Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1687552 > unrolled thread

[PATCH] perf/x86/intel: Fix debug_store reset field for freq events

Started byJiri Olsa <jolsa@kernel.org>
First post2017-07-14 18:40 +0200
Last post2017-07-18 14:30 +0200
Articles 7 — 4 participants

Back to article view | Back to linux.kernel

Contents

  [PATCH] perf/x86/intel: Fix debug_store reset field for freq events Jiri Olsa <jolsa@kernel.org> - 2017-07-14 18:40 +0200
    Re: [PATCH] perf/x86/intel: Fix debug_store reset field for freq events Andi Kleen <andi@firstfloor.org> - 2017-07-14 19:30 +0200
      Re: [PATCH] perf/x86/intel: Fix debug_store reset field for freq  events Jiri Olsa <jolsa@redhat.com> - 2017-07-17 09:40 +0200
      Re: [PATCH] perf/x86/intel: Fix debug_store reset field for freq  events Jiri Olsa <jolsa@redhat.com> - 2017-07-17 13:40 +0200
        Re: [PATCH] perf/x86/intel: Fix debug_store reset field for freq  events Andi Kleen <andi@firstfloor.org> - 2017-07-17 20:10 +0200
    [tip:perf/urgent] perf/x86/intel: Fix debug_store reset field for  freq events tip-bot for Jiri Olsa <tipbot@zytor.com> - 2017-07-18 13:00 +0200
    [tip:perf/urgent] perf/x86/intel: Fix debug_store reset field for  freq events tip-bot for Jiri Olsa <tipbot@zytor.com> - 2017-07-18 14:30 +0200

#1687552 — [PATCH] perf/x86/intel: Fix debug_store reset field for freq events

FromJiri Olsa <jolsa@kernel.org>
Date2017-07-14 18:40 +0200
Subject[PATCH] perf/x86/intel: Fix debug_store reset field for freq events
Message-ID<u3bNf-600-11@gated-at.bofh.it>
There's a bug in PEBs event enabling code, that prevents PEBS
freq events to work properly after non freq PEBS event was run.

freq events - perf_event_attr::freq set
              -F <freq> option of perf record

PEBS events - perf_event_attr::precise_ip > 0
              default for perf record

Like in following example with cpu 0 busy, we expect ~10000 samples
for following perf tool run:

  # perf record -F 10000 -C 0 sleep 1
  [ perf record: Woken up 2 times to write data ]
  [ perf record: Captured and wrote 0.640 MB perf.data (10031 samples) ]

Everything's fine, but once we run non freq PEBS event like:

  # perf record -c 10000 -C 0 sleep 1
  [ perf record: Woken up 4 times to write data ]
  [ perf record: Captured and wrote 1.053 MB perf.data (20061 samples) ]

the freq events start to fail like this:

  # perf record -F 10000 -C 0 sleep 1
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 0.185 MB perf.data (40 samples) ]

The issue is in non freq PEBs event initialization of debug_store reset
field, which value is used to auto-reload the counter value after PEBS
event drain. This value is not being used for PEBS freq events, but once
we run non freq event it stays in debug_store data and screws the
sample_freq counting for PEBS freq events.

Setting the reset field to 0 for freq events.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
---
 arch/x86/events/intel/ds.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/x86/events/intel/ds.c b/arch/x86/events/intel/ds.c
index c6d23ffe422d..2244bd8c09b1 100644
--- a/arch/x86/events/intel/ds.c
+++ b/arch/x86/events/intel/ds.c
@@ -889,6 +889,8 @@ void intel_pmu_pebs_enable(struct perf_event *event)
 	if (hwc->flags & PERF_X86_EVENT_AUTO_RELOAD) {
 		ds->pebs_event_reset[hwc->idx] =
 			(u64)(-hwc->sample_period) & x86_pmu.cntval_mask;
+	} else {
+		ds->pebs_event_reset[hwc->idx] = 0;
 	}
 }
 
-- 
2.9.4

[toc] | [next] | [standalone]

#1687578

FromAndi Kleen <andi@firstfloor.org>
Date2017-07-14 19:30 +0200
Message-ID<u3czE-6A0-13@gated-at.bofh.it>
In reply to#1687552
Jiri Olsa <jolsa@kernel.org> writes:
>
> Setting the reset field to 0 for freq events.

Looks good to me.

Reviewed-by: Andi Kleen <ak@linux.intel.com>

BTW I suspect there's a related bug that

perf record -e '{cycles:pp,branches}:S' ..

would enable multi record PEBS, even though it shouldn't because
we need the PMI to read the other events.

-Andi

[toc] | [prev] | [next] | [standalone]

#1688722 — Re: [PATCH] perf/x86/intel: Fix debug_store reset field for freq events

FromJiri Olsa <jolsa@redhat.com>
Date2017-07-17 09:40 +0200
SubjectRe: [PATCH] perf/x86/intel: Fix debug_store reset field for freq events
Message-ID<u48Nj-232-3@gated-at.bofh.it>
In reply to#1687578
On Fri, Jul 14, 2017 at 10:22:49AM -0700, Andi Kleen wrote:
> Jiri Olsa <jolsa@kernel.org> writes:
> >
> > Setting the reset field to 0 for freq events.
> 
> Looks good to me.
> 
> Reviewed-by: Andi Kleen <ak@linux.intel.com>
> 
> BTW I suspect there's a related bug that
> 
> perf record -e '{cycles:pp,branches}:S' ..
> 
> would enable multi record PEBS, even though it shouldn't because
> we need the PMI to read the other events.

thanks, I'll check on that

jirka

[toc] | [prev] | [next] | [standalone]

#1688945 — Re: [PATCH] perf/x86/intel: Fix debug_store reset field for freq events

FromJiri Olsa <jolsa@redhat.com>
Date2017-07-17 13:40 +0200
SubjectRe: [PATCH] perf/x86/intel: Fix debug_store reset field for freq events
Message-ID<u4cxz-4pb-13@gated-at.bofh.it>
In reply to#1687578
On Fri, Jul 14, 2017 at 10:22:49AM -0700, Andi Kleen wrote:
> Jiri Olsa <jolsa@kernel.org> writes:
> >
> > Setting the reset field to 0 for freq events.
> 
> Looks good to me.
> 
> Reviewed-by: Andi Kleen <ak@linux.intel.com>
> 
> BTW I suspect there's a related bug that
> 
> perf record -e '{cycles:pp,branches}:S' ..
> 
> would enable multi record PEBS, even though it shouldn't because
> we need the PMI to read the other events.

there's PERF_SAMPLE_READ om cycles's sample_type for this example
so it won't pass the x86_pmu::free_running_flags filter

also PERF_SAMPLE_TIME and PERF_SAMPLE_PERIOD will be set
in your example which will prevent that, but those
could be unset via record's '-c xxxx' and '--no-timestamp'

jirka

[toc] | [prev] | [next] | [standalone]

#1689317 — Re: [PATCH] perf/x86/intel: Fix debug_store reset field for freq events

FromAndi Kleen <andi@firstfloor.org>
Date2017-07-17 20:10 +0200
SubjectRe: [PATCH] perf/x86/intel: Fix debug_store reset field for freq events
Message-ID<u4iD0-8sS-21@gated-at.bofh.it>
In reply to#1688945
On Mon, Jul 17, 2017 at 01:37:58PM +0200, Jiri Olsa wrote:
> On Fri, Jul 14, 2017 at 10:22:49AM -0700, Andi Kleen wrote:
> > Jiri Olsa <jolsa@kernel.org> writes:
> > >
> > > Setting the reset field to 0 for freq events.
> > 
> > Looks good to me.
> > 
> > Reviewed-by: Andi Kleen <ak@linux.intel.com>
> > 
> > BTW I suspect there's a related bug that
> > 
> > perf record -e '{cycles:pp,branches}:S' ..
> > 
> > would enable multi record PEBS, even though it shouldn't because
> > we need the PMI to read the other events.
> 
> there's PERF_SAMPLE_READ om cycles's sample_type for this example
> so it won't pass the x86_pmu::free_running_flags filter

Good thanks for checking.

> 
> also PERF_SAMPLE_TIME and PERF_SAMPLE_PERIOD will be set
> in your example which will prevent that, but those
> could be unset via record's '-c xxxx' and '--no-timestamp'

PERF_SAMPLE_TIME works with Skylake/goldmont, but yes forgot 
the -c.

-Andi

[toc] | [prev] | [next] | [standalone]

#1690106 — [tip:perf/urgent] perf/x86/intel: Fix debug_store reset field for freq events

Fromtip-bot for Jiri Olsa <tipbot@zytor.com>
Date2017-07-18 13:00 +0200
Subject[tip:perf/urgent] perf/x86/intel: Fix debug_store reset field for freq events
Message-ID<u4yor-1lh-39@gated-at.bofh.it>
In reply to#1687552
Commit-ID:  b32fd2f3f667c8a94f2f6bc2fb88a607be8f3229
Gitweb:     http://git.kernel.org/tip/b32fd2f3f667c8a94f2f6bc2fb88a607be8f3229
Author:     Jiri Olsa <jolsa@kernel.org>
AuthorDate: Fri, 14 Jul 2017 18:35:51 +0200
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Tue, 18 Jul 2017 11:06:12 +0200

perf/x86/intel: Fix debug_store reset field for freq events

There's a bug in PEBs event enabling code, that prevents PEBS
freq events to work properly after non freq PEBS event was run.

freq events - perf_event_attr::freq set
              -F <freq> option of perf record

PEBS events - perf_event_attr::precise_ip > 0
              default for perf record

Like in following example with CPU 0 busy, we expect ~10000 samples
for following perf tool run:

  # perf record -F 10000 -C 0 sleep 1
  [ perf record: Woken up 2 times to write data ]
  [ perf record: Captured and wrote 0.640 MB perf.data (10031 samples) ]

Everything's fine, but once we run non freq PEBS event like:

  # perf record -c 10000 -C 0 sleep 1
  [ perf record: Woken up 4 times to write data ]
  [ perf record: Captured and wrote 1.053 MB perf.data (20061 samples) ]

the freq events start to fail like this:

  # perf record -F 10000 -C 0 sleep 1
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 0.185 MB perf.data (40 samples) ]

The issue is in non freq PEBs event initialization of debug_store reset
field, which value is used to auto-reload the counter value after PEBS
event drain. This value is not being used for PEBS freq events, but once
we run non freq event it stays in debug_store data and screws the
sample_freq counting for PEBS freq events.

Setting the reset field to 0 for freq events.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20170714163551.19459-1-jolsa@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/events/intel/ds.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/x86/events/intel/ds.c b/arch/x86/events/intel/ds.c
index 2ca4d2d..6dc8a59 100644
--- a/arch/x86/events/intel/ds.c
+++ b/arch/x86/events/intel/ds.c
@@ -895,6 +895,8 @@ void intel_pmu_pebs_enable(struct perf_event *event)
 	if (hwc->flags & PERF_X86_EVENT_AUTO_RELOAD) {
 		ds->pebs_event_reset[hwc->idx] =
 			(u64)(-hwc->sample_period) & x86_pmu.cntval_mask;
+	} else {
+		ds->pebs_event_reset[hwc->idx] = 0;
 	}
 }

[toc] | [prev] | [next] | [standalone]

#1690224 — [tip:perf/urgent] perf/x86/intel: Fix debug_store reset field for freq events

Fromtip-bot for Jiri Olsa <tipbot@zytor.com>
Date2017-07-18 14:30 +0200
Subject[tip:perf/urgent] perf/x86/intel: Fix debug_store reset field for freq events
Message-ID<u4zNw-2nc-31@gated-at.bofh.it>
In reply to#1687552
Commit-ID:  dc853e26f73e903e0c87e24f2695b5dcf33b3bc1
Gitweb:     http://git.kernel.org/tip/dc853e26f73e903e0c87e24f2695b5dcf33b3bc1
Author:     Jiri Olsa <jolsa@kernel.org>
AuthorDate: Fri, 14 Jul 2017 18:35:51 +0200
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Tue, 18 Jul 2017 14:13:41 +0200

perf/x86/intel: Fix debug_store reset field for freq events

There's a bug in PEBs event enabling code, that prevents PEBS
freq events to work properly after non freq PEBS event was run.

freq events - perf_event_attr::freq set
              -F <freq> option of perf record

PEBS events - perf_event_attr::precise_ip > 0
              default for perf record

Like in following example with CPU 0 busy, we expect ~10000 samples
for following perf tool run:

  # perf record -F 10000 -C 0 sleep 1
  [ perf record: Woken up 2 times to write data ]
  [ perf record: Captured and wrote 0.640 MB perf.data (10031 samples) ]

Everything's fine, but once we run non freq PEBS event like:

  # perf record -c 10000 -C 0 sleep 1
  [ perf record: Woken up 4 times to write data ]
  [ perf record: Captured and wrote 1.053 MB perf.data (20061 samples) ]

the freq events start to fail like this:

  # perf record -F 10000 -C 0 sleep 1
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 0.185 MB perf.data (40 samples) ]

The issue is in non freq PEBs event initialization of debug_store reset
field, which value is used to auto-reload the counter value after PEBS
event drain. This value is not being used for PEBS freq events, but once
we run non freq event it stays in debug_store data and screws the
sample_freq counting for PEBS freq events.

Setting the reset field to 0 for freq events.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20170714163551.19459-1-jolsa@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/events/intel/ds.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/x86/events/intel/ds.c b/arch/x86/events/intel/ds.c
index 2ca4d2d..6dc8a59 100644
--- a/arch/x86/events/intel/ds.c
+++ b/arch/x86/events/intel/ds.c
@@ -895,6 +895,8 @@ void intel_pmu_pebs_enable(struct perf_event *event)
 	if (hwc->flags & PERF_X86_EVENT_AUTO_RELOAD) {
 		ds->pebs_event_reset[hwc->idx] =
 			(u64)(-hwc->sample_period) & x86_pmu.cntval_mask;
+	} else {
+		ds->pebs_event_reset[hwc->idx] = 0;
 	}
 }

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web