Re: [PATCH v8 04/18] perf ftrace: add option '-m/--buffer-size' to set per-cpu buffer size
From: Arnaldo Carvalho de Melo
Date: Fri Aug 14 2020 - 08:12:59 EST
Em Fri, Aug 14, 2020 at 08:53:18AM -0300, Arnaldo Carvalho de Melo escreveu:
> Em Sat, Aug 08, 2020 at 10:31:27AM +0800, Changbin Du escreveu:
> > This adds an option '-m/--buffer-size' to allow us set the size of per-cpu
> > tracing buffer.
> >
> > Signed-off-by: Changbin Du <changbin.du@xxxxxxxxx>
>
> So this is a little bit confusing, i.e. if I don't specify a suffix (B,
> K, M, G) I get:
>
> # perf ftrace -m 2048
>
> Usage: perf ftrace [<options>] [<command>]
> or: perf ftrace [<options>] -- <command> [<options>]
>
> -m, --buffer-size <size>
> size of per cpu buffer
> #
>
> Which is not very helpful, and you also forgot to add an entry to 'man
> perf-ftrace' (tools/perf/Documentation/perf-ftrace.txt), so only by
I must be blind, the entry _was_ added to perf-ftrace.txt, nevermind :-)
But the other points still stand :)
> looking at the patch I was able to figure out that a suffix is
> required.
>
> Now look how 'perf trace' works:
>
> # perf trace -m 2048 -e open* --max-events 1
> 0.772 sh/343169 openat(dfd: CWD, filename: /etc/ld.so.cache, flags: RDONLY|CLOEXEC) = 3
> #
>
> I.e. it doesn't require a unit suffix, which is more natural.
>
> # perf trace -h -m
>
> -m, --mmap-pages <pages>
> number of mmap data pages
>
> #
>
> It indicates that the argument is in units of 'pages', while you don't
> state that in:
>
> $ perf ftrace -h -m
>
> -m, --buffer-size <size>
> size of per cpu buffer
>
> $
>
> So please improve the option description and add its longer explanation
> in the man page, also I suggest you interpret:
>
> # perf ftrace -m 2048
>
> as asking for 2048K since the file as _kb in its name, so people used to
> ftrace will get what they ask for, i.e. N kb. Please add this to the one
> line description in 'perf ftrace -h', namely that the default unit os
> 'K'.
>
> All this can be made on top of this series, so I'm applying it now to
> make progress.
>
> - Arnaldo
>
> > ---
> > v2: support units as a suffix.
> > ---
> > tools/perf/Documentation/perf-ftrace.txt | 5 +++
> > tools/perf/builtin-ftrace.c | 55 ++++++++++++++++++++++++
> > 2 files changed, 60 insertions(+)
> >
> > diff --git a/tools/perf/Documentation/perf-ftrace.txt b/tools/perf/Documentation/perf-ftrace.txt
> > index 4f5628445a63..7a5d915f60b0 100644
> > --- a/tools/perf/Documentation/perf-ftrace.txt
> > +++ b/tools/perf/Documentation/perf-ftrace.txt
> > @@ -53,6 +53,11 @@ OPTIONS
> > Ranges of CPUs are specified with -: 0-2.
> > Default is to trace on all online CPUs.
> >
> > +-m::
> > +--buffer-size::
> > + Set the size of per-cpu tracing buffer, <size> is expected to
> > + be a number with appended unit character - B/K/M/G.
> > +
> > -T::
> > --trace-funcs=::
> > Select function tracer and set function filter on the given
> > diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
> > index 4b3fcee5725a..a3a4f4be9dde 100644
> > --- a/tools/perf/builtin-ftrace.c
> > +++ b/tools/perf/builtin-ftrace.c
> > @@ -26,6 +26,7 @@
> > #include "thread_map.h"
> > #include "util/cap.h"
> > #include "util/config.h"
> > +#include "util/units.h"
> >
> > #define DEFAULT_TRACER "function_graph"
> >
> > @@ -39,6 +40,7 @@ struct perf_ftrace {
> > struct list_head graph_funcs;
> > struct list_head nograph_funcs;
> > int graph_depth;
> > + unsigned long percpu_buffer_size;
> > };
> >
> > struct filter_entry {
> > @@ -324,6 +326,21 @@ static int set_tracing_depth(struct perf_ftrace *ftrace)
> > return 0;
> > }
> >
> > +static int set_tracing_percpu_buffer_size(struct perf_ftrace *ftrace)
> > +{
> > + int ret;
> > +
> > + if (ftrace->percpu_buffer_size == 0)
> > + return 0;
> > +
> > + ret = write_tracing_file_int("buffer_size_kb",
> > + ftrace->percpu_buffer_size / 1024);
> > + if (ret < 0)
> > + return ret;
> > +
> > + return 0;
> > +}
> > +
> > static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
> > {
> > char *trace_file;
> > @@ -388,6 +405,11 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
> > goto out_reset;
> > }
> >
> > + if (set_tracing_percpu_buffer_size(ftrace) < 0) {
> > + pr_err("failed to set tracing per-cpu buffer size\n");
> > + goto out_reset;
> > + }
> > +
> > if (write_tracing_file("current_tracer", ftrace->tracer) < 0) {
> > pr_err("failed to set current_tracer to %s\n", ftrace->tracer);
> > goto out_reset;
> > @@ -506,6 +528,37 @@ static void delete_filter_func(struct list_head *head)
> > }
> > }
> >
> > +static int parse_buffer_size(const struct option *opt,
> > + const char *str, int unset)
> > +{
> > + unsigned long *s = (unsigned long *)opt->value;
> > + static struct parse_tag tags_size[] = {
> > + { .tag = 'B', .mult = 1 },
> > + { .tag = 'K', .mult = 1 << 10 },
> > + { .tag = 'M', .mult = 1 << 20 },
> > + { .tag = 'G', .mult = 1 << 30 },
> > + { .tag = 0 },
> > + };
> > + unsigned long val;
> > +
> > + if (unset) {
> > + *s = 0;
> > + return 0;
> > + }
> > +
> > + val = parse_tag_value(str, tags_size);
> > + if (val != (unsigned long) -1) {
> > + if (val < 1024) {
> > + pr_err("buffer size too small, must larger than 1KB.");
> > + return -1;
> > + }
> > + *s = val;
> > + return 0;
> > + }
> > +
> > + return -1;
> > +}
> > +
> > static void select_tracer(struct perf_ftrace *ftrace)
> > {
> > bool graph = !list_empty(&ftrace->graph_funcs) ||
> > @@ -560,6 +613,8 @@ int cmd_ftrace(int argc, const char **argv)
> > "Set nograph filter on given functions", parse_filter_func),
> > OPT_INTEGER('D', "graph-depth", &ftrace.graph_depth,
> > "Max depth for function graph tracer"),
> > + OPT_CALLBACK('m', "buffer-size", &ftrace.percpu_buffer_size, "size",
> > + "size of per cpu buffer", parse_buffer_size),
> > OPT_END()
> > };
> >
> > --
> > 2.25.1
> >
>
> --
>
> - Arnaldo
--
- Arnaldo