* [PATCH 0/2] lib: move Netlink wrapper to lib
@ 2018-03-13 12:28 Nelio Laranjeiro
2018-03-13 12:28 ` [PATCH 1/2] lib: move Netlink code into a common library Nelio Laranjeiro
` (2 more replies)
0 siblings, 3 replies; 11+ messages in thread
From: Nelio Laranjeiro @ 2018-03-13 12:28 UTC (permalink / raw)
To: Pascal Mazon; +Cc: dev
TAP PMD uses a wrapper to communicate with Netlink through a socket. This
wrapper can be useful for other PMD which may need to communicate also with
the Linux kernel via Netlink.
This series moves this wrapper to the lib level to let any PMD implement use
it when necessary.
Nelio Laranjeiro (2):
lib: move Netlink code into a common library
lib: add request Netlink messages
MAINTAINERS | 3 +
config/common_base | 6 ++
config/common_linuxapp | 1 +
drivers/net/tap/Makefile | 3 +-
drivers/net/tap/rte_eth_tap.c | 16 +--
drivers/net/tap/tap_flow.c | 118 ++++++++++-----------
drivers/net/tap/tap_netlink.h | 42 --------
drivers/net/tap/tap_tcmsgs.c | 28 ++---
drivers/net/tap/tap_tcmsgs.h | 2 +-
lib/Makefile | 2 +
lib/librte_netlink/Makefile | 26 +++++
lib/librte_netlink/meson.build | 13 +++
.../librte_netlink/rte_netlink.c | 87 +++++++++++----
lib/librte_netlink/rte_netlink.h | 43 ++++++++
lib/librte_netlink/rte_netlink_version.map | 18 ++++
lib/meson.build | 2 +-
mk/rte.app.mk | 1 +
17 files changed, 263 insertions(+), 148 deletions(-)
delete mode 100644 drivers/net/tap/tap_netlink.h
create mode 100644 lib/librte_netlink/Makefile
create mode 100644 lib/librte_netlink/meson.build
rename drivers/net/tap/tap_netlink.c => lib/librte_netlink/rte_netlink.c (77%)
create mode 100644 lib/librte_netlink/rte_netlink.h
create mode 100644 lib/librte_netlink/rte_netlink_version.map
--
2.11.0
^ permalink raw reply [flat|nested] 11+ messages in thread
* [PATCH 1/2] lib: move Netlink code into a common library
2018-03-13 12:28 [PATCH 0/2] lib: move Netlink wrapper to lib Nelio Laranjeiro
@ 2018-03-13 12:28 ` Nelio Laranjeiro
2018-03-13 13:58 ` Bruce Richardson
2018-03-13 12:28 ` [PATCH 2/2] lib: add request Netlink messages Nelio Laranjeiro
2018-03-13 21:20 ` [PATCH 0/2] lib: move Netlink wrapper to lib Stephen Hemminger
2 siblings, 1 reply; 11+ messages in thread
From: Nelio Laranjeiro @ 2018-03-13 12:28 UTC (permalink / raw)
To: Pascal Mazon; +Cc: dev
Moves TAP PMD generic Netlink library into lib directory to let other PMD
use Netlink to communicate with kernel.
As this library uses a socket to communicate with Netlink it does not bring
any dependency on the libnl.
Cc: Pascal Mazon <pascal.mazon@6wind.com>
Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
MAINTAINERS | 3 +
config/common_base | 6 ++
config/common_linuxapp | 1 +
drivers/net/tap/Makefile | 3 +-
drivers/net/tap/rte_eth_tap.c | 16 +--
drivers/net/tap/tap_flow.c | 118 ++++++++++-----------
drivers/net/tap/tap_netlink.h | 42 --------
drivers/net/tap/tap_tcmsgs.c | 28 ++---
drivers/net/tap/tap_tcmsgs.h | 2 +-
lib/Makefile | 2 +
lib/librte_netlink/Makefile | 26 +++++
lib/librte_netlink/meson.build | 13 +++
.../librte_netlink/rte_netlink.c | 42 ++++----
lib/librte_netlink/rte_netlink.h | 42 ++++++++
lib/librte_netlink/rte_netlink_version.map | 17 +++
lib/meson.build | 2 +-
mk/rte.app.mk | 1 +
17 files changed, 216 insertions(+), 148 deletions(-)
delete mode 100644 drivers/net/tap/tap_netlink.h
create mode 100644 lib/librte_netlink/Makefile
create mode 100644 lib/librte_netlink/meson.build
rename drivers/net/tap/tap_netlink.c => lib/librte_netlink/rte_netlink.c (88%)
create mode 100644 lib/librte_netlink/rte_netlink.h
create mode 100644 lib/librte_netlink/rte_netlink_version.map
diff --git a/MAINTAINERS b/MAINTAINERS
index a646ca3e1..f836f4c5d 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -325,6 +325,9 @@ F: drivers/raw/skeleton_rawdev/
F: test/test/test_rawdev.c
F: doc/guides/prog_guide/rawdev.rst
+Netlink Library
+M: Pascal Mazon <pascal.mazon@6wind.com>
+F: lib/librte_netlink/
Bus Drivers
-----------
diff --git a/config/common_base b/config/common_base
index ad03cf433..6ee715b21 100644
--- a/config/common_base
+++ b/config/common_base
@@ -823,3 +823,9 @@ CONFIG_RTE_APP_CRYPTO_PERF=y
# Compile the eventdev application
#
CONFIG_RTE_APP_EVENTDEV=y
+
+#
+# Compile netlink library
+# It is enabled by default for Linux only.
+#
+CONFIG_RTE_LIBRTE_NETLINK=n
diff --git a/config/common_linuxapp b/config/common_linuxapp
index ff98f2355..97e9d6dfe 100644
--- a/config/common_linuxapp
+++ b/config/common_linuxapp
@@ -23,3 +23,4 @@ CONFIG_RTE_LIBRTE_NFP_PMD=y
CONFIG_RTE_LIBRTE_POWER=y
CONFIG_RTE_VIRTIO_USER=y
CONFIG_RTE_PROC_INFO=y
+CONFIG_RTE_LIBRTE_NETLINK=y
diff --git a/drivers/net/tap/Makefile b/drivers/net/tap/Makefile
index ccc5c5fc0..05fd392ef 100644
--- a/drivers/net/tap/Makefile
+++ b/drivers/net/tap/Makefile
@@ -24,7 +24,7 @@ CFLAGS += -I.
CFLAGS += $(WERROR_FLAGS)
LDLIBS += -lrte_eal -lrte_mbuf -lrte_mempool -lrte_ring
LDLIBS += -lrte_ethdev -lrte_net -lrte_kvargs -lrte_hash
-LDLIBS += -lrte_bus_vdev
+LDLIBS += -lrte_bus_vdev -lrte_netlink
CFLAGS += -DTAP_MAX_QUEUES=$(TAP_MAX_QUEUES)
@@ -33,7 +33,6 @@ CFLAGS += -DTAP_MAX_QUEUES=$(TAP_MAX_QUEUES)
#
SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += rte_eth_tap.c
SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += tap_flow.c
-SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += tap_netlink.c
SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += tap_tcmsgs.c
SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += tap_bpf_api.c
SRCS-$(CONFIG_RTE_LIBRTE_PMD_TAP) += tap_intr.c
diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index f09db0ea9..05a6beabb 100644
--- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c
@@ -15,6 +15,7 @@
#include <rte_net.h>
#include <rte_debug.h>
#include <rte_ip.h>
+#include <rte_netlink.h>
#include <sys/types.h>
#include <sys/stat.h>
@@ -36,7 +37,6 @@
#include <rte_eth_tap.h>
#include <tap_flow.h>
-#include <tap_netlink.h>
#include <tap_tcmsgs.h>
/* Linux based path to the TUN device */
@@ -1197,7 +1197,7 @@ tap_set_mc_addr_list(struct rte_eth_dev *dev __rte_unused,
}
static int
-tap_nl_msg_handler(struct nlmsghdr *nh, void *arg)
+rte_nl_msg_handler(struct nlmsghdr *nh, void *arg)
{
struct rte_eth_dev *dev = arg;
struct pmd_internals *pmd = dev->data->dev_private;
@@ -1216,7 +1216,7 @@ tap_dev_intr_handler(void *cb_arg)
struct rte_eth_dev *dev = cb_arg;
struct pmd_internals *pmd = dev->data->dev_private;
- tap_nl_recv(pmd->intr_handle.fd, tap_nl_msg_handler, dev);
+ rte_nl_recv(pmd->intr_handle.fd, rte_nl_msg_handler, dev);
}
static int
@@ -1227,20 +1227,20 @@ tap_lsc_intr_handle_set(struct rte_eth_dev *dev, int set)
/* In any case, disable interrupt if the conf is no longer there. */
if (!dev->data->dev_conf.intr_conf.lsc) {
if (pmd->intr_handle.fd != -1) {
- tap_nl_final(pmd->intr_handle.fd);
+ rte_nl_final(pmd->intr_handle.fd);
rte_intr_callback_unregister(&pmd->intr_handle,
tap_dev_intr_handler, dev);
}
return 0;
}
if (set) {
- pmd->intr_handle.fd = tap_nl_init(RTMGRP_LINK);
+ pmd->intr_handle.fd = rte_nl_init(RTMGRP_LINK);
if (unlikely(pmd->intr_handle.fd == -1))
return -EBADF;
return rte_intr_callback_register(
&pmd->intr_handle, tap_dev_intr_handler, dev);
}
- tap_nl_final(pmd->intr_handle.fd);
+ rte_nl_final(pmd->intr_handle.fd);
return rte_intr_callback_unregister(&pmd->intr_handle,
tap_dev_intr_handler, dev);
}
@@ -1436,7 +1436,7 @@ eth_dev_tap_create(struct rte_vdev_device *vdev, char *tap_name,
* - rte_flow actual/implicit lists
* - implicit rules
*/
- pmd->nlsk_fd = tap_nl_init(0);
+ pmd->nlsk_fd = rte_nl_init(0);
if (pmd->nlsk_fd == -1) {
RTE_LOG(WARNING, PMD, "%s: failed to create netlink socket.\n",
pmd->name);
@@ -1673,7 +1673,7 @@ rte_pmd_tap_remove(struct rte_vdev_device *dev)
if (internals->nlsk_fd) {
tap_flow_flush(eth_dev, NULL);
tap_flow_implicit_flush(internals, NULL);
- tap_nl_final(internals->nlsk_fd);
+ rte_nl_final(internals->nlsk_fd);
}
for (i = 0; i < RTE_PMD_TAP_MAX_QUEUES; i++) {
if (internals->rxq[i].fd != -1) {
diff --git a/drivers/net/tap/tap_flow.c b/drivers/net/tap/tap_flow.c
index 551b2d83d..90a105d47 100644
--- a/drivers/net/tap/tap_flow.c
+++ b/drivers/net/tap/tap_flow.c
@@ -538,16 +538,16 @@ tap_flow_create_eth(const struct rte_flow_item *item, void *data)
return 0;
msg = &flow->msg;
if (!is_zero_ether_addr(&spec->dst)) {
- tap_nlattr_add(&msg->nh, TCA_FLOWER_KEY_ETH_DST, ETHER_ADDR_LEN,
+ rte_nlattr_add(&msg->nh, TCA_FLOWER_KEY_ETH_DST, ETHER_ADDR_LEN,
&spec->dst.addr_bytes);
- tap_nlattr_add(&msg->nh,
+ rte_nlattr_add(&msg->nh,
TCA_FLOWER_KEY_ETH_DST_MASK, ETHER_ADDR_LEN,
&mask->dst.addr_bytes);
}
if (!is_zero_ether_addr(&mask->src)) {
- tap_nlattr_add(&msg->nh, TCA_FLOWER_KEY_ETH_SRC, ETHER_ADDR_LEN,
+ rte_nlattr_add(&msg->nh, TCA_FLOWER_KEY_ETH_SRC, ETHER_ADDR_LEN,
&spec->src.addr_bytes);
- tap_nlattr_add(&msg->nh,
+ rte_nlattr_add(&msg->nh,
TCA_FLOWER_KEY_ETH_SRC_MASK, ETHER_ADDR_LEN,
&mask->src.addr_bytes);
}
@@ -599,10 +599,10 @@ tap_flow_create_vlan(const struct rte_flow_item *item, void *data)
uint8_t vid = VLAN_ID(tci);
if (prio)
- tap_nlattr_add8(&msg->nh,
+ rte_nlattr_add8(&msg->nh,
TCA_FLOWER_KEY_VLAN_PRIO, prio);
if (vid)
- tap_nlattr_add16(&msg->nh,
+ rte_nlattr_add16(&msg->nh,
TCA_FLOWER_KEY_VLAN_ID, vid);
}
return 0;
@@ -646,19 +646,19 @@ tap_flow_create_ipv4(const struct rte_flow_item *item, void *data)
if (!spec)
return 0;
if (spec->hdr.dst_addr) {
- tap_nlattr_add32(&msg->nh, TCA_FLOWER_KEY_IPV4_DST,
+ rte_nlattr_add32(&msg->nh, TCA_FLOWER_KEY_IPV4_DST,
spec->hdr.dst_addr);
- tap_nlattr_add32(&msg->nh, TCA_FLOWER_KEY_IPV4_DST_MASK,
+ rte_nlattr_add32(&msg->nh, TCA_FLOWER_KEY_IPV4_DST_MASK,
mask->hdr.dst_addr);
}
if (spec->hdr.src_addr) {
- tap_nlattr_add32(&msg->nh, TCA_FLOWER_KEY_IPV4_SRC,
+ rte_nlattr_add32(&msg->nh, TCA_FLOWER_KEY_IPV4_SRC,
spec->hdr.src_addr);
- tap_nlattr_add32(&msg->nh, TCA_FLOWER_KEY_IPV4_SRC_MASK,
+ rte_nlattr_add32(&msg->nh, TCA_FLOWER_KEY_IPV4_SRC_MASK,
mask->hdr.src_addr);
}
if (spec->hdr.next_proto_id)
- tap_nlattr_add8(&msg->nh, TCA_FLOWER_KEY_IP_PROTO,
+ rte_nlattr_add8(&msg->nh, TCA_FLOWER_KEY_IP_PROTO,
spec->hdr.next_proto_id);
return 0;
}
@@ -702,19 +702,19 @@ tap_flow_create_ipv6(const struct rte_flow_item *item, void *data)
if (!spec)
return 0;
if (memcmp(spec->hdr.dst_addr, empty_addr, 16)) {
- tap_nlattr_add(&msg->nh, TCA_FLOWER_KEY_IPV6_DST,
+ rte_nlattr_add(&msg->nh, TCA_FLOWER_KEY_IPV6_DST,
sizeof(spec->hdr.dst_addr), &spec->hdr.dst_addr);
- tap_nlattr_add(&msg->nh, TCA_FLOWER_KEY_IPV6_DST_MASK,
+ rte_nlattr_add(&msg->nh, TCA_FLOWER_KEY_IPV6_DST_MASK,
sizeof(mask->hdr.dst_addr), &mask->hdr.dst_addr);
}
if (memcmp(spec->hdr.src_addr, empty_addr, 16)) {
- tap_nlattr_add(&msg->nh, TCA_FLOWER_KEY_IPV6_SRC,
+ rte_nlattr_add(&msg->nh, TCA_FLOWER_KEY_IPV6_SRC,
sizeof(spec->hdr.src_addr), &spec->hdr.src_addr);
- tap_nlattr_add(&msg->nh, TCA_FLOWER_KEY_IPV6_SRC_MASK,
+ rte_nlattr_add(&msg->nh, TCA_FLOWER_KEY_IPV6_SRC_MASK,
sizeof(mask->hdr.src_addr), &mask->hdr.src_addr);
}
if (spec->hdr.proto)
- tap_nlattr_add8(&msg->nh,
+ rte_nlattr_add8(&msg->nh,
TCA_FLOWER_KEY_IP_PROTO, spec->hdr.proto);
return 0;
}
@@ -753,14 +753,14 @@ tap_flow_create_udp(const struct rte_flow_item *item, void *data)
if (!flow)
return 0;
msg = &flow->msg;
- tap_nlattr_add8(&msg->nh, TCA_FLOWER_KEY_IP_PROTO, IPPROTO_UDP);
+ rte_nlattr_add8(&msg->nh, TCA_FLOWER_KEY_IP_PROTO, IPPROTO_UDP);
if (!spec)
return 0;
if (spec->hdr.dst_port & mask->hdr.dst_port)
- tap_nlattr_add16(&msg->nh, TCA_FLOWER_KEY_UDP_DST,
+ rte_nlattr_add16(&msg->nh, TCA_FLOWER_KEY_UDP_DST,
spec->hdr.dst_port);
if (spec->hdr.src_port & mask->hdr.src_port)
- tap_nlattr_add16(&msg->nh, TCA_FLOWER_KEY_UDP_SRC,
+ rte_nlattr_add16(&msg->nh, TCA_FLOWER_KEY_UDP_SRC,
spec->hdr.src_port);
return 0;
}
@@ -799,14 +799,14 @@ tap_flow_create_tcp(const struct rte_flow_item *item, void *data)
if (!flow)
return 0;
msg = &flow->msg;
- tap_nlattr_add8(&msg->nh, TCA_FLOWER_KEY_IP_PROTO, IPPROTO_TCP);
+ rte_nlattr_add8(&msg->nh, TCA_FLOWER_KEY_IP_PROTO, IPPROTO_TCP);
if (!spec)
return 0;
if (spec->hdr.dst_port & mask->hdr.dst_port)
- tap_nlattr_add16(&msg->nh, TCA_FLOWER_KEY_TCP_DST,
+ rte_nlattr_add16(&msg->nh, TCA_FLOWER_KEY_TCP_DST,
spec->hdr.dst_port);
if (spec->hdr.src_port & mask->hdr.src_port)
- tap_nlattr_add16(&msg->nh, TCA_FLOWER_KEY_TCP_SRC,
+ rte_nlattr_add16(&msg->nh, TCA_FLOWER_KEY_TCP_SRC,
spec->hdr.src_port);
return 0;
}
@@ -912,43 +912,43 @@ add_action(struct rte_flow *flow, size_t *act_index, struct action_data *adata)
{
struct nlmsg *msg = &flow->msg;
- if (tap_nlattr_nested_start(msg, (*act_index)++) < 0)
+ if (rte_nlattr_nested_start(msg, (*act_index)++) < 0)
return -1;
- tap_nlattr_add(&msg->nh, TCA_ACT_KIND,
+ rte_nlattr_add(&msg->nh, TCA_ACT_KIND,
strlen(adata->id) + 1, adata->id);
- if (tap_nlattr_nested_start(msg, TCA_ACT_OPTIONS) < 0)
+ if (rte_nlattr_nested_start(msg, TCA_ACT_OPTIONS) < 0)
return -1;
if (strcmp("gact", adata->id) == 0) {
- tap_nlattr_add(&msg->nh, TCA_GACT_PARMS, sizeof(adata->gact),
+ rte_nlattr_add(&msg->nh, TCA_GACT_PARMS, sizeof(adata->gact),
&adata->gact);
} else if (strcmp("mirred", adata->id) == 0) {
if (adata->mirred.eaction == TCA_EGRESS_MIRROR)
adata->mirred.action = TC_ACT_PIPE;
else /* REDIRECT */
adata->mirred.action = TC_ACT_STOLEN;
- tap_nlattr_add(&msg->nh, TCA_MIRRED_PARMS,
+ rte_nlattr_add(&msg->nh, TCA_MIRRED_PARMS,
sizeof(adata->mirred),
&adata->mirred);
} else if (strcmp("skbedit", adata->id) == 0) {
- tap_nlattr_add(&msg->nh, TCA_SKBEDIT_PARMS,
+ rte_nlattr_add(&msg->nh, TCA_SKBEDIT_PARMS,
sizeof(adata->skbedit.skbedit),
&adata->skbedit.skbedit);
- tap_nlattr_add16(&msg->nh, TCA_SKBEDIT_QUEUE_MAPPING,
+ rte_nlattr_add16(&msg->nh, TCA_SKBEDIT_QUEUE_MAPPING,
adata->skbedit.queue);
} else if (strcmp("bpf", adata->id) == 0) {
- tap_nlattr_add32(&msg->nh, TCA_ACT_BPF_FD, adata->bpf.bpf_fd);
- tap_nlattr_add(&msg->nh, TCA_ACT_BPF_NAME,
+ rte_nlattr_add32(&msg->nh, TCA_ACT_BPF_FD, adata->bpf.bpf_fd);
+ rte_nlattr_add(&msg->nh, TCA_ACT_BPF_NAME,
strlen(adata->bpf.annotation) + 1,
adata->bpf.annotation);
- tap_nlattr_add(&msg->nh, TCA_ACT_BPF_PARMS,
+ rte_nlattr_add(&msg->nh, TCA_ACT_BPF_PARMS,
sizeof(adata->bpf.bpf),
&adata->bpf.bpf);
} else {
return -1;
}
- tap_nlattr_nested_finish(msg); /* nested TCA_ACT_OPTIONS */
- tap_nlattr_nested_finish(msg); /* nested act_index */
+ rte_nlattr_nested_finish(msg); /* nested TCA_ACT_OPTIONS */
+ rte_nlattr_nested_finish(msg); /* nested act_index */
return 0;
}
@@ -978,12 +978,12 @@ add_actions(struct rte_flow *flow, int nb_actions, struct action_data *data,
size_t act_index = 1;
int i;
- if (tap_nlattr_nested_start(msg, classifier_action) < 0)
+ if (rte_nlattr_nested_start(msg, classifier_action) < 0)
return -1;
for (i = 0; i < nb_actions; i++)
if (add_action(flow, &act_index, data + i) < 0)
return -1;
- tap_nlattr_nested_finish(msg); /* nested TCA_FLOWER_ACT */
+ rte_nlattr_nested_finish(msg); /* nested TCA_FLOWER_ACT */
return 0;
}
@@ -1067,8 +1067,8 @@ priv_flow_process(struct pmd_internals *pmd,
TC_H_MAKE(MULTIQ_MAJOR_HANDLE, 0);
}
/* use flower filter type */
- tap_nlattr_add(&flow->msg.nh, TCA_KIND, sizeof("flower"), "flower");
- if (tap_nlattr_nested_start(&flow->msg, TCA_OPTIONS) < 0)
+ rte_nlattr_add(&flow->msg.nh, TCA_KIND, sizeof("flower"), "flower");
+ if (rte_nlattr_nested_start(&flow->msg, TCA_OPTIONS) < 0)
goto exit_item_not_supported;
}
for (; items->type != RTE_FLOW_ITEM_TYPE_END; ++items) {
@@ -1104,14 +1104,14 @@ priv_flow_process(struct pmd_internals *pmd,
}
if (flow) {
if (data.vlan) {
- tap_nlattr_add16(&flow->msg.nh, TCA_FLOWER_KEY_ETH_TYPE,
+ rte_nlattr_add16(&flow->msg.nh, TCA_FLOWER_KEY_ETH_TYPE,
htons(ETH_P_8021Q));
- tap_nlattr_add16(&flow->msg.nh,
+ rte_nlattr_add16(&flow->msg.nh,
TCA_FLOWER_KEY_VLAN_ETH_TYPE,
data.eth_type ?
data.eth_type : htons(ETH_P_ALL));
} else if (data.eth_type) {
- tap_nlattr_add16(&flow->msg.nh, TCA_FLOWER_KEY_ETH_TYPE,
+ rte_nlattr_add16(&flow->msg.nh, TCA_FLOWER_KEY_ETH_TYPE,
data.eth_type);
}
}
@@ -1224,7 +1224,7 @@ priv_flow_process(struct pmd_internals *pmd,
}
end:
if (flow)
- tap_nlattr_nested_finish(&flow->msg); /* nested TCA_OPTIONS */
+ rte_nlattr_nested_finish(&flow->msg); /* nested TCA_OPTIONS */
return 0;
exit_item_not_supported:
rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM,
@@ -1368,13 +1368,13 @@ tap_flow_create(struct rte_eth_dev *dev,
tap_flow_set_handle(flow);
if (priv_flow_process(pmd, attr, items, actions, error, flow, 0))
goto fail;
- err = tap_nl_send(pmd->nlsk_fd, &msg->nh);
+ err = rte_nl_send(pmd->nlsk_fd, &msg->nh);
if (err < 0) {
rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_HANDLE,
NULL, "couldn't send request to kernel");
goto fail;
}
- err = tap_nl_recv_ack(pmd->nlsk_fd);
+ err = rte_nl_recv_ack(pmd->nlsk_fd);
if (err < 0) {
RTE_LOG(ERR, PMD,
"Kernel refused TC filter rule creation (%d): %s\n",
@@ -1412,14 +1412,14 @@ tap_flow_create(struct rte_eth_dev *dev,
NULL, "rte flow rule validation failed");
goto fail;
}
- err = tap_nl_send(pmd->nlsk_fd, &msg->nh);
+ err = rte_nl_send(pmd->nlsk_fd, &msg->nh);
if (err < 0) {
rte_flow_error_set(
error, ENOMEM, RTE_FLOW_ERROR_TYPE_HANDLE,
NULL, "Failure sending nl request");
goto fail;
}
- err = tap_nl_recv_ack(pmd->nlsk_fd);
+ err = rte_nl_recv_ack(pmd->nlsk_fd);
if (err < 0) {
RTE_LOG(ERR, PMD,
"Kernel refused TC filter rule creation (%d): %s\n",
@@ -1465,13 +1465,13 @@ tap_flow_destroy_pmd(struct pmd_internals *pmd,
flow->msg.nh.nlmsg_flags = NLM_F_REQUEST | NLM_F_ACK;
flow->msg.nh.nlmsg_type = RTM_DELTFILTER;
- ret = tap_nl_send(pmd->nlsk_fd, &flow->msg.nh);
+ ret = rte_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 = tap_nl_recv_ack(pmd->nlsk_fd);
+ ret = rte_nl_recv_ack(pmd->nlsk_fd);
/* If errno is ENOENT, the rule is already no longer in the kernel. */
if (ret < 0 && errno == ENOENT)
ret = 0;
@@ -1489,14 +1489,14 @@ tap_flow_destroy_pmd(struct pmd_internals *pmd,
remote_flow->msg.nh.nlmsg_flags = NLM_F_REQUEST | NLM_F_ACK;
remote_flow->msg.nh.nlmsg_type = RTM_DELTFILTER;
- ret = tap_nl_send(pmd->nlsk_fd, &remote_flow->msg.nh);
+ ret = rte_nl_send(pmd->nlsk_fd, &remote_flow->msg.nh);
if (ret < 0) {
rte_flow_error_set(
error, ENOMEM, RTE_FLOW_ERROR_TYPE_HANDLE,
NULL, "Failure sending nl request");
goto end;
}
- ret = tap_nl_recv_ack(pmd->nlsk_fd);
+ ret = rte_nl_recv_ack(pmd->nlsk_fd);
if (ret < 0 && errno == ENOENT)
ret = 0;
if (ret < 0) {
@@ -1709,12 +1709,12 @@ int tap_flow_implicit_create(struct pmd_internals *pmd,
RTE_LOG(ERR, PMD, "rte flow rule validation failed\n");
goto fail;
}
- err = tap_nl_send(pmd->nlsk_fd, &msg->nh);
+ err = rte_nl_send(pmd->nlsk_fd, &msg->nh);
if (err < 0) {
RTE_LOG(ERR, PMD, "Failure sending nl request\n");
goto fail;
}
- err = tap_nl_recv_ack(pmd->nlsk_fd);
+ err = rte_nl_recv_ack(pmd->nlsk_fd);
if (err < 0) {
/* Silently ignore re-entering remote promiscuous rule */
if (errno == EEXIST && idx == TAP_REMOTE_PROMISC)
@@ -1882,13 +1882,13 @@ static int rss_enable(struct pmd_internals *pmd,
msg->t.tcm_info = TC_H_MAKE(prio << 16, msg->t.tcm_info);
msg->t.tcm_parent = TC_H_MAKE(MULTIQ_MAJOR_HANDLE, 0);
- tap_nlattr_add(&msg->nh, TCA_KIND, sizeof("bpf"), "bpf");
- if (tap_nlattr_nested_start(msg, TCA_OPTIONS) < 0)
+ rte_nlattr_add(&msg->nh, TCA_KIND, sizeof("bpf"), "bpf");
+ if (rte_nlattr_nested_start(msg, TCA_OPTIONS) < 0)
return -1;
- tap_nlattr_add32(&msg->nh, TCA_BPF_FD, pmd->bpf_fd[i]);
+ rte_nlattr_add32(&msg->nh, TCA_BPF_FD, pmd->bpf_fd[i]);
snprintf(annotation, sizeof(annotation), "[%s%d]",
SEC_NAME_CLS_Q, i);
- tap_nlattr_add(&msg->nh, TCA_BPF_NAME, strlen(annotation) + 1,
+ rte_nlattr_add(&msg->nh, TCA_BPF_NAME, strlen(annotation) + 1,
annotation);
/* Actions */
{
@@ -1904,12 +1904,12 @@ static int rss_enable(struct pmd_internals *pmd,
if (add_actions(rss_flow, 1, &adata, TCA_BPF_ACT) < 0)
return -1;
}
- tap_nlattr_nested_finish(msg); /* nested TCA_OPTIONS */
+ rte_nlattr_nested_finish(msg); /* nested TCA_OPTIONS */
/* Netlink message is now ready to be sent */
- if (tap_nl_send(pmd->nlsk_fd, &msg->nh) < 0)
+ if (rte_nl_send(pmd->nlsk_fd, &msg->nh) < 0)
return -1;
- err = tap_nl_recv_ack(pmd->nlsk_fd);
+ err = rte_nl_recv_ack(pmd->nlsk_fd);
if (err < 0) {
RTE_LOG(ERR, PMD,
"Kernel refused TC filter rule creation (%d): %s\n",
diff --git a/drivers/net/tap/tap_netlink.h b/drivers/net/tap/tap_netlink.h
deleted file mode 100644
index fafef8401..000000000
--- a/drivers/net/tap/tap_netlink.h
+++ /dev/null
@@ -1,42 +0,0 @@
-/* SPDX-License-Identifier: BSD-3-Clause
- * Copyright 2017 6WIND S.A.
- * Copyright 2017 Mellanox.
- */
-
-#ifndef _TAP_NETLINK_H_
-#define _TAP_NETLINK_H_
-
-#include <ctype.h>
-#include <inttypes.h>
-#include <linux/rtnetlink.h>
-#include <linux/netlink.h>
-#include <stdio.h>
-
-#include <rte_log.h>
-
-#define NLMSG_BUF 512
-
-struct nlmsg {
- struct nlmsghdr nh;
- struct tcmsg t;
- char buf[NLMSG_BUF];
- struct nested_tail *nested_tails;
-};
-
-#define NLMSG_TAIL(nlh) (void *)((char *)(nlh) + NLMSG_ALIGN((nlh)->nlmsg_len))
-
-int tap_nl_init(uint32_t nl_groups);
-int tap_nl_final(int nlsk_fd);
-int tap_nl_send(int nlsk_fd, struct nlmsghdr *nh);
-int tap_nl_recv(int nlsk_fd, int (*callback)(struct nlmsghdr *, void *),
- void *arg);
-int tap_nl_recv_ack(int nlsk_fd);
-void tap_nlattr_add(struct nlmsghdr *nh, unsigned short type,
- unsigned int data_len, const void *data);
-void tap_nlattr_add8(struct nlmsghdr *nh, unsigned short type, uint8_t data);
-void tap_nlattr_add16(struct nlmsghdr *nh, unsigned short type, uint16_t data);
-void tap_nlattr_add32(struct nlmsghdr *nh, unsigned short type, uint32_t data);
-int tap_nlattr_nested_start(struct nlmsg *msg, uint16_t type);
-void tap_nlattr_nested_finish(struct nlmsg *msg);
-
-#endif /* _TAP_NETLINK_H_ */
diff --git a/drivers/net/tap/tap_tcmsgs.c b/drivers/net/tap/tap_tcmsgs.c
index 954f13ebb..69638dfc7 100644
--- a/drivers/net/tap/tap_tcmsgs.c
+++ b/drivers/net/tap/tap_tcmsgs.c
@@ -79,7 +79,7 @@ qdisc_del(int nlsk_fd, uint16_t ifindex, struct qdisc *qinfo)
msg.t.tcm_parent = qinfo->parent;
/* if no netlink socket is provided, create one */
if (!nlsk_fd) {
- fd = tap_nl_init(0);
+ fd = rte_nl_init(0);
if (fd < 0) {
RTE_LOG(ERR, PMD,
"Could not delete QDISC: null netlink socket\n");
@@ -88,16 +88,16 @@ qdisc_del(int nlsk_fd, uint16_t ifindex, struct qdisc *qinfo)
} else {
fd = nlsk_fd;
}
- if (tap_nl_send(fd, &msg.nh) < 0)
+ if (rte_nl_send(fd, &msg.nh) < 0)
goto error;
- if (tap_nl_recv_ack(fd) < 0)
+ if (rte_nl_recv_ack(fd) < 0)
goto error;
if (!nlsk_fd)
- return tap_nl_final(fd);
+ return rte_nl_final(fd);
return 0;
error:
if (!nlsk_fd)
- tap_nl_final(fd);
+ rte_nl_final(fd);
return -1;
}
@@ -122,11 +122,11 @@ qdisc_add_multiq(int nlsk_fd, uint16_t ifindex)
NLM_F_REQUEST | NLM_F_ACK | NLM_F_EXCL | NLM_F_CREATE);
msg.t.tcm_handle = TC_H_MAKE(MULTIQ_MAJOR_HANDLE, 0);
msg.t.tcm_parent = TC_H_ROOT;
- tap_nlattr_add(&msg.nh, TCA_KIND, sizeof("multiq"), "multiq");
- tap_nlattr_add(&msg.nh, TCA_OPTIONS, sizeof(opt), &opt);
- if (tap_nl_send(nlsk_fd, &msg.nh) < 0)
+ rte_nlattr_add(&msg.nh, TCA_KIND, sizeof("multiq"), "multiq");
+ rte_nlattr_add(&msg.nh, TCA_OPTIONS, sizeof(opt), &opt);
+ if (rte_nl_send(nlsk_fd, &msg.nh) < 0)
return -1;
- if (tap_nl_recv_ack(nlsk_fd) < 0)
+ if (rte_nl_recv_ack(nlsk_fd) < 0)
return -1;
return 0;
}
@@ -151,10 +151,10 @@ qdisc_add_ingress(int nlsk_fd, uint16_t ifindex)
NLM_F_REQUEST | NLM_F_ACK | NLM_F_EXCL | NLM_F_CREATE);
msg.t.tcm_handle = TC_H_MAKE(TC_H_INGRESS, 0);
msg.t.tcm_parent = TC_H_INGRESS;
- tap_nlattr_add(&msg.nh, TCA_KIND, sizeof("ingress"), "ingress");
- if (tap_nl_send(nlsk_fd, &msg.nh) < 0)
+ rte_nlattr_add(&msg.nh, TCA_KIND, sizeof("ingress"), "ingress");
+ if (rte_nl_send(nlsk_fd, &msg.nh) < 0)
return -1;
- if (tap_nl_recv_ack(nlsk_fd) < 0)
+ if (rte_nl_recv_ack(nlsk_fd) < 0)
return -1;
return 0;
}
@@ -218,9 +218,9 @@ qdisc_iterate(int nlsk_fd, uint16_t ifindex,
};
tc_init_msg(&msg, ifindex, RTM_GETQDISC, NLM_F_REQUEST | NLM_F_DUMP);
- if (tap_nl_send(nlsk_fd, &msg.nh) < 0)
+ if (rte_nl_send(nlsk_fd, &msg.nh) < 0)
return -1;
- if (tap_nl_recv(nlsk_fd, callback, &args) < 0)
+ if (rte_nl_recv(nlsk_fd, callback, &args) < 0)
return -1;
return 0;
}
diff --git a/drivers/net/tap/tap_tcmsgs.h b/drivers/net/tap/tap_tcmsgs.h
index f72f8c5c2..73f74502a 100644
--- a/drivers/net/tap/tap_tcmsgs.h
+++ b/drivers/net/tap/tap_tcmsgs.h
@@ -20,7 +20,7 @@
#include <inttypes.h>
#include <rte_ether.h>
-#include <tap_netlink.h>
+#include <rte_netlink.h>
#define MULTIQ_MAJOR_HANDLE (1 << 16)
diff --git a/lib/Makefile b/lib/Makefile
index ec965a606..699f962d9 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -103,5 +103,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_KNI) += librte_kni
endif
DEPDIRS-librte_kni := librte_eal librte_mempool librte_mbuf librte_ether
DEPDIRS-librte_kni += librte_pci
+DIRS-$(CONFIG_RTE_LIBRTE_NETLINK) += librte_netlink
+DEPDIRS-librte_netlink := librte_eal
include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/lib/librte_netlink/Makefile b/lib/librte_netlink/Makefile
new file mode 100644
index 000000000..89300284f
--- /dev/null
+++ b/lib/librte_netlink/Makefile
@@ -0,0 +1,26 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2010-2017 Intel Corporation
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+#
+# library name
+#
+LIB = librte_netlink.a
+
+CFLAGS += -O3
+CFLAGS += $(WERROR_FLAGS)
+LDLIBS += -lrte_eal
+
+EXPORT_MAP := rte_netlink_version.map
+
+LIBABIVER := 1
+
+SRCS-y += rte_netlink.c
+
+#
+# Export include files
+#
+SYMLINK-y-include += rte_netlink.h
+
+include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/lib/librte_netlink/meson.build b/lib/librte_netlink/meson.build
new file mode 100644
index 000000000..77592c7d4
--- /dev/null
+++ b/lib/librte_netlink/meson.build
@@ -0,0 +1,13 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright 2018 6WIND S.A.
+# Copyright 2018 Mellanox Technologies, Ltd.
+
+
+name = 'netlink'
+version = 1
+allow_experimental_apis = true
+sources = files('rte_netlink.c',)
+
+headers = files('rte_netlink.h')
+
+deps += ['net']
diff --git a/drivers/net/tap/tap_netlink.c b/lib/librte_netlink/rte_netlink.c
similarity index 88%
rename from drivers/net/tap/tap_netlink.c
rename to lib/librte_netlink/rte_netlink.c
index 82c8dc0e6..a4ed07d30 100644
--- a/drivers/net/tap/tap_netlink.c
+++ b/lib/librte_netlink/rte_netlink.c
@@ -11,7 +11,7 @@
#include <unistd.h>
#include <rte_malloc.h>
-#include <tap_netlink.h>
+#include <rte_netlink.h>
#include <rte_random.h>
/* Must be quite large to support dumping a huge list of QDISC or filters. */
@@ -35,7 +35,7 @@ struct nested_tail {
* netlink socket file descriptor on success, -1 otherwise.
*/
int
-tap_nl_init(uint32_t nl_groups)
+rte_nl_init(uint32_t nl_groups)
{
int fd, sndbuf_size = SNDBUF_SIZE, rcvbuf_size = RCVBUF_SIZE;
struct sockaddr_nl local = {
@@ -73,7 +73,7 @@ tap_nl_init(uint32_t nl_groups)
* 0 on success, -1 otherwise.
*/
int
-tap_nl_final(int nlsk_fd)
+rte_nl_final(int nlsk_fd)
{
if (close(nlsk_fd)) {
RTE_LOG(ERR, PMD, "Failed to close netlink socket: %s (%d)\n",
@@ -95,7 +95,7 @@ tap_nl_final(int nlsk_fd)
* the number of sent bytes on success, -1 otherwise.
*/
int
-tap_nl_send(int nlsk_fd, struct nlmsghdr *nh)
+rte_nl_send(int nlsk_fd, struct nlmsghdr *nh)
{
/* man 7 netlink EXAMPLE */
struct sockaddr_nl sa = {
@@ -126,7 +126,7 @@ tap_nl_send(int nlsk_fd, struct nlmsghdr *nh)
/**
* Check that the kernel sends an appropriate ACK in response
- * to an tap_nl_send().
+ * to an rte_nl_send().
*
* @param[in] nlsk_fd
* The netlink socket file descriptor used for communication.
@@ -135,14 +135,14 @@ tap_nl_send(int nlsk_fd, struct nlmsghdr *nh)
* 0 on success, -1 otherwise with errno set.
*/
int
-tap_nl_recv_ack(int nlsk_fd)
+rte_nl_recv_ack(int nlsk_fd)
{
- return tap_nl_recv(nlsk_fd, NULL, NULL);
+ return rte_nl_recv(nlsk_fd, NULL, NULL);
}
/**
* Receive a message from the kernel on the netlink socket, following an
- * tap_nl_send().
+ * rte_nl_send().
*
* @param[in] nlsk_fd
* The netlink socket file descriptor used for communication.
@@ -155,7 +155,7 @@ tap_nl_recv_ack(int nlsk_fd)
* 0 on success, -1 otherwise with errno set.
*/
int
-tap_nl_recv(int nlsk_fd, int (*cb)(struct nlmsghdr *, void *arg), void *arg)
+rte_nl_recv(int nlsk_fd, int (*cb)(struct nlmsghdr *, void *arg), void *arg)
{
/* man 7 netlink EXAMPLE */
struct sockaddr_nl sa;
@@ -220,7 +220,7 @@ tap_nl_recv(int nlsk_fd, int (*cb)(struct nlmsghdr *, void *arg), void *arg)
* The data to append.
*/
void
-tap_nlattr_add(struct nlmsghdr *nh, unsigned short type,
+rte_nlattr_add(struct nlmsghdr *nh, unsigned short type,
unsigned int data_len, const void *data)
{
/* see man 3 rtnetlink */
@@ -244,9 +244,9 @@ tap_nlattr_add(struct nlmsghdr *nh, unsigned short type,
* The data to append.
*/
void
-tap_nlattr_add8(struct nlmsghdr *nh, unsigned short type, uint8_t data)
+rte_nlattr_add8(struct nlmsghdr *nh, unsigned short type, uint8_t data)
{
- tap_nlattr_add(nh, type, sizeof(uint8_t), &data);
+ rte_nlattr_add(nh, type, sizeof(uint8_t), &data);
}
/**
@@ -260,9 +260,9 @@ tap_nlattr_add8(struct nlmsghdr *nh, unsigned short type, uint8_t data)
* The data to append.
*/
void
-tap_nlattr_add16(struct nlmsghdr *nh, unsigned short type, uint16_t data)
+rte_nlattr_add16(struct nlmsghdr *nh, unsigned short type, uint16_t data)
{
- tap_nlattr_add(nh, type, sizeof(uint16_t), &data);
+ rte_nlattr_add(nh, type, sizeof(uint16_t), &data);
}
/**
@@ -276,14 +276,14 @@ tap_nlattr_add16(struct nlmsghdr *nh, unsigned short type, uint16_t data)
* The data to append.
*/
void
-tap_nlattr_add32(struct nlmsghdr *nh, unsigned short type, uint32_t data)
+rte_nlattr_add32(struct nlmsghdr *nh, unsigned short type, uint32_t data)
{
- tap_nlattr_add(nh, type, sizeof(uint32_t), &data);
+ rte_nlattr_add(nh, type, sizeof(uint32_t), &data);
}
/**
* Start a nested netlink attribute.
- * It must be followed later by a call to tap_nlattr_nested_finish().
+ * It must be followed later by a call to rte_nlattr_nested_finish().
*
* @param[in, out] msg
* The netlink message where to edit the nested_tails metadata.
@@ -294,7 +294,7 @@ tap_nlattr_add32(struct nlmsghdr *nh, unsigned short type, uint32_t data)
* -1 if adding a nested netlink attribute failed, 0 otherwise.
*/
int
-tap_nlattr_nested_start(struct nlmsg *msg, uint16_t type)
+rte_nlattr_nested_start(struct nlmsg *msg, uint16_t type)
{
struct nested_tail *tail;
@@ -308,7 +308,7 @@ tap_nlattr_nested_start(struct nlmsg *msg, uint16_t type)
tail->tail = (struct rtattr *)NLMSG_TAIL(&msg->nh);
- tap_nlattr_add(&msg->nh, type, 0, NULL);
+ rte_nlattr_add(&msg->nh, type, 0, NULL);
tail->prev = msg->nested_tails;
@@ -319,7 +319,7 @@ tap_nlattr_nested_start(struct nlmsg *msg, uint16_t type)
/**
* End a nested netlink attribute.
- * It follows a call to tap_nlattr_nested_start().
+ * It follows a call to rte_nlattr_nested_start().
* In effect, it will modify the nested attribute length to include every bytes
* from the nested attribute start, up to here.
*
@@ -327,7 +327,7 @@ tap_nlattr_nested_start(struct nlmsg *msg, uint16_t type)
* The netlink message where to edit the nested_tails metadata.
*/
void
-tap_nlattr_nested_finish(struct nlmsg *msg)
+rte_nlattr_nested_finish(struct nlmsg *msg)
{
struct nested_tail *tail = msg->nested_tails;
diff --git a/lib/librte_netlink/rte_netlink.h b/lib/librte_netlink/rte_netlink.h
new file mode 100644
index 000000000..29f7d64c5
--- /dev/null
+++ b/lib/librte_netlink/rte_netlink.h
@@ -0,0 +1,42 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2017 6WIND S.A.
+ * Copyright 2017 Mellanox.
+ */
+
+#ifndef _RTE_NETLINK_H_
+#define _RTE_NETLINK_H_
+
+#include <ctype.h>
+#include <inttypes.h>
+#include <linux/rtnetlink.h>
+#include <linux/netlink.h>
+#include <stdio.h>
+
+#include <rte_log.h>
+
+#define NLMSG_BUF 512
+
+struct nlmsg {
+ struct nlmsghdr nh;
+ struct tcmsg t;
+ char buf[NLMSG_BUF];
+ struct nested_tail *nested_tails;
+};
+
+#define NLMSG_TAIL(nlh) (void *)((char *)(nlh) + NLMSG_ALIGN((nlh)->nlmsg_len))
+
+int rte_nl_init(uint32_t nl_groups);
+int rte_nl_final(int nlsk_fd);
+int rte_nl_send(int nlsk_fd, struct nlmsghdr *nh);
+int rte_nl_recv(int nlsk_fd, int (*callback)(struct nlmsghdr *, void *),
+ void *arg);
+int rte_nl_recv_ack(int nlsk_fd);
+void rte_nlattr_add(struct nlmsghdr *nh, unsigned short type,
+ unsigned int data_len, const void *data);
+void rte_nlattr_add8(struct nlmsghdr *nh, unsigned short type, uint8_t data);
+void rte_nlattr_add16(struct nlmsghdr *nh, unsigned short type, uint16_t data);
+void rte_nlattr_add32(struct nlmsghdr *nh, unsigned short type, uint32_t data);
+int rte_nlattr_nested_start(struct nlmsg *msg, uint16_t type);
+void rte_nlattr_nested_finish(struct nlmsg *msg);
+
+#endif /* _RTE_NETLINK_H_ */
diff --git a/lib/librte_netlink/rte_netlink_version.map b/lib/librte_netlink/rte_netlink_version.map
new file mode 100644
index 000000000..5aa9b6228
--- /dev/null
+++ b/lib/librte_netlink/rte_netlink_version.map
@@ -0,0 +1,17 @@
+DPDK_18.05 {
+ global:
+
+ rte_nl_init;
+ rte_nl_final;
+ rte_nl_send;
+ rte_nl_recv;
+ rte_nl_recv_ack;
+ rte_nlattr_add;
+ rte_nlattr_add8;
+ rte_nlattr_add16;
+ rte_nlattr_add32;
+ rte_nlattr_nested_start;
+ rte_nlattr_nested_finish;
+
+ local: *;
+};
diff --git a/lib/meson.build b/lib/meson.build
index ef6159170..873a39785 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -23,7 +23,7 @@ libraries = [ 'compat', # just a header, used for versioning
# add pkt framework libs which use other libs from above
'port', 'table', 'pipeline',
# flow_classify lib depends on pkt framework table lib
- 'flow_classify']
+ 'flow_classify', 'netlink']
foreach l:libraries
build = true
diff --git a/mk/rte.app.mk b/mk/rte.app.mk
index 3eb41d176..5cc8ed622 100644
--- a/mk/rte.app.mk
+++ b/mk/rte.app.mk
@@ -106,6 +106,7 @@ _LDLIBS-$(CONFIG_RTE_LIBRTE_EAL) += -lrte_eal
_LDLIBS-$(CONFIG_RTE_LIBRTE_CMDLINE) += -lrte_cmdline
_LDLIBS-$(CONFIG_RTE_LIBRTE_REORDER) += -lrte_reorder
_LDLIBS-$(CONFIG_RTE_LIBRTE_SCHED) += -lrte_sched
+_LDLIBS-$(CONFIG_RTE_LIBRTE_NETLINK) += -lrte_netlink
ifeq ($(CONFIG_RTE_EXEC_ENV_LINUXAPP),y)
_LDLIBS-$(CONFIG_RTE_LIBRTE_KNI) += -lrte_kni
--
2.11.0
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PATCH 2/2] lib: add request Netlink messages
2018-03-13 12:28 [PATCH 0/2] lib: move Netlink wrapper to lib Nelio Laranjeiro
2018-03-13 12:28 ` [PATCH 1/2] lib: move Netlink code into a common library Nelio Laranjeiro
@ 2018-03-13 12:28 ` Nelio Laranjeiro
2018-03-13 21:20 ` [PATCH 0/2] lib: move Netlink wrapper to lib Stephen Hemminger
2 siblings, 0 replies; 11+ messages in thread
From: Nelio Laranjeiro @ 2018-03-13 12:28 UTC (permalink / raw)
To: Pascal Mazon; +Cc: dev
rte_nl_send() generate a message without request. In some situation, the
drivers needs to retrieve information from the Kernel before requesting
modifications.
Cc: Pascal Mazon <pascal.mazon@6wind.com>
Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
lib/librte_netlink/rte_netlink.c | 45 ++++++++++++++++++++++++++++++
lib/librte_netlink/rte_netlink.h | 1 +
lib/librte_netlink/rte_netlink_version.map | 1 +
3 files changed, 47 insertions(+)
diff --git a/lib/librte_netlink/rte_netlink.c b/lib/librte_netlink/rte_netlink.c
index a4ed07d30..a2020f6f8 100644
--- a/lib/librte_netlink/rte_netlink.c
+++ b/lib/librte_netlink/rte_netlink.c
@@ -125,6 +125,51 @@ rte_nl_send(int nlsk_fd, struct nlmsghdr *nh)
}
/**
+ * Send a request message to the kernel on the netlink socket.
+ *
+ * @param[in] nlsk_fd
+ * The netlink socket file descriptor used for communication.
+ * @param[in] nh
+ * The netlink message send to the kernel.
+ * @param[in] req
+ * Pointer to the request structure.
+ * @param[in] len
+ * Length of the request in bytes.
+ *
+ * @return
+ * the number of sent bytes on success, -1 otherwise.
+ */
+int
+rte_nl_request(int nlsk_fd, struct nlmsghdr *nh, void *req, int len)
+{
+ /* man 7 netlink EXAMPLE */
+ struct sockaddr_nl sa = {
+ .nl_family = AF_NETLINK,
+ };
+ struct iovec iov[2] = {
+ { .iov_base = nh, .iov_len = sizeof(*nh), },
+ { .iov_base = req, .iov_len = len, },
+ };
+ struct msghdr msg = {
+ .msg_name = &sa,
+ .msg_namelen = sizeof(sa),
+ .msg_iov = iov,
+ .msg_iovlen = 2,
+ };
+ int send_bytes;
+
+ nh->nlmsg_pid = 0; /* communication with the kernel uses pid 0 */
+ nh->nlmsg_seq = (uint32_t)rte_rand();
+ send_bytes = sendmsg(nlsk_fd, &msg, 0);
+ if (send_bytes < 0) {
+ RTE_LOG(ERR, PMD, "Failed to send netlink message: %s (%d)\n",
+ strerror(errno), errno);
+ return -1;
+ }
+ return send_bytes;
+}
+
+/**
* Check that the kernel sends an appropriate ACK in response
* to an rte_nl_send().
*
diff --git a/lib/librte_netlink/rte_netlink.h b/lib/librte_netlink/rte_netlink.h
index 29f7d64c5..9af03e738 100644
--- a/lib/librte_netlink/rte_netlink.h
+++ b/lib/librte_netlink/rte_netlink.h
@@ -28,6 +28,7 @@ struct nlmsg {
int rte_nl_init(uint32_t nl_groups);
int rte_nl_final(int nlsk_fd);
int rte_nl_send(int nlsk_fd, struct nlmsghdr *nh);
+int rte_nl_request(int nlsk_fd, struct nlmsghdr *nh, void *req, int len);
int rte_nl_recv(int nlsk_fd, int (*callback)(struct nlmsghdr *, void *),
void *arg);
int rte_nl_recv_ack(int nlsk_fd);
diff --git a/lib/librte_netlink/rte_netlink_version.map b/lib/librte_netlink/rte_netlink_version.map
index 5aa9b6228..a52254b36 100644
--- a/lib/librte_netlink/rte_netlink_version.map
+++ b/lib/librte_netlink/rte_netlink_version.map
@@ -4,6 +4,7 @@ DPDK_18.05 {
rte_nl_init;
rte_nl_final;
rte_nl_send;
+ rte_nl_request;
rte_nl_recv;
rte_nl_recv_ack;
rte_nlattr_add;
--
2.11.0
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH 1/2] lib: move Netlink code into a common library
2018-03-13 12:28 ` [PATCH 1/2] lib: move Netlink code into a common library Nelio Laranjeiro
@ 2018-03-13 13:58 ` Bruce Richardson
2018-03-13 14:58 ` Nélio Laranjeiro
0 siblings, 1 reply; 11+ messages in thread
From: Bruce Richardson @ 2018-03-13 13:58 UTC (permalink / raw)
To: Nelio Laranjeiro; +Cc: Pascal Mazon, dev
On Tue, Mar 13, 2018 at 01:28:26PM +0100, Nelio Laranjeiro wrote:
> Moves TAP PMD generic Netlink library into lib directory to let other PMD
> use Netlink to communicate with kernel.
>
> As this library uses a socket to communicate with Netlink it does not bring
> any dependency on the libnl.
>
> Cc: Pascal Mazon <pascal.mazon@6wind.com>
>
> Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
> ---
> MAINTAINERS | 3 +
> config/common_base | 6 ++
> config/common_linuxapp | 1 +
> drivers/net/tap/Makefile | 3 +-
> drivers/net/tap/rte_eth_tap.c | 16 +--
> drivers/net/tap/tap_flow.c | 118 ++++++++++-----------
> drivers/net/tap/tap_netlink.h | 42 --------
> drivers/net/tap/tap_tcmsgs.c | 28 ++---
> drivers/net/tap/tap_tcmsgs.h | 2 +-
> lib/Makefile | 2 +
> lib/librte_netlink/Makefile | 26 +++++
> lib/librte_netlink/meson.build | 13 +++
> .../librte_netlink/rte_netlink.c | 42 ++++----
> lib/librte_netlink/rte_netlink.h | 42 ++++++++
> lib/librte_netlink/rte_netlink_version.map | 17 +++
> lib/meson.build | 2 +-
> mk/rte.app.mk | 1 +
> 17 files changed, 216 insertions(+), 148 deletions(-)
> delete mode 100644 drivers/net/tap/tap_netlink.h
> create mode 100644 lib/librte_netlink/Makefile
> create mode 100644 lib/librte_netlink/meson.build
> rename drivers/net/tap/tap_netlink.c => lib/librte_netlink/rte_netlink.c (88%)
> create mode 100644 lib/librte_netlink/rte_netlink.h
> create mode 100644 lib/librte_netlink/rte_netlink_version.map
>
<snip>
> diff --git a/lib/librte_netlink/meson.build b/lib/librte_netlink/meson.build
> new file mode 100644
> index 000000000..77592c7d4
> --- /dev/null
> +++ b/lib/librte_netlink/meson.build
> @@ -0,0 +1,13 @@
> +# SPDX-License-Identifier: BSD-3-Clause
> +# Copyright 2018 6WIND S.A.
> +# Copyright 2018 Mellanox Technologies, Ltd.
> +
> +
> +name = 'netlink'
> +version = 1
Not a big issue, but the above two lines are unnecessary, and can be
omitted. I'd rather than the name of the lib *not* be explicitly called
out, to prevent the temptation of giving the library a different name to
the name of the directory it is in.
> +allow_experimental_apis = true
> +sources = files('rte_netlink.c',)
> +
> +headers = files('rte_netlink.h')
> +
> +deps += ['net']
<snip>
> diff --git a/lib/meson.build b/lib/meson.build
> index ef6159170..873a39785 100644
> --- a/lib/meson.build
> +++ b/lib/meson.build
> @@ -23,7 +23,7 @@ libraries = [ 'compat', # just a header, used for versioning
> # add pkt framework libs which use other libs from above
> 'port', 'table', 'pipeline',
> # flow_classify lib depends on pkt framework table lib
> - 'flow_classify']
> + 'flow_classify', 'netlink']
Given that this doesn't depend on anything apart from the core libs,
I think it should be placed earlier in the list, where the majority of
the libs are in alphabetical order. [Yes, I should have put a comment
just before the "acl" lib entry calling out the fact that it was meant
to be an alphabetical list - I forgot, sorry! :-)]
/Bruce
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH 1/2] lib: move Netlink code into a common library
2018-03-13 13:58 ` Bruce Richardson
@ 2018-03-13 14:58 ` Nélio Laranjeiro
0 siblings, 0 replies; 11+ messages in thread
From: Nélio Laranjeiro @ 2018-03-13 14:58 UTC (permalink / raw)
To: Bruce Richardson; +Cc: Pascal Mazon, dev
On Tue, Mar 13, 2018 at 01:58:28PM +0000, Bruce Richardson wrote:
> On Tue, Mar 13, 2018 at 01:28:26PM +0100, Nelio Laranjeiro wrote:
> > Moves TAP PMD generic Netlink library into lib directory to let other PMD
> > use Netlink to communicate with kernel.
> >
> > As this library uses a socket to communicate with Netlink it does not bring
> > any dependency on the libnl.
> >
> > Cc: Pascal Mazon <pascal.mazon@6wind.com>
> >
> > Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
> > ---
> > MAINTAINERS | 3 +
> > config/common_base | 6 ++
> > config/common_linuxapp | 1 +
> > drivers/net/tap/Makefile | 3 +-
> > drivers/net/tap/rte_eth_tap.c | 16 +--
> > drivers/net/tap/tap_flow.c | 118 ++++++++++-----------
> > drivers/net/tap/tap_netlink.h | 42 --------
> > drivers/net/tap/tap_tcmsgs.c | 28 ++---
> > drivers/net/tap/tap_tcmsgs.h | 2 +-
> > lib/Makefile | 2 +
> > lib/librte_netlink/Makefile | 26 +++++
> > lib/librte_netlink/meson.build | 13 +++
> > .../librte_netlink/rte_netlink.c | 42 ++++----
> > lib/librte_netlink/rte_netlink.h | 42 ++++++++
> > lib/librte_netlink/rte_netlink_version.map | 17 +++
> > lib/meson.build | 2 +-
> > mk/rte.app.mk | 1 +
> > 17 files changed, 216 insertions(+), 148 deletions(-)
> > delete mode 100644 drivers/net/tap/tap_netlink.h
> > create mode 100644 lib/librte_netlink/Makefile
> > create mode 100644 lib/librte_netlink/meson.build
> > rename drivers/net/tap/tap_netlink.c => lib/librte_netlink/rte_netlink.c (88%)
> > create mode 100644 lib/librte_netlink/rte_netlink.h
> > create mode 100644 lib/librte_netlink/rte_netlink_version.map
> >
>
> <snip>
>
> > diff --git a/lib/librte_netlink/meson.build b/lib/librte_netlink/meson.build
> > new file mode 100644
> > index 000000000..77592c7d4
> > --- /dev/null
> > +++ b/lib/librte_netlink/meson.build
> > @@ -0,0 +1,13 @@
> > +# SPDX-License-Identifier: BSD-3-Clause
> > +# Copyright 2018 6WIND S.A.
> > +# Copyright 2018 Mellanox Technologies, Ltd.
> > +
> > +
> > +name = 'netlink'
> > +version = 1
>
> Not a big issue, but the above two lines are unnecessary, and can be
> omitted. I'd rather than the name of the lib *not* be explicitly called
> out, to prevent the temptation of giving the library a different name to
> the name of the directory it is in.
>
> > +allow_experimental_apis = true
> > +sources = files('rte_netlink.c',)
> > +
> > +headers = files('rte_netlink.h')
> > +
> > +deps += ['net']
>
> <snip>
>
> > diff --git a/lib/meson.build b/lib/meson.build
> > index ef6159170..873a39785 100644
> > --- a/lib/meson.build
> > +++ b/lib/meson.build
> > @@ -23,7 +23,7 @@ libraries = [ 'compat', # just a header, used for versioning
> > # add pkt framework libs which use other libs from above
> > 'port', 'table', 'pipeline',
> > # flow_classify lib depends on pkt framework table lib
> > - 'flow_classify']
> > + 'flow_classify', 'netlink']
>
> Given that this doesn't depend on anything apart from the core libs,
> I think it should be placed earlier in the list, where the majority of
> the libs are in alphabetical order. [Yes, I should have put a comment
> just before the "acl" lib entry calling out the fact that it was meant
> to be an alphabetical list - I forgot, sorry! :-)]
>
> /Bruce
Hi Bruce,
I will modify it for the v2.
Thanks,
--
Nélio Laranjeiro
6WIND
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH 0/2] lib: move Netlink wrapper to lib
2018-03-13 12:28 [PATCH 0/2] lib: move Netlink wrapper to lib Nelio Laranjeiro
2018-03-13 12:28 ` [PATCH 1/2] lib: move Netlink code into a common library Nelio Laranjeiro
2018-03-13 12:28 ` [PATCH 2/2] lib: add request Netlink messages Nelio Laranjeiro
@ 2018-03-13 21:20 ` Stephen Hemminger
2018-03-14 12:08 ` Nélio Laranjeiro
2 siblings, 1 reply; 11+ messages in thread
From: Stephen Hemminger @ 2018-03-13 21:20 UTC (permalink / raw)
To: Nelio Laranjeiro; +Cc: Pascal Mazon, dev
On Tue, 13 Mar 2018 13:28:25 +0100
Nelio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
> TAP PMD uses a wrapper to communicate with Netlink through a socket. This
> wrapper can be useful for other PMD which may need to communicate also with
> the Linux kernel via Netlink.
>
> This series moves this wrapper to the lib level to let any PMD implement use
> it when necessary.
>
> Nelio Laranjeiro (2):
> lib: move Netlink code into a common library
> lib: add request Netlink messages
>
> MAINTAINERS | 3 +
> config/common_base | 6 ++
> config/common_linuxapp | 1 +
> drivers/net/tap/Makefile | 3 +-
> drivers/net/tap/rte_eth_tap.c | 16 +--
> drivers/net/tap/tap_flow.c | 118 ++++++++++-----------
> drivers/net/tap/tap_netlink.h | 42 --------
> drivers/net/tap/tap_tcmsgs.c | 28 ++---
> drivers/net/tap/tap_tcmsgs.h | 2 +-
> lib/Makefile | 2 +
> lib/librte_netlink/Makefile | 26 +++++
> lib/librte_netlink/meson.build | 13 +++
> .../librte_netlink/rte_netlink.c | 87 +++++++++++----
> lib/librte_netlink/rte_netlink.h | 43 ++++++++
> lib/librte_netlink/rte_netlink_version.map | 18 ++++
> lib/meson.build | 2 +-
> mk/rte.app.mk | 1 +
> 17 files changed, 263 insertions(+), 148 deletions(-)
> delete mode 100644 drivers/net/tap/tap_netlink.h
> create mode 100644 lib/librte_netlink/Makefile
> create mode 100644 lib/librte_netlink/meson.build
> rename drivers/net/tap/tap_netlink.c => lib/librte_netlink/rte_netlink.c (77%)
> create mode 100644 lib/librte_netlink/rte_netlink.h
> create mode 100644 lib/librte_netlink/rte_netlink_version.map
>
I might have raised this before; but having yet another netlink library is
not a great advantage. It would be much better to use a common external library
libmnl which is already available on every distribution.
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH 0/2] lib: move Netlink wrapper to lib
2018-03-13 21:20 ` [PATCH 0/2] lib: move Netlink wrapper to lib Stephen Hemminger
@ 2018-03-14 12:08 ` Nélio Laranjeiro
2018-03-14 15:17 ` Stephen Hemminger
0 siblings, 1 reply; 11+ messages in thread
From: Nélio Laranjeiro @ 2018-03-14 12:08 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: Pascal Mazon, dev
On Tue, Mar 13, 2018 at 02:20:31PM -0700, Stephen Hemminger wrote:
> On Tue, 13 Mar 2018 13:28:25 +0100
> Nelio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
>
> > TAP PMD uses a wrapper to communicate with Netlink through a socket. This
> > wrapper can be useful for other PMD which may need to communicate also with
> > the Linux kernel via Netlink.
> >
> > This series moves this wrapper to the lib level to let any PMD implement use
> > it when necessary.
> >
> > Nelio Laranjeiro (2):
> > lib: move Netlink code into a common library
> > lib: add request Netlink messages
> >
> > MAINTAINERS | 3 +
> > config/common_base | 6 ++
> > config/common_linuxapp | 1 +
> > drivers/net/tap/Makefile | 3 +-
> > drivers/net/tap/rte_eth_tap.c | 16 +--
> > drivers/net/tap/tap_flow.c | 118 ++++++++++-----------
> > drivers/net/tap/tap_netlink.h | 42 --------
> > drivers/net/tap/tap_tcmsgs.c | 28 ++---
> > drivers/net/tap/tap_tcmsgs.h | 2 +-
> > lib/Makefile | 2 +
> > lib/librte_netlink/Makefile | 26 +++++
> > lib/librte_netlink/meson.build | 13 +++
> > .../librte_netlink/rte_netlink.c | 87 +++++++++++----
> > lib/librte_netlink/rte_netlink.h | 43 ++++++++
> > lib/librte_netlink/rte_netlink_version.map | 18 ++++
> > lib/meson.build | 2 +-
> > mk/rte.app.mk | 1 +
> > 17 files changed, 263 insertions(+), 148 deletions(-)
> > delete mode 100644 drivers/net/tap/tap_netlink.h
> > create mode 100644 lib/librte_netlink/Makefile
> > create mode 100644 lib/librte_netlink/meson.build
> > rename drivers/net/tap/tap_netlink.c => lib/librte_netlink/rte_netlink.c (77%)
> > create mode 100644 lib/librte_netlink/rte_netlink.h
> > create mode 100644 lib/librte_netlink/rte_netlink_version.map
> >
>
> I might have raised this before; but having yet another netlink library is
> not a great advantage. It would be much better to use a common external library
> libmnl which is already available on every distribution.
The question is more do we really want to have a dependency on a so
small wrapper for a socket interface?
--
Nélio Laranjeiro
6WIND
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH 0/2] lib: move Netlink wrapper to lib
2018-03-14 12:08 ` Nélio Laranjeiro
@ 2018-03-14 15:17 ` Stephen Hemminger
2018-03-15 15:39 ` Thomas Monjalon
0 siblings, 1 reply; 11+ messages in thread
From: Stephen Hemminger @ 2018-03-14 15:17 UTC (permalink / raw)
To: Nélio Laranjeiro; +Cc: Pascal Mazon, dev
On Wed, 14 Mar 2018 13:08:55 +0100
Nélio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
> On Tue, Mar 13, 2018 at 02:20:31PM -0700, Stephen Hemminger wrote:
> > On Tue, 13 Mar 2018 13:28:25 +0100
> > Nelio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
> >
> > > TAP PMD uses a wrapper to communicate with Netlink through a socket. This
> > > wrapper can be useful for other PMD which may need to communicate also with
> > > the Linux kernel via Netlink.
> > >
> > > This series moves this wrapper to the lib level to let any PMD implement use
> > > it when necessary.
> > >
> > > Nelio Laranjeiro (2):
> > > lib: move Netlink code into a common library
> > > lib: add request Netlink messages
> > >
> > > MAINTAINERS | 3 +
> > > config/common_base | 6 ++
> > > config/common_linuxapp | 1 +
> > > drivers/net/tap/Makefile | 3 +-
> > > drivers/net/tap/rte_eth_tap.c | 16 +--
> > > drivers/net/tap/tap_flow.c | 118 ++++++++++-----------
> > > drivers/net/tap/tap_netlink.h | 42 --------
> > > drivers/net/tap/tap_tcmsgs.c | 28 ++---
> > > drivers/net/tap/tap_tcmsgs.h | 2 +-
> > > lib/Makefile | 2 +
> > > lib/librte_netlink/Makefile | 26 +++++
> > > lib/librte_netlink/meson.build | 13 +++
> > > .../librte_netlink/rte_netlink.c | 87 +++++++++++----
> > > lib/librte_netlink/rte_netlink.h | 43 ++++++++
> > > lib/librte_netlink/rte_netlink_version.map | 18 ++++
> > > lib/meson.build | 2 +-
> > > mk/rte.app.mk | 1 +
> > > 17 files changed, 263 insertions(+), 148 deletions(-)
> > > delete mode 100644 drivers/net/tap/tap_netlink.h
> > > create mode 100644 lib/librte_netlink/Makefile
> > > create mode 100644 lib/librte_netlink/meson.build
> > > rename drivers/net/tap/tap_netlink.c => lib/librte_netlink/rte_netlink.c (77%)
> > > create mode 100644 lib/librte_netlink/rte_netlink.h
> > > create mode 100644 lib/librte_netlink/rte_netlink_version.map
> > >
> >
> > I might have raised this before; but having yet another netlink library is
> > not a great advantage. It would be much better to use a common external library
> > libmnl which is already available on every distribution.
>
> The question is more do we really want to have a dependency on a so
> small wrapper for a socket interface?
>
Lots of drivers already have dependencies.
And I trust libmnl rather than a reinvented library.
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH 0/2] lib: move Netlink wrapper to lib
2018-03-14 15:17 ` Stephen Hemminger
@ 2018-03-15 15:39 ` Thomas Monjalon
2018-03-15 16:19 ` Nélio Laranjeiro
0 siblings, 1 reply; 11+ messages in thread
From: Thomas Monjalon @ 2018-03-15 15:39 UTC (permalink / raw)
To: Stephen Hemminger, Nélio Laranjeiro; +Cc: dev, Pascal Mazon
14/03/2018 16:17, Stephen Hemminger:
> On Wed, 14 Mar 2018 13:08:55 +0100
> Nélio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
> > On Tue, Mar 13, 2018 at 02:20:31PM -0700, Stephen Hemminger wrote:
> > > On Tue, 13 Mar 2018 13:28:25 +0100
> > > Nelio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
> > > > drivers/net/tap/Makefile | 3 +-
> > > > drivers/net/tap/rte_eth_tap.c | 16 +--
> > > > drivers/net/tap/tap_flow.c | 118 ++++++++++-----------
> > > > drivers/net/tap/tap_netlink.h | 42 --------
> > > > drivers/net/tap/tap_tcmsgs.c | 28 ++---
> > > > drivers/net/tap/tap_tcmsgs.h | 2 +-
> > > > lib/Makefile | 2 +
> > > > lib/librte_netlink/Makefile | 26 +++++
> > > > lib/librte_netlink/meson.build | 13 +++
> > > > .../librte_netlink/rte_netlink.c | 87 +++++++++++----
> > > > lib/librte_netlink/rte_netlink.h | 43 ++++++++
> > > > lib/librte_netlink/rte_netlink_version.map | 18 ++++
> > >
> > > I might have raised this before; but having yet another netlink library is
> > > not a great advantage. It would be much better to use a common external library
> > > libmnl which is already available on every distribution.
> >
> > The question is more do we really want to have a dependency on a so
> > small wrapper for a socket interface?
>
> Lots of drivers already have dependencies.
> And I trust libmnl rather than a reinvented library.
Are we sure it is commonly installed in distributions?
This is needed only for TAP and mlx drivers.
Do we really want to add this dependency as mandatory for everybody
to compile DPDK with default PMDs?
I see 4 options:
1/ keep duplicated few functions in PMDs
2/ use libmnl
3/ add the functions in a DPDK lib (this patch)
4/ add the functions in EAL
I would vote for the first option because mlx5 has very minimal
requirement regarding netlink.
The first option can give us some time to think about the right way
to use netlink in the next releases.
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH 0/2] lib: move Netlink wrapper to lib
2018-03-15 15:39 ` Thomas Monjalon
@ 2018-03-15 16:19 ` Nélio Laranjeiro
2018-03-15 16:26 ` Stephen Hemminger
0 siblings, 1 reply; 11+ messages in thread
From: Nélio Laranjeiro @ 2018-03-15 16:19 UTC (permalink / raw)
To: Thomas Monjalon; +Cc: Stephen Hemminger, dev, Pascal Mazon
On Thu, Mar 15, 2018 at 04:39:46PM +0100, Thomas Monjalon wrote:
> 14/03/2018 16:17, Stephen Hemminger:
> > On Wed, 14 Mar 2018 13:08:55 +0100
> > Nélio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
> > > On Tue, Mar 13, 2018 at 02:20:31PM -0700, Stephen Hemminger wrote:
> > > > On Tue, 13 Mar 2018 13:28:25 +0100
> > > > Nelio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
> > > > > drivers/net/tap/Makefile | 3 +-
> > > > > drivers/net/tap/rte_eth_tap.c | 16 +--
> > > > > drivers/net/tap/tap_flow.c | 118 ++++++++++-----------
> > > > > drivers/net/tap/tap_netlink.h | 42 --------
> > > > > drivers/net/tap/tap_tcmsgs.c | 28 ++---
> > > > > drivers/net/tap/tap_tcmsgs.h | 2 +-
> > > > > lib/Makefile | 2 +
> > > > > lib/librte_netlink/Makefile | 26 +++++
> > > > > lib/librte_netlink/meson.build | 13 +++
> > > > > .../librte_netlink/rte_netlink.c | 87 +++++++++++----
> > > > > lib/librte_netlink/rte_netlink.h | 43 ++++++++
> > > > > lib/librte_netlink/rte_netlink_version.map | 18 ++++
> > > >
> > > > I might have raised this before; but having yet another netlink library is
> > > > not a great advantage. It would be much better to use a common external library
> > > > libmnl which is already available on every distribution.
> > >
> > > The question is more do we really want to have a dependency on a so
> > > small wrapper for a socket interface?
> >
> > Lots of drivers already have dependencies.
> > And I trust libmnl rather than a reinvented library.
>
> Are we sure it is commonly installed in distributions?
>
> This is needed only for TAP and mlx drivers.
> Do we really want to add this dependency as mandatory for everybody
> to compile DPDK with default PMDs?
>
> I see 4 options:
> 1/ keep duplicated few functions in PMDs
> 2/ use libmnl
> 3/ add the functions in a DPDK lib (this patch)
> 4/ add the functions in EAL
>
> I would vote for the first option because mlx5 has very minimal
> requirement regarding netlink.
> The first option can give us some time to think about the right way
> to use netlink in the next releases.
I also vote for option 1 until those 4 and maybe more options are
discussed.
--
Nélio Laranjeiro
6WIND
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH 0/2] lib: move Netlink wrapper to lib
2018-03-15 16:19 ` Nélio Laranjeiro
@ 2018-03-15 16:26 ` Stephen Hemminger
0 siblings, 0 replies; 11+ messages in thread
From: Stephen Hemminger @ 2018-03-15 16:26 UTC (permalink / raw)
To: Nélio Laranjeiro; +Cc: Thomas Monjalon, dev, Pascal Mazon
On Thu, 15 Mar 2018 17:19:05 +0100
Nélio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
> On Thu, Mar 15, 2018 at 04:39:46PM +0100, Thomas Monjalon wrote:
> > 14/03/2018 16:17, Stephen Hemminger:
> > > On Wed, 14 Mar 2018 13:08:55 +0100
> > > Nélio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
> > > > On Tue, Mar 13, 2018 at 02:20:31PM -0700, Stephen Hemminger wrote:
> > > > > On Tue, 13 Mar 2018 13:28:25 +0100
> > > > > Nelio Laranjeiro <nelio.laranjeiro@6wind.com> wrote:
> > > > > > drivers/net/tap/Makefile | 3 +-
> > > > > > drivers/net/tap/rte_eth_tap.c | 16 +--
> > > > > > drivers/net/tap/tap_flow.c | 118 ++++++++++-----------
> > > > > > drivers/net/tap/tap_netlink.h | 42 --------
> > > > > > drivers/net/tap/tap_tcmsgs.c | 28 ++---
> > > > > > drivers/net/tap/tap_tcmsgs.h | 2 +-
> > > > > > lib/Makefile | 2 +
> > > > > > lib/librte_netlink/Makefile | 26 +++++
> > > > > > lib/librte_netlink/meson.build | 13 +++
> > > > > > .../librte_netlink/rte_netlink.c | 87 +++++++++++----
> > > > > > lib/librte_netlink/rte_netlink.h | 43 ++++++++
> > > > > > lib/librte_netlink/rte_netlink_version.map | 18 ++++
> > > > >
> > > > > I might have raised this before; but having yet another netlink library is
> > > > > not a great advantage. It would be much better to use a common external library
> > > > > libmnl which is already available on every distribution.
> > > >
> > > > The question is more do we really want to have a dependency on a so
> > > > small wrapper for a socket interface?
> > >
> > > Lots of drivers already have dependencies.
> > > And I trust libmnl rather than a reinvented library.
> >
> > Are we sure it is commonly installed in distributions?
> >
iproute2 now mostly requires libmnl. It is used for several networking types
and extended error reporting.
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2018-03-15 16:27 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-03-13 12:28 [PATCH 0/2] lib: move Netlink wrapper to lib Nelio Laranjeiro
2018-03-13 12:28 ` [PATCH 1/2] lib: move Netlink code into a common library Nelio Laranjeiro
2018-03-13 13:58 ` Bruce Richardson
2018-03-13 14:58 ` Nélio Laranjeiro
2018-03-13 12:28 ` [PATCH 2/2] lib: add request Netlink messages Nelio Laranjeiro
2018-03-13 21:20 ` [PATCH 0/2] lib: move Netlink wrapper to lib Stephen Hemminger
2018-03-14 12:08 ` Nélio Laranjeiro
2018-03-14 15:17 ` Stephen Hemminger
2018-03-15 15:39 ` Thomas Monjalon
2018-03-15 16:19 ` Nélio Laranjeiro
2018-03-15 16:26 ` Stephen Hemminger
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.