[clang] [analyzer] Suppress out of bounds reports after weak loop assumptions (PR #109804)
DonĂ¡t Nagy via cfe-commits
cfe-commits at lists.llvm.org
Wed Sep 25 07:26:22 PDT 2024
================
@@ -2808,27 +2825,63 @@ void ExprEngine::processBranch(const Stmt *Condition,
std::tie(StTrue, StFalse) = *KnownCondValueAssumption;
else {
assert(!isa<ObjCForCollectionStmt>(Condition));
+ // TODO: instead of this shortcut perhaps it would be better to "rejoin"
+ // the common execution path with
+ // StTrue = StFalse = PrevState;
builder.generateNode(PrevState, true, PredN);
builder.generateNode(PrevState, false, PredN);
continue;
}
if (StTrue && StFalse)
assert(!isa<ObjCForCollectionStmt>(Condition));
+ const Expr *EagerlyAssumeExpr =
+ PrevState->get<LastEagerlyAssumeAssumptionAt>();
+ const Expr *ConditionExpr = dyn_cast<Expr>(Condition);
+ if (ConditionExpr)
+ ConditionExpr = ConditionExpr->IgnoreParenCasts();
+ bool DidEagerlyAssume = EagerlyAssumeExpr == ConditionExpr;
+ bool BothFeasible = (DidEagerlyAssume || (StTrue && StFalse)) &&
+ builder.isFeasible(true) && builder.isFeasible(false);
+
// Process the true branch.
if (builder.isFeasible(true)) {
- if (StTrue)
+ if (StTrue) {
+ if (BothFeasible && IterationsFinishedInLoop &&
+ *IterationsFinishedInLoop >= 2) {
+ // When programmers write a loop, they imply that at least two
+ // iterations are possible (otherwise they would just write an `if`),
+ // but the third iteration is not implied: there are situations where
+ // the programmer knows that there won't be a third iteration (e.g.
+ // they iterate over a structure that has <= 2 elements) but this is
+ // not marked in the source code.
----------------
NagyDonat wrote:
I'm speaking about a situation when they have an array with a static size 2 (or 3) but the loop condition look like `i < some_struct.num_channels` where the upper bound is an opaque value (which is presumably not more than the static array size, but this is not checked).
I know that this sounds like a niche situation, but analyzing FFMPEG floods the user with dozens of issues like this.
https://github.com/llvm/llvm-project/pull/109804
More information about the cfe-commits
mailing list