如何调试 python 核心转储？
How to debug python core dump?
我的 python 程序调用共享对象中的函数。有两个核心转储文件。但是我无法从他们那里得到任何有用的信息。我的 python 程序调用共享对象内的函数。有两个核心转储文件。但是我无法从他们那里得到任何有用的信息。
(gdb) bt
#0 0x00007f9a436b6d44 in PyObject_stgdict () from /home/ubuntu/bingo/lib/python2.7/lib-dynload/_ctypes.x86_64-linux-gnu.so
#1 0x00007f9a436b66e5 in _ctypes_callproc () from /home/ubuntu/bingo/lib/python2.7/lib-dynload/_ctypes.x86_64-linux-gnu.so
#2 0x00007f9a436b7f9e in ?? () from /home/ubuntu/bingo/lib/python2.7/lib-dynload/_ctypes.x86_64-linux-gnu.so
#3 0x0000000000499be5 in PyEval_EvalFrameEx ()
#4 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#5 0x00000000004a090c in PyEval_EvalCodeEx ()
#6 0x000000000049ab45 in PyEval_EvalFrameEx ()
#7 0x00000000004a090c in PyEval_EvalCodeEx ()
#8 0x000000000049ab45 in PyEval_EvalFrameEx ()
#9 0x00000000004a1c9a in ?? ()
#10 0x0000000000505f96 in PyObject_Call ()
#11 0x000000000049b07a in PyEval_EvalFrameEx ()
#12 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#13 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#14 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#15 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#16 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#17 0x00000000004a090c in PyEval_EvalCodeEx ()
#18 0x000000000049ab45 in PyEval_EvalFrameEx ()
#19 0x00000000004a090c in PyEval_EvalCodeEx ()
#20 0x000000000049ab45 in PyEval_EvalFrameEx ()
#21 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#22 0x00000000004a1c9a in ?? ()
#23 0x00000000004dfe94 in ?? ()
#24 0x00000000004dc9cb in PyEval_CallObjectWithKeywords ()
#25 0x0000000000588b12 in ?? ()
#26 0x00007f9a449de184 in start_thread (arg=0x7f99f1ffb700) at pthread_create.c:312
#27 0x00007f9a4470b03d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111
(gdb) bt
#0 0x000000000054eee0 in ?? ()
#1 0x000000000057392b in ?? ()
#2 0x0000000000536476 in ?? ()
#3 0x0000000000537827 in _PyObject_GC_New ()
#4 0x000000000052fcb2 in PyList_New ()
#5 0x00000000005ae57d in ?? ()
#6 0x000000000051eb00 in ?? ()
#7 0x000000000049a3b5 in PyEval_EvalFrameEx ()
#8 0x00000000004a090c in PyEval_EvalCodeEx ()
#9 0x000000000049ab45 in PyEval_EvalFrameEx ()
#10 0x00000000004a090c in PyEval_EvalCodeEx ()
#11 0x000000000049ab45 in PyEval_EvalFrameEx ()
#12 0x00000000004a090c in PyEval_EvalCodeEx ()
#13 0x000000000049ab45 in PyEval_EvalFrameEx ()
#14 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#15 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#16 0x00000000004a090c in PyEval_EvalCodeEx ()
#17 0x000000000049ab45 in PyEval_EvalFrameEx ()
#18 0x00000000004a090c in PyEval_EvalCodeEx ()
#19 0x000000000049ab45 in PyEval_EvalFrameEx ()
#20 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#21 0x00000000004a1c9a in ?? ()
#22 0x00000000004dfe94 in ?? ()
#23 0x00000000004dc9cb in PyEval_CallObjectWithKeywords ()
#24 0x0000000000588b12 in ?? ()
#25 0x00007fad2dbdb184 in start_thread (arg=0x7fad255d3700) at pthread_create.c:312
#26 0x00007fad2d90803d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111
I can't get any useful information from them.
相反：您确实从他们那里得到了有用的信息。
两次崩溃都在 Python 内，其中一次发生在分配例程内。这通常堆损坏的迹象。
My python program calls functions inside a shared object.
如果“共享对象”指的是您（或其他人）开发的库，则该库可能存在堆损坏错误。此类错误非常严重，因为它们的影响（例如崩溃）通常会在 1000 条指令之后发生，而且众所周知，如果没有专门的工具，很难找到它们。
幸运的是我们有这样的工具：Valgrind 和 Address Sanitizer。
我将从 运行 您在 Valgrind 下的程序开始。请注意，您需要 special Python 版本才能生效。
我的 python 程序调用共享对象中的函数。有两个核心转储文件。但是我无法从他们那里得到任何有用的信息。我的 python 程序调用共享对象内的函数。有两个核心转储文件。但是我无法从他们那里得到任何有用的信息。
(gdb) bt
#0 0x00007f9a436b6d44 in PyObject_stgdict () from /home/ubuntu/bingo/lib/python2.7/lib-dynload/_ctypes.x86_64-linux-gnu.so
#1 0x00007f9a436b66e5 in _ctypes_callproc () from /home/ubuntu/bingo/lib/python2.7/lib-dynload/_ctypes.x86_64-linux-gnu.so
#2 0x00007f9a436b7f9e in ?? () from /home/ubuntu/bingo/lib/python2.7/lib-dynload/_ctypes.x86_64-linux-gnu.so
#3 0x0000000000499be5 in PyEval_EvalFrameEx ()
#4 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#5 0x00000000004a090c in PyEval_EvalCodeEx ()
#6 0x000000000049ab45 in PyEval_EvalFrameEx ()
#7 0x00000000004a090c in PyEval_EvalCodeEx ()
#8 0x000000000049ab45 in PyEval_EvalFrameEx ()
#9 0x00000000004a1c9a in ?? ()
#10 0x0000000000505f96 in PyObject_Call ()
#11 0x000000000049b07a in PyEval_EvalFrameEx ()
#12 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#13 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#14 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#15 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#16 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#17 0x00000000004a090c in PyEval_EvalCodeEx ()
#18 0x000000000049ab45 in PyEval_EvalFrameEx ()
#19 0x00000000004a090c in PyEval_EvalCodeEx ()
#20 0x000000000049ab45 in PyEval_EvalFrameEx ()
#21 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#22 0x00000000004a1c9a in ?? ()
#23 0x00000000004dfe94 in ?? ()
#24 0x00000000004dc9cb in PyEval_CallObjectWithKeywords ()
#25 0x0000000000588b12 in ?? ()
#26 0x00007f9a449de184 in start_thread (arg=0x7f99f1ffb700) at pthread_create.c:312
#27 0x00007f9a4470b03d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111
(gdb) bt
#0 0x000000000054eee0 in ?? ()
#1 0x000000000057392b in ?? ()
#2 0x0000000000536476 in ?? ()
#3 0x0000000000537827 in _PyObject_GC_New ()
#4 0x000000000052fcb2 in PyList_New ()
#5 0x00000000005ae57d in ?? ()
#6 0x000000000051eb00 in ?? ()
#7 0x000000000049a3b5 in PyEval_EvalFrameEx ()
#8 0x00000000004a090c in PyEval_EvalCodeEx ()
#9 0x000000000049ab45 in PyEval_EvalFrameEx ()
#10 0x00000000004a090c in PyEval_EvalCodeEx ()
#11 0x000000000049ab45 in PyEval_EvalFrameEx ()
#12 0x00000000004a090c in PyEval_EvalCodeEx ()
#13 0x000000000049ab45 in PyEval_EvalFrameEx ()
#14 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#15 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#16 0x00000000004a090c in PyEval_EvalCodeEx ()
#17 0x000000000049ab45 in PyEval_EvalFrameEx ()
#18 0x00000000004a090c in PyEval_EvalCodeEx ()
#19 0x000000000049ab45 in PyEval_EvalFrameEx ()
#20 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#21 0x00000000004a1c9a in ?? ()
#22 0x00000000004dfe94 in ?? ()
#23 0x00000000004dc9cb in PyEval_CallObjectWithKeywords ()
#24 0x0000000000588b12 in ?? ()
#25 0x00007fad2dbdb184 in start_thread (arg=0x7fad255d3700) at pthread_create.c:312
#26 0x00007fad2d90803d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111
I can't get any useful information from them.
相反：您确实从他们那里得到了有用的信息。
两次崩溃都在 Python 内，其中一次发生在分配例程内。这通常堆损坏的迹象。
My python program calls functions inside a shared object.
如果“共享对象”指的是您（或其他人）开发的库，则该库可能存在堆损坏错误。此类错误非常严重，因为它们的影响（例如崩溃）通常会在 1000 条指令之后发生，而且众所周知，如果没有专门的工具，很难找到它们。
幸运的是我们有这样的工具：Valgrind 和 Address Sanitizer。
我将从 运行 您在 Valgrind 下的程序开始。请注意，您需要 special Python 版本才能生效。