[PATCH 7/8] trace: remove limitation on trace point name

[David Marchand]

On Tue, Oct 11, 2022 at 4:49 PM Jerin Jacob <jerinjacobk at gmail.com> wrote:
>
> On Wed, Sep 21, 2022 at 5:35 PM David Marchand
> <david.marchand at redhat.com> wrote:
> >
> > The name of a trace point is provided as a constant string via the
> > RTE_TRACE_POINT_REGISTER macro.
> > We can rely on the constant string in the binary and simply point at it.
>
> I am not sure about this. If we compile with -Os (optimized for space)
> compiler may decide to use stack instead of rodata.
> If so, we will have segfaults with specific compiler or compiler build options.
> Thoughts?

Good point.

We need an explicit storage for the string.
What do you think of:

@@ -20,9 +20,10 @@ RTE_DECLARE_PER_LCORE(volatile int, trace_point_sz);

 #define RTE_TRACE_POINT_REGISTER(trace, name) \
 rte_trace_point_t __attribute__((section("__rte_trace_point"))) __##trace; \
+static const char __##trace##_name[] = RTE_STR(name); \
 RTE_INIT(trace##_init) \
 { \
-       __rte_trace_point_register(&__##trace, RTE_STR(name), \
+       __rte_trace_point_register(&__##trace, __##trace##_name, \
                (void (*)(void)) trace); \
 }



-- 
David Marchand