From 7cdb10f1aabfe57b3e1dcb1a8e5a064b1c7343dc Mon Sep 17 00:00:00 2001
From: Sriraman Tallam <tmsriram@google.com>
Date: Fri, 3 Nov 2017 00:10:19 +0000
Subject: [PATCH] Avoid PLT for external calls when attribute nonlazybind is
 used.

Differential Revision: https://reviews.llvm.org/D39065

llvm-svn: 317292
---
 llvm/lib/Target/X86/X86Subtarget.cpp | 11 +++++++++--
 llvm/test/CodeGen/X86/no-plt.ll      | 23 +++++++++++++++++++++++
 2 files changed, 32 insertions(+), 2 deletions(-)
 create mode 100644 llvm/test/CodeGen/X86/no-plt.ll

diff --git a/llvm/lib/Target/X86/X86Subtarget.cpp b/llvm/lib/Target/X86/X86Subtarget.cpp
index b0ce133..9e060f9 100644
--- a/llvm/lib/Target/X86/X86Subtarget.cpp
+++ b/llvm/lib/Target/X86/X86Subtarget.cpp
@@ -144,6 +144,15 @@ X86Subtarget::classifyGlobalFunctionReference(const GlobalValue *GV) const {
 unsigned char
 X86Subtarget::classifyGlobalFunctionReference(const GlobalValue *GV,
                                               const Module &M) const {
+  const Function *F = dyn_cast_or_null<Function>(GV);
+
+  // Do not use the PLT when explicitly told to do so for ELF 64-bit
+  // target.
+  if (isTargetELF() && is64Bit() && F &&
+      F->hasFnAttribute(Attribute::NonLazyBind) &&
+      GV->isDeclarationForLinker())
+    return X86II::MO_GOTPCREL;
+
   if (TM.shouldAssumeDSOLocal(M, GV))
     return X86II::MO_NO_FLAG;
 
@@ -153,8 +162,6 @@ X86Subtarget::classifyGlobalFunctionReference(const GlobalValue *GV,
     return X86II::MO_DLLIMPORT;
   }
 
-  const Function *F = dyn_cast_or_null<Function>(GV);
-
   if (isTargetELF()) {
     if (is64Bit() && F && (CallingConv::X86_RegCall == F->getCallingConv()))
       // According to psABI, PLT stub clobbers XMM8-XMM15.
diff --git a/llvm/test/CodeGen/X86/no-plt.ll b/llvm/test/CodeGen/X86/no-plt.ll
new file mode 100644
index 0000000..77ef686
--- /dev/null
+++ b/llvm/test/CodeGen/X86/no-plt.ll
@@ -0,0 +1,23 @@
+; RUN: llc < %s -mcpu=generic -mtriple=x86_64-linux-gnu -relocation-model=pic \
+; RUN:   | FileCheck -check-prefix=X64 %s
+; RUN: llc < %s -mcpu=generic -mtriple=x86_64-linux-gnu \
+; RUN:   | FileCheck -check-prefix=X64 %s
+
+define i32 @main() #0 {
+; X64: callq *_Z3foov@GOTPCREL(%rip)
+; X64: callq _Z3barv
+
+entry:
+  %retval = alloca i32, align 4
+  store i32 0, i32* %retval, align 4
+  %call1 = call i32 @_Z3foov()
+  %call2 = call i32 @_Z3barv()
+  ret i32 0
+}
+
+; Function Attrs: nonlazybind
+declare i32 @_Z3foov() #1
+
+declare i32 @_Z3barv() #2
+
+attributes #1 = { nonlazybind }
-- 
2.7.4