Re: [RFC PATCH v2 4/6] tracing/fprobe: Enable fprobe events with CONFIG_DYNAMIC_FTRACE_WITH_ARGS

[Google]

On Wed, 9 Aug 2023 12:31:00 +0200
Florent Revest <revest@xxxxxxxxxxxx> wrote:

> On Mon, Aug 7, 2023 at 8:49 AM Masami Hiramatsu (Google)
> <mhiramat@xxxxxxxxxx> wrote:
> >
> > From: Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx>
> >
> > Allow fprobe events to be enabled with CONFIG_DYNAMIC_FTRACE_WITH_ARGS.
> > With this change, fprobe events mostly use ftrace_regs instead of pt_regs.
> > Note that if the arch doesn't enable HAVE_PT_REGS_COMPAT_FTRACE_REGS,
> > fprobe events will not be able to use from perf.
> 
> nit: "to be used from perf" ?

OK.

> 
> > --- a/kernel/trace/trace_fprobe.c
> > +++ b/kernel/trace/trace_fprobe.c
> > @@ -132,25 +132,30 @@ static int
> >  process_fetch_insn(struct fetch_insn *code, void *rec, void *dest,
> >                    void *base)
> >  {
> > -       struct pt_regs *regs = rec;
> > -       unsigned long val;
> > +       struct ftrace_regs *fregs = rec;
> > +       unsigned long val, *stackp;
> >         int ret;
> >
> >  retry:
> >         /* 1st stage: get value from context */
> >         switch (code->op) {
> >         case FETCH_OP_STACK:
> > -               val = regs_get_kernel_stack_nth(regs, code->param);
> > +               stackp = (unsigned long *)ftrace_regs_get_stack_pointer(fregs);
> > +               if (((unsigned long)(stackp + code->param) & ~(THREAD_SIZE - 1)) ==
> > +                   ((unsigned long)stackp & ~(THREAD_SIZE - 1)))
> 
> Maybe it'd be worth extracting a local
> "ftrace_regs_get_kernel_stack_nth_addr" helper function and/or
> "ftrace_regs_within_kernel_stack" ?

Yeah, maybe we can make it a generic inline function in linux/ftrace.h.

Thank you!


-- 
Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx>