Messages in this thread | | | From | Namhyung Kim <> | Date | Fri, 3 Jul 2020 15:43:38 +0900 | Subject | Re: [PATCH v2 11/15] perf ftrace: add support for trace option funcgraph-irqs |
| |
On Sat, Jun 27, 2020 at 10:38 PM Changbin Du <changbin.du@gmail.com> wrote: > > This adds an option '--graph-noirqs' to filter out functions executed > in irq context.
Ditto.
> > Signed-off-by: Changbin Du <changbin.du@gmail.com> > > --- > v2: option name '--nofuncgraph-irqs' -> '--graph-noirqs'. > --- > tools/perf/Documentation/perf-ftrace.txt | 3 +++ > tools/perf/builtin-ftrace.c | 20 ++++++++++++++++++++ > 2 files changed, 23 insertions(+) > > diff --git a/tools/perf/Documentation/perf-ftrace.txt b/tools/perf/Documentation/perf-ftrace.txt > index a3000436f80b..b616e05d5156 100644 > --- a/tools/perf/Documentation/perf-ftrace.txt > +++ b/tools/perf/Documentation/perf-ftrace.txt > @@ -107,6 +107,9 @@ OPTIONS > --graph-nosleep-time:: > Measure on-CPU time only for function_graph tracer. > > +--graph-noirqs:: > + Ignore functions that happen inside interrupt for function_graph tracer. > + > SEE ALSO > -------- > linkperf:perf-record[1], linkperf:perf-trace[1] > diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c > index eba125a60820..876c8e800425 100644 > --- a/tools/perf/builtin-ftrace.c > +++ b/tools/perf/builtin-ftrace.c > @@ -43,6 +43,7 @@ struct perf_ftrace { > bool inherit; > bool func_stack_trace; > bool graph_nosleep_time; > + bool graph_noirqs; > }; > > struct filter_entry { > @@ -202,6 +203,7 @@ static void reset_tracing_options(struct perf_ftrace *ftrace __maybe_unused) > write_tracing_option_file("function-fork", "0"); > write_tracing_option_file("func_stack_trace", "0"); > write_tracing_option_file("sleep-time", "1"); > + write_tracing_option_file("funcgraph-irqs", "1"); > } > > static int reset_tracing_files(struct perf_ftrace *ftrace __maybe_unused) > @@ -393,6 +395,17 @@ static int set_tracing_sleep_time(struct perf_ftrace *ftrace) > return 0; > } > > +static int set_tracing_funcgraph_irqs(struct perf_ftrace *ftrace) > +{ > + if (!ftrace->graph_noirqs) > + return 0; > + > + if (write_tracing_option_file("funcgraph-irqs", "0") < 0) > + return -1; > + > + return 0; > +} > + > static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv) > { > char *trace_file; > @@ -477,6 +490,11 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv) > goto out_reset; > } > > + if (set_tracing_funcgraph_irqs(ftrace) < 0) { > + pr_err("failed to set tracing option funcgraph-irqs\n"); > + goto out_reset; > + }
Why not add set_tracing_options() which calls these individual functions to handle option processing altogether? It seems consistent to its 'reset' counterpart.
Thanks Namhyung
> + > if (write_tracing_file("current_tracer", ftrace->tracer) < 0) { > pr_err("failed to set current_tracer to %s\n", ftrace->tracer); > goto out_reset; > @@ -658,6 +676,8 @@ int cmd_ftrace(int argc, const char **argv) > "trace children processes"), > OPT_BOOLEAN(0, "graph-nosleep-time", &ftrace.graph_nosleep_time, > "measure on-CPU time only for function_graph tracer"), > + OPT_BOOLEAN(0, "graph-noirqs", &ftrace.graph_noirqs, > + "ignore functions that happen inside interrupt for function_graph tracer"), > OPT_END() > }; > > -- > 2.25.1 >
| |