From: Ian Rogers <irogers@google.com>
To: tcwzxx <tcwzxx@gmail.com>
Cc: peterz@infradead.org, mingo@redhat.com, acme@kernel.org,
alexander.shishkin@linux.intel.com, jolsa@kernel.org,
linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] perf tool: add skip_not_exec_file_map_events option
Date: Wed, 17 Aug 2022 11:13:23 -0700 [thread overview]
Message-ID: <CAP-5=fVrpsid2_Ku=5hHeFSQ5Q1zE--5qszWw38ZVciq9Rxd5A@mail.gmail.com> (raw)
In-Reply-To: <20220817134324.702278-1-tcwzxx@gmail.com>
On Wed, Aug 17, 2022 at 6:43 AM tcwzxx <tcwzxx@gmail.com> wrote:
>
> When generate the flame graph, the perf-script subcommand will
> process all mmap event and add them to the rbtree. The 240,000
> mmap region takes about 5 minutes, which is not useful for flame
> graph. So add the skip-not-exec-file-map-events option to skip
> not PROT_EXEC flag memory regions.
>
> Signed-off-by: tcwzxx <tcwzxx@gmail.com>
Could you provide more context? A reproduction?
When we synthesize mmap events we drop non-executable pages:
https://github.com/torvalds/linux/blob/master/tools/perf/util/synthetic-events.c#L466
Similarly in the kernel for the "dummy" event:
https://github.com/torvalds/linux/blob/master/kernel/events/core.c#L8258
Thanks,
Ian
> ---
> tools/perf/builtin-report.c | 2 ++
> tools/perf/builtin-script.c | 3 +++
> tools/perf/util/machine.c | 3 +++
> tools/perf/util/map.c | 5 +++++
> tools/perf/util/map.h | 2 ++
> 5 files changed, 15 insertions(+)
>
> diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
> index 91ed41cc7d88..c28eb9450a66 100644
> --- a/tools/perf/builtin-report.c
> +++ b/tools/perf/builtin-report.c
> @@ -1364,6 +1364,8 @@ int cmd_report(int argc, const char **argv)
> "Disable raw trace ordering"),
> OPT_BOOLEAN(0, "skip-empty", &report.skip_empty,
> "Do not display empty (or dummy) events in the output"),
> + OPT_BOOLEAN(0, "skip-not-exec-file-map_events", &skip_not_exec_file_map_events,
> + "skip not exec map events"),
> OPT_END()
> };
> struct perf_data data = {
> diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
> index 13580a9c50b8..e3f4e5e654c9 100644
> --- a/tools/perf/builtin-script.c
> +++ b/tools/perf/builtin-script.c
> @@ -32,6 +32,7 @@
> #include "util/time-utils.h"
> #include "util/path.h"
> #include "util/event.h"
> +#include "util/map.h"
> #include "ui/ui.h"
> #include "print_binary.h"
> #include "archinsn.h"
> @@ -3936,6 +3937,8 @@ int cmd_script(int argc, const char **argv)
> "Guest code can be found in hypervisor process"),
> OPT_BOOLEAN('\0', "stitch-lbr", &script.stitch_lbr,
> "Enable LBR callgraph stitching approach"),
> + OPT_BOOLEAN(0, "skip-not-exec-map_events", &skip_not_exec_file_map_events,
> + "skip not exec map events"),
> OPTS_EVSWITCH(&script.evswitch),
> OPT_END()
> };
> diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c
> index 2a16cae28407..21dde9f9935c 100644
> --- a/tools/perf/util/machine.c
> +++ b/tools/perf/util/machine.c
> @@ -1904,6 +1904,9 @@ int machine__process_mmap2_event(struct machine *machine,
> if (thread == NULL)
> goto out_problem;
>
> + if (skip_not_exec_file_map_events && !(event->mmap2.prot & PROT_EXEC))
> + goto out_problem;
> +
> map = map__new(machine, event->mmap2.start,
> event->mmap2.len, event->mmap2.pgoff,
> &dso_id, event->mmap2.prot,
> diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c
> index e0aa4a254583..2b51ca012c91 100644
> --- a/tools/perf/util/map.c
> +++ b/tools/perf/util/map.c
> @@ -16,6 +16,8 @@
> #include "thread.h"
> #include "vdso.h"
>
> +bool skip_not_exec_file_map_events;
> +
> static inline int is_android_lib(const char *filename)
> {
> return strstarts(filename, "/data/app-lib/") ||
> @@ -168,6 +170,9 @@ struct map *map__new(struct machine *machine, u64 start, u64 len,
> if (dso == NULL)
> goto out_delete;
>
> + if (skip_not_exec_file_map_events && dso__type(dso, machine) == DSO__TYPE_UNKNOWN)
> + goto out_delete;
> +
> map__init(map, start, start + len, pgoff, dso);
>
> if (anon || no_dso) {
> diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h
> index 3dcfe06db6b3..67b0e0f9f0ae 100644
> --- a/tools/perf/util/map.h
> +++ b/tools/perf/util/map.h
> @@ -11,6 +11,8 @@
> #include <stdbool.h>
> #include <linux/types.h>
>
> +extern bool skip_not_exec_file_map_events;
> +
> struct dso;
> struct maps;
> struct machine;
> --
> 2.34.1
>
next prev parent reply other threads:[~2022-08-17 18:13 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-08-17 13:43 [PATCH] perf tool: add skip_not_exec_file_map_events option tcwzxx
2022-08-17 18:13 ` Ian Rogers [this message]
2022-08-18 3:26 ` lika you
2022-08-18 15:00 ` Ian Rogers
2022-08-19 3:18 ` zhizhi xu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAP-5=fVrpsid2_Ku=5hHeFSQ5Q1zE--5qszWw38ZVciq9Rxd5A@mail.gmail.com' \
--to=irogers@google.com \
--cc=acme@kernel.org \
--cc=alexander.shishkin@linux.intel.com \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=tcwzxx@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).