One other topic that we didn't really discuss at the Usenix Security BOF was the when and how of submitting the LSM kernel patch to the kernel developers for initial consideration. At present, the most critical tasks I know of that should probably precede the initial submission are: 1) Resolve the authoritative hooks issue. 2) Resolve the syscall interface issue. 3) Add hooks to the Unix domain socket code to allow control over socket IPC using the abstract namespace (currently under investigation here). See my explanation of the issue from http://mail.wirex.com/pipermail/linux-security-module/2001-August/001665.html. 4) Measure the performance overhead of LSM and LSM+capabilities relative to the unmodified Linux kernel. 5) Write up a little documentation about LSM as a whole and about the individual hooks (explain the origins and rationale for LSM as a whole; explain the rationale for each hook; explain the idea behind the capabilities module, its current state, and its possibilities for future work, e.g. that we could even move the capability bits themselves out of the base kernel, but didn't in the initial version to more easily support composing with the capabilities plug). -- Stephen D. Smalley, NAI Labs ssmalleyat_private _______________________________________________ linux-security-module mailing list linux-security-moduleat_private http://mail.wirex.com/mailman/listinfo/linux-security-module
This archive was generated by hypermail 2b30 : Wed Aug 22 2001 - 07:49:50 PDT