ESR_EL2, Exception Syndrome Register (EL2)

The ESR_EL2 characteristics are:

Purpose

Holds syndrome information for an exception taken to EL2.

Configuration

AArch64 System register ESR_EL2 bits [31:0] are architecturally mapped to AArch32 System register HSR[31:0].

If EL2 is not implemented, this register is RES0 from EL3.

This register has no effect if EL2 is not enabled in the current Security state.

Attributes

ESR_EL2 is a 64-bit register.

Field descriptions

6362616059585756555453525150494847464544434241403938373635343332
313029282726252423222120191817161514131211109876543210
RES0ISS2
ECILISS

ESR_EL2 is made UNKNOWN as a result of an exception return from EL2.

When an UNPREDICTABLE instruction is treated as UNDEFINED, and the exception is taken to EL2, the value of ESR_EL2 is UNKNOWN. The value written to ESR_EL2 must be consistent with a value that could be created as a result of an exception from the same Exception level that generated the exception as a result of a situation that is not UNPREDICTABLE at that Exception level, in order to avoid the possibility of a privilege violation.

Bits [63:56]

Reserved, RES0.

ISS2, bits [55:32]

ISS2 encoding for an exception, the bit assignments are:

ISS2 encoding for an exception from a Data Abort

23222120191817161514131211109876543210
RES0HDBSSFTnDTagAccessGCSAssuredOnlyOverlayDirtyBitXs

Bits [23:12]

Reserved, RES0.

HDBSSF, bit [11]
When FEAT_HDBSS is implemented:

Indicates that the fault was caused by the HDBSS.

When DFSC indicates a Permission fault, this field indicates whether that fault was caused by the HDBSS being full.

When DFSC indicates a synchronous External abort on translation table walk or hardware update of translation table, this field indicates whether the fault was caused by a write to the HDBSS.

When DFSC indicates a Granule Protection Fault on translation table walk or hardware update of translation table, this field indicates whether the fault was caused by a write to the HDBSS.

HDBSSFMeaning
0b0

Fault was not caused by HDBSS.

0b1

Fault was caused by HDBSS.

For any other fault, this field is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

TnD, bit [10]
When FEAT_MTE_CANONICAL_TAGS is implemented:

Tag not Data.

If a memory access generates a Data Abort for a stage 1 Permission fault, this field indicates whether the fault is due to an Allocation Tag access.

TnDMeaning
0b0

Permission fault is not due to a write of an Allocation Tag to Canonically Tagged memory.

0b1

Permission fault is due to a write of an Allocation Tag to Canonically Tagged memory.

For any other fault, this field is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

TagAccess, bit [9]
When FEAT_MTE_PERM is implemented:

NoTagAccess fault.

If a memory access generates a Data Abort for a Permission fault, this field indicates whether the fault is due to the NoTagAccess memory attribute.

TagAccessMeaning
0b0

Permission fault is not due to the NoTagAccess memory attribute.

0b1

Permission fault is due to the NoTagAccess memory attribute.

For any other fault, this field is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

GCS, bit [8]
When FEAT_GCS is implemented:

Guarded Control Stack data access.

If a memory access generates a Data Abort, this field indicates whether the fault is due to a Guarded Control Stack data access.

GCSMeaning
0b0

The Data Abort is not due to a Guarded control stack data access.

0b1

The Data Abort is due to a Guarded control stack data access.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

AssuredOnly, bit [7]
When FEAT_THE is implemented:

AssuredOnly flag.

If a memory access generates a stage 2 Data Abort for a Permission fault, this field holds information about the fault.

AssuredOnlyMeaning
0b0

The Data Abort is not due to AssuredOnly.

0b1

The Data Abort is due to AssuredOnly.

For any other fault, this field is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Overlay, bit [6]
When FEAT_S1POE is implemented or FEAT_S2POE is implemented:

Overlay flag.

If a memory access generates a Data Abort for a Permission fault, this field holds information about the fault.

OverlayMeaning
0b0

The Data Abort is not due to Overlay Permissions.

0b1

The Data Abort is due to Overlay Permissions.

For any other fault, this field is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

DirtyBit, bit [5]
When FEAT_S1PIE is implemented or FEAT_S2PIE is implemented:

DirtyBit flag.

If a write access to memory generates a Data Abort for a Permission fault using Indirect Permission, this field holds information about the fault.

DirtyBitMeaning
0b0

Permission Fault is not due to dirty state.

0b1

Permission Fault is due to dirty state.

For any other fault or Access, this field is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Xs, bits [4:0]
When FEAT_LS64 is implemented:

When FEAT_LS64_V is implemented, if a memory access generated by an ST64BV instruction generates a Data Abort exception for a Translation fault, Access flag fault, or Permission fault, then this field holds register specifier, Xs.

When FEAT_LS64_ACCDATA is implemented, if a memory access generated by an ST64BV0 instruction generates a Data Abort exception for a Translation fault, Access flag fault, or Permission fault, then this field holds register specifier, Xs.

Otherwise, this field is RES0.


Otherwise:

Reserved, RES0.

ISS2 encoding for an exception from an Instruction Abort

23222120191817161514131211109876543210
RES0HDBSSFRES0AssuredOnlyOverlayRES0

Bits [23:12]

Reserved, RES0.

HDBSSF, bit [11]
When FEAT_HDBSS is implemented:

Indicates that the fault was caused by the HDBSS.

When IFSC indicates a Permission fault, this field indicates whether the fault was caused by the HDBSS being full.

When IFSC indicates a synchronous External abort on translation table walk or hardware update of translation table, this field indicates whether the fault was caused by a write to the HDBSS.

When IFSC indicates a Granule Protection Fault on translation table walk or hardware update of translation table, this field indicates whether the fault was caused by a write to the HDBSS.

HDBSSFMeaning
0b0

Fault was not caused by HDBSS.

0b1

Fault was caused by HDBSS.

For any other fault, this field is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bits [10:8]

Reserved, RES0.

AssuredOnly, bit [7]
When FEAT_THE is implemented:

AssuredOnly flag.

If a memory access generates a Instruction Abort for a Permission fault, then this field holds information about the fault.

AssuredOnlyMeaning
0b0

Instruction Abort is not due to AssuredOnly.

0b1

Instruction Abort is due to stage 2 AssuredOnly attribute.

For any other fault, this field is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Overlay, bit [6]
When FEAT_S1POE is implemented or FEAT_S2POE is implemented:

Overlay flag.

If a memory access generates a Instruction Abort for a Permission fault, then this field holds information about the fault.

OverlayMeaning
0b0

Instruction Abort is not due to Overlay Permissions.

0b1

Instruction Abort is due to Overlay Permissions.

For any other fault, this field is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bits [5:0]

Reserved, RES0.

ISS2 encoding for an exception from a Watchpoint exception

23222120191817161514131211109876543210
RES0GCSRES0

Bits [23:9]

Reserved, RES0.

GCS, bit [8]
When FEAT_GCS is implemented:

Guarded control stack data access.

Indicates that the Watchpoint exception is due to a Guarded control stack data access.

GCSMeaning
0b0

The Watchpoint exception is not due to a Guarded control stack data access.

0b1

The Watchpoint exception is due to a Guarded control stack data access.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bits [7:0]

Reserved, RES0.

ISS2 encoding for all other exceptions

23222120191817161514131211109876543210
RES0

Bits [23:0]

Reserved, RES0.

EC, bits [31:26]

Exception Class. Indicates the reason for the exception that this register holds information about.

For each EC value, the table references a subsection that gives information about:

Possible values of the EC field are:

ECMeaningISSISS2Applies when
0b000000

Unknown reason.

ISS encoding for exceptions with an unknown reasonISS2 encoding for all other exceptions
0b000001

Trapped WF* instruction execution.

Conditional WF* instructions that fail their condition code check do not cause an exception.

ISS encoding for an exception from a WF* instructionISS2 encoding for all other exceptions
0b000011

Trapped MCR or MRC access with (coproc==0b1111) that is not reported using EC 0b000000.

ISS encoding for an exception from an MCR or MRC accessISS2 encoding for all other exceptionsWhen AArch32 is supported
0b000100

Trapped MCRR or MRRC access with (coproc==0b1111) that is not reported using EC 0b000000.

ISS encoding for an exception from an MCRR or MRRC accessISS2 encoding for all other exceptionsWhen AArch32 is supported
0b000101

Trapped MCR or MRC access with (coproc==0b1110).

ISS encoding for an exception from an MCR or MRC accessISS2 encoding for all other exceptionsWhen AArch32 is supported
0b000110

Trapped LDC or STC access.

The only architected uses of these instruction are:

ISS encoding for an exception from an LDC or STC instructionISS2 encoding for all other exceptionsWhen AArch32 is supported
0b000111

Access to SME, SVE, Advanced SIMD or floating-point functionality trapped by CPACR_EL1.FPEN, CPTR_EL2.FPEN, CPTR_EL2.TFP, or CPTR_EL3.TFP control.

Excludes exceptions resulting from CPACR_EL1 when the value of HCR_EL2.TGE is 1, or because SVE or Advanced SIMD and floating-point are not implemented. These are reported with EC value 0b000000.

ISS encoding for an exception from an access to SVE, Advanced SIMD or floating-point functionality, resulting from the FPEN and TFP trapsISS2 encoding for all other exceptions
0b001000

Trapped VMRS access, from ID group trap, that is not reported using EC 0b000111.

ISS encoding for an exception from an MCR or MRC accessISS2 encoding for all other exceptionsWhen AArch32 is supported
0b001001

Trapped use of a Pointer authentication instruction because HCR_EL2.API == 0 || SCR_EL3.API == 0.

ISS encoding for an exception from a Pointer Authentication instruction when HCR_EL2.API == 0 || SCR_EL3.API == 0ISS2 encoding for all other exceptionsWhen FEAT_PAuth is implemented
0b001010

An exception from an LD64B or ST64B* instruction.

ISS encoding for an exception from an LD64B or ST64B* instructionISS2 encoding for all other exceptionsWhen FEAT_LS64 is implemented
0b001100

Trapped MRRC access with (coproc==0b1110).

ISS encoding for an exception from an MCRR or MRRC accessISS2 encoding for all other exceptionsWhen AArch32 is supported
0b001101

Branch Target Exception.

ISS encoding for an exception from Branch Target Identification instructionISS2 encoding for all other exceptionsWhen FEAT_BTI is implemented
0b001110

Illegal Execution state.

ISS encoding for an exception from an Illegal Execution state, or a PC or SP alignment faultISS2 encoding for all other exceptions
0b010001

SVC instruction execution in AArch32 state.

This is reported in ESR_EL2 only when the exception is generated because the value of HCR_EL2.TGE is 1.

ISS encoding for an exception from HVC or SVC instruction executionISS2 encoding for all other exceptionsWhen AArch32 is supported
0b010010

HVC instruction execution in AArch32 state, when HVC is not disabled.

ISS encoding for an exception from HVC or SVC instruction executionISS2 encoding for all other exceptionsWhen AArch32 is supported
0b010011

SMC instruction execution in AArch32 state, when SMC is not disabled.

This is reported in ESR_EL2 only when the exception is generated because the value of HCR_EL2.TSC is 1.

ISS encoding for an exception from SMC instruction execution in AArch32 stateISS2 encoding for all other exceptionsWhen AArch32 is supported
0b010100

Trapped MSRR, MRRS or System instruction execution in AArch64 state, that is not reported using EC 0b000000.

ISS encoding for an exception from MSRR, MRRS, or 128-bit System instruction execution in AArch64 stateISS2 encoding for all other exceptionsWhen FEAT_SYSREG128 is implemented or FEAT_SYSINSTR128 is implemented
0b010101

SVC instruction execution in AArch64 state.

ISS encoding for an exception from HVC or SVC instruction executionISS2 encoding for all other exceptionsWhen AArch64 is supported
0b010110

HVC instruction execution in AArch64 state, when HVC is not disabled.

ISS encoding for an exception from HVC or SVC instruction executionISS2 encoding for all other exceptionsWhen AArch64 is supported
0b010111

SMC instruction execution in AArch64 state, when SMC is not disabled.

This is reported in ESR_EL2 only when the exception is generated because the value of HCR_EL2.TSC is 1.

ISS encoding for an exception from SMC instruction execution in AArch64 stateISS2 encoding for all other exceptionsWhen AArch64 is supported
0b011000

Trapped MSR, MRS or System instruction execution in AArch64 state, that is not reported using EC 0b000000, 0b000001 or 0b000111.

This includes all instructions that cause exceptions that are part of the encoding space defined in 'System instruction class encoding overview', except for those exceptions reported using EC values 0b000000, 0b000001, or 0b000111.

ISS encoding for an exception from MSR, MRS, or System instruction execution in AArch64 stateISS2 encoding for all other exceptionsWhen AArch64 is supported
0b011001

Access to SVE functionality trapped as a result of CPACR_EL1.ZEN, CPTR_EL2.ZEN, CPTR_EL2.TZ, or CPTR_EL3.EZ, that is not reported using EC 0b000000.

ISS encoding for an exception from an access to SVE functionality, resulting from CPACR_EL1.ZEN, CPTR_EL2.ZEN, CPTR_EL2.TZ, or CPTR_EL3.EZISS2 encoding for all other exceptionsWhen FEAT_SVE is implemented
0b011010

Trapped ERET, ERETAA, or ERETAB instruction execution.

ISS encoding for an exception from an ERET, ERETAA, or ERETAB instructionISS2 encoding for all other exceptionsWhen FEAT_FGT is implemented or FEAT_NV is implemented
0b011011

Exception from an access to a TSTART instruction at EL0 when SCTLR_EL1.TME0 == 0, EL0 when SCTLR_EL2.TME0 == 0, at EL1 when SCTLR_EL1.TME == 0, at EL2 when SCTLR_EL2.TME == 0 or at EL3 when SCTLR_EL3.TME == 0.

ISS encoding for an exception from a TSTART instructionISS2 encoding for all other exceptionsWhen FEAT_TME is implemented
0b011100

Exception from a PAC Fail

ISS encoding for a PAC Fail exceptionISS2 encoding for all other exceptionsWhen FEAT_FPAC is implemented
0b011101

Access to SME functionality trapped as a result of CPACR_EL1.SMEN, CPTR_EL2.SMEN, CPTR_EL2.TSM, CPTR_EL3.ESM, or an attempted execution of an instruction that is illegal because of the value of PSTATE.SM or PSTATE.ZA, that is not reported using EC 0b000000.

ISS encoding for an exception due to SME functionalityISS2 encoding for all other exceptionsWhen FEAT_SME is implemented
0b100000

Instruction Abort from a lower Exception level.

Used for MMU faults generated by instruction accesses and synchronous External aborts, including synchronous parity or ECC errors. Not used for debug-related exceptions.

ISS encoding for an exception from an Instruction AbortISS2 encoding for an exception from an Instruction Abort
0b100001

Instruction Abort taken without a change in Exception level.

Used for MMU faults generated by instruction accesses and synchronous External aborts, including synchronous parity or ECC errors. Not used for debug-related exceptions.

ISS encoding for an exception from an Instruction AbortISS2 encoding for an exception from an Instruction Abort
0b100010

PC alignment fault exception.

ISS encoding for an exception from an Illegal Execution state, or a PC or SP alignment faultISS2 encoding for all other exceptions
0b100100

Data Abort exception from a lower Exception level, excluding Data Abort exceptions taken to EL2 as a result of accesses generated associated with VNCR_EL2 as part of nested virtualization support.

These Data Abort exceptions might be generated from Exception levels in any Execution state.

Used for MMU faults generated by data accesses, alignment faults other than those caused by Stack Pointer misalignment, and synchronous External aborts, including synchronous parity or ECC errors. Not used for debug-related exceptions.

ISS encoding for an exception from a Data AbortISS2 encoding for an exception from a Data Abort
0b100101

Data Abort exception without a change in Exception level, or Data Abort exceptions taken to EL2 as a result of accesses generated associated with VNCR_EL2 as part of nested virtualization support.

Used for MMU faults generated by data accesses, alignment faults other than those caused by Stack Pointer misalignment, and synchronous External aborts, including synchronous parity or ECC errors. Not used for debug-related exceptions.

ISS encoding for an exception from a Data AbortISS2 encoding for an exception from a Data Abort
0b100110

SP alignment fault exception.

ISS encoding for an exception from an Illegal Execution state, or a PC or SP alignment faultISS2 encoding for all other exceptions
0b100111

Memory Operation Exception.

ISS encoding for an exception from the Memory Copy and Memory Set instructionsISS2 encoding for all other exceptionsWhen FEAT_MOPS is implemented
0b101000

Trapped floating-point exception taken from AArch32 state.

This EC value is valid if the implementation supports trapping of floating-point exceptions, otherwise it is reserved. Whether a floating-point implementation supports trapping of floating-point exceptions is IMPLEMENTATION DEFINED.

ISS encoding for an exception from a trapped floating-point exceptionISS2 encoding for all other exceptionsWhen AArch32 is supported
0b101100

Trapped floating-point exception taken from AArch64 state.

This EC value is valid if the implementation supports trapping of floating-point exceptions, otherwise it is reserved. Whether a floating-point implementation supports trapping of floating-point exceptions is IMPLEMENTATION DEFINED.

ISS encoding for an exception from a trapped floating-point exceptionISS2 encoding for all other exceptionsWhen AArch64 is supported
0b101101

GCS exception.

ISS encoding for a GCS exceptionISS2 encoding for all other exceptionsWhen FEAT_GCS is implemented
0b101111

SError exception.

ISS encoding for an SError exceptionISS2 encoding for all other exceptions
0b110000

Breakpoint exception from a lower Exception level.

ISS encoding for an exception from a Breakpoint or Vector Catch debug exceptionISS2 encoding for all other exceptions
0b110001

Breakpoint exception taken without a change in Exception level.

ISS encoding for an exception from a Breakpoint or Vector Catch debug exceptionISS2 encoding for all other exceptions
0b110010

Software Step exception from a lower Exception level.

ISS encoding for an exception from a Software Step exceptionISS2 encoding for all other exceptions
0b110011

Software Step exception taken without a change in Exception level.

ISS encoding for an exception from a Software Step exceptionISS2 encoding for all other exceptions
0b110100

Watchpoint from a lower Exception level, excluding Watchpoint Exceptions taken to EL2 as a result of accesses generated associated with VNCR_EL2 as part of nested virtualization support.

These Watchpoint Exceptions might be generated from Exception levels using any Execution state.

ISS encoding for an exception from a Watchpoint exceptionISS2 encoding for an exception from a Watchpoint exception
0b110101

Watchpoint exceptions without a change in Exception level, or Watchpoint exceptions taken to EL2 as a result of accesses generated associated with VNCR_EL2 as part of nested virtualization support.

ISS encoding for an exception from a Watchpoint exceptionISS2 encoding for an exception from a Watchpoint exception
0b111000

BKPT instruction execution in AArch32 state.

ISS encoding for an exception from execution of a Breakpoint instructionISS2 encoding for all other exceptionsWhen AArch32 is supported
0b111010

Vector Catch exception from AArch32 state.

The only case where a Vector Catch exception is taken to an Exception level that is using AArch64 is when the exception is routed to EL2 and EL2 is using AArch64.

ISS encoding for an exception from a Breakpoint or Vector Catch debug exceptionISS2 encoding for all other exceptionsWhen AArch32 is supported
0b111100

BRK instruction execution in AArch64 state.

ISS encoding for an exception from execution of a Breakpoint instructionISS2 encoding for all other exceptionsWhen AArch64 is supported
0b111101

PMU exception

ISS encoding for a PMU exceptionISS2 encoding for all other exceptionsWhen FEAT_EBEP is implemented

All other EC values are reserved by Arm, and:

The effect of programming this field to a reserved value is that behavior is CONSTRAINED UNPREDICTABLE.

The reset behavior of this field is:

IL, bit [25]

Instruction Length for synchronous exceptions. Possible values of this bit are:

ILMeaning
0b0

16-bit instruction trapped.

0b1

32-bit instruction trapped. This value is also used when the exception is one of the following:

  • An SError exception.

  • An Instruction Abort exception.

  • A PC alignment fault exception.

  • An SP alignment fault exception.

  • A Data Abort exception for which the value of the ISV bit is 0.

  • An Illegal Execution state exception.

  • Any debug exception except for Breakpoint instruction exceptions. For Breakpoint instruction exceptions, this bit has its standard meaning:

    • 0b0: 16-bit T32 BKPT instruction.

    • 0b1: 32-bit A32 BKPT instruction or A64 BRK instruction.

  • An exception reported using EC value 0b000000.

The reset behavior of this field is:

ISS, bits [24:0]

Instruction Specific Syndrome. Architecturally, this field can be defined independently for each defined Exception class. However, in practice, some ISS encodings are used for more than one Exception class.

Typically, an ISS encoding has a number of subfields. When an ISS subfield holds a register number, the value returned in that field is the AArch64 view of the register number.

For an exception taken from AArch32 state, see 'Mapping of the general-purpose registers between the Execution states'.

If the AArch32 register descriptor is 0b1111, then:

ISS encoding for exceptions with an unknown reason

2423222120191817161514131211109876543210
RES0

Bits [24:0]

Reserved, RES0.

Additional information for the ISS encoding for exceptions with an unknown reason

When an exception is reported using this EC value, the IL field is set to 1.

This EC value is used for all exceptions that are not covered by any other EC value. This includes exceptions that are generated in the following situations:

  • The attempted execution of an instruction bit pattern that has no allocated instruction or that is not accessible at the current Exception level and Security state, including:
    • A read access using a System register pattern that is not allocated for reads or that does not permit reads at the current Exception level and Security state.
    • A write access using a System register pattern that is not allocated for writes or that does not permit writes at the current Exception level and Security state.
    • Instruction encodings that are unallocated.
    • Instruction encodings for instructions or System registers that are not implemented in the implementation.
  • In Debug state, the attempted execution of an instruction bit pattern that is not accessible in Debug state.
  • In Non-debug state, the attempted execution of an instruction bit pattern that is not accessible in Non-debug state.
  • In AArch32 state, attempted execution of a short vector floating-point instruction.
  • In an implementation that does not include Advanced SIMD and floating-point functionality, an attempted access to Advanced SIMD or floating-point functionality under conditions where that access would be permitted if that functionality was present. This includes the attempted execution of an Advanced SIMD or floating-point instruction, and attempted accesses to Advanced SIMD and floating-point System registers.
  • An exception generated because of the value of one of the SCTLR_EL1.{ITD, SED, CP15BEN} control bits.
  • Attempted execution of:
    • An HVC instruction when disabled by HCR_EL2.HCD or SCR_EL3.HCE.
    • An SMC instruction when disabled by SCR_EL3.SMD.
    • An HLT instruction when disabled by EDSCR.HDE.
  • Attempted execution of an MSR or MRS instruction to access SP_EL0 when the value of SPSel.SP is 0.
  • Attempted execution of an MSR or MRS instruction using a _EL12 register name when the Effective value of HCR_EL2.E2H is not 1.
  • Attempted execution, in Debug state, of:
    • A DCPS1 instruction when the value of HCR_EL2.TGE is 1 and EL2 is disabled or not implemented in the current Security state.
    • A DCPS2 instruction from EL1 or EL0 when EL2 is disabled or not implemented in the current Security state.
    • A DCPS3 instruction when the value of EDSCR.SDD is 1, or when EL3 is not implemented.
  • When EL3 is using AArch64, attempted execution from Secure EL1 of an SRS instruction using R13_mon.
  • In Debug state when the value of EDSCR.SDD is 1, the attempted execution at EL2, EL1, or EL0 of an instruction that is configured to trap to EL3.
  • In AArch32 state, the attempted execution of an MRS (banked register) or an MSR (banked register) instruction to SPSR_mon, SP_mon, or LR_mon.
  • An exception that is taken to EL2 because the value of HCR_EL2.TGE is 1. If the value of HCR_EL2.TGE is 0, this exception is reported using an ESR_EL2.EC value of 0b000111.
  • In Non-transactional state, attempted execution of a TCOMMIT instruction.

ISS encoding for an exception from a WF* instruction

2423222120191817161514131211109876543210
CVCONDRES0RNRES0RVTI

CV, bit [24]

Condition code valid.

CVMeaning
0b0

The COND field is not valid.

0b1

The COND field is valid.

For exceptions taken from AArch64, CV is set to 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether CV is set to 1 or set to 0. See the description of the COND field for more information.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

COND, bits [23:20]

For exceptions taken from AArch64, this field is set to 0b1110.

The condition code for the trapped instruction. This field is valid only for exceptions taken from AArch32, and only when the value of CV is 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1 and:
    • If the instruction is conditional, COND is set to the condition code field value from the instruction.
    • If the instruction is unconditional, COND is set to 0b1110.
  • A conditional A32 instruction that is known to pass its condition code check can be presented either:
    • With COND set to 0b1110, the value for unconditional.
    • With the COND value held in the instruction.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether:
    • CV is set to 0 and COND is set to an UNKNOWN value. Software must examine the SPSR.IT field to determine the condition, if any, of the T32 instruction.
    • CV is set to 1 and COND is set to the condition code for the condition that applied to the instruction.
  • For an implementation that, for both A32 and T32 instructions, takes an exception on a trapped conditional instruction only if the instruction passes its condition code check, these definitions mean that when CV is set to 1 it is IMPLEMENTATION DEFINED whether the COND field is set to 0b1110, or to the value of any condition that applied to the instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits [19:10]

Reserved, RES0.

RN, bits [9:5]
When FEAT_WFxT is implemented:

Register Number. Indicates the register number supplied for a WFET or WFIT instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bits [4:3]

Reserved, RES0.

RV, bit [2]
When FEAT_WFxT is implemented:

Register field Valid.

If TI[1] == 1, then this field indicates whether RN holds a valid register number for the register argument to the trapped WFET or WFIT instruction.

RVMeaning
0b0

Register field invalid.

0b1

Register field valid.

If TI[1] == 0, then this field is RES0.

This field is set to 1 on a trap on WFET or WFIT.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

TI, bits [1:0]

Trapped instruction. Possible values of this bit are:

TIMeaningApplies when
0b00

WFI trapped.

0b01

WFE trapped.

0b10

WFIT trapped.

When FEAT_WFxT is implemented
0b11

WFET trapped.

When FEAT_WFxT is implemented

When FEAT_WFxT is implemented, this is a two bit field as shown. Otherwise, bit[1] is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from a WF* instruction

The following fields describe configuration settings for generating this exception:

ISS encoding for an exception from an MCR or MRC access

2423222120191817161514131211109876543210
CVCONDOpc2Opc1CRnRtCRmDirection

CV, bit [24]

Condition code valid.

CVMeaning
0b0

The COND field is not valid.

0b1

The COND field is valid.

For exceptions taken from AArch64, CV is set to 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether CV is set to 1 or set to 0. See the description of the COND field for more information.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

COND, bits [23:20]

For exceptions taken from AArch64, this field is set to 0b1110.

The condition code for the trapped instruction. This field is valid only for exceptions taken from AArch32, and only when the value of CV is 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1 and:
    • If the instruction is conditional, COND is set to the condition code field value from the instruction.
    • If the instruction is unconditional, COND is set to 0b1110.
  • A conditional A32 instruction that is known to pass its condition code check can be presented either:
    • With COND set to 0b1110, the value for unconditional.
    • With the COND value held in the instruction.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether:
    • CV is set to 0 and COND is set to an UNKNOWN value. Software must examine the SPSR.IT field to determine the condition, if any, of the T32 instruction.
    • CV is set to 1 and COND is set to the condition code for the condition that applied to the instruction.
  • For an implementation that, for both A32 and T32 instructions, takes an exception on a trapped conditional instruction only if the instruction passes its condition code check, these definitions mean that when CV is set to 1 it is IMPLEMENTATION DEFINED whether the COND field is set to 0b1110, or to the value of any condition that applied to the instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Opc2, bits [19:17]

The Opc2 value from the issued instruction.

For a trapped VMRS access, holds the value 0b000.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Opc1, bits [16:14]

The Opc1 value from the issued instruction.

For a trapped VMRS access, holds the value 0b111.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

CRn, bits [13:10]

The CRn value from the issued instruction.

For a trapped VMRS access, holds the reg field from the VMRS instruction encoding.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Rt, bits [9:5]

The Rt value from the issued instruction, the general-purpose register used for the transfer.

If the Rt value is not 0b1111, then the reported value gives the AArch64 view of the register. Otherwise, if the Rt value is 0b1111:

  • If the instruction that generated the exception is not UNPREDICTABLE, then the register specifier takes the value 0b11111.

  • If the instruction that generated the exception is UNPREDICTABLE, then the register specifier takes an UNKNOWN value, which is restricted to either:

    • The AArch64 view of one of the registers that could have been used in AArch32 state at the Exception level that the instruction was executed at.

    • The value 0b11111.

See 'Mapping of the general-purpose registers between the Execution states'.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

CRm, bits [4:1]

The CRm value from the issued instruction.

For a trapped VMRS access, holds the value 0b0000.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Direction, bit [0]

Indicates the direction of the trapped instruction.

DirectionMeaning
0b0

Write to System register space. MCR instruction.

0b1

Read from System register space. MRC or VMRS instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from an MCR or MRC access

The following fields describe configuration settings for generating exceptions from an MCR or MRC access using coproc 0b1111, that are reported using EC value 0b000011:

  • If FEAT_TIDCP1 is implemented, SCTLR_EL1.TIDCP, for EL0 accesses to IMPLEMENTATION DEFINED functionality using AArch32 state, trapped to EL1.
  • CNTKCTL_EL1.{EL0PTEN, EL0VTEN, EL0PCTEN, EL0VCTEN}, for accesses to the Generic Timer Registers from EL0 using AArch32 state, trapped to EL1 or EL2.
  • PMUSERENR_EL0.{ER, CR, SW, EN}, for accesses to Performance Monitor registers from EL0 using AArch32 state, trapped to EL1 or EL2.
  • AMUSERENR_EL0.EN, for accesses to Activity Monitors registers from EL0 using AArch32 state, trapped to EL1 or EL2.
  • HCR.{TRVM, TVM} and HCR_EL2.{TRVM, TVM}, for accesses to virtual memory control registers from EL1 using AArch32 state, trapped to EL2.
  • HCR.TTLB and HCR_EL2.TTLB, for execution of TLB maintenance instructions at EL1 using AArch32 state, trapped to EL2.
  • HCR.{TSW, TPC, TPU} and HCR_EL2.{TSW, TPC, TPU} for execution of cache maintenance instructions at EL0 and EL1 using AArch32 state, trapped to EL2.
  • HCR.TAC and HCR_EL2.TACR, for accesses to the Auxiliary Control Register at EL1 using AArch32 state, trapped to EL2.
  • HCR.TIDCP and HCR_EL2.TIDCP, for accesses to lockdown, DMA, and TCM operations at EL0 and EL1 using AArch32 state, trapped to EL2.
  • If FEAT_TIDCP1 is implemented, SCTLR_EL2.TIDCP, for EL0 accesses to IMPLEMENTATION DEFINED functionality using AArch32 state, trapped to EL2.
  • HCR.{TID1, TID2, TID3} and HCR_EL2.{TID1, TID2, TID3}, for accesses to ID registers at EL0 and EL1 using AArch32 state, trapped to EL2.
  • HCR2.TERR, for Non-secure accesses to error record registers at EL1 using AArch32 state, trapped to EL2.
  • HCPTR.TCPAC and CPTR_EL2.TCPAC, for accesses to CPACR_EL1 or CPACR using AArch32 state, trapped to EL2.
  • HSTR.T<n> and HSTR_EL2.T<n>, for accesses to System registers using AArch32 state, trapped to EL2.
  • CNTHCTL.PL1PCEN and CNTHCTL_EL2.EL1PCEN, for accesses to the Generic Timer registers from EL0 and EL1 using AArch32 state, trapped to EL2.
  • HDCR.TTRF, for Non-secure accesses to trace filter control registers from system registers using AArch32 state, trapped to EL2.
  • HDCR.{TPM, TPMCR} and MDCR_EL2.{TPM, TPMCR}, for accesses to Performance Monitor registers from EL0 and EL1 using AArch32 state, trapped to EL2.
  • HCPTR.TAM and CPTR_EL2.TAM, for accesses to Activity Monitors registers from EL0 and EL1 using AArch32 state, trapped to EL2.
  • CPTR_EL3.TCPAC, for accesses to CPACR from EL1 and EL2, and accesses to HCPTR from EL2 using AArch32 state, trapped to EL3.
  • MDCR_EL3.TPM, for accesses to Performance Monitor registers from EL0, EL1 and EL2 using AArch32 state, trapped to EL3.
  • CPTR_EL3.TAM, for accesses to Activity Monitors registers from EL0, EL1 and EL2 using AArch32 state, trapped to EL3.
  • If FEAT_FGT is implemented, access to some registers at EL0, trapped to EL2.

The following fields describe configuration settings for generating exceptions from an MCR or MRC access using coproc 0b1110, that are reported using EC value 0b000101:

  • CPACR_EL1.TTA for accesses to trace registers, trapped to EL1 or EL2.
  • MDSCR_EL1.TDCC, for accesses to the Debug Communications Channel (DCC) registers at EL0 and EL1 using AArch32 state, trapped to EL1 or EL2.
  • If FEAT_FGT is implemented, MDCR_EL2.TDCC for accesses to the DCC registers at EL0 and EL1 trapped to EL2, and MDCR_EL3.TDCC for accesses to the DCC registers at EL0, EL1, and EL2 trapped to EL3.
  • HCR.TID0 and HCR_EL2.TID0, for accesses to the JIDR register in the ID group 0 at EL0 and EL1 using AArch32, trapped to EL2.
  • HCPTR.TTA and CPTR_EL2.TTA, for accesses to trace registers using AArch32, trapped to EL2.
  • HDCR.TDRA and MDCR_EL2.TDRA, for accesses to Debug ROM registers DBGDRAR and DBGDSAR using AArch32, trapped to EL2.
  • HDCR.TDOSA and MDCR_EL2.TDOSA, for accesses to powerdown debug registers, using AArch32 state, trapped to EL2.
  • HDCR.TDA and MDCR_EL2.TDA, for accesses to other debug registers, using AArch32 state, trapped to EL2.
  • CPTR_EL3.TTA, for accesses to trace registers using AArch32, trapped to EL3.
  • MDCR_EL3.TDOSA, for accesses to powerdown debug registers using AArch32, trapped to EL3.
  • MDCR_EL3.TDA, for accesses to other debug registers, using AArch32, trapped to EL3.

The following fields describe configuration settings for generating exceptions from a VMSR or VMRS access, that are reported using EC value 0b001000:

ISS encoding for an exception from an LD64B or ST64B* instruction

2423222120191817161514131211109876543210
ISS

ISS, bits [24:0]

ISSMeaningApplies when
0b0000000000000000000000000

ST64BV instruction trapped.

When FEAT_LS64_V is implemented
0b0000000000000000000000001

ST64BV0 instruction trapped.

When FEAT_LS64_ACCDATA is implemented
0b0000000000000000000000010

LD64B or ST64B instruction trapped.

When FEAT_LS64 is implemented

All other values are reserved.

ISS encoding for an exception from an MCRR or MRRC access

2423222120191817161514131211109876543210
CVCONDOpc1RES0Rt2RtCRmDirection

CV, bit [24]

Condition code valid.

CVMeaning
0b0

The COND field is not valid.

0b1

The COND field is valid.

For exceptions taken from AArch64, CV is set to 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether CV is set to 1 or set to 0. See the description of the COND field for more information.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

COND, bits [23:20]

For exceptions taken from AArch64, this field is set to 0b1110.

The condition code for the trapped instruction. This field is valid only for exceptions taken from AArch32, and only when the value of CV is 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1 and:
    • If the instruction is conditional, COND is set to the condition code field value from the instruction.
    • If the instruction is unconditional, COND is set to 0b1110.
  • A conditional A32 instruction that is known to pass its condition code check can be presented either:
    • With COND set to 0b1110, the value for unconditional.
    • With the COND value held in the instruction.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether:
    • CV is set to 0 and COND is set to an UNKNOWN value. Software must examine the SPSR.IT field to determine the condition, if any, of the T32 instruction.
    • CV is set to 1 and COND is set to the condition code for the condition that applied to the instruction.
  • For an implementation that, for both A32 and T32 instructions, takes an exception on a trapped conditional instruction only if the instruction passes its condition code check, these definitions mean that when CV is set to 1 it is IMPLEMENTATION DEFINED whether the COND field is set to 0b1110, or to the value of any condition that applied to the instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Opc1, bits [19:16]

The Opc1 value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bit [15]

Reserved, RES0.

Rt2, bits [14:10]

The Rt2 value from the issued instruction, the second general-purpose register used for the transfer.

If the Rt2 value is not 0b1111, then the reported value gives the AArch64 view of the register. Otherwise, if the Rt2 value is 0b1111:

  • If the instruction that generated the exception is not UNPREDICTABLE, then the register specifier takes the value 0b11111.

  • If the instruction that generated the exception is UNPREDICTABLE, then the register specifier takes an UNKNOWN value, which is restricted to either:

    • The AArch64 view of one of the registers that could have been used in AArch32 state at the Exception level that the instruction was executed at.

    • The value 0b11111.

See 'Mapping of the general-purpose registers between the Execution states'.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Rt, bits [9:5]

The Rt value from the issued instruction, the first general-purpose register used for the transfer.

If the Rt value is not 0b1111, then the reported value gives the AArch64 view of the register. Otherwise, if the Rt value is 0b1111:

  • If the instruction that generated the exception is not UNPREDICTABLE, then the register specifier takes the value 0b11111.

  • If the instruction that generated the exception is UNPREDICTABLE, then the register specifier takes an UNKNOWN value, which is restricted to either:

    • The AArch64 view of one of the registers that could have been used in AArch32 state at the Exception level that the instruction was executed at.

    • The value 0b11111.

See 'Mapping of the general-purpose registers between the Execution states'.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

CRm, bits [4:1]

The CRm value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Direction, bit [0]

Indicates the direction of the trapped instruction.

DirectionMeaning
0b0

Write to System register space. MCRR instruction.

0b1

Read from System register space. MRRC instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from an MCRR or MRRC access

The following fields describe configuration settings for generating exceptions from an MCRR or MRRC access using coproc 0b1111, that are reported using EC value 0b000100:

  • CNTKCTL_EL1.{EL0PTEN, EL0VTEN, EL0PCTEN, EL0VCTEN}, for accesses to the Generic Timer Registers from EL0 using AArch32 state, trapped to EL1 or EL2.
  • PMUSERENR_EL0.{CR, EN}, for accesses to Performance Monitor registers from EL0 using AArch32 state, trapped to EL1 or EL2.
  • AMUSERENR_EL0.{EN}, for accesses to Activity Monitors registers AMEVCNTR0<n> and AMEVCNTR1<n> from EL0 using AArch32 state, trapped to EL1 or EL2.
  • HCR.{TRVM, TVM} and HCR_EL2.{TRVM, TVM}, for accesses to virtual memory control registers from EL1 using AArch32 state, trapped to EL2.
  • HCR2.TERR, for Non-secure accesses to error record registers at EL1 using AArch32 state, trapped to EL2.
  • HSTR.T<n> and HSTR_EL2.T<n>, for accesses to System registers using AArch32 state, trapped to EL2.
  • CNTHCTL.{PL1PCEN, PL1PCTEN} and CNTHCTL_EL2.{EL1PCEN, EL1PCTEN}, for accesses to the Generic Timer registers from EL0 and EL1 using AArch32 state, trapped to EL2.
  • HDCR.TPM and MDCR_EL2.{TPM, TPMCR}, for accesses to Performance Monitor registers from EL0 and EL1 using AArch32 state, trapped to EL2.
  • HCPTR.TAM and CPTR_EL2.TAM, for accesses to Activity Monitors registers AMEVCNTR0<n> and AMEVCNTR1<n> from EL0 and EL1 using AArch32 state, trapped to EL2.
  • MDCR_EL3.TPM, for accesses to Performance Monitor registers from EL0, EL1 and EL2 using AArch32 state, trapped to EL3.
  • CPTR_EL3.TAM, for accesses to Activity Monitors registers from EL0, EL1 and EL2 using AArch32 state, trapped to EL3.
  • If FEAT_FGT is implemented, HDFGRTR_EL2.PMCCNTR_EL0 for MRRC access and HDFGWTR_EL2.PMCCNTR_EL0 for MCRR access to PMCCNTR at EL0, trapped to EL2.

The following fields describe configuration settings for generating exceptions from an MCRR or MRRC access using coproc 0b1110, that are reported using EC value 0b001100:

  • MDSCR_EL1.TDCC, for accesses to the Debug ROM registers DBGDSAR and DBGDRAR at EL0 using AArch32 state, trapped to EL1 or EL2.
  • HDCR.TDRA and MDCR_EL2.TDRA, for accesses to Debug ROM registers DBGDRAR and DBGDSAR using AArch32, trapped to EL2.
  • MDCR_EL3.TDA, for accesses to debug registers, using AArch32, trapped to EL3.
  • CPACR_EL1.TTA for accesses to trace registers using AArch32, trapped to EL1 or EL2.
  • HCPTR.TTA and CPTR_EL2.TTA, for accesses to trace registers using AArch32, trapped to EL2.
  • CPTR_EL3.TTA, for accesses to trace registers using AArch32, trapped to EL3.
Note

If the Armv8-A architecture is implemented with an ETMv4 implementation, MCRR and MRRC accesses to trace registers are UNDEFINED and the resulting exception is higher priority than an exception due to these traps.

ISS encoding for an exception from an LDC or STC instruction

2423222120191817161514131211109876543210
CVCONDimm8RES0RnOffsetAMDirection

CV, bit [24]

Condition code valid.

CVMeaning
0b0

The COND field is not valid.

0b1

The COND field is valid.

For exceptions taken from AArch64, CV is set to 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether CV is set to 1 or set to 0. See the description of the COND field for more information.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

COND, bits [23:20]

For exceptions taken from AArch64, this field is set to 0b1110.

The condition code for the trapped instruction. This field is valid only for exceptions taken from AArch32, and only when the value of CV is 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1 and:
    • If the instruction is conditional, COND is set to the condition code field value from the instruction.
    • If the instruction is unconditional, COND is set to 0b1110.
  • A conditional A32 instruction that is known to pass its condition code check can be presented either:
    • With COND set to 0b1110, the value for unconditional.
    • With the COND value held in the instruction.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether:
    • CV is set to 0 and COND is set to an UNKNOWN value. Software must examine the SPSR.IT field to determine the condition, if any, of the T32 instruction.
    • CV is set to 1 and COND is set to the condition code for the condition that applied to the instruction.
  • For an implementation that, for both A32 and T32 instructions, takes an exception on a trapped conditional instruction only if the instruction passes its condition code check, these definitions mean that when CV is set to 1 it is IMPLEMENTATION DEFINED whether the COND field is set to 0b1110, or to the value of any condition that applied to the instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

imm8, bits [19:12]

The immediate value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits [11:10]

Reserved, RES0.

Rn, bits [9:5]

The Rn value from the issued instruction, the general-purpose register used for the transfer.

If the Rn value is not 0b1111, then the reported value gives the AArch64 view of the register. Otherwise, if the Rn value is 0b1111:

  • If the instruction that generated the exception is not UNPREDICTABLE, then the register specifier takes the value 0b11111.

  • If the instruction that generated the exception is UNPREDICTABLE, then the register specifier takes an UNKNOWN value, which is restricted to either:

    • The AArch64 view of one of the registers that could have been used in AArch32 state at the Exception level that the instruction was executed at.

    • The value 0b11111.

See 'Mapping of the general-purpose registers between the Execution states'.

This field is valid only when AM[2] is 0, indicating an immediate form of the LDC or STC instruction. When AM[2] is 1, indicating a literal form of the LDC or STC instruction, this field is UNKNOWN.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Offset, bit [4]

Indicates whether the offset is added or subtracted:

OffsetMeaning
0b0

Subtract offset.

0b1

Add offset.

This bit corresponds to the U bit in the instruction encoding.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

AM, bits [3:1]

Addressing mode. The permitted values of this field are:

AMMeaning
0b000

Immediate unindexed.

0b001

Immediate post-indexed.

0b010

Immediate offset.

0b011

Immediate pre-indexed.

0b100

For a trapped STC instruction or a trapped T32 LDC instruction this encoding is reserved.

0b110

For a trapped STC instruction, this encoding is reserved.

The values 0b101 and 0b111 are reserved. The effect of programming this field to a reserved value is that behavior is CONSTRAINED UNPREDICTABLE, as described in 'Reserved values in System and memory-mapped registers and translation table entries'.

Bit [2] in this subfield indicates the instruction form, immediate or literal.

Bits [1:0] in this subfield correspond to the bits {P, W} in the instruction encoding.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Direction, bit [0]

Indicates the direction of the trapped instruction.

DirectionMeaning
0b0

Write to memory. STC instruction.

0b1

Read from memory. LDC instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from an LDC or STC instruction

The following fields describe the configuration settings from an LDC or STC access for the traps that are reported using EC value 0b000110:

ISS encoding for an exception from an access to SVE, Advanced SIMD or floating-point functionality, resulting from the FPEN and TFP traps

2423222120191817161514131211109876543210
CVCONDRES0

The accesses covered by this trap include:

  • Execution of SVE or Advanced SIMD and floating-point instructions.
  • Accesses to the Advanced SIMD and floating-point System registers.
  • Execution of SME instructions.

For an implementation that does not include either SVE or support for Advanced SIMD and floating-point, the exception is reported using the EC value 0b000000.

CV, bit [24]

Condition code valid.

CVMeaning
0b0

The COND field is not valid.

0b1

The COND field is valid.

For exceptions taken from AArch64, CV is set to 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether CV is set to 1 or set to 0. See the description of the COND field for more information.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

COND, bits [23:20]

For exceptions taken from AArch64, this field is set to 0b1110.

The condition code for the trapped instruction. This field is valid only for exceptions taken from AArch32, and only when the value of CV is 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1 and:
    • If the instruction is conditional, COND is set to the condition code field value from the instruction.
    • If the instruction is unconditional, COND is set to 0b1110.
  • A conditional A32 instruction that is known to pass its condition code check can be presented either:
    • With COND set to 0b1110, the value for unconditional.
    • With the COND value held in the instruction.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether:
    • CV is set to 0 and COND is set to an UNKNOWN value. Software must examine the SPSR.IT field to determine the condition, if any, of the T32 instruction.
    • CV is set to 1 and COND is set to the condition code for the condition that applied to the instruction.
  • For an implementation that, for both A32 and T32 instructions, takes an exception on a trapped conditional instruction only if the instruction passes its condition code check, these definitions mean that when CV is set to 1 it is IMPLEMENTATION DEFINED whether the COND field is set to 0b1110, or to the value of any condition that applied to the instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits [19:0]

Reserved, RES0.

Additional information for the ISS encoding for an exception from an access to SVE, Advanced SIMD or floating-point functionality, resulting from the FPEN and TFP traps

The following fields describe the configuration settings for the traps that are reported using EC value 0b000111:

  • HCPTR.{TCP10, TCP11}, for Non-secure accesses to Advanced SIMD and floating-point registers and instructions, trapped to EL2.
  • HCPTR.TASE, for Non-secure accesses to Advanced SIMD functionality, trapped to EL2.
  • CPACR_EL1.FPEN, for accesses to SIMD and floating-point registers trapped to EL1.
  • CPTR_EL2.FPEN and CPTR_EL2.TFP, for accesses to SIMD and floating-point registers trapped to EL2.
  • CPTR_EL3.TFP, for accesses to SIMD and floating-point registers trapped to EL3.

ISS encoding for an exception from an access to SVE functionality, resulting from CPACR_EL1.ZEN, CPTR_EL2.ZEN, CPTR_EL2.TZ, or CPTR_EL3.EZ

2423222120191817161514131211109876543210
RES0

The accesses covered by this trap include:

  • Execution of SVE instructions when the PE is not in Streaming SVE mode.
  • Accesses to the SVE System registers, ZCR_ELx.

For an implementation that does not include SVE, the exception is reported using the EC value 0b000000.

Bits [24:0]

Reserved, RES0.

Additional information for the ISS encoding for an exception from an access to SVE functionality, resulting from CPACR_EL1.ZEN, CPTR_EL2.ZEN, CPTR_EL2.TZ, or CPTR_EL3.EZ

The following fields describe the configuration settings for the traps that are reported using EC value 0b011001:

  • CPACR_EL1.ZEN, for execution of SVE instructions and accesses to SVE registers at EL0 or EL1, trapped to EL1.
  • CPTR_EL2.ZEN and CPTR_EL2.TZ, for execution of SVE instructions and accesses to SVE registers at EL0, EL1, or EL2, trapped to EL2.
  • CPTR_EL3.EZ, for execution of SVE instructions and accesses to SVE registers from all Exception levels, trapped to EL3.

ISS encoding for a PMU exception

2423222120191817161514131211109876543210
RES0SYNC

Bits [24:1]

Reserved, RES0.

SYNC, bit [0]

Indicates whether the exception was taken synchronously or asynchronously.

SYNCMeaningApplies when
0b0

The exception was taken asynchronously because an overflow status flag was set.

0b1

The exception was taken synchronously because PSTATE.PPEND was set.

When FEAT_SEBEP is implemented

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

ISS encoding for an exception from an Illegal Execution state, or a PC or SP alignment fault

2423222120191817161514131211109876543210
RES0

Bits [24:0]

Reserved, RES0.

Additional information for the ISS encoding for an exception from an Illegal Execution state, or a PC or SP alignment fault

There are no configuration settings for generating Illegal Execution state exceptions and PC alignment fault exceptions. For more information about PC alignment fault exceptions, see 'PC alignment checking'.

'SP alignment checking' describes the configuration settings for generating SP alignment fault exceptions.

ISS encoding for an exception from the Memory Copy and Memory Set instructions

2423222120191817161514131211109876543210
MemInstisSETGOptionsFromEpilogueWrongOptionOptionARES0destregsrcregsizereg

MemInst, bit [24]

Indicates the memory instruction class causing the exception.

MemInstMeaning
0b0

CPYFE*, CPYFM*, CPYE*, and CPYM* instructions.

0b1

SETE*, SETM*, SETGE*, and SETGM* instructions.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

isSETG, bit [23]

Indicates whether the instruction belongs to SETGM* or SETGE* class of instruction.

isSETGMeaning
0b0

Not a SETGM* or SETGE* instruction.

0b1

SETGM* or SETGE* instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Options, bits [22:19]

Options : the Options field of the instruction.

For Memory Copy instructions, bits[22:19] forms the Options field, which holds the bits[15:12] of the instruction.

For Memory Set instructions:

  • Bits[22:21] are RES0.
  • Bits[20:19] form the Options field, which holds the bits[13:12] of the instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

FromEpilogue, bit [18]

Indicates whether the instruction belongs to the epilogue class of Memory Copy or Memory Set instructions.

FromEpilogueMeaning
0b0

Not an epilogue instruction.

0b1

CPYE*, CPYFE*, SETE*, or SETGE* instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

WrongOption, bit [17]

Algorithm option.

WrongOptionMeaning
0b0

WrongOption is false.

0b1

WrongOption is true.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

OptionA, bit [16]

Algorithm type indicated by the PSTATE.C bit.

OptionAMeaning
0b0

OptionB indicated by PSTATE.C is 0.

0b1

OptionA indicated by PSTATE.C is 1.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bit [15]

Reserved, RES0.

destreg, bits [14:10]

The destination register value from the issued instruction, containing the destination address.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

srcreg, bits [9:5]

The source register value from the issued instruction, containing either the source address or the source data.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

sizereg, bits [4:0]

The size register value from the issued instruction, containing the number of bytes to be transfered or set.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

ISS encoding for an exception from HVC or SVC instruction execution

2423222120191817161514131211109876543210
RES0imm16

Bits [24:16]

Reserved, RES0.

imm16, bits [15:0]

The value of the immediate field from the HVC or SVC instruction.

For an HVC instruction, and for an A64 SVC instruction, this is the value of the imm16 field of the issued instruction.

For an A32 or T32 SVC instruction:

  • If the instruction is unconditional, then:
    • For the T32 instruction, this field is zero-extended from the imm8 field of the instruction.
    • For the A32 instruction, this field is the bottom 16 bits of the imm24 field of the instruction.
  • If the instruction is conditional, this field is UNKNOWN.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from HVC or SVC instruction execution

In AArch32 state, the HVC instruction is unconditional, and a conditional SVC instruction generates an exception only if it passes its condition code check. Therefore, the syndrome information for these exceptions does not require conditionality information.

For T32 and A32 instructions, see 'SVC' and 'HVC'.

For A64 instructions, see 'SVC' and 'HVC'.

If FEAT_FGT is implemented, HFGITR_EL2.{SVC_EL1, SVC_EL0} control fine-grained traps on SVC execution.

ISS encoding for an exception from SMC instruction execution in AArch32 state

2423222120191817161514131211109876543210
CVCONDCCKNOWNPASSRES0

For an SMC instruction that completes normally and generates an exception that is taken to EL3, the ISS encoding is RES0.

For an SMC instruction that is trapped to EL2 from EL1 because HCR_EL2.TSC is 1, the ISS encoding is as shown in the diagram.

CV, bit [24]

Condition code valid.

CVMeaning
0b0

The COND field is not valid.

0b1

The COND field is valid.

For exceptions taken from AArch64, CV is set to 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether CV is set to 1 or set to 0. See the description of the COND field for more information.

This field is valid only if CCKNOWNPASS is 1, otherwise it is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

COND, bits [23:20]

For exceptions taken from AArch64, this field is set to 0b1110.

The condition code for the trapped instruction. This field is valid only for exceptions taken from AArch32, and only when the value of CV is 1.

For exceptions taken from AArch32:

  • When an A32 instruction is trapped, CV is set to 1 and:
    • If the instruction is conditional, COND is set to the condition code field value from the instruction.
    • If the instruction is unconditional, COND is set to 0b1110.
  • A conditional A32 instruction that is known to pass its condition code check can be presented either:
    • With COND set to 0b1110, the value for unconditional.
    • With the COND value held in the instruction.
  • When a T32 instruction is trapped, it is IMPLEMENTATION DEFINED whether:
    • CV is set to 0 and COND is set to an UNKNOWN value. Software must examine the SPSR.IT field to determine the condition, if any, of the T32 instruction.
    • CV is set to 1 and COND is set to the condition code for the condition that applied to the instruction.
  • For an implementation that, for both A32 and T32 instructions, takes an exception on a trapped conditional instruction only if the instruction passes its condition code check, these definitions mean that when CV is set to 1 it is IMPLEMENTATION DEFINED whether the COND field is set to 0b1110, or to the value of any condition that applied to the instruction.

This field is valid only if CCKNOWNPASS is 1, otherwise it is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

CCKNOWNPASS, bit [19]

Indicates whether the instruction might have failed its condition code check.

CCKNOWNPASSMeaning
0b0

The instruction was unconditional, or was conditional and passed its condition code check.

0b1

The instruction was conditional, and might have failed its condition code check.

Note

In an implementation in which an SMC instruction that fails it code check is not trapped, this field can always return the value 0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits [18:0]

Reserved, RES0.

Additional information for the ISS encoding for an exception from SMC instruction execution in AArch32 state

HCR.TSC describes the configuration settings for trapping SMC instructions to EL2.

HCR_EL2.TSC describes the configuration settings for trapping SMC instructions to EL2.

ISS encoding for an exception from SMC instruction execution in AArch64 state

2423222120191817161514131211109876543210
RES0imm16

Bits [24:16]

Reserved, RES0.

imm16, bits [15:0]

The value of the immediate field from the issued SMC instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from SMC instruction execution in AArch64 state

The value of ISS[24:0] described here is used both:

  • When an SMC instruction is trapped from EL1 modes.
  • When an SMC instruction is not trapped, so completes normally and generates an exception that is taken to EL3.

HCR_EL2.TSC describes the configuration settings for trapping SMC from EL1 modes.

ISS encoding for an exception from MSR, MRS, or System instruction execution in AArch64 state

2423222120191817161514131211109876543210
RES0Op0Op2Op1CRnRtCRmDirection

Bits [24:22]

Reserved, RES0.

Op0, bits [21:20]

The Op0 value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Op2, bits [19:17]

The Op2 value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Op1, bits [16:14]

The Op1 value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

CRn, bits [13:10]

The CRn value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Rt, bits [9:5]

The Rt value from the issued instruction, the general-purpose register used for the transfer.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

CRm, bits [4:1]

The CRm value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Direction, bit [0]

Indicates the direction of the trapped instruction.

DirectionMeaning
0b0

Write access, including MSR instructions.

0b1

Read access, including MRS instructions.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from MSR, MRS, or System instruction execution in AArch64 state

For exceptions caused by System instructions, see 'System instructions' subsection of 'Branches, exception generating and System instructions' for the encoding values returned by an instruction.

The following fields describe configuration settings for generating the exception that is reported using EC value 0b011000:

  • If FEAT_TIDCP1 is implemented, SCTLR_EL1.TIDCP, for EL0 accesses to IMPLEMENTATION DEFINED functionality using AArch64 state, MSR or MRS access trapped to EL1.
  • SCTLR_EL1.UCI, for execution of cache maintenance instructions using AArch64 state, MSR or MRS access trapped to EL1 or EL2.
  • SCTLR_EL1.UCT, for accesses to CTR_EL0 using AArch64 state, MSR or MRS access trapped to EL1 or EL2.
  • SCTLR_EL1.DZE, for execution of DC ZVA instructions using AArch64 state, MSR or MRS access trapped to EL1 or EL2.
  • SCTLR_EL1.UMA, for accesses to the PSTATE interrupt masks using AArch64 state, MSR or MRS access trapped to EL1 or EL2.
  • CPACR_EL1.TTA, for accesses to the trace registers using AArch64 state, MSR or MRS access trapped to EL1 or EL2.
  • MDSCR_EL1.TDCC, for accesses to the Debug Communications Channel (DCC) registers using AArch64 state, MSR or MRS access trapped to EL1 or EL2.
  • If FEAT_FGT is implemented, MDCR_EL2.TDCC for accesses to the DCC registers at EL0 and EL1 trapped to EL2, and MDCR_EL3.TDCC for accesses to the DCC registers at EL0, EL1, and EL2 trapped to EL3.
  • CNTKCTL_EL1.{EL0PTEN, EL0VTEN, EL0PCTEN, EL0VCTEN} accesses to the Generic Timer registers using AArch64 state, MSR or MRS access trapped to EL1 or EL2.
  • PMUSERENR_EL0, for accesses to the Performance Monitor registers using AArch64 state, MSR or MRS access trapped to EL1 or EL2.
  • AMUSERENR_EL0.EN, for accesses to Activity Monitors registers using AArch64 state, MSR or MRS access trapped to EL1 or EL2.
  • HCR_EL2.{TRVM, TVM}, for accesses to virtual memory control registers using AArch64 state, MSR or MRS access trapped to EL2.
  • HCR_EL2.TDZ, for execution of DC ZVA instructions using AArch64 state, MSR or MRS access trapped to EL2.
  • HCR_EL2.TTLB, for execution of TLB maintenance instructions using AArch64 state, MSR or MRS access trapped to EL2.
  • HCR_EL2.{TSW, TPC, TPU}, for execution of cache maintenance instructions using AArch64 state, MSR or MRS access trapped to EL2.
  • HCR_EL2.TACR, for accesses to the Auxiliary Control Register, ACTLR_EL1, using AArch64 state, MSR or MRS access trapped to EL2.
  • HCR_EL2.TIDCP, for accesses to lockdown, DMA, and TCM operations using AArch64 state, MSR or MRS access trapped to EL2.
  • If FEAT_TIDCP1 is implemented, SCTLR_EL2.TIDCP, for EL0 accesses to IMPLEMENTATION DEFINED functionality using AArch64 state, MSR or MRS access trapped to EL2.
  • HCR_EL2.{TID1, TID2, TID3}, for accesses to ID group 1, ID group 2 or ID group 3 registers, using AArch64 state, MSR or MRS access trapped to EL2.
  • CPTR_EL2.TCPAC, for accesses to CPACR_EL1, using AArch64 state, MSR or MRS access trapped to EL2.
  • CPTR_EL2.TTA, for accesses to the trace registers, using AArch64 state, MSR or MRS access trapped to EL2.
  • MDCR_EL2.TTRF, for accesses to the trace filter control register, TRFCR_EL1, using AArch64 state, MSR or MRS access trapped to EL2.
  • MDCR_EL2.TDRA, for accesses to Debug ROM registers, using AArch64 state, MSR or MRS access trapped to EL2.
  • MDCR_EL2.TDOSA, for accesses to powerdown debug registers using AArch64 state, MSR or MRS access trapped to EL2.
  • CNTHCTL_EL2.{EL1PCEN, EL1PCTEN}, for accesses to the Generic Timer registers using AArch64 state, MSR or MRS access trapped to EL2.
  • MDCR_EL2.TDA, for accesses to debug registers using AArch64 state, MSR or MRS access trapped to EL2.
  • MDCR_EL2.{TPM, TPMCR}, for accesses to Performance Monitor registers, using AArch64 state, MSR or MRS access trapped to EL2.
  • CPTR_EL2.TAM, for accesses to Activity Monitors registers, using AArch64 state, MSR or MRS access trapped to EL2.
  • HCR_EL2.APK, for accesses to Pointer authentication key registers, using AArch64 state, MSR or MRS access trapped to EL2.
  • HCR_EL2.{NV, NV1}, for Nested virtualization register access, using AArch64 state, MSR or MRS access, trapped to EL2.
  • HCR_EL2.AT, for execution of AT S1E* instructions, using AArch64 state, MSR or MRS access, trapped to EL2.
  • HCR_EL2.{TERR, FIEN}, for accesses to RAS registers, using AArch64 state, MSR or MRS access, trapped to EL2.
  • SCR_EL3.APK, for accesses to Pointer authentication key registers, using AArch64 state, MSR or MRS access trapped to EL3.
  • SCR_EL3.ST, for accesses to the Counter-timer Physical Secure timer registers, using AArch64 state, MSR or MRS access trapped to EL3.
  • SCR_EL3.{TERR, FIEN}, for accesses to RAS registers, using AArch64 state, MSR or MRS access trapped to EL3.
  • CPTR_EL3.TCPAC, for accesses to CPTR_EL2 and CPACR_EL1 using AArch64 state, MSR or MRS access trapped to EL3.
  • CPTR_EL3.TTA, for accesses to the trace registers, using AArch64 state, MSR or MRS access trapped to EL3.
  • MDCR_EL3.TTRF, for accesses to the trace filter control registers, TRFCR_EL1 and TRFCR_EL2, using AArch64 state, MSR or MRS access trapped to EL3.
  • MDCR_EL3.TDA, for accesses to debug registers, using AArch64 state, MSR or MRS access trapped to EL3.
  • MDCR_EL3.TDOSA, for accesses to powerdown debug registers, using AArch64 state, MSR or MRS access trapped to EL3.
  • MDCR_EL3.TPM, for accesses to Performance Monitor registers, using AArch64 state, MSR or MRS access trapped to EL3.
  • If FEAT_SPE is implemented:
    • MDCR_EL3.NSPB for accesses to Statistical Profiling and Profiling Buffer control registers, using AArch64 state, MSR or MRS access at EL1 and EL2 trapped to EL3.
    • MDCR_EL2.TPMS for accesses to SPE registers, using AArch64 state, MSR or MRS access at EL1 trapped to EL2.
  • CPTR_EL3.TAM, for accesses to Activity Monitors registers, using AArch64 state, MSR or MRS access, trapped to EL3.
  • If FEAT_EVT is implemented, the following registers control traps for EL1 and EL0 Cache controls that use this EC value:
    • HCR_EL2.{TTLBOS, TTLBIS, TICAB, TOCU, TID4}.
    • HCR2.{TTLBIS, TICAB, TOCU, TID4}.
  • If FEAT_FGT is implemented:
    • SCR_EL3.FGTEn, for accesses to the fine-grained trap registers, MSR or MRS access at EL2 trapped to EL3.
    • HFGRTR_EL2 for reads and HFGWTR_EL2 for writes of registers, using AArch64 state, MSR or MRS access at EL0 and EL1 trapped to EL2.
    • HFGITR_EL2 for execution of system instructions, MSR or MRS access trapped to EL2.
    • HDFGRTR_EL2 for reads and HDFGWTR_EL2 for writes of registers, using AArch64 state, MSR or MRS access at EL0 and EL1 state trapped to EL2.
    • HAFGRTR_EL2 for reads of Activity Monitor counters, using AArch64 state, MRS access at EL0 and EL1 trapped to EL2.
  • If FEAT_RNG_TRAP is implemented, SCR_EL3.TRNDR for reads of RNDR and RNDRRS using AArch64 state, MRS access trapped to EL3.
  • If FEAT_SME is implemented:
  • If FEAT_FPMR is implemented:
    • SCTLR_EL1.EnFPM, for accesses to FPMR at EL0, trapped to EL1 or EL2.
    • SCTLR_EL2.EnFPM, for accesses to FPMR at EL0, trapped to EL2.
    • HCRX_EL2.EnFPM, for accesses to FPMR at EL0 and EL1, trapped to EL2.
    • SCR_EL3.EnFPM, for accesses to FPMR at EL0, EL1, and EL2, trapped to EL3.
  • If FEAT_NMI is implemented, HCRX_EL2.TALLINT, for MSR writes of ALLINT at EL1, trapped to EL2.
  • If FEAT_FGT2 is implemented:
    • SCR_EL3.FGTEn2, for accesses to the fine-grained trap registers, MSR or MRS access at EL2 trapped to EL3.
    • HFGRTR2_EL2 for reads and HFGWTR2_EL2 for writes of registers, using AArch64 state, using MSR or MRS access at EL1 trapped to EL2.
    • HDFGRTR2_EL2 for reads and HDFGWTR2_EL2 for writes of registers, using AArch64 state, using MSR or MRS access at EL0 and EL1 trapped to EL2.
    • HFGITR2_EL2 for execution of system instructions, MSR or MRS access trapped to EL2.
  • If FEAT_ITE is implemented, MDCR_EL3.EnITE, for accesses to Instrumentation trace registers, using AArch64 state, MSR or MRS access, trapped to EL3.
  • If FEAT_MEC is implemented, SCR_EL3.MECEn, for accesses to MECID registers at EL2, trapped to EL3.
  • If FEAT_SPE_FDS is implemented, MDCR_EL3.EnPMS3 for accesses to SPE registers, using AArch64 state, MSR or MRS access at EL1 and EL2 trapped to EL3.
  • If FEAT_RASv2 is implemented, SCR_EL3.TWERR, for accesses to Error Record registers, MSR or MRS access at EL1 and EL2 trapped to EL3.
  • If FEAT_Debugv8p9 is implemented, MDCR_EL3.EBWE for accesses of MDSELR_EL1, using AArch64 state, MRS or MSR access at EL2 and EL1 trapped to EL3.
  • If FEAT_PMUv3p9, FEAT_SPMU, FEAT_EBEP, or FEAT_PMUv3_SS is implemented, MDCR_EL3.EnPM2, for accesses to PMU registers, using AArch64 state, MSR or MRS access at EL2, EL1, and EL0, trapped to EL3.
  • If FEAT_PMUv3_SS is implemented, MDCR_EL3.EnPMSS, for accesses to PMU Snapshot registers, using AArch64 state, MSR or MRS access at EL2 and EL1 trapped to EL3.
  • If FEAT_THE is implemented, SCR_EL3.RCWMASKEn for accesses to RCWMASK_EL1 and RCWSMASK_EL1, using AArch64 state, MSR or MRS access at EL2 and EL1 trapped to EL3.
  • If FEAT_AIE is implemented, SCR_EL3.AIEn for accesses to Extended Memory Attribute registers, MSR or MRS access at EL2 and EL1 trapped to EL3.
  • If FEAT_S1PIEx, FEAT_S2PIEx, FEAT_S1POEx, or FEAT_S2POEx is implemented, SCR_EL3.PIEn for accesses to Permission Indirection, Overlay registers, MSR or MRS access at EL2, EL1 and EL0 trapped to EL3.

ISS encoding for an exception from MSRR, MRRS, or 128-bit System instruction execution in AArch64 state

2423222120191817161514131211109876543210
RES0Op0Op2Op1CRnRtRES0CRmDirection

Bits [24:22]

Reserved, RES0.

Op0, bits [21:20]

The Op0 value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Op2, bits [19:17]

The Op2 value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Op1, bits [16:14]

The Op1 value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

CRn, bits [13:10]

The CRn value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Rt, bits [9:6]

The Rt value from the issued instruction, the general-purpose register used for the transfer.

Note

This value represents register pair of X[Rt:0], X[Rt:1].

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bit [5]

Reserved, RES0.

CRm, bits [4:1]

The CRm value from the issued instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Direction, bit [0]

Indicates the direction of the trapped instruction.

DirectionMeaning
0b0

Write access, MSRR instructions.

0b1

Read access, MRRS instructions.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from MSRR, MRRS, or 128-bit System instruction execution in AArch64 state

The following fields describe configuration settings for generating exceptions from an MSRR or MRRS access that are reported using EC value 0b010100:

  • If FEAT_FGT is implemented:
    • HFGRTR_EL2 for reads and HFGWTR_EL2 for writes of registers, using AArch64 state, accesses at EL1 trapped to EL2.
  • If FEAT_FGT2 is implemented:
  • If FEAT_SYSREG128 is implemented:
    • SCTLR2_EL1.EnIDCP128 for accesses to 128-bit IMPLEMENTATION DEFINED System registers, accesses at EL0 trapped to EL1.
    • SCTLR2_EL2.EnIDCP128 for accesses to 128-bit IMPLEMENTATION DEFINED System registers, accesses at EL0 trapped to EL2.
    • HCRX_EL2.EnIDCP128 for accesses to 128-bit IMPLEMENTATION DEFINED System registers, accesses at EL1 and EL0 trapped to EL2.
    • SCR_EL3.EnIDCP128 for accesses to 128-bit IMPLEMENTATION DEFINED System registers, accesses at EL2, EL1, and EL0 trapped to EL3.
  • If FEAT_D128 is implemented:
    • HCR_EL2.{TRVM, TVM} for accesses to TTBR0_EL1 and TTBR1_EL1, accesses at EL1 and EL0 trapped to EL2.
    • HCRX_EL2.D128En for accesses to 128-bit IMPLEMENTATION DEFINED System registers, accesses at EL1 trapped to EL2.
    • SCR_EL3.D128En for accesses to 128-bit IMPLEMENTATION DEFINED System registers, accesses at EL2 and EL1 trapped to EL3.
  • If FEAT_THE is implemented, SCR_EL3.RCWMASKEn for accesses to RCWMASK_EL1 and RCWSMASK_EL1, using AArch64 state, accesses at EL2 and EL1 trapped to EL3.

ISS encoding for an exception from an Instruction Abort

2423222120191817161514131211109876543210
RES0TopLevelRES0PFVRES0SETFnVEARES0S1PTWRES0IFSC

When FEAT_THE is implemented, if a memory access generates an Instruction Abort for AssuredOnly, the ISS2 encoding for an exception from an Instruction Abort includes further information about the exception.

When FEAT_S1POE or FEAT_S2POE is implemented, if a memory access generates an Instruction Abort due to a Permission fault, the ISS2 encoding for an exception from an Instruction Abort includes further information about the exception.

Bits [24:22]

Reserved, RES0.

TopLevel, bit [21]
When FEAT_THE is implemented:

Indicates if the fault was due to TopLevel.

TopLevelMeaning
0b0

Fault is not due to TopLevel.

0b1

Fault is due to TopLevel.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bits [20:15]

Reserved, RES0.

PFV, bit [14]
When FEAT_PFAR is implemented:

FAR Valid. Describes whether the PFAR_EL2 is valid.

PFVMeaning
0b0

PFAR_EL2 is UNKNOWN.

0b1

PFAR_EL2 is valid.

This field is valid only if the IFSC code is 0b10000, 0b01001x, or 0b0101xx.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bit [13]

Reserved, RES0.

SET, bits [12:11]
When FEAT_RAS is implemented and IFSC == 0b010000:

Synchronous Error Type. When IFSC is 0b010000, describes the PE error state after taking the Instruction Abort exception.

SETMeaningApplies when
0b00

Recoverable state (UER).

0b10

Uncontainable (UC).

When FEAT_RASv2 is not implemented
0b11

Restartable state (UEO).

All other values are reserved.

Note

Software can use this information to determine what recovery might be possible. Taking a synchronous External abort exception might result in a PE state that is not recoverable.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

FnV, bit [10]
When IFSC == 0b010000:

FAR not Valid, for a synchronous External abort other than a synchronous External abort on a translation table walk.

FnVMeaning
0b0

FAR is valid.

0b1

FAR is not valid, and holds an UNKNOWN value.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

EA, bit [9]

External abort type. This bit can provide an IMPLEMENTATION DEFINED classification of External aborts.

For any abort other than an External abort this bit returns a value of 0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bit [8]

Reserved, RES0.

S1PTW, bit [7]

For a stage 2 fault, indicates whether the fault was a stage 2 fault on an access made for a stage 1 translation table walk:

S1PTWMeaning
0b0

Fault not on a stage 2 translation for a stage 1 translation table walk.

0b1

Fault on the stage 2 translation of an access for a stage 1 translation table walk.

For any abort other than a stage 2 fault this bit is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bit [6]

Reserved, RES0.

IFSC, bits [5:0]

Instruction Fault Status Code.

IFSCMeaningApplies when
0b000000

Address size fault, level 0 of translation or translation table base register.

0b000001

Address size fault, level 1.

0b000010

Address size fault, level 2.

0b000011

Address size fault, level 3.

0b000100

Translation fault, level 0.

0b000101

Translation fault, level 1.

0b000110

Translation fault, level 2.

0b000111

Translation fault, level 3.

0b001001

Access flag fault, level 1.

0b001010

Access flag fault, level 2.

0b001011

Access flag fault, level 3.

0b001000

Access flag fault, level 0.

When FEAT_LPA2 is implemented
0b001100

Permission fault, level 0.

When FEAT_LPA2 is implemented
0b001101

Permission fault, level 1.

0b001110

Permission fault, level 2.

0b001111

Permission fault, level 3.

0b010000

Synchronous External abort, not on translation table walk or hardware update of translation table.

0b010010

Synchronous External abort on translation table walk or hardware update of translation table, level -2.

When FEAT_D128 is implemented
0b010011

Synchronous External abort on translation table walk or hardware update of translation table, level -1.

When FEAT_LPA2 is implemented
0b010100

Synchronous External abort on translation table walk or hardware update of translation table, level 0.

0b010101

Synchronous External abort on translation table walk or hardware update of translation table, level 1.

0b010110

Synchronous External abort on translation table walk or hardware update of translation table, level 2.

0b010111

Synchronous External abort on translation table walk or hardware update of translation table, level 3.

0b011000

Synchronous parity or ECC error on memory access, not on translation table walk.

When FEAT_RAS is not implemented
0b011011

Synchronous parity or ECC error on memory access on translation table walk or hardware update of translation table, level -1.

When FEAT_LPA2 is implemented and FEAT_RAS is not implemented
0b011100

Synchronous parity or ECC error on memory access on translation table walk or hardware update of translation table, level 0.

When FEAT_RAS is not implemented
0b011101

Synchronous parity or ECC error on memory access on translation table walk or hardware update of translation table, level 1.

When FEAT_RAS is not implemented
0b011110

Synchronous parity or ECC error on memory access on translation table walk or hardware update of translation table, level 2.

When FEAT_RAS is not implemented
0b011111

Synchronous parity or ECC error on memory access on translation table walk or hardware update of translation table, level 3.

When FEAT_RAS is not implemented
0b100010

Granule Protection Fault on translation table walk or hardware update of translation table, level -2.

When FEAT_D128 is implemented and FEAT_RME is implemented
0b100011

Granule Protection Fault on translation table walk or hardware update of translation table, level -1.

When FEAT_RME is implemented and FEAT_LPA2 is implemented
0b100100

Granule Protection Fault on translation table walk or hardware update of translation table, level 0.

When FEAT_RME is implemented
0b100101

Granule Protection Fault on translation table walk or hardware update of translation table, level 1.

When FEAT_RME is implemented
0b100110

Granule Protection Fault on translation table walk or hardware update of translation table, level 2.

When FEAT_RME is implemented
0b100111

Granule Protection Fault on translation table walk or hardware update of translation table, level 3.

When FEAT_RME is implemented
0b101000

Granule Protection Fault, not on translation table walk or hardware update of translation table.

When FEAT_RME is implemented
0b101001

Address size fault, level -1.

When FEAT_LPA2 is implemented
0b101010

Translation fault, level -2.

When FEAT_D128 is implemented
0b101011

Translation fault, level -1.

When FEAT_LPA2 is implemented
0b101100

Address Size fault, level -2.

When FEAT_D128 is implemented
0b110000

TLB conflict abort.

0b110001

Unsupported atomic hardware update fault.

When FEAT_HAFDBS is implemented

All other values are reserved.

For more information about the lookup level associated with a fault, see 'The lookup level associated with MMU faults'.

If the S1PTW bit is set, then the level refers the level of the stage2 translation that is translating a stage 1 translation walk.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

ISS encoding for an exception due to SME functionality

2423222120191817161514131211109876543210
RES0SMTC

The accesses covered by this trap include:

  • Execution of SME instructions.
  • Execution of SVE and Advanced SIMD instructions, when the PE is in Streaming SVE mode.
  • Direct accesses of SVCR, SMCR_EL1, SMCR_EL2, SMCR_EL3.

Bits [24:3]

Reserved, RES0.

SMTC, bits [2:0]

SME Trap Code. Identifies the reason for instruction trapping.

SMTCMeaningApplies when
0b000

Access to SME functionality trapped as a result of CPACR_EL1.SMEN, CPTR_EL2.SMEN, CPTR_EL2.TSM, or CPTR_EL3.ESM, that is not reported using EC 0b000000.

0b001

Advanced SIMD, SVE, or SVE2 instruction trapped because PSTATE.SM is 1.

0b010

SME instruction trapped because PSTATE.SM is 0.

0b011

SME instruction trapped because PSTATE.ZA is 0.

0b100

Access to the SME2 ZT0 register trapped as a result of SMCR_EL1.EZT0, SMCR_EL2.EZT0, or SMCR_EL3.EZT0.

When FEAT_SME2 is implemented

All other values are reserved.

Additional information for the ISS encoding for an exception due to SME functionality

The following fields describe the configuration settings for the traps that are reported using the EC value 0b011101:

  • CPACR_EL1.SMEN, for execution of SME instructions, SVE instructions when the PE is in Streaming SVE mode, and instructions that directly access SVCR and SMCR_EL1 System registers at EL1 and EL0, trapped to EL1 or EL2.
  • CPTR_EL2.SMEN and CPTR_EL2.TSM, for execution of SME instructions, SVE instructions when the PE is in Streaming SVE mode, and instructions that directly access SVCR, SMCR_EL1, SMCR_EL2 at EL2, EL1, or EL0, trapped to EL2.
  • CPTR_EL3.ESM, for execution of SME instructions, SVE instructions when the PE is in Streaming SVE mode, and instructions that directly access SVCR, SMCR_EL1, SMCR_EL2, SMCR_EL3 from all Exception levels and any Security state, trapped to EL3.
  • If FEAT_SME2 is implemented:
    • SMCR_EL1.EZT0, for accesses to ZT0 at EL1 and EL0, trapped to EL1 or EL2.
    • SMCR_EL2.EZT0, for accesses to ZT0 at EL2, EL1, and EL0, trapped to EL2.
    • SMCR_EL3.EZT0, for accesses to ZT0 at any Exception level, trapped to EL3.

ISS encoding for an exception from a Data Abort

2423222120191817161514131211109876543210
ISVSASBit[21]Bits[20:16]Bit[15]Bit[14]VNCRBits[12:11]FnVEACMS1PTWWnRDFSC

The ISS2 encoding for an exception from a Data Abort includes further information about the exception when any of the following features are implemented:

  • FEAT_LS64_V.

  • FEAT_LS64_ACCDATA.

  • FEAT_THE.

  • FEAT_S1POE or FEAT_S2POE.

  • FEAT_S1PIE or FEAT_S2PIE.

  • FEAT_GCS.

  • FEAT_MTE_CANONICAL_TAGS.

  • FEAT_MTE_PERM.

ISV, bit [24]

Instruction Syndrome Valid. Indicates whether the syndrome information in ISS[23:14] is valid.

ISVMeaning
0b0

No valid instruction syndrome. ISS[23:14] are RES0.

0b1

ISS[23:14] hold a valid instruction syndrome.

In ESR_EL2, ISV is 1 when FEAT_LS64 is implemented and a memory access generated by an LD64B or ST64B instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault.

In ESR_EL2, ISV is 1 when FEAT_LS64_V is implemented and a memory access generated by an ST64BV instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault.

In ESR_EL2, ISV is 1 when FEAT_LS64_ACCDATA is implemented and a memory access generated by an ST64BV0 instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault.

For other faults reported in ESR_EL2, ISV is 0 except for the following stage 2 aborts:

  • AArch64 loads and stores of a single general-purpose register (including the register specified with 0b11111, including those with Acquire/Release semantics, but excluding Load Exclusive or Store Exclusive and excluding those with writeback).
  • AArch32 instructions where the instruction:
    • Is an LDR, LDA, LDRT, LDRSH, LDRSHT, LDRH, LDAH, LDRHT, LDRSB, LDRSBT, LDRB, LDAB, LDRBT, STR, STL, STRT, STRH, STLH, STRHT, STRB, STLB, or STRBT instruction.
    • Is not performing register writeback.
    • Is not using R15 as a source or destination register.

For these stage 2 aborts, ISV is UNKNOWN if the exception was generated in Debug state in memory access mode, and otherwise indicates whether ISS[23:14] hold a valid syndrome.

For faults reported in ESR_EL1 or ESR_EL3, ISV is 1 when FEAT_LS64 is implemented and a memory access generated by an LD64B or ST64B instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault.

For faults reported in ESR_EL1 or ESR_EL3, ISV is 1 when FEAT_LS64_V is implemented and a memory access generated by an ST64BV instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault.

For faults reported in ESR_EL1 or ESR_EL3, ISV is 1 when FEAT_LS64_ACCDATA is implemented and a memory access generated by an ST64BV0 instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault.

When FEAT_RAS is implemented, ISV is 0 for any synchronous External abort.

When FEAT_RAS is not implemented, it is IMPLEMENTATION DEFINED whether ISV is set to 1 or 0 on a synchronous External abort on a stage 2 translation table walk.

For ISS reporting, a stage 2 abort on a stage 1 translation table walk does not return a valid instruction syndrome, and therefore ISV is 0 for these aborts.

When FEAT_MOPS is implemented, for a synchronous Data Abort on a Memory Copy and Memory Set instruction, ISV is 0.

When FEAT_MTE is implemented, for a synchronous Data Abort on an instruction that directly accesses Allocation Tags, ISV is 0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

SAS, bits [23:22]
When ISV == 1:

Syndrome Access Size. Indicates the size of the access attempted by the faulting operation.

SASMeaning
0b00

Byte

0b01

Halfword

0b10

Word

0b11

Doubleword

When FEAT_LS64 is implemented, if a memory access generated by an LD64B or ST64B instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 0b11.

When FEAT_LS64_V is implemented, if a memory access generated by an ST64BV instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 0b11.

When FEAT_LS64_ACCDATA is implemented, if a memory access generated by an ST64BV0 instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 0b11.

This field is UNKNOWN when the value of ISV is UNKNOWN.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bit[21]
When ISV == 1:

SSE, bit [21]

Syndrome Sign Extend. For a byte, halfword, or word load operation, indicates whether the data item must be sign extended.

SSEMeaning
0b0

Sign-extension not required.

0b1

Data item must be sign-extended.

When FEAT_LS64 is implemented, if a memory access generated by an LD64B or ST64B instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 0.

When FEAT_LS64_V is implemented, if a memory access generated by an ST64BV instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 0.

When FEAT_LS64_ACCDATA is implemented, if a memory access generated by an ST64BV0 instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 0.

For all other operations, this field is 0.

This field is UNKNOWN when the value of ISV is UNKNOWN.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


When ISV == 0 and FEAT_THE is implemented:

TopLevel, bit [21]

Indicates if the fault was due to TopLevel.

TopLevelMeaning
0b0

Fault is not due to TopLevel.

0b1

Fault is due to TopLevel.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bits[20:16]
When ISV == 1:

SRT, bits [4:0] of bits [20:16]

Syndrome Register Transfer. The register number of the Wt/Xt/Rt operand of the faulting instruction.

If the exception was taken from an Exception level that is using AArch32, then this is the AArch64 view of the register. See 'Mapping of the general-purpose registers between the Execution states'.

This field is UNKNOWN when the value of ISV is UNKNOWN.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


When ISV == 0, FEAT_RASv2 is implemented and (DFSC == 0b010000, or DFSC IN {0b01001x} or DFSC IN {0b0101xx}):

Bits [4:2] of bits [20:16]

Reserved, RES0.

WU, bits [1:0] of bits [20:16]

Write Update. Describes whether a store instruction that generated an External abort updated the location.

WUMeaning
0b00

Not a store instruction or translation table update, or the location might have been updated.

0b10

Store instruction or translation table update that did not update the location.

0b11

Store instruction or translation table update that updated the location.

In the description of this field, a store instruction is any memory-writing instruction that explicitly performs a store. This includes instructions that both read and write memory.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bit[15]
When ISV == 1:

SF, bit [15]

Sixty Four bit general-purpose register transfer. Width of the register accessed by the instruction is 64-bit.

SFMeaning
0b0

Instruction loads/stores a 32-bit general-purpose register.

0b1

Instruction loads/stores a 64-bit general-purpose register.

Note

This field specifies the register width identified by the instruction, not the Execution state.

When FEAT_LS64 is implemented, if a memory access generated by an LD64B or ST64B instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 1.

When FEAT_LS64_V is implemented, if a memory access generated by an ST64BV instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 1.

When FEAT_LS64_ACCDATA is implemented, if a memory access generated by an ST64BV0 instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 1.

This field is UNKNOWN when the value of ISV is UNKNOWN.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


When ISV == 0:

FnP, bit [15]

FAR not Precise.

FnPMeaningApplies when
0b0

The FAR holds the faulting virtual address that generated the Data Abort.

0b1

The FAR holds any virtual address within the naturally-aligned granule that contains the faulting virtual address that generated a Data Abort due to an SVE contiguous vector load/store instruction, or an SME load/store instruction.

For more information about the naturally-aligned fault granule, see FAR_ELx (for example, FAR_EL1).

When FEAT_SME is implemented or FEAT_SVE is implemented

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bit[14]
When ISV == 1:

AR, bit [14]

Acquire/Release.

ARMeaning
0b0

Instruction did not have acquire/release semantics.

0b1

Instruction did have acquire/release semantics.

When FEAT_LS64 is implemented, if a memory access generated by an LD64B or ST64B instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 0.

When FEAT_LS64_V is implemented, if a memory access generated by an ST64BV instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 0.

When FEAT_LS64_ACCDATA is implemented, if a memory access generated by an ST64BV0 instruction generates a Data Abort for a Translation fault, Access flag fault, or Permission fault, then this field is 0.

This field is UNKNOWN when the value of ISV is UNKNOWN.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


When FEAT_PFAR is implemented, ISV == 0 and (DFSC == 0b010000, or DFSC IN {0b01001x} or DFSC IN {0b0101xx}):

PFV, bit [14]

FAR Valid. Describes whether the PFAR_EL2 is valid.

PFVMeaning
0b0

PFAR_EL2 is UNKNOWN.

0b1

PFAR_EL2 is valid.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

VNCR, bit [13]

Indicates that the fault came from use of VNCR_EL2 register by EL1 code.

VNCRMeaningApplies when
0b0

The fault was not generated by the use of VNCR_EL2 by EL1 code.

0b1

The fault was generated by the use of VNCR_EL2 by EL1 code.

When FEAT_NV2 is implemented

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits[12:11]
When (DFSC IN {0b00xxxx} || DFSC IN {0b10101x}) && !(DFSC IN {0b0000xx}):

LST, bits [1:0] of bits [12:11]

Load/Store Type. Used when a Translation fault, Access flag fault, or Permission fault generates a Data Abort.

LSTMeaningApplies when
0b00

The instruction that generated the Data Abort is not specified.

0b01

An ST64BV instruction generated the Data Abort.

When FEAT_LS64_V is implemented
0b10

An LD64B or ST64B instruction generated the Data Abort.

When FEAT_LS64 is implemented
0b11

An ST64BV0 instruction generated the Data Abort.

When FEAT_LS64_ACCDATA is implemented

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


When FEAT_RAS is implemented and (DFSC == 0b010000, or DFSC IN {0b01001x} or DFSC IN {0b0101xx}):

SET, bits [1:0] of bits [12:11]

Synchronous Error Type. Used when a synchronous External abort, not on a Translation table walk or hardware update of the Translation table, generated the Data Abort. Describes the PE error state after taking the Data Abort exception.

SETMeaningApplies when
0b00

Recoverable state (UER).

0b10

Uncontainable (UC).

If FEAT_RASv2 is implemented, this value is reserved.

When FEAT_RASv2 is implemented
0b11

Restartable state (UEO).

Note

Software can use this information to determine what recovery might be possible. Taking a synchronous External abort exception might result in a PE state that is not recoverable.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

FnV, bit [10]

FAR not Valid, for a synchronous External abort other than a synchronous External abort on a translation table walk.

FnVMeaning
0b0

FAR is valid.

0b1

FAR is not valid, and holds an UNKNOWN value.

This field is valid only if the DFSC code is 0b010000. It is RES0 for all other aborts.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

EA, bit [9]

External abort type. This bit can provide an IMPLEMENTATION DEFINED classification of External aborts.

For any abort other than an External abort this bit returns a value of 0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

CM, bit [8]

Cache maintenance. Indicates whether the Data Abort came from a cache maintenance or address translation instruction:

CMMeaning
0b0

The Data Abort was not generated by the execution of one of the System instructions identified in the description of value 1.

0b1

The Data Abort was generated by either the execution of a cache maintenance instruction or by a synchronous fault on the execution of an address translation instruction. The DC ZVA, DC GVA, and DC GZVA instructions are not classified as cache maintenance instructions, and therefore their execution cannot cause this field to be set to 1.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

S1PTW, bit [7]

For a stage 2 fault, indicates whether the fault was a stage 2 fault on an access made for a stage 1 translation table walk:

S1PTWMeaning
0b0

Fault not on a stage 2 translation for a stage 1 translation table walk.

0b1

Fault on the stage 2 translation of an access for a stage 1 translation table walk.

For any abort other than a stage 2 fault this bit is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

WnR, bit [6]

Write not Read. Indicates whether a synchronous abort was caused by an instruction writing to a memory location, or by an instruction reading from a memory location.

WnRMeaning
0b0

Abort caused by an instruction reading from a memory location.

0b1

Abort caused by an instruction writing to a memory location.

For faults on cache maintenance and address translation instructions, this bit always returns a value of 1.

For faults from an atomic instruction that both reads and writes from a memory location, this bit is set to 0 if a read of the address specified by the instruction would have generated the fault which is being reported, otherwise it is set to 1. The architecture permits, but does not require, a relaxation of this requirement such that for all stage 2 aborts on stage 1 translation table walks for atomic instructions, the WnR bit is always 0.

This field is UNKNOWN for:

  • If FEAT_RASv2 is implemented, an External abort on an Atomic access, reported with ESR_EL2.WU set to 0b00.
  • A fault reported using a DFSC value of 0b110101 or 0b110001, indicating an unsupported Exclusive or atomic access.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

DFSC, bits [5:0]

Data Fault Status Code.

DFSCMeaningApplies when
0b000000

Address size fault, level 0 of translation or translation table base register.

0b000001

Address size fault, level 1.

0b000010

Address size fault, level 2.

0b000011

Address size fault, level 3.

0b000100

Translation fault, level 0.

0b000101

Translation fault, level 1.

0b000110

Translation fault, level 2.

0b000111

Translation fault, level 3.

0b001001

Access flag fault, level 1.

0b001010

Access flag fault, level 2.

0b001011

Access flag fault, level 3.

0b001000

Access flag fault, level 0.

When FEAT_LPA2 is implemented
0b001100

Permission fault, level 0.

When FEAT_LPA2 is implemented
0b001101

Permission fault, level 1.

0b001110

Permission fault, level 2.

0b001111

Permission fault, level 3.

0b010000

Synchronous External abort, not on translation table walk or hardware update of translation table.

0b010001

Synchronous Tag Check Fault.

When FEAT_MTE2 is implemented
0b010010

Synchronous External abort on translation table walk or hardware update of translation table, level -2.

When FEAT_D128 is implemented
0b010011

Synchronous External abort on translation table walk or hardware update of translation table, level -1.

When FEAT_LPA2 is implemented
0b010100

Synchronous External abort on translation table walk or hardware update of translation table, level 0.

0b010101

Synchronous External abort on translation table walk or hardware update of translation table, level 1.

0b010110

Synchronous External abort on translation table walk or hardware update of translation table, level 2.

0b010111

Synchronous External abort on translation table walk or hardware update of translation table, level 3.

0b011000

Synchronous parity or ECC error on memory access, not on translation table walk.

When FEAT_RAS is not implemented
0b011011

Synchronous parity or ECC error on memory access on translation table walk or hardware update of translation table, level -1.

When FEAT_LPA2 is implemented and FEAT_RAS is not implemented
0b011100

Synchronous parity or ECC error on memory access on translation table walk or hardware update of translation table, level 0.

When FEAT_RAS is not implemented
0b011101

Synchronous parity or ECC error on memory access on translation table walk or hardware update of translation table, level 1.

When FEAT_RAS is not implemented
0b011110

Synchronous parity or ECC error on memory access on translation table walk or hardware update of translation table, level 2.

When FEAT_RAS is not implemented
0b011111

Synchronous parity or ECC error on memory access on translation table walk or hardware update of translation table, level 3.

When FEAT_RAS is not implemented
0b100001

Alignment fault.

0b100010

Granule Protection Fault on translation table walk or hardware update of translation table, level -2.

When FEAT_D128 is implemented and FEAT_RME is implemented
0b100011

Granule Protection Fault on translation table walk or hardware update of translation table, level -1.

When FEAT_RME is implemented and FEAT_LPA2 is implemented
0b100100

Granule Protection Fault on translation table walk or hardware update of translation table, level 0.

When FEAT_RME is implemented
0b100101

Granule Protection Fault on translation table walk or hardware update of translation table, level 1.

When FEAT_RME is implemented
0b100110

Granule Protection Fault on translation table walk or hardware update of translation table, level 2.

When FEAT_RME is implemented
0b100111

Granule Protection Fault on translation table walk or hardware update of translation table, level 3.

When FEAT_RME is implemented
0b101000

Granule Protection Fault, not on translation table walk or hardware update of translation table.

When FEAT_RME is implemented
0b101001

Address size fault, level -1.

When FEAT_LPA2 is implemented
0b101010

Translation fault, level -2.

When FEAT_D128 is implemented
0b101011

Translation fault, level -1.

When FEAT_LPA2 is implemented
0b101100

Address Size fault, level -2.

When FEAT_D128 is implemented
0b110000

TLB conflict abort.

0b110001

Unsupported atomic hardware update fault.

When FEAT_HAFDBS is implemented
0b110100

IMPLEMENTATION DEFINED fault (Lockdown).

0b110101

IMPLEMENTATION DEFINED fault (Unsupported Exclusive or Atomic access).

All other values are reserved.

For more information about the lookup level associated with a fault, see 'The lookup level associated with MMU faults'.

If the S1PTW bit is set, then the level refers the level of the stage2 translation that is translating a stage 1 translation walk.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

ISS encoding for an exception from a trapped floating-point exception

2423222120191817161514131211109876543210
RES0TFVRES0VECITRIDFRES0IXFUFFOFFDZFIOF

Bit [24]

Reserved, RES0.

TFV, bit [23]

Trapped Fault Valid bit. Indicates whether the IDF, IXF, UFF, OFF, DZF, and IOF bits hold valid information about trapped floating-point exceptions.

TFVMeaning
0b0

The IDF, IXF, UFF, OFF, DZF, and IOF bits do not hold valid information about trapped floating-point exceptions and are UNKNOWN.

0b1

One or more floating-point exceptions occurred during an operation performed while executing the reported instruction. The IDF, IXF, UFF, OFF, DZF, and IOF bits indicate trapped floating-point exceptions that occurred. For more information, see 'Floating-point exceptions and exception traps'.

It is IMPLEMENTATION DEFINED whether this field is set to 0 on an exception generated by a trapped floating-point exception from an instruction that is performing floating-point operations on more than one lane of a vector.

Note

This is not a requirement. Implementations can set this field to 1 on a trapped floating-point exception from an instruction and return valid information in the {IDF, IXF, UFF, OFF, DZF, IOF} fields.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits [22:11]

Reserved, RES0.

VECITR, bits [10:8]

For a trapped floating-point exception from an instruction executed in AArch32 state this field is RES1.

For a trapped floating-point exception from an instruction executed in AArch64 state this field is UNKNOWN.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

IDF, bit [7]

Input Denormal floating-point exception trapped bit. If the TFV field is 0, this bit is UNKNOWN. Otherwise, the possible values of this bit are:

IDFMeaning
0b0

Input denormal floating-point exception has not occurred.

0b1

Input denormal floating-point exception occurred during execution of the reported instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits [6:5]

Reserved, RES0.

IXF, bit [4]

Inexact floating-point exception trapped bit. If the TFV field is 0, this bit is UNKNOWN. Otherwise, the possible values of this bit are:

IXFMeaning
0b0

Inexact floating-point exception has not occurred.

0b1

Inexact floating-point exception occurred during execution of the reported instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

UFF, bit [3]

Underflow floating-point exception trapped bit. If the TFV field is 0, this bit is UNKNOWN. Otherwise, the possible values of this bit are:

UFFMeaning
0b0

Underflow floating-point exception has not occurred.

0b1

Underflow floating-point exception occurred during execution of the reported instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

OFF, bit [2]

Overflow floating-point exception trapped bit. If the TFV field is 0, this bit is UNKNOWN. Otherwise, the possible values of this bit are:

OFFMeaning
0b0

Overflow floating-point exception has not occurred.

0b1

Overflow floating-point exception occurred during execution of the reported instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

DZF, bit [1]

Divide by Zero floating-point exception trapped bit. If the TFV field is 0, this bit is UNKNOWN. Otherwise, the possible values of this bit are:

DZFMeaning
0b0

Divide by Zero floating-point exception has not occurred.

0b1

Divide by Zero floating-point exception occurred during execution of the reported instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

IOF, bit [0]

Invalid Operation floating-point exception trapped bit. If the TFV field is 0, this bit is UNKNOWN. Otherwise, the possible values of this bit are:

IOFMeaning
0b0

Invalid Operation floating-point exception has not occurred.

0b1

Invalid Operation floating-point exception occurred during execution of the reported instruction.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from a trapped floating-point exception

In an implementation that supports the trapping of floating-point exceptions:

  • From an Exception level using AArch64, the FPCR.{IDE, IXE, UFE, OFE, DZE, IOE} bits enable each of the floating-point exception traps.
  • From an Exception level using AArch32, the FPSCR.{IDE, IXE, UFE, OFE, DZE, IOE} bits enable each of the floating-point exception traps.

ISS encoding for a GCS exception

2423222120191817161514131211109876543210
RES0ExTypeRES0RaddrBits[9:5]IT

Bit [24]

Reserved, RES0.

ExType, bits [23:20]

The first level classification of GCS exceptions.

ExTypeMeaning
0b0000

The exception reported is a Guarded Control Stack Data Check Exception.

0b0001

The exception reported is an EXLOCK Exception.

0b0010

The exception reported is a trap exception on GCSSTR or GCSSTTR instruction execution.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits [19:15]

Reserved, RES0.

Raddr, bits [14:10]
When ExType == 0b0010 :

Indicates the data address register number supplied in the instruction that has been trapped.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bits[9:5]
When ExType == 0b0000 :

Rn, bits [4:0] of bits [9:5]

Indicates a register number used by the instruction that caused the Guarded Control Stack Data Check Exception.

For a procedure return instruction reported with ESR_EL2.ISS.IT as 0b00000, contains the register number for the register which contains the target address of the branch.

For a GCSPOPM instruction reported with ESR_EL2.ISS.IT as 0b00001, contains the register number for the register which is the destination register of the instruction.

For a procedure return instruction reported with ESR_EL2.ISS.IT as 0b00010 or 0b00011, contains the value 0b11110, indicating X30.

For a GCSSS1 instruction reported with ESR_EL2.ISS.IT as 0b00100, contains the register number for the register which is the input register of the instruction.

If ESR_EL2.ISS.IT is reported as 0b00101 or 0b01000, this field is UNKNOWN

If ESR_EL2.ISS.IT is reported as 0b01001, this field is 0b11111

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


When ExType == 0b0010 :

Rvalue, bits [4:0] of bits [9:5]

Indicates the data value register number supplied in the instruction that has been trapped.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

IT, bits [4:0]
When ExType == 0b0000 :

Type of the instruction that caused the Guarded Control Stack Data Check Exception.

ITMeaning
0b00000

Guarded Control Stack Data Check Exception is from a procedure return instruction without Pointer authentication.

0b00001

Guarded Control Stack Data Check Exception is from a GCSPOPM instruction.

0b00010

Guarded Control Stack Data Check Exception is from a procedure return instruction with Pointer authentication that uses key A.

0b00011

Guarded Control Stack Data Check Exception is from a procedure return instruction with Pointer authentication that uses key B.

0b00100

Guarded Control Stack Data Check Exception is from a GCSSS1 instruction.

0b00101

Guarded Control Stack Data Check Exception is from a GCSSS2 instruction.

0b01000

Guarded Control Stack Data Check Exception is from a GCSPOPCX instruction.

0b01001

Guarded Control Stack Data Check Exception is from a GCSPOPX instruction.

All other values are reserved

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Additional information for the ISS encoding for a GCS exception

The following fields describe the configuration settings for the traps that are reported using EC value 0b101101 and ExType value 0b0010:

ISS encoding for an SError exception

2423222120191817161514131211109876543210
IDSRES0ELSWUVFVPFVIESBAETEARES0WnRVWnRDFSC
Note

In earlier versions of the architecture, an SError exception is referred to as an SError interrupt or an asynchronous External abort exception.

IDS, bit [24]

IMPLEMENTATION DEFINED syndrome.

IDSMeaning
0b0

Bits [23:0] of the ISS field holds the fields described in this encoding.

Note

If FEAT_RAS is not implemented, bits [23:0] of the ISS field are RES0.

0b1

Bits [23:0] of the ISS field holds IMPLEMENTATION DEFINED syndrome information that can be used to provide additional information about the SError exception.

Note

This field was previously called ISV.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits [23:19]

Reserved, RES0.

ELS, bit [18]
When FEAT_RASv2 is implemented and DFSC == 0b010001:

Meaning of ELR_ELx.

ELSMeaning
0b0

Asynchronous. Does not indicate the trigger for the exception.

0b1

Synchronous. The exception was triggered by the instruction at ELR_ELx.

SError exceptions that report this field is 1 are not required to be precise.

The ESR_EL2.AET field describes whether the exception is precise or imprecise.

Corrected, Recoverable or Restartable exceptions are precise. Unrecoverable or Uncontainable exceptions are imprecise.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

WU, bits [17:16]
When FEAT_RASv2 is implemented and DFSC == 0b010001:

Write Update. Describes whether a store instruction that generated an External abort updated the location.

WUMeaning
0b00

Not a store instruction or translation table update, or the location might have been updated.

0b10

Store instruction or translation table update that did not update the location.

0b11

Store instruction or translation table update that updated the location.

In the description of this field, a store instruction is any memory-writing instruction that explicitly performs a store. This includes instructions that both read and write memory.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

VFV, bit [15]
When FEAT_RASv2 is implemented and DFSC == 0b010001:

FAR Valid. Indicates the FAR_EL2 register contains a valid virtual address.

VFVMeaning
0b0

FAR_EL2 is not valid, and holds an UNKNOWN value.

0b1

FAR_EL2 contains a valid virtual address associated with the error.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

PFV, bit [14]
When FEAT_PFAR is implemented and DFSC == 0b010001:

FAR Valid. Describes whether the PFAR_EL2 is valid.

PFVMeaning
0b0

PFAR_EL2 is UNKNOWN.

0b1

PFAR_EL2 is valid.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

IESB, bit [13]
When FEAT_IESB is implemented and DFSC == 0b010001:

Implicit error synchronization event.

IESBMeaning
0b0

The SError exception was either not synchronized by the implicit error synchronization event or not taken immediately.

0b1

The SError exception was synchronized by the implicit error synchronization event and taken immediately.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

AET, bits [12:10]
When FEAT_RAS is implemented and DFSC == 0b010001:

Asynchronous Error Type.

Describes the PE error state after taking the SError exception.

AETMeaning
0b000

Uncontainable (UC).

0b001

Unrecoverable state (UEU).

0b010

Restartable state (UEO).

0b011

Recoverable state (UER).

0b110

Corrected (CE).

All other values are reserved.

If multiple errors are taken as a single SError exception, the overall PE error state is reported.

Note

Software can use this information to determine what recovery might be possible. The recovery software must also examine any implemented fault records to determine the location and extent of the error.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

EA, bit [9]
When FEAT_RAS is implemented and DFSC == 0b010001:

External abort type. Provides an IMPLEMENTATION DEFINED classification of External aborts.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

Bit [8]

Reserved, RES0.

WnRV, bit [7]
When FEAT_RASv2 is implemented and DFSC == 0b010001:

ESR_EL2.WnR valid.

WnRVMeaning
0b0

ESR_EL2.WnR is not valid and has been set to 0b0.

0b1

ESR_EL2.WnR is valid.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

WnR, bit [6]
When FEAT_RASv2 is implemented and DFSC == 0b010001:

Write-not-Read. When the WnRV field is 0b1, indicates whether an exception was caused by an instruction writing to a memory location, or by an instruction reading from a memory location.

WnRMeaning
0b0

Exception was caused by an instruction reading from a memory location.

0b1

Exception was caused by an instruction writing to a memory location.

Accessing this bit has the following behavior:

  • This bit is RES0 if ESR_EL2.WnRV==0b0.
  • This bit is not valid and reads UNKNOWN if an External abort on a Atomic access, reported with ESR_EL2.WU == 0b00.
  • Otherwise RW.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

DFSC, bits [5:0]
When FEAT_RAS is implemented:

Data Fault Status Code.

DFSCMeaning
0b000000

Uncategorized error.

0b010001

Asynchronous SError exception.

All other values are reserved.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.


Otherwise:

Reserved, RES0.

ISS encoding for an exception from a Breakpoint or Vector Catch debug exception

2423222120191817161514131211109876543210
RES0IFSC

Bits [24:6]

Reserved, RES0.

IFSC, bits [5:0]

Instruction Fault Status Code.

IFSCMeaning
0b100010

Debug exception.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from a Breakpoint or Vector Catch debug exception

For more information about generating these exceptions:

  • For exceptions from AArch64, see 'Breakpoint exceptions'.
  • For exceptions from AArch32, see 'Breakpoint exceptions' and 'Vector Catch exceptions'.

ISS encoding for an exception from a Software Step exception

2423222120191817161514131211109876543210
ISVRES0EXIFSC

ISV, bit [24]

Instruction syndrome valid. Indicates whether the EX bit, ISS[6], is valid, as follows:

ISVMeaning
0b0

EX bit is RES0.

0b1

EX bit is valid.

See the EX bit description for more information.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits [23:7]

Reserved, RES0.

EX, bit [6]

Exclusive operation. If the ISV bit is set to 1, this bit indicates whether a Load-Exclusive instruction was stepped.

EXMeaning
0b0

An instruction other than a Load-Exclusive instruction was stepped.

0b1

A Load-Exclusive instruction was stepped.

If the ISV bit is set to 0, this bit is RES0, indicating no syndrome data is available.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

IFSC, bits [5:0]

Instruction Fault Status Code.

IFSCMeaning
0b100010

Debug exception.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from a Software Step exception

For more information about generating these exceptions, see 'Software Step exceptions'.

ISS encoding for an exception from a Watchpoint exception

2423222120191817161514131211109876543210
RES0WPTWPTVWPFFnPRES0VNCRRES0FnVRES0CMRES0WnRDFSC

Bit [24]

Reserved, RES0.

WPT, bits [23:18]
When FEAT_Debugv8p2 is implemented:

Watchpoint number.

All other values are reserved.


Otherwise:

Reserved, RES0.

WPTV, bit [17]
When FEAT_Debugv8p2 is implemented:

Watchpoint number Valid.

WPTVMeaning
0b0

The WPT field is invalid, and holds an UNKNOWN value.

0b1

The WPT field is valid, and holds the number of a watchpoint that triggered a Watchpoint exception.

If FEAT_Debugv8p9 is implemented, value 0b0 is not permitted.

When a Watchpoint exception is triggered by a watchpoint match:

  • If FEAT_Debugv8p9 is implemented or the PE sets any of FnV, FnP, or WPF to 1, then the PE sets WPTV to 1.
  • Otherwise, the PE sets WPTV to an IMPLEMENTATION DEFINED value, 0 or 1.


Otherwise:

Reserved, RES0.

WPF, bit [16]

Watchpoint might be false-positive.

WPFMeaningApplies when
0b0

The watchpoint matched an address or address range that was accessed by the instruction.

0b1

The watchpoint matched an address or address range that might not have been accessed by the instruction.

When FEAT_SVE is implemented or FEAT_SME is implemented

Arm strongly recommends that this bit is set to 0, other than when one of the following instructions might generate a watchpoint match for an address or address range that the instruction does not access:

  • An SVE contiguous vector load/store instruction, when the PE is in Streaming SVE mode.
  • An SME load/store instruction.

FnP, bit [15]

FAR not Precise.

This field only has meaning if the FAR is valid; that is, when the FnV field is 0. If the FnV field is 1, the FnP field is 0.

FnPMeaningApplies when
0b0

If the FnV field is 0, the FAR holds the virtual address of an access or set of contiguous accesses that triggered a Watchpoint exception.

0b1

The FAR holds any address within the smallest implemented translation granule that contains the virtual address of an access or set of contiguous accesses that triggered a Watchpoint exception.

When FEAT_SVE is implemented or FEAT_SME is implemented

Bit [14]

Reserved, RES0.

VNCR, bit [13]

Indicates that the watchpoint came from use of VNCR_EL2 register by EL1 code.

VNCRMeaningApplies when
0b0

The watchpoint was not generated by the use of VNCR_EL2 by EL1 code.

0b1

The watchpoint was generated by the use of VNCR_EL2 by EL1 code.

When FEAT_NV2 is implemented

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bits [12:11]

Reserved, RES0.

FnV, bit [10]

FAR not Valid.

FnVMeaningApplies when
0b0

The FAR is valid, and its value is as described by the FnP field.

0b1

The FAR is invalid, and holds an UNKNOWN value.

When FEAT_SVE is implemented or FEAT_SME is implemented

Bit [9]

Reserved, RES0.

CM, bit [8]

Cache maintenance. Indicates whether the Watchpoint exception came from a cache maintenance instruction:

CMMeaning
0b0

The Watchpoint exception was not generated by the execution of one of the System instructions identified in the description of value 1.

0b1

The Watchpoint exception was generated by the execution of a cache maintenance instruction. The DC ZVA, DC GVA, and DC GZVA instructions are not classified as a cache maintenance instructions, and therefore their execution does not cause this field to be set to 1.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bit [7]

Reserved, RES0.

WnR, bit [6]

Write not Read. Indicates whether the Watchpoint exception was caused by an instruction writing to a memory location, or by an instruction reading from a memory location.

WnRMeaning
0b0

Watchpoint exception caused by an instruction reading from a memory location.

0b1

Watchpoint exception caused by an instruction writing to a memory location.

For Watchpoint exceptions on cache maintenance instructions, this bit always returns a value of 1.

For Watchpoint exceptions from an atomic instruction, this field is set to 0 if a read of the location would have generated the Watchpoint exception, otherwise it is set to 1.

If multiple watchpoints match on the same access, it is UNPREDICTABLE which watchpoint generates the Watchpoint exception.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

DFSC, bits [5:0]

Data Fault Status Code.

DFSCMeaning
0b100010

Debug exception.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from a Watchpoint exception

For more information about generating these exceptions, see 'Watchpoint exceptions'.

ISS encoding for an exception from execution of a Breakpoint instruction

2423222120191817161514131211109876543210
RES0Comment

Bits [24:16]

Reserved, RES0.

Comment, bits [15:0]

Set to the instruction comment field value, zero extended as necessary.

For the AArch32 BKPT instructions, the comment field is described as the immediate field.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from execution of a Breakpoint instruction

For more information about generating these exceptions, see 'Breakpoint instruction exceptions'.

ISS encoding for an exception from an ERET, ERETAA, or ERETAB instruction

2423222120191817161514131211109876543210
RES0ERETERETA

This EC value applies when FEAT_FGT is implemented, or when the Effective value of HCR_EL2.NV is 1.

Bits [24:2]

Reserved, RES0.

ERET, bit [1]

Indicates whether an ERET or ERETA* instruction was trapped to EL2.

ERETMeaning
0b0

ERET instruction trapped to EL2.

0b1

ERETAA or ERETAB instruction trapped to EL2.

If this bit is 0, the ERETA field is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

ERETA, bit [0]

Indicates whether an ERETAA or ERETAB instruction was trapped to EL2.

ERETAMeaning
0b0

ERETAA instruction trapped to EL2.

0b1

ERETAB instruction trapped to EL2.

When the ERET field is 0, this bit is RES0.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for an exception from an ERET, ERETAA, or ERETAB instruction

For more information about generating these exceptions, see HCR_EL2.NV.

If FEAT_FGT is implemented, HFGITR_EL2.ERET controls fine-grained trap exceptions from ERET, ERETAA and ERETAB execution.

ISS encoding for an exception from a TSTART instruction

2423222120191817161514131211109876543210
RES0RdRES0

Bits [24:10]

Reserved, RES0.

Rd, bits [9:5]

The Rd value from the issued instruction, the general purpose register used for the destination.

Bits [4:0]

Reserved, RES0.

ISS encoding for an exception from Branch Target Identification instruction

2423222120191817161514131211109876543210
RES0BTYPE

Bits [24:2]

Reserved, RES0.

BTYPE, bits [1:0]

This field is set to the PSTATE.BTYPE value that generated the Branch Target Exception.

Additional information for the ISS encoding for an exception from Branch Target Identification instruction

For more information about generating these exceptions, see 'The AArch64 application level programmers model'.

ISS encoding for an exception from a Pointer Authentication instruction when HCR_EL2.API == 0 || SCR_EL3.API == 0

2423222120191817161514131211109876543210
RES0

Bits [24:0]

Reserved, RES0.

Additional information for the ISS encoding for an exception from a Pointer Authentication instruction when HCR_EL2.API == 0 || SCR_EL3.API == 0

For more information about generating these exceptions, see:

  • HCR_EL2.API, for exceptions from Pointer authentication instructions, using AArch64 state, trapped to EL2.
  • SCR_EL3.API, for exceptions from Pointer authentication instructions, using AArch64 state, trapped to EL3.

ISS encoding for a PAC Fail exception

2423222120191817161514131211109876543210
RES0Exception as a result of an Instruction key or a Data keyException as a result of an A key or a B key

Bits [24:2]

Reserved, RES0.

Bit [1]

This field indicates whether the exception is as a result of an Instruction key or a Data key.

Meaning
0b0

Instruction Key.

0b1

Data Key.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Bit [0]

This field indicates whether the exception is as a result of an A key or a B key.

Meaning
0b0

A key.

0b1

B key.

The reset behavior of this field is:

  • On a Warm reset, this field resets to an architecturally UNKNOWN value.

Additional information for the ISS encoding for a PAC Fail exception

The following instructions generate a PAC Fail exception when the Pointer Authentication Code (PAC) is incorrect:

  • AUTDA, AUTDZA.
  • AUTDB, AUTDZB.
  • AUTIA, AUTIA1716, AUTIASP, AUTIAZ, AUTIZA.
  • AUTIB, AUTIB1716, AUTIBSP, AUTIBZ, AUTIZB.

If FEAT_FPACCOMBINE is implemented, the following instructions generate a PAC Fail exception when the Pointer Authentication Code (PAC) is incorrect:

  • RETAA, RETAB.
  • BLRAA, BLRAAZ, BLRAB, BLRABZ.
  • BRAA, BRAB, BRAAZ, BRABZ.
  • ERETAA, ERETAB.
  • LDRAA, LDRAB, whether the authenticated address is written back to the base register or not.

Accessing ESR_EL2

When the Effective value of HCR_EL2.E2H is 1, without explicit synchronization, accesses from EL2 using the accessor name ESR_EL2 or ESR_EL1 are not guaranteed to be ordered with respect to accesses using the other accessor name.

Accesses to this register use the following encodings in the System register encoding space:

MRS <Xt>, ESR_EL2

op0op1CRnCRmop2
0b110b1000b01010b00100b000

if PSTATE.EL == EL0 then UNDEFINED; elsif PSTATE.EL == EL1 then if EffectiveHCR_EL2_NVx() IN {'1x1'} then X[t, 64] = ESR_EL1; elsif EffectiveHCR_EL2_NVx() IN {'xx1'} then AArch64.SystemAccessTrap(EL2, 0x18); else UNDEFINED; elsif PSTATE.EL == EL2 then X[t, 64] = ESR_EL2; elsif PSTATE.EL == EL3 then X[t, 64] = ESR_EL2;

MSR ESR_EL2, <Xt>

op0op1CRnCRmop2
0b110b1000b01010b00100b000

if PSTATE.EL == EL0 then UNDEFINED; elsif PSTATE.EL == EL1 then if EffectiveHCR_EL2_NVx() IN {'1x1'} then ESR_EL1 = X[t, 64]; elsif EffectiveHCR_EL2_NVx() IN {'xx1'} then AArch64.SystemAccessTrap(EL2, 0x18); else UNDEFINED; elsif PSTATE.EL == EL2 then ESR_EL2 = X[t, 64]; elsif PSTATE.EL == EL3 then ESR_EL2 = X[t, 64];

MRS <Xt>, ESR_EL1

op0op1CRnCRmop2
0b110b0000b01010b00100b000

if PSTATE.EL == EL0 then UNDEFINED; elsif PSTATE.EL == EL1 then if EL2Enabled() && HCR_EL2.TRVM == '1' then AArch64.SystemAccessTrap(EL2, 0x18); elsif EL2Enabled() && IsFeatureImplemented(FEAT_FGT) && (!HaveEL(EL3) || SCR_EL3.FGTEn == '1') && HFGRTR_EL2.ESR_EL1 == '1' then AArch64.SystemAccessTrap(EL2, 0x18); elsif EffectiveHCR_EL2_NVx() IN {'111'} then X[t, 64] = NVMem[0x138]; else X[t, 64] = ESR_EL1; elsif PSTATE.EL == EL2 then if ELIsInHost(EL2) then X[t, 64] = ESR_EL2; else X[t, 64] = ESR_EL1; elsif PSTATE.EL == EL3 then X[t, 64] = ESR_EL1;

MSR ESR_EL1, <Xt>

op0op1CRnCRmop2
0b110b0000b01010b00100b000

if PSTATE.EL == EL0 then UNDEFINED; elsif PSTATE.EL == EL1 then if EL2Enabled() && HCR_EL2.TVM == '1' then AArch64.SystemAccessTrap(EL2, 0x18); elsif EL2Enabled() && IsFeatureImplemented(FEAT_FGT) && (!HaveEL(EL3) || SCR_EL3.FGTEn == '1') && HFGWTR_EL2.ESR_EL1 == '1' then AArch64.SystemAccessTrap(EL2, 0x18); elsif EffectiveHCR_EL2_NVx() IN {'111'} then NVMem[0x138] = X[t, 64]; else ESR_EL1 = X[t, 64]; elsif PSTATE.EL == EL2 then if ELIsInHost(EL2) then ESR_EL2 = X[t, 64]; else ESR_EL1 = X[t, 64]; elsif PSTATE.EL == EL3 then ESR_EL1 = X[t, 64];


26/03/2024 09:49; 67c0ae5282a7629ba0ea0ba7267b43cd4f7939f6

Copyright © 2010-2024 Arm Limited or its affiliates. All rights reserved. This document is Non-Confidential.