[ python-Bugs-1077106 ] Negative numbers to os.read() cause segfault

[SourceForge.net]

Bugs item #1077106, was opened at 2004-12-02 06:40
Message generated for change (Comment added) made by quiver
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=105470&aid=1077106&group_id=5470

Category: Python Interpreter Core
Group: Python 2.4
Status: Open
Resolution: None
Priority: 5
Submitted By: Jp Calderone (kuran)
Assigned to: Nobody/Anonymous (nobody)
Summary: Negative numbers to os.read() cause segfault

Initial Comment:
Python 2.3.4 (#2, Sep 24 2004, 08:39:09)
[GCC 3.3.4 (Debian 1:3.3.4-12)] on linux2
Type "help", "copyright", "credits" or "license" for
more information.
>>> import sys, os
>>> stdin = sys.stdin.fileno()
>>> os.read(stdin, 0)
''
>>> os.read(stdin, 0)
''
>>> os.read(stdin, -1)
asdkljasd
'asdk\x01\x00\x00\x00\x00\x00'
>>> os.read(stdin, 100)
Segmentation fault
exarkun at boson:~$

This problem persists in Python 2.4, although the
resulting incorrect behavior differs slightly (at least
on my build), as is to be expected of a memory
corrupting bug.

Note that the value returned from os.read(stdin, -1) is
incorrect in addition to the subsequent read segfaulting.


----------------------------------------------------------------------

Comment By: George Yoshida (quiver)
Date: 2004-12-04 20:43

Message:
Logged In: YES 
user_id=671362

On Win2k(Python 2.3.4 & 2.4), I get:
>>> os.read(si, -1)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
OSError: [Errno 12] Not enough space

On Linux(SUSE 9.2 & kernel 2.6.5-7.108-smp & gcc 3.3.3) 
in Python 2.4 debug built, I get:

>>>os.read(si, -1)
asd
Debug memory block at address p=0x4024d6b8:
    31 bytes originally requested
    The 4 pad bytes at p-4 are FORBIDDENBYTE, as expected.
    The 4 pad bytes at tail=0x4024d6d7 are not all 
FORBIDDENBYTE
(0xfb):
        at tail+0: 0x0a *** OUCH
        at tail+1: 0xfb
        at tail+2: 0xfb
        at tail+3: 0xfb
    The block was made by call #10310 to debug 
malloc/realloc.
    Data at p: 00 00 00 00 00 00 00 00 ... ff 00 00 00 00 61 
73 64
Fatal Python error: bad trailing pad byte
Aborted

In a normal built, same as Jp.

----------------------------------------------------------------------

Comment By: Raymond Hettinger (rhettinger)
Date: 2004-12-04 13:29

Message:
Logged In: YES 
user_id=80475

In both Py2.3.4 and Py2.4, I get the following correct
behavior on WinME:

>>> os.read(si, -1)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
OSError: [Errno 22] Invalid argument


----------------------------------------------------------------------

Comment By: James Y Knight (foom)
Date: 2004-12-02 08:11

Message:
Logged In: YES 
user_id=1104715

This appears to be because PyString_FromStringAndSize takes a signed int 
for size, doesn't verify that it is > 0, and then adds it to 
sizeof(PyStringObject):
    op = (PyStringObject *)PyObject_MALLOC(sizeof(PyStringObject) + 
size);

PyObject_MALLOC will fail if given a < 0 size, but, if size is > 
-sizeof(PyStringObject), the object will be allocated, but too small. Then, 
memory gets clobbered.

If it returned NULL like it should, posix_read's error handling would be 
fine.


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=105470&aid=1077106&group_id=5470