1741292faSPeter MaydellA-profile CPU architecture support 2741292faSPeter Maydell================================== 3741292faSPeter Maydell 4741292faSPeter MaydellQEMU's TCG emulation includes support for the Armv5, Armv6, Armv7 and 5741292faSPeter MaydellArmv8 versions of the A-profile architecture. It also has support for 6741292faSPeter Maydellthe following architecture extensions: 7741292faSPeter Maydell 8741292faSPeter Maydell- FEAT_AA32BF16 (AArch32 BFloat16 instructions) 9741292faSPeter Maydell- FEAT_AA32HPD (AArch32 hierarchical permission disables) 10741292faSPeter Maydell- FEAT_AA32I8MM (AArch32 Int8 matrix multiplication instructions) 11741292faSPeter Maydell- FEAT_AES (AESD and AESE instructions) 1275d08a40SPeter Maydell- FEAT_BBM at level 2 (Translation table break-before-make levels) 13741292faSPeter Maydell- FEAT_BF16 (AArch64 BFloat16 instructions) 14741292faSPeter Maydell- FEAT_BTI (Branch Target Identification) 15*74b17e16SRichard Henderson- FEAT_CSV2 (Cache speculation variant 2) 16741292faSPeter Maydell- FEAT_DIT (Data Independent Timing instructions) 17741292faSPeter Maydell- FEAT_DPB (DC CVAP instruction) 18033a4f15SRichard Henderson- FEAT_Debugv8p2 (Debug changes for v8.2) 198fc756b6SRichard Henderson- FEAT_Debugv8p4 (Debug changes for v8.4) 20741292faSPeter Maydell- FEAT_DotProd (Advanced SIMD dot product instructions) 21741292faSPeter Maydell- FEAT_FCMA (Floating-point complex number instructions) 22741292faSPeter Maydell- FEAT_FHM (Floating-point half-precision multiplication instructions) 23741292faSPeter Maydell- FEAT_FP16 (Half-precision floating-point data processing) 24741292faSPeter Maydell- FEAT_FRINTTS (Floating-point to integer instructions) 25741292faSPeter Maydell- FEAT_FlagM (Flag manipulation instructions v2) 26741292faSPeter Maydell- FEAT_FlagM2 (Enhancements to flag manipulation instructions) 27741292faSPeter Maydell- FEAT_HPDS (Hierarchical permission disables) 28741292faSPeter Maydell- FEAT_I8MM (AArch64 Int8 matrix multiplication instructions) 29880cd10eSRichard Henderson- FEAT_IESB (Implicit error synchronization event) 30741292faSPeter Maydell- FEAT_JSCVT (JavaScript conversion instructions) 31741292faSPeter Maydell- FEAT_LOR (Limited ordering regions) 327a928f43SRichard Henderson- FEAT_LPA (Large Physical Address space) 33ef56c242SRichard Henderson- FEAT_LPA2 (Large Physical and virtual Address space v2) 34741292faSPeter Maydell- FEAT_LRCPC (Load-acquire RCpc instructions) 35741292faSPeter Maydell- FEAT_LRCPC2 (Load-acquire RCpc instructions v2) 36741292faSPeter Maydell- FEAT_LSE (Large System Extensions) 370af312b6SRichard Henderson- FEAT_LVA (Large Virtual Address space) 38741292faSPeter Maydell- FEAT_MTE (Memory Tagging Extension) 39741292faSPeter Maydell- FEAT_MTE2 (Memory Tagging Extension) 4086f0d4c7SPeter Collingbourne- FEAT_MTE3 (MTE Asymmetric Fault Handling) 41741292faSPeter Maydell- FEAT_PAN (Privileged access never) 42741292faSPeter Maydell- FEAT_PAN2 (AT S1E1R and AT S1E1W instruction variants affected by PSTATE.PAN) 43741292faSPeter Maydell- FEAT_PAuth (Pointer authentication) 44741292faSPeter Maydell- FEAT_PMULL (PMULL, PMULL2 instructions) 45741292faSPeter Maydell- FEAT_PMUv3p1 (PMU Extensions v3.1) 46741292faSPeter Maydell- FEAT_PMUv3p4 (PMU Extensions v3.4) 47e95c74c5SRichard Henderson- FEAT_RAS (Reliability, availability, and serviceability) 48741292faSPeter Maydell- FEAT_RDM (Advanced SIMD rounding double multiply accumulate instructions) 49741292faSPeter Maydell- FEAT_RNG (Random number generator) 50741292faSPeter Maydell- FEAT_SB (Speculation Barrier) 51741292faSPeter Maydell- FEAT_SEL2 (Secure EL2) 52741292faSPeter Maydell- FEAT_SHA1 (SHA1 instructions) 53741292faSPeter Maydell- FEAT_SHA256 (SHA256 instructions) 54741292faSPeter Maydell- FEAT_SHA3 (Advanced SIMD SHA3 instructions) 55741292faSPeter Maydell- FEAT_SHA512 (Advanced SIMD SHA512 instructions) 56741292faSPeter Maydell- FEAT_SM3 (Advanced SIMD SM3 instructions) 57741292faSPeter Maydell- FEAT_SM4 (Advanced SIMD SM4 instructions) 58741292faSPeter Maydell- FEAT_SPECRES (Speculation restriction instructions) 59741292faSPeter Maydell- FEAT_SSBS (Speculative Store Bypass Safe) 60741292faSPeter Maydell- FEAT_TLBIOS (TLB invalidate instructions in Outer Shareable domain) 61741292faSPeter Maydell- FEAT_TLBIRANGE (TLB invalidate range instructions) 62741292faSPeter Maydell- FEAT_TTCNP (Translation table Common not private translations) 63f81c60c2SPeter Maydell- FEAT_TTL (Translation Table Level) 64741292faSPeter Maydell- FEAT_TTST (Small translation tables) 65741292faSPeter Maydell- FEAT_UAO (Unprivileged Access Override control) 66741292faSPeter Maydell- FEAT_VHE (Virtualization Host Extensions) 67741292faSPeter Maydell- FEAT_VMID16 (16-bit VMID) 68741292faSPeter Maydell- FEAT_XNX (Translation table stage 2 Unprivileged Execute-never) 69741292faSPeter Maydell- SVE (The Scalable Vector Extension) 70741292faSPeter Maydell- SVE2 (The Scalable Vector Extension v2) 71741292faSPeter Maydell 72741292faSPeter MaydellFor information on the specifics of these extensions, please refer 73741292faSPeter Maydellto the `Armv8-A Arm Architecture Reference Manual 74741292faSPeter Maydell<https://developer.arm.com/documentation/ddi0487/latest>`_. 75741292faSPeter Maydell 76741292faSPeter MaydellWhen a specific named CPU is being emulated, only those features which 77741292faSPeter Maydellare present in hardware for that CPU are emulated. (If a feature is 78741292faSPeter Maydellnot in the list above then it is not supported, even if the real 79741292faSPeter Maydellhardware should have it.) The ``max`` CPU enables all features. 80741292faSPeter Maydell 81741292faSPeter MaydellR-profile CPU architecture support 82741292faSPeter Maydell================================== 83741292faSPeter Maydell 84741292faSPeter MaydellQEMU's TCG emulation support for R-profile CPUs is currently limited. 85741292faSPeter MaydellWe emulate only the Cortex-R5 and Cortex-R5F CPUs. 86741292faSPeter Maydell 87741292faSPeter MaydellM-profile CPU architecture support 88741292faSPeter Maydell================================== 89741292faSPeter Maydell 90741292faSPeter MaydellQEMU's TCG emulation includes support for Armv6-M, Armv7-M, Armv8-M, and 91741292faSPeter MaydellArmv8.1-M versions of the M-profile architucture. It also has support 92741292faSPeter Maydellfor the following architecture extensions: 93741292faSPeter Maydell 94741292faSPeter Maydell- FP (Floating-point Extension) 95741292faSPeter Maydell- FPCXT (FPCXT access instructions) 96741292faSPeter Maydell- HP (Half-precision floating-point instructions) 97741292faSPeter Maydell- LOB (Low Overhead loops and Branch future) 98741292faSPeter Maydell- M (Main Extension) 99741292faSPeter Maydell- MPU (Memory Protection Unit Extension) 100741292faSPeter Maydell- PXN (Privileged Execute Never) 101741292faSPeter Maydell- RAS (Reliability, Serviceability and Availability): "minimum RAS Extension" only 102741292faSPeter Maydell- S (Security Extension) 103741292faSPeter Maydell- ST (System Timer Extension) 104741292faSPeter Maydell 105741292faSPeter MaydellFor information on the specifics of these extensions, please refer 106741292faSPeter Maydellto the `Armv8-M Arm Architecture Reference Manual 107741292faSPeter Maydell<https://developer.arm.com/documentation/ddi0553/latest>`_. 108741292faSPeter Maydell 109741292faSPeter MaydellWhen a specific named CPU is being emulated, only those features which 110741292faSPeter Maydellare present in hardware for that CPU are emulated. (If a feature is 111741292faSPeter Maydellnot in the list above then it is not supported, even if the real 112741292faSPeter Maydellhardware should have it.) There is no equivalent of the ``max`` CPU for 113741292faSPeter MaydellM-profile. 114