https://github.com/arsenm created
https://github.com/llvm/llvm-project/pull/129834
The previous checks missed the new metadata at the end of the line.
Regenerate to avoid future spurious diffs.
>From 677c5bdaecc01020e00abd1814701680091fa367 Mon Sep 17 00:00:00 2001
From: Matt Arsenault <matthew.arsena...@amd.com>
Date: Wed, 5 Mar 2025 12:58:34 +0700
Subject: [PATCH] clang: Regenerate test checks
The previous checks missed the new metadata at the end of the line.
Regenerate to avoid future spurious diffs.
---
.../builtin-amdgcn-atomic-inc-dec.cpp | 80 +++++++++----------
1 file changed, 40 insertions(+), 40 deletions(-)
diff --git a/clang/test/CodeGenCXX/builtin-amdgcn-atomic-inc-dec.cpp
b/clang/test/CodeGenCXX/builtin-amdgcn-atomic-inc-dec.cpp
index 4f3b5a4ce7760..5920ceda4a811 100644
--- a/clang/test/CodeGenCXX/builtin-amdgcn-atomic-inc-dec.cpp
+++ b/clang/test/CodeGenCXX/builtin-amdgcn-atomic-inc-dec.cpp
@@ -13,12 +13,12 @@
// CHECK-NEXT: [[TMP0:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP1:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr [[TMP1]], align 4
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw uinc_wrap ptr [[TMP0]], i32
[[TMP2]] syncscope("workgroup") seq_cst, align 4
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw uinc_wrap ptr [[TMP0]], i32
[[TMP2]] syncscope("workgroup") seq_cst, align 4,
!amdgpu.no.fine.grained.memory [[META4:![0-9]+]]
// CHECK-NEXT: store i32 [[TMP3]], ptr [[RES_ASCAST]], align 4
// CHECK-NEXT: [[TMP4:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP5:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
-// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr [[TMP4]], i32
[[TMP6]] syncscope("workgroup") seq_cst, align 4
+// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr [[TMP4]], i32
[[TMP6]] syncscope("workgroup") seq_cst, align 4,
!amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i32 [[TMP7]], ptr [[RES_ASCAST]], align 4
// CHECK-NEXT: ret void
//
@@ -39,12 +39,12 @@ __attribute__((device)) void
test_non_volatile_parameter32(__UINT32_TYPE__ *ptr)
// CHECK-NEXT: [[TMP0:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP1:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP2:%.*]] = load i64, ptr [[TMP1]], align 8
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw uinc_wrap ptr [[TMP0]], i64
[[TMP2]] syncscope("workgroup") seq_cst, align 8
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw uinc_wrap ptr [[TMP0]], i64
[[TMP2]] syncscope("workgroup") seq_cst, align 8,
!amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i64 [[TMP3]], ptr [[RES_ASCAST]], align 8
// CHECK-NEXT: [[TMP4:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP5:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP6:%.*]] = load i64, ptr [[TMP5]], align 8
-// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr [[TMP4]], i64
[[TMP6]] syncscope("workgroup") seq_cst, align 8
+// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr [[TMP4]], i64
[[TMP6]] syncscope("workgroup") seq_cst, align 8,
!amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i64 [[TMP7]], ptr [[RES_ASCAST]], align 8
// CHECK-NEXT: ret void
//
@@ -65,12 +65,12 @@ __attribute__((device)) void
test_non_volatile_parameter64(__UINT64_TYPE__ *ptr)
// CHECK-NEXT: [[TMP0:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP1:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP2:%.*]] = load volatile i32, ptr [[TMP1]], align 4
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw volatile uinc_wrap ptr [[TMP0]],
i32 [[TMP2]] syncscope("workgroup") seq_cst, align 4
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw volatile uinc_wrap ptr [[TMP0]],
i32 [[TMP2]] syncscope("workgroup") seq_cst, align 4,
!amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i32 [[TMP3]], ptr [[RES_ASCAST]], align 4
// CHECK-NEXT: [[TMP4:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP5:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP6:%.*]] = load volatile i32, ptr [[TMP5]], align 4
-// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw volatile udec_wrap ptr [[TMP4]],
i32 [[TMP6]] syncscope("workgroup") seq_cst, align 4
+// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw volatile udec_wrap ptr [[TMP4]],
i32 [[TMP6]] syncscope("workgroup") seq_cst, align 4,
!amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i32 [[TMP7]], ptr [[RES_ASCAST]], align 4
// CHECK-NEXT: ret void
//
@@ -91,12 +91,12 @@ __attribute__((device)) void
test_volatile_parameter32(volatile __UINT32_TYPE__
// CHECK-NEXT: [[TMP0:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP1:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP2:%.*]] = load volatile i64, ptr [[TMP1]], align 8
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw volatile uinc_wrap ptr [[TMP0]],
i64 [[TMP2]] syncscope("workgroup") seq_cst, align 8
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw volatile uinc_wrap ptr [[TMP0]],
i64 [[TMP2]] syncscope("workgroup") seq_cst, align 8,
!amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i64 [[TMP3]], ptr [[RES_ASCAST]], align 8
// CHECK-NEXT: [[TMP4:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP5:%.*]] = load ptr, ptr [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT: [[TMP6:%.*]] = load volatile i64, ptr [[TMP5]], align 8
-// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw volatile udec_wrap ptr [[TMP4]],
i64 [[TMP6]] syncscope("workgroup") seq_cst, align 8
+// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw volatile udec_wrap ptr [[TMP4]],
i64 [[TMP6]] syncscope("workgroup") seq_cst, align 8,
!amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i64 [[TMP7]], ptr [[RES_ASCAST]], align 8
// CHECK-NEXT: ret void
//
@@ -110,10 +110,10 @@ __attribute__((device)) void
test_volatile_parameter64(volatile __UINT64_TYPE__
// CHECK-LABEL: @_Z13test_shared32v(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ13test_shared32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ13test_shared32vE3val to ptr), i32 [[TMP0]]
syncscope("workgroup") seq_cst, align 4
+// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ13test_shared32vE3val to ptr), i32 [[TMP0]]
syncscope("workgroup") seq_cst, align 4, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i32 [[TMP1]], ptr addrspacecast (ptr addrspace(3)
@_ZZ13test_shared32vE3val to ptr), align 4
// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ13test_shared32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ13test_shared32vE3val to ptr), i32 [[TMP2]]
syncscope("workgroup") seq_cst, align 4
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ13test_shared32vE3val to ptr), i32 [[TMP2]]
syncscope("workgroup") seq_cst, align 4, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i32 [[TMP3]], ptr addrspacecast (ptr addrspace(3)
@_ZZ13test_shared32vE3val to ptr), align 4
// CHECK-NEXT: ret void
//
@@ -128,10 +128,10 @@ __attribute__((device)) void test_shared32() {
// CHECK-LABEL: @_Z13test_shared64v(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ13test_shared64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ13test_shared64vE3val to ptr), i64 [[TMP0]]
syncscope("workgroup") seq_cst, align 8
+// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ13test_shared64vE3val to ptr), i64 [[TMP0]]
syncscope("workgroup") seq_cst, align 8, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i64 [[TMP1]], ptr addrspacecast (ptr addrspace(3)
@_ZZ13test_shared64vE3val to ptr), align 8
// CHECK-NEXT: [[TMP2:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ13test_shared64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ13test_shared64vE3val to ptr), i64 [[TMP2]]
syncscope("workgroup") seq_cst, align 8
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ13test_shared64vE3val to ptr), i64 [[TMP2]]
syncscope("workgroup") seq_cst, align 8, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i64 [[TMP3]], ptr addrspacecast (ptr addrspace(3)
@_ZZ13test_shared64vE3val to ptr), align 8
// CHECK-NEXT: ret void
//
@@ -147,10 +147,10 @@ __attribute__((device)) __UINT32_TYPE__ global_val32;
// CHECK-LABEL: @_Z13test_global32v(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspacecast (ptr addrspace(1)
@global_val32 to ptr), align 4
-// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(1) @global_val32 to ptr), i32 [[TMP0]] syncscope("workgroup")
seq_cst, align 4
+// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(1) @global_val32 to ptr), i32 [[TMP0]] syncscope("workgroup")
seq_cst, align 4, !amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i32 [[TMP1]], ptr addrspacecast (ptr addrspace(1)
@global_val32 to ptr), align 4
// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspacecast (ptr addrspace(1)
@global_val32 to ptr), align 4
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(1) @global_val32 to ptr), i32 [[TMP2]] syncscope("workgroup")
seq_cst, align 4
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(1) @global_val32 to ptr), i32 [[TMP2]] syncscope("workgroup")
seq_cst, align 4, !amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i32 [[TMP3]], ptr addrspacecast (ptr addrspace(1)
@global_val32 to ptr), align 4
// CHECK-NEXT: ret void
//
@@ -164,10 +164,10 @@ __attribute__((device)) __UINT64_TYPE__ global_val64;
// CHECK-LABEL: @_Z13test_global64v(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr addrspacecast (ptr addrspace(1)
@global_val64 to ptr), align 8
-// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(1) @global_val64 to ptr), i64 [[TMP0]] syncscope("workgroup")
seq_cst, align 8
+// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(1) @global_val64 to ptr), i64 [[TMP0]] syncscope("workgroup")
seq_cst, align 8, !amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i64 [[TMP1]], ptr addrspacecast (ptr addrspace(1)
@global_val64 to ptr), align 8
// CHECK-NEXT: [[TMP2:%.*]] = load i64, ptr addrspacecast (ptr addrspace(1)
@global_val64 to ptr), align 8
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(1) @global_val64 to ptr), i64 [[TMP2]] syncscope("workgroup")
seq_cst, align 8
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(1) @global_val64 to ptr), i64 [[TMP2]] syncscope("workgroup")
seq_cst, align 8, !amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i64 [[TMP3]], ptr addrspacecast (ptr addrspace(1)
@global_val64 to ptr), align 8
// CHECK-NEXT: ret void
//
@@ -183,10 +183,10 @@ __attribute__((constant)) __UINT32_TYPE__ cval32;
// CHECK-NEXT: [[LOCAL_VAL:%.*]] = alloca i32, align 4, addrspace(5)
// CHECK-NEXT: [[LOCAL_VAL_ASCAST:%.*]] = addrspacecast ptr addrspace(5)
[[LOCAL_VAL]] to ptr
// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspacecast (ptr addrspace(4)
@cval32 to ptr), align 4
-// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(4) @cval32 to ptr), i32 [[TMP0]] syncscope("workgroup") seq_cst,
align 4
+// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(4) @cval32 to ptr), i32 [[TMP0]] syncscope("workgroup") seq_cst,
align 4, !amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i32 [[TMP1]], ptr [[LOCAL_VAL_ASCAST]], align 4
// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspacecast (ptr addrspace(4)
@cval32 to ptr), align 4
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(4) @cval32 to ptr), i32 [[TMP2]] syncscope("workgroup") seq_cst,
align 4
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(4) @cval32 to ptr), i32 [[TMP2]] syncscope("workgroup") seq_cst,
align 4, !amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i32 [[TMP3]], ptr [[LOCAL_VAL_ASCAST]], align 4
// CHECK-NEXT: ret void
//
@@ -204,10 +204,10 @@ __attribute__((constant)) __UINT64_TYPE__ cval64;
// CHECK-NEXT: [[LOCAL_VAL:%.*]] = alloca i64, align 8, addrspace(5)
// CHECK-NEXT: [[LOCAL_VAL_ASCAST:%.*]] = addrspacecast ptr addrspace(5)
[[LOCAL_VAL]] to ptr
// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr addrspacecast (ptr addrspace(4)
@cval64 to ptr), align 8
-// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(4) @cval64 to ptr), i64 [[TMP0]] syncscope("workgroup") seq_cst,
align 8
+// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(4) @cval64 to ptr), i64 [[TMP0]] syncscope("workgroup") seq_cst,
align 8, !amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i64 [[TMP1]], ptr [[LOCAL_VAL_ASCAST]], align 8
// CHECK-NEXT: [[TMP2:%.*]] = load i64, ptr addrspacecast (ptr addrspace(4)
@cval64 to ptr), align 8
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(4) @cval64 to ptr), i64 [[TMP2]] syncscope("workgroup") seq_cst,
align 8
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(4) @cval64 to ptr), i64 [[TMP2]] syncscope("workgroup") seq_cst,
align 8, !amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i64 [[TMP3]], ptr [[LOCAL_VAL_ASCAST]], align 8
// CHECK-NEXT: ret void
//
@@ -222,22 +222,22 @@ __attribute__((device)) void test_constant64() {
// CHECK-LABEL: @_Z12test_order32v(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP0]]
syncscope("workgroup") monotonic, align 4
+// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP0]]
syncscope("workgroup") monotonic, align 4, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i32 [[TMP1]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP2]]
syncscope("workgroup") acquire, align 4
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP2]]
syncscope("workgroup") acquire, align 4, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i32 [[TMP3]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP5:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP4]]
syncscope("workgroup") acquire, align 4
+// CHECK-NEXT: [[TMP5:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP4]]
syncscope("workgroup") acquire, align 4, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i32 [[TMP5]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
// CHECK-NEXT: [[TMP6:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP6]]
syncscope("workgroup") release, align 4
+// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP6]]
syncscope("workgroup") release, align 4, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i32 [[TMP7]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
// CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP9:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP8]]
syncscope("workgroup") acq_rel, align 4
+// CHECK-NEXT: [[TMP9:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP8]]
syncscope("workgroup") acq_rel, align 4, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i32 [[TMP9]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
// CHECK-NEXT: [[TMP10:%.*]] = load i32, ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP11:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP10]]
syncscope("workgroup") seq_cst, align 4
+// CHECK-NEXT: [[TMP11:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order32vE3val to ptr), i32 [[TMP10]]
syncscope("workgroup") seq_cst, align 4, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i32 [[TMP11]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order32vE3val to ptr), align 4
// CHECK-NEXT: ret void
//
@@ -260,22 +260,22 @@ __attribute__((device)) void test_order32() {
// CHECK-LABEL: @_Z12test_order64v(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP0]]
syncscope("workgroup") monotonic, align 8
+// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP0]]
syncscope("workgroup") monotonic, align 8, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i64 [[TMP1]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
// CHECK-NEXT: [[TMP2:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP2]]
syncscope("workgroup") acquire, align 8
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP2]]
syncscope("workgroup") acquire, align 8, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i64 [[TMP3]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
// CHECK-NEXT: [[TMP4:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP5:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP4]]
syncscope("workgroup") acquire, align 8
+// CHECK-NEXT: [[TMP5:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP4]]
syncscope("workgroup") acquire, align 8, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i64 [[TMP5]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
// CHECK-NEXT: [[TMP6:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP6]]
syncscope("workgroup") release, align 8
+// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP6]]
syncscope("workgroup") release, align 8, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i64 [[TMP7]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
// CHECK-NEXT: [[TMP8:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP9:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP8]]
syncscope("workgroup") acq_rel, align 8
+// CHECK-NEXT: [[TMP9:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP8]]
syncscope("workgroup") acq_rel, align 8, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i64 [[TMP9]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
// CHECK-NEXT: [[TMP10:%.*]] = load i64, ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP11:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP10]]
syncscope("workgroup") seq_cst, align 8
+// CHECK-NEXT: [[TMP11:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_order64vE3val to ptr), i64 [[TMP10]]
syncscope("workgroup") seq_cst, align 8, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i64 [[TMP11]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_order64vE3val to ptr), align 8
// CHECK-NEXT: ret void
//
@@ -298,16 +298,16 @@ __attribute__((device)) void test_order64() {
// CHECK-LABEL: @_Z12test_scope32v(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope32vE3val to ptr), i32 [[TMP0]] seq_cst, align 4
+// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope32vE3val to ptr), i32 [[TMP0]] seq_cst, align 4,
!amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i32 [[TMP1]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope32vE3val to ptr), align 4
// CHECK-NEXT: [[TMP2:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope32vE3val to ptr), i32 [[TMP2]]
syncscope("workgroup") seq_cst, align 4
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope32vE3val to ptr), i32 [[TMP2]]
syncscope("workgroup") seq_cst, align 4, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i32 [[TMP3]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope32vE3val to ptr), align 4
// CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP5:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope32vE3val to ptr), i32 [[TMP4]] syncscope("agent")
seq_cst, align 4
+// CHECK-NEXT: [[TMP5:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope32vE3val to ptr), i32 [[TMP4]] syncscope("agent")
seq_cst, align 4, !amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i32 [[TMP5]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope32vE3val to ptr), align 4
// CHECK-NEXT: [[TMP6:%.*]] = load i32, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope32vE3val to ptr), align 4
-// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope32vE3val to ptr), i32 [[TMP6]]
syncscope("wavefront") seq_cst, align 4
+// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope32vE3val to ptr), i32 [[TMP6]]
syncscope("wavefront") seq_cst, align 4, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i32 [[TMP7]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope32vE3val to ptr), align 4
// CHECK-NEXT: ret void
//
@@ -326,16 +326,16 @@ __attribute__((device)) void test_scope32() {
// CHECK-LABEL: @_Z12test_scope64v(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope64vE3val to ptr), i64 [[TMP0]] seq_cst, align 8
+// CHECK-NEXT: [[TMP1:%.*]] = atomicrmw uinc_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope64vE3val to ptr), i64 [[TMP0]] seq_cst, align 8,
!amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i64 [[TMP1]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope64vE3val to ptr), align 8
// CHECK-NEXT: [[TMP2:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope64vE3val to ptr), i64 [[TMP2]]
syncscope("workgroup") seq_cst, align 8
+// CHECK-NEXT: [[TMP3:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope64vE3val to ptr), i64 [[TMP2]]
syncscope("workgroup") seq_cst, align 8, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i64 [[TMP3]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope64vE3val to ptr), align 8
// CHECK-NEXT: [[TMP4:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP5:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope64vE3val to ptr), i64 [[TMP4]] syncscope("agent")
seq_cst, align 8
+// CHECK-NEXT: [[TMP5:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope64vE3val to ptr), i64 [[TMP4]] syncscope("agent")
seq_cst, align 8, !amdgpu.no.fine.grained.memory [[META4]]
// CHECK-NEXT: store i64 [[TMP5]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope64vE3val to ptr), align 8
// CHECK-NEXT: [[TMP6:%.*]] = load i64, ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope64vE3val to ptr), align 8
-// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope64vE3val to ptr), i64 [[TMP6]]
syncscope("wavefront") seq_cst, align 8
+// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw udec_wrap ptr addrspacecast (ptr
addrspace(3) @_ZZ12test_scope64vE3val to ptr), i64 [[TMP6]]
syncscope("wavefront") seq_cst, align 8, !amdgpu.no.fine.grained.memory
[[META4]]
// CHECK-NEXT: store i64 [[TMP7]], ptr addrspacecast (ptr addrspace(3)
@_ZZ12test_scope64vE3val to ptr), align 8
// CHECK-NEXT: ret void
//
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
