
This issue trackerhas been migrated toGitHub, and is currentlyread-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.
Created on2011-06-02 11:57 byNiels.Heinen, last changed2022-04-11 14:57 byadmin.
| Messages (17) | |||
|---|---|---|---|
| msg137473 -(view) | Author: Niels Heinen (Niels.Heinen) | Date: 2011-06-02 11:57 | |
Running the python binary without a script or using the -i flag willstart the process in interactive mode. The interactive mode requires anexternal module to be loaded: readline.Per default behavior, Python also tries to load this module from the current working directory (see also trace below)strcpy(0x7fff17609ed8, ".so") = 0x7fff17609ed8fopen64("readline.so", "rb" <unfinished ...>SYS_open("readline.so", 0, 0666) = -2<... fopen64 resumed> ) = 0strcpy(0x7fff17609ed8, "module.so") = 0x7fff17609ed8fopen64("readlinemodule.so", "rb" <unfinished ...>SYS_open("readlinemodule.so", 0, 0666)The module is imported inModules/main.c line 663: if ((Py_InspectFlag || ...... isatty(fileno(stdin))) { PyObject *v; v = PyImport_ImportModule("readline");Why consider this a security bug: basically because you don't expect aprogram to import a shared library from your current directory _unless_you explicitly tell it to (e.g. import blah).On a multi user system, someone could plant a malicious shared librariesnamed "readline.so" in an attempt to hack a user that runs python ininteractive mode.The risk obviously _very_ low but nevertheless worth to consider improving by, for example, loading readline with a more strict path? (e.g. python lib directories only?)NielsAN EXAMPLE:-----------The code below is compiled to readline.so and stored in /tmp: void __attribute__ ((constructor)) _load(); void _load() { printf("DING DONG!\n"); }foo@foo:/tmp$ ls -l /tmp/readline.so -rwxr-x--- 1 nnnnn nnn 7952 Mar 29 16:24 /tmp/readline.sofoo@foo:/tmp$ pythonPython 2.6.5 (r265:79063, Apr 16 2010, 13:57:41) [GCC 4.4.3] on linux2Type "help", "copyright", "credits" or "license" for more information.DING DONG!>>> | |||
| msg137475 -(view) | Author: R. David Murray (r.david.murray)*![]() | Date: 2011-06-02 12:38 | |
This is a general principle of how Python runs in interactive mode and is not confined to loading readline. The same would be true for any module loaded during startup, and there are quite a few that are so loaded. Since loading modules from the current working directory is an important feature of using python in interactive mode, this is not something that is likely to be changed. | |||
| msg137529 -(view) | Author: Éric Araujo (eric.araujo)*![]() | Date: 2011-06-03 15:42 | |
+1 to what David said. See also#5753. | |||
| msg137741 -(view) | Author: Niels Heinen (Niels.Heinen) | Date: 2011-06-06 15:09 | |
Hi Eric, David,This means that you cannot type "python" and press <enter> in any shared directory without the risk of a malicious readlinemodule.so being imported and executed. I think this is different from a scenario where someone explicitly runs a script or imports a module in interactive mode where it is also reasonable that such a person understands the importing mechanism.Thanks for the quick responses btw!Niels | |||
| msg137752 -(view) | Author: R. David Murray (r.david.murray)*![]() | Date: 2011-06-06 16:58 | |
I've done a little poking around, and it looks like you are correct and I'm wrong. It appears that readline.so is or should be a special case. I've added some people to nosy to see what they think.Specifically, it appears that if I put a file that should shadow a library module that is imported at python startup time (eg: os.py) into my current working directory I still get the os.py from the appropriate lib directory, even though '' is first in my sys.path. This is not how I thought it worked, but it is my observation. I tested this on 2.6.6, 2.7.1 and 3.3 tip. | |||
| msg137804 -(view) | Author: Antoine Pitrou (pitrou)*![]() | Date: 2011-06-07 11:08 | |
I don't think readline is "special-cased":$ echo "1/0" > logging.py$ cpython/default/pythonPython 3.3a0 (default:d8502fee4638+, Jun 6 2011, 19:13:58) [GCC 4.4.3] on linux2Type "help", "copyright", "credits" or "license" for more information.>>> import loggingTraceback (most recent call last): File "<stdin>", line 1, in <module> File "logging.py", line 1, in <module> 1/0ZeroDivisionError: division by zero | |||
| msg137821 -(view) | Author: R. David Murray (r.david.murray)*![]() | Date: 2011-06-07 13:13 | |
Python 3.3a0 (default:7323a865457a+, Jun 5 2011, 19:22:38) [GCC 4.5.2] on linux2Type "help", "copyright", "credits" or "license" for more information.>>> import sys>>> sys.modules['logging']Traceback (most recent call last): File "<stdin>", line 1, in <module>KeyError: 'logging'>>> sys.modules['os']<module 'os' from '/home/rdmurray/python/p33/Lib/os.py'>The difference is that logging is not imported at startup. So, however os (and friends, there are a lot of modules in sys.modules at startup) is imported, it is different from how readline.so is imported. | |||
| msg137824 -(view) | Author: Antoine Pitrou (pitrou)*![]() | Date: 2011-06-07 13:21 | |
> The difference is that logging is not imported at startup. So, however> os (and friends, there are a lot of modules in sys.modules at startup)> is imported, it is different from how readline.so is imported.For the record, os is imported by the _io module: /* put os in the module state */ state->os_module = PyImport_ImportModule("os"); if (state->os_module == NULL) goto fail;(inModules/_io/_iomodule.c)This probably happens before sys.path isadjusted/tweaked/fixed/garbled/whatever. | |||
| msg137828 -(view) | Author: R. David Murray (r.david.murray)*![]() | Date: 2011-06-07 14:28 | |
Yeah, that would be my guess. And readline.so is imported in main at a point where it has decided we are going into interactive mode, which is presumably after all other initialization has taken place, including the path munging.Thus my suggestion that that particular import of readline.so should be special cased... | |||
| msg173544 -(view) | Author: Serhiy Storchaka (serhiy.storchaka)*![]() | Date: 2012-10-22 16:49 | |
This issue was fixed in 3.3, but not in 2.7 or 3.2.$ strace ./python -i </dev/null 2>&1 | grep readlinestat64("/home/serhiy/py/cpython3.3/build/lib.linux-i686-3.3/readline.cpython-33m.so", {st_mode=S_IFREG|0755, st_size=52511, ...}) = 0open("/home/serhiy/py/cpython3.3/build/lib.linux-i686-3.3/readline.cpython-33m.so", O_RDONLY) = 4open("/lib/libreadline.so.6", O_RDONLY) = 4 | |||
| msg173551 -(view) | Author: Antoine Pitrou (pitrou)*![]() | Date: 2012-10-22 18:23 | |
Serhiy, I don't think it's fixed in 3.3, or perhaps I'm misunderstanding what you mean by "fixed". If you create readline.cpython-33m.so in your cwd and then run python, the fake readline will still be loaded instead of the real one.For example (here with 3.4):$ touch readline.cpython-34dm.so$ ./pythonPython 3.4.0a0 (default:2a0c9472c89c, Oct 21 2012, 23:24:06) [GCC 4.5.2] on linuxType "help", "copyright", "credits" or "license" for more information.>>> import readlineTraceback (most recent call last): File "<stdin>", line 1, in <module>ImportError: ./readline.cpython-34dm.so: file too short | |||
| msg173552 -(view) | Author: Antoine Pitrou (pitrou)*![]() | Date: 2012-10-22 18:27 | |
Regardless, I'm not sure what we should do about this issue. Loading readline is obviously provided as a convenience to make the interpreter prompt easier to use. Several of us would probably like to go a bit further and also add tab-completion (seeissue5845).It stands that while the -S and -E option allow to disable any customization a user might have done (which is necessary for e.g. suid scripts), the automatic insertion of '' into sys.path has no way of being undone. | |||
| msg173561 -(view) | Author: Serhiy Storchaka (serhiy.storchaka)*![]() | Date: 2012-10-22 19:10 | |
I understand what happens. Python 3.3+ uses getdents(), not stat() for module search. Therefore stat() is not called for non-existed files.> It stands that while the -S and -E option allow to disable any customization a user might have done (which is necessary for e.g. suid scripts), the automatic insertion of '' into sys.path has no way of being undone.Python used not only for suid scripts. | |||
| msg187597 -(view) | Author: STINNER Victor (vstinner)*![]() | Date: 2013-04-22 22:42 | |
We may add a command line option and/or an environment variable to not add the current directory to sys.path.Changing the current behaviour may break many applications / use cases. | |||
| msg187598 -(view) | Author: STINNER Victor (vstinner)*![]() | Date: 2013-04-22 22:46 | |
> We may add a command line option and/or an environment variable to not add the current directory to sys.path.Oh, this is exactly what the issue#16499 proposes. | |||
| msg252029 -(view) | Author: R. David Murray (r.david.murray)*![]() | Date: 2015-10-01 14:05 | |
This issue was reported again inissue 25288.To summarize: the cwd should only be used for imports *after* the command prompt is displayed, and readline is imported *before* the prompt is displayed but currently is imported from the cwd. This should be fixed. | |||
| msg277330 -(view) | Author: Christian Heimes (christian.heimes)*![]() | Date: 2016-09-24 19:22 | |
Steve took care of the readline import for isolated mode in#28192. We can't change the default behavior. If you want to prevent Python from important files from either cwd, user packages or env vars, you have to use isolated mode. System scripts should use the isolated mode flag, too. | |||
| History | |||
|---|---|---|---|
| Date | User | Action | Args |
| 2022-04-11 14:57:18 | admin | set | github: 56447 |
| 2020-03-06 20:51:30 | brett.cannon | set | status: pending -> open nosy: -brett.cannon |
| 2016-09-24 19:22:02 | christian.heimes | set | status: open -> pending nosy: +christian.heimes,steve.dower messages: +msg277330 superseder:Don't import readline in isolated mode |
| 2015-10-01 23:29:55 | eric.snow | set | nosy: +eric.snow |
| 2015-10-01 14:05:34 | r.david.murray | set | messages: +msg252029 |
| 2015-10-01 14:03:50 | r.david.murray | set | stage: needs patch versions: + Python 3.5, Python 3.6, - Python 3.2, Python 3.3, Python 3.4 |
| 2015-10-01 14:03:10 | r.david.murray | link | issue25288 superseder |
| 2013-04-24 11:43:33 | yaccz | set | nosy: +yaccz |
| 2013-04-22 22:46:17 | vstinner | set | messages: +msg187598 |
| 2013-04-22 22:42:13 | vstinner | set | messages: +msg187597 |
| 2012-10-22 19:10:03 | serhiy.storchaka | set | messages: +msg173561 |
| 2012-10-22 18:27:59 | pitrou | set | nosy: +loewis messages: +msg173552 |
| 2012-10-22 18:23:56 | pitrou | set | messages: +msg173551 versions: + Python 3.3, Python 3.4 |
| 2012-10-22 16:49:57 | serhiy.storchaka | set | versions: - Python 2.6, Python 3.1, Python 3.3 nosy: +serhiy.storchaka,ncoghlan messages: +msg173544 components: + Interpreter Core |
| 2011-06-07 14:28:12 | r.david.murray | set | messages: +msg137828 |
| 2011-06-07 13:21:53 | pitrou | set | messages: +msg137824 |
| 2011-06-07 13:13:00 | r.david.murray | set | messages: +msg137821 |
| 2011-06-07 11:08:03 | pitrou | set | messages: +msg137804 |
| 2011-06-06 16:58:41 | r.david.murray | set | nosy: +brett.cannon,pitrou,vstinner messages: +msg137752 versions: + Python 3.1, Python 2.7, Python 3.2, Python 3.3 |
| 2011-06-06 15:09:15 | Niels.Heinen | set | messages: +msg137741 |
| 2011-06-03 15:42:37 | eric.araujo | set | nosy: +eric.araujo messages: +msg137529 |
| 2011-06-02 15:38:18 | jcea | set | nosy: +jcea |
| 2011-06-02 12:38:20 | r.david.murray | set | nosy: +r.david.murray messages: +msg137475 |
| 2011-06-02 11:57:39 | Niels.Heinen | create | |