René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"SIGSEGV using `journal.Reader`, thread safety issue?","articleBody":"### Env:\n```\n$ python3 --version\nPython 3.11.2\n$ systemctl --version\nsystemd 252 (252.36-1~deb12u1)\n+PAM +AUDIT +SELINUX +APPARMOR +IMA +SMACK +SECCOMP +GCRYPT -GNUTLS +OPENSSL +ACL +BLKID +CURL +ELFUTILS +FIDO2 +IDN2 -IDN +IPTC +KMOD +LIBCRYPTSETUP +LIBFDISK +PCRE2 -PWQUALITY +P11KIT +QRENCODE +TPM2 +BZIP2 +LZ4 +XZ +ZLIB +ZSTD -BPF_FRAMEWORK -XKBCOMMON +UTMP +SYSVINIT default-hierarchy=unified\n$ python3 -c \"from systemd import journal; print(journal.__version__)\"\n235\n```\n### Issue:\nI got sporadically a segfault using journal.Reader from python-systemd module.\nThereby it seemed to happen by stop of thread, monitoring journal with Reader.\n\nI could reproduce it also under dgb, here is the callstack:\n\u003cdetails\u003e\n\n```\n[New Thread 0x7ffff5f876c0 (LWP 89339)]\n\nThread 2 \"f2b/f.DummyJail\" received signal SIGSEGV, Segmentation fault.\n[Switching to Thread 0x7ffff5f876c0 (LWP 89339)]\n0x00007ffff6f86d63 in ?? () from /lib/x86_64-linux-gnu/libsystemd.so.0\n(gdb) where\n#0  0x00007ffff6f86d63 in ?? () from /lib/x86_64-linux-gnu/libsystemd.so.0\n#1  0x00007ffff6f877f5 in ?? () from /lib/x86_64-linux-gnu/libsystemd.so.0\n#2  0x00007ffff6f6a848 in ?? () from /lib/x86_64-linux-gnu/libsystemd.so.0\n#3  0x00007ffff6f99502 in ?? () from /lib/x86_64-linux-gnu/libsystemd.so.0\n#4  0x00007ffff6f69403 in ?? () from /lib/x86_64-linux-gnu/libsystemd.so.0\n#5  0x00007ffff6f6961e in ?? () from /lib/x86_64-linux-gnu/libsystemd.so.0\n#6  0x00007ffff6f6d11a in sd_journal_get_fd () from /lib/x86_64-linux-gnu/libsystemd.so.0\n#7  0x00007ffff6f6f3c9 in sd_journal_wait () from /lib/x86_64-linux-gnu/libsystemd.so.0\n#8  0x00007ffff6cadf2c in ?? () from /usr/lib/python3/dist-packages/systemd/_reader.cpython-311-x86_64-linux-gnu.so\n#9  0x0000000000548ac8 in ?? ()\n#10 0x00000000005188b3 in _PyObject_MakeTpCall ()\n#11 0x000000000052c2f0 in _PyEval_EvalFrameDefault ()\n#12 0x0000000000585e54 in ?? ()\n#13 0x0000000000584e98 in ?? ()\n#14 0x000000000053034e in _PyEval_EvalFrameDefault ()\n#15 0x0000000000585e54 in ?? ()\n#16 0x0000000000584e98 in ?? ()\n#17 0x000000000067b32c in ?? ()\n#18 0x0000000000656384 in ?? ()\n#19 0x00007ffff7d381f5 in start_thread (arg=\u003coptimized out\u003e) at ./nptl/pthread_create.c:442\n#20 0x00007ffff7db889c in clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:81\n```\n\u003c/details\u003e\n\nIt can be either some missing lock or atomic handling somewhere in journal.Reader, because if I rewrote the code to avoid too earlier clean-up, closing the journal with `Reader.close()` before join of thread (now it ensures that the thread working with Reader is really stopped, before clean-up happens), it doesn't happen anymore.\nSo my assumption is if TH2 is somewhere within `reader.wait()` and TH1 would invoke `reader.close()`, it could segfault.\nSure, it can be avoided on the user side, just SF is SF and shall not happen normally.\n\nNaive attempts to reproduce it with simple wait/close in different threads did not work, but I could find a bit complex scenario (with a busy-wait \u0026 lock using to synchronize the timing a bit), where it is pretty reproducible, almost every time not later than on 10 iteration:\n```python\nimport time\nfrom systemd import journal\nfrom threading import Thread, Lock\nl = Lock(); # just used for \"better\" timing here to get SF\n\ndef test():\n  stat = {};\n  jr = journal.Reader()\n  def mon():\n    print('start')\n    try:\n      jr.seek_head()\n      print('work')\n      stat['started'] = 1\n      l.acquire(); l.release()\n      # *-* [A], thread is here as another thread is on *-* [B]\n      while (j := jr):\n        if j.wait(3) in (journal.NOP, journal.APPEND):\n          j.get_next()\n    except Exception as e:\n      print('ERR: %s' % e)\n    if jr: jr.close()\n  th = Thread(target=mon); \n  th.start()\n  # just \"sync\" wait and close calls:\n  while not stat: time.sleep(0.00001);\n  l.acquire(); l.release()\n  # *-* [B], thread is here as another thread is on *-* [A]\n  # do evil close:\n  jr.close(); jr = None\n  th.join()\n  print('OK')\n\nfor i in range(100): # just 100 attempts, because of sporadic matter\n  test()\n```","author":{"url":"https://github.com/sebres","@type":"Person","name":"sebres"},"datePublished":"2025-03-30T03:29:49.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":3},"url":"https://github.com/143/python-systemd/issues/143"}