diff options
| author | Chandler Carruth <chandlerc@gmail.com> | 2018-08-14 23:30:32 +0000 |
|---|---|---|
| committer | Chandler Carruth <chandlerc@gmail.com> | 2018-08-14 23:30:32 +0000 |
| commit | 66654b72c9e6b2aea51d20f9949a010df864d0f9 (patch) | |
| tree | 729a349a925aa9a8d7e1a4a970c811df18d04418 /llvm/lib/Target/MSP430 | |
| parent | 44fdf2dad6ac2434630cb122418cfc8bf418cc83 (diff) | |
| download | bcm5719-llvm-66654b72c9e6b2aea51d20f9949a010df864d0f9.tar.gz bcm5719-llvm-66654b72c9e6b2aea51d20f9949a010df864d0f9.zip | |
[SDAG] Remove the reliance on MI's allocation strategy for
`MachineMemOperand` pointers attached to `MachineSDNodes` and instead
have the `SelectionDAG` fully manage the memory for this array.
Prior to this change, the memory management was deeply confusing here --
The way the MI was built relied on the `SelectionDAG` allocating memory
for these arrays of pointers using the `MachineFunction`'s allocator so
that the raw pointer to the array could be blindly copied into an
eventual `MachineInstr`. This creates a hard coupling between how
`MachineInstr`s allocate their array of `MachineMemOperand` pointers and
how the `MachineSDNode` does.
This change is motivated in large part by a change I am making to how
`MachineFunction` allocates these pointers, but it seems like a layering
improvement as well.
This would run the risk of increasing allocations overall, but I've
implemented an optimization that should avoid that by storing a single
`MachineMemOperand` pointer directly instead of allocating anything.
This is expected to be a net win because the vast majority of uses of
these only need a single pointer.
As a side-effect, this makes the API for updating a `MachineSDNode` and
a `MachineInstr` reasonably different which seems nice to avoid
unexpected coupling of these two layers. We can map between them, but we
shouldn't be *surprised* at where that occurs. =]
Differential Revision: https://reviews.llvm.org/D50680
llvm-svn: 339740
Diffstat (limited to 'llvm/lib/Target/MSP430')
| -rw-r--r-- | llvm/lib/Target/MSP430/MSP430ISelDAGToDAG.cpp | 5 |
1 files changed, 2 insertions, 3 deletions
diff --git a/llvm/lib/Target/MSP430/MSP430ISelDAGToDAG.cpp b/llvm/lib/Target/MSP430/MSP430ISelDAGToDAG.cpp index 005f5f44a63..b196c013902 100644 --- a/llvm/lib/Target/MSP430/MSP430ISelDAGToDAG.cpp +++ b/llvm/lib/Target/MSP430/MSP430ISelDAGToDAG.cpp @@ -362,12 +362,11 @@ bool MSP430DAGToDAGISel::tryIndexedBinOp(SDNode *Op, SDValue N1, SDValue N2, MVT VT = LD->getMemoryVT().getSimpleVT(); unsigned Opc = (VT == MVT::i16 ? Opc16 : Opc8); - MachineSDNode::mmo_iterator MemRefs0 = MF->allocateMemRefsArray(1); - MemRefs0[0] = cast<MemSDNode>(N1)->getMemOperand(); + MachineMemOperand *MemRef = cast<MemSDNode>(N1)->getMemOperand(); SDValue Ops0[] = { N2, LD->getBasePtr(), LD->getChain() }; SDNode *ResNode = CurDAG->SelectNodeTo(Op, Opc, VT, MVT::i16, MVT::Other, Ops0); - cast<MachineSDNode>(ResNode)->setMemRefs(MemRefs0, MemRefs0 + 1); + CurDAG->setNodeMemRefs(cast<MachineSDNode>(ResNode), {MemRef}); // Transfer chain. ReplaceUses(SDValue(N1.getNode(), 2), SDValue(ResNode, 2)); // Transfer writeback. |

