Fix reordering of instructions during VirtRegRewriter unbundling

[critson]

Summary

When unbundling COPY bundles in VirtRegRewriter the start of the
bundle is not correctly referenced in the unbundling loop.

The effect of this is that unbundled instructions are sometimes
inserted out-of-order, particular in cases where multiple
reordering have been applied to avoid clobbering dependencies.
The resulting instruction sequence clobbers dependencies.

[critson]

Pre-commit test.

[foad]

llvm/test/CodeGen/AMDGPU/splitkit-copy-bundle.mir

301

So the bug here is that this overwrites sgpr16 before the copy from sgpr16_sgpr17?

[critson]

llvm/test/CodeGen/AMDGPU/splitkit-copy-bundle.mir

301

Yes

[foad]

LGTM, though I can't help feeling the whole function could be simpler.

llvm/lib/CodeGen/VirtRegMap.cpp

if (MI.isBundledWithPred() && !MI.isBundledWithSucc()) {
 
     MachineInstr *BundleStart = FirstMI;
     for (MachineInstr *BundledMI : llvm::reverse(MIs)) {
       // If instruction is in the middle of the bundle, move it before the
       // bundle starts, otherwise, just unbundle it. When we get to the last
       // instruction, the bundle will have been completely undone.
       if (BundledMI != BundleStart) {
         BundledMI->removeFromBundle();
-        MBB.insert(FirstMI, BundledMI);
+        MBB.insert(BundleStart, BundledMI);
       } else if (BundledMI->isBundledWithSucc()) {
         BundledMI->unbundleFromSucc();
         BundleStart = &*std::next(BundledMI->getIterator());
       }
 
       if (Indexes && BundledMI != FirstMI)
         Indexes->insertMachineInstrInMaps(*BundledMI);
     }

llvm/test/CodeGen/AMDGPU/splitkit-copy-bundle.mir

bb.0:
     ; VR: renamable $sgpr23 = S_MOV_B32 -1
     ; VR: renamable $sgpr19 = S_MOV_B32 -1
     ; VR: renamable $sgpr20 = S_MOV_B32 -1
     ; VR: renamable $sgpr25 = S_MOV_B32 -1
     ; VR: renamable $sgpr26 = S_MOV_B32 -1
     ; VR: SI_SPILL_S512_SAVE killed renamable $sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27, %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32 :: (store 64 into %stack.0, align 4, addrspace 5)
     ; VR: S_NOP 0, implicit-def $sgpr8, implicit-def $sgpr12, implicit-def $sgpr16, implicit-def $sgpr20, implicit-def $sgpr24, implicit-def $sgpr28, implicit-def $sgpr32, implicit-def $sgpr36, implicit-def $sgpr40, implicit-def $sgpr44, implicit-def $sgpr48, implicit-def $sgpr52, implicit-def $sgpr56, implicit-def $sgpr60, implicit-def $sgpr64, implicit-def $sgpr68, implicit-def $sgpr72, implicit-def $sgpr74, implicit-def $sgpr78, implicit-def $sgpr82, implicit-def $sgpr86, implicit-def $sgpr90, implicit-def $sgpr94, implicit-def $sgpr98
     ; VR: renamable $sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27 = SI_SPILL_S512_RESTORE %stack.0, implicit $exec, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr32 :: (load 64 from %stack.0, align 4, addrspace 5)
-    ; VR: renamable $sgpr16 = COPY killed renamable $sgpr20
+    ; VR: renamable $sgpr12_sgpr13 = COPY killed renamable $sgpr16_sgpr17
     ; VR: renamable $sgpr15 = COPY killed renamable $sgpr19
     ; VR: renamable $sgpr18_sgpr19 = COPY killed renamable $sgpr22_sgpr23
+    ; VR: renamable $sgpr16 = COPY killed renamable $sgpr20
     ; VR: renamable $sgpr21 = COPY killed renamable $sgpr25
-    ; VR: renamable $sgpr12_sgpr13 = COPY renamable $sgpr16_sgpr17
     ; VR: renamable $sgpr22 = COPY killed renamable $sgpr26
     ; VR: renamable $sgpr4_sgpr5_sgpr6_sgpr7 = IMPLICIT_DEF
     ; VR: renamable $sgpr8 = S_BUFFER_LOAD_DWORD_SGPR renamable $sgpr4_sgpr5_sgpr6_sgpr7, killed renamable $sgpr12, 0, 0 :: (dereferenceable invariant load 4)
     ; VR: renamable $sgpr9 = S_BUFFER_LOAD_DWORD_SGPR renamable $sgpr4_sgpr5_sgpr6_sgpr7, killed renamable $sgpr13, 0, 0 :: (dereferenceable invariant load 4)
     ; VR: renamable $sgpr14 = S_BUFFER_LOAD_DWORD_SGPR renamable $sgpr4_sgpr5_sgpr6_sgpr7, killed renamable $sgpr15, 0, 0 :: (dereferenceable invariant load 4)
     ; VR: renamable $sgpr15 = S_BUFFER_LOAD_DWORD_SGPR renamable $sgpr4_sgpr5_sgpr6_sgpr7, killed renamable $sgpr16, 0, 0 :: (dereferenceable invariant load 4)
     ; VR: renamable $sgpr10_sgpr11 = IMPLICIT_DEF
     ; VR: renamable $sgpr12 = S_BUFFER_LOAD_DWORD_SGPR renamable $sgpr4_sgpr5_sgpr6_sgpr7, killed renamable $sgpr18, 0, 0 :: (dereferenceable invariant load 4)