From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:48610) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1defJA-0000hN-Bb for qemu-devel@nongnu.org; Mon, 07 Aug 2017 06:30:50 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1defJ6-00015t-7Z for qemu-devel@nongnu.org; Mon, 07 Aug 2017 06:30:48 -0400 Received: from userp1040.oracle.com ([156.151.31.81]:27346) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1defJ5-00015h-JF for qemu-devel@nongnu.org; Mon, 07 Aug 2017 06:30:44 -0400 Message-ID: <1502101838.3624.8.camel@oracle.com> From: Knut Omang Date: Mon, 07 Aug 2017 12:30:38 +0200 In-Reply-To: <20170807102729.GG18259@redhat.com> References: <7c4ddf95a57f7dfa546c11f8a08a9091b8bd1ceb.1501362828.git-series.knut.omang@oracle.com> <20170807084524.GF18259@redhat.com> <1502097685.3624.6.camel@oracle.com> <20170807102729.GG18259@redhat.com> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v6 1/4] tests: Add test-listen - a stress test for QEMU socket listen List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: "Daniel P. Berrange" Cc: Gerd Hoffmann , Paolo Bonzini , qemu-devel@nongnu.org On Mon, 2017-08-07 at 11:27 +0100, Daniel P. Berrange wrote: > On Mon, Aug 07, 2017 at 11:21:25AM +0200, Knut Omang wrote: > > On Mon, 2017-08-07 at 09:45 +0100, Daniel P. Berrange wrote: > > > On Sat, Jul 29, 2017 at 11:18:15PM +0200, Knut Omang wrote: > > > > There's a potential race condition between multiple bind()'s > > > > attempting to bind to the same port, which occasionally > > > > allows more than one bind to succeed against the same port. > > > >=20 > > > > When a subsequent listen() call is made with the same socket > > > > only one will succeed. > > > >=20 > > > > The current QEMU code does however not take this situation into acc= ount > > > > and the listen will cause the code to break out and fail even > > > > when there are actually available ports to use. > > > >=20 > > > > This test exposes two subtests: > > > >=20 > > > > /socket/listen-serial > > > > /socket/listen-compete > > > >=20 > > > > The "compete" subtest creates a number of threads and have them all= trying to bind > > > > to the same port with a large enough offset input to > > > > allow all threads to get it's own port. > > > > The "serial" subtest just does the same, except in series in a > > > > single thread. > > > >=20 > > > > The serial version passes, probably in most versions of QEMU. > > > >=20 > > > > The parallel version exposes the problem in a relatively reliable w= ay, > > > > eg. it fails a majority of times, but not with a 100% rate, occasio= nal > > > > passes can be seen. Nevertheless this is quite good given that > > > > the bug was tricky to reproduce and has been left undetected for > > > > a while. > > > >=20 > > > > The problem seems to be present in all versions of QEMU. > > > >=20 > > > > The original failure scenario occurred with VNC port allocation > > > > in a traditional Xen based build, in different code > > > > but with similar functionality. > > > >=20 > > > > Reported-by: Bhavesh Davda > > > > Signed-off-by: Knut Omang > > > > Reviewed-by: Yuval Shaia > > > > Reviewed-by: Bhavesh Davda > > > > Reviewed-by: Girish Moodalbail > > > > --- > > > > =C2=A0tests/Makefile.include |=C2=A0=C2=A0=C2=A02 +- > > > > =C2=A0tests/test-listen.c=C2=A0=C2=A0=C2=A0=C2=A0| 253 ++++++++++++= ++++++++++++++++++++++++++++++- > > > > =C2=A02 files changed, 255 insertions(+) > > > > =C2=A0create mode 100644 tests/test-listen.c > > > >=20 > > > > diff --git a/tests/Makefile.include b/tests/Makefile.include > > > > index 7af278d..b37c0c8 100644 > > > > --- a/tests/Makefile.include > > > > +++ b/tests/Makefile.include > > > > @@ -128,6 +128,7 @@ check-unit-y +=3D tests/test-bufferiszero$(EXES= UF) > > > > =C2=A0gcov-files-check-bufferiszero-y =3D util/bufferiszero.c > > > > =C2=A0check-unit-y +=3D tests/test-uuid$(EXESUF) > > > > =C2=A0check-unit-y +=3D tests/ptimer-test$(EXESUF) > > > > +#check-unit-y +=3D tests/test-listen$(EXESUF) > > > > =C2=A0gcov-files-ptimer-test-y =3D hw/core/ptimer.c > > > > =C2=A0check-unit-y +=3D tests/test-qapi-util$(EXESUF) > > > > =C2=A0gcov-files-test-qapi-util-y =3D qapi/qapi-util.c > > > > @@ -769,6 +770,7 @@ tests/test-arm-mptimer$(EXESUF): tests/test-arm= -mptimer.o > > > > =C2=A0tests/test-qapi-util$(EXESUF): tests/test-qapi-util.o $(test-= util-obj-y) > > > > =C2=A0tests/numa-test$(EXESUF): tests/numa-test.o > > > > =C2=A0tests/vmgenid-test$(EXESUF): tests/vmgenid-test.o tests/boot-= sector.o tests/acpi-utils.o > > > > +tests/test-listen$(EXESUF): tests/test-listen.o $(test-util-obj-y) > > > > =C2=A0 > > > > =C2=A0tests/migration/stress$(EXESUF): tests/migration/stress.o > > > > =C2=A0 $(call quiet-command, $(LINKPROG) -static -O3 $(PTHREAD_LIB)= -o $@ $< ,"LINK","$(TARGET_DIR)$@") > > > > diff --git a/tests/test-listen.c b/tests/test-listen.c > > > > new file mode 100644 > > > > index 0000000..5c07537 > > > > --- /dev/null > > > > +++ b/tests/test-listen.c > > > > @@ -0,0 +1,253 @@ > > > > +/* > > > > + * Copyright (c) 2017, Oracle and/or its affiliates. All rights re= served. > > > > + *=C2=A0=C2=A0=C2=A0=C2=A0Author: Knut Omang > > > > + * > > > > + * This program is free software; you can redistribute it and/or m= odify > > > > + * it under the terms of the GNU General Public License version 2 > > > > + * as published by the Free Software Foundation. > > >=20 > > > Can you change that to "version 2 or later" - per the LICENSE file, w= e don't > > > accept contributions under "version 2 only" except for 4 specific sub= dirs: > > >=20 > > >=20 > > > =C2=A0 "As of July 2013, contributions under version 2 of the GNU Gen= eral Public > > > =C2=A0=C2=A0=C2=A0License (and no later version) are only accepted fo= r the following files > > > =C2=A0=C2=A0=C2=A0or directories: bsd-user/, linux-user/, hw/vfio/, h= w/xen/xen_pt*." > >=20 > > Oh, sorry - I wasn't aware of this, +"...or later" is fine with me. > > Would you like me to send a v7 of the set with only that change, or can= you amend=C2=A0 > > it as part of the merge? >=20 > Since its a copyright statement, I'd prefer if you just sent a v7. I've n= o > other comments, so will queue it for merge once you resend, and send a pu= ll > request once 2.11 opens up. Ok, will do, Thanks, Knut > Regards, > Daniel