[IfConversion] Correctly handle cases where analyzeBranch fails.

If analyzeBranch fails, on some targets, the out parameters point to
some blocks in the function. But we can't use that information, so make
sure to clear it out.  (In some places in IfConversion, we assume that
any block with a TrueBB is analyzable.)

The change to the testcase makes it trigger a bug on builds without this
fix: IfConvertDiamond tries to perform a followup "merge" operation,
which isn't legal, and we somehow end up with a branch to a deleted MBB.
I'm not sure how this doesn't crash the compiler.

Differential Revision: https://reviews.llvm.org/D67306

llvm-svn: 371434

llvm/lib/CodeGen/IfConversion.cpp

@@ -912,6 +912,12 @@ void IfConverter::AnalyzeBranches(BBInfo &BBI) {
   BBI.BrCond.clear();
   BBI.IsBrAnalyzable =
       !TII->analyzeBranch(*BBI.BB, BBI.TrueBB, BBI.FalseBB, BBI.BrCond);
+  if (!BBI.IsBrAnalyzable) {
+    BBI.TrueBB = nullptr;
+    BBI.FalseBB = nullptr;
+    BBI.BrCond.clear();
+  }
+
   SmallVector<MachineOperand, 4> RevCond(BBI.BrCond.begin(), BBI.BrCond.end());
   BBI.IsBrReversible = (RevCond.size() == 0) ||
       !TII->reverseBranchCondition(RevCond);

llvm/test/CodeGen/ARM/ifcvt-diamond-unanalyzable-common.mir

@@ -1,4 +1,4 @@
-# RUN: llc %s -o - -run-pass=if-converter | FileCheck %s
+# RUN: llc %s -o - -run-pass=if-converter -verify-machineinstrs | FileCheck %s
 # Make sure we correctly if-convert blocks containing an INLINEASM_BR.
 # CHECK: t2CMPri killed renamable $r2, 34
 # CHECK-NEXT: $r0 = t2MOVi 2, 1, $cpsr, $noreg
@@ -48,9 +48,8 @@ body:             |
     t2B %bb.3, 14, $noreg
   
   bb.3:
-    successors: %bb.4(0x80000000)
-  
     INLINEASM &"", 1
+    $sp = t2LDMIA_RET $sp, 14, $noreg, def $r4, def $pc
   
   bb.4.l_yes (address-taken):
     $sp = t2LDMIA_RET $sp, 14, $noreg, def $r4, def $pc