Date: Wed, 5 Jul 2017 21:18:58 +0200
From: Willy Tarreau <w@1wt.eu>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Ben Hutchings <ben@decadent.org.uk>, Michal Hocko <mhocko@kernel.org>,
        Hugh Dickins <hughd@google.com>, Oleg Nesterov <oleg@redhat.com>,
        "Jason A. Donenfeld" <Jason@zx2c4.com>, Rik van Riel <riel@redhat.com>,
        Larry Woodman <lwoodman@redhat.com>,
        "Kirill A. Shutemov" <kirill@shutemov.name>,
        Tony Luck <tony.luck@intel.com>,
        "James E.J. Bottomley" <jejb@parisc-linux.org>,
        Helge Diller <deller@gmx.de>, James Hogan <james.hogan@imgtec.com>,
        Laura Abbott <labbott@redhat.com>, Greg KH <greg@kroah.com>,
        "security@kernel.org" <security@kernel.org>,
        Qualys Security Advisory <qsa@qualys.com>,
        LKML <linux-kernel@vger.kernel.org>, Ximin Luo <infinity0@debian.org>
Subject: Re: [PATCH] mm: larger stack guard gap, between vmas
Message-ID: <20170705191858.GF24459@1wt.eu>
References: <20170704093538.GF14722@dhcp22.suse.cz>
 <20170704094728.GB22013@1wt.eu>
 <20170704104211.GG14722@dhcp22.suse.cz>
 <20170704113611.GA4732@decadent.org.uk>
 <1499209315.2707.29.camel@decadent.org.uk>
 <CA+55aFzjibuaV=GBZQN8KOaCpf1P3B1abJCYGaxEM9RVpf9fXg@mail.gmail.com>
 <1499257180.2707.34.camel@decadent.org.uk>
 <CA+55aFzV+CEtFhj-+t31249rcEud3DQqqY4fwJYsO3oT0RJ_PQ@mail.gmail.com>
 <20170705185937.GB24459@1wt.eu>
 <CA+55aFz1t2CHY5wpsd2K2Wi9DaY8S_+on4V+XU2cCkxmtBA3oA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CA+55aFz1t2CHY5wpsd2K2Wi9DaY8S_+on4V+XU2cCkxmtBA3oA@mail.gmail.com>
User-Agent: Mutt/1.6.1 (2016-04-27)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 824
Lines: 20

On Wed, Jul 05, 2017 at 12:17:20PM -0700, Linus Torvalds wrote:
> On Wed, Jul 5, 2017 at 11:59 AM, Willy Tarreau <w@1wt.eu> wrote:
> >
> > Don't you think that the option of having a sysctl to relax the check
> > per task wouldn't be easier for distros and safer overall ? Ie, emit
> > a warning the first time the gap is hit instead of segfaulting, then
> > reduce it to something that used to work (4k or 64k, I don't remember)
> > and try again ?
> 
> It used to be just 4k.
> 
> .. and I think that might be a valid way to find these things, but
> would it be safer? It basically disables the new stack gap entirely
> apart from the warning.

But only if the sysctl is set. It can simply be recommended to set it
if any program fails. We've done this for many years with other ones
like min_mmap_addr or tcp_ecn.

Willy