MPIDR, Multiprocessor Affinity Register

The MPIDR characteristics are:

Purpose

In a multiprocessor system, provides an additional PE identification mechanism.

Configuration

AArch32 System register MPIDR bits [31:0] are architecturally mapped to AArch64 System register MPIDR_EL1[31:0].

This register is present only when EL1 is capable of using AArch32. Otherwise, direct accesses to MPIDR are UNDEFINED.

In a uniprocessor system, Arm recommends that each Aff<n> field of this register returns a value of 0.

Attributes

MPIDR is a 32-bit register.

Field descriptions

313029282726252423222120191817161514131211109876543210
MURES0MTAff2Aff1Aff0

M, bit [31]

Indicates whether this implementation includes the functionality introduced by the Armv7 Multiprocessing Extensions.

The value of this field is an IMPLEMENTATION DEFINED choice of:

MMeaning
0b0

This implementation does not include the Armv7 Multiprocessing Extensions functionality.

0b1

This implementation includes the Armv7 Multiprocessing Extensions functionality.

Access to this field is RAO/WI.

U, bit [30]

Indicates a Uniprocessor system, as distinct from PE 0 in a multiprocessor system.

The value of this field is an IMPLEMENTATION DEFINED choice of:

UMeaning
0b0

Processor is part of a multiprocessor system.

0b1

Processor is part of a uniprocessor system.

Access to this field is RO.

Bits [29:25]

Reserved, RES0.

MT, bit [24]

Indicates whether the lowest level of affinity consists of logical PEs that are implemented using an interdependent approach, such as multithreading. See the description of Aff0 for more information about affinity levels.

The value of this field is an IMPLEMENTATION DEFINED choice of:

MTMeaning
0b0

Performance of PEs with different affinity level 0 values, and the same values for affinity level 1 and higher, is largely independent.

0b1

Performance of PEs with different affinity level 0 values, and the same values for affinity level 1 and higher, is very interdependent.

Note

This field does not indicate that multithreading is implemented and does not indicate that PEs with different affinity level 0 values, and the same values for affinity level 1 and higher are implemented.

Access to this field is RO.

Aff2, bits [23:16]

Affinity level 2. See the description of Aff0 for more information.

This field has an IMPLEMENTATION DEFINED value.

Access to this field is RO.

Aff1, bits [15:8]

Affinity level 1. See the description of Aff0 for more information.

This field has an IMPLEMENTATION DEFINED value.

Access to this field is RO.

Aff0, bits [7:0]

Affinity level 0. The value of the MPIDR.{Aff2, Aff1, Aff0} or MPIDR_EL1.{Aff3, Aff2, Aff1, Aff0} set of fields of each PE must be unique within the system as a whole.

This field has an IMPLEMENTATION DEFINED value.

Access to this field is RO.

Accessing MPIDR

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

MRC{<c>}{<q>} <coproc>, {#}<opc1>, <Rt>, <CRn>, <CRm>{, {#}<opc2>}

coprocopc1CRnCRmopc2
0b11110b0000b00000b00000b101

if PSTATE.EL == EL0 then UNDEFINED; elsif PSTATE.EL == EL1 then if EL2Enabled() && !ELUsingAArch32(EL2) && HSTR_EL2.T0 == '1' then AArch64.AArch32SystemAccessTrap(EL2, 0x03); elsif EL2Enabled() && ELUsingAArch32(EL2) && HSTR.T0 == '1' then AArch32.TakeHypTrapException(0x03); elsif EL2Enabled() && !ELUsingAArch32(EL2) then R[t] = VMPIDR_EL2<31:0>; elsif EL2Enabled() && ELUsingAArch32(EL2) then R[t] = VMPIDR; else R[t] = MPIDR; elsif PSTATE.EL == EL2 then R[t] = MPIDR; elsif PSTATE.EL == EL3 then R[t] = MPIDR;


26/03/2024 09:49; 67c0ae5282a7629ba0ea0ba7267b43cd4f7939f6

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