Choices (class hierarchical open interface for custom embedded systems)
ACM SIGOPS Operating Systems Review
The Rio file cache: surviving operating system crashes
Proceedings of the seventh international conference on Architectural support for programming languages and operating systems
Improving the reliability of commodity operating systems
ACM Transactions on Computer Systems (TOCS)
Nooks: an architecture for reliable device drivers
EW 10 Proceedings of the 10th workshop on ACM SIGOPS European workshop
ACM Transactions on Computer Systems (TOCS)
CuriOS: improving reliability through operating system structure
OSDI'08 Proceedings of the 8th USENIX conference on Operating systems design and implementation
Hi-index | 0.00 |
The current researches to improve the reliability of operating systems have been focusing on the evolution of kernel architecture or protecting against device driver errors. In particularly, the device driver errors are critical to the most of the complementary operating systems that have a kernel level device driver. Especially on special purpose embedded system, because of its limited resources and variety of devices, more serious problems are induced. Preventing data corruption or blocking the arrogation of operational level is not enough to cover the entire problems. For examples, when using device drivers, the violation of function's call sequence can cause a malfunction. Also a violation of behavior rules on system level involves the same problem. This type of errors is difficult to be detected by the previous methods. Accordingly, we designed a system that traces system behavior at runtime and recovers optimally when errors are detected. We experimented in Linux 2.6.24 kernel operating on GP2X-WIZ mobile game player.