Merged master:bb26838ceffb into amd-gfx:c8f92d00575c

Local branch amd-gfx c8f92d0 Merged master:1b10c618e928 into amd-gfx:8504aad041f6
Remote branch master bb26838 [clang driver] Move default module cache from system temporary directory

clang/docs/ReleaseNotes.rst

@@ -104,6 +104,10 @@ New Compiler Flags
   simplify access to the many single purpose floating point options. The default
   setting is ``precise``.
 
+- The default module cache has moved from /tmp to a per-user cache directory.
+  By default, this is ~/.cache but on some platforms or installations, this
+  might be elsewhere. The -fmodules-cache-path=... flag continues to work.
+
 Deprecated Compiler Flags
 -------------------------
 

clang/include/clang/Basic/DiagnosticASTKinds.td

@@ -346,6 +346,9 @@ def err_experimental_clang_interp_failed : Error<
 def warn_integer_constant_overflow : Warning<
   "overflow in expression; result is %0 with type %1">,
   InGroup<DiagGroup<"integer-overflow">>;
+def warn_fixedpoint_constant_overflow : Warning<
+  "overflow in expression; result is %0 with type %1">,
+  InGroup<DiagGroup<"fixed-point-overflow">>;
 
 // This is a temporary diagnostic, and shall be removed once our
 // implementation is complete, and like the preceding constexpr notes belongs

clang/include/clang/Basic/FixedPoint.h

@@ -28,8 +28,8 @@ class QualType;
 /// The fixed point semantics work similarly to llvm::fltSemantics. The width
 /// specifies the whole bit width of the underlying scaled integer (with padding
 /// if any). The scale represents the number of fractional bits in this type.
-/// When HasUnsignedPadding is true and this type is signed, the first bit
-/// in the value this represents is treaded as padding.
+/// When HasUnsignedPadding is true and this type is unsigned, the first bit
+/// in the value this represents is treated as padding.
 class FixedPointSemantics {
 public:
   FixedPointSemantics(unsigned Width, unsigned Scale, bool IsSigned,
@@ -125,9 +125,12 @@ class APFixedPoint {
 
    // Perform binary operations on a fixed point type. The resulting fixed point
    // value will be in the common, full precision semantics that can represent
-   // the precision and ranges os both input values. See convert() for an
+   // the precision and ranges of both input values. See convert() for an
    // explanation of the Overflow parameter.
    APFixedPoint add(const APFixedPoint &Other, bool *Overflow = nullptr) const;
+   APFixedPoint sub(const APFixedPoint &Other, bool *Overflow = nullptr) const;
+   APFixedPoint mul(const APFixedPoint &Other, bool *Overflow = nullptr) const;
+   APFixedPoint div(const APFixedPoint &Other, bool *Overflow = nullptr) const;
 
    /// Perform a unary negation (-X) on this fixed point type, taking into
    /// account saturation if applicable.

clang/include/clang/Basic/arm_sve.td

@@ -563,6 +563,11 @@ def SVST1H_VNUM_U : MInst<"svst1h_vnum[_{d}]", "vPFld", "UiUl",            [IsSt
 def SVST1W_VNUM_S : MInst<"svst1w_vnum[_{d}]", "vPCld", "l",               [IsStore], MemEltTyInt32,   "aarch64_sve_st1">;
 def SVST1W_VNUM_U : MInst<"svst1w_vnum[_{d}]", "vPGld", "Ul",              [IsStore], MemEltTyInt32,   "aarch64_sve_st1">;
 
+let ArchGuard = "defined(__ARM_FEATURE_SVE_BF16)" in {
+  def SVST1_BF      : MInst<"svst1[_{d}]",      "vPpd",  "b", [IsStore], MemEltTyDefault, "aarch64_sve_st1">;
+  def SVST1_VNUM_BF : MInst<"svst1_vnum[_{d}]", "vPpld", "b", [IsStore], MemEltTyDefault, "aarch64_sve_st1">;
+}
+
 // Store one vector (vector base)
 def SVST1_SCATTER_BASES_U     : MInst<"svst1_scatter[_{2}base_{d}]",  "vPud",  "ilUiUlfd", [IsScatterStore], MemEltTyDefault, "aarch64_sve_st1_scatter_scalar_offset">;
 def SVST1B_SCATTER_BASES_U    : MInst<"svst1b_scatter[_{2}base_{d}]", "vPud",  "ilUiUl",   [IsScatterStore], MemEltTyInt8,    "aarch64_sve_st1_scatter_scalar_offset">;
@@ -654,6 +659,11 @@ def SVSTNT1 : MInst<"svstnt1[_{d}]", "vPpd", "csilUcUsUiUlhfd", [IsStore], MemEl
 // Store one vector, with no truncation, non-temporal (scalar base, VL displacement)
 def SVSTNT1_VNUM : MInst<"svstnt1_vnum[_{d}]", "vPpld", "csilUcUsUiUlhfd", [IsStore], MemEltTyDefault, "aarch64_sve_stnt1">;
 
+let ArchGuard = "defined(__ARM_FEATURE_SVE_BF16)" in {
+  def SVSTNT1_BF      : MInst<"svstnt1[_{d}]",      "vPpd",  "b", [IsStore], MemEltTyDefault, "aarch64_sve_stnt1">;
+  def SVSTNT1_VNUM_BF : MInst<"svstnt1_vnum[_{d}]", "vPpld", "b", [IsStore], MemEltTyDefault, "aarch64_sve_stnt1">;
+}
+
 ////////////////////////////////////////////////////////////////////////////////
 // Prefetches
 
@@ -1971,7 +1981,7 @@ def SVWHILEWR_S : SInst<"svwhilewr[_{1}]", "Pcc", "iUif", MergeNone, "aarch64_sv
 def SVWHILEWR_D : SInst<"svwhilewr[_{1}]", "Pcc", "lUld", MergeNone, "aarch64_sve_whilewr_d", [IsOverloadWhileRW]>;
 }
 
-let ArchGuard = "defined(__ARM_FEATURE_SVE2) && defined(__ARM_FEATURE_SVE_BF16)" in {
+let ArchGuard = "defined(__ARM_FEATURE_SVE2) && defined(__ARM_FEATURE_BF16_SCALAR_ARITHMETIC)" in {
 def SVWHILERW_H_BF16 : SInst<"svwhilerw[_{1}]", "Pcc", "b", MergeNone, "aarch64_sve_whilerw_h", [IsOverloadWhileRW]>;
 def SVWHILEWR_H_BF16 : SInst<"svwhilewr[_{1}]", "Pcc", "b", MergeNone, "aarch64_sve_whilewr_h", [IsOverloadWhileRW]>;
 }

clang/include/clang/Driver/Driver.h

@@ -621,7 +621,8 @@ class Driver {
   static bool GetReleaseVersion(StringRef Str,
                                 MutableArrayRef<unsigned> Digits);
   /// Compute the default -fmodule-cache-path.
-  static void getDefaultModuleCachePath(SmallVectorImpl<char> &Result);
+  /// \return True if the system provides a default cache directory.
+  static bool getDefaultModuleCachePath(SmallVectorImpl<char> &Result);
 };
 
 /// \return True if the last defined optimization level is -Ofast.

clang/lib/AST/ExprConstant.cpp

@@ -12881,8 +12881,14 @@ bool FixedPointExprEvaluator::VisitCastExpr(const CastExpr *E) {
       return false;
     bool Overflowed;
     APFixedPoint Result = Src.convert(DestFXSema, &Overflowed);
-    if (Overflowed && !HandleOverflow(Info, E, Result, DestType))
-      return false;
+    if (Overflowed) {
+      if (Info.checkingForUndefinedBehavior())
+        Info.Ctx.getDiagnostics().Report(E->getExprLoc(),
+                                         diag::warn_fixedpoint_constant_overflow)
+          << Result.toString() << E->getType();
+      else if (!HandleOverflow(Info, E, Result, E->getType()))
+        return false;
+    }
     return Success(Result, E);
   }
   case CK_IntegralToFixedPoint: {
@@ -12894,8 +12900,14 @@ bool FixedPointExprEvaluator::VisitCastExpr(const CastExpr *E) {
     APFixedPoint IntResult = APFixedPoint::getFromIntValue(
         Src, Info.Ctx.getFixedPointSemantics(DestType), &Overflowed);
 
-    if (Overflowed && !HandleOverflow(Info, E, IntResult, DestType))
-      return false;
+    if (Overflowed) {
+      if (Info.checkingForUndefinedBehavior())
+        Info.Ctx.getDiagnostics().Report(E->getExprLoc(),
+                                         diag::warn_fixedpoint_constant_overflow)
+          << IntResult.toString() << E->getType();
+      else if (!HandleOverflow(Info, E, IntResult, E->getType()))
+        return false;
+    }
 
     return Success(IntResult, E);
   }
@@ -12908,6 +12920,9 @@ bool FixedPointExprEvaluator::VisitCastExpr(const CastExpr *E) {
 }
 
 bool FixedPointExprEvaluator::VisitBinaryOperator(const BinaryOperator *E) {
+  if (E->isPtrMemOp() || E->isAssignmentOp() || E->getOpcode() == BO_Comma)
+    return ExprEvaluatorBaseTy::VisitBinaryOperator(E);
+
   const Expr *LHS = E->getLHS();
   const Expr *RHS = E->getRHS();
   FixedPointSemantics ResultFXSema =
@@ -12920,20 +12935,41 @@ bool FixedPointExprEvaluator::VisitBinaryOperator(const BinaryOperator *E) {
   if (!EvaluateFixedPointOrInteger(RHS, RHSFX, Info))
     return false;
 
+  bool OpOverflow = false, ConversionOverflow = false;
+  APFixedPoint Result(LHSFX.getSemantics());
   switch (E->getOpcode()) {
   case BO_Add: {
-    bool AddOverflow, ConversionOverflow;
-    APFixedPoint Result = LHSFX.add(RHSFX, &AddOverflow)
-                              .convert(ResultFXSema, &ConversionOverflow);
-    if ((AddOverflow || ConversionOverflow) &&
-        !HandleOverflow(Info, E, Result, E->getType()))
-      return false;
-    return Success(Result, E);
+    Result = LHSFX.add(RHSFX, &OpOverflow)
+                  .convert(ResultFXSema, &ConversionOverflow);
+    break;
+  }
+  case BO_Sub: {
+    Result = LHSFX.sub(RHSFX, &OpOverflow)
+                  .convert(ResultFXSema, &ConversionOverflow);
+    break;
+  }
+  case BO_Mul: {
+    Result = LHSFX.mul(RHSFX, &OpOverflow)
+                  .convert(ResultFXSema, &ConversionOverflow);
+    break;
+  }
+  case BO_Div: {
+    Result = LHSFX.div(RHSFX, &OpOverflow)
+                  .convert(ResultFXSema, &ConversionOverflow);
+    break;
   }
   default:
     return false;
   }
-  llvm_unreachable("Should've exited before this");
+  if (OpOverflow || ConversionOverflow) {
+    if (Info.checkingForUndefinedBehavior())
+      Info.Ctx.getDiagnostics().Report(E->getExprLoc(),
+                                       diag::warn_fixedpoint_constant_overflow)
+        << Result.toString() << E->getType();
+    else if (!HandleOverflow(Info, E, Result, E->getType()))
+      return false;
+  }
+  return Success(Result, E);
 }
 
 //===----------------------------------------------------------------------===//

clang/lib/Basic/FixedPoint.cpp

@@ -173,6 +173,142 @@ APFixedPoint APFixedPoint::add(const APFixedPoint &Other,
   return APFixedPoint(Result, CommonFXSema);
 }
 
+APFixedPoint APFixedPoint::sub(const APFixedPoint &Other,
+                               bool *Overflow) const {
+  auto CommonFXSema = Sema.getCommonSemantics(Other.getSemantics());
+  APFixedPoint ConvertedThis = convert(CommonFXSema);
+  APFixedPoint ConvertedOther = Other.convert(CommonFXSema);
+  llvm::APSInt ThisVal = ConvertedThis.getValue();
+  llvm::APSInt OtherVal = ConvertedOther.getValue();
+  bool Overflowed = false;
+
+  llvm::APSInt Result;
+  if (CommonFXSema.isSaturated()) {
+    Result = CommonFXSema.isSigned() ? ThisVal.ssub_sat(OtherVal)
+                                     : ThisVal.usub_sat(OtherVal);
+  } else {
+    Result = ThisVal.isSigned() ? ThisVal.ssub_ov(OtherVal, Overflowed)
+                                : ThisVal.usub_ov(OtherVal, Overflowed);
+  }
+
+  if (Overflow)
+    *Overflow = Overflowed;
+
+  return APFixedPoint(Result, CommonFXSema);
+}
+
+APFixedPoint APFixedPoint::mul(const APFixedPoint &Other,
+                               bool *Overflow) const {
+  auto CommonFXSema = Sema.getCommonSemantics(Other.getSemantics());
+  APFixedPoint ConvertedThis = convert(CommonFXSema);
+  APFixedPoint ConvertedOther = Other.convert(CommonFXSema);
+  llvm::APSInt ThisVal = ConvertedThis.getValue();
+  llvm::APSInt OtherVal = ConvertedOther.getValue();
+  bool Overflowed = false;
+
+  // Widen the LHS and RHS so we can perform a full multiplication.
+  unsigned Wide = CommonFXSema.getWidth() * 2;
+  if (CommonFXSema.isSigned()) {
+    ThisVal = ThisVal.sextOrSelf(Wide);
+    OtherVal = OtherVal.sextOrSelf(Wide);
+  } else {
+    ThisVal = ThisVal.zextOrSelf(Wide);
+    OtherVal = OtherVal.zextOrSelf(Wide);
+  }
+
+  // Perform the full multiplication and downscale to get the same scale.
+  //
+  // Note that the right shifts here perform an implicit downwards rounding.
+  // This rounding could discard bits that would technically place the result
+  // outside the representable range. We interpret the spec as allowing us to
+  // perform the rounding step first, avoiding the overflow case that would
+  // arise.
+  llvm::APSInt Result;
+  if (CommonFXSema.isSigned())
+    Result = ThisVal.smul_ov(OtherVal, Overflowed)
+                    .ashr(CommonFXSema.getScale());
+  else
+    Result = ThisVal.umul_ov(OtherVal, Overflowed)
+                    .lshr(CommonFXSema.getScale());
+  assert(!Overflowed && "Full multiplication cannot overflow!");
+  Result.setIsSigned(CommonFXSema.isSigned());
+
+  // If our result lies outside of the representative range of the common
+  // semantic, we either have overflow or saturation.
+  llvm::APSInt Max = APFixedPoint::getMax(CommonFXSema).getValue()
+                                                       .extOrTrunc(Wide);
+  llvm::APSInt Min = APFixedPoint::getMin(CommonFXSema).getValue()
+                                                       .extOrTrunc(Wide);
+  if (CommonFXSema.isSaturated()) {
+    if (Result < Min)
+      Result = Min;
+    else if (Result > Max)
+      Result = Max;
+  } else
+    Overflowed = Result < Min || Result > Max;
+
+  if (Overflow)
+    *Overflow = Overflowed;
+
+  return APFixedPoint(Result.sextOrTrunc(CommonFXSema.getWidth()),
+                      CommonFXSema);
+}
+
+APFixedPoint APFixedPoint::div(const APFixedPoint &Other,
+                               bool *Overflow) const {
+  auto CommonFXSema = Sema.getCommonSemantics(Other.getSemantics());
+  APFixedPoint ConvertedThis = convert(CommonFXSema);
+  APFixedPoint ConvertedOther = Other.convert(CommonFXSema);
+  llvm::APSInt ThisVal = ConvertedThis.getValue();
+  llvm::APSInt OtherVal = ConvertedOther.getValue();
+  bool Overflowed = false;
+
+  // Widen the LHS and RHS so we can perform a full division.
+  unsigned Wide = CommonFXSema.getWidth() * 2;
+  if (CommonFXSema.isSigned()) {
+    ThisVal = ThisVal.sextOrSelf(Wide);
+    OtherVal = OtherVal.sextOrSelf(Wide);
+  } else {
+    ThisVal = ThisVal.zextOrSelf(Wide);
+    OtherVal = OtherVal.zextOrSelf(Wide);
+  }
+
+  // Upscale to compensate for the loss of precision from division, and
+  // perform the full division.
+  ThisVal = ThisVal.shl(CommonFXSema.getScale());
+  llvm::APSInt Result;
+  if (CommonFXSema.isSigned()) {
+    llvm::APInt Rem;
+    llvm::APInt::sdivrem(ThisVal, OtherVal, Result, Rem);
+    // If the quotient is negative and the remainder is nonzero, round
+    // towards negative infinity by subtracting epsilon from the result.
+    if (Result.isNegative() && !Rem.isNullValue())
+      Result = Result - 1;
+  } else
+    Result = ThisVal.udiv(OtherVal);
+  Result.setIsSigned(CommonFXSema.isSigned());
+
+  // If our result lies outside of the representative range of the common
+  // semantic, we either have overflow or saturation.
+  llvm::APSInt Max = APFixedPoint::getMax(CommonFXSema).getValue()
+                                                       .extOrTrunc(Wide);
+  llvm::APSInt Min = APFixedPoint::getMin(CommonFXSema).getValue()
+                                                       .extOrTrunc(Wide);
+  if (CommonFXSema.isSaturated()) {
+    if (Result < Min)
+      Result = Min;
+    else if (Result > Max)
+      Result = Max;
+  } else
+    Overflowed = Result < Min || Result > Max;
+
+  if (Overflow)
+    *Overflow = Overflowed;
+
+  return APFixedPoint(Result.sextOrTrunc(CommonFXSema.getWidth()),
+                      CommonFXSema);
+}
+
 void APFixedPoint::toString(llvm::SmallVectorImpl<char> &Str) const {
   llvm::APSInt Val = getValue();
   unsigned Scale = getScale();