Taint tracking via non-intrusive bytecode instrumentation
Various embodiments of systems and methods to track tainting information via non-intrusive bytecode instrumentation are described herein. The described techniques include, at one aspect, defining a taint-aware class to shadow an original data class. The taint-aware class includes a payload field to store objects of the original data class, a metadata field to store tainting information corresponding to the objects of the original data class, and a method proxying a corresponding method of the original data class. In another aspect, the instances of the original data class are replaced with corresponding instances of the taint-aware class in an application bytecode. Further, in a yet another aspect, when executing the application in a runtime environment, the method propagates the content of the metadata filed and calls the corresponding method of the original data class to manage the content of the payload field.
1. A computer system to track tainted data flows comprising:
at least one operative memory device to store instructions; and
at least one processor coupled to the at least one operative memory device to execute the instructions to:
generate a taint-aware class comprising:
a payload field to store objects of an original data class;
a metadata field to store tainting information correspondingly assigned to the objects of the original data class, wherein the metadata field comprises a binary string, each bit of the binary string comprising an indication of whether a corresponding character of at least one of the objects is potentially damaging; and
perform a method to:
access the metadata field,
proxy a corresponding method of the original data class to access the payload field, and
replace instances of the original data class of a bytecode of an application with corresponding instances of the taint aware class; and wherein
in response to a call during an execution of the application in a runtime environment, the method of the taint aware class manages the content of the metadata field and calls the corresponding method of the original data class to manage the content of the payload field.
2. The system of claim 1 , wherein the at least one processor to execute the instructions further to:
instrument one or more bytecode files to replace each reference to the original data class with a corresponding reference to the taint-aware class, wherein the one or more bytecode files store the bytecode of the application; and
load the bytecode of the application to the runtime environment from the instrumented one or more bytecode files.
3. The system of claim 1 , wherein the at least one processor to execute the instructions further to:
replace each reference to the original data class with a corresponding reference to the taint-aware class at loading the bytecode of the application to the runtime from one or more bytecode files.
4. The system of claim 3 , wherein the replacing is executed by a configurable agent of the runtime environment.
5. The system of claim 1 , wherein managing the content of the metadata field comprises one or more of:
receiving the metadata from a source;
sending the metadata to a sink; and
propagating the metadata between instances of the taint-aware class.
6. A non-transitory machine-readable medium to store instructions, which when executed by a computer, cause the computer to perform operations comprising:
replace instances of an original data class with instances of a taint aware class in a bytecode of an application, wherein the taint-aware class comprises:
a payload field to store objects of the original data class;
a metadata field to store tainting information correspondingly assigned to the objects of the original data class, wherein the metadata field comprises a binary string, each bit of the binary string comprising an indication of whether a corresponding character of at least one of the objects is potentially damaging; and
perform a method to:
access the metadata field and proxy a corresponding method of the original data class to access the payload field; and
in response to a call during an execution of the application in a runtime environment, the method of the taint aware class is to:
manage the content of the metadata field; and
call the corresponding method of the original data class to manage the content of the payload field.
7. The machine-readable medium of claim 6 , wherein replacing the instances comprises:
instrumenting one or more bytecode files to replace each reference to the original data class with a corresponding reference to the taint-aware class, wherein the one or more bytecode files store the bytecode of the application; and
loading the bytecode of the application to the runtime environment from the instrumented one or more bytecode files.
8. The machine-readable medium of claim 6 , wherein replacing the instances comprises:
replacing each reference to the original data class with a corresponding reference to the taint-aware class at loading the bytecode of the application to the runtime from one or more bytecode files.
9. The machine-readable medium of claim 8 , wherein the replacing is executed by a configurable agent of the runtime environment.
10. The machine-readable medium of claim 6 , wherein the metadata field is of a data type determined by a preconfigured tainting mode.
11. The machine-readable medium of claim 6 , wherein managing the content of the metadata field comprises one or more of:
receiving the metadata from a source;
sending the metadata to a sink; and
propagating the metadata between instances of the taint-aware class.
12. A computerized method to track tainted data flows comprising:
replacing instances of an original data class with instances of a taint aware class in a bytecode of an application, wherein the taint-aware class comprises:
a payload field to store objects of the original data class;
a metadata field to store tainting information correspondingly assigned to the objects of the original data class, wherein the metadata field comprises a binary string, each bit of the binary string comprising an indication of whether a corresponding character of at least one of the objects is potentially damaging; and
performing a method to access the metadata field and proxy a corresponding method of the original data class to access the payload field; and
in response to a call during an execution of the application in a runtime environment, the method of the taint aware class is to:
manage the content of the metadata field; and
call the corresponding method of the original data class to manage the content of the payload field.
13. The method of claim 12 , wherein replacing the instances comprises:
instrumenting one or more bytecode files to replace each reference to the original data class with a corresponding reference to the taint-aware class, wherein the one or more bytecode files store the bytecode of the application; and
loading the bytecode of the application to the runtime environment from the instrumented one or more bytecode files.
14. The method of claim 12 , wherein replacing the instances comprises:
replacing each reference to the original data class with a corresponding reference to the taint-aware class at loading the bytecode of the application to the runtime from one or more bytecode files.
15. The method of claim 14 , wherein the replacing is executed by a configurable agent of the runtime environment.
16. The method of claim 12 , wherein the metadata field is of a data type determined by a tainting mode.
17. The method of claim 12 , wherein managing the content of the metadata field comprises one or more of:
receiving the metadata from a source;
sending the metadata to a sink; and
propagating the metadata between instances of the taint-aware class.