Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20190331214020.836098943@linutronix.de> <20190331215136.039902969@linutronix.de>
 <alpine.DEB.2.21.1904021214340.1676@nanos.tec.linutronix.de>
 <20190402154329.scp7i7uqevubgwrz@treble> <alpine.DEB.2.21.1904021746080.1676@nanos.tec.linutronix.de>
 <7789E14E-C623-4DB7-B076-76B931957C9C@amacapital.net> <alpine.DEB.2.21.1904022027170.1812@nanos.tec.linutronix.de>
 <alpine.DEB.2.21.1904022127210.1812@nanos.tec.linutronix.de>
 <6205D576-694A-4C7D-B1B7-A9FF2E1F9E77@amacapital.net> <alpine.DEB.2.21.1904031823100.1967@nanos.tec.linutronix.de>
 <alpine.DEB.2.21.1904032038500.1833@nanos.tec.linutronix.de>
In-Reply-To: <alpine.DEB.2.21.1904032038500.1833@nanos.tec.linutronix.de>
From:   Andy Lutomirski <luto@kernel.org>
Date:   Wed, 3 Apr 2019 17:03:18 -0700
Message-ID: <CALCETrWfSytPj4a34DWFi+R3v+xGsaDuLg75GQ0n-5-y+Y8nSg@mail.gmail.com>
Subject: Re: [patch 15/14] x86/dumpstack/64: Speedup in_exception_stack()
To:     Thomas Gleixner <tglx@linutronix.de>
Cc:     Josh Poimboeuf <jpoimboe@redhat.com>,
        LKML <linux-kernel@vger.kernel.org>, X86 ML <x86@kernel.org>,
        Andy Lutomirski <luto@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Apr 3, 2019 at 12:42 PM Thomas Gleixner <tglx@linutronix.de> wrote:
>
> On Wed, 3 Apr 2019, Thomas Gleixner wrote:
> > On Tue, 2 Apr 2019, Andy Lutomirski wrote:
> > > > On Apr 2, 2019, at 1:29 PM, Thomas Gleixner <tglx@linutronix.de> wr=
ote:
> > > >>> How about a much better fix: make the DB stack be the same size a=
s all
> > > >>> the others and just have 4 of them (DB0, DB1, DB2, and DB3.  Afte=
r all,
> > > >>> overflowing from one debug stack into another is just as much of =
a bug as
> > > >>> overflowing into a different IST stack.
> > > >>
> > > >> That makes sense.
> > > >
> > > > Except that we just have two not four.
> > > >
> > > > It needs some tweaking of the ist_shift stuff in entry_64.S but tha=
t's not
> > > > rocket science. Famous last words....
> > > >
> > >
> > > The ist_shift mess should probably be in C, but that=E2=80=99s a big =
can of
> > > worms. That being said, why do we have it at all?  Once upon a time, =
we=E2=80=99d
> > > do ICEBP from user mode (or a legit breakpoint), then send a signal a=
nd
> > > hit a data breakpoint, and we=E2=80=99d recurse.  But we don=E2=80=99=
t run user debug
> > > handlers on the IST stack at all anymore.
> > >
> > > Maybe we can convince ourselves it=E2=80=99s safe?
> >
> > Maybe. Need to think about it for a while.
>
> What about kprobes. It has nasty reentrancy stuff as well...
>

Hmm.  We used to have #BP on the same stack, and I bet there were
plenty of ways to get #DB and #BP inside each other.

I bet the best solution is to set dr7 to 0 before we do anything
complicated in do_debug() (at least if we got there from kernel mode).
But we should probably make this a whole separate project after your
series is done.