[llvm] Add NCD search on Array of basic blocks (NFC)

[enoskova-sc]

Shrink-Wrap points split Part 2.
RFC: https://discourse.llvm.org/t/shrink-wrap-save-restore-points-splitting/83581

Part 1: #117862
Part 3: #119357
Part 4: #119358
Part 5: #119359

[llvmbot]

@llvm/pr-subscribers-llvm-support

@llvm/pr-subscribers-backend-risc-v

Author: Elizaveta Noskova (enoskova-sc)

Changes

Shrink-Wrap points split Part 2.

Part 1: #117862 (comment)

---

Full diff: https://github.com/llvm/llvm-project/pull/119355.diff

3 Files Affected:

• (modified) llvm/include/llvm/CodeGen/MachineDominators.h (+5)
• (modified) llvm/lib/CodeGen/MachineDominators.cpp (+16)
• (modified) llvm/lib/Target/RISCV/RISCVFrameLowering.cpp (+4)

diff --git a/llvm/include/llvm/CodeGen/MachineDominators.h b/llvm/include/llvm/CodeGen/MachineDominators.h
index 74cf94398736dd..88800d91ef51a9 100644
--- a/llvm/include/llvm/CodeGen/MachineDominators.h
+++ b/llvm/include/llvm/CodeGen/MachineDominators.h
@@ -185,6 +185,11 @@ class MachineDominatorTree : public DomTreeBase<MachineBasicBlock> {
     return Base::findNearestCommonDominator(A, B);
   }
 
+  /// Returns the nearest common dominator of the given blocks.
+  /// If that tree node is a virtual root, a nullptr will be returned.
+  MachineBasicBlock *
+  findNearestCommonDominator(ArrayRef<MachineBasicBlock *> Blocks) const;
+
   MachineDomTreeNode *operator[](MachineBasicBlock *BB) const {
     applySplitCriticalEdges();
     return Base::getNode(BB);
diff --git a/llvm/lib/CodeGen/MachineDominators.cpp b/llvm/lib/CodeGen/MachineDominators.cpp
index a2cc8fdfa7c9f9..384f90c6da66c0 100644
--- a/llvm/lib/CodeGen/MachineDominators.cpp
+++ b/llvm/lib/CodeGen/MachineDominators.cpp
@@ -189,3 +189,19 @@ void MachineDominatorTree::applySplitCriticalEdges() const {
   NewBBs.clear();
   CriticalEdgesToSplit.clear();
 }
+
+MachineBasicBlock *MachineDominatorTree::findNearestCommonDominator(
+    ArrayRef<MachineBasicBlock *> Blocks) const {
+  assert(!Blocks.empty());
+
+  MachineBasicBlock *NCD = Blocks.front();
+  for (MachineBasicBlock *BB : Blocks.drop_front()) {
+    NCD = Base::findNearestCommonDominator(NCD, BB);
+
+    // Stop when the root is reached.
+    if (Base::isVirtualRoot(Base::getNode(NCD)))
+      return nullptr;
+  }
+
+  return NCD;
+}
diff --git a/llvm/lib/Target/RISCV/RISCVFrameLowering.cpp b/llvm/lib/Target/RISCV/RISCVFrameLowering.cpp
index deb0b627225c64..0de1f1d821a6e2 100644
--- a/llvm/lib/Target/RISCV/RISCVFrameLowering.cpp
+++ b/llvm/lib/Target/RISCV/RISCVFrameLowering.cpp
@@ -1607,6 +1607,8 @@ bool RISCVFrameLowering::assignCalleeSavedSpillSlots(
         int FrameIdx = MFI.CreateFixedSpillStackObject(Size, Offset);
         assert(FrameIdx < 0);
         CS.setFrameIdx(FrameIdx);
+        if (RISCVRegisterInfo::isRVVRegClass(RC))
+          MFI.setStackID(FrameIdx, TargetStackID::ScalableVector);
         continue;
       }
     }
@@ -1623,6 +1625,8 @@ bool RISCVFrameLowering::assignCalleeSavedSpillSlots(
     if ((unsigned)FrameIdx > MaxCSFrameIndex)
       MaxCSFrameIndex = FrameIdx;
     CS.setFrameIdx(FrameIdx);
+    if (RISCVRegisterInfo::isRVVRegClass(RC))
+      MFI.setStackID(FrameIdx, TargetStackID::ScalableVector);
   }
 
   // Allocate a fixed object that covers the full push or libcall size.

[enoskova-sc]

@paperchalice, could you, please, review.

[paperchalice]

assert with error message if possible: https://llvm.org/docs/CodingStandards.html#id45

[enoskova-sc]

addressed

[paperchalice]

Base::getNode here is unfortunate, use wrapped version here to get the latest dominator tree.
Addressed by #115111.

[enoskova-sc]

addressed

[preames]

Can you find a spot in existing code to use this factored out routine? A few ideas:
FindIDom in ShrinkWrap
DeleteUnreachable in GenericDomConstruction.
SILowerI1Copies.cpp
HexagonFrameLowering.cpp

Basically, there's a whole bunch of these. Please common a few so that we're not adding untested code.

This appears to be identical code to findNearestCommonDominator. Can we sink this into the GenericIDom layer instead?

[michaelmaitland]

Reverse ping. Also, I think this could be rebased now that the base patch has landed.

[enoskova-sc]

@preames, @michaelmaitland, I have addressed your comments.

[michaelmaitland]

LGTM

I think you can mark the title as NFC?

[enoskova-sc]

@preames, @paperchalice, any other comments?

[preames]

@preames, @paperchalice, any other comments?

You've got an LGTM and we're a one approval culture. My prior comments about reuse still stands, but that can easily be extended in a follow up commit.