Bugtraq mailing list archives

Re: gzip TOCTOU file-permissions vulnerability


From: Scott Gifford <sgifford () suspectclass com>
Date: Fri, 15 Apr 2005 01:33:12 -0400

"Mark Senior" <Mark.Senior () gov ab ca> writes:

 

From: Derek Martin [mailto:code () pizzashack org] 
Sent: April 13, 2005 08:50


The open() call is at fault here.  If instead of being called 
with a mode of RW_USER, it is called with the final intended 
access mode, there is no need to later call chmod(), and the 
problem is averted.

One wrinkle - if the file is not intended to have user write permission
on it, and gzip (unzip/cpio/pax...) initially created it with the
intended permissions, there would be no way to then write the file.

In a quick test, this seems not to be true, at least on my Linux
system.  It may be true over NFS. 
    
    [gifford@gifford tmp]$ ls -ld testfile
    ls: testfile: No such file or directory
    [gifford@gifford tmp]$ cat t.c
    #include <stdio.h>
    #include <stdlib.h>
    #include <unistd.h>
    #include <fcntl.h>
     
    void die(char *msg)
    {
      perror(msg);
      exit(1);
    }
     
    int main()
    {
      int fd;
     
      if ((fd = open("testfile",O_CREAT|O_WRONLY,0)) < 0)
        die("open failed");
      if (write(fd,"output\n",7) < 0)
        die("write failed");
      if (close(fd) < 0)
        die("close failed");
     
      return 0;
    }
    [gifford@gifford tmp]$ gcc -Wall t.c
    [gifford@gifford tmp]$ ./a.out
    [gifford@gifford tmp]$ ls -ld testfile
    ----------  1 gifford gifford 7 Apr 15 01:28 testfile
    [gifford@gifford tmp]$ chmod +r testfile
    [gifford@gifford tmp]$ cat testfile
    output
    [gifford@gifford tmp]$

----ScottG.


Current thread: