[llvm] [coro][CoroSplit] Use `llvm.lifetime.end` to compute putting objects on the frame vs the stack (PR #90265)

[Alan Zhao via llvm-commits]

================
@@ -1614,6 +1621,24 @@ struct AllocaUseVisitor : PtrUseVisitor<AllocaUseVisitor> {
       // suspend point between lifetime markers. This should also cover the
       // case of a single lifetime.start intrinsic in a loop with suspend point.
       if (PI.isEscaped()) {
+        // If there is no explicit lifetime.end, then assume the address can
+        // cross suspension points.
+        if (LifetimeEndBBs.empty())
+          return true;
+
+        // If there is a path from a lifetime.start to a suspend without a
+        // corresponding lifetime.end, then the alloca's lifetime persists
+        // beyond that suspension point and the alloca must go on the frame.
+        for (AnyCoroSuspendInst *SuspendInst : CoroShape.CoroSuspends) {
+          SmallVector<BasicBlock *> LifetimeStartsBB;
+          for (IntrinsicInst *II : LifetimeStarts)
+            LifetimeStartsBB.push_back(II->getParent());
+          if (isPotentiallyReachableFromMany(LifetimeStartsBB,
----------------
alanzhao1 wrote:

Any solution to the problem would require answering the question "Is there a path from `BasicBlock` A to `BasicBlock` B **without** going through `BasicBlock`s { C, D, E, ...}. The current logic caches whether a path _may_ go through a suspend point - it doesn't answer the reverse question and I'm not sure if that's even cacheable since we have to provide a new set of `BasicBlock`s for each `alloca` (since each `alloca` has its own set of `llvm.coro.end`s)

https://github.com/llvm/llvm-project/pull/90265