[RISCV] Insert VSETVLI at the end of a basic block if we didn't produce BlockInfo.Exit.

[craig.topper]

Summary

This is an alternative to D118667 that instead of fixing the store
to match phase 1, it tries to detect the mismatch with the expected
value at the end of the block. This inserts a vsetvli after the vse
to satisfy the requirement of the other basic block.

We still have serious design issues in the pass, that is going to
require some rethinking.

We definitely need something for the LLVM 14 branch as the previous
fix introduced a worse bug.

[frasercrmck]

LGTM, at least in the short term. We definitely need to revisit this pass though.

Can we get this set of patches into LLVM 14, do you think?

llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp

bool operator==(const VSETVLIInfo &Other) const {
     // If the full VTYPE is valid, check that it is the same.
     if (!SEWLMULRatioOnly && !Other.SEWLMULRatioOnly)
       return hasSameVTYPE(Other);
 
     // If the SEWLMULRatioOnly bits are different, then they aren't equal.
     return false;
   }
 
+  bool operator!=(const VSETVLIInfo &Other) const {
+    return !(*this == Other);
+  }
+
   // Calculate the VSETVLIInfo visible to a block assuming this and Other are
   // both predecessors.
   VSETVLIInfo intersect(const VSETVLIInfo &Other) const {
     // If the new value isn't valid, ignore it.
     if (!Other.isValid())
       return *this;
 
     // If this value isn't valid, this must be the first predecessor, use it.
for (MachineInstr &MI : MBB) {
 
     // If this is something updates VL/VTYPE that we don't know about, set
     // the state to unknown.
     if (MI.isCall() || MI.isInlineAsm() || MI.modifiesRegister(RISCV::VL) ||
         MI.modifiesRegister(RISCV::VTYPE)) {
       CurInfo = VSETVLIInfo::getUnknown();
       PrevVSETVLIMI = nullptr;
     }
+
+    // If we reach the end of the block and our current info doesn't match the
+    // expected info, insert a vsetvli to correct.
+    if (MI.isTerminator()) {
+      const VSETVLIInfo &ExitInfo = BlockInfo[MBB.getNumber()].Exit;
+      if (CurInfo.isValid() && ExitInfo.isValid() && !ExitInfo.isUnknown() &&
+          CurInfo != ExitInfo) {
+        insertVSETVLI(MBB, MI, ExitInfo, CurInfo);
+        CurInfo = ExitInfo;
+      }
+    }
   }
 }
 
 bool RISCVInsertVSETVLI::runOnMachineFunction(MachineFunction &MF) {
   // Skip if the vector extension is not enabled.
   const RISCVSubtarget &ST = MF.getSubtarget<RISCVSubtarget>();
   if (!ST.hasVInstructions())
     return false;

llvm/test/CodeGen/RISCV/rvv/vsetvli-insert-crossbb.mir

body:             |
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.1:
   ; CHECK-NEXT:   successors: %bb.1(0x40000000), %bb.2(0x40000000)
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[PHI:%[0-9]+]]:gpr = PHI [[COPY2]], %bb.0, %10, %bb.1
   ; CHECK-NEXT:   [[PseudoVADD_VX_M1_:%[0-9]+]]:vr = PseudoVADD_VX_M1 [[PseudoVID_V_M1_]], [[PHI]], -1, 6, implicit $vl, implicit $vtype
   ; CHECK-NEXT:   [[MUL:%[0-9]+]]:gpr = MUL [[PHI]], [[SRLI]]
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:gpr = ADD [[COPY]], [[MUL]]
-  ; FIXME: We insert a SEW=32,LMUL=1/2 VSETVLI here but no SEW=64,LMUL=1
-  ; VSETVLI before the VADD above. This misconfigures the VADD in the case that
-  ; the loop takes its backedge.
   ; CHECK-NEXT:   dead $x0 = PseudoVSETVLIX0 killed $x0, 87, implicit-def $vl, implicit-def $vtype, implicit $vl
   ; CHECK-NEXT:   PseudoVSE32_V_MF2 killed [[PseudoVADD_VX_M1_]], killed [[ADD]], -1, 5, implicit $vl, implicit $vtype
   ; CHECK-NEXT:   [[ADDI:%[0-9]+]]:gpr = ADDI [[PHI]], 1
+  ; CHECK-NEXT:   dead $x0 = PseudoVSETVLIX0 killed $x0, 88, implicit-def $vl, implicit-def $vtype, implicit $vl
   ; CHECK-NEXT:   BLTU [[ADDI]], [[COPY1]], %bb.1
   ; CHECK-NEXT:   PseudoBR %bb.2
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.2:
   ; CHECK-NEXT:   PseudoRET
   bb.0:
     liveins: $x10, $x11
     %0:gpr = COPY $x10