Granted Patent B2

US 9,870,466 · App. 14/866,928 · Granted Jan 16, 2018

Abstract

Claims (31)

1. A computing apparatus, comprising:

first one or more logic elements, including at least a hardware logic element, comprising a module providing executable code including a cryptographically signed private routine having a public entry point, the public entry point configured to be accessible by an unprivileged process, the private routine having associated therewith a branching policy comprising a requirement that entry to the private routine be from a signed entry point; and

second one or more logic elements, including at least a hardware logic element, comprising a policy engine, operable to:

receive a first branching instruction to the private routine;

determine that the first branching instruction is not a signed entry point for the private routine, comprising a branch check selected from predictive branch checking, comprising evaluating the branching policy before branching, and reactive branch checking, comprising evaluating the branching policy after branching; and

take a security action.

2. The computing apparatus of claim 1 , wherein the module further comprises a public routine, and wherein the policy engine is further operable to receive a second branching instruction from an external source to the module, and allow the second branching instruction to execute.

3. The computing apparatus of claim 1 , wherein the policy engine is further operable to receive a third branching instruction to the private routine, determine that the third branching instruction comes from an authorized source internal to the module, and allow the third branching instruction to execute.

4. The computing apparatus of claim 1 , wherein the policy engine is implemented within a trusted execution environment (TEE).

5. The computing apparatus of claim 1 , wherein the security action comprises throwing an exception.

6. The computing apparatus of claim 1 , wherein the branching policy comprises a private routine designation.

7. The computing apparatus of claim 1 , wherein the branching policy comprises a public routine designation.

8. The computing apparatus of claim 1 , wherein the policy engine is implemented at least partly in processor hardware.

9. The computing apparatus of claim 1 , wherein the policy engine is implemented at least partly in a hypervisor or memory controller.

10. The computing apparatus of claim 1 , wherein determining that the first branching instruction is not a signed entry point for the private routine comprises determining that a calling procedure of the first branching instruction does not include a call-to marker for the routine.

11. One or more tangible, non-transitory computer-readable storage mediums having stored thereon executable instructions for providing a policy engine, operable to:

execute a module providing executable code, including a cryptographically signed private routine having a public entry point, the public entry point configured to be accessible by an unprivileged process, the private routine having associated therewith a branching policy comprising a requirement that entry to the private routine be from a signed entry point;

receive a first branching instruction to the private routine;

determine that the first branching instruction is not a signed entry point for the private routine, comprising a branch check selected from predictive branch checking, comprising evaluating the branching policy before branching, and reactive branch checking, comprising evaluating the branching policy after branching; and

take a security action.

12. The one or more tangible, non-transitory computer-readable storage mediums of claim 11 , wherein the module further comprises a public routine, and wherein the policy engine is further operable to receive a second branching instruction from an external source to the public routine, and allow the second branching instruction to execute.

13. The one or more tangible, non-transitory computer-readable storage mediums of claim 11 , wherein the branching policy comprises a private routine designation.

14. The one or more tangible, non-transitory computer-readable storage mediums of claim 11 , wherein the branching policy comprises a public routine designation.

15. The one or more tangible, non-transitory computer-readable storage mediums of claim 11 , comprising processor hardware instructions or microcode.

16. The one or more tangible, non-transitory computer-readable storage mediums of claim 11 , comprising a hypervisor or memory controller.

17. The one or more tangible, non-transitory computer-readable storage mediums of claim 11 , wherein determining that the first branching instruction is not a signed entry point for the private routine comprises determining that a calling procedure of the first branching instruction does not include a call-to marker for the routine.

18. A computer-implemented method of providing a policy engine for branching instructions, comprising:

executing a module providing executable code, including a cryptographically signed private routine having a public entry point, the public entry point configured to be accessible by an unprivileged process, the private routine having associated therewith a branching policy comprising a requirement that entry to the private routine be from a signed entry point;

receiving a first branching instruction to the private routine;

determining that the first branching instruction is not a signed entry point for the private routine, comprising a branch check selected from predictive branch checking, comprising evaluating the branching policy before branching, and reactive branch checking, comprising evaluating the branching policy after branching; and

taking a security action.

Assignments (21)