Extracting code patches from binary code for independent fuzz testing

1 . A method comprising:

obtaining a binary code of an embedded device;

performing static analysis of the binary code to identify a potentially vulnerable code region, wherein the potentially vulnerable code region is associated with providing input or output from code units of the binary code;

identifying propagation paths associated with the potentially vulnerable code region in the binary code;

analyzing the propagation paths to identify a potentially exploitable propagation path, wherein the potentially exploitable propagation path is potentially affected by external input that is inputted to a device executing the binary code;

determining a code patch associated with the potentially exploitable propagation path;

performing fuzz testing on the code patch independently from the binary code, whereby detecting vulnerabilities of the code patch; and

wherein said performing the fuzz testing comprises:

generating an emulation of the code patch independently from the binary code;

detecting vulnerabilities of the code patch by fuzz testing the emulation;

wherein the static analysis comprises:

determining a characterization of the potentially vulnerable code region based on influences between the potentially vulnerable code region and at least another code region in the binary code, wherein the characterization of the potentially vulnerable code region indicates a role of the potentially vulnerable code region in the binary code;

based on the characterization of the potentially vulnerable code region, determining that the potentially vulnerable code region is potentially affected by external input that is inputted to the device executing the binary code;

wherein said identifying the potentially exploitable propagation path comprises:

determining for the potentially vulnerable code region a corresponding potentially exploitable propagation path within the binary code, wherein the corresponding potentially exploitable propagation path indicates a reachability of the potentially vulnerable code region within the binary code;

locating in the binary code the code patch, wherein the code patch is associated with a functionality of interest of the binary code, wherein the code patch is associated with the potentially exploitable propagation path of the code region, wherein the code patch comprises the code region, wherein the code patch can be executed independently from the binary code.

2 . The method of claim 1 , wherein the emulation is generated based on the characterization of the code region indicating the role of the code region in the binary code, wherein the emulation is generated based on influences between the code region and other code regions in the binary code.

3 . The method of claim 1 wherein the code patch comprises at least a first code element from a first code unit associated with a first hierarchy level of the binary code and a second code element from a second code unit associated with a second hierarchy level of the binary code, whereby crossing through different code hierarchy levels, wherein the code patch excludes at least a third code element of the first code unit.

4 . The method of claim 3 , wherein the first code unit is a first function, wherein the second code unit is a second function, whereby the code patch comprises at least two portions of two separate functions of the binary code.

5 . The method of claim 3 , wherein the first code unit is a first code file, wherein the second code unit is a second code file, whereby the code patch comprises at least two portions of two separate files of the binary code.

6 . The method of claim 3 , wherein each of the influences comprises an influence between an influencing code region and an influenced code region, wherein the influence indicates a change of a first value that is associated with the influenced code region with a second value that is associated with the influencing code region.

7 . The method of claim 1 , comprising determining the characterization of the code region to indicate that the code region is an input, in response to a determination that the code region only influences other code regions without being influenced by the other code regions, and determining the characterization of the code region to indicate that the code region is an output in response to a determination that the code region is only influenced by other code region s without influencing the other code regions.

8 . The method of claim 1 , wherein the code patch comprises one or more code regions that match a pattern of interest.

9 . The method of claim 1 , wherein the code patch comprises the potentially exploitable propagation path.

10 . The method of claim 1 , wherein the code region comprises at least one of: a function argument, a return value, a condition branch, a global variable, and an offset of a buffer.

11 . The method of claim 1 , wherein the code region is characterized by a role thereof in the binary code.

12 . The method of claim 1 , wherein the potentially exploitable propagation path indicates a reachability of the code region within the binary code.

13 . The method of claim 1 , wherein the code patch excludes at least one code element from the binary code that is not utilized for performing the functionality of interest of the code patch.

14 . A system comprising a processor and coupled memory, the processor being adapted to:

obtain a binary code of an embedded device;

performing static analysis of the binary code to identify a potentially vulnerable code region, wherein the potentially vulnerable code region is associated with providing input or output from code units of the binary code;

identify propagation paths associated with the potentially vulnerable code region in the binary code;

analyze the propagation paths to identify a potentially exploitable propagation path, wherein the potentially exploitable propagation path is potentially affected by external input that is inputted to a device executing the binary code;

determine a code patch associated with the potentially exploitable propagation path;

perform fuzz testing on the code patch independently from the binary code, whereby detecting embedded code vulnerabilities; and

wherein said perform the fuzz testing comprises:

generating an emulation of the code patch independently from the binary code;

detecting vulnerabilities of the code patch by fuzz testing the emulation;

wherein the static analysis comprises:

determining a characterization of the code region based on influences between the code region and at least another code region in the binary code, wherein the characterization of the code region indicates a role of the code region in the binary code;

based on the characterization of the code region, determining that the code region is potentially affected by external input that is inputted to the device executing the binary code;

wherein said identifying the potentially exploitable propagation path comprises:

determining for the code region a corresponding potentially exploitable propagation path within the binary code, wherein the corresponding potentially exploitable propagation path indicates a reachability of the code region within the binary code;

locating in the binary code the code patch, wherein the code patch is associated with a functionality of interest of the binary code, wherein the code patch is associated with the potentially exploitable propagation path of the code region, wherein the code patch comprises the code region, wherein the code patch can be executed independently from the binary code.

15 . The system of claim 14 , wherein the emulation is generated based on a characterization of the code region indicating the role of the code region in the binary code, wherein the emulation is generated based on influences between the code region and other code regions in the binary code.

16 . A computer program product comprising a non-transitory computer readable storage medium retaining program instructions, which program instructions when read by a processor, cause the processor to perform: obtaining a binary code of an embedded device;

performing static analysis of the binary code to identify a potentially vulnerable code region, wherein the potentially vulnerable code region is associated with providing input or output from code units of the binary code;

identifying propagation paths associated with the potentially vulnerable code region in the binary code;

analyzing the propagation paths to identify a potentially exploitable propagation path, wherein the potentially exploitable propagation path is potentially affected by external input that is inputted to a device executing the binary code;

determining a code patch associated with the potentially exploitable propagation path;

performing fuzz testing on the code patch independently from the binary code, whereby detecting embedded code vulnerabilities; and

wherein said performing the fuzz testing comprises:

generating an emulation of the code patch independently from the binary code;

detecting vulnerabilities of the code patch by fuzz testing the emulation;

wherein the static analysis comprises:

determining a characterization of the code region based on influences between the code region and at least another code region in the binary code, wherein the characterization of the code region indicates a role of the code region in the binary code;

based on the characterization of the code region, determining that the code region is potentially affected by external input that is inputted to the device executing the binary code;

wherein said identifying the potentially exploitable propagation path comprises:

determining for the code region a corresponding potentially exploitable propagation path within the binary code, wherein the corresponding potentially exploitable propagation path indicates a reachability of the code region within the binary code;

locating in the binary code the code patch, wherein the code patch is associated with a functionality of interest of the binary code, wherein the code patch is associated with the potentially exploitable propagation path of the code region, wherein the code patch comprises the code region, wherein the code patch can be executed independently from the binary code.