A synchronization mechanism that communicates information from a writer to a reader without permitting information flow in the reverse direction is presented. The synchronization mechanism takes advantage of a priori knowledge of the semantics of communicated information and is optimal because it does not require blocking, busy wait states, an unbounded number of rereads of data, or inefficient use of system resources. The mechanism has been implemented in the Trusted Mach (TMach) operating system's file system. The TMach's file system is described, and it is shown how performance is enhanced through the use of the synchronization mechanism
Published in:
Research in Security and Privacy, 1992. Proceedings., 1992 IEEE Computer Society Symposium on
Date of Conference: 4-6 May 1992