Add extra operand to CALLSEQ_START to keep frame part set up previously

Using arguments with attribute inalloca creates problems for verification
of machine representation. This attribute instructs the backend that the
argument is prepared in stack prior to  CALLSEQ_START..CALLSEQ_END
sequence (see http://llvm.org/docs/InAlloca.htm for details). Frame size
stored in CALLSEQ_START in this case does not count the size of this
argument. However CALLSEQ_END still keeps total frame size, as caller can
be responsible for cleanup of entire frame. So CALLSEQ_START and
CALLSEQ_END keep different frame size and the difference is treated by
MachineVerifier as stack error. Currently there is no way to distinguish
this case from actual errors.

This patch adds additional argument to CALLSEQ_START and its
target-specific counterparts to keep size of stack that is set up prior to
the call frame sequence. This argument allows MachineVerifier to calculate
actual frame size associated with frame setup instruction and correctly
process the case of inalloca arguments.

The changes made by the patch are:
- Frame setup instructions get the second mandatory argument. It
  affects all targets that use frame pseudo instructions and touched many
  files although the changes are uniform.
- Access to frame properties are implemented using special instructions
  rather than calls getOperand(N).getImm(). For X86 and ARM such
  replacement was made previously.
- Changes that reflect appearance of additional argument of frame setup
  instruction. These involve proper instruction initialization and
  methods that access instruction arguments.
- MachineVerifier retrieves frame size using method, which reports sum of
  frame parts initialized inside frame instruction pair and outside it.

The patch implements approach proposed by Quentin Colombet in
https://bugs.llvm.org/show_bug.cgi?id=27481#c1.
It fixes 9 tests failed with machine verifier enabled and listed
in PR27481.

Differential Revision: https://reviews.llvm.org/D32394

llvm-svn: 302527

4 files changed, 18 insertions, 18 deletions

diff --git a/llvm/lib/Target/PowerPC/PPCFastISel.cpp b/llvm/lib/Target/PowerPC/PPCFastISel.cpp
index 9b91b9ab8f8..5230e729b7b 100644
--- a/llvm/lib/Target/PowerPC/PPCFastISel.cpp
+++ b/llvm/lib/Target/PowerPC/PPCFastISel.cpp
@@ -1330,7 +1330,7 @@ bool PPCFastISel::processCallArgs(SmallVectorImpl<Value*> &Args,
   // Issue CALLSEQ_START.
   BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
           TII.get(TII.getCallFrameSetupOpcode()))
-    .addImm(NumBytes);
+    .addImm(NumBytes).addImm(0);
 
   // Prepare to assign register arguments.  Every argument uses up a
   // GPR protocol register even if it's passed in a floating-point
diff --git a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
index 685f24cb502..46cd75bc9b2 100644
--- a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
+++ b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
@@ -4949,8 +4949,7 @@ SDValue PPCTargetLowering::LowerCall_32SVR4(
 
   // Adjust the stack pointer for the new arguments...
   // These operations are automatically eliminated by the prolog/epilog pass
-  Chain = DAG.getCALLSEQ_START(Chain, DAG.getIntPtrConstant(NumBytes, dl, true),
-                               dl);
+  Chain = DAG.getCALLSEQ_START(Chain, NumBytes, 0, dl);
   SDValue CallSeqStart = Chain;
 
   // Load the return address and frame pointer so it can be moved somewhere else
@@ -5000,9 +4999,8 @@ SDValue PPCTargetLowering::LowerCall_32SVR4(
                                   Flags, DAG, dl);
 
       // This must go outside the CALLSEQ_START..END.
-      SDValue NewCallSeqStart = DAG.getCALLSEQ_START(MemcpyCall,
-                           CallSeqStart.getNode()->getOperand(1),
-                           SDLoc(MemcpyCall));
+      SDValue NewCallSeqStart = DAG.getCALLSEQ_START(MemcpyCall, NumBytes, 0,
+                                                     SDLoc(MemcpyCall));
       DAG.ReplaceAllUsesWith(CallSeqStart.getNode(),
                              NewCallSeqStart.getNode());
       Chain = CallSeqStart = NewCallSeqStart;
@@ -5083,9 +5081,9 @@ SDValue PPCTargetLowering::createMemcpyOutsideCallSeq(
                         CallSeqStart.getNode()->getOperand(0),
                         Flags, DAG, dl);
   // The MEMCPY must go outside the CALLSEQ_START..END.
-  SDValue NewCallSeqStart = DAG.getCALLSEQ_START(MemcpyCall,
-                             CallSeqStart.getNode()->getOperand(1),
-                             SDLoc(MemcpyCall));
+  int64_t FrameSize = CallSeqStart.getConstantOperandVal(1);
+  SDValue NewCallSeqStart = DAG.getCALLSEQ_START(MemcpyCall, FrameSize, 0,
+                                                 SDLoc(MemcpyCall));
   DAG.ReplaceAllUsesWith(CallSeqStart.getNode(),
                          NewCallSeqStart.getNode());
   return NewCallSeqStart;
@@ -5268,8 +5266,7 @@ SDValue PPCTargetLowering::LowerCall_64SVR4(
   // Adjust the stack pointer for the new arguments...
   // These operations are automatically eliminated by the prolog/epilog pass
   if (!IsSibCall)
-    Chain = DAG.getCALLSEQ_START(Chain,
-                                 DAG.getIntPtrConstant(NumBytes, dl, true), dl);
+    Chain = DAG.getCALLSEQ_START(Chain, NumBytes, 0, dl);
   SDValue CallSeqStart = Chain;
 
   // Load the return address and frame pointer so it can be move somewhere else
@@ -5828,8 +5825,7 @@ SDValue PPCTargetLowering::LowerCall_Darwin(
 
   // Adjust the stack pointer for the new arguments...
   // These operations are automatically eliminated by the prolog/epilog pass
-  Chain = DAG.getCALLSEQ_START(Chain, DAG.getIntPtrConstant(NumBytes, dl, true),
-                               dl);
+  Chain = DAG.getCALLSEQ_START(Chain, NumBytes, 0, dl);
   SDValue CallSeqStart = Chain;
 
   // Load the return address and frame pointer so it can be move somewhere else
diff --git a/llvm/lib/Target/PowerPC/PPCInstrInfo.td b/llvm/lib/Target/PowerPC/PPCInstrInfo.td
index f004ce49cac..7064776efcb 100644
--- a/llvm/lib/Target/PowerPC/PPCInstrInfo.td
+++ b/llvm/lib/Target/PowerPC/PPCInstrInfo.td
@@ -33,7 +33,8 @@ def SDT_PPCVexts  : SDTypeProfile<1, 2, [
   SDTCisVT<0, f64>, SDTCisVT<1, f64>, SDTCisPtrTy<2>
 ]>;
 
-def SDT_PPCCallSeqStart : SDCallSeqStart<[ SDTCisVT<0, i32> ]>;
+def SDT_PPCCallSeqStart : SDCallSeqStart<[ SDTCisVT<0, i32>,
+                                           SDTCisVT<1, i32> ]>;
 def SDT_PPCCallSeqEnd   : SDCallSeqEnd<[ SDTCisVT<0, i32>,
                                          SDTCisVT<1, i32> ]>;
 def SDT_PPCvperm   : SDTypeProfile<1, 3, [
@@ -1099,9 +1100,11 @@ multiclass AForm_3r<bits<6> opcode, bits<5> xo, dag OOL, dag IOL,
 
 let hasCtrlDep = 1 in {
 let Defs = [R1], Uses = [R1] in {
-def ADJCALLSTACKDOWN : Pseudo<(outs), (ins u16imm:$amt), "#ADJCALLSTACKDOWN $amt",
-                              [(callseq_start timm:$amt)]>;
-def ADJCALLSTACKUP   : Pseudo<(outs), (ins u16imm:$amt1, u16imm:$amt2), "#ADJCALLSTACKUP $amt1 $amt2",
+def ADJCALLSTACKDOWN : Pseudo<(outs), (ins u16imm:$amt1, u16imm:$amt2),
+                              "#ADJCALLSTACKDOWN $amt1 $amt2",
+                              [(callseq_start timm:$amt1, timm:$amt2)]>;
+def ADJCALLSTACKUP   : Pseudo<(outs), (ins u16imm:$amt1, u16imm:$amt2),
+                              "#ADJCALLSTACKUP $amt1 $amt2",
                               [(callseq_end timm:$amt1, timm:$amt2)]>;
 }
 
diff --git a/llvm/lib/Target/PowerPC/PPCTLSDynamicCall.cpp b/llvm/lib/Target/PowerPC/PPCTLSDynamicCall.cpp
index 0c1260a2965..c7aa4cb78b7 100644
--- a/llvm/lib/Target/PowerPC/PPCTLSDynamicCall.cpp
+++ b/llvm/lib/Target/PowerPC/PPCTLSDynamicCall.cpp
@@ -99,7 +99,8 @@ protected:
         // Don't really need to save data to the stack - the clobbered
         // registers are already saved when the SDNode (e.g. PPCaddiTlsgdLAddr)
         // gets translated to the pseudo instruction (e.g. ADDItlsgdLADDR).
-        BuildMI(MBB, I, DL, TII->get(PPC::ADJCALLSTACKDOWN)).addImm(0);
+        BuildMI(MBB, I, DL, TII->get(PPC::ADJCALLSTACKDOWN)).addImm(0)
+                                                            .addImm(0);
 
         // Expand into two ops built prior to the existing instruction.
         MachineInstr *Addi = BuildMI(MBB, I, DL, TII->get(Opc1), GPR3)