N. Appendix N - Common Platform Error Record (CPER)

N.1. Introduction

This appendix describes the common platform error record (CPER) format for representing platform hardware errors.

N.2. Format

The general format of the common platform error record is illustrated in the Figure below . The record consists of a header; followed by one or more section descriptors; and for each descriptor, an associated section which may contain either error or informational data.

_images/Common_Platform_Error_Record-2.png

Fig. N.1 Error Record Format

N.2.1. Record Header

The record header includes information which uniquely identifies a hardware error record on a given system. The contents of the record header are described in the Table below. The header is immediately followed by an array of one or more section descriptors. Sections may be either error sections, which contain error information retrieved from hardware, or they may be informational sections, which contain contextual information relevant to the error. An error record must contain at least one section.

Table N.1 Error record header

Mnemonic

Byte Offset

Byte Length

Description

Signature Start

0

4

ASCII 4-character array “CPER” (0x43,0 x50,0x45,0x52). Identifies this structure as a hardware error record.

Revision

4

2
This is a 2-byte field representing a major and minor version number for the error record definition in BCD format. The interpretation of the major and minor version number is as follows:
• Byte 0 - Minor (01): An increase in this revision indicates that changes to the headers and sections are backward compatible with software that use earlier revisions. Addition of new GUID types, errata fixes or clarifications are covered by a bump up.
• Byte 1 - Major (01): An increase in this revision indicates that the changes are not backward compatible from a software perspective.

Signature End

6

4

Must be 0xFFFFFFFF

Section Count

10

2

This field indicates the number of valid sections associated with the record, corresponding to each of the following section descriptors.

Error Severity

12

4
Indicates the severity of the error condition. The severity of the error record corresponds to the most severe error section.
0 - Recoverable (also called non-fatal uncorrected)
1 - Fatal
2 - Corrected
3 - Informational
All other values are reserved.
Note that severity of “Informational” indicates that the record could be safely ignored by error handling software.

Validation Bits

16

4
This field indicates the validity of the following fields:
• Bit 0 - If 1, the PlatformID field contains valid information
• Bit 1 - If 1, the TimeStamp field contains valid information
• Bit 2 - If 1, the PartitionID field contains valid information
• Bits 3-31: Reserved, must be zero.

Record Length

20

4

Indicates the size of the actual error record, including the size of the record header, all section descriptors, and section bodies. The size may include extra buffer space to allow for the dynamic addition of error sections descriptors and bodies.

Timestamp

24

8
The timestamp correlates to the time when the error information was collected by the system software and may not necessarily represent the time of the error event. The timestamp contains the local time in BCD format.
• Byte 7 - Byte 0:
• Byte 0: Seconds
• Byte 1: Minutes
• Byte 2: Hours
• Byte 3:
• Bit 0 - Timestamp is precise if this bit is set and correlates to the time of the error event.
• Bit 7:1 - Reserved
• Byte 4: Day
• Byte 5: Month
• Byte 6: Year
• Byte 7: Century

Platform ID

32

16

This field uniquely identifies the platform with a GUID. The platform’s SMBIOS UUID should be used to populate this field. Error analysis software may use this value to uniquely identify a platform.

Partition ID

48

16

If the platform has multiple software partitions, system software may associate a GUID with the partition on which the error occurred.

Creator ID

64

16

This field contains a GUID indicating the creator of the error record. This value may be overwritten by subsequent owners of the record.

Notification Type

80

16
This field holds a pre-assigned GUID value indicating the record association with an error event notification type. The defined types are:

CMC
{0x2DCE8BB1, 0xBDD7, 0x450e, {0xB9, 0xAD, 0x9C, 0xF4, 0xEB, 0xD4, 0xF8, 0x90}}

CPE
{0x4E292F96, 0xD843, 0x4a55, {0xA8, 0xC2, 0xD4, 0x81, 0xF2, 0x7E, 0xBE, 0xEE}}

MCE
{0xE8F56FFE, 0x919C, 0x4cc5, {0xBA, 0x88, 0x65, 0xAB, 0xE1, 0x49, 0x13, 0xBB}}

PCIe
{0xCF93C01F, 0x1A16, 0x4dfc, {0xB8, 0xBC, 0x9C, 0x4D, 0xAF, 0x67, 0xC1, 0x04}}

INIT
{0xCC5263E8, 0x9308, 0x454a, {0x89, 0xD0, 0x34, 0x0B, 0xD3, 0x9B, 0xC9, 0x8E}}

NMI
{0x5BAD89FF, 0xB7E6, 0x42c9, {0x81, 0x4A, 0xCF, 0x24, 0x85, 0xD6, 0xE9, 0x8A}}

Boot
{0x3D61A466, 0xAB40, 0x409a, {0xA6, 0x98, 0xF3, 0x62, 0xD4, 0x64, 0xB3, 0x8F}}

DMAr
{0x667DD791, 0xC6B3, 0x4c27, {0x8A, 0x6B, 0x0F, 0x8E,0x72, 0x2D, 0xEB, 0x41}}

SEA
{0x9A78788A, 0xBBE8, 0x11E4, {0x80, 0x9E, 0x67, 0x61, 0x1E, 0x5D, 0x46, 0xB0}}

SEI
{0x5C284C81, 0xB0AE, 0x4E87, {0xA3, 0x22, 0xB0, 0x4C, 0x85, 0x62, 0x43, 0x23}}

PEI
{0x09A9D5AC, 0x5204, 0x4214, {0x96, 0xE5, 0x94, 0x99, 0x2E, 0x75, 0x2B, 0xCD}}

CXL Component
{0x69293BC9, 0x41DF, 0x49A3 {0xB4, 0xBD, 0x4F, 0xB0, 0xDB, 0x30, 0x41, 0xF6}}

Record ID

96

8

This value, when combined with the Creator ID, uniquely identifies the error record across other error records on a given system.

Flags

104

4

Flags field contains information that describes the error record. See Table 2 for defined flags.

Persistence Information

108

8

This field is produced and consumed by the creator of the error record identified in the Creator ID field. The format of this field is defined by the creator and it is out of scope of this specification.

Reserved

116

12

Reserved. Must be zero.

Section Descriptor

128

Nx72

An array of SectionCount descriptors for the associated sections. The number of valid sections is equivalent to the SectionCount. The buffer size of the record may include more space to dynamically add additional Section Descriptors to the error record.

Error Record Header Flags

The following table lists flags that can be used to qualify an error record in the Error Record Header’s Flags field.

Table N.2 Error Record Header Flags

Value

Description

1

HW_ERROR_FLAGS_RECOVERED: Qualifies an error condition as one that has been recovered by system software.

2

HW_ERROR_FLAGS_PREVERR: Qualifies an error condition as one that occurred during a previous session. For instance, of the OS detects an error and determines that the system must be reset; it will save the error record before stopping the system. Upon restarting the OS marks the error record with this flag to know that the error is not live.

4

HW_ERROR_FLAGS_SIMULATED: Qualifies an error condition as one that was intentionally caused. This allows system software to recognize errors that are injected as a means of validating or testing error handling mechanisms.

N.2.1.1. Notification Type

A notification type identifies the mechanism by which an error event is reported to system software. This information helps consumers of error information (e.g. management applications or humans) by identifying the source of the error information. This allows, for instance, all CMC error log entries to be filtered from an error event log.

Listed below are the standard notification types. Each standard notification type is identified by a GUID. For error notification types that do not conform to one of the standard types, a platform-specific GUID may be defined to identify the notification type.

  • Machine Check Exception (MCE): {0xE8F56FFE, 0x919C, 0x4cc5, {0xBA, 0x88, 0x65, 0xAB, 0xE1, 0x49, 0x13, 0xBB}} A Machine Check Exception is a processor-generated exception class interrupt used to system software of the presence of a fatal or recoverable error condition.

  • Corrected Machine Check (CMC): {0x2DCE8BB1, 0xBDD7, 0x450e, {0xB9, 0xAD, 0x9C, 0xF4,0xEB, 0xD4, 0xF8, 0x90}} Corrected Machine Checks identify error conditions that have been corrected by hardware or system firmware. CMCs are reported by the processor and may be reported via interrupt or by polling error status registers.

  • Corrected Platform Error (CPE): {0x4E292F96, 0xD843, 0x4a55, {0xA8, 0xC2, 0xD4, 0x81, 0xF2, 0x7E, 0xBE, 0xEE}} Corrected Platform Errors identify corrected errors from the platform (i.e., external memory controller, system bus, etc.). CPEs can be reported via interrupt or by polling error status registers.

  • Non-Maskable Interrupt (NMI): {0x5BAD89FF, 0xB7E6, 0x42c9, {0x81, 0x4A, 0xCF, 0x24, 0x85, 0xD6, 0xE9, 0x8A}} Non-Maskable Interrupts are used on X64 platforms to report fatal or recoverable platform error conditions. NMIs are reported via interrupt vector 2 on IA32 and X64 processor architecture platforms.

  • PCI Express Error (PCIe): {0xCF93C01F, 0x1A16, 0x4dfc, {0xB8, 0xBC, 0x9C, 0x4D, 0xAF, 0x67, 0xC1, 0x04}} See the PCI Express standard v1.1 for details regarding PCI Express Error Reporting. This notification type identifies errors that were reported to the system via an interrupt on a PCI Express root port.

  • INIT Record (INIT): {0xCC5263E8, 0x9308, 0x454a, {0x89, 0xD0, 0x34, 0x0B, 0xD3, 0x9B, 0xC9, 0x8E}} IPF Platforms optionally implement a mechanism (switch or button on the chassis) by which an operator may reset a system and have the system generate an INIT error record. This error record is documented in the IPF SAL specification. System software retrieves an INIT error record by querying the SAL for existing INIT records.

  • BOOT Error Record (BOOT): {0x3D61A466, 0xAB40, 0x409a, {0xA6, 0x98, 0xF3, 0x62, 0xD4, 0x64, 0xB3, 0x8F}}

    The BOOT Notification Type represents error conditions which are unhandled by system software and which result in a system shutdown/reset. System software retrieves a BOOT error record during boot by querying the platform for existing BOOT records. As an example, consider an x64 platform which implements a service processor. In some scenarios, the service processor may detect that the system is either hung or is in such a state that it cannot safely proceed without risking data corruption. In such a scenario the service processor may record some minimal error information in its system event log (SEL) and unilaterally reset the machine without notifying the OS or other system software. In such scenarios, system software is unaware of the condition that caused the system reset. A BOOT error record would contain information that describes the error condition that led to the reset so system software can log the information and use it for health monitoring.

  • DMA Remapping Error (DMAr): {0x667DD791, 0xC6B3, 0x4c27, {0x8A, 0x6B, 0x0F, 0x8E, 0x72, 0x2D, 0xEB, 0x41}} The DMA Remapping Notification Type identifies fault conditions generated by the DMAr unit when processing un-translated, translation and translated DMA requests. The fault conditions are reported to the system using a message signaled interrupt.

  • Synchronous External Abort (SEA): {0x9A78788A, 0xBBE8, 0x11E4, {0x80, 0x9E, 0x67, 0x61, 0x1E, 0x5D, 0x46, 0xB0}}

    Synchronous External Aborts represent precise processor error conditions on ARM systems (uncorrectable and/or recoverable) as described in D3.5 of the ARMv8 ARM reference manual. This notification may be triggered by one of the following scenarios: cache parity error, cache ECC error, external bus error, micro-architectural error, data poisoning, and other platform errors.

  • SError Interrupt (SEI): {0x5C284C81, 0xB0AE, 0x4E87, {0xA3, 0x22, 0xB0, 0x4C, 0x85, 0x62, 0x43, 0x23}} SError Interrupts represent asynchronous imprecise (or possibly precise) processor error conditions on ARM systems (corrected, uncorrectable, and recoverable) as described in D3.5 of the ARM ARM reference manual. This notification may be triggered by one of the following scenarios: cache parity error, cache ECC error, external bus error, micro-architectural error, data poisoning, and other platform errors.

  • Platform Error Interrupt (PEI): {0x09A9D5AC, 0x5204, 0x4214, {0x96, 0xE5, 0x94, 0x99, 0x2E, 0x75, 0x2B, 0xCD} Platform Error Interrupt represent asynchronous imprecise platform error conditions on ARM systems that may be triggered by the following scenarios: system memory ECC error, ECC errors in system cache (e.g. shared high-level caches), vendor specific chip errors, external platform errors.

  • Compute Express Link (CXL) Component: {0x69293BC9, 0x41DF, 0x49A3 {0xB4, 0xBD, 0x4F, 0xB0, 0xDB, 0x30, 0x41, 0xF6}} This Notification Type identifies errors that were reported to the system by CXL components that support error reporting via the CXL RAS Mailbox interface. See the CXL Specification, Rev 2.0 or later, for details regarding CXL Error Reporting.

N.2.1.2. Error Status

The error status definition provides the capability to abstract information from implementation-specific error registers into generic error codes.

Table N.3 Error Status Fields

Bit Position

Description

7:0

Reserved

15:8

Encoded value for the Error_Type. See Table 20 Error Types for details.

16

Address: Error was detected on the address signals or on the address portion of the transaction.

17

Control: Error was detected on the control signals or in the control portion of the transaction.

18

Data: Error was detected on the data signals or in the data portion of the transaction.

19

Responder: Error was detected by the responder of the transaction.

20

Requester: Error was detected by the requester of the transaction.

21

First Error: If multiple errors are logged for a section type, this is the first error in the chronological sequence. Setting of this bit is optional.

22

Overflow: Additional errors occurred and were not logged due to lack of logging resources.

63:23

Reserved.
Table N.4 Error Types

Encoding

Description

1

ERR_INTERNAL Error detected internal to the component.

16

ERR_BUS Error detected in the bus.

Detailed Internal Errors

4

ERR_MEM Storage error in memory (DRAM).

5

ERR_TLB Storage error in TLB.

6

ERR_CACHE Storage error in cache.

7

ERR_FUNCTION Error in one or more functional units.

8

ERR_SELFTEST component failed self test.

9

ERR_FLOW Overflow or undervalue of internal queue.

Detailed Bus Errors

17

ERR_MAP Virtual address not found on IO-TLB or IO-PDIR.

18

ERR_IMPROPER Improper access error.

19

ERR_UNIMPL Access to a memory address which is not mapped to any component.

20

ERR_LOL Loss of Lockstep

21

ERR_RESPONSE Response not associated with a request

22

ERR_PARITY Bus parity error (must also set the A, C, or D Bits).

23

ERR_PROTOCOL Detection of a protocol error.

24

ERR_ERROR Detection of a PATH_ERROR

25

ERR_TIMEOUT Bus operation timeout.

26

ERR_POISONED A read was issued to data that has been poisoned.

All Others

Reserved

N.2.2. Section Descriptor

Table N.5 Section Descriptor

Mnemonic

Byte Offset

Byte Length

Description

Section Offset

0

4

Offset in bytes of the section body from the base of the record header.

Section Length

4

4

The length in bytes of the section body.

Revision

8

2
This is a 2-byte field representing a major and minor version number for the error record definition in BCD format. The interpretation of the major and minor version number is as follows:
• Byte 0 — Minor (00): An increase in this revision indicates that changes to the headers and sections are backward compatible with software that uses earlier revisions. Addition of new GUID types, errata fixes or clarifications are covered by a bump up.
• Byte 1 — Major (01): An increase in this revision indicates that the changes are not backward compatible from a software perspective

Validation Bits

10

1
This field indicates the validity of the following fields:
• Bit 0 - If 1, the FRUId field contains valid information
• Bit 1 - If 1, the FRUString field contains valid information
• Bits 7:2 - Reserved, must be zero.

Reserved

11

1

Must be zero.

Flags

12

4
Flag field contains information that describes the error section as follows:
Bit 0 - Primary: If set, identifies the section as the section to be associated with the error condition. This allows for FRU determination and for error recovery operations. By identifying a primary section, the consumer of an error record can determine which section to focus on. It is not always possible to identify a primary section so this flag should be taken as a hint.
Bit 1 - Containment Warning: If set, the error was not contained within the processor or memory hierarchy and the error may have propagated to persistent storage or network.
Bit 2 - Reset: If set, the component has been reset and must be re-initialized or re-enabled by the operating system prior to use.
Bit 3 - Error threshold exceeded: If set, OS may choose to discontinue use of this resource.
Bit 4 - Resource not accessible: If set, the resource could not be queried for error information due to conflicts with other system software or resources. Some fields of the section will be invalid.
Bit 5 - Latent error: If set this flag indicates that action has been taken to ensure error containment (such a poisoning data), but the error has not been fully corrected and the data has not been consumed. System software may choose to take further corrective action before the data is consumed.
Bit 6 - Propagated: If set this flag indicates the section is to be associated with an error that has been propagated due to hardware poisoning. This implies the error is a symptom of another error. It is not always possible to ascertain whether this is the case for an error, therefore if the flag is not set, it is unknown whether the error was propagated. this helps determining FRU when dealing with HW failures.
Bit 7 - Overflow: If set this flag indicates the firmware has detected an overflow of buffers/queues that are used to accumulate, collect, or report errors (e.g. the error status control block exposed to the OS). When this occurs, some error records may be lost.

Bit 8 through 31 - Reserved.

Section Type

16

16
This field holds a pre-assigned GUID value indicating that it is a section of a particular error. The different error section types are as defined below:

Processor Generic
• {0x9876CCAD, 0x47B4, 0x4bdb, {0xB6, 0x5E, 0x16, 0xF1, 0x93, 0xC4, 0xF3, 0xDB}}

Processor Specific
• IA32/X 64:{0xDC3EA0B0, 0xA144, 0x4797, {0xB9, 0x5B, 0x53, 0xFA, 0x24, 0x2B, 0x6E, 0x1D}}
• IPF: {0xe429faf1, 0x3cb7, 0x11d4, {0xb, 0xca, 0x7, 0x00, 0x80, 0xc7, 0x3c, 0x88, 0x81}} (see footnote 1 at the end of Appendix N)
• ARM: { 0 xE19E3D16,0xBC1 1,0x11E4,{0x9C, 0xAA, 0xC2, 0x05, 0x1D, 0x5D, 0x46, 0xB0}}
NOTE: In addition to the types listed above, there may exist vendor specific GUIDs that describe vendor specific section types.

Platform Memory
• {0xA5BC1114, 0x6F64, 0x4EDE, {0xB8, 0x63, 0x3E, 0x83, 0xED, 0x7C, 0x83, 0xB1}}

PCIe
• {0xD995E954, 0xBBC1, 0x430F, {0xAD, 0x91, 0xB4, 0x4D, 0xCB, 0x3C, 0x6F, 0x35}}

Firmware Error Record Reference
• {0x81212A96, 0x09ED, 0x4996, {0x94, 0x71, 0x8D, 0x72, 0x9C, 0x8E, 0x69, 0xED}}

PCI/PCI-X Bus
• {0xC5753963, 0x3B84, 0x4095, {0xBF, 0x78, 0xED, 0xDA, 0xD3, 0xF9, 0xC9, 0xDD}}

PCI Component/Device
• {0xEB5E4685, 0xCA66, 0x4769, {0xB6, 0xA2, 0x26, 0x06, 0x8B, 0x00, 0x13, 0x26}}

DMAr Generic
• {0x5B51FEF7, 0xC79D, 0x4434, {0x8F, 0x1B, 0xAA,
• 0x62, 0xDE, 0x3E, 0x2C, 0x64}}

Intel® VT for Directed I/O specific DMAr section
• {0x71761D37, 0x32B2, 0x45cd, {0xA7, 0xD0, 0xB0,
• 0xFE 0xDD, 0x93, 0xE8, 0xCF}}

IOMMU specific DMAr section
• {0x036F84E1, 0x7F37, 0x428c, {0xA7, 0x9E, 0x57,
• 0x5F, 0xDF, 0xAA, 0x84, 0xEC}}

Note 1 CXL Component Events CXL Component Event Log Record in the CXL Component Events Section

FRU Id

32

16

GUID representing the FRU ID, if it exists, for the section reporting the error. The default value is zero indicating an invalid FRU ID. System software can use this to uniquely identify a physical device for tracking purposes. Association of a GUID to a physical device is done by the platform in an implemen tation-specific way (i.e., PCIe Device can lock a GUID to a PCIe Device ID).

Section Severity

48

4

This field indicates the severity associated with the error section. 0 - Recoverable (also called non-fatal uncorrected) 1 - Fatal 2 - Corrected 3 - Informational All other values are reserved. Note that severity of “Informational” indicates that the section contains extra information that can be safely ignored by error handling software.

FRU Text

52

20

ASCII string identifying the FRU hardware.

Note

  1. For an IPF processor-specific error section, the GUID listed is the value from the SAL specification. The format of the data for this section is same as the Processor Device Error Info in the SAL specification.

N.2.3. Non-standard Section Body

Information that does not conform to one the standard formats (i.e., those defined in sections 2.4 through 2.9 of this document) may be recorded in the error record in a non-standard section. The type (e.g. format) of a non-standard section is identified by the GUID populated in the Section Descriptor’s Section Type field. This allows the information to be decoded by consumers if the format is externally documented. Examples of information that might be placed in a non-standard section include the IPF raw SAL error record, Error information recorded in implementation-specific PCI configuration space, and IPMI error information recorded in an IPMI SEL.

N.2.4. Processor Error Sections

The processor error sections are divided into two different components as described below:

  1. Processor Generic Error Section: This section holds information about processor errors in a generic form and will be common across all processor architectures. An example or error information provided is the generic information of cache, tlb, etc., errors.

  2. Processor Specific Error Section: This section consists of error information, which is specific to a processor architecture. In addition, certain processor architecture state at the time of error may also be captured in this section. This section is unique to each processor architecture (Itanium Processor Family, IA32/X64, ARM).

N.2.4.1. Generic Processor Error Section

The Generic Processor Error Section describes processor reported hardware errors for logical processors in the system.

Section Type: {0x9876CCAD, 0x47B4, 0x4bdb, {0xB6, 0x5E, 0x16, 0xF1, 0x93, 0xC4, 0xF3, 0xDB}}

Table N.6 Processor Generic Error Section

Name

Byte Offset

Byte Length

Description

Validation Bits

0

8
The validation bit mask indicates whether or not each of the following fields is valid in this section.
Bit 0 - Processor Type Valid
Bit 1 - Processor ISA Valid
Bit 2 - Processor Error Type Valid
Bit 3 - Operation Valid
Bit 4 - Flags Valid
Bit 5 - Level Valid
Bit 6 - CPU Version Valid
Bit 7 - CPU Brand Info Valid
Bit 8 - CPU Id Valid
Bit 9 - Target Address Valid
Bit 10 - Requester Identifier Valid
Bit 11 - Responder Identifier Valid
Bit 12 - Instruction IP Valid
All other bits are reserved and must be zero.

Processor Type

8

1
Identifies the type of the processor architecture.
0: IA32/X64
1: IA64
2: ARM
All other values reserved.

Processor ISA

9

1
Identifies the type of the instruction set executing when the error occurred:
0: IA32
1: IA64
2: X64
3: ARM A32/T32
4: ARM A64
All other values are reserved.

Processor Error Type

10

1
Indicates the type of error that occurred:
0x00: Unknown
0x01: Cache Error
0x02: TLB Error
0x04: Bus Error
0x08: Micro-Architectural Error
All other values reserved.

Operation

11

1
Indicates the type of operation:
0: Unknown or generic
1: Data Read
2: Data Write
3: Instruction Execution
All other values reserved.

Flags

12

1
Indicates additional information about the error:
Bit 0: Restartable - If 1, program execution can be restarted reliably after the error.
Bit 1: Precise IP - If 1, the instruction IP captured is directly associated with the error.
Bit 2: Overflow - If 1, a machine check overflow occurred (a second error occurred while the results of a previous error were still in the error reporting resources).
Bit 3: Corrected - If 1, the error was corrected by hardware and/or firmware.
All other bits are reserved and must be zero.

Level

13

1

Level of the structure where the error occurred, with 0 being the lowest level of cache.

Reserved

14

2

Must be zero.

CPU Version Info

16

8
This field represents the CPU Version Information and returns Family, Model, and stepping information (e.g. As provided by CPUID instruction with EAX=1 input with output values from EAX on the IA32/X64 processor or as provided by CPUID Register 3 register - Version Information on IA64 processors).

On ARM processors, this field will be provided as:
Bits 127:64 - Reserved and must be zero
Bits 63:0 - MIDR_EL1 of the processor

CPU Brand String

24

128
This field represents the null-terminated ASCII Processor Brand String (e.g. As provided by the CPUID instruction with EAX=0x80000002 and ECX=0x80000003 for IA32/X64 processors or the return from PAL_BRAND_INFO for IA64 processors).
This field is optional for ARM processors.

Processor ID

152

8
This value uniquely identifies the logical processor (e.g. As programmed into the local APIC ID register on IA32/X64 processors or programmed into the LID register on IA64 processors).

On ARM processors, this field will be provided as programmed in the architected MPIDR_EL1.

Target Address

160

8

Identifies the target address associated with the error.

Requestor Identifier

168

8

Identifies the requestor associated with the error.

Responder Identifier

176

8

Identifies the responder associated with the error.

Instruction IP

184

8

Identifies the instruction pointer when the error occurred.

N.2.4.2. IA32/X64 Processor Error Section

Type:{0xDC3EA0B0, 0xA144, 0x4797, {0xB9, 0x5B, 0x53, 0xFA, 0x24, 0x2B, 0x6E, 0x1D}}

Table N.7 Processor Error Record

Mnemonic

Byte Offset

Byte Length

Description

Validation Bits

0

8
The validation bit mask indicates each of the following field is valid in this section:

Bit0 - LocalAPIC_ID Valid
Bit1 - CPUID Info Valid
Bits 2-7 - Number of Processor Error Information Structure
(PRO C_ERR_INFO_NUM)
Bit 8- 13 Number of Processor Context Information Structure
(PROC_CO NTEXT_INFO_NUM)
Bits 14-63 - Reserved

Local APIC_ID

8

8

This is the processor APIC ID programmed into the APIC ID registers.

CPUID Info

16

48

This field represents the CPU ID structure of 48 bytes and returns Model, Family, and stepping information as provided by the CPUID instruction with EAX=1 input and output values from EAX, EBX, ECX, and EDX null extended to 64-bits.

Processor Error Info

64

Nx64

This is a variable-length structure consisting of N different 64 byte structures, each representing a single processor error information structure. The value of N ranges from 0-63 and is as indicated by PRO C_ERR_INFO_NUM.

Processor Context

64+Nx64

NxX

This is a variable size field providing the information for the processor context state such as MC Bank MSRs and general registers. The value of N ranges from 0-63 and is as indicated by PROC_CO NTEXT_INFO_NUM. Each processor context information structure is padded with zeros if the size is not a multiple of 16 bytes.
N.2.4.2.1. IA32/X64 Processor Error Information Structure

As described above, the processor error section contains a collection of structures called Processor Error Information Structures that contain processor structure specific error information. This section details the layout of the Processor Error Information Structure and the detailed check information which is contained within.

Table N.8 IA32/X64 Processor Error Information Structure

Mnemonic

Byte Offset

Byte Length

Description

Error Structure Type

0

16
This field holds a pre-assigned GUID indicating the type of Processor Error Information structure. The following Processor Error Information Structure Types have pre-defined GUID.
• Cache Error Information (Cache Check)
• TLB Error Information (TLB Check)
• Bus Error Information (Bus Check)
• Micro-architecture Specific Error Information (MS Check)

Validation Bits

16

8
Bit 0 - Check Info Valid
Bit 1 - Target Address Identifier Valid
Bit 2 - Requestor Identifier Valid
Bit 3 - Responder Identifier Valid
Bit 4 - Instruction Pointer Valid
Bits 5-63 - Reserved

Check Information

24

8

Str uctureErrorType specific error check structure.

Target Identifier

32

8

Identifies the target associated with the error.

Requestor Identifier

40

8

Identifies the requestor associated with the error.

Responder Identifier

48

8

Identifies the responder associated with the error.

Instruction Pointer

56

8

Identifies the instruction executing when the error occurred.
N.2.4.2.2. IA32/X64 Cache Check Structure

Type:{0xA55701F5, 0xE3EF, 0x43de, {0xAC, 0x72, 0x24, 0x9B, 0x57, 0x3F, 0xAD, 0x2C}}

Table N.9 IA32/X64 Cache Check Structure

Field Name

Bits

Description

ValidationBits

15:0
Indicates which fields in the Cache Check structure are valid:
Bit 0 - Transaction Type Valid
Bit 1 - Operation Valid
Bit 2 - Level Valid
Bit 3 - Processor Context Corrupt Valid
Bit 4 - Uncorrected Valid
Bit 5 - Precise IP Valid
Bit 6 - Restartable Valid
Bit 7- Overflow Valid
Bits 8 - 15 Reserved

TransactionType

17:16
Type of cache error:
0 - Instruction
1 - Data Access
2 - Generic
All other values are reserved

Operation

21:18
Type of cache operation that caused the error:
0 - generic error (type of error cannot be determined)
1 - generic read (type of instruction or data request cannot be determined)
2 - generic write (type of instruction or data request cannot be determined)
3 - data read
4 - data write
5 - instruction fetch
6 - prefetch
7 - eviction
8 - snoop
All other values are reserved.

Level

24:22

Cache Level

Processor Context Corrupt

25
This field indicates that the processor context might have been corrupted.
0 - Processor context not corrupted
1 - Processor context corrupted

Uncorrected

26
This field indicates whether the error was corrected or uncorrected:
0: Corrected
1: Uncorrected

Precise IP

27

This field indicates that the instruction pointer pushed onto the stack is directly associated with the error

Restartable IP

28

This field indicates that program execution can be restarted reliably at the instruction pointer pushed onto the stack

Overflow

29
This field indicates an error overflow occurred
0 - Overflow not occurred
1 - Overflow occurred

63:30

Reserved
N.2.4.2.3. IA32/X64 TLB Check Structure

Type:{0xFC06B535, 0x5E1F, 0x4562, {0x9F, 0x25, 0x0A, 0x3B, 0x9A, 0xDB, 0x63, 0xC3}}

Table N.10 IA32/X64 TLB Check Structure

Field Name

Bits

Description

Validation Bits

15:0
Indicate which fields in the Cache_Check structure are valid
Bit 0 - Transaction Type Valid
Bit 1 - Operation Valid
Bit 2 - Level Valid
Bit 3 - Processor Context Corrupt Valid
Bit 4 - Uncorrected Valid
Bit 5 - Precise IP Valid
Bit 6 - Restartable IP Valid
Bit 7 - Overflow Valid
Bit 8 - 15 Reserved

Transaction Type

17:16
Type of TLB error
0 - Instruction
1 - Data Access
2 - Generic
All other values are reserved

Operation

21:18
Type of TLB access operation that caused the machine check:
0 - generic error (type of error cannot be determined)
1 - generic read (type of instruction or data request cannot be determined)
2 - generic write (type of instruction or data request cannot be determined)
3 - data read
4 - data write
5 - instruction fetch
6 - prefetch
All other values are reserved.

Level

24:22

TLB Level

Processor Context Corrupt

25
This field indicates that the processor context might have been corrupted.
0 - Processor context not corrupted
1 - Processor context corrupted

Uncorrected

26
This field indicates whether the error was corrected or uncorrected:
0: Corrected
1: Uncorrected

PreciseIP

27

This field indicates that the instruction pointer pushed onto the stack is directly associated with the error.

Restartable IP

28

This field indicates the program execution can be restarted reliably at the instruction pointer pushed onto the stack.

Overflow

29
This field indicates an error overflow occurred
0 - Overflow not occurred
1 - Overflow occurred

63:30

Reserved
N.2.4.2.4. IA32/X64 Bus Check Structure

Type:{0x1CF3F8B3, 0xC5B1, 0x49a2, {0xAA, 0x59, 0x5E, 0xEF, 0x92, 0xFF, 0xA6, 0x3C}}

Table N.11 IA32/X64 Bus Check Structure

Field Name

Bits

Description

Validation Bits

15:0
Indicate which fields in the Bus_Check structure are valid
Bit 0 - Transaction Type Valid
Bit 1 - Operation Valid
Bit 2 - Level Valid
Bit 3 - Processor Context Corrupt Valid
Bit 4 - Uncorrected Valid
Bit 5 - Precise IP Valid
Bit 6 - Restartable IP Valid
Bit 7 - Overflow Valid
Bit 8 - Participation Type Valid
Bit 9 - Time Out Valid
Bit 10 - Address Space Valid
Bit 11 - 15 Reserved

Transaction Type

17:16
Type of Bus error
0 - Instruction
1 - Data Access
2 - Generic
All other values are reserved

Operation

21:18
Type of bus access operation that caused the machine check:
0 - generic error (type of error cannot be determined)
1 - generic read (type of instruction or data request cannot be determined)
2 - generic write (type of instruction or data request cannot be determined)
3 - data read
4 - data write
5 - instruction fetch
6 - prefetch All other values are reserved.

Level

24:22

Indicate which level of the bus hierarchy the error occurred in.

Processor Context Corrupt

25
This field indicates that the processor context might have been corrupted.
0 - Processor context not corrupted
1 - Processor context corrupted

Uncorrected

26
This field indicates whether the error was corrected or uncorrected:
0: Corrected
1: Uncorrected

PreciseIP

27

This field indicates that the instruction pointer pushed onto the stack is directly associated with the error.

Restartable IP

28

This field indicates the program execution can be restarted reliably at the instruction pointer pushed onto the stack.

Overflow

29
This field indicates an error overflow occurred
0 - Overflow not occurred
1 - Overflow occurred

Participation Type

31:30
Type of Participation
0 - Local Processor originated request
1 - Local processor Responded to request
2 - Local processor Observed
3 - Generic

Time Out

32

This field indicates that the request timed out.

Address Space

34:33
0 - Memory Access
1 - Reserved
2 - I/O
3 - Other Transaction

63:35

Reserved
N.2.4.2.5. IA32/X64 MS Check Field Description

Type: {0x48AB7F57, 0xDC34, 0x4f6c, {0xA7, 0xD3, 0xB0, 0xB5, 0xB0, 0xA7, 0x43, 0x14}}

Table N.12 IA32/X64 MS Check Field Description

Field Name

Bits

Description

Validation Bits

15:0
Indicate which fields in the Cache_Check structure are valid
Bit 0 - Error Type Valid
Bit 1 - Processor Context Corrupt Valid
Bit 2 - Uncorrected Valid
Bit 3 - Precise IP Valid
Bit 4 - Restartable IP Valid
Bit 5 - Overflow Valid
Bit 6 - 15 Reserved

Error Type

18:16
Identifies the operation that caused the error:
0 - No Error
1 - Unclassified
2 - Microcode ROM Parity Error
3 - External Error
4 - FRC Error
5 - Internal Unclassified
All other value are processor specific.

Processor Context Corrupt

19
This field indicates that the processor context might have been corrupted.
0 - Processor context not corrupted
1 - Processor context corrupted

Uncorrected

20
This field indicates whether the error was corrected or uncorrected:
0: Corrected
1: Uncorrected

Precise IP

21

This field indicates that the instruction pointer pushed onto the stack is directly associated with the error.

Restartable IP

22

This field indicates the program execution can be restarted reliably at the instruction pointer pushed onto the stack.

Overflow

23
This field indicates an error overflow occurred
0 - Overflow not occurred
1 - Overflow occurred

63:24

Reserved
N.2.4.2.6. IA32/X64 Processor Context Information Structure

As described above, the processor error section contains a collection of structures called Processor Context Information that contain processor context state specific to the IA32/X64 processor architecture. This section details the layout of the Processor Context Information Structure and the detailed processor context type information.

Table N.13 IA32/X64 Processor Context Information

Mnemonic

Byte Offset

Byte Length

Description

Register Context Type

0

2 bytes
Value indicating the type of processor context state being reported:
0 - Unclassified Data
1 - MSR Registers (Machine Check and other MSRs)
2 - 32-bit Mode Execution Context
3 - 64-bit Mode Execution Context
4 - FXSAVE Context
5 - 32-bit Mode Debug Registers (DR0-DR7)
6 - 64-bit Mode Debug Registers (DR0-DR7)
7 - Memory Mapped Registers
Others - Reserved

Register Array Size

2

2 bytes

Represents the total size of the array for the Data Type being reported in bytes.

MSR Address

4

4 bytes

This field contains the starting MSR address for the type 1 register context.

MM Register Address

8

8 bytes

This field contains the starting memory address for the type 7 register context.

Register Array

16

N bytes

This field will provide the contents of the actual registers or raw data. The number of Registers or size of the raw data reported is determined by (Array Size / 8) or otherwise specified by the context structure type definition.

The Table below shows the register context type 2, 32-bit mode execution context.

Table N.14 IA32 Register State

Offset

Length

Field

0

4 bytes

EAX

4

4 bytes

EBX

8

4 bytes

ECX

12

4 bytes

EDX

16

4 bytes

ESI

20

4 bytes

EDI

24

4 bytes

EBP

28

4 bytes

ESP

32

2 bytes

CS

34

2 bytes

DS

36

2 bytes

SS

38

2 bytes

ES

40

2 bytes

FS

42

2 bytes

GS

44

4 bytes

EFLAGS

48

4 bytes

EIP

52

4 bytes

CR0

56

4 bytes

CR1

60

4 bytes

CR2

64

4 bytes

CR3

68

4 bytes

CR4

72

8 bytes

GDTR

80

8 bytes

IDTR

88

2 bytes

LDTR

90

2 bytes

TR

See the Table below for the register context type 3, 64-bit mode execution context.

Table N.15 X64 Register State

Offset

Length

Field

0

8 bytes

RAX

8

8 bytes

RBX

16

8 bytes

RCX

24

8 bytes

RDX

32

8 bytes

RSI

40

8 bytes

RDI

48

8 bytes

RBP

56

8 bytes

RSP

64

8 bytes

R8

72

8 bytes

R9

80

8 bytes

R10

88

8 bytes

R11

96

8 bytes

R12

104

8 bytes

R13

112

8 bytes

R14

120

8 bytes

R15

128

2 bytes

CS

130

2 bytes

DS

132

2 bytes

SS

134

2 bytes

ES

136

2 bytes

FS

138

2 bytes

GS

140

4 bytes

Reserved

144

8 bytes

RFLAGS

152

8 bytes

EIP

160

8 bytes

CR0

168

8 bytes

CR1

176

8 bytes

CR2

184

8 bytes

CR3

192

8 bytes

CR4

200

8 bytes

CR8

208

16 bytes

GDTR

224

16 bytes

IDTR

240

2 bytes

LDTR

242

2 bytes

TR

N.2.4.3. IA64 Processor Error Section

Refer to the Intel Itanium Processor Family System Abstraction Layer specification for finding the IA64 specific error section body definition.

N.2.4.4. ARM Processor Error Section

Type: {0xE19E3D16, 0xBC11, 0x11E4, {0x9C, 0xAA, 0xC2, 0x05, 0x1D, 0x5D, 0x46, 0xB0}}

The ARM Processor Error Section may contain multiple instances of error information structures associated to a single error event. An error may propagate to other hardware components (e.g. poisoned data) or cause subsequent errors, all of which may be captured in a single ARM processor error section. The processor context information describes the observed state of the processor at the point of error detection.

It is optional for vendors to capture processor context information. The specifics of capturing processor context is vendor specific. Vendors must take care when handling errors that have originated whilst a processor was executing in a secure exception level. In those cases providing processor context information to non-secure agents could be unsafe and lead to security attacks.

Table N.16 ARM Processor Error Section

Mnemonic

Byte Offset

Byte Length

Description

Validation Bit

0

4
The validation bit mask indicates whether or not each of the following fields is valid in this section.
Bit 0 - MPIDR Valid
Bit 1 - Error affinity level Valid
Bit 2 - Running State
Bit 3 - Vendor Specific Info Valid
All other bits are reserved and must be zero.

ERR_INFO_NUM

4

2

ERR_INFO_NUM is the number of Processor Error Information Structures (must be 1 or greater)

CONTEXT_INFO_NUM

6

2

C ONTEXT_INFO_NUM is the number of Context Information Structures

Section Length

8

4

This describes the total size of the ARM processor error section

Error affinity level

12

1
For errors that can be attributed to a specific affinity level, this field defines the affinity level at which the error was produced, detected, and/or consumed. This is a value between 0 and 3. All other values (4-255) are reserved

For example, a vendor may choose to define affinity levels as follows:
Level 0: errors that can be precisely attributed to a specific CPU
(e.g. due to a synchronous external abort)
Level 1: Cache parity and/or ECC errors detected at cache of affinity level 1 (e.g. only attributed to higher level cache due to prefetching and/or error propagation)

NOTE: Detailed meanings and groupings of affinity level are chip and/or platform specific. The affinity level described here must be consistent with the platform definitions used MPIDR. For cache/TLB errors, the cache/TLB level is provided by the cache/TLB error structure, which may differ from affinity level.

Reserved

13

3

Must be zero

MPIDR_EL1

16

8

This field is valid for “attributable errors” that can be attributed to a specific CPU, cache, or cluster. This is the processor’s unique ID in the system.

MIDR_EL1

24

8

This field provides identification information of the chip, including an implementer code for the device and a device ID number

Running State

32

4

Bit 0 - Processor running. If this bit is set, “PSCI State” field must be zero. All other bits are reserved and must be zero.

PSCI State

36

4

This field provides PSCI state of the processor, as defined in ARM PSCI document. This field is valid when bit 32 of “Running State” field is zero.

Processor Error Information Structure

40

Nx32

This is a variable-length structure consisting of N different 32 byte structures (reference the Table below, ARM Processor Error Information Structure) , each representing a single processor error information structure. The value of N ranges from 1-255 and is as indicated by ERR_INFO_NUM field in this table.

Processor Context

40 + Nx32

MxP

This is a variable size field consisting of M different P byte structures providing the information for the processor context state such as general purpose registers (GPRs) and special purpose registers (SPRs) as defined in Table 266 or 267 (depending on the context type). The value of M ranges from 0-65536 and is indicated by the C ONTEXT_INFO_NUM field in this table. Each processor context information structure is padded with zeros if the size is not a multiple of 16 bytes. The value of P is a variable length defined by the processor context structure per Table 266 and 267.

Vendor Specific Error Info

40 + Nx32 + MxP

vendor specific

This is an optional variable field provided by vendors that prefer to provide additional details.
N.2.4.4.1. ARM Processor Error Information

As described above, the processor error section contains a collection of Processor Error Information structures that contain processor specific error information. This section details the layout of the Processor Error Information structure and the detailed information which is contained within.

Table N.17 ARM Processor Error Information Structure

Mnemonic

Byte Offset

Byte Length

Description

Version

0

1

0 (revision of this table)

Length

1

1

32 (length in bytes)

Validation Bit

2

2
The validation bit mask indicates whether or not each of the following fields is valid in this section.
Bit 0 - Multiple Error (Error Count) Valid
Bit 1 - Flags Valid
Bit 2 - Error Information Valid
Bit 3 - Virtual Fault Address
Bit 4 - Physical Fault Address
All other bits are reserved and must be zero.

Type

4

1
Bit 1 - Cache Error
Bit 2 - TLB Error
Bit 3 - Bus Error
Bit 4 - Micro-architectural Error
All other values are reserved

Multiple Error (Error Count)

5

2
This field indicates whether multiple errors have occurred. In the case of multiple error with a valid count, this field will specify the error count. The value of this field is defined as follows:
0: Single Error
1: Multiple Errors
2-65535: Error Count (if known)

Flags

7

1
This field indicates flags that describe the error attributes. The value of this field is defined as follows:
Bit 0 - First error captured
Bit 1 - Last error captured
Bit 2 - Propagated
Bit 3 - Overflow
All other bits are reserved and must be zero
Note: The overflow bit indicates when firmware/hardware error buffers experience an overflow, so it is possible that some error information has been lost.

Error Information

8

8

The error information structure is specific to each error type (described in tables below)

Virtual Fault Address

16

8

If known, this field indicates a virtual fault address associated with the error (e.g. when an error occurs in virtually indexed cache)

Physical Fault Address

24

8

If known, this field indicates a physical fault address associated with the error

See the 4 Tables directly below for more error information: Arm Cache Error Structure, ARM TLB Error Structure, ARM Bus Error Structure, and ARM Processor Error Context Information Header Structure .

Table N.18 ARM Cache Error Structure

Name

Bits

Description

Validation Bit

15:0
Indicates which fields in the Cache Check structure are valid:
Bit 0 - Transaction Type Valid
Bit 1 - Operation Valid
Bit 2 - Level Valid
Bit 3 - Processor Context Corrupt Valid
Bit 4 - Corrected Valid
Bit 5 - Precise PC Valid
Bit 6 - Restartable PC Valid
All other bits are reserved and must be zero.

Transaction Type

17:16
Type of cache error:
0 - Instruction
1 - Data Access
2 - Generic
All other values are reserved

Operation

21:18
Type of cache operation that caused the error:
0 - generic error (type of error cannot be determined)
1 - generic read (type of instruction or data request cannot be determined)
2 - generic write (type of instruction or data request cannot be determined)
3 - data read
4 - data write
5 - instruction fetch
6 - prefetch
7 - eviction
8 - snooping (the processor described in this record initiated a cache snoop that resulted in an error)
9 - snooped (The processor described in this record raised a cache error caused by another processor or device snooping into its cache)
10 - management
All other values are reserved.

Level

24:22

Cache level

Processor Context Corrupt

25
This field indicates that the processor context might have been corrupted.
0 - Processor context not corrupted
1 - Processor context corrupted

Corrected

26
This field indicates whether the error was corrected or uncorrected:
1: Corrected
0: Uncorrected

Precise PC

27

This field indicates that the program counter that is directly associated with the error

Restartable PC

28

This field indicates that program execution can be restarted reliably at the PC associated with the error.

Reserved

63:29

Must be zero
Table N.19 ARM TLB Error Structure

Name

Bits

Description

Validation Bit

15:0

Indicates which fields in the TLB error structure are valid: | Bit 0 - Transaction Type Valid | Bit 1 - Operation Valid | Bit 2 - Level Valid | Bit 3 - Processor Context Corrupt Valid | Bit 4 - Corrected Valid | Bit 5 - Precise PC Valid | Bit 6 - Restartable PC Valid | All other bits are reserved and must be zero.

Transaction Type

17:16

Type of TLB error: 0 - Instruction 1 - Data Access 2 - Generic All other values are reserved

Operation

21:18
Type of TLB operation that caused the error:
0 - generic error (type of error cannot be determined)
1 - generic read (type of instruction or data request cannot be determined)
2 - generic write (type of instruction or data request cannot be determined)
3 - data read
4 - data write
5 - instruction fetch
6 - prefetch
7 - local management operation (the processor described in this record initiated a TLB management operation that resulted in an error)
8 - external management operation (the processor described in this record raised a TLB error caused by another processor or device broadcasting TLB operations)
All other values are reserved.

Level

24:22

TLB level

Processor Context Corrupt

25
This field indicates that the processor context might have been corrupted.
0 - Processor context not corrupted
1 - Processor context corrupted

Corrected

26

This field indicates whether the error was corrected or uncorrected: 1: Corrected 0: Uncorrected

Precise PC

27

This field indicates that the program counter that is directly associated with the error

Restartable PC

28

This field indicates that program execution can be restarted reliably at the PC associated with the error.

Reserved

63:29

Must be zero.
Table N.20 ARM Bus Error Structure

Name

Bits

Description

Validation Bit

15:0
Indicates which fields in the Bus error structure are valid: Bit
0 - Transaction Type Valid
Bit 1 - Operation Valid
Bit 2 - Level Valid
Bit 3 - Processor Context Corrupt Valid
Bit 4 - Corrected Valid
Bit 5 - Precise PC Valid
Bit 6 - Restartable PC Valid
Bit 7 - Participation Type Valid
Bit 8 - Time Out Valid
Bit 9 - Address Space Valid
Bit 10 - Memory Attributes Valid
Bit 11 - Access Mode valid
All other bits are reserved and must be zero.

Transaction Type

17:16
Type of bus error:
0 - Instruction
1 - Data Access
2 - Generic
All other values are reserved

Operation

21:18
Type of bus operation that caused the error:
0 - generic error (type of error cannot be determined)
1 - generic read (type of instruction or data request cannot be determined)
2 - generic write (type of instruction or data request cannot be determined)
3 - data read
4 - data write
5 - instruction fetch
6 - prefetch
All other values are reserved.

Level

24:22

Affinity level at which the bus error occurred

Processor Context Corrupt

25
This field indicates that the processor context might have been corrupted.
0 - Processor context not corrupted
1 - Processor context corrupted

Corrected

26

This field indicates whether the error was corrected or uncorrected: 1: Corrected 0: Uncorrected

Precise PC

27

This field indicates that the program counter that is directly associated with the error

Restartable PC

28

This field indicates that program execution can be restarted reliably at the PC associated with the error.

Participation Type

30:29
Type of Participation
0 - Local Processor originated request
1 - Local processor Responded to request
2 - Local processor Observed
3 - Generic

The usage of this field depends on the vendor, but the examples below provide some guidance on how this field is to be used:
If bus error occurs on an LDR instruction, the local processor originated the request.
If the bus error occurs due to a snoop operation, local processor responded to the request
If a bus error occurs due to cache prefetching and an SEI was sent to a particular CPU to notify this bus error has occurred, then the local processor only observed the error.

Time Out

31

This field indicates that the request timed out.

Address Space

33:32
0 - External Memory Access (e.g. DDR)
1 - Internal Memory Access (e.g. internal chip ROM)
3 - Device Memory Access

Memory Access Attributes

42:34

Memory attribute as described in the ARM ARM specification.

Access Mode

43
Indicates whether the access was a secure or normal bus request
0 - secure
1 - normal

Note: A platform may choose to hide some or all of the error information for errors that are consumed/detected in the secure context.

Reserved

63:44

Must be zero.
N.2.4.4.1.1. ARM Vendor Specific Micro-Architecture ErrorStructure

This is a vendor specific structure. Please refer to your hardware vendor documentation for the format of this structure.

N.2.4.4.2. ARM Processor Context Information

As described above, the processor error section contains a collection of structures called Processor Context Information. These provide processor context state specific to the ARM processor architecture. This section details the layout of the Processor Error Context Information Header Structure ( See Table N-21, ARM Processor Error Context Information HeaderStructure ) and the detailed processor context type information structures ( See Table N-21 through Table N-30).

Care must be taken when reporting context information structures. The amount of context reported depends on the agent that is going to observe the data. The following are recommended guidelines:

  1. If the error happens whilst the processor is in the secure world, EL3, Secure EL1 or secure EL0, context information can contain sensitive data, and should not be exposed to unauthorized parties.

  2. If the error information is being provided to a software agent running at EL2, then the context information should only include any registers visible in EL2, e.g. GPR, EL1 and EL2 registers.

  3. If the error information is being provided to a software agent running at EL1, then the context information should only include any registers visible in EL1, e.g. GPR, EL1 and registers.

For context information on processor running in AArch64 mode, even though some registers are defined as 4 bytes in length, following tables provide 8 bytes space to account for possible future expansion.

Table N.21 ARM Processor Error Context Information HeaderStructure

Name

Byte Offset

Byte Length

Description

Version

0

2

0 (revision of this table)

Register Context Type

2

2
Value indicating the type of processor context state being reported:
0 — AArch32 GPRs (General Purpose Registers).
1 — AArch32 EL1 context registers
2 — AArch32 EL2 context registers
3 — Aarch32 secure context registers
4 — AArch64 GPRs
5 — AArch64 EL1 context registers
6 — Aarch64 EL2 context registers
7 — AArch64 EL3 context registers
8 — Misc. System Register Structure
All other values are reserved.

Register Array Size

4

4

Represents the total size of the array for the Data Type being reported in bytes.

Register Array

8

N

This field will provide the contents of the actual registers or raw data. The contents of the array depends on the Type, with the structures described in Tables 266 - 274.
Table N.22 ARMv8 AArch32 GPRs (Type 0)

Byte Offset

Byte Length

Field

0

4

R0

4

4

R1

8

4

R2

12

4

R3

16

4

R4

20

4

R5

24

4

R6

28

4

R7

32

4

R8

36

4

R9

40

4

R10

44

4

R11

48

4

R12

52

4

R13 (SP)

56

4

R14 (LR)

60

4

R15 (PC)
Table N.23 ARM AArch32 EL1 Context System Registers (Type 1)

Byte Offset

Byte Length

Field

0

4

DFAR

4

4

DFSR

8

4

IFAR

12

4

ISR

16

4

MAIR0

20

4

MAIR1

24

4

MIDR

28

4

MPIDR

32

4

NMRR

36

4

PRRR

40

4

SCTLR (NS)

44

4

SPSR

48

4

SPSR_abt

52

4

SPSR_fiq

56

4

SPSR_irq

60

4

SPSR_svc

64

4

SPSR_und

68

4

TPIDRPRW

72

4

TPIDRURO

76

4

TPIDRURW

80

4

TTBCR

84

4

TTBR0

88

4

TTBR1

92

4

DACR
Table N.24 ARM AArch32 EL2 Context System Registers (Type 2)

Byte Offset

Byte Length

Field

0

4

ELR_hyp

4

4

HAMAIR0

8

4

HAMAIR1

12

4

HCR

16

4

HCR2

20

4

HDFAR

24

4

HIFAR

28

4

HPFAR

32

4

HSR

36

4

HTCR

40

4

HTPIDR

44

4

HTTBR

48

4

SPSR_hyp

52

4

VTCR

56

4

VTTBR

60

4

DACR32_EL2
Table N.25 ARM AArch32 secure Context System Registers (Type3)

Byte Offset

Byte Length

Field

0

4

SCTLR (S)

4

4

SPSR_mon
Table N.26 ARMv8 AArch64 GPRs (Type 4)

Byte Offset

Byte Length

Field

0

8

X0

8

8

X1

16

8

X2

24

8

X3

32

8

X4

40

8

X5

48

8

X6

56

8

X7

64

8

X8

72

8

X9

80

8

X10

88

8

X11

96

8

X12

104

8

X13

112

8

X14

120

8

X15

128

8

X16

136

8

X17

144

8

X18

152

8

X19

160

8

X20

168

8

X21

176

8

X22

184

8

X23

192

8

X24

200

8

X25

208

8

X26

216

8

X27

224

8

X28

232

8

X29

240

8

X30

248

8

SP
Table N.27 ARM AArch64 EL1 Context System Registers (Type 5)

Byte Offset

Byte Length

Field

0

8

ELR_EL1

8

8

ESR_EL1

16

8

FAR_EL1

24

8

ISR_EL1

32

8

MAIR_EL1

40

8

MIDR_EL1

48

8

MPIDR_EL1

56

8

SCTLR_EL1

64

8

SP_EL0

72

8

SP_EL1

80

8

SPSR_EL1

88

8

TCR_EL1

96

8

TPIDR_EL0

104

8

TPIDR_EL1

112

8

TPIDRRO_EL0

120

8

TTBR0_EL1

128

8

TTBR1_EL1
Table N.28 ARM AArch64 EL2 Context System Registers (Type 6)

Byte Offset

Byte Length

Field

0

8

ELR_EL2

8

8

ESR_EL2

16

8

FAR_EL2

24

8

HACR_EL2

32

8

HCR_EL2

40

8

HPFAR_EL2

48

8

MAIR_EL2

56

8

SCTLR_EL2

64

8

SP_EL2

72

8

SPSR_EL2

80

8

TCR_EL2

88

8

TPIDR_EL2

96

8

TTBR0_EL2

104

8

VTCR_EL2

112

8

VTTBR_EL2
Table N.29 ARM AArch64 EL3 Context System Registers (Type 7)

Byte Offset

Byte Length

Field

0

8

ELR_EL3

8

8

ESR_EL3

16

8

FAR_EL3

24

8

MAIR_EL3

32

8

SCTLR_EL3

40

8

SP_EL3

48

8

SPSR_EL3

56

8

TCR_EL3

64

8

TPIDR_EL3

72

8

TTBR0_EL3

The following structure (Table 275) describes additional AArch64/AArch32 miscellaneous system registers captured from the perspective of the processor that took the hardware error exception. Each register array entry will be per the following table. The number of register entries present in the register array is based on the register array size (i.e. N/10).

Table N.30 ARM Misc. Context System Register (Type 8) - SingleRegister Entry

Name

Byte Offset

Byte Length

Description

MRS encoding

0

2
This field defines MRS instruction encoding.
Bit 0:2 – Op2
Bit 3:6 - CRm
Bit 7:10 - CRn
Bit 11:13 - Op1
Bit 14 - O0

Value

2

8

Value read from system register

N.2.5. Memory Error Section

Type: {0xA5BC1114, 0x6F64, 0x4EDE, {0xB8, 0x63, 0x3E, 0x83, 0xED, 0x7C, 0x83, 0xB1}}

Table N.31 Memory Error Record

Mnemonic

Byte Offset

Byte Length

Description

Validation Bits

0

8
Indicates which fields in the memory error record are valid.
Bit 0 - Error Status Valid
Bit 1 - Physical Address Valid
Bit 2 - Physical Address Mask Valid
Bit 3 - Node Valid
Bit 4 - Card Valid
Bit 5 - Module Valid
Bit 6 - Bank Valid (When Bank is addressed via group/address, refer to Bit 19 and 20)
Bit 7 - Device Valid
Bit 8 - Row Valid
1 - the Row field at Offset 42 contains row number (15:0) and row number (17:16) are 00b
0 - the Row field at Offset 42 is not used, or is defined by
Bit 18 (Extended Row Bit 16 and 17 Valid).
Bit 9 - Column Valid
Bit 10 - Bit Position Valid
Bit 11 - Platform Requestor Id Valid
Bit 12 - Platform Responder Id Valid
Bit 13 - Memory Platform Target Valid
Bit 14 - Memory Error Type Valid
Bit 15 - Rank Number Valid
Bit 16 - Card Handle Valid
Bit 17 - Module Handle Valid
Bit 18 - Extended Row Bit 16 and 17 Valid (refer to Byte Offset 42 and 73 below)
1 - the Row field at Offset 42 contains row number (15:0) and the Extended field at Offset 73 contains row number (17:16)
0 - the Extended field at Offset 73 and the Row field at Offset 42 are not used, or the Rowfield at Offset 42 is defined by Bit 8 (Row Valid). When this bit is set to 1, Bit 8 (Row Valid) must be set to 0.
Bit 19 - Bank Group Valid
Bit 20 - Bank Address Valid
Bit 21 - Chip Identification Valid
Bit 22-63 Reserved

Error Status

8

8

Memory error status information. See See Error Status for error status details.

Physical Address

16

8

The physical address at which the memory error occurred.

Physical Address Mask

24

8

Defines the valid address bits in the Physical Address field. The mask specifies the granularity of the physical address which is dependent on the hw/ implementation factors such as interleaving.

Node

32

2

In a multi-node system, this value identifies the node containing the memory in error.

Card

34

2

The card number of the memory error location.

Module

36

2

The module or rank number of the memory error location. (NODE, CARD, and MODULE should provide the information necessary to identify the failing FRU).

Bank

38

2
The bank number of the memory associated with the error.
When Bank is addressed via group/address
Bit 7:0 - Bank Address
Bit 15:8 - Bank Group

Device

40

2

The device number of the memory associated with the error.

Row

42

2

First 16 bits (15:0) of the row number of the memory error location. This field is valid if either “Row Valid” or “Extended Row Bit 16 and 17” Validation Bits at Offset 0 is set to 1..

Column

44

2

The column number of the memory error location.

Bit Position

46

2

The bit position at which the memory error occurred.

Requestor ID

48

8

Hardware address of the device that initiated the transaction that took the error.

Responder ID

56

8

Hardware address of the device that responded to the transaction.

Target ID

64

8

Hardware address of the intended target of the transaction.

Memory Error Type

72

1
Identifies the type of error that occurred:
0 - Unknown
1 - No error
2 - Single-bit ECC
3 - Multi-bit ECC
4 - Single-symbol ChipKill ECC
5 - Multi-symbol ChipKill ECC
6 - Master abort
7 - Target abort
8 - Parity Error
9 - Watchdog timeout
10 - Invalid address
11 - Mirror Broken
12 - Memory Sparing
13 - Scrub corrected error
14 - Scrub uncorrected error
15 - Physical Memory Map-out event
All other values reserved.

Extended

73

1
Bit 0 - Bit 16 of the row number of the memory error location.
- This field is valid if “Extended Row Bit 16 and 17” Validation Bits at Offset 0 is set to 1.
Bit 1 - Bit 17 of the row number of the memory error location.
- This field is valid if “Extended Row Bit 16 and 17” Validation Bits at Offset 0 is set to 1.
Bit 4:2 - Reserved
Bit 7:5 - Chip Identification.

Rank Number

74

2

The Rank number of the memory error location.

Card Handle

76

2

If bit 16 in Validation Bits is 1, this field contains the SMBIOS handle for the Type 16 Memory Array Structure that represents the memory card.

Module Handle

78

2

If bit 17 in Validation Bits is 1, this field contains the SMBIOS handle for the Type 17 Memory Device Structure that represents the Memory Module.

N.2.6. Memory Error Section 2

Type: { 0x61EC04FC, 0x48E6, 0xD813, { 0x25, 0xC9, 0x8D, 0xAA, 0x44, 0x75, 0x0B, 0x12 } };

Table N.32 Memory Error Record 2

Mnemonic

Byte Offset

Byte Length

Description

Validation Bits

0

8
Indicates which fields in the memory error record are valid.
Bit 0 - Error Status Valid
Bit 1 - Physical Address Valid
Bit 2 - Physical Address Mask Valid
Bit 3 - Node Valid
Bit 4 - Card Valid
Bit 5 - Module Valid
Bit 6 - Bank Valid
(When Bank is addressed via group/address, refer to Bit 20 and 21)
Bit 7 - Device Valid
Bit 8 - Row Valid
Bit 9 - Column Valid
Bit 10 - Rank Valid
Bit 11 - Bit Position Valid
Bit 12 - Chip Identification Valid
Bit 13 - Memory Error Type Valid
Bit 14 - Status Valid
Bit 15 - Requestor ID Valid
Bit 16 - Responder ID Valid
Bit 17 - Target ID Valid
Bit 18 - Card Handle Valid
Bit 19 - Module Handle Valid
Bit 20 - Bank Group Valid
Bit 21 - Bank Address Valid
Bit 22-63 Reserved

Error Status

8

8

Memory error status information. See See Error Status for error status details.

Physical Address

16

8

The physical address at which the memory error occurred.

Physical Address Mask

24

8

Defines the valid address bits in the Physical Address field. The mask specifies the granularity of the physical address which is dependent on the hardware implementation factors such as interleaving.

Node

32

2

In a multi-node system, this value identifies the node containing the memory in error.

Card

34

2

The card number of the memory error location.

Module

36

2

The module number of the memory error location. (NODE, CARD, and MODULE should provide the information necessary to identify the failing FRU).

Bank

38

2

The bank number of the memory associated with the error. When Bank is addressed via group/address (e.g., DDR4) Bit 7:0 - Bank Address Bit 15:8 - Bank Group

Device

40

4

The device number of the memory associated with the error.

Row

44

4

The row number of the memory error location.

Column

48

4

The column number of the memory error location.

Rank

52

4

The rank number of the memory error location.

Bit Position

56

4

The bit position at which the memory error occurred.

Chip Identification

60

1

The Chip Identification. This is an encoded field used to address the die in 3DS packages.

Memory Error Type

61

1
Identifies the type of error that occurred:
0 - Unknown
1 - No error
2 - Single-bit ECC
3 - Multi-bit ECC
4 - Single-symbol ChipKill ECC
5 - Multi-symbol ChipKill ECC
6 - Master abort
7 - Target abort
8 - Parity Error
9 - Watchdog timeout
10 - Invalid address
11 - Mirror Broken
12 - Memory Sparing
13 - Scrub corrected error
14 - Scrub uncorrected error
15 - Physical Memory Map-out event
All other values reserved.
16 - 255 Reserved

Status

62

1
Bit 0:
If set to 0, the memory error is corrected; if set to 1, the memory error is uncorrected
Bit 1-7: Reserved values are 0

Reserved

63

1

Reserved values are 0

Requestor ID

64

8

Hardware address of the device that initiated the transaction that took the error.

Responder ID

72

8

Hardware address of the device that responded to the transaction.

Target ID

80

8

Hardware address of the intended target of the transaction.

Card Handle

88

4

This field contains the SMBIOS handle for the Type 16 Memory Array Structure that represents the memory card.

Module Handle

92

4

This field contains the SMBIOS handle for the Type 17 Memory Device Structure that represents the Memory Module.

N.2.7. PCI Express Error Section

Type: {0xD995E954, 0xBBC1, 0x430F, {0xAD, 0x91, 0xB4, 0x4D, 0xCB, 0x3C, 0x6F, 0x35}}

Table N.33 PCI Express Error Record

Mnemonic

Byte Offset

Byte Length

Description

Validation Bits

0

8
Indicates which of the following fields is valid:
Bit 0 -Port Type Valid
Bit 1 - Version Valid
Bit 2 - Command Status Valid
Bit 3 - Device ID Valid
Bit 4 - Device Serial Number Valid
Bit 5 - Bridge Control Status Valid
Bit 6 - Capability Structure Status Valid
Bit 7 - AER Info Valid
Bit 8-63 - Reserved

Port Type

8

4
PCIe Device/Port Type as defined in the PCI Express capabilities register:
0: PCI Express End Point
1: Legacy PCI End Point Device
4: Root Port 5: Upstream Switch Port
6: Downstream Switch Port
7: PCI Express to PCI/PCI-X Bridge
8: PCI/PCI-X to PCI Express Bridge
9: Root Complex Integrated Endpoint Device
10: Root Complex Event Collector

Version

12

4
PCIe Spec. version supported by the platform:
Byte 0-1: PCIe Spec. Version Number
• Byte0: Minor Version in BCD
• Byte1: Major Version in BCD
Byte2-3: Reserved

Command Status

16

4
Byte0-1: PCI Command Register
Byte2-3: PCI Status Register

Reserved

20

4

Must be zero

Device ID

24

16
PCIe Root Port PCI/bridge PCI compatible device number and bus number information to uniquely identify the root port or bridge.
Default values for both the bus numbers is zero.
Byte 0-1: Vendor ID
Byte 2-3: Device ID
Byte 4-6: Class Code
Byte 7: Function Number
Byte 8: Device Number
Byte 9-10: Segment Number
Byte 11: Root Port/Bridge Primary Bus Number or device bus number
Byte 12: Root Port/Bridge Secondary Bus Number
Byte 13-14: Bit0:2: Reserved Bit3:15 Slot Number
Byte 15 Reserved

Device Serial Number

40

8
Byte 0-3: PCIe Device Serial Number Lower DW
Byte 4-7: PCIe Device Serial Number Upper DW

Bridge Control Status

48

4
This field is valid for bridges only.
Byte 0-1: Bridge Secondary Status Register
Byte 2-3: Bridge Control Register

Capability Structure

52

60
PCIe Capability Structure.
- The 60-byte structure is used to report device capabilities. This structure is used to report the 36-byte PCIe 1.1 Capability Structure (See Figure 7-9 of the PCI Express Base Specification, Rev 1.1) with the last 24 bytes padded.
- This structure is also used to report the 60-byte PCIe 2.0 Capability Structure (See Figure 7-9 of the PCI Express 2.0 Base Specification.)
- The fields in the structure vary with different device types.
- The “Next CAP pointer” field should be considered invalid and any reserved fields of the structure are reserved for future use.
Note that PCIe devices without AER (PCI e_AER_INFO_STRU CT_VALID_BIT=0) may report status using this structure.

AER Info

112

96

PCIe Advanced Error Reporting Extended Capability Structure.

N.2.8. PCI/PCI-X Bus Error Section

Type: {0xC5753963, 0x3B84, 0x4095, {0xBF, 0x78, 0xED, 0xDA, 0xD3, 0xF9, 0xC9, 0xDD}}

Table N.34 PCI/PCI-X Bus Error Section

Mnemonic

Byte Offset

Byte Length

Description

Validation Bits

0

8
Indicates which of the following fields is valid:
Bit 0 -Error Status Valid
Bit 1 - Error Type Valid
Bit 2 - Bus Id Valid
Bit 3 - Bus Address Valid
Bit 4 - Bus Data Valid
Bit 5 - Command Valid
Bit 6 - Requestor Id Valid
Bit 7 - Completer Id Valid
Bit 8 - Target Id Valid
Bit 9-63 Reserved

Error Status

8

8

PCI Bus Error Status. See See Error Status for details.

Error Type

16

2
PCI Bus error Type
Byte 0:
0 - Unknown or OEM system specific error
1 - Data Parity Error
2 - System Error
3 - Master Abort
4 - Bus Timeout or No Device Present (No DEVSEL#)
5 - Master Data Parity Error
6 - Address Parity Error
7 - Command Parity Error
Others - Reserved
Byte 1: Reserved

Bus Id

18

2
Bits 0:7 - Bus Number
Bits 8:15 - Segment Number

Reserved

20

4

Bus Address

24

8

Memory or I/O address on the bus at the time of the error.

Bus Data

32

8

Data on the PCI bus at the time of the error.

Bus Command

40

8
Bus command or operation at the time of the error.
Byte 7: Bits 7-1: Reserved (should be zero)
Byte 7: Bit 0: If 0, then the command is a PCI command.
If 1, the command is a PCI-X command.

Bus Requestor Id

48

8

PCI Bus Requestor Id.

Bus Completer Id

56

8

PCI Bus Responder Id.

Target Id

64

8

PCI Bus intended target identifier.

N.2.9. PCI/PCI-X Component Error Section

Type: {0xEB5E4685, 0xCA66, 0x4769, {0xB6, 0xA2, 0x26, 0x06, 0x8B, 0x00, 0x13, 0x26}}

Table N.35 PCI/PCI-X Component Error Section

Mnemonic

Byte Offset

Byte Length

Description

Validation Bits

0

8
Indicate which fields are valid:
Bit 0 - Error Status Valid
Bit 1 - Id Info Valid
Bit 2 - Memory Number Valid
Bit 3 - IO Number Valid
Bit 4 - Register Data Pair Valid
Bit 5-63 Reserved

Error Status

8

8

PCI Component Error Status. See Error Status for details.

Id Info

16

16
Identification Information:
Bytes 0-1: Vendor Id
Bytes 1-2: Device Id
Bytes 4-6: Class Code
Byte 7: Function Number
Byte 8: Device Number
Byte 9: Bus Number
Byte 10: Segment Number
Bytes 11-15: Reserved

Memory Number

32

4

Number of PCI Component Memory Mapped register address/data pair values present in this structure.

IO Number

36

4

Number of PCI Component Programmed IO register address/data pair values present in this structure.

Register Data Pairs

40

2x8xN

An array of address/data pair values. The address and data information may be from 2 to 8 bytes of actual data represented in the 8 byte array locations.

N.2.10. Firmware Error Record Reference

Type: {0x81212A96, 0x09ED, 0x4996, {0x94, 0x71, 0x8D, 0x72, 0x9C, 0x8E, 0x69, 0xED}}

Table N.36 Firmware Error Record Reference

Mnemonic

Byte Offset

Byte Length

Description

Firmware Error Record Type

0

1
Identifies the type of firmware error record that is referenced by this section:
0: IPF SAL Error Record
1: SOC Firmware error record Type1 is reserved and used by Legacy CrashLog support
2: SOC Firmware error record Type2
All other values reserved

Revision

1

1

Indicates the Header Revision. For this Revision of the specification value is 2.

Reserved

1

7

Must be zero.

Record Identifier

8

8
This value uniquely identifies the firmware error record referenced by this section. This value may be used to retrieve the referenced firmware error record using means appropriate for the error record type.
Note: value is ignored for Revision >=1 of the header and must be set to NULL.

Record identifier GUID extension

16

16
This value uniquely identifies the firmware error record referenced by this section. This value may be used to retrieve the referenced firmware error record using means appropriate for the error record type.
Note: in case if Error Record Type == 2 then this filed indicates the GUID.
For Error Record Type 0 and Type 1 this field is ignored.

N.2.11. DMAr Error Sections

The DMAr error sections are divided into two different components as described below:

DMAr Generic Error Section:

This section holds information about DMAr errors in a generic form and will be common across all DMAr unit architectures.

Architecture specific DMAr Error Section:

This section consists of DMA remapping errors specific to the architecture. In addition, certain state information of the DMAr unit is captured at the time of error. This section is unique for each DMAr architecture (VT-d, IOMMU).

N.2.11.1. DMAr Generic Error Section

Type: {0x5B51FEF7, 0xC79D, 0x4434, {0x8F, 0x1B, 0xAA, 0x62, 0xDE, 0x3E, 0x2C, 0x64}}

Table N.37 DMAr Generic Errors

Mnemonic

Byte Offset

Byte Length

Description

Requester-ID

0

2

Device ID associated with a fault condition

Segment Number

2

2

PCI segment associated with a device

Fault Reason

4

1
1h: Domain mapping table entry is not present
2h: Invalid domain mapping table entry
3h: DMAr unit’s attempt to access the domain mapping table resulted in an error
4h: Reserved bit set to non-zero value in the domain mapping table
5h: DMA request to access an address beyond the device address width
6h: Invalid read or write access
7h: Invalid device request
8h: DMAr unit’s attempt to access the address translation table resulted in an error
9h: Reserved bit set to non-zero value in the address translation table
Ah: Illegal command error
Bh: DMAr unit’s attempt to access the command buffer resulted in an error
Other values are reserved

Access Type

5

1
0h: DMA Write
1h: DMA Read
Other values are reserved

Address Type

6

1
0h: Untranslated request
1h: Translation request
Other values are reserved

Architecture Type

7

1

1h: VT-d architecture 2h: IOMMU architecture Other values are reserved

Device Address

8

8

This field contains the 64-bit device virtual address in the faulted DMA request.

Reserved

16

16

Must be 0

N.2.11.2. Intel® VT for Directed I/O specific DMAr Error Section

Type: {0x71761D37, 0x32B2, 0x45cd, {0xA7, 0xD0, 0xB0, 0xFE 0xDD, 0x93, 0xE8, 0xCF}} All fields in this error section are specific to Intel’s VT-d architecture. This error section has a fixed size.

Table N.38 Intel® VT for Directed I/O specific DMAr Errors

Mnemonic

Byte Offset

Byte Length

Description

Version

0

1

Value of version register as defined in VT-d architecture

Revision

1

1

Value of revision field in VT-d specific DMA remapping reporting structure

OemId

2

6

Value of OEM ID field in VT-d specific DMA remapping reporting structure

Capability

8

8

Value of capability register in VT-d architecture

Extended Capability

16

8

Value of extended capability register in VT-d architecture

Global Command

24

4

Value of Global Command register in VT-d architecture programmed by the operating system

Global Status

28

4

Value of Global Status register in VT-d architecture

Fault Status

32

4

Value of Fault Status register in VT-d architecture

Reserved

36

12

Must be 0

Fault record

48

16

Fault record as defined in the VT-d specification

Root Entry

64

16

Value from the root entry table for the given requester-ID

Context Entry

80

16

Value from the context entry table for the given requester-ID.

Level 6 Page Table Entry

96

8

PTE entry for device virtual address in page level 6

Level 5 Page Table Entry

104

8

PTE entry for device virtual address in page level 5

Level 4 Page Table Entry

112

8

PTE entry for device virtual address in page level 4

Level 3 Page Table Entry

120

8

PTE entry for device virtual address in page level 3

Level 2 Page Table Entry

128

8

PTE entry for device virtual address in page level 2.

Level 1 Page Table Entry

136

8

PTE entry for device virtual address in page level 1

N.2.11.3. IOMMU Specific DMAr Error Section

Type: {0x036F84E1, 0x7F37, 0x428c, {0xA7, 0x9E, 0x57, 0x5F, 0xDF, 0xAA, 0x84, 0xEC}}

All fields in this error record are specific to AMD’s IOMMU specification. This error section has a fixed size.

Table N.39 IOMMU-specific DMAr Errors

Mnemonic

Byte Offset

Byte Length

Description

Revision

0

1

Specifies the IOMMU specification revision

Reserved

1

7

Must be 0

Control

8

8

IOMMU control register

Status

16

8

IOMMU status register

Reserved

24

8

Must be 0

Event Log Entry

32

16

IOMMU fault related event log entry as defined in the IOMMU specification

Reserved

48

16

Must be 0

Device Table Entry

64

32

Value from the device table for a given Requester ID

Level 6 Page Table Entry

96

8

PTE entry for device virtual address in page level 6

Level 5 Page Table Entry

104

8

PTE entry for device virtual address in page level 5

Level 4 Page Table Entry

112

8

PTE entry for device virtual address in page level 4

Level 3 Page Table Entry

120

8

PTE entry for device virtual address in page level 3

Level 2 Page Table Entry

128

8

PTE entry for device virtual address in page level 2

Level 1 Page Table Entry

136

8

PTE entry for device virtual address in page level 1

N.2.12. CCIX PER Log Error Section

Type:{0x91335EF6, 0xEBFB, 0x4478, {0xA6, 0xA6, 0x88, 0xB7, 0x28, 0xCF, 0x75, 0xD7}}

Table N.40 CCIX PER Log Error Record

Mnemonic

Byte Offset

Byte Length

Description

Length

0

4

Length in bytes for entire structure.

Validation Bits

4

8
Indicates which of the following fields is valid:
Bit 0 - CCIX Source ID Valid
Bit 1 - CCIX Port ID Valid
Bit 2 - CCIX PER Log Valid
Bit 3-63 - Reserved

CCIX Source ID

12

1
If the agent type is an HA, SA, or RA: This field indicates the CCIX Agent ID of the component that reported this error. In this case bits 7:6 must be zero, since Agent ID is only 6 bits.
Otherwise, this field this specifies the CCIX Device ID (i.e. in the case of Port, CCIX Link, or device errors).

CCIX Port ID

13

1

This field indicates the CCIX Port ID that reported this error. Bits 7:5 must be zero, since CCIX Port ID is only 5 bits.

Reserved

14

2

Must be zero.

CCIX PER Log

16

20…n
DWORD (32-bit) entries in CCIX PER Log Structure, as described in Section 7.3.2 of the CCIX Base Specification - Revision 1.0.
NOTE: The Per Log Structure contains a header describing the number of DWORDs in the error record.

N.2.14. CXL Component Events Section

Refer to the Events Record Format for CXL components in the CXL Specification, Rev 2.0 or later.

  • For the Section Type GUID: Refer to the Event Record Identifier field (Offset 0) of the Events Record Format for each CXL component.

  • For the CXL Component Event Log: Refer to the Common Event Record field (Offset 16) of the Events Record Format for each CXL component.

Table N.42 CXL Component Event Log Record

Mnemonic

Byte Offset

Byte Length

Description

Length

0

4

Length in bytes for entire structure.

Validation Bits

4

8
Bit 0 - Device ID Valid
Bit 1 - Device Serial Number Valid
Bit 2 - CXL Component Event Log Valid
Bit 3-63 - Reserved

Device ID

12

12
PCIe Device Identifiers of CXL Component:

Byte 0-1: Vendor ID
Byte 2-3: Device ID
Byte 4: Function Number
Byte 5: Device Number
Byte 6: Bus Number
Byte 7-8: Segment Number
Byte 9-10:
Bit 0:2: Reserved
Bit 3:15 Slot Number
Byte 11 Reserved

Device Serial Number

24

8

Byte 0-3: PCIe Device Serial Number Lower DW Byte 4-7: PCIe Device Serial Number Upper DW

CXL Component Event Log

32

CXL Component Event Log, starting with the Common Event Record field corresponding to the Component specified by the Section Type GUID. The Length of this field may vary.