Typically an unbundled product is designed to be installed in a stand-alone, unique location. This product is composed of binaries, shared object dependencies, and associated configuration files. For example, the unbundled product ABC might have the following layout:
Assume that the product is designed for installation under /opt. Normally a user will be required to augment their PATH with /opt/ABC/bin so as to locate the product's binaries. Each binary will locate their dependencies using a hard-coded runpath within the binary. For the application abc it would be:
% dump -Lv abc [1] NEEDED libA.so.1 [2] RPATH /opt/ABC/lib |
and similarly for the dependency libA.so.1 this would be:
% dump -Lv libA.so.1 [1] NEEDED libB.so.1 [2] RPATH /opt/ABC/lib |
This dependency representation works fine until the product is installed in some directory other than the recommended default. If a different installation is created, users must resort to LD_LIBRARY_PATH to run the product's applications. Often this is brought about by the introduction of wrappers for each binary, and sometimes inventive souls have even tried modifying the runpath within the appropriate objects.
The dynamic token $ORIGIN expands to the directory in which an object originated. This feature keys off of an auxiliary vector provided by the kernel to the runtime linker on process start-up (see also getexecname(3C)). Using this technology, we can now redefine our unbundled application to locate its dependencies in terms of $ORIGIN:
% dump -Lv abc [1] NEEDED libA.so.1 [2] RPATH $ORIGIN/../lib |
and the dependency libA.so.1 can also be defined in terms of $ORIGIN:
% dump -Lv libA.so.1 [1] NEEDED libB.so.1 [2] RPATH $ORIGIN |
Therefore, if this product is now installed under /usr/local/ABC, and the user's PATH is augmented with /usr/local/ABC/bin, invocation of the application abc will result in a pathname lookup for its dependencies as follows:
% ldd -s abc find object=libA.so.1; required by abc search path=$ORIGIN/../lib (RPATH from file abc) trying path=/usr/local/ABC/lib/libA.so.1 libA.so.1 => /usr/local/ABC/lib/libA.so.1 find object=libB.so.1; required by /usr/local/ABC/lib/libA.so.1 search path=$ORIGIN (RPATH from file /usr/local/ABC/lib/libA.so.1) trying path=/usr/local/ABC/lib/libB.so.1 libB.so.1 => /usr/local/ABC/lib/libB.so.1 |
The next issue to confront dependency location is how to establish a model whereby one unbundled product might have dependencies on the shared objects of another unbundled product.
For example, the unbundled product XYZ might have dependencies on the product ABC. This dependency can be established by a host package installation script that generates a symbolic link to the installation point of the ABC product, as shown in the following figure.
The binaries and shared objects of the XYZ product can represent their dependencies on the ABC product using the symbolic link as a stable reference point. For the application xyz this would be:
% dump -Lv xyz [1] NEEDED libX.so.1 [2] NEEDED libA.so.1 [3] RPATH $ORIGIN/../lib:$ORIGIN/../ABC/lib |
and similarly for the dependency libX.so.1 this would be:
% dump -Lv libX.so.1 [1] NEEDED libY.so.1 [2] NEEDED libC.so.1 [3] RPATH $ORIGIN:$ORIGIN/../ABC/lib |
Therefore, if this product is now installed under /usr/local/XYZ, its post-install script would be required to establish a symbolic link of:
% ln -s ../ABC /usr/local/XYZ/ABC |
If the user's PATH is augmented with /usr/local/XYZ/bin, then invocation of the application xyz will result in a pathname lookup for its dependencies as follows:
% ldd -s xyz find object=libX.so.1; required by xyz search path=$ORIGIN/../lib:$ORIGIN/../ABC/lib (RPATH from file xyz) trying path=/usr/local/XYZ/lib/libX.so.1 libX.so.1 => /usr/local/XYZ/lib/libX.so.1 find object=libA.so.1; required by xyz search path=$ORIGIN/../lib:$ORIGIN/../ABC/lib (RPATH from file xyz) trying path=/usr/local/XYZ/lib/libA.so.1 trying path=/usr/local/ABC/lib/libA.so.1 libA.so.1 => /usr/local/ABC/lib/libA.so.1 find object=libY.so.1; required by /usr/local/XYZ/lib/libX.so.1 search path=$ORIGIN:$ORIGIN/../ABC/lib \ (RPATH from file /usr/local/XYZ/lib/libX.so.1) trying path=/usr/local/XYZ/lib/libY.so.1 libY.so.1 => /usr/local/XYZ/lib/libY.so.1 find object=libC.so.1; required by /usr/local/XYZ/lib/libX.so.1 search path=$ORIGIN:$ORIGIN/../ABC/lib \ (RPATH from file /usr/local/XYZ/lib/libX.so.1) trying path=/usr/local/XYZ/lib/libC.so.1 trying path=/usr/local/ABC/lib/libC.so.1 libC.so.1 => /usr/local/ABC/lib/libC.so.1 find object=libB.so.1; required by /usr/local/ABC/lib/libA.so.1 search path=$ORIGIN (RPATH from file /usr/local/ABC/lib/libA.so.1) trying path=/usr/local/ABC/lib/libB.so.1 libB.so.1 => /usr/local/ABC/lib/libB.so.1 |
In a secure process (see "Security"), the expansion of the $ORIGIN string is allowed only of it expands to a trusted directory. The occurrence of other relative pathnames poses a security risk.
At first glance, it would appear that a path like $ORIGIN/../lib points to a fixed location; fixed by the location of the executable. Unfortunately, this is not true, and all that is needed to exploit a secure program with $ORIGIN is a writable directory in the same file system.
The following example shows how this could be achieved if $ORIGIN was arbitrarily expanded within a secure process.
% cd /worldwritable/dir/in/same/fs % mkdir bin lib % ln $ORIGIN/bin/program bin/program % cp ~/crooked-libc.so.1 lib/libc.so.1 % bin/program ..... using crooked-libc.so.1 |
The utility crle(1) may be used to specify trusted directories applicable for secure applications to use $ORIGIN. Administrators who use this technique should ensure that the target directories are suitably protected from malicious intrusion.