[ovs-dev] [PATCH 1/3] dpif-netlink: Generate ufids for installing TC flowers

Tonghao Zhang xiangxia.m.yue at gmail.com
Mon Sep 14 05:54:35 UTC 2020


On Fri, Sep 11, 2020 at 5:46 PM Eelco Chaudron <echaudro at redhat.com> wrote:
>
> Hi Tonghao, at al.
>
> This patch is breaking basic flow adding/deletion with ovs-dpctl. I did
> not look at the root cause yet, maybe you can take a peek. This is
> present since 13.0, so the potential fix needs backporting.
>
> Here is how to replicate:
>
> ip link add name vport0 type veth peer name vport1
> modprobe openvswitch
> ovs-dpctl add-dp test
> ovs-dpctl add-if test vport0
> ovs-dpctl add-flow test
> "in_port(0),eth(),eth_type(0x800),ipv4(src=100.1.0.1)" 0
> ovs-dpctl del-flow test
> "in_port(0),eth(),eth_type(0x800),ipv4(src=100.1.0.1)"
>
> The last command will give:
>
> 2020-09-11T09:40:06Z|00001|dpif|WARN|system at test: failed to flow_del (No
> such file or directory) ufid:e4457189-3990-4a01-bdcf-1e5f8b208711
> in_port(0),eth(src=00:00:00:00:00:00,dst=00:00:00:00:00:00),eth_type(0x0800),ipv4(src=100.1.0.1,dst=0.0.0.0,proto=0,tos=0,ttl=0,frag=no)
> ovs-dpctl: deleting flow (No such file or directory)
> Perhaps you need to specify a UFID?
Hi Eelco
In odp_flow_key_hash, OVS will generate a random key for hashing flow.
The commands ovs-appctl dpctl/add-flow and ovs-appctl dpctl/del-flow use the
same key, but ovs-dpctl use the different key for hashing flow.

I haven't found a good way to fix it.
1. don't generate a random key for hashing flow
2. use ovs-appctl instead of ovs-dpctl
3. revert that patch

> Cheers,
>
> Eelco
>
> On 18 May 2020, at 3:44, xiangxia.m.yue at gmail.com wrote:
>
> > From: Tonghao Zhang <xiangxia.m.yue at gmail.com>
> >
> > To support installing the TC flowers to HW, via "ovs-appctl
> > dpctl/add-flow"
> > command, there should be an ufid. This patch will check whether ufid
> > exists,
> > if not, generate an ufid. Should to know that when processing upcall
> > packets,
> > ufid is generated in parse_odp_packet for kernel datapath.
> >
> > Configuring the max-idle/max-revalidator, may help testing this patch.
> >
> > Cc: Simon Horman <simon.horman at netronome.com>
> > Cc: Paul Blakey <paulb at mellanox.com>
> > Cc: Roi Dayan <roid at mellanox.com>
> > Cc: Ben Pfaff <blp at ovn.org>
> > Cc: William Tu <u9012063 at gmail.com>
> > Cc: Ilya Maximets <i.maximets at ovn.org>
> > Signed-off-by: Tonghao Zhang <xiangxia.m.yue at gmail.com>
> > ---
> >  lib/dpif-netlink.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
> >  1 file changed, 45 insertions(+)
> >
> > diff --git a/lib/dpif-netlink.c b/lib/dpif-netlink.c
> > index dc642100fc58..a19ed7e53566 100644
> > --- a/lib/dpif-netlink.c
> > +++ b/lib/dpif-netlink.c
> > @@ -2231,12 +2231,55 @@ dpif_netlink_operate_chunks(struct
> > dpif_netlink *dpif, struct dpif_op **ops,
> >      }
> >  }
> >
> > +static void
> > +dpif_netlink_try_update_ufid__(struct dpif_op *op, ovs_u128 *ufid)
> > +{
> > +    switch (op->type) {
> > +    case DPIF_OP_FLOW_PUT:
> > +        if (!op->flow_put.ufid) {
> > +            odp_flow_key_hash(op->flow_put.key, op->flow_put.key_len,
> > +                              ufid);
> > +            op->flow_put.ufid = ufid;
> > +        }
> > +        break;
> > +    case DPIF_OP_FLOW_DEL:
> > +        if (!op->flow_del.ufid) {
> > +            odp_flow_key_hash(op->flow_del.key, op->flow_del.key_len,
> > +                              ufid);
> > +            op->flow_del.ufid = ufid;
> > +        }
> > +        break;
> > +    case DPIF_OP_FLOW_GET:
> > +        if (!op->flow_get.ufid) {
> > +            odp_flow_key_hash(op->flow_get.key, op->flow_get.key_len,
> > +                              ufid);
> > +            op->flow_get.ufid = ufid;
> > +        }
> > +        break;
> > +    case DPIF_OP_EXECUTE:
> > +    default:
> > +        break;
> > +    }
> > +}
> > +
> > +static void
> > +dpif_netlink_try_update_ufid(struct dpif_op **ops, ovs_u128 *ufid,
> > +                             size_t n_ops)
> > +{
> > +    int i;
> > +
> > +    for (i = 0; i < n_ops; i++) {
> > +        dpif_netlink_try_update_ufid__(ops[i], &ufid[i]);
> > +    }
> > +}
> > +
> >  static void
> >  dpif_netlink_operate(struct dpif *dpif_, struct dpif_op **ops, size_t
> > n_ops,
> >                       enum dpif_offload_type offload_type)
> >  {
> >      struct dpif_netlink *dpif = dpif_netlink_cast(dpif_);
> >      struct dpif_op *new_ops[OPERATE_MAX_OPS];
> > +    ovs_u128 ufids[OPERATE_MAX_OPS];
> >      int count = 0;
> >      int i = 0;
> >      int err = 0;
> > @@ -2246,6 +2289,8 @@ dpif_netlink_operate(struct dpif *dpif_, struct
> > dpif_op **ops, size_t n_ops,
> >          return;
> >      }
> >
> > +    dpif_netlink_try_update_ufid(ops, ufids, n_ops);
> > +
> >      if (offload_type != DPIF_OFFLOAD_NEVER &&
> > netdev_is_flow_api_enabled()) {
> >          while (n_ops > 0) {
> >              count = 0;
> > --
> > 2.26.1
> >
> > _______________________________________________
> > dev mailing list
> > dev at openvswitch.org
> > https://mail.openvswitch.org/mailman/listinfo/ovs-dev
>


-- 
Best regards, Tonghao


More information about the dev mailing list