Intel® 460GX Chipset System
Software Developer’s Manual
June 2001
Document Number: 248704-001
Contents
1
Introduction......................................................................................................................1-1
1.1
1.1.1 Component Overview............................................................................1-2
I/O Support.........................................................................................................1-4
1.5.1 PXB Features........................................................................................1-4
1.5.2 WXB Features.......................................................................................1-6
1.5.3 GXB Features........................................................................................1-6
1.7.1 I/O & Firmware Bridge (IFB)..................................................................1-6
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2
Register Descriptions ......................................................................................................2-1
2.1
2.2
Access Restrictions............................................................................................2-2
2.2.1 Partitioning ............................................................................................2-2
2.2.2 Register Attributes.................................................................................2-3
2.2.5 Default Upon Reset...............................................................................2-3
Error Handling Registers....................................................................................2-5
2.4.1 SAC.......................................................................................................2-5
2.4.3 MAC ....................................................................................................2-21
2.4.4 PXB .....................................................................................................2-22
2.4.6 WXB ....................................................................................................2-27
2.5.1 SAC.....................................................................................................2-30
2.5.3 PXB .....................................................................................................2-36
2.5.5 WXB ....................................................................................................2-43
2.6.1 SAC.....................................................................................................2-44
2.6.3 PID Indirect Access Registers.............................................................2-46
2.3
2.4
2.5
2.6
Intel® 460GX Chipset System Software Developer’s Manual
iii
3
System Architecture........................................................................................................3-1
3.1
3.1.1 Processor Coherency............................................................................3-1
3.1.2 PCI Coherency......................................................................................3-2
3.1.3 AGP Coherency ....................................................................................3-2
3.6.1 Processor Locks....................................................................................3-4
3.6.2 Inbound PCI Locks................................................................................3-5
3.6.3 Atomic Writes........................................................................................3-5
3.6.4 Atomic Reads........................................................................................3-5
3.8.1 Slot Power-up and Enable ....................................................................3-7
3.2
3.3
3.4
3.5
3.6
3.7
3.8
4
System Address Map......................................................................................................4-1
4.1
4.1.1 Compatibility Region .............................................................................4-1
4.2
4.3
4.4
5
Memory Subsystem ........................................................................................................5-1
5.1
Organization.......................................................................................................5-1
5.1.1 DIMM Types..........................................................................................5-3
5.5.1 Auto Detection.......................................................................................5-6
5.5.2 Removing a Bad Row ...........................................................................5-6
5.5.3 Hardware Initialization...........................................................................5-7
5.5.4 Memory Scrubbing................................................................................5-7
5.2
5.3
5.4
5.5
6
Data Integrity and Error Handling...................................................................................6-1
6.1
6.1.1 System Bus...........................................................................................6-1
6.1.3 Expander Buses....................................................................................6-2
6.1.4 PCI Buses .............................................................................................6-2
iv
Intel® 460GX Chipset System Software Developer’s Manual
Data Poisoning...................................................................................................6-3
Usage of First-error and Next-error....................................................................6-3
6.4.1 Masked Bits...........................................................................................6-4
6.4.2 BERR#/BINIT# Generation ...................................................................6-4
6.4.5 XSERR#................................................................................................6-5
6.5.1 Data ECC or Parity Errors.....................................................................6-5
6.5.2 System Bus Errors ................................................................................6-6
6.5.3 SAC to SDC Interface Errors.................................................................6-6
6.5.4 SAC to MAC Interface Errors ................................................................6-7
6.5.6 SDC/System Bus Errors........................................................................6-8
6.5.7 SDC Internal Errors...............................................................................6-8
Error Determination............................................................................................6-8
6.6.1 SAC Address on an Error......................................................................6-9
6.6.2 SDC Logging Registers.......................................................................6-10
6.7.1 SAC/SDC Error Clearing.....................................................................6-11
6.8.1 SDC Multiple Errors.............................................................................6-12
6.8.2 SAC Multiple Errors.............................................................................6-13
6.8.4 Error Anomalies...................................................................................6-13
6.10.1 Table of Errors.....................................................................................6-15
6.11.1 PCI Bus Monitoring .............................................................................6-20
6.11.2 PXB as Master ....................................................................................6-20
6.11.3 PXB as Target.....................................................................................6-21
6.11.4 GXB Error Flow ...................................................................................6-22
6.12.2 Data Parity Poisoning..........................................................................6-26
6.12.4 Error Mask Bits....................................................................................6-27
6.12.5 Error Steering/Signaling ......................................................................6-27
6.12.6 INTRQ# Interrupt.................................................................................6-29
6.12.8 Error Conditions ..................................................................................6-30
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
6.10
6.11
6.12
7
AGP Subsystem..............................................................................................................7-1
7.1
7.1.1 GART Implementation...........................................................................7-3
7.1.2 Programming GART..............................................................................7-4
7.1.3 GART Implementation...........................................................................7-5
7.1.4 Coherency.............................................................................................7-5
7.1.5 Interrupt Handling..................................................................................7-6
Intel® 460GX Chipset System Software Developer’s Manual
v
7.2
7.3
7.2.2 Traffic Priority........................................................................................7-7
7.2.4 Ordering Rules......................................................................................7-8
7.2.7 PCI Semantics Traffic ...........................................................................7-9
7.3.1 Inbound Read Prefetching ..................................................................7-14
7.4
7.5
8
WXB Hot-Plug.................................................................................................................8-1
8.1
IHPC Configuration Registers............................................................................8-1
8.1.3 DID: Device Identification Register........................................................8-3
8.1.5 PCISTS: PCI Status Register................................................................8-5
8.1.7 CLASS: Class Register.........................................................................8-6
8.1.8 CLS: Cache Line Size...........................................................................8-6
8.1.10 HDR: Header Register ..........................................................................8-6
8.1.11 Base Address........................................................................................8-7
8.1.13 SID: Subsystem ID................................................................................8-7
8.1.14 Interrupt Line.........................................................................................8-7
8.1.15 Interrupt Pin...........................................................................................8-8
8.1.16 Hot-Plug Slot Identifier ..........................................................................8-8
8.1.18 Hot-Plug Features.................................................................................8-9
8.1.20 Power Fault SERR Status.....................................................................8-9
8.1.21 Arbiter SERR Status ...........................................................................8-10
8.1.22 Memory Access Index.........................................................................8-10
8.2.2 Slot Enable..........................................................................................8-12
8.2.3 Hot-Plug Miscellaneous ......................................................................8-13
8.2.4 LED Control.........................................................................................8-13
8.2.6 Hot-Plug Interrupt Mask ......................................................................8-15
8.2.7 Serial Input Byte Data .........................................................................8-16
8.2.8 Serial Input Byte Pointer .....................................................................8-17
8.2.9 General Purpose Output .....................................................................8-17
8.2.10 Hot-Plug Non-interrupt Inputs .............................................................8-17
8.2.11 Hot-Plug Slot Identifier ........................................................................8-17
8.2.13 Slot Power Control ..............................................................................8-18
8.2
vi
Intel® 460GX Chipset System Software Developer’s Manual
9
IFB Register Mapping......................................................................................................9-1
9.1
9.2
9.3
9.4
IDE Configuration...............................................................................................9-3
10
IFB Usage Considerations ............................................................................................10-1
10.5.1 UDMAC–Ultra DMA Control Register (IFB Function 1 PCI
10.5.2 UDMATIM–Ultra DMA Timing Register (IFB Function 1 PCI
10.5.5 Determining a Drive’s Best Multi Word DMA/Single Word DMA
10.5.6 IFB Timing Settings.............................................................................10-9
10.5.8 Settings Checklist..............................................................................10-13
10.5.9 Example Configurations ....................................................................10-14
10.5.11 Additional Ultra DMA/PCI Bus Master IDE Device Driver
11
LPC/FWH Interface Configuration.................................................................................11-1
11.1.9 ACPI Enable (Function 0)....................................................................11-4
11.1.10 SCI IRQ Routing Control.....................................................................11-4
Intel® 460GX Chipset System Software Developer’s Manual
vii
11.1.19 DDMABP–Distributed DMA Slave Base Pointer
11.1.22 GPIO Enable (Function 0).................................................................11-10
11.1.27 LPC Enables (Function 0).................................................................11-13
PCI to LPC I/O Space Registers....................................................................11-15
11.2.1 DMA Registers..................................................................................11-15
11.2.2 Interrupt Controller Registers............................................................11-20
11.2.3 Counter/Timer Registers...................................................................11-25
11.2.4 NMI Registers ...................................................................................11-28
11.2.5 Real Time Clock Registers................................................................11-29
11.2.7 ACPI Registers..................................................................................11-31
11.2.8 SMI Registers....................................................................................11-35
11.2
IDE Configuration..........................................................................................................12-1
12.1
12.2
12.2.7 BMIBA–Bus Master Interface Base Address Register
(Function 1).........................................................................................12-4
12.3
Universal Serial Bus (USB) Configuration.....................................................................13-1
13.1
13.2
viii
Intel® 460GX Chipset System Software Developer’s Manual
13.2.13 INTPN–Interrupt Pin (Function 2)........................................................13-6
14
SM Bus Controller Configuration...................................................................................14-1
14.2.11 INTPN–Interrupt Pin (Function 3)........................................................14-5
Intel® 460GX Chipset System Software Developer’s Manual
ix
PCI/LPC Bridge Description..........................................................................................15-1
15.1
15.1.1 Transaction Termination .....................................................................15-1
15.1.2 Parity Support .....................................................................................15-1
15.1.3 PCI Arbitration.....................................................................................15-1
15.2.2 End of Interrupt Operation...................................................................15-3
15.2.3 Modes of Operation.............................................................................15-4
15.2.4 Cascade Mode....................................................................................15-5
15.2.6 Interrupt Masks ...................................................................................15-6
15.2.8 Interrupt Steering ................................................................................15-7
15.3.1 Protocol...............................................................................................15-8
15.5.1 RTC Registers and RAM...................................................................15-14
15.5.2 RTC Update Cycle ............................................................................15-17
15.5.3 RTC Interrupts...................................................................................15-17
15.5.4 Lockable RAM Ranges .....................................................................15-17
15.2
15.3
15.4
15.5
IFB Power Management ...............................................................................................16-1
16.1
16.2
16.2.1 Power Plane Descriptions...................................................................16-2
16.2.2 SMI# Generation.................................................................................16-2
16.2.3 SCI Generation ...................................................................................16-3
16.2.4 Sleep States........................................................................................16-3
16.3
Figures
1-1
4-1
4-2
4-3
4-4
System Memory Address Space........................................................................4-2
System I/O Address Space................................................................................4-6
System Memory Address Space as Viewed from an Expander
GXB Error Flow................................................................................................6-25
GART Table Usage for 4 MB Pages..................................................................7-2
5-1
5-2
6-1
6-2
6-3
7-1
7-2
7-3
x
Intel® 460GX Chipset System Software Developer’s Manual
7-4
7-5
Tables
1-1
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
2-9
2-10
4-1
5-1
5-2
5-3
5-4
6-1
6-2
Intel® 460GX Chipset Components ...................................................................1-2
I/O Window Register Format............................................................................2-46
I/O (x)APIC Version Register Format...............................................................2-49
List of WXB Error Sources Selectively Routable to XBINIT#, SERR_OUT#,
Identify Device Information Used for Determining Ultra DMA
6-3
6-4
6-5
7-1
7-2
7-3
7-4
7-5
8-1
8-2
9-1
9-2
9-3
9-4
10-2
10-3
10-4
Identify Device Information Used for Determining Multi/Single Word DMA
Drive Multi Word DMA/Single Word DMA Capability as a Function
Intel® 460GX Chipset System Software Developer’s Manual
xi
10-12 PIO Transfer/Mode Values.............................................................................10-12
10-13 Drive Capabilities Checklist............................................................................10-13
10-14 IFB Settings Checklist....................................................................................10-14
12-2
12-3
12-4
13-2
15-1
15-2
16-1
16-2
16-3
16-4
PCI Configuration Registers–Function 2..........................................................13-1
IFB Power States and Consumption................................................................16-1
Causes of SCI#................................................................................................16-3
ACPI Bits Not Implemented in IFB...................................................................16-4
xii
Intel® 460GX Chipset System Software Developer’s Manual
Introduction
1
This document provides information about the Intel® 460GX chipset components. The 460GX
chipset is a high performance memory and I/O chipset for the Intel Itanium™ processor, targeted
for multiprocessor server and high-end workstation designs.
This document describes the software programmer's interface to the 460GX chipset. It provides a
brief summary of the system architecture supported by the 460GX chipset, a list of features within
the chipset and a detailed description of software or other externally visible segments.
1.1
System Overview
The Intel 460GX chipset is a high performance chipset for Intel Itanium processor-based systems,
targeted for multiprocessor servers and high-performance workstations. It provides the memory
controller interface and appropriate bridges to PCI, AGP 4X, and other standard I/O buses.
Figure 1-1 illustrates the basic system configuration of a four-processor platform.
Figure 1-1. Diagram of a Typical Intel® 460GX Chipset-based System with AGP
Cache
Cache
Cache
Cache
Processor
Processor
Processor
Processor
Itanium™Processor System Bus
Private Bus
SAC
SystemAddress
Controller
SDC
SystemData
Controller
MAC
MAC
MDC
MDC
MDC
MDC
MDC
MDC
MDC
MDC
Memory Subsystem
MAC
Memory Subsystem
Expander
Buses
Memory Data and Control Bus
GXB
Graphics
Expansion
Bridge
PXB
PCI
Expansion
Bridge
PID
Programmable
Interrupt Device
USB
Compatibility
PCI Bus
IDEHDD
IDECD-ROM
IFB
I/Oand Firmware
Bridge
AGP
4X Mode
Slot
GART
SRAM
Firmware
Hub Interface
LPC Interface
2 PCI Buses
3.3V/5.0V,
32-bit, 33 MHz
FWH
Firmware
Hub
Super
I/O
WXB
Wide PCI
Expansion
Bridge
2 PCI Buses
3.3V, 64-bit, 66MHz
000346e
Intel® 460GX Chipset Software Developer’s Manual
1-1
Introduction
1.1.1
Component Overview
Table 1-1 lists the 460GX chipset components.
Table 1-1. Intel® 460GX Chipset Components
Component
Name
Function
SAC
82461GX
Interfaces the address and control portion of the Itanium™ processor
system bus and the memory bus. Acts as a host bridge interface to
peripheral I/O devices through four Expander busses.
System Address
Controller
SDC
MAC
MDC
GXB
WXB
PXB
82462GX
Interfaces the data portion of the Itanium processor system bus and
the memory bus.
System Data path
Controller
82463GX
Provides the SDRAM RAS/CAS/WE/CS generation as well as
redriving the address to the SDRAMs. It is capable of buffering several
commands from the SAC.
Memory Address
Controller
82464GX
Multiplexes the data from the SDRAM to the SDC. On reads, it latches
data from the SDRAM, then transfers the data to the SDC. On writes, it
latches the data from the SDC, then writes the data to the SDRAM.
Memory Data path
Controller
82465GX
Provides the control and data interface for an AGP 4X graphics port.
This device attaches to the SAC via two Expander busses which utilize
a special configuration.
Graphics Expander
Bridge
82466GX
Provides the primary control and data interface for two independent
64-bit, 66 MHz PCI interfaces. This device attaches to the SAC via an
Expander bus.
Wide and fast PCI
Expander Bridge
82467GX
Provides the primary control and data interface for two independent
32-bit, 33-MHz PCI interfaces. These two 32-bit interfaces may
operate together to produce a single 64-bit, 33-MHz interface via a
configuration option. This device attaches to the SAC via an Expander
bus.
PCI Expander Bridge
IFB
82468GX
The IFB is a multi-function PCI device implementing a PCI-to-LPC
bridge function, a PCI IDE function, a Universal Serial Bus Host/Hub
function, an SMBus Interface function, Power Management function
and the Firmware Hub interface.
I/O and Firmware
Bridge
FWH
PID
82802AC
The FWH component interfaces to the IFB component and provides
firmware storage and security features. Further FWH information can
be found at http://developer.intel.com/design/chipsets/datasheets or by
ordering document 290658.
Firmware Hub 8Mb
NEC# UPD66566S1- The PID is an interrupt controller that provides interrupt steering
016
functions. The PID contains the logic required to support 8259A mode,
APIC mode, and SAPIC mode interrupt controller operations. The PID
interfaces include a PCI bus interface, an APIC bus interface, a serial
IRQ interface, and an interrupt input interface.
Programmable
Interrupt Device
1-2
Intel® 460GX Chipset Software Developer’s Manual
Introduction
1.2
Product Features
•
High performance hardware based on IA-64
architecture
•
High bandwidth system bus for multiprocessor
scalability
— 4.2 GB/s memory bandwidth can
simultaneously support both the full system
bus and the full I/O bus bandwidths
— Architectural support for 64 MB to 64 GB of
SDRAM
— Support of the Intel® Itanium™ processor
64-bit data bus
— Full support for 4-way multiprocessing
— 266 MHz data bus frequency
— Cache line size of 64 bytes
— Support for up to four bridge chips that
interface to the 82461GX (SAC) through
four Expander channels, each 30 bits wide
and providing 533 MB/s peak bandwidth
— AGP 4X compatible, via the 82465GX
(GXB) and two Expander channels running
at 266 MHz totaling 1 GB/s peak bandwidth
— Support for two 64-bit, 66-MHz PCI buses
using one 82466GX (WXB) component per
Expander channel
— Support for two independent 32-bit, 33-MHz
PCI buses or one 64-bit, 33-MHz PCI bus via
the 82467GX (PXB) per Expander channel
— Data streaming support between Expanders
and DRAM, up to 533 MB/s per Expander
channel
— Enhanced defer feature for out-of-order data
delivery using IDS#
— AGTL+ bus driver technology
Features to support flexible platform
environments
— Hardware compatible with IA-32 binaries
— AGP address space up to 32 GB supported
— Support for Auto Detection of SDRAM
memory type and mixed memory sizes
allowed between rows
•
— Supports 16-, 64-, 128- and 256-Mbit
DRAM devices
— Full support for the PCI Configuration Space
Enable (CSE) protocol to devices on all
Expander channels
— WXB supports 3.3 volt PCI bus operation
(supports universal and 3.3 volt PCI cards)
and has an Integrated Hot-Plug Controller**
— PCI Rev. 2.2 compliant on the WXB and
PXB
•
Extensive RAS features for mission-critical
needs
— ECC protection on the system bus data
signals
— Memory ECC with single-bit error
correction, double and nibble error detection
— Address and data flows protected by parity
throughout chipset
— ECC bits in DRAM accessible by diagnostics
— Fault recording of multiple errors; sticky
through reset
— GXB supports fast writes and 1x, 2x and 4x
data rates
— 1 MB or greater of firmware storage
provided by the 82802AC (FWH)
— Interrupt controller, bus-mastering IDE and
Universal Serial Bus supported by the
82468GX (IFB)
— JTAG TAP port for debug and boundary scan
— Support of 8259A mode, APIC mode and
capability
SAPIC mode interrupts via the
— I2C slave interface for viewing and
modifying specific error and configuration
registers
UPD55566S1-016 (PID) provided by NEC*
— Bus, memory and I/O performance counters
— Support of ACPI/DMI functions (support is
provided in the IFB)
**Based on technology licensed from Compaq Computer Corp.
1.3
Itanium™ Processor System Bus Support
• Full support for the Itanium processor system bus.
— 64-bit data bus.
— 266 MHz data bus frequency.
— Cache line size of 64 bytes.
— Supports SAPIC interrupt protocol.
• Full support for 4-way multiprocessing.
• Parity protection on address and control signals, ECC protection on the data signals.
• GTL+ bus driver technology.
Intel® 460GX Chipset Software Developer’s Manual
1-3
Introduction
1.4
DRAM Interface Support
• SDRAM 3.3 volt, 168-pin DIMM’s are the only memory type supported.
• Support for 64 MB to 64 GB of DRAM.
• Minimum memory size is 64 MB using 16 MB DIMM’s.
• Minimum incremental size is 64 MB using 16 MB DIMM’s.
• Maximum memory size is 16 GB using 128 MB DIMM’s.
• Maximum memory size is 64 GB using 1 GB DIMM’s.
• Only 3.3 volt memory is supported.
• Support for Auto Detection of SDRAM Memory Type.
• Supports 16, 64, 128 and 256 Mbit DRAM devices.
• Mixed memory sizes allowed between rows.
• Staggered CAS-before-RAS refresh (standard SDRAM refresh).
• ECC with single-bit error correction, double and nibble error detection.
• Extensive processor-to-Memory and PCI-to-Memory write data buffering, thus minimizing
the interference of writes on read latency.
1.5
I/O Support
• 4 Expander ports, each 30 bits wide and providing 533 MB/s peak bandwidth.
• Each Expander bus supports a single PXB or WXB. Two Expander busses can be configured
to support a GXB.
• Full support for the PCI Configuration Space Enable (CSE) protocol to devices on all
Expander ports.
• Data streaming support between Expanders and DRAM, up to 533 MB/s per Expander port.
• All outbound memory and I/O reads (except locked reads) are deferred.
• All outbound memory space writes are posted. Outbound I/O space writes are optionally
posted (unless targeting an address with side effects, in which case they are deferred).
• All inbound memory reads are delayed.
• All inbound memory space writes are posted.
• Supports concurrent processor and I/O initiated transactions to main memory.
• Maintains coherency with processors by snooping all inbound transactions to the system bus.
• Supports non-coherent traffic (for AGP), with a direct path to memory bypassing the system
bus.
1.5.1
PXB Features
• Can be configured to provide two independent 32 bit, 33 MHz PCI buses or one 64 bit, 33
MHz PCI bus.
• PCI Rev. 2.2, 5V tolerant (PXB drives 3.3 volts, but is 5.0 volt tolerant).
1-4
Intel® 460GX Chipset Software Developer’s Manual
Introduction
• Parity protection on all PCI signals.
• Data collection & write assembly.
— Combines back-to-back sequential processor-to-PCI memory writes to PCI burst writes.
— Processor to PCI write assembly of full/partial line writes.
• Two outbound read requests containing a total of two cache lines of read data for each PCI bus.
• Supports six outbound write requests containing a total of three cache lines of write data for
each 32 bit PCI bus. Supports 12 outbound write requests containing a total of six cache lines
of write data for a 64 bit PCI bus.
• Supports two delayed inbound read requests.
• Supports the I/O and Firmware Bridge (IFB).
• Supports either internal or external arbitration, allowing additional bus masters, on the PCI
bus.
1.5.2
1.5.3
WXB Features
• Support for two 64 bit, 66 MHz PCI busses.
• 3.3 Volt PCI bus operation (supports Universal and 3.3 Volt PCI cards).
• PCI Specification, Revision 2.2.
• Integrated Hot-Plug controller.
GXB Features
• The GXB is AGP and AGP 4X mode compatible, nominal 66 MHz, 266 MHz, 1 GB/s peak
bandwidth.
• The GXB supports pipelined operation or sideband signals on AGP 4X mode bus.
• AGP address space of 1 GB or 256 MB supported. Also supports 32 GB of GART window, if
4 MB pages are used.
• Supports Fast Writes and 1x, 2x and 4x data rates.
1.6
RAS Features
• ECC coverage of system data bus using the Itanium™ processor SEC/DED ECC code.
Memory is protected using a SEC/DED code which also provides nibble detection capabilities
of 4 bits. All control and address signals are parity protected. Local control buses are parity
protected. The Expander is covered by parity.
• Data flows protected by parity throughout chipset.
• ECC bits in DRAM accessible by diagnostics.
• Fault recording of multiple errors; sticky through reset, but NOT through power-down.
• Memory scrubbing implemented in hardware.
• Boundary test capability through JTAG.
• JTAG TAP port for debug.
Intel® 460GX Chipset Software Developer’s Manual
1-5
Introduction
• I2C Slave Interface will allow viewing and modifying of specific error and configuration
registers.
1.7
Other Platform Components
These 460GX devices provide access to flash space, interrupt collection and legacy features.
1.7.1
I/O & Firmware Bridge (IFB)
The 460GX chipset is designed to work with the IFB south bridge. As part of this support, the PXB
includes an internal PCI arbiter as well as support for an external PCI arbiter. The IFB consists of
an 8259C Interrupt controller, a bus-mastering IDE interface, and a Universal Serial Bus interface.
Devices using IFB are limited to a 32 bit addressing space available for DMA, not the full 44 bits
supported by the Itanium™ processor.
1.7.2
Programmable Interrupt Device (PID)
The PID is a PCI device that gathers interrupts and delivers them from the PCI bus to the system
bus using the SAPIC interrupt protocol. The interrupt will be presented to one of the processors on
the bus for servicing. A 460GX chipset based platform requires at least one PID located on the
compatibility PCI bus. The compatibility PID will handshake with the IFB before delivering a
south bridge/compatibility device interrupt. The same PID may also be used to deliver some
portion of the PCI based interrupts.
The system implementor can choose how many PIDs are used in the platform. If enough interrupt
lines are shared, there need be only one PID in the system; all interrupts in the system would then
be routed to that PID. Each PID has enough interrupt inputs to handle dedicated interrupts from the
cards on two PCI buses. Therefore, using one PID per PXB provides a high performance solution
with minimum routing between PCI buses.
1.8
Reference Documents
In addition to this document, the reader should be familiar with the following reference documents:
• Intel® 460GX Chipset Datasheet
(Document Number: 248703)
• Intel® Itanium™ Processor Hardware Developer’s Manual
(Document Number: 248701)
• Intel® Itanium™ Processor at 800 MHz and 733 MHz Datasheet
(Document Number: 245481)
• Intel® 82460GX Chipset OLGA1 Package, Manufacturing, Mechanical, and Thermal Design
Guide
• PCI Local Bus Specification, Rev 2.2
(http://www.pcisig.com/)
• Accelerated Graphics Port Interface Specification
(http://www.intel.com/technology/agp/agp_index.htm)
1-6
Intel® 460GX Chipset Software Developer’s Manual
Introduction
• JTAG IEEE 1149.1 Specification
(http://www.ieee.com)
• Universal Serial Bus Specification
(http://www.usb.org)
• System Management Bus Specification, Rev. 1.0
• Low Pin Count (LPC) Interface Specification, Rev 1.0
Note: Contact your Intel representative for the latest revision of the documents without document
numbers.
1.9
Revision History
Date
Description
June 2001
Initial release.
Intel® 460GX Chipset Software Developer’s Manual
1-7
Introduction
1-8
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
2
The 460GX chipset has both memory mapped and PCI configuration space mapped registers. The
460GX chipset supports access mechanism #1 as defined in the PCI specification. Two 32-bit
register locations (CONFIG_ADDRESS and CONFIG_DATA) are defined in the Itanium
processor’s I/O space; I/O accesses to these registers are translated by the 460GX chipset into
appropriate PCI configuration cycles.
To access a configuration register in the 460GX chipset (or any other I/O device), software first
writes a value to the CONFIG_ADDRESS location consisting of the bus number, Device Number,
function number and register number. These writes are claimed and saved by the 460GX chipset.
Subsequent reads or writes to the CONFIG_DATA location result in the 460GX chipset using the
information stored in CONFIG_ADDRESS to deliver a PCI configuration read or write cycle to
the appropriate address on the appropriate PCI bus.
Upon reset, the 460GX chipset sets its internal configuration registers to predetermined default
states, representing the minimum feature set required to successfully bring up the system. It is
expected that the firmware will properly determine and program the optimal configuration settings.
The 460GX chipset implements a PCI-compatible configuration space for each PCI bus under the
PXBs, for each AGP bus under the GXB, and for each 460GX chipset component. Each
configuration space provides hardwired device identification, address range registers, operation
control registers, status and error registers. This chapter describes how the configuration spaces are
accessed, then provides detailed descriptions of each register.
2.1
Access Mechanism
The PCI specification defines two bus cycles to access PCI configuration space: Configuration
Read and Configuration Write. While memory and I/O spaces are supported by the microprocessor,
configuration space is not directly supported. The PCI specification defines two mechanisms to
access configuration space, Mechanism #1 and Mechanism #2. The 460GX chipset supports only
Mechanism #1.
Mechanism #1 defines two I/O-space locations: an address register (CONFIG_ADDRESS) at
location 0CF8h, and a data register (CONFIG_DATA) at location 0CFCh. Dword I/O Writes to the
configuration address are latched and held; they specify the PCI Bus Number, Device Number
within the bus, and Register Number within the device. Subsequent I/O reads and writes to the
configuration data location cause a configuration space access the register specified by the address
stored in the configuration address location.
Note: The AGP bus under the GXB looks like a standard PCI bus for configuration purposes. The term
xXB refers to the PXB, WXB, or GXB. In general, any reference to an access to PCI bus includes
accesses to an AGP bus.
Configuration space accesses are processed as follows:
• If the SAC detects that the I/O request is a configuration access to its own configuration space,
it will service that request entirely within the SAC or the other chipset components. Reads
result in data being returned to the system bus.
• If the SAC detects that the I/O request is a configuration access to a xXB configuration space,
it will forward the request to the appropriate xXB for servicing. The request is not forwarded
Intel® 460GX Chipset Software Developer’s Manual
2-1
Register Descriptions
to a PCI bus. Reads result in data being returned by the xXB through the SAC to the system
bus.
• Otherwise, the access is forwarded to the xXB to be placed on the PCI bus (or AGP bus) as a
Configuration Read or Configuration Write cycle. Reads will result in data being returned
through the xXB and SAC back to the system bus, just as in normal Outbound Read
operations.
2.2
Access Restrictions
The 460GX chipset supports PCI configuration space access using the mechanism denoted as
Configuration Mechanism #1 in the PCI specification.
The 460GX chipset internal registers (both I/O Mapped and Configuration registers) are accessible
by the Host CPU. The registers can be accessed as Byte, Word (16-bit), or Dword (32-bit)
quantities, with the exception of CONFIG_ADDRESS which can only be accessed as a Dword. All
multi-byte numeric fields use “little-endian” ordering (i.e. lower addresses contain the least
significant parts of the field).
2.2.1
Partitioning
Each SAC, SDC, MAC, PXB, WXB, GXB, each AGP bus below an GXB, and each PCI bus below
an PXB or WXB, has an independent configuration space. None of the registers are shared between
the spaces; that is, the SAC, and each PCI bus in the PXB, have separate control and status
registers.
Configuration registers are accessed using an “address” comprised of the PCI Bus Number, the
Device Number within the bus, and the Register Number within the Device.
Accesses to devices on Bus #0 and Bus #(CBN) are serviced by the 460GX chipset depending on
their device number. Device 10h on Bus #0 is mapped to the SAC; it contains the programmable
Chipset Bus Number. All other chipset devices reside on bus CBN.
mapping information.
Table 2-1. Device Mapping on Bus CBN
No.
Device
No.
Device
a
00h
01h
02h
03h
04h
05h
06h
07h
SAC
SAC
10h
11h
12h
13h
14h
15h
16h
17h
Expander 0, Bus a
Expander 0, Bus b
Expander 1, Bus a
Expander 1, Bus b
Expander 2, Bus a
Expander 2, Bus b
Expander 3, Bus a
Expander 3, Bus b
reserved
reserved
reserved
SDC
Memory Card A
Memory Card B
reserved
08h-0Fh
reserved
18h-1Fh
a. This is the compatibility bus (where the boot vector is always directed).
Configuration registers located in the SDC are accessed over the private data bus. The SAC
translates CF8/CFC accesses to SDC registers into configuration commands over the PDB.
Configuration registers located on the memory boards are accessed over the I2C port. The SAC
2-2
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
translates CF8/CFC accesses to the MAC registers into read/write commands over the I2C port.
The SAC also contains an IIADR pointer register that can be used in conjunction with a CF8/CFC
access to generate I2C commands to generic I2C devices on the memory boards.
2.2.2
Register Attributes
Registers have designated “access attributes”, with the following definitions:
Read Only
Read/Write
Writes to this register have no effect.
Data may be read from and written to this register. Selected bits in the register may
be designated as “hardwired” or “read-only”; such bits are not affected by data writes
to the register.
Read/Clear
Sticky
Data may be read from the register. A data write operates strictly as a clear: a “1”-bit
in the data field clears the corresponding bit in the register, while a “0”-bit in the data
field has no effect on the corresponding bit in the register. Selected bits in the register
may be designated as “hardwired” or “read-only”; such bits are not affected by data
writes to the register.
Data in this register remains valid and unchanged, during and following a hard reset.
Typically, these registers contain special configuration information or error logs.
2.2.3
2.2.4
Reserved Bits Defined in Registers
Most 460GX chipset registers described in this section contain reserved bits. The PCI specification
requires that software correctly handle reserved fields, as follows. On reads, software must use
appropriate masks to extract the defined bits and not rely on reserved bits being any particular
value. On writes, software must ensure that the values of reserved bit positions are preserved. That
is, the values of reserved bit positions must first be read, merged with the new values for other bit
positions and then written back. Note the software does not need to perform read, merge, write
operation for the CONFIG_ADDRESS register.
Reserved or Undefined Register Locations
In addition to reserved bits within a register, the 460GX chipset contains address locations in the
PCI configuration space that are marked “Reserved” or are simply undefined. Several of the
460GX chipset devices are multi-function devices; all registers in the unused functions are
considered “Reserved”. Reserved registers can be 8-, 16-, or 32-bit in size. The PCI specification
requires that the 460GX chipset respond to accesses to these address locations by completing the
host cycle. Reserved register locations must be treated by software the same as reserved fields are
treated: software can not rely on reads returning any particular value, and must not attempt to
change the value returned when read.
2.2.5
Default Upon Reset
Upon reset, the 460GX chipset sets its internal configuration registers to predetermined default
states. The default state represents the minimum functionality feature set required to successfully
bring up the system. Hence, it does not represent the optimal system configuration. It is the
responsibility of the system initialization software (firmware) to properly determine the DRAM
configurations, operating parameters and optional system features that are applicable, and to
program the 460GX chipset registers accordingly.
Intel® 460GX Chipset Software Developer’s Manual
2-3
Register Descriptions
2.2.6
Consistency
There are a number of registers that are repeated in both the SAC and xXB/PCI spaces. It is
software’s responsibility to insure that these registers are programmed in a consistent fashion.
Failure to insure consistency can produce indeterminate results. See the Initialization Chapter for
an overview on initializing all chipset components.
When the address decode ranges of 460GX chipset devices are being updated, no other bus traffic
is allowed over the address ranges being affected by the update. This means that the code that
updates initial configuration must be executing from a location that will not be affected by the
update. Furthermore in a multiprocessor system, precautions should be taken to assure that only
one CPU is accessing configuration space at a time.
2.2.7
GART Programming Region
The region starting at FE20_0000h is used for programming the GARTs. This region is accessible
2.3
I/O Mapped Registers
The 460GX chipset contains two registers that reside in the CPU I/O address space: the
Configuration Address (CONFIG_ADDRESS) Register and the Configuration Data
(CONFIG_DATA) Register. The Configuration Address Register enables/disables the
configuration space and determines what portion of configuration space is visible through the
Configuration Data window. The following sections define the fields within the
CONFIG_ADDRESS and CONFIG_DATA registers. The 460GX chipset’s device ID mapping
2.3.1
CONFIG_ADDRESS: Configuration Address Register
I/O Address:
Default Value:
Sticky:
CF8h [Dword]
00000000h
No
Size:
Attribute:
Locked:
32 bits
Read/Write
No
CONFIG_ADDRESS is a 32 bit register accessed only when referenced as a Dword. A Byte or
Word reference will “pass through” the Configuration Address Register onto the PCI bus as an I/O
cycle. The CONFIG_ADDRESS register contains the Bus Number, Device Number, Function
Number, and Register Number for which a subsequent configuration access is intended.
Bits
Description
31
Configuration Enable(CFGE).
When this bit is set to 1 accesses to PCI configuration space are enabled. If this bit is reset
to 0 accesses to PCI configuration space are disabled.
30:24
23:16
reserved (0)
Bus Number.
When the Bus Number is programmed to match the Chipset Bus Number (CBN), the
target of the Configuration Cycle is the 460GX chipset. If the Bus Number is not CBN,
the destination and type of access is determined by the Bus Number and Subordinate Bus
Number of each PCI port in each PXB. A type 0 access is generated on the appropriate
PCI bus if one of the PXB port’s bus number is matched. Otherwise, a type 1
configuration cycle is generated on the appropriate PCI bus below the PXB port whose
2-4
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
subordinate bus number is in that range. For a type 1 cycle, the Bus Number is mapped
to AD[23:16] during the address phase.
15:11
Device Number.
This field selects one agent on the PCI bus selected by the Bus Number. Device 16 (10h)
on Bus #0 is always reserved for programming the CBN. On the bus that the chipset is
mapped into (determined by the CBN register), Device Numbers 0-31 are reserved for
forwarded to the selected bus.
1 0:8
7:2
Function Number.
This field is mapped to AD[10:8] during PCI configuration cycles. This allows the
configuration registers of a particular function in a multi-function device to be accessed.
Register Number.
This field selects one register within a particular Bus, Device, and Function as specified
by the other fields in the Configuration Address Register. This field is mapped to AD[7:2]
during PCI configuration cycles.
1:0
reserved (0)
2.3.2
CONFIG_DATA: Configuration Data Register
I/O Address:
Default Value:
Sticky:
CFCh
00000000h
No
Size:
Attribute:
Locked:
32 bits
Read/Write
No
CONFIG_DATA is a 32 bit read/write window into configuration space. The portion of
configuration space that is referenced by CONFIG_DATA is determined by the contents of
CONFIG_ADDRESS.
Bits
Description
31:0
Configuration Data Window (CDW).
If bit 31 of CONFIG_ADDRESS is 1 any I/O reference that falls in the CONFIG_DATA
I/O space will be mapped to configuration space using the contents of
CONFIG_ADDRESS.
2.4
Error Handling Registers
2.4.1
SAC
2.4.1.1
SECTID: SEC ITID
Bus CBN, Device Number: 00h
Function:
Size:
Attribute:
0
Address Offset:
Default Value:
80h
00h
8 bits
Read Only/Write
Clear, Read/Write
No
Sticky:
Yes
Locked:
This register is used to capture the ITID for a single bit memory ECC error. The ITID can then be
used to determine the address of the failure. To force the ITID to be cleared and re-used, write a 1
to bit 6. This register is set anytime that the SEC bit is sent from the SDC to the SAC on a ’Retire
ITID’ command.
Intel® 460GX Chipset Software Developer’s Manual
2-5
Register Descriptions
Bits
Description
Disable
7
This bit can be written by software. When set, the ITID is retired immediately and not
of this bit.
6
Valid
If set then the ITID in bits 5:0 is valid and shows the address of a single-bit memory error.
Writing a 1 to this bit will clear the ITID and reset the valid bit.
5:0
ITID
The ITID of the SEC error. These bits are read-only.
2.4.1.2
DEDTID: DED ITID
Bus CBN, Device Number: 00h
Function:
Size:
Attribute:
0
Address Offset:
Default Value:
81h
00h
8 bits
Read Only/Write
Clear, Read/Write
No
Sticky:
Yes
Locked:
This register is used to capture the ITID for a double bit memory ECC error. The ITID can then be
used to determine the address of the failure. To force the ITID to be cleared and re-used, write a 1
to bit 6. This register is set anytime that the DED bit is sent from the SDC to the SAC on a ‘Retire
ITID’ command.
Bits
Description
Disable
7
This bit can be written by software. When set, the ITID is retired immediately and not
of this bit.
6
Valid
If set then the ITID in bits 5:0 is valid and shows the address of a double-bit memory
error. Writing a 1 to this bit will clear the ITID and reset the valid bit.
5:0
ITID
The ITID of the double-bit error. These bits are read-only.
2.4.1.3
FSETID: FSE ITID
Bus CBN, Device Number: 00h
Function:
Size:
Attribute:
0
Address Offset:
Default Value:
82h
00h
8 bits
Read Only/Write
Clear, Read/Write
No
Sticky:
Yes
Locked:
This register is used to capture the ITID for a system bus data error. The ITID can then be used to
determine the address of the failure. To force the ITID to be cleared and re-used, write a 1 to bit 6.
This register is set anytime that the ADE bit is asserted and both SEC and DED are deasserted on a
’Retire ITID’ command from the SDC to the SAC. NOTE: this register is set for both processor-
bus errors and errors on the SAC-to-SDC data bus.
2-6
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
Bits
Description
Disable
7
This bit can be written by software. When set, the ITID is retired immediately and not
of this bit.
6
Valid
If set then the ITID in bits 5:0 is valid and shows the address of a system bus data error.
Writing a 1 to this bit will clear the ITID and reset the valid bit.
5:0
ITID
The ITID of the system bus error. These bits are read-only.
2.4.1.4
FERR_SAC: First Error Status Register
Bus CBN, Device Number: 00h
Function:
Size:
Attribute:
Locked:
1
Address Offset:
Default Value:
Sticky:
40h
000000h
Yes
32 bits
Read/Write Clear
No
This register records the first error condition detected in the SAC/SDC.
Bits
Description
31
Memory Card B Error (MBE)
Set when the memory card B signals a fatal error.
30
29
Memory Card A Error (MAE)
Set when the memory card A signals a fatal error.
XSERR# Asserted (XSA)
Set when the SAC sees the signal XSERR# active.
28
27
26
25
‘Store-Write’ Command Underflow, card A, Stack L (SCAL)
‘Store-Write’ Command Underflow, card A, Stack R (SCAR)
‘Store-Write’ Command Underflow, card B, Stack L (SCBL)
‘Store-Write’ Command Underflow, card B, Stack R (SCBR)
One of these 4 is asserted when a signal is sent from the SDC to the SAC indicating write
data was sent to the MDC, and there is no outstanding write in the SAC.
24
23
SDC Correctable Memory Error (SCME)
Reports correctable DRAM errors (single-bit ECC errors). This bit does not mask other
bits in the FERR register from being set. It is the one exception to the rule that only one
bit in FERR may be set at a time.
SDC Non-Fatal Error (SNE)
Reports non-fatal errors that are uncorrectable such as double-bit ECC error, or parity
errors. This also reports a single-bit correctable error on the system bus. This bit will also
be set if there is a second correctable error from memory in the SDC, and the first one has
not been cleared by the time the second one occurs. The first correctable memory error
would have set the SCME bit, and all later correctable memory errors (until the SDC’s
error registers are cleared) are reported as SNE in the FERR or NERR.
22
SDC Fatal Error (SFE)
Fatal error in SDC.
21
20
19
‘Completion’ Command Underflow; MAC A, Stack L (CCAL)
‘Completion’ Command Underflow; MAC A, Stack R (CCAR)
‘Completion’ Command Underflow; MAC B, Stack L (CCBL)
Intel® 460GX Chipset Software Developer’s Manual
2-7
Register Descriptions
18
‘Completion’ Command Underflow; MAC B, Stack R (CCBR)
One of these 4 bits is set when the SAC receives a completion from the MAC and the
SAC has no outstanding transaction.
17
16
BERR# Observed (BER)
BERR# seen on the system bus. Set whenever BERR# is observed active.
IOQ Underflow/Overflow (IUE)
Set when the IOQ is empty and the SDC sends out a signal saying it popped something
from the top of the queue. Or set when the IOQ is 8 (or 1 when the IOQ depth is set to 1)
and an ADS# is seen on the bus.
15
14
reserved(0)
External XBINIT# Active. (XBE)
Set when XBINIT# is seen active. This signal is from an Expander port or other external
agent.
13
12
False Retirement (FRE)
Retirement from SDC that doesn’t match an outstanding ITID in the SAC.
Address above TOM (TE)
Asserted when an address on the system bus is above TOM and not inside the I/O gap
below 4 GB.
11
10
9
Illegal HITM# (IHS)
HITM# on non-memory access.
Unsupported ASZ[1:0]# (ASE)
Processor access to an address above 64 GB, so that ASZ# = 10b or 11b.
System Bus Address Parity Error (AE)
Parity error on A[36:3]#.
8
System Bus Request Parity Error (RQE)
Parity error on REQ[4:0]#.
7
PDB ITID Parity Error (IPE)
Parity error on the ITID bus from SDC to SAC.
6
Retirement Bus Parity Error (RPE)
Parity error on the retirement bus from the SDC to the SAC.
5
Lock# Transaction With No Resources Available (LTE)
Set when a LOCK# transaction occurs and there are no outbound resources available in
which to place the lock.
4:1
0
reserved(0)
Resource Counter Overflow/Underflow (RCE)
Set if the resource counter has an underflow or overflow.
2.4.1.5
NERR_SAC: All Error Status Register
Bus CBN, Device Number: 00h
Function:
Size:
Attribute:
Locked:
1
Address Offset:
Default Value:
Sticky:
44h
000000h
Yes
32 bits
Read/Write Clear
No
This register records all error conditions detected in the SAC/SDC.
Bits
Description
31:0
See FERR_SAC for bit definitions.
2-8
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
2.4.1.6
SA_FERR: System Address on First Error
Bus CBN, Device Number: 00h
Function:
1
Address Offset:
Default Value:
Sticky:
60h
Size:
Attribute:
Locked:
128 bits
Read Only
No
undefined after
Yes
This register records and latches the address for the first system bus error detected.
Bits Description
127:107 Reserved (0)
106
105
104
103:99
98
LOCK, ’b’ phase.
ADS, ’b’ phase.
RP#, ’b’ phase.
REQ, ’b’ phase.
AP1; ’b’ phase.
AP0; ’b’ phase.
A[35:3]#, ’b’ phase.
Reserved (0)
97
96:64
63:43
42
LOCK#, ’a’ phase.
ADS#, ’a’ phase.
41
40
RP# for REQa#.
Parity on REQa# signals.
39:35
34:33
32:0
REQa#.
REQa signals on error.
AP[1:0]#, ’a’ phase.
Address parity for failing address.
Aa[35:3]#, ’a’ phase.
System Bus - System Address of Error.
2.4.1.7
BIUITID: BIU ITID Register
Bus CBN, Device Number: 00h
Function:
Size:
Attribute:
Locked:
1
Address Offset:
Default Value:
Sticky:
80h
0
No
8 bits
Read/Write
No
A write to this register causes the SAC to update the BIUDATA register with the contents of the
CAM and RAM associated with the ITID that is written into this register.
Bits
7:6
Description
reserved (0)
5:0
ITID
This is the ITID that is used to address the CAM/RAM structure.
Intel® 460GX Chipset Software Developer’s Manual
2-9
Register Descriptions
2.4.1.8
BIUDATA: BIU Data Register
Bus CBN, Device Number: 00h
Function:
Size:
Attribute:
Locked:
1
Address Offset:
Default Value:
Sticky:
90h
undefined
No
128 bits
Read Only
No
This is the contents of the CAM concatenated with the contents of the RAM associated with the
ITID in BIUITID.
Bits
Description
127:116 reserved(0)
115:82
Address bits [35:2].
This is the contents of the CAM with address bits [5:2] from the RAM (bit 2 is only of
interest if the transaction came from an Expander bus).
81:76
75:71
70:63
62:55
54
reserved(0)
Reqa. Request phase a[4:0].
DID. The DID for the transaction.
BE. The byte enables for the transaction.
reserved (0)
53
OWN. OWN# active.
52
DPS. DPS# active.
51:49
48
Reqb. Request phase b bits [4:2].
Lock. The transaction.had LOCK# asserted.
LockLoad. The transaction is the first occurrence of an OB lock sequence.
Dst. The destination of the transaction.
ORetry. A retry due to HITO.
47
46:43
42
41:36
35
CMD. The command for the transaction.
P2P. Set for peer-to-peer transactions.
FEorR. The end-of-request bit from the Expander port.
FRoute. The Expander bus route.
34
33:30
29:22
21:12
11:9
8
FLEN. The length on the Expander bus.
FTID. The Expander id.
Len. The length of the transaction.
System Bus Retry. The transaction was retried on the bus.
Dfr. The transaction is deferred.
7
6
MEM. The target for the transaction if memory.
System Bus. The target for the transaction is the system bus.
CLINE. The transaction is for a full line.
Zero. If set then this is a 0-length transaction.
5
4
3
2:0
RS. The Response generated for the transaction by the BIU. This may not match the
system bus response sent, since the BIU’s response may be changed by the MIU. This
may be for a HITM# or other reasons.
Note: Note: if the P2P bit is not set, then bits [34:12] and [76] are not defined, since the transaction
originated on the system bus and not the Expander bus.
2-10
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
2.4.2
SDC
2.4.2.1
SEC0_D_FERR: Data on First Memory Card B SEC
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
40-47h
0
Size:
Attribute:
64 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the data corresponding to the first SEC detected by memory
interface 0 in the SDC.
Bits
Description
63:0
DE - System Data of Error.
2.4.2.2
SEC0_ECC_FERR: ECC on First Memory Card B SEC
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
48h
00h
Size:
Attribute:
8 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the ECC checkbits corresponding to the first SEC detected by
memory interface 0 in the SDC
Bits
Description
7:0
ECC - ECC of Error.
2.4.2.3
SEC0_TXINFO_FERR: TXINFO on First Memory Card B SEC
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
49-4Ah
00h
Size:
Attribute:
16 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records the ITID and failing chunk corresponding to the first SEC detected by
memory interface 0 in the SDC.
Bits
15:9
8:6
Description
reserved(0)
DC - Data Chunk of ITID.
ITID - ITID of error.
5:0
Intel® 460GX Chipset Software Developer’s Manual
2-11
Register Descriptions
2.4.2.4
2.4.2.5
2.4.2.6
DED0_D_FERR: Data on First Memory Card B DED
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
50-57h
0
Size:
Attribute:
64 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the data corresponding to the first DED detected by memory
interface 0 in the SDC.
Bits
Description
63:0
DE - System Data of Error.
DED0_ECC_FERR: ECC on First Memory Card B DED
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
58h
00h
Size:
Attribute:
8 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the ECC checkbits corresponding to the first SEC detected by
memory interface 0 in the SDC
Bits
Description
7:0
ECC - ECC of Error.
DED0_TXINFO_FERR: TXINFO on First Memory Card B DED
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
59-5Ah
00h
Size:
Attribute:
16 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records the ITID and failing chunk corresponding to the first DED detected by
memory interface 0 in the SDC.
Bits
15:9
8:6
Description
reserved(0)
DC - Data Chunk of ITID.
ITID - ITID of error.
5:0
2.4.2.7
SEC1_D_FERR: Data on First Memory Card A SEC
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
60-67h
0
Size:
Attribute:
64 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
2-12
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
This register records and latches the data corresponding to the first SEC detected by memory
interface 1 in the SDC.
Bits
Description
63:0
DE - System Data of Error.
2.4.2.8
SEC1_ECC_FERR: ECC on First Memory Card A SEC
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
68h
00h
Size:
Attribute:
8 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the ECC checkbits corresponding to the first SEC detected by
memory interface 1 in the SDC.
Bits
Description
7:0
ECC - ECC of Error.
2.4.2.9
SEC1_TXINFO_FERR: TXINFO on First Memory Card A SEC
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
69-6Ah
00h
Size:
Attribute:
16 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records the ITID and failing chunk corresponding to the first SEC detected by
memory interface 1 in the SDC.
Bits
15:9
8:6
Description
reserved(0)
DC - Data Chunk of error.
ITID - ITID of error.
5:0
2.4.2.10
DED1_D_FERR: Data on First Memory Card A DED
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
70-77h
0
Size:
Attribute:
64 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the data corresponding to the first DED detected by memory
interface 1 in the SDC.
Bits
Description
63:0
DE - System Data of Error.
Intel® 460GX Chipset Software Developer’s Manual
2-13
Register Descriptions
2.4.2.11
DED1_ECC_FERR: ECC on First Memory Card A DED
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
78h
00h
Size:
Attribute:
8 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the ECC checkbits corresponding to the first DED detected by
memory interface 0 in the SDC.
Bits
Description
7:0
ECC - ECC of Error.
2.4.2.12
DED1_TXINFO_FERR: TXINFO on First Memory Card A DED
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
79-7Ah
00h
Size:
Attribute:
16 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records the ITID and failing chunk corresponding to the first DED detected by
memory interface 1 in the SDC.
Bits
15:9
8:6
Description
reserved(0)
DC - Data Chunk of ITID.
ITID - ITID of error.
5:0
2.4.2.13
SDC_FERR: First Error Status Register
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
80-83h
0000h
Size:
Attribute:
32 bits
Read/Write to Clear
This register records the first error condition detected in the SDC. Writing a ’1’ to this register will
clear the bit in both SDC_FERR and the same bit in SDC_NERR.
Bits
Description
31
Simultaneous S/W write-one-to-clear and H/W error detected in the same cycle. This bit
will only be set if another bit is also set. This implies that the ERROR>_<TYPE>_FERR
data registers associated with the other asserted bit contain stale data.
30
29
28
27
PDB Receive Length Error (RLE)
Private Bus receive length error
DRDY# Protocol Error (FS2)
Asserted when a protocol error is found involving DRDY#, SBUSY# and DBUSY#.
Write Data Protocol Error (FS1)
Asserted on write protocol errors.
LEN# Protocol Error (FS0)
Asserted on mismatches of LEN# field and actual data transmitted.
2-14
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
26
25
24
’Forward’ Overlapping ’Forward’; Card A (FWMDI1)
Indicates FWMDI sampled asserted while a store transaction is in progress
’Load’ Overlapping ’Load’; Card A (LRMDI1)
Indicates LRMDI sampled asserted while a store transaction is in progress
’Load’ Overlapping ’Forward’; Card A (WrRd1)
Memory interface 1 detected simultaneous read and write operation. Write and Read
collision.
23
’Forward’ Overlapping ’Load’; Card A (RdWr1)
Memory interface 1 detected simultaneous read and write operation. Read and write
collision.
22
21
20
19
18
17
’Forward’ Underflow; Card A Right Stack Error (FR1)
Memory interface 0 received Forward right Bank without corresponding Store command
’Forward’ Underflow; Card A Left Stack Error (FL1)
Memory interface received Forward left Bank without corresponding Store command
’Accept Underflow’; Card A (AE1)
Memory interface 0 received data without corresponding Accept command
’Forward’ Overlapping ’Forward’; Card B (FWMDI0)
Indicates FWMDI sampled asserted while a store transaction is in progress
’Load’ Overlapping ’Load’; Card B (LRMDI0)
Indicates LRMDI sampled asserted while a store transaction is in progress
’Load’ Overlapping ’Forward’; Card B (WrRd0)
Memory interface 1 detected simultaneous read and write operation. Write and Read
collision.
16
’Forward’ Overlapping ’Load’; Card B (RdWr0)
Memory interface 1 detected simultaneous read and write operation. Read and write
collision.
15
14
13
12
11
10
9
’Forward’ Underflow; Card B Right Stack Error (FR0)
Memory interface 0 received Forward right Bank without corresponding Store command
’Forward’ Underflow; Card B Left Stack Error (FL0)
Memory interface received Forward left Bank without corresponding Store command
’Accept’ Underflow; Card B (AE0)
Memory interface 0 received data without corresponding Accept command
Configuration Information Parity Error (CIE)
Data buffer detected data parity while reading config address or data from SDC RAM.
Response Bus Transmission Error (RTE)
Indicates that the SDCRSP bus detected a transmission error.
PDB - ITID Parity Error (IPE)
Look in ITID_FERR Register to isolate.
PDB - Command Parity Error (CPE)
Look in CMD_FERR Register to isolate.
8
PDB Byte Enable Parity Error (BPE)
Parity error on the Byte-enables from the SAC.
7
SDC Data Buffer RAM Parity Error (RPE)
SDC detected bad parity on good data stored in its data buffer. Indicates potential RAM
cell disturbance due to alpha or cosmic hit. All four data port map to this bit.
6
PDB - Data Parity Error (DPE)
Parity Error Detected on transfer of Data from SAC to SDC.
Intel® 460GX Chipset Software Developer’s Manual
2-15
Register Descriptions
5
4
3
2
1
0
System Bus Double Bit Error (DEDF)
ECC Double Bit Error Detected on system bus.
System Bus Single Bit Error (SECF)
ECC Single Bit Error Detected on system bus.
SDC Card A Double Bit Error (DED1)
ECC Double Bit Error Detected from Memory Card A.
SDC Card A Single Bit Error (SEC1)
ECC Single Bit Error Detected from Memory Card A.
SDC Card B Double Bit Error (DED0)
ECC Double Bit Error Detected from Memory Card B.
SDC Card B Single Bit Error (SEC0)
ECC Single Bit Error Detected from Memory Card B.
2.4.2.14
SDC_NERR: SDC Next Error Status Register
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
84-87h
0000h
Size:
Attribute:
32 bits
Read/Write to Clear
This register records the next error status within the SDC. Writing a ’1’ to this register will clear
the bit in both SDC_NERR and the same bit in SDC_FERR.
Bits
Description
31:0
See SDC_FERR for bit definitions.
2.4.2.15
PCMD_FERR: Command on First PCMD Parity Error
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
88-8Bh
00h
Size:
Attribute:
32 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the data associated with the first parity error detected on the
PCMD bus.
Bits
31:19
18
Description
reserved(0)
If set then the error was detected on the 1st half of the double–pumped transfer.
Otherwise, these fields contain the information from the 2nd half of the double-pumped
transfer.
17
Parity of Error
16:0
PCMD - Private Data Command value of Error.
2.4.2.16
PITID_FERR: Data on First PITID Parity Error
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
8Ch
0h
Size:
Attribute:
8 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
2-16
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
This register records and latches the data associated with the first parity error detected on the
PITID bus.
Bits
Description
7
If set then the error was detected on the 1st half of the double-pumped transfer. Otherwise,
these fields contain the information from the 2nd half of the double-pumped transfer.
6
Parity of Error
5:0
PITID - Private ITID bus value of Error.
2.4.2.17
SDCRSP_FERR: Response on First SDCRSP Error
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
8Dh
0h
Size:
Attribute:
8 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the data and inverted data associated with the first transmission
error detected on the SDCRSP bus.
Bits
7:4
Description
Response Bus for 2nd half of double–pumped transfer.
Response Bus for 1st half of double–pumped transfer.
3:0
2.4.2.18
DPBRLE_FERR: Private Data Bus Receive Length Error
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
8Eh
0h
Size:
Attribute:
8 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register indicates that the amount of data transferred from the SAC to the SDC for a given
transfer did not match the expected transfer length.
Bits
7:3
2
Description
reserved(0)
Data packet longer than expected (LDP)
Data packet shorter than expected (SDP)
No data packet shipped as expected (NDP)
1
0
2.4.2.19
ECCMSK0: ECC Mask Register - Card B
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
C8h
00h
Size:
Attribute:
8 bits
Read/Write
This register is used to test the ECC error detection logic in the memory subsystem for memory
card 0. To test, this register is written with a masking function. All subsequent writes into memory
will store a masked version of the computed ECC. Subsequent reads of memory locations written
Intel® 460GX Chipset Software Developer’s Manual
2-17
Register Descriptions
while masked will return an invalid ECC code. To disable testing, the mask value is left at 0h (the
default). The mask is a bit-wise XOR with the computed ECC.
Bits
Description
7:0
ECC Generation Mask - For 64 bits of data.
2.4.2.20
ECCMSK1: ECC Mask Register - Card A
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
C9h
00h
Size:
Attribute:
8 bits
Read/Write
This register is used to test the ECC error detection logic in the memory subsystem for memory
card 1. To test, this register is written with a masking function. All subsequent writes into memory
will store a masked version of the computed ECC. Subsequent reads of memory locations written
while masked will return an invalid ECC code. To disable testing, the mask value is left at 0h (the
default). The mask is a bit-wise XOR with the computed ECC.
Bits
Description
7:0
ECC Generation Mask - For 64 bits of data.
2.4.2.21
ECCMSKF: ECC Mask Register
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
CAh
00h
Size:
Attribute:
8 bits
Read/Write
This register is used to test the ECC error detection logic of the host processor bus. To test, this
register is written with a masking function. All subsequent processor reads will received a masked
version of ECC code. To disable testing, the mask value is left at 0h (the default). The mask is a
bit-wise XOR with the computed ECC.
Bits
Description
7:0
ECC Generation Mask - For 64 bits of data.
2.4.2.22
ParMskP: PB Parity Mask and IB Correction Enable Register
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
CBh
00h
Size:
Attribute:
8 bits
Read/Write
The first 4 bits of this register are used to test the data parity error detection logic of the private bus.
To test, bits 3:0 are written with a masking function. All subsequent private bus reads will receive a
masked version of double byte parity. To disable testing, the mask value is left at 0h (the default).
The mask is a bit-wise XOR with the computed parity.
Bits 7:4 are used to enable ECC or parity checking for the different busses. Note that the SDC
defaults to no parity or ECC checking at power-on.
Bits
7
Description
Private Bus parity detection enable.
Front Side Bus ECC correction/detection enable.
6
2-18
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
5
Memory Bus A ECC correction/detection enable.
Memory Bus B ECC correction/detection enable.
Double byte parity mask for 128 bits of data.
4
3:0
2.4.2.23
PVD_D_FERR: Data on First PVD Parity Error
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
D0-D7h
0
Size:
Attribute:
64 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the data associated with the first parity error detected on the PVD
bus.
Bits
Description
63:0
PVD - Private Data Bus data.
2.4.2.24
PVD_PAR_FERR: Parity on First PVD Parity Error
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
D8h
0
Size:
Attribute:
8 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the data associated with the first parity error detected on the PVD
bus.
Bits
7:4
Description
reserved(0)
3:0
Double–byte parity of error
2.4.2.25
PVD_TXINFO_FERR: TXINFO on First PVD Parity Error
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
D9-DAh
00h
Size:
Attribute:
16 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records the ITID and failing chunk corresponding to the first double-byte parity
detected by private bus interface in the SDC.
Bits
15:9
8:6
Description
reserved(0)
DC - Data Chunk of ITID.
ITID - ITID of error.
5:0
Intel® 460GX Chipset Software Developer’s Manual
2-19
Register Descriptions
2.4.2.26
2.4.2.27
2.4.2.28
SECF_D_FERR: Data on First System Bus SEC
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
E0-E7h
0
Size:
Attribute:
64 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the data corresponding to the first SEC detected by system bus
interface in the SDC.
Bits
Description
63:0
DE - System Data of Error.
SECF_ECC_FERR: ECC on First System Bus SEC
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
E8h
00h
Size:
Attribute:
8 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the ECC checkbits corresponding to the first SEC detected by
system bus interface in the SDC.
Bits
Description
7:0
ECC - ECC of Error.
SECF_TXINFO_FERR: TXINFO on First System Bus SEC
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
E9-EAh
00h
Size:
Attribute:
16 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records the ITID and failing chunk corresponding to the first SEC detected by system
bus interface in the SDC.
Bits
15:9
8:6
Description
reserved(0)
DC - Data Chunk of error.
ITID - ITID of error.
5:0
2.4.2.29
DEDF_D_FERR: Data on First System Bus DED
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
F0-F7h
0
Size:
Attribute:
64 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
2-20
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
This register records and latches the data corresponding to the first DED detected by system bus
interface in the SDC.
Bits
Description
63:0
DE - System Data of Error.
2.4.2.30
DEDF_ECC_FERR: ECC on First System Bus DED
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
F8h
00h
Size:
Attribute:
8 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records and latches the ECC checkbits corresponding to the first DED detected by
system bus interface in the SDC.
Bits
Description
7:0
ECC - ECC of Error.
2.4.2.31
DEDF_TXINFO_FERR: TXINFO on First System Bus DED
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
F9-FAh
00h
Size:
Attribute:
16 bits
Read Only, New Value Latched
anytime appropriate FERR register
bit is set
This register records the ITID and failing chunk corresponding to the first DED detected by system
bus interface in the SDC.
Bits
15:9
8:6
Description
reserved(0)
DC - Data Chunk of ITID.
ITID - ITID of error.
5:0
2.4.3
MAC
2.4.3.1
FERR_MAC: First Error Status Register
Bus CBN, Device Number: 05h,06h
Function Number: 00h,01h
Address Offset:
Default Value:
98h
00h
Size:
Attribute:
8 bits
Read
This register records the first error condition detected in the MAC.
Bits
7:2
1
Description
reserved(0)
Que-Overflow Error
Signals that the MAC received too many commands from the SAC.
Intel® 460GX Chipset Software Developer’s Manual
2-21
Register Descriptions
0
Parity Error - CMND
Parity Error Detected on SAC-MAC CMND Bus. Look in CMND_FERR Register to
isolate. When the error is detected, the MAC will complete those operations which have
a RAS pending, and stop. No new RAS cycles will be issued after the parity error and that
card is effectively dead.
2.4.3.2
CMND_FERR: Command on First Error
Bus CBN, Device Number: 05h,06h
Function Number: 00h,01h
Address Offset:
Default Value:
9Ch
0000h
Size:
Attribute:
24 bits
Read
This register records and latches the data on the SAC-MAC Command Bus for the first error
detected.
Bits
Description
23:22
21:19
18:17
16:0
reserved (0)
Row address [2:0]
Command [1:0]
MA[16:0]
2.4.4
PXB
2.4.4.1
ERRSTS: Error Status Register
Address Offset:
Default Value:
44h
00h
Size:
Attribute:
8 bits
Read/Write Clear, Sticky
This register records error conditions detected from the PCI bus (not already covered in PCISTS),
from the Expander Bus, and performance monitoring events.
The register is sticky through reset; that is, the contents of the register remain unchanged during
and following the assertion of X(0,1)RST#. This allows system recovery software invoked
following a forced reset to examine the flags to determine the cause of an error. Once set, the flags
remain set until explicitly cleared by software or a power-good reset.
Bits
7
Description
reserved(0).
6
PERR# observed on PCI Bus
This flag is set if the PXB detects the PERR# input asserted, and the PXB was not the
asserting agent. This flag may be configured to assert SERR# or PERR# in the ERRCMD
register. This bit remains set until explicitly cleared by software writing a 1 to this bit.
5
4
Parity Error on Received PCI Data
This flag is set if the PXB detects a parity error on data being read from the PCI bus. This
flag may be configured to assert SERR# or PERR# in the ERRCMD register. This bit
remains set until explicitly cleared by software writing a 1 to this bit.
Parity Error on PCI Address
This flag is set if the PXB detects a parity error on the PCI address. This flag may be
configured to assert SERR# in the ERRCMD register. This bit remains set until explicitly
cleared by software writing a 1 to this bit.
2-22
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
3
Inbound Delayed Read Time-out Flag
Each inbound read request that is accepted and serviced as a delayed read (i.e. the PXB
retries the request) will initiate a watchdog timer (215 cycles, per the PCI spec). If the data
has been returned and the timer expires before the requesting master initiates its repeat
request, this flag will be set. This flag may be configured to assert SERR# or PERR# in
the ERRCMD register. This bit remains set until explicitly cleared by software writing
a 1 to this bit.
2
1
reserved(0)
Performance Monitor #1 Event Flag
This flag is set when the Performance Monitor #1 requests that an interrupt request be
conditions that can cause this to occur. While this bit is set, the INT(A,B)RQ# line will
be asserted. This bit remains set until explicitly cleared by software writing a 1 to this bit.
0
Performance Monitor #0 Event Flag
This flag is set when the Performance Monitor #0 requests that an interrupt request be
conditions that can cause this to occur. While this bit is set, the INT(A,B)RQ# line will
be asserted. This bit remains set until explicitly cleared by software writing a 1 to this bit.
2.4.4.2
ERRCMD: Error Command Register
Address Offset:
Default Value:
46h
00h
Size:
Attribute:
8 bits
Read/Write
This register provides extended control over the assertion of SERR# beyond the basic controls
specified in the PCI-standard PCICMD register.
Bits
7
Description
reserved(0)
6
Assert SERR# on Observed Parity Error
If set, the PXB asserts SERR# if PERR# is observed asserted, and the PXB was not the
asserting agent.
5
Assert SERR# on Received Data with Parity Error
If set, the PXB asserts SERR# upon receiving PCI data (i.e. an inbound write or outbound
read) with a parity error. This occurs regardless of whether PXB asserts it’s PERR# pin.
4
3
Assert SERR# on Address Parity Error
If set, the PXB asserts SERR# on detecting a PCI address parity error.
Assert PERR# on Data Parity Error
If set, and the PERRE bit is set in the PCICMD register, the PXB asserts PERR# upon
receiving PCI data with parity errors.
2
Assert SERR# on Inbound Delayed Read Time-out
Each inbound read request that is accepted and serviced as a delayed read (i.e. the PXB
retries the request) will initiate a watchdog timer (215 cycles, per the PCI spec). If this
enable is set, the PXB will assert SERR# if the data has been returned and the timer
expires before the requesting master initiates its repeat request. Default=0.
1
0
reserved(0)
Return Hard Fail Upon Generating Master Abort
If set, the PXB will return a Hard Fail response through the SAC to the system bus after
generating a master abort time-out for an outbound transaction placed on the PCI bus. If
cleared, the PXB will return a normal response (with data of all 1’s for a read). In either
case, an error flag is set in the PCISTS register. Default=0.
Intel® 460GX Chipset Software Developer’s Manual
2-23
Register Descriptions
2.4.5
GXB
2.4.5.1
FERR_GXB
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
80h
00h
Size:
Attribute:
Locked:
8 bits
Read/Write Clear
No
Yes
These registers record the first error detected by the GXB. For the order to clear this register with
respect to the other GXB error registers.
Bits
7:3
2
Description
reserved (0)
FERR_PCI bit asserted
FERR_AGP bit asserted
FERR_GART bit asserted
1
0
2.4.5.2
FERR_PCI
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
84h
00h each
Yes
Size:
Attribute:
Locked:
8 bits
Read/Write Clear
No
These registers record the first error detected in GPI.
Bits
Description
7
PCISTS Error Logged
This bit is asserted when an error, except for a master abort, has been logged in the PCI
Status register.
6
Non-Configuration Master Abort
This bit is asserted when a master abort occurs on any transaction other than a
configuration read or configuration write.
reported the same as a master abort - see PCISTS register
5
Discard Timer Expiration
This is the 215 clock timeout.
4
3
2
SERR# Observed
PERR# Observed
PCI Inbound Read Que Data Parity Error
Parity error detected as read data is retrieved from buffer.
1
0
PCI Outbound Write Que Data Parity Error
Parity error detected as write data is retrieved from buffer.
Illegal OB GART Access
Access may continue or abort, results undefined.
2.4.5.3
FERR_AGP: First Error Status Register for AGP
Function Number:
Address Offset:
BFN+1
85h
Size:
8 bits
2-24
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
Default Value:
Sticky:
00h
Yes
Attribute:
Locked:
Read/Write Clear
No
These registers record and latch the first error detected in the AGP interface.
Bits
7:6
5
Description
reserved (0)
Lo-priority Read Data Que Parity Error
This is data returned to the graphics card out of the Low-priority buffer.
4
Hi-priority Read Data Que Parity Error
This is data returned to the graphics card out of the Hi-priority buffer.
3
2
1
Use of Pipe with Sideband Enabled
AGP address from graphics card [63:40] not equal to 0
AGP Request Queue Overflow
The GXB supports 16 outstanding requests. This bit is set if a new request is sent by the
AGP card when the GXB already has 16 requests.
0
Illegal AGP Command
2.4.5.4
FERR_GART: First Error Status Register for GART
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
86h
00h
Size:
Attribute:
Locked:
8 bits
Read/Write Clear
No
Yes
These registers record and latch the first error detected in the AGP interface.
Bits
7:4
3
Description
reserved (0)
GART Parity Error.
GART Entry Invalid
2
1
Illegal Address (after GART translation) in range between GAPBAS and GAPTOP, or
in VGA range and VGAGE is asserted, or directed by MARG to PCI instead of memory,
or above TOM.
0
reserved (0)
2.4.5.5
NERR_AGP: Next Errors Status Register for AGP
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
8Dh
00h
Size:
Attribute:
Locked:
8 bits
Read/Write Clear
No
Yes
This register records all error conditions detected in the AGP interface after the first error. Errors
recorded in FERR_AGP are not recorded here.
Bits
Description
7:0
See FERR_AGP for definition of these bits.
Intel® 460GX Chipset Software Developer’s Manual
2-25
Register Descriptions
2.4.5.6
NERR_GART
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
8Eh
00h each
Yes
Size:
Attribute:
Locked:
8 bits
Read/Write Clear
No
This register records all error conditions detected in the GART logic after the first error. Errors
recorded in FERR_GART are not recorded here.
Bits
Description
7:0
See FERR_GART for definitions of these bits.
2.4.5.7
PAC_ERR: PCI Address & Cmd First Error
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
A0h
0000000000h each
Yes
Size:
Attribute:
Locked:
64 bits
Read/Write
No
These registers record and latch the Address and Command information on the PCI Bus for the first
error detected.
Bits
63:46
45
Description
reserved(0)
PCI Parity (2nd phase of DAC, not defined for non-DAC address).
PCI Parity (if DAC, this is the parity of the first half of the address).
PCI Command - Command of Error.
44
43:40
39:0
PCI Address - Address Received on Error. (possible DAC address).
2.4.5.8
PD_ERR: PCI Data First Error
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
A8h
00h each
Yes
Size:
Attribute:
Locked:
64 bits
Read/Write
No
These registers record and latch the Data and Byte Enable information on the PCI Bus for the first
error detected.
Bits
63:37
36
Description
reserved(0)
PCI Parity.
35:32
31:0
PCI Byte Enable [3:0] - Byte Enable of Error.
PCI Data - Data of Error.
2-26
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
2.4.6
WXB
2.4.6.1
ERRSTS: Error Status Register
Address Offset:
Default Value:
Sticky
44h
00h
Size:
Attribute:
8 bits
Read/Write Clear,
This register records certain error conditions detected from the PCI bus. This register is sticky
through reset; that is, the contents of the register remain unchanged during and following the
assertion of XRST#. This allows system recovery software invoked following a forced reset to
examine the flags to determine the cause of an error. Once set, the flags remain set until explicitly
cleared by software or a power-good reset.
Note: Bits 6, 4, and 2 are Reserved on side-b and records a zero value only!
Bits
Description
7
INTRQ Asserted Flag
This flag is set if the WXB has initiated an INTRQ interrupt event. This bit remains set,
and the event signaled, until explicitly cleared by software writing a 1 to this bit.
Default = 0.
6
XBINIT Asserted Flag
This flag is set if the WXB has asserted XBINIT#. This bit remains set, and the event
signaled, until explicitly cleared by software writing a 1 to this bit. Default = 0.
Note: This bit is Reserved on side-b and records a zero value only!
5
NEPCI Register Records an PCI Bus Error Flag
This flag is set when the PCI bus reports an error (e.g. data parity error) on transactions
to/from other PCI bus agents. This bit remains set until explicitly cleared by software
writing a 1 to this bit. This bit is only set when the error is reported through the NEPCI
Next Error register, indicating that the error is not the first error occurrence since the
First Error register was last cleared. Default = 0.
4
3
reserved (0)
FEPCI Register Records an PCI Bus Error Flag
This flag is set when the PCI bus reports an error (e.g. data parity error) on transactions
to/from other PCI bus agents. This bit remains set until explicitly cleared by software
writing a 1 to this bit. This bit is only set when the error is reported through the FEPCI
register, indicating that the error is the first error occurrence since the FEPCI register
was last cleared. Default = 0.
2
1
reserved (0)
Performance Monitor #1 Event Flag
This flag is set when the Performance Monitor #1 detects an event. The
PCI_WXB_PMC1 registers describes the conditions that can cause this to occur. This
bit remains set until explicitly cleared by software writing a 1 to this bit. Default = 0.
0
Performance Monitor #0 Event Flag
This flag is set when the Performance Monitor #0 detects an event. The
PCI_WXB_PMC0 registers describes the conditions that can cause this to occur. This
bit remains set until explicitly cleared by software writing a 1 to this bit. Default = 0.
Intel® 460GX Chipset Software Developer’s Manual
2-27
Register Descriptions
2.4.6.2
ERRCMD: Error Command Register
Address Offset:
Default Value:
45h– 46h
8040h
Size:
Attribute:
16 bits
Read/Write
This register provides extended control over the signalling of errors through SERR_OUT#,
XBINIT#, and INTRQ#. These controls are in addition to the defined controls specified in the PCI-
standard PCICMD register for SERR# assertion.
Bits
Description
15
XBINITO: XBinit Override Enable
This bit should always be initially set to 0 by software. If set to 0, XBINIT# may be
asserted by the WXB. The WXB will automatically set this bit after an XBINIT# is
signaled. Default = 1
Note: Software should verify that there are no errors pending (by evaluating the ERRSTS register) before
clearing this bit.
Note: This bit is Reserved on side-b and records a value of one only!
14
reserved(0)
Note: This bit is Reserved on side-b and records a value of zero only!
13
IRQE: INTRQ Enable
Controls the reporting of WXB transmitted data errors. If set, the WXB will assert an
INTRQ interrupt for observed PERR# (including IHPC-driven parity errors) and data
parity errors detected in outbound transactions (e.g. Internal Queue Error detected
during read by PCI interface). Default = 0.
12
11
10
ASAPE: Assert SERR# on Address Parity Error
This bit should always be set to 1. When the WXB detects a PCI Address Parity Error
and both SERRE and PERRE are set, SERR# (and SERR_OUT#) will be signaled.
Default = 0.
ASDPE: Assert SERR# on any Data Parity Error
If set, the WXB will assert SERR# (and SERR_OUT#) whenever a data parity error is
detected in an inbound transaction. The SERRE bit in the PCICMD register must also be
set for SERR# (and SERR_OUT#) to be signaled. Default = 0.
ASDTE: Assert SERR# on Discard Timer Expiration
This bit should always be set to 1. When an inbound read Discard Timer Expiration
occurs and SERRE is set, SERR# (and SERR_OUT#) will be signaled. Default = 0.
9:7
6
reserved(0)
reserved(1)
reserved (0)
5:0
2.4.6.3
FEPCI: PCI Bus First Error Status Register
Address Offset:
Default Value:
83h
00h
Size:
Attribute:
8 bits
Read/Write Clear,
Sticky
This register records and latches the first error observed on the PCI bus. Once an error has been
noted in this register, no further updates are allowed. This register is a write-1-to-clear register,
meaning that software must write a 1 to the specific bit location it wishes cleared. The response to
2-28
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
each of these errors varies and is (generally) controlled through a combination of the PCICMD and
errors via the SERR#, XBINIT#, or INTRQ# outputs.
Note, if multiple errors are observed very close in time multiple errors may be signaled as a first
error.
Bits
Description
7
PCILV: PCI Error Logs Valid
This flag is set when an error has been logged in the FEPCIAL and FEPCIDL log
registers. The FEPCI status bit that lead to the logging must be cleared prior to clearing
the PCILV flag. Default = 0.
6
5
4
3
UMATA: Unexpected Master or Target Abort
This flag is set when an unexpected master abort or any target abort occurs. Master
aborts are reported as described for the RMA bit in the PCISTS register. Target aborts
through the RTA bit. Default = 0.
DTE: Discard Timer Expiration
This flag is set when the 215 clock timeout timer expires. This flag may be configured to
assert SERR#, XBINIT#, or an INTRQ interrupt through the ERRCMD register.
Default = 0.
SES: System Error Signaled
This flag is set when the a PCI agent other than the WXB asserts SERR#. This flag may
be configured to assert XBINIT#, or an INTRQ interrupt through the ERRCMD
register. Default = 0
PODT: PERR# Observed on PCI Data Transfer
This flag is set if the WXB detects the PERR# input asserted, and the WXB was not the
asserting agent. This flag may be configured to assert SERR#, XBINIT#, or an INTRQ
interrupt through the ERRCMD register. Default = 0.
2
1
reserved (0)
PEOD: Parity Error on Received PCI Data
This flag is set if the WXB detects an parity error (i.e. calculates a parity different from
what is provided with the data) on data being sent to the WXB on the PCI bus (from a
master read or target write). Default = 0.
0
PEPA: Parity Error on PCI Address
This flag is set if the WXB detects a parity error on the PCI address. Default = 0.
2.4.6.4
NEPCI: PCI Bus Next Error Status Register
Address Offset:
Default Value:
Sticky
87h
00h
Size:
Attribute:
8 bits
Read/Write Clear,
This register records any PCI bus errors detected after the first error is observed and recorded in the
FEPCI register. This register is a write-1-to-clear register, meaning that software must write a 1 to
the specific bit location it wishes cleared. The response to each of these errors varies and is
(generally) controlled through a combination of the PCICMD and ERRCMD registers. Error
logging is not performed for Next Error occurrences.
Bits
Description
7:0
See the FEPCI register description for definitions. Error logging is not performed for
Next Error occurrences.
Intel® 460GX Chipset Software Developer’s Manual
2-29
Register Descriptions
2.4.6.5
FEPCIAL: PCI First Error Address/Command Log
Address Offset:
Default Value:
Sticky
A5h–ADh
000000000000000000h
Size:
Attribute:
72 bits
Read/Write Clear,
These registers record and latch the address/command information, sent or received, associated
with a PCI Bus error for the first error detected.
Bits
Description
reserved (0)
C/BE[3:0]
AD[63:32]
AD[31:0]
71:68
67:64
63:32
31:0
2.4.6.6
FEPCIDL: PCI First Error Data Log
Address Offset:
Default Value:
Sticky
AFh – B3h
0000000000h
Size:
Attribute:
40 bits
Read/Write Clear,
These registers record and latch the PCI information, sent or received, specifically associated with
the PCI bus error for the first error detected. The recorded data contains the upper or lower AD and
C/BE and PAR signals.
Bits
39:37
36
Description
reserved (0)
PAR
35:32
31:0
C/BE
AD
2.5
Performance Monitor Registers
2.5.1
SAC
2.5.1.1
IT_MON_PMD_[0 to 5]: Internal Transaction Performance Monitor
Data Register
Bus CBN, Device Number: 00h
Function:
Size:
2
Address Offset:
90-97h, 98-9Fh,
64 bits each
A0-A7h, A8-AFh,
B0-B7h, B8-BFh
0 each
Default Value:
Sticky:
Attribute:
Locked:
Read/Write
No
No
The value written to this address, loads the counter and is also saved in a reload register. Each
counter can be configured to reload the data when it overflows.
2-30
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
The IT_MON_PMD_[0 to 5] registers hold the performance monitoring count values. 39-bits of
the counter are used for event counting, the 40th-bit is used as a overflow detection bit. The 39-bit
count value allows up to 70 minutes of event collection at 133 MHz. Event selection is controlled
Each counter may be stopped/started independently, using the controls available in the associated
PMC register.
Bits
63:40
39
Description
reserved(0)
Overflow
This bit is asserted when the Event Count bit 38 carries into bit 39.
38:0
Count Value
This register contains the Performance Monitor Data Register. You may preset the value
of the performance counter by writing to this register. You may read back the value of the
performance counter by reading this register.
2.5.1.2
IT_MON_PMC_[0 to 5]: Internal Transaction Performance Monitor
Config. Register
Bus CBN, Device Number: 00h
Function:
Size:
2
Address Offset:
D0-D7h, D8-DFh,
64 bits each
E0-E7h, E8-EFh,
F0-F7h, F8-FFh
0h each
Default Value:
Sticky:
Attribute:
Locked:
Read/Write
No
No
The IT_MON_PMC_[0 to 5] Registers specify the configuration of the Internal Transaction
Performance Monitors. This includes specifying Event Selection, Unit Mask, Enable & Disable
Source and Reload Control.
Bits
Description
63:41
40:33
reserved(0)
Length Encodings
0000 0000b Any Length Transaction
0111 0000b 0 - 8 Bytes
0111 0001b 16 Bytes
0111 0010b 32 Bytes
0111 0111b 48 Bytes
0111 0011b 64 Bytes
0110 0000b < 32 Bytes
1000 0011b < 64 Bytes
Note: Length Encodings only used during the following Performance Monitor Setting:
Mem Read - Delayed
Mem Read - Deferred Reply
Memory Write
32:24
DMASK Encodings
0 0001 1010 - Configuration Space - Monitor transactions Destined for Config Block
0 0000 1100 - Memory - Monitor transactions Destined for Memory
0 0001 1110 - Broadcasts - Monitor Broadcasts
1 0000 0000 - All Destinations - Monitor all Destinations
Intel® 460GX Chipset Software Developer’s Manual
2-31
Register Descriptions
23:15
UMASK Encodings
0 0000 0000 - Processor 0 - Monitor transactions originating from Processor 0
0 0000 0010 - Processor 1 - Monitor transactions originating from Processor 1
0 0000 0100 - Processor 2 - Monitor transactions originating from Processor 2
0 0000 0110 - Processor 3 - Monitor transactions originating from Processor 3
0 1000 0000 - All Processors - Monitor transactions originating from all Processors
0 0001 1010 - Configuration Space - Monitor transactions originating from Config Block
1 0000 0000 - All Initiators - Monitor all transactions
14:8
Event Select
Note: In the fields below bit 13 sometimes has an ’R’. This stands for retry. If ’R’ is set to a 1 then the
count is the number of occurrences of the event that were retried. If ’R’ is set to a 0, then the count
is for non-retried occurrences of the event. For example 110_0000b would be retried Interrupt Acks
and 100_0000b is for non-retried Interrupt Acks.
Selects the event to be monitored.
000 0000b Monitoring Disabled.
011 1111b All Clocks.
1R0 0000b Interrupt Ack
1R1 0001b Special Transaction and Defer Reply to Write.
1R0 0010b I/O Read
1R0 0011b I/O Write - Deferred Reply
1R1 0011b I/O Write - Posted.
1R0 0101b Purge TC and reserved and Branch Trace Messages
1R1 0101b Outbound Interrupt or Hard Fail Write Completion
1R0 0110b Mem Read
1R1 0111b Memory Write
110 1000b Check Connection
1R0 1010b Cfg Read
1R0 1011b Cfg Write - Deferred Reply
1R1 1011b Cfg Write - Posted
000 1100b Inbound Interrupt
1R0 1110b Locked Read - In Order
1R1 1110b Locked Read - Delayed
1R1 1111b Mem Write-Line
1R1 0000b Normal Read Completion
1R1 0100b Memory Scrub
1R0 1100b SSBR
1R1 1101b RSBR
000 0001b Snoop events
000 0010b Snoop Stall events caused by 460GX chipset
000 0011b Snoop Stall events caused by processors
000 0100b Snoop Stall events by either CPU’s or GX
000 0101b Hit events from CPU
000 0110b BNR events from GX
000 0111b BNR events from CPU
000 1000b BNR events from GX or CPU
000 1001b BPRI Clocks from GX
000 1010b BPRI Events from GX
000 1011b BPRI ADS’s from GX
001 0001b Speculative reads that are not used because of HITW
001 0010b Speculative reads that are not used because of HITM
001 0011b Speculative reads that are not used because of retry
001 0100b Speculative reads that are not used because read was restarted
001 0101b Speculative reads that were successful
001 0110b Speculative reads that were not done speculatively
2-32
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
001 0111b Memory Read that was to be retried and received a HITM
001 1000b Memory Read with active OWN# and received a HITM
001 1001b Memory Read from a CPU that received a HITW
001 1010b Memory Read from a CPU that received a HITM
001 1011b Memory Read from PCI that received a HITW
001 1100b Memory Read from PCI that received a HITM
001 1101b Memory Write from PCI that received a HITM
010 0001b EV0 Events
010 0010b EV0 Clocks
010 0011b EV1 Events
010 0100b EV1 Clocks
7
reserved(0).
6:5
Disable Source
Selects event that will disable the performance monitor.
00b Never Disable.
01b Disable when this counter is overflowed (Note that if this setting is used, and bits
[4:3] are set to ’Enable Always’, then when this counter is overflowed, counting
will not resume until the counter is loaded with a non-overflow value).
10b Disable on falling edge of SAC Event 0.
11b Disable on falling edge of SAC Event 1.
4:3
2:0
Enable Source
Selects event that will enable the performance monitor.
00b Never Enable.
01b Enable Always (Disable events overrides this setting and will disable counting).
10b Enable on rising edge of SAC Event 0.
11b Enable on rising edge of SAC Event 1.
Reload Control
Selects event that will control the Reloading of the performance monitor with the value
written into the associated PMD2 register.
000b Never Reload.
001b Reload when counter overflows.
010b Reload on SAC Event 0 Asserted.
011b Reload on SAC Event 1 Asserted.
100b Reload on SAC Event 0 Asserting edge.
101b Reload on SAC Event 1 Asserting edge.
Note: When counting retried reads (code of 110_0110b) the logic will count reads that were to be retried,
but got a HITM# and therefore the HITM# overrides the retry. For an exact count of reads that are
truly retried, count retried reads (110_0110b) and subtract out the number of reads that were to be
retried but got a HITM# (code of 001_0111b). This give the exact number of reads that were retried
on the bus.
Intel® 460GX Chipset Software Developer’s Manual
2-33
Register Descriptions
2.5.2
SDC
2.5.2.1
FSB_D_PMC_[1,0]: System Bus Performance Monitor Configuration
Register
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
98-9Ah, 9C-9Eh
000000h each
Size:
Attribute:
24 bits each
Read/Write
The FSB_D_PMC_[1,0] Registers specify the configuration of the SDC system bus performance
monitors. This includes specifying Event Selection, Unit Mask, Enable & Divisible Source &
Reload Control.
Bits
Description
23:17
16:15
reserved(0).
Mask.
This field contains the Event Specific Mask Bits. This allows qualifying event collection
by the issuing agent of the transaction.
00b reserved.
01b Monitor only if 460GX chipset initiated the transaction.
10b Monitor only if 460GX chipset did not initiate the transaction.
11b Monitor all transactions regardless of issuing agent.
14:8
Event Select.
Selects the event to be monitored.
000 0000b Monitoring Disabled.
000 0001b System Bus Clocks.
000 0010b DBSY# Clocks.
100 0010b DBSY# Events.
000 0011b DRDY# Clocks.
100 0011b DRDY# Events.
100 0100b DBSY# and not(DRDY#) Events.
000 0101b TRDY# Clocks.
100 0101b TRDY# Events.
000 0110b TRDY# asserted when DBUSY# asserted (clocks)
100 0110b TRDY# asserted when DBUSY# asserted (events)
000 0111b Read from SDC waiting on processor write (clocks)
100 0111b Read from SDC waiting on processor write (events)
000 1000b Read from SDC waiting on an SDC read to complete (clocks)
100 1000b Read from SDC waiting on an SDC read to complete (events)
100 1001b reserved
100 1010b reserved
000 1011bEvent Logic 0 Active Clocks
100 1011b Event Logic 0 Events
000 1100b Event Logic 1 Active Clocks
100 1100b Event Logic 1 Events
7
reserved(0).
6:5
Disable Source.
Selects event that will disable the performance monitor. Note that the disable and enable
sources are edge-triggered for Event 0 or 1.
00b Never Disable.
2-34
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
01b Disable when counter overflows.
10b Disable on falling edge (Deassertion) of SDC Event 0.
11b Disable on falling edge (Deassertion) of SDC Event 1.
4:3
Enable Source.
Selects event that will enable the performance monitor.
00b Never Enable (in this mode the counter will never count).
01b Enable Always (note that if this setting is used, the disable events in bits [6:5] will
only disable counting for one clock, and then the counter resumes. When bits [4:3]
are set to ’Enable Always’, the only meaningful setting for bits [6:5] is
’Never Disable’.
10b Enable on rising edge (Assertion) of SDC Event 0.
11b Enable on rising edge (Assertion) of SDC Event 1.
2:0
Reload Control.
Selects event that will control the Reloading of the performance monitor with the value
written into the associated PMD1 register.
000b Never Reload.
001b Reload when counter overflows.
010b Reload on SDC Event 0 Asserted.
011b Reload on SDC Event 1 Asserted.
100b Reload on SDC Event 0 Asserting edge.
101b Reload on SDC Event 1 Asserting edge.
2.5.2.2
FSB_D_PMD_[1,0]: System Bus Performance Monitor Data Registers
Bus CBN, Device Number: 04h
Address Offset:
Default Value:
A0-A7h, A8-AFh
0 each
Size:
Attribute:
64 bits each
Read/Write
Two performance monitoring counters, with associated event selection and control registers, is
provided in the SDC component. These counters may be configured to track system bus events.
Event detection may be configured to increment a counter, affect performance monitoring pins, and
issue an interrupt request on counter overflow.
The value written to this address, loads the counter and is also saved in a reload register. Each
counter can be configured to reload the data when it overflows.
The FSB_D_PMD_[1,0] registers hold the performance monitoring count values. 39-bits of the
counter are used for event counting, the 40th-bit is used as a overflow detection bit. The 39-bit
count value allows up to 70 minutes of event collection at 133 MHz. Event selection is controlled
by the PMC registers.
Each counter may be stopped/started independently, using the controls available in the associated
PMD register.
Bits
63:40
39
Description
reserved(0)
Overflow
This bit is asserted when the Event Count bit 38 carries into bit 39.
38:0
Count Value
This register contains the Performance Monitor Data Register. You may preset the value
of the performance counter by writing to this register. You may read back the value of the
performance counter by reading this register.
Intel® 460GX Chipset Software Developer’s Manual
2-35
Register Descriptions
2.5.3
PXB
2.5.3.1
PMD[1:0]: Performance Monitoring Data Register
Address Offset:
Default Value:
D8-DBh, E0-E3h
0000_0000h each
Size:
Attribute:
32 bits each
Read/Write
Two performance monitoring counters, with associated event selection and control registers, are
provided for each PCI bus in the PXB. Each counter may be configured to track PCI bus events.
Event detection may be configured to increment a counter, toggle a pin on event or counter
overflow, and issue an interrupt request on counter overflow.
The PMD registers hold the performance monitoring count values. These registers are 32-bit
counters. Event selection is controlled by the PME registers, and the action performed on event
detection is controlled by the PMR registers.
Each counter may be stopped/started independently, using the controls available in the associated
PMR register.
Bits
Description
Count Value
31:0
2.5.3.2
PMR[1:0]: Performance Monitoring Response
Address Offset:
Default Value:
DDh, E5h
0000h each
Size:
Attribute:
8 bits each
Read/Write
There are two PMR registers for each PCI bus, one for each PMD counter. Each PMR register
specifies how the event selected by the corresponding PME register affects the associated PMD
register, P(A,B)MON# pins, and the INT(A,B)RQ# pins.
Bits
Description
7:6
Interrupt Assertion
Defines how selected event affects INTRQ# assertion. Whenever INTRQ# is asserted, a
flag for this counter is set in the Error Status Register, so that software can determine the
cause of the interrupt. This flag is reset by writing the Error Status Register.
0
1
2
3
Selected event does not assert INTRQ #
reserved
Assert INTRQ# pin when event occurs
Assert INTRQ# pin when counter overflows
5:4
3:2
Performance Monitoring pin assertion
Defines how the selected event affects the PMON# pin for this counter.
0
1
2
3
PMON# pin is tristated. Selected event has no effect.
reserved
Assert this counter’s PMON# pin when event occurs
Assert this counter’s PMON# pin when counter overflows
Count Mode
Selects when the counter is updated for the detected event.
0
1
2
3
Stop counting.
Count each cycle selected event is active.
Count on each rising edge of the selected event.
Trigger. Start counting on the first rising edge of the selected event, and
continue counting each clock cycle.
2-36
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
Once configured to count, all counters in the SAC and each PXB can be (nearly)
simultaneously started and stopped using a separate enable.
1:0
Reload Mode
Reload has priority over increment. That is, if a Reload event and a count event happen
simultaneously, the count event has no effect.
0
1
2
3
Never Reload
Reload when this counter overflows.
Reload when the other counter overflows.
Reload unless the other counter increments.
2.5.3.3
PME[1:0]: Performance Monitoring Event Selection
Address Offset:
Default Value:
E8 - EBh
0000h each
Size:
Attribute:
16 bits each
Read/Write
The PME registers specify the particular event to track in the performance monitoring counters.
The PXB supports tracking of PCI bus transactions (both specific and generic), and PCI bus signal
assertion. Bus transactions may be qualified by the originating agent and transaction destination.
Accumulated event counts are held in the PMD registers, while the PMR registers specify the
action to performed on event detection.
Bits
15
Description
reserved (0)
14
Count Data Cycles
1: Count data cycles associated with selected event
0: Count the selected event
13:10
Initiating Agent Selection
This field qualifies the tracking of bus transactions by limiting event detection to those
transactions issued by specific agents. That is, unless otherwise noted for the specific
event selected (below), the agent initiating the bus transaction must match the selection
specified here for the transaction to be tracked.
0000
0001
0010
0011
0100
0101
0110
Agent 0
Agent 1
Agent 2
Agent 3
Agent 4
Agent 5
reserved
1000
1001
1010
1011
1100
1101
1110
reserved
reserved
reserved
reserved
reserved
South bridge
460GX chipset agent (i.e.
outbound)
Any agent
0111
reserved
1111
Note: This field is applicable only if the PCI bus is operated in internal-arbiter mode. If the bus is
operated using an external arbiter, this field must be set to Any Agent to trigger any events.
9:8
Transaction Destination Selection
This field qualifies the tracking of bus transactions by limiting event detection to those
transactions directed to a specific resource. That is, unless otherwise noted for the
specific event selected (below), the source or destination of the data must match the
selection specified here for the transaction to be tracked.
00
01
any
10
11
PCI Target
Parallel Segment Peer
Main Memory
7:6
reserved
Intel® 460GX Chipset Software Developer’s Manual
2-37
Register Descriptions
5:0
Event Selection
This field specifies the basic PCI bus transaction or PCI bus signal to be monitored.
Individual Bus Transactions
00 0000 reserved
00 0001 reserved
00 0010 I/O Read
00 0011 I/O Write
00 0100 reserved
00 0101 reserved
00 0110 Memory Read
00 0111 Memory Write
00 1000 reserved
00 1001 reserved
00 1010 reserved
00 1011 reserved
00 1100 Memory Read Multiple
00 1101 Dual Address Cycle
00 1110 Memory Read Line
00 1111 Memory Write & Invalidate
Generic (Grouped) Bus Transactions
010 000 Any bus transaction
010 001 Any memory transaction
010 100 Any I/O transaction
010 101 Any I/O or memory
transactions
010 010 Any memory read
010 011 Any memory write
010 110 Any I/O or memory read
010 111 Any I/O or memory write
Bus Signal Assertions
011 000 reserved
011 001 reserved
011 010 RETRY1
011 011 reserved
011 100 reserved
011 101 reserved
011 110 LOCK
011 111 ACK64
All other encodings are reserved.
NOTE:1. Counting data cycles is undefined for this selection.
2.5.4
GXB
2.5.4.1
AGP_PMD_0,1: AGP Performance Monitor Data Registers
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
50h, 58h
0
Size:
Attribute:
Locked:
64 bits
Read/Write
No
No
This counter may be configured to track AGP bus events as well as events internal to the GXB.
Event detection may be configured to increment a counter, affect performance monitoring pins, and
issue an interrupt request on counter overflow.
The value written to this address, loads the counter and is also saved in a reload register. Each
counter can be configured to reload the data.
The AGP_G_PMD register holds the performance monitoring count value. 39-bits of the counter
are used for event counting, the 40th-bit is used as a overflow detection bit. The 39-bit count value
allows up to 70 minutes of event collection at 133 MHz. Event selection is controlled by the PMC
registers.
Each counter may be stopped/started independently, using the controls available in the associated
PMD register.
Bits
63:40
39
Description
reserved(0)
Overflow
This bit is asserted when the Event Count bit 38 carries into bit 39.
2-38
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
38:0
Count Value
This register contains the Performance Monitor Data Register. You may preset the value
of the performance counter by writing to this register. You may read back the value of the
performance counter by reading this register.
2.5.4.2
PCI_PMD: PCI Performance Monitor Data Registers
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
60h
0
Size:
Attribute:
Locked:
64 bits
Read/Write
No
No
This counter may be configured to track PCI bus events as well as events internal to the GXB.
Event detection may be configured to increment a counter, affect performance monitoring pins, and
issue an interrupt request on counter overflow.
The value written to this address, loads the counter and is also saved in a reload register. Each
counter can be configured to reload the data.
The PCI_PMD register holds the performance monitoring count value. 39-bits of the counter are
used for event counting, the 40th-bit is used as a overflow detection bit. The 39-bit count value
allows up to 70 minutes of event collection at 133 MHz. Event selection is controlled by the PMC
registers.
Each counter may be stopped/started independently, using the controls available in the associated
PMD register.
Bits
63:40
39
Description
reserved(0)
Overflow
This bit is asserted when the Event Count bit 38 carries into bit 39.
38:0
Count Value
This register contains the Performance Monitor Data Register. You may preset the value
of the performance counter by writing to this register. You may read back the value of the
performance counter by reading this register.
2.5.4.3
PERCON: Performance Monitor Control Register
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
E0h
00h
Size:
Attribute:
Locked:
8 bits
Read/Write
No
No
The PERCON Register allows one software write to start and stop the performance monitors. The 2
bits are fed into the Event Logic generation. Bit 0 feeds Event 0 and bit 1 feeds Event 1. Since the
counters can be enabled and disabled by Events 0 or 1, then one write can start or stop all the
counters together.
Bits
7:2
1
Description
reserved (0)
Event 1 Input
This bit is fed as an input into Event 1 logic. This bit is OR’ed with any other logic
generating Event 1, guaranteeing that if this bit is set, then Event 1 will be asserted.
Intel® 460GX Chipset Software Developer’s Manual
2-39
Register Descriptions
0
Event 0 Input
This bit is fed an input into Event 0 logic. This bit is OR’ed with any other logic
generating Event 0, guaranteeing that if this bit is set, then Event 0 will be asserted.
2.5.4.4
AGP_PMC_[0,1]: AGP Performance Monitor Configuration Register
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
ECh, F0h
000000h
No
Size:
Attribute:
Locked:
32 bits
Read/Write
No
The AGP_G_PMC Register specifies the configuration of the GXB AGP Performance Monitor.
This includes specifying Event Selection, Unit Mask, Enable & Disable Source and Reload
Control.
Bits
Description
31:24
23:20
19:18
’n’ for threshold Monitoring
reserved(0).
Events to monitor, when Event Select Code points to events. Data Transfer is one
QW (8 bytes). Used also for code 11 0000
00 - All events
01 - Events with ’n’ Data Transfer Cycles
02 - Events with < ’n’ Data Transfer Cycles
03 - Events with > ’n’ Data Transfer Cycles
17:16
Pipe or Sideband Request Mask
00b reserved.
01b Selects Low Priority.
10b Selects High Priority.
11b Selects Both.
15:14
13:8
reserved(0).
Event Select
Selects the event to be monitored.
00 0000b Monitoring Disabled.
00 0001b AGP Read Request Events.
00 0010b AGP Write Request Events.
00 0011b All AGP Request Events (does not include Flush or Fence requests).
00 0101b Singe Read that splits 4k page boundary.
00 0110b Single Write that splits line boundary.
01 0000b Flushes (events)
01 0001b Fences (events)
01 0010b RBF# active events.
01 0011b Events of write wait states inserted by AGP card
01 0100b Events of LPTT timeout with another request active.
01 0101b Events of MTT timeout with another request active.
01 0111b GART Aperture Misses
10 0000b AGP Clocks.
10 0100b Count AGP clocks that RBF# is stalled
10 0101b Count AGP clocks that low-priority read buffer is physically empty.
10 0110b Count AGP clocks that high-priority read buffer is physically empty.
10 0111b Count AGP clocks that both the low and high priority buffers are empty.
11 0000b Count AGP clocks that there are <=> ’n’ requests queued; using bits 31:24 and
19:18.
2-40
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
7
EVENT1 Count Enable
If set, then this bit over-rides bits 13:8. If set, then AGP_PMD_0 will count the number
of occurrences of EVENT1 and AGP_PMD_1 will count the number of clocks that
EVENT1 is active. When this bit is not set, then the 2 counters are controlled by bits 13:8.
6:5
Disable Source
Selects event that will disable the performance monitor.
00b Never Disable.
01b Disable when counter overflows.
10b Disable on falling edge of GXB Event 0.
11b Disable on falling edge of GXB Event 1.
4:3
2:0
Enable Source
Selects event that will enable the performance monitor.
00b Never Enable.
01b Enable Always (with this setting, the only meaningful setting for bits 6:5 is 00b).
10b Enable on rising edge of GXB Event 0.
11b Enable on rising edge of GXB Event 1.
Reload Control.
Selects event that will control the Reloading of the performance monitor with the value
written into the associated PMD register.
000b Never Reload.
001b Reload when counter overflows.
010b Reload when GXB Event 0 Asserted.
011b Reload when GXB Event 1 Asserted.
100b Reload on GXB Event 0 Asserting edge.
101b Reload on GXB Event 1 Asserting edge.
2.5.4.5
PCI_PMC: PCI Performance Monitor Configuration Register
Function Number:
Address Offset:
Default Value:
Sticky:
BFN+1
F4h
000000h
No
Size:
Attribute:
Locked:
24 bits
Read/Write
No
The PCI_PMC Register specifies the configuration of the GXB PCI Performance Monitor. This
includes specifying Event Selection, Unit Mask, Enable & Disable Source and Reload Control.
NOTE: the Event Select field determines whether clocks or events are counted. In the case of
events, then all events are counted. In the case of clocks, all clocks in which the selection is active
are counted.
Bits
Description
23:18
17:16
reserved(0).
Initiating Agent
00b reserved.
01b Outbound.
10b Inbound.
11b Selects Both.
15:14
13:8
reserved(0).
Event Select
Selects the event to be monitored.
00 0000b Monitoring Disabled
Intel® 460GX Chipset Software Developer’s Manual
2-41
Register Descriptions
00 0010b All PCI Clocks
00 0100b Idle Bus Cycles
00 0111b All Disconnect Events
00 1000b Lock Asserted Clocks
00 1001b Lock Asserted Events
00 1011b I/O Reads - Events
00 1101b I/O Writes - Events
00 1111b Memory Read Events
01 0001b Memory Write Events
01 0011b SRAM Read Events
01 0101b SRAM Write Events
01 0111b Write Combining Events
01 1000b WBF# Active - Clocks
01 1001b WBF# Active - Events
01 1011b Retry Read that’s not Delayed - Events
01 1101b Retry Write because no Write Slot available - Events
01 1110b Count PCI clocks waiting (Devsel and !IRDY and TRDY)
10 0000b Count PCI clocks waiting (Devsel and IRDY and !TRDY)
10 0010b Count PCI clocks data is transferring
7
reserved(0).
6:5
Disable Source
Selects event that will disable the performance monitor.
00b Never Disable.
01b Disable when counter overflows.
10b Disable on falling edge of GXB Event 0.
11b Disable on falling edge of GXB Event 1.
4:3
2:0
Enable Source
Selects event that will enable the performance monitor.
00b Never Enable.
01b Enable Always (with this setting, the only meaningful setting for bits 6:5 is 00b).
10b Enable on rising edge of GXB Event 0.
11b Enable on rising edge of GXB Event 1.
Reload Control
Selects event that will control the Reloading of the performance monitor with the value
written into the associated PMD register.
000b Never Reload.
001b Reload when counter overflows.
010b Reload on GXB Event 0 Asserted.
011b Reload on GXB Event 1 Asserted.
100b Reload on GXB Event 0 Asserting edge.
101b Reload on GXB Event 1 Asserting edge.
2-42
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
2.5.5
WXB
2.5.5.1
PCI_WXB_PMC0: PCI Performance Monitor Configuration Register
Address Offset:
Default Value:
DCh – DFh
00000000h
Size:
Attribute:
32bits
Read/Write
This register controls the PCI performance monitors. There are two performance monitors for each
PCI bus. This register defines the events to be monitored, and when monitoring should start and
stop. The selected event can be qualified by data transferred, and issuing agent.
Bits
Description
31:24
23:21
reserved (0)
Data Transfer and Transaction Qualifier
Qualifies a selected “Packet Type” by the presence or absence of data transferred.
000bAll events
001bRetry - for any reason
010bRetry - no buffers available (inbound read or write transactions only)
011bRetry - no data available (inbound read transactions only)
101bLocked
110bDual Address Cycles
20:19
18:17
reserved (0)
Issuing Agent Qualifier
Monitor only those selected events issued by the following agent:
00breserved
01bOutbound: Issued by WXB
10bInbound: Not issued by WXB
11bAll:
Issued by any agent
16:11
Event Select
Selects the event(s) to be monitored.
Measurement
00 0000b
Monitoring Disabled
Transaction Types
01 0001bI/O Reads
01 0010bMemory Reads
01 0100bMem Read Lines
01 1000bMem Read Multiples
01 1111bAny Read
10 0001bI/O Writes
10 0010bMemory Writes
10 0100bMem Wr & Invalidates
10 1111bAny Write
11 1111bAny Transaction
10:4
3
reserved (0)
Enable Source
When this bit is set to 1, the performance monitoring logic is enabled. Default=0.
2:0
reserved (0)
Intel® 460GX Chipset Software Developer’s Manual
2-43
Register Descriptions
2.5.5.2
PCI_WXB_PMC1: PCI Performance Monitor Configuration Register
Address Offset:
Default Value:
E8h – EBh
00000000h
Size:
Attribute:
32bits
Read/Write
This register controls the PCI performance monitors. There are two performance monitors for each
PCI bus. This register defines the events to be monitored and when monitoring starts and stops.
The selected event can be qualified by data transferred, and issuing agent.
Bits
Description
31:0
See PCI_WXB_PMC0 definitions.
2.6
Interrupt Related Registers
2.6.1
SAC
2.6.1.1
XTPRS: External Task Priority Registers
Bus CBN, Device Number: 00h
Function:
Size:
Attribute:
Locked:
0
Address Offset:
Default Value:
Sticky:
C0-C7h
80h
No
64 bits
Read Only
No
The XTPRS are used to support redirectable interrupts. These registers are made observable to
software primarily for test and debug purposes. These registers will be updated by XTPR Update
Special Cycle on the system bus. The second cycle of the XTPR Update Special Cycle’s address
determines the value to load into the register. Ab[27:24]# is the 4 bit XTPR value. Ab[23:20]#
determines which register to update. Since the high priority agent reserves the uppermost agent ID
bit, only Ab[22:20]# are used. These 3 bits decode to one of the 8 registers. Ab[31] is the enable bit
for the register.
All registers default to 1000_0000b, which is the disabled state.
Each register is defined as:
Bit 7
- enable (1=disable, 0=enable)
Bits 6:4 - reserved (0)
Bits 3:0 - value of XTPR
Bits
Description
63:56
XTPR 7
These bits represent the external task priority for symmetric agent ID 07h.
55:48
47:40
39:32
31:24
23:16
XTPR 6
These bits represent the external task priority for symmetric agent ID 06h.
XTPR 5
These bits represent the external task priority for symmetric agent ID 05h.
XTPR 4
These bits represent the external task priority for symmetric agent ID 04h.
XTPR 3
These bits represent the external task priority for symmetric agent ID 03h.
XTPR 2
These bits represent the external task priority for symmetric agent ID 02h.
2-44
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
15:8
7:0
XTPR 1
These bits represent the external task priority for symmetric agent ID 01h.
XTPR 0
These bits represent the external task priority for symmetric agent ID 00h.
2.6.2
PID PCI Memory-mapped Registers
The PID uses two 32-bit memory-mapped registers to provide the indirect addressing access to its
(x)APIC interrupt redirection registers as well as to its ID, version, and arbitration ID registers. The
memory-mapped registers are placed at default addresses of FEC0_0000h, FEC0_0010h, and
FEC0_0040h for (x)APIC compatibility.
The I/O select register is used to provide the index of the internal register being accessed. The
register being accessed is determined by bits 7 through 0 of this register. The I/O window register is
used to provide/receive data associated with the access.
Table 2-2 summarizes the memory-mapped registers. Detailed descriptions of each register follow.
Note: The default base address FEC00 which is mapped to A[31:12] can be changed by reprogramming
the (x)APIC base address register via PCI configuration space access.
Table 2-2. Memory-Mapped Register Summary
Address
Name
Access
Default Value
FEC00000h
FEC00010h
FEC00040h
I/O Register Select Register
I/O Window Register
R/W
R/W
R/W
00000000h
00000000h
00000000h
(x)APIC EOI Register
2.6.2.1
I/O Register Select Register (FEC00000h)
The I/O register select register selects which indirect access register appears in the I/O window
register where it can be manipulated by software. The selector values for the indirect access
registers are listed in Section 2.6.3. Software programs bits 7 through 0 of this register to select the
desired internal register. The contents of the selected 32-bit register can be manipulated via the I/O
window register. The I/O register select register is read/write by software and its default is listed in
Section 2.6.2.2. This starting address, where the I/O register select register and I/O window register
reside, can be relocated to a different address via the APIC base address register. The format of the
Note: This register is defined as a 32-bit register, but only the lower eight bits are used. This register must
be accessed using 32-bit memory reads or writes. Bits 31 through 8 should be set to 0s.
Table 2-3. I/O Select Register Format
Register Offset: FEC00000hDefault Value: [00000000h]Attribute: Read/Write
Bit(s)
Name
Reserved
Description
31:8
7:0
These 24 bits are reserved.
REGISTER
ADDRESS
These eight bits provide the address offset of the internal 32-bit register. This
number is used to select consecutive 32-bit internal registers via the I/O
Intel® 460GX Chipset Software Developer’s Manual
2-45
Register Descriptions
2.6.2.2
I/O Window Register (FEC00010h)
This register is mapped onto the PID’s internal register that is selected by the I/O register select
register. Readability/writeability by software is determined by the characteristics of the internal
register that is currently selected. The format of the I/O window register is shown in Table 2-4
below. This register must be accessed using 32-bit read or write operations.
Table 2-4. I/O Window Register Format
Register Offset: FEC00010hDefault Value: [00000000h]Attribute: Read/Write
Bit(s)
Name
Description
31:0
IOWIN[31:0]
This 32-bit register contains the 32-bit write or read data value.
2.6.2.3
(x)APIC EOI Register (FEC00040h)
The (x)APIC EOI register provides a means of informing the PID that interrupt service has been
completed by the processor for a given interrupt vector (mapped to 1 of 64 interrupt inputs received
by the PID). The PID will compare this vector value with the vector field of each entry in the RT.
When a match is found, the RIRR bit for that entry will be cleared. In the case of level-triggered
interrupts, clearing the remote IRR bit will initiate a resampling of the corresponding interrupt
input signal. If the interrupt input is still asserted, the PID will issue a new level-triggered interrupt
message. The PID will therefore issue a single new interrupt message upon receiving an EOI write,
corresponding to the still asserted interrupt input pin. The PID only uses the (x)APIC EOI register
in SAPIC mode.
Note: If multiple redirection entries assign the same vector for more than one interrupt pin, each of those
pins will be resampled and new interrupt messages issued for those that are still asserted. This
register must be accessed using 32-bit read or write operations.
Table 2-5. (x)APIC EOI Register Format
Register Offset: FEC00040hDefault Value: [00000000h]Attribute: Read/Write
Bit(s)
Name
Description
31:8
7:0
Reserved
SAPICEOI[7:0]
Reserved
Interrupt vector
2.6.3
PID Indirect Access Registers
The PID provides several indirect access registers. These registers are accessed via the I/O select
and I/O window registers described above. The indirect access registers include the (x)APIC ID,
version, and arbitration ID registers, and 64 RTEs for each of the 64 interrupt inputs to the PID.
Registers at offsets 03h-0Fh are reserved and will return a 00h value when read.
Table 2-6 summarizes the indirect access registers. Detailed descriptions of each register follow.
2-46
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
2.6.3.1
I/O (x)APIC ID Register (00h)
Table 2-6. Memory-mapped Register Summary
Offset
Name
I/O (x)APIC ID Register
Access
Default Value
00h
01h
02h
R/W
R/O
R/O
R/O
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
00000000h
a
I/O (x)APIC Version Register
003F00vvh
I/O (x)APIC Arbitration ID Register
Reserved
RTE 0
00000000h
03h-0Fh
10h
12h
14h
16h
18h
1Ah
1Ch
1Eh
20h
22h
24h
26h
28h
2Ah
2Ch
2Eh
30h
32h
34h
36h
38h
3Ah
3Ch
3Eh
40h
42h
44h
46h
48h
4Ah
4Ch
4Eh
50h
00000000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
RTE 1
RTE 2
RTE 3
RTE 4
RTE 5
RTE 6
RTE 7
RTE 8
RTE 9
RTE 10
RTE 11
RTE 12
RTE 13
RTE 14
RTE 15
RTE 16
RTE 17
RTE 18
RTE 19
RTE 20
RTE 21
RTE 22
RTE 23
RTE 24
RTE 25
RTE 26
RTE 27
RTE 28
RTE 29
RTE 30
RTE 31
RTE 32
Intel® 460GX Chipset Software Developer’s Manual
2-47
Register Descriptions
Table 2-6. Memory-mapped Register Summary (Cont’d)
Offset
Name
Access
Default Value
52h
54h
56h
58h
5Ah
5Ch
5Eh
60h
62h
64h
66h
68h
6Ah
6Ch
6Eh
70h
72h
74h
76h
78h
7Ah
7Ch
7Eh
80h
82h
84h
86h
88h
8Ah
8Ch
8Eh
RTE 33
RTE 34
RTE 35
RTE 36
RTE 37
RTE 38
RTE 39
RTE 40
RTE 41
RTE 42
RTE 43
RTE 44
RTE 45
RTE 46
RTE 47
RTE 48
RTE 49
RTE 50
RTE 51
RTE 52
RTE 53
RTE 54
RTE 55
RTE 56
RTE 57
RTE 58
RTE 59
RTE 60
RTE 61
RTE 62
RTE 63
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
00000000_00010000h
a. vv is 13h in APIC mode of operation, 21h in SAPIC mode of operation.
The I/O (x)APIC ID register is read/write by software. On reset, this register’s contents are reset to
zero. This register is provided for APIC compatibility only and it does not serve any other purpose.
The PID’s (x)APIC ID register has a default value of 00000000h. This register should be
programmed with the correct (x)APIC ID value before using the PID in APIC mode. The (x)APIC
ARBID register is also written during a write to this register.
2-48
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
Table 2-7. I/O APIC ID Register Format
Register Offset: 00h
Default Value: [00000000h]Attribute: Read/Write
Description
Bit(s)
Name
Reserved
31:28
27:24
These four bits are reserved.
ID[3:0]
These four bits provide the APIC ID. This field is used by the I/O APIC unit of
the PID. In SAPIC or compatibility mode of operation, these bits are ignored.
23:16
15
Reserved
DT
These 24 bits are reserved.
This bit defines the delivery type. A ‘0’ in this field indicates APIC delivery
mechanism. A ‘1’ indicates SAPIC delivery mechanism. This bit reflects the
PICMODE strap pin.
14
LTS
Level deassert message support. This bit is always ‘0’ since the PID does not
support “level deassert” messages.
13:0
Reserved
Reserved.
2.6.3.2
I/O (x)APIC Version Register (01h)
The PID contains an I/O (x)APIC version register that identifies the type of I/O (x)APIC it
implements. Software can use this to provide compatibility between different I/O (x)APIC
implementations and their versions. The version register also contains the maximum RTE.
Table 2-8. I/O (x)APIC Version Register Format
Register Offset: 01h
Default Value: [003F00vvh]Attribute: Read-Only
Description
Bit(s)
Name
Reserved
31:24
23:16
These eight bits are reserved
MAX REDIR
This is the entry number (0 being the lowest entry) of the highest entry in the I/
O RT. It is equal to the number of interrupt input pins minus one that the PID
supports. This field is hardwired and is read-only. The PID sets this field to
3FH, indicating that it supports 64 RTEs.
15:8
7:0
Reserved
VERSION
These eight bits are reserved.
This is a version number that identifies the implementation version of the APIC
or SAPIC units in the PID. This field is hardwired and is read-only. When the
PID powers-up in APIC mode this field will return a value of 13H. When the
PID powers-up in SAPIC mode this field will return a value of 21H.
2.6.3.3
I/O (x)APIC Arbitration ID Register (02h)
This register contains the APIC bus arbitration priority for the PID. This register is loaded
whenever the PID’s (x)APIC ID register is loaded. A rotating priority scheme is used for APIC bus
arbitration. The winner of the arbitration becomes the lowest-priority agent and assumes an
arbitration ID of 0. All other agents except the agent whose ARBID is 15, increment their ARBID
by 1. The agent whose ARBID is 15 will take the winner’s ARBID and will increment it by 1.
ARBIDs are changed (incremented or assumed) only for messages that are successfully
transmitted. A message transmitted successfully means that no CS error or acceptance error was
reported for that message. The PID APIC ARBID is always loaded with the PID APIC ID during a
“INIT-level deassert” message.
Note: Only four bits are required for the APIC ARBID. Bits 27:24 are used for this ID.
Intel® 460GX Chipset Software Developer’s Manual
2-49
Register Descriptions
Table 2-9. I/O (x)APIC Arbitration ID Register Format
Register Offset: 02h
Default Value: [00000000h]Attribute: Read-Only
Description
Bit(s)
Name
Reserved
31:28
27:24
23:0
These four bits are reserved.
APIC Arbitration ID.
ARBID
Reserved
These 24 bits are reserved.
2.6.3.4
I/O (x)APIC RTE (10h-8Fh)
The interrupt RT has a dedicated entry for each interrupt input pin. Software can individually
choose the interrupt vector number for input pins. For each individual pin, the operating system can
also specify the signal polarity (low-active or high-active), whether the interrupt is signaled as
edges or levels, as well as the destination and delivery mode of the interrupt. The information in the
RT is used to generate the inter-(x)APIC message upon assertion/deassertion of the interrupt pin.
The PID’s version register contains the number of entries in the interrupt RT. Offsets 10h through
8Fh are used for the interrupt RT. Each entry in the table is 64-bits. The format of each entry is
Note: The interrupt RT is shared between SAPIC and APIC modes. Some of the fields have different
meanings in the two modes as indicated in Columns 2 and 3. The PID implements only those bits
that have valid functional fields associated with them. Reserved bits cannot be written and will
return 0s when read.
Since each RTE is 64 bits wide, it must be accessed using two 32-bit memory read or write
operations to consecutive dword-aligned addresses. The lower half of the entry, bits 31 through 0,
is located at the even offset (such as 10h) and the upper half of the entry, bits 63 through 32, is
located at the odd offset (such as 11h). While programming the RTEs, it is recommended that the
lower half be programmed first, followed by the upper half.
Table 2-10. I/O (x)APIC RTE Format
Register Offset: 10-8Fh
Default Value: Undefined except mask bit is 1Attribute: Read/Write
SAPIC Mode
APIC Mode
Description
Name
Bit(s)
Name
63:56 DEST ID
DEST ID
This field contains an (x)APIC ID. Local (x)APIC units receiving an interrupt
message compare the DEST ID and DEST EID fields to the corresponding
fields in their LID registers to determine if the interrupt is to be serviced by
them.
55:48 DEST EID
Reserved
This field contains the extended (x)APIC ID. Local (x)APIC units receiving an
interrupt message will use this field along with the DEST ID field to determine if
the interrupt is to be accepted by them. This field is not used during APIC
mode.
47:32 Reserved
31:18 Reserved
Reserved
Reserved
FLUSHEN
These 16 bits are reserved
These 14 bits are reserved
17
FLUSHEN
This bit controls the flushing of the I/O buffer on a per-interrupt basis.
A 0 indicates that the buffer must be flushed before the interrupt is sent to the
local (x)APIC. This setting will cause the hardware flush control signals to be
used.
A 1 indicates that the buffer does not need to be flushed before the interrupt is
sent out to the local (x)APIC. This setting will cause the hardware flush control
signals to be ignored.
2-50
Intel® 460GX Chipset Software Developer’s Manual
Register Descriptions
Table 2-10. I/O (x)APIC RTE Format (Cont’d)
Register Offset: 10-8Fh
Default Value: Undefined except mask bit is 1Attribute: Read/Write
SAPIC Mode
APIC Mode
Name
Bit(s)
Name
Description
16
MASK
MASK
This bit masks the (x)APIC delivery of this interrupt.
A 0 indicates that delivery of this interrupt is not masked. An edge or level on
an interrupt pin that is not masked results in the delivery of the interrupt to the
destination.
A 1 indicates that delivery of this interrupt is masked. It is the software’s
responsibility to deal with the case where the mask bit is set after the interrupt
message has been accepted by a local (x)APIC unit but before the interrupt is
dispensed to the processor.
When the mask bit is 1, interrupts coming into the PID are routed to the INTIO,
I2O_INT#, or I2B pins, provided they are correctly mapped. No delivery status
bit latching is performed in this mode.
When the mask bit is a 0, interrupts are latched in the delivery status bit. If the
mask bit is set to a 1 after the interrupt is latched but before it is delivered, the
latched value will be held until the interrupt is unmasked and delivered.
15
14
TRIGGER
MODE
TRIGGER
MODE
The trigger mode field indicates the type of signal on the interrupt pin that
triggers an interrupt.
A 0 indicates edge sensitive.
A 1 indicates level sensitive.
RIRR
RIRR
This bit is read-only. During (x)APIC mode, this bit is used for level-triggered
interrupts. Its meaning is undefined for edge-triggered interrupts. For level-
triggered interrupts, this bit is set when the local (x)APIC(s) accepts the level
interrupt sent by the PID. The RIRR bit is reset when an EOI message is
received from the local (x)APIC.
13
12
POLARITY
POLARITY
This bit specifies the polarity of each interrupt signal connected to the interrupt
pins of the PID. A value of 0 means the signal is high-active and a value of 1
means the signal is low-active. In the case of level, high-active means if the pin
is sampled high, it is considered active. In the case of an edge, high-active
means the low-to-high edge is considered active.
DELIVERY
STATUS
DELIVERY
STATUS
This bit is read-only. It holds the current status of interrupt delivery to the
processor. This bit functions differently depending on the Trigger Mode bit.
When the Trigger Mode bit indicates edge sensitive:
The Delivery Status bit is set when an active edge is detected on the interrupt
pin.
The Delivery Status bit is reset when the interrupt message is successfully sent
to the processor.
When the Trigger Mode bit indicates level sensitive:
The Delivery Status bit reflects the assertion of the pin. i.e. If the pin is at its
active level, according to its Polarity bit, the Delivery Status bit is set; If the pin
is at its inactive level, according to its Polarity bit, the Delivery status bit is
reset.
To accurately determine the delivery status of an interrupt in level mode the
RIRR bit must be considered as well as the Delivery Status bit.
Intel® 460GX Chipset Software Developer’s Manual
2-51
Register Descriptions
Table 2-10. I/O (x)APIC RTE Format (Cont’d)
Register Offset: 10-8Fh
Default Value: Undefined except mask bit is 1Attribute: Read/Write
SAPIC Mode
APIC Mode
Name
Bit(s)
Name
Description
11
DESTINATION DESTINATION This bit determines the interpretation of the destination field.
MODE
MODE
A 0 indicates physical mode. In physical APIC mode, a destination APIC is
identified by its ID. Bits 56 through 59 of the destination field specify the 4-bit
APIC ID. In physical SAPIC mode, the destination ID is defined by bit 63
through 48 of the destination field.
A 1 indicates logical mode. In logical APIC mode, destinations are identified by
matching on logical destination under the control of the destination format
register and logical destination register in each local APIC. Bits 56 through 63
(8 MSB) of the destination field specify the 8-bit APIC ID. In logical SAPIC
mode, the destination field bits 63 through 56 make up 8 bits of the destination
ID. The remaining 8 bits are bits 55 through 48 of the destination field and must
always be programmed to 0.
10:8 DELIVERY
MODE
DELIVERY
MODE
The delivery mode is a 3-bit field that specifies how the local (x)APIC units
listed in the destination field should act upon reception of this signal. Note that
certain delivery modes will only operate as intended when used in conjunction
with a specific trigger mode. These restrictions are indicated in the table below
for each delivery mode. Delivery mode encodings are shown below. Modes
001 and 010&Vector Field apply to APIC mode only:
00x (Fixed SAPIC Mode): This means deliver the interrupt on the INTR signal
of the processor listed in the destination. Trigger mode can be edge or level.
The least significant bit of this delivery mode is a HINT bit that is
communicated to the platform to allow it to redirect the interrupt to another
processor on the same processor system bus as the destination. The way this
redirection occurs is independent of platform implementation. Otherwise, the
processor can ignore the least significant bit.
000 (Fixed APIC Mode): This means deliver the signal on the INTR signal of
all processor cores listed in the destination field. Trigger mode for “fixed”
delivery mode can be edge or level.
001 (Lowest-Priority APIC Mode): This means deliver the interrupt on the
INTR signal of the processor core that is executing at the lowest-priority among
all the processors listed in the specified destination. Trigger mode for lowest-
priority delivery mode can be edge or level.
010 (PMI or SMI): Delivery mode is edge only. For systems that rely on SMI
semantics, the vector is ignored but must be all zeros for future upgrades. For
systems that rely on PMI semantics, the vector number has meaning and is not
ignored.
011 (Reserved).
100 (NMI): This means deliver the interrupt on the NMI signal of the processor
listed in the destination. Vector information is ignored. The NMI is always
delivered as an edge-triggered interrupt. The trigger mode field is ignored.
101 (INIT): This means deliver the interrupt on the INIT signal of the processor
listed in the destination. Vector information is ignored.
110 (Reserved).
111 (ExtINT): This means deliver the interrupt to the processor listed in the
destination as an interrupt that originated in an externally connected (8259A-
compatible) interrupt controller. The INTA cycle that corresponds to this ExtINT
delivery will be routed to the external controller that is expected to supply the
vector. A delivery mode of ExtINT requires an edge-triggered mode. ExtINT
should be targeted for only one processor.
7:0
VECTOR
VECTOR
This is the vector number identifying the interrupt being sent.
2-52
Intel® 460GX Chipset Software Developer’s Manual
System Architecture
3
This chapter provides an explanation of the 460GX chipset’s handling of various aspects of the
system architecture. It covers coherency, ordering, interrupts and related issues.
3.1
Coherency
For any computer system, data coherency between processor caches and other elements within the
system is one of the main concerns of the system designer. In a Symmetric Multiprocessor (SMP)
system, hardware is usually required to maintain this coherency. In some instances, though,
software may assume responsibility for coherency. The use of WC (write-combining) memory
requires software to manage coherency. Coherency applies to the processor caches, the I/O sub-
system and graphics traffic. The programming model for a particular architecture determines
precisely how the hardware must behave. A loosely coupled system may require that software
guarantee that different processors which need to use the same piece of data are able to do so. For
most SMP systems, the hardware in the system guarantees that a piece of data is updated to all
processors if it is shared. For coherent data or code, there can exist only one valid version. There
may be shared copies of this data or code, but all elements in the system have the same value. If the
data is not coherent, then different elements may have different values for the identical address. For
example, non-coherent AGP traffic is not kept coherent with processor caches, since non-coherent
AGP addresses are not snooped on the bus. Coherency is related to, but different than cacheability.
3.1.1
Processor Coherency
Intel processors do not have a specific bit to specify coherency for each transaction. Data and code
are usually considered fully coherent with respect to other processors and to each other. There are
exceptions to this - such as the WC (write combining) memory type. Data that is marked as WC in
the page table will not be coherent between processors. The Itanium processor uses the standard
MESI (Modified/Exclusive/Shared/Invalid) protocol for its caches. As well, the Itanium processor,
when in EM mode, does not guarantee in hardware that the instruction caches are coherent with
data transfers. Software must flush the i-cache and reload it when self-modifying code is running.
Processors maintain coherency with each other by placing the addresses for referenced memory on
the common system bus. Each of the other processors snoops this address to see if it has the data in
any of its own caches. If one has the data unmodified, then it signals this using the HIT# pin, and
memory, since it has the latest version of that address, provides the data. Both the responding and
the requesting processors will then go to the shared state in the cache. If the snooping processor has
the data modified, it asserts the HITM pin and provides the data, since it owns the most up-to-date
copy of that address. The memory system will also take the data provided and update the copy in
SDRAM, unless the OWN# signal is active, in which case the memory is not updated.
For WB memory space the MESI protocol requires that only one processor has a copy of modified
data. Therefore if processor A is writing the data, it was required to gain ownership of the line
before updating it, and as it gained ownership the other processors in the system would have gone
to the invalid state for that line.
Intel® 460GX Chipset Software Developer’s Manual
3-1
System Architecture
For WC memory, one processor may write to an address that is marked WC in its page table and
hold the write in its own data buffer, while waiting to write to the bus. If a different processor were
to read this address, the first processor does not snoop the WC holding registers and therefore
would not provide the newly written data. So the 2nd processor would get the data from memory
which is stale or un-updated with respect to the first processor.
Processor caches use physical addresses for indexing into the cache. The memory attributes are
marked in the page tables which are based on the virtual address. If the O/S points two virtual
addresses to the same physical page and provides different attributes for each, then coherency may
not be maintained between processors, since one may have used the virtual address that pointed to
the page as WB and the other one used a different virtual address that pointed to the same physical
page, but was marked WC. This is usually considered a programming error. It may be useful in
some applications, but software takes full responsibility for the results.
3.1.2
PCI Coherency
As well as processor to processor coherency, the I/O subsystem maintains coherency between
processor caches and transactions initiated on PCI or other I/O devices that are directed to memory.
This is done in hardware and therefore software is not required to flush caches before I/O
operations (as long as that page is not marked WC).
Reads originating from the I/O sub-system are presented to the system bus for snooping. If a
processor has the modified (dirty) data, then it provides it on the data bus and the SAC presents this
data back to I/O. If no cache has the data modified, then the data is provided by the SDRAMs.
Writes are presented to the system bus as well. This allows a new code or data page to be brought
in and the old page to be invalidated as the write is being done for each line. Writes may get a
HITM# snoop which causes the processor to write back the entire line and then having the I/O
write to overwrite the particular bytes it wishes to update.
3.1.3
AGP Coherency
AGP transactions originate from a graphics (or other) device residing in the one AGP slot provided
when using the GXB. There are 2 different types of transactions originating from a device in this
slot. The graphics card may do coherent or non-coherent transactions.
Non-coherent transactions are not required to be placed on the system bus (although they could be,
with some loss of bus bandwidth). The 460GX chipset implementation does not pass non-coherent
AGP traffic to the system bus. These addresses are sent directly to the memory queue. The
processor could still have these addresses cached. If the application wishes to do this, then it must
handle coherency itself. Software must flush the processor’s cache.
Coherent traffic from the graphics card is treated like coherent PCI traffic. Addresses are sent to the
system bus before being serviced by the memory system.
3.2
Ordering
Intel processors prior to the Itanium processor have used processor consistency for their ordering
model. The Itanium processor allows both a strongly ordered and a weakly ordered programming
model to be implemented.
3-2
Intel® 460GX Chipset Software Developer’s Manual
System Architecture
New EM code may be weakly ordered. To allow the processor to take advantage of this, the 460GX
chipset defers all reads and returns the data out-of-order to the processor. By returning data in an
out-of-order fashion, the DRAM’s may be accessed in an optimal manner. Accesses are sent out of
the memory queue to free banks of SDRAM’s. Thus, if consecutive addresses are to the same bank,
instead of holding up all later accesses while doing the first 2 in order, later accesses may move
around the second access and allow data to move continuously from SDRAM to the system bus.
To maintain ordering in the system, the processor issues an address and must wait until that address
has been accepted by the system, or in other words become globally visible. If the operation is kept
in the in-order queue, then visibility occurs at the snoop phase with no defer. This means that
operation is not retried and visibility has been met. A read becomes visible when no later store can
change the value seen by the reading processor. A write becomes visible when all later reads will
see the result of that write. With this definition, the 460GX chipset is able to guarantee visibility is
met when the access is deferred, since it prevents any later access from affecting that read. Any
coherent write to that line would be retried until the read is complete. Writes to memory are posted
and so are immediately visible and complete from the system perspective.
There is no ordering relationship between the PCI command streams of an AGP card and its AGP
command streams. The AGP spec mandates certain ordering rules within each stream that are
visible by the graphics card, but the order in which the system does the transactions is not
specified. Therefore, the typical producer-consumer model can not be guaranteed by doing simple
reads and writes across command streams. The AGP card must first issue a “flush” command in
order to guarantee the AGP low-priority stream is observable in memory before sending a flag
(which indicates all the writes are visible to the processor) up the PCI stream.
3.3
Processor to PCI Traffic and PCI to PCI (Peer-to-
Peer) Traffic
Due to the limited number of resources for transactions directed to PCI, a transaction may be
retried if all the resources are utilized. It is possible for one processor or one PCI agent to keep
taking all the resources and preventing a different processor or PCI agent from making any forward
progress.
3.4
WXB Arbitration
The following topics highlight a few of the methods employed within the WXB for starvation
prevention.
3.4.0.1
WXB Arbitration at the PCI Bus
Arbitration for Inbound Transactions
The WXB implements a simple two-level PCI arbitration scheme in the same vein as that of the
PXB-C0. Access to inbound resources is subject to the PCI arbitration scheme and to inbound
resource management algorithms. Inbound Write Request (IWR) acceptance is subject to an IWR
starvation prevention mechanism while Inbound Read Request acceptance is subject to the
availability of either an invalid stream slot or to space in the Delayed Transaction Reservation
Buffer.
Intel® 460GX Chipset Software Developer’s Manual
3-3
System Architecture
Arbitration for Outbound Transactions
The WXB relies heavily on the PCIset core and the PCI Specification regarding transaction
ordering for dealing with starvation on outbound transactions. Once the WXB has won PCI
arbitration for an outbound transaction, the WXB will initiate the request at the top of the
Outbound Transaction Queue (OTQ) unless there is a read within the Outbound Read Request
FIFO (ORRF). In this case, if the read at the top of the ORRF has already received some data, and
the transaction at the top of the OTQ is a read, it will be moved behind the read in the ORRF. The
read at the head of the ORRF will then be initiated on the PCI bus. The read at the head of the
ORRF will be moved to the back if it hasn’t already received some data. If, however, the
transaction at the top of the OTQ is a write, then the arbitration policy will be to choose, on a
round-robin basis, between the read at the head of the ORRF and the write. The write will attempt
to burst to the end of a cache line; if there are no reads in the ORRF at the end of the cache line
burst, the write will continue to burst to the end of the next cache line unless the MLT causes the
WXB to disconnect.
The IHPC will also participate in arbitration with the purpose of idling the PCI bus. When the
IHPC has won grant, and FRAME# has deasserted, there will be moments of inactivity as the IHPC
alters the state of various external signal and power control registers for one or more PCI slots.
3.5
Big-endian Support
The Itanium processor supports both little-endian and big-endian accesses. The chipset does not
need to know which mode the processor is in. The chipset provides data in the same manner in both
cases. There is no indication on the system bus which mode the processor is using.
3.6
Indivisible Operations
3.6.1
Processor Locks
The 460GX chipset supports locks on the system bus, done by the processor. These locked
transactions are either a series of atomic Read-Write or Read-Read-Write-Write transactions. They
may be targeted to I/O devices or to SDRAM. See ‘Transactions’ Chapter for the flow for locks.
During the sequence, the system bus is locked and no other traffic can occur on that bus. Traffic
may be flowing throughout the rest of the system, such as AGP to memory or transactions that stay
within the non-locked PCI buses.
The 460GX chipset does not support locks that cross device boundaries. In other words, if the first
read in a locked sequence targets device X, then the remaining transactions in the lock (either R-W-
W or W) must also target device X. The only exception to this rule is when device firmware has
been “in-line shadowed” using the MAR registers. In this case the Read(s) in a locked sequence
can be mapped to the compatibility PCI bus, and the Writes(s) could be mapped to memory. When
a MAR has been mapped to write protect memory, a locked sequence to that MAR region is
completely redirected to PCI, in order to avoid the resource allocation problems associated with
crossing memory/PCI device boundaries. The 460GX chipset does not provide any special checks
to detect locks that cross device boundaries outside of the MARs. If software attempts to establish
such a lock, indeterminate results will occur: either the lock will appear to work, even thought the
access was not performed atomically, or a deadlock will result.
3-4
Intel® 460GX Chipset Software Developer’s Manual
System Architecture
AGP LOCKS
There is no LOCK signal on the AGP bus. However, legacy code that issues read-modify-write
(RMW) transactions could still be converted for use with an AGP device. The GXB will attempt to
establish a “pseudo-lock” to cover such an event. However, there is still a deadlock case within the
AGP controller that the GXB can not address. This case is covered in a “special design
considerations” section of the AGP specification.
The deadlock occurs when the device internally posts an inbound write after the first read
completes in a “R-R-W-W” locked sequence. The specification requires that the AGP master
resolve the problem in either software or hardware:
• Software must prevent the device driver from accessing internal registers with misaligned
reads while there are posted writes in the PCI interface. This works if the device never posts
writes, implements a unified interface (i.e.: no “internally posted” writes), or disallows
misaligned read access (no multi-word registers).
• Hardware allows the read to proceed even in the presence of the posted write. Technically this
is a violation of protocol, but the master is at liberty to insure that internal status doesn’t get
updated on behalf of the “posted” write until that data actually leaves the part.
3.6.2
3.6.3
Inbound PCI Locks
The 460GX chipset does not support inbound locks.
Atomic Writes
Some system bus operations such as Write 8 bytes, Write 16 bytes and Write 32 bytes, are
indivisible operations on the system bus. However, since the PCI protocol allows target device to
disconnect at any point in a transfer sequence, these operations are not indivisible on the PCI bus.
Furthermore, these accesses cannot be locked because PCI specification allows use of locked
cycles only if the first transaction of the locked operation is a read. Therefore software must not
rely upon atomicity of system bus write transactions which are greater than 32 bits or Dword
misaligned once they are translated to the PCI bus.
3.6.4
3.6.5
Atomic Reads
The system bus memory read operations to PCI can request more than 32-bits of data (i.e. 8 byte,
16 byte and 32 byte). The problem of indivisibility of operations is very critical for this type of
transaction. The PXB does NOT Lock multi-cycle reads to guarantee atomicity. Note that
ATOMICITY of Host-PCI reads which are greater than 32 bits or are Dword misaligned is NOT
GUARANTEED.
The PXB can accept inbound reads and writes while an outbound read or write is in a partially
completed state.
Locks with AGP Non-coherent Traffic
AGP has a non-coherent stream that does not go over the system bus. Therefore processor locks,
whether established to memory or I/O, do not prevent the non-coherent AGP accesses from
occurring. If a processor reads a location in memory with a locked read, and then writes the same
Intel® 460GX Chipset Software Developer’s Manual
3-5
System Architecture
location, there is no guarantee that AGP has not written the location while the lock was active on
the system bus. AGP may read or write those locations or any other memory location, independent
of the processor lock.
3.7
Interrupt Delivery
Interrupts may be delivered to the processors over the system bus. The interrupts may come from
an I/O device or from another processor. The new system bus delivery method for interrupts to
Itanium processor is referred to as SAPIC.
For SAPIC, the interrupts appear on the bus with a specific encoding on the REQa/REQb signals.
The address used is 0x000FEEzzzzy. The 16 bits ZZZZ determine the target to which the interrupt
is being sent. Both I/O interrupts and inter-processor interrupts are delivered in this manner. The
data field contains the interrupt vector.
Devices may send their interrupts to one specific processor, or have the system choose the
processor to which to deliver the interrupt. The algorithm used by the system to deliver interrupts
depends on the software. The chipset will have a register for each processor. This register, called
the XTPR (eXternal Task Priority Register), is programmable and may be set as software wishes.
One could use this feature for lowest-priority delivery, as one example. The processor updates its
XTPR when it updates its own internal priority register. An interrupt which is received by the
chipset with the redirectable hint bit set, will be sent to the processor with the lowest value in the
XTPR. If 2 or more processors tie for the lowest value, the processor with the lowest processor ID
will be selected.
The 4 XTPR registers in the 460GX chipset are updated when the processor does a special cycle on
the bus. When the special cycle is decoded, the low order 3 bits of the DID are used to determine
which register to update. Each XTPR register is disabled at reset, and requires a special cycle
XPTR-update to be enabled.
3.8
WXB PCI Hot-Plug Support
“Hot-Plug” is the term given to describe the capability to insert and remove PCI add-in cards into a
computer while the PCI bus itself and other subsystems in the computer are fully operational. Hot-
Plug logic in the WXB supports system hot-plug capability by providing the state machines and
programming interface to individually power up and power down PCI slots in a controlled fashion.
An Integrated Hot-Plug Controller (IHPC) comprises the hot-plug logic for one PCI bus.
Hot-Plug functionality in a complete system which includes the WXB requires three subsystems to
work together: (1) the software subsystem, including firmware and drivers, (2) the hardware logic
subsystem in the WXB, including hot-plug registers and memory and state machines, and (3) the
external hardware subsystem, including interlock switches, shift registers, FETs, and LEDs on each
PCI bus.
The PCI hot-plug logic performs three sets of operations: reset, slot enable (power-up and connect
to bus), and slot disable (disconnect from bus and power-down). In normal operation, slot enable
and slot disable occur only under software direction. Slot disable may occur automatically if a PCI
card generates a power fault or if the user opens an interlock switch to remove a powered-up PCI
card. The slot enable and disable sequences are briefly described below.
3-6
Intel® 460GX Chipset Software Developer’s Manual
System Architecture
3.8.1
Slot Power-up and Enable
To power-up a PCI slot, software sets a command bit in a register. Then the hot-plug logic performs
the following steps:
1. Set PWREN active to the slot and clock the parallel latch.
2. Set CLKEN# active to the slot but do not clock the parallel latch.
3. Wait 200 msec for slot power to stabilize, except in test mode.
4. Gain ownership of the PCI bus through arbitration.
5. Clock the parallel latch.
6. Release ownership of the bus after 480 nsec.
7. Set RESET# inactive to the slot but do not clock the parallel latch.
8. Wait 200 msec for slot clock to stabilize, except in test mode.
9. Gain ownership of the PCI bus through arbitration.
10. Clock the parallel latch.
11. Release ownership of the bus after 480 nsec.
12. Set BUSEN# active to the slot but do not clock the parallel latch.
13. Wait 1000 msec for PCI card initialization, except in test mode.
14. Gain ownership of the PCI bus through arbitration.
15. Clock the parallel latch.
16. Release ownership of the bus after 480 nsec.
3.8.2
Slot Power-down and Disable
To power-down a PCI slot, software sets a command bit in a register. Then the hot-plug logic
performs the following steps:
1. Set BUSEN# inactive to the slot but do not clock the parallel latch.
2. Gain ownership of the PCI bus through arbitration.
3. Clock the parallel latch.
4. Release ownership of the bus after 480 nsec.
5. Set RESET# active to the slot but do not clock the parallel latch.
6. Gain ownership of the PCI bus through arbitration.
7. Clock the parallel latch.
8. Release ownership of the bus after 480 nsec.
9. Set CLKEN# inactive to the slot but do not clock the parallel latch.
10. Gain ownership of the PCI bus through arbitration.
11. Clock the parallel latch.
12. Release ownership of the bus after 480 nsec.
13. Set PWREN inactive to the slot and clock the parallel latch.
Intel® 460GX Chipset Software Developer’s Manual
3-7
System Architecture
3-8
Intel® 460GX Chipset Software Developer’s Manual
System Address Map
4
4.1
Memory Map
The Itanium™ processor supports a 44 bit address space. The 460GX chipset supports only 36 bits
of the address bus for a 64 GB of physical memory and must address up to several GB of memory
mapped I/O space. The 460GX chipset attaches to A#[35:3].
The memory address space is divided into four regions: the 1 MB Compatibility Area, the 15 MB
Low Extended Memory region, the (4 GB minus 16 MB) Medium Extended Memory region, and
the (16 TB minus 4 GB) High Extended Memory region. The first three regions are divided into
multiple subregions, with dedicated purpose and semantics. The memory address map is illustrated
4.1.1
Compatibility Region
This is the range from 0-1 MB (0_0000 to F_FFFF). Addresses here may be directed to a PCI bus
(the compatibility bus) or main memory. Any DRAM located in this area that is not used as main
memory is not recovered. This region is divided into four subregions, some of which are further
subdivided. Regions below 1M that are mapped to memory are accessible by the processors and by
any PCI bus.
Regions below 1M that are mapped to PCI are accessible by the processor, and by PCI devices on
the targeted PCI bus (in the case of the regions mapped by the MAR this means the compatibility
PCI bus; for the VGA region this means the PCI bus to which VGA is mapped). Parallel segment
peer-to-peer accesses are not supported below 1M; a PXB will either forward the access to memory
or let it be claimed/master abort on the PCI bus below it.
4.1.1.1
4.1.1.2
4.1.1.3
DOS Region
The DOS Region is the lowest 640 KB, in the address range 0h to 9_FFFFh. DOS applications
execute here. The lower 512K of this region is always mapped to main memory, and is accessible
by the processors and by any PCI bus. The upper 128K of this region can be mapped to either PCI
or main memory, and is mapped using one of the MAR registers. The range defaults to memory.
VGA Memory
The 128 KB Video Graphics Adapter Memory subregion (A_0000h to B_FFFFh) is normally
mapped to a video device on the compatibility PCI bus. Typically, this is a VGA controller. The
460GX chipset supports mapping this region to any of its logical PCI segments. At power-on this
space is mapped to the compatibility PCI bus.
C, D, and E Segments
The 192 KB C, D, and E Segments are divided into smaller blocks that can be independently
programmed as Mapped to Memory, Memory Write Protect, In-line Shadowed, or Mapped to PCI.
These regions are used to provide memory to PCI devices requiring memory space below 1 MB.
The MAR registers determine how each range is used. The default for these segments at power-on
is that they are mapped read/write to the PCI compatibility bus.
Intel® 460GX Chipset Software Developer’s Manual
4-1
System Address Map
Figure 4-1. System Memory Address Space
FFF_FFFF_FFFF
FFFF_FFFF
System
Memory
16 MB
High
System
Firmware
High
Extended
FF00_0000
FEC0_0000
(16 TB-4GB)
Processor
Specific
1_0000_0000
Memory
4 MB
Firmware,
Processor &
Chipset
Chipset
Specific
32
MB
12 MB
Specific
FE00_0000
n x 32M
PCI Gaps
(4 GB-16MB)
Medium
Extended
memory
FE00_0000 -
(n x 32M)
System
Memory
Low
Extended
15 MB
1 MB
Memory
10_0000
F_0000
Compatibility
Area
System
Firmware
64K
C, D, and E
Segments
192K
Areas are not
drawn to scale.
C_0000
A_0000
128K
VGA
Memory
512K-640K
DOS
Region
640K
0
4.1.1.4
System Firmware
The 64 KB region from F_0000h to F_FFFFh is treated as a single block. Read/Write attribute
enables defined in the MAR registers may be used to direct accesses to the compatibility PCI bus
or main memory. At power-on, this area is mapped by default to the PCI compatibility bus.
4-2
Intel® 460GX Chipset Software Developer’s Manual
System Address Map
4.1.2
4.1.3
Low Extended Memory Region
The 15 MB Low Extended Memory region is always mapped to main memory. Since the 460GX
chipset does not support ISA cards, there is no gap provided in this region.
Medium Extended Memory Region
The Medium Extended Memory region is divided into two primary regions: A fixed gap containing
the firmware area along with gaps for Itanium processor and chipset specific functions, and a
variable gap to support memory mapped I/O.
The fixed gap is between 4 GB and (4 GB minus 32 MB) and is always enabled. This region must
not be defined as WB. DRAM supported by the 460GX chipset that is masked by this hole is
remapped to an area over 4 GB. The fixed gap is further divided into three regions:
• The 4 GB to (4 GB minus 16 MB) region is reserved for system firmware. Addresses directed
to this area are always directed to the compatibility PCI bus.
• The (4 GB minus 16 MB) to (4 GB minus 20 MB) region is reserved for processor specific
functions. This region can be thought of as four 1 MB segments:
— FEF0_0000 - FEFF_FFFF: This segment is used for Local APIC messages. Neither
inbound nor outbound accesses to this region should be seen by the chipset. The chipset
claims outbound accesses to this region, and will forward reads to PCI 0a to be master
aborted and drops the writes. Inbound accesses to this region can only occur due to a
programming error or address parity error; firmware programs the Expander bridges to
prevent programming errors from reaching the SAC; address parity errors are protected
against at each of the major bus interfaces. Therefore the SAC does not require a defined
response for inbound requests that reach this area; the results are unpredictable.
— FEE0_0000 - FEEF_FFFF: This segment is used to deliver interrupts. The chipset claims
outbound accesses to this region, and will forward reads to PCI 0a to be master aborted
and drops the writes. Inbound writes to this region are translated to an interrupt command
encoding, forwarded to the system bus, and then claimed by the chipset. Reads to this area
are illegal.
— FED0_0000 - FEDF_FFFF: This segment is reserved. The chipset claims outbound
accesses to this region, and will forward reads to PCI 0a to be master aborted and drops
the writes. Inbound accesses to this region are illegal.
— FEC0_0000 - FECF_FFFF: This segment is used for SAPIC messages. The chipset claims
outbound accesses to this region and forward them to the appropriate PCI bus. The
processors use this region to program the SAPIC or IOAPIC registers and for targeted
EOI writes. Inbound accesses to this region are illegal.
• The (4G minus 20M) to (4G minus 32M) region is reserved for chipset specific functions.
Inbound accesses to this region are expected only from the compatibility PCI bus from either a
server management or a validation card. IB accesses are directed to the system bus and then
forwarded to the appropriate PCI segment. This region is segmented as follows (unlisted
regions are reserved):
— FE20_0000 - FE3F_FFFF: This range is for programming the GART. Reads and writes
are sent to Expander port-2 to be forwarded to the GXB. If the DEVNPRES bit for Device
14 is set (meaning that there is no xXB attached to the Expander bus), then accesses to this
region will be forwarded to the compatibility bus for termination.
— FEB0_0CB0: This address is for a memory-mapped register in the SAC.
Intel® 460GX Chipset Software Developer’s Manual
4-3
System Address Map
— FEB0_0CC0: This address is used for BSP selection. It is a write once register in the
SAC.
Figure 4-1 shows how the SAPIC and GART spaces are allocated. There may be up to 255 I/O
SAPICs in the system. There is one region defined for the GART space.
4.1.3.1
Variable GAP
The variable gap starts at 4G-32M, and can grow downward in 32M increments. This gap is used to
provide memory mapped I/O spaces to all of the logical PCI buses (this includes AGP buses) in the
system. Each logical PCI bus is allowed n x 32M of contiguous space. PCI bus #0 is allowed the
first a*32M below (4G-32M), PCI bus #1 is allowed the next b*32M, PCI bus #2 is allowed the
next c*32M, etc. The total gap size, n, is equal to a + b + c and so on. The combined size of the
variable and the fixed gaps must equal a multiple of 64 MB. Since the variable gap starts on a
32 MB boundary, the variable gap must total to an odd multiple of 32 MB. This limit is set up by
firmware and is a function of the memory controller design.
4.1.4
High Extended Memory (above 4G)
The entire address space above 4 GB is treated by the 460GX chipset as ordinary memory. The top
of system memory is calculated by firmware. Processor accesses above the top of system memory
are still claimed by the chipset, but are not forwarded to memory or PCI; instead they cause a
BINIT#. Inbound accesses to this region can only occur due to a programming or address parity
error. Firmware programs both the PXBs and GXBs with the Top of Memory value. A
programming error that results in a PXB access above the Top of Memory causes the PXB to route
the request as if it were to a peer PCI bus. Therefore the request goes through the SAC decoder and
causes a BINIT#. A programming error that results in a GXB access above the Top of Memory
(detected after GART translation) causes the GXB to force a BINIT#. Address parity errors are
detected at each of the major bus interfaces.
4.1.5
Re-mapped Memory Areas
Any DRAM that lies behind an address that is mapped to PCI or is reserved in the region below
4 GB and above 1 MB is recovered. The memory that lies in the Medium Extended Memory
Region is moved so that it is addressed above 4 GB. This covers all addresses in the Medium
Extended Memory Region. For example, if there is 3 GB of memory and the Medium Extended
Memory Region is 2 GB total (covering PCI, AGP and the reserved area), then the first 2 GB of
addresses are directed to DRAM, the next 2 GB of addresses are directed to PCI or are AGP
addresses or reserved, and the address range between 4 GB and 5 GB would be directed to DRAM.
In other words, to access the 2 GB+1 byte of memory, the processor would use address 4 GB+1,
since the physical address of 2 GB+1 is now mapped to PCI.
Any DRAM not used in the Low Extended Memory Region, i.e. the region below 1 MB is not
recovered. The DRAM in this region which has its address directed to PCI is simply lost to the
system.
4-4
Intel® 460GX Chipset Software Developer’s Manual
System Address Map
Figure 4-2. Itanium™ Processor and Chipset-specific Memory Space
FEFF_FFFF
I/O reserved
1 MB
FEF0_0000
FEE0_0000
FED0_0000
FECF_F000
Interrupt Delivery
On system bus - 1MB
PCI Bus mapping
of SAPIC addresses.
I/O reserved
1 MB
I/O SAPIC #255
FFFF_FFFF
I/O SAPIC
#3 to # 254
4KB each
PCI Bus 2A
PCI Bus 1A
PCI Bus 0B
System
FEC0_3000
FEC0_2000
FEC0_1000
Firmware
I/O SAPIC #2
I/O SAPIC #1
16 MB
FF00_0000
FEC0_0000
PCI Bus 0A
Processor
Specific
4 MB
I/O SAPIC #0
FEC0_0000
FEBF_FFFF
Chipset
Specific
Chipset
FE60_0000
FE40_0000
12 MB
Reserved
FE00_0000
8 MB
GART Table
2 MB
FE20_0000
FE00_0000
Chipset
Reserved
2 MB
4.2
I/O Address Map
The 460GX chipset allows I/O addresses to be mapped to resources supported on the I/O buses
underneath the 460GX chipset controller. This I/O space is partitioned into sixteen 4K byte
segments. Each of the segments can be individually configured to any I/O bus. Segment 0 is always
assigned to the compatibility I/O bus (of which there is only one per system).
There are four classes of I/O addresses that are specifically decoded by the 460GX chipset:
• All I/O addresses less than 100h: These addresses are specifically decoded as “defer-only”
addresses. The SAC does not post any I/O accesses to this range, regardless of the state of the
I/O posting enable bit. This is necessary because I/O accesses below 100h have historically
had ordering side effects: e.g. accesses to the 8259 Interrupt Masks.
Intel® 460GX Chipset Software Developer’s Manual
4-5
System Address Map
• I/O addresses used for VGA controllers: 03B0h-03BBh and 03C0h-03DFh. These addresses
are specifically decoded so they can be mapped to the PCI bus specified by the VGA Space
Register. An I/O access must be contained fully within the VGA I/O range to be remapped
(e.g. an I/O read spanning 03BBh and 03BCh would not be remapped because it crosses the
VGA I/O range). Posting of this range for writes is controlled by the state of the I/O posting
enable bit in the Software-Defined Configuration Register.
• I/O addresses used for the PCI Configuration Space Enable (CSE) protocol. The I/O addresses
0CF8h and 0CFCh are specifically decoded as part of the CSE protocol. These addresses, like
the I/O addresses less than 100h, are treated as “defer only” addresses.
• Posting for all other I/O addresses is controlled by the state of the I/O posting enable bit in the
Software-Defined Configuration Register. If this bit is set, I/O writes are posted. If this bit is
not set, all I/O writes are deferred. I/O reads are always deferred.
Note, the 460GX chipset does not support ISA expansion aliasing. The IFB supports a full I/O
space decode, so the compatibility issue will be drivers that rely on the I/O aliasing behavior.
Historically, the 64k I/O space actually was 64k+3 bytes. For the extra 3 bytes, A#[16] is asserted.
The 460GX chipset decodes only A#[15:3] when the request encoding indicates an I/O cycle.
Therefore accesses with A#[16] asserted are decoded as if they were accesses to address 0 and will
be forwarded to the compatibility bus. Since they look like accesses less than 100h, they are always
deferred rather than posted. The full address is sent to the PXB and on to the compatibility PCI bus,
which therefore has PCI address bit A#[16] active.
At power-on, all I/O accesses are mapped to the compatibility bus. An I/O access is never
Figure 4-3. System I/O Address Space
1_0003
+3 bytes
(Decoded
as 0_000X)
FFFF
F000
Segment 15
4000
3000
Segment 3
Segment 2
Segment 1
2000
1000
0000
Segment 0
Compatibility
Bus Only
4-6
Intel® 460GX Chipset Software Developer’s Manual
System Address Map
4.3
Devices View of the System Memory Map
invalid accesses at the expander bridge level, since different expander bridge devices are allowed
to access different regions. For example, PXBs are allowed to access other logical PCI segments
and GXBs are not. The SAC does not perform special checking to prevent this, and therefore the
expander bridges must be set up by firmware accordingly. An exception to this rule can be made if
the request is being routed to the system bus rather than directly to memory. For instance, accesses
above the Top of Memory are not blocked at the PXB level, but instead cause a BINIT# in the SAC
because they are guaranteed to go through the SAC’s decode logic since the PXB routes these
accesses to the system bus
Note: Figure 4-1 shows that parallel segment peer-to-peer accesses are only supported when initiated by
a PXB (GXBs are only allowed to initiate accesses directed to system memory) and directed to one
of the nx32M logical PCI segments (this segment can be anywhere except below the compatibility
PCI bus). Parallel segment peer accesses are not permitted anywhere below 1 MB, not even to the
VGA region. This means that if VGA is relocated below a GXB, a server management card on one
of the PCI buses in the system can no longer access the VGA range.
Figure 4-4. System Memory Address Space as Viewed from an Expander Bridge (PXB/GXB)
PXB accesses above the Top of Memory
FFF_FFFF_FFFF
Top of Memory
are routed as if to peer and therefore
go through the SAC decode and cause
a BINIT#; GXB must cause a BINIT#
after the GART
FFFF_FFFF
System
Memory
PXB must ignore
GXB must BINIT# after GART
High
System Firmware
PXB must ignore; GXB must BINIT# after GART
PXB/GXB allow to interrupt delivery area
FF00_0000
4G-16M to 4G-17M
4G-17M to 4G-18M
4G-18M to 4G-19M
4G-19M to 4G-20M
1_0000_0000
PXB defaults to memory so SAC must trap;
GXB must BINIT# after GART
FEC0_0000
FE00_0000
Firmware,
Processor &
Chipset
32
MB
Chipset
Specific
PXB must ignore; GXB must BINIT# after GART
Specific
PXB allows (only recommended to use
from compatibility bus for server management)
GXB must BINIT# after GART
n x 32M
PCI Gaps
PXBs may support peer-to-peer; accesses
to the PXB’s own PCI bus must be ignored;
accesses to other PCI buses can be directed
to the system bus so the SAC forwards
them to the targeted peer bus
FE00_0000 -
(n x 32M)
System
Memory
GXBs can not allow peer-to-peer; accesses
to the GXB’s own PCI bus must be ignored
(allows AGP card to “talk to itself” in that)
region); accesses to any PCI bus that are found
after the GART must cause a BINIT#
10_0000
F_0000
System
Firmware
if enabled as memory GAP (directed to a logical PCI segment),
all PXB’s must ignore and all GXBs must BINIT# after GART
C, D, and E
Segments
if enabled as memory GAP (directed to the compatibility PCI bus),
all PXB’s must ignore and all GXBs must BINIT# after GART
C_0000
A_0000
if enabled as memory GAP (directed to a logical PCI segment),
all PXB’s must ignore and all GXBs must ignore in the PCI stream
but BINIT# after GART (allows AGP card to “talk to itself” in that region)
VGA
Memory
512K-640K
if enabled as a memory GAP (directed to the compatibility PCI bus),
all PXB’s must ignore and all GXB’s must BINIT# after GART
DOS
Region
The term “ignore” from a PCI perspective means do not assert DEVSEL#;
instead allow request to master abort
0
All remaining regions are mapped to main memory and are always forwarded inbound
Notes: using the “memory” route encoding
Intel® 460GX Chipset Software Developer’s Manual
4-7
System Address Map
4.4
Legal and Illegal Address Disposition
Below is the disposition of addresses done by the Bus Interface Unit (BIU).
Table 4-1. Address Disposition
Address Range
Outbound
DRAM
Inbound
DRAM
Dest. Decision
0-07FFFFh
080000h-09FFFFh
DRAM
DRAM
MAR=11 or (Read and MAR= 01 and
no system bus LOCK#) or (Write and
MAR = 10)
PCI0a
unclaimed
MAR=00 or (Read and MAR=01 and
system bus LOCK#) or (Read and
MAR=10) or (Write and MAR=01)
0A0000h-0BFFFFh
DRAM
PCIx
DRAM
VGASE=0
VGASE=1
unclaimed
DRAM
0C0000h-0EFFFFh
(divided into 12 regions of 4k
bytes)
DRAM
MAR=11 or (Read and MAR= 01 and
no LOCK#) or (Write and MAR = 10)
PCI0a
unclaimed
MAR=00 or (Read and MAR=01 and
LOCK#) or (Read and MAR=10) or
(Write and MAR=01)
0F_0000h-0F_FFFFh
DRAM
PCI0a
DRAM
MAR=11 or (Read and MAR= 01 and
no LOCK#) or (Write and MAR = 10)
unclaimed
MAR=00 or (Read and MAR=01 and
LOCK#) or (Read and MAR=10) or
(Write and MAR=01)
10_0000h - PCIS[7]
DRAM
PCIx
DRAM
PCIx
PXB uses LXGB instead of PCIS[7]
PCIS[7] - FDFF_FFFFh
PCIS register determines target PCI
bus
On PCI if
MMBASE<=address<=MMT, then not
claimed. GXB: unclaimed
FE00_0000h-FE1F_FFFFh
FE20_0000h-FE3F_FFFFh
undefined
undefined
This region is reserved.
Expander port- Expander port
If DEVNPRES[14]=0, send to
2 or PCI0A
2 or PCI0A
Expander port 2, else send to PCI-0a
FE40_0000h-FE5F_FFFFh
undefined
undefined
This region is reserved.
FE60_0000h to FEBF_FFFFh
Config unit or
PCI-0a
Config unit or
PCI-0a
If <=8B then: if one of defined
registers read or write value, if not
then return all 1’s or terminate writes.
If >8B, then send to PCI-0a. NOTE:
Locks to this range are forbidden and
will hang the system
FEC0_0000 to FECF_FFFFh
FED0_0000h to FEDF_FFFFh
PCI x
unclaimed
PXB will
SAR, IOABASE (PXB)
PCI 0a or
dropped
Reads are sent to PCI-0a for master
forward IB and abort. Writes get No-Data response
mark it to
memory.
and are dropped
FEE0_0000h to FEEF_FFFFh
Interrupt
interrupt for IB
write, PCI-0a
for read
Reads are sent to PCI-0a for master
abort. IB writes are turned to
interrupts. OB writes get No-Data
response and are dropped
Transaction;
Reads sent to
PCI 0a, writes
dropped
4-8
Intel® 460GX Chipset Software Developer’s Manual
System Address Map
Table 4-1. Address Disposition (Cont’d)
Address Range
Outbound
PCI0a
Inbound
Dest. Decision
FEF0_0000h to FEFF_FFFFh
unclaimed
Reads are sent to PCI-0a for master
abort. Writes get No-Data response
and are dropped. PXB never claims
this range
FF00_0000h to FFFF_FFFFh
1_0000_0000h to TOM
Above TOM
PCI0a
unclaimed
DRAM
na
Firmware region always enabled
main memory (if present) above 4 GB
BINIT
DRAM
na
Note: Accesses listed as “unclaimed” in the table for inbound transactions assume the PXB is
programmed correctly. If an access were received up the Expander bus that hits in the listed address
range, then its behavior is the same as outbound transactions to the same range.
Note: The PXB will never respond to an access that it is the master for. This means that an OB access will
not be claimed by the PXB, even if that access hits a range to which the PXB would normally
respond with DEVSEL#.
Note: The only ranges the PXB doesn’t claim are MMBASE to MMT, FEF0_0000h to FEFF_FFFFh, and
4G-16M to 4G. If the PCI card initiates a request to any other address, it will be sent up as TPA or
memory.
Intel® 460GX Chipset Software Developer’s Manual
4-9
System Address Map
4-10
Intel® 460GX Chipset Software Developer’s Manual
Memory Subsystem
5
The Intel 460GX chipset’s memory subsystem consists of the SAC’s DRAM controller, the SDC’s
buffering and datapath access, the MAC and MDC components, and the DRAMs themselves.
Table 5-1. General Memory Characteristics
DRAM Types
Synchronous DRAM (SDRAM)
Maximum Memory Size
Up to 4 GB using 16MB DIMM’s, up to 16 GB using 128 MB DIMM’s, or 64 GB
using 1 GB DIMM’s
Minimum Memory
Memory Increment
Memory Modules
DRAM Sizes
64 MB using 16MB DIMM’s, 256 MB using 1 GB DIMM’s
64 MB is the smallest increment
168 pin (x72) 3.3 volt DIMMs
16 Mbit, 64 Mbit, 128Mbit, 256 Mbit
DIMMs/Row
4 DIMMs per row which must be populated as a unit
Up to 4 rows per stack; may populate any number of rows
2 stacks per card; may populate either 1 or 2 stacks
Rows/Stack
Stacks/Card
Cards/System
2 cards per system; 1 card per memory port, may have either one or two cards
in the system
DRAM Types
Synchronous DRAM (SDRAM)
The SAC’s DRAM controller provides addresses and commands to the 2 MAC’s on each memory
card. The MAC generates row and column addresses for the DRAM array and controls the data
flow through the MDC. For processor-memory cycles, the address is received from the system bus,
and data flows through the SDC to/from the memory cards. For PCI-DRAM cycles, the address is
presented on the bus. The data moves between PCI and DRAM within the SAC/SDC, without
appearing on the system bus. AGP-DRAM cycles are done non-coherently (unless coherent AGP is
being used) and therefore do not have their addresses placed on the bus. The data for AGP
transactions, like PCI cycles, flows to/from the DRAM within the SAC/SDC.
5.1
Organization
The 460GX chipset supports 1 or 2 memory cards. Each card supports up to 8 GB of memory using
128 MB DIMM’s (32 GB with 1 GB DIMM’s); 2 cards provide up to 16 GB of memory (64 GB
with 1 GB DIMM’s). There are 2 independent interfaces to the SAC/SDC from the memory
subsystem, running simultaneously. Each memory interface supports 1 card and has a 72-bit
datapath and a separate control path. Running at 266 MHz, each interface allows a 2.13 GB/s peak
configuration. platform.
Intel® 460GX Chipset Software Developer’s Manual
5-1
Memory Subsystem
Figure 5-1. Maximum Memory Configuration Using Two Cards
processor data bus
processor address bus
SDC
PB[71:0]
controls
MDb[71:0]
MDa[71:0]
MAb[24:0]
MDC MDC
288 bits
MDC MDC
SAC
28bits
MDC MDC
MAa[24:0]
MDC MDC
288 bits
x 72 x 72
288 bits
row 1 of 4
row 1 of 4
x 72
x 72
x 72
x 72
x 72 x 72
x 72 x 72
to PCI via expander bridge
row 4 of 4
row 4 of 4
Stack L
Stack R
memory card A
Each card is organized as 2 stacks of up to 4 rows each. A stack consists of 1 to 4 rows of DRAM
which share a common data bus. A row consists of the 4 DIMM sockets which have a common
address/control bus. A row is the minimum atomic unit that can be accessed. Each stack has a
separate data path from the MDC to the DRAMs. Data may be transferring on both stacks
simultaneously. For instance, on the same card, stack L could be doing a read while stack R is
doing a write, or both could be doing a read.
Each row represents a set of memory devices simultaneously selected by a RAS signal and having
a common address bus. Each row generates 288 bits (256 data, 32 ECC) of data. Each row uses 4 of
the x72 (168 pin) DIMMs. There may be multiple banks per row. For instance, SDRAMs have 2, 4
or 8 banks internal to the chip. Or the DIMM may be organized such that there are 2 groups of
memory on the DIMM itself (double-sided DIMMs), with 2 chips dotted together on the data pin.
Since a row provides 256 bits (32 bytes) of data, the 64 byte line transfers of the processor will
access a row twice to read or write the entire line. Data is interleaved by the MDCs to exchange 72
bits of data per transfer with the SDCs at the rate of one cache line every 30ns (2.13 GB/s) per
interface.
There is a separate address and control bus for each memory port, with 1 card on each port. These
are independent and may be driving addresses at the same time. While one bus is driving a read, the
other could be driving a write from the write queue. This allows greater bandwidth and allows
writes to be done without interfering with read traffic.
A system may be built with only one card, and provide only half the possible bandwidth, or have
the memory on the motherboard itself. If there are 2 memory cards, then one is placed on each
interface. Average memory latency is a function of how many interleaves are available and
bandwidth. The present structure interleaves across cards first, then across stacks, and then rows.
With 2 cards, each having 2 stacks of memory, Line 0 would be on cardB/stackR, Line 1 would be
on cardA/stackR, Line 2 would be on card B/stackL, and Line 3 would be on cardA/stackL. See
5-2
Intel® 460GX Chipset Software Developer’s Manual
Memory Subsystem
Figure 5-2 for an illustration. In theory all 4 of these lines could be transferring data at the same
time. It would then be muxed by the MDC to the SDC and then by the SDC to the bus. This allows
data to be moved with no dead cycles on consecutive reads. Another possibility is that the data
from cardA is going to the system bus, while the data from cardB is going to a PCI port. Since each
memory interface can support the transfer rate of 266 MT/s, data transfers to PCI are completely
overlapping transfers to the system bus such that there are no holes in the data traffic.
Table 5-2 gives a summary of the characteristics of memory configurations supported by the
460GX chipset.
Table 5-2. Minimum/Maximum Memory Size per Configuration
Memory Size
Min Max
Double
Sided?
(Yes/no)
Number of
Chips/
DIMM
DRAM
DIMM
Size
Technology & Configuration
16 Mb
64 Mb
128Mb
256 Mb
2M x 8
2M x 72
No
9
64 MB
128 MB
128 MB
256 MB
256 MB
512 MB
512 MB
1 GB
1 GB
2M x 8
2Mx72x2
Yes
No
18
18
36
9
2 GB
4M x 4
4M x 72
2 GB
4M x 4
4M x 72 x 2
8M x 72
Yes
No
4 GB
8M x 8
4 GB
8M x 8
8M x 72 x 2
16M x 72
Yes
No
18
18
36
9
8 GB
16M x 4
16M x 4
16M x 8
16M x 8
32M x 4
32M x 4
32M x 8
32M x 8
64M x 4
64M x 4
8 GB
16M x 72 x 2
16M x 72
Yes
No
16 GB
8GB
512 MB
1 GB
16M x 72 x 2
32M x 72
Yes
No
18
18
36
9
16 GB
16 GB
32 GB
16 GB
32 GB
32 GB
64 GB
1 GB
32M x 72 x 2
32M x 72
Yes
No
2 GB
1 GB
32M x 72 x 2
64Mx72
Yes
No
18
18
36
2 GB
2 GB
64M x72 x 2
Yes
4 GB
5.1.1
DIMM Types
The only DIMMs supported are 3.3 volt 168 pin (x72) parts. These may be composed of 16Mb,
64Mb or 256Mb DRAMs. The 256Mb DRAMs must have a supply voltage of 3.3 volts.
The DIMMs used must have the serial presence detect (SPD) feature, since this is used to program
the chipset configuration registers. DIMMs not having SPD will be considered as not-present in the
system, since they are not visible to firmware.
DIMMs may have a buffer on the DIMM itself. The buffer can be used in a registered mode or a
pass-through mode. The 460GX will support both buffered and unbuffered DIMMs. It will support
the buffered DIMM in the pass-through mode, not the registered mode. Thus the timings of the
state machines in the MAC will be the same for both types of DIMM. DIMMs with 36 components
will be buffered. DIMMs with 9 or 18 components may be buffered.
Intel® 460GX Chipset Software Developer’s Manual
5-3
Memory Subsystem
5.2
Interleaving/Configurations
Maximum system bandwidth is obtainable in several ways. If the address patterns are well-behaved
then one can use the page mode of the DRAM itself to obtain high bandwidths. Generally page hits
can sustain about 5 times the bandwidth of page misses with a one-bank memory system. In
systems with only several memory banks, designs tend to try and optimize the page hit rate to
increase bandwidth.
A second approach is to have as many parallel operations within the memory system as possible.
One can spread the addresses out across multiple DRAMs and have the data transfers in parallel.
This lends itself well to designs which require a large memory system of many gigabytes.
The 460GX will implement the second approach. It will attempt to increase the amount of
address layout. It assumes that 2Mx72 DIMMs are used, so that each row is 64 MB. With all the
rows populated evenly we have 16 x 64 MB or 1 GB total memory space.
For sequential accesses, the addresses are laid out so that lines 0 and 1 can be accessed
simultaneously, and can be transferred in parallel up to the final data transfer on the system bus. As
0 and 1 are being transferred, 2 and 3 can be started to the left stacks of each card and their data
transfer will be done immediately following that of 0 and 1. The SDC buffers the data and sends it
to the system bus with no dead cycles.
SDRAMs have at least two internal banks; 64Mb chips will generally have 4, and 256 Mb chips
rows are split into 2 halves. Since there are at least 2 banks in all SDRAMs, the system will
interleave assuming all DRAMs have only 2 banks, and be split as shown. So with 16 rows, each
split in 2, there is a 32 way interleaving scheme in a totally populated system.
The first 256 MB lies in the bottom 4 rows. The next 256 MB lies above that and so on up. This
allows multiple processes, which may be spread throughout memory, to also be interleaved.
Figure 5-2. Address Interleaving
768M+7,768M+15,...
768M+5,768M+13,...
768M+1,768M+9,...
768M+6,768M+14,...
768M+2,768M+10,...
768M+4,768M+12,...
768M,768M+8,...
768M+3,768M+11,...
512M+7,512M+15,...
512M+3,512M+11,...
512M+5,512M+13,...
512M+1,512M+9,...
512M+6,512M+14,...
512M+2,512M+10,...
512M+4,512M+12,...
512M,512M+8,...
256M+7,256M+15,...
256M+3,256M+11,...
256M+5, 256M+13,...
256M+1, 256M+9,...
256M+6,256M+14,...
256M+2,256M+10,...
256M+4, 256M+12,...
256M,256M+8,...
Line 7,15,...
Line 3,11,...
Line 5,13,...
Line 1,9,...
Line 6,14,...
Line 2,10,...
Line 4,12,...
Line 0,8,...
Card B
Card A
5-4
Intel® 460GX Chipset Software Developer’s Manual
Memory Subsystem
5.2.1
Summary of Configuration Rules
The memory system may populate any row in any order. There are preferred ways of populating
the memory subsystem for performance, but all configurations will work.
The following rules summarize the way the memory system may be built up. The one hard rule is
that a given row must be populated with 4 of the same DIMMs. There is no mixing allowed within
a row. If the 4 DIMMs within a row are not the same, there is no guarantee as to system behavior,
or that the system will even work.
• For each memory row, 4 DIMMs must be populated as a unit.
• The entire row must be populated with the exact same type of DIMM i.e. the same size,
number of sides, technology (16Mb vs. 64 Mb), etc.
• Different rows may use different size DIMMs (2Mx72 in row 1 and 4Mx72 in row 2).
• Different rows may mix x4 and x8 DRAMs.
• Different rows may mix double sided and single sided DIMMs.
• Any combination of rows may be populated in any order, though performance will be affected
by how the rows are populated.
• For highest performance, the total amount of memory in each stack should be the same.
• Either one or both memory cards can be populated in the system.
• Any number from 0 to 8 memory rows on a card can be populated in the system.
5.2.2
Non-uniform Memory Configurations
DIMMs. There is no requirement that memory be populated evenly. Some stacks may have fewer
populated rows than others and the sizes within each stack may differ. Performance will be optimal
with evenly populated rows. Knowing that users may not populate the card optimally, the 460GX
will attempt to spread addresses out as best it can in an unevenly populated system.
For an easy example, use the example above, and assume that there are 4 rows of memory such that
the first row of the first 3 stacks are populated with 4Mx72 DIMMs and the last stack has only
2Mx72 DIMMs; for a total of 448 MB. The addresses are broken up such that the first 256 MB are
interleaved on a 4 way basis. The remaining 192 MB is interleaved on a 3 way basis; since all the
addresses to the last (2Mx72 row) row have been used.
This algorithm is extended for multiple sizes and arrangements of populated rows. Each row is
broken into multiple chunks and the least common denominator is found across stacks. Whatever
the system configuration, there will be some level of interleaving between cards to increase
parallelism.
5.3
Bandwidth
Sustained bandwidth is a function of traffic patterns as well as the system design and configuration.
Each memory port can transfer 16 bytes per clock. This is a peak of 2.13 GB/s per port.
Intel® 460GX Chipset Software Developer’s Manual
5-5
Memory Subsystem
5.4
Memory Subsystem Clocking
The DIMMs are clocked at half the system bus frequency. For the Itanium processor, this means
the DRAMs are clocked at 15 ns. Data is clocked out at the rate of 32B per 15 ns.
The following table lists the DRAM parameters used for the 460GX chipset.
Table 5-3. Required DRAM Parameters
Min.
(clocks)
Max.
(clocks)
Parameter
Symbol
Clock cycle time at CL=2
Access time from CLK
CAS Latency
Tck
15 ns.
Tac
6 ns.
TCL
2
4
6
2
2
1
RAS latency
RAS cycle time
Trc
RAS to CAS delay
RAS precharge
Trcd
Trp
a
Data to precharge
Tdpl
a. The sum of Tdpl and Trp are equal to Tdal as defined in the PC SDRAM Specification.
5.5
Supporting Features
5.5.1
Auto Detection
The memory controller provides capability for auto-detection of SDRAM type installed in the
system during the system configuration and initialization, providing a Plug-and-Play DRAM
interface to the user. This is done through the Serial Presence Detect logic on the DIMM. Firmware
will read the Serial Presence Detect (SPD) for each row to determine the size of the memory in that
row. Firmware will then write the size and interleaving information into the SAC and MAC
through configuration cycles. Firmware will not have to go write data and see if it is there and do
any addressing schemes to understand the system configuration. It will simply read a configuration
register and then write a different configuration register with the chipset mapping. At the same time
Firmware can calculate total system memory.
5.5.2
Removing a Bad Row
A row of memory may have a chip or DIMM fail. If an un-correctable error occurs, the system will
machine-check, usually resulting in a reset. The 460GX will report which row failed. During the
next re-boot or at power on, if the memory test fails, firmware may map the failing row as if it
didn’t exist. Since firmware goes through and reads the SPD on each row to determine its size,
firmware can just set that particular row to a size of 0, as if it weren’t there. No addresses will then
be sent to that row. The entire row is removed, even if only one side of a double-sided DIMM were
bad. But only the failing row is disabled. All other rows are still present, and the interleaving
scheme will make the maximum use of the remaining rows. For example, a system with 8 rows
populated that has one go bad will be restarted with 7 rows available.
5-6
Intel® 460GX Chipset Software Developer’s Manual
Memory Subsystem
5.5.3
5.5.4
Hardware Initialization
In order to decrease boot time of systems with large amounts of DRAM installed, hardware
initialization of memory will be supported. Since multiple rows will be initialized simultaneously,
the memory system will be able to initialize to 0 about 8 times faster than having the processor
looping through memory with writes. The MDC will force all zeroes on the data lines, with good
ECC, and the MAC will cycle through the memory addresses generating writes. The main limiter
to the number of rows being simultaneously initialized is current draw on the DRAM. One row
from each of the 4 stacks across the 2 cards will be initialized concurrently.
Memory Scrubbing
Scrubbing is the operation of walking through all installed DRAM and looking for errors. Each line
is read and then written back, whether there is an error or not. Within the SAC there is an engine to
generate addresses to be placed in the memory queue. These addresses are placed directly into the
SAC memory queue and are not snooped on the system bus, nor are they checked for address
conflicts, since the read-modify-write is treated as an atomic operation.
A scrub address is generated every 65K (65536) clocks. For a system with 32 GB of memory, this
would walk through all memory every 3.2 days. The following table shows the approximate time to
scrub memory based on memory size. Scrubbing may be disabled through a configuration bit.
Table 5-4. Scrubbing Time
Memory Size
Time to Scrub
10 minutes
64 MB
128 MB
256 MB
512 MB
1 GB
20 minutes
40 minutes
1.2 hours
3 hours
2 GB
5 hours
4 GB
10 hours
20 hours
1.6 days
3.2 days
6.4 days
8 GB
16 GB
32 GB
64 GB
Intel® 460GX Chipset Software Developer’s Manual
5-7
Memory Subsystem
5-8
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling 6
6.1
Integrity
This chapter explains the various errors in the chipset. Error handling requires catching the error,
containing it, notifying the system, and recovery or system restart. Different platforms have
different requirements for error handling. A server is most interested in containment. It wants bad
data to be stopped before it reaches the network or the disk. On the other hand workstations with
graphics may be less interested in containment. If the screen blips for one frame and the blip is
gone in the next frame, the error is transient, and may not even be noticed.
The 460GX chipset will attempt to accommodate both philosophies. It will allow certain errors to
be masked off, or will turn them into simple interrupts instead of fatal errors. Fatal errors are those
which require a re-boot, e.g. BINIT#. Some errors will always be fatal, such as protocol errors or
when the chipset has lost synchronization of queues or events. The user (OEM, O.S.) can decide
the behavior for data errors. These may be considered as fatal, for maximum containment, or they
may simply be reported as an interrupt while the system continues as best it can. If the data is
moving to graphics, then an error may be unnoticed. It is possible that data entering memory as bad
is never used, and therefore never shows up as an error to any user.
Each error will not be individually maskable. In general there are only 2 modes - aggressive and
non-aggressive. In aggressive mode, every error - parity, protocol, queue management - will be
considered fatal and lead to a BINIT#. In the non-aggressive mode, many errors will be reported as
interrupts and not cause BINIT#. Even in non-aggressive mode, when the chipset has certain errors
and doesn’t know what to do with a transaction or seems out of sync across the chips, it will
BINIT#.
The chipset will report errors at their use, instead of their generation. Both the processor and the
chipset may ‘poison’ data. If the processor has an internal cache error, it may write out the data
with bad ECC. If the chipset has bad parity on I/O data, it will corrupt the data as it is passed along.
In both cases the data will be put in memory with bad ECC. If it isn’t used, then no error is
reported. If it is used, then the error is found at that point.
The 460GX chipset will isolate the error reporting as close to the error itself as possible. In some
cases this can be to a failing DRAM or PCI card. In others it will be for a PCI bus or Expander port.
6.1.1
System Bus
• The 460GX chipset provides ECC generation on data delivered to the system bus, and ECC
checking of data accepted from the system bus. Single-bit errors are corrected; multi-bit errors
will write the data with bad ECC into the DRAM’s (poisoned data) or to I/O with bad parity.
• Parity bits are generated and checked independently for the system bus address lines, the
system bus request group, and the system bus response group. Errors typically result in the
assertion of BINIT#.
• A variety of system bus protocol errors are also detected, and will result in assertion of
BINIT#.
• The first instance of a bus error is logged with the address and error type. Additional status
flags indicate subsequent errors occurred.
• For I/O accesses, good ECC is always generated for data with no parity errors. For data with
bad parity, the data is poisoned with bad ECC as it’s returned to the processor.
Intel® 460GX Chipset Software Developer’s Manual
6-1
Data Integrity and Error Handling
6.1.2
DRAM
• The 460GX chipset provides ECC generation on all writes into the DRAM, and ECC checking
on all reads from the DRAM. Single-bit errors are corrected. Multi-bit errors will return
poisoned data. Both types of errors are logged, with the address and ECC bits for the data
being recorded. The row which failed, as well as the bit for single-bit errors, can be identified
by software.
• The first instance of a single-bit error is logged. After the first error, additional status flags
indicate subsequent errors occurred. The first multi-bit error is logged, with a status bit
indicating there were more uncorrectable errors. In both cases, software can clear the error
register and reset the error capture logic.
• Single-bit errors are corrected as they are received.
• To facilitate component debug and diagnostics, the ECC code generated on writes into the
DRAM can be forced incorrect. A configuration bit, when set, will force the ECC bits that are
written into memory to be XOR’ed with the correct value. This will allow either single or
double bit failures to be generated in memory. When the data is read, the system should correct
the data and report the error for single-bit errors, or report the error for double bit errors while
passing bad ECC to the processor.
6.1.3
Expander Buses
• Parity bits are generated and checked independently for each Expander bus. For error behavior
• Hard Fail responses are supported.
• A mechanism for elevating fatal errors to BINIT# (XBINIT#) and non-fatal errors to BERR#
(XBERR#) is provided.
6.1.4
6.1.5
PCI Buses
• Parity bits are generated and checked independently for each PCI bus.
• Standard PCI checking for aborts, PERR# and SERR# are also done.
AGP
• There is no parity on the bus between the graphics card and the GXB when using AGP
protocol. Transactions using PCI protocol have parity as defined for the standard PCI bus.
• The GXB checks for illegal or unknown operations, Expander bus parity errors, or internal
parity errors.
• There is parity on the GART table.
6.1.6
Private Bus between SAC and SDC
• There will be parity on the 64 bit data bus. Errors on data into the SDC will poison the data in
the DB. Errors on data into the SAC will always be passed on without correction, with an
option to BINIT#.
• The command bus and ITID bus are parity protected. Parity errors on this bus cause a BINIT#.
6-2
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
6.2
Memory ECC Routing
The ECC code used in DRAM is the same code as used in the Itanium processor, requiring 8 check
bits to cover 64 bits of data. On the system bus, this code detects and corrects all single-bit errors,
and detects double-bit errors.
The system designer has the option of wiring the boards such that the following is true:
• Using x4 DRAM’s, multiple errors within one chip are 100% corrected
• Using x8 chips, all errors within a single chip are 100% detected
This is done by wiring the board so that each x4 DRAM has one bit in each of the 4 ECC words of
a half-line. Since a half-line is 256 bits and the ECC is on 64 bits, there are 4 ECC words per half-
line. For x8 chips, the bits are sliced across the 4 words, so that at most 2 bits from any one chip are
in one ECC word. The ECC used on the processor will detect all 4-bit nibble errors.
6.3
Data Poisoning
When data is received that is uncorrectable, it will be passed on to the next interface as poisoned.
The data may have come from memory or from the system bus with uncorrectable ECC errors. All
data passes through the data buffer in the SDC. As uncorrectable data is placed in the data buffer it
is marked that it was received as bad. When the data is read out of the data buffer and sent on, then
the parity or ECC generated will be deliberately forced bad. Data is checked on a ‘chunk’
boundary, with a chunk being 64 bits of data.
Data to the system bus or to DRAM will have 2 bits of ECC corrupted for each failed chunk of
data. These are bits 0 and 1 of the ECC bits, or bits 63 and 71 if looking at the entire 72 bits of data/
ECC. Data passed to the private data bus will invert all the calculated parity bits associated with the
failing chunk, thus passing bad parity to the private data bus.
6.4
Usage of First-error and Next-error
The first instance of an error is latched in the first-error status register (FERR). The first error does
NOT set the bit in the next-error register (NERR). When an error is found, it is latched into the
FERR if the FERR has no other bit set. If any bit is already set, then the appropriate bit in NERR is
set.
Since the system needs to know if only one error has occurred or many, setting the FERR does not
set the NERR. If there is another error of any type, including a second occurrence of the first-error,
then the NERR is set. Software can read both FERR and NERR. If FERR is set but NERR is not,
then only one error occurred in the system. If both are set, then multiple errors have occurred.
For the first error, as much information as possible is captured. The data, address and command
information is captured if available. This allows isolation of errors and possible recovery.
In the case of 2 errors occurring in the same cycle, then 2 bits may be set in FERR. This should be
a rare case. The other exception is for FERR_SAC. If there is a single-bit correctable ECC error
from DRAM, then bit SCME will be set. This bit will not block other bits in FERR_SAC from
being set. This allows software to poll periodically looking for single bit errors while not
preventing other errors from being logged. Other than these two conditions, there should never be
more than one bit set in any FERR.
Intel® 460GX Chipset Software Developer’s Manual
6-3
Data Integrity and Error Handling
Note: In the SAC if there is a single-bit error and a double-bit error reported from the SDC on the same
cycle, then only the double-bit error is reported and only the double-bit error has its ITID captured
in the SAC. The SDC will have its SEC bit set and so software must read and clear all the errors in
the SDC after clearing the SAC.
A bit in FERR may be set that signals a minor error, such as correctable ECC error or other non-
fatal error. Another error may occur before FERR is serviced, thus forcing a bit in NERR to be set.
Since this next error may be fatal, both the FERR and NERR bits must be used to generate BINIT#,
BERR#, INTREQ# or whatever the appropriate action is.
Both the FERR and NERR registers are write-1-to-clear registers. This means that software must
write a one to the bits it wishes to clear.
For the GXB, all NERR and FERR registers must be cleared before FERR_GXB is cleared,
otherwise FERR_GXB will be set again.
6.4.1
6.4.2
Masked Bits
Many of the errors have conditional reporting behavior. The error always sets the FERR register or
NERR register. If the error is masked then a BINIT# or whatever is supposed to happen will not
occur, but the bit is still set. The mask will not prevent the error from appearing in FERR/NERR.
This allows software to poll, looking for errors that are not fatal to the system. If an error is
masked, it will still set FERR and force all other errors to appear in NERR, thus losing logging
information regarding later errors.
BERR#/BINIT# Generation
When an error occurs that forces BINIT#, then an enable bit in CONFIG2 is cleared as BINIT# is
driven to the bus. The enable bit is automatically cleared in order to mask further BINIT#
assertions. Software may also explicitly clear the enable bit to prevent BINIT# from occurring.
When an error occurs, software should go out, clean up the error, clear the error status registers and
only then set the enable bit so that new errors will be seen. When the bit is set, which can only be
done by software writing the bit, then the SAC will assert BINIT# on an error. If a new error occurs
after the first one is handled and is pending in FERR or NERR, then it will be reported when
software re-enables BINIT# reporting.
The same behavior is true for BERR# as well. It is enabled by a separate bit in CONFIG2.
6.4.3
INTREQ#
The INTREQ# signal is driven by the SAC when it wishes to cause an interrupt and signal the
operations system that an error has occurred that is non-fatal, but that may need to be logged. The
signal is held asserted as long as the condition that caused the error exists. All the errors that cause
an interrupt are OR-ed together to drive this signal. Software will reset the bit in FERR or NERR
that caused the error. When the bit is reset, INTREQ# should be deasserted unless there is another
error which holds the signal active.
After software clears the NERR/FERR bit that caused the error, it will do an EOI to the PID to re-
enable interrupt reporting by the PID. If INTREQ# is still active after the EOI, then a new interrupt
is generated.
6-4
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
6.4.4
XBINIT#
XBINIT# is an input to the SAC and an output from one of the xXB’s or can also be generated by
platform logic. XBINIT# is GTL+ level, and therefore all the outputs from the xXB’s can be tied
together and fed into the SAC’s input. XBINIT# is held by the xXB until the xXB is reset, so if it
takes multiple clocks to drive from the xXB to the SAC, there should be no problem.
6.4.5
XSERR#
XSERR# is an input to the SAC. It is generated by the system from the PCI SERR# signals or for
other reasons. The system must OR all the generators together, either doing a wire-OR or adding
logic for the OR’ing function.
6.5
SAC/SDC Errors
Many errors require reading both the SAC and SDC to isolate the cause. This is because most
transactions involve both chips with a fair amount of handshaking between them. Refer to
Table 6-1 for which register holds the status information and which ones hold error logging
information. Also this table shows which errors are fatal and cause BINIT# and which are
interrupts. Also some errors are maskable.
6.5.1
Data ECC or Parity Errors
The following errors are captured by the SAC and SDC together:
• SDC Non-Fatal Error (SNE). This is set for any of the following conditions: a) double-bit ECC
error on memory b) double-bit ECC error from the system bus c) parity error on the private bus
for data d) parity error on the private bus for byte-enables e) an internal SDC ram parity error
f) a single bit correctable error on the system bus or g) the 2nd or subsequent single-bit
memory ECC errors that are not recorded by the SDC as the first 1x error. On this error
software must read the SDC to determine the type of error that was found. If the SDC reports
only single bit errors, then the SNE bit was set for a 2nd or later single-bit memory error. Once
the SDC’s error registers are cleared, the first single-bit memory error is recorded as a SCME,
and then later ones are SNE.
• SDC Correctable Memory Error (SCME). This is set on the first single-bit ECC error when
reading DRAM. The error is always corrected before being passed on to the next interface.
Since soft memory errors are not un-common, this bit has its own mask enable/disable. NOTE:
if the SDC were to receive a single bit failure from both the A and B memory boards on the
same clock, then FERR(SCME) and NERR(SCME) would be set. Otherwise, because only the
first correctable error is reported as SCME, NERR(SCME) should not be set. This assumes
that the SAC and SDC have their error registers cleared correctly. If the SDC has it’s error
registers cleared before the SAC and in between this time there is a new 1x memory error, then
the SAC_NERR(SCME) bit would get set.
• SDC Fatal Error. This is set when the SDC detects conditions that will cause an unconditional
BINIT#. The exact cause must be read from the SDC. This is set for parity errors on the
control interface between the SAC and SDC or for protocol errors.
Intel® 460GX Chipset Software Developer’s Manual
6-5
Data Integrity and Error Handling
6.5.2
System Bus Errors
There are several errors that are detected by the SAC.
• System Bus Address Parity Error. Parity is checked on both address phases.
— On A[43:24]#, detected by AP[1]#.
— On A[23:3]#, detected by AP[0]#.
• System Bus Request Parity Error. Parity on both phases of the request bus is checked.
— On REQ[4:0]#, detected by RP#.
• Address above TOM. Set for addresses above TOM and not in a PCI range. These addresses
can’t be sent to PCI0, since they may be greater than 4 GB, so are fatal. Out-bound addresses
required a DAC on the PCI bus are not supported and would cause a fault at the xXB.
• Unsupported ASZ#. Since the GX only supports 36bits of address space, accesses which have
ASZ# = 10b or 11b are an error.
• IOQ Underflow/Overflow. This occurs when there are no entries in the IOQ and the SDC
attempts to do a Response Phase. Since the IOQ is empty, there should not be a Response
phase. Or it can occur when either of the following 2 conditions are met: a) when there are 8
entries in the IOQ and a new ADS is seen b) when the IOQ depth is set to 1 for the system and
the IOQ has one entry and a new ADS is seen.
• BERR# Observed#. When the GX sees BERR# active on the bus, whether driven by the
processor or the SAC, it will elevate that to BINIT# if the ‘BERR# to BINIT# Enable’ bit in
CONFIG register is set. If the enable bit is not set, then BERR# is ignored as an input. BERR#
is driven active on the bus for 3 clocks. Each time a new BERR# assertion is sampled, BINIT#
will be driven, unless the error is masked off.
• LOCK# Transaction with No Resources Available. Set when a LOCK# occurs and there are no
outbound resources for the transaction. Since the lock can’t be retried and there is no place to
put the transaction, it gets dropped and lost.
• Resource Counter Overflow/Underflow. Set when the resource counter has an overflow or
underflow. This occurs if there is a retirement to a counter that is empty or a transaction is not
retried when the counter is full.
There is one logging register in the SAC for recording the actual error information. This is
SA_FERR. It captures the system bus address and request for both the a and b phases. This register
can be used to determine which bit is bad on parity errors.
6.5.3
SAC to SDC Interface Errors
The SAC will detect the following errors on the interface between the SDC and the SAC. They are
all flagged in FERR_SAC.
• PDB ITID parity error. Set on a data transfer from SDC to SAC that has bad parity on the ITID
sent with the data.
• Retirement Bus parity error. Set when the SDC attempts to retire an ITID and a parity error
occurs.
• False retirement seen by SAC. Set when the SAC sees a retirement to an ITID that is not in
use.
• ‘Store-Retire’ Command Underflow. Invalid write data sent from SDC to MDC. When the
SDC sends data to the MDC, it signals that it did so to the SAC. This error is flagged when the
‘data-sent’ signal is seen by the SAC, but there are no writes to that stack pending in the SAC.
6-6
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
The SDC will capture the following errors on its side of the interface.
• PDB Data Parity Error. On data received from the SAC, parity is checked. If parity is bad, the
data is sent to memory or the system bus with poisoned ECC.
• PDB Byte Enable Parity Error. On parity errors for the byte enables, the entire data transfer is
sent to memory or the system bus with bad ECC. If the transfer is 64B then the entire line will
have bad ECC.
• PDB Command Parity Error. Parity is checked over the Command bus from the SAC to SDC.
• PDB ITID Parity Error. Parity is checked on the ITID bus which accompanies the data from
the SAC. The data is dropped.
• PDB Receive Length Error. Occurs when the SDC receives data and the length of the data
transfer doesn’t match the length indicated by the command for that transfer.
• Configuration Information Parity Error. Config accesses from the SAC to the SDC go through
the data buffer in the SDC for both the address and data of the targeted register. Parity errors
on reading out either the address or data from the buffer are detected.
6.5.4
SAC to MAC Interface Errors
The SAC will detect the following errors on the memory card interface.
• Completion Command Underflow. When the MAC has completed a transaction (either a read
or a write), it will send a ‘Complete’ back to the SAC. Since each stack is kept ordered, the
SAC can pop the top of the queue for the stack sending the ‘Complete’. This error is flagged
when the stack queue is empty and a ‘Complete’ is received for that stack.
The MAC will detect the following error in its interface with the SAC.
• Memory Card Error. This is set on a parity error on the command sent from the SAC. The
command bus is a 23 bit bus (22 address bits and one parity bit). This error is flagged in the
MAC when there is bad parity on this bus. The bus should always be driven with good parity.
Parity checking is done every cycle. The MAC will complete any outstanding transactions, i.e.
those with RAS already started. It will not start any new transactions for either stack until it is
reset. Refreshes will continue, though no new accesses will start until the MAC is reset. The
error is signaled to the system on the MAC’s ERR# pin, which sets FERR_SAC [MAE or
MBE].
6.5.5
SDC/Memory Card Interface Errors
The following errors are detected.
• Card x SEC. Data corrected and placed into the data buffer.
• Card x DED. Data placed into the data buffer with bad parity on each chunk that had bad ECC.
Data is sent to the system bus with poisoned ECC or to the PDB with bad parity.
• ‘Accept’ Underflow. Set when the SDC received data from the memory card without having
an ‘Accept’ command pending.
• ‘Forward’ Underflow. Set when the SDC receives a Forward signal from the MAC without a
corresponding ‘Store’ command.
• ‘Forward’ Overlapping ‘Forward’. Set when the SDC is doing a ‘Forward’ by sending data
and then another ‘Forward’ is seen before the first finishes. The SDC does not queue up these
commands and the timings would imply that two lines are being transferred at the same time.
• ‘Load’ overlapping ‘Load’. Set when the SDC is doing a ‘Load’ by receiving data and a
second ‘Load’ is seen before the first ‘Load’ finishes. This implies the MDC is sending data
for two different lines at the same time.
Intel® 460GX Chipset Software Developer’s Manual
6-7
Data Integrity and Error Handling
• ‘Load’ Overlapping ‘Forward’. Set when the SDC is doing a ‘Forward’ by sending data to the
MDC and the MDC starts to send the SDC data before the ‘Forward’ is complete.
• ‘Forward’ overlapping ‘Load’. Set when the SDC is receiving data from the MDC, and then is
told to send data to the MDC at the same time it is receiving data.
6.5.6
SDC/System Bus Errors
• LEN# Protocol Error. Set when the actual write data received not matching length given by the
system bus transaction. Set when the processor transfers either more or less data than is
indicated by the LEN# encoding on the bus transaction.
• Write Data Protocol Error. Set when data is received from processor without a TRDY# having
been given. Set when the processor places data on the bus without the correct TRDY#
assertion. This may be for writes or implicit writebacks. It is also set if the processor drives a
DRDY# on a zero-length write transaction.
• DRDY# Protocol Error. This error occurs and is flagged if a) DRDY# is not deasserted 1 clock
after DBSY# deassertion OR b) DRDY# is not deasserted 2 clocks after SBUSY# deassertion.
Either of these implies that the processor is holding onto the data bus longer than it is supposed
to.
6.5.7
SDC Internal Errors
The SDC has two internal error bits that it detects.
• Data Buffer Parity Error. This is set when the data buffer detects a parity error on data that was
placed into the buffer as good. Data with uncorrectable errors, such as double-bit ECC errors
or data parity errors from the private bus, is placed in the buffer with good parity and has a bit
set to indicate that the data is uncorrectable. If the data is read out of the buffer and parity is
bad, then an internal alpha-hit or other error occurred.
• Simultaneous write-one-to-clear and hardware set. When the SDC_FERR register is set, it can
only be cleared by writing a one to the set bit. If on the cycle that SDC_FERR is cleared, there
is an error for that same bit, then 2 things happen a) the bit is set to a one and behaves as
normal and b) the Simultaneous Clear and Set error bit is set. This bit indicates that there was
not time to reset the logging registers associated with the incoming error, so that the log
registers are stale and cannot be considered as valid.
If software is writing a single one to the unique asserted bit in SDC_FERR, then the
Simultaneous bit is only set if the new error is to the bit being cleared by software. If software
is clearing the register by writing all 1’s to every location, then any new incoming error will
also set the simultaneous bit. Software should only write to the single location (or locations)
that it wishes to clear, and not write all 1’s to the entire register.
6.6
Error Determination
the actual error. This is listed under ‘Log Register’. Parity or ECC errors generally capture the data
and the parity/ECC bits for the failing transfer. This information can be used for debug and
diagnostic purposes. The log register is updated when the appropriate bit is set in the status register.
Only after the status register (FERR register) is cleared will a new value be captured on subsequent
errors.
6-8
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
Other errors capture the address associated with the failure. This is also for debug and diagnostic
purposes, but also has the potential for use in system recovery. For instance, if there is an
uncorrectable error on a data read, and the access can be isolated, then instead of re-starting the
whole system, it might be possible to kill only the failing process and allow other users to continue
running.
6.6.1
SAC Address on an Error
The SAC has several registers used to access the address for a failure. After FERR_SAC is read to
determine the precise error that occurred, the address can then be determined for certain errors. The
method is somewhat indirect. The SAC is the only chip that tracks the original address, so is used
to get the address even when another chip may have detected the error.
The GX keeps track of all transactions using an ITID (Internal Transaction ID). All the chips use
this tag to track a specific transaction through the system. When the transaction is compete the
ITID is retired and a later transaction may re-use the same value. On an error the ITID for that
failing transaction is captured and not retired. There are 3 registers used by the SAC to capture the
failing ITID:
• SECTID - captures the ITID of the first Single-bit ECC error from memory.
• DEDTID - captures the ITID of the first Double-bit ECC error from memory.
• FSETID - captures the ITID of the first system bus or PDB data error seen by the SDC.
SECTID, DEDTID, and FSETID all have ‘Valid’ bits in the register. The ITID is captured until the
error is logged and the system is ready to clear all the error indication. Single, Double and system
bus errors are recoverable, and therefore the system can clear those errors and continue running. To
do so software must write a one to the ‘Valid’ bit of the register. This will cause the system to retire
the ITID and that transaction is now complete.
Note that these 3 registers are sticky through reset, so that the information is preserved. After
BINIT# or reset (but not power-on), the SECTID, DEDTID and FSETID registers are valid and the
failing address can be retrieved from the RAM.
Multiple of these registers may be active. After the status register is read then the log register can
be read to determine the ITID. The SDC signals the first of each type of error. If the first error the
SDC sees is a Double-bit ECC then the FERR_SAC and SDC_FERR bits are set for that along
with DEDTID in the SAC. If the SDC sees a single-bit error next, before software has cleared out
the error logging, then the SDC_NERR and NERR_SAC registers are set, as well as SECTID.
Anytime the SDC sends an indication of a Single, Double or system bus error, the appropriate
SECTID, DEDTID, FSETID register is set. The SDC is responsible for not sending a 2nd
indication of one of these errors, until it’s FERR and NERR registers are cleared.
When there is an error in the SDC and software has finished processing it, it should follow the
following procedure in the order given:
• Write to the ‘Valid’ bit of SECTID, DEDTID or FSETID to release the ITID and unlock the
register.
• Write ‘1’ to clear those bits that software has read as asserted in FERR_SAC and the
NERR_SAC registers. Software should not just write 1’s to the entire register.
• Since any of these 3 types of errors are reported through interrupt and not BERR# or BINIT#,
the interrupt must be cleared so that the next error can be visible. Write an EOI to the PID.
• Write ‘1’ to clear the SDC_FERR and SDC_NERR registers. Writing a ‘1’ to either register in
the SDC will clear both the SDC_FERR and SDC_NERR at the same time.
Intel® 460GX Chipset Software Developer’s Manual
6-9
Data Integrity and Error Handling
After this the error reporting is in the clean state.
After the ITID is found, the actual address is needed. Again this is somewhat indirect. There are 2
locations in the SAC in which the address may be found. One is the Bus Interface Unit’s (BIU)
CAM and RAM, and the second is the MIU’s RAM. The BIU’s CAM contains the address for
coherent transactions. The RAM contains the address, command and other information. See the
register definition for BIUDATA for the bit definitions of the information captured. For the MIU,
the RAM contains the address of the transaction before it has been translated into a memory
address. This address may have come from the system bus for a processor or a 460GX-initiated
coherent access, or it may have come from AGP or other non-coherent source. The RAM/CAM
may only be used for coherent transactions. The MIU tracks both coherent and non-coherent
transactions. Non-coherent transactions are those sent by an AGP card. Software can simply read
the MIU for all transactions. If software reads the MIU and RAM/CAM and gets the same value for
the address, then the transaction was coherent and the rest of the RAM/CAM data is valid. If the 2
addresses are different, then software should not read the rest of the RAM/CAM.
To access the system bus’s address, the ITID which was read from one of the registers listed above
is written into BIUITID. The write to this register causes the register BIUDATA to be updated with
the RAM and CAM contents associated with the ITID written into BIUITID. BIUDATA can be
read and the address determined.
To access the memory’s address buffer, the procedure is slightly different. This buffer is directly
readable, instead of using the indirect approach used by the BIU. To read the MIU address do:
• Read ITID from one of the registers above:
— If the ITID is less than or equal to 31, then do a configuration read from BUS: CBN,
Device: 1, Function: 2, Address: (80h+4*ITID). This is MEMTID0 register.
— If the ITID is greater than 31, then do a configuration read from BUS: CBN, Device: 1,
Function: 3, Address: (80h+4*(ITID-32)). This is MEMTID1 register.
6.6.1.1
Special Notes on Usage of SECTID, DEDTID, FSETID Registers
The SAC’s SECTID, DEDTID and FSETID registers all define bit [7] as a Disable bit. When set,
the register will not capture the ITID and the transaction is retired immediately. There is one side-
effect that isn’t apparent when setting the Disable bit to ‘1’. Not only is the ITID not captured, the
FERR or NERR registers are not set either. Therefore if the Disable bit is a ‘1’, the SAC will not
indicate any error, and there will be no interrupt or signal to the system. The SDC is not affected
by setting the Disable bit, so it will continue to log the errors.
— If software sets the Disable bit in any of these 3 registers, it is responsible for polling the
SDC for errors or accept that the errors will not be reported to the system. It is
recommended that the Disable bit be set only for special applications or usage.
6.6.2
SDC Logging Registers
The SDC has logging registers to capture single and multiple bit errors on all the interfaces. These
are:
• SEC0 - first single-bit ECC error on memory card 0.
• DED0 - first double-bit ECC error on memory card 0.
• SEC1 - first single-bit ECC error on memory card 1.
• DED1 - first double-bit ECC error on memory card 1.
• PVD - first parity error on the private data bus.
• SECF - first single-bit ECC error on the system bus.
6-10
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
• DEDF - first double-bit ECC error on the system bus.
• PCMD - first parity error on the command bus.
• PITID - first parity error on the ITID bus.
• SDCRSP - first failing transmission on the response bus. The response bus does not have
parity, instead it sends the response in clock x and the inversion of the response in clock x+1.
All of these registers are independent. Having one of these valid doe not block any other from
being valid. They record the first error of each type. Each one is locked only when the bit which is
associated with the particular error in either SDC_FERR or SDC_NERR is set. In theory all the
above registers could contain valid error information. There may be many cases where there are
both single and double bit errors set as valid, especially if the single-bit errors are not scrubbed. If
one whole line is bad both FERR and NERR get set, since each data chunk of 16B is considered as
an independent unit.
Note: The error logging registers are enabled anytime the FERR and NERR registers both have the
appropriate bit cleared. If an error is found and the FERR and NERR registers are cleared, then the
next error will overwrite the old value in the logging register. If FERR or NERR is not cleared, then
new errors will not overwrite the logging register.
6.7
Clearing Errors
Firmware or the operating system must clear out all the error registers when returning from a
BINIT# or a reset. Leaving error bits set in the error registers will cause the system to flag that an
error from an earlier time is still present. Firmware should read each FERR and NERR register and
log any bits that are set. It should then clear those bits and continue to the next group of error
registers.
6.7.1
SAC/SDC Error Clearing
In the case of a data error, both the SAC and SDC will log the event. After handling the error,
software should clear the error logging in the following order
• Read the SECTID, DEDTID, or FSETID register as above
• Clear the SECTID, DEDTID or FSETID registers by writing a ‘1’ to bit 6 of the appropriate
register
• Clear the SAC_NERR register if set
• Clear the SAC_FERR register
• Clear the SDC error registers only after the SAC’s registers are cleared
• Re-read the error registers to make sure they are still clear
• Re-enable interrupts or other system signaling
6.8
Multiple Errors
With the number of errors that are detected in the 460GX chipset, there are many possible multiple
error cases. There is no way to specify what can happen in the case of every combination. In
general the first error that is found is the important one. If there is only one error then it may be
possible for the operating system to recover and continue without a full re-boot of the system. If
there is more than one error, except for correctable ECC errors, then the system is probably not
Intel® 460GX Chipset Software Developer’s Manual
6-11
Data Integrity and Error Handling
capable of any recovery. The first error, especially if it is fatal, may itself have caused downstream
errors to be flagged. The error that is flagged as first should be considered as correct and an
indication of some real problem. The problem may have been a transient condition or a true
hardware failure. Errors that are flagged in the NERR register should not be considered as caused
by a hardware failure, since the first error may have actually precipitated the later errors.
6.8.1
SDC Multiple Errors
There are several important cases of multiple errors in the SDC. These are:
The data buffer in the SDC is 16B wide. Each 16B is checked for errors. Therefore if a single line
has errors in multiple 16B chunks, then both the SDC_FERR and SDC_NERR registers will be set.
ECC checking is done on 8B, and it is possible that a 16B chunk has both 8B chunks as bad; this
case will be reported as a single error and only SDC_FERR is set. Since this is one line, there is
only one error reported from the SDC to the SAC, so only FERR_SAC is set (or NERR_SAC if
FERR_SAC were previously set by a different error).
Multiple lines in the SDC may have errors. For example, memory ECC errors are marked in the
Data Buffer as the data is read from memory and placed in the buffer. As the data is sent from the
SDC to the appropriate target, an indication is sent to the SAC as to whether there was an error or
not on that transfer. If multiple lines have a 2x memory ECC error, only the first line that was
retrieved by the SDC will have the SAC FERR error set. Later 2x memory errors are flagged as
generic errors. For the SAC_FERR, generic errors set the SNE bit. Therefore, if SAC_NERR has
the SNE bit set, this simply means that the SDC had multiple errors that were non-fatal. This may
be multiple errors of the same type or different types. Even 1x correctable memory errors will set
the SNE bit of SAC_NERR if there are more than one.
One transaction may have any combination of front-side bus, single-bit memory ECC error, and/or
double-bit memory ECC error. The following shows what is captured for the different cases, where
FSE is front-side bus, DED is 2x memory ECC error, and SEC is 1x memory ECC error:
FSE | SEC | DED | Status
ERR | ERR | ERR |
0
0
0
0
1
1
1
1
|
|
|
|
|
|
|
|
0
1
0
1
0
0
1
1
|
|
|
|
|
|
|
|
0
0
1
1
0
1
0
1
|
|
|
|
|
|
|
|
Store nothing, no errors
Store ITID in SECTID, log SCME in FERR
Store ITID in DEDTID, log SNE in FERR
Store ITID in DEDTID, log only SNE in FERR
Store ITID in FSETID, log SNE in FERR
Store ITID in DEDTID, log SNE in FERR
Store ITID in FSETID, log only SNE in FERR
Store ITID in DEDTID, log only SNE in FERR
• If there are multiple 1x Memory ECC errors, then the FERR[SCME] bit is set in the SAC.
Also the FERR[SNE] bit will be set on the second 1x Memory ECC error (since SCME does
not prevent FERR from being written. If there are more than 2 errors, then NERR[SNE] will
also be set. Software must look at the SDC to determine that all the errors were single bit
errors.
6-12
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
6.8.2
6.8.3
SAC Multiple Errors
There are several important cases of multiple errors in the SAC. Some of these are caused by the
SAC and SDC not being in one chip and therefore having delays in the handshaking paths that will
allow events that occur after a fatal problem to appear as errors, even when they are not real.
Some of the multiple errors and the behavior for the SAC are:
• If the system has a fatal error and there is a BINIT#, then the NERR register may contain
indications of other errors. For instance if there is a retirement from the SDC during the clock
the BINIT# is clearing the system bus queues, then a retirement underflow may be flagged.
Single Errors with Multiple Reporting
There are some cases where multiple error bits are set as data is passed along.
Data moving to the Expander bus from the SDC may cause multiple error bits to be set. For
instance, there is a PCI read that gets a system bus or memory ECC error. The SDC will flag the
error and the expander unit will also flag that the error occurred. Either may come in first, so that
FERR may have the expander bus error and NERR have SDC system bus error, or vice-versa.
A second case is for data from the expander bus to memory. The data is poisoned across the private
data bus. Therefore FERR_SAC will be set for an expander bus parity error, and NERR_SAC will
be set for a SNE error from the SDC.
The GXB may have an address from the graphics card that is in the aperture range, but finds the
Valid bit cleared in the GART entry. This will set FERR_GART[2]. Since the address won’t be
translated, it is likely to be outside a valid memory range, thus an Illegal Address error is likely.
Since this is one operation, both bits 2 and 1 of FERR_GART are set.
Data in the SDC is handled on an 8-byte basis, so that poisoned data causes 8 bytes to have bad
parity. The expander interface in the SAC handles the data on a 4-byte basis. Therefore a single
8B data transfer that is on an 8B aligned address will cause both the FERR and NERR bits to be
set, because each 4B is flagged as a separate error. If the access starts on an odd 4B boundary, then
only the FERR register is set, since only 4 bytes are accessed out of the bad 8 bytes.
6.8.4
Error Anomalies
There are several cases that may have unexpected behavior. These are listed below.
Assume there is a 2x error in a certain 8B chunk of DRAM. There is a partial write which writes
all 8 bytes of the bad chunk. One might expect that, since the partial write overwrote all 8 bytes of
the chunk, the error would not be reported. The GX design is such that the data buffer in the SDC
tracks errors on 8B chunks. The data is marked as having a 2x error when the line is read out of
DRAM. Even though the entire 8B is overwritten, the chunk is still marked as having a 2x error.
When the line with the merged data is written back into DRAM, the chunk that has the 2x error will
be poisoned and later reads will report a 2x error on that line. NOTE: if there is a full line write,
then any line in DRAM with a 2x ECC error will be overwritten, and the 2x error will never be
seen.
An expander bus MWI writes an entire line into DRAM. On the system bus there is an implicit
writeback. The data from the processor for the IWB has a 2x error. Even though the processor data
is not used, since the expander bustransaction was a full line write, the data will be sent to DRAM
as poisoned.
Intel® 460GX Chipset Software Developer’s Manual
6-13
Data Integrity and Error Handling
Take the case where processor 1 reads a line from memory and there are no errors, and then does a
write into its cache. Later processor 2 does a read, getting an IWB. The SAC starts a speculative
read for the line before seeing the HITM#. If the data in DRAM has a 2x error on this read the
SDC data buffer will be marked as having a 2x error on this line. In this case, assuming the
speculative read is done (in some cases the HITM# is seen before the memory cycle is attempted,
so no speculative read is done), the IWB data that is written back into memory will be poisoned.
6.9
Data Flow Errors
DRAM errors for either single-bit (S) or double-bit (D) fails. The SDC will never signal that an
error was fatal (causing BINIT#). The processor will handle data that it receives with bad ECC.
The memory will have bad ECC written to it, if incoming data is bad. Data to the SAC will be
handled by the SAC itself. Either the SAC will BINIT# on seeing bad data or it will pass it on to
the xXB via an Expander port.
Figure 6-1. SAC Error Flow on Data
System Bus
ITID
Address parity.
Check parity.
Request parity
.
Bus protocol checking
RETIREMENT
BUS
Check parity.
.
Pass data with
good/bad parity to
SDC or Expander (peer).
No checking done.
TO
SDC
1 parity per 18 bits (16 data and 2 BE)
68b bus
1 parity per
16b of data
1 parity per 16 bits (16 data)
From
BIU
Pass data from SDC
to Expander bus with
good/bad parity.
From
SDC
1 parity, 16 data,
2 BE
2b BE
From
BIU
No checking done.
16b data,
1Parity
Generate
parity as DP xor
BE0 xor BE1, not
18-way XOR
Check parity on
header and data.
Check for HF.
Check outgoing
parity from data
buffer.
Expander: From xXB
Expander: To xXB
6-14
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
Figure 6-2. SDC Error Data Flow
From Processor
To Processor
Check parity, if bad
poison ECC of chunk.
Retire ITID and report error.
Correct if 1x; if 2x, then write
into DB with bad parity on chunk
that is bad. Set status. No reporting.
Check and flag
status;
1b/16b parity
regenerate
new parity, if
incoming is bad
then generate new
parity as bad.
On data to SAC
send with bad
parity if parity
if data is bad in
buffer. Report
error.
1b/8b parity
To
SAC
Data Buffer
S
D
Correct if 1x;
if 2x write to
DB with bad
parity, flag
status. Set S
or D bit if error
Check, if bad
poison ECC of
bad chunk. Retire
ITID and report
error.
6.10
Error Conditions
6.10.1
Table of Errors
Table 6-1 is a list of possible errors found in the system. The table shows the error and the system
action. It also shows the information that is captured on any failure. The captured info is sticky
through reset and can be read through configuration accesses. If the system action is conditional,
then the qualifier column shows what the action is conditional on. Note that for the BINIT#,
XBINIT#, and the interrupts there is a driver enable. If the driver is disabled, then these signals
won’t be active, even if it says Unconditional in the table.
Intel® 460GX Chipset Software Developer’s Manual
6-15
Data Integrity and Error Handling
Table 6-1. Error Cases
Chip
Error
System
Action
Status
Register
Log
Register
Qualifier
Detecting
System Bus 1x
ECC
SDC
SDC
SDC
SDC
Correct the data and
pass to bus.
SDC_FERR[SECF],
FERR_SAC[SNE]
SECF_D_FERR,
SECF_ECC_FERR,
SECF_TXINFO_FERR,
FSETID
SNE
Enable
Conditional Interrupt.
System Bus 2x
ECC
Write into memory with
bad ECC (poison data).
SDC_FERR[DEDF],
FERR_SAC[SNE]
DEDF_D_FERR,
DEDF_ECC_FERR,
DEDF_TXINFO_FERR,
FSETID
SNE
Enable
Conditional Interrupt.
Mem 1x ECC
Mem 2x ECC
Correct the data and
pass to bus.
SDC_FERR[SECx],
FERR_SAC[SCME]
SECx_D_FERR,
SECx_ECC_FERR,
SECx_TXINFO_FERR,
SECTID
SCME
Enable
Conditional Interrupt.
Pass to system bus with SDC_FERR[DEDx],
bad ECC (poison data).
DEDx_D_FERR,
DEDx_ECC_FERR,
DEDx_TXINFO_FERR,
DEDTID
SNE
Enable
FERR_SAC[SNE]
Conditional Interrupt.
SAC to SDC Interface Errors
PDB Command
Parity Error
SDC
SDC
SDC
SDC
Unconditional BINIT#
Unconditional BINIT#
Unconditional BINIT#
Unconditional BINIT#
SDC_FERR[CPE],
FERR_SAC[SFE]
PCMD_FERR
SDCRSP_FERR
PITID_FERR
Nothing
RSP Bus
Transmission Error
SDC_FERR[RTE],
FERR_SAC[SFE]
ITID Parity Error
SDC_FERR[IPE],
FERR_SAC[SFE]
PDB Receive
Length Error
SDC_FERR[RLE],
DPBRLE_FERR,
FERR_SAC[SFE]
PDB Data Parity
Error
SDC
Failing Chunk of Data
will be put in memory
with bad ECC
SDC_FERR[DPE],
FERR_SAC[SNE]
PVD_D_FERR,
PVD_PAR_FERR,
PVD_TXINFO_FERR
SNE
Enable
(poisoned),
Conditional Interrupt.
PDB Byte Enables SDC
Parity Error
Entire Line of Data will
be put in memory with
bad ECC (poisoned),
Conditional Interrupt.
SDC_FERR[BPE],
FERR_SAC[SNE]
PVD_D_FERR,
PVD_PAR_FERR,
PVD_TXINFO_FERR
SNE
Enable
Configuration
Information Parity
Error
SDC
Unconditional BINIT#
SDC_FERR[CIE],
FERR_SAC[SFE]
Nothing
False retirement
seen by SAC
SAC
SAC
SAC
SAC
Unconditional BINIT#
Unconditional BINIT#
Unconditional BINIT#
Unconditional BINIT#
FERR_SAC[FRE]
FERR_SAC[RPE]
FERR_SAC[IPE]
FERR_SAC[SCxx]
Nothing
Nothing
Nothing
Nothing
Retirement Bus
Parity Error
PDB ITID Parity
Error
‘Store-Retire’
Underflow
System Bus ADD/CMND
Address Parity
Error
SAC
Conditional BINIT#
FERR_SAC[AE]
SA_FERR
Flag
System
Bus
Parity
Errors
6-16
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
Table 6-1. Error Cases (Cont’d)
Chip
Error
System
Action
Status
Register
Log
Qualifier
Register
Detecting
Request Parity
Error
SAC
Conditional BINIT#
FERR_SAC[RQE]
SA_FERR
Flag
System
Bus
Parity
Errors
Unsupported ASZ
SAC
Conditional BINIT#
FERR_SAC[ASE]
SA_FERR
Flag
System
Bus
Parity
Errors
Illegal HITM (HITM SAC
on any Expander
bus access)
Unconditional BINIT#
Unconditional BINIT#
FERR_SAC[IHS]
FERR_SAC[TE]
Nothing
Nothing
Address above
TOM
SAC
LEN# Protocol
Error
SDC
SDC
Unconditional BINIT#
Unconditional BINIT#
Unconditional BINIT#
Convert to BINIT#
SDC_FERR[FS0],
FERR_SAC[SFE]
Nothing
Nothing
Nothing
Nothing
DRDY# Protocol
Error
SDC_FERR[FS2],
FERR_SAC[SFE]
Write Data Protocol SDC
Error
SDC_FERR[FS1],
FERR_SAC[SFE]
BERR# Observed
SAC
FERR_SAC
BERR# to
BINIT#
Enable
IOQ Overflow/
Underflow
SAC
SAC
SAC
Unconditional BINIT#
Unconditional BINIT#
Conditional BERR#
FERR_SAC[IUE]
FERR_SAC[XBE]
FERR_SAC[XSA]
Nothing
Nothing
Nothing
External XBINIT#
Active
External XSERR#
Active
BERR#
Driver
Enable
LOCK# when no
resources are
available
SAC
SAC
Unconditional BINIT#
Unconditional BINIT#
FERR_SAC[LTE]
FERR_SAC[RCE]
Nothing
Nothing
Resource Counter
Overflow/
Underflow
Memory System Protocol Errors
SAC-MAC
Command Parity
Error
MAC
ERR# pin asserted on
MAC.
Unconditional BINIT# by
SAC.
FERR_SAC[MxE],
FERR_MAC
CMND_FERR
‘Complete’
Command
Underflow
SAC
SDC
Unconditional BINIT#
Unconditional BINIT#
FERR_SAC[CCxx]
Nothing
Nothing
‘Forward’
overlapping
‘Forward’
SDC_FERR[FWMDIx],
FERR_SAC[SFE]
‘Forward’
overlapping ‘Load’
SDC
SDC
SDC
SDC
Unconditional BINIT#
Unconditional BINIT#
Unconditional BINIT#
Unconditional BINIT#
SDC_FERR[RdWrx],
FERR_SAC[SFE]
Nothing
Nothing
Nothing
Nothing
‘Load’ overlapping
‘Load’
SDC_FERR[LRMDIx],
FERR_SAC[SFE]
‘Load’ overlapping
‘Forward’
SDC_FERR[WrRdx],
FERR_SAC[SFE]
‘Forward’
Underflow
SDC_FERR[FRx, FLx],
FERR_SAC[SFE]
Intel® 460GX Chipset Software Developer’s Manual
6-17
Data Integrity and Error Handling
Table 6-1. Error Cases (Cont’d)
Chip
Error
System
Action
Status
Register
Log
Register
Qualifier
Detecting
‘Accept’ Underflow SDC
Unconditional BINIT#
Unconditional Interrupt
SDC_FERR[AEx],
FERR_SAC[SFE]
Nothing
Nothing
Internal SDC Error
Data Buffer Ram
Parity Error
SDC
SDC_FERR[RPE],
FERR_SAC[SNE]
GXB ERRORS
AGP Request
Queue Overflow
GXB
GXB
GXB
GXB
Unconditional XBINIT#
Unconditional XBINIT#
Unconditional XBINIT#
Unconditional XBINIT#
FERR_AGP
FERR_AGP
FERR_AGP
FERR_AGP
Nothing
Nothing
Nothing
Nothing
Use of Pipe with
Sideband Enabled
AGP Address [
..36] not = 0
Unsupported
command using
AGP semantics
PCI IB Read Que
Data Parity Error
GXB
GXB
GXB
Conditional XINTR#
Conditional XBINIT#
FERR_PCI
FERR_PCI
Nothing
Nothing
Nothing
TXDERR
_INTE,
TXDERR
_BINITE
PCI OB Write Que
Data Parity Error
Conditional XINTR#
Conditional XBINIT#
TXDERR
_INTE,
TXDERR
_BINITE
Discard timer
expiration
Unconditional XINTR#
SERR# if SERRE set
FERR_PCI
FERR_PCI
SERRE
SERR# Observed
PERR# Observed
GXB
GXB
Unconditional XINTR#
Unconditional XINTR#
Nothing
Nothing
FERR_PCI,
possibly PCISTS[DPE]
PCI Parity Error on GXB
Address from Card
Let card master abort;
SERR# and XINTR# if
SERRE set; if SERRE
not set then neither
SERR# nor XINTR#
driven.
PCISTS[PE],
FERR_PCI,
possibly PCISTS[SSE]
PAC_ERR
SERRE
PCI Parity Error on GXB
Data from Card
Data placed into queue
with bad parity.
Conditional PERR#
PCISTS[PE],
PCISTS[DPE] if PERRE
PD_ERR, PAC_ERR
Nothing
PCICM
[PERRE]
Master Abort on
Read by GXB
GXB
GXB
1DW: Return all 1’s
> 1DW: Hard Fail (HF)
completion.
PCISTS [RMA],
FERR_PCI
Master Abort on
1 DW: normal
completion
PCISTS [RMA],
FERR_PCI
Nothing
Write done by GXB
>1 DW: Hard Fail (HF)
completion.
Master Abort on
Configuration
Cycle
GXB
GXB
Normal completion
PCISTS [RMA]
(FERR_PCI is not set)
Nothing
Nothing
Nothing
Target Abort on
Transaction
Mastered by GXB
Return HF to either read PCISTS [RTA],
or write.
FERR_PCI
GART Entry Invalid GXB
Unconditional XINTR#,
Conditional XBINIT#.
(NOTE: if XBINIT# is
driven, then it is not
required to drive
FERR_GART
GARTINV
_BINITE
XINTR#)
6-18
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
Table 6-1. Error Cases (Cont’d)
Chip
Error
System
Action
Status
Register
Log
Qualifier
Register
Detecting
GART Parity Error
GXB
Continue, use address
as read from GART,
Unconditional XINTR#,
Conditional XBINIT#.
FERR_GART
Nothing
GARTER
R_BINITE
(NOTE: if XBINIT# is
driven, then it is not
required to drive
XINTR#)
Illegal SMM
Access
GXB
Unconditional XBINIT#
FERR_GART
Nothing
Illegal Address
GXB
GXB
Unconditional XBINIT#
FERR_GART
FERR_PCI
Nothing
Nothing
Illegal OB GART
Access
Unconditional XINTR#,
Conditional XBINIT#,
results undefined.
GARTER
R_BINITE
PXB Errors
Detected at
Expander Port
Expander Par-err
on any Header
PXB
PXB
Set Status. If outbound
PCISTS [SSE],
Nothing
Nothing
MODES
[3]
error handling is enabled ERRSTS[2]
then assert BINIT#, else
then SERR#, drop
request (no cmplt
returned).
Expander Par-err
on Write Data from
Expander
Set Status. If outbound
error handling in enabled ERRSTS[2]
then poison data as
PCISTS [SSE],
MODES
[3]
passed to PCI, else then
SERR# and write out
data as good to PCI.
Expander HF Read PXB
Cmplt. from SAC
PXB: Set Status. Target
abort read to card.
(Received in peer-to-
peer)
PCISTS [STA]
ERRSTS[2],
Nothing
Nothing
Expander Par-err
on IB Read Data
PXB
Set Status. If outbound
error handling is enabled PCISTS[SSE]
then poison data as
MODES
[3]
passed to PCI, else then
SERR# and pass read
data as good to PCI.
Detected as PCI Master
PCI Par-err on OB
Read Data
Received from
Card
PXB
Set Status. Drive
ERRSTS[5],PCISTS[PE], Nothing
PCISTS[DPE]
PERR#. Pass data with
good parity to Expander.
(PERR# is optionally
elevated to SERR#)
Master Abort on
Read Done by PXB
PXB
PXB
PXB
Return all 1’s
PCISTS [RMA]
PCISTS [RMA]
Nothing
Nothing
Nothing
Master Abort on
Write Done by PXB
Drop data, normal CMP
Target Abort
received by xXB
Return HF to either read PCISTS [RTA]
or write. If failed access
is in middle of
transaction then the
remainder of the
transaction is discarded.
PERR# Asserted
by Card
card
Optionally turned to
SERR# by PXB.
ERRSTS[6],
PCISTS[DPE]
Nothing
Intel® 460GX Chipset Software Developer’s Manual
6-19
Data Integrity and Error Handling
Table 6-1. Error Cases (Cont’d)
Chip
Error
System
Action
Status
Register
Log
Register
Qualifier
Detecting
Detected as PCI Target
PCI Par-err on
Address from Card
PXB
Accept address as sent
and process as if parity
were good.
PCISTS[PE], PCISTS
[SSE], ERRSTS[2]
Nothing
SERRE
Conditionally assert
SERR#.
PCI Par-err on
Data for an IB Write
PXB
card
Set Status. Drive
PERR#. Pass data with
good parity to Expander.
ERRSTS [5]
Nothing
Nothing
PERR# Asserted
by Card
Optionally turned to
SERR# by PXB.
ERRSTS[6],
PCISTS[DPE]
6.11
PCI Integrity
The PCI bus provides a single even-parity bit (PAR) that covers the AD[31:0] and C/BE#[3:0]
lines. The agent that drives the AD[31:0] lines is responsible for driving PAR. Any undefined
signals must still be driven to a valid logic level and included in the parity calculation.
6.11.1
PCI Bus Monitoring
When the PXB is not the PCI bus master, it will still check the address parity for all PCI address
phases. If a parity error is detected and the Parity Error Response is enabled in the PCI Command
(PCICMD) register, the event is logged in the Parity Error Detected bit in the PCI Status
(PCISTS) register. If the PCICMD register’s SERR# Enable bit (SERRE) is set, SERR# will
also be asserted. When the PXB asserts SERR#, the PCISTS register’s Signalled System Error bit
(SSE) is also set.
Note that the expected settings for the GX system are: SERRE=1, PERRE=1, ERRCMD[3]=1, and
ERRCMD[4]=1. ERRCMD[6:5] are to be programmed as each OEM, operating system or system
user desires in order to enable SERR reporting of data parity errors. For data which has bad parity
and is received from the PCI bus, the PXB will forward this up the Expander bus as if there were no
parity error, so that ERRCMD[5] may need to be set for sufficient error containment, since PERR#
asserted to the card will not prevent the data which had a parity error from being placed in memory
6.11.2
PXB as Master
6.11.2.1
Master Abort
If the PXB initiates a PCI transaction and no target responds, the PXB will terminate the
transaction with a master-abort. The PXB will wait five PCI clocks after asserting FRAME# for a
target to respond with DEVSEL#. If no target responds, the PXB will perform a master abort to
terminate the cycle on the PCI bus. Special Cycle commands, which are broadcast to all PCI
targets, will always be terminated with master abort. Therefore master-aborts during Special Cycle
commands are not considered errors, and are never logged or reported.
When the PXB performs a master abort, if the command was not a Special cycle, it will log the
event by setting the PCISTS register’s Received Master Abort (RMA) bit. The PXB then has two
options for generating a response.
6-20
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
The default option is to return a “normal” response. If the aborted transaction was a read, the PXB
will return all 1’s for the data. If the aborted transaction was a write, the PXB will discard the write
data. SERR# is not asserted in either case.
6.11.2.2
6.11.2.3
Received Target Disconnect
A PCI target may issue a disconnect to indicate it is unable to respond within the PCI latency
guidelines. Disconnect is signalled when the target asserts both STOP# and DEVSEL#. The target
controls whether another data transfer may occur by whether TRDY# is asserted when STOP# is
asserted. A target disconnect is not usually issued on the first data phase of the transaction. Target
disconnects are not considered errors, and are not logged or reported in any way. After a target
disconnect, the PXB will wait at least two PCI clocks before re-arbitrating for the PCI bus to
complete the transfer.
Received Target Retry
A PCI target may issue a retry to indicate that it is currently unable to process the transaction. Retry
is signalled when the target asserts STOP# and DEVSEL# and does not assert TRDY#. Retry is
actually a special case of disconnect that occurs before the first data transfer.
After receiving a retry for a transaction, the PXB will wait at least two PCI clocks before re-
arbitrating for the PCI bus to retry the transaction. If the transaction is a write, the PXB will retry
the transaction until it succeeds. If the transaction is a read, the PXB will retry the transaction until
it succeeds, but may allow writes to pass it. Note, in all of these cases the retries are not considered
errors. There is no logging or error reporting of any kind.
6.11.2.4
Received Target Abort
A PCI target may issue an abort to indicate that the current transaction should be terminated and
should not be attempted again. This is a catastrophic failure. Target abort is signalled when
STOP# is asserted and DEVSEL# is deasserted. The PXB will log the target abort by setting the
PCISTS register’s Received Target Abort (RTA) bit. The PXB then returns a hard fail response to
the SAC.
6.11.2.5
6.11.2.6
Data Parity Errors
When the PXB is the PCI bus master, it will check the data parity provided during read data cycles,
and watch for the assertion of PERR# during write data cycles. See the earlier tables for exact
details.
Other Violations
The PCI specification identifies numerous cases that are violations of the PCI protocol. Other than
the cases identified above, the PXB makes no attempt to check for such violations. Response to
such violations is undefined. Refer to the PCI specification for a complete description of the
required PCI protocol.
6.11.3
PXB as Target
6.11.3.1
Target Disconnect
The PXB will issue a target disconnect under the following circumstances:
Intel® 460GX Chipset Software Developer’s Manual
6-21
Data Integrity and Error Handling
• After the first data transfer if the transaction is using an unrecognized addressing mode (the
PXB will only support linear incrementing as a target),
• On reads, when no more data is available in the read buffers, and
• On writes, when the write crosses a 4 KB boundary.
These conditions are not treated as an error, and will not be logged or reported.
6.11.3.2
Target Retry
The PXB will issue a target retry when:
• A read request is to an address that has already been accepted as a delayed transaction (i.e. the
request is already being serviced, but data has not arrived),
• A write request has insufficient buffering in the PXB to allow it to be posted (a full line is not
available for an MWI), or
• The PCI interface is LOCKED from the host side.
6.11.3.3
6.11.3.4
6.11.3.5
Target Abort
The PXB will issue a target abort if a hard fail response is returned over the Expander bus. This
response is limited to inbound read requests.
Data Parity Errors
When the PXB is a PCI bus target, it will check the data parity provided during write data cycles.
For exact details on data errors, see the earlier tables.
Other Violations
The PCI specification identifies numerous cases that are violations of the PCI protocol. Other than
the cases identified above, the PXB makes no attempt to check for such violations. Response to
such violations is undefined. This includes, but is not limited to:
• MWI to a misaligned (non-cache-line-boundary) address.
• MWI to an aligned address, but with one or more byte enables not asserted.
Refer to the PCI specification for a complete description of the required PCI protocol.
Note: When multiple errors which cause an SERR# assertion occur within a few cycles of each other,
there may not be a separate SERR# assertion for each error.
6.11.4
GXB Error Flow
Figure 6-3 shows a block diagram of the GXB. On the PCI side, the first error is latched in
FERR_PCI. Subsequent errors are latched into NERR_PCI. Errors from PCI will cause the PCI
address and data to be latched in PAC_ERR and PD_ERR.
The internal buffers of the GXB are parity protected. Data coming from the graphics card is placed
in an inbound buffer with parity generated as for the Expander bus (1 parity bit over 16 bits of data
and 2 byte enables). At the top of the inbound queue, the data, BE and parity is sent to the
Expander bus as-is out of the queue. There is no checking done in the GXB. If there is bad parity at
the SAC, the SAC will flag the error.
6-22
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
6.11.4.1
GXB Error Signals
The GXB has 2 error signals: XBINIT# and XINTR#.
6.11.4.1.1 GXB_XBINIT#
XBINIT# is used to signal a fatal error. All header errors are fatal, since the GXB and SAC are out
of sync with each other at that point. Data parity errors may be considered fatal in some systems.
For graphics, the error may be in a texture or some field that is a transient screen blip. The OEM
may configure the system to BINIT# on data parity errors or not.
XBINIT# is held active until XRST# is asserted. On XRST#, GXBCTL[7] (XBINIT# enable) is
reset and thus XBINIT# is deasserted. After reset, firmware will log and clear any pending error
registers. It will then enable GXBCTL[7] so that future errors will be reported. Pending bits in the
FERR or NERR registers that were not cleared by firmware will not cause a new BINIT# when
GXBCTL[7] is enabled. Only errors that occur after GXBCTL[7] is enabled will cause the
XBINIT# pin to be asserted.
6.11.4.1.2 XINTR#
XINTR# is used to signal non-fatal errors on the bus. The SERR# signal on the AGP bus will be
OR’ed with internal GXB signals to create XINTR#. The system designer may take this signal and
use it to force an interrupt, handle it like an NMI or SERR#, or even choose to ignore it.
XINTR# is held asserted until all internal conditions which cause INT# are cleared. So software
must handle all errors and then do an EOI to the interrupt controller.
6.11.4.2
GXB Errors
6.11.4.2.1 PCI Interface Errors
• SERR# Observed - Set when the GXB sees SERR# that was asserted by the graphics card.
This is not set if the GXB drove SERR#.
• PERR# Observed - Set when the GXB sees PERR# that was asserted by the graphics card.
This is not set if the GXB drove PERR#.
• Discard Timer Expiration - Set when the 215 timer expires. The timer starts when the data for a
delayed read is returned to the GXB. If the card doesn’t re-access the data in 215 clocks, then
an error is flagged.
• Non-Configuration Master Abort - Sets bits in both PCISTS and in FERR_PCI. This error
occurs when the GXB is the master for a PCI transaction and there is a master abort (the card
doesn’t assert DEVSEL#). This only occurs on non-configuration cycles, since a master abort
is not considered an error on configuration cycles. This error will not cause an interrupt or
BINIT#. Reads will return 1’s for DW accesses, or a Hard Fail (HF) for reads that are greater
than a DW. Writes of 1 DW will return a normal write-complete. Writes of more than 1 DW
will return a HF on the write-complete. If there is a master abort on the write then the rest of
the data for that transaction is dropped. Writes using the Fast-write protocol may generate a
master abort.
• PCISTS Error Logged - Set when any error bit, except RMA (bit 13), in PCISTS is set. This
includes bits 15, 12 or 8. Setting this bit in FERR_PCI does not cause an interrupt or BINIT#.
Intel® 460GX Chipset Software Developer’s Manual
6-23
Data Integrity and Error Handling
6.11.4.2.2 GART Interface Errors
• GART Parity Error - There is one parity bit covering each GART entry. When the GART is
accessed, parity is checked. If an error occurs, then this bit is set. Parity errors are only
reported when the access falls within the GART aperture range. This prevents errors being
reported when the GART entry was not used.
• GART Entry Invalid - Each GART entry has a valid bit associated with it. If the GART entry
associated with an AGP address is not marked as valid, then this bit is set. This error is only
reported when the access falls within the GART aperture range. This prevents errors being
reported when the GART entry was not used.
• Illegal Address - After translation is done, the address is checked. If it is in the range between
GAPBAS and GAPTOP, or in the VGA range with VGAGE asserted, or directed by the
MARG’s to PCI instead of memory, or above TOM; then the access is illegal and considered a
fault.
• Illegal Outbound GART Access - Any programming access, read or write, to the GART that is
not 4B or is not aligned on a 4B boundary. The behavior is undefined for the error.
6.11.4.2.3 AGP Interface Errors
• Use of Pipe with Sideband Enabled - The card must not mix sideband and Pipe requests. If
sideband is enabled and the card attempts an access with PIPE, this error occurs.
• AGP Address[63:40] not Equal to Zero - This address is outside the legal graphics aperture
and is over the TOM, therefore is an error.
• AGP Request Queue Overflow - The GXB supports 16 AGP transactions. If the card attempts
to do a new transaction when there are 16 already outstanding, this error is flagged.
Illegal AGP Command - Set whenever the GXB receives an unknown or undefined command
from the graphics card.
6.11.4.2.4 Data Errors
• AGP Hi-priority Write Que Data Parity Error - AGP write data was placed in the que with
good parity. If this error is set, then the write que itself was corrupted. The GXB will not report
this error with either interrupt or BINIT#. The error is actually reported by the SAC or allowed
to continue to memory where it will be poisoned.
• AGP Low-priority Write Que Data Parity Error - AGP write data was placed in the que with
good parity. If this error is set, then the write que itself was corrupted. The GXB will not report
this error with either interrupt or BINIT#. The error is actually reported by the SAC or allowed
to continue to memory where it will be poisoned.
• PCI Inbound Write Que Data Parity Error - PCI write data was placed in the que with parity as
received from the PCI bus. This error may have occurred because of bad parity on the PCI bus
or because the que itself was corrupted. The GXB will not report this error with either interrupt
or BINIT#. The error is actually reported by the SAC or allowed to continue to memory where
it will be poisoned.
Note: On data transfers from the card to the GXB, there may be wait states. If there is a parity error on the
AD bus when there is a wait state, then the GXB will not flag any error. It only flags an error when
data is actually captured by the GXB’s PCI interface.
• PCI Outbound Read Que Data Parity Error - PCI read data was placed in the que with parity as
received from the PCI bus. The GXB will not report this error with either interrupt or BINIT#.
The error is actually reported by the SAC or allowed to continue to memory where it will be
poisoned.
6-24
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
• PCI Outbound Write Que Data Parity Error - This error signifies that either a) data was
received from the Expander bus with bad parity or b) the OB write Que was corrupted. As data
is read from the queue and passed to the PCI bus, the parity is checked.
• AGP Low-priority Read Data Que Parity Error - This error signifies that either a) data was
received from the Expander bus with bad parity or b) the read Que was corrupted. As data is
read from the queue and passed to the AGP bus, the parity is checked.
• AGP Hi-priority Read Data Que Parity Error - This error signifies that either a) data was
received from the Expander bus with bad parity or b) the read Que was corrupted. As data is
read from the queue and passed to the AGP bus, the parity is checked.
• PCI Inbound Read Que Data Parity Error - This error signifies that either a) data was received
from the Expander bus with bad parity or b) the read Que was corrupted. As data is read from
the queue and passed to the AGP bus, the parity is checked. This is for PCI reads done by the
graphics card. The parity on the PCI data will be poisoned out to the card as the data is
returned.
6.11.4.3
Multiple Errors
In the case that 2 or more errors occur at the same cycle, multiple bits are set in the FERR register.
This should be an extremely rare case. Software can read the register and check that only one bit is
set. The data that is captured along with the error is indeterminate. Since there are multiple
registers for error data, 2 errors may cause the Expander and PCI error registers to have valid data.
Or, if there were 2 errors from Expander bus at the same time, then the Expander data-error register
may have been set by either error.
Figure 6-3. GXB Error Flow
SAC Interconnect
Check parity.
SAC
UpS
Data
SAC
DnS
Data
On HDR: BINIT#. Drop
PKT and data. Set FERR.
SAC
IB
Pass data to SAC with
parity from queue, don’t
check outgoing parity.
Place data into buffer
with parity as received from
Expander (16 data, 2 BE, 1 par.)
SAC
OB
AGP
HI
AGP
LO
PCI
IB
PCI
OB
Read
Data
AGP
Low
Write
Data
AGP
High
Write
Data
PCI
IB
Write
Chk parity.
SRAM
Data
GART
AGP
High
PCI
IB
AGP
Low
AGP
Low
Read
Data
PCI
IB
Read Write
Data Data
PCI
OB
AGP
High
Read
Data
Generate parity
in Expander format,
data and BE. If
PCI data is bad,
poison into
CFG
Ctrl
PCI
OB
AGP
Que
buffer.
Check Parity.
If PCI, then corrupt outgoing parity.
AGP Bus
Intel® 460GX Chipset Software Developer’s Manual
6-25
Data Integrity and Error Handling
6.12
WXB Data Integrity and Error Handling
6.12.1
Integrity
Error handling in the context of a chipset component requires observing the error, containing it,
notifying the system, and recovery or system restart. Different platforms have different
requirements for error handling. A server is generally most interested in containment. It attempts to
stop bad data before it reaches the network or the disk. On the other hand, workstations with
graphics may be less interested in containment. If the screen blips for one frame and the blip is
gone in the next frame, the error is transient and may not even be noticed.
The WXB accommodates both philosophies by allowing certain errors to be masked off
completely, or by turning them into simple interrupts instead of signalling SERR# or XBINIT#.
Certain errors can only be directed to cause an XBINIT#.
In general the WXB will report errors upon their observation. In the case of data parity errors on
data transiting the WXB, however, the reporting can be deferred until the data is actually used. In
the case of data heading inbound to memory, this could result in deferring the reporting of the error
until it is read from memory or completely if the data is never used.
6.12.2
Data Parity Poisoning
When data is received by the WXB that is in error, it will be passed on to the next interface with
bad (or “poisoned”) parity. Data received over the PCI bus that has bad parity will always be sent
on to the chipset core with bad parity.
Note: There is no mode in the WXB to forward good parity if the data was received as bad. If the data
comes in with bad parity it is always sent out with bad parity.
Note: When addressed to the IHPC, outbound writes containing data with bad parity will be “dropped on
the floor.”
6.12.3
Usage of First Error and Next Error Registers
When an error is identified, and it is the first occurrence of an error, it is latched into the “First
Error” register (see FEPCI register). If an error has been previously identified (and not cleared),
the associated First Error bits will already be set. In such a case, subsequent errors of any type will
be recorded through “Next Error” registers.
Since the system really only needs to know if just one error has occurred or many, setting a First
Error bit does not affect the Next Error bits. If there is another error of any type, including a second
occurrence of the first error, then the appropriate Next Error bit is set. Software can read both First
Error and Next Error registers to determine a complete error status.
In general, as much information as possible is logged for the first error encountered; data, address,
and/or command information are captured if available. This allows better isolation of errors and
possible recovery.
Note: Multiple errors occurring in (nearly) the same cycle may result in multiple bits being set in the First
Error register.
6-26
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
Note: Additionally, error responses such as SERR#, XBINIT# and INTRQ# are predicated on both First
Error and Next Error contents since a second error may occur while the first error is in the process
of being serviced by the WXB hardware. Thus, for example, the system may observe an interrupt
associated with an error at almost the same instant XBINIT# is asserted for another.
For ease of isolation, the ERRSTS register records which of these registers have recorded an error.
6.12.4
Error Mask Bits
Many of the errors that the WXB handles have conditional reporting behavior. The error always
sets the appropriate bit(s) within the First Error or Next Error registers. For a subset of errors
additional information is also recorded in the PCI required PCISTS register. If the error is masked,
then an XBINIT# (or other specified action) will not occur, but the status bit indicating the
occurrence of the error is always set. This allows software to poll if it chooses, looking for errors
that are not considered as fatal to the system. The error logging information (associated only with
First Error) will always represent the first error identified since the register was last cleared
regardless of the masking of the particular error.
Table 6-2. List of WXB Error Sources Selectively Routable to XBINIT#, SERR_OUT#,
and P(A/B)INTRQ#
Abbreviation
Error
APE
PCI address parity error
CUIQ
WXB RAM-generated data parity errors in transactions directed at WXB configuration space
Inbound read discard timer expiration event
DTE
HPIQ
WXB RAM-generated data parity errors in transactions destined to the IHPC
IHPC-generated SERR
HPSERR
OPERR
OSERR
PCIDPE
PUIQ
WXB-observed PERR# assertion
WXB-observed SERR# assertion
PCI bus data parity error
WXB RAM-generated data parity errors in transactions destined to the PCI bus
6.12.5
Error Steering/Signaling
All observed errors are recorded through status bits. In most cases these errors can be (optionally)
caused to signal the system that the error has occurred. Methods for signaling an error include
SERR#, XBINIT# and a P(A/B)INTRQ1 interrupt. In the case of data parity errors, the minimal
response is to record the error and to simply forward the data on across the next interface with bad
signaling outputs. The error routing to SERR_OUT#, XBINIT#, and P(A/B)INTRQ as a function
Table 6-3. Supported Error Escalation to XBINIT#a
XBINIT0
XBINIT#
ERRCMD(15)
Escalation
0
1
CUDPE, CUIQ, HPIQ, HP, IT, HLM, and MXBLK
OFF
a. To obtain the listed escalation, the following settings are required: XBINITE=’0’’.
Intel® 460GX Chipset Software Developer’s Manual
6-27
Data Integrity and Error Handling
Table 6-4. Supported Error Escalation to SERR_OUT#a
ASDPE
ERRCMD(11) ERRCMD(9)
ASFPE
SERRE
PCICMD(8)
PERRE
PCICMD(6)
SERR_OUT#
Escalation
X
0
0
0
X
0
0
1
0
X
HPSERR, OSERR
1
1
1
0
1
0
HPSERR, OSERR, DTE
HPSERR, OSERR, DTE, APE
HPSERR, OSERR, DTE, OPERR, HPPERR,
PUIQ
0
1
1
1
HPSERR, OSERR, DTE, APE, OPERR,
HPPERR, PUIQ
1
1
1
0
0
1
1
1
1
0
1
0
HPSERR, OSERR, DTE, FUIQ, PCIDPE
HPSERR, OSERR, DTE, APE FUIQ, PCIDPE
HPSERR, OSERR, DTE, FUIQ, PCIDPE,
OPERR, HPPERR, PUIQ
1
1
1
1
HPSERR, OSERR, DTE, APE FUIQ, PCIDPE,
OPERR, HPPERR, PUIQ
a. To obtain the listed escalation, the following settings are required: ASAPE=’1’, and ASDTE=’1’.
Table 6-5. Supported Error Escalation to P(A/B)INTRQ#
IRQE
P(A/B)INTRQ#
ERRCMD(13)
0
1
OFF
OPERR, HPPERR, PUIQ
6.12.5.1
SERR# Generation
Most errors can be caused to steer the observation of the error to the signaling of an SERR#. The
system then has a chance to respond to the event while it continues to run. Often, SERR# results in
an NMI which itself often results in a system hang. If the WXB is configured to cause an SERR#
the system should be configured with a usable NMI handler or wired in such a way that an SERR#
is not escalated into an NMI.
Alternatively, the WXB can be configured to cause an INTRQ# interrupt or an XBINIT# instead of
(or in addition to) an SERR#. This is achieved by having any unmasked errors (errors configured
such that they might cause an SERR#) directed to cause the specified event, be it an INTRQ# or
an XBINIT#.
SERR# signaling itself is ultimately controlled by appropriately setting or clearing the SERRE bit
in the PCICMD register. If set then unmasked errors will result in SERR# being signaled.
Otherwise, even unmasked errors will not cause an SERR#.
Whenever the WXB actually signals an SERR# the SSE bit in the PCISTS register will be set. If
another agent has caused an SERR# then the SES bit in the FEPCI (or NEPCI) register will be
set.
Note: When multiple errors, which cause an SERR# assertion, occur within a few cycles of each other,
there may not be a separate SERR# assertion for each.
6-28
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
6.12.5.2
6.12.6
6.12.7
XBINIT# Generation
A certain subset of errors within the WXB will always result in the WXB attempting to signal an
XBINIT#. Whenever an error occurs that forces an XBINIT#, an internal “override” bit is set as
XBINIT# is driven to the bus. While the override bit is set, XBINIT# will no longer be reasserted,
even for qualifying errors. This prevents the presence of a single error from causing XBINIT# to be
signaled multiple times. Software may also set the override bit and thus prevent XBINIT# from
occurring at all. When an XBINIT# is signaled, software should (1) clean up the error, (2) record
and clear the error status registers, and (only) then, (3) clear the XBINIT# override bit so that new
errors may be signaled. The override bit is always set on a Power Good reset.
Anytime an XBINIT# is signaled by the WXB, the “XBINIT Asserted” bit in the ERRSTS register
will be set. Clearing this bit will cause the deassertion of the XBINIT# pin.
Errors that can be caused to signal an SERR# can also be caused to signal an XBINIT# (refer to the
INTRQ# Interrupt
The WXB has an internal mechanism to cause a level interrupt under various error conditions. The
signaling of this INTRQ interrupt is directly visible through the INTRQ# pin.
An INTRQ interrupt would generally be an expected outcome anytime that a “less serious” error
has occurred. The INTRQ# signal is held asserted as long as the record of the error persists. All the
errors that cause such an event are wired together to drive the internal signal. Software is expected
to reset the bit causing the interrupt in the First Error or Next Error register where the error is
recorded. In addition, software is expected to then clear the “INTRQ Asserted” bit in the
ERRSTS register. When this bit is reset, INTRQ# will be deasserted unless there is another error
which holds the signal active.
It is possible to configure the WXB to cause all errors to result in an INTRQ# interrupt1. In this
way there is some flexibility to have a soft response (i.e. interrupt) for all errors as well as a more
harsh response for specific “more serious” errors.
Error Determination and Logging
The various error status registers identify which error(s) have occurred. In certain cases key
information is logged in conjunction with the error status being set. For example, parity errors log
the data and parity for the failing transfer if it is the first error occurrence. Other errors capture the
address associated with the particular failure. This information can be used for debug and
diagnostic purposes and may be used in system recovery. For instance, if there is an parity error on
a data read, and the access can be isolated, instead of restarting the whole system it might be
possible to kill only the failing process and allow other users to continue running.
The error log registers are updated at virtually the same time that the associated bit is set in the
status register.
1. The single exception to this rule is Hard Fail Completion which will not initiate any sideband error signal (INTRQ#, SERR# or XBINIT#).
However, an in-band PCI Target Abort will occur as a result of a Hard Fail Completion.
Intel® 460GX Chipset Software Developer’s Manual
6-29
Data Integrity and Error Handling
6.12.8
Error Conditions
6.12.8.1
WXB as Bus Master
6.12.8.1.1 Master Abort
If the WXB initiates a PCI transaction and no target responds, the WXB will terminate the
transaction with a master-abort. The WXB will wait five PCI clocks after asserting FRAME# for a
target to respond with DEVSEL#. If no target responds, the WXB will perform a master abort to
terminate the cycle on the PCI bus. Special Cycle commands, which are broadcast to all PCI
targets, will always be terminated with master abort and, therefore, are never considered an error.
Master aborts during other transactions (configuration, memory map, or I/O cycles) also are
generally not considered as errors. For writes, a normal completion packet will be returned to the
Chipset core. Reads will additionally return all 1’s as data.
When the WXB performs a master abort, it will log the event by setting the PCISTS register’s
Received Master Abort (RMA) bit, unless the transaction was a Special Cycle.
6.12.8.1.2 Received Target Disconnect
A PCI target may issue a disconnect to indicate it is unable to respond within the PCI latency
guidelines. Disconnect is signaled when the target asserts both STOP# and DEVSEL#. The target
controls whether another data transfer may occur by whether TRDY# is asserted when STOP# is
asserted. A target disconnect is not usually issued on the first data phase of the transaction. Target
disconnects are not considered errors, and are not logged or reported in any way. After a target
disconnect, the WXB will deassert its request signal and wait at least two PCI clocks before re-
arbitrating for the PCI bus to complete the transfer.
6.12.8.1.3 Received Target Retry
A PCI target may issue a retry to indicate that it is currently unable to process the transaction. Retry
is signaled when the target asserts STOP# and DEVSEL# and does not assert TRDY#. Retry is
actually a special case of disconnect that occurs before the first data transfer.
After receiving a retry for a transaction, the WXB will deassert its request line and wait at least two
PCI clocks before re-arbitrating for the PCI bus to retry the transaction. If the transaction is a write,
the WXB will retry the transaction until it succeeds. If the transaction is a read, the WXB will
reattempt the transaction until it succeeds, but may allow other reads and writes to pass it. Note, in
all of these cases the retries are not considered errors. There is no logging or error reporting of any
kind.
6.12.8.1.4 Received Target Abort
A PCI target may issue an abort to indicate that the current transaction should be terminated and
should not be attempted again. This is a catastrophic failure. Target abort is signaled when STOP#
is asserted and DEVSEL# is deasserted. The WXB will log the target abort by setting the PCISTS
register’s Received Target Abort (RTA) bit. The WXB then returns a hard fail response to the
Chipset core.
6.12.8.1.5 Data Parity Errors
When the WXB is the PCI bus master, it will check the data parity provided during read data
cycles, and watch for the assertion of PERR# during write data cycles. If a parity error is detected,
and the PCICMD registers PERRE bit is set, then the PCISTS register’s Detected Parity Error
6-30
Intel® 460GX Chipset Software Developer’s Manual
Data Integrity and Error Handling
(DPE) bit is asserted. Regardless, if the transaction is a read, the PCISTS register’s Parity Error
(PE) bit will be set. Additionally, address, command, and data related information is logged in the
FEPCIAL and FEPCIL registers if the error is the first error observed by the WXB.
6.12.8.1.6 Other Violations
The PCI specification identifies numerous cases that are violations of the PCI protocol. Other than
the cases identified above, the WXB makes no attempt to check for such violations. Response to
such violations is undefined. Refer to the PCI specification for a complete description of the
required PCI protocol.
6.12.8.2
WXB as Target
6.12.8.2.1 Illegal PCI Request Type
The WXB will not claim transactions that use an illegal or unrecognized request type.
6.12.8.2.2 Target Disconnect
The WXB will issue a target disconnect under the following circumstances:
• After the first data transfer if the transaction is using an unrecognized addressing mode (the
WXB will only support linear incrementing as a target),
• On reads, when no more data is available in the read buffers, and
These conditions are not treated as an error, and will not be logged or reported.
6.12.8.2.3 Target Retry
The WXB will issue a target retry when:
• A read request is to an address that has already been accepted as a delayed transaction (i.e. the
request is already being serviced, but data has not arrived).
• A read request to this address has not yet been accepted by the WXB as a delayed transaction,
there is room to enqueue a new delayed transaction, and the request is enqueued.
• A read request is to an address that has not yet been accepted by the WXB and there is no more
room to enqueue a new delayed transaction.
• A write request has insufficient buffering in the WXB to allow it to be posted (e.g. a full line is
not available for MWI).
• The PCI interface is LOCKED from the host side, unless the transaction is a read request and
the data has already been fetched by the WXB.
6.12.8.2.4 Target Abort
The WXB will issue a target abort if a hard fail response is returned over the Expander bus. This
response is limited to inbound read requests.
6.12.8.2.5 Other Violations
The PCI specification identifies numerous cases that are violations of the PCI protocol. Other than
the cases identified above, the WXB makes no attempt to check for such violations. Response to
such violations is undefined. This includes, but is not limited to:
Intel® 460GX Chipset Software Developer’s Manual
6-31
Data Integrity and Error Handling
• MWI to a misaligned (non-cache-line-boundary) address
• MWI to an aligned address, but with one or more byte enables not asserted
Refer to the PCI specification for a complete description of the required PCI protocol.
6.12.8.3
PCI Interface Errors
Other PCI interface errors that are handled by the WXB are:
• System Error Signaled
Set within the FEPCI register when the WXB sees an SERR# asserted by another PCI agent.
This is not set when the WXB drives SERR#.
• Discard Timer Expiration
Set when the 215 timer expires. The timer starts approximately when the data for a delayed
read is requested by the WXB. If the card doesn’t re-access the data in 215 PCI clocks, then an
error is flagged.
6-32
Intel® 460GX Chipset Software Developer’s Manual
AGP Subsystem
7
AGP is a new port defined for graphics adapters. In the initial implementation it is a 500 MB/s port.
There is also an extension called AGP 4X mode, which has a bandwidth of 1 GB/s. AGP 2X mode
cards will work in an AGP 4X mode slot. The 460GX chipset is designed to work at the AGP 4X
mode bandwidths. It will support 3.3V AGP 1X and 2X mode cards as well.
AGP 4X mode is a high-bandwidth port targeted to workstation graphics needs. It provides 1 GB of
data bandwidth. There are several protocol modes. One mode is 66 MHz PCI. The other mode is
the new AGP protocol, which comes in two flavors. The first is an extension of PCI using PIPE#
instead of FRAME#. The second uses sideband signals to improve bandwidth. Only one AGP
protocol may be in effect at a time. Both of the modes using AGP protocol and the mode using PCI
protocol are fully supported.
This chapter explains the 460GX chipset implementation of AGP and related issues. It assumes
that the reader is familiar with the AGP and AGP 4X mode specifications. This chapter explains
the GX implementation of the AGP interface to the GXB and the interface between the GXB and
the rest of the chipset, which is the expansion bridge chip containing the AGP port. It will not
explain AGP protocol or the AGP bus itself.
7.1
Graphics Address Relocation Table (GART)
Graphics cards, especially for texture traffic, would like to see a large contiguous view of memory.
This allows them to make random accesses into a large area which contains the entire texture map
for the displayed image. Processors have the same view of a flat contiguous memory range in their
virtual space, where a program may have many megabytes of contiguous address space to use. The
memory pointed to by those virtual addresses will most likely be discontiguous in main memory.
The processor, through page tables and TLBs, remaps the contiguous virtual addresses to the
discontiguous physical DRAM addresses. For graphics cards, the chipset will perform the same
translation. The graphics card uses, in essence, a virtual address. This address is translated by the
chipset’s Graphics Address Relocation Table (GART) from the virtual address used by the graphics
card to the physical DRAM address the memory controller understands.
The GART table is loaded by the graphics device driver at initialization time. The table resides in
the GXB, so that translations will be fast and not interfere with main memory traffic. The table will
contain a new page address for each address coming in from the graphics card. It will also contain
a bit to determine whether addresses to that page need to be snooped on the system bus or not. The
following figures show how the GART operates. There is a base address for AGP addresses and a
length for the AGP range. Addresses are received from the graphics card. If the address is in the
AGP range, the page number within that range is used as a pointer into the GART and a new page
address is combined with the 12 (22 for 4 MB pages) least significant bits of the address from the
graphics card. This new address is used by the memory system. If the address doesn’t hit inside the
AGP-space, then no translation is done. Note that the AGP_BASE address bits 27:12 are always a
0 (hardwired in config register).
Intel® 460GX Chipset Software Developer’s Manual
7-1
AGP Subsystem
Figure 7-1. GART Table Usage for 4k Pages
39
11
0
AGP address
Offset
12b
AGP address (39:12) - APBASE(39:12)
(If less than APBASE+Aperture)
24b
24 Bit
+
18b
GART Entry
(16b if256 MB
of GART space)
GART Table
36b Main Memory Address
Figure 7-2. GART Table Usage for 4 MB Pages
39
21
0
AGP address
Offset
22b
AGP address (39:22) - APBASE(39:22)
(If less than APBASE+Aperture)
14b
14 Bit
+
13b
GART Entry
(
8b if 1 GB GART,
6b if 256 MB GART
)
GART Table
36b Main Memory Address
7-2
Intel® 460GX Chipset Software Developer’s Manual
AGP Subsystem
7.1.1
GART Implementation
The GX implementation will support 256 MB, 1 GB, or 32 GB (32 GB requires 4 MB pages by the
O.S.) of translation space. This limit is implementation-based not architectural. Each entry in the
GART requires 24 bits for the address, 1 bit for coherency, 1 parity bit and 1 valid bit for a total of
27 bits. The width of each GART entry will be 32 bits. Each entry covers one 4kB (or 4 MB) page.
One megabyte is sufficient to cover 256k entries or a total of 1 GB of translatable space for 4kB
entry that does not have the valid bit set are treated as GART misses and the address is passed on
untranslated. Optionally, accesses that hit an entry with the valid bit marked invalid can cause an
error.
Figure 7-3. GART Entry Format for 4kB Pages
1bit
1bit
5bits
Reserved
1bit
24bits
Coherency Bit
1 = Coherent
0 = Non-coherent
Valid
1=valid
0=invalid
Parity
New Page
address
Must be 0
Figure 7-4. GART Entry Format for 4 MB Pages
1bit
1bit
5bits
Reserved
1bit
14bits
10bits
Coherency Bit
1 = Coherent
0 = Non-coherent
New Page Rsvd.
Valid
1=valid
0=invalid
Parity
Must
be 0
Must be 0
address
The table itself will be kept in SRAM attached to the GXB. It is possible that the table could be in
main memory and the GXB could fetch the translation from memory when it needs a new entry.
Having the GART in local SRAM allows the translations to be done much faster and is a simpler
implementation, since the interface is local and accesses to fetch GART entries don’t compete with
other traffic. A fetch from local SRAM should take 30 ns. or less instead of the roughly 300-500
ns. required to get to DRAM from the GXB.
Both 1 MB and 4 MB SRAMs will be used. To get the 1 MB required, 2 of these parts are required.
For systems that don’t require the full 1 GB translation space, the SRAM scales with the required
space. If 256 MB of translated graphics area were sufficient, then the system could simply populate
256k of SRAM. The GX will support the following sizes of AGP space and SRAM sizes.
• 1 MB of SRAM - 1 GB of AGP space for 4kB pages, requires two 4 MB SRAMs; using 4 MB
pages, the AGP space is still limited to 32 GB.
• 256kB of SRAM - 256 MB of AGP space for 4kB pages, requires two 1MB SRAMs; using 4
MB pages, the AGP space is still limited to 32 GB.
There is no requirement that there be any GART table. If the graphics card doesn’t require
translation, then there is no need of a GART and the SRAM need not be present at all. The GXB
will fully work when there is no SRAM. Of course, in this case, there is no translation so the AGP
card must put out the physical memory address itself.
Intel® 460GX Chipset Software Developer’s Manual
7-3
AGP Subsystem
7.1.1.1
Page Sizes
The Itanium processor supports both a 4kB and a 4 MB page size. The AGP programming model is
designed using 4kB pages for GART entries. Using the larger page size would greatly reduce the
number of misses and reduce the number of entries needed in the GART, thus lowering system
cost.
The GXB will support 4 MB pages. Using 4 MB pages, 256kB of SRAM could provide 256 GB of
AGP space. The 460GX chipset will not support more than 32 GB of translatable graphics space.
7.1.1.2
7.1.1.3
GTLB
The GXB does not implement any GTLB or hold any old translations. Each address received by the
GXB will be translated through the SRAM.
Parity
There is one parity bit which covers each GART entry. The parity covers the entire GART entry,
including reserved bits. The parity bit is generated by hardware. Software does not need to
compute parity for the GART entry. Software should mask out the parity bit and treat it as a
reserved bit on GART reads. The parity bit is returned on a read, so that it can be read by
diagnostics. Writing a zero or a one to the parity bit has no affect since the hardware will compute
correct parity and write that into the SRAM, which may cause the bit to be read as different from
what was written. Parity will be done such that the total number of 1’s, including the parity bit, total
to an even number.
7.1.2
Programming GART
The addresses to load or read the GART itself lie in the chipset specific area below 4 GB. There is
a 12 MB area for the chipset to use. The GART will exist in this range.
The GART lies at address range FE20_0000h to FE3F_FFFFh. GART entries may be read or
written from the processor. Accesses by the processor must be 4 byte accesses on an aligned 4B
boundary. Any other type of access is considered an error and the system will fault. For the Dual-
Expander bus GXB which attaches to Expander ports 2 and 3, it will be programmed through port
2 addresses only.
To the GXB, address bits 22 and 21 are don’t cares and shouldn’t be used in the decode for any
access at FEzx_xxxxh, where z is 0xx0 (binary). Address bit 23 must be a 0 for the address to point
to the GART. Address bit 20 must be a 0 also. The range FE30_0000h to FE3F_FFFFh is not
decoded by the GXB as a GART access, but is present in the address map for future growth.
When programming the GART, software must read the last location written, to guarantee that all
writes have taken affect. All writes to the GART table are done in order, so only the last write
address needs to be read. This ensures the GART table is updated in SRAM before allowing the
graphics cards to try and access translatable space. The processor must map the area from (4 GB-32
MB) to (4 GB-20 MB) as UC.
7-4
Intel® 460GX Chipset Software Developer’s Manual
AGP Subsystem
7.1.3
GART Implementation
Figure 7-5 shows the timings for the SRAM interface. Synchronous SRAM will be clocked at 7.5
ns. The SRAM will be used in the pipelined mode. This allows addresses to be presented to the
SRAM every cycle. The data is valid to be latched by the GXB on the 3rd clock edge after the
clock edge which drove the address. Writes require the data to be driven with the address. The
entire 32 bits of data are read or written at one time.
Figure 7-5. GART SRAM Timings
0ns
25ns
2
50ns
6
75ns
100ns
1
3
4
5
7
8
9
10 11 12 13 14 15 16
K
SA(17:0)
SE3#
Rd1
Rd2 Rd3 Rd4
WrA
WrB WrC WrD
SGW#
G#
DQ(31:0)
Q(1)
Q(2) Q(3) Q(4)
D(A)
D(B) D(C)
D(D)
SE1#
SE2
ADSC#
ADSP#
ADV#
LBO#
SB#
SW#
NOTES:
1. LBO# is a don’t care, but must be tied either high or low.
2. SE1#, SE2, ADSC#, ADSP#, ADV#, LBO#, SB# and SW# are not connected to the GXB.
These signals must be tied to the required level by the board.
7.1.4
Coherency
Traffic from the graphics card may or may not want to be coherent with the system bus. For the
discussion here, coherency means that addresses will appear on the system bus so that the
processors may snoop their caches for that address. If the texture map or other image data is
marked WC by the processor, then that data is not coherent. Addresses on the bus which hit in a
processor’s WC buffer are not snooped and even if there is dirty data in the buffer, no implicit
writeback is done. Or the application may know that the data in memory was not used by the
processor (e.g. it came from disk) and wants the graphics card to fetch the data without using
address bus bandwidth, so forces the access to be non-coherent.
Intel® 460GX Chipset Software Developer’s Manual
7-5
AGP Subsystem
For all AGP-type accesses which hit in the AGP range, there is a bit per GART entry which
determines whether the address is coherent. For AGP-type accesses outside the AGP range, there is
a bit in a configuration register of the GXB which determines the coherency. Coherency or non-
coherency applies to accesses using AGP protocol only. Accesses using PCI protocol are always
done coherently, whether they hit the translation table or not.
Accesses using AGP protocol which are non-coherent are truly not coherent. There is no ordering
between them and processor traffic at all. The traffic goes directly to memory regardless of
processor cache state or bus locks. However accesses within the AGP stream must follow the AGP
ordering rules regardless of whether or not they are coherent. Coherent low-priority writes will
push non-coherent low-priority writes, as the AGP spec calls for. Coherent traffic is not a separate
stream from non-coherent traffic within the GXB.
7.1.4.1
3.3V AGP 1X and 2X Mode Compatibility
The GXB will be compatible with 3.3V AGP 1X and 2X mode cards. If they do accesses of less
than 32 bytes, then their performance may not be adequate. As well, some AGP cards might expect
lower latencies than the 460GX chipset will guarantee. These cards must not have a watchdog
timer or other time-out mechanism which requires a latency lower than the 460GX chipset is able
to guarantee.
7.1.5
Interrupt Handling
Before an interrupt is delivered to the processor, the system must ensure visibility or completion of
all operations from that device which were generated before that interrupt. For PCI, device drivers
are required to read a status register on the card that caused the interrupt. If the card had done a
write before the interrupt, then the read completion would force the write to have been seen by the
system.
AGP, with its hi-priority, low-priority and PCI streams, has no concept of ordering between the
streams. Therefore reading an AGP cards register will not enforce pending hi or low-priority writes
to have been completed. If the graphics card needs to force all writes to be visible to the system
before the interrupt is visible, then the AGP card must issue a flush operation to guarantee that all
writes are complete, wait for the return of the flush acknowledgment, and then issue the interrupt.
Even using Itanium processor’s SAPIC interrupt protocol, whereby a write on the PCI bus
becomes an interrupt to the processor, does not guarantee that writes from other streams have
become visible to the system by the time the interrupt is received.
The interrupt signals on the AGP 4X mode interface will be routed to an I/O APIC chip (PID).
The GXB will not receive these signals. The GXB will have an interrupt pin that it will drive for
error reporting.
7.2
AGP Traffic
7.2.1
Addresses Used by the Graphics Card
AGP introduced the concept of a contiguous virtual address range that the graphics card could use.
This address range lies outside the physical memory space. It also lies outside the range of
addresses mapped to I/O. This range may exist in one of two places in the 460GX chipset system
map. The virtual address is limited to 40 bits for the GXB, even when in 64-bit addressing mode.
7-6
Intel® 460GX Chipset Software Developer’s Manual
AGP Subsystem
The range may lie above the top of physical memory. Or the range may be placed in one of the gaps
used to map addresses to PCI, and have that gap marked as reserved and not usable for addressing
PCI devices.
In the first case, the virtual range used by the graphics card may be above or below the 4 GB
boundary. If it is above, then it can be placed anywhere in the 40 bit address range supported by the
GXB. This requires dual-address cycles (DAC) on PCI-type accesses or for AGP-type accesses
which use the sideband signals or the DAC command with PIPE#. Having the virtual range above
the top of memory means that s/w can allocate the full GART space per GXB at boot time and not
worry about it again. The address must be translated by the GART to a 36-bit address, since only
36 physical bits are supported after the GART. In this case, there is a special BAR that is used by
the firmware to know how much space to allocate for AGP. This special BAR is the BAPBASE
register in the GXB and is exactly the same as the standard PCI-defined BAR, except that it is not
in the PCI-defined Header space.
In the second case, one can map out the required region by using the standard BAR configuration
register in the SAC. Firmware will read the BAR on the AGP card itself and give it the address
space it needs. It will then read the BAR on the GXB to see how much AGP space is being
mapped. It will then use the PCIS register for Device 14h to allocate this space. The PCIS registers
for Devices 15h, 16h, and 17h will all be set to that of 14h, which in effect disables 15-17h.
The graphics card can issue addresses to the non-AGP space using AGP protocol. The addresses it
puts out are the physical memory addresses, the same as PCI cards or the processor does today.
The physical memory used for graphics may change during run-time. One application may need
only a few megabytes of AGP space and would only get that many bytes of physical space from the
o.s. It would only program the few entries in the GART that it needed. A later application might ask
for hundreds of MBs of physical memory and then map more entries in the GART. When that
application finished it could release the physical memory back to the O.S. In all cases the virtual
addresses used are allocated at boot/config time. They do not change during runtime, only the
mapping of the physical memory they point to changes.
7.2.2
Traffic Priority
The AGP specification has both hi-priority and low-priority AGP transfers as well as PCI transfers.
The GXB will do hi-priority requests ahead of any low-priority or PCI requests. The SAC will not
prioritize any AGP traffic over other system traffic. So, except for the GXB, the chipset has no
concept of traffic priority. Since the chipset’s intent is to provide the best service to all agents,
trying to give one agent priority does not fit well. Since memory is highly interleaved, having the
GXB move hi-priority traffic ahead of low-priority traffic should be sufficient to provide the
latency needed by the traffic.
Since the processor and PCI traffic must appear on the system bus, a new address may appear every
3 clocks from any coherent agent. The other 2 clocks are used for moving non-coherent traffic to
the memory controller. Thus the processor can never hold off AGP traffic.
7.2.3
Coherency, Translation and Types of AGP Traffic
Table 7-1 shows the types of traffic versus address range and the resulting behavior. AGP-space
means an address that hits in the translation window. DRAM-space is an address which is a
physical DRAM location. PCI-space is an address that would normally be directed to a PXB and a
PCI device. Both the DRAM-space and the PCI-space lie outside the AGP-space. These areas must
not overlap. The destination of the request is determined by address registers in configuration
space.
Intel® 460GX Chipset Software Developer’s Manual
7-7
AGP Subsystem
Note: Accesses from an AGP card that are directed to a PCI bus are a system fault and cause a BINIT#
(system reboot). The 460GX chipset does NOT support any access originating from the AGP port
to another PCI bus. This is true for PCI cycles (FRAME# active) as well as AGP cycles.
PCI accesses are always disconnected at a line boundary for writes and a read will at most prefetch
to the end of the line. Therefore each PCI request will have only one translation.
An AGP request may be more than a line in length and therefore could cross a page-boundary. In
that case the request may be broken into 2 parts, each with its own translation and targeted location
(DRAM-space or AGP-space). Each part of the request is handled individually and then the data
recombined before delivering back to the graphics device.
Table 7-1. Coherency for AGP/PCI Streams
Transaction Type
Traffic Type Address Range
System Effect
System Action
Translate
Coherent
PCI
AGP-space
DRAM-space
PCI-space
Yes
No
No
Always
Always
Translated access to memory
Untranslated access to memory
(Frame
Active)
Not
applicable
Peer-to-peer PCI access (THIS IS ILLEGAL
AND CAUSES SYSTEM BINIT#)
AGP
AGP-space
Yes
No
GART
Sets.
Translated access to memory
(Pipe# or SB)
DRAM-space
Depends
on Config
bit
Untranslated access to memory
PCI-space
No
Not
applicable
Peer-to-peer PCI access (THIS IS ILLEGAL
AND CAUSES SYSTEM BINIT#)
7.2.4
7.2.5
Ordering Rules
The GX will obey AGP ordering rules as spelled out in the AGP specification. The coherent stream
will follow the ordering rules for the type of transaction it is. Low-priority coherent traffic is
treated as being in the same stream as low-priority non-coherent traffic.
Processor Locks and AGP Traffic
AGP non-coherent traffic is not affected by locks on the system bus that target memory or other I/
O devices. The accesses continue to memory independent of whether the LOCK# bit is set on the
bus. However, since legacy code may issue a locked transaction to an AGP device, the GXB must
information on processor locks). The lock flow for a single Read-Modify-Write to the AGP bus is
outlined below.
1. An outbound Locked Read request is transferred across the Expander bus.
2. When the locked read request reaches the head of the outbound PCI request queue, it is issued
on the AGP bus with PCI semantics.
3. Once the locked read transaction is complete the read completion is placed in the inbound PCI
transaction queue. At this point the PCI interface establishes a pseudo-lock and begin retrying
all inbound transactions.
4. When the Read completion reaches the head of the PCI inbound PCI transaction queue,
assuming the higher priority queues are empty, it is transferred across the Expander ring. At
7-8
Intel® 460GX Chipset Software Developer’s Manual
AGP Subsystem
this point the inbound Expander logic establishes a pseudo-lock and will no-longer send
coherent requests from the AGP streams. Non-coherent requests can still be issued, but
anything that can block the PCI stream in the SAC’s queues must be held in the GXB.
5. The outbound Locked Write request is transferred across the Expander bus.
6. An Unlock transactions is transferred across the Expander bus (locked writes are posted; the
Unlock will occur once the system bus lock goes away, so this will occur even if the data has
not been delivered to its destination). The Unlock request, which throughout the Expander
queues has the same semantics as a posted write, must be pushed into the outbound PCI queue
behind the Locked Write.
7. When the Locked Write reaches the head of the outbound PCI queue it is performed on the
AGP Bus with PCI semantics.
8. When the write completes on the AGP bus, a Write Completion is pushed into the inbound PCI
transaction queue.
9. The Unlock transaction reaches the head of the outbound PCI queue.
10. A write completion for the Unlock transaction is placed in the inbound PCI transaction queue.
At this point the PCI interface’s pseudo-lock is released.
11. Both the completion for the locked write and for the Unlock transaction are delivered over the
Expander bus. When the Expander logic transmits the completion for the Unlock transaction,
its pseudo-lock is released.
7.2.6
Address Alignment and Transfer Sizes
The AGP specification allows the graphics card to request reads of size 8-64 bytes in eight-byte
increments or of size 32 to 256 bytes in 32-byte increments. Reads and writes are both aligned on
any 8-byte boundary. There is no concept of cache-line aligned or even page aligned. The request
may cross a page boundary and require 2 translations. There is no requirement that the physical
addresses resulting from the translation be contiguous. There is also no requirement that the
coherency be the same on these two pages.
7.2.6.1
Address Faults
See the ‘ERROR’ chapter for the GXB behavior on addresses that are considered illegal.
The graphics card may do AGP transfers to areas outside the GART range. There is no translation
on these addresses, so the graphics card must use the physical DRAM address for the object it is
attempting to access. This is perfectly acceptable, but the graphics card and device driver must be
able to provide the correct physical address to be used.
7.2.7
PCI Semantics Traffic
Inbound Reads
7.2.7.1
Delayed Transactions
The GXB supports the Delayed Transaction mechanism as defined in the PCI 2.2 Specification.
The process of latching all information required to complete the transaction, terminating with
Retry, and completing the request without holding the master in wait-states is called a Delayed
Transaction. The GXB will delay all memory space read requests (unless a delayed slot is
unavailable); no other request types are delayed.
Intel® 460GX Chipset Software Developer’s Manual
7-9
AGP Subsystem
Delayed transactions are issued and serviced as follows:
1. Upon receiving a read request, the address is compared against the GXB’s internal buffers.
Unless the data corresponding to this request is already available in the buffers (i.e. from a
previously retried request), the read cycle is immediately retried (the GXB retries the read
cycle in three PCI clocks from FRAME# driven active).
2. If there is an available Delayed Transaction Request buffer (there is only one per GXB), then
the GXB will latch the pertinent information and forward the read request.
3. When the data is returned, the GXB will store it in its internal buffers, and flag the buffers as
“available” for a subsequent read to that address (i.e. GOTO #1).
The GXB allows only one outstanding Delayed Transaction access. After accepting a Delayed
Transaction, the GXB maintains system concurrency by continuing to accept and process
additional writes as buffering allows. If the Delayed Transaction Request slot is full, the GXB will
retry all reads that do not match the address and byte enables of the pending Delayed Transaction.
For PCI Stream AGP to DRAM writes, the GXB will post the cycle unless the posted write slots
are full (in which case the write is retried).
The GXB does not differentiate between read transactions from different masters. If a different
master, other than the one that initiated the Delayed Transaction, attempts to read the same DRAM
location as the Delayed Transaction, the GXB will respond with the data and complete the Delayed
Transaction on the AGP bus.
7.2.7.2
PCI Stream Read Prefetching
The GXB delays all inbound reads. A Memory Read targeting memory will fetch 8B, unless the
transaction begins 4B from the end of a cache line, in which case the transaction only fetches 4B. A
Memory Read Line or Memory Read Multiple will always prefetch up to one cache line of data.
When the read data is available in the GXB, the next matching read attempt from the controller is
accepted and the data is streamed to the AGP bus until the controller disconnects or the prefetched
data is consumed. Any data left when the controller disconnects is discarded. If the controller
disconnects half way through a line, and comes back with the next address; then a new read is sent
to DRAM for the data.
7.2.7.3
Inbound Reads Directed To Memory
All PCI reads-to-memory are propagated up through the SAC and placed on the system bus to
allow snooping. The snoops are performed using the Memory Read (Length=0) transaction, which
does not transfer data on the system bus. Instead, the data is transferred directly from the memory
to the SAC, and finally back to the GXB. This path means that PCI Stream AGP-DRAM reads
need consume no host data bus bandwidth (except in the case of implicit writebacks).
Note: The GXB as a PCI target supports only a linear incrementing burst (denoted by AD[1:0]=00 during
the address phase). A PCI burst request specifying any type of sequence other than linear results in
a disconnect after the first data phase.
7.2.7.4
Inbound Delayed Read Matching Rules
The completion response returned to the GXB for a Delayed Read transaction is matched to the
requesting PCI master by using information that was latched into the target during the initial
which read request information is used for matching Delayed Read Completions to the requesting
agent.
7-10
Intel® 460GX Chipset Software Developer’s Manual
AGP Subsystem
Table 7-2. Delayed Read Matching Criteria
Command
Address
Match
BEs
Any Memory Read
Match
When a DRC is valid in the GXB, a 215 PCI clock timer is started as described in the PCI 2.2
Specification. When the timer expires the DRC is discarded and the associated delayed read
matching registers are cleared. This condition is optionally treated as an error. See the “Error”
Chapter for details.
7.2.7.5
7.2.7.6
Inbound I/O Reads
I/O reads on the PCI bus are not claimed by the GXB.
Inbound Writes
Memory write requests may be directed to the memory only. All write requests are posted. A series
of buffers in both the GXB and SAC allow posted writes to be accumulated and serviced as
convenient. The actual data is forwarded to buffers in the SAC, where it is held until the request
can be snooped on the system bus.
Writes to Memory
The snoops on the system bus for PCI Stream AGP-DRAM writes will be initiated by using the
Memory Read and Invalidate of Length=0 system bus transaction. Any implicit write back due to
snoop hit to a modified line will be snarfed by the SDC. The write back data will be merged with
the AGP write data within the SDC.
Writes to memory use the linear burst ordering provided on the AGP bus.
The Write and Write & Invalidate commands have small differences, as follows:
Write
Accumulates posted data until (a) a cache line boundary is reached, or (b) the
master disconnects, before forwarding the request to the SAC. The SAC
therefore deals with a single packet that represents up to a cache line of data.
The SAC places a cache-line snoop on the bus, and the SDC handles the writing/
merging of the partial data into memory. The Expander bus command is a write.
Write & Invalidate By definition this must be an aligned cache line worth of data. It is forwarded
to the SAC as a cache line of data. The SAC places a cache-line snoop on the
bus for each line of data, and the SDC writes the full cache line into memory.
In the event the snoop results in a HITM#, the write back data from the
processor is discarded by the SDC.
Note that the actual writing of the data into memory may be delayed. However, in effect, the data
can be counted as being written in the memory, since no other write to that location may pass this
write.
7.2.7.7
Inbound I/O Writes
I/O writes on the PCI bus are not claimed by the GXB.
Intel® 460GX Chipset Software Developer’s Manual
7-11
AGP Subsystem
7.2.7.8
Retry/Disconnect Conditions
The GXB as a PCI target retries the initial data phase of inbound access when:
• The read request is to an address that has already been accepted as a delayed transaction (i.e.
the request is already being serviced, but data has not arrived).
• A write request has insufficient buffering in the gxb to allow it to be posted. (a full line is not
available for mwi).
• The pci interface is locked from the host side.
• No delayed read buffer is available.
The GXB as a target will issue a disconnect to a read when no more data is available (at the end of
the cacheline), or if linear addressing is not used. The GXB as a target will issue a disconnect to a
write when no more posting buffers are available, or when the write crosses a cacheline boundary.
7.2.7.9
Outbound Reads
Outbound reads are enqueued in the GXB. The GXB will hold up to four cache line reads at one
time. The GXB will continuously retry an outbound read until it completes successfully on the
AGP bus. Outbound posted writes must be allowed to pass the read(s) between retry attempts. PCI
2.2 ordering rules allow reads to pass each other. However, this is optional in the specification and
is not required to guarantee forward progress. For the GXB, outbound reads are done in order. If a
read is retried, then all other reads after it will wait until the first has completed.
7.2.7.10
Outbound Writes
Write Combining
The GXB optimizes outbound write performance by combining writes to sequential locations (if
enabled) into a single write burst on the AGP bus. This holds true for all memory attributes, not
just WC. The GXB only provides write combining; no collapsing or byte merging is performed.
The source of the transaction is not checked. Accesses from processors could be combined with
peer accesses.
The GXB will combine an access if the next data is valid at the head of the queue in time; this puts
a restriction on the size of requests that can be combined (if a request is too short the GXB does not
Notice that even if use of the Memory Write and Invalidate command is enabled, the GXB will not
terminate a burst that was started with a Memory Write command in order to switch to the Memory
Write Invalidate Command when it receives a full line. However, if a Memory Write Invalidate
burst has been started and the next sequential access is less than a line, the GXB must terminate and
7-12
Intel® 460GX Chipset Software Developer’s Manual
AGP Subsystem
Table 7-3. Burst Write Combining Modes
Write Command
Used
Transfer
Mode
Data Length
Combining Supported
Memory Write
Memory Write
Memory Write
1X
< 4 DW
Can not be combined with the next access.
>= 4 DW
Can combine next access if it is sequential, regardless
of next access size.
2X
4X
All
< 8 DW
Can not be combined with the next access.
>= 8 DW
Can combine next access if it is sequential, regardless
of next access size.
< 16 DW
Can not be combined with the next access.
>= 16 DW
Can combine next access if it is sequential, regardless
of next access size.
Memory Write
Invalidate
must be Line
Can only combine sequential lines.
Table 7-4. Burst Write Combining Examples with 3 Writes in 1X Transfer Mode
1st write
2nd write
y DW
3rd write
z DW
Transferred as:
x DW
(x<4)
if y < 4
x DW followed by y DW followed by z DW
else
x DW followed by (y+z) DW
x DW
y DW
z DW
if y < 4
(x>=4)
(x+y) DW followed by z DW
else
(x+y+z) DW
7.2.7.11
Fast Back-to-Back Transactions
The GXB as a PCI target will accept fast back-to-back cycles from a PCI master accessing different
agents during back-to-back sequence. As an initiator the GXB does not generate a fast back-to-
back cycle.
7.3
Bandwidth
The bandwidth of AGP 4X mode is 1 GB peak. The sustained bandwidth will be less than this. The
sustainable bandwidth obtainable with the GX is dependent on several things including the size of
a fairly broad range for each of the sizes. The bandwidth will depend on the mix of traffic. Traffic
consisting of all reads by the graphics card or all writes by the processor to the card will result in
higher bandwidths than a 50/50 mix of reads by the graphics card and writes by the processor to the
card. This table will change as the implementation is completed and should be viewed as a
guideline for the graphics card designer for relative performance trade-offs.
Intel® 460GX Chipset Software Developer’s Manual
7-13
AGP Subsystem
Table 7-5. Bandwidth Estimates for Various Request Sizes
Request Size
(in bytes)
Sustainable Bandwidth
(MB/s)
Latency (in ns.) for First Data
Returned on AGP Bus
8
100-300
500
16
32
64
128
256
250-500
450-700
750-850
850-925
800-850
500
530
600
750
1000
7.3.1
Inbound Read Prefetching
The PCI protocol has no transfer size explicitly spelled out. Reads begin and continue until the
device has the data it needs. For performance, a PCI bridge could prefetch data ahead of when the
PCI device requests it. AGP, as opposed to PCI, has a length field in its protocol. Therefore the
graphics bridge does know how much data to fetch. Also the GXB has 1 GB/s peak into it, which is
the AGP peak, so there is no extra bandwidth in case the bridge prefetched the wrong data. For this
reason, the 460GX chipset will have no prefetching of graphics data. The graphics card should
have enough buffering and outstanding transactions to keep its pipes full. With pipelining, the card
can fetch the data it needs in advance of using it. Knowing what it wants to do, the graphics card
can prefetch only the data it expects to use. This allows the best utilization of the graphics port.
7.4
7.5
Latency
Latency of reads from the graphics card is an important parameter. Unfortunately it is difficult to
specify usefully. For an idle system, the latency of a graphics read will be 450-500 ns. For a
moderately loaded system, the latency of a read should be 600-900 ns. Defining ‘moderately
loaded’ is the difficult part. If the GXB queues are backed up servicing many small AGP requests,
then the latency may be much higher.
GXB Address Map
The ‘System Address Map’ chapter contains a section describing the what the system address map
looks like from the perspective of an expander bridge. This section is included to explicitly state
which address ranges must be checked at the AGP interface when using PCI protocol and which
ranges must be checked after GART translation.
There are two address regions which must be checked at the AGP bus interface when the card is
using PCI protocol: the GXB’s nx32M logical PCI space and the VGA space. This is required since
the AGP card may attempt to “talk to itself” using the PCI protocol in these regions, so the GXB
must not assert DEVSEL# to these accesses. The GXB’s nx32M region is specified using its
MMBASE and MMTOP registers. The VGA gap is enabled as a memory hole using the VGASE
register in the GXB. All other accesses using PCI protocol will receive DEVSEL#. This includes
accesses above TOM, in a MARG, etc. that may cause XBINIT# after the GART (see below).
7-14
Intel® 460GX Chipset Software Developer’s Manual
AGP Subsystem
All regions, including the two described above, must be checked after GART translation. The GXB
must only allow accesses that are directed to physical memory to reach the SAC. Therefore, the
GXB must force a BINIT#, by asserting its “XBINIT#” output, when it detects an access falling in
any of the following regions:
• Above the TOM register value (firmware sets TOM to top of physical system memory - this
may be anywhere in memory, as low as 64 MB or as high as 64 GB)
• Between GAPBASE and GAPTOP registers and not in the SAPIC interrupt delivery region -
firmware must set these two registers to cover the following ranges:
— In High System firmware (fixed range from 4G to 4G-16M).
— In first megabyte of Itanium processor specific (fixed range from 4G-16M to 4G-17M).
— In Itanium processor specific below the Interrupt region (fixed range from 4G-18M to
4G-20M).
— In chipset specific (fixed range from 4G-20M to 4G-32M).
— In any of the nx32M PCI spaces.
Therefore GAPTOP=4G and GAPBASE is the lowest PXB’s/GXB’s MMBASE in the system.
Note, the SAPIC interrupt delivery area from 4G-17M to 4G-18M must be allowed and specifically
decoded so it will *not* cause a fault.
• In a MARG with DRAM accesses disabled (the MARGs cover the C, D, E, and F segments).
• In the VGA region (A_0000-BFFFF) with VGAGE enabled so that this region is directed to
PCI, a PCI access in this range would not have gotten DEVSEL#.
Intel® 460GX Chipset Software Developer’s Manual
7-15
AGP Subsystem
7-16
Intel® 460GX Chipset Software Developer’s Manual
WXB Hot-Plug
8
8.1
IHPC Configuration Registers
Each WXB supports two independent Integrated Hot-Plug1 Controllers (IHPCs). The A-side
controller (IHPA) and the B-side controller (IHPB) are configured independently. Each IHPC
therefore has its own configuration space. Both configuration spaces are identical.
Each IHPC reserves a 256 byte configuration space. A list of the configuration registers specific to
hot-plug operation follows. The default power-up value is included in each register description
heading. All registers return to their default values upon the assertion of XRST# or PCI reset (for
the particular side in question) unless otherwise noted in the description. The standard PCI
Configuration registers are not described here when implemented exactly as described in the PCI
Specification 2.2.
The following map shows registers specific to the IHPC.
1. WXB Integrated Hot-Plug controller derived from technology licensed from Compaq Corporation.
Intel® 460GX Chipset Software Developer’s Manual
8-1
WXB Hot-Plug
Table 8-1. IHPC Configuration Register Space
DID
VID
00h
04h
08h
0Ch
10h
14h
18h
1Ch
20h
24h
28h
2Ch
30h
34h
38h
3Ch
80h
84h
88h
8Ch
90h
94h
98h
9Ch
A0h
A4h
A8h
ACh
B0h
B4h
B8h
BCh
PCISTS
PCICMD
CLASS
HDR
Base Address
RID
MLT
CLS
SID
SVID
Interrupt
Pin
Interrupt
Line
Miscellaneous
Configuration (RW)
Slot ID
40h
C0h
Hot-Plug Features
44h
48h
C4h
C8h
Arbiter
SERR
status
Power
Fault
Switch
Change
SERR
status
SERR
status
4Ch
50h
54h
58h
5Ch
60h
64h
68h
6Ch
70h
74h
78h
7Ch
CCh
D0h
D4h
D8h
DCh
E0h
E4h
E8h
ECh
F0h
F4h
F8h
FCh
Memory Access Index
Memory Mapped Register Access Port
NOTE: The first 64 bytes are predefined in the PCI specification. All other locations are defined specifically for the
component of interest.
8-2
Intel® 460GX Chipset Software Developer’s Manual
WXB Hot-Plug
8.1.1
Page Number List for the IHPC PCI Register Descriptions
Register
Page
Arbiter SERR Status......................................................................................8-10
Base Address...................................................................................................8-7
CLASS: Class Register ..................................................................................8-6
CLS: Cache Line Size .....................................................................................8-6
Device Identification Register.........................................................................8-3
HDR: Header Register ....................................................................................8-6
Hot-Plug Features............................................................................................8-9
Hot-Plug Slot Identifier...................................................................................8-8
Interrupt Pin.....................................................................................................8-8
Interrupt Line...................................................................................................8-7
Master Latency Timer .....................................................................................8-6
Memory Access Index...................................................................................8-10
PCISTS: PCI Status Register ..........................................................................8-5
Power Fault SERR Status................................................................................8-9
RID: Revision Identification Register.............................................................8-5
Subsystem Vendor ID .....................................................................................8-7
Subsystem ID ..................................................................................................8-7
Switch Change SERR Status...........................................................................8-9
VID: Vendor Identification Register...............................................................8-3
8.1.2
VID: Vendor Identification Register
Address Offset:
Default Value:
00 – 01h
8086h
Size:
Attributes:
16 bits
Read-Only
The VID Register contains the vendor identification number, identifying the manufacturer of the
device. This 16-bit register combined with the Device Identification Register uniquely identifies
any PCI device. Writes to this register have no effect.
Bits
Description
15:0
Vendor Identification Number
This is a 16-bit value assigned to Intel. Intel VID = 8086h
8.1.3
DID: Device Identification Register
Address Offset:
Default Value:
02 – 03h
123Fh
Size:
Attributes:
16 bits
Read-Only
This 16-bit register combined with the Vendor Identification register uniquely identifies any PCI
device. Writes to this register have no effect.
Bits
Description
15:0
Device Identification Number
The value 123F indicates the IHPC
Intel® 460GX Chipset Software Developer’s Manual
8-3
WXB Hot-Plug
8.1.4
PCICMD: PCI Command Register
Address Offset:
Default Value:
04h-05h
0000h
Size:
Attribute:
16 bits
Partial Read/Write
The PCI command register provides control over the IHPC’s ability to generate and respond to PCI
cycles. When a zero (0) is written to this register, the IHPC is logically disconnected from the PCI
bus for all accesses except configuration. The register bit list below shows the layout of the
register, explains the meanings of the different bits in the command register, gives the default value
of this register upon power-up, and gives the specific implementation of individual bits in the IHPC
(i.e. R/O or R/W).
Bits
15:10
9
Description
reserved (0)
Fast Back-to-Back Enable
This bit is not supported. Hardwired Value = 0.
8
SERR# Enable
This bit is an enable bit for the SERR# driver. A value of zero (0) disables the SERR#
driver. A value of one (1) enables the SERR# driver. WXB configuration space
PCICMD(6) must be set to 1 to permit WXB SERR# pin assertion. Since the IHPC is
integrated into the WXB and is not addressable from the local PCI bus segment, the IHPC
will not report address parity errors via SERR#. This bit is configurable in the IHPC with
a default value of zero (0).
7
6
Wait Cycle Control
This bit is not supported. Hardwired Value = 0.
Parity Error Enable
This bit controls a device’s response to IHPC configuration and memory write parity
errors. When the bit is 1, the IHPC will assert both the Detected Parity Error status bit and
the PERR# output upon detection of an error. When the bit is 0, the IHPC will assert the
Detected Parity Error status on an error but will not assert PERR#. This bit must be set to
zero (0) after RST#. This bit is configurable in the IHPC with the default value zero (0).
5
4
3
VGA Palette Snoop
This bit controls how VGA-compatible devices handle accesses to their palette registers.
This bit is always set to zero (0) in the IHPC (disabled).
Memory Write and Invalidate Enable
This is an enable bit for using the memory write and invalidate command. This bit is
always set to zero (0) in the IHPC (disabled).
Special Cycle
This bit controls a device’s action on special cycle operations. A value of zero (0) causes
the device to ignore all special cycle operations. This bit is always set to zero (0) in the
IHPC.
2
1
0
Bus Master
This bit controls a device’s ability to act as a master on the PCI bus. A value of zero (0)
disables the device from generating PCI accesses. A value of one (1) allows the device to
behave as a bus master. Hardwired Value = 0.
Memory Space
This bit controls a device’s response to memory space accesses. A value of zero (0)
disables the device response. A value of one (1) allows the device to respond to memory
space accesses. This bit is configurable in the IHPC with the default value of zero (0).
I/O Space
This bit controls a device’s response to I/O space accesses. The IHPC does not respond to
I/O space accesses. Hardwired Value = 0.
8-4
Intel® 460GX Chipset Software Developer’s Manual
WXB Hot-Plug
8.1.5
PCISTS: PCI Status Register
Address Offset:
Default Value:
06h – 07h
0200h
Size:
Attribute:
16 bits
Partial Read/Write, Sticky
The PCI status register is used to record status information for PCI bus-related events. The
definition of each of the bits is given in the register bit list below. The specific implementation of
each bit in the IHPC is also given. Reads to this register behave normally. Writes are slightly
different in that bits can be reset, but not set. A bit is reset whenever the register is written and the
data in the corresponding bit location is a one (1). For instance, to clear bit 14 and not affect any
other bits, write the value 0100_0000_0000_0000b to the register.
Bits
Description
15
Detected Parity Error
This bit is set by the IHPC whenever it detects a parity error, even if parity error handling
is disabled (as controlled by bit 6 in the command register). The default value of this bit
in the IHPC is zero (0).
14
Signaled System Error
This bit is set whenever the IHPC asserts SERR#. The default value of this bit in the IHPC
is zero (0).
13
Received Master Abort
Not supported. Hardwired Value = 0
12
Received Target Abort
Not supported. Hardwired Value = 0
11
Signaled Target Abort
Not supported. Hardwired Value = 0
10:9
DEVSEL# Timing
These bits encode the timing of DEVSEL#. There are three allowable timings for
assertion of DEVSEL#. These are encoded as 00b for fast, 01b for medium, and 10b for
slow (11b is reserved). The value of these bits are always set to medium (01).
8
Data Parity Error
Not supported. Hardwired Value = 0
7
Fast Back-to-Back Capable
Not supported. Hardwired Value = 0
6:0
reserved(0)
8.1.6
RID: Revision Identification Register
Address Offset:
Default Value:
08h
Size:
Attribute:
8 bits
Read-Only
Stepping-Dependent
This register contains the revision number of the IHPC. These bits are read-only and writes to this
register have no effect.
Intel® 460GX Chipset Software Developer’s Manual
8-5
WXB Hot-Plug
Bits
Description
7:0
Revision Identification Number
This is an 8-bit value that indicates the revision identification number for the IHPC
WXB A Steppings:
WXB B0 Step:
Hardwired Value = 00h
Hardwired Value = 01h
8.1.7
CLASS: Class Register
Address Offset:
Default Value:
09 – 0Bh
080400h
Size:
Attribute:
24 bits
Read-Only
This register contains the Class Code for the IHPC, specifying the device function. Writes to this
register will have no effect.
Bits
Description
23:16
Base Class
This field indicates the general device category. The IHPC is a Base System Peripheral.
Hardwired Value = 08h.
15:8
7:0
Sub-Class
This field qualifies the Base Class, providing a more detailed specification of the device
function. For the IHPC this field indicates a Generic PCI Hot-Plug Controller. Hardwired
Value = 04h.
Register-level Programming Interface
This field identifies a specific programming interface (if any), that device independent
software can use to interact with the device. The Interface is not defined. Hardwired
Value = 00h.
8.1.8
8.1.9
8.1.10
CLS: Cache Line Size
Address Offset:
Default Value:
0Ch
00h
Size:
Attribute:
8 bits
Read/Write
See PCI Specification, Rev. 2.2.
MLT: Master Latency Timer Register
Address Offset:
Default Value:
0Dh
00h
Size:
Attribute:
8 bits
Read/Write
See PCI Specification, Rev. 2.2. This register is not applicable to the IHPC’s within the WXB.
HDR: Header Register
Address Offset:
Default Value:
0Eh
00h
Size:
Attribute:
8 bits
Read-Only
This register identifies the header layout of the configuration space. Writes to this register will have
no effect.
8-6
Intel® 460GX Chipset Software Developer’s Manual
WXB Hot-Plug
Bits
Description
7
Multi-function Device
Selects whether this is a multi-function device, that may have alternative configuration
layouts. The IHPC is not a multifunction device. Hardwired Value = 0.
6:0
Configuration Layout
This field identifies the format of the 10h through 3Fh space. This field specifies the
“standard” or “default” PCI configuration layout. Hardwired Value = 00h
8.1.11
Base Address
Address Offset:
Default Value:
10h-13h
00000000h
Size:
Attribute:
32 bits
Read/Write, Read-Only
This is a standard PCI configuration register which provides power-up software with the ability to
build a consistent address map before booting the machine.
Bits
Description
31:8
Base Address
Read/Write.
7:4
3
Indicate 256-byte Address Space Requested
Hardwired Value = 0h.
Not Prefetchable
Hardwired Value = 0.
2:1
0
Type: Located anywhere in 32-bit address space.
Hardwired Value = 00.
Memory Space Indicator
Hardwired Value = 0.
8.1.12
8.1.13
8.1.14
SVID: Subsystem Vendor Identification
Address Offset:
Default Value:
2Ch –2Dh
0000h
Size:
Attributes:
16 bits
Write-Once (Pwr Good Reset Only)
See PCI Specification, Revision 2.2.
SID: Subsystem ID
Address Offset:
Default Value:
2Eh - 2Fh
0000h
Size:
Attributes:
16 bits
Write-Once (Pwr Good Reset Only)
See PCI Specification, Revision 2.2.
Interrupt Line
Address Offset:
Default Value:
3Ch
FFh
Size:
Attribute:
8 bits
Read-Write (Pwr Good Reset Only)
Intel® 460GX Chipset Software Developer’s Manual
8-7
WXB Hot-Plug
This is a standard PCI configuration register which defines which interrupt request line on the
interrupt controller this function’s interrupt pin (see register 3DH) is connected to. The power-up
default value is FFh.
8.1.15
Interrupt Pin
Address Offset:
Default Value:
3Dh
01h
Size:
Attribute:
8 bits
Read Only
This is a standard PCI configuration register which defines which of the four PCI interrupt pins,
INTA# through INTD#, this function is connected to. PCI Hot-Plug is connected to INTA#, making
this value hard-wired to 01h.
8.1.16
Hot-Plug Slot Identifier
Address Offset:
Default Value:
40h-41h
0000h
Size:
Attribute:
16 bits
Partial Read/Write
This register indicates which slots support hot-plug. Not used by the IHPC, this register is set by
the boot ROM for later reference by the hot-plug drivers. This register is also mapped to memory
space. System designers are required to follow a convention whereby hot-plug slots are always
implemented in sequential PCI device number order. IHPC slot A should be the lowest numbered
slot (i.e. the device number in this register).
Bits
15:8
7:4
Description
reserved(0)
The PCI device number for the first slot that supports hot-plug.
Number of hot-plug slots controlled.
3:0
8.1.17
Miscellaneous Hot-Plug Configuration
Address Offset:
Default Value:
42h-43h
0002h 1
Size:
Attribute:
16 bits
Read/Write, Write-Once, Read-Only
This is a hot-plug specific register used to configure many features of the IHPC.
Bits
Description
15
Change device ID. Write Once/Read Only. Changes the device ID from 123Fh to 123Eh
in order to “hide” the IHPC from industry-standard hot-plug drivers and device
enumerators in non-hot-plug systems.
14
13
12
Inhibit hot-plug registers. Write Once/Read Only. This bit prohibits access to the IHPC
registers, effectively disabling the IHPC.
Enable Power Fault functions. This bit must be set to enable other power-fault related
SERR signaling.
Enable Lock of the Auto Power-Down Disable bit. Write Once/Read Only. This bit
should be written to a logic 1 (by firmware) after the Auto-Power-Down Disable bit of
the Hot-Plug Miscellaneous register has been written. Once written, this bit causes the
Auto-Power-Down Disable bit to be unchangeable until XRST# or PCI reset. This
prevents driver software from changing the auto-power-down feature.
1. Bit 0 may be changed to a 1 after Reset in future implementation.
8-8
Intel® 460GX Chipset Software Developer’s Manual
WXB Hot-Plug
11:8
7
reserved(0)
Enable PCI Configuration Space Access to Hot-Plug Registers. Enables IHPC memory-
mapped register access through the index register (configuration offset 50h) and data port
(configuration offset 54h).
6:2
1
reserved (0)
reserved (1)
0
On / Off Busy (OOBS) status. Read Only. Same as bit 24 of the memory-mapped Hot-
Plug Miscellaneous register.
8.1.18
8.1.19
Hot-Plug Features
Address Offset:
Default Value:
44h-45
0000h
Size:
Attribute:
16 bits
Read Only
Definitions of each bit within this register are expected to be constant through the industry but are,
as of yet, undefined.
Bits
Description
15:0
reserved (0)
Switch Change SERR Status
Address Offset:
Default Value:
48h
00h
Size:
Attribute:
8 bits
Partial Read/Write
Bits
7:6
Description
reserved (0)
5:0
Switch Change SERR Status. Slot F is MSB. Slot A is LSB. Similar to the Power Fault
SERR status register, but applicable to switch changes when the switch interrupt redirect
bit for a slot is set and the associated interrupt mask bit is logic 0. Unlike the Power Fault
SERR status register, clearing this bit will also clear the associated interrupt.
8.1.20
Power Fault SERR Status
Address Offset:
Default Value:
49h
00h
Size:
Attribute:
8 bits
Partial Read/Write
Bits
7:6
Description
reserved.
5:0
Power Fault SERR Status. If the power fault function enable bit and the SERR on power
fault bit are both set, then these six bits will indicate (by a logic 1, one for each slot) if a
power fault has occurred while a slot was connected to the bus or clock. Slot F is MSB.
Slot A is LSB. These bits can be cleared by writing a logic 1 to the appropriate position.
This register does not effect PCI interrupts.
Intel® 460GX Chipset Software Developer’s Manual
8-9
WXB Hot-Plug
8.1.21
Arbiter SERR Status
Address Offset:
Default Value:
4A
00h
Size:
Attribute:
8 bits
Partial Read/Write
Bits
Description
reserved (0)
7:0
8.1.22
Memory Access Index
Address Offset:
Default Value:
50h-53h
00000000h
Size:
Attribute:
32 bits
Partial Read/Write
When the “Enable PCI Config Space Access to Hot-Plug Registers” bit in the Miscellaneous Hot-
Plug Configuration Register is set, this register becomes a pointer into the IHPC memory mapped
register space (divided into 64, 32-bit Dwords).
Bits
31:8
7:2
Description
reserved (0)
Hot-Plug Memory Access Index
Read/Write
1:0
reserved (0)
8.1.23
Memory Mapped Register Access Port
Address Offset:
Default Value:
54h-57h
00000000h
Size:
Attribute:
32 bits
Read/Write
When the “Enable PCI Config Space Access to Hot-Plug Registers” bit in the Miscellaneous Hot-
Plug Configuration Register is set, this register becomes mapped into the IHPC memory mapped
register space at the location pointed to by the Memory Index Register.
Bits
Description
31:0
Memory Mapped Register Access Port
8.2
IHPC Memory Mapped Registers
Each IHPC reserves 256 bytes of memory mapped registers. A list of those registers follows.
Unlike other register descriptions in this document, the offset listed is always Dword aligned with
bit offset provided. This nomenclature is used to be consistent with the Memory Access Index
Register in IHPC configuration space. The default power-up value is included in each register
description heading.
8-10
Intel® 460GX Chipset Software Developer’s Manual
WXB Hot-Plug
Table 8-2. IHPC Memor Mapped Register Space
Hot-Plug Miscellaneous SlotEnable
00h
80h
(RW)
(RW)
LED Control (RW)
04h
08h
84h
88h
Hot-Plug Interrupt Input and Clear (RW)
Present 1
bits
Present 2
bits
Power
Faults
Switches
Hot-Plug Interrupt Mask(RW)
0Ch
10h
8Ch
90h
Present 1
mask
Present 2
mask
Power
Fault mask
Switch
Mask
User-
Defined
Reserved SerialInput SerialInput
Byte
Pointer
(RW)
Data (RO)
Outputs
(RW)
Hot-Plug Non-Interrupt Inputs (RW)
14h
94h
Reserved
Reserved
Reserved
M66EN
18h
1Ch
20h
24h
28h
98h
9Ch
A0h
A4h
A8h
Reserved
Slot ID
(RW)
Slot Power Switch Int.
2Ch
ACh
(RW)
Redirect
Enable
(RW)
ExtendedHot-PlugMisc.
(RW)
30h
B0h
34h
38h
3Ch
40h
44h
48h
4Ch
50h
54h
58h
5Ch
60h
64h
68h
6Ch
70h
74h
78h
7Ch
B4h
B8h
BCh
C0h
C4h
C8h
CCh
D0h
D4h
D8h
DCh
E0h
E4h
E8h
ECh
F0h
F4h
F8h
FCh
NOTE: The first 64 bytes are predefined in the PCI specification. All other locations are defined specifically for the
component of interest.
Intel® 460GX Chipset Software Developer’s Manual
8-11
WXB Hot-Plug
8.2.1
Page Number List for IHPC Memory Mapped Register
Descriptions
Register
Page
Extended Hot-Plug Miscellaneous................................................................ 8-18
LED Control.................................................................................................. 8-13
General Purpose Output................................................................................ 8-17
Hot-Plug Interrupt and Clear......................................................................... 8-14
Hot-Plug Interrupt Mask............................................................................... 8-15
Hot-Plug Miscellaneous................................................................................ 8-13
Hot-Plug Non-interrupt Inputs...................................................................... 8-17
Hot-Plug Slot Identifier................................................................................. 8-17
Hot-Plug Switch Interrupt Redirect Enable .................................................. 8-18
Serial Input Byte Data................................................................................... 8-16
Serial Input Byte Pointer............................................................................... 8-17
Slot Enable.................................................................................................... 8-12
Slot Power Control........................................................................................ 8-18
8.2.2
Slot Enable
Address Offset:
Default Value:
01h
00h
Size:
Attribute:
8 bits
Partial Read/Write (Pwr Good Rst Only)
Used to power-on a slot and connect it to the bus (or disconnect and power-down). The SOGO bit
must be set to start the output sequence. The set of usable Enable Slot bits is determined by the
strapping values on the P(A,B)HSIL, P(A,B)HSOL, and P(A,B)HSOC inputs. Unsupported slots in
a system do not have writeable Enable Slot bits. Writing a zero to a Enable Slot bit will clear the
associated Slot Power register bit.
Bits
7:6
5
Description
reserved(0)
Enable Slot F
When 1, Slot F is powered and connected to the PCI bus
4
3
2
1
0
Enable Slot E
When 1, Slot E is powered and connected to the PCI bus
Enable Slot D
When 1, Slot D is powered and connected to the PCI bus
Enable Slot C
When 1, Slot C is powered and connected to the PCI bus
Enable Slot B
When 1, Slot B is powered and connected to the PCI bus
Enable Slot A
When 1, Slot A is powered and connected to the PCI bus
8-12
Intel® 460GX Chipset Software Developer’s Manual
WXB Hot-Plug
8.2.3
Hot-Plug Miscellaneous
Address Offset:
Default Value:
02h - 03h
0040h
Size:
Attribute:
16 bits
Partial Read/Write
Bits
15
Description
reserved (0)
14
Enable SERR on Power Fault. When set, the assertion of a slot power fault causes a
SERR# to be asserted if SERR# generation is enabled in the PCI device command
register, the slot is connected to the bus or PCI clock, and IHPC power fault functions are
enabled (bit 10 of this register).
13
12
reserved (0)
Input Scan Complete. This bit is cleared at the conclusion of each input cycle. Software
can use this bit to determine when fresh data is available by setting it and waiting until a
logic 0 is seen.
11
10
9
66 MHz Prescaler Enable, Read Only, set according to the PCI frequency, a 1 indicates
the PCI frequency is 66 MHz, a 0 indicates 33 MHz.
Enable Power Fault Functions. This bit is also mapped to bit 13 of the IHPC Misc.
Configuration Register (configuration offset 42h).
Auto Power Down Disable. Controls whether or not opening a slot switch will cause a
powered slot to auto-power-down.
8
7
6
5
4
3
On / Off Busy status. (Same as Configuration Register 42, bit 0)
reserved (0)
reserved (1)
reserved (0)
Dummy Cycle Enable, this feature is not supported on this stepping.
General-interrupt-input Interrupt Pending. Set to a logic 1 when an interrupt is generated.
Cleared when the interrupt is cleared.
2
Shift Output Interrupt Pending / Clear. When read as logic 1, a hot-plug interrupt was
generated by SOBS changing from 1 to 0 while the Serial Output Interrupt Enable bit was
set. Writing a logic 1 clears this bit and its interrupt.
1
0
Shift Output Interrupt Enable. When set to a logic 1, an interrupt will be generated when
SOBS changes from 1 to 0, indicating completion of a serial output sequence.
Shift Output Go / Busy Status. Writing a logic 1 to this bit (after it has been read as a logic
0) initiates a serial output sequence (e.g. power-down of a slot). When read as a logic 0,
the previous serial output sequence has completed. Ensure that LED blinking has
completed (or is forced to complete) prior to initiating a SOGO for a slot power-up or
power-down.
8.2.4
LED Control
Address Offset:
Default Value:
04h
Size:
Attribute:
32 bits
sampled at PWRGD
Read/Write (Pwr Good Rst Only)
If both the MSB and LSB for an LED are logic 0, then the LED is turned off. The LSB is for Blink
phase A and the MSB is for Blink phase B. Setting either phase A or phase B will program the
LED to blink. Setting both the MSB and the LSB will turn the LED on. An auto-power-down
sequence will turn off the LEDs for that slot. It is intended that the green LED is a power indicator
and the amber LED is the attention indicator.Following chip power on, all LED bits will be cleared
Intel® 460GX Chipset Software Developer’s Manual
8-13
WXB Hot-Plug
for unpopulated slots and slots with open switches. The set of usable LED Control bits is
determined by the strapping values on the P(A,B)HSIL, P(A,B)HSOL, and P(A,B)HSOC inputs.
Unsupported slots in a system do not have writeable LED Control bits. LEDs are not affected by
changes to these register bits until a SOGO is initiated following the changes.
Bits
31:30
29
28
27
26
25
24
23:22
21
20
19
18
17
16
15:14
13
12
11
10
9
Description
reserved (0)
Slot F Amber LED, msb
Slot E Amber LED, msb
Slot D Amber LED, msb
Slot C Amber LED, msb
Slot B Amber LED, msb
Slot A Amber LED, msb
reserved (0)
Slot F Amber LED, lsb
Slot E Amber LED, lsb
Slot D Amber LED, lsb
Slot C Amber LED, lsb
Slot B Amber LED, lsb
Slot A Amber LED, lsb
reserved (0)
Slot F Green LED, msb
Slot E Green LED, msb
Slot D Green LED, msb
Slot C Green LED, msb
Slot B Green LED, msb
Slot A Green LED, msb
reserved(0)
8
7:6
5
Slot F Green LED, lsb
Slot E Green LED, lsb
Slot D Green LED, lsb
Slot C Green LED, lsb
Slot B Green LED, lsb
Slot A Green LED, lsb
4
3
2
1
0
8.2.5
Hot-Plug Interrupt Input and Clear
Address Offset:
Default Value:
08h
Varies
Size:
Attribute:
32 bits
Read/Write Clear (Pwr Gd Rst Only)
These bits indicate the state of the interrupt capable inputs when no interrupt is pending. When a
hot-plug general interrupt changes state (either high to low or low to high), and the bit is not
masked in the Interrupt Mask register, an interrupt is generated and the state of that input is latched
8-14
Intel® 460GX Chipset Software Developer’s Manual
WXB Hot-Plug
into this register. Writing a logic 1 will clear the pending interrupt. If there are no other pending
interrupts on the bit, the bit will clear. This register takes on a value based on the monitored status
of the slots and therefore has no particular default value.
Bits
31:30
29
28
27
26
25
24
23:22
21
20
19
18
17
16
15:14
13
12
11
10
9
Description
reserved (0)
Slot F PRSNT(0)#, PCI Present Signal 1
Slot E PRSNT(0)#, PCI Present Signal 1
Slot D PRSNT(0)#, PCI Present Signal 1
Slot C PRSNT(0)#, PCI Present Signal 1
Slot B PRSNT(0)#, PCI Present Signal 1
Slot A PRSNT(0)#, PCI Present Signal 1
reserved (0)
Slot F PRSNT(1)#, PCI Present Signal 2
Slot E PRSNT(1)#, PCI Present Signal 2
Slot D PRSNT(1)#, PCI Present Signal 2
Slot C PRSNT(1)#, PCI Present Signal 2
Slot B PRSNT(1)#, PCI Present Signal 2
Slot A PRSNT(1)#, PCI Present Signal 2
reserved (0)
Slot F FAULT#, PCI Power Fault Signal
Slot E FAULT#, PCI Power Fault Signal
Slot D FAULT#, PCI Power Fault Signal
Slot C FAULT#, PCI Power Fault Signal
Slot B FAULT#, PCI Power Fault Signal
Slot A FAULT#, PCI Power Fault Signal
reserved(0)
8
7:6
5
Slot F Hot-Plug Switch, 0 = lever closed (board installed)
Slot E Hot-Plug Switch, 0 = lever closed (board installed)
Slot D Hot-Plug Switch, 0 = lever closed (board installed)
Slot C Hot-Plug Switch, 0 = lever closed (board installed)
Slot B Hot-Plug Switch, 0 = lever closed (board installed)
Slot A Hot-Plug Switch, 0 = lever closed (board installed)
4
3
2
1
0
8.2.6
Hot-Plug Interrupt Mask
Address Offset:
Default Value:
0Ch
FFFFFFFFh
Size:
Attribute:
32 bits
Read/Write
This read/write mask register is used to indicate which inputs should generate interrupts and which
should not. The mask bits in this register map one-for-one with the HIICR (Interrupt Input and
Clear Register) bits. If a state change occurs on an input while the mask bit for that input is set to
one, then no interrupt will be generated for that state change. If the mask bit is cleared, then an
interrupt will be generated on the next state change.
Intel® 460GX Chipset Software Developer’s Manual
8-15
WXB Hot-Plug
Bits
31:30
29
28
27
26
25
24
23:22
21
20
19
18
17
16
15:14
13
12
11
10
9
Description
reserved (0)
Slot F PRSNT(0)#, PCI Present Signal 1
Slot E PRSNT(0)#, PCI Present Signal 1
Slot D PRSNT(0)#, PCI Present Signal 1
Slot C PRSNT(0)#, PCI Present Signal 1
Slot B PRSNT(0)#, PCI Present Signal 1
Slot A PRSNT(0)#, PCI Present Signal 1
reserved (0)
Slot F PRSNT(1)#, PCI Present Signal 2
Slot E PRSNT(1)#, PCI Present Signal 2
Slot D PRSNT(1)#, PCI Present Signal 2
Slot C PRSNT(1)#, PCI Present Signal 2
Slot B PRSNT(1)#, PCI Present Signal 2
Slot A PRSNT(1)#, PCI Present Signal 2
reserved (0)
Slot F FAULT#, PCI Power Fault Signal
Slot E FAULT#, PCI Power Fault Signal
Slot D FAULT#, PCI Power Fault Signal
Slot C FAULT#, PCI Power Fault Signal
Slot B FAULT#, PCI Power Fault Signal
Slot A FAULT#, PCI Power Fault Signal
reserved(0)
8
7:6
5
Slot F Hot-Plug Switch, 0 = lever closed (board installed)
Slot E Hot-Plug Switch, 0 = lever closed (board installed)
Slot D Hot-Plug Switch, 0 = lever closed (board installed)
Slot C Hot-Plug Switch, 0 = lever closed (board installed)
Slot B Hot-Plug Switch, 0 = lever closed (board installed)
Slot A Hot-Plug Switch, 0 = lever closed (board installed)
4
3
2
1
0
8.2.7
Serial Input Byte Data
Address Offset:
Default Value:
10h
00h
Size:
Attribute:
8 bits
Read-Only (Pwr Good Rst Only)
After the Serial Input Busy Status is read as a logic zero, the requested byte will be in the Serial
Input Data Register. The contents will be a copy of the corresponding byte in the Interrupting Input
and Clear Register (least significant) or Non-Interrupting Inputs Register (most significant).
Bits
Description
7:0
Serial Input Data Register. RO
8-16
Intel® 460GX Chipset Software Developer’s Manual
WXB Hot-Plug
8.2.8
Serial Input Byte Pointer
Address Offset:
Default Value:
11h
00h
Size:
Attribute:
16 bits
Read/Write, Read-Only
Used to input a byte into the IHPC input registers. The byte number is written to the pointer. After
the Serial Input Busy Status is read as a logic zero, the requested byte will be in the Serial Input
Data Register. The contents will be a copy of the corresponding byte in the Interrupting Input and
Clear Register (least significant) or Non-Interrupting Inputs Register (most significant).
Bits
7
Description
Serial Input Busy Status: RO
reserved (0)
6:4
3:0
Serial Input Byte Pointer. RW
8.2.9
General Purpose Output
Address Offset:
Default Value:
13h
00h
Size:
Attribute:
8 bits
Read/Write (Pwr Good Rst Only)
These bits are only driven out following SOGO/LED cycles.
Bits
7:6
Description
reserved (0)
5:0
General Purpose Output bits (one bit per supported slot). Slot F is MSB. Slot A is LSB.
8.2.10
Hot-Plug Non-interrupt Inputs
Address Offset:
Default Value:
14h
Size:
Attribute:
32 bits
Read Only
Bits
Description
reserved (0)
31:6
5
4
3
2
1
0
Slot F M66EN, PCI 66 MHz Clock Enable
Slot E M66EN, PCI 66 MHz Clock Enable
Slot D M66EN, PCI 66 MHz Clock Enable
Slot C M66EN, PCI 66 MHz Clock Enable
Slot B M66EN, PCI 66 MHz Clock Enable
Slot A M66EN, PCI 66 MHz Clock Enable
8.2.11
Hot-Plug Slot Identifier
Address Offset:
Default Value:
28h
0000h
Size:
Attribute:
16 bits
Read/Write
This register is a copy of Configuration Register 40h, byte 0.
Intel® 460GX Chipset Software Developer’s Manual
8-17
WXB Hot-Plug
8.2.12
Hot-Plug Switch Interrupt Redirect Enable
Address Offset:
Default Value:
2Ch
00h
Size:
Attribute:
8 bits
Read/Write
This register allows the slot switch change interrupts to be redirected to the SERR# instead of the
INTA#.
Bits
7:6
5
Description
reserved (0)
Slot F INTR Redirect Enable
Slot E INTR Redirect Enable
Slot D INTR Redirect Enable
Slot C INTR Redirect Enable
Slot B INTR Redirect Enable
Slot A INTR Redirect Enable
4
3
2
1
0
8.2.13
Slot Power Control
Address Offset:
Default Value:
2Dh
Size:
8 bits
Sampled at PWRGD
Attribute: Read/Write (Pwr Good Rst Only)
This register is used to power a slot without connecting it to the bus. SOGO must be set to cause
the output sequence. Slots that are connected to the bus cannot be powered down through this
register (the Slot Enable register is used instead). The set of usable Slot Power Control bits is
determined by the strapping values on the P(A,B)HSIL, P(A,B)HSOL, and P(A,B)HSOC inputs.
Unsupported slots in a system do not have writeable Slot Power Control bits.
Bits
7:6
5
Description
reserved (0)
Enable Power to Slot F. When 1, slot F is powered up. When 0, slot F is powered down
Enable Power to Slot E. When 1, slot E is powered up. When 0, slot E is powered down
Enable Power to Slot D. When 1, slot D is powered up. When 0, slot D is powered down
Enable Power to Slot C. When 1, slot C is powered up. When 0, slot C is powered down
Enable Power to Slot B. When 1, slot B is powered up. When 0, slot B is powered down
Enable Power to Slot A. When 1, slot A is powered up. When 0, slot A is powered down
4
3
2
1
0
8.2.14
Extended Hot-Plug Miscellaneous
Address Offset:
Default Value:
32h
0000h
Size:
Attribute:
16 bits
Partial Read/Write
Bits
Description
reserved (0)
15:0
8-18
Intel® 460GX Chipset Software Developer’s Manual
IFB Register Mapping
9
The IFB internal registers are organized into four Functions–LPC/FWH interface bridge, IDE
Controller, USB Host Controller, and Enhanced Power Management. Each Function has its
registers divided into 1 set of PCI Configuration Registers and one or more register sets located in
system I/O space.
Software should not map programmable memory or I/O address registers such that any part of the
range overlaps addresses decoded by other IFB devices.
Some of the IFB registers contain reserved bits. Software must deal correctly with fields that are
reserved. On reads, software must use appropriate masks to extract the defined bits and not rely on
reserved bits being any particular value. On writes, software must ensure that the values of
reserved bit positions are preserved. That is, the values of reserved bit positions must first be read,
merged with the new values for other bit positions and then written back.
In addition to reserved bits within a register, the IFB contains address locations in the PCI
configuration space that are marked “Reserved”. The IFB responds to accesses to these address
locations by completing the Host cycle. Software should not write to reserved IFB configuration
locations in the device-specific region (above address offset 3Fh).
During a hard reset, the IFB sets its internal registers to predetermined default states. The default
values are indicated in the individual register descriptions.
The following notation is used to describe register access attributes:
RO
Read Only. If a register is read only, writes have no effect.
WO
R/W
Write Only. If a register is write only, reads have no effect.
Read/Write. A register with this attribute can be read and written. Note that individual
bits in some read/write registers may be read only.
R/WC
Read/Write Clear. A register bit with this attribute can be read and written. However, a
write of a 1 clears (sets to 0) the corresponding bit and a write of a 0 has no effect.
9.1
PCI / LPC / FWH Configuration
The IFB PCI Function 0 contains a LPC/FWH interface, interrupt controller and counter / timers,
including the real time clock. The register set associated with this Functionality and associated
logic is shown below with actual register descriptions given in this section.
9.1.1
PCI Configuration Registers (Function 0)
Table 9-1. PCI Configuration Registers–Function 0(PCI to LPC/FWH Interface Bridge)
Configuration Offset
Mnemonic
VID
Register
Vendor Identification
Register Access
00–01h
02–03h
04–05h
RO
RO
DID
Device Identification
PCI Command
PCICMD
R/W
Intel® 460GX Chipset Software Developer’s Manual
9-1
IFB Register Mapping
Table 9-1. PCI Configuration Registers–Function 0(PCI to LPC/FWH Interface Bridge)
(Cont’d)
Configuration Offset
Mnemonic
PCISTS
Register
PCI Device Status
Register Access
06–07h
08h
R/W
RO
RO
–
RID
Revision Identification
Class Code
09-0Bh
0C–0Dh
0Eh
CLASSC
–
Reserved
HEDT
Header Type
RO
–
0F–3Fh
40–43h
44h
–
Reserved
ACPIBR
ACPI Base Address Register
ACPI Enable
R/W
R/W
R/W
–
ACPIEN
45h
SCIRC
Reserved
46–4Bh
4Ch
–
Reserved
–
Reserved
R/W
–
4Dh
–
Reserved
4E–4Fh
50–5Fh
60–63h
64h
BIOSEN
Reserved
R/W
–
–
Reserved
PIRQRC[A:D]
Reserved
R/W
R/W
–
SERIRQC
Serial IRQ Control
Reserved
65–68h
69h
–
TOM
Top of Memory
Miscellaneous Status
Reserved
R/W
R/W
–
6A–6Bh
6C–75h
76–77h
78–7Bh
7C–7Fh
80h
MSTAT
–
–
Reserved
R/W
R/W
–
–
Reserved
–
Reserved
–
APIC Base Address Relocation
Reserved
R/W
–
81h
–
82h
DLC
Deterministic Latency Control
Reserved
R/W
–
83h
–
84–85h
86–8Fh
90-91h
92-95h
96-C7h
C8h
MGPIOC
–
Muxed GPIO Control
Reserved
R/W
–
PDMACFG
DDMABASE
–
PCI DMA Configuration
Distributed DMA Slave Base Pointer
Reserved
R/W
R/W
–
RTCCFG
–
Real Time Clock Configuration
Reserved
R/W
–
C9–CFh
D0–D3h
D4h
GPIOBA
GPIOE
–
GPIO Base Address Register
GPIO Enable
R/W
R/W
–
D5–DFh
E0h
Reserved
LPCCD
LPCFD
LPCSD
FWHDE
LPCGD
LPCDE
FWHS
–
LPC COM Decode
LPC FDD/LPT Decode
LPC Sound Decode
Firmware Hub Decode Enable
LPC Generic Decode Range
LPC Decode Enables
Firmware Hub Select
Reserved
R/W
R/W
R/W
–
E1h
E2h
E3h
E4–E5h
E6–E7h
E8h
R/W
R/W
–
E9–FFh
–
9-2
Intel® 460GX Chipset Software Developer’s Manual
IFB Register Mapping
9.2
IDE Configuration
The IFB PCI function 1 contains an IDE Controller capable of standard Programmed I/O (PIO)
transfers as well as Bus Master transfer capability. It also supports the “Ultra DMA/33”
synchronous DMA mode of data transfer.
9.2.1
PCI Configuration Registers (Function 1)
Table 9-2. PCI Configuration Registers–Function 1 (IDE Interface)
Configuration Offset
Mnemonic
VID
Register
Vendor Identification
Register Access
00–01h
02–03h
04–05h
06–07h
08h
RO
RO
R/W
R/W
RO
RO
–
DID
Device Identification
PCI Command
PCI Device Status
Revision Identification
Class Code
PCICMD
PCISTS
RID
09-0Bh
0Ch
CLASSC
–
Reserved
0Dh
MLT
Master Latency Timer
Header Type
R/W
RO
–
0Eh
HEDT
0F–1Fh
20–23h
24–3Fh
40–43h
44h
–
Reserved
BMIBA
Bus Master Interface Base Address
Reserved
R/W
–
–
IDETIM
IDE Timing
R/W
R/W
–
SIDETIM
Slave IDE Timing
Reserved
45–47h
48h
–
SDMACTL
Synchronous DMA Control
Reserved
R/W
–
49h
–
4A–4Bh
4C–F7h
F8-FBh
FC-FFh
SDMATIM
Synchronous DMA Timing
Reserved
R/W
–
–
---
---
Manufacturer’s ID
Reserved
---
---
Intel® 460GX Chipset Software Developer’s Manual
9-3
IFB Register Mapping
9.3
Universal Serial Bus (USB) Configuration
The IFB integrates an USB Controllers. The USB Controller is UHCI 1.1 compliant. It implements
the root hub of the USB, which contains two ports.
The IFB PCI function 2 reflects both the Universal Serial Bus Host and Root Hubs, with 2
connected USB ports.
9.3.1
PCI Configuration Registers (Function 2)
Table 9-3. PCI Configuration Registers–Function 2 (USB Interface)
Configuration Offset
Mnemonic
VID
Register
Vendor Identification
Register Access
00–01h
02–03h
04–05h
06–07h
08h
RO
RO
R/W
R/W
RO
RO
–
DID
Device Identification
PCI Command
PCICMD
PCISTS
PCI Device Status
Revision Identification
Class Code
RID
09-0Bh
0Ch
CLASSC
–
Reserved
0Dh
MLT
Latency Timer
R/W
RO
–
0Eh
HEDT
Header Type
0F–1Fh
20–23h
24–3Bh
3Ch
–
Reserved
USBBA
USB I/O Space Base Address
Reserved
R/W
–
–
INTLN
Interrupt Line
R/W
RO
–
3Dh
INTPN
Interrupt Pin
3E–5Fh
60h
–
Reserved.
SBRNUM
Serial Bus Release Number
Reserved
RO
–
61–69h
6A–6Bh
6C–BFh
C0–C1h
C2-C3h
C4h
–
MCR
Miscellaneous Control Register
Reserved
R/W
–
–
LEGSUP
USB Legacy Keyboard/Mouse Control
Reserved
R/W
-----
R/W
–
---
USBRES
USB Resume Enable
Reserved
C5-F3h
F4-F7h
F8-FBh
FC-FFh
–
–
–
–
Reserved
–
Manufacturer’s ID Register
Reserved
–
–
9-4
Intel® 460GX Chipset Software Developer’s Manual
IFB Register Mapping
9.4
SMBus Controller Configuration
The IFB PCI function 3 contains the SMBus Controller configuration space.
9.4.1
SMBus Configuration Registers (Function 3)
Table 9-4. PCI Configuration Registers–Function 3 (SMBus Controller Interface)
Configuration Offset
Mnemonic
VID
Register
Vendor Identification
Register Access
00–01h
02–03h
04–05h
06–07h
08h
RO
RO
R/W
R/WC
RO
RO
–
DID
Device Identification
PCI Command
PCICMD
PCISTS
PCI Device Status
Revision Identification
Class Code
RID
09-0Bh
0C-1Fh
20-23h
24-3Bh
3Ch
CLASSC
–
Reserved
BAR
Base Address Register
Reserved
R/W
–
–
IL
Interrupt Line
RW
RO
–
3Dh
IP
Interrupt Pin
3E-3Fh
40h
–
Reserved
HC
Host Configuration
Slave Command Port
Slave Shadow Address 1
Slave Shadow Address 2
Reserved
RW
RW
RW
RW
–
41h
SCOM
42h
SS1
SS2
–
43h
41-F3h
F4-F7h
F8-FBh
FC-FFh
–
Reserved
–
–
Manufacturer’s ID Register
Reserved
–
–
–
Intel® 460GX Chipset Software Developer’s Manual
9-5
IFB Register Mapping
9-6
Intel® 460GX Chipset Software Developer’s Manual
IFB Usage Considerations
10
This section talks about the normal usage for some of the features in the IFB component.
10.1
10.2
Usage of 1MIN Timer in Power Management
IFB does not support the global standby timer concept. The determination of a system inactivity
can be done by using the 1MIN Timer that can be used by the SMI handler to generate an SMI#
every minute. The SMM handler can check all the appropriate power management status registers
in IFB to see if there is any system inactivity. After n minutes of no system activity (where n is
determined by the SMM handler, time based setup option for the user to indicate system inactivity),
the SMM handler can decide to put the system into a lower power state.
Usage of the SW SMI# Timer
The Software SMI# Timer provides a mechanism for the SMM code to temporarily exit SMM
space. This is required when some application code is required to complete before the SMM code
can complete. An example of this may be when a Soft-Off or Suspend-To-Disk is requested, and
some application code must complete.
After starting the timer, the SMM code exits. When the timer expires, another SMI# is generated.
At that point, the SMM code has the option to restart the timer, if needed.
If the SMM handler desires, it can halt the timer by setting its enable bit to 0. This will prevent the
SMI# generation at the next time-out.
10.3
10.4
CD-ROM AUTO RUN Feature of the OS
Whenever a CD-ROM controller is detected in a system, the OS may use the auto run feature if
enabled to indicate to the user whether a CD-ROM has been plugged-in into the CD-ROM
controller. The OS driver issues a Test Unit Ready (TUR) command to check for the insertion of
media. The firmware or the OS driver needs to ensure that the IRQ on which the CD-ROM
controller is being serviced is not being monitored with the help of the enabled Wake/Break enable
registers. Also the PM timers available in the IFB should be used appropriately before the firmware
goes to the low power state. All these changes need to be done if the system wishes to use the auto
run feature in the OS and still be able to use the power management features of the IFB.
ACPI, SMBus, GPIO Base Address Reporting to the
OS
The firmware should report all of the register space that is being used by the ACPI, SMBus and
GPIO functionalities to the OS by using the PNP device node entries. They should be declared as
Plug and Play motherboard resources with the Device ID of PNP0C02. If the ACPI tables are used
instead of the PNP Device node entries for reporting resources the change has to be incorporated
Intel® 460GX Chipset Software Developer’s Manual
10-1
IFB Usage Considerations
into the system firmware by the vendor. This reporting will make these register locations safe and
the OS will not use these locations randomly if a PNP conflicting device is relocatable in those I/O
or memory locations. These locations also got to be reported to the OS whenever the OEM sends
the systems for their WHQL suite test.
10.5
Ultra DMA Configuration
The following registers are programmed in systems that contain devices that implement the Ultra
DMA Protocol. These registers allow Ultra DMA to be used when PCI Bus Master IDE operation
is initiated by the device driver.
10.5.1
UDMAC–Ultra DMA Control Register (IFB Function 1 PCI
Configuration Offset 48h)
7
6
5
4
3
2
1
0
Reserved
Secondary Secondary Primary
Primary
Drive 0
Ultra DMA
Mode
Drive 1
Ultra DMA
Mode
Drive 0
Ultra DMA
Mode
Drive 1
Ultra DMA
Mode
Enable
Enable
Enable
Enable
(SSDE1)
(SSDE0)
(PSDE1)
(PSDE0)
0: Disabled 0: Disabled 0: Disabled 0: Disabled
1: Enabled 1: Enabled 1: Enabled 1: Enabled
10.5.2
UDMATIM–Ultra DMA Timing Register (IFB Function 1 PCI
Configuration Offsets 4A-4Bh)
15
14
13
12
11
10
9
8
Reserved
Secondary Drive 1
Reserved
Secondary Drive 0
Ultra DMA Cycle Time
Ultra DMA Cycle Time
(SCT1)
(SCT1)
00: CT=4 clks, RP=6 clks
01: CT=3 clks, RP=5 clks
10: CT=2 clks, RP=4 clks
11: Reserved
00: CT=4 clks, RP=6 clks
01: CT=3 clks, RP=5 clks
10: CT=2 clks, RP=4 clks
11: Reserved
7
6
5
4
3
2
1
0
Reserved
Primary Drive 1
Reserved
Primary Drive 0
Ultra DMA Cycle Time
Ultra DMA Cycle Time
(PCT1)
(PCT1)
00: CT=4 clks, RP=6 clks
01: CT=3 clks, RP=5 clks
10: CT=2 clks, RP=4 clks
11: Reserved
00: CT=4 clks, RP=6 clks
01: CT=3 clks, RP=5 clks
10: CT=2 clks, RP=4 clks
11: Reserved
10-2
Intel® 460GX Chipset Software Developer’s Manual
IFB Usage Considerations
NOTES:
• The Ultra DMA Enable bit specifies the current Ultra DMA enabled status:
— Disabled by default: This field needs to be enabled in order to take advantage of the IFB
Ultra DMA timings. When this field is disabled, the IFB Ultra DMA Timing Register is
disabled.
• The Ultra DMA Cycle Time Field specifies the current Ultra DMA timing mode.
Note that this field only applies if the corresponding Ultra DMA Enable field is set.
10.5.3
Determining a Drive’s Transfer Rate Capabilities
10.5.3.1
Overview
The following section provides the information that allows a drive’s capabilities to be determined.
The ATA IDENTIFY_DRIVE (0xEC) and ATAPI IDENTIFY_DEVICE(0x1A) can be used to
determine the capability. Each of these commands returns a 256 word buffer with fields that can
allow the drive’s capabilities to be determined.
Refer to the ATA Specification for more information regarding the IDENTIFY_DEVICE
command.
Table 10-1. Identify Device Information Used for Determining Drive Capabilities
Word
Offset
Capability
Bits
Field
Device Type
Device Type
Fields that Indicate Device Type
0
15
2
General Configuration:
0: ATA Device
1: ATAPI Device
Ultra DMA
Fields that Indicate Ultra DMA Drive Capabilities
Ultra DMA
53
88
Field Validity
0: the fields reported in word 88 are not valid
1: the fields reported in word 88 are valid
Ultra DMA
Ultra DMA
PIO, DMA
10:8
2:0
1
Ultra DMA Modes Active*
bit 10: 1: Ultra DMA Mode 2 is active
0: Ultra DMA Mode 2 is not active
bit 9: 1: Ultra DMA Mode 1 is active
0: Ultra DMA Mode 1 is not active
bit 8: 1: Ultra DMA Mode 0 is active
0: Ultra DMA Mode 0 is not active
88
53
Ultra DMA Modes Supported*
bit 2: 1: Ultra DMA Mode 2 is supported
0: Ultra DMA Mode 2 is not supported
bit 1: 1: Ultra DMA Mode 1 is supported
0: Ultra DMA Mode 1 is not supported
bit 0: 1: Ultra DMA Mode 0 is supported
0: Ultra DMA Mode 0 is not supported
Field Validity
0: the fields reported in words 64-70 are not valid
1: the fields reported in word 64-70 are valid
Intel® 460GX Chipset Software Developer’s Manual
10-3
IFB Usage Considerations
Table 10-1. Identify Device Information Used for Determining Drive Capabilities (Cont’d)
Word
Offset
Capability
DMA
Bits
10:8
Field
62
Single Word DMA Modes Active
bit 10: 1: Single Word DMA Mode 2 is active
0: Single Word DMA Mode 2 is not active
bit 9: 1: Single Word DMA Mode 1 is active
0: Single Word DMA Mode 1 is not active
bit 8: 1: Single Word DMA Mode 0 is active
0: Single Word DMA Mode 0 is not active
DMA
DMA
DMA
62
63
63
2:0
Single Word DMA Modes Supported
bit 2: 1: Single Word DMA Mode 2 is supported
0: Single Word DMA Mode 2 is not supported
bit 1: 1: Single Word DMA Mode 1 is supported
0: Single Word DMA Mode 1 is not supported
bit 0: 1: Single Word DMA Mode 0 is supported
0: Single Word DMA Mode 0 is not supported
10:8
Multi Word DMA Modes Active
bit 10: 1: Multi Word DMA Mode 2 is active
0: Multi Word DMA Mode 2 is not active
bit 9: 1: Multi Word DMA Mode 1 is active
0: Multi Word DMA Mode 1 is not active
bit 8: 1: Multi Word DMA Mode 0 is active
0: Multi Word DMA Mode 0 is not active
2:0
Multi Word DMA Modes Supported
bit 2: 1: Multi Word DMA Mode 2 is supported
0: Multi Word DMA Mode 2 is not supported
bit 1: 1: Multi Word DMA Mode 1 is supported
0: Multi Word DMA Mode 1 is not supported
bit 0: 1: Multi Word DMA Mode 0 is supported
0: Multi Word DMA Mode 0 is not supported
DMA
PIO
65
51
15:0
15:8
Minimum Multi Word DMA Transfer Cycle Time Per Word
PIO Data Transfer Cycle Timing Mode Supported
00h: PIO0
01h: PIO1
02h: PIO2
PIO, DMA
PIO
53
64
68
1
Field Validity
0: the fields reported in words 64-70 are not valid
1: the fields reported in word 64-70 are valid
7:0
15:0
Advanced Flow Control PIO Transfer Modes Supported
bit 0: PIO3 (w/IORDY Flow Control)
bit 1: PIO4 (w/IORDY Flow Control)
PIO
Minimum PIO Transfer Cycle Time with IORDY Flow Control
For IFB IDE Timing Configuration, each of the following things must be determined:
• Drive Type: ATAPI or ATA (non-ATAPI)
• Best DMA Capability
• Best Ultra DMA Capability OR
• Best Multi Word DMA Capability (if Ultra DMA not supported) OR
• Best Single Word DMA Capability (if neither Ultra DMA nor Multi Word DMA supported)
• Best PIO Capability
10-4
Intel® 460GX Chipset Software Developer’s Manual
IFB Usage Considerations
10.5.4
Determining a Drive’s Best Ultra DMA Capability
The drive’s ultra DMA mode capability and current configuration are specified in the
IDENTIFY_DRIVE buffer, Word 88. Software must first check to see that the Word 88 is valid
before determining the Ultra DMA drive capability.
Table 10-2. Identify Device Information Used for Determining Ultra DMA Drive Capabilities
Word
Offset
Capability
Ultra DMA
Bits
Field
53
88
2
Field Validity
0: the fields reported in word 88 are not valid
1: the fields reported in word 88 are valid
Ultra DMA
Ultra DMA
10:8
Ultra DMA Modes Active*
bit 10: 1: Ultra DMA Mode 2 is active
0: Ultra DMA Mode 2 is not active
bit 9: 1: Ultra DMA Mode 1 is active
0: Ultra DMA Mode 1 is not active
bit 8: 1: Ultra DMA Mode 0 is active
0: Ultra DMA Mode 0 is not active
88
2:0
Ultra DMA Modes Supported*
bit 2: 1: Ultra DMA Mode 2 is supported
0: Ultra DMA Mode 2 is not supported
bit 1: 1: Ultra DMA Mode 1 is supported
0: Ultra DMA Mode 1 is not supported
bit 0: 1: Ultra DMA Mode 0 is supported
0: Ultra DMA Mode 0 is not supported
The following Ultra DMA drive capabilities are supported by the IFB, from fastest to slowest:
• Ultra DMA Mode 2
• Ultra DMA Mode 1
• Ultra DMA Mode 0
• Disabled (Drive does not support any of the above Ultra DMA Modes.)
10.5.5
Determining a Drive’s Best Multi Word DMA/Single Word
DMA (Non-ultra DMA) Capability
This section describes how to determine a drive’s multi word DMA and single word DMA
capabilities.
The following DMA drive capabilities are supported by the IFB, from fastest to slowest:
• Multi Word DMA Mode 2
• Multi Word DMA Mode 1
• Single Word DMA Mode 2
• Disabled (Drive does not support any of the above Multi/Single DMA Modes.)
Software at this stage needs to determine if at least one of the above modes is supported by the
drive. Software should initially determine a drive’s best Mult Word DMA capability initially. If a
drive doesn’t support multi word DMA Modes 0 or 1, then software should check if single word
DMA Mode 2 is supported.
Intel® 460GX Chipset Software Developer’s Manual
10-5
IFB Usage Considerations
The drive’s multi word DMA mode capability and current configuration are specified in the
IDENTIFY_DRIVE buffer, Words 63 and 65 Software must first check to see that the Words 64-70
are valid before determining the drive’s multi word DMA drive capability.
The drive’s single word DMA mode capability and current configuration are specified in the
IDENTIFY_DRIVE buffer, Word 62.
Table 10-3. Identify Device Information Used for Determining Multi/Single Word DMA Drive
Capabilities
Capability
PIO, DMA
Word Offset
Bits
Field
53
1
Field Validity
0: the fields reported in words 64-70 are not valid
1: the fields reported in word 64-70 are valid
DMA
DMA
DMA
DMA
DMA
62
62
63
63
65
10:8
Single Word DMA Modes Active
bit 10: 1: Single Word DMA Mode 2 is active
0: Single Word DMA Mode 2 is not active
bit 9: 1: Single Word DMA Mode 1 is active
0: Single Word DMA Mode 1 is not active
bit 8: 1: Single Word DMA Mode 0 is active
0: Single Word DMA Mode 0 is not active
2:0
Single Word DMA Modes Supported
bit 2: 1: Single Word DMA Mode 2 is supported
0: Single Word DMA Mode 2 is not supported
bit 1: 1: Single Word DMA Mode 1 is supported
0: Single Word DMA Mode 1 is not supported
bit 0: 1: Single Word DMA Mode 0 is supported
0: Single Word DMA Mode 0 is not supported
10:8
Multi Word DMA Modes Active
bit 10: 1: Multi Word DMA Mode 2 is active
0: Multi Word DMA Mode 2 is not active
bit 9: 1: Multi Word DMA Mode 1 is active
0: Multi Word DMA Mode 1 is not active
bit 8: 1: Multi Word DMA Mode 0 is active
0: Multi Word DMA Mode 0 is not active
2:0
Multi Word DMA Modes Supported
bit 2: 1: Multi Word DMA Mode 2 is supported
0: Multi Word DMA Mode 2 is not supported
bit 1: 1: Multi Word DMA Mode 1 is supported
0: Multi Word DMA Mode 1 is not supported
bit 0: 1: Multi Word DMA Mode 0 is supported
0: Multi Word DMA Mode 0 is not supported
15:0
Minimum Multi Word DMA Transfer Cycle Time per
Word
To determine the drive’s best DMA capability ensure that the drive reports the capability and is
10-6
Intel® 460GX Chipset Software Developer’s Manual
IFB Usage Considerations
Table 10-4. Drive Multi Word DMA/Single Word DMA Capability as a Function of Cycle Time
Drives Reported DMA
Drive’s Reported DMA Mode Setting
Drive’s Best DMA Mode
1
Cycle Time
Multi Word DMA Mode 2
Multi Word DMA Mode 2
Multi Word DMA Mode 2
Multi Word DMA Mode 2
Multi Word DMA Mode 1
Multi Word DMA Mode 1
Multi Word DMA Mode 1
Multi Word DMA Mode 0
Single Word DMA Mode 2
Single Word DMA Mode 2
Single Word DMA Mode 1
Single Word DMA Mode 0
Drive does not support DMA transfers.
t <= 120ns
Multi Word DMA Mode 2
Multi Word DMA Mode 1
Single Word DMA Mode 2
N/A (Disabled)
120 < t <= 180ns
180 < t <= 240ns
t > 240ns
t <= 180ns
180 < t <= 240ns
t > 240ns
N/A
Multi Word DMA Mode 1
Single Word DMA Mode 2
N/A (Disabled)
N/A (Disabled)
t <= 240ns
t > 114ns
N/A
Single Word DMA Mode 2
N/A (Disabled)
N/A (Disabled)
N/A
N/A (Disabled)
N/A
N/A
NOTE: Timing cycle times are defined by the ATA specification. A device that reports a given DMA mode
capability must be capable of supporting the minimum DMA cycle time.
A drive’s Multi Word DMA or Single Word DMA speed is the fastest Multi Word DMA speed, if
supported, that is supported for the drive. If Multi Word DMA is NOT supported for the reported
timing cycle time then the drive’s best non Ultra DMA speed is based on its Single Word DMA 2
capability.
If N/A is used, the drive is configured for PIO-only. The PIIX mode follows the PIO mode only. If
a drive does not report a DMA cycle time that is consistent with the Target DMA Cycle Time, a
slower speed should be chosen.
The cycle times associated with the various timing modes are defined by the ATA Specification. A
drive that reports a given Multi/Single Word DMA capability must be capable of supporting the
minimum cycle time for that mode.
10.5.5.1
Determining a Drive’s Best PIO Capability
This section describes how to determine a drive’s PIO Capabilities.
The following PIO drive capabilities are supported by the IFB, from fastest to slowest:
• PIO4 w/IORDY
• PIO3 w/IORDY
• PIO2 w/IORDY
• PIO2 (without IORDY)
• Compatible (Drive does not support any of the above PIO Modes.)
Intel® 460GX Chipset Software Developer’s Manual
10-7
IFB Usage Considerations
Software at this stage needs to determine if at least one of the above modes is supported by the
drive. Software should initially determine a drive’s best PIO w/IORDY capability (PIO4 w/IORDY
or PIO3 w/IORDY) initially. If these PIO w/IORDY modes are not supported, the drive should
determine the PIO2 mode support with IORDY or PIO2 mode support without IORDY. Otherwise,
Compatible timings should be applied to the drive.
The drive’s PIO w/IORDY mode capability is specified in the IDENTIFY_DRIVE buffer, Words
64 and 68. Software must first check to see that the Words 64-70 are valid before determining the
drive’s PIO w/IORDY drive capability.
The drive’s PIO2 mode capability and current configuration are specified in the
IDENTIFY_DRIVE buffer, Word 51.
Table 10-5. Identify Device Information Used for Determining PIO Drive Capabilities
Capability
Word Offset
Bits
Field
PIO
51
15:8
PIO Data Transfer Cycle Timing Mode Supported
00h: PIO0
01h: PIO1
02h: PIO2
PIO, DMA
PIO
53
64
68
1
Field Validity
0: the fields reported in words 64-70 are not valid
1: the fields reported in word 64-70 are valid
7:0
Advanced Flow Control PIO Transfer Modes Supported
bit 0: PIO3 (w/IORDY Flow Control)
bit 1: PIO4 (w/IORDY Flow Control)
PIO
15:0
Minimum PIO Transfer Cycle Time with IORDY Flow Control
To determine the drive’s best PIO capability ensure that the drive reports the capability and is able
to transfer at the target cycle time:
Table 10-6. Drive PIO Capability as a Function of Cycle Time
Drives Reported PIO
Drive’s Reported PIO Mode Setting
Drive’s Best PIO Mode
Cycle Time
PIO4
PIO4
t <= 120ns
120 < t <= 180ns
180 < t <= 240ns
t > 240ns
PIO4
PIO3
PIO4
PIO2
PIO4
PIO0/Compatible
PIO3
PIO3
t <= 180ns
PIO3
180 < t <= 240ns
t > 240ns
PIO2
PIO3
PIO0/Compatible
PIO2
PIO2
N/A (drive must support t<= 240ns)
N/A
PIO1
PIO0/Compatible
PIO0/Compatible
PIO0/Compatible
N/A
Note: The cycle times associated with the various timing modes are defined by the ATA Specification. A
drive that reports a given PIO capability must be capable of supporting the minimum cycle time for
that mode.
Note: If a drive does not report a PIO cycle time that is consistent with the Target PIO Cycle Time, a
slower speed should be chosen.
10-8
Intel® 460GX Chipset Software Developer’s Manual
IFB Usage Considerations
10.5.6
IFB Timing Settings
10.5.6.1
DMA/PIO Timing Settings
Ultra DMA mode settings are completely independent of the following timings.
Table 10-7. IFB Drive Mode Based on DMA/PIO Capabilities
Fast PIO
Non Ultra DMA
Supported?
Best DMA
Mode is {SW2, Only Select Enable Select
MW1, MW2}
DMA
Timing
Enable
Supported?
Best PIO
Mode >=
Best DMA
Mode
IORDY
Sample
Point Enable
Select
Drive’s Best
DMA
Capability
Drive’s Best
PIO
Capability
IFB
Timing
Mode
Prefetch and
Posting
Fast Timing
Bank Drive
Select
a
N/A
(DMA not
supported)
PIO0/1/
Compatible
Mode 0
Mode 2
Mode 3
Mode 4
Mode 2
No
No
No
No
No
Yes
Disabled
Disabled
Disabled
Disabled
Enabled
Enabled
(if fixed disk)
Disabled
Disabled
N/A
(DMA not
supported)
PIO2
Yes
Yes
Yes
Enabled
(if fixed disk) Drive
Dependson Enabled
N/A
(DMA not
supported)
PIO3
(w/IORDY)
Enabled
(if fixed disk)
Enabled
Enabled
Enabled
N/A
(DMA not
supported)
PIO4
(w/IORDY)
Enabled
(if fixed disk)
Enabled
Single Word PIO0/1/
DMA Mode 2 Compatible
No (special
config.
Enabled
(if fixed disk) Drive
Dependson Enabled
needed)
Single Word PIO2
DMA Mode 2
Mode 2
Mode 2
Mode 2
Mode 3
Yes
Yes
Yes
No
No
No
Yes
Disabled
Disabled
Disabled
Enabled
Enabled
(if fixed disk) Drive
Dependson Enabled
Single Word PIO3
DMA Mode 2 (w/IORDY)
Enabled
(if fixed disk) Drive
Dependson Enabled
Single Word PIO4
DMA Mode 2 (w/IORDY)
Enabled
(if fixed disk) Drive
Dependson Enabled
Multi Word
PIO0/1/
No (special
config.
needed)
Enabled
(if fixed disk)
Enabled
Enabled
Enabled
DMA Mode 1 Compatible/
PIO2
Multi Word
PIO3
Mode 3
Mode 4
Yes
No
Disabled
Enabled
Enabled
(if fixed disk)
Enabled
Enabled
DMA Mode 1 (w/IORDY)/
PIO4
(w/IORDY)
Multi Word
PIO0/1/
No (special
config.
needed)
Yes
Enabled
(if fixed disk)
Enabled
DMA Mode 2 Compatible/
PIO2
Multi Word
DMA Mode 2 (w/IORDY)
PIO3
Mode 3
Mode 4
Yes
No
No
Disabled
Disabled
Enabled
(if fixed disk)
Enabled
Enabled
Enabled
Enabled
Multi Word PIO4
Yes
Enabled
DMA Mode 2 (w/IORDY)
(if fixed disk)
a. Configurations where a drive reports a PIO speed much slower than its reported DMA speed require the DMA Timing Enable Only Select bit to be
Enabled.
Intel® 460GX Chipset Software Developer’s Manual
10-9
IFB Usage Considerations
Configurations where a drive reports a PIO speed much slower than its reported DMA speed
require the DMA Timing Enable Only Select bit to be Enabled.
Table 10-8. IDE Mode/Drive Feature Settings for Optimal DMA/PIO Operation
Slave IDE
Timing
Register
Value
IDE
IDE
DMA
Timing
Enable
Only
IORDY
Sample
Point
Fast
Timing
Bank
Drive
Select 0
DMA
Timing
Enable
Only
IORDY
Sample
Point
Enable
Select 1
Fast
Timing
Bank
Drive
Select 1
IFB IDE
Drive 0
Mode
IFB IDE
Drive 1
Mode
Pre-Fetch
and Posting
Enable
Pre-Fetch
andPosting
Enable
Timing
Register
Value
bits15:8
(hex)
Timing
Register
Value
bits 7:0
(binary)
(hex)
bits 3:0
(Primary)
OR
bits 7:4
(Second-
ary)
Enable
Settings Settings
Select 0
Select 1
a
b
Select 0
Select 0
Select 1
c
All
speeds
All
speeds
Mode 4
Mode 4
Mode 4
Mode 4
Mode 4
Mode 3
Mode 0
Depends
on Drive
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Disabled
Disabled
Enabled
Disabled
Disabled
Depends
Disabled
Disabled
Enabled
Enabled
Enabled
Disabled
A3
A3
E3
E3
E3
A1
0
0
4
9
B
0
0x00xx11
0000xx11
xxx1xx11
xx11xx11
xx11xx11
0x00xx11
(if fixed disk)
(if fixed disk)
Not
Present
Depends
on Drive
Enabled
(if fixed disk)
Disabled
Mode 2
Mode 3
Mode 4
Mode 0
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Enabled
(if fixed disk) on Drive
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Enabled
(if fixed disk)
Enabled
Enabled
Disabled
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Enabled
(if fixed disk)
Disabled
Enabled
(if fixed disk)
Mode 3
Mode 3
Mode 3
Mode 3
Mode 2
Mode 2
Mode 2
Mode 2
Mode 2
Mode 0
Mode 0
Mode 0
Mode 0
Mode 0
Not
Depends
on Drive
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Enabled
Disabled
Disabled
Disabled
Disabled
Disabled
Disabled
Disabled
Disabled
Depends
Disabled
Enabled
Enabled
Enabled
Disabled
Disabled
Enabled
Enabled
Enabled
Disabled
Disabled
Enabled
Enabled
Enabled
A1
E1
E1
E1
90
0
4
9
B
0
0
4
9
B
0
0
4
9
B
0000xx11
xxx1xx11
xx11xx11
xx11xx11
0x00xxx1
0000xxx1
xxx1xxx1
xx11xxx1
xx11xxx1
Present
(if fixed disk)
Mode 2
Mode 3
Mode 4
Mode 0
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Enabled
(if fixed disk) on Drive
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Enabled
(if fixed disk)
Enabled
Enabled
Disabled
Disabled
Depends
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Disabled
Disabled
Enabled
(if fixed disk)
Not
Present
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Disabled
Enabled
90
Mode 2
Mode 3
Mode 4
Mode 0
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Depends
on Drive
D0
D0
D0
80
(if fixed disk) on Drive
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Depends
on Drive
Enabled
(if fixed disk)
Enabled
Enabled
Disabled
Disabled
Depends
Depends
on Drive
Enabled
(if fixed disk)
Depends
on Drive
Depends
on Drive
Enabled
(if fixed disk)
Disabled
Disabled
Disabled
Disabled
Disabled
Enabled
(if fixed disk)
Disabled
Disabled
Disabled
Disabled
Disabled
Disabled
Disabled
Enabled
(if fixed disk)
0000000
0
Not
Present
Enabled
(if fixed disk)
Disabled
Enabled
80
0000000
0
Mode 2
Mode 3
Mode 4
Enabled
(if fixed disk)
Depends
on Drive
C0
C0
C0
xxx10000
xx110000
xx110000
(if fixed disk) on Drive
Enabled
(if fixed disk)
Depends
on Drive
Enabled
(if fixed disk)
Enabled
Enabled
Enabled
(if ifxed disk)
Depends
on Drive
Enabled
(if fixed disk)
abc.
NOTES:
1. DMA Timing Enable Only field is in general Disabled. It is only Enabled in certain cases if the DMA Mode capability of the drive
is much greater than the PIO Mode capability of the drive.
2. The IORDY Sample Point field must be Enabled for PIO Modes 3 and 4. It is Enabled on PIO2 drives if and only if IORDY
capability is supported in the drive.
3. The above recommendations assume that if the attached slave drive is Mode 0 or not present, SITRE bit is ’0’.
10-10
Intel® 460GX Chipset Software Developer’s Manual
IFB Usage Considerations
Table 10-9. DMA/PIO Timing Values Based on PIIX Cable Mode/System Speed
IDETIMx
Value
Drive 0
(Master) if
no Slave
Attached or
Slave is
Mode 0
SIDETIM
Value
Drive 1
(Slave)
bits 3:0
(Primary) or
bits 7:4
(Secondary)
IDETIMx
Value
Drive 0
(Master)
if Slave
Attached
bits 15:8
IORDY
Sample
Point (ISP)
bits 1:0
Resultant Cycle
Time (Total
Clocks Base
Recovery
Time (RCT)
bits 1:0
IFB Drive
Mode
Operating Freq)
bits 15:8
PIO0/
Default
Default
4 clocks
3 clocks
1 clock
C0h
80h
90h
A1h
A3h
0
30 MHz: 900ns
33 MHz: 900ns
Compatible
PIO2/SW2
4 clocks
3 clocks
3 clocks
D0h
E1h
E3h
4
9
B
30 MHz: 256ns
33 MHz: 240ns
PIO3/MW1
PIO4/MW2
30 MHz: 198ns
33 MHz: 180ns
30 MHz: 132ns
33 MHz: 120ns
10.5.6.2
Ultra DMA Timing Settings
The following settings apply to Ultra DMA Mode Settings only.
Table 10-10. Ultra DMA Timing Value Based on Drive Mode
SDMAC Value
Ultra DMA Mode Enable x:
Drive 0: bit 0
SDMATIMx Value
Ultra DMA Cycle Time x:
Drive 0: bits 1:0
DMA Speed Used on
DMA Based Data
Transfer Commands
IFB Drive Mode
Drive 1: bit 1
Drive 2: bit 2
Drive 1: bits 5:4
Drive 2: bits 9:8
Drive 3: bit 3
Drive 3: bits 13:12
N/A (Ultra DMA Not Non-ultra DMA
Disabled
Default
Supported)
if supported
Ultra DMA Mode 0
Ultra DMA Mode 1
Ultra DMA Mode 2
Ultra DMA Mode 0
Ultra DMA Mode 1
Ultra DMA Mode 2
Enabled
Enabled
Enabled
00b: CT=4 clks, RP=6 clks
01b: CT=3 clks, RP=5 clks
10b: CT=2 clks, RP=4 clks
10.5.7
Drive Configuration for Selected Timings
Once the IFB Timing Modes for DMA, PIO and Ultra DMA have been selected, the Set Features
Command (0 x EF) with Set Transfer Mode (subcommand 0 x 03) can be issued to set the drives on
the system to the optimal speeds:
Intel® 460GX Chipset Software Developer’s Manual
10-11
IFB Usage Considerations
Table 10-11. Ultra DMA/Multi Word DMA/Single Word Transfer/Mode Values
ATA SET_FEATURES -
Command Set Transfer
Mode Sub Command
Parameter for Selected
Speed
Drive’s Selected Ultra
Drive’s Selected Non-
ultra DMA Capability
Selected Speed
DMA Capability
Ultra DMA Mode 2
any
any
any
Ultra DMA Mode 2
Ultra DMA Mode 1
Ultra DMA Mode 0
42h
41h
40h
22h
21h
12h
Ultra DMA Mode 1
Ultra DMA Mode 0
N/A
N/A
N/A
Multi Word DMA Mode 2 Multi Word DMA Mode 2
Multi Word DMA Mode 1 Multi Word DMA Mode 1
Single Word DMA Mode
2
Single Word DMA Mode
2
N/A
N/A
Disabled
N/A
Refer to the Set Features Command description in the ATA Specification for more information.
Table 10-12. PIO Transfer/Mode Values
ATA SET_FEATURES -Command Set Transfer Mode
Sub Command Parameter for Selected Speed
Drive’s Selected PIO Speed Capability
PIO0/PIO1/PIO2/Compatible
PIO3 w/IORDY Flow Control
PIO4 w/IORDY Flow Control
N/A
C3
C4
A drive may only be enabled for a Single DMA capability. In general, if a drive supports a
supported Ultra DMA speed, then Ultra DMA is configured for the drive. If a drive does not
support Ultra DMA, then it should be configured with it’s selected DMA speed, if it exists. If a
drive supports only PIO (does not have support for either Ultra DMA or DMA speeds), then the
drive shall only be accessed in a PIO mode only.
If a drive is configured for an Ultra DMA speed or a DMA speed, its corresponding DMA-capable
bit in the PCI Bus Master I/O Status Register (Primary: PCI Bus Master IDE I/O Offset + 02h;
Secondary - PCI Bus Master I/O Offset + 0Ah) MUST be set. This will allow PCI Bus Master IDE
capable device drivers to recognize the fact that this drive has been identified and configured by the
firmware for PCI Bus Master IDE operation.
10-12
Intel® 460GX Chipset Software Developer’s Manual
IFB Usage Considerations
10.5.7.1
10.5.7.2
BMIS1 - Bus Master IDE Status Register 1
(Primary: Bus Master IDE Base I/O Address + Offset 02h)
BMIS2 - Bus Master IDE Status Register 2
(Secondary: Bus Master IDE Base I/O Address + Offset 0Ah)
7
6
5
4
3
2
1
0
Reserved Drive 1 DMA Drive 0 DMA Reserved
Capable Capable
(DMACAP1) (DMACAP0)
IDE
Interrupt
Status
IDE DMA
Error
Bus Master
IDE Active
(BMIDEA)
(IDEINTS)
0: Drive is
PIO only.
0: Drive is
PIO only.
0: R/W/C
Software
0: R/W/C
Software
0: RO
1: Drive is
capable and capable and
1: Drive is
writes 1’ to writes 1’ to
clear. clear.
configured
for DMA
configured
for DMA
transfers.
transfers.
• The Drive 0 DMA Capable bit shall be:
— Set to ‘1’ when the Drive 0 (Master) has been identified and configured for DMA
transfers (Ultra DMA, Multi Word DMA or Single Word DMA).
— Set to ‘0’ if Drive 0 is PIO only and/or not configured for DMA operation.
• The Drive 1 DMA Capable bit shall be:
— Set to ‘1’ when the Drive 1 (Slave) has been identified and configured for DMA transfers.
— Set to ‘0’ if Drive 1 is PIO only and/or not configured for DMA operation.
It is the responsibility of initialization software to ensure that these DMA capable bits are set so
that a PCI Bus Master IDE device driver can determine which drives have been configured for
DMA operation.
10.5.8
Settings Checklist
The following checklists can be used in determining drive modes. Refer to the “Determining a
Drive’s Transfer Rate Capabilities” and “IFB Timing Settings” sections for more information.
Table 10-13. Drive Capabilities Checklist
Best DMA
Best Ultra
DMA Mode
(Ultra DMA
Mode 0, 1,
2, or N/A)
Best PIO
Mode (Fast
PIO Mode 2,
3, 4 or
Non Ultra DMA
Supported?
Best DMA Mode
is {SW2, MW1,
MW2}
Fast PIO
Supported?
Best PIO Mode
>= Best DMA
Mode
Type
(ATA Fixed
Disk or
Mode
(Single
IFB
Ultra
DMA
Mode
IFB
Mode
Drive
Position
Word 2,
Multi Word
1, 2, or N/A)
ATAPI)
Compatible)
Drive 0
Drive 1
Drive 2
Drive 3
Primary
Master
Primary
Slave
Secondary
Master
Secondary
Slave
Intel® 460GX Chipset Software Developer’s Manual
10-13
IFB Usage Considerations
Table 10-14. IFB Settings Checklist
Register
Type
Offset
Value
0005h
Comments
PCI Command Register
PCI
PCI
PCI
PCI
PCI
PCI
PCI
PCI
04h
0Dh
Ensure that bits 0 and 2 are ‘1’
PCI Master Latency Timer
PCI Bus Master IDE Base I/O Address
IDE Timing Register 1
20-23h
40-41h
42-43h
44h
Ensure that bit 0 (of register value) is ‘1’
IDE Timing Register 2
Secondary IDE Timing Register
Ultra DMA Control Register
Ultra DMA Timing Register
48h
4A-4Bh
10.5.9
Example Configurations
This section provides examples of drive configurations on a IFB-based system.
10.5.9.1
Example #1: Ultra DMA/33 Configuration
Non Ultra DMA
Best
Fast PIO
IFB Ultra
DMA
Mode
Supported?
Best DMA
Ultra
DMA
Mode
Best DMA
Mode
Best PIO
Mode
IFB
Mode
Supported?
Drive
Type
Position
Best PIO Mode >=
Best DMA Mode
Mode is {SW2,
MW1, MW2}
Drive 0
Drive 2
Fixed Disk
Primary
Single
Ultra
Multi Word PIO4
DMA Mode
2
Ultra
Mode 4
Mode 3
yes
yes
yes
DMA
DMA
Mode 2
Mode 2
ATAPI
CDROM
Secondary
Single
Ultra
Multi Word PIO3
DMA Mode
1
Ultra
yes
DMA
DMA
Mode 1
Mode 1
In the above configuration, since both drives support Ultra DMA, Ultra DMA will be enabled on
each of the drives: Ultra DMA Mode 2 for Drive 0 and Ultra DMA Mode 1 for Drive 1. Non-ultra
DMA and Fast PIO support will be enabled on each drive as well.
Register
Type
Offset
Value
Comments
PCI Command Register
PCI Master Latency Timer
PCI
PCI
PCI
04h
0Dh
0005h
Ensure that bits 0 and 2 are ‘1’.
system dependent
system dependent
PCI Bus Master IDE Base I/O
Address
20-23h
Ensure that bit 0
(of register value) is ‘1’.
IDE Timing Register 1
IDE Timing Register 2
PCI
PCI
40-41h
42-43h
44h
A307h
A303h
00h
mode config. for Primary
mode config. for Secondary
Secondary IDE Timing Register PCI
Ultra DMA Control Register
PCI
48h
05h
Drive 0 and 2 are Ultra DMA
capable.
Ultra DMA Timing Register
PCI
4A-4Bh
0102h
Ultra DMA mode config.
10-14
Intel® 460GX Chipset Software Developer’s Manual
IFB Usage Considerations
10.5.9.2
Example #2: Mixed Ultra DMA/33 and Non-ultra DMA/33 Configuration
Non Ultra DMA
Best
Ultra
DMA
Mode
Fast PIO
IFB Ultra
DMA
Mode
Supported?
Best DMA
Best DMA
Mode
Best PIO
Mode
IFB
Mode
Supported?
Drive
Type
Position
Best PIO Mode >=
Best DMA Mode
Mode is {SW2,
MW1, MW2}
Drive 0
Drive 1
Drive 2
Fixed Disk
Fixed Disk
Primary
Master
Ultra
Multi Word PIO4
DMA Mode
2
Ultra
Mode 4
yes
yes
yes
yes
yes
yes
DMA
DMA
Mode 2
Mode 2
Primary
Slave
N/A
Multi Word PIO4
DMA Mode
2
N/A (Ultra Mode 4
DMA
disabled)
ATAPI
CDROM
Secondary
Single
N/A
Multi Word PIO3
DMA Mode
1
N/A (Ultra Mode 3
DMA
disabled)
In the above configuration, Ultra DMA Mode 2 will only be enabled on Drive 0. Non-ultra DMA
and Fast PIO support will be enabled on each drive as well.
.
Register
Type
Offset
Value
Comments
PCI Command Register
PCI Master Latency Timer
PCI
PCI
PCI
04h
0Dh
0005h
Ensure that bits 0 and 2 are ‘1’.
System dependent
System dependent
PCI Bus Master IDE Base I/O
Address
20-23h
Ensure that bit 0 (of register value)
is ‘1’.
IDE Timing Register 1
PCI
PCI
PCI
PCI
PCI
40-41h
42-43h
44h
E377h
A103h
0Bh
Mode config. for Primary
IDE Timing Register 2
Mode config. for Secondary
Secondary IDE Timing Register
Ultra DMA Control Register
Ultra DMA Timing Register
48h
01h
Drive 0 is Ultra DMA capable.
Ultra DMA mode config.
4A-4Bh
0002h
10.5.9.3
Example #3: Non Ultra DMA/33 Drive Configuration
Non Ultra DMA
Fast PIO
Supported?
Best PIO Mode
>= Best DMA
Mode
Best
Ultra
DMA
Mode
Supported?
Best DMA
Best DMA
Mode
Best PIO
Mode
IFB Ultra
DMA Mode
IFB
Mode
Drive
Type
Position
Mode is {SW2,
MW1, MW2}
Drive 0
Drive 1
Drive 2
Fixed Disk
Fixed Disk
Primary
Master
N/A
Multi Word
PIO4
PIO4
PIO3
N/A (Ultra
DMA
Mode 4 yes
Mode 4 yes
Mode 3 yes
yes
yes
yes
DMA Mode 2
disabled)
Primary Slave N/A
Multi Word
DMA Mode 2
N/A (Ultra
DMA
disabled)
ATAPI
CDROM
Secondary
Single
N/A
Multi Word
DMA Mode 1
N/A (Ultra
DMA
disabled)
Intel® 460GX Chipset Software Developer’s Manual
10-15
IFB Usage Considerations
In the above configuration, none of the drives supports Ultra DMA. Only Non-ultra DMA and Fast
PIO support will be enabled on each drive.
Register
Type
Offset
Value
Comments
PCI Command Register
PCI Master Latency Timer
PCI
PCI
PCI
04h
0Dh
0005h
Ensure that bits 0 and 2 are ‘1’.
System dependent
System dependent
PCI Bus Master IDE Base I/O
Address
20-23h
Ensure that bit 0 (of register value)
is ‘1’.
IDE Timing Register 1
PCI
PCI
PCI
PCI
PCI
40-41h
42-43h
44h
E377h
A103h
0Bh
Mode config. for Primary
IDE Timing Register 2
Mode config. for Secondary
Secondary IDE Timing Register
Ultra DMA Control Register
Ultra DMA Timing Register
48h
00h
Ultra DMA is disabled for all drives.
4A-4Bh
0000h
10.5.10 Ultra DMA System Software Considerations
This section outlines some of the key system considerations for systems where Ultra DMA
operation is enabled.
The following components shall comprise a Ultra DMA/33 System:
1. Ultra DMA/33-capable Host IDE Controller.
2. Ultra DMA/33-capable ATA/ATAPI Devices.
3. Ultra DMA/33 Aware firmware and /or System Initialization Software.
4. Ultra DMA/33 Aware Device Driver.
Two of the aforementioned components are software based with specific requirements: Item #3,
Ultra DMA/33 Aware firmware and/or System Initialization Software; and Item #4, Ultra DMA/33
Aware Device Driver .
Ultra DMA/33 Aware firmware and/or System Initialization Software shall:
1. Identify Ultra DMA/33 capable devices and host controllers.
2. Configure Ultra DMA/33 operation for all Ultra DMA/33 capable devices and host
controllers.
3. Preserve Ultra DMA/33 configuration across reset states, restoring Ultra DMA/33 operation
(described in Item #2) as necessary.
Ultra DMA/33 Aware Device Drivers shall:
I. Provide support for PCI Bus Master IDE Operation (SFF8038i).
A. Identify system configured for PCI Bus Master IDE operation.
1. Identify PCI Bus Master IDE and Ultra DMA/33 capable devices and host
controllers.
2. Utilize PCI Bus Master IDE when it determines that Host Controller and ATA/
ATAPI devices have fulfilled device driver-specific, configuration requirements for
PCI Bus Master IDE operation.
B. Perform data transfer commands with PCI Bus Master IDE on devices, host controllers
that support PCI Bus Master IDE.
10-16
Intel® 460GX Chipset Software Developer’s Manual
IFB Usage Considerations
II. Provide recovery for data transfers that fail as the result of Ultra DMA/33 Interface CRC
Errors:
A. Determine that the data transfer command’s error source is Ultra DMA/33 Interface CRC
error.
B. Retry data transfer command when Ultra DMA/33 Interface CRC is the source of error.
III. Ensure that the Ultra DMA/33 configuration of the devices and host controller is restored
when events that clear the Ultra DMA/33 enabled status are encountered.
•
•
Ensure that Hard Resets are never issued to the device during a power cycle -OR-
Provide path to Ultra DMA/33 Aware firmware and/or System Initialization Software in
the case of a Hard/Power-On reset.
When enabled on the host controller and devices, Ultra DMA/33 operation shall be used for all data
transfer commands issued by the Ultra DMA/33 Aware Device Driver with PCI Bus Master IDE,
DMA operation. PIO or Multi Word DMA shall be the mode of access used with devices and host
controllers that do not support Ultra DMA/33.
10.5.11 Additional Ultra DMA/PCI Bus Master IDE Device Driver
Considerations
This section provides information regarding Terminating Transfers performed with Ultra DMA/
PCI Bus Master IDE device drivers or system software.
In normal bus master operations, at the end of a data transfer, the IDE device signals an interrupt.
In response to the interrupt, software verifies that the bus is idle and then writes the Stop Bus
Master Command. It then reads the controller status register to determine if the transfer completed
successfully. For a detailed description of the Bus Master IDE Status Register refer to the last
section of this document called Bus Master IDE Command and Status Registers.
If the IDE device does not signal the interrupt, the last bus master transfer did not complete. In this
case, it is necessary to read the Bus Master IDE Status Register to check if the bus is idle or active.
If the bus is active it is necessary to send the Stop Bus Master Command and reset the IDE
controller and drives connected to the IDE cable prior to sending out the next ATA/ATAPI drive
command to the cable. This is necessary because the cable (Primary or Secondary IDE) may not be
ready to receive new commands unless the Bus Master state machine has stopped. All the drives on
the cable should be reset.
In general, a prematurely terminated command on the IDE bus implies that some of the state
machines in the drive and/or in the IFB are still in an “active” condition. By performing a drive
reset immediately following the burst “stop”, the IFB will be in a state such that the IDE DMA
engines can be programmed to perform the transfer once again.
10.5.11.1 Bus Master IDE Command and Status Register
10.5.11.2 BMICX–Bus Master IDE Command Register (I/O)
Address Offset:
Default Value:
Attribute:
Primary Channel–Base + 00h; Secondary Channel–Base + 08h
00h
Read/Write
Intel® 460GX Chipset Software Developer’s Manual
10-17
IFB Usage Considerations
This register enables/disables bus master capability for the IDE function and provides direction
control for the IDE DMA transfers. This register also provides bits that software uses to indicate
DMA capability of the IDE device.
Bit
Description
7:4
3
Reserved.
Bus Master Read/Write Control (RWCON). 0=Reads; 1=Writes. This bit must NOT be changed
when the bus master function is active. While a Ultra DMA transfer is in progress, this bit will be
READ ONLY. The bit will return to read/write once the synchronous DMA transfer has been
completed or halted.
2:1
0
Reserved.
Start/Stop Bus Master (SSBM). 1=Start; 0=Stop. When this bit is set to 1, bus master operation
starts. The controller transfers data between the IDE device and memory only while this bit is set.
Master operation can be stopped by writing a 0 to this bit. This results in all state information
being lost (i.e. master mode operation cannot be stopped and then resumed).
If this bit is set to 0 while bus master operation is still active (i.e. Bit 0=1 in the Bus Master IDE
Status Register for that IDE channel) and the drive has not yet finished its data transfer (bit 2=0 in
the channel’s Bus Master IDE Status Register), the bus master command is aborted and data
transferred from the drive may be discarded by IFB rather than being written to system memory.
This bit is intended to be set to 0 after the data transfer is completed, as indicated by either bit 0 or
bit 2 being set in the IDE Channel’s Bus Master IDE Status Register.
10.5.11.3 BMISX–Bus Master IDE Status Register (I/O)
Address Offset:
Default Value:
Attribute:
Primary Channel–Base + 02h; Secondary Channel–Base + 0Ah
00h
Read/Write Clear
This register provides status information about the IDE device and state of the IDE DMA transfer.
Bit
Description
7
6
Reserved. This bit is hardwired to 0.
Drive 1 DMA Capable (DMA1CAP)–R/W. 1=Drive 1 is capable of DMA transfers. This bit is a
software controlled status bit that indicates IDE DMA device capability and does not affect hardware
operation.
5
Drive 0 DMA Capable (DMA0CAP)–R/W. 1=Drive 0 is capable of DMA transfers. This bit is a
software controlled status bit that indicates IDE DMA device capability and does not affect hardware
operation.
4:3
2
Reserved.
IDE Interrupt Status (IDEINTS)–R/WC. This bit, when set to a 1, indicates when an IDE device has
asserted its interrupt line. When bit 2=1, all read data from the IDE device has been transferred to
main memory and all write data has been transferred to the IDE device. Software sets this bit to a 0
by writing a 1 to it. IRQ14 is used for the primary channel and IRQ15 is used for the secondary
channel. Note that, if the interrupt status bit is set to a 0 by writing a 1 to this bit while the interrupt
line is still at the active level, this bit remains 0 until another assertion edge is detected on the
interrupt line.
1
0
IDE DMA Error–R/WC. This bit is set to 1 when IFB encounters a target abort or master abort while
transferring data on the PCI Bus. Software sets this bit to a 0 by writing a 1 to it.
Bus Master IDE Active (BMIDEA)–RO. IFB sets this bit to 1 when bit 0 in the BMICx Register is set
to 1. IFB sets this bit to 0 when the last transfer for a region is performed (where EOT for that region
is set in the region descriptor). IFB also sets this bit to 0 when bit 0 of the BMICx Register is set to 0.
When this bit is read as a zero, all data transferred from the drive during the previous bus master
command is visible in system memory, unless the bus master command was aborted.
10-18
Intel® 460GX Chipset Software Developer’s Manual
IFB Usage Considerations
Interrupt/Activity Status Combinations
Bit 2
Bit 0
Description
0
1
0
DMA transfer is in progress. No interrupt has been generated by the IDE device.
1
1
0
The IDE device generated an interrupt and the Physical Region Descriptors
exhausted. This is normal completion where the size of the physical memory regions is
equal to the IDE device transfer size.
1
0
The IDE device generated an interrupt. The controller has not reached the end of the
physical memory regions. This is a valid completion case when the size of the physical
memory regions is larger than the IDE device transfer size.
Error condition. If the IDE DMA Error bit is 1, there is a problem transferring data to/
from memory. Specifics of the error have to be determined using bus-specific
information. If the Error bit is 0, the PRD specified a smaller buffer size than the
programmed IDE transfer size.
10.6
USB Resume Enable Bit
Two bits have been added to the USB Host controller functionality in function 2 of IFB (PCI
Register configuration space at Offset C4h). This register is in the resume well of this function.
USB Resume Enable: (IFB Function 2 Config Register)
Address Offset:
Attribute:
Default Value:
Size:
C4h
Read/Write
00h
8 bits
Bit
7:2
Type
RO
Description
Reserved
1
RW
PORT1EN: Enable port 1 of the USB controller to look at wakeup events. When set, the
USB controller will monitor port 1 for remote wakeup and connect/disconnect events.
When cleared, the USB controller will not look at this port for a wakeup event. For function
2, this bit applies to port 1.
0
RW
PORT0EN: Enable port 0 of the USB controller to look at wakeup events. When set, the
USB controller will monitor port 0 for remote wakeup and connect/disconnect events.
When cleared, the USB controller will not look at this port for a wakeup event. For function
2, this bit applies to port 0.
For performing legacy power management, the firmware has to set these two bits in each of the
functions, if it wants the USB Host controller to monitor these ports.
Intel® 460GX Chipset Software Developer’s Manual
10-19
IFB Usage Considerations
10-20
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration 11
The IFB PCI Function 0 contains a LPC/FWH interface, interrupt controller and counter / timers,
including the real time clock. The register set associated with this Functionality and associated
logic is shown below with actual register descriptions given in this section.
11.1
PCI to LPC/FWH Interface Configuration Space
Registers (PCI Function 0)
11.1.1
VID–Vendor Identification Register (Function 0)
Address Offset:
Default Value:
Attribute:
00–01h
8086h
Read Only
The VID Register contains the vendor identification number. This register, along with the Device
Identification Register, uniquely identifies any PCI device. Writes to this register have no effect.
Bit
Description
15:0
Vendor Identification Number. This is a 16-bit value assigned to Intel.
11.1.2
DID–Device Identification Register (Function 0)
Address Offset:
Default Value:
Attribute:
02–03h
7600h
Read Only
The DID Register contains the device identification number. This register, along with the VID
Register, defines the IFB. Writes to this register have no effect.
Bit
Description
15:0
Device Identification Number. This is a 16-bit value assigned to the IFB.
Intel® 460GX Chipset Software Developer’s Manual
11-1
LPC/FWH Interface Configuration
11.1.3
PCICMD–PCI Command Register (Function 0)
Address Offset:
Default Value:
Attribute:
04–05h
0007h
Read/Write
This 16-bit register provides basic control over the IFB's ability to respond to PCI cycles.
Bit
Description
15:10
Reserved.
9
8
Fast Back-to-Back Enable (Not Implemented). This bit is hardwired to 0.
SERR# Enable (SERRE). 1=Enable. 0=Disable. When enabled (and DLC Register, bit 3=1),
a delayed transaction time-out causes the IFB to assert the SERR# signal. The PCISTS
register reports the status of the SERR# signal.
7:5
4
Reserved.
Postable Memory Write Enable (Not Implemented). This bit is hardwired to 0.
3
Special Cycle Enable (SCE). 1=Enable, the IFB recognizes Shutdown special cycle.
0=Disable, the IFB ignores all PCI Special Cycles.
2
1
0
Bus Master Enable (Not Implemented). The IFB does not support disabling its Function 0
bus master capability. This bit is hardwired to 1.
Memory Access (Not Implemented). The IFB does not support disabling Function 0 access
to memory. This bit is hardwired to 1.
I/O Space Access Enable (Not Implemented). The IFB does not support disabling its
Function 0 response to PCI I/O cycles. This bit is hardwired to 1.
11.1.4
PCISTS–PCI Device Status Register (Function 0)
Address Offset:
Default Value:
Attribute:
06–07h
0280h
Read/Write
The PCISTS Register reports the occurrence of a PCI master-abort by the IFB or a PCI target-abort
when the IFB is a master. The register also indicates the IFB DEVSEL# signal timing.
Bit
Description
15
14
Detected Parity Error (Not Implemented). Read as 0.
Signaled SERR# Status (SERRS)–R/WC. When the IFB asserts the SERR# signal, this bit is
set to 1. Software clears this bit by writing a 1 to it.
13
12
Master-Abort Status (MAS)–R/WC. When the IFB, as a master (for Function 0), generates a
master-abort, MAS is set to a 1. Software sets MAS to 0 by writing a 1 to this bit location.
Received Target-Abort Status (RTA)–R/WC. When the IFB is a master on the PCI Bus (for
Function 0) and receives a target-abort, this bit is set to a 1. Software sets RTA to 0 by writing
a 1 to this bit location.
11
Signaled Target-Abort Status (STA)–R/WC. This bit is set when the IFB LPC bridge Function
is targeted with a transaction that the IFB terminates with a target abort. Software sets STA to
0 by writing a 1 to this bit location.
10:9
DEVSEL# Timing Status (DEVT)–RO. The IFB always generates DEVSEL# with medium
timing for Function 0 I/O cycles. Thus, DEVT=01. This DEVSEL# timing does not include
Configuration cycles.
11-2
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
Bit
Description
8
7
PERR# Response (Not Implemented). Read as 0.
Fast Back to Back–RO. This bit indicates to the PCI Master that IFB as a target is capable of
accepting fast back-to-back transactions. This bit is hardwired to 1.
6:0
Reserved.
11.1.5
RID–Revision Identification Register (Function 0)
Address Offset:
Default Value:
Attribute:
08h
Stepping Dependent
Read Only
This 8 bit register contains device stepping information. Writes to this register have no effect.
Bit
Description
7:0
Revision ID Byte.
11.1.6
CLASSC–Class Code Register (Function 0)
Address Offset:
Default Value:
Attribute:
09h-0Bh
060100h
Read Only
This register identifies the Base Class Code, Sub-class Code, and Device Programming interface
for the IFB PCI Function 0.
Bit
Description
Base Class Code (BASEC). 06h=Bridge device.
23:16
15:8
Sub-Class Code (SCC). 01h=PCI-to-ISA Bridge. ISA is not supported, IFB forwards cycles to
the LPC interface.
7:0
Programming Interface (PI). 00h=No register level programming interface defined.
11.1.7
HEDT–Header Type Register (Function 0)
Address Offset:
Default Value:
Attribute:
0Eh
80h
Read Only
The HEDT Register identifies the IFB as a multi-Function device.
Bit
Description
7:0
Device Type (DEVICET). 80h=multi-Function device.
Intel® 460GX Chipset Software Developer’s Manual
11-3
LPC/FWH Interface Configuration
11.1.8
ACPI Base Address (Function 0)
Address:
40-43h
Default Value:
Attribute:
00000001h
Read/Write
Bit
Description
31:16
15:6
5:1
Reserved.
Base Address: Provides the 64 bytes of I/O space.
Reserved.
0
Resource Indicator: Tied to 1 to indicate I/O space
11.1.9
ACPI Enable (Function 0)
Address:
44h
Default Value
Attribute:
00
Read/Write
Bit
Description
7:1
0
Reserved.
ACPI Enable: When this bit is set to ‘1’, decode of the I/O range pointed to by the ACPI base
register is enabled, and the ACPI power management Function is enabled. Note that the APM
power management ranges (B2/B3h) are always enabled and are not affected by this bit.
11.1.10 SCI IRQ Routing Control
Address:
45h
Default Value:
Attributes:
00h
Read/Write
Bit
Description
7:3
2:0
Reserved.
SCI IRQ Map: Specifies on which pin the SCI will appear on internally. If not using the APIC,
software must program this register to “000”. This interrupt is not sharable. When this interrupt is
chosen, the corresponding interrupt pin is blocked and the SERIRQ frame entry is blocked. This
interrupt can be shared with a PCI interrupt.
Bits
000
001
010
011
100
101
110
111
SCI Map
IRQ9
SCI
FEMPTY#
IRQ8
IRQ0
IRQ10
IRQ11
Reserved
11-4
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
11.1.11 BIOSEN–BIOS Enable Register (FUNCTION 0)
Address Offset:
Default Value:
Attribute:
4E-4Fh
07C1h
Read/Write
This register is used to implement protections to writes to firmware (BIOS) ranges.
Bit
Description
15
BLE - BIOS Lock Enable: When the bit is a “1”, setting BIOS_WEN bit will cause SMIs. When this
bit is a “0”, setting BIOS_WEN will not cause SMIs. Once set, this bit can only be cleared by a
PCIRST#.
14:11
10:3
2
Reserved.
Reserved. Software must preserve these register values.
BIOS_WEN - BIOS Write Enable. When this bit is set to a “1”, writes to firmware (BIOS) ranges
are allowed. When this bit is a “0”, writes to firmware (BIOS) ranges are not allowed and write
cycles will be master-aborted on PCI. There is no protection on writes to FWH register space.
When this bit is written from a “0” to a “1”, and bit 15 (BIOS Lock Enable) is set, an SMI# is
generated.
1:0
Reserved. Software must preserve these register values.
11.1.12 PIRQRC[A:D]–PIRQx Route Control Registers (Function 0)
Address Offset: 60h (PIRQRCA#)–63h (PIRQRCD#)
Default Value: 80h
Attribute:
R/W
These registers control the routing of the PIRQ[A:D]# signals to the IRQ inputs of the interrupt
controller. Each PIRQx# can be independently routed to any one of 11 interrupts. All four PIRQx#
lines can be routed to the same IRQx input. Note that the IRQ that is selected through bits [3:0]
must be set to level sensitive mode in the corresponding ELCR Register. When a PIRQ signal is
routed to an interrupt controller IRQ, the IFB masks the corresponding IRQ signal.
Bit
Description
7
Interrupt Routing Enable. When this bit is a “0”, the corresponding PIRQ is routed to one of the
compatibility mode interrupts specified in bits[3:0]. When this bit is a “1”, the PIRQ is not routed.
6:4
3:0
Reserved.
Interrupt Routing. When bit 7=0, this field selects the routing of the PIRQx to one of the interrupt
controller interrupt inputs.
Bits[3:0]
0000
0001
0010
0011
IRQ Routing
Reserved
Reserved
Reserved
IRQ3
Bits[3:0]
0110
0111
1000
1001
IRQ Routing
IRQ6
IRQ7
Reserved
IRQ9
Bits[3:0]
1011
1100
1101
1110
IRQ Routing
IRQ11
IRQ12
Reserved
IRQ14
0100
IRW4
1010
IRQ10
1111
IRQ15
0101
IRQ5
Intel® 460GX Chipset Software Developer’s Manual
11-5
LPC/FWH Interface Configuration
11.1.13 SerIRQC–Serial IRQ Control Register (Function 0)
Address Offset:
Default Value:
Attribute:
64h
10h
R/W
This register controls the Start Frame Pulse Width generated on the Serial Interrupt signal
(SERIRQ).
Bit
Description
7
6
Serial IRQ Enable. 1=Serial Interrupts are enabled. 0=Serial Interrupts disabled.
Serial IRQ Mode Select. When this bit is a “1”, the serial IRQ machine will be in continuous mode.
When this bit is a “0”, the serial IRQ machine will be in quiet mode. When setting the SIRQEN bit,
this bit must also be written as a one so that the first action of the serial IRQ machine will be a start
frame.
5:2
1:0
Serial IRQ Frame Size. These bits select the frame size used by the Serial IRQ logic. The default
is 0100b indicating a frame size of 21 (17+4). These bits are readable and writeable, however the
only programmed value supported by the IFB is 0100b. All other frame sizes are unsupported.
Start Frame Pulse Width. These bits define the Start Frame pulse width generated by the Serial
Interrupt control logic.
Bits[1:0] Pulse Width (PCI Clocks)
00 4 Clocks
01 6 Clocks
10 8 Clocks
11 Reserved
11.1.14 TOM–Top of Memory Register (Function 0)
Address Offset:
Default Value:
Attribute:
69h
02h
Read/Write
This register enables the forwarding of DMA memory cycles to the PCI Bus and sets the top of
main memory accessible by DMA devices. In addition, this register controls the forwarding of
DMA accesses to the lower firmware region (E0000–EFFFFh) and the 512–640 KByte main
memory region (80000–9FFFFh).
Bit
Description
7:4
Top Of Memory. The top of memory can be assigned in 1 Mbyte increments from 1–16 Mbytes.
DMA accesses within this region, and not in the memory hole region, are forwarded to PCI.
Bits[7:4]
0000
Top of Memory
1 Mbyte
Bits[7:4]
0110
Top of Memory
7 Mbyte
Bits[7:4]
1011
Top of Memory
12 Mbyte
0001
2 Mbyte
0111
8 Mbyte
1100
13 Mbyte
0010
3 Mbyte
1000
9 Mbyte
1101
14 Mbyte
0011
0100
4 Mbyte
5 Mbyte
1001
1010
10 Mbyte
11 Mbyte
1110
1111
15 Mbyte
16 Mbyte
0101
6 Mbyte
Note that the IFB only supports a main memory hole at the top of 16 Mbytes. Thus, If a 1 Mbyte
memory hole is created for the Host-to-PCI Bridge DRAM controller between 15 and 16 Mbytes, the
IFB Top Of Memory should be set at 15 Mbytes.
3
2
1
0
Reserved. Must be set to ‘0’.
Reserved. Must be set to ‘0’.
Reserved. Must be set to ‘0’.
Reserved.
11-6
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
11.1.15 MSTAT–Miscellaneous Status Register (Function 0)
Address Offset:
Default Value:
Attribute:
6A–6Bh
0000h
Read/Write
This register provides miscellaneous status and control Functions.
Bit
Description
15
SERR# Generation Due To Delayed Transaction Time-out–R/WC. This status bit is set
whenever the IFB times out a cycle it was running as a delayed transaction, bit 3 (Delayed
Transaction SERR# enable) of offset 82h is set, and bit 8 (SERR# enable) of offset 4h is set.
When a master does not return for the data within 1 ms of the cycle’s completion, the IFB
asserts SERR#, clears the delayed transaction, and sets this bit. If either bit 3 of 82h or bit 8
of 4h is not set then this bit will not get set and SERR# will not be generated. The IFB will still
discard the delayed transaction cycle. The bit can be cleared by writing “1” to it via software.
14:1
0
Reserved.
ECC SERR# Enable: When this bit is a “1”, it enables a one clock pulse on ECCINT# to
instead be routed to generate SERR#. This allows ECC scrubbing through SMI (greater than
one clock) or NMI (one clock).
11.1.16 Deterministic Latency Control Register (Function 0)
Address Offset:
Default Value:
Attribute:
82h
00h
Read/Write
This register enables and disables the Delayed Transaction and Passive Release Functions. When
enabled, these Functions make the IFB PCI revision 2.1 compliant.
The PCI specification requires much tighter controls on target and master latency. Targets must
respond with TRDY# or STOP# within 16 clocks of FRAME#, and masters must assert IRDY#
within 8 PCI clocks for any data phase. PCI cycles to or from LPC typically take longer than this.
The IFB provides a programmable delayed completion mechanism described in the PCI
specification to meet the required target latencies.
Bit
Description
7:4
3
Reserved.
SERR# Generation Due To Delayed Transaction Time-out Enable. When this bit is set,
and bit 8 (SERR# enable) of the Command Register (offset 4h) is set, the IFB will generate
SERR# when a delayed transaction cycle times out.
2
USB Passive Release Enable (USBPR). When this bit is set, the IFB enables the passive
release mechanism for USB. Normally, if the IFB gets retried under USB, it will not remove
PHOLD#. This bit is needed due to posting of USWC cycles, which means a retry might be
due to a USWC flush.
1
0
Passive Release Enable. When this bit is a “1”, it enables the Passive Release mechanism
encoded on the PHOLD# signal when IFB is a PCI Master. When this bit is a “0”, Passive
Release is disabled.
Delayed Transaction Enable. When this bit is a “1”, it enables the Delayed Transaction
mechanism when the IFB is the target of a PCI transaction. When this bit is a “0”, the Delayed
Transaction mechanism is disabled.
Intel® 460GX Chipset Software Developer’s Manual
11-7
LPC/FWH Interface Configuration
11.1.17 MGPIOC–Muxed GPIO Control (Function 0)
Offset:
84-85h
Default Value:
Attribute:
0500h
Read/Write
Bit
Description
16:13
12
11
10
9
Reserved.
Reserved. Must be set to ‘1’.
Reserved.
Reserved. Must be set to ‘1’.
Reserved.
8
Reserved. Must be set to ‘1’.
Reserved. Must be set to ‘1’.
Reserved. Must be set to ‘1’.
Reserved. Must be set to ‘1’.
Reserved. Must be set to ‘1’.
Reserved.
7
6
5
4
3:0
11.1.18 PDMACFG–PCI DMA Configuration Resister (Function O)
Address Offset:
Default Value:
Attribute:
90-91h
0000h
Read/Write
Bits
Description
15:14
13:12
11:10
9:8
Reserved. Must be set to ‘11’.
Reserved. Must be set to ‘11’.
Reserved. Must be set to ‘11’.
Reserved.
7:6
Reserved. Must be set to ‘11’.
Reserved. Must be set to ‘11’.
Reserved. Must be set to ‘11’.
Reserved. Must be set to ‘11’.
5:4
3:2
1:0
11.1.19 DDMABP–Distributed DMA Slave Base Pointer Registers
(Function 0)
Address Offset:
Default Value:
Attribute:
92-93h (CH0-3), 94-95h (CH5-7)
0000h
Read/Write
11-8
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
These registers provide the base address for distributed DMA slave channel registers, one for each
DMA controller. Bits 5:0 are reserved to provide access to a 64 byte I/O space (16 bytes per
channel). The channels are accessed using offset from base address as follows (Note that Channel 4
is reserved and is not accessible).
Base Offset
Channel
00 - 0Fh
10 - 1Fh
20 - 2Fh
30 - 3Fh
0,4
1,5
2,6
3,7
Bits
Description
15:6
Base Pointer: I/O Address pointer to DMA Slave Channel registers. Corresponds to PCI
address AD[15:6].
5:0
Reserved.
11.1.20 RTCCFG–Real Time Clock Configuration Register
(Function 0)
Address Offset:
Default Value:
Attribute:
C8h
00h
Read/Write
This register is used to configure the internal Real Time Clock.
Bits
Description
7:5
4
Reserved.
Lock Upper RAM Bytes: 0 = Upper RAM data bytes 38h-3Fh in the extended bank are
readable and write-able (default). 1 = Upper RAM data bytes 38h-3Fh in the extended bank
are neither readable nor write-able. This is used to lock bytes 38h-3Fh in the upper 128-byte
bank of RAM. Write cycles will have no effect and read cycles will not return an expected
value.
Warning: This is a write-once register that can only be reset by a hardware reset. No
software means is possible to reset this bit.
3
Lock Lower RAM Bytes: 0 = Lower RAM data bytes 38h-3Fh in the standard bank are
readable and writeable (default). 1 = Lower RAM data bytes 38h-3Fh in the standard bank are
neither readable nor writeable. This is used to lock bytes 38h-3Fh in the lower 128-byte bank
of RAM. Write cycles will have no effect and read cycles will not return an expected value.
Warning: This is a write-once register that can only be reset by a hardware reset. No
software means is possible to reset this bit.
2
Upper RAM Enable: 0 = Accesses to RTC Upper 128 byte extended bank at I/O address 72-
73h is disabled. 1 = Accesses to 72-73h are forwarded to RTC Upper 128 byte extended
bank.
1:0
Reserved.
Intel® 460GX Chipset Software Developer’s Manual
11-9
LPC/FWH Interface Configuration
11.1.21 GPIO Base Address (FUNCTION 0)
Address:
D0-D3h
Default Value:
Attributes:
00000001h
Read/Write
Bit
Description
31:16
15:6
5:1
Reserved.
Base Address: Provides the 64 bytes of I/O space.
Reserved.
0
Resource Indicator: Tied to 1 to indicate I/O space
11.1.22 GPIO Enable (FUNCTION 0)
Address:
D4h
Default Value
Attributes:
00h
Read/Write
Bit
Description
7:1
0
Reserved.
GPIO Enable: When this bit is set to ‘1’, decode of the I/O range pointed to by the GPIO base
register is enabled, and the GPIO Function is enabled.
11.1.23 LPC COM Decode Ranges (Function 0)
Address:
E0h
Default Value:
Attributes:
00h
Read/Write
Bit
Description
7
Reserved.
Decode Range: The following table describes which range to decode for the COMB Port
6:4
Bits
Decode Range
3F8 - 3FF (COM 1)
000
001
010
011
100
101
110
111
2F8 - 2FF (COM 2)
220 - 227
228 - 22F
238 - 23F
2E8 - 2EF (COM 4)
338 - 33F
3E8 - 3EF (COM 3)
11-10
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
Bit
Description
3
Reserved.
Decode Range: The following table describes which range to decode for the COMA Port
2:0
Bits
Decode Range
000
001
010
011
100
101
110
111
3F8 - 3FF (COM 1)
2F8 - 2FF (COM 2)
220 - 227
228 - 22F
238 - 23F
2E8 - 2EF (COM 4)
338 - 33F
3E8 - 3EF (COM 3)
11.1.24 LPC FDD/LPT Decode Ranges (Function 0)
Address:
E1h
Default Value:
Attributes:
00h
Read/Write
Bit
Description
7:5
4
Reserved.
Decode Range: The following table describes which range to decode for the FDD Port
Bits
0
Decode Range
3F0 - 3F5, 3F7 (Primary)
370 - 375, 377 (Secondary)
1
3:2
1:0
Reserved.
Decode Range: The following table describes which range to decode for the LPT Port
Bits
00
Decode Range
378 - 37F, 778-77F
01
278 - 27F (port 279 is read only), 678-
67F
10
11
3BC - 3BE, 7BC-7BE
Reserved.
Intel® 460GX Chipset Software Developer’s Manual
11-11
LPC/FWH Interface Configuration
11.1.25 LPC Sound Decode Ranges (Function 0)
Address:
E2h
Default Value:
Attributes:
00h
Read/Write
Bit
Description
Reserved.
7:6
5:4
Microsoft Sound System Decode Range: The following table describes which range to
decode for the MSS port:
Bits
Decode Range
00
530 - 537
604 - 60B
E80 - E87
F40 - F47
01
10
11
MIDI Decode Range: The following table describes which range to decode for the Midi Port
3
Bits
Decode Range
0
330 - 331
300 - 301
1
2
Reserved.
1:0
SB16 Decode Range: The following table describes which range to decode for the Sound
Blaster 16 Port
Bits
Decode Range
00
220 - 233
240 - 253
260 - 273
280 - 293
01
10
11
11.1.26 LPC Generic Decode Range (Function 0)
Address:
E4-E5h
Default Value:
Attributes:
0000h
Read/Write
Bit
Description
15:9
Base Address: Base Address for the generic decode range. This address is aligned on a 512
byte boundary, and being I/O, must have address lines 31:16 as “0”.
8:1
0
Reserved.
Enable: When set, the range defined by the Base Address register is enabled for decode.
11-12
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
11.1.27 LPC Enables (Function 0)
Address:
Default Value:
Attributes:
E6-E7h
0000h
Read/Write
Bit
Description
15
14:13
12
Reserved. This bit must be a “0”.
Reserved.
Secondary Configuration Enable: Enables I/O locations 4Eh and 4Fh to be sent to the LPC
bus. Super I/Os use these addresses as an alternate index/data register pair for Super I/O
configuration.
11
Configuration Enable: Enables I/O locations 2Eh and 2Fh to be sent to the LPC bus. Super
I/Os use these addresses as an index/data register pair for Super I/O configuration.
10
9
ACPI µController Enable: This enables decoding of the ports 62h and 66h to the LPC Bus.
MSS Enable: This enables decoding of the Microsoft Sound System range to the LPC Bus.
8
Keyboard Enable: This enables decoding of the keyboard ports at 60h and 64h to the LPC
Bus.
7
Game Port Enable: This enables decoding of the Game Port range at 200h - 20Fh to the
LPC Bus.
6
5
4
3
2
1
0
ADLIB Enable: This enables decoding of the ADLIB range at 388h - 38Bh to the LPC Bus.
MIDI Enable: This enables decoding of the MIDI range to the LPC Bus.
SB16 Enable: This enables decoding of the SB16 range to the LPC Bus.
FDD Enable: This enables decoding of the FDD range to the LPC Bus.
LPT Enable: This enables decoding of the LPT range to the LPC Bus.
COM B Enable: This enables decoding of the COMB range to the LPC Bus.
COM A Enable: This enables decoding of the COMA range to the LPC Bus.
11.1.27.1 Firmware Hub (FWH) Decode Enable Register
Address:
Default Value:
Attributes:
E3H
00H1
Read/Write
Bit
Description
7
FWH_F8_EN: This enables decoding of 512 KB of the FWH memory range starting at 4 GB – 512
KB (FFF80000H) to the top 4 GB (FFFFFFFFH). In addition, the upper 128 KB of this range is
shadowed at the top of 1MB (000E0000H – 000FFFFFH). Additionally, this enables decoding of
512K of register space starting at (4 GB – 4 MB) - 512KB (FFB80000h) to the top 4 GB – 4 MB
(FFBFFFFFh).
6
FWH_F0_EN: This enables decoding 512 KB of the FWH memory range starting at 4 GB – 1 MB
(FFF00000H) to 4 GB – 512 KB (FFF7FFFFH). Additionally, this enables decoding of 512K of
register space starting at (4 GB – 4 MB) - 1MB (FFB00000h) to (4 GB – 4 MB) - 512KB
(FFB7FFFFh).
1. LFRAME is defined to be tri-stated at reset. This pin is sampled on PWROK. If the pin is sampled as a logic ‘1’, the firmware does not exist, and
all the registers defined by E3H are cleared to ‘0’. If the pin is sampled as a logic ‘0’, then the firmware does exist, and all the bits in E3H are set
to ‘1’.
Intel® 460GX Chipset Software Developer’s Manual
11-13
LPC/FWH Interface Configuration
Bit
Description
5
4
3
2
1
0
FWH_E8_EN: This enables decoding 512KB of the FWH memory range starting at 4 GB – 1.5 MB
(FFE80000H) to 4 GB – 1 MB (FFEFFFFFH). Additionally, this enables decoding of 512K of
register space starting at (4 GB – 4 MB) - 1.5MB (FFA80000h) to (4 GB – 4 MB) - 1MB
(FFAFFFFFh).
FWH_E0_EN: This enables decoding 512 KB of the FWH memory range starting at 4 GB – 2 MB
(FFE00000H) to 4 GB - 1.5 MB (FFE7FFFFH). Additionally, this enables decoding of 512K of
register space starting at (4 GB – 4 MB) – 2MB (FFA00000h) to (4 GB – 4 MB) – 1.5MB
(FFA7FFFFh).
FWH_D8_EN: This enables decoding 512 KB of the FWH memory range starting at 4 GB - 2.5 MB
(FFD80000H) to 4 GB - 2 MB (FFDFFFFFH). Additionally, this enables decoding of 512K of
register space starting at (4 GB – 4 MB) - 2.5MB (FF980000h) to (4 GB – 4 MB) - 2 MB
(FF9FFFFFh).
FWH_D0_EN: This enables decoding 512 KB of the FWH memory range starting at 4 GB - 3 MB
(FFD00000H) to 4 GB - 2.5 MB (FFD7FFFFH). Additionally, this enables decoding of 512K of
register space starting at (4 GB – 4 MB) - 3.0MB (FF900000h) to 4 GB – 4 MB) - 2.5 MB
8
(FF97FFFFh) .
FWH_C8_EN: This enables decoding 512 KB of the FWH memory range starting at 4 GB - 3.5 MB
(FFC80000H) to 4 GB - 3 MB (FFCFFFFFH). Additionally, this enables decoding of 512K of
register space starting at (4 GB – 4 MB) - 3.5MB (FF880000h) to (4 GB – 4 MB) - 3MB
8
(FF8FFFFFh) .
FWH_C0_EN: This enables decoding 512 KB of the FWH memory range starting at 4 GB – 4 MB
(FFC00000H) to 4 GB - 3.5 MB (FFC7FFFFH). Additionally, this enables decoding of 512K of
register space starting at (4 GB – 4 MB) – 4MB (FF800000h) to (4 GB – 4 MB) - 3.5MB
(FF87FFFFh).
11.1.27.2 Firmware Hub (FWH) Select Register
Address:
E8H
Default Value:
Attributes:
00112233H
Read/Write
Bit
Description
31:28
FWH_F8_IDSEL: This dictates the IDSEL of 512 KB of the FWH memory range starting at 4 GB -
512 KB (FFF80000H) to the top 4 GB (FFFFFFFFH) as well as register space starting at (4 GB-
4MB) - 512KB (FFB80000h) to the top 4 GB - 4MB (FFBFFFFFh). In addition, the upper 128 KB of
this range is shadowed at the top of 1 MB (000E0000H - 000FFFFFH). The enable for this range is
controlled through bit 7 of the FWH Decode Enable Register at E3H.
FWH_F0_IDSEL: This dictates the IDSEL of 512 KB of the FWH memory range starting at 4 GB -
1 MB (FFF00000H) to 4 GB - 512 KB (FFF7FFFFH) as well as register space starting at (4 GB-
4MB) - 1MB (FFB00000h) to (4 GB-4MB) - 512KB (FFB7FFFFh). The enable for this range is
controlled through bit 6 of the FWH Decode Enable Register at E3H.
27:24
23:20
19:16
15:12
11:8
FWH_E8_IDSEL: This dictates the IDSEL of 512 KB of the FWH memory range starting at 4 GB -
1.5 MB (FFE80000H) to 4 GB - 1 MB (FFEFFFFFH) as well as register space starting at (4 GB-
4MB) - 1.5MB (FFA80000h) to (4 GB-4MB) - 1MB (FFAFFFFFh). The enable for this range is
controlled through bit 5 of the FWH Decode Enable Register at E3H.
FWH_E0_IDSEL: This dictates the IDSEL of 512 KB of the FWH memory range starting at 4 GB -
2 MB (FFE00000H) to 4 GB - 1.5 MB (FFE7FFFFH) as well as register space starting at (4 GB-
4MB) - 2MB (FFA00000h) to (4 GB-4MB) - 1.5MB (FFA7FFFFh). The enable for this range is
controlled through bit 4 of the FWH Decode Enable Register at E3H.
FWH_D8_IDSEL: This dictates the IDSEL of 512 KB of the FWH memory range starting at 4 GB -
2.5 MB (FFD80000H) to 4 GB - 2 MB (FFDFFFFFH) as well as register space starting at (4 GB-
4MB) - 2.5MB (FF980000h) to (4 GB-4MB) - 2MB (FF9FFFFFh). The enable for this range is
controlled through bit 3 of the FWH Decode Enable Register at E3H.
FWH_D0_IDSEL: This dictates the IDSEL of 512 KB of the FWH memory range starting at 4 GB -
3 MB (FFD00000H) to 4 GB - 2.5 MB (FFD7FFFFH) as well as register space starting at (4 GB-
4MB) - 3.0MB (FF900000h) to (4 GB-4MB) - 2.5MB (FF97FFFFh). The enable for this range is
controlled through bit 2 of the FWH Decode Enable Register at E3H.
11-14
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
Bit
Description
FWH_C8_IDSEL: This dictates the IDSEL of 512 KB of the FWH memory range starting at 4 GB -
3.5 MB (FFC80000H) to 4 GB - 3 MB (FFCFFFFFH) as well as register space starting at (4 GB-
4MB) - 3.5MB (FF880000h) to (4 GB-4MB) - 3MB (FF8FFFFFh). The enable for this range is
controlled through bit 1 of the FWH Decode Enable Register at E3H.
7:4
FWH_C0_IDSEL: This dictates the IDSEL of 512 KB of the FWH memory range starting at 4 GB -
4 MB (FFC00000H) to 4 GB - 3.5 MB (FFC7FFFFH) as well as register space starting at (4 GB-
4MB) – 4MB (FF800000h) to (4 GB-4MB) - 3.5MB (FF87FFFFh). The enable for this range is
controlled through bit 0 of the FWH Decode Enable Register at E3H.
3:0
11.1.27.3 Test Mode Register
Address:
FC-FFh
Default Value:
Attributes:
00000000h
Read/Write
Bit
Description
31:1
0
Reserved.
Alternate Access Mode Enable: When set, the part enters alternate access mode. This allows
reads to certain write-only registers and writes to certain read-only registers. Read to Port 70h will
return the NMI mask value. See Section 10.3 of the RS-IFB (I/O & Firmware Bridge) External
Design Specification, Rev 2.0, for more detail.
11.2
PCI to LPC I/O Space Registers
11.2.1
DMA Registers
The IFB contains DMA circuitry that incorporates the functionality of two 82C37 DMA controllers
(DMA1 and DMA2). The DMA registers control the operation of the DMA controllers and are all
accessible from the Host CPU via the PCI Bus interface.
11.2.1.1
Dcom–Dma Command Register (I/O)
I/O Address:
Default Value:
Attribute:
Channels 0-3–08h; Channels 4-7–0D0h
00h (CPURST or Master Clear)
Write Only
This 8-bit register controls the configuration of the DMA. Note that disabling channels 4-7 also
disables channels 0-3, since channels 0-3 are cascaded onto Channel 4.
Bit
Description
7
6
Reserved. Must be 0.
Reserved. Must be 0.
Reserved. Must be 0.
5
4
DMA Group Arbitration Priority. 1=Rotating priority; 0=Fixed priority
Reserved. Must be 0
3
2
DMA Channel Group Enable. 1=Disable; 0 = Enable.
Reserved. Must be 0.
1:0
Intel® 460GX Chipset Software Developer’s Manual
11-15
LPC/FWH Interface Configuration
11.2.1.2
Dcm–Dma Channel Mode Register (I/O)
I/O Address:
Default Value:
Attribute:
Channels 0-3=0Bh; Channels 4-7=0D6h
Bits[7:2]=0; Bits[1:0]=undefined (CPURST or Master Clear)
Write Only
Each channel has a 6-bit DMA Channel Mode Register. The Channel Mode Registers provide
control over DMA transfer type, transfer mode, address increment/decrement, and auto-
initialization.
Bit
Description
7:6
DMA Transfer Mode. Each DMA channel can be programmed in one of four different modes:
Bits[7:6] Transfer Mode
00 Demand Mode
01 Single Mode
10 Block Mode
11 I Cascade Mode
5
4
Address Increment/Decrement Select. 0=Increment; 1=Decrement.
Auto-initialize Enable. 1=Enable; 0=Disable.
3:2
DMA Transfer Type. When Bits [7:6]=11, the transfer type bits are irrelevant.
Bits[3:2] Transfer Type
00 Verify transfer
01 Write transfer
10 Read transfer
11 I Illegal
1:0
DMA Channel Select. Bits [1:0] select the DMA Channel Mode Register written to by bits [7:2].
Bits[1:0] Channel
00 Channel 0 (4)
01 Channel 1 (5)
10 Channel 2 (6)
11 Channel 3 (7)
11.2.1.3
Dr–Dma Request Register (I/O)
I/O Address:
Default Value:
Attribute:
Channels 0-3–09h; Channels 4-7–0D2h
Bits[1:0]=undefined; Bits[7:2]=0 (CPURST or Master Clear)
Write Only
The Request Register is used by software to initiate a DMA request. The DMA responds to the
software request as though DREQx is asserted. These requests are non-maskable and subject to
prioritization by the priority encoder network. For a software request, the channel must be in Block
Mode. The Request Register status for DMA1 and DMA2 is output on bits [7:4] of a Status
Register read.
Bit
Description
7:3
2
Reserved. Must be 0
DMA Channel Service Request. 0=Resets the individual software DMA channel request bit.
1=Sets the request bit. Generation of a TC also sets this bit to 0.
1:0
DMA Channel Select. Bits [1:0] select the DMA channel mode register to program with bit 2.
Bits[1:0] Channel
00 Channel 0
01 Channel 1 (5)
10 Channel 2 (6)
11 Channel 3 (7)
11-16
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
11.2.1.4
WSMB–Write Single Mask Bit (I/O)
I/O Address:
Default Value:
Attribute:
Channels 0-3–0Ah; Channels 4-7–0D4h
Bits[1:0]=undefined; Bit 2=1; Bits[7:3]=0 (CPURST or a Master Clear)
Write Only
A channel's mask bit is automatically set when the Current Byte/Word Count Register reaches
terminal count (unless the channel is programmed for auto-initialization). Setting the entire register
disables all DMA requests until a clear mask register instruction allows them to occur. This
instruction format is similar to the format used with the DMA Request Register. Masking DMA
channel 4 (DMA controller 2, channel 0) also masks DMA channels [3:0].
Bit
Description
7:3
2
Reserved. Must be 0.
Channel Mask Select. 1=Disable DREQ for the selected channel. 0=Enable DREQ for the
selected channel.
1:0
DMA Channel Select. Bits [1:0] select the DMA Channel Mode Register for bit 2.
Bits[1:0] Channel
00 Channel 0 (4)
01 Channel 1 (5)
10 Channel 2 (6)
11 Channel 3 (7)
11.2.1.5
RWAMB–Read / Write All Mask Bits (I/O)
I/O Address:
Default Value:
Attribute:
Channels 0-3–0Fh; Channels 4-7–0DEh
Bit[3:0]=1111; Bit[7:4]=0000 (CPURST or Master Clear)
Read/Write
A channel's mask bit is automatically set to 1 when the Current Byte/Word Count Register reaches
terminal count (unless the channel is programmed for auto-initialization). Setting bits [3:0] to 1
disables all DMA requests until a clear mask register instruction enables the requests. Note that,
masking DMA channel 4 (DMA controller 2, channel 0), masks DMA channels [3:0]. Also note
that, Masking DMA controller 2 with a write to port 0DEh also masks DREQ assertions from
DMA controller 1.
Bit
Description
7:4
3:0
Reserved. Must be 0.
Channel Mask Bits. 1=Disable the corresponding DREQ(s); 0=Enable the corresponding
DREQ(s).
Bit Channel
0 0 (4)
1 1 (5)
2 2 (6)
3 3 (7)
Intel® 460GX Chipset Software Developer’s Manual
11-17
LPC/FWH Interface Configuration
11.2.1.6
Ds–Dma Status Register (I/O)
I/O Address:
Default Value:
Attribute:
Channels 0-3–08h; Channels 4-7–0D0h
00h
Read Only
Each DMA controller has a read-only DMA Status Register that indicates which channels have
reached terminal count and which channels have a pending DMA request.
Bit
Description
7:4
Channel Request Status. When a valid DMA request is pending for a channel (on its DREQ
signal line), the corresponding bit is set to 1. When a DMA request is not pending for a
particular channel, the corresponding bit is set to 0. The source of the DREQ may be hardware
or a software request. Note that channel 4 does not have DREQ or DACK lines, so the
response for a read of DMA2 status for channel 4 is irrelevant.
Bit Channel
4 0
5 1 (5)
6 2 (6)
7 3 (7)
3:0
Channel Terminal Count Status. 1=TC is reached; 0=TC is not reached.
Bit Channel
0 0
1 1 (5)
2 2 (6)
3 3 (7)
11.2.1.7
DBADDR–DMA Base and Current Address Registers (I/O)
I/O Address:
DMA Channel 0–000h
DMA Channel 1–002h
DMA Channel 2–004h
DMA Channel 3–006h
DMA Channel 4–0C0h
DMA Channel 5–0C4h
DMA Channel 6–0C8h
DMA Channel 7–0CCh
Default Value:
Attribute:
Undefined (CPURST or Master Clear)
Read/Write
This Register works in conjunction with the Low Page Register. After an auto-initialization, this
register retains the original programmed value. Auto-initialize takes place after a TC. The address
register is automatically incremented or decrement after each transfer. This register is read/written
in successive 8-bit bytes. The programmer must issue the Clear Byte Pointer Flip-Flop command to
reset the internal byte pointer and correctly align the write prior to programming the Current
Address Register. Auto-initialize takes place only after a TC.
Bit
Description
15:0
Base and Current Address [15:0]. These bits represent address bits [15:0] used when
forming the 24-bit address for DMA transfers.
11-18
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
11.2.1.8
DBCNT–Dma Base and Current Count Registers (I/O)
I/O Address:
DMA Channel 0–001h
DMA Channel 1–003h
DMA Channel 2–005h
DMA Channel 3–007h
DMA Channel 4–0C2h
DMA Channel 5–0C6h
DMA Channel 6–0CAh
DMA Channel 7–0CEh
Default Value:
Attribute:
Undefined (CPURST or Master Clear)
Read/Write
This register determines the number of transfers to be performed. The actual number of transfers is
one more than the number programmed in the Current Byte/Word Count Register When the value
in the register is decremented from zero to FFFFh, a TC is generated. Auto-initialize can only
occur when a TC occurs. If it is not auto-initialized, this register has a count of FFFFh after TC.
For transfers to/from an 8-bit I/O, the Byte/Word count indicates the number of bytes to be
transferred. This applies to DMA channels 0-3. For transfers to/from a 16-bit I/O, with shifted
address, the Byte/Word count indicates the number of 16-bit words to be transferred. This applies
to DMA channels 5-7.
Bit
Description
15:0
Base and Current Byte/ Word Count. These bits represent the 16 byte/word count bits used
when counting down a DMA transfer.
11.2.1.9
DLPAGE–DMA Low Page Registers (I/O)
I/O Address:
DMA Channel 0–087h
DMA Channel 1–083h
DMA Channel 2–081h
DMA Channel 3–082h
DMA Channel 5–08Bh
DMA Channel 6–089h
DMA Channel 7–08Ah
Default Value:
Attribute:
Undefined (CPURST or Master Clear)
Read/Write
This register works in conjunction with the Current Address Register. After an auto-initialization,
this register retains the original programmed value. Auto-initialize takes place after a TC.
Bit
Description
7:0
DMA Low Page [23:16]. These bits represent address bits [23:16] of the 24-bit DMA address.
11.2.1.10 DCBP–Dma Clear Byte Pointer Register (I/O)
I/O Address:
Default Value:
Attribute:
Channels 0-3–00Ch; Channels 4-7–0D8h
All bits undefined
Write Only
Writing to this register executes the Clear Byte Pointer Command. This command is executed prior
to reading/writing a new address or word count to the DMA. The command initializes the byte
pointer flip-flop to a known state so that subsequent accesses to register contents address upper and
lower bytes in the correct sequence. The Clear Byte Pointer Command (or CPURST or the Master
Clear Command) clears the internal latch used to address the upper or lower byte of the 16-bit
Address and Word Count Registers.
Intel® 460GX Chipset Software Developer’s Manual
11-19
LPC/FWH Interface Configuration
Bit
Description
7:0
Clear Byte Pointer. No specific pattern. Command enabled with a write to the I/O port
address.
11.2.1.11 Dmc–Dma Master Clear Register (I/O)
I/O Address:
Default Value:
Attribute:
Channel 0-3–00Dh; Channel 4-7–0DAh
All bits undefined
Write Only
This software instruction has the same effect as the hardware Reset.
Bit
Description
7:0
Master Clear. No specific pattern. Command enabled with a write to the I/O port address
11.2.1.12 Dclm–Dma Clear Mask Register (I/O)
I/O Address:
Default Value:
Attribute:
Channel 0-3–00Eh; Channel 4-7–0DCh
All bits undefined
Write Only
This command clears the mask bits of all four channels, enabling them to accept DMA requests.
Bit
Description
7:0
Clear Mask Register. No specific pattern. Command enabled with a write to the I/O port
address.
11.2.2
Interrupt Controller Registers
The IFB contains an interrupt controller that incorporates the Functionality of two 82C59 interrupt
controllers. The interrupt registers control the operation of the interrupt controller.
11.2.2.1
Icw1–Initialization Command Word 1 Register (I/O)
I/O Address:
Default Value:
Attribute:
INT CNTRL-1–020h; INT CNTRL-2–0A0h
All bits undefined
Write Only
A write to Initialization Command Word 1 starts the interrupt controller initialization sequence.
Addresses 020h and 0A0h are referred to as the base addresses of CNTRL-1 and CNTRL-2,
respectively. An I/O write to the CNTRL-1 or CNTRL-2 base address with bit 4 equal to 1 is
interpreted as ICW1. For IFB-based systems, three I/O writes to “base address + 1" must follow the
ICW1. The first write to “base address + 1" performs ICW2, the second write performs ICW3, and
the third write performs ICW4.
ICW1 starts the initialization sequence during which the following automatically occur:
1. The Interrupt Mask register is cleared.
2. IRQ7 input is assigned priority 7.
3. The slave mode address is set to 7.
4. Special Mask Mode is cleared and Status Read is set to IRR.
11-20
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
Bit
Description
7:5
4
ICW/OCW select. These bits should be 000 when programming the IFB.
ICW/OCW select. Bit 4 must be a 1 to select ICW1. After the fixed initialization sequence to
ICW1, ICW2, ICW3, and ICW4, the controller base address is used to write to OCW2 and
OCW3. Bit 4 is a 0 on writes to these registers. A 1 on this bit at any time will force the interrupt
controller to interpret the write as an ICW1. The controller will then expect to see ICW2, ICW3,
and ICW4.
3
Edge/Level Bank Select (LTIM). This bit is disabled. Its Function is replaced by the Edge/
Level Triggered Control (ELCR) Registers.
2
1
0
ADI. Ignored for the IFB. This bit should be programmed to ‘0’.
Single or Cascade (SNGL). This bit must be programmed to a 0.
ICW4 Write Required (IC4). This bit must be set to a 1.
11.2.2.2
Icw2–Initialization Command Word 2 Register (I/O)
I/O Address:
Default Value:
Attribute:
INT CNTRL-1–021h; INT CNTRL-2–0A1h
All bits undefined
Write Only
ICW2 is used to initialize the interrupt controller with the five most significant bits of the interrupt
vector address.
Bit
Description
7:3
Interrupt Vector Base Address. Bits [7:3] define the base address in the interrupt vector table
for the interrupt routines associated with each interrupt request level input.
2:0
Interrupt Request Level. Must be programmed to all 0s.
11.2.2.3
Icw3–Initialization Command Word 3 Register (I/O)
I/O Address:
Default Value:
Attribute:
INT CNTRL-1–021h
All bits undefined
Write Only
The meaning of ICW3 differs between CNTRL-1 and CNTRL-2. On CNTRL-1, the master
controller, ICW3 indicates which CNTRL-1 IRQ line physically connects the INTR output of
CNTRL-2 to CNTRL-1.
Bit
Description
Reserved. Must be programmed to all 0s.
7:3
2
Cascaded Mode Enable. This bit must be programmed to 1 selecting cascade mode.
1:0
Reserved. Must be programmed to all 0s.
Intel® 460GX Chipset Software Developer’s Manual
11-21
LPC/FWH Interface Configuration
11.2.2.4
Icw3–Initialization Command Word 3 Register (I/O)
I/O Address:
Default Value:
Attribute:
INT CNTRL-2–0A1h
All bits undefined
Write Only
On CNTRL-2 (the slave controller), ICW3 is the slave identification code broadcast by CNTRL-1.
Bit
Description
7:3
2:0
Reserved. Must be programmed to all 0s.
Slave Identification Code. Must be programmed to 02h.
11.2.2.5
Icw4–Initialization Command Word 4 Register (I/O)
I/O Address:
Default Value:
Attribute:
INT CNTRL-1–021h; INT CNTRL-2–0A1h
01h
Write Only
Both IFB interrupt controllers must have ICW4 programmed as part of their initialization
sequence.
Bit
Description
Reserved. Must be programmed to all 0s.
7:5
4
Special Fully Nested Mode (SFNM). Bit 4, SFNM, should normally be disabled by writing a 0
to this bit. If SFNM=1, the special fully nested mode is programmed.
3
2
1
Buffered mode (BUF). Must be programmed to 0 selecting non-buffered mode.
Master/Slave in Buffered Mode. Should always be programmed to 0. Bit not used.
AEOI (Automatic End of Interrupt). This bit should normally be programmed to 0. This is the
normal end of interrupt. If this bit is 1, the automatic end of interrupt mode is programmed.
0
Microprocessor Mode. Must be programmed to 1 indicating an Intel Architecture-based
system.
11.2.2.6
Ocw1–Operational Control Word 1 Register (I/O)
I/O Address:
Default Value:
Attribute:
INT CNTRL-1–021h; INT CNTRL-2–0A1h
00h
Read/Write
OCW1 sets and clears the mask bits in the Interrupt Mask Register (IMR). Each interrupt request
line may be selectively masked or unmasked any time after initialization. The IMR stores the
interrupt line mask bits. The IMR operates on the IRR. Masking of a higher priority input does not
affect the interrupt request lines of lower priority. Unlike status reads of the ISR and IRR, for
reading the IMR, no OCW3 is needed. The output data bus contains the IMR when an I/O read is
active and the I/O address is 021h or 0A1h (OCW1). All writes to OCW1 must occur following the
ICW1-ICW4 initialization sequence, since the same I/O ports are used for OCW1, ICW2, ICW3
and ICW4.
11-22
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
Bit
Description
7:0
Interrupt Request Mask (Mask [7:0]). When a 1 is written to any bit in this register, the
corresponding IRQx line is masked. For example, if bit 4 is set to a 1, then IRQ4 is masked.
Interrupt requests on IRQ4 do not set channel 4’s interrupt request register (IRR) bit as long is
the channel is masked. When a 0 is written to any bit in this register, the corresponding IRQx is
unmasked. Note that masking IRQ2 on CNTRL-1 also masks the interrupt requests from
CNTRL-2, which is physically cascaded to IRQ2.
11.2.2.7
Ocw2–Operational Control Word 2 Register (I/O)
I/O Address:
Default Value:
Attribute:
INT CNTRL-1–020h; INT CNTRL-2–0A0h
Bit[4:0]=undefined; Bit[7:5]=001
Write Only
OCW2 controls both the Rotate Mode and the End of Interrupt Mode. Following a CPURST or
ICW initialization, the controller enters the fully nested mode of operation. Both rotation mode and
specific EOI mode are disabled following initialization.
Bit
Description
7:5
Rotate and EOI Codes. R, SL, EOI - These three bits control the Rotate and End of Interrupt
modes and combinations of the two.
Bits[7:5] Function Bits[7:5] Function
001 Non-specific EOI Cmd 000 Rotate in Auto EOI Mode (Clear)
011 Specific EOI Cmd 111 *Rotate on Specific EOI Cmd
101 Rotate on Non-Spec EOI Cmd 110 *Set Priority Cmd
100 Rotate in Auto EOI Mode (Set) 010 No Operation
* L0 - L2 Are Used
4:3
2:0
OCW2 Select. Must be programmed to 00 selecting OCW2.
Interrupt Level Select (L2, L1, L0). L2, L1, and L0 determine the interrupt level acted upon
when the SL bit is active (bit 6). When the SL bit is inactive, bits [2:0] do not have a defined
Function; programming L2, L1 and L0 to 0 is sufficient in this case.
Bit[2:0] Interrupt Level Bit[2:0] Interrupt Level
000 IRQ 0(8) 100 IRQ 4(12)
001 IRQ 1(9) 101 IRQ 5(13)
010 IRQ 2(10) 110 IRQ 6(14)
011 IRQ 3(11) 111 IRQ 7(15)
11.2.2.8
Ocw3–Operational Control Word 3 Register (I/O)
I/O Address:
Default Value:
Attribute:
INT CNTRL-1–020h; INT CNTRL-2–0A0h
Bit[6,0]=0; Bit[7,4:2]=Undefined; Bit[5,1]=1
Read/Write
OCW3 serves three important Functions–Enable Special Mask Mode, Poll Mode control, and IRR/
ISR register read control.
Bit
Description
7
6
Reserved. Must be 0.
Special Mask Mode (SMM). If ESMM=1 and SMM=1, the interrupt controller enters Special
Mask Mode. If ESMM=1 and SMM=0, the interrupt controller is in normal mask mode. When
ESMM=0, SMM has no effect.
Intel® 460GX Chipset Software Developer’s Manual
11-23
LPC/FWH Interface Configuration
Bit
Description
5
4:3
2
Enable Special Mask Mode (ESMM). 1=Enable SMM bit; 0=Disable SMM bit.
OCW3 Select. Must be programmed to 01 selecting OCW3.
Poll Mode Command. 0=Disable Poll Mode Command. When bit 2=1, the next I/O read to the
interrupt controller is treated as an interrupt acknowledge cycle indicating highest priority
request.
1:0
Register Read Command. Bits [1:0] provide control for reading the In-Service Register (ISR)
and the Interrupt Request Register (IRR). When bit 1=0, bit 0 does not affect the register read
selection. When bit 1=1, bit 0 selects the register status returned following an OCW3 read. If bit
0=0, the IRR will be read. If bit 0=1, the ISR will be read. Following ICW initialization, the
default OCW3 port address read will be “read IRR”. To retain the current selection (read ISR or
read IRR), always write a 0 to bit 1 when programming this register. The selected register can
be read repeatedly without reprogramming OCW3. To select a new status register, OCW3
must be reprogrammed prior to attempting the read.
Bit[1:0] Function
00 No Action
01 No Action
10 Read IRQ Register
11 Read IS Register
11.2.2.9
Elcr1–Edge/Level Control Register (I/O)
I/O Address:
Default Value:
Attribute:
INT CNTRL-1–4D0h
00h
Read/Write
ELCR1 register allows IRQ3 - IRQ7 to be edge or level programmable on an interrupt by interrupt
basis. IRQ0, IRQ1 and IRQ2 are not programmable and are always edge sensitive. When level
triggered, the interrupt is signaled active when input IRQ signal is high.
Bit
Description
7
6
IRQ7 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
IRQ6 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
IRQ5 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
IRQ4 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
IRQ3 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
Reserved. Must be 0.
5
4
3
2:0
11-24
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
11.2.2.10 Elcr2–Edge/Level Control Register (I/O)
I/O Address:
Default Value:
Attribute:
INT CNTRL-2–4D1h
00h
Read/Write
ELCR2 register allows IRQ[15,14,12:9] to be edge or level programmable on an interrupt by
interrupt basis. Note that, IRQ[13,8#] are not programmable and are always edge sensitive. When
level triggered, the interrupt is signaled active when input IRQ signal is high.
Bit
Description
7
6
5
4
3
2
1
0
IRQ15 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
IRQ14 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
Reserved. Must be 0.
IRQ12 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
IRQ11 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
IRQ10 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
IRQ9 ECL. 0 = Edge Triggered mode; 1 = Level Triggered mode.
Reserved. Must be 0.
11.2.3
Counter/Timer Registers
11.2.3.1
Tcw–Timer Control Word Register (I/O)
I/O Address:
Default Value:
Attribute:
043h
All bits undefined
Write Only
Bit
Decription
Counter Select. The Read Back Command is selected when bits[7:6] are both 1.
7:6
Bit[7:6] Function
00 Counter 0 select
01 Counter 1 select
Bit[7:6] Function
10 Counter 2 select
11 Read Back Command
5:4
3:1
Read/Write Select. The Counter Latch Command is selected when bits[5:4] are both 0.
Bit[5:4] Function
Bit[5:4] Function
00 Counter Latch Command
10 R/W Most Significant Byte
01 R/W Least Significant Byte 11 R/W LSB then MSB
Counter Mode Selection. Bits [3:1] select one of six possible counter modes.
Bit[3:1] Mode Function
000 0 Out signal on end of count (=0)
001 1 Hardware re-triggerable one-shot
X10 2 Rate generator (divide by n counter)
X11 3 Square wave output
100 4 Software triggered strobe
101 5 Hardware triggered strobe
0
Binary/BCD Countdown Select. 0=Binary countdown. The largest possible binary count is
216. 1=Binary coded decimal (BCD) count is used. The largest BCD count allowed is 104.
Intel® 460GX Chipset Software Developer’s Manual
11-25
LPC/FWH Interface Configuration
The Timer Control Word Register specifies the counter selection, the operating mode, the counter
byte programming order and size of the count value, and whether the counter counts down in a 16-
bit or binary-coded decimal (BCD) format. After writing the control word, a new count can be
written at any time. The new value takes effect according to the programmed mode.
Read Back Command
The Read Back Command is used to determine the count value, programmed mode, and current
states of the OUT pin and Null count flag of the selected counter or counters. The Read Back
Command is written to the Timer Control Word Register which latches the current states of the
above mentioned variables. The value of the counter and its status may then be read by I/O access
to the counter address. Note that The Timer Counter Register bit definitions are different during the
Read Back Command than for a normal Timer Counter Register write.
Bit
Description
7:6
Read Back Command. When bits[7:6]=11, the Read Back Command is selected during a write
to the Timer Control Word Register. Following the Read Back Command, I/O reads from the
selected counter’s I/O addresses produce the current latch status, the current latched count, or
both if bits 4 and 5 are both 0.
5
4
3
2
1
0
Latch Count of Selected Counters. When bit 5=0, the current count value of the selected
counters will be latched. When bit 5=1, the count will not be latched.
Latch Status of Selected Counters. When bit 4=0, the status of the selected counters will be
latched. When bit 4=1, the status will not be latched.
Counter 2 Select. When bit 3=1, Counter 2 is selected for the latch command selected with bits
4 and 5. When bit 3=0, status and/or count will not be latched.
Counter 1 Select. When bit 2=1, Counter 1 is selected for the latch command selected with bits
4 and 5. When bit 2=0, status and/or count will not be latched.
Counter 0 Select. When bit 1=1, Counter 0 is selected for the latch command selected with bits
4 and 5. When bit 1=0, status and/or count will not be latched.
Reserved. Must be 0.
Counter Latch Command
Bit
Description
7:6
Counter Selection. Bits 6 and 7 are used to select the counter for latching.
Bit[7:6] Function
00 latch counter 0 select
01 latch counter 1 select
10 latch counter 2 select
11 Read Back Command select
5:4
3:0
Counter Latch Command. When bits[5:4]=00, the Counter Latch Command is selected during
a write to the Timer Control Word Register. Following the Counter Latch Command, I/O reads
from the selected counter’s I/O addresses produce the current latched count.
Reserved. Must be 0.
The Counter Latch Command latches the current count value at the time the command is received.
If a Counter is latched once and then, some time later, latched again before the count is read, the
second Counter Latch Command is ignored. The count read will be the count at the time the first
Counter Latch Command was issued. If the counter is programmed for two byte counts, two bytes
must be read. The two bytes do not have to be read successively (read, write, or programming
operations for other counters may be inserted between the reads). Note that the Timer Counter
11-26
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
Register bit definitions are different during the Counter Latch Command than for a normal Timer
Counter Register write. Note that, If a counter is programmed to read/write two-byte counts, a
program must not transfer control between reading the first and second byte to another routine that
also reads from that same counter. Otherwise, an incorrect count will be read.
11.2.3.2
TMRSTS–Timer Status Registers (I/O)
I/O Address:
Default Value:
Attribute:
Counter 0–040h; Counter 1–041h; Counter 2–042h
Bits[6:0]=X; Bit 7=0
Read Only
Each counter's status byte can be read following an Interval Timer Read Back Command. If latch
status is chosen (bit 4=0, Read Back Command) as a read back option for a given counter, the next
read from the counter's Counter Access Ports Register returns the status byte.
Bit
Description
Counter OUT Pin State. 1=Pin is 1; 0=Pin is 0.
7
6
Count Register Status. This bit indicates when the last count written to the Count Register
(CR) has been loaded into the counting element (CE). 0=Count has been transferred from CR
to CE and is available for reading. 1=Count has not been transferred from CR to CE and is not
yet available for reading.
5:4
Read/Write Selection Status. Bits[5:4] reflect the read/write selection made through bits[5:4] of
the Control Register.
Bit[5:4] Function
00 Counter Latch Command
01 R/W Least Significant Byte (LSB)
10 R/W Most Significant Byte (MSB)
11 R/W LSB then MSB
3:1
Mode Selection Status. Bits[3:1] return the counter mode programming.
Bit[3:1] Mode Selected Bit[3:1] Mode Selected
000 0 X11 3
001 1 100 4
X10 2 101 5
0
Countdown Type Status. 0=Binary countdown; 1=Binary coded decimal (BCD) countdown.
11.2.3.3
TMRCNT–Timer Count Registers (I/O)
I/O Address:
Default Value:
Attribute:
Counter 0–040h; Counter 1–041h; Counter 2–042h
All bits undefined
Read/Write
Each of these I/O ports is used for writing count values to the Count Registers; reading the current
count value from the counter by either an I/O read, after a counter-latch command, or after a Read
Back Command; and reading the status byte following a Read Back Command.
Bit
Description
7:0
Counter Port Bits]. Each counter I/O port address is used to program the 16-bit Count
Register. The order of programming, either LSB only, MSB only, or LSB then MSB, is defined
with the Interval Counter Control Register. The counter I/O port is also used to read the current
count from the Count Register and return counter programming status following a Read Back
Command.
Intel® 460GX Chipset Software Developer’s Manual
11-27
LPC/FWH Interface Configuration
11.2.4
NMI Registers
The NMI logic incorporates two different 8-bit registers. The CPU reads the NMISC Register to
determine the NMI source (bits set to a 1). After the NMI interrupt routine processes the interrupt,
software clears the NMI status bits by setting the corresponding enable/disable bit to a 1. The NMI
Enable and Real-Time Clock Register can mask the NMI signal and disable/enable all NMI
sources.
To ensure that all NMI requests are serviced, the NMI service routine software flow should be as
follows:
1. NMI is detected by the processor on the rising edge of the NMI input.
2. The processor will read the status stored in port 061h to determine what sources caused the
NMI. The processor may then set to 0 the register bits controlling the sources that it has
determined to be active. Between the time the processor reads the NMI sources and sets them
to a 0, an NMI may have been generated by another source. The level of NMI will then remain
active. This new NMI source will not be recognized by the processor because there was no
edge on NMI.
3. The processor must then disable all NMIs by setting bit 7 of port 070H to a 1 and then enable
all NMIs by setting bit 7 of port 070H to a 0. This will cause the NMI output to transition low
then high if there are any pending NMI sources. The CPU’s NMI input logic will then register
a new NMI.
11.2.4.1
Nmisc–Nmi Status and Control Register (I/O)
I/O Address:
Default Value:
Attribute:
061h
00h
Read/Write
This register reports the status of different system components, controls the output of the speaker
counter (Counter 2), and gates the counter output that drives the SPKR signal.
Bit
Description
7
SERR# NMI Source Status–RO. Bit 7 is set if a system board agent (PCI devices or main
memory) detects a system board error and pulses the PCI SERR# line. This interrupt source is
enabled by setting bit 2 to 0. To reset the interrupt, set bit 2 to 0 and then set it to 1. When writing
to port 061h, bit 7 must be 0.
6
5
reserved (0)
Timer Counter 2 OUT Status–RO. The Counter 2 OUT signal state is reflected in bit 5. The
value on this bit following a read is the current state of the Counter 2 OUT signal. Counter 2
must be programmed following a CPURST for this bit to have a determinate value. When writing
to port 061h, bit 5 must be a 0.
4
Refresh Cycle Toggle–RO. The Refresh Cycle Toggle signal toggles from either 0 to 1 or 1 to 0
following every refresh cycle. When writing to port 061h, bit 4 must be a 0.
3
2
Must be programmed to ‘1’.
PCI SERR# Enable–R/W. 1=Clear and Disable; 0=Enable.
For the IFB, the SERR# signal can be for a special protocol between the host-to-PCI bridge and
the IFB (see MSTAT Register description, 6Ah-6Bh, Function 0).
1
0
Speaker Data Enable–R/W. 0=SPKR output is 0; 1= the SPKR output is the Counter 2 OUT
signal value.
Timer Counter 2 Enable–R/W. 0=Disable; 1=Enable.
11-28
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
11.2.4.2
NmiEN–Nmi Enable Register (Shared with Real-time Clock Index
Register) (I/O)
I/O Address:
Default Value:
Attribute:
070h
Bit[6:0]=undefined; Bit 7=1
Write Only
This port is shared with the real-time clock. Do not modify the contents of this register without
considering the effects on the state of the other bits.
Bit
Description
7
NMI Enable. 1=Disable generation of NMI; 0=Enable generation of NMI.
6:0
Real Time Clock Address. Used by the Real Time Clock to address memory locations. Not
11.2.5
Real Time Clock Registers
11.2.5.1
RTCI–Real-time Clock Index Register (Shared with NMI Enable
Register) (I/O)
I/O Address:
Default Value:
Attribute:
070h
Bit[6:0]=Undefined; Bit 7=1
Write Only
This port is shared with the NMI enable. Do not modify the contents of this register without
considering the effects on the state of the other bits.
Bit
Description
7
NMI Enable. Used by IFB NMI logic.
6:0
Real Time Clock Address. Latched by the Real Time Clock to address memory locations
within the standard RAM bank accessed via the Real Time Clock Data Register (071h).
11.2.5.2
RTCD–Real-time Clock Data Register (I/O)
I/O Address:
Default Value:
Attribute:
071h
Undefined
Read/Write
The data port for accesses to the RTC standard RAM bank.
Bit
Description
7:0
Standard RAM Data Port. Data written to standard RAM bank address selected via RTC Index
Register (070h).
Intel® 460GX Chipset Software Developer’s Manual
11-29
LPC/FWH Interface Configuration
11.2.5.3
RTCEI–Real-time Clock Extended Index Register (I/O)
I/O Address:
Default Value:
Attribute:
072h
Unknown
Write Only
The index port for accesses to the RTC extended RAM bank.
Bit
Description
7
Reserved.
6:0
Real Time Clock Extended Address. Latched by the Real Time Clock to address memory
locations within the extended RAM bank accessed via the Real Time Clock Extended Data
Register (073h).
11.2.5.4
RTCED–Real-time Clock Extended Data Register (I/O)
I/O Address:
Default Value:
Attribute:
073h
Unknown
Read/Write
The data port for accesses to the RTC extended RAM bank.
Bit
Description
7:0
Extended RAM Data Port. Data written to standard RAM bank address selected via RTC
Extended Index Register (072h).
11.2.6
Advanced Power Management (APM) Registers
This section describes two power management registers–APMC and APMS Registers. These
registers are located in normal I/O space and must be accessed (via the PCI Bus) with 8 bit
accesses.
11.2.6.1
APMC–Advanced Power Management Control Port (I/O)
I/O Address:
Default Value:
Attribute:
0B2h
00h
Read/Write
This register passes data (APM Commands) between the OS and the SMI handler. In addition,
writes can generate an SMI. The IFB operation is not effected by the data in this register.
Bit
Description
7:0
APM Control Port (APMC). Writes to this register store data in the APMC Register and reads
return the last data written. In addition, writes generate an SMI, if the APMC_EN bit (PCI
Function 3, offset58h, bit 25) is set to 1. Reads do not generate an SMI.
11-30
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
11.2.6.2
APMS–Advanced Power Management Status Port (I/O)
I/O Address:
Default Value:
Attribute:
0B3h
00h
Read/Write
This register passes status information between the OS and the SMI handler. The IFB operation is
not effected by the data in this register.
Bit
Description
7:0
APM Status Port (APMS). Writes store data in this register and reads return the last data
written.
11.2.7
ACPI Registers
The ACPI registers are I/O mapped. The base address is set via PCI Function 0 configuration
register 40h. The registers are defined to be compliant with the ACPI 1.0 specification, and use the
same bit names. All reserved bits will always return 0 when read, and will have no effect when
written.
11.2.7.1
Power Management 1 Status
Address Offset:
Attributes:
Default Value:
Size:
00-01h
Read/Write
Bit 11:Undefined, All other bits ‘0’
16 bits
Bit
Description
15
WAK_STS: This bit is set when the system is in one of the Sleep states (via the SLP_EN bit)
and an enabled IFB wake/break event occurs. Upon setting this bit, IFB will transition the
system to the ON state. This bit can only be set by hardware and can only be cleared by writing
a one to this bit position. This bit is not affected by a hard reset caused by a CF9 write.
14:12
11
Reserved.
PWRBTNOR_STS: This bit is set any time a Power Button Override Event occurs. The override
event occurs when the power button is pressed for 4 consecutive seconds. The power button
override will cause an unconditional transition to the S5 state, as well as set the AFTERG3 bit.
The firmware or SCI handler can clear this bit by writing a 1 to it. This bit is not affected by a
hard reset caused by a CF9 write. Upon reset, this bit is undefined.
10
RTC_STS: This bit is set when the RTC generates an alarm (assertion of the IRQ8# signal).
Additionally if the RTC_EN bit is set then the setting of the RTC_STS bit will generate wake
event. This bit is only set by hardware and can only be reset by writing a one to this bit position.
This bit is not affected by a hard reset caused by a CF9 write.
9
8
Reserved.
PWRBTN_STS: This bit is set by the hardware when the power button is pressed. The firmware
or SCI handler should clear this bit by writing a 1 to it. This bit is not affected by a hard reset
caused by a CF9 write.
7:6
5
Reserved.
GBL_STS: This bit is set when an SCI is generated due to the firmware wanting the attention of
the SCI handler. Firmware has a corresponding bit, BIOS_RLS, which will cause an SCI and set
this bit. The SCI handler should then clear this bit by writing a 1 to it.
4:1
Reserved.
Intel® 460GX Chipset Software Developer’s Manual
11-31
LPC/FWH Interface Configuration
Bit
Description
nd
0
TMROF_STS: This is the timer overflow status bit. This bit gets set anytime the 22 bit of the
24 bit timer goes from high to low (bits are counted from 0 to 23). This will occur every 2.3435
seconds. When the TMROF_EN bit is set, the setting of the TMROF_STS bit will generate an
SCI or SMI. SMI will be generated if ACPI_TMR_EN, SMI_EN and TMROF_EN are set, and
SCI_EN is not set. SCI will be generated if SCI_EN and TMROF_EN are set. The SCI or SMI
handler clears this bit by writing a 1 to it.
11.2.7.2
Power Management 1 Enable
Address Offset:
Attributes:
Default Value:
Size:
02-03h
Read/Write
Bit 10: Undefined, All other bits ‘0’
16 bits
Bit
Description
15:11
10
Reserved.
RTC_EN: This bit is used to enable the setting of the RTC_STS bit to generate a wake event.
The RTC_STS bit is set anytime the RTC generates an alarm (asserts the IRQ8# signal active).
The enable bit does not have to be set to enable the setting of the RTC_STS bit by the assertion
of the RTC alarm. The value of this bit must be maintained, even through a G3 state. The IFB
will not resume, from RTC, after power failure (RSMRST# low) even if this bit is set. This bit is
automatically cleared by a power button override. The IFB can resume from RTC if only
PWROK goes low with RSMRST# high. Upon reset, this bit is undefined.
9
8
Reserved.
PWRBTN_EN: This bit is set to 1 to enable the setting of the PWRBTN_STS bit to also
generate an SCI or SMI. Even if the PWRBTN_EN bit is set to 0, the power button can always
generate a Wake event (if in a n S1-S5 state), and the power button override can cause an
unconditional transition to the S5 state.
7:6
5
Reserved.
GBL_EN: Global enable bit. When both GBL_EN and GBL_STS are set, an SCI event is
generated.
4:1
0
Reserved.
TMROF_EN: This is the timer overflow interrupt enable bit. When this bit is set then an SCI or
SMI is generated anytime the TMOF_STS bit is also set. When this bit is reset, then no interrupt
is generated when the TMROF_STS bit is set.
11.2.7.3
Power Management 1 Control
Address Offset:
Type:
04-05h
Read/Write
Default Value:
Size:
Bits 12:10 Undefined, All other bits ‘0’
16 bits
Bit
Description
15:14
13
Reserved.
SLP_EN: This is a write-only bit and reads to it always return a zero. Setting this bit causes the
system to sequence into the Sleep state defined by the SLP_TYP field.
12:10
SLP_TYP: This 3-bit field defines the type of Sleep the system should enter when the SLP_EN bit
is set to 1. IFB doesn’t directly support any modes other than those listed in this table. Other Sleep
states can be supported in external logic.
11-32
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
Bit
Description
Mode
Bits
000
ON
001
010
011
100
Typically mapped to S1 state. STPCLK# active. CPU in Stop-
Grant state. Equivalent to Level 2.
Typically mapped to S1 state. Both STPCLK# and SLP#
signals active. CPU in Sleep state. Equivalent to Level 3.
S3 state. This is also known as Suspend-To-RAM (STR).
The SUSB signal will go active.
S4/S5 state. The S4 state is also known as Suspend-To-Disk
(STD). The S5 state is also known as Soft-Off.
All other combinations are reserved. If a value is written other than those shown above, the IFB
will ignore the value and stay in the ON (000) state. However, the last written value will be
readable. For example, if software writes 111 (reserved value), IFB will stay in the ON state, but
the next read to the SLP_TYP field will return 111, not 000. Upon reset, this bit is undefined.
9:3
2
Reserved.
GBL_RLS: This bit is used by the ACPI software to generate an SMI to the firmware. Firmware
has corresponding enable and status bits to control its ability to receive ACPI events.
1
0
Reserved.
SCI_EN: Selects the SCI interrupt for the THRM_STS and Timer. When this bit is 1, then these
events will generate an SCI interrupt. When this bit is 0, these events will generate an SMI#.
11.2.7.4
Power Management 1 Timer
Address Offset:
Attributes:
Default Value:
Size:
08-0Bh
Read Only
00000000h
32 bits
Bit
Description
31:24
23:0
Reserved.
TMR_VAL: This read-only field returns the running count of the power management timer. This is a
24-bit counter that runs off a 3.579545 MHz clock. The timer is reset to an initial value of zero during
a PCI reset, starts running immediately after PCI reset, and then continues counting until the
14.31818 MHz input to the chip is stopped (which can only occur in STR/STD and the power is
nd
removed to the IFB). Anytime the 22 bit of the timer goes HIGH to LOW (bits referenced from 0 to
23), the TMROF_STS bit is set. If the TMROF_EN bit is set, an SCI or SMI is also generated.
11.2.7.5
General Purpose 0 Status
Address Offset:
Attributes:
Default Value:
Size:
0C-0Dh
Read/Write
0800h
16 bits
When any bit is set in this register, and the corresponding bit is enabled in the General Purpose 0
Enable register, an SCI and a wake event will be generated.
Intel® 460GX Chipset Software Developer’s Manual
11-33
LPC/FWH Interface Configuration
Bit
Description
15:12
11
Reserved.
PWR_FAIL: This bit will be set to 1 when a power failure occurs. This is defined as either PWROK
or RSMRST# going inactive unexpectedly. This bit is only set by hardware and can only be reset by
writing a one to this bit position. This bit is not affected by a hard reset caused by a CF9 write. Upon
power up, this bit is set to ‘1’.
10
RI_STS: This bit will be set by 1 to hardware when the RI# input signal goes active. This bit can be
reset only by writing a one to this bit position. This bit is not affected by a hard reset caused by a
CF9 write.
9
8
Reserved.
USB1_STS: This bit is set when the USB controller for ports 0 and 1 needs to cause a wake/break
event. Additionally if the USB1_EN bit is set, the setting of the USB1_STS bit will generate a wake/
break event. This bit is only set by hardware and can only be reset by writing a one to this bit
position. This bit is not affected by a hard reset caused by a CF9 write.
7
THRMOR_STS: This is the thermal interrupt over-ride status. This bit is 1 anytime a thermal over-
ride condition occurs and starts throttling the CPU’s clock at the THRM_DTY ratio. This bit is set by
hardware and can only be cleared by writing a one to this bit position.
6:2
1
Reserved.
NMI_STS: This indicates that an external device generated SERR# or IOCHK#. If the NMI_EN bit is
set in the General Purpose 0 Enable Register, then the setting of this bit will generate an SCI. This
bit is set by hardware and can only be cleared by writing a ‘1’ to this bit position.
0
THRM_STS: This is the thermal interrupt status bit. This bit gets set anytime the THRM# signal is
driven active as defined by the THRM_POL bit. Additionally if the THRM_EN bit is set then the
setting of the THRM_STS bit will additionally generate a power management event (SCI or SMI).
This bit is only set through hardware (the THRM# signal being driven active), and is cleared by
software writing a one to this bit position.
11.2.7.6
General Purpose 0 Enable
Address Offset:
Attributes:
Default Value:
Size:
OE-0Fh
Read/Write
0000h
16 bits
Bit
Description
15:13
12
Reserved.
AFTERG3: If set to 0, the IFB will boot the system after the power is returned after a power
failure. If set to 1, the IFB will not boot the system after a power failure and will wait for a wake
event (such as the Power Button being pressed). This bit is automatically set to 1 if a Power
Button Override occurs. Upon power up, this bit is undefined.
11
10
Reserved.
RI_EN: When RI_EN and RI_STS are both set, a Wake event will occur. If RI_EN is not set,
then when RI_STS is set, no Wake event will occur. This bit must be maintained even in the G3
state. This bit is automatically set to 0 if a Power Button Override occurs. Upon power up, this
bit is undefined.
9
8
Reserved.
USB1_EN: This bit is used to enable the setting of the USB1_STS bit to generate a wake/break
event. The USB1_STS bit is set anytime the USB controller for ports 0 and 1 signals a wake/
break event. The value of this bit must be maintained, even through a G3 state. The IFB will not
resume, from USB, after power failure (RSMRST# low). This bit is automatically set to 0 if a
Power Button Override occurs. Upon power up, this bit is undefined.
7:2
Reserved.
11-34
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
Bit
Description
1
NMI_EN: This bit enables an SCI to be generated on a NMI event. Upon power up, this bit is set
to ‘0’.
0
THRM_EN: This is the thermal enable bit. When this bit is set an active level assertion of the
THRM# signal (as defined by the THRM_POL bit) will set the THRM_STS bit and generate a
power management event (an SCI or SMI). Upon power up, this bit is set to ‘0’.
11.2.7.7
General Purpose 1 Status
Address Offset:
Attributes:
Default Value:
Size:
16-17h
Read/Write
0000h
16 bits
When any bit is set in this register, and the corresponding bit is enabled in the General Purpose 1
Enable register, an SCI and a wake event will be generated.
Bit
Description
15:9
8:0
Reserved.
GPIO_STS: Each bit corresponds to a single GPIO bit (Ex: Bit 8 refers to GPIO[8]. This bit is set
when the data bit in the GPIO data register is set to a ‘1’, and cleared by software writing a ‘1’ to
this bit location. Additionally, if the GPIO_EN bit in the General Purpose 1 Enable Register is
set, then the setting of this bit will generate an SCI or wake event.
11.2.7.8
General Purpose 1 Enable
Address Offset:
Attributes:
Default Value:
Size:
18-19h
Read/Write
0000h
16 bits
Bit
Description
15:9
8:0
Reserved.
GPIO_EN: Each bit corresponds to a single GPIO bit (Ex: Bit 8 refers to GPIO8). When set, the
setting of the corresponding bit in the General Purpose 1 Status Register will generate an SCI or
wake event. These bits are not cleared on a CF9 reset.
11.2.8
SMI Registers
In addition to the ACPI I/O registers, there are some generic registers that are defined for legacy
based SMI# logic. These I/O registers are added to the end of the I/O register space defined by the
ACPI block. All reserved bits will always return 0 when read, and will have no effect when written.
Intel® 460GX Chipset Software Developer’s Manual
11-35
LPC/FWH Interface Configuration
11.2.8.1
Global Control and Enable
Address Offset:
Attributes:
Default Value:
Size:
1A-1Bh
Read/Write
Bits 8 Undefined, Bit 3 ‘1’, All other bits ‘0’
16 bits
Bit
Description
15:13
12
Reserved.
ACPI_TMR_EN: If not using ACPI (SCI) mode, as indicated by SCI_EN not set, then the
ACPI_TMR_EN bit can be set to cause an SMI#. SMI will be generated if this bit, SMI_EN and
TMROF_EN are set, and SCI_EN is not set.
11
10
Reserved.
APMC_EN: Software sets this bit to 1 to enable SMIs based upon accesses to the APM control
port at B2h in I/O space.
9
8
ECC_EN: This bit is used to enable the generation of SMI when the ECC_STS bit in the Global
Status Register is set.
EXTSMI_EN: This bit is used to enable the setting of the EXTSMI_STS bit to generate an SMI
event (SMI# or Wake/Break). The EXTSMI_STS bit is set anytime a GPIO is programmed as an
input, is routed to generate an SMI#, and has its register bit set. Upon power up, this bit is
undefined.
7
SWSMI_TMR_EN: Software sets this bit to a 1 to start the Software SMI# Timer. When the 128
ms timer expires (±4.57 ms), it will generate an SMI# and set the SWSMI_TMR_STS bit. The
SWSMI_TMR_EN bit will remain at 1 until software sets it back to 0. Once the timer has been
started, if SWSMI_TMR_EN bit is set to 0 before the timer expires, the timer will not expire, and
the SMI# will not be generated. The default for this bit is 0.
6
1MIN_EN: Enables the 1 minute timer (±1.17s) to count. When it reaches its timeout, it will
generate an SMI#.
5
4
3
Reserved.
BIOS_EN: Enables the generation of SMI# when ACPI software writes a 1 to the GBL_RLS bit.
EOS: End of SMI. This bit controls the arbitration of the SMI signal to the processor. When this bit
is set, SMI# will be deasserted. Also this bit will be automatically cleared once IFB asserts SMI#
low. In the SMI handler, the processor should clear all pending SMIs (by servicing them and then
clearing their respective status bits) and then setting this bit and resuming. This will allow the SMI
arbiter to re-assert SMI upon detection of an SMI event and the setting of a SMI status bit.
2
THRM_POL: This bit controls the polarity of the THRM# pin needed to set the THRM_STS bit.
When the THRM_POL bit is LOW then a HIGH value on the THRM# signal will set the
THRM_STS bit.
1
0
BIOS_RLS: Enables the generation of an SCI interrupt for ACPI software when a one is written to
this bit position by firmware software. This bit always reads a zero.
SMI_EN: When set, this bit enables the generation of SMIs in the system upon any enabled SMI
event. The exception is when the BIOS Write Enable bit of the BOISEN register is set, an SMI is
generated regardless of the state of the SMI_EN bit. This bit is reset by a PCI reset event.
11-36
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
11.2.8.2
Global Status Register
Address Offset:
Attributes:
Default Value:
Size:
1Ch-1Dh
Read/Write
0000h
16 bits
Bit
Description
15:11
10
Reserved.
ECC_STS: This bit is set when ECCINT# asserted for more than 1 PCI clock with the ECC SERR#
bit in configuration space set, or for one or more clocks if the ECC SERR# bit is not set. The
setting of this bit will cause an SMI if the ECC_EN bit in the Global Control and Enable register
(offset 1A-1Bh) is set. This bit is only set by hardware and can only be cleared by writing a 1 to this
bit position.
9
8
BIOSWEN_STS: This bit is set when the BIOSWEN bit in configuration register 4Eh of Function 0
is set. The setting of this bit will cause an SMI. This bit is only set by hardware and can only be
reset by writing a one to this bit position.
EXTSMI_STS: This bit is set when a GPIO is programmed as an input, is routed to generate an
SMI#, and its register bit is set. Additionally if the EXTSMI_EN bit is set, the setting of the
EXTSMI_STS bit will cause the following, depending on the current state:
If in the S0/C0, S0/C1 or S0/C2 states: An SMI event will be generated.
If in an S1 state, then a Wake event will be generated, and SMI# will also be generated.
If in an S3-S5 state, a Wake event will be generated, but no SMI# will be generated.
This bit is only set by hardware and can only be reset by writing a one.
7
6
SWSMI_TMR_STS: This bit will be set to 1 by the hardware when the Software SMI# Timer
expires. This bit will remain 1 until software writes a 1 to this bit. The default of this bit is 0.
1MIN_STS: This hardware will set this bit to 1 when the 1-minute timer expires. This will also
cause an SMI#. This bit will remain a 1 until the software writes a 1 to this bit. The default of this bit
is 0.
5
4
3
GP_REG_STS: This bit is an OR of the bits in the ACPI General Purpose Register (offset 0Ch).
PM1_REG_STS: This bit is an OR of the Power Management 1 Status Register (offset 00h) bits.
APM_STS: SMI# was generated by a write access to the APM control register. An SMI is
generated when the APMC_EN bit is set. This bit is cleared by writing a 1.
2
1
Reserved.
LEGACY_USB_STS: This bit is an OR of the SMI status bits in the USB Legacy Keyboard
Register ANDed with the enable bits.
0
BIOS_STS: SMI# was generated due to ACPI software requesting attention (writing a 1 to the
GBL_RLS bit with the BIOS_EN bit set). This bit is set by hardware and cleared by software writing
a 1.
11.2.9
General Purpose I/O Registers
For the following GPIO Registers, bits 28:16 refer to the Muxed GPIO signals, and bits 8:0 refer to
the dedicated GPIO signal.
Bit
GPIO
Bit
GPIO
28
27
26
GPIO[22]
GPIO[21]
GPIO[20]
8
7
6
GPIO[8]
GPIO[7]
GPIO[6]
Intel® 460GX Chipset Software Developer’s Manual
11-37
LPC/FWH Interface Configuration
Bit
GPIO
Bit
GPIO
25
24
23
22
21
20
19
18
17
16
GPIO[19]
GPIO[18]
Reserved
Reserved
Reserved
Reserved
GPIO[13]
GPIO[12]
GPIO[11]
GPIO[10]
5
4
3
2
1
0
GPIO[5]
GPIO[4]
GPIO[3]
GPIO[2]
GPIO[1]
GPIO[0]
11.2.9.1
GP Output
Offset:
00-03h
Attribute:
Default Value:
Size:
Read/Write
00000000h
32 bits
Bit
Description
31:29
28:24
Reserved.
Mux Output: When set to a ‘0’, the muxed GPIO pin is programmed as an input. When set to ‘1’,
the muxed GPIO pin is programmed as an output. In the GPO mode, this bit cannot be changed
once the GP Lock bit is set. The setting of this bit only has effect if the muxed GPIO is
programmed to be a GPIO.
23:20
19:16
Reserved.
Mux Output: When set to a ‘0’, the muxed GPIO pin is programmed as an input. When set to ‘1’,
the muxed GPIO pin is programmed as an output. In the GPO mode, this bit cannot be changed
once the GP Lock bit is set. The setting of this bit only has effect if the muxed GPIO is
programmed to be a GPIO.
15:9
8:0
Reserved.
Output: When set to a ‘0’, the GPIO pin is programmed as an input. When set to ‘1’, the GPIO pin
is programmed as an output. This bit cannot be changed once the GP Lock bit is set.
11.2.9.2
GP Data
Offset:
04-07h
Attribute:
Read/Write
Default Value:
Size:
00000000h
32 bits
Bit
Description
31:29
28:24
Reserved.
Muxed Data: If a data bit is programmed to be an output, then this bit can be updated by software
to drive a value on the output pin. If the data bit is programmed as an input, then this bit reflects
the state of the input pin and cannot be updated by software. This bit cannot be changed once the
GP Lock bit is set. The value of this bit only has meaning if the muxed GPIO is enabled as a GPIO.
23:20
Reserved.
11-38
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
Bit
Description
19:16
Muxed Data: If a data bit is programmed to be an output, then this bit can be updated by software
to drive a value on the output pin. If the data bit is programmed as an input, then this bit reflects
the state of the input pin and cannot be updated by software. This bit cannot be changed once the
GP Lock bit is set. The value of this bit only has meaning if the muxed GPIO is enabled as a GPIO.
15:9
8:0
Reserved.
Data: If a data bit is programmed to be an output, then this bit can be updated by software to drive
a value on the output pin. If the data bit is programmed as an input, then this bit reflects the state
of the input pin and cannot be updated by software. This bit cannot be changed once the GP Lock
bit is set.
11.2.9.3
GP TTL
Offset:
08-0Bh
Attribute:
Read/Write
Default Value:
Size:
00000000h
32 bits
Bit
Description
31:29
28:24
Reserved.
Muxed TTL: When set to a ‘1’, and the data bit is programmed as an output, the pin will be
driven with the value of the pin. When set to a ‘0’, and the data bit is programmed as an output,
the pin will be tri-stated when the pin is to be driven to a ‘1’, and driven when the pin is to be
driven to a ‘0’. The setting of this bit has no effect if the pin is programmed as an input. This bit
cannot be changed once the GP Lock bit is set. The value of this bit only has meaning if the
muxed GPIO is enabled as a GPIO.
23:20
19:16
Reserved.
Muxed TTL: When set to a ‘1’, and the data bit is programmed as an output, the pin will be
driven with the value of the pin. When set to a ‘0’, and the data bit is programmed as an output,
the pin will be tri-stated when the pin is to be driven to a ‘1’, and driven when the pin is to be
driven to a ‘0’. The setting of this bit has no effect if the pin is programmed as an input. This bit
cannot be changed once the GP Lock bit is set. The value of this bit only has meaning if the
muxed GPIO is enabled as a GPIO.
15:9
8:0
Reserved.
TTL: When set to a ‘1’, and the data bit is programmed as an output, the pin will be driven with
the value of the pin. When set to a ‘0’, and the data bit is programmed as an output, the pin will
be tri-stated when the pin is to be driven to a ‘1’, and driven when the pin is to be driven to a ‘0’.
The value of the pin is determined by XORing the data bit with the invert bit. The setting of this
bit has no effect if the pin is programmed as an input. This bit cannot be changed once the GP
Lock bit is set.
Intel® 460GX Chipset Software Developer’s Manual
11-39
LPC/FWH Interface Configuration
11.2.9.4
GP Blink
Offset:
0C-0Fh
Attribute:
Default Value:
Size:
Read/Write
00000000h
32 bits
Bit
Description
31:9
8:0
Reserved.
Blink: When set to a ‘1’, and the GP pin is programmed as an output, it will blink at a rate of
once per second. The value of the data bit remains unchanged during the blink process. If it was
set, it remains set. The setting of this bit has no effect if the pin is programmed as an input. This
bit cannot be changed once the GP Lock bit is set.
11.2.9.5
GP Lock
Offset:
10-13h
Attribute:
Default Value:
Size:
Read/Write
00000000h
32 bits
Bit
Description
31:29
28:24
Reserved.
Muxed Lock: When set, and the pin is programmed as an output, the data bit cannot be
changed. Once this bit is set, it can only be cleared by a PCIRST#. Once this bit is set, all other
register bits at this bit location cannot be changed.
23:20
19:16
Reserved.
Muxed Lock: When set, and the pin is programmed as an output, the data bit cannot be
changed. Once this bit is set, it can only be cleared by a PCIRST#. Once this bit is set, all other
register bits at this bit location cannot be changed.
15:9
8:0
Reserved.
Lock: When set, and the pin is programmed as an output, the data bit cannot be changed. Once
this bit is set, it can only be cleared by a PCIRST#. Once this bit is set, all other register bits at
this bit location cannot be changed. This is applicable to GPO mode only.
11.2.9.6
GP Invert
Offset:
14-17h
Attribute:
Default Value:
Size:
Read/Write
00000000h
32 bits
Bit
Description
31:9
8:0
Reserved.
Invert: When set to a ‘0’, the GPIO pin is not inverted. When set to ‘1’, the GPIO pin is inverted.
When set to an output, the data bit is inverted before it is driven on the pin. When set to an input,
the data bit is inverted before entering the data register. This bit cannot be changed once the GP
Lock bit is set.
11-40
Intel® 460GX Chipset Software Developer’s Manual
LPC/FWH Interface Configuration
11.2.9.7
GP SMI
Offset:
1C-1Fh
Attribute:
Default Value:
Size:
Read/Write
00000000h
32 bits
Bit
Description
31:9
8:0
Reserved.
SMI Rout: When set to a ‘1’, and the corresponding data bit is set to an input, a ‘1’ in the data bit
register will be routed to an SMI. If the data bit is set to an output, this value of this bit has no
effect. When cleared, no routing is performed. This bit cannot be changed once the GP Lock bit
is set.
11.2.9.8
GP Pulse
Offset:
20-23h
Attribute:
Read/Write
Default Value:
Size:
00000000h
32 bits
Bit
Description
31:9
8:0
Reserved.
Pulse: When set to a ‘1’, and the data bit (after the invert bit) is programmed as an input, a ‘0’ to
‘1’ transition that is longer than 2 RTC clocks will cause the data bit to be set. A ‘1’ to ‘0’
transition will not clear the bit. Only a write of ‘1’ to the data bit can clear the data bit. If the data
bit is not set to an output, this value of this bit has no effect. When cleared, edge triggering is not
performed. This bit cannot be changed once the GP Lock bit is set.
11.2.9.9
GP Core
Offset:
24-27h
Attribute:
Read/Write
Default Value:
Size:
00000000h
32 bits
Bit
Description
31:9
8:0
Reserved.
Core: When set to a ‘1’, and the data bit is programmed as an output, a ‘0’ will be driven on the
pin when core power loss is detected (PWROK low). This allows the pin, while in the resume
well, to be connected to a device in the core well. When cleared, this check is not performed.
This bit cannot be changed once the GP Lock bit is set.
11.2.9.10 GP Pull-up
Offset:
28-2Bh
Read/Write
03FFh
Attribute:
Default Value:
Size:
32 bits
Bit
Description
31:10
9:0
Reserved.
Pull-up: When set, an internal pull-up will be enabled on the pin. When disabled, the pull-up is
disabled. This bit cannot be changed once the GP Lock bit is set.
Intel® 460GX Chipset Software Developer’s Manual
11-41
LPC/FWH Interface Configuration
11-42
Intel® 460GX Chipset Software Developer’s Manual
IDE Configuration
12
The IFB PCI Function 1 contains an IDE Controller capable of Programmed I/O (PIO) transfers as
well as Bus Master transfer capability. It also supports the “Ultra DMA/33” synchronous DMA
mode of data transfer. The register set associated with IDE Controller is shown below.
12.1
PCI Configuration Registers (Function 1)
Table 12-1. PCI Configuration Registers–Function 1 (IDE Interface)
Configuration
Mnemonic
Register
Register Access
Offset
00–01h
02–03h
04–05h
06–07h
08h
VID
DID
Vendor Identification
Device Identification
PCI Command
RO
RO
R/W
R/W
RO
RO
–
PCICMD
PCISTS
RID
PCI Device Status
Revision Identification
Class Code
09-0Bh
0Ch
CLASSC
–
Reserved
0Dh
MLT
Master Latency Timer
Header Type
R/W
RO
–
0Eh
HEDT
–
0F–1Fh
20–23h
24–3Bh
2C–2Dh
2E–2Fh
30–3Fh
40–43h
44h
Reserved
BMIBA
–
Bus Master Interface Base Address
Reserved
R/W
–
SVID
SID
Subsystem Vendor ID
Subsystem ID
RO
RO
–
–
Reserved
IDETIM
SIDETIM
–
IDE Timing
R/W
R/W
–
Slave IDE Timing
Reserved
45–47h
48h
SDMACTL
–
Synchronous DMA Control
Reserved
R/W
–
49h
4A–4Bh
4C–FFh
SDMATIM
–
Synchronous DMA Timing
Reserved
R/W
–
12.2
IDE Controller Register Descriptions (PCI Function 1)
This section describes in detail the registers associated with the IFB IDE Controller Function. This
includes Programmed I/O (PIO), Bus Master, and “Ultra DMA/33” synchronous DMA
Functionality.
Intel® 460GX Chipset Software Developer’s Manual
12-1
IDE Configuration
12.2.1
12.2.2
12.2.3
VID–Vendor Identification Register (Function 1)
Address Offset:
Default Value:
Attribute:
00–01h
8086h
Read only
The VID Register contains the vendor identification number. This register, along with the Device
Identification Register, uniquely identifies any PCI device. Writes to this register have no effect.
Bit
Description
15:0
Vendor Identification Number. This is a 16-bit value assigned to Intel
DID–Device Identification Register (Function 1)
Address Offset:
Default Value:
Attribute:
02–03h
7601h
Read only
The DID Register contains the device identification number. This register, along with the VID
Register, defines the IFB Function. Writes to this register have no effect.
Bit
Description
15:0
Device Identification Number. This is a 16-bit value assigned to the IFB IDE Controller Function.
PCICMD–PCI Command Register (Function 1)
Address Offset:
Default Value:
Attribute:
04–05h
0000h
Read/Write
The PCICMD Register controls access to the I/O space registers.
Bit
Description
15:10
Reserved.
9
8
7
6
5
4
3
2
1
0
Fast Back to Back Enable (FBE). This bit is hardwired to 0.
SERR# Enable. This bit is hardwired to 0.
Wait Cycle Control. This bit is hardwired to 0.
Parity Error Response. This bit is hardwired to 0.
VGA Palette Snoop. This bit is hardwired to 0.
Postable Memory Write and Invalidate Enable. This bit is hardwired to 0.
Special Cycle Enable. This bit is hardwired to 0.
Bus Master Function Enable (BME). 1=Enable. 0=Disable.
Memory Space Enable. This bit is hardwired to 0.
I/O Space Enable (IOSE). This bit controls access to the I/O space registers. When IOSE=1,
access to the Legacy IDE ports (both primary and secondary) and the PCI Bus Master IDE I/O
Registers is enabled. The Base Address Register for the PCI Bus Master IDE I/O Registers should
be programmed before this bit is set to 1.
12-2
Intel® 460GX Chipset Software Developer’s Manual
IDE Configuration
12.2.4
PCISTS–PCI Device Status Register (Function 1)
Address Offset:
Default Value:
Attribute:
06–07h
0280h
Read/Write
PCISTS is a 16-bit status register for the IDE interface Function. The register also indicates the
IFB's DEVSEL# signal timing.
Bit
Description
15
14
13
Detected Parity Error. Read as 0.
SERR# Status. Read as 0.
Master-Abort Status (MAS)–R/WC. When the Bus Master IDE interface Function, as a master,
generates a master abort, MAS is set to a 1. Software sets MAS to 0 by writing a 1 to this bit.
12
11
Received Target-Abort Status (RTA)–R/WC. When the Bus Master IDE interface Function is a
master on the PCI Bus and receives a target abort, this bit is set to a 1. Software sets RTA to 0
by writing a 1 to this bit.
Signaled Target Abort Status (STA)–R/WC. This bit is set when the IFB IDE interface Function
is targeted with a transaction that the IFB terminates with a target abort. Software resets STA to
0 by writing a 1 to this bit.
10:9
DEVSEL# Timing Status (DEVT)–RO. For the IFB, DEVT=01 indicating medium timing for
DEVSEL# assertion when performing a positive decode. DEVSEL# timing does not include
configuration cycles.
8
7
Data Parity Detected (DPD). Read as 0.
Fast Back to back Capable (FBC)–RO. Hardwired to a 1.
Reserved.
6:0
12.2.5
CLASSC–Class Code Register (Function 1)
Address Offset:
Default Value:
Attribute:
09-0Bh
010180h
Read only
This register identifies the Base Class Code, Sub-Class Code, and Device Programming interface
for IFB PCI Function 1.
Bit
Description
23:16
15:8
7:0
Base Class Code (BASEC). 01h=Mass storage device.
Sub-Class Code (SCC). 01h=IDE controller.
Programming Interface (PI). 80h=Capable of IDE bus master operation.
Intel® 460GX Chipset Software Developer’s Manual
12-3
IDE Configuration
12.2.6
MLT–Master Latency Timer Register (Function 1)
Address Offset:
Default Value:
Attribute:
0Dh
00h
Read/Write
MLT controls the amount of time IFB, as a bus master, can burst data on the PCI Bus. The count
value is an 8 bit quantity. However, MLT[3:0] are reserved and 0 when determining the count
value. The Master Latency Timer is cleared and suspended when IFB is not asserting FRAME#.
When IFB asserts FRAME#, the counter begins counting. If the IFB finishes its transaction before
the count expires, the MLT count is ignored. If the count expires before the transaction completes
(count = # of clocks programmed in MLT), IFB initiates a transaction termination as soon as its
PHLDA# is removed. The number of clocks programmed in the MLT represents the time slice
(measured in PCI clocks) allotted to IFB. The default value of MLT is 00h or 0 PCI clocks.
Bit
Description
7:4
Master Latency Timer Count Value (MLTC). IFB-initiated PCI burst cycles can last indefinitely,
as long as PHLDA# remains active. However, if PHLDA# is negated after the burst cycle is
initiated, IFB limits the burst cycle to the number of PCI Bus clocks specified by this field.
3:0
Reserved.
12.2.7
BMIBA–Bus Master Interface Base Address Register
(Function 1)
Address Offset:
Default Value:
Attribute:
20–23h
00000001h
Read/Write
This register selects the base address of a 16 byte I/O space to provide a software interface to the
Bus Master Functions. Only 12 bytes are actually used (6 bytes for primary and 6 bytes for
secondary).
This register selects the base address of a 16 byte I/O space to provide a software interface to the
Bus Master Functions. Only 12 bytes are actually used (6 bytes for primary and 6 bytes for
secondary).
Bit
Description
31:16
15:4
Reserved.
Bus Master Interface Base Address (BMIBA). These bits provide the base address for the
Bus Master interface registers and correspond to AD[15:4].
3:1
0
Reserved.
Resource Type Indicator (RTE)–RO. This bit is hardwired to 1 indicating that the base address
field in this register maps to I/O space.
12-4
Intel® 460GX Chipset Software Developer’s Manual
IDE Configuration
12.2.8
12.2.9
SVID–Subsystem Vendor ID (Function 1)
Address:
Default Value:
Attribute:
2C-2Dh
0000h
Read only
Bit
Description
15:0
Subsystem Vendor ID.
SID–Subsystem ID (Function 1)
Address:
Default Value:
2E-2Fh
0000h
Attribute:
Read only
Bit
Description
15:0
Subsystem ID.
12.2.10 IDETIM–IDE Timing Register (Function 1)
Address Offset:
Default Value:
Attribute:
40–41h = Primary Channel 42–43h = Secondary Channel
0000h
Read/Write only
This register controls the IFB's IDE interface and selects the timing characteristics of the PCI Local
Bus IDE cycle for PIO and Bus Master transfers. Note that primary and secondary denotations
distinguish between the cables and the 0/1 denotations distinguish between master (0) and slave
(1).
Bit
Description
15
IDE Decode Enable (IDE). 1=Enable. 0=Disable. When enabled, I/O transactions on PCI
targeting the IDE ATA register blocks (command block and control block) are positively decoded
on PCI and driven on the IDE interface. When disabled, these accesses are subtractively decoded
to LPC.
14
Slave IDE Timing Register Enable (SITRE).
0
1
Use bits 13:12, 9:8 for both drive 0 and drive 1.
Use bits 13:12, 9:8 for drive 0, Slave IDE timing register for drive 1.
13:12
IORDY Sample Point (ISP). This field selects the number of PCI clocks between IOR#/IOW#
assertion and the first IORDY sample point.
Bits[13:12] Number of Clocks
00 5
01 4
10 3
11 2
11:10
Reserved.
Intel® 460GX Chipset Software Developer’s Manual
12-5
IDE Configuration
Bit
Description
9:8
Recovery Time (RTC). This field selects the minimum number of PCI clocks between the last
IORDY# sample point and the DIOx# strobe of the next cycle.
Bits[9:8] Number of Clocks
00 4
01 3
10 2
11 1
7
DMA Timing Enable Only (DTE1). When DTE1=1, fast timing mode is enabled for DMA data
transfers for drive 1. PIO transfers to the IDE data port will run in compatible timing. When DTE1 =
0, both DMA and PIO data transfers for drive 1 will use the fast timing mode.
6
5
Prefetch and Posting Enable (PPE1). When PPE1=1, prefetch and posting to the IDE data port
is enabled for drive 1. When PPE1 = 0, prefetch and posting is disabled for drive 1.
IORDY Sample Point Enable Drive Select 1 (IE1). When IE1=0, IORDY sampling is disabled for
Drive 1. When the internal IORDY signal is forced asserted, IORDY is sampled asserted at the first
sample point as specified by the ISP field in this register.
When IE1=1 and the currently selected drive (via a copy of bit 4 of 1x6h) is Drive 1, all accesses to
the enabled I/O address range sample IORDY. The IORDY sample point is specified by the ISP
field in this register.
4
3
Fast Timing Bank Drive Select 1 (TIME1). When cleared, accesses to the data port will use
compatible timings for this drive. When set and bit 14 cleared, accesses to the data port will use
bits 13:12 for the IORDY sample point, and bits 9:8 for the recovery time. When set and bit 14 set,
accesses to the data port will use the IORDY sample point and recover time specified in the slave
IDE timing register.
DMA Timing Enable Only (DTE0). When DTE0=1, fast timing mode is enabled for DMA data
transfers for drive 0. PIO transfers to the IDE data port will run in compatible timing. When DTE1 =
0, both DMA and PIO data transfers for drive 0 will use the fast timing mode.
2
1
Prefetch and Posting Enable (PPE0). When PPE0 = 1, prefetch and posting to the IDE data port
is enabled for drive 0. When PPE0 = 0, prefetch and posting is disabled for drive 0.
IORDY Sample Point Enable Drive Select 0 (IE0). When IE0=0, IORDY sampling is disabled for
Drive 0. When the internal IORDY signal is forced asserted, IORDY is sampled asserted at the first
sample point as specified by the ISP field in this register.
When IE0=1 and the currently selected drive (via a copy of bit 4 of 1x6h) is Drive 0, all accesses to
the enabled I/O address range sample IORDY. The IORDY sample point is specified by the ISP
field in this register.
0
Fast Timing Bank Drive Select 0 (TIME0). When TIME0=0, accesses to the data port of the
enabled I/O address range uses the 16 bit compatible timing.
When TIME0=1 and the currently selected drive (via a copy of bit 4 of 1x6h) is Drive 0, accesses to
the data port of the enabled I/O address range use fast timings. PIO accesses to the data port use
fast timing only if bit 3 of this register (DTE0) is 0. Accesses to all non-data ports of the enabled I/
O address range always use the 8 bit compatible timings.
12.2.11 SIDETIM–Slave IDE Timing Register (Function 1)
Address Offset:
Default Value:
Attribute:
44h
00h
Read/Write only
This register controls the IFB’s IDE interface and selects the timing characteristics for the slave
drives on each IDE channel. This allows for programming of independent operating modes for
each IDE agent. This register has no affect unless the SITRE bit is enabled in the IDETIM Register.
12-6
Intel® 460GX Chipset Software Developer’s Manual
IDE Configuration
Bit
Description
7:6
Secondary Drive 1 IORDY Sample Point (SISP1). This field selects the number of PCI clocks
between SDIOx# assertion and the first SIORDY sample point for the slave drive on the secondary
channel.
Bits[7:6] Number of Clocks
00 5
01 4
10 3
11 2
5:4
3:2
1:0
Secondary Drive 1 Recovery Time (SRTC1). This field selects the minimum number of PCI
clocks between the last SIORDY# sample point and the SDIOx# strobe of the next cycle for the
slave drive on the secondary channel.
Bits[5:4] Number of Clocks
00 4
01 3
10 2
11 1
Primary Drive 1 IORDY Sample Point (PISP1). This field selects the number of PCI clocks
between PDIOx# assertion and the first PIORDY sample point for the slave drive on the primary
channel.
Bits[3:2] Number of Clocks
00 5
01 4
10 3
11 2
Primary Drive 1 Recovery Time (PRTC1). This field selects the minimum number of PCI clocks
between the last PIORDY# sample point and the PDIOx# strobe of the next cycle for the slave
drive on the primary channel.
Bits[1:0] Number of Clocks
00 4
01 3
10 2
11 1
12.2.12 DMACTL–Synchronous DMA Control Register (Function 1)
Address Offset:
Default Value:
Attribute:
48h
00h
Read/Write
This register enables each individual channel and drive for Synchronous DMA transfers. For non-
synchronous DMA operation, this register should be left programmed to its default value.
Bit
Description
7:4
3
Reserved.
Secondary Drive 1 SDMA Enable (SSDE1). 1 = Enable Synchronous DMA mode for
secondary channel drive 1. 0 = Disable (default).
2
1
0
Secondary Drive 0 SDMA Enable (SSDE0). 1 = Enable Synchronous DMA mode for
secondary channel drive 0. 0 = Disable (default).
Primary Drive 1 SDMA Enable (PSDE1). 1 = Enable Synchronous DMA mode for primary
channel drive 1. 0 = Disable (default).
Primary Drive 0 SDMA Enable (PSDE0). 1 = Enable Synchronous DMA mode for primary
channel drive 0. 0 = Disable (default).
Intel® 460GX Chipset Software Developer’s Manual
12-7
IDE Configuration
12.2.13 SDMATIM–Synchronous DMA Timing Register (Function 1)
Address Offset:
Default Value:
Attribute:
4A-4Bh
0000h
Read/Write only
This register controls the timings used by each Synchronous DMA enabled device. For non-
synchronous DMA operation, this register should be left programmed to its default value.
Bit
Description
15:14
13:12
Reserved.
Secondary Drive 1 Cycle Time (SCT1): These bit settings determine the minimum data write
strobe Cycle Time (CT) and minimum Ready to Pause time (RP).
00: CT = 4 PCICLK, RP = 6 PCICLK
01: CT = 3 PCICLK, RP = 5 PCICLK
10: CT = 2 PCICLK, RP = 4 PCICLK
11: Reserved
11:10
9:8
Reserved.
Secondary Drive 0 Cycle Time (SCT0): These bit settings determine the minimum data write
strobe Cycle Time (CT) and minimum Ready to Pause time (RP).
00: CT = 4 PCICLK, RP = 6 PCICLK
01: CT = 3 PCICLK, RP = 5 PCICLK
10: CT = 2 PCICLK, RP = 4 PCICLK
11: Reserved
7:6
5:4
Reserved.
Primary Drive 1 Cycle Time (PCT1): These bit settings determine the minimum data write
strobe Cycle Time (CT) and minimum Ready to Pause time (RP).
00: CT = 4 PCICLK, RP = 6 PCICLK
01: CT = 3 PCICLK, RP = 5 PCICLK
10: CT = 2 PCICLK, RP = 4 PCICLK
11: Reserved
3:2
1:0
Reserved.
Primary Drive 0 Cycle Time (PCT0): These bit settings determine the minimum data write
strobe Cycle Time (CT) and minimum Ready to Pause time (RP).
00: CT = 4 PCICLK, RP = 6 PCICLK
01: CT = 3 PCICLK, RP = 5 PCICLK
10: CT = 2 PCICLK, RP = 4 PCICLK
11: Reserved
12-8
Intel® 460GX Chipset Software Developer’s Manual
IDE Configuration
Table 12-2. Ultra DMA/33 Timing Mode Settings
Ultra DMA/33 Timing Modes
Cycle Time Bit Settings
Mode 0 (120 ns)
00
Mode 1 (90 ns)
01
Mode 2 (60 ns)
10
Table 12-3. DMA/PIO Timing Values Based on IFB Cable Mode and System Speed
IDETIM [15:8]
IDETIM[15:8]
Drive 0
(Master)
If Slave
SIDETIM
Pri [3:0]
Sec [7:4]
Drive 1
Resultant
Cycle Time
IORDY
Sample
Point
Drive 0 (Master)
If no Slave
Attached or
Slave is
Recovery
Time
(RCT)
IFB Drive
Mode
Base Operating
Frequency and
(ISP)
2
Attached
(Slave)
Cycle Time
1
Mode 0
PIO0/
6 clocks
1 clocks
C0h
D0h
E1h
E3h
80h
90h
A1h
A3h
0
4
9
B
30 MHz: 660ns
33 MHz: 600ns
Compatible (Default) (Default)
PIO2/SW2
PIO3/MW1
PIO4/MW2
4 clocks
3 clocks
3 clocks
4 clocks
3 clocks
1 clock
30 MHz: 256ns
33 MHz: 240ns
30 MHz: 198ns
33 MHz: 180ns
30 MHz: 132ns
33 MHz: 120ns
NOTES:
1. Table 12-3 assumes that if the attached slave drive is Mode 0 or not present, the SITRE bit is ‘0’.
2. Table 12-3 assumes that 25 MHz is not supported as a target PCI system speed. If the DMA Timing Enable
Only (DTE) bit has been enabled for that drive, this resultant cycle time applies to data transfers performed
with DMA only.
12.3
IDE Controller I/O Space Registers
The PCI IDE Function uses 16 bytes of I/O space, allocated via the BMIBA register. All bus master
IDE I/O space registers can be accessed as byte, word, or DWord quantities. The description of the
16 bytes of I/O registers follows.
12.3.1
BMICx–Bus Master IDE Command Register (I/O)
Address Offset:
Default Value:
Attribute:
Primary Channel–Base + 00h; Secondary Channel–Base + 08h
00h
Read/Write
This register enables/disables bus master capability for the IDE Function and provides direction
control for the IDE DMA transfers. This register also provides bits that software uses to indicate
DMA capability of the IDE device.
Intel® 460GX Chipset Software Developer’s Manual
12-9
IDE Configuration
Bit
Description
7:4
3
Reserved.
Bus Master Read/Write Control (RWCON). 0=Reads; 1=Writes. This bit must NOT be changed
when the bus master Function is active.
While a synchronous DMA transfer is in progress, this bit will be READ ONLY. The bit will return to
read/write once the synchronous DMA transfer has been completed or halted.
2:1
0
Reserved.
Start/Stop Bus Master (SSBM). 1=Start; 0=Stop. When this bit is set to 1, bus master operation
starts. The controller transfers data between the IDE device and memory only while this bit is set.
Master operation can be stopped by writing a 0 to this bit. This results in all state information being
lost (i.e. master mode operation cannot be stopped and then resumed).
If this bit is set to 0 while bus master operation is still active (i.e. Bit 0=1 in the Bus Master IDE Status
Register for that IDE channel) and the drive has not yet finished its data transfer (bit 2=0 in the
channel’s Bus Master IDE Status Register), the bus master command is aborted and data
transferred from the drive may be discarded by the IFB rather than being written to system memory.
This bit is intended to be set to 0 after the data transfer is completed, as indicated by either bit 0 or
bit 2 being set in the IDE Channel’s Bus Master IDE Status Register.
12.3.2
BMISx–Bus Master IDE Status Register (I/O)
Address Offset:
Default Value:
Attribute:
Primary Channel–Base + 02h; Secondary Channel–Base + 0Ah
00h
Read/Write Clear
This register provides status information about the IDE device and state of the IDE DMA transfer.
Bit
Description
7
6
Reserved. This bit is hardwired to 0.
Drive 1 DMA Capable (DMA1CAP)–R/W. 1=Drive 1 is capable of DMA transfers. This bit is a
software controlled status bit that indicates IDE DMA device capability and does not affect
hardware operation.
5
Drive 0 DMA Capable (DMA0CAP)–R/W. 1=Drive 0 is capable of DMA transfers. This bit is a
software controlled status bit that indicates IDE DMA device capability and does not affect
hardware operation.
4:3
2
Reserved.
IDE Interrupt Status (IDEINTS)–R/WC. This bit, when set to a 1, indicates when an IDE device
has asserted its interrupt line. When bit 2=1, all read data from the IDE device has been
transferred to main memory and all write data has been transferred to the IDE device. Software
sets this bit to a 0 by writing a 1 to it. IRQ14 is used for the primary channel and IRQ15 is used
for the secondary channel. Note that, if the interrupt status bit is set to a 0 by writing a 1 to this
bit while the interrupt line is still at the active level, this bit remains 0 until another assertion edge
is detected on the interrupt line.
1
0
IDE DMA Error–R/WC. This bit is set to 1 when the IFB encounters a target abort or master
abort while transferring data on the PCI Bus. Software sets this bit to a 0 by writing a 1 to it.
Bus Master IDE Active (BMIDEA)–RO. The IFB sets this bit to 1 when bit 0 in the BMICx
Register is set to 1. The IFB sets this bit to 0 when the last transfer for a region is performed
(where EOT for that region is set in the region descriptor). The IFB also sets this bit to 0 when
bit 0 of the BMICx Register is set to 0. When this bit is read as a zero, all data transferred from
the drive during the previous bus master command is visible in system memory, unless the bus
master command was aborted.
12-10
Intel® 460GX Chipset Software Developer’s Manual
IDE Configuration
Table 12-4. Interrupt/Activity Status Combinations
Bit 2
Bit 0
Description
0
1
1
0
DMA transfer is in progress. No interrupt has been generated by the IDE device.
The IDE device generated an interrupt and the Physical Region Descriptors
exhausted. This is normal completion where the size of the physical memory regions
is equal to the IDE device transfer size.
1
0
1
0
The IDE device generated an interrupt. The controller has not reached the end of
the physical memory regions. This is a valid completion case when the size of the
physical memory regions is larger than the IDE device transfer size.
Error condition. If the IDE DMA Error bit is 1, there is a problem transferring data to/
from memory. Specifics of the error have to be determined using bus-specific
information. If the Error bit is 0, the PRD specified a smaller buffer size than the
programmed IDE transfer size.
12.3.3
BMIDTPx–Bus Master IDE Descriptor Table Pointer Register
(I/O)
Address Offset:
Default Value:
Attribute:
Primary Channel–Base + 04h; Secondary Channel–Base + 0Ch
00000000h
Read/Write
This register provides the base memory address of the Descriptor Table. The Descriptor Table must
be DWord aligned and not cross a 4 KByte boundary in memory.
Bit
Description
31:2
1:0
Descriptor Table Base Address (DTBA). Bits [31:2] correspond to A[31:2].
Reserved.
Intel® 460GX Chipset Software Developer’s Manual
12-11
IDE Configuration
12-12
Intel® 460GX Chipset Software Developer’s Manual
Universal Serial Bus (USB)
Configuration
13
The IFB integrates one USB Controller. The USB Controller is UHCI 1.1 compliant and
implements the root hub of the USB, which contains two ports.
The IFB PCI Function 2 reflects the USB Host and Root Hubs, with 2 connected USB ports. The
register set associated with USB Host Controller is shown below with actual register descriptions
13.1
PCI Configuration Registers (Function 2)
Table 13-1. PCI Configuration Registers–Function 2
Configuration
Mnemonic
Offset
Register
Access
Register
00–01h
02–03h
04–05h
06–07h
08h
VID
DID
Vendor Identification
Device Identification
PCI Command
PCI Device Status
Revision Identification
Class Code
RO
RO
R/W
R/W
RO
RO
–
PCICMD
PCISTS
RID
09-0Bh
0Ch
CLASSC
–
Reserved
0Dh
MLT
Latency Timer
Header Type
R/W
RO
–
0Eh
HEDT
–
0F–1Fh
20–23h
24–3Bh
2C–2Dh
2E–2Fh
30–3Fh
3Ch
Reserved
USBBA
–
USB I/O Space Base Address
Reserved
R/W
–
SVID
SID
Subsystem Vendor ID
Subsystem ID
RO
RO
–
–
Reserved
INTLN
INTPN
–
Interrupt Line
R/W
RO
–
3Dh
Interrupt Pin
3E–5Fh
60h
Reserved
SBRNUM
–
Serial Bus Release Number
Reserved
RO
–
61–69h
6A–6Bh
6C–BFh
C0–C1h
C2-C3h
C4h
MCR
–
Miscellaneous Control Register
Reserved
R/W
–
LEGSUP
---
Legacy Support
Reserved
R/W
---
USBREN
---
USB Resume Enable
Reserved
R/W
---
C5-FF
Intel® 460GX Chipset Software Developer’s Manual
13-1
Universal Serial Bus (USB) Configuration
13.2
USB Host Controller Register Descriptions (PCI
Function 2)
This section describes in detail the registers associated with the IFB USB Host Controller
Functions. This includes UHCI compatible registers and Legacy Keyboard registers.
13.2.1
VID–Vendor Identification Register (Function 2)
Address Offset:
Default Value:
Attribute:
00–01h
8086h
Read only
The VID Register contains the vendor identification number. This register, along with the Device
Identification Register, uniquely identifies any PCI device. Writes to this register have no effect.
Bit
Description
15:0
Vendor Identification Number. This is a 16-bit value assigned to Intel.
13.2.2
DID–Device Identification Register (Function 2)
Address Offset:
Default Value:
Attribute:
02-03h
7602h
Read only
The DID Register contains the device identification number. This register, along with the VID
Register, defines the IFB USB Host Controller. Writes to this register have no effect.
Bit
Description
15:0
Device Identification Number. This is a 16-bit value assigned to the IFB USB Host Controller.
13.2.3
PCICMD–PCI Command Register (Function 2)
Address Offset:
Default Value:
Attribute:
04-05h
00h
Read/Write
This register controls access to the I/O space registers.
Bit
Description
15:10
Reserved. Read 0.
9
8:5
4
Fast Back to Back Enable (Not Implemented). This bit is hardwired to 0.
Reserved. Read as 0.
Memory Write and Invalidate Enable (Not Implemented). This bit is hardwired to 0.
Special Cycle Enable (Not Implemented). This bit is hardwired to 0.
3
2
Bus Master Enable (BME). This bit controls the IFB’s ability to act as a master on the PCI bus for
the host controller transfers. A value of 0 disables the device from generating PCI accesses. A
value of 1 allows the device to behave as a USB host controller bus master. This bit must be set to
1 before USB transactions can start.
13-2
Intel® 460GX Chipset Software Developer’s Manual
Universal Serial Bus (USB) Configuration
Bit
Description
1
0
Memory Space Enable (Not Implemented). This bit is hardwired to 0.
I/O Space Enable (IOSE). 1=Enable. 0=Disable. This bit controls the access to the I/O space
registers. If this bit is set, access to the host controller I/O registers is enabled. The base register for
the I/O registers must be programmed before this bit is set.
13.2.4
PCISTS–PCI Device Status Register (Function 2)
Address Offset:
Default Value:
Attribute:
06-07h
0280h
Read/Write
DSR is a 16-bit status register that reports the occurrence of a PCI master-abort by the USB HC
module or a PCI target-abort when the Serial Bus module is a master. The register also indicates the
USB HC module DEVSEL# signal timing that is hardwired in the USB HC module.
Bit
Description
Detected Parity (Not Implemented). Read as 0.
15
14
13
SERR# Status (Not Implemented). Read as 0.
Master-Abort Status (MAS)–R/WC. When the Serial Bus module receives a master-abort from a
PCI transaction, MAS is set to a 1. Software sets MAS to 0 by writing a 1 to this bit.
12
11
Received Target-Abort Status (RTA)–R/WC. When the Serial Bus module is a master on the PCI
Bus and receives a target-abort, this bit is set to a 1. Software resets RTA to 0 by writing a 1 to this
bit.
Signaled Target-Abort Status (STA)–R/WC. This bit is set when the Serial Bus module Function is
targeted with a transaction that the Serial Bus module terminates with a target abort. Software
resets STA to 0 by writing a 1 to this bit.
10:9
DEVSEL# Timing Status (DEVT)–RO. This 2-bit field defines the timing for DEVSEL# assertion.
These read only bits indicate the IFB’s DEVSEL# timing when performing a positive decode. Since
the IFB always generate the DEVSEL# with medium timing, DEVT=01. This DEVSEL# timing does
not include Configuration cycles.
8
7
Data Parity Detected (Not Implemented). Read as 0.
Fast Back to Back Capable (FBC)–RO. Hardwired to 1. This bit indicates to the PCI Master that
Serial Bus module as a target is capable of accepting fast back-to-back transactions.
6:0
Reserved. Read as 0’s.
13.2.5
RID–Revision Identification Register (Function 2)
Address Offset:
Default Value:
08h
Initial Stepping = 00h. Refer to IFB Specification Updates
for other values programmed here.
Read only
Attribute:
This 8 bit register contains device stepping information. Writes to this register have no effect.
Bit
Description
7:0
Revision ID Byte. The register is hardwired to the default value.
Intel® 460GX Chipset Software Developer’s Manual
13-3
Universal Serial Bus (USB) Configuration
13.2.6
CLASSC–Class Code Register (Function 2)
Address Offset:
Default Value:
Attribute:
0A-0Bh
0C03h
Read only
This register identifies the Base Class Code, Sub-Class Code, and Device Programming interface
for IFB PCI Function 2.
Bit
Description
23:16
15:8
7:0
Base Class Code (BASEC). 0Ch=Serial Bus controller.
Sub-Class Code (SCC). 03h=USB Host Controller..
Programming Interface (PI). 00h=Universal Host Controller Interface.
13.2.7
MLT–Master Latency Timer Register (Function 2)
Address Offset:
Default Value:
Attribute:
0Dh
00h
Read/Write
MLT is an 8-bit register that controls the amount of time (in terms of PCI clocks) the USB module
can do transactions on the PCI bus. The count value is an 8 bit quantity, however MLT[3:0] are
reserved and assumed to be 0 when determining the count value. MLT is used when the USB
module becomes the PCI bus master and is cleared and suspended when IFB is not asserting
FRAME#. When IFB asserts FRAME#, the counter is enabled and begins counting. If the serial
bus module finishes its transaction before count is expired, the MLT value is ignored. If the count
expires before the transaction completes, IFB initiates a transaction termination as soon as the
current transaction is completed.. The number of clocks programmed in the MLT represents the
time slice (measured in PCI clocks) allotted to IFB, after which it must surrender the bus as soon as
the current transaction is completed.
Bit
Description
7:4
Master Latency Counter Value. IFB initiated PCI cycles (including multiple transactions) can last
indefinitely as long as PHLDA# remains active. However, if PHLDA# is negated after a transaction
is initiated, IFB limits the duration of the transactions to the number of PCI bus clocks specified by
this field.
3:0
Reserved.
13.2.8
HEDT–Header Type Register (Function 2)
Address Offset:
Default Value:
Attribute:
0Eh
00h
Read only
This register identifies the Serial Bus module as a single Function device.
Bit
Description
7:0
Device Type (DEVICET). 00. Multi-Function device capability for IFB is defined by the HEDT
register in Function 0.
13-4
Intel® 460GX Chipset Software Developer’s Manual
Universal Serial Bus (USB) Configuration
13.2.9
USBBA–USB I/O Space Base Address (Function 2)
Address Offset:
Default Value:
Attribute:
20-23h
00000001h
Read/Write
This register contains the base address of the USB I/O Registers.
Bit
Description
Reserved. Hardwired to 0s. Must be written as 0S.
31:16
15:5
Index Register Base Address. Bits [15:5] correspond to I/O address signals AD [15:5],
respectively.
4:1
0
Reserved. Read as 0.
Resource Type Indicator (RTE)–RO. This bit is hardwired to 1 indicating that the base address
field in this register maps to I/O space.
13.2.10 SVID–Subsystem Vendor ID (Function 2)
Address:
Default Value:
2C-2Dh
0000h
Attribute:
Read only
Bit
Description
15:0
Subsystem Vendor ID.
13.2.11 SID–Subsystem ID (Function 2)
Address:
Default Value:
Attribute:
2E-2Fh
0000h
Read only
Bit
Description
15:0
Subsystem ID.
13.2.12 INTLN–Interrupt Line Register (Function 2)
Address Offset:
Default Value:
Attribute:
3Ch
00h
Read/Write
Software programs this register with interrupt information concerning the USB.
Bit
Description
7:0
Interrupt Line. The value in this register has no affect on IFB hardware operations.
Intel® 460GX Chipset Software Developer’s Manual
13-5
Universal Serial Bus (USB) Configuration
13.2.13 INTPN–Interrupt Pin (Function 2)
Address Offset:
Default Value:
Attribute:
3Dh
04h
Read only
This register indicates which PCI interrupt pin is used for the USB module interrupt. The USB
interrupt is internally ORed to the interrupt controller with the PIRQD# signal.
Bit
Description
7:3
2:0
Reserved.
Serial Bus Module Interrupt Routing. The value of 04h in Function 2 indicates that the IFB will
drive PIRQD# as its interrupt line for the USB controller.
13.2.14 Miscellaneous Control (Function 2)
Address Offset:
Default Value:
Attribute:
6A-6Bh
0000h
Read/Write
Bit
Description
15:2
1
Reserved.
Low Speed PreSOF Disable. This bit should be set to ‘1’ to disable.
0
Reserved.
13.2.15 SBRNUM–Serial Bus Release Number (Function 2)
Address Offset:
Default Value:
Attribute:
60h
10h
Read only
This register contains the release of the USB Specification with which this USB Host Controller
module is compliant.
Bit
Description
7:0
Serial Bus Specification Release Number. All other combinations are reserved.
Bits[7:0] Release Number
00h Pre-release 1.0
10h Release 1.0
13.2.16 LEGSUP–Legacy Support Register (Function 2)
PCI Address Offset:
Default:
C0-C1h
2000h
Attribute:
Read/Write Clear
This register provides control and status capability for the legacy keyboard and mouse Functions.
13-6
Intel® 460GX Chipset Software Developer’s Manual
Universal Serial Bus (USB) Configuration
Bit
Description
15
End OF A20GATE Pass Through Status (A20PTS)–R/WC. This bit is set to 1 to indicate that the
A20GATE pass-through sequence has ended. This bit will only be set if bit 7 of this register is also
set. Software must use the enable bits to determine the exact cause of an SMI#. Software clears
this bit by writing a 1 to it.
14
13
Reserved.
USB PIRQ Enable (USBPIRQDEN)–R/W. 1 (default) = USB interrupt is routed to PIRQD. 0 = USB
interrupt does not route to PIRQD. This bit prevents the USB controller from generating an
interrupt. Note that it will probably be configured to generate an SMI using bit 4 of this register.
Default to 1 for compatibility with older USB software.
12
11
10
9
USB IRQ Status (USBIRQS)–RO. This bit is set to 1 to indicate that the USB IRQ is active.
Software must use the enable bits to determine the exact cause of an SMI#. Writing a 1 to this bit
will have no effect. Software must clear the IRQ via the USB controller.
Trap By 64h Write Status (TBY64W)–R/WC. This bit is set to 1 to indicate that a write to port 64h
occurred. Software must use the enable bits to determine the exact cause of an SMI#. Software
clears this bit by writing a 1 to it.
Trap By 64h Read Status (TBY64R)–R/WC. This bit is set to 1 to indicate that a read to port 64h
occurred. Software must use the enable bits to determine the exact cause of an SMI#. Software
clears this bit by writing a 1 to it.
Trap By 60h Write Status (TBY60W)–R/WC. This bit is set to 1 to indicate that a write to port 60h
occurred. Software must use the enable bits to determine the exact cause of an SMI#. Software
clears this bit by writing a 1 to it.
8
Trap By 60h Read Status (TBY60R)–R/WC. This bit is set to 1 to indicate that a read to port 60h
occurred. Software must use the enable bits to determine the exact cause of an SMI#. Software
clears this bit by writing a 1 to it.
7
SMI At End Of Pass Through Enable (SMIEPTE)–R/W. 1=Enable the generation of an SMI when
the A20GATE pass-through sequence has ended. 0 (default) = Disable. This may be required if an
SMI is generated by a USB interrupt in the middle of an A20GATE pass through sequence and
needs to be serviced later.
6
5
Pass Through Status (PSS)–RO. 1 =A20GATE pass through sequence is currently in progress. 0
(default) = Not currently executing the A20GATE pass-through sequence. This bit indicates that
the host controller is executing the A20GATE pass-through sequence. If software needs to reset
this bit, it should set Bit 5 to 0 causing the host controller to immediately end the A20GATE pass
through sequence.
A20Gate Pass Through Enable (A20PTEN)–R/W. 1=Enable A20GATE pass through sequence.
0 (default) = Disable. When enabled, the logic will pass through the following A20GATE command
sequence:
Cycle Address Data
Write 64h D1h ( 1 or more) (Starts the Sequence)
Write 60h xxh
Read 64h N/A ( 0 or more)
Write 64h FFh (End of A20GATE Pass Through Sequence)
Any deviation seen in the above sequence will cause the host controller to immediately exit the
sequence and return to typical operation, performing an I/O trap and generating an SMI# if
appropriate enable bits are set.
When enabled, SMI# will not be generated during the sequence, even if the various enable bits are
set. Note that during a Pass-through sequence, the above status bits will not be set for the I/O
accesses that are part of the sequence.
4
3
2
1
0
Trap/SMI ON IRQ Enable (USBSMIEN)–R/W. 1 = Enable SMI# generation on USB IRQ. 0
(default) = Disable.
Trap/SMI On 64h Write Enable (64WEN)–R/W. 1 = Enable I/O Trap and SMI# generation on port
64h write. 0 (default) = Disable.
Trap/SMI On 64h Read Enable (64REN)–R/W. 1 = Enable I/O Trap and SMI# generation on port
64h read. 0 (default) = Disable.
Trap/SMI On 60h Write Enable (60WEN)–R/W. 1 = Enable I/O Trap and SMI# generation on port
60h write. 0 (default) = Disable.
Trap/SMI On 60h Read Enable (60REN)–R/W. 1 = Enable I/O Trap and SMI# generation on port
60h read. 0 (default) = Disable.
Intel® 460GX Chipset Software Developer’s Manual
13-7
Universal Serial Bus (USB) Configuration
13.2.17 USBREN–USB Resume Enable
Address Offset:
Default Value:
Attribute:
C4h
00h
Read/Write
Bit
Description
7:2
1
Reserved.
PORT1EN: Enable port 1 of the USB controller to look at wakeup events. When set, the USB
controller will monitor port 1 for a connect/disconnect, which is a resume event for USB. When
cleared, the USB controller will not look at this port for a wakeup event. This bit applies to port 1.
0
PORT0EN: Enable port 0 of the USB controller to look at wakeup events. When set, the USB
controller will monitor port 0 for a connect/disconnect, which is a resume event for USB. When
cleared, the USB controller will not look at this port for a wakeup event. This bit applies to port 0.
13.3
USB Host Controller I/O Space Registers
13.3.1
USBCMD–USB Command Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (00-01h)
0000h
Read/Write (WORD write-able only)
The Command Register indicates the command to be executed by the serial bus host controller.
Writing to the register causes a command to be executed. The table following the bit description
provides additional information on the operation of the Run/Stop and Debug bits.
Bit
Description
15:9
8
Reserved.
Loop Back Test Mode: When set, the IFB is in loop back test mode. When both ports are connected
together, a write to one port will be seen on the other port and the data will be store in I/O offset 18h.
When clear, this will not occur.
7
Max Packet (MAXP). 1=64 bytes. 0=32 bytes. This bit selects the maximum packet size that can be
used for full speed bandwidth reclamation at the end of a frame. This value is used by the Host
Controller to determine whether it should initiate another transaction based on the time remaining in
the SOF counter. Use of reclamation packets larger than the programmed size will cause a Babble
error if executed during the critical window at frame end. The Babble error results in the offending
endpoint being stalled. Software is responsible for ensuring that any packet which could be executed
under bandwidth reclamation be within this size limit.
6
5
Configure Flag (CF). HCD software sets this bit as the last action in its process of configuring the
Host Controller. This bit has no effect on the hardware. It is provided only as a semaphore service for
software.
Software Debug (SWDBG). 1=Debug mode. 0=Normal Mode. In SW Debug mode, the Host
Controller clears the Run/Stop bit after the completion of each USB transaction. The next transaction
is executed when software sets the Run/Stop bit back to 1. The SWDBG bit must only be
manipulated when the controller is in the stopped state. This can be determined by checking the
HCHalted bit in the USBSTS register.
4
Force Global Resume (FGR). 1=Host Controller sends the Global Resume signal on the USB.
Software sets this bit to 0 after 20 ms have elapsed to stop sending the Global Resume signal. At
that time all USB devices should be ready for bus activity. The Host Controller sets this bit to 1 when
a resume event (connect, disconnect, or K-state) is detected while in global suspend mode. Software
resets this bit to 0 to end Global Resume signaling. The 1 to 0 transition causes the port to send a
low speed EOP signal. This bit will remain a 1 until the EOP has completed.
13-8
Intel® 460GX Chipset Software Developer’s Manual
Universal Serial Bus (USB) Configuration
Bit
Description
3
Enter Global Suspend Mode (EGSM). 1=Host Controller enters the Global Suspend mode. No USB
transactions occur during this time. The Host Controller is able to receive resume signals from USB
and interrupt the system. Software resets this bit to 0 to come out of Global Suspend mode. Software
writes this bit to 0 at the same time that Force Global Resume (bit 4) is written to 0 or after writing bit
4 to 0. Software must also ensure that the Run/Stop bit (bit 0) is cleared prior to setting this bit.
2
1
Global Reset (GRESET). When this bit is set, the Host Controller sends the global reset signal on
the USB and then resets all its logic, including the internal hub registers. The hub registers are reset
to their power on state. This bit is reset by the software after a minimum of 10 ms has elapsed as
specified in Chapter 7 of the USB Specification.
Note: Chip Hardware Reset has the same effect as Global Reset (bit 2), except that the Host
Controller does not send the Global Reset on USB.
Host Controller Reset (HCRESET). When this bit is set, the Host Controller module resets its
internal timers, counters, state machines, etc. to their initial value. Any transaction currently in
progress on USB is immediately terminated. This bit is reset by the Host Controller when the reset
process is complete.
The HCReset effects on Hub registers are slightly different from Chip Hardware Reset and Global
USB Reset. The HCReset affects bits [8,3:0] of the Port Status and Control Register (PORTSC) of
each port. HCReset resets the state machines of the Host Controller including the Connect/
Disconnect state machine (one for each port). When the Connect/Disconnect state machine is reset,
the output that signals connect/disconnect are negated to 0, effectively signaling a disconnect, even
if a device is attached to the port. This virtual disconnect causes the port to be disabled. This
disconnect and disabling of the port causes bit 1 (connect status change) and bit 3 (port enable/
disable change) of the PORTSC to get set. The disconnect also causes bit 8 of PORTSC to reset.
About 64 bit times after HCReset goes to 0, the connect and low-speed detect will take place and bits
0 and 8 of the PORTSC will change accordingly.
0
Run/Stop (RS). 1=Run. 0=Stop. When set to a 1, the Host Controller proceeds with execution of the
schedule. The Host Controller continues execution as long as this bit is set. When this bit is set to 0,
the Host Controller completes the current transaction on the USB and then halts. The HC Halted bit
in the status register indicates when the Host Controller has finished the transaction and has entered
the stopped state. The Host Controller clears this bit when the following fatal errors occur:
consistency check failure, PCI Bus errors.
Table 13-2. Run/Stop, Debug Bit Interaction
SWDBG (Bit 5) Run/Stop (Bit 0)
Operation
0
0
If executing a command, the Host Controller completes the command
and then stops. The 1.0 ms frame counter is reset and command list
execution resumes from start of frame using the frame list pointer
selected by the current value in the FRNUM register. (While Run/
Stop=0, the FRNUM register can be reprogrammed).
0
1
1
1
0
1
Execution of the command list resumes from Start Of Frame using the
frame list pointer selected by the current value in the FRNUM register.
The Host Controller remains running until the Run/Stop bit is cleared
(by Software or Hardware).
If executing a command, the Host Controller completes the command
and then stops and the 1.0 ms frame counter is frozen at its current
value. All status is preserved. The Host Controller begins execution of
the command list from where it left off when the Run/Stop bit is set.
Execution of the command list resumes from where the previous
execution stopped. The Run/Stop bit is set to 0 by the Host Controller
when a TD is being fetched. This causes the Host Controller to stop
again after the execution of the TD (single step). When the Host
Controller has completed execution, the HC Halted bit in the Status
Register is set.
Intel® 460GX Chipset Software Developer’s Manual
13-9
Universal Serial Bus (USB) Configuration
13.3.2
USBSTS–USB Status Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (02-03h)
0000h
Read/Write Clear
This register indicates pending interrupts and various states of the Host Controller. The status
resulting from a transaction on the serial bus is not indicated in this register. Software sets a bit to 0
in this register by writing a 1 to it.
Bit
Description
15:6
5
Reserved.
HCHalted. The Host Controller sets this bit to 1 after it has stopped executing as a result of the
Run/Stop bit being set to 0, either by software or by the Host Controller hardware (debug mode or
an internal error).
4
3
Host Controller Process Error. The Host Controller sets this bit to 1 when it detects a fatal error
and indicates that the Host Controller suffered a consistency check failure while processing a
Transfer Descriptor. An example of a consistency check failure would be finding an illegal PID field
while processing the packet header portion of the TD. When this error occurs, the Host Controller
clears the Run/Stop bit in the Command register to prevent further schedule execution. A hardware
interrupt is generated to the system.
Host System Error. The Host Controller sets this bit to 1 when a serious error occurs during a host
system access involving the Host Controller module. In a PCI system, conditions that set this bit to
1 include PCI Parity error, PCI Master Abort, and PCI Target Abort. When this error occurs, the Host
Controller clears the Run/Stop bit in the Command register to prevent further execution of the
scheduled TDs. A hardware interrupt is generated to the system.
2
1
0
Resume Detect. The Host Controller sets this bit to 1 when it receives a “RESUME” signal from a
USB device. This is only valid if the Host Controller is in a global suspend state (bit 3 of Command
register = 1).
USB Error Interrupt. The Host Controller sets this bit to 1 when completion of a USB transaction
results in an error condition (e.g. error counter underflow). If the TD on which the error interrupt
occurred also had its IOC bit set, both this bit and Bit 0 are set.
USB Interrupt (USBINT). The Host Controller sets this bit to 1 when the cause of an interrupt is a
completion of a USB transaction whose Transfer Descriptor had its IOC bit set.
The Host Controller also sets this bit to 1 when a short packet is detected (actual length field in TD
is less than maximum length field in TD), and short packet detection is enabled in that TD.
13.3.3
USBINTR–USB Interrupt Enable Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (04-05h)
0000h
Read/Write
This register enables and disables reporting of the corresponding interrupt to the software. When a
bit is set and the corresponding interrupt is active, an interrupt is generated to the host. Fatal errors
(Host Controller Processor Error- bit 4, USBSTS Register) cannot be disabled by the host
controller. Interrupt sources that are disabled in this register still appear in the Status Register to
allow the software to poll for events.
Bit
Description
15:4
Reserved.
3
2
1
0
Short Packet Interrupt Enable. 1=Enabled. 0=Disabled.
Interrupt On Complete (IOC) Enable. 1= Enabled. 0=Disabled.
Resume Interrupt Enable. 1= Enabled. 0=Disabled.
Time-out/CRC Interrupt Enable. 1= Enabled. 0=Disabled.
13-10
Intel® 460GX Chipset Software Developer’s Manual
Universal Serial Bus (USB) Configuration
13.3.4
FRNUM–Frame Number Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (06-07h)
0000h
Read/Write (Writes must be Word Writes)
Bits [10:0] of this register contain the current frame number which is included in the frame SOF
packet. This register reflects the count value of the internal frame number counter. Bits [9:0] are
used to select a particular entry in the Frame List during schedule execution. This register is
updated at the end of each frame time.
This register must be written as a word. Byte writes are not supported. This register cannot be
written unless the Host Controller is in the STOPPED state as indicated by the HCHalted bit
(USBSTS register). A write to this register while the Run/Stop bit is set (USBCMD register) is
ignored.
Bit
Description
15:11
10:0
Reserved.
Frame List Current Index/Frame Number. Bits [10:0] provide the frame number in the SOF
Frame. The value in this register increments at the end of each time frame (approximately every 1
ms). In addition, bits [9:0] are used for the Frame List current index and correspond to memory
address signals [11:2].
13.3.5
FLBASEADD–Frame List Base Address Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (08-0Bh)
Undefined
Read/Write
This 32-bit register contains the beginning address of the Frame List in the system memory. HCD
loads this register prior to starting the schedule execution by the Host Controller. When written,
only the upper 20 bits are used. The lower 12 bits are written as zero (4-Kbyte alignment). The
contents of this register are combined with the frame number counter to enable the Host Controller
to step through the Frame List in sequence. The two least significant bits are always 00. This
requires Dword alignment for all list entries. This configuration supports 1024 Frame List entries.
Bit
Description
31:12
11:0
Base Address. These bits correspond to memory address signals [31:12], respectively.
Reserved. Must be written as 0s.
13.3.6
SOFMOD–Start of Frame (SOF) Modify Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (0Ch)
40h
Read/Write
This 1-byte register is used to modify the value used in the generation of SOF timing on the USB.
Only the 7 least significant bits are used. When a new value is written into the these 7 bits, the SOF
timing of the next frame will be adjusted. This feature can be used to adjust out any offset from the
clock source that generates the clock that drives the SOF counter. This register can also be used to
maintain real time synchronization with the rest of the system so that all devices have the same
sense of real time. Using this register, the frame length can be adjusted across the full range
Intel® 460GX Chipset Software Developer’s Manual
13-11
Universal Serial Bus (USB) Configuration
required by the USB specification. It’s initial programmed value is system dependent based on the
accuracy of hardware USB clock and is initialized by system BIOS. It may be reprogrammed by
USB system software at any time. Its value will take effect from the beginning of the next frame.
This register is reset upon a Host Controller Reset or Global Reset. Software must maintain a copy
of its value for reprogramming if necessary.
Bit
Description
7
Reserved.
6:0
SOF Timing Value. Guidelines for the modification of frame time are contained in Chapter 7 of the
USB Specification. The SOF cycle time (number of SOF counter clock periods to generate a SOF
frame length) is equal to 11936 + value in this field. The default value is decimal 64 which gives a
SOF cycle time of 12000. For a 12 MHz SOF counter clock input, this produces a 1 ms Frame
period. The following table indicates what SOF Timing Value to program into this field for a certain
frame period.
Frame Length
(# 12 MHz Clocks) SOF Reg. Value
(decimal) (decimal)
11936
11937
0
1
.
.
.
.
11999 63
12000 64
12001 65
.
.
.
.
12062 126
12063 127
13.3.7
PORTSC–Port Status and Control Register (I/O)
I/O Address:
Base + (10-11h)–Port 0
Base + (12-13h)–Port 1
0080h
Default:
Access:
Read/Write (WORD writeable only)
After a Power-up reset, Global reset, or Host Controller reset, the initial conditions of a port are:
No device connected, Port disabled, and the bus line status is 00 (single-ended zero). Note: If a
device is attached, the port state will transition to the attached state and system software will
process this as with any status change notification. It may take up to 64 USB bit times for the port
transition to occur. If the Host Controller is in global suspend mode, then, if any of bits [6,3,1] gets
set, the Host Controller will signal a global resume. Refer to Chapter 11 of the USB Specification
for details on hub operation.
Bit
Description
15:13
Reserved. Must written as 0s when writing this register.
13-12
Intel® 460GX Chipset Software Developer’s Manual
Universal Serial Bus (USB) Configuration
Bit
Description
12
Suspend–R/W. 1=Port in suspend state. 0=Port not in suspend state. This bit should not be
written to a 1 if global suspend is active (bit 3=1 in the USBCMD register). Bit 2 and bit 12 of this
register define the hub states as follows:
Bits [12,2] Hub Port State
x0 Disable
01 Enable
11 Suspend
When in suspend state, downstream propagation of data is blocked on this port, except for single-
ended 0 resets (global reset and port reset). The blocking occurs at the end of the current
transaction, if a transaction was in progress when this bit was written to 1. In the suspend state,
the port is sensitive to resume detection. Note that the bit status does not change until the port is
suspended and that there may be a delay in suspending a port if there is a transaction currently in
progress on the USB.
11
10
9
Over-current Indicator Change–R/WC. 1=A change from 1 to 0 has been detected on the Over-
current (OC[X]#) pin for this port. 0=No change has been detected. Software sets this bit to 0 by
writing a 1 to it.
Over-current Indicator–RO. 1=Overcurrent pin (OC[X]#) for this port is at logic 0 indicating over-
current condition. 0=Overcurrent pin for this port is at logic 1 indicating a normal condition. If
asserted, the corresponding port is disabled.
Port Reset–R/W. 1=Port is in Reset. 0=Port is not in Reset. When in the Reset State, the port is
disabled and sends the USB Reset signaling. Note that host software must ensure that the
RESET signaling is active for the proper amount of time as specified in the USB Specification.
8
Low Speed Device Attached–RO. 1=Low speed device is attached to this port. 0=Full speed
device. Writes have no effect.
7
6
Reserved–RO. Always read as 1.
Resume Detect–R/W. 1= Resume detected/driven on port. 0=No resume (K-state) detected/
driven on port. Software sets this bit to a 1 to drive resume signaling. The Host Controller sets this
bit to a 1 if a J-to-K transition is detected while the port is in the Suspend state. Note that when this
bit is 1, a K-state is driven on the port as long as this bit remains 1 and the port is still in suspend
state. Writing a 0 (from 1) causes the port to send a low speed EOP. This bit will remain a 1 until
the EOP has completed.
5:4
3
Line Status–RO. These bits reflect the D+ (bit 4) and D- (bit 5) signals lines’ logical levels. These
bits are used for fault detect and recovery as well as for USB diagnostics. This field is updated at
EOF2 time (See Chapter 11 of the USB Specification).
Port Enable/Disable Change–R/WC. 1=Port enabled/disabled status has changed. 0=No
change. For the root hub, this bit gets set only when a port is disabled due to disconnect on the
that port or due to the appropriate conditions existing at the EOF2 point (See Chapter 11 of the
USB Specification). Software clears this bit by writing a 1 to it.
2
1
Port Enabled/Disabled–R/W. 1=Enable. 0=Disable. Ports can be enabled by host software only.
Ports can be disabled by either a fault condition (disconnect event, overcurrent condition, or other
fault condition) or by host software. Note that the bit status does not change until the port state
actually changes and that there may be a delay in disabling or enabling a port if there is a
transaction currently in progress on the USB.
Connect Status Change–R/WC. 1=Change in Current Connect Status. 0=No change.
Indicates a change has occurred in the port’s Current Connect Status (see bit 0). The hub device
sets this bit for any changes to the port device connect status, even if system software has not
cleared a connect status change. If, for example, the insertion status changes twice before system
software has cleared the changed condition, hub hardware will be “setting” an already-set bit (i.e.
the bit will remain set). However, the hub transfers the change bit only once when the Host
Controller requests a data transfer to the Status Change endpoint. System software is responsible
for determining state change history in such a case. Software sets this bit to 0 by writing a 1 to it.
0
Current Connect Status–RO. 1=Device is present on port. 0=No device is present. This value
reflects the current state of the port, and may not correspond directly to the event that caused the
Connect Status Change bit (Bit 1) to be set.
Intel® 460GX Chipset Software Developer’s Manual
13-13
Universal Serial Bus (USB) Configuration
13-14
Intel® 460GX Chipset Software Developer’s Manual
SM Bus Controller Configuration 14
The IFB PCI Function 3 contains the SMBus Controller configuration space.
14.1
SM Bus Configuration Registers (Function 3)
Configuration
Register
Access
Mnemonic
Register
Offset
00–01h
02–03h
04–05h
06–07h
08h
VID
DID
PCICMD
PCISTS
RID
CLASSC
–
Vendor Identification
Device Identification
PCI Command
RO
RO
R/W
R/WC
RO
RO
–
PCI Device Status
Revision Identification
Class Code
09-0Bh
0C-1Fh
20-23h
24–3Bh
2C–2Dh
2E–2Fh
30–3Fh
3Ch
Reserved
BAR
–
Base Address Register
Reserved
R/W
–
SVID
SID
Subsystem Vendor ID
Subsystem ID
RO
RO
–
–
Reserved
IL
Interrupt Line
RW
RO
–
3Dh
IP
Interrupt Pin
3E-3Fh
40h
–
Reserved
HC
Host Configuration
Slave Command Port
Slave Shadow Address 1
Slave Shadow Address 2
Reserved
RW
RW
RW
RW
–
41h
SCOM
SS1
SS2
–
42h
43h
41-FFh
Intel® 460GX Chipset Software Developer’s Manual
14-1
SM Bus Controller Configuration
14.2
System Management Register Descriptions
This section describes in detail the registers associated with the IFB System Management Function.
14.2.1
VID–Vendor Identification Register (Function 3)
Address Offset:
Default Value:
Attribute:
00–01h
8086h
Read only
The VID Register contains the vendor identification number. This register, along with the Device
Identification Register, uniquely identifies any PCI device. Writes to this register have no effect.
Bit
Description
15:0
Vendor Identification Number. This is a 16-bit value assigned to Intel.
14.2.2
DID–Device Identification Register (Function 3)
Address Offset:
Default Value:
Attribute:
02-03h
7603h
Read only
The DID Register contains the device identification number. This register, along with the VID
Register, defines the IFB Power Management Controller. Writes to this register have no effect.
Bit
Description
15:0
Device Identification Number. This is a 16-bit value assigned to the IFB System Management
Controller.
14.2.3
PCICMD–PCI Command Register (Function 3)
Address Offset:
Default Value:
Attribute:
04-05h
00h
Read/Write
This register controls access to the I/O space registers.
Bit
Description
15:10
Reserved. Read 0.
9
8:5
4
Fast Back to Back Enable (Not Implemented). This bit is hardwired to 0.
Reserved. Read as 0.
Memory Write and Invalidate Enable (Not Implemented). This bit is hardwired to 0.
3
Special Cycle Enable (SCE). 1 = Enable, the IFB recognizes the Stop Grant special cycle.
0=Disable. The SCE bit in Function 1 PCI Command register controls IFB response to the
Shutdown special cycle.
2
Bus Master Enable (Not Implemented). This bit is hardwired to 0.
14-2
Intel® 460GX Chipset Software Developer’s Manual
SM Bus Controller Configuration
Bit
Description
1
Memory Space Enable (Not Implemented). 1=Enable. 0=Disable. This bit controls the access to
memory space. If this bit is set, access to the memory space by power management logic is
enabled.
0
I/O Space Enable (IOSE). 1=Enable. 0=Disable. This bit controls the access to the I/O space
registers. If this bit is set, access to the power management I/O registers is enabled. The base
register for the I/O registers must be programmed before this bit is set.
14.2.4
PCISTS–PCI Device Status Register (Function 3)
Address Offset:
Default Value:
Attribute:
06-07h
0280h
Read/Write
DSR is a 16-bit status register that reports the occurrence of a PCI target-abort when the System
Management Function is a target device. The register also indicates the System Management
DEVSEL# signal timing that is hardwired in the module. The DSR fields are shown in the table
below.
Bit
Description
15
14
13
12
11
Detected Parity (Not Implemented). This bit is hardwired to 0.
SERR# Status (Not Implemented). This bit is hardwired to 0.
Master-Abort Status (Not Implemented). This bit is hardwired to 0.
Received Target-Abort Status (Not Implemented). This bit is hardwired to 0.
Signaled Target-Abort Status (STA)–R/WC. This bit is set when the Power Management
Function is targeted with a transaction that the it terminates with a target abort. Software resets
STA to 0 by writing a 1 to this bit.
10:9
DEVSEL# Timing Status (DEVT)–RO. This 2-bit field defines the timing for DEVSEL#
assertion. These read only bits indicate the IFB’s DEVSEL# timing when performing a positive
decode. Since the IFB always generate the DEVSEL# with medium timing, DEVT=01. This
DEVSEL# timing does not include Configuration cycles.
8
7
Data Parity Detected (Not Implemented). This bit is hardwired to 0.
Fast Back to Back Capable (FBC)–RO. Hardwired to 1. This bit indicates to the PCI Master
that Power Management as a target is capable of accepting fast back-to-back transactions.
6:0
Reserved. Read as 0’s.
14.2.5
RID–Revision Identification Register (Function 3)
Address Offset:
Default Value:
08h
Initial Stepping = 00h. Refer to IFB Specification Updates
for other values programmed here.
Read only
Attribute:
This 8 bit register contains device stepping information. Writes to this register have no effect.
Bit
Description
7:0
Revision ID Byte. The register is hardwired to the default value.
Intel® 460GX Chipset Software Developer’s Manual
14-3
SM Bus Controller Configuration
14.2.6
CLASSC–Class Code Register (Function 3)
Address Offset:
Default Value:
Attribute:
09-0Bh
0C0500h
Read only
This register identifies the Base Class Code, Sub-Class Code, and Device Programming interface
for IFB PCI Function 3.
Bit
Description
23:16
15:8
7:0
Base Class Code (BASEC). 0Ch = Serial Bus Controller.
Sub-Class Code (SCC). 05h = System Management Bus (SMBus) Controller.
Programming Interface (PI). 00h = No specific register level programming defined.
14.2.7
SMBBA–SMBus Base Address (Function 3)
Address Offset:
Default Value:
Attribute:
20-23h
00000001h
Read/Write
This register contains the base address of the SMBus I/O Registers.
Bit
Description
Reserved. Hardwired to 0s. Must be written as 0s.
31:16
15:4
Index Register Base Address. Bits [15:4] correspond to I/O address signals AD [15:4],
respectively.
3:1
0
Reserved. Read as 0.
Resource Type Indicator (RTE)–RO. This bit is hardwired to 1 indicating that the base address
field in this register maps to I/O space.
14.2.8
SVID–Subsystem Vendor ID (Function 3)
Address:
Default Value:
Attribute:
2C-2Dh
0000h
Read only
Bit
Description
15:0
Subsystem Vendor ID.
14-4
Intel® 460GX Chipset Software Developer’s Manual
SM Bus Controller Configuration
14.2.9
SID–Subsystem ID (Function 3)
Address:
Default Value:
Attribute:
2E-2Fh
0000h
Read only
Bit
Description
15:0
Subsystem ID.
14.2.10 INTLN–Interrupt Line Register (Function 3)
Address Offset:
Default Value:
Attribute:
3Ch
00h
Read/Write
Software programs this register with interrupt information concerning the Power Management
module.
Bit
Description
7:0
Interrupt Line. The value in this register has no affect on IFB hardware operations.
14.2.11 INTPN–Interrupt Pin (Function 3)
Address Offset:
Default Value:
Attribute:
3Dh
02h
Read only
This register indicates that PCI interrupt PIRQB# is used for the Power Management module.
Bit
Description
7:3
2:0
Reserved.
Serial Bus Module Interrupt Routing. This field is hardwired to 02h to indicate that PCI
interrupt pin PIRQB# is used.
14.2.12 Host Configuration
Address Offset:
Default Value:
Attribute:
40h
00h
Read/Write
Bit
Description
7:2
1
Reserved.
SMI_EN: When this bit is set, any source of an SMB interrupt will instead be routed to generate
an SMI#. This bit will only take effect if the INTREN bit is set in I/O space.
0
HST_EN: When set, the SMB Host Controller interface is enabled to execute commands. The
HST_INT_EN bit needs to be enabled in order for the SMB Host Controller to interrupt or SMI#.
Additionally, the SMB Host Controller will not respond to any new requests until all interrupt
requests have been. The HST_EN bit does not affect the SMB Slave Port.
Intel® 460GX Chipset Software Developer’s Manual
14-5
SM Bus Controller Configuration
14.2.13 smbslvc–SMBus Slave Command (Function 3)
Address Offset:
Default Value:
Attribute:
41h
00h
Read/Write
Bit
Description
7:0
SMBus Host Slave Command (SMBCMD)–R/W. Specifies the command values to be
matched for SMBus master accesses to the SMBus controller host slave interface (SMBus port
10h).
14.2.14 smbshdw1–SMBus Slave Shadow Port 1 (Function 3)
Address Offset:
Default Value:
Attribute:
42h
00h
Read/Write
Bit
Description
7:0
SHDW1_ADD: Slave shadow address 1. When an SMB master generates an access to the port
defined by this register and the SHDW1_EN bit is set in I/O space, then the SHDW1_STS bit is
set and an interrupt or resume event is generated.
14.2.15 smbshdw2–SMBus Slave Shadow Port 2 (Function 3)
Address Offset:
Default Value:
Attribute:
43h
00h
Read/Write
Bit
Description
7:0
SHDW2_ADD: Slave shadow address 2. When an SMB master generates an access to the port
defined by this register and the SHDW2_EN bit is set in I/O space, then the SHDW2_STS bit is
set and an interrupt or resume event is generated.
14.3
SMBus I/O Space Registers
The “Base” address is programmed in the IFB PCI Configuration Space for Function 3, Offset 20h-
23h.
14-6
Intel® 460GX Chipset Software Developer’s Manual
SM Bus Controller Configuration
14.3.1
smbhststs–SMBus Host Status Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (00h)
00h
Read/Write
This register provides status information concerning the SMBus controller host interface.
Bit
Description
7:5
4
Reserved.
Failed (FAILED)–R/WC. 1 = Indicates that the source of SMBus interrupt was a failed bus
transaction, set when KILL bit is set (SMBHSTCNT register). 0 = SMBus interrupt not caused by
KILL bit. This bit is only set by hardware and can only be reset by writing a 1 to this bit position.
3
2
BUS COLLISION(BUS_ERR)–R/WC. 1 = Indicates that the source of SMBus interrupt was a
transaction collision. 0 = SMBus interrupt not caused by transaction collision. This bit is only set
by hardware and can only be reset by writing a 1 to this bit position.
Device Error (DEV_ERR)–R/WC. 1 = Indicates that the source of SMBus interrupt was the
generation of an SMBus transaction error. 0 = SMBus interrupt not caused by transaction error.
This bit is only set by hardware and can only be reset by writing a 1 to this bit position.
Transaction errors are caused by:
Illegal Command Field
Unclaimed Cycle (host initiated)
Host Device Time-out
1
0
SMBus Interrupt (INTER)–R/WC. 1 = Indicates that the source of SMBus interrupt was the
completion of the last host command. 0 = SMBus interrupt not caused by host command
completion. This bit is only set by hardware and can only be reset by writing a 1 to this bit
position.
Host Busy (HOST_BUSY)–RO. 1 = Indicates that the SMBus controller host interface is in the
process of completing a command. 0 = SMBus controller host interface is not processing a
command. None of the other registers should be accessed if this bit is set.
14.3.2
smbslvsts–SMBus Slave Status Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (01h)
00h
Read/Write
This register provides status information concerning the SMBus controller slave interface.
Bit
Description
7:5
4
Reserved.
Shadow2 Status (SHDW2_STS)–R/WC. 1 = Indicates that the source of SMBus interrupt or
resume event was a slave cycle address match of the SMBSHDW2 port. 0 = SMBus interrupt
not caused by address match to SMBSHDW2 port. This bit is only set by hardware and can only
be reset by writing a 1 to this bit position.
3
Shadow1 Status (SHDW1_STS)–R/WC. 1 = Indicates that the source of SMBus interrupt or
resume event was a slave cycle address match of the SMBSHDW1 port. 0 = SMBus interrupt
not caused by address match to SMBSHDW1 port. This bit is only set by hardware and can only
be reset by writing a 1 to this bit position.
Intel® 460GX Chipset Software Developer’s Manual
14-7
SM Bus Controller Configuration
Bit
Description
2
Slave Status (SLV_STS)–R/WC. 1 = Indicates that the source of SMBus interrupt or resume
event was a slave cycle event match of the SMBSLVC (command match) and SMBSLVEVT
(data event match). 0 = SMBus interrupt not caused by slave event match. This bit is only set by
hardware and can only be reset by writing a 1 to this bit position.
1
0
Reserved.
Slave Busy (SLV_BSY)–RO. 1 = Indicates that the SMBus controller slave interface is in the
process of receiving data. 0 = SMBus controller slave interface is not processing data. None of
the other registers should be accessed if this bit is set.
14.3.3
smbhstcnt–SMBus Host Control Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (02h)
00h
Read/Write
The control register is used to enable SMBus controller host interface Functions. Reads to this
register clears the host interface’s index pointer to the block data storage array.
Bit
Description
7
6
Reserved.
Start (START)–R/W. 1 = Writing a 1 to this bit initiates the SMBus controller host interface to
execute the command programmed in the SMB_CMD_PROT field. All necessary registers
should be setup prior to writing a 1 to this bit position. 0 = Writing a zero has no effect. This bit
always reads zero. The HOST_BUSY bit can be used to identify when the SMBus host
controller has finished executing the command.
5
SMB_IDX_CLR: Any read to this register clears the slave interface’s internal index pointer to
the block SRAM array. This bit always reads zero, but exists as a reminder that reads to this
register clear the pointer index to the block SRAM array.
4:2
SMBus Command Protocol (SMB_CMD_PROT)–R/W. Selects the type of command the
SMBus controller host interface will execute. Reads or writes are determined by bit 0 of
SMBHSTADD register. This field is decoded as follows:
[4:2] Protocol
0,0,0 Quick Read or Write
0,0,1 Byte Read or Write
0,1,0 Byte Data Read or Write
0,1,1 Word Data Read or Write
1,0,0 Block Read or Write
1,0,1 Reserved
1,1,0 Reserved
1,1,1 Reserved
1
0
Kill (KILL)–R/W. 1 = Stop the current in process SMBus controller host transaction. This sets
the FAILED status bit and asserts the interrupt selected by the SMB_INTRSEL field. 0 = Allows
the SMBus controller host interface to Function normally.
Interrupt Enable (INTEREN)–R/W. 1 = Enable the generation of interrupts upon the completion
of the current host transaction. 0 = Disable.
14-8
Intel® 460GX Chipset Software Developer’s Manual
SM Bus Controller Configuration
14.3.4
smbhstcmd–SMBus Host Command Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (03h)
00h
Read/Write
This register is transmitted by the SMBus controller host interface in the command field of the
SMBus protocol.
Bit
Description
7:0
SMBus Host Command (HST_CMD)–R/W. This field contains the data transmitted in the
command field of SMBus host transaction.
14.3.5
smbhstadd–SMBus Host Address Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (04h)
00h
Read/Write
This register is transmitted by the SMBus controller host interface in the slave address field of the
SMBus protocol.
Bit
Description
7:1
SMBus Address (SMB_ADDRESS)–R/W. This field contains the 7-bit address of the targeted
slave device.
0
SMBus Read or Write (SMB_RW)–R/W. 1 = Execute a READ command. 0 = Execute a WRITE
command.
14.3.6
smbhstdat0–SMBus Host Data 0 Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (05h)
00h
Read/Write
This register is transmitted by the SMBus controller host interface in the Data0 field of the SMBus
protocol.
Bit
Description
7:0
SMBus Data 0 (SMBD0)–R/W. This register should be programmed with the value to be
transmitted in the Data0 field of an SMBus host interface transaction. For a block write
command, the count of the memory block should be stored in this field. The value of this register
is loaded into the block transfer count field. This register must be programmed to a value
between 1 and 32 for block command counts. A count of 0 or a count above 32 will result in
unpredictable behavior. For block reads, the count received from the SMBus device is stored
here.
Intel® 460GX Chipset Software Developer’s Manual
14-9
SM Bus Controller Configuration
14.3.7
smbhstdat1–SMBus Host Data 1 Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (06h)
00h
Read/Write
This register is transmitted by the SMBus controller host interface in the Data1 field of the SMBus
protocol.
Bit
Description
7:0
SMBus Data 1 (SMBD1)–R/W. This register should be programmed with the value to be
transmitted in the Data1 field of an SMBus host interface transaction.
14.3.8
smbblkdat–SMBus Block Data Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (07h)
00h
Read/Write
Reads and writes to this register are used to access the 32 byte block data storage array. An internal
index pointer is used to address the array. It is reset to 0 by reading the SMBHSTCNT register. The
index pointer then increments automatically upon each access to this register. The transfer of block
data into (read) or out of (write) this storage array during an SMBus transaction always starts at
index address 0.
Bit
Description
7:0
SMBus Block Data (BLK_DAT)–R/W. This register is used to transfer data into or out of the
block data storage array.
14.3.9
smbslvcnt–SMBus Slave Control Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (08h)
00h
Read/Write
The control register is used to enable SMBus controller slave interface Functions.
Bit
Description
7
SLV_INT_EN: When set to a ‘1’, the generation of a slave interrupt or SMI# based on a master
SMB device generating an access to the host controller’s slave port is enabled. The slave port
will set the SLV_STS bit in the Slave Status register (offset 01h). The data will placed in the
Slave Data register (offset 0Ah).
6:3
2
Reserved.
SMBus Shadow Port 2 Enable(SHDW2_EN)–R/W. SLV_INT_EN and SHDW2_EN =1 will
enable the generation of an interrupt or resume event upon an external SMBus master
generating a transaction with an address that matches the SMBSHDW2 register. 0 = Disable.
1
SMBus Shadow Port 1 Enable(SHDW1_EN)–R/W. SLV_INT_EN and SHDW1_EN =1 will
enable the generation of an interrupt or resume event upon an external SMBus master
generating a transaction with an address that matches the SMBSHDW1 register. 0 = Disable.
14-10
Intel® 460GX Chipset Software Developer’s Manual
SM Bus Controller Configuration
Bit
Description
0
Slave Enable (SLV_EN)–R/W. 1 = Enable the generation of an interrupt or resume event upon
an external SMBus master generating a transaction with an address that matches the host
controller slave port of 10h, a command field which matches the SMBSLVC register, and a
match of one of the corresponding enabled events in the SMBSLVEVT register. 0 = Disable.
14.3.9.1
10.3.10.smbshdwcmd–SMBus Shadow Command Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (09h)
00h
Read only
This register is used to store command values for external SMBus master accesses to the host slave
and slave shadow ports.
Bit
Description
7:0
Shadow Command (SHDW_CMD)–RO. This field contains the command value which was
received during an external SMBus master access whose address field matched the host slave
address (10h) or one of the slave shadow port addresses.
14.3.9.2
10.3.11.smbslvevt–SMBus Slave Event Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (0Ah)
0000h
Read/Write
This register is used to enable generation of interrupt or resume events for accesses to the host
controller’s slave port.
Bit
Description
15:0
SM BUS Slave Event (SMB_SLV_EVT)–R/W. This field contains data bits used to compare
against incoming data to the SMBSLVDAT register. When a bit in this register is a 1 and the
corresponding bit in the SMBSLVDAT register is set, then an interrupt or resume event will be
generated if the command value matches the value in the SMBSLVC register and the access
was to SMBus host address 10h.
14.3.10 smbslvdat–SMBus Slave Data Register (I/O)
I/O Address:
Default Value:
Attribute:
Base + (0Ch)
0000h
Read only
This register is used to store data values for external SMBus master accesses to the shadow ports or
the SMBus host controller’s slave port.
Bit
Description
15:0
SLAVE DATA (SMB_SLV_DATA)–RO. This field contains the data value which was transmitted
during an external SMBus master access whose address field matched one of the slave shadow
port addresses or the SMBus host controller slave port address of 10h.
Intel® 460GX Chipset Software Developer’s Manual
14-11
SM Bus Controller Configuration
14-12
Intel® 460GX Chipset Software Developer’s Manual
PCI/LPC Bridge Description
15
15.1
PCI Interface
The IFB incorporates a fully PCI Bus compatible master and slave interface. As a PCI master, the
IFB runs cycles on behalf of DMA, Bus Master IDE, and USB. The IFB implements an internal
arbiter to request the PCI bus IDE and USB for these master Functions.
All memory cycles run by the IFB master interface target system DRAM.
As a PCI slave, the IFB responds to memory cycles destined for the firmware and I/O cycles to the
integrated legacy Functions (8237, 8254, 8259), the integrated IDE controller, and the relocate-able
I/O spaces for ACPI, IDE, USB, and SM Bus.
15.1.1
Transaction Termination
The IFB supports the PCI cycle termination as described in the PCI Local Bus specification. IFB
As Master–Master-initiated Termination: The IFB supports three forms of master-initiated
termination: 1.) Normal termination of a completed transaction, 2.) Normal termination of an
incomplete transaction due to time-out (applies to line buffer operations-IDE Bus Master, 3.)
Abnormal termination due to the slave not responding to the transaction (Abort) The IFB As a
Master–Response to Target-initiated Termination: As a master, the IFB responds correctly to
the target-termination– Target-Abort, Retry, or Disconnect.
IFB as a Target–Target-initiated Termination: The IFB supports three forms of Target-initiated
Termination– Disconnect, Retry, and Target Abort.
15.1.2
Parity Support
As a master, the IFB generates address parity for read/write cycles and data parity when the IFB is
providing the data. As a slave, the IFB generates data parity for read cycles. The IFB does not
check parity and does not generate SERR# due to an address parity error. However, the IFB does
generate an NMI when another PCI device asserts SERR# (if enabled).
PAR is the calculated parity signal. PAR is even parity and is calculated on 36 bits–AD[31:0]
signals plus C/BE[3:0]#. PAR is always calculated on 36 bits, regardless of the valid byte enables.
PAR is valid one PCI clock after the corresponding address or data phase.
15.1.3
PCI Arbitration
The IFB arbitrates for the PCI Bus through the PHOLD# and PHLDA# signals.
15.2
Interrupt Controller
The IFB provides the functionality of two 82C59 interrupt controllers. The two controllers are
cascaded so that 13 external and three internal interrupts are possible. The master interrupt
controller provides IRQ [7:0] and the slave interrupt controller provides IRQ [15:8]. The three
Intel® 460GX Chipset Software Developer’s Manual
15-1
PCI/LPC Bridge Description
internal interrupts are used for internal Functions only. IRQ2 is used to cascade the two controllers
together and is not available to the user. IRQ0 is used as a system timer interrupt and is tied to
Interval Timer 1, Counter 0. IRQ13 is connected internally to FERR#. The remaining 13 interrupt
lines (IRQ1, IRQ3-IRQ12, IRQ14, and IRQ15) are available for external system interrupts. Edge
or level sense selection is programmable on an individual channel by channel basis, except for
IRQ0, IRQ2, IRQ8#, and IRQ13.
The Interrupt unit also supports interrupt steering. The IFB can be programmed to allow the four
PCI active low interrupts (PIRQA#-PIRQD#) to be internally routed to one of 11 interrupts: 3 - 7,
9-12, 14 or 15.
The Interrupt Controller consists of two separate 82C59 cores. Interrupt Controller 1 (CNTRL-1)
and Interrupt Controller 2 (CNTRL-2) are initialized separately and can be programmed to operate
in different modes. The default settings are: IA-32 Compatibility Mode, Edge Sensitive (IRQ0-15)
Detection, Normal EOI, Non-Buffered Mode, Special Fully Nested Mode disabled, and Cascade
Mode. CNTRL-1 is connected as the Master Interrupt Controller and CNTRL-2 is connected as the
Slave Interrupt Controller.
Note that IRQ13 is generated internally (as part of the coprocessor error support) by the IFB.
15.2.1
Programming the Interrupt Controller
The Interrupt Controller accepts two types of command words generated by the CPU or bus
master:
15.2.1.1
Initialization Command Words (ICWs)
Before normal operation can begin, each Interrupt Controller in the system must be initialized. In
the 82C59, this is a two to four byte sequence. However, for the IFB, each controller must be
initialized with a four byte sequence. This four byte sequence is required to configure the interrupt
controller correctly for the IFB implementation.
The four initialization command words are referred to by their acronyms: ICW1, ICW2, ICW3, and
ICW4.
The base address for each interrupt controller is a fixed location in the I/O memory space, at 0020h
for CNTRL-1 and at 00A0h for CNTRL-2.
An I/O write to the CNTRL-1 or CNTRL-2 base address with data bit 4 equal to 1 is interpreted as
ICW1. For IFB-based systems, three I/O writes to “base address + 1" (021h for CNTRL-1 and
0A1h for CNTRL-2) must follow the ICW1. The first write to “base address + 1” (021h/0A1h)
performs ICW2, the second write performs ICW3, and the third write performs ICW4.
ICW1 starts the initialization sequence.
ICW2 is programmed to provide bits [7:3] of the interrupt vector that will be released onto the data
bus by the interrupt controller during an interrupt acknowledge. A different base [7:3] is selected
for each interrupt controller.
ICW3 is programmed differently for CNTRL-1 and CNTRL-2, and has a different meaning for
each controller.
For CNTRL-1, the master controller, ICW3 is used to indicate which IRQx input line is used to
cascade CNTRL-2, the slave controller. Within the IFB interrupt unit, IRQ2 on CNTRL-1 is used
to cascade the INTR output of CNTRL-2. Consequently, bit-2 of ICW3 on CNTRL-1 is set to a 1,
and the other bits are set to 0's.
15-2
Intel® 460GX Chipset Software Developer’s Manual
PCI/LPC Bridge Description
For CNTRL-2, ICW3 is the slave identification code used during an interrupt acknowledge cycle.
CNTRL-1 broadcasts a code to CNTRL-2 over three internal cascade lines if an IRQ[x] line from
CNTRL-2 won the priority arbitration on the master controller and was granted an interrupt
acknowledge by the CPU. CNTRL-2 compares this identification code to the value stored in
ICW3, and if the code is equal to bits [2:0] of ICW3, CNTRL-2 assumes responsibility for
broadcasting the interrupt vector during the second interrupt acknowledge cycle pulse.
ICW4 must be programmed on both controllers. At the very least, bit 0 must be set to a 1 to
indicate that the controllers are operating in an Intel Architecture-based system.
15.2.1.2
Operation Command Words (OCWs)
These are the command words which dynamically reprogram the Interrupt Controller to operate in
various interrupt modes. Any interrupt lines can be masked by writing an OCW1. A 1 written in
any bit of this command word will mask incoming interrupt requests on the corresponding IRQx
line.
OCW2 is used to control the rotation of interrupt priorities when operating in the rotating priority
mode and to control the End of Interrupt (EOI) Function of the controller.
OCW3 is used to set up reads of the ISR and IRR, to enable or disable the Special Mask Mode
(SMM), and to set up the interrupt controller in polled interrupt mode. The OCWs can be written
into the Interrupt Controller any time after initialization.
15.2.2
End of Interrupt Operation
15.2.2.1
End of Interrupt (EOI)
The In Service (IS) bit can be set to 0 automatically following the trailing edge of the second
INTA# pulse (when AEOI bit in ICW1 is set to 1) or by a command word that must be issued to the
Interrupt Controller before returning from a service routine (EOI command). An EOI command
must be issued twice with this cascaded interrupt controller configuration, once for the master and
once for the slave.
There are two forms of EOI commands: Specific and Non-specific. When the Interrupt Controller
is operated in modes which preserve the fully nested structure, it can determine which IS bit to set
to 0 on EOI. When a non-Specific EOI command is issued, the Interrupt Controller will
automatically set to 0 the highest IS bit of those that are set to 1, since in the fully nested mode the
highest IS level was necessarily the last level acknowledged and serviced. A non-specific EOI can
be issued with OCW2 (EOI=1, SL=0, R=0).
When a mode is used which may disturb the fully nested structure, the Interrupt Controller may no
longer be able to determine the last level acknowledged. In this case a Specific End of Interrupt
must be issued which includes as part of the command the IS level to be reset. A specific EOI can
be issued with OCW2 (EOI=1, SL=1, R=0, and LO-L2 is the binary level of the IS bit to be set to
0).
It should be noted that an IS bit that is masked by an IMR bit will not be cleared by a non-specific
EOI if the Interrupt Controller is in the Special Mask Mode.
15.2.2.2
Automatic End of Interrupt (AEOI) Mode
If AEOI=1 in ICW4, then the Interrupt Controller will operate in AEOI mode continuously until
reprogrammed by ICW4. Note that reprogramming ICW4 implies that ICW1, ICW2, and ICW3
must be reprogrammed first, in sequence. In this mode, the Interrupt Controller will automatically
Intel® 460GX Chipset Software Developer’s Manual
15-3
PCI/LPC Bridge Description
perform a non-specific EOI operation at the trailing edge of the last interrupt acknowledge pulse.
Note that from a system standpoint, this mode should be used only when a nested multi-level
interrupt structure is not required within a single Interrupt Controller. The AEOI mode can only be
used in a master Interrupt Controller and not a slave (on CNTRL-1 but not CNTRL-2).
15.2.3
Modes of Operation
15.2.3.1
Fully Nested Mode
This mode is entered after initialization unless another mode is programmed. The interrupt requests
are ordered in priority from 0 through 7 (0 being the highest). When an interrupt is acknowledged,
the highest priority request is determined and its vector placed on the bus. Additionally, a bit of the
Interrupt Service Register (IS[0:7]) is set. This IS bit remains set until the processor issues an End
of Interrupt (EOI) command immediately before returning from the service routine. If the AEOI
(Automatic End of Interrupt) is set, the IS bit remains set until the trailing edge of the second
INTA#. With the IS bit set, all further interrupts of same or lower priority are inhibited, while
higher levels will generate an interrupt (which will be acknowledged if the processor internal
interrupt enable flip-flop has been re-enabled through software).
After the initialization sequence, IRQ0 has the highest priority and IRQ7 the lowest. Priorities can
be changed, as will be explained, in the rotating priority mode.
15.2.3.2
The Special Fully Nested Mode
This mode will be used in the case of a system where cascading is used, and the priority has to be
conserved within each slave. In this case, the special fully nested mode will be programmed to the
master (using ICW4). This mode is similar to the normal nested mode with the following
exceptions:
When an interrupt request from a certain slave is in service, this slave is not locked out from the
master’s priority logic and further interrupt requests from higher priority IRQs within the slave will
be recognized by the master and will initiate interrupts to the processor. (In the normal nested
mode, a slave is masked out when its request is in service and no higher requests from the same
slave can be serviced.)
When exiting the Interrupt Service routine, the software has to check whether the interrupt serviced
was the only one from that slave. This is done by sending a non-specific End of Interrupt (EOI)
command to the slave and then reading its In-service Register and checking for zero. If it is empty,
a non-specific EOI can be sent to the master too. If not, no EOI should be sent.
15.2.3.3
Automatic Rotation (Equal Priority Devices)
In some applications, there are a number of interrupting devices of equal priority. Automatic
rotation mode provides for a sequential 8-way rotation. In this mode, a device receives the lowest
priority after being serviced. In the worst case, a device requesting an interrupt will have to wait
until each of seven other devices are serviced at most once.
There are two ways to accomplish automatic rotation using OCW2; the Rotation on Non-Specific
EOI Command (R=1, SL=0, EOI=1) and the Rotate in Automatic EOI Mode which is set by (R=1,
SL=0, EOI=0) and cleared by (R=0, SL=0, EOI=0).
15-4
Intel® 460GX Chipset Software Developer’s Manual
PCI/LPC Bridge Description
15.2.3.4
Specific Rotation (Specific Priority)
The programmer can change priorities by programming the bottom priority and thus fixing all
other priorities. For example, if IRQ5 is programmed as the bottom priority device, then IRQ6 will
be the highest priority device.
The Set Priority Command is issued in OCW2 where: R=1, SL=1; LO-L2 is the binary priority
level code of the bottom priority device. See the register description for the bit definitions.
Note that, in this mode, internal status is updated by software control during OCW2. However, it is
independent of the End of Interrupt (EOI) command (also executed by OCW2). Priority changes
can be executed during an EOI command by using the Rotate on Specific EOI Command in OCW2
(R=1, SL=1, EOI=1 and LO-L2=IRQ level to receive bottom priority).
15.2.3.5
Poll Command
The Polled Mode can be used to conserve space in the interrupt vector table. Multiple interrupts
that can be serviced by one interrupt service routine do not need separate vectors if the service
routine uses the poll command.
The Polled Mode can also be used to expand the number of interrupts. The polling interrupt service
routine can call the appropriate service routine, instead of providing the interrupt vectors in the
vector table.
In this mode, the INTR output is not used and the microprocessor internal Interrupt Enable flip-flop
is reset, disabling its interrupt input. Service to devices is achieved by software using a Poll
Command.
The Poll command is issued by setting P=1 in OCW3. The Interrupt Controller treats the next I/O
read pulse to the Interrupt Controller as an interrupt acknowledge, sets the appropriate IS bit if
there is a request, and reads the priority level. Interrupts are frozen from the I/O write to the I/O
read.
This mode is useful if there is a routine command common to several levels so that the INTA#
sequence is not needed (saves ROM space).
15.2.4
Cascade Mode
The Interrupt Controllers in the IFB are interconnected in a cascade configuration with one master
and one slave. This configuration can handle up to 15 separate priority levels.
The master controls the slaves through a three line internal cascade bus. When the master drives
010b on the cascade bus, this bus acts like a chip select to the slave controller.
In a cascade configuration, the slave interrupt outputs are connected to the master interrupt request
inputs. When a slave request line is activated and afterwards acknowledged, the master will enable
the corresponding slave to release the interrupt vector address during the second INTA# cycle of
the interrupt acknowledge sequence.
Each Interrupt Controller in the cascaded system must follow a separate initialization sequence and
can be programmed to work in a different mode. An EOI Command must be issued twice: once for
the master and once for the slave.
Intel® 460GX Chipset Software Developer’s Manual
15-5
PCI/LPC Bridge Description
15.2.5
Edge and Level Triggered Mode
This mode is programmed using bit 3 in ICW1. With IFB this bit is disabled and a new register for
edge and level triggered mode selection, per interrupt input, is included. This is the Edge/Level
control Registers ELCR1 and ELCR2. The default programming is equivalent to programming the
LTIM bit (ICW1 bit 3) to a 0 (all interrupts selected for edge triggered mode). Note, that IRQ0, 1,
2, 8#, and 13 can not be programmed for level sensitive mode and can not be modified by software.
If an ELCR bit = “0”, an interrupt request will be recognized by a low to high transition on the
corresponding IRQx input. The IRQ input can remain high without generating another interrupt.
If an ELCR bit = “1”, an interrupt request will be recognized by a high level on the corresponding
IRQ input and there is no need for an edge detection. The interrupt request must be removed before
the EOI command is issued to prevent a second interrupt from occurring.
In both the edge and level triggered modes, the IRQ inputs must remain active until after the falling
edge of the first INTA#. If the IRQ input goes inactive before this time, a default IRQ7 will occur
when the CPU acknowledges the interrupt. This can be a useful safeguard for detecting interrupts
caused by spurious noise glitches on the IRQ inputs. To implement this feature, the IRQ7 routine is
used for “clean up” simply executing a return instruction, thus ignoring the interrupt. If IRQ7 is
needed for other purposes, a default IRQ7 can still be detected by reading the ISR. A normal IRQ7
interrupt will set the corresponding ISR bit; a default IRQ7 will not set this bit. If a default IRQ7
routine occurs during a normal IRQ7 routine, however, the ISR will remain set. In this case, it is
necessary to keep track of whether or not the IRQ7 routine was previously entered. If another IRQ7
occurs, it is a default.
15.2.6
Interrupt Masks
15.2.6.1
Masking on an Individual Interrupt Request Basis
Each interrupt request input can be masked individually by the Interrupt Mask Register (IMR).
This register is programmed through OCW1. Each bit in the IMR masks one interrupt channel, if it
is set to a 1. Bit 0 masks IRQ0, Bit 1 masks IRQ1 and so forth. Masking an IRQ channel does not
affect the other channel's operation, with one exception. Masking IRQ2 on CNTRL-1 will mask off
all requests for service from CNTRL-2. The CNTRL-2 INTR output is physically connected to the
CNTRL-1 IRQ2 input.
15.2.6.2
Special Mask Mode
Some applications may require an interrupt service routine to dynamically alter the system priority
structure during its execution under software control. For example, the routine may wish to inhibit
lower priority requests for a portion of its execution but enable some of them for another portion.
The difficulty is that if an Interrupt Request is acknowledged and an End of Interrupt command did
not reset its IS bit (i.e. while executing a service routine), the Interrupt Controller would have
inhibited all lower priority requests with no easy way for the routine to enable them.
The Special Mask Mode enables all interrupts not masked by a bit set in the Mask Register.
Interrupt service routines that require dynamic alteration of interrupt priorities can take advantage
of the Special Mask Mode. For example, a service routine can inhibit lower priority requests during
a part of the interrupt service, then enable some of them during another part.
In the Special Mask Mode, when a mask bit is set to 1 in OCW1, it inhibits further interrupts at that
level and enables interrupts from all other levels (lower as well as higher) that are not masked.
15-6
Intel® 460GX Chipset Software Developer’s Manual
PCI/LPC Bridge Description
Thus, any interrupts may be selectively enabled by loading the Mask Register with the appropriate
pattern.
Without Special Mask Mode, if an interrupt service routine acknowledges an interrupt without
issuing an EOI to clear the IS bit, the interrupt controller inhibits all lower priority requests. The
Special Mask Mode provides an easy way for the interrupt service routine to selectively enable
only the interrupts needed by loading the Mask register.
The special Mask Mode is set by OCW3 where: SSMM=1, SMM=1, and cleared where SSMM=1,
SMM=0.
15.2.7
Reading the Interrupt Controller Status
The input status of several internal registers can be read to update the user information on the
system. The Interrupt Request Register (IRR) and In-Service Register (ISR) can be read via
OCW3. The Interrupt Mask Register (IMR) is read via a read of OCW1. Here are brief descriptions
of the ISR, the IRR, and the IMR.
Interrupt Request Register (IRR): 8-bit register which contains the status of each interrupt request
line. Bits that are clear indicate interrupts that have not requested service. The Interrupt Controller
clears the IRR’s highest priority bit during an interrupt acknowledge cycle. (Not affected by IMR).
In-Service Register (ISR): 8-bit register indicating the priority levels currently receiving service.
Bits that are set indicate interrupts that have been acknowledged and their interrupt service routine
started. Bits that are cleared indicate interrupt requests that have not been acknowledged, or
interrupt request lines that have not been asserted. Only the highest priority interrupt service
routine executes at any time. The lower priority interrupt services are suspended while higher
priority interrupts are serviced. The ISR is updated when an End of Interrupt Command is issued.
Interrupt Mask Register (IMR): 8-bit register indicating which interrupt request lines are masked.
The IRR can be read when, prior to the I/O read cycle, a Read Register Command is issued with
OCW3 (RR=1, RIS=0).
The ISR can be read when, prior to the I/O read cycle, a Read Register Command is issued with
OCW3 (RR=1, RIS=1).
The interrupt controller retains the ISR/IRR status read selection following each write to OCW3.
Therefore, there is no need to write an OCW3 before every status read operation, as long as the
current status read corresponds to the previously selected register. For example, if the ISR is
selected for status read by an OCW3 write, the ISR can be read over and over again without writing
to OCW3 again. However, to read the IRR, OCW3 will have to be reprogrammed for this status
read prior to the OCW3 read to check the IRR. This is not true when poll mode is used. Polling
Mode overrides status read when P=1, RR=1 in OCW3.
After initialization the Interrupt Controller is set to read the IRR.
As stated, OCW1 is used for reading the IMR. The output data bus will contain the IMR status
whenever I/O read is active the address is 021h or 061h (OCW1).
15.2.8
Interrupt Steering
The IFB can be programmed to allow 4 PCI programmable interrupts (PIRQA#-PIRQD#) to be
internally routed to one of 11 interrupts: 3 - 7, 9-12, 14 or 15. PCLK is used to synchronize the
PIRQx# inputs. The PIRQx# lines are run through an internal multiplexer that assigns, or routes, an
Intel® 460GX Chipset Software Developer’s Manual
15-7
PCI/LPC Bridge Description
individual PIRQx# line to any one of 11 IRQ inputs. The assignment is programmable through the
PIRQx Route Control registers. One or more PIRQx# lines can be routed to the same IRQx input.
If interrupt steering is not required, the Route Registers can be programmed to disable steering.
Bits 0-3 in each PIRQx Route Control register are used to route the associated PIRQx# line to an
internal IRQ input. Bit 7 in each register is used to disable routing of the associated PIRQx#.
The PIRQx# lines are defined as active low, level sensitive to allow multiple interrupts on a PCI
Board to share a single line across the connector. When a PIRQx# is routed to specified IRQ line,
the software must change the IRQ’s corresponding ELCR bit to level sensitive mode.
15.3
Serial Interrupts
The IFB supports a serial IRQ scheme. Because more than one device may need to share the single
serial IRQ signal, an Open Collector signaling scheme is used. Timing is based on the PCI Clock.
If the PCI clock is inactive when a device needs to signal an interrupt, the CLKRUN# signal must
first be asserted by the device to restart the PCI clock.
The serial IRQ configuration is handled via the PCI configuration space. No other registers are
associated with the scheme.
15.3.1
Protocol
Serial interrupt information is transferred using three types of frames: a Start Frame, one or more
IRQ Data frames, and one Stop frame. There are also two modes of operation: Quiet Mode and
Continuous Mode.
15.3.1.1
Quiet (Active) Mode
To indicate an interrupt, the peripheral brings the SERIRQ signal low for one clock, and then tri-
states it. This brings all the state machines from IDLE to the ACTIVE states.
The IFB will then take control of the SERIRQ signal by driving it low on the next clock, and will
continue driving it low for 3-7 clocks more (programmable). Thus the total number of clocks low
will be 4-8. After those clocks, the IFB will drive SERIRQ high for one clock and then tri-state the
signal.
15.3.1.2
15.3.1.3
Continuous (Idle) Mode
In this mode, the IFB initiates the START frame, rather than the peripherals. Typically this is done
to update IRQ status (acknowledges). The IFB will drive SERIRQ low for 4, 6, 8 clocks depending
on bits SERIRQC register (Function 0), bits [1:0].
This is the default mode after reset, and can be used to enter the Quiet mode.
Data Frame
Once the Start frame has been initiated, all of the serial interrupt peripherals must start counting
frames based on the rising edge of SERIRQ. Each of the IRQ/DATA frames has exactly 3 phases of
1 clock each: a Sample phase, a Recovery Phase, and a Turn-around phase.
15-8
Intel® 460GX Chipset Software Developer’s Manual
PCI/LPC Bridge Description
During the Sample phase, the device drives SERIRQ low if the corresponding interrupt signal is
low. If the corresponding interrupt is high, then the devices will tri-state the SERIRQ signal. It will
remain high due to pull-up resistors.
During the other two phases (turnaround and recovery), no device should drive the SERIRQ signal.
If an SMI# is activated on frame 3, the IFB will drive its SMI# signal low. This will then generate
an SMI# to the microprocessor if enabled.
Table 15-1. SERIRQ Frames
Data Frame Number
Usage
# Clocks Past Start
1
2
UNASSIGNED
IRQ1
2
5
3
SMI#
8
4
IRQ3
11
14
17
20
23
26
29
32
35
38
41
44
47
50
53
56
59
62
96
5
IRQ4
6
IRQ5
7
IRQ6
8
IRQ7
9
UNASSIGNED
IRQ9
10
11
12
13
14
15
16
17
18
19
20
21
32:22
IRQ10
IRQ11
IRQ12
UNASSIGNED
IRQ14
IRQ15
IOCHCK#
PCI INTA#
PCI INTB#
PCI INTC#
PCI INTD#
UNASSIGNED
15.3.1.4
Stop Frame
After all of the data frames, a Stop Frame will be done by the IFB. The IFB will drive SERIRQ low
for 2-3 PCI clocks. The number of clocks determines the next mode:
• If SERIRQ is low for 2 clocks, then the next mode is the Quite Mode. Any device may initiate
a Start Frame in the second clock (or more) after the rising edge of the Stop Frame.
• If SERIRQ is low for 3 clocks, then the next mode is the Continuous mode. Only the IFB may
initiate a Start Frame in the second clock (or more) after the rising edge of the Stop Frame.
Intel® 460GX Chipset Software Developer’s Manual
15-9
PCI/LPC Bridge Description
15.4
Timer/Counters
The IFB contains three counters that are equivalent to those found in the 82C54 programmable
interval timer. The three counters are contained in one IFB timer unit, referred to as Timer-1. Each
counter output provides a key system Function. Counter 0 is connected to interrupt controller IRQ0
and provides a system timer interrupt for a time-of-day, diskette time-out, or other system timing
Functions. Counter 1 generates a refresh request signal and Counter 2 generates the tone for the
speaker. The 14.31818 MHz counters normally use OSC as a clock source.
Counter 0, System Timer
This counter Functions as the system timer by controlling the state of IRQ0 and is typically
programmed for Mode 3 operation. The counter produces a square wave with a period equal to the
product of the counter period (838 ns) and the initial count value. The counter loads the initial
count value one counter period after software writes the count value to the counter I/O address. The
counter initially asserts IRQ0 and decrements the count value by two each counter period. The
counter negates IRQ0 when the count value reaches 0. It then reloads the initial count value and
again decrements the initial count value by two each counter period. The counter then asserts IRQ0
when the count value reaches 0, reloads the initial count value, and repeats the cycle, alternately
asserting and negating IRQ0.
Counter 1, Refresh Request Signal
This counter provides the refresh request signal and is typically programmed for Mode 2 operation.
The counter negates refresh request for one counter period (838 ns) during each count cycle. The
initial count value is loaded one counter period after being written to the counter I/O address. The
counter initially asserts refresh request, and negates it for 1 counter period when the count value
reaches 1. The counter then asserts refresh request and continues counting from the initial count
value.
Counter 2, Speaker Tone
This counter provides the speaker tone and is typically programmed for Mode 3 operation. The
counter provides a speaker frequency equal to the counter clock frequency (1.193 MHz) divided by
the initial count value. The speaker must be enabled by a write to port 061h (see NMI Status and
Control ports).
15.4.1
Programming the Interval Timer
The counter/timers are programmed by I/O accesses and are addressed as though they are
contained in one 82C54 interval timer. A single Control Word Register controls the operation of all
three counters.
The interval timer is an I/O-mapped device. Several commands are available:
The Control Word Command specifies:
• Which counter to read or write.
• The operating mode.
• The count format (binary or BCD).
15-10
Intel® 460GX Chipset Software Developer’s Manual
PCI/LPC Bridge Description
The Counter Latch Command latches the current count so that it can be read by the system. The
countdown process continues.
The Read Back Command reads the count value, programmed mode, the current state of the OUT
pins, and the state of the Null Count Flag of the selected counter.
The Read/Write Logic selects the Control Word Register during an I/O write when address lines
A[1:0]=11. This condition occurs during an I/O write to port address 043h, the address for the
Control Word Register on Timer 1. If the CPU writes to port 043h, the data is stored in the Control
Word Register and is interpreted as the Control Word used to define the operation of the Counters.
The Control Word Register is write-only. Counter status information is available with the read back
Command.
Because the timer counters wake up in an unknown state after power up, multiple refresh requests
may be queued. To avoid possible multiple refresh cycles after power up, program the timer
counter immediately after power up.
15.4.1.1
Write Operations
Programming the interval timer is a simple process:
1. Write a control word.
2. Write an initial count for each counter.
3. Load the least and/or most significant bytes (as required by Control Word bits 5, 4) of the 16-
bit counter.
The programming procedure for the IFB timer is very flexible. Only two conventions need to be
observed. First, for each counter, the control word must be written before the initial count is
written. Second, the initial count must follow the count format specified in the control word (least
significant byte only, most significant byte only, or least significant byte and then most significant
byte).
Since the Control Word Register and the three counters have separate addresses (selected by the
A1, A0 inputs), and each control word specifies the counter it applies to (SC0, SC1 bits), no special
instruction sequence is required. Any programming sequence that follows the conventions above is
acceptable.
A new initial count may be written to a counter at any time without affecting the counter’s
programmed mode. Counting will be affected as described in the mode definitions. The new count
must follow the programmed count format.
If a counter is programmed to read/write two-byte counts, the following precaution applies: A
program must not transfer control between writing the first and second byte to another routine
which also writes into that same counter. Otherwise, the counter will be loaded with an incorrect
count.
15.4.1.2
Interval Timer Control Word Format
The control word specifies the counter, the operating mode, the order and size of the count value,
and whether it counts down in a 16-bit or binary-coded decimal (BCD) format. After writing the
control word, a new count may be written at any time. The new value will take effect according to
the programmed mode.
Intel® 460GX Chipset Software Developer’s Manual
15-11
PCI/LPC Bridge Description
If a counter is programmed to read/write two-byte counts, the following precaution applies: A
program must not transfer control between writing the first and second byte to another routine
which also writes into that same counter. Otherwise, the counter will be loaded with an incorrect
count. The count must always be completely loaded with both bytes.
15.4.1.3
15.4.1.4
Read Operations
It is often desirable to read the value of a counter without disturbing the count in progress. This is
easily done in the IFB timer unit. There are three possible methods for reading the counters: a
simple read operation, the Counter Latch Command, and the Read-Back Command.
Counter I/O Port Read
The first method is to perform a simple read operation. To read the counter, which is selected with
the A1, A0 inputs (port 040h, 041h, or 042h), the CLK input of the selected counter must be
inhibited by using either the GATE input or external logic. Otherwise, the count may be in the
process of changing when it is read, giving an undefined result. When reading the count value
directly, follow the format programmed in the control register: read LSB, read MSB, or read LSB
then MSB. Within the IFB timer unit, the GATE input on Counter 0 and Counter 1 is tied high.
Therefore, the direct register read should not be used on these two counters. The GATE input of
Counter 2 is controlled through I/O port 061h. If the GATE is disabled through this register, direct
I/O reads of port 042h will return the current count value.
15.4.1.5
Counter Latch Command
The Counter Latch Command latches the count at the time the command is received. This
command is used to ensure that the count read from the counter is accurate (particularly when
reading a two-byte count). The count value is then read from each counter’s Count Register as was
programmed by the Control Register.
The selected counter’s output latch (OL) latches the count at the time the Counter Latch Command
is received. This count is held in the latch until it is read by the CPU (or until the Counter is
reprogrammed). The count is then unlatched automatically and the OL returns to “following” the
counting element (CE). This allows reading the contents of the counters “on the fly” without
affecting counting in progress. Multiple Counter Latch Commands may be used to latch more than
one counter. Each latched counter's OL holds its count until it is read. Counter Latch Commands do
not affect the programmed mode of the counter in any way. The Counter Latch Command can be
used for each counter in the IFB timer unit.
If a Counter is latched and then, some time later, latched again before the count is read, the second
Counter Latch Command is ignored. The count read will be the count at the time the first Counter
Latch Command was issued.
With either method, the count must be read according to the programmed format; specifically, if
the counter is programmed for two byte counts, two bytes must be read. The two bytes do not have
to be read one right after the other. Read, write, or programming operations for other counters may
be inserted between them.
Another feature of the IFB timer is that reads and writes of the same counter may be interleaved.
For example, if the Counter is programmed for two byte counts, the following sequence is valid:
1. Read least significant byte.
2. Write new least significant byte.
3. Read most significant byte.
4. Write new most significant byte.
15-12
Intel® 460GX Chipset Software Developer’s Manual
PCI/LPC Bridge Description
If a counter is programmed to read/write two-byte counts, a program must not transfer control
between reading the first and second byte to another routine which also reads from that same
counter. Otherwise, an incorrect count will be read.
15.4.1.6
Read Back Command
The third method uses the Read Back Command. The Read Back Command is used to determine
the count value, programmed mode, and current states of the OUT pin and Null Count flag of the
selected counter or counters. The Read Back Command is written to the Control Word Register,
which causes the current states of the above mentioned variables to be latched. The value of the
counter and its status may then be read by I/O access to the counter address.
The Read Back Command may be used to latch multiple counter output latches (OL) by setting the
COUNT# bit D5=0 and selecting the desired counter(s). This single command is Functionally
equivalent to several counter latch commands, one for each counter latched. Each counter’s latched
count is held until it is read (or the counter is reprogrammed). Once read, a counter is automatically
unlatched. The other counters remain latched until they are read. If multiple count Read-Back
Commands are issued to the same counter without reading the count, all but the first are ignored
(i.e. the count which will be read is the count at the time the first Read-Back Command was
issued).
The Read Back Command may also be used to latch status information of selected counter(s) by
setting STATUS# bit D4=0. Status must be latched to be read. The status of a counter is accessed
by a read from that counter’s I/O port address.
If multiple counter status latch operations are performed without reading the status, all but the first
are ignored. The status returned from the read is the counter status at the time the first status Read
Back Command was issued.
Both count and status of the selected counter(s) may be latched simultaneously by setting both the
COUNT# and STATUS# bits [5:4]=00. This is functionally the same as issuing two consecutive,
separate Read Back Commands. The above discussions apply here also. Specifically, if multiple
count and/or status Read Back Commands are issued to the same counter(s) without any
intervening reads, all but the first are ignored.
If both count and status of a counter are latched, the first read operation from that counter will
return the latched status, regardless of which was latched first. The next one or two reads
(depending on whether the counter is programmed for one or two type counts) return the latched
count. Subsequent reads return unlatched count.
15.5
Real Time Clock
The Real Time Clock (RTC) module provides a battery backed-up date and time keeping device
with two banks of static RAM with 128 bytes each, although the first bank has 114 bytes for
general purpose usage. Three interrupt features are available: time of day alarm with once a second
to once a month range, periodic rates of 122 µs to 500 ms, and end of update cycle notification.
Seconds, minutes, hours, days, day of week, month, and year are counted. Daylight savings
compensation is optional. The hour is represented in twelve or twenty-four hour format, and data
can be represented in BCD or binary format. The time keeping comes from a 32.768 kHz
oscillating source, which is divided to achieve an update every second. The lower 14 bytes on the
lower RAM block have very specific Functions. The first ten are for time and date information.
The next four (0Ah to 0Dh) are registers, which configure and report RTC Functions.
Intel® 460GX Chipset Software Developer’s Manual
15-13
PCI/LPC Bridge Description
The time and calendar data should match the data mode (BCD or binary) and hour mode (12 or 24
hour) as selected in register B. It is up to the programmer to make sure that data stored in these
locations is within the reasonable values ranges and represents a possible date and time. The
exception to these ranges is to store a value of C0 - FF in the Alarm bytes to indicate a don’t care
situation. All Alarm conditions must match to trigger an Alarm Flag, which could trigger an Alarm
Interrupt if enabled. The UIP bit should be read as 0 before each access to these registers. The SET
bit of register B should be one while programming these locations to avoid clashes with an update
cycle. Access to time and date information is done through the RAM locations. If a RAM read
from the ten time and date bytes is attempted during an update cycle, the value read will not
necessarily represent the true contents of those locations. Any RAM writes under the same
conditions will be ignored.
15.5.1
RTC Registers and RAM
The RTC internal registers and RAM are organized as two banks of 128 bytes each, called the
standard and extended banks. The first 14 bytes of the standard bank contain the RTC time and date
information along with four registers, A - D, that are used for configuration of the RTC. The
extended bank contains a full 128 bytes of battery backed SRAM, and will be accessible even when
the RTC module is disabled (via the RTC configuration register).
All data movement between the host CPU and the real-time clock is done through registers mapped
to the I/O space at locations 70-73h.
I/O locations 70h and 71h are the standard RAM location for the real-time clock. I/O locations 72h
and 73h are the extended RAM, and may be disabled. When disabled, 72h and 73h become aliases
for 70h and 71h respectively. The addressing is done by a indexing scheme: 70h (72h) is written
with the index, and 71h (73h) is written with data or read for data. This scheme is shown in
Table 15-2. RTC (Standard) RAM Bank
Index Address
Name
00h
01h
Seconds
Seconds Alarm
Minutes
02h
03h
Minutes Alarm
Hours
04h
05h
Hours Alarm
Day of Week
Date of Month
Month
06h
07h
08h
09h
Year
0Ah
Register A
0Bh
Register B
0Ch
0Dh
0Eh - 7Fh
Register C
Register D
114 Bytes of User RAM
15-14
Intel® 460GX Chipset Software Developer’s Manual
PCI/LPC Bridge Description
The extended RAM bank is also accessed using an indexed scheme. I/O address 72h is used as the
address pointer and I/O address 73h is used as the data register. Index addresses above 127h are not
valid.
15.5.1.1
Register A
Address Offset:
Default Value:
Attribute:
0Ah
NA - This register is not affected by any system reset signal.
Read/Write
This register is used for general configuration of the RTC Functions.
Bit(s)
Description
7
UPDATE IN PROGRESS (UIP): This bit may be monitored as a status flag. When asserted as a 1,
the update is soon to occur or is in progress. If 0, the update cycle will not start for at least 244 µs.
The time, calendar, and alarm information in RAM is always available when the UIP bit is 0.
6:4
Division Chain Select (DVx): These three bits control the divider chain for the oscillator.
DV2 DV1 DV0 Function
0 1 0 Normal Operation
1 1 X Divider Reset
1 0 1 Bypass 15 stages (test mode only)
1 0 0 Bypass 10 stages (test mode only)
0 1 1 Bypass 5 stages (test mode only)
0 1 1 Invalid
0 0 0 Invalid
3:0
Rate Select Bits (RSx): Selects one of 13 taps of the 15 stage divider chain. The selected tap can
generate a periodic interrupt if the PIE bit is set in register B. Otherwise this tap will set the PF flag
of register C. If the periodic interrupt is not to be used, these bits should all be set to zero.
RS3
RS2
RS1
RS0
Periodic Rate
0
0
0
0
Interrupt never toggles
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
3.90625 ms
7.8125 ms
122.070 µs
244.141 µs
488.281 µs
976.5625µs
1.953125 ms
3.90625 ms
7.8125 ms
15.625 ms
31.25 ms
62.5 ms
125 ms
250 ms
500 ms
Intel® 460GX Chipset Software Developer’s Manual
15-15
PCI/LPC Bridge Description
15.5.1.2
Register B
Address Offset:
Default Value:
Attribute:
0Bh
X0000XXXb
Read/Write
This register is used for general configuration of the RTC Functions.
Bit(s)
Description
7
SET: Enables the update cycles. When is zero, update cycle occurs normally once a second. If
set to one, a current update cycle will abort and subsequent update cycles will not occur until SET
is returned to zero. When set is one, the firmware may initialize time and calendar bytes safely.
This bit is not affected by RSMRST#.
6
5
Periodic Interrupt Enable (PIE): If set to one, the Periodic Interrupt Enable (PIE) bit allows an
interrupt to occur with a time base set with the RS bits of register A. This bit is cleared (set to
zero) on active RSMRST#.
Alarm Interrupt Enable (AIE): If set to one, the Alarm Interrupt Enable (AIE) bit allows an
interrupt to occur when the AF is one as set from an alarm match from the update cycle. An alarm
can occur once a second, one an hour, once a day, or one a month. This bit is cleared on active
RSMRST#.
4
3
Update-ended Interrupt Enable (UIE): If set to one, the Update-ended Interrupt Enable (UIE) bit
allows an interrupt to occur when the update cycle ends. This bit is cleared on active RSMRST#.
Square Wave Enable (SQWE): The Square Wave Enable bit serves no Function in this device,
yet is left in this register bank to provide compatibility with the Motorola 146818B. There is not
SQW pin on this device. This bit is cleared on active RSMRST#.
2
1
Data Mode (DM): The Data Mode (DM) bit specifies either binary or BCD data representation. A
one denotes binary, and zero denotes BCD. This bit is not affected by RSMRST#.
Hour Format (HF): This bit indicates the hour byte format. If one, twenty-four hour mode is
selected. If zero, twelve-hour mode is selected. In twelve hour mode, the seventh bit represents
AM as zero and PM as one. This bit is not affected by RSMRST#.
0
Daylight Savings Enable (DSE): The Daylight Savings Enable bit triggers two special hour
updates per year when set to one. One is on the first Sunday in April, where time increments from
1:59:59 AM to 3:00:00 AM. The other is the last Sunday in October when the time first reaches
1:59:59 AM, it is changed to 1:00:00 AM. The time must increment normally for at least two
update cycles (seconds) previous to these conditions for the time change to occur properly.
These special update conditions do not occur when the DSE bit is set to zero. The days for the
hour adjustment are those specified in United States federal law as of 1987, which is different
than previous years. This bit is not affected by RSMRST#.
15.5.1.3
Register C
Address Offset:
Default Value:
Attribute:
0Ch
00h
Read/Write
This register is used for various flags. All flag bits are cleared upon active RSMRST# or a read of
Register C.
Bit(s)
Description
7
Interrupt Request Flag (IRQF): Interrupt Request Flag = PF * PIE + AF * AIE + UF *UFE. This
also causes the CH_IRQ_B signal to be asserted.
6
Periodic Interrupt Flag (PF): Periodic interrupt Flag will be one when the tap as specified by the
RS bits of register A is one. If no taps are specified, this flag bit will remain at zero.
5
4
Alarm Flag (AF): Alarm Flag will be high after all Alarm values match the current time.
Update-ended Flag (UF): Updated-ended flag will be high immediately following an update cycle
for each second.
3:0
Reserved. Read as 0.
15-16
Intel® 460GX Chipset Software Developer’s Manual
PCI/LPC Bridge Description
15.5.1.4
Register D
Address Offset:
Default Value:
Attribute:
0Dh
NA - This register is not affected by any system reset signal.
Read/Write
This register is used for various flags.
Bit(s)
Description
7
Valid RAM and TIME Bit (VRT): The Valid Ram and Time bit is set to one when the PWRGD
(power good) signal provided is high. This feature is not typically used. This bit should always be
set to 0 for write to this register.
6
Reserved. This bit always returns a 0 and should be set to 0 for write cycles.
5:0
Date Alarm (DA): These bits store the date of month alarm value. If set to 000000, then a don’t
care state is assumed. The host must configure the date alarm for these bits to do anything, yet
they can be written at any time. If the date alarm is not enabled, these bits will return zeros to
mimic the Functionality of the Motorola 146818B. These bits are not affected by RSMRST#.
15.5.2
RTC Update Cycle
An update cycle occurs once a second, if the SET bit of register B is not asserted and the divide
chain is properly configured. During this procedure, the stored time and date will be incremented,
overflow will be checked, a matching alarm condition will be checked, and the time and date will
be rewritten to the RAM locations. The update cycle will start at least 244µs after the UIP bit of
register A is asserted, and the entire cycle will not take more than 1984µs to complete. The time
and date RAM locations (0-9) will be disconnected from the external bus during this time. To avoid
update and data conditions, external RAM access to these locations can safely occur at two times.
When a updated-ended interrupt is detected, almost 999ms are available to read and write valid
time and date data. If the UIP bit of register A is detected to be low, there is at least 244µs before
the update cycle begins. Because the overflow conditions for leap years and daylight savings
adjustments are based on more than one date or time item, the time before one of these conditions
should be set (when adjusting) at least two seconds before one of these conditions to ensure proper
operation.
15.5.3
15.5.4
RTC Interrupts
The real-time clock interrupt is internally routed within the IFB both to the I/O APIC and the 8259.
It is mapped to interrupt vector 8. This interrupt is not shared with any other interrupt. IRQ8# from
the serial stream is ignored.
Lockable RAM Ranges
The real-time clock battery-backed RAM supports two 8-byte ranges that can be enabled via the
configuration space. If the configuration bits are set, the corresponding range in the RAM will not
be readable or writeable. A write cycle to those locations will have no effect. A read cycle to those
locations will not return the actual location value.
Once enabled, this Function can only be disabled by a hard reset.
Intel® 460GX Chipset Software Developer’s Manual
15-17
PCI/LPC Bridge Description
15-18
Intel® 460GX Chipset Software Developer’s Manual
IFB Power Management
16
16.1
Overview
IFB is designed for desktop systems, and includes the following power management features for
the desktop design:
1. Compliance with industry standard specifications:
APM Rev 1.2
ACPI Rev 1.0
Energy Star (30W idle)
PCI Power Management Rev 1.0
2. ACPI S1 Sleep State with STPCLK# and/or SLP# active.
3. ACPI S4/S5 Sleep States (Suspend-to-Disk and Soft-Off).
4. ACPI Power management timer.
5. APCI Compliant Power Button and Thermal Input signals.
6. Ability for firmware to cause SCI and ACPI software to cause SMI#: ACPI Requirement.
7. SMI# and/or SCI generation from various sources, including power management timer.
8. Alt Access Mode: Needed to allow for Suspend-to-Disk.
Table 16-1 shows the power states defined for platforms using the IFB.
Table 16-1. IFB Power States and Consumption
ACPI State/
Substate
Description
S0/C0
S0/C1
ON: CPU operating a full speed with no latencies.
Auto Halt: CPU has executed a Halt instruction. Returns to the S0/C0 state based on
a break event.
S1
Stop-Grant or Sleep: IFB supports two versions of the S1 Sleep state. In the first the
STPCLK# signal is active, much like the S0 state. In the 2 , both STPCLK# and SLP#
nd
are active. This puts the CPU in an even lower power state, but it cannot maintain
cache coherency. Entrance to the S1 state is performed by a write to the SLP_EN bit
with the appropriate SLP_TYP field. Will return to S0 state based on Wake event.
S4
S5
Suspend to Disk (STD): The context of the system is maintained on the disk. All
power is then shut to the system except for the logic required to resume. Entrance to
the S4 state is performed by a write to the SLP_EN bit with the appropriate SLP_TYP
field. Will return to S0 state based on Wake event.
Soft Off (SOFF): No system context saved.These two states are similar. For STD
(ACPI S4 state), SOFF (ACPI S5 or G2 states) is the same, except the context is not
saved. Entrance to the S5 state is performed either by a write to the SLP_EN bit with
the appropriate SLP_TYP field, or by a power button override. Will return to S0 state
based on Wake event.
G3
Power Failure (PFAIL): Power is lost to the system (typically because it has been
unplugged). IFB still maintains the RTC and CMOS RAM with the external backup
battery. Return after power resumption depends on AFTERG3 bit and state prior to the
power failure.
Intel® 460GX Chipset Software Developer’s Manual
16-1
IFB Power Management
16.2
IFB Power Planes
16.2.1
Power Plane Descriptions
The IFB contains three power planes:
RTC Plane
This plane includes the RTC, as well as some of the power management
logic. It is intended to be backed up by a battery, even when all other
power to the system is shut.
Resume Plane
Main (Core) Plane
This plane contains additional power management logic, as well as other
circuits that can wake the systems from the S4-S5 states. The resume
plane will typically be powered by the main power supply’s trickle output.
This includes all other signals.
16.2.2
SMI# Generation
Table 16-2 shows which sources can cause the IFB to drive SMI# active. When operating with an
ACPI-based Operating System, some of the causes of SMI# will instead be routed to cause an SCI.
Upon any SMI# event taking place, the IFB will assert SMI#. SMI# remains active until the EOS
bit is set. When the EOS bit is set, SMI# will go inactive for a minimum of 1 PCICLK. If another
SMI event occurs, SMI# will be driven active again.
Table 16-2. Causes of SMI#
SMI Event
Comment
ACPI SMI# bit (GBL_RLS)
GPIO Assertion
ACPI sets bit to cause SMI#, SMM handler clears the bit. ACPI I/O offset
04h, Bit 2
When a GPIO is programmed as an input and is set to a ‘1’, an SMI# will be
generated. The bit is cleared when the SMM handler clears the asserting
device.
Overflow of ACPI Timer
THRM# signal
Time-out every 2.34 seconds. If SCI_EN is set, the timer overflow will instead
cause an SCI.
The THRM# can cause an SMI# on either the rising or falling edge. If the
SCI_EN is set, the THRM# signal will instead cause an SCI.
Master Aborts of the IFB
Internal bus master state machine sets bit. SMM handler will typically clear
DDMA Logic, USB Controller, the bit, then retry the cycle.
or IDE Controller
Legacy USB Support
Bit set based on address decode or incoming USB IRQ. SMM handler will
clear bits.
1MIN Timer
Needed for legacy power management, this time will generate an SMI# every
minute. The SMM handler can check the Wake/Break status register to see if
there is any system activity. After n minutes of no system activity (where n is
determine by the SMM handler), the SMM handler can decide to put the
system into a lower power state.
SW SMI# Timer
Not to be confused with the above periodic SMI timer.
16-2
Intel® 460GX Chipset Software Developer’s Manual
IFB Power Management
16.2.3
SCI Generation
In an ACPI environment, an SCI (system control interrupt) must be generated for any event that
must be handled by ACPI software. If the SCI_EN bit is set, the IFB will generate an SCI based on
Table 16-3. Causes of SCI#
SCI Event
Comment
Overflow of ACPI Timer
Time-out every 2.34 seconds. If SCI_EN is not set, the timer
overflow will instead cause an SMI#.
THRM# Signal
The THRM# can cause an SCI# on either the rising or falling
edge. If the SCI_EN is not set, the THRM# signal will instead
cause an SMI#.
Setting of the BIOS_RLS Bit
GPIO Assertion
This bit is set by the firmware to cause an SCI. The ACPI
handler will clear the bit.
When a GPIO bit is programmed as an input, and the register bit
is set to a ‘1’. The bit will be cleared when the ACPI handler
clears the asserting device.
SCI is a level mode interrupt. In non-APIC systems (default), the SCI IRQ is routed to IRQ9. The
8259 interrupt controller must be programmed to level mode for that interrupt. In APIC systems,
the SCI IRQ can still be IRQ9, or can be routed to one of the APIC interrupts 20-23. In either case,
the interrupt generated internally is active high level. The interrupt will remain high until all SCI
sources are removed.
16.2.4
Sleep States
The IFB directly supports several sleep states (two of which will typically be mapped to the ACPI
S1 state). From a IFB perspective, the two S1 states only differ on whether the SLP# signal is
active. Additional Sleep states, such as S4 and S5 are also supported.
The entry to the Sleep states are based on several assumptions:
• After setting the SLP_EN bit, the IFB will assert the STPCLK# pin. The IFB will not continue
to step through the sleep sequence unless the Stop Grant special cycle is received on the PCI
bus.
• Entering the Sleep state without at least one wake event set is not recommended. This could
lock up the system. However, the power button will always be a wake event.
• If using the S1 Sleep state with SLP# active, the PCI masters must be prevented from
accessing memory, because the CPU cannot maintain cache coherence.
• In either of the IFB S1 Sleep states, the PCI clock will still be running, so the Serial IRQ
stream will still be available. DMA, USB, and IDE will not be available, because the IFB bus
masters will not be able to access main memory. The USB controllers, can still generate WAK
event, however.
• In the S4 or S5 Sleep states, the PCI clock will NOT be running (and the PCI bus unpowered),
so the Serial IRQ stream will NOT be available. If an external device is still powered during
S4-S5, it should use some other mechanism to request that the IFB wakes the system.
• Upon exit from any Sleep states, the WAK_STS bit will be set.
• Upon exit from any Sleep state, the SLP_TYP bits will contain the originally programmed
values.
Intel® 460GX Chipset Software Developer’s Manual
16-3
IFB Power Management
16.2.5
ACPI Bits Not Implemented by IFB
Many ACPI registers and bits are optional, and do not have to be implemented for a standard
Table 16-4. ACPI Bits Not Implemented in IFB
Offset
Register Name/Function
Comment
00-01h
PM1 Status
4
BM_STS
Not needed for standard desktop.
04-05h
PM1 Control
BM_RLD
1
Stopping CPU clock not supported.
0C-0Dh General Purpose Status
9
GPI_STS
LID_STS
GPI not needed for desktop
Lid not needed for desktop
11
OE-0Fh General Purpose Enables
9
GPI_EN
LID_EN
GPI not needed for desktop
Lid not needed for desktop
11
15h
Level 3 Register
Power state not needed for desktop
16.2.6
Entry/Exit for the S4 and S5 States
As part of the ACPI spec, as well as PC’97 specs, all new desktop systems must support the SOFT
OFF (ACPI S5) state. The state will have the following characteristics:
• No system context preserved.
• Power shut to all subsystems except RTC and wake logic (typically just the power button).
• All system clocks shut except 32.768 kHz internal to the RTC logic.
There are two ways to enter the Soft-Off state:
1. The CPU will write a value of 100 to the SLP_TYP field and set the SLP_EN bit to 1.
2. The power button is pressed for 4 seconds. This is known as a power-button override event.
In either case, the entry to the Soft-Off state is done by asserting the SUSB and SUSC signals. This
will cause the power to be shut and the PWROK signal is assumed to go low.
Note also that there is no need to enter the Soft-Off state gracefully. The STPCLK# and
SUS_STAT# signals don’t have to be asserted in any particular order, since the CPU and memory
controller will be reset after the system is rebooted.
16-4
Intel® 460GX Chipset Software Developer’s Manual
IFB Power Management
A Wake event will cause an exit from the Soft-Off state. The wake events that can wake from the
S5 state are:
S5 Wake Event
Comment
RTC Alarm cause RTC_STS bit set
Power button press causes PWRBTN_STS bit set
Ring Indicate causes RI_STS bit set
GPIO Routed to SCI goes active
RTC_EN must be set for the wake event
Power Button is unconditional wake event
RI_EN must be set for the wake event
EXTSCI_EN must be set for the wake event
EXTSMI_EN must be set for the wake event
GPIO Routed to SMI goes active
16.3
Handling of Power Failures in IFB
A power failure is defined as any one of the following:
• PWROK goes low and the IFB did not yet cause SUSB and/or SUSC to go inactive.
• Power to the Resume plane (as detected by RSMRST# or VccRESUME) goes inactive.
Upon detection of a power failure, the state machine will go to the G3 state and the IFB will set a
new status bit called PWR_FAIL. That bit can be cleared only by writing a 1 to that bit position.
Add new config bit, AFTERG3, powered off RTC well to indicate what should be done after power
failure:
• 1 = Return to S4/S5 state (see table below), 0 = Cold Boot.
• Software should set this bit prior to going to an S4 or S5 state if it desires the system to return
there after the power failure. This bit is automatically set to 1 due to a Power Button Override
event.
When RSMRST# (or the Resume Vcc) is inactive after a power failure, the state machine will
transition based on the following table:
State Prior to Power Failure
AFTERG3
Action after Power Returns
S0, S1
S0, S1
0
1
Cold boot to return to S0 state. SUSB and SUSC go inactive.
No boot. SUSB and SUSC stay active (S5 state) until wake
event.
S4
S4
0
1
Cold boot to return to S0 state. SUSB and SUSC go inactive.
No boot. SUSB and SUSC stay active (S4 state) until wake
event.
S5
S5
0
1
Cold boot to return to S0 state. SUSB and SUSC go inactive.
No boot. SUSB and SUSC stay active (S5 state) until wake
event.
If USB devices are attached, setting the AFTERG3 bit to 1 may not be a good idea, since the USB
devices will lose their configuration during the power failure and may no longer be able to generate
a wake events when the power is restored. In this case, the system may have to always boot after a
power failure (except if placed into the S5 state due to Power Button Override).
Intel® 460GX Chipset Software Developer’s Manual
16-5
IFB Power Management
16-6
Intel® 460GX Chipset Software Developer’s Manual
|