Re: [PATCH] Version 6 (2.6.23) Smack: Simplified Mandatory Access Control Kernel

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

--- Al Viro <viro@ftp.linux.org.uk> wrote:

> On Fri, Oct 12, 2007 at 10:01:17PM -0700, Casey Schaufler wrote:
> 
> What do you need smk_sb for?  Looks like dead weight...

Reminant of an abandoned thought process. Cleaning.

> smk_read_load(): obvious seq_file candidate.
> smk_read_cipso(): ditto.
> 
> What protects smk_cipso_written?  BTW, its use on the read side is an
> atrocity...
> 
> smk_write_doi() - WTF would NUL-terminate temp[]?  You run sscanf on
> it...
> smk_write_direct() - ditto
> smk_write_ambient() - ditto
> smk_read_ambient() - who said that you won't get write between strlen()
> and simple_read_from_buffer()?

Eek. The smackfs code does look pretty crufty in the
context of seq_file, doesn't it? I will have a go at
cleaning that up some.

> smk_import_entry():
> > +	for (skp = smack_known; skp != NULL; skp = skp->smk_next)
> > +		if (skp->smk_known == smack)
> > +			break;
> Really?  With smack[] being an auto array?

Nope. What you see there is a flaw, a mistake, a bug.
And a serious memory use problem, too. Fix in test.
Thank you.

> That's from quick look through the read/write in there; I hadn't really
> looked into parsers or deeper into the guts.

Your suggestions regarding seq_file look helpful. Thank you.


Casey Schaufler
casey@schaufler-ca.com