This is the mail archive of the libc-alpha@sourceware.cygnus.com mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Ad PR libc/1730: glibc bug in memmem()



The appended bug report has been reported for glibc's memmem
implementation.

I propose the appended patch.  What do you think?

Andreas

2000-05-17  Andreas Jaeger  <aj@suse.de>

	* sysdeps/generic/memmem.c (memmem): Check for invalid parameter.
	Closes PR libc/1730, reported by Greg Hudson <ghudson@mit.edu>.

============================================================
Index: sysdeps/generic/memmem.c
--- sysdeps/generic/memmem.c	1998/01/30 13:39:22	1.10
+++ sysdeps/generic/memmem.c	2000/05/17 07:42:03
@@ -1,4 +1,4 @@
-/* Copyright (C) 1991, 92, 93, 94, 96, 97, 98 Free Software Foundation, Inc.
+/* Copyright (C) 1991,92,93,94,96,97,98,2000 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
 
    The GNU C Library is free software; you can redistribute it and/or
@@ -37,6 +37,11 @@
     /* The first occurrence of the empty string is deemed to occur at
        the beginning of the string.  */
     return (void *) haystack;
+
+  /* Sanity check, otherwise the loop will search through the whole
+     memory.  */
+  if (haystack_len < needle_len)
+    return NULL;
 
   for (begin = (const char *) haystack; begin <= last_possible; ++begin)
     if (begin[0] == ((const char *) needle)[0] &&


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Greg Hudson <ghudson@mit.edu>
To: bugs@gnu.org
Subject: glibc bug in memmem()

>Number:         1730
>Category:       libc
>Synopsis:       glibc bug in memmem()
>Confidential:   yes
>Severity:       serious
>Priority:       medium
>Responsible:    libc-gnats
>State:          open
>Class:          sw-bug
>Submitter-Id:   unknown
>Arrival-Date:   Wed May 10 21:50:01 EDT 2000
>Last-Modified:  Wed May 17 03:20:19 EDT 2000
>Originator:     
>Organization:
>Release:        
>Environment:
>Description:
>How-To-Repeat:
>Fix:
>Audit-Trail:

Responsible-Changed-From-To: gnats-admin->libc-gnats
Responsible-Changed-By: jaeger
Responsible-Changed-When: Wed May 17 03:20:10 2000
Responsible-Changed-Why:
This is a glibc bug.

Andreas
>Unformatted:
(I'd use the glibcbug shell script, but that only exists if you've
built glibc, which I haven't; I merely cribbed some of its code.)

memmem() begins with the following computation:

        last_possible = (const char *) haystack + haystack_len - needle_len;

This computation computes an invalid pointer value if haystack_len is
less than needle_len.  In particular, if haystack is 0 and
haystack_len is 0 (which is perfectly valid and reasonable on a system
where malloc(0) returns 0), then the value computed will be some large
pointer, and memmem() will dereference invalid memory in its loop.

You need a special-case check for haystack_len < needle_len.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.inka.de


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]