python 2.5.1 segfault, multithreading & dual core issue?

[Paul Sijben]

I am running a multi-threaded python application in a dual core intel
running Ubuntu.

I am using python 2.5.1 that I compiled myself. At random points I am
getting segmentation faults (sometimes indicating a duplicate free).
Below is the backtrace of the latest segfault.

I am thinking this might be an issue related to the dual core CPU so I
am now running the app with affinity to one CPU to test this hypothesis.

Where can I put such a bugreport?

----

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1486967920 (LWP 4579)]
0x009196ed in fclose@@GLIBC_2.1 () from /lib/libc.so.6
(gdb) bt
#0 0x009196ed in fclose@@GLIBC_2.1 () from /lib/libc.so.6
#1 0x0806af9b in file_close (f=0xa259e30) at Objects/fileobject.c:446
#2 0x080c51b0 in PyEval_EvalFrameEx (f=0xa3c790c, throwflag=0) at
Python/ceval.c:3548
#3 0x080c5795 in PyEval_EvalFrameEx (f=0xa3c77a4, throwflag=0) at
Python/ceval.c:3650
#4 0x080c5795 in PyEval_EvalFrameEx (f=0xa3afccc, throwflag=0) at
Python/ceval.c:3650
#5 0x080c5795 in PyEval_EvalFrameEx (f=0xa3c8804, throwflag=0) at
Python/ceval.c:3650
#6 0x080c65a5 in PyEval_EvalCodeEx (co=0xb7f315c0, globals=0xb7f29824,
locals=0x0, args=0xa39b9a8, argcount=3,
kws=0xa39b9b4, kwcount=0, defs=0xb7c8fb98, defcount=1, closure=0x0)
at Python/ceval.c:2831
#7 0x080c4a59 in PyEval_EvalFrameEx (f=0xa39b864, throwflag=0) at
Python/ceval.c:3660
#8 0x080c65a5 in PyEval_EvalCodeEx (co=0xb7f2ec38, globals=0xb7f29824,
locals=0x0, args=0xa399920, argcount=3,
kws=0xa39992c, kwcount=0, defs=0xb7c8fad8, defcount=1, closure=0x0)
at Python/ceval.c:2831
#9 0x080c4a59 in PyEval_EvalFrameEx (f=0xa3997c4, throwflag=0) at
Python/ceval.c:3660
#10 0x080c65a5 in PyEval_EvalCodeEx (co=0xb7f2eda0, globals=0xb7f29824,
locals=0x0, args=0xa340e30, argcount=3,
kws=0xa340e3c, kwcount=0, defs=0xb7cea330, defcount=4, closure=0x0)
at Python/ceval.c:2831
#11 0x080c4a59 in PyEval_EvalFrameEx (f=0xa340cb4, throwflag=0) at
Python/ceval.c:3660
#12 0x080c65a5 in PyEval_EvalCodeEx (co=0xb7f317b8, globals=0xb7f29824,
locals=0x0, args=0xa39cedc, argcount=2,
kws=0xa39cee4, kwcount=2, defs=0xb7c905b0, defcount=3, closure=0x0)
at Python/ceval.c:2831
#13 0x080c4a59 in PyEval_EvalFrameEx (f=0xa39cd7c, throwflag=0) at
Python/ceval.c:3660
#14 0x080c65a5 in PyEval_EvalCodeEx (co=0xb7f316e0, globals=0xb7f29824,
locals=0x0, args=0xa3aff8c, argcount=4,
kws=0xa3aff9c, kwcount=1, defs=0xb7c8fbb8, defcount=2, closure=0x0)
at Python/ceval.c:2831
#15 0x080c4a59 in PyEval_EvalFrameEx (f=0xa3afe34, throwflag=0) at
Python/ceval.c:3660
#16 0x080c65a5 in PyEval_EvalCodeEx (co=0xb7f07728, globals=0xb7f59acc,
locals=0x0, args=0xa32b9ac, argcount=3,
kws=0xa32b9b8, kwcount=0, defs=0xb7c0a1f8, defcount=2, closure=0x0)
at Python/ceval.c:2831
#17 0x080c4a59 in PyEval_EvalFrameEx (f=0xa32b86c, throwflag=0) at
Python/ceval.c:3660
#18 0x080c5795 in PyEval_EvalFrameEx (f=0xa3c84f4, throwflag=0) at
Python/ceval.c:3650
#19 0x080c5795 in PyEval_EvalFrameEx (f=0xa3a4ef4, throwflag=0) at
Python/ceval.c:3650
#20 0x080c65a5 in PyEval_EvalCodeEx (co=0xb7f1d1d0, globals=0xb7f11dfc,
locals=0x0, args=0xa2d8cd8, argcount=1,
kws=0x0, kwcount=0, defs=0x0, defcount=0, closure=0x0) at
Python/ceval.c:2831
#21 0x0810d6a1 in function_call (func=0xb7f26e2c, arg=0xa2d8ccc, kw=0x0)
at Objects/funcobject.c:517
#22 0x0805a257 in PyObject_Call (func=0xa48ff4, arg=0xa2d8ccc, kw=0x0)
at Objects/abstract.c:1860
#23 0x08060387 in instancemethod_call (func=0xa23b2fc, arg=0xa2d8ccc,
kw=0x0) at Objects/classobject.c:2497
#24 0x0805a257 in PyObject_Call (func=0xa48ff4, arg=0xb7f4102c, kw=0x0)
at Objects/abstract.c:1860
#25 0x080be79c in PyEval_CallObjectWithKeywords (func=0xa23b2fc,
arg=0xb7f4102c, kw=0x0) at Python/ceval.c:3433
#26 0x080f01b8 in t_bootstrap (boot_raw=0xa3c8048) at
../Modules/threadmodule.c:424
#27 0x00a3a45b in start_thread () from /lib/libpthread.so.0
#28 0x0099223e in clone () from /lib/libc.so.6

 

[Peter Otten]

I am running a multi-threaded python application in a dual core intel
running Ubuntu.

I am using python 2.5.1 that I compiled myself. At random points I am
getting segmentation faults (sometimes indicating a duplicate free).
Below is the backtrace of the latest segfault.

I am thinking this might be an issue related to the dual core CPU so I
am now running the app with affinity to one CPU to test this hypothesis.

Where can I put such a bugreport?

http://sourceforge.net/bugs/?group_id=5470

if you do it before August 23, when the bugtracker will be moved to
bugs.python.org.

Note that a thread-related bug causing a segfault has been reported today at

http://sourceforge.net/tracker/index.php?func=detail&aid=1778376&group_id=5470&atid=105470

I could reproduce this one on my machine, though, so I'm sure that it has
nothing to do with a dual core CPU.

Peter

 

[Hrvoje Niksic]

I am running a multi-threaded python application in a dual core
intel running Ubuntu.

[...]

Judging from the stack trace, this patch has a good chance of fixing
your problem:

http://mail.python.org/pipermail/python-dev/2007-August/074232.html

 

[Paul Sijben]

thanks very much! I am currently compiling python with the patch and
will test it over the coming days.

Paul

I am running a multi-threaded python application in a dual core
intel running Ubuntu.

[...]

Judging from the stack trace, this patch has a good chance of fixing
your problem:

http://mail.python.org/pipermail/python-dev/2007-August/074232.html