New options -fexceptions-fp-math and -fdenormal-fp-math

[SjoerdMeijer]

Summary

This adds options -fexceptions-fp-math and -fdenormal-fp-math, which are translated to function attributes. This is intended to be mapped on build attributes FP_exceptions and FP_denormal. Setting these build attributes allows better selection of floating point libraries.

[jmolloy]

Hi Sjoerd,

These make sense - we currently lack any way to inform the backend of the user's FP strictness requirements for exceptions and denormals which forces us to be conservative in the backend build attribute generation.

Your new -fno-exceptions-fp-math appears to be the same as GCC's "-ftrapping-math" which we already implement in the Driver. In fact, it looks like this code never passes -ftrapping-math through to CC1; it only uses it to determine if -menable-unsafe-fpmath should be set.

Therefore I think passing -ftrapping-math through to CC1 is the right thing here.

It seems to me that GCC doesn't have an equivalent flag to -fdenormal-fp-math, so inventing a flag sounds like the right thing to do. This patch should provide a docs patch for http://clang.llvm.org/docs/UsersManual.html#controlling-code-generation , which is generated from docs/UsersManual.rst.

Cheers,

James

[SjoerdMeijer]

Hi James,

Thanks for the review. I got rid of that new option -fexceptions-fp-math and now use -ftrapping-math for that which indeed was not passed through to CC1.

Cheers.

[jmolloy]

Hi Sjoerd,

This still needs a docs patch. I'd also really appreciate someone else to help sign this off - While it looks fine, I don't commit enough to clang to give the go-ahead.

Cheers,

James

[rengolin]

Hi James,

Also looks good to me, though it need docs (in the same patch).

Is there an LLVM counter part to this patch? Where is this going to be used?

From the fact that trapping-math wasn't passed down, its documentation may be missing/outdated/wrong, and probably nothing down the pipe is using it either. What are the plans for that?

cheers,
--renato

[SjoerdMeijer]

Sorry, forgot to add the doc patch, but it has been included now.

This will be used in llvm to emit and appropriately set build attributes ABI_FP_denormal and ABI_FP_exceptions. I am now trying to get a llvm patch ready for that (not uploaded yet, but it will mainly affect only ARMAsmPrinter.cpp and perhaps TargetOptions to query buildattributes).

[jmolloy]

LGTM.

Cheers,

James

cfe/trunk/docs/UsersManual.rst

.. option:: -ffast-math
 
    * Floating-point math obeys regular algebraic rules for real numbers (e.g.
      ``+`` and ``*`` are associative, ``x/y == x * (1/y)``, and
      ``(a + b) * c == a * c + b * c``),
    * operands to floating-point operations are not equal to ``NaN`` and
      ``Inf``, and
    * ``+0`` and ``-0`` are interchangeable.
 
+.. option:: -fdenormal-fp-math=[values]
+
+   Select which denormal numbers the code is permitted to require.
+
+   Valid values are: ``ieee``, ``preserve-sign``, and ``positive-zero``,
+   which correspond to IEEE 754 denormal numbers, the sign of a
+   flushed-to-zero number is preserved in the sign of 0, denormals are
+   flushed to positive zero, respectively.
+
 .. option:: -fwhole-program-vtables
 
    Enable whole-program vtable optimizations, such as single-implementation
    devirtualization and virtual constant propagation, for classes with
    :doc:`hidden LTO visibility <LTOVisibility>`. Requires ``-flto``.
 
 .. option:: -fno-assume-sane-operator-new
 

cfe/trunk/include/clang/Driver/Options.td

 def : Flag<["-"], "fexpensive-optimizations">, Group<clang_ignored_gcc_optimization_f_Group>;
 def : Flag<["-"], "fno-expensive-optimizations">, Group<clang_ignored_gcc_optimization_f_Group>;
 def fextdirs_EQ : Joined<["-"], "fextdirs=">, Group<f_Group>;
 def : Flag<["-"], "fdefer-pop">, Group<clang_ignored_gcc_optimization_f_Group>;
 def : Flag<["-"], "fno-defer-pop">, Group<clang_ignored_gcc_optimization_f_Group>;
 def : Flag<["-"], "fextended-identifiers">, Group<clang_ignored_f_Group>;
 def : Flag<["-"], "fno-extended-identifiers">, Group<f_Group>, Flags<[Unsupported]>;
 def fhosted : Flag<["-"], "fhosted">, Group<f_Group>;
+def fdenormal_fp_math_EQ : Joined<["-"], "fdenormal-fp-math=">, Group<f_Group>, Flags<[CC1Option]>;
 def ffast_math : Flag<["-"], "ffast-math">, Group<f_Group>, Flags<[CC1Option]>,
   HelpText<"Allow aggressive, lossy floating-point optimizations">;
 def fno_fast_math : Flag<["-"], "fno-fast-math">, Group<f_Group>;
 def fmath_errno : Flag<["-"], "fmath-errno">, Group<f_Group>, Flags<[CC1Option]>,
   HelpText<"Require math functions to indicate errors by setting errno">;
 def fno_math_errno : Flag<["-"], "fno-math-errno">, Group<f_Group>;
 def fbracket_depth_EQ : Joined<["-"], "fbracket-depth=">, Group<f_Group>;
 def fsignaling_math : Flag<["-"], "fsignaling-math">, Group<f_Group>;
def fno_signed_zeros :
   HelpText<"Allow optimizations that ignore the sign of floating point zeros">;
 def fhonor_nans : Flag<["-"], "fhonor-nans">, Group<f_Group>;
 def fno_honor_nans : Flag<["-"], "fno-honor-nans">, Group<f_Group>;
 def fhonor_infinities : Flag<["-"], "fhonor-infinities">, Group<f_Group>;
 def fno_honor_infinities : Flag<["-"], "fno-honor-infinities">, Group<f_Group>;
 // This option was originally misspelt "infinites" [sic].
 def : Flag<["-"], "fhonor-infinites">, Alias<fhonor_infinities>;
 def : Flag<["-"], "fno-honor-infinites">, Alias<fno_honor_infinities>;
-def ftrapping_math : Flag<["-"], "ftrapping-math">, Group<f_Group>;
-def fno_trapping_math : Flag<["-"], "fno-trapping-math">, Group<f_Group>;
+def ftrapping_math : Flag<["-"], "ftrapping-math">, Group<f_Group>, Flags<[CC1Option]>;
+def fno_trapping_math : Flag<["-"], "fno-trapping-math">, Group<f_Group>, Flags<[CC1Option]>;
 def ffp_contract : Joined<["-"], "ffp-contract=">, Group<f_Group>,
   Flags<[CC1Option]>, HelpText<"Form fused FP ops (e.g. FMAs): fast (everywhere)"
   " | on (according to FP_CONTRACT pragma, default) | off (never fuse)">;
 
 def ffor_scope : Flag<["-"], "ffor-scope">, Group<f_Group>;
 def fno_for_scope : Flag<["-"], "fno-for-scope">, Group<f_Group>;
 
 def frewrite_includes : Flag<["-"], "frewrite-includes">, Group<f_Group>,

cfe/trunk/include/clang/Frontend/CodeGenOptions.h

public:
   /// non-empty.
   std::string DwarfDebugFlags;
 
   std::map<std::string, std::string> DebugPrefixMap;
 
   /// The ABI to use for passing floating point arguments.
   std::string FloatABI;
 
+  /// The floating-point denormal mode to use.
+  std::string FPDenormalMode;
+
   /// The float precision limit to use, if non-empty.
   std::string LimitFloatPrecision;
 
   /// The name of the bitcode file to link before optzns.
   std::vector<std::pair<unsigned, std::string>> LinkBitcodeFiles;
 
   /// The user provided name for the "main file", if non-empty. This is useful
   /// in situations where the input file name does not match the original input

cfe/trunk/include/clang/Frontend/CodeGenOptions.def

 CODEGENOPT(NoExecStack       , 1, 0) ///< Set when -Wa,--noexecstack is enabled.
 CODEGENOPT(FatalWarnings     , 1, 0) ///< Set when -Wa,--fatal-warnings is
                                      ///< enabled.
 CODEGENOPT(EnableSegmentedStacks , 1, 0) ///< Set when -fsplit-stack is enabled.
 CODEGENOPT(NoImplicitFloat   , 1, 0) ///< Set when -mno-implicit-float is enabled.
 CODEGENOPT(NoInfsFPMath      , 1, 0) ///< Assume FP arguments, results not +-Inf.
 CODEGENOPT(NoSignedZeros     , 1, 0) ///< Allow ignoring the signedness of FP zero
 CODEGENOPT(ReciprocalMath    , 1, 0) ///< Allow FP divisions to be reassociated.
+CODEGENOPT(NoTrappingMath    , 1, 0) ///< Set when -fno-trapping-math is enabled.
 CODEGENOPT(NoInline          , 1, 0) ///< Set when -fno-inline is enabled.
                                      ///< Disables use of the inline keyword.
 CODEGENOPT(NoNaNsFPMath      , 1, 0) ///< Assume FP arguments, results not NaN.
 CODEGENOPT(FlushDenorm       , 1, 0) ///< Allow FP denorm numbers to be flushed to zero
 CODEGENOPT(CorrectlyRoundedDivSqrt, 1, 0) ///< -cl-fp32-correctly-rounded-divide-sqrt
 CODEGENOPT(NoZeroInitializedInBSS , 1, 0) ///< -fno-zero-initialized-in-bss.
 /// \brief Method of Objective-C dispatch to use.
 ENUM_CODEGENOPT(ObjCDispatchMethod, ObjCDispatchMethodKind, 2, Legacy)

cfe/trunk/lib/CodeGen/CGCall.cpp

bool DisableTailCalls =
         CodeGenOpts.DisableTailCalls || HasAnyX86InterruptAttr ||
         (TargetDecl && TargetDecl->hasAttr<DisableTailCallsAttr>());
     FuncAttrs.addAttribute(
         "disable-tail-calls",
         llvm::toStringRef(DisableTailCalls));
 
     FuncAttrs.addAttribute("less-precise-fpmad",
                            llvm::toStringRef(CodeGenOpts.LessPreciseFPMAD));
+
+    if (!CodeGenOpts.FPDenormalMode.empty())
+      FuncAttrs.addAttribute("denormal-fp-math",
+                             CodeGenOpts.FPDenormalMode);
+
+    FuncAttrs.addAttribute("no-trapping-math",
+                           llvm::toStringRef(CodeGenOpts.NoTrappingMath));
     FuncAttrs.addAttribute("no-infs-fp-math",
                            llvm::toStringRef(CodeGenOpts.NoInfsFPMath));
     FuncAttrs.addAttribute("no-nans-fp-math",
                            llvm::toStringRef(CodeGenOpts.NoNaNsFPMath));
     FuncAttrs.addAttribute("unsafe-fp-math",
                            llvm::toStringRef(CodeGenOpts.UnsafeFPMath));
     FuncAttrs.addAttribute("use-soft-float",
                            llvm::toStringRef(CodeGenOpts.SoftFloat));

cfe/trunk/lib/Driver/Tools.cpp

if (!MathErrno && AssociativeMath && ReciprocalMath && !SignedZeros &&
     CmdArgs.push_back("-menable-unsafe-fp-math");
 
   if (!SignedZeros)
     CmdArgs.push_back("-fno-signed-zeros");
 
   if (ReciprocalMath)
     CmdArgs.push_back("-freciprocal-math");
 
+  if (!TrappingMath) 
+    CmdArgs.push_back("-fno-trapping-math");
+
+  if (Args.hasArg(options::OPT_fdenormal_fp_math_EQ))
+    Args.AddLastArg(CmdArgs, options::OPT_fdenormal_fp_math_EQ);
+
   // Validate and pass through -fp-contract option.
   if (Arg *A = Args.getLastArg(options::OPT_ffast_math, FastMathAliasOption,
                                options::OPT_fno_fast_math,
                                options::OPT_ffp_contract)) {
     if (A->getOption().getID() == options::OPT_ffp_contract) {
       StringRef Val = A->getValue();
       if (Val == "fast" || Val == "on" || Val == "off") {
         CmdArgs.push_back(Args.MakeArgString("-ffp-contract=" + Val));

cfe/trunk/lib/Frontend/CompilerInvocation.cpp

Opts.NoNaNsFPMath = (Args.hasArg(OPT_menable_no_nans) ||
                        Args.hasArg(OPT_cl_finite_math_only) ||
                        Args.hasArg(OPT_cl_fast_relaxed_math));
   Opts.NoSignedZeros = (Args.hasArg(OPT_fno_signed_zeros) ||
                         Args.hasArg(OPT_cl_no_signed_zeros));
   Opts.FlushDenorm = Args.hasArg(OPT_cl_denorms_are_zero);
   Opts.CorrectlyRoundedDivSqrt =
       Args.hasArg(OPT_cl_fp32_correctly_rounded_divide_sqrt);
   Opts.ReciprocalMath = Args.hasArg(OPT_freciprocal_math);
+  Opts.NoTrappingMath = Args.hasArg(OPT_fno_trapping_math);
   Opts.NoZeroInitializedInBSS = Args.hasArg(OPT_mno_zero_initialized_in_bss);
   Opts.BackendOptions = Args.getAllArgValues(OPT_backend_option);
   Opts.NumRegisterParameters = getLastArgIntValue(Args, OPT_mregparm, 0, Diags);
   Opts.NoExecStack = Args.hasArg(OPT_mno_exec_stack);
   Opts.FatalWarnings = Args.hasArg(OPT_massembler_fatal_warnings);
   Opts.EnableSegmentedStacks = Args.hasArg(OPT_split_stacks);
   Opts.RelaxAll = Args.hasArg(OPT_mrelax_all);
   Opts.IncrementalLinkerCompatible =
if (Arg *A = Args.getLastArg(OPT_ffp_contract)) {
     else if (Val == "on")
       Opts.setFPContractMode(CodeGenOptions::FPC_On);
     else if (Val == "off")
       Opts.setFPContractMode(CodeGenOptions::FPC_Off);
     else
       Diags.Report(diag::err_drv_invalid_value) << A->getAsString(Args) << Val;
   }
 
+  if (Arg *A = Args.getLastArg(OPT_fdenormal_fp_math_EQ)) {
+    StringRef Val = A->getValue();
+    if (Val == "ieee")
+      Opts.FPDenormalMode = "ieee";
+    else if (Val == "preserve-sign")
+      Opts.FPDenormalMode = "preserve-sign";
+    else if (Val == "positive-zero")
+      Opts.FPDenormalMode = "positive-zero";
+    else
+      Diags.Report(diag::err_drv_invalid_value) << A->getAsString(Args) << Val;
+  }
+
   if (Arg *A = Args.getLastArg(OPT_fpcc_struct_return, OPT_freg_struct_return)) {
     if (A->getOption().matches(OPT_fpcc_struct_return)) {
       Opts.setStructReturnConvention(CodeGenOptions::SRCK_OnStack);
     } else {
       assert(A->getOption().matches(OPT_freg_struct_return));
       Opts.setStructReturnConvention(CodeGenOptions::SRCK_InRegs);
     }
   }

cfe/trunk/test/CodeGen/denormalfpmode.c

+// RUN: %clang_cc1 -S -fdenormal-fp-math=ieee %s -emit-llvm -o - | FileCheck %s --check-prefix=CHECK-IEEE
+// RUN: %clang_cc1 -S -fdenormal-fp-math=preserve-sign %s -emit-llvm -o - | FileCheck %s --check-prefix=CHECK-PS
+// RUN: %clang_cc1 -S -fdenormal-fp-math=positive-zero %s -emit-llvm -o - | FileCheck %s --check-prefix=CHECK-PZ
+
+// CHECK-LABEL: main
+// CHECK-IEEE: attributes #0 = {{.*}}"denormal-fp-math"="ieee"{{.*}}
+// CHECK-PS: attributes #0 = {{.*}}"denormal-fp-math"="preserve-sign"{{.*}}
+// CHECK-PZ: attributes #0 = {{.*}}"denormal-fp-math"="positive-zero"{{.*}}
+
+int main() {
+  return 0;
+}

cfe/trunk/test/CodeGen/noexceptionsfpmath.c

+// RUN: %clang_cc1 -S -fno-trapping-math %s -emit-llvm -o - | FileCheck %s
+
+// CHECK-LABEL: main
+// CHECK: attributes #0 = {{.*}}"no-trapping-math"="true"{{.*}}
+
+int main() {
+  return 0;
+}

cfe/trunk/test/Driver/fast-math.c

 // RUN:   | FileCheck --check-prefix=CHECK-NO-UNSAFE-MATH %s
 // RUN: %clang -### -ffast-math -ftrapping-math -c %s 2>&1 \
 // RUN:   | FileCheck --check-prefix=CHECK-NO-UNSAFE-MATH %s
 // RUN: %clang -### -ffast-math -fno-unsafe-math-optimizations -c %s 2>&1 \
 // RUN:   | FileCheck --check-prefix=CHECK-NO-UNSAFE-MATH %s
 // CHECK-NO-UNSAFE-MATH: "-cc1"
 // CHECK-NO-UNSAFE-MATH-NOT: "-menable-unsafe-fp-math"
 // CHECK-NO-UNSAFE-MATH: "-o"
+//
+// RUN: %clang -### -fexceptions-fp-math -fno-exceptions-fp-math -c %s 2>&1 \
+// RUN:   | FileCheck --check-prefix=CHECK-NO-FP-EXCEPTIONS %s
+// RUN: %clang -### -fno-exceptions-fp-math -fexceptions-fp-math -c %s 2>&1 \
+// RUN:   | FileCheck --check-prefix=CHECK-FP-EXCEPTIONS %s
+// CHECK-NO-FP-EXCEPTIONS: "-fno-exceptions-fp-math"
+// CHECK-FP-EXCEPTIONS-NOT: "-fno-exceptions-fp-math"
+//
+// RUN: %clang -### -fdenormal-fp-math=ieee -c %s 2>&1 \
+// RUN:   | FileCheck --check-prefix=CHECK-FP-DENORMAL-IEEE %s
+// RUN: %clang -### -fdenormal-fp-math=preserve-sign -c %s 2>&1 \
+// RUN:   | FileCheck --check-prefix=CHECK-FP-DENORMAL-PS %s
+// RUN: %clang -### -fdenormal-fp-math=positive-zero -c %s 2>&1 \
+// RUN:   | FileCheck --check-prefix=CHECK-FP-DENORMAL-PZ %s
+// CHECK-FP-DENORMAL-IEEE: "-fdenormal-fp-math=ieee"
+// CHECK-FP-DENORMAL-PS: "-fdenormal-fp-math=preserve-sign"
+// CHECK-FP-DENORMAL-PZ: "-fdenormal-fp-math=positive-zero"