[compiler-rt] [scudo] Allow to resize allocation ring buffer (PR #82683)

[Evgenii Stepanov via llvm-commits]

================
@@ -1594,12 +1600,33 @@ class Allocator {
     RB->StackDepotSize = StackDepotSize;
     RB->RawStackDepotMap = DepotMap;
 
-    atomic_store(&RingBufferAddress, reinterpret_cast<uptr>(RB),
-                 memory_order_release);
+    swapOutRingBuffer(RB);
     static_assert(sizeof(AllocationRingBuffer) %
                           alignof(typename AllocationRingBuffer::Entry) ==
                       0,
                   "invalid alignment");
+    return true;
+  }
+
+  void swapOutRingBuffer(AllocationRingBuffer *NewRB) {
+    // To allow resizeRingBuffer to be called in a multi-threaded context by apps,
+    // we do not actually unmap, but only madvise(DONTNEED) the pages. That way,
+    // straggler threads will not crash.
----------------
eugenis wrote:

This DONTNEED usage looks safe to me. A trailing access with either repopulate the page or substitute a zero page, according to the spec. Not completely sure, but the latter might be the only actual possibility in Linux. In any case, the worst that may happen is we waste (leak) a page or two.

https://github.com/llvm/llvm-project/pull/82683