From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [140.186.70.92] (port=38122 helo=eggs.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1Pn5bQ-0008Mf-3K for qemu-devel@nongnu.org; Wed, 09 Feb 2011 03:40:45 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Pn5bO-0006ru-V8 for qemu-devel@nongnu.org; Wed, 09 Feb 2011 03:40:43 -0500 Received: from fmmailgate01.web.de ([217.72.192.221]:54129) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Pn5bO-0006rp-IH for qemu-devel@nongnu.org; Wed, 09 Feb 2011 03:40:42 -0500 Message-ID: <4D525308.8090307@web.de> Date: Wed, 09 Feb 2011 09:40:40 +0100 From: Jan Kiszka MIME-Version: 1.0 References: <1297185509-20996-1-git-send-email-pbonzini@redhat.com> <1297185509-20996-7-git-send-email-pbonzini@redhat.com> <4D51A67A.7080306@web.de> <4D524134.8040504@redhat.com> In-Reply-To: <4D524134.8040504@redhat.com> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="------------enig766CCBB188BDF67E5D7A039F" Sender: jan.kiszka@web.de Subject: [Qemu-devel] Re: [CFT PATCH 06/12] exit round-robin vcpu loop if cpu->stopped is true List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Paolo Bonzini Cc: qemu-devel@nongnu.org This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --------------enig766CCBB188BDF67E5D7A039F Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: quoted-printable On 2011-02-09 08:24, Paolo Bonzini wrote: > On 02/08/2011 09:24 PM, Jan Kiszka wrote: >> Hmm, does this path actually trigger? If yes, does it happen to obsole= te >> the global exit_request hack? >=20 > No idea, I wanted to make the initial work as mechanical as possible. B= y > inspection, cpu_stop_current is doing >=20 > cpu_single_env->stopped =3D 1; > cpu_exit(cpu_single_env); >=20 > without setting ->stop. (I'm changing that in patch 7, but I'm setting= > ->stop to 0, not 1). Checked my own patches again: :) exit_request is not obsoleted this way, at least as long as we have !CONFIG_IOTHREAD hanging around. Also, I don't see any compelling reason now why that test should be unneeded. Rather, this looks like a sleeping race between asynchronous and synchronous vcpu stop, probably papered over by polling the cond variables so far. Jan --------------enig766CCBB188BDF67E5D7A039F Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.15 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/ iEYEARECAAYFAk1SUwgACgkQitSsb3rl5xQg/gCbBZHvE9swmybDXE+voaQBgBZE XY4AmgOea6GDGjs8MV0gjAq/JgxLWR/J =ndi4 -----END PGP SIGNATURE----- --------------enig766CCBB188BDF67E5D7A039F--