Recovering source code structure from program binaries
Recovering structure from binaries is disclosed. A binary file having components including a plurality of linker objects is received. A cross reference map of linker objects is created. The linker objects are associated based on calls. An address space distance for each call is determined. Boundaries are defined in the cross reference map based on the address space distance. Subsets are defined as portions of the file that are separated by the boundaries.
1. A system, comprising:
a set of one or more processors programmed with instructions to:
receive a pre-existing binary file previously built from a plurality of intermediate components previously created from source code;
create a cross reference map of at least some of the intermediate components within the pre-existing binary file, wherein the intermediate components are associated based on calls;
determine an address space distance for each call in the cross reference map;
define a set of estimated boundaries in the cross reference map by at least one of (1) inserting links in the cross reference map based on the address space distance and (2) pruning links in the cross reference map based on the address space distance;
define subsets as portions of the pre-existing binary file that are separated by the set of estimated boundaries; and
provide as output, a listing of at least some of the defined subsets, wherein at least some of the output subsets correspond to program structure; and
a memory coupled to the set of one or more processors and configured to provide the set of one or more processors with instructions.
2. The system of claim 1 wherein the cross reference map comprises a function call graph.
3. The system of claim 1 wherein the cross reference map comprises function-data references.
4. The system of claim 1 wherein one or more intermediate components included in the pre-existing binary file are full linker objects.
5. The system of claim 1 wherein one or more intermediate components included in the pre-existing binary file are a necessary subset of code and data present in a full linker object.
6. The system of claim 1 wherein at least one estimated boundary comprises a linker object boundary.
7. The system of claim 1 wherein at least one estimated boundary comprises a static library function boundary.
8. The system of claim 1 wherein the set of one or more processors is further configured to recover function scope information.
9. The system of claim 1 wherein the set of one or more processors is further configured to recover data scope information.
10. The system of claim 1 wherein the set of one or more processors is further configured to recover data boundaries for linker objects.
11. The system of claim 1 wherein the set of one or more processors is further configured to recover data boundaries for static libraries.
12. The system of claim 1 wherein the set of one or more processors is further configured to recover source code layout.
13. The system of claim 1 wherein the set of one or more processors is further configured to recover linker command line order.
14. The system of claim 1 wherein the set of one or more processors is further configured to recover a list of functions that are part of an API.
15. The system of claim 1 wherein the set of one or more processors is further configured to recover boundaries at least in part by using a clustering technique.
16. The system of claim 1 , wherein the set of one or more processors is further configured to define the set of estimated boundaries at least in part by identifying articulation nodes in a graph.
17. The system of claim 1 , wherein the set of one or more processors is further configured to define the set of estimated boundaries at least in part by computing a minimum cut in a graph.
18. A method, comprising:
receiving a pre-existing binary file previously built from a plurality of intermediate components previously created from source code;
creating a cross reference map of at least some of the intermediate components within the pre-existing binary file, wherein the intermediate components are associated based on calls;
determining an address space distance for each call in the cross reference map;
defining, using a set of one or more processors, a set of estimated boundaries in the cross reference map by at least one of (1) inserting links in the cross reference map based on the address space distance and (2) pruning links in the cross reference map based on the address space distance;
defining subsets as portions of the pre-existing binary file that are separated by the set of estimated boundaries; and
providing as output, a listing of at least some of the defined subsets, wherein at least some of the output subsets correspond to program structure.
19. The method of claim 18 wherein the cross reference map comprises a function call graph.
20. The method of claim 18 wherein the cross reference map comprises function-data references.