Re: [PATCH] ftrace/fgraph: fix increased missing-prototypes warnings

[Catalin Marinas]

On Tue, May 24, 2022 at 07:17:18PM -0400, Steven Rostedt wrote:
> On Fri,  6 May 2022 11:27:37 +0800
> Chengming Zhou <zhouchengming@xxxxxxxxxxxxx> wrote:
> 
> > After commit e999995c84c3 ("ftrace: cleanup ftrace_graph_caller enable
> > and disable") merged into the linux-next tree, the kernel test robot
> > (lkp@xxxxxxxxx) has send out report that there are increased missing-prototypes
> > warnings caused by that commit.
> > 
> > COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-11.3.0 make.cross W=1 \
> > 	O=build_dir ARCH=sh SHELL=/bin/bash kernel/trace/
> > warning: no previous prototype for 'ftrace_enable_ftrace_graph_caller' [-Wmissing-prototypes]
> > warning: no previous prototype for 'ftrace_disable_ftrace_graph_caller' [-Wmissing-prototypes]
> > warning: no previous prototype for 'ftrace_return_to_handler' [-Wmissing-prototypes]
> > warning: no previous prototype for 'ftrace_graph_sleep_time_control' [-Wmissing-prototypes]
> > 
> > BTW there are so many missing-prototypes warnings if build kernel with "W=1".
> > 
> > The increased warnings for 'ftrace_[enable,disable]_ftrace_graph_caller'
> > is caused by CONFIG_FUNCTION_GRAPH_TRACER && !CONFIG_DYNAMIC_FTRACE,
> > so the declarations in <linux/ftrace.h> can't be seen in fgraph.c.
> > 
> > And this warning can't reproduce on x86_64 since x86_64 select
> > HAVE_FUNCTION_GRAPH_TRACER only when DYNAMIC_FTRACE, so fgraph.c will
> > always see the declarations in <linux/ftrace.h>.
> > 
> > This patch fix the increased warnings by put the definitions in
> > CONFIG_DYNAMIC_FTRACE although there are no real problems exist.
> > 
> > Signed-off-by: Chengming Zhou <zhouchengming@xxxxxxxxxxxxx>
> 
> The change this fixes was merged into the ARM tree and I do not have that
> code, so I can not apply this.

Ah, I missed this patch. The commit mentioned is in mainline now but I'm
happy to add it to the arm64 tree and send before -rc1.

-- 
Catalin