linux-kernel - Re: [PATCH 1/2] ubifs: Remove dead xattr code

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Wed, 26 Aug 2015 09:15:09 -0500
From:	Josh Cartwright <joshc@...com>
To:	Dongsheng Yang <yangds.fnst@...fujitsu.com>
Cc:	Richard Weinberger <richard@....at>, linux-mtd@...ts.infradead.org,
	linux-kernel@...r.kernel.org, linux-fsdevel@...r.kernel.org,
	Subodh Nijsure <snijsure@...d-net.com>,
	Marc Kleine-Budde <mkl@...gutronix.de>,
	Brad Mouring <brad.mouring@...com>,
	Gratian Crisan <gratian.crisan@...com>,
	Artem Bityutskiy <artem.bityutskiy@...ux.intel.com>,
	Artem Bityutskiy <dedekind1@...il.com>
Subject: Re: [PATCH 1/2] ubifs: Remove dead xattr code

On Thu, Aug 20, 2015 at 10:48:38AM +0800, Dongsheng Yang wrote:
> On 08/20/2015 04:35 AM, Richard Weinberger wrote:
> >This is a partial revert of commit d7f0b70d30ffb9bbe6b8a3e1035cf0b79965ef53
> >("UBIFS: Add security.* XATTR support for the UBIFS").
> 
> Hi Richard,
> 	What about a full reverting of this commit. In ubifs, we
> *can* support any namespace of xattr including user, trusted, security
> or other anyone prefixed by any words. But we have a check_namespace()
> in xattr.c to limit what we want to support. That said, if we want to
> "Add security.* XATTR support for the UBIFS", what we need to do is
> just extending the check_namespace() to allow security namespace pass.
> And yes, check_namespace() have been supporting security namespace.

Is this good enough?  Yes, it'd mean that the xattrs end up on disk, but
then who's responsible for invoking the selected LSMs inode_init_security() hooks?
AFAICT, we'd still need to invoke security_inode_init_security for newly
created inodes (which, Richard's proposed commit still does).

Thanks,

  Josh (who, admittedly, is neither a filesystem nor security module guy :)

Download attachment "signature.asc" of type "application/pgp-signature" (474 bytes)