Re: [dpdk-dev] [PATCH 4/4] net/tap: add basic flow API patterns and actions

[Pascal Mazon <pascal.mazon@6wind.com>]

On Fri, 3 Mar 2017 15:47:58 +0000
"Wiles, Keith" <keith.wiles@intel.com> wrote:

> 
> > On Mar 3, 2017, at 4:45 AM, Pascal Mazon <pascal.mazon@6wind.com>
> > wrote:
> > 
> > Supported flow rules are now mapped to TC rules on the tap
> > netdevice. The netlink message used for creating the TC rule is
> > stored in struct rte_flow. That way, by simply changing a metadata
> > in it, we can require for the rule deletion without further parsing.
> > 
> > Supported items:
> > - eth: src and dst (with variable masks), and eth_type (0xffff
> > mask).
> > - vlan: vid, pcp, tpid, but not eid.
> > - ipv4/6: src and dst (with variable masks), and ip_proto (0xffff
> > mask).
> > - udp/tcp: src and dst port (0xffff) mask.
> > 
> > Supported actions:
> > - DROP
> > - QUEUE
> > - PASSTHRU
> > 
> > It is generally not possible to provide a "last" item. However, if
> > the "last" item, once masked, is identical to the masked spec, then
> > it is supported.
> > 
> > Only IPv4/6 and MAC addresses can use a variable mask. All other
> > items need a full mask (exact match).
> > 
> > Support for VLAN requires kernel headers >= 4.9, checked using
> > auto-config.sh.
> > 
> > Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com>
> > Acked-by: Olga Shern <olgas@mellanox.com>
> > —
> 
> …
> 
> > /**
> > @@ -120,13 +965,33 @@ tap_flow_create(struct rte_eth_dev *dev,
> >  * @see rte_flow_ops
> >  */
> > static int
> > -tap_flow_destroy(struct rte_eth_dev *dev __rte_unused,
> > +tap_flow_destroy(struct rte_eth_dev *dev,
> > 		 struct rte_flow *flow,
> > -		 struct rte_flow_error *error __rte_unused)
> > +		 struct rte_flow_error *error)
> > {
> > +	struct pmd_internals *pmd = dev->data->dev_private;
> > +	int ret = 0;
> > +
> > 	LIST_REMOVE(flow, next);
> > +	flow->msg.nh.nlmsg_flags = NLM_F_REQUEST | NLM_F_ACK;
> > +	flow->msg.nh.nlmsg_type = RTM_DELTFILTER;
> > +
> > +	ret = nl_send(pmd->nlsk_fd, &flow->msg.nh);
> > +	if (ret < 0) {
> > +		rte_flow_error_set(error, ENOTSUP,
> > RTE_FLOW_ERROR_TYPE_HANDLE,
> > +				   NULL, "couldn't send request to
> > kernel");
> > +		goto end;
> > +	}
> > +	ret = nl_recv_ack(pmd->nlsk_fd);
> > +	if (ret < 0) {
> > +		rte_flow_error_set(error, ENOTSUP,
> > RTE_FLOW_ERROR_TYPE_HANDLE,
> > +				   NULL,
> > +				   "couldn't receive kernel ack to
> > our request");
> > +		goto end;
> 
> This goto is not required.
> 

Indeed!
I'll fix that in v2.

Regards,
Pascal

> > +	}
> > +end:
> > 	rte_free(flow);
> > -	return 0;
> > +	return ret;
> > }
> > 
> > /**
> > diff --git a/drivers/net/tap/tap_flow.h b/drivers/net/tap/tap_flow.h
> > index 377a9f7b758a..a05e945df523 100644
> > --- a/drivers/net/tap/tap_flow.h
> > +++ b/drivers/net/tap/tap_flow.h
> > @@ -37,6 +37,18 @@
> > #include <rte_flow.h>
> > #include <rte_flow_driver.h>
> > 
> > +/**
> > + * In TC, priority 0 means we require the kernel to allocate one
> > for us.
> > + * In rte_flow, however, we want the priority 0 to be the most
> > important one.
> > + * Use an offset to have the most important priority being 1 in TC.
> > + */
> > +#define PRIORITY_OFFSET 1
> > +#define PRIORITY_MASK (0xfff)
> > +#define MAX_PRIORITY (PRIORITY_MASK - PRIORITY_OFFSET)
> > +#define GROUP_MASK (0xf)
> > +#define GROUP_SHIFT 12
> > +#define MAX_GROUP GROUP_MASK
> > +
> > int tap_dev_filter_ctrl(struct rte_eth_dev *dev,
> > 			enum rte_filter_type filter_type,
> > 			enum rte_filter_op filter_op,
> > -- 
> > 2.8.0.rc0
> > 
> 
> Regards,
> Keith
>