Title :
Software-based control flow error detection and correction using branch triplication
Author :
Ghalaty, Nahid Farhady ; Fazeli, Mahdi ; Rad, Hossein Izadi ; Miremadi, Seyed Ghassem
Author_Institution :
Dept. of Comput. Eng., Sharif Univ. of Technol., Tehran, Iran
Abstract :
Ever Increasing use of commercial off-the-shelf (COTS) processors to reduce cost and time to market in embedded systems has brought significant challenges in error detection and recovery methods employing in such systems. This paper presents a software based control flow error detection and correction technique, so called branch TMR (BTMR), suitable for use in COTS-based embedded systems. In BTMR method, each branch instruction is triplicated and a software interrupt routine is invoked to check the correctness of the branch instruction. During the execution of a program, when a branch instruction is executed, it is compared with the second redundant branch in the interrupt routine. If a mismatch is detected, the third redundant branch instruction is considered as the error-free branch instruction; otherwise, no error has occurred. The main advantage of BTMR over previously proposed control flow checking (CFC) methods is its ability to correct CFEs as well as protection of indirect branch instructions. The BTMR method is evaluated on LEON2 embedded processor. The results show that, error correction coverage is about 96%, while memory overhead and performance overhead of BTMR is about 28% and 10%, respectively.
Keywords :
embedded systems; error detection; software engineering; time to market; branch TMR; branch triplication; commercial off-the-shelf processor; control flow checking method; embedded system; recovery method; redundant branch instruction; software based control flow error detection; software interrupt routine; time to market; Data structures; Error correction; Fault tolerance; Fault tolerant systems; Program processors; Registers; COTS processor; Control Flow Checking; error correction; fault tolerance; software-based error detection;
Conference_Titel :
On-Line Testing Symposium (IOLTS), 2011 IEEE 17th International
Conference_Location :
Athens
Print_ISBN :
978-1-4577-1053-7
DOI :
10.1109/IOLTS.2011.5993847