linux-kernel - Re: [PATCH] vfs: fix IMA lockdep circular locking dependency

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]

Message-ID: <CA+55aFyUnwjvS=WGOPxns8F1Ar=pYC3-4XFac41dZ14FFmdEUQ@mail.gmail.com>
Date:	Wed, 16 May 2012 08:23:19 -0700
From:	Linus Torvalds <torvalds@...ux-foundation.org>
To:	Eric Paris <eparis@...isplace.org>
Cc:	Mimi Zohar <zohar@...ux.vnet.ibm.com>,
	James Morris <jmorris@...ei.org>,
	Al Viro <viro@...iv.linux.org.uk>,
	Mimi Zohar <zohar@...ibm.com>,
	linux-security-module@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] vfs: fix IMA lockdep circular locking dependency

On Wed, May 16, 2012 at 7:06 AM, Eric Paris <eparis@...isplace.org> wrote:
>
> But it would be flat wrong to put it there.  True historically we did
> things differently, but this patch is right given Linus is doing it a
> different way.

Indeed. As Eric points out, I'm forcing the address capability checks
on everybody, so defaulting to them is actively the wrong thing to do.

I think the NULL pointer games are from back when we inlined the
security hooks into the code directly, rather than have out-of-line
helper functions. At that point, the code expansion overhead of
loading the pointer, checking it for NULL, and then calling it was
quite noticeable.

But we uninlined them in 2007 (commit 20510f2f4e2d: "security: Convert
LSM into a static interface"), so I think the "set to default so that
you can do an unconditional jump" is not all that useful any more.

Of course, if all security models do export the fn, not having the
conditional is still useful. But defaulting to the capability one is
not in this case, since we will call the capability version later
anyway. So it should be defaulted to something that just returns 0.

                 Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/