Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[FuncAttrs] Deduce noundef attributes for return values #76553

Merged
merged 2 commits into from
Dec 31, 2023
Merged

[FuncAttrs] Deduce noundef attributes for return values #76553

merged 2 commits into from
Dec 31, 2023

Conversation

dtcxzyw
Copy link
Member

@dtcxzyw dtcxzyw commented Dec 29, 2023 •
edited
Loading

This patch deduces noundef attributes for return values.
IIUC, a function returns noundef values iff all of its return values are guaranteed not to be undef or poison.
Definition of noundef from LangRef:

noundef
This attribute applies to parameters and return values. If the value representation contains any 
undefined or poison bits, the behavior is undefined. Note that this does not refer to padding 
introduced by the type’s storage representation.

Alive2: https://alive2.llvm.org/ce/z/g8Eis6

Compile-time impact: http://llvm-compile-time-tracker.com/compare.php?from=30dcc33c4ea3ab50397a7adbe85fe977d4a400bd&to=c5e8738d4bfbf1e97e3f455fded90b791f223d74&stat=instructions:u

stage1-O3 stage1-ReleaseThinLTO stage1-ReleaseLTO-g stage1-O0-g stage2-O3 stage2-O0-g stage2-clang
+0.01% +0.01% -0.01% +0.01% +0.03% -0.04% +0.01%

The motivation of this patch is to reduce the number of freeze insts and enable more optimizations.
Example:

diff --git a/bench/flac/optimized/replaygain.c.ll b/bench/flac/optimized/replaygain.c.ll
index fa826475..413bd717 100644
--- a/bench/flac/optimized/replaygain.c.ll
+++ b/bench/flac/optimized/replaygain.c.ll
@@ -63,7 +63,7 @@ entry:
 declare i32 @InitGainAnalysis(i64 noundef) local_unnamed_addr #1
 
 ; Function Attrs: nounwind sspstrong uwtable
-define dso_local i32 @grabbag__replaygain_analyze(ptr nocapture noundef readonly %input, i32 noundef %is_stereo, i32 noundef %bps, i32 noundef %samples) local_unnamed_addr #0 {
+define dso_local noundef i32 @grabbag__replaygain_analyze(ptr nocapture noundef readonly %input, i32 noundef %is_stereo, i32 noundef %bps, i32 noundef %samples) local_unnamed_addr #0 {
 entry:
   %cmp = icmp eq i32 %bps, 16
   br i1 %cmp, label %if.then, label %if.else71
@@ -337,7 +337,7 @@ entry:
 declare float @GetTitleGain() local_unnamed_addr #1
 
 ; Function Attrs: nounwind sspstrong uwtable
-define dso_local ptr @grabbag__replaygain_analyze_file(ptr noundef %filename, ptr nocapture noundef writeonly %title_gain, ptr nocapture noundef writeonly %title_peak) local_unnamed_addr #0 {
+define dso_local noundef ptr @grabbag__replaygain_analyze_file(ptr noundef %filename, ptr nocapture noundef writeonly %title_gain, ptr nocapture noundef writeonly %title_peak) local_unnamed_addr #0 {
 entry:
   %instance = alloca %struct.DecoderInstance, align 4
   %call = tail call ptr @FLAC__stream_decoder_new() #15
@@ -392,7 +392,7 @@ declare i32 @FLAC__stream_decoder_set_metadata_respond(ptr noundef, i32 noundef)
 declare i32 @FLAC__stream_decoder_init_file(ptr noundef, ptr noundef, ptr noundef, ptr noundef, ptr noundef, ptr noundef) local_unnamed_addr #1
 
 ; Function Attrs: nounwind sspstrong uwtable
-define internal i32 @write_callback_(ptr nocapture readnone %decoder, ptr nocapture noundef readonly %frame, ptr nocapture noundef readonly %buffer, ptr nocapture noundef %client_data) #0 {
+define internal noundef i32 @write_callback_(ptr nocapture readnone %decoder, ptr nocapture noundef readonly %frame, ptr nocapture noundef readonly %buffer, ptr nocapture noundef %client_data) #0 {
 entry:
   %bits_per_sample1 = getelementptr inbounds %struct.FLAC__FrameHeader, ptr %frame, i64 0, i32 4
   %0 = load i32, ptr %bits_per_sample1, align 8
@@ -429,23 +429,16 @@ land.lhs.true14:                                  ; preds = %land.lhs.true11
   %cmp16 = icmp eq i32 %2, %8
   br i1 %cmp16, label %if.end, label %if.end.thread
 
-if.end.thread:                                    ; preds = %land.lhs.true, %land.lhs.true14, %land.lhs.true11, %land.lhs.true8, %entry
-  store i32 1, ptr %error, align 4
-  br label %9
-
 if.end:                                           ; preds = %land.lhs.true14
   %conv = zext i1 %cmp to i32
   %call = tail call i32 @grabbag__replaygain_analyze(ptr noundef %buffer, i32 noundef %conv, i32 noundef %0, i32 noundef %3), !range !14
-  %call.fr = freeze i32 %call
-  %lnot.ext = xor i32 %call.fr, 1
-  store i32 %lnot.ext, ptr %error, align 4
-  %tobool22.not = icmp ne i32 %lnot.ext, 0
-  %spec.select = zext i1 %tobool22.not to i32
-  br label %9
-
-9:                                                ; preds = %if.end, %if.end.thread
-  %10 = phi i32 [ 1, %if.end.thread ], [ %spec.select, %if.end ]
-  ret i32 %10
+  %lnot.ext = xor i32 %call, 1
+  br label %if.end.thread
+
+if.end.thread:                                    ; preds = %entry, %land.lhs.true8, %land.lhs.true11, %land.lhs.true14, %land.lhs.true, %if.end
+  %storemerge = phi i32 [ %lnot.ext, %if.end ], [ 1, %land.lhs.true ], [ 1, %land.lhs.true14 ], [ 1, %land.lhs.true11 ], [ 1, %land.lhs.true8 ], [ 1, %entry ]
+  store i32 %storemerge, ptr %error, align 4
+  ret i32 %storemerge
 }
 
 ; Function Attrs: nounwind sspstrong uwtable

@llvmbot llvmbot added clang Clang issues not falling into any other category backend:X86 debuginfo LTO Link time optimization (regular/full LTO or ThinLTO) llvm:transforms labels Dec 29, 2023
@llvmbot
Copy link
Member

llvmbot commented Dec 29, 2023 •
edited
Loading

@llvm/pr-subscribers-lld
@llvm/pr-subscribers-lld-coff
@llvm/pr-subscribers-lld-elf
@llvm/pr-subscribers-debuginfo
@llvm/pr-subscribers-llvm-transforms

@llvm/pr-subscribers-clang

Author: Yingwei Zheng (dtcxzyw)

Changes

This patch deduces noundef attributes for return values.
IIUC, a function returns noundef values iff all of its return values are guaranteed not to be undef or poison.
Definition of noundef from LangRef:

noundef
This attribute applies to parameters and return values. If the value representation contains any 
undefined or poison bits, the behavior is undefined. Note that this does not refer to padding 
introduced by the type’s storage representation.

Alive2: https://alive2.llvm.org/ce/z/g8Eis6

Compile-time impact: http://llvm-compile-time-tracker.com/compare.php?from=7f69c8b3a6c02ea32fefb16c2016dfa1ba994858&to=1dafc281ff8c04bb0a968fb3d898f08876dc59e0&stat=instructions:u

stage1-O3 stage1-ReleaseThinLTO stage1-ReleaseLTO-g stage1-O0-g stage2-O3 stage2-O0-g stage2-clang
+0.01% +0.01% -0.01% +0.00% +0.03% +0.02% +0.01%

The motivation of this patch is to reduce the number of freeze insts and enable more optimizations.
Example:

diff --git a/bench/flac/optimized/replaygain.c.ll b/bench/flac/optimized/replaygain.c.ll
index fa826475..413bd717 100644
--- a/bench/flac/optimized/replaygain.c.ll
+++ b/bench/flac/optimized/replaygain.c.ll
@@ -63,7 +63,7 @@ entry:
 declare i32 @<!-- -->InitGainAnalysis(i64 noundef) local_unnamed_addr #<!-- -->1
 
 ; Function Attrs: nounwind sspstrong uwtable
-define dso_local i32 @<!-- -->grabbag__replaygain_analyze(ptr nocapture noundef readonly %input, i32 noundef %is_stereo, i32 noundef %bps, i32 noundef %samples) local_unnamed_addr #<!-- -->0 {
+define dso_local noundef i32 @<!-- -->grabbag__replaygain_analyze(ptr nocapture noundef readonly %input, i32 noundef %is_stereo, i32 noundef %bps, i32 noundef %samples) local_unnamed_addr #<!-- -->0 {
 entry:
   %cmp = icmp eq i32 %bps, 16
   br i1 %cmp, label %if.then, label %if.else71
@@ -337,7 +337,7 @@ entry:
 declare float @<!-- -->GetTitleGain() local_unnamed_addr #<!-- -->1
 
 ; Function Attrs: nounwind sspstrong uwtable
-define dso_local ptr @<!-- -->grabbag__replaygain_analyze_file(ptr noundef %filename, ptr nocapture noundef writeonly %title_gain, ptr nocapture noundef writeonly %title_peak) local_unnamed_addr #<!-- -->0 {
+define dso_local noundef ptr @<!-- -->grabbag__replaygain_analyze_file(ptr noundef %filename, ptr nocapture noundef writeonly %title_gain, ptr nocapture noundef writeonly %title_peak) local_unnamed_addr #<!-- -->0 {
 entry:
   %instance = alloca %struct.DecoderInstance, align 4
   %call = tail call ptr @<!-- -->FLAC__stream_decoder_new() #<!-- -->15
@@ -392,7 +392,7 @@ declare i32 @<!-- -->FLAC__stream_decoder_set_metadata_respond(ptr noundef, i32 noundef)
 declare i32 @<!-- -->FLAC__stream_decoder_init_file(ptr noundef, ptr noundef, ptr noundef, ptr noundef, ptr noundef, ptr noundef) local_unnamed_addr #<!-- -->1
 
 ; Function Attrs: nounwind sspstrong uwtable
-define internal i32 @<!-- -->write_callback_(ptr nocapture readnone %decoder, ptr nocapture noundef readonly %frame, ptr nocapture noundef readonly %buffer, ptr nocapture noundef %client_data) #<!-- -->0 {
+define internal noundef i32 @<!-- -->write_callback_(ptr nocapture readnone %decoder, ptr nocapture noundef readonly %frame, ptr nocapture noundef readonly %buffer, ptr nocapture noundef %client_data) #<!-- -->0 {
 entry:
   %bits_per_sample1 = getelementptr inbounds %struct.FLAC__FrameHeader, ptr %frame, i64 0, i32 4
   %0 = load i32, ptr %bits_per_sample1, align 8
@@ -429,23 +429,16 @@ land.lhs.true14:                                  ; preds = %land.lhs.true11
   %cmp16 = icmp eq i32 %2, %8
   br i1 %cmp16, label %if.end, label %if.end.thread
 
-if.end.thread:                                    ; preds = %land.lhs.true, %land.lhs.true14, %land.lhs.true11, %land.lhs.true8, %entry
-  store i32 1, ptr %error, align 4
-  br label %9
-
 if.end:                                           ; preds = %land.lhs.true14
   %conv = zext i1 %cmp to i32
   %call = tail call i32 @<!-- -->grabbag__replaygain_analyze(ptr noundef %buffer, i32 noundef %conv, i32 noundef %0, i32 noundef %3), !range !14
-  %call.fr = freeze i32 %call
-  %lnot.ext = xor i32 %call.fr, 1
-  store i32 %lnot.ext, ptr %error, align 4
-  %tobool22.not = icmp ne i32 %lnot.ext, 0
-  %spec.select = zext i1 %tobool22.not to i32
-  br label %9
-
-9:                                                ; preds = %if.end, %if.end.thread
-  %10 = phi i32 [ 1, %if.end.thread ], [ %spec.select, %if.end ]
-  ret i32 %10
+  %lnot.ext = xor i32 %call, 1
+  br label %if.end.thread
+
+if.end.thread:                                    ; preds = %entry, %land.lhs.true8, %land.lhs.true11, %land.lhs.true14, %land.lhs.true, %if.end
+  %storemerge = phi i32 [ %lnot.ext, %if.end ], [ 1, %land.lhs.true ], [ 1, %land.lhs.true14 ], [ 1, %land.lhs.true11 ], [ 1, %land.lhs.true8 ], [ 1, %entry ]
+  store i32 %storemerge, ptr %error, align 4
+  ret i32 %storemerge
 }
 
 ; Function Attrs: nounwind sspstrong uwtable

Patch is 55.83 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/76553.diff

38 Files Affected:

  • (modified) clang/test/CodeGen/X86/ms-x86-intrinsics.c (+2-2)
  • (modified) clang/test/CodeGen/arm-bf16-params-returns.c (+4-4)
  • (modified) clang/test/CodeGen/arm-vector_type-params-returns.c (+12-12)
  • (modified) clang/test/CodeGen/ifunc.c (+4-4)
  • (modified) clang/test/CodeGen/isfpclass.c (+14-14)
  • (modified) clang/test/CodeGen/ms-mixed-ptr-sizes.c (+4-4)
  • (modified) clang/test/CodeGenCUDA/link-builtin-bitcode-denormal-fp-mode.cu (+8-8)
  • (modified) clang/test/CodeGenOpenCL/as_type.cl (+6-6)
  • (modified) llvm/lib/Transforms/IPO/FunctionAttrs.cpp (+39)
  • (modified) llvm/test/DebugInfo/X86/array2.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt2.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_check.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_promote.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_promote_legacy.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_pure_virtual_base.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_single_hybrid.ll (+2-2)
  • (modified) llvm/test/ThinLTO/X86/devirt_vcall_vis_hidden.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_vcall_vis_public.ll (+2-2)
  • (modified) llvm/test/ThinLTO/X86/funcimport.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/globals-import-const-fold.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/import-constant.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/index-const-prop-alias.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/index-const-prop.ll (+1-1)
  • (modified) llvm/test/Transforms/FunctionAttrs/nocapture.ll (+5-5)
  • (modified) llvm/test/Transforms/FunctionAttrs/nonnull.ll (+23-14)
  • (added) llvm/test/Transforms/FunctionAttrs/noundef.ll (+145)
  • (modified) llvm/test/Transforms/FunctionAttrs/out-of-bounds-iterator-bug.ll (+2-2)
  • (modified) llvm/test/Transforms/Inline/devirtualize-3.ll (+1-1)
  • (modified) llvm/test/Transforms/Inline/devirtualize-5.ll (+1-1)
  • (modified) llvm/test/Transforms/Inline/launder.invariant.group.ll (+2-2)
  • (modified) llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll (+11-11)
  • (modified) llvm/test/Transforms/PhaseOrdering/X86/merge-functions.ll (+1-1)
  • (modified) llvm/test/Transforms/PhaseOrdering/bitcast-store-branch.ll (+1-1)
  • (modified) llvm/test/Transforms/PhaseOrdering/dce-after-argument-promotion-loads.ll (+1-1)
  • (modified) llvm/test/Transforms/PhaseOrdering/early-arg-attrs-inference.ll (+1-1)
  • (modified) llvm/test/Transforms/PhaseOrdering/gep-null-compare-in-loop.ll (+2-2)
  • (modified) llvm/test/Transforms/SampleProfile/ctxsplit.ll (+6-6)
diff --git a/clang/test/CodeGen/X86/ms-x86-intrinsics.c b/clang/test/CodeGen/X86/ms-x86-intrinsics.c
index b4b8f5dc0b8c55..a1c90d71c8ebf5 100644
--- a/clang/test/CodeGen/X86/ms-x86-intrinsics.c
+++ b/clang/test/CodeGen/X86/ms-x86-intrinsics.c
@@ -145,7 +145,7 @@ unsigned __int64 test__shiftleft128(unsigned __int64 l, unsigned __int64 h,
                                     unsigned char d) {
   return __shiftleft128(l, h, d);
 }
-// CHECK-X64-LABEL: define dso_local i64 @test__shiftleft128(i64 noundef %l, i64 noundef %h, i8 noundef %d)
+// CHECK-X64-LABEL: define dso_local noundef i64 @test__shiftleft128(i64 noundef %l, i64 noundef %h, i8 noundef %d)
 // CHECK-X64: = zext i8 %{{.*}} to i64
 // CHECK-X64: = tail call i64 @llvm.fshl.i64(i64 %h, i64 %l, i64 %{{.*}})
 // CHECK-X64:  ret i64 %
@@ -154,7 +154,7 @@ unsigned __int64 test__shiftright128(unsigned __int64 l, unsigned __int64 h,
                                      unsigned char d) {
   return __shiftright128(l, h, d);
 }
-// CHECK-X64-LABEL: define dso_local i64 @test__shiftright128(i64 noundef %l, i64 noundef %h, i8 noundef %d)
+// CHECK-X64-LABEL: define dso_local noundef i64 @test__shiftright128(i64 noundef %l, i64 noundef %h, i8 noundef %d)
 // CHECK-X64: = zext i8 %{{.*}} to i64
 // CHECK-X64: = tail call i64 @llvm.fshr.i64(i64 %h, i64 %l, i64 %{{.*}})
 // CHECK-X64:  ret i64 %
diff --git a/clang/test/CodeGen/arm-bf16-params-returns.c b/clang/test/CodeGen/arm-bf16-params-returns.c
index 2d33e3f45eacf7..21be0bb151697a 100644
--- a/clang/test/CodeGen/arm-bf16-params-returns.c
+++ b/clang/test/CodeGen/arm-bf16-params-returns.c
@@ -11,7 +11,7 @@
 __bf16 test_ret_bf16(__bf16 v) {
   return v;
 }
-// CHECK32-HARD: define{{.*}} arm_aapcs_vfpcc bfloat @test_ret_bf16(bfloat noundef returned %v) {{.*}} {
+// CHECK32-HARD: define{{.*}} arm_aapcs_vfpcc noundef bfloat @test_ret_bf16(bfloat noundef returned %v) {{.*}} {
 // CHECK32-HARD: ret bfloat %v
 // CHECK32-SOFTFP: define{{.*}} bfloat @test_ret_bf16(bfloat noundef returned %v) {{.*}} {
 // CHECK32-SOFTFP: ret bfloat %v
@@ -23,11 +23,11 @@ __bf16 test_ret_bf16(__bf16 v) {
 bfloat16x4_t test_ret_bf16x4_t(bfloat16x4_t v) {
   return v;
 }
-// CHECK32-HARD: define{{.*}} arm_aapcs_vfpcc <4 x bfloat> @test_ret_bf16x4_t(<4 x bfloat> noundef returned %v) {{.*}} {
+// CHECK32-HARD: define{{.*}} arm_aapcs_vfpcc noundef <4 x bfloat> @test_ret_bf16x4_t(<4 x bfloat> noundef returned %v) {{.*}} {
 // CHECK32-HARD: ret <4 x bfloat> %v
-// CHECK32-SOFTFP: define{{.*}} <2 x i32> @test_ret_bf16x4_t(<2 x i32> [[V0:.*]]) {{.*}} {
+// CHECK32-SOFTFP: define{{.*}} noundef <2 x i32> @test_ret_bf16x4_t(<2 x i32> [[V0:.*]]) {{.*}} {
 // CHECK32-SOFTFP: ret <2 x i32> %v
-// CHECK64NEON: define{{.*}} <4 x bfloat> @test_ret_bf16x4_t(<4 x bfloat> noundef returned %v) {{.*}} {
+// CHECK64NEON: define{{.*}} noundef <4 x bfloat> @test_ret_bf16x4_t(<4 x bfloat> noundef returned %v) {{.*}} {
 // CHECK64NEON: ret <4 x bfloat> %v
 
 #endif
\ No newline at end of file
diff --git a/clang/test/CodeGen/arm-vector_type-params-returns.c b/clang/test/CodeGen/arm-vector_type-params-returns.c
index 14c3512ab81a9f..a55aba9ce06651 100644
--- a/clang/test/CodeGen/arm-vector_type-params-returns.c
+++ b/clang/test/CodeGen/arm-vector_type-params-returns.c
@@ -27,7 +27,7 @@
 #endif
 
 // function return types
-// CHECK-LABEL: define dso_local <8 x half> @test_ret_v8f16(
+// CHECK-LABEL: define dso_local noundef <8 x half> @test_ret_v8f16(
 // CHECK-SAME: <8 x half> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <8 x half> [[V]]
@@ -36,7 +36,7 @@ float16x8_t test_ret_v8f16(float16x8_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <4 x float> @test_ret_v4f32(
+// CHECK-LABEL: define dso_local noundef <4 x float> @test_ret_v4f32(
 // CHECK-SAME: <4 x float> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <4 x float> [[V]]
@@ -45,7 +45,7 @@ float32x4_t test_ret_v4f32(float32x4_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <2 x double> @test_ret_v2f64(
+// CHECK-LABEL: define dso_local noundef <2 x double> @test_ret_v2f64(
 // CHECK-SAME: <2 x double> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <2 x double> [[V]]
@@ -54,7 +54,7 @@ float64x2_t test_ret_v2f64(float64x2_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <8 x bfloat> @test_ret_v8bf16(
+// CHECK-LABEL: define dso_local noundef <8 x bfloat> @test_ret_v8bf16(
 // CHECK-SAME: <8 x bfloat> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <8 x bfloat> [[V]]
@@ -63,7 +63,7 @@ bfloat16x8_t test_ret_v8bf16(bfloat16x8_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <16 x i8> @test_ret_v16s8(
+// CHECK-LABEL: define dso_local noundef <16 x i8> @test_ret_v16s8(
 // CHECK-SAME: <16 x i8> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <16 x i8> [[V]]
@@ -72,7 +72,7 @@ int8x16_t test_ret_v16s8(int8x16_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <8 x i16> @test_ret_v8s16(
+// CHECK-LABEL: define dso_local noundef <8 x i16> @test_ret_v8s16(
 // CHECK-SAME: <8 x i16> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <8 x i16> [[V]]
@@ -81,7 +81,7 @@ int16x8_t test_ret_v8s16(int16x8_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <4 x i32> @test_ret_v32s4(
+// CHECK-LABEL: define dso_local noundef <4 x i32> @test_ret_v32s4(
 // CHECK-SAME: <4 x i32> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <4 x i32> [[V]]
@@ -90,7 +90,7 @@ int32x4_t test_ret_v32s4(int32x4_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <2 x i64> @test_ret_v64s2(
+// CHECK-LABEL: define dso_local noundef <2 x i64> @test_ret_v64s2(
 // CHECK-SAME: <2 x i64> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <2 x i64> [[V]]
@@ -99,7 +99,7 @@ int64x2_t test_ret_v64s2(int64x2_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <16 x i8> @test_ret_v16u8(
+// CHECK-LABEL: define dso_local noundef <16 x i8> @test_ret_v16u8(
 // CHECK-SAME: <16 x i8> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <16 x i8> [[V]]
@@ -108,7 +108,7 @@ uint8x16_t test_ret_v16u8(uint8x16_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <8 x i16> @test_ret_v8u16(
+// CHECK-LABEL: define dso_local noundef <8 x i16> @test_ret_v8u16(
 // CHECK-SAME: <8 x i16> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <8 x i16> [[V]]
@@ -117,7 +117,7 @@ uint16x8_t test_ret_v8u16(uint16x8_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <4 x i32> @test_ret_v32u4(
+// CHECK-LABEL: define dso_local noundef <4 x i32> @test_ret_v32u4(
 // CHECK-SAME: <4 x i32> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <4 x i32> [[V]]
@@ -126,7 +126,7 @@ uint32x4_t test_ret_v32u4(uint32x4_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <2 x i64> @test_ret_v64u2(
+// CHECK-LABEL: define dso_local noundef <2 x i64> @test_ret_v64u2(
 // CHECK-SAME: <2 x i64> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <2 x i64> [[V]]
diff --git a/clang/test/CodeGen/ifunc.c b/clang/test/CodeGen/ifunc.c
index a29b500e80bd50..3aa29f7dff74de 100644
--- a/clang/test/CodeGen/ifunc.c
+++ b/clang/test/CodeGen/ifunc.c
@@ -51,11 +51,11 @@ void* goo_ifunc(void) {
 // CHECK: call i32 @foo(i32
 // CHECK: call void @goo()
 
-// SAN: define internal nonnull ptr @foo_ifunc() #[[#FOO_IFUNC:]] {
-// MACSAN: define internal nonnull ptr @foo_ifunc() #[[#FOO_IFUNC:]] {
+// SAN: define internal nonnull {{(noundef )?}}ptr @foo_ifunc() #[[#FOO_IFUNC:]] {
+// MACSAN: define internal nonnull {{(noundef )?}}ptr @foo_ifunc() #[[#FOO_IFUNC:]] {
 
-// SAN: define dso_local noalias ptr @goo_ifunc() #[[#GOO_IFUNC:]] {
-// MACSAN: define noalias ptr @goo_ifunc() #[[#GOO_IFUNC:]] {
+// SAN: define dso_local noalias {{(noundef )?}}ptr @goo_ifunc() #[[#GOO_IFUNC:]] {
+// MACSAN: define noalias {{(noundef )?}}ptr @goo_ifunc() #[[#GOO_IFUNC:]] {
 
 // SAN-DAG: attributes #[[#FOO_IFUNC]] = {{{.*}} disable_sanitizer_instrumentation {{.*}}
 // MACSAN-DAG: attributes #[[#FOO_IFUNC]] = {{{.*}} disable_sanitizer_instrumentation {{.*}}
diff --git a/clang/test/CodeGen/isfpclass.c b/clang/test/CodeGen/isfpclass.c
index 34873c08e04f87..88d7a21b9733d8 100644
--- a/clang/test/CodeGen/isfpclass.c
+++ b/clang/test/CodeGen/isfpclass.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
 // RUN: %clang_cc1 -triple aarch64-linux-gnu -S -O1 -emit-llvm %s -o - | FileCheck %s
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_finite
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_finite
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call float @llvm.fabs.f32(float [[X]])
@@ -12,7 +12,7 @@ _Bool check_isfpclass_finite(float x) {
   return __builtin_isfpclass(x, 504 /*Finite*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_finite_strict
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_finite_strict
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2:[0-9]+]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 504) #[[ATTR6:[0-9]+]]
@@ -23,7 +23,7 @@ _Bool check_isfpclass_finite_strict(float x) {
   return __builtin_isfpclass(x, 504 /*Finite*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_nan_f32
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_nan_f32
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR3:[0-9]+]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = fcmp uno float [[X]], 0.000000e+00
@@ -33,7 +33,7 @@ _Bool check_isfpclass_nan_f32(float x) {
   return __builtin_isfpclass(x, 3 /*NaN*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_nan_f32_strict
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_nan_f32_strict
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 3) #[[ATTR6]]
@@ -44,7 +44,7 @@ _Bool check_isfpclass_nan_f32_strict(float x) {
   return __builtin_isfpclass(x, 3 /*NaN*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_snan_f64
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_snan_f64
 // CHECK-SAME: (double noundef [[X:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f64(double [[X]], i32 1)
@@ -54,7 +54,7 @@ _Bool check_isfpclass_snan_f64(double x) {
   return __builtin_isfpclass(x, 1 /*SNaN*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_snan_f64_strict
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_snan_f64_strict
 // CHECK-SAME: (double noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f64(double [[X]], i32 1) #[[ATTR6]]
@@ -65,7 +65,7 @@ _Bool check_isfpclass_snan_f64_strict(double x) {
   return __builtin_isfpclass(x, 1 /*NaN*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_zero_f16
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_zero_f16
 // CHECK-SAME: (half noundef [[X:%.*]]) local_unnamed_addr #[[ATTR3]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = fcmp oeq half [[X]], 0xH0000
@@ -75,7 +75,7 @@ _Bool check_isfpclass_zero_f16(_Float16 x) {
   return __builtin_isfpclass(x, 96 /*Zero*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_zero_f16_strict
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_zero_f16_strict
 // CHECK-SAME: (half noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f16(half [[X]], i32 96) #[[ATTR6]]
@@ -86,7 +86,7 @@ _Bool check_isfpclass_zero_f16_strict(_Float16 x) {
   return __builtin_isfpclass(x, 96 /*Zero*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isnan
+// CHECK-LABEL: define dso_local noundef i1 @check_isnan
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 3) #[[ATTR6]]
@@ -97,7 +97,7 @@ _Bool check_isnan(float x) {
   return __builtin_isnan(x);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isinf
+// CHECK-LABEL: define dso_local noundef i1 @check_isinf
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 516) #[[ATTR6]]
@@ -108,7 +108,7 @@ _Bool check_isinf(float x) {
   return __builtin_isinf(x);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfinite
+// CHECK-LABEL: define dso_local noundef i1 @check_isfinite
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 504) #[[ATTR6]]
@@ -119,7 +119,7 @@ _Bool check_isfinite(float x) {
   return __builtin_isfinite(x);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isnormal
+// CHECK-LABEL: define dso_local noundef i1 @check_isnormal
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 264) #[[ATTR6]]
@@ -136,7 +136,7 @@ typedef double __attribute__((ext_vector_type(4))) double4;
 typedef int __attribute__((ext_vector_type(4))) int4;
 typedef long __attribute__((ext_vector_type(4))) long4;
 
-// CHECK-LABEL: define dso_local <4 x i32> @check_isfpclass_nan_v4f32
+// CHECK-LABEL: define dso_local noundef <4 x i32> @check_isfpclass_nan_v4f32
 // CHECK-SAME: (<4 x float> noundef [[X:%.*]]) local_unnamed_addr #[[ATTR3]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = fcmp uno <4 x float> [[X]], zeroinitializer
@@ -147,7 +147,7 @@ int4 check_isfpclass_nan_v4f32(float4 x) {
   return __builtin_isfpclass(x, 3 /*NaN*/);
 }
 
-// CHECK-LABEL: define dso_local <4 x i32> @check_isfpclass_nan_strict_v4f32
+// CHECK-LABEL: define dso_local noundef <4 x i32> @check_isfpclass_nan_strict_v4f32
 // CHECK-SAME: (<4 x float> noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <4 x i1> @llvm.is.fpclass.v4f32(<4 x float> [[X]], i32 3) #[[ATTR6]]
diff --git a/clang/test/CodeGen/ms-mixed-ptr-sizes.c b/clang/test/CodeGen/ms-mixed-ptr-sizes.c
index a853c6083cb8ba..89d05fd30b72c2 100644
--- a/clang/test/CodeGen/ms-mixed-ptr-sizes.c
+++ b/clang/test/CodeGen/ms-mixed-ptr-sizes.c
@@ -49,7 +49,7 @@ void test_other(struct Foo *f, __attribute__((address_space(10))) int *i) {
 }
 
 int test_compare1(int *__ptr32 __uptr i, int *__ptr64 j) {
-  // ALL-LABEL: define dso_local i32 @test_compare1
+  // ALL-LABEL: define dso_local noundef i32 @test_compare1
   // X64: %{{.+}} = addrspacecast ptr %j to ptr addrspace(271)
   // X64: %cmp = icmp eq ptr addrspace(271) %{{.+}}, %i
   // X86: %{{.+}} = addrspacecast ptr addrspace(272) %j to ptr addrspace(271)
@@ -58,7 +58,7 @@ int test_compare1(int *__ptr32 __uptr i, int *__ptr64 j) {
 }
 
 int test_compare2(int *__ptr32 __sptr i, int *__ptr64 j) {
-  // ALL-LABEL: define dso_local i32 @test_compare2
+  // ALL-LABEL: define dso_local noundef i32 @test_compare2
   // X64: %{{.+}} = addrspacecast ptr %j to ptr addrspace(270)
   // X64: %cmp = icmp eq ptr addrspace(270) %{{.+}}, %i
   // X86: %{{.+}} = addrspacecast ptr addrspace(272) %j to ptr
@@ -67,7 +67,7 @@ int test_compare2(int *__ptr32 __sptr i, int *__ptr64 j) {
 }
 
 int test_compare3(int *__ptr32 __uptr i, int *__ptr64 j) {
-  // ALL-LABEL: define dso_local i32 @test_compare3
+  // ALL-LABEL: define dso_local noundef i32 @test_compare3
   // X64: %{{.+}} = addrspacecast ptr addrspace(271) %i to ptr
   // X64: %cmp = icmp eq ptr %{{.+}}, %j
   // X86: %{{.+}} = addrspacecast ptr addrspace(271) %i to ptr addrspace(272)
@@ -76,7 +76,7 @@ int test_compare3(int *__ptr32 __uptr i, int *__ptr64 j) {
 }
 
 int test_compare4(int *__ptr32 __sptr i, int *__ptr64 j) {
-  // ALL-LABEL: define dso_local i32 @test_compare4
+  // ALL-LABEL: define dso_local noundef i32 @test_compare4
   // X64: %{{.+}} = addrspacecast ptr addrspace(270) %i to ptr
   // X64: %cmp = icmp eq ptr %{{.+}}, %j
   // X86: %{{.+}} = addrspacecast ptr %i to ptr addrspace(272)
diff --git a/clang/test/CodeGenCUDA/link-builtin-bitcode-denormal-fp-mode.cu b/clang/test/CodeGenCUDA/link-builtin-bitcode-denormal-fp-mode.cu
index f12eacfb53fab3..ef02668c3697bf 100644
--- a/clang/test/CodeGenCUDA/link-builtin-bitcode-denormal-fp-mode.cu
+++ b/clang/test/CodeGenCUDA/link-builtin-bitcode-denormal-fp-mode.cu
@@ -111,16 +111,16 @@ __global__ void kernel_f64(double* out, double* a, double* b, double* c) {
 }
 }
 
-// INTERNALIZE: define internal half @do_f16_stuff({{.*}}) #[[$FUNCATTR:[0-9]+]]
-// INTERNALIZE: define internal float @do_f32_stuff({{.*}}) #[[$FUNCATTR]]
-// INTERNALIZE: define internal double @do_f64_stuff({{.*}}) #[[$FUNCATTR]]
-// INTERNALIZE: define internal float @weak_do_f32_stuff({{.*}}) #[[$WEAK_FUNCATTR:[0-9]+]]
+// INTERNALIZE: define internal {{(noundef )?}}half @do_f16_stuff({{.*}}) #[[$FUNCATTR:[0-9]+]]
+// INTERNALIZE: define internal {{(noundef )?}}float @do_f32_stuff({{.*}}) #[[$FUNCATTR]]
+// INTERNALIZE: define internal {{(noundef )?}}double @do_f64_stuff({{.*}}) #[[$FUNCATTR]]
+// INTERNALIZE: define internal {{(noundef )?}}float @weak_do_f32_stuff({{.*}}) #[[$WEAK_FUNCATTR:[0-9]+]]
 
 
-// NOINTERNALIZE: define dso_local half @do_f16_stuff({{.*}}) #[[$FUNCATTR:[0-9]+]]
-// NOINTERNALIZE: define dso_local float @do_f32_stuff({{.*}}) #[[$FUNCATTR]]
-// NOINTERNALIZE: define dso_local double @do_f64_stuff({{.*}}) #[[$FUNCATTR]]
-// NOINTERNALIZE: define weak float @weak_do_f32_stuff({{.*}}) #[[$WEAK_FUNCATTR:[0-9]+]]
+// NOINTERNALIZE: define dso_local {{(noundef )?}}half @do_f16_stuff({{.*}}) #[[$FUNCATTR:[0-9]+]]
+// NOINTERNALIZE: define dso_local {{(noundef )?}}float @do_f32_stuff({{.*}}) #[[$FUNCATTR]]
+// NOINTERNALIZE: define dso_local {{(noundef )?}}double @do_f64_stuff({{.*}}) #[[$FUNCATTR]]
+// NOINTERNALIZE: define weak {{(noundef )?}}float @weak_do_f32_stuff({{.*}}) #[[$WEAK_FUNCATTR:[0-9]+]]
 
 
 
diff --git a/clang/test/CodeGenOpenCL/as_type.cl b/clang/test/CodeGenOpenCL/as_type.cl
index d8e75db936f732..1fe26fbeafdb4b 100644
--- a/clang/test/CodeGenOpenCL/as_type.cl
+++ b/clang/test/CodeGenOpenCL/as_type.cl
@@ -27,7 +27,7 @@ char3 f3(int x) {
   return __builtin_astype(x, char3);
 }
 
-//CHECK: define{{.*}} spir_func <4 x i8> @f4(i32 noundef %[[x:.*]])
+//CHECK: define{{.*}} spir_func noundef <4 x i8> @f4(i32 noundef %[[x:.*]])
 //CHECK: %[[astype:.*]] = bitcast i32 %[[x]] to <4 x i8>
 //CHECK-NOT: shufflevector
 //CHECK: ret <4 x i8> %[[astype]]
@@ -43,7 +43,7 @@ int f5(char3 x) {
   return __builtin_astype(x, int);
 }
 
-//CHECK: define{{.*}} spir_func i32 @f6(<4 x i8> noundef %[[x:.*]])
+//CHECK: define{{.*}} spir_func noundef i32 @f6(<4 x i8> noundef %[[x:.*]])
 //CHECK: %[[astype:.*]] = bitcast <4 x i8> %[[x]] to i32
 //CHECK-NOT: shufflevector
 //CHECK: ret i32 %[[astype]]
@@ -51,7 +51,7 @@ int f6(char4 x) {
   return __builtin_astype(x, int);
 }
 
-//CHECK: define{{.*}} spir_func <3 x i8> @f7(<3 x i8> noundef returned %[[x:.*]])
+//CHECK: define{{.*}} spir_func noundef <3 x i8> @f7(<3 x i8> noundef returned %[[x:.*]])
 //CHECK-NOT: bitcast
 //CHECK-NOT: shufflevector
 //CHECK: ret <3 x i8> %[[x]]
@@ -67,21 +67,21 @@ int3 f8(char16 x) {
   return __builtin_astype(x, int3);
 }
 
-//CHECK: define{{.*}} spir_func ptr addrspace(1) @addr_cast(ptr noundef readnone %[[x:.*]])
+//CHECK: define{{.*}} spir_func noundef ptr a...
[truncated]

@llvmbot
Copy link
Member

llvmbot commented Dec 29, 2023

@llvm/pr-subscribers-backend-x86

Author: Yingwei Zheng (dtcxzyw)

Changes

This patch deduces noundef attributes for return values.
IIUC, a function returns noundef values iff all of its return values are guaranteed not to be undef or poison.
Definition of noundef from LangRef:

noundef
This attribute applies to parameters and return values. If the value representation contains any 
undefined or poison bits, the behavior is undefined. Note that this does not refer to padding 
introduced by the type’s storage representation.

Alive2: https://alive2.llvm.org/ce/z/g8Eis6

Compile-time impact: http://llvm-compile-time-tracker.com/compare.php?from=7f69c8b3a6c02ea32fefb16c2016dfa1ba994858&amp;to=1dafc281ff8c04bb0a968fb3d898f08876dc59e0&amp;stat=instructions:u

stage1-O3 stage1-ReleaseThinLTO stage1-ReleaseLTO-g stage1-O0-g stage2-O3 stage2-O0-g stage2-clang
+0.01% +0.01% -0.01% +0.00% +0.03% +0.02% +0.01%

The motivation of this patch is to reduce the number of freeze insts and enable more optimizations.
Example:

diff --git a/bench/flac/optimized/replaygain.c.ll b/bench/flac/optimized/replaygain.c.ll
index fa826475..413bd717 100644
--- a/bench/flac/optimized/replaygain.c.ll
+++ b/bench/flac/optimized/replaygain.c.ll
@@ -63,7 +63,7 @@ entry:
 declare i32 @<!-- -->InitGainAnalysis(i64 noundef) local_unnamed_addr #<!-- -->1
 
 ; Function Attrs: nounwind sspstrong uwtable
-define dso_local i32 @<!-- -->grabbag__replaygain_analyze(ptr nocapture noundef readonly %input, i32 noundef %is_stereo, i32 noundef %bps, i32 noundef %samples) local_unnamed_addr #<!-- -->0 {
+define dso_local noundef i32 @<!-- -->grabbag__replaygain_analyze(ptr nocapture noundef readonly %input, i32 noundef %is_stereo, i32 noundef %bps, i32 noundef %samples) local_unnamed_addr #<!-- -->0 {
 entry:
   %cmp = icmp eq i32 %bps, 16
   br i1 %cmp, label %if.then, label %if.else71
@@ -337,7 +337,7 @@ entry:
 declare float @<!-- -->GetTitleGain() local_unnamed_addr #<!-- -->1
 
 ; Function Attrs: nounwind sspstrong uwtable
-define dso_local ptr @<!-- -->grabbag__replaygain_analyze_file(ptr noundef %filename, ptr nocapture noundef writeonly %title_gain, ptr nocapture noundef writeonly %title_peak) local_unnamed_addr #<!-- -->0 {
+define dso_local noundef ptr @<!-- -->grabbag__replaygain_analyze_file(ptr noundef %filename, ptr nocapture noundef writeonly %title_gain, ptr nocapture noundef writeonly %title_peak) local_unnamed_addr #<!-- -->0 {
 entry:
   %instance = alloca %struct.DecoderInstance, align 4
   %call = tail call ptr @<!-- -->FLAC__stream_decoder_new() #<!-- -->15
@@ -392,7 +392,7 @@ declare i32 @<!-- -->FLAC__stream_decoder_set_metadata_respond(ptr noundef, i32 noundef)
 declare i32 @<!-- -->FLAC__stream_decoder_init_file(ptr noundef, ptr noundef, ptr noundef, ptr noundef, ptr noundef, ptr noundef) local_unnamed_addr #<!-- -->1
 
 ; Function Attrs: nounwind sspstrong uwtable
-define internal i32 @<!-- -->write_callback_(ptr nocapture readnone %decoder, ptr nocapture noundef readonly %frame, ptr nocapture noundef readonly %buffer, ptr nocapture noundef %client_data) #<!-- -->0 {
+define internal noundef i32 @<!-- -->write_callback_(ptr nocapture readnone %decoder, ptr nocapture noundef readonly %frame, ptr nocapture noundef readonly %buffer, ptr nocapture noundef %client_data) #<!-- -->0 {
 entry:
   %bits_per_sample1 = getelementptr inbounds %struct.FLAC__FrameHeader, ptr %frame, i64 0, i32 4
   %0 = load i32, ptr %bits_per_sample1, align 8
@@ -429,23 +429,16 @@ land.lhs.true14:                                  ; preds = %land.lhs.true11
   %cmp16 = icmp eq i32 %2, %8
   br i1 %cmp16, label %if.end, label %if.end.thread
 
-if.end.thread:                                    ; preds = %land.lhs.true, %land.lhs.true14, %land.lhs.true11, %land.lhs.true8, %entry
-  store i32 1, ptr %error, align 4
-  br label %9
-
 if.end:                                           ; preds = %land.lhs.true14
   %conv = zext i1 %cmp to i32
   %call = tail call i32 @<!-- -->grabbag__replaygain_analyze(ptr noundef %buffer, i32 noundef %conv, i32 noundef %0, i32 noundef %3), !range !14
-  %call.fr = freeze i32 %call
-  %lnot.ext = xor i32 %call.fr, 1
-  store i32 %lnot.ext, ptr %error, align 4
-  %tobool22.not = icmp ne i32 %lnot.ext, 0
-  %spec.select = zext i1 %tobool22.not to i32
-  br label %9
-
-9:                                                ; preds = %if.end, %if.end.thread
-  %10 = phi i32 [ 1, %if.end.thread ], [ %spec.select, %if.end ]
-  ret i32 %10
+  %lnot.ext = xor i32 %call, 1
+  br label %if.end.thread
+
+if.end.thread:                                    ; preds = %entry, %land.lhs.true8, %land.lhs.true11, %land.lhs.true14, %land.lhs.true, %if.end
+  %storemerge = phi i32 [ %lnot.ext, %if.end ], [ 1, %land.lhs.true ], [ 1, %land.lhs.true14 ], [ 1, %land.lhs.true11 ], [ 1, %land.lhs.true8 ], [ 1, %entry ]
+  store i32 %storemerge, ptr %error, align 4
+  ret i32 %storemerge
 }
 
 ; Function Attrs: nounwind sspstrong uwtable

Patch is 55.83 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/76553.diff

38 Files Affected:

  • (modified) clang/test/CodeGen/X86/ms-x86-intrinsics.c (+2-2)
  • (modified) clang/test/CodeGen/arm-bf16-params-returns.c (+4-4)
  • (modified) clang/test/CodeGen/arm-vector_type-params-returns.c (+12-12)
  • (modified) clang/test/CodeGen/ifunc.c (+4-4)
  • (modified) clang/test/CodeGen/isfpclass.c (+14-14)
  • (modified) clang/test/CodeGen/ms-mixed-ptr-sizes.c (+4-4)
  • (modified) clang/test/CodeGenCUDA/link-builtin-bitcode-denormal-fp-mode.cu (+8-8)
  • (modified) clang/test/CodeGenOpenCL/as_type.cl (+6-6)
  • (modified) llvm/lib/Transforms/IPO/FunctionAttrs.cpp (+39)
  • (modified) llvm/test/DebugInfo/X86/array2.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt2.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_check.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_promote.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_promote_legacy.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_pure_virtual_base.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_single_hybrid.ll (+2-2)
  • (modified) llvm/test/ThinLTO/X86/devirt_vcall_vis_hidden.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/devirt_vcall_vis_public.ll (+2-2)
  • (modified) llvm/test/ThinLTO/X86/funcimport.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/globals-import-const-fold.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/import-constant.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/index-const-prop-alias.ll (+1-1)
  • (modified) llvm/test/ThinLTO/X86/index-const-prop.ll (+1-1)
  • (modified) llvm/test/Transforms/FunctionAttrs/nocapture.ll (+5-5)
  • (modified) llvm/test/Transforms/FunctionAttrs/nonnull.ll (+23-14)
  • (added) llvm/test/Transforms/FunctionAttrs/noundef.ll (+145)
  • (modified) llvm/test/Transforms/FunctionAttrs/out-of-bounds-iterator-bug.ll (+2-2)
  • (modified) llvm/test/Transforms/Inline/devirtualize-3.ll (+1-1)
  • (modified) llvm/test/Transforms/Inline/devirtualize-5.ll (+1-1)
  • (modified) llvm/test/Transforms/Inline/launder.invariant.group.ll (+2-2)
  • (modified) llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll (+11-11)
  • (modified) llvm/test/Transforms/PhaseOrdering/X86/merge-functions.ll (+1-1)
  • (modified) llvm/test/Transforms/PhaseOrdering/bitcast-store-branch.ll (+1-1)
  • (modified) llvm/test/Transforms/PhaseOrdering/dce-after-argument-promotion-loads.ll (+1-1)
  • (modified) llvm/test/Transforms/PhaseOrdering/early-arg-attrs-inference.ll (+1-1)
  • (modified) llvm/test/Transforms/PhaseOrdering/gep-null-compare-in-loop.ll (+2-2)
  • (modified) llvm/test/Transforms/SampleProfile/ctxsplit.ll (+6-6)
diff --git a/clang/test/CodeGen/X86/ms-x86-intrinsics.c b/clang/test/CodeGen/X86/ms-x86-intrinsics.c
index b4b8f5dc0b8c55..a1c90d71c8ebf5 100644
--- a/clang/test/CodeGen/X86/ms-x86-intrinsics.c
+++ b/clang/test/CodeGen/X86/ms-x86-intrinsics.c
@@ -145,7 +145,7 @@ unsigned __int64 test__shiftleft128(unsigned __int64 l, unsigned __int64 h,
                                     unsigned char d) {
   return __shiftleft128(l, h, d);
 }
-// CHECK-X64-LABEL: define dso_local i64 @test__shiftleft128(i64 noundef %l, i64 noundef %h, i8 noundef %d)
+// CHECK-X64-LABEL: define dso_local noundef i64 @test__shiftleft128(i64 noundef %l, i64 noundef %h, i8 noundef %d)
 // CHECK-X64: = zext i8 %{{.*}} to i64
 // CHECK-X64: = tail call i64 @llvm.fshl.i64(i64 %h, i64 %l, i64 %{{.*}})
 // CHECK-X64:  ret i64 %
@@ -154,7 +154,7 @@ unsigned __int64 test__shiftright128(unsigned __int64 l, unsigned __int64 h,
                                      unsigned char d) {
   return __shiftright128(l, h, d);
 }
-// CHECK-X64-LABEL: define dso_local i64 @test__shiftright128(i64 noundef %l, i64 noundef %h, i8 noundef %d)
+// CHECK-X64-LABEL: define dso_local noundef i64 @test__shiftright128(i64 noundef %l, i64 noundef %h, i8 noundef %d)
 // CHECK-X64: = zext i8 %{{.*}} to i64
 // CHECK-X64: = tail call i64 @llvm.fshr.i64(i64 %h, i64 %l, i64 %{{.*}})
 // CHECK-X64:  ret i64 %
diff --git a/clang/test/CodeGen/arm-bf16-params-returns.c b/clang/test/CodeGen/arm-bf16-params-returns.c
index 2d33e3f45eacf7..21be0bb151697a 100644
--- a/clang/test/CodeGen/arm-bf16-params-returns.c
+++ b/clang/test/CodeGen/arm-bf16-params-returns.c
@@ -11,7 +11,7 @@
 __bf16 test_ret_bf16(__bf16 v) {
   return v;
 }
-// CHECK32-HARD: define{{.*}} arm_aapcs_vfpcc bfloat @test_ret_bf16(bfloat noundef returned %v) {{.*}} {
+// CHECK32-HARD: define{{.*}} arm_aapcs_vfpcc noundef bfloat @test_ret_bf16(bfloat noundef returned %v) {{.*}} {
 // CHECK32-HARD: ret bfloat %v
 // CHECK32-SOFTFP: define{{.*}} bfloat @test_ret_bf16(bfloat noundef returned %v) {{.*}} {
 // CHECK32-SOFTFP: ret bfloat %v
@@ -23,11 +23,11 @@ __bf16 test_ret_bf16(__bf16 v) {
 bfloat16x4_t test_ret_bf16x4_t(bfloat16x4_t v) {
   return v;
 }
-// CHECK32-HARD: define{{.*}} arm_aapcs_vfpcc <4 x bfloat> @test_ret_bf16x4_t(<4 x bfloat> noundef returned %v) {{.*}} {
+// CHECK32-HARD: define{{.*}} arm_aapcs_vfpcc noundef <4 x bfloat> @test_ret_bf16x4_t(<4 x bfloat> noundef returned %v) {{.*}} {
 // CHECK32-HARD: ret <4 x bfloat> %v
-// CHECK32-SOFTFP: define{{.*}} <2 x i32> @test_ret_bf16x4_t(<2 x i32> [[V0:.*]]) {{.*}} {
+// CHECK32-SOFTFP: define{{.*}} noundef <2 x i32> @test_ret_bf16x4_t(<2 x i32> [[V0:.*]]) {{.*}} {
 // CHECK32-SOFTFP: ret <2 x i32> %v
-// CHECK64NEON: define{{.*}} <4 x bfloat> @test_ret_bf16x4_t(<4 x bfloat> noundef returned %v) {{.*}} {
+// CHECK64NEON: define{{.*}} noundef <4 x bfloat> @test_ret_bf16x4_t(<4 x bfloat> noundef returned %v) {{.*}} {
 // CHECK64NEON: ret <4 x bfloat> %v
 
 #endif
\ No newline at end of file
diff --git a/clang/test/CodeGen/arm-vector_type-params-returns.c b/clang/test/CodeGen/arm-vector_type-params-returns.c
index 14c3512ab81a9f..a55aba9ce06651 100644
--- a/clang/test/CodeGen/arm-vector_type-params-returns.c
+++ b/clang/test/CodeGen/arm-vector_type-params-returns.c
@@ -27,7 +27,7 @@
 #endif
 
 // function return types
-// CHECK-LABEL: define dso_local <8 x half> @test_ret_v8f16(
+// CHECK-LABEL: define dso_local noundef <8 x half> @test_ret_v8f16(
 // CHECK-SAME: <8 x half> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <8 x half> [[V]]
@@ -36,7 +36,7 @@ float16x8_t test_ret_v8f16(float16x8_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <4 x float> @test_ret_v4f32(
+// CHECK-LABEL: define dso_local noundef <4 x float> @test_ret_v4f32(
 // CHECK-SAME: <4 x float> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <4 x float> [[V]]
@@ -45,7 +45,7 @@ float32x4_t test_ret_v4f32(float32x4_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <2 x double> @test_ret_v2f64(
+// CHECK-LABEL: define dso_local noundef <2 x double> @test_ret_v2f64(
 // CHECK-SAME: <2 x double> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <2 x double> [[V]]
@@ -54,7 +54,7 @@ float64x2_t test_ret_v2f64(float64x2_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <8 x bfloat> @test_ret_v8bf16(
+// CHECK-LABEL: define dso_local noundef <8 x bfloat> @test_ret_v8bf16(
 // CHECK-SAME: <8 x bfloat> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <8 x bfloat> [[V]]
@@ -63,7 +63,7 @@ bfloat16x8_t test_ret_v8bf16(bfloat16x8_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <16 x i8> @test_ret_v16s8(
+// CHECK-LABEL: define dso_local noundef <16 x i8> @test_ret_v16s8(
 // CHECK-SAME: <16 x i8> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <16 x i8> [[V]]
@@ -72,7 +72,7 @@ int8x16_t test_ret_v16s8(int8x16_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <8 x i16> @test_ret_v8s16(
+// CHECK-LABEL: define dso_local noundef <8 x i16> @test_ret_v8s16(
 // CHECK-SAME: <8 x i16> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <8 x i16> [[V]]
@@ -81,7 +81,7 @@ int16x8_t test_ret_v8s16(int16x8_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <4 x i32> @test_ret_v32s4(
+// CHECK-LABEL: define dso_local noundef <4 x i32> @test_ret_v32s4(
 // CHECK-SAME: <4 x i32> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <4 x i32> [[V]]
@@ -90,7 +90,7 @@ int32x4_t test_ret_v32s4(int32x4_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <2 x i64> @test_ret_v64s2(
+// CHECK-LABEL: define dso_local noundef <2 x i64> @test_ret_v64s2(
 // CHECK-SAME: <2 x i64> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <2 x i64> [[V]]
@@ -99,7 +99,7 @@ int64x2_t test_ret_v64s2(int64x2_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <16 x i8> @test_ret_v16u8(
+// CHECK-LABEL: define dso_local noundef <16 x i8> @test_ret_v16u8(
 // CHECK-SAME: <16 x i8> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <16 x i8> [[V]]
@@ -108,7 +108,7 @@ uint8x16_t test_ret_v16u8(uint8x16_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <8 x i16> @test_ret_v8u16(
+// CHECK-LABEL: define dso_local noundef <8 x i16> @test_ret_v8u16(
 // CHECK-SAME: <8 x i16> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <8 x i16> [[V]]
@@ -117,7 +117,7 @@ uint16x8_t test_ret_v8u16(uint16x8_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <4 x i32> @test_ret_v32u4(
+// CHECK-LABEL: define dso_local noundef <4 x i32> @test_ret_v32u4(
 // CHECK-SAME: <4 x i32> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <4 x i32> [[V]]
@@ -126,7 +126,7 @@ uint32x4_t test_ret_v32u4(uint32x4_t v) {
   return v;
 }
 
-// CHECK-LABEL: define dso_local <2 x i64> @test_ret_v64u2(
+// CHECK-LABEL: define dso_local noundef <2 x i64> @test_ret_v64u2(
 // CHECK-SAME: <2 x i64> noundef returned [[V:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    ret <2 x i64> [[V]]
diff --git a/clang/test/CodeGen/ifunc.c b/clang/test/CodeGen/ifunc.c
index a29b500e80bd50..3aa29f7dff74de 100644
--- a/clang/test/CodeGen/ifunc.c
+++ b/clang/test/CodeGen/ifunc.c
@@ -51,11 +51,11 @@ void* goo_ifunc(void) {
 // CHECK: call i32 @foo(i32
 // CHECK: call void @goo()
 
-// SAN: define internal nonnull ptr @foo_ifunc() #[[#FOO_IFUNC:]] {
-// MACSAN: define internal nonnull ptr @foo_ifunc() #[[#FOO_IFUNC:]] {
+// SAN: define internal nonnull {{(noundef )?}}ptr @foo_ifunc() #[[#FOO_IFUNC:]] {
+// MACSAN: define internal nonnull {{(noundef )?}}ptr @foo_ifunc() #[[#FOO_IFUNC:]] {
 
-// SAN: define dso_local noalias ptr @goo_ifunc() #[[#GOO_IFUNC:]] {
-// MACSAN: define noalias ptr @goo_ifunc() #[[#GOO_IFUNC:]] {
+// SAN: define dso_local noalias {{(noundef )?}}ptr @goo_ifunc() #[[#GOO_IFUNC:]] {
+// MACSAN: define noalias {{(noundef )?}}ptr @goo_ifunc() #[[#GOO_IFUNC:]] {
 
 // SAN-DAG: attributes #[[#FOO_IFUNC]] = {{{.*}} disable_sanitizer_instrumentation {{.*}}
 // MACSAN-DAG: attributes #[[#FOO_IFUNC]] = {{{.*}} disable_sanitizer_instrumentation {{.*}}
diff --git a/clang/test/CodeGen/isfpclass.c b/clang/test/CodeGen/isfpclass.c
index 34873c08e04f87..88d7a21b9733d8 100644
--- a/clang/test/CodeGen/isfpclass.c
+++ b/clang/test/CodeGen/isfpclass.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 2
 // RUN: %clang_cc1 -triple aarch64-linux-gnu -S -O1 -emit-llvm %s -o - | FileCheck %s
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_finite
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_finite
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call float @llvm.fabs.f32(float [[X]])
@@ -12,7 +12,7 @@ _Bool check_isfpclass_finite(float x) {
   return __builtin_isfpclass(x, 504 /*Finite*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_finite_strict
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_finite_strict
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2:[0-9]+]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 504) #[[ATTR6:[0-9]+]]
@@ -23,7 +23,7 @@ _Bool check_isfpclass_finite_strict(float x) {
   return __builtin_isfpclass(x, 504 /*Finite*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_nan_f32
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_nan_f32
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR3:[0-9]+]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = fcmp uno float [[X]], 0.000000e+00
@@ -33,7 +33,7 @@ _Bool check_isfpclass_nan_f32(float x) {
   return __builtin_isfpclass(x, 3 /*NaN*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_nan_f32_strict
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_nan_f32_strict
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 3) #[[ATTR6]]
@@ -44,7 +44,7 @@ _Bool check_isfpclass_nan_f32_strict(float x) {
   return __builtin_isfpclass(x, 3 /*NaN*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_snan_f64
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_snan_f64
 // CHECK-SAME: (double noundef [[X:%.*]]) local_unnamed_addr #[[ATTR0]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f64(double [[X]], i32 1)
@@ -54,7 +54,7 @@ _Bool check_isfpclass_snan_f64(double x) {
   return __builtin_isfpclass(x, 1 /*SNaN*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_snan_f64_strict
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_snan_f64_strict
 // CHECK-SAME: (double noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f64(double [[X]], i32 1) #[[ATTR6]]
@@ -65,7 +65,7 @@ _Bool check_isfpclass_snan_f64_strict(double x) {
   return __builtin_isfpclass(x, 1 /*NaN*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_zero_f16
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_zero_f16
 // CHECK-SAME: (half noundef [[X:%.*]]) local_unnamed_addr #[[ATTR3]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = fcmp oeq half [[X]], 0xH0000
@@ -75,7 +75,7 @@ _Bool check_isfpclass_zero_f16(_Float16 x) {
   return __builtin_isfpclass(x, 96 /*Zero*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfpclass_zero_f16_strict
+// CHECK-LABEL: define dso_local noundef i1 @check_isfpclass_zero_f16_strict
 // CHECK-SAME: (half noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f16(half [[X]], i32 96) #[[ATTR6]]
@@ -86,7 +86,7 @@ _Bool check_isfpclass_zero_f16_strict(_Float16 x) {
   return __builtin_isfpclass(x, 96 /*Zero*/);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isnan
+// CHECK-LABEL: define dso_local noundef i1 @check_isnan
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 3) #[[ATTR6]]
@@ -97,7 +97,7 @@ _Bool check_isnan(float x) {
   return __builtin_isnan(x);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isinf
+// CHECK-LABEL: define dso_local noundef i1 @check_isinf
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 516) #[[ATTR6]]
@@ -108,7 +108,7 @@ _Bool check_isinf(float x) {
   return __builtin_isinf(x);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isfinite
+// CHECK-LABEL: define dso_local noundef i1 @check_isfinite
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 504) #[[ATTR6]]
@@ -119,7 +119,7 @@ _Bool check_isfinite(float x) {
   return __builtin_isfinite(x);
 }
 
-// CHECK-LABEL: define dso_local i1 @check_isnormal
+// CHECK-LABEL: define dso_local noundef i1 @check_isnormal
 // CHECK-SAME: (float noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call i1 @llvm.is.fpclass.f32(float [[X]], i32 264) #[[ATTR6]]
@@ -136,7 +136,7 @@ typedef double __attribute__((ext_vector_type(4))) double4;
 typedef int __attribute__((ext_vector_type(4))) int4;
 typedef long __attribute__((ext_vector_type(4))) long4;
 
-// CHECK-LABEL: define dso_local <4 x i32> @check_isfpclass_nan_v4f32
+// CHECK-LABEL: define dso_local noundef <4 x i32> @check_isfpclass_nan_v4f32
 // CHECK-SAME: (<4 x float> noundef [[X:%.*]]) local_unnamed_addr #[[ATTR3]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = fcmp uno <4 x float> [[X]], zeroinitializer
@@ -147,7 +147,7 @@ int4 check_isfpclass_nan_v4f32(float4 x) {
   return __builtin_isfpclass(x, 3 /*NaN*/);
 }
 
-// CHECK-LABEL: define dso_local <4 x i32> @check_isfpclass_nan_strict_v4f32
+// CHECK-LABEL: define dso_local noundef <4 x i32> @check_isfpclass_nan_strict_v4f32
 // CHECK-SAME: (<4 x float> noundef [[X:%.*]]) local_unnamed_addr #[[ATTR2]] {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <4 x i1> @llvm.is.fpclass.v4f32(<4 x float> [[X]], i32 3) #[[ATTR6]]
diff --git a/clang/test/CodeGen/ms-mixed-ptr-sizes.c b/clang/test/CodeGen/ms-mixed-ptr-sizes.c
index a853c6083cb8ba..89d05fd30b72c2 100644
--- a/clang/test/CodeGen/ms-mixed-ptr-sizes.c
+++ b/clang/test/CodeGen/ms-mixed-ptr-sizes.c
@@ -49,7 +49,7 @@ void test_other(struct Foo *f, __attribute__((address_space(10))) int *i) {
 }
 
 int test_compare1(int *__ptr32 __uptr i, int *__ptr64 j) {
-  // ALL-LABEL: define dso_local i32 @test_compare1
+  // ALL-LABEL: define dso_local noundef i32 @test_compare1
   // X64: %{{.+}} = addrspacecast ptr %j to ptr addrspace(271)
   // X64: %cmp = icmp eq ptr addrspace(271) %{{.+}}, %i
   // X86: %{{.+}} = addrspacecast ptr addrspace(272) %j to ptr addrspace(271)
@@ -58,7 +58,7 @@ int test_compare1(int *__ptr32 __uptr i, int *__ptr64 j) {
 }
 
 int test_compare2(int *__ptr32 __sptr i, int *__ptr64 j) {
-  // ALL-LABEL: define dso_local i32 @test_compare2
+  // ALL-LABEL: define dso_local noundef i32 @test_compare2
   // X64: %{{.+}} = addrspacecast ptr %j to ptr addrspace(270)
   // X64: %cmp = icmp eq ptr addrspace(270) %{{.+}}, %i
   // X86: %{{.+}} = addrspacecast ptr addrspace(272) %j to ptr
@@ -67,7 +67,7 @@ int test_compare2(int *__ptr32 __sptr i, int *__ptr64 j) {
 }
 
 int test_compare3(int *__ptr32 __uptr i, int *__ptr64 j) {
-  // ALL-LABEL: define dso_local i32 @test_compare3
+  // ALL-LABEL: define dso_local noundef i32 @test_compare3
   // X64: %{{.+}} = addrspacecast ptr addrspace(271) %i to ptr
   // X64: %cmp = icmp eq ptr %{{.+}}, %j
   // X86: %{{.+}} = addrspacecast ptr addrspace(271) %i to ptr addrspace(272)
@@ -76,7 +76,7 @@ int test_compare3(int *__ptr32 __uptr i, int *__ptr64 j) {
 }
 
 int test_compare4(int *__ptr32 __sptr i, int *__ptr64 j) {
-  // ALL-LABEL: define dso_local i32 @test_compare4
+  // ALL-LABEL: define dso_local noundef i32 @test_compare4
   // X64: %{{.+}} = addrspacecast ptr addrspace(270) %i to ptr
   // X64: %cmp = icmp eq ptr %{{.+}}, %j
   // X86: %{{.+}} = addrspacecast ptr %i to ptr addrspace(272)
diff --git a/clang/test/CodeGenCUDA/link-builtin-bitcode-denormal-fp-mode.cu b/clang/test/CodeGenCUDA/link-builtin-bitcode-denormal-fp-mode.cu
index f12eacfb53fab3..ef02668c3697bf 100644
--- a/clang/test/CodeGenCUDA/link-builtin-bitcode-denormal-fp-mode.cu
+++ b/clang/test/CodeGenCUDA/link-builtin-bitcode-denormal-fp-mode.cu
@@ -111,16 +111,16 @@ __global__ void kernel_f64(double* out, double* a, double* b, double* c) {
 }
 }
 
-// INTERNALIZE: define internal half @do_f16_stuff({{.*}}) #[[$FUNCATTR:[0-9]+]]
-// INTERNALIZE: define internal float @do_f32_stuff({{.*}}) #[[$FUNCATTR]]
-// INTERNALIZE: define internal double @do_f64_stuff({{.*}}) #[[$FUNCATTR]]
-// INTERNALIZE: define internal float @weak_do_f32_stuff({{.*}}) #[[$WEAK_FUNCATTR:[0-9]+]]
+// INTERNALIZE: define internal {{(noundef )?}}half @do_f16_stuff({{.*}}) #[[$FUNCATTR:[0-9]+]]
+// INTERNALIZE: define internal {{(noundef )?}}float @do_f32_stuff({{.*}}) #[[$FUNCATTR]]
+// INTERNALIZE: define internal {{(noundef )?}}double @do_f64_stuff({{.*}}) #[[$FUNCATTR]]
+// INTERNALIZE: define internal {{(noundef )?}}float @weak_do_f32_stuff({{.*}}) #[[$WEAK_FUNCATTR:[0-9]+]]
 
 
-// NOINTERNALIZE: define dso_local half @do_f16_stuff({{.*}}) #[[$FUNCATTR:[0-9]+]]
-// NOINTERNALIZE: define dso_local float @do_f32_stuff({{.*}}) #[[$FUNCATTR]]
-// NOINTERNALIZE: define dso_local double @do_f64_stuff({{.*}}) #[[$FUNCATTR]]
-// NOINTERNALIZE: define weak float @weak_do_f32_stuff({{.*}}) #[[$WEAK_FUNCATTR:[0-9]+]]
+// NOINTERNALIZE: define dso_local {{(noundef )?}}half @do_f16_stuff({{.*}}) #[[$FUNCATTR:[0-9]+]]
+// NOINTERNALIZE: define dso_local {{(noundef )?}}float @do_f32_stuff({{.*}}) #[[$FUNCATTR]]
+// NOINTERNALIZE: define dso_local {{(noundef )?}}double @do_f64_stuff({{.*}}) #[[$FUNCATTR]]
+// NOINTERNALIZE: define weak {{(noundef )?}}float @weak_do_f32_stuff({{.*}}) #[[$WEAK_FUNCATTR:[0-9]+]]
 
 
 
diff --git a/clang/test/CodeGenOpenCL/as_type.cl b/clang/test/CodeGenOpenCL/as_type.cl
index d8e75db936f732..1fe26fbeafdb4b 100644
--- a/clang/test/CodeGenOpenCL/as_type.cl
+++ b/clang/test/CodeGenOpenCL/as_type.cl
@@ -27,7 +27,7 @@ char3 f3(int x) {
   return __builtin_astype(x, char3);
 }
 
-//CHECK: define{{.*}} spir_func <4 x i8> @f4(i32 noundef %[[x:.*]])
+//CHECK: define{{.*}} spir_func noundef <4 x i8> @f4(i32 noundef %[[x:.*]])
 //CHECK: %[[astype:.*]] = bitcast i32 %[[x]] to <4 x i8>
 //CHECK-NOT: shufflevector
 //CHECK: ret <4 x i8> %[[astype]]
@@ -43,7 +43,7 @@ int f5(char3 x) {
   return __builtin_astype(x, int);
 }
 
-//CHECK: define{{.*}} spir_func i32 @f6(<4 x i8> noundef %[[x:.*]])
+//CHECK: define{{.*}} spir_func noundef i32 @f6(<4 x i8> noundef %[[x:.*]])
 //CHECK: %[[astype:.*]] = bitcast <4 x i8> %[[x]] to i32
 //CHECK-NOT: shufflevector
 //CHECK: ret i32 %[[astype]]
@@ -51,7 +51,7 @@ int f6(char4 x) {
   return __builtin_astype(x, int);
 }
 
-//CHECK: define{{.*}} spir_func <3 x i8> @f7(<3 x i8> noundef returned %[[x:.*]])
+//CHECK: define{{.*}} spir_func noundef <3 x i8> @f7(<3 x i8> noundef returned %[[x:.*]])
 //CHECK-NOT: bitcast
 //CHECK-NOT: shufflevector
 //CHECK: ret <3 x i8> %[[x]]
@@ -67,21 +67,21 @@ int3 f8(char16 x) {
   return __builtin_astype(x, int3);
 }
 
-//CHECK: define{{.*}} spir_func ptr addrspace(1) @addr_cast(ptr noundef readnone %[[x:.*]])
+//CHECK: define{{.*}} spir_func noundef ptr a...
[truncated]

dtcxzyw added a commit to dtcxzyw/llvm-opt-benchmark that referenced this pull request Dec 29, 2023
@dtcxzyw
pre-commit: test PR76553
6afeb5f 
PR Link: llvm/llvm-project#76553
@dtcxzyw dtcxzyw mentioned this pull request Dec 29, 2023
Closed
@nikic
Copy link
Contributor

nikic commented Dec 29, 2023

There are lld test failures.

nikic
nikic reviewed Dec 29, 2023
View reviewed changes
clang/test/CodeGen/arm-vector_type-params-returns.c
@@ -27,7 +27,7 @@
#endif

// function return types
// CHECK-LABEL: define dso_local <8 x half> @test_ret_v8f16(
// CHECK-LABEL: define dso_local noundef <8 x half> @test_ret_v8f16(
Copy link
Contributor

@nikic nikic Dec 29, 2023 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you know why clang places the noundef on the argument but not the return for these, despite them having the same type? Is this due to some weird C rule or an implementation oversight?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah, it seems to indeed be a "weird C rule":

llvm-project/clang/lib/CodeGen/CGCall.cpp

Lines 1816 to 1819 in 3c92011

// C++ explicitly makes returning undefined values UB. C's rule only applies
// to used values, so we never mark them noundef for now.
if (!Module.getLangOpts().CPlusPlus)
return false;

nikic
nikic reviewed Dec 29, 2023
View reviewed changes
llvm/lib/Transforms/IPO/FunctionAttrs.cpp Outdated Show resolved Hide resolved
llvm/lib/Transforms/IPO/FunctionAttrs.cpp Outdated Show resolved Hide resolved
@dtcxzyw dtcxzyw force-pushed the perf/infer-noundef-retattr branch from 1dafc28 to 46188df Compare December 31, 2023 10:58
@dtcxzyw dtcxzyw force-pushed the perf/infer-noundef-retattr branch from 46188df to fe11127 Compare December 31, 2023 11:08
@dtcxzyw
Copy link
Member Author

dtcxzyw commented Dec 31, 2023

There are lld test failures.

Done.

nikic
nikic approved these changes Dec 31, 2023
View reviewed changes
Copy link
Contributor

@nikic nikic left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM, assuming Linux CI passes now.

@dtcxzyw
Copy link
Member Author

dtcxzyw commented Dec 31, 2023

Failed Tests (3):
LLVM :: CodeGen/BPF/loop-exit-cond.ll
LLVM :: CodeGen/NVPTX/nvvm-reflect-opaque.ll
LLVM :: CodeGen/NVPTX/nvvm-reflect.ll

@dtcxzyw dtcxzyw force-pushed the perf/infer-noundef-retattr branch from fe11127 to c5e8738 Compare December 31, 2023 12:25
@dtcxzyw dtcxzyw merged commit 1228bec into llvm:main Dec 31, 2023
3 of 4 checks passed
@dtcxzyw dtcxzyw deleted the perf/infer-noundef-retattr branch December 31, 2023 12:44
@vitalybuka
Copy link
Collaborator

Can you please fix or revert https://lab.llvm.org/buildbot/#/builders/74/builds/24592 ?

@vitalybuka
Copy link
Collaborator

vitalybuka commented Jan 1, 2024 •
edited
Loading 

This patch deduces noundef attributes for return values.
IIUC, a function returns noundef values iff all of its return values are guaranteed not to be undef or poison.

What exactly does this mean? I assume we do load from some random pointer and return that value, it will not be marked as noundef?

@vitalybuka
Copy link
Collaborator

Another problem is that there was assumption that definition and declaration will be consistent, for msan is a part of ABI.
Clang will generate them consistently for all modules, with declaration and definition of the function.

Is possible that after this patch some definitions will get noundef, but declarations in other modules will still have none? If that is so, it breaks msan.

@dtcxzyw
Copy link
Member Author

dtcxzyw commented Jan 1, 2024

Can you please fix or revert https://lab.llvm.org/buildbot/#/builders/74/builds/24592 ?

Thank you for reporting this! I will check the error log.

@nikic
Copy link
Contributor

nikic commented Jan 1, 2024

Yeah, we should skip this inference for functions with the sanitize_memory attribute.

@dtcxzyw
Copy link
Member Author

dtcxzyw commented Jan 1, 2024

Yeah, we should skip this inference for functions with the sanitize_memory attribute.

I will post a patch later.

@dtcxzyw dtcxzyw mentioned this pull request Jan 1, 2024
Merged
@dtcxzyw
Copy link
Member Author

dtcxzyw commented Jan 1, 2024

Yeah, we should skip this inference for functions with the sanitize_memory attribute.

I will post a patch later.

Candidate patch: #76691

dtcxzyw added a commit that referenced this pull request Jan 1, 2024
@dtcxzyw
[FuncAttrs] Don't infer noundef for functions with `sanitize_memory…
7e405eb 
…` attribute (#76691)

MemorySanitizer assumes that the definition and declaration of a
function will be consistent. If we add `noundef` for some definitions,
it will break msan.

Fix buildbot failure caused by #76553.
@dtcxzyw
Copy link
Member Author

dtcxzyw commented Jan 2, 2024

Can you please fix or revert https://lab.llvm.org/buildbot/#/builders/74/builds/24592 ?

Should be fixed by 7e405eb.
https://lab.llvm.org/buildbot/#/builders/74/builds/24613

nikic added a commit that referenced this pull request Jan 2, 2024
@nikic
[gold] Fix tests after #76553 (NFC)
5c458ed
@nikic nikic mentioned this pull request Jan 2, 2024
Closed
maurer added a commit to maurer/rust that referenced this pull request Jan 2, 2024
@maurer
llvm: Allow noundef on llvm-18
13b3403 
LLVM 18 will automatically infer @noundef in some situations. Adjust
codegen tests to accept this.

See llvm/llvm-project#76553 for why @noundef is being generated now.
@aeubanks aeubanks mentioned this pull request Jan 2, 2024
Closed
maurer added a commit to maurer/rust that referenced this pull request Jan 2, 2024
@maurer
llvm: Allow noundef in sparc codegen tests
96ce14b 
LLVM 18 will automatically infer `noundef` in some situations on SPARC.
Adjust codegen tests to accept this.

See llvm/llvm-project#76553 for why `noundef` is being generated now.
@maurer maurer mentioned this pull request Jan 2, 2024
Merged
maurer added a commit to maurer/rust that referenced this pull request Jan 2, 2024
@maurer
llvm: Allow noundef in codegen tests
ee86b1f 
LLVM 18 will automatically infer `noundef` in some situations.
Adjust codegen tests to accept this.

See llvm/llvm-project#76553 for why `noundef` is being generated now.
fmease added a commit to fmease/rust that referenced this pull request Jan 3, 2024
@fmease
Rollup merge of rust-lang#119523 - maurer:fix-sparc-llvm-18, r=nikic
7be0135 
llvm: Allow `noundef` in codegen tests

LLVM 18 will automatically infer `noundef` in some situations. Adjust codegen tests to accept this.

See llvm/llvm-project#76553 for why `noundef` is being generated now.

`@rustbot` label:+llvm-main
fmease added a commit to fmease/rust that referenced this pull request Jan 3, 2024
@fmease
Rollup merge of rust-lang#119523 - maurer:fix-sparc-llvm-18, r=nikic
fcec407 
llvm: Allow `noundef` in codegen tests

LLVM 18 will automatically infer `noundef` in some situations. Adjust codegen tests to accept this.

See llvm/llvm-project#76553 for why `noundef` is being generated now.

``@rustbot`` label:+llvm-main
rust-timer added a commit to rust-lang-ci/rust that referenced this pull request Jan 3, 2024
@rust-timer
Unrolled build for rust-lang#119523
eaab8b5 
Rollup merge of rust-lang#119523 - maurer:fix-sparc-llvm-18, r=nikic

llvm: Allow `noundef` in codegen tests

LLVM 18 will automatically infer `noundef` in some situations. Adjust codegen tests to accept this.

See llvm/llvm-project#76553 for why `noundef` is being generated now.

``@rustbot`` label:+llvm-main
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@nikic nikic nikic approved these changes

@fhahn fhahn Awaiting requested review from fhahn

@nunoplopes nunoplopes Awaiting requested review from nunoplopes

Assignees
No one assigned
Labels
backend:X86 clang Clang issues not falling into any other category debuginfo lld:COFF lld:ELF lld llvm:transforms LTO Link time optimization (regular/full LTO or ThinLTO)
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@dtcxzyw @llvmbot @nikic @vitalybuka