Quellcodebibliothek Statistik Leitseite products/Sources/formale Sprachen/C/LibreOffice/bin/   (Office von Apache Version 25.8.3.2©)  Datei vom 5.10.2025 mit Größe 5 kB image not shown  

Quelle  refcount_leak.py   Sprache: Python

 
#!/usr/bin/python3
# -*- tab-width: 4; indent-tabs-mode: nil; py-indent-offset: 4 -*-
#
# This file is part of the LibreOffice project.
#
# This Source Code Form is subject to the terms of the Mozilla Public
# License, v. 2.0. If a copy of the MPL was not distributed with this
# file, You can obtain one at http://mozilla.org/MPL/2.0/.
#

### script to help debug leaks of reference counted objects

## I. to use it, first override acquire() and release()

# Foo * g_pTrackedFoo = 0;

# Foo::Foo()
#   static int nFoos = 0;
#     if (++nFoos == 42) // track instance #42
#       g_pTrackedFoo = this;

# void Foo::acquire()
#   if (this == g_pTrackedFoo)
#     ; // set gdb breakpoint here
#   Foo_Base::acquire()

# void Foo::release()
#   if (this == g_pTrackedFoo)
#      ; // set gdb breakpoint here
#   Foo_Base::release()

## II. run test/soffice in gdb and set breakpoints in acquire/release
##     with a command to print the backtrace

# set logging enabled on
# break foo.cxx:123
# break foo.cxx:234

# command 1 2
# bt
# c
# end
# run

## III. now feed logfile gdb.txt into this script

# bin/refcount_leak.py < gdb.txt

###

from operator import itemgetter
import re
import sys

threshold = 2

class Trace:
    clock = 0 # global counter
    # frames: list of stack frames, beginning with outermost
    def __init__(self, lines):
        lines.reverse()
        self.frames = lines
        Trace.clock += 1
        self.clock = Trace.clock

def addTrace(traces, lines):
    if traces is not None and len(lines) > 0:
        traces.append(Trace(lines))

def readGdbLog(infile):
    traces_acquire = []
    traces_release = []
    current = None
    lines = []
    apattern = re.compile("^Breakpoint.*::acquire")
    rpattern = re.compile("^Breakpoint.*::release")
    for line in infile:
        if apattern.match(line):
            addTrace(current, lines)
            lines = []
            current = traces_acquire
        if rpattern.match(line):
            addTrace(current, lines)
            lines = []
            current = traces_release
        if line.startswith("#"):
            # strip #123 stack frame number, and newline
            lines.append(line[line.index("0x"):-1])
    addTrace(current, lines)
    print("# parsed traces acquire: ", len(traces_acquire))
    print("# parsed traces release: ", len(traces_release))
    return (traces_acquire, traces_release)

def getFunction(frame):
    start = frame.index(" in ") + len(" in ")
    try:
        end = frame.index(" at ", start)
    except ValueError:
        # argh... stack frames may be split across multiple lines if
        # a parameter has a fancy pretty printer
        return frame[start:]
    return frame[start:end]


def matchStack(trace_acquire, trace_release):
    if trace_release.clock < trace_acquire.clock:
        return None # acquire must precede release
    common = 0
    refpattern = re.compile(r"::Reference<.*>::Reference\(")
    for (frame1, frame2) in zip(trace_release.frames, trace_acquire.frames):
        if frame1 == frame2:
            common += 1
        else:
            if getFunction(frame1) == getFunction(frame2):
                common += 1
            acquireframes = len(trace_acquire.frames)
            # there is sometimes a dozen frames of UNO type related junk
            # on the stack where the acquire() happens, which breaks the
            # matching; try to avoid that
            for i in range(common, acquireframes):
                if refpattern.search(trace_acquire.frames[i]):
                    acquireframes = i+1 # cut off junk above Reference ctor
                    break
            score = max(len(trace_release.frames), acquireframes) - common
            # smaller score is better
            return (score, trace_release.clock - trace_acquire.clock)

# brute force greedy n^2 matching
def matchStacks(traces_acquire, traces_release):
    matches = []
    for release in traces_release:
        for acquire in traces_acquire:
            score = matchStack(acquire, release)
            if score is not None:
                matches.append((score, acquire, release))
    matches.sort(key=itemgetter(0))
    return matches

def bestMatches(traces_acquire, traces_release, matches):
    traces_aunmatched = traces_acquire
    traces_runmatched = traces_release
    bestmatches = []
    for (score,acquire,release) in matches:
        if not(acquire in traces_aunmatched and release in traces_runmatched):
            continue
        traces_aunmatched.remove(acquire)
        traces_runmatched.remove(release)
        bestmatches.append((score,acquire,release))
    print("# unmatched acquire: ", len(traces_aunmatched))
    print("# unmatched release: ", len(traces_runmatched))
    return (bestmatches,traces_aunmatched,traces_runmatched)

def printTrace(trace):
    for frame in reversed(trace.frames):
        print(" ", frame)

def printMatched(bestmatches):
    for (score,acquire,release) in reversed(bestmatches):
        print("\n*** Matched trace with score: ", score)
        print(" acquire: ")
        printTrace(acquire)
        print(" release: ")
        printTrace(release)

def printUnmatched(traces, prefix):
    for trace in traces:
        print("\n*** Unmatched trace (", prefix, "):")
        printTrace(trace)

if __name__ == "__main__":
    (traces_acquire, traces_release) = readGdbLog(sys.stdin)
    matches = matchStacks(traces_acquire, traces_release)
    (bestmatches,traces_au,traces_ru) = bestMatches(traces_acquire, traces_release, matches)
    # print output, sorted with the most suspicious stuff first:
    printUnmatched(traces_au, "acquire")
    printUnmatched(traces_ru, "release")
    printMatched(bestmatches)

# vim:set shiftwidth=4 softtabstop=4 expandtab:

72%


¤ Dauer der Verarbeitung: 0.14 Sekunden  (vorverarbeitet)  ¤

*© Formatika GbR, Deutschland






Wurzel

Suchen

Beweissystem der NASA

Beweissystem Isabelle

NIST Cobol Testsuite

Cephes Mathematical Library

Wiener Entwicklungsmethode

Haftungshinweis

Die Informationen auf dieser Webseite wurden nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit, noch Qualität der bereit gestellten Informationen zugesichert.

Bemerkung:

Die farbliche Syntaxdarstellung ist noch experimentell.