[test][Transforms] Remove unsafe-fp-math uses part 3 (NFC) #164787

paperchalice · 2025-10-23T09:59:27Z

Post cleanup for #164534.

llvmbot · 2025-10-23T12:57:40Z

@llvm/pr-subscribers-llvm-transforms
@llvm/pr-subscribers-pgo

@llvm/pr-subscribers-backend-amdgpu

Author: None (paperchalice)

Changes

Post cleanup for #164534.

Patch is 144.07 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/164787.diff

59 Files Affected:

(modified) llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll (+7-11)
(modified) llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll (+20-25)
(modified) llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll (+38-42)
(modified) llvm/test/Transforms/PhaseOrdering/ARM/arm_add_q7.ll (+2-2)
(modified) llvm/test/Transforms/PhaseOrdering/ARM/arm_fill_q7.ll (+10-15)
(modified) llvm/test/Transforms/PhaseOrdering/ARM/arm_mean_q7.ll (+1-1)
(modified) llvm/test/Transforms/PhaseOrdering/ARM/arm_mult_q15.ll (+2-2)
(modified) llvm/test/Transforms/PhaseOrdering/X86/fmaddsub.ll (+16-18)
(modified) llvm/test/Transforms/PhaseOrdering/X86/vdiv-nounroll.ll (+5-9)
(modified) llvm/test/Transforms/PhaseOrdering/X86/vdiv.ll (+1-1)
(modified) llvm/test/Transforms/PhaseOrdering/X86/vector-reductions-expanded.ll (+8-10)
(modified) llvm/test/Transforms/SLPVectorizer/NVPTX/vectorizable-intrinsic.ll (+6-9)
(modified) llvm/test/Transforms/SLPVectorizer/X86/crash_bullet3.ll (+1-3)
(modified) llvm/test/Transforms/SLPVectorizer/X86/crash_flop7.ll (+1-3)
(modified) llvm/test/Transforms/SLPVectorizer/X86/debug_info.ll (+2-5)
(modified) llvm/test/Transforms/SLPVectorizer/X86/pr16899.ll (+1-3)
(modified) llvm/test/Transforms/SLPVectorizer/X86/vector_gep.ll (+1-4)
(modified) llvm/test/Transforms/SLPVectorizer/X86/vectorize-pair-path.ll (+1-3)
(modified) llvm/test/Transforms/SLPVectorizer/consecutive-access.ll (+8-10)
(modified) llvm/test/Transforms/SLPVectorizer/insert-element-build-vector-inseltpoison.ll (+13-15)
(modified) llvm/test/Transforms/SLPVectorizer/insert-element-build-vector.ll (+14-16)
(modified) llvm/test/Transforms/SROA/mem-par-metadata-sroa.ll (+2-6)
(modified) llvm/test/Transforms/SafeStack/ARM/debug.ll (+3-4)
(modified) llvm/test/Transforms/SafeStack/X86/debug-loc.ll (+2-3)
(modified) llvm/test/Transforms/SafeStack/X86/debug-loc2.ll (+2-3)
(modified) llvm/test/Transforms/SampleProfile/Inputs/profile-symbol-list.ll (+4-4)
(modified) llvm/test/Transforms/SampleProfile/branch.ll (+7-11)
(modified) llvm/test/Transforms/SampleProfile/csspgo-import-list.ll (+1-1)
(modified) llvm/test/Transforms/SampleProfile/csspgo-inline-debug.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/csspgo-inline.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/csspgo-summary.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/csspgo-use-preinliner.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/entry_counts_cold.ll (+2-3)
(modified) llvm/test/Transforms/SampleProfile/entry_counts_missing_dbginfo.ll (+2-3)
(modified) llvm/test/Transforms/SampleProfile/fsafdo_test.ll (+3-5)
(modified) llvm/test/Transforms/SampleProfile/gcc-simple.ll (+5-8)
(modified) llvm/test/Transforms/SampleProfile/inline-act.ll (+2-4)
(modified) llvm/test/Transforms/SampleProfile/misexpect.ll (+8-14)
(modified) llvm/test/Transforms/SampleProfile/norepeated-icp-2.ll (+4-7)
(modified) llvm/test/Transforms/SampleProfile/norepeated-icp-3.ll (+3-4)
(modified) llvm/test/Transforms/SampleProfile/norepeated-icp-4.ll (+1-1)
(modified) llvm/test/Transforms/SampleProfile/norepeated-icp.ll (+3-5)
(modified) llvm/test/Transforms/SampleProfile/offset.ll (+1-1)
(modified) llvm/test/Transforms/SampleProfile/profile-context-order.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/profile-context-tracker-debug.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/profile-context-tracker.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/profile-topdown-order.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/propagate.ll (+4-7)
(modified) llvm/test/Transforms/SampleProfile/pseudo-probe-discriminator.ll (+2-7)
(modified) llvm/test/Transforms/SampleProfile/pseudo-probe-icp-factor.ll (+1-1)
(modified) llvm/test/Transforms/SampleProfile/pseudo-probe-no-debug-info.ll (+2-6)
(modified) llvm/test/Transforms/SampleProfile/remarks.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/uniqname.ll (+6-7)
(modified) llvm/test/Transforms/Scalarizer/dbginfo.ll (+2-5)
(modified) llvm/test/Transforms/SimplifyCFG/Hexagon/switch-to-lookup-table.ll (+1-3)
(modified) llvm/test/Transforms/SimplifyCFG/X86/merge-cleanuppads.ll (+8-12)
(modified) llvm/test/Transforms/SimplifyCFG/pr50060-constantfold-loopid.ll (+11-15)
(modified) llvm/test/Transforms/StraightLineStrengthReduce/AMDGPU/pr23975.ll (+1-3)
(modified) llvm/test/Transforms/StructurizeCFG/nested-loop-order.ll (-4)

diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll
index 27e8fd0ece55f..3b617502ed955 100644
--- a/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll
+++ b/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll
@@ -6,7 +6,7 @@ target triple = "aarch64"
 
 ; This function (a 16x reduction of a[i] * b[i]) should be vectorized successfully.
 
-define dso_local nofpclass(nan inf) float @vmlaq(ptr noundef %0, ptr noundef %1) #0 {
+define dso_local nofpclass(nan inf) float @vmlaq(ptr noundef %0, ptr noundef %1) {
 ; CHECK-LABEL: define dso_local nofpclass(nan inf) float @vmlaq
 ; CHECK-SAME: (ptr noundef readonly captures(none) [[TMP0:%.*]], ptr noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
 ; CHECK-NEXT:    [[TMP3:%.*]] = load <16 x float>, ptr [[TMP0]], align 4, !tbaa [[TBAA4:![0-9]+]]
@@ -21,9 +21,9 @@ define dso_local nofpclass(nan inf) float @vmlaq(ptr noundef %0, ptr noundef %1)
   %6 = alloca i32, align 4
   store ptr %0, ptr %3, align 8, !tbaa !4
   store ptr %1, ptr %4, align 8, !tbaa !4
-  call void @llvm.lifetime.start.p0(ptr %5) #2
+  call void @llvm.lifetime.start.p0(ptr %5)
   store float 0.000000e+00, ptr %5, align 4, !tbaa !9
-  call void @llvm.lifetime.start.p0(ptr %6) #2
+  call void @llvm.lifetime.start.p0(ptr %6)
   store i32 0, ptr %6, align 4, !tbaa !11
   br label %7
 
@@ -33,7 +33,7 @@ define dso_local nofpclass(nan inf) float @vmlaq(ptr noundef %0, ptr noundef %1)
   br i1 %9, label %11, label %10
 
 10:                                               ; preds = %7
-  call void @llvm.lifetime.end.p0(ptr %6) #2
+  call void @llvm.lifetime.end.p0(ptr %6)
   br label %28
 
 11:                                               ; preds = %7
@@ -61,16 +61,12 @@ define dso_local nofpclass(nan inf) float @vmlaq(ptr noundef %0, ptr noundef %1)
 
 28:                                               ; preds = %10
   %29 = load float, ptr %5, align 4, !tbaa !9
-  call void @llvm.lifetime.end.p0(ptr %5) #2
+  call void @llvm.lifetime.end.p0(ptr %5)
   ret float %29
 }
 
-declare void @llvm.lifetime.start.p0(ptr captures(none)) #1
-declare void @llvm.lifetime.end.p0(ptr captures(none)) #1
-
-attributes #0 = { nounwind uwtable "frame-pointer"="non-leaf" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="generic" "target-features"="+fp-armv8,+neon,+v8a,-fmv" "unsafe-fp-math"="true" }
-attributes #1 = { nocallback nofree nosync nounwind willreturn memory(argmem: readwrite) }
-attributes #2 = { nounwind }
+declare void @llvm.lifetime.start.p0(ptr captures(none))
+declare void @llvm.lifetime.end.p0(ptr captures(none))
 
 !llvm.module.flags = !{!0, !1, !2}
 !llvm.ident = !{!3}
diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll
index 68bfbc1344f23..eefde9ddfdc1d 100644
--- a/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll
+++ b/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll
@@ -6,7 +6,7 @@ target triple = "aarch64"
 
 ; This function (a more complex reduction of (a[i] - b[i]) * itself) should be vectorized successfully.
 
-define dso_local noundef nofpclass(nan inf) float @_Z4testPKfS0_ii(ptr noundef %0, ptr noundef %1, i32 noundef %2, i32 noundef %3) #0 {
+define dso_local noundef nofpclass(nan inf) float @_Z4testPKfS0_ii(ptr noundef %0, ptr noundef %1, i32 noundef %2, i32 noundef %3) {
 ; CHECK-LABEL: define dso_local noundef nofpclass(nan inf) float @_Z4testPKfS0_ii
 ; CHECK-SAME: (ptr noundef readonly captures(none) [[TMP0:%.*]], ptr noundef readonly captures(none) [[TMP1:%.*]], i32 noundef [[TMP2:%.*]], i32 noundef [[TMP3:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
 ; CHECK-NEXT:  .preheader.i:
@@ -125,7 +125,7 @@ define dso_local noundef nofpclass(nan inf) float @_Z4testPKfS0_ii(ptr noundef %
   ret float %13
 }
 
-define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr noundef %0, ptr noundef %1, i32 noundef %2, i32 noundef %3) #1 {
+define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr noundef %0, ptr noundef %1, i32 noundef %2, i32 noundef %3) {
   %5 = alloca ptr, align 8
   %6 = alloca ptr, align 8
   %7 = alloca i32, align 4
@@ -143,15 +143,15 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
   store ptr %1, ptr %6, align 8, !tbaa !4
   store i32 %2, ptr %7, align 4, !tbaa !9
   store i32 %3, ptr %8, align 4, !tbaa !9
-  call void @llvm.lifetime.start.p0(ptr %9) #3
+  call void @llvm.lifetime.start.p0(ptr %9)
   store i32 3, ptr %9, align 4, !tbaa !9
-  call void @llvm.lifetime.start.p0(ptr %10) #3
+  call void @llvm.lifetime.start.p0(ptr %10)
   store i32 3, ptr %10, align 4, !tbaa !9
-  call void @llvm.lifetime.start.p0(ptr %11) #3
+  call void @llvm.lifetime.start.p0(ptr %11)
   store i32 7, ptr %11, align 4, !tbaa !9
-  call void @llvm.lifetime.start.p0(ptr %12) #3
+  call void @llvm.lifetime.start.p0(ptr %12)
   store float 0.000000e+00, ptr %12, align 4, !tbaa !11
-  call void @llvm.lifetime.start.p0(ptr %13) #3
+  call void @llvm.lifetime.start.p0(ptr %13)
   store i32 0, ptr %13, align 4, !tbaa !9
   br label %18
 
@@ -162,13 +162,13 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
 
 21:                                               ; preds = %18
   store i32 2, ptr %14, align 4
-  call void @llvm.lifetime.end.p0(ptr %13) #3
+  call void @llvm.lifetime.end.p0(ptr %13)
   br label %62
 
 22:                                               ; preds = %18
-  call void @llvm.lifetime.start.p0(ptr %15) #3
+  call void @llvm.lifetime.start.p0(ptr %15)
   store float 0.000000e+00, ptr %15, align 4, !tbaa !11
-  call void @llvm.lifetime.start.p0(ptr %16) #3
+  call void @llvm.lifetime.start.p0(ptr %16)
   store i32 0, ptr %16, align 4, !tbaa !9
   br label %23
 
@@ -179,11 +179,11 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
 
 26:                                               ; preds = %23
   store i32 5, ptr %14, align 4
-  call void @llvm.lifetime.end.p0(ptr %16) #3
+  call void @llvm.lifetime.end.p0(ptr %16)
   br label %47
 
 27:                                               ; preds = %23
-  call void @llvm.lifetime.start.p0(ptr %17) #3
+  call void @llvm.lifetime.start.p0(ptr %17)
   %28 = load ptr, ptr %5, align 8, !tbaa !4
   %29 = load i32, ptr %16, align 4, !tbaa !9
   %30 = sext i32 %29 to i64
@@ -202,7 +202,7 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
   %42 = load float, ptr %15, align 4, !tbaa !11
   %43 = fadd fast float %42, %41
   store float %43, ptr %15, align 4, !tbaa !11
-  call void @llvm.lifetime.end.p0(ptr %17) #3
+  call void @llvm.lifetime.end.p0(ptr %17)
   br label %44
 
 44:                                               ; preds = %27
@@ -226,7 +226,7 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
   %57 = load float, ptr %12, align 4, !tbaa !11
   %58 = fadd fast float %57, %56
   store float %58, ptr %12, align 4, !tbaa !11
-  call void @llvm.lifetime.end.p0(ptr %15) #3
+  call void @llvm.lifetime.end.p0(ptr %15)
   br label %59
 
 59:                                               ; preds = %47
@@ -238,20 +238,15 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
 62:                                               ; preds = %21
   %63 = load float, ptr %12, align 4, !tbaa !11
   store i32 1, ptr %14, align 4
-  call void @llvm.lifetime.end.p0(ptr %12) #3
-  call void @llvm.lifetime.end.p0(ptr %11) #3
-  call void @llvm.lifetime.end.p0(ptr %10) #3
-  call void @llvm.lifetime.end.p0(ptr %9) #3
+  call void @llvm.lifetime.end.p0(ptr %12)
+  call void @llvm.lifetime.end.p0(ptr %11)
+  call void @llvm.lifetime.end.p0(ptr %10)
+  call void @llvm.lifetime.end.p0(ptr %9)
   ret float %63
 }
 
-declare void @llvm.lifetime.start.p0(ptr captures(none)) #2
-declare void @llvm.lifetime.end.p0(ptr captures(none)) #2
-
-attributes #0 = { mustprogress uwtable "frame-pointer"="non-leaf" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="generic" "target-features"="+fp-armv8,+neon,+v8a,-fmv" "unsafe-fp-math"="true" }
-attributes #1 = { inlinehint mustprogress nounwind uwtable "frame-pointer"="non-leaf" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="generic" "target-features"="+fp-armv8,+neon,+v8a,-fmv" "unsafe-fp-math"="true" }
-attributes #2 = { nocallback nofree nosync nounwind willreturn memory(argmem: readwrite) }
-attributes #3 = { nounwind }
+declare void @llvm.lifetime.start.p0(ptr captures(none))
+declare void @llvm.lifetime.end.p0(ptr captures(none))
 
 !llvm.module.flags = !{!0, !1, !2}
 !llvm.ident = !{!3}
diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll
index 92e625deb11b1..82ecc3a0ed172 100644
--- a/llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll
+++ b/llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll
@@ -10,7 +10,7 @@ target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128-Fn32"
 target triple = "aarch64"
 
 ; Function Attrs: nounwind uwtable
-define i32 @slpordering(ptr noundef %p1, i32 noundef %ip1, ptr noundef %p2, i32 noundef %ip2) #0 {
+define i32 @slpordering(ptr noundef %p1, i32 noundef %ip1, ptr noundef %p2, i32 noundef %ip2) {
 ; CHECK-LABEL: define range(i32 0, 65536) i32 @slpordering(
 ; CHECK-SAME: ptr noundef readonly captures(none) [[P1:%.*]], i32 noundef [[IP1:%.*]], ptr noundef readonly captures(none) [[P2:%.*]], i32 noundef [[IP2:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
 ; CHECK-NEXT:  [[ENTRY:.*:]]
@@ -136,14 +136,14 @@ entry:
   store i32 %ip1, ptr %ip1.addr, align 4, !tbaa !8
   store ptr %p2, ptr %p2.addr, align 8, !tbaa !4
   store i32 %ip2, ptr %ip2.addr, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %emp) #2
-  call void @llvm.lifetime.start.p0(ptr %r0) #2
-  call void @llvm.lifetime.start.p0(ptr %r1) #2
-  call void @llvm.lifetime.start.p0(ptr %r2) #2
-  call void @llvm.lifetime.start.p0(ptr %r3) #2
-  call void @llvm.lifetime.start.p0(ptr %sum) #2
+  call void @llvm.lifetime.start.p0(ptr %emp)
+  call void @llvm.lifetime.start.p0(ptr %r0)
+  call void @llvm.lifetime.start.p0(ptr %r1)
+  call void @llvm.lifetime.start.p0(ptr %r2)
+  call void @llvm.lifetime.start.p0(ptr %r3)
+  call void @llvm.lifetime.start.p0(ptr %sum)
   store i32 0, ptr %sum, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %i) #2
+  call void @llvm.lifetime.start.p0(ptr %i)
   store i32 0, ptr %i, align 4, !tbaa !8
   br label %for.cond
 
@@ -153,7 +153,7 @@ for.cond:                                         ; preds = %for.inc, %entry
   br i1 %cmp, label %for.body, label %for.cond.cleanup
 
 for.cond.cleanup:                                 ; preds = %for.cond
-  call void @llvm.lifetime.end.p0(ptr %i) #2
+  call void @llvm.lifetime.end.p0(ptr %i)
   br label %for.end
 
 for.body:                                         ; preds = %for.cond
@@ -241,22 +241,22 @@ for.body:                                         ; preds = %for.cond
   %shl42 = shl i32 %sub41, 16
   %rdd43 = add nsw i32 %sub36, %shl42
   store i32 %rdd43, ptr %r3, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e0) #2
+  call void @llvm.lifetime.start.p0(ptr %e0)
   %33 = load i32, ptr %r0, align 4, !tbaa !8
   %34 = load i32, ptr %r1, align 4, !tbaa !8
   %rdd44 = add i32 %33, %34
   store i32 %rdd44, ptr %e0, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e1) #2
+  call void @llvm.lifetime.start.p0(ptr %e1)
   %35 = load i32, ptr %r0, align 4, !tbaa !8
   %36 = load i32, ptr %r1, align 4, !tbaa !8
   %sub45 = sub i32 %35, %36
   store i32 %sub45, ptr %e1, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e2) #2
+  call void @llvm.lifetime.start.p0(ptr %e2)
   %37 = load i32, ptr %r2, align 4, !tbaa !8
   %38 = load i32, ptr %r3, align 4, !tbaa !8
   %rdd46 = add i32 %37, %38
   store i32 %rdd46, ptr %e2, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e3) #2
+  call void @llvm.lifetime.start.p0(ptr %e3)
   %39 = load i32, ptr %r2, align 4, !tbaa !8
   %40 = load i32, ptr %r3, align 4, !tbaa !8
   %sub47 = sub i32 %39, %40
@@ -293,10 +293,10 @@ for.body:                                         ; preds = %for.cond
   %rrrayidx61 = getelementptr inbounds [4 x [4 x i32]], ptr %emp, i64 0, i64 %idxprom60
   %rrrayidx62 = getelementptr inbounds [4 x i32], ptr %rrrayidx61, i64 0, i64 3
   store i32 %sub59, ptr %rrrayidx62, align 4, !tbaa !8
-  call void @llvm.lifetime.end.p0(ptr %e3) #2
-  call void @llvm.lifetime.end.p0(ptr %e2) #2
-  call void @llvm.lifetime.end.p0(ptr %e1) #2
-  call void @llvm.lifetime.end.p0(ptr %e0) #2
+  call void @llvm.lifetime.end.p0(ptr %e3)
+  call void @llvm.lifetime.end.p0(ptr %e2)
+  call void @llvm.lifetime.end.p0(ptr %e1)
+  call void @llvm.lifetime.end.p0(ptr %e0)
   br label %for.inc
 
 for.inc:                                          ; preds = %for.body
@@ -316,7 +316,7 @@ for.inc:                                          ; preds = %for.body
   br label %for.cond, !llvm.loop !11
 
 for.end:                                          ; preds = %for.cond.cleanup
-  call void @llvm.lifetime.start.p0(ptr %i65) #2
+  call void @llvm.lifetime.start.p0(ptr %i65)
   store i32 0, ptr %i65, align 4, !tbaa !8
   br label %for.cond66
 
@@ -326,11 +326,11 @@ for.cond66:                                       ; preds = %for.inc114, %for.en
   br i1 %cmp67, label %for.body70, label %for.cond.cleanup69
 
 for.cond.cleanup69:                               ; preds = %for.cond66
-  call void @llvm.lifetime.end.p0(ptr %i65) #2
+  call void @llvm.lifetime.end.p0(ptr %i65)
   br label %for.end116
 
 for.body70:                                       ; preds = %for.cond66
-  call void @llvm.lifetime.start.p0(ptr %e071) #2
+  call void @llvm.lifetime.start.p0(ptr %e071)
   %rrrayidx72 = getelementptr inbounds [4 x [4 x i32]], ptr %emp, i64 0, i64 0
   %59 = load i32, ptr %i65, align 4, !tbaa !8
   %idxprom73 = sext i32 %59 to i64
@@ -343,7 +343,7 @@ for.body70:                                       ; preds = %for.cond66
   %62 = load i32, ptr %rrrayidx77, align 4, !tbaa !8
   %rdd78 = add i32 %60, %62
   store i32 %rdd78, ptr %e071, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e179) #2
+  call void @llvm.lifetime.start.p0(ptr %e179)
   %rrrayidx80 = getelementptr inbounds [4 x [4 x i32]], ptr %emp, i64 0, i64 0
   %63 = load i32, ptr %i65, align 4, !tbaa !8
   %idxprom81 = sext i32 %63 to i64
@@ -356,7 +356,7 @@ for.body70:                                       ; preds = %for.cond66
   %66 = load i32, ptr %rrrayidx85, align 4, !tbaa !8
   %sub86 = sub i32 %64, %66
   store i32 %sub86, ptr %e179, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e287) #2
+  call void @llvm.lifetime.start.p0(ptr %e287)
   %rrrayidx88 = getelementptr inbounds [4 x [4 x i32]], ptr %emp, i64 0, i64 2
   %67 = load i32, ptr %i65, align 4, !tbaa !8
   %idxprom89 = sext i32 %67 to i64
@@ -369,7 +369,7 @@ for.body70:                                       ; preds = %for.cond66
   %70 = load i32, ptr %rrrayidx93, align 4, !tbaa !8
   %rdd94 = add i32 %68, %70
   store i32 %rdd94, ptr %e287, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e395) #2
+  call void @llvm.lifetime.start.p0(ptr %e395)
   %rrrayidx96 = getelementptr inbounds [4 x [4 x i32]], ptr %emp, i64 0, i64 2
   %71 = load i32, ptr %i65, align 4, !tbaa !8
   %idxprom97 = sext i32 %71 to i64
@@ -398,10 +398,10 @@ for.body70:                                       ; preds = %for.cond66
   %82 = load i32, ptr %e395, align 4, !tbaa !8
   %sub106 = sub nsw i32 %81, %82
   store i32 %sub106, ptr %r3, align 4, !tbaa !8
-  call void @llvm.lifetime.end.p0(ptr %e395) #2
-  call void @llvm.lifetime.end.p0(ptr %e287) #2
-  call void @llvm.lifetime.end.p0(ptr %e179) #2
-  call void @llvm.lifetime.end.p0(ptr %e071) #2
+  call void @llvm.lifetime.end.p0(ptr %e395)
+  call void @llvm.lifetime.end.p0(ptr %e287)
+  call void @llvm.lifetime.end.p0(ptr %e179)
+  call void @llvm.lifetime.end.p0(ptr %e071)
   %83 = load i32, ptr %r0, align 4, !tbaa !8
   %call = call i32 @twoabs(i32 noundef %83)
   %84 = load i32, ptr %r1, align 4, !tbaa !8
@@ -432,28 +432,28 @@ for.end116:                                       ; preds = %for.cond.cleanup69
   %shr = lshr i32 %90, 16
   %rdd119 = add i32 %conv118, %shr
   %shr120 = lshr i32 %rdd119, 1
-  call void @llvm.lifetime.end.p0(ptr %sum) #2
-  call void @llvm.lifetime.end.p0(ptr %r3) #2
-  call void @llvm.lifetime.end.p0(ptr %r2) #2
-  call void @llvm.lifetime.end.p0(ptr %r1) #2
-  call void @llvm.lifetime.end.p0(ptr %r0) #2
-  call void @llvm.lifetime.end.p0(ptr %emp) #2
+  call void @llvm.lifetime.end.p0(ptr %sum)
+  call void @llvm.lifetime.end.p0(ptr %r3)
+  call void @llvm.lifetime.end.p0(ptr %r2)
+  call void @llvm.lifetime.end.p0(ptr %r1)
+  call void @llvm.lifetime.end.p0(ptr %r0)
+  call void @llvm.lifetime.end.p0(ptr %emp)
   ret i32 %shr120
 }
 
 ; Function Attrs: nocallback nofree nosync nounwind willreturn memory(argmem: readwrite)
-declare void @llvm.lifetime.start.p0(ptr nocapture) #1
+declare void @llvm.lifetime.start.p0(ptr nocapture)
 
 ; Function Attrs: nocallback nofree nosync nounwind willreturn memory(argmem: readwrite)
-declare void @llvm.lifetime.end.p0(ptr nocapture) #1
+declare void @llvm.lifetime.end.p0(ptr nocapture)
 
 ; Function Attrs: nounwind uwtable
-define internal i32 @twoabs(i32 noundef %r) #0 {
+define internal i32 @twoabs(i32 noundef %r) {
 entry:
   %r.addr = alloca i32, align 4
   %s = alloca i32, align 4
   store i32 %r, ptr %r.addr, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %s) #2
+  call void @llvm.lifetime.start.p0(ptr %s)
   %0 = load i32, ptr %r.addr, align 4, !tbaa !8
   %shr = lshr i32 %0, 15
   %rnd = and i32 %shr, 65537
@@ -464,14 +464,10 @@ entry:
   %rdd = add i32 %1, %2
   %3 = load i32, ptr %s, align 4, !tbaa !8
   %xor = xor i32 %rdd, %3
-  call void @llvm.lifetime.end.p0(ptr %s) #2
+  call void @llvm.lifetime.end.p0(ptr %s)
   ret i32 %xor
 }
 
-attributes #0 = { nounwind uwtable "frame-pointer"="non-leaf" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="generic" "target-features"="+fp-armv8,+neon,+v8a,-fmv" "unsafe-fp-math"="true" }
-attributes #1 = { nocallback nofree nosync nounwind willreturn memory(argmem: readwrite) }
-attributes #2 = { nounwind }
-
 !4 = !{!5, !5, i64 0}
 !5 = !{!"any pointer", !6, i64 0}
 !6 = !{!"omnipotent char", !7, i64 0}
diff --git a/llvm/test/Transforms/PhaseOrdering/ARM/arm_add_q7.ll b/llvm/test/Transforms/PhaseOrdering/ARM/arm_add_q7.ll
index 7fb72e6c5147a..811957c7ea5d3 100644
--- a/llvm/test/Transforms/PhaseOrdering/ARM/arm_add_q7.ll
+++ b/llvm/test/Transforms/PhaseOrdering/ARM/arm_add_q7.ll
@@ -170,5 +170,5 @@ unreachable:                                      ; preds = %cleanup
 declare void @llvm.lifetime.start.p0(ptr nocapture)
 declare void @llvm.lifetime.end.p0(ptr nocapture)
 
-attributes #0 = { nounwind "frame-pointer"="all" "min-legal-vector-width"="0" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="cortex-m55" "target-features"="+armv8.1-m.main,+dsp,+fp-armv8d16,+fp-armv8d16sp,+fp16,+fp64,+fullfp16,+hwdiv,+lob,+mve,+mve.fp,+ras,+strict-align,+thumb-mode,+vfp2,+vfp2sp,+vfp3d16,+vfp3d16sp,+vfp4d16,+vfp4d16sp,-aes,-bf16,-cdecp0,-cdecp1,-cdecp2,-cdecp3,-cdecp4,-cdecp5,-cdecp6,-cdecp7,-crc,-crypto,-d32,-dotprod,-fp-armv8,-fp-armv8sp,-fp16fml,-hwdiv-arm,-i8mm,-neon,-sb,-sha2,-vfp3,-vfp3sp,-vfp4,-vfp4sp" "unsafe-fp-math"="true" }
-attributes #1 = { alwaysinline nounwind "frame-pointer"="all" "min-legal-vector-width"="0" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "targe...
[truncated]

llvmbot · 2025-10-23T12:57:42Z

@llvm/pr-subscribers-backend-nvptx

Author: None (paperchalice)

Changes

Post cleanup for #164534.

Patch is 144.07 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/164787.diff

59 Files Affected:

(modified) llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll (+7-11)
(modified) llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll (+20-25)
(modified) llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll (+38-42)
(modified) llvm/test/Transforms/PhaseOrdering/ARM/arm_add_q7.ll (+2-2)
(modified) llvm/test/Transforms/PhaseOrdering/ARM/arm_fill_q7.ll (+10-15)
(modified) llvm/test/Transforms/PhaseOrdering/ARM/arm_mean_q7.ll (+1-1)
(modified) llvm/test/Transforms/PhaseOrdering/ARM/arm_mult_q15.ll (+2-2)
(modified) llvm/test/Transforms/PhaseOrdering/X86/fmaddsub.ll (+16-18)
(modified) llvm/test/Transforms/PhaseOrdering/X86/vdiv-nounroll.ll (+5-9)
(modified) llvm/test/Transforms/PhaseOrdering/X86/vdiv.ll (+1-1)
(modified) llvm/test/Transforms/PhaseOrdering/X86/vector-reductions-expanded.ll (+8-10)
(modified) llvm/test/Transforms/SLPVectorizer/NVPTX/vectorizable-intrinsic.ll (+6-9)
(modified) llvm/test/Transforms/SLPVectorizer/X86/crash_bullet3.ll (+1-3)
(modified) llvm/test/Transforms/SLPVectorizer/X86/crash_flop7.ll (+1-3)
(modified) llvm/test/Transforms/SLPVectorizer/X86/debug_info.ll (+2-5)
(modified) llvm/test/Transforms/SLPVectorizer/X86/pr16899.ll (+1-3)
(modified) llvm/test/Transforms/SLPVectorizer/X86/vector_gep.ll (+1-4)
(modified) llvm/test/Transforms/SLPVectorizer/X86/vectorize-pair-path.ll (+1-3)
(modified) llvm/test/Transforms/SLPVectorizer/consecutive-access.ll (+8-10)
(modified) llvm/test/Transforms/SLPVectorizer/insert-element-build-vector-inseltpoison.ll (+13-15)
(modified) llvm/test/Transforms/SLPVectorizer/insert-element-build-vector.ll (+14-16)
(modified) llvm/test/Transforms/SROA/mem-par-metadata-sroa.ll (+2-6)
(modified) llvm/test/Transforms/SafeStack/ARM/debug.ll (+3-4)
(modified) llvm/test/Transforms/SafeStack/X86/debug-loc.ll (+2-3)
(modified) llvm/test/Transforms/SafeStack/X86/debug-loc2.ll (+2-3)
(modified) llvm/test/Transforms/SampleProfile/Inputs/profile-symbol-list.ll (+4-4)
(modified) llvm/test/Transforms/SampleProfile/branch.ll (+7-11)
(modified) llvm/test/Transforms/SampleProfile/csspgo-import-list.ll (+1-1)
(modified) llvm/test/Transforms/SampleProfile/csspgo-inline-debug.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/csspgo-inline.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/csspgo-summary.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/csspgo-use-preinliner.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/entry_counts_cold.ll (+2-3)
(modified) llvm/test/Transforms/SampleProfile/entry_counts_missing_dbginfo.ll (+2-3)
(modified) llvm/test/Transforms/SampleProfile/fsafdo_test.ll (+3-5)
(modified) llvm/test/Transforms/SampleProfile/gcc-simple.ll (+5-8)
(modified) llvm/test/Transforms/SampleProfile/inline-act.ll (+2-4)
(modified) llvm/test/Transforms/SampleProfile/misexpect.ll (+8-14)
(modified) llvm/test/Transforms/SampleProfile/norepeated-icp-2.ll (+4-7)
(modified) llvm/test/Transforms/SampleProfile/norepeated-icp-3.ll (+3-4)
(modified) llvm/test/Transforms/SampleProfile/norepeated-icp-4.ll (+1-1)
(modified) llvm/test/Transforms/SampleProfile/norepeated-icp.ll (+3-5)
(modified) llvm/test/Transforms/SampleProfile/offset.ll (+1-1)
(modified) llvm/test/Transforms/SampleProfile/profile-context-order.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/profile-context-tracker-debug.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/profile-context-tracker.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/profile-topdown-order.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/propagate.ll (+4-7)
(modified) llvm/test/Transforms/SampleProfile/pseudo-probe-discriminator.ll (+2-7)
(modified) llvm/test/Transforms/SampleProfile/pseudo-probe-icp-factor.ll (+1-1)
(modified) llvm/test/Transforms/SampleProfile/pseudo-probe-no-debug-info.ll (+2-6)
(modified) llvm/test/Transforms/SampleProfile/remarks.ll (+2-2)
(modified) llvm/test/Transforms/SampleProfile/uniqname.ll (+6-7)
(modified) llvm/test/Transforms/Scalarizer/dbginfo.ll (+2-5)
(modified) llvm/test/Transforms/SimplifyCFG/Hexagon/switch-to-lookup-table.ll (+1-3)
(modified) llvm/test/Transforms/SimplifyCFG/X86/merge-cleanuppads.ll (+8-12)
(modified) llvm/test/Transforms/SimplifyCFG/pr50060-constantfold-loopid.ll (+11-15)
(modified) llvm/test/Transforms/StraightLineStrengthReduce/AMDGPU/pr23975.ll (+1-3)
(modified) llvm/test/Transforms/StructurizeCFG/nested-loop-order.ll (-4)

diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll
index 27e8fd0ece55f..3b617502ed955 100644
--- a/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll
+++ b/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_muladd.ll
@@ -6,7 +6,7 @@ target triple = "aarch64"
 
 ; This function (a 16x reduction of a[i] * b[i]) should be vectorized successfully.
 
-define dso_local nofpclass(nan inf) float @vmlaq(ptr noundef %0, ptr noundef %1) #0 {
+define dso_local nofpclass(nan inf) float @vmlaq(ptr noundef %0, ptr noundef %1) {
 ; CHECK-LABEL: define dso_local nofpclass(nan inf) float @vmlaq
 ; CHECK-SAME: (ptr noundef readonly captures(none) [[TMP0:%.*]], ptr noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
 ; CHECK-NEXT:    [[TMP3:%.*]] = load <16 x float>, ptr [[TMP0]], align 4, !tbaa [[TBAA4:![0-9]+]]
@@ -21,9 +21,9 @@ define dso_local nofpclass(nan inf) float @vmlaq(ptr noundef %0, ptr noundef %1)
   %6 = alloca i32, align 4
   store ptr %0, ptr %3, align 8, !tbaa !4
   store ptr %1, ptr %4, align 8, !tbaa !4
-  call void @llvm.lifetime.start.p0(ptr %5) #2
+  call void @llvm.lifetime.start.p0(ptr %5)
   store float 0.000000e+00, ptr %5, align 4, !tbaa !9
-  call void @llvm.lifetime.start.p0(ptr %6) #2
+  call void @llvm.lifetime.start.p0(ptr %6)
   store i32 0, ptr %6, align 4, !tbaa !11
   br label %7
 
@@ -33,7 +33,7 @@ define dso_local nofpclass(nan inf) float @vmlaq(ptr noundef %0, ptr noundef %1)
   br i1 %9, label %11, label %10
 
 10:                                               ; preds = %7
-  call void @llvm.lifetime.end.p0(ptr %6) #2
+  call void @llvm.lifetime.end.p0(ptr %6)
   br label %28
 
 11:                                               ; preds = %7
@@ -61,16 +61,12 @@ define dso_local nofpclass(nan inf) float @vmlaq(ptr noundef %0, ptr noundef %1)
 
 28:                                               ; preds = %10
   %29 = load float, ptr %5, align 4, !tbaa !9
-  call void @llvm.lifetime.end.p0(ptr %5) #2
+  call void @llvm.lifetime.end.p0(ptr %5)
   ret float %29
 }
 
-declare void @llvm.lifetime.start.p0(ptr captures(none)) #1
-declare void @llvm.lifetime.end.p0(ptr captures(none)) #1
-
-attributes #0 = { nounwind uwtable "frame-pointer"="non-leaf" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="generic" "target-features"="+fp-armv8,+neon,+v8a,-fmv" "unsafe-fp-math"="true" }
-attributes #1 = { nocallback nofree nosync nounwind willreturn memory(argmem: readwrite) }
-attributes #2 = { nounwind }
+declare void @llvm.lifetime.start.p0(ptr captures(none))
+declare void @llvm.lifetime.end.p0(ptr captures(none))
 
 !llvm.module.flags = !{!0, !1, !2}
 !llvm.ident = !{!3}
diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll
index 68bfbc1344f23..eefde9ddfdc1d 100644
--- a/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll
+++ b/llvm/test/Transforms/PhaseOrdering/AArch64/reduce_submuladd.ll
@@ -6,7 +6,7 @@ target triple = "aarch64"
 
 ; This function (a more complex reduction of (a[i] - b[i]) * itself) should be vectorized successfully.
 
-define dso_local noundef nofpclass(nan inf) float @_Z4testPKfS0_ii(ptr noundef %0, ptr noundef %1, i32 noundef %2, i32 noundef %3) #0 {
+define dso_local noundef nofpclass(nan inf) float @_Z4testPKfS0_ii(ptr noundef %0, ptr noundef %1, i32 noundef %2, i32 noundef %3) {
 ; CHECK-LABEL: define dso_local noundef nofpclass(nan inf) float @_Z4testPKfS0_ii
 ; CHECK-SAME: (ptr noundef readonly captures(none) [[TMP0:%.*]], ptr noundef readonly captures(none) [[TMP1:%.*]], i32 noundef [[TMP2:%.*]], i32 noundef [[TMP3:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
 ; CHECK-NEXT:  .preheader.i:
@@ -125,7 +125,7 @@ define dso_local noundef nofpclass(nan inf) float @_Z4testPKfS0_ii(ptr noundef %
   ret float %13
 }
 
-define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr noundef %0, ptr noundef %1, i32 noundef %2, i32 noundef %3) #1 {
+define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr noundef %0, ptr noundef %1, i32 noundef %2, i32 noundef %3) {
   %5 = alloca ptr, align 8
   %6 = alloca ptr, align 8
   %7 = alloca i32, align 4
@@ -143,15 +143,15 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
   store ptr %1, ptr %6, align 8, !tbaa !4
   store i32 %2, ptr %7, align 4, !tbaa !9
   store i32 %3, ptr %8, align 4, !tbaa !9
-  call void @llvm.lifetime.start.p0(ptr %9) #3
+  call void @llvm.lifetime.start.p0(ptr %9)
   store i32 3, ptr %9, align 4, !tbaa !9
-  call void @llvm.lifetime.start.p0(ptr %10) #3
+  call void @llvm.lifetime.start.p0(ptr %10)
   store i32 3, ptr %10, align 4, !tbaa !9
-  call void @llvm.lifetime.start.p0(ptr %11) #3
+  call void @llvm.lifetime.start.p0(ptr %11)
   store i32 7, ptr %11, align 4, !tbaa !9
-  call void @llvm.lifetime.start.p0(ptr %12) #3
+  call void @llvm.lifetime.start.p0(ptr %12)
   store float 0.000000e+00, ptr %12, align 4, !tbaa !11
-  call void @llvm.lifetime.start.p0(ptr %13) #3
+  call void @llvm.lifetime.start.p0(ptr %13)
   store i32 0, ptr %13, align 4, !tbaa !9
   br label %18
 
@@ -162,13 +162,13 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
 
 21:                                               ; preds = %18
   store i32 2, ptr %14, align 4
-  call void @llvm.lifetime.end.p0(ptr %13) #3
+  call void @llvm.lifetime.end.p0(ptr %13)
   br label %62
 
 22:                                               ; preds = %18
-  call void @llvm.lifetime.start.p0(ptr %15) #3
+  call void @llvm.lifetime.start.p0(ptr %15)
   store float 0.000000e+00, ptr %15, align 4, !tbaa !11
-  call void @llvm.lifetime.start.p0(ptr %16) #3
+  call void @llvm.lifetime.start.p0(ptr %16)
   store i32 0, ptr %16, align 4, !tbaa !9
   br label %23
 
@@ -179,11 +179,11 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
 
 26:                                               ; preds = %23
   store i32 5, ptr %14, align 4
-  call void @llvm.lifetime.end.p0(ptr %16) #3
+  call void @llvm.lifetime.end.p0(ptr %16)
   br label %47
 
 27:                                               ; preds = %23
-  call void @llvm.lifetime.start.p0(ptr %17) #3
+  call void @llvm.lifetime.start.p0(ptr %17)
   %28 = load ptr, ptr %5, align 8, !tbaa !4
   %29 = load i32, ptr %16, align 4, !tbaa !9
   %30 = sext i32 %29 to i64
@@ -202,7 +202,7 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
   %42 = load float, ptr %15, align 4, !tbaa !11
   %43 = fadd fast float %42, %41
   store float %43, ptr %15, align 4, !tbaa !11
-  call void @llvm.lifetime.end.p0(ptr %17) #3
+  call void @llvm.lifetime.end.p0(ptr %17)
   br label %44
 
 44:                                               ; preds = %27
@@ -226,7 +226,7 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
   %57 = load float, ptr %12, align 4, !tbaa !11
   %58 = fadd fast float %57, %56
   store float %58, ptr %12, align 4, !tbaa !11
-  call void @llvm.lifetime.end.p0(ptr %15) #3
+  call void @llvm.lifetime.end.p0(ptr %15)
   br label %59
 
 59:                                               ; preds = %47
@@ -238,20 +238,15 @@ define internal noundef nofpclass(nan inf) float @_ZL6reduceILi7EEfPKfS1_ii(ptr
 62:                                               ; preds = %21
   %63 = load float, ptr %12, align 4, !tbaa !11
   store i32 1, ptr %14, align 4
-  call void @llvm.lifetime.end.p0(ptr %12) #3
-  call void @llvm.lifetime.end.p0(ptr %11) #3
-  call void @llvm.lifetime.end.p0(ptr %10) #3
-  call void @llvm.lifetime.end.p0(ptr %9) #3
+  call void @llvm.lifetime.end.p0(ptr %12)
+  call void @llvm.lifetime.end.p0(ptr %11)
+  call void @llvm.lifetime.end.p0(ptr %10)
+  call void @llvm.lifetime.end.p0(ptr %9)
   ret float %63
 }
 
-declare void @llvm.lifetime.start.p0(ptr captures(none)) #2
-declare void @llvm.lifetime.end.p0(ptr captures(none)) #2
-
-attributes #0 = { mustprogress uwtable "frame-pointer"="non-leaf" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="generic" "target-features"="+fp-armv8,+neon,+v8a,-fmv" "unsafe-fp-math"="true" }
-attributes #1 = { inlinehint mustprogress nounwind uwtable "frame-pointer"="non-leaf" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="generic" "target-features"="+fp-armv8,+neon,+v8a,-fmv" "unsafe-fp-math"="true" }
-attributes #2 = { nocallback nofree nosync nounwind willreturn memory(argmem: readwrite) }
-attributes #3 = { nounwind }
+declare void @llvm.lifetime.start.p0(ptr captures(none))
+declare void @llvm.lifetime.end.p0(ptr captures(none))
 
 !llvm.module.flags = !{!0, !1, !2}
 !llvm.ident = !{!3}
diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll
index 92e625deb11b1..82ecc3a0ed172 100644
--- a/llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll
+++ b/llvm/test/Transforms/PhaseOrdering/AArch64/slpordering.ll
@@ -10,7 +10,7 @@ target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128-Fn32"
 target triple = "aarch64"
 
 ; Function Attrs: nounwind uwtable
-define i32 @slpordering(ptr noundef %p1, i32 noundef %ip1, ptr noundef %p2, i32 noundef %ip2) #0 {
+define i32 @slpordering(ptr noundef %p1, i32 noundef %ip1, ptr noundef %p2, i32 noundef %ip2) {
 ; CHECK-LABEL: define range(i32 0, 65536) i32 @slpordering(
 ; CHECK-SAME: ptr noundef readonly captures(none) [[P1:%.*]], i32 noundef [[IP1:%.*]], ptr noundef readonly captures(none) [[P2:%.*]], i32 noundef [[IP2:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
 ; CHECK-NEXT:  [[ENTRY:.*:]]
@@ -136,14 +136,14 @@ entry:
   store i32 %ip1, ptr %ip1.addr, align 4, !tbaa !8
   store ptr %p2, ptr %p2.addr, align 8, !tbaa !4
   store i32 %ip2, ptr %ip2.addr, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %emp) #2
-  call void @llvm.lifetime.start.p0(ptr %r0) #2
-  call void @llvm.lifetime.start.p0(ptr %r1) #2
-  call void @llvm.lifetime.start.p0(ptr %r2) #2
-  call void @llvm.lifetime.start.p0(ptr %r3) #2
-  call void @llvm.lifetime.start.p0(ptr %sum) #2
+  call void @llvm.lifetime.start.p0(ptr %emp)
+  call void @llvm.lifetime.start.p0(ptr %r0)
+  call void @llvm.lifetime.start.p0(ptr %r1)
+  call void @llvm.lifetime.start.p0(ptr %r2)
+  call void @llvm.lifetime.start.p0(ptr %r3)
+  call void @llvm.lifetime.start.p0(ptr %sum)
   store i32 0, ptr %sum, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %i) #2
+  call void @llvm.lifetime.start.p0(ptr %i)
   store i32 0, ptr %i, align 4, !tbaa !8
   br label %for.cond
 
@@ -153,7 +153,7 @@ for.cond:                                         ; preds = %for.inc, %entry
   br i1 %cmp, label %for.body, label %for.cond.cleanup
 
 for.cond.cleanup:                                 ; preds = %for.cond
-  call void @llvm.lifetime.end.p0(ptr %i) #2
+  call void @llvm.lifetime.end.p0(ptr %i)
   br label %for.end
 
 for.body:                                         ; preds = %for.cond
@@ -241,22 +241,22 @@ for.body:                                         ; preds = %for.cond
   %shl42 = shl i32 %sub41, 16
   %rdd43 = add nsw i32 %sub36, %shl42
   store i32 %rdd43, ptr %r3, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e0) #2
+  call void @llvm.lifetime.start.p0(ptr %e0)
   %33 = load i32, ptr %r0, align 4, !tbaa !8
   %34 = load i32, ptr %r1, align 4, !tbaa !8
   %rdd44 = add i32 %33, %34
   store i32 %rdd44, ptr %e0, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e1) #2
+  call void @llvm.lifetime.start.p0(ptr %e1)
   %35 = load i32, ptr %r0, align 4, !tbaa !8
   %36 = load i32, ptr %r1, align 4, !tbaa !8
   %sub45 = sub i32 %35, %36
   store i32 %sub45, ptr %e1, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e2) #2
+  call void @llvm.lifetime.start.p0(ptr %e2)
   %37 = load i32, ptr %r2, align 4, !tbaa !8
   %38 = load i32, ptr %r3, align 4, !tbaa !8
   %rdd46 = add i32 %37, %38
   store i32 %rdd46, ptr %e2, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e3) #2
+  call void @llvm.lifetime.start.p0(ptr %e3)
   %39 = load i32, ptr %r2, align 4, !tbaa !8
   %40 = load i32, ptr %r3, align 4, !tbaa !8
   %sub47 = sub i32 %39, %40
@@ -293,10 +293,10 @@ for.body:                                         ; preds = %for.cond
   %rrrayidx61 = getelementptr inbounds [4 x [4 x i32]], ptr %emp, i64 0, i64 %idxprom60
   %rrrayidx62 = getelementptr inbounds [4 x i32], ptr %rrrayidx61, i64 0, i64 3
   store i32 %sub59, ptr %rrrayidx62, align 4, !tbaa !8
-  call void @llvm.lifetime.end.p0(ptr %e3) #2
-  call void @llvm.lifetime.end.p0(ptr %e2) #2
-  call void @llvm.lifetime.end.p0(ptr %e1) #2
-  call void @llvm.lifetime.end.p0(ptr %e0) #2
+  call void @llvm.lifetime.end.p0(ptr %e3)
+  call void @llvm.lifetime.end.p0(ptr %e2)
+  call void @llvm.lifetime.end.p0(ptr %e1)
+  call void @llvm.lifetime.end.p0(ptr %e0)
   br label %for.inc
 
 for.inc:                                          ; preds = %for.body
@@ -316,7 +316,7 @@ for.inc:                                          ; preds = %for.body
   br label %for.cond, !llvm.loop !11
 
 for.end:                                          ; preds = %for.cond.cleanup
-  call void @llvm.lifetime.start.p0(ptr %i65) #2
+  call void @llvm.lifetime.start.p0(ptr %i65)
   store i32 0, ptr %i65, align 4, !tbaa !8
   br label %for.cond66
 
@@ -326,11 +326,11 @@ for.cond66:                                       ; preds = %for.inc114, %for.en
   br i1 %cmp67, label %for.body70, label %for.cond.cleanup69
 
 for.cond.cleanup69:                               ; preds = %for.cond66
-  call void @llvm.lifetime.end.p0(ptr %i65) #2
+  call void @llvm.lifetime.end.p0(ptr %i65)
   br label %for.end116
 
 for.body70:                                       ; preds = %for.cond66
-  call void @llvm.lifetime.start.p0(ptr %e071) #2
+  call void @llvm.lifetime.start.p0(ptr %e071)
   %rrrayidx72 = getelementptr inbounds [4 x [4 x i32]], ptr %emp, i64 0, i64 0
   %59 = load i32, ptr %i65, align 4, !tbaa !8
   %idxprom73 = sext i32 %59 to i64
@@ -343,7 +343,7 @@ for.body70:                                       ; preds = %for.cond66
   %62 = load i32, ptr %rrrayidx77, align 4, !tbaa !8
   %rdd78 = add i32 %60, %62
   store i32 %rdd78, ptr %e071, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e179) #2
+  call void @llvm.lifetime.start.p0(ptr %e179)
   %rrrayidx80 = getelementptr inbounds [4 x [4 x i32]], ptr %emp, i64 0, i64 0
   %63 = load i32, ptr %i65, align 4, !tbaa !8
   %idxprom81 = sext i32 %63 to i64
@@ -356,7 +356,7 @@ for.body70:                                       ; preds = %for.cond66
   %66 = load i32, ptr %rrrayidx85, align 4, !tbaa !8
   %sub86 = sub i32 %64, %66
   store i32 %sub86, ptr %e179, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e287) #2
+  call void @llvm.lifetime.start.p0(ptr %e287)
   %rrrayidx88 = getelementptr inbounds [4 x [4 x i32]], ptr %emp, i64 0, i64 2
   %67 = load i32, ptr %i65, align 4, !tbaa !8
   %idxprom89 = sext i32 %67 to i64
@@ -369,7 +369,7 @@ for.body70:                                       ; preds = %for.cond66
   %70 = load i32, ptr %rrrayidx93, align 4, !tbaa !8
   %rdd94 = add i32 %68, %70
   store i32 %rdd94, ptr %e287, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %e395) #2
+  call void @llvm.lifetime.start.p0(ptr %e395)
   %rrrayidx96 = getelementptr inbounds [4 x [4 x i32]], ptr %emp, i64 0, i64 2
   %71 = load i32, ptr %i65, align 4, !tbaa !8
   %idxprom97 = sext i32 %71 to i64
@@ -398,10 +398,10 @@ for.body70:                                       ; preds = %for.cond66
   %82 = load i32, ptr %e395, align 4, !tbaa !8
   %sub106 = sub nsw i32 %81, %82
   store i32 %sub106, ptr %r3, align 4, !tbaa !8
-  call void @llvm.lifetime.end.p0(ptr %e395) #2
-  call void @llvm.lifetime.end.p0(ptr %e287) #2
-  call void @llvm.lifetime.end.p0(ptr %e179) #2
-  call void @llvm.lifetime.end.p0(ptr %e071) #2
+  call void @llvm.lifetime.end.p0(ptr %e395)
+  call void @llvm.lifetime.end.p0(ptr %e287)
+  call void @llvm.lifetime.end.p0(ptr %e179)
+  call void @llvm.lifetime.end.p0(ptr %e071)
   %83 = load i32, ptr %r0, align 4, !tbaa !8
   %call = call i32 @twoabs(i32 noundef %83)
   %84 = load i32, ptr %r1, align 4, !tbaa !8
@@ -432,28 +432,28 @@ for.end116:                                       ; preds = %for.cond.cleanup69
   %shr = lshr i32 %90, 16
   %rdd119 = add i32 %conv118, %shr
   %shr120 = lshr i32 %rdd119, 1
-  call void @llvm.lifetime.end.p0(ptr %sum) #2
-  call void @llvm.lifetime.end.p0(ptr %r3) #2
-  call void @llvm.lifetime.end.p0(ptr %r2) #2
-  call void @llvm.lifetime.end.p0(ptr %r1) #2
-  call void @llvm.lifetime.end.p0(ptr %r0) #2
-  call void @llvm.lifetime.end.p0(ptr %emp) #2
+  call void @llvm.lifetime.end.p0(ptr %sum)
+  call void @llvm.lifetime.end.p0(ptr %r3)
+  call void @llvm.lifetime.end.p0(ptr %r2)
+  call void @llvm.lifetime.end.p0(ptr %r1)
+  call void @llvm.lifetime.end.p0(ptr %r0)
+  call void @llvm.lifetime.end.p0(ptr %emp)
   ret i32 %shr120
 }
 
 ; Function Attrs: nocallback nofree nosync nounwind willreturn memory(argmem: readwrite)
-declare void @llvm.lifetime.start.p0(ptr nocapture) #1
+declare void @llvm.lifetime.start.p0(ptr nocapture)
 
 ; Function Attrs: nocallback nofree nosync nounwind willreturn memory(argmem: readwrite)
-declare void @llvm.lifetime.end.p0(ptr nocapture) #1
+declare void @llvm.lifetime.end.p0(ptr nocapture)
 
 ; Function Attrs: nounwind uwtable
-define internal i32 @twoabs(i32 noundef %r) #0 {
+define internal i32 @twoabs(i32 noundef %r) {
 entry:
   %r.addr = alloca i32, align 4
   %s = alloca i32, align 4
   store i32 %r, ptr %r.addr, align 4, !tbaa !8
-  call void @llvm.lifetime.start.p0(ptr %s) #2
+  call void @llvm.lifetime.start.p0(ptr %s)
   %0 = load i32, ptr %r.addr, align 4, !tbaa !8
   %shr = lshr i32 %0, 15
   %rnd = and i32 %shr, 65537
@@ -464,14 +464,10 @@ entry:
   %rdd = add i32 %1, %2
   %3 = load i32, ptr %s, align 4, !tbaa !8
   %xor = xor i32 %rdd, %3
-  call void @llvm.lifetime.end.p0(ptr %s) #2
+  call void @llvm.lifetime.end.p0(ptr %s)
   ret i32 %xor
 }
 
-attributes #0 = { nounwind uwtable "frame-pointer"="non-leaf" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="generic" "target-features"="+fp-armv8,+neon,+v8a,-fmv" "unsafe-fp-math"="true" }
-attributes #1 = { nocallback nofree nosync nounwind willreturn memory(argmem: readwrite) }
-attributes #2 = { nounwind }
-
 !4 = !{!5, !5, i64 0}
 !5 = !{!"any pointer", !6, i64 0}
 !6 = !{!"omnipotent char", !7, i64 0}
diff --git a/llvm/test/Transforms/PhaseOrdering/ARM/arm_add_q7.ll b/llvm/test/Transforms/PhaseOrdering/ARM/arm_add_q7.ll
index 7fb72e6c5147a..811957c7ea5d3 100644
--- a/llvm/test/Transforms/PhaseOrdering/ARM/arm_add_q7.ll
+++ b/llvm/test/Transforms/PhaseOrdering/ARM/arm_add_q7.ll
@@ -170,5 +170,5 @@ unreachable:                                      ; preds = %cleanup
 declare void @llvm.lifetime.start.p0(ptr nocapture)
 declare void @llvm.lifetime.end.p0(ptr nocapture)
 
-attributes #0 = { nounwind "frame-pointer"="all" "min-legal-vector-width"="0" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="cortex-m55" "target-features"="+armv8.1-m.main,+dsp,+fp-armv8d16,+fp-armv8d16sp,+fp16,+fp64,+fullfp16,+hwdiv,+lob,+mve,+mve.fp,+ras,+strict-align,+thumb-mode,+vfp2,+vfp2sp,+vfp3d16,+vfp3d16sp,+vfp4d16,+vfp4d16sp,-aes,-bf16,-cdecp0,-cdecp1,-cdecp2,-cdecp3,-cdecp4,-cdecp5,-cdecp6,-cdecp7,-crc,-crypto,-d32,-dotprod,-fp-armv8,-fp-armv8sp,-fp16fml,-hwdiv-arm,-i8mm,-neon,-sb,-sha2,-vfp3,-vfp3sp,-vfp4,-vfp4sp" "unsafe-fp-math"="true" }
-attributes #1 = { alwaysinline nounwind "frame-pointer"="all" "min-legal-vector-width"="0" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "targe...
[truncated]

arsenm · 2025-10-24T06:37:08Z

llvm/test/Transforms/PhaseOrdering/X86/vdiv-nounroll.ll

-declare void @llvm.lifetime.start.p0(ptr nocapture) #1
-declare void @llvm.lifetime.end.p0(ptr nocapture) #1
-
-attributes #0 = { nounwind ssp uwtable "frame-pointer"="all" "min-legal-vector-width"="0" "no-infs-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="penryn" "target-features"="+cx16,+cx8,+fxsr,+mmx,+sahf,+sse,+sse2,+sse3,+sse4.1,+ssse3,+x87" "tune-cpu"="generic" "unsafe-fp-math"="true" }


Shouldn't be touching all of these other attribute uses

#164605 suggested removing attributes sections, so I applied the same rule to Transform, but SampleProfile is an exception.

Post cleanup for llvm#164534.

paperchalice force-pushed the unsafe-fp-math-trans-3 branch 2 times, most recently from b49523f to 9253d7f Compare October 23, 2025 12:04

[test][Transforms] Remove unsafe-fp-math uses part 3 (NFC)

fcd47a3

paperchalice force-pushed the unsafe-fp-math-trans-3 branch from 9253d7f to fcd47a3 Compare October 23, 2025 12:28

paperchalice changed the title ~~[test][Transforms] Remove unsafe-fp-math uses part 3(NFC)~~ [test][Transforms] Remove unsafe-fp-math uses part 3 (NFC) Oct 23, 2025

paperchalice marked this pull request as ready for review October 23, 2025 12:56

paperchalice requested review from RKSimon, davemgreen and nikic October 23, 2025 12:57

llvmbot added backend:AMDGPU PGO Profile Guided Optimizations backend:NVPTX llvm:transforms labels Oct 23, 2025

arsenm reviewed Oct 24, 2025

View reviewed changes

nikic approved these changes Oct 24, 2025

View reviewed changes

paperchalice merged commit f8b81b4 into llvm:main Oct 24, 2025
16 checks passed

paperchalice deleted the unsafe-fp-math-trans-3 branch October 24, 2025 11:20

dvbuka pushed a commit to dvbuka/llvm-project that referenced this pull request Oct 27, 2025

[test][Transforms] Remove unsafe-fp-math uses part 3 (NFC) (llvm#164787)

02bfec1

Post cleanup for llvm#164534.

Lukacma pushed a commit to Lukacma/llvm-project that referenced this pull request Oct 29, 2025

[test][Transforms] Remove unsafe-fp-math uses part 3 (NFC) (llvm#164787)

1941a48

Post cleanup for llvm#164534.

aokblast pushed a commit to aokblast/llvm-project that referenced this pull request Oct 30, 2025

[test][Transforms] Remove unsafe-fp-math uses part 3 (NFC) (llvm#164787)

fa2448b

Post cleanup for llvm#164534.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[test][Transforms] Remove unsafe-fp-math uses part 3 (NFC) #164787

[test][Transforms] Remove unsafe-fp-math uses part 3 (NFC) #164787

Uh oh!

paperchalice commented Oct 23, 2025

Uh oh!

llvmbot commented Oct 23, 2025 •

edited

Loading

Uh oh!

llvmbot commented Oct 23, 2025

Uh oh!

arsenm Oct 24, 2025

Uh oh!

paperchalice Oct 24, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

[test][Transforms] Remove unsafe-fp-math uses part 3 (NFC) #164787

[test][Transforms] Remove unsafe-fp-math uses part 3 (NFC) #164787

Uh oh!

Conversation

paperchalice commented Oct 23, 2025

Uh oh!

llvmbot commented Oct 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

llvmbot commented Oct 23, 2025

Uh oh!

arsenm Oct 24, 2025

Choose a reason for hiding this comment

Uh oh!

paperchalice Oct 24, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

llvmbot commented Oct 23, 2025 •

edited

Loading