, , On Thu, Mar 10, 2022 at 6:04 PM Steven Rostedt <rostedt@xxxxxxxxxxx> wrote: > > On Fri, 11 Mar 2022 09:44:11 +0800 > kernel test robot <lkp@xxxxxxxxx> wrote: > > > Hi Ard, > > > > FYI, the error/warning still remains. > > > > tree: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master > > head: 71941773e143369a73c9c4a3b62fbb60736a1182 > > commit: 41918ec82eb6f80c8b401422f27ca76c85aa0cb7 [4103/11953] ARM: ftrace: enable the graph tracer with the EABI unwinder > > config: arm-aspeed_g4_defconfig (https://download.01.org/0day-ci/archive/20220311/202203110903.3xDyTUVl-lkp@xxxxxxxxx/config) > > compiler: clang version 15.0.0 (https://github.com/llvm/llvm-project 276ca87382b8f16a65bddac700202924228982f6) > > reproduce (this is a W=1 build): > > wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross > > chmod +x ~/bin/make.cross > > # install arm cross compiling tool for clang build > > # apt-get install binutils-arm-linux-gnueabi > > # https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=41918ec82eb6f80c8b401422f27ca76c85aa0cb7 > > git remote add linux-next https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git > > git fetch --no-tags linux-next master > > git checkout 41918ec82eb6f80c8b401422f27ca76c85aa0cb7 > > # save the config file to linux build tree > > mkdir build_dir > > COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=arm SHELL=/bin/bash arch/arm/kernel/ > > > > If you fix the issue, kindly add following tag as appropriate > > Reported-by: kernel test robot <lkp@xxxxxxxxx> > > > > All warnings (new ones prefixed by >>): > > > > >> arch/arm/kernel/ftrace.c:229:6: warning: no previous prototype for function 'prepare_ftrace_return' [-Wmissing-prototypes] > > void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr, > > ^ > > arch/arm/kernel/ftrace.c:229:1: note: declare 'static' if the function is not intended to be used outside of this translation unit > > void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr, > > ^ > > static > > 1 warning generated. > > > > Hmm, this is only called from assembly code, which doesn't care about > prototypes. Is there a way to quiet this warning to say "this is called > only by assembly"? > > Or do we just add a useless prototype to shut it up? So this has come up a few times before, particularly around the usage in the kernel of `asmlinkage`. This 0day report also only comes up for W=1 builds, which enable -Wmissing-prototypes (see scripts/Makefile.extrawarn). The problem is that `asmlinkage` is not a language keyword; it's a preprocessor define (see include/linux/linkage.h, arch/ia64/include/asm/linkage.h, arch/x86/include/asm/linkage.h), so the compiler has no idea that this function is only called from assembly only. The warning is meant to catch implicit defines (-Wimplicit-function-declaration) where the signature the caller is expecting doesn't match the callee's actual definition (because the callee wasn't declared in a shared header). I think IF we had a function attribute added to the compiler that disabled -Wmissing-prototype warnings just for that definition, then we could do something like: #define asmlinkage CPP_ASMLINKAGE \ __attribute__((no_wmissing_prototypes,error( \ "Called an asmlinkage function from C; remove " \ "asmlinkage or dont call this fn")) or something. We wouldn't be able to always enable -Wmissing-prototypes until we had such an fn attr though, due to cases like this. I think -Wmissing-prototypes is still worthwhile, and that Arnd previously mentioned interest in this warning. For now, I think it's worth forward declaring the function, in the .c file (rather than a header) with a comment that this is expected to be called from asm only. If that ever changes, then the declaration can be moved to a header. -- Thanks, ~Nick Desaulniers
