Generic Opcodes

Note

This documentation does not yet fully account for vectors. Many of the scalar/integer/floating-point operations can also take vectors.

Constants

G_IMPLICIT_DEF

An undefined value.

%0:_(s32) = G_IMPLICIT_DEF

G_CONSTANT

An integer constant.

%0:_(s32) = G_CONSTANT i32 1

G_FCONSTANT

A floating point constant.

%0:_(s32) = G_FCONSTANT float 1.0

G_FRAME_INDEX

The address of an object in the stack frame.

%1:_(p0) = G_FRAME_INDEX %stack.0.ptr0

G_GLOBAL_VALUE

The address of a global value.

%0(p0) = G_GLOBAL_VALUE @var_local

G_BLOCK_ADDR

The address of a basic block.

%0:_(p0) = G_BLOCK_ADDR blockaddress(@test_blockaddress, %ir-block.block)

G_CONSTANT_POOL

The address of an object in the constant pool.

%0:_(p0) = G_CONSTANT_POOL %const.0

Integer Extension and Truncation

G_ANYEXT

Extend the underlying scalar type of an operation, leaving the high bits unspecified.

%1:_(s32) = G_ANYEXT %0:_(s16)

G_SEXT

Sign extend the underlying scalar type of an operation, copying the sign bit into the newly-created space.

%1:_(s32) = G_SEXT %0:_(s16)

G_SEXT_INREG

Sign extend the value from an arbitrary bit position, copying the sign bit into all bits above it. This is equivalent to a shl + ashr pair with an appropriate shift amount. $sz is an immediate (MachineOperand::isImm() returns true) to allow targets to have some bitwidths legal and others lowered. This opcode is particularly useful if the target has sign-extension instructions that are cheaper than the constituent shifts as the optimizer is able to make decisions on whether it’s better to hang on to the G_SEXT_INREG or to lower it and optimize the individual shifts.

%1:_(s32) = G_SEXT_INREG %0:_(s32), 16

G_ZEXT

Zero extend the underlying scalar type of an operation, putting zero bits into the newly-created space.

%1:_(s32) = G_ZEXT %0:_(s16)

G_TRUNC

Truncate the underlying scalar type of an operation. This is equivalent to G_EXTRACT for scalar types, but acts elementwise on vectors.

%1:_(s16) = G_TRUNC %0:_(s32)

Type Conversions

G_INTTOPTR

Convert an integer to a pointer.

%1:_(p0) = G_INTTOPTR %0:_(s32)

G_PTRTOINT

Convert a pointer to an integer.

%1:_(s32) = G_PTRTOINT %0:_(p0)

G_BITCAST

Reinterpret a value as a new type. This is usually done without changing any bits but this is not always the case due a subtlety in the definition of the LLVM-IR Bitcast Instruction. It is allowed to bitcast between pointers with the same size, but different address spaces.

%1:_(s64) = G_BITCAST %0:_(<2 x s32>)

G_ADDRSPACE_CAST

Convert a pointer to an address space to a pointer to another address space.

%1:_(p1) = G_ADDRSPACE_CAST %0:_(p0)

Caution

‘addrspacecast .. to’ Instruction doesn’t mention what happens if the cast is simply invalid (i.e. if the address spaces are disjoint).

Scalar Operations

G_EXTRACT

Extract a register of the specified size, starting from the block given by index. This will almost certainly be mapped to sub-register COPYs after register banks have been selected.

%3:_(s32) = G_EXTRACT %2:_(s64), 32

G_INSERT

Insert a smaller register into a larger one at the specified bit-index.

%2:_(s64) = G_INSERT %0:(_s64), %1:_(s32), 0

G_MERGE_VALUES

Concatenate multiple registers of the same size into a wider register. The input operands are always ordered from lowest bits to highest:

%0:(s32) = G_MERGE_VALUES %bits_0_7:(s8), %bits_8_15:(s8),
                          %bits_16_23:(s8), %bits_24_31:(s8)

G_UNMERGE_VALUES

Extract multiple registers of the specified size, starting from blocks given by indexes. This will almost certainly be mapped to sub-register COPYs after register banks have been selected. The output operands are always ordered from lowest bits to highest:

%bits_0_7:(s8), %bits_8_15:(s8),
    %bits_16_23:(s8), %bits_24_31:(s8) = G_UNMERGE_VALUES %0:(s32)

G_BSWAP

Reverse the order of the bytes in a scalar.

%1:_(s32) = G_BSWAP %0:_(s32)

G_BITREVERSE

Reverse the order of the bits in a scalar.

%1:_(s32) = G_BITREVERSE %0:_(s32)

G_SBFX, G_UBFX

Extract a range of bits from a register.

The source operands are registers as follows:

• Source

• The least-significant bit for the extraction

• The width of the extraction

The least-significant bit (lsb) and width operands are in the range:

0 <= lsb < lsb + width <= source bitwidth, where all values are unsigned

G_SBFX sign-extends the result, while G_UBFX zero-extends the result.

; Extract 5 bits starting at bit 1 from %x and store them in %a.
; Sign-extend the result.
;
; Example:
; %x = 0...0000[10110]1 ---> %a = 1...111111[10110]
%lsb_one = G_CONSTANT i32 1
%width_five = G_CONSTANT i32 5
%a:_(s32) = G_SBFX %x, %lsb_one, %width_five

; Extract 3 bits starting at bit 2 from %x and store them in %b. Zero-extend
; the result.
;
; Example:
; %x = 1...11111[100]11 ---> %b = 0...00000[100]
%lsb_two = G_CONSTANT i32 2
%width_three = G_CONSTANT i32 3
%b:_(s32) = G_UBFX %x, %lsb_two, %width_three

Integer Operations

G_ADD, G_SUB, G_MUL, G_AND, G_OR, G_XOR, G_SDIV, G_UDIV, G_SREM, G_UREM

These each perform their respective integer arithmetic on a scalar.

%dst:_(s32) = G_ADD %src0:_(s32), %src1:_(s32)

The above example adds %src1 to %src0 and stores the result in %dst.

G_SDIVREM, G_UDIVREM

Perform integer division and remainder thereby producing two results.

%div:_(s32), %rem:_(s32) = G_SDIVREM %0:_(s32), %1:_(s32)

G_SADDSAT, G_UADDSAT, G_SSUBSAT, G_USUBSAT, G_SSHLSAT, G_USHLSAT

Signed and unsigned addition, subtraction and left shift with saturation.

%2:_(s32) = G_SADDSAT %0:_(s32), %1:_(s32)

G_SHL, G_LSHR, G_ASHR

Shift the bits of a scalar left or right inserting zeros (sign-bit for G_ASHR).

G_ROTR, G_ROTL

Rotate the bits right (G_ROTR) or left (G_ROTL).

G_ICMP

Perform integer comparison producing non-zero (true) or zero (false). It’s target specific whether a true value is 1, ~0U, or some other non-zero value.

G_SELECT

Select between two values depending on a zero/non-zero value.

%5:_(s32) = G_SELECT %4(s1), %6, %2

G_PTR_ADD

Add a scalar offset in addressible units to a pointer. Addressible units are typically bytes but this may vary between targets.

%1:_(p0) = G_PTR_ADD %0:_(p0), %1:_(s32)

Caution

There are currently no in-tree targets that use this with addressable units not equal to 8 bit.

G_PTRMASK

Zero out an arbitrary mask of bits of a pointer. The mask type must be an integer, and the number of vector elements must match for all operands. This corresponds to i_intr_llvm_ptrmask.

%2:_(p0) = G_PTRMASK %0, %1

G_SMIN, G_SMAX, G_UMIN, G_UMAX

Take the minimum/maximum of two values.

%5:_(s32) = G_SMIN %6, %2

G_ABS

Take the absolute value of a signed integer. The absolute value of the minimum negative value (e.g. the 8-bit value 0x80) is defined to be itself.

%1:_(s32) = G_ABS %0

G_UADDO, G_SADDO, G_USUBO, G_SSUBO, G_SMULO, G_UMULO

Perform the requested arithmetic and produce a carry output in addition to the normal result.

%3:_(s32), %4:_(s1) = G_UADDO %0, %1

G_UADDE, G_SADDE, G_USUBE, G_SSUBE

Perform the requested arithmetic and consume a carry input in addition to the normal input. Also produce a carry output in addition to the normal result.

%4:_(s32), %5:_(s1) = G_UADDE %0, %1, %3:_(s1)

G_UMULH, G_SMULH

Multiply two numbers at twice the incoming bit width (unsigned or signed) and return the high half of the result.

%3:_(s32) = G_UMULH %0, %1

G_CTLZ, G_CTTZ, G_CTPOP

Count leading zeros, trailing zeros, or number of set bits.

%2:_(s33) = G_CTLZ_ZERO_UNDEF %1
%2:_(s33) = G_CTTZ_ZERO_UNDEF %1
%2:_(s33) = G_CTPOP %1

G_CTLZ_ZERO_UNDEF, G_CTTZ_ZERO_UNDEF

Count leading zeros or trailing zeros. If the value is zero then the result is undefined.

%2:_(s33) = G_CTLZ_ZERO_UNDEF %1
%2:_(s33) = G_CTTZ_ZERO_UNDEF %1

Floating Point Operations

G_FCMP

Perform floating point comparison producing non-zero (true) or zero (false). It’s target specific whether a true value is 1, ~0U, or some other non-zero value.

G_FNEG

Floating point negation.

G_FPEXT

Convert a floating point value to a larger type.

G_FPTRUNC

Convert a floating point value to a narrower type.

G_FPTOSI, G_FPTOUI, G_SITOFP, G_UITOFP

Convert between integer and floating point.

G_FABS

Take the absolute value of a floating point value.

G_FCOPYSIGN

Copy the value of the first operand, replacing the sign bit with that of the second operand.

G_FCANONICALIZE

See ‘llvm.canonicalize.*’ Intrinsic.

G_IS_FPCLASS

Tests if the first operand, which must be floating-point scalar or vector, has floating-point class specified by the second operand. Returns non-zero (true) or zero (false). It’s target specific whether a true value is 1, ~0U, or some other non-zero value. If the first operand is a vector, the returned value is a vector of the same length.

G_FMINNUM

Perform floating-point minimum on two values.

In the case where a single input is a NaN (either signaling or quiet), the non-NaN input is returned.

The return value of (FMINNUM 0.0, -0.0) could be either 0.0 or -0.0.

G_FMAXNUM

Perform floating-point maximum on two values.

In the case where a single input is a NaN (either signaling or quiet), the non-NaN input is returned.

The return value of (FMAXNUM 0.0, -0.0) could be either 0.0 or -0.0.

G_FMINNUM_IEEE

Perform floating-point minimum on two values, following the IEEE-754 2008 definition. This differs from FMINNUM in the handling of signaling NaNs. If one input is a signaling NaN, returns a quiet NaN.

G_FMAXNUM_IEEE

Perform floating-point maximum on two values, following the IEEE-754 2008 definition. This differs from FMAXNUM in the handling of signaling NaNs. If one input is a signaling NaN, returns a quiet NaN.

G_FMINIMUM

NaN-propagating minimum that also treat -0.0 as less than 0.0. While FMINNUM_IEEE follow IEEE 754-2008 semantics, FMINIMUM follows IEEE 754-2018 draft semantics.

G_FMAXIMUM

NaN-propagating maximum that also treat -0.0 as less than 0.0. While FMAXNUM_IEEE follow IEEE 754-2008 semantics, FMAXIMUM follows IEEE 754-2018 draft semantics.

G_FADD, G_FSUB, G_FMUL, G_FDIV, G_FREM

Perform the specified floating point arithmetic.

G_FMA

Perform a fused multiply add (i.e. without the intermediate rounding step).

G_FMAD

Perform a non-fused multiply add (i.e. with the intermediate rounding step).

G_FPOW

Raise the first operand to the power of the second.

G_FEXP, G_FEXP2

Calculate the base-e or base-2 exponential of a value

G_FLOG, G_FLOG2, G_FLOG10

Calculate the base-e, base-2, or base-10 respectively.

G_FCEIL, G_FCOS, G_FSIN, G_FSQRT, G_FFLOOR, G_FRINT, G_FNEARBYINT

These correspond to the standard C functions of the same name.

G_INTRINSIC_TRUNC

Returns the operand rounded to the nearest integer not larger in magnitude than the operand.

G_INTRINSIC_ROUND

Returns the operand rounded to the nearest integer.

G_LROUND, G_LLROUND

Returns the source operand rounded to the nearest integer with ties away from zero.

See the LLVM LangRef entry on ‘llvm.lround.*' for details on behaviour.

%rounded_32:_(s32) = G_LROUND %round_me:_(s64)
%rounded_64:_(s64) = G_LLROUND %round_me:_(s64)

Vector Specific Operations

G_CONCAT_VECTORS

Concatenate two vectors to form a longer vector.

G_BUILD_VECTOR, G_BUILD_VECTOR_TRUNC

Create a vector from multiple scalar registers. No implicit conversion is performed (i.e. the result element type must be the same as all source operands)

The _TRUNC version truncates the larger operand types to fit the destination vector elt type.

G_INSERT_VECTOR_ELT

Insert an element into a vector

G_EXTRACT_VECTOR_ELT

Extract an element from a vector

G_SHUFFLE_VECTOR

Concatenate two vectors and shuffle the elements according to the mask operand. The mask operand should be an IR Constant which exactly matches the corresponding mask for the IR shufflevector instruction.

Vector Reduction Operations

These operations represent horizontal vector reduction, producing a scalar result.

G_VECREDUCE_SEQ_FADD, G_VECREDUCE_SEQ_FMUL

The SEQ variants perform reductions in sequential order. The first operand is an initial scalar accumulator value, and the second operand is the vector to reduce.

G_VECREDUCE_FADD, G_VECREDUCE_FMUL

These reductions are relaxed variants which may reduce the elements in any order.

G_VECREDUCE_FMAX, G_VECREDUCE_FMIN, G_VECREDUCE_FMAXIMUM, G_VECREDUCE_FMINIMUM

FMIN/FMAX/FMINIMUM/FMAXIMUM nodes can have flags, for NaN/NoNaN variants.

Integer/bitwise reductions

• G_VECREDUCE_ADD

• G_VECREDUCE_MUL

• G_VECREDUCE_AND

• G_VECREDUCE_OR

• G_VECREDUCE_XOR

• G_VECREDUCE_SMAX

• G_VECREDUCE_SMIN

• G_VECREDUCE_UMAX

• G_VECREDUCE_UMIN

Integer reductions may have a result type larger than the vector element type. However, the reduction is performed using the vector element type and the value in the top bits is unspecified.

Memory Operations

G_LOAD, G_SEXTLOAD, G_ZEXTLOAD

Generic load. Expects a MachineMemOperand in addition to explicit operands. If the result size is larger than the memory size, the high bits are undefined, sign-extended, or zero-extended respectively.

Only G_LOAD is valid if the result is a vector type. If the result is larger than the memory size, the high elements are undefined (i.e. this is not a per-element, vector anyextload)

Unlike in SelectionDAG, atomic loads are expressed with the same opcodes as regular loads. G_LOAD, G_SEXTLOAD and G_ZEXTLOAD may all have atomic memory operands.

G_INDEXED_LOAD

Generic indexed load. Combines a GEP with a load. $newaddr is set to $base + $offset. If $am is 0 (post-indexed), then the value is loaded from $base; if $am is 1 (pre-indexed) then the value is loaded from $newaddr.

G_INDEXED_SEXTLOAD

Same as G_INDEXED_LOAD except that the load performed is sign-extending, as with G_SEXTLOAD.

G_INDEXED_ZEXTLOAD

Same as G_INDEXED_LOAD except that the load performed is zero-extending, as with G_ZEXTLOAD.

G_STORE

Generic store. Expects a MachineMemOperand in addition to explicit operands. If the stored value size is greater than the memory size, the high bits are implicitly truncated. If this is a vector store, the high elements are discarded (i.e. this does not function as a per-lane vector, truncating store)

G_INDEXED_STORE

Combines a store with a GEP. See description of G_INDEXED_LOAD for indexing behaviour.

G_ATOMIC_CMPXCHG_WITH_SUCCESS

Generic atomic cmpxchg with internal success check. Expects a MachineMemOperand in addition to explicit operands.

G_ATOMIC_CMPXCHG

Generic atomic cmpxchg. Expects a MachineMemOperand in addition to explicit operands.

G_ATOMICRMW_XCHG, G_ATOMICRMW_ADD, G_ATOMICRMW_SUB, G_ATOMICRMW_AND, G_ATOMICRMW_NAND, G_ATOMICRMW_OR, G_ATOMICRMW_XOR, G_ATOMICRMW_MAX, G_ATOMICRMW_MIN, G_ATOMICRMW_UMAX, G_ATOMICRMW_UMIN, G_ATOMICRMW_FADD, G_ATOMICRMW_FSUB, G_ATOMICRMW_FMAX, G_ATOMICRMW_FMIN ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Generic atomicrmw. Expects a MachineMemOperand in addition to explicit operands.

G_FENCE

Generic fence. The first operand is the memory ordering. The second operand is the syncscope.

See the LLVM LangRef entry on the ‘fence' instruction for more details.

G_MEMCPY

Generic memcpy. Expects two MachineMemOperands covering the store and load respectively, in addition to explicit operands.

G_MEMCPY_INLINE

Generic inlined memcpy. Like G_MEMCPY, but it is guaranteed that this version will not be lowered as a call to an external function. Currently the size operand is required to evaluate as a constant (not an immediate), though that is expected to change when llvm.memcpy.inline is taught to support dynamic sizes.

G_MEMMOVE

Generic memmove. Similar to G_MEMCPY, but the source and destination memory ranges are allowed to overlap.

G_MEMSET

Generic memset. Expects a MachineMemOperand in addition to explicit operands.

G_BZERO

Generic bzero. Expects a MachineMemOperand in addition to explicit operands.

Control Flow

G_PHI

Implement the φ node in the SSA graph representing the function.

%dst(s8) = G_PHI %src1(s8), %bb.<id1>, %src2(s8), %bb.<id2>

G_BR

Unconditional branch

G_BR %bb.<id>

G_BRCOND

Conditional branch

G_BRCOND %condition, %basicblock.<id>

G_BRINDIRECT

Indirect branch

G_BRINDIRECT %src(p0)

G_BRJT

Indirect branch to jump table entry

G_BRJT %ptr(p0), %jti, %idx(s64)

G_JUMP_TABLE

Generates a pointer to the address of the jump table specified by the source operand. The source operand is a jump table index. G_JUMP_TABLE can be used in conjunction with G_BRJT to support jump table codegen with GlobalISel.

%dst:_(p0) = G_JUMP_TABLE %jump-table.0

The above example generates a pointer to the source jump table index.

G_INVOKE_REGION_START

A marker instruction that acts as a pseudo-terminator for regions of code that may throw exceptions. Being a terminator, it prevents code from being inserted after it during passes like legalization. This is needed because calls to exception throw routines do not return, so no code that must be on an executable path must be placed after throwing.

G_INTRINSIC, G_INTRINSIC_CONVERGENT

Call an intrinsic that has no side-effects.

The _CONVERGENT variant corresponds to an LLVM IR intrinsic marked convergent.

Note

Unlike SelectionDAG, there is no _VOID variant. Both of these are permitted to have zero, one, or multiple results.

G_INTRINSIC_W_SIDE_EFFECTS, G_INTRINSIC_CONVERGENT_W_SIDE_EFFECTS

Call an intrinsic that is considered to have unknown side-effects and as such cannot be reordered across other side-effecting instructions.

The _CONVERGENT variant corresponds to an LLVM IR intrinsic marked convergent.

Note

Unlike SelectionDAG, there is no _VOID variant. Both of these are permitted to have zero, one, or multiple results.

Variadic Arguments

G_VASTART

Caution

I found no documentation for this instruction at the time of writing.

G_VAARG

Caution

I found no documentation for this instruction at the time of writing.

Other Operations

G_DYN_STACKALLOC

Dynamically realigns the stack pointer to the specified size and alignment. An alignment value of 0 or 1 means no specific alignment.

%8:_(p0) = G_DYN_STACKALLOC %7(s64), 32

Optimization Hints

These instructions do not correspond to any target instructions. They act as hints for various combines.

G_ASSERT_SEXT, G_ASSERT_ZEXT

This signifies that the contents of a register were previously extended from a smaller type.

The smaller type is denoted using an immediate operand. For scalars, this is the width of the entire smaller type. For vectors, this is the width of the smaller element type.

%x_was_zexted:_(s32) = G_ASSERT_ZEXT %x(s32), 16
%y_was_zexted:_(<2 x s32>) = G_ASSERT_ZEXT %y(<2 x s32>), 16

%z_was_sexted:_(s32) = G_ASSERT_SEXT %z(s32), 8

G_ASSERT_SEXT and G_ASSERT_ZEXT act like copies, albeit with some restrictions.

The source and destination registers must

• Be virtual

• Belong to the same register class

• Belong to the same register bank

It should always be safe to

• Look through the source register

• Replace the destination register with the source register

Miscellaneous

G_CONSTANT_FOLD_BARRIER

This operation is used as an opaque barrier to prevent constant folding. Combines and other transformations should not look through this. These have no other semantics and can be safely eliminated if a target chooses.