[dpdk-dev] [PATCH 1/2] net/tap: UDP/TCP port mask not supported in flow

[dpdk-dev] [PATCH 1/2] net/tap: UDP/TCP port mask not supported in flow

[Pascal Mazon]

Only full mask (0xffff) is accepted, there is no way to specify a mask
for layer 4 ports to the kernel using TC rules.

Fixes: 1c71189ab9b7 ("net/tap: add basic flow API patterns and actions")

Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com>
---
 drivers/net/tap/tap_flow.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/drivers/net/tap/tap_flow.c b/drivers/net/tap/tap_flow.c
index 514e3fae5c38..cf1c8a26c8ff 100644
--- a/drivers/net/tap/tap_flow.c
+++ b/drivers/net/tap/tap_flow.c
@@ -614,18 +614,20 @@ tap_flow_create_udp(const struct rte_flow_item *item, void *data)
 	/* check that previous ip_proto is compatible with udp */
 	if (info->ip_proto && info->ip_proto != IPPROTO_UDP)
 		return -1;
+	/* TC does not support UDP port masking. Only accept if exact match. */
+	if ((mask->hdr.src_port && mask->hdr.src_port != 0xffff) ||
+	    (mask->hdr.dst_port && mask->hdr.dst_port != 0xffff))
+		return -1;
 	if (!flow)
 		return 0;
 	msg = &flow->msg;
 	nlattr_add8(&msg->nh, TCA_FLOWER_KEY_IP_PROTO, IPPROTO_UDP);
 	if (!spec)
 		return 0;
-	if (spec->hdr.dst_port &&
-	    (spec->hdr.dst_port & mask->hdr.dst_port) == spec->hdr.dst_port)
+	if (spec->hdr.dst_port & mask->hdr.dst_port)
 		nlattr_add16(&msg->nh, TCA_FLOWER_KEY_UDP_DST,
 			     spec->hdr.dst_port);
-	if (spec->hdr.src_port &&
-	    (spec->hdr.src_port & mask->hdr.src_port) == spec->hdr.src_port)
+	if (spec->hdr.src_port & mask->hdr.src_port)
 		nlattr_add16(&msg->nh, TCA_FLOWER_KEY_UDP_SRC,
 			     spec->hdr.src_port);
 	return 0;
@@ -658,18 +660,20 @@ tap_flow_create_tcp(const struct rte_flow_item *item, void *data)
 	/* check that previous ip_proto is compatible with tcp */
 	if (info->ip_proto && info->ip_proto != IPPROTO_TCP)
 		return -1;
+	/* TC does not support TCP port masking. Only accept if exact match. */
+	if ((mask->hdr.src_port && mask->hdr.src_port != 0xffff) ||
+	    (mask->hdr.dst_port && mask->hdr.dst_port != 0xffff))
+		return -1;
 	if (!flow)
 		return 0;
 	msg = &flow->msg;
 	nlattr_add8(&msg->nh, TCA_FLOWER_KEY_IP_PROTO, IPPROTO_TCP);
 	if (!spec)
 		return 0;
-	if (spec->hdr.dst_port &&
-	    (spec->hdr.dst_port & mask->hdr.dst_port) == spec->hdr.dst_port)
+	if (spec->hdr.dst_port & mask->hdr.dst_port)
 		nlattr_add16(&msg->nh, TCA_FLOWER_KEY_TCP_DST,
 			     spec->hdr.dst_port);
-	if (spec->hdr.src_port &&
-	    (spec->hdr.src_port & mask->hdr.src_port) == spec->hdr.src_port)
+	if (spec->hdr.src_port & mask->hdr.src_port)
 		nlattr_add16(&msg->nh, TCA_FLOWER_KEY_TCP_SRC,
 			     spec->hdr.src_port);
 	return 0;
-- 
2.12.0.306.g4a9b9b3

[dpdk-dev] [PATCH 2/2] net/tap: remove minimum packet size in Rx

[Pascal Mazon]

With support for segmented packets, it is now possible to easily receive
packets of many sizes, given an adequate number of descriptors.

Remove limitation on the minimum size of mbuf: on reception, if a packet
won't fit in the queue's mbufs, it will be detected in the packet info
and the packet will be discarded.

Fixes: 4a6bb33dc67c ("net/tap: support segmented mbufs")

Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com>
---
 drivers/net/tap/rte_eth_tap.c | 13 -------------
 1 file changed, 13 deletions(-)

diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index 069200199573..eef6e6cc2828 100644
--- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c
@@ -875,7 +875,6 @@ tap_rx_queue_setup(struct rte_eth_dev *dev,
 	struct rte_mbuf **tmp = &rxq->pool;
 	struct iovec (*iovecs)[nb_rx_desc + 1];
 	int data_off = RTE_PKTMBUF_HEADROOM;
-	uint16_t buf_size;
 	int ret = 0;
 	int fd;
 	int i;
@@ -901,18 +900,6 @@ tap_rx_queue_setup(struct rte_eth_dev *dev,
 	}
 	rxq->iovecs = iovecs;
 
-	/* Now get the space available for data in the mbuf */
-	buf_size = (uint16_t)(rte_pktmbuf_data_room_size(mp) -
-				RTE_PKTMBUF_HEADROOM);
-
-	if (buf_size < ETH_FRAME_LEN) {
-		RTE_LOG(WARNING, PMD,
-			"%s: %d bytes will not fit in mbuf (%d bytes)\n",
-			dev->data->name, ETH_FRAME_LEN, buf_size);
-		ret = -ENOMEM;
-		goto error;
-	}
-
 	fd = rx_setup_queue(dev, internals, rx_queue_id);
 	if (fd == -1) {
 		ret = fd;
-- 
2.12.0.306.g4a9b9b3

Re: [dpdk-dev] [PATCH 2/2] net/tap: remove minimum packet size in Rx

[Wiles, Keith]

> On Mar 30, 2017, at 3:52 AM, Pascal Mazon <pascal.mazon@6wind.com> wrote:
> 
> With support for segmented packets, it is now possible to easily receive
> packets of many sizes, given an adequate number of descriptors.
> 
> Remove limitation on the minimum size of mbuf: on reception, if a packet
> won't fit in the queue's mbufs, it will be detected in the packet info
> and the packet will be discarded.
> 
> Fixes: 4a6bb33dc67c ("net/tap: support segmented mbufs")
> 
> Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com>

Acked-by: Keith Wiles <keith.wiles@intel.com>

Re: [dpdk-dev] [PATCH 1/2] net/tap: UDP/TCP port mask not supported in flow

[Wiles, Keith]

> On Mar 30, 2017, at 3:52 AM, Pascal Mazon <pascal.mazon@6wind.com> wrote:
> 
> Only full mask (0xffff) is accepted, there is no way to specify a mask
> for layer 4 ports to the kernel using TC rules.
> 
> Fixes: 1c71189ab9b7 ("net/tap: add basic flow API patterns and actions")
> 
> Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com>

Acked-by: Keith Wiles <keith.wiles@intel.com>

Regards,
Keith

Re: [dpdk-dev] [PATCH 1/2] net/tap: UDP/TCP port mask not supported in flow

[Ferruh Yigit]

On 3/30/2017 2:08 PM, Wiles, Keith wrote:
> 
>> On Mar 30, 2017, at 3:52 AM, Pascal Mazon <pascal.mazon@6wind.com> wrote:
>>
>> Only full mask (0xffff) is accepted, there is no way to specify a mask
>> for layer 4 ports to the kernel using TC rules.
>>
>> Fixes: 1c71189ab9b7 ("net/tap: add basic flow API patterns and actions")
>>
>> Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com>
> 
> Acked-by: Keith Wiles <keith.wiles@intel.com>

Series applied to dpdk-next-net/master, thanks.

(This patch fixes commits in next-net, when integrated to main tree
"Fixes:" will be with wrong commit id, please confirm the commit log of
integrated commit.)