All of lore.kernel.org
 help / color / mirror / Atom feed
From: Rob Herring <robh@kernel.org>
To: Jiri Olsa <jolsa@redhat.com>
Cc: Will Deacon <will@kernel.org>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@redhat.com>,
	Arnaldo Carvalho de Melo <acme@kernel.org>,
	linux-arm-kernel <linux-arm-kernel@lists.infradead.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Namhyung Kim <namhyung@kernel.org>,
	Raphael Gault <raphael.gault@arm.com>,
	Mark Rutland <mark.rutland@arm.com>,
	Jonathan Cameron <Jonathan.Cameron@huawei.com>,
	Ian Rogers <irogers@google.com>,
	Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>,
	Itaru Kitayama <itaru.kitayama@gmail.com>
Subject: Re: [PATCH v4 4/9] libperf: Add libperf_evsel__mmap()
Date: Fri, 6 Nov 2020 15:56:11 -0600	[thread overview]
Message-ID: <CAL_JsqJzeCebq4VP+xBtfh=fbomvaJoVMp35AQQDGTYD-fRWgw@mail.gmail.com> (raw)
In-Reply-To: <20201105224121.GA4112111@krava>

On Thu, Nov 5, 2020 at 4:41 PM Jiri Olsa <jolsa@redhat.com> wrote:
>
> On Thu, Nov 05, 2020 at 10:19:24AM -0600, Rob Herring wrote:
>
> SNIP
>
> > > > >
> > > > > that maps page for each event, then perf_evsel__read
> > > > > could go through the fast code, no?
> > > >
> > > > No, because we're not self-monitoring (pid == 0 and cpu == -1). With
> > > > the following change:
> > > >
> > > > diff --git a/tools/lib/perf/tests/test-evsel.c
> > > > b/tools/lib/perf/tests/test-evsel.c
> > > > index eeca8203d73d..1fca9c121f7c 100644
> > > > --- a/tools/lib/perf/tests/test-evsel.c
> > > > +++ b/tools/lib/perf/tests/test-evsel.c
> > > > @@ -17,6 +17,7 @@ static int test_stat_cpu(void)
> > > >  {
> > > >         struct perf_cpu_map *cpus;
> > > >         struct perf_evsel *evsel;
> > > > +       struct perf_event_mmap_page *pc;
> > > >         struct perf_event_attr attr = {
> > > >                 .type   = PERF_TYPE_SOFTWARE,
> > > >                 .config = PERF_COUNT_SW_CPU_CLOCK,
> > > > @@ -32,6 +33,15 @@ static int test_stat_cpu(void)
> > > >         err = perf_evsel__open(evsel, cpus, NULL);
> > > >         __T("failed to open evsel", err == 0);
> > > >
> > > > +       pc = perf_evsel__mmap(evsel, 0);
> > > > +       __T("failed to mmap evsel", pc);
> > > > +
> > > > +#if defined(__i386__) || defined(__x86_64__) || defined(__aarch64__)
> > > > +       __T("userspace counter access not supported", pc->cap_user_rdpmc);
> > > > +       __T("userspace counter access not enabled", pc->index);
> > > > +       __T("userspace counter width not set", pc->pmc_width >= 32);
> > > > +#endif
> > >
> > > I'll need to check, I'm surprised this would depend on the way
> > > you open the event
> >
> > Any more thoughts on this?
>
> sry I got stuck with other stuff.. I tried your change
> and pc->cap_user_rdpmc is 0 because the test creates
> software event, which does not support that

Sigh, yes, of course.

> when I change that to:
>
>         .type   = PERF_TYPE_HARDWARE,
>         .config = PERF_COUNT_HW_CPU_CYCLES,
>
> I don't see any of those warning you added

So I've now implemented the per fd mmap. It seems to run and get some
data, but for the above case the counts don't look right.

cpu0: count = 0x10883, ena = 0xbf42, run = 0xbf42
cpu1: count = 0x1bc65, ena = 0xa278, run = 0xa278
cpu2: count = 0x1fab2, ena = 0x91ea, run = 0x91ea
cpu3: count = 0x23d61, ena = 0x81ac, run = 0x81ac
cpu4: count = 0x2936a, ena = 0x7149, run = 0x7149
cpu5: count = 0x2cd4e, ena = 0x634f, run = 0x634f
cpu6: count = 0x3139f, ena = 0x53e7, run = 0x53e7
cpu7: count = 0x35350, ena = 0x4690, run = 0x4690

For comparison, this is what I get using the slow path read():
cpu0: count = 0x1c40, ena = 0x188b5, run = 0x188b5
cpu1: count = 0x18e0, ena = 0x1b8f4, run = 0x1b8f4
cpu2: count = 0x745e, ena = 0x1ab9e, run = 0x1ab9e
cpu3: count = 0x2416, ena = 0x1a280, run = 0x1a280
cpu4: count = 0x19c7, ena = 0x19b00, run = 0x19b00
cpu5: count = 0x1737, ena = 0x19262, run = 0x19262
cpu6: count = 0x11d0e, ena = 0x18944, run = 0x18944
cpu7: count = 0x20dbe, ena = 0x181f4, run = 0x181f4

The difference is we get a sequentially increasing count rather than 1
random CPU (the one running the test) with a much higher count. That
seems to me we're just reading the count for the calling process, not
each CPU.

For this to work correctly, cap_user_rdpmc would have to be set only
for the CPU's mmap that matches the calling process's CPU. I'm not
sure whether that can be done. Even if it can, is it really worth
doing so? You're accelerating reading an event on 1 out of N CPUs. And
what do we do on every kernel up til now this won't work on? Another
cap bit?

Rob

P.S. I did find one bug with all this. The shifts by pmc_width in the
read seq need to be a signed count. This test happens to have raw
counter values starting at 2^47.

WARNING: multiple messages have this Message-ID
From: Rob Herring <robh@kernel.org>
To: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>,
	Ian Rogers <irogers@google.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Catalin Marinas <catalin.marinas@arm.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	Arnaldo Carvalho de Melo <acme@kernel.org>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Raphael Gault <raphael.gault@arm.com>,
	Ingo Molnar <mingo@redhat.com>,
	Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>,
	Jonathan Cameron <Jonathan.Cameron@huawei.com>,
	Namhyung Kim <namhyung@kernel.org>,
	Itaru Kitayama <itaru.kitayama@gmail.com>,
	Will Deacon <will@kernel.org>,
	linux-arm-kernel <linux-arm-kernel@lists.infradead.org>
Subject: Re: [PATCH v4 4/9] libperf: Add libperf_evsel__mmap()
Date: Fri, 6 Nov 2020 15:56:11 -0600	[thread overview]
Message-ID: <CAL_JsqJzeCebq4VP+xBtfh=fbomvaJoVMp35AQQDGTYD-fRWgw@mail.gmail.com> (raw)
In-Reply-To: <20201105224121.GA4112111@krava>

On Thu, Nov 5, 2020 at 4:41 PM Jiri Olsa <jolsa@redhat.com> wrote:
>
> On Thu, Nov 05, 2020 at 10:19:24AM -0600, Rob Herring wrote:
>
> SNIP
>
> > > > >
> > > > > that maps page for each event, then perf_evsel__read
> > > > > could go through the fast code, no?
> > > >
> > > > No, because we're not self-monitoring (pid == 0 and cpu == -1). With
> > > > the following change:
> > > >
> > > > diff --git a/tools/lib/perf/tests/test-evsel.c
> > > > b/tools/lib/perf/tests/test-evsel.c
> > > > index eeca8203d73d..1fca9c121f7c 100644
> > > > --- a/tools/lib/perf/tests/test-evsel.c
> > > > +++ b/tools/lib/perf/tests/test-evsel.c
> > > > @@ -17,6 +17,7 @@ static int test_stat_cpu(void)
> > > >  {
> > > >         struct perf_cpu_map *cpus;
> > > >         struct perf_evsel *evsel;
> > > > +       struct perf_event_mmap_page *pc;
> > > >         struct perf_event_attr attr = {
> > > >                 .type   = PERF_TYPE_SOFTWARE,
> > > >                 .config = PERF_COUNT_SW_CPU_CLOCK,
> > > > @@ -32,6 +33,15 @@ static int test_stat_cpu(void)
> > > >         err = perf_evsel__open(evsel, cpus, NULL);
> > > >         __T("failed to open evsel", err == 0);
> > > >
> > > > +       pc = perf_evsel__mmap(evsel, 0);
> > > > +       __T("failed to mmap evsel", pc);
> > > > +
> > > > +#if defined(__i386__) || defined(__x86_64__) || defined(__aarch64__)
> > > > +       __T("userspace counter access not supported", pc->cap_user_rdpmc);
> > > > +       __T("userspace counter access not enabled", pc->index);
> > > > +       __T("userspace counter width not set", pc->pmc_width >= 32);
> > > > +#endif
> > >
> > > I'll need to check, I'm surprised this would depend on the way
> > > you open the event
> >
> > Any more thoughts on this?
>
> sry I got stuck with other stuff.. I tried your change
> and pc->cap_user_rdpmc is 0 because the test creates
> software event, which does not support that

Sigh, yes, of course.

> when I change that to:
>
>         .type   = PERF_TYPE_HARDWARE,
>         .config = PERF_COUNT_HW_CPU_CYCLES,
>
> I don't see any of those warning you added

So I've now implemented the per fd mmap. It seems to run and get some
data, but for the above case the counts don't look right.

cpu0: count = 0x10883, ena = 0xbf42, run = 0xbf42
cpu1: count = 0x1bc65, ena = 0xa278, run = 0xa278
cpu2: count = 0x1fab2, ena = 0x91ea, run = 0x91ea
cpu3: count = 0x23d61, ena = 0x81ac, run = 0x81ac
cpu4: count = 0x2936a, ena = 0x7149, run = 0x7149
cpu5: count = 0x2cd4e, ena = 0x634f, run = 0x634f
cpu6: count = 0x3139f, ena = 0x53e7, run = 0x53e7
cpu7: count = 0x35350, ena = 0x4690, run = 0x4690

For comparison, this is what I get using the slow path read():
cpu0: count = 0x1c40, ena = 0x188b5, run = 0x188b5
cpu1: count = 0x18e0, ena = 0x1b8f4, run = 0x1b8f4
cpu2: count = 0x745e, ena = 0x1ab9e, run = 0x1ab9e
cpu3: count = 0x2416, ena = 0x1a280, run = 0x1a280
cpu4: count = 0x19c7, ena = 0x19b00, run = 0x19b00
cpu5: count = 0x1737, ena = 0x19262, run = 0x19262
cpu6: count = 0x11d0e, ena = 0x18944, run = 0x18944
cpu7: count = 0x20dbe, ena = 0x181f4, run = 0x181f4

The difference is we get a sequentially increasing count rather than 1
random CPU (the one running the test) with a much higher count. That
seems to me we're just reading the count for the calling process, not
each CPU.

For this to work correctly, cap_user_rdpmc would have to be set only
for the CPU's mmap that matches the calling process's CPU. I'm not
sure whether that can be done. Even if it can, is it really worth
doing so? You're accelerating reading an event on 1 out of N CPUs. And
what do we do on every kernel up til now this won't work on? Another
cap bit?

Rob

P.S. I did find one bug with all this. The shifts by pmc_width in the
read seq need to be a signed count. This test happens to have raw
counter values starting at 2^47.

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

  reply	other threads:[~2020-11-06 21:56 UTC|newest]

Thread overview: 58+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-01 14:01 [PATCH v4 0/9] libperf and arm64 userspace counter access support Rob Herring
2020-10-01 14:01 ` Rob Herring
2020-10-01 14:01 ` [PATCH v4 1/9] arm64: pmu: Add function implementation to update event index in userpage Rob Herring
2020-10-01 14:01   ` Rob Herring
2020-10-01 14:01 ` [PATCH v4 2/9] arm64: perf: Enable pmu counter direct access for perf event on armv8 Rob Herring
2020-10-01 14:01   ` Rob Herring
2020-11-13 18:06   ` Mark Rutland
2020-11-13 18:06     ` Mark Rutland
2020-11-19 18:35     ` Rob Herring
2020-11-19 18:35       ` Rob Herring
2020-11-19 19:15     ` Will Deacon
2020-11-19 19:15       ` Will Deacon
2020-11-20 20:03       ` Rob Herring
2020-11-20 20:03         ` Rob Herring
2020-11-20 22:08         ` Rob Herring
2020-11-20 22:08           ` Rob Herring
2020-12-02 14:57         ` Rob Herring
2020-12-02 14:57           ` Rob Herring
2021-01-07  0:17           ` Rob Herring
2021-01-07  0:17             ` Rob Herring
2020-10-01 14:01 ` [PATCH v4 3/9] tools/include: Add an initial math64.h Rob Herring
2020-10-01 14:01   ` Rob Herring
2020-10-01 14:01 ` [PATCH v4 4/9] libperf: Add libperf_evsel__mmap() Rob Herring
2020-10-01 14:01   ` Rob Herring
2020-10-14 11:05   ` Jiri Olsa
2020-10-14 11:05     ` Jiri Olsa
2020-10-16 21:39     ` Rob Herring
2020-10-16 21:39       ` Rob Herring
2020-10-19 20:15       ` Jiri Olsa
2020-10-19 20:15         ` Jiri Olsa
2020-10-20 14:38         ` Rob Herring
2020-10-20 14:38           ` Rob Herring
2020-10-20 15:35           ` Jiri Olsa
2020-10-20 15:35             ` Jiri Olsa
2020-10-20 17:11             ` Rob Herring
2020-10-20 17:11               ` Rob Herring
2020-10-21 11:24               ` Jiri Olsa
2020-10-21 11:24                 ` Jiri Olsa
2020-11-05 16:19                 ` Rob Herring
2020-11-05 16:19                   ` Rob Herring
2020-11-05 22:41                   ` Jiri Olsa
2020-11-05 22:41                     ` Jiri Olsa
2020-11-06 21:56                     ` Rob Herring [this message]
2020-11-06 21:56                       ` Rob Herring
2020-11-11 12:00                       ` Jiri Olsa
2020-11-11 12:00                         ` Jiri Olsa
2020-11-11 14:50                         ` Rob Herring
2020-11-11 14:50                           ` Rob Herring
2020-10-01 14:01 ` [PATCH v4 5/9] libperf: tests: Add support for verbose printing Rob Herring
2020-10-01 14:01   ` Rob Herring
2020-10-01 14:01 ` [PATCH v4 6/9] libperf: Add support for user space counter access Rob Herring
2020-10-01 14:01   ` Rob Herring
2020-10-01 14:01 ` [PATCH v4 7/9] libperf: Add arm64 support to perf_mmap__read_self() Rob Herring
2020-10-01 14:01   ` Rob Herring
2020-10-01 14:01 ` [PATCH v4 8/9] perf: arm64: Add test for userspace counter access on heterogeneous systems Rob Herring
2020-10-01 14:01   ` Rob Herring
2020-10-01 14:01 ` [PATCH v4 9/9] Documentation: arm64: Document PMU counters access from userspace Rob Herring
2020-10-01 14:01   ` Rob Herring

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='CAL_JsqJzeCebq4VP+xBtfh=fbomvaJoVMp35AQQDGTYD-fRWgw@mail.gmail.com' \
    --to=robh@kernel.org \
    --cc=Jonathan.Cameron@huawei.com \
    --cc=acme@kernel.org \
    --cc=alexander.shishkin@linux.intel.com \
    --cc=catalin.marinas@arm.com \
    --cc=honnappa.nagarahalli@arm.com \
    --cc=irogers@google.com \
    --cc=itaru.kitayama@gmail.com \
    --cc=jolsa@redhat.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mark.rutland@arm.com \
    --cc=mingo@redhat.com \
    --cc=namhyung@kernel.org \
    --cc=peterz@infradead.org \
    --cc=raphael.gault@arm.com \
    --cc=will@kernel.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.