Bugs item #1429585, was opened at 2006-02-11 00:16 Message generated for change (Settings changed) made by nnorwitz You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=105470&aid=1429585&group_id=5470
Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Extension Modules Group: Python 2.4 >Status: Closed Resolution: Works For Me Priority: 5 Submitted By: aix-d (aix-d) Assigned to: Neal Norwitz (nnorwitz) Summary: segfault in FreeBSD Initial Comment: Dear developers, My program has many TCP connections (~30 per second), MySQL connections (~100 per second) and threads (~280). It works long time and receives this segmentation fault: exited on signal 11 (core dumped) =========== GDB OUTPUT ================================================= Program received signal SIGSEGV, Segmentation fault. 0x283199e8 in ?? () from /usr/local/lib/python2.4/lib-dynload/_socket.so (gdb) backtrace #0 0x283199e8 in ?? () from /usr/local/lib/python2.4/lib-dynload/_socket.so #1 0x2831b3d5 in internal_connect () from /usr/local/lib/python2.4/lib-dynload/_socket.so #2 0x2831b440 in sock_connect () from /usr/local/lib/python2.4/lib-dynload/_socket.so #3 0x0805a828 in PyObject_Call () #4 0x080a01dd in PyEval_EvalFrame () #5 0x080a294f in PyEval_EvalCodeEx () #6 0x080a138b in PyEval_EvalFrame () #7 0x080a212d in PyEval_EvalFrame () #8 0x080a212d in PyEval_EvalFrame () #9 0x080a212d in PyEval_EvalFrame () #10 0x080a212d in PyEval_EvalFrame () #11 0x080a212d in PyEval_EvalFrame () #12 0x080a294f in PyEval_EvalCodeEx () #13 0x080a138b in PyEval_EvalFrame () #14 0x080a212d in PyEval_EvalFrame () #15 0x080a294f in PyEval_EvalCodeEx () #16 0x080da027 in PyFunction_SetClosure () #17 0x0805a828 in PyObject_Call () #18 0x080a01dd in PyEval_EvalFrame () #19 0x080a294f in PyEval_EvalCodeEx () #20 0x080a138b in PyEval_EvalFrame () #21 0x080a294f in PyEval_EvalCodeEx () #22 0x080a138b in PyEval_EvalFrame () #23 0x080a294f in PyEval_EvalCodeEx () #24 0x080a138b in PyEval_EvalFrame () #25 0x080a294f in PyEval_EvalCodeEx () #26 0x080a138b in PyEval_EvalFrame () #27 0x080a294f in PyEval_EvalCodeEx () #28 0x080a138b in PyEval_EvalFrame () #29 0x080a294f in PyEval_EvalCodeEx () #30 0x080a138b in PyEval_EvalFrame () #31 0x080a212d in PyEval_EvalFrame () #32 0x080a294f in PyEval_EvalCodeEx () #33 0x080da027 in PyFunction_SetClosure () #34 0x0805a828 in PyObject_Call () #35 0x080601e9 in PyMethod_New () #36 0x0805a828 in PyObject_Call () #37 0x0809cd31 in PyEval_CallObjectWithKeywords () #38 0x080c21a5 in _PyObject_GC_Del () #39 0x2821e0cc in pthread_create () from /usr/lib/libpthread.so.2 #40 0x282db757 in _ctx_start () from /lib/libc.so.6 (gdb) ==== END OF GDB OUTPUT ================================================= Imported modules: socket, atexit, time, pickle, os, random, traceback, MySQLdb, re, urllib, urllib2, threading Python version: 2.4.2 Operating system: FreeBSD 6.0-STABLE Can you help with debugging? Best regards, Alexander. ---------------------------------------------------------------------- >Comment By: Neal Norwitz (nnorwitz) Date: 2006-02-12 15:20 Message: Logged In: YES user_id=33168 Closing this bug report since it's fixed. As to python not compiling on FreeBSD, open a new bug report with all the info about what is causing the failure. It would be great if you can help diagnose the problem as not many python developers have access to FreeBSD. Is the FreeBSD port up-to-date? ---------------------------------------------------------------------- Comment By: Uyamba (uyamba) Date: 2006-02-12 15:14 Message: Logged In: YES user_id=1450360 Thank you very much!!! :-) It works fine: python-devel from FreeBSD ports with three files updated by SVN to latest version. But i can't compile full last SVN release on FreeBSD :-( Could you help with updating FreeBSD port please. ---------------------------------------------------------------------- Comment By: Neal Norwitz (nnorwitz) Date: 2006-02-12 13:11 Message: Logged In: YES user_id=33168 The 2.4 fix was committed in revision 42339. ---------------------------------------------------------------------- Comment By: Neal Norwitz (nnorwitz) Date: 2006-02-12 13:09 Message: Logged In: YES user_id=33168 I get exceptions when running normally. I had to do ulimit -n 2000 to be able to run without exception. I can confirm that with an unpatched 2.4.2, python crashed. With the 2.4.2+ patch, no crash. On SVN head (2.5), I didn't get a crash, but got a thread exception. I think this bug may be fixed. To get the new features, you need to update at least these files: http://svn.python.org/projects/python/branches/release24-maint/Modules/socketmodule.c http://svn.python.org/projects/python/branches/release24-maint/Modules/_ssl.c http://svn.python.org/projects/python/branches/release24-maint/PC/pyconfig.h Please verify if this is fixed for you. ---------------------------------------------------------------------- Comment By: Neal Norwitz (nnorwitz) Date: 2006-02-12 12:54 Message: Logged In: YES user_id=33168 Can you "print *s" ? My guess is that s->sock_fd > FD_SETSIZE on your box. I will try your test code with the current versions and see if they still fail. ---------------------------------------------------------------------- Comment By: Uyamba (uyamba) Date: 2006-02-12 09:48 Message: Logged In: YES user_id=1450360 This is detailed backtrace (from python compiled with CFLAGS+="-g"): ====================================== #0 0xa8348959 in ?? () #1 0x08ebb800 in ?? () #2 0x00000001 in ?? () #3 0x77881688 in ?? () #4 0x28348947 in internal_connect (s=0x8ebb800, addr=0x8ebb818, addrlen=16, timeoutp=0x778816b4) at /usr/ports/lang/python/work/Python-2.4.2/Modules/ socketmodule.c:1740 #5 0x28348a00 in sock_connect (s=0x8ebb800, addro=0x8ebd72c) at /usr/ports/lang/python/work/Python-2.4.2/Modules/ socketmodule.c:1771 ====================================== This is the /usr/ports/lang/python/work/Python-2.4.2/ Modules/socketmodule.c: ====================================== if (s->sock_timeout > 0.0) { 1740: if (res < 0 && errno == EINPROGRESS) { timeout = internal_select(s, 1); res = connect(s->sock_fd, addr, addrlen); if (res < 0 && errno == EISCONN) res = 0; } } ====================================== No ideas :-( ---------------------------------------------------------------------- Comment By: Uyamba (uyamba) Date: 2006-02-12 08:47 Message: Logged In: YES user_id=1450360 This is test program that generates fault: ---------------------------- #!/usr/local/bin/python import socket, time from threading import Thread class cthread(Thread): def __init__ (self): Thread.__init__(self) def run(self): socket.setdefaulttimeout(5) res = socket.getaddrinfo('19.54.123.2', 33, 0, socket.SOCK_STREAM)[0] sock = socket.socket(res[0], res[1], res[2]) try: sock.connect(res[4]) except: pass def main(): for i in range(1900): print "starting " + str(i) ct = cthread() ct.start() if __name__ == '__main__': main() ---------------------------- IP 19.54.123.2 is random. Results are various for some IPs and hostnames. ---------------------------------------------------------------------- Comment By: Neal Norwitz (nnorwitz) Date: 2006-02-11 22:54 Message: Logged In: YES user_id=33168 Can you build python from source? You can try checking out the current 2.4 SVN. This problem may have been fixed recently. I'm guessing that the socket descriptor value is larger than FD_SETSIZE (probably 1k or 4k). You also may want to build python with debugging. If that doesn't solve your problem, you will need to provide more information about how to reproduce your problem. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=105470&aid=1429585&group_id=5470 _______________________________________________ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
