[BuildLibCalls] Add noundef to standard I/O functions

This patch adds noundef to return value and arguments of standard I/O functions.
With this patch, passing undef or poison to the functions becomes undefined
behavior in LLVM IR. Since undef/poison is lowered from operations having UB in C/C++,
passing undef to them was already UB in source.

With this patch, the functions cannot return undef or poison anymore as well.
According to C17 standard, ungetc/ungetwc/fgetpos/ftell can generate unspecified
value; 3.19.3 says unspecified value is a valid value of the relevant type,
and using unspecified value is unspecified behavior, which is not UB, so it
cannot be undef (using undef is UB when e.g. it is used at branch condition).

— The value of the file position indicator after a successful call to the ungetc function for a text stream, or the ungetwc function for any stream, until all pushed-back characters are read or discarded (7.21.7.10, 7.29.3.10).
— The details of the value stored by the fgetpos function (7.21.9.1).
— The details of the value returned by the ftell function for a text stream (7.21.9.4).

In the long run, most of the functions listed in BuildLibCalls should have noundefs; to remove redundant diffs which will anyway disappear in the future, I added noundef to a few more non-I/O functions as well.

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D85345

Author: aqjune

clang/test/CodeGen/PR3589-freestanding-libcalls.c

@@ -1,6 +1,6 @@
 // RUN: %clang_cc1 -triple i386-unknown-unknown -emit-llvm %s -o - | grep 'declare i32 @printf' | count 1
-// RUN: %clang_cc1 -triple i386-unknown-unknown -O2 -emit-llvm %s -o - | grep 'declare i32 @puts' | count 1
-// RUN: %clang_cc1 -triple i386-unknown-unknown -ffreestanding -O2 -emit-llvm %s -o - | not grep 'declare i32 @puts'
+// RUN: %clang_cc1 -triple i386-unknown-unknown -O2 -emit-llvm %s -o - | grep 'declare noundef i32 @puts' | count 1
+// RUN: %clang_cc1 -triple i386-unknown-unknown -ffreestanding -O2 -emit-llvm %s -o - | not grep 'declare noundef i32 @puts'
 
 int printf(const char *, ...);
 

llvm/lib/Transforms/Utils/BuildLibCalls.cpp

@@ -37,6 +37,7 @@ STATISTIC(NumNoUnwind, "Number of functions inferred as nounwind");
 STATISTIC(NumNoCapture, "Number of arguments inferred as nocapture");
 STATISTIC(NumReadOnlyArg, "Number of arguments inferred as readonly");
 STATISTIC(NumNoAlias, "Number of function returns inferred as noalias");
+STATISTIC(NumNoUndef, "Number of function returns inferred as noundef returns");
 STATISTIC(NumNonNull, "Number of function returns inferred as nonnull returns");
 STATISTIC(NumReturnedArg, "Number of arguments inferred as returned");
 
@@ -104,6 +105,24 @@ static bool setOnlyReadsMemory(Function &F, unsigned ArgNo) {
   return true;
 }
 
+static bool setRetAndArgsNoUndef(Function &F) {
+  bool Changed = false;
+  if (!F.getReturnType()->isVoidTy() &&
+      !F.hasAttribute(AttributeList::ReturnIndex, Attribute::NoUndef)) {
+    F.addAttribute(AttributeList::ReturnIndex, Attribute::NoUndef);
+    ++NumNoUndef;
+    Changed = true;
+  }
+  for (unsigned ArgNo = 0; ArgNo < F.arg_size(); ++ArgNo) {
+    if (!F.hasParamAttribute(ArgNo, Attribute::NoUndef)) {
+      F.addParamAttr(ArgNo, Attribute::NoUndef);
+      ++NumNoUndef;
+      Changed = true;
+    }
+  }
+  return Changed;
+}
+
 static bool setRetNonNull(Function &F) {
   assert(F.getReturnType()->isPointerTy() &&
          "nonnull applies only to pointers");
@@ -227,6 +246,7 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_scanf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setOnlyReadsMemory(F, 0);
@@ -251,20 +271,23 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     Changed |= setOnlyReadsMemory(F, 0);
     return Changed;
   case LibFunc_sscanf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 1);
     Changed |= setOnlyReadsMemory(F, 0);
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_sprintf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotAlias(F, 0);
     Changed |= setDoesNotCapture(F, 1);
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_snprintf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotAlias(F, 0);
@@ -347,9 +370,11 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     return Changed;
   case LibFunc_read:
     // May throw; "read" is a valid pthread cancellation point.
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotCapture(F, 1);
     return Changed;
   case LibFunc_rewind:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     return Changed;
@@ -375,6 +400,7 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     return Changed;
   case LibFunc_write:
     // May throw; "write" is a valid pthread cancellation point.
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotCapture(F, 1);
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
@@ -428,6 +454,7 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     Changed |= setOnlyReadsMemory(F, 0);
     return Changed;
   case LibFunc_fopen:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setRetDoesNotAlias(F);
     Changed |= setDoesNotCapture(F, 0);
@@ -436,13 +463,21 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_fdopen:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setRetDoesNotAlias(F);
     Changed |= setDoesNotCapture(F, 1);
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_feof:
+    Changed |= setRetAndArgsNoUndef(F);
+    Changed |= setDoesNotThrow(F);
+    Changed |= setDoesNotCapture(F, 0);
+    return Changed;
   case LibFunc_free:
+    Changed |= setDoesNotThrow(F);
+    Changed |= setDoesNotCapture(F, 0);
+    return Changed;
   case LibFunc_fseek:
   case LibFunc_ftell:
   case LibFunc_fgetc:
@@ -456,64 +491,88 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
   case LibFunc_flockfile:
   case LibFunc_funlockfile:
   case LibFunc_ftrylockfile:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     return Changed;
   case LibFunc_ferror:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setOnlyReadsMemory(F);
     return Changed;
   case LibFunc_fputc:
   case LibFunc_fputc_unlocked:
   case LibFunc_fstat:
+    Changed |= setRetAndArgsNoUndef(F);
+    Changed |= setDoesNotThrow(F);
+    Changed |= setDoesNotCapture(F, 1);
+    return Changed;
   case LibFunc_frexp:
   case LibFunc_frexpf:
   case LibFunc_frexpl:
+    Changed |= setDoesNotThrow(F);
+    Changed |= setDoesNotCapture(F, 1);
+    return Changed;
   case LibFunc_fstatvfs:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 1);
     return Changed;
   case LibFunc_fgets:
   case LibFunc_fgets_unlocked:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 2);
     return Changed;
   case LibFunc_fread:
   case LibFunc_fread_unlocked:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 3);
     return Changed;
   case LibFunc_fwrite:
   case LibFunc_fwrite_unlocked:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 3);
     // FIXME: readonly #1?
     return Changed;
   case LibFunc_fputs:
   case LibFunc_fputs_unlocked:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 1);
     Changed |= setOnlyReadsMemory(F, 0);
     return Changed;
   case LibFunc_fscanf:
   case LibFunc_fprintf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 1);
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_fgetpos:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 1);
     return Changed;
   case LibFunc_getc:
+    Changed |= setRetAndArgsNoUndef(F);
+    Changed |= setDoesNotThrow(F);
+    Changed |= setDoesNotCapture(F, 0);
+    return Changed;
   case LibFunc_getlogin_r:
+    Changed |= setDoesNotThrow(F);
+    Changed |= setDoesNotCapture(F, 0);
+    return Changed;
   case LibFunc_getc_unlocked:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     return Changed;
@@ -525,6 +584,7 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
   case LibFunc_gets:
   case LibFunc_getchar:
   case LibFunc_getchar_unlocked:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     return Changed;
   case LibFunc_getitimer:
@@ -537,6 +597,7 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     Changed |= setOnlyReadsMemory(F, 0);
     return Changed;
   case LibFunc_ungetc:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 1);
     return Changed;
@@ -564,27 +625,32 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     return Changed;
   case LibFunc_putc:
   case LibFunc_putc_unlocked:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 1);
     return Changed;
   case LibFunc_puts:
   case LibFunc_printf:
   case LibFunc_perror:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setOnlyReadsMemory(F, 0);
     return Changed;
   case LibFunc_pread:
     // May throw; "pread" is a valid pthread cancellation point.
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotCapture(F, 1);
     return Changed;
   case LibFunc_pwrite:
     // May throw; "pwrite" is a valid pthread cancellation point.
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotCapture(F, 1);
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_putchar:
   case LibFunc_putchar_unlocked:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     return Changed;
   case LibFunc_popen:
@@ -600,18 +666,21 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     Changed |= setDoesNotCapture(F, 0);
     return Changed;
   case LibFunc_vscanf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setOnlyReadsMemory(F, 0);
     return Changed;
   case LibFunc_vsscanf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 1);
     Changed |= setOnlyReadsMemory(F, 0);
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_vfscanf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 1);
@@ -622,25 +691,29 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     Changed |= setRetDoesNotAlias(F);
     return Changed;
   case LibFunc_vprintf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setOnlyReadsMemory(F, 0);
     return Changed;
   case LibFunc_vfprintf:
   case LibFunc_vsprintf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 1);
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_vsnprintf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 2);
     Changed |= setOnlyReadsMemory(F, 2);
     return Changed;
   case LibFunc_open:
     // May throw; "open" is a valid pthread cancellation point.
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setOnlyReadsMemory(F, 0);
     return Changed;
@@ -693,14 +766,17 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_under_IO_getc:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     return Changed;
   case LibFunc_under_IO_putc:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 1);
     return Changed;
   case LibFunc_dunder_isoc99_scanf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setOnlyReadsMemory(F, 0);
@@ -714,13 +790,15 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     Changed |= setOnlyReadsMemory(F, 0);
     return Changed;
   case LibFunc_dunder_isoc99_sscanf:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setDoesNotCapture(F, 1);
     Changed |= setOnlyReadsMemory(F, 0);
     Changed |= setOnlyReadsMemory(F, 1);
     return Changed;
   case LibFunc_fopen64:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setRetDoesNotAlias(F);
     Changed |= setDoesNotCapture(F, 0);
@@ -730,6 +808,7 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     return Changed;
   case LibFunc_fseeko64:
   case LibFunc_ftello64:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 0);
     return Changed;
@@ -739,11 +818,13 @@ bool llvm::inferLibFuncAttributes(Function &F, const TargetLibraryInfo &TLI) {
     return Changed;
   case LibFunc_fstat64:
   case LibFunc_fstatvfs64:
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotThrow(F);
     Changed |= setDoesNotCapture(F, 1);
     return Changed;
   case LibFunc_open64:
     // May throw; "open" is a valid pthread cancellation point.
+    Changed |= setRetAndArgsNoUndef(F);
     Changed |= setDoesNotCapture(F, 0);
     Changed |= setOnlyReadsMemory(F, 0);
     return Changed;

llvm/test/CodeGen/X86/no-plt-libcalls.ll

@@ -17,7 +17,7 @@ define void @printf_call() {
 }
 
 ; CHECK: Function Attrs: nofree nounwind nonlazybind
-; CHECK-NEXT: declare i32 @puts(i8* nocapture readonly)
+; CHECK-NEXT: declare noundef i32 @puts(i8* nocapture noundef readonly)
 
 !llvm.module.flags = !{!0}
 !0 = !{i32 7, !"RtLibUseGOT", i32 1}