[ovs-dev] [PATCH 1/3] netdev-dpdk: Fix and document vhost tx retries.

[Ilya Maximets]

On 25.06.2019 12:32, Kevin Traynor wrote:
> On 24/06/2019 22:40, Flavio Leitner wrote:
>> On Fri, Jun 21, 2019 at 02:41:56PM +0100, Kevin Traynor wrote:
>>> Fix minor issue of one additional retry and add
>>> documentation about vhost tx retries and ways to
>>> reduce/remove them.
>>>
>>> Signed-off-by: Kevin Traynor <ktraynor at redhat.com>
>>>
>>> ---
>>> There is a checkpatch warning that one of the libvirt
>>> lines in docs is a few characters too long. It is similar
>>> for some others in the file and I think it makes sense to
>>> leave it as is but can wrap if required.
>>> ---
>>>  Documentation/topics/dpdk/vhost-user.rst | 31 ++++++++++++++++++++++++
>>>  lib/netdev-dpdk.c                        |  2 +-
>>>  2 files changed, 32 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/Documentation/topics/dpdk/vhost-user.rst b/Documentation/topics/dpdk/vhost-user.rst
>>> index f7b4b338e..d1682fd05 100644
>>> --- a/Documentation/topics/dpdk/vhost-user.rst
>>> +++ b/Documentation/topics/dpdk/vhost-user.rst
>>> @@ -76,4 +76,35 @@ mode ports require QEMU version 2.7.  Ports of type vhost-user are currently
>>>  deprecated and will be removed in a future release.
>>>  
>>> +vhost tx retries
>>> +~~~~~~~~~~~~~~~~
>>> +
>>> +When sending a batch of packets (max is 32) to a vhost-user or
>>
>> Maybe use the actual define name (NETDEV_MAX_BURST) since it might change?
>>
> 
> I agree the doc could go stale if it ever changed. OTOH, I want a user
> to be able to read the doc part as a standalone without having to search
> the code. I think best to say NETDEV_MAX_BURST and add it's currently
> 32. If it ever changed in the code, an author or reviewer would find the
> comment through a search and be able to update.
> 
>>> +vhost-user-client interface, it may happen that some but not all of the packets
>>> +in the batch are able to be sent to the guest. This is often because there is
>>> +not enough free descriptors in the virtqueue for all the packets to be sent.
>>> +In this case there will be a retry, with a default maximum of 8 occurring. If
>>> +at any time no packets can be sent, it may mean the guest is not accepting
>>> +packets, so there are no (more) retries.
>>> +
>>> +Tx Retries may be reduced or removed by some external configuration, such
>>
>> s/removed/even avoided/ ?
>>
> 
> That sounds better, will change it.
> 
>>> +as increasing the virtqueue size through the ``rx_queue_size`` parameter in
>>> +libvirt::
>>> +
>>> +  <interface type='vhostuser'>
>>> +      <mac address='56:48:4f:53:54:01'/>
>>> +      <source type='unix' path='/tmp/dpdkvhostclient0' mode='server'/>
>>> +      <model type='virtio'/>
>>> +      <driver name='vhost' rx_queue_size='1024' tx_queue_size='1024'/>
>>> +      <address type='pci' domain='0x0000' bus='0x00' slot='0x10' function='0x0'/>
>>> +  </interface>
>>
>> Do we need to worry about the qemu version?
>>
> 
> hmm, yeah, it was definitely not always available. I will check the
> minimum version and add.
> 
>>> +
>>> +The guest application will also need need to provide enough descriptors. For
>>> +example with ``testpmd`` the command line argument can be used::
>>> +
>>> + --rxd=1024 --txd=1024
>>> +
>>> +The guest should also have sufficient cores dedicated for consuming and
>>> +processing packets at the required rate.
>>> +
>>>  .. _dpdk-vhost-user:
>>>  
>>> diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c
>>> index 4856c56aa..0f3b9c6f4 100644
>>> --- a/lib/netdev-dpdk.c
>>> +++ b/lib/netdev-dpdk.c
>>> @@ -2353,5 +2353,5 @@ __netdev_dpdk_vhost_send(struct netdev *netdev, int qid,
>>>              break;
>>>          }
>>> -    } while (cnt && (retries++ <= VHOST_ENQ_RETRY_NUM));
>>> +    } while (cnt && (retries++ < VHOST_ENQ_RETRY_NUM));
>>
>> You're removing the packet's last hope to reach home, are you sure? :-)
> 
> pop-up dialog box coming in v2 :-)

Maybe we could just increase VHOST_ENQ_RETRY_NUM by 1 to avoid any
functional changes in this patch?