[llvm] [AA] Improve precision for monotonic atomic load/store operations (PR #158169)

[Jin Huang via llvm-commits]

================
@@ -431,6 +431,14 @@ ModRefInfo AAResults::getModRefInfo(const LoadInst *L,
     if (AR == AliasResult::NoAlias)
       return ModRefInfo::NoModRef;
   }
+
+  // At this point, the load's ordering is at most `Monotonic` (i.e., Monotonic,
+  // Unordered, or non-atomic), and it aliases with `Loc`. The condition
----------------
jinhuang1102 wrote:

Replace the comments by 
```c++
assert(!isStrongerThanMonotonic(L->getOrdering()) &&
            "Stronger atomic orderings should have been handled above!" );
```

https://github.com/llvm/llvm-project/pull/158169