Implement _ExtInt as an extended int type specifier.

Introduction/Motivation:
LLVM-IR supports integers of non-power-of-2 bitwidth, in the iN syntax.
Integers of non-power-of-two aren't particularly interesting or useful
on most hardware, so much so that no language in Clang has been
motivated to expose it before.

However, in the case of FPGA hardware normal integer types where the
full bitwidth isn't used, is extremely wasteful and has severe
performance/space concerns.  Because of this, Intel has introduced this
functionality in the High Level Synthesis compiler[0]
under the name "Arbitrary Precision Integer" (ap_int for short). This
has been extremely useful and effective for our users, permitting them
to optimize their storage and operation space on an architecture where
both can be extremely expensive.

We are proposing upstreaming a more palatable version of this to the
community, in the form of this proposal and accompanying patch.  We are
proposing the syntax _ExtInt(N).  We intend to propose this to the WG14
committee[1], and the underscore-capital seems like the active direction
for a WG14 paper's acceptance.  An alternative that Richard Smith
suggested on the initial review was __int(N), however we believe that
is much less acceptable by WG14.  We considered _Int, however _Int is
used as an identifier in libstdc++ and there is no good way to fall
back to an identifier (since _Int(5) is indistinguishable from an
unnamed initializer of a template type named _Int).

[0]https://www.intel.com/content/www/us/en/software/programmable/quartus-prime/hls-compiler.html)
[1]http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2472.pdf

Differential Revision: https://reviews.llvm.org/D73967

1 files changed, 22 insertions, 4 deletions

diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp
index 97e96941ec2f..343c62273ec4 100644
--- a/clang/lib/CodeGen/CGExprScalar.cpp
+++ b/clang/lib/CodeGen/CGExprScalar.cpp
@@ -760,6 +760,11 @@ public:
                                                   llvm::Value *Zero,bool isDiv);
   // Common helper for getting how wide LHS of shift is.
   static Value *GetWidthMinusOneValue(Value* LHS,Value* RHS);
+
+  // Used for shifting constraints for OpenCL, do mask for powers of 2, URem for
+  // non powers of two.
+  Value *ConstrainShiftValue(Value *LHS, Value *RHS, const Twine &Name);
+
   Value *EmitDiv(const BinOpInfo &Ops);
   Value *EmitRem(const BinOpInfo &Ops);
   Value *EmitAdd(const BinOpInfo &Ops);
@@ -3770,6 +3775,21 @@ Value *ScalarExprEmitter::GetWidthMinusOneValue(Value* LHS,Value* RHS) {
   return llvm::ConstantInt::get(RHS->getType(), Ty->getBitWidth() - 1);
 }
 
+Value *ScalarExprEmitter::ConstrainShiftValue(Value *LHS, Value *RHS,
+                                              const Twine &Name) {
+  llvm::IntegerType *Ty;
+  if (auto *VT = dyn_cast<llvm::VectorType>(LHS->getType()))
+    Ty = cast<llvm::IntegerType>(VT->getElementType());
+  else
+    Ty = cast<llvm::IntegerType>(LHS->getType());
+
+  if (llvm::isPowerOf2_64(Ty->getBitWidth()))
+        return Builder.CreateAnd(RHS, GetWidthMinusOneValue(LHS, RHS), Name);
+
+  return Builder.CreateURem(
+      RHS, llvm::ConstantInt::get(RHS->getType(), Ty->getBitWidth()), Name);
+}
+
 Value *ScalarExprEmitter::EmitShl(const BinOpInfo &Ops) {
   // LLVM requires the LHS and RHS to be the same type: promote or truncate the
   // RHS to the same size as the LHS.
@@ -3784,8 +3804,7 @@ Value *ScalarExprEmitter::EmitShl(const BinOpInfo &Ops) {
   bool SanitizeExponent = CGF.SanOpts.has(SanitizerKind::ShiftExponent);
   // OpenCL 6.3j: shift values are effectively % word size of LHS.
   if (CGF.getLangOpts().OpenCL)
-    RHS =
-        Builder.CreateAnd(RHS, GetWidthMinusOneValue(Ops.LHS, RHS), "shl.mask");
+    RHS = ConstrainShiftValue(Ops.LHS, RHS, "shl.mask");
   else if ((SanitizeBase || SanitizeExponent) &&
            isa<llvm::IntegerType>(Ops.LHS->getType())) {
     CodeGenFunction::SanitizerScope SanScope(&CGF);
@@ -3847,8 +3866,7 @@ Value *ScalarExprEmitter::EmitShr(const BinOpInfo &Ops) {
 
   // OpenCL 6.3j: shift values are effectively % word size of LHS.
   if (CGF.getLangOpts().OpenCL)
-    RHS =
-        Builder.CreateAnd(RHS, GetWidthMinusOneValue(Ops.LHS, RHS), "shr.mask");
+    RHS = ConstrainShiftValue(Ops.LHS, RHS, "shr.mask");
   else if (CGF.SanOpts.has(SanitizerKind::ShiftExponent) &&
            isa<llvm::IntegerType>(Ops.LHS->getType())) {
     CodeGenFunction::SanitizerScope SanScope(&CGF);