================
@@ -1496,7 +1522,24 @@ void
Sema::checkFortifiedBuiltinMemoryFunction(FunctionDecl *FD,
Checker.checkSourceOverread(/*SrcArgIdx=*/0, /*SizeArgIdx=*/2);
break;
}
-
+ case Builtin::BIfread: {
+ DiagID = diag::warn_fortify_source_overflow;
+ SourceSize = Checker.ComputeExplicitObjectSizeArgumentProduct(1, 2);
+ DestinationSize = Checker.ComputeSizeArgument(0);
+ break;
+ }
+ case Builtin::BIfwrite: {
+ DiagID = diag::warn_fortify_source_overread;
+ SourceSize = Checker.ComputeExplicitObjectSizeArgumentProduct(1, 2);
+ DestinationSize = Checker.ComputeSizeArgument(0);
----------------
bozicrHT wrote:
I agree that the names are confusing for the stdio cases. They are also used by
exsiting string and memory checks, so I did not want to rename them globally.
Using separate `BufferSize`/`AccessSize` variables would either duplicate the
common diagnostic logic or require extracting it into a helper. I can refactor
that logic into lambda like `DiagnoseSizeMismatch` and use the clearer names
for `fwrite`, `fread` cases. Does that sounds good to you?
https://github.com/llvm/llvm-project/pull/204337
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
