Redundant transactions for detection of timing sensitive errors

Patent metadata

Field	Value
Publication number	US-9928158-B2
Application number	US-201615011557-A
Country	US
Kind code	B2
Filing date	Jan 30, 2016
Priority date	Aug 8, 2013
Publication date	Mar 27, 2018
Grant date	Mar 27, 2018

Abstract

A method for detecting a software-race condition in a program includes copying a state of a transaction of the program from a first core of a multi-core processor to at least one additional core of the multi-core processor, running the transaction, redundantly, on the first core and the at least one additional core given the state, outputting a result of the first core and the at least one additional core, and detecting a difference in the results between the first core and the at least one additional core, wherein the difference indicates the software-race condition.

First claim

What is claimed is:

1. A method for detecting a software-race condition in a program, the method comprising:
copying a state of a transaction of the program from a first core of a multi-core processor to at least one additional core of the multi-core processor;
running the transaction, redundantly, on the first core and the at least one additional core given the state, wherein the transaction is run by at least one of the first core and the at least one additional core with a shifting delay starting the transaction at different times on the first core and the at least one additional core;
outputting a result of the first core and the at least one additional core;
detecting a difference in the results between the first core and the at least one additional core, wherein the difference indicates the software-race condition;
maintaining a log of the difference detected while continuously executing the transaction, wherein the log includes start and end addresses of the transaction;
obtaining a failing case based on the software-race condition;
performing a rollback of the program; and
executing the failing case using a debugger.

2. A non-transitory computer readable medium comprising computer executable instructions which when executed by a computer cause the computer to perform a method for detecting a software-race condition in a program, the method comprising:
copying a state of a transaction of the program from a first core of a multi-core processor to at least one additional core of the multi-core processor;
running the transaction, redundantly, on the first core and the at least one additional core given the state, wherein the transaction is run by at least one of the first core and the at least one additional core with a shifting delay starting the transaction at different times on the first core and the at least one additional core;
outputting a result of the first core and the at least one additional core;
detecting a difference in the results between the first core and the at least one additional core, wherein the difference indicates the software-race condition;
maintaining a log of the difference detected while continuously executing the transaction, wherein the log includes start and end addresses of the transaction;
obtaining a failing case based on the software-race condition;
performing a rollback of the program; and
executing the failing case using a debugger.

3. An apparatus for detecting a software-race condition in a program, the apparatus comprising:
a multi-core processor having at least first and second cores; and
a memory coupled to said multi-core processor;

wherein said multi-core processor and said memory are cooperatively configured to:
copy a state of a transaction of the program from said first core to at least said second core;
run the transaction, redundantly, on the first core and the at least second core given the state, wherein the transaction is run by at least one of the first core and the at least second core with a shifting delay starting the transaction at different times on the first core and the at least second core;
output a result of the first core and the at least second core;
detect a difference in the results between the first core and the at least second core, wherein the difference indicates the software-race condition;
maintain a log of the difference detected while continuously executing the transaction, wherein the log includes start and end addresses of the transaction;
obtain a failing case based on the software-race condition;
perform a rollback of the program; and
execute the failing case using a debugger.

Assignees

• IBM

Inventors

• Cain Iii Harold W
• Daly David M
• Huang Michael C
• Ekanadham Kattamuri
• Moreira Jose E
• Serrano Mauricio J

Classifications

• G06F2201/865

  Monitoring of software · CPC title

• G06F11/3668

  Testing of software · CPC title

• G06F11/3636 Primary

  by tracing the execution of the program · CPC title

• G06F11/1492

  using run-time replication performed by the application software, e.g. N-modular type · CPC title

• G06F11/1497

  Time redundant execution of software on a single processing unit · CPC title