pr6652: Use LDM to restore PC to the return address on ARMv4.

Patch by John Tytgat!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@99096 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/ARM/ARMLoadStoreOptimizer.cpp b/lib/Target/ARM/ARMLoadStoreOptimizer.cpp
index d0a3179b996f26e72d1701231b6293b8d072d073..bdbec30d859691ed40ede579ff6102b9135e38ef 100644 (file)
--- a/lib/Target/ARM/ARMLoadStoreOptimizer.cpp
+++ b/lib/Target/ARM/ARMLoadStoreOptimizer.cpp
@@ -1155,19 +1155,24 @@ namespace {
   };
 }
 
-/// MergeReturnIntoLDM - If this is a exit BB, try merging the return op
-/// (bx lr) into the preceeding stack restore so it directly restore the value
-/// of LR into pc.
-///   ldmfd sp!, {r7, lr}
+/// MergeReturnIntoLDM - If this is a exit BB, try merging the return ops
+/// ("bx lr" and "mov pc, lr") into the preceeding stack restore so it
+/// directly restore the value of LR into pc.
+///   ldmfd sp!, {..., lr}
 ///   bx lr
+/// or
+///   ldmfd sp!, {..., lr}
+///   mov pc, lr
 /// =>
-///   ldmfd sp!, {r7, pc}
+///   ldmfd sp!, {..., pc}
 bool ARMLoadStoreOpt::MergeReturnIntoLDM(MachineBasicBlock &MBB) {
   if (MBB.empty()) return false;
 
   MachineBasicBlock::iterator MBBI = prior(MBB.end());
   if (MBBI != MBB.begin() &&
-      (MBBI->getOpcode() == ARM::BX_RET || MBBI->getOpcode() == ARM::tBX_RET)) {
+      (MBBI->getOpcode() == ARM::BX_RET ||
+       MBBI->getOpcode() == ARM::tBX_RET ||
+       MBBI->getOpcode() == ARM::MOVPCLR)) {
     MachineInstr *PrevMI = prior(MBBI);
     if (PrevMI->getOpcode() == ARM::LDM_UPD ||
         PrevMI->getOpcode() == ARM::t2LDM_UPD) {

diff --git a/test/CodeGen/ARM/2010-03-18-ldm-rtrn.ll b/test/CodeGen/ARM/2010-03-18-ldm-rtrn.ll
new file mode 100644 (file)
index 0000000..31525ef
--- /dev/null
+++ b/test/CodeGen/ARM/2010-03-18-ldm-rtrn.ll
@@ -0,0 +1,13 @@
+; RUN: llc < %s -mtriple=armv4-unknown-eabi | FileCheck %s
+; RUN: llc < %s -mtriple=armv5-unknown-eabi | FileCheck %s
+; RUN: llc < %s -mtriple=armv6-unknown-eabi | FileCheck %s
+
+define i32 @bar(i32 %a) nounwind {
+entry:
+  %0 = tail call i32 @foo(i32 %a) nounwind ; <i32> [#uses=1]
+  %1 = add nsw i32 %0, 3                          ; <i32> [#uses=1]
+; CHECK: ldmia sp!, {r11, pc}
+  ret i32 %1
+}
+
+declare i32 @foo(i32)