AMDGPU: Make infinite loop clear, NFC

[dexonsmith]

Summary

Change a while loop that was checking for nullptr on an
iterator-to-pointer conversion to an infinite for loop. Now it's clear
that the condition doesn't terminate.

The only change in behaviour is if an invalid iterator (holding nullptr)
was passed into AMDGPUCFGStructurizer::reversePredicateSetter. There
are only two callers, and they both dereference the iterator before
sending it in, so rather than adding an early return to avoid the loop
I've just asserted (using a static_cast, to avoid an implicit conversion
to pointer).

[dexonsmith]

Is this the right fix, or is the infinite loop accidental?

[dexonsmith]

Ping!

[arsenm]

LGTM

[dexonsmith]

Committed in r274902. Thanks!

lib/Target/AMDGPU/AMDILCFGStructurizer.cpp

if(!MultiplePreds)
     return false;
   unsigned BlkSize = MBB->size();
   return ((BlkSize > BlockSizeThreshold) &&
       (BlkSize * (MBB->pred_size() - 1) > CloneInstrThreshold));
 }
 
 void AMDGPUCFGStructurizer::reversePredicateSetter(
     MachineBasicBlock::iterator I) {
-  while (I--) {
+  assert(static_cast<MachineInstr *>(I) && "Expected valid iterator");
+  for (;; --I) {
     if (I->getOpcode() == AMDGPU::PRED_X) {
       switch (static_cast<MachineInstr *>(I)->getOperand(2).getImm()) {
       case OPCODE_IS_ZERO_INT:
         static_cast<MachineInstr *>(I)->getOperand(2)
             .setImm(OPCODE_IS_NOT_ZERO_INT);
         return;
       case OPCODE_IS_NOT_ZERO_INT:
         static_cast<MachineInstr *>(I)->getOperand(2)