def int_amdgcn_sffbh :
Intrinsic<[llvm_anyint_ty], [LLVMMatchType<0>], [IntrNoMem]>;
-// TODO: Do we want an ordering for these?
-def int_amdgcn_atomic_inc : Intrinsic<[llvm_anyint_ty],
- [llvm_anyptr_ty, LLVMMatchType<0>],
- [IntrArgMemOnly, NoCapture<0>]
->;
-def int_amdgcn_atomic_dec : Intrinsic<[llvm_anyint_ty],
- [llvm_anyptr_ty, LLVMMatchType<0>],
+// Fields should mirror atomicrmw
+class AMDGPUAtomicIncIntrin : Intrinsic<[llvm_anyint_ty],
+ [llvm_anyptr_ty,
+ LLVMMatchType<0>,
+ llvm_i32_ty, // ordering
+ llvm_i32_ty, // scope
+ llvm_i1_ty], // isVolatile
[IntrArgMemOnly, NoCapture<0>]
>;
+def int_amdgcn_atomic_inc : AMDGPUAtomicIncIntrin;
+def int_amdgcn_atomic_dec : AMDGPUAtomicIncIntrin;
+
class AMDGPUImageLoad : Intrinsic <
[llvm_anyfloat_ty], // vdata(VGPR)
[llvm_anyint_ty, // vaddr(VGPR)
unsigned IntrID) const {
switch (IntrID) {
case Intrinsic::amdgcn_atomic_inc:
- case Intrinsic::amdgcn_atomic_dec:
+ case Intrinsic::amdgcn_atomic_dec: {
Info.opc = ISD::INTRINSIC_W_CHAIN;
Info.memVT = MVT::getVT(CI.getType());
Info.ptrVal = CI.getOperand(0);
Info.align = 0;
- Info.vol = false;
+
+ const ConstantInt *Vol = dyn_cast<ConstantInt>(CI.getOperand(4));
+ Info.vol = !Vol || !Vol->isNullValue();
Info.readMem = true;
Info.writeMem = true;
return true;
+ }
default:
return false;
}
Module *M = II->getParent()->getParent()->getParent();
switch (II->getIntrinsicID()) {
- case Intrinsic::objectsize:
case Intrinsic::amdgcn_atomic_inc:
- case Intrinsic::amdgcn_atomic_dec: {
+ case Intrinsic::amdgcn_atomic_dec:{
+ const ConstantInt *IsVolatile = dyn_cast<ConstantInt>(II->getArgOperand(4));
+ if (!IsVolatile || !IsVolatile->isNullValue())
+ return false;
+
+ LLVM_FALLTHROUGH;
+ }
+ case Intrinsic::objectsize: {
Type *DestTy = II->getType();
Type *SrcTy = NewV->getType();
Function *NewDecl =
ret {i32, i1} %orig
}
-; CHECK: DIVERGENT: %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %ptr, i32 %val)
+; CHECK: DIVERGENT: %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %ptr, i32 %val, i32 0, i32 0, i1 false)
define i32 @test_atomic_inc_i32(i32 addrspace(1)* %ptr, i32 %val) #0 {
- %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %ptr, i32 %val)
+ %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %ptr, i32 %val, i32 0, i32 0, i1 false)
ret i32 %ret
}
-; CHECK: DIVERGENT: %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %ptr, i64 %val)
+; CHECK: DIVERGENT: %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %ptr, i64 %val, i32 0, i32 0, i1 false)
define i64 @test_atomic_inc_i64(i64 addrspace(1)* %ptr, i64 %val) #0 {
- %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %ptr, i64 %val)
+ %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %ptr, i64 %val, i32 0, i32 0, i1 false)
ret i64 %ret
}
-; CHECK: DIVERGENT: %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %ptr, i32 %val)
+; CHECK: DIVERGENT: %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %ptr, i32 %val, i32 0, i32 0, i1 false)
define i32 @test_atomic_dec_i32(i32 addrspace(1)* %ptr, i32 %val) #0 {
- %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %ptr, i32 %val)
+ %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %ptr, i32 %val, i32 0, i32 0, i1 false)
ret i32 %ret
}
-; CHECK: DIVERGENT: %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %ptr, i64 %val)
+; CHECK: DIVERGENT: %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %ptr, i64 %val, i32 0, i32 0, i1 false)
define i64 @test_atomic_dec_i64(i64 addrspace(1)* %ptr, i64 %val) #0 {
- %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %ptr, i64 %val)
+ %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %ptr, i64 %val, i32 0, i32 0, i1 false)
ret i64 %ret
}
-declare i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* nocapture, i32) #1
-declare i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* nocapture, i64) #1
-declare i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* nocapture, i32) #1
-declare i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* nocapture, i64) #1
+declare i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* nocapture, i32, i32, i32, i1) #1
+declare i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* nocapture, i64, i32, i32, i1) #1
+declare i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* nocapture, i32, i32, i32, i1) #1
+declare i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* nocapture, i64, i32, i32, i1) #1
attributes #0 = { nounwind }
attributes #1 = { nounwind argmemonly }
; OPT: %sunkaddr = ptrtoint i32 addrspace(3)* %in to i32
; OPT: %sunkaddr1 = add i32 %sunkaddr, 28
; OPT: %sunkaddr2 = inttoptr i32 %sunkaddr1 to i32 addrspace(3)*
-; OPT: %tmp1 = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %sunkaddr2, i32 2)
+; OPT: %tmp1 = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %sunkaddr2, i32 2, i32 0, i32 0, i1 false)
define amdgpu_kernel void @test_sink_local_small_offset_atomic_inc_i32(i32 addrspace(3)* %out, i32 addrspace(3)* %in) {
entry:
%out.gep = getelementptr i32, i32 addrspace(3)* %out, i32 999999
br i1 %tmp0, label %endif, label %if
if:
- %tmp1 = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %in.gep, i32 2)
+ %tmp1 = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %in.gep, i32 2, i32 0, i32 0, i1 false)
br label %endif
endif:
; OPT: %sunkaddr = ptrtoint i32 addrspace(3)* %in to i32
; OPT: %sunkaddr1 = add i32 %sunkaddr, 28
; OPT: %sunkaddr2 = inttoptr i32 %sunkaddr1 to i32 addrspace(3)*
-; OPT: %tmp1 = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %sunkaddr2, i32 2)
+; OPT: %tmp1 = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %sunkaddr2, i32 2, i32 0, i32 0, i1 false)
define amdgpu_kernel void @test_sink_local_small_offset_atomic_dec_i32(i32 addrspace(3)* %out, i32 addrspace(3)* %in) {
entry:
%out.gep = getelementptr i32, i32 addrspace(3)* %out, i32 999999
br i1 %tmp0, label %endif, label %if
if:
- %tmp1 = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %in.gep, i32 2)
+ %tmp1 = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %in.gep, i32 2, i32 0, i32 0, i1 false)
br label %endif
endif:
}
declare i32 @llvm.amdgcn.mbcnt.lo(i32, i32) #0
-declare i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* nocapture, i32) #2
-declare i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* nocapture, i32) #2
+declare i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* nocapture, i32, i32, i32, i1) #2
+declare i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* nocapture, i32, i32, i32, i1) #2
attributes #0 = { nounwind readnone }
attributes #1 = { nounwind }
; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=CI %s
; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
-declare i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* nocapture, i32) #2
-declare i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* nocapture, i32) #2
-declare i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* nocapture, i32) #2
+declare i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* nocapture, i32, i32, i32, i1) #2
+declare i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* nocapture, i32, i32, i32, i1) #2
+declare i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* nocapture, i32, i32, i32, i1) #2
-declare i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* nocapture, i64) #2
-declare i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* nocapture, i64) #2
-declare i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* nocapture, i64) #2
+declare i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* nocapture, i64, i32, i32, i1) #2
+declare i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* nocapture, i64, i32, i32, i1) #2
+declare i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* nocapture, i64, i32, i32, i1) #2
declare i32 @llvm.amdgcn.workitem.id.x() #1
+; Make sure no crash on invalid non-constant
+; GCN-LABEL: {{^}}invalid_variable_order_lds_atomic_dec_ret_i32:
+define amdgpu_kernel void @invalid_variable_order_lds_atomic_dec_ret_i32(i32 addrspace(1)* %out, i32 addrspace(3)* %ptr, i32 %order.var) #0 {
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %ptr, i32 42, i32 %order.var, i32 0, i1 false)
+ store i32 %result, i32 addrspace(1)* %out
+ ret void
+}
+
+; Make sure no crash on invalid non-constant
+; GCN-LABEL: {{^}}invalid_variable_scope_lds_atomic_dec_ret_i32:
+define amdgpu_kernel void @invalid_variable_scope_lds_atomic_dec_ret_i32(i32 addrspace(1)* %out, i32 addrspace(3)* %ptr, i32 %scope.var) #0 {
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %ptr, i32 42, i32 0, i32 %scope.var, i1 false)
+ store i32 %result, i32 addrspace(1)* %out
+ ret void
+}
+
+; Make sure no crash on invalid non-constant
+; GCN-LABEL: {{^}}invalid_variable_volatile_lds_atomic_dec_ret_i32:
+define amdgpu_kernel void @invalid_variable_volatile_lds_atomic_dec_ret_i32(i32 addrspace(1)* %out, i32 addrspace(3)* %ptr, i1 %volatile.var) #0 {
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %ptr, i32 42, i32 0, i32 0, i1 %volatile.var)
+ store i32 %result, i32 addrspace(1)* %out
+ ret void
+}
+
; GCN-LABEL: {{^}}lds_atomic_dec_ret_i32:
; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
; GCN: ds_dec_rtn_u32 v{{[0-9]+}}, v{{[0-9]+}}, [[K]]
define amdgpu_kernel void @lds_atomic_dec_ret_i32(i32 addrspace(1)* %out, i32 addrspace(3)* %ptr) #0 {
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %ptr, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out
ret void
}
; GCN: ds_dec_rtn_u32 v{{[0-9]+}}, v{{[0-9]+}}, [[K]] offset:16
define amdgpu_kernel void @lds_atomic_dec_ret_i32_offset(i32 addrspace(1)* %out, i32 addrspace(3)* %ptr) #0 {
%gep = getelementptr i32, i32 addrspace(3)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out
ret void
}
; GCN: v_mov_b32_e32 [[VPTR:v[0-9]+]], [[SPTR]]
; GCN: ds_dec_u32 [[VPTR]], [[DATA]]
define amdgpu_kernel void @lds_atomic_dec_noret_i32(i32 addrspace(3)* %ptr) nounwind {
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %ptr, i32 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: ds_dec_u32 v{{[0-9]+}}, [[K]] offset:16
define amdgpu_kernel void @lds_atomic_dec_noret_i32_offset(i32 addrspace(3)* %ptr) nounwind {
%gep = getelementptr i32, i32 addrspace(3)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
; GCN: buffer_atomic_dec [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
define amdgpu_kernel void @global_atomic_dec_ret_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %ptr) #0 {
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %ptr, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out
ret void
}
; GCN: buffer_atomic_dec [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16 glc{{$}}
define amdgpu_kernel void @global_atomic_dec_ret_i32_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %ptr) #0 {
%gep = getelementptr i32, i32 addrspace(1)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}global_atomic_dec_noret_i32:
; GCN: buffer_atomic_dec [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
define amdgpu_kernel void @global_atomic_dec_noret_i32(i32 addrspace(1)* %ptr) nounwind {
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %ptr, i32 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: buffer_atomic_dec [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
define amdgpu_kernel void @global_atomic_dec_noret_i32_offset(i32 addrspace(1)* %ptr) nounwind {
%gep = getelementptr i32, i32 addrspace(1)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
%gep.tid = getelementptr i32, i32 addrspace(1)* %ptr, i32 %id
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id
%gep = getelementptr i32, i32 addrspace(1)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out.gep
ret void
}
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i32, i32 addrspace(1)* %ptr, i32 %id
%gep = getelementptr i32, i32 addrspace(1)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
; GCN: flat_atomic_dec v{{[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, [[K]] glc{{$}}
define amdgpu_kernel void @flat_atomic_dec_ret_i32(i32 addrspace(4)* %out, i32 addrspace(4)* %ptr) #0 {
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %ptr, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(4)* %out
ret void
}
; GCN: flat_atomic_dec v{{[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, [[K]] glc{{$}}
define amdgpu_kernel void @flat_atomic_dec_ret_i32_offset(i32 addrspace(4)* %out, i32 addrspace(4)* %ptr) #0 {
%gep = getelementptr i32, i32 addrspace(4)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(4)* %out
ret void
}
; FUNC-LABEL: {{^}}flat_atomic_dec_noret_i32:
; GCN: flat_atomic_dec v{{\[[0-9]+:[0-9]+\]}}, [[K]]{{$}}
define amdgpu_kernel void @flat_atomic_dec_noret_i32(i32 addrspace(4)* %ptr) nounwind {
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %ptr, i32 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: flat_atomic_dec v{{\[[0-9]+:[0-9]+\]}}, [[K]]{{$}}
define amdgpu_kernel void @flat_atomic_dec_noret_i32_offset(i32 addrspace(4)* %ptr) nounwind {
%gep = getelementptr i32, i32 addrspace(4)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
%gep.tid = getelementptr i32, i32 addrspace(4)* %ptr, i32 %id
%out.gep = getelementptr i32, i32 addrspace(4)* %out, i32 %id
%gep = getelementptr i32, i32 addrspace(4)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(4)* %out.gep
ret void
}
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i32, i32 addrspace(4)* %ptr, i32 %id
%gep = getelementptr i32, i32 addrspace(4)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: flat_atomic_dec_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} glc{{$}}
define amdgpu_kernel void @flat_atomic_dec_ret_i64(i64 addrspace(4)* %out, i64 addrspace(4)* %ptr) #0 {
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %ptr, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(4)* %out
ret void
}
; GCN: flat_atomic_dec_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} glc{{$}}
define amdgpu_kernel void @flat_atomic_dec_ret_i64_offset(i64 addrspace(4)* %out, i64 addrspace(4)* %ptr) #0 {
%gep = getelementptr i64, i64 addrspace(4)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(4)* %out
ret void
}
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: flat_atomic_dec_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]$}}
define amdgpu_kernel void @flat_atomic_dec_noret_i64(i64 addrspace(4)* %ptr) nounwind {
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %ptr, i64 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: flat_atomic_dec_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]$}}
define amdgpu_kernel void @flat_atomic_dec_noret_i64_offset(i64 addrspace(4)* %ptr) nounwind {
%gep = getelementptr i64, i64 addrspace(4)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
%gep.tid = getelementptr i64, i64 addrspace(4)* %ptr, i32 %id
%out.gep = getelementptr i64, i64 addrspace(4)* %out, i32 %id
%gep = getelementptr i64, i64 addrspace(4)* %gep.tid, i32 5
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(4)* %out.gep
ret void
}
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i64, i64 addrspace(4)* %ptr, i32 %id
%gep = getelementptr i64, i64 addrspace(4)* %gep.tid, i32 5
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
%tid.x = tail call i32 @llvm.amdgcn.workitem.id.x() #1
%idx.0 = add nsw i32 %tid.x, 2
%arrayidx0 = getelementptr inbounds [512 x i32], [512 x i32] addrspace(3)* @lds0, i32 0, i32 %idx.0
- %val0 = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %arrayidx0, i32 9)
+ %val0 = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %arrayidx0, i32 9, i32 0, i32 0, i1 false)
store i32 %idx.0, i32 addrspace(1)* %add_use
store i32 %val0, i32 addrspace(1)* %out
ret void
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: ds_dec_rtn_u64 v{{\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}}{{$}}
define amdgpu_kernel void @lds_atomic_dec_ret_i64(i64 addrspace(1)* %out, i64 addrspace(3)* %ptr) #0 {
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %ptr, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out
ret void
}
; GCN: ds_dec_rtn_u64 v{{\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} offset:32
define amdgpu_kernel void @lds_atomic_dec_ret_i64_offset(i64 addrspace(1)* %out, i64 addrspace(3)* %ptr) #0 {
%gep = getelementptr i64, i64 addrspace(3)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out
ret void
}
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: ds_dec_u64 v{{[0-9]+}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}}{{$}}
define amdgpu_kernel void @lds_atomic_dec_noret_i64(i64 addrspace(3)* %ptr) nounwind {
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %ptr, i64 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: ds_dec_u64 v{{[0-9]+}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} offset:32{{$}}
define amdgpu_kernel void @lds_atomic_dec_noret_i64_offset(i64 addrspace(3)* %ptr) nounwind {
%gep = getelementptr i64, i64 addrspace(3)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
define amdgpu_kernel void @global_atomic_dec_ret_i64(i64 addrspace(1)* %out, i64 addrspace(1)* %ptr) #0 {
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %ptr, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out
ret void
}
; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32 glc{{$}}
define amdgpu_kernel void @global_atomic_dec_ret_i64_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %ptr) #0 {
%gep = getelementptr i64, i64 addrspace(1)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out
ret void
}
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
define amdgpu_kernel void @global_atomic_dec_noret_i64(i64 addrspace(1)* %ptr) nounwind {
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %ptr, i64 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32{{$}}
define amdgpu_kernel void @global_atomic_dec_noret_i64_offset(i64 addrspace(1)* %ptr) nounwind {
%gep = getelementptr i64, i64 addrspace(1)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
%gep.tid = getelementptr i64, i64 addrspace(1)* %ptr, i32 %id
%out.gep = getelementptr i64, i64 addrspace(1)* %out, i32 %id
%gep = getelementptr i64, i64 addrspace(1)* %gep.tid, i32 5
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out.gep
ret void
}
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i64, i64 addrspace(1)* %ptr, i32 %id
%gep = getelementptr i64, i64 addrspace(1)* %gep.tid, i32 5
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
%tid.x = tail call i32 @llvm.amdgcn.workitem.id.x() #1
%idx.0 = add nsw i32 %tid.x, 2
%arrayidx0 = getelementptr inbounds [512 x i64], [512 x i64] addrspace(3)* @lds1, i32 0, i32 %idx.0
- %val0 = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %arrayidx0, i64 9)
+ %val0 = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %arrayidx0, i64 9, i32 0, i32 0, i1 false)
store i32 %idx.0, i32 addrspace(1)* %add_use
store i64 %val0, i64 addrspace(1)* %out
ret void
; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=CI %s
; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
-declare i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* nocapture, i32) #2
-declare i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* nocapture, i32) #2
-declare i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* nocapture, i32) #2
+declare i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* nocapture, i32, i32, i32, i1) #2
+declare i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* nocapture, i32, i32, i32, i1) #2
+declare i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* nocapture, i32, i32, i32, i1) #2
-declare i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* nocapture, i64) #2
-declare i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* nocapture, i64) #2
-declare i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* nocapture, i64) #2
+declare i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* nocapture, i64, i32, i32, i1) #2
+declare i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* nocapture, i64, i32, i32, i1) #2
+declare i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* nocapture, i64, i32, i32, i1) #2
declare i32 @llvm.amdgcn.workitem.id.x() #1
; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
; GCN: ds_inc_rtn_u32 v{{[0-9]+}}, v{{[0-9]+}}, [[K]]
define amdgpu_kernel void @lds_atomic_inc_ret_i32(i32 addrspace(1)* %out, i32 addrspace(3)* %ptr) #0 {
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %ptr, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out
ret void
}
; GCN: ds_inc_rtn_u32 v{{[0-9]+}}, v{{[0-9]+}}, [[K]] offset:16
define amdgpu_kernel void @lds_atomic_inc_ret_i32_offset(i32 addrspace(1)* %out, i32 addrspace(3)* %ptr) #0 {
%gep = getelementptr i32, i32 addrspace(3)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out
ret void
}
; GCN: v_mov_b32_e32 [[VPTR:v[0-9]+]], [[SPTR]]
; GCN: ds_inc_u32 [[VPTR]], [[DATA]]
define amdgpu_kernel void @lds_atomic_inc_noret_i32(i32 addrspace(3)* %ptr) nounwind {
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %ptr, i32 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: ds_inc_u32 v{{[0-9]+}}, [[K]] offset:16
define amdgpu_kernel void @lds_atomic_inc_noret_i32_offset(i32 addrspace(3)* %ptr) nounwind {
%gep = getelementptr i32, i32 addrspace(3)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
; GCN: buffer_atomic_inc [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
define amdgpu_kernel void @global_atomic_inc_ret_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %ptr) #0 {
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %ptr, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out
ret void
}
; GCN: buffer_atomic_inc [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16 glc{{$}}
define amdgpu_kernel void @global_atomic_inc_ret_i32_offset(i32 addrspace(1)* %out, i32 addrspace(1)* %ptr) #0 {
%gep = getelementptr i32, i32 addrspace(1)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}global_atomic_inc_noret_i32:
; GCN: buffer_atomic_inc [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
define amdgpu_kernel void @global_atomic_inc_noret_i32(i32 addrspace(1)* %ptr) nounwind {
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %ptr, i32 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: buffer_atomic_inc [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
define amdgpu_kernel void @global_atomic_inc_noret_i32_offset(i32 addrspace(1)* %ptr) nounwind {
%gep = getelementptr i32, i32 addrspace(1)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
%gep.tid = getelementptr i32, i32 addrspace(1)* %ptr, i32 %id
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id
%gep = getelementptr i32, i32 addrspace(1)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out.gep
ret void
}
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i32, i32 addrspace(1)* %ptr, i32 %id
%gep = getelementptr i32, i32 addrspace(1)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
%tid.x = tail call i32 @llvm.amdgcn.workitem.id.x() #1
%idx.0 = add nsw i32 %tid.x, 2
%arrayidx0 = getelementptr inbounds [512 x i32], [512 x i32] addrspace(3)* @lds0, i32 0, i32 %idx.0
- %val0 = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %arrayidx0, i32 9)
+ %val0 = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %arrayidx0, i32 9, i32 0, i32 0, i1 false)
store i32 %idx.0, i32 addrspace(1)* %add_use
store i32 %val0, i32 addrspace(1)* %out
ret void
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: ds_inc_rtn_u64 v{{\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}}{{$}}
define amdgpu_kernel void @lds_atomic_inc_ret_i64(i64 addrspace(1)* %out, i64 addrspace(3)* %ptr) #0 {
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %ptr, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out
ret void
}
; GCN: ds_inc_rtn_u64 v{{\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} offset:32
define amdgpu_kernel void @lds_atomic_inc_ret_i64_offset(i64 addrspace(1)* %out, i64 addrspace(3)* %ptr) #0 {
%gep = getelementptr i64, i64 addrspace(3)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out
ret void
}
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: ds_inc_u64 v{{[0-9]+}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}}{{$}}
define amdgpu_kernel void @lds_atomic_inc_noret_i64(i64 addrspace(3)* %ptr) nounwind {
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %ptr, i64 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: ds_inc_u64 v{{[0-9]+}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} offset:32{{$}}
define amdgpu_kernel void @lds_atomic_inc_noret_i64_offset(i64 addrspace(3)* %ptr) nounwind {
%gep = getelementptr i64, i64 addrspace(3)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 glc{{$}}
define amdgpu_kernel void @global_atomic_inc_ret_i64(i64 addrspace(1)* %out, i64 addrspace(1)* %ptr) #0 {
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %ptr, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out
ret void
}
; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32 glc{{$}}
define amdgpu_kernel void @global_atomic_inc_ret_i64_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %ptr) #0 {
%gep = getelementptr i64, i64 addrspace(1)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out
ret void
}
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0{{$}}
define amdgpu_kernel void @global_atomic_inc_noret_i64(i64 addrspace(1)* %ptr) nounwind {
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %ptr, i64 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32{{$}}
define amdgpu_kernel void @global_atomic_inc_noret_i64_offset(i64 addrspace(1)* %ptr) nounwind {
%gep = getelementptr i64, i64 addrspace(1)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
%gep.tid = getelementptr i64, i64 addrspace(1)* %ptr, i32 %id
%out.gep = getelementptr i64, i64 addrspace(1)* %out, i32 %id
%gep = getelementptr i64, i64 addrspace(1)* %gep.tid, i32 5
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out.gep
ret void
}
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i64, i64 addrspace(1)* %ptr, i32 %id
%gep = getelementptr i64, i64 addrspace(1)* %gep.tid, i32 5
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: v_mov_b32_e32 [[K:v[0-9]+]], 42
; GCN: flat_atomic_inc v{{[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, [[K]] glc{{$}}
define amdgpu_kernel void @flat_atomic_inc_ret_i32(i32 addrspace(4)* %out, i32 addrspace(4)* %ptr) #0 {
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %ptr, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(4)* %out
ret void
}
; GCN: flat_atomic_inc v{{[0-9]+}}, v{{\[[0-9]+:[0-9]+\]}}, [[K]] glc{{$}}
define amdgpu_kernel void @flat_atomic_inc_ret_i32_offset(i32 addrspace(4)* %out, i32 addrspace(4)* %ptr) #0 {
%gep = getelementptr i32, i32 addrspace(4)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(4)* %out
ret void
}
; FUNC-LABEL: {{^}}flat_atomic_inc_noret_i32:
; GCN: flat_atomic_inc v{{\[[0-9]+:[0-9]+\]}}, [[K]]{{$}}
define amdgpu_kernel void @flat_atomic_inc_noret_i32(i32 addrspace(4)* %ptr) nounwind {
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %ptr, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %ptr, i32 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: flat_atomic_inc v{{\[[0-9]+:[0-9]+\]}}, [[K]]{{$}}
define amdgpu_kernel void @flat_atomic_inc_noret_i32_offset(i32 addrspace(4)* %ptr) nounwind {
%gep = getelementptr i32, i32 addrspace(4)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
%gep.tid = getelementptr i32, i32 addrspace(4)* %ptr, i32 %id
%out.gep = getelementptr i32, i32 addrspace(4)* %out, i32 %id
%gep = getelementptr i32, i32 addrspace(4)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(4)* %out.gep
ret void
}
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i32, i32 addrspace(4)* %ptr, i32 %id
%gep = getelementptr i32, i32 addrspace(4)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
%tid.x = tail call i32 @llvm.amdgcn.workitem.id.x() #1
%idx.0 = add nsw i32 %tid.x, 2
%arrayidx0 = getelementptr inbounds [512 x i64], [512 x i64] addrspace(3)* @lds1, i32 0, i32 %idx.0
- %val0 = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %arrayidx0, i64 9)
+ %val0 = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %arrayidx0, i64 9, i32 0, i32 0, i1 false)
store i32 %idx.0, i32 addrspace(1)* %add_use
store i64 %val0, i64 addrspace(1)* %out
ret void
}
-attributes #0 = { nounwind }
-attributes #1 = { nounwind readnone }
-attributes #2 = { nounwind argmemonly }
-
-
-
-
-
-
; GCN-LABEL: {{^}}flat_atomic_inc_ret_i64:
; GCN-DAG: v_mov_b32_e32 v[[KLO:[0-9]+]], 42
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: flat_atomic_inc_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} glc{{$}}
define amdgpu_kernel void @flat_atomic_inc_ret_i64(i64 addrspace(4)* %out, i64 addrspace(4)* %ptr) #0 {
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %ptr, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(4)* %out
ret void
}
; GCN: flat_atomic_inc_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} glc{{$}}
define amdgpu_kernel void @flat_atomic_inc_ret_i64_offset(i64 addrspace(4)* %out, i64 addrspace(4)* %ptr) #0 {
%gep = getelementptr i64, i64 addrspace(4)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(4)* %out
ret void
}
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; GCN: flat_atomic_inc_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]$}}
define amdgpu_kernel void @flat_atomic_inc_noret_i64(i64 addrspace(4)* %ptr) nounwind {
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %ptr, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %ptr, i64 42, i32 0, i32 0, i1 false)
ret void
}
; GCN: flat_atomic_inc_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]$}}
define amdgpu_kernel void @flat_atomic_inc_noret_i64_offset(i64 addrspace(4)* %ptr) nounwind {
%gep = getelementptr i64, i64 addrspace(4)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
%gep.tid = getelementptr i64, i64 addrspace(4)* %ptr, i32 %id
%out.gep = getelementptr i64, i64 addrspace(4)* %out, i32 %id
%gep = getelementptr i64, i64 addrspace(4)* %gep.tid, i32 5
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(4)* %out.gep
ret void
}
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i64, i64 addrspace(4)* %ptr, i32 %id
%gep = getelementptr i64, i64 addrspace(4)* %gep.tid, i32 5
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
+
+attributes #0 = { nounwind }
+attributes #1 = { nounwind readnone }
+attributes #2 = { nounwind argmemonly }
}
; CHECK-LABEL: @atomicinc_global_to_flat_i32(
-; CHECK: call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %global.ptr, i32 %y)
+; CHECK: call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %global.ptr, i32 %y, i32 0, i32 0, i1 false)
define i32 @atomicinc_global_to_flat_i32(i32 addrspace(1)* %global.ptr, i32 %y) #0 {
%cast = addrspacecast i32 addrspace(1)* %global.ptr to i32 addrspace(4)*
- %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %cast, i32 %y)
+ %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %cast, i32 %y, i32 0, i32 0, i1 false)
ret i32 %ret
}
; CHECK-LABEL: @atomicinc_group_to_flat_i32(
-; CHECK: %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %group.ptr, i32 %y)
+; CHECK: %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %group.ptr, i32 %y, i32 0, i32 0, i1 false)
define i32 @atomicinc_group_to_flat_i32(i32 addrspace(3)* %group.ptr, i32 %y) #0 {
%cast = addrspacecast i32 addrspace(3)* %group.ptr to i32 addrspace(4)*
- %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %cast, i32 %y)
+ %ret = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %cast, i32 %y, i32 0, i32 0, i1 false)
ret i32 %ret
}
; CHECK-LABEL: @atomicinc_global_to_flat_i64(
-; CHECK: call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %global.ptr, i64 %y)
+; CHECK: call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %global.ptr, i64 %y, i32 0, i32 0, i1 false)
define i64 @atomicinc_global_to_flat_i64(i64 addrspace(1)* %global.ptr, i64 %y) #0 {
%cast = addrspacecast i64 addrspace(1)* %global.ptr to i64 addrspace(4)*
- %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %cast, i64 %y)
+ %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %cast, i64 %y, i32 0, i32 0, i1 false)
ret i64 %ret
}
; CHECK-LABEL: @atomicinc_group_to_flat_i64(
-; CHECK: call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %group.ptr, i64 %y)
+; CHECK: call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %group.ptr, i64 %y, i32 0, i32 0, i1 false)
define i64 @atomicinc_group_to_flat_i64(i64 addrspace(3)* %group.ptr, i64 %y) #0 {
%cast = addrspacecast i64 addrspace(3)* %group.ptr to i64 addrspace(4)*
- %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %cast, i64 %y)
+ %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %cast, i64 %y, i32 0, i32 0, i1 false)
ret i64 %ret
}
; CHECK-LABEL: @atomicdec_global_to_flat_i32(
-; CHECK: call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %global.ptr, i32 %val)
+; CHECK: call i32 @llvm.amdgcn.atomic.dec.i32.p1i32(i32 addrspace(1)* %global.ptr, i32 %val, i32 0, i32 0, i1 false)
define i32 @atomicdec_global_to_flat_i32(i32 addrspace(1)* %global.ptr, i32 %val) #0 {
%cast = addrspacecast i32 addrspace(1)* %global.ptr to i32 addrspace(4)*
- %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %cast, i32 %val)
+ %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %cast, i32 %val, i32 0, i32 0, i1 false)
ret i32 %ret
}
; CHECK-LABEL: @atomicdec_group_to_flat_i32(
-; CHECK: %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %group.ptr, i32 %val)
+; CHECK: %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %group.ptr, i32 %val, i32 0, i32 0, i1 false)
define i32 @atomicdec_group_to_flat_i32(i32 addrspace(3)* %group.ptr, i32 %val) #0 {
%cast = addrspacecast i32 addrspace(3)* %group.ptr to i32 addrspace(4)*
- %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %cast, i32 %val)
+ %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %cast, i32 %val, i32 0, i32 0, i1 false)
ret i32 %ret
}
; CHECK-LABEL: @atomicdec_global_to_flat_i64(
-; CHECK: call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %global.ptr, i64 %y)
+; CHECK: call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %global.ptr, i64 %y, i32 0, i32 0, i1 false)
define i64 @atomicdec_global_to_flat_i64(i64 addrspace(1)* %global.ptr, i64 %y) #0 {
%cast = addrspacecast i64 addrspace(1)* %global.ptr to i64 addrspace(4)*
- %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %cast, i64 %y)
+ %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %cast, i64 %y, i32 0, i32 0, i1 false)
ret i64 %ret
}
; CHECK-LABEL: @atomicdec_group_to_flat_i64(
-; CHECK: call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %group.ptr, i64 %y)
+; CHECK: call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %group.ptr, i64 %y, i32 0, i32 0, i1 false
define i64 @atomicdec_group_to_flat_i64(i64 addrspace(3)* %group.ptr, i64 %y) #0 {
%cast = addrspacecast i64 addrspace(3)* %group.ptr to i64 addrspace(4)*
- %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %cast, i64 %y)
+ %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %cast, i64 %y, i32 0, i32 0, i1 false)
+ ret i64 %ret
+}
+
+; CHECK-LABEL: @volatile_atomicinc_group_to_flat_i64(
+; CHECK-NEXT: %1 = addrspacecast i64 addrspace(3)* %group.ptr to i64 addrspace(4)*
+; CHECK-NEXT: %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %1, i64 %y, i32 0, i32 0, i1 true)
+define i64 @volatile_atomicinc_group_to_flat_i64(i64 addrspace(3)* %group.ptr, i64 %y) #0 {
+ %cast = addrspacecast i64 addrspace(3)* %group.ptr to i64 addrspace(4)*
+ %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %cast, i64 %y, i32 0, i32 0, i1 true)
+ ret i64 %ret
+}
+
+; CHECK-LABEL: @volatile_atomicdec_global_to_flat_i32(
+; CHECK-NEXT: %1 = addrspacecast i32 addrspace(1)* %global.ptr to i32 addrspace(4)*
+; CHECK-NEXT: %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %1, i32 %val, i32 0, i32 0, i1 true)
+define i32 @volatile_atomicdec_global_to_flat_i32(i32 addrspace(1)* %global.ptr, i32 %val) #0 {
+ %cast = addrspacecast i32 addrspace(1)* %global.ptr to i32 addrspace(4)*
+ %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %cast, i32 %val, i32 0, i32 0, i1 true)
+ ret i32 %ret
+}
+
+; CHECK-LABEL: @volatile_atomicdec_group_to_flat_i32(
+; CHECK-NEXT: %1 = addrspacecast i32 addrspace(3)* %group.ptr to i32 addrspace(4)*
+; CHECK-NEXT: %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %1, i32 %val, i32 0, i32 0, i1 true)
+define i32 @volatile_atomicdec_group_to_flat_i32(i32 addrspace(3)* %group.ptr, i32 %val) #0 {
+ %cast = addrspacecast i32 addrspace(3)* %group.ptr to i32 addrspace(4)*
+ %ret = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %cast, i32 %val, i32 0, i32 0, i1 true)
+ ret i32 %ret
+}
+
+; CHECK-LABEL: @volatile_atomicdec_global_to_flat_i64(
+; CHECK-NEXT: %1 = addrspacecast i64 addrspace(1)* %global.ptr to i64 addrspace(4)*
+; CHECK-NEXT: %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %1, i64 %y, i32 0, i32 0, i1 true)
+define i64 @volatile_atomicdec_global_to_flat_i64(i64 addrspace(1)* %global.ptr, i64 %y) #0 {
+ %cast = addrspacecast i64 addrspace(1)* %global.ptr to i64 addrspace(4)*
+ %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %cast, i64 %y, i32 0, i32 0, i1 true)
+ ret i64 %ret
+}
+
+; CHECK-LABEL: @volatile_atomicdec_group_to_flat_i64(
+; CHECK-NEXT: %1 = addrspacecast i64 addrspace(3)* %group.ptr to i64 addrspace(4)*
+; CHECK-NEXT: %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %1, i64 %y, i32 0, i32 0, i1 true)
+define i64 @volatile_atomicdec_group_to_flat_i64(i64 addrspace(3)* %group.ptr, i64 %y) #0 {
+ %cast = addrspacecast i64 addrspace(3)* %group.ptr to i64 addrspace(4)*
+ %ret = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %cast, i64 %y, i32 0, i32 0, i1 true)
+ ret i64 %ret
+}
+
+; CHECK-LABEL: @invalid_variable_volatile_atomicinc_group_to_flat_i64(
+; CHECK-NEXT: %1 = addrspacecast i64 addrspace(3)* %group.ptr to i64 addrspace(4)*
+; CHECK-NEXT: %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %1, i64 %y, i32 0, i32 0, i1 %volatile.var)
+define i64 @invalid_variable_volatile_atomicinc_group_to_flat_i64(i64 addrspace(3)* %group.ptr, i64 %y, i1 %volatile.var) #0 {
+ %cast = addrspacecast i64 addrspace(3)* %group.ptr to i64 addrspace(4)*
+ %ret = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %cast, i64 %y, i32 0, i32 0, i1 %volatile.var)
ret i64 %ret
}
declare i32 @llvm.objectsize.i32.p4i8(i8 addrspace(4)*, i1, i1) #1
declare i64 @llvm.objectsize.i64.p4i8(i8 addrspace(4)*, i1, i1) #1
-declare i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* nocapture, i32) #2
-declare i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* nocapture, i64) #2
-declare i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* nocapture, i32) #2
-declare i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* nocapture, i64) #2
+declare i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* nocapture, i32, i32, i32, i1) #2
+declare i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* nocapture, i64, i32, i32, i1) #2
+declare i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* nocapture, i32, i32, i32, i1) #2
+declare i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* nocapture, i64, i32, i32, i1) #2
attributes #0 = { nounwind }
attributes #1 = { nounwind readnone }