02376-00 ATCA-2210 SCM Reference

PROMENTUM
SWITCH AND CONTROL MODULE

REFERENCE
ATCA-2210
www.radisys.com
007-02376-0000 December 2006
Release history Release -0000 Date December 2006 Description First release.
Copyright 2006 by RadiSys Corporation. All rights reserved. RadiSys is a registered trademark and Promentum is a trademark of RadiSys Corporation. MontaVista is a registered trademark of MontaVista Software Inc. Red Hat and Red Hat Linux are registered trademarks of Red Hat, Inc. Linux is a registered trademark of Linus Torvalds. All other trademarks, registered trademarks, service marks, and trade names are the property of their respective owners.
TABLE OF CONTENTS
Preface ...........................................................................................................................................................
About this manual.................................................................................................................................................................. Notational conventions......................................................................................................................................................... Electrostatic discharge .......................................................................................................................................................... Where to get more product information..........................................................................................................................
7
7 7 7 8
Chapter 1: Features Overview...................................................................................................................... 11

Local management processor............................................................................................................................................. Ethernet switches ................................................................................................................................................................... Network timing subsystem .................................................................................................................................................. Intelligent Platform Management Controller ................................................................................................................... Installed software ................................................................................................................................................................... External connectivity.............................................................................................................................................................. 11 11 12 12 12 12
Chapter 2: Theory of Operation................................................................................................................... 13

Possible applications of the SCM ....................................................................................................................................... Overview of functionality ..................................................................................................................................................... 13 13
Chapter 3: Subsystems.................................................................................................................................. 19
Local management processor............................................................................................................................................. Clock synthesizer subsystem ............................................................................................................................................... Base Ethernet switch subsystem......................................................................................................................................... Fabric Ethernet switch subsystem ...................................................................................................................................... LED controller interface subsystem ................................................................................................................................... Network timing subsystem .................................................................................................................................................. Rear transition module ......................................................................................................................................................... Hardware management subsystem ................................................................................................................................... Power supplies........................................................................................................................................................................ 20 23 24 28 31 35 37 41 43
Chapter 4: Physical Interfaces ...................................................................................................................... 45

Mechanical specification....................................................................................................................................................... 45
3
Table of Contents
Front panel interfaces............................................................................................................................................................ 46 Backplane interfaces.............................................................................................................................................................. 49 Internal interfaces................................................................................................................................................................... 56
Chapter 5: Initial Verification and Configuration ....................................................................................... 59

Configuration checklist.......................................................................................................................................................... Connecting to the SCM......................................................................................................................................................... Logging in to the SCM .......................................................................................................................................................... Rebooting the SCM................................................................................................................................................................ Configuring IP addresses...................................................................................................................................................... Adding user accounts and setting passwords.................................................................................................................. Configuring SNMP trap destinations and security access............................................................................................. Enabling the SNMP agent .................................................................................................................................................... Enabling the SCMs DHCP server....................................................................................................................................... Enabling and disabling the Telnet server and TFTP server........................................................................................... Enabling and disabling syslog and other services .......................................................................................................... Copying a configuration from existing CLI configuration files..................................................................................... Copying CLI configuration files to a TFTP server............................................................................................................ Copying any file to the SCM................................................................................................................................................ 59 60 60 60 61 63 63 64 65 66 67 68 68 69
Chapter 6: Software Features....................................................................................................................... 71

Shelf Manager......................................................................................................................................................................... Ethernet switching.................................................................................................................................................................. IP routing.................................................................................................................................................................................. Linux and boot loader........................................................................................................................................................... RAM and file system usage.................................................................................................................................................. IPv4 interface and Ethernet switch port identifiers......................................................................................................... Pre-defined IP addressing scheme..................................................................................................................................... DHCP server............................................................................................................................................................................ Telnet and SSH ....................................................................................................................................................................... Network timing subsystem (NTS) software ..................................................................................................................... NTP server ............................................................................................................................................................................... Module state management.................................................................................................................................................. Flash and file system management ................................................................................................................................... Software image management ............................................................................................................................................. Firmware and software upgrades....................................................................................................................................... Logging service.......................................................................................................................................................................
4
72 72 72 73 74 77 79 82 83 83 86 86 86 86 86 87
E-Key authorization and notification..................................................................................................................................
87
Chapter 7: SNMP Agent Support ................................................................................................................. 89

Base or Fabric switch selection ........................................................................................................................................... 89 ifIndex values for IP interfaces ............................................................................................................................................ 89 MIB module support............................................................................................................................................................. 90 Notification support............................................................................................................................................................... 91 MIB modules supplied.......................................................................................................................................................... 92 MIB browser applications..................................................................................................................................................... 93
Chapter 8: API Overview............................................................................................................................... 95

Where to find related files and additional information ................................................................................................. Ethernet switching API .......................................................................................................................................................... NTS API..................................................................................................................................................................................... Module state management API .......................................................................................................................................... Flash and file system management API............................................................................................................................ Software image management API...................................................................................................................................... Management plane framework (MPF) API ...................................................................................................................... 95 96 96 97 97 98 98
Chapter 9: Using the Command Line Interface ......................................................................................... 99

Accessing the master CLI...................................................................................................................................................... Displaying available commands and options.................................................................................................................. Master CLI command modes.............................................................................................................................................. No commands .................................................................................................................................................................... Show commands................................................................................................................................................................ Saving configurations persistently...................................................................................................................................... Command line completion.................................................................................................................................................. Editing and history keys........................................................................................................................................................ Master CLI commands.......................................................................................................................................................... 99 99 100 100 101 101 101 101 102
Chapter 10: Blade Management Commands ............................................................................................. 105

Accessing blade-management command modes .......................................................................................................... 105 SNMP commands.................................................................................................................................................................. 106 Network timing subsystem (NTS) commands ................................................................................................................ 106
Chapter 11: Maintenance.............................................................................................................................. 119

Installing the SCM .................................................................................................................................................................. 119 Removing the SCM ............................................................................................................................................................... 119
Table of Contents
Troubleshooting...................................................................................................................................................................... 120
Appendix A: Specifications............................................................................................................................ 125

Environmental......................................................................................................................................................................... Safety ........................................................................................................................................................................................ Mechanical dimensions ........................................................................................................................................................ Electromagnetic compatibility (EMC)................................................................................................................................ Power........................................................................................................................................................................................ 125 126 126 127 128
Appendix B: Managed Sensors .................................................................................................................... 129
PREFACE
About this manual

This manual describes the Promentum ATCA-2210, a switch and control module (SCM), which is fully compliant with AdvancedTCA (ATCA). The SCM is designed to be incorporated into High Availability (HA) systems such as the Promentum platforms SYS-6006 and SYS-6010. Use this manual as a reference for the operation and the maintenance of the ATCA-2210 SCM. The material presented here is not introductory; it is assumed that you are already familiar with the intended use of the ATCA-2210 SCM in your organizations ATCA platform. For instructions on initial setup of the ATCA-2210 SCM, refer to the ATCA-2210 Switch and Control Module Installation Guide. The simplified names SCM and module will be used in place of ATCA-2210 SCM for the rest of the manual.
Notational conventions
This manual uses the following conventions: BoldText
MonoText ItalicText
A command or keyword. Screen text and syntax strings. Variable parameters.
All numbers are decimal unless otherwise stated.
Electrostatic discharge
WARNING This product contains static-sensitive components and should be handled with care. Failure to employ adequate anti-static measures can cause irreparable damage to components. Electrostatic discharge (ESD) damage can result in partial or complete device failure, performance degradation, or reduced operating life. To avoid ESD damage, the following precautions are strongly recommended. Keep each module/PCB in its ESD shielding bag until you are ready to install it. Before touching a module, attach an ESD wrist strap to your wrist and connect its other end to a known ground.
Preface
Handle the module only in an area that has its working surfaces, floor coverings, and chairs connected to a known ground. Hold modules only by their edges and mounting hardware. Avoid touching PCB components and connector pins. For further information on ESD, visit www.esda.org.
Where to get more product information

Please visit the RadiSys Web site at www.radisys.com for product information and other resources. Downloads (manuals, release notes, software, etc.) are available via the Technical Support Library product links at www.radisys.com/support or the product pages at www.radisys.com/products.
About related RadiSys Products

The SCM is part of the Promentum SYS-6006 and SYS-6010 platforms. For information on the Promentum product family and other RadiSys products, see the RadiSys Web site at www.radisys.com.
Standards information
For information about the PCI Industrial Computer Manufacturers Group (PICMG) and the AdvancedTCA standard, consult the PICMG Web site at this URL: http://www.picmg.org
Related Documents
PICMG 3.0 R2.0 Advanced Telecommunications Computing Architecture (AdvancedTCA), PICMG PICMG 3.1 R1.0 Ethernet/Fibre Channel over PICMG 3.0, PICMG PICMG COM.0 R1.0 COM Express Module Base Specification, PICMG Small Form-factor Pluggable (SFP) Transceiver MultiSource Agreement (MSA), Cooperation Agreement for Small Form-Factor Pluggable Transceivers, September 14, 2000, Agilent Technologies, Blaze Network Products, E2O Communications, Inc., ExceLight Communications, Finisar Corporation, Fujikura Technology America Corp., Hitachi Cable, Infineon Technologies Corp., IBM Corp., Lucent Technologies, Molex, Inc., Optical Communication Products, Inc., Picolight, Inc., Stratos Lightwave, Tyco Electronics. INF-8077i 10 Gigabit Small Form Factor Pluggable Module (XFP), Revision 4.5, XFP Multi Source Agreement (MSA) Group UL 60950-1 Safety for Information Technology Equipment CAN/CSA 22.2 #60950-1-03 Safety for Information Technology Equipment, CSA EN 60950-1:2002 Safety for Information Technology Equipment, CENELEC IEC 60950-1 Safety for Information Technology Equipment, IEC
Where to get more product information
GR-63-CORE NEBS Requirements Physical Protection, Issue 2, April 2002, Telcordia GR-78-CORE Generic Requirements for the Physical Design and Manufacture of Telecommunications Products and Equipment, Issue 1, September 1997, Telcordia GR-1089-CORE Electromagnetic Compatibility and Electrical Safety Generic Criteria for Network Telecommunications Equipment, Issue 3, October 2002, Telcordia
Preface
10
FEATURES OVERVIEW
This guide describes the ATCA-2210 switch and control module (SCM) and serves as a hardware and software reference for its electrical, mechanical, and environmental aspects. The SCM is fully compliant with the ATCA standard and is designed to be incorporated into high availability (HA) systems, which meet 99.999% availability. The module provides highly integrated, centralized common equipment functionspacket switching, shelf management, network timing, and provides a site for a daughter card processor, which can be used for system management. It has comprehensive hardware management subsystem, managed by the IPMC, which interfaces to the Shelf Manager and functions as a fully managed building block. The HA infrastructure uses two SCMs that each contain a local management processor (LMP), a 1 Gigabit Base Ethernet switch, a 10 Gigabit Fabric Ethernet switch, a network timing subsystem (NTS), an Intelligent Platform Management Controller (IPMC), and a Shelf Manager. Duplication of the SCM and its supporting components allows for redundancya means of providing continuing functionality even if a failure occurs on one of the SCMs. This chapter provides a features overview of the SCM and its components. Detailed information on the theory and the operation of the SCM is covered in the chapters that follow.
Local management processor

The SCM incorporates a powerful on-board PowerQUICCIII processor for its local management functions. The local management processor (LMP) is used to manage the Ethernet switches, manage the NTS, and to provide access to the hardware management subsystem. It may also function as the Shelf Manager, depending on the software configuration you implement through your application software. See page 20 for an overview of the LMPs features, and subset of features, as they apply to the implementation of the SCM.
Ethernet switches
The SCM is has two Ethernet switches: 1 Gigabit Base Ethernet switch The Base Ethernet is comprised of a single Broadcom BCM56304 Ethernet switch. See page 24 for more information on the Base Ethernet switch subsystem. 10 Gigabit Fabric Ethernet switch The Fabric Ethernet interface uses a single Broadcom BCM56800 Ethernet switch, which has twenty ports configurable for 1Gb or 10 Gb operation. See page 28 for more information on the Fabric Ethernet switch subsystem.
11
Features Overview
Network timing subsystem

The network timing subsystem (NTS) is comprised of a circuit block and software. The circuit block provides a centralized timing source for the shelf based on the AdvancedTCA synchronization clock Interface specification. The NTS selects between several internal and external timing references, and provides a phase-locked, highly stable system clock to the platform modules managed by the SCMs Shelf Manager and the NTS software. See page 35 for information on the circuit block and page 83 for details about the software aspects of the NTS.
Intelligent Platform Management Controller

The SCM uses a Renesas H8S/2166 micro-controller for the Intelligent platform management controller (IPMC). This device manages commands and data as part of the hardware management subsystem, which is comprised of the IPMCs and the sensors from installed modules and field replaceable units (FRUs) and their communication with the Shelf Manager. See page 41 for more information on the hardware management system and the IPMC.
Installed software
The SCMs Linux-based software includes many features and several management interfaces, including Ethernet switching protocol support, a PICMG-compliant Shelf Manager, a command line interface (CLI), application programming interfaces (APIs), and an SNMP agent. For a complete overview, see Software Features on page 71.
External connectivity
The SCM provides connectivity for communicating with external equipment through three interfaces: Front panel interfaces Front panel interfaces are accessed through each SCMs faceplate. The connection types are labeled on the faceplate. See page 46 for an illustration of the front panel and a description of all its components, including LEDs and port pinouts. Backplane interfaces The backplane is the primary electrical interconnection between the modules in a platform. The Zone 2 connectors of the backplane provide connectivity for the Base, the Fabric, the update channel interfaces, and synchronization clocks. See page 49 for details on the backplane contact assignments and connectors. Rear transition module interfaces When installed in a RadiSys shelf, each SCM has an associated rear transition module (RTM) site, which can provide external access to SCMs signals and connections. See page 37 for descriptions of the RTMs electrical connections and signal interfaces.
12
THEORY OF OPERATION
Possible applications of the SCM

The SCM is ideal for next generation network elements in third generation (3G) wireless and wire-line infrastructures because it provides highly integrated, centralized common equipment functionsswitching, shelf management, and network timing. Possible applications include the following: Carrier grade billing and network servers IPTV switching infrastructure Media gateway (MGW) Push-to-connect/Push-to-talk (POC/PTT) server Radio network controller (RNC) Softswitches Session border controllers (SBC) Web servers
Overview of functionality
The SCM provides system interfaces through Base Ethernet, Fabric Ethernet, clock I/O, and two Intelligent Platform Management Buses (IPMB). An LMP with a PowerQUICCIII chip has two serial interfaces, a local Inter-integrated circuit (I2C) bus, a PCI bus, and a memory bus. A 10/100 Ethernet channel from the PowerQUICCIII is routed to the front panel to provide an interface for switch management. The memory bus provides interfaces to the DRAM and boot flash memory. The PCI bus provides the interface to the Base Ethernet and the Fabric Ethernet switches. Figure 1 shows a block diagram.
13
Theory of Operation
Figure 1. SCM system block diagram

RJ45 USB RJ45 RS -232 USB 1Gb Eth 1Gb 1Gb RJ45 RJ45 RS -232 FE RTC XFP XFP PCI XFP XFP 10Gb 3-4 10Gb Base Ethernet Switch 1 FE Switch 1 1Gb PCI Fabric Ethernet Switch PQIII Local Management Processor NTS Sync Sync Chan Upd Chan Clk 2 2-8 Clk Clk 10Gb 10Gb 1Gb 1Gb 1-4 Zone 3 1Gb SFP SFP SFP SFP COM Express * SAS/ SATA* Hard Drive SFP* SFP SFP SFP
1Gb 100Mb
10Gb 12-14 FE 2 FE
Node Slots E1/DS1 Framer
RJ45 RJ45 RJ45 RJ45
XFP
Sh Mgr
SFP SFP SFP SFP
1Gb 0-4
Node 1Gb-Tx Slots 12-14 Zone 2 IPMB 48V Zone 1
Serial Local Power Conversion 12V -48V Power Conversion
RTM (Illustration)
ShMC IPMC
SCM
* Reserved for future implementation.
Blue = Base Ethernet Red = Fabric Ethernet
The SCM as a Shelf Manager

The SCM is configured to act as the Shelf Manager by default. The shelf management duties are divided between the LMP and the Intelligent Platform Management Controller (IPMC). The IPMC on the SCM is equivalent to a shelf management controller (ShMC). The Shelf Manager is responsible for monitoring and controlling the behavior of the components within the shelf and the ShMC is the device that communicates with the IPMCs on the field replaceable units (FRUs). The high-level aspects of shelf management are handled by the Shelf Manager software, which is described in detail in the Shelf Management Software Reference.
14
Intelligent Platform Management

Intelligent Platform Management (IPM) is a subset of shelf management. IPM handles aspects of shelf management that involve communication between the Shelf Manager and the IPMCs on the FRUs. The Shelf Manager and all the IPMCs communicate through the Intelligent Platform Management Bus (IPMB). This communication uses the Intelligent Platform Management Interface (IPMI) protocol. Figure 2 provides an overview of the IPMB architecture on RadiSys 14-slot shelf. Figure 2. Example IPMB architecture on a RadiSys 14-slot shelf
Fan IPMC Fan IPMC
Fan
IPMC
Fan
IPMC
Shelf display or alarm panel
Sensors
SPM IPMC
I2C
IPMC
IPMC
ShMC
ShMC
IPMC
IPMC
Module
Module
SCM
SCM
Module
Module
(Modules in some slots not shown)
(Modules in some slots not shown)
Note: IPMC Intelligent platform managment controller SCM Switch and control module ShMC Shelf management controller SPM Shelf peripheral module PEM Power entry module
PEM
IPMC
PEM
IPMC
The Shelf Manager retrieves FRU information from the IPMCs and stores this information. Possible types of information that can be retrieved from a FRU include information on its capabilities, sensor readings, and hot-swap state. To provide redundancy, the active Shelf
15
Theory of Operation
Manager also sends the FRU information to the inactive Shelf Manager, which also stores the FRU information. To implement full redundancy, the IPMB is set up as two I2C buses. This dual-bus architecture prevents shelf management from being interrupted by the failure of either bus. Each ShMC and each IPMC connects to both I2C buses, so if one bus becomes unavailable, communication takes place on the other bus. To prevent a failed ShMC or an IPMC from significantly disrupting either or both buses, each ShMC and IPMC has associated circuitry that isolates the ShMC or the IPMC from both buses if its associated watchdog timer fails to be triggered periodically.
Base Ethernet interface

When the SCMs are installed on a platform, like the RadiSys SYS-6010, they provide a Base interface that supports the 10/100/1000Base-T Ethernet standard on the backplane. Each SCM also provides 10/100/1000Base-T Ethernet uplinks on its front panel. The topology of the Base interface forms a dual star. Figure 3 illustrates this topology, showing the relationship between the node and the hub modules. Figure 3. Dual star topology
Node
Node
Node
Hub (SCM)
Hub (SCM)
Node
Node
Node
16
Figure 4 illustrates the dual star topology in terms of the backplane connections between modules on a RadiSys 14-slot shelf. Figure 4. Backplane connections through modules on a RadiSys 14-slot shelf
Ethernet interface
2 1 Channel
2 1 Channel
Channel
Channel
Module
Module
(Modules not shown in some slots)
SCM
SCM
(Modules not shown in some slots)
Module
Module
Uplinks Slot 1 Slot 2 Slot 7
Uplinks Slot 8 Slot 13 Slot 14
Node slots
Hub slots
Node slots
Each Ethernet backplane link consists of four differential signal pairs. Each SCM is installed in a hub slot and acts as a switch that connects to every node (non-SCM) module to form the star topology. This redundant topology provides an alternate path if one Ethernet path fails, or if one SCM fails. The SCMs connect to each other through a direct Ethernet connection (Base channel 2), which the high-level shelf-management software uses. Traffic through the SCM-to-SCM link is controlled by the Base Ethernet switch, which can also pass user traffic. For information on the software managing the Base Ethernet interface, see The Ethernet Switching Software Reference.
Fabric Ethernet interface

The SCM supports a 10 Gb Ethernet switch fabric for the Fabric Interconnect using the full fabric interface channel in a 10 Gb XAUI configuration as specified in the PICMG 3.1 specification. The Fabric Ethernet interface also implements a dual star topology similar to the topology of the Base Ethernet interface. However, there is no direct switch-to-switch link.
17
Theory of Operation
For information on the software managing the Fabric Ethernet interface, see The Ethernet Switching Software Reference.
Redundancy
Typically, two SCMs are installed on a platform for redundancy purposes. Redundancy provides a means of continuing functionality even if a failure occurs on one of the SCMs: Each SCM contains switches for both the Base Ethernet and the Fabric Ethernet. If either SCM fails or is hot-swapped, the other SCM switches all the traffic for both the Base and the Fabric interface. The SCMs operate on an active/active basis, which means the switches in both SCMs are always operating. Both SCMs run the same switch configuration and management software. If one SCM is removed or fails, the copy in the other SCM continues to control the switches in the remaining active SCM. Each SCM has a Shelf Manager. If one SCM is removed or fails, the Shelf Manager in the other SCM operates as the active Shelf Manager. The dual Shelf Managers operate on an active/standby basis, which means only one Shelf Manager is active at any time.
18
SUBSYSTEMS
The SCM is comprised of a series of subsystems that work together to provide the network elements required for third generation wireless and wire-line infrastructures. This chapter covers the details of the subsystems that make up the SCM. Figure 5. Detailed block diagram of SCM
RJ45 RJ45 232 Buf 232 Buf LPC USB USB COM Express* 12V Switch 2 1 RJ -45 Mag RJ -45 Mag 1Gb Eth 2 4 2 10/100-Tx 2 10/10 10/10 PHY PHY MII Serial XFP XAUI to XFI XAUI to XFI XAUI to XFI XAUI to XFI
Dual -Boot Flash
SIO
SAS/SATA Hard Drive*
12V Clock/ calendar 3.3V
Switch Switch I2C Buf Buf PCI-E Zone 3 Conn

Buf Buf
Local Power Conversion
SAS/SATA PCI-E
Dual 1 Gb MAC SERDES
ID ID & Temp
SFP*
EE PROM
I2C
SFP
SFP 1Gb SD SFP
SPI 4 3 NTM
2 ClkIn RefClk
SPI ClkIn RefClk 1Gb SD
SFP
MII
SFP
1Gb-Tx Sync Clk

10/100-Tx
Upd Chan 1,2 Upd Chan 3,4 Sync I/F Upd Chan 0
SFP
DDR1 SODIMM
XFP
PQIII Local Management Processor Mag
SFP
XFP
Serial
BITS/ SSU Clk BITS/ SSU Clk
Mag I/F Mag I/F Mag I/F Mag I/F
RJ -45
10 Gb XAUI 10 Gb XAUI TBI X 20-port 10Gb Eth Switch X 2

10Gb XAUI Node Slots
XFP
RJ -45
BITS/ SSU Clk BITS/ SSU Clk
RJ -45
XFP
XAUI to XFI
SER
14 PCI-32 SPI 2 FE Switch 1 2
DES
RJ -45
1Gb SD SFP 10Gb XAUI 10Gb XAUI SFP 2 1Gb SFP X 4 2

24-port Gb + 2 port 10Gb Eth Switch
ShMC 10/100 -Tx Base Ch1 2
14 1Gb SGMII COM -E Type Det
16-port Gb Eth PHY
16-port Mag
14 1Gb-Tx 1 EE PROM
Node Slots F-Link Base Ch2
RTM (Illustration)
Temp Sensors
I2C
SFP Flsh Sel IPMC IPMB 2 Local Power Conversion 12V DC/DC Power Converter 48V Tyco PIM-200 Power Cond Fuses
Zone 2 Conn
Zone 1 Conn
SCM3
* Reserved for future implementation. Blue = Base Ethernet Red = Fabric Ethernet
19
3 Local management processor

The local management processor (LMP) is used to manage the Ethernet switches, to manage the network timing subsystem (NTS), and to provide access to the hardware management subsystem. The LMP provides Ethernet ports, serial communications, an inter-integrated circuit (I2C) controller, a serial peripheral interface (SPI), memory, and peripheral component interconnect (PCI) bus interfaces. A 1Gigabit Ethernet port provides a high-speed interface to the Base switch. A 10/100 Mb Ethernet channel from the LMP is routed to the front panel and provides an interface for switch management. A 1Gigabit Ethernet channel is routed across the backplane to the redundant SCM. The I2C controller and the SPI buses control on-board peripheral devices. The memory buses provide interfaces to the local synchronous dynamic random access memory (SDRAM), and the boot flash memories. The PCI bus connects to the management ports of the Base and the Fabric Ethernet switches. The LMP is comprised of the following components: A PowerQUICCIII processor chip with 833 MHz core frequency A 200-pin SODIMM memory connector for connecting to a memory module with a capacity of 1Gb DDR SDRAM A redundant pair of 64 MB Spansion flash memory components (128MB total) for PowerQUICCIII configuration, Linux operating system boot image, and file system Clock generation circuitry Power good generation circuitry Two 10/100 Fast Ethernet interfaces to a front panel connector and the update channel A serial interface to a front panel connector A serial bus interface to the Intelligent Platform Management controller (IPMC) An I2C bus interface to the serial EEPROMs, the clock/calendar, and the rear transition module (RTM) A SPI bus to the NTS, the Fast Ethernet switch, and the RTM A peripheral component interconnect (PCI) bus configuration interface to the Base and the Fabric Ethernet switches A 1 Gb Ethernet port to the Base switch
The block diagram in Figure 6 shows the relationship between the LMP and the SCM.
20
Local management processor
Figure 6. Local management processor with SCM

Local Management Processor
Base interface switch PCI bus 0 10/100/1000 MB Ethernet Base interface switch
Fabric interface switch
PCI bus 1
10/100/1000 MB Ethernet
Front panel mgmt port
10/100Base-T Ethernet Local bus 10/100Base-T Ethernet MDIO DRAM bus
64MB Flash ROM 0
64MB Flash ROM 1
512MB DDR1 SDRAM
Update channel link
MAC address
User
Front panel console port
EEPROM UART0 I2C bus
EEPROM
Clock/ calendar
RTM
Network Timing module Fast Ethernet switch
Serial port to IPMC
UART1 SPI bus
RTM
COP debug header
JTAG/COP
PowerQUICCIII processor
The PowerQUICCIII is the processor chip used to run the LMP. The PowerQUICCIII is packaged in a 783-pin ball grid array and is highly integrated with an embedded e500 core; integrated instruction and data caches, a system interface unit, and an integrated reduced instruction set computer-based (RISC) communications processor. The PowerQUICCIII used in the LMP has the following characteristics: Processor MPC8541 Core Freq. (MHz) 833 PCI Bus SYSCLK VDD/Core Freq. (MHz) Freq. (MHz) 66 66 3.3V/1.2V I Cache (Kbytes) 32 DCache (Kbytes) 32 Processor Version 0x0080
21
3
The PowerQUICCIII processor chip incorporates the following elements: CPU with e5000 core that implements Book E 32-bit architecture 256KB on-chip memory DDR memory controller Programmable interrupt controller (PIC) compliant with OpenPIC architecture Local Bus Controller (LBC), which connects to the two 64MB flash components Two PCI bus controllers used to configure the Base and the Fabric switches I2C controller providing access to the SDRAM module, two serial EEPROMs, the clock/calendar, and devices on the RTM Two UART interfaces for: Serial console port routed through front panel Serial port providing link between the PowerQUICCIII and the on-board IPMC Two 10/100Base-T Ethernet interfaces (Fast Ethernet) One interface is used for management and maintenance purposes One interface is used for connecting to the redundant SCM through the update channel One 10/100/1000 interface configured as a ten-bit interface (TBI), which is linked to a serializer/deserializer (SERDES). The interface is connected to the Base Ethernet switch as an alternate port to the PCI bus. Serial management data input/output (MDIO) port for managing the physical layer devices (PHYs) on the 10/100Base-T interfaces SPI controller providing a four-wire serial bus for accessing the NTS, the 10/100Base-T Ethernet switch, and devices on the RTM EEPROM used to store media access control (MAC) addresses for Ethernet ports EEPROM used for storing user-defined information Clock/calendar to keep a running time and date. Includes lithium battery to maintain timekeeping for a specific minimum period of time when power is not available to the SCM. General purpose input/output (I/O) signals to and from various functions on the SCM for control and status Synchronous dynamic random access memory (SDRAM) interface for the small outline dual in-line memory module (SODIMM). Two Spansion flash memory devices used as a redundant set of 64MB flash stores. Reads I/O bus information on the SCM version and revision history Integrated reset logic
22
Clock synthesizer subsystem
Communications controller assignments

The PowerQUICCIII communications functions as they pertain to the SCM are mapped out in the following table. Table 1. Communication controller assignments PowerQUICCIII communications function PCI0 PCI1 TSEC1 UART0 UART1 I2C CPM:FCC0 CPM:FCC1 CPM:SPI CPM:I2C SCM function Base interface switch Fabric interface switch Gb link to Base interface switch RS232 console port IPMC interface SMBus to SDRAM, EEPROM, clock/calendar 10/1000Base-T Ethernet port (front panel) 10/100Base-T Ethernet port (update channel) NTS, Fast Ethernet switch, RTM interface Not used
Clock synthesizer subsystem

The SCM has a clock synthesizer subsystem, which includes two clocks used by the PowerQUICCIII. The other clocks are buffered and distributed to several Ethernet chips. These include Ethernet switch, PHY reference, XAUI-to-XFI interface, and the serial communications clocks. All system clocks are series-terminated at the clock source.
23
3 Base Ethernet switch subsystem

One of the primary functions of the SCM is to provide the Ethernet switch function for ATCA platforms. This is implemented using a Broadcom 28-port Gigabit switch. Twenty-four of the ports are 1 Gb and four of the ports are 10 Gb. The key features of the 28-port Gigabit Ethernet switch include the following: Twenty-four 10/100/1000Mb ports Four 10 Gb ports Ethernet switch/router with integrated MACs, packet buffer engine and switching engine Supports line-rate Layer 2 switching and Layer 2 multicast for all packet sizes and conditions. Line-rate switching for all packet sizes and conditions Eight layer Quality of Service (QoS) on a per-port basis Supports multiple (256) Spanning Trees (IEEE 802.1D) Supports Flow Control (IEEE 802.3x) VLAN Support based on MAC, port, subnet, and protocol (IEEE 802.1Q) Supports link aggregation (up to 8 ports) (IEEE 802.3ad) IPSec engine supports DES, 3DES, HMAC-SHA-1, and HMAC-MD5, and is compatible with IPSec, SSL, TLS, 802.1x, and 802.11i. Head-of-line blocking prevention Per-port packet rate (storm) control Supports port mirroring PCI interface to PowerQUICCIII host processor SGMII interface support to external 1000 Mb PHYs Integrated SERDES Thermally enhanced 1156-pin FCBGA package
Figure 7 is a block diagram showing how the Base Ethernet switch communicates with the rest of the SCM.
24
Base Ethernet switch subsystem
Figure 7. Base Ethernet switch

Base Ethernet Switch
PowerQUICCIII
PCI Bus 1 GPIO TSEC1
PCI Bus Reset
Port Port Port Port Port Port Port Port
0 1 2 3 4 5 6 7
Qua d GbE PHY
Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics FE Switch w/ Mags
BCh2 BCh3 BCh4 BCh5 BCh6 BCh7 BCh8 BCh9 BCh10 BCh11 BCh12 BCh13 BCh14 BCh15 BCh16 BCh1 ShMC1 ShMC2 Backplane
Quad GbE PHY
TBI Gb SERDES
Port 8 Port 9 Port 10 Port 11 Port 23 Port Port Port Port 12 13 14 15
Quad GbE PHY
Quad GbE PHY
COM Express *
PCI -Express
Port 17 Port 22 Port 25 Port 27
Zone 3 Conn
RTM
GbE MAC SERDES
Port 16
Port Port Port Port
18 19 20 21 XAUI to XFI
Quad GbE SFP
Front Panel
Port 24 Port 26
XFP
Front Panel
* Reserved for future implementation
Configuration
The switch has built in media access controllers (MACs) for each port that interface to the external physical layer devices (PHYs). The 1 Gigabit Ethernet ports can each be configured to operate in: Serial Gigabit Media Independent Interface (SGMII) mode, which has Gigabit Ethernet ports going to the backplane, faceplate RJ-45 ports, and RTM. Serializer/deserializer (SERDES) modes, which have Gigabit Ethernet ports going to optical SFP modules or internal links. The 10 Gigabit Ethernet ports are configured to operate in 10 Gigabit attachment unit interface (XAUI) mode, with ports routed to Zone 3 and the faceplate XFP ports.
25
3
Port mapping
Table 2. Base Ethernet port mapping MDIO interface Port destination address 0 0/0 0x00 FE Switch (Shelf Manager) 1 0/1 0x01 Base channel 2 (link to redundant SCM) 2 0/2 0x02 Base channel 3 3 0/3 0x03 Base channel 4 4 1/2 0x06 Base channel 5 5 1/3 0x07 Base channel 6 6 1/0 0x04 Base channel 7 7 1/1 0x05 Base channel 8 8 2/2 0x0A Base channel 9 9 2/3 0x0B Base channel 10 10 2/0 0x08 Base channel 11 11 2/1 0x09 Base channel 12 12 3/2 0x0E Base channel 13 13 3/3 0x0F Base channel 14 14 SERDES/SGMII N/A RTM 15 SERDES/SGMII N/A RTM 16* SERDES N/A COM Express 17 SERDES/SGMII N/A RTM 18 SFP N/A Front panel 1G 1/6 19 SFP N/A Front panel 1G 1/7 20 SFP N/A Front panel 1G 1/8 21 SFP N/A Front panel 1G 1/9 22 SERDES/SGMII N/A RTM 23 SERDES N/A PowerQUICCIII X0 XAUI IF 0x10 Front panel XFP 1/5 X1 XAUI IF TBD RTM XAUI X2 XAUI IF N/A Not used X3 XAUI IF TBD RTM XAUI * Reserved for future implementation. Port # PHY/Port IPMI hardware Logical slot address N/A N/A 0x41 / 42 1/2 0x43 0x44 0x45 0x46 0x47 0x48 0x49 0x4A 0x4B 0x4C 0x4D 0x4E N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 3 4 5 6 7 8 9 10 11 12 13 14 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A
26
Base Ethernet switch subsystem
10/100/1000Base-T Gigabit Ethernet ports

Sixteen Gigabit Ethernet PHYs provide 10/100/1000Base-T connectivity to the backplane and the Fast Ethernet (FE) switch. All of the 1000Base-T ports use a Gigabit PHY channel along with Gigabit Ethernet transformers to couple to the backplane. 12 ports to backplane node slots. One port to an FE switch, allowing two FE connections to an external Shelf Manager (if present) through Base channel 1. One port to the Base interface of a redundant SCM (if present) through Base channel 2.
SERDES Gigabit Ethernet ports

Six ports are configured as serializer/deserializer (SERDES) ports for a 14-slot (default) backplane. Two of the SERDES ports go to the on-board LMP and the COM Express site. Four of the SERDES ports can be directed to SFP sockets on the SCM front panel or routed to the RTM, depending on the build option. The default is four ports routed to the SFP sockets on the SCM front panel.
10 Gigabit Ethernet ports

Three 10Gb Base Ethernet ports provide external access on the front and rear faceplates. 1 XFP port on the front faceplate, XFI signaling converted from XAUI 2 XAUI signaling to RTM The XFP 10Gb connector uses XFI signaling, composed of single transmit and receive 10Gb differential pairs. A XAUI-to-XFI interface chip converts the signaling from the XAUI signals (connected to the switch) to XFI signals (connected to the XFP port) at the front panel.
Fast Ethernet switch

A 5-port 10/100Base-T Ethernet switch connects one Base interface port to one or two offboard Shelf Managers through Base channel 1. The Fast Ethernet (FE) switch is managed by the PowerQUICCIII using the SPI port. Table 3. Fast Ethernet switch port mapping Port # 1 2 3 4 5 Port destination Base interface Base channel 1a Base channel 1b Not used Not used Port states Port 1 should always be enabled. Ports 2 and 3 may be enabled or disabled (together), depending on SCM configuration for Base connectivity to an off-board Shelf Manager. Ports 4 and 5 are not used, and should be disabled.
27
3
Port status LEDs
This table describes the LEDs specific to the Base Ethernet. Table 4. Base Ethernet port status LEDS Port
1 Gb SFP ports
Each Base Ethernet SFP port on the front panel or on the RTM has two LEDs
10 Gb XFP ports
Each Base Ethernet XFP port on the front panel or on the RTM has two associated LEDs.
LED and possible states 1 green link status LED Green link established Blinking green link activity Off link fail or port is disabled 1 green port status LED Green port is enabled Off port is disabled 1 green link/activity status LED Green link established Blinking green link activity Off link fail or port is disabled 1 green port status LED Green port is enabled Off port is disabled
Fabric Ethernet switch subsystem

For its standard configuration, the SCM provides a 10 Gb Fabric Ethernet connection to 14 node slots based on AdvancedTCA 3.1. Four ports of the Fabric interface are also available on the front panel and three ports are routed to the RTM through the Zone 3 area. The Fabric interface is implemented using a Broadcom 20-port 10 Gb Ethernet switch. The backplane ports use XAUI, made up of four lanes (Tx and Rx differential pairs) operating at 3.125 Gb. The faceplate uses XFI on an XFP connector, converted from the XAUI signaling on the switch ports by a XAUI-to-XFI transceiver. The Fabric Interface connects: Node slots as 10Gb XAUI Front panel XFP connector ports as 10Gb XFI 3 ports to the RTM as 10Gb XAUI or 1Gb SERDES One port to the COM Express site as 1 Gb SERDES (reserved for future implementation) The LMP on PCI Bus 2
Note that there is no direct Base-to-Fabric connection on an SCM and no Fabric-to-Fabric connection between the two SCM blades. Figure 8 is a block diagram showing how the Fabric Ethernet switch communicates with the rest of the SCM.
28
Fabric Ethernet switch subsystem
Figure 8. Fabric Ethernet switch

Fabric Ethernet PowerQUICCIII
PCI Bus 2
PCI Bus Reset
GPIO TSEC2
Port 0 Port 1 Port 2 Port 3 Port 4 Port 5 Port 6 Port 7 Port 8 Port 9 Port 10 Port 11 Port 12 Port 13
FCh2 FCh3 FCh4 FCh5 FCh6 FCh7 FCh8 FCh9 FCh10 FCh11 FCh12 FCh13 FCh14 FCh15
Backplane
COM Express
Port 19
RTM
PCI-Express
Port 14
XAUI to XFI XAUI to XFI XAUI to XFI XAUI to XFI
XFP
Front panel
Port 15
XFP
GbE MAC SERDES
Port 16 Port 18 Port 17
XFP
XFP
29
3
Port mapping
Table 5. Fabric Ethernet port mapping MDIO Port # PHY/Port interface address 0 XAUI N/A 1 XAUI N/A 2 XAUI N/A 3 XAUI N/A 4 XAUI N/A 5 XAUI N/A 6 XAUI N/A 7 XAUI N/A 8 XAUI N/A 9 XAUI N/A 10 XAUI N/A 11 XAUI N/A 12 XAUI N/A 13 XAUI N/A 14 XAUI 0x10 15 XAUI 0x11 16 XAUI 0x12 17 XAUI 0x13 18* SERDES N/A 19 XAUI N/A * Reserved for future implementation. Port destination Fabric Channel 2 Fabric Channel 3 Fabric Channel 4 Fabric Channel 5 Fabric Channel 6 Fabric Channel 7 Fabric Channel 8 Fabric Channel 9 Fabric Channel 10 Fabric Channel 11 Fabric Channel 12 Fabric Channel 13 RTM RTM Front Panel XFP 1/1 Front Panel XFP 1/2 Front Panel XFP 1/3 Front Panel XFP 1/4 COM Express RTM IPMI hardware address 0x43 0x44 0x45 0x46 0x47 0x48 0x49 0x4A 0x4B 0x4C 0x4D 0x4E N/A N/A N/A N/A N/A N/A N/A N/A Logical slot 3 4 5 6 7 8 9 10 11 12 13 14 N/A N/A N/A N/A N/A N/A N/A N/A
Configuration
The Fabric interface on the SCM is a managed 10Gb Ethernet switch with the following features: Switch configuration via the PowerQUICCIII PCI bus Transmit (Tx) port disable based on E-Key port state The Fabric Ethernet configuration port is connected to the PowerQUICCIII through PCI bus 2, a 32-bits wide bus operating at 66MHz.
30
LED controller interface subsystem
10 Gigabit Ethernet ports

For the standard SCM configuration, twenty 10Gb Fabric Interface ports are available on the Fabric Ethernet. One is configured for 1Gb operation, connecting to the COM Express site, which is reserved for future implementation. Twelve connect directly to backplane node slots, four provide faceplate XFP connections, and one or three are routed to the RTM. 12 direct XAUI node-slot connections, configurable as 10 Gb XAUI or 1 Gb SERDES. 4 XFP ports on the front faceplate. Each XFP 10Gb connector uses XFI signaling, composed of single transmit and receive 10Gb differential pairs. A XAUI to XFI interface chip converts the signaling from the XAUI signals at each switch port to the XFI signals on the XFP connector. 3 ports routed to the backplane, configurable as 10 Gb XAUI or 1 Gb SERDES.
XFP port LEDs

This table describes the LEDs specific to the Fabric Ethernet. The LEDs are controlled by the Fabric Ethernet switch through a serial LED bus. Table 6. XFP port LEDs Port LED and possible states XFP port LEDs 1 green link/activity status LED Each of the Fabric Ethernet XFP ports on the Green link established front panel or the RTM has two LEDs. Blinking green transmit/receive activity Off no link or port disabled 1 green port status LED Green port is enabled Off port is disabled

Ethernet port status LED indicators are controlled from the Base Ethernet and the Fabric Ethernet switches through serial LED data streams consisting of clock and data signals. One stream comes from the Fabric Ethernet switch and one comes from the Base Ethernet switch. The streams are routed to a complex programmable logic device (CPLD), which acts as a hardware LED controller-driver. The stream is then routed to the Zone 3 connector to support a LED controller on an optional rear I/O module (e.g. RTM). A three-wire serial peripheral interface port (SPI) is available on the LED control CPLD. The SPI supports read operations on the CPLD Revision ID register. Write operations are not supported.
31
3
Front panel LED positions
Figure 9. LED positions on SCM front panel
LMP serial port (LMP SER), RJ45
LMP Fabric Ethernet (LMP ETH), RJ45

1/1 ACT 1/1 STAT
1/2 ACT 1/2 STAT
Fabric 10 GE, XFP Qty 4

1/3 ACT 1/3 STAT 1/4 ACT 1/4 STAT
Base 1 GE, XFP
1/5 ACT 1/5 STAT 1/6 ACT 1/6 STAT 1/7 1/7 1/8 1/8 ACT ACT ACT STAT
Base 1 GE, SFP Qty4
1/9 ACT 1/9 STAT
32
Front panel LED descriptions and states

Table 7. Front panel LEDs Description
Bi-color out of service (OOS)
States Yellow or amber = OOS Off = normal operation Green = power is good Yellow flashing = booting Yellow on = functioning Off = inactive Blue = ready for hot swap Green = link Off = no link Yellow blink = activity Off = no link or port disabled Green = link is good Off = no link or port disabled Green = enabled Green blink = activity Off = no link or port disabled Green = link is good Off = no link or port disabled Green = enabled Green blink = activity Off = no link or port disabled Green = link is good Green blink = activity Off = no link or port disabled Green = enabled Off = port disabled
LED1 controlled by GPO signals on the IPMC.

Green PowerGood

Yellow module state

Blue hot swap ready
Blue LED controlled by an IPMC output signal.

2 Green COM Express and LMP serial console port Links
LED part of the RJ-45 connector - indicates PHY port activity. Note: The COM Express connections are reserved for future implementation.
2 Yellow COM Express and LMP serial console port Activity
LED part of the RJ-45 connector - indicates PHY port activity. Note: The COM Express connections are reserved for future implementation.
4 green Fabric Ethernet XFP Port Status
LEDs controlled by the LMP through the 10 G PHY LED signals.

4 green Fabric Ethernet XFP Link / Activity

1 green Base Ethernet XFP Port Status

1 green Base Ethernet XFP Link / Activity

4 green Base Ethernet SFP Link / Activity
LEDs part of the RJ-45 connector indicates GE PHY port status, driven through the PHY LED signals.
4 green Base Ethernet SFP Port Status
LEDs part of the RJ-45 connector indicates GE PHY port status, driven through the PHY LED signals.
33
3
LED controller Base assignments
Table 8. LED controller Base assignments Eth switch Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Bit stream position 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Faceplate port label 1/6 1/6 1/6 1/6 1/7 1/7 1/7 1/7 1/8 1/8 1/8 1/8 1/9 1/9 1/9 1/9 1/5 1/5 1/5 1/5 Switch port GE18 GE18 GE18 GE18 GE19 GE19 GE19 GE19 GE20 GE20 GE20 GE20 GE21 GE21 GE21 GE21 XG0 XG0 XG0 XG0 Indicator label Act Act Stat Stat Act Act Stat Stat Act Act Stat Stat Act Act Stat Stat Act Act Stat Stat Position color Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow
34
LED controller Fabric assignments

Table 9. LED controller Fabric assignments Eth switch Fabric Fabric Fabric Fabric Fabric Fabric Fabric Fabric Fabric Fabric Fabric Fabric Fabric Fabric Fabric Fabric Bit stream position 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Faceplate port label 1/1 1/1 1/1 1/1 1/2 1/2 1/2 1/2 1/3 1/3 1/3 1/3 1/4 1/4 1/4 1/4 Switch port XG15 XG15 XG15 XG15 XG16 XG16 XG16 XG16 XG17 XG17 XG17 XG17 XG18 XG18 XG18 XG18 Indicator label Act Act Stat Stat Act Act Stat Stat Act Act Stat Stat Act Act Stat Stat Position color Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow

The NTS is the centralized system timing function provided by the SCM. The timing function is based on the synchronization clock interface described in the AdvancedTCA Base Specification. The list below and Figure 10 provide an overview of the operation of the NTS. Provides a Synchronous Equipment Timing Source (SETS) for all shelf node blades Accepts system timing up to four (redundant) Building Integrated Timing Source (BITS) frames, and generates framed synchronization signals for shelf-to-shelf chaining Accepts and generates Synchronization Supply Unit (SSU) frames per ITU-T Recommendation G.812 Accepts timing references from backplane node module synchronization clocks (CLK3-A and CLK3B) Generates 8 KHz and 19.44 MHz synchronization clocks on backplane (CLK1-A/B and CLK2-A/B, respectively) Generates 8 KHz, 2.048 MHz, 1.544 Mhz, and 19.44 Mhz synchronization clocks for the RTM Can provide automatic or manual hitless input switchover on loss of input, if configured using input priority levels Provides automatic hitless protection of the system clock when operating with a redundant SCM
35
3
Figure 10. Network timing subsystem
EXT CLK IN 4 EXT CLK IN 3 EXT CLK IN 2 EXT CLK IN 1 TO2 EXT CLK OUT DS26502 Framer / Deframer DS26502 Framer / Deframer DS26502 Framer / Deframer CLK1 -A TO4 12.8MHz TCXO REFCLK 12.8MHz OCXO 12.8MHz XO I3 I4 CLK3 -A CLK3 -B TO11 TO1 SYNC2K I11 8 KHz CLK1 -B CLK2 -A TO5 19.44 MHz 2 KHz OUT 6.48 MHZ OUT 2 KHz IN 6.48 MHz IN CLK2 -B UPDCH3Tx UPDCH4Tx UPDCH3Rx UPDCH4Rx BITS/SSU-3 REF-3 DS26502 Framer / Deframer BITS/SSU-4 REF-4
I5 I6 I9 I10
BITS/SSU-2 REF-2 BITS/SSU-1 REF-1
TO3
19.44 MHz
PowerQUICCIII SPI GPIO

SPI CSB
RTM
(Illustration)
SCM
The NTS can select reference timing sources from the synchronization bus, the RTM, or the update channel and be configured with priority levels for each input. By monitoring each input, the NTS can automatically switch between valid sources based on those priorities. Information on configuring the different timing sources is covered in Network timing subsystem (NTS) software on page 83.
Synchronization clock and update channel buffer enables

PowerQUICCIII can enable or disable the clock buffers to the synchronization clock and update channel outputs. Permission to drive these signals comes from the Shelf Manager through the IPMC, as part of E-Keying.
36
Rear transition module

Three connectors in the Zone 3 area provide connectivity to the rear transition module (RTM). Electrical connections between the SCM and the associated RTM include: Two to four 1 Gb SGMII/SERDES ports from the Base interface Two 10 Gb XAUI port from Base interface One to three 10Gb XAUI ports from Fabric switch Four external reference clock inputs Three external system clock outputs I2C interface to IPMC for ID, temperature, temperature sensors, and control management I2C interface to the LMP for ID SPI to the LMP PCI Express interface from COM Express site (the site is reserved for future implementation) Switched 12 V from payload supply Current limited 3.3 V from IPMC supply RTM hot-swap switch closure signal Hot-swap RTM LED output from the IPMC
A shelf peripheral module (SPM) acts as the shelf interface when installed in a 14-slot RadiSys shelves. The SPM slides into the rear of the platform and connects to the SCM in the corresponding front module slot (i.e., hub slot) using Zone 3 connectors. SPMs are optional for SCMs installed in RadiSys 6-slot shelves or non-RadiSys shelves. Figure 11 shows a block diagram example of an SPM designed for the 14-slot RadiSys shelf.
37
3
Figure 11. SPM block diagram
Local Power Conversion
Zone 3 Conn
Buf Buf
ID ID & Temp
SFP*
SFP
SFP 1Gb SD SFP
SPI ClkIn RefClk 1Gb SD
SFP
SFP
SFP
SFP
BITS/ SSU Clk BITS/ SSU Clk BITS/ SSU Clk BITS/ SSU Clk
Mag I/F Mag I/F Mag I/F Mag I/F
RJ -45
RJ -45
RJ -45
RJ -45
SPM (Illustration)
* Reserved for future implementation.
RTM hot-swap support buffers

In order to protect the RTM interface signals from disruption during insertion or removal of the RTM, buffers have been added to many of the RTM signals. These buffers are controlled through RTM_BUF_ENA~ on GPIO PC12 of the LMP. The buffer should be disabled whenever the RTM is not installed, or is being changed.
38
RTM Interface Signals

Space has been allocated in the Zone 3 connectors for signals passing between the SCM and the RTM, including data path, configuration, and hot-swap signals. This represents a superset of the signals that may be used for various combinations of the SCM and the RTM. Table 10. RTM interface signals Signal name +12V_RTM +3.3V_IPMC_LTD RTM_EJECT_IN~ RTM_Present~ RTM_HOTSWAP~ RTM_ENA~ RTM_INTR~ IPMC_I2C_CLK IPMC_I2C_DAT PQ_12C_CLK PQ_12C_DAT SPI_RTM_SEL~ Direction from SCM Pins Description to RTM Out 6 Power source for RTM circuits, 3 Amp limit. This is current limited, and can be turned on and off by the IPMC. Out 1 Power source for RTM hot-swap control circuits, 0.5 Amp limit. This is current limited, and can be turned on and off by the IPMC. In 1 Signal is grounded on RTM when ejector handle is closed (RTM fully seated). Pull-up to +3.3V_IPMI on SCM. In 1 Indicates RTM is connected. This pin should be grounded on the RTM. Pulled up to +3.3V_IPMI on SCM. Out 1 Drive signal for RTM hot-swap indicator LED. 3.3V Signal goes low to turn on LED. Pull up to +3.3V_IPMC on RTM. Out 1 Low-true Enable signal for RTM on-board controller. Low enables, high disables (resets). Pull up to +3.3V_IPMC on RTM. In 1 Low-true interrupt signal from the RTM to the IPMC. Pulled up to +3.3V_IPMI on SCM. I/O 1 IPMC control bus clock. Pulled up to +3.3V_IPMI on SCM. I/O 1 IPMC control bus data. Pulled up to +3.3V_IPMI on SCM. I/O 1 PowerQUICCIII control bus clock. Pulled up to Payload +3.3V on SCM. I/O 1 PowerQUICCIII control bus data. Pulled up to Payload +3.3V on SCM Out 1 SPI target select, low-true. RTM SPI bus slave devices will use this select along with address bits to select a certain device. 3.3V signal from LMP goes low to select target device. Pull up to Payload +3.3V on RTM. Out 3 SPI bus target device address. The MSB indicates SMC3 (0) or RTM3 (1) as the target board. 3.3V signal from LMP. Out 1 Base switch LED control stream clock signal. 3.3 V signal from LMP controls RTM faceplate LEDs. Out 1 Base switch LED control stream data signal. 3.3 V signal from LMP controls RTM faceplate LEDs. Out 1 Fabric switch LED control stream clock signal. 3.3 V signal from LMP controls RTM faceplate LEDs. Out 1 Fabric switch LED control stream data signal. 3.3 V signal from LMP controls RTM faceplate LEDs. Out 1 Reset output to RTM peripherals (active low). Pull up to +3.3V_IPMC on RTM. Out 1 SPI data transfer clock. 3.3 V signal from LMP. Out 1 SPI data out (from LMP to target). 3.3V signal from LMP. In 1 SPI data in (3.3 V signal from RTM target to LMP). I/O 1 Base 1Gb interface MDIO clock. 3.3 V signal from Base switch. 39
SPI_SLVA[0..2] BLED_CLK BLED_DAT FLED_CLK FLED_DAT RTM_RST~ SPI_CLK SPI_SDO SPI_SDI BASE_MDC
3
Table 10. RTM interface signals (continued) Signal name BASE_MDIO BASE_XMDC BASE_XMDIO FABRIC_XMDC FABRIC_XMDIO EXTCLK1_IN+/EXTCLK2_IN+/EXTCLK3_IN+/EXTCLK4_IN+/EXTCLK_OUT+/RTM_LMP_INTR5~ RTM_LMP_INTR6~ NTS_8K_OUT+/NTS_19M_OUT+/PCI_E_TX+/PCI_E_RX+/BASE_GE_TX[0..1]+/BASE_GE_RX[0..1]+/BASE_10G_TX[0..3]+/BASE_10G_RX[0..3]+/FABRIC_10G_TX[0..3]+/FABRIC_10G_RX[0..3]+/Direction from SCM to RTM Out I/O Out Out I/O In In In In Out In In Out Out Out In Out In Out In Out In Pins Description 1 1 1 1 1 2 2 2 2 2 1 1 2 2 2 2 164 16 16 16 32 32 Base 1Gb interface MDIO data. 3.3 V bidirectional signal from Base switch. Base 10 Gb interface MDIO clock. 2.5 V signal from Base switch. Base 10 Gb interface MDIO clock. 2.5 V bidirectional signal from Base switch. Fabric 10 Gb interface MDIO clock. 2.5 V signal from Base switch. Fabric 10 Gb interface MDIO data. 2.5 V bidirectional signal from Base switch. Recovered clock 1 in, LVDS differential Recovered clock 2 in, LVDS differential Recovered clock 3 in, LVDS differential Recovered clock 4 in, LVDS differential Clock out, LVDS differential Interrupt (low-true) from the RTM to the LMP. Pulled up on SCM. Interrupt (low-true) from the RTM to the LMP. Pulled up on SCM. 8KHz external clock out, LVDS differential 19.44MHz external clock out, LVDS differential One PCI Express Tx lane, differential One PCI Express Rx lane, differential Base interface 1Gb Tx port, differential. Base interface SERDES signals. The number of Base ports available depends on SCM component loading (stuffing). Base interface 1Gb Rx port, differential. Base interface SERDES signals. The number of Base ports available depends on SCM component loading (stuffing). Base interface 10 Gb Tx port, differential. Base interface XAUI signals. The number of Base ports available depends on SCM component loading (stuffing). Base interface 10 Gb Rx port, differential. Base interface XAUI signals. The number of Base ports available depends on SCM component loading (stuffing). Fabric interface 10 Gb Tx port, differential. Fabric interface XAUI signals. The number of Fabric ports available depends on SCM component loading (stuffing). Fabric interface 10 Gb Rx port, differential. Fabric interface XAUI signals. The number of Fabric ports available depends on SCM component loading (stuffing).
40
Hardware management subsystem
Hardware management subsystem

The function and operation of the SCM hardware management subsystem is controlled by the Intelligent Platform Management Controller (IPMC). The hardware management subsystem is the collection of IPMCs and sensors on the installed modules and field replaceable units (FRUs), and the communication between these devices and the Shelf Manager. The SCMs IPMC manages the commands and the data portion of this subsystem. The IPMC has the following features: Dual I2C interface to backplane intelligent platform management bus (IPMB) with programmable pullups Serial interfaces to LMP I2C interfaces to I/O expander, field-programmable gate array (FPGA), temperature sensors, voltage sensors, and RTM System event log (SEL) capabilities Remote executable flash and micro-controller software update support Electronic keying (E-Keying) support for Base interface, Fabric interface, synchronization clocks, and update channels Remote micro-controller software update support
Figure 12 shows how the IPMC communicates with other subsystems like the LMP and the RTM.
41
3
Figure 12. IPMC control interface to payload and RTM
+12V RTM_+12V_ENA~ +3.3V_IPMI +12V_RTM
RTM:+12V_FAIL~
RTM_+3.3V_ENA~ RTM:+3.3V_FAIL~ RTM_INTR~ RTM_PRESENT~ RTM_EJECT_IN~ RTM_RST~ RTM:HOTSWAP RTM_ENA~ IPMC:I2C_SCL IPMC:I2C_SDA
Power switch Power switch
+3.3V_RTM
+12V +3.3V COME:PWR_ENA~ COME:PWR_FAIL~ COM_E_PRST~ COM_E_TYPE2 COM_E_TYPE1 COM_E_TYPE0 COME:RESET# CECOM1_TX CECOM1_RX CE_UC_ENA CE_GBE1 +12V_CE +3.3V_CE
RTM
Power switch
1 GbE Ena
COM Express*
Update channel
PL-RST#0 FLSH_SEL PORESET# BUF_MRESET# SRESET# IRQ 0 COM0_TX COM0_RX IPMC:MSG_FLAG_STB
GPIO5
Flip Flop
IPMC:MSG_FLAG IPMC:MSG_FLAG_CLR
-48 V
+3.3V_IPMC PAYLOAD_PWR#
LMP
48 V DC/DC
IPMC
*Reserved for future implementation
+12V
42
Power supplies
Non-volatile random access memory

The non-volatile random access memory (NVRAM) in the IPMC on the SCM is a 64 KB device. The NVRAM stores the following types of information about the field replaceable units (FRUs): Serial number Part number Manufacturer Date and time of manufacture Product name FRU capabilities Point to point connectivity records for modules that plug into the backplane, which is used for E-Keying.
The SCM acting as an active Shelf Manager, retrieves this information from the IPMC and stores the information. To provide redundancy, the active Shelf Manager also sends the FRU information to the inactive Shelf Manager, which also stores the information. This information is recorded in the system event log (SEL) to monitor system events.
IPMC watchdog timer

The IPMC entity includes a hardware watchdog timer. Once the watchdog is activated, the IPMC must strobe it in order to keep it from timing out. If a firmware or a hardware problem on the IPMC causes it to stop strobing the watchdog timer, the IPMC is automatically isolated from the IPMB. This isolation of the IPMC keeps the IPMI buses functional for the remaining IPMC devices. A two-pin header can be used to disable the watchdog timer during debug.
Power supplies
Information on the SCMs power consumption and emissions is provided in Appendix A: Specifications on page 125.
Power input module (PIM)

The SCM uses a power input module (PIM), which provides input conditioning of the -48V backplane power feeds. This module is between the input protection fuses and the main 48V to 12V DC/DC converter. Features of the PIM include: Inrush current limit protection Integrated filter designed to meet CISPR class B EMI limits Isolated auxiliary 3.3V power for IPMI circuit ORing field effect transistors (FETs) on A and B -48V feeds A/B feed loss alarm Hot-swap control Glitch hold-up circuit based on external capacitor and 72V charging circuit Input under voltage and over voltage protection Over current and thermal protection
43
3
-48V to 12V primary voltage conversion
A DC/DC converter is used to generate 12V for the SCM. This power supply provides up to 18 Amps of current to power the SCM payload and RTM. The RTM power buses are switched from this supply through integrated FET switches, controlled by the IPMC. Output voltage tolerance: 5%. The DC/DC is switched on and off through an optically isolated control signal from the IPMC.
Power converters
The following non-isolated supplies are generated from the 12V intermediate supply by switching regulators: 5.0 V DC for components such as the hard drive and XFP modules. 3.3 V DC for components such as the hard drive, the XFP modules, and the NTS block. 2.5 V DC for use components such as the PowerQUICCIII, the DRAM, the 1Gb SERDES, and the FE switch. 1.8 V DC for components such as the XFP modules, the XAUI-to-XFI interfaces, the PCI Express 1Gb controllers. 1.25 V DDR for the DDR DRAM module. 1.25 V DC for the Base and the Fabric Ethernet switches. 1.2 V DC for the PowerQUICCIII, Gigabit Ethernet Quad PHYs, and the XAUI-XFI interfaces. 1.1 V DC for use by the PCI Express 1Gb controllers. 1.0 V DC for the 10 Gb Ethernet switch.
44
PHYSICAL INTERFACES
This chapter describes the physical interfaces for the SCM.
Mechanical specification
The SCM dimensions conform to the PICMG AdvancedTCA 3.0 R2.0 Specification. Table 11. Base board layout
COM Express site* Clock/ calendar SAS/SATA site*
Console ports DRAM SODIMM
Base and Fabric XFP sockets
SFP sockets
Fabric Ethernet switch
Base Ethernet switch
DC/DC Converter PowerQUICCIII
Power input module
* Reserved for future implementation
45
Physical Interfaces
Thermal design
Heat sinks are used in the SCMs thermal design. Extruded and crosscut aluminum heat sinks with thermal interface pads are used to cool the Base Ethernet switch, the Fabric Ethernet switch, the Base Ethernet PHYs, and the PowerQUICCIII processor. The heat sinks for the XFPs are a pin fin configuration and are compliant with the specifications defined in the XFP MSA.
Front panel interfaces

Each switch includes a sheet metal front panel that will serve as an EMI/RFI barrier and complies with PICMG 3.0 Revision 2.0. The front panel provides access to the LMP, the Base Ethernet switch, and the Fabric Ethernet switch. The front panel includes: RJ-45 connector for the RS-232 serial port (labeled LMP SER) from the LMP. RJ-45 connector (with integrated LEDs) for the LMP 10/100Base-T copper port (labeled LMP ETH). Two LEDs on each connector indicate link status and activity. Four 10 Gb XFP sockets with 10Gb XFI-to-XAUI interfaces to the Fabric Ethernet switch. XFP module loading is optional. Two LEDs next to connector indicate link, activity, and port status. One XFP socket with a 10Gb XFI-to-XAUI interface to the Base Ethernet switch. XFP module loading is optional. Two LEDs next to connector indicate link, activity, and port status Four SFP sockets with 1Gb SERDES interfaces (and integrated status LEDs) to the Base Ethernet switch. SFP loading is optional. Two LEDs on each connector indicate link, activity, and port status. Three connection types on the front panel are reserved for future implementation of the COM Express: one USB port, one RJ-45 connector for a RS-232 serial port, and one RJ-45 connector for a 10/100/1000 Base-T copper interface. Reset button (recessed to prevent accidental actuation) Blue hot-swap LED (ATCA "Blue LED") Red or amber Out Of Service LED (ATCA "LED 1") Green Power Good LED (ATCA "LED 2") Amber SCM Status LED (ATCA "LED 3") Five green/amber 10Gb Ethernet port Link status LEDs Five green/amber 10Gb Ethernet port Activity status LEDs Hot-swap ejector latches
OOS COM-E ETH PWR ACT LINK XFP 1/1 PORT LINK XFP 1/2 PORT LINK XFP 1/3 PORT LINK XFP 1/4 PORT LINK PORT
LINK PORT LINK PORT
ATCA - 2210
COM-E USB COM-E SER
LMP serial port
LMP SER LMP ETH
LMP Ethernet port
Fabric Ethernet XFPs
FABRIC BASE
Base Ethernet XFP
XFP 1/5 SFP 1/6 SFP 1/7
For descriptions of the front panel LEDs and their various states, see LED controller interface subsystem on page 31.
H/S
RESET
Base Ethernet SFPs

SFP 1/8 SFP 1/9
AdvancedTCA
LINK PORT LINK PORT
Reset
46
Front panel interfaces
Push Buttons/Switches
The SCM includes a reset button and a hot-swap switch. The recessed reset push button only resets the payload section of the SCM when pressed. The hot-swap switch closes when the ejector handle is fully latched.
LMP serial console port

An RJ-45 connector provides serial access to the console port of the LMP. This may be used as a console port for setup and management of applications running on the PowerQUICCIII processor chip. Table 12. LMP serial console port connector pinouts Pin 1 2 3 4 Description LMP_RTS NC LMP_TXD GND Pin 5 6 7 8 Description GND LMP_RXD NC LMP_CTS
LMP Ethernet maintenance port

An RJ-45 connector is used to provide a maintenance port for the SCM. This Ethernet port interfaces to the PowerQUICCIII processor chip. This connector includes the isolation magnetics and LEDs for Link (green) and Activity (yellow). Table 13. LMP maintenance port connector pinouts Pin 1 2 3 4 Description TX+ TXRX+ Unused (DC+)* Pin 5 6 7 8 Description Unused (DC-)* RXUnused (DD+)* Unused (DD-)*
*Unused pairs are terminated.
SFP Base Ethernet interface ports

The SFP Base Ethernet connectors are used to provide front panel access to the Base Ethernet switch. Each connector is inside a metal cage, which is directly coupled to a port of the Base Ethernet switch.
47
Physical Interfaces
See the ATCA-2210 SCM Installation Guide for the list of SFP configurations and products that have been tested and confirmed by RadiSys to work with the SCM. The connector pinouts are included in the following table. Table 14. SFP Base Ethernet connector pinouts Pin 1 2 3 4 5 6 7 8 9 10 11 Description GND TXFAULT TXDIS MODDEF2 MODDEF1 MODDEF0 RATE SEL LOS GND GND GND Pin 11 12 13 14 15 16 17 18 19 20 20 Description GND RDRD+ GND +3,3V +3.3V GND TD+ TDGND GND
XFP Base Ethernet and Fabric Ethernet ports

The XFP connectors are used to provide front panel access to the 10Gb ports of the Base and the Fabric Ethernet switches. Each connector is coupled to a switch port through a XAUI-to-XFI interface. The connector is inside a metal cage, which is connected to the SCM. See the ATCA-2210 SCM Installation Guide for the list of XFP configurations and products that have been tested and confirmed by RadiSys to work with the SCM. The connector pinouts are included in the following table. Table 15. XFP Base Ethernet & Fabric Ethernet connector pinouts Pin 1 2 3 4 5 6 7 8 9 10 Description GND NC MOD_DESEL INTERRUPT TX_DIS +5.0V GND +3.3V +3.3V SCL Pin 11 12 13 14 15 16 17 18 19 20 Description SDA MOD_ABS MOD_NR RX_LOS GND GND RDRD+ GND +1.8V Pin 21 22 23 24 25 26 27 28 29 30 Description P_DOWN +1.8V GND REFCLK+ REFCLKGND GND TDTD+ GND
48
Backplane interfaces
The following SCM interfaces are available through the backplane: Twelve 10/100/1000Base-T Base interface node ports One inter-LMP link to a redundant SCM through the update channel One Base Ethernet switch-to-switch link port through Base channel 2 Dual 10/100Base-T Base interface port to one or two external Shelf Managers through Base channel 1 Twelve 10 Gb XAUI Fabric interface node ports Six synchronization channel clock input/outputs to the NTS Master/slave link to NTS module clock through the update channel Two -48V power rails IPMB interface (two I 2C ports) with programmable pull-ups Eight hardware address lines, seven address / one parity
Update channels
Update channels are the backplane connections that exist between a pair of SCMs operating on a redundant basis. Application software can use update channels for redundancy interlock. If you configure a pair of SCMs to use the update channel for redundancy support, you must install the SCMs into slots linked by an update channel. The update channel for SCMs exists between physical slots 1 and 2 in RadiSys 6-slot shelf and 7 and 8 in a RadiSys 14-slot shelf. For more information on update channels and the slots that link redundant modules, see the appropriate Platform Hardware Reference manual (SYS-6000/6010 or SYS-6006) for the platform you are using.
Power connector signals (P10)

The Zone 1 contact assignments for P10 are included in the following table. Table 16. Zone 1 contact assignments P10 Contact number 1 2 3 4 5 6 7 8 9 10 11 12 13 Designation Reserved Reserved Reserved Reserved HA0 HA1 HA2 HA3 HA4 HA5 HA6 HA7/P SCL_A Description Reserved Reserved Reserved Reserved Hardware address bit 0 Hardware address bit 1 Hardware address bit 2 Hardware address bit 3 Hardware address bit 4 Hardware address bit 5 Hardware address bit 6 Hardware address bit 7 (odd parity bit) IPMB clock, port A Mating sequence N/A N/A N/A N/A Third Third Third Third Third Third Third Third Third
49
Physical Interfaces
Table 16. Zone 1 contact assignments P10 (continued) Contact number 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Designation SDA_A SCL_B SDA_B MT1_TIP MT2_TIP -RING_A -RING_A MT1_RING MT2_RING RRTN_A RRTN_B SHELF_GND LOGIC_GND ENABLE_B VRTN_A VRTN_B EARLY_A EARLY_B ENABLE_A -48V_A -48V_B Unused pin on SCM Description IPMB data, port A IPMB clock, port B IPMB data, port B Metallic test # 1 tip Metallic test # 2 tip Ringing generator A negative ringing Ringing generator B negative ringing Metallic test #1ringing Metallic test #2 ringing Ringing generator A return Ringing generator B return Shelf ground and safety ground Logic ground (reference and return for front module-tofront module logic signals) Enable B (short pin for power sequencing, tied to VRTN_B on backplane) Voltage return A Voltage return B -48 volt early (pre-charge) A -48 volt early (pre-charge) B Enable A (short pin for power sequencing, tied to VRTN_A on backplane) -48 volt A, uses ENABLE_A to enable converters -48 volt B, uses ENABLE_B to enable converters Mating sequence Third Third Third Third Third Third First Third Third Third Third First Fourth First First First First First Fourth Second Third
50
Backplane connector J20 signals

Table 17. Backplane connector J20 signals Row # Interface designation Clks Update channel and clks ab cd CLK2A+ CLK3A+ Tx3(UP)+ Tx1(UP)+ Tx3[15]+ Tx1[15]+ Tx3[14]+ Tx1[14]+ Tx3[13]+ Tx1[13]+ ef CLK2ACLK3ATx3(UP)Tx1(UP)Tx3[15]Tx1[15]Tx3[14]Tx1[14]Tx3[13]Tx1[13]CLK2B+ CLK3B+ Rx3(UP)+ Rx1(UP)+ Rx3[15]+ Rx1[15]+ Rx3[14]+ Rx1[14]+ Rx3[13]+ Rx1[13]+ gh CLK2BCLK3BRx3(UP)Rx1(UP)Rx3[15]Rx1[15]Rx3[14]Rx1[14]Rx3[13]Rx1[13]-
1 CLK1A+ CLK1ACLK1B+ CLK1B2 Tx4(UP)+ Tx4(UP)- Rx4(UP)+ Rx4(UP)3 Tx2(UP)+ Tx2(UP)- Rx2(UP)+ Rx2(UP)4 Tx0(UP)+ Tx0(UP)- Rx0(UP)+ Rx0(UP)5 Fabric channel 15 Tx2[15]+ Tx2[15]Rx2[15]+ Rx2[15]6 Tx0[15]+ Tx0[15]Rx0[15]+ Rx0[15]7 Fabric channel 14 Tx2[14]+ Tx2[14]Rx2[14]+ Rx2[14]8 Tx0[14]+ Tx0[14]Rx0[14]+ Rx0[14]9 Fabric channel 13 Tx2[13]+ Tx2[13]Rx2[13]+ Rx2[13]10 Tx0[13]+ Tx0[13]Rx0[13]+ Rx0[13]Note: Each differential pair has an individual L-shaped ground contact (not shown).

Table 18. Backplane connector J21 signals Row # Interface designation Fabric channel 12 ab cd Tx3[12]+ Tx1[12]+ Tx3[11]+ Tx1[11]+ Tx3[10]+ Tx1[10]+ Tx3[9]+ Tx1[9]+ Tx3[8]+ Tx1[8]+ ef Tx3[12]Tx1[12]Tx3[11]Tx1[11]Tx3[10]Tx1[10]Tx3[9]Tx1[9]Tx3[8]Tx1[8]Rx3[12]+ Rx1[12]+ Rx3[11]+ Rx1[11]+ Rx3[10]+ Rx1[10]+ Rx3[9]+ Rx1[9]+ Rx3[8]+ Rx1[8]+ gh Rx3[12]Rx1[12]Rx3[11]Rx1[11]Rx3[10]Rx1[10]Rx3[9]Rx1[9]Rx3[8]Rx1[8]-
1 Tx2[12]+ Tx2[12]Rx2[12]+ Rx2[12]2 Tx0[12]+ Tx0[12]Rx0[12]+ Rx0[12]3 Fabric channel 11 Tx2[11]+ Tx2[11]Rx2[11]+ Rx2[11]4 Tx0[11]+ Tx0[11]Rx0[11]+ Rx0[11]5 Fabric channel 10 Tx2[10]+ Tx2[10]Rx2[10]+ Rx2[10]6 Tx0[10]+ Tx0[10]Rx0[10]+ Rx0[10]7 Fabric channel 9 Tx2[9]+ Tx2[9]Rx2[9]+ Rx2[9]8 Tx0[9]+ Tx0[9]Rx0[9]+ Rx0[9]9 Fabric channel 8 Tx2[8]+ Tx2[8]Rx2[8]+ Rx2[8]10 Tx0[8]+ Tx0[8]Rx0[8]+ Rx0[8]Note: Each differential pair has an individual L-shaped ground contact (not shown).
51
Physical Interfaces

Table 19. Backplane connector J22 signals Row # Interface designation Fabric channel 7 ab cd Tx3[7]+ Tx1[7]+ Tx3[6]+ Tx1[6]+ Tx3[5]+ Tx1[5]+ Tx3[4]+ Tx1[4]+ Tx3[3]+ Tx1[3]+ ef Tx3[7]Tx1[7]Tx3[6]Tx1[6]Tx3[5]Tx1[5]Tx3[4]Tx1[4]Tx3[3]Tx1[3]Rx3[7]+ Rx1[7]+ Rx3[6]+ Rx1[6]+ Rx3[5]+ Rx1[5]+ Rx3[4]+ Rx1[4]+ Rx3[3]+ Rx1[3]+ gh Rx3[7]Rx1[7]Rx3[6]Rx1[6]Rx3[5]Rx1[5]Rx3[4]Rx1[4]Rx3[3]Rx1[3]-
1 Tx2[7]+ Tx2[7]Rx2[7]+ Rx2[7]2 Tx0[7]+ Tx0[7]Rx0[7]+ Rx0[7]3 Fabric channel 8 Tx2[6]+ Tx2[6]Rx2[6]+ Rx2[6]4 Tx0[6]+ Tx0[6]Rx0[6]+ Rx0[6]5 Fabric channel 5 Tx2[5]+ Tx2[5]Rx2[5]+ Rx2[5]6 Tx0[5]+ Tx0[5]Rx0[5]+ Rx0[5]7 Fabric channel 4 Tx2[4]+ Tx2[4]Rx2[4]+ Rx2[4]8 Tx0[4]+ Tx0[4]Rx0[4]+ Rx0[4]9 Fabric channel 3 Tx2[3]+ Tx2[3]Rx2[3]+ Rx2[3]10 Tx0[3]+ Tx0[3]Rx0[3]+ Rx0[3]Note: Each differential pair has an individual L-shaped ground contact (not shown).

Table 20. Backplane connector J23 signals Row # 1 2 3 4 5 Interface designation Fabric channel 2 ab cd ef Tx3[2]Tx1[2]Tx3[1]Tx1[1]BI_SHMC _Tx1BI_DC2BI_DC3BI_DC4BI_DC5BI_DC6Rx3[2]+ Rx1[2]+ Rx3[1]+ Rx1[1]+ BI_SHMC _Rx1+ BI_DD2+ BI_DD3+ BI_DD4+ BI_DD5+ BI_DD6+ gh Rx3[2]Rx1[2]Rx3[1]Rx1[1]BI_SHMC _Rx1BI_DD2BI_DD3BI_DD4BI_DD5BI_DD6-
Tx2[2]+ Tx2[2]Rx2[2]+ Rx2[2]Tx3[2]+ Tx0[2]+ Tx0[2]Rx0[2]+ Rx0[2]Tx1[2]+ Fabric channel 1 Tx2[1]+ Tx2[1]Rx2[1]+ Rx2[1]Tx3[1]+ Tx0[1]+ Tx0[1]Rx0[1]+ Rx0[1]Tx1[1]+ ShMC Port BI_SHMC BI_SHMC BI_SHMC BI_SHMC BI_SHMC _Tx0+ _Tx0_Rx0+ _Rx0_Tx1+ 6 Base channel 2 BI_DA2+ BI_DA2BI_DB2+ BI_DB2BI_DC2+ 7 Base channel 3 BI_DA3+ BI_DA3BI_DB3+ BI_DB3BI_DC3+ 8 Base channel 4 BI_DA4+ BI_DA4BI_DB4+ BI_DB4BI_DC4+ 9 Base channel 5 BI_DA5+ BI_DA5BI_DB5+ BI_DB5BI_DC5+ 10 Base channel 6 BI_DA6+ BI_DA6BI_DB6+ BI_DB6BI_DC6+ Note: Each differential pair has an individual L-shaped ground contact (not shown). Unused pin on SCM
52

Table 21. Backplane connector J24 signals Row # Interface ab cd ef designation 1 Base channel 7 BI_DA7+ BI_DA7BI_DB7+ BI_DB7BI_DC7+ BI_DC72 Base channel 8 BI_DA8+ BI_DA8BI_DB8+ BI_DB8BI_DC8+ BI_DC83 Base channel 9 BI_DA9+ BI_DA9BI_DB9+ BI_DB9BI_DC9+ BI_DC94 Base channel 10 BI_DA10+ BI_DA10- BI_DB10+ BI_DB10- BI_DC10+ BI_DC105 Base channel 11 BI_DA11+ BI_DA11- BI_DB11+ BI_DB11- BI_DC11+ BI_DC116 Base channel 12 BI_DA12+ BI_DA12- BI_DB12+ BI_DB12- BI_DC12+ BI_DC127 Base channel 13 BI_DA13+ BI_DA13- BI_DB13+ BI_DB13- BI_DC13+ BI_DC138 Base channel 14 BI_DA14+ BI_DA14- BI_DB14+ BI_DB14- BI_DC14+ BI_DC149 Base channel 15 BI_DA15+ BI_DA15- BI_DB15+ BI_DB15- BI_DC15+ BI_DC1510 Base channel 16 BI_DA16+ BI_DA16- BI_DB16+ BI_DB16- BI_DC16+ BI_DC16Note: Each differential pair has an individual L-shaped ground contact (not shown). gh BI_DD7+ BI_DD8+ BI_DD9+ BI_DD10+ BI_DD11+ BI_DD12+ BI_DD13+ BI_DD14+ BI_DD15+ BI_DD16+ BI_DD7BI_DD8BI_DD9BI_DD10BI_DD11BI_DD12BI_DD13BI_DD14BI_DD15BI_DD16-
Zone 3 connector J30 signals

Table 22. Zone 3 connector J30 signals Row # 1 2 3 ab +12V_RTM +12V_RTM +12V_RTM +12V_RTM +12V_RTM +12V_RTM cd +3.3V_IPMC _LTD RTM_INTR~ RTM_EJECT IPMC_I2C_ CLK BLED_CLK ef RTM_ PRESENT~ IPMC_I2C_ DAT BLED_DAT BASE_MDIO EXTCLK2_INEXTCLK4_INFAB:XE RTMxTX2FAB:XE RTMxRX2BASE:GE RTM7TXBASE:GE RTM5TXRTM_HOT SWAP~ PQ_I2C_ CLK FLED_CLK gh RTM_ENA~ PQ_I2C_ DAT FLED_DAT
SPI_RTM_ SLVA2 SLVA1 SLVA0 SEL~ 4 RTM_RST~ SPI_CLK SPI_SDO SPI_SDI BASE_MDC 5 EXTCLK_ EXTCLK_ RTM_LMP_ RTM_LMP_ EXTCLK2_IN+ OUTOUT+ IRQ5 IRQ6 6 NTS_8K_ NTS_8K_ NTS_19M_ NTS_19M_ EXTCLK4_IN+ OUT+ OUTOUT+ OUT7 FAB:XE FAB:XE FAB:XE FAB:XE FAB:XE RTMxTX0+ RTMxTX0RTMxTX1+ RTMxTX1RTMxTX2+ 8 FAB:XE FAB:XE FAB:XE FAB:XE FAB:XE RTMxRX0+ RTMxRX0RTMxRX1+ RTMxRX1RTMxRX2+ 9 BASE:GE BASE:GE BASE:GE BASE:GE BASE:GE RTM6TX+ RTM6TXRTM6RX+ RTM6RXRTM7TX+ 10 BASE:GE BASE:GE BASE:GE BASE:GE BASE:GE RTM4TX+ RTM4TXRTM4RX+ RTM4RXRTM5TX+ Note: Each differential pair has an individual L-shaped ground contact (not shown).
FAB_XMDC FAB_XMDIO EXTCLK1_IN+ EXTCLK1_

IN-
EXTCLK3_IN+ EXTCLK3_INFAB:XE RTMxTX3+ FAB:XE RTMxRX3+ BASE:GE RTM7RX+ BASE:GE RTM5RX+ FAB:XE RTMxTX3FAB:XE RTMxRX3BASE:GE RTM7RXBASE:GE RTM5RX-
53
Physical Interfaces

Table 23. Zone 3 connector J31 signals ab cd FAB:XE FAB:XE FAB:XE FAB:XE FAB:XE RTM1TX0+ RTM1TX0RTM1TX1+ RTM1TX1RTM1TX2+ 2 FAB:XE FAB:XE FAB:XE FAB:XE FAB:XE RTM1RX0+ RTM1RX0RTM1RX1+ RTM1RX1RTM1RX2+ 3 FAB:XE FAB:XE FAB:XE FAB:XE FAB:XE RTM3TX0+ RTM3TX0RTM3TX1+ RTM3TX1RTM3TX2+ 4 FAB:XE FAB:XE FAB:XE FAB:XE FAB:XE RTM3RX0+ RTM3RX0RTM3RX1+ RTM3RX1RTM3RX2+ 5 COM_PCIE_ COM_PCIE_ COM_PCIE_ COM_PCIE_ NC TX1+ TX1RX1+ RX16 NC NC NC NC NC 7 FAB:XE FAB:XE FAB:XE FAB:XE FAB:XE RTMzTX0+ RTMzTX0RTMzTX1+ RTMzTX1RTMzTX2+ 8 FAB:XE FAB:XE FAB:XE FAB:XE FAB:XE RTMzRX0+ RTMzRX0RTMzRX1+ RTMzRX1RTMzRX2+ 9 BASE:XE BASE:XE BASE:XE BASE:XE BASE:XE RTM1TX0+ RTM1TX0RTM1TX1+ RTM1TX1RTM1TX2+ 10 BASE:XE BASE:XE BASE:XE BASE:XE BASE:XE RTM1RX0+ RTM1RX0RTM1RX1+ RTM1RX1RTM1RX2+ Note: Each differential pair has an individual L-shaped ground contact (not shown). NC = No connection Row # 1 ef FAB:XE RTM1TX2FAB:XE RTM1RX2FAB:XE RTM3TX2FAB:XE RTM3RX2NC NC FAB:XE RTMzTX2FAB:XE RTMzRX2BASE:XE RTM1TX2BASE:XE RTM1RX2gh FAB:XE FAB:XE RTM1TX3+ RTM1TX3FAB:XE FAB:XE RTM1RX3+ RTM1RX3FAB:XE FAB:XE RTM3TX3+ RTM3TX3FAB:XE FAB:XE RTM3RX3+ RTM3RX3BASE_ BASE_ XMDC XMDIO NC NC FAB:XE FAB:XE RTMzTX3+ RTMzTX3FAB:XE FAB:XE RTMzRX3+ RTMzRX3BASE:XE BASE:XE RTM1TX3+ RTM1TX3BASE:XE BASE:XE RTM1RX3+ RTM1RX3-
54

Table 24. Zone 3 connector J32 signals ab cd BASE:XE BASE:XE BASE:XE BASE:XE BASE:XE RTM3TX0+ RTM3TX0RTM3TX1+ RTM3TX1RTM3TX2+ 2 BASE:XE BASE:XE BASE:XE BASE:XE BASE:XE RTM3RX0+ RTM3RX0RTM3RX1+ RTM3RX1RTM3RX2+ 3 BASE:GE BASE:GE BASE:GE BASE:GE BASE:GE RTM18TX+ RTM18TXRTM18RX+ RTM18RXRTM19TX+ 4 BASE:GE BASE:GE BASE:GE BASE:GE BASE:GE RTM20TX+ RTM20TXRTM20RX+ RTM20RXRTM21TX+ 5 NC NC NC NC NC 6 NC NC NC NC NC 7 NC NC NC NC NC 8 NC NC NC NC NC 9 NC NC NC NC NC 10 NC NC NC NC NC Note: Each differential pair has an individual L-shaped ground contact (not shown). NC = No connection Row # 1 ef BASE:XE RTM3TX2BASE:XE RTM3RX2BASE:GE RTM19TXBASE:GE RTM21TXNC NC NC NC NC NC gh BASE:XE BASE:XE RTM3TX3+ RTM3TX3BASE:XE BASE:XE RTM3RX3+ RTM3RX3BASE:GE BASE:GE RTM19RX+ RTM19RXBASE:GE BASE:GE RTM21RX+ RTM21RXNC NC NC NC NC NC NC NC NC NC NC NC
55
Physical Interfaces
Internal interfaces
PowerQUICCIII COP/JTAG header
The common on-chip debug processor/Joint Test Action Group (COP/JTAG) port provides a full duplex serial interface for communications between the internal PowerQUICCIII development support logic and an external development tool. The development port signals are available for use on a 16-pin surface mount COP/JTAG header located on the SCM. The following tables provide the connector and pin functions for the header. Table 25. PowerQUICCIII COP/JTAG connector J3
1 3 5 7 9 11 13 15 2 4 6 8 10 12 14 16
Pin # 1 3 5 7 9 11 13 15
Description JTAG_TDO JTAG_TDI QREQ~ JTAG_TCK JTAG_TMS COP_SRESET~ COPHRESET~ CKSTP_OUT~
Pin # 2 4 6 8 10 12 14 16
Description GND JTAG_TRST~ PULL-UP CKSTP_IN~ N/C GND N/C GND
Table 26. PowerQUICCIII COP/JTAG pin functions Pin # 1 2 3 4 5 6 7 8 9 10, 14 11 12 13 14 15 Name TDO QACK TDI TRST QREQ~ +3.3V TCK CKSTP_IN~ TMS N.C. SOFTREST~ GND HREST# Key CKSTP_OUT~ Description Development serial data out Quiescent Acknowledge (GND) Development serial data in (Pull-up) Test (soft) reset (Pull-down) Quiescent request (Pull-up) +3.3V supply through 1K) Development serial clock (Pull-up) Check Stop Input (Pull-up) Test mode select input (Pull-up) No Connect Soft Reset on PowerQUICCIII (Pull-up) Digital GND Hard Reset (Pull-up) Mechanical Key (No connect) Check stop output (Pull-up)
56
Internal interfaces
IPMC headers
The following headers on the SCM are embedded in the Intelligent Platform Management Controller (IPMC) entity, and are documented in the IPMC Reusable Entity Hardware Specification. J6, H8 Serial Debug and Programming Header J8, H8 JTAG Debugger Header J10, Auxiliary Function and IO FPGA Programming Header J11, Watchdog and Power Header
PowerQUICCIII port 80 interface

The PowerQUICCIII Port 80 interface is used to connect to a RadiSys 84-6000-01 hex display assembly. It connects to the GPIO signals of the parallel bus of the PowerQUICCIII. Figure 13. Port 80 debug socket J4
1 3 5 7 9 11 13
2 4 6 8 10 12 14
Pin # 1 3 5 7 9 11 13
Description +3.3V PRT80D1 PRT80D3 PRT80D5 PRT80D7 GND GND
Pin # 2 4
12 14
Description PRT80D0 PRT80D2 PRT80D4 PRT80D6 MRESET# LCS4# N/C
57
Physical Interfaces
58
INITIAL VERIFICATION AND CONFIGURATION
Configuration checklist
This checklist identifies tasks to be done soon after installing the SCM, and points to information that can help with each task. Some of these tasks are optional. Task: _____ Verifying Base interface operation _____ Verifying Fabric interface operation _____ Verifying Shelf Manager operation _____ Configuring IP addresses for SCM interfaces _____ Adding user accounts and setting passwords Configuring SNMP trap destinations and security _____ access (optional) _____ Enabling the SNMP agent (optional) _____ Enabling the SCMs DHCP server (optional) _____ Enabling and disabling the Telnet server and TFTP server (optional) Enabling and disabling syslog and other services _____ (optional) Copying a configuration from existing CLI _____ configuration files (optional) Copying CLI configuration files to a TFTP server _____ (optional) _____ Copying any file to the SCM (optional) This chapter begins with the following essential procedures: Connecting to the SCM on page 60 Logging in to the SCM on page 60 Rebooting the SCM on page 60 Where to find information: SCM Installation Guide SCM Installation Guide SCM Installation Guide page 61 page 63 page 63 page 64 page 65 page 66 page 67 page 68 page 68 page 69
59
Initial Verification and Configuration
Connecting to the SCM

Serial connection
1. Connect the serial cable to the serial port of the SCM (labeled LMP SER). The appropriate 9-pin D-shell to RJ-45 cable was included in SCM package. 2. Connect the cables other end to the COM1 or COM2 serial port of an external computer (or to a dumb terminal). 3. Start a terminal-emulator application on the external computer. Specify 115200 baud, 8 data bits, no parity, one stop bit, and no hardware or software flow control. When the terminal emulator is set up and connected, you should receive a login prompt.
Telnet connection
1. Connect an Ethernet cable to the SCMs Ethernet maintenance port (labeled LMP ETH). 2. Connect the cables other end to the Ethernet port of a computer that is connected to your network. 3. Configure your computer to a subnet IP address (10.0.0.x, but not 10.0.0.1) and set the netmask to 255.255.255.0. 4. Telnet to IP address 10.0.0.1 (port 23). You should receive a login prompt.
Logging in to the SCM

From the ATCA-2210 login prompt, the user name you enter determines whether you begin in the command line interface (CLI) or in the Linux shell. To access the... CLI Linux shell ...from the... ATCA-2210 login prompt Linux shell ATCA-2210 login prompt CLI ...enter:
admin and password mcli root and password linux-shell
If you have not yet set passwords, no password is required for the root and admin accounts.
Rebooting the SCM

To cause a reboot of the SCM from the Linux shell, enter:
reboot
To physically initiate a reboot of the SCM, use a pen or narrow instrument (not a pencil) and firmly push the recessed reset button on the front panel. See Effects of rebooting the SCM on page 73.
60
Configuring IP addresses
Configuring IP addresses
The SCM IP interfaces are configured by default to minimize your work in setting them up. Some interfaces are set up to get IP addresses from a DHCP server. Others are assigned unique static addresses by including variables for the logical shelf, physical slot, or logical slot numbers in the IP address. A few interfaces have static, literal IP addresses assigned to them that should not be changed. We recommend that you: Keep the default IP addresses for certain interfaces. See Keeping default IP addresses for certain interfaces on page 61. Obtain dynamic IP addresses for interfaces configured for DHCP by connecting the SCM to a network that has a DHCP server and running the DHCP client. See Requesting IP addresses through DHCP on page 61. Add static, literal IP addresses as alternates to some of the literal addresses already assigned (as subinterfaces). See Adding static IP addresses for management ports on page 62. For more information and other possibilities, see Pre-defined IP addressing scheme on page 79.
Keeping default IP addresses for certain interfaces

Do not change the default IP addresses for these interfaces: eth1 - This update channel is used to communicate with the redundant SCM to synchronize the Shelf Managers. Changing this address would reduce performance of Shelf Manager communications with its peer. lo - This internal loopback interface is not visible to any network. Changing it may disrupt occasional communications within the SCM between applications and the local management processor (LMP). For the default addresses, see Table 30 on page 79.
Requesting IP addresses through DHCP

The dtl0 and dtl1 interfaces are configured to request addresses through DHCP, with RadiSysdefined client IDs. To modify the client IDs, see DHCP client on page 81. To get addresses: 1. Physically connect the SCM to a network that has a DHCP server. 2. Connect to the SCM through the serial port or with Telnet. 3. Ensure that the interfaces on the Base and Fabric switches that are expected to communicate with the DHCP server are administratively enabled. See the Ethernet Switching Software Reference for details about the CLI port numbering and the no shutdown command that enables ports. 4. Access the Linux shell.
61
5. Verify that the interfaces have received IP addresses by entering:

ifconfig dtl0 ifconfig dtl1
Look for the inet addr values. If they do not show, wait five minutes and look again. When the SCM reboots, the addresses will be reused or new ones will be requested.
Adding static IP addresses for management ports

This procedure tells how to configure subinterfaces with alternate IP addresses for certain ports: eth0:1subinterface for the front panel Ethernet maintenance port dtl0:2subinterface for the Base interface connection to the local management processor (LMP) dtl1:0subinterface for the Fabric interface connection to the LMP To configure alternate IP addresses for the ports: 1. Connect to the SCMs serial port. 2. Access the CLI. 3. Select the Base interface switch by entering:
base-ethernet
4. Configure the front panel port (eth0) with a new IP address, subnet mask, and (optional) default gateway, using this syntax:
serviceport ip <ipaddr> <netmask> [gateway]
This establishes the subinterface eth0:1 with the IP address you assigned. 5. Configure dtl0 with a new IP address, subnet mask, and (optional) default gateway, using this syntax:
network parms <ipaddr> <netmask> [gateway]
This establishes the subinterface dtl0:2 with the IP address you assigned. 6. Exit the Base interface by entering:
exit
7. Select the Fabric interface switch by entering:

fabric-ethernet
8. Configure the Fabric interface connection to the LMP with a new IP address, subnet mask, and (optional) default gateway, using this syntax:
network parms <ipaddr> <netmask> [gateway]
This establishes the subinterface dtl1:0 with the IP address you assigned. 9. Exit the switch portion of the CLI by typing: exit
62
Adding user accounts and setting passwords
10. When prompted to save changes, say yes (y) to save the changes to the SCMs startup configuration.
Adding user accounts and setting passwords

The SCM initially has two main Linux accounts: root and admin. Neither account is configured with a password by default. Logging in to the blade as root takes you to the Linux prompt. Logging in as admin brings you to the master CLI. You should immediately establish passwords for the root and admin users. In addition, we recommend that you set up non-root user accounts. Use standard Linux methods to administer accounts. To set up another account that goes directly to the CLI, set the login script for the account to /usr/bin/mcli. This mimics how the admin user is set up in the /etc/passwd file. Once logged in, users are not prompted for a password when moving from the CLI to the Linux shell and back.
Configuring SNMP trap destinations and security access

To configure the SNMP notification (trap and inform) destinations and community names: 1. Connect to the SCM. 2. Access the Linux prompt. 3. Make a backup copy of the configuration file:
cp /etc/snmp/snmpd.conf /etc/snmp/snmpd.conf.orig
4. Edit the configuration file /etc/snmp/snmpd.conf. 5. (Optional) Configure a trap destination by uncommenting a template line and filling in values as follows: A. Locate the line that begins with #trap2sink. B. Fill in the host IP address and community name in the correct positions as follows:
trap2sink <ip_addr> <community_name> 162
C. Uncomment the line by removing the # character, and add more lines as necessary. 6. (Optional) Configure an inform destination by uncommenting a line and filling in values as follows: A. Locate the line that begins with #informsink. B. Fill in the host IP address and community name in the correct positions as follows:
informsink <ip_addr> <community_name> 162
C. Uncomment the line and add more lines as necessary.
63
7. Configure community names by uncommenting lines and filling in values as follows: A. Locate the lines that begin with #com2sec. The lines that specify the context name as -Cn fi configure access to the Fabric interface switch objects. The other lines configure access to all other objects, including Base interface switch objects. The items ending in RO or RW are the security names. WARNING Do not change the context name fi, or you will lose access to many Fabric interface switch objects. B. Change the community names. This step is optional, but highly recommended. C. Uncomment these lines and add more lines as appropriate. 8. Save the file. The changes take place when you enable or restart the SNMP agent, or cause it to re-read its configuration file. See Enabling the SNMP agent on page 64 or the CLI command service snmpd on page 106. Changes to the file are saved persistently, so they will be restored after the SCM is rebooted. If you have installed the Net-SNMP 5.2.1 man pages onto a system, you can find more details on the snmpd.conf file options using the command:
man snmpd.conf
Enabling the SNMP agent

This procedure assumes that you already performed the steps in Configuring SNMP trap destinations and security access on page 63. To enable the SNMP agent: 1. Connect to the SCM. 2. Access the CLI. 3. Access blade-management mode by entering:
blade-mgmt
4. Enable the master agent by entering: service snmpd 5. Exit blade-management mode by entering: exit 6. Make the change to the agents enable status persistent across reboots by entering: copy system:running-config nvram:startup-config For more information, see the service snmpd command description on page 106 and the copy system:running-config description on page 102.
64
Enabling the SCMs DHCP server
Enabling the SCMs DHCP server

The SCMs DHCP server is set up to work with a redundant SCM to provide primary and secondary DHCP servers. Configuration files are supplied in /etc as pri.dhcpd.conf.example, sec.dhcpd.conf.example, and common.dhcpd.conf.example, but they must be modified to fit your network. If you want to use the SCMs DHCP server to provide IP addresses, configure and enable the DHCP server as follows: 1. Connect to the SCM. 2. Access the Linux prompt. 3. Change to the directory containing the DHCP server configuration files:
cd /etc
4. Copy the example configuration files or your replacement files to the required file names as follows:
cp pri.dhcpd.conf.example pri.dhcpd.conf cp sec.dhcpd.conf.example sec.dhcpd.conf cp common.dhcpd.conf.example common.dhcpd.conf
5. Modify the interface configuration for the SCM interface that will serve addresses. The DHCP server configuration uses dlt0 by default (as specified in the /etc/default/dhcp file), but dtl0 uses a dynamic IP address by default. 6. Verify that the interface that will serve addresses is administratively enabled through the Base or Fabric Ethernet CLI. 7. Modify the configuration files as necessary for your network. Be sure to make the address and peer address match the IP addresses that the two SCMs will use to communicate with each other in both the primary and secondary files. See page 79 for information about the eth1 update channel IP address, which can be used for this purpose. 8. If this SCM will host the secondary DHCP server, change the configuration as follows: A. Edit the DHCP server initialization script /etc/rc.d/init.d/dhcpd. B. Locate the FAILOVER_PRIMARY line that appears after the commented Read Me text. The line should look like this:
FAILOVER_PRIMARY=1
The default value is 1, indicating the primary, so no changes are needed on the SCM hosting the primary DHCP server. C. Change the FAILOVER_PRIMARY value to 0 to indicate the secondary. The line should look like this when you are done:
FAILOVER_PRIMARY=0
D. Save the file. 9. Create an empty lease database file:

mkdir -p /mnt/RAM touch dhcpd.leases
65
10. Make the lease database file persistent by copying it to the bind mount directory:
cd /rsys/onboot.data/bindmount mkdir -p mnt/RAM cp /mnt/RAM/dhcpd.leases mnt/RAM
11. Create a symbolic link from the /etc/rc.d/rc2.d directory to the startup script:
ln -s ../init.d/dhcpd /etc/rc.d/rc2.d/<link_name>
For example:
ln -s ../init.d/dhcpd /etc/rc.d/rc2.d/S53dhcpd
This causes the DHCP server to be started automatically at SCM bootup. The value of <link_name> determines the order of startup relative to other scripts in the rc2.d directory and should place this service after all other scripts upon which this service depends. 12. Enable the DHCP server immediately by running the startup script:
/etc/rc.d/init.d/dhcpd start
Enabling and disabling the Telnet server and TFTP server

If you want to restrict access to the SCM, you can disable the Telnet server and the TFTP server, which are enabled by default. These services are controlled by the /etc/inetd.conf file. WARNING Do not disable the Telnet server until you confirm access to the SCM through other means. For example, the SSH service is enabled by default, but a user account with a password must be set up before SSH can be used. To enable or disable the Telnet server or TFTP server: 1. Connect to the SCM. 2. Access the Linux prompt. 3. Change to the directory containing the DHCP server configuration files:
cd /etc
4. Edit the inetd.conf file as appropriate: A. To disable the Telnet server, comment out the line beginning with telnet. (You can comment out the line by inserting a pound character (#) at the beginning of the line.) B. To disable the TFTP server, comment out the line beginning with tftp. C. To enable the Telnet server, uncomment the line beginning with telnet (by removing the pound character). D. To enable the TFTP server, uncomment the line beginning with tftp. This changes the enable states of the services the next time the SCM boots up. 5. Enable or disable the services immediately by restarting inetd:
/etc/init.d/inetd restart
66
Enabling and disabling syslog and other services
Enabling and disabling syslog and other services

If you want the SCM to log messages to a file, enable the syslog service. You may want to enable other servicessuch as NTPthat are disabled by default, or to re-enable a service after it has been disabled. To enable a service: 1. Prepare a configuration file for the service with appropriate contents (for example, syslog.conf). 2. Copy the configuration file to the correct location (/etc) on the SCM. 3. Create a symbolic link from the /etc/rc.d/rc2.d directory to the services startup script:
ln -s ../init.d/<startup_script> /etc/rc.d/rc2.d/<link_name>
For example: ln -s ../init.d/syslog /etc/rc.d/rc2.d/S12syslog This causes the service to be started automatically at SCM bootup. The value of <link_name> determines the order of startup relative to other scripts in the rc2.d directory and should place this service after all other scripts upon which this service depends. 4. Enable the service immediately by running the startup script:
/etc/rc.d/init.d/<startup_script> start
For example: /etc/rc.d/init.d/syslog start
To disable a service: 1. Change to the /etc/rc.d/rc2.d directory:

cd /etc/rc.d/rc2.d
2. Identify the symbolic link to the appropriate startup script. For example, listing the details of the syslog link might show:
lrwxrwxrwx 1 root root 16 May 29 08:57 S12syslog -> ../init.d/syslog
3. Disable the service immediately by running the startup script with the stop option:
/etc/rc.d/init.d/<startup_script> stop
For example: /etc/rc.d/init.d/syslog stop 4. Keep the service from restarting when the SCM reboots by removing the symbolic link:
rm <link_name>
For example: rm S12syslog The service can be re-enabled later by adding the link again as describe above.
67
Copying a configuration from existing CLI configuration files

You can change the configuration of the SCM using CLI configuration files created on another SCM or created on this SCM but stored remotely. The files should be placed on an accessible TFTP server (for example, in the /tftpboot directory). This procedure copies the files to the SCM and makes them the current and persistent configurations. To change to configuration settings created by CLI commands: 1. Connect to the SCM. 2. Access the CLI. 3. Retrieve the Base Ethernet switch configuration file and make it the running configuration: 4. base-ethernet
copy tftp://<IP address>/<file-1> nvram:startup-config
Executing the command returns you to the master CLI. 5. Retrieve the Fabric Ethernet switch configuration file and make it the running configuration:
fabric-ethernet copy tftp://<IP address>/<file-2> nvram:startup-config
Executing the command returns you to the master CLI. 6. Retrieve the blade-management configuration file and make it the running configuration (from the master CLI):
copy tftp://<IP address>/<file-3> nvram:startup-config copy nvram:startup-config system:running-config
The running configuration and the persistent startup configuration has changed to reflect the content of the configuration files.
Copying CLI configuration files to a TFTP server

You can copy the CLI configuration files from the SCM to a TFTP server. This procedure also saves the current running configuration to be the SCMs persistent startup configuration, if it wasnt already. Tip Most TFTP servers require that the target files you specify in copy commands (below) already exist on the target TFTP server before you copy them. The target files must also have global write permissions. To avoid creating files in advance, start the TFTP server with the -c option. If you have the Linux man pages installed on a system, see the tftpd man page for details. To save the current configuration to be the startup configuration and to copy the configuration files to a TFTP server: 1. Connect to the SCM. 2. Access the CLI.
68
Copying any file to the SCM
3. Save the Base Ethernet switch configuration file to persistent storage and copy it to the server:
base-ethernet copy system:running-config nvram:startup-config copy nvram:startup-config tftp://<IP address>/<file-1> exit
4. Save the Fabric Ethernet switch configuration file and copy it to the server:
fabric-ethernet copy system:running-config nvram:startup-config copy nvram:startup-config tftp://<IP address>/<file-2> exit
5. Save the blade-management configuration file and copy it to the server (from the master CLI):
copy tftp://<IP address>/<file-3> nvram:startup-config copy system:running-config nvram:startup-config copy nvram:startup-config tftp://<IP address>/<file-3>
The configuration now resides on the TFTP server, and will be restored from persistent storage when the SCM is rebooted.
Copying any file to the SCM

You can copy any file to the SCM using TFTP. The files should be placed on an accessible TFTP server (for example, in the /tftpboot directory). To copy a file to the SCM: 1. Change to the directory where the files should be placed on the SCM. 2. Connect to the TFTP server:
tftp <IP address of TFTP server>
3. (Optional) For binary files, specify binary mode:

bin
4. Retrieve the file:

get <myfile>
Repeat this step as necessary to get multiple files. 5. Exit TFTP:

quit
69
70
SOFTWARE FEATURES
This chapter describes the major software features of the SCM. In some cases, it describes usage of features. It also indicates where to find additional information. The SCM supports the management interfaces and features described below. Management Interface Command Line Interface (CLI) SNMP agent APIs (application programming interfaces) HPI (Hardware Platform Interface) IPMI (Intelligent Platform Management Interface) Feature PICMG-compliant Shelf Manager Ethernet switching (VLANs, spanning tree, link aggregation, Class of Service queuing, storm control, and more) IP routing Described in This manual, starting in Using the Command Line Interface on page 99. Details on the some portions of the CLI appear in other manuals as identified on page 100. This manual in SNMP Agent Support on page 89. In addition, the Ethernet Switching Software Reference describes controls for Layer 2 trap notifications. This manual in API Overview on page 95. The Shelf Management Software Reference. The Shelf Management Software Reference.
Description page 72 page 72 and Ethernet Switching Software Reference page 72 and IP Routing Software Reference page 73 page 74 page 77 page 79 page 82 page 83 page 83 and page 96 page 86 page 86 and page 97 page 86 and page 97
Procedures and Command Reference Shelf Management Software Reference Ethernet Switching Software Reference IP Routing Software Reference Linux and U-Boot documentation N/A N/A page 61 and page 79 page 63 N/A page 106 N/A API Reference page 101 and API Reference
Management Interfaces CLI, HPI, IPMI CLI, API, SNMP MIBs for certain aspects CLI
Linux and boot loader RAM and file system usage (JFFS, persistent files) IPv4 interface and Ethernet switch port identifiers Pre-defined IP addressing scheme (DHCP client, defaults) DHCP server Telnet and SSH Network timing subsystem (NTS) software (system clock) NTP server (time of day) Module state management Flash and file system management (moving files, persistency)
standard Linux and U-Boot standard Linux N/A standard Linux standard Linux standard Linux CLI, API standard Linux API CLI, API
71
6
Feature
Software Features
Description page 86 and page 98 page 86 and page 97 page 87 page 87
Software image management (versions) Firmware and software upgrades Logging service (syslog errors and messages) E-Key authorization and notification
Procedures and Command Reference API Reference Firmware and Software Upgrades page 67 and page 87 N/A
Management Interfaces API Linux utilities, API standard Linux, CLI None
Shelf Manager
The SCM provides a PICMG-compliant Shelf Manager and the IPMI infrastructure for managing the shelf and FRU data. The Shelf Manager keeps track of alarms and provides E-Key services to modules to authorize them to power up (if appropriate) when inserted into the shelf. The Shelf Manager also corrects many error conditions, such as controlling overheating situations by increasing fan speeds. The platform-management CLI, the standard hardware platform interface (HPI), and an example HPI application provide interactive and programmatic management access to the shelf. For more information, see the Shelf Management Software Reference.
Ethernet switching
The SCM provides Ethernet switching software that runs on both the Base and Fabric switches. It contains support for many Layer 2 switching protocols such as virtual LANs, spanning tree, and Class of Service queuing. The SCM provides a CLI, an API, and some SNMP MIB support to manage the switch. For information about the Layer 2 protocols supported and the CLI that controls them, see the Ethernet Switching Software Reference. For information about the SNMP MIB support, see Chapter 7 on page 89 and the Ethernet Switching Software Reference. For information about the API, see API Overview on page 95.
IP routing
The SCM includes IPv4 routing features as a licensed software option. These unicast and multicast routing features are supported on the Fabric interface only. With the IP routing feature set enabled, the SCM supports: OSPF (open shortest path first) version 2 RIP (router information protocol) version 1 and 2 DVMRP (distance vector multicast routing protocol) VRRP (virtual router redundancy protocol) The SCM provides CLI commands to manage the Layer 3 configuration. For more information on the IP routing features and the CLI, see the IP Routing Software Reference.
72
Linux and boot loader
Linux and boot loader

The installed SCM software includes the Wind River Platform for Network Equipment, Linux edition, which supplies the Linux 2.6 kernel along with user space applications and utilities. For information on Linux, see http://www.windriver.com or http://www.linux.org/docs. The SCM uses the RadiSys onboot utility to load its software into RAM so it can run in a way that resembles diskless Linux. For details, see RAM and file system usage on page 74. The software includes the das U-Boot Universal Bootloader, version 1.1.4. U-Boot initializes the SCM hardware and jumps to the Linux operating system. It also provides an environment for debugging, changing the flash contents, and scripting. For information on U-Boot, see http://u-boot.sourceforge.net.
Linux services
All standard services of this Linux version are included on the SCM. While RadiSys provides additional management interfaces and documentation for some services, the following are included without additional RadiSys support: NFS client FTP client TFTP server and client Watchdog daemon BootP
Effects of rebooting the SCM

Rebooting the SCM causes the following to occur: Base and Fabric interface connections from node blades may change over to the redundant SCM and possibly lose traffic. Spanning tree topology changes occur, possibly resulting in lost traffic. If the rebooted SCM hosts the active Shelf Manager or master network timing subsystem (NTS), they will fail over. For more information, see the Shelf Management Software Reference and Network timing subsystem (NTS) software on page 83. The SCM terminates all of its Telnet sessions. Configuration settings revert to those saved as startup configurations (configuration changes not saved are lost). For more information, see Saving configurations persistently on page 101. The file system is recreated, so files and directories revert to those saved persistently (changes not made persistent are lost). For more information, see File persistence on page 74.
73
Software Features
Linux prompt
The default Linux shell prompt can be very informative. It follows this format:
<userName>@<hostName>@<LCAB>-<LCHAS>-<SLOT>:<pwd><symbol>
where
<userName> = the name of the user currently logged in <hostName> = the system name, ATCA-2210 <LCAB> = the logical cabinet number stored in the Shelf FRU Information <LCHAS> = the logical chassis (shelf address) stored in the Shelf FRU Information <SLOT> = the physical slot number in the chassis <pwd> = the present working directory <symbol> = # for the root user or $ for any other user.
For example:
root@ATCA-2210@1-3-8:/etc/sysconfig#
If the Shelf FRU Information including the logical cabinet cannot be read at startup, it is considered an error condition, and the prompt uses this format:
<userName>@<hostName>@ DEBUGMODE :<pwd><symbol>
For information about debug mode, see the troubleshooting information in the Shelf Management Software Reference. The prompt is defined in /etc/profile.
RAM and file system usage

The SCM uses the RadiSys onboot utility to load its software into RAM so it can run in a way that resembles diskless Linux. The SCM runs its software from the RAM disk and uses a JFFS2 (journaling flash file system version 2) for persistent file storage. At SCM bootup, U-Boot copies the kernel and initial RAM disk (initrd) from flash memory into RAM. Then Linux converts the initial RAM disk into a normal RAM disk mounted as the root file system. During the boot process, certain files in the JFFS are linked (bind mounted) to the RAM disk, causing the files in the JFFS to appear in RAM. Certain other files are copied (overlayed) to RAM. The linking and copying of files to RAM occurs before the initialization scripts are run.
File persistence
Because the SCM file system is recreated in RAM at each reboot, changes made to files must be saved persistently to the JFFS or be lost upon reboot. Changes to a RadiSys-defined set of configuration files can be preserved by deliberately saving them persistently through the CLI or API. (See Saving configurations persistently on page 101 and Flash and file system management on page 86.) Other files must be included in a persistency directory.
74
RAM and file system usage
The persistency directories include files that should replace existing files or be added to the RAM file system upon bootup. The types of persistency directories are: Bind mount. The files are linked to the RAM disk, and any changes made to these files in RAM are persistent in the JFFS and relinked in RAM upon reboot. Use the bind mount directory for files that will change over time and must be persistent. The bind mount directory is /rsys/onboot.data/bindmount. Its contents are linked to the RAM file system at the root level (/). For example, /rsys/onboot.data/bindmount/sbin/make-client-id is linked to /sbin/make-client-id. Overlay. The files are copied to the RAM disk, and any changes made to these files in RAM are lost upon reboot. Use the overlay directory for files such as executables that are normally not changed. Overlaying is typically used to replace existing files. The overlay directory is /rsys/onboot.data/overlay. Its contents are recreated at the root level (/) in RAM. For example, /rsys/onboot.data/overlay/usr/bin/myfile.txt replaces the /usr/bin/myfile.txt file in RAM. The JFFS partition is limited to 15.5 MB. Bind mounting and overlaying should be used conservatively, because the partition must hold the factory-installed persistent files, your persistent files, configuration files, and any changes made to the persistent files in RAM. In most cases, individual files should be made persistent instead of whole directories. The factory default configuration includes a number of files and directories that are bind mounted and overlayed. You can view the current overlay and bind mount lists of persistent filesgenerated each time the SCM bootsin /etc/version. All of the /etc directory, which includes all configuration files generated by the CLI, is bind mounted for persistence. In some portions of the CLI, before you request that the configuration be saved or copied persistently, the configuration change exists only within the application. When you request that the current configuration be made persistent, the configuration file is written to the RAM disk, which is also saved to the JFFS because of the bind mount settings. See Saving configurations persistently on page 101.
Adding a persistent file

This procedure explains how to bind mount a file to the RAM file system. Overlaying a file is very similar, but instead uses the overlay directory (/rsys/onboot.data/overlay). This procedure assumes that the original file exists in RAM on the SCM, but files can be moved directly to the persistency directories from another system. To make a file persistent: 1. Copy the file into the bind mount tree:
cp -a <path><file> /rsys/onboot.data/bindmount<path><file>
For example, when <path>=/sbin/ and <file>=make-client-id:

cp -a /sbin/make-client-id /rsys/onboot.data/bindmount/sbin/makeclient-id
2. Cause this persistency configuration to take effect by rebooting the SCM:

reboot
75
Software Features
After reboot, the file appears in RAM. After this point, any changes to a bind mounted file are saved persistently. Any changes to an overlayed file are not.
Adding a persistent directory

This procedure explains how to bind mount a directory to the RAM file system. This procedure assumes that the original directory exists in RAM on the SCM, but directories can be moved directly to the persistency directories from another system. To make a new or existing directory and its contents persistent: 1. Copy the directory and its contents into the bind mount tree:
cp -a <path><dir> /rsys/onboot.data/bindmount<path><dir>
For example, when <path>=/var/lib/ and <dir>=misc:

cp -a /var/lib/misc /rsys/onboot.data/bindmount/var/lib/misc
2. Indicate that the directory should be made persistent and not the individual files inside it by creating a .bindmount file in the directory:
touch /rsys/onboot.data/bindmount<path><dir>/.bindmount
For example, when <path>=/var/lib/ and <dir>=misc:

touch /rsys/onboot.data/bindmount/var/lib/misc/.bindmount
3. Cause this persistency configuration to take effect by rebooting the SCM:

reboot
After reboot, the directory and its contents immediately appear in RAM. After this point, any changes within a bind mounted directory are saved persistently. Tip If you accidentally made an empty directory persistent and overwrote a directory in RAM that had valuable contents, you can retrieve the contents by unmounting the directory (with the umount command). Then you can repeat the above procedure. If you have rebooted the SCM, your contents are lost.
76
IPv4 interface and Ethernet switch port identifiers
IPv4 interface and Ethernet switch port identifiers

This section describes the IPv4 interfaces on the local management processor (LMP) and the Ethernet switch port interfaces on both the Base and Fabric switches. An SNMP ifIndex is listed for each interface, and a CLI slot/port number is listed for each Ethernet switch port interface. Table 27 lists the LMP IPv4 interfaces. Table 28 and Table 29 list the Ethernet switch port interfaces. Table 27. LMP IPv4 interface identifiers and descriptions LMP IPv4 Interface eth2 lo eth0 eth1 dtl0 dtl1 SNMP ifIndex 1 2 3 4 5 6 Description Ethernet connection to the Base switch port 2/2. Software loopback interface. Ethernet connection to the SCM front panel maintenance port labeled LMP ETH. Update channel between the two slots designated for SCMs, used by the Shelf Manager to synchronize with its peer. Pseudo-Ethernet connection to the Base switch, implemented over PCI. Pseudo-Ethernet connection to the Fabric switch, implemented over PCI. Table 28. Base Ethernet switch port identifiers and descriptions Base Switch Port SNMP ifIndex To front slots 1000 +
<slotnum>
To front panel ports
To other SCM To LMP To non-RadiSys Shelf Manager To RTM
For example, 1003 for slot 3. 1021 1022 1023 1024 1025 1033 1034 1036
CLI Slot/Port Description 0/<slotnum> Backplane Base Ethernet connections to front slots, where For example, <slotnum> is the physical chassis slot number. 0/3 for slot 3. 1/5 1/6 1/7 1/8 1/9 2/1 2/2 2/4 10 Gbps Base Ethernet connection to front panel port 1/5. 1 Gbps Base Ethernet connection to front panel port 1/6. 1 Gbps Base Ethernet connection to front panel port 1/7. 1 Gbps Base Ethernet connection to front panel port 1/8. 1 Gbps Base Ethernet connection to front panel port 1/9. Backplane Base Ethernet connection to the other, redundant SCM (AdvancedTCA 3.0 Base channel 2). Base Ethernet connection to the LMP eth2 interface. Base Ethernet connection to a non-RadiSys Shelf Manager (AdvancedTCA 3.0 Base channel 1), to be used when the SCM is in a non-RadiSys shelf. Base Ethernet connections to the RTM through the zone 3 connector.
To LMP
1053 1054 1055 1056 1065
3/5 3/6 3/7 3/8 N/A
Base pseudo-Ethernet connection to the LMP dtl0 interface, implemented over PCI.
77
Software Features
Table 28. Base Ethernet switch port identifiers and descriptions Base Switch Port SNMP ifIndex Link aggregation 1066 connections 1067 1068 1069 1070 1071 1072 1073 CLI Slot/Port 4/1 4/2 4/3 4/4 4/5 4/6 4/7 4/8 Description Base Ethernet link aggregation (LAG) logical portchannels. These exist only if created by users and are numbered in the order in which they were created.
Table 29. Fabric interface Ethernet interface identifiers and descriptions Fabric Switch Port SNMP ifIndex To front slots 2000 +
<slotnum>
For example, 2005 for slot 5. To front panel ports 2017 2018 2019 2020 To RTM 2049 2050 2051 To LMP 2065 Link aggregation connections 2066 2067 2068 2069 2070 2071 2072 2073 N/A
CLI Slot/Port Description 0/<slotnum> Backplane Fabric Ethernet connections to front slots, where For example, <slotnum> is the physical chassis slot number. 0/5 for slot 5. 1/1 1/2 1/3 1/4 3/1 3/2 3/3 N/A 4/1 4/2 4/3 4/4 4/5 4/6 4/7 4/8 5/1 to 5/28 10 Gbps Fabric Ethernet connection to front panel port 1/1. 10 Gbps Fabric Ethernet connection to front panel port 1/2. 10 Gbps Fabric Ethernet connection to front panel port 1/3. 10 Gbps Fabric Ethernet connection to front panel port 1/4. Fabric Ethernet connections to the RTM.
Fabric pseudo-Ethernet connection to the LMP dtl1 interface, implemented over PCI. Fabric Ethernet link aggregation logical port-channels. These exist only if created by users and are numbered in the order in which they were created.
VLAN routing interfaces
Fabric Ethernet VLAN router interfaces. These exist only for VLANs on which routing is enabled. They are numbered in the order in which the VLANs had routing enabled.
78
Pre-defined IP addressing scheme

The SCM supports a number of IP interfaces that are used as management interfaces to the blade. The interfaces are configured by default to minimize your work in setting them up. This section describes the default IP addresses, the pre-defined address assignment scheme, and the options for customizing address assignment if necessary. For a basic overview and the initially recommended steps, see Configuring IP addresses on page 61.
Default IPv4 addresses

Table 30 summarizes the IPv4 interfaces and their default IP addresses. Table 30. Default IP addresses and address usage for IP interfaces Interface Default IPv4 Address/ Name Subnet Mask lo 127.0.0.1/8 eth0 10.0.0.1/24 Description Software loopback interface on the LMP. Ethernet maintenance port labeled LMP ETH on the SCM front panel. Interface and IP Address Usage IPv4 datagrams sent to this interface by the LMP are received immediately by the LMP. Do not change this address. This interface can be used as a maintenance port for the entire ATCA shelf. From an attached remote console, a user can update the SCMs firmware and software, and also perform configuration and service of other blades that have IPv4 connectivity to the SCM. This private IPv4 address allows for easy connectivity from a service terminal connected to it, provided that the terminal is also on the 10.0.0.0/24 subnet. If IPv4 Masquerade is enabled on the SCM, this allows the terminal to easily communicate with other blades beyond the SCMs switch fabric, since the SCM translates the private addresses on the 10.0.0.0/24 network to the public or private address assigned on its dtl0 interface. This subinterface allows users and validators to use a shared LAN segment to maintain concurrent IPv4 connectivity to SCMs via their eth0:0 subinterfaces. It also allows multiple SCMs to be networked together via their eth0 physical interfaces without reconfiguring the eth0 IP addresses. This subinterface is configured through the Base interface CLI using the serviceport ip command. See page 62. This interface is used by the Shelf Manager to synchronize with its peer on the redundant SCM. The interface could also be used for management. Changing this address reduces performance of Shelf Manager synchronizations, which then use a slower path. This is a pseudo-Ethernet interface.
eth0:0
eth0:1 eth1
10.1.<LCHAS>.<SLOT>/ Subinterface. 16 (where <LCHAS> is the logical chassis and <SLOT> is the physical slot) None, interface not created Subinterface. 10.0.1.<LSLOT>/24 (where <LSLOT> is the logical slot) PICMG 3.0 update channel to the redundant SCM, connecting logical slots 1 and 2. PCI bus connection between the Base switch and the LMP. Subinterface.
dtl0
Dynamically assigned by DHCP 10.2.<LCHAS>.<SLOT>/ 16
dtl0:0
This subinterface can be used to configure the SCM or the shelf.
79
Software Features
Table 30. Default IP addresses and address usage for IP interfaces Interface Default IPv4 Address/ Name Subnet Mask dtl0:1 <ShelfMgrIPAddr>/ 24, which is 192.168.16.17/24 by default Description Used by the LMP to communicate with the Shelf Manager through the Base interface. Interface and IP Address Usage This subinterface is dedicated to providing access to the Shelf Manager server. This subinterface exists only on the SCM with the active Shelf Manager, and is enabled and disabled by the Shelf Manager server. This setup allows the system manager to always reach the Shelf Manager, regardless of which SCM is active at a given time. The Shelf Manager IP address is stored in the Shelf FRU Device and can be configured via the HPI example application, hpiapp. Updating the Shelf Manager IP address immediately updates the address of dtl0:1, bringing the interface down and then up with the new address. This subinterface is configured through the Base interface CLI using the network parms command. See page 62. This is a pseudo-Ethernet interface.
dtl0:2 dtl1
None, interface not created Subinterface. Dynamically assigned by DHCP PCI bus connection between the Fabric switch and the LMP. None, interface not created Subinterface.
dtl1:0
This subinterface is configured through the Fabric interface CLI using the network parms command. See page 62.
Static IP address assignment configuration file setup

The static IP address assignments are determined by configuration files. The /etc/sysconfig/network-devices directory contains the configuration files that determine the IP assignments at SCM bootup. Files called template.ifconfig.<interfaceName> are interpreted and used to create ifconfig.<interfaceName> files. The interpretation process replaces variables with appropriate values. Interfaces that receive literal IP addresses or that use DHCP have just an ifconfig.<interfaceName> file and no template file, because there is nothing to interpret. The IP address configuration files are saved persistently because the /etc directory is bind mounted as a factory default setting.
80
Variables used to assign IP addresses

Table 31 explains the use of specific variables to assign static IP addresses. Table 31. Use of variables in IP address templates Variable How It is Useful <LCHAS> - The logical shelf ID, or shelf logical address, can identify the physical chassis location of the shelf within a central office. It can also be used to generate unique IP addresses and prompts to help users distinguish between shelves. How the Value Is Set and Obtained RadiSys assigns the default value of 255 to the shelf. During the commissioning process, a shelf should be assigned a shelf address that is unique within a particular central office, or at least within the domain in which the SCM communicates. This information is stored in the Shelf FRU Device on the shelf, and is read during host processor initialization. If the shelf address does not conform to the RadiSys-defined format (described in the Shelf Management Software Reference), the SCM cannot read it, and the default value is used. The slot number is read during host processor initialization.
<SLOT> physical slot in shelf <LSLOT> logical slot
The physical slot location uniquely identifies this SCMs location in the shelf. The logical slot is an easy way for The logical slot identifies the position of the SCM in the chassis the SCM to distinguish itself from in relation to the other SCM. The lower-numbered physical slot the other SCM. is logical slot 1. The higher numbered slot is logical slot 2.
Note The factory default Logical Shelf address is 255. If two shelves are connected which both have factory default settings, there will be multiple SCMs with the same eth0:0 IP address (for example, 10.1.255.7 for the slot 7 SCM and 10.1.255.8 for the slot 8 SCM). To avoid this, add a shelf to the network and commission it with a new shelf address before adding another shelf.
DHCP client
Some interfaces are set up to use the DHCP client (/sbin/dhclient) to request an IP address from a DHCP server on the network. The shelf address and the physical slot number are used to derive the DHCP client identifier string sent in DHCP client requests. The client ID is formatted as a colon-delimited hexadecimal string containing the logical shelf and physical slot. The ID always begins with a null character (for example, 00:34:2d:38 for logical shelf 4 and physical slot 8). To modify the format of the client ID, edit the /etc/sysconfig/networkdevices/make-interfaces script. The client ID appears as the variable CLIENT_ID in /etc/template.dhclient.conf and is replaced by the actual value in the dhclient.conf file. By default the DHCP client waits 300 seconds for a DHCP server to respond before timing out. If a timeout occurs, the interface is not initialized. The blade continues to boot up, and the DHCP client attempts to acquire an IP address in the background. The timeout can be configured in the /etc/template.dhclient.conf file. A value of 0 represents an infinite timeout. If you have Linux man pages installed on a system, refer to the dhclient.conf man page for more information.
81
Software Features
To request addresses for the first time, see Requesting IP addresses through DHCP on page 61.
Customizing IP address assignments

Addresses to avoid changing
We recommend keeping the default addresses for: lo eth1 In addition, it may not be necessary to change the other statically configured IP addresses. For default values, see Table 30 on page 79. For more explanation, see Keeping default IP addresses for certain interfaces on page 61. For a procedure, see Adding static IP addresses for management ports on page 62.
Customization options
Users who wish to further change the default configuration of the IPv4 interfaces have these choices: Change the /etc/sysconfig/network-devices/template.ifconfig.<interface> template files to modify the variables used in assigning addresses. Change the /etc/sysconfig/network-devices/ifconfig.<interface> files to change the static address assignments. Change the external DHCP server configuration to change the dynamic address assignments based on the client identifier. Change the run level startup script (/etc/rc.d/init.d/network) to modify the default behavior for configuring the static and dynamic address assignments. Users who wish to change the DHCP client behavior have these choices: Change the /etc/template.dhclient.conf files to change the protocol timers and options used by the DHCP client. See DHCP client on page 81. Change the run level startup script (/etc/rc.d/init.d/network) to modify the default behavior for launching the DHCP client application. Change the DHCP client application itself and the scripts that invoke it.
DHCP server
The SCM supports the standard Linux DHCP server, which can supply IP addresses for blades and other network equipment on the same network as the SCM. For more information on setting up the DHCP server, see Enabling the SCMs DHCP server on page 65.
82
Telnet and SSH
Telnet and SSH

The SCM supports the standard Linux Telnet server and client to enable connections to and from the SCM. The Telnet server daemon is enabled by default, responding to requests on TCP port 23 and bringing you to the blade login prompt. (See Logging in to the SCM on page 60.) The SCM also supports secure shell (SSH) client and server, along with the rest of the full OpenSSH distribution. SSH allows secure connections to and from the SCM, and requires a password. SSH protocol version 2.0 is supported. The SSH server is enabled by default, responding to requests on TCP port 22. To disable the Telnet server, see Enabling and disabling the Telnet server and TFTP server on page 66. To disable the SSH server, see Enabling and disabling syslog and other services on page 67.
Network timing subsystem (NTS) software

The SCMs network timing subsystem provides centralized system timing based on the synchronization clock interface defined in the PICMG 3.0 Revision 2.0 AdvancedTCA Base Specification. The SCM can receive timing reference signals from the network interfaces and external clock sources and can provide a network-synchronized system clock to the other interconnected blades in the shelf. Working together with capable node blades, the SCM can provide uninterrupted (hitless failover) network timing for the whole shelf. Primary reference sources (PRSs) provide the highest level of time and frequency accuracy for telecommunications networking. International standards require that all clocks in a synchronization network be referenced or traceable to a PRS. PRS timing can come from a variety of sources, such as a GPS, synchronous supply unit (SSU), or building integrated timing supply (BITS). These sources distribute precise time and frequency information to equipment in central offices, enabling voice, video, and data traffic to be transported reliably and efficiently throughout the network.
Clock bus signals

The SCM requires that the backplane synchronization clock interface buses be used as follows: CLK1A and CLK1B - 8 kHz system clock distributed by an SCM to node blades. CLK2A and CLK2B - 19.44 MHz system clock distributed by an SCM to node blades. CLK3A and CLK3B - 8 kHz, 1.544 MHz, 2.048 MHz, or 19.44 MHz network timing reference input from node blades. The SCM allows flexibility in the usage of the clock buses, allowing a clock buss A and B halves to be driven by two separate sources. For example, to provide redundancy, the CLK1 and CLK2 pairs can be split with the A sides being driven by one SCM and the B sides by the other SCM. For details, see System clock redundancy on page 85.
83
Software Features
Selectable timing sources

The SCM can select reference timing sources from: Synchronization clock interface. CLK3A and CLK3B can each carry a timing signal from a node blade. RTM. The RTM may have external unframed sources or framed sources that it de-frames and provides to the SCM. The SCM receives these reference clock inputs at multiples of 8 kHz up to 19.44 MHz. Update channel interface. The SCM receives high speed and frame alignment clock signals from the redundant SCM over dedicated links.
System clock outputs

The SCM can output network-synchronized system clocks to: Synchronization clock interface. CLK1A, CLK1B, CLK2A, and CLK2B can each carry a system clock signal to the node blades. RTM. The SCM drives system clock signals to the RTM that may be used by the RTM to generate framed or unframed shelf outputs. The SCM supplies the clocks at 8 kHz, 19.44 MHz, and either 1.544 MHz or 2.048 MHz. Update channel interface. The SCM outputs high speed and frame alignment clock signals to the redundant SCM over dedicated links.
Source priorities
The SCM can be configured with source priority levels for each input clock. The SCM can automatically respond to a faulty clock source by selecting the next best source based on configured priorities. Alternatively, an application such as a system manager can explicitly specify the input to use as the reference clock source.
Hitless Failover Ability

The SCMs and the node blades work together to provide hitless failover network timing for the ATCA shelf. The SCM provides a consistent system clock, even when its primary reference source becomes invalid or it transitions to a secondary source. Two SCMs acting in concert provide redundant system clock sources for the node blades. In case of an interruption to an SCM's system clock output, node blades must also maintain stable clocking by transitioning to the redundant SCM as the system clock source. All of these safeguards together provide for hitless failover for the entire shelf.
Transition to a secondary input timing source

The SCM monitors incoming timing sources and qualifies their signals against network timing standards. Timing sources are selected based on user-assigned priorities. The SCM locks to its primary reference and uses it to distribute precise, network-synchronized, 8 kHz and 19.44 MHz system clock signals to the node blades.
84
Network timing subsystem (NTS) software
If the SCM's primary reference clock fails, the SCM selects a secondary source based on the user-configured priorities. The internal digital PLL (phase-locked loop) immediately enters digital holdover mode and then locks to the new timing source. Digital holdover mode maintains system clock stability while not locked to a timing source. It accomplishes this by using consistent timing from the SCM's 12.8 MHz local oscillator in conjunction with stored, averaged frequency data collected while locked to a qualified source. If all incoming timing references become disqualified, the PLL immediately enters digital holdover mode to maintain a stable system clock until it identifies and locks to a qualified source.
System clock redundancy

By default, the pair of redundant SCMs are configured to provide hitless failover, with one acting as the network timing master and one as a slave. The slave SCM locks to the masters system clock output, based on the clock signals received through the update channel. Since the slave can output the same signals as the master, seamlessly redundant clocking is possible if you configure your system as follows: The master SCM drives CLK1A and CLK2A, and the slave drives CLK1B and CLK2B. Node blades prioritize the master SCM clocks as top priority, with the slave clocks as second priority. With this configuration, the slave provides an immediate failover alternative without any clock cycle loss. For example, if the master SCM clock on CLK1A becomes invalid, the node blade can switch to CLK1B and lock to the new source. The roles of the SCMs can be reversed, or they can be autonomous, where they do not exchange timing information.
Node blade transition to redundant system clock

Hitless failover of timing relies on node blades that can seamlessly transition to a secondary system clock upon failure of the primary clock. If the node blades primary reference clock (master SCM system clock) becomes disqualified, the node blades must remain active and maintain stable clocking while transitioning to the slave SCM system clock.
E-Key management of clock buses

The SCM can restrict entities from driving each clock bus by requiring E-Key authorization. With the SCM configured to require authorization, blades and applications must request permission to drive a clock bus. The Shelf Manager handles the details of enabling the bus by granting tokens to requesters. A system manager or custom application can control the granting of tokens through the NTS API. See NTS API on page 96 and network-clock-select request-busedresource on page 111.
85
Software Features
NTP server
The SCM supports the Linux standard NTP server, which allows it to provide time of day services to other blades or systems. The NTP server is disabled by default. To enable it, see Enabling and disabling syslog and other services on page 67.
Module state management

The SCM provides an API to display and set the status of the SCMs rear transition module. For more information on the API, see Module state management API on page 97. Additionally, the SCM provides control over the hot-swap state of any shelf FRU through the platform-management CLI. See the Shelf Management Software Reference for details.
Flash and file system management

The SCM provides CLI commands and an API to help you manage configuration files and other files on the SCM. These help you to: Copy RadiSys-defined configuration files to new locations, on or off the blade. Retrieve configuration files from another system to the blade. Save the current configuration persistently. Reset the configuration to the one in persistent storage (the startup configuration). The RadiSys-defined configuration files are those operated on by the CLI. They cover all configuration performed in the CLI. For more details on performing these functions through CLI commands see Saving configurations persistently on page 101 and Master CLI commands on page 102. For more information about the API, see Flash and file system management API on page 97.
Software image management

The SCM provides an API to report the software images residing on the SCM. For more information about the API, see Software image management API on page 98.
Firmware and software upgrades

The SCM provides Linux utilities to help you upgrade the firmware. It provides a Linux utility to help you upgrade the software. An API can also handle certain aspects of software upgrades, and a software upgrade example application is provided. For details on the Linux utilities, see Firmware and Software Upgrades, which is included in the Promentum software distribution. For an introduction to the API, see Flash and file system management API on page 97.
86
Logging service
Logging service
The SCM supports the Linux standard syslog service, which allows it to collect messages from various SCM software components and control the output. By default, the blade-wide syslog service is disabled, but logging is enabled for several components. Messages from the components are ignored unless the blade-wide service is enabled. The blades syslog configuration also determines the destination for the messages. See Enabling and disabling syslog and other services on page 67. The Ethernet switching software is an example of a component that enables syslog generation by default. Generation of syslog messages is controlled separately on the Base and Fabric switches. To disable it or re-enable it, see the logging syslog command in the Ethernet Switching Software Reference.
E-Key authorization and notification

The SCMs Shelf Manager provides E-Key authorization for blades to enable and disable the backplane ports between blades and modules (such as AMCs). The E-Key notification handler communicates with the Shelf Manager and notifies interested applications when ports change E-Key states. For example, the Ethernet switching software needs to know the E-Key state of relevant ports. The Ethernet switching software registers itself with the E-Key notification handler through the internal E-Key API to receive notifications when its switch ports become enabled or disabled. Other software components may also register themselves. In addition, the Shelf Manager tracks current authorization for the synchronization clock interface using bused E-Key commands, giving permission to blades as appropriate to drive clock signals onto the clock buses. For more information, see Network timing subsystem (NTS) software on page 83.
87
Software Features
88
SNMP AGENT SUPPORT
The SCM supports an SNMP agent that enables remote SNMP management of the SCM. The SNMP agent implementation is modular and extensible, dividing the functionality into a master agent and two subagents that represent the Base interface switch and the Fabric interface switch. The master agent communicates with the subagents using the AgentX protocol. The master agent receives requests from the SNMP manager and completes the requested action, coordinating with its subagents when necessary. The master agent also generates notifications and relays notifications generated by the subagents to the registered notification destinations. The SNMP agent must be configured and enabled as described in Configuring SNMP trap destinations and security access on page 63 and Enabling the SNMP agent on page 64. The SCM implements its SNMP agent by incorporating Net-SNMP open-source software, version 5.2.1. Many usage details are covered in Net-SNMP man pages and on the http://www.net-snmp.com Web site. The SCM currently supports SNMPv2c.
Base or Fabric switch selection

The SNMP objects in the BRIDGE-MIB, P-BRIDGE-MIB, and Q-BRIDGE-MIB modules are instantiated separately for the Fabric interface and Base interface switches. You must use community names that map to the "fi" context name to access these objects for the Fabric interface switch. Use community names that map to the null context name to access all other objects. Spanning tree notifications (newRoot and topologyChange) include the dot1dBaseBridgeAddress.0 variable to indicate the appropriate switch. Its value is the 6-byte MAC address used in the switch's bridge priority value. The dot1dBaseBridgeAddress.0 value for the switches is shown as output of the Ethernet CLI show spanning-tree brief command. When accessing either the Base or Fabric switch CLI, and running show spanning-tree brief, look at the last six bytes of the Bridge Priority field to determine the value.
ifIndex values for IP interfaces

For a list of ifIndex values that identify the SCM ports, see IPv4 interface and Ethernet switch port identifiers on page 77.
89
SNMP Agent Support
MIB module support

This section lists the MIB modules that the SCM generally supports. The major exceptions to total support are noted in the sections that follow. The SCM supports: RFC1213-MIB (see page 90) SNMPv2-MIB (see page 90) IF-MIB (see page 90) IP-MIB (from RFC 2011, no major exceptions) TCP-MIB (from RFC 2012, no major exceptions) UDP-MIB (from RFC 2013, no major exceptions) BRIDGE-MIB (see page 90) P-BRIDGE-MIB (see page 91) Q-BRIDGE-MIB (see page 91) EtherLike-MIB (see page 90) IEEE8023-LAG-MIB (see page 91)
RFC1213-MIB support details

The following objects are not supported: egp scalars egpNeighTable The following tables are supported as originally specified in RFC1213-MIB: atTable ipRouteTable The remaining scalars and tables are supported as re-specified in SNMPv2-MIB, IF-MIB, IP-MIB, TCP-MIB, and UDP-MIB.
SNMPv2-MIB support details (from RFC 3418)

The warmStart notification is not supported.
IF-MIB support details (from RFC 2863)

ifTestTable is not supported.
BRIDGE-MIB support details (from RFC 1493)

dot1dStaticTable is not supported.
90
Notification support
P-BRIDGE-MIB support details (from RFC 2674)

The following tables are not supported: dot1dUserPriorityRegenTable dot1dPortOutboundAccessPriorityTable
Q-BRIDGE-MIB support details (from RFC 2674)

The following tables are not supported: dot1qForwardAllTable dot1qForwardUnregisteredTable dot1qStaticUnicastTable dot1qPortVlanStatisticsTable dot1qPortVlanHCStatisticsTable dot1qLearningConstraintsTable
EtherLike-MIB support details (from RFC 2665)

dot3CollTable is not supported.
IEEE8023-LAG-MIB support details

dot3adAggPortDebugTable is not supported.
Notification support
The SCM supports the following notifications: coldStart (SNMPv2-MIB) authenticationFailure (SNMPv2-MIB) linkDown (IF-MIB) linkUp (IF-MIB) newRoot (BRIDGE-MIB) topologyChange (BRIDGE-MIB)
91
SNMP Agent Support
MIB modules supplied

The following MIB modules are supplied as text files installed on the blade: RFC1213-MIB (from RFC 1213) SNMPv2-MIB (from RFC 3418) IF-MIB (from RFC 2863) IP-MIB (from RFC 2011) TCP-MIB (from RFC 2012) UDP-MIB (from RFC 2013) BRIDGE-MIB (from RFC 1493) P-BRIDGE-MIB (from RFC 2674) Q-BRIDGE-MIB (from RFC 2674) EtherLike-MIB (from RFC 2665) IEEE8023-LAG-MIB (from IEEE) RMON-MIB (from RFC 2819) SNMPv2-SMI (from RFC 2578) SNMPv2-TC (from RFC 2579) SNMPv2-CONF (from RFC 2580) SNMPv2-TM (from RFC 3417) RFC1155-SMI (from RFC 1155) SNMP-FRAMEWORK-MIB (from RFC 3411) SNMP-MPD-MIB (from RFC 3412) SNMP-NOTIFICATION-MIB (from RFC 3413) SNMP-TARGET-MIB (from RFC 3413) SNMP-USER-BASED-SM-MIB (from RFC 3414) SNMP-VIEW-BASED-ACM-MIB (from RFC 3415) NET-SNMP-AGENT-MIB (from Net-SNMP package) NET-SNMP-MIB (from Net-SNMP package) NET-SNMP-TC (from Net-SNMP package) IANAifType-MIB (from http://www.iana.org on May 19, 2006) RADISYS-ATCA-MIB (from proprietary RadiSys MIB) RADISYS-MIB (from proprietary RadiSys MIB)
92
MIB browser applications
These files are located on the blade at /usr/share/snmp/mibs. To make it easy to copy these files elsewhere, the same directory contains a zipped archive of these files called mibs.tgz. A README.txt file located there lists these files and their sources.
MIB browser applications

The SCM provides the Net-SNMP MIB browser tools, which are utility applications that you can use to simulate an SNMP managers interaction with the master agent. They are installed directly on the SCM in the /usr/bin directory. A few of the utilities are: snmpget snmpset snmpwalk If you have the Net-SNMP 5.2.1 man pages installed on a system, execute this command for more information:
man snmpcmd
93
SNMP Agent Support
94
API OVERVIEW
This chapter provides an introduction to the C-language APIs provided for management of the SCM. The APIs initially cover a portion of the configuration and management features found in the CLI, with more features to be added incrementally. Currently, applications must run locally on the SCMs LMP. The APIs introduced here are the: Ethernet switching API on page 96 NTS API on page 96 Module state management API on page 97 Flash and file system management API on page 97 Software image management API on page 98 Management plane framework (MPF) API on page 98
Where to find related files and additional information

To support application development, the RadiSys Promentum software distribution includes the following resources. Resource README.ATCA-2210 file explaining: Distribution contents Development host requirements Instructions for library installation Instructions for building provided examples API libraries and header files in RPM packages API client example source code API reference manuals ATCA-2210/api-devel/<release_ID>/RPMs/ API_Development API_Development Location in Software Distribution API_Development
Reference information for the Module state management API, Flash and file system management API, and Flash and file system management API is located in the Blade Management API Reference provided in the software distribution. 95
8 Ethernet switching API

The Ethernet switching API allows you to configure and monitor statistics on the Base and Fabric switches. Each switch is treated as a forwarding element (FE). A function call returns the array of FE information. Each FE is identified by a handle, which can be used in subsequent calls to identify the FE (switch). The Base switch is assigned the product designator type of zero (0), while the Fabric switch has a non-zero type. A few relevant details are as follows: Function name prefix: Shared library: Header file to include: RSYS_Esw librsyseswapi.so.0 rsys_esw_api.h
NTS API
The network timing subsystem API allows you to configure the SCMs clock generation and select the network timing sources. It also allows you to retrieve the status and receive asynchronous NTS events. These events include: Loss of signal on any input Change of reference source state between valid and invalid Change in operating state, such as locked or holdover Temporary interruption to the currently selected reference source A few relevant details are as follows: Function name prefix: Shared library: Header file to include: RSYS_Nts librsysntsapi.so nts/rsys_nts_api.h
96
Module state management API
Module state management API

The module state management API is useful for monitoring and controlling the state of the SCMs RTM. This API can keep both the controlling SCM and the system manager informed about the RTM. Through the module state management API, an application can: Retrieve the modules operational state Change the modules administrative state Retrieve the modules current usage state A few relevant details are as follows: Function name prefix: Shared library: Header file to include: RSYS_Mm librsysmmsapi.so rsys_mm_api_d0.5.h
Flash and file system management API

The flash and file system management API allows loading and management of configuration and other files. It provides capabilities to: Report the CLI-created configuration files stored on the blade Restore the startup configuration Copy a file to a local or remote location Rename, delete, or list details about a file A few relevant details are as follows: Function name prefix: Shared library: Header file to include: RSYS_Fm librsysmmsapi.so rsys_mm_api_d0.5.h
A provided example application copies a software image to the SCM and then upgrades the running software to that image by invoking the Linux software upgrade utility.
97
8 Software image management API

The software image management API reports the software images and versions residing on the SCM. A few relevant details are as follows: Function name prefix: Shared library: Header file to include: RSYS_Swm librsysmmsapi.so rsys_mm_api_d0.5.h
Management plane framework (MPF) API

The RadiSys MPF API provides access to general purpose features that can be useful in application development. It provides: An event loop architecture Multiple socket connection support, signal, and timer events Common error definition and logging A syslog wrapper A runtime debug logging framework A few relevant details are as follows: Function name prefix: Shared library: Header file to include: RSYS_Mpf libmpf.so mpf/rsys_mpf_api.h
98
USING THE COMMAND LINE INTERFACE
The SCMs command line interface (CLI) allows you to interactively configure and monitor a wide range of services. This chapter describes how the CLI is divided and provides basic usage information that is useful throughout the different areas of the CLI. It refers to other locations for reference material on specific commands.
Accessing the master CLI

To access the master CLI: Connect to the SCM as described on page 60. Log in to the CLI as described on page 60. Once logged in to the CLI, you should see the ATCA-2210 prompt.
Displaying available commands and options

To access online help while typing commands, enter a question mark (?). The CLI returns a list of the commands or options available in your current mode. To get a list of available commands, enter a question mark (?):
ATCA-2210# ? base-ethernet : fabric-ethernet: blade-mgmt : linux-shell : show : platform-mgmt : copy : erase : exit : logout : help : Start base ethernet CLI Start fabric ethernet CLI Blade management configuration and status Start the Linux shell Show configuration and status Platform management configuration and status Copy information to and from the blade Erase startup configuration Logout from CLI session Logout from CLI session Operation instructions
To get help on a specific command, type as much as you know of the command, followed by a question mark. Many commands require multiple arguments, each one shown as the next option when you type a question mark. After entering an option, type a question mark again to see if there are more options. The <CR> notation indicates that a carriage return is an option. This means that you have typed all required arguments, and you can press Enter to execute the command.
99
Using the Command Line Interface
Master CLI command modes

The top-level (master) CLI is divided into several command modes that give access to different subsets of commands. Some of the modes are documented in detail in this manual, while other modes are documented in manuals that describe those portions of the software in more detail. As indicated from the list of commands above, the master CLI has the following main command modes: Mode Access Method Base Ethernet base-ethernet Fabric Ethernet fabric-ethernet Prompt
ATCA-2210-Base# ATCA-2210-Fabric#
Mode Description Configure the Base Ethernet switch. Configure the Fabric Ethernet switch, including optional IP routing features. provides, including Linux services, SNMP, and network timing. Access the Linux shell, which is not part of the CLI. When you exit the shell, you return to the CLI. Access shelf management, alarm, and FRU information.
Blade management Linux shell
blade-mgmt
ATCA-2210(blade-man)# Configure the services the blade
Described in Ethernet Switching Software Reference and IP Routing Software Reference This manual on page 105 Standard Linux documentation Shelf Management Software Reference
linux-shell
Linux prompt (see page 74)
Platform management
platform-mgmt
platform-mgmt#
The master CLI also contains some commands that can be executed directly, without entering a mode. These are described in Master CLI commands on page 102. Many of these modes contain additional modes within them. The additional modes are described with the command reference information for each command mode. To exit any of the modes or the master CLI, enter:
exit
No commands
Many configuration commands have a no form that is used to disable, delete, or reset a configuration back to the factory defaults. For example, the following command administratively disables a port:
shutdown
The no form of this command administratively enables a port:

no shutdown
100
Show commands
Show commands
A show command displays information about resources. You can use the applicable show command to show information for individual resources. The show commands are executed outside of configuration modes, usually directly within a main command mode such as Base Ethernet.
Saving configurations persistently

Some CLI configuration changes are kept within the application until you save them to persistent storage. If you do not save, your unsaved changes remain in effect until the module is rebooted, which restores the startup configuration. These methods for saving CLI configurations are currently required for the various CLI command modes: Mode Persistent Save Method Base Ethernet Two alternatives: Fabric Ethernet Save when exiting these modes, as prompted. From within these modes, use this command:
copy system:running-config nvram:startup-config
See also Ethernet Switching Software Reference page 102 N/A
Blade management Platform management
From the master CLI, use this command:

copy system:running-config nvram:startup-config
Not necessary. The configuration is saved persistently when you execute the commands.
Command line completion

Pressing the tab key completes a partially typed command keyword. For example, typing a partial command, con, and pressing the tab key completes the keyword configure. Typing a question mark (?) lists valid entries after a command keyword. For example, typing the keyword configure and then a question mark brings up a list of valid options that follow configure.
Editing and history keys

The editing key sequences are similar to those used by EMACS text editors. Any character typed is inserted into the command line at the current cursor position, and all characters to the right of the typed character shift to the right. The history key sequence is similar to that used by the UNIX C shell. You can obtain a list of key sequences by entering help at the main CLI prompt.
101
Master CLI commands

show version
Shows the software versions of the master CLI and various CLI components. Syntax:
show version
show running-config
Shows the current configuration, displayed as a series of commands necessary to reproduce the current configuration. This command currently shows only the blade-management mode configuration. Syntax:
show running-config
copy tftp
Copies the specified file from a TFTP server and makes it the startup configuration in flash memory or the running configuration. This command currently replaces only the blademanagement mode configuration. Syntax: Options:
nvram:startup-config copy tftp:<url> { nvram:startup-config | system:running-config }
Loads the configuration file so it becomes the persistent startup configuration.

system:running-config
Loads the configuration file so it becomes the running configuration.
copy system:running-config
Saves the current running configuration to persistent storage on the SCM or to another system. If saved to persistent storage, the saved configuration is loaded again when the blade is rebooted. This command currently saves only the blade-management mode configuration. Syntax: Options:
tftp:<URL> copy system:running-config {tftp:<URL> | nvram:startup-config}
Saves the configuration to another system.

nvram:startup-config
Saves the configuration to the SCMs non-volatile memory. The saved configuration is reloaded when the SCM is rebooted.
102
copy nvram:startup-config
copy nvram:startup-config
Copies the persistently saved startup configuration to the specified URL, or reloads it so it becomes the running configuration. This command currently copies only the blademanagement mode configuration. Syntax: Options:
tftp:<URL> copy nvram:startup-config {tftp:<URL> | system:running-config}
Saves the startup configuration to another system.

system:running-config
Reloads the configuration to RAM, making it the current configuration.
erase nvram:startup-config
Erases the saved CLI configuration files from persistent storage. This returns aspects of the blade that are controlled through the CLI to the factory defaults. Syntax:
erase nvram:startup-config
This command currently only affects the blade-management mode configuration.
exit
Leaves the current CLI command mode, and goes to the next higher level. If you are already in the master CLI (at the top level), this is the same as the logout command, which exits the CLI entirely and logs out from the blade. Syntax:
exit
logout
Exits the CLI and logs out from the blade. This command works only from the master CLI (top level). Syntax:
logout
help
Shows the editing and history keys. This is similar to what is shown on page 101. Syntax:
help
103
104
BLADE MANAGEMENT COMMANDS
10
This chapter provides reference material for the commands within the blade-management CLI mode. These commands help you enable, disable, configure, and show status of various services that run on the SCM. The command types include: SNMP commands on page 106 Network timing subsystem (NTS) commands on page 106
Accessing blade-management command modes

To enter blade-management commands: 1. Log in to the SCM CLI. 2. Access blade-management mode by entering:
blade-mgmt
When the command reference pages below do not indicate a mode, use the top-level blademanagement mode.
Config
To access configuration mode from blade-management mode, enter:
config
The prompt changes each time you change modes. The command reference pages below indicate when to use the configuration mode.
105
10
Blade Management Commands
SNMP commands
service snmpd
Enables the SNMP master agent by enabling it to listen on the UDP port 161 (without any options or with the start option), or disables it by keeping it from listening on any port (stop option). The restart option causes the agent to reset itself by performing both a stop and start, deleting any log files and starting with the current configuration files. The reconfigure option forces the agent to re-read its configuration files and change its operation based on their contents, without fully re-initializing. Syntax:
service snmpd [ start | stop | restart | reconfigure ]
Tip To ensure seamless communication with the subagents, the snmpd daemon is always running, and the stop option does not stop the daemon. Do not check for running processes to determine whether the SNMP agent is running. Instead, use the show running-config command from the master CLI and check the output for the service snmpd start command. See also: Enabling the SNMP agent on page 64, SNMP Agent Support on page 89
Network timing subsystem (NTS) commands

The commands in this section configure the NTS behavior and show the current settings. For a general description of NTS, see Network timing subsystem (NTS) software on page 83.
Master/slave interactions for NTS configuration

When using two SCMs as an NTS master/slave pair, you should generally execute configuration commands on the SCM that is currently the master. The configuration is automatically copied to the running configuration of the partner SCM, with adjustments to the configured role of the SCM. For SCMs configured to use NTS autonomously, no configuration information is shared between them. Tip Save the configuration persistently by executing the copy system:running-config command on each SCM. To set and display the master/slave roles, see the commands network-clock-select role and show network-clocks brief.
106
Network timing subsystem (NTS) commands
10
Input clock definitions

The following are the network timing reference sources the SCM can accept as input. Clock
clk3a
Description The a half of the CLK3 bus, which is a backplane bus that is part of the PICMG-defined synchronization clock interface. A node blade can input a signal on CLK3. The b half of the CLK3 bus.
Frequency Input 8 kHz, 1.544 MHz (T1 and J1 modes only), 2.048 MHz (E1 mode only), or 19.44 MHz 8 kHz, 1.544 MHz (T1 and J1 modes only), 2.048 MHz (E1 mode only), or 19.44 MHz 8 kHz 8 kHz 8 kHz 6.48 MHz
clk3b
extclk1 extclk2 extclk3 extclk4 xlink1*
The first input from the rear transition module (RTM). The third input from the rear transition module (RTM). The fourth input from the rear transition module (RTM). The first update channel clock received from the partner SCM. Update channel signals between partner SCMs maintain synchronization by providing phase alignment and allows redundancy by providing quick detection of peer loss. The second update channel clock received from the partner SCM.
The second input from the rear transition module (RTM). 8 kHz
xlink2*
2 kHz
xlink1 and xlink2 here could be called xlink1(rx) and xlink2(rx) to indicate that they are received clocks. The same names are used for transmitted clocks, but the paths are electrically distinct. Both SCMs transmit and receive a pair of xlink clocks at all times.
Output clock definitions

The following are the network clock outputs the SCM can provide. Clock
clk1a clk1b clk2a clk2b
Description
Frequency Output
The a half of the CLK1 bus, which is a backplane bus that 8 kHz is part of the PICMG-defined synchronization clock interface. The b half of the CLK1 bus. 8 kHz The a half of the CLK2 bus, which is a backplane bus that 19.44 MHz is part of the PICMG-defined synchronization clock interface. The b half of the CLK2 bus. 19.44 MHz
107
10
sysclk1
The first output to the rear transition module (RTM). The outputs to the RTM are always being driven. The second output to the RTM. The third output to the RTM. SCM. Update channel signals between partner SCMs maintain synchronization by providing phase alignment and allows redundancy by providing quick detection of peer loss. The outputs to the update channel are always being driven.
2.048 MHz in E1 mode or 1.544 MHz in T1 or J1 mode 8 kHz 19.44 MHz 6.48 MHz
sysclk2 sysclk3
xlink1** The first update channel clock transmitted to the partner
xlink2** The second update channel clock to the partner SCM.
2 kHz
** xlink1 and xlink2 here could be called xlink1(tx) and xlink2(tx) to indicate that they are transmitted clocks. The same names are used for received clocks, but the paths are electrically distinct. Both SCMs transmit and receive a pair of xlink clocks at all times.
network-clock-select
Sets the priority of a clock input, from 2 to 7, with 2 indicating the highest priority. The SCM uses priorities to select a reference clock when the mode is configured to automatic (see network-clock-select mode). Each clock must have a unique priority. If clocks are set to the same priority, the last clock configured retains that priority, while the others are set to zero (ignored). Priority 1 is reserved for the update channel clock (xlink1) from the SCM in the NTS master role, used as the input by the slave SCM. Syntax: Mode: Defaults:
network-clock-select <priority> { clk3a | clk3b | extclk1 | extclk2 | extclk3 | extclk4 }
Config Source:
xlink1 clk3a clk3b extclk1 extclk2 extclk3 extclk4
Priority: 0 (ignored) if master, 1 if slave 2 3 4 5 6 7
See also:
show network-clocks priority, show network-clocks source priority, Input clock definitions
108
network-clock-select mode
10
network-clock-select mode
Selects the reference clock source. The automatic option selects the highest priority, validated clock, while the force mode with a clock specified selects a particular clock source. If a clock source specified in force mode is not valid, the SCM goes into digital holdover mode in the absence of a valid clock source. Syntax: Mode: Default: See also:
network-clock-select mode [ automatic | force { clk3a | clk3b | extclk1 | extclk2 | extclk3 | extclk4 } ]
Config automatic show network-clocks brief
no network-clock-select mode
Sets the reference clock source to automatic, which selects the highest priority clock. Syntax: Mode:
no network-clock-select mode
Config
network-clock-select role
Configures the SCM to act as the master, slave, or autonomous in relationship to the other SCM in the shelf. The master SCM provides the system clocks, while the slave stays synchronized with the master and outputs the same system clocks, if so configured, providing a seamless failover source. In the autonomous role, the SCMs do not communicate their timing or configuration information to each other, and the SCM may be under manual control of a management application such as a system manager. Syntax: Mode: Default: See also:
network-clock-select role [ master | slave | autonomous ]
Config the SCM in lower-numbered slot is the master, and the other SCM is the slave show network-clocks brief
no network-clock-select role
Sets the SCM roles to the default, making the SCM in the lower-numbered slot the master, and the other the slave. Syntax: Mode:
no network-clock-select role
Config
109
10
network-clock-select revertive
Sets the reference clock to revert to the original source if the signal is lost, the SCM fails over to another source, and the signal is restored. With revertive mode enabled, when a newly validated or revalidated source has a higher priority than the current reference source, the SCM transitions to the higher priority source. With revertive mode disabled, a source transition does not occur in that same situation, and the number of source transitions is minimized. Syntax: Mode: Default: See also:
network-clock-select revertive
Config Does not revert, stays with current source show network-clocks configuration
no network-clock-select revertive
Sets the reference clock to stay with the current source, even if the original source is restored after a failover. Syntax: Mode:
no network-clock-select revertive
Config
network-clock-select manage-bused-resource
Enables E-Key management of the clock bus resources (CLK1 and CLK2). With E-Key management enabled, the Shelf Manager responds to CLI or NTS API requests to drive a particular clock bus by granting or denying a token. Unless this SCM or its partner owns a token, the SCM cant enable output on a bus. With management enabled, the Shelf Manager tracks token ownership, and the running configuration reflects the ownership, allowing it to be saved persistently to the startup configuration. If the NTS master fails over to the slave, the SCM notifies the Shelf Manager that a new token owner is driving the bus, and the configuration value is automatically changed. With management disabled, the SCM can drive output to a clock bus without requesting a token. Syntax: Mode: Default: See also:
network-clock-select manage-bused-resource
Config Disabled network-clock-select request-bused-resource, show network-clocks bused-resource
no network-clock-select manage-bused-resource
Disables E-Key management of the clock bus resources. Syntax: Mode:
no network-clock-select manage-bused-resource
Config
110
network-clock-select request-bused-resource
10
network-clock-select request-bused-resource
Requests permission to drive a clock bus pair. The Shelf Manager provides a token to the entity authorized to drive the clock. A result message indicates whether the token was granted. If E-Key management of clock bus resources is enabled, token ownership is enforced when allowing output to be enabled, and ownership is reflected in the running configuration, allowing it to be saved persistently to the startup configuration. Even though one SCM requests permission to drive both halves of the bus, the SCMs can each drive half. Output can be enabled if the SCM or its partner SCM owns the token. See the example in network-clock-select output. Syntax: Mode: See also:
network-clock-select request-bused-resource {clk1 | clk2 | clk3}
Config network-clock-select manage-bused-resource, show network-clocks bused-resource
no network-clock-select request-bused-resource
Releases a token previously granted to the SCM to allow it to drive a clock bus pair. This frees the bus resource for use by another entity. Syntax: Mode:
no network-clock-select request-bused-resource { clk1 | clk2 | clk3 }
Config
network-clock-select output
Enables output of the PICMG-defined system clock signal on the selected clock bus. Selecting a or b enables output on only one of the bus pairs for the selected bus resource. Otherwise, both pairs are driven. When management of bused resources is enabled, output can be enabled only for buses for which this SCM or its partner owns the token. This command does not affect output to the RTM or to the update channel between SCMs. Output to the RTM and to the update channel is always enabled. Syntax: Mode: Default: Example:
network-clock-select output { clk1 | clk2 } [ a | b ]
Config No output To configure the partner SCMs to provide redundancy as described in System clock redundancy on page 85, use this setup: On the master SCM:
network-clock-select network-clock-select network-clock-select network-clock-select request-bused-resource clk1 request-bused-resource clk2 output clk1 a output clk2 a
111
10
On the slave SCM:

network-clock-select output clk1 b network-clock-select output clk2 b
Either SCM can request the bus tokens and share them with the other SCM. See also: show network-clocks
no network-clock-select output
Disables output of the system clock signal on the selected clock bus. Syntax: Mode:
no network-clock-select output { clk1 | clk2 } [ a | b ]
Config
network-clock-select framer
Sets the expected framing format of the signal received from the RTM framer clock. The format also affects the CLK3 input frequency and the sysclk1 output frequency. With the format set to e1, 2.048 MHz is the expected RTM and CLK3 input frequency and sysclk1 output frequency. With the format set to t1 or j1, 1.544 MHz is the expected frequency. Syntax: Mode: Default: See also:
network-clock-select framer [ e1 | t1 | j1 ]
Config e1 show network-clocks configuration
no network-clock-select framer
Resets the framing format expected from the framer clock to the default. Syntax: Mode:
no network-clock-select framer
Config
network-clock-select frequency
Specifies the source frequency for the specified CLK3 input clock. The 2048-kHz option is available only when the framing format is set to e1. The 1544-kHz option is available only when the framing format is set to t1 or j1. Syntax: Mode: Default: See also:
network-clock-select frequency { clk3a | clk3b } [ 8-kHz | 1544-kHz | 2048-kHz | 19.44-MHz ]
Config 19.44-MHz network-clock-select framer, show network-clocks source status
112
show network-clocks
10
no network-clock-select frequency
Sets the selected input clock back to the default frequency. Syntax: Mode:
no network-clock-select frequency { clk3a | clk3b }
Config
show network-clocks
Shows the frequency and enable status of each clock that the SCM may output. Syntax: Output:
Network Clock Output show network-clocks
The output clock name. The output clocks are described in Output clock definitions on page 107.
Output Frequency
The output frequency of each network clock.

Status Enabled
Whether the SCM is driving each clock output.
show network-clocks brief

Shows a summary of the current SCMs NTS role and operating status. Syntax: Output:
Mode show network-clocks brief
Whether NTS is using the defined priorities to automatically select the highest priority, valid input clock or whether a specific input selection was forced by a user. The current NTS role of the SCM. The possible values are master, slave, and autonomous. When configured as a master or slave, the value may not be what was configured, because an automatic failover may have changed the role. sources are described in Input clock definitions on page 107.
Role
Selected The source clock currently being used as a network timing reference. The Operating Mode
The mode of the PLL. The possible values are:

locked - The PLL is locked to a valid source. holdover - The PLL is operating based on stored phase information from
a previous source to maintain clock stability until it locks to another source.

freerun - The PLL phase is independent of any source.
113
10
pre-locked - After being in free-run mode, the SCM has selected a reference source and the PLL is attempting to lock to it. If the source is of good quality, the PLL enters locked state within 100 seconds. Otherwise, it reverts to free-run mode and the SCM selects another reference source. pre-locked2 - After being in holdover mode, the SCM has selected a
reference source and the PLL is attempting to lock to it. The PLL also enters this mode if the SCM is in revertive mode and is transitioning to a higher priority source that has been restored.
phase-lost - The PLL has lost phase lock and is attempting to lock to the input reference clock, if it exists. If the anomaly causing the loss of lock is serious, the SCM disqualifies the reference clock. The clock is automatically disqualified and generates an alarm if the PLL remains in phase-lost mode for more than 100 seconds. When the reference clock is disqualifies, the PLL proceeds to pre-locked2 mode if a valid reference clock is available. Otherwise, it goes into holdover mode. Operating State
The current alarms for the PLL. The possible values are:
t0-soft-alarm - The PLL is tracking its input reference clock beyond the
11.43 ppm drift limit that generates a soft alarm.

sync2k-alarm - The xlink2 input from the redundant SCM has a
problem.
RTM Present
Whether a rear transition module has registered its presence with the NTS API.
Partner Visible
Whether this SCMs master/slave partner is visible.
show network-clocks configuration

Summarizes the NTS configuration settings for the current SCM. Syntax: Output:
Mode Role show network-clocks configuration
Whether NTS is using the defined priorities to automatically select the input clock. Off indicates the forced selection of a user-specified input. The current NTS role of the SCM. The possible values are master, slave, and autonomous. The value may or may not be what was configured, because an automatic failover may have changed the role. Whether the selected clock source reverts to the original if the original source signal is lost and subsequently restored.
Revertive
114
show network-clocks priority
10
Output Framer Sa Bit SSM
The output clocks the SCM is currently driving. For source clock descriptions, see Input clock definitions on page 107. The framing format expected on RTM or CLK3 source clocks. The bit used to receive and transmit the synchronization status message (SSM). The synchronization status message to transmit. Whether the ability to drive the clock buses (CLK1 - CLK3) is restricted by ownership of E-Key tokens.
Manage EKey Token
Bused Resource
The clock bus E-Key tokens this SCM currently owns, allowing it (and its partner SCM) to drive clock signals on those buses.

Shows the configured priorities of the source network clocks. The highest priority is 1, with 0 (zero) being ignored. If the blade is in a forced mode, the priorities are ignored (see networkclock-select mode). Syntax: See also: Output:
network-clock-select For a description of the source clocks, see Input clock definitions on page 107.
show network-clocks source priority

Shows the actual priorities of the source network clocks, including the source currently used. The priorities shown here are based on configured priorities and clock validity status. The priorities may vary from the configured priorities because some inputs may not be valid. If the blade is in a forced mode, the priorities are ignored (see network-clock-select mode). Syntax: Output:
Selected show network-clocks source priority
The source clock currently in use.

1st Priority
The source with the highest priority. This may match the selected clock.
2nd Priority
The source with the 2nd priority.

3rd Priority
The source with the 3rd priority. See also: For source clock descriptions, see Input clock definitions on page 107.
115
10
show network-clocks source status

Shows the current status of the source network clock signals. Syntax: Output:
(Actual priority and usage) show network-clocks source priority
Briefly indicates the actual priorities and usage of the source clocks, also shown by show network-clocks source priority. The possible characters next to the corresponding clock are:
* - The validated input clock to which the PLL is currently locked. This
matches the first priority validated input clock (denoted by 1) when configured with revertive mode enabled. With revertive mode disabled, they may or may not match.
1 - The first priority validated input clock. 2 - The second priority validated input clock. 3 - The third priority validated input clock. Input Local Partner
The input source, described in Input clock definitions on page 107. The validity status of the input from the perspective of this SCM. The validity status of the input from the perspective of the partner SCM in the protection pair. In the case of the xlink1 and xlink2 clocks, the partner SCM reports on the validity of clocks it receives from the local SCM. The synchronization status message (SSM) value received on the input. The frequency of the input clock signal.
SSM Frequency
See also:
show network-clocks source priority
show network-clocks source alarms

Shows the current alarms for each source network clock. Not every alarm invalidates a source clock. An alarm is cleared when the signal improves. Syntax: Output: For each source clock (described in Input clock definitions on page 107), zero or more of the following alarms are displayed:
soft out-of-band show network-clocks source alarms
The source clock phase has drifted by more than 11.43 ppm. This alarm does not invalidate the source clock.
hard out-of-band
The source clock phase has drifted by more than 15.24 ppm. This alarm invalidates the source clock.
116
show network-clocks bused-resource
10
no-activity
The source is within the hard out-of-band limits, but shows no activity.
phase-lock
The PLL could not achieve phase lock on this selected source within 100 seconds. The SCM checks for phase lock only on a source currently selected and in use. Once the alarm has been raised and the SCM selects another source, the SCM is no longer checking for phase lock on this source. The alarm is cleared after 128 seconds.
peer-loss
The xlink1 or xlink2 update channel from the partner SCM no longer has a valid signal. If the slave SCM loses the masters xlink1 signal when it was previously valid, it causes a failover and becomes the master.
loss of frame
The SCM received a loss of frame condition for this source.

loss of signal
The SCM received a loss of signal condition for this source.

alarm indication
The SCM received an alarm indication signal condition (AIS) for this source.
yellow alarm
The SCM received a yellow alarm condition for this source.
show network-clocks bused-resource

Shows the E-Key token ownership of the clock buses and summarizes the last communications between this SCM and the Shelf Manager in regard to the tokens. Syntax: Output:
Bused Resource show network-clocks bused-resource
A backplane clock bus that is part of the PICMG-defined synchronization clock interface.
Token Owner
The current entity authorized to drive signals on this clock bus, according to the Shelf Manager. The possible values are:
Local - This SCM holds the E-Key token. Partner - The other SCM in this master/slave pair holds the token. Unknown - An unknown entity or no one holds the token. Board2Shelf
The last communication this SCM sent to the Shelf Manager regarding this token.
117
10
Shelf Reply
The last reply from the Shelf Manager to this SCM regarding this token, as a standard IPMI return code.
Shelf2Board
The last communication this SCM received from the Shelf Manager.
Board Reply
The last reply from this SCM to the Shelf Manager regarding this token.
118
MAINTENANCE
11
The SCM and the shelf are shipped with the jumper, connector, and pin configurations already set for the Promentum platform. Using two SCMs in the platform creates a redundancy that provides continued availability, even if a failure occurs in one of the modules. The Promentum platform includes interface management software from which the SCM can configure other platform supported modules and connect to external network systems. The interface management software is accessed during an initial installation through the serial port cable included with the SCM. The front panel 10/100/Base-T Ethernet connection can be used after the initial installation verification of interface operations. For additional information on the Ethernet connection and CLI commands, see the Ethernet Software Reference manual. WARNING This product contains static-sensitive components and should be handled with care. Failure to employ adequate anti-static measures can cause irreparable damage to components. Handle this product only when wearing a grounded wrist strap at a grounded work area. Additional precautions and ESD information are provided at www.radisys.com/esd.
Installing the SCM

For information on installing or replacing an SCM, please see the ATCA-2210 Switch and Control Module Installation Guide.
Removing the SCM

To remove the SCM from the shelf: 1. Disconnect all cables from the SCM front panel. 2. Loosen the SCM screws from the shelf. 3. Release the ejector latch in the position noted and then stop: If the SCM is oriented vertically, release the lower ejector latch. If the SCM is oriented horizontally, release the right ejector latch.
119
11
Maintenance
4. The H/S blue light starts flashing. When the H/S blue light turns solid blue you are ready to remove the SCM. Note When only one SCM is present in the shelf, the H/S blue light will never turn solid blue. 5. Release the ejector latch in the position noted: If the SCM is oriented vertically, release the top ejector latch. If the SCM is oriented horizontally, release the left ejector latch. 6. Simultaneously pull both extractor latches to release the module from the slot. 7. Pull the SCM out of the slot. 8. Place the SCM on a flat, static-free surface.
Troubleshooting
When you encounter a situation in which the Promentum platform does not perform as you expect, look for symptoms that might clarify the cause. Performing the following actions can aid you in diagnosing symptoms: Check the state of the LEDs on all the modules in the platform, especially the power entry modules (PEMs) and rear transition modules (RTMs). Check the shelf-management events logged in the system event log (SEL), which is accessible through the system manager. The Shelf Management Software Reference provides details on how to use the SEL. Information about the Base Ethernet configuration, which can be generated using the show commands in the base-interface management CLI. Information about the Fabric Ethernet configuration, which can be generated using show commands in the fabric-interface management CLI. Verify the IP address and the subnet mask assignment. See Chapter 5, Initial Verification and Configuration on page 59 for details on verifying these assignments. Verify the Fabric Ethernet interface is communicating properly. See the Ethernet Switching Software Reference for information on diagnosing communication and connection issues. Check temperatures on the module. The hpiapp application can be used to query the Alarm Table, which provides a way to query the actual temperature readings of devices. See the Shelf Management Software Reference for information on how to use the hpiapp application and how to perform these queries. Verify SCMs are of the same type and are using the same software version, when installing two on the same platform for redundancy purposes. For example, you cannot use an ATCA-2210 SCM with an ATCA-2100 SCM.
120
Troubleshooting
11
Symptoms and recommended actions

Look in this table for a symptom that applies to your situation, then follow the recommended action (or actions) for that symptom. When an action reveals the cause of the problem, resolve the problem as indicated. Table 32. Troubleshooting actions based on symptoms Symptom Recommendation The power LED on the SCM or another module is Verify that the module is fully inserted. not lit. Inspect the module connector pins for damage. If connector pins show no sign of damage, you can try the following: carefully insert the module into a different slot carefully insert a different module in the original slot Do not force modules during insertion. If insertion is not easy, the pins on the backplane connectors may be damaged, which could potentially damage the module connector pins. Check to see if the modules Shelf Manager is disabled and whether the module is waiting for an external Shelf Manager to enable it. Communication cannot be established between Connect to the SCM through the serial (LMP SER) or the Ethernet nodes within the platform. Ethernet maintenance (LMP ETH) ports. Use base-interface management CLI commands to diagnose the cause. In particular, verify that an IP address has been assigned to the node, an appropriate subnet mask is assigned to the node, the node is within a reachable VLAN, and routing instructions do not prevent the communication. If switch control CLI commands fail to work with one of the nodes, check the system event log to verify that the node's Ethernet backplane interface was activated. The SCM is in debug mode according to the login Restore valid Shelf FRU Information using the frurw utility as banner. described in the Shelf Management Software Reference. A module does not work correctly. Check the system event log for significant events related to the module. In particular, verify that the module worked correctly when it was installed, and look for any events since then that would account for why the module stopped working correctly. If the log information does not reveal useful symptoms, power down the module, slide it out, perhaps move it to a different slot, slide it back in, and power it back up. If appropriate, install a different module of the same kind to help determine whether the module might be defective. The SCMs embedded software is corrupted or it See Loading software through U-Boot on page 122 or fails to boot properly with newly upgraded Loading software from the secondary flash device on software. page 124.
121
11
Maintenance
Table 32. Troubleshooting actions based on symptoms (Continued) Symptom Intermittently, the SCM and other installed modules experience random data errors. Recommendation Verify that the platforms frame-ground connection is properly connected to a high-quality earth-ground connection. Check for electrical noise at the backplane power connections and at the power entry module power inputs. Consider the possibility of a malfunctioning module causing electrical noise on backplane connections. Verify no cover plates are installed. Empty slots must have air management panels installed to properly maintain airflow and emissions. Use shelf management software to check temperatures at the air intake, on the module, and at the platforms air exhaust. Use the information to determine whether the overheating may be caused by warm facility air, a module failure, or a failed fan module. Try moving the module to a different slot to see if that resolves the overheating. Verify there is at least two inches of clearance between the side of the shelf and the side of the rack cabinet. Check the air filter for obstructions and dirt.
The SCM or another module overheats
Loading software through U-Boot

If the SCMs software kernel becomes corrupted, you can load a software image using U-Boot if U-Boot is functioning. This procedure requires: A serial connection to the SCMs front panel serial port. A valid SCM software image (unzipped, with a file name ending in .bin) placed on an accessible TFTP server (for example in the /tftpboot directory). A network connection to the SCM front panel Ethernet maintenance port (LMP ETH). To load software using U-Boot: 1. Establish a serial connection to the SCM. See Serial connection on page 60. You may not see a prompt at this point, because the SCM has stopped its attempts to boot. Tip After the next step, you must enter a key combination within two seconds to break into U-Boot. Read the next steps carefully before performing them. If you dont enter the keystrokes in time, you can perform the steps again until you are successful. 2. Reboot the SCM by pressing the recessed reset button on the SCMs front panel.
122
Troubleshooting
11
3. At the prompt to enter U-Boot, enter these key combinations within 2 seconds: Ctrl-x Ctrl-w Tip If your terminal doesnt show any activity after pushing the reset button, the terminal may not be configured correctly. Check your terminal emulator settings again. If continued reboots show no activity, U-Boot isnt working properly. Read Loading software from the secondary flash device on page 124. 4. Specify the IP address of the TFTP server by entering: setenv serverip <server_ip_addr> 5. Configure the SCM Ethernet maintenance port to a subnet IP address accessible to the TFTP server by entering: setenv ipaddr <SCM_ip_addr> 6. Transfer the software image to the SCMs RAM by entering: tftp 1000000 <image_name> For example: tftp 1000000 /ATCA-2210/release1.7.13_scm3-wr.bin for the file /tftpboot/ATCA-2210/release1.7.13_scm3-wr.bin on the TFTP server.
WARNING U-Boot does not verify the software image you supply, so be sure you selected a valid image before proceeding. You are about to overwrite memory addresses, so type the next commands carefully. 7. Load the image from RAM to the primary flash device by entering: protect off 2:0-507 erase 2:0-507 cp 1000000 fc000000 fe0000 This loads everything in the software image except U-Boot. Preserving U-Boot allows you to try the procedure again if the image is corrupted or if the reload doesnt work. 8. Make the new software take effect by rebooting the SCM: reset If the software doesnt boot successfully, try the procedure again, or read Loading software from the secondary flash device.
123
11
Maintenance
Loading software from the secondary flash device

If the SCMs software in the primary flash device becomes corrupted, you can boot the SCM from the secondary flash device if the secondarys software is functioning. From there, you can load a new software image to the primary flash device. The procedure requires: Removing the SCM from the shelf and reinstalling it multiple times in order to add a jumper to particular pins and to remove it again. A valid SCM software image and the latest software upgrade (reflash) utility placed on an accessible TFTP server. If you need to perform this procedure, call RadiSys Technical Support for information.
124
SPECIFICATIONS
Environmental
RadiSys does not provide environmental certification testing, because any meaningful emissions agency certification must include the entire system. Thus, the SCM is designed and tested to pass the environmental specifications noted below, but is not certified. Please note the operating environment must provide sufficient airflow across the SCM to keep it within its temperature specification. Table 33. Environmental Characteristic Temperature (ambient) State Operating Short term operating 1 Storage Value +5 C to +45 C 30 C/hr rate of change 5C to +55 C 30 C/hr rate of change 40 C to +70 C Rate of change: 23 C to -40 C at 30 C/hr -40 C to 23 C at 13 C/min +23o C to +70o C at 30 C/hr +70 C to +23 C at 10 C/min 5% to 85% RH non-condensing 5% to 90% RH non-condensing at +30 C 5% to 90% RH non-condensing at +40 o C 5% to 95% RH non-condensing at +40 o C Up to 1800 meters (5,905 feet), +55C > 1800 meters up to 4000 meters (13,123 feet), derated linearly to +45C 0 to < 10 kg = 100 mm drop 10 to < 25 kg = 75 mm drop 0 to < 10 kg = 750 mm drop 10 to < 25 kg = 600 mm drop 300 mm free fall drop
Relative humidity Operating Short term operating1 Storage Altitude Short term storage1 Operating2
Shock (drop)
Unpacked Free fall, corners and edges Packaged (Unpalletized) Free fall, corners and edges Palletized
125
Specifications
Table 33. Environmental (continued) Characteristic Vibration State Operating Transportation (packaged) Value 0.1g, 5 to 100 Hz and back, 0.1 octave/min sine sweep 2 0.5g, 5 to 50 Hz and back, 0.1 octave/min sine sweep 2
3.0g, 50 to 500 Hz and back, 0.25 octave/min sine sweep 2 Seismic Operating Per Zone 4 test method, GR-63-CORE 1 Short term defined as a maximum of 96 consecutive hours. 2 For a maximum of eight consecutive hours.
Safety
The safety specifications are measured under laboratory ambient temperature and humidity (approximately 25C and humidity between 30% and 50%). Testing was performed in partnership with a Nationally Recognized Testing Laboratory (NRTL) accredited to provide the required certifications. Table 34. Safety Characteristic US Canada EU Other Certification Accessory Listing Approval Conformance with the Low Voltage Directive CB Report Standard and Test Criteria UL 60950-1 Safety for Information Technology Equipment CSA 22.2 #60950-1 Safety for Information Technology Equipment EN 60950-1 Safety for Information Technology Equipment IEC 60950-1 Safety for Information Technology Equipment
Mechanical dimensions
Table 35. Mechanical dimensions Characteristic PCB board Dimensions Board thickness Value 322.25mm x 280.0mm +0, -0.3mm (12.687 x 11.023 +0.0, -0.012) 2.0mm 0.2mm (0.079 0.007)
126
Electromagnetic compatibility (EMC)
Electromagnetic compatibility (EMC)

EMC and the immunity specifications are only measured under ambient temperature and humidity (at any point between 20 C to 30 C, and humidity at any point between 30% to 50%). Table 36. EMC emissions Characteristic Radiated emissions Conducted emissions State Operating Operating Standard and Criteria FCC Part 15, Class A EN 55022: 1998, Class A FCC Part 15, Class A EN 55022: 1998, Class A Table 37. EMC immunity Characteristic ESD State Operating Standard and Criteria EN 61000-4-2 8 KV direct contact, performance criteria B 15 KV air discharge, performance criteria C EN 61000-4-3 10 V/m, 30 MHz 10 GHz, 80% AM Performance criteria A EN 61000-4-4 0.5 kV, 5/50 ns, 5 kHz repetition frequency Performance criteria B EN 61000-4-5 Data ports - 1 kV, 1.2/50 s or 8/20 s DC power port 0.5 kV, 1.2/50 s or 8/20 s Performance criteria B EN 61000-4-6 0.15 80 MHz, 3 V, 80% AM Performance criteria A EN 61000-4-8 50 Hz / 1 A/m Performance criteria A
Radiated
Operating
Fast transient/burst
Operating
Surge voltages
Operating
Conducted
Operating
Magnetic field immunity
Operating
127
Specifications
Power
Table 38. Power emissions Characteristic Nominal operating voltage State -48V Standard and Criteria ETS 300 132-2 Static voltage levels
Table 39. Power consumption summary Supply Total converted from 12V DC/DC Conversion Loss (85%) COM Express 12V RTM 12V Total 12V brick power output Brick Conversion Loss (93%) IPMC and RTM 3.3V & 2.5V Total PIM output PIM Loss (97%)
Total -48V Power Required
Amp
4.47 2.80
Watt 86.9 15.3 26.0 33.6 161.8 12.2 3.0 177.0 5.5
182.5
128
MANAGED SENSORS
The SCM supports the following sensors, with sensor data record (SDR) support. The SCM implements a Type 0x12 record for the management controller and Type 0x01 records for the sensor descriptions. Type 0x12 is a management controller record and Type 0x01 is full sensor record. The following table provides a summary of sensors that are managed by the SCM. Table 40. Managed sensors Sensor Name # 1 ATCA FRU hot swap 2 ATCA physical IPMB Reading type ATCA FRU hot Sensorswap specific ATCA physical SensorIPMB-0 specific Type Normal reading 0x00 -0x07 0x00-0xFF Notes Return M0 to M7 ATCA hotswap states. Bit [7] = IMPB B Override state Bit [6:4] = IPMB B Local state Bit [3] = IPMB A Override state Bit [2:0} = IPMB A Local status 0 1 Eject latch is open Eject latch is closed Fault status asserted Identity status asserted 0 1 Power supply A Power supply A Presence detected Failure detected 0 1 Power supply B Power supply B Presence detected Failure detected 0 1 Both A & B fuses OK Either A or B fuse blown Presence detected Failure detected 0 1 No SEL overflow SEL overflow detected 0 1 IPMC watchdog not expired IPMC watchdog expired Hardware version number. Firmware is specific to each hardware version.
Eject open
Slot or connector Power supply
Digital
-48V absent A
Digital
-48V absent B
Power supply
Digital
-48V fuse fault
Power supply
Digital
7 8 9
SEL overflow IPMC WDOG TIMER HW Version
Event logging disabled System event
Digital Digital
0 0 0x00
Slot/connector Discrete
129
Managed Sensors
Table 40. Managed sensors (continued) Sensor Name # 10 +3.3_IPMI Type Voltage Reading Normal type reading Threshold Notes This sensor measures voltage in Volts. Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures temperature in C Default thresholds LC LNC UNC UC 0.0 0.0 55 75
LNR 0.0 11 +12V Voltage Threshold LNR 0.0 12 +5V Voltage Threshold LNR 0.0 13 +3.3V Voltage Threshold LNR 0.0 14 +2.5V Voltage Threshold LNR 0.0 15 +1.8V Voltage Threshold LNR 0.0 16 +1.25V Voltage Threshold LNR 0.0 17 +1.2V Voltage Threshold LNR 0.0 18 Board temp Temp Threshold 25 LNR 0.0
UNR
UNR
UNR
UNR
UNR
UNR
UNR
UNR
UNR 85
130
Managed sensors
Table 40. Managed sensors (continued) Sensor Name # 19 COM Express temp* Type Temp Reading Normal type reading Threshold 25 Notes This sensor measures temperature in C Default thresholds LNR LC LNC UNC UC UNR 0.0 0.0 0.0 55 75 85 This sensor measures temperature in C Default thresholds LNR LC LNC UNC UC UNR 0.0 0.0 0.0 55 75 85 0 1 PIM alarm not asserted PIM alarm asserted 0 1 COM Express present COM Express not present 0 1 COM Express power COM Express power Fail note asserted Fail asserted 000-111 Module type 0 1 Asserted Not asserted 0 No failover 1 Communication loss with active 2 Active not found / active 3 Active issued M5 4 Set per status command 5 Active / re-arbitrated 6 Forced failover 0x00 Pending 0x01 Standby0 0x02 Standby1 0x03 Standby2 0x04 Active 0 1 RTM is present RTM is not present 0 1 RTM latch is open RTM latch is closed 0 1 2C is not ready 2C is ready RTM I RTM I
20
Fabric/HD temp
Temp
Threshold 25
21 22 23
Power input module (PIM) temp
Power supply
Digital
0 or 1 0 or 1 0 or 1
COM Express present Module/board Digital COM Express power fail COM Express type PB reset Shelf management controller (ShMC) failover Slot or connector Slot or connector Slot or connector Digital
24 25 26
Discrete Digital
110 or 100 0 or 1 0
Sensor-specific Discrete
27
ShMC STATECHG
Sensor-specific Discrete
3 or 4
(ShMC)
2
(Non-ShMC)
28 29 30
RTM PRSNT# RTM eject open RTM I2C
Slot or connector Slot or connector Slot of connector
Digital Digital Digital
0 or 1 1 1
131
Managed Sensors
Table 40. Managed sensors (continued) Sensor Name # 31 RTM temp 0 Type Temp Reading Normal type reading Threshold 25 Notes This temperature measures in C Default thresholds LNR LC LNC UNC UC UNR 0.0 0.0 0.0 55 62 72 This temperature measures in C Default thresholds LNR LC LNC UNC UC UNR 0.0 0.0 0.0 55 62 72 0 1 RTM 12V and 3.3 power is RTM 3.3V or 12V power good failed 0 1 2 No pull-up Half pull-up Full pull-up 0 1 2 No pull-up Half pull-up Full pull-up
32
RTM temp 1
Temp
Threshold 25
33
RTM PWR Fail_12V_3V IPMB PU Ch A IPMB PU Ch B
Slot or connector Slot or connector Slot or connector
Digital
34 35
Discrete Discrete
TBD 0 (non-ShMC) 2 (ShMC)
Table 41. Managed sensorsevent generation and board health Sensor Name # 1 2 3 4 5 6 7 8 9 10 11 12 13 ATCA FRU hot swap ATCA physical IPMB Eject open -48V absent A -48V absent B -48V fuse SEL overflow IPMC watchdog timer HW version +3.3V IPMI +12V +5V +3.3V Board healthy Local check scanning No No No No No No No No No No No No No Local event generation disabled No No No No No No No No No No Varies* Varies* Varies* Remote event Remote scanning generation disabled disabled No No No No No No No No No No No No No No No No No Varies* No No No Varies* No No Varies* Varies* Alarm level (if applicable) N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Minor, Major, Critical Minor, Major, Critical Minor, Major, Critical
132
Managed sensorsevent generation and board health
Table 41. Managed sensorsevent generation and board health (continued) Sensor Name # 14 15 16 17 18 19 20 21 22 23 +2.5 +1.8V +1.25V +1.2V Board temp COM Express temp Fabric/HD temp Board healthy Local check scanning No No No No No No No Local event generation disabled Varies* Varies* Varies* Varies* No No No Remote event Remote scanning Alarm level generation disabled (if applicable) disabled No Varies* Minor, Major, Critical No Varies* Minor, Major, Critical No Varies* Minor, Major, Critical No Varies* Minor, Major, Critical No No Minor, Major, Critical No No Minor, Major, Critical No No Minor, Major, Critical No No N/A No Varies* N/A No Varies* N/A No No No No No No No No No No No No Varies No No No Varies** Varies** Varies** Varies** Varies* Varies* No No N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A
PIM status No No COM Express present No Varies* COM Express power No Varies* fail 24 COM Express Type No Varies* 25 PB reset No No 26 ShMC failover No No 27 ShMC STATECHG No No 28 RTM PRSNT# No Varies** 29 RTM eject open No Varies** 2C ready 30 No Varies** RTM 1 31 RTM temp0 No Varies** 32 RTM temp1 No Varies* 33 RTM PWR No Varies* Fail_12V_3V 34 IPMB PU Ch A No No 35 IPMB PU Ch B No No * Event generation is enabled when the payload power is turned on. ** Event generation is enabled if an RTM is detected.
133
Managed Sensors
134

02376-00 ATCA-2210 SCM Reference

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

02376-00 ATCA-2210 SCM Reference

Transféré par

Droits d'auteur :

Formats disponibles

PROMENTUM

SWITCH AND CONTROL MODULE

007-02376-0000 December 2006

Chapter 1: Features Overview...................................................................................................................... 11

Chapter 2: Theory of Operation................................................................................................................... 13

Chapter 4: Physical Interfaces ...................................................................................................................... 45

Chapter 5: Initial Verification and Configuration ....................................................................................... 59

Chapter 6: Software Features....................................................................................................................... 71

E-Key authorization and notification..................................................................................................................................

Chapter 7: SNMP Agent Support ................................................................................................................. 89

Chapter 8: API Overview............................................................................................................................... 95

Chapter 9: Using the Command Line Interface ......................................................................................... 99

Chapter 10: Blade Management Commands ............................................................................................. 105

Chapter 11: Maintenance.............................................................................................................................. 119

Appendix A: Specifications............................................................................................................................ 125

Appendix B: Managed Sensors .................................................................................................................... 129

About this manual

A command or keyword. Screen text and syntax strings. Variable parameters.

All numbers are decimal unless otherwise stated.

Where to get more product information

About related RadiSys Products

Where to get more product information

Local management processor

Network timing subsystem

Intelligent Platform Management Controller

Possible applications of the SCM

Figure 1. SCM system block diagram

Node Slots E1/DS1 Framer

RJ45 RJ45 RJ45 RJ45

SFP SFP SFP SFP

Node 1Gb-Tx Slots 12-14 Zone 2 IPMB 48V Zone 1

Serial Local Power Conversion 12V -48V Power Conversion

Blue = Base Ethernet Red = Fabric Ethernet

The SCM as a Shelf Manager

Intelligent Platform Management

Shelf display or alarm panel

(Modules in some slots not shown)

(Modules in some slots not shown)

Base Ethernet interface

(Modules not shown in some slots)

(Modules not shown in some slots)

Uplinks Slot 1 Slot 2 Slot 7

Uplinks Slot 8 Slot 13 Slot 14

Fabric Ethernet interface

SAS/SATA Hard Drive*

12V Clock/ calendar 3.3V

Switch Switch I2C Buf Buf PCI-E Zone 3 Conn

Local Power Conversion

SFP 1Gb SD SFP

SPI ClkIn RefClk 1Gb SD

1Gb-Tx Sync Clk

PQIII Local Management Processor Mag

BITS/ SSU Clk BITS/ SSU Clk

Mag I/F Mag I/F Mag I/F Mag I/F

10 Gb XAUI 10 Gb XAUI TBI X 20-port 10Gb Eth Switch X 2

BITS/ SSU Clk BITS/ SSU Clk

14 PCI-32 SPI 2 FE Switch 1 2

1Gb SD SFP 10Gb XAUI 10Gb XAUI SFP 2 1Gb SFP X 4 2

ShMC 10/100 -Tx Base Ch1 2

14 1Gb SGMII COM -E Type Det

16-port Gb Eth PHY

Node Slots F-Link Base Ch2

3 Local management processor