Symbolic execution and program testing
Communications of the ACM
BASE: using abstraction to improve fault tolerance
SOSP '01 Proceedings of the eighteenth ACM symposium on Operating systems principles
An Analysis of Error Behaviour in a Large Storage System
An Analysis of Error Behaviour in a Large Storage System
DART: directed automated random testing
Proceedings of the 2005 ACM SIGPLAN conference on Programming language design and implementation
Automatically Generating Malicious Disks using Symbolic Execution
SP '06 Proceedings of the 2006 IEEE Symposium on Security and Privacy
Using model checking to find serious file system errors
ACM Transactions on Computer Systems (TOCS)
EXPLODE: a lightweight, general system for finding serious storage system errors
OSDI '06 Proceedings of the 7th USENIX Symposium on Operating Systems Design and Implementation - Volume 7
Failure trends in a large disk drive population
FAST '07 Proceedings of the 5th USENIX conference on File and Storage Technologies
FAST'08 Proceedings of the 6th USENIX Conference on File and Storage Technologies
An analysis of data corruption in the storage stack
ACM Transactions on Storage (TOS)
Understanding latent sector errors and how to protect against them
ACM Transactions on Storage (TOS)
SQCK: a declarative file system checker
OSDI'08 Proceedings of the 8th USENIX conference on Operating systems design and implementation
KLEE: unassisted and automatic generation of high-coverage tests for complex systems programs
OSDI'08 Proceedings of the 8th USENIX conference on Operating systems design and implementation
Tolerating file-system mistakes with EnvyFS
USENIX'09 Proceedings of the 2009 conference on USENIX Annual technical conference
S2E: a platform for in-vivo multi-path analysis of software systems
Proceedings of the sixteenth international conference on Architectural support for programming languages and operating systems
FATE and DESTINI: a framework for cloud recovery testing
Proceedings of the 8th USENIX conference on Networked systems design and implementation
Execution generated test cases: how to make systems code crash itself
SPIN'05 Proceedings of the 12th international conference on Model Checking Software
Lightweight snapshots and system-level backtracking
HotOS'13 Proceedings of the 14th USENIX conference on Hot Topics in Operating Systems
On fault resilience of OpenStack
Proceedings of the 4th annual Symposium on Cloud Computing
Ffsck: The Fast File-System Checker
ACM Transactions on Storage (TOS)
Ffsck: the fast file system checker
FAST'13 Proceedings of the 11th USENIX conference on File and Storage Technologies
Checking the integrity of transactional mechanisms
FAST'14 Proceedings of the 12th USENIX conference on File and Storage Technologies
| Hi-index | 0.00 |
File system checkers (like e2fsck) are critical, complex, and hard to develop, and developers today rely on hand-written tests to exercise this intricate code. Test suites for file system checkers take a lot of effort to develop and require careful reasoning to cover a sufficiently comprehensive set of inputs and recovery mechanisms. We present a tool and methodology for testing file system checkers that reduces the need for a specification of the recovery process and the development of a test suite. Our methodology splits the correctness of the checker into two objectives: consistency and completeness of recovery. For each objective, we leverage either the file system checker code itself or a comparison among the outputs of multiple checkers to extract an implicit specification of correct behavior. Our methodology is embodied in a testing tool called SWIFT, which uses a mix of symbolic and concrete execution; it introduces two new techniques: a specific concretization strategy and a corruption model that leverages test suites of file system checkers. We used SWIFT to test the file system checkers of ext2, ext3, ext4, ReiserFS, and Minix; we found bugs in all checkers, including cases leading to data loss. Additionally, we automatically generated test suites achieving code coverage on par with manually constructed test suites shipped with the checkers.