Add movabs -> mov shortening optimization. Add peephole optimization …

…pass that does instruction shortening.

Summary:
Shorten when a mov instruction has a 64-bit immediate that can be repesented as
a sign extended 32-bit number, use the smaller mov instruction (MOV64ri -> MOV64ri32).

Add peephole optimization pass that does instruction shortening.

(cherry picked from FBD3603099)

bolt/BinaryPassManager.cpp

@@ -36,6 +36,12 @@ SimplifyConditionalTailCalls("simplify-conditional-tail-calls",
                                             "by removing unnecessary jumps"),
                              llvm::cl::Optional);
 
+static llvm::cl::opt<bool>
+Peepholes("peepholes",
+          llvm::cl::desc("run peephole optimizations"),
+          llvm::cl::init(true),
+          llvm::cl::Optional);
+
 } // namespace opts
 
 namespace llvm {
@@ -82,6 +88,8 @@ void BinaryFunctionPassManager::runAllPasses(
 
   Manager.registerPass(std::move(llvm::make_unique<FixupFunctions>()));
 
+  Manager.registerPass(llvm::make_unique<Peepholes>(), opts::Peepholes);
+
   Manager.runPasses();
 }
 

bolt/BinaryPasses.cpp

@@ -21,6 +21,7 @@ extern llvm::cl::opt<bool> DumpDotAll;
 extern llvm::cl::opt<bool> PrintReordered;
 extern llvm::cl::opt<bool> PrintEHRanges;
 extern llvm::cl::opt<bool> PrintUCE;
+extern llvm::cl::opt<bool> PrintPeepholes;
 extern llvm::cl::opt<llvm::bolt::BinaryFunction::SplittingType> SplitFunctions;
 extern bool shouldProcess(const llvm::bolt::BinaryFunction &Function);
 
@@ -538,5 +539,33 @@ void SimplifyConditionalTailCalls::runOnFunctions(
          << " from a total of " << NumTailCallCandidates << "\n";
 }
 
+void Peepholes::shortenInstructions(BinaryContext &BC,
+                                    BinaryFunction &Function) {
+  for (auto &BB : Function) {
+    for (auto &Inst : BB) {
+      BC.MIA->shortenInstruction(Inst);
+    }
+  }
+}
+
+void Peepholes::runOnFunctions(BinaryContext &BC,
+                               std::map<uint64_t, BinaryFunction> &BFs,
+                               std::set<uint64_t> &LargeFunctions) {
+  for (auto &It : BFs) {
+    auto &Function = It.second;
+    if (Function.isSimple() && opts::shouldProcess(Function)) {
+      shortenInstructions(BC, Function);
+
+      if (opts::PrintAll || opts::PrintPeepholes) {
+        Function.print(errs(), "after peepholes", true);
+      }
+
+      if (opts::DumpDotAll) {
+        Function.dumpGraphForPass("peepholes");
+      }
+    }
+  }
+}
+
 } // namespace bolt
 } // namespace llvm

bolt/BinaryPasses.h

@@ -148,6 +148,15 @@ class SimplifyConditionalTailCalls : public BinaryFunctionPass {
                       std::set<uint64_t> &LargeFunctions) override;
 };
 
+/// Perform simple peephole optimizations.
+class Peepholes : public BinaryFunctionPass {
+  void shortenInstructions(BinaryContext &BC, BinaryFunction &Function);
+ public:
+  void runOnFunctions(BinaryContext &BC,
+                      std::map<uint64_t, BinaryFunction> &BFs,
+                      std::set<uint64_t> &LargeFunctions) override;
+};
+
 } // namespace bolt
 } // namespace llvm
 

bolt/RewriteInstance.cpp

@@ -157,6 +157,11 @@ PrintUCE("print-uce",
          cl::desc("print functions after unreachable code elimination"),
          cl::Hidden);
 
+cl::opt<bool>
+PrintPeepholes("print-peepholes",
+               cl::desc("print functions after peephole optimization"),
+               cl::Hidden);
+
 static cl::opt<bool>
 PrintDisasm("print-disasm", cl::desc("print function after disassembly"),
             cl::Hidden);