FPSCR, Floating-Point Status and Control Register

The FPSCR characteristics are:

Purpose

Provides floating-point system status information and control.

Configuration

AArch32 System register FPSCR bits [31:27] are architecturally mapped to AArch64 System register FPSR[31:27].

AArch32 System register FPSCR bit [7] is architecturally mapped to AArch64 System register FPSR[7].

AArch32 System register FPSCR bits [4:0] are architecturally mapped to AArch64 System register FPSR[4:0].

AArch32 System register FPSCR bits [26:15] are architecturally mapped to AArch64 System register FPCR[26:15].

AArch32 System register FPSCR bits [12:8] are architecturally mapped to AArch64 System register FPCR[12:8].

This register is present only when AArch32 is supported. Otherwise, direct accesses to FPSCR are UNDEFINED.

It is IMPLEMENTATION DEFINED whether the Len and Stride fields can be programmed to nonzero values, which will cause some AArch32 floating-point instruction encodings to be UNDEFINED, or whether these fields are RAZ.

Implemented only if the implementation includes the Advanced SIMD and floating-point functionality.

Attributes

FPSCR is a 32-bit register.

Field descriptions

313029282726252423222120191817161514131211109876543210
NZCVQCAHPDNFZRModeStrideFZ16LenIDERES0IXEUFEOFEDZEIOEIDCRES0IXCUFCOFCDZCIOC

N, bit [31]

Negative condition flag. This is updated by floating-point comparison operations.

The reset behavior of this field is:

Z, bit [30]

Zero condition flag. This is updated by floating-point comparison operations.

The reset behavior of this field is:

C, bit [29]

Carry condition flag. This is updated by floating-point comparison operations.

The reset behavior of this field is:

V, bit [28]

Overflow condition flag. This is updated by floating-point comparison operations.

The reset behavior of this field is:

QC, bit [27]

Cumulative saturation bit, Advanced SIMD only. This bit is set to 1 to indicate that an Advanced SIMD integer operation has saturated since 0 was last written to this bit.

The reset behavior of this field is:

AHP, bit [26]

Alternative half-precision control bit:

AHPMeaning
0b0

IEEE half-precision format selected.

0b1

Alternative half-precision format selected.

This bit is used only for conversions between half-precision floating-point and other floating-point formats.

The data-processing instructions added as part of the FEAT_FP16 extension always use the IEEE half-precision format, and ignore the value of this bit.

The reset behavior of this field is:

DN, bit [25]

Default NaN mode control bit:

DNMeaning
0b0

NaN operands propagate through to the output of a floating-point operation.

0b1

Any operation involving one or more NaNs returns the Default NaN.

The value of this bit controls only scalar floating-point arithmetic. Advanced SIMD arithmetic always uses the Default NaN setting, regardless of the value of the DN bit.

The reset behavior of this field is:

FZ, bit [24]

Flush-to-zero mode control bit:

FZMeaning
0b0

Flush-to-zero mode disabled. Behavior of the floating-point system is fully compliant with the IEEE 754 standard.

0b1

Flush-to-zero mode enabled.

The value of this bit controls only scalar floating-point arithmetic. Advanced SIMD arithmetic always uses the Flush-to-zero setting, regardless of the value of the FZ bit.

This bit has no effect on half-precision calculations.

The reset behavior of this field is:

RMode, bits [23:22]

Rounding Mode control field. The encoding of this field is:

RModeMeaning
0b00

Round to Nearest (RN) mode.

0b01

Round towards Plus Infinity (RP) mode.

0b10

Round towards Minus Infinity (RM) mode.

0b11

Round towards Zero (RZ) mode.

The specified rounding mode is used by almost all scalar floating-point instructions. Advanced SIMD arithmetic always uses the Round to Nearest setting, regardless of the value of the RMode bits.

The reset behavior of this field is:

Stride, bits [21:20]

If this field is RW and is set to a value other than zero, some floating-point instruction encodings are UNDEFINED. The instruction pseudocode identifies these instructions.

Arm strongly recommends that software never sets this field to a value other than zero.

The value of this field is ignored when processing Advanced SIMD instructions.

The reset behavior of this field is:

When an implementation implements FPSCR.LEN,STRIDE as RAZ, access to this field is RAZ/WI.

FZ16, bit [19]
When FEAT_FP16 is implemented:

Flush-to-zero mode control bit on half-precision data-processing instructions:

FZ16Meaning
0b0

Flush-to-zero mode disabled. Behavior of the floating-point system is fully compliant with the IEEE 754 standard.

0b1

Flush-to-zero mode enabled.

The value of this bit applies to both scalar and Advanced SIMD floating-point half-precision calculations.

The reset behavior of this field is:


Otherwise:

Reserved, RES0.

Len, bits [18:16]

If this field is RW and is set to a value other than zero, some floating-point instruction encodings are UNDEFINED. The instruction pseudocode identifies these instructions.

Arm strongly recommends that software never sets this field to a value other than zero.

The value of this field is ignored when processing Advanced SIMD instructions.

The reset behavior of this field is:

When an implementation implements FPSCR.LEN,STRIDE as RAZ, access to this field is RAZ/WI.

IDE, bit [15]

Input Denormal floating-point exception trap enable.

IDEMeaning
0b0

Untrapped exception handling selected. If the floating-point exception occurs, the IDC bit is set to 1.

0b1

Trapped exception handling selected. If the floating-point exception occurs, the PE does not update the IDC bit.

When this bit is RW, it applies only to floating-point operations. Advanced SIMD operations always use untrapped floating-point exception handling in AArch32 state.

The reset behavior of this field is:

When an implementation does not implement trapping of Input Denormal floating-point exceptions, access to this field is RAZ/WI.

Bits [14:13]

Reserved, RES0.

IXE, bit [12]

Inexact floating-point exception trap enable.

IXEMeaning
0b0

Untrapped exception handling selected. If the floating-point exception occurs, the IXC bit is set to 1.

0b1

Trapped exception handling selected. If the floating-point exception occurs, the PE does not update the IXC bit.

When this bit is RW, it applies only to floating-point operations. Advanced SIMD operations always use untrapped floating-point exception handling in AArch32 state.

The reset behavior of this field is:

When an implementation does not implement trapping of Inexact floating-point exceptions, access to this field is RAZ/WI.

UFE, bit [11]

Underflow floating-point exception trap enable.

UFEMeaning
0b0

Untrapped exception handling selected. If the floating-point exception occurs, the UFC bit is set to 1.

0b1

Trapped exception handling selected. If the floating-point exception occurs and Flush-to-zero is not enabled, the PE does not update the UFC bit.

When this bit is RW, it applies only to floating-point operations. Advanced SIMD operations always use untrapped floating-point exception handling in AArch32 state.

The reset behavior of this field is:

When an implementation does not implement trapping of Underflow floating-point exceptions, access to this field is RAZ/WI.

OFE, bit [10]

Overflow floating-point exception trap enable.

OFEMeaning
0b0

Untrapped exception handling selected. If the floating-point exception occurs, the OFC bit is set to 1.

0b1

Trapped exception handling selected. If the floating-point exception occurs, the PE does not update the OFC bit.

When this bit is RW, it applies only to floating-point operations. Advanced SIMD operations always use untrapped floating-point exception handling in AArch32 state.

The reset behavior of this field is:

When an implementation does not implement trapping of Overflow floating-point exceptions, access to this field is RAZ/WI.

DZE, bit [9]

Divide by Zero floating-point exception trap enable.

DZEMeaning
0b0

Untrapped exception handling selected. If the floating-point exception occurs, the DZC bit is set to 1.

0b1

Trapped exception handling selected. If the floating-point exception occurs, the PE does not update the DZC bit.

When this bit is RW, it applies only to floating-point operations. Advanced SIMD operations always use untrapped floating-point exception handling in AArch32 state.

The reset behavior of this field is:

When an implementation does not implement trapping of Divide by Zero floating-point exceptions, access to this field is RAZ/WI.

IOE, bit [8]

Invalid Operation floating-point exception trap enable.

IOEMeaning
0b0

Untrapped exception handling selected. If the floating-point exception occurs, the IOC bit is set to 1.

0b1

Trapped exception handling selected. If the floating-point exception occurs, the PE does not update the IOC bit.

When this bit is RW, it applies only to floating-point operations. Advanced SIMD operations always use untrapped floating-point exception handling in AArch32 state.

The reset behavior of this field is:

When an implementation does not implement trapping of Invalid Operation floating-point exceptions, access to this field is RAZ/WI.

IDC, bit [7]

Input Denormal cumulative floating-point exception bit. This bit is set to 1 to indicate that the Input Denormal floating-point exception has occurred since 0 was last written to this bit.

How VFP instructions update this bit depends on the value of the IDE bit.

Advanced SIMD instructions set this bit if the Input Denormal floating-point exception occurs in one or more of the floating-point calculations performed by the instruction, regardless of the value of the IDE bit.

The reset behavior of this field is:

Bits [6:5]

Reserved, RES0.

IXC, bit [4]

Inexact cumulative floating-point exception bit. This bit is set to 1 to indicate that the Inexact floating-point exception has occurred since 0 was last written to this bit.

How VFP instructions update this bit depends on the value of the IXE bit.

Advanced SIMD instructions set this bit if the Inexact floating-point exception occurs in one or more of the floating-point calculations performed by the instruction, regardless of the value of the IXE bit.

The criteria for the Inexact floating-point exception to occur are different in Flush-to-zero mode. For more information, see 'Flush-to-zero'.

The reset behavior of this field is:

UFC, bit [3]

Underflow cumulative floating-point exception bit. This bit is set to 1 to indicate that the Underflow floating-point exception has occurred since 0 was last written to this bit.

How VFP instructions update this bit depends on the value of the UFE bit.

Advanced SIMD instructions set this bit if the Underflow floating-point exception occurs in one or more of the floating-point calculations performed by the instruction, if FPSCR.UFE is 0 or if Flush-to-zero is enabled.

The criteria for the Underflow floating-point exception to occur are different in Flush-to-zero mode. For more information, see 'Flush-to-zero'.

The reset behavior of this field is:

OFC, bit [2]

Overflow cumulative floating-point exception bit. This bit is set to 1 to indicate that the Overflow floating-point exception has occurred since 0 was last written to this bit.

How VFP instructions update this bit depends on the value of the OFE bit.

Advanced SIMD instructions set this bit if the Overflow floating-point exception occurs in one or more of the floating-point calculations performed by the instruction, regardless of the value of the OFE bit.

The reset behavior of this field is:

DZC, bit [1]

Divide by Zero cumulative floating-point exception bit. This bit is set to 1 to indicate that the Divide by Zero floating-point exception has occurred since 0 was last written to this bit.

How VFP instructions update this bit depends on the value of the DZE bit.

Advanced SIMD instructions set this bit if the Divide by Zero floating-point exception occurs in one or more of the floating-point calculations performed by the instruction, regardless of the value of the DZE bit.

The reset behavior of this field is:

IOC, bit [0]

Invalid Operation cumulative floating-point exception bit. This bit is set to 1 to indicate that the Invalid Operation floating-point exception has occurred since 0 was last written to this bit.

How VFP instructions update this bit depends on the value of the IOE bit.

Advanced SIMD instructions set this bit if the Invalid Operation floating-point exception occurs in one or more of the floating-point calculations performed by the instruction, regardless of the value of the IOE bit.

The reset behavior of this field is:

Accessing FPSCR

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

VMRS{<c>}{<q>} <Rt>, <spec_reg>

reg
0b0001

if PSTATE.EL == EL0 then if HaveEL(EL3) && EL3SDDUndefPriority() && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then UNDEFINED; elsif !ELUsingAArch32(EL1) && !ELIsInHost(EL0) && CPACR_EL1.FPEN != '11' then if EL2Enabled() && !ELUsingAArch32(EL2) && HCR_EL2.TGE == '1' then AArch64.AArch32SystemAccessTrap(EL2, 0x00); else AArch64.AArch32SystemAccessTrap(EL1, 0x07); elsif ELUsingAArch32(EL1) && ((ELUsingAArch32(EL3) && SCR.NS == '1' && NSACR.cp10 == '0') || CPACR.cp10 IN {'0x'}) then UNDEFINED; elsif ELIsInHost(EL0) && CPTR_EL2.FPEN != '11' then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif ELIsInHost(EL2) && CPTR_EL2.FPEN IN {'x0'} then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif EL2Enabled() && !ELUsingAArch32(EL2) && !ELIsInHost(EL2) && CPTR_EL2.TFP == '1' then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif EL2Enabled() && ELUsingAArch32(EL1) && ((ELUsingAArch32(EL3) && SCR.NS == '1' && NSACR.cp10 == '0') || HCPTR.TCP10 == '1') then AArch32.TakeHypTrapException(0x08); elsif HaveEL(EL3) && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then if EL3SDDUndef() then UNDEFINED; else AArch64.AArch32SystemAccessTrap(EL3, 0x07); else R[t] = FPSCR; elsif PSTATE.EL == EL1 then if HaveEL(EL3) && EL3SDDUndefPriority() && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then UNDEFINED; elsif CPACR_EL1.FPEN IN {'x0'} then AArch64.AArch32SystemAccessTrap(EL1, 0x07); elsif (ELUsingAArch32(EL3) && SCR.NS == '1' && NSACR.cp10 == '0') || CPACR.cp10 == '00' then UNDEFINED; elsif EL2Enabled() && !ELUsingAArch32(EL2) && !ELIsInHost(EL2) && CPTR_EL2.TFP == '1' then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif ELIsInHost(EL2) && CPTR_EL2.FPEN IN {'x0'} then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif EL2Enabled() && ELUsingAArch32(EL2) && ((ELUsingAArch32(EL3) && SCR.NS == '1' && NSACR.cp10 == '0') || HCPTR.TCP10 == '1') then AArch32.TakeHypTrapException(0x08); elsif HaveEL(EL3) && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then if EL3SDDUndef() then UNDEFINED; else AArch64.AArch32SystemAccessTrap(EL3, 0x07); else R[t] = FPSCR; elsif PSTATE.EL == EL2 then if HaveEL(EL3) && EL3SDDUndefPriority() && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then UNDEFINED; elsif ELIsInHost(EL2) && CPTR_EL2.FPEN IN {'x0'} then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif EL2Enabled() && ((ELUsingAArch32(EL3) && SCR.NS == '1' && NSACR.cp10 == '0') || HCPTR.TCP10 == '1') then AArch32.TakeHypTrapException(0x00); elsif HaveEL(EL3) && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then if EL3SDDUndef() then UNDEFINED; else AArch64.AArch32SystemAccessTrap(EL3, 0x07); else R[t] = FPSCR; elsif PSTATE.EL == EL3 then if CPACR.cp10 == '00' then UNDEFINED; else R[t] = FPSCR;

VMSR{<c>}{<q>} <spec_reg>, <Rt>

reg
0b0001

if PSTATE.EL == EL0 then if HaveEL(EL3) && EL3SDDUndefPriority() && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then UNDEFINED; elsif !ELUsingAArch32(EL1) && !ELIsInHost(EL0) && CPACR_EL1.FPEN != '11' then if EL2Enabled() && !ELUsingAArch32(EL2) && HCR_EL2.TGE == '1' then AArch64.AArch32SystemAccessTrap(EL2, 0x00); else AArch64.AArch32SystemAccessTrap(EL1, 0x07); elsif ELUsingAArch32(EL1) && ((ELUsingAArch32(EL3) && SCR.NS == '1' && NSACR.cp10 == '0') || CPACR.cp10 IN {'0x'}) then UNDEFINED; elsif ELIsInHost(EL0) && CPTR_EL2.FPEN != '11' then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif ELIsInHost(EL2) && CPTR_EL2.FPEN IN {'x0'} then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif EL2Enabled() && !ELUsingAArch32(EL2) && !ELIsInHost(EL2) && CPTR_EL2.TFP == '1' then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif EL2Enabled() && ELUsingAArch32(EL1) && ((ELUsingAArch32(EL3) && SCR.NS == '1' && NSACR.cp10 == '0') || HCPTR.TCP10 == '1') then AArch32.TakeHypTrapException(0x08); elsif HaveEL(EL3) && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then if EL3SDDUndef() then UNDEFINED; else AArch64.AArch32SystemAccessTrap(EL3, 0x07); else FPSCR = R[t]; elsif PSTATE.EL == EL1 then if HaveEL(EL3) && EL3SDDUndefPriority() && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then UNDEFINED; elsif CPACR_EL1.FPEN IN {'x0'} then AArch64.AArch32SystemAccessTrap(EL1, 0x07); elsif (ELUsingAArch32(EL3) && SCR.NS == '1' && NSACR.cp10 == '0') || CPACR.cp10 == '00' then UNDEFINED; elsif EL2Enabled() && !ELUsingAArch32(EL2) && !ELIsInHost(EL2) && CPTR_EL2.TFP == '1' then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif ELIsInHost(EL2) && CPTR_EL2.FPEN IN {'x0'} then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif EL2Enabled() && ELUsingAArch32(EL2) && ((ELUsingAArch32(EL3) && SCR.NS == '1' && NSACR.cp10 == '0') || HCPTR.TCP10 == '1') then AArch32.TakeHypTrapException(0x08); elsif HaveEL(EL3) && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then if EL3SDDUndef() then UNDEFINED; else AArch64.AArch32SystemAccessTrap(EL3, 0x07); else FPSCR = R[t]; elsif PSTATE.EL == EL2 then if HaveEL(EL3) && EL3SDDUndefPriority() && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then UNDEFINED; elsif ELIsInHost(EL2) && CPTR_EL2.FPEN IN {'x0'} then AArch64.AArch32SystemAccessTrap(EL2, 0x07); elsif EL2Enabled() && ((ELUsingAArch32(EL3) && SCR.NS == '1' && NSACR.cp10 == '0') || HCPTR.TCP10 == '1') then AArch32.TakeHypTrapException(0x00); elsif HaveEL(EL3) && !ELUsingAArch32(EL3) && CPTR_EL3.TFP == '1' then if EL3SDDUndef() then UNDEFINED; else AArch64.AArch32SystemAccessTrap(EL3, 0x07); else FPSCR = R[t]; elsif PSTATE.EL == EL3 then if CPACR.cp10 == '00' then UNDEFINED; else FPSCR = R[t];


26/03/2024 09:49; 67c0ae5282a7629ba0ea0ba7267b43cd4f7939f6

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