AMDGPU: Add all atomicrmw fields to atomic.inc/dec
Add scope, order, isVolatile
llvm-svn: 299122
diff --git a/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes.ll b/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes.ll
index 0ebff89..46bd194 100644
--- a/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes.ll
+++ b/llvm/test/CodeGen/AMDGPU/cgp-addressing-modes.ll
@@ -626,7 +626,7 @@
; 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
@@ -636,7 +636,7 @@
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:
@@ -652,7 +652,7 @@
; 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
@@ -662,7 +662,7 @@
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:
@@ -675,8 +675,8 @@
}
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 }
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.dec.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.dec.ll
index 4a620e0d..405534e 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.dec.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.dec.ll
@@ -1,21 +1,45 @@
; 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
}
@@ -25,7 +49,7 @@
; 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
}
@@ -36,7 +60,7 @@
; 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
}
@@ -45,7 +69,7 @@
; 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
}
@@ -53,7 +77,7 @@
; 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
}
@@ -63,7 +87,7 @@
; 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
}
@@ -71,7 +95,7 @@
; 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
}
@@ -80,7 +104,7 @@
; 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
}
@@ -93,7 +117,7 @@
%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
}
@@ -106,7 +130,7 @@
%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
}
@@ -114,7 +138,7 @@
; 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
}
@@ -124,7 +148,7 @@
; 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
}
@@ -132,7 +156,7 @@
; 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
}
@@ -141,7 +165,7 @@
; GCN: flat_atomic_dec v{{\[[0-9]+:[0-9]+\]}}, [[K]]{{$}}
define amdgpu_kernel void @flat_atomic_dec_noret_i32_offset(i32 addrspace(4)* %ptr) nounwind {
%gep = getelementptr i32, i32 addrspace(4)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
@@ -153,7 +177,7 @@
%gep.tid = getelementptr i32, i32 addrspace(4)* %ptr, i32 %id
%out.gep = getelementptr i32, i32 addrspace(4)* %out, i32 %id
%gep = getelementptr i32, i32 addrspace(4)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(4)* %out.gep
ret void
}
@@ -165,7 +189,7 @@
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i32, i32 addrspace(4)* %ptr, i32 %id
%gep = getelementptr i32, i32 addrspace(4)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.dec.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
@@ -174,7 +198,7 @@
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; 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
}
@@ -185,7 +209,7 @@
; GCN: flat_atomic_dec_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} glc{{$}}
define amdgpu_kernel void @flat_atomic_dec_ret_i64_offset(i64 addrspace(4)* %out, i64 addrspace(4)* %ptr) #0 {
%gep = getelementptr i64, i64 addrspace(4)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(4)* %out
ret void
}
@@ -195,7 +219,7 @@
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; 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
}
@@ -205,7 +229,7 @@
; GCN: flat_atomic_dec_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]$}}
define amdgpu_kernel void @flat_atomic_dec_noret_i64_offset(i64 addrspace(4)* %ptr) nounwind {
%gep = getelementptr i64, i64 addrspace(4)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
@@ -218,7 +242,7 @@
%gep.tid = getelementptr i64, i64 addrspace(4)* %ptr, i32 %id
%out.gep = getelementptr i64, i64 addrspace(4)* %out, i32 %id
%gep = getelementptr i64, i64 addrspace(4)* %gep.tid, i32 5
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(4)* %out.gep
ret void
}
@@ -231,7 +255,7 @@
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i64, i64 addrspace(4)* %ptr, i32 %id
%gep = getelementptr i64, i64 addrspace(4)* %gep.tid, i32 5
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
@@ -244,7 +268,7 @@
%tid.x = tail call i32 @llvm.amdgcn.workitem.id.x() #1
%idx.0 = add nsw i32 %tid.x, 2
%arrayidx0 = getelementptr inbounds [512 x i32], [512 x i32] addrspace(3)* @lds0, i32 0, i32 %idx.0
- %val0 = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %arrayidx0, i32 9)
+ %val0 = call i32 @llvm.amdgcn.atomic.dec.i32.p3i32(i32 addrspace(3)* %arrayidx0, i32 9, i32 0, i32 0, i1 false)
store i32 %idx.0, i32 addrspace(1)* %add_use
store i32 %val0, i32 addrspace(1)* %out
ret void
@@ -255,7 +279,7 @@
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; 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
}
@@ -266,7 +290,7 @@
; 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
}
@@ -276,7 +300,7 @@
; 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
}
@@ -286,7 +310,7 @@
; GCN: ds_dec_u64 v{{[0-9]+}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} offset:32{{$}}
define amdgpu_kernel void @lds_atomic_dec_noret_i64_offset(i64 addrspace(3)* %ptr) nounwind {
%gep = getelementptr i64, i64 addrspace(3)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p3i64(i64 addrspace(3)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
@@ -295,7 +319,7 @@
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; 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
}
@@ -306,7 +330,7 @@
; GCN: buffer_atomic_dec_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32 glc{{$}}
define amdgpu_kernel void @global_atomic_dec_ret_i64_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %ptr) #0 {
%gep = getelementptr i64, i64 addrspace(1)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.dec.i64.p1i64(i64 addrspace(1)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out
ret void
}
@@ -316,7 +340,7 @@
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; 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
}
@@ -326,7 +350,7 @@
; 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
}
@@ -340,7 +364,7 @@
%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
}
@@ -354,7 +378,7 @@
%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
}
@@ -367,7 +391,7 @@
%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
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.inc.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.inc.ll
index a4e3dc3..8334c0c 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.inc.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.inc.ll
@@ -1,13 +1,13 @@
; 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
@@ -15,7 +15,7 @@
; 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
}
@@ -25,7 +25,7 @@
; 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
}
@@ -36,7 +36,7 @@
; 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
}
@@ -45,7 +45,7 @@
; 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
}
@@ -53,7 +53,7 @@
; 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
}
@@ -63,7 +63,7 @@
; 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
}
@@ -71,7 +71,7 @@
; 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
}
@@ -80,7 +80,7 @@
; GCN: buffer_atomic_inc [[K]], off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:16{{$}}
define amdgpu_kernel void @global_atomic_inc_noret_i32_offset(i32 addrspace(1)* %ptr) nounwind {
%gep = getelementptr i32, i32 addrspace(1)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
@@ -93,7 +93,7 @@
%gep.tid = getelementptr i32, i32 addrspace(1)* %ptr, i32 %id
%out.gep = getelementptr i32, i32 addrspace(1)* %out, i32 %id
%gep = getelementptr i32, i32 addrspace(1)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(1)* %out.gep
ret void
}
@@ -106,7 +106,7 @@
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i32, i32 addrspace(1)* %ptr, i32 %id
%gep = getelementptr i32, i32 addrspace(1)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1i32(i32 addrspace(1)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
@@ -119,7 +119,7 @@
%tid.x = tail call i32 @llvm.amdgcn.workitem.id.x() #1
%idx.0 = add nsw i32 %tid.x, 2
%arrayidx0 = getelementptr inbounds [512 x i32], [512 x i32] addrspace(3)* @lds0, i32 0, i32 %idx.0
- %val0 = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %arrayidx0, i32 9)
+ %val0 = call i32 @llvm.amdgcn.atomic.inc.i32.p3i32(i32 addrspace(3)* %arrayidx0, i32 9, i32 0, i32 0, i1 false)
store i32 %idx.0, i32 addrspace(1)* %add_use
store i32 %val0, i32 addrspace(1)* %out
ret void
@@ -130,7 +130,7 @@
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; 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
}
@@ -141,7 +141,7 @@
; 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
}
@@ -151,7 +151,7 @@
; 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
}
@@ -161,7 +161,7 @@
; GCN: ds_inc_u64 v{{[0-9]+}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} offset:32{{$}}
define amdgpu_kernel void @lds_atomic_inc_noret_i64_offset(i64 addrspace(3)* %ptr) nounwind {
%gep = getelementptr i64, i64 addrspace(3)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %gep, i64 42, i32 0, i32 0, i1 false)
ret void
}
@@ -170,7 +170,7 @@
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; 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
}
@@ -181,7 +181,7 @@
; GCN: buffer_atomic_inc_x2 v{{\[}}[[KLO]]:[[KHI]]{{\]}}, off, s{{\[[0-9]+:[0-9]+\]}}, 0 offset:32 glc{{$}}
define amdgpu_kernel void @global_atomic_inc_ret_i64_offset(i64 addrspace(1)* %out, i64 addrspace(1)* %ptr) #0 {
%gep = getelementptr i64, i64 addrspace(1)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1i64(i64 addrspace(1)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(1)* %out
ret void
}
@@ -191,7 +191,7 @@
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; 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
}
@@ -201,7 +201,7 @@
; 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
}
@@ -215,7 +215,7 @@
%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
}
@@ -229,7 +229,7 @@
%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
}
@@ -237,7 +237,7 @@
; 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
}
@@ -247,7 +247,7 @@
; 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
}
@@ -255,7 +255,7 @@
; 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
}
@@ -264,7 +264,7 @@
; GCN: flat_atomic_inc v{{\[[0-9]+:[0-9]+\]}}, [[K]]{{$}}
define amdgpu_kernel void @flat_atomic_inc_noret_i32_offset(i32 addrspace(4)* %ptr) nounwind {
%gep = getelementptr i32, i32 addrspace(4)* %ptr, i32 4
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
@@ -276,7 +276,7 @@
%gep.tid = getelementptr i32, i32 addrspace(4)* %ptr, i32 %id
%out.gep = getelementptr i32, i32 addrspace(4)* %out, i32 %id
%gep = getelementptr i32, i32 addrspace(4)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
store i32 %result, i32 addrspace(4)* %out.gep
ret void
}
@@ -288,7 +288,7 @@
%id = call i32 @llvm.amdgcn.workitem.id.x()
%gep.tid = getelementptr i32, i32 addrspace(4)* %ptr, i32 %id
%gep = getelementptr i32, i32 addrspace(4)* %gep.tid, i32 5
- %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42)
+ %result = call i32 @llvm.amdgcn.atomic.inc.i32.p4i32(i32 addrspace(4)* %gep, i32 42, i32 0, i32 0, i1 false)
ret void
}
@@ -301,27 +301,18 @@
%tid.x = tail call i32 @llvm.amdgcn.workitem.id.x() #1
%idx.0 = add nsw i32 %tid.x, 2
%arrayidx0 = getelementptr inbounds [512 x i64], [512 x i64] addrspace(3)* @lds1, i32 0, i32 %idx.0
- %val0 = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %arrayidx0, i64 9)
+ %val0 = call i64 @llvm.amdgcn.atomic.inc.i64.p3i64(i64 addrspace(3)* %arrayidx0, i64 9, i32 0, i32 0, i1 false)
store i32 %idx.0, i32 addrspace(1)* %add_use
store i64 %val0, i64 addrspace(1)* %out
ret void
}
-attributes #0 = { nounwind }
-attributes #1 = { nounwind readnone }
-attributes #2 = { nounwind argmemonly }
-
-
-
-
-
-
; GCN-LABEL: {{^}}flat_atomic_inc_ret_i64:
; GCN-DAG: v_mov_b32_e32 v[[KLO:[0-9]+]], 42
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; 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
}
@@ -332,7 +323,7 @@
; GCN: flat_atomic_inc_x2 v{{\[[0-9]+:[0-9]+\]}}, v{{\[[0-9]+:[0-9]+\]}}, v{{\[}}[[KLO]]:[[KHI]]{{\]}} glc{{$}}
define amdgpu_kernel void @flat_atomic_inc_ret_i64_offset(i64 addrspace(4)* %out, i64 addrspace(4)* %ptr) #0 {
%gep = getelementptr i64, i64 addrspace(4)* %ptr, i32 4
- %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %gep, i64 42)
+ %result = call i64 @llvm.amdgcn.atomic.inc.i64.p4i64(i64 addrspace(4)* %gep, i64 42, i32 0, i32 0, i1 false)
store i64 %result, i64 addrspace(4)* %out
ret void
}
@@ -342,7 +333,7 @@
; GCN-DAG: v_mov_b32_e32 v[[KHI:[0-9]+]], 0{{$}}
; 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
}
@@ -352,7 +343,7 @@
; 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
}
@@ -365,7 +356,7 @@
%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
}
@@ -378,6 +369,10 @@
%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 }