From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:34957)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <paolo.bonzini@gmail.com>) id 1bHqob-0008Vg-Mm
	for qemu-devel@nongnu.org; Tue, 28 Jun 2016 07:04:26 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <paolo.bonzini@gmail.com>) id 1bHqoV-0006uq-10
	for qemu-devel@nongnu.org; Tue, 28 Jun 2016 07:04:24 -0400
Received: from mail-wm0-x243.google.com ([2a00:1450:400c:c09::243]:33384)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <paolo.bonzini@gmail.com>) id 1bHqoU-0006ue-Qu
	for qemu-devel@nongnu.org; Tue, 28 Jun 2016 07:04:18 -0400
Received: by mail-wm0-x243.google.com with SMTP id r201so4953056wme.0
	for <qemu-devel@nongnu.org>; Tue, 28 Jun 2016 04:04:18 -0700 (PDT)
Sender: Paolo Bonzini <paolo.bonzini@gmail.com>
References: <1467104499-27517-1-git-send-email-pl@kamp.de>
	<1467104499-27517-2-git-send-email-pl@kamp.de>
From: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <ce254183-faf2-e2ad-1a9c-da0985494116@redhat.com>
Date: Tue, 28 Jun 2016 13:04:13 +0200
MIME-Version: 1.0
In-Reply-To: <1467104499-27517-2-git-send-email-pl@kamp.de>
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 8bit
Subject: Re: [Qemu-devel] [PATCH 01/15] coroutine-ucontext: mmap stack memory
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel/>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Peter Lieven <pl@kamp.de>, qemu-devel@nongnu.org
Cc: kwolf@redhat.com, peter.maydell@linaro.org, mst@redhat.com, dgilbert@redhat.com, mreitz@redhat.com, kraxel@redhat.com



On 28/06/2016 11:01, Peter Lieven wrote:
> +#ifdef MAP_GROWSDOWN
> +    co->stack = mmap(NULL, COROUTINE_STACK_SIZE, PROT_READ | PROT_WRITE,
> +                     MAP_PRIVATE | MAP_ANONYMOUS | MAP_GROWSDOWN, -1, 0);
> +    if (co->stack == MAP_FAILED) {
> +        abort();
> +    }
> +    /* add a guard page at bottom of the stack */
> +    if (mmap(co->stack, getpagesize(), PROT_NONE,
> +        MAP_PRIVATE | MAP_ANONYMOUS | MAP_GROWSDOWN, -1, 0) == MAP_FAILED) {
> +        abort();
> +    }

Nevermind, you added a guard page!  Good! :)  And actually it looks like
the stack usage has been mostly tamed, at least for the block layer.

On the other hand MAP_GROWSDOWN automatically adds a guard page since
Linux 3.9 (see commit 09884964335e, "mm: do not grow the stack vma just
because of an overrun on preceding vma", 2013-02-27), so as it turns out
you don't even need the guard page!

Paolo

> +#else
>      co->stack = g_malloc(stack_size);
> +#endif