Date: Tue, 22 Jan 2008 01:30:54 GMT From: Robert Watson <rwatson@FreeBSD.org> To: Perforce Change Reviews <perforce@freebsd.org> Subject: PERFORCE change 133832 for review Message-ID: <200801220130.m0M1Usav046211@repoman.freebsd.org>
next in thread | raw e-mail | index | archive | help
http://perforce.freebsd.org/chv.cgi?CH=133832 Change 133832 by rwatson@rwatson_freebsd_capabilities on 2008/01/22 01:29:57 Define cap_enter(2), a system call to enter "capability mode", in which a process is prevented from accessing system resources via global namespaces, and must instead use only file descriptors and capabilities exposed prior to the call to cap_enter(2). Not yet implemented, need to think a lot about the semantics. Affected files ... .. //depot/projects/trustedbsd/capabilities/src/sys/kern/sys_capability.c#4 edit .. //depot/projects/trustedbsd/capabilities/src/sys/kern/syscalls.master#4 edit .. //depot/projects/trustedbsd/capabilities/src/sys/sys/capability.h#6 edit Differences ... ==== //depot/projects/trustedbsd/capabilities/src/sys/kern/sys_capability.c#4 (text+ko) ==== @@ -54,10 +54,11 @@ * - Should there be a privilege to expand capability rights? * - Should different underlying object sets have different valid capability * rights? I.e., CAP_SOCK_*, CAP_VNODE_*, etc? + * - cap_enter(2) is not implemented. */ #include <sys/cdefs.h> -__FBSDID("$P4: //depot/projects/trustedbsd/capabilities/src/sys/kern/sys_capability.c#3 $"); +__FBSDID("$P4: //depot/projects/trustedbsd/capabilities/src/sys/kern/sys_capability.c#4 $"); #include <sys/param.h> #include <sys/capability.h> @@ -161,6 +162,18 @@ } /* + * Enter capability mode for the process. + */ +int +cap_enter(struct thread *td, struct cap_enter_args *uap) +{ + + /* XXXRW: Not implemented. */ + + return (0); +} + +/* * Create a new capability reference to either an existing file object or an * an existing capability. */ ==== //depot/projects/trustedbsd/capabilities/src/sys/kern/syscalls.master#4 (text+ko) ==== @@ -853,5 +853,6 @@ 484 AUE_NULL STD { int cap_new(int fd, u_int64_t rights); } 485 AUE_NULL STD { int cap_getrights(int fd, \ u_int64_t *rightsp); } +486 AUE_NULL STD { int cap_enter(void); } ; Please copy any additions and changes to the following compatability tables: ; sys/compat/freebsd32/syscalls.master ==== //depot/projects/trustedbsd/capabilities/src/sys/sys/capability.h#6 (text+ko) ==== @@ -23,7 +23,7 @@ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * - * $P4: //depot/projects/trustedbsd/capabilities/src/sys/sys/capability.h#5 $ + * $P4: //depot/projects/trustedbsd/capabilities/src/sys/sys/capability.h#6 $ */ /* @@ -120,6 +120,18 @@ #else /* !_KERNEL */ /* + * cap_enter(): Cause the process to enter capability mode, which will + * prevent it from directly accessing global namespaces. System calls will + * be limited to those performed on file descriptors, and subject to the + * restrictions imposed by the objects referenced and the rights specified in + * the file descriptor and possibly a protecting capability. If already in + * the capability mode, a no-op. + * + * XXXRW: This isn't implemented yet. + */ +int cap_enter(void); + +/* * cap_new(): Create a new capability derived from an existing file * descriptor with the specified rights. If the existing file descriptor is * a capability, then the new rights must be a subset of the existing rights.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200801220130.m0M1Usav046211>