Re: Latest episode from "the most unpopular thread on LSM"

sarnoldat_private

On Tue, Sep 25, 2001 at 04:40:38PM -0700, richard offer wrote:
> Not much has changed recently,

*sigh* Including a huge problem I pointed out before...

> I re-worked ptrace() to follow the pre-LSM
> logic to avoid any race conditions that Greg was worried about.

Thanks.

> +	/* messier than it should be to ensure we have a code path identical to 
> +	 * that pre-LSM */

... honestly ...

> ===== security/capability_plug.c 1.66 vs edited =====
> --- 1.66/security/capability_plug.c	Sun Sep 23 18:37:19 2001
> +++ edited/security/capability_plug.c	Tue Sep 25 10:41:04 2001
> @@ -21,27 +21,27 @@
>  /* flag to keep track of how we were registered */
>  static int secondary;
>  
> -static int cap_sethostname (char *hostname)
> +static int cap_sethostname (char *hostname, int kerror)
>  {
>  	return 0;
>  }

And here start the problems.

If you just blanket return 0 from all these authoritative functions, you
have given root away. While I know we all want to push our individual
security modules, neutering the ones that come with the kernel won't
make us real popular. :)

Fix all these, and I think the patch might have potential.

_______________________________________________
linux-security-module mailing list
linux-security-moduleat_private
http://mail.wirex.com/mailman/listinfo/linux-security-module