[clang] [clang] constexpr atomic builtins (__c11_atomic_OP and __atomic_OP) (PR #98756)

[Aaron Ballman via cfe-commits]

Hana =?utf-8?q?Dusíková?= <hanicka at hanicka.net>,
Hana =?utf-8?q?Dusíková?= <hanicka at hanicka.net>
Message-ID:
In-Reply-To: <llvm.org/llvm/llvm-project/pull/98756 at github.com>


================
@@ -1921,6 +1921,22 @@ static bool EvaluateFixedPointOrInteger(const Expr *E, APFixedPoint &Result,
 static bool EvaluateFixedPoint(const Expr *E, APFixedPoint &Result,
                                EvalInfo &Info);
 
+/// Support for atomic builtins
+static bool EvaluateAtomicOrder(const AtomicExpr *E, EvalInfo &Info);
+static bool EvaluateAtomicWeak(const AtomicExpr *E, EvalInfo &Info);
+static bool EvaluateAtomicLoad(const AtomicExpr *E, APValue &Result,
+                               EvalInfo &Info);
+static bool EvaluatePointerAndStoreValueInto(Expr *ResultPtr,
+                                             APValue &ValueToStore,
+                                             EvalInfo &Info);
+static bool EvaluateAtomicLoadInto(const AtomicExpr *E, EvalInfo &Info);
+static bool EvaluateAtomicStore(const AtomicExpr *E, EvalInfo &Info);
+static bool EvaluateAtomicExchangeInto(const AtomicExpr *E, EvalInfo &Info);
+static bool EvaluateAtomicFetchOp(const AtomicExpr *E, APValue &Result,
+                                  EvalInfo &Info, bool StoreToResultAfter);
+static bool EvaluateAtomicCompareExchange(const AtomicExpr *E, APValue &Result,
+                                          EvalInfo &Info);
+
----------------
AaronBallman wrote:

I don't have a strong opinion either way currently, but because this is constant expression evaluation and C++ already makes processing headers painfully slow, we prefer reduced readability if it measurably improves performance in this case. So measuring overhead may not be a bad idea.

https://github.com/llvm/llvm-project/pull/98756