Re: [PATCH] tracing: fprobe-event: Fix to check tracepoint event and return

[Masami Hiramatsu (Google) <mhiramat@kernel.org>]

On Mon, 6 Nov 2023 17:28:11 -0500
Steven Rostedt <rostedt@goodmis.org> wrote:

> On Sat,  4 Nov 2023 01:05:34 +0900
> "Masami Hiramatsu (Google)" <mhiramat@kernel.org> wrote:
> 
> > From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> > 
> > Fix to check the tracepoint event is not valid with $retval.
> > The commit 08c9306fc2e3 ("tracing/fprobe-event: Assume fprobe is
> > a return event by $retval") introduced automatic return probe
> > conversion with $retval. But since tracepoint event does not
> > support return probe, $retval is not acceptable.
> 
> Can you add the command that causes this to fail.

Actually this doesn't command fail but it expected to fail.

$ echo "t kfree $retval" >> dynamic_events

This should be rejected because 't' (tracepoint) should not be
a return probe. Without this fix, this is accepted.
(But it traces a random return value because the tracepoint function
will return 'void'.)

Thank you,

> 
> -- Steve
> 
> > 
> > Fixes: 08c9306fc2e3 ("tracing/fprobe-event: Assume fprobe is a return event by $retval")
> > Cc: stable@vger.kernel.org
> > Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> > ---
> >  kernel/trace/trace_fprobe.c |    9 +++++----
> >  1 file changed, 5 insertions(+), 4 deletions(-)
> > 
> > diff --git a/kernel/trace/trace_fprobe.c b/kernel/trace/trace_fprobe.c
> > index 8bfe23af9c73..7d2ddbcfa377 100644
> > --- a/kernel/trace/trace_fprobe.c
> > +++ b/kernel/trace/trace_fprobe.c
> > @@ -927,11 +927,12 @@ static int parse_symbol_and_return(int argc, const char *argv[],
> >  	for (i = 2; i < argc; i++) {
> >  		tmp = strstr(argv[i], "$retval");
> >  		if (tmp && !isalnum(tmp[7]) && tmp[7] != '_') {
> > +			if (is_tracepoint) {
> > +				trace_probe_log_set_index(i);
> > +				trace_probe_log_err(tmp - argv[i], RETVAL_ON_PROBE);
> > +				return -EINVAL;
> > +			}
> >  			*is_return = true;
> > -			/*
> > -			 * NOTE: Don't check is_tracepoint here, because it will
> > -			 * be checked when the argument is parsed.
> > -			 */
> >  			break;
> >  		}
> >  	}
> 


-- 
Masami Hiramatsu (Google) <mhiramat@kernel.org>