ARM: correct Dwarf output for non-contiguous VFP saves.

When the list of VFP registers to be saved was non-contiguous (so multiple
vpush/vpop instructions were needed) these were being ordered oddly, as in:
    vpush {d8, d9}
    vpush {d11}

This led to the layout in memory being [d11, d8, d9] which is ugly and doesn't
match the CFI_INSTRUCTIONs we're generating either (so Dwarf info would be
broken).

This switches the order of vpush/vpop (in both prologue and epilogue,
obviously) so that the Dwarf locations are correct again.

rdar://problem/16264856

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@203655 91177308-0d34-0410-b5e6-96231b3b80d8

lib/Target/ARM/ARMFrameLowering.cpp

@@ -803,6 +803,11 @@ void ARMFrameLowering::emitPushInst(MachineBasicBlock &MBB,
       AddDefaultPred(MIB);
     }
     Regs.clear();
+
+    // Put any subsequent vpush instructions before this one: they will refer to
+    // higher register numbers so need to be pushed first in order to preserve
+    // monotonicity.
+    --MI;
   }
 }
 
@@ -886,6 +891,10 @@ void ARMFrameLowering::emitPopInst(MachineBasicBlock &MBB,
       AddDefaultPred(MIB);
     }
     Regs.clear();
+
+    // Put any subsequent vpop instructions after this one: they will refer to
+    // higher register numbers so need to be popped afterwards.
+    ++MI;
   }
 }