Re: [PATCH v5 1/1] tracing: Remove "__attribute__()" from the type field of event format

From: Google
Date: Tue Jul 29 2025 - 20:20:16 EST


On Tue, 29 Jul 2025 12:33:49 -0400
Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:

> On Tue, 29 Jul 2025 14:10:35 +0900
> "Masami Hiramatsu (Google)" <mhiramat@xxxxxxxxxx> wrote:
>
> > static void update_event_fields(struct trace_event_call *call,
> > struct trace_eval_map *map)
> > {
> > struct ftrace_event_field *field;
> > + const char *eval_string = NULL;
> > struct list_head *head;
> > + int len = 0;
> > char *ptr;
> > char *str;
> > - int len = strlen(map->eval_string);
> >
> > /* Dynamic events should never have field maps */
> > - if (WARN_ON_ONCE(call->flags & TRACE_EVENT_FL_DYNAMIC))
> > + if (call->flags & TRACE_EVENT_FL_DYNAMIC)
> > return;
>
> Is there a reason you removed the WARN_ON_ONCE()?

Yeah, I removed that, but maybe I misunderstood.

>
> Even for the attribute, this shouldn't be called by dynamic events. Or
> do you know of a case?

Yeah, with this change, trace_event_update_all() is changed to call
update_event_fields() on all events even if there is no eval map.
And this trace_event_update_all() is called from the below call chain;

trace_module_notify() -> trace_module_add_evals() -> trace_event_update_with_eval_map().

So, IIUC, if there is any dynamic events on the list, and a module is
loaded, trace_event_update_all() is called.

Hmm, if so, this is really inefficient. We need to introduce build-time
sanitize routine.

Thank you,

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