Updated version of r96634 (which was reverted due to failing 176.gcc and

126.gcc nightly tests. These failures uncovered latent bugs that machine DCE
could remove one half of a stack adjust down/up pair, causing PEI to assert.
This update fixes that, and the tests now pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96822 91177308-0d34-0410-b5e6-96231b3b80d8

6 files changed, 34 insertions, 5 deletions

diff --git a/include/llvm/Target/TargetRegisterInfo.h b/include/llvm/Target/TargetRegisterInfo.h
index 65b60f7a93..212cc93401 100644
--- a/include/llvm/Target/TargetRegisterInfo.h
+++ b/include/llvm/Target/TargetRegisterInfo.h
@@ -587,6 +587,17 @@ public:
     return !hasFP(MF);
   }
 
+  /// canSimplifyCallFramePseudos - When possible, it's best to simplify the
+  /// call frame pseudo ops before doing frame index elimination. This is
+  /// possible only when frame index references between the pseudos won't
+  /// need adjusted for the call frame adjustments. Normally, that's true
+  /// if the function has a reserved call frame or a frame pointer. Some
+  /// targets (Thumb2, for example) may have more complicated criteria,
+  /// however, and can override this behavior.
+  virtual bool canSimplifyCallFramePseudos(MachineFunction &MF) const {
+    return hasReservedCallFrame(MF) || hasFP(MF);
+  }
+
   /// hasReservedSpillSlot - Return true if target has reserved a spill slot in
   /// the stack frame of the given function for the specified register. e.g. On
   /// x86, if the frame register is required, the first fixed stack object is
diff --git a/lib/CodeGen/PrologEpilogInserter.cpp b/lib/CodeGen/PrologEpilogInserter.cpp
index 040259e15c..138e711030 100644
--- a/lib/CodeGen/PrologEpilogInserter.cpp
+++ b/lib/CodeGen/PrologEpilogInserter.cpp
@@ -175,9 +175,10 @@ void PEI::calculateCallsInformation(MachineFunction &Fn) {
     MachineBasicBlock::iterator I = *i;
 
     // If call frames are not being included as part of the stack frame, and
-    // there is no dynamic allocation (therefore referencing frame slots off
-    // sp), leave the pseudo ops alone. We'll eliminate them later.
-    if (RegInfo->hasReservedCallFrame(Fn) || RegInfo->hasFP(Fn))
+    // the target doesn't indicate otherwise, remove the call frame pseudos
+    // here. The sub/add sp instruction pairs are still inserted, but we don't
+    // need to track the SP adjustment for frame index elimination.
+    if (RegInfo->canSimplifyCallFramePseudos(Fn))
       RegInfo->eliminateCallFramePseudoInstr(Fn, *I->getParent(), I);
   }
 }
diff --git a/lib/Target/ARM/ARMBaseRegisterInfo.cpp b/lib/Target/ARM/ARMBaseRegisterInfo.cpp
index b1dad39939..361a90fd48 100644
--- a/lib/Target/ARM/ARMBaseRegisterInfo.cpp
+++ b/lib/Target/ARM/ARMBaseRegisterInfo.cpp
@@ -1085,6 +1085,16 @@ hasReservedCallFrame(MachineFunction &MF) const {
   return !MF.getFrameInfo()->hasVarSizedObjects();
 }
 
+// canSimplifyCallFramePseudos - If there is a reserved call frame, the
+// call frame pseudos can be simplified. Unlike most targets, having a FP
+// is not sufficient here since we still may reference some objects via SP
+// even when FP is available in Thumb2 mode.
+bool ARMBaseRegisterInfo::
+canSimplifyCallFramePseudos(MachineFunction &MF) const {
+  ARMFunctionInfo *AFI = MF.getInfo<ARMFunctionInfo>();
+  return hasReservedCallFrame(MF) || (AFI->isThumb1OnlyFunction() && hasFP(MF));
+}
+
 static void
 emitSPUpdate(bool isARM,
              MachineBasicBlock &MBB, MachineBasicBlock::iterator &MBBI,
diff --git a/lib/Target/ARM/ARMBaseRegisterInfo.h b/lib/Target/ARM/ARMBaseRegisterInfo.h
index 33ba21dcb8..64f6ff1cb3 100644
--- a/lib/Target/ARM/ARMBaseRegisterInfo.h
+++ b/lib/Target/ARM/ARMBaseRegisterInfo.h
@@ -138,6 +138,7 @@ public:
   virtual bool requiresFrameIndexScavenging(const MachineFunction &MF) const;
 
   virtual bool hasReservedCallFrame(MachineFunction &MF) const;
+  virtual bool canSimplifyCallFramePseudos(MachineFunction &MF) const;
 
   virtual void eliminateCallFramePseudoInstr(MachineFunction &MF,
                                              MachineBasicBlock &MBB,
diff --git a/lib/Target/ARM/ARMInstrInfo.td b/lib/Target/ARM/ARMInstrInfo.td
index 262695b03c..ca917fa120 100644
--- a/lib/Target/ARM/ARMInstrInfo.td
+++ b/lib/Target/ARM/ARMInstrInfo.td
@@ -635,7 +635,10 @@ PseudoInst<(outs), (ins cpinst_operand:$instid, cpinst_operand:$cpidx,
                     i32imm:$size), NoItinerary,
            "${instid:label} ${cpidx:cpentry}", []>;
 
-let Defs = [SP], Uses = [SP] in {
+// FIXME: Marking these as hasSideEffects is necessary to prevent machine DCE
+// from removing one half of the matched pairs. That breaks PEI, which assumes
+// these will always be in pairs, and asserts if it finds otherwise. Better way?
+let Defs = [SP], Uses = [SP], hasSideEffects = 1 in {
 def ADJCALLSTACKUP :
 PseudoInst<(outs), (ins i32imm:$amt1, i32imm:$amt2, pred:$p), NoItinerary,
            "@ ADJCALLSTACKUP $amt1",
diff --git a/lib/Target/ARM/ARMInstrThumb.td b/lib/Target/ARM/ARMInstrThumb.td
index b1f0459501..e8d3e228ec 100644
--- a/lib/Target/ARM/ARMInstrThumb.td
+++ b/lib/Target/ARM/ARMInstrThumb.td
@@ -120,7 +120,10 @@ def t_addrmode_sp : Operand<i32>,
 //  Miscellaneous Instructions.
 //
 
-let Defs = [SP], Uses = [SP] in {
+// FIXME: Marking these as hasSideEffects is necessary to prevent machine DCE
+// from removing one half of the matched pairs. That breaks PEI, which assumes
+// these will always be in pairs, and asserts if it finds otherwise. Better way?
+let Defs = [SP], Uses = [SP], hasSideEffects = 1 in {
 def tADJCALLSTACKUP :
 PseudoInst<(outs), (ins i32imm:$amt1, i32imm:$amt2), NoItinerary,
            "@ tADJCALLSTACKUP $amt1",