net/tap: support flow API

In the next patch, access to struct pmd_internals will be necessary in tap_flow.c to store the flows.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

---

drivers/net/tap/Makefile      |  1 +
drivers/net/tap/rte_eth_tap.c | 34 ++------------------
drivers/net/tap/tap.h         | 73 +++++++++++++++++++++++++++++++++++++++++++
3 files changed, 76 insertions(+), 32 deletions(-)

create mode 100644 drivers/net/tap/tap.h

drivers/net/tap/Makefile +0 -0

index e18f30c56f52..bdbe69e62a4e 100644
@@ -40,6 +40,7 @@ EXPORT_MAP := rte_pmd_tap_version.map

LIBABIVER := 1

CFLAGS += -O3
+CFLAGS += -I$(SRCDIR)
CFLAGS += $(WERROR_FLAGS)

#

drivers/net/tap/rte_eth_tap.c +2 -32

index 3fd057225ab3..fa57d645f3b1 100644
@@ -51,6 +51,8 @@
 #include <linux/if_ether.h>
 #include <fcntl.h>
 
+#include <tap.h>
+
 /* Linux based path to the TUN device */
 #define TUN_TAP_DEV_PATH        "/dev/net/tun"
 #define DEFAULT_TAP_NAME        "dtap"

@@ -83,38 +85,6 @@ static struct rte_eth_link pmd_link = {
 	.link_autoneg = ETH_LINK_SPEED_AUTONEG
 };
 
-struct pkt_stats {
-	uint64_t opackets;		/* Number of output packets */
-	uint64_t ipackets;		/* Number of input packets */
-	uint64_t obytes;		/* Number of bytes on output */
-	uint64_t ibytes;		/* Number of bytes on input */
-	uint64_t errs;			/* Number of error packets */
-};
-
-struct rx_queue {
-	struct rte_mempool *mp;		/* Mempool for RX packets */
-	uint16_t in_port;		/* Port ID */
-	int fd;
-
-	struct pkt_stats stats;		/* Stats for this RX queue */
-};
-
-struct tx_queue {
-	int fd;
-	struct pkt_stats stats;		/* Stats for this TX queue */
-};
-
-struct pmd_internals {
-	char name[RTE_ETH_NAME_MAX_LEN];	/* Internal Tap device name */
-	uint16_t nb_queues;		/* Number of queues supported */
-	struct ether_addr eth_addr;	/* Mac address of the device port */
-
-	int if_index;			/* IF_INDEX for the port */
-
-	struct rx_queue rxq[RTE_PMD_TAP_MAX_QUEUES];	/* List of RX queues */
-	struct tx_queue txq[RTE_PMD_TAP_MAX_QUEUES];	/* List of TX queues */
-};
-
 /* Tun/Tap allocation routine
 *
 * name is the number of the interface to use, unless NULL to take the host

drivers/net/tap/tap.h new file +61 -0

new file mode 100644
index 000000000000..88f62b895feb
@@ -0,0 +1,73 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright 2017 6WIND S.A.
+ *   Copyright 2017 Mellanox.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ *     * Redistributions of source code must retain the above copyright
+ *       notice, this list of conditions and the following disclaimer.
+ *     * Redistributions in binary form must reproduce the above copyright
+ *       notice, this list of conditions and the following disclaimer in
+ *       the documentation and/or other materials provided with the
+ *       distribution.
+ *     * Neither the name of 6WIND S.A. nor the names of its
+ *       contributors may be used to endorse or promote products derived
+ *       from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _TAP_H_
+#define _TAP_H_
+
+#include <inttypes.h>
+
+#include <rte_ethdev.h>
+#include <rte_ether.h>
+
+#define RTE_PMD_TAP_MAX_QUEUES 16
+
+struct pkt_stats {
+	uint64_t opackets; /* Number of output packets */
+	uint64_t ipackets; /* Number of input packets */
+	uint64_t obytes; /* Number of bytes on output */
+	uint64_t ibytes; /* Number of bytes on input */
+	uint64_t errs; /* Number of error packets */
+};
+
+struct rx_queue {
+	struct rte_mempool *mp; /* Mempool for RX packets */
+	uint16_t in_port; /* Port ID */
+	int fd;
+	struct pkt_stats stats; /* Stats for this RX queue */
+};
+
+struct tx_queue {
+	int fd;
+	struct pkt_stats stats; /* Stats for this TX queue */

drivers/net/tap/tap.h new file +2 -0

+
+#endif /* _TAP_H_ */

Signature · 1 line

2.8.0.rc0

Quoted 185 lines

On Mar 3, 2017, at 4:45 AM, Pascal Mazon <[email protected]> wrote:

In the next patch, access to struct pmd_internals will be necessary in tap_flow.c to store the flows.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

---

drivers/net/tap/Makefile      |  1 +
drivers/net/tap/rte_eth_tap.c | 34 ++------------------
drivers/net/tap/tap.h         | 73

+++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 76 insertions(+), 32 deletions(-) create mode 100644 drivers/net/tap/tap.h

drivers/net/tap/Makefile +0 -0

index e18f30c56f52..bdbe69e62a4e 100644
@@ -40,6 +40,7 @@ EXPORT_MAP := rte_pmd_tap_version.map

LIBABIVER := 1

CFLAGS += -O3
+CFLAGS += -I$(SRCDIR)
CFLAGS += $(WERROR_FLAGS)

#

(quoted diff) +0 -0

b/drivers/net/tap/rte_eth_tap.c index 3fd057225ab3..fa57d645f3b1 100644 --- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c

drivers/net/tap/Makefile +2 -2

@@ -51,6 +51,8 @@
 #include <linux/if_ether.h>
 #include <fcntl.h>
 
+#include <tap.h>
+
 /* Linux based path to the TUN device */
 #define TUN_TAP_DEV_PATH        "/dev/net/tun"
 #define DEFAULT_TAP_NAME        "dtap"

@@ -83,38 +85,6 @@ static struct rte_eth_link pmd_link = {
 	.link_autoneg = ETH_LINK_SPEED_AUTONEG
 };
 
-struct pkt_stats {
-	uint64_t opackets;		/* Number of output

packets */

drivers/net/tap/Makefile +0 -1

-	uint64_t ipackets;		/* Number of input

packets */

drivers/net/tap/Makefile +0 -1

-	uint64_t obytes;		/* Number of bytes on

output */

drivers/net/tap/Makefile +0 -1

-	uint64_t ibytes;		/* Number of bytes on

input */

drivers/net/tap/Makefile +0 -1

-	uint64_t errs;			/* Number of error

packets */ -};

drivers/net/tap/Makefile +0 -3

-
-struct rx_queue {
-	struct rte_mempool *mp;		/* Mempool for RX

packets */

drivers/net/tap/Makefile +0 -4

-	uint16_t in_port;		/* Port ID */
-	int fd;
-
-	struct pkt_stats stats;		/* Stats for this

RX queue */ -};

drivers/net/tap/Makefile +0 -4

-
-struct tx_queue {
-	int fd;
-	struct pkt_stats stats;		/* Stats for this

TX queue */ -};

drivers/net/tap/Makefile +0 -3

-
-struct pmd_internals {
-	char name[RTE_ETH_NAME_MAX_LEN];	/* Internal Tap

device name */

drivers/net/tap/Makefile +0 -1

-	uint16_t nb_queues;		/* Number of queues

supported */

drivers/net/tap/Makefile +0 -1

-	struct ether_addr eth_addr;	/* Mac address of the

device port */ -

drivers/net/tap/Makefile +0 -1

-	int if_index;			/* IF_INDEX for the

port */ -

drivers/net/tap/Makefile +0 -1

-	struct rx_queue rxq[RTE_PMD_TAP_MAX_QUEUES];	/*

List of RX queues */

drivers/net/tap/Makefile +0 -1

-	struct tx_queue txq[RTE_PMD_TAP_MAX_QUEUES];	/*

List of TX queues */ -};
-
/* Tun/Tap allocation routine
*
* name is the number of the interface to use, unless NULL to take
the host diff --git a/drivers/net/tap/tap.h b/drivers/net/tap/tap.h new file mode 100644
index 000000000000..88f62b895feb

drivers/net/tap/tap.h new file +7 -0

@@ -0,0 +1,73 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright 2017 6WIND S.A.
+ *   Copyright 2017 Mellanox.
+ *
+ *   Redistribution and use in source and binary forms, with or

without
+ * modification, are permitted provided that the following
conditions
+ * are met:
+ *
+ * * Redistributions of source code must retain the above
copyright
+ * notice, this list of conditions and the following
disclaimer.
+ * * Redistributions in binary form must reproduce the above
copyright
+ * notice, this list of conditions and the following
disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ * * Neither the name of 6WIND S.A. nor the names of its
+ * contributors may be used to endorse or promote products
derived
+ * from this software without specific prior written
permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT
NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
DAMAGE.
+ */

drivers/net/tap/tap.h new file +33 -0

+
+#ifndef _TAP_H_
+#define _TAP_H_
+
+#include <inttypes.h>
+
+#include <rte_ethdev.h>
+#include <rte_ether.h>
+
+#define RTE_PMD_TAP_MAX_QUEUES 16
+
+struct pkt_stats {
+	uint64_t opackets; /* Number of output packets */
+	uint64_t ipackets; /* Number of input packets */
+	uint64_t obytes; /* Number of bytes on output */
+	uint64_t ibytes; /* Number of bytes on input */
+	uint64_t errs; /* Number of error packets */
+};
+
+struct rx_queue {
+	struct rte_mempool *mp; /* Mempool for RX packets */
+	uint16_t in_port; /* Port ID */
+	int fd;
+	struct pkt_stats stats; /* Stats for this RX queue */
+};
+
+struct tx_queue {
+	int fd;
+	struct pkt_stats stats; /* Stats for this TX queue */
+};
+
+struct pmd_internals {
+	char name[RTE_ETH_NAME_MAX_LEN]; /* Internal Tap device

name */

drivers/net/tap/tap.h new file +2 -0

+	uint16_t nb_queues; /* Number of queues supported */
+	struct ether_addr eth_addr; /* Mac address of the device

port */

drivers/net/tap/tap.h new file +2 -0

+	int if_index; /* IF_INDEX for the port */
+	struct rx_queue rxq[RTE_PMD_TAP_MAX_QUEUES]; /* List of RX

Quoted 3 lines

queues */

drivers/net/tap/tap.h new file +1 -0

+	struct tx_queue txq[RTE_PMD_TAP_MAX_QUEUES]; /* List of TX

queues */ +};

I guess I am going to be a bit picky here on the formatting. Moving the code from .c to .h you compress a lot of white space out and now I think it is very hard to read. Can you add back some of the white space for readability.

Quoted 5 lines

drivers/net/tap/tap.h new file +2 -0

+
+#endif /* _TAP_H_ */

Signature · 1 line

2.8.0.rc0

Regards,
Keith

Quoted 185 lines · Keith

Quoted 185 lines

On Mar 3, 2017, at 4:45 AM, Pascal Mazon <[email protected]> wrote:

In the next patch, access to struct pmd_internals will be necessary in tap_flow.c to store the flows.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

---

drivers/net/tap/Makefile      |  1 +
drivers/net/tap/rte_eth_tap.c | 34 ++------------------
drivers/net/tap/tap.h         | 73

+++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 76 insertions(+), 32 deletions(-) create mode 100644 drivers/net/tap/tap.h

drivers/net/tap/Makefile +0 -0

index e18f30c56f52..bdbe69e62a4e 100644
@@ -40,6 +40,7 @@ EXPORT_MAP := rte_pmd_tap_version.map

LIBABIVER := 1

CFLAGS += -O3
+CFLAGS += -I$(SRCDIR)
CFLAGS += $(WERROR_FLAGS)

#

(quoted diff) +0 -0

b/drivers/net/tap/rte_eth_tap.c index 3fd057225ab3..fa57d645f3b1 100644 --- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c

drivers/net/tap/Makefile +2 -2

@@ -51,6 +51,8 @@
 #include <linux/if_ether.h>
 #include <fcntl.h>
 
+#include <tap.h>
+
 /* Linux based path to the TUN device */
 #define TUN_TAP_DEV_PATH        "/dev/net/tun"
 #define DEFAULT_TAP_NAME        "dtap"

@@ -83,38 +85,6 @@ static struct rte_eth_link pmd_link = {
 	.link_autoneg = ETH_LINK_SPEED_AUTONEG
 };
 
-struct pkt_stats {
-	uint64_t opackets;		/* Number of output

packets */

drivers/net/tap/Makefile +0 -1

-	uint64_t ipackets;		/* Number of input

packets */

drivers/net/tap/Makefile +0 -1

-	uint64_t obytes;		/* Number of bytes on

output */

drivers/net/tap/Makefile +0 -1

-	uint64_t ibytes;		/* Number of bytes on

input */

drivers/net/tap/Makefile +0 -1

-	uint64_t errs;			/* Number of error

packets */ -};

drivers/net/tap/Makefile +0 -3

-
-struct rx_queue {
-	struct rte_mempool *mp;		/* Mempool for RX

packets */

drivers/net/tap/Makefile +0 -4

-	uint16_t in_port;		/* Port ID */
-	int fd;
-
-	struct pkt_stats stats;		/* Stats for this

RX queue */ -};

drivers/net/tap/Makefile +0 -4

-
-struct tx_queue {
-	int fd;
-	struct pkt_stats stats;		/* Stats for this

TX queue */ -};

drivers/net/tap/Makefile +0 -3

-
-struct pmd_internals {
-	char name[RTE_ETH_NAME_MAX_LEN];	/* Internal Tap

device name */

drivers/net/tap/Makefile +0 -1

-	uint16_t nb_queues;		/* Number of queues

supported */

drivers/net/tap/Makefile +0 -1

-	struct ether_addr eth_addr;	/* Mac address of the

device port */ -

drivers/net/tap/Makefile +0 -1

-	int if_index;			/* IF_INDEX for the

port */ -

drivers/net/tap/Makefile +0 -1

-	struct rx_queue rxq[RTE_PMD_TAP_MAX_QUEUES];	/*

List of RX queues */

drivers/net/tap/Makefile +0 -1

-	struct tx_queue txq[RTE_PMD_TAP_MAX_QUEUES];	/*

List of TX queues */ -};
-
/* Tun/Tap allocation routine
*
* name is the number of the interface to use, unless NULL to take the host diff --git a/drivers/net/tap/tap.h b/drivers/net/tap/tap.h new file mode 100644
index 000000000000..88f62b895feb

drivers/net/tap/tap.h new file +7 -0

@@ -0,0 +1,73 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright 2017 6WIND S.A.
+ *   Copyright 2017 Mellanox.
+ *
+ *   Redistribution and use in source and binary forms, with or

without
+ * modification, are permitted provided that the following
conditions
+ * are met:
+ *
+ * * Redistributions of source code must retain the above
copyright
+ * notice, this list of conditions and the following
disclaimer.
+ * * Redistributions in binary form must reproduce the above
copyright
+ * notice, this list of conditions and the following
disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ * * Neither the name of 6WIND S.A. nor the names of its
+ * contributors may be used to endorse or promote products
derived
+ * from this software without specific prior written
permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT
NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
DAMAGE.
+ */

drivers/net/tap/tap.h new file +33 -0

+
+#ifndef _TAP_H_
+#define _TAP_H_
+
+#include <inttypes.h>
+
+#include <rte_ethdev.h>
+#include <rte_ether.h>
+
+#define RTE_PMD_TAP_MAX_QUEUES 16
+
+struct pkt_stats {
+	uint64_t opackets; /* Number of output packets */
+	uint64_t ipackets; /* Number of input packets */
+	uint64_t obytes; /* Number of bytes on output */
+	uint64_t ibytes; /* Number of bytes on input */
+	uint64_t errs; /* Number of error packets */
+};
+
+struct rx_queue {
+	struct rte_mempool *mp; /* Mempool for RX packets */
+	uint16_t in_port; /* Port ID */
+	int fd;
+	struct pkt_stats stats; /* Stats for this RX queue */
+};
+
+struct tx_queue {
+	int fd;
+	struct pkt_stats stats; /* Stats for this TX queue */
+};
+
+struct pmd_internals {
+	char name[RTE_ETH_NAME_MAX_LEN]; /* Internal Tap device

name */

drivers/net/tap/tap.h new file +2 -0

+	uint16_t nb_queues; /* Number of queues supported */
+	struct ether_addr eth_addr; /* Mac address of the device

port */

drivers/net/tap/tap.h new file +2 -0

+	int if_index; /* IF_INDEX for the port */
+	struct rx_queue rxq[RTE_PMD_TAP_MAX_QUEUES]; /* List of RX

Quoted 9 lines

Quoted 3 lines

queues */

drivers/net/tap/tap.h new file +1 -0

+	struct tx_queue txq[RTE_PMD_TAP_MAX_QUEUES]; /* List of TX

queues */ +};

I guess I am going to be a bit picky here on the formatting. Moving the code from .c to .h you compress a lot of white space out and now I think it is very hard to read. Can you add back some of the white space for readability.

Do you mean whitespaces between ";" and the comment, or do you mean line jumps? Would you rather have it look like what's described there?
http://dpdk.org/doc/guides/contributing/coding_style.html#structure-declarations

Regards,
Pascal

Quoted 9 lines

Quoted 5 lines

drivers/net/tap/tap.h new file +2 -0

+
+#endif /* _TAP_H_ */

Signature · 1 line

2.8.0.rc0

Regards,
Keith

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 <[email protected]>

Acked-by Olga Shern <[email protected]>

—

/**

(quoted diff) +0 -0

@@ -120,13 +965,33 @@ tap_flow_create(struct rte_eth_dev *dev,
 * @see rte_flow_ops
 */

static int

drivers/net/tap/tap_flow.c +1 -1

-tap_flow_destroy(struct rte_eth_dev *dev __rte_unused,
+tap_flow_destroy(struct rte_eth_dev *dev,

struct rte_flow *flow,

drivers/net/tap/tap_flow.c +1 -1

-		 struct rte_flow_error *error __rte_unused)
+		 struct rte_flow_error *error)

{

drivers/net/tap/tap_flow.c +3 -0

+	struct pmd_internals *pmd = dev->data->dev_private;
+	int ret = 0;
+

	LIST_REMOVE(flow, next);

drivers/net/tap/tap_flow.c +5 -0

+	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) {

drivers/net/tap/tap_flow.c +2 -0

+	}
+end:

	rte_free(flow);

drivers/net/tap/tap_flow.c +1 -1

-	return 0;
+	return ret;

}

/**

drivers/net/tap/tap_flow.h +12 -0

index 377a9f7b758a..a05e945df523 100644
@@ -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,

Signature · 1 line

2.8.0.rc0

Quoted 35 lines

On Mar 3, 2017, at 4:45 AM, Pascal Mazon <[email protected]> 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 <[email protected]>

Acked-by Olga Shern <[email protected]>

—

…

Quoted 28 lines

/**

(quoted diff) +0 -0

@@ -120,13 +965,33 @@ tap_flow_create(struct rte_eth_dev *dev,
 * @see rte_flow_ops
 */

static int

drivers/net/tap/tap_flow.c +1 -1

-tap_flow_destroy(struct rte_eth_dev *dev __rte_unused,
+tap_flow_destroy(struct rte_eth_dev *dev,

struct rte_flow *flow,

drivers/net/tap/tap_flow.c +1 -1

-		 struct rte_flow_error *error __rte_unused)
+		 struct rte_flow_error *error)

{

drivers/net/tap/tap_flow.c +3 -0

+	struct pmd_internals *pmd = dev->data->dev_private;
+	int ret = 0;
+

	LIST_REMOVE(flow, next);

drivers/net/tap/tap_flow.c +6 -0

+	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,

(quoted diff) +1 -0

+				   NULL, "couldn't send request to

kernel");

drivers/net/tap/tap_flow.c +4 -0

+		goto end;
+	}
+	ret = nl_recv_ack(pmd->nlsk_fd);
+	if (ret < 0) {

Review

Quoted 6 lines

(quoted diff) +1 -0

+		rte_flow_error_set(error, ENOTSUP,

RTE_FLOW_ERROR_TYPE_HANDLE,

(quoted diff) +2 -0

+				   NULL,
+				   "couldn't receive kernel ack to

our request");

(quoted diff) +1 -0

+		goto end;

This goto is not required.

Quoted 30 lines

drivers/net/tap/tap_flow.c +2 -0

+	}
+end:

	rte_free(flow);

drivers/net/tap/tap_flow.c +1 -1

-	return 0;
+	return ret;

}

/**

drivers/net/tap/tap_flow.h +2 -0

index 377a9f7b758a..a05e945df523 100644
@@ -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.
+ */

drivers/net/tap/tap_flow.h +7 -0

+#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
+

Quoted 6 lines

int tap_dev_filter_ctrl(struct rte_eth_dev *dev,
enum rte_filter_type filter_type, enum rte_filter_op filter_op,

Signature · 1 line

2.8.0.rc0

Regards,
Keith

This series add support for the flow API in tap PMD.

It enables filtering specific packets incoming on the tap netdevice, to process only desired ones. Under the hood, it uses kernel TC (traffic control), which takes place very early in the stack, and supports most common pattern items and actions defined in the flow API.

This series applies on top of:

[PATCH 0/6] net/tap: add additional management ops

Pascal Mazon (4):
net/tap: move private elements to external header net/tap: add preliminary support for rte_flow net/tap: add netlink back-end for flow API net/tap: add basic flow API patterns and actions

doc/guides/nics/features/tap.ini |    1 +
drivers/net/tap/Makefile         |   39 ++
drivers/net/tap/rte_eth_tap.c    |   73 +--
drivers/net/tap/tap.h            |   76 +++
drivers/net/tap/tap_flow.c       | 1050 ++++++++++++++++++++++++++++++++++++++
drivers/net/tap/tap_flow.h       |   58 +++
drivers/net/tap/tap_netlink.c    |  367 +++++++++++++
drivers/net/tap/tap_netlink.h    |   69 +++
drivers/net/tap/tap_tcmsgs.c     |  378 ++++++++++++++
drivers/net/tap/tap_tcmsgs.h     |   63 +++
10 files changed, 2140 insertions(+), 34 deletions(-)

create mode 100644 drivers/net/tap/tap.h
create mode 100644 drivers/net/tap/tap_flow.c
create mode 100644 drivers/net/tap/tap_flow.h
create mode 100644 drivers/net/tap/tap_netlink.c
create mode 100644 drivers/net/tap/tap_netlink.h
create mode 100644 drivers/net/tap/tap_tcmsgs.c
create mode 100644 drivers/net/tap/tap_tcmsgs.h

Signature · 2 lines

-- 
2.8.0.rc0

This series add support for the flow API in tap PMD.

It enables filtering specific packets incoming on the tap netdevice, to process only desired ones. Under the hood, it uses kernel TC (traffic control), which takes place very early in the stack, and supports most common pattern items and actions defined in the flow API.

This series applies on top of:

[PATCH 0/6] net/tap: add additional management ops

v2 changes:
- support compilation on kernels < 4.2 (where flower support
appeared)
- set whitespaces in tap.h
- remove unnecessary goto

Pascal Mazon (4):
net/tap: move private elements to external header net/tap: add preliminary support for rte_flow net/tap: add netlink back-end for flow API net/tap: add basic flow API patterns and actions

 doc/guides/nics/features/tap.ini |    1 +
 drivers/net/tap/Makefile         |   44 ++
 drivers/net/tap/rte_eth_tap.c    |   94 ++--
 drivers/net/tap/tap.h            |   77 +++
 drivers/net/tap/tap_flow.c       | 1084
++++++++++++++++++++++++++++++++++++++
drivers/net/tap/tap_flow.h       |   58 ++
drivers/net/tap/tap_netlink.c    |  367 +++++++++++++
drivers/net/tap/tap_netlink.h    |   69 +++
drivers/net/tap/tap_tcmsgs.c     |  378 +++++++++++++
drivers/net/tap/tap_tcmsgs.h     |   63 +++ 10 files changed, 2202
insertions(+), 33 deletions(-) create mode 100644
drivers/net/tap/tap.h create mode 100644 drivers/net/tap/tap_flow.c
 create mode 100644 drivers/net/tap/tap_flow.h
 create mode 100644 drivers/net/tap/tap_netlink.c
 create mode 100644 drivers/net/tap/tap_netlink.h
 create mode 100644 drivers/net/tap/tap_tcmsgs.c
 create mode 100644 drivers/net/tap/tap_tcmsgs.h

Hi Keith,

I'm working on a v3 for that series:

- I added info regarding flow API support in tap.rst doc
- I fixed support for kernels where flower/vlan was not supported.

Do you have any other remarks, or can I send the v3 (hopefully ok for integration)?

I'll send a v2 for the latest series (introducing remote capture), because struct pmd_internals changed (whitespaces). Same question there, do you have remarks?

Thank you.

Best regards,
Pascal

Quoted 43 lines · On Mon, 6 Mar 2017 18:05:26 +0100 Pascal Mazon <[email protected]> wrote:

This series add support for the flow API in tap PMD.

It enables filtering specific packets incoming on the tap netdevice, to process only desired ones. Under the hood, it uses kernel TC (traffic control), which takes place very early in the stack, and supports most common pattern items and actions defined in the flow API.

This series applies on top of:

[PATCH 0/6] net/tap: add additional management ops

v2 changes:
- support compilation on kernels < 4.2 (where flower support
appeared)
- set whitespaces in tap.h
- remove unnecessary goto

Pascal Mazon (4):
net/tap: move private elements to external header net/tap: add preliminary support for rte_flow net/tap: add netlink back-end for flow API net/tap: add basic flow API patterns and actions

doc/guides/nics/features/tap.ini |    1 +
drivers/net/tap/Makefile         |   44 ++
drivers/net/tap/rte_eth_tap.c    |   94 ++--
drivers/net/tap/tap.h            |   77 +++
drivers/net/tap/tap_flow.c       | 1084
++++++++++++++++++++++++++++++++++++++
drivers/net/tap/tap_flow.h       |   58 ++
drivers/net/tap/tap_netlink.c    |  367 +++++++++++++
drivers/net/tap/tap_netlink.h    |   69 +++
drivers/net/tap/tap_tcmsgs.c     |  378 +++++++++++++
drivers/net/tap/tap_tcmsgs.h     |   63 +++ 10 files changed, 2202

insertions(+), 33 deletions(-) create mode 100644 drivers/net/tap/tap.h create mode 100644 drivers/net/tap/tap_flow.c create mode 100644 drivers/net/tap/tap_flow.h
create mode 100644 drivers/net/tap/tap_netlink.c
create mode 100644 drivers/net/tap/tap_netlink.h
create mode 100644 drivers/net/tap/tap_tcmsgs.c
create mode 100644 drivers/net/tap/tap_tcmsgs.h

In the next patch, access to struct pmd_internals will be necessary in tap_flow.c to store the flows.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

---
 drivers/net/tap/Makefile      |  1 +
 drivers/net/tap/rte_eth_tap.c | 34 ++------------------
 drivers/net/tap/tap.h         | 73 +++++++++++++++++++++++++++++++++++++++++++

Quoted 10 lines · Pascal Mazon

In the next patch, access to struct pmd_internals will be necessary in tap_flow.c to store the flows.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

---
 drivers/net/tap/Makefile      |  1 +
 drivers/net/tap/rte_eth_tap.c | 34 ++------------------
 drivers/net/tap/tap.h         | 73

+++++++++++++++++++++++++++++++++++++++++++

tap.h is a generic name, I think rte_eth_tap.h fits better here.

<...>

Each kernel netdevice may have queueing disciplines set for it, which determine how to handle the packet (mostly on egress). That's part of the TC (Traffic Control) mechanism.

Through TC, it is possible to set filter rules that match specific packets, and act according to what is in the rule. This is a perfect candidate to implement the flow API for the tap PMD, as it has an associated kernel netdevice automatically.

Each flow API rule will be translated into its TC counterpart.

To leverage TC, it is necessary to communicate with the kernel using netlink. This patch introduces a library to help that communication.

Inside netlink.c, functions are generic for any netlink messaging. Inside tcmsgs.c, functions are specific to deal with TC rules.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

Quoted 3 lines · Pascal Mazon

Each kernel netdevice may have queueing disciplines set for it, which determine how to handle the packet (mostly on egress). That's part of the TC (Traffic Control) mechanism.

This is nice.
qdisc is egress part of the network stack right, is there any ingress part of it?

Quoted 7 lines

Through TC, it is possible to set filter rules that match specific packets, and act according to what is in the rule. This is a perfect candidate to implement the flow API for the tap PMD, as it has an associated kernel netdevice automatically.

Each flow API rule will be translated into its TC counterpart.

What can be use cases here?

Quoted 5 lines

To leverage TC, it is necessary to communicate with the kernel using netlink. This patch introduces a library to help that communication.

What do you think implementing these out of tap PMD? These can be used by KNI too.

Quoted 5 lines

Inside netlink.c, functions are generic for any netlink messaging. Inside tcmsgs.c, functions are specific to deal with TC rules.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

<...>

In the next patch, access to struct pmd_internals will be necessary in tap_flow.c to store the flows.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

Each kernel netdevice may have queueing disciplines set for it, which determine how to handle the packet (mostly on egress). That's part of the TC (Traffic Control) mechanism.

Through TC, it is possible to set filter rules that match specific packets, and act according to what is in the rule. This is a perfect candidate to implement the flow API for the tap PMD, as it has an associated kernel netdevice automatically.

Each flow API rule will be translated into its TC counterpart.

To leverage TC, it is necessary to communicate with the kernel using netlink. This patch introduces a library to help that communication.

Inside netlink.c, functions are generic for any netlink messaging. Inside tcmsgs.c, functions are specific to deal with TC rules.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

In the next patch, access to struct pmd_internals will be necessary in tap_flow.c to store the flows.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

---

drivers/net/tap/Makefile      |  1 +
drivers/net/tap/rte_eth_tap.c | 36 ++-------------------
drivers/net/tap/rte_eth_tap.h | 75 +++++++++++++++++++++++++++++++++++++++++++
3 files changed, 78 insertions(+), 34 deletions(-)

create mode 100644 drivers/net/tap/rte_eth_tap.h

drivers/net/tap/Makefile +0 -0

index ddf87232d335..fa4658bd1e75 100644
@@ -40,6 +40,7 @@ EXPORT_MAP := rte_pmd_tap_version.map

LIBABIVER := 1

CFLAGS += -O3
+CFLAGS += -I$(SRCDIR)
CFLAGS += $(WERROR_FLAGS)

#

drivers/net/tap/rte_eth_tap.c +2 -34

index f8d9cc7dc3b2..6bb63e5ec873 100644
@@ -55,6 +55,8 @@
 #include <linux/if_ether.h>
 #include <fcntl.h>
 
+#include <rte_eth_tap.h>
+
 /* Linux based path to the TUN device */
 #define TUN_TAP_DEV_PATH        "/dev/net/tun"
 #define DEFAULT_TAP_NAME        "dtap"

@@ -87,40 +89,6 @@ static struct rte_eth_link pmd_link = {
 	.link_autoneg = ETH_LINK_SPEED_AUTONEG
 };
 
-struct pkt_stats {
-	uint64_t opackets;		/* Number of output packets */
-	uint64_t ipackets;		/* Number of input packets */
-	uint64_t obytes;		/* Number of bytes on output */
-	uint64_t ibytes;		/* Number of bytes on input */
-	uint64_t errs;			/* Number of error packets */
-};
-
-struct rx_queue {
-	struct rte_mempool *mp;		/* Mempool for RX packets */
-	uint32_t trigger_seen;		/* Last seen Rx trigger value */
-	uint16_t in_port;		/* Port ID */
-	int fd;
-
-	struct pkt_stats stats;		/* Stats for this RX queue */
-};
-
-struct tx_queue {
-	int fd;
-	struct pkt_stats stats;		/* Stats for this TX queue */
-};
-
-struct pmd_internals {
-	char name[RTE_ETH_NAME_MAX_LEN];	/* Internal Tap device name */
-	uint16_t nb_queues;		/* Number of queues supported */
-	struct ether_addr eth_addr;	/* Mac address of the device port */
-
-	int if_index;			/* IF_INDEX for the port */
-	int ioctl_sock;			/* socket for ioctl calls */
-
-	struct rx_queue rxq[RTE_PMD_TAP_MAX_QUEUES];	/* List of RX queues */
-	struct tx_queue txq[RTE_PMD_TAP_MAX_QUEUES];	/* List of TX queues */
-};
-

static void
tap_trigger_cb(int sig __rte_unused)
{

drivers/net/tap/rte_eth_tap.h new file +30 -0

new file mode 100644
index 000000000000..aafdef1faa99
@@ -0,0 +1,75 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright 2017 6WIND S.A.
+ *   Copyright 2017 Mellanox.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ *     * Redistributions of source code must retain the above copyright
+ *       notice, this list of conditions and the following disclaimer.
+ *     * Redistributions in binary form must reproduce the above copyright
+ *       notice, this list of conditions and the following disclaimer in
+ *       the documentation and/or other materials provided with the
+ *       distribution.
+ *     * Neither the name of 6WIND S.A. nor the names of its
+ *       contributors may be used to endorse or promote products derived
+ *       from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE

drivers/net/tap/rte_eth_tap.h new file +35 -0

+
+#define RTE_PMD_TAP_MAX_QUEUES 16
+
+struct pkt_stats {
+	uint64_t opackets;              /* Number of output packets */
+	uint64_t ipackets;              /* Number of input packets */
+	uint64_t obytes;                /* Number of bytes on output */
+	uint64_t ibytes;                /* Number of bytes on input */
+	uint64_t errs;                  /* Number of TX error packets */
+};
+
+struct rx_queue {
+	struct rte_mempool *mp;         /* Mempool for RX packets */
+	uint32_t trigger_seen;          /* Last seen Rx trigger value */
+	uint16_t in_port;               /* Port ID */
+	int fd;
+	struct pkt_stats stats;         /* Stats for this RX queue */
+};
+
+struct tx_queue {
+	int fd;
+	struct pkt_stats stats;         /* Stats for this TX queue */
+};
+
+struct pmd_internals {
+	char name[RTE_ETH_NAME_MAX_LEN];  /* Internal Tap device name */
+	uint16_t nb_queues;               /* Number of queues supported */
+	struct ether_addr eth_addr;       /* Mac address of the device port */
+	int if_index;                     /* IF_INDEX for the port */
+	int ioctl_sock;                   /* socket for ioctl calls */
+	struct rx_queue rxq[RTE_PMD_TAP_MAX_QUEUES]; /* List of RX queues */
+	struct tx_queue txq[RTE_PMD_TAP_MAX_QUEUES]; /* List of TX queues */
+};
+
+#endif /* _RTE_ETH_TAP_H_ */

Signature · 1 line

2.8.0.rc0

Just noticed this new header does not have the C++ ifdefs. Create a new patch to fix this problem, unless you need to update this patch series. Just starting my review of this one, sorry was traveling last week.

Regards,
Keith

The flow API provides the ability to classify packets received by a tap netdevice.

This patch only implements skeleton functions for flow API support, no patterns are supported yet.

Signed-off-by Pascal Mazon <[email protected]>

Acked-by Olga Shern <[email protected]>

drivers/net/tap/tap_flow.h new file +8 -0

+
+int tap_dev_filter_ctrl(struct rte_eth_dev *dev,
+			enum rte_filter_type filter_type,
+			enum rte_filter_op filter_op,
+			void *arg);
+int tap_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error);
+
+#endif /* _TAP_FLOW_H_ */

Signature · 1 line

2.8.0.rc0

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 <[email protected]>

Acked-by Olga Shern <[email protected]>

drivers/net/tap/tap_flow.c +17 -0

+/**
+ * Set a unique handle in a flow.
+ *
+ * The kernel supports TC rules with equal priority, as long as they use the
+ * same matching fields (e.g.: dst mac and ipv4) with different values (and
+ * full mask to ensure no collision is possible).
+ * In those rules, the handle (uint32_t) is the part that would identify
+ * specifically each rule.
+ *
+ * On 32-bit architectures, the handle can simply be the flow's pointer address.
+ * On 64-bit architectures, we rely on jhash(flow) to find a (sufficiently)
+ * unique handle.
+ *
+ * @param[in, out] flow
+ *   The flow that needs its handle set.
+ */
+static void

drivers/net/tap/tap_flow.c +10 -0

+
+		handle = (uint32_t)h;
+	} else {
+		handle = rte_jhash(&flow, sizeof(flow), 1);
+	}
+#endif
+	/* must be at least 1 to avoid letting the kernel choose one for us */
+	if (!handle)
+		handle = 1;
+	flow->msg.t.tcm_handle = handle;
 }

This series add support for the flow API in tap PMD.

It enables filtering specific packets incoming on the tap netdevice, to process only desired ones. Under the hood, it uses kernel TC (traffic control), which takes place very early in the stack, and supports most common pattern items and actions defined in the flow API.

This series applies on top of:

[PATCH 0/6] net/tap: add additional management ops

v2 changes:
- support compilation on kernels < 4.2 (where flower support appeared)
- set whitespaces in tap.h
- remove unnecessary goto

v3 changes:
- vlan patterns enabled depending on running kernel (4.9+)
- update doc/guides/nics/tap.rst for Flow API support
- rebase on top of "net/tap: add additional management ops" series

v4 changes:
- rebase on top of "net/tap: add additional management ops" series
- fix a few netlink doxygen comments
- rename tap.h -> rte_eth_tap.h
- flush flow rules only when applicable

v5 changes:
- rebase after adrien's patches on Tx poll and Rx signaling
- better spaces for comments in rte_eth_tap.h

Pascal Mazon (4):
net/tap: move private elements to external header net/tap: add preliminary support for rte_flow net/tap: add netlink back-end for flow API net/tap: add basic flow API patterns and actions

doc/guides/nics/features/tap.ini |    1 +
doc/guides/nics/tap.rst          |   23 +
drivers/net/tap/Makefile         |   44 ++
drivers/net/tap/rte_eth_tap.c    |  101 ++--
drivers/net/tap/rte_eth_tap.h    |   80 +++
drivers/net/tap/tap_flow.c       | 1078 ++++++++++++++++++++++++++++++++++++++
drivers/net/tap/tap_flow.h       |   58 ++
drivers/net/tap/tap_netlink.c    |  367 +++++++++++++
drivers/net/tap/tap_netlink.h    |   69 +++
drivers/net/tap/tap_tcmsgs.c     |  378 +++++++++++++
drivers/net/tap/tap_tcmsgs.h     |   63 +++
11 files changed, 2227 insertions(+), 35 deletions(-)

create mode 100644 drivers/net/tap/rte_eth_tap.h
create mode 100644 drivers/net/tap/tap_flow.c
create mode 100644 drivers/net/tap/tap_flow.h
create mode 100644 drivers/net/tap/tap_netlink.c
create mode 100644 drivers/net/tap/tap_netlink.h
create mode 100644 drivers/net/tap/tap_tcmsgs.c
create mode 100644 drivers/net/tap/tap_tcmsgs.h

Signature · 1 line

2.8.0.rc0

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 <[email protected]>

Acked-by Olga Shern <[email protected]>

doc/guides/nics/tap.rst +15 -0

index c4f207be3b47..cdb528b5eae4 100644
@@ -82,6 +82,29 @@ can utilize that stack to handle the network protocols. Plus you would be able
 to address the interface using an IP address assigned to the internal
 interface.
 
+Flow API support
+----------------
+
+The tap PMD supports major flow API pattern items and actions, when running on
+linux kernels above 4.2 ("Flower" classifier required). Supported items:
+
+- eth: src and dst (with variable masks), and eth_type (0xffff mask).
+- vlan: vid, pcp, tpid, but not eid. (requires kernel 4.9)
+- 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

This series add support for the flow API in tap PMD.

It enables filtering specific packets incoming on the tap netdevice, to process only desired ones. Under the hood, it uses kernel TC (traffic control), which takes place very early in the stack, and supports most common pattern items and actions defined in the flow API.

v6 changes:
- fix compilation issue on i686 (wrong cast for rte flow handle)

v5 changes:
- rebase after adrien's patches on Tx poll and Rx signaling
- better spaces for comments in rte_eth_tap.h

v4 changes:
- rebase on top of "net/tap: add additional management ops" series
- fix a few netlink doxygen comments
- rename tap.h -> rte_eth_tap.h
- flush flow rules only when applicable

v3 changes:
- vlan patterns enabled depending on running kernel (4.9+)
- update doc/guides/nics/tap.rst for Flow API support
- rebase on top of "net/tap: add additional management ops" series

v2 changes:
- support compilation on kernels < 4.2 (where flower support appeared)
- set whitespaces in tap.h
- remove unnecessary goto

Pascal Mazon (4):
net/tap: move private elements to external header net/tap: add preliminary support for rte_flow net/tap: add netlink back-end for flow API net/tap: add basic flow API patterns and actions

doc/guides/nics/features/tap.ini |    1 +
doc/guides/nics/tap.rst          |   23 +
drivers/net/tap/Makefile         |   44 ++
drivers/net/tap/rte_eth_tap.c    |  101 ++--
drivers/net/tap/rte_eth_tap.h    |   80 +++
drivers/net/tap/tap_flow.c       | 1070 ++++++++++++++++++++++++++++++++++++++
drivers/net/tap/tap_flow.h       |   58 +++
drivers/net/tap/tap_netlink.c    |  367 +++++++++++++
drivers/net/tap/tap_netlink.h    |   69 +++
drivers/net/tap/tap_tcmsgs.c     |  378 ++++++++++++++
drivers/net/tap/tap_tcmsgs.h     |   63 +++
11 files changed, 2219 insertions(+), 35 deletions(-)