From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:59605)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <imammedo@redhat.com>) id 1c5BOZ-0003Uf-7R
	for qemu-devel@nongnu.org; Fri, 11 Nov 2016 07:57:30 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <imammedo@redhat.com>) id 1c5BOU-0007e7-Cv
	for qemu-devel@nongnu.org; Fri, 11 Nov 2016 07:57:27 -0500
Received: from mx1.redhat.com ([209.132.183.28]:51442)
	by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32)
	(Exim 4.71) (envelope-from <imammedo@redhat.com>) id 1c5BOU-0007dv-7t
	for qemu-devel@nongnu.org; Fri, 11 Nov 2016 07:57:22 -0500
Date: Fri, 11 Nov 2016 13:57:17 +0100
From: Igor Mammedov <imammedo@redhat.com>
Message-ID: <20161111135717.18eaa6dc@nial.brq.redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Subject: [Qemu-devel] FW_CFG_NB_CPUS vs fwcfg file 'etc/boot-cpus'
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: qemu-devel@nongnu.org
Cc: Laszlo Ersek <lersek@redhat.com>, mst@redhat.com, Stefan Hajnoczi <stefanha@gmail.com>, Kevin O'Connor <kevin@koconnor.net>, Gerd Hoffmann <kraxel@redhat.com>, Eduardo Habkost <ehabkost@redhat.com>

While looking at OVMF and how it handles CPUs (ACPI/AP wakeup),
I've noticed that it uses legacy FW_CFG_NB_CPUS(0x05) to get
the number of present at start CPUs.
Variable was introduced back in 2008 by fbfcf955ba and is/was used
by ppc/sparc/arm/x86 and a bunch of firmwares (but not by SeaBIOS).

However in 2.8 I've just added similar fwcfg file 'etc/boot-cpus'
which is used for the same purpose \-)

Both variables are UINT16 and the only difference that FW_CFG_NB_CPUS
doesn't account for CPUs added with -device which might make
a firmware to wait indefinitely in FW_CFG_NB_CPUS == Woken-up-APs loop
due to extra not expected APs being woken up.

FW_CFG_NB_CPUS should be fixed to mimic 'etc/boot-cpus' behavior anyway,
so question arises why not to reuse fixed legacy FW_CFG_NB_CPUS and
drop just added but not yet released 'etc/boot-cpus' from QEMU and SeaBIOS.

Any opinions?