kernel-hardening - Re: [PATCH v2012.1] fs: symlink restrictions on sticky directories

Follow @Openwall on Twitter for new release announcements and other news

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

Message-ID: <20120106073456.GB14188@elte.hu>
Date: Fri, 6 Jan 2012 08:34:56 +0100
From: Ingo Molnar <mingo@...e.hu>
To: Rik van Riel <riel@...hat.com>
Cc: Kees Cook <keescook@...omium.org>,
	Nick Bowler <nbowler@...iptictech.com>,
	linux-kernel@...r.kernel.org,
	Alexander Viro <viro@...iv.linux.org.uk>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Federica Teodori <federica.teodori@...glemail.com>,
	Lucian Adrian Grijincu <lucian.grijincu@...il.com>,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Eric Paris <eparis@...hat.com>, Randy Dunlap <rdunlap@...otime.net>,
	Dan Rosenberg <drosenberg@...curity.com>, linux-doc@...r.kernel.org,
	linux-fsdevel@...r.kernel.org, kernel-hardening@...ts.openwall.com
Subject: Re: [PATCH v2012.1] fs: symlink restrictions on sticky directories


* Rik van Riel <riel@...hat.com> wrote:

> On 01/05/2012 03:55 PM, Kees Cook wrote:
> >On Thu, Jan 5, 2012 at 12:08 PM, Nick Bowler<nbowler@...iptictech.com>  wrote:
> 
> >>But this is a brand new feature that changes longstanding behaviour of
> >>various syscalls.  Making it default to enabled is rather mean to users
> >>(since it will tend to get enabled by "oldconfig") and seems almost
> >>guaranteed to cause regressions.
> >
> > I couldn't disagree more. There has been zero evidence of 
> > this change causing anything but regressions in _attacks_. 
> > :P If anything, I think there should be no CONFIG and no 
> > sysctl, and it should be entirely non-optional. But since 
> > this patch needs consensus, I have provided knobs to control 
> > it.
> 
> I agree with you, Kees.
> 
> The behaviour introduced by this patch should produce so few 
> issues, that the new behaviour should probably be on by 
> default.

Up to the point people report regressions.

And yes, I think Kees is perfectly right that the setting of the 
default should be evidence based. (Assuming Al and Linus is fine 
with the whole concept.)

The only specific counter-argument I can see is the spinlock 
performance impact I raised during review. I think we can (and 
should) live with that, and it's probably fixable, BYMMV.

Thanks,

	Ingo

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.