On Fri, 03 Aug 2001 12:58:44 EDT, Stephen Smalley said: > Again, euid==0 vs. capabilities is NOT the same as the DAC logic. > It relates to the ability to move the capabilities implementation > out of the base kernel and make it optional, with the alternative > being the traditional superuser tests. That is exactly what LSM > is doing. > > Keep in mind that Linus' comments were a response to the > presentation of SELinux at the Linux Kernel Summit. The > focus was on adding support for implementing additional access > control schemes like MLS, TE, and RBAC, not on replacing the base > DAC logic. He also said he wanted something "generic" and get rid of euid==0 and all that (as one option). Could somebody who knows Linus better than I do ask him what he meant? On the one hand, his comments *were* spurred by SELinux, and as such some of us believe he meant something not TOO revolutionary. On the other hand, what he *wrote* certainly reads (at least to some of us) as being open to discussion of a more complete overhaul of the security model. -- Valdis Kletnieks Operating Systems Analyst Virginia Tech
This archive was generated by hypermail 2b30 : Fri Aug 03 2001 - 10:10:32 PDT