Memory access bounds checking for a programmable atomic operator
Devices and techniques for memory access bounds checking for a programmable atomic operator are described herein. A processor can execute a programmable atomic operator with a base memory address. The processor can obtain a memory interleave size indicator corresponding to the programmable atomic operator and calculate a contiguous memory address range from the base memory address and the memory interleave size. The processor can then detect that a memory request from the programmable atomic operator is outside the contiguous memory address range and deny the memory request when it is outside of the contiguous memory address range and allow the memory request otherwise.
1. An apparatus comprising:
an interface to connect the apparatus to a memory when in operation; and
processing circuitry that, when in operation, is configured to:
determine that a memory request from an executing program is outside a contiguous memory address range in a memory, the contiguous memory address range defined by a base memory address for the program and an interleave size; and
refuse execution of the memory request based on the memory request being outside the contiguous memory address range in the memory.
2. The method of claim 1 , wherein the program is a programmable atomic operator.
3. The method of claim 2 , wherein the memory is part of the apparatus.
4. The method of claim 1 , wherein the apparatus is a chiplet in a chiplet system.
5. The method of claim 1 , wherein the processing circuitry is configured to receive a request to execute the program, the request including the base memory address.
6. The method of claim 5 , wherein the request includes the interleave size.
7. The method of claim 6 , wherein the interleave size is specified by an indicator that is one of four values.
8. The method of claim 7 , wherein the four values respectively correspond to two-hundred fifty-six bytes, sixteen kilobytes, one megabyte, and sixty-four megabytes.
9. The method of claim 7 wherein the four values are integers between zero and three inclusive.
10. The method of claim 5 , wherein, to refuse execution of the memory request, the processing circuitry is configured to communicate a failure status to an entity that requested execution of the program.
11. The method of claim 1 , wherein the memory interleave size indicator is a bit mask.
12. The method of claim 11 , wherein, to determine that the memory request is outside the contiguous memory address range, the processing circuitry is configured to:
perform an exclusive OR (XOR) operation on the base memory address and a memory address of the memory request to produce a result; and
apply the bit mask to the result to cover lower bits of the result; and
evaluate uncovered bits of the result to find a non-zero bit, the memory request being within the contiguous memory address range when the uncovered bits of the result are all zero, and outside otherwise.
13. A method comprising:
determining that a memory request from an executing program is outside a contiguous memory address range in a memory, the contiguous memory address range defined by a base memory address for the program and an interleave size; and
refusing execution of the memory request based on the memory request being outside the contiguous memory address range in the memory.
14. The method of claim 13 , wherein the program is a programmable atomic operator.
15. The method of claim 14 , wherein the memory is part of a memory device that includes processing circuitry upon which the programmable atomic operator is executing.
16. The method of claim 15 , wherein the memory device is a chiplet in a chiplet system.
17. The method of claim 13 , comprising receiving, at a device that includes the memory, a request to execute the program, the request including the base memory address.
18. The method of claim 17 , wherein the request includes the interleave size.
19. The method of claim 18 , wherein the interleave size is specified by an indicator that is one of four values.
20. The method of claim 19 , wherein the four values respectively correspond to two-hundred fifty-six bytes, sixteen kilobytes, one megabyte, and sixty-four megabytes.
21. The method of claim 19 wherein the four values are integers between zero and three inclusive.
22. The method of claim 17 , wherein refusing execution of the memory request includes communicating a failure status to an entity that requested execution of the program.
23. The method of claim 13 , wherein the memory interleave size indicator is a bit mask.
24. The method of claim 23 , wherein determining that the memory request is outside the contiguous memory address range includes:
performing an exclusive OR (XOR) operation on the base memory address and a memory address of the memory request to produce a result; and
applying the bit mask to the result to cover lower bits of the result; and
evaluating uncovered bits of the result to find a non-zero bit, the memory request being within the contiguous memory address range when the uncovered bits of the result are all zero, and outside otherwise.