public inbox for isar-users@googlegroups.com
 help / color / mirror / Atom feed
From: Henning Schild <henning.schild@siemens.com>
To: Uladzimir Bely <ubely@ilbers.de>
Cc: isar-users@googlegroups.com
Subject: Re: [PATCH v2 1/4] buildstats: Borrow buildstats and pybootchartgui from OE
Date: Tue, 5 Oct 2021 11:13:33 +0200	[thread overview]
Message-ID: <20211005111333.6153efc5@md1za8fc.ad001.siemens.net> (raw)
In-Reply-To: <20210927141705.25386-2-ubely@ilbers.de>

Am Mon, 27 Sep 2021 16:17:02 +0200
schrieb Uladzimir Bely <ubely@ilbers.de>:

> Buildstats is a module in OpenEmbedded that collects build statistics.
> After that, it can be converted with `pybootchartgui` script into
> human-readable graphic form.
> 
> This patch just borrows the required files from openembedded-core
> (commit b71d30aef5dc: pybootchart/draw: Avoid divide by zero error).

I hope that is _all_ exactly that.

Henning
 
> Signed-off-by: Uladzimir Bely <ubely@ilbers.de>
> ---
>  meta/classes/buildstats.bbclass               | 295 ++++++
>  meta/lib/buildstats.py                        | 161 +++
>  scripts/pybootchartgui/AUTHORS                |  11 +
>  scripts/pybootchartgui/COPYING                | 340 ++++++
>  scripts/pybootchartgui/MAINTAINERS            |   3 +
>  scripts/pybootchartgui/NEWS                   | 204 ++++
>  scripts/pybootchartgui/README.pybootchart     |  37 +
>  scripts/pybootchartgui/pybootchartgui.py      |  23 +
>  .../pybootchartgui/pybootchartgui/__init__.py |   0
>  .../pybootchartgui/pybootchartgui/batch.py    |  46 +
>  scripts/pybootchartgui/pybootchartgui/draw.py | 975
> ++++++++++++++++++ scripts/pybootchartgui/pybootchartgui/gui.py  |
> 348 +++++++ scripts/pybootchartgui/pybootchartgui/main.py |   1 +
>  .../pybootchartgui/pybootchartgui/main.py.in  | 183 ++++
>  .../pybootchartgui/pybootchartgui/parsing.py  | 821 +++++++++++++++
>  .../pybootchartgui/process_tree.py            | 292 ++++++
>  .../pybootchartgui/pybootchartgui/samples.py  | 178 ++++
>  .../pybootchartgui/tests/parser_test.py       | 105 ++
>  .../pybootchartgui/tests/process_tree_test.py |  92 ++
>  19 files changed, 4115 insertions(+)
>  create mode 100644 meta/classes/buildstats.bbclass
>  create mode 100644 meta/lib/buildstats.py
>  create mode 100644 scripts/pybootchartgui/AUTHORS
>  create mode 100644 scripts/pybootchartgui/COPYING
>  create mode 100644 scripts/pybootchartgui/MAINTAINERS
>  create mode 100644 scripts/pybootchartgui/NEWS
>  create mode 100644 scripts/pybootchartgui/README.pybootchart
>  create mode 100755 scripts/pybootchartgui/pybootchartgui.py
>  create mode 100644 scripts/pybootchartgui/pybootchartgui/__init__.py
>  create mode 100644 scripts/pybootchartgui/pybootchartgui/batch.py
>  create mode 100644 scripts/pybootchartgui/pybootchartgui/draw.py
>  create mode 100644 scripts/pybootchartgui/pybootchartgui/gui.py
>  create mode 120000 scripts/pybootchartgui/pybootchartgui/main.py
>  create mode 100644 scripts/pybootchartgui/pybootchartgui/main.py.in
>  create mode 100644 scripts/pybootchartgui/pybootchartgui/parsing.py
>  create mode 100644
> scripts/pybootchartgui/pybootchartgui/process_tree.py create mode
> 100644 scripts/pybootchartgui/pybootchartgui/samples.py create mode
> 100644 scripts/pybootchartgui/pybootchartgui/tests/parser_test.py
> create mode 100644
> scripts/pybootchartgui/pybootchartgui/tests/process_tree_test.py
> 
> diff --git a/meta/classes/buildstats.bbclass
> b/meta/classes/buildstats.bbclass new file mode 100644
> index 0000000..0de6052
> --- /dev/null
> +++ b/meta/classes/buildstats.bbclass
> @@ -0,0 +1,295 @@
> +BUILDSTATS_BASE = "${TMPDIR}/buildstats/"
> +
> +################################################################################
> +# Build statistics gathering.
> +#
> +# The CPU and Time gathering/tracking functions and bbevent
> inspiration +# were written by Christopher Larson.
> +#
> +################################################################################
> +
> +def get_buildprocess_cputime(pid):
> +    with open("/proc/%d/stat" % pid, "r") as f:
> +        fields = f.readline().rstrip().split()
> +    # 13: utime, 14: stime, 15: cutime, 16: cstime
> +    return sum(int(field) for field in fields[13:16])
> +
> +def get_process_cputime(pid):
> +    import resource
> +    with open("/proc/%d/stat" % pid, "r") as f:
> +        fields = f.readline().rstrip().split()
> +    stats = { 
> +        'utime'  : fields[13],
> +        'stime'  : fields[14], 
> +        'cutime' : fields[15], 
> +        'cstime' : fields[16],  
> +    }
> +    iostats = {}
> +    if os.path.isfile("/proc/%d/io" % pid):
> +        with open("/proc/%d/io" % pid, "r") as f:
> +            while True:
> +                i = f.readline().strip()
> +                if not i:
> +                    break
> +                if not ":" in i:
> +                    # one more extra line is appended (empty or
> containing "0")
> +                    # most probably due to race condition in kernel
> while
> +                    # updating IO stats
> +                    break
> +                i = i.split(": ")
> +                iostats[i[0]] = i[1]
> +    resources = resource.getrusage(resource.RUSAGE_SELF)
> +    childres = resource.getrusage(resource.RUSAGE_CHILDREN)
> +    return stats, iostats, resources, childres
> +
> +def get_cputime():
> +    with open("/proc/stat", "r") as f:
> +        fields = f.readline().rstrip().split()[1:]
> +    return sum(int(field) for field in fields)
> +
> +def set_timedata(var, d, server_time):
> +    d.setVar(var, server_time)
> +
> +def get_timedata(var, d, end_time):
> +    oldtime = d.getVar(var, False)
> +    if oldtime is None:
> +        return
> +    return end_time - oldtime
> +
> +def set_buildtimedata(var, d):
> +    import time
> +    time = time.time()
> +    cputime = get_cputime()
> +    proctime = get_buildprocess_cputime(os.getpid())
> +    d.setVar(var, (time, cputime, proctime))
> +
> +def get_buildtimedata(var, d):
> +    import time
> +    timedata = d.getVar(var, False)
> +    if timedata is None:
> +        return
> +    oldtime, oldcpu, oldproc = timedata
> +    procdiff = get_buildprocess_cputime(os.getpid()) - oldproc
> +    cpudiff = get_cputime() - oldcpu
> +    end_time = time.time()
> +    timediff = end_time - oldtime
> +    if cpudiff > 0:
> +        cpuperc = float(procdiff) * 100 / cpudiff
> +    else:
> +        cpuperc = None
> +    return timediff, cpuperc
> +
> +def write_task_data(status, logfile, e, d):
> +    with open(os.path.join(logfile), "a") as f:
> +        elapsedtime = get_timedata("__timedata_task", d, e.time)
> +        if elapsedtime:
> +            f.write(d.expand("${PF}: %s\n" % e.task))
> +            f.write(d.expand("Elapsed time: %0.2f seconds\n" %
> elapsedtime))
> +            cpu, iostats, resources, childres =
> get_process_cputime(os.getpid())
> +            if cpu:
> +                f.write("utime: %s\n" % cpu['utime'])
> +                f.write("stime: %s\n" % cpu['stime'])
> +                f.write("cutime: %s\n" % cpu['cutime'])
> +                f.write("cstime: %s\n" % cpu['cstime'])
> +            for i in iostats:
> +                f.write("IO %s: %s\n" % (i, iostats[i]))
> +            rusages = ["ru_utime", "ru_stime", "ru_maxrss",
> "ru_minflt", "ru_majflt", "ru_inblock", "ru_oublock", "ru_nvcsw",
> "ru_nivcsw"]
> +            for i in rusages:
> +                f.write("rusage %s: %s\n" % (i, getattr(resources,
> i)))
> +            for i in rusages:
> +                f.write("Child rusage %s: %s\n" % (i,
> getattr(childres, i)))
> +        if status == "passed":
> +            f.write("Status: PASSED \n")
> +        else:
> +            f.write("Status: FAILED \n")
> +        f.write("Ended: %0.2f \n" % e.time)
> +
> +def write_host_data(logfile, e, d, type):
> +    import subprocess, os, datetime
> +    # minimum time allowed for each command to run, in seconds
> +    time_threshold = 0.5
> +    limit = 10
> +    # the total number of commands
> +    num_cmds = 0
> +    msg = ""
> +    if type == "interval":
> +        # interval at which data will be logged
> +        interval = d.getVar("BB_HEARTBEAT_EVENT", False)
> +        if interval is None:
> +            bb.warn("buildstats: Collecting host data at intervals
> failed. Set BB_HEARTBEAT_EVENT=\"<interval>\" in conf/local.conf for
> the interval at which host data will be logged.")
> +            d.setVar("BB_LOG_HOST_STAT_ON_INTERVAL", "0")
> +            return
> +        interval = int(interval)
> +        cmds = d.getVar('BB_LOG_HOST_STAT_CMDS_INTERVAL')
> +        msg = "Host Stats: Collecting data at %d second
> intervals.\n" % interval
> +        if cmds is None:
> +            d.setVar("BB_LOG_HOST_STAT_ON_INTERVAL", "0")
> +            bb.warn("buildstats: Collecting host data at intervals
> failed. Set BB_LOG_HOST_STAT_CMDS_INTERVAL=\"command1 ; command2 ;
> ... \" in conf/local.conf.")
> +            return
> +    if type == "failure":
> +        cmds = d.getVar('BB_LOG_HOST_STAT_CMDS_FAILURE')
> +        msg = "Host Stats: Collecting data on failure.\n"
> +        msg += "Failed at task: " + e.task + "\n"
> +        if cmds is None:
> +            d.setVar("BB_LOG_HOST_STAT_ON_FAILURE", "0")
> +            bb.warn("buildstats: Collecting host data on failure
> failed. Set BB_LOG_HOST_STAT_CMDS_FAILURE=\"command1 ; command2 ; ...
> \" in conf/local.conf.")
> +            return
> +    c_san = []
> +    for cmd in cmds.split(";"):
> +        if len(cmd) == 0:
> +            continue
> +        num_cmds += 1
> +        c_san.append(cmd)
> +    if num_cmds == 0:
> +        if type == "interval":
> +            d.setVar("BB_LOG_HOST_STAT_ON_INTERVAL", "0")
> +        if type == "failure":
> +            d.setVar("BB_LOG_HOST_STAT_ON_FAILURE", "0")
> +        return
> +
> +    # return if the interval is not enough to run all commands
> within the specified BB_HEARTBEAT_EVENT interval
> +    if type == "interval":
> +        limit = interval / num_cmds
> +        if limit <= time_threshold:
> +            d.setVar("BB_LOG_HOST_STAT_ON_INTERVAL", "0")
> +            bb.warn("buildstats: Collecting host data failed.
> BB_HEARTBEAT_EVENT interval not enough to run the specified commands.
> Increase value of BB_HEARTBEAT_EVENT in conf/local.conf.")
> +            return
> +
> +    # set the environment variables 
> +    path = d.getVar("PATH")
> +    opath = d.getVar("BB_ORIGENV", False).getVar("PATH")
> +    ospath = os.environ['PATH']
> +    os.environ['PATH'] = path + ":" + opath + ":" + ospath
> +    with open(logfile, "a") as f:
> +        f.write("Event Time: %f\nDate: %s\n" % (e.time,
> datetime.datetime.now()))
> +        f.write("%s" % msg)
> +        for c in c_san:
> +            try:
> +                output = subprocess.check_output(c.split(),
> stderr=subprocess.STDOUT, timeout=limit).decode('utf-8')
> +            except (subprocess.CalledProcessError,
> subprocess.TimeoutExpired, FileNotFoundError) as err:
> +                output = "Error running command: %s\n%s\n" % (c, err)
> +            f.write("%s\n%s\n" % (c, output))
> +    # reset the environment
> +    os.environ['PATH'] = ospath
> +
> +python run_buildstats () {
> +    import bb.build
> +    import bb.event
> +    import time, subprocess, platform
> +
> +    bn = d.getVar('BUILDNAME')
> +
> ########################################################################
> +    # bitbake fires HeartbeatEvent even before a build has been
> +    # triggered, causing BUILDNAME to be None
> +
> ########################################################################
> +    if bn is not None:
> +        bsdir = os.path.join(d.getVar('BUILDSTATS_BASE'), bn)
> +        taskdir = os.path.join(bsdir, d.getVar('PF'))
> +        if isinstance(e, bb.event.HeartbeatEvent) and
> bb.utils.to_boolean(d.getVar("BB_LOG_HOST_STAT_ON_INTERVAL")):
> +            bb.utils.mkdirhier(bsdir)
> +            write_host_data(os.path.join(bsdir,
> "host_stats_interval"), e, d, "interval") +
> +    if isinstance(e, bb.event.BuildStarted):
> +
> ########################################################################
> +        # If the kernel was not configured to provide I/O
> statistics, issue
> +        # a one time warning.
> +
> ########################################################################
> +        if not os.path.isfile("/proc/%d/io" % os.getpid()):
> +            bb.warn("The Linux kernel on your build host was not
> configured to provide process I/O statistics.
> (CONFIG_TASK_IO_ACCOUNTING is not set)") +
> +
> ########################################################################
> +        # at first pass make the buildstats hierarchy and then
> +        # set the buildname
> +
> ########################################################################
> +        bb.utils.mkdirhier(bsdir)
> +        set_buildtimedata("__timedata_build", d)
> +        build_time = os.path.join(bsdir, "build_stats")
> +        # write start of build into build_time
> +        with open(build_time, "a") as f:
> +            host_info = platform.uname()
> +            f.write("Host Info: ")
> +            for x in host_info:
> +                if x:
> +                    f.write(x + " ")
> +            f.write("\n")
> +            f.write("Build Started: %0.2f \n" %
> d.getVar('__timedata_build', False)[0]) +
> +    elif isinstance(e, bb.event.BuildCompleted):
> +        build_time = os.path.join(bsdir, "build_stats")
> +        with open(build_time, "a") as f:
> +
> ########################################################################
> +            # Write build statistics for the build
> +
> ########################################################################
> +            timedata = get_buildtimedata("__timedata_build", d)
> +            if timedata:
> +                time, cpu = timedata
> +                # write end of build and cpu used into build_time
> +                f.write("Elapsed time: %0.2f seconds \n" % (time))
> +                if cpu:
> +                    f.write("CPU usage: %0.1f%% \n" % cpu)
> +
> +    if isinstance(e, bb.build.TaskStarted):
> +        set_timedata("__timedata_task", d, e.time)
> +        bb.utils.mkdirhier(taskdir)
> +        # write into the task event file the name and start time
> +        with open(os.path.join(taskdir, e.task), "a") as f:
> +            f.write("Event: %s \n" % bb.event.getName(e))
> +            f.write("Started: %0.2f \n" % e.time)
> +
> +    elif isinstance(e, bb.build.TaskSucceeded):
> +        write_task_data("passed", os.path.join(taskdir, e.task), e,
> d)
> +        if e.task == "do_rootfs":
> +            bs = os.path.join(bsdir, "build_stats")
> +            with open(bs, "a") as f:
> +                rootfs = d.getVar('IMAGE_ROOTFS')
> +                if os.path.isdir(rootfs):
> +                    try:
> +                        rootfs_size = subprocess.check_output(["du",
> "-sh", rootfs],
> +
> stderr=subprocess.STDOUT).decode('utf-8')
> +                        f.write("Uncompressed Rootfs size: %s" %
> rootfs_size)
> +                    except subprocess.CalledProcessError as err:
> +                        bb.warn("Failed to get rootfs size: %s" %
> err.output.decode('utf-8')) +
> +    elif isinstance(e, bb.build.TaskFailed):
> +        # Can have a failure before TaskStarted so need to mkdir
> here too
> +        bb.utils.mkdirhier(taskdir)
> +        write_task_data("failed", os.path.join(taskdir, e.task), e,
> d)
> +
> ########################################################################
> +        # Lets make things easier and tell people where the build
> failed in
> +        # build_status. We do this here because BuildCompleted
> triggers no
> +        # matter what the status of the build actually is
> +
> ########################################################################
> +        build_status = os.path.join(bsdir, "build_stats")
> +        with open(build_status, "a") as f:
> +            f.write(d.expand("Failed at: ${PF} at task: %s \n" %
> e.task))
> +        if
> bb.utils.to_boolean(d.getVar("BB_LOG_HOST_STAT_ON_FAILURE")):
> +            write_host_data(os.path.join(bsdir,
> "host_stats_%s_failure" % e.task), e, d, "failure") +}
> +
> +addhandler run_buildstats
> +run_buildstats[eventmask] = "bb.event.BuildStarted
> bb.event.BuildCompleted bb.event.HeartbeatEvent bb.build.TaskStarted
> bb.build.TaskSucceeded bb.build.TaskFailed" + +python runqueue_stats
> () {
> +    import buildstats
> +    from bb import event, runqueue
> +    # We should not record any samples before the first task has
> started,
> +    # because that's the first activity shown in the process chart.
> +    # Besides, at that point we are sure that the build variables
> +    # are available that we need to find the output directory.
> +    # The persistent SystemStats is stored in the datastore and
> +    # closed when the build is done.
> +    system_stats = d.getVar('_buildstats_system_stats', False)
> +    if not system_stats and isinstance(e,
> (bb.runqueue.sceneQueueTaskStarted, bb.runqueue.runQueueTaskStarted)):
> +        system_stats = buildstats.SystemStats(d)
> +        d.setVar('_buildstats_system_stats', system_stats)
> +    if system_stats:
> +        # Ensure that we sample at important events.
> +        done = isinstance(e, bb.event.BuildCompleted)
> +        system_stats.sample(e, force=done)
> +        if done:
> +            system_stats.close()
> +            d.delVar('_buildstats_system_stats')
> +}
> +
> +addhandler runqueue_stats
> +runqueue_stats[eventmask] = "bb.runqueue.sceneQueueTaskStarted
> bb.runqueue.runQueueTaskStarted bb.event.HeartbeatEvent
> bb.event.BuildCompleted bb.event.MonitorDiskEvent" diff --git
> a/meta/lib/buildstats.py b/meta/lib/buildstats.py new file mode
> 100644 index 0000000..8627ed3 --- /dev/null
> +++ b/meta/lib/buildstats.py
> @@ -0,0 +1,161 @@
> +#
> +# SPDX-License-Identifier: GPL-2.0-only
> +#
> +# Implements system state sampling. Called by buildstats.bbclass.
> +# Because it is a real Python module, it can hold persistent state,
> +# like open log files and the time of the last sampling.
> +
> +import time
> +import re
> +import bb.event
> +
> +class SystemStats:
> +    def __init__(self, d):
> +        bn = d.getVar('BUILDNAME')
> +        bsdir = os.path.join(d.getVar('BUILDSTATS_BASE'), bn)
> +        bb.utils.mkdirhier(bsdir)
> +
> +        self.proc_files = []
> +        for filename, handler in (
> +                ('diskstats', self._reduce_diskstats),
> +                ('meminfo', self._reduce_meminfo),
> +                ('stat', self._reduce_stat),
> +        ):
> +            # The corresponding /proc files might not exist on the
> host.
> +            # For example, /proc/diskstats is not available in
> virtualized
> +            # environments like Linux-VServer. Silently skip
> collecting
> +            # the data.
> +            if os.path.exists(os.path.join('/proc', filename)):
> +                # In practice, this class gets instantiated only
> once in
> +                # the bitbake cooker process.  Therefore 'append'
> mode is
> +                # not strictly necessary, but using it makes the
> class
> +                # more robust should two processes ever write
> +                # concurrently.
> +                destfile = os.path.join(bsdir, '%sproc_%s.log' %
> ('reduced_' if handler else '', filename))
> +                self.proc_files.append((filename, open(destfile,
> 'ab'), handler))
> +        self.monitor_disk = open(os.path.join(bsdir,
> 'monitor_disk.log'), 'ab')
> +        # Last time that we sampled /proc data resp. recorded disk
> monitoring data.
> +        self.last_proc = 0
> +        self.last_disk_monitor = 0
> +        # Minimum number of seconds between recording a sample. This
> +        # becames relevant when we get called very often while many
> +        # short tasks get started. Sampling during quiet periods
> +        # depends on the heartbeat event, which fires less often.
> +        self.min_seconds = 1
> +
> +        self.meminfo_regex =
> re.compile(b'^(MemTotal|MemFree|Buffers|Cached|SwapTotal|SwapFree):\s*(\d+)')
> +        self.diskstats_regex =
> re.compile(b'^([hsv]d.|mtdblock\d|mmcblk\d|cciss/c\d+d\d+.*)$')
> +        self.diskstats_ltime = None
> +        self.diskstats_data = None
> +        self.stat_ltimes = None
> +
> +    def close(self):
> +        self.monitor_disk.close()
> +        for _, output, _ in self.proc_files:
> +            output.close()
> +
> +    def _reduce_meminfo(self, time, data):
> +        """
> +        Extracts 'MemTotal', 'MemFree', 'Buffers', 'Cached',
> 'SwapTotal', 'SwapFree'
> +        and writes their values into a single line, in that order.
> +        """
> +        values = {}
> +        for line in data.split(b'\n'):
> +            m = self.meminfo_regex.match(line)
> +            if m:
> +                values[m.group(1)] = m.group(2)
> +        if len(values) == 6:
> +            return (time,
> +                    b' '.join([values[x] for x in
> +                               (b'MemTotal', b'MemFree', b'Buffers',
> b'Cached', b'SwapTotal', b'SwapFree')]) + b'\n') +
> +    def _diskstats_is_relevant_line(self, linetokens):
> +        if len(linetokens) != 14:
> +            return False
> +        disk = linetokens[2]
> +        return self.diskstats_regex.match(disk)
> +
> +    def _reduce_diskstats(self, time, data):
> +        relevant_tokens = filter(self._diskstats_is_relevant_line,
> map(lambda x: x.split(), data.split(b'\n')))
> +        diskdata = [0] * 3
> +        reduced = None
> +        for tokens in relevant_tokens:
> +            # rsect
> +            diskdata[0] += int(tokens[5])
> +            # wsect
> +            diskdata[1] += int(tokens[9])
> +            # use
> +            diskdata[2] += int(tokens[12])
> +        if self.diskstats_ltime:
> +            # We need to compute information about the time interval
> +            # since the last sampling and record the result as sample
> +            # for that point in the past.
> +            interval = time - self.diskstats_ltime
> +            if interval > 0:
> +                sums = [ a - b for a, b in zip(diskdata,
> self.diskstats_data) ]
> +                readTput = sums[0] / 2.0 * 100.0 / interval
> +                writeTput = sums[1] / 2.0 * 100.0 / interval
> +                util = float( sums[2] ) / 10 / interval
> +                util = max(0.0, min(1.0, util))
> +                reduced = (self.diskstats_ltime, (readTput,
> writeTput, util)) +
> +        self.diskstats_ltime = time
> +        self.diskstats_data = diskdata
> +        return reduced
> +
> +
> +    def _reduce_nop(self, time, data):
> +        return (time, data)
> +
> +    def _reduce_stat(self, time, data):
> +        if not data:
> +            return None
> +        # CPU times {user, nice, system, idle, io_wait, irq,
> softirq} from first line
> +        tokens = data.split(b'\n', 1)[0].split()
> +        times = [ int(token) for token in tokens[1:] ]
> +        reduced = None
> +        if self.stat_ltimes:
> +            user = float((times[0] + times[1]) -
> (self.stat_ltimes[0] + self.stat_ltimes[1]))
> +            system = float((times[2] + times[5] + times[6]) -
> (self.stat_ltimes[2] + self.stat_ltimes[5] + self.stat_ltimes[6]))
> +            idle = float(times[3] - self.stat_ltimes[3])
> +            iowait = float(times[4] - self.stat_ltimes[4])
> +
> +            aSum = max(user + system + idle + iowait, 1)
> +            reduced = (time, (user/aSum, system/aSum, iowait/aSum))
> +
> +        self.stat_ltimes = times
> +        return reduced
> +
> +    def sample(self, event, force):
> +        now = time.time()
> +        if (now - self.last_proc > self.min_seconds) or force:
> +            for filename, output, handler in self.proc_files:
> +                with open(os.path.join('/proc', filename), 'rb') as
> input:
> +                    data = input.read()
> +                    if handler:
> +                        reduced = handler(now, data)
> +                    else:
> +                        reduced = (now, data)
> +                    if reduced:
> +                        if isinstance(reduced[1], bytes):
> +                            # Use as it is.
> +                            data = reduced[1]
> +                        else:
> +                            # Convert to a single line.
> +                            data = (' '.join([str(x) for x in
> reduced[1]]) + '\n').encode('ascii')
> +                        # Unbuffered raw write, less overhead and
> useful
> +                        # in case that we end up with concurrent
> writes.
> +                        os.write(output.fileno(),
> +                                 ('%.0f\n' %
> reduced[0]).encode('ascii') +
> +                                 data +
> +                                 b'\n')
> +            self.last_proc = now
> +
> +        if isinstance(event, bb.event.MonitorDiskEvent) and \
> +           ((now - self.last_disk_monitor > self.min_seconds) or
> force):
> +            os.write(self.monitor_disk.fileno(),
> +                     ('%.0f\n' % now).encode('ascii') +
> +                     ''.join(['%s: %d\n' % (dev, sample.total_bytes
> - sample.free_bytes)
> +                              for dev, sample in
> event.disk_usage.items()]).encode('ascii') +
> +                     b'\n')
> +            self.last_disk_monitor = now
> diff --git a/scripts/pybootchartgui/AUTHORS
> b/scripts/pybootchartgui/AUTHORS new file mode 100644
> index 0000000..672b7e9
> --- /dev/null
> +++ b/scripts/pybootchartgui/AUTHORS
> @@ -0,0 +1,11 @@
> +Michael Meeks <michael.meeks@novell.com>
> +Anders Norgaard <anders.norgaard@gmail.com>
> +Scott James Remnant <scott@ubuntu.com>
> +Henning Niss <henningniss@gmail.com>
> +Riccardo Magliocchetti <riccardo.magliocchetti@gmail.com>
> +
> +Contributors:
> +	Brian Ewins
> +
> +Based on work by:
> +	Ziga Mahkovec
> diff --git a/scripts/pybootchartgui/COPYING
> b/scripts/pybootchartgui/COPYING new file mode 100644
> index 0000000..ed87acf
> --- /dev/null
> +++ b/scripts/pybootchartgui/COPYING
> @@ -0,0 +1,340 @@
> +		    GNU GENERAL PUBLIC LICENSE
> +		       Version 2, June 1991
> +
> + Copyright (C) 1989, 1991 Free Software Foundation, Inc.
> +    51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA 
> + Everyone is permitted to copy and distribute verbatim copies
> + of this license document, but changing it is not allowed.
> +
> +			    Preamble
> +
> +  The licenses for most software are designed to take away your
> +freedom to share and change it.  By contrast, the GNU General Public
> +License is intended to guarantee your freedom to share and change
> free +software--to make sure the software is free for all its users.
> This +General Public License applies to most of the Free Software
> +Foundation's software and to any other program whose authors commit
> to +using it.  (Some other Free Software Foundation software is
> covered by +the GNU Library General Public License instead.)  You can
> apply it to +your programs, too.
> +
> +  When we speak of free software, we are referring to freedom, not
> +price.  Our General Public Licenses are designed to make sure that
> you +have the freedom to distribute copies of free software (and
> charge for +this service if you wish), that you receive source code
> or can get it +if you want it, that you can change the software or
> use pieces of it +in new free programs; and that you know you can do
> these things. +
> +  To protect your rights, we need to make restrictions that forbid
> +anyone to deny you these rights or to ask you to surrender the
> rights. +These restrictions translate to certain responsibilities for
> you if you +distribute copies of the software, or if you modify it.
> +
> +  For example, if you distribute copies of such a program, whether
> +gratis or for a fee, you must give the recipients all the rights that
> +you have.  You must make sure that they, too, receive or can get the
> +source code.  And you must show them these terms so they know their
> +rights.
> +
> +  We protect your rights with two steps: (1) copyright the software,
> and +(2) offer you this license which gives you legal permission to
> copy, +distribute and/or modify the software.
> +
> +  Also, for each author's protection and ours, we want to make
> certain +that everyone understands that there is no warranty for this
> free +software.  If the software is modified by someone else and
> passed on, we +want its recipients to know that what they have is not
> the original, so +that any problems introduced by others will not
> reflect on the original +authors' reputations.
> +
> +  Finally, any free program is threatened constantly by software
> +patents.  We wish to avoid the danger that redistributors of a free
> +program will individually obtain patent licenses, in effect making
> the +program proprietary.  To prevent this, we have made it clear
> that any +patent must be licensed for everyone's free use or not
> licensed at all. +
> +  The precise terms and conditions for copying, distribution and
> +modification follow.
> +\f
> +		    GNU GENERAL PUBLIC LICENSE
> +   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
> +
> +  0. This License applies to any program or other work which contains
> +a notice placed by the copyright holder saying it may be distributed
> +under the terms of this General Public License.  The "Program",
> below, +refers to any such program or work, and a "work based on the
> Program" +means either the Program or any derivative work under
> copyright law: +that is to say, a work containing the Program or a
> portion of it, +either verbatim or with modifications and/or
> translated into another +language.  (Hereinafter, translation is
> included without limitation in +the term "modification".)  Each
> licensee is addressed as "you". +
> +Activities other than copying, distribution and modification are not
> +covered by this License; they are outside its scope.  The act of
> +running the Program is not restricted, and the output from the
> Program +is covered only if its contents constitute a work based on
> the +Program (independent of having been made by running the Program).
> +Whether that is true depends on what the Program does.
> +
> +  1. You may copy and distribute verbatim copies of the Program's
> +source code as you receive it, in any medium, provided that you
> +conspicuously and appropriately publish on each copy an appropriate
> +copyright notice and disclaimer of warranty; keep intact all the
> +notices that refer to this License and to the absence of any
> warranty; +and give any other recipients of the Program a copy of
> this License +along with the Program.
> +
> +You may charge a fee for the physical act of transferring a copy, and
> +you may at your option offer warranty protection in exchange for a
> fee. +
> +  2. You may modify your copy or copies of the Program or any portion
> +of it, thus forming a work based on the Program, and copy and
> +distribute such modifications or work under the terms of Section 1
> +above, provided that you also meet all of these conditions:
> +
> +    a) You must cause the modified files to carry prominent notices
> +    stating that you changed the files and the date of any change.
> +
> +    b) You must cause any work that you distribute or publish, that
> in
> +    whole or in part contains or is derived from the Program or any
> +    part thereof, to be licensed as a whole at no charge to all third
> +    parties under the terms of this License.
> +
> +    c) If the modified program normally reads commands interactively
> +    when run, you must cause it, when started running for such
> +    interactive use in the most ordinary way, to print or display an
> +    announcement including an appropriate copyright notice and a
> +    notice that there is no warranty (or else, saying that you
> provide
> +    a warranty) and that users may redistribute the program under
> +    these conditions, and telling the user how to view a copy of this
> +    License.  (Exception: if the Program itself is interactive but
> +    does not normally print such an announcement, your work based on
> +    the Program is not required to print an announcement.)
> +\f
> +These requirements apply to the modified work as a whole.  If
> +identifiable sections of that work are not derived from the Program,
> +and can be reasonably considered independent and separate works in
> +themselves, then this License, and its terms, do not apply to those
> +sections when you distribute them as separate works.  But when you
> +distribute the same sections as part of a whole which is a work based
> +on the Program, the distribution of the whole must be on the terms of
> +this License, whose permissions for other licensees extend to the
> +entire whole, and thus to each and every part regardless of who
> wrote it. +
> +Thus, it is not the intent of this section to claim rights or contest
> +your rights to work written entirely by you; rather, the intent is to
> +exercise the right to control the distribution of derivative or
> +collective works based on the Program.
> +
> +In addition, mere aggregation of another work not based on the
> Program +with the Program (or with a work based on the Program) on a
> volume of +a storage or distribution medium does not bring the other
> work under +the scope of this License.
> +
> +  3. You may copy and distribute the Program (or a work based on it,
> +under Section 2) in object code or executable form under the terms of
> +Sections 1 and 2 above provided that you also do one of the
> following: +
> +    a) Accompany it with the complete corresponding machine-readable
> +    source code, which must be distributed under the terms of
> Sections
> +    1 and 2 above on a medium customarily used for software
> interchange; or, +
> +    b) Accompany it with a written offer, valid for at least three
> +    years, to give any third party, for a charge no more than your
> +    cost of physically performing source distribution, a complete
> +    machine-readable copy of the corresponding source code, to be
> +    distributed under the terms of Sections 1 and 2 above on a medium
> +    customarily used for software interchange; or,
> +
> +    c) Accompany it with the information you received as to the offer
> +    to distribute corresponding source code.  (This alternative is
> +    allowed only for noncommercial distribution and only if you
> +    received the program in object code or executable form with such
> +    an offer, in accord with Subsection b above.)
> +
> +The source code for a work means the preferred form of the work for
> +making modifications to it.  For an executable work, complete source
> +code means all the source code for all modules it contains, plus any
> +associated interface definition files, plus the scripts used to
> +control compilation and installation of the executable.  However, as
> a +special exception, the source code distributed need not include
> +anything that is normally distributed (in either source or binary
> +form) with the major components (compiler, kernel, and so on) of the
> +operating system on which the executable runs, unless that component
> +itself accompanies the executable.
> +
> +If distribution of executable or object code is made by offering
> +access to copy from a designated place, then offering equivalent
> +access to copy the source code from the same place counts as
> +distribution of the source code, even though third parties are not
> +compelled to copy the source along with the object code.
> +\f
> +  4. You may not copy, modify, sublicense, or distribute the Program
> +except as expressly provided under this License.  Any attempt
> +otherwise to copy, modify, sublicense or distribute the Program is
> +void, and will automatically terminate your rights under this
> License. +However, parties who have received copies, or rights, from
> you under +this License will not have their licenses terminated so
> long as such +parties remain in full compliance.
> +
> +  5. You are not required to accept this License, since you have not
> +signed it.  However, nothing else grants you permission to modify or
> +distribute the Program or its derivative works.  These actions are
> +prohibited by law if you do not accept this License.  Therefore, by
> +modifying or distributing the Program (or any work based on the
> +Program), you indicate your acceptance of this License to do so, and
> +all its terms and conditions for copying, distributing or modifying
> +the Program or works based on it.
> +
> +  6. Each time you redistribute the Program (or any work based on the
> +Program), the recipient automatically receives a license from the
> +original licensor to copy, distribute or modify the Program subject
> to +these terms and conditions.  You may not impose any further
> +restrictions on the recipients' exercise of the rights granted
> herein. +You are not responsible for enforcing compliance by third
> parties to +this License.
> +
> +  7. If, as a consequence of a court judgment or allegation of patent
> +infringement or for any other reason (not limited to patent issues),
> +conditions are imposed on you (whether by court order, agreement or
> +otherwise) that contradict the conditions of this License, they do
> not +excuse you from the conditions of this License.  If you cannot
> +distribute so as to satisfy simultaneously your obligations under
> this +License and any other pertinent obligations, then as a
> consequence you +may not distribute the Program at all.  For example,
> if a patent +license would not permit royalty-free redistribution of
> the Program by +all those who receive copies directly or indirectly
> through you, then +the only way you could satisfy both it and this
> License would be to +refrain entirely from distribution of the
> Program. +
> +If any portion of this section is held invalid or unenforceable under
> +any particular circumstance, the balance of the section is intended
> to +apply and the section as a whole is intended to apply in other
> +circumstances.
> +
> +It is not the purpose of this section to induce you to infringe any
> +patents or other property right claims or to contest validity of any
> +such claims; this section has the sole purpose of protecting the
> +integrity of the free software distribution system, which is
> +implemented by public license practices.  Many people have made
> +generous contributions to the wide range of software distributed
> +through that system in reliance on consistent application of that
> +system; it is up to the author/donor to decide if he or she is
> willing +to distribute software through any other system and a
> licensee cannot +impose that choice.
> +
> +This section is intended to make thoroughly clear what is believed to
> +be a consequence of the rest of this License.
> +\f
> +  8. If the distribution and/or use of the Program is restricted in
> +certain countries either by patents or by copyrighted interfaces, the
> +original copyright holder who places the Program under this License
> +may add an explicit geographical distribution limitation excluding
> +those countries, so that distribution is permitted only in or among
> +countries not thus excluded.  In such case, this License incorporates
> +the limitation as if written in the body of this License.
> +
> +  9. The Free Software Foundation may publish revised and/or new
> versions +of the General Public License from time to time.  Such new
> versions will +be similar in spirit to the present version, but may
> differ in detail to +address new problems or concerns.
> +
> +Each version is given a distinguishing version number.  If the
> Program +specifies a version number of this License which applies to
> it and "any +later version", you have the option of following the
> terms and conditions +either of that version or of any later version
> published by the Free +Software Foundation.  If the Program does not
> specify a version number of +this License, you may choose any version
> ever published by the Free Software +Foundation.
> +
> +  10. If you wish to incorporate parts of the Program into other free
> +programs whose distribution conditions are different, write to the
> author +to ask for permission.  For software which is copyrighted by
> the Free +Software Foundation, write to the Free Software Foundation;
> we sometimes +make exceptions for this.  Our decision will be guided
> by the two goals +of preserving the free status of all derivatives of
> our free software and +of promoting the sharing and reuse of software
> generally. +
> +			    NO WARRANTY
> +
> +  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO
> WARRANTY +FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.
>  EXCEPT WHEN +OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS
> AND/OR OTHER PARTIES +PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF
> ANY KIND, EITHER EXPRESSED +OR IMPLIED, INCLUDING, BUT NOT LIMITED
> TO, THE IMPLIED WARRANTIES OF +MERCHANTABILITY AND FITNESS FOR A
> PARTICULAR PURPOSE.  THE ENTIRE RISK AS +TO THE QUALITY AND
> PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE +PROGRAM PROVE
> DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, +REPAIR OR
> CORRECTION. +
> +  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN
> WRITING +WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY
> AND/OR +REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU
> FOR DAMAGES, +INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR
> CONSEQUENTIAL DAMAGES ARISING +OUT OF THE USE OR INABILITY TO USE THE
> PROGRAM (INCLUDING BUT NOT LIMITED +TO LOSS OF DATA OR DATA BEING
> RENDERED INACCURATE OR LOSSES SUSTAINED BY +YOU OR THIRD PARTIES OR A
> FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER +PROGRAMS), EVEN IF
> SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE +POSSIBILITY OF
> SUCH DAMAGES. +
> +		     END OF TERMS AND CONDITIONS
> +\f
> +	    How to Apply These Terms to Your New Programs
> +
> +  If you develop a new program, and you want it to be of the greatest
> +possible use to the public, the best way to achieve this is to make
> it +free software which everyone can redistribute and change under
> these terms. +
> +  To do so, attach the following notices to the program.  It is
> safest +to attach them to the start of each source file to most
> effectively +convey the exclusion of warranty; and each file should
> have at least +the "copyright" line and a pointer to where the full
> notice is found. +
> +    <one line to give the program's name and a brief idea of what it
> does.>
> +    Copyright (C) <year>  <name of author>
> +
> +    This program is free software; you can redistribute it and/or
> modify
> +    it under the terms of the GNU General Public License as
> published by
> +    the Free Software Foundation; either version 2 of the License, or
> +    (at your option) any later version.
> +
> +    This program is distributed in the hope that it will be useful,
> +    but WITHOUT ANY WARRANTY; without even the implied warranty of
> +    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +    GNU General Public License for more details.
> +
> +    You should have received a copy of the GNU General Public License
> +    along with this program; if not, write to the Free Software
> +    Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
> 02110-1301 USA +
> +
> +Also add information on how to contact you by electronic and paper
> mail. +
> +If the program is interactive, make it output a short notice like
> this +when it starts in an interactive mode:
> +
> +    Gnomovision version 69, Copyright (C) year  name of author
> +    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type
> `show w'.
> +    This is free software, and you are welcome to redistribute it
> +    under certain conditions; type `show c' for details.
> +
> +The hypothetical commands `show w' and `show c' should show the
> appropriate +parts of the General Public License.  Of course, the
> commands you use may +be called something other than `show w' and
> `show c'; they could even be +mouse-clicks or menu items--whatever
> suits your program. +
> +You should also get your employer (if you work as a programmer) or
> your +school, if any, to sign a "copyright disclaimer" for the
> program, if +necessary.  Here is a sample; alter the names:
> +
> +  Yoyodyne, Inc., hereby disclaims all copyright interest in the
> program
> +  `Gnomovision' (which makes passes at compilers) written by James
> Hacker. +
> +  <signature of Ty Coon>, 1 April 1989
> +  Ty Coon, President of Vice
> +
> +This General Public License does not permit incorporating your
> program into +proprietary programs.  If your program is a subroutine
> library, you may +consider it more useful to permit linking
> proprietary applications with the +library.  If this is what you want
> to do, use the GNU Library General +Public License instead of this
> License. diff --git a/scripts/pybootchartgui/MAINTAINERS
> b/scripts/pybootchartgui/MAINTAINERS new file mode 100644
> index 0000000..c65e131
> --- /dev/null
> +++ b/scripts/pybootchartgui/MAINTAINERS
> @@ -0,0 +1,3 @@
> +Riccardo Magliocchetti <riccardo.magliocchetti@gmail.com>
> +Michael Meeks <michael.meeks@novell.com>
> +Harald Hoyer <harald@redhat.com>
> diff --git a/scripts/pybootchartgui/NEWS b/scripts/pybootchartgui/NEWS
> new file mode 100644
> index 0000000..7c5b2fc
> --- /dev/null
> +++ b/scripts/pybootchartgui/NEWS
> @@ -0,0 +1,204 @@
> +bootchart2 0.14.5:
> +	+ pybootchartgui (Riccardo)
> +		+ Fix tests with python3
> +		+ Fix parsing of files with non-ascii bytes
> +		+ Robustness fixes to taskstats and meminfo parsing
> +		+ More python3 fixes
> +
> +bootchart2 0.14.4:
> +	+ bootchartd
> +		+ Add relevant EXIT_PROC for GNOME3, XFCE4, openbox
> +		  (Justin Lecher, Ben Eills)
> +	+ pybootchartgui (Riccardo)
> +		+ Fix some issues in --crop-after and --annotate
> +		+ Fix pybootchartgui process_tree tests
> +		+ More python3 fixes
> +
> +bootchart2 0.14.2:
> +	+ pybootchartgui
> +		+ Fix some crashes in parsing.py (Jakub Czaplicki,
> Riccardo)
> +		+ speedup a bit meminfo parsing (Riccardo)
> +		+ Fix indentation for python3.2 (Riccardo)
> +
> +bootchart2 0.14.1:
> +	+ bootchartd
> +		+ Expect dmesg only if started as init (Henry Yei)
> +		+ look for bootchart_init in the environment (Henry
> Gebhardt)
> +	+ pybootchartgui
> +		+ Fixup some tests (Riccardo)
> +		+ Support hp smart arrays block devices (Anders
> Norgaard,
> +		  Brian Murray)
> +		+ Fixes for -t, -o and -f options (Mladen Kuntner,
> Harald, Riccardo) +
> +bootchart2 0.14.0:
> +	+ bootchartd
> +		+ Add ability to define custom commands
> +		  (Lucian Muresan, Peter Hjalmarsson)
> +	+ collector
> +		+ fix tmpfs mount leakage (Peter Hjalmarsson)
> +	+ pybootchartgui
> +		+ render cumulative I/O time chart (Sankar P)
> +		+ python3 compatibility fixes (Riccardo)
> +	+ Misc (Michael)
> +		+ remove confusing, obsolete setup.py
> +		+ install docs to /usr/share/
> +		+ lot of fixes for easier packaging (Peter
> Hjalmarsson)
> +		+ add bootchart2, bootchartd and pybootchartgui
> manpages
> +		  (Francesca Ciceri, David Paleino)
> +
> +bootchart2 0.12.6:
> +	+ bootchartd
> +		+ better check for initrd (Riccardo Magliocchetti)
> +		+ code cleanup (Riccardo)
> +		+ make the list of processes we are waiting for
> editable
> +		  in config file by EXIT_PROC (Riccardo)
> +		+ fix parsing of cmdline for alternative init system
> (Riccardo)
> +		+ fixed calling init in initramfs (Harald)
> +		+ exit 0 for start, if the collector is already
> running (Harald)
> +	+ collector
> +		+ try harder with taskstats (Michael)
> +		+ plug some small leaks (Riccardo)
> +		+ fix missing PROC_EVENTS detection (Harald)
> +	+ pybootchartgui (Michael)
> +		+ add kernel bootchart tab to interactive gui
> +		+ report bootchart version in cli interface
> +		+ improve rendering performance
> +		+ GUI improvements
> +		+ lot of cleanups
> +	+ Makefile
> +		+ do not python compile if NO_PYTHON_COMPILE is set
> (Harald)
> +	+ systemd service files
> +		+ added them and install (Harald, Wulf C. Krueger)
> +
> +bootchart2 0.12.5:
> +	+ administrative snafu version; pull before pushing...
> +
> +bootchart2 0.12.4:
> +	+ bootchartd
> +		+ reduce overhead caused by pidof (Riccardo
> Magliocchetti)
> +	+ collector
> +		+ attempt to retry ptrace to avoid bogus ENOSYS
> (Michael)
> +		+ add meminfo polling (Dave Martin)
> +	+ pybootchartgui
> +		+ handle dmesg timestamps with big delta (Riccardo)
> +		+ avoid divide by zero when rendering I/O
> utilization (Riccardo)
> +		+ add process grouping in the cumulative chart
> (Riccardo)
> +		+ fix cpu time calculation in cumulative chart
> (Riccardo)
> +		+ get i/o statistics for flash based devices
> (Riccardo)
> +		+ prettier coloring for the cumulative graphs
> (Michael)
> +		+ fix interactive CPU rendering (Michael)
> +		+ render memory usage graph (Dave Martin)
> +
> +bootchart2 0.12.3
> +	+ collector
> +		+ pclose after popen (Riccardo Magliocchetti (xrmx))
> +		+ fix buffer overflow (xrmx)
> +		+ count 'processor:' in /proc/cpuinfo for ARM
> (Michael)
> +		+ get model name from that line too for ARM (xrmx)
> +		+ store /proc/cpuinfo in the boot-chart archive
> (xrmx)
> +		+ try harder to detect missing TASKSTATS (Michael)
> +		+ sanity-check invalid domain names (Michael)
> +		+ detect missing PROC_EVENTS more reliably (Michael)
> +		+ README fixes (xrmx, Michael)
> +	+ pybootchartgui
> +		+ make num_cpu parsing robust (Michael)
> +
> +bootchart2 0.12.2
> + 	+ fix pthread compile / linking bug
> +
> +bootchart2 0.12.1
> +	+ pybootchartgui
> +		+ pylint cleanup
> +		+ handle empty traces more elegantly
> +		+ add '-t' / '--boot-time' argument (Matthew Bauer)
> +	+ collector
> +		+ now GPLv2
> +		+ add rdinit support for very early initrd tracing
> +		+ cleanup / re-factor code into separate modules
> +		+ re-factor arg parsing, and parse remote process
> args
> +		+ handle missing bootchartd.conf cleanly
> +		+ move much of bootchartd from shell -> C
> +			+ drop dmesg and uname usage
> +			+ avoid rpm/dpkg with native version
> reporting +
> +bootchart2 0.12.0 (Michael Meeks)
> +	+ collector
> +		+ use netlink PROC_EVENTS to generate parentage data
> +			+ finally kills any need for 'acct' et. al.
> +			+ also removes need to poll /proc => faster
> +		+ cleanup code to K&R, 8 stop tabs.
> +	+ pybootchartgui
> +		+ consume thread parentage data
> +
> +bootchart2 0.11.4 (Michael Meeks)
> +	+ collector
> +		+ if run inside an initrd detect when /dev is
> writable
> +		  and remount ourselves into that.
> +		+ overflow buffers more elegantly in extremis
> +		+ dump full process path and command-line args
> +		+ calm down debugging output
> +	+ pybootchartgui
> +		+ can render logs in a directory again
> +		+ has a 'show more' option to show command-lines
> +
> +bootchart2 0.11.3 (Michael Meeks)
> +	+ add $$ display to the bootchart header
> +	+ process command-line bits
> +		+ fix collection code, and rename stream to match
> +		+ enable parsing, add check button to UI, and
> --show-all
> +		  command-line option
> +	+ fix parsing of directories full of files.
> +
> +bootchart2 0.11.2 (Michael Meeks)
> +	+ fix initrd sanity check to use the right proc path
> +	+ don't return a bogus error value when dumping state
> +	+ add -c to aid manual console debugging
> +
> +bootchart2 0.11.1 (Michael Meeks)
> +	+ even simpler initrd setup
> +		+ create a single directory: /lib/bootchart/tmpfs
> +
> +bootchart2 0.11 (Michael Meeks)
> +	+ bootchartd
> +		+ far, far simpler, less shell, more robustness etc.
> +	+ bootchart-collector
> +		+ remove the -p argument - we always mount proc
> +		+ requires /lib/bootchart (make install-chroot) to
> +		  be present (also in the initrd) [ with a kmsg
> +		  node included ]
> +		+ add a --probe-running mode
> +	+ ptrace re-write
> +		+ gives -much- better early-boot-time resolution
> +		+ unconditional chroot /lib/bootchart/chroot
> +			+ we mount proc there ourselves
> +		+ log extraction requires no common file-system view
> +
> +
> +bootchart2 0.10.1 (Kel Modderman)
> +	+ collector arg -m should mount /proc
> +	+ remove bogus vcsid code
> +	+ split collector install in Makefile
> +	+ remove bogus debug code
> +	+ accept process names containing spaces
> +
> +bootchart2 0.10.0
> +	+ rendering (Anders Norgaard)
> +		+ fix for unknown exceptions
> +	+ interactive UI (Michael)
> +		+ much faster rendering by manual clipping
> +		+ horizontal scaling
> +		+ remove annoying page-up/down bindings
> +	+ initrd portability & fixes (Federic Crozat)
> +		+ port to Mandriva
> +		+ improved process waiting
> +		+ inittab commenting fix
> +		+ improved initrd detection / jail tagging
> +		+ fix for un-detectable accton behaviour change
> +		+ implement a built-in usleep to help initrd deps
> (Michael) +
> +bootchart2 0.0.9
> +	+ fix initrd bug
> +
> +bootchart2 0.0.8
> +	+ add a filename string to the window title in interactive
> mode
> +	+ add a NEWS file
> diff --git a/scripts/pybootchartgui/README.pybootchart
> b/scripts/pybootchartgui/README.pybootchart new file mode 100644
> index 0000000..8642e64
> --- /dev/null
> +++ b/scripts/pybootchartgui/README.pybootchart
> @@ -0,0 +1,37 @@
> +			    PYBOOTCHARTGUI
> +			   ----------------
> +
> +pybootchartgui is a tool (now included as part of bootchart2) for
> +visualization and analysis of the GNU/Linux boot process. It renders
> +the output of the boot-logger tool bootchart (see
> +http://www.bootchart.org/) to either the screen or files of various
> +formats. Bootchart collects information about the processes, their
> +dependencies, and resource consumption during boot of a GNU/Linux
> +system. The pybootchartgui tools visualizes the process tree and
> +overall resource utilization.
> +
> +pybootchartgui is a port of the visualization part of bootchart from
> +Java to Python and Cairo.
> +
> +Adapted from the bootchart-documentation:
> +
> +  The CPU and disk statistics are used to render stacked area and
> line
> +  charts. The process information is used to create a Gantt chart
> +  showing process dependency, states and CPU usage.
> +
> +  A typical boot sequence consists of several hundred processes.
> Since
> +  it is difficult to visualize such amount of data in a
> comprehensible
> +  way, tree pruning is utilized. Idle background processes and
> +  short-lived processes are removed. Similar processes running in
> +  parallel are also merged together.
> +
> +  Finally, the performance and dependency charts are rendered as a
> +  single image to either the screen or in PNG, PDF or SVG format.
> +
> +
> +To get help for pybootchartgui, run
> +
> +$ pybootchartgui --help
> +
> +This code was originally hosted at:
> +	http://code.google.com/p/pybootchartgui/
> diff --git a/scripts/pybootchartgui/pybootchartgui.py
> b/scripts/pybootchartgui/pybootchartgui.py new file mode 100755
> index 0000000..1c4062b
> --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui.py
> @@ -0,0 +1,23 @@
> +#!/usr/bin/env python3
> +#
> +#  This file is part of pybootchartgui.
> +
> +#  pybootchartgui is free software: you can redistribute it and/or
> modify +#  it under the terms of the GNU General Public License as
> published by +#  the Free Software Foundation, either version 3 of
> the License, or +#  (at your option) any later version.
> +
> +#  pybootchartgui is distributed in the hope that it will be useful,
> +#  but WITHOUT ANY WARRANTY; without even the implied warranty of
> +#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +#  GNU General Public License for more details.
> +
> +#  You should have received a copy of the GNU General Public License
> +#  along with pybootchartgui. If not, see
> <http://www.gnu.org/licenses/>. +
> +
> +import sys
> +from pybootchartgui.main import main
> +
> +if __name__ == '__main__':
> +	sys.exit(main())
> diff --git a/scripts/pybootchartgui/pybootchartgui/__init__.py
> b/scripts/pybootchartgui/pybootchartgui/__init__.py new file mode
> 100644 index 0000000..e69de29
> diff --git a/scripts/pybootchartgui/pybootchartgui/batch.py
> b/scripts/pybootchartgui/pybootchartgui/batch.py new file mode 100644
> index 0000000..05c714e
> --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui/batch.py
> @@ -0,0 +1,46 @@
> +#  This file is part of pybootchartgui.
> +
> +#  pybootchartgui is free software: you can redistribute it and/or
> modify +#  it under the terms of the GNU General Public License as
> published by +#  the Free Software Foundation, either version 3 of
> the License, or +#  (at your option) any later version.
> +
> +#  pybootchartgui is distributed in the hope that it will be useful,
> +#  but WITHOUT ANY WARRANTY; without even the implied warranty of
> +#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +#  GNU General Public License for more details.
> +
> +#  You should have received a copy of the GNU General Public License
> +#  along with pybootchartgui. If not, see
> <http://www.gnu.org/licenses/>. +
> +import cairo
> +from . import draw
> +from .draw import RenderOptions
> +
> +def render(writer, trace, app_options, filename):
> +    handlers = {
> +        "png": (lambda w, h: cairo.ImageSurface(cairo.FORMAT_ARGB32,
> w, h), \
> +                lambda sfc: sfc.write_to_png(filename)),
> +        "pdf": (lambda w, h: cairo.PDFSurface(filename, w, h),
> lambda sfc: 0),
> +        "svg": (lambda w, h: cairo.SVGSurface(filename, w, h),
> lambda sfc: 0)
> +    }
> +
> +    if app_options.format is None:
> +        fmt = filename.rsplit('.', 1)[1]
> +    else:
> +        fmt = app_options.format
> +
> +    if not (fmt in handlers):
> +        writer.error ("Unknown format '%s'." % fmt)
> +        return 10
> +
> +    make_surface, write_surface = handlers[fmt]
> +    options = RenderOptions (app_options)
> +    (w, h) = draw.extents (options, 1.0, trace)
> +    w = max (w, draw.MIN_IMG_W)
> +    surface = make_surface (w, h)
> +    ctx = cairo.Context (surface)
> +    draw.render (ctx, options, 1.0, trace)
> +    write_surface (surface)
> +    writer.status ("bootchart written to '%s'" % filename)
> +
> diff --git a/scripts/pybootchartgui/pybootchartgui/draw.py
> b/scripts/pybootchartgui/pybootchartgui/draw.py new file mode 100644
> index 0000000..29eb750
> --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui/draw.py
> @@ -0,0 +1,975 @@
> +#  This file is part of pybootchartgui.
> +
> +#  pybootchartgui is free software: you can redistribute it and/or
> modify +#  it under the terms of the GNU General Public License as
> published by +#  the Free Software Foundation, either version 3 of
> the License, or +#  (at your option) any later version.
> +
> +#  pybootchartgui is distributed in the hope that it will be useful,
> +#  but WITHOUT ANY WARRANTY; without even the implied warranty of
> +#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +#  GNU General Public License for more details.
> +
> +#  You should have received a copy of the GNU General Public License
> +#  along with pybootchartgui. If not, see
> <http://www.gnu.org/licenses/>. +
> +
> +import cairo
> +import math
> +import re
> +import random
> +import colorsys
> +import functools
> +from operator import itemgetter
> +
> +class RenderOptions:
> +
> +    def __init__(self, app_options):
> +        # should we render a cumulative CPU time chart
> +        self.cumulative = True
> +        self.charts = True
> +        self.kernel_only = False
> +        self.app_options = app_options
> +
> +    def proc_tree (self, trace):
> +        if self.kernel_only:
> +            return trace.kernel_tree
> +        else:
> +            return trace.proc_tree
> +
> +# Process tree background color.
> +BACK_COLOR = (1.0, 1.0, 1.0, 1.0)
> +
> +WHITE = (1.0, 1.0, 1.0, 1.0)
> +# Process tree border color.
> +BORDER_COLOR = (0.63, 0.63, 0.63, 1.0)
> +# Second tick line color.
> +TICK_COLOR = (0.92, 0.92, 0.92, 1.0)
> +# 5-second tick line color.
> +TICK_COLOR_BOLD = (0.86, 0.86, 0.86, 1.0)
> +# Annotation colour
> +ANNOTATION_COLOR = (0.63, 0.0, 0.0, 0.5)
> +# Text color.
> +TEXT_COLOR = (0.0, 0.0, 0.0, 1.0)
> +
> +# Font family
> +FONT_NAME = "Bitstream Vera Sans"
> +# Title text font.
> +TITLE_FONT_SIZE = 18
> +# Default text font.
> +TEXT_FONT_SIZE = 12
> +# Axis label font.
> +AXIS_FONT_SIZE = 11
> +# Legend font.
> +LEGEND_FONT_SIZE = 12
> +
> +# CPU load chart color.
> +CPU_COLOR = (0.40, 0.55, 0.70, 1.0)
> +# IO wait chart color.
> +IO_COLOR = (0.76, 0.48, 0.48, 0.5)
> +# Disk throughput color.
> +DISK_TPUT_COLOR = (0.20, 0.71, 0.20, 1.0)
> +# CPU load chart color.
> +FILE_OPEN_COLOR = (0.20, 0.71, 0.71, 1.0)
> +# Mem cached color
> +MEM_CACHED_COLOR = CPU_COLOR
> +# Mem used color
> +MEM_USED_COLOR = IO_COLOR
> +# Buffers color
> +MEM_BUFFERS_COLOR = (0.4, 0.4, 0.4, 0.3)
> +# Swap color
> +MEM_SWAP_COLOR = DISK_TPUT_COLOR
> +
> +# Process border color.
> +PROC_BORDER_COLOR = (0.71, 0.71, 0.71, 1.0)
> +# Waiting process color.
> +PROC_COLOR_D = (0.76, 0.48, 0.48, 0.5)
> +# Running process color.
> +PROC_COLOR_R = CPU_COLOR
> +# Sleeping process color.
> +PROC_COLOR_S = (0.94, 0.94, 0.94, 1.0)
> +# Stopped process color.
> +PROC_COLOR_T = (0.94, 0.50, 0.50, 1.0)
> +# Zombie process color.
> +PROC_COLOR_Z = (0.71, 0.71, 0.71, 1.0)
> +# Dead process color.
> +PROC_COLOR_X = (0.71, 0.71, 0.71, 0.125)
> +# Paging process color.
> +PROC_COLOR_W = (0.71, 0.71, 0.71, 0.125)
> +
> +# Process label color.
> +PROC_TEXT_COLOR = (0.19, 0.19, 0.19, 1.0)
> +# Process label font.
> +PROC_TEXT_FONT_SIZE = 12
> +
> +# Signature color.
> +SIG_COLOR = (0.0, 0.0, 0.0, 0.3125)
> +# Signature font.
> +SIG_FONT_SIZE = 14
> +# Signature text.
> +SIGNATURE = "http://github.com/mmeeks/bootchart"
> +
> +# Process dependency line color.
> +DEP_COLOR = (0.75, 0.75, 0.75, 1.0)
> +# Process dependency line stroke.
> +DEP_STROKE = 1.0
> +
> +# Process description date format.
> +DESC_TIME_FORMAT = "mm:ss.SSS"
> +
> +# Cumulative coloring bits
> +HSV_MAX_MOD = 31
> +HSV_STEP = 7
> +
> +# Configure task color
> +TASK_COLOR_CONFIGURE = (1.0, 1.0, 0.00, 1.0)
> +# Compile task color.
> +TASK_COLOR_COMPILE = (0.0, 1.00, 0.00, 1.0)
> +# Install task color
> +TASK_COLOR_INSTALL = (1.0, 0.00, 1.00, 1.0)
> +# Sysroot task color
> +TASK_COLOR_SYSROOT = (0.0, 0.00, 1.00, 1.0)
> +# Package task color
> +TASK_COLOR_PACKAGE = (0.0, 1.00, 1.00, 1.0)
> +# Package Write RPM/DEB/IPK task color
> +TASK_COLOR_PACKAGE_WRITE = (0.0, 0.50, 0.50, 1.0)
> +
> +# Distinct colors used for different disk volumnes.
> +# If we have more volumns, colors get re-used.
> +VOLUME_COLORS = [
> +    (1.0, 1.0, 0.00, 1.0),
> +    (0.0, 1.00, 0.00, 1.0),
> +    (1.0, 0.00, 1.00, 1.0),
> +    (0.0, 0.00, 1.00, 1.0),
> +    (0.0, 1.00, 1.00, 1.0),
> +]
> +
> +# Process states
> +STATE_UNDEFINED = 0
> +STATE_RUNNING   = 1
> +STATE_SLEEPING  = 2
> +STATE_WAITING   = 3
> +STATE_STOPPED   = 4
> +STATE_ZOMBIE    = 5
> +
> +STATE_COLORS = [(0, 0, 0, 0), PROC_COLOR_R, PROC_COLOR_S,
> PROC_COLOR_D, \
> +        PROC_COLOR_T, PROC_COLOR_Z, PROC_COLOR_X, PROC_COLOR_W]
> +
> +# CumulativeStats Types
> +STAT_TYPE_CPU = 0
> +STAT_TYPE_IO = 1
> +
> +# Convert ps process state to an int
> +def get_proc_state(flag):
> +    return "RSDTZXW".find(flag) + 1
> +
> +def draw_text(ctx, text, color, x, y):
> +    ctx.set_source_rgba(*color)
> +    ctx.move_to(x, y)
> +    ctx.show_text(text)
> +
> +def draw_fill_rect(ctx, color, rect):
> +    ctx.set_source_rgba(*color)
> +    ctx.rectangle(*rect)
> +    ctx.fill()
> +
> +def draw_rect(ctx, color, rect):
> +    ctx.set_source_rgba(*color)
> +    ctx.rectangle(*rect)
> +    ctx.stroke()
> +
> +def draw_legend_box(ctx, label, fill_color, x, y, s):
> +    draw_fill_rect(ctx, fill_color, (x, y - s, s, s))
> +    draw_rect(ctx, PROC_BORDER_COLOR, (x, y - s, s, s))
> +    draw_text(ctx, label, TEXT_COLOR, x + s + 5, y)
> +
> +def draw_legend_line(ctx, label, fill_color, x, y, s):
> +    draw_fill_rect(ctx, fill_color, (x, y - s/2, s + 1, 3))
> +    ctx.arc(x + (s + 1)/2.0, y - (s - 3)/2.0, 2.5, 0, 2.0 * math.pi)
> +    ctx.fill()
> +    draw_text(ctx, label, TEXT_COLOR, x + s + 5, y)
> +
> +def draw_label_in_box(ctx, color, label, x, y, w, maxx):
> +    label_w = ctx.text_extents(label)[2]
> +    label_x = x + w / 2 - label_w / 2
> +    if label_w + 10 > w:
> +        label_x = x + w + 5
> +    if label_x + label_w > maxx:
> +        label_x = x - label_w - 5
> +    draw_text(ctx, label, color, label_x, y)
> +
> +def draw_sec_labels(ctx, options, rect, sec_w, nsecs):
> +    ctx.set_font_size(AXIS_FONT_SIZE)
> +    prev_x = 0
> +    for i in range(0, rect[2] + 1, sec_w):
> +        if ((i / sec_w) % nsecs == 0) :
> +            if options.app_options.as_minutes :
> +                label = "%.1f" % (i / sec_w / 60.0)
> +            else :
> +                label = "%d" % (i / sec_w)
> +            label_w = ctx.text_extents(label)[2]
> +            x = rect[0] + i - label_w/2
> +            if x >= prev_x:
> +                draw_text(ctx, label, TEXT_COLOR, x, rect[1] - 2)
> +                prev_x = x + label_w
> +
> +def draw_box_ticks(ctx, rect, sec_w):
> +    draw_rect(ctx, BORDER_COLOR, tuple(rect))
> +
> +    ctx.set_line_cap(cairo.LINE_CAP_SQUARE)
> +
> +    for i in range(sec_w, rect[2] + 1, sec_w):
> +        if ((i / sec_w) % 10 == 0) :
> +            ctx.set_line_width(1.5)
> +        elif sec_w < 5 :
> +            continue
> +        else :
> +            ctx.set_line_width(1.0)
> +        if ((i / sec_w) % 30 == 0) :
> +            ctx.set_source_rgba(*TICK_COLOR_BOLD)
> +        else :
> +            ctx.set_source_rgba(*TICK_COLOR)
> +        ctx.move_to(rect[0] + i, rect[1] + 1)
> +        ctx.line_to(rect[0] + i, rect[1] + rect[3] - 1)
> +        ctx.stroke()
> +    ctx.set_line_width(1.0)
> +
> +    ctx.set_line_cap(cairo.LINE_CAP_BUTT)
> +
> +def draw_annotations(ctx, proc_tree, times, rect):
> +    ctx.set_line_cap(cairo.LINE_CAP_SQUARE)
> +    ctx.set_source_rgba(*ANNOTATION_COLOR)
> +    ctx.set_dash([4, 4])
> +
> +    for time in times:
> +        if time is not None:
> +            x = ((time - proc_tree.start_time) * rect[2] /
> proc_tree.duration) +
> +            ctx.move_to(rect[0] + x, rect[1] + 1)
> +            ctx.line_to(rect[0] + x, rect[1] + rect[3] - 1)
> +            ctx.stroke()
> +
> +    ctx.set_line_cap(cairo.LINE_CAP_BUTT)
> +    ctx.set_dash([])
> +
> +def draw_chart(ctx, color, fill, chart_bounds, data, proc_tree,
> data_range):
> +    ctx.set_line_width(0.5)
> +    x_shift = proc_tree.start_time
> +
> +    def transform_point_coords(point, x_base, y_base, \
> +                   xscale, yscale, x_trans, y_trans):
> +        x = (point[0] - x_base) * xscale + x_trans
> +        y = (point[1] - y_base) * -yscale + y_trans + chart_bounds[3]
> +        return x, y
> +
> +    max_x = max (x for (x, y) in data)
> +    max_y = max (y for (x, y) in data)
> +    # avoid divide by zero
> +    if max_y == 0:
> +        max_y = 1.0
> +    xscale = float (chart_bounds[2]) / (max_x - x_shift)
> +    # If data_range is given, scale the chart so that the value
> range in
> +    # data_range matches the chart bounds exactly.
> +    # Otherwise, scale so that the actual data matches the chart
> bounds.
> +    if data_range and (data_range[1] - data_range[0]):
> +        yscale = float(chart_bounds[3]) / (data_range[1] -
> data_range[0])
> +        ybase = data_range[0]
> +    else:
> +        yscale = float(chart_bounds[3]) / max_y
> +        ybase = 0
> +
> +    first = transform_point_coords (data[0], x_shift, ybase, xscale,
> yscale, \
> +                        chart_bounds[0], chart_bounds[1])
> +    last =  transform_point_coords (data[-1], x_shift, ybase,
> xscale, yscale, \
> +                        chart_bounds[0], chart_bounds[1])
> +
> +    ctx.set_source_rgba(*color)
> +    ctx.move_to(*first)
> +    for point in data:
> +        x, y = transform_point_coords (point, x_shift, ybase,
> xscale, yscale, \
> +                           chart_bounds[0], chart_bounds[1])
> +        ctx.line_to(x, y)
> +    if fill:
> +        ctx.stroke_preserve()
> +        ctx.line_to(last[0], chart_bounds[1]+chart_bounds[3])
> +        ctx.line_to(first[0], chart_bounds[1]+chart_bounds[3])
> +        ctx.line_to(first[0], first[1])
> +        ctx.fill()
> +    else:
> +        ctx.stroke()
> +    ctx.set_line_width(1.0)
> +
> +bar_h = 55
> +meminfo_bar_h = 2 * bar_h
> +header_h = 60
> +# offsets
> +off_x, off_y = 220, 10
> +sec_w_base = 1 # the width of a second
> +proc_h = 16 # the height of a process
> +leg_s = 10
> +MIN_IMG_W = 800
> +CUML_HEIGHT = 2000 # Increased value to accommodate CPU and I/O
> Graphs +OPTIONS = None
> +
> +def extents(options, xscale, trace):
> +    start = min(trace.start.keys())
> +    end = start
> +
> +    processes = 0
> +    for proc in trace.processes:
> +        if not options.app_options.show_all and \
> +               trace.processes[proc][1] - trace.processes[proc][0] <
> options.app_options.mintime:
> +            continue
> +
> +        if trace.processes[proc][1] > end:
> +            end = trace.processes[proc][1]
> +        processes += 1
> +
> +    if trace.min is not None and trace.max is not None:
> +        start = trace.min
> +        end = trace.max
> +
> +    w = int ((end - start) * sec_w_base * xscale) + 2 * off_x
> +    h = proc_h * processes + header_h + 2 * off_y
> +
> +    if options.charts:
> +        if trace.cpu_stats:
> +            h += 30 + bar_h
> +        if trace.disk_stats:
> +            h += 30 + bar_h
> +        if trace.monitor_disk:
> +            h += 30 + bar_h
> +        if trace.mem_stats:
> +            h += meminfo_bar_h
> +
> +    # Allow for width of process legend and offset
> +    if w < (720 + off_x):
> +        w = 720 + off_x
> +
> +    return (w, h)
> +
> +def clip_visible(clip, rect):
> +    xmax = max (clip[0], rect[0])
> +    ymax = max (clip[1], rect[1])
> +    xmin = min (clip[0] + clip[2], rect[0] + rect[2])
> +    ymin = min (clip[1] + clip[3], rect[1] + rect[3])
> +    return (xmin > xmax and ymin > ymax)
> +
> +def render_charts(ctx, options, clip, trace, curr_y, w, h, sec_w):
> +    proc_tree = options.proc_tree(trace)
> +
> +    # render bar legend
> +    if trace.cpu_stats:
> +        ctx.set_font_size(LEGEND_FONT_SIZE)
> +
> +        draw_legend_box(ctx, "CPU (user+sys)", CPU_COLOR, off_x,
> curr_y+20, leg_s)
> +        draw_legend_box(ctx, "I/O (wait)", IO_COLOR, off_x + 120,
> curr_y+20, leg_s) +
> +        # render I/O wait
> +        chart_rect = (off_x, curr_y+30, w, bar_h)
> +        if clip_visible (clip, chart_rect):
> +            draw_box_ticks (ctx, chart_rect, sec_w)
> +            draw_annotations (ctx, proc_tree, trace.times,
> chart_rect)
> +            draw_chart (ctx, IO_COLOR, True, chart_rect, \
> +                    [(sample.time, sample.user + sample.sys +
> sample.io) for sample in trace.cpu_stats], \
> +                    proc_tree, None)
> +            # render CPU load
> +            draw_chart (ctx, CPU_COLOR, True, chart_rect, \
> +                    [(sample.time, sample.user + sample.sys) for
> sample in trace.cpu_stats], \
> +                    proc_tree, None)
> +
> +        curr_y = curr_y + 30 + bar_h
> +
> +    # render second chart
> +    if trace.disk_stats:
> +        draw_legend_line(ctx, "Disk throughput", DISK_TPUT_COLOR,
> off_x, curr_y+20, leg_s)
> +        draw_legend_box(ctx, "Disk utilization", IO_COLOR, off_x +
> 120, curr_y+20, leg_s) +
> +        # render I/O utilization
> +        chart_rect = (off_x, curr_y+30, w, bar_h)
> +        if clip_visible (clip, chart_rect):
> +            draw_box_ticks (ctx, chart_rect, sec_w)
> +            draw_annotations (ctx, proc_tree, trace.times,
> chart_rect)
> +            draw_chart (ctx, IO_COLOR, True, chart_rect, \
> +                    [(sample.time, sample.util) for sample in
> trace.disk_stats], \
> +                    proc_tree, None)
> +
> +        # render disk throughput
> +        max_sample = max (trace.disk_stats, key = lambda s: s.tput)
> +        if clip_visible (clip, chart_rect):
> +            draw_chart (ctx, DISK_TPUT_COLOR, False, chart_rect, \
> +                    [(sample.time, sample.tput) for sample in
> trace.disk_stats], \
> +                    proc_tree, None)
> +
> +        pos_x = off_x + ((max_sample.time - proc_tree.start_time) *
> w / proc_tree.duration) +
> +        shift_x, shift_y = -20, 20
> +        if (pos_x < off_x + 245):
> +            shift_x, shift_y = 5, 40
> +
> +        label = "%dMB/s" % round ((max_sample.tput) / 1024.0)
> +        draw_text (ctx, label, DISK_TPUT_COLOR, pos_x + shift_x,
> curr_y + shift_y) +
> +        curr_y = curr_y + 30 + bar_h
> +
> +    # render disk space usage
> +    #
> +    # Draws the amount of disk space used on each volume relative to
> the
> +    # lowest recorded amount. The graphs for each volume are stacked
> above
> +    # each other so that total disk usage is visible.
> +    if trace.monitor_disk:
> +        ctx.set_font_size(LEGEND_FONT_SIZE)
> +        # Determine set of volumes for which we have
> +        # information and the minimal amount of used disk
> +        # space for each. Currently samples are allowed to
> +        # not have a values for all volumes; drawing could be
> +        # made more efficient if that wasn't the case.
> +        volumes = set()
> +        min_used = {}
> +        for sample in trace.monitor_disk:
> +            for volume, used in sample.records.items():
> +                volumes.add(volume)
> +                if volume not in min_used or min_used[volume] > used:
> +                    min_used[volume] = used
> +        volumes = sorted(list(volumes))
> +        disk_scale = 0
> +        for i, volume in enumerate(volumes):
> +            volume_scale = max([sample.records[volume] -
> min_used[volume]
> +                                for sample in trace.monitor_disk
> +                                if volume in sample.records])
> +            # Does not take length of volume name into account, but
> fixed offset
> +            # works okay in practice.
> +            draw_legend_box(ctx, '%s (max: %u MiB)' % (volume,
> volume_scale / 1024 / 1024),
> +                            VOLUME_COLORS[i % len(VOLUME_COLORS)],
> +                            off_x + i * 250, curr_y+20, leg_s)
> +            disk_scale += volume_scale
> +
> +        # render used amount of disk space
> +        chart_rect = (off_x, curr_y+30, w, bar_h)
> +        if clip_visible (clip, chart_rect):
> +            draw_box_ticks (ctx, chart_rect, sec_w)
> +            draw_annotations (ctx, proc_tree, trace.times,
> chart_rect)
> +            for i in range(len(volumes), 0, -1):
> +                draw_chart (ctx, VOLUME_COLORS[(i - 1) %
> len(VOLUME_COLORS)], True, chart_rect, \
> +                            [(sample.time,
> +                              # Sum up used space of all volumes
> including the current one
> +                              # so that the graphs appear as stacked
> on top of each other.
> +                              functools.reduce(lambda x,y: x+y,
> +                                     [sample.records[volume] -
> min_used[volume]
> +                                      for volume in volumes[0:i]
> +                                      if volume in sample.records],
> +                                     0))
> +                             for sample in trace.monitor_disk], \
> +                            proc_tree, [0, disk_scale])
> +
> +        curr_y = curr_y + 30 + bar_h
> +
> +    # render mem usage
> +    chart_rect = (off_x, curr_y+30, w, meminfo_bar_h)
> +    mem_stats = trace.mem_stats
> +    if mem_stats and clip_visible (clip, chart_rect):
> +        mem_scale = max(sample.buffers for sample in mem_stats)
> +        draw_legend_box(ctx, "Mem cached (scale: %u MiB)" %
> (float(mem_scale) / 1024), MEM_CACHED_COLOR, off_x, curr_y+20, leg_s)
> +        draw_legend_box(ctx, "Used", MEM_USED_COLOR, off_x + 240,
> curr_y+20, leg_s)
> +        draw_legend_box(ctx, "Buffers", MEM_BUFFERS_COLOR, off_x +
> 360, curr_y+20, leg_s)
> +        draw_legend_line(ctx, "Swap (scale: %u MiB)" %
> max([(sample.swap)/1024 for sample in mem_stats]), \
> +                 MEM_SWAP_COLOR, off_x + 480, curr_y+20, leg_s)
> +        draw_box_ticks(ctx, chart_rect, sec_w)
> +        draw_annotations(ctx, proc_tree, trace.times, chart_rect)
> +        draw_chart(ctx, MEM_BUFFERS_COLOR, True, chart_rect, \
> +               [(sample.time, sample.buffers) for sample in
> trace.mem_stats], \
> +               proc_tree, [0, mem_scale])
> +        draw_chart(ctx, MEM_USED_COLOR, True, chart_rect, \
> +               [(sample.time, sample.used) for sample in mem_stats],
> \
> +               proc_tree, [0, mem_scale])
> +        draw_chart(ctx, MEM_CACHED_COLOR, True, chart_rect, \
> +               [(sample.time, sample.cached) for sample in
> mem_stats], \
> +               proc_tree, [0, mem_scale])
> +        draw_chart(ctx, MEM_SWAP_COLOR, False, chart_rect, \
> +               [(sample.time, float(sample.swap)) for sample in
> mem_stats], \
> +               proc_tree, None)
> +
> +        curr_y = curr_y + meminfo_bar_h
> +
> +    return curr_y
> +
> +def render_processes_chart(ctx, options, trace, curr_y, w, h, sec_w):
> +    chart_rect = [off_x, curr_y+header_h, w, h - curr_y - 1 * off_y
> - header_h  ] +
> +    draw_legend_box (ctx, "Configure", \
> +             TASK_COLOR_CONFIGURE, off_x  , curr_y + 45, leg_s)
> +    draw_legend_box (ctx, "Compile", \
> +             TASK_COLOR_COMPILE, off_x+120, curr_y + 45, leg_s)
> +    draw_legend_box (ctx, "Install", \
> +             TASK_COLOR_INSTALL, off_x+240, curr_y + 45, leg_s)
> +    draw_legend_box (ctx, "Populate Sysroot", \
> +             TASK_COLOR_SYSROOT, off_x+360, curr_y + 45, leg_s)
> +    draw_legend_box (ctx, "Package", \
> +             TASK_COLOR_PACKAGE, off_x+480, curr_y + 45, leg_s)
> +    draw_legend_box (ctx, "Package Write", \
> +             TASK_COLOR_PACKAGE_WRITE, off_x+600, curr_y + 45, leg_s)
> +
> +    ctx.set_font_size(PROC_TEXT_FONT_SIZE)
> +
> +    draw_box_ticks(ctx, chart_rect, sec_w)
> +    draw_sec_labels(ctx, options, chart_rect, sec_w, 30)
> +
> +    y = curr_y+header_h
> +
> +    offset = trace.min or min(trace.start.keys())
> +    for start in sorted(trace.start.keys()):
> +        for process in sorted(trace.start[start]):
> +            if not options.app_options.show_all and \
> +                    trace.processes[process][1] - start <
> options.app_options.mintime:
> +                continue
> +            task = process.split(":")[1]
> +
> +            #print(process)
> +            #print(trace.processes[process][1])
> +            #print(s)
> +
> +            x = chart_rect[0] + (start - offset) * sec_w
> +            w = ((trace.processes[process][1] - start) * sec_w)
> +
> +            #print("proc at %s %s %s %s" % (x, y, w, proc_h))
> +            col = None
> +            if task == "do_compile":
> +                col = TASK_COLOR_COMPILE
> +            elif task == "do_configure":
> +                col = TASK_COLOR_CONFIGURE
> +            elif task == "do_install":
> +                col = TASK_COLOR_INSTALL
> +            elif task == "do_populate_sysroot":
> +                col = TASK_COLOR_SYSROOT
> +            elif task == "do_package":
> +                col = TASK_COLOR_PACKAGE
> +            elif task == "do_package_write_rpm" or \
> +                     task == "do_package_write_deb" or \
> +                     task == "do_package_write_ipk":
> +                col = TASK_COLOR_PACKAGE_WRITE
> +            else:
> +                col = WHITE
> +
> +            if col:
> +                draw_fill_rect(ctx, col, (x, y, w, proc_h))
> +            draw_rect(ctx, PROC_BORDER_COLOR, (x, y, w, proc_h))
> +
> +            draw_label_in_box(ctx, PROC_TEXT_COLOR, process, x, y +
> proc_h - 4, w, proc_h)
> +            y = y + proc_h
> +
> +    return curr_y
> +
> +#
> +# Render the chart.
> +#
> +def render(ctx, options, xscale, trace):
> +    (w, h) = extents (options, xscale, trace)
> +    global OPTIONS
> +    OPTIONS = options.app_options
> +
> +    # x, y, w, h
> +    clip = ctx.clip_extents()
> +
> +    sec_w = int (xscale * sec_w_base)
> +    ctx.set_line_width(1.0)
> +    ctx.select_font_face(FONT_NAME)
> +    draw_fill_rect(ctx, WHITE, (0, 0, max(w, MIN_IMG_W), h))
> +    w -= 2*off_x
> +    curr_y = off_y;
> +
> +    if options.charts:
> +        curr_y = render_charts (ctx, options, clip, trace, curr_y,
> w, h, sec_w) +
> +    curr_y = render_processes_chart (ctx, options, trace, curr_y, w,
> h, sec_w) +
> +    return
> +
> +    proc_tree = options.proc_tree (trace)
> +
> +    # draw the title and headers
> +    if proc_tree.idle:
> +        duration = proc_tree.idle
> +    else:
> +        duration = proc_tree.duration
> +
> +    if not options.kernel_only:
> +        curr_y = draw_header (ctx, trace.headers, duration)
> +    else:
> +        curr_y = off_y;
> +
> +    # draw process boxes
> +    proc_height = h
> +    if proc_tree.taskstats and options.cumulative:
> +        proc_height -= CUML_HEIGHT
> +
> +    draw_process_bar_chart(ctx, clip, options, proc_tree,
> trace.times,
> +                   curr_y, w, proc_height, sec_w)
> +
> +    curr_y = proc_height
> +    ctx.set_font_size(SIG_FONT_SIZE)
> +    draw_text(ctx, SIGNATURE, SIG_COLOR, off_x + 5, proc_height - 8)
> +
> +    # draw a cumulative CPU-time-per-process graph
> +    if proc_tree.taskstats and options.cumulative:
> +        cuml_rect = (off_x, curr_y + off_y, w, CUML_HEIGHT/2 - off_y
> * 2)
> +        if clip_visible (clip, cuml_rect):
> +            draw_cuml_graph(ctx, proc_tree, cuml_rect, duration,
> sec_w, STAT_TYPE_CPU) +
> +    # draw a cumulative I/O-time-per-process graph
> +    if proc_tree.taskstats and options.cumulative:
> +        cuml_rect = (off_x, curr_y + off_y * 100, w, CUML_HEIGHT/2 -
> off_y * 2)
> +        if clip_visible (clip, cuml_rect):
> +            draw_cuml_graph(ctx, proc_tree, cuml_rect, duration,
> sec_w, STAT_TYPE_IO) +
> +def draw_process_bar_chart(ctx, clip, options, proc_tree, times,
> curr_y, w, h, sec_w):
> +    header_size = 0
> +    if not options.kernel_only:
> +        draw_legend_box (ctx, "Running (%cpu)",
> +                 PROC_COLOR_R, off_x    , curr_y + 45, leg_s)
> +        draw_legend_box (ctx, "Unint.sleep (I/O)",
> +                 PROC_COLOR_D, off_x+120, curr_y + 45, leg_s)
> +        draw_legend_box (ctx, "Sleeping",
> +                 PROC_COLOR_S, off_x+240, curr_y + 45, leg_s)
> +        draw_legend_box (ctx, "Zombie",
> +                 PROC_COLOR_Z, off_x+360, curr_y + 45, leg_s)
> +        header_size = 45
> +
> +    chart_rect = [off_x, curr_y + header_size + 15,
> +              w, h - 2 * off_y - (curr_y + header_size + 15) +
> proc_h]
> +    ctx.set_font_size (PROC_TEXT_FONT_SIZE)
> +
> +    draw_box_ticks (ctx, chart_rect, sec_w)
> +    if sec_w > 100:
> +        nsec = 1
> +    else:
> +        nsec = 5
> +    draw_sec_labels (ctx, options, chart_rect, sec_w, nsec)
> +    draw_annotations (ctx, proc_tree, times, chart_rect)
> +
> +    y = curr_y + 60
> +    for root in proc_tree.process_tree:
> +        draw_processes_recursively(ctx, root, proc_tree, y, proc_h,
> chart_rect, clip)
> +        y = y + proc_h * proc_tree.num_nodes([root])
> +
> +
> +def draw_header (ctx, headers, duration):
> +    toshow = [
> +      ('system.uname', 'uname', lambda s: s),
> +      ('system.release', 'release', lambda s: s),
> +      ('system.cpu', 'CPU', lambda s: re.sub('model name\s*:\s*',
> '', s, 1)),
> +      ('system.kernel.options', 'kernel options', lambda s: s),
> +    ]
> +
> +    header_y = ctx.font_extents()[2] + 10
> +    ctx.set_font_size(TITLE_FONT_SIZE)
> +    draw_text(ctx, headers['title'], TEXT_COLOR, off_x, header_y)
> +    ctx.set_font_size(TEXT_FONT_SIZE)
> +
> +    for (headerkey, headertitle, mangle) in toshow:
> +        header_y += ctx.font_extents()[2]
> +        if headerkey in headers:
> +            value = headers.get(headerkey)
> +        else:
> +            value = ""
> +        txt = headertitle + ': ' + mangle(value)
> +        draw_text(ctx, txt, TEXT_COLOR, off_x, header_y)
> +
> +    dur = duration / 100.0
> +    txt = 'time : %02d:%05.2f' % (math.floor(dur/60), dur - 60 *
> math.floor(dur/60))
> +    if headers.get('system.maxpid') is not None:
> +        txt = txt + '      max pid: %s' %
> (headers.get('system.maxpid')) +
> +    header_y += ctx.font_extents()[2]
> +    draw_text (ctx, txt, TEXT_COLOR, off_x, header_y)
> +
> +    return header_y
> +
> +def draw_processes_recursively(ctx, proc, proc_tree, y, proc_h,
> rect, clip) :
> +    x = rect[0] +  ((proc.start_time - proc_tree.start_time) *
> rect[2] / proc_tree.duration)
> +    w = ((proc.duration) * rect[2] / proc_tree.duration)
> +
> +    draw_process_activity_colors(ctx, proc, proc_tree, x, y, w,
> proc_h, rect, clip)
> +    draw_rect(ctx, PROC_BORDER_COLOR, (x, y, w, proc_h))
> +    ipid = int(proc.pid)
> +    if not OPTIONS.show_all:
> +        cmdString = proc.cmd
> +    else:
> +        cmdString = ''
> +    if (OPTIONS.show_pid or OPTIONS.show_all) and ipid is not 0:
> +        cmdString = cmdString + " [" + str(ipid // 1000) + "]"
> +    if OPTIONS.show_all:
> +        if proc.args:
> +            cmdString = cmdString + " '" + "' '".join(proc.args) +
> "'"
> +        else:
> +            cmdString = cmdString + " " + proc.exe
> +
> +    draw_label_in_box(ctx, PROC_TEXT_COLOR, cmdString, x, y + proc_h
> - 4, w, rect[0] + rect[2]) +
> +    next_y = y + proc_h
> +    for child in proc.child_list:
> +        if next_y > clip[1] + clip[3]:
> +            break
> +        child_x, child_y = draw_processes_recursively(ctx, child,
> proc_tree, next_y, proc_h, rect, clip)
> +        draw_process_connecting_lines(ctx, x, y, child_x, child_y,
> proc_h)
> +        next_y = next_y + proc_h * proc_tree.num_nodes([child])
> +
> +    return x, y
> +
> +
> +def draw_process_activity_colors(ctx, proc, proc_tree, x, y, w,
> proc_h, rect, clip): +
> +    if y > clip[1] + clip[3] or y + proc_h + 2 < clip[1]:
> +        return
> +
> +    draw_fill_rect(ctx, PROC_COLOR_S, (x, y, w, proc_h))
> +
> +    last_tx = -1
> +    for sample in proc.samples :
> +        tx = rect[0] + round(((sample.time - proc_tree.start_time) *
> rect[2] / proc_tree.duration)) +
> +        # samples are sorted chronologically
> +        if tx < clip[0]:
> +            continue
> +        if tx > clip[0] + clip[2]:
> +            break
> +
> +        tw = round(proc_tree.sample_period * rect[2] /
> float(proc_tree.duration))
> +        if last_tx != -1 and abs(last_tx - tx) <= tw:
> +            tw -= last_tx - tx
> +            tx = last_tx
> +        tw = max (tw, 1) # nice to see at least something
> +
> +        last_tx = tx + tw
> +        state = get_proc_state( sample.state )
> +
> +        color = STATE_COLORS[state]
> +        if state == STATE_RUNNING:
> +            alpha = min (sample.cpu_sample.user +
> sample.cpu_sample.sys, 1.0)
> +            color = tuple(list(PROC_COLOR_R[0:3]) + [alpha])
> +#            print "render time %d [ tx %d tw %d ], sample state %s
> color %s alpha %g" % (sample.time, tx, tw, state, color, alpha)
> +        elif state == STATE_SLEEPING:
> +            continue
> +
> +        draw_fill_rect(ctx, color, (tx, y, tw, proc_h))
> +
> +def draw_process_connecting_lines(ctx, px, py, x, y, proc_h):
> +    ctx.set_source_rgba(*DEP_COLOR)
> +    ctx.set_dash([2, 2])
> +    if abs(px - x) < 3:
> +        dep_off_x = 3
> +        dep_off_y = proc_h / 4
> +        ctx.move_to(x, y + proc_h / 2)
> +        ctx.line_to(px - dep_off_x, y + proc_h / 2)
> +        ctx.line_to(px - dep_off_x, py - dep_off_y)
> +        ctx.line_to(px, py - dep_off_y)
> +    else:
> +        ctx.move_to(x, y + proc_h / 2)
> +        ctx.line_to(px, y + proc_h / 2)
> +        ctx.line_to(px, py)
> +    ctx.stroke()
> +    ctx.set_dash([])
> +
> +# elide the bootchart collector - it is quite distorting
> +def elide_bootchart(proc):
> +    return proc.cmd == 'bootchartd' or proc.cmd == 'bootchart-colle'
> +
> +class CumlSample:
> +    def __init__(self, proc):
> +        self.cmd = proc.cmd
> +        self.samples = []
> +        self.merge_samples (proc)
> +        self.color = None
> +
> +    def merge_samples(self, proc):
> +        self.samples.extend (proc.samples)
> +        self.samples.sort (key = lambda p: p.time)
> +
> +    def next(self):
> +        global palette_idx
> +        palette_idx += HSV_STEP
> +        return palette_idx
> +
> +    def get_color(self):
> +        if self.color is None:
> +            i = self.next() % HSV_MAX_MOD
> +            h = 0.0
> +            if i is not 0:
> +                h = (1.0 * i) / HSV_MAX_MOD
> +            s = 0.5
> +            v = 1.0
> +            c = colorsys.hsv_to_rgb (h, s, v)
> +            self.color = (c[0], c[1], c[2], 1.0)
> +        return self.color
> +
> +
> +def draw_cuml_graph(ctx, proc_tree, chart_bounds, duration, sec_w,
> stat_type):
> +    global palette_idx
> +    palette_idx = 0
> +
> +    time_hash = {}
> +    total_time = 0.0
> +    m_proc_list = {}
> +
> +    if stat_type is STAT_TYPE_CPU:
> +        sample_value = 'cpu'
> +    else:
> +        sample_value = 'io'
> +    for proc in proc_tree.process_list:
> +        if elide_bootchart(proc):
> +            continue
> +
> +        for sample in proc.samples:
> +            total_time += getattr(sample.cpu_sample, sample_value)
> +            if not sample.time in time_hash:
> +                time_hash[sample.time] = 1
> +
> +        # merge pids with the same cmd
> +        if not proc.cmd in m_proc_list:
> +            m_proc_list[proc.cmd] = CumlSample (proc)
> +            continue
> +        s = m_proc_list[proc.cmd]
> +        s.merge_samples (proc)
> +
> +    # all the sample times
> +    times = sorted(time_hash)
> +    if len (times) < 2:
> +        print("degenerate boot chart")
> +        return
> +
> +    pix_per_ns = chart_bounds[3] / total_time
> +#    print "total time: %g pix-per-ns %g" % (total_time, pix_per_ns)
> +
> +    # FIXME: we have duplicates in the process list too [!] - why !?
> +
> +    # Render bottom up, left to right
> +    below = {}
> +    for time in times:
> +        below[time] = chart_bounds[1] + chart_bounds[3]
> +
> +    # same colors each time we render
> +    random.seed (0)
> +
> +    ctx.set_line_width(1)
> +
> +    legends = []
> +    labels = []
> +
> +    # render each pid in order
> +    for cs in m_proc_list.values():
> +        row = {}
> +        cuml = 0.0
> +
> +        # print "pid : %s -> %g samples %d" % (proc.cmd, cuml, len
> (cs.samples))
> +        for sample in cs.samples:
> +            cuml += getattr(sample.cpu_sample, sample_value)
> +            row[sample.time] = cuml
> +
> +        process_total_time = cuml
> +
> +        # hide really tiny processes
> +        if cuml * pix_per_ns <= 2:
> +            continue
> +
> +        last_time = times[0]
> +        y = last_below = below[last_time]
> +        last_cuml = cuml = 0.0
> +
> +        ctx.set_source_rgba(*cs.get_color())
> +        for time in times:
> +            render_seg = False
> +
> +            # did the underlying trend increase ?
> +            if below[time] != last_below:
> +                last_below = below[last_time]
> +                last_cuml = cuml
> +                render_seg = True
> +
> +            # did we move up a pixel increase ?
> +            if time in row:
> +                nc = round (row[time] * pix_per_ns)
> +                if nc != cuml:
> +                    last_cuml = cuml
> +                    cuml = nc
> +                    render_seg = True
> +
> +#            if last_cuml > cuml:
> +#                assert fail ... - un-sorted process samples
> +
> +            # draw the trailing rectangle from the last time to
> +            # before now, at the height of the last segment.
> +            if render_seg:
> +                w = math.ceil ((time - last_time) * chart_bounds[2]
> / proc_tree.duration) + 1
> +                x = chart_bounds[0] + round((last_time -
> proc_tree.start_time) * chart_bounds[2] / proc_tree.duration)
> +                ctx.rectangle (x, below[last_time] - last_cuml, w,
> last_cuml)
> +                ctx.fill()
> +#                ctx.stroke()
> +                last_time = time
> +                y = below [time] - cuml
> +
> +            row[time] = y
> +
> +        # render the last segment
> +        x = chart_bounds[0] + round((last_time -
> proc_tree.start_time) * chart_bounds[2] / proc_tree.duration)
> +        y = below[last_time] - cuml
> +        ctx.rectangle (x, y, chart_bounds[2] - x, cuml)
> +        ctx.fill()
> +#        ctx.stroke()
> +
> +        # render legend if it will fit
> +        if cuml > 8:
> +            label = cs.cmd
> +            extnts = ctx.text_extents(label)
> +            label_w = extnts[2]
> +            label_h = extnts[3]
> +#            print "Text extents %g by %g" % (label_w, label_h)
> +            labels.append((label,
> +                       chart_bounds[0] + chart_bounds[2] - label_w -
> off_x * 2,
> +                       y + (cuml + label_h) / 2))
> +            if cs in legends:
> +                print("ARGH - duplicate process in list !")
> +
> +        legends.append ((cs, process_total_time))
> +
> +        below = row
> +
> +    # render grid-lines over the top
> +    draw_box_ticks(ctx, chart_bounds, sec_w)
> +
> +    # render labels
> +    for l in labels:
> +        draw_text(ctx, l[0], TEXT_COLOR, l[1], l[2])
> +
> +    # Render legends
> +    font_height = 20
> +    label_width = 300
> +    LEGENDS_PER_COL = 15
> +    LEGENDS_TOTAL = 45
> +    ctx.set_font_size (TITLE_FONT_SIZE)
> +    dur_secs = duration / 100
> +    cpu_secs = total_time / 1000000000
> +
> +    # misleading - with multiple CPUs ...
> +#    idle = ((dur_secs - cpu_secs) / dur_secs) * 100.0
> +    if stat_type is STAT_TYPE_CPU:
> +        label = "Cumulative CPU usage, by process; total CPU: " \
> +            " %.5g(s) time: %.3g(s)" % (cpu_secs, dur_secs)
> +    else:
> +        label = "Cumulative I/O usage, by process; total I/O: " \
> +            " %.5g(s) time: %.3g(s)" % (cpu_secs, dur_secs)
> +
> +    draw_text(ctx, label, TEXT_COLOR, chart_bounds[0] + off_x,
> +          chart_bounds[1] + font_height)
> +
> +    i = 0
> +    legends = sorted(legends, key=itemgetter(1), reverse=True)
> +    ctx.set_font_size(TEXT_FONT_SIZE)
> +    for t in legends:
> +        cs = t[0]
> +        time = t[1]
> +        x = chart_bounds[0] + off_x + int (i/LEGENDS_PER_COL) *
> label_width
> +        y = chart_bounds[1] + font_height * ((i % LEGENDS_PER_COL) +
> 2)
> +        str = "%s - %.0f(ms) (%2.2f%%)" % (cs.cmd, time/1000000,
> (time/total_time) * 100.0)
> +        draw_legend_box(ctx, str, cs.color, x, y, leg_s)
> +        i = i + 1
> +        if i >= LEGENDS_TOTAL:
> +            break
> diff --git a/scripts/pybootchartgui/pybootchartgui/gui.py
> b/scripts/pybootchartgui/pybootchartgui/gui.py new file mode 100644
> index 0000000..e1fe915
> --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui/gui.py
> @@ -0,0 +1,348 @@
> +#  This file is part of pybootchartgui.
> +
> +#  pybootchartgui is free software: you can redistribute it and/or
> modify +#  it under the terms of the GNU General Public License as
> published by +#  the Free Software Foundation, either version 3 of
> the License, or +#  (at your option) any later version.
> +
> +#  pybootchartgui is distributed in the hope that it will be useful,
> +#  but WITHOUT ANY WARRANTY; without even the implied warranty of
> +#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +#  GNU General Public License for more details.
> +
> +#  You should have received a copy of the GNU General Public License
> +#  along with pybootchartgui. If not, see
> <http://www.gnu.org/licenses/>. +
> +import gi
> +gi.require_version('Gtk', '3.0') 
> +from gi.repository import Gtk as gtk
> +from gi.repository import Gtk
> +from gi.repository import Gdk
> +from gi.repository import GObject as gobject
> +from gi.repository import GObject
> +
> +from . import draw
> +from .draw import RenderOptions
> +
> +class PyBootchartWidget(gtk.DrawingArea, gtk.Scrollable):
> +    __gsignals__ = {
> +            'clicked' : (gobject.SIGNAL_RUN_LAST, gobject.TYPE_NONE,
> (gobject.TYPE_STRING, Gdk.Event)),
> +            'position-changed' : (gobject.SIGNAL_RUN_LAST,
> gobject.TYPE_NONE, (gobject.TYPE_INT, gobject.TYPE_INT)),
> +            'set-scroll-adjustments' : (gobject.SIGNAL_RUN_LAST,
> gobject.TYPE_NONE, (gtk.Adjustment, gtk.Adjustment))
> +    }
> +
> +    hadjustment = GObject.property(type=Gtk.Adjustment,
> +                                   default=Gtk.Adjustment(),
> +                                   flags=GObject.PARAM_READWRITE)
> +    hscroll_policy = GObject.property(type=Gtk.ScrollablePolicy,
> +
> default=Gtk.ScrollablePolicy.MINIMUM,
> +                                      flags=GObject.PARAM_READWRITE)
> +    vadjustment = GObject.property(type=Gtk.Adjustment,
> +                                   default=Gtk.Adjustment(),
> +                                   flags=GObject.PARAM_READWRITE)
> +    vscroll_policy = GObject.property(type=Gtk.ScrollablePolicy,
> +
> default=Gtk.ScrollablePolicy.MINIMUM,
> +                                      flags=GObject.PARAM_READWRITE)
> +
> +    def __init__(self, trace, options, xscale):
> +        gtk.DrawingArea.__init__(self)
> +
> +        self.trace = trace
> +        self.options = options
> +
> +        self.set_can_focus(True)
> +
> +        self.add_events(Gdk.EventMask.BUTTON_PRESS_MASK |
> Gdk.EventMask.BUTTON_RELEASE_MASK)
> +        self.connect("button-press-event", self.on_area_button_press)
> +        self.connect("button-release-event",
> self.on_area_button_release)
> +        self.add_events(Gdk.EventMask.POINTER_MOTION_MASK |
> Gdk.EventMask.POINTER_MOTION_HINT_MASK |
> Gdk.EventMask.BUTTON_RELEASE_MASK)
> +        self.connect("motion-notify-event",
> self.on_area_motion_notify)
> +        self.connect("scroll-event", self.on_area_scroll_event)
> +        self.connect('key-press-event', self.on_key_press_event)
> +
> +        self.connect("size-allocate",
> self.on_allocation_size_changed)
> +        self.connect("position-changed", self.on_position_changed)
> +
> +        self.connect("draw", self.on_draw)
> +
> +        self.zoom_ratio = 1.0
> +        self.xscale = xscale
> +        self.x, self.y = 0.0, 0.0
> +
> +        self.chart_width, self.chart_height =
> draw.extents(self.options, self.xscale, self.trace)
> +        self.our_width, self.our_height = self.chart_width,
> self.chart_height +
> +        self.hadj = gtk.Adjustment(0.0, 0.0, 0.0, 0.0, 0.0, 0.0)
> +        self.vadj = gtk.Adjustment(0.0, 0.0, 0.0, 0.0, 0.0, 0.0)
> +        self.vadj.connect('value-changed',
> self.on_adjustments_changed)
> +        self.hadj.connect('value-changed',
> self.on_adjustments_changed) +
> +    def bound_vals(self):
> +        self.x = max(0, self.x)
> +        self.y = max(0, self.y)
> +        self.x = min(self.chart_width - self.our_width, self.x)
> +        self.y = min(self.chart_height - self.our_height, self.y)
> +
> +    def on_draw(self, darea, cr):
> +        # set a clip region
> +        #cr.rectangle(
> +        #        self.x, self.y,
> +        #        self.chart_width, self.chart_height
> +        #)
> +        #cr.clip()
> +        cr.set_source_rgba(1.0, 1.0, 1.0, 1.0)
> +        cr.paint()
> +        cr.scale(self.zoom_ratio, self.zoom_ratio)
> +        cr.translate(-self.x, -self.y)
> +        draw.render(cr, self.options, self.xscale, self.trace)
> +
> +    def position_changed(self):
> +        self.emit("position-changed", self.x, self.y)
> +
> +    ZOOM_INCREMENT = 1.25
> +
> +    def zoom_image (self, zoom_ratio):
> +        self.zoom_ratio = zoom_ratio
> +        self._set_scroll_adjustments()
> +        self.queue_draw()
> +
> +    def zoom_to_rect (self, rect):
> +        zoom_ratio = float(rect.width)/float(self.chart_width)
> +        self.zoom_image(zoom_ratio)
> +        self.x = 0
> +        self.position_changed()
> +
> +    def set_xscale(self, xscale):
> +        old_mid_x = self.x + self.hadj.page_size / 2
> +        self.xscale = xscale
> +        self.chart_width, self.chart_height =
> draw.extents(self.options, self.xscale, self.trace)
> +        new_x = old_mid_x
> +        self.zoom_image (self.zoom_ratio)
> +
> +    def on_expand(self, action):
> +        self.set_xscale (int(self.xscale * 1.5 + 0.5))
> +
> +    def on_contract(self, action):
> +        self.set_xscale (max(int(self.xscale / 1.5), 1))
> +
> +    def on_zoom_in(self, action):
> +        self.zoom_image(self.zoom_ratio * self.ZOOM_INCREMENT)
> +
> +    def on_zoom_out(self, action):
> +        self.zoom_image(self.zoom_ratio / self.ZOOM_INCREMENT)
> +
> +    def on_zoom_fit(self, action):
> +        self.zoom_to_rect(self.get_allocation())
> +
> +    def on_zoom_100(self, action):
> +        self.zoom_image(1.0)
> +        self.set_xscale(1.0)
> +
> +    def show_toggled(self, button):
> +        self.options.app_options.show_all = button.get_property
> ('active')
> +        self.chart_width, self.chart_height =
> draw.extents(self.options, self.xscale, self.trace)
> +        self._set_scroll_adjustments()
> +        self.queue_draw()
> +
> +    POS_INCREMENT = 100
> +
> +    def on_key_press_event(self, widget, event):
> +        if event.keyval == Gdk.keyval_from_name("Left"):
> +            self.x -= self.POS_INCREMENT/self.zoom_ratio
> +        elif event.keyval == Gdk.keyval_from_name("Right"):
> +            self.x += self.POS_INCREMENT/self.zoom_ratio
> +        elif event.keyval == Gdk.keyval_from_name("Up"):
> +            self.y -= self.POS_INCREMENT/self.zoom_ratio
> +        elif event.keyval == Gdk.keyval_from_name("Down"):
> +            self.y += self.POS_INCREMENT/self.zoom_ratio
> +        else:
> +            return False
> +        self.bound_vals()
> +        self.queue_draw()
> +        self.position_changed()
> +        return True
> +
> +    def on_area_button_press(self, area, event):
> +        if event.button == 2 or event.button == 1:
> +            window = self.get_window()
> +            window.set_cursor(Gdk.Cursor(Gdk.CursorType.FLEUR))
> +            self.prevmousex = event.x
> +            self.prevmousey = event.y
> +        if event.type not in (Gdk.EventType.BUTTON_PRESS,
> Gdk.EventType.BUTTON_RELEASE):
> +            return False
> +        return False
> +
> +    def on_area_button_release(self, area, event):
> +        if event.button == 2 or event.button == 1:
> +            window = self.get_window()
> +            window.set_cursor(Gdk.Cursor(Gdk.CursorType.ARROW))
> +            self.prevmousex = None
> +            self.prevmousey = None
> +            return True
> +        return False
> +
> +    def on_area_scroll_event(self, area, event):
> +        if event.state & Gdk.CONTROL_MASK:
> +            if event.direction == Gdk.SCROLL_UP:
> +                self.zoom_image(self.zoom_ratio *
> self.ZOOM_INCREMENT)
> +                return True
> +            if event.direction == Gdk.SCROLL_DOWN:
> +                self.zoom_image(self.zoom_ratio /
> self.ZOOM_INCREMENT)
> +                return True
> +            return False
> +
> +    def on_area_motion_notify(self, area, event):
> +        state = event.state
> +        if state & Gdk.ModifierType.BUTTON2_MASK or state &
> Gdk.ModifierType.BUTTON1_MASK:
> +            x, y = int(event.x), int(event.y)
> +            # pan the image
> +            self.x += (self.prevmousex - x)/self.zoom_ratio
> +            self.y += (self.prevmousey - y)/self.zoom_ratio
> +            self.bound_vals()
> +            self.queue_draw()
> +            self.prevmousex = x
> +            self.prevmousey = y
> +            self.position_changed()
> +        return True
> +
> +    def on_allocation_size_changed(self, widget, allocation):
> +        self.hadj.page_size = allocation.width
> +        self.hadj.page_increment = allocation.width * 0.9
> +        self.vadj.page_size = allocation.height
> +        self.vadj.page_increment = allocation.height * 0.9
> +        self.our_width = allocation.width
> +        if self.chart_width < self.our_width:
> +            self.our_width = self.chart_width
> +        self.our_height = allocation.height
> +        if self.chart_height < self.our_height:
> +            self.our_height = self.chart_height
> +        self._set_scroll_adjustments()
> +
> +    def _set_adj_upper(self, adj, upper):
> +
> +        if adj.get_upper() != upper:
> +            adj.set_upper(upper)
> +
> +    def _set_scroll_adjustments(self):
> +        self._set_adj_upper (self.hadj, self.zoom_ratio *
> (self.chart_width - self.our_width))
> +        self._set_adj_upper (self.vadj, self.zoom_ratio *
> (self.chart_height - self.our_height)) +
> +    def on_adjustments_changed(self, adj):
> +        self.x = self.hadj.get_value() / self.zoom_ratio
> +        self.y = self.vadj.get_value() / self.zoom_ratio
> +        self.queue_draw()
> +
> +    def on_position_changed(self, widget, x, y):
> +        self.hadj.set_value(x * self.zoom_ratio)
> +        #self.hadj.value_changed()
> +        self.vadj.set_value(y * self.zoom_ratio)
> +
> +class PyBootchartShell(gtk.VBox):
> +    ui = '''
> +    <ui>
> +            <toolbar name="ToolBar">
> +                    <toolitem action="Expand"/>
> +                    <toolitem action="Contract"/>
> +                    <separator/>
> +                    <toolitem action="ZoomIn"/>
> +                    <toolitem action="ZoomOut"/>
> +                    <toolitem action="ZoomFit"/>
> +                    <toolitem action="Zoom100"/>
> +            </toolbar>
> +    </ui>
> +    '''
> +    def __init__(self, window, trace, options, xscale):
> +        gtk.VBox.__init__(self)
> +
> +        self.widget2 = PyBootchartWidget(trace, options, xscale)
> +
> +        # Create a UIManager instance
> +        uimanager = self.uimanager = gtk.UIManager()
> +
> +        # Add the accelerator group to the toplevel window
> +        accelgroup = uimanager.get_accel_group()
> +        window.add_accel_group(accelgroup)
> +
> +        # Create an ActionGroup
> +        actiongroup = gtk.ActionGroup('Actions')
> +        self.actiongroup = actiongroup
> +
> +        # Create actions
> +        actiongroup.add_actions((
> +                ('Expand', gtk.STOCK_ADD, None, None, None,
> self.widget2.on_expand),
> +                ('Contract', gtk.STOCK_REMOVE, None, None, None,
> self.widget2.on_contract),
> +                ('ZoomIn', gtk.STOCK_ZOOM_IN, None, None, None,
> self.widget2.on_zoom_in),
> +                ('ZoomOut', gtk.STOCK_ZOOM_OUT, None, None, None,
> self.widget2.on_zoom_out),
> +                ('ZoomFit', gtk.STOCK_ZOOM_FIT, 'Fit Width', None,
> None, self.widget2.on_zoom_fit),
> +                ('Zoom100', gtk.STOCK_ZOOM_100, None, None, None,
> self.widget2.on_zoom_100),
> +        ))
> +
> +        # Add the actiongroup to the uimanager
> +        uimanager.insert_action_group(actiongroup, 0)
> +
> +        # Add a UI description
> +        uimanager.add_ui_from_string(self.ui)
> +
> +        # Scrolled window
> +        scrolled = gtk.ScrolledWindow(self.widget2.hadj,
> self.widget2.vadj)
> +        scrolled.add(self.widget2)
> +
> +        #scrolled.set_hadjustment()
> +        #scrolled.set_vadjustment(self.widget2.vadj)
> +        scrolled.set_policy(gtk.PolicyType.ALWAYS,
> gtk.PolicyType.ALWAYS) +
> +        # toolbar / h-box
> +        hbox = gtk.HBox(False, 8)
> +
> +        # Create a Toolbar
> +        toolbar = uimanager.get_widget('/ToolBar')
> +        hbox.pack_start(toolbar, True, True, 0)
> +
> +        if not options.kernel_only:
> +            # Misc. options
> +            button = gtk.CheckButton("Show more")
> +            button.connect ('toggled', self.widget2.show_toggled)
> +            button.set_active(options.app_options.show_all)
> +            hbox.pack_start (button, False, True, 0)
> +
> +        self.pack_start(hbox, False, True, 0)
> +        self.pack_start(scrolled, True, True, 0)
> +        self.show_all()
> +
> +    def grab_focus(self, window):
> +        window.set_focus(self.widget2)
> +
> +
> +class PyBootchartWindow(gtk.Window):
> +
> +    def __init__(self, trace, app_options):
> +        gtk.Window.__init__(self)
> +
> +        window = self
> +        window.set_title("Bootchart %s" % trace.filename)
> +        window.set_default_size(750, 550)
> +
> +        tab_page = gtk.Notebook()
> +        tab_page.show()
> +        window.add(tab_page)
> +
> +        full_opts = RenderOptions(app_options)
> +        full_tree = PyBootchartShell(window, trace, full_opts, 1.0)
> +        tab_page.append_page (full_tree, gtk.Label("Full tree"))
> +
> +        if trace.kernel is not None and len (trace.kernel) > 2:
> +            kernel_opts = RenderOptions(app_options)
> +            kernel_opts.cumulative = False
> +            kernel_opts.charts = False
> +            kernel_opts.kernel_only = True
> +            kernel_tree = PyBootchartShell(window, trace,
> kernel_opts, 5.0)
> +            tab_page.append_page (kernel_tree, gtk.Label("Kernel
> boot")) +
> +        full_tree.grab_focus(self)
> +        self.show()
> +
> +
> +def show(trace, options):
> +    win = PyBootchartWindow(trace, options)
> +    win.connect('destroy', gtk.main_quit)
> +    gtk.main()
> diff --git a/scripts/pybootchartgui/pybootchartgui/main.py
> b/scripts/pybootchartgui/pybootchartgui/main.py new file mode 120000
> index 0000000..b45ae0a
> --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui/main.py
> @@ -0,0 +1 @@
> +main.py.in
> \ No newline at end of file
> diff --git a/scripts/pybootchartgui/pybootchartgui/main.py.in
> b/scripts/pybootchartgui/pybootchartgui/main.py.in new file mode
> 100644 index 0000000..a954b12
> --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui/main.py.in
> @@ -0,0 +1,183 @@
> +#
> +#
> ***********************************************************************
> +#  Warning: This file is auto-generated from main.py.in - edit it
> there. +#
> ***********************************************************************
> +# +#  pybootchartgui is free software: you can redistribute it
> and/or modify +#  it under the terms of the GNU General Public
> License as published by +#  the Free Software Foundation, either
> version 3 of the License, or +#  (at your option) any later version. +
> +#  pybootchartgui is distributed in the hope that it will be useful,
> +#  but WITHOUT ANY WARRANTY; without even the implied warranty of
> +#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +#  GNU General Public License for more details.
> +
> +#  You should have received a copy of the GNU General Public License
> +#  along with pybootchartgui. If not, see
> <http://www.gnu.org/licenses/>. +
> +import sys
> +import os
> +import optparse
> +
> +from . import parsing
> +from . import batch
> +
> +def _mk_options_parser():
> +	"""Make an options parser."""
> +	usage = "%prog [options]
> /path/to/tmp/buildstats/<recipe-machine>/<BUILDNAME>/"
> +	version = "%prog v1.0.0"
> +	parser = optparse.OptionParser(usage, version=version)
> +	parser.add_option("-i", "--interactive",
> action="store_true", dest="interactive", default=False, 
> +			  help="start in active mode")
> +	parser.add_option("-f", "--format", dest="format",
> default="png", choices=["png", "svg", "pdf"],
> +			  help="image format (png, svg, pdf);
> default format png")
> +	parser.add_option("-o", "--output", dest="output",
> metavar="PATH", default=None,
> +			  help="output path (file or directory)
> where charts are stored")
> +	parser.add_option("-s", "--split", dest="num", type=int,
> default=1,
> +			  help="split the output chart into <NUM>
> charts, only works with \"-o PATH\"")
> +	parser.add_option("-m", "--mintime", dest="mintime",
> type=int, default=8,
> +			  help="only tasks longer than this time
> will be displayed")
> +	parser.add_option("-M", "--minutes", action="store_true",
> dest="as_minutes", default=False,
> +			  help="display time in minutes instead of
> seconds") +#	parser.add_option("-n", "--no-prune",
> action="store_false", dest="prune", default=True, +#
> 	  help="do not prune the process tree")
> +	parser.add_option("-q", "--quiet", action="store_true",
> dest="quiet", default=False,
> +			  help="suppress informational messages")
> +#	parser.add_option("-t", "--boot-time", action="store_true",
> dest="boottime", default=False, +#
> help="only display the boot time of the boot in text format
> (stdout)")	
> +	parser.add_option("--very-quiet", action="store_true",
> dest="veryquiet", default=False,
> +			  help="suppress all messages except errors")
> +	parser.add_option("--verbose", action="store_true",
> dest="verbose", default=False,
> +			  help="print all messages")
> +#	parser.add_option("--profile", action="store_true",
> dest="profile", default=False, +#
> help="profile rendering of chart (only useful when in batch mode
> indicated by -f)") +#	parser.add_option("--show-pid",
> action="store_true", dest="show_pid", default=False, +#
> 		  help="show process ids in the bootchart as
> 'processname [pid]'")
> +	parser.add_option("--show-all", action="store_true",
> dest="show_all", default=False,
> +			  help="show all processes in the chart")
> +#	parser.add_option("--crop-after", dest="crop_after",
> metavar="PROCESS", default=None, +#
> help="crop chart when idle after PROCESS is started") +#
> parser.add_option("--annotate", action="append", dest="annotate",
> metavar="PROCESS", default=None, +#
> help="annotate position where PROCESS is started; can be specified
> multiple times. " + +#			       "To create a
> single annotation when any one of a set of processes is started, use
> commas to separate the names") +#
> parser.add_option("--annotate-file", dest="annotate_file",
> metavar="FILENAME", default=None, +#
> help="filename to write annotation points to")
> +	parser.add_option("-T", "--full-time", action="store_true",
> dest="full_time", default=False,
> +			  help="display the full time regardless of
> which processes are currently shown")
> +	return parser
> +
> +class Writer:
> +	def __init__(self, write, options):
> +		self.write = write
> +		self.options = options
> +		
> +	def error(self, msg):
> +		self.write(msg)
> +
> +	def warn(self, msg):
> +		if not self.options.quiet:
> +			self.write(msg)
> +
> +	def info(self, msg):
> +		if self.options.verbose:
> +			self.write(msg)
> +
> +	def status(self, msg):
> +		if not self.options.quiet:
> +			self.write(msg)
> +
> +def _mk_writer(options):
> +	def write(s):
> +		print(s)
> +	return Writer(write, options)
> +	
> +def _get_filename(path):
> +	"""Construct a usable filename for outputs"""
> +	dname = "."
> +	fname = "bootchart"
> +	if path != None:
> +		if os.path.isdir(path):
> +			dname = path
> +		else:
> +			fname = path
> +	return os.path.join(dname, fname)
> +
> +def main(argv=None):
> +	try:
> +		if argv is None:
> +			argv = sys.argv[1:]
> +	
> +		parser = _mk_options_parser()
> +		options, args = parser.parse_args(argv)
> +
> +		# Default values for disabled options
> +		options.prune = True
> +		options.boottime = False
> +		options.profile = False
> +		options.show_pid = False
> +		options.crop_after = None
> +		options.annotate = None
> +		options.annotate_file = None
> +
> +		writer = _mk_writer(options)
> +
> +		if len(args) == 0:
> +			print("No path given, trying
> /var/log/bootchart.tgz")
> +			args = [ "/var/log/bootchart.tgz" ]
> +
> +		res = parsing.Trace(writer, args, options)
> +
> +		if options.interactive or options.output == None:
> +			from . import gui
> +			gui.show(res, options)
> +		elif options.boottime:
> +			import math
> +			proc_tree = res.proc_tree
> +			if proc_tree.idle:
> +			    duration = proc_tree.idle
> +			else:
> +			    duration = proc_tree.duration
> +			dur = duration / 100.0
> +			print('%02d:%05.2f' % (math.floor(dur/60),
> dur - 60 * math.floor(dur/60)))
> +		else:
> +			if options.annotate_file:
> +				f = open (options.annotate_file, "w")
> +				try:
> +					for time in res[4]:
> +						if time is not None:
> +							# output as
> ms
> +							f.write(time
> * 10)
> +				finally:
> +					f.close()
> +			filename = _get_filename(options.output)
> +			res_list = parsing.split_res(res, options)
> +			n = 1
> +			width = len(str(len(res_list)))
> +			s = "_%%0%dd." % width
> +			for r in res_list:
> +				if len(res_list) == 1:
> +					f = filename + "." +
> options.format
> +				else:
> +					f = filename + s % n +
> options.format
> +					n = n + 1
> +				def render():
> +					batch.render(writer, r,
> options, f)
> +				if options.profile:
> +					import cProfile
> +					import pstats
> +					profile = '%s.prof' %
> os.path.splitext(filename)[0]
> +					cProfile.runctx('render()',
> globals(), locals(), profile)
> +					p = pstats.Stats(profile)
> +
> p.strip_dirs().sort_stats('time').print_stats(20)
> +				else:
> +					render()
> +
> +		return 0
> +	except parsing.ParseError as ex:
> +		print(("Parse error: %s" % ex))
> +		return 2
> +
> +
> +if __name__ == '__main__':
> +	sys.exit(main())
> diff --git a/scripts/pybootchartgui/pybootchartgui/parsing.py
> b/scripts/pybootchartgui/pybootchartgui/parsing.py new file mode
> 100644 index 0000000..b42dac6
> --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui/parsing.py
> @@ -0,0 +1,821 @@
> +#  This file is part of pybootchartgui.
> +
> +#  pybootchartgui is free software: you can redistribute it and/or
> modify +#  it under the terms of the GNU General Public License as
> published by +#  the Free Software Foundation, either version 3 of
> the License, or +#  (at your option) any later version.
> +
> +#  pybootchartgui is distributed in the hope that it will be useful,
> +#  but WITHOUT ANY WARRANTY; without even the implied warranty of
> +#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +#  GNU General Public License for more details.
> +
> +#  You should have received a copy of the GNU General Public License
> +#  along with pybootchartgui. If not, see
> <http://www.gnu.org/licenses/>. +
> +import os
> +import string
> +import re
> +import sys
> +import tarfile
> +import time
> +from collections import defaultdict
> +from functools import reduce
> +
> +from .samples import *
> +from .process_tree import ProcessTree
> +
> +if sys.version_info >= (3, 0):
> +    long = int
> +
> +# Parsing produces as its end result a 'Trace'
> +
> +class Trace:
> +    def __init__(self, writer, paths, options):
> +        self.processes = {}
> +        self.start = {}
> +        self.end = {}
> +        self.min = None
> +        self.max = None
> +        self.headers = None
> +        self.disk_stats =  []
> +        self.ps_stats = None
> +        self.taskstats = None
> +        self.cpu_stats = []
> +        self.cmdline = None
> +        self.kernel = None
> +        self.kernel_tree = None
> +        self.filename = None
> +        self.parent_map = None
> +        self.mem_stats = []
> +        self.monitor_disk = None
> +        self.times = [] # Always empty, but expected by draw.py when
> drawing system charts. +
> +        if len(paths):
> +            parse_paths (writer, self, paths)
> +            if not self.valid():
> +                raise ParseError("empty state: '%s' does not contain
> a valid bootchart" % ", ".join(paths)) +
> +            if options.full_time:
> +                self.min = min(self.start.keys())
> +                self.max = max(self.end.keys())
> +
> +
> +        # Rendering system charts depends on start and end
> +        # time. Provide them where the original drawing code expects
> +        # them, i.e. in proc_tree.
> +        class BitbakeProcessTree:
> +            def __init__(self, start_time, end_time):
> +                self.start_time = start_time
> +                self.end_time = end_time
> +                self.duration = self.end_time - self.start_time
> +        self.proc_tree = BitbakeProcessTree(min(self.start.keys()),
> +                                            max(self.end.keys()))
> +
> +
> +        return
> +
> +        # Turn that parsed information into something more useful
> +        # link processes into a tree of pointers, calculate
> statistics
> +        self.compile(writer)
> +
> +        # Crop the chart to the end of the first idle period after
> the given
> +        # process
> +        if options.crop_after:
> +            idle = self.crop (writer, options.crop_after)
> +        else:
> +            idle = None
> +
> +        # Annotate other times as the first start point of given
> process lists
> +        self.times = [ idle ]
> +        if options.annotate:
> +            for procnames in options.annotate:
> +                names = [x[:15] for x in procnames.split(",")]
> +                for proc in self.ps_stats.process_map.values():
> +                    if proc.cmd in names:
> +                        self.times.append(proc.start_time)
> +                        break
> +                    else:
> +                        self.times.append(None)
> +
> +        self.proc_tree = ProcessTree(writer, self.kernel,
> self.ps_stats,
> +                                     self.ps_stats.sample_period,
> +
> self.headers.get("profile.process"),
> +                                     options.prune, idle,
> self.taskstats,
> +                                     self.parent_map is not None)
> +
> +        if self.kernel is not None:
> +            self.kernel_tree = ProcessTree(writer, self.kernel,
> None, 0,
> +
> self.headers.get("profile.process"),
> +                                           False, None, None, True)
> +
> +    def valid(self):
> +        return len(self.processes) != 0
> +        return self.headers != None and self.disk_stats != None and \
> +               self.ps_stats != None and self.cpu_stats != None
> +
> +    def add_process(self, process, start, end):
> +        self.processes[process] = [start, end]
> +        if start not in self.start:
> +            self.start[start] = []
> +        if process not in self.start[start]:
> +            self.start[start].append(process)
> +        if end not in self.end:
> +            self.end[end] = []
> +        if process not in self.end[end]:
> +            self.end[end].append(process)
> +
> +    def compile(self, writer):
> +
> +        def find_parent_id_for(pid):
> +            if pid is 0:
> +                return 0
> +            ppid = self.parent_map.get(pid)
> +            if ppid:
> +                # many of these double forks are so short lived
> +                # that we have no samples, or process info for them
> +                # so climb the parent hierarcy to find one
> +                if int (ppid * 1000) not in
> self.ps_stats.process_map: +#                    print "Pid '%d'
> short lived with no process" % ppid
> +                    ppid = find_parent_id_for (ppid)
> +#                else:
> +#                    print "Pid '%d' has an entry" % ppid
> +            else:
> +#                print "Pid '%d' missing from pid map" % pid
> +                return 0
> +            return ppid
> +
> +        # merge in the cmdline data
> +        if self.cmdline is not None:
> +            for proc in self.ps_stats.process_map.values():
> +                rpid = int (proc.pid // 1000)
> +                if rpid in self.cmdline:
> +                    cmd = self.cmdline[rpid]
> +                    proc.exe = cmd['exe']
> +                    proc.args = cmd['args']
> +#                else:
> +#                    print "proc %d '%s' not in cmdline" % (rpid,
> proc.exe) +
> +        # re-parent any stray orphans if we can
> +        if self.parent_map is not None:
> +            for process in self.ps_stats.process_map.values():
> +                ppid = find_parent_id_for (int(process.pid // 1000))
> +                if ppid:
> +                    process.ppid = ppid * 1000
> +
> +        # stitch the tree together with pointers
> +        for process in self.ps_stats.process_map.values():
> +            process.set_parent (self.ps_stats.process_map)
> +
> +        # count on fingers variously
> +        for process in self.ps_stats.process_map.values():
> +            process.calc_stats (self.ps_stats.sample_period)
> +
> +    def crop(self, writer, crop_after):
> +
> +        def is_idle_at(util, start, j):
> +            k = j + 1
> +            while k < len(util) and util[k][0] < start + 300:
> +                k += 1
> +            k = min(k, len(util)-1)
> +
> +            if util[j][1] >= 0.25:
> +                return False
> +
> +            avgload = sum(u[1] for u in util[j:k+1]) / (k-j+1)
> +            if avgload < 0.25:
> +                return True
> +            else:
> +                return False
> +        def is_idle(util, start):
> +            for j in range(0, len(util)):
> +                if util[j][0] < start:
> +                    continue
> +                return is_idle_at(util, start, j)
> +            else:
> +                return False
> +
> +        names = [x[:15] for x in crop_after.split(",")]
> +        for proc in self.ps_stats.process_map.values():
> +            if proc.cmd in names or proc.exe in names:
> +                writer.info("selected proc '%s' from list (start %d)"
> +                            % (proc.cmd, proc.start_time))
> +                break
> +        if proc is None:
> +            writer.warn("no selected crop proc '%s' in list" %
> crop_after) +
> +
> +        cpu_util = [(sample.time, sample.user + sample.sys +
> sample.io) for sample in self.cpu_stats]
> +        disk_util = [(sample.time, sample.util) for sample in
> self.disk_stats] +
> +        idle = None
> +        for i in range(0, len(cpu_util)):
> +            if cpu_util[i][0] < proc.start_time:
> +                continue
> +            if is_idle_at(cpu_util, cpu_util[i][0], i) \
> +               and is_idle(disk_util, cpu_util[i][0]):
> +                idle = cpu_util[i][0]
> +                break
> +
> +        if idle is None:
> +            writer.warn ("not idle after proc '%s'" % crop_after)
> +            return None
> +
> +        crop_at = idle + 300
> +        writer.info ("cropping at time %d" % crop_at)
> +        while len (self.cpu_stats) \
> +                    and self.cpu_stats[-1].time > crop_at:
> +            self.cpu_stats.pop()
> +        while len (self.disk_stats) \
> +                    and self.disk_stats[-1].time > crop_at:
> +            self.disk_stats.pop()
> +
> +        self.ps_stats.end_time = crop_at
> +
> +        cropped_map = {}
> +        for key, value in self.ps_stats.process_map.items():
> +            if (value.start_time <= crop_at):
> +                cropped_map[key] = value
> +
> +        for proc in cropped_map.values():
> +            proc.duration = min (proc.duration, crop_at -
> proc.start_time)
> +            while len (proc.samples) \
> +                        and proc.samples[-1].time > crop_at:
> +                proc.samples.pop()
> +
> +        self.ps_stats.process_map = cropped_map
> +
> +        return idle
> +
> +
> +
> +class ParseError(Exception):
> +    """Represents errors during parse of the bootchart."""
> +    def __init__(self, value):
> +        self.value = value
> +
> +    def __str__(self):
> +        return self.value
> +
> +def _parse_headers(file):
> +    """Parses the headers of the bootchart."""
> +    def parse(acc, line):
> +        (headers, last) = acc
> +        if '=' in line:
> +            last, value = map (lambda x: x.strip(), line.split('=',
> 1))
> +        else:
> +            value = line.strip()
> +        headers[last] += value
> +        return headers, last
> +    return reduce(parse, file.read().split('\n'),
> (defaultdict(str),''))[0] +
> +def _parse_timed_blocks(file):
> +    """Parses (ie., splits) a file into so-called timed-blocks. A
> +    timed-block consists of a timestamp on a line by itself followed
> +    by zero or more lines of data for that point in time."""
> +    def parse(block):
> +        lines = block.split('\n')
> +        if not lines:
> +            raise ParseError('expected a timed-block consisting a
> timestamp followed by data lines')
> +        try:
> +            return (int(lines[0]), lines[1:])
> +        except ValueError:
> +            raise ParseError("expected a timed-block, but timestamp
> '%s' is not an integer" % lines[0])
> +    blocks = file.read().split('\n\n')
> +    return [parse(block) for block in blocks if block.strip() and
> not block.endswith(' not running\n')] +
> +def _parse_proc_ps_log(writer, file):
> +    """
> +     * See proc(5) for details.
> +     *
> +     * {pid, comm, state, ppid, pgrp, session, tty_nr, tpgid, flags,
> minflt, cminflt, majflt, cmajflt, utime, stime,
> +     *  cutime, cstime, priority, nice, 0, itrealvalue, starttime,
> vsize, rss, rlim, startcode, endcode, startstack,
> +     *  kstkesp, kstkeip}
> +    """
> +    processMap = {}
> +    ltime = 0
> +    timed_blocks = _parse_timed_blocks(file)
> +    for time, lines in timed_blocks:
> +        for line in lines:
> +            if not line: continue
> +            tokens = line.split(' ')
> +            if len(tokens) < 21:
> +                continue
> +
> +            offset = [index for index, token in
> enumerate(tokens[1:]) if token[-1] == ')'][0]
> +            pid, cmd, state, ppid = int(tokens[0]), '
> '.join(tokens[1:2+offset]), tokens[2+offset], int(tokens[3+offset])
> +            userCpu, sysCpu, stime = int(tokens[13+offset]),
> int(tokens[14+offset]), int(tokens[21+offset]) +
> +            # magic fixed point-ness ...
> +            pid *= 1000
> +            ppid *= 1000
> +            if pid in processMap:
> +                process = processMap[pid]
> +                process.cmd = cmd.strip('()') # why rename after
> latest name??
> +            else:
> +                process = Process(writer, pid, cmd.strip('()'),
> ppid, min(time, stime))
> +                processMap[pid] = process
> +
> +            if process.last_user_cpu_time is not None and
> process.last_sys_cpu_time is not None and ltime is not None:
> +                userCpuLoad, sysCpuLoad = process.calc_load(userCpu,
> sysCpu, max(1, time - ltime))
> +                cpuSample = CPUSample('null', userCpuLoad,
> sysCpuLoad, 0.0)
> +                process.samples.append(ProcessSample(time, state,
> cpuSample)) +
> +            process.last_user_cpu_time = userCpu
> +            process.last_sys_cpu_time = sysCpu
> +        ltime = time
> +
> +    if len (timed_blocks) < 2:
> +        return None
> +
> +    startTime = timed_blocks[0][0]
> +    avgSampleLength = (ltime - startTime)/(len (timed_blocks) - 1)
> +
> +    return ProcessStats (writer, processMap, len (timed_blocks),
> avgSampleLength, startTime, ltime) +
> +def _parse_taskstats_log(writer, file):
> +    """
> +     * See bootchart-collector.c for details.
> +     *
> +     * { pid, ppid, comm, cpu_run_real_total, blkio_delay_total,
> swapin_delay_total }
> +     *
> +    """
> +    processMap = {}
> +    pidRewrites = {}
> +    ltime = None
> +    timed_blocks = _parse_timed_blocks(file)
> +    for time, lines in timed_blocks:
> +        # we have no 'stime' from taskstats, so prep 'init'
> +        if ltime is None:
> +            process = Process(writer, 1, '[init]', 0, 0)
> +            processMap[1000] = process
> +            ltime = time
> +#                       continue
> +        for line in lines:
> +            if not line: continue
> +            tokens = line.split(' ')
> +            if len(tokens) != 6:
> +                continue
> +
> +            opid, ppid, cmd = int(tokens[0]), int(tokens[1]),
> tokens[2]
> +            cpu_ns, blkio_delay_ns, swapin_delay_ns =
> long(tokens[-3]), long(tokens[-2]), long(tokens[-1]), +
> +            # make space for trees of pids
> +            opid *= 1000
> +            ppid *= 1000
> +
> +            # when the process name changes, we re-write the pid.
> +            if opid in pidRewrites:
> +                pid = pidRewrites[opid]
> +            else:
> +                pid = opid
> +
> +            cmd = cmd.strip('(').strip(')')
> +            if pid in processMap:
> +                process = processMap[pid]
> +                if process.cmd != cmd:
> +                    pid += 1
> +                    pidRewrites[opid] = pid
> +#                                       print "process mutation !
> '%s' vs '%s' pid %s -> pid %s\n" % (process.cmd, cmd, opid, pid)
> +                    process = process.split (writer, pid, cmd, ppid,
> time)
> +                    processMap[pid] = process
> +                else:
> +                    process.cmd = cmd;
> +            else:
> +                process = Process(writer, pid, cmd, ppid, time)
> +                processMap[pid] = process
> +
> +            delta_cpu_ns = (float) (cpu_ns - process.last_cpu_ns)
> +            delta_blkio_delay_ns = (float) (blkio_delay_ns -
> process.last_blkio_delay_ns)
> +            delta_swapin_delay_ns = (float) (swapin_delay_ns -
> process.last_swapin_delay_ns) +
> +            # make up some state data ...
> +            if delta_cpu_ns > 0:
> +                state = "R"
> +            elif delta_blkio_delay_ns + delta_swapin_delay_ns > 0:
> +                state = "D"
> +            else:
> +                state = "S"
> +
> +            # retain the ns timing information into a CPUSample -
> that tries
> +            # with the old-style to be a %age of CPU used in this
> time-slice.
> +            if delta_cpu_ns + delta_blkio_delay_ns +
> delta_swapin_delay_ns > 0: +#                               print
> "proc %s cpu_ns %g delta_cpu %g" % (cmd, cpu_ns, delta_cpu_ns)
> +                cpuSample = CPUSample('null', delta_cpu_ns, 0.0,
> +                                      delta_blkio_delay_ns,
> +                                      delta_swapin_delay_ns)
> +                process.samples.append(ProcessSample(time, state,
> cpuSample)) +
> +            process.last_cpu_ns = cpu_ns
> +            process.last_blkio_delay_ns = blkio_delay_ns
> +            process.last_swapin_delay_ns = swapin_delay_ns
> +        ltime = time
> +
> +    if len (timed_blocks) < 2:
> +        return None
> +
> +    startTime = timed_blocks[0][0]
> +    avgSampleLength = (ltime - startTime)/(len(timed_blocks)-1)
> +
> +    return ProcessStats (writer, processMap, len (timed_blocks),
> avgSampleLength, startTime, ltime) +
> +def _parse_proc_stat_log(file):
> +    samples = []
> +    ltimes = None
> +    for time, lines in _parse_timed_blocks(file):
> +        # skip emtpy lines
> +        if not lines:
> +            continue
> +        # CPU times {user, nice, system, idle, io_wait, irq, softirq}
> +        tokens = lines[0].split()
> +        times = [ int(token) for token in tokens[1:] ]
> +        if ltimes:
> +            user = float((times[0] + times[1]) - (ltimes[0] +
> ltimes[1]))
> +            system = float((times[2] + times[5] + times[6]) -
> (ltimes[2] + ltimes[5] + ltimes[6]))
> +            idle = float(times[3] - ltimes[3])
> +            iowait = float(times[4] - ltimes[4])
> +
> +            aSum = max(user + system + idle + iowait, 1)
> +            samples.append( CPUSample(time, user/aSum, system/aSum,
> iowait/aSum) ) +
> +        ltimes = times
> +        # skip the rest of statistics lines
> +    return samples
> +
> +def _parse_reduced_log(file, sample_class):
> +    samples = []
> +    for time, lines in _parse_timed_blocks(file):
> +        samples.append(sample_class(time, *[float(x) for x in
> lines[0].split()]))
> +    return samples
> +
> +def _parse_proc_disk_stat_log(file):
> +    """
> +    Parse file for disk stats, but only look at the whole device,
> eg. sda,
> +    not sda1, sda2 etc. The format of relevant lines should be:
> +    {major minor name rio rmerge rsect ruse wio wmerge wsect wuse
> running use aveq}
> +    """
> +    disk_regex_re = re.compile
> ('^([hsv]d.|mtdblock\d|mmcblk\d|cciss/c\d+d\d+.*)$') +
> +    # this gets called an awful lot.
> +    def is_relevant_line(linetokens):
> +        if len(linetokens) != 14:
> +            return False
> +        disk = linetokens[2]
> +        return disk_regex_re.match(disk)
> +
> +    disk_stat_samples = []
> +
> +    for time, lines in _parse_timed_blocks(file):
> +        sample = DiskStatSample(time)
> +        relevant_tokens = [linetokens for linetokens in map (lambda
> x: x.split(),lines) if is_relevant_line(linetokens)] +
> +        for tokens in relevant_tokens:
> +            disk, rsect, wsect, use = tokens[2], int(tokens[5]),
> int(tokens[9]), int(tokens[12])
> +            sample.add_diskdata([rsect, wsect, use])
> +
> +        disk_stat_samples.append(sample)
> +
> +    disk_stats = []
> +    for sample1, sample2 in zip(disk_stat_samples[:-1],
> disk_stat_samples[1:]):
> +        interval = sample1.time - sample2.time
> +        if interval == 0:
> +            interval = 1
> +        sums = [ a - b for a, b in zip(sample1.diskdata,
> sample2.diskdata) ]
> +        readTput = sums[0] / 2.0 * 100.0 / interval
> +        writeTput = sums[1] / 2.0 * 100.0 / interval
> +        util = float( sums[2] ) / 10 / interval
> +        util = max(0.0, min(1.0, util))
> +        disk_stats.append(DiskSample(sample2.time, readTput,
> writeTput, util)) +
> +    return disk_stats
> +
> +def _parse_reduced_proc_meminfo_log(file):
> +    """
> +    Parse file for global memory statistics with
> +    'MemTotal', 'MemFree', 'Buffers', 'Cached', 'SwapTotal',
> 'SwapFree' values
> +    (in that order) directly stored on one line.
> +    """
> +    used_values = ('MemTotal', 'MemFree', 'Buffers', 'Cached',
> 'SwapTotal', 'SwapFree',) +
> +    mem_stats = []
> +    for time, lines in _parse_timed_blocks(file):
> +        sample = MemSample(time)
> +        for name, value in zip(used_values, lines[0].split()):
> +            sample.add_value(name, int(value))
> +
> +        if sample.valid():
> +            mem_stats.append(DrawMemSample(sample))
> +
> +    return mem_stats
> +
> +def _parse_proc_meminfo_log(file):
> +    """
> +    Parse file for global memory statistics.
> +    The format of relevant lines should be: ^key: value( unit)?
> +    """
> +    used_values = ('MemTotal', 'MemFree', 'Buffers', 'Cached',
> 'SwapTotal', 'SwapFree',) +
> +    mem_stats = []
> +    meminfo_re = re.compile(r'([^ \t:]+):\s*(\d+).*')
> +
> +    for time, lines in _parse_timed_blocks(file):
> +        sample = MemSample(time)
> +
> +        for line in lines:
> +            match = meminfo_re.match(line)
> +            if not match:
> +                raise ParseError("Invalid meminfo line \"%s\"" %
> line)
> +            sample.add_value(match.group(1), int(match.group(2)))
> +
> +        if sample.valid():
> +            mem_stats.append(DrawMemSample(sample))
> +
> +    return mem_stats
> +
> +def _parse_monitor_disk_log(file):
> +    """
> +    Parse file with information about amount of diskspace used.
> +    The format of relevant lines should be: ^volume path:
> number-of-bytes?
> +    """
> +    disk_stats = []
> +    diskinfo_re = re.compile(r'^(.+):\s*(\d+)$')
> +
> +    for time, lines in _parse_timed_blocks(file):
> +        sample = DiskSpaceSample(time)
> +
> +        for line in lines:
> +            match = diskinfo_re.match(line)
> +            if not match:
> +                raise ParseError("Invalid monitor_disk line \"%s\""
> % line)
> +            sample.add_value(match.group(1), int(match.group(2)))
> +
> +        if sample.valid():
> +            disk_stats.append(sample)
> +
> +    return disk_stats
> +
> +
> +# if we boot the kernel with: initcall_debug printk.time=1 we can
> +# get all manner of interesting data from the dmesg output
> +# We turn this into a pseudo-process tree: each event is
> +# characterised by a
> +# we don't try to detect a "kernel finished" state - since the kernel
> +# continues to do interesting things after init is called.
> +#
> +# sample input:
> +# [    0.000000] ACPI: FACP 3f4fc000 000F4 (v04 INTEL  Napa
> 00000001 MSFT 01000013) +# ...
> +# [    0.039993] calling  migration_init+0x0/0x6b @ 1
> +# [    0.039993] initcall migration_init+0x0/0x6b returned 1 after 0
> usecs +def _parse_dmesg(writer, file):
> +    timestamp_re = re.compile ("^\[\s*(\d+\.\d+)\s*]\s+(.*)$")
> +    split_re = re.compile ("^(\S+)\s+([\S\+_-]+) (.*)$")
> +    processMap = {}
> +    idx = 0
> +    inc = 1.0 / 1000000
> +    kernel = Process(writer, idx, "k-boot", 0, 0.1)
> +    processMap['k-boot'] = kernel
> +    base_ts = False
> +    max_ts = 0
> +    for line in file.read().split('\n'):
> +        t = timestamp_re.match (line)
> +        if t is None:
> +#                       print "duff timestamp " + line
> +            continue
> +
> +        time_ms = float (t.group(1)) * 1000
> +        # looks like we may have a huge diff after the clock
> +        # has been set up. This could lead to huge graph:
> +        # so huge we will be killed by the OOM.
> +        # So instead of using the plain timestamp we will
> +        # use a delta to first one and skip the first one
> +        # for convenience
> +        if max_ts == 0 and not base_ts and time_ms > 1000:
> +            base_ts = time_ms
> +            continue
> +        max_ts = max(time_ms, max_ts)
> +        if base_ts:
> +#                       print "fscked clock: used %f instead of %f"
> % (time_ms - base_ts, time_ms)
> +            time_ms -= base_ts
> +        m = split_re.match (t.group(2))
> +
> +        if m is None:
> +            continue
> +#               print "match: '%s'" % (m.group(1))
> +        type = m.group(1)
> +        func = m.group(2)
> +        rest = m.group(3)
> +
> +        if t.group(2).startswith ('Write protecting the') or \
> +           t.group(2).startswith ('Freeing unused kernel memory'):
> +            kernel.duration = time_ms / 10
> +            continue
> +
> +#               print "foo: '%s' '%s' '%s'" % (type, func, rest)
> +        if type == "calling":
> +            ppid = kernel.pid
> +            p = re.match ("\@ (\d+)", rest)
> +            if p is not None:
> +                ppid = float (p.group(1)) // 1000
> +#                               print "match: '%s' ('%g') at '%s'" %
> (func, ppid, time_ms)
> +            name = func.split ('+', 1) [0]
> +            idx += inc
> +            processMap[func] = Process(writer, ppid + idx, name,
> ppid, time_ms / 10)
> +        elif type == "initcall":
> +#                       print "finished: '%s' at '%s'" % (func,
> time_ms)
> +            if func in processMap:
> +                process = processMap[func]
> +                process.duration = (time_ms / 10) -
> process.start_time
> +            else:
> +                print("corrupted init call for %s" % (func))
> +
> +        elif type == "async_waiting" or type == "async_continuing":
> +            continue # ignore
> +
> +    return processMap.values()
> +
> +#
> +# Parse binary pacct accounting file output if we have one
> +# cf. /usr/include/linux/acct.h
> +#
> +def _parse_pacct(writer, file):
> +    # read LE int32
> +    def _read_le_int32(file):
> +        byts = file.read(4)
> +        return (ord(byts[0]))       | (ord(byts[1]) << 8) | \
> +               (ord(byts[2]) << 16) | (ord(byts[3]) << 24)
> +
> +    parent_map = {}
> +    parent_map[0] = 0
> +    while file.read(1) != "": # ignore flags
> +        ver = file.read(1)
> +        if ord(ver) < 3:
> +            print("Invalid version 0x%x" % (ord(ver)))
> +            return None
> +
> +        file.seek (14, 1)     # user, group etc.
> +        pid = _read_le_int32 (file)
> +        ppid = _read_le_int32 (file)
> +#               print "Parent of %d is %d" % (pid, ppid)
> +        parent_map[pid] = ppid
> +        file.seek (4 + 4 + 16, 1) # timings
> +        file.seek (16, 1)         # acct_comm
> +    return parent_map
> +
> +def _parse_paternity_log(writer, file):
> +    parent_map = {}
> +    parent_map[0] = 0
> +    for line in file.read().split('\n'):
> +        if not line:
> +            continue
> +        elems = line.split(' ') # <Child> <Parent>
> +        if len (elems) >= 2:
> +#                       print "paternity of %d is %d" %
> (int(elems[0]), int(elems[1]))
> +            parent_map[int(elems[0])] = int(elems[1])
> +        else:
> +            print("Odd paternity line '%s'" % (line))
> +    return parent_map
> +
> +def _parse_cmdline_log(writer, file):
> +    cmdLines = {}
> +    for block in file.read().split('\n\n'):
> +        lines = block.split('\n')
> +        if len (lines) >= 3:
> +#                       print "Lines '%s'" % (lines[0])
> +            pid = int (lines[0])
> +            values = {}
> +            values['exe'] = lines[1].lstrip(':')
> +            args = lines[2].lstrip(':').split('\0')
> +            args.pop()
> +            values['args'] = args
> +            cmdLines[pid] = values
> +    return cmdLines
> +
> +def _parse_bitbake_buildstats(writer, state, filename, file):
> +    paths = filename.split("/")
> +    task = paths[-1]
> +    pn = paths[-2]
> +    start = None
> +    end = None
> +    for line in file:
> +        if line.startswith("Started:"):
> +            start = int(float(line.split()[-1]))
> +        elif line.startswith("Ended:"):
> +            end = int(float(line.split()[-1]))
> +    if start and end:
> +        state.add_process(pn + ":" + task, start, end)
> +
> +def get_num_cpus(headers):
> +    """Get the number of CPUs from the system.cpu header property.
> As the
> +    CPU utilization graphs are relative, the number of CPUs
> currently makes
> +    no difference."""
> +    if headers is None:
> +        return 1
> +    if headers.get("system.cpu.num"):
> +        return max (int (headers.get("system.cpu.num")), 1)
> +    cpu_model = headers.get("system.cpu")
> +    if cpu_model is None:
> +        return 1
> +    mat = re.match(".*\\((\\d+)\\)", cpu_model)
> +    if mat is None:
> +        return 1
> +    return max (int(mat.group(1)), 1)
> +
> +def _do_parse(writer, state, filename, file):
> +    writer.info("parsing '%s'" % filename)
> +    t1 = time.process_time()
> +    name = os.path.basename(filename)
> +    if name == "proc_diskstats.log":
> +        state.disk_stats = _parse_proc_disk_stat_log(file)
> +    elif name == "reduced_proc_diskstats.log":
> +        state.disk_stats = _parse_reduced_log(file, DiskSample)
> +    elif name == "proc_stat.log":
> +        state.cpu_stats = _parse_proc_stat_log(file)
> +    elif name == "reduced_proc_stat.log":
> +        state.cpu_stats = _parse_reduced_log(file, CPUSample)
> +    elif name == "proc_meminfo.log":
> +        state.mem_stats = _parse_proc_meminfo_log(file)
> +    elif name == "reduced_proc_meminfo.log":
> +        state.mem_stats = _parse_reduced_proc_meminfo_log(file)
> +    elif name == "cmdline2.log":
> +        state.cmdline = _parse_cmdline_log(writer, file)
> +    elif name == "monitor_disk.log":
> +        state.monitor_disk = _parse_monitor_disk_log(file)
> +    elif not filename.endswith('.log'):
> +        _parse_bitbake_buildstats(writer, state, filename, file)
> +    t2 = time.process_time()
> +    writer.info("  %s seconds" % str(t2-t1))
> +    return state
> +
> +def parse_file(writer, state, filename):
> +    if state.filename is None:
> +        state.filename = filename
> +    basename = os.path.basename(filename)
> +    with open(filename, "r") as file:
> +        return _do_parse(writer, state, filename, file)
> +
> +def parse_paths(writer, state, paths):
> +    for path in paths:
> +        if state.filename is None:
> +            state.filename = path
> +        root, extension = os.path.splitext(path)
> +        if not(os.path.exists(path)):
> +            writer.warn("warning: path '%s' does not exist,
> ignoring." % path)
> +            continue
> +        #state.filename = path
> +        if os.path.isdir(path):
> +            files = sorted([os.path.join(path, f) for f in
> os.listdir(path)])
> +            state = parse_paths(writer, state, files)
> +        elif extension in [".tar", ".tgz", ".gz"]:
> +            if extension == ".gz":
> +                root, extension = os.path.splitext(root)
> +                if extension != ".tar":
> +                    writer.warn("warning: can only handle zipped tar
> files, not zipped '%s'-files; ignoring" % extension)
> +                    continue
> +            tf = None
> +            try:
> +                writer.status("parsing '%s'" % path)
> +                tf = tarfile.open(path, 'r:*')
> +                for name in tf.getnames():
> +                    state = _do_parse(writer, state, name,
> tf.extractfile(name))
> +            except tarfile.ReadError as error:
> +                raise ParseError("error: could not read tarfile
> '%s': %s." % (path, error))
> +            finally:
> +                if tf != None:
> +                    tf.close()
> +        else:
> +            state = parse_file(writer, state, path)
> +    return state
> +
> +def split_res(res, options):
> +    """ Split the res into n pieces """
> +    res_list = []
> +    if options.num > 1:
> +        s_list = sorted(res.start.keys())
> +        frag_size = len(s_list) / float(options.num)
> +        # Need the top value
> +        if frag_size > int(frag_size):
> +            frag_size = int(frag_size + 1)
> +        else:
> +            frag_size = int(frag_size)
> +
> +        start = 0
> +        end = frag_size
> +        while start < end:
> +            state = Trace(None, [], None)
> +            if options.full_time:
> +                state.min = min(res.start.keys())
> +                state.max = max(res.end.keys())
> +            for i in range(start, end):
> +                # Add this line for reference
> +                #state.add_process(pn + ":" + task, start, end)
> +                for p in res.start[s_list[i]]:
> +                    state.add_process(p, s_list[i],
> res.processes[p][1])
> +            start = end
> +            end = end + frag_size
> +            if end > len(s_list):
> +                end = len(s_list)
> +            res_list.append(state)
> +    else:
> +        res_list.append(res)
> +    return res_list
> diff --git a/scripts/pybootchartgui/pybootchartgui/process_tree.py
> b/scripts/pybootchartgui/pybootchartgui/process_tree.py new file mode
> 100644 index 0000000..cf88110
> --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui/process_tree.py
> @@ -0,0 +1,292 @@
> +#  This file is part of pybootchartgui.
> +
> +#  pybootchartgui is free software: you can redistribute it and/or
> modify +#  it under the terms of the GNU General Public License as
> published by +#  the Free Software Foundation, either version 3 of
> the License, or +#  (at your option) any later version.
> +
> +#  pybootchartgui is distributed in the hope that it will be useful,
> +#  but WITHOUT ANY WARRANTY; without even the implied warranty of
> +#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +#  GNU General Public License for more details.
> +
> +#  You should have received a copy of the GNU General Public License
> +#  along with pybootchartgui. If not, see
> <http://www.gnu.org/licenses/>. +
> +class ProcessTree:
> +    """ProcessTree encapsulates a process tree.  The tree is built
> from log files
> +       retrieved during the boot process.  When building the process
> tree, it is
> +       pruned and merged in order to be able to visualize it in a
> comprehensible
> +       manner.
> +
> +       The following pruning techniques are used:
> +
> +        * idle processes that keep running during the last process
> sample
> +          (which is a heuristic for a background processes) are
> removed,
> +        * short-lived processes (i.e. processes that only live for
> the
> +          duration of two samples or less) are removed,
> +        * the processes used by the boot logger are removed,
> +        * exploders (i.e. processes that are known to spawn huge
> meaningless
> +          process subtrees) have their subtrees merged together,
> +        * siblings (i.e. processes with the same command line living
> +          concurrently -- thread heuristic) are merged together,
> +        * process runs (unary trees with processes sharing the
> command line)
> +          are merged together.
> +
> +    """
> +    LOGGER_PROC = 'bootchart-colle'
> +    EXPLODER_PROCESSES = set(['hwup'])
> +
> +    def __init__(self, writer, kernel, psstats, sample_period,
> +                 monitoredApp, prune, idle, taskstats,
> +                 accurate_parentage, for_testing = False):
> +        self.writer = writer
> +        self.process_tree = []
> +        self.taskstats = taskstats
> +        if psstats is None:
> +            process_list = kernel
> +        elif kernel is None:
> +            process_list = psstats.process_map.values()
> +        else:
> +            process_list = list(kernel) +
> list(psstats.process_map.values())
> +        self.process_list = sorted(process_list, key = lambda p:
> p.pid)
> +        self.sample_period = sample_period
> +
> +        self.build()
> +        if not accurate_parentage:
> +            self.update_ppids_for_daemons(self.process_list)
> +
> +        self.start_time = self.get_start_time(self.process_tree)
> +        self.end_time = self.get_end_time(self.process_tree)
> +        self.duration = self.end_time - self.start_time
> +        self.idle = idle
> +
> +        if for_testing:
> +            return
> +
> +        removed = self.merge_logger(self.process_tree,
> self.LOGGER_PROC, monitoredApp, False)
> +        writer.status("merged %i logger processes" % removed)
> +
> +        if prune:
> +            p_processes = self.prune(self.process_tree, None)
> +            p_exploders = self.merge_exploders(self.process_tree,
> self.EXPLODER_PROCESSES)
> +            p_threads = self.merge_siblings(self.process_tree)
> +            p_runs = self.merge_runs(self.process_tree)
> +            writer.status("pruned %i process, %i exploders, %i
> threads, and %i runs" % (p_processes, p_exploders, p_threads,
> p_runs)) +
> +        self.sort(self.process_tree)
> +
> +        self.start_time = self.get_start_time(self.process_tree)
> +        self.end_time = self.get_end_time(self.process_tree)
> +        self.duration = self.end_time - self.start_time
> +
> +        self.num_proc = self.num_nodes(self.process_tree)
> +
> +    def build(self):
> +        """Build the process tree from the list of top samples."""
> +        self.process_tree = []
> +        for proc in self.process_list:
> +            if not proc.parent:
> +                self.process_tree.append(proc)
> +            else:
> +                proc.parent.child_list.append(proc)
> +
> +    def sort(self, process_subtree):
> +        """Sort process tree."""
> +        for p in process_subtree:
> +            p.child_list.sort(key = lambda p: p.pid)
> +            self.sort(p.child_list)
> +
> +    def num_nodes(self, process_list):
> +        "Counts the number of nodes in the specified process tree."""
> +        nodes = 0
> +        for proc in process_list:
> +            nodes = nodes + self.num_nodes(proc.child_list)
> +        return nodes + len(process_list)
> +
> +    def get_start_time(self, process_subtree):
> +        """Returns the start time of the process subtree.  This is
> the start
> +           time of the earliest process.
> +
> +        """
> +        if not process_subtree:
> +            return 100000000
> +        return min( [min(proc.start_time,
> self.get_start_time(proc.child_list)) for proc in process_subtree] ) +
> +    def get_end_time(self, process_subtree):
> +        """Returns the end time of the process subtree.  This is the
> end time
> +           of the last collected sample.
> +
> +        """
> +        if not process_subtree:
> +            return -100000000
> +        return max( [max(proc.start_time + proc.duration,
> self.get_end_time(proc.child_list)) for proc in process_subtree] ) +
> +    def get_max_pid(self, process_subtree):
> +        """Returns the max PID found in the process tree."""
> +        if not process_subtree:
> +            return -100000000
> +        return max( [max(proc.pid,
> self.get_max_pid(proc.child_list)) for proc in process_subtree] ) +
> +    def update_ppids_for_daemons(self, process_list):
> +        """Fedora hack: when loading the system services from rc,
> runuser(1)
> +           is used.  This sets the PPID of all daemons to 1, skewing
> +           the process tree.  Try to detect this and set the PPID of
> +           these processes the PID of rc.
> +
> +        """
> +        rcstartpid = -1
> +        rcendpid = -1
> +        rcproc = None
> +        for p in process_list:
> +            if p.cmd == "rc" and p.ppid // 1000 == 1:
> +                rcproc = p
> +                rcstartpid = p.pid
> +                rcendpid = self.get_max_pid(p.child_list)
> +        if rcstartpid != -1 and rcendpid != -1:
> +            for p in process_list:
> +                if p.pid > rcstartpid and p.pid < rcendpid and
> p.ppid // 1000 == 1:
> +                    p.ppid = rcstartpid
> +                    p.parent = rcproc
> +            for p in process_list:
> +                p.child_list = []
> +            self.build()
> +
> +    def prune(self, process_subtree, parent):
> +        """Prunes the process tree by removing idle processes and
> processes
> +           that only live for the duration of a single top sample.
> Sibling
> +           processes with the same command line (i.e. threads) are
> merged
> +           together. This filters out sleepy background processes,
> short-lived
> +           processes and bootcharts' analysis tools.
> +        """
> +        def is_idle_background_process_without_children(p):
> +            process_end = p.start_time + p.duration
> +            return not p.active and \
> +                   process_end >= self.start_time + self.duration
> and \
> +                   p.start_time > self.start_time and \
> +                   p.duration > 0.9 * self.duration and \
> +                   self.num_nodes(p.child_list) == 0
> +
> +        num_removed = 0
> +        idx = 0
> +        while idx < len(process_subtree):
> +            p = process_subtree[idx]
> +            if parent != None or len(p.child_list) == 0:
> +
> +                prune = False
> +                if is_idle_background_process_without_children(p):
> +                    prune = True
> +                elif p.duration <= 2 * self.sample_period:
> +                    # short-lived process
> +                    prune = True
> +
> +                if prune:
> +                    process_subtree.pop(idx)
> +                    for c in p.child_list:
> +                        process_subtree.insert(idx, c)
> +                    num_removed += 1
> +                    continue
> +                else:
> +                    num_removed += self.prune(p.child_list, p)
> +            else:
> +                num_removed += self.prune(p.child_list, p)
> +            idx += 1
> +
> +        return num_removed
> +
> +    def merge_logger(self, process_subtree, logger_proc,
> monitored_app, app_tree):
> +        """Merges the logger's process subtree.  The logger will
> typically
> +           spawn lots of sleep and cat processes, thus polluting the
> +           process tree.
> +
> +        """
> +        num_removed = 0
> +        for p in process_subtree:
> +            is_app_tree = app_tree
> +            if logger_proc == p.cmd and not app_tree:
> +                is_app_tree = True
> +                num_removed += self.merge_logger(p.child_list,
> logger_proc, monitored_app, is_app_tree)
> +                # don't remove the logger itself
> +                continue
> +
> +            if app_tree and monitored_app != None and monitored_app
> == p.cmd:
> +                is_app_tree = False
> +
> +            if is_app_tree:
> +                for child in p.child_list:
> +                    self.merge_processes(p, child)
> +                    num_removed += 1
> +                p.child_list = []
> +            else:
> +                num_removed += self.merge_logger(p.child_list,
> logger_proc, monitored_app, is_app_tree)
> +        return num_removed
> +
> +    def merge_exploders(self, process_subtree, processes):
> +        """Merges specific process subtrees (used for processes
> which usually
> +           spawn huge meaningless process trees).
> +
> +        """
> +        num_removed = 0
> +        for p in process_subtree:
> +            if processes in processes and len(p.child_list) > 0:
> +                subtreemap = self.getProcessMap(p.child_list)
> +                for child in subtreemap.values():
> +                    self.merge_processes(p, child)
> +                    num_removed += len(subtreemap)
> +                    p.child_list = []
> +                    p.cmd += " (+)"
> +            else:
> +                num_removed += self.merge_exploders(p.child_list,
> processes)
> +        return num_removed
> +
> +    def merge_siblings(self, process_subtree):
> +        """Merges thread processes.  Sibling processes with the same
> command
> +           line are merged together.
> +
> +        """
> +        num_removed = 0
> +        idx = 0
> +        while idx < len(process_subtree)-1:
> +            p = process_subtree[idx]
> +            nextp = process_subtree[idx+1]
> +            if nextp.cmd == p.cmd:
> +                process_subtree.pop(idx+1)
> +                idx -= 1
> +                num_removed += 1
> +                p.child_list.extend(nextp.child_list)
> +                self.merge_processes(p, nextp)
> +            num_removed += self.merge_siblings(p.child_list)
> +            idx += 1
> +        if len(process_subtree) > 0:
> +            p = process_subtree[-1]
> +            num_removed += self.merge_siblings(p.child_list)
> +        return num_removed
> +
> +    def merge_runs(self, process_subtree):
> +        """Merges process runs.  Single child processes which share
> the same
> +           command line with the parent are merged.
> +
> +        """
> +        num_removed = 0
> +        idx = 0
> +        while idx < len(process_subtree):
> +            p = process_subtree[idx]
> +            if len(p.child_list) == 1 and p.child_list[0].cmd ==
> p.cmd:
> +                child = p.child_list[0]
> +                p.child_list = list(child.child_list)
> +                self.merge_processes(p, child)
> +                num_removed += 1
> +                continue
> +            num_removed += self.merge_runs(p.child_list)
> +            idx += 1
> +        return num_removed
> +
> +    def merge_processes(self, p1, p2):
> +        """Merges two process' samples."""
> +        p1.samples.extend(p2.samples)
> +        p1.samples.sort( key = lambda p: p.time )
> +        p1time = p1.start_time
> +        p2time = p2.start_time
> +        p1.start_time = min(p1time, p2time)
> +        pendtime = max(p1time + p1.duration, p2time + p2.duration)
> +        p1.duration = pendtime - p1.start_time
> diff --git a/scripts/pybootchartgui/pybootchartgui/samples.py
> b/scripts/pybootchartgui/pybootchartgui/samples.py new file mode
> 100644 index 0000000..9fc309b
> --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui/samples.py
> @@ -0,0 +1,178 @@
> +#  This file is part of pybootchartgui.
> +
> +#  pybootchartgui is free software: you can redistribute it and/or
> modify +#  it under the terms of the GNU General Public License as
> published by +#  the Free Software Foundation, either version 3 of
> the License, or +#  (at your option) any later version.
> +
> +#  pybootchartgui is distributed in the hope that it will be useful,
> +#  but WITHOUT ANY WARRANTY; without even the implied warranty of
> +#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +#  GNU General Public License for more details.
> +
> +#  You should have received a copy of the GNU General Public License
> +#  along with pybootchartgui. If not, see
> <http://www.gnu.org/licenses/>. +
> +
> +class DiskStatSample:
> +    def __init__(self, time):
> +        self.time = time
> +        self.diskdata = [0, 0, 0]
> +    def add_diskdata(self, new_diskdata):
> +        self.diskdata = [ a + b for a, b in zip(self.diskdata,
> new_diskdata) ] +
> +class CPUSample:
> +    def __init__(self, time, user, sys, io = 0.0, swap = 0.0):
> +        self.time = time
> +        self.user = user
> +        self.sys = sys
> +        self.io = io
> +        self.swap = swap
> +
> +    @property
> +    def cpu(self):
> +        return self.user + self.sys
> +
> +    def __str__(self):
> +        return str(self.time) + "\t" + str(self.user) + "\t" + \
> +               str(self.sys) + "\t" + str(self.io) + "\t" + str
> (self.swap) +
> +class MemSample:
> +    used_values = ('MemTotal', 'MemFree', 'Buffers', 'Cached',
> 'SwapTotal', 'SwapFree',) +
> +    def __init__(self, time):
> +        self.time = time
> +        self.records = {}
> +
> +    def add_value(self, name, value):
> +        if name in MemSample.used_values:
> +            self.records[name] = value
> +
> +    def valid(self):
> +        keys = self.records.keys()
> +        # discard incomplete samples
> +        return [v for v in MemSample.used_values if v not in keys]
> == [] +
> +class DrawMemSample:
> +    """
> +    Condensed version of a MemSample with exactly the values used by
> the drawing code.
> +    Initialized either from a valid MemSample or
> +    a tuple/list of buffer/used/cached/swap values.
> +    """
> +    def __init__(self, mem_sample):
> +        self.time = mem_sample.time
> +        if isinstance(mem_sample, MemSample):
> +            self.buffers = mem_sample.records['MemTotal'] -
> mem_sample.records['MemFree']
> +            self.used = mem_sample.records['MemTotal'] -
> mem_sample.records['MemFree'] - mem_sample.records['Buffers']
> +            self.cached = mem_sample.records['Cached']
> +            self.swap = mem_sample.records['SwapTotal'] -
> mem_sample.records['SwapFree']
> +        else:
> +            self.buffers, self.used, self.cached, self.swap =
> mem_sample +
> +class DiskSpaceSample:
> +    def __init__(self, time):
> +        self.time = time
> +        self.records = {}
> +
> +    def add_value(self, name, value):
> +        self.records[name] = value
> +
> +    def valid(self):
> +        return bool(self.records)
> +
> +class ProcessSample:
> +    def __init__(self, time, state, cpu_sample):
> +        self.time = time
> +        self.state = state
> +        self.cpu_sample = cpu_sample
> +
> +    def __str__(self):
> +        return str(self.time) + "\t" + str(self.state) + "\t" +
> str(self.cpu_sample) +
> +class ProcessStats:
> +    def __init__(self, writer, process_map, sample_count,
> sample_period, start_time, end_time):
> +        self.process_map = process_map
> +        self.sample_count = sample_count
> +        self.sample_period = sample_period
> +        self.start_time = start_time
> +        self.end_time = end_time
> +        writer.info ("%d samples, avg. sample length %f" %
> (self.sample_count, self.sample_period))
> +        writer.info ("process list size: %d" % len
> (self.process_map.values())) +
> +class Process:
> +    def __init__(self, writer, pid, cmd, ppid, start_time):
> +        self.writer = writer
> +        self.pid = pid
> +        self.cmd = cmd
> +        self.exe = cmd
> +        self.args = []
> +        self.ppid = ppid
> +        self.start_time = start_time
> +        self.duration = 0
> +        self.samples = []
> +        self.parent = None
> +        self.child_list = []
> +
> +        self.active = None
> +        self.last_user_cpu_time = None
> +        self.last_sys_cpu_time = None
> +
> +        self.last_cpu_ns = 0
> +        self.last_blkio_delay_ns = 0
> +        self.last_swapin_delay_ns = 0
> +
> +    # split this process' run - triggered by a name change
> +    def split(self, writer, pid, cmd, ppid, start_time):
> +        split = Process (writer, pid, cmd, ppid, start_time)
> +
> +        split.last_cpu_ns = self.last_cpu_ns
> +        split.last_blkio_delay_ns = self.last_blkio_delay_ns
> +        split.last_swapin_delay_ns = self.last_swapin_delay_ns
> +
> +        return split
> +
> +    def __str__(self):
> +        return " ".join([str(self.pid), self.cmd, str(self.ppid), '[
> ' + str(len(self.samples)) + ' samples ]' ]) +
> +    def calc_stats(self, samplePeriod):
> +        if self.samples:
> +            firstSample = self.samples[0]
> +            lastSample = self.samples[-1]
> +            self.start_time = min(firstSample.time, self.start_time)
> +            self.duration = lastSample.time - self.start_time +
> samplePeriod +
> +        activeCount = sum( [1 for sample in self.samples if
> sample.cpu_sample and sample.cpu_sample.sys + sample.cpu_sample.user
> + sample.cpu_sample.io > 0.0] )
> +        activeCount = activeCount + sum( [1 for sample in
> self.samples if sample.state == 'D'] )
> +        self.active = (activeCount>2)
> +
> +    def calc_load(self, userCpu, sysCpu, interval):
> +        userCpuLoad = float(userCpu - self.last_user_cpu_time) /
> interval
> +        sysCpuLoad = float(sysCpu - self.last_sys_cpu_time) /
> interval
> +        cpuLoad = userCpuLoad + sysCpuLoad
> +        # normalize
> +        if cpuLoad > 1.0:
> +            userCpuLoad = userCpuLoad / cpuLoad
> +            sysCpuLoad = sysCpuLoad / cpuLoad
> +        return (userCpuLoad, sysCpuLoad)
> +
> +    def set_parent(self, processMap):
> +        if self.ppid != None:
> +            self.parent = processMap.get (self.ppid)
> +            if self.parent == None and self.pid // 1000 > 1 and \
> +                not (self.ppid == 2000 or self.pid == 2000): #
> kernel threads: ppid=2
> +                self.writer.warn("Missing CONFIG_PROC_EVENTS: no
> parent for pid '%i' ('%s') with ppid '%i'" \
> +                                 % (self.pid,self.cmd,self.ppid))
> +
> +    def get_end_time(self):
> +        return self.start_time + self.duration
> +
> +class DiskSample:
> +    def __init__(self, time, read, write, util):
> +        self.time = time
> +        self.read = read
> +        self.write = write
> +        self.util = util
> +        self.tput = read + write
> +
> +    def __str__(self):
> +        return "\t".join([str(self.time), str(self.read),
> str(self.write), str(self.util)]) diff --git
> a/scripts/pybootchartgui/pybootchartgui/tests/parser_test.py
> b/scripts/pybootchartgui/pybootchartgui/tests/parser_test.py new file
> mode 100644 index 0000000..00fb3bf --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui/tests/parser_test.py
> @@ -0,0 +1,105 @@
> +import sys, os, re, struct, operator, math
> +from collections import defaultdict
> +import unittest
> +
> +sys.path.insert(0, os.getcwd())
> +
> +import pybootchartgui.parsing as parsing
> +import pybootchartgui.main as main
> +
> +debug = False
> +
> +def floatEq(f1, f2):
> +	return math.fabs(f1-f2) < 0.00001
> +
> +bootchart_dir = os.path.join(os.path.dirname(sys.argv[0]),
> '../../examples/1/') +parser = main._mk_options_parser()
> +options, args = parser.parse_args(['--q', bootchart_dir])
> +writer = main._mk_writer(options)
> +
> +class TestBCParser(unittest.TestCase):
> +    
> +	def setUp(self):
> +		self.name = "My first unittest"
> +		self.rootdir = bootchart_dir
> +
> +	def mk_fname(self,f):
> +		return os.path.join(self.rootdir, f)
> +
> +	def testParseHeader(self):
> +		trace = parsing.Trace(writer, args, options)
> +		state = parsing.parse_file(writer, trace,
> self.mk_fname('header'))
> +		self.assertEqual(6, len(state.headers))
> +		self.assertEqual(2,
> parsing.get_num_cpus(state.headers)) +
> +	def test_parseTimedBlocks(self):
> +		trace = parsing.Trace(writer, args, options)
> +		state = parsing.parse_file(writer, trace,
> self.mk_fname('proc_diskstats.log'))
> +		self.assertEqual(141, len(state.disk_stats))
> +
> +	def testParseProcPsLog(self):
> +		trace = parsing.Trace(writer, args, options)
> +		state = parsing.parse_file(writer, trace,
> self.mk_fname('proc_ps.log'))
> +		samples = state.ps_stats
> +		processes = samples.process_map
> +		sorted_processes = [processes[k] for k in
> sorted(processes.keys())] +
> +		ps_data = open(self.mk_fname('extract2.proc_ps.log'))
> +		for index, line in enumerate(ps_data):
> +			tokens = line.split();
> +			process = sorted_processes[index]
> +			if debug:
> +				print(tokens[0:4])
> +				print(process.pid / 1000,
> process.cmd, process.ppid, len(process.samples))
> +				print('-------------------')
> +
> +			self.assertEqual(tokens[0], str(process.pid
> // 1000))
> +			self.assertEqual(tokens[1], str(process.cmd))
> +			self.assertEqual(tokens[2], str(process.ppid
> // 1000))
> +			self.assertEqual(tokens[3],
> str(len(process.samples)))
> +		ps_data.close()
> +
> +	def testparseProcDiskStatLog(self):
> +		trace = parsing.Trace(writer, args, options)
> +		state_with_headers = parsing.parse_file(writer,
> trace, self.mk_fname('header'))
> +		state_with_headers.headers['system.cpu'] = 'xxx (2)'
> +		samples = parsing.parse_file(writer,
> state_with_headers, self.mk_fname('proc_diskstats.log')).disk_stats
> +		self.assertEqual(141, len(samples))
> +
> +		diskstats_data =
> open(self.mk_fname('extract.proc_diskstats.log'))
> +		for index, line in enumerate(diskstats_data):
> +			tokens = line.split('\t')
> +			sample = samples[index]
> +			if debug:		
> +				print(line.rstrip())
> +				print(sample)
> +				print('-------------------')
> +			
> +			self.assertEqual(tokens[0], str(sample.time))
> +			self.assert_(floatEq(float(tokens[1]),
> sample.read))
> +			self.assert_(floatEq(float(tokens[2]),
> sample.write))
> +			self.assert_(floatEq(float(tokens[3]),
> sample.util))
> +		diskstats_data.close()
> +	
> +	def testparseProcStatLog(self):
> +		trace = parsing.Trace(writer, args, options)
> +		samples = parsing.parse_file(writer, trace,
> self.mk_fname('proc_stat.log')).cpu_stats
> +		self.assertEqual(141, len(samples))
> +
> +		stat_data =
> open(self.mk_fname('extract.proc_stat.log'))
> +		for index, line in enumerate(stat_data):
> +			tokens = line.split('\t')
> +			sample = samples[index]
> +			if debug:
> +				print(line.rstrip())
> +				print(sample)
> +				print('-------------------')
> +			self.assert_(floatEq(float(tokens[0]),
> sample.time))
> +			self.assert_(floatEq(float(tokens[1]),
> sample.user))
> +			self.assert_(floatEq(float(tokens[2]),
> sample.sys))
> +			self.assert_(floatEq(float(tokens[3]),
> sample.io))
> +		stat_data.close()
> +
> +if __name__ == '__main__':
> +    unittest.main()
> +
> diff --git
> a/scripts/pybootchartgui/pybootchartgui/tests/process_tree_test.py
> b/scripts/pybootchartgui/pybootchartgui/tests/process_tree_test.py
> new file mode 100644 index 0000000..6f46a1c --- /dev/null
> +++ b/scripts/pybootchartgui/pybootchartgui/tests/process_tree_test.py
> @@ -0,0 +1,92 @@
> +import sys
> +import os
> +import unittest
> +
> +sys.path.insert(0, os.getcwd())
> +
> +import pybootchartgui.parsing as parsing
> +import pybootchartgui.process_tree as process_tree
> +import pybootchartgui.main as main
> +
> +if sys.version_info >= (3, 0):
> +    long = int
> +
> +class TestProcessTree(unittest.TestCase):
> +
> +    def setUp(self):
> +        self.name = "Process tree unittest"
> +        self.rootdir = os.path.join(os.path.dirname(sys.argv[0]),
> '../../examples/1/') +
> +        parser = main._mk_options_parser()
> +        options, args = parser.parse_args(['--q', self.rootdir])
> +        writer = main._mk_writer(options)
> +        trace = parsing.Trace(writer, args, options)
> +
> +        parsing.parse_file(writer, trace,
> self.mk_fname('proc_ps.log'))
> +        trace.compile(writer)
> +        self.processtree = process_tree.ProcessTree(writer, None,
> trace.ps_stats, \
> +            trace.ps_stats.sample_period, None, options.prune, None,
> None, False, for_testing = True) +
> +    def mk_fname(self,f):
> +        return os.path.join(self.rootdir, f)
> +
> +    def flatten(self, process_tree):
> +        flattened = []
> +        for p in process_tree:
> +            flattened.append(p)
> +            flattened.extend(self.flatten(p.child_list))
> +        return flattened
> +
> +    def checkAgainstJavaExtract(self, filename, process_tree):
> +        test_data = open(filename)
> +        for expected, actual in zip(test_data,
> self.flatten(process_tree)):
> +            tokens = expected.split('\t')
> +            self.assertEqual(int(tokens[0]), actual.pid // 1000)
> +            self.assertEqual(tokens[1], actual.cmd)
> +            self.assertEqual(long(tokens[2]), 10 * actual.start_time)
> +            self.assert_(long(tokens[3]) - 10 * actual.duration < 5,
> "duration")
> +            self.assertEqual(int(tokens[4]), len(actual.child_list))
> +            self.assertEqual(int(tokens[5]), len(actual.samples))
> +        test_data.close()
> +
> +    def testBuild(self):
> +        process_tree = self.processtree.process_tree
> +
> self.checkAgainstJavaExtract(self.mk_fname('extract.processtree.1.log'),
> process_tree) +
> +    def testMergeLogger(self):
> +        self.processtree.merge_logger(self.processtree.process_tree,
> 'bootchartd', None, False)
> +        process_tree = self.processtree.process_tree
> +
> self.checkAgainstJavaExtract(self.mk_fname('extract.processtree.2.log'),
> process_tree) +
> +    def testPrune(self):
> +        self.processtree.merge_logger(self.processtree.process_tree,
> 'bootchartd', None, False)
> +        self.processtree.prune(self.processtree.process_tree, None)
> +        process_tree = self.processtree.process_tree
> +
> self.checkAgainstJavaExtract(self.mk_fname('extract.processtree.3b.log'),
> process_tree) +
> +    def testMergeExploders(self):
> +        self.processtree.merge_logger(self.processtree.process_tree,
> 'bootchartd', None, False)
> +        self.processtree.prune(self.processtree.process_tree, None)
> +
> self.processtree.merge_exploders(self.processtree.process_tree,
> set(['hwup']))
> +        process_tree = self.processtree.process_tree
> +
> self.checkAgainstJavaExtract(self.mk_fname('extract.processtree.3c.log'),
> process_tree) +
> +    def testMergeSiblings(self):
> +        self.processtree.merge_logger(self.processtree.process_tree,
> 'bootchartd', None, False)
> +        self.processtree.prune(self.processtree.process_tree, None)
> +
> self.processtree.merge_exploders(self.processtree.process_tree,
> set(['hwup']))
> +
> self.processtree.merge_siblings(self.processtree.process_tree)
> +        process_tree = self.processtree.process_tree
> +
> self.checkAgainstJavaExtract(self.mk_fname('extract.processtree.3d.log'),
> process_tree) +
> +    def testMergeRuns(self):
> +        self.processtree.merge_logger(self.processtree.process_tree,
> 'bootchartd', None, False)
> +        self.processtree.prune(self.processtree.process_tree, None)
> +
> self.processtree.merge_exploders(self.processtree.process_tree,
> set(['hwup']))
> +
> self.processtree.merge_siblings(self.processtree.process_tree)
> +        self.processtree.merge_runs(self.processtree.process_tree)
> +        process_tree = self.processtree.process_tree
> +
> self.checkAgainstJavaExtract(self.mk_fname('extract.processtree.3e.log'),
> process_tree) + +if __name__ == '__main__':
> +    unittest.main()


  reply	other threads:[~2021-10-05  9:13 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-27 14:17 [PATCH v2 0/4] Use buildstats from OE to generate build charts Uladzimir Bely
2021-09-27 14:17 ` [PATCH v2 1/4] buildstats: Borrow buildstats and pybootchartgui from OE Uladzimir Bely
2021-10-05  9:13   ` Henning Schild [this message]
2021-09-27 14:17 ` [PATCH v2 2/4] buildstats: Fix bbclass to work with ISAR Uladzimir Bely
2021-10-05  9:19   ` Henning Schild
2021-09-27 14:17 ` [PATCH v2 3/4] buildstats: Manage buildstats via USER_CLASSES variable Uladzimir Bely
2021-10-05  9:21   ` Henning Schild
2021-09-27 14:17 ` [PATCH v2 4/4] doc: Add buildstats section in user manual Uladzimir Bely

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=20211005111333.6153efc5@md1za8fc.ad001.siemens.net \
    --to=henning.schild@siemens.com \
    --cc=isar-users@googlegroups.com \
    --cc=ubely@ilbers.de \
    /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