Re: selinux_inode_setsecurity and fsetxattr

[Casey Schaufler <casey@schaufler-ca.com>]

--- Stephen Smalley <sds@tycho.nsa.gov> wrote:

> 
> On Fri, 2008-01-18 at 10:38 -0800, Casey Schaufler wrote:
> > --- Stephen Smalley <sds@tycho.nsa.gov> wrote:
> > 
> > > 
> > > On Fri, 2008-01-18 at 08:40 -0800, Casey Schaufler wrote:
> > > > I was looking in selinux_inode_setsecurity for my own
> > > > neffarious purposes and was curious what prevents a program
> > > > that has a file open from setting the context on the file
> > > > using fsetxattr. I confess that I haven't tried it to see
> > > > how it actually behaves.
> > > 
> > > Sorry, what's your question?  Of course you can relabel a file via
> > > fsetxattr (if you pass the permission check, which is a different hook
> > > called earlier).
> > 
> > In the case of fsetxattr, which hook would that be?
> 
> All three setxattr system calls ultimately call vfs_setxattr(), which
> calls security_inode_setxattr() to check security module permissions.
> If it passes that check, then it proceeds to the actual processing,
> which follows different paths depending on whether the fs implements a
> setxattr operation or not.
> 
> security_inode_setsecurity() is just the low level primitive for setting
> the value in the in-core security structure and can be called either
> from the vfs (as the fallback) or from the individual fs op (as in
> tmpfs).  Originally only the latter before the vfs fallbacks were
> introduced.

Thank you. That was what I was looking for.


Casey Schaufler
casey@schaufler-ca.com

--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.