[memprof] Use ArrayRef instead of std::vector<LinearFrameId> (NFC)

[kazutakahirata]

LLVM Programmer's Manual prefers ArrayRef over actual sequential
container types for read accesses.

[llvmbot]

@llvm/pr-subscribers-pgo

Author: Kazu Hirata (kazutakahirata)

Changes

LLVM Programmer's Manual prefers ArrayRef over actual sequential
container types for read accesses.

---

Full diff: https://github.com/llvm/llvm-project/pull/116279.diff

1 Files Affected:

• (modified) llvm/include/llvm/ProfileData/MemProf.h (+1-1)

diff --git a/llvm/include/llvm/ProfileData/MemProf.h b/llvm/include/llvm/ProfileData/MemProf.h
index ff05bb7da2f799..bfd91407769bb3 100644
--- a/llvm/include/llvm/ProfileData/MemProf.h
+++ b/llvm/include/llvm/ProfileData/MemProf.h
@@ -1134,7 +1134,7 @@ class CallStackRadixTreeBuilder {
              const llvm::DenseMap<FrameId, LinearFrameId> &MemProfFrameIndexes,
              llvm::DenseMap<FrameId, FrameStat> &FrameHistogram);
 
-  const std::vector<LinearFrameId> &getRadixArray() const { return RadixArray; }
+  ArrayRef<LinearFrameId> getRadixArray() const { return RadixArray; }
 
   llvm::DenseMap<CallStackId, LinearCallStackId> takeCallStackPos() {
     return std::move(CallStackPos);