All of lore.kernel.org
 help / color / mirror / Atom feed
From: Daniel Wagner <dwagner@suse.de>
To: Clark Williams <williams@redhat.com>, John Kacur <jkacur@redhat.com>
Cc: linux-rt-users@vger.kernel.org, Daniel Wagner <dwagner@suse.de>
Subject: [RFC rt-tests v1 0/4] Generate machine-readable output
Date: Fri, 18 Dec 2020 20:35:36 +0100	[thread overview]
Message-ID: <20201218193540.6168-1-dwagner@suse.de> (raw)

The current output of cyclictest is optimized for humans to read. This
is all good when working directly with the tools. But for CI
integration it's a bit of pain. Furthermore, not all rt-tests use the
same output format.

By using some easy to parse existing machine-readable format we can use
standard libraries to parse the data. For example in jitterdebug there
is a short Python program[1] to visualize either the histogram[2] or
all samples[3].

The implementation for JSON output for this is very simple. The last
patch adds a version of jitterdebugs's JSON output, which looks like

{
  "file_version": 1,
  "version:": "cyclictest V 1.90",
  "num_threads": 2,
  "resolution_in_ns": 0,
  "cmdline:": "./cyclictest --affinity=1-2 --duration=1s --output=dump.json -h 1000 -p 80",
  "sysinfo": {
    "sysname": "Linux",
    "nodename": "beryllium",
    "release": "5.9.14-1-default",
    "version": "#1 SMP Sat Dec 12 06:57:32 UTC 2020 (c648a46)",
    "machine": "x86_64"
  },
  "thread": {
    "0": {
      "histogram": {
        "0": 16,
        "1": 853,
        "2": 80,
        "3": 50,
        "4": 1
      },
      "cycles": 1000,
      "min": 0,
      "max": 4,
      "avg": 1.17,
      "cpu": 1,
      "node": 0
    },
    "1": {
      "histogram": {
        "0": 14,
        "1": 833,
        "2": 93,
        "3": 56,
        "4": 4
      },
      "cycles": 1000,
      "min": 0,
      "max": 4,
      "avg": 1.20,
      "cpu": 2,
      "node": 0
    }
  }
}

It's just a rough version. I didn't try to make it generic for the
other rt-tests or make it as plugin as John was suggesting. I'd think
we could make this feature as compile option, if you want to keep the
program small. Obviously, we could also make the terminal output a
compile option, to keep it small.

Anyway, what do you think about it?

Thanks,
Daniel

[1] https://github.com/igaw/jitterdebugger/blob/master/jitterplot
[2] https://www.monom.org/data/jitterdebug/jitterplot-hist.png
[3] https://www.monom.org/data/jitterdebug/jitterplot-samples.png

Daniel Wagner (4):
  rt-tests: Rename error.h to rt-error.h
  cyclictest: Move thread data to struct thread_param
  signaltest: Move thread data to struct thread_param
  cyclictest: Add JSON output feature

 src/cyclictest/cyclictest.c         | 182 +++++++++++++++++++++++++---
 src/include/pip_stress.h            |   5 +-
 src/include/{error.h => rt-error.h} |   0
 src/lib/error.c                     |   2 +-
 src/lib/rt-numa.c                   |   2 +-
 src/lib/rt-utils.c                  |   3 +-
 src/oslat/oslat.c                   |   2 +-
 src/pi_tests/pi_stress.c            |   3 +-
 src/pmqtest/pmqtest.c               |   6 +-
 src/ptsematest/ptsematest.c         |   6 +-
 src/signaltest/signaltest.c         |  46 +++----
 src/sigwaittest/sigwaittest.c       |   2 +-
 src/svsematest/svsematest.c         |   3 +-
 13 files changed, 204 insertions(+), 58 deletions(-)
 rename src/include/{error.h => rt-error.h} (100%)

-- 
2.29.2


             reply	other threads:[~2020-12-18 19:36 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-12-18 19:35 Daniel Wagner [this message]
2020-12-18 19:35 ` [RFC rt-tests v1 1/4] rt-tests: Rename error.h to rt-error.h Daniel Wagner
2020-12-18 19:35 ` [RFC rt-tests v1 2/4] cyclictest: Move thread data to struct thread_param Daniel Wagner
2020-12-18 19:35 ` [RFC rt-tests v1 3/4] signaltest: " Daniel Wagner
2020-12-18 19:35 ` [RFC rt-tests v1 4/4] cyclictest: Add JSON output feature Daniel Wagner

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=20201218193540.6168-1-dwagner@suse.de \
    --to=dwagner@suse.de \
    --cc=jkacur@redhat.com \
    --cc=linux-rt-users@vger.kernel.org \
    --cc=williams@redhat.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 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.