International Business Machines Corporation

Brad Lee Herold

Nitin Kumar Gupta

Abhishek Bansal

Jayakumar N Sankarannair

Sangram Alapati

Alexandru Adrian Patrascu

Shakti Kapoor

Allon Adir

Gil Eliezer Shurek

John Martin Ludden

Charles Meissner

Pedro Martin-de-Nicolas

Harish P Omkar

A process for processor testing includes dividing a memory space into first and second portions and generating first and second sets of test instructions for the respective first and second portion. The first and second sets each include a plurality of counter increment/branch instruction pairs where each branch instruction of the first set branches to a backward instruction location and each branch instruction of the second set branches to a forward instruction location.

Processor testing using pairs of counter incrementing and branch instructions

A process for processor testing includes generating a set of test instructions having a first portion and a second portion. A first branch instruction is randomly generated for the first portion where the first branch instruction branches to a respective instruction in a second portion by a branching location offset. A second branch instruction is randomly generated for the second portion where the second branch instruction branches to a respective instruction in the first portion by the branching location offset. If additional instructions are to be added to the set of test instructions, a value of the branching location offset is incrementing by a predetermined amount.

Processor testing using randomly generated branch instructions

Processors may be tested according to various implementations. In one general implementation, a process for processor testing may include randomly generating a first plurality of branch instructions for a first portion of an instruction set, each branch instruction in the first portion branching to a respective instruction in a second portion of the instruction set. The process may also include randomly generating a second plurality of branch instructions for the second portion of the instruction set, each branch instruction in the second portion branching to a respective instruction in the first portion of the instruction set. The process may additionally include generating a plurality of instructions to increment a counter when each branch instruction is encountered during execution.

Processor testing

Processors may be tested according to various implementations. In one general implementation, a process for processor testing may include randomly generating a first plurality of branch instructions for a first portion of an instruction set, each branch instruction in the first portion branching to a respective instruction in a second portion of the instruction set, the branching of the branch instructions to the respective instructions being arranged in a sequential manner. The process may also include randomly generating a second plurality of branch instructions for the second portion of the instruction set, each branch instruction in the second portion branching to a respective instruction in the first portion of the instruction set, the branching of the branch instructions to the respective instructions being arranged in a sequential manner. The process may additionally include generating a plurality of instructions to increment a counter when each branch instruction is encountered during execution.

In one implementation, processor testing may include the ability to randomly generate a first plurality of branch instructions for a first portion of an instruction set, each branch instruction in the first portion branching to a respective instruction in a second portion of the instruction set, the branching of the branch instructions to the respective instructions being arranged in a sequential manner. Processor testing may also include the ability to randomly generate a second plurality of branch instructions for the second portion of the instruction set, each branch instruction in the second portion branching to a respective instruction in the first portion of the instruction set, the branching of the branch instructions to the respective instructions being arranged in a sequential manner. Processor testing may additionally include the ability to generate a plurality of instructions to increment a counter when each branch instruction is encountered during execution.

A test code generation technique that replaces instructions having a machine state dependent result with special redirection instructions provides generation of test code in which state dependent execution choices are made without a state model. Redirection instructions cause execution of a handler than examines the machine state and replaces the redirection instruction with a replacement instruction having a desired result resolved in accordance with the current machine state. The instructions that are replaced may be conditional branch instructions and the result a desired execution path. The examination of the machine state permits determination of a branch condition for the replacement instruction so that the next pass of the test code executes along the desired path. Alternatively, the handler can execute a jump to the branch instruction, causing immediate execution of the desired branch path. The re-direction instructions may be illegal instructions, which cause execution of an interrupt handler that performs the replacement.

Two pass test case generation using self-modifying instruction replacement

A hypervisor receives an interrupt that includes a target address and, in turn, branches to an administrating interrupt vector. Next, the administrating interrupt vector determines whether to branch to a piggyback interrupt handler corresponding to a piggyback interrupt vector. Based upon the determination, the hypervisor either branches to the piggyback interrupt handler or to an administrating interrupt handler that corresponds to the administrating interrupt vector.

Interrupt vector piggybacking

A system for stealing interrupt vectors from an operating system. Custom interrupt handler extensions are copied into an allocated block of memory from a kernel module. Also, operating system interrupt handlers are copied into a reserved space in the allocated block of memory from an interrupt vector memory location. In response to copying the operating system interrupt handlers into the reserved space in the allocated block of memory, custom interrupt handlers from the kernel module are copied over the operating system interrupt handlers in the interrupt vector memory location. The custom interrupt handlers after being copied into the interrupt vector memory location handle all interrupts received by the operating system.

Growing community of inventors

Austin, TX, United States of America

Brad Lee Herold