTIFY (ENABLE SPINUP) primitive received through SAS TARG port A is treated as the wakeup for the START STOP UNIT command received through SAS TARG port B).

The NOTIFY (RESERVED 0), NOTIFY (RESERVED 1), and NOTIFY (RESERVED 2) primitives are ignored by all devices.

1.4.3.10 OPEN_ACCEPT

The OPEN_ACCEPT primitive indicates the acceptance of a connection request. The recipient of an OPEN frame must send the OPEN_ACCEPT primitive within 1 ms, when it can accept the connection request.

1.4.3.11 OPEN_REJECT

An OPEN_REJECT primitive indicates the rejection of a connection request. Among the OPEN_REJECT primitives, there are those indicating that the connection request cannot be continued (OPEN_REJECT abandon primitives) and those indicating that the connection request can be retried (OPEN_REJECT retry primitives).

The OPEN_REJECT primitives defined in Table 1.13 direct the recipients to abandon the connection request.

50 C141-C013

1.4 Link layer

Table 1.13 OPEN_REJECT abandon primitives

Primitive Issuer

OPEN_REJECT (BAD DESTINATION)

OPEN_REJECT (CONNECTION RATE NOT SUPPORTED)

OPEN_REJECT (PROTOCOL NOT SUPPORTED)

OPEN_REJECT (RESERVED ABANDON 0)

Processing by the drive that receives the OPEN_REJECT

primitive

Expander phy

Any phy 1) When the physical link rate is 3.0 Gbps and the connection

Destination phy

Unknown

Cancels the connection request, and aborts the TASK (command).

link rate is 3.0 Gbps, the drive changes the connection rate to

1.5 Gbps, and resends the OPEN frame.

2) When the physical link rate is 3.0 Gbps and the connection link rate is 1.5 Gbps, the drive changes the connection rate to

3.0 Gbps, and resends the OPEN frame.

3) When the physical link rate is 1.5 Gbps and the connection link rate is 1.5 Gbps, the drive cancels connection request, and aborts the TASK (command).

* After sending an OPEN frame in case 1) or 2), if the drive

receives this OPEN_REJECT primitive again, the drive cancels the connection request, and aborts the TASK (command).

If the driver receives this OPEN_REJECT primitive, it cancels the connection request, and aborts the TASK (command).

Performs the same processing as the OPEN_REJECT (WRONG DESTINATION).

OPEN_REJECT (RESERVED ABANDON 0)

OPEN_REJECT (STP RESOURCES BUSY)

OPEN_REJECT (WRONG DESTINATION)

All of the OPEN_REJECT versions defined in Table 1.14 shall result in the originating device retrying the connection request.

Unknown

Destination phy

Performs the same processing as the OPEN_REJECT (WRONG DESTINATION).

If the driver receives this OPEN_REJECT primitive, it cancels the connection request, and aborts the TASK (command).

C141-C013 51

SAS Interface

Table 1.14 OPEN_REJECT retry primitives

Primitive Issuer

OPEN_REJECT (NO DESTINATION)

OPEN_REJECT (PATHWAY BLOCKED)

OPEN_REJECT (RESERVED CONTINUE 0)

OPEN_REJECT (RESERVED CONTINUE 1)

Expander phy

Unknown

Processing by the drive that receives the OPEN_REJECT

primitive

1) If the I_Tnexus loss timer is working, the drive leaves the timer working. If the I_Tnexus loss timer has expired, the drive aborts all the commands of the relevant initiator. Otherwise, if the timer has not expired, the drive resends the OPEN frame.

2) If the I_Tnexus loss timer is stopped and it has expired, the drive aborts all the commands of the relevant initiator. Otherwise, if the timer has not expired, the drive initializes the timer and resends the OPEN frame.

Performs the same processing as the OPEN_REJECT (RETRY).

OPEN_REJECT (RESERVED INITIALIZE 0)

OPEN_REJECT (RESERVED INITIALIZE 1)

OPEN_REJECT (RESERVED STOP 0)

OPEN_REJECT (RESERVED STOP 1)

OPEN_REJECT (RETRY)

When an INIT sends an OPEN frame to a drive but the drive cannot accept the connection request, the drive sends an OPEN_REJECT within 1ms. A drive sends an OPEN_REJECT primitive in any of the following cases:

1) When the INIT port bit in a received OPEN frame is 0, the drive sends the OPEN_REJECT (PROTOCOL NOT SUPPORTED) primitive.

2) When the PROTOCOL field in a received OPEN frame is not SSP, the drive sends the OPEN_REJECT (PROTOCOL NOT SUPPORTED) primitive.

Unknown

Destination phy

Performs the same processing as the OPEN_REJECT (NO DESTINATION).

Performs the same processing as the OPEN_REJECT (PATHWAY BLOCKED).

Otherwise, if the I_Tnexus loss timer is working, the drive stops the timer and resends the OPEN frame.

3) When the CONNECTION RATE field in the received OPEN frame is not "1.5G or 3.0G", the drive sends the OPEN_REJECT (CONNECTION RATE NOT SUPPORTED) primitive.

52 C141-C013

1.4 Link layer

4) When, in contradiction to the fact that the physical link rate is 1.5 Gbps, the CONNECTION RATE field in a received OPEN frame is 3.0G, the drive sends the OPEN_REJECT (CONNECTION RATE NOT SUPPORTED) primitive.

5) A dual port drive may process requests from up to four INITs at a time for each port. When a dual port drive holds commands from four INITs for one of its ports, and if it receives a connection request from a fifth INIT through the same port, the drive sends the OPEN_REJECT (RETRY) primitive.

6) When the drive cannot receive an SSP frame temporarily due to the recovery processing of an internal drive error, the drive sends the OPEN_REJECT (RETRY) primitive.

7) When the SAS address of a drive does not match the DESTINATION SAS ADDRESS field in a received OPEN frame, the drive sends the OPEN_REJECT (WRONG DESTINATION) primitive.

1.4.3.12 SOAF (Start of address frame)

SOAF specifies the start of an address frame.

C141-C013 53

SAS Interface

1.4.4 Primitives used only inside SSP and SMP connections

1.4.4.1 ACK (acknowledge)

The ACK primitive indicates that an SSP frame is received normally. When a drive receives an SSP frame, and the CRC field in the received SSP frame is correct, the drive sends the ACK primitive within 1 ms.

1.4.4.2 CREDIT_BLOCKED

The CREDIT_BLOCKED primitive indicates that no more credits can be sent during this connection. When a drive determines that it can give out no more credits during the connection, the drive sends the CREDIT_BLOCKED primitive to avoid credit timeout. A drive sends the CREDIT_BLOCKED primitive in any of the following cases:

1) When the drive sends an OPEN frame and, then, receives the OPEN_ACCEPT primitive from the initiator during the abort processing in the drive, the drive sends the CREDIT_BLOCKED primitive instead of an RRDY primitive.

2) When the drive determines during the connection that it has entered a

After sending the CREDIT_BLOCKED primitive, the drive must wait until the connection is closed before it can send an RRDY primitive.

Meanwhile, when a drive receives the CREDIT_BLOCKED primitive from the initiator, the drive performs any of the following processing:

1) At the time of the receipt of the CREDIT_BLOCKED primitive, if the drive

2) At the time of the receipt of the CREDIT_BLOCKED primitive, if the drive holds

3) At the time of the receipt of the CREDIT_BLOCKED primitive, if the drive

1.4.4.3 DONE

condition where abort processing inside the drive is necessary, if an SSP frame is received from the initiator, the drive sends the CREDIT_BLOCKED primitive instead of an RRDY primitive.

holds credits that have been given before and has frames to send, the drive can send as many frames as there are credits being held at this time.

credits that have been given before but has no frames to send, the drive sends the DONE (CREDIT TIMEOUT) primitive and starts the close processing.

holds no credits, it sends the DONE (CREDIT TIMEOUT) primitive and starts the close processing.

A DONE primitive indicates the start of the close processing of an SSP connection. In the close processing, in principle, when no frame for processing remains on the side of the OPEN frame sender (recipient of the OPEN_ACCEPT primitive), a DONE primitive is sent. However, there is no rule concerning which side should first start the close processing.

Table 1.15 contains DONE primitive definitions.

54 C141-C013

1.4 Link layer

Table 1.15 DONE primitives

primitive Description

DONE (ACK/NAK TIMEOUT)

DONE (RESERVED TIMEOUT 0)

DONE (RESERVED TIMEOUT 1)

When the sender of an SSP frame does neither receive the ACK nor the NAK primitive after sending the SSP frame, the sender sends the DONE (ACK/NAK TIMEOUT) primitive.

This primitive is processed in the same manner as the DONE (ACK/NAK TIMEOUT) primitive.

DONE (NORMAL) Indicates that all frames have been sent.

DONE (RESERVED 0)

DONE (RESERVED 1)

This primitive is processed in the same manner as the DONE (NORMAL) primitive.

When the wait time for an RRDY primitive has exceeded 1

DONE (CREDIT TIMEOUT)

ms, or the CREDIT BLOCKED primitive is received before an RRDY primitive is received, if the close processing is performed, the DONE (CREDIT TIMEOUT) primitive is sent.

A drive sends a DONE primitive in any of the following cases:

1) When the drive sends an OPEN frame and establishes connection, and it becomes that there are no frame to be sent, the drive sends the DONE (NORMAL) primitive.

2) When credit timeout is detected before the drive sends an SSP frame, the drive sends the DONE (CREDIT TIMEOUT) primitive.

3) When the drive receives the CREDIT_BLOCKED primitive during an inside connection, the drive sends the DONE (CREDIT TIMEOUT) primitive during the close processing.

4) When an OPEN frame is sent by an initiator, connection is established, and no SSP frame has been sent for more than 1 ms, the drive sends the DONE (NORMAL) primitive.

5) When the drive sends a read DATA frame, and data burst size of the data that has been sent by the drive since connection establishment reaches 64 Kbytes, the drive sends the DONE (NORMAL) primitive to temporarily close the connection.

6) When the drive detects Ack/Nak timeout, the drive sends the DONE (ACK/NAK TIMEOUT) primitive.

Before sending the DONE primitive, a drive starts the done timer (1 ms). Later, when receiving an SSP frame from the initiator, the drive initializes the done timer and restarts it. If the drive has received neither the SSP frame nor the DONE primitive for 1 ms since sending the DONE primitive, the drive sends the BREAK primitive.

C141-C013 55

SAS Interface

Meanwhile, when a drive receives the DONE primitive, the drive performs the following processing:

1) If the drive has SSP frames to send at the time of the receipt of the DONE primitive from the initiator, the drive can send as many SSP frames as there are credits being held at this time.

2) If the drive has no SSP frames to send at the time of the receipt of the DONE primitive from the initiator, the drive sends the DONE (NORMAL) primitive.

The side that has sent the DONE primitive cannot send the SSP primitive until connection is established again after the current connection is temporarily closed.

1.4.4.4 EOF (End of frame)

EOF specifies the end of an SSP or SMP frame.

1.4.4.5 NAK (negative acknowledgement)

The NAK primitive indicates that an SSP frame cannot be received normally. When a drive receives an SSP frame, and the drive detects a CRC error after checking the CRC field in the SSP frame, the drive sends the NAK primitive.

If a received SSP frame contains the ERROR primitive or an invalid dword, the drive sends the NAK primitive even if the CRC field in the SSP field is correct.

1.4.4.6 RRDY

During connection, when a drive detects the state where it can receive an SSP frame, the drive sends an RRDY primitive to indicate that the connecting companion can send the SSP frame.

When the OPEN_ACCEPT primitive is transferred, a drive will send up to three RRDY primitives.

The sending side of the SSP frame cannot receive the SSP frame before sending RRDY primitives (before the sending side obtains credits). If a drive receives more SSP frames than the RRDY primitives it has sent, the drive discards the surplus SSP frames.

1.4.4.7 SOF (Start of frame)

SOF specifies the start of an SSP or SMP frame.

56 C141-C013

1.4 Link layer

1.4.5 Clock skew management

The internal clock for a device is typically based on a PLL with its own clock generator and is used when transmitting dwords on the physical link. When receiving, however, dwords need to be latched based on a clock derived from the input bit stream itself. Although the input clock is nominally a fixed frequency, it may differ slightly from the internal clock frequency up to the physical link rate tolerance. Over time, if the input clock is faster than the internal clock, the device may receive a dword and not be able to forward it to an internal buffer; this is called an overrun. If the input clock is slower than the internal clock, the device may not have a dword when needed in an internal buffer; this is called an underrun.

To solve this problem, transmitting devices insert ALIGNs or NOTIFYs in the dword stream.

A phy that is the original source for the dword stream (i.e., a phy that is not an expander phy forwarding dwords from another expander phy) shall insert one ALIGN or NOTIFY for clock skew management as described in Table 1.16.

Table 1.16 Clock skew management ALIGN insertion requirement

Physical link rate requirement

1.5 Gbps One ALIGN or NOTIFY within every 2 048 dwords

3.0 Gbps Two ALIGNs or NOTIFYs within every 4 096 dwords

In the clock skew management of a drive, one ALIGN primitive is always inserted every 1024 dwords regardless of the physical link rate.

1.4.6 Idle physical link

In any of the conditions below, when there is no dword to be sent, an idle dword is sent. This idle dword is scrambled.

1) In an outside connection

2) The SSP or SMP connection is an inside connection.

C141-C013 57

SAS Interface

1.4.7 Scrambling

Scrambling is used to reduce the probability of long strings of repeated patterns appearing on the physical link.

All data dwords are scrambled. Table 1.17 lists the scrambling for different types of data dwords.

Table 1.17 Scrambling for different data dword types

connection state Data dword type Description of scrambling

Outside connections

Inside SSP connection

SAS idle dword When a connection is not open and there are no

other dwords to transmit, vendor-specific scrambled data dwords shall be transmitted.

Address frame After an SOAF, all data dwords shall be

scrambled until the EOAF.

SSP frame After an SOF, all data dwords shall be

scrambled until the EOF.

SSP idle dword When there are no other dwords to transmit,

vendor-specific scrambled data dwords shall be transmitted.

To generate scrambled values that are to be encoded and transferred on the physical link layer, dwords to be sent are XOR-ed with the predefined pattern. If there is no transmission error, the received data of the dwords is XOR-ed with the same pattern after being decoded so that the dword values of the original data can be obtained.

The pattern that is XOR-ed with the data dwords is defined by the output of a linear feedback shift register implemented with the following polynomial:

G(x) = x 16 + x 15 + x

+ x 4 + 1

The value of the linear feedback shift register shall be initialized at each SOF and SOAF to FFFFh.

58 C141-C013

1.5 Address frames

1.5.1 Address frames overview

Address frames are used for the identification sequence and for connection requests. The address frame follows an SOAF and ends with an EOAF. Address frames shall only be sent outside connections. Address frames shall not be terminated early. All data dwords in an address frame shall be scrambled.

Table 1.18 defines the address frame format.

Table 1.18 Address frame format

Bit

Byte

0 ADDRESS FRAME TYPE

1 27 28 (MSB) 31

7 6 5 4 3 2 1 0

Frame Type dependent bytes

CRC

(LSB)

The ADDRESS FRAME TYPE field indicates the type of the address frame as follows:

• An ADDRESS FRAME TYPE field value of "0" indicates an IDENTIFY

address frame.

• An ADDRESS FRAME TYPE field value of "1" indicates an OPEN address

frame.

The CRC field contains the CRC value calculated for all the data of an address frame (data from the beginning of the frame to the byte preceding the CRC field). If an address frame has an undefined ADDRESS FRAME TYPE, has an incorrect frame length (longer or shorter than 7 dwords + the CRC length), or has a CRC error, the recipient discards the frame.

C141-C013 59

SAS Interface

(for Open

Restricted (for Open

1.5.2 IDENTIFY address frame

Table 1.19 defines the IDENTIFY address frame format used for the identification sequence. The IDENTIFY address frame is sent after the phy reset sequence completes if the physical link is a SAS physical link.

Table 1.19 IDENTIFY address frame format

Bit

Byte

1 Restricted (for OPEN address frame)

2 Reserved

3 Reserved

4 11 12 19 20 PHY IDENTIFIYR 21 27 28 (MSB) 31

7 6 5 4 3 2 1 0

Restricted

(for Open

address

frame)

DEVICE TYPE ADDRESS FRAME TYPE (0h)

SSP

INITIATOR

PORT

SSP

TARGET

PORT

Restricted (for Open address frame)

SAS ADDRESS

Reserved

CRC

STP

INITIATOR

PORT

STP

TARGET

PORT

SMP

INITIATOR

PORT

SMP

TARGET

PORT

Restricted

address

frame)

address

frame)

(LSB)

The DEVICE TYPE field specifies the type of device containing the phy, and is defined in Table 1.20.

Table 1.20 DEVICE TYPE field

Code Description

001b End device 010b Edge expander device 011b Fanout expander device

All others Reserved

The ADDRESS FRAME TYPE field shall be set to 0h.

60 C141-C013

1.5 Address frames

An SSP INITIATOR PORT bit set to one specifies that an SSP INIT port is present. An SSP INITIATOR PORT bit set to zero specifies that an SSP INIT port is not present. Expander devices shall set the SSP INITIATOR PORT bit to zero.

An SSP TARGET PORT bit set to one specifies that an SSP TARG port is present. An SSP TARGET PORT bit set to zero specifies that an SSP TARG port is not present. Expander devices shall set the SSP TARGET PORT bit to zero.

For SAS ports, the SAS ADDRESS field specifies the port identifier of the SAS port transmitting the IDENTIFY address frame. For expander ports, the SAS ADDRESS field specifies the device name of the expander device transmitting the IDENTIFY address frame.

The PHY IDENTIFIER field specifies the phy identifier of the phy transmitting the IDENTIFY address frame.

A wide port shall set the DEVICE TYPE field, SSP INITIATOR PORT bit, STP INITIATOR PORT bit, SMP INITIATOR PORT bit, SSP TARGET PORT bit, STP TARGET PORT bit, and SMP TARGET PORT bit to the same set of values on each phy in the wide port. Recipient ports need not check the consistency of these fields across phys.

The contents of the IDENTIFY address frame sent by a drive are as follows:

a) ADDRESS FRAME TYPE: 0h b) DEVICE TYPE: 001b c) SSP INITIATOR PORT bit: 0b d) STP INITIATOR PORT bit: 0b e) SMP INITIATOR PORT bit: 0b f) SSP TARGET PORT bit: 1b g) STP TARGET PORT bit: 0b h) SMP TARGET PORT bit: 0h i) SAS ADDRESS: Refer to the Section 1.2.1, "SAS address." j) PHY IDENTIFIER: port A (0h)/port B (1h) k) Reserved: 0h

C141-C013 61

SAS Interface

1.5.3 OPEN address frame

Table 1.21 defines the OPEN address frame format used for connection requests.

Table 1.21 OPEN address frame format

Bit

Byte

1 FEATURES CONNECTION RATE

2 (MSB)

4 11 12 19 20 COMPATIBLE FEATURES

21 PATHWAY BLOCKED COUNT 22 (MSB) 23 24 27 28 (MSB) 31

7 6 5 4 3 2 1 0

INITIATO

R PORT

PROTOCOL ADDRESS FRAME TYPE (1h)

INITIATOR CONNECTION TAG

DESTINATION SAS ADDRESS

SOURCE SAS ADDRESS

ARBITRARION WAIT TIME

MORE COMPATIBLE FEATURES

CRC

(LSB)

An INITIATOR PORT bit set to one specifies that the source port is acting as a SAS INIT port. An INITIATOR PORT bit set to zero specifies that the source port is acting as a SAS TARG port.

The PROTOCOL field specifies the protocol for the connection being requested and is defined in Table 1.22.

Table 1.22 PROTOCOL field

Code Description

000b SMP 001b SSP 010b STP

All others Reserved

62 C141-C013

1.5 Address frames

The ADDRESS FRAME TYPE field shall be set to 1h.

The FEATURES field shall be set to zero.

The CONNECTION RATE field specifies the connection rate being requested between the source and destination, and is defined in Table 1.23.

Table 1.23 CONNECTION RATE field

Code Description

8h 1,5 Gbps 9h 3,0 Gbps

All others Reserved

When requesting a connection to a SAS TARG port, a SAS INIT port shall set the CONNECTION RATE field to a value supported by at least one potential pathway.

When requesting an SSP connection to an SSP INIT port, an SSP TARG port shall set the CONNECTION RATE field to the connection rate in effect when the command was received unless it has received an OPEN_REJECT (CONNECTION RATE NOT SUPPORTED).

The INITIATOR CONNECTION TAG field is used when the SAS TARG port issues a connection request. This field is used as a substitute for the SAS address of the SAS TARG port when the SAS INIT port searches for processing in progress. If the initiator port does not require field support from the SAS TARG port, the INITIATOR CONNECTION TAG field is set to FFFFh. If the initiator port requires field support, the INITIATOR CONNECTION TAG field should be set to the value that is peculiar to the SAS TARG port. When the SAS TARG port issues a connection request to the SAS INIT port, the SAS TARG port must set the INITIATOR CONNECTION TAG field to any of the following: the last received value, the uncompleted command that has been received from the SAS INIT port, or the value that has been set at the time of the connection request of the task management. For all connection requests, the SAS INIT port must use the same INITIATOR CONNECTION TAG value for the same SAS TARG port. The SAS INIT port can change the INITIATOR CONNECTION TAG value only when there is no command processing in progress or task management in the SAS TARG port. The SAS TARG port does not check for consistency among INITIATOR CONNECTION TAG fields of different connection requests from the same SAS INIT port.

The DESTINATION SAS ADDRESS field indicates the port identifier of the SAS port to which a connection request is being made.

The SOURCE SAS ADDRESS field indicates the port identifier of the SAS port that has been issued an OPEN address frame.

C141-C013 63

SAS Interface

The COMPATIBLE FEATURES field is set to zero. The destination device ignores the COMPATIBLE FEATURES field.

The PATHWAY BLOCKED COUNT field specifies the number of times the port has retried this connection request due to receiving OPEN_REJECT (PATHWAY BLOCKED). The port shall not increment the PATHWAY BLOCKED COUNT value past FFh. If the port changes connection requests, it shall use a PATHWAY BLOCKED COUNT of 00h.

The ARBITRATION WAIT TIME field specifies how long the port transmitting the OPEN address frame has been waiting for a connection request to be accepted or rejected. This time is maintained by the port layer in an Arbitration Wait Time timer. For values from 0000h to 7FFFh, the Arbitration Wait Time timer increments in one microsecond steps. For values from 8000h to FFFFh, the Arbitration Wait Time timer increments in one millisecond steps. The maximum value represents 32767 ms + 32768 µs. Table 1.24 describes several values of the ARBITRATION WAIT TIME field.

Table 1.24 ARBITRATION WAIT TIME field

Code Description

0000h 0001h

0 µs 1 µs

... ...

7FFFh

8000h 8001h

32 767 µs 0 ms + 32 768 µs 1 ms + 32 768 µs

... ...

FFFFh

32 767 ms + 32 768 µs

The MORE COMPATIBLE FEATURES field shall be set to zero. The destination device shall ignore the MORE COMPATIBLE FEATURES field.

The contents of an OPEN address frame sent by a drive are as follows:

• INITIATOR PORT bit: 0b

• PROTOCOL: 001b

• CONNECTION RATE:

The CONNECTION RATE value when an initiator makes a connection request

• INITIATOR CONNECTION TAG:

The INITIATOR CONNECTION TAG value when an initiator makes a connection request

64 C141-C013

1.5 Address frames

• DESTINATION SAS ADDRESS:

The SAS address when an initiator makes a connection request

• SOURCE SAS ADDRESS:

The SAS address of the drive

• PATHWAY BLOCKED COUNT:

At the time of the first connection request, this field is set to 0h. Otherwise, it indicates how many times the OPEN_REJECT (PATHWAY BLOCKED) primitive has been received.

• ARBITRATION WAIT TIME:

At the time of the first connection request, this field is set to 0h. When a drive's connection request is unsuccessful due to arbitration fairness at the time of sending an OPEN frame, this field of the next OPEN frame indicates the time that has elapsed since the first connection request.

1.5.4 Identification and hard reset sequence

After the phy reset sequence has been completed indicating the physical link is using SAS rather than SATA, each phy transmits either:

a) an IDENTIFY address frame; or b) a HARD_RESET. Each phy receives an IDENTIFY address frame or a HARD_RESET from the phy

to which it is attached. The combination of a phy reset sequence, an optional hard reset sequence, and an identification sequence is called a link reset sequence (See Section 1.3.2).

If a device supports more than one phy, it shall transmit the same SAS address on all phys for which it is capable of sharing within a port.

If a device detects the same SAS address incoming on different phys, it shall consider those phys part of the same wide port.

If a device detects different SAS addresses incoming on different physical links, it shall consider those physical links as independent physical links and consider those phys part of different ports.

If a device does not receive a valid IDENTIFY address frame within 1 ms of phy reset sequence completion, it shall restart the phy reset sequence.

If a device receives an additional IDENTIFY address frame after receiving the first one, without an intervening phy reset sequence, it shall ignore the additional IDENTIFY address frame.

If a phy receives a HARD_RESET, it shall be considered a reset event and cause a hard reset (device initialization of the port) of the port containing that phy. After a hard reset, a phy begins a link reset sequence.

C141-C013 65

SAS Interface

1.5.5 Connections

1.5.5.1 Connections overview

A connection is opened between a SAS INIT port and a SAS TARG port before communication begins. A connection is established between one SAS INIT phy in the SAS INIT port and one SAS TARG phy in the SAS TARG port.

SSP INIT ports open SSP connections to transmit SCSI commands, task management functions, or transfer data. SSP TARG ports open SSP connections to transfer data or transmit status.

Also at the SSP TARG port, an SSP connection must be established before data or a status is sent.

The OPEN address frame is used to request that a connection be opened. AIP, OPEN_ACCEPT and OPEN_REJECT are the responses to an OPEN address frame. BREAK is used to abort connection requests and to unilaterally break a connection. CLOSE is used for orderly closing a connection.

Connections use a single pathway from the SAS INIT phy to the SAS TARG phy. While a connection is open, only one pathway shall be used for that connection.

A wide port may have separate connections on each of its phys.

1.5.5.2 Connection request

The OPEN address frame is used to open a connection from a source port to a destination port using one source phy and one destination phy.

To make a connection request, the source port shall transmit an OPEN address frame through an available phy. The source phy shall transmit idle dwords after the OPEN address frame until it receives a response or aborts the connection request with BREAK.

After transmitting an OPEN address frame, the source phy shall initialize and start a 1 ms Open Timeout timer. Whenever an AIP is received, the source phy shall reinitialize and restart the Open Timeout timer. Source phys are not required to enforce a limit on the number of AIPs received before aborting the connection request. When any connection response is received, the source phy shall reinitialize the Open Timeout timer. If the Open Timeout timer expires before a connection response is received, the source phy may assume the destination port does not exist and shall transmit BREAK to abort the connection request.

The OPEN address frame flows through expander devices onto intermediate physical links. If an expander device on the pathway is unable to forward the connect request because none of the prospective physical links support the requested connection rate, the expander device shall return OPEN_REJECT (CONNECTION RATE NOT SUPPORTED). If the OPEN address frame reaches the destination, it shall return either OPEN_ACCEPT or OPEN_REJECT. Rate matching shall be used on any physical links in the pathway with negotiated physical link rates that are faster than the requested connection rate.

66 C141-C013

1.5 Address frames

1.5.5.3 Connection responses

Table 1.25 lists the responses to an OPEN address frame being transmitted.

Table 1.25 Connection responses

Response Description

AIP Arbitration in progress. When an expander device is trying

to open a connection to the selected destination port, it returns an AIP to the source phy. The source phy shall reinitialize and restart its Open Timeout timer when it receives an AIP. AIP is sent by an expander device while it is internally arbitrating for access to an expander port.

OPEN_ACCEPT Connection request accepted. This is sent by the destination

phy.

OPEN_REJECT Connection request rejected. This is sent in response by the

destination phy or by an expander device.

OPEN address frame

If the AIP primitive has been detected just before the OPEN address frame, this indicates a top-priority connection request, meaning that the OPEN_ACCEPT primitive or an OPEN_REJECT primitive must be sent. If an AIP primitive has not been detected yet, this indicates that two connection requests have crossed over the physical link. At this time, arbitration fairness is used to determine which request should take priority.

BREAK The destination port or expander port may reply with

BREAK indicating the connection is not being established.

Open Timeout timer expires

The source phy shall abort the connection request by transmitting BREAK.

After an OPEN_REJECT (CONNECTION RATE NOT SUPPORTED) has been received by a SAS TARG port, the SAS TARG device shall set the connection rate for future requests for that I_T_L_Q nexus to:

a) the last value received in a connection request from the SAS INIT port; b) 1,5 Gbps; or c) the connection rate in effect when the command was received.

1.5.5.4 Arbitration fairness

SAS supports least-recently used arbitration fairness.

(This is a service equalization method that lowers the priority of the most recently used device, that is to say, gives higher priority to devices that have been in the wait state for a long period of time.)

C141-C013 67

SAS Interface

Each SAS port and expander port shall include an Arbitration Wait Time timer which counts the time from the moment when the port makes a connection request until the request is accepted or rejected. The Arbitration Wait Time timer shall count in microseconds from 0 µs to 32 767 µs and in milliseconds from 32 768 µs to 32 767 ms + 32 768 µs. The Arbitration Wait Time timer shall stop incrementing when its value reaches 32 767 ms + 32 768 µs.

SAS ports (i.e., SAS INIT ports and SAS TARG ports) shall start the Arbitration Wait Time timer when they transmit the first OPEN address frame for the connection request. When the SAS port retransmits the OPEN address frame (e.g., after losing arbitration and handling an inbound OPEN address frame), it shall set the ARBITRATION WAIT TIME field to the current value of the Arbitration Wait Time timer.

SAS ports should set the Arbitration Wait Time timer to zero when they transmit the first OPEN address frame for the connection request. A SAS INIT port or SAS TARG port may be unfair by setting the ARBITRATION WAIT TIME field in the OPEN address frame to a higher value than its Arbitration Wait Time timer indicates. However, unfair SAS ports shall not set the ARBITRATION WAIT TIME field to a value greater than or equal to 8000h; this limits the amount of unfairness and helps prevent livelocks.

The expander port that receives an OPEN address frame shall set the Arbitration Wait Time timer to the value of the incoming ARBITRATION WAIT TIME field and start the Arbitration Wait Time timer as it arbitrates for internal access to the outgoing expander port. When the expander device transmits the OPEN address frame out another expander port, it shall set the outgoing ARBITRATION WAIT TIME field to the current value of the Arbitration Wait Time timer maintained by the incoming expander port.

A port shall stop the Arbitration Wait Time timer and set it to zero when it wins arbitration (i.e., it receives either OPEN_ACCEPT or OPEN_REJECT from the destination SAS port rather than from an intermediate expander device). If a port receives a connection request that satisfies its arbitration request (i.e., it receives an OPEN address frame from the destination SAS port with the INITIATOR PORT bit set to the opposite value and a matching PROTOCOL field), it shall stop the Arbitration Wait Time timer unless it accepts the request (i.e., transmits an OPEN_ACCEPT rather than an OPEN_REJECT).

If two connection requests pass on a physical link, the phy shall determine the winner by comparing OPEN address frame field contents using the arbitration priority described in Table 1.26.

Table 1.26 Arbitration priority for OPEN address frames passing on a physical link

Bits 79-64 (79 is MSB) Bits 63-0 (0 is LSB)

ARBITRATION WAIT

TIME field value

SOURCE SAS ADDRESS

field value

68 C141-C013

1.5 Address frames

1.5.5.5 Aborting a connection request

BREAK may be used to abort a connection request. The source phy shall transmit a BREAK after the Open Timeout timer expires or if it chooses to abort its request for any other reason.

After transmitting BREAK, the source phy shall initialize a Break Timeout timer to 1 ms and start the Break Timeout timer. Table 1.27 lists the responses to a BREAK being transmitted before a connection response has been received.

Table 1.27 Abort connection responses

Response Description

BREAK This confirms that the connection request has been aborted. Break Timeout

timer expires

The originating phy shall assume the connection request has been aborted.

Figure 1.15 shows the sequence for a connection request where the Open Timeout timer expires.

Figure 1.15 Connection request timeout example

C141-C013 69

SAS Interface

1.5.5.6 Closing a connection

CLOSE is used to close a connection of any protocol. After transmitting CLOSE, the source phy shall initialize a Close Timeout timer to 1 ms and start the Close Timeout timer. Table 1.28 lists the responses to a CLOSE being transmitted.

Table 1.28 Close connection responses

Response Description

CLOSE This confirms that the connection has been closed. Close Timeout

timer expires

Figure 1.16 shows example sequences for closing a connection.

The originating phy shall attempt to break the connection.

Figure 1.16 Closing a connection example

70 C141-C013

1.5 Address frames

1.5.5.7 Breaking a connection

In addition to aborting a connection request, BREAK may also be used to break a connection, in cases where CLOSE is not available. After transmitting BREAK, the originating phy shall ignore all incoming dwords except for BREAKs.

After transmitting BREAK, the source phy shall initialize a Break Timeout timer to 1 ms and start the Break Timeout timer. Table 1.29 lists the responses to a BREAK being transmitted after a connection has been established.

Table 1.29 Break connection responses

Response Description

BREAK This confirms that the connection has been broken. Break Timeout

timer expires

In addition to a BREAK, a connection is considered broken due to loss of dword synchronization.

The following shall be the responses by an SSP phy to a broken connection:

a) Received frames having no CRC error may be considered valid regardless of

whether an ACK has been transmitted in response to the frame prior to the broken connection;

b) Transmitted frames for which an ACK has been received prior to a broken

connection shall be considered successfully transmitted; and

c) Transmitted frames for which an ACK or NAK has not been received prior to

a broken connection shall be considered not successfully transmitted.

1.5.5.8 Rate matching

In spite of the fact that the physical link is set to 3.0 Gbps, the connection rate (indicated in an OPEN address frame) may be 1.5 Gbps. As shown in Table 1.30, if the transfer rate of the physical link is higher than the connection rate, ALIGN or NOTIFY primitives must be inserted between dwords for each phy of the pathway.

The originating phy shall assume the connection has been broken. The originating phy may perform a link reset sequence.

C141-C013 71

SAS Interface

Table 1.30 Rate matching ALIGN and/or NOTIFY insertion requirements

Physical link rate

1,5 Gbps 1,5 Gbps None

3,0 Gbps

Connection

rate

1,5 Gbps

3,0 Gbps None

A phy shall start inserting ALIGNs and/or NOTIFYs for rate matching at the selected connection rate with the first dword that is not an ALIGN or NOTIFY inserted for clock skew management following:

a) transmitting the EOAF for an OPEN address frame; or b) transmitting an OPEN_ACCEPT. The source phy transmits idle dwords including ALIGNs and NOTIFYs at the

selected connection rate while waiting for the connection response. This enables each expander device to start forwarding dwords from the source phy to the destination phy after forwarding an OPEN_ACCEPT.

Requirement

One ALIGN or NOTIFY within every 2 dwords that are not clock skew management ALIGNs or NOTIFYs (i.e., every overlapping window of 2 dwords)

A phy shall stop inserting ALIGNs and/or NOTIFYs for rate matching after:

a) transmitting the first dword in a CLOSE; b) transmitting the first dword in a BREAK; c) receiving an OPEN_REJECT for a connection request; or d) losing arbitration to a received OPEN address frame.

1.5.6 SSP link layer

An SSP phy that accepts an OPEN address frame shall transmit at least one RRDY in that connection within 1 ms of transmitting an OPEN_ACCEPT. If the SSP phy is not able to grant credit, it shall respond with OPEN_REJECT (RETRY) and not accept the connection request.

SSP is a full duplex protocol. An SSP phy may receive an SSP frame or primitive in a connection while it is transmitting an SSP frame or primitive in the same connection. A wide SSP port may send and/or receive SSP frames or primitives concurrently on different connections (i.e., on different phys).

When a connection is open and an SSP phy has no more SSP frames to transmit on that connection, it transmits a DONE to start closing the connection. The other direction may still be active, so the DONE may be followed by one or more of the following primitives: CREDIT_BLOCKED, RRDY, ACK, or NAK.

72 C141-C013

1.5 Address frames

1.5.6.1 SSP frame transmission and reception

During an SSP connection, SSP frames are preceded by SOF and followed by EOF. The last data dword after the SOF prior to the EOF always contains a CRC.

An SSP phy checks the frame length and validity of the CRC.

Receiving phys must acknowledge ACK or NAK within 1 ms if there is no condition for discarding frames. The ACK primitive indicates that SSP frames have been received into a frame buffer without errors. The NAK (CRC ERROR) primitive indicates that SSP frames have been received with a CRC error.

1.5.6.2 SSP flow control

An SSP phy uses RRDY to grant credit for permission for the other SSP phy in the connection to transmit frames. Each RRDY increments credit by one frame. Frame transmission decrements credit by one frame. Credit of zero frames is established at the beginning of each connection.

SSP phys shall not increment credit past 255 frames.

To prevent deadlocks where an SSP INIT port and SSP TARG port are both waiting on each other to provide credit, an SSP INIT port shall never refuse to provide credit by withholding RRDY because it needs to transmit a frame itself. It may refuse to provide credit for other reasons (e.g., temporary buffer full conditions).

When the drive sends an OPEN frame and, then, receives the OPEN_ACCEPT primitive from the initiator due to abort processing in the drive (e.g., when receiving the TASK frame), the drive sends the CREDIT_BLOCKED primitive instead of an RRDY primitive. In this way, the drive may refuse to offer credits.

If credit is zero, SSP phys that are going to be unable to provide credit for 1 ms may send CREDIT_BLOCKED. The other phy may use this to avoid waiting 1 ms to transmit DONE (CREDIT TIMEOUT).

If credit is nonzero, SSP phys that are going to be unable to provide additional credit for 1 ms, even if they receive frames per the existing credit, may transmit CREDIT_BLOCKED.

After sending CREDIT_BLOCKED, an SSP phy shall not transmit any additional RRDYs in the connection.

1.5.6.3 Interlocked frames

Table 1.31 shows which SSP frames shall be interlocked and which are noninterlocked.

One approach for successively processing frames and corresponding ACK or NAK (responses to these frames) in a single processing unit consists of ensuring that no new frame is sent until the response to the preceding frame is returned. A frame processed in this way is called an interlocked frame.

C141-C013 73

SAS Interface

Table 1.31 SSP frame interlock requirements

SSP frame type Interlock requirement

COMMAND Interlocked TASK Interlocked XFER_RDY Interlocked DATA Non-interlocked RESPONSE Interlocked

Before transmitting an interlocked frame, an SSP phy shall wait for all SSP frames to be acknowledged with ACK or NAK, even if credit is available. After transmitting an interlocked frame, an SSP phy shall not transmit another SSP frame until it has been acknowledged with ACK or NAK, even if credit is available.

Before sending a non-interlocked frame, an SSP phy must wait for the following frames to be acknowledged with ACK or NAK, even if credit is available:

a) all non-interlocked frames with different tags; and b) all interlocked frames; After sending a non-interlocked frame, an SSP phy may send another non-

interlocked frame with the same tag if credit is available. The SSP phy must not send the following frames until all SSP frames have been acknowledged with ACK or NAK, even if credit is available:

a) a non-interlocked frame with a different tag; or b) an interlocked frame; Interlocking does not prevent sending and receiving interlocked frames

simultaneously (e.g., an SSP INIT phy can send a COMMAND frame while receiving XFER_RDY, DATA, or RESPONSE frames for a different command).

An SSP phy may transmit primitives responding to traffic it is receiving (e.g., an ACK or NAK to acknowledge an SSP frame, an RRDY to grant more receive credit, or a CREDIT_BLOCKED to specify that no more RRDYs are going to be transmitted in the connection) while waiting for an interlocked frame it transmitted to be acknowledged. These primitives may also be interspersed within an SSP frame.

74 C141-C013

1.5 Address frames

Of the primitives that may be inserted in SSP frames sent from the SSP INIT port and expander port, a drive can detect the following:

a) ACK b) NAK c) RRDY d) CREDIT_BLOCKED e) NOTIFY f) BREAK g) ERROR Meanwhile, the primitives that may be inserted in SSP frames sent from the drive

are as follows:

a) ACK b) NAK c) RRDY d) CREDIT_BLOCKED Figure 1.17 shows an example of the transfer of interlocked frames.

Figure 1.18 shows an example of the transfer of non-interlocked frames with the same tag.

Figure 1.19 shows an example of the transfer of non-interlocked frames with different tags.

Figure 1.17 Interlocked frames

C141-C013 75

SAS Interface

Figure 1.18 Non-interlocked frames with the same tag

Figure 1.19 Non-interlocked frames with different tags

1.5.6.4 Closing an SSP connection

DONE shall be exchanged prior to closing an SSP connection. There are several versions of the DONE primitive indicating additional information about why the SSP connection is being closed:

a) DONE (NORMAL) specifies normal completion; the transmitter has no more

SSP frames to transmit;

76 C141-C013

1.5 Address frames

b) DONE (CREDIT TIMEOUT) specifies that the transmitter still has SSP

frames to transmit but did not receive an RRDY granting frame credit within 1 ms, or the transmitter has received a CREDIT_BLOCKED and has consumed all RRDYs received; and

c) DONE (ACK/NAK TIMEOUT) specifies that the transmitter transmitted an

SSP frame but did not receive the corresponding ACK or NAK within 1 ms. As a result, the ACK/NAK count is not balanced and the transmitter is going to transmit a BREAK in 1 ms unless the recipient replies with DONE and the connection is closed.

If the transmitter has no more SSP frames to transmit and receives a CREDIT_BLOCKED, it may transmit either DONE (NORMAL) or DONE (CREDIT TIMEOUT).

After transmitting DONE, the transmitting phy initializes and starts a 1 ms DONE Timeout timer.

After transmitting DONE, the transmitting phy shall not transmit any more SSP frames during this connection. However, the phy may transmit ACK, NAK, RRDY, and CREDIT_BLOCKED as needed after transmitting DONE if the other phy is still transmitting SSP frames in the reverse direction. Once an SSP phy has both transmitted and received DONE, it shall close the connection by transmitting CLOSE (NORMAL).

Figure 1.20 shows the sequence for a closing an SSP connection.

Figure 1.20 Closing an SSP connection example

C141-C013 77

SAS Interface

1.6 Transport layer

1.6.1 SSP frame format

This layer defines the processing of each of a COMMAND frame, TASK frame, DATA frame, XFER_RDY frame and RESPONSE frame.

Table 1.32 contains the definitions of the SSP frame format. An SSP frame consists of the header part, information unit part, and CRC.

Table 1.32 SSP frame format

Bit

Byte

0 FRAME TYPE 1 (MSB) 3 4 Reserved 5 (MSB) 7 8 Reserved 9 Reserved

10 Reserved

11 Reserved 12 Reserved

13 15 16 (MSB) 17 18 (MSB) 19 20 (MSB) 23 24 (MSB)

Fill bytes (0 to 3bytes)

n-3 (MSB)

7 6 5 4 3 2 1 0

HASHED DESTINATION SAS ADDRESS

HASHED SOURCE SAS ADDRESS

RETRY

DATA

FRAMES

Reserved

TAG

TARGET PORT TRANSFER TAG

DATA OFFSET

INFORMATION UNIT

CRC

RETRA-

NSMIT

NUMBER OF FILL

(LSB)

CHANGING

DATA

POINTER

BYTES

(LSB)

78 C141-C013

1.6 Transport layer

Table 1.33 defines the FRAME TYPE field.

Table 1.33 FRAME TYPE field

Code Name of frame Information Unit Originator

Information unit size

01h DATA frame DATA INIT or TARG 1 to 1024

05h XFER_RDY

Transfer ready SSP TARG port 12

frame

06h COMMAND

Command SSP INIT port 28 to 280

frame

07h RESPONSE frame Response SSP TARG port

NO_DATA: 24 RESPONSE_DATA: 28 SENSE_DATA: 72

16h TASK frame Task

SSP INIT port 28 management function

All others Reserved (unknown frame)

The HASHED DESTINATION SAS ADDRESS field contains the hashed value of the destination SAS address.

(bytes)

The HASHED SOURCE SAS ADDRESS field contains the hashed value of the source SAS address

The RETRY DATA FRAMES bit is set to one for XFER_RDY frames and shall be set to zero for all other frame types. When set to one this bit specifies that the SSP INIT port may retry write DATA frames that fail.

The RETRANSMIT bit is set to one for TASK frames, RESPONSE frames, and XFER_RDY frames and shall be set to zero for all other frame types. This bit specifies that the frame is a retransmission after the SSP port failed in its previous attempt to transmit the frame.

The CHANGING DATA POINTER bit is set to one for DATA frames and shall be set to zero for all other frame types. When set to one this bit specifies that the frame is a retransmission after the SSP TARG port failed in its previous attempt to transmit the frame or a subsequent frame and the dataoffset field of the frame may not be sequentially increased from that of the previous frame.

The NUMBER OF FILL BYTES field specifies the number of fill bytes between the INFORMATION UNIT field and the CRC field. The NUMBER OF FILL BYTES field shall be set to zero for all frame types except DATA frames (i.e., all other frame types are already four-byte aligned).

C141-C013 79

SAS Interface

The TAG field contains a value that allows the SSP INIT port to establish a context for commands and task management functions.

For COMMAND frames and TASK frames, the SSP INIT port shall set the TAG field to a value that is unique for the I_T nexus established by the connection. An SSP INIT port shall not reuse the same tag when transmitting COMMAND frames or TASK frames to different LUNs in the same SSP TARG port. An SSP INIT port may reuse a tag when transmitting frames to different SSP TARG ports. The TAG field in a COMMAND frame contains the task tag defined in SAM-3. The TAG field in a TASK frame does not correspond to a SAM-3 task tag, but corresponds to an SAM-3 association. The tag space used in the TAG fields is shared across COMMAND frames and TASK frames (e.g., if a tag is used for a COMMAND frame, it is not also used for a concurrent TASK frame).

For DATA, XFER_RDY, and RESPONSE frames, the SSP TARG port shall set the TAG field to the tag of the command or task management function to which the frame pertains.

The TARGET PORT TRANSFER TAG field provides an optional method for an SSP TARG port to establish the write data context when receiving a write DATA frame (i.e., determine the command to which the write data corresponds).

SSP TARG ports may set the TARGET PORT TRANSFER TAG field to any value when transmitting any SSP frame. SSP TARG ports that use this field should set the TARGET PORT TRANSFER TAG field in every XFER_RDY frame to a value that is unique for the L_Q portion of the I_T_L_Q nexus (i.e., that is unique for every XFER_RDY that is outstanding from the SSP TARG port).

SSP INIT ports shall set the TARGET PORT TRANSFER TAG field as follows:

a) For each write DATA frame that is sent in response to an XFER_RDY frame,

the SSP INIT port shall set the TARGET PORT TRANSFER TAG field to the value that was in the corresponding XFER_RDY frame;

b) For frames other than DATA frames, the SSP INIT port sets the TARGET

PORT TRANSFER TAG field to FFFFh. However, this setting is not forcibly made.

Remarks: Drives do not support the FIRST BURST SIZE field of the

Disconnect-Reconnect mode page (page code 02h).

For DATA frames, the DATA OFFSET field is described later. For all other frame types, the DATA OFFSET field shall be ignored.

The INFORMATION UNIT field contains the information unit, the format of which is defined by the FRAME TYPE field. The maximum size of the INFORMATION UNIT field is 1024 bytes, making the maximum size of the frame 1052 bytes (1024 bytes of data + 24 bytes of header + 4 bytes of CRC).

80 C141-C013

1.6 Transport layer

If necessary, fill bytes are placed after the INFORMATION UNIT field so that the CRC field is aligned on a four-byte boundary. The number of the fill bytes is specified by the NUMBER OF FILL BYTES field. If the transferred bytes in the last DATA frame are not aligned on a four byte boundary, this field is set to the number of the extra bytes (4 - the remainder of (the number of transferred bytes/4) bytes).

The CRC field contains a CRC value that is computed over the entire SSP frame prior to the CRC field including the fill bytes (i.e., all data dwords between the SOF and EOF).

Drives ignore the Reserved fields.

1.6.2 Information units

1.6.2.1 COMMAND information unit

Table 1.34 defines the command IU. The COMMAND frame is sent by an SSP INIT port to request that a command be processed by the device server in a logical unit.

Table 1.34 COMMAND information unit

Bit

Byte

0 (MSB) 7 8 Reserved

10 Reserved 11 12 (MSB)

27 28

27+ n x 4

7 6 5 4 3 2 1 0

LOGICAL UNIT NUMBER

ENABLE

FIRST

BURST

ADDITIONAL CDB LENGTH

(n dwords)

Reserved TASK ATTRIBUTE

Reserved

CDB

ADDITIONAL CDB BYTES

(LSB)

The LOGICAL UNIT NUMBER field contains the address of the logical unit.

An ENABLE FIRST BURST bit set to "1" specifies that the first burst data item is transferred as defined by the FIRST BURST SIZE field in the DisconnectReconnect mode page. However, drives do not support this bit. Therefore, a drive ignores the FIRST BURST SIZE field of the Disconnect-Reconnect mode page.

Table 1.35 contains the definitions of the TASK ATTRIBUTE field. However, drives do not support ACA in Table 1.35. Moreover, when a reserved TASK

C141-C013 81

SAS Interface

ATTRIBUTE code is specified, a drive will treat the task as one with the SIMPLE task attribute.

Table 1.35 TASK ATTRIBUTE field

Code Task attribute

000b SIMPLE 001b HEAD OF QUEUE 010b ORDERED 011b Reserved 100b ACA (not supported)

101b-111b Reserved

The ADDITIONAL CDB LENGTH field contains the length of the ADDITIONAL CDB field in dwords (four bytes). A drive expects a max 64-byte long COMMAND frame.

The CDB and ADDITIONAL CDB BYTES fields together contain the CDB to be interpreted by the addressed logical unit.

The contents of the CDB are defined in the SCSI command standards (e.g., SPC-3).

When a drive receives a COMMAND frame, the following items are checked:

• HASHED DESTINATION SAS ADDRESS field

When this field contains a value different from the Hashed SAS address of the drive, the drive sends the ACK and CREDIT_BLOCKED primitives, and discards the COMMAND frame.

• HASHED SOURCE SAS ADDRESS field

When this field contains a value different from the SOURCE SAS ADDRESS field of the OPEN frame, the drive sends the CREDIT_BLOCKED primitive, and discards the COMMAND frame.

82 C141-C013

1.6 Transport layer

• TAG field

The drive checks for duplication of tags. If a tag is found to be duplicated in the SAS address of the port, the drive sends the ACK primitive to temporarily close the connection. Then, after re-establishing a connection, the drive reports the CHECK CONDITION status with the OVERLAPPED COMMANDS DETECTS error. At this time, the drive aborts processing of all the commands received from the SAS address of the port.

• CRC field

When a CRC error is detected, the drive sends the NAK primitive, and discards the COMMAND frame.

• Frame length

When the IU length is other than 28 bytes, the drive reports INVALID FRAME (Response Code=0x02).

• Command queue count value

If the drive receives a command when the queue is full, the drive sends the ACK primitive to temporarily close the connection. Then, after reestablishing a connection, the drive reports the QUEUE FULL status.

• LOGICAL UNIT NUMBER field

• TASK ATTRIBUTE field

• CDB field

Refer to Chapter 2, "Command Processing," and Chapter 4, "Command Specification."

• Reserved field

The drive ignores the Reserved fields except the CDB field.

• Management of the command queue count (supplementary explanation)

The maximum command queue count (the upper limit of the number of commands to be accepted) of a drive is 128. However, if the drive is a dual port drive and, for example, has already queued 128 commands at port A, the drive can receive another command at port B without causing the Queue full error. In other words, at least one command can be received at each port of the drive even when the drive is in the Queue full state.

C141-C013 83

SAS Interface

1.6.2.2 TASK information unit

Table 1.36 defines the task management function IU. The TASK frame is sent by an SSP INIT port to request that a task management function be processed by the task manager in a logical unit.

Table 1.36 TASK information unit

Bit

Byte

0 (MSB) 7 8 Reserved

9 Reserved 10 TASK MANAGEMENT FUNCTION 11 Reserved 12 (MSB) 13 14 27

7 6 5 4 3 2 1 0

LOGICAL UNIT NUMBER

TAG OF TASK TO BE MANAGED

Reserved

The LOGICAL UNIT NUMBER field contains the address of the logical unit.

Table 1.37 contains the definition of the TASK MANAGEMENT FUNCTION field. However, drives do not support the CLEAR ACA task management function listed in Table 1.37.

(LSB)

84 C141-C013

1.6 Transport layer

Table 1.37 TASK MANAGEMENT FUNCTION field

Task

Code

01h ABORT TASK yes yes

02h ABORT TASK SET yes no

04h CLEAR TASK SET yes no

08h

20h Reserved 40h CLEAR ACA yes no Not supported

80h QUERY TASK yes yes

Management

function

LOGICAL UNIT RESET

LOGICAL

UNIT

NUMBER field

yes no

TAG OF

TASK TO BE

MANAGED

Field

Description

Aborts a task that matches with respect to I-T-L-Q.

Aborts all the tasks that match with respect to I-T-L.

Aborts all the tasks of the INIT. Reports UNIT ATTENTION

(06/2F/00) to the INIT other than nexus that performed abort processing.

Aborts all the tasks of the INIT. Reports UNIT ATTENTION

(06/29/03) to all the INITs.

When maintaining a task that matches with respect to I-T-L-Q, reports TASK MANAGEMENT FUNCTION SUCCEEDED. When not maintaining a task that matches with respect to I-T-L-Q, reports TASK MANAGEMENT FUNCTION COMPLETE.

All others Reserved

If TASK MANAGEMENT FUNCTION is set to ABORT TASK or QUERY TASK, the TAG OF TASK TO BE MANAGED field specifies the TAG value from the COMMAND frame that contained the task to be aborted or checked. For all other task management functions, the TAG OF TASK TO BE MANAGED field shall be ignored.

When a drive receives a TASK frame, the following fields are checked:

• HASHED DESTINATION SAS ADDRESS field

When this field contains a value different from the hashed SAS address of the drive, the drive sends the ACK and CREDIT_BLOCKED primitives, and discards the TASK frame.

• HASHED SOURCE SAS ADDRESS field

When this field contains a value different from that in the SOURCE SAS ADDRESS field of the OPEN frame, the drive sends the CREDIT_BLOCKED primitive, and discards the TASK frame.

• TAG field

C141-C013 85

SAS Interface

The drive checks for duplication of tags between multiple TASK frames or between a COMMAND frame and a TASK frame.

If a tag is found to be duplicated in the INIT, the drive sends the ACK primitive to temporarily close the connection. Then the drive reports the RESPONSE CODE of the OVERLAPPED TAG ATTEMPTED (0x0A).

• CRC field

When a CRC error is detected, the drive sends the NAK primitive, and discards the TASK frame.

• Frame length

When the IU length is other than 28 bytes, the drive reports INVALID FRAME (Response Code=0x02).

• LOGICAL UNIT NUMBER field

If this field is other than zero, the drive sets the response code to INVALID LOGICAL UNIT NUMBER (0x09), and sends the RESPONSE frame of RESPONSE DATA.

• TASK MANAGEMENT FUNCTION field

If this field is other than ABORT TASK, ABORT TASK SET, CLEAR TASK SET, LOGICAL UNIT RESET, or QUERY TASK, the drive sets the response code to TASK MANAGEMENT FUNCTION NOT SUPPORTED (0x04), and sends the RESPONSE frame of RESPONSE DATA.

1.6.2.3 XFER_RDY information unit

Table 1.38 defines the transfer ready IU. The XFER_RDY frame is sent by an SSP TARG port to request write data from the SSP INIT port.

Table 1.38 XFER_RDY information unit

Bit

Byte

0 (MSB)

4 (MSB)

8 11

7 6 5 4 3 2 1 0

WRITE DATA LENGTH

REQUESTED OFFSET

Reserved

(LSB)

The REQUESTED OFFSET field contains the application client buffer offset of the segment of write data that is sent to the SSP (initiator port as a logical unit). The requested offset is a multiple of four (i.e., each DATA frame begins transferring data on a dword boundary). The REQUESTED OFFSET field is always zero for the first XFER_RDY frame of a command (drives do not support the ENABLE FIRST BURST field of the COMMAND frame).

86 C141-C013

1.6 Transport layer

If any additional XFER_RDY frames are required, the REQUESTED OFFSET field must be set to the value of the REQUESTED OFFSET in the previous XFER_RDY frame plus the value of the WRITE DATA LENGTH field in the previous XFER_RDY frame.

The WRITE DATA LENGTH field contains the number of bytes of write data the SSP INIT port will send to the logical unit from the application client buffer starting at the requested offset. The SSP TARG port sets the WRITE DATA LENGTH field to a value greater than or equal to 00000001h.

If an SSP TARG port sends a XFER_RDY frame containing a WRITE DATA LENGTH field that is not divisible by four, the SSP TARG port must not send any subsequent XFER_RDY frames for that command (i.e., only the last XFER_RDY for a command may request a non-dword multiple write data length).

When the drive sends a XFER_RDY frame, the following values are set:

• HASHED DESTINATION SAS ADDRESS field

The hashed SAS address of the COMMAND frame is set.

• HASHED SOURCE SAS ADDRESS field

The hashed SAS address of the drive is set.

• NUMBER OF FILL BYTES field

Always set to zero.

• TAG field

The tag value of the COMMAND frame is set.

• TARGET PORT TRANSFER TAG field

The drive sets a value between 0 and 7.

• DATA OFFSET field

Always set to zero.

• REQUESTED OFFSET field

The first XFER_RDY frame for a command is always set to zero. If multiple XFER_RDY frames are to be sent for a single command, the value set in the field of the second XFER_RDY is increased by the value in the WRITE DATA LENGTH field of the first XFER_RDY frame. However, when transferring multiple XFER_RDY frames for a single command, the drive does not send the additional XFER_RDY frame until it receives the data requested with the first XFER_RDY frame. The value of the REQUESTED OFFSET field is divisible by four.

Table 1.39 An example of requested offset

XFER_RDY frame REQUESTED OFFSET

WRITE DATA

C141-C013 87

SAS Interface

LENGTH

First frame 0000h 0400h Second frame 0400h 0400h Third frame 0800h 0400h

• WRITE DATA LENGTH field

This field is used to set the number of data bytes that the drive can receive. The WRITE DATA LENGTH field indicates the actual number of data bytes to be received by the drive, and may not be divisible by four.

Figure 1.21 shows an example case where multiple XFER_RDY frames are sent.

Figure 1.21 Example of XFER_RDY frames

88 C141-C013

1.6 Transport layer

1.6.2.4 DATA information unit

Table 1.40 defines the data IU. The DATA frame is sent by an SSP INIT port to deliver write data (i.e., a write DATA frame) and is sent by an SSP TARG port to deliver read data (i.e., a read DATA frame). The maximum size of the data IU is the maximum size of any IU in an SSP frame. The minimum size of the data IU is one byte.

Table 1.40 DATA information unit

Bit

Byte

n-1

7 6 5 4 3 2 1 0

DATA (1byte to 1024bytes)

An SSP INIT port shall only transmit a write DATA frame:

a) in response to an XFER_RDY frame; or b) after transmitting a COMMAND frame if the ENABLE FIRST BURST field

in the COMMAND frame was set to one and the FIRST BURST SIZE field in the Disconnect-Reconnect mode page is not zero.

However, since drives do not support the FIRST BURST SIZE field, a drive can receive data only when it requests data with XFER_RDY frames.

If the value in the MAXIMUM BURST SIZE field on the Disconnect-Reconnect mode page is not zero, the maximum amount of data that is transferred at one time by an SSP TARG port per I_T_L_Q nexus is limited by the value in the MAXIMUM BURST SIZE field.

Drives do not support the MAXIMUM BURST SIZE field. However, the maximum data amount a drive can request with a single XFER_RDY frame is limited to 64 Kbytes.

An SSP TARG port shall set the NUMBER OF FILL BYTES field to zero in the frame header in all read DATA frames for a command except the last read DATA frame for that command. The SSP TARG port may set the NUMBER OF FILL BYTES field to a non-zero value in the last read DATA frame for a command (i.e., only the last read DATA frame for a command may contain data with a length that is not a multiple of four).

An SSP INIT port may set the NUMBER OF FILL BYTES field to a non-zero value contained in the last DATA frame that is sent as a response to an XFER_RDY frame. In this case, the SSP INIT port must set zeros in the NUMBER OF FILL BYTES fields in the frame headers of all the other data frames to be sent.

C141-C013 89

SAS Interface

The initial read DATA frame for a given command shall set the DATA OFFSET field to zero. If any additional read DATA frames are required, the DATA OFFSET field shall be set to the value of the previous read DATA frame's data offset plus the previous read DATA frame's data length.

The initial write DATA frame for a given command shall set the DATA OFFSET field to zero. If any additional write DATA frames are required, the DATA OFFSET field shall be set to the value of the previous write DATA frame's data offset plus the previous write DATA frame's data length.

When a drive receives a write DATA frame, the following items are checked:

• HASHED DESTINATION SAS ADDRESS field

When this field contains a value different from the Hashed SAS address of the drive, the drive sends the ACK and CREDIT_BLOCKED primitives, and discards the DATA frame.

• HASHED SOURCE SAS ADDRESS field

When this field contains a value different from the SOURCE SAS ADDRESS field of the OPEN frame, the drive sends the CREDIT_BLOCKED primitive, and discards the DATA frame.

• CRC field

When a CRC error is detected, the drive sends the NAK primitive, and discards the DATA frame. At this time, the SENSE DATA for INFORMATION UNIT iuCRC ERROR DETECTED is reported.

• Frame length

When the IU length is greater than 1024 bytes, the drive discards the DATA frame. At this time, the SENSE DATA for INFORMATION UNIT TOO LONG is reported.

• TAG field

When a DATA frame in which the same value as that set in the TARGET PORT TRANSFER TAG field of the XFER_RDY frame is received, but the tag is different from the one of the XFER_RDY frame, the drive sends the ACK primitive and CREDIT_BLOCKED primitive, and discards the DATA frame.

• TARGET PORT TRANSFER TAG (TPTT) field

When a DATA frame with the TPTT field that does not match with the TPTT field of the XFER_RDY frame is received, the drive sends the ACK primitive and CREDIT_BLOCKED primitive, and discards the DATA frame. At this time, after temporarily closing the connection, the drive re-opens a connection and sends the SENSE DATA for INVALID TARGET PORT TRANSFER TAG RECEIVED.

90 C141-C013

1.6 Transport layer

• DATA OFFSET field

The DATA OFFSET field is used to check whether frames are being transferred in the correct order. When the drive receives a frame with an offset value that the drive does not expect, the drive sends the ACK primitive and CREDIT_BLOCKED primitive, and discards the DATA frame. At this time, after temporarily closing the connection, the drive re-opens a connection and sends the SENSE DATA for DATA OFFSET ERROR.

• Received data length

When a data item with a data length exceeding the value specified in the WRITE DATA LENGTH field of the XFER_RDY frame is received, the drive sends the ACK primitive and CREDIT_BLOCKED primitive, and discards the DATA frame. At this time, after temporarily closing the connection, the drive re-opens a connection and sends the SENSE DATA for TOO MUCH WRITE DATA.

Meanwhile, when a drive sends a read DATA frame, the following values are set:

• HASHED DESTINATION SAS ADDRESS field

The hashed SAS address of the COMMAND frame is set.

• HASHED SOURCE SAS ADDRESS field

The hashed SAS address of the drive is set.

• NUMBER OF FILL BYTES field

If the transferred bytes in the last DATA frame for the command are not aligned on a four byte boundary, the number of the extra bytes (4 - the remainder of [the number of all transferred bytes/4] bytes) is set in this field.

• TAG field

The tag value of the COMMAND frame is set.

• TARGET PORT TRANSFER TAG field

Set to 0xFFFF by the drive.

• DATA OFFSET field

For the first DATA frame of the command, this field is set to zero. For the second DATA frame, the DATA OFFSET field is increased by the number of the transferred bytes of the first DATA frame.

• Number of transferred bytes

When the number of bytes to be transferred is greater than or equal to 1024 bytes, this field is set to 1024. When the number of bytes to be transferred is less than 1024 bytes, this field is set to that number of bytes.

C141-C013 91

SAS Interface

1.6.2.5 RESPONSE information unit

Table 1.41 defines the response IU. The RESPONSE frame is sent by an SSP TARG port to deliver SCSI status (e.g., GOOD or CHECK CONDITION) and sense data, or to deliver SSP-specific status (e.g., illegal frame format). The maximum size of the RESPONSE frame is the maximum size of any IU in an SSP frame.

Table 1.41 RESPONSE information unit

Bit

Byte

10 Reserved DATAPRES 11 STATUS 12 15 16 (MSB) 19 20 (MSB) 23

24 RESPONSE DATA (DATAPRES: 01h) 23+m 24+m SENSE DATA (DATAPRES: 10b)

23+m+n

7 6 5 4 3 2 1 0

Reserved

SENSE DATA LENGTH (n bytes)

REPONSE DATA LENGTH (m bytes)

Table 1.42 defines the DATAPRES field.

(LSB)

Table 1.42 DATAPRES field

Code Name Description

00b NO_DATA The SENSE DATA LENGTH field is zero, and the

RESPONSE DATA LENGTH is zero.

01b RESPONSE_DATA The SENSE DATA LENGTH field is zero, and the

RESPONSE DATA LENGTH is four.

10b SENSE_DATA The SENSE DATA LENGTH field is 48, and the

RESPONSE DATA LENGTH is zero.

11b Reserved

The SSP TARG port shall return a RESPONSE frame with the DATAPRES field set to NO_DATA if a command completes without response data or sense data to return.

92 C141-C013

1.6 Transport layer

The SSP TARG port shall return a RESPONSE frame with the DATAPRES field set to RESPONSE_DATA in response to every TASK frame and in response to errors that occur while the transport layer is processing a COMMAND frame.

The SSP TARG port shall return a RESPONSE frame with the DATAPRES field set to SENSE_DATA if a command completes with sense data to return (e.g., CHECK CONDITION status).

• RESPONSE information unit NO_DATA format

If the DATAPRES field is set to NO_DATA, then:

a) the SSP TARG port shall set the STATUS field to the status code for a

command that has ended (see SAM-3 for a list of status codes);

b) the SSP TARG port shall set the SENSE DATA LENGTH field to zero and

the RESPONSE DATA LENGTH field to zero;

c) the SSP TARG port shall not include the SENSE DATA field and the

RESPONSE DATA field.

• RESPONSE information unit RESPONSE_DATA format

If the DATAPRES field is set to RESPONSE_DATA, then:

a) the SSP TARG port shall set the STATUS field to zero and the SENSE

DATA LENGTH field to zero; b) the SSP TARG port shall not include the SENSE DATA field; c) the SSP TARG port shall set the RESPONSE DATA LENGTH field to four. d) the SSP TARG port shall include the RESPONSE DATA field. Table 1.43 defines the RESPONSE DATA field, which contains information

describing protocol failures detected during processing of a request received by the SSP TARG port. The RESPONSE DATA field shall be present if the SSP TARG port detects any of the conditions described by a non-zero value in the RESPONSE CODE field and shall be present for a RESPONSE frame sent in response to a TASK frame.

C141-C013 93

SAS Interface

Table 1.43 RESPONSE DATA field

Bit

Byte

0 Reserved 1 Reserved 2 Reserved 3 RESPONSE CODE

7 6 5 4 3 2 1 0

Table 1.44 defines the RESPONSE CODE field, which specifies the error condition or the completion status of a task management function.

Table 1.44 RESPONSE CODE field

Code Description

00h TASK MANAGEMENT FUNCTION COMPLETE 02h INVAILD FRAME 04h TASK MANAGEMENT FUNCTION NOT SUPPORTED 05h TASK MANAGEMENT FUNCTION FAILED 08h TASK MANAGEMENT FUNCTION SUCCEEDED 09h INVAILD LOGICAL UNIT NUMBER

0Ah OVERLAPPED TAG ATTEMPTED

All others Reserved

• RESPONSE information unit SENSE_DATA format

If the DATAPRES field is set to SENSE_DATA, then:

a) The STATUS field contains the status code for a command whose processing

has ended. b) The RESPONSE DATA LENGTH field is set to zero. c) No RESPONSE DATA field exists. d) The SENSE DATA LENGTH field indicates the length of the SENSE DATA

field in bytes by a non-zero value. For a drive, the SENSE DATA LENGTH

field is set to 48 bytes. e) The SENSE DATA field contains sense information (see Chapter 5).

94 C141-C013

1.6 Transport layer

When the drive sends a RESPONSE frame, the following values are set:

• HASHED DESTINATION SAS ADDRESS field

The hashed SAS address of the COMMAND frame is set.

• HASHED SOURCE SAS ADDRESS field

The hashed SAS address of the drive is set.

• NUMBER OF FILL BYTES field

Set to zero.

• RETRANSMIT bit

When Ack timeout occurs after the drive sends a RESPONSE frame, the drive resends the RESPONSE frame with this bit set to 1.

• TAG field

The tag value of the COMMAND frame is set.

• TARGET PORT TRANSFER TAG field

This field is set to 0xFFFF by the drive.

• DATA OFFSET field

Set to zero.

• DATAPRES field

Refer to Table 1.42.

• STATUS field

The SCSI status byte is set (see Chapter 2).

• SENSE DATA LENGTH field

When sense data is sent, this field is set to 0x30.

• RESPONSE DATA LENGTH field

When RESPONSE data is sent, this field is set to 0x04.

• RESPONSE DATA field

Refer to Table 1.43.

• SENSE DATA field

Refer to Chapter 5.

C141-C013 95

SAS Interface

1.6.3 Sequences of SSP frames

Figure 1.22, Figure 1.23, and Figure 1.24 show examples of the sequences of frames for single task management functions and commands. Frames may be interleaved in any order when multiple commands and/or task management functions are outstanding. Frames may be transmitted during one or more connections (e.g., the COMMAND frame could be transmitted in a connection originated by the SSP INIT port, and the DATA frames and RESPONSE frame transmitted in one or more connections originated by the SSP TARG port).

Figure 1.22 shows the sequence of SSP frames for a task management function.

Figure 1.22 Example of TASK frame

96 C141-C013

Fujitsu MBA3147RC, MBA3073RC, MBA3300RC, MBC2036RC, MBC2073RC User Manual

Specifications and Main Features

Frequently Asked Questions

User Manual

FOR SAFE OPERATION

Revision History

Preface

Contents

Illustrations

CHAPTER 1 SAS Interface

1.1 Topologies in SAS Interface

1.1.1 SAS Layering

1.1.2 Physical links and phys

1.1.3 Ports (narrow ports and wide ports)

1.1.4 SAS devices

1.1.5 Pathways

1.1.6 Connections

1.2.2 Hashed SAS address

1.3.4 Out of band (OOB) signals

1.3.5 SAS OOB sequency

1.3.5.1 Exception handling in the OOB sequence

1.3.6 SAS speed negotiation sequence

1.3.6.2 When the speed negotiation sequence is successful

1.3.6.3 Exception handling in the speed negotiation sequence

1.3.7 Phy layer dword synchronization (DWS)

1.4 Link layer

1.4.1 Primitives

1.4.2 Primitive sequences

1.4.3 Primitives not specific to type of connections

1.4.3.1 AIP (Arbitration in progress)

1.4.3.2 ALIGN

1.4.3.4 BROADCAST

1.4.3.5 CLOSE

1.4.3.6 EOAF (End of address frame)

1.4.3.7 ERROR

1.4.3.8 HARD_RESET

1.4.3.10 OPEN_ACCEPT

1.4.3.11 OPEN_REJECT

1.4.3.12 SOAF (Start of address frame)

1.4.4 Primitives used only inside SSP and SMP connections

1.4.4.1 ACK (acknowledge)

1.4.4.2 CREDIT_BLOCKED

1.4.4.3 DONE

1.4.4.4 EOF (End of frame)

1.4.4.5 NAK (negative acknowledgement)

1.4.4.6 RRDY

1.4.4.7 SOF (Start of frame)

1.4.5 Clock skew management

1.4.6 Idle physical link

1.4.7 Scrambling

1.5.2 IDENTIFY address frame

1.5.3 OPEN address frame

1.5.4 Identification and hard reset sequence

1.5.5 Connections

1.5.5.1 Connections overview

1.5.5.2 Connection request

1.5.5.3 Connection responses

1.5.5.4 Arbitration fairness

1.5.5.5 Aborting a connection request

1.5.5.6 Closing a connection

1.5.5.7 Breaking a connection

1.5.5.8 Rate matching

1.5.6 SSP link layer

1.5.6.1 SSP frame transmission and reception

1.5.6.2 SSP flow control

1.5.6.3 Interlocked frames

1.6.2 Information units

1.6.2.1 COMMAND information unit

1.6.2.2 TASK information unit

1.6.2.3 XFER_RDY information unit

1.6.2.4 DATA information unit

1.6.3 Sequences of SSP frames