From: Adrian Hunter <adrian.hunter-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
To: Pawel Moll <pawel.moll-5wv7dgnIgG8@public.gmane.org>,
Richard Cochran
<richardcochran-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>,
Steven Rostedt <rostedt-nx8X9YLhiw1AfugRpC6u6w@public.gmane.org>,
Ingo Molnar <mingo-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>,
Peter Zijlstra <peterz-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>,
Paul Mackerras <paulus-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org>,
Arnaldo Carvalho de Melo
<acme-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>,
John Stultz <john.stultz-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>,
Masami Hiramatsu
<masami.hiramatsu.pt-FCd8Q96Dh0JBDgjK7y7TUQ@public.gmane.org>,
Christopher Covington
<cov-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org>,
Namhyung Kim <namhyung-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>,
David Ahern <dsahern-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>,
Thomas Gleixner <tglx-hfZtesqFncYOwBW4kG4KsQ@public.gmane.org>,
Tomeu Vizoso <tomeu-XCtybt49RKsYaV1qd6yewg@public.gmane.org>
Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-api-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Subject: Re: [PATCH v4 1/3] perf: Use monotonic clock as a source for timestamps
Date: Fri, 16 Jan 2015 14:41:55 +0200 [thread overview]
Message-ID: <54B90713.1090503@intel.com> (raw)
In-Reply-To: <1415292718-19785-2-git-send-email-pawel.moll-5wv7dgnIgG8@public.gmane.org>
On 06/11/14 18:51, Pawel Moll wrote:
> Until now, perf framework never defined the meaning of the timestamps
> captured as PERF_SAMPLE_TIME sample type. The values were obtaining
> from local (sched) clock, which is unavailable in userspace. This made
> it impossible to correlate perf data with any other events. Other
> tracing solutions have the source configurable (ftrace) or just share
> a common time domain between kernel and userspace (LTTng).
>
> Follow the trend by using monotonic clock, which is readily available
> as POSIX CLOCK_MONOTONIC.
>
> Also add a sysctl "perf_sample_time_clk_id" attribute which can be used
> by the user to obtain the clk_id to be used with POSIX clock API (eg.
> clock_gettime()) to obtain a time value comparable with perf samples.
>
> Old behaviour can be restored by using "perf_use_local_clock" kernel
> parameter.
>
Don't forget this breaks the relationship to TSC. So you will need something
like below. Also patch 3 needs to be done first and extended to cover TSC so
there is no gap when we cannot get TSC.
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 143e5f5..b6e833d 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1922,21 +1922,6 @@ void arch_perf_update_userpage(struct
perf_event_mmap_page *userpg, u64 now)
userpg->cap_user_time_zero = 0;
userpg->cap_user_rdpmc = x86_pmu.attr_rdpmc;
userpg->pmc_width = x86_pmu.cntval_bits;
-
- if (!sched_clock_stable())
- return;
-
- data = cyc2ns_read_begin();
-
- userpg->cap_user_time = 1;
- userpg->time_mult = data->cyc2ns_mul;
- userpg->time_shift = data->cyc2ns_shift;
- userpg->time_offset = data->cyc2ns_offset - now;
-
- userpg->cap_user_time_zero = 1;
- userpg->time_zero = data->cyc2ns_offset;
-
- cyc2ns_read_end(data);
}
/*
WARNING: multiple messages have this Message-ID (diff)
From: Adrian Hunter <adrian.hunter@intel.com>
To: Pawel Moll <pawel.moll@arm.com>,
Richard Cochran <richardcochran@gmail.com>,
Steven Rostedt <rostedt@goodmis.org>,
Ingo Molnar <mingo@redhat.com>,
Peter Zijlstra <peterz@infradead.org>,
Paul Mackerras <paulus@samba.org>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
John Stultz <john.stultz@linaro.org>,
Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>,
Christopher Covington <cov@codeaurora.org>,
Namhyung Kim <namhyung@kernel.org>,
David Ahern <dsahern@gmail.com>,
Thomas Gleixner <tglx@linutronix.de>,
Tomeu Vizoso <tomeu@tomeuvizoso.net>
Cc: linux-kernel@vger.kernel.org, linux-api@vger.kernel.org
Subject: Re: [PATCH v4 1/3] perf: Use monotonic clock as a source for timestamps
Date: Fri, 16 Jan 2015 14:41:55 +0200 [thread overview]
Message-ID: <54B90713.1090503@intel.com> (raw)
In-Reply-To: <1415292718-19785-2-git-send-email-pawel.moll@arm.com>
On 06/11/14 18:51, Pawel Moll wrote:
> Until now, perf framework never defined the meaning of the timestamps
> captured as PERF_SAMPLE_TIME sample type. The values were obtaining
> from local (sched) clock, which is unavailable in userspace. This made
> it impossible to correlate perf data with any other events. Other
> tracing solutions have the source configurable (ftrace) or just share
> a common time domain between kernel and userspace (LTTng).
>
> Follow the trend by using monotonic clock, which is readily available
> as POSIX CLOCK_MONOTONIC.
>
> Also add a sysctl "perf_sample_time_clk_id" attribute which can be used
> by the user to obtain the clk_id to be used with POSIX clock API (eg.
> clock_gettime()) to obtain a time value comparable with perf samples.
>
> Old behaviour can be restored by using "perf_use_local_clock" kernel
> parameter.
>
Don't forget this breaks the relationship to TSC. So you will need something
like below. Also patch 3 needs to be done first and extended to cover TSC so
there is no gap when we cannot get TSC.
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 143e5f5..b6e833d 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1922,21 +1922,6 @@ void arch_perf_update_userpage(struct
perf_event_mmap_page *userpg, u64 now)
userpg->cap_user_time_zero = 0;
userpg->cap_user_rdpmc = x86_pmu.attr_rdpmc;
userpg->pmc_width = x86_pmu.cntval_bits;
-
- if (!sched_clock_stable())
- return;
-
- data = cyc2ns_read_begin();
-
- userpg->cap_user_time = 1;
- userpg->time_mult = data->cyc2ns_mul;
- userpg->time_shift = data->cyc2ns_shift;
- userpg->time_offset = data->cyc2ns_offset - now;
-
- userpg->cap_user_time_zero = 1;
- userpg->time_zero = data->cyc2ns_offset;
-
- cyc2ns_read_end(data);
}
/*
next prev parent reply other threads:[~2015-01-16 12:41 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-11-06 16:51 [PATCH v4 0/3] perf: User/kernel time correlation and event generation Pawel Moll
2014-11-06 16:51 ` [PATCH v4 1/3] perf: Use monotonic clock as a source for timestamps Pawel Moll
[not found] ` <1415292718-19785-2-git-send-email-pawel.moll-5wv7dgnIgG8@public.gmane.org>
2014-11-27 15:05 ` Pawel Moll
2014-11-27 15:05 ` Pawel Moll
[not found] ` <1417100750.4371.1.camel-5wv7dgnIgG8@public.gmane.org>
2014-12-11 13:39 ` Pawel Moll
2014-12-11 13:39 ` Pawel Moll
[not found] ` <1418305153.4037.1.camel-5wv7dgnIgG8@public.gmane.org>
2015-01-05 13:01 ` Peter Zijlstra
2015-01-05 13:01 ` Peter Zijlstra
2015-01-21 15:47 ` Pawel Moll
2015-01-16 12:41 ` Adrian Hunter [this message]
2015-01-16 12:41 ` Adrian Hunter
2015-01-21 20:27 ` [PATCH v5] " Pawel Moll
2015-01-21 20:27 ` Pawel Moll
[not found] ` <1421872037-12559-1-git-send-email-pawel.moll-5wv7dgnIgG8@public.gmane.org>
2015-02-02 16:52 ` Pawel Moll
2015-02-02 16:52 ` Pawel Moll
[not found] ` <CAN+dfcT_6zZZ4oeyngUE5N0Wtx2B9CvXsfU71m+cuyXpq2KBdw@mail.gmail.com>
[not found] ` <CAN+dfcT_6zZZ4oeyngUE5N0Wtx2B9CvXsfU71m+cuyXpq2KBdw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2015-02-03 9:20 ` Pawel Moll
2015-02-03 9:20 ` Pawel Moll
[not found] ` <1422955245.4944.26.camel-5wv7dgnIgG8@public.gmane.org>
2015-02-11 16:12 ` Peter Zijlstra
2015-02-11 16:12 ` Peter Zijlstra
[not found] ` <20150211161256.GH2896-IIpfhp3q70z/8w/KjCw3T+5/BudmfyzbbVWyRVo5IupeoWH0uzbU5w@public.gmane.org>
2015-02-12 10:04 ` Adrian Hunter
2015-02-12 10:04 ` Adrian Hunter
[not found] ` <54DC7AC6.5010605-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2015-02-12 10:28 ` Peter Zijlstra
2015-02-12 10:28 ` Peter Zijlstra
[not found] ` <20150212102814.GK2896-IIpfhp3q70z/8w/KjCw3T+5/BudmfyzbbVWyRVo5IupeoWH0uzbU5w@public.gmane.org>
2015-02-12 15:38 ` Peter Zijlstra
2015-02-12 15:38 ` Peter Zijlstra
[not found] ` <20150212153808.GR24151-ndre7Fmf5hadTX5a5knrm8zTDFooKrT+cvkQGrU6aU0@public.gmane.org>
2015-02-13 0:25 ` John Stultz
2015-02-13 0:25 ` John Stultz
2015-02-13 7:07 ` Adrian Hunter
2015-02-13 7:07 ` Adrian Hunter
2015-01-05 13:00 ` [PATCH v4 1/3] " Peter Zijlstra
[not found] ` <20150105130035.GP30905-ndre7Fmf5hadTX5a5knrm8zTDFooKrT+cvkQGrU6aU0@public.gmane.org>
2015-01-21 15:52 ` Pawel Moll
2015-01-21 15:52 ` Pawel Moll
[not found] ` <1421855543.14076.68.camel-5wv7dgnIgG8@public.gmane.org>
2015-01-21 19:48 ` Pawel Moll
2015-01-21 19:48 ` Pawel Moll
[not found] ` <1421869684.14076.105.camel-5wv7dgnIgG8@public.gmane.org>
2015-01-21 20:07 ` Pawel Moll
2015-01-21 20:07 ` Pawel Moll
[not found] ` <1415292718-19785-1-git-send-email-pawel.moll-5wv7dgnIgG8@public.gmane.org>
2014-11-06 16:51 ` [PATCH v4 2/3] perf: Userspace event Pawel Moll
2014-11-06 16:51 ` Pawel Moll
[not found] ` <1415292718-19785-3-git-send-email-pawel.moll-5wv7dgnIgG8@public.gmane.org>
2015-01-05 13:12 ` Peter Zijlstra
2015-01-05 13:12 ` Peter Zijlstra
[not found] ` <20150105131237.GR30905-ndre7Fmf5hadTX5a5knrm8zTDFooKrT+cvkQGrU6aU0@public.gmane.org>
2015-01-21 16:01 ` Pawel Moll
2015-01-21 16:01 ` Pawel Moll
2014-11-06 16:51 ` [PATCH v4 3/3] perf: Sample additional clock value Pawel Moll
[not found] ` <1415292718-19785-4-git-send-email-pawel.moll-5wv7dgnIgG8@public.gmane.org>
2015-01-05 13:45 ` Peter Zijlstra
2015-01-05 13:45 ` Peter Zijlstra
2015-01-05 19:17 ` John Stultz
[not found] ` <20150105134514.GS30905-ndre7Fmf5hadTX5a5knrm8zTDFooKrT+cvkQGrU6aU0@public.gmane.org>
2015-01-21 17:12 ` Pawel Moll
2015-01-21 17:12 ` Pawel Moll
[not found] ` <1421860365.14076.91.camel-5wv7dgnIgG8@public.gmane.org>
2015-01-21 17:44 ` John Stultz
2015-01-21 17:44 ` John Stultz
[not found] ` <CALAqxLV6ggCO81ntWuYuDjrqNeMePkZBpq92G9-iwMHm7ONuFg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2015-01-21 17:54 ` Pawel Moll
2015-01-21 17:54 ` Pawel Moll
[not found] ` <1421862883.14076.99.camel-5wv7dgnIgG8@public.gmane.org>
2015-01-21 18:05 ` John Stultz
2015-01-21 18:05 ` John Stultz
2015-01-23 17:06 ` Pawel Moll
2015-01-23 17:06 ` Pawel Moll
[not found] ` <1422032767.14076.151.camel-5wv7dgnIgG8@public.gmane.org>
2015-01-23 18:05 ` David Ahern
2015-01-23 18:05 ` David Ahern
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=54B90713.1090503@intel.com \
--to=adrian.hunter-ral2jqcrhueavxtiumwx3w@public.gmane.org \
--cc=acme-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org \
--cc=cov-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org \
--cc=dsahern-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org \
--cc=john.stultz-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org \
--cc=linux-api-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
--cc=linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
--cc=masami.hiramatsu.pt-FCd8Q96Dh0JBDgjK7y7TUQ@public.gmane.org \
--cc=mingo-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
--cc=namhyung-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org \
--cc=paulus-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org \
--cc=pawel.moll-5wv7dgnIgG8@public.gmane.org \
--cc=peterz-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org \
--cc=richardcochran-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org \
--cc=rostedt-nx8X9YLhiw1AfugRpC6u6w@public.gmane.org \
--cc=tglx-hfZtesqFncYOwBW4kG4KsQ@public.gmane.org \
--cc=tomeu-XCtybt49RKsYaV1qd6yewg@public.gmane.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.