[Qemu-devel] Re: [PATCH] qdev: Catch attempt to attach more than one device to a netdev

[Markus Armbruster <armbru@redhat.com>]

"Michael S. Tsirkin" <mst@redhat.com> writes:

> On Thu, Feb 25, 2010 at 11:10:15AM +0100, Markus Armbruster wrote:
>> Guest device and host netdev are peers, i.e. it's a 1:1 relation.
>> However, we fail to enforce that:
>> 
>>     $ qemu -nodefaults --nographic -netdev user,id=net0 -device e1000,netdev=net0 -device virtio-net-pci,netdev=net0 -monitor stdio
>>     QEMU 0.12.50 monitor - type 'help' for more information
>>     (qemu) info network
>>     Devices not on any VLAN:
>>       net0: net=10.0.2.0, restricted=n peer=virtio-net-pci.0
>>       e1000.0: model=e1000,macaddr=52:54:00:12:34:56 peer=net0
>>       virtio-net-pci.0: model=virtio-net-pci,macaddr=52:54:00:12:34:57 peer=net0
>> 
>> It's all downhill from there.
>> 
>> Signed-off-by: Markus Armbruster <armbru@redhat.com>
>> ---
>>  hw/qdev-properties.c |    3 +++
>>  net.c                |    1 +
>>  2 files changed, 4 insertions(+), 0 deletions(-)
>> 
>> diff --git a/hw/qdev-properties.c b/hw/qdev-properties.c
>> index 277ff9e..89efd91 100644
>> --- a/hw/qdev-properties.c
>> +++ b/hw/qdev-properties.c
>> @@ -341,6 +341,9 @@ static int parse_netdev(DeviceState *dev, Property *prop, const char *str)
>>      *ptr = qemu_find_netdev(str);
>>      if (*ptr == NULL)
>>          return -1;
>> +    if ((*ptr)->peer) {
>> +        return -1;
>> +    }
>
> Not even -EBUSY?
> Can this produce a helpful diagnostic message?

Callers take care of that.  It's how property parse methods work.
In this case:

    property "virtio-net-pci.netdev": failed to parse "net0"
    can't set property "netdev" to "net0" for "virtio-net-pci"

>>      return 0;
>>  }
>>  
>> diff --git a/net.c b/net.c
>> index a1bf49f..e6c96d3 100644
>> --- a/net.c
>> +++ b/net.c
>> @@ -245,6 +245,7 @@ VLANClientState *qemu_new_net_client(NetClientInfo *info,
>>          QTAILQ_INSERT_TAIL(&vc->vlan->clients, vc, next);
>>      } else {
>>          if (peer) {
>> +            assert(!peer->peer);
>
> Do we ever get herE?

If "get here" means "conditional entered": yes.  For instance, with the
command line from the commit message, we get here for each of the
-device, like this:

#0  qemu_new_net_client (info=0x8b2ce0, vlan=0x0, peer=0xcd83f0, 
    model=0x5ea826 "e1000", name=0x0) at /work/armbru/qemu/net.c:227
#1  0x000000000047682f in qemu_new_nic (info=0x8b2ce0, conf=0x7fffece4a270, 
    model=0x5ea826 "e1000", name=0x0, opaque=0x7fffece4a010)
    at /work/armbru/qemu/net.c:273
#2  0x0000000000447c74 in pci_e1000_init (pci_dev=0x7fffece4a010)
    at /work/armbru/qemu/hw/e1000.c:1125
#3  0x0000000000422335 in pci_qdev_init (qdev=0x7fffece4a010, base=0x8b2dc0)
    at /work/armbru/qemu/hw/pci.c:1647
#4  0x00000000004bbb29 in qdev_init (dev=0x7fffece4a010)
    at /work/armbru/qemu/hw/qdev.c:265

If "get here" means "assertion fails": only if somebody breaks something
elsewhere.  And then assertion will slap his wrist.

>>              vc->peer = peer;
>>              peer->peer = vc;
>>          }
>> -- 
>> 1.6.6
>> 
>>