From bfb00e3d536e4a907f257684eba7836951677864 Mon Sep 17 00:00:00 2001
From: Jessica Paquette <jpaquette@apple.com>
Date: Mon, 9 Sep 2019 17:15:56 +0000
Subject: [PATCH] [GlobalISel][AArch64] Handle tail calls with non-void return
 types

Just return once you emit the call, which is exactly what SelectionDAG does in
this situation.

Update call-translator-tail-call.ll.

Also update dllimport.ll to show that we tail call here in GISel again. Add
-verify-machineinstrs to the GISel line too, to defend against verifier
failures.

Differential revision: https://reviews.llvm.org/D67282

llvm-svn: 371425
---
 llvm/lib/Target/AArch64/AArch64CallLowering.cpp    | 24 ++++++++--------------
 .../GlobalISel/call-translator-tail-call.ll        |  7 +------
 llvm/test/CodeGen/AArch64/dllimport.ll             |  4 ++--
 3 files changed, 12 insertions(+), 23 deletions(-)

diff --git a/llvm/lib/Target/AArch64/AArch64CallLowering.cpp b/llvm/lib/Target/AArch64/AArch64CallLowering.cpp
index ffca0e7..03f20a2 100644
--- a/llvm/lib/Target/AArch64/AArch64CallLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64CallLowering.cpp
@@ -450,14 +450,6 @@ bool AArch64CallLowering::isEligibleForTailCallOptimization(
     return false;
   }
 
-  if (!Info.OrigRet.Ty->isVoidTy()) {
-    // TODO: lowerCall will insert COPYs to handle the call's return value.
-    // This needs some refactoring to avoid this with tail call returns. For
-    // now, just don't handle that case.
-    LLVM_DEBUG(dbgs() << "... Cannot handle non-void return types yet.\n");
-    return false;
-  }
-
   if (!mayTailCallThisCC(CalleeCC)) {
     LLVM_DEBUG(dbgs() << "... Calling convention cannot be tail called.\n");
     return false;
@@ -649,6 +641,11 @@ bool AArch64CallLowering::lowerCall(MachineIRBuilder &MIRBuilder,
         *MF.getSubtarget().getRegBankInfo(), *MIB, MIB->getDesc(), Info.Callee,
         0));
 
+  // If we're tail calling, then we're the return from the block. So, we don't
+  // want to copy anything.
+  if (IsSibCall)
+    return true;
+
   // Finally we can copy the returned value back into its virtual-register. In
   // symmetry with the arugments, the physical register must be an
   // implicit-define of the call instruction.
@@ -668,13 +665,10 @@ bool AArch64CallLowering::lowerCall(MachineIRBuilder &MIRBuilder,
     MIRBuilder.buildCopy(Info.SwiftErrorVReg, Register(AArch64::X21));
   }
 
-  if (!IsSibCall) {
-    // If we aren't sibcalling, we need to move the stack.
-    CallSeqStart.addImm(Handler.StackSize).addImm(0);
-    MIRBuilder.buildInstr(AArch64::ADJCALLSTACKUP)
-        .addImm(Handler.StackSize)
-        .addImm(0);
-  }
+  CallSeqStart.addImm(Handler.StackSize).addImm(0);
+  MIRBuilder.buildInstr(AArch64::ADJCALLSTACKUP)
+      .addImm(Handler.StackSize)
+      .addImm(0);
 
   return true;
 }
diff --git a/llvm/test/CodeGen/AArch64/GlobalISel/call-translator-tail-call.ll b/llvm/test/CodeGen/AArch64/GlobalISel/call-translator-tail-call.ll
index 120bead..fb937c3 100644
--- a/llvm/test/CodeGen/AArch64/GlobalISel/call-translator-tail-call.ll
+++ b/llvm/test/CodeGen/AArch64/GlobalISel/call-translator-tail-call.ll
@@ -47,12 +47,7 @@ declare i32 @nonvoid_ret()
 define i32 @test_nonvoid_ret() {
   ; COMMON-LABEL: name: test_nonvoid_ret
   ; COMMON: bb.1 (%ir-block.0):
-  ; COMMON:   ADJCALLSTACKDOWN 0, 0, implicit-def $sp, implicit $sp
-  ; COMMON:   BL @nonvoid_ret, csr_aarch64_aapcs, implicit-def $lr, implicit $sp, implicit-def $w0
-  ; COMMON:   [[COPY:%[0-9]+]]:_(s32) = COPY $w0
-  ; COMMON:   ADJCALLSTACKUP 0, 0, implicit-def $sp, implicit $sp
-  ; COMMON:   $w0 = COPY [[COPY]](s32)
-  ; COMMON:   RET_ReallyLR implicit $w0
+  ; COMMON:   TCRETURNdi @nonvoid_ret, 0, csr_aarch64_aapcs, implicit $sp
   %call = tail call i32 @nonvoid_ret()
   ret i32 %call
 }
diff --git a/llvm/test/CodeGen/AArch64/dllimport.ll b/llvm/test/CodeGen/AArch64/dllimport.ll
index 281c847..61db9ea 100644
--- a/llvm/test/CodeGen/AArch64/dllimport.ll
+++ b/llvm/test/CodeGen/AArch64/dllimport.ll
@@ -1,6 +1,6 @@
 ; RUN: llc -mtriple aarch64-unknown-windows-msvc -filetype asm -o - %s | FileCheck %s -check-prefixes=CHECK,DAG-ISEL
 ; RUN: llc -mtriple aarch64-unknown-windows-msvc -fast-isel -filetype asm -o - %s | FileCheck %s -check-prefixes=CHECK,FAST-ISEL
-; RUN: llc -mtriple aarch64-unknown-windows-msvc -O0 -filetype asm -o - %s | FileCheck %s -check-prefixes=CHECK,GLOBAL-ISEL,GLOBAL-ISEL-FALLBACK
+; RUN: llc -mtriple aarch64-unknown-windows-msvc -verify-machineinstrs -O0 -filetype asm -o - %s | FileCheck %s -check-prefixes=CHECK,GLOBAL-ISEL,GLOBAL-ISEL-FALLBACK
 
 @var = external dllimport global i32
 @ext = external global i32
@@ -59,4 +59,4 @@ define i32 @call_internal() {
 ; CHECK-LABEL: call_internal
 ; DAG-ISEL: b internal
 ; FAST-ISEL: b internal
-; GLOBAL-ISEL: bl internal
+; GLOBAL-ISEL: b internal
-- 
2.7.4