[libvirt PATCH 8/9] qemu: hook up passt config to qemu domains

[Laine Stump]

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 at 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.