[PATCH 0/2] net: Fix announce_self with vhost

[PATCH 0/2] net: Fix announce_self with vhost

[Laurent Vivier]

announce_self that sends a RARP packet after migration
or with announce_self QMP/HMP command doesn't work with
vhost because of the vnet headers.

announce_self is the only user of QEMU_NET_PACKET_FLAG_RAW and
this flag is not correctly managed.

This series fix a problem with this flag in filter-dump and in
qemu_deliver_packet_iov().

Laurent Vivier (2):
  net: Fix announce_self
  net/dump: Correctly compute Ethernet packet offset

 net/dump.c | 3 ++-
 net/net.c  | 1 +
 2 files changed, 3 insertions(+), 1 deletion(-)

-- 
2.47.1

[PATCH 1/2] net: Fix announce_self

[Laurent Vivier]

b9ad513e1876 ("net: Remove receive_raw()") adds an iovec entry
in qemu_deliver_packet_iov() to add the virtio-net header
in the data when QEMU_NET_PACKET_FLAG_RAW is set but forgets
to increase the number of iovec entries in the array, so
receive_iov() will only send the first entry (the virtio-net
entry, full of 0) and no data. The packet will be discarded.

The only user of QEMU_NET_PACKET_FLAG_RAW is announce_self.

We can see the problem with tcpdump:

- QEMU parameters:

  .. -monitor stdio \
     -netdev bridge,id=netdev0,br=virbr0 \
     -device virtio-net,mac=9a:2b:2c:2d:2e:2f,netdev=netdev0 \

- HMP command:

  (qemu) announce_self

- TCP dump:

  $ sudo tcpdump -nxi virbr0

  without the fix:

    <nothing>

  with the fix:

   ARP, Reverse Request who-is 9a:2b:2c:2d:2e:2f tell 9a:2b:2c:2d:2e:2f, length 46
        0x0000:  0001 0800 0604 0003 9a2b 2c2d 2e2f 0000
        0x0010:  0000 9a2b 2c2d 2e2f 0000 0000 0000 0000
        0x0020:  0000 0000 0000 0000 0000 0000 0000

Reported-by: Xiaohui Li <xiaohli@redhat.com>
Bug: https://issues.redhat.com/browse/RHEL-73891
Fixes: b9ad513e1876 ("net: Remove receive_raw()")
Cc: akihiko.odaki@daynix.com
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
---
 net/net.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/net/net.c b/net/net.c
index c1bb19a52373..9cded70dde74 100644
--- a/net/net.c
+++ b/net/net.c
@@ -822,6 +822,7 @@ static ssize_t qemu_deliver_packet_iov(NetClientState *sender,
         iov_copy[0].iov_len =  nc->vnet_hdr_len;
         memcpy(&iov_copy[1], iov, iovcnt * sizeof(*iov));
         iov = iov_copy;
+        iovcnt++;
     }
 
     if (nc->info->receive_iov) {
-- 
2.47.1

Re: [PATCH 1/2] net: Fix announce_self

[Akihiko Odaki]

On 2025/01/17 20:17, Laurent Vivier wrote:
> b9ad513e1876 ("net: Remove receive_raw()") adds an iovec entry
> in qemu_deliver_packet_iov() to add the virtio-net header
> in the data when QEMU_NET_PACKET_FLAG_RAW is set but forgets
> to increase the number of iovec entries in the array, so
> receive_iov() will only send the first entry (the virtio-net
> entry, full of 0) and no data. The packet will be discarded.
> 
> The only user of QEMU_NET_PACKET_FLAG_RAW is announce_self.
> 
> We can see the problem with tcpdump:
> 
> - QEMU parameters:
> 
>    .. -monitor stdio \
>       -netdev bridge,id=netdev0,br=virbr0 \
>       -device virtio-net,mac=9a:2b:2c:2d:2e:2f,netdev=netdev0 \
> 
> - HMP command:
> 
>    (qemu) announce_self
> 
> - TCP dump:
> 
>    $ sudo tcpdump -nxi virbr0
> 
>    without the fix:
> 
>      <nothing>
> 
>    with the fix:
> 
>     ARP, Reverse Request who-is 9a:2b:2c:2d:2e:2f tell 9a:2b:2c:2d:2e:2f, length 46
>          0x0000:  0001 0800 0604 0003 9a2b 2c2d 2e2f 0000
>          0x0010:  0000 9a2b 2c2d 2e2f 0000 0000 0000 0000
>          0x0020:  0000 0000 0000 0000 0000 0000 0000
> 
> Reported-by: Xiaohui Li <xiaohli@redhat.com>
> Bug: https://issues.redhat.com/browse/RHEL-73891
> Fixes: b9ad513e1876 ("net: Remove receive_raw()")
> Cc: akihiko.odaki@daynix.com
> Signed-off-by: Laurent Vivier <lvivier@redhat.com>

Thanks for finding out this bug:

Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>

I believe this should have:

Cc: qemu-stable@nongnu.org

> ---
>   net/net.c | 1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/net/net.c b/net/net.c
> index c1bb19a52373..9cded70dde74 100644
> --- a/net/net.c
> +++ b/net/net.c
> @@ -822,6 +822,7 @@ static ssize_t qemu_deliver_packet_iov(NetClientState *sender,
>           iov_copy[0].iov_len =  nc->vnet_hdr_len;
>           memcpy(&iov_copy[1], iov, iovcnt * sizeof(*iov));
>           iov = iov_copy;
> +        iovcnt++;
>       }
>   
>       if (nc->info->receive_iov) {

Re: [PATCH 1/2] net: Fix announce_self

[Jason Wang]

On Fri, Jan 17, 2025 at 7:17 PM Laurent Vivier <lvivier@redhat.com> wrote:
>
> b9ad513e1876 ("net: Remove receive_raw()") adds an iovec entry
> in qemu_deliver_packet_iov() to add the virtio-net header
> in the data when QEMU_NET_PACKET_FLAG_RAW is set but forgets
> to increase the number of iovec entries in the array, so
> receive_iov() will only send the first entry (the virtio-net
> entry, full of 0) and no data. The packet will be discarded.
>
> The only user of QEMU_NET_PACKET_FLAG_RAW is announce_self.
>
> We can see the problem with tcpdump:
>
> - QEMU parameters:
>
>   .. -monitor stdio \
>      -netdev bridge,id=netdev0,br=virbr0 \
>      -device virtio-net,mac=9a:2b:2c:2d:2e:2f,netdev=netdev0 \
>
> - HMP command:
>
>   (qemu) announce_self
>
> - TCP dump:
>
>   $ sudo tcpdump -nxi virbr0
>
>   without the fix:
>
>     <nothing>
>
>   with the fix:
>
>    ARP, Reverse Request who-is 9a:2b:2c:2d:2e:2f tell 9a:2b:2c:2d:2e:2f, length 46
>         0x0000:  0001 0800 0604 0003 9a2b 2c2d 2e2f 0000
>         0x0010:  0000 9a2b 2c2d 2e2f 0000 0000 0000 0000
>         0x0020:  0000 0000 0000 0000 0000 0000 0000
>
> Reported-by: Xiaohui Li <xiaohli@redhat.com>
> Bug: https://issues.redhat.com/browse/RHEL-73891
> Fixes: b9ad513e1876 ("net: Remove receive_raw()")
> Cc: akihiko.odaki@daynix.com
> Signed-off-by: Laurent Vivier <lvivier@redhat.com>
> ---

Acked-by: Jason Wang <jasowang@redhat.com>

Thanks

[PATCH 2/2] net/dump: Correctly compute Ethernet packet offset

[Laurent Vivier]

When a packet is sent with QEMU_NET_PACKET_FLAG_RAW by QEMU it
never includes virtio-net header even if qemu_get_vnet_hdr_len()
is not 0, and filter-dump is not managing this case.

The only user of QEMU_NET_PACKET_FLAG_RAW is announce_self,
we can show the problem using it and tcpddump:

- QEMU parameters:

  .. -monitor stdio \
     -netdev bridge,id=netdev0,br=virbr0 \
     -device virtio-net,mac=9a:2b:2c:2d:2e:2f,netdev=netdev0 \
     -object filter-dump,netdev=netdev0,file=log.pcap,id=pcap0

- HMP command:

  (qemu) announce_self

- TCP dump:

  $ tcpdump -nxr log.pcap

  without the fix:

    08:00:06:04:00:03 > 2e:2f:80:35:00:01, ethertype Unknown (0x9a2b), length 50:
         0x0000:  2c2d 2e2f 0000 0000 9a2b 2c2d 2e2f 0000
         0x0010:  0000 0000 0000 0000 0000 0000 0000 0000
         0x0020:  0000 0000

  with the fix:

    ARP, Reverse Request who-is 9a:2b:2c:2d:2e:2f tell 9a:2b:2c:2d:2e:2f, length 46
         0x0000:  0001 0800 0604 0003 9a2b 2c2d 2e2f 0000
         0x0010:  0000 9a2b 2c2d 2e2f 0000 0000 0000 0000
         0x0020:  0000 0000 0000 0000 0000 0000 0000

Fixes: 481c52320a26 ("net: Strip virtio-net header when dumping")
Cc: akihiko.odaki@daynix.com
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
---
 net/dump.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/net/dump.c b/net/dump.c
index d7dd2ce461de..140215aa1054 100644
--- a/net/dump.c
+++ b/net/dump.c
@@ -155,7 +155,8 @@ static ssize_t filter_dump_receive_iov(NetFilterState *nf, NetClientState *sndr,
 {
     NetFilterDumpState *nfds = FILTER_DUMP(nf);
 
-    dump_receive_iov(&nfds->ds, iov, iovcnt, qemu_get_vnet_hdr_len(nf->netdev));
+    dump_receive_iov(&nfds->ds, iov, iovcnt, flags & QEMU_NET_PACKET_FLAG_RAW ?
+                     0 : qemu_get_vnet_hdr_len(nf->netdev));
     return 0;
 }
 
-- 
2.47.1

Re: [PATCH 2/2] net/dump: Correctly compute Ethernet packet offset

[Akihiko Odaki]

On 2025/01/17 20:17, Laurent Vivier wrote:
> When a packet is sent with QEMU_NET_PACKET_FLAG_RAW by QEMU it
> never includes virtio-net header even if qemu_get_vnet_hdr_len()
> is not 0, and filter-dump is not managing this case.
> 
> The only user of QEMU_NET_PACKET_FLAG_RAW is announce_self,
> we can show the problem using it and tcpddump:
> 
> - QEMU parameters:
> 
>    .. -monitor stdio \
>       -netdev bridge,id=netdev0,br=virbr0 \
>       -device virtio-net,mac=9a:2b:2c:2d:2e:2f,netdev=netdev0 \
>       -object filter-dump,netdev=netdev0,file=log.pcap,id=pcap0
> 
> - HMP command:
> 
>    (qemu) announce_self
> 
> - TCP dump:
> 
>    $ tcpdump -nxr log.pcap
> 
>    without the fix:
> 
>      08:00:06:04:00:03 > 2e:2f:80:35:00:01, ethertype Unknown (0x9a2b), length 50:
>           0x0000:  2c2d 2e2f 0000 0000 9a2b 2c2d 2e2f 0000
>           0x0010:  0000 0000 0000 0000 0000 0000 0000 0000
>           0x0020:  0000 0000
> 
>    with the fix:
> 
>      ARP, Reverse Request who-is 9a:2b:2c:2d:2e:2f tell 9a:2b:2c:2d:2e:2f, length 46
>           0x0000:  0001 0800 0604 0003 9a2b 2c2d 2e2f 0000
>           0x0010:  0000 9a2b 2c2d 2e2f 0000 0000 0000 0000
>           0x0020:  0000 0000 0000 0000 0000 0000 0000
> 
> Fixes: 481c52320a26 ("net: Strip virtio-net header when dumping")
> Cc: akihiko.odaki@daynix.com
> Signed-off-by: Laurent Vivier <lvivier@redhat.com>

Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>

Re: [PATCH 2/2] net/dump: Correctly compute Ethernet packet offset

[Jason Wang]

On Fri, Jan 17, 2025 at 7:17 PM Laurent Vivier <lvivier@redhat.com> wrote:
>
> When a packet is sent with QEMU_NET_PACKET_FLAG_RAW by QEMU it
> never includes virtio-net header even if qemu_get_vnet_hdr_len()
> is not 0, and filter-dump is not managing this case.
>
> The only user of QEMU_NET_PACKET_FLAG_RAW is announce_self,
> we can show the problem using it and tcpddump:
>
> - QEMU parameters:
>
>   .. -monitor stdio \
>      -netdev bridge,id=netdev0,br=virbr0 \
>      -device virtio-net,mac=9a:2b:2c:2d:2e:2f,netdev=netdev0 \
>      -object filter-dump,netdev=netdev0,file=log.pcap,id=pcap0
>
> - HMP command:
>
>   (qemu) announce_self
>
> - TCP dump:
>
>   $ tcpdump -nxr log.pcap
>
>   without the fix:
>
>     08:00:06:04:00:03 > 2e:2f:80:35:00:01, ethertype Unknown (0x9a2b), length 50:
>          0x0000:  2c2d 2e2f 0000 0000 9a2b 2c2d 2e2f 0000
>          0x0010:  0000 0000 0000 0000 0000 0000 0000 0000
>          0x0020:  0000 0000
>
>   with the fix:
>
>     ARP, Reverse Request who-is 9a:2b:2c:2d:2e:2f tell 9a:2b:2c:2d:2e:2f, length 46
>          0x0000:  0001 0800 0604 0003 9a2b 2c2d 2e2f 0000
>          0x0010:  0000 9a2b 2c2d 2e2f 0000 0000 0000 0000
>          0x0020:  0000 0000 0000 0000 0000 0000 0000
>
> Fixes: 481c52320a26 ("net: Strip virtio-net header when dumping")
> Cc: akihiko.odaki@daynix.com
> Signed-off-by: Laurent Vivier <lvivier@redhat.com>
> ---

Acked-by: Jason Wang <jasowang@redhat.com>

Thanks

Re: [PATCH 0/2] net: Fix announce_self with vhost

[Laurent Vivier]

CC: qemu-stable and qemu-trivial.

On 17/01/2025 12:17, Laurent Vivier wrote:
> announce_self that sends a RARP packet after migration
> or with announce_self QMP/HMP command doesn't work with
> vhost because of the vnet headers.
> 
> announce_self is the only user of QEMU_NET_PACKET_FLAG_RAW and
> this flag is not correctly managed.
> 
> This series fix a problem with this flag in filter-dump and in
> qemu_deliver_packet_iov().
> 
> Laurent Vivier (2):
>    net: Fix announce_self
>    net/dump: Correctly compute Ethernet packet offset
> 
>   net/dump.c | 3 ++-
>   net/net.c  | 1 +
>   2 files changed, 3 insertions(+), 1 deletion(-)
> 

Re: [PATCH 0/2] net: Fix announce_self with vhost

[Michael Tokarev]

17.01.2025 17:05, Laurent Vivier wrote:
> CC: qemu-stable and qemu-trivial.
> 
> On 17/01/2025 12:17, Laurent Vivier wrote:
>> announce_self that sends a RARP packet after migration
>> or with announce_self QMP/HMP command doesn't work with
>> vhost because of the vnet headers.
>>
>> announce_self is the only user of QEMU_NET_PACKET_FLAG_RAW and
>> this flag is not correctly managed.
>>
>> This series fix a problem with this flag in filter-dump and in
>> qemu_deliver_packet_iov().
>>
>> Laurent Vivier (2):
>>    net: Fix announce_self
>>    net/dump: Correctly compute Ethernet packet offset

Is this one - "Correctly compute Ethernet packet offset" - needed for
older stable series too, like 8.2 and 7.2 (currently active ones)?

Since it doesn't apply without 4b52d63249 (tap: Remove qemu_using_vnet_hdr()).

Thanks,

/mjt

Re: [PATCH 0/2] net: Fix announce_self with vhost

[Laurent Vivier]

Le 01/02/2025 à 20:36, Michael Tokarev a écrit :
> 17.01.2025 17:05, Laurent Vivier wrote:
>> CC: qemu-stable and qemu-trivial.
>>
>> On 17/01/2025 12:17, Laurent Vivier wrote:
>>> announce_self that sends a RARP packet after migration
>>> or with announce_self QMP/HMP command doesn't work with
>>> vhost because of the vnet headers.
>>>
>>> announce_self is the only user of QEMU_NET_PACKET_FLAG_RAW and
>>> this flag is not correctly managed.
>>>
>>> This series fix a problem with this flag in filter-dump and in
>>> qemu_deliver_packet_iov().
>>>
>>> Laurent Vivier (2):
>>>    net: Fix announce_self
>>>    net/dump: Correctly compute Ethernet packet offset
> 
> Is this one - "Correctly compute Ethernet packet offset" - needed for
> older stable series too, like 8.2 and 7.2 (currently active ones)?
> 
> Since it doesn't apply without 4b52d63249 (tap: Remove qemu_using_vnet_hdr()).

It fixes 481c52320a26 ("net: Strip virtio-net header when dumping") that is only in 8.0.
But I'm not sure the bug exists without 4b52d63249 (tap: Remove qemu_using_vnet_hdr()), so it should 
be tested first.

Thanks,
Laurent

Re: [PATCH 0/2] net: Fix announce_self with vhost

[Michael Tokarev]

17.01.2025 14:17, Laurent Vivier wrote:
> announce_self that sends a RARP packet after migration
> or with announce_self QMP/HMP command doesn't work with
> vhost because of the vnet headers.
> 
> announce_self is the only user of QEMU_NET_PACKET_FLAG_RAW and
> this flag is not correctly managed.
> 
> This series fix a problem with this flag in filter-dump and in
> qemu_deliver_packet_iov().
> 
> Laurent Vivier (2):
>    net: Fix announce_self
>    net/dump: Correctly compute Ethernet packet offset
> 
>   net/dump.c | 3 ++-
>   net/net.c  | 1 +
>   2 files changed, 3 insertions(+), 1 deletion(-)

Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>

Applied to trivial-patches and queued for qemu-stable.

Thanks,

/mjt

Re: [PATCH 0/2] net: Fix announce_self with vhost

[Laurent Vivier]

On 18/01/2025 21:20, Michael Tokarev wrote:
> 17.01.2025 14:17, Laurent Vivier wrote:
>> announce_self that sends a RARP packet after migration
>> or with announce_self QMP/HMP command doesn't work with
>> vhost because of the vnet headers.
>>
>> announce_self is the only user of QEMU_NET_PACKET_FLAG_RAW and
>> this flag is not correctly managed.
>>
>> This series fix a problem with this flag in filter-dump and in
>> qemu_deliver_packet_iov().
>>
>> Laurent Vivier (2):
>>    net: Fix announce_self
>>    net/dump: Correctly compute Ethernet packet offset
>>
>>   net/dump.c | 3 ++-
>>   net/net.c  | 1 +
>>   2 files changed, 3 insertions(+), 1 deletion(-)
> 
> Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
> 
> Applied to trivial-patches and queued for qemu-stable.
> 

When do you plan to send the PR for them?

Thanks,
Laurent

Re: [PATCH 0/2] net: Fix announce_self with vhost

[Michael Tokarev]

30.01.2025 12:29, Laurent Vivier wrote:
> When do you plan to send the PR for them?

Um, sorry for the delay.  Will do ASAP.

Thank you for the reminder!

/mjt