8:14 a.m.
On 1/9/23 2:32 AM, Ján Tomko wrote:
On a Sunday in 2023, Laine Stump wrote:
> This consists of (1) adding the necessary args to the qemu commandline
> netdev option, and (2) starting a passt process prior to starting
> qemu, and making sure that it is terminated when it's no longer
> needed. Under normal circumstances, passt will terminate itself as
> soon as qemu closes its socket, but in case of some error where qemu
> is never started, or fails to startup completely, we need to terminate
> passt manually.
>
> Signed-off-by: Laine Stump <laine(a)redhat.com>
> ---
> meson.build | 1 +
> po/POTFILES | 1 +
> src/qemu/meson.build | 2 +
> src/qemu/qemu_command.c | 11 +-
> src/qemu/qemu_command.h | 3 +-
> src/qemu/qemu_domain.c | 5 +-
> src/qemu/qemu_domain.h | 3 +-
> src/qemu/qemu_extdevice.c | 25 +-
> src/qemu/qemu_hotplug.c | 26 +-
> src/qemu/qemu_passt.c | 284 ++++++++++++++++++
> src/qemu/qemu_passt.h | 38 +++
> src/qemu/qemu_process.c | 1 +
> src/qemu/qemu_validate.c | 9 +-
> tests/qemuxml2argvdata/net-user-passt.args | 34 +++
> .../net-user-passt.x86_64-latest.args | 37 +++
> tests/qemuxml2argvtest.c | 2 +
> 16 files changed, 470 insertions(+), 12 deletions(-)
> create mode 100644 src/qemu/qemu_passt.c
> create mode 100644 src/qemu/qemu_passt.h
> create mode 100644 tests/qemuxml2argvdata/net-user-passt.args
> create mode 100644
> tests/qemuxml2argvdata/net-user-passt.x86_64-latest.args
>
> diff --git a/src/qemu/qemu_passt.c b/src/qemu/qemu_passt.c
> new file mode 100644
> index 0000000000..5941594811
> --- /dev/null
> +++ b/src/qemu/qemu_passt.c
> @@ -0,0 +1,284 @@
> +/*
> + * qemu_passt.c: QEMU passt support
> + *
> + * Copyright (C) 2022 Red Hat, Inc.
> + *
> + * This library is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU Lesser General Public
> + * License as published by the Free Software Foundation; either
> + * version 2.1 of the License, or (at your option) any later version.
> + *
> + * This library is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + * Lesser General Public License for more details.
> + *
> + * You should have received a copy of the GNU Lesser General Public
> + * License along with this library. If not, see
> + * <http://www.gnu.org/licenses/>;.
> + */
> +
> +#include <config.h>
> +
> +#include "qemu_dbus.h"
> +#include "qemu_extdevice.h"
> +#include "qemu_security.h"
> +#include "qemu_passt.h"
> +#include "virenum.h"
> +#include "virerror.h"
> +#include "virjson.h"
> +#include "virlog.h"
> +#include "virpidfile.h"
> +
> +#define VIR_FROM_THIS VIR_FROM_NONE
> +
> +VIR_LOG_INIT("qemu.passt");
> +
> +
> +static char *
> +qemuPasstCreatePidFilename(virDomainObj *vm,
> + virDomainNetDef *net)
> +{
> + qemuDomainObjPrivate *priv = vm->privateData;
> + virQEMUDriver *driver = priv->driver;
> + g_autoptr(virQEMUDriverConfig) cfg = virQEMUDriverGetConfig(driver);
> + g_autofree char *name = NULL;
> +
> + name = g_strdup_printf("%s-%s-passt", vm->def->name,
> net->info.alias);
Please use virDomainDefGetShortName for filename purposes.
Why? If I use GetShortName, then there's the possibility that two
domains would want to use the same name for the pidfile.
Would it be better to use the domain's UUID (as I did for the socket
path?) The advantage of using the name is that it's easier for a human
to find, but while the uuid is usually longer, its length is at least
predictable/consistent, and I suppose a human will probably never need
to find the pidfile anyway...
> +
> + return virPidFileBuildPath(cfg->passtStateDir, name);
> +}
> +
> +
[...]
> +int
> +qemuPasstAddNetProps(virDomainObj *vm,
> + virDomainNetDef *net,
> + virJSONValue **netprops)
> +{
> + g_autofree char *passtSocketName = qemuPasstCreateSocketPath(vm,
> net);
> + g_autoptr(virJSONValue) addrprops = NULL;
> +
> + if (virJSONValueObjectAdd(&addrprops,
> + "s:type", "unix",
> + "s:path", passtSocketName,
> + NULL) < 0) {
> + return -1;
> + }
> +
> + if (virJSONValueObjectAdd(netprops,
> + "s:type", "stream",
> + "a:addr", &addrprops,
> + "b:server", false,
> + /* "u:reconnect", 5, */
Debugging leftover?
It's a new option that hasn't been pushed to QEMU upstream yet. I had
meant to completely remove it, but forgot. I'll get rid of it before I push.