[CodeGen] Fix partial phi input removal in TailDuplicator.

[ivafanas]

Tail duplicator removes the first PHI income from the predecessor basic block, while it should remove all operands for this block.

PHI instructions happen to have duplicated values for the same predecessor block:

• UnreachableMachineBlockElim assumes that PHI instruction might have duplicates:

  llvm-project/llvm/lib/CodeGen/UnreachableBlockElim.cpp

  Line 160 in 7289f2c

  for (MachineInstr &Phi : succ->phis()) {

• AArch64 directly states that PHI instruction might have duplicates:

  llvm-project/llvm/lib/Target/AArch64/AArch64ConditionalCompares.cpp

  Line 244 in 7289f2c

  // I is a PHI. It can have multiple entries for CmpBB.

• And Hexagon:

  llvm-project/llvm/lib/Target/Hexagon/HexagonConstPropagation.cpp

  Line 844 in 7289f2c

  // Remove all corresponding PHI operands in the To block.

We have caught the bug on custom out-of-tree backend. TailDuplicator should remove all operands corresponding to the removing block.

Please note, that bug likely does not affect in-tree backends, because:

• It happens only in scenario of partial tail duplication (i.e. tail block is duplicated in some predecessors, but not in all of them)
• It happens in Pre-RA tail duplication only (Post-RA does not contain PHIs, obviously)
• The only backend (I know) uses Pre-RA tail duplication is X86. It uses tail duplication via early-tailduplication pass which declines partial tail duplication via canCompletelyDuplicateBB check, because it uses TailDuplicator::tailDuplicateBlocks public API.

So, bug happens only in the case of pre-ra partial tail duplication if backend uses TailDuplicator::tailDuplicate public API directly.

That's why I can not add reproducer test for in-tree backends.

[ivafanas]

@fhahn, @arsenm

Could you please review the PR?

If it is ok, could you please merge it? I do not have commit access.

[arsenm]

I thought there was a helper for this somewhere? The IR has one, and isn't there one used in the MachineBasicBlock CFG helpers?

[ivafanas]

I've run grep by getMBB() and seems like there is no helper. PHI input removal code is seems to be re-implemented from scratch each time (or copy-pasted)

[ivafanas]

Thank you!