From mboxrd@z Thu Jan 1 00:00:00 1970
Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124])
by passt.top (Postfix) with ESMTP id B7C1D5A0265
for <passt-dev@passt.top>; Tue, 1 Nov 2022 08:59:36 +0100 (CET)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com;
s=mimecast20190719; t=1667289575;
h=from:from:reply-to:subject:subject:date:date:message-id:message-id:
to:to:cc:cc:mime-version:mime-version:content-type:content-type:
content-transfer-encoding:content-transfer-encoding:
in-reply-to:in-reply-to:references:references;
bh=UWMpornfRcI/PrljtoRV9usaT8LIWQvXUD/oM+X5b1U=;
b=UGHfQ5L12Uq83Y73VK5SrHAb2JjVgH6is7AocqJD6p63W27Svbq8F23u0HOhZodoGySqv0
hAA0nxY70kPru6pgpKO6uqWgFWyc2t8cWjtR+wRGA2FK5m1jT3cbw5QtroZonEKlNGJ3br
pfXT1u+FUdnVzD89mV1Wz/g/xF4eoNU=
Received: from mimecast-mx02.redhat.com (mx3-rdu2.redhat.com
[66.187.233.73]) by relay.mimecast.com with ESMTP with STARTTLS
(version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id
us-mta-410-ylwL1g7YNXWqOZ0D3DoDww-1; Tue, 01 Nov 2022 03:59:34 -0400
X-MC-Unique: ylwL1g7YNXWqOZ0D3DoDww-1
Received: from smtp.corp.redhat.com (int-mx03.intmail.prod.int.rdu2.redhat.com [10.11.54.3])
(using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits))
(No client certificate requested)
by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 0759F3815D21;
Tue, 1 Nov 2022 07:59:34 +0000 (UTC)
Received: from maya.cloud.tilaa.com (ovpn-208-31.brq.redhat.com [10.40.208.31])
by smtp.corp.redhat.com (Postfix) with ESMTPS id 8D46E111F3BB;
Tue, 1 Nov 2022 07:59:33 +0000 (UTC)
Date: Tue, 1 Nov 2022 08:59:30 +0100
From: Stefano Brivio <sbrivio@redhat.com>
To: David Gibson <david@gibson.dropbear.id.au>
Subject: Re: [PATCH v2 5/5] test: Add memory/passt test cases
Message-ID: <20221101085930.30e9b7bf@elisabeth>
In-Reply-To: <Y2BvMovjBiFLA1e1@yekko>
References: <20221031112059.170269-1-sbrivio@redhat.com>
<20221031112059.170269-6-sbrivio@redhat.com>
<Y2BvMovjBiFLA1e1@yekko>
Organization: Red Hat
MIME-Version: 1.0
X-Scanned-By: MIMEDefang 3.1 on 10.11.54.3
X-Mimecast-Spam-Score: 0
X-Mimecast-Originator: redhat.com
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Message-ID-Hash: 4JD2PKAUQSXSLCOCYXPZL4ROP3B2LTGS
X-Message-ID-Hash: 4JD2PKAUQSXSLCOCYXPZL4ROP3B2LTGS
X-MailFrom: sbrivio@redhat.com
X-Mailman-Rule-Misses: dmarc-mitigation; no-senders; approved; emergency; loop; banned-address; member-moderation; nonmember-moderation; administrivia; implicit-dest; max-recipients; max-size; news-moderation; no-subject; digests; suspicious-header
CC: passt-dev@passt.top
X-Mailman-Version: 3.3.3
Precedence: list
List-Id: Development discussion and patches for passt <passt-dev.passt.top>
Archived-At: <>
Archived-At: <https://archives.passt.top/passt-dev/20221101085930.30e9b7bf@elisabeth/>
List-Archive: <>
List-Archive: <https://archives.passt.top/passt-dev/>
List-Help: <mailto:passt-dev-request@passt.top?subject=help>
List-Owner: <mailto:passt-dev-owner@passt.top>
List-Post: <mailto:passt-dev@passt.top>
List-Subscribe: <mailto:passt-dev-join@passt.top>
List-Unsubscribe: <mailto:passt-dev-leave@passt.top>
On Tue, 1 Nov 2022 11:58:26 +1100
David Gibson <david@gibson.dropbear.id.au> wrote:
> On Mon, Oct 31, 2022 at 12:20:59PM +0100, Stefano Brivio wrote:
> > These show a summary of memory usage in kernel and userspace with
> > different port forwarding configurations, details of userspace usage
> > using 'nm' (passt only uses statically allocated memory), and details
> > of kernel memory from slab reporting facilities.
> >=20
> > This adds a new test image, mbuto.mem.img, with harcoded IPv4 and
> > IPv6 addresses and routes, and just the tools we need to start and
> > stop passt, to report from /proc/slabinfo, /proc/meminfo, and to
> > print and parse symbol sizes using nm(1).
> >=20
> > passt can't pivot_root() for sandboxing purposes on ramfs, so we need
> > to create another filesystem and chroot into it, first. =20
>=20
> Huh.. weird.
Yeah, it also surprised me, but man 2 pivot_root says:
The rootfs (initial ramfs) cannot be pivot_root()ed. The recommen=
ded
method of changing the root filesystem in this case is to delete eve=
ry=E2=80=90
thing in rootfs, overmount rootfs with the new root, attach stdin/s=
td=E2=80=90
out/stderr to the new /dev/console, and exec the new init(1). Hel=
per
programs for this process exist; see switch_root(8).
(of course I got the documented case of EINVAL before reading that).
> > We don't want to use pane context functions, as we're checking memory
> > usage for sockets: resort to screen-scraping. =20
>=20
> So, we don't do it at the moment, but it should be pretty easy to make
> the ssh-over-vsock stuff use a persistent control socket
> (ControlPersist / ControlMaster). That would mean we'd just establish
> one connection then all the subsequent guest bound commands would go
> over that one initial connection.
>=20
> It wouldn't take zero socket memory of course, but it should be small
> and more importantly stable. So, if we take a diff of usage as you
> already do, the results should still be good.
Ah, yes, I guess so. I would track this as a ticket instead of trying
to implement it now.
> This patch is not working for me, for reasons I'm still debugging.
> Some more minor comments in the interim.
>=20
> >=20
> > Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
> > ---
> > test/Makefile | 5 +-
> > test/lib/layout_ugly | 39 +++++++++
> > test/lib/setup_ugly | 37 +++++++++
> > test/memory/passt | 187 +++++++++++++++++++++++++++++++++++++++++++
> > test/passt.mem.mbuto | 42 ++++++++++
> > test/run | 4 +
> > 6 files changed, 313 insertions(+), 1 deletion(-)
> > create mode 100644 test/memory/passt
> > create mode 100755 test/passt.mem.mbuto
> >=20
> > diff --git a/test/Makefile b/test/Makefile
> > index 91498ff..9f2bc42 100644
> > --- a/test/Makefile
> > +++ b/test/Makefile
> > @@ -55,7 +55,7 @@ UBUNTU_IMGS =3D $(UBUNTU_OLD_IMGS) $(UBUNTU_NEW_IMGS)
> > DOWNLOAD_ASSETS =3D mbuto \
> > =09$(DEBIAN_IMGS) $(FEDORA_IMGS) $(OPENSUSE_IMGS) $(UBUNTU_IMGS)
> > TESTDATA_ASSETS =3D small.bin big.bin medium.bin
> > -LOCAL_ASSETS =3D mbuto.img QEMU_EFI.fd \
> > +LOCAL_ASSETS =3D mbuto.img mbuto.mem.img QEMU_EFI.fd \
> > =09$(DEBIAN_IMGS:%=3Dprepared-%) $(FEDORA_IMGS:%=3Dprepared-%) \
> > =09$(UBUNTU_NEW_IMGS:%=3Dprepared-%) \
> > =09nsholder guest-key guest-key.pub \
> > @@ -76,6 +76,9 @@ guest-key guest-key.pub:
> > mbuto.img: passt.mbuto mbuto guest-key.pub $(TESTDATA_ASSETS)
> > =09./mbuto/mbuto -p ./$< -c lz4 -f $@
> > =20
> > +mbuto.mem.img: passt.mem.mbuto mbuto ../passt.avx2
> > +=09./mbuto/mbuto -p ./$< -c lz4 -f $@
> > +
> > nsholder: nsholder.c
> > =09$(CC) $(CFLAGS) -o $@ $^
> > =20
> > diff --git a/test/lib/layout_ugly b/test/lib/layout_ugly
> > index 9397b7d..d62d337 100644
> > --- a/test/lib/layout_ugly
> > +++ b/test/lib/layout_ugly
> > @@ -81,3 +81,42 @@ layout_pasta_simple() {
> > =20
> > =09sleep 1
> > }
> > +
> > +# layout_memory() - Screen-scraped panes for memory usage tests, big g=
uest pane
> > +layout_memory() {
> > +=09sleep 3
> > +
> > +=09tmux kill-pane -a -t 0
> > +=09cmd_write 0 clear
> > +
> > +=09tmux split-window -h -l '35%' -t passt_test
> > +=09tmux split-window -v -l '15%' -t passt_test
> > +
> > +=09PANE_PASST=3D2
> > +=09PANE_GUEST=3D0
> > +=09PANE_INFO=3D1
> > +
> > +=09get_info_cols
> > +
> > +=09tmux send-keys -l -t ${PANE_INFO} 'while cat '"$STATEBASE/log_pipe"=
'; do :; done'
> > +=09tmux send-keys -t ${PANE_INFO} -N 100 C-m
> > +=09tmux select-pane -t ${PANE_INFO} -T "test log"
> > +
> > +=09if context_exists guest; then
> > +=09=09pane_watch_contexts ${PANE_GUEST} guest guest
> > +=09else
> > +=09=09tmux pipe-pane -O -t ${PANE_GUEST} "cat >> ${LOGDIR}/pane_guest.=
log"
> > +=09=09tmux select-pane -t ${PANE_GUEST} -T "guest"
> > +=09fi
> > +
> > +=09if context_exists passt; then =20
>=20
> AFAICT you're always using a context pane for passt, so you don't need
> this conditional.
Right, I'll drop this.
> > +=09=09pane_watch_contexts ${PANE_PASST} passt passt
> > +=09else
> > +=09=09tmux pipe-pane -O -t ${PANE_PASST} "cat >> ${LOGDIR}/pane_passt.=
log"
> > +=09=09tmux select-pane -t ${PANE_PASST} -T "passt"
> > +=09fi
> > +
> > +=09info_layout "memory usage"
> > +
> > +=09sleep 1
> > +}
> > diff --git a/test/lib/setup_ugly b/test/lib/setup_ugly
> > index 764177e..f2e07ba 100755
> > --- a/test/lib/setup_ugly
> > +++ b/test/lib/setup_ugly
> > @@ -13,6 +13,8 @@
> > # Copyright (c) 2022 Red Hat GmbH
> > # Author: Stefano Brivio <sbrivio@redhat.com>
> > =20
> > +INITRAMFS_MEM=3D"${BASEPATH}/mbuto.mem.img"
> > +
> > # setup_distro() - Set up pane layout for distro tests
> > setup_distro() {
> > =09layout_host
> > @@ -25,6 +27,30 @@ setup_pasta_options() {
> > =09layout_pasta_simple
> > }
> > =20
> > +# setup_memory() - Start qemu in guest pane, and passt in passt contex=
t
> > +setup_memory() {
> > +=09context_setup_host passt
> > +
> > +=09layout_memory
> > +
> > +=09context_run passt "./passt -P ${STATESETUP}/passt.pid"
> > +
> > +=09# pidfile isn't created until passt is listening
> > +=09wait_for [ -f "${STATESETUP}/passt.pid" ]
> > +
> > +=09pane_or_context_run guest './qrap 5 qemu-system-$(uname -m)'=09 \
> > +=09=09' -machine accel=3Dkvm' \
> > +=09=09' -m '${VMEM}' -cpu host -smp '${VCPUS} \
> > +=09=09' -kernel ' "/boot/vmlinuz-$(uname -r)"=09=09=09 \
> > +=09=09' -initrd '${INITRAMFS_MEM}' -nographic -serial stdio'=09 \
> > +=09=09' -nodefaults'=09=09=09=09=09=09 \
> > +=09=09' -append "console=3DttyS0 mitigations=3Doff apparmor=3D0 '=09 =
\
> > +=09=09'virtio-net.napi_tx=3D1"'=09=09=09=09=09 \
> > +=09=09" -device virtio-net-pci,netdev=3Dhostnet0,x-txburst=3D16384" \
> > +=09=09" -netdev socket,fd=3D5,id=3Dhostnet0"=09=09=09 \
> > +=09=09" -pidfile ${STATESETUP}/qemu.pid"
> > +}
> > +
> > # teardown_distro() - Nothing to do, yet
> > teardown_distro() {
> > =09:
> > @@ -36,3 +62,14 @@ teardown_pasta_options() {
> > =09teardown_context_watch ${PANE_PASST} passt
> > }
> > =20
> > +# teardown_passt() - Kill qemu with ^C, remove passt PID file
> > +teardown_memory() {
> > +=09kill $(cat "${STATESETUP}/qemu.pid")
> > +
> > +=09rm "${STATESETUP}/passt.pid"
> > +
> > +=09tmux send-keys -t ${PANE_PASST} "C-c" =20
>=20
> You shouldn't need this for the passt pane, since it is context based,
> the teardown_context_watch will do it.
Oops, yes, I added this in a previous version where it was also not
context based. I'll drop this,
> > +=09teardown_context_watch ${PANE_PASST} passt
> > +=09teardown_context_watch ${PANE_GUEST} qemu =20
>=20
> Likewise you don't want this for the guest pane since it is *not*
> context based. =20
and this, of course.
--=20
Stefano