[ovs-dev] [PATCH] netdev_afxdp: Detects combined channels and aborts wrong config

[Ilya Maximets]

On 06.01.2020 22:48, William Tu wrote:
> On Mon, Dec 23, 2019 at 11:34 AM Yi-Hung Wei <yihung.wei at gmail.com> wrote:
>>
>> This patches detects the number of combined channels on a AF_XDP port
>> via using ethtool interface.  If the number of combined channels
>> is different from the n_rxq, netdev_afxdp will not be able to get
>> all the packets from that port.  Thus, abort the port setup when
>> the case is detected.
>>
>> Signed-off-by: Yi-Hung Wei <yihung.wei at gmail.com>
> 
> Looks good to me, thanks!
> CC Ilya to see if he has more insight/comments.
> 
> Acked-by: William Tu <u9012063 at gmail.com>
> 
>> ---
>> Travis CI: https://travis-ci.org/YiHungWei/ovs/builds/627972465
>>
>> ---
>>  lib/netdev-afxdp.c | 15 +++++++++++++++
>>  lib/netdev-linux.c | 27 +++++++++++++++++++++++++++
>>  lib/netdev-linux.h |  2 ++
>>  3 files changed, 44 insertions(+)
>>
>> diff --git a/lib/netdev-afxdp.c b/lib/netdev-afxdp.c
>> index 58365ed483e3..6d002882d355 100644
>> --- a/lib/netdev-afxdp.c
>> +++ b/lib/netdev-afxdp.c
>> @@ -601,6 +601,14 @@ netdev_afxdp_set_config(struct netdev *netdev, const struct smap *args,
>>      enum afxdp_mode xdp_mode;
>>      bool need_wakeup;
>>      int new_n_rxq;
>> +    int combined_channels;
>> +
>> +    if (netdev_linux_ethtool_get_combined_channels(netdev,
>> +                                                   &combined_channels)) {
>> +        VLOG_INFO("Cannot get the number of combined channels of %s. "
>> +                  "Defaults it to 1.", netdev_get_name(netdev));
>> +        combined_channels = 1;
>> +    }
>>
>>      ovs_mutex_lock(&dev->mutex);
>>      new_n_rxq = MAX(smap_get_int(args, "n_rxq", NR_QUEUE), 1);
>> @@ -611,6 +619,13 @@ netdev_afxdp_set_config(struct netdev *netdev, const struct smap *args,
>>          return EINVAL;
>>      }
>>
>> +    if (new_n_rxq != combined_channels) {
>> +        ovs_mutex_unlock(&dev->mutex);
>> +        VLOG_ERR("%s: n_rxq: %d != combined channels: %d.",
>> +                 netdev_get_name(netdev), new_n_rxq, combined_channels);
>> +        return EINVAL;
>> +    }
>> +
>>      str_xdp_mode = smap_get_def(args, "xdp-mode", "best-effort");
>>      for (xdp_mode = OVS_AF_XDP_MODE_BEST_EFFORT;
>>           xdp_mode < OVS_AF_XDP_MODE_MAX;
>> diff --git a/lib/netdev-linux.c b/lib/netdev-linux.c
>> index f8e59bacfb13..e3086fc1cbb6 100644
>> --- a/lib/netdev-linux.c
>> +++ b/lib/netdev-linux.c
>> @@ -2166,6 +2166,33 @@ out:
>>      netdev->get_features_error = error;
>>  }
>>
>> +int
>> +netdev_linux_ethtool_get_combined_channels(struct netdev *netdev_,
>> +                                           int *combined_channels)
>> +{
>> +    struct netdev_linux *netdev = netdev_linux_cast(netdev_);
>> +    struct ethtool_channels echannels;
>> +    int err;
>> +
>> +    ovs_mutex_lock(&netdev->mutex);
>> +
>> +    COVERAGE_INC(netdev_get_ethtool);
>> +
>> +    memset(&echannels, 0, sizeof echannels);
>> +    err = netdev_linux_do_ethtool(netdev_get_name(netdev_),
>> +                                  (struct ethtool_cmd *) &echannels,
>> +                                  ETHTOOL_GCHANNELS, "ETHTOOL_GCHANNELS");

Does this command return maximum possible number of channels or
number of currently enabled channels?

In first case we'll end up with a need to configure huge number of queues.
In second case it's not guaranteed that number of channels will not be
changed later.  (Does enabling more channels generates if-notifier event?)

Another point is why we need a configurable parameter that is allowed to
be configured with a single value only?

And there is a possible usecase for not having all the queues attached
to OVS.  For example, if you have a custom xdp program loaded, you may
redirect specific traffic to fast afxdp queues and other traffic to
kernel or system datapath.  This might be useful for quick handling
of undesired or malicious traffic.

However, this still make sense to limit the maximum number of queues
with the number of actually available channels instead of having a
hardcoded constant (MAX_XSKQ).

Best regards, Ilya Maximets.