Re: Question about security system call in LSM ?

From: Stephen Smalley (sds@private)
Date: Mon Apr 18 2005 - 05:13:46 PDT


On Sun, 2005-04-17 at 20:22 -0700, Dave Peterson wrote:
> It looks to me like the main reason the syscall was removed has to do
> with architectures in which the kernel executes in 64-bit mode and
> user-space code may execute in 32-bit mode.  I think x86_64, ia64,
> and sparc64 is a comprehensive list of such architectures that are
> supported by Linux; is this correct?

IMHO, the more important reason was that it opened up a gateway for
arbitrary interfaces into the kernel without peer review, much like
ioctl.  The trend in kernel development seems to be away from ioctl-like
interfaces and toward much more well-defined interfaces, with a
particular emphasis on implementing new interfaces via pseudo
filesystems rather than syscalls.  For SELinux, we re-implemented (a
subset of) its API via a combination of a new /proc/pid/attr API for
process security attribute manipulation, the existing xattr API (with
the addition of a new security namespace) for file security attribute
manipulation, and a selinuxfs pseudo filesystem for SELinux-specific
interfaces.  See 
http://www.nsa.gov/selinux/papers/module/x360.html

-- 
Stephen Smalley <sds@private>
National Security Agency



This archive was generated by hypermail 2.1.3 : Mon Apr 18 2005 - 05:22:58 PDT