[flang][Driver] Add support for -f[no-]wrapv and -f[no]-strict-overflow in the frontend #110061

yus3710-fj · 2024-09-26T00:21:45Z

This patch introduces the options for integer overflow flags into Flang.
The behavior is similar to that of Clang.

…ow in the frontend This patch introduces the options for integer overflow flags into Flang.

llvmbot · 2024-10-03T04:15:47Z

@llvm/pr-subscribers-flang-driver

@llvm/pr-subscribers-clang-driver

Author: Yusuke MINATO (yus3710-fj)

Changes

This patch introduces the options for integer overflow flags into Flang.
The behavior is similar to that of Clang.

Full diff: https://github.com/llvm/llvm-project/pull/110061.diff

8 Files Affected:

(modified) clang/include/clang/Driver/Options.td (+7-4)
(modified) clang/lib/Driver/ToolChains/Flang.cpp (+11)
(modified) flang/include/flang/Common/LangOptions.def (+2)
(modified) flang/include/flang/Common/LangOptions.h (+8)
(modified) flang/include/flang/Lower/LoweringOptions.def (+2-3)
(modified) flang/lib/Frontend/CompilerInvocation.cpp (+33-3)
(modified) flang/test/Driver/frontend-forwarding.f90 (+2)
(added) flang/test/Driver/integer-overflow.f90 (+10)

diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
index 9d183ff2d69b3c..4610cd917f685d 100644
--- a/clang/include/clang/Driver/Options.td
+++ b/clang/include/clang/Driver/Options.td
@@ -3451,7 +3451,8 @@ def fno_strict_aliasing : Flag<["-"], "fno-strict-aliasing">, Group<f_Group>,
 def fstruct_path_tbaa : Flag<["-"], "fstruct-path-tbaa">, Group<f_Group>;
 def fno_struct_path_tbaa : Flag<["-"], "fno-struct-path-tbaa">, Group<f_Group>;
 def fno_strict_enums : Flag<["-"], "fno-strict-enums">, Group<f_Group>;
-def fno_strict_overflow : Flag<["-"], "fno-strict-overflow">, Group<f_Group>;
+def fno_strict_overflow : Flag<["-"], "fno-strict-overflow">, Group<f_Group>,
+  Visibility<[ClangOption, FlangOption]>;
 def fno_pointer_tbaa : Flag<["-"], "fno-pointer-tbaa">, Group<f_Group>;
 def fno_temp_file : Flag<["-"], "fno-temp-file">, Group<f_Group>,
   Visibility<[ClangOption, CC1Option, CLOption, DXCOption]>, HelpText<
@@ -3467,7 +3468,8 @@ def fno_verbose_asm : Flag<["-"], "fno-verbose-asm">, Group<f_Group>,
   Visibility<[ClangOption, CC1Option]>,
   MarshallingInfoNegativeFlag<CodeGenOpts<"AsmVerbose">>;
 def fno_working_directory : Flag<["-"], "fno-working-directory">, Group<f_Group>;
-def fno_wrapv : Flag<["-"], "fno-wrapv">, Group<f_Group>;
+def fno_wrapv : Flag<["-"], "fno-wrapv">, Group<f_Group>,
+  Visibility<[ClangOption, FlangOption]>;
 def fobjc_arc : Flag<["-"], "fobjc-arc">, Group<f_Group>,
   Visibility<[ClangOption, CC1Option]>,
   HelpText<"Synthesize retain and release calls for Objective-C pointers">;
@@ -3963,7 +3965,8 @@ defm strict_vtable_pointers : BoolFOption<"strict-vtable-pointers",
           "Enable optimizations based on the strict rules for"
             " overwriting polymorphic C++ objects">,
   NegFlag<SetFalse>>;
-def fstrict_overflow : Flag<["-"], "fstrict-overflow">, Group<f_Group>;
+def fstrict_overflow : Flag<["-"], "fstrict-overflow">, Group<f_Group>,
+  Visibility<[ClangOption, FlangOption]>;
 def fpointer_tbaa : Flag<["-"], "fpointer-tbaa">, Group<f_Group>;
 def fdriver_only : Flag<["-"], "fdriver-only">, Flags<[NoXarchOption]>,
   Visibility<[ClangOption, CLOption, DXCOption]>,
@@ -4232,7 +4235,7 @@ defm virtual_function_elimination : BoolFOption<"virtual-function-elimination",
   NegFlag<SetFalse>, BothFlags<[], [ClangOption, CLOption]>>;
 
 def fwrapv : Flag<["-"], "fwrapv">, Group<f_Group>,
-  Visibility<[ClangOption, CC1Option]>,
+  Visibility<[ClangOption, CC1Option, FlangOption, FC1Option]>,
   HelpText<"Treat signed integer overflow as two's complement">;
 def fwritable_strings : Flag<["-"], "fwritable-strings">, Group<f_Group>,
   Visibility<[ClangOption, CC1Option]>,
diff --git a/clang/lib/Driver/ToolChains/Flang.cpp b/clang/lib/Driver/ToolChains/Flang.cpp
index 98350690f8d20e..1865fd57a2d893 100644
--- a/clang/lib/Driver/ToolChains/Flang.cpp
+++ b/clang/lib/Driver/ToolChains/Flang.cpp
@@ -866,6 +866,17 @@ void Flang::ConstructJob(Compilation &C, const JobAction &JA,
     }
   }
 
+  // -fno-strict-overflow implies -fwrapv if it isn't disabled, but
+  // -fstrict-overflow won't turn off an explicitly enabled -fwrapv.
+  if (Arg *A = Args.getLastArg(options::OPT_fwrapv, options::OPT_fno_wrapv)) {
+    if (A->getOption().matches(options::OPT_fwrapv))
+      CmdArgs.push_back("-fwrapv");
+  } else if (Arg *A = Args.getLastArg(options::OPT_fstrict_overflow,
+                                      options::OPT_fno_strict_overflow)) {
+    if (A->getOption().matches(options::OPT_fno_strict_overflow))
+      CmdArgs.push_back("-fwrapv");
+  }
+
   assert((Output.isFilename() || Output.isNothing()) && "Invalid output.");
   if (Output.isFilename()) {
     CmdArgs.push_back("-o");
diff --git a/flang/include/flang/Common/LangOptions.def b/flang/include/flang/Common/LangOptions.def
index d3e1e972d1519f..1bfdba9cc2c1c7 100644
--- a/flang/include/flang/Common/LangOptions.def
+++ b/flang/include/flang/Common/LangOptions.def
@@ -20,6 +20,8 @@ LANGOPT(Name, Bits, Default)
 #endif
 
 ENUM_LANGOPT(FPContractMode, FPModeKind, 2, FPM_Fast) ///< FP Contract Mode (off/fast)
+/// signed integer overflow handling
+ENUM_LANGOPT(SignedOverflowBehavior, SignedOverflowBehaviorTy, 1, SOB_Undefined)
 
 /// Indicate a build without the standard GPU libraries.
 LANGOPT(NoGPULib  , 1, false)
diff --git a/flang/include/flang/Common/LangOptions.h b/flang/include/flang/Common/LangOptions.h
index 52a45047deb0e2..83f25cfbe26142 100644
--- a/flang/include/flang/Common/LangOptions.h
+++ b/flang/include/flang/Common/LangOptions.h
@@ -27,6 +27,14 @@ namespace Fortran::common {
 class LangOptionsBase {
 
 public:
+  enum SignedOverflowBehaviorTy {
+    // -fno-wrapv (default behavior in Flang)
+    SOB_Undefined,
+
+    // -fwrapv
+    SOB_Defined,
+  };
+
   enum FPModeKind {
     // Do not fuse FP ops
     FPM_Off,
diff --git a/flang/include/flang/Lower/LoweringOptions.def b/flang/include/flang/Lower/LoweringOptions.def
index d3f17c3f939c16..231de533fbd30a 100644
--- a/flang/include/flang/Lower/LoweringOptions.def
+++ b/flang/include/flang/Lower/LoweringOptions.def
@@ -35,9 +35,8 @@ ENUM_LOWERINGOPT(NoPPCNativeVecElemOrder, unsigned, 1, 0)
 ENUM_LOWERINGOPT(Underscoring, unsigned, 1, 1)
 
 /// If true, assume the behavior of integer overflow is defined
-/// (i.e. wraps around as two's complement). On by default.
-/// TODO: make the default off
-ENUM_LOWERINGOPT(IntegerWrapAround, unsigned, 1, 1)
+/// (i.e. wraps around as two's complement). Off by default.
+ENUM_LOWERINGOPT(IntegerWrapAround, unsigned, 1, 0)
 
 /// If true, add nsw flags to loop variable increments.
 /// Off by default.
diff --git a/flang/lib/Frontend/CompilerInvocation.cpp b/flang/lib/Frontend/CompilerInvocation.cpp
index 2154b9ab2fbf47..c8160a4236b466 100644
--- a/flang/lib/Frontend/CompilerInvocation.cpp
+++ b/flang/lib/Frontend/CompilerInvocation.cpp
@@ -1109,6 +1109,24 @@ static bool parseOpenMPArgs(CompilerInvocation &res, llvm::opt::ArgList &args,
   return diags.getNumErrors() == numErrorsBefore;
 }
 
+/// Parses signed integer overflow options and populates the
+/// CompilerInvocation accordingly.
+/// Returns false if new errors are generated.
+///
+/// \param [out] invoc Stores the processed arguments
+/// \param [in] args The compiler invocation arguments to parse
+/// \param [out] diags DiagnosticsEngine to report erros with
+static bool parseIntegerOverflowArgs(CompilerInvocation &invoc,
+                                     llvm::opt::ArgList &args,
+                                     clang::DiagnosticsEngine &diags) {
+  Fortran::common::LangOptions &opts = invoc.getLangOpts();
+
+  if (args.getLastArg(clang::driver::options::OPT_fwrapv))
+    opts.setSignedOverflowBehavior(Fortran::common::LangOptions::SOB_Defined);
+
+  return true;
+}
+
 /// Parses all floating point related arguments and populates the
 /// CompilerInvocation accordingly.
 /// Returns false if new errors are generated.
@@ -1249,6 +1267,18 @@ static bool parseLinkerOptionsArgs(CompilerInvocation &invoc,
   return true;
 }
 
+static bool parseLangOptionsArgs(CompilerInvocation &invoc,
+                                 llvm::opt::ArgList &args,
+                                 clang::DiagnosticsEngine &diags) {
+  bool success = true;
+
+  success &= parseIntegerOverflowArgs(invoc, args, diags);
+  success &= parseFloatingPointArgs(invoc, args, diags);
+  success &= parseVScaleArgs(invoc, args, diags);
+
+  return success;
+}
+
 bool CompilerInvocation::createFromArgs(
     CompilerInvocation &invoc, llvm::ArrayRef<const char *> commandLineArgs,
     clang::DiagnosticsEngine &diags, const char *argv0) {
@@ -1357,9 +1387,7 @@ bool CompilerInvocation::createFromArgs(
   invoc.frontendOpts.mlirArgs =
       args.getAllArgValues(clang::driver::options::OPT_mmlir);
 
-  success &= parseFloatingPointArgs(invoc, args, diags);
-
-  success &= parseVScaleArgs(invoc, args, diags);
+  success &= parseLangOptionsArgs(invoc, args, diags);
 
   success &= parseLinkerOptionsArgs(invoc, args, diags);
 
@@ -1571,6 +1599,8 @@ void CompilerInvocation::setLoweringOptions() {
   loweringOpts.setUnderscoring(codegenOpts.Underscoring);
 
   const Fortran::common::LangOptions &langOptions = getLangOpts();
+  loweringOpts.setIntegerWrapAround(langOptions.getSignedOverflowBehavior() ==
+                                    Fortran::common::LangOptions::SOB_Defined);
   Fortran::common::MathOptionsBase &mathOpts = loweringOpts.getMathOptions();
   // TODO: when LangOptions are finalized, we can represent
   //       the math related options using Fortran::commmon::MathOptionsBase,
diff --git a/flang/test/Driver/frontend-forwarding.f90 b/flang/test/Driver/frontend-forwarding.f90
index 35adb47b56861e..382c1aa5d350b7 100644
--- a/flang/test/Driver/frontend-forwarding.f90
+++ b/flang/test/Driver/frontend-forwarding.f90
@@ -14,6 +14,7 @@
 ! RUN:     -fno-signed-zeros \
 ! RUN:     -fassociative-math \
 ! RUN:     -freciprocal-math \
+! RUN:     -fno-strict-overflow \
 ! RUN:     -fomit-frame-pointer \
 ! RUN:     -fpass-plugin=Bye%pluginext \
 ! RUN:     -fversion-loops-for-stride \
@@ -63,4 +64,5 @@
 ! CHECK: "-Rpass=inline"
 ! CHECK: "-mframe-pointer=none"
 ! CHECK: "-mllvm" "-print-before-all"
+! CHECK: "-fwrapv"
 ! CHECK: "-save-temps=obj"
diff --git a/flang/test/Driver/integer-overflow.f90 b/flang/test/Driver/integer-overflow.f90
new file mode 100644
index 00000000000000..2cfb36527a30ad
--- /dev/null
+++ b/flang/test/Driver/integer-overflow.f90
@@ -0,0 +1,10 @@
+! Test for correct forwarding of integer overflow flags from the compiler driver
+! to the frontend driver
+
+! RUN: %flang -### -fno-strict-overflow %s 2>&1 | FileCheck %s --check-prefixes CHECK,INDUCED
+! RUN: %flang -### -fstrict-overflow %s 2>&1 | FileCheck %s
+! RUN: %flang -### -fno-wrapv %s 2>&1 | FileCheck %s
+! RUN: %flang -### -fno-wrapv -fno-strict-overflow %s 2>&1 | FileCheck %s
+
+! CHECK-NOT: "-fno-wrapv"
+! INDUCED: "-fwrapv"

llvmbot · 2024-10-03T04:15:47Z

@llvm/pr-subscribers-flang-fir-hlfir

Author: Yusuke MINATO (yus3710-fj)

Changes

This patch introduces the options for integer overflow flags into Flang.
The behavior is similar to that of Clang.

Full diff: https://github.com/llvm/llvm-project/pull/110061.diff

8 Files Affected:

(modified) clang/include/clang/Driver/Options.td (+7-4)
(modified) clang/lib/Driver/ToolChains/Flang.cpp (+11)
(modified) flang/include/flang/Common/LangOptions.def (+2)
(modified) flang/include/flang/Common/LangOptions.h (+8)
(modified) flang/include/flang/Lower/LoweringOptions.def (+2-3)
(modified) flang/lib/Frontend/CompilerInvocation.cpp (+33-3)
(modified) flang/test/Driver/frontend-forwarding.f90 (+2)
(added) flang/test/Driver/integer-overflow.f90 (+10)

diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
index 9d183ff2d69b3c..4610cd917f685d 100644
--- a/clang/include/clang/Driver/Options.td
+++ b/clang/include/clang/Driver/Options.td
@@ -3451,7 +3451,8 @@ def fno_strict_aliasing : Flag<["-"], "fno-strict-aliasing">, Group<f_Group>,
 def fstruct_path_tbaa : Flag<["-"], "fstruct-path-tbaa">, Group<f_Group>;
 def fno_struct_path_tbaa : Flag<["-"], "fno-struct-path-tbaa">, Group<f_Group>;
 def fno_strict_enums : Flag<["-"], "fno-strict-enums">, Group<f_Group>;
-def fno_strict_overflow : Flag<["-"], "fno-strict-overflow">, Group<f_Group>;
+def fno_strict_overflow : Flag<["-"], "fno-strict-overflow">, Group<f_Group>,
+  Visibility<[ClangOption, FlangOption]>;
 def fno_pointer_tbaa : Flag<["-"], "fno-pointer-tbaa">, Group<f_Group>;
 def fno_temp_file : Flag<["-"], "fno-temp-file">, Group<f_Group>,
   Visibility<[ClangOption, CC1Option, CLOption, DXCOption]>, HelpText<
@@ -3467,7 +3468,8 @@ def fno_verbose_asm : Flag<["-"], "fno-verbose-asm">, Group<f_Group>,
   Visibility<[ClangOption, CC1Option]>,
   MarshallingInfoNegativeFlag<CodeGenOpts<"AsmVerbose">>;
 def fno_working_directory : Flag<["-"], "fno-working-directory">, Group<f_Group>;
-def fno_wrapv : Flag<["-"], "fno-wrapv">, Group<f_Group>;
+def fno_wrapv : Flag<["-"], "fno-wrapv">, Group<f_Group>,
+  Visibility<[ClangOption, FlangOption]>;
 def fobjc_arc : Flag<["-"], "fobjc-arc">, Group<f_Group>,
   Visibility<[ClangOption, CC1Option]>,
   HelpText<"Synthesize retain and release calls for Objective-C pointers">;
@@ -3963,7 +3965,8 @@ defm strict_vtable_pointers : BoolFOption<"strict-vtable-pointers",
           "Enable optimizations based on the strict rules for"
             " overwriting polymorphic C++ objects">,
   NegFlag<SetFalse>>;
-def fstrict_overflow : Flag<["-"], "fstrict-overflow">, Group<f_Group>;
+def fstrict_overflow : Flag<["-"], "fstrict-overflow">, Group<f_Group>,
+  Visibility<[ClangOption, FlangOption]>;
 def fpointer_tbaa : Flag<["-"], "fpointer-tbaa">, Group<f_Group>;
 def fdriver_only : Flag<["-"], "fdriver-only">, Flags<[NoXarchOption]>,
   Visibility<[ClangOption, CLOption, DXCOption]>,
@@ -4232,7 +4235,7 @@ defm virtual_function_elimination : BoolFOption<"virtual-function-elimination",
   NegFlag<SetFalse>, BothFlags<[], [ClangOption, CLOption]>>;
 
 def fwrapv : Flag<["-"], "fwrapv">, Group<f_Group>,
-  Visibility<[ClangOption, CC1Option]>,
+  Visibility<[ClangOption, CC1Option, FlangOption, FC1Option]>,
   HelpText<"Treat signed integer overflow as two's complement">;
 def fwritable_strings : Flag<["-"], "fwritable-strings">, Group<f_Group>,
   Visibility<[ClangOption, CC1Option]>,
diff --git a/clang/lib/Driver/ToolChains/Flang.cpp b/clang/lib/Driver/ToolChains/Flang.cpp
index 98350690f8d20e..1865fd57a2d893 100644
--- a/clang/lib/Driver/ToolChains/Flang.cpp
+++ b/clang/lib/Driver/ToolChains/Flang.cpp
@@ -866,6 +866,17 @@ void Flang::ConstructJob(Compilation &C, const JobAction &JA,
     }
   }
 
+  // -fno-strict-overflow implies -fwrapv if it isn't disabled, but
+  // -fstrict-overflow won't turn off an explicitly enabled -fwrapv.
+  if (Arg *A = Args.getLastArg(options::OPT_fwrapv, options::OPT_fno_wrapv)) {
+    if (A->getOption().matches(options::OPT_fwrapv))
+      CmdArgs.push_back("-fwrapv");
+  } else if (Arg *A = Args.getLastArg(options::OPT_fstrict_overflow,
+                                      options::OPT_fno_strict_overflow)) {
+    if (A->getOption().matches(options::OPT_fno_strict_overflow))
+      CmdArgs.push_back("-fwrapv");
+  }
+
   assert((Output.isFilename() || Output.isNothing()) && "Invalid output.");
   if (Output.isFilename()) {
     CmdArgs.push_back("-o");
diff --git a/flang/include/flang/Common/LangOptions.def b/flang/include/flang/Common/LangOptions.def
index d3e1e972d1519f..1bfdba9cc2c1c7 100644
--- a/flang/include/flang/Common/LangOptions.def
+++ b/flang/include/flang/Common/LangOptions.def
@@ -20,6 +20,8 @@ LANGOPT(Name, Bits, Default)
 #endif
 
 ENUM_LANGOPT(FPContractMode, FPModeKind, 2, FPM_Fast) ///< FP Contract Mode (off/fast)
+/// signed integer overflow handling
+ENUM_LANGOPT(SignedOverflowBehavior, SignedOverflowBehaviorTy, 1, SOB_Undefined)
 
 /// Indicate a build without the standard GPU libraries.
 LANGOPT(NoGPULib  , 1, false)
diff --git a/flang/include/flang/Common/LangOptions.h b/flang/include/flang/Common/LangOptions.h
index 52a45047deb0e2..83f25cfbe26142 100644
--- a/flang/include/flang/Common/LangOptions.h
+++ b/flang/include/flang/Common/LangOptions.h
@@ -27,6 +27,14 @@ namespace Fortran::common {
 class LangOptionsBase {
 
 public:
+  enum SignedOverflowBehaviorTy {
+    // -fno-wrapv (default behavior in Flang)
+    SOB_Undefined,
+
+    // -fwrapv
+    SOB_Defined,
+  };
+
   enum FPModeKind {
     // Do not fuse FP ops
     FPM_Off,
diff --git a/flang/include/flang/Lower/LoweringOptions.def b/flang/include/flang/Lower/LoweringOptions.def
index d3f17c3f939c16..231de533fbd30a 100644
--- a/flang/include/flang/Lower/LoweringOptions.def
+++ b/flang/include/flang/Lower/LoweringOptions.def
@@ -35,9 +35,8 @@ ENUM_LOWERINGOPT(NoPPCNativeVecElemOrder, unsigned, 1, 0)
 ENUM_LOWERINGOPT(Underscoring, unsigned, 1, 1)
 
 /// If true, assume the behavior of integer overflow is defined
-/// (i.e. wraps around as two's complement). On by default.
-/// TODO: make the default off
-ENUM_LOWERINGOPT(IntegerWrapAround, unsigned, 1, 1)
+/// (i.e. wraps around as two's complement). Off by default.
+ENUM_LOWERINGOPT(IntegerWrapAround, unsigned, 1, 0)
 
 /// If true, add nsw flags to loop variable increments.
 /// Off by default.
diff --git a/flang/lib/Frontend/CompilerInvocation.cpp b/flang/lib/Frontend/CompilerInvocation.cpp
index 2154b9ab2fbf47..c8160a4236b466 100644
--- a/flang/lib/Frontend/CompilerInvocation.cpp
+++ b/flang/lib/Frontend/CompilerInvocation.cpp
@@ -1109,6 +1109,24 @@ static bool parseOpenMPArgs(CompilerInvocation &res, llvm::opt::ArgList &args,
   return diags.getNumErrors() == numErrorsBefore;
 }
 
+/// Parses signed integer overflow options and populates the
+/// CompilerInvocation accordingly.
+/// Returns false if new errors are generated.
+///
+/// \param [out] invoc Stores the processed arguments
+/// \param [in] args The compiler invocation arguments to parse
+/// \param [out] diags DiagnosticsEngine to report erros with
+static bool parseIntegerOverflowArgs(CompilerInvocation &invoc,
+                                     llvm::opt::ArgList &args,
+                                     clang::DiagnosticsEngine &diags) {
+  Fortran::common::LangOptions &opts = invoc.getLangOpts();
+
+  if (args.getLastArg(clang::driver::options::OPT_fwrapv))
+    opts.setSignedOverflowBehavior(Fortran::common::LangOptions::SOB_Defined);
+
+  return true;
+}
+
 /// Parses all floating point related arguments and populates the
 /// CompilerInvocation accordingly.
 /// Returns false if new errors are generated.
@@ -1249,6 +1267,18 @@ static bool parseLinkerOptionsArgs(CompilerInvocation &invoc,
   return true;
 }
 
+static bool parseLangOptionsArgs(CompilerInvocation &invoc,
+                                 llvm::opt::ArgList &args,
+                                 clang::DiagnosticsEngine &diags) {
+  bool success = true;
+
+  success &= parseIntegerOverflowArgs(invoc, args, diags);
+  success &= parseFloatingPointArgs(invoc, args, diags);
+  success &= parseVScaleArgs(invoc, args, diags);
+
+  return success;
+}
+
 bool CompilerInvocation::createFromArgs(
     CompilerInvocation &invoc, llvm::ArrayRef<const char *> commandLineArgs,
     clang::DiagnosticsEngine &diags, const char *argv0) {
@@ -1357,9 +1387,7 @@ bool CompilerInvocation::createFromArgs(
   invoc.frontendOpts.mlirArgs =
       args.getAllArgValues(clang::driver::options::OPT_mmlir);
 
-  success &= parseFloatingPointArgs(invoc, args, diags);
-
-  success &= parseVScaleArgs(invoc, args, diags);
+  success &= parseLangOptionsArgs(invoc, args, diags);
 
   success &= parseLinkerOptionsArgs(invoc, args, diags);
 
@@ -1571,6 +1599,8 @@ void CompilerInvocation::setLoweringOptions() {
   loweringOpts.setUnderscoring(codegenOpts.Underscoring);
 
   const Fortran::common::LangOptions &langOptions = getLangOpts();
+  loweringOpts.setIntegerWrapAround(langOptions.getSignedOverflowBehavior() ==
+                                    Fortran::common::LangOptions::SOB_Defined);
   Fortran::common::MathOptionsBase &mathOpts = loweringOpts.getMathOptions();
   // TODO: when LangOptions are finalized, we can represent
   //       the math related options using Fortran::commmon::MathOptionsBase,
diff --git a/flang/test/Driver/frontend-forwarding.f90 b/flang/test/Driver/frontend-forwarding.f90
index 35adb47b56861e..382c1aa5d350b7 100644
--- a/flang/test/Driver/frontend-forwarding.f90
+++ b/flang/test/Driver/frontend-forwarding.f90
@@ -14,6 +14,7 @@
 ! RUN:     -fno-signed-zeros \
 ! RUN:     -fassociative-math \
 ! RUN:     -freciprocal-math \
+! RUN:     -fno-strict-overflow \
 ! RUN:     -fomit-frame-pointer \
 ! RUN:     -fpass-plugin=Bye%pluginext \
 ! RUN:     -fversion-loops-for-stride \
@@ -63,4 +64,5 @@
 ! CHECK: "-Rpass=inline"
 ! CHECK: "-mframe-pointer=none"
 ! CHECK: "-mllvm" "-print-before-all"
+! CHECK: "-fwrapv"
 ! CHECK: "-save-temps=obj"
diff --git a/flang/test/Driver/integer-overflow.f90 b/flang/test/Driver/integer-overflow.f90
new file mode 100644
index 00000000000000..2cfb36527a30ad
--- /dev/null
+++ b/flang/test/Driver/integer-overflow.f90
@@ -0,0 +1,10 @@
+! Test for correct forwarding of integer overflow flags from the compiler driver
+! to the frontend driver
+
+! RUN: %flang -### -fno-strict-overflow %s 2>&1 | FileCheck %s --check-prefixes CHECK,INDUCED
+! RUN: %flang -### -fstrict-overflow %s 2>&1 | FileCheck %s
+! RUN: %flang -### -fno-wrapv %s 2>&1 | FileCheck %s
+! RUN: %flang -### -fno-wrapv -fno-strict-overflow %s 2>&1 | FileCheck %s
+
+! CHECK-NOT: "-fno-wrapv"
+! INDUCED: "-fwrapv"

yus3710-fj linked an issue Sep 26, 2024 that may be closed by this pull request

[Flang] TSVC s243: some nsw flags for address calculations will help vectorization #78934

Closed

[flang][Driver] Add support for -f[no-]wrapv and -f[no]-strict-overfl…

aea2cfa

…ow in the frontend This patch introduces the options for integer overflow flags into Flang.

yus3710-fj force-pushed the options-for-overflow branch from d3a02bc to aea2cfa Compare October 3, 2024 03:51

yus3710-fj marked this pull request as ready for review October 3, 2024 04:15

llvmbot added clang Clang issues not falling into any other category clang:driver 'clang' and 'clang++' user-facing binaries. Not 'clang-cl' flang:driver flang Flang issues not falling into any other category flang:fir-hlfir labels Oct 3, 2024

yus3710-fj requested review from tblah, tarunprabhu and jeanPerier October 3, 2024 04:16

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[flang][Driver] Add support for -f[no-]wrapv and -f[no]-strict-overflow in the frontend #110061

[flang][Driver] Add support for -f[no-]wrapv and -f[no]-strict-overflow in the frontend #110061

yus3710-fj commented Sep 26, 2024

llvmbot commented Oct 3, 2024 •

edited

Loading

llvmbot commented Oct 3, 2024

[flang][Driver] Add support for -f[no-]wrapv and -f[no]-strict-overflow in the frontend #110061

Are you sure you want to change the base?

[flang][Driver] Add support for -f[no-]wrapv and -f[no]-strict-overflow in the frontend #110061

Conversation

yus3710-fj commented Sep 26, 2024

llvmbot commented Oct 3, 2024 • edited Loading

llvmbot commented Oct 3, 2024

llvmbot commented Oct 3, 2024 •

edited

Loading