Re: [RFC] ftrace: Add support to keep some functions out of ftrace

[Jiri Olsa]

On Thu, Aug 18, 2022 at 02:00:21PM -0700, Alexei Starovoitov wrote:
> On Thu, Aug 18, 2022 at 1:50 PM Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
> >
> > On Thu, 18 Aug 2022 22:27:07 +0200
> > Jiri Olsa <olsajiri@xxxxxxxxx> wrote:
> >
> > > ok, so the problem with __attribute__((patchable_function_entry(5))) is that
> > > it puts function address into __patchable_function_entries section, which is
> > > one of ftrace locations source:
> > >
> > >   #define MCOUNT_REC()    . = ALIGN(8);     \
> > >     __start_mcount_loc = .;                 \
> > >     KEEP(*(__mcount_loc))                   \
> > >     KEEP(*(__patchable_function_entries))   \
> > >     __stop_mcount_loc = .;                  \
> > >    ...
> > >
> > >
> > > it looks like __patchable_function_entries is used for other than x86 archs,
> > > so we perhaps we could have x86 specific MCOUNT_REC macro just with
> > > __mcount_loc section?
> >
> > So something like this:
> >
> > #ifdef CONFIG_X86
> > # define NON_MCOUNT_PATCHABLE KEEP(*(__patchable_function_entries))
> > # define MCOUNT_PATCHABLE
> > #else
> > # define NON_MCOUNT_PATCHABLE
> > # define MCOUNT_PATCHABLE  KEEP(*(__patchable_function_entries))
> > #endif
> >
> >   #define MCOUNT_REC()    . = ALIGN(8);     \
> >     __start_mcount_loc = .;                 \
> >     KEEP(*(__mcount_loc))                   \
> >     MCOUNT_PATCHABLE                        \
> >     __stop_mcount_loc = .;                  \
> >     NON_MCOUNT_PATCHABLE                    \
> >    ...
> >
> > ??
> 
> That's what more or less Peter's patch is doing:
> Here it is again for reference:
> https://git.kernel.org/pub/scm/linux/kernel/git/peterz/queue.git/commit/?id=8d075bdf11193f1d276bf19fa56b4b8dfe24df9e

ah nice, and discards the __patchable_function_entries section, great

jirka