Messages in this thread | | | Date | Fri, 15 May 2009 02:20:28 +0200 | From | Frederic Weisbecker <> | Subject | Re: [RFC] tracing: adding flags to events |
| |
On Thu, May 14, 2009 at 08:01:20PM -0400, Steven Rostedt wrote: > > > > On Fri, 15 May 2009, Frederic Weisbecker wrote: > > > On Thu, May 14, 2009 at 07:18:23PM -0400, Steven Rostedt wrote: > > > > > + > > > > > #undef TRACE_EVENT > > > > > #define TRACE_EVENT(call, proto, args, tstruct, assign, print) \ > > > > > enum print_line_t \ > > > > > @@ -127,6 +132,7 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ > > > > > struct trace_seq *s = &iter->seq; \ > > > > > struct ftrace_raw_##call *field; \ > > > > > struct trace_entry *entry; \ > > > > > + struct trace_seq *p; \ > > > > > int ret; \ > > > > > \ > > > > > entry = iter->ent; \ > > > > > @@ -138,7 +144,9 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ > > > > > \ > > > > > field = (typeof(field))entry; \ > > > > > \ > > > > > + p = &get_cpu_var(ftrace_event_seq); \ > > > > > ret = trace_seq_printf(s, #call ": " print); \ > > > > > + put_cpu(); \ > > > > > > > > > > > > > > > > I don't understand the role of this per-cpu trace_seq variable. > > > > It doesn't seem to be used. > > > > > > See it now? ;-) > > > > > > Still not :-) > > > > I don't understand, it doesn't seem to be used. May be I'm too > > much sleepy to understand... > > That's because you cut out an important detail ;-) > > > > +#define __print_flags(flag, x...) ftrace_print_flags_seq(p, flag, x) >
Aah, ok. Then it returns the seq buffer and matches the %s of the real trace_seq_printf.
Wow tricky! And by disabling preemption and using per_cpu, we ensure it won't be concurrently modified because we are in the read callback/always in user context.
Thanks, now I can go to sleep without this pain of doubt :)
| |