On Tue, Mar 19, 2024 at 01:13:33PM -0400, Steven Rostedt wrote:
> On Tue, 19 Mar 2024 13:06:53 -0400
> Steven Rostedt <rostedt@goodmis.org> wrote:
> 
> > This is in the fast path (where the trace event happens), but I can make it
> > always do strcmp(), even though it will slow down what is being recorded,
> > as I plan on removing the parameter in the next merge window anyway.
> 
> I'll change it to this:
> 
> #define __assign_str(dst, src)						\
> 	do {								\
> 		char *__str__ = __get_str(dst);				\
> 		int __len__ = __get_dynamic_array_len(dst) - 1;		\
> 		WARN_ON_ONCE(!src != !__data_offsets.dst##_ptr_);	\
> 		WARN_ON_ONCE(src && strcmp((src), __data_offsets.dst##_ptr_)); \
> 		memcpy(__str__, __data_offsets.dst##_ptr_ ? :		\
> 		       EVENT_NULL_STR, __len__);			\
> 		__str__[__len__] = '\0';				\
> 	} while (0)
> 
> The first WARN_ON() will report the bug if src is not NULL and the pointer
> is even though the strcmp() will likely crash in that case. But that's a
> bug if that happens anyway.

For what it's worth, I applied that change and built ARCH=x86_64
defconfig with LLVM 18.1.1 from [1] but it does not appear to help the
instances of -Wstring-compare; in fact, it adds some additional warnings
that I have not seen before. I have attached the full build log.

[1]: https://mirrors.edge.kernel.org/pub/tools/llvm/

Cheers,
Nathan