* [PATCH 0/9] net/cnxk: support for port representors
@ 2023-08-11 16:34 Harman Kalra
2023-08-11 16:34 ` [PATCH 1/9] common/cnxk: debug log type for representors Harman Kalra
` (13 more replies)
0 siblings, 14 replies; 142+ messages in thread
From: Harman Kalra @ 2023-08-11 16:34 UTC (permalink / raw)
To: jerinj; +Cc: dev, Harman Kalra
Introducing CNXK driver support for port representor model by adding virtual
ethernet ports providing a logical representation in DPDK for SR-IOV virtual
function (VF) devices for control and monitoring.
These port representor ethdev instances can be spawned on an as needed basis
through configuration parameters passed to the driver of the underlying
base device using devargs ``-a pci:dbdf,representor=[0]``
Harman Kalra (9):
common/cnxk: debug log type for representors
net/cnxk: probing representor ports
common/cnxk: maintaining representor state
net/cnxk: callbacks for representor state
net/cnxk: add representor control plane
net/cnxk: representor ethdev ops
net/cnxk: representor flow ops
common/cnxk: support represented port for cnxk
net/cnxk: add represented port for cnxk
doc/guides/nics/cnxk.rst | 39 ++
doc/guides/nics/features/cnxk.ini | 1 +
doc/guides/nics/features/cnxk_vec.ini | 2 +
doc/guides/nics/features/cnxk_vf.ini | 2 +
drivers/common/cnxk/roc_dev.c | 167 ++++--
drivers/common/cnxk/roc_dev_priv.h | 7 +-
drivers/common/cnxk/roc_nix.c | 23 +
drivers/common/cnxk/roc_nix.h | 22 +-
drivers/common/cnxk/roc_npc.c | 54 +-
drivers/common/cnxk/roc_npc.h | 16 +-
drivers/common/cnxk/roc_npc_mcam.c | 69 ++-
drivers/common/cnxk/roc_npc_parse.c | 14 +
drivers/common/cnxk/roc_npc_priv.h | 1 +
drivers/common/cnxk/roc_platform.c | 1 +
drivers/common/cnxk/roc_platform.h | 2 +
drivers/common/cnxk/version.map | 4 +
drivers/net/cnxk/cn10k_ethdev.c | 4 +-
drivers/net/cnxk/cn9k_ethdev.c | 4 +-
drivers/net/cnxk/cnxk_ethdev.c | 50 +-
drivers/net/cnxk/cnxk_ethdev.h | 15 +
drivers/net/cnxk/cnxk_flow.c | 188 ++++---
drivers/net/cnxk/cnxk_flow.h | 9 +-
drivers/net/cnxk/cnxk_rep.c | 340 ++++++++++++
drivers/net/cnxk/cnxk_rep.h | 108 ++++
drivers/net/cnxk/cnxk_rep_flow.c | 715 +++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.c | 559 ++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 151 ++++++
drivers/net/cnxk/cnxk_rep_ops.c | 718 ++++++++++++++++++++++++++
drivers/net/cnxk/meson.build | 4 +
29 files changed, 3122 insertions(+), 167 deletions(-)
create mode 100644 drivers/net/cnxk/cnxk_rep.c
create mode 100644 drivers/net/cnxk/cnxk_rep.h
create mode 100644 drivers/net/cnxk/cnxk_rep_flow.c
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.c
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.h
create mode 100644 drivers/net/cnxk/cnxk_rep_ops.c
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH 1/9] common/cnxk: debug log type for representors
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
@ 2023-08-11 16:34 ` Harman Kalra
2023-08-11 16:34 ` [PATCH 2/9] net/cnxk: probing representor ports Harman Kalra
` (12 subsequent siblings)
13 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-08-11 16:34 UTC (permalink / raw)
To: jerinj, Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao
Cc: dev, Harman Kalra
Registering exclusive debug log type for representors
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_platform.c | 1 +
drivers/common/cnxk/roc_platform.h | 2 ++
drivers/common/cnxk/version.map | 1 +
3 files changed, 4 insertions(+)
diff --git a/drivers/common/cnxk/roc_platform.c b/drivers/common/cnxk/roc_platform.c
index f91b95ceab..2016be8354 100644
--- a/drivers/common/cnxk/roc_platform.c
+++ b/drivers/common/cnxk/roc_platform.c
@@ -70,4 +70,5 @@ RTE_LOG_REGISTER(cnxk_logtype_npc, pmd.net.cnxk.flow, NOTICE);
RTE_LOG_REGISTER(cnxk_logtype_sso, pmd.event.cnxk, NOTICE);
RTE_LOG_REGISTER(cnxk_logtype_tim, pmd.event.cnxk.timer, NOTICE);
RTE_LOG_REGISTER(cnxk_logtype_tm, pmd.net.cnxk.tm, NOTICE);
+RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_rep, NOTICE);
RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_ree, NOTICE);
diff --git a/drivers/common/cnxk/roc_platform.h b/drivers/common/cnxk/roc_platform.h
index 9884398a99..a8077cd7bc 100644
--- a/drivers/common/cnxk/roc_platform.h
+++ b/drivers/common/cnxk/roc_platform.h
@@ -243,6 +243,7 @@ extern int cnxk_logtype_sso;
extern int cnxk_logtype_tim;
extern int cnxk_logtype_tm;
extern int cnxk_logtype_ree;
+extern int cnxk_logtype_rep;
#define plt_err(fmt, args...) \
RTE_LOG(ERR, PMD, "%s():%u " fmt "\n", __func__, __LINE__, ##args)
@@ -271,6 +272,7 @@ extern int cnxk_logtype_ree;
#define plt_tim_dbg(fmt, ...) plt_dbg(tim, fmt, ##__VA_ARGS__)
#define plt_tm_dbg(fmt, ...) plt_dbg(tm, fmt, ##__VA_ARGS__)
#define plt_ree_dbg(fmt, ...) plt_dbg(ree, fmt, ##__VA_ARGS__)
+#define plt_rep_dbg(fmt, ...) plt_dbg(rep, fmt, ##__VA_ARGS__)
/* Datapath logs */
#define plt_dp_err(fmt, args...) \
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index 8c71497df8..1d6e306848 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -13,6 +13,7 @@ INTERNAL {
cnxk_logtype_npa;
cnxk_logtype_npc;
cnxk_logtype_ree;
+ cnxk_logtype_rep;
cnxk_logtype_sso;
cnxk_logtype_tim;
cnxk_logtype_tm;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH 2/9] net/cnxk: probing representor ports
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
2023-08-11 16:34 ` [PATCH 1/9] common/cnxk: debug log type for representors Harman Kalra
@ 2023-08-11 16:34 ` Harman Kalra
2023-08-11 16:34 ` [PATCH 3/9] common/cnxk: maintaining representor state Harman Kalra
` (11 subsequent siblings)
13 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-08-11 16:34 UTC (permalink / raw)
To: jerinj, Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori,
Satha Rao, Anatoly Burakov
Cc: dev, Harman Kalra
Basic skeleton for probing representor devices. If PF device is
passed with "representor" devargs, representor ports gets probed
as a separate ethdev device.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
doc/guides/nics/cnxk.rst | 39 +++++
drivers/net/cnxk/cn10k_ethdev.c | 4 +-
drivers/net/cnxk/cn9k_ethdev.c | 4 +-
drivers/net/cnxk/cnxk_ethdev.c | 42 ++++-
drivers/net/cnxk/cnxk_ethdev.h | 12 ++
drivers/net/cnxk/cnxk_rep.c | 262 ++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep.h | 51 +++++++
drivers/net/cnxk/cnxk_rep_ops.c | 112 ++++++++++++++
drivers/net/cnxk/meson.build | 2 +
9 files changed, 516 insertions(+), 12 deletions(-)
create mode 100644 drivers/net/cnxk/cnxk_rep.c
create mode 100644 drivers/net/cnxk/cnxk_rep.h
create mode 100644 drivers/net/cnxk/cnxk_rep_ops.c
diff --git a/doc/guides/nics/cnxk.rst b/doc/guides/nics/cnxk.rst
index 9229056f6f..dd14102efa 100644
--- a/doc/guides/nics/cnxk.rst
+++ b/doc/guides/nics/cnxk.rst
@@ -37,6 +37,8 @@ Features of the CNXK Ethdev PMD are:
- Inline IPsec processing support
- Ingress meter support
- Queue based priority flow control support
+- Virtual function representors
+- Represented port pattern matching and action
Prerequisites
-------------
@@ -581,6 +583,41 @@ Runtime Config Options for inline device
With the above configuration, driver would poll for soft expiry events every
1000 usec.
+Virtual Function Representors
+-----------------------------
+
+The CNXK driver supports port representor model by adding virtual ethernet
+ports providing a logical representation in DPDK for SR-IOV virtual function
+(VF) devices for control and monitoring.
+
+These port representor ethdev instances can be spawned on an as needed basis
+through configuration parameters passed to the driver of the underlying
+base device using devargs ``-a pci:dbdf,representor=[0]``
+
+.. note::
+
+ Base device is the PF whose VFs will be represented by these representors
+
+ Above devarg parameters can be provided as a range of representor device
+ ``-a pci:dbdf,representor=[0-3]`` or a single representor device on need
+ basis ``-a pci:dbdf,representor=[0]``
+
+In case of exception path (i.e. until the flow definition is offloaded to the
+hardware), packets transmitted by the VFs shall be received by these
+representor port, while packets transmitted by representor ports shall be
+received by respective VFs.
+
+On receiving the VF traffic via these representor ports, applications holding
+these representor ports can decide to offload the traffic flow into the HW.
+Henceforth the matching traffic shall be directly steered to the respective
+VFs without being received by the application.
+
+Current virtual representor port PMD supports following operations:
+
+- Get and clear VF statistics
+- Set mac address
+- Flow operations - create, validate, destroy, query, flush, dump
+
Debugging Options
-----------------
@@ -595,3 +632,5 @@ Debugging Options
+---+------------+-------------------------------------------------------+
| 2 | NPC | --log-level='pmd\.net.cnxk\.flow,8' |
+---+------------+-------------------------------------------------------+
+ | 3 | REP | --log-level='pmd\.net.cnxk\.rep,8' |
+ +---+------------+-------------------------------------------------------+
diff --git a/drivers/net/cnxk/cn10k_ethdev.c b/drivers/net/cnxk/cn10k_ethdev.c
index 4c4acc7cf0..a6a4665af1 100644
--- a/drivers/net/cnxk/cn10k_ethdev.c
+++ b/drivers/net/cnxk/cn10k_ethdev.c
@@ -912,8 +912,8 @@ static const struct rte_pci_id cn10k_pci_nix_map[] = {
static struct rte_pci_driver cn10k_pci_nix = {
.id_table = cn10k_pci_nix_map,
- .drv_flags = RTE_PCI_DRV_NEED_MAPPING | RTE_PCI_DRV_NEED_IOVA_AS_VA |
- RTE_PCI_DRV_INTR_LSC,
+ .drv_flags = RTE_PCI_DRV_NEED_MAPPING | RTE_PCI_DRV_NEED_IOVA_AS_VA | RTE_PCI_DRV_INTR_LSC |
+ RTE_PCI_DRV_PROBE_AGAIN,
.probe = cn10k_nix_probe,
.remove = cn10k_nix_remove,
};
diff --git a/drivers/net/cnxk/cn9k_ethdev.c b/drivers/net/cnxk/cn9k_ethdev.c
index bae4dda5e2..0448d7e219 100644
--- a/drivers/net/cnxk/cn9k_ethdev.c
+++ b/drivers/net/cnxk/cn9k_ethdev.c
@@ -834,8 +834,8 @@ static const struct rte_pci_id cn9k_pci_nix_map[] = {
static struct rte_pci_driver cn9k_pci_nix = {
.id_table = cn9k_pci_nix_map,
- .drv_flags = RTE_PCI_DRV_NEED_MAPPING | RTE_PCI_DRV_NEED_IOVA_AS_VA |
- RTE_PCI_DRV_INTR_LSC,
+ .drv_flags = RTE_PCI_DRV_NEED_MAPPING | RTE_PCI_DRV_NEED_IOVA_AS_VA | RTE_PCI_DRV_INTR_LSC |
+ RTE_PCI_DRV_PROBE_AGAIN,
.probe = cn9k_nix_probe,
.remove = cn9k_nix_remove,
};
diff --git a/drivers/net/cnxk/cnxk_ethdev.c b/drivers/net/cnxk/cnxk_ethdev.c
index 4b98faa729..902e6df72d 100644
--- a/drivers/net/cnxk/cnxk_ethdev.c
+++ b/drivers/net/cnxk/cnxk_ethdev.c
@@ -2102,6 +2102,10 @@ cnxk_eth_dev_uninit(struct rte_eth_dev *eth_dev, bool reset)
rte_free(eth_dev->data->mac_addrs);
eth_dev->data->mac_addrs = NULL;
+ /* Remove representor devices associated with PF */
+ if (dev->num_reps)
+ cnxk_rep_dev_remove(eth_dev);
+
rc = roc_nix_dev_fini(nix);
/* Can be freed later by PMD if NPA LF is in use */
if (rc == -EAGAIN) {
@@ -2180,18 +2184,40 @@ cnxk_nix_remove(struct rte_pci_device *pci_dev)
int
cnxk_nix_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
{
- int rc;
+ struct rte_eth_devargs eth_da = {.nb_representor_ports = 0};
+ struct rte_eth_dev *pf_ethdev;
+ uint16_t num_rep;
+ int rc = 0;
RTE_SET_USED(pci_drv);
- rc = rte_eth_dev_pci_generic_probe(pci_dev, sizeof(struct cnxk_eth_dev),
- cnxk_eth_dev_init);
+ if (pci_dev->device.devargs) {
+ rc = rte_eth_devargs_parse(pci_dev->device.devargs->args, ð_da);
+ if (rc)
+ return rc;
+ }
+
+ num_rep = eth_da.nb_representor_ports;
+ plt_rep_dbg("nb_representor_ports = %d\n", num_rep);
- /* On error on secondary, recheck if port exists in primary or
- * in mid of detach state.
+ /* This probing API may get invoked even after first level of probe is
+ * done, as part of an application bringup(OVS-DPDK vswitchd), checking
+ * if eth_dev is allocated for the PF device
*/
- if (rte_eal_process_type() != RTE_PROC_PRIMARY && rc)
- if (!rte_eth_dev_allocated(pci_dev->device.name))
- return 0;
+ pf_ethdev = rte_eth_dev_allocated(pci_dev->device.name);
+ if (pf_ethdev == NULL) {
+ rc = rte_eth_dev_pci_generic_probe(pci_dev, sizeof(struct cnxk_eth_dev),
+ cnxk_eth_dev_init);
+ if (rc || !num_rep)
+ return rc;
+
+ pf_ethdev = rte_eth_dev_allocated(pci_dev->device.name);
+ }
+
+ if (!num_rep)
+ return rc;
+
+ rc = cnxk_rep_dev_probe(pci_dev, pf_ethdev, ð_da);
+
return rc;
}
diff --git a/drivers/net/cnxk/cnxk_ethdev.h b/drivers/net/cnxk/cnxk_ethdev.h
index ed531fb277..3896db38e1 100644
--- a/drivers/net/cnxk/cnxk_ethdev.h
+++ b/drivers/net/cnxk/cnxk_ethdev.h
@@ -22,7 +22,9 @@
#include <rte_tm_driver.h>
#include "roc_api.h"
+
#include <cnxk_ethdev_dp.h>
+#include <cnxk_rep.h>
#define CNXK_ETH_DEV_PMD_VERSION "1.0"
@@ -307,6 +309,10 @@ struct cnxk_macsec_sess {
};
TAILQ_HEAD(cnxk_macsec_sess_list, cnxk_macsec_sess);
+struct cnxk_rep_info {
+ struct rte_eth_dev *rep_eth_dev;
+};
+
struct cnxk_eth_dev {
/* ROC NIX */
struct roc_nix nix;
@@ -414,6 +420,12 @@ struct cnxk_eth_dev {
/* MCS device */
struct cnxk_mcs_dev *mcs_dev;
struct cnxk_macsec_sess_list mcs_list;
+
+ /* Port representor fields */
+ uint16_t switch_domain_id;
+ uint16_t num_reps;
+ uint16_t rep_xport_vdev;
+ struct cnxk_rep_info *rep_info;
};
struct cnxk_eth_rxq_sp {
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
new file mode 100644
index 0000000000..ebefc34ac8
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -0,0 +1,262 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+#include <cnxk_rep.h>
+
+/* CNXK platform representor dev ops */
+struct eth_dev_ops cnxk_rep_dev_ops = {
+ .dev_infos_get = cnxk_rep_dev_info_get,
+ .dev_configure = cnxk_rep_dev_configure,
+ .dev_start = cnxk_rep_dev_start,
+ .rx_queue_setup = cnxk_rep_rx_queue_setup,
+ .rx_queue_release = cnxk_rep_rx_queue_release,
+ .tx_queue_setup = cnxk_rep_tx_queue_setup,
+ .tx_queue_release = cnxk_rep_tx_queue_release,
+ .link_update = cnxk_rep_link_update,
+ .dev_close = cnxk_rep_dev_close,
+ .dev_stop = cnxk_rep_dev_stop,
+ .stats_get = cnxk_rep_stats_get,
+ .stats_reset = cnxk_rep_stats_reset,
+ .flow_ops_get = cnxk_rep_flow_ops_get
+};
+
+int
+cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev)
+{
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return 0;
+
+ plt_rep_dbg("Representor port:%d uninit", ethdev->data->port_id);
+ rte_free(ethdev->data->mac_addrs);
+ ethdev->data->mac_addrs = NULL;
+
+ return 0;
+}
+
+int
+cnxk_rep_dev_remove(struct rte_eth_dev *pf_ethdev)
+{
+ struct cnxk_eth_dev *pf_dev = cnxk_eth_pmd_priv(pf_ethdev);
+ int rc = 0;
+
+ rc = rte_eth_switch_domain_free(pf_dev->switch_domain_id);
+ if (rc)
+ plt_err("Failed to alloc switch domain: %d", rc);
+
+ return rc;
+}
+
+static int
+hotplug_rep_xport_vdev(struct cnxk_eth_dev *pf_dev)
+{
+ char rep_xport_devargs[] = CNXK_REP_XPORT_VDEV_DEVARGS;
+ char name[] = CNXK_REP_XPORT_VDEV_NAME;
+ uint16_t portid;
+ int rc = 0;
+
+ rc = rte_eth_dev_get_port_by_name(name, &portid);
+ if (rc != 0) {
+ if (rc == -ENODEV) {
+ /* rep_xport device should get added once during first PF probe */
+ rc = rte_eal_hotplug_add("vdev", name, rep_xport_devargs);
+ if (rc) {
+ plt_err("rep base hotplug failed %d", -rte_errno);
+ goto fail;
+ }
+
+ /* Get the portID of rep_xport port */
+ if (rte_eth_dev_get_port_by_name(name, &portid)) {
+ plt_err("cannot find added vdev %s", name);
+ goto free;
+ }
+ } else {
+ plt_err("cannot find added vdev %s", name);
+ goto free;
+ }
+ }
+
+ plt_rep_dbg("rep_xport vdev port %d, name %s", portid, name);
+ pf_dev->rep_xport_vdev = portid;
+
+ return 0;
+free:
+ rte_eal_hotplug_remove("vdev", name);
+fail:
+ return rc;
+}
+
+static int
+cnxk_init_rep_internal(struct cnxk_eth_dev *pf_dev)
+{
+ int rc;
+
+ if (pf_dev->rep_info)
+ return 0;
+
+ pf_dev->rep_info =
+ plt_zmalloc(sizeof(pf_dev->rep_info[0]) * CNXK_MAX_REP_PORTS, 0);
+ if (!pf_dev->rep_info) {
+ plt_err("Failed to alloc memory for rep info");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ /* Allocate switch domain for this PF */
+ rc = rte_eth_switch_domain_alloc(&pf_dev->switch_domain_id);
+ if (rc) {
+ plt_err("Failed to alloc switch domain: %d", rc);
+ goto fail;
+ }
+
+ rc = hotplug_rep_xport_vdev(pf_dev);
+ if (rc) {
+ plt_err("Failed to hotplug representor base port, err %d", rc);
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static uint16_t
+cnxk_rep_tx_burst(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(tx_queue);
+ PLT_SET_USED(tx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
+
+static uint16_t
+cnxk_rep_rx_burst(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(rx_queue);
+ PLT_SET_USED(rx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
+
+static int
+cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ struct cnxk_rep_dev *rep_params = (struct cnxk_rep_dev *)params;
+ struct rte_eth_link *link;
+ struct cnxk_eth_dev *pf_dev;
+
+ rep_dev->vf_id = rep_params->vf_id;
+ rep_dev->switch_domain_id = rep_params->switch_domain_id;
+ rep_dev->parent_dev = rep_params->parent_dev;
+
+ eth_dev->data->dev_flags |= RTE_ETH_DEV_REPRESENTOR;
+ eth_dev->data->representor_id = rep_params->vf_id;
+ eth_dev->data->backer_port_id = rep_params->parent_dev->data->port_id;
+
+ eth_dev->data->mac_addrs = plt_zmalloc(RTE_ETHER_ADDR_LEN, 0);
+ if (!eth_dev->data->mac_addrs) {
+ plt_err("Failed to allocate memory for mac addr");
+ return -ENOMEM;
+ }
+
+ rte_eth_random_addr(rep_dev->mac_addr);
+ memcpy(eth_dev->data->mac_addrs, rep_dev->mac_addr, RTE_ETHER_ADDR_LEN);
+
+ /* Set the device operations */
+ eth_dev->dev_ops = &cnxk_rep_dev_ops;
+
+ /* Rx/Tx functions stubs to avoid crashing */
+ eth_dev->rx_pkt_burst = cnxk_rep_rx_burst;
+ eth_dev->tx_pkt_burst = cnxk_rep_tx_burst;
+
+ /* Link state. Inherited from PF */
+ pf_dev = cnxk_eth_pmd_priv(rep_dev->parent_dev);
+ link = &pf_dev->eth_dev->data->dev_link;
+
+ eth_dev->data->dev_link.link_speed = link->link_speed;
+ eth_dev->data->dev_link.link_duplex = link->link_duplex;
+ eth_dev->data->dev_link.link_status = link->link_status;
+ eth_dev->data->dev_link.link_autoneg = link->link_autoneg;
+
+ return 0;
+}
+
+int
+cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct rte_eth_dev *pf_ethdev,
+ struct rte_eth_devargs *eth_da)
+{
+ char name[RTE_ETH_NAME_MAX_LEN];
+ struct rte_eth_dev *rep_eth_dev;
+ struct cnxk_eth_dev *pf_dev;
+ uint16_t num_rep;
+ int i, rc;
+
+ /* Get the PF device */
+ pf_dev = cnxk_eth_pmd_priv(pf_ethdev);
+
+ /* Check the representor devargs */
+ if (eth_da->type == RTE_ETH_REPRESENTOR_NONE)
+ return 0;
+ if (eth_da->type != RTE_ETH_REPRESENTOR_VF) {
+ plt_err("unsupported representor type %d\n", eth_da->type);
+ return -ENOTSUP;
+ }
+ num_rep = eth_da->nb_representor_ports;
+ if (num_rep > CNXK_MAX_REP_PORTS) {
+ plt_err("nb_representor_ports = %d > %d MAX VF REPS\n", num_rep,
+ CNXK_MAX_REP_PORTS);
+ return -EINVAL;
+ }
+
+ if (num_rep >= RTE_MAX_ETHPORTS) {
+ plt_err("nb_representor_ports = %d > %d MAX ETHPORTS\n", num_rep, RTE_MAX_ETHPORTS);
+ return -EINVAL;
+ }
+
+ /* Initialize the internals of representor ports */
+ if (cnxk_init_rep_internal(pf_dev))
+ return 0;
+
+ for (i = 0; i < num_rep; i++) {
+ struct cnxk_rep_dev representor = {.vf_id = eth_da->representor_ports[i],
+ .switch_domain_id = pf_dev->switch_domain_id,
+ .parent_dev = pf_ethdev};
+
+ if (representor.vf_id >= pci_dev->max_vfs) {
+ plt_err("VF-Rep id %d >= %d pci dev max vfs\n", representor.vf_id,
+ pci_dev->max_vfs);
+ continue;
+ }
+
+ /* Representor port net_bdf_port */
+ snprintf(name, sizeof(name), "net_%s_representor_%d", pci_dev->device.name,
+ eth_da->representor_ports[i]);
+
+ rc = rte_eth_dev_create(&pci_dev->device, name, sizeof(struct cnxk_rep_dev), NULL,
+ NULL, cnxk_rep_dev_init, &representor);
+ if (rc) {
+ plt_err("failed to create cnxk vf representor %s", name);
+ rc = -EINVAL;
+ goto err;
+ }
+
+ rep_eth_dev = rte_eth_dev_allocated(name);
+ if (!rep_eth_dev) {
+ plt_err("Failed to find the eth_dev for VF-Rep: %s.", name);
+ rc = -ENODEV;
+ goto err;
+ }
+
+ plt_rep_dbg("PF portid %d switch domain %d representor portid %d (%s) probe done",
+ pf_ethdev->data->port_id, pf_dev->switch_domain_id,
+ rep_eth_dev->data->port_id, name);
+ pf_dev->rep_info[representor.vf_id].rep_eth_dev = rep_eth_dev;
+ pf_dev->num_reps++;
+ }
+
+ return 0;
+err:
+ return rc;
+}
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
new file mode 100644
index 0000000000..24adb9649b
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -0,0 +1,51 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+#include <cnxk_ethdev.h>
+
+#ifndef __CNXK_REP_H__
+#define __CNXK_REP_H__
+
+#define CNXK_REP_XPORT_VDEV_DEVARGS "role=server"
+#define CNXK_REP_XPORT_VDEV_NAME "net_memif"
+#define CNXK_MAX_REP_PORTS 128
+
+/* Common ethdev ops */
+extern struct eth_dev_ops cnxk_rep_dev_ops;
+
+struct cnxk_rep_dev {
+ uint16_t vf_id;
+ uint16_t switch_domain_id;
+ struct rte_eth_dev *parent_dev;
+ uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
+};
+
+static inline struct cnxk_rep_dev *
+cnxk_rep_pmd_priv(const struct rte_eth_dev *eth_dev)
+{
+ return eth_dev->data->dev_private;
+}
+
+int cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct rte_eth_dev *pf_ethdev,
+ struct rte_eth_devargs *eth_da);
+int cnxk_rep_dev_remove(struct rte_eth_dev *pf_ethdev);
+int cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev);
+int cnxk_rep_dev_info_get(struct rte_eth_dev *eth_dev, struct rte_eth_dev_info *dev_info);
+int cnxk_rep_dev_configure(struct rte_eth_dev *eth_dev);
+
+int cnxk_rep_link_update(struct rte_eth_dev *eth_dev, int wait_to_compl);
+int cnxk_rep_dev_start(struct rte_eth_dev *eth_dev);
+int cnxk_rep_rx_queue_setup(struct rte_eth_dev *eth_dev, uint16_t queue_idx, uint16_t nb_desc,
+ unsigned int socket_id, const struct rte_eth_rxconf *rx_conf,
+ struct rte_mempool *mp);
+int cnxk_rep_tx_queue_setup(struct rte_eth_dev *eth_dev, uint16_t queue_idx, uint16_t nb_desc,
+ unsigned int socket_id, const struct rte_eth_txconf *tx_conf);
+void cnxk_rep_rx_queue_release(struct rte_eth_dev *dev, uint16_t queue_idx);
+void cnxk_rep_tx_queue_release(struct rte_eth_dev *dev, uint16_t queue_idx);
+int cnxk_rep_dev_stop(struct rte_eth_dev *eth_dev);
+int cnxk_rep_dev_close(struct rte_eth_dev *eth_dev);
+int cnxk_rep_stats_get(struct rte_eth_dev *eth_dev, struct rte_eth_stats *stats);
+int cnxk_rep_stats_reset(struct rte_eth_dev *eth_dev);
+int cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops);
+
+#endif /* __CNXK_REP_H__ */
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
new file mode 100644
index 0000000000..3f1aab077b
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -0,0 +1,112 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <cnxk_rep.h>
+
+int
+cnxk_rep_link_update(struct rte_eth_dev *ethdev, int wait_to_complete)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(wait_to_complete);
+ return 0;
+}
+
+int
+cnxk_rep_dev_info_get(struct rte_eth_dev *ethdev, struct rte_eth_dev_info *devinfo)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(devinfo);
+ return 0;
+}
+
+int
+cnxk_rep_dev_configure(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_dev_start(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_dev_close(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_dev_stop(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_rx_queue_setup(struct rte_eth_dev *ethdev, uint16_t rx_queue_id, uint16_t nb_rx_desc,
+ unsigned int socket_id, const struct rte_eth_rxconf *rx_conf,
+ struct rte_mempool *mb_pool)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(rx_queue_id);
+ PLT_SET_USED(nb_rx_desc);
+ PLT_SET_USED(socket_id);
+ PLT_SET_USED(rx_conf);
+ PLT_SET_USED(mb_pool);
+ return 0;
+}
+
+void
+cnxk_rep_rx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(queue_id);
+}
+
+int
+cnxk_rep_tx_queue_setup(struct rte_eth_dev *ethdev, uint16_t tx_queue_id, uint16_t nb_tx_desc,
+ unsigned int socket_id, const struct rte_eth_txconf *tx_conf)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(tx_queue_id);
+ PLT_SET_USED(nb_tx_desc);
+ PLT_SET_USED(socket_id);
+ PLT_SET_USED(tx_conf);
+ return 0;
+}
+
+void
+cnxk_rep_tx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(queue_id);
+}
+
+int
+cnxk_rep_stats_get(struct rte_eth_dev *ethdev, struct rte_eth_stats *stats)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(stats);
+ return 0;
+}
+
+int
+cnxk_rep_stats_reset(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(ops);
+ return 0;
+}
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index e83f3c9050..38dde54ce9 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -32,6 +32,8 @@ sources = files(
'cnxk_lookup.c',
'cnxk_ptp.c',
'cnxk_flow.c',
+ 'cnxk_rep.c',
+ 'cnxk_rep_ops.c',
'cnxk_stats.c',
'cnxk_tm.c',
)
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH 3/9] common/cnxk: maintaining representor state
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
2023-08-11 16:34 ` [PATCH 1/9] common/cnxk: debug log type for representors Harman Kalra
2023-08-11 16:34 ` [PATCH 2/9] net/cnxk: probing representor ports Harman Kalra
@ 2023-08-11 16:34 ` Harman Kalra
2023-08-11 16:34 ` [PATCH 4/9] net/cnxk: callbacks for " Harman Kalra
` (10 subsequent siblings)
13 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-08-11 16:34 UTC (permalink / raw)
To: jerinj, Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao
Cc: dev, Harman Kalra
Maintaining the state of representor which describes whether it
has an active VF and what is the func id of the representee.
Implement a mbox between VF and PF for the VF to know if
representors are available.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_dev.c | 167 +++++++++++++++++++++++------
drivers/common/cnxk/roc_dev_priv.h | 7 +-
drivers/common/cnxk/roc_nix.c | 23 ++++
drivers/common/cnxk/roc_nix.h | 22 ++--
drivers/common/cnxk/version.map | 3 +
5 files changed, 182 insertions(+), 40 deletions(-)
diff --git a/drivers/common/cnxk/roc_dev.c b/drivers/common/cnxk/roc_dev.c
index 4b0ba218ed..4057380eb7 100644
--- a/drivers/common/cnxk/roc_dev.c
+++ b/drivers/common/cnxk/roc_dev.c
@@ -216,15 +216,120 @@ af_pf_wait_msg(struct dev *dev, uint16_t vf, int num_msg)
return req_hdr->num_msgs;
}
+static int
+foward_msg_to_af(struct dev *dev, struct mbox_msghdr *msg, size_t size)
+{
+ struct mbox_msghdr *af_req;
+
+ /* Reserve AF/PF mbox message */
+ size = PLT_ALIGN(size, MBOX_MSG_ALIGN);
+ af_req = mbox_alloc_msg(dev->mbox, 0, size);
+ if (af_req == NULL)
+ return -ENOSPC;
+ mbox_req_init(msg->id, af_req);
+
+ /* Copy message from VF<->PF mbox to PF<->AF mbox */
+ mbox_memcpy((uint8_t *)af_req + sizeof(struct mbox_msghdr),
+ (uint8_t *)msg + sizeof(struct mbox_msghdr),
+ size - sizeof(struct mbox_msghdr));
+ af_req->pcifunc = msg->pcifunc;
+
+ return 0;
+}
+
+static int
+process_vf_ready_msg(struct dev *dev, struct mbox *mbox, struct mbox_msghdr *msg,
+ uint16_t vf)
+{
+ uint16_t max_bits = sizeof(dev->active_vfs[0]) * 8;
+ struct ready_msg_rsp *rsp;
+ int rc;
+
+ /* Handle READY message in PF */
+ dev->active_vfs[vf / max_bits] |= BIT_ULL(vf % max_bits);
+ rsp = (struct ready_msg_rsp *)mbox_alloc_msg(mbox, vf, sizeof(*rsp));
+ if (!rsp) {
+ plt_err("Failed to alloc VF%d READY message", vf);
+ return -1;
+ }
+
+ mbox_rsp_init(msg->id, rsp);
+
+ /* PF/VF function ID */
+ rsp->hdr.pcifunc = msg->pcifunc;
+ rsp->hdr.rc = 0;
+
+ /* Set pffunc value to its representor, op = 0 */
+ if (dev->ops && dev->ops->rep_state) {
+ rc = dev->ops->rep_state(dev->roc_nix, msg->pcifunc, 0);
+ if (rc < 0)
+ plt_err("Failed to set repr status, pcifunc 0x%x",
+ msg->pcifunc);
+ }
+
+ return 0;
+}
+
+static int
+process_vf_read_base_rule_msg(struct dev *dev, struct mbox *mbox, struct mbox_msghdr *msg,
+ uint16_t vf, size_t size, int *routed)
+{
+ struct npc_mcam_read_base_rule_rsp *rsp;
+ int route = *routed;
+ int rc = 0;
+
+ /* Check if pcifunc has representor, op = 1 */
+ if (dev->ops && dev->ops->rep_state) {
+ rc = dev->ops->rep_state(dev->roc_nix, msg->pcifunc, 1);
+ if (rc < 0) {
+ plt_err("Failed to get repr status, pcifunc 0x%x",
+ msg->pcifunc);
+ return rc;
+ }
+ }
+
+ /* If ret is 1 meaning pci func has a representor,
+ * return without forwarding base rule mbox
+ */
+ if (rc == 1) {
+ rsp = (struct npc_mcam_read_base_rule_rsp *)mbox_alloc_msg(
+ mbox, vf, sizeof(*rsp));
+ if (!rsp) {
+ plt_err("Failed to alloc VF%d rep status message", vf);
+ return -1;
+ }
+
+ mbox_rsp_init(msg->id, rsp);
+
+ /* PF/VF function ID */
+ rsp->hdr.pcifunc = msg->pcifunc;
+ rsp->hdr.rc = 0;
+ } else {
+ /* If ret is 0, default case i.e. forwarding to AF
+ * should happen.
+ */
+ rc = foward_msg_to_af(dev, msg, size);
+ if (rc) {
+ plt_err("Failed to forward msg ID %d to af, err %d",
+ msg->id, rc);
+ return rc;
+ }
+ route++;
+ }
+ *routed = route;
+
+ return 0;
+}
+
/* PF receives mbox DOWN messages from VF and forwards to AF */
static int
vf_pf_process_msgs(struct dev *dev, uint16_t vf)
{
struct mbox *mbox = &dev->mbox_vfpf;
struct mbox_dev *mdev = &mbox->dev[vf];
+ int offset, routed = 0, ret = 0;
struct mbox_hdr *req_hdr;
struct mbox_msghdr *msg;
- int offset, routed = 0;
size_t size;
uint16_t i;
@@ -242,42 +347,31 @@ vf_pf_process_msgs(struct dev *dev, uint16_t vf)
/* RVU_PF_FUNC_S */
msg->pcifunc = dev_pf_func(dev->pf, vf);
- if (msg->id == MBOX_MSG_READY) {
- struct ready_msg_rsp *rsp;
- uint16_t max_bits = sizeof(dev->active_vfs[0]) * 8;
-
- /* Handle READY message in PF */
- dev->active_vfs[vf / max_bits] |=
- BIT_ULL(vf % max_bits);
- rsp = (struct ready_msg_rsp *)mbox_alloc_msg(
- mbox, vf, sizeof(*rsp));
- if (!rsp) {
- plt_err("Failed to alloc VF%d READY message",
- vf);
+ switch (msg->id) {
+ case MBOX_MSG_READY:
+ ret = process_vf_ready_msg(dev, mbox, msg, vf);
+ if (ret) {
+ plt_err("Failed to process ready msg for vf %d", vf);
continue;
}
- mbox_rsp_init(msg->id, rsp);
+ break;
+ case MBOX_MSG_NPC_MCAM_READ_BASE_RULE:
+ ret = process_vf_read_base_rule_msg(dev, mbox, msg, vf, size, &routed);
+ if (ret) {
+ plt_err("Failed to process base rule for vf %d, err %d", vf, ret);
+ continue;
+ }
- /* PF/VF function ID */
- rsp->hdr.pcifunc = msg->pcifunc;
- rsp->hdr.rc = 0;
- } else {
- struct mbox_msghdr *af_req;
- /* Reserve AF/PF mbox message */
- size = PLT_ALIGN(size, MBOX_MSG_ALIGN);
- af_req = mbox_alloc_msg(dev->mbox, 0, size);
- if (af_req == NULL)
- return -ENOSPC;
- mbox_req_init(msg->id, af_req);
-
- /* Copy message from VF<->PF mbox to PF<->AF mbox */
- mbox_memcpy((uint8_t *)af_req +
- sizeof(struct mbox_msghdr),
- (uint8_t *)msg + sizeof(struct mbox_msghdr),
- size - sizeof(struct mbox_msghdr));
- af_req->pcifunc = msg->pcifunc;
+ break;
+ default: {
+ ret = foward_msg_to_af(dev, msg, size);
+ if (ret) {
+ plt_err("Failed to forward msg ID %d to af, err %d", msg->id, ret);
+ return ret;
+ }
routed++;
+ } break;
}
offset = mbox->rx_start + msg->next_msgoff;
}
@@ -1051,6 +1145,7 @@ vf_flr_handle_msg(void *param, dev_intr_t *flr)
{
uint16_t vf, max_vf, max_bits;
struct dev *dev = param;
+ int ret;
max_bits = sizeof(flr->bits[0]) * sizeof(uint64_t);
max_vf = max_bits * MAX_VFPF_DWORD_BITS;
@@ -1063,6 +1158,14 @@ vf_flr_handle_msg(void *param, dev_intr_t *flr)
vf_flr_send_msg(dev, vf);
flr->bits[vf / max_bits] &= ~(BIT_ULL(vf % max_bits));
+ /* Reset VF representors state, op = 2 */
+ if (dev->ops && dev->ops->rep_state) {
+ ret = dev->ops->rep_state(dev->roc_nix, dev_pf_func(dev->pf, vf),
+ 2);
+ if (ret < 0)
+ plt_err("Failed to set repr status, for vf %x", vf);
+ }
+
/* Signal FLR finish */
plt_write64(BIT_ULL(vf % max_bits),
dev->bar2 + RVU_PF_VFTRPENDX(vf / max_bits));
diff --git a/drivers/common/cnxk/roc_dev_priv.h b/drivers/common/cnxk/roc_dev_priv.h
index 1f84f74ff3..50a7a67d42 100644
--- a/drivers/common/cnxk/roc_dev_priv.h
+++ b/drivers/common/cnxk/roc_dev_priv.h
@@ -34,14 +34,17 @@ typedef int (*ptp_info_t)(void *roc_nix, bool enable);
typedef void (*q_err_cb_t)(void *roc_nix, void *data);
/* Link status get callback */
-typedef void (*link_status_get_t)(void *roc_nix,
- struct cgx_link_user_info *link);
+typedef void (*link_status_get_t)(void *roc_nix, struct cgx_link_user_info *link);
+
+/* Process representor status callback */
+typedef int (*rep_state_t)(void *roc_nix, uint16_t pf_func, uint8_t op);
struct dev_ops {
link_info_t link_status_update;
ptp_info_t ptp_info_update;
link_status_get_t link_status_get;
q_err_cb_t q_err_cb;
+ rep_state_t rep_state;
};
#define dev_is_vf(dev) ((dev)->hwcap & DEV_HWCAP_F_VF)
diff --git a/drivers/common/cnxk/roc_nix.c b/drivers/common/cnxk/roc_nix.c
index 152ef7269e..0ee534f188 100644
--- a/drivers/common/cnxk/roc_nix.c
+++ b/drivers/common/cnxk/roc_nix.c
@@ -522,3 +522,26 @@ roc_nix_dev_fini(struct roc_nix *roc_nix)
rc |= dev_fini(&nix->dev, nix->pci_dev);
return rc;
}
+
+int
+roc_nix_process_rep_state_cb_register(struct roc_nix *roc_nix,
+ process_rep_state_t proc_rep_st)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+
+ if (proc_rep_st == NULL)
+ return NIX_ERR_PARAM;
+
+ dev->ops->rep_state = (rep_state_t)proc_rep_st;
+ return 0;
+}
+
+void
+roc_nix_process_rep_state_cb_unregister(struct roc_nix *roc_nix)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+
+ dev->ops->rep_state = NULL;
+}
diff --git a/drivers/common/cnxk/roc_nix.h b/drivers/common/cnxk/roc_nix.h
index 9c2ba9a685..47ab3560ea 100644
--- a/drivers/common/cnxk/roc_nix.h
+++ b/drivers/common/cnxk/roc_nix.h
@@ -443,8 +443,14 @@ typedef int (*ptp_info_update_t)(struct roc_nix *roc_nix, bool enable);
typedef void (*q_err_get_t)(struct roc_nix *roc_nix, void *data);
/* Link status get callback */
-typedef void (*link_info_get_t)(struct roc_nix *roc_nix,
- struct roc_nix_link_info *link);
+typedef void (*link_info_get_t)(struct roc_nix *roc_nix, struct roc_nix_link_info *link);
+
+/* Process representor status callback:
+ * op = 0 update pffunc of vf being represented
+ * op = 1 check if any representor is representing pffunc
+ * op = 2 vf is going down, reset rep state
+ */
+typedef int (*process_rep_state_t)(void *roc_nix, uint16_t pf_func, uint8_t op);
TAILQ_HEAD(roc_nix_list, roc_nix);
@@ -520,6 +526,9 @@ roc_nix_tm_max_shaper_burst_get(void)
/* Dev */
int __roc_api roc_nix_dev_init(struct roc_nix *roc_nix);
int __roc_api roc_nix_dev_fini(struct roc_nix *roc_nix);
+int __roc_api roc_nix_process_rep_state_cb_register(struct roc_nix *roc_nix,
+ process_rep_state_t proc_rep_st);
+void __roc_api roc_nix_process_rep_state_cb_unregister(struct roc_nix *roc_nix);
/* Type */
bool __roc_api roc_nix_is_lbk(struct roc_nix *roc_nix);
@@ -532,13 +541,14 @@ int __roc_api roc_nix_get_vf(struct roc_nix *roc_nix);
uint16_t __roc_api roc_nix_get_pf_func(struct roc_nix *roc_nix);
uint16_t __roc_api roc_nix_get_vwqe_interval(struct roc_nix *roc_nix);
int __roc_api roc_nix_max_pkt_len(struct roc_nix *roc_nix);
+bool __roc_api roc_nix_has_rep(struct roc_nix *roc_nix);
/* LF ops */
-int __roc_api roc_nix_lf_alloc(struct roc_nix *roc_nix, uint32_t nb_rxq,
- uint32_t nb_txq, uint64_t rx_cfg);
+int __roc_api roc_nix_lf_alloc(struct roc_nix *roc_nix, uint32_t nb_rxq, uint32_t nb_txq,
+ uint64_t rx_cfg);
int __roc_api roc_nix_lf_free(struct roc_nix *roc_nix);
-int __roc_api roc_nix_lf_inl_ipsec_cfg(struct roc_nix *roc_nix,
- struct roc_nix_ipsec_cfg *cfg, bool enb);
+int __roc_api roc_nix_lf_inl_ipsec_cfg(struct roc_nix *roc_nix, struct roc_nix_ipsec_cfg *cfg,
+ bool enb);
int __roc_api roc_nix_cpt_ctx_cache_sync(struct roc_nix *roc_nix);
int __roc_api roc_nix_rx_drop_re_set(struct roc_nix *roc_nix, bool ena);
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index 1d6e306848..327840429f 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -205,6 +205,8 @@ INTERNAL {
roc_nix_cqe_dump;
roc_nix_dev_fini;
roc_nix_dev_init;
+ roc_nix_process_rep_state_cb_register;
+ roc_nix_process_rep_state_cb_unregister;
roc_nix_dump;
roc_nix_err_intr_ena_dis;
roc_nix_fc_config_get;
@@ -217,6 +219,7 @@ INTERNAL {
roc_nix_get_pf_func;
roc_nix_get_vf;
roc_nix_get_vwqe_interval;
+ roc_nix_has_rep;
roc_nix_inl_cb_register;
roc_nix_inl_cb_unregister;
roc_nix_inl_ctx_write;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH 4/9] net/cnxk: callbacks for representor state
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
` (2 preceding siblings ...)
2023-08-11 16:34 ` [PATCH 3/9] common/cnxk: maintaining representor state Harman Kalra
@ 2023-08-11 16:34 ` Harman Kalra
2023-08-11 16:34 ` [PATCH 5/9] net/cnxk: add representor control plane Harman Kalra
` (9 subsequent siblings)
13 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-08-11 16:34 UTC (permalink / raw)
To: jerinj, Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao
Cc: dev, Harman Kalra
Implementing the callbacks for processing representor state.
Three operations currently supported:
- set a representor to be active if its VF is enabled and set
it appropriate pf func value.
- check if the VF which sent a mbox has a representor
- clear representor state if its VF goes down.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_rep.c | 65 +++++++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep.h | 4 +++
2 files changed, 69 insertions(+)
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
index ebefc34ac8..4dd564058c 100644
--- a/drivers/net/cnxk/cnxk_rep.c
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -39,6 +39,7 @@ cnxk_rep_dev_remove(struct rte_eth_dev *pf_ethdev)
struct cnxk_eth_dev *pf_dev = cnxk_eth_pmd_priv(pf_ethdev);
int rc = 0;
+ roc_nix_process_rep_state_cb_unregister(&pf_dev->nix);
rc = rte_eth_switch_domain_free(pf_dev->switch_domain_id);
if (rc)
plt_err("Failed to alloc switch domain: %d", rc);
@@ -183,6 +184,63 @@ cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
return 0;
}
+static int
+cnxk_process_representor_status(void *roc_nix, uint16_t pf_func, uint8_t op)
+{
+ struct cnxk_eth_dev *pf_dev = (struct cnxk_eth_dev *)roc_nix;
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct rte_eth_dev *rep_eth_dev;
+ uint16_t match = 0, func_val;
+ bool is_vf_active;
+ int i, rc = 0;
+
+ if (!pf_dev) {
+ plt_err("Failed to get PF ethdev handle");
+ return -1;
+ }
+
+ switch (op) {
+ case 0: /* update pffunc of vf being represented */
+ match = 0;
+ func_val = pf_func;
+ is_vf_active = true;
+ break;
+ case 1: /* check if any representor is representing pffunc */
+ match = pf_func;
+ func_val = pf_func;
+ is_vf_active = true;
+ break;
+ case 2: /* vf is going down, reset rep state */
+ match = pf_func;
+ func_val = 0;
+ is_vf_active = false;
+ break;
+ default:
+ plt_err("Invalid op received %d pf_func %x", op, pf_func);
+ return -1;
+ };
+
+ for (i = 0; i < pf_dev->num_reps; i++) {
+ rep_eth_dev = pf_dev->rep_info[i].rep_eth_dev;
+ if (!rep_eth_dev) {
+ plt_err("Failed to get rep ethdev handle");
+ return -1;
+ }
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ if (rep_dev->pf_func == match) {
+ plt_base_dbg("Representor port %d op %d match %d func_val %d vf_active %d",
+ i, op, match, func_val, is_vf_active);
+ rep_dev->pf_func = func_val;
+ rep_dev->is_vf_active = is_vf_active;
+ rc = 1;
+ break;
+ }
+ }
+
+ return rc;
+}
+
int
cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct rte_eth_dev *pf_ethdev,
struct rte_eth_devargs *eth_da)
@@ -256,6 +314,13 @@ cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct rte_eth_dev *pf_ethdev
pf_dev->num_reps++;
}
+ /* Register up msg callbacks for processing representor information */
+ if (roc_nix_process_rep_state_cb_register(&pf_dev->nix, cnxk_process_representor_status)) {
+ plt_err("Failed to register callback for representor status");
+ rc = -EINVAL;
+ goto err;
+ }
+
return 0;
err:
return rc;
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index 24adb9649b..e3fc717a58 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -17,6 +17,10 @@ struct cnxk_rep_dev {
uint16_t vf_id;
uint16_t switch_domain_id;
struct rte_eth_dev *parent_dev;
+ struct rte_mempool *ctrl_chan_pool;
+ uint16_t rep_xport_vdev;
+ bool is_vf_active;
+ uint16_t pf_func;
uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
};
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH 5/9] net/cnxk: add representor control plane
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
` (3 preceding siblings ...)
2023-08-11 16:34 ` [PATCH 4/9] net/cnxk: callbacks for " Harman Kalra
@ 2023-08-11 16:34 ` Harman Kalra
2023-08-11 16:34 ` [PATCH 6/9] net/cnxk: representor ethdev ops Harman Kalra
` (8 subsequent siblings)
13 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-08-11 16:34 UTC (permalink / raw)
To: jerinj, Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao
Cc: dev, Harman Kalra
Implementing the control path for representor ports, where represented
ports can be configured using TLV messaging.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_ethdev.c | 8 +
drivers/net/cnxk/cnxk_ethdev.h | 3 +
drivers/net/cnxk/cnxk_rep.c | 13 +-
drivers/net/cnxk/cnxk_rep.h | 1 +
drivers/net/cnxk/cnxk_rep_msg.c | 559 ++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 78 +++++
drivers/net/cnxk/meson.build | 1 +
7 files changed, 662 insertions(+), 1 deletion(-)
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.c
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.h
diff --git a/drivers/net/cnxk/cnxk_ethdev.c b/drivers/net/cnxk/cnxk_ethdev.c
index 902e6df72d..a63c020c0e 100644
--- a/drivers/net/cnxk/cnxk_ethdev.c
+++ b/drivers/net/cnxk/cnxk_ethdev.c
@@ -1645,6 +1645,14 @@ cnxk_nix_dev_stop(struct rte_eth_dev *eth_dev)
memset(&link, 0, sizeof(link));
rte_eth_linkstatus_set(eth_dev, &link);
+ /* Exiting the rep msg ctrl thread */
+ if (dev->num_reps) {
+ if (dev->start_rep_thread) {
+ dev->start_rep_thread = false;
+ pthread_join(dev->rep_ctrl_msg_thread, NULL);
+ }
+ }
+
return 0;
}
diff --git a/drivers/net/cnxk/cnxk_ethdev.h b/drivers/net/cnxk/cnxk_ethdev.h
index 3896db38e1..0a1a4e377d 100644
--- a/drivers/net/cnxk/cnxk_ethdev.h
+++ b/drivers/net/cnxk/cnxk_ethdev.h
@@ -425,6 +425,9 @@ struct cnxk_eth_dev {
uint16_t switch_domain_id;
uint16_t num_reps;
uint16_t rep_xport_vdev;
+ rte_spinlock_t rep_lock;
+ bool start_rep_thread;
+ pthread_t rep_ctrl_msg_thread;
struct cnxk_rep_info *rep_info;
};
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
index 4dd564058c..e6f5790adc 100644
--- a/drivers/net/cnxk/cnxk_rep.c
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -2,6 +2,7 @@
* Copyright(C) 2023 Marvell.
*/
#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
/* CNXK platform representor dev ops */
struct eth_dev_ops cnxk_rep_dev_ops = {
@@ -203,7 +204,7 @@ cnxk_process_representor_status(void *roc_nix, uint16_t pf_func, uint8_t op)
case 0: /* update pffunc of vf being represented */
match = 0;
func_val = pf_func;
- is_vf_active = true;
+ is_vf_active = false;
break;
case 1: /* check if any representor is representing pffunc */
match = pf_func;
@@ -314,6 +315,9 @@ cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct rte_eth_dev *pf_ethdev
pf_dev->num_reps++;
}
+ /* Spinlock for synchronization between the control messages */
+ plt_spinlock_init(&pf_dev->rep_lock);
+
/* Register up msg callbacks for processing representor information */
if (roc_nix_process_rep_state_cb_register(&pf_dev->nix, cnxk_process_representor_status)) {
plt_err("Failed to register callback for representor status");
@@ -321,6 +325,13 @@ cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct rte_eth_dev *pf_ethdev
goto err;
}
+ /* Launch a thread to handle control messages */
+ rc = cnxk_rep_control_thread_launch(pf_dev);
+ if (rc) {
+ plt_err("Failed to launch message ctrl thread");
+ goto err;
+ }
+
return 0;
err:
return rc;
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index e3fc717a58..8825fa1cf2 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -8,6 +8,7 @@
#define CNXK_REP_XPORT_VDEV_DEVARGS "role=server"
#define CNXK_REP_XPORT_VDEV_NAME "net_memif"
+#define CNXK_REP_VDEV_CTRL_QUEUE 0
#define CNXK_MAX_REP_PORTS 128
/* Common ethdev ops */
diff --git a/drivers/net/cnxk/cnxk_rep_msg.c b/drivers/net/cnxk/cnxk_rep_msg.c
new file mode 100644
index 0000000000..ca3b6b014e
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_msg.c
@@ -0,0 +1,559 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
+
+#define CTRL_MSG_RCV_TIMEOUT_MS 2000
+#define CTRL_MSG_READY_WAIT_US 2000
+#define CTRL_MSG_THRD_NAME_LEN 35
+#define CTRL_MSG_BUFFER_SZ 1500
+#define CTRL_MSG_SIGNATURE 0xcdacdeadbeefcadc
+
+static int
+send_message(void *buffer, size_t len, struct rte_mempool *mb_pool, uint16_t portid)
+{
+ struct rte_mbuf *m = NULL;
+ uint8_t nb_pkt;
+ int rc = 0;
+ char *data;
+
+ m = rte_pktmbuf_alloc(mb_pool);
+ if (m == NULL) {
+ plt_err("Cannot allocate mbuf");
+ rc = -rte_errno;
+ goto fail;
+ }
+
+ if (rte_pktmbuf_pkt_len(m) != 0) {
+ plt_err("Bad length");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* append data */
+ data = rte_pktmbuf_append(m, len);
+ if (data == NULL) {
+ plt_err("Cannot append data\n");
+ rc = -EINVAL;
+ goto fail;
+ }
+ if (rte_pktmbuf_pkt_len(m) != len) {
+ plt_err("Bad pkt length\n");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ if (rte_pktmbuf_data_len(m) != len) {
+ plt_err("Bad data length\n");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ rte_memcpy(data, buffer, len);
+
+ /* Send the control message */
+ nb_pkt = rte_eth_tx_burst(portid, CNXK_REP_VDEV_CTRL_QUEUE, (struct rte_mbuf **)&m, 1);
+ if (nb_pkt == 0) {
+ plt_err("Failed to send message");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+void
+cnxk_rep_msg_populate_msg_end(void *buffer, uint32_t *length)
+{
+ cnxk_rep_msg_populate_command(buffer, length, CNXK_REP_MSG_END, 0);
+}
+
+void
+cnxk_rep_msg_populate_type(void *buffer, uint32_t *length, cnxk_type_t type, uint32_t sz)
+{
+ uint32_t len = *length;
+ cnxk_type_data_t data;
+
+ /* Prepare type data */
+ data.type = type;
+ data.length = sz;
+
+ /* Populate the type data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), &data, sizeof(cnxk_type_data_t));
+ len += sizeof(cnxk_type_data_t);
+
+ *length = len;
+}
+
+void
+cnxk_rep_msg_populate_header(void *buffer, uint32_t *length)
+{
+ cnxk_header_t hdr;
+ int len;
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_HEADER, sizeof(cnxk_header_t));
+
+ len = *length;
+ /* Prepare header data */
+ hdr.signature = CTRL_MSG_SIGNATURE;
+
+ /* Populate header data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), &hdr, sizeof(cnxk_header_t));
+ len += sizeof(cnxk_header_t);
+
+ *length = len;
+}
+
+void
+cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type, uint32_t size)
+{
+ cnxk_rep_msg_data_t msg_data;
+ uint32_t len;
+ uint16_t sz = sizeof(cnxk_rep_msg_data_t);
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_MSG, sz);
+
+ len = *length;
+ /* Prepare command data */
+ msg_data.type = type;
+ msg_data.length = size;
+
+ /* Populate the command */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), &msg_data, sz);
+ len += sz;
+
+ *length = len;
+}
+
+void
+cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
+ cnxk_rep_msg_t msg)
+{
+ uint32_t len;
+
+ cnxk_rep_msg_populate_command(buffer, length, msg, sz);
+
+ len = *length;
+ /* Populate command data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), msg_meta, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static int
+parse_validate_header(void *msg_buf, uint32_t *buf_trav_len)
+{
+ cnxk_type_data_t *tdata = NULL;
+ cnxk_header_t *hdr = NULL;
+ void *data = NULL;
+ uint16_t len = 0;
+
+ /* Read first bytes of type data */
+ data = msg_buf;
+ tdata = (cnxk_type_data_t *)data;
+ if (tdata->type != CNXK_TYPE_HEADER) {
+ plt_err("Invalid type %d, type header expected", tdata->type);
+ goto fail;
+ }
+
+ /* Get the header value */
+ data = RTE_PTR_ADD(msg_buf, sizeof(cnxk_type_data_t));
+ len += sizeof(cnxk_type_data_t);
+
+ /* Validate the header */
+ hdr = (cnxk_header_t *)data;
+ if (hdr->signature != CTRL_MSG_SIGNATURE) {
+ plt_err("Invalid signature detected: 0x%lx", hdr->signature);
+ goto fail;
+ }
+
+ /* Update length read till point */
+ len += tdata->length;
+
+ *buf_trav_len = len;
+ return 0;
+fail:
+ return errno;
+}
+
+static cnxk_rep_msg_data_t *
+message_data_extract(void *msg_buf, uint32_t *buf_trav_len)
+{
+ cnxk_type_data_t *tdata = NULL;
+ cnxk_rep_msg_data_t *msg = NULL;
+ uint16_t len = *buf_trav_len;
+ void *data;
+
+ tdata = (cnxk_type_data_t *)RTE_PTR_ADD(msg_buf, len);
+ if (tdata->type != CNXK_TYPE_MSG) {
+ plt_err("Invalid type %d, type MSG expected", tdata->type);
+ goto fail;
+ }
+
+ /* Get the message type */
+ len += sizeof(cnxk_type_data_t);
+ data = RTE_PTR_ADD(msg_buf, len);
+ msg = (cnxk_rep_msg_data_t *)data;
+
+ /* Advance to actual message data */
+ len += tdata->length;
+ *buf_trav_len = len;
+
+ return msg;
+fail:
+ return NULL;
+}
+
+static void
+process_ack_message(void *msg_buf, uint32_t *buf_trav_len, uint32_t msg_len, void *data)
+{
+ cnxk_rep_msg_ack_data_t *adata = (cnxk_rep_msg_ack_data_t *)data;
+ uint16_t len = *buf_trav_len;
+ void *buf;
+
+ /* Get the message type data viz ack data */
+ buf = RTE_PTR_ADD(msg_buf, len);
+ adata->u.data = rte_zmalloc("Ack data", msg_len, 0);
+ adata->size = msg_len;
+ if (adata->size == sizeof(uint64_t))
+ rte_memcpy(&adata->u.data, buf, msg_len);
+ else
+ rte_memcpy(adata->u.data, buf, msg_len);
+ plt_rep_dbg("Address %p val 0x%lx sval %ld msg_len %d", adata->u.data, adata->u.val,
+ adata->u.sval, msg_len);
+
+ /* Advance length to nex message */
+ len += msg_len;
+ *buf_trav_len = len;
+}
+
+static int
+notify_rep_dev_ready(void *data, bool state)
+{
+ struct cnxk_eth_dev *pf_dev = (struct cnxk_eth_dev *)data;
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct rte_eth_dev *rep_eth_dev;
+ int i;
+
+ for (i = 0; i < pf_dev->num_reps; i++) {
+ rep_eth_dev = pf_dev->rep_info[i].rep_eth_dev;
+ if (!rep_eth_dev)
+ continue;
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ rep_dev->is_vf_active = state;
+ }
+
+ return 0;
+}
+
+static void
+process_ready_message(void *msg_buf, uint32_t *buf_trav_len, uint32_t msg_len, void *data)
+{
+ cnxk_rep_msg_ready_data_t *rdata = NULL;
+ uint16_t len = *buf_trav_len;
+ void *buf;
+
+ /* Get the message type data viz ready data */
+ buf = RTE_PTR_ADD(msg_buf, len);
+ rdata = (cnxk_rep_msg_ready_data_t *)buf;
+
+ plt_rep_dbg("Ready data received %d", rdata->val);
+
+ /* Wait required to ensure other side ready for recieving the ack */
+ usleep(CTRL_MSG_READY_WAIT_US);
+ /* Update all representor about ready message */
+ if (rdata->val)
+ notify_rep_dev_ready(data, true);
+
+ /* Advance length to nex message */
+ len += msg_len;
+ *buf_trav_len = len;
+}
+
+static void
+process_exit_message(void *msg_buf, uint32_t *buf_trav_len, uint32_t msg_len, void *data)
+{
+ cnxk_rep_msg_exit_data_t *edata = NULL;
+ uint16_t len = *buf_trav_len;
+ void *buf;
+
+ /* Get the message type data viz exit data */
+ buf = RTE_PTR_ADD(msg_buf, len);
+ edata = (cnxk_rep_msg_exit_data_t *)buf;
+
+ plt_rep_dbg("Exit data received %d", edata->val);
+
+ /* Update all representor about ready/exit message */
+ if (edata->val)
+ notify_rep_dev_ready(data, false);
+
+ /* Advance length to nex message */
+ len += msg_len;
+ *buf_trav_len = len;
+}
+
+static void
+populate_ack_msg(void *buffer, uint32_t *length, cnxk_rep_msg_ack_data_t *adata)
+{
+ uint32_t sz = sizeof(cnxk_rep_msg_ack_data_t);
+ uint32_t len;
+
+ cnxk_rep_msg_populate_command(buffer, length, CNXK_REP_MSG_ACK, sz);
+
+ len = *length;
+
+ rte_memcpy(RTE_PTR_ADD(buffer, len), adata, sz);
+
+ len += sz;
+
+ *length = len;
+}
+
+static int
+send_ack_message(cnxk_rep_msg_ack_data_t *adata, struct rte_mempool *mb_pool, uint16_t portid)
+{
+ uint32_t len = 0, size;
+ void *buffer;
+ int rc = 0;
+
+ /* Allocate memory for preparing a message */
+ size = CTRL_MSG_BUFFER_SZ;
+ buffer = rte_zmalloc("ACK msg", size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ return -ENOMEM;
+ }
+
+ /* Prepare the ACK message */
+ cnxk_rep_msg_populate_header(buffer, &len);
+ populate_ack_msg(buffer, &len, adata);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ /* Send it to the peer */
+ rc = send_message(buffer, len, mb_pool, portid);
+ if (rc)
+ plt_err("Failed send ack");
+
+ return rc;
+}
+
+static int
+process_message(void *msg_buf, uint32_t *buf_trav_len, void *data, struct rte_mempool *mb_pool,
+ uint16_t portid)
+{
+ cnxk_rep_msg_data_t *msg = NULL;
+ cnxk_rep_msg_ack_data_t adata;
+ bool send_ack;
+ int rc = 0;
+
+ /* Get the message data */
+ msg = message_data_extract(msg_buf, buf_trav_len);
+ if (!msg) {
+ plt_err("Failed to get message data");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Different message type processing */
+ while (msg->type != CNXK_REP_MSG_END) {
+ send_ack = true;
+ switch (msg->type) {
+ case CNXK_REP_MSG_ACK:
+ process_ack_message(msg_buf, buf_trav_len, msg->length, data);
+ send_ack = false;
+ break;
+ case CNXK_REP_MSG_READY:
+ process_ready_message(msg_buf, buf_trav_len, msg->length, data);
+ adata.type = CNXK_REP_MSG_READY;
+ adata.u.val = 0;
+ adata.size = sizeof(uint64_t);
+ break;
+ case CNXK_REP_MSG_EXIT:
+ process_exit_message(msg_buf, buf_trav_len, msg->length, data);
+ adata.type = CNXK_REP_MSG_EXIT;
+ adata.u.val = 0;
+ adata.size = sizeof(uint64_t);
+ break;
+ default:
+ plt_err("Invalid message type: %d", msg->type);
+ rc = -EINVAL;
+ };
+
+ /* Send ACK */
+ if (send_ack)
+ send_ack_message(&adata, mb_pool, portid);
+
+ /* Advance to next message */
+ msg = message_data_extract(msg_buf, buf_trav_len);
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+process_control_packet(struct rte_mbuf *mbuf, void *data, uint16_t portid)
+{
+ uint32_t len = mbuf->data_len;
+ uint32_t buf_trav_len = 0;
+ void *msg_buf;
+ int rc;
+
+ msg_buf = plt_zmalloc(len, 0);
+ if (!msg_buf) {
+ plt_err("Failed to allocate mem for msg_buf");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ /* Extract the packet data which contains the message */
+ rte_memcpy(msg_buf, rte_pktmbuf_mtod(mbuf, void *), len);
+
+ /* Validate the validity of the received message */
+ parse_validate_header(msg_buf, &buf_trav_len);
+
+ /* Detect message and process */
+ rc = process_message(msg_buf, &buf_trav_len, data, mbuf->pool, portid);
+ if (rc) {
+ plt_err("Failed to process message");
+ goto fail;
+ }
+
+ /* Ensuring entire message has been processed */
+ if (len != buf_trav_len) {
+ plt_err("Out of %d bytes %d bytes of msg_buf processed", len, buf_trav_len);
+ rc = -EFAULT;
+ goto fail;
+ }
+
+ rte_free(msg_buf);
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+receive_control_msg_resp(uint16_t portid, void *data)
+{
+ uint32_t wait_us = CTRL_MSG_RCV_TIMEOUT_MS * 1000;
+ uint32_t timeout = 0, sleep = 1;
+ struct rte_mbuf *m = NULL;
+ uint8_t rx = 0;
+ int rc = -1;
+
+ do {
+ rx = rte_eth_rx_burst(portid, CNXK_REP_VDEV_CTRL_QUEUE, &m, 1);
+ if (rx != 0)
+ break;
+
+ /* Timeout after CTRL_MSG_RCV_TIMEOUT_MS */
+ if (timeout >= wait_us) {
+ plt_err("Control message wait timedout");
+ return -ETIMEDOUT;
+ }
+
+ plt_delay_us(sleep);
+ timeout += sleep;
+ } while ((rx == 0) || (timeout < wait_us));
+
+ if (rx) {
+ rc = process_control_packet(m, data, portid);
+ /* Freeing the allocated buffer */
+ rte_pktmbuf_free(m);
+ }
+
+ return rc;
+}
+
+int
+cnxk_rep_msg_send_process(struct cnxk_rep_dev *rep_dev, void *buffer, uint32_t len,
+ cnxk_rep_msg_ack_data_t *adata)
+{
+ struct cnxk_eth_dev *pf_dev;
+ int rc = 0;
+
+ pf_dev = cnxk_eth_pmd_priv(rep_dev->parent_dev);
+ if (!pf_dev) {
+ plt_err("Failed to get parent pf handle");
+ rc = -1;
+ goto fail;
+ }
+
+ plt_spinlock_lock(&pf_dev->rep_lock);
+ rc = send_message(buffer, len, rep_dev->ctrl_chan_pool, rep_dev->rep_xport_vdev);
+ if (rc) {
+ plt_err("Failed to send the message, err %d", rc);
+ goto free;
+ }
+
+ rc = receive_control_msg_resp(rep_dev->rep_xport_vdev, adata);
+ if (rc) {
+ plt_err("Failed to receive the response, err %d", rc);
+ goto free;
+ }
+ plt_spinlock_unlock(&pf_dev->rep_lock);
+
+ return 0;
+free:
+ plt_spinlock_unlock(&pf_dev->rep_lock);
+fail:
+ return rc;
+}
+
+static void
+poll_for_control_msg(void *data)
+{
+ struct cnxk_eth_dev *pf_dev = (struct cnxk_eth_dev *)data;
+ uint16_t portid = pf_dev->rep_xport_vdev;
+ struct rte_mbuf *m = NULL;
+ uint8_t rx = 0;
+
+ do {
+ rx = rte_eth_rx_burst(portid, CNXK_REP_VDEV_CTRL_QUEUE, &m, 1);
+ if (rx != 0)
+ break;
+ } while (rx == 0 && pf_dev->start_rep_thread);
+
+ if (rx) {
+ plt_spinlock_lock(&pf_dev->rep_lock);
+ process_control_packet(m, data, portid);
+ /* Freeing the allocated buffer */
+ rte_pktmbuf_free(m);
+ plt_spinlock_unlock(&pf_dev->rep_lock);
+ }
+}
+
+static void *
+rep_ctrl_msg_thread_main(void *arg)
+{
+ struct cnxk_eth_dev *pf_dev = (struct cnxk_eth_dev *)arg;
+
+ while (pf_dev->start_rep_thread)
+ poll_for_control_msg(pf_dev);
+
+ return NULL;
+}
+
+int
+cnxk_rep_control_thread_launch(struct cnxk_eth_dev *pf_dev)
+{
+ char name[CTRL_MSG_THRD_NAME_LEN];
+ int rc = 0;
+
+ rte_strscpy(name, "rep_ctrl_msg_hndlr", CTRL_MSG_THRD_NAME_LEN);
+ pf_dev->start_rep_thread = true;
+ rc = plt_ctrl_thread_create(&pf_dev->rep_ctrl_msg_thread, name, NULL,
+ rep_ctrl_msg_thread_main, pf_dev);
+ if (rc != 0)
+ plt_err("Failed to create rep control message handling");
+
+ return rc;
+}
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
new file mode 100644
index 0000000000..a28c63f762
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -0,0 +1,78 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#ifndef __CNXK_REP_MSG_H__
+#define __CNXK_REP_MSG_H__
+
+#include <stdint.h>
+
+#define CNXK_REP_MSG_MAX_BUFFER_SZ 1500
+
+typedef enum CNXK_TYPE {
+ CNXK_TYPE_HEADER = 0,
+ CNXK_TYPE_MSG,
+} cnxk_type_t;
+
+typedef enum CNXK_REP_MSG {
+ /* General sync messages */
+ CNXK_REP_MSG_READY = 0,
+ CNXK_REP_MSG_ACK,
+ CNXK_REP_MSG_EXIT,
+ /* End of messaging sequence */
+ CNXK_REP_MSG_END,
+} cnxk_rep_msg_t;
+
+/* Types */
+typedef struct cnxk_type_data {
+ cnxk_type_t type;
+ uint32_t length;
+ uint64_t data[];
+} __rte_packed cnxk_type_data_t;
+
+/* Header */
+typedef struct cnxk_header {
+ uint64_t signature;
+ uint16_t nb_hops;
+} __rte_packed cnxk_header_t;
+
+/* Message meta */
+typedef struct cnxk_rep_msg_data {
+ cnxk_rep_msg_t type;
+ uint32_t length;
+ uint64_t data[];
+} __rte_packed cnxk_rep_msg_data_t;
+
+/* Ack msg */
+typedef struct cnxk_rep_msg_ack_data {
+ cnxk_rep_msg_t type;
+ uint32_t size;
+ union {
+ void *data;
+ uint64_t val;
+ int64_t sval;
+ } u;
+} __rte_packed cnxk_rep_msg_ack_data_t;
+
+/* Ready msg */
+typedef struct cnxk_rep_msg_ready_data {
+ uint8_t val;
+} __rte_packed cnxk_rep_msg_ready_data_t;
+
+/* Exit msg */
+typedef struct cnxk_rep_msg_exit_data {
+ uint8_t val;
+} __rte_packed cnxk_rep_msg_exit_data_t;
+
+void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
+ uint32_t size);
+void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
+ cnxk_rep_msg_t msg);
+void cnxk_rep_msg_populate_msg_end(void *buffer, uint32_t *length);
+void cnxk_rep_msg_populate_type(void *buffer, uint32_t *length, cnxk_type_t type, uint32_t sz);
+void cnxk_rep_msg_populate_header(void *buffer, uint32_t *length);
+int cnxk_rep_msg_send_process(struct cnxk_rep_dev *rep_dev, void *buffer, uint32_t len,
+ cnxk_rep_msg_ack_data_t *adata);
+int cnxk_rep_control_thread_launch(struct cnxk_eth_dev *pf_dev);
+
+#endif /* __CNXK_REP_MSG_H__ */
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index 38dde54ce9..0e7334f5cd 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -33,6 +33,7 @@ sources = files(
'cnxk_ptp.c',
'cnxk_flow.c',
'cnxk_rep.c',
+ 'cnxk_rep_msg.c',
'cnxk_rep_ops.c',
'cnxk_stats.c',
'cnxk_tm.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH 6/9] net/cnxk: representor ethdev ops
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
` (4 preceding siblings ...)
2023-08-11 16:34 ` [PATCH 5/9] net/cnxk: add representor control plane Harman Kalra
@ 2023-08-11 16:34 ` Harman Kalra
2023-08-11 16:34 ` [PATCH 7/9] net/cnxk: representor flow ops Harman Kalra
` (7 subsequent siblings)
13 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-08-11 16:34 UTC (permalink / raw)
To: jerinj, Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao
Cc: dev, Harman Kalra
Implementing ethernet device operation callbacks for
port representors PMD
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_rep.c | 62 +--
drivers/net/cnxk/cnxk_rep.h | 36 ++
drivers/net/cnxk/cnxk_rep_msg.h | 15 +
drivers/net/cnxk/cnxk_rep_ops.c | 655 ++++++++++++++++++++++++++++++--
4 files changed, 713 insertions(+), 55 deletions(-)
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
index e6f5790adc..5ee7e93ab9 100644
--- a/drivers/net/cnxk/cnxk_rep.c
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -13,6 +13,9 @@ struct eth_dev_ops cnxk_rep_dev_ops = {
.rx_queue_release = cnxk_rep_rx_queue_release,
.tx_queue_setup = cnxk_rep_tx_queue_setup,
.tx_queue_release = cnxk_rep_tx_queue_release,
+ .promiscuous_enable = cnxk_rep_promiscuous_enable,
+ .promiscuous_disable = cnxk_rep_promiscuous_disable,
+ .mac_addr_set = cnxk_rep_mac_addr_set,
.link_update = cnxk_rep_link_update,
.dev_close = cnxk_rep_dev_close,
.dev_stop = cnxk_rep_dev_stop,
@@ -24,14 +27,36 @@ struct eth_dev_ops cnxk_rep_dev_ops = {
int
cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev)
{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ rep_xport_vdev_cfg_t *rep_xport_vdev_cfg = NULL;
+ const struct plt_memzone *mz;
+
if (rte_eal_process_type() != RTE_PROC_PRIMARY)
return 0;
+ mz = plt_memzone_lookup(CNXK_REP_XPORT_VDEV_CFG_MZ);
+ if (!mz) {
+ plt_err("Failed to lookup a memzone, rep id %d, err %d",
+ rep_dev->vf_id, rte_errno);
+ goto fail;
+ }
+
+ rep_xport_vdev_cfg = mz->addr;
plt_rep_dbg("Representor port:%d uninit", ethdev->data->port_id);
rte_free(ethdev->data->mac_addrs);
ethdev->data->mac_addrs = NULL;
+ rep_xport_vdev_cfg->nb_rep_ports--;
+ /* Once all representors are closed, cleanup rep base vdev config */
+ if (!rep_xport_vdev_cfg->nb_rep_ports) {
+ plt_free(rep_xport_vdev_cfg->q_bmap_mem);
+ plt_free(rep_xport_vdev_cfg->mdevinfo);
+ plt_memzone_free(mz);
+ }
+
return 0;
+fail:
+ return rte_errno;
}
int
@@ -121,26 +146,6 @@ cnxk_init_rep_internal(struct cnxk_eth_dev *pf_dev)
return rc;
}
-static uint16_t
-cnxk_rep_tx_burst(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
-{
- PLT_SET_USED(tx_queue);
- PLT_SET_USED(tx_pkts);
- PLT_SET_USED(nb_pkts);
-
- return 0;
-}
-
-static uint16_t
-cnxk_rep_rx_burst(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
-{
- PLT_SET_USED(rx_queue);
- PLT_SET_USED(rx_pkts);
- PLT_SET_USED(nb_pkts);
-
- return 0;
-}
-
static int
cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
{
@@ -152,6 +157,11 @@ cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
rep_dev->vf_id = rep_params->vf_id;
rep_dev->switch_domain_id = rep_params->switch_domain_id;
rep_dev->parent_dev = rep_params->parent_dev;
+ rep_dev->u.rxq = UINT16_MAX;
+ rep_dev->u.txq = UINT16_MAX;
+
+ pf_dev = cnxk_eth_pmd_priv(rep_dev->parent_dev);
+ rep_dev->rep_xport_vdev = pf_dev->rep_xport_vdev;
eth_dev->data->dev_flags |= RTE_ETH_DEV_REPRESENTOR;
eth_dev->data->representor_id = rep_params->vf_id;
@@ -170,11 +180,10 @@ cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
eth_dev->dev_ops = &cnxk_rep_dev_ops;
/* Rx/Tx functions stubs to avoid crashing */
- eth_dev->rx_pkt_burst = cnxk_rep_rx_burst;
- eth_dev->tx_pkt_burst = cnxk_rep_tx_burst;
+ eth_dev->rx_pkt_burst = cnxk_rep_rx_burst_dummy;
+ eth_dev->tx_pkt_burst = cnxk_rep_tx_burst_dummy;
/* Link state. Inherited from PF */
- pf_dev = cnxk_eth_pmd_priv(rep_dev->parent_dev);
link = &pf_dev->eth_dev->data->dev_link;
eth_dev->data->dev_link.link_speed = link->link_speed;
@@ -325,13 +334,6 @@ cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct rte_eth_dev *pf_ethdev
goto err;
}
- /* Launch a thread to handle control messages */
- rc = cnxk_rep_control_thread_launch(pf_dev);
- if (rc) {
- plt_err("Failed to launch message ctrl thread");
- goto err;
- }
-
return 0;
err:
return rc;
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index 8825fa1cf2..2b6403f003 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -6,6 +6,7 @@
#ifndef __CNXK_REP_H__
#define __CNXK_REP_H__
+#define CNXK_REP_XPORT_VDEV_CFG_MZ "rep_xport_vdev_cfg"
#define CNXK_REP_XPORT_VDEV_DEVARGS "role=server"
#define CNXK_REP_XPORT_VDEV_NAME "net_memif"
#define CNXK_REP_VDEV_CTRL_QUEUE 0
@@ -14,6 +15,18 @@
/* Common ethdev ops */
extern struct eth_dev_ops cnxk_rep_dev_ops;
+/* Representor base device configurations */
+typedef struct rep_xport_vdev_cfg_s {
+ struct plt_bitmap *q_map;
+ void *q_bmap_mem;
+ uint8_t nb_rep_ports;
+ uint8_t nb_rep_started;
+ struct rte_mempool *ctrl_chan_pool;
+ struct rte_eth_dev_info *mdevinfo;
+ bool rep_xport_configured;
+} rep_xport_vdev_cfg_t;
+
+/* Representor port configurations */
struct cnxk_rep_dev {
uint16_t vf_id;
uint16_t switch_domain_id;
@@ -22,15 +35,33 @@ struct cnxk_rep_dev {
uint16_t rep_xport_vdev;
bool is_vf_active;
uint16_t pf_func;
+ union {
+ uint16_t rxq;
+ uint16_t txq;
+ uint16_t rep_portid;
+ } u;
uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
};
+/* Inline functions */
static inline struct cnxk_rep_dev *
cnxk_rep_pmd_priv(const struct rte_eth_dev *eth_dev)
{
return eth_dev->data->dev_private;
}
+static inline struct rte_eth_dev *
+cnxk_rep_xport_eth_dev(uint16_t portid)
+{
+ if (!rte_eth_dev_is_valid_port(portid)) {
+ plt_err("Invalid port_id=%u", portid);
+ return NULL;
+ }
+
+ return &rte_eth_devices[portid];
+}
+
+/* Prototypes */
int cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct rte_eth_dev *pf_ethdev,
struct rte_eth_devargs *eth_da);
int cnxk_rep_dev_remove(struct rte_eth_dev *pf_ethdev);
@@ -52,5 +83,10 @@ int cnxk_rep_dev_close(struct rte_eth_dev *eth_dev);
int cnxk_rep_stats_get(struct rte_eth_dev *eth_dev, struct rte_eth_stats *stats);
int cnxk_rep_stats_reset(struct rte_eth_dev *eth_dev);
int cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops);
+int cnxk_rep_promiscuous_enable(struct rte_eth_dev *ethdev);
+int cnxk_rep_promiscuous_disable(struct rte_eth_dev *ethdev);
+int cnxk_rep_mac_addr_set(struct rte_eth_dev *eth_dev, struct rte_ether_addr *addr);
+uint16_t cnxk_rep_tx_burst_dummy(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts);
+uint16_t cnxk_rep_rx_burst_dummy(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts);
#endif /* __CNXK_REP_H__ */
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
index a28c63f762..554122d7f8 100644
--- a/drivers/net/cnxk/cnxk_rep_msg.h
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -19,6 +19,10 @@ typedef enum CNXK_REP_MSG {
CNXK_REP_MSG_READY = 0,
CNXK_REP_MSG_ACK,
CNXK_REP_MSG_EXIT,
+ /* Ethernet operation msgs */
+ CNXK_REP_MSG_ETH_SET_MAC,
+ CNXK_REP_MSG_ETH_STATS_GET,
+ CNXK_REP_MSG_ETH_STATS_CLEAR,
/* End of messaging sequence */
CNXK_REP_MSG_END,
} cnxk_rep_msg_t;
@@ -64,6 +68,17 @@ typedef struct cnxk_rep_msg_exit_data {
uint8_t val;
} __rte_packed cnxk_rep_msg_exit_data_t;
+/* Ethernet op - set mac */
+typedef struct cnxk_rep_msg_eth_mac_set_meta {
+ uint16_t portid;
+ uint8_t addr_bytes[RTE_ETHER_ADDR_LEN];
+} __rte_packed cnxk_rep_msg_eth_set_mac_meta_t;
+
+/* Ethernet op - get/clear stats */
+typedef struct cnxk_rep_msg_eth_stats_meta {
+ uint16_t portid;
+} __rte_packed cnxk_rep_msg_eth_stats_meta_t;
+
void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
uint32_t size);
void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
index 3f1aab077b..022a5137df 100644
--- a/drivers/net/cnxk/cnxk_rep_ops.c
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -3,6 +3,54 @@
*/
#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
+
+#define MEMPOOL_CACHE_SIZE 256
+#define TX_DESC_PER_QUEUE 512
+#define RX_DESC_PER_QUEUE 256
+#define NB_REP_VDEV_MBUF 1024
+
+static uint16_t
+cnxk_rep_tx_burst(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
+{
+ struct cnxk_rep_dev *rep_dev = tx_queue;
+
+ nb_pkts = rte_eth_tx_burst(rep_dev->rep_xport_vdev, rep_dev->u.txq, tx_pkts, nb_pkts);
+
+ return nb_pkts;
+}
+
+static uint16_t
+cnxk_rep_rx_burst(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
+{
+ struct cnxk_rep_dev *rep_dev = rx_queue;
+
+ nb_pkts = rte_eth_rx_burst(rep_dev->rep_xport_vdev, rep_dev->u.txq, rx_pkts, 32);
+ if (nb_pkts == 0)
+ return 0;
+
+ return nb_pkts;
+}
+
+uint16_t
+cnxk_rep_tx_burst_dummy(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(tx_queue);
+ PLT_SET_USED(tx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
+
+uint16_t
+cnxk_rep_rx_burst_dummy(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(rx_queue);
+ PLT_SET_USED(rx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
int
cnxk_rep_link_update(struct rte_eth_dev *ethdev, int wait_to_complete)
@@ -13,39 +61,379 @@ cnxk_rep_link_update(struct rte_eth_dev *ethdev, int wait_to_complete)
}
int
-cnxk_rep_dev_info_get(struct rte_eth_dev *ethdev, struct rte_eth_dev_info *devinfo)
+cnxk_rep_dev_info_get(struct rte_eth_dev *ethdev, struct rte_eth_dev_info *dev_info)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(devinfo);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ rep_xport_vdev_cfg_t *rep_xport_vdev_cfg = NULL;
+ struct rte_eth_dev_info mdevinfo;
+ const struct plt_memzone *mz;
+ int rc = 0;
+
+ mz = plt_memzone_lookup(CNXK_REP_XPORT_VDEV_CFG_MZ);
+ if (!mz) {
+ mz = plt_memzone_reserve_cache_align(CNXK_REP_XPORT_VDEV_CFG_MZ,
+ sizeof(rep_xport_vdev_cfg_t));
+ if (!mz) {
+ plt_err("Failed to reserve a memzone, rep id %d, err %d",
+ rep_dev->vf_id, rte_errno);
+ goto fail;
+ }
+ }
+
+ rep_xport_vdev_cfg = mz->addr;
+ /* Get the rep base vdev devinfo */
+ if (!rep_xport_vdev_cfg->mdevinfo) {
+ rc = rte_eth_dev_info_get(rep_dev->rep_xport_vdev, &mdevinfo);
+ if (rc) {
+ plt_err("Failed to get rep_xport port dev info, err %d", rc);
+ goto fail;
+ }
+ rep_xport_vdev_cfg->mdevinfo = plt_zmalloc(sizeof(struct rte_eth_dev_info), 0);
+ if (!rep_xport_vdev_cfg->mdevinfo) {
+ plt_err("Failed to alloc memory for dev info");
+ goto fail;
+ }
+ rte_memcpy(rep_xport_vdev_cfg->mdevinfo, &mdevinfo,
+ sizeof(struct rte_eth_dev_info));
+ }
+
+ /* Use rep_xport device info */
+ dev_info->max_mac_addrs = rep_xport_vdev_cfg->mdevinfo->max_mac_addrs;
+ dev_info->max_rx_pktlen = rep_xport_vdev_cfg->mdevinfo->max_rx_pktlen;
+ dev_info->min_rx_bufsize = rep_xport_vdev_cfg->mdevinfo->min_rx_bufsize;
+ dev_info->tx_offload_capa = rep_xport_vdev_cfg->mdevinfo->tx_offload_capa;
+
+ /* For the sake of symmetry, max_rx_queues = max_tx_queues */
+ dev_info->max_rx_queues = 1;
+ dev_info->max_tx_queues = 1;
+
+ /* MTU specifics */
+ dev_info->max_mtu = rep_xport_vdev_cfg->mdevinfo->max_mtu;
+ dev_info->min_mtu = rep_xport_vdev_cfg->mdevinfo->min_mtu;
+
+ /* Switch info specific */
+ dev_info->switch_info.name = ethdev->device->name;
+ dev_info->switch_info.domain_id = rep_dev->switch_domain_id;
+ dev_info->switch_info.port_id = rep_dev->vf_id;
+
return 0;
+fail:
+ return rc;
+}
+
+static inline int
+bitmap_ctzll(uint64_t slab)
+{
+ if (slab == 0)
+ return 0;
+
+ return __builtin_ctzll(slab);
+}
+
+static uint16_t
+alloc_rep_xport_qid(struct plt_bitmap *bmp)
+{
+ uint16_t idx, rc;
+ uint64_t slab;
+ uint32_t pos;
+
+ pos = 0;
+ slab = 0;
+ /* Scan from the beginning */
+ plt_bitmap_scan_init(bmp);
+ /* Scan bitmap to get the free pool */
+ rc = plt_bitmap_scan(bmp, &pos, &slab);
+ /* Empty bitmap */
+ if (rc == 0)
+ return UINT16_MAX;
+
+ idx = pos + bitmap_ctzll(slab);
+ plt_bitmap_clear(bmp, idx);
+ return idx;
+}
+
+static int
+configure_rep_xport_queues_map(rep_xport_vdev_cfg_t *rep_xport_vdev_cfg)
+{
+ int id, rc = 0, q_max;
+ uint32_t bmap_sz;
+ void *bmap_mem;
+
+ q_max = CNXK_MAX_REP_PORTS + 1;
+ /* Return success on no-pci case */
+ if (!q_max)
+ return 0;
+
+ bmap_sz = plt_bitmap_get_memory_footprint(q_max);
+
+ /* Allocate memory for rep_xport queue bitmap */
+ bmap_mem = plt_zmalloc(bmap_sz, RTE_CACHE_LINE_SIZE);
+ if (bmap_mem == NULL) {
+ plt_err("Failed to allocate memory for worker lmt bmap");
+ rc = -ENOMEM;
+ goto exit;
+ }
+ rep_xport_vdev_cfg->q_bmap_mem = bmap_mem;
+
+ /* Initialize worker lmt bitmap */
+ rep_xport_vdev_cfg->q_map = plt_bitmap_init(q_max, bmap_mem, bmap_sz);
+ if (!rep_xport_vdev_cfg->q_map) {
+ plt_err("Failed to initialize rep_xport queue bitmap");
+ rc = -EIO;
+ goto exit;
+ }
+
+ /* Set all the queue initially */
+ for (id = 0; id < q_max; id++)
+ plt_bitmap_set(rep_xport_vdev_cfg->q_bmap_mem, id);
+
+ return 0;
+exit:
+ return rc;
+}
+
+static uint16_t
+cnxk_rep_eth_dev_count_total(void)
+{
+ uint16_t port, count = 0;
+ struct rte_eth_dev *ethdev;
+
+ RTE_ETH_FOREACH_DEV(port) {
+ ethdev = &rte_eth_devices[port];
+ if (ethdev->data->dev_flags & RTE_ETH_DEV_REPRESENTOR)
+ count++;
+ }
+
+ return count;
+}
+
+static int
+configure_control_channel(rep_xport_vdev_cfg_t *rep_xport_vdev_cfg, uint16_t portid)
+{
+ struct rte_mempool *ctrl_chan_pool = NULL;
+ int rc;
+
+ /* Allocate a qid for control channel */
+ alloc_rep_xport_qid(rep_xport_vdev_cfg->q_map);
+
+ /* Create the mbuf pool. */
+ ctrl_chan_pool = rte_pktmbuf_pool_create("rep_xport_ctrl_pool", NB_REP_VDEV_MBUF,
+ MEMPOOL_CACHE_SIZE, RTE_CACHE_LINE_SIZE,
+ RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id());
+
+ if (ctrl_chan_pool == NULL) {
+ plt_err("Cannot init mbuf pool");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ /* Setup a RX queue for control channel */
+ rc = rte_eth_rx_queue_setup(portid, CNXK_REP_VDEV_CTRL_QUEUE, RX_DESC_PER_QUEUE,
+ rte_eth_dev_socket_id(portid), NULL, ctrl_chan_pool);
+ if (rc < 0) {
+ plt_err("rte_eth_rx_queue_setup:err=%d, port=%u\n", rc, portid);
+ goto fail;
+ }
+
+ /* Setup a TX queue for control channel */
+ rc = rte_eth_tx_queue_setup(portid, CNXK_REP_VDEV_CTRL_QUEUE, TX_DESC_PER_QUEUE,
+ rte_eth_dev_socket_id(portid), NULL);
+ if (rc < 0) {
+ plt_err("TX queue setup failed, err %d port %d", rc, portid);
+ goto fail;
+ }
+
+ rep_xport_vdev_cfg->ctrl_chan_pool = ctrl_chan_pool;
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+configure_rep_xport_dev(rep_xport_vdev_cfg_t *rep_xport_vdev_cfg, uint16_t portid)
+{
+ struct rte_eth_dev *rep_xport_ethdev = cnxk_rep_xport_eth_dev(portid);
+ static struct rte_eth_conf port_conf_default;
+ uint16_t nb_rxq, nb_txq, nb_rep_ports;
+ int rc = 0;
+
+ /* If rep_xport port already started, stop it and reconfigure */
+ if (rep_xport_ethdev->data->dev_started)
+ rte_eth_dev_stop(portid);
+
+ /* Get the no of representors probed */
+ nb_rep_ports = cnxk_rep_eth_dev_count_total();
+ if (nb_rep_ports > CNXK_MAX_REP_PORTS) {
+ plt_err("Representors probed %d > Max supported %d", nb_rep_ports,
+ CNXK_MAX_REP_PORTS);
+ goto fail;
+ }
+
+ /* Each queue of rep_xport describes representor port. 1 additional queue is
+ * configured as control channel to configure flows, etc.
+ */
+ nb_rxq = CNXK_MAX_REP_PORTS + 1;
+ nb_txq = CNXK_MAX_REP_PORTS + 1;
+
+ rc = rte_eth_dev_configure(portid, nb_rxq, nb_txq, &port_conf_default);
+ if (rc) {
+ plt_err("Failed to configure rep_xport port: %d", rc);
+ goto fail;
+ }
+
+ rep_xport_vdev_cfg->rep_xport_configured = true;
+ rep_xport_vdev_cfg->nb_rep_ports = nb_rep_ports;
+
+ return 0;
+fail:
+ return rc;
}
int
cnxk_rep_dev_configure(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ rep_xport_vdev_cfg_t *rep_xport_vdev_cfg = NULL;
+ const struct plt_memzone *mz;
+ int rc = -1;
+
+ mz = plt_memzone_lookup(CNXK_REP_XPORT_VDEV_CFG_MZ);
+ if (!mz) {
+ mz = plt_memzone_reserve_cache_align(CNXK_REP_XPORT_VDEV_CFG_MZ,
+ sizeof(rep_xport_vdev_cfg_t));
+ if (!mz) {
+ plt_err("Failed to reserve a memzone, rep id %d, err %d",
+ rep_dev->vf_id, rte_errno);
+ goto fail;
+ }
+ }
+
+ rep_xport_vdev_cfg = mz->addr;
+ /* Return if rep_xport dev already configured */
+ if (rep_xport_vdev_cfg->rep_xport_configured) {
+ rep_dev->ctrl_chan_pool = rep_xport_vdev_cfg->ctrl_chan_pool;
+ return 0;
+ }
+
+ /* Configure rep_xport pmd */
+ rc = configure_rep_xport_dev(rep_xport_vdev_cfg, rep_dev->rep_xport_vdev);
+ if (rc) {
+ plt_err("Configuring rep_xport port failed");
+ goto free;
+ }
+
+ /* Setup a bitmap for rep_xport queues */
+ rc = configure_rep_xport_queues_map(rep_xport_vdev_cfg);
+ if (rc != 0) {
+ plt_err("Failed to setup rep_xport queue map, err %d", rc);
+ goto free;
+ }
+
+ /* Setup a queue for control channel */
+ rc = configure_control_channel(rep_xport_vdev_cfg, rep_dev->rep_xport_vdev);
+ if (rc != 0) {
+ plt_err("Failed to setup control channgel, err %d", rc);
+ goto free;
+ }
+ rep_dev->ctrl_chan_pool = rep_xport_vdev_cfg->ctrl_chan_pool;
+
return 0;
+free:
+ plt_memzone_free(mz);
+fail:
+ return rc;
}
int
-cnxk_rep_dev_start(struct rte_eth_dev *ethdev)
+cnxk_rep_promiscuous_enable(struct rte_eth_dev *ethdev)
{
PLT_SET_USED(ethdev);
return 0;
}
int
-cnxk_rep_dev_close(struct rte_eth_dev *ethdev)
+cnxk_rep_promiscuous_disable(struct rte_eth_dev *ethdev)
{
PLT_SET_USED(ethdev);
return 0;
}
+int
+cnxk_rep_dev_start(struct rte_eth_dev *ethdev)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ rep_xport_vdev_cfg_t *rep_xport_vdev_cfg = NULL;
+ const struct plt_memzone *mz;
+ int rc = 0;
+
+ mz = plt_memzone_lookup(CNXK_REP_XPORT_VDEV_CFG_MZ);
+ if (!mz) {
+ plt_err("Failed to lookup a memzone, rep id %d, err %d",
+ rep_dev->vf_id, rte_errno);
+ goto fail;
+ }
+
+ rep_xport_vdev_cfg = mz->addr;
+ ethdev->rx_pkt_burst = cnxk_rep_rx_burst;
+ ethdev->tx_pkt_burst = cnxk_rep_tx_burst;
+
+ /* Start rep_xport device only once after first representor gets active */
+ if (!rep_xport_vdev_cfg->nb_rep_started) {
+ rc = rte_eth_dev_start(rep_dev->rep_xport_vdev);
+ if (rc) {
+ plt_err("Rep base vdev portid %d start failed, err %d",
+ rep_dev->rep_xport_vdev, rc);
+ goto fail;
+ }
+
+ /* Launch a thread to handle control messages */
+ rc = cnxk_rep_control_thread_launch(cnxk_eth_pmd_priv(rep_dev->parent_dev));
+ if (rc) {
+ plt_err("Failed to launch message ctrl thread");
+ goto fail;
+ }
+ }
+
+ rep_xport_vdev_cfg->nb_rep_started++;
+
+ return 0;
+fail:
+ return rc;
+}
+
+int
+cnxk_rep_dev_close(struct rte_eth_dev *ethdev)
+{
+ return cnxk_rep_dev_uninit(ethdev);
+}
+
int
cnxk_rep_dev_stop(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ rep_xport_vdev_cfg_t *rep_xport_vdev_cfg = NULL;
+ const struct plt_memzone *mz;
+
+ mz = plt_memzone_lookup(CNXK_REP_XPORT_VDEV_CFG_MZ);
+ if (!mz) {
+ plt_err("Failed to lookup a memzone, rep id %d, err %d",
+ rep_dev->vf_id, rte_errno);
+ goto fail;
+ }
+
+ rep_xport_vdev_cfg = mz->addr;
+ ethdev->rx_pkt_burst = cnxk_rep_rx_burst_dummy;
+ ethdev->tx_pkt_burst = cnxk_rep_tx_burst_dummy;
+ rep_xport_vdev_cfg->nb_rep_started--;
+
+ /* Stop rep_xport device only after all other devices stopped */
+ if (!rep_xport_vdev_cfg->nb_rep_started)
+ rte_eth_dev_stop(rep_dev->rep_xport_vdev);
+
return 0;
+fail:
+ return rte_errno;
}
int
@@ -53,54 +441,220 @@ cnxk_rep_rx_queue_setup(struct rte_eth_dev *ethdev, uint16_t rx_queue_id, uint16
unsigned int socket_id, const struct rte_eth_rxconf *rx_conf,
struct rte_mempool *mb_pool)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(rx_queue_id);
- PLT_SET_USED(nb_rx_desc);
- PLT_SET_USED(socket_id);
- PLT_SET_USED(rx_conf);
- PLT_SET_USED(mb_pool);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ rep_xport_vdev_cfg_t *rep_xport_vdev_cfg = NULL;
+ const struct plt_memzone *mz;
+ int rc = 0;
+
+ mz = plt_memzone_lookup(CNXK_REP_XPORT_VDEV_CFG_MZ);
+ if (!mz) {
+ plt_err("Failed to lookup a memzone, rep id %d, err %d",
+ rep_dev->vf_id, rte_errno);
+ goto fail;
+ }
+
+ rep_xport_vdev_cfg = mz->addr;
+ /* Allocate a qid, if tx queue setup already done use the same qid */
+ if (rep_dev->u.rxq == UINT16_MAX && rep_dev->u.txq == UINT16_MAX)
+ rep_dev->u.rxq = alloc_rep_xport_qid(rep_xport_vdev_cfg->q_map);
+ else
+ rep_dev->u.rxq = rep_dev->u.txq;
+
+ /* Setup the RX queue */
+ rc = rte_eth_rx_queue_setup(rep_dev->rep_xport_vdev, rep_dev->u.rxq, nb_rx_desc, socket_id,
+ rx_conf, mb_pool);
+ if (rc < 0) {
+ plt_err("rte_eth_rx_queue_setup:err=%d, port=%u\n", rc, rep_dev->rep_xport_vdev);
+ goto fail;
+ }
+
+ ethdev->data->rx_queues[rx_queue_id] = rep_dev;
+ plt_info("Representor id %d portid %d rxq %d", rep_dev->vf_id, ethdev->data->port_id,
+ rep_dev->u.rxq);
+
return 0;
+fail:
+ return rc;
}
void
cnxk_rep_rx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(queue_id);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ rep_xport_vdev_cfg_t *rep_xport_vdev_cfg = NULL;
+ const struct plt_memzone *mz;
+ RTE_SET_USED(queue_id);
+
+ mz = plt_memzone_lookup(CNXK_REP_XPORT_VDEV_CFG_MZ);
+ if (!mz) {
+ plt_err("Failed to lookup a memzone, rep id %d, err %d",
+ rep_dev->vf_id, rte_errno);
+ return;
+ }
+
+ rep_xport_vdev_cfg = mz->addr;
+ plt_bitmap_clear(rep_xport_vdev_cfg->q_bmap_mem, rep_dev->u.rxq);
}
int
cnxk_rep_tx_queue_setup(struct rte_eth_dev *ethdev, uint16_t tx_queue_id, uint16_t nb_tx_desc,
unsigned int socket_id, const struct rte_eth_txconf *tx_conf)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(tx_queue_id);
- PLT_SET_USED(nb_tx_desc);
- PLT_SET_USED(socket_id);
- PLT_SET_USED(tx_conf);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ rep_xport_vdev_cfg_t *rep_xport_vdev_cfg = NULL;
+ const struct plt_memzone *mz;
+ int rc = 0;
+
+ mz = plt_memzone_lookup(CNXK_REP_XPORT_VDEV_CFG_MZ);
+ if (!mz) {
+ plt_err("Failed to lookup a memzone, rep id %d, err %d",
+ rep_dev->vf_id, rte_errno);
+ goto fail;
+ }
+
+ rep_xport_vdev_cfg = mz->addr;
+ /* Allocate a qid, if rx queue setup already done use the same qid */
+ if (rep_dev->u.rxq == UINT16_MAX && rep_dev->u.txq == UINT16_MAX)
+ rep_dev->u.txq = alloc_rep_xport_qid(rep_xport_vdev_cfg->q_map);
+ else
+ rep_dev->u.txq = rep_dev->u.rxq;
+
+ /* Setup the TX queue */
+ rc = rte_eth_tx_queue_setup(rep_dev->rep_xport_vdev, rep_dev->u.txq, nb_tx_desc, socket_id,
+ tx_conf);
+ if (rc < 0) {
+ plt_err("TX queue setup failed, err %d port %d", rc, rep_dev->rep_xport_vdev);
+ goto fail;
+ }
+
+ ethdev->data->tx_queues[tx_queue_id] = rep_dev;
+ plt_info("Representor id %d portid %d txq %d", rep_dev->vf_id, ethdev->data->port_id,
+ rep_dev->u.txq);
+
return 0;
+fail:
+ return rc;
}
void
cnxk_rep_tx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
{
- PLT_SET_USED(ethdev);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ rep_xport_vdev_cfg_t *rep_xport_vdev_cfg = NULL;
+ const struct plt_memzone *mz;
PLT_SET_USED(queue_id);
+
+ mz = plt_memzone_lookup(CNXK_REP_XPORT_VDEV_CFG_MZ);
+ if (!mz) {
+ plt_err("Failed to lookup a memzone, rep id %d, err %d",
+ rep_dev->vf_id, rte_errno);
+ return;
+ }
+
+ rep_xport_vdev_cfg = mz->addr;
+ plt_bitmap_clear(rep_xport_vdev_cfg->q_bmap_mem, rep_dev->u.txq);
+}
+
+static int
+process_eth_stats(struct cnxk_rep_dev *rep_dev, cnxk_rep_msg_ack_data_t *adata, cnxk_rep_msg_t msg)
+{
+ cnxk_rep_msg_eth_stats_meta_t msg_st_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ size = CNXK_REP_MSG_MAX_BUFFER_SZ;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_st_meta.portid = rep_dev->u.rxq;
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_st_meta,
+ sizeof(cnxk_rep_msg_eth_stats_meta_t), msg);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ rte_free(buffer);
+
+ return 0;
+fail:
+ rte_free(buffer);
+ return rc;
}
int
cnxk_rep_stats_get(struct rte_eth_dev *ethdev, struct rte_eth_stats *stats)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(stats);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ struct rte_eth_stats vf_stats;
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ rc = process_eth_stats(rep_dev, &adata, CNXK_REP_MSG_ETH_STATS_GET);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ plt_err("Failed to clear stats for vf rep %x, err %d", rep_dev->vf_id, rc);
+ }
+
+ if (adata.size != sizeof(struct rte_eth_stats)) {
+ rc = -EINVAL;
+ plt_err("Incomplete stats received for vf rep %d", rep_dev->vf_id);
+ goto fail;
+ }
+
+ rte_memcpy(&vf_stats, adata.u.data, adata.size);
+
+ stats->q_ipackets[0] = vf_stats.ipackets;
+ stats->q_ibytes[0] = vf_stats.ibytes;
+ stats->ipackets = vf_stats.ipackets;
+ stats->ibytes = vf_stats.ibytes;
+
+ stats->q_opackets[0] = vf_stats.opackets;
+ stats->q_obytes[0] = vf_stats.obytes;
+ stats->opackets = vf_stats.opackets;
+ stats->obytes = vf_stats.obytes;
+
return 0;
+fail:
+ return rc;
}
int
cnxk_rep_stats_reset(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
- return 0;
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc = 0;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ rc = process_eth_stats(rep_dev, &adata, CNXK_REP_MSG_ETH_STATS_CLEAR);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ plt_err("Failed to clear stats for vf rep %x, err %d", rep_dev->vf_id, rc);
+ }
+
+ return rc;
}
int
@@ -110,3 +664,54 @@ cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **op
PLT_SET_USED(ops);
return 0;
}
+
+int
+cnxk_rep_mac_addr_set(struct rte_eth_dev *eth_dev, struct rte_ether_addr *addr)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_eth_set_mac_meta_t msg_sm_meta;
+ cnxk_rep_msg_ack_data_t adata;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ /* If representor not representing any VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ size = CNXK_REP_MSG_MAX_BUFFER_SZ;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_sm_meta.portid = rep_dev->u.rxq;
+ rte_memcpy(&msg_sm_meta.addr_bytes, addr->addr_bytes, RTE_ETHER_ADDR_LEN);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_sm_meta,
+ sizeof(cnxk_rep_msg_eth_set_mac_meta_t),
+ CNXK_REP_MSG_ETH_SET_MAC);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, &adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ if (adata.u.sval < 0) {
+ rc = adata.u.sval;
+ plt_err("Failed to set mac address, err %d", rc);
+ goto fail;
+ }
+
+ rte_free(buffer);
+
+ return 0;
+fail:
+ rte_free(buffer);
+ return rc;
+}
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH 7/9] net/cnxk: representor flow ops
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
` (5 preceding siblings ...)
2023-08-11 16:34 ` [PATCH 6/9] net/cnxk: representor ethdev ops Harman Kalra
@ 2023-08-11 16:34 ` Harman Kalra
2023-08-11 16:34 ` [PATCH 8/9] common/cnxk: support represented port for cnxk Harman Kalra
` (6 subsequent siblings)
13 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-08-11 16:34 UTC (permalink / raw)
To: jerinj, Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao
Cc: dev, Harman Kalra
Implementing flow operation callbacks for port representors PMD
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_flow.h | 9 +-
drivers/net/cnxk/cnxk_rep.h | 3 +
drivers/net/cnxk/cnxk_rep_flow.c | 715 +++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 58 +++
drivers/net/cnxk/cnxk_rep_ops.c | 3 +-
drivers/net/cnxk/meson.build | 1 +
6 files changed, 786 insertions(+), 3 deletions(-)
create mode 100644 drivers/net/cnxk/cnxk_rep_flow.c
diff --git a/drivers/net/cnxk/cnxk_flow.h b/drivers/net/cnxk/cnxk_flow.h
index bb23629819..303002176b 100644
--- a/drivers/net/cnxk/cnxk_flow.h
+++ b/drivers/net/cnxk/cnxk_flow.h
@@ -16,8 +16,13 @@ struct cnxk_rte_flow_term_info {
uint16_t item_size;
};
-struct roc_npc_flow *cnxk_flow_create(struct rte_eth_dev *dev,
- const struct rte_flow_attr *attr,
+struct cnxk_rte_flow_action_info {
+ uint16_t conf_size;
+};
+
+extern const struct cnxk_rte_flow_term_info term[];
+
+struct roc_npc_flow *cnxk_flow_create(struct rte_eth_dev *dev, const struct rte_flow_attr *attr,
const struct rte_flow_item pattern[],
const struct rte_flow_action actions[],
struct rte_flow_error *error);
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index 2b6403f003..4886527f83 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -15,6 +15,9 @@
/* Common ethdev ops */
extern struct eth_dev_ops cnxk_rep_dev_ops;
+/* Flow ops for representor ports */
+extern struct rte_flow_ops cnxk_rep_flow_ops;
+
/* Representor base device configurations */
typedef struct rep_xport_vdev_cfg_s {
struct plt_bitmap *q_map;
diff --git a/drivers/net/cnxk/cnxk_rep_flow.c b/drivers/net/cnxk/cnxk_rep_flow.c
new file mode 100644
index 0000000000..9e181f5173
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_flow.c
@@ -0,0 +1,715 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <fcntl.h>
+#include <unistd.h>
+
+#include <cnxk_flow.h>
+#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
+
+#define DEFAULT_DUMP_FILE_NAME "/tmp/fdump"
+#define MAX_BUFFER_SIZE 1500
+
+const struct cnxk_rte_flow_action_info action_info[] = {
+ [RTE_FLOW_ACTION_TYPE_MARK] = {sizeof(struct rte_flow_action_mark)},
+ [RTE_FLOW_ACTION_TYPE_VF] = {sizeof(struct rte_flow_action_vf)},
+ [RTE_FLOW_ACTION_TYPE_REPRESENTED_PORT] = {sizeof(struct rte_flow_action_port_id)},
+ [RTE_FLOW_ACTION_TYPE_PORT_ID] = {sizeof(struct rte_flow_action_port_id)},
+ [RTE_FLOW_ACTION_TYPE_QUEUE] = {sizeof(struct rte_flow_action_queue)},
+ [RTE_FLOW_ACTION_TYPE_RSS] = {sizeof(struct rte_flow_action_rss)},
+ [RTE_FLOW_ACTION_TYPE_SECURITY] = {sizeof(struct rte_flow_action_security)},
+ [RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID] = {sizeof(struct rte_flow_action_of_set_vlan_vid)},
+ [RTE_FLOW_ACTION_TYPE_OF_PUSH_VLAN] = {sizeof(struct rte_flow_action_of_push_vlan)},
+ [RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_PCP] = {sizeof(struct rte_flow_action_of_set_vlan_pcp)},
+ [RTE_FLOW_ACTION_TYPE_METER] = {sizeof(struct rte_flow_action_meter)},
+ [RTE_FLOW_ACTION_TYPE_OF_POP_MPLS] = {sizeof(struct rte_flow_action_of_pop_mpls)},
+ [RTE_FLOW_ACTION_TYPE_OF_PUSH_MPLS] = {sizeof(struct rte_flow_action_of_push_mpls)},
+ [RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP] = {sizeof(struct rte_flow_action_vxlan_encap)},
+ [RTE_FLOW_ACTION_TYPE_NVGRE_ENCAP] = {sizeof(struct rte_flow_action_nvgre_encap)},
+ [RTE_FLOW_ACTION_TYPE_RAW_ENCAP] = {sizeof(struct rte_flow_action_raw_encap)},
+ [RTE_FLOW_ACTION_TYPE_RAW_DECAP] = {sizeof(struct rte_flow_action_raw_decap)},
+ [RTE_FLOW_ACTION_TYPE_COUNT] = {sizeof(struct rte_flow_action_count)},
+};
+
+static void
+cnxk_flow_params_count(const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ uint16_t *n_pattern, uint16_t *n_action)
+{
+ int i = 0;
+
+ for (; pattern->type != RTE_FLOW_ITEM_TYPE_END; pattern++)
+ i++;
+
+ *n_pattern = ++i;
+ plt_rep_dbg("Total patterns is %d", *n_pattern);
+
+ i = 0;
+ for (; actions->type != RTE_FLOW_ACTION_TYPE_END; actions++)
+ i++;
+ *n_action = ++i;
+ plt_rep_dbg("Total actions is %d", *n_action);
+}
+
+static void
+populate_attr_data(void *buffer, uint32_t *length, const struct rte_flow_attr *attr)
+{
+ uint32_t sz = sizeof(struct rte_flow_attr);
+ uint32_t len;
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_ATTR, sz);
+
+ len = *length;
+ /* Populate the attribute data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), attr, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static uint16_t
+prepare_pattern_data(const struct rte_flow_item *pattern, uint16_t nb_pattern,
+ uint64_t *pattern_data)
+{
+ cnxk_pattern_hdr_t hdr;
+ uint16_t len = 0;
+ int i = 0;
+
+ for (i = 0; i < nb_pattern; i++) {
+ /* Populate the pattern type hdr */
+ memset(&hdr, 0, sizeof(cnxk_pattern_hdr_t));
+ hdr.type = pattern->type;
+ if (pattern->spec) {
+ hdr.spec_sz = term[pattern->type].item_size;
+ hdr.last_sz = 0;
+ hdr.mask_sz = term[pattern->type].item_size;
+ }
+
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), &hdr, sizeof(cnxk_pattern_hdr_t));
+ len += sizeof(cnxk_pattern_hdr_t);
+
+ /* Copy pattern spec data */
+ if (pattern->spec) {
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), pattern->spec,
+ term[pattern->type].item_size);
+ len += term[pattern->type].item_size;
+ }
+
+ /* Copy pattern last data */
+ if (pattern->last) {
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), pattern->last,
+ term[pattern->type].item_size);
+ len += term[pattern->type].item_size;
+ }
+
+ /* Copy pattern mask data */
+ if (pattern->mask) {
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), pattern->mask,
+ term[pattern->type].item_size);
+ len += term[pattern->type].item_size;
+ }
+ pattern++;
+ }
+
+ return len;
+}
+
+static void
+populate_pattern_data(void *buffer, uint32_t *length, const struct rte_flow_item *pattern,
+ uint16_t nb_pattern)
+{
+ uint64_t pattern_data[BUFSIZ];
+ uint32_t len;
+ uint32_t sz;
+
+ /* Prepare pattern_data */
+ sz = prepare_pattern_data(pattern, nb_pattern, pattern_data);
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_PATTERN, sz);
+
+ len = *length;
+ /* Populate the pattern data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), pattern_data, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static uint16_t
+populate_rss_action_conf(const struct rte_flow_action_rss *conf, void
+ *rss_action_conf)
+{
+ int len, sz;
+
+ len = sizeof(struct rte_flow_action_rss) - sizeof(conf->key) -
+ sizeof(conf->queue);
+
+ if (rss_action_conf)
+ rte_memcpy(rss_action_conf, conf, len);
+
+ if (conf->key) {
+ sz = conf->key_len;
+ if (rss_action_conf)
+ rte_memcpy(RTE_PTR_ADD(rss_action_conf, len), conf->key,
+ sz);
+ len += sz;
+ }
+
+ if (conf->queue) {
+ sz = conf->queue_num * sizeof(conf->queue);
+ if (rss_action_conf)
+ rte_memcpy(RTE_PTR_ADD(rss_action_conf, len),
+ conf->queue, sz);
+ len += sz;
+ }
+
+ return len;
+}
+
+static uint16_t
+prepare_action_data(const struct rte_flow_action *action, uint16_t nb_action, uint64_t *action_data)
+{
+ void *action_conf_data = NULL;
+ cnxk_action_hdr_t hdr;
+ uint16_t len = 0, sz = 0;
+ int i = 0;
+
+ for (i = 0; i < nb_action; i++) {
+ if (action->conf) {
+ if (action->type == RTE_FLOW_ACTION_TYPE_RSS) {
+ sz = populate_rss_action_conf(action->conf, NULL);
+ action_conf_data = plt_zmalloc(sz, 0);
+ if (populate_rss_action_conf(action->conf,
+ action_conf_data) != sz) {
+ plt_err("Populating RSS action config failed");
+ return 0;
+ }
+ } else {
+ sz = action_info[action->type].conf_size;
+ action_conf_data = plt_zmalloc(sz, 0);
+ rte_memcpy(action_conf_data, action->conf, sz);
+ }
+ }
+
+ /* Populate the action type hdr */
+ memset(&hdr, 0, sizeof(cnxk_action_hdr_t));
+ hdr.type = action->type;
+ hdr.conf_sz = sz;
+
+ rte_memcpy(RTE_PTR_ADD(action_data, len), &hdr, sizeof(cnxk_action_hdr_t));
+ len += sizeof(cnxk_action_hdr_t);
+
+ /* Copy action conf data */
+ if (action_conf_data) {
+ rte_memcpy(RTE_PTR_ADD(action_data, len),
+ action_conf_data, sz);
+ len += sz;
+ plt_free(action_conf_data);
+ action_conf_data = NULL;
+ }
+
+ action++;
+ }
+
+ return len;
+}
+
+static void
+populate_action_data(void *buffer, uint32_t *length, const struct rte_flow_action *action,
+ uint16_t nb_action)
+{
+ uint64_t action_data[BUFSIZ];
+ uint32_t len;
+ uint32_t sz;
+
+ /* Prepare action_data */
+ sz = prepare_action_data(action, nb_action, action_data);
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_ACTION, sz);
+
+ len = *length;
+ /* Populate the action data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), action_data, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static int
+process_flow_destroy(struct cnxk_rep_dev *rep_dev, void *flow, cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_destroy_meta_t msg_fd_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_fd_meta.portid = rep_dev->u.rep_portid;
+ msg_fd_meta.flow = (uint64_t)flow;
+ plt_rep_dbg("Flow Destroy: flow 0x%lx, portid %d", msg_fd_meta.flow, msg_fd_meta.portid);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_fd_meta,
+ sizeof(cnxk_rep_msg_flow_destroy_meta_t),
+ CNXK_REP_MSG_FLOW_DESTROY);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+copy_flow_dump_file(FILE *target)
+{
+ FILE *source = NULL;
+ int pos;
+ char ch;
+
+ source = fopen(DEFAULT_DUMP_FILE_NAME, "r");
+ if (source == NULL) {
+ plt_err("Failed to read default dump file: %s, err %d", DEFAULT_DUMP_FILE_NAME,
+ errno);
+ return errno;
+ }
+
+ fseek(source, 0L, SEEK_END);
+ pos = ftell(source);
+ fseek(source, 0L, SEEK_SET);
+ while (pos--) {
+ ch = fgetc(source);
+ fputc(ch, target);
+ }
+
+ fclose(source);
+
+ /* Remove the default file after reading */
+ remove(DEFAULT_DUMP_FILE_NAME);
+
+ return 0;
+}
+
+static int
+process_flow_dump(struct cnxk_rep_dev *rep_dev, struct rte_flow *flow, FILE *file,
+ cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_dump_meta_t msg_fp_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_fp_meta.portid = rep_dev->u.rep_portid;
+ msg_fp_meta.flow = (uint64_t)flow;
+ msg_fp_meta.is_stdout = (file == stdout) ? 1 : 0;
+
+ plt_rep_dbg("Flow Dump: flow 0x%lx, portid %d stdout %d", msg_fp_meta.flow,
+ msg_fp_meta.portid, msg_fp_meta.is_stdout);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_fp_meta,
+ sizeof(cnxk_rep_msg_flow_dump_meta_t),
+ CNXK_REP_MSG_FLOW_DUMP);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ /* Copy contents from default file to user file */
+ if (file != stdout)
+ copy_flow_dump_file(file);
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+process_flow_flush(struct cnxk_rep_dev *rep_dev, cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_flush_meta_t msg_ff_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_ff_meta.portid = rep_dev->u.rep_portid;
+ plt_rep_dbg("Flow Flush: portid %d", msg_ff_meta.portid);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_ff_meta,
+ sizeof(cnxk_rep_msg_flow_flush_meta_t),
+ CNXK_REP_MSG_FLOW_FLUSH);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+process_flow_query(struct cnxk_rep_dev *rep_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_query_meta_t *msg_fq_meta;
+ uint32_t len = 0, rc, sz, total_sz;
+ uint64_t action_data[BUFSIZ];
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ sz = prepare_action_data(action, 1, action_data);
+ total_sz = sz + sizeof(cnxk_rep_msg_flow_query_meta_t);
+
+ msg_fq_meta = plt_zmalloc(total_sz, 0);
+ if (!msg_fq_meta) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ msg_fq_meta->portid = rep_dev->u.rep_portid;
+ msg_fq_meta->flow = (uint64_t)flow;
+ /* Populate the action data */
+ rte_memcpy(msg_fq_meta->action_data, action_data, sz);
+ msg_fq_meta->action_data_sz = sz;
+
+ plt_rep_dbg("Flow query: flow 0x%lx, portid %d, action type %d total sz %d action sz %d",
+ msg_fq_meta->flow, msg_fq_meta->portid, action->type, total_sz, sz);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, msg_fq_meta, total_sz,
+ CNXK_REP_MSG_FLOW_QUERY);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto free;
+ }
+
+ rte_free(msg_fq_meta);
+
+ return 0;
+
+free:
+ rte_free(msg_fq_meta);
+fail:
+ return rc;
+}
+
+static int
+process_flow_rule(struct cnxk_rep_dev *rep_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error, cnxk_rep_msg_ack_data_t *adata, cnxk_rep_msg_t msg)
+{
+ cnxk_rep_msg_flow_create_meta_t msg_fc_meta;
+ uint16_t n_pattern, n_action;
+ uint32_t len = 0, rc = 0;
+ void *buffer;
+ size_t size;
+
+ RTE_SET_USED(error);
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ /* Get no of actions and patterns */
+ cnxk_flow_params_count(pattern, actions, &n_pattern, &n_action);
+
+ /* Adding the header */
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ /* Representor port identified as rep_xport queue */
+ msg_fc_meta.portid = rep_dev->u.rep_portid;
+ msg_fc_meta.nb_pattern = n_pattern;
+ msg_fc_meta.nb_action = n_action;
+
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_fc_meta,
+ sizeof(cnxk_rep_msg_flow_create_meta_t), msg);
+
+ /* Populate flow create parameters data */
+ populate_attr_data(buffer, &len, attr);
+ populate_pattern_data(buffer, &len, pattern, n_pattern);
+ populate_action_data(buffer, &len, actions, n_action);
+
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static struct rte_flow *
+cnxk_rep_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ struct rte_flow *flow = NULL;
+ cnxk_rep_msg_ack_data_t adata;
+ int rc = 0;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ rc = process_flow_rule(rep_dev, attr, pattern, actions, error, &adata,
+ CNXK_REP_MSG_FLOW_CREATE);
+ if (!rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0) {
+ rc = (int)adata.u.sval;
+ rte_flow_error_set(error, adata.u.sval, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+ NULL, "Failed to validate flow");
+ goto fail;
+ }
+
+ flow = adata.u.data;
+ if (!flow) {
+ rte_flow_error_set(error, adata.u.sval, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+ NULL, "Failed to create flow");
+ goto fail;
+ }
+ } else {
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to create flow");
+ }
+ plt_rep_dbg("Flow %p created successfully", adata.u.data);
+
+ return flow;
+fail:
+ return NULL;
+}
+
+static int
+cnxk_rep_flow_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc = 0;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ rc = process_flow_rule(rep_dev, attr, pattern, actions, error, &adata,
+ CNXK_REP_MSG_FLOW_VALIDATE);
+ if (!rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0) {
+ rc = (int)adata.u.sval;
+ rte_flow_error_set(error, adata.u.sval, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+ NULL, "Failed to validate flow");
+ goto fail;
+ }
+ } else {
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to validate flow");
+ }
+
+ plt_rep_dbg("Flow %p validated successfully", adata.u.data);
+
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_destroy(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ RTE_SET_USED(error);
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ rc = process_flow_destroy(rep_dev, flow, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to destroy flow");
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data, struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ if (action->type != RTE_FLOW_ACTION_TYPE_COUNT) {
+ rc = -ENOTSUP;
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Only COUNT is supported in query");
+ goto fail;
+ }
+
+ rc = process_flow_query(rep_dev, flow, action, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to query the flow");
+ goto fail;
+ }
+
+ rte_memcpy(data, adata.u.data, adata.size);
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_flush(struct rte_eth_dev *eth_dev, struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ rc = process_flow_flush(rep_dev, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to destroy flow");
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow, FILE *file,
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ RTE_SET_USED(error);
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ rc = process_flow_dump(rep_dev, flow, file, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to destroy flow");
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_isolate(struct rte_eth_dev *eth_dev __rte_unused, int enable __rte_unused,
+ struct rte_flow_error *error)
+{
+ /* If we support, we need to un-install the default mcam
+ * entry for this port.
+ */
+
+ rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Flow isolation not supported");
+
+ return -rte_errno;
+}
+
+struct rte_flow_ops cnxk_rep_flow_ops = {
+ .validate = cnxk_rep_flow_validate,
+ .create = cnxk_rep_flow_create,
+ .destroy = cnxk_rep_flow_destroy,
+ .query = cnxk_rep_flow_query,
+ .flush = cnxk_rep_flow_flush,
+ .isolate = cnxk_rep_flow_isolate,
+ .dev_dump = cnxk_rep_flow_dev_dump,
+};
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
index 554122d7f8..23fd72434c 100644
--- a/drivers/net/cnxk/cnxk_rep_msg.h
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -12,6 +12,10 @@
typedef enum CNXK_TYPE {
CNXK_TYPE_HEADER = 0,
CNXK_TYPE_MSG,
+ CNXK_TYPE_ATTR,
+ CNXK_TYPE_PATTERN,
+ CNXK_TYPE_ACTION,
+ CNXK_TYPE_FLOW
} cnxk_type_t;
typedef enum CNXK_REP_MSG {
@@ -23,6 +27,13 @@ typedef enum CNXK_REP_MSG {
CNXK_REP_MSG_ETH_SET_MAC,
CNXK_REP_MSG_ETH_STATS_GET,
CNXK_REP_MSG_ETH_STATS_CLEAR,
+ /* Flow operation msgs */
+ CNXK_REP_MSG_FLOW_CREATE,
+ CNXK_REP_MSG_FLOW_DESTROY,
+ CNXK_REP_MSG_FLOW_VALIDATE,
+ CNXK_REP_MSG_FLOW_FLUSH,
+ CNXK_REP_MSG_FLOW_DUMP,
+ CNXK_REP_MSG_FLOW_QUERY,
/* End of messaging sequence */
CNXK_REP_MSG_END,
} cnxk_rep_msg_t;
@@ -79,6 +90,53 @@ typedef struct cnxk_rep_msg_eth_stats_meta {
uint16_t portid;
} __rte_packed cnxk_rep_msg_eth_stats_meta_t;
+/* Flow create msg meta */
+typedef struct cnxk_rep_msg_flow_create_meta {
+ uint16_t portid;
+ uint16_t nb_pattern;
+ uint16_t nb_action;
+} __rte_packed cnxk_rep_msg_flow_create_meta_t;
+
+/* Flow destroy msg meta */
+typedef struct cnxk_rep_msg_flow_destroy_meta {
+ uint64_t flow;
+ uint16_t portid;
+} __rte_packed cnxk_rep_msg_flow_destroy_meta_t;
+
+/* Flow flush msg meta */
+typedef struct cnxk_rep_msg_flow_flush_meta {
+ uint16_t portid;
+} __rte_packed cnxk_rep_msg_flow_flush_meta_t;
+
+/* Flow dump msg meta */
+typedef struct cnxk_rep_msg_flow_dump_meta {
+ uint64_t flow;
+ uint16_t portid;
+ uint8_t is_stdout;
+} __rte_packed cnxk_rep_msg_flow_dump_meta_t;
+
+/* Flow query msg meta */
+typedef struct cnxk_rep_msg_flow_query_meta {
+ uint64_t flow;
+ uint16_t portid;
+ uint32_t action_data_sz;
+ uint8_t action_data[];
+} __rte_packed cnxk_rep_msg_flow_query_meta_t;
+
+/* Type pattern meta */
+typedef struct cnxk_pattern_hdr {
+ uint16_t type;
+ uint16_t spec_sz;
+ uint16_t last_sz;
+ uint16_t mask_sz;
+} __rte_packed cnxk_pattern_hdr_t;
+
+/* Type action meta */
+typedef struct cnxk_action_hdr {
+ uint16_t type;
+ uint16_t conf_sz;
+} __rte_packed cnxk_action_hdr_t;
+
void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
uint32_t size);
void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
index 022a5137df..c418ecf383 100644
--- a/drivers/net/cnxk/cnxk_rep_ops.c
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -661,7 +661,8 @@ int
cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops)
{
PLT_SET_USED(ethdev);
- PLT_SET_USED(ops);
+ *ops = &cnxk_rep_flow_ops;
+
return 0;
}
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index 0e7334f5cd..d4b1110f38 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -35,6 +35,7 @@ sources = files(
'cnxk_rep.c',
'cnxk_rep_msg.c',
'cnxk_rep_ops.c',
+ 'cnxk_rep_flow.c',
'cnxk_stats.c',
'cnxk_tm.c',
)
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH 8/9] common/cnxk: support represented port for cnxk
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
` (6 preceding siblings ...)
2023-08-11 16:34 ` [PATCH 7/9] net/cnxk: representor flow ops Harman Kalra
@ 2023-08-11 16:34 ` Harman Kalra
2023-08-11 16:34 ` [PATCH 9/9] net/cnxk: add " Harman Kalra
` (5 subsequent siblings)
13 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-08-11 16:34 UTC (permalink / raw)
To: jerinj, Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao
Cc: dev, Harman Kalra
Adding represented port item an action support for cnxk
device. Flow operations can be performed via representor
ports as well as represented ports.
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_npc.c | 54 ++++++++++++++++------
drivers/common/cnxk/roc_npc.h | 16 ++++++-
drivers/common/cnxk/roc_npc_mcam.c | 69 ++++++++++++++---------------
drivers/common/cnxk/roc_npc_parse.c | 14 ++++++
drivers/common/cnxk/roc_npc_priv.h | 1 +
drivers/net/cnxk/cnxk_flow.c | 27 +++++++++--
drivers/net/cnxk/cnxk_rep.h | 13 ++++++
7 files changed, 140 insertions(+), 54 deletions(-)
diff --git a/drivers/common/cnxk/roc_npc.c b/drivers/common/cnxk/roc_npc.c
index 586bc55791..0a5bc5c2b1 100644
--- a/drivers/common/cnxk/roc_npc.c
+++ b/drivers/common/cnxk/roc_npc.c
@@ -779,10 +779,14 @@ npc_parse_pattern(struct npc *npc, const struct roc_npc_item_info pattern[],
struct roc_npc_flow *flow, struct npc_parse_state *pst)
{
npc_parse_stage_func_t parse_stage_funcs[] = {
- npc_parse_meta_items, npc_parse_mark_item, npc_parse_pre_l2, npc_parse_cpt_hdr,
- npc_parse_higig2_hdr, npc_parse_tx_queue, npc_parse_la, npc_parse_lb,
- npc_parse_lc, npc_parse_ld, npc_parse_le, npc_parse_lf,
- npc_parse_lg, npc_parse_lh,
+ npc_parse_meta_items, npc_parse_represented_port_id,
+ npc_parse_mark_item, npc_parse_pre_l2,
+ npc_parse_cpt_hdr, npc_parse_higig2_hdr,
+ npc_parse_tx_queue, npc_parse_la,
+ npc_parse_lb, npc_parse_lc,
+ npc_parse_ld, npc_parse_le,
+ npc_parse_lf, npc_parse_lg,
+ npc_parse_lh,
};
uint8_t layer = 0;
int key_offset;
@@ -843,11 +847,11 @@ npc_parse_attr(struct npc *npc, const struct roc_npc_attr *attr,
return NPC_ERR_PARAM;
else if (attr->priority >= npc->flow_max_priority)
return NPC_ERR_PARAM;
- else if ((!attr->egress && !attr->ingress) ||
- (attr->egress && attr->ingress))
+ else if ((!attr->egress && !attr->ingress && !attr->transfer) ||
+ (attr->egress && attr->ingress && attr->transfer))
return NPC_ERR_PARAM;
- if (attr->ingress)
+ if (attr->ingress || attr->transfer)
flow->nix_intf = ROC_NPC_INTF_RX;
else
flow->nix_intf = ROC_NPC_INTF_TX;
@@ -1002,15 +1006,18 @@ npc_rss_action_program(struct roc_npc *roc_npc,
struct roc_npc_flow *flow)
{
const struct roc_npc_action_rss *rss;
+ struct roc_npc *npc = roc_npc;
uint32_t rss_grp;
uint8_t alg_idx;
int rc;
+ if (flow->has_rep)
+ npc = roc_npc->rep_npc;
+
for (; actions->type != ROC_NPC_ACTION_TYPE_END; actions++) {
if (actions->type == ROC_NPC_ACTION_TYPE_RSS) {
rss = (const struct roc_npc_action_rss *)actions->conf;
- rc = npc_rss_action_configure(roc_npc, rss, &alg_idx,
- &rss_grp, flow->mcam_id);
+ rc = npc_rss_action_configure(npc, rss, &alg_idx, &rss_grp, flow->mcam_id);
if (rc)
return rc;
@@ -1448,6 +1455,17 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
memset(flow, 0, sizeof(*flow));
memset(&parse_state, 0, sizeof(parse_state));
+ flow->port_id = -1;
+ if (roc_npc->rep_npc) {
+ flow->rep_channel = roc_nix_to_nix_priv(roc_npc->rep_npc->roc_nix)->rx_chan_base;
+ flow->rep_pf_func = roc_npc->rep_pf_func;
+ flow->rep_mbox = roc_npc_to_npc_priv(roc_npc->rep_npc)->mbox;
+ flow->has_rep = true;
+ flow->is_rep_vf = !roc_nix_is_pf(roc_npc->rep_npc->roc_nix);
+ flow->port_id = roc_npc->rep_port_id;
+ flow->rep_npc = roc_npc_to_npc_priv(roc_npc->rep_npc);
+ }
+
parse_state.dst_pf_func = dst_pf_func;
rc = npc_parse_rule(roc_npc, attr, pattern, actions, flow, &parse_state);
@@ -1475,6 +1493,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
*errcode = rc;
goto set_rss_failed;
}
+ roc_npc->rep_npc = NULL;
if (flow->use_pre_alloc == 0)
list = &npc->flow_list[flow->priority];
@@ -1484,6 +1503,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
TAILQ_FOREACH(flow_iter, list, next) {
if (flow_iter->mcam_id > flow->mcam_id) {
TAILQ_INSERT_BEFORE(flow_iter, flow, next);
+ roc_npc->rep_npc = NULL;
return flow;
}
}
@@ -1491,6 +1511,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
return flow;
set_rss_failed:
+ roc_npc->rep_npc = NULL;
if (flow->use_pre_alloc == 0) {
rc = roc_npc_mcam_free_entry(roc_npc, flow->mcam_id);
if (rc != 0) {
@@ -1502,6 +1523,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
npc_inline_dev_ipsec_action_free(npc, flow);
}
err_exit:
+ roc_npc->rep_npc = NULL;
plt_free(flow);
return NULL;
}
@@ -1509,15 +1531,19 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
int
npc_rss_group_free(struct npc *npc, struct roc_npc_flow *flow)
{
+ struct npc *lnpc = npc;
uint32_t rss_grp;
+ if (flow->has_rep)
+ lnpc = flow->rep_npc;
+
if ((flow->npc_action & 0xF) == NIX_RX_ACTIONOP_RSS) {
rss_grp = (flow->npc_action >> NPC_RSS_ACT_GRP_OFFSET) &
NPC_RSS_ACT_GRP_MASK;
if (rss_grp == 0 || rss_grp >= npc->rss_grps)
return -EINVAL;
- plt_bitmap_clear(npc->rss_grp_entries, rss_grp);
+ plt_bitmap_clear(lnpc->rss_grp_entries, rss_grp);
}
return 0;
@@ -1591,7 +1617,7 @@ roc_npc_flow_destroy(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
}
void
-roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc)
+roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc, int rep_port_id)
{
struct npc *npc = roc_npc_to_npc_priv(roc_npc);
struct roc_npc_flow *flow_iter;
@@ -1605,12 +1631,14 @@ roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc)
/* List in ascending order of mcam entries */
TAILQ_FOREACH(flow_iter, list, next) {
- roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
+ if (rep_port_id == -1 || rep_port_id == flow_iter->port_id)
+ roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
}
}
TAILQ_FOREACH(flow_iter, &npc->ipsec_list, next) {
- roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
+ if (rep_port_id == -1 || rep_port_id == flow_iter->port_id)
+ roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
}
}
diff --git a/drivers/common/cnxk/roc_npc.h b/drivers/common/cnxk/roc_npc.h
index 2ada774934..0f50c55175 100644
--- a/drivers/common/cnxk/roc_npc.h
+++ b/drivers/common/cnxk/roc_npc.h
@@ -41,6 +41,7 @@ enum roc_npc_item_type {
ROC_NPC_ITEM_TYPE_MARK,
ROC_NPC_ITEM_TYPE_TX_QUEUE,
ROC_NPC_ITEM_TYPE_IPV6_ROUTING_EXT,
+ ROC_NPC_ITEM_TYPE_REPRESENTED_PORT,
ROC_NPC_ITEM_TYPE_END,
};
@@ -272,7 +273,8 @@ struct roc_npc_attr {
uint32_t priority; /**< Rule priority level within group. */
uint32_t ingress : 1; /**< Rule applies to ingress traffic. */
uint32_t egress : 1; /**< Rule applies to egress traffic. */
- uint32_t reserved : 30; /**< Reserved, must be zero. */
+ uint32_t transfer : 1; /**< Rule applies to transfer traffic. */
+ uint32_t reserved : 29; /**< Reserved, must be zero. */
};
struct roc_npc_flow_dump_data {
@@ -312,6 +314,13 @@ struct roc_npc_flow {
uint16_t match_id;
uint8_t is_inline_dev;
bool use_pre_alloc;
+ uint16_t rep_pf_func;
+ uint16_t rep_channel;
+ struct mbox *rep_mbox;
+ bool has_rep;
+ bool is_rep_vf;
+ struct npc *rep_npc;
+ int port_id;
TAILQ_ENTRY(roc_npc_flow) next;
};
@@ -366,6 +375,9 @@ struct roc_npc {
bool is_sdp_mask_set;
uint16_t sdp_channel;
uint16_t sdp_channel_mask;
+ struct roc_npc *rep_npc;
+ uint16_t rep_pf_func;
+ int rep_port_id;
#define ROC_NPC_MEM_SZ (6 * 1024)
uint8_t reserved[ROC_NPC_MEM_SZ];
@@ -401,7 +413,7 @@ int __roc_api roc_npc_mcam_clear_counter(struct roc_npc *roc_npc, uint32_t ctr_i
int __roc_api roc_npc_inl_mcam_read_counter(uint32_t ctr_id, uint64_t *count);
int __roc_api roc_npc_inl_mcam_clear_counter(uint32_t ctr_id);
int __roc_api roc_npc_mcam_free_all_resources(struct roc_npc *roc_npc);
-void __roc_api roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc);
+void __roc_api roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc, int rep_port_id);
void __roc_api roc_npc_flow_mcam_dump(FILE *file, struct roc_npc *roc_npc,
struct roc_npc_flow *mcam);
int __roc_api roc_npc_mark_actions_get(struct roc_npc *roc_npc);
diff --git a/drivers/common/cnxk/roc_npc_mcam.c b/drivers/common/cnxk/roc_npc_mcam.c
index 62e0ce21b2..d25e82c652 100644
--- a/drivers/common/cnxk/roc_npc_mcam.c
+++ b/drivers/common/cnxk/roc_npc_mcam.c
@@ -143,8 +143,8 @@ npc_lid_lt_in_kex(struct npc *npc, uint8_t lid, uint8_t lt)
}
static void
-npc_construct_ldata_mask(struct npc *npc, struct plt_bitmap *bmap, uint8_t lid,
- uint8_t lt, uint8_t ld)
+npc_construct_ldata_mask(struct npc *npc, struct plt_bitmap *bmap, uint8_t lid, uint8_t lt,
+ uint8_t ld)
{
struct npc_xtract_info *x_info, *infoflag;
int hdr_off, keylen;
@@ -197,8 +197,7 @@ npc_construct_ldata_mask(struct npc *npc, struct plt_bitmap *bmap, uint8_t lid,
* @param len length of the match
*/
static bool
-npc_is_kex_enabled(struct npc *npc, uint8_t lid, uint8_t lt, int offset,
- int len)
+npc_is_kex_enabled(struct npc *npc, uint8_t lid, uint8_t lt, int offset, int len)
{
struct plt_bitmap *bmap;
uint32_t bmap_sz;
@@ -349,8 +348,8 @@ npc_mcam_alloc_entries(struct mbox *mbox, int ref_mcam, int *alloc_entry, int re
}
int
-npc_mcam_alloc_entry(struct npc *npc, struct roc_npc_flow *mcam,
- struct roc_npc_flow *ref_mcam, int prio, int *resp_count)
+npc_mcam_alloc_entry(struct npc *npc, struct roc_npc_flow *mcam, struct roc_npc_flow *ref_mcam,
+ int prio, int *resp_count)
{
struct npc_mcam_alloc_entry_req *req;
struct npc_mcam_alloc_entry_rsp *rsp;
@@ -450,22 +449,17 @@ npc_mcam_write_entry(struct mbox *mbox, struct roc_npc_flow *mcam)
static void
npc_mcam_process_mkex_cfg(struct npc *npc, struct npc_get_kex_cfg_rsp *kex_rsp)
{
- volatile uint64_t(
- *q)[NPC_MAX_INTF][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD];
+ volatile uint64_t(*q)[NPC_MAX_INTF][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD];
struct npc_xtract_info *x_info = NULL;
int lid, lt, ld, fl, ix;
npc_dxcfg_t *p;
uint64_t keyw;
uint64_t val;
- npc->keyx_supp_nmask[NPC_MCAM_RX] =
- kex_rsp->rx_keyx_cfg & 0x7fffffffULL;
- npc->keyx_supp_nmask[NPC_MCAM_TX] =
- kex_rsp->tx_keyx_cfg & 0x7fffffffULL;
- npc->keyx_len[NPC_MCAM_RX] =
- npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_RX]);
- npc->keyx_len[NPC_MCAM_TX] =
- npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_TX]);
+ npc->keyx_supp_nmask[NPC_MCAM_RX] = kex_rsp->rx_keyx_cfg & 0x7fffffffULL;
+ npc->keyx_supp_nmask[NPC_MCAM_TX] = kex_rsp->tx_keyx_cfg & 0x7fffffffULL;
+ npc->keyx_len[NPC_MCAM_RX] = npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_RX]);
+ npc->keyx_len[NPC_MCAM_TX] = npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_TX]);
keyw = (kex_rsp->rx_keyx_cfg >> 32) & 0x7ULL;
npc->keyw[NPC_MCAM_RX] = keyw;
@@ -485,8 +479,7 @@ npc_mcam_process_mkex_cfg(struct npc *npc, struct npc_get_kex_cfg_rsp *kex_rsp)
/* Update LID, LT and LDATA cfg */
p = &npc->prx_dxcfg;
- q = (volatile uint64_t(*)[][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD])(
- &kex_rsp->intf_lid_lt_ld);
+ q = (volatile uint64_t(*)[][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD])(&kex_rsp->intf_lid_lt_ld);
for (ix = 0; ix < NPC_MAX_INTF; ix++) {
for (lid = 0; lid < NPC_MAX_LID; lid++) {
for (lt = 0; lt < NPC_MAX_LT; lt++) {
@@ -539,8 +532,7 @@ npc_mcam_fetch_kex_cfg(struct npc *npc)
goto done;
}
- mbox_memcpy((char *)npc->profile_name, kex_rsp->mkex_pfl_name,
- MKEX_NAME_LEN);
+ mbox_memcpy((char *)npc->profile_name, kex_rsp->mkex_pfl_name, MKEX_NAME_LEN);
npc->exact_match_ena = (kex_rsp->rx_keyx_cfg >> 40) & 0xF;
npc_mcam_process_mkex_cfg(npc, kex_rsp);
@@ -551,9 +543,8 @@ npc_mcam_fetch_kex_cfg(struct npc *npc)
}
static void
-npc_mcam_set_channel(struct roc_npc_flow *flow,
- struct npc_mcam_write_entry_req *req, uint16_t channel,
- uint16_t chan_mask, bool is_second_pass)
+npc_mcam_set_channel(struct roc_npc_flow *flow, struct npc_mcam_write_entry_req *req,
+ uint16_t channel, uint16_t chan_mask, bool is_second_pass)
{
uint16_t chan = 0, mask = 0;
@@ -733,6 +724,14 @@ npc_mcam_alloc_and_write(struct npc *npc, struct roc_npc_flow *flow, struct npc_
npc_mcam_set_channel(flow, req, inl_dev->channel, inl_dev->chan_mask,
false);
+ } else if (flow->has_rep) {
+ pf_func = flow->rep_pf_func;
+ req->entry_data.action &= ~(GENMASK(19, 4));
+ req->entry_data.action |= (uint64_t)pf_func << 4;
+ flow->npc_action &= ~(GENMASK(19, 4));
+ flow->npc_action |= (uint64_t)pf_func << 4;
+ npc_mcam_set_channel(flow, req, flow->rep_channel, (BIT_ULL(12) - 1),
+ false);
} else if (npc->is_sdp_link) {
npc_mcam_set_channel(flow, req, npc->sdp_channel, npc->sdp_channel_mask,
pst->is_second_pass_rule);
@@ -789,9 +788,8 @@ npc_set_vlan_ltype(struct npc_parse_state *pst)
uint64_t val, mask;
uint8_t lb_offset;
- lb_offset =
- __builtin_popcount(pst->npc->keyx_supp_nmask[pst->nix_intf] &
- ((1ULL << NPC_LTYPE_LB_OFFSET) - 1));
+ lb_offset = __builtin_popcount(pst->npc->keyx_supp_nmask[pst->nix_intf] &
+ ((1ULL << NPC_LTYPE_LB_OFFSET) - 1));
lb_offset *= 4;
mask = ~((0xfULL << lb_offset));
@@ -811,9 +809,8 @@ npc_set_ipv6ext_ltype_mask(struct npc_parse_state *pst)
uint8_t lc_offset, lcflag_offset;
uint64_t val, mask;
- lc_offset =
- __builtin_popcount(pst->npc->keyx_supp_nmask[pst->nix_intf] &
- ((1ULL << NPC_LTYPE_LC_OFFSET) - 1));
+ lc_offset = __builtin_popcount(pst->npc->keyx_supp_nmask[pst->nix_intf] &
+ ((1ULL << NPC_LTYPE_LC_OFFSET) - 1));
lc_offset *= 4;
mask = ~((0xfULL << lc_offset));
@@ -903,13 +900,11 @@ npc_program_mcam(struct npc *npc, struct npc_parse_state *pst, bool mcam_alloc)
data_off = 0;
index++;
}
- key_data[index] |=
- ((uint64_t)data << data_off);
+ key_data[index] |= ((uint64_t)data << data_off);
if (lt == 0)
mask = 0;
- key_mask[index] |=
- ((uint64_t)mask << data_off);
+ key_mask[index] |= ((uint64_t)mask << data_off);
data_off += 4;
}
}
@@ -934,8 +929,12 @@ npc_program_mcam(struct npc *npc, struct npc_parse_state *pst, bool mcam_alloc)
(pst->flow->npc_action & NIX_RX_ACTIONOP_UCAST_IPSEC))
skip_base_rule = true;
- if (pst->is_vf && pst->flow->nix_intf == NIX_INTF_RX && !skip_base_rule) {
- mbox = mbox_get(npc->mbox);
+ if ((pst->is_vf || pst->flow->is_rep_vf) && pst->flow->nix_intf == NIX_INTF_RX &&
+ !skip_base_rule) {
+ if (pst->flow->has_rep)
+ mbox = mbox_get(pst->flow->rep_mbox);
+ else
+ mbox = mbox_get(npc->mbox);
(void)mbox_alloc_msg_npc_read_base_steer_rule(mbox);
rc = mbox_process_msg(mbox, (void *)&base_rule_rsp);
if (rc) {
diff --git a/drivers/common/cnxk/roc_npc_parse.c b/drivers/common/cnxk/roc_npc_parse.c
index ecd1b3e13b..f59a053761 100644
--- a/drivers/common/cnxk/roc_npc_parse.c
+++ b/drivers/common/cnxk/roc_npc_parse.c
@@ -35,6 +35,20 @@ npc_parse_mark_item(struct npc_parse_state *pst)
return 0;
}
+int
+npc_parse_represented_port_id(struct npc_parse_state *pst)
+{
+ if (pst->pattern->type != ROC_NPC_ITEM_TYPE_REPRESENTED_PORT)
+ return 0;
+
+ if (pst->flow->nix_intf != NIX_INTF_RX)
+ return -EINVAL;
+
+ pst->pattern++;
+
+ return 0;
+}
+
static int
npc_flow_raw_item_prepare(const struct roc_npc_flow_item_raw *raw_spec,
const struct roc_npc_flow_item_raw *raw_mask,
diff --git a/drivers/common/cnxk/roc_npc_priv.h b/drivers/common/cnxk/roc_npc_priv.h
index 593dca353b..50c54b895d 100644
--- a/drivers/common/cnxk/roc_npc_priv.h
+++ b/drivers/common/cnxk/roc_npc_priv.h
@@ -448,6 +448,7 @@ int npc_mask_is_supported(const char *mask, const char *hw_mask, int len);
int npc_parse_item_basic(const struct roc_npc_item_info *item, struct npc_parse_item_info *info);
int npc_parse_meta_items(struct npc_parse_state *pst);
int npc_parse_mark_item(struct npc_parse_state *pst);
+int npc_parse_represented_port_id(struct npc_parse_state *pst);
int npc_parse_pre_l2(struct npc_parse_state *pst);
int npc_parse_higig2_hdr(struct npc_parse_state *pst);
int npc_parse_cpt_hdr(struct npc_parse_state *pst);
diff --git a/drivers/net/cnxk/cnxk_flow.c b/drivers/net/cnxk/cnxk_flow.c
index 970daec035..c08b09338d 100644
--- a/drivers/net/cnxk/cnxk_flow.c
+++ b/drivers/net/cnxk/cnxk_flow.c
@@ -187,9 +187,28 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
plt_err("Output port not under same driver");
goto err_exit;
}
- hw_dst = portid_eth_dev->data->dev_private;
- roc_npc_dst = &hw_dst->npc;
- *dst_pf_func = roc_npc_dst->pf_func;
+
+ if (cnxk_ethdev_is_representor(if_name)) {
+ struct rte_eth_dev *rep_eth_dev = portid_eth_dev;
+ struct rte_flow_action_mark *act_mark;
+ struct cnxk_rep_dev *rep_dev;
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ *dst_pf_func = rep_dev->pf_func;
+ plt_rep_dbg("Representor port %d act port %d rep_dev->pf_func 0x%x",
+ port_act->id, act_ethdev->port_id, rep_dev->pf_func);
+
+ /* Add Mark action */
+ i++;
+ act_mark = plt_zmalloc(sizeof(struct rte_flow_action_mark), 0);
+ act_mark->id = 1;
+ in_actions[i].type = ROC_NPC_ACTION_TYPE_MARK;
+ in_actions[i].conf = (struct rte_flow_action_mark *)act_mark;
+ } else {
+ hw_dst = portid_eth_dev->data->dev_private;
+ roc_npc_dst = &hw_dst->npc;
+ *dst_pf_func = roc_npc_dst->pf_func;
+ }
break;
case RTE_FLOW_ACTION_TYPE_QUEUE:
@@ -477,7 +496,7 @@ cnxk_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
return -EINVAL;
}
- roc_npc_flow_dump(file, npc);
+ roc_npc_flow_dump(file, npc, -1);
return 0;
}
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index 4886527f83..1b3de65329 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -6,6 +6,8 @@
#ifndef __CNXK_REP_H__
#define __CNXK_REP_H__
+#include <regex.h>
+
#define CNXK_REP_XPORT_VDEV_CFG_MZ "rep_xport_vdev_cfg"
#define CNXK_REP_XPORT_VDEV_DEVARGS "role=server"
#define CNXK_REP_XPORT_VDEV_NAME "net_memif"
@@ -64,6 +66,17 @@ cnxk_rep_xport_eth_dev(uint16_t portid)
return &rte_eth_devices[portid];
}
+static inline int
+cnxk_ethdev_is_representor(const char *if_name)
+{
+ regex_t regex;
+ int val;
+
+ val = regcomp(®ex, "net_.*_representor_.*", 0);
+ val = regexec(®ex, if_name, 0, NULL, 0);
+ return (val == 0);
+}
+
/* Prototypes */
int cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct rte_eth_dev *pf_ethdev,
struct rte_eth_devargs *eth_da);
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH 9/9] net/cnxk: add represented port for cnxk
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
` (7 preceding siblings ...)
2023-08-11 16:34 ` [PATCH 8/9] common/cnxk: support represented port for cnxk Harman Kalra
@ 2023-08-11 16:34 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (4 subsequent siblings)
13 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-08-11 16:34 UTC (permalink / raw)
To: jerinj, Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao
Cc: dev, Harman Kalra
Adding represented port item matching and action support
for cnxk device.
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
doc/guides/nics/features/cnxk.ini | 1 +
doc/guides/nics/features/cnxk_vec.ini | 2 +
doc/guides/nics/features/cnxk_vf.ini | 2 +
drivers/net/cnxk/cnxk_flow.c | 161 ++++++++++++++++----------
4 files changed, 107 insertions(+), 59 deletions(-)
diff --git a/doc/guides/nics/features/cnxk.ini b/doc/guides/nics/features/cnxk.ini
index f5ff692c27..3b74c6739e 100644
--- a/doc/guides/nics/features/cnxk.ini
+++ b/doc/guides/nics/features/cnxk.ini
@@ -79,6 +79,7 @@ udp = Y
vlan = Y
vxlan = Y
vxlan_gpe = Y
+represented_port = Y
[rte_flow actions]
count = Y
diff --git a/doc/guides/nics/features/cnxk_vec.ini b/doc/guides/nics/features/cnxk_vec.ini
index e2cac64e4b..b33eaaf5d9 100644
--- a/doc/guides/nics/features/cnxk_vec.ini
+++ b/doc/guides/nics/features/cnxk_vec.ini
@@ -73,6 +73,7 @@ udp = Y
vlan = Y
vxlan = Y
vxlan_gpe = Y
+represented_port = Y
[rte_flow actions]
count = Y
@@ -86,5 +87,6 @@ of_set_vlan_vid = Y
pf = Y
queue = Y
rss = Y
+represented_port = Y
security = Y
vf = Y
diff --git a/doc/guides/nics/features/cnxk_vf.ini b/doc/guides/nics/features/cnxk_vf.ini
index 5579007831..70d7772b92 100644
--- a/doc/guides/nics/features/cnxk_vf.ini
+++ b/doc/guides/nics/features/cnxk_vf.ini
@@ -70,6 +70,7 @@ udp = Y
vlan = Y
vxlan = Y
vxlan_gpe = Y
+represented_port = Y
[rte_flow actions]
count = Y
@@ -83,6 +84,7 @@ of_set_vlan_pcp = Y
of_set_vlan_vid = Y
pf = Y
queue = Y
+represented_port = Y
rss = Y
security = Y
skip_cman = Y
diff --git a/drivers/net/cnxk/cnxk_flow.c b/drivers/net/cnxk/cnxk_flow.c
index c08b09338d..d8213f92e8 100644
--- a/drivers/net/cnxk/cnxk_flow.c
+++ b/drivers/net/cnxk/cnxk_flow.c
@@ -4,69 +4,48 @@
#include <cnxk_flow.h>
const struct cnxk_rte_flow_term_info term[] = {
- [RTE_FLOW_ITEM_TYPE_ETH] = {ROC_NPC_ITEM_TYPE_ETH,
- sizeof(struct rte_flow_item_eth)},
- [RTE_FLOW_ITEM_TYPE_VLAN] = {ROC_NPC_ITEM_TYPE_VLAN,
- sizeof(struct rte_flow_item_vlan)},
- [RTE_FLOW_ITEM_TYPE_E_TAG] = {ROC_NPC_ITEM_TYPE_E_TAG,
- sizeof(struct rte_flow_item_e_tag)},
- [RTE_FLOW_ITEM_TYPE_IPV4] = {ROC_NPC_ITEM_TYPE_IPV4,
- sizeof(struct rte_flow_item_ipv4)},
- [RTE_FLOW_ITEM_TYPE_IPV6] = {ROC_NPC_ITEM_TYPE_IPV6,
- sizeof(struct rte_flow_item_ipv6)},
- [RTE_FLOW_ITEM_TYPE_IPV6_FRAG_EXT] = {
- ROC_NPC_ITEM_TYPE_IPV6_FRAG_EXT,
- sizeof(struct rte_flow_item_ipv6_frag_ext)},
- [RTE_FLOW_ITEM_TYPE_ARP_ETH_IPV4] = {
- ROC_NPC_ITEM_TYPE_ARP_ETH_IPV4,
- sizeof(struct rte_flow_item_arp_eth_ipv4)},
- [RTE_FLOW_ITEM_TYPE_MPLS] = {ROC_NPC_ITEM_TYPE_MPLS,
- sizeof(struct rte_flow_item_mpls)},
- [RTE_FLOW_ITEM_TYPE_ICMP] = {ROC_NPC_ITEM_TYPE_ICMP,
- sizeof(struct rte_flow_item_icmp)},
- [RTE_FLOW_ITEM_TYPE_UDP] = {ROC_NPC_ITEM_TYPE_UDP,
- sizeof(struct rte_flow_item_udp)},
- [RTE_FLOW_ITEM_TYPE_TCP] = {ROC_NPC_ITEM_TYPE_TCP,
- sizeof(struct rte_flow_item_tcp)},
- [RTE_FLOW_ITEM_TYPE_SCTP] = {ROC_NPC_ITEM_TYPE_SCTP,
- sizeof(struct rte_flow_item_sctp)},
- [RTE_FLOW_ITEM_TYPE_ESP] = {ROC_NPC_ITEM_TYPE_ESP,
- sizeof(struct rte_flow_item_esp)},
- [RTE_FLOW_ITEM_TYPE_GRE] = {ROC_NPC_ITEM_TYPE_GRE,
- sizeof(struct rte_flow_item_gre)},
- [RTE_FLOW_ITEM_TYPE_NVGRE] = {ROC_NPC_ITEM_TYPE_NVGRE,
- sizeof(struct rte_flow_item_nvgre)},
- [RTE_FLOW_ITEM_TYPE_VXLAN] = {ROC_NPC_ITEM_TYPE_VXLAN,
- sizeof(struct rte_flow_item_vxlan)},
- [RTE_FLOW_ITEM_TYPE_GTPC] = {ROC_NPC_ITEM_TYPE_GTPC,
- sizeof(struct rte_flow_item_gtp)},
- [RTE_FLOW_ITEM_TYPE_GTPU] = {ROC_NPC_ITEM_TYPE_GTPU,
- sizeof(struct rte_flow_item_gtp)},
+ [RTE_FLOW_ITEM_TYPE_ETH] = {ROC_NPC_ITEM_TYPE_ETH, sizeof(struct rte_flow_item_eth)},
+ [RTE_FLOW_ITEM_TYPE_VLAN] = {ROC_NPC_ITEM_TYPE_VLAN, sizeof(struct rte_flow_item_vlan)},
+ [RTE_FLOW_ITEM_TYPE_E_TAG] = {ROC_NPC_ITEM_TYPE_E_TAG, sizeof(struct rte_flow_item_e_tag)},
+ [RTE_FLOW_ITEM_TYPE_IPV4] = {ROC_NPC_ITEM_TYPE_IPV4, sizeof(struct rte_flow_item_ipv4)},
+ [RTE_FLOW_ITEM_TYPE_IPV6] = {ROC_NPC_ITEM_TYPE_IPV6, sizeof(struct rte_flow_item_ipv6)},
+ [RTE_FLOW_ITEM_TYPE_IPV6_FRAG_EXT] = {ROC_NPC_ITEM_TYPE_IPV6_FRAG_EXT,
+ sizeof(struct rte_flow_item_ipv6_frag_ext)},
+ [RTE_FLOW_ITEM_TYPE_ARP_ETH_IPV4] = {ROC_NPC_ITEM_TYPE_ARP_ETH_IPV4,
+ sizeof(struct rte_flow_item_arp_eth_ipv4)},
+ [RTE_FLOW_ITEM_TYPE_MPLS] = {ROC_NPC_ITEM_TYPE_MPLS, sizeof(struct rte_flow_item_mpls)},
+ [RTE_FLOW_ITEM_TYPE_ICMP] = {ROC_NPC_ITEM_TYPE_ICMP, sizeof(struct rte_flow_item_icmp)},
+ [RTE_FLOW_ITEM_TYPE_UDP] = {ROC_NPC_ITEM_TYPE_UDP, sizeof(struct rte_flow_item_udp)},
+ [RTE_FLOW_ITEM_TYPE_TCP] = {ROC_NPC_ITEM_TYPE_TCP, sizeof(struct rte_flow_item_tcp)},
+ [RTE_FLOW_ITEM_TYPE_SCTP] = {ROC_NPC_ITEM_TYPE_SCTP, sizeof(struct rte_flow_item_sctp)},
+ [RTE_FLOW_ITEM_TYPE_ESP] = {ROC_NPC_ITEM_TYPE_ESP, sizeof(struct rte_flow_item_esp)},
+ [RTE_FLOW_ITEM_TYPE_GRE] = {ROC_NPC_ITEM_TYPE_GRE, sizeof(struct rte_flow_item_gre)},
+ [RTE_FLOW_ITEM_TYPE_NVGRE] = {ROC_NPC_ITEM_TYPE_NVGRE, sizeof(struct rte_flow_item_nvgre)},
+ [RTE_FLOW_ITEM_TYPE_VXLAN] = {ROC_NPC_ITEM_TYPE_VXLAN, sizeof(struct rte_flow_item_vxlan)},
+ [RTE_FLOW_ITEM_TYPE_GTPC] = {ROC_NPC_ITEM_TYPE_GTPC, sizeof(struct rte_flow_item_gtp)},
+ [RTE_FLOW_ITEM_TYPE_GTPU] = {ROC_NPC_ITEM_TYPE_GTPU, sizeof(struct rte_flow_item_gtp)},
[RTE_FLOW_ITEM_TYPE_GENEVE] = {ROC_NPC_ITEM_TYPE_GENEVE,
sizeof(struct rte_flow_item_geneve)},
- [RTE_FLOW_ITEM_TYPE_VXLAN_GPE] = {
- ROC_NPC_ITEM_TYPE_VXLAN_GPE,
- sizeof(struct rte_flow_item_vxlan_gpe)},
+ [RTE_FLOW_ITEM_TYPE_VXLAN_GPE] = {ROC_NPC_ITEM_TYPE_VXLAN_GPE,
+ sizeof(struct rte_flow_item_vxlan_gpe)},
[RTE_FLOW_ITEM_TYPE_IPV6_EXT] = {ROC_NPC_ITEM_TYPE_IPV6_EXT,
sizeof(struct rte_flow_item_ipv6_ext)},
[RTE_FLOW_ITEM_TYPE_VOID] = {ROC_NPC_ITEM_TYPE_VOID, 0},
[RTE_FLOW_ITEM_TYPE_ANY] = {ROC_NPC_ITEM_TYPE_ANY, 0},
- [RTE_FLOW_ITEM_TYPE_GRE_KEY] = {ROC_NPC_ITEM_TYPE_GRE_KEY,
- sizeof(uint32_t)},
+ [RTE_FLOW_ITEM_TYPE_GRE_KEY] = {ROC_NPC_ITEM_TYPE_GRE_KEY, sizeof(uint32_t)},
[RTE_FLOW_ITEM_TYPE_HIGIG2] = {ROC_NPC_ITEM_TYPE_HIGIG2,
sizeof(struct rte_flow_item_higig2_hdr)},
- [RTE_FLOW_ITEM_TYPE_RAW] = {ROC_NPC_ITEM_TYPE_RAW,
- sizeof(struct rte_flow_item_raw)},
- [RTE_FLOW_ITEM_TYPE_MARK] = {ROC_NPC_ITEM_TYPE_MARK,
- sizeof(struct rte_flow_item_mark)},
+ [RTE_FLOW_ITEM_TYPE_RAW] = {ROC_NPC_ITEM_TYPE_RAW, sizeof(struct rte_flow_item_raw)},
+ [RTE_FLOW_ITEM_TYPE_MARK] = {ROC_NPC_ITEM_TYPE_MARK, sizeof(struct rte_flow_item_mark)},
[RTE_FLOW_ITEM_TYPE_IPV6_ROUTING_EXT] = {ROC_NPC_ITEM_TYPE_IPV6_ROUTING_EXT,
- sizeof(struct rte_flow_item_ipv6_routing_ext)},
+ sizeof(struct rte_flow_item_ipv6_routing_ext)},
[RTE_FLOW_ITEM_TYPE_TX_QUEUE] = {ROC_NPC_ITEM_TYPE_TX_QUEUE,
- sizeof(struct rte_flow_item_tx_queue)}};
+ sizeof(struct rte_flow_item_tx_queue)},
+ [RTE_FLOW_ITEM_TYPE_REPRESENTED_PORT] = {ROC_NPC_ITEM_TYPE_REPRESENTED_PORT,
+ sizeof(struct rte_flow_item_ethdev)}};
static int
-npc_rss_action_validate(struct rte_eth_dev *eth_dev,
- const struct rte_flow_attr *attr,
+npc_rss_action_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
const struct rte_flow_action *act)
{
const struct rte_flow_action_rss *rss;
@@ -274,28 +253,92 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
}
static int
-cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
- const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
- struct roc_npc_attr *in_attr, struct roc_npc_item_info in_pattern[],
- struct roc_npc_action in_actions[], uint32_t *flowkey_cfg, uint16_t *dst_pf_func)
+cnxk_map_pattern(struct rte_eth_dev *eth_dev, const struct rte_flow_item pattern[],
+ struct roc_npc_item_info in_pattern[])
{
+ struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
+ const struct rte_flow_item_ethdev *rep_eth_dev;
+ struct rte_eth_dev *portid_eth_dev;
+ char if_name[RTE_ETH_NAME_MAX_LEN];
+ struct cnxk_eth_dev *hw_dst;
int i = 0;
- in_attr->priority = attr->priority;
- in_attr->ingress = attr->ingress;
- in_attr->egress = attr->egress;
-
while (pattern->type != RTE_FLOW_ITEM_TYPE_END) {
in_pattern[i].spec = pattern->spec;
in_pattern[i].last = pattern->last;
in_pattern[i].mask = pattern->mask;
in_pattern[i].type = term[pattern->type].item_type;
in_pattern[i].size = term[pattern->type].item_size;
+ if (pattern->type == RTE_FLOW_ITEM_TYPE_REPRESENTED_PORT) {
+ rep_eth_dev = (const struct rte_flow_item_ethdev *)pattern->spec;
+ if (rte_eth_dev_get_name_by_port(rep_eth_dev->port_id, if_name)) {
+ plt_err("Name not found for output port id");
+ goto fail;
+ }
+ portid_eth_dev = rte_eth_dev_allocated(if_name);
+ if (!portid_eth_dev) {
+ plt_err("eth_dev not found for output port id");
+ goto fail;
+ }
+ if (strcmp(portid_eth_dev->device->driver->name,
+ eth_dev->device->driver->name) != 0) {
+ plt_err("Output port not under same driver");
+ goto fail;
+ }
+ if (cnxk_ethdev_is_representor(if_name)) {
+ /* Case where represented port not part of same
+ * app and represented by a representor port.
+ */
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eth_dev *pf_dev;
+
+ rep_dev = cnxk_rep_pmd_priv(portid_eth_dev);
+ pf_dev = cnxk_eth_pmd_priv(rep_dev->parent_dev);
+ dev->npc.rep_npc = &pf_dev->npc;
+ dev->npc.rep_port_id = rep_eth_dev->port_id;
+ dev->npc.rep_pf_func = rep_dev->pf_func;
+ plt_rep_dbg("Represented port %d act port %d rep_dev->pf_func 0x%x",
+ rep_eth_dev->port_id, eth_dev->data->port_id,
+ rep_dev->pf_func);
+ } else {
+ /* Case where represented port part of same app
+ * as PF.
+ */
+ hw_dst = portid_eth_dev->data->dev_private;
+ dev->npc.rep_npc = &hw_dst->npc;
+ dev->npc.rep_port_id = rep_eth_dev->port_id;
+ dev->npc.rep_pf_func = hw_dst->npc.pf_func;
+ }
+ }
pattern++;
i++;
}
in_pattern[i].type = ROC_NPC_ITEM_TYPE_END;
+ return 0;
+fail:
+ return -EINVAL;
+}
+
+static int
+cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct roc_npc_attr *in_attr, struct roc_npc_item_info in_pattern[],
+ struct roc_npc_action in_actions[], uint32_t *flowkey_cfg, uint16_t *dst_pf_func)
+{
+ int rc;
+
+ in_attr->priority = attr->priority;
+ in_attr->ingress = attr->ingress;
+ in_attr->egress = attr->egress;
+ in_attr->transfer = attr->transfer;
+
+ rc = cnxk_map_pattern(eth_dev, pattern, in_pattern);
+ if (rc) {
+ plt_err("Failed to map pattern list");
+ return rc;
+ }
+
return cnxk_map_actions(eth_dev, attr, actions, in_actions, flowkey_cfg, dst_pf_func);
}
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 00/24] net/cnxk: support for port representors
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
` (8 preceding siblings ...)
2023-08-11 16:34 ` [PATCH 9/9] net/cnxk: add " Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 01/24] common/cnxk: add support for representors Harman Kalra
` (23 more replies)
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (3 subsequent siblings)
13 siblings, 24 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
Cc: dev, jerinj, Harman Kalra
Introducing port representor support to CNXK drivers by adding virtual ethernet
ports providing a logical representation in DPDK for physical function(PF) or
SR-IOV virtual function (VF) devices for control and monitoring.
These port representor ethdev instances can be spawned on an as needed basis
through configuration parameters passed to the driver of the underlying
base device using devargs ``-a <base PCI BDF>,representor=pf*vf*``
In case of exception path (i.e. until the flow definition is offloaded to the
hardware), packets transmitted by the VFs shall be received by these
representor port, while packets transmitted by representor ports shall be
received by respective VFs.
On receiving the VF traffic via these representor ports, applications holding
these representor ports can decide to offload the traffic flow into the HW.
Henceforth the matching traffic shall be directly steered to the respective
VFs without being received by the application.
Current virtual representor port PMD supports following operations:
- Get represented port statistics
- Set mac address
- Flow operations - create, validate, destroy, query, flush, dump
Changes since V1:
* Updated communication layer between representor and represented port.
* Added support for native represented ports
* Port representor and represented port item and action support
* Build failure fixes
--
Harman Kalra (24):
common/cnxk: add support for representors
net/cnxk: implementing eswitch device
net/cnxk: eswitch HW resource configuration
net/cnxk: eswitch devargs parsing
net/cnxk: probing representor ports
common/cnxk: common NPC changes for eswitch
common/cnxk: interface to update VLAN TPID
net/cnxk: eswitch flow configurations
net/cnxk: eswitch fastpath routines
net/cnxk: add representor control plane
common/cnxk: representee notification callback
net/cnxk: handling representee notification
net/cnxk: representor ethdev ops
common/cnxk: get representees ethernet stats
net/cnxk: ethernet statistic for representor
common/cnxk: base support for eswitch VF
net/cnxk: eswitch VF as ethernet device
common/cnxk: support port representor and represented port
net/cnxk: add represented port pattern and action
net/cnxk: add port representor pattern and action
net/cnxk: generalize flow operation APIs
net/cnxk: flow create on representor ports
net/cnxk: other flow operations
doc: port representors in cnxk
MAINTAINERS | 1 +
doc/guides/nics/cnxk.rst | 58 ++
doc/guides/nics/features/cnxk.ini | 3 +
doc/guides/nics/features/cnxk_vf.ini | 4 +
drivers/common/cnxk/meson.build | 1 +
drivers/common/cnxk/roc_api.h | 3 +
drivers/common/cnxk/roc_constants.h | 2 +
drivers/common/cnxk/roc_dev.c | 25 +
drivers/common/cnxk/roc_dev_priv.h | 3 +
drivers/common/cnxk/roc_eswitch.c | 368 ++++++++++
drivers/common/cnxk/roc_eswitch.h | 33 +
drivers/common/cnxk/roc_mbox.c | 2 +
drivers/common/cnxk/roc_mbox.h | 73 +-
drivers/common/cnxk/roc_nix.c | 46 +-
drivers/common/cnxk/roc_nix.h | 4 +
drivers/common/cnxk/roc_nix_priv.h | 5 +-
drivers/common/cnxk/roc_nix_vlan.c | 23 +-
drivers/common/cnxk/roc_npc.c | 89 ++-
drivers/common/cnxk/roc_npc.h | 18 +-
drivers/common/cnxk/roc_npc_mcam.c | 64 +-
drivers/common/cnxk/roc_npc_parse.c | 28 +-
drivers/common/cnxk/roc_npc_priv.h | 5 +-
drivers/common/cnxk/roc_platform.c | 2 +
drivers/common/cnxk/roc_platform.h | 4 +
drivers/common/cnxk/version.map | 14 +
drivers/net/cnxk/cn10k_ethdev.c | 1 +
drivers/net/cnxk/cnxk_eswitch.c | 871 ++++++++++++++++++++++++
drivers/net/cnxk/cnxk_eswitch.h | 213 ++++++
drivers/net/cnxk/cnxk_eswitch_devargs.c | 237 +++++++
drivers/net/cnxk/cnxk_eswitch_flow.c | 445 ++++++++++++
drivers/net/cnxk/cnxk_eswitch_rxtx.c | 212 ++++++
drivers/net/cnxk/cnxk_ethdev.c | 39 +-
drivers/net/cnxk/cnxk_ethdev.h | 3 +
drivers/net/cnxk/cnxk_ethdev_ops.c | 4 +
drivers/net/cnxk/cnxk_flow.c | 521 +++++++++++---
drivers/net/cnxk/cnxk_flow.h | 28 +-
drivers/net/cnxk/cnxk_link.c | 3 +-
drivers/net/cnxk/cnxk_rep.c | 555 +++++++++++++++
drivers/net/cnxk/cnxk_rep.h | 141 ++++
drivers/net/cnxk/cnxk_rep_flow.c | 813 ++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.c | 823 ++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 169 +++++
drivers/net/cnxk/cnxk_rep_ops.c | 715 +++++++++++++++++++
drivers/net/cnxk/meson.build | 8 +
44 files changed, 6498 insertions(+), 181 deletions(-)
create mode 100644 drivers/common/cnxk/roc_eswitch.c
create mode 100644 drivers/common/cnxk/roc_eswitch.h
create mode 100644 drivers/net/cnxk/cnxk_eswitch.c
create mode 100644 drivers/net/cnxk/cnxk_eswitch.h
create mode 100644 drivers/net/cnxk/cnxk_eswitch_devargs.c
create mode 100644 drivers/net/cnxk/cnxk_eswitch_flow.c
create mode 100644 drivers/net/cnxk/cnxk_eswitch_rxtx.c
create mode 100644 drivers/net/cnxk/cnxk_rep.c
create mode 100644 drivers/net/cnxk/cnxk_rep.h
create mode 100644 drivers/net/cnxk/cnxk_rep_flow.c
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.c
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.h
create mode 100644 drivers/net/cnxk/cnxk_rep_ops.c
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 01/24] common/cnxk: add support for representors
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 02/24] net/cnxk: implementing eswitch device Harman Kalra
` (22 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
- Mailbox for registering base device behind all representors
- Registering debug log type for representors
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_constants.h | 1 +
drivers/common/cnxk/roc_mbox.h | 8 ++++++++
drivers/common/cnxk/roc_nix.c | 31 +++++++++++++++++++++++++++++
drivers/common/cnxk/roc_nix.h | 3 +++
drivers/common/cnxk/roc_platform.c | 2 ++
drivers/common/cnxk/roc_platform.h | 4 ++++
drivers/common/cnxk/version.map | 3 +++
7 files changed, 52 insertions(+)
diff --git a/drivers/common/cnxk/roc_constants.h b/drivers/common/cnxk/roc_constants.h
index 291b6a4bc9..cb4edbea58 100644
--- a/drivers/common/cnxk/roc_constants.h
+++ b/drivers/common/cnxk/roc_constants.h
@@ -43,6 +43,7 @@
#define PCI_DEVID_CNXK_RVU_NIX_INL_VF 0xA0F1
#define PCI_DEVID_CNXK_RVU_REE_PF 0xA0f4
#define PCI_DEVID_CNXK_RVU_REE_VF 0xA0f5
+#define PCI_DEVID_CNXK_RVU_ESWITCH_PF 0xA0E0
#define PCI_DEVID_CN9K_CGX 0xA059
#define PCI_DEVID_CN10K_RPM 0xA060
diff --git a/drivers/common/cnxk/roc_mbox.h b/drivers/common/cnxk/roc_mbox.h
index 3257a370bc..b7e2f43d45 100644
--- a/drivers/common/cnxk/roc_mbox.h
+++ b/drivers/common/cnxk/roc_mbox.h
@@ -68,6 +68,7 @@ struct mbox_msghdr {
M(NDC_SYNC_OP, 0x009, ndc_sync_op, ndc_sync_op, msg_rsp) \
M(LMTST_TBL_SETUP, 0x00a, lmtst_tbl_setup, lmtst_tbl_setup_req, \
msg_rsp) \
+ M(GET_REP_CNT, 0x00d, get_rep_cnt, msg_req, get_rep_cnt_rsp) \
/* CGX mbox IDs (range 0x200 - 0x3FF) */ \
M(CGX_START_RXTX, 0x200, cgx_start_rxtx, msg_req, msg_rsp) \
M(CGX_STOP_RXTX, 0x201, cgx_stop_rxtx, msg_req, msg_rsp) \
@@ -546,6 +547,13 @@ struct lmtst_tbl_setup_req {
uint64_t __io rsvd[2]; /* Future use */
};
+#define MAX_PFVF_REP 64
+struct get_rep_cnt_rsp {
+ struct mbox_msghdr hdr;
+ uint16_t __io rep_cnt;
+ uint16_t __io rep_pfvf_map[MAX_PFVF_REP];
+};
+
/* CGX mbox message formats */
/* CGX mailbox error codes
* Range 1101 - 1200.
diff --git a/drivers/common/cnxk/roc_nix.c b/drivers/common/cnxk/roc_nix.c
index f64933a1d9..7e327a7e6e 100644
--- a/drivers/common/cnxk/roc_nix.c
+++ b/drivers/common/cnxk/roc_nix.c
@@ -531,3 +531,34 @@ roc_nix_dev_fini(struct roc_nix *roc_nix)
rc |= dev_fini(&nix->dev, nix->pci_dev);
return rc;
}
+
+int
+roc_nix_max_rep_count(struct roc_nix *roc_nix)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+ struct mbox *mbox = mbox_get(dev->mbox);
+ struct get_rep_cnt_rsp *rsp;
+ struct msg_req *req;
+ int rc, i;
+
+ req = mbox_alloc_msg_get_rep_cnt(mbox);
+ if (!req) {
+ rc = -ENOSPC;
+ goto exit;
+ }
+
+ req->hdr.pcifunc = roc_nix_get_pf_func(roc_nix);
+
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ roc_nix->rep_cnt = rsp->rep_cnt;
+ for (i = 0; i < rsp->rep_cnt; i++)
+ roc_nix->rep_pfvf_map[i] = rsp->rep_pfvf_map[i];
+
+exit:
+ mbox_put(mbox);
+ return rc;
+}
diff --git a/drivers/common/cnxk/roc_nix.h b/drivers/common/cnxk/roc_nix.h
index 84e6fc3df5..b369335fc4 100644
--- a/drivers/common/cnxk/roc_nix.h
+++ b/drivers/common/cnxk/roc_nix.h
@@ -483,6 +483,8 @@ struct roc_nix {
uint32_t buf_sz;
uint64_t meta_aura_handle;
uintptr_t meta_mempool;
+ uint16_t rep_cnt;
+ uint16_t rep_pfvf_map[MAX_PFVF_REP];
TAILQ_ENTRY(roc_nix) next;
#define ROC_NIX_MEM_SZ (6 * 1070)
@@ -1013,4 +1015,5 @@ int __roc_api roc_nix_mcast_list_setup(struct mbox *mbox, uint8_t intf, int nb_e
uint16_t *pf_funcs, uint16_t *channels, uint32_t *rqs,
uint32_t *grp_index, uint32_t *start_index);
int __roc_api roc_nix_mcast_list_free(struct mbox *mbox, uint32_t mcast_grp_index);
+int __roc_api roc_nix_max_rep_count(struct roc_nix *roc_nix);
#endif /* _ROC_NIX_H_ */
diff --git a/drivers/common/cnxk/roc_platform.c b/drivers/common/cnxk/roc_platform.c
index 15cbb6d68f..181902a585 100644
--- a/drivers/common/cnxk/roc_platform.c
+++ b/drivers/common/cnxk/roc_platform.c
@@ -96,4 +96,6 @@ RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_sso, NOTICE);
RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_tim, NOTICE);
RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_tm, NOTICE);
RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_dpi, NOTICE);
+RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_rep, NOTICE);
+RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_esw, NOTICE);
RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_ree, NOTICE);
diff --git a/drivers/common/cnxk/roc_platform.h b/drivers/common/cnxk/roc_platform.h
index ba23b2e0d7..e08eb7f6ba 100644
--- a/drivers/common/cnxk/roc_platform.h
+++ b/drivers/common/cnxk/roc_platform.h
@@ -264,6 +264,8 @@ extern int cnxk_logtype_tim;
extern int cnxk_logtype_tm;
extern int cnxk_logtype_ree;
extern int cnxk_logtype_dpi;
+extern int cnxk_logtype_rep;
+extern int cnxk_logtype_esw;
#define plt_err(fmt, args...) \
RTE_LOG(ERR, PMD, "%s():%u " fmt "\n", __func__, __LINE__, ##args)
@@ -293,6 +295,8 @@ extern int cnxk_logtype_dpi;
#define plt_tm_dbg(fmt, ...) plt_dbg(tm, fmt, ##__VA_ARGS__)
#define plt_ree_dbg(fmt, ...) plt_dbg(ree, fmt, ##__VA_ARGS__)
#define plt_dpi_dbg(fmt, ...) plt_dbg(dpi, fmt, ##__VA_ARGS__)
+#define plt_rep_dbg(fmt, ...) plt_dbg(rep, fmt, ##__VA_ARGS__)
+#define plt_esw_dbg(fmt, ...) plt_dbg(esw, fmt, ##__VA_ARGS__)
/* Datapath logs */
#define plt_dp_err(fmt, args...) \
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index 7b6afa63a9..bd28803013 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -8,12 +8,14 @@ INTERNAL {
cnxk_logtype_base;
cnxk_logtype_cpt;
cnxk_logtype_dpi;
+ cnxk_logtype_esw;
cnxk_logtype_mbox;
cnxk_logtype_ml;
cnxk_logtype_nix;
cnxk_logtype_npa;
cnxk_logtype_npc;
cnxk_logtype_ree;
+ cnxk_logtype_rep;
cnxk_logtype_sso;
cnxk_logtype_tim;
cnxk_logtype_tm;
@@ -216,6 +218,7 @@ INTERNAL {
roc_nix_get_base_chan;
roc_nix_get_pf;
roc_nix_get_pf_func;
+ roc_nix_max_rep_count;
roc_nix_get_rx_chan_cnt;
roc_nix_get_vf;
roc_nix_get_vwqe_interval;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 02/24] net/cnxk: implementing eswitch device
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
2023-12-19 17:39 ` [PATCH v2 01/24] common/cnxk: add support for representors Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2024-01-04 12:30 ` Jerin Jacob
2023-12-19 17:39 ` [PATCH v2 03/24] net/cnxk: eswitch HW resource configuration Harman Kalra
` (21 subsequent siblings)
23 siblings, 1 reply; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra, Anatoly Burakov
Cc: dev, jerinj
Eswitch device is a parent or base device behind all the representors,
acting as transport layer between representors and representees
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 465 ++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_eswitch.h | 103 +++++++
drivers/net/cnxk/meson.build | 1 +
3 files changed, 569 insertions(+)
create mode 100644 drivers/net/cnxk/cnxk_eswitch.c
create mode 100644 drivers/net/cnxk/cnxk_eswitch.h
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
new file mode 100644
index 0000000000..51110a762d
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -0,0 +1,465 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <cnxk_eswitch.h>
+
+#define CNXK_NIX_DEF_SQ_COUNT 512
+
+static int
+cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
+{
+ struct cnxk_eswitch_dev *eswitch_dev;
+ struct roc_nix *nix;
+ int rc = 0;
+
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return 0;
+
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+
+ /* Check if this device is hosting common resource */
+ nix = roc_idev_npa_nix_get();
+ if (!nix || nix->pci_dev != pci_dev) {
+ rc = -EINVAL;
+ goto exit;
+ }
+
+ /* Try nix fini now */
+ rc = roc_nix_dev_fini(&eswitch_dev->nix);
+ if (rc == -EAGAIN) {
+ plt_info("%s: common resource in use by other devices", pci_dev->name);
+ goto exit;
+ } else if (rc) {
+ plt_err("Failed in nix dev fini, rc=%d", rc);
+ goto exit;
+ }
+
+ rte_free(eswitch_dev);
+exit:
+ return rc;
+}
+
+static int
+eswitch_dev_nix_flow_ctrl_set(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ /* TODO enable flow control */
+ return 0;
+ enum roc_nix_fc_mode mode_map[] = {ROC_NIX_FC_NONE, ROC_NIX_FC_RX, ROC_NIX_FC_TX,
+ ROC_NIX_FC_FULL};
+ struct roc_nix *nix = &eswitch_dev->nix;
+ struct roc_nix_fc_cfg fc_cfg;
+ uint8_t rx_pause, tx_pause;
+ struct roc_nix_sq *sq;
+ struct roc_nix_cq *cq;
+ struct roc_nix_rq *rq;
+ uint8_t tc;
+ int rc, i;
+
+ rx_pause = 1;
+ tx_pause = 1;
+
+ /* Check if TX pause frame is already enabled or not */
+ tc = tx_pause ? 0 : ROC_NIX_PFC_CLASS_INVALID;
+
+ for (i = 0; i < eswitch_dev->nb_rxq; i++) {
+ memset(&fc_cfg, 0, sizeof(struct roc_nix_fc_cfg));
+
+ rq = &eswitch_dev->rxq[i].rqs;
+ cq = &eswitch_dev->cxq[i].cqs;
+
+ fc_cfg.type = ROC_NIX_FC_RQ_CFG;
+ fc_cfg.rq_cfg.enable = !!tx_pause;
+ fc_cfg.rq_cfg.tc = tc;
+ fc_cfg.rq_cfg.rq = rq->qid;
+ fc_cfg.rq_cfg.pool = rq->aura_handle;
+ fc_cfg.rq_cfg.spb_pool = rq->spb_aura_handle;
+ fc_cfg.rq_cfg.cq_drop = cq->drop_thresh;
+ fc_cfg.rq_cfg.pool_drop_pct = ROC_NIX_AURA_THRESH;
+
+ rc = roc_nix_fc_config_set(nix, &fc_cfg);
+ if (rc)
+ return rc;
+ }
+
+ /* Check if RX pause frame is enabled or not */
+ tc = rx_pause ? 0 : ROC_NIX_PFC_CLASS_INVALID;
+ for (i = 0; i < eswitch_dev->nb_txq; i++) {
+ memset(&fc_cfg, 0, sizeof(struct roc_nix_fc_cfg));
+
+ sq = &eswitch_dev->txq[i].sqs;
+
+ fc_cfg.type = ROC_NIX_FC_TM_CFG;
+ fc_cfg.tm_cfg.sq = sq->qid;
+ fc_cfg.tm_cfg.tc = tc;
+ fc_cfg.tm_cfg.enable = !!rx_pause;
+ rc = roc_nix_fc_config_set(nix, &fc_cfg);
+ if (rc && rc != EEXIST)
+ return rc;
+ }
+
+ rc = roc_nix_fc_mode_set(nix, mode_map[ROC_NIX_FC_FULL]);
+ if (rc)
+ return rc;
+
+ return rc;
+}
+
+int
+cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ int rc;
+
+ /* Update Flow control configuration */
+ rc = eswitch_dev_nix_flow_ctrl_set(eswitch_dev);
+ if (rc) {
+ plt_err("Failed to enable flow control. error code(%d)", rc);
+ goto done;
+ }
+
+ /* Enable Rx in NPC */
+ rc = roc_nix_npc_rx_ena_dis(&eswitch_dev->nix, true);
+ if (rc) {
+ plt_err("Failed to enable NPC rx %d", rc);
+ goto done;
+ }
+
+ rc = roc_npc_mcam_enable_all_entries(&eswitch_dev->npc, 1);
+ if (rc) {
+ plt_err("Failed to enable NPC entries %d", rc);
+ goto done;
+ }
+
+done:
+ return 0;
+}
+
+int
+cnxk_eswitch_txq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_sq *sq = &eswitch_dev->txq[qid].sqs;
+ int rc = -EINVAL;
+
+ if (eswitch_dev->txq[qid].state == CNXK_ESWITCH_QUEUE_STATE_STARTED)
+ return 0;
+
+ if (eswitch_dev->txq[qid].state != CNXK_ESWITCH_QUEUE_STATE_CONFIGURED) {
+ plt_err("Eswitch txq %d not configured yet", qid);
+ goto done;
+ }
+
+ rc = roc_nix_tm_sq_aura_fc(sq, true);
+ if (rc) {
+ plt_err("Failed to enable sq aura fc, txq=%u, rc=%d", qid, rc);
+ goto done;
+ }
+
+ eswitch_dev->txq[qid].state = CNXK_ESWITCH_QUEUE_STATE_STARTED;
+done:
+ return rc;
+}
+
+int
+cnxk_eswitch_txq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_sq *sq = &eswitch_dev->txq[qid].sqs;
+ int rc = -EINVAL;
+
+ if (eswitch_dev->txq[qid].state == CNXK_ESWITCH_QUEUE_STATE_STOPPED ||
+ eswitch_dev->txq[qid].state == CNXK_ESWITCH_QUEUE_STATE_RELEASED)
+ return 0;
+
+ if (eswitch_dev->txq[qid].state != CNXK_ESWITCH_QUEUE_STATE_STARTED) {
+ plt_err("Eswitch txq %d not started", qid);
+ goto done;
+ }
+
+ rc = roc_nix_tm_sq_aura_fc(sq, false);
+ if (rc) {
+ plt_err("Failed to disable sqb aura fc, txq=%u, rc=%d", qid, rc);
+ goto done;
+ }
+
+ eswitch_dev->txq[qid].state = CNXK_ESWITCH_QUEUE_STATE_STOPPED;
+done:
+ return rc;
+}
+
+int
+cnxk_eswitch_rxq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_rq *rq = &eswitch_dev->rxq[qid].rqs;
+ int rc = -EINVAL;
+
+ if (eswitch_dev->rxq[qid].state == CNXK_ESWITCH_QUEUE_STATE_STARTED)
+ return 0;
+
+ if (eswitch_dev->rxq[qid].state != CNXK_ESWITCH_QUEUE_STATE_CONFIGURED) {
+ plt_err("Eswitch rxq %d not configured yet", qid);
+ goto done;
+ }
+
+ rc = roc_nix_rq_ena_dis(rq, true);
+ if (rc) {
+ plt_err("Failed to enable rxq=%u, rc=%d", qid, rc);
+ goto done;
+ }
+
+ eswitch_dev->rxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_STARTED;
+done:
+ return rc;
+}
+
+int
+cnxk_eswitch_rxq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_rq *rq = &eswitch_dev->rxq[qid].rqs;
+ int rc = -EINVAL;
+
+ if (eswitch_dev->rxq[qid].state == CNXK_ESWITCH_QUEUE_STATE_STOPPED ||
+ eswitch_dev->rxq[qid].state == CNXK_ESWITCH_QUEUE_STATE_RELEASED)
+ return 0;
+
+ if (eswitch_dev->rxq[qid].state != CNXK_ESWITCH_QUEUE_STATE_STARTED) {
+ plt_err("Eswitch rxq %d not started", qid);
+ goto done;
+ }
+
+ rc = roc_nix_rq_ena_dis(rq, false);
+ if (rc) {
+ plt_err("Failed to disable rxq=%u, rc=%d", qid, rc);
+ goto done;
+ }
+
+ eswitch_dev->rxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_STOPPED;
+done:
+ return rc;
+}
+
+int
+cnxk_eswitch_rxq_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_rq *rq;
+ struct roc_nix_cq *cq;
+ int rc;
+
+ if (eswitch_dev->rxq[qid].state == CNXK_ESWITCH_QUEUE_STATE_RELEASED)
+ return 0;
+
+ /* Cleanup ROC SQ */
+ rq = &eswitch_dev->rxq[qid].rqs;
+ rc = roc_nix_rq_fini(rq);
+ if (rc) {
+ plt_err("Failed to cleanup sq, rc=%d", rc);
+ goto fail;
+ }
+
+ eswitch_dev->rxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_RELEASED;
+
+ /* Cleanup ROC CQ */
+ cq = &eswitch_dev->cxq[qid].cqs;
+ rc = roc_nix_cq_fini(cq);
+ if (rc) {
+ plt_err("Failed to cleanup cq, rc=%d", rc);
+ goto fail;
+ }
+
+ eswitch_dev->cxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_RELEASED;
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_rxq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint16_t nb_desc,
+ const struct rte_eth_rxconf *rx_conf, struct rte_mempool *mp)
+{
+ struct roc_nix *nix = &eswitch_dev->nix;
+ struct rte_mempool *lpb_pool = mp;
+ struct rte_mempool_ops *ops;
+ const char *platform_ops;
+ struct roc_nix_rq *rq;
+ struct roc_nix_cq *cq;
+ uint16_t first_skip;
+ int rc = -EINVAL;
+
+ if (eswitch_dev->rxq[qid].state != CNXK_ESWITCH_QUEUE_STATE_RELEASED ||
+ eswitch_dev->cxq[qid].state != CNXK_ESWITCH_QUEUE_STATE_RELEASED) {
+ plt_err("Queue %d is in invalid state %d, cannot be setup", qid,
+ eswitch_dev->txq[qid].state);
+ goto fail;
+ }
+
+ RTE_SET_USED(rx_conf);
+ platform_ops = rte_mbuf_platform_mempool_ops();
+ /* This driver needs cnxk_npa mempool ops to work */
+ ops = rte_mempool_get_ops(lpb_pool->ops_index);
+ if (strncmp(ops->name, platform_ops, RTE_MEMPOOL_OPS_NAMESIZE)) {
+ plt_err("mempool ops should be of cnxk_npa type");
+ goto fail;
+ }
+
+ if (lpb_pool->pool_id == 0) {
+ plt_err("Invalid pool_id");
+ goto fail;
+ }
+
+ /* Setup ROC CQ */
+ cq = &eswitch_dev->cxq[qid].cqs;
+ memset(cq, 0, sizeof(struct roc_nix_cq));
+ cq->qid = qid;
+ cq->nb_desc = nb_desc;
+ rc = roc_nix_cq_init(nix, cq);
+ if (rc) {
+ plt_err("Failed to init roc cq for rq=%d, rc=%d", qid, rc);
+ goto fail;
+ }
+ eswitch_dev->cxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_CONFIGURED;
+
+ /* Setup ROC RQ */
+ rq = &eswitch_dev->rxq[qid].rqs;
+ memset(rq, 0, sizeof(struct roc_nix_rq));
+ rq->qid = qid;
+ rq->cqid = cq->qid;
+ rq->aura_handle = lpb_pool->pool_id;
+ rq->flow_tag_width = 32;
+ rq->sso_ena = false;
+
+ /* Calculate first mbuf skip */
+ first_skip = (sizeof(struct rte_mbuf));
+ first_skip += RTE_PKTMBUF_HEADROOM;
+ first_skip += rte_pktmbuf_priv_size(lpb_pool);
+ rq->first_skip = first_skip;
+ rq->later_skip = sizeof(struct rte_mbuf) + rte_pktmbuf_priv_size(lpb_pool);
+ rq->lpb_size = lpb_pool->elt_size;
+ if (roc_errata_nix_no_meta_aura())
+ rq->lpb_drop_ena = true;
+
+ rc = roc_nix_rq_init(nix, rq, true);
+ if (rc) {
+ plt_err("Failed to init roc rq for rq=%d, rc=%d", qid, rc);
+ goto cq_fini;
+ }
+ eswitch_dev->rxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_CONFIGURED;
+
+ return 0;
+cq_fini:
+ rc |= roc_nix_cq_fini(cq);
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_txq_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_sq *sq;
+ int rc = 0;
+
+ if (eswitch_dev->txq[qid].state == CNXK_ESWITCH_QUEUE_STATE_RELEASED)
+ return 0;
+
+ /* Cleanup ROC SQ */
+ sq = &eswitch_dev->txq[qid].sqs;
+ rc = roc_nix_sq_fini(sq);
+ if (rc) {
+ plt_err("Failed to cleanup sq, rc=%d", rc);
+ goto fail;
+ }
+
+ eswitch_dev->txq[qid].state = CNXK_ESWITCH_QUEUE_STATE_RELEASED;
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_txq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint16_t nb_desc,
+ const struct rte_eth_txconf *tx_conf)
+{
+ struct roc_nix_sq *sq;
+ int rc = 0;
+
+ if (eswitch_dev->txq[qid].state != CNXK_ESWITCH_QUEUE_STATE_RELEASED) {
+ plt_err("Queue %d is in invalid state %d, cannot be setup", qid,
+ eswitch_dev->txq[qid].state);
+ rc = -EINVAL;
+ goto fail;
+ }
+ RTE_SET_USED(tx_conf);
+ /* Setup ROC SQ */
+ sq = &eswitch_dev->txq[qid].sqs;
+ memset(sq, 0, sizeof(struct roc_nix_sq));
+ sq->qid = qid;
+ sq->nb_desc = nb_desc;
+ /* TODO: Revisit to enable MSEG nix_sq_max_sqe_sz(dev) */
+ sq->max_sqe_sz = NIX_MAXSQESZ_W8;
+ if (sq->nb_desc >= CNXK_NIX_DEF_SQ_COUNT)
+ sq->fc_hyst_bits = 0x1;
+
+ rc = roc_nix_sq_init(&eswitch_dev->nix, sq);
+ if (rc)
+ plt_err("Failed to init sq=%d, rc=%d", qid, rc);
+
+ eswitch_dev->txq[qid].state = CNXK_ESWITCH_QUEUE_STATE_CONFIGURED;
+
+fail:
+ return rc;
+}
+
+static int
+cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
+{
+ struct cnxk_eswitch_dev *eswitch_dev;
+ const struct rte_memzone *mz = NULL;
+ int rc = -ENOMEM;
+
+ RTE_SET_USED(pci_drv);
+
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ if (!eswitch_dev) {
+ rc = roc_plt_init();
+ if (rc) {
+ plt_err("Failed to initialize platform model, rc=%d", rc);
+ return rc;
+ }
+
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return 0;
+
+ mz = rte_memzone_reserve_aligned(CNXK_REP_ESWITCH_DEV_MZ, sizeof(*eswitch_dev),
+ SOCKET_ID_ANY, 0, RTE_CACHE_LINE_SIZE);
+ if (mz == NULL) {
+ plt_err("Failed to reserve a memzone");
+ goto fail;
+ }
+
+ eswitch_dev = mz->addr;
+ eswitch_dev->pci_dev = pci_dev;
+ }
+
+ /* Spinlock for synchronization between representors traffic and control
+ * messages
+ */
+ rte_spinlock_init(&eswitch_dev->rep_lock);
+
+ return rc;
+fail:
+ return rc;
+}
+
+static const struct rte_pci_id cnxk_eswitch_pci_map[] = {
+ {RTE_PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVID_CNXK_RVU_ESWITCH_PF)},
+ {
+ .vendor_id = 0,
+ },
+};
+
+static struct rte_pci_driver cnxk_eswitch_pci = {
+ .id_table = cnxk_eswitch_pci_map,
+ .drv_flags =
+ RTE_PCI_DRV_NEED_MAPPING | RTE_PCI_DRV_NEED_IOVA_AS_VA | RTE_PCI_DRV_PROBE_AGAIN,
+ .probe = cnxk_eswitch_dev_probe,
+ .remove = cnxk_eswitch_dev_remove,
+};
+
+RTE_PMD_REGISTER_PCI(cnxk_eswitch, cnxk_eswitch_pci);
+RTE_PMD_REGISTER_PCI_TABLE(cnxk_eswitch, cnxk_eswitch_pci_map);
+RTE_PMD_REGISTER_KMOD_DEP(cnxk_eswitch, "vfio-pci");
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
new file mode 100644
index 0000000000..331397021b
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -0,0 +1,103 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#ifndef __CNXK_ESWITCH_H__
+#define __CNXK_ESWITCH_H__
+
+#include <sys/socket.h>
+#include <sys/un.h>
+
+#include <cnxk_ethdev.h>
+
+#include "cn10k_tx.h"
+
+#define CNXK_ESWITCH_CTRL_MSG_SOCK_PATH "/tmp/cxk_rep_ctrl_msg_sock"
+#define CNXK_REP_ESWITCH_DEV_MZ "cnxk_eswitch_dev"
+#define CNXK_ESWITCH_VLAN_TPID 0x8100 /* TODO change */
+#define CNXK_ESWITCH_MAX_TXQ 256
+#define CNXK_ESWITCH_MAX_RXQ 256
+#define CNXK_ESWITCH_LBK_CHAN 63
+#define CNXK_ESWITCH_VFPF_SHIFT 8
+
+#define CNXK_ESWITCH_QUEUE_STATE_RELEASED 0
+#define CNXK_ESWITCH_QUEUE_STATE_CONFIGURED 1
+#define CNXK_ESWITCH_QUEUE_STATE_STARTED 2
+#define CNXK_ESWITCH_QUEUE_STATE_STOPPED 3
+
+struct cnxk_rep_info {
+ struct rte_eth_dev *rep_eth_dev;
+};
+
+struct cnxk_eswitch_txq {
+ struct roc_nix_sq sqs;
+ uint8_t state;
+};
+
+struct cnxk_eswitch_rxq {
+ struct roc_nix_rq rqs;
+ uint8_t state;
+};
+
+struct cnxk_eswitch_cxq {
+ struct roc_nix_cq cqs;
+ uint8_t state;
+};
+
+TAILQ_HEAD(eswitch_flow_list, roc_npc_flow);
+struct cnxk_eswitch_dev {
+ /* Input parameters */
+ struct plt_pci_device *pci_dev;
+ /* ROC NIX */
+ struct roc_nix nix;
+
+ /* ROC NPC */
+ struct roc_npc npc;
+
+ /* ROC NPA */
+ struct rte_mempool *ctrl_chan_pool;
+ const struct plt_memzone *pktmem_mz;
+ uint64_t pkt_aura;
+
+ /* Eswitch RQs, SQs and CQs */
+ struct cnxk_eswitch_txq *txq;
+ struct cnxk_eswitch_rxq *rxq;
+ struct cnxk_eswitch_cxq *cxq;
+
+ /* Configured queue count */
+ uint16_t nb_rxq;
+ uint16_t nb_txq;
+ uint16_t rep_cnt;
+ uint8_t configured;
+
+ /* Port representor fields */
+ rte_spinlock_t rep_lock;
+ uint16_t switch_domain_id;
+ uint16_t eswitch_vdev;
+ struct cnxk_rep_info *rep_info;
+};
+
+static inline struct cnxk_eswitch_dev *
+cnxk_eswitch_pmd_priv(void)
+{
+ const struct rte_memzone *mz;
+
+ mz = rte_memzone_lookup(CNXK_REP_ESWITCH_DEV_MZ);
+ if (!mz)
+ return NULL;
+
+ return mz->addr;
+}
+
+int cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_eswitch_txq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint16_t nb_desc,
+ const struct rte_eth_txconf *tx_conf);
+int cnxk_eswitch_txq_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+int cnxk_eswitch_rxq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint16_t nb_desc,
+ const struct rte_eth_rxconf *rx_conf, struct rte_mempool *mp);
+int cnxk_eswitch_rxq_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+int cnxk_eswitch_rxq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+int cnxk_eswitch_rxq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+int cnxk_eswitch_txq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+int cnxk_eswitch_txq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+#endif /* __CNXK_ESWITCH_H__ */
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index e83f3c9050..012d098f80 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -28,6 +28,7 @@ sources = files(
'cnxk_ethdev_sec.c',
'cnxk_ethdev_telemetry.c',
'cnxk_ethdev_sec_telemetry.c',
+ 'cnxk_eswitch.c',
'cnxk_link.c',
'cnxk_lookup.c',
'cnxk_ptp.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 03/24] net/cnxk: eswitch HW resource configuration
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
2023-12-19 17:39 ` [PATCH v2 01/24] common/cnxk: add support for representors Harman Kalra
2023-12-19 17:39 ` [PATCH v2 02/24] net/cnxk: implementing eswitch device Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2024-01-04 12:34 ` Jerin Jacob
2023-12-19 17:39 ` [PATCH v2 04/24] net/cnxk: eswitch devargs parsing Harman Kalra
` (20 subsequent siblings)
23 siblings, 1 reply; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Configuring the hardware resources used by the eswitch device.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 206 ++++++++++++++++++++++++++++++++
1 file changed, 206 insertions(+)
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index 51110a762d..306edc6037 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -6,6 +6,30 @@
#define CNXK_NIX_DEF_SQ_COUNT 512
+static int
+eswitch_hw_rsrc_cleanup(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ struct roc_nix *nix;
+ int rc = 0;
+
+ nix = &eswitch_dev->nix;
+
+ roc_nix_unregister_queue_irqs(nix);
+ roc_nix_tm_fini(nix);
+ rc = roc_nix_lf_free(nix);
+ if (rc) {
+ plt_err("Failed to cleanup sq, rc %d", rc);
+ goto exit;
+ }
+
+ rte_free(eswitch_dev->txq);
+ rte_free(eswitch_dev->rxq);
+ rte_free(eswitch_dev->cxq);
+
+exit:
+ return rc;
+}
+
static int
cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
{
@@ -18,6 +42,7 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
eswitch_dev = cnxk_eswitch_pmd_priv();
+ eswitch_hw_rsrc_cleanup(eswitch_dev);
/* Check if this device is hosting common resource */
nix = roc_idev_npa_nix_get();
if (!nix || nix->pci_dev != pci_dev) {
@@ -404,6 +429,178 @@ cnxk_eswitch_txq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint1
return rc;
}
+static int
+nix_lf_setup(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ uint16_t nb_rxq, nb_txq, nb_cq;
+ struct roc_nix_fc_cfg fc_cfg;
+ struct roc_nix *nix;
+ uint64_t rx_cfg;
+ void *qs;
+ int rc;
+
+ /* Initialize base roc nix */
+ nix = &eswitch_dev->nix;
+ nix->pci_dev = eswitch_dev->pci_dev;
+ nix->hw_vlan_ins = true;
+ nix->reta_sz = ROC_NIX_RSS_RETA_SZ_256;
+ rc = roc_nix_dev_init(nix);
+ if (rc) {
+ plt_err("Failed to init nix eswitch device, rc=%d(%s)", rc, roc_error_msg_get(rc));
+ goto fail;
+ }
+
+ /* Get the representors count */
+ rc = roc_nix_max_rep_count(&eswitch_dev->nix);
+ if (rc) {
+ plt_err("Failed to get rep cnt, rc=%d(%s)", rc, roc_error_msg_get(rc));
+ goto free_cqs;
+ }
+
+ /* Allocating an NIX LF */
+ nb_rxq = CNXK_ESWITCH_MAX_RXQ;
+ nb_txq = CNXK_ESWITCH_MAX_TXQ;
+ nb_cq = CNXK_ESWITCH_MAX_RXQ;
+ rx_cfg = ROC_NIX_LF_RX_CFG_DIS_APAD;
+ rc = roc_nix_lf_alloc(nix, nb_rxq, nb_txq, rx_cfg);
+ if (rc) {
+ plt_err("lf alloc failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto dev_fini;
+ }
+
+ if (nb_rxq) {
+ /* Allocate memory for eswitch rq's and cq's */
+ qs = plt_zmalloc(sizeof(struct cnxk_eswitch_rxq) * nb_rxq, 0);
+ if (!qs) {
+ plt_err("Failed to alloc eswitch rxq");
+ goto lf_free;
+ }
+ eswitch_dev->rxq = qs;
+ }
+
+ if (nb_txq) {
+ /* Allocate memory for roc sq's */
+ qs = plt_zmalloc(sizeof(struct cnxk_eswitch_txq) * nb_txq, 0);
+ if (!qs) {
+ plt_err("Failed to alloc eswitch txq");
+ goto free_rqs;
+ }
+ eswitch_dev->txq = qs;
+ }
+
+ if (nb_cq) {
+ qs = plt_zmalloc(sizeof(struct cnxk_eswitch_cxq) * nb_cq, 0);
+ if (!qs) {
+ plt_err("Failed to alloc eswitch cxq");
+ goto free_sqs;
+ }
+ eswitch_dev->cxq = qs;
+ }
+
+ eswitch_dev->nb_rxq = nb_rxq;
+ eswitch_dev->nb_txq = nb_txq;
+
+ /* Re-enable NIX LF error interrupts */
+ roc_nix_err_intr_ena_dis(nix, true);
+ roc_nix_ras_intr_ena_dis(nix, true);
+
+ rc = roc_nix_lso_fmt_setup(nix);
+ if (rc) {
+ plt_err("lso setup failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto free_cqs;
+ }
+
+ rc = roc_nix_switch_hdr_set(nix, 0, 0, 0, 0);
+ if (rc) {
+ plt_err("switch hdr set failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto free_cqs;
+ }
+
+ rc = roc_nix_rss_default_setup(nix,
+ FLOW_KEY_TYPE_IPV4 | FLOW_KEY_TYPE_TCP | FLOW_KEY_TYPE_UDP);
+ if (rc) {
+ plt_err("rss default setup failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto free_cqs;
+ }
+
+ rc = roc_nix_tm_init(nix);
+ if (rc) {
+ plt_err("tm failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto free_cqs;
+ }
+
+ /* Register queue IRQs */
+ rc = roc_nix_register_queue_irqs(nix);
+ if (rc) {
+ plt_err("Failed to register queue interrupts rc=%d", rc);
+ goto tm_fini;
+ }
+
+ /* Enable default tree */
+ rc = roc_nix_tm_hierarchy_enable(nix, ROC_NIX_TM_DEFAULT, false);
+ if (rc) {
+ plt_err("tm default hierarchy enable failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto q_irq_fini;
+ }
+
+ /* TODO: Revisit Enable flow control */
+ memset(&fc_cfg, 0, sizeof(struct roc_nix_fc_cfg));
+ fc_cfg.rxchan_cfg.enable = false;
+ rc = roc_nix_fc_config_set(nix, &fc_cfg);
+ if (rc) {
+ plt_err("Failed to setup flow control, rc=%d(%s)", rc, roc_error_msg_get(rc));
+ goto q_irq_fini;
+ }
+
+ roc_nix_fc_mode_get(nix);
+
+ return rc;
+q_irq_fini:
+ roc_nix_unregister_queue_irqs(nix);
+tm_fini:
+ roc_nix_tm_fini(nix);
+free_cqs:
+ rte_free(eswitch_dev->cxq);
+free_sqs:
+ rte_free(eswitch_dev->txq);
+free_rqs:
+ rte_free(eswitch_dev->rxq);
+lf_free:
+ roc_nix_lf_free(nix);
+dev_fini:
+ roc_nix_dev_fini(nix);
+fail:
+ return rc;
+}
+
+static int
+eswitch_hw_rsrc_setup(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ struct roc_nix *nix;
+ int rc;
+
+ nix = &eswitch_dev->nix;
+ rc = nix_lf_setup(eswitch_dev);
+ if (rc) {
+ plt_err("Failed to setup hw rsrc, rc=%d(%s)", rc, roc_error_msg_get(rc));
+ goto fail;
+ }
+
+ /* Initialize roc npc */
+ eswitch_dev->npc.roc_nix = nix;
+ eswitch_dev->npc.flow_max_priority = 3;
+ eswitch_dev->npc.flow_prealloc_size = 1;
+ rc = roc_npc_init(&eswitch_dev->npc);
+ if (rc)
+ goto rsrc_cleanup;
+
+ return rc;
+rsrc_cleanup:
+ eswitch_hw_rsrc_cleanup(eswitch_dev);
+fail:
+ return rc;
+}
+
static int
cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
{
@@ -433,6 +630,12 @@ cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pc
eswitch_dev = mz->addr;
eswitch_dev->pci_dev = pci_dev;
+
+ rc = eswitch_hw_rsrc_setup(eswitch_dev);
+ if (rc) {
+ plt_err("Failed to setup hw rsrc, rc=%d(%s)", rc, roc_error_msg_get(rc));
+ goto free_mem;
+ }
}
/* Spinlock for synchronization between representors traffic and control
@@ -441,6 +644,9 @@ cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pc
rte_spinlock_init(&eswitch_dev->rep_lock);
return rc;
+free_mem:
+ if (mz)
+ rte_memzone_free(mz);
fail:
return rc;
}
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 04/24] net/cnxk: eswitch devargs parsing
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (2 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 03/24] net/cnxk: eswitch HW resource configuration Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 05/24] net/cnxk: probing representor ports Harman Kalra
` (19 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Implementing the devargs parsing logic via which the representors
pattern is provided. These patterns define for which representies
representors shall be created.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 88 +++++++++
drivers/net/cnxk/cnxk_eswitch.h | 52 ++++++
drivers/net/cnxk/cnxk_eswitch_devargs.c | 236 ++++++++++++++++++++++++
drivers/net/cnxk/meson.build | 1 +
4 files changed, 377 insertions(+)
create mode 100644 drivers/net/cnxk/cnxk_eswitch_devargs.c
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index 306edc6037..739a09c034 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -456,6 +456,7 @@ nix_lf_setup(struct cnxk_eswitch_dev *eswitch_dev)
plt_err("Failed to get rep cnt, rc=%d(%s)", rc, roc_error_msg_get(rc));
goto free_cqs;
}
+ eswitch_dev->repr_cnt.max_repr = eswitch_dev->nix.rep_cnt;
/* Allocating an NIX LF */
nb_rxq = CNXK_ESWITCH_MAX_RXQ;
@@ -601,11 +602,73 @@ eswitch_hw_rsrc_setup(struct cnxk_eswitch_dev *eswitch_dev)
return rc;
}
+int
+cnxk_eswitch_representor_info_get(struct cnxk_eswitch_dev *eswitch_dev,
+ struct rte_eth_representor_info *info)
+{
+ struct cnxk_eswitch_devargs *esw_da;
+ int rc = 0, n_entries, i, j = 0, k = 0;
+
+ for (i = 0; i < eswitch_dev->nb_esw_da; i++) {
+ for (j = 0; j < eswitch_dev->esw_da[i].nb_repr_ports; j++)
+ k++;
+ }
+ n_entries = k;
+
+ if (info == NULL)
+ goto out;
+
+ if ((uint32_t)n_entries > info->nb_ranges_alloc)
+ n_entries = info->nb_ranges_alloc;
+
+ k = 0;
+ info->controller = 0;
+ info->pf = 0;
+ for (i = 0; i < eswitch_dev->nb_esw_da; i++) {
+ esw_da = &eswitch_dev->esw_da[i];
+ info->ranges[k].type = esw_da->da.type;
+ switch (esw_da->da.type) {
+ case RTE_ETH_REPRESENTOR_PF:
+ info->ranges[k].controller = 0;
+ info->ranges[k].pf = esw_da->repr_hw_info[0].pfvf;
+ info->ranges[k].vf = 0;
+ info->ranges[k].id_base = info->ranges[i].pf;
+ info->ranges[k].id_end = info->ranges[i].pf;
+ snprintf(info->ranges[k].name, sizeof(info->ranges[k].name), "pf%d",
+ info->ranges[k].pf);
+ k++;
+ break;
+ case RTE_ETH_REPRESENTOR_VF:
+ for (j = 0; j < esw_da->nb_repr_ports; j++) {
+ info->ranges[k].controller = 0;
+ info->ranges[k].pf = esw_da->da.ports[0];
+ info->ranges[k].vf = esw_da->repr_hw_info[j].pfvf;
+ info->ranges[k].id_base = esw_da->repr_hw_info[j].port_id;
+ info->ranges[k].id_end = esw_da->repr_hw_info[j].port_id;
+ snprintf(info->ranges[k].name, sizeof(info->ranges[k].name),
+ "pf%dvf%d", info->ranges[k].pf, info->ranges[k].vf);
+ k++;
+ }
+ break;
+ default:
+ plt_err("Invalid type %d", esw_da->da.type);
+ rc = 0;
+ goto fail;
+ };
+ }
+ info->nb_ranges = k;
+fail:
+ return rc;
+out:
+ return n_entries;
+}
+
static int
cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
{
struct cnxk_eswitch_dev *eswitch_dev;
const struct rte_memzone *mz = NULL;
+ uint16_t num_reps;
int rc = -ENOMEM;
RTE_SET_USED(pci_drv);
@@ -638,12 +701,37 @@ cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pc
}
}
+ if (pci_dev->device.devargs) {
+ rc = cnxk_eswitch_repr_devargs(pci_dev, eswitch_dev);
+ if (rc)
+ goto rsrc_cleanup;
+ }
+
+ if (eswitch_dev->repr_cnt.nb_repr_created > eswitch_dev->repr_cnt.max_repr) {
+ plt_err("Representors to be created %d can be greater than max allowed %d",
+ eswitch_dev->repr_cnt.nb_repr_created, eswitch_dev->repr_cnt.max_repr);
+ rc = -EINVAL;
+ goto rsrc_cleanup;
+ }
+
+ num_reps = eswitch_dev->repr_cnt.nb_repr_created;
+ if (!num_reps) {
+ plt_err("No representors enabled");
+ goto fail;
+ }
+
+ plt_esw_dbg("Max no of reps %d reps to be created %d Eswtch pfunc %x",
+ eswitch_dev->repr_cnt.max_repr, eswitch_dev->repr_cnt.nb_repr_created,
+ roc_nix_get_pf_func(&eswitch_dev->nix));
+
/* Spinlock for synchronization between representors traffic and control
* messages
*/
rte_spinlock_init(&eswitch_dev->rep_lock);
return rc;
+rsrc_cleanup:
+ eswitch_hw_rsrc_cleanup(eswitch_dev);
free_mem:
if (mz)
rte_memzone_free(mz);
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index 331397021b..dcb787cf02 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -25,6 +25,47 @@
#define CNXK_ESWITCH_QUEUE_STATE_STARTED 2
#define CNXK_ESWITCH_QUEUE_STATE_STOPPED 3
+enum cnxk_esw_da_pattern_type {
+ CNXK_ESW_DA_TYPE_LIST = 0,
+ CNXK_ESW_DA_TYPE_PFVF,
+};
+
+struct cnxk_esw_repr_hw_info {
+ /* Representee pcifunc value */
+ uint16_t hw_func;
+ /* rep id in sync with kernel */
+ uint16_t rep_id;
+ /* pf or vf id */
+ uint16_t pfvf;
+ /* representor port id assigned to representee */
+ uint16_t port_id;
+};
+
+/* Structure representing per devarg information - this can be per representee
+ * or range of representee
+ */
+struct cnxk_eswitch_devargs {
+ /* Devargs populated */
+ struct rte_eth_devargs da;
+ /* HW info of representee */
+ struct cnxk_esw_repr_hw_info *repr_hw_info;
+ /* No of representor ports */
+ uint16_t nb_repr_ports;
+ /* Devargs pattern type */
+ enum cnxk_esw_da_pattern_type type;
+};
+
+struct cnxk_eswitch_repr_cnt {
+ /* Max possible representors */
+ uint16_t max_repr;
+ /* Representors to be created as per devargs passed */
+ uint16_t nb_repr_created;
+ /* Representors probed successfully */
+ uint16_t nb_repr_probed;
+ /* Representors started representing a representee */
+ uint16_t nb_repr_started;
+};
+
struct cnxk_rep_info {
struct rte_eth_dev *rep_eth_dev;
};
@@ -70,6 +111,14 @@ struct cnxk_eswitch_dev {
uint16_t rep_cnt;
uint8_t configured;
+ /* Eswitch Representors Devargs */
+ uint16_t nb_esw_da;
+ uint16_t last_probed;
+ struct cnxk_eswitch_devargs esw_da[RTE_MAX_ETHPORTS];
+
+ /* No of representors */
+ struct cnxk_eswitch_repr_cnt repr_cnt;
+
/* Port representor fields */
rte_spinlock_t rep_lock;
uint16_t switch_domain_id;
@@ -90,6 +139,9 @@ cnxk_eswitch_pmd_priv(void)
}
int cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_eswitch_repr_devargs(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_eswitch_representor_info_get(struct cnxk_eswitch_dev *eswitch_dev,
+ struct rte_eth_representor_info *info);
int cnxk_eswitch_txq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint16_t nb_desc,
const struct rte_eth_txconf *tx_conf);
int cnxk_eswitch_txq_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
diff --git a/drivers/net/cnxk/cnxk_eswitch_devargs.c b/drivers/net/cnxk/cnxk_eswitch_devargs.c
new file mode 100644
index 0000000000..f1a1b05a99
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_eswitch_devargs.c
@@ -0,0 +1,236 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <cnxk_eswitch.h>
+
+#define PF_SHIFT 10
+static inline int
+get_hw_func(uint16_t pf, uint16_t vf)
+{
+ return (pf << PF_SHIFT) | vf;
+}
+
+static int
+devargs_enlist(uint16_t *list, uint16_t *len_list, const uint16_t max_list, uint16_t val)
+{
+ uint16_t i;
+
+ for (i = 0; i < *len_list; i++) {
+ if (list[i] == val)
+ return 0;
+ }
+ if (*len_list >= max_list)
+ return -1;
+ list[(*len_list)++] = val;
+ return 0;
+}
+
+static char *
+devargs_process_range(char *str, uint16_t *list, uint16_t *len_list, const uint16_t max_list)
+{
+ uint16_t lo, hi, val;
+ int result, n = 0;
+ char *pos = str;
+
+ result = sscanf(str, "%hu%n-%hu%n", &lo, &n, &hi, &n);
+ if (result == 1) {
+ if (devargs_enlist(list, len_list, max_list, lo) != 0)
+ return NULL;
+ } else if (result == 2) {
+ if (lo > hi)
+ return NULL;
+ for (val = lo; val <= hi; val++) {
+ if (devargs_enlist(list, len_list, max_list, val) != 0)
+ return NULL;
+ }
+ } else {
+ return NULL;
+ }
+
+ return pos + n;
+}
+
+static char *
+devargs_process_list(char *str, uint16_t *list, uint16_t *len_list, const uint16_t max_list)
+{
+ char *pos = str;
+
+ if (*pos == '[')
+ pos++;
+ while (1) {
+ pos = devargs_process_range(pos, list, len_list, max_list);
+ if (pos == NULL)
+ return NULL;
+ if (*pos != ',') /* end of list */
+ break;
+ pos++;
+ }
+ if (*str == '[' && *pos != ']')
+ return NULL;
+ if (*pos == ']')
+ pos++;
+ return pos;
+}
+
+static int
+devargs_parse_representor_ports(char *str, void *data)
+{
+ struct rte_eth_devargs *eth_da = data;
+
+ if (str[0] == 'p' && str[1] == 'f') {
+ eth_da->type = RTE_ETH_REPRESENTOR_PF;
+ str += 2;
+ str = devargs_process_list(str, eth_da->ports, ð_da->nb_ports,
+ RTE_DIM(eth_da->ports));
+ if (str == NULL || str[0] == '\0')
+ goto done;
+ }
+
+ if (str[0] == 'v' && str[1] == 'f') {
+ eth_da->type = RTE_ETH_REPRESENTOR_VF;
+ str += 2;
+ } else if (str[0] == 's' && str[1] == 'f') {
+ eth_da->type = RTE_ETH_REPRESENTOR_SF;
+ str += 2;
+ } else {
+ /* 'pf' must followed by 'vf' or 'sf'. */
+ if (eth_da->type == RTE_ETH_REPRESENTOR_PF) {
+ str = NULL;
+ goto done;
+ }
+ eth_da->type = RTE_ETH_REPRESENTOR_VF;
+ }
+ str = devargs_process_list(str, eth_da->representor_ports, ð_da->nb_representor_ports,
+ RTE_DIM(eth_da->representor_ports));
+done:
+ if (str == NULL)
+ plt_err("wrong representor format: %s", str);
+ return str == NULL ? -1 : 0;
+}
+
+static int
+populate_repr_hw_info(struct cnxk_eswitch_dev *eswitch_dev, struct rte_eth_devargs *eth_da,
+ uint16_t idx)
+{
+ struct cnxk_eswitch_devargs *esw_da = &eswitch_dev->esw_da[idx];
+ uint16_t nb_repr_ports, hw_func;
+ int rc, i, j;
+
+ if (eth_da->type == RTE_ETH_REPRESENTOR_NONE) {
+ plt_err("No representor type found");
+ return -EINVAL;
+ }
+
+ if (eth_da->type != RTE_ETH_REPRESENTOR_VF && eth_da->type != RTE_ETH_REPRESENTOR_PF &&
+ eth_da->type != RTE_ETH_REPRESENTOR_SF) {
+ plt_err("unsupported representor type %d\n", eth_da->type);
+ return -ENOTSUP;
+ }
+
+ nb_repr_ports = (eth_da->type == RTE_ETH_REPRESENTOR_PF) ? eth_da->nb_ports :
+ eth_da->nb_representor_ports;
+ esw_da->nb_repr_ports = nb_repr_ports;
+ /* If plain list is provided as representor pattern */
+ if (eth_da->nb_ports == 0)
+ return 0;
+
+ esw_da->repr_hw_info = plt_zmalloc(nb_repr_ports * sizeof(struct cnxk_esw_repr_hw_info), 0);
+ if (!esw_da->repr_hw_info) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ plt_esw_dbg("Representor param %d has %d pfvf", idx, nb_repr_ports);
+ /* Check if representor can be created for PFVF and populating HW func list */
+ for (i = 0; i < nb_repr_ports; i++) {
+ if (eth_da->type == RTE_ETH_REPRESENTOR_PF)
+ hw_func = get_hw_func(eth_da->ports[0], 0);
+ else
+ hw_func = get_hw_func(eth_da->ports[0], eth_da->representor_ports[i] + 1);
+
+ for (j = 0; j < eswitch_dev->repr_cnt.max_repr; j++) {
+ if (eswitch_dev->nix.rep_pfvf_map[j] == hw_func)
+ break;
+ }
+
+ /* HW func which does not match the map table received from AF, no
+ * representor port is assigned.
+ */
+ if (j == eswitch_dev->repr_cnt.max_repr) {
+ plt_err("Representor port cant be created for PF%dVF%d", eth_da->ports[0],
+ eth_da->representor_ports[i]);
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ esw_da->repr_hw_info[i].hw_func = hw_func;
+ esw_da->repr_hw_info[i].rep_id = j;
+ esw_da->repr_hw_info[i].pfvf = (eth_da->type == RTE_ETH_REPRESENTOR_PF) ?
+ eth_da->ports[0] :
+ eth_da->representor_ports[i];
+ plt_esw_dbg(" HW func %x index %d", hw_func, j);
+ }
+
+ esw_da->type = CNXK_ESW_DA_TYPE_PFVF;
+
+ return 0;
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_repr_devargs(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev)
+{
+ struct rte_devargs *devargs = pci_dev->device.devargs;
+ struct rte_eth_devargs *eth_da;
+ struct rte_kvargs *kvlist;
+ uint32_t i;
+ int rc, j;
+
+ if (devargs == NULL) {
+ plt_err("No devargs passed");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ kvlist = rte_kvargs_parse(devargs->args, NULL);
+ if (kvlist == NULL) {
+ plt_err("Failed to find representor key in devargs list");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ if (rte_kvargs_count(kvlist, "representor") <= 0) {
+ plt_err("Invalid representor key count");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ j = eswitch_dev->nb_esw_da;
+ for (i = 0; i < kvlist->count; i++) {
+ eth_da = &eswitch_dev->esw_da[j].da;
+ memset(eth_da, 0, sizeof(*eth_da));
+ rc = devargs_parse_representor_ports(kvlist->pairs[i].value, eth_da);
+ if (rc) {
+ plt_err("Failed to parse the representor devargs, err %d", rc);
+ goto fail;
+ }
+
+ rc = populate_repr_hw_info(eswitch_dev, eth_da, j);
+ if (rc) {
+ plt_err("Failed to populate representer hw funcs, err %d", rc);
+ goto fail;
+ }
+
+ /* No of representor ports to be created */
+ eswitch_dev->repr_cnt.nb_repr_created += eswitch_dev->esw_da[j].nb_repr_ports;
+ j++;
+ }
+ eswitch_dev->nb_esw_da += kvlist->count;
+
+ return 0;
+fail:
+ return rc;
+}
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index 012d098f80..ea7e363e89 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -29,6 +29,7 @@ sources = files(
'cnxk_ethdev_telemetry.c',
'cnxk_ethdev_sec_telemetry.c',
'cnxk_eswitch.c',
+ 'cnxk_eswitch_devargs.c',
'cnxk_link.c',
'cnxk_lookup.c',
'cnxk_ptp.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 05/24] net/cnxk: probing representor ports
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (3 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 04/24] net/cnxk: eswitch devargs parsing Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 06/24] common/cnxk: common NPC changes for eswitch Harman Kalra
` (18 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra, Anatoly Burakov
Cc: dev, jerinj
Basic skeleton for probing representor devices. If PF device is
passed with "representor" devargs, representor ports gets probed
as a separate ethdev device.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 12 ++
drivers/net/cnxk/cnxk_eswitch.h | 8 +-
drivers/net/cnxk/cnxk_rep.c | 256 ++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep.h | 50 +++++++
drivers/net/cnxk/cnxk_rep_ops.c | 129 ++++++++++++++++
drivers/net/cnxk/meson.build | 2 +
6 files changed, 456 insertions(+), 1 deletion(-)
create mode 100644 drivers/net/cnxk/cnxk_rep.c
create mode 100644 drivers/net/cnxk/cnxk_rep.h
create mode 100644 drivers/net/cnxk/cnxk_rep_ops.c
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index 739a09c034..563b224a6c 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -3,6 +3,7 @@
*/
#include <cnxk_eswitch.h>
+#include <cnxk_rep.h>
#define CNXK_NIX_DEF_SQ_COUNT 512
@@ -42,6 +43,10 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
eswitch_dev = cnxk_eswitch_pmd_priv();
+ /* Remove representor devices associated with PF */
+ if (eswitch_dev->repr_cnt.nb_repr_created)
+ cnxk_rep_dev_remove(eswitch_dev);
+
eswitch_hw_rsrc_cleanup(eswitch_dev);
/* Check if this device is hosting common resource */
nix = roc_idev_npa_nix_get();
@@ -724,6 +729,13 @@ cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pc
eswitch_dev->repr_cnt.max_repr, eswitch_dev->repr_cnt.nb_repr_created,
roc_nix_get_pf_func(&eswitch_dev->nix));
+ /* Probe representor ports */
+ rc = cnxk_rep_dev_probe(pci_dev, eswitch_dev);
+ if (rc) {
+ plt_err("Failed to probe representor ports");
+ goto rsrc_cleanup;
+ }
+
/* Spinlock for synchronization between representors traffic and control
* messages
*/
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index dcb787cf02..4908c3ba95 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -66,6 +66,11 @@ struct cnxk_eswitch_repr_cnt {
uint16_t nb_repr_started;
};
+struct cnxk_eswitch_switch_domain {
+ uint16_t switch_domain_id;
+ uint16_t pf;
+};
+
struct cnxk_rep_info {
struct rte_eth_dev *rep_eth_dev;
};
@@ -121,7 +126,8 @@ struct cnxk_eswitch_dev {
/* Port representor fields */
rte_spinlock_t rep_lock;
- uint16_t switch_domain_id;
+ uint16_t nb_switch_domain;
+ struct cnxk_eswitch_switch_domain sw_dom[RTE_MAX_ETHPORTS];
uint16_t eswitch_vdev;
struct cnxk_rep_info *rep_info;
};
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
new file mode 100644
index 0000000000..295bea3724
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -0,0 +1,256 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+#include <cnxk_rep.h>
+
+#define PF_SHIFT 10
+#define PF_MASK 0x3F
+
+static uint16_t
+get_pf(uint16_t hw_func)
+{
+ return (hw_func >> PF_SHIFT) & PF_MASK;
+}
+
+static uint16_t
+switch_domain_id_allocate(struct cnxk_eswitch_dev *eswitch_dev, uint16_t pf)
+{
+ int i = 0;
+
+ for (i = 0; i < eswitch_dev->nb_switch_domain; i++) {
+ if (eswitch_dev->sw_dom[i].pf == pf)
+ return eswitch_dev->sw_dom[i].switch_domain_id;
+ }
+
+ return RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID;
+}
+
+int
+cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev)
+{
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return 0;
+
+ plt_rep_dbg("Representor port:%d uninit", ethdev->data->port_id);
+ rte_free(ethdev->data->mac_addrs);
+ ethdev->data->mac_addrs = NULL;
+
+ return 0;
+}
+
+int
+cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ int i, rc = 0;
+
+ for (i = 0; i < eswitch_dev->nb_switch_domain; i++) {
+ rc = rte_eth_switch_domain_free(eswitch_dev->sw_dom[i].switch_domain_id);
+ if (rc)
+ plt_err("Failed to alloc switch domain: %d", rc);
+ }
+
+ return rc;
+}
+
+static int
+cnxk_rep_parent_setup(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ uint16_t pf, prev_pf = 0, switch_domain_id;
+ int rc, i, j = 0;
+
+ if (eswitch_dev->rep_info)
+ return 0;
+
+ eswitch_dev->rep_info =
+ plt_zmalloc(sizeof(eswitch_dev->rep_info[0]) * eswitch_dev->repr_cnt.max_repr, 0);
+ if (!eswitch_dev->rep_info) {
+ plt_err("Failed to alloc memory for rep info");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ /* Allocate switch domain for all PFs (VFs will be under same domain as PF) */
+ for (i = 0; i < eswitch_dev->repr_cnt.max_repr; i++) {
+ pf = get_pf(eswitch_dev->nix.rep_pfvf_map[i]);
+ if (pf == prev_pf)
+ continue;
+
+ rc = rte_eth_switch_domain_alloc(&switch_domain_id);
+ if (rc) {
+ plt_err("Failed to alloc switch domain: %d", rc);
+ goto fail;
+ }
+ plt_rep_dbg("Allocated switch domain id %d for pf %d\n", switch_domain_id, pf);
+ eswitch_dev->sw_dom[j].switch_domain_id = switch_domain_id;
+ eswitch_dev->sw_dom[j].pf = pf;
+ prev_pf = pf;
+ j++;
+ }
+ eswitch_dev->nb_switch_domain = j;
+
+ return 0;
+fail:
+ return rc;
+}
+
+static uint16_t
+cnxk_rep_tx_burst(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(tx_queue);
+ PLT_SET_USED(tx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
+
+static uint16_t
+cnxk_rep_rx_burst(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(rx_queue);
+ PLT_SET_USED(rx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
+
+static int
+cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
+{
+ struct cnxk_rep_dev *rep_params = (struct cnxk_rep_dev *)params;
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+
+ rep_dev->port_id = rep_params->port_id;
+ rep_dev->switch_domain_id = rep_params->switch_domain_id;
+ rep_dev->parent_dev = rep_params->parent_dev;
+ rep_dev->hw_func = rep_params->hw_func;
+ rep_dev->rep_id = rep_params->rep_id;
+
+ eth_dev->data->dev_flags |= RTE_ETH_DEV_REPRESENTOR;
+ eth_dev->data->representor_id = rep_params->port_id;
+ eth_dev->data->backer_port_id = eth_dev->data->port_id;
+
+ eth_dev->data->mac_addrs = plt_zmalloc(RTE_ETHER_ADDR_LEN, 0);
+ if (!eth_dev->data->mac_addrs) {
+ plt_err("Failed to allocate memory for mac addr");
+ return -ENOMEM;
+ }
+
+ rte_eth_random_addr(rep_dev->mac_addr);
+ memcpy(eth_dev->data->mac_addrs, rep_dev->mac_addr, RTE_ETHER_ADDR_LEN);
+
+ /* Set the device operations */
+ eth_dev->dev_ops = &cnxk_rep_dev_ops;
+
+ /* Rx/Tx functions stubs to avoid crashing */
+ eth_dev->rx_pkt_burst = cnxk_rep_rx_burst;
+ eth_dev->tx_pkt_burst = cnxk_rep_tx_burst;
+
+ /* Only single queues for representor devices */
+ eth_dev->data->nb_rx_queues = 1;
+ eth_dev->data->nb_tx_queues = 1;
+
+ eth_dev->data->dev_link.link_speed = RTE_ETH_SPEED_NUM_NONE;
+ eth_dev->data->dev_link.link_duplex = RTE_ETH_LINK_FULL_DUPLEX;
+ eth_dev->data->dev_link.link_status = RTE_ETH_LINK_UP;
+ eth_dev->data->dev_link.link_autoneg = RTE_ETH_LINK_FIXED;
+
+ return 0;
+}
+
+static int
+create_representor_ethdev(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev,
+ struct cnxk_eswitch_devargs *esw_da, int idx)
+{
+ char name[RTE_ETH_NAME_MAX_LEN];
+ struct rte_eth_dev *rep_eth_dev;
+ uint16_t hw_func;
+ int rc = 0;
+
+ struct cnxk_rep_dev rep = {.port_id = eswitch_dev->repr_cnt.nb_repr_probed,
+ .parent_dev = eswitch_dev};
+
+ if (esw_da->type == CNXK_ESW_DA_TYPE_PFVF) {
+ hw_func = esw_da->repr_hw_info[idx].hw_func;
+ rep.switch_domain_id = switch_domain_id_allocate(eswitch_dev, get_pf(hw_func));
+ if (rep.switch_domain_id == RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID) {
+ plt_err("Failed to get a valid switch domain id");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ esw_da->repr_hw_info[idx].port_id = rep.port_id;
+ /* Representor port net_bdf_port */
+ snprintf(name, sizeof(name), "net_%s_hw_%x_representor_%d", pci_dev->device.name,
+ hw_func, rep.port_id);
+
+ rep.hw_func = hw_func;
+ rep.rep_id = esw_da->repr_hw_info[idx].rep_id;
+
+ } else {
+ snprintf(name, sizeof(name), "net_%s_representor_%d", pci_dev->device.name,
+ rep.port_id);
+ rep.switch_domain_id = RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID;
+ }
+
+ rc = rte_eth_dev_create(&pci_dev->device, name, sizeof(struct cnxk_rep_dev), NULL, NULL,
+ cnxk_rep_dev_init, &rep);
+ if (rc) {
+ plt_err("Failed to create cnxk vf representor %s", name);
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ rep_eth_dev = rte_eth_dev_allocated(name);
+ if (!rep_eth_dev) {
+ plt_err("Failed to find the eth_dev for VF-Rep: %s.", name);
+ rc = -ENODEV;
+ goto fail;
+ }
+
+ plt_rep_dbg("Representor portid %d (%s) type %d probe done", rep_eth_dev->data->port_id,
+ name, esw_da->da.type);
+ eswitch_dev->rep_info[rep.port_id].rep_eth_dev = rep_eth_dev;
+ eswitch_dev->repr_cnt.nb_repr_probed++;
+
+ return 0;
+fail:
+ return rc;
+}
+
+int
+cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev)
+{
+ struct cnxk_eswitch_devargs *esw_da;
+ uint16_t num_rep;
+ int i, j, rc;
+
+ if (eswitch_dev->repr_cnt.nb_repr_created > RTE_MAX_ETHPORTS) {
+ plt_err("nb_representor_ports %d > %d MAX ETHPORTS\n",
+ eswitch_dev->repr_cnt.nb_repr_created, RTE_MAX_ETHPORTS);
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Initialize the internals of representor ports */
+ rc = cnxk_rep_parent_setup(eswitch_dev);
+ if (rc) {
+ plt_err("Failed to setup the parent device, err %d", rc);
+ goto fail;
+ }
+
+ for (i = eswitch_dev->last_probed; i < eswitch_dev->nb_esw_da; i++) {
+ esw_da = &eswitch_dev->esw_da[i];
+ /* Check the representor devargs */
+ num_rep = esw_da->nb_repr_ports;
+ for (j = 0; j < num_rep; j++) {
+ rc = create_representor_ethdev(pci_dev, eswitch_dev, esw_da, j);
+ if (rc)
+ goto fail;
+ }
+ }
+ eswitch_dev->last_probed = i;
+
+ return 0;
+fail:
+ return rc;
+}
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
new file mode 100644
index 0000000000..2cb3ae8ac5
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+#include <cnxk_eswitch.h>
+#include <cnxk_ethdev.h>
+
+#ifndef __CNXK_REP_H__
+#define __CNXK_REP_H__
+
+/* Common ethdev ops */
+extern struct eth_dev_ops cnxk_rep_dev_ops;
+
+struct cnxk_rep_dev {
+ uint16_t port_id;
+ uint16_t rep_id;
+ uint16_t switch_domain_id;
+ struct cnxk_eswitch_dev *parent_dev;
+ uint16_t hw_func;
+ uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
+};
+
+static inline struct cnxk_rep_dev *
+cnxk_rep_pmd_priv(const struct rte_eth_dev *eth_dev)
+{
+ return eth_dev->data->dev_private;
+}
+
+int cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev);
+int cnxk_rep_dev_info_get(struct rte_eth_dev *eth_dev, struct rte_eth_dev_info *dev_info);
+int cnxk_rep_representor_info_get(struct rte_eth_dev *dev, struct rte_eth_representor_info *info);
+int cnxk_rep_dev_configure(struct rte_eth_dev *eth_dev);
+
+int cnxk_rep_link_update(struct rte_eth_dev *eth_dev, int wait_to_compl);
+int cnxk_rep_dev_start(struct rte_eth_dev *eth_dev);
+int cnxk_rep_rx_queue_setup(struct rte_eth_dev *eth_dev, uint16_t queue_idx, uint16_t nb_desc,
+ unsigned int socket_id, const struct rte_eth_rxconf *rx_conf,
+ struct rte_mempool *mp);
+int cnxk_rep_tx_queue_setup(struct rte_eth_dev *eth_dev, uint16_t queue_idx, uint16_t nb_desc,
+ unsigned int socket_id, const struct rte_eth_txconf *tx_conf);
+void cnxk_rep_rx_queue_release(struct rte_eth_dev *dev, uint16_t queue_idx);
+void cnxk_rep_tx_queue_release(struct rte_eth_dev *dev, uint16_t queue_idx);
+int cnxk_rep_dev_stop(struct rte_eth_dev *eth_dev);
+int cnxk_rep_dev_close(struct rte_eth_dev *eth_dev);
+int cnxk_rep_stats_get(struct rte_eth_dev *eth_dev, struct rte_eth_stats *stats);
+int cnxk_rep_stats_reset(struct rte_eth_dev *eth_dev);
+int cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops);
+
+#endif /* __CNXK_REP_H__ */
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
new file mode 100644
index 0000000000..67dcc422e3
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -0,0 +1,129 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <cnxk_rep.h>
+
+int
+cnxk_rep_link_update(struct rte_eth_dev *ethdev, int wait_to_complete)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(wait_to_complete);
+ return 0;
+}
+
+int
+cnxk_rep_dev_info_get(struct rte_eth_dev *ethdev, struct rte_eth_dev_info *devinfo)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(devinfo);
+ return 0;
+}
+
+int
+cnxk_rep_dev_configure(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_dev_start(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_dev_close(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_dev_stop(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_rx_queue_setup(struct rte_eth_dev *ethdev, uint16_t rx_queue_id, uint16_t nb_rx_desc,
+ unsigned int socket_id, const struct rte_eth_rxconf *rx_conf,
+ struct rte_mempool *mb_pool)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(rx_queue_id);
+ PLT_SET_USED(nb_rx_desc);
+ PLT_SET_USED(socket_id);
+ PLT_SET_USED(rx_conf);
+ PLT_SET_USED(mb_pool);
+ return 0;
+}
+
+void
+cnxk_rep_rx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(queue_id);
+}
+
+int
+cnxk_rep_tx_queue_setup(struct rte_eth_dev *ethdev, uint16_t tx_queue_id, uint16_t nb_tx_desc,
+ unsigned int socket_id, const struct rte_eth_txconf *tx_conf)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(tx_queue_id);
+ PLT_SET_USED(nb_tx_desc);
+ PLT_SET_USED(socket_id);
+ PLT_SET_USED(tx_conf);
+ return 0;
+}
+
+void
+cnxk_rep_tx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(queue_id);
+}
+
+int
+cnxk_rep_stats_get(struct rte_eth_dev *ethdev, struct rte_eth_stats *stats)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(stats);
+ return 0;
+}
+
+int
+cnxk_rep_stats_reset(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(ops);
+ return 0;
+}
+
+/* CNXK platform representor dev ops */
+struct eth_dev_ops cnxk_rep_dev_ops = {
+ .dev_infos_get = cnxk_rep_dev_info_get,
+ .dev_configure = cnxk_rep_dev_configure,
+ .dev_start = cnxk_rep_dev_start,
+ .rx_queue_setup = cnxk_rep_rx_queue_setup,
+ .rx_queue_release = cnxk_rep_rx_queue_release,
+ .tx_queue_setup = cnxk_rep_tx_queue_setup,
+ .tx_queue_release = cnxk_rep_tx_queue_release,
+ .link_update = cnxk_rep_link_update,
+ .dev_close = cnxk_rep_dev_close,
+ .dev_stop = cnxk_rep_dev_stop,
+ .stats_get = cnxk_rep_stats_get,
+ .stats_reset = cnxk_rep_stats_reset,
+ .flow_ops_get = cnxk_rep_flow_ops_get
+};
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index ea7e363e89..fcd5d3d569 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -34,6 +34,8 @@ sources = files(
'cnxk_lookup.c',
'cnxk_ptp.c',
'cnxk_flow.c',
+ 'cnxk_rep.c',
+ 'cnxk_rep_ops.c',
'cnxk_stats.c',
'cnxk_tm.c',
)
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 06/24] common/cnxk: common NPC changes for eswitch
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (4 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 05/24] net/cnxk: probing representor ports Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 07/24] common/cnxk: interface to update VLAN TPID Harman Kalra
` (17 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
- adding support for installing flow using npc_install_flow mbox
- rss action configuration for eswitch
- new mcam helper apis
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/meson.build | 1 +
drivers/common/cnxk/roc_api.h | 3 +
drivers/common/cnxk/roc_eswitch.c | 285 +++++++++++++++++++++++++++++
drivers/common/cnxk/roc_eswitch.h | 21 +++
drivers/common/cnxk/roc_mbox.h | 25 +++
drivers/common/cnxk/roc_npc.c | 26 ++-
drivers/common/cnxk/roc_npc.h | 5 +-
drivers/common/cnxk/roc_npc_mcam.c | 2 +-
drivers/common/cnxk/roc_npc_priv.h | 3 +-
drivers/common/cnxk/version.map | 6 +
10 files changed, 368 insertions(+), 9 deletions(-)
create mode 100644 drivers/common/cnxk/roc_eswitch.c
create mode 100644 drivers/common/cnxk/roc_eswitch.h
diff --git a/drivers/common/cnxk/meson.build b/drivers/common/cnxk/meson.build
index 56eea52909..e0e4600989 100644
--- a/drivers/common/cnxk/meson.build
+++ b/drivers/common/cnxk/meson.build
@@ -20,6 +20,7 @@ sources = files(
'roc_cpt_debug.c',
'roc_dev.c',
'roc_dpi.c',
+ 'roc_eswitch.c',
'roc_hash.c',
'roc_idev.c',
'roc_irq.c',
diff --git a/drivers/common/cnxk/roc_api.h b/drivers/common/cnxk/roc_api.h
index f630853088..6a86863c57 100644
--- a/drivers/common/cnxk/roc_api.h
+++ b/drivers/common/cnxk/roc_api.h
@@ -117,4 +117,7 @@
/* MACsec */
#include "roc_mcs.h"
+/* Eswitch */
+#include "roc_eswitch.h"
+
#endif /* _ROC_API_H_ */
diff --git a/drivers/common/cnxk/roc_eswitch.c b/drivers/common/cnxk/roc_eswitch.c
new file mode 100644
index 0000000000..42a27e7442
--- /dev/null
+++ b/drivers/common/cnxk/roc_eswitch.c
@@ -0,0 +1,285 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <arpa/inet.h>
+
+#include "roc_api.h"
+#include "roc_priv.h"
+
+static int
+eswitch_vlan_rx_cfg(uint16_t pcifunc, struct mbox *mbox)
+{
+ struct nix_vtag_config *vtag_cfg;
+ int rc;
+
+ vtag_cfg = mbox_alloc_msg_nix_vtag_cfg(mbox_get(mbox));
+
+ /* config strip, capture and size */
+ vtag_cfg->hdr.pcifunc = pcifunc;
+ vtag_cfg->vtag_size = NIX_VTAGSIZE_T4;
+ vtag_cfg->cfg_type = VTAG_RX; /* rx vlan cfg */
+ vtag_cfg->rx.vtag_type = NIX_RX_VTAG_TYPE0;
+ vtag_cfg->rx.strip_vtag = true;
+ vtag_cfg->rx.capture_vtag = true;
+
+ rc = mbox_process(mbox);
+ if (rc)
+ goto exit;
+
+ rc = 0;
+exit:
+ mbox_put(mbox);
+ return rc;
+}
+
+static int
+eswitch_vlan_tx_cfg(struct roc_npc_flow *flow, uint16_t pcifunc, struct mbox *mbox,
+ uint16_t vlan_tci, uint16_t *vidx)
+{
+ struct nix_vtag_config *vtag_cfg;
+ struct nix_vtag_config_rsp *rsp;
+ int rc;
+
+ union {
+ uint64_t reg;
+ struct nix_tx_vtag_action_s act;
+ } tx_vtag_action;
+
+ vtag_cfg = mbox_alloc_msg_nix_vtag_cfg(mbox_get(mbox));
+
+ /* Insert vlan tag */
+ vtag_cfg->hdr.pcifunc = pcifunc;
+ vtag_cfg->vtag_size = NIX_VTAGSIZE_T4;
+ vtag_cfg->cfg_type = VTAG_TX; /* tx vlan cfg */
+ vtag_cfg->tx.cfg_vtag0 = true;
+ vtag_cfg->tx.vtag0 = (((uint32_t)ROC_ESWITCH_VLAN_TPID << 16) | vlan_tci);
+
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ if (rsp->vtag0_idx < 0) {
+ plt_err("Failed to config TX VTAG action");
+ rc = -EINVAL;
+ goto exit;
+ }
+
+ *vidx = rsp->vtag0_idx;
+ tx_vtag_action.reg = 0;
+ tx_vtag_action.act.vtag0_def = rsp->vtag0_idx;
+ tx_vtag_action.act.vtag0_lid = NPC_LID_LA;
+ tx_vtag_action.act.vtag0_op = NIX_TX_VTAGOP_INSERT;
+ tx_vtag_action.act.vtag0_relptr = NIX_TX_VTAGACTION_VTAG0_RELPTR;
+
+ flow->vtag_action = tx_vtag_action.reg;
+
+ rc = 0;
+exit:
+ mbox_put(mbox);
+ return rc;
+
+ return 0;
+}
+
+int
+roc_eswitch_npc_mcam_tx_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow, uint16_t pcifunc,
+ uint32_t vlan_tci)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+ struct npc_install_flow_req *req;
+ struct npc_install_flow_rsp *rsp;
+ struct mbox *mbox = npc->mbox;
+ uint16_t vidx = 0, lbkid;
+ int rc;
+
+ rc = eswitch_vlan_tx_cfg(flow, roc_npc->pf_func, mbox, vlan_tci, &vidx);
+ if (rc) {
+ plt_err("Failed to configure VLAN TX, err %d", rc);
+ goto fail;
+ }
+
+ req = mbox_alloc_msg_npc_install_flow(mbox_get(mbox));
+
+ lbkid = 0;
+ req->hdr.pcifunc = roc_npc->pf_func; /* Eswitch PF is requester */
+ req->vf = pcifunc;
+ req->entry = flow->mcam_id;
+ req->intf = NPC_MCAM_TX;
+ req->op = NIX_TX_ACTIONOP_UCAST_CHAN;
+ req->index = (lbkid << 8) | ROC_ESWITCH_LBK_CHAN;
+ req->set_cntr = 1;
+ req->vtag0_def = vidx;
+ req->vtag0_op = 1;
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ flow->nix_intf = NIX_INTF_TX;
+exit:
+ mbox_put(mbox);
+fail:
+ return rc;
+}
+
+static int
+eswitch_vtag_cfg_delete(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+ struct nix_vtag_config *vtag_cfg;
+ struct nix_vtag_config_rsp *rsp;
+ struct mbox *mbox = npc->mbox;
+ int rc = 0;
+
+ union {
+ uint64_t reg;
+ struct nix_tx_vtag_action_s act;
+ } tx_vtag_action;
+
+ tx_vtag_action.reg = flow->vtag_action;
+ vtag_cfg = mbox_alloc_msg_nix_vtag_cfg(mbox_get(mbox));
+
+ if (vtag_cfg == NULL) {
+ rc = -ENOSPC;
+ goto exit;
+ }
+
+ vtag_cfg->cfg_type = VTAG_TX;
+ vtag_cfg->vtag_size = NIX_VTAGSIZE_T4;
+ vtag_cfg->tx.vtag0_idx = tx_vtag_action.act.vtag0_def;
+ vtag_cfg->tx.free_vtag0 = true;
+
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ rc = rsp->hdr.rc;
+exit:
+ mbox_put(mbox);
+ return rc;
+}
+
+int
+roc_eswitch_npc_mcam_delete_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+ struct npc_delete_flow_req *req;
+ struct msg_rsp *rsp;
+ struct mbox *mbox = npc->mbox;
+ int rc = 0;
+
+ /* Removing the VLAN TX config */
+ if (flow->nix_intf == NIX_INTF_TX) {
+ rc = eswitch_vtag_cfg_delete(roc_npc, flow);
+ if (rc)
+ plt_err("Failed to delete TX vtag config");
+ }
+
+ req = mbox_alloc_msg_npc_delete_flow(mbox_get(mbox));
+
+ req->entry = flow->mcam_id;
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ rc = rsp->hdr.rc;
+exit:
+ mbox_put(mbox);
+ return rc;
+}
+
+int
+roc_eswitch_npc_mcam_rx_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow, uint16_t pcifunc,
+ uint16_t vlan_tci, uint16_t vlan_tci_mask)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+ struct npc_install_flow_req *req;
+ struct npc_install_flow_rsp *rsp;
+ struct mbox *mbox = npc->mbox;
+ bool is_esw_dev;
+ int rc;
+
+ /* For ESW PF/VF */
+ is_esw_dev = (dev_get_pf(roc_npc->pf_func) == dev_get_pf(pcifunc));
+ /* VLAN Rx config */
+ if (is_esw_dev) {
+ rc = eswitch_vlan_rx_cfg(roc_npc->pf_func, mbox);
+ if (rc) {
+ plt_err("Failed to configure VLAN RX rule, err %d", rc);
+ goto fail;
+ }
+ }
+
+ req = mbox_alloc_msg_npc_install_flow(mbox_get(mbox));
+ req->vf = pcifunc;
+ /* Action */
+ req->op = NIX_RX_ACTIONOP_DEFAULT;
+ req->index = 0;
+ req->entry = flow->mcam_id;
+ req->hdr.pcifunc = roc_npc->pf_func; /* Eswitch PF is requester */
+ req->features = BIT_ULL(NPC_OUTER_VID) | BIT_ULL(NPC_VLAN_ETYPE_CTAG);
+ req->vtag0_valid = true;
+ /* For ESW PF/VF using configured vlan rx cfg while for other
+ * representees using standard vlan_type = 7 which is strip.
+ */
+ req->vtag0_type = is_esw_dev ? NIX_RX_VTAG_TYPE0 : NIX_RX_VTAG_TYPE7;
+ req->packet.vlan_etype = ROC_ESWITCH_VLAN_TPID;
+ req->mask.vlan_etype = 0xFFFF;
+ req->packet.vlan_tci = ntohs(vlan_tci & 0xFFFF);
+ req->mask.vlan_tci = ntohs(vlan_tci_mask);
+
+ req->channel = ROC_ESWITCH_LBK_CHAN;
+ req->chan_mask = 0xffff;
+ req->intf = NPC_MCAM_RX;
+ req->set_cntr = 1;
+ req->cntr_val = flow->ctr_id;
+
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ flow->nix_intf = NIX_INTF_RX;
+exit:
+ mbox_put(mbox);
+fail:
+ return rc;
+}
+
+int
+roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc, struct roc_npc_flow *flow,
+ uint32_t flowkey_cfg, uint16_t *reta_tbl)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+ struct roc_nix *roc_nix = roc_npc->roc_nix;
+ uint32_t rss_grp_idx;
+ uint8_t flowkey_algx;
+ int rc;
+
+ rc = npc_rss_free_grp_get(npc, &rss_grp_idx);
+ /* RSS group :0 is not usable for flow rss action */
+ if (rc < 0 || rss_grp_idx == 0)
+ return -ENOSPC;
+
+ /* Populating reta table for the specific RSS group */
+ rc = roc_nix_rss_reta_set(roc_nix, rss_grp_idx, reta_tbl);
+ if (rc) {
+ plt_err("Failed to init rss table rc = %d", rc);
+ return rc;
+ }
+
+ rc = roc_nix_rss_flowkey_set(roc_nix, &flowkey_algx, flowkey_cfg, rss_grp_idx,
+ flow->mcam_id);
+ if (rc) {
+ plt_err("Failed to set rss hash function rc = %d", rc);
+ return rc;
+ }
+
+ plt_bitmap_set(npc->rss_grp_entries, rss_grp_idx);
+
+ flow->npc_action &= (~(0xfULL));
+ flow->npc_action |= NIX_RX_ACTIONOP_RSS;
+ flow->npc_action |=
+ ((uint64_t)(flowkey_algx & NPC_RSS_ACT_ALG_MASK) << NPC_RSS_ACT_ALG_OFFSET) |
+ ((uint64_t)(rss_grp_idx & NPC_RSS_ACT_GRP_MASK) << NPC_RSS_ACT_GRP_OFFSET);
+ return 0;
+}
diff --git a/drivers/common/cnxk/roc_eswitch.h b/drivers/common/cnxk/roc_eswitch.h
new file mode 100644
index 0000000000..35976b7ff6
--- /dev/null
+++ b/drivers/common/cnxk/roc_eswitch.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#ifndef __ROC_ESWITCH_H__
+#define __ROC_ESWITCH_H__
+
+#define ROC_ESWITCH_VLAN_TPID 0x8100
+#define ROC_ESWITCH_LBK_CHAN 63
+
+/* NPC */
+int __roc_api roc_eswitch_npc_mcam_rx_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow,
+ uint16_t pcifunc, uint16_t vlan_tci,
+ uint16_t vlan_tci_mask);
+int __roc_api roc_eswitch_npc_mcam_tx_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow,
+ uint16_t pcifunc, uint32_t vlan_tci);
+int __roc_api roc_eswitch_npc_mcam_delete_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow);
+int __roc_api roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc,
+ struct roc_npc_flow *flow, uint32_t flowkey_cfg,
+ uint16_t *reta_tbl);
+#endif /* __ROC_ESWITCH_H__ */
diff --git a/drivers/common/cnxk/roc_mbox.h b/drivers/common/cnxk/roc_mbox.h
index b7e2f43d45..4c846f0757 100644
--- a/drivers/common/cnxk/roc_mbox.h
+++ b/drivers/common/cnxk/roc_mbox.h
@@ -386,6 +386,18 @@ enum rvu_af_status {
RVU_INVALID_VF_ID = -256,
};
+/* For NIX RX vtag action */
+enum nix_rx_vtag0_type {
+ NIX_RX_VTAG_TYPE0,
+ NIX_RX_VTAG_TYPE1,
+ NIX_RX_VTAG_TYPE2,
+ NIX_RX_VTAG_TYPE3,
+ NIX_RX_VTAG_TYPE4,
+ NIX_RX_VTAG_TYPE5,
+ NIX_RX_VTAG_TYPE6,
+ NIX_RX_VTAG_TYPE7,
+};
+
struct ready_msg_rsp {
struct mbox_msghdr hdr;
uint16_t __io sclk_freq; /* SCLK frequency */
@@ -2442,6 +2454,8 @@ enum header_fields {
NPC_DMAC,
NPC_SMAC,
NPC_ETYPE,
+ NPC_VLAN_ETYPE_CTAG, /* 0x8100 */
+ NPC_VLAN_ETYPE_STAG, /* 0x88A8 */
NPC_OUTER_VID,
NPC_TOS,
NPC_SIP_IPV4,
@@ -2476,6 +2490,14 @@ struct flow_msg {
uint8_t __io tc;
uint16_t __io sport;
uint16_t __io dport;
+ union {
+ uint8_t __io ip_flag;
+ uint8_t __io next_header;
+ };
+ uint16_t __io vlan_itci;
+ uint32_t __io gtpu_teid;
+ uint32_t __io gtpc_teid;
+ uint16_t __io sq_id;
};
struct npc_install_flow_req {
@@ -2485,6 +2507,7 @@ struct npc_install_flow_req {
uint64_t __io features;
uint16_t __io entry;
uint16_t __io channel;
+ uint16_t __io chan_mask;
uint8_t __io intf;
uint8_t __io set_cntr;
uint8_t __io default_rule;
@@ -2507,6 +2530,8 @@ struct npc_install_flow_req {
uint8_t __io vtag0_op;
uint16_t __io vtag1_def;
uint8_t __io vtag1_op;
+ /* old counter value */
+ uint16_t __io cntr_val;
};
struct npc_install_flow_rsp {
diff --git a/drivers/common/cnxk/roc_npc.c b/drivers/common/cnxk/roc_npc.c
index 9a0fe5f4e2..67a660a2bc 100644
--- a/drivers/common/cnxk/roc_npc.c
+++ b/drivers/common/cnxk/roc_npc.c
@@ -77,8 +77,23 @@ roc_npc_inl_mcam_clear_counter(uint32_t ctr_id)
}
int
-roc_npc_mcam_read_counter(struct roc_npc *roc_npc, uint32_t ctr_id,
- uint64_t *count)
+roc_npc_mcam_alloc_counter(struct roc_npc *roc_npc, uint16_t *ctr_id)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+
+ return npc_mcam_alloc_counter(npc->mbox, ctr_id);
+}
+
+int
+roc_npc_get_free_mcam_entry(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+
+ return npc_get_free_mcam_entry(npc->mbox, flow, npc);
+}
+
+int
+roc_npc_mcam_read_counter(struct roc_npc *roc_npc, uint32_t ctr_id, uint64_t *count)
{
struct npc *npc = roc_npc_to_npc_priv(roc_npc);
@@ -157,14 +172,13 @@ roc_npc_mcam_free_all_resources(struct roc_npc *roc_npc)
}
int
-roc_npc_mcam_alloc_entries(struct roc_npc *roc_npc, int ref_entry,
- int *alloc_entry, int req_count, int priority,
- int *resp_count)
+roc_npc_mcam_alloc_entries(struct roc_npc *roc_npc, int ref_entry, int *alloc_entry, int req_count,
+ int priority, int *resp_count, bool is_conti)
{
struct npc *npc = roc_npc_to_npc_priv(roc_npc);
return npc_mcam_alloc_entries(npc->mbox, ref_entry, alloc_entry, req_count, priority,
- resp_count, 0);
+ resp_count, is_conti);
}
int
diff --git a/drivers/common/cnxk/roc_npc.h b/drivers/common/cnxk/roc_npc.h
index e880a7fa67..349c7f9d22 100644
--- a/drivers/common/cnxk/roc_npc.h
+++ b/drivers/common/cnxk/roc_npc.h
@@ -431,7 +431,8 @@ int __roc_api roc_npc_mcam_enable_all_entries(struct roc_npc *roc_npc, bool enab
int __roc_api roc_npc_mcam_alloc_entry(struct roc_npc *roc_npc, struct roc_npc_flow *mcam,
struct roc_npc_flow *ref_mcam, int prio, int *resp_count);
int __roc_api roc_npc_mcam_alloc_entries(struct roc_npc *roc_npc, int ref_entry, int *alloc_entry,
- int req_count, int priority, int *resp_count);
+ int req_count, int priority, int *resp_count,
+ bool is_conti);
int __roc_api roc_npc_mcam_ena_dis_entry(struct roc_npc *roc_npc, struct roc_npc_flow *mcam,
bool enable);
int __roc_api roc_npc_mcam_write_entry(struct roc_npc *roc_npc, struct roc_npc_flow *mcam);
@@ -442,6 +443,8 @@ int __roc_api roc_npc_get_low_priority_mcam(struct roc_npc *roc_npc);
int __roc_api roc_npc_mcam_free_counter(struct roc_npc *roc_npc, uint16_t ctr_id);
int __roc_api roc_npc_mcam_read_counter(struct roc_npc *roc_npc, uint32_t ctr_id, uint64_t *count);
int __roc_api roc_npc_mcam_clear_counter(struct roc_npc *roc_npc, uint32_t ctr_id);
+int __roc_api roc_npc_mcam_alloc_counter(struct roc_npc *roc_npc, uint16_t *ctr_id);
+int __roc_api roc_npc_get_free_mcam_entry(struct roc_npc *roc_npc, struct roc_npc_flow *flow);
int __roc_api roc_npc_inl_mcam_read_counter(uint32_t ctr_id, uint64_t *count);
int __roc_api roc_npc_inl_mcam_clear_counter(uint32_t ctr_id);
int __roc_api roc_npc_mcam_free_all_resources(struct roc_npc *roc_npc);
diff --git a/drivers/common/cnxk/roc_npc_mcam.c b/drivers/common/cnxk/roc_npc_mcam.c
index 3ef189e184..2de988a44b 100644
--- a/drivers/common/cnxk/roc_npc_mcam.c
+++ b/drivers/common/cnxk/roc_npc_mcam.c
@@ -4,7 +4,7 @@
#include "roc_api.h"
#include "roc_priv.h"
-static int
+int
npc_mcam_alloc_counter(struct mbox *mbox, uint16_t *ctr)
{
struct npc_mcam_alloc_counter_req *req;
diff --git a/drivers/common/cnxk/roc_npc_priv.h b/drivers/common/cnxk/roc_npc_priv.h
index c0809407a6..50b62b1244 100644
--- a/drivers/common/cnxk/roc_npc_priv.h
+++ b/drivers/common/cnxk/roc_npc_priv.h
@@ -432,6 +432,7 @@ roc_npc_to_npc_priv(struct roc_npc *npc)
return (struct npc *)npc->reserved;
}
+int npc_mcam_alloc_counter(struct mbox *mbox, uint16_t *ctr);
int npc_mcam_free_counter(struct mbox *mbox, uint16_t ctr_id);
int npc_mcam_read_counter(struct mbox *mbox, uint32_t ctr_id, uint64_t *count);
int npc_mcam_clear_counter(struct mbox *mbox, uint32_t ctr_id);
@@ -480,7 +481,6 @@ uint64_t npc_get_kex_capability(struct npc *npc);
int npc_process_ipv6_field_hash(const struct roc_npc_flow_item_ipv6 *ipv6_spec,
const struct roc_npc_flow_item_ipv6 *ipv6_mask,
struct npc_parse_state *pst, uint8_t type);
-int npc_rss_free_grp_get(struct npc *npc, uint32_t *grp);
int npc_rss_action_configure(struct roc_npc *roc_npc, const struct roc_npc_action_rss *rss,
uint8_t *alg_idx, uint32_t *rss_grp, uint32_t mcam_id);
int npc_rss_action_program(struct roc_npc *roc_npc, const struct roc_npc_action actions[],
@@ -496,4 +496,5 @@ void npc_aged_flows_bitmap_free(struct roc_npc *roc_npc);
int npc_aging_ctrl_thread_create(struct roc_npc *roc_npc, const struct roc_npc_action_age *age,
struct roc_npc_flow *flow);
void npc_aging_ctrl_thread_destroy(struct roc_npc *roc_npc);
+int npc_rss_free_grp_get(struct npc *npc, uint32_t *pos);
#endif /* _ROC_NPC_PRIV_H_ */
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index bd28803013..feda34b852 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -91,6 +91,10 @@ INTERNAL {
roc_dpi_disable;
roc_dpi_enable;
roc_error_msg_get;
+ roc_eswitch_npc_mcam_delete_rule;
+ roc_eswitch_npc_mcam_rx_rule;
+ roc_eswitch_npc_mcam_tx_rule;
+ roc_eswitch_npc_rss_action_configure;
roc_hash_md5_gen;
roc_hash_sha1_gen;
roc_hash_sha256_gen;
@@ -443,6 +447,7 @@ INTERNAL {
roc_npc_flow_dump;
roc_npc_flow_mcam_dump;
roc_npc_flow_parse;
+ roc_npc_get_free_mcam_entry;
roc_npc_get_low_priority_mcam;
roc_npc_init;
roc_npc_kex_capa_get;
@@ -450,6 +455,7 @@ INTERNAL {
roc_npc_mark_actions_sub_return;
roc_npc_vtag_actions_get;
roc_npc_vtag_actions_sub_return;
+ roc_npc_mcam_alloc_counter;
roc_npc_mcam_alloc_entries;
roc_npc_mcam_alloc_entry;
roc_npc_mcam_clear_counter;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 07/24] common/cnxk: interface to update VLAN TPID
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (5 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 06/24] common/cnxk: common NPC changes for eswitch Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2024-01-04 12:47 ` Jerin Jacob
2023-12-19 17:39 ` [PATCH v2 08/24] net/cnxk: eswitch flow configurations Harman Kalra
` (16 subsequent siblings)
23 siblings, 1 reply; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Introducing eswitch variant of set vlan tpid api which can be
using for PF and VF
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_eswitch.c | 15 +++++++++++++++
drivers/common/cnxk/roc_eswitch.h | 4 ++++
drivers/common/cnxk/roc_nix_priv.h | 4 ++--
drivers/common/cnxk/roc_nix_vlan.c | 23 ++++++++++++++++++-----
drivers/common/cnxk/version.map | 1 +
5 files changed, 40 insertions(+), 7 deletions(-)
diff --git a/drivers/common/cnxk/roc_eswitch.c b/drivers/common/cnxk/roc_eswitch.c
index 42a27e7442..7f2a8e6c06 100644
--- a/drivers/common/cnxk/roc_eswitch.c
+++ b/drivers/common/cnxk/roc_eswitch.c
@@ -283,3 +283,18 @@ roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc, struct roc_npc_flo
((uint64_t)(rss_grp_idx & NPC_RSS_ACT_GRP_MASK) << NPC_RSS_ACT_GRP_OFFSET);
return 0;
}
+
+int
+roc_eswitch_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t tpid, bool is_vf)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+ int rc = 0;
+
+ /* Configuring for PF/VF */
+ rc = nix_vlan_tpid_set(dev->mbox, dev->pf_func | is_vf, type, tpid);
+ if (rc)
+ plt_err("Failed to set tpid for PF, rc %d", rc);
+
+ return rc;
+}
diff --git a/drivers/common/cnxk/roc_eswitch.h b/drivers/common/cnxk/roc_eswitch.h
index 35976b7ff6..0dd23ff76a 100644
--- a/drivers/common/cnxk/roc_eswitch.h
+++ b/drivers/common/cnxk/roc_eswitch.h
@@ -18,4 +18,8 @@ int __roc_api roc_eswitch_npc_mcam_delete_rule(struct roc_npc *roc_npc, struct r
int __roc_api roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc,
struct roc_npc_flow *flow, uint32_t flowkey_cfg,
uint16_t *reta_tbl);
+
+/* NIX */
+int __roc_api roc_eswitch_nix_vlan_tpid_set(struct roc_nix *nix, uint32_t type, uint16_t tpid,
+ bool is_vf);
#endif /* __ROC_ESWITCH_H__ */
diff --git a/drivers/common/cnxk/roc_nix_priv.h b/drivers/common/cnxk/roc_nix_priv.h
index a582b9df33..8767a62577 100644
--- a/drivers/common/cnxk/roc_nix_priv.h
+++ b/drivers/common/cnxk/roc_nix_priv.h
@@ -473,9 +473,9 @@ int nix_lf_stat_reg_dump(uintptr_t nix_lf_base, uint64_t *data,
uint8_t lf_tx_stats, uint8_t lf_rx_stats);
int nix_lf_int_reg_dump(uintptr_t nix_lf_base, uint64_t *data, uint16_t qints,
uint16_t cints);
-int nix_q_ctx_get(struct dev *dev, uint8_t ctype, uint16_t qid,
- __io void **ctx_p);
+int nix_q_ctx_get(struct dev *dev, uint8_t ctype, uint16_t qid, __io void **ctx_p);
uint8_t nix_tm_lbk_relchan_get(struct nix *nix);
+int nix_vlan_tpid_set(struct mbox *mbox, uint16_t pcifunc, uint32_t type, uint16_t tpid);
/*
* Telemetry
diff --git a/drivers/common/cnxk/roc_nix_vlan.c b/drivers/common/cnxk/roc_nix_vlan.c
index abd2eb0571..db218593ad 100644
--- a/drivers/common/cnxk/roc_nix_vlan.c
+++ b/drivers/common/cnxk/roc_nix_vlan.c
@@ -211,18 +211,17 @@ roc_nix_vlan_insert_ena_dis(struct roc_nix *roc_nix,
}
int
-roc_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t tpid)
+nix_vlan_tpid_set(struct mbox *mbox, uint16_t pcifunc, uint32_t type, uint16_t tpid)
{
- struct nix *nix = roc_nix_to_nix_priv(roc_nix);
- struct dev *dev = &nix->dev;
- struct mbox *mbox = mbox_get(dev->mbox);
struct nix_set_vlan_tpid *tpid_cfg;
int rc = -ENOSPC;
- tpid_cfg = mbox_alloc_msg_nix_set_vlan_tpid(mbox);
+ /* Configuring for PF */
+ tpid_cfg = mbox_alloc_msg_nix_set_vlan_tpid(mbox_get(mbox));
if (tpid_cfg == NULL)
goto exit;
tpid_cfg->tpid = tpid;
+ tpid_cfg->hdr.pcifunc = pcifunc;
if (type & ROC_NIX_VLAN_TYPE_OUTER)
tpid_cfg->vlan_type = NIX_VLAN_TYPE_OUTER;
@@ -234,3 +233,17 @@ roc_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t tpid)
mbox_put(mbox);
return rc;
}
+
+int
+roc_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t tpid)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+ int rc;
+
+ rc = nix_vlan_tpid_set(dev->mbox, dev->pf_func, type, tpid);
+ if (rc)
+ plt_err("Failed to set tpid for PF, rc %d", rc);
+
+ return rc;
+}
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index feda34b852..78c421677d 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -91,6 +91,7 @@ INTERNAL {
roc_dpi_disable;
roc_dpi_enable;
roc_error_msg_get;
+ roc_eswitch_nix_vlan_tpid_set;
roc_eswitch_npc_mcam_delete_rule;
roc_eswitch_npc_mcam_rx_rule;
roc_eswitch_npc_mcam_tx_rule;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 08/24] net/cnxk: eswitch flow configurations
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (6 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 07/24] common/cnxk: interface to update VLAN TPID Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 09/24] net/cnxk: eswitch fastpath routines Harman Kalra
` (15 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
- Adding flow rules for eswitch PF and VF
- Interfaces to delete shift flow rules
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 43 ++-
drivers/net/cnxk/cnxk_eswitch.h | 25 +-
drivers/net/cnxk/cnxk_eswitch_devargs.c | 1 +
drivers/net/cnxk/cnxk_eswitch_flow.c | 445 ++++++++++++++++++++++++
drivers/net/cnxk/meson.build | 1 +
5 files changed, 511 insertions(+), 4 deletions(-)
create mode 100644 drivers/net/cnxk/cnxk_eswitch_flow.c
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index 563b224a6c..1cb0f0310a 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -2,11 +2,30 @@
* Copyright(C) 2023 Marvell.
*/
+#include <rte_thash.h>
+
#include <cnxk_eswitch.h>
#include <cnxk_rep.h>
#define CNXK_NIX_DEF_SQ_COUNT 512
+struct cnxk_esw_repr_hw_info *
+cnxk_eswitch_representor_hw_info(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func)
+{
+ struct cnxk_eswitch_devargs *esw_da;
+ int i, j;
+
+ /* Traversing the initialized represented list */
+ for (i = 0; i < eswitch_dev->nb_esw_da; i++) {
+ esw_da = &eswitch_dev->esw_da[i];
+ for (j = 0; j < esw_da->nb_repr_ports; j++) {
+ if (esw_da->repr_hw_info[j].hw_func == hw_func)
+ return &esw_da->repr_hw_info[j];
+ }
+ }
+ return NULL;
+}
+
static int
eswitch_hw_rsrc_cleanup(struct cnxk_eswitch_dev *eswitch_dev)
{
@@ -48,6 +67,10 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
cnxk_rep_dev_remove(eswitch_dev);
eswitch_hw_rsrc_cleanup(eswitch_dev);
+
+ /* Cleanup NPC rxtx flow rules */
+ cnxk_eswitch_flow_rules_remove_list(eswitch_dev, &eswitch_dev->esw_flow_list);
+
/* Check if this device is hosting common resource */
nix = roc_idev_npa_nix_get();
if (!nix || nix->pci_dev != pci_dev) {
@@ -58,7 +81,7 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
/* Try nix fini now */
rc = roc_nix_dev_fini(&eswitch_dev->nix);
if (rc == -EAGAIN) {
- plt_info("%s: common resource in use by other devices", pci_dev->name);
+ plt_esw_dbg("%s: common resource in use by other devices", pci_dev->name);
goto exit;
} else if (rc) {
plt_err("Failed in nix dev fini, rc=%d", rc);
@@ -154,6 +177,21 @@ cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev)
goto done;
}
+ /* Install eswitch PF mcam rules */
+ rc = cnxk_eswitch_pfvf_flow_rules_install(eswitch_dev, false);
+ if (rc) {
+ plt_err("Failed to install rxtx rules, rc %d", rc);
+ goto done;
+ }
+
+ /* Configure TPID for Eswitch PF LFs */
+ rc = roc_eswitch_nix_vlan_tpid_set(&eswitch_dev->nix, ROC_NIX_VLAN_TYPE_OUTER,
+ CNXK_ESWITCH_VLAN_TPID, false);
+ if (rc) {
+ plt_err("Failed to configure tpid, rc %d", rc);
+ goto done;
+ }
+
rc = roc_npc_mcam_enable_all_entries(&eswitch_dev->npc, 1);
if (rc) {
plt_err("Failed to enable NPC entries %d", rc);
@@ -600,6 +638,9 @@ eswitch_hw_rsrc_setup(struct cnxk_eswitch_dev *eswitch_dev)
if (rc)
goto rsrc_cleanup;
+ /* List for eswitch default flows */
+ TAILQ_INIT(&eswitch_dev->esw_flow_list);
+
return rc;
rsrc_cleanup:
eswitch_hw_rsrc_cleanup(eswitch_dev);
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index 4908c3ba95..470e4035bf 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -13,11 +13,10 @@
#include "cn10k_tx.h"
#define CNXK_ESWITCH_CTRL_MSG_SOCK_PATH "/tmp/cxk_rep_ctrl_msg_sock"
+#define CNXK_ESWITCH_VLAN_TPID ROC_ESWITCH_VLAN_TPID
#define CNXK_REP_ESWITCH_DEV_MZ "cnxk_eswitch_dev"
-#define CNXK_ESWITCH_VLAN_TPID 0x8100 /* TODO change */
#define CNXK_ESWITCH_MAX_TXQ 256
#define CNXK_ESWITCH_MAX_RXQ 256
-#define CNXK_ESWITCH_LBK_CHAN 63
#define CNXK_ESWITCH_VFPF_SHIFT 8
#define CNXK_ESWITCH_QUEUE_STATE_RELEASED 0
@@ -25,6 +24,7 @@
#define CNXK_ESWITCH_QUEUE_STATE_STARTED 2
#define CNXK_ESWITCH_QUEUE_STATE_STOPPED 3
+TAILQ_HEAD(eswitch_flow_list, roc_npc_flow);
enum cnxk_esw_da_pattern_type {
CNXK_ESW_DA_TYPE_LIST = 0,
CNXK_ESW_DA_TYPE_PFVF,
@@ -39,6 +39,9 @@ struct cnxk_esw_repr_hw_info {
uint16_t pfvf;
/* representor port id assigned to representee */
uint16_t port_id;
+ uint16_t num_flow_entries;
+
+ TAILQ_HEAD(flow_list, roc_npc_flow) repr_flow_list;
};
/* Structure representing per devarg information - this can be per representee
@@ -90,7 +93,6 @@ struct cnxk_eswitch_cxq {
uint8_t state;
};
-TAILQ_HEAD(eswitch_flow_list, roc_npc_flow);
struct cnxk_eswitch_dev {
/* Input parameters */
struct plt_pci_device *pci_dev;
@@ -116,6 +118,13 @@ struct cnxk_eswitch_dev {
uint16_t rep_cnt;
uint8_t configured;
+ /* NPC rxtx rules */
+ struct flow_list esw_flow_list;
+ uint16_t num_entries;
+ bool eswitch_vf_rules_setup;
+ uint16_t esw_pf_entry;
+ uint16_t esw_vf_entry;
+
/* Eswitch Representors Devargs */
uint16_t nb_esw_da;
uint16_t last_probed;
@@ -144,7 +153,10 @@ cnxk_eswitch_pmd_priv(void)
return mz->addr;
}
+/* HW Resources */
int cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev);
+struct cnxk_esw_repr_hw_info *cnxk_eswitch_representor_hw_info(struct cnxk_eswitch_dev *eswitch_dev,
+ uint16_t hw_func);
int cnxk_eswitch_repr_devargs(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
int cnxk_eswitch_representor_info_get(struct cnxk_eswitch_dev *eswitch_dev,
struct rte_eth_representor_info *info);
@@ -158,4 +170,11 @@ int cnxk_eswitch_rxq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
int cnxk_eswitch_rxq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
int cnxk_eswitch_txq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
int cnxk_eswitch_txq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+/* Flow Rules */
+int cnxk_eswitch_flow_rules_install(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func);
+int cnxk_eswitch_flow_rules_delete(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func);
+int cnxk_eswitch_pfvf_flow_rules_install(struct cnxk_eswitch_dev *eswitch_dev, bool is_vf);
+int cnxk_eswitch_flow_rule_shift(uint16_t hw_func, uint16_t *new_entry);
+int cnxk_eswitch_flow_rules_remove_list(struct cnxk_eswitch_dev *eswitch_dev,
+ struct flow_list *list);
#endif /* __CNXK_ESWITCH_H__ */
diff --git a/drivers/net/cnxk/cnxk_eswitch_devargs.c b/drivers/net/cnxk/cnxk_eswitch_devargs.c
index f1a1b05a99..aaefad2085 100644
--- a/drivers/net/cnxk/cnxk_eswitch_devargs.c
+++ b/drivers/net/cnxk/cnxk_eswitch_devargs.c
@@ -170,6 +170,7 @@ populate_repr_hw_info(struct cnxk_eswitch_dev *eswitch_dev, struct rte_eth_devar
esw_da->repr_hw_info[i].pfvf = (eth_da->type == RTE_ETH_REPRESENTOR_PF) ?
eth_da->ports[0] :
eth_da->representor_ports[i];
+ TAILQ_INIT(&esw_da->repr_hw_info[i].repr_flow_list);
plt_esw_dbg(" HW func %x index %d", hw_func, j);
}
diff --git a/drivers/net/cnxk/cnxk_eswitch_flow.c b/drivers/net/cnxk/cnxk_eswitch_flow.c
new file mode 100644
index 0000000000..f2ad87c75a
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_eswitch_flow.c
@@ -0,0 +1,445 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <rte_thash.h>
+
+#include <cnxk_eswitch.h>
+
+const uint8_t eswitch_vlan_rss_key[ROC_NIX_RSS_KEY_LEN] = {
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE,
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE,
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE,
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE,
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE,
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE};
+
+int
+cnxk_eswitch_flow_rules_remove_list(struct cnxk_eswitch_dev *eswitch_dev, struct flow_list *list)
+{
+ struct roc_npc_flow *flow, *tvar;
+ int rc = 0;
+
+ RTE_TAILQ_FOREACH_SAFE(flow, list, next, tvar) {
+ plt_esw_dbg("Removing flow %d", flow->mcam_id);
+ rc = roc_eswitch_npc_mcam_delete_rule(&eswitch_dev->npc, flow);
+ if (rc)
+ plt_err("Failed to delete rule %d", flow->mcam_id);
+ rc = roc_npc_mcam_free(&eswitch_dev->npc, flow);
+ if (rc)
+ plt_err("Failed to free entry %d", flow->mcam_id);
+ TAILQ_REMOVE(list, flow, next);
+ rte_free(flow);
+ }
+
+ return rc;
+}
+
+static int
+eswitch_npc_vlan_rss_configure(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
+{
+ struct roc_nix *roc_nix = roc_npc->roc_nix;
+ uint32_t qid, idx, hash, vlan_tci;
+ uint16_t *reta, reta_sz, id;
+ int rc = 0;
+
+ id = flow->mcam_id;
+ /* Setting up the key */
+ roc_nix_rss_key_set(roc_nix, eswitch_vlan_rss_key);
+
+ reta_sz = roc_nix->reta_sz;
+ reta = plt_zmalloc(reta_sz * sizeof(uint16_t), 0);
+ if (!reta) {
+ plt_err("Failed to allocate mem for reta table");
+ rc = -ENOMEM;
+ goto fail;
+ }
+ for (qid = 0; qid < reta_sz; qid++) {
+ vlan_tci = (1 << CNXK_ESWITCH_VFPF_SHIFT) | qid;
+ hash = rte_softrss(&vlan_tci, 1, eswitch_vlan_rss_key);
+ idx = hash & 0xFF;
+ reta[idx] = qid;
+ }
+ flow->mcam_id = id;
+ rc = roc_eswitch_npc_rss_action_configure(roc_npc, flow, FLOW_KEY_TYPE_VLAN, reta);
+ if (rc) {
+ plt_err("Failed to configure rss action, err %d", rc);
+ goto done;
+ }
+
+done:
+ plt_free(reta);
+fail:
+ return rc;
+}
+
+static int
+eswitch_pfvf_mcam_install_rules(struct cnxk_eswitch_dev *eswitch_dev, struct roc_npc_flow *flow,
+ bool is_vf)
+{
+ uint16_t vlan_tci = 0, hw_func;
+ int rc;
+
+ hw_func = eswitch_dev->npc.pf_func | is_vf;
+ if (!is_vf) {
+ /* Eswitch PF RX VLAN rule */
+ vlan_tci = 1ULL << CNXK_ESWITCH_VFPF_SHIFT;
+ rc = roc_eswitch_npc_mcam_rx_rule(&eswitch_dev->npc, flow, hw_func, vlan_tci,
+ 0xFF00);
+ if (rc) {
+ plt_err("Failed to install RX rule for ESW PF to ESW VF, rc %d", rc);
+ goto exit;
+ }
+ plt_esw_dbg("Installed eswitch PF RX rule %d", flow->mcam_id);
+ rc = eswitch_npc_vlan_rss_configure(&eswitch_dev->npc, flow);
+ if (rc)
+ goto exit;
+ flow->enable = true;
+ } else {
+ /* Eswitch VF RX VLAN rule */
+ rc = roc_eswitch_npc_mcam_rx_rule(&eswitch_dev->npc, flow, hw_func, vlan_tci,
+ 0xFF00);
+ if (rc) {
+ plt_err("Failed to install RX rule for ESW VF to ESW PF, rc %d", rc);
+ goto exit;
+ }
+ flow->enable = true;
+ plt_esw_dbg("Installed eswitch PF RX rule %d", flow->mcam_id);
+ }
+
+ return 0;
+exit:
+ return rc;
+}
+
+static int
+eswitch_npc_get_counter(struct roc_npc *npc, struct roc_npc_flow *flow)
+{
+ uint16_t ctr_id;
+ int rc;
+
+ rc = roc_npc_mcam_alloc_counter(npc, &ctr_id);
+ if (rc < 0) {
+ plt_err("Failed to allocate counter, rc %d", rc);
+ goto fail;
+ }
+ flow->ctr_id = ctr_id;
+ flow->use_ctr = true;
+
+ rc = roc_npc_mcam_clear_counter(npc, flow->ctr_id);
+ if (rc < 0) {
+ plt_err("Failed to clear counter idx %d, rc %d", flow->ctr_id, rc);
+ goto free;
+ }
+ return 0;
+free:
+ roc_npc_mcam_free_counter(npc, ctr_id);
+fail:
+ return rc;
+}
+
+static int
+eswitch_npc_get_counter_entry_ref(struct roc_npc *npc, struct roc_npc_flow *flow,
+ struct roc_npc_flow *ref_flow)
+{
+ int rc = 0, resp_count;
+
+ rc = eswitch_npc_get_counter(npc, flow);
+ if (rc)
+ goto free;
+
+ /* Allocate an entry viz higher priority than ref flow */
+ rc = roc_npc_mcam_alloc_entry(npc, flow, ref_flow, NPC_MCAM_HIGHER_PRIO, &resp_count);
+ if (rc) {
+ plt_err("Failed to allocate entry, err %d", rc);
+ goto free;
+ }
+ plt_esw_dbg("New entry %d ref entry %d resp_count %d", flow->mcam_id, ref_flow->mcam_id,
+ resp_count);
+
+ return 0;
+free:
+ roc_npc_mcam_free_counter(npc, flow->ctr_id);
+ return rc;
+}
+
+int
+cnxk_eswitch_flow_rule_shift(uint16_t hw_func, uint16_t *entry)
+{
+ struct cnxk_esw_repr_hw_info *repr_info;
+ struct cnxk_eswitch_dev *eswitch_dev;
+ struct roc_npc_flow *ref_flow, *flow;
+ uint16_t curr_entry, new_entry;
+ int rc = 0, resp_count;
+
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ repr_info = cnxk_eswitch_representor_hw_info(eswitch_dev, hw_func);
+ if (!repr_info) {
+ plt_warn("Failed to get representor group for %x", hw_func);
+ rc = -ENOENT;
+ goto fail;
+ }
+
+ ref_flow = TAILQ_FIRST(&repr_info->repr_flow_list);
+ if (*entry > ref_flow->mcam_id) {
+ flow = plt_zmalloc(sizeof(struct roc_npc_flow), 0);
+ if (!flow) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ /* Allocate a higher priority flow rule */
+ rc = roc_npc_mcam_alloc_entry(&eswitch_dev->npc, flow, ref_flow,
+ NPC_MCAM_HIGHER_PRIO, &resp_count);
+ if (rc < 0) {
+ plt_err("Failed to allocate a newmcam entry, rc %d", rc);
+ goto fail;
+ }
+
+ if (flow->mcam_id > ref_flow->mcam_id) {
+ plt_err("New flow %d is still at higher priority than ref_flow %d",
+ flow->mcam_id, ref_flow->mcam_id);
+ rc = -EINVAL;
+ goto free_entry;
+ }
+
+ plt_info("Before shift: HW_func %x curr_entry %d ref flow id %d new_entry %d",
+ hw_func, *entry, ref_flow->mcam_id, flow->mcam_id);
+
+ curr_entry = *entry;
+ new_entry = flow->mcam_id;
+
+ rc = roc_npc_mcam_move(&eswitch_dev->npc, curr_entry, new_entry);
+ if (rc) {
+ plt_err("Failed to shift the new index %d to curr index %d, err %d", *entry,
+ curr_entry, rc);
+ goto free_entry;
+ }
+ *entry = flow->mcam_id;
+
+ /* Freeing the current entry */
+ rc = roc_npc_mcam_free_entry(&eswitch_dev->npc, curr_entry);
+ if (rc) {
+ plt_err("Failed to free the old entry. err %d", rc);
+ goto free_entry;
+ }
+
+ plt_free(flow);
+ plt_info("After shift: HW_func %x old_entry %d new_entry %d", hw_func, curr_entry,
+ *entry);
+ }
+
+ return 0;
+free_entry:
+
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_flow_rules_delete(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func)
+{
+ struct cnxk_esw_repr_hw_info *repr_info;
+ struct flow_list *list;
+ int rc = 0;
+
+ repr_info = cnxk_eswitch_representor_hw_info(eswitch_dev, hw_func);
+ if (!repr_info) {
+ plt_warn("Failed to get representor group for %x", hw_func);
+ rc = -ENOENT;
+ goto fail;
+ }
+ list = &repr_info->repr_flow_list;
+
+ plt_esw_dbg("Deleting flows for %x", hw_func);
+ rc = cnxk_eswitch_flow_rules_remove_list(eswitch_dev, list);
+ if (rc)
+ plt_err("Failed to delete rules for hw func %x", hw_func);
+
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_flow_rules_install(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func)
+{
+ struct roc_npc_flow *rx_flow, *tx_flow, *flow_iter, *esw_pf_flow = NULL;
+ struct cnxk_esw_repr_hw_info *repr_info;
+ struct flow_list *list;
+ uint16_t vlan_tci;
+ int rc = 0;
+
+ repr_info = cnxk_eswitch_representor_hw_info(eswitch_dev, hw_func);
+ if (!repr_info) {
+ plt_err("Failed to get representor group for %x", hw_func);
+ rc = -EINVAL;
+ goto fail;
+ }
+ list = &repr_info->repr_flow_list;
+
+ /* Taking ESW PF as reference entry for installing new rules */
+ TAILQ_FOREACH(flow_iter, &eswitch_dev->esw_flow_list, next) {
+ if (flow_iter->mcam_id == eswitch_dev->esw_pf_entry) {
+ esw_pf_flow = flow_iter;
+ break;
+ }
+ }
+
+ if (!esw_pf_flow) {
+ plt_err("Failed to get the ESW PF flow");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Installing RX rule */
+ rx_flow = plt_zmalloc(sizeof(struct roc_npc_flow), 0);
+ if (!rx_flow) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ rc = eswitch_npc_get_counter_entry_ref(&eswitch_dev->npc, rx_flow, esw_pf_flow);
+ if (rc) {
+ plt_err("Failed to get counter and mcam entry, rc %d", rc);
+ goto free_rx_flow;
+ }
+
+ /* VLAN TCI value for this representee is the rep id from AF driver */
+ vlan_tci = repr_info->rep_id;
+ rc = roc_eswitch_npc_mcam_rx_rule(&eswitch_dev->npc, rx_flow, hw_func, vlan_tci, 0xFFFF);
+ if (rc) {
+ plt_err("Failed to install RX rule for ESW PF to ESW VF, rc %d", rc);
+ goto free_rx_entry;
+ }
+ rx_flow->enable = true;
+ /* List in ascending order of mcam entries */
+ TAILQ_FOREACH(flow_iter, list, next) {
+ if (flow_iter->mcam_id > rx_flow->mcam_id) {
+ TAILQ_INSERT_BEFORE(flow_iter, rx_flow, next);
+ goto done_rx;
+ }
+ }
+ TAILQ_INSERT_TAIL(list, rx_flow, next);
+done_rx:
+ repr_info->num_flow_entries++;
+ plt_esw_dbg("Installed RX flow rule %d for representee %x with vlan tci %x MCAM id %d",
+ eswitch_dev->num_entries, hw_func, vlan_tci, rx_flow->mcam_id);
+
+ /* Installing TX rule */
+ tx_flow = plt_zmalloc(sizeof(struct roc_npc_flow), 0);
+ if (!tx_flow) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto remove_rx_rule;
+ }
+
+ rc = eswitch_npc_get_counter_entry_ref(&eswitch_dev->npc, tx_flow, esw_pf_flow);
+ if (rc) {
+ plt_err("Failed to get counter and mcam entry, rc %d", rc);
+ goto free_tx_flow;
+ }
+
+ vlan_tci = (1ULL << CNXK_ESWITCH_VFPF_SHIFT) | repr_info->rep_id;
+ rc = roc_eswitch_npc_mcam_tx_rule(&eswitch_dev->npc, tx_flow, hw_func, vlan_tci);
+ if (rc) {
+ plt_err("Failed to install RX rule for ESW PF to ESW VF, rc %d", rc);
+ goto free_tx_entry;
+ }
+ tx_flow->enable = true;
+ /* List in ascending order of mcam entries */
+ TAILQ_FOREACH(flow_iter, list, next) {
+ if (flow_iter->mcam_id > tx_flow->mcam_id) {
+ TAILQ_INSERT_BEFORE(flow_iter, tx_flow, next);
+ goto done_tx;
+ }
+ }
+ TAILQ_INSERT_TAIL(list, tx_flow, next);
+done_tx:
+ repr_info->num_flow_entries++;
+ plt_esw_dbg("Installed TX flow rule %d for representee %x with vlan tci %x MCAM id %d",
+ repr_info->num_flow_entries, hw_func, vlan_tci, tx_flow->mcam_id);
+
+ return 0;
+free_tx_entry:
+ roc_npc_mcam_free(&eswitch_dev->npc, tx_flow);
+free_tx_flow:
+ rte_free(tx_flow);
+remove_rx_rule:
+ TAILQ_REMOVE(list, rx_flow, next);
+free_rx_entry:
+ roc_npc_mcam_free(&eswitch_dev->npc, rx_flow);
+free_rx_flow:
+ rte_free(rx_flow);
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_pfvf_flow_rules_install(struct cnxk_eswitch_dev *eswitch_dev, bool is_vf)
+{
+ struct roc_npc_flow *flow, *flow_iter;
+ struct flow_list *list;
+ int rc = 0;
+
+ list = &eswitch_dev->esw_flow_list;
+ flow = plt_zmalloc(sizeof(struct roc_npc_flow), 0);
+ if (!flow) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ rc = eswitch_npc_get_counter(&eswitch_dev->npc, flow);
+ if (rc) {
+ plt_err("Failed to get counter and mcam entry, rc %d", rc);
+ goto free_flow;
+ }
+ if (!is_vf) {
+ /* Reserving an entry for esw VF but will not be installed */
+ rc = roc_npc_get_free_mcam_entry(&eswitch_dev->npc, flow);
+ if (rc < 0) {
+ plt_err("Failed to allocate entry for vf, err %d", rc);
+ goto free_flow;
+ }
+ eswitch_dev->esw_vf_entry = flow->mcam_id;
+ /* Allocate an entry for esw PF */
+ rc = eswitch_npc_get_counter_entry_ref(&eswitch_dev->npc, flow, flow);
+ if (rc) {
+ plt_err("Failed to allocate entry for pf, err %d", rc);
+ goto free_flow;
+ }
+ eswitch_dev->esw_pf_entry = flow->mcam_id;
+ plt_esw_dbg("Allocated entries for esw: PF %d and VF %d", eswitch_dev->esw_pf_entry,
+ eswitch_dev->esw_vf_entry);
+ } else {
+ flow->mcam_id = eswitch_dev->esw_vf_entry;
+ }
+
+ rc = eswitch_pfvf_mcam_install_rules(eswitch_dev, flow, is_vf);
+ if (rc) {
+ plt_err("Failed to install entries, rc %d", rc);
+ goto free_flow;
+ }
+
+ /* List in ascending order of mcam entries */
+ TAILQ_FOREACH(flow_iter, list, next) {
+ if (flow_iter->mcam_id > flow->mcam_id) {
+ TAILQ_INSERT_BEFORE(flow_iter, flow, next);
+ goto done;
+ }
+ }
+ TAILQ_INSERT_TAIL(list, flow, next);
+done:
+ eswitch_dev->num_entries++;
+ plt_esw_dbg("Installed new eswitch flow rule %d with MCAM id %d", eswitch_dev->num_entries,
+ flow->mcam_id);
+
+ return 0;
+
+free_flow:
+ cnxk_eswitch_flow_rules_remove_list(eswitch_dev, &eswitch_dev->esw_flow_list);
+fail:
+ return rc;
+}
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index fcd5d3d569..488e89253d 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -30,6 +30,7 @@ sources = files(
'cnxk_ethdev_sec_telemetry.c',
'cnxk_eswitch.c',
'cnxk_eswitch_devargs.c',
+ 'cnxk_eswitch_flow.c',
'cnxk_link.c',
'cnxk_lookup.c',
'cnxk_ptp.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 09/24] net/cnxk: eswitch fastpath routines
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (7 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 08/24] net/cnxk: eswitch flow configurations Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 10/24] net/cnxk: add representor control plane Harman Kalra
` (14 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Implementing fastpath RX and TX fast path routines which can be
invoked from respective representors rx burst and tx burst
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.h | 5 +
drivers/net/cnxk/cnxk_eswitch_rxtx.c | 212 +++++++++++++++++++++++++++
drivers/net/cnxk/meson.build | 1 +
3 files changed, 218 insertions(+)
create mode 100644 drivers/net/cnxk/cnxk_eswitch_rxtx.c
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index 470e4035bf..d92c4f4778 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -177,4 +177,9 @@ int cnxk_eswitch_pfvf_flow_rules_install(struct cnxk_eswitch_dev *eswitch_dev, b
int cnxk_eswitch_flow_rule_shift(uint16_t hw_func, uint16_t *new_entry);
int cnxk_eswitch_flow_rules_remove_list(struct cnxk_eswitch_dev *eswitch_dev,
struct flow_list *list);
+/* RX TX fastpath routines */
+uint16_t cnxk_eswitch_dev_tx_burst(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid,
+ struct rte_mbuf **pkts, uint16_t nb_tx, const uint16_t flags);
+uint16_t cnxk_eswitch_dev_rx_burst(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid,
+ struct rte_mbuf **pkts, uint16_t nb_pkts);
#endif /* __CNXK_ESWITCH_H__ */
diff --git a/drivers/net/cnxk/cnxk_eswitch_rxtx.c b/drivers/net/cnxk/cnxk_eswitch_rxtx.c
new file mode 100644
index 0000000000..b5a69e3338
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_eswitch_rxtx.c
@@ -0,0 +1,212 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <cnxk_eswitch.h>
+
+static __rte_always_inline struct rte_mbuf *
+eswitch_nix_get_mbuf_from_cqe(void *cq, const uint64_t data_off)
+{
+ rte_iova_t buff;
+
+ /* Skip CQE, NIX_RX_PARSE_S and SG HDR(9 DWORDs) and peek buff addr */
+ buff = *((rte_iova_t *)((uint64_t *)cq + 9));
+ return (struct rte_mbuf *)(buff - data_off);
+}
+
+static inline uint64_t
+eswitch_nix_rx_nb_pkts(struct roc_nix_cq *cq, const uint64_t wdata, const uint32_t qmask)
+{
+ uint64_t reg, head, tail;
+ uint32_t available;
+
+ /* Update the available count if cached value is not enough */
+
+ /* Use LDADDA version to avoid reorder */
+ reg = roc_atomic64_add_sync(wdata, cq->status);
+ /* CQ_OP_STATUS operation error */
+ if (reg & BIT_ULL(NIX_CQ_OP_STAT_OP_ERR) || reg & BIT_ULL(NIX_CQ_OP_STAT_CQ_ERR))
+ return 0;
+
+ tail = reg & 0xFFFFF;
+ head = (reg >> 20) & 0xFFFFF;
+ if (tail < head)
+ available = tail - head + qmask + 1;
+ else
+ available = tail - head;
+
+ return available;
+}
+
+static inline void
+nix_cn9k_xmit_one(uint64_t *cmd, void *lmt_addr, const plt_iova_t io_addr)
+{
+ uint64_t lmt_status;
+
+ do {
+ roc_lmt_mov(lmt_addr, cmd, 0);
+ lmt_status = roc_lmt_submit_ldeor(io_addr);
+ } while (lmt_status == 0);
+}
+
+uint16_t
+cnxk_eswitch_dev_tx_burst(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid,
+ struct rte_mbuf **pkts, uint16_t nb_xmit, const uint16_t flags)
+{
+ struct roc_nix_sq *sq = &eswitch_dev->txq[qid].sqs;
+ struct roc_nix_rq *rq = &eswitch_dev->rxq[qid].rqs;
+ uint16_t lmt_id, pkt = 0, nb_tx = 0;
+ struct nix_send_ext_s *send_hdr_ext;
+ uint64_t aura_handle, cmd[6], data;
+ struct nix_send_hdr_s *send_hdr;
+ uint16_t vlan_tci = qid;
+ union nix_send_sg_s *sg;
+ uintptr_t lmt_base, pa;
+ int64_t fc_pkts, dw_m1;
+ rte_iova_t io_addr;
+
+ if (unlikely(eswitch_dev->txq[qid].state != CNXK_ESWITCH_QUEUE_STATE_STARTED))
+ return 0;
+
+ lmt_base = sq->roc_nix->lmt_base;
+ io_addr = sq->io_addr;
+ aura_handle = rq->aura_handle;
+ /* Get LMT base address and LMT ID as per thread ID */
+ lmt_id = roc_plt_control_lmt_id_get();
+ lmt_base += ((uint64_t)lmt_id << ROC_LMT_LINE_SIZE_LOG2);
+ /* Double word minus 1: LMTST size-1 in units of 128 bits */
+ /* 2(HDR) + 2(EXT_HDR) + 1(SG) + 1(IOVA) = 6/2 - 1 = 2 */
+ dw_m1 = cn10k_nix_tx_ext_subs(flags) + 1;
+
+ memset(cmd, 0, sizeof(cmd));
+ send_hdr = (struct nix_send_hdr_s *)&cmd[0];
+ send_hdr->w0.sizem1 = dw_m1;
+ send_hdr->w0.sq = sq->qid;
+
+ if (dw_m1 >= 2) {
+ send_hdr_ext = (struct nix_send_ext_s *)&cmd[2];
+ send_hdr_ext->w0.subdc = NIX_SUBDC_EXT;
+ if (flags & NIX_TX_OFFLOAD_VLAN_QINQ_F) {
+ send_hdr_ext->w1.vlan0_ins_ena = true;
+ /* 2B before end of l2 header */
+ send_hdr_ext->w1.vlan0_ins_ptr = 12;
+ send_hdr_ext->w1.vlan0_ins_tci = 0;
+ }
+ sg = (union nix_send_sg_s *)&cmd[4];
+ } else {
+ sg = (union nix_send_sg_s *)&cmd[2];
+ }
+
+ sg->subdc = NIX_SUBDC_SG;
+ sg->segs = 1;
+ sg->ld_type = NIX_SENDLDTYPE_LDD;
+
+ /* Tx */
+ fc_pkts = ((int64_t)sq->nb_sqb_bufs_adj - *((uint64_t *)sq->fc)) << sq->sqes_per_sqb_log2;
+
+ if (fc_pkts < 0)
+ nb_tx = 0;
+ else
+ nb_tx = PLT_MIN(nb_xmit, (uint64_t)fc_pkts);
+
+ for (pkt = 0; pkt < nb_tx; pkt++) {
+ send_hdr->w0.total = pkts[pkt]->pkt_len;
+ /* TODO: revsit */
+ if (pkts[pkt]->pool) {
+ aura_handle = pkts[pkt]->pool->pool_id;
+ send_hdr->w0.aura = roc_npa_aura_handle_to_aura(aura_handle);
+ } else {
+ send_hdr->w0.df = 1;
+ }
+ if (flags & NIX_TX_OFFLOAD_VLAN_QINQ_F)
+ send_hdr_ext->w1.vlan0_ins_tci = vlan_tci;
+ sg->seg1_size = pkts[pkt]->pkt_len;
+ *(plt_iova_t *)(sg + 1) = rte_mbuf_data_iova(pkts[pkt]);
+
+ plt_esw_dbg("Transmitting pkt %d (%p) vlan tci %x on sq %d esw qid %d", pkt,
+ pkts[pkt], vlan_tci, sq->qid, qid);
+ if (roc_model_is_cn9k()) {
+ nix_cn9k_xmit_one(cmd, sq->lmt_addr, sq->io_addr);
+ } else {
+ cn10k_nix_xmit_mv_lmt_base(lmt_base, cmd, flags);
+ /* PA<6:4> = LMTST size-1 in units of 128 bits. Size of the first LMTST in
+ * burst.
+ */
+ pa = io_addr | (dw_m1 << 4);
+ data &= ~0x7ULL;
+ /*<15:12> = CNTM1: Count minus one of LMTSTs in the burst */
+ data = (0ULL << 12);
+ /* *<10:0> = LMT_ID: Identifies which LMT line is used for the first LMTST
+ */
+ data |= (uint64_t)lmt_id;
+
+ /* STEOR0 */
+ roc_lmt_submit_steorl(data, pa);
+ rte_io_wmb();
+ }
+ }
+
+ return nb_tx;
+}
+
+uint16_t
+cnxk_eswitch_dev_rx_burst(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid,
+ struct rte_mbuf **pkts, uint16_t nb_pkts)
+{
+ struct roc_nix_rq *rq = &eswitch_dev->rxq[qid].rqs;
+ struct roc_nix_cq *cq = &eswitch_dev->cxq[qid].cqs;
+ const union nix_rx_parse_u *rx;
+ struct nix_cqe_hdr_s *cqe;
+ uint64_t pkt = 0, nb_rx;
+ struct rte_mbuf *mbuf;
+ uint64_t wdata;
+ uint32_t qmask;
+ uintptr_t desc;
+ uint32_t head;
+
+ if (unlikely(eswitch_dev->rxq[qid].state != CNXK_ESWITCH_QUEUE_STATE_STARTED))
+ return 0;
+
+ wdata = cq->wdata;
+ qmask = cq->qmask;
+ desc = (uintptr_t)cq->desc_base;
+ nb_rx = eswitch_nix_rx_nb_pkts(cq, wdata, qmask);
+ nb_rx = RTE_MIN(nb_rx, nb_pkts);
+ head = cq->head;
+
+ /* Nothing to receive */
+ if (!nb_rx)
+ return 0;
+
+ /* Rx */
+ for (pkt = 0; pkt < nb_rx; pkt++) {
+ /* Prefetch N desc ahead */
+ rte_prefetch_non_temporal((void *)(desc + (CQE_SZ((head + 2) & qmask))));
+ cqe = (struct nix_cqe_hdr_s *)(desc + CQE_SZ(head));
+ rx = (const union nix_rx_parse_u *)((const uint64_t *)cqe + 1);
+
+ /* Skip QE, NIX_RX_PARSE_S and SG HDR(9 DWORDs) and peek buff addr */
+ mbuf = eswitch_nix_get_mbuf_from_cqe(cqe, rq->first_skip);
+ mbuf->pkt_len = rx->pkt_lenm1 + 1;
+ mbuf->data_len = rx->pkt_lenm1 + 1;
+ mbuf->data_off = 128;
+ /* Rx parse to capture vlan info */
+ if (rx->vtag0_valid)
+ mbuf->vlan_tci = rx->vtag0_tci;
+ /* Populate RSS hash */
+ mbuf->hash.rss = cqe->tag;
+ mbuf->ol_flags |= RTE_MBUF_F_RX_RSS_HASH;
+ pkts[pkt] = mbuf;
+ roc_prefetch_store_keep(mbuf);
+ plt_esw_dbg("Packet %d rec on queue %d esw qid %d hash %x mbuf %p vlan tci %d",
+ (uint32_t)pkt, rq->qid, qid, mbuf->hash.rss, mbuf, mbuf->vlan_tci);
+ head++;
+ head &= qmask;
+ }
+
+ /* Free all the CQs that we've processed */
+ rte_write64_relaxed((wdata | nb_rx), (void *)cq->door);
+ cq->head = head;
+
+ return nb_rx;
+}
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index 488e89253d..7121845dc6 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -31,6 +31,7 @@ sources = files(
'cnxk_eswitch.c',
'cnxk_eswitch_devargs.c',
'cnxk_eswitch_flow.c',
+ 'cnxk_eswitch_rxtx.c',
'cnxk_link.c',
'cnxk_lookup.c',
'cnxk_ptp.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 10/24] net/cnxk: add representor control plane
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (8 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 09/24] net/cnxk: eswitch fastpath routines Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 11/24] common/cnxk: representee notification callback Harman Kalra
` (13 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Implementing the control path for representor ports, where represented
ports can be configured using TLV messaging.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 67 ++-
drivers/net/cnxk/cnxk_eswitch.h | 8 +
drivers/net/cnxk/cnxk_rep.c | 52 ++
drivers/net/cnxk/cnxk_rep.h | 3 +
drivers/net/cnxk/cnxk_rep_msg.c | 823 ++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 95 ++++
drivers/net/cnxk/meson.build | 1 +
7 files changed, 1041 insertions(+), 8 deletions(-)
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.c
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.h
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index 1cb0f0310a..ffcf89b1b1 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -9,6 +9,27 @@
#define CNXK_NIX_DEF_SQ_COUNT 512
+int
+cnxk_eswitch_representor_id(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func,
+ uint16_t *rep_id)
+{
+ struct cnxk_esw_repr_hw_info *repr_info;
+ int rc = 0;
+
+ repr_info = cnxk_eswitch_representor_hw_info(eswitch_dev, hw_func);
+ if (!repr_info) {
+ plt_warn("Failed to get representor group for %x", hw_func);
+ rc = -ENOENT;
+ goto fail;
+ }
+
+ *rep_id = repr_info->rep_id;
+
+ return 0;
+fail:
+ return rc;
+}
+
struct cnxk_esw_repr_hw_info *
cnxk_eswitch_representor_hw_info(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func)
{
@@ -63,8 +84,38 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
eswitch_dev = cnxk_eswitch_pmd_priv();
/* Remove representor devices associated with PF */
- if (eswitch_dev->repr_cnt.nb_repr_created)
+ if (eswitch_dev->repr_cnt.nb_repr_created) {
+ /* Exiting the rep msg ctrl thread */
+ if (eswitch_dev->start_ctrl_msg_thrd) {
+ uint32_t sunlen;
+ struct sockaddr_un sun = {0};
+ int sock_fd;
+
+ eswitch_dev->start_ctrl_msg_thrd = false;
+ if (!eswitch_dev->client_connected) {
+ plt_esw_dbg("Establishing connection for teardown");
+ sock_fd = socket(AF_UNIX, SOCK_STREAM, 0);
+ if (sock_fd == -1) {
+ plt_err("Failed to open socket. err %d", -errno);
+ return -errno;
+ }
+ sun.sun_family = AF_UNIX;
+ sunlen = sizeof(struct sockaddr_un);
+ strncpy(sun.sun_path, CNXK_ESWITCH_CTRL_MSG_SOCK_PATH,
+ sizeof(sun.sun_path) - 1);
+
+ if (connect(sock_fd, (struct sockaddr *)&sun, sunlen) < 0) {
+ plt_err("Failed to connect socket: %s, err %d",
+ CNXK_ESWITCH_CTRL_MSG_SOCK_PATH, errno);
+ return -errno;
+ }
+ }
+ rte_thread_join(eswitch_dev->rep_ctrl_msg_thread, NULL);
+ }
+
+ /* Remove representor devices associated with PF */
cnxk_rep_dev_remove(eswitch_dev);
+ }
eswitch_hw_rsrc_cleanup(eswitch_dev);
@@ -170,13 +221,6 @@ cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev)
goto done;
}
- /* Enable Rx in NPC */
- rc = roc_nix_npc_rx_ena_dis(&eswitch_dev->nix, true);
- if (rc) {
- plt_err("Failed to enable NPC rx %d", rc);
- goto done;
- }
-
/* Install eswitch PF mcam rules */
rc = cnxk_eswitch_pfvf_flow_rules_install(eswitch_dev, false);
if (rc) {
@@ -192,6 +236,13 @@ cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev)
goto done;
}
+ /* Enable Rx in NPC */
+ rc = roc_nix_npc_rx_ena_dis(&eswitch_dev->nix, true);
+ if (rc) {
+ plt_err("Failed to enable NPC rx %d", rc);
+ goto done;
+ }
+
rc = roc_npc_mcam_enable_all_entries(&eswitch_dev->npc, 1);
if (rc) {
plt_err("Failed to enable NPC entries %d", rc);
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index d92c4f4778..a2f4aa0fcc 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -133,6 +133,12 @@ struct cnxk_eswitch_dev {
/* No of representors */
struct cnxk_eswitch_repr_cnt repr_cnt;
+ /* Representor control channel field */
+ bool start_ctrl_msg_thrd;
+ rte_thread_t rep_ctrl_msg_thread;
+ bool client_connected;
+ int sock_fd;
+
/* Port representor fields */
rte_spinlock_t rep_lock;
uint16_t nb_switch_domain;
@@ -155,6 +161,8 @@ cnxk_eswitch_pmd_priv(void)
/* HW Resources */
int cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_eswitch_representor_id(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func,
+ uint16_t *rep_id);
struct cnxk_esw_repr_hw_info *cnxk_eswitch_representor_hw_info(struct cnxk_eswitch_dev *eswitch_dev,
uint16_t hw_func);
int cnxk_eswitch_repr_devargs(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
index 295bea3724..f8e1d5b965 100644
--- a/drivers/net/cnxk/cnxk_rep.c
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -2,6 +2,7 @@
* Copyright(C) 2023 Marvell.
*/
#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
#define PF_SHIFT 10
#define PF_MASK 0x3F
@@ -25,6 +26,48 @@ switch_domain_id_allocate(struct cnxk_eswitch_dev *eswitch_dev, uint16_t pf)
return RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID;
}
+int
+cnxk_rep_state_update(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, uint16_t *rep_id)
+{
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct rte_eth_dev *rep_eth_dev;
+ int i, rc = 0;
+
+ /* Delete the individual PFVF flows as common eswitch VF rule will be used. */
+ rc = cnxk_eswitch_flow_rules_delete(eswitch_dev, hw_func);
+ if (rc) {
+ if (rc != -ENOENT) {
+ plt_err("Failed to delete %x flow rules", hw_func);
+ goto fail;
+ }
+ }
+ /* Rep ID for respective HW func */
+ rc = cnxk_eswitch_representor_id(eswitch_dev, hw_func, rep_id);
+ if (rc) {
+ if (rc != -ENOENT) {
+ plt_err("Failed to get rep info for %x", hw_func);
+ goto fail;
+ }
+ }
+ /* Update the state - representee is standalone or part of companian app */
+ for (i = 0; i < eswitch_dev->repr_cnt.nb_repr_probed; i++) {
+ rep_eth_dev = eswitch_dev->rep_info[i].rep_eth_dev;
+ if (!rep_eth_dev) {
+ plt_err("Failed to get rep ethdev handle");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ if (rep_dev->hw_func == hw_func && rep_dev->is_vf_active)
+ rep_dev->native_repte = false;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
int
cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev)
{
@@ -250,6 +293,15 @@ cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswi
}
eswitch_dev->last_probed = i;
+ /* Launch a thread to handle control messages */
+ if (!eswitch_dev->start_ctrl_msg_thrd) {
+ rc = cnxk_rep_msg_control_thread_launch(eswitch_dev);
+ if (rc) {
+ plt_err("Failed to launch message ctrl thread");
+ goto fail;
+ }
+ }
+
return 0;
fail:
return rc;
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index 2cb3ae8ac5..a62d9b0ae8 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -16,6 +16,8 @@ struct cnxk_rep_dev {
uint16_t switch_domain_id;
struct cnxk_eswitch_dev *parent_dev;
uint16_t hw_func;
+ bool is_vf_active;
+ bool native_repte;
uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
};
@@ -46,5 +48,6 @@ int cnxk_rep_dev_close(struct rte_eth_dev *eth_dev);
int cnxk_rep_stats_get(struct rte_eth_dev *eth_dev, struct rte_eth_stats *stats);
int cnxk_rep_stats_reset(struct rte_eth_dev *eth_dev);
int cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops);
+int cnxk_rep_state_update(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, uint16_t *rep_id);
#endif /* __CNXK_REP_H__ */
diff --git a/drivers/net/cnxk/cnxk_rep_msg.c b/drivers/net/cnxk/cnxk_rep_msg.c
new file mode 100644
index 0000000000..f538c3f27f
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_msg.c
@@ -0,0 +1,823 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
+
+#define CTRL_MSG_RCV_TIMEOUT_MS 2000
+#define CTRL_MSG_READY_WAIT_US 2000
+#define CTRL_MSG_THRD_NAME_LEN 35
+#define CTRL_MSG_BUFFER_SZ 1500
+#define CTRL_MSG_SIGNATURE 0xcdacdeadbeefcadc
+
+static void
+close_socket(int fd)
+{
+ close(fd);
+ unlink(CNXK_ESWITCH_CTRL_MSG_SOCK_PATH);
+}
+
+static int
+receive_control_message(int socketfd, void *data, uint32_t len)
+{
+ char ctl[CMSG_SPACE(sizeof(int)) + CMSG_SPACE(sizeof(struct ucred))] = {0};
+ struct ucred *cr __rte_unused;
+ struct msghdr mh = {0};
+ struct cmsghdr *cmsg;
+ static uint64_t rec;
+ struct iovec iov[1];
+ ssize_t size;
+ int afd = -1;
+
+ iov[0].iov_base = data;
+ iov[0].iov_len = len;
+ mh.msg_iov = iov;
+ mh.msg_iovlen = 1;
+ mh.msg_control = ctl;
+ mh.msg_controllen = sizeof(ctl);
+
+ size = recvmsg(socketfd, &mh, MSG_DONTWAIT);
+ if (size < 0) {
+ if (errno == EAGAIN)
+ return 0;
+ plt_err("recvmsg err %d invalid size %ld", errno, size);
+ return -errno;
+ } else if (size == 0) {
+ return 0;
+ }
+
+ rec++;
+ plt_rep_dbg("Packet %" PRId64 " Received %" PRId64 " bytes over socketfd %d",
+ rec, size, socketfd);
+
+ cr = 0;
+ cmsg = CMSG_FIRSTHDR(&mh);
+ while (cmsg) {
+ if (cmsg->cmsg_level == SOL_SOCKET) {
+ if (cmsg->cmsg_type == SCM_CREDENTIALS) {
+ cr = (struct ucred *)CMSG_DATA(cmsg);
+ } else if (cmsg->cmsg_type == SCM_RIGHTS) {
+ rte_memcpy(&afd, CMSG_DATA(cmsg), sizeof(int));
+ plt_rep_dbg("afd %d", afd);
+ }
+ }
+ cmsg = CMSG_NXTHDR(&mh, cmsg);
+ }
+ return size;
+}
+
+static int
+send_message_on_socket(int socketfd, void *data, uint32_t len, int afd)
+{
+ char ctl[CMSG_SPACE(sizeof(int))];
+ struct msghdr mh = {0};
+ struct cmsghdr *cmsg;
+ static uint64_t sent;
+ struct iovec iov[1];
+ int size;
+
+ iov[0].iov_base = data;
+ iov[0].iov_len = len;
+ mh.msg_iov = iov;
+ mh.msg_iovlen = 1;
+
+ if (afd > 0) {
+ memset(&ctl, 0, sizeof(ctl));
+ mh.msg_control = ctl;
+ mh.msg_controllen = sizeof(ctl);
+ cmsg = CMSG_FIRSTHDR(&mh);
+ cmsg->cmsg_len = CMSG_LEN(sizeof(int));
+ cmsg->cmsg_level = SOL_SOCKET;
+ cmsg->cmsg_type = SCM_RIGHTS;
+ rte_memcpy(CMSG_DATA(cmsg), &afd, sizeof(int));
+ }
+
+ size = sendmsg(socketfd, &mh, MSG_DONTWAIT);
+ if (size < 0) {
+ if (errno == EAGAIN)
+ return 0;
+ plt_err("Failed to send message, err %d", -errno);
+ return -errno;
+ } else if (size == 0) {
+ return 0;
+ }
+ sent++;
+ plt_rep_dbg("Sent %" PRId64 " packets of size %d on socketfd %d", sent, size, socketfd);
+
+ return size;
+}
+
+static int
+open_socket_ctrl_channel(void)
+{
+ struct sockaddr_un un;
+ int sock_fd;
+
+ sock_fd = socket(AF_UNIX, SOCK_STREAM, 0);
+ if (sock_fd < 0) {
+ RTE_LOG(ERR, EAL, "failed to create unix socket\n");
+ return -1;
+ }
+
+ /* Set unix socket path and bind */
+ memset(&un, 0, sizeof(un));
+ un.sun_family = AF_UNIX;
+
+ if (strlen(CNXK_ESWITCH_CTRL_MSG_SOCK_PATH) > sizeof(un.sun_path) - 1) {
+ plt_err("Server socket path too long: %s", CNXK_ESWITCH_CTRL_MSG_SOCK_PATH);
+ close(sock_fd);
+ return -E2BIG;
+ }
+
+ if (remove(CNXK_ESWITCH_CTRL_MSG_SOCK_PATH) == -1 && errno != ENOENT) {
+ plt_err("remove-%s", CNXK_ESWITCH_CTRL_MSG_SOCK_PATH);
+ close(sock_fd);
+ return -errno;
+ }
+
+ memset(&un, 0, sizeof(struct sockaddr_un));
+ un.sun_family = AF_UNIX;
+ strncpy(un.sun_path, CNXK_ESWITCH_CTRL_MSG_SOCK_PATH, sizeof(un.sun_path) - 1);
+
+ if (bind(sock_fd, (struct sockaddr *)&un, sizeof(un)) < 0) {
+ plt_err("Failed to bind %s: %s", un.sun_path, strerror(errno));
+ close(sock_fd);
+ return -errno;
+ }
+
+ if (listen(sock_fd, 1) < 0) {
+ plt_err("Failed to listen, err %s", strerror(errno));
+ close(sock_fd);
+ return -errno;
+ }
+
+ plt_rep_dbg("Unix socket path %s", un.sun_path);
+ return sock_fd;
+}
+
+static int
+send_control_message(struct cnxk_eswitch_dev *eswitch_dev, void *buffer, uint32_t len)
+{
+ int sz;
+ int rc = 0;
+
+ sz = send_message_on_socket(eswitch_dev->sock_fd, buffer, len, 0);
+ if (sz < 0) {
+ plt_err("Error sending message, err %d", sz);
+ rc = sz;
+ goto done;
+ }
+
+ /* Ensuring entire message has been processed */
+ if (sz != (int)len) {
+ plt_err("Out of %d bytes only %d bytes sent", sz, len);
+ rc = -EFAULT;
+ goto done;
+ }
+ plt_rep_dbg("Sent %d bytes of buffer", sz);
+done:
+ return rc;
+}
+
+void
+cnxk_rep_msg_populate_msg_end(void *buffer, uint32_t *length)
+{
+ cnxk_rep_msg_populate_command(buffer, length, CNXK_REP_MSG_END, 0);
+}
+
+void
+cnxk_rep_msg_populate_type(void *buffer, uint32_t *length, cnxk_type_t type, uint32_t sz)
+{
+ uint32_t len = *length;
+ cnxk_type_data_t data;
+
+ /* Prepare type data */
+ data.type = type;
+ data.length = sz;
+
+ /* Populate the type data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), &data, sizeof(cnxk_type_data_t));
+ len += sizeof(cnxk_type_data_t);
+
+ *length = len;
+}
+
+void
+cnxk_rep_msg_populate_header(void *buffer, uint32_t *length)
+{
+ cnxk_header_t hdr;
+ int len;
+
+ memset(&hdr, 0, sizeof(cnxk_header_t));
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_HEADER, sizeof(cnxk_header_t));
+
+ len = *length;
+ /* Prepare header data */
+ hdr.signature = CTRL_MSG_SIGNATURE;
+
+ /* Populate header data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), &hdr, sizeof(cnxk_header_t));
+ len += sizeof(cnxk_header_t);
+
+ *length = len;
+}
+
+void
+cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type, uint32_t size)
+{
+ cnxk_rep_msg_data_t msg_data;
+ uint32_t len;
+ uint16_t sz = sizeof(cnxk_rep_msg_data_t);
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_MSG, sz);
+
+ len = *length;
+ /* Prepare command data */
+ msg_data.type = type;
+ msg_data.length = size;
+
+ /* Populate the command */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), &msg_data, sz);
+ len += sz;
+
+ *length = len;
+}
+
+void
+cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
+ cnxk_rep_msg_t msg)
+{
+ uint32_t len;
+
+ cnxk_rep_msg_populate_command(buffer, length, msg, sz);
+
+ len = *length;
+ /* Populate command data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), msg_meta, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static int
+parse_validate_header(void *msg_buf, uint32_t *buf_trav_len)
+{
+ cnxk_type_data_t *tdata = NULL;
+ cnxk_header_t *hdr = NULL;
+ void *data = NULL;
+ uint16_t len = 0;
+
+ /* Read first bytes of type data */
+ data = msg_buf;
+ tdata = (cnxk_type_data_t *)data;
+ if (tdata->type != CNXK_TYPE_HEADER) {
+ plt_err("Invalid type %d, type header expected", tdata->type);
+ goto fail;
+ }
+
+ /* Get the header value */
+ data = RTE_PTR_ADD(msg_buf, sizeof(cnxk_type_data_t));
+ len += sizeof(cnxk_type_data_t);
+
+ /* Validate the header */
+ hdr = (cnxk_header_t *)data;
+ if (hdr->signature != CTRL_MSG_SIGNATURE) {
+ plt_err("Invalid signature %" PRIu64 " detected", hdr->signature);
+ goto fail;
+ }
+
+ /* Update length read till point */
+ len += tdata->length;
+
+ *buf_trav_len = len;
+ return 0;
+fail:
+ return errno;
+}
+
+static cnxk_rep_msg_data_t *
+message_data_extract(void *msg_buf, uint32_t *buf_trav_len)
+{
+ cnxk_type_data_t *tdata = NULL;
+ cnxk_rep_msg_data_t *msg = NULL;
+ uint16_t len = *buf_trav_len;
+ void *data;
+
+ tdata = (cnxk_type_data_t *)RTE_PTR_ADD(msg_buf, len);
+ if (tdata->type != CNXK_TYPE_MSG) {
+ plt_err("Invalid type %d, type MSG expected", tdata->type);
+ goto fail;
+ }
+
+ /* Get the message type */
+ len += sizeof(cnxk_type_data_t);
+ data = RTE_PTR_ADD(msg_buf, len);
+ msg = (cnxk_rep_msg_data_t *)data;
+
+ /* Advance to actual message data */
+ len += tdata->length;
+ *buf_trav_len = len;
+
+ return msg;
+fail:
+ return NULL;
+}
+
+static void
+process_ack_message(void *msg_buf, uint32_t *buf_trav_len, uint32_t msg_len, void *data)
+{
+ cnxk_rep_msg_ack_data_t *adata = (cnxk_rep_msg_ack_data_t *)data;
+ uint16_t len = *buf_trav_len;
+ void *buf;
+
+ /* Get the message type data viz ack data */
+ buf = RTE_PTR_ADD(msg_buf, len);
+ adata->u.data = rte_zmalloc("Ack data", msg_len, 0);
+ adata->size = msg_len;
+ if (adata->size == sizeof(uint64_t))
+ rte_memcpy(&adata->u.data, buf, msg_len);
+ else
+ rte_memcpy(adata->u.data, buf, msg_len);
+ plt_rep_dbg("Address %p val 0x%" PRIu64 " sval %" PRId64 " msg_len %d",
+ adata->u.data, adata->u.val, adata->u.sval, msg_len);
+
+ /* Advance length to nex message */
+ len += msg_len;
+ *buf_trav_len = len;
+}
+
+static int
+notify_rep_dev_ready(cnxk_rep_msg_ready_data_t *rdata, void *data,
+ cnxk_rep_msg_ack_data1_t **padata)
+{
+ struct cnxk_eswitch_dev *eswitch_dev;
+ uint64_t rep_id_arr[RTE_MAX_ETHPORTS];
+ cnxk_rep_msg_ack_data1_t *adata;
+ uint16_t rep_id, sz, total_sz;
+ int rc, i, j = 0;
+
+ PLT_SET_USED(data);
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ if (!eswitch_dev) {
+ plt_err("Failed to get PF ethdev handle");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* For ready state */
+ if ((rdata->nb_ports / 2) > eswitch_dev->repr_cnt.nb_repr_probed) {
+ rc = CNXK_REP_CTRL_MSG_NACK_INV_REP_CNT;
+ goto fail;
+ }
+
+ for (i = 0; i < rdata->nb_ports / 2; i++) {
+ rep_id = UINT16_MAX;
+ rc = cnxk_rep_state_update(eswitch_dev, rdata->data[i], &rep_id);
+ if (rc) {
+ rc = CNXK_REP_CTRL_MSG_NACK_REP_STAT_UP_FAIL;
+ goto fail;
+ }
+ if (rep_id != UINT16_MAX)
+ rep_id_arr[j++] = rep_id;
+ }
+
+ /* Send Rep Id array to companian app */
+ sz = j * sizeof(uint64_t);
+ total_sz = sizeof(cnxk_rep_msg_ack_data1_t) + sz;
+ adata = plt_zmalloc(total_sz, 0);
+ rte_memcpy(adata->data, rep_id_arr, sz);
+ adata->size = sz;
+ *padata = adata;
+
+ plt_rep_dbg("Installing NPC rules for Eswitch VF");
+ /* Install RX VLAN rule for eswitch VF */
+ if (!eswitch_dev->eswitch_vf_rules_setup) {
+ rc = cnxk_eswitch_pfvf_flow_rules_install(eswitch_dev, true);
+ if (rc) {
+ plt_err("Failed to install rxtx rules, rc %d", rc);
+ goto fail;
+ }
+
+ /* Configure TPID for Eswitch PF LFs */
+ rc = roc_eswitch_nix_vlan_tpid_set(&eswitch_dev->nix, ROC_NIX_VLAN_TYPE_OUTER,
+ CNXK_ESWITCH_VLAN_TPID, true);
+ if (rc) {
+ plt_err("Failed to configure tpid, rc %d", rc);
+ goto fail;
+ }
+ eswitch_dev->eswitch_vf_rules_setup = true;
+ }
+
+ return 0;
+fail:
+ sz = sizeof(cnxk_rep_msg_ack_data1_t) + sizeof(uint64_t);
+ adata = plt_zmalloc(sz, 0);
+ adata->data[0] = rc;
+ adata->size = sizeof(uint64_t);
+ *padata = adata;
+
+ return rc;
+}
+
+static int
+process_ready_message(void *msg_buf, uint32_t *buf_trav_len, uint32_t msg_len, void *data,
+ cnxk_rep_msg_ack_data1_t **padata)
+{
+ cnxk_rep_msg_ready_data_t *rdata = NULL;
+ cnxk_rep_msg_ack_data1_t *adata;
+ uint16_t len = *buf_trav_len;
+ void *buf;
+ int rc = 0, sz;
+
+ /* Get the message type data viz ready data */
+ buf = RTE_PTR_ADD(msg_buf, len);
+ rdata = (cnxk_rep_msg_ready_data_t *)buf;
+
+ plt_rep_dbg("Ready data received %d, nb_ports %d", rdata->val, rdata->nb_ports);
+
+ /* Wait required to ensure other side ready for receiving the ack */
+ usleep(CTRL_MSG_READY_WAIT_US);
+
+ /* Update all representor about ready message */
+ if (rdata->val) {
+ rc = notify_rep_dev_ready(rdata, data, padata);
+ } else {
+ sz = sizeof(cnxk_rep_msg_ack_data1_t) + sizeof(uint64_t);
+ adata = plt_zmalloc(sz, 0);
+ adata->data[0] = CNXK_REP_CTRL_MSG_NACK_INV_RDY_DATA;
+ adata->size = sizeof(uint64_t);
+ *padata = adata;
+ }
+
+ /* Advance length to nex message */
+ len += msg_len;
+ *buf_trav_len = len;
+
+ return rc;
+}
+
+static int
+notify_rep_dev_exit(cnxk_rep_msg_exit_data_t *edata, void *data)
+{
+ struct cnxk_eswitch_dev *eswitch_dev;
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct rte_eth_dev *rep_eth_dev;
+ int i, rc = 0;
+
+ PLT_SET_USED(data);
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ if (!eswitch_dev) {
+ plt_err("Failed to get PF ethdev handle");
+ rc = -EINVAL;
+ goto fail;
+ }
+ if ((edata->nb_ports / 2) > eswitch_dev->repr_cnt.nb_repr_probed) {
+ rc = CNXK_REP_CTRL_MSG_NACK_INV_REP_CNT;
+ goto fail;
+ }
+
+ for (i = 0; i < eswitch_dev->repr_cnt.nb_repr_probed; i++) {
+ rep_eth_dev = eswitch_dev->rep_info[i].rep_eth_dev;
+ if (!rep_eth_dev) {
+ plt_err("Failed to get rep ethdev handle");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ if (!rep_dev->native_repte)
+ rep_dev->is_vf_active = false;
+ }
+ /* For Exit message */
+ eswitch_dev->client_connected = false;
+ return 0;
+fail:
+ return rc;
+}
+
+static void
+process_exit_message(void *msg_buf, uint32_t *buf_trav_len, uint32_t msg_len, void *data)
+{
+ cnxk_rep_msg_exit_data_t *edata = NULL;
+ uint16_t len = *buf_trav_len;
+ void *buf;
+
+ /* Get the message type data viz exit data */
+ buf = RTE_PTR_ADD(msg_buf, len);
+ edata = (cnxk_rep_msg_exit_data_t *)buf;
+
+ plt_rep_dbg("Exit data received %d", edata->val);
+
+ /* Update all representor about ready/exit message */
+ if (edata->val)
+ notify_rep_dev_exit(edata, data);
+
+ /* Advance length to nex message */
+ len += msg_len;
+ *buf_trav_len = len;
+}
+
+static void
+populate_ack_msg(void *buffer, uint32_t *length, cnxk_rep_msg_ack_data1_t *adata)
+{
+ uint32_t sz = sizeof(cnxk_rep_msg_ack_data1_t) + adata->size;
+ uint32_t len;
+
+ cnxk_rep_msg_populate_command(buffer, length, CNXK_REP_MSG_ACK, sz);
+
+ len = *length;
+
+ /* Populate ACK message data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), adata, sz);
+
+ len += sz;
+
+ *length = len;
+}
+
+static int
+send_ack_message(void *data, cnxk_rep_msg_ack_data1_t *adata)
+{
+ struct cnxk_eswitch_dev *eswitch_dev = (struct cnxk_eswitch_dev *)data;
+ uint32_t len = 0, size;
+ void *buffer;
+ int rc = 0;
+
+ /* Allocate memory for preparing a message */
+ size = CTRL_MSG_BUFFER_SZ;
+ buffer = rte_zmalloc("ACK msg", size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ return -ENOMEM;
+ }
+
+ /* Prepare the ACK message */
+ cnxk_rep_msg_populate_header(buffer, &len);
+ populate_ack_msg(buffer, &len, adata);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ /* Length check to avoid buffer overflow */
+ if (len > CTRL_MSG_BUFFER_SZ) {
+ plt_err("Invalid length %d for max sized buffer %d", len, CTRL_MSG_BUFFER_SZ);
+ rc = -EFAULT;
+ goto done;
+ }
+
+ /* Send it to the peer */
+ rc = send_control_message(eswitch_dev, buffer, len);
+ if (rc)
+ plt_err("Failed send ack");
+
+done:
+ return rc;
+}
+
+static int
+process_message(void *msg_buf, uint32_t *buf_trav_len, void *data)
+{
+ cnxk_rep_msg_data_t *msg = NULL;
+ cnxk_rep_msg_ack_data1_t *adata = NULL;
+ bool send_ack;
+ int rc = 0, sz;
+
+ /* Get the message data */
+ msg = message_data_extract(msg_buf, buf_trav_len);
+ if (!msg) {
+ plt_err("Failed to get message data");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Different message type processing */
+ while (msg->type != CNXK_REP_MSG_END) {
+ send_ack = true;
+ switch (msg->type) {
+ case CNXK_REP_MSG_ACK:
+ plt_rep_dbg("Received ack response");
+ process_ack_message(msg_buf, buf_trav_len, msg->length, data);
+ send_ack = false;
+ break;
+ case CNXK_REP_MSG_READY:
+ plt_rep_dbg("Received ready message");
+ process_ready_message(msg_buf, buf_trav_len, msg->length, data, &adata);
+ adata->type = CNXK_REP_MSG_READY;
+ break;
+ case CNXK_REP_MSG_EXIT:
+ plt_rep_dbg("Received exit message");
+ process_exit_message(msg_buf, buf_trav_len, msg->length, data);
+ sz = sizeof(cnxk_rep_msg_ack_data1_t) + sizeof(uint64_t);
+ adata = plt_zmalloc(sz, 0);
+ adata->type = CNXK_REP_MSG_EXIT;
+ adata->data[0] = 0;
+ adata->size = sizeof(uint64_t);
+ break;
+ default:
+ send_ack = false;
+ plt_err("Invalid message type: %d", msg->type);
+ rc = -EINVAL;
+ };
+
+ /* Send ACK */
+ if (send_ack)
+ send_ack_message(data, adata);
+
+ /* Advance to next message */
+ msg = message_data_extract(msg_buf, buf_trav_len);
+ if (!msg) {
+ plt_err("Failed to get message data");
+ rc = -EINVAL;
+ goto fail;
+ }
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+process_control_message(void *msg_buf, void *data, size_t sz)
+{
+ uint32_t buf_trav_len = 0;
+ int rc;
+
+ /* Validate the validity of the received message */
+ parse_validate_header(msg_buf, &buf_trav_len);
+
+ /* Detect message and process */
+ rc = process_message(msg_buf, &buf_trav_len, data);
+ if (rc) {
+ plt_err("Failed to process message");
+ goto fail;
+ }
+
+ /* Ensuring entire message has been processed */
+ if (sz != buf_trav_len) {
+ plt_err("Out of %" PRId64 " bytes %d bytes of msg_buf processed", sz, buf_trav_len);
+ rc = -EFAULT;
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+receive_control_msg_resp(struct cnxk_eswitch_dev *eswitch_dev, void *data)
+{
+ uint32_t wait_us = CTRL_MSG_RCV_TIMEOUT_MS * 1000;
+ uint32_t timeout = 0, sleep = 1;
+ int sz = 0;
+ int rc = -1;
+ uint32_t len = BUFSIZ;
+ void *msg_buf;
+
+ msg_buf = plt_zmalloc(len, 0);
+
+ do {
+ sz = receive_control_message(eswitch_dev->sock_fd, msg_buf, len);
+ if (sz != 0)
+ break;
+
+ /* Timeout after CTRL_MSG_RCV_TIMEOUT_MS */
+ if (timeout >= wait_us) {
+ plt_err("Control message wait timedout");
+ return -ETIMEDOUT;
+ }
+
+ plt_delay_us(sleep);
+ timeout += sleep;
+ } while ((sz == 0) || (timeout < wait_us));
+
+ if (sz > 0) {
+ plt_rep_dbg("Received %d sized response packet", sz);
+ rc = process_control_message(msg_buf, data, sz);
+ plt_free(msg_buf);
+ }
+
+ return rc;
+}
+
+int
+cnxk_rep_msg_send_process(struct cnxk_rep_dev *rep_dev, void *buffer, uint32_t len,
+ cnxk_rep_msg_ack_data_t *adata)
+{
+ struct cnxk_eswitch_dev *eswitch_dev;
+ int rc = 0;
+
+ eswitch_dev = rep_dev->parent_dev;
+ if (!eswitch_dev) {
+ plt_err("Failed to get parent eswitch handle");
+ rc = -1;
+ goto fail;
+ }
+
+ plt_spinlock_lock(&eswitch_dev->rep_lock);
+ rc = send_control_message(eswitch_dev, buffer, len);
+ if (rc) {
+ plt_err("Failed to send the message, err %d", rc);
+ goto free;
+ }
+
+ /* Get response of the command sent */
+ rc = receive_control_msg_resp(eswitch_dev, adata);
+ if (rc) {
+ plt_err("Failed to receive the response, err %d", rc);
+ goto free;
+ }
+ plt_spinlock_unlock(&eswitch_dev->rep_lock);
+
+ return 0;
+free:
+ plt_spinlock_unlock(&eswitch_dev->rep_lock);
+fail:
+ return rc;
+}
+
+static void
+poll_for_control_msg(void *data)
+{
+ struct cnxk_eswitch_dev *eswitch_dev = (struct cnxk_eswitch_dev *)data;
+ uint32_t len = BUFSIZ;
+ int sz = 0;
+ void *msg_buf;
+
+ while (eswitch_dev->client_connected) {
+ msg_buf = plt_zmalloc(len, 0);
+ do {
+ plt_spinlock_lock(&eswitch_dev->rep_lock);
+ sz = receive_control_message(eswitch_dev->sock_fd, msg_buf, len);
+ plt_spinlock_unlock(&eswitch_dev->rep_lock);
+ if (sz != 0)
+ break;
+ plt_delay_us(2000);
+ } while (sz == 0);
+
+ if (sz > 0) {
+ plt_rep_dbg("Received new %d bytes control message", sz);
+ plt_spinlock_lock(&eswitch_dev->rep_lock);
+ process_control_message(msg_buf, data, sz);
+ plt_spinlock_unlock(&eswitch_dev->rep_lock);
+ plt_free(msg_buf);
+ }
+ }
+ plt_rep_dbg("Exiting poll for control message loop");
+}
+
+static uint32_t
+rep_ctrl_msg_thread_main(void *arg)
+{
+ struct cnxk_eswitch_dev *eswitch_dev = (struct cnxk_eswitch_dev *)arg;
+ struct sockaddr_un client;
+ int addr_len;
+ int ssock_fd;
+ int sock_fd;
+
+ ssock_fd = open_socket_ctrl_channel();
+ if (ssock_fd < 0) {
+ plt_err("Failed to open socket for ctrl channel, err %d", ssock_fd);
+ return UINT32_MAX;
+ }
+
+ addr_len = sizeof(client);
+ while (eswitch_dev->start_ctrl_msg_thrd) {
+ /* Accept client connection until the thread is running */
+ sock_fd = accept(ssock_fd, (struct sockaddr *)&client, (socklen_t *)&addr_len);
+ if (sock_fd < 0) {
+ plt_err("Failed to accept connection request on socket fd %d", ssock_fd);
+ break;
+ }
+
+ plt_rep_dbg("Client %s: Connection request accepted.", client.sun_path);
+ eswitch_dev->sock_fd = sock_fd;
+ if (eswitch_dev->start_ctrl_msg_thrd) {
+ eswitch_dev->client_connected = true;
+ poll_for_control_msg(eswitch_dev);
+ }
+ eswitch_dev->sock_fd = -1;
+ }
+
+ /* Closing the opened socket */
+ close_socket(ssock_fd);
+ plt_rep_dbg("Exiting representor ctrl thread");
+
+ return 0;
+}
+
+int
+cnxk_rep_msg_control_thread_launch(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ char name[CTRL_MSG_THRD_NAME_LEN];
+ int rc = 0;
+
+ rte_strscpy(name, "rep_ctrl_msg_hndlr", CTRL_MSG_THRD_NAME_LEN);
+ eswitch_dev->start_ctrl_msg_thrd = true;
+ rc = rte_thread_create_internal_control(&eswitch_dev->rep_ctrl_msg_thread, name,
+ rep_ctrl_msg_thread_main, eswitch_dev);
+ if (rc)
+ plt_err("Failed to create rep control message handling");
+
+ return rc;
+}
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
new file mode 100644
index 0000000000..fb84d58848
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -0,0 +1,95 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#ifndef __CNXK_REP_MSG_H__
+#define __CNXK_REP_MSG_H__
+
+#include <stdint.h>
+
+#define CNXK_REP_MSG_MAX_BUFFER_SZ 1500
+
+typedef enum CNXK_TYPE {
+ CNXK_TYPE_HEADER = 0,
+ CNXK_TYPE_MSG,
+} cnxk_type_t;
+
+typedef enum CNXK_REP_MSG {
+ /* General sync messages */
+ CNXK_REP_MSG_READY = 0,
+ CNXK_REP_MSG_ACK,
+ CNXK_REP_MSG_EXIT,
+ /* End of messaging sequence */
+ CNXK_REP_MSG_END,
+} cnxk_rep_msg_t;
+
+typedef enum CNXK_NACK_CODE {
+ CNXK_REP_CTRL_MSG_NACK_INV_RDY_DATA = 0x501,
+ CNXK_REP_CTRL_MSG_NACK_INV_REP_CNT = 0x502,
+ CNXK_REP_CTRL_MSG_NACK_REP_STAT_UP_FAIL = 0x503,
+} cnxk_nack_code_t;
+
+/* Types */
+typedef struct cnxk_type_data {
+ cnxk_type_t type;
+ uint32_t length;
+ uint64_t data[];
+} __rte_packed cnxk_type_data_t;
+
+/* Header */
+typedef struct cnxk_header {
+ uint64_t signature;
+ uint16_t nb_hops;
+} __rte_packed cnxk_header_t;
+
+/* Message meta */
+typedef struct cnxk_rep_msg_data {
+ cnxk_rep_msg_t type;
+ uint32_t length;
+ uint64_t data[];
+} __rte_packed cnxk_rep_msg_data_t;
+
+/* Ack msg */
+typedef struct cnxk_rep_msg_ack_data {
+ cnxk_rep_msg_t type;
+ uint32_t size;
+ union {
+ void *data;
+ uint64_t val;
+ int64_t sval;
+ } u;
+} __rte_packed cnxk_rep_msg_ack_data_t;
+
+/* Ack msg */
+typedef struct cnxk_rep_msg_ack_data1 {
+ cnxk_rep_msg_t type;
+ uint32_t size;
+ uint64_t data[];
+} __rte_packed cnxk_rep_msg_ack_data1_t;
+
+/* Ready msg */
+typedef struct cnxk_rep_msg_ready_data {
+ uint8_t val;
+ uint16_t nb_ports;
+ uint16_t data[];
+} __rte_packed cnxk_rep_msg_ready_data_t;
+
+/* Exit msg */
+typedef struct cnxk_rep_msg_exit_data {
+ uint8_t val;
+ uint16_t nb_ports;
+ uint16_t data[];
+} __rte_packed cnxk_rep_msg_exit_data_t;
+
+void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
+ uint32_t size);
+void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
+ cnxk_rep_msg_t msg);
+void cnxk_rep_msg_populate_msg_end(void *buffer, uint32_t *length);
+void cnxk_rep_msg_populate_type(void *buffer, uint32_t *length, cnxk_type_t type, uint32_t sz);
+void cnxk_rep_msg_populate_header(void *buffer, uint32_t *length);
+int cnxk_rep_msg_send_process(struct cnxk_rep_dev *rep_dev, void *buffer, uint32_t len,
+ cnxk_rep_msg_ack_data_t *adata);
+int cnxk_rep_msg_control_thread_launch(struct cnxk_eswitch_dev *eswitch_dev);
+
+#endif /* __CNXK_REP_MSG_H__ */
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index 7121845dc6..9ca7732713 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -37,6 +37,7 @@ sources = files(
'cnxk_ptp.c',
'cnxk_flow.c',
'cnxk_rep.c',
+ 'cnxk_rep_msg.c',
'cnxk_rep_ops.c',
'cnxk_stats.c',
'cnxk_tm.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 11/24] common/cnxk: representee notification callback
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (9 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 10/24] net/cnxk: add representor control plane Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 12/24] net/cnxk: handling representee notification Harman Kalra
` (12 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Setting up a callback which gets invoked every time a representee
comes up or goes down. Later this callback gets handled by network
conterpart.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_dev.c | 24 ++++++++++++++++++++++++
drivers/common/cnxk/roc_dev_priv.h | 3 +++
drivers/common/cnxk/roc_eswitch.c | 23 +++++++++++++++++++++++
drivers/common/cnxk/roc_eswitch.h | 6 ++++++
drivers/common/cnxk/roc_mbox.c | 2 ++
drivers/common/cnxk/roc_mbox.h | 10 +++++++++-
drivers/common/cnxk/version.map | 2 ++
7 files changed, 69 insertions(+), 1 deletion(-)
diff --git a/drivers/common/cnxk/roc_dev.c b/drivers/common/cnxk/roc_dev.c
index e7e89bf3d6..b12732de34 100644
--- a/drivers/common/cnxk/roc_dev.c
+++ b/drivers/common/cnxk/roc_dev.c
@@ -538,6 +538,29 @@ pf_vf_mbox_send_up_msg(struct dev *dev, void *rec_msg)
}
}
+static int
+mbox_up_handler_esw_repte_notify(struct dev *dev, struct esw_repte_req *req, struct msg_rsp *rsp)
+{
+ int rc = 0;
+
+ plt_base_dbg("pf:%d/vf:%d msg id 0x%x (%s) from: pf:%d/vf:%d", dev_get_pf(dev->pf_func),
+ dev_get_vf(dev->pf_func), req->hdr.id, mbox_id2name(req->hdr.id),
+ dev_get_pf(req->hdr.pcifunc), dev_get_vf(req->hdr.pcifunc));
+
+ plt_base_dbg("repte pcifunc %x, enable %d", req->repte_pcifunc, req->enable);
+
+ if (dev->ops && dev->ops->repte_notify) {
+ rc = dev->ops->repte_notify(dev->roc_nix, req->repte_pcifunc,
+ req->enable);
+ if (rc < 0)
+ plt_err("Failed to sent new representee %x notification to %s",
+ req->repte_pcifunc, (req->enable == true) ? "enable" : "disable");
+ }
+
+ rsp->hdr.rc = rc;
+ return rc;
+}
+
static int
mbox_up_handler_mcs_intr_notify(struct dev *dev, struct mcs_intr_info *info, struct msg_rsp *rsp)
{
@@ -712,6 +735,7 @@ mbox_process_msgs_up(struct dev *dev, struct mbox_msghdr *req)
}
MBOX_UP_CGX_MESSAGES
MBOX_UP_MCS_MESSAGES
+ MBOX_UP_ESW_MESSAGES
#undef M
}
diff --git a/drivers/common/cnxk/roc_dev_priv.h b/drivers/common/cnxk/roc_dev_priv.h
index 5b2c5096f8..dd694b8572 100644
--- a/drivers/common/cnxk/roc_dev_priv.h
+++ b/drivers/common/cnxk/roc_dev_priv.h
@@ -36,12 +36,15 @@ typedef void (*q_err_cb_t)(void *roc_nix, void *data);
/* Link status get callback */
typedef void (*link_status_get_t)(void *roc_nix,
struct cgx_link_user_info *link);
+/* Representee notification callback */
+typedef int (*repte_notify_t)(void *roc_nix, uint16_t pf_func, bool enable);
struct dev_ops {
link_info_t link_status_update;
ptp_info_t ptp_info_update;
link_status_get_t link_status_get;
q_err_cb_t q_err_cb;
+ repte_notify_t repte_notify;
};
#define dev_is_vf(dev) ((dev)->hwcap & DEV_HWCAP_F_VF)
diff --git a/drivers/common/cnxk/roc_eswitch.c b/drivers/common/cnxk/roc_eswitch.c
index 7f2a8e6c06..31bdba3985 100644
--- a/drivers/common/cnxk/roc_eswitch.c
+++ b/drivers/common/cnxk/roc_eswitch.c
@@ -298,3 +298,26 @@ roc_eswitch_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t t
return rc;
}
+
+int
+roc_eswitch_nix_process_repte_notify_cb_register(struct roc_nix *roc_nix,
+ process_repte_notify_t proc_repte_nt)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+
+ if (proc_repte_nt == NULL)
+ return NIX_ERR_PARAM;
+
+ dev->ops->repte_notify = (repte_notify_t)proc_repte_nt;
+ return 0;
+}
+
+void
+roc_eswitch_nix_process_repte_notify_cb_unregister(struct roc_nix *roc_nix)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+
+ dev->ops->repte_notify = NULL;
+}
diff --git a/drivers/common/cnxk/roc_eswitch.h b/drivers/common/cnxk/roc_eswitch.h
index 0dd23ff76a..8837e19b22 100644
--- a/drivers/common/cnxk/roc_eswitch.h
+++ b/drivers/common/cnxk/roc_eswitch.h
@@ -8,6 +8,9 @@
#define ROC_ESWITCH_VLAN_TPID 0x8100
#define ROC_ESWITCH_LBK_CHAN 63
+/* Process representee notification callback */
+typedef int (*process_repte_notify_t)(void *roc_nix, uint16_t pf_func, bool enable);
+
/* NPC */
int __roc_api roc_eswitch_npc_mcam_rx_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow,
uint16_t pcifunc, uint16_t vlan_tci,
@@ -22,4 +25,7 @@ int __roc_api roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc,
/* NIX */
int __roc_api roc_eswitch_nix_vlan_tpid_set(struct roc_nix *nix, uint32_t type, uint16_t tpid,
bool is_vf);
+int __roc_api roc_eswitch_nix_process_repte_notify_cb_register(struct roc_nix *roc_nix,
+ process_repte_notify_t proc_repte_nt);
+void __roc_api roc_eswitch_nix_process_repte_notify_cb_unregister(struct roc_nix *roc_nix);
#endif /* __ROC_ESWITCH_H__ */
diff --git a/drivers/common/cnxk/roc_mbox.c b/drivers/common/cnxk/roc_mbox.c
index 7b734fcd24..cb486b2505 100644
--- a/drivers/common/cnxk/roc_mbox.c
+++ b/drivers/common/cnxk/roc_mbox.c
@@ -499,6 +499,7 @@ mbox_id2name(uint16_t id)
return #_name;
MBOX_MESSAGES
MBOX_UP_CGX_MESSAGES
+ MBOX_UP_ESW_MESSAGES
#undef M
}
}
@@ -514,6 +515,7 @@ mbox_id2size(uint16_t id)
return sizeof(struct _req_type);
MBOX_MESSAGES
MBOX_UP_CGX_MESSAGES
+ MBOX_UP_ESW_MESSAGES
#undef M
}
}
diff --git a/drivers/common/cnxk/roc_mbox.h b/drivers/common/cnxk/roc_mbox.h
index 4c846f0757..2bedf1fb81 100644
--- a/drivers/common/cnxk/roc_mbox.h
+++ b/drivers/common/cnxk/roc_mbox.h
@@ -355,9 +355,11 @@ struct mbox_msghdr {
#define MBOX_UP_MCS_MESSAGES M(MCS_INTR_NOTIFY, 0xE00, mcs_intr_notify, mcs_intr_info, msg_rsp)
+#define MBOX_UP_ESW_MESSAGES M(ESW_REPTE_NOTIFY, 0xF00, esw_repte_notify, esw_repte_req, msg_rsp)
+
enum {
#define M(_name, _id, _1, _2, _3) MBOX_MSG_##_name = _id,
- MBOX_MESSAGES MBOX_UP_CGX_MESSAGES MBOX_UP_MCS_MESSAGES
+ MBOX_MESSAGES MBOX_UP_CGX_MESSAGES MBOX_UP_MCS_MESSAGES MBOX_UP_ESW_MESSAGES
#undef M
};
@@ -2778,4 +2780,10 @@ struct nix_spi_to_sa_delete_req {
uint16_t __io hash_index;
uint8_t __io way;
};
+
+struct esw_repte_req {
+ struct mbox_msghdr hdr;
+ uint16_t __io repte_pcifunc;
+ bool __io enable;
+};
#endif /* __ROC_MBOX_H__ */
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index 78c421677d..e170a6a63a 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -91,6 +91,8 @@ INTERNAL {
roc_dpi_disable;
roc_dpi_enable;
roc_error_msg_get;
+ roc_eswitch_nix_process_repte_notify_cb_register;
+ roc_eswitch_nix_process_repte_notify_cb_unregister;
roc_eswitch_nix_vlan_tpid_set;
roc_eswitch_npc_mcam_delete_rule;
roc_eswitch_npc_mcam_rx_rule;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 12/24] net/cnxk: handling representee notification
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (10 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 11/24] common/cnxk: representee notification callback Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 13/24] net/cnxk: representor ethdev ops Harman Kalra
` (11 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
In case of any representee coming up or going down, kernel sends a
mbox up call which signals a thread to process these messages and
enable/disable HW resources accordingly.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 8 +
drivers/net/cnxk/cnxk_eswitch.h | 20 +++
drivers/net/cnxk/cnxk_rep.c | 263 ++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep.h | 36 +++++
4 files changed, 327 insertions(+)
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index ffcf89b1b1..35c517f124 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -113,6 +113,14 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
rte_thread_join(eswitch_dev->rep_ctrl_msg_thread, NULL);
}
+ if (eswitch_dev->repte_msg_proc.start_thread) {
+ eswitch_dev->repte_msg_proc.start_thread = false;
+ pthread_cond_signal(&eswitch_dev->repte_msg_proc.repte_msg_cond);
+ rte_thread_join(eswitch_dev->repte_msg_proc.repte_msg_thread, NULL);
+ pthread_mutex_destroy(&eswitch_dev->repte_msg_proc.mutex);
+ pthread_cond_destroy(&eswitch_dev->repte_msg_proc.repte_msg_cond);
+ }
+
/* Remove representor devices associated with PF */
cnxk_rep_dev_remove(eswitch_dev);
}
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index a2f4aa0fcc..8aab3e8a72 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -30,6 +30,23 @@ enum cnxk_esw_da_pattern_type {
CNXK_ESW_DA_TYPE_PFVF,
};
+struct cnxk_esw_repte_msg {
+ uint16_t hw_func;
+ bool enable;
+
+ TAILQ_ENTRY(cnxk_esw_repte_msg) next;
+};
+
+struct cnxk_esw_repte_msg_proc {
+ bool start_thread;
+ uint8_t msg_avail;
+ rte_thread_t repte_msg_thread;
+ pthread_cond_t repte_msg_cond;
+ pthread_mutex_t mutex;
+
+ TAILQ_HEAD(esw_repte_msg_list, cnxk_esw_repte_msg) msg_list;
+};
+
struct cnxk_esw_repr_hw_info {
/* Representee pcifunc value */
uint16_t hw_func;
@@ -139,6 +156,9 @@ struct cnxk_eswitch_dev {
bool client_connected;
int sock_fd;
+ /* Representee notification */
+ struct cnxk_esw_repte_msg_proc repte_msg_proc;
+
/* Port representor fields */
rte_spinlock_t rep_lock;
uint16_t nb_switch_domain;
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
index f8e1d5b965..3b01856bc8 100644
--- a/drivers/net/cnxk/cnxk_rep.c
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -4,6 +4,8 @@
#include <cnxk_rep.h>
#include <cnxk_rep_msg.h>
+#define REPTE_MSG_PROC_THRD_NAME_MAX_LEN 30
+
#define PF_SHIFT 10
#define PF_MASK 0x3F
@@ -86,6 +88,7 @@ cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev)
{
int i, rc = 0;
+ roc_eswitch_nix_process_repte_notify_cb_unregister(&eswitch_dev->nix);
for (i = 0; i < eswitch_dev->nb_switch_domain; i++) {
rc = rte_eth_switch_domain_free(eswitch_dev->sw_dom[i].switch_domain_id);
if (rc)
@@ -95,6 +98,236 @@ cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev)
return rc;
}
+static int
+cnxk_representee_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func)
+{
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct rte_eth_dev *rep_eth_dev;
+ int i, rc = 0;
+
+ for (i = 0; i < eswitch_dev->repr_cnt.nb_repr_probed; i++) {
+ rep_eth_dev = eswitch_dev->rep_info[i].rep_eth_dev;
+ if (!rep_eth_dev) {
+ plt_err("Failed to get rep ethdev handle");
+ rc = -EINVAL;
+ goto done;
+ }
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ if (rep_dev->hw_func == hw_func &&
+ (!rep_dev->native_repte || rep_dev->is_vf_active)) {
+ rep_dev->is_vf_active = false;
+ rc = cnxk_rep_dev_stop(rep_eth_dev);
+ if (rc) {
+ plt_err("Failed to stop repr port %d, rep id %d", rep_dev->port_id,
+ rep_dev->rep_id);
+ goto done;
+ }
+
+ cnxk_rep_rx_queue_release(rep_eth_dev, 0);
+ cnxk_rep_tx_queue_release(rep_eth_dev, 0);
+ plt_rep_dbg("Released representor ID %d representing %x", rep_dev->rep_id,
+ hw_func);
+ break;
+ }
+ }
+done:
+ return rc;
+}
+
+static int
+cnxk_representee_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, uint16_t rep_id)
+{
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct rte_eth_dev *rep_eth_dev;
+ int i, rc = 0;
+
+ for (i = 0; i < eswitch_dev->repr_cnt.nb_repr_probed; i++) {
+ rep_eth_dev = eswitch_dev->rep_info[i].rep_eth_dev;
+ if (!rep_eth_dev) {
+ plt_err("Failed to get rep ethdev handle");
+ rc = -EINVAL;
+ goto done;
+ }
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ if (rep_dev->hw_func == hw_func && !rep_dev->is_vf_active) {
+ rep_dev->is_vf_active = true;
+ rep_dev->native_repte = true;
+ if (rep_dev->rep_id != rep_id) {
+ plt_err("Rep ID assigned during init %d does not match %d",
+ rep_dev->rep_id, rep_id);
+ rc = -EINVAL;
+ goto done;
+ }
+
+ rc = cnxk_rep_rx_queue_setup(rep_eth_dev, rep_dev->rxq->qid,
+ rep_dev->rxq->nb_desc, 0,
+ rep_dev->rxq->rx_conf, rep_dev->rxq->mpool);
+ if (rc) {
+ plt_err("Failed to setup rxq repr port %d, rep id %d",
+ rep_dev->port_id, rep_dev->rep_id);
+ goto done;
+ }
+
+ rc = cnxk_rep_tx_queue_setup(rep_eth_dev, rep_dev->txq->qid,
+ rep_dev->txq->nb_desc, 0,
+ rep_dev->txq->tx_conf);
+ if (rc) {
+ plt_err("Failed to setup txq repr port %d, rep id %d",
+ rep_dev->port_id, rep_dev->rep_id);
+ goto done;
+ }
+
+ rc = cnxk_rep_dev_start(rep_eth_dev);
+ if (rc) {
+ plt_err("Failed to start repr port %d, rep id %d", rep_dev->port_id,
+ rep_dev->rep_id);
+ goto done;
+ }
+ break;
+ }
+ }
+done:
+ return rc;
+}
+
+static int
+cnxk_representee_msg_process(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, bool enable)
+{
+ struct cnxk_eswitch_devargs *esw_da;
+ uint16_t rep_id = UINT16_MAX;
+ int rc = 0, i, j;
+
+ /* Traversing the initialized represented list */
+ for (i = 0; i < eswitch_dev->nb_esw_da; i++) {
+ esw_da = &eswitch_dev->esw_da[i];
+ for (j = 0; j < esw_da->nb_repr_ports; j++) {
+ if (esw_da->repr_hw_info[j].hw_func == hw_func) {
+ rep_id = esw_da->repr_hw_info[j].rep_id;
+ break;
+ }
+ }
+ if (rep_id != UINT16_MAX)
+ break;
+ }
+ /* No action on PF func for which representor has not been created */
+ if (rep_id == UINT16_MAX)
+ goto done;
+
+ if (enable) {
+ rc = cnxk_representee_setup(eswitch_dev, hw_func, rep_id);
+ if (rc) {
+ plt_err("Failed to setup representee, err %d", rc);
+ goto fail;
+ }
+ plt_rep_dbg(" Representor ID %d representing %x", rep_id, hw_func);
+ rc = cnxk_eswitch_flow_rules_install(eswitch_dev, hw_func);
+ if (rc) {
+ plt_err("Failed to install rxtx flow rules for %x", hw_func);
+ goto fail;
+ }
+ } else {
+ rc = cnxk_eswitch_flow_rules_delete(eswitch_dev, hw_func);
+ if (rc) {
+ plt_err("Failed to delete flow rules for %x", hw_func);
+ goto fail;
+ }
+ rc = cnxk_representee_release(eswitch_dev, hw_func);
+ if (rc) {
+ plt_err("Failed to release representee, err %d", rc);
+ goto fail;
+ }
+ }
+
+done:
+ return 0;
+fail:
+ return rc;
+}
+
+static uint32_t
+cnxk_representee_msg_thread_main(void *arg)
+{
+ struct cnxk_eswitch_dev *eswitch_dev = (struct cnxk_eswitch_dev *)arg;
+ struct cnxk_esw_repte_msg_proc *repte_msg_proc;
+ struct cnxk_esw_repte_msg *msg, *next_msg;
+ int count, rc;
+
+ repte_msg_proc = &eswitch_dev->repte_msg_proc;
+ pthread_mutex_lock(&eswitch_dev->repte_msg_proc.mutex);
+ while (eswitch_dev->repte_msg_proc.start_thread) {
+ do {
+ rc = pthread_cond_wait(&eswitch_dev->repte_msg_proc.repte_msg_cond,
+ &eswitch_dev->repte_msg_proc.mutex);
+ } while (rc != 0);
+
+ /* Go through list pushed from interrupt context and process each message */
+ next_msg = TAILQ_FIRST(&repte_msg_proc->msg_list);
+ count = 0;
+ while (next_msg) {
+ msg = next_msg;
+ next_msg = TAILQ_NEXT(msg, next);
+ count++;
+ plt_rep_dbg(" Processing msg %d: hw_func %x action %s", count,
+ msg->hw_func, msg->enable ? "enable" : "disable");
+
+ /* Unlocking for interrupt thread to grab lock
+ * while thread process the message.
+ */
+ pthread_mutex_unlock(&eswitch_dev->repte_msg_proc.mutex);
+ /* Processing the message */
+ cnxk_representee_msg_process(eswitch_dev, msg->hw_func, msg->enable);
+ TAILQ_REMOVE(&repte_msg_proc->msg_list, msg, next);
+ rte_free(msg);
+ /* Locking as cond wait will unlock before wait */
+ pthread_mutex_lock(&eswitch_dev->repte_msg_proc.mutex);
+ }
+ }
+
+ pthread_mutex_unlock(&eswitch_dev->repte_msg_proc.mutex);
+
+ return 0;
+}
+
+static int
+cnxk_representee_notification(void *roc_nix, uint16_t hw_func, bool enable)
+{
+ struct cnxk_esw_repte_msg_proc *repte_msg_proc;
+ struct cnxk_eswitch_dev *eswitch_dev;
+ struct cnxk_esw_repte_msg *msg;
+ int rc = 0;
+
+ RTE_SET_USED(roc_nix);
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ if (!eswitch_dev) {
+ plt_err("Failed to get PF ethdev handle");
+ rc = -EINVAL;
+ goto done;
+ }
+
+ repte_msg_proc = &eswitch_dev->repte_msg_proc;
+ msg = rte_zmalloc("msg", sizeof(struct cnxk_esw_repte_msg), 0);
+ if (!msg) {
+ plt_err("Failed to allocate memory for repte msg");
+ rc = -ENOMEM;
+ goto done;
+ }
+
+ msg->hw_func = hw_func;
+ msg->enable = enable;
+
+ plt_rep_dbg("Pushing new notification : hw_func %x enable %d\n", msg->hw_func, enable);
+ pthread_mutex_lock(&eswitch_dev->repte_msg_proc.mutex);
+ TAILQ_INSERT_TAIL(&repte_msg_proc->msg_list, msg, next);
+ /* Signal vf message handler thread */
+ pthread_cond_signal(&eswitch_dev->repte_msg_proc.repte_msg_cond);
+ pthread_mutex_unlock(&eswitch_dev->repte_msg_proc.mutex);
+
+done:
+ return rc;
+}
+
static int
cnxk_rep_parent_setup(struct cnxk_eswitch_dev *eswitch_dev)
{
@@ -263,6 +496,7 @@ create_representor_ethdev(struct rte_pci_device *pci_dev, struct cnxk_eswitch_de
int
cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev)
{
+ char name[REPTE_MSG_PROC_THRD_NAME_MAX_LEN];
struct cnxk_eswitch_devargs *esw_da;
uint16_t num_rep;
int i, j, rc;
@@ -302,7 +536,36 @@ cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswi
}
}
+ if (!eswitch_dev->repte_msg_proc.start_thread) {
+ /* Register callback for representee notification */
+ if (roc_eswitch_nix_process_repte_notify_cb_register(&eswitch_dev->nix,
+ cnxk_representee_notification)) {
+ plt_err("Failed to register callback for representee notification");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Create a thread for handling msgs from VFs */
+ TAILQ_INIT(&eswitch_dev->repte_msg_proc.msg_list);
+ pthread_cond_init(&eswitch_dev->repte_msg_proc.repte_msg_cond, NULL);
+ pthread_mutex_init(&eswitch_dev->repte_msg_proc.mutex, NULL);
+
+ rte_strscpy(name, "repte_msg_proc_thrd", REPTE_MSG_PROC_THRD_NAME_MAX_LEN);
+ eswitch_dev->repte_msg_proc.start_thread = true;
+ rc =
+ rte_thread_create_internal_control(&eswitch_dev->repte_msg_proc.repte_msg_thread,
+ name, cnxk_representee_msg_thread_main,
+ eswitch_dev);
+ if (rc != 0) {
+ plt_err("Failed to create thread for VF mbox handling\n");
+ goto thread_fail;
+ }
+ }
+
return 0;
+thread_fail:
+ pthread_mutex_destroy(&eswitch_dev->repte_msg_proc.mutex);
+ pthread_cond_destroy(&eswitch_dev->repte_msg_proc.repte_msg_cond);
fail:
return rc;
}
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index a62d9b0ae8..9172fae641 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -10,6 +10,40 @@
/* Common ethdev ops */
extern struct eth_dev_ops cnxk_rep_dev_ops;
+struct cnxk_rep_queue_stats {
+ uint64_t pkts;
+ uint64_t bytes;
+};
+
+struct cnxk_rep_rxq {
+ /* Parent rep device */
+ struct cnxk_rep_dev *rep_dev;
+ /* Queue ID */
+ uint16_t qid;
+ /* No of desc */
+ uint16_t nb_desc;
+ /* mempool handle */
+ struct rte_mempool *mpool;
+ /* RX config parameters */
+ const struct rte_eth_rxconf *rx_conf;
+ /* Per queue TX statistics */
+ struct cnxk_rep_queue_stats stats;
+};
+
+struct cnxk_rep_txq {
+ /* Parent rep device */
+ struct cnxk_rep_dev *rep_dev;
+ /* Queue ID */
+ uint16_t qid;
+ /* No of desc */
+ uint16_t nb_desc;
+ /* TX config parameters */
+ const struct rte_eth_txconf *tx_conf;
+ /* Per queue TX statistics */
+ struct cnxk_rep_queue_stats stats;
+};
+
+/* Representor port configurations */
struct cnxk_rep_dev {
uint16_t port_id;
uint16_t rep_id;
@@ -18,6 +52,8 @@ struct cnxk_rep_dev {
uint16_t hw_func;
bool is_vf_active;
bool native_repte;
+ struct cnxk_rep_rxq *rxq;
+ struct cnxk_rep_txq *txq;
uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
};
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 13/24] net/cnxk: representor ethdev ops
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (11 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 12/24] net/cnxk: handling representee notification Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 14/24] common/cnxk: get representees ethernet stats Harman Kalra
` (10 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Implementing ethernet device operation callbacks for
port representors PMD
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_rep.c | 28 +-
drivers/net/cnxk/cnxk_rep.h | 35 +++
drivers/net/cnxk/cnxk_rep_msg.h | 8 +
drivers/net/cnxk/cnxk_rep_ops.c | 495 ++++++++++++++++++++++++++++++--
4 files changed, 523 insertions(+), 43 deletions(-)
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
index 3b01856bc8..6e2424db40 100644
--- a/drivers/net/cnxk/cnxk_rep.c
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -73,6 +73,8 @@ cnxk_rep_state_update(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, ui
int
cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev)
{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+
if (rte_eal_process_type() != RTE_PROC_PRIMARY)
return 0;
@@ -80,6 +82,8 @@ cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev)
rte_free(ethdev->data->mac_addrs);
ethdev->data->mac_addrs = NULL;
+ rep_dev->parent_dev->repr_cnt.nb_repr_probed--;
+
return 0;
}
@@ -369,26 +373,6 @@ cnxk_rep_parent_setup(struct cnxk_eswitch_dev *eswitch_dev)
return rc;
}
-static uint16_t
-cnxk_rep_tx_burst(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
-{
- PLT_SET_USED(tx_queue);
- PLT_SET_USED(tx_pkts);
- PLT_SET_USED(nb_pkts);
-
- return 0;
-}
-
-static uint16_t
-cnxk_rep_rx_burst(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
-{
- PLT_SET_USED(rx_queue);
- PLT_SET_USED(rx_pkts);
- PLT_SET_USED(nb_pkts);
-
- return 0;
-}
-
static int
cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
{
@@ -418,8 +402,8 @@ cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
eth_dev->dev_ops = &cnxk_rep_dev_ops;
/* Rx/Tx functions stubs to avoid crashing */
- eth_dev->rx_pkt_burst = cnxk_rep_rx_burst;
- eth_dev->tx_pkt_burst = cnxk_rep_tx_burst;
+ eth_dev->rx_pkt_burst = cnxk_rep_rx_burst_dummy;
+ eth_dev->tx_pkt_burst = cnxk_rep_tx_burst_dummy;
/* Only single queues for representor devices */
eth_dev->data->nb_rx_queues = 1;
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index 9172fae641..266dd4a688 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -7,6 +7,13 @@
#ifndef __CNXK_REP_H__
#define __CNXK_REP_H__
+#define CNXK_REP_TX_OFFLOAD_CAPA \
+ (RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE | RTE_ETH_TX_OFFLOAD_VLAN_INSERT | \
+ RTE_ETH_TX_OFFLOAD_MULTI_SEGS)
+
+#define CNXK_REP_RX_OFFLOAD_CAPA \
+ (RTE_ETH_RX_OFFLOAD_SCATTER | RTE_ETH_RX_OFFLOAD_RSS_HASH | RTE_ETH_RX_OFFLOAD_VLAN_STRIP)
+
/* Common ethdev ops */
extern struct eth_dev_ops cnxk_rep_dev_ops;
@@ -57,12 +64,33 @@ struct cnxk_rep_dev {
uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
};
+/* Inline functions */
+static inline void
+cnxk_rep_lock(struct cnxk_rep_dev *rep)
+{
+ rte_spinlock_lock(&rep->parent_dev->rep_lock);
+}
+
+static inline void
+cnxk_rep_unlock(struct cnxk_rep_dev *rep)
+{
+ rte_spinlock_unlock(&rep->parent_dev->rep_lock);
+}
+
static inline struct cnxk_rep_dev *
cnxk_rep_pmd_priv(const struct rte_eth_dev *eth_dev)
{
return eth_dev->data->dev_private;
}
+static __rte_always_inline void
+cnxk_rep_pool_buffer_stats(struct rte_mempool *pool)
+{
+ plt_rep_dbg(" pool %s size %d buffer count in use %d available %d\n", pool->name,
+ pool->size, rte_mempool_in_use_count(pool), rte_mempool_avail_count(pool));
+}
+
+/* Prototypes */
int cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
int cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev);
int cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev);
@@ -85,5 +113,12 @@ int cnxk_rep_stats_get(struct rte_eth_dev *eth_dev, struct rte_eth_stats *stats)
int cnxk_rep_stats_reset(struct rte_eth_dev *eth_dev);
int cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops);
int cnxk_rep_state_update(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, uint16_t *rep_id);
+int cnxk_rep_promiscuous_enable(struct rte_eth_dev *ethdev);
+int cnxk_rep_promiscuous_disable(struct rte_eth_dev *ethdev);
+int cnxk_rep_mac_addr_set(struct rte_eth_dev *eth_dev, struct rte_ether_addr *addr);
+uint16_t cnxk_rep_tx_burst_dummy(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts);
+uint16_t cnxk_rep_rx_burst_dummy(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts);
+void cnxk_rep_tx_queue_stop(struct rte_eth_dev *ethdev, uint16_t queue_id);
+void cnxk_rep_rx_queue_stop(struct rte_eth_dev *ethdev, uint16_t queue_id);
#endif /* __CNXK_REP_H__ */
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
index fb84d58848..37953ac74f 100644
--- a/drivers/net/cnxk/cnxk_rep_msg.h
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -19,6 +19,8 @@ typedef enum CNXK_REP_MSG {
CNXK_REP_MSG_READY = 0,
CNXK_REP_MSG_ACK,
CNXK_REP_MSG_EXIT,
+ /* Ethernet operation msgs */
+ CNXK_REP_MSG_ETH_SET_MAC,
/* End of messaging sequence */
CNXK_REP_MSG_END,
} cnxk_rep_msg_t;
@@ -81,6 +83,12 @@ typedef struct cnxk_rep_msg_exit_data {
uint16_t data[];
} __rte_packed cnxk_rep_msg_exit_data_t;
+/* Ethernet op - set mac */
+typedef struct cnxk_rep_msg_eth_mac_set_meta {
+ uint16_t portid;
+ uint8_t addr_bytes[RTE_ETHER_ADDR_LEN];
+} __rte_packed cnxk_rep_msg_eth_set_mac_meta_t;
+
void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
uint32_t size);
void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
index 67dcc422e3..4b3fe28acc 100644
--- a/drivers/net/cnxk/cnxk_rep_ops.c
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -3,25 +3,221 @@
*/
#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
+
+#define MEMPOOL_CACHE_SIZE 256
+#define TX_DESC_PER_QUEUE 512
+#define RX_DESC_PER_QUEUE 256
+#define NB_REP_VDEV_MBUF 1024
+
+static uint16_t
+cnxk_rep_tx_burst(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
+{
+ struct cnxk_rep_txq *txq = tx_queue;
+ struct cnxk_rep_dev *rep_dev;
+ uint16_t n_tx;
+
+ if (unlikely(!txq))
+ return 0;
+
+ rep_dev = txq->rep_dev;
+ plt_rep_dbg("Transmitting %d packets on eswitch queue %d", nb_pkts, txq->qid);
+ n_tx = cnxk_eswitch_dev_tx_burst(rep_dev->parent_dev, txq->qid, tx_pkts, nb_pkts,
+ NIX_TX_OFFLOAD_VLAN_QINQ_F);
+ return n_tx;
+}
+
+static uint16_t
+cnxk_rep_rx_burst(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
+{
+ struct cnxk_rep_rxq *rxq = rx_queue;
+ struct cnxk_rep_dev *rep_dev;
+ uint16_t n_rx;
+
+ if (unlikely(!rxq))
+ return 0;
+
+ rep_dev = rxq->rep_dev;
+ n_rx = cnxk_eswitch_dev_rx_burst(rep_dev->parent_dev, rxq->qid, rx_pkts, nb_pkts);
+ if (n_rx == 0)
+ return 0;
+
+ plt_rep_dbg("Received %d packets on eswitch queue %d", n_rx, rxq->qid);
+ return n_rx;
+}
+
+uint16_t
+cnxk_rep_tx_burst_dummy(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(tx_queue);
+ PLT_SET_USED(tx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
+
+uint16_t
+cnxk_rep_rx_burst_dummy(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(rx_queue);
+ PLT_SET_USED(rx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
int
cnxk_rep_link_update(struct rte_eth_dev *ethdev, int wait_to_complete)
{
- PLT_SET_USED(ethdev);
+ struct rte_eth_link link;
PLT_SET_USED(wait_to_complete);
+
+ memset(&link, 0, sizeof(link));
+ if (ethdev->data->dev_started)
+ link.link_status = RTE_ETH_LINK_UP;
+ else
+ link.link_status = RTE_ETH_LINK_DOWN;
+
+ link.link_duplex = RTE_ETH_LINK_FULL_DUPLEX;
+ link.link_autoneg = RTE_ETH_LINK_FIXED;
+ link.link_speed = RTE_ETH_SPEED_NUM_UNKNOWN;
+
+ return rte_eth_linkstatus_set(ethdev, &link);
+}
+
+int
+cnxk_rep_dev_info_get(struct rte_eth_dev *ethdev, struct rte_eth_dev_info *dev_info)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ uint32_t max_rx_pktlen;
+
+ max_rx_pktlen = (roc_nix_max_pkt_len(&rep_dev->parent_dev->nix) + RTE_ETHER_CRC_LEN -
+ CNXK_NIX_MAX_VTAG_ACT_SIZE);
+
+ dev_info->min_rx_bufsize = NIX_MIN_HW_FRS + RTE_ETHER_CRC_LEN;
+ dev_info->max_rx_pktlen = max_rx_pktlen;
+ dev_info->max_mac_addrs = roc_nix_mac_max_entries_get(&rep_dev->parent_dev->nix);
+
+ dev_info->rx_offload_capa = CNXK_REP_RX_OFFLOAD_CAPA;
+ dev_info->tx_offload_capa = CNXK_REP_TX_OFFLOAD_CAPA;
+ dev_info->rx_queue_offload_capa = 0;
+ dev_info->tx_queue_offload_capa = 0;
+
+ /* For the sake of symmetry, max_rx_queues = max_tx_queues */
+ dev_info->max_rx_queues = 1;
+ dev_info->max_tx_queues = 1;
+
+ /* MTU specifics */
+ dev_info->max_mtu = dev_info->max_rx_pktlen - (RTE_ETHER_HDR_LEN + RTE_ETHER_CRC_LEN);
+ dev_info->min_mtu = dev_info->min_rx_bufsize - CNXK_NIX_L2_OVERHEAD;
+
+ /* Switch info specific */
+ dev_info->switch_info.name = ethdev->device->name;
+ dev_info->switch_info.domain_id = rep_dev->switch_domain_id;
+ dev_info->switch_info.port_id = rep_dev->port_id;
+
return 0;
}
int
-cnxk_rep_dev_info_get(struct rte_eth_dev *ethdev, struct rte_eth_dev_info *devinfo)
+cnxk_rep_representor_info_get(struct rte_eth_dev *ethdev, struct rte_eth_representor_info *info)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+
+ return cnxk_eswitch_representor_info_get(rep_dev->parent_dev, info);
+}
+
+static int
+rep_eth_conf_chk(const struct rte_eth_conf *conf, uint16_t nb_rx_queues)
+{
+ const struct rte_eth_rss_conf *rss_conf;
+ int ret = 0;
+
+ if (conf->link_speeds != 0) {
+ plt_err("specific link speeds not supported");
+ ret = -EINVAL;
+ }
+
+ switch (conf->rxmode.mq_mode) {
+ case RTE_ETH_MQ_RX_RSS:
+ if (nb_rx_queues != 1) {
+ plt_err("Rx RSS is not supported with %u queues", nb_rx_queues);
+ ret = -EINVAL;
+ break;
+ }
+
+ rss_conf = &conf->rx_adv_conf.rss_conf;
+ if (rss_conf->rss_key != NULL || rss_conf->rss_key_len != 0 ||
+ rss_conf->rss_hf != 0) {
+ plt_err("Rx RSS configuration is not supported");
+ ret = -EINVAL;
+ }
+ break;
+ case RTE_ETH_MQ_RX_NONE:
+ break;
+ default:
+ plt_err("Rx mode MQ modes other than RSS not supported");
+ ret = -EINVAL;
+ break;
+ }
+
+ if (conf->txmode.mq_mode != RTE_ETH_MQ_TX_NONE) {
+ plt_err("Tx mode MQ modes not supported");
+ ret = -EINVAL;
+ }
+
+ if (conf->lpbk_mode != 0) {
+ plt_err("loopback not supported");
+ ret = -EINVAL;
+ }
+
+ if (conf->dcb_capability_en != 0) {
+ plt_err("priority-based flow control not supported");
+ ret = -EINVAL;
+ }
+
+ if (conf->intr_conf.lsc != 0) {
+ plt_err("link status change interrupt not supported");
+ ret = -EINVAL;
+ }
+
+ if (conf->intr_conf.rxq != 0) {
+ plt_err("receive queue interrupt not supported");
+ ret = -EINVAL;
+ }
+
+ if (conf->intr_conf.rmv != 0) {
+ plt_err("remove interrupt not supported");
+ ret = -EINVAL;
+ }
+
+ return ret;
+}
+
+int
+cnxk_rep_dev_configure(struct rte_eth_dev *ethdev)
+{
+ struct rte_eth_dev_data *ethdev_data = ethdev->data;
+ int rc = -1;
+
+ rc = rep_eth_conf_chk(ðdev_data->dev_conf, ethdev_data->nb_rx_queues);
+ if (rc)
+ goto fail;
+
+ return 0;
+fail:
+ return rc;
+}
+
+int
+cnxk_rep_promiscuous_enable(struct rte_eth_dev *ethdev)
{
PLT_SET_USED(ethdev);
- PLT_SET_USED(devinfo);
return 0;
}
int
-cnxk_rep_dev_configure(struct rte_eth_dev *ethdev)
+cnxk_rep_promiscuous_disable(struct rte_eth_dev *ethdev)
{
PLT_SET_USED(ethdev);
return 0;
@@ -30,21 +226,73 @@ cnxk_rep_dev_configure(struct rte_eth_dev *ethdev)
int
cnxk_rep_dev_start(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ int rc = 0, qid;
+
+ ethdev->rx_pkt_burst = cnxk_rep_rx_burst;
+ ethdev->tx_pkt_burst = cnxk_rep_tx_burst;
+
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ if (!rep_dev->rxq || !rep_dev->txq) {
+ plt_err("Invalid rxq or txq for representor id %d", rep_dev->rep_id);
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Start rx queues */
+ qid = rep_dev->rxq->qid;
+ rc = cnxk_eswitch_rxq_start(rep_dev->parent_dev, qid);
+ if (rc) {
+ plt_err("Failed to start rxq %d, rc=%d", qid, rc);
+ goto fail;
+ }
+
+ /* Start tx queues */
+ qid = rep_dev->txq->qid;
+ rc = cnxk_eswitch_txq_start(rep_dev->parent_dev, qid);
+ if (rc) {
+ plt_err("Failed to start txq %d, rc=%d", qid, rc);
+ goto fail;
+ }
+
+ /* Start rep_xport device only once after first representor gets active */
+ if (!rep_dev->parent_dev->repr_cnt.nb_repr_started) {
+ rc = cnxk_eswitch_nix_rsrc_start(rep_dev->parent_dev);
+ if (rc) {
+ plt_err("Failed to start nix dev, rc %d", rc);
+ goto fail;
+ }
+ }
+
+ ethdev->data->tx_queue_state[0] = RTE_ETH_QUEUE_STATE_STARTED;
+ ethdev->data->rx_queue_state[0] = RTE_ETH_QUEUE_STATE_STARTED;
+
+ rep_dev->parent_dev->repr_cnt.nb_repr_started++;
+
return 0;
+fail:
+ return rc;
}
int
cnxk_rep_dev_close(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
- return 0;
+ return cnxk_rep_dev_uninit(ethdev);
}
int
cnxk_rep_dev_stop(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+
+ ethdev->rx_pkt_burst = cnxk_rep_rx_burst_dummy;
+ ethdev->tx_pkt_burst = cnxk_rep_tx_burst_dummy;
+ cnxk_rep_rx_queue_stop(ethdev, 0);
+ cnxk_rep_tx_queue_stop(ethdev, 0);
+ rep_dev->parent_dev->repr_cnt.nb_repr_started--;
+
return 0;
}
@@ -53,39 +301,189 @@ cnxk_rep_rx_queue_setup(struct rte_eth_dev *ethdev, uint16_t rx_queue_id, uint16
unsigned int socket_id, const struct rte_eth_rxconf *rx_conf,
struct rte_mempool *mb_pool)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(rx_queue_id);
- PLT_SET_USED(nb_rx_desc);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ struct cnxk_rep_rxq *rxq = NULL;
+ uint16_t qid = 0;
+ int rc;
+
PLT_SET_USED(socket_id);
- PLT_SET_USED(rx_conf);
- PLT_SET_USED(mb_pool);
+ /* If no representee assigned, store the respective rxq parameters */
+ if (!rep_dev->is_vf_active && !rep_dev->rxq) {
+ rxq = plt_zmalloc(sizeof(*rxq), RTE_CACHE_LINE_SIZE);
+ if (!rxq) {
+ rc = -ENOMEM;
+ plt_err("Failed to alloc RxQ for rep id %d", rep_dev->rep_id);
+ goto fail;
+ }
+
+ rxq->qid = qid;
+ rxq->nb_desc = nb_rx_desc;
+ rxq->rep_dev = rep_dev;
+ rxq->mpool = mb_pool;
+ rxq->rx_conf = rx_conf;
+ rep_dev->rxq = rxq;
+ ethdev->data->rx_queues[rx_queue_id] = NULL;
+
+ return 0;
+ }
+
+ qid = rep_dev->rep_id;
+ rc = cnxk_eswitch_rxq_setup(rep_dev->parent_dev, qid, nb_rx_desc, rx_conf, mb_pool);
+ if (rc) {
+ plt_err("failed to setup eswitch queue id %d", qid);
+ goto fail;
+ }
+
+ rxq = rep_dev->rxq;
+ if (!rxq) {
+ plt_err("Invalid RXQ handle for representor port %d rep id %d", rep_dev->port_id,
+ rep_dev->rep_id);
+ goto free_queue;
+ }
+
+ rxq->qid = qid;
+ ethdev->data->rx_queues[rx_queue_id] = rxq;
+ ethdev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
+ plt_rep_dbg("representor id %d portid %d rxq id %d", rep_dev->port_id,
+ ethdev->data->port_id, rxq->qid);
+
return 0;
+free_queue:
+ cnxk_eswitch_rxq_release(rep_dev->parent_dev, qid);
+fail:
+ return rc;
+}
+
+void
+cnxk_rep_rx_queue_stop(struct rte_eth_dev *ethdev, uint16_t queue_id)
+{
+ struct cnxk_rep_rxq *rxq = ethdev->data->rx_queues[queue_id];
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ int rc;
+
+ if (!rxq)
+ return;
+
+ plt_rep_dbg("Stopping rxq %u", rxq->qid);
+
+ rc = cnxk_eswitch_rxq_stop(rep_dev->parent_dev, rxq->qid);
+ if (rc)
+ plt_err("Failed to stop rxq %d, rc=%d", rc, rxq->qid);
+
+ ethdev->data->rx_queue_state[queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
}
void
cnxk_rep_rx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(queue_id);
+ struct cnxk_rep_rxq *rxq = ethdev->data->rx_queues[queue_id];
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ int rc;
+
+ if (!rxq) {
+ plt_err("Invalid rxq retrieved for rep_id %d", rep_dev->rep_id);
+ return;
+ }
+
+ plt_rep_dbg("Releasing rxq %u", rxq->qid);
+
+ rc = cnxk_eswitch_rxq_release(rep_dev->parent_dev, rxq->qid);
+ if (rc)
+ plt_err("Failed to release rxq %d, rc=%d", rc, rxq->qid);
}
int
cnxk_rep_tx_queue_setup(struct rte_eth_dev *ethdev, uint16_t tx_queue_id, uint16_t nb_tx_desc,
unsigned int socket_id, const struct rte_eth_txconf *tx_conf)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(tx_queue_id);
- PLT_SET_USED(nb_tx_desc);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ struct cnxk_rep_txq *txq = NULL;
+ int rc = 0, qid = 0;
+
PLT_SET_USED(socket_id);
- PLT_SET_USED(tx_conf);
+ /* If no representee assigned, store the respective rxq parameters */
+ if (!rep_dev->is_vf_active && !rep_dev->txq) {
+ txq = plt_zmalloc(sizeof(*txq), RTE_CACHE_LINE_SIZE);
+ if (!txq) {
+ rc = -ENOMEM;
+ plt_err("failed to alloc txq for rep id %d", rep_dev->rep_id);
+ goto free_queue;
+ }
+
+ txq->qid = qid;
+ txq->nb_desc = nb_tx_desc;
+ txq->tx_conf = tx_conf;
+ txq->rep_dev = rep_dev;
+ rep_dev->txq = txq;
+
+ ethdev->data->tx_queues[tx_queue_id] = NULL;
+
+ return 0;
+ }
+
+ qid = rep_dev->rep_id;
+ rc = cnxk_eswitch_txq_setup(rep_dev->parent_dev, qid, nb_tx_desc, tx_conf);
+ if (rc) {
+ plt_err("failed to setup eswitch queue id %d", qid);
+ goto fail;
+ }
+
+ txq = rep_dev->txq;
+ if (!txq) {
+ plt_err("Invalid TXQ handle for representor port %d rep id %d", rep_dev->port_id,
+ rep_dev->rep_id);
+ goto free_queue;
+ }
+
+ txq->qid = qid;
+ ethdev->data->tx_queues[tx_queue_id] = txq;
+ ethdev->data->tx_queue_state[tx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
+ plt_rep_dbg("representor id %d portid %d txq id %d", rep_dev->port_id,
+ ethdev->data->port_id, txq->qid);
+
return 0;
+free_queue:
+ cnxk_eswitch_txq_release(rep_dev->parent_dev, qid);
+fail:
+ return rc;
+}
+
+void
+cnxk_rep_tx_queue_stop(struct rte_eth_dev *ethdev, uint16_t queue_id)
+{
+ struct cnxk_rep_txq *txq = ethdev->data->tx_queues[queue_id];
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ int rc;
+
+ if (!txq)
+ return;
+
+ plt_rep_dbg("Releasing txq %u", txq->qid);
+
+ rc = cnxk_eswitch_txq_stop(rep_dev->parent_dev, txq->qid);
+ if (rc)
+ plt_err("Failed to stop txq %d, rc=%d", rc, txq->qid);
+
+ ethdev->data->tx_queue_state[queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
}
void
cnxk_rep_tx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(queue_id);
+ struct cnxk_rep_txq *txq = ethdev->data->tx_queues[queue_id];
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ int rc;
+
+ if (!txq) {
+ plt_err("Invalid txq retrieved for rep_id %d", rep_dev->rep_id);
+ return;
+ }
+
+ plt_rep_dbg("Releasing txq %u", txq->qid);
+
+ rc = cnxk_eswitch_txq_release(rep_dev->parent_dev, txq->qid);
+ if (rc)
+ plt_err("Failed to release txq %d, rc=%d", rc, txq->qid);
}
int
@@ -111,15 +509,70 @@ cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **op
return 0;
}
+int
+cnxk_rep_mac_addr_set(struct rte_eth_dev *eth_dev, struct rte_ether_addr *addr)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_eth_set_mac_meta_t msg_sm_meta;
+ cnxk_rep_msg_ack_data_t adata;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ /* If representor not representing any VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ size = CNXK_REP_MSG_MAX_BUFFER_SZ;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_sm_meta.portid = rep_dev->rep_id;
+ rte_memcpy(&msg_sm_meta.addr_bytes, addr->addr_bytes, RTE_ETHER_ADDR_LEN);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_sm_meta,
+ sizeof(cnxk_rep_msg_eth_set_mac_meta_t),
+ CNXK_REP_MSG_ETH_SET_MAC);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, &adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ if (adata.u.sval < 0) {
+ rc = adata.u.sval;
+ plt_err("Failed to set mac address, err %d", rc);
+ goto fail;
+ }
+
+ rte_free(buffer);
+
+ return 0;
+fail:
+ rte_free(buffer);
+ return rc;
+}
+
/* CNXK platform representor dev ops */
struct eth_dev_ops cnxk_rep_dev_ops = {
.dev_infos_get = cnxk_rep_dev_info_get,
+ .representor_info_get = cnxk_rep_representor_info_get,
.dev_configure = cnxk_rep_dev_configure,
.dev_start = cnxk_rep_dev_start,
.rx_queue_setup = cnxk_rep_rx_queue_setup,
.rx_queue_release = cnxk_rep_rx_queue_release,
.tx_queue_setup = cnxk_rep_tx_queue_setup,
.tx_queue_release = cnxk_rep_tx_queue_release,
+ .promiscuous_enable = cnxk_rep_promiscuous_enable,
+ .promiscuous_disable = cnxk_rep_promiscuous_disable,
+ .mac_addr_set = cnxk_rep_mac_addr_set,
.link_update = cnxk_rep_link_update,
.dev_close = cnxk_rep_dev_close,
.dev_stop = cnxk_rep_dev_stop,
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 14/24] common/cnxk: get representees ethernet stats
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (12 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 13/24] net/cnxk: representor ethdev ops Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 15/24] net/cnxk: ethernet statistic for representor Harman Kalra
` (9 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Implementing an mbox interface to fetch the representees's ethernet
stats from the kernel.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_eswitch.c | 45 +++++++++++++++++++++++++++++++
drivers/common/cnxk/roc_eswitch.h | 2 ++
drivers/common/cnxk/roc_mbox.h | 30 +++++++++++++++++++++
drivers/common/cnxk/version.map | 1 +
4 files changed, 78 insertions(+)
diff --git a/drivers/common/cnxk/roc_eswitch.c b/drivers/common/cnxk/roc_eswitch.c
index 31bdba3985..034a5e6c92 100644
--- a/drivers/common/cnxk/roc_eswitch.c
+++ b/drivers/common/cnxk/roc_eswitch.c
@@ -321,3 +321,48 @@ roc_eswitch_nix_process_repte_notify_cb_unregister(struct roc_nix *roc_nix)
dev->ops->repte_notify = NULL;
}
+
+int
+roc_eswitch_nix_repte_stats(struct roc_nix *roc_nix, uint16_t pf_func, struct roc_nix_stats *stats)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+ struct nix_get_lf_stats_req *req;
+ struct nix_lf_stats_rsp *rsp;
+ struct mbox *mbox;
+ int rc;
+
+ mbox = mbox_get(dev->mbox);
+ req = mbox_alloc_msg_nix_get_lf_stats(mbox);
+ if (!req) {
+ rc = -ENOSPC;
+ goto exit;
+ }
+
+ req->hdr.pcifunc = roc_nix_get_pf_func(roc_nix);
+ req->pcifunc = pf_func;
+
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ stats->rx_octs = rsp->rx.octs;
+ stats->rx_ucast = rsp->rx.ucast;
+ stats->rx_bcast = rsp->rx.bcast;
+ stats->rx_mcast = rsp->rx.mcast;
+ stats->rx_drop = rsp->rx.drop;
+ stats->rx_drop_octs = rsp->rx.drop_octs;
+ stats->rx_drop_bcast = rsp->rx.drop_bcast;
+ stats->rx_drop_mcast = rsp->rx.drop_mcast;
+ stats->rx_err = rsp->rx.err;
+
+ stats->tx_ucast = rsp->tx.ucast;
+ stats->tx_bcast = rsp->tx.bcast;
+ stats->tx_mcast = rsp->tx.mcast;
+ stats->tx_drop = rsp->tx.drop;
+ stats->tx_octs = rsp->tx.octs;
+
+exit:
+ mbox_put(mbox);
+ return rc;
+}
diff --git a/drivers/common/cnxk/roc_eswitch.h b/drivers/common/cnxk/roc_eswitch.h
index 8837e19b22..907e6c37c6 100644
--- a/drivers/common/cnxk/roc_eswitch.h
+++ b/drivers/common/cnxk/roc_eswitch.h
@@ -25,6 +25,8 @@ int __roc_api roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc,
/* NIX */
int __roc_api roc_eswitch_nix_vlan_tpid_set(struct roc_nix *nix, uint32_t type, uint16_t tpid,
bool is_vf);
+int __roc_api roc_eswitch_nix_repte_stats(struct roc_nix *roc_nix, uint16_t pf_func,
+ struct roc_nix_stats *stats);
int __roc_api roc_eswitch_nix_process_repte_notify_cb_register(struct roc_nix *roc_nix,
process_repte_notify_t proc_repte_nt);
void __roc_api roc_eswitch_nix_process_repte_notify_cb_unregister(struct roc_nix *roc_nix);
diff --git a/drivers/common/cnxk/roc_mbox.h b/drivers/common/cnxk/roc_mbox.h
index 2bedf1fb81..1a6bb2f5a2 100644
--- a/drivers/common/cnxk/roc_mbox.h
+++ b/drivers/common/cnxk/roc_mbox.h
@@ -304,6 +304,7 @@ struct mbox_msghdr {
M(NIX_MCAST_GRP_DESTROY, 0x802c, nix_mcast_grp_destroy, nix_mcast_grp_destroy_req, msg_rsp)\
M(NIX_MCAST_GRP_UPDATE, 0x802d, nix_mcast_grp_update, nix_mcast_grp_update_req, \
nix_mcast_grp_update_rsp) \
+ M(NIX_GET_LF_STATS, 0x802e, nix_get_lf_stats, nix_get_lf_stats_req, nix_lf_stats_rsp) \
/* MCS mbox IDs (range 0xa000 - 0xbFFF) */ \
M(MCS_ALLOC_RESOURCES, 0xa000, mcs_alloc_resources, mcs_alloc_rsrc_req, \
mcs_alloc_rsrc_rsp) \
@@ -1846,6 +1847,35 @@ struct nix_mcast_grp_update_rsp {
uint32_t __io mce_start_index;
};
+struct nix_get_lf_stats_req {
+ struct mbox_msghdr hdr;
+ uint16_t __io pcifunc;
+ uint64_t __io rsvd;
+};
+
+struct nix_lf_stats_rsp {
+ struct mbox_msghdr hdr;
+ struct {
+ uint64_t __io octs;
+ uint64_t __io ucast;
+ uint64_t __io bcast;
+ uint64_t __io mcast;
+ uint64_t __io drop;
+ uint64_t __io drop_octs;
+ uint64_t __io drop_mcast;
+ uint64_t __io drop_bcast;
+ uint64_t __io err;
+ uint64_t __io rsvd[5];
+ } rx;
+ struct {
+ uint64_t __io ucast;
+ uint64_t __io bcast;
+ uint64_t __io mcast;
+ uint64_t __io drop;
+ uint64_t __io octs;
+ } tx;
+};
+
/* Global NIX inline IPSec configuration */
struct nix_inline_ipsec_cfg {
struct mbox_msghdr hdr;
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index e170a6a63a..87c9d7511f 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -93,6 +93,7 @@ INTERNAL {
roc_error_msg_get;
roc_eswitch_nix_process_repte_notify_cb_register;
roc_eswitch_nix_process_repte_notify_cb_unregister;
+ roc_eswitch_nix_repte_stats;
roc_eswitch_nix_vlan_tpid_set;
roc_eswitch_npc_mcam_delete_rule;
roc_eswitch_npc_mcam_rx_rule;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 15/24] net/cnxk: ethernet statistic for representor
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (13 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 14/24] common/cnxk: get representees ethernet stats Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 16/24] common/cnxk: base support for eswitch VF Harman Kalra
` (8 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Adding representor ethernet statistics support which can fetch stats
for representees which are operating independently or part of
companian app.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_rep_msg.h | 7 ++
drivers/net/cnxk/cnxk_rep_ops.c | 140 +++++++++++++++++++++++++++++++-
2 files changed, 143 insertions(+), 4 deletions(-)
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
index 37953ac74f..3236de50ad 100644
--- a/drivers/net/cnxk/cnxk_rep_msg.h
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -21,6 +21,8 @@ typedef enum CNXK_REP_MSG {
CNXK_REP_MSG_EXIT,
/* Ethernet operation msgs */
CNXK_REP_MSG_ETH_SET_MAC,
+ CNXK_REP_MSG_ETH_STATS_GET,
+ CNXK_REP_MSG_ETH_STATS_CLEAR,
/* End of messaging sequence */
CNXK_REP_MSG_END,
} cnxk_rep_msg_t;
@@ -89,6 +91,11 @@ typedef struct cnxk_rep_msg_eth_mac_set_meta {
uint8_t addr_bytes[RTE_ETHER_ADDR_LEN];
} __rte_packed cnxk_rep_msg_eth_set_mac_meta_t;
+/* Ethernet op - get/clear stats */
+typedef struct cnxk_rep_msg_eth_stats_meta {
+ uint16_t portid;
+} __rte_packed cnxk_rep_msg_eth_stats_meta_t;
+
void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
uint32_t size);
void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
index 4b3fe28acc..e07c63dcb2 100644
--- a/drivers/net/cnxk/cnxk_rep_ops.c
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -486,19 +486,151 @@ cnxk_rep_tx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
plt_err("Failed to release txq %d, rc=%d", rc, txq->qid);
}
+static int
+process_eth_stats(struct cnxk_rep_dev *rep_dev, cnxk_rep_msg_ack_data_t *adata, cnxk_rep_msg_t msg)
+{
+ cnxk_rep_msg_eth_stats_meta_t msg_st_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ size = CNXK_REP_MSG_MAX_BUFFER_SZ;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_st_meta.portid = rep_dev->rep_id;
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_st_meta,
+ sizeof(cnxk_rep_msg_eth_stats_meta_t), msg);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ rte_free(buffer);
+
+ return 0;
+fail:
+ rte_free(buffer);
+ return rc;
+}
+
+static int
+native_repte_eth_stats(struct cnxk_rep_dev *rep_dev, struct rte_eth_stats *stats)
+{
+ struct roc_nix_stats nix_stats;
+ int rc = 0;
+
+ rc = roc_eswitch_nix_repte_stats(&rep_dev->parent_dev->nix, rep_dev->hw_func, &nix_stats);
+ if (rc) {
+ plt_err("Failed to get stats for representee %x, err %d", rep_dev->hw_func, rc);
+ goto fail;
+ }
+
+ memset(stats, 0, sizeof(struct rte_eth_stats));
+ stats->opackets = nix_stats.tx_ucast;
+ stats->opackets += nix_stats.tx_mcast;
+ stats->opackets += nix_stats.tx_bcast;
+ stats->oerrors = nix_stats.tx_drop;
+ stats->obytes = nix_stats.tx_octs;
+
+ stats->ipackets = nix_stats.rx_ucast;
+ stats->ipackets += nix_stats.rx_mcast;
+ stats->ipackets += nix_stats.rx_bcast;
+ stats->imissed = nix_stats.rx_drop;
+ stats->ibytes = nix_stats.rx_octs;
+ stats->ierrors = nix_stats.rx_err;
+
+ return 0;
+fail:
+ return rc;
+}
+
int
cnxk_rep_stats_get(struct rte_eth_dev *ethdev, struct rte_eth_stats *stats)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(stats);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ struct rte_eth_stats vf_stats;
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ if (rep_dev->native_repte) {
+ /* For representees which are independent */
+ rc = native_repte_eth_stats(rep_dev, &vf_stats);
+ if (rc) {
+ plt_err("Failed to get stats for vf rep %x (hw_func %x), err %d",
+ rep_dev->port_id, rep_dev->hw_func, rc);
+ goto fail;
+ }
+ } else {
+ /* For representees which are part of companian app */
+ rc = process_eth_stats(rep_dev, &adata, CNXK_REP_MSG_ETH_STATS_GET);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ plt_err("Failed to get stats for vf rep %x, err %d", rep_dev->port_id, rc);
+ }
+
+ if (adata.size != sizeof(struct rte_eth_stats)) {
+ rc = -EINVAL;
+ plt_err("Incomplete stats received for vf rep %d", rep_dev->port_id);
+ goto fail;
+ }
+
+ rte_memcpy(&vf_stats, adata.u.data, adata.size);
+ }
+
+ stats->q_ipackets[0] = vf_stats.ipackets;
+ stats->q_ibytes[0] = vf_stats.ibytes;
+ stats->ipackets = vf_stats.ipackets;
+ stats->ibytes = vf_stats.ibytes;
+
+ stats->q_opackets[0] = vf_stats.opackets;
+ stats->q_obytes[0] = vf_stats.obytes;
+ stats->opackets = vf_stats.opackets;
+ stats->obytes = vf_stats.obytes;
+
+ plt_rep_dbg("Input packets %" PRId64 " Output packets %" PRId64 "", stats->ipackets,
+ stats->opackets);
+
return 0;
+fail:
+ return rc;
}
int
cnxk_rep_stats_reset(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
- return 0;
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc = 0;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ rc = process_eth_stats(rep_dev, &adata, CNXK_REP_MSG_ETH_STATS_CLEAR);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ plt_err("Failed to clear stats for vf rep %x, err %d", rep_dev->port_id, rc);
+ }
+
+ return rc;
}
int
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 16/24] common/cnxk: base support for eswitch VF
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (14 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 15/24] net/cnxk: ethernet statistic for representor Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 17/24] net/cnxk: eswitch VF as ethernet device Harman Kalra
` (7 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
- ROC layer changes for supporting eswitch VF
- NIX lbk changes for esw
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_constants.h | 1 +
drivers/common/cnxk/roc_dev.c | 1 +
drivers/common/cnxk/roc_nix.c | 15 +++++++++++++--
drivers/common/cnxk/roc_nix.h | 1 +
drivers/common/cnxk/roc_nix_priv.h | 1 +
drivers/common/cnxk/version.map | 1 +
6 files changed, 18 insertions(+), 2 deletions(-)
diff --git a/drivers/common/cnxk/roc_constants.h b/drivers/common/cnxk/roc_constants.h
index cb4edbea58..21b3998cee 100644
--- a/drivers/common/cnxk/roc_constants.h
+++ b/drivers/common/cnxk/roc_constants.h
@@ -44,6 +44,7 @@
#define PCI_DEVID_CNXK_RVU_REE_PF 0xA0f4
#define PCI_DEVID_CNXK_RVU_REE_VF 0xA0f5
#define PCI_DEVID_CNXK_RVU_ESWITCH_PF 0xA0E0
+#define PCI_DEVID_CNXK_RVU_ESWITCH_VF 0xA0E1
#define PCI_DEVID_CN9K_CGX 0xA059
#define PCI_DEVID_CN10K_RPM 0xA060
diff --git a/drivers/common/cnxk/roc_dev.c b/drivers/common/cnxk/roc_dev.c
index b12732de34..4d4cfeaaca 100644
--- a/drivers/common/cnxk/roc_dev.c
+++ b/drivers/common/cnxk/roc_dev.c
@@ -1225,6 +1225,7 @@ dev_vf_hwcap_update(struct plt_pci_device *pci_dev, struct dev *dev)
case PCI_DEVID_CNXK_RVU_VF:
case PCI_DEVID_CNXK_RVU_SDP_VF:
case PCI_DEVID_CNXK_RVU_NIX_INL_VF:
+ case PCI_DEVID_CNXK_RVU_ESWITCH_VF:
dev->hwcap |= DEV_HWCAP_F_VF;
break;
}
diff --git a/drivers/common/cnxk/roc_nix.c b/drivers/common/cnxk/roc_nix.c
index 7e327a7e6e..f1eaca3ab4 100644
--- a/drivers/common/cnxk/roc_nix.c
+++ b/drivers/common/cnxk/roc_nix.c
@@ -13,6 +13,14 @@ roc_nix_is_lbk(struct roc_nix *roc_nix)
return nix->lbk_link;
}
+bool
+roc_nix_is_esw(struct roc_nix *roc_nix)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+
+ return nix->esw_link;
+}
+
int
roc_nix_get_base_chan(struct roc_nix *roc_nix)
{
@@ -156,7 +164,7 @@ roc_nix_max_pkt_len(struct roc_nix *roc_nix)
if (roc_model_is_cn9k())
return NIX_CN9K_MAX_HW_FRS;
- if (nix->lbk_link)
+ if (nix->lbk_link || nix->esw_link)
return NIX_LBK_MAX_HW_FRS;
return NIX_RPM_MAX_HW_FRS;
@@ -349,7 +357,7 @@ roc_nix_get_hw_info(struct roc_nix *roc_nix)
rc = mbox_process_msg(mbox, (void *)&hw_info);
if (rc == 0) {
nix->vwqe_interval = hw_info->vwqe_delay;
- if (nix->lbk_link)
+ if (nix->lbk_link || nix->esw_link)
roc_nix->dwrr_mtu = hw_info->lbk_dwrr_mtu;
else if (nix->sdp_link)
roc_nix->dwrr_mtu = hw_info->sdp_dwrr_mtu;
@@ -366,6 +374,7 @@ sdp_lbk_id_update(struct plt_pci_device *pci_dev, struct nix *nix)
{
nix->sdp_link = false;
nix->lbk_link = false;
+ nix->esw_link = false;
/* Update SDP/LBK link based on PCI device id */
switch (pci_dev->id.device_id) {
@@ -374,7 +383,9 @@ sdp_lbk_id_update(struct plt_pci_device *pci_dev, struct nix *nix)
nix->sdp_link = true;
break;
case PCI_DEVID_CNXK_RVU_AF_VF:
+ case PCI_DEVID_CNXK_RVU_ESWITCH_VF:
nix->lbk_link = true;
+ nix->esw_link = true;
break;
default:
break;
diff --git a/drivers/common/cnxk/roc_nix.h b/drivers/common/cnxk/roc_nix.h
index b369335fc4..ffea84dae8 100644
--- a/drivers/common/cnxk/roc_nix.h
+++ b/drivers/common/cnxk/roc_nix.h
@@ -527,6 +527,7 @@ int __roc_api roc_nix_dev_fini(struct roc_nix *roc_nix);
/* Type */
bool __roc_api roc_nix_is_lbk(struct roc_nix *roc_nix);
+bool __roc_api roc_nix_is_esw(struct roc_nix *roc_nix);
bool __roc_api roc_nix_is_sdp(struct roc_nix *roc_nix);
bool __roc_api roc_nix_is_pf(struct roc_nix *roc_nix);
bool __roc_api roc_nix_is_vf_or_sdp(struct roc_nix *roc_nix);
diff --git a/drivers/common/cnxk/roc_nix_priv.h b/drivers/common/cnxk/roc_nix_priv.h
index 8767a62577..e2f65a49c8 100644
--- a/drivers/common/cnxk/roc_nix_priv.h
+++ b/drivers/common/cnxk/roc_nix_priv.h
@@ -170,6 +170,7 @@ struct nix {
uintptr_t base;
bool sdp_link;
bool lbk_link;
+ bool esw_link;
bool ptp_en;
bool is_nix1;
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index 87c9d7511f..cdb46d8739 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -276,6 +276,7 @@ INTERNAL {
roc_nix_inl_outb_cpt_lfs_dump;
roc_nix_cpt_ctx_cache_sync;
roc_nix_is_lbk;
+ roc_nix_is_esw;
roc_nix_is_pf;
roc_nix_is_sdp;
roc_nix_is_vf_or_sdp;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 17/24] net/cnxk: eswitch VF as ethernet device
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (15 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 16/24] common/cnxk: base support for eswitch VF Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 18/24] common/cnxk: support port representor and represented port Harman Kalra
` (6 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Adding support for eswitch VF to probe as normal cnxk ethernet device
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cn10k_ethdev.c | 1 +
drivers/net/cnxk/cnxk_ethdev.c | 39 ++++++++++++++++++++++--------
drivers/net/cnxk/cnxk_ethdev.h | 3 +++
drivers/net/cnxk/cnxk_ethdev_ops.c | 4 +++
drivers/net/cnxk/cnxk_link.c | 3 ++-
5 files changed, 39 insertions(+), 11 deletions(-)
diff --git a/drivers/net/cnxk/cn10k_ethdev.c b/drivers/net/cnxk/cn10k_ethdev.c
index a2e943a3d0..9a072b72a7 100644
--- a/drivers/net/cnxk/cn10k_ethdev.c
+++ b/drivers/net/cnxk/cn10k_ethdev.c
@@ -963,6 +963,7 @@ static const struct rte_pci_id cn10k_pci_nix_map[] = {
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CN10KB, PCI_DEVID_CNXK_RVU_PF),
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CNF10KB, PCI_DEVID_CNXK_RVU_PF),
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CN10KA, PCI_DEVID_CNXK_RVU_VF),
+ CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CN10KA, PCI_DEVID_CNXK_RVU_ESWITCH_VF),
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CN10KAS, PCI_DEVID_CNXK_RVU_VF),
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CNF10KA, PCI_DEVID_CNXK_RVU_VF),
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CN10KB, PCI_DEVID_CNXK_RVU_VF),
diff --git a/drivers/net/cnxk/cnxk_ethdev.c b/drivers/net/cnxk/cnxk_ethdev.c
index 2372a4e793..50f1641c38 100644
--- a/drivers/net/cnxk/cnxk_ethdev.c
+++ b/drivers/net/cnxk/cnxk_ethdev.c
@@ -1449,12 +1449,14 @@ cnxk_nix_configure(struct rte_eth_dev *eth_dev)
goto cq_fini;
/* Init flow control configuration */
- fc_cfg.type = ROC_NIX_FC_RXCHAN_CFG;
- fc_cfg.rxchan_cfg.enable = true;
- rc = roc_nix_fc_config_set(nix, &fc_cfg);
- if (rc) {
- plt_err("Failed to initialize flow control rc=%d", rc);
- goto cq_fini;
+ if (!roc_nix_is_esw(nix)) {
+ fc_cfg.type = ROC_NIX_FC_RXCHAN_CFG;
+ fc_cfg.rxchan_cfg.enable = true;
+ rc = roc_nix_fc_config_set(nix, &fc_cfg);
+ if (rc) {
+ plt_err("Failed to initialize flow control rc=%d", rc);
+ goto cq_fini;
+ }
}
/* Update flow control configuration to PMD */
@@ -1688,10 +1690,12 @@ cnxk_nix_dev_start(struct rte_eth_dev *eth_dev)
}
/* Update Flow control configuration */
- rc = nix_update_flow_ctrl_config(eth_dev);
- if (rc) {
- plt_err("Failed to enable flow control. error code(%d)", rc);
- return rc;
+ if (!roc_nix_is_esw(&dev->nix)) {
+ rc = nix_update_flow_ctrl_config(eth_dev);
+ if (rc) {
+ plt_err("Failed to enable flow control. error code(%d)", rc);
+ return rc;
+ }
}
/* Enable Rx in NPC */
@@ -1976,6 +1980,16 @@ cnxk_eth_dev_init(struct rte_eth_dev *eth_dev)
TAILQ_INIT(&dev->mcs_list);
}
+ /* Reserve a switch domain for eswitch device */
+ if (pci_dev->id.device_id == PCI_DEVID_CNXK_RVU_ESWITCH_VF) {
+ eth_dev->data->dev_flags |= RTE_ETH_DEV_REPRESENTOR;
+ rc = rte_eth_switch_domain_alloc(&dev->switch_domain_id);
+ if (rc) {
+ plt_err("Failed to alloc switch domain: %d", rc);
+ goto free_mac_addrs;
+ }
+ }
+
plt_nix_dbg("Port=%d pf=%d vf=%d ver=%s hwcap=0x%" PRIx64
" rxoffload_capa=0x%" PRIx64 " txoffload_capa=0x%" PRIx64,
eth_dev->data->port_id, roc_nix_get_pf(nix),
@@ -2046,6 +2060,11 @@ cnxk_eth_dev_uninit(struct rte_eth_dev *eth_dev, bool reset)
}
}
+ /* Free switch domain ID reserved for eswitch device */
+ if ((eth_dev->data->dev_flags & RTE_ETH_DEV_REPRESENTOR) &&
+ rte_eth_switch_domain_free(dev->switch_domain_id))
+ plt_err("Failed to free switch domain");
+
/* Disable and free rte_meter entries */
nix_meter_fini(dev);
diff --git a/drivers/net/cnxk/cnxk_ethdev.h b/drivers/net/cnxk/cnxk_ethdev.h
index 4d3ebf123b..d8eba5e1dd 100644
--- a/drivers/net/cnxk/cnxk_ethdev.h
+++ b/drivers/net/cnxk/cnxk_ethdev.h
@@ -424,6 +424,9 @@ struct cnxk_eth_dev {
/* MCS device */
struct cnxk_mcs_dev *mcs_dev;
struct cnxk_macsec_sess_list mcs_list;
+
+ /* Eswitch domain ID */
+ uint16_t switch_domain_id;
};
struct cnxk_eth_rxq_sp {
diff --git a/drivers/net/cnxk/cnxk_ethdev_ops.c b/drivers/net/cnxk/cnxk_ethdev_ops.c
index 5de2919047..67fbf7c269 100644
--- a/drivers/net/cnxk/cnxk_ethdev_ops.c
+++ b/drivers/net/cnxk/cnxk_ethdev_ops.c
@@ -71,6 +71,10 @@ cnxk_nix_info_get(struct rte_eth_dev *eth_dev, struct rte_eth_dev_info *devinfo)
RTE_ETH_DEV_CAPA_FLOW_RULE_KEEP;
devinfo->max_rx_mempools = CNXK_NIX_NUM_POOLS_MAX;
+ if (eth_dev->data->dev_flags & RTE_ETH_DEV_REPRESENTOR) {
+ devinfo->switch_info.name = eth_dev->device->name;
+ devinfo->switch_info.domain_id = dev->switch_domain_id;
+ }
return 0;
}
diff --git a/drivers/net/cnxk/cnxk_link.c b/drivers/net/cnxk/cnxk_link.c
index 127c9e72e7..903b44de2c 100644
--- a/drivers/net/cnxk/cnxk_link.c
+++ b/drivers/net/cnxk/cnxk_link.c
@@ -13,7 +13,8 @@ cnxk_nix_toggle_flag_link_cfg(struct cnxk_eth_dev *dev, bool set)
dev->flags &= ~CNXK_LINK_CFG_IN_PROGRESS_F;
/* Update link info for LBK */
- if (!set && (roc_nix_is_lbk(&dev->nix) || roc_nix_is_sdp(&dev->nix))) {
+ if (!set &&
+ (roc_nix_is_lbk(&dev->nix) || roc_nix_is_sdp(&dev->nix) || roc_nix_is_esw(&dev->nix))) {
struct rte_eth_link link;
link.link_status = RTE_ETH_LINK_UP;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 18/24] common/cnxk: support port representor and represented port
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (16 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 17/24] net/cnxk: eswitch VF as ethernet device Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 19/24] net/cnxk: add represented port pattern and action Harman Kalra
` (5 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj, Satheesh Paul
Implementing the common infrastructural changes for supporting port
representors and represented ports used as action and pattern in net
layer.
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Satheesh Paul <psatheesh@marvell.com>
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_npc.c | 63 +++++++++++++++++++++++------
drivers/common/cnxk/roc_npc.h | 13 +++++-
drivers/common/cnxk/roc_npc_mcam.c | 62 +++++++++++++++-------------
drivers/common/cnxk/roc_npc_parse.c | 28 ++++++++++++-
drivers/common/cnxk/roc_npc_priv.h | 2 +
drivers/net/cnxk/cnxk_flow.c | 2 +-
6 files changed, 125 insertions(+), 45 deletions(-)
diff --git a/drivers/common/cnxk/roc_npc.c b/drivers/common/cnxk/roc_npc.c
index 67a660a2bc..5a836f16f5 100644
--- a/drivers/common/cnxk/roc_npc.c
+++ b/drivers/common/cnxk/roc_npc.c
@@ -570,6 +570,8 @@ npc_parse_actions(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
flow->ctr_id = NPC_COUNTER_NONE;
flow->mtr_id = ROC_NIX_MTR_ID_INVALID;
pf_func = npc->pf_func;
+ if (flow->has_rep)
+ pf_func = flow->rep_pf_func;
for (; actions->type != ROC_NPC_ACTION_TYPE_END; actions++) {
switch (actions->type) {
@@ -898,10 +900,14 @@ npc_parse_pattern(struct npc *npc, const struct roc_npc_item_info pattern[],
struct roc_npc_flow *flow, struct npc_parse_state *pst)
{
npc_parse_stage_func_t parse_stage_funcs[] = {
- npc_parse_meta_items, npc_parse_mark_item, npc_parse_pre_l2, npc_parse_cpt_hdr,
- npc_parse_higig2_hdr, npc_parse_tx_queue, npc_parse_la, npc_parse_lb,
- npc_parse_lc, npc_parse_ld, npc_parse_le, npc_parse_lf,
- npc_parse_lg, npc_parse_lh,
+ npc_parse_meta_items, npc_parse_port_representor_id,
+ npc_parse_mark_item, npc_parse_pre_l2,
+ npc_parse_cpt_hdr, npc_parse_higig2_hdr,
+ npc_parse_tx_queue, npc_parse_la,
+ npc_parse_lb, npc_parse_lc,
+ npc_parse_ld, npc_parse_le,
+ npc_parse_lf, npc_parse_lg,
+ npc_parse_lh,
};
uint8_t layer = 0;
int key_offset;
@@ -1140,15 +1146,20 @@ npc_rss_action_program(struct roc_npc *roc_npc,
struct roc_npc_flow *flow)
{
const struct roc_npc_action_rss *rss;
+ struct roc_npc *npc = roc_npc;
uint32_t rss_grp;
uint8_t alg_idx;
int rc;
+ if (flow->has_rep) {
+ npc = roc_npc->rep_npc;
+ npc->flowkey_cfg_state = roc_npc->flowkey_cfg_state;
+ }
+
for (; actions->type != ROC_NPC_ACTION_TYPE_END; actions++) {
if (actions->type == ROC_NPC_ACTION_TYPE_RSS) {
rss = (const struct roc_npc_action_rss *)actions->conf;
- rc = npc_rss_action_configure(roc_npc, rss, &alg_idx,
- &rss_grp, flow->mcam_id);
+ rc = npc_rss_action_configure(npc, rss, &alg_idx, &rss_grp, flow->mcam_id);
if (rc)
return rc;
@@ -1171,7 +1182,7 @@ npc_vtag_cfg_delete(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
struct roc_nix *roc_nix = roc_npc->roc_nix;
struct nix_vtag_config *vtag_cfg;
struct nix_vtag_config_rsp *rsp;
- struct mbox *mbox;
+ struct mbox *mbox, *ombox;
struct nix *nix;
int rc = 0;
@@ -1181,7 +1192,10 @@ npc_vtag_cfg_delete(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
} tx_vtag_action;
nix = roc_nix_to_nix_priv(roc_nix);
- mbox = mbox_get((&nix->dev)->mbox);
+ ombox = (&nix->dev)->mbox;
+ if (flow->has_rep)
+ ombox = flow->rep_mbox;
+ mbox = mbox_get(ombox);
tx_vtag_action.reg = flow->vtag_action;
vtag_cfg = mbox_alloc_msg_nix_vtag_cfg(mbox);
@@ -1400,6 +1414,7 @@ npc_vtag_strip_action_configure(struct mbox *mbox,
rx_vtag_action |= (NIX_RX_VTAGACTION_VTAG_VALID << 15);
rx_vtag_action |= ((uint64_t)NPC_LID_LB << 8);
+ rx_vtag_action |= (NIX_RX_VTAG_TYPE7 << 12);
rx_vtag_action |= NIX_RX_VTAGACTION_VTAG0_RELPTR;
if (*strip_cnt == 2) {
@@ -1432,6 +1447,8 @@ npc_vtag_action_program(struct roc_npc *roc_npc,
nix = roc_nix_to_nix_priv(roc_nix);
mbox = (&nix->dev)->mbox;
+ if (flow->has_rep)
+ mbox = flow->rep_mbox;
memset(vlan_info, 0, sizeof(vlan_info));
@@ -1448,6 +1465,7 @@ npc_vtag_action_program(struct roc_npc *roc_npc,
if (rc)
return rc;
+ plt_npc_dbg("VLAN strip action, strip_cnt %d", strip_cnt);
if (strip_cnt == 2)
actions++;
@@ -1587,6 +1605,17 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
memset(flow, 0, sizeof(*flow));
memset(&parse_state, 0, sizeof(parse_state));
+ flow->port_id = -1;
+ if (roc_npc->rep_npc) {
+ flow->rep_channel = roc_nix_to_nix_priv(roc_npc->rep_npc->roc_nix)->rx_chan_base;
+ flow->rep_pf_func = roc_npc->rep_pf_func;
+ flow->rep_mbox = roc_npc_to_npc_priv(roc_npc->rep_npc)->mbox;
+ flow->has_rep = true;
+ flow->is_rep_vf = !roc_nix_is_pf(roc_npc->rep_npc->roc_nix);
+ flow->port_id = roc_npc->rep_port_id;
+ flow->rep_npc = roc_npc_to_npc_priv(roc_npc->rep_npc);
+ }
+
parse_state.dst_pf_func = dst_pf_func;
rc = npc_parse_rule(roc_npc, attr, pattern, actions, flow, &parse_state);
@@ -1629,6 +1658,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
*errcode = rc;
goto set_rss_failed;
}
+ roc_npc->rep_npc = NULL;
if (flow->has_age_action)
npc_age_flow_list_entry_add(roc_npc, flow);
@@ -1641,6 +1671,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
TAILQ_FOREACH(flow_iter, list, next) {
if (flow_iter->mcam_id > flow->mcam_id) {
TAILQ_INSERT_BEFORE(flow_iter, flow, next);
+ roc_npc->rep_npc = NULL;
return flow;
}
}
@@ -1649,6 +1680,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
return flow;
set_rss_failed:
+ roc_npc->rep_npc = NULL;
if (flow->use_pre_alloc == 0) {
rc = roc_npc_mcam_free_entry(roc_npc, flow->mcam_id);
if (rc != 0) {
@@ -1660,6 +1692,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
npc_inline_dev_ipsec_action_free(npc, flow);
}
err_exit:
+ roc_npc->rep_npc = NULL;
plt_free(flow);
return NULL;
}
@@ -1667,15 +1700,19 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
int
npc_rss_group_free(struct npc *npc, struct roc_npc_flow *flow)
{
+ struct npc *lnpc = npc;
uint32_t rss_grp;
+ if (flow->has_rep)
+ lnpc = flow->rep_npc;
+
if ((flow->npc_action & 0xF) == NIX_RX_ACTIONOP_RSS) {
rss_grp = (flow->npc_action >> NPC_RSS_ACT_GRP_OFFSET) &
NPC_RSS_ACT_GRP_MASK;
if (rss_grp == 0 || rss_grp >= npc->rss_grps)
return -EINVAL;
- plt_bitmap_clear(npc->rss_grp_entries, rss_grp);
+ plt_bitmap_clear(lnpc->rss_grp_entries, rss_grp);
}
return 0;
@@ -1770,7 +1807,7 @@ roc_npc_flow_destroy(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
}
void
-roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc)
+roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc, int rep_port_id)
{
struct npc *npc = roc_npc_to_npc_priv(roc_npc);
struct roc_npc_flow *flow_iter;
@@ -1784,12 +1821,14 @@ roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc)
/* List in ascending order of mcam entries */
TAILQ_FOREACH(flow_iter, list, next) {
- roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
+ if (rep_port_id == -1 || rep_port_id == flow_iter->port_id)
+ roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
}
}
TAILQ_FOREACH(flow_iter, &npc->ipsec_list, next) {
- roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
+ if (rep_port_id == -1 || rep_port_id == flow_iter->port_id)
+ roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
}
}
diff --git a/drivers/common/cnxk/roc_npc.h b/drivers/common/cnxk/roc_npc.h
index 349c7f9d22..03432909c7 100644
--- a/drivers/common/cnxk/roc_npc.h
+++ b/drivers/common/cnxk/roc_npc.h
@@ -42,6 +42,7 @@ enum roc_npc_item_type {
ROC_NPC_ITEM_TYPE_MARK,
ROC_NPC_ITEM_TYPE_TX_QUEUE,
ROC_NPC_ITEM_TYPE_IPV6_ROUTING_EXT,
+ ROC_NPC_ITEM_TYPE_REPRESENTED_PORT,
ROC_NPC_ITEM_TYPE_END,
};
@@ -339,6 +340,13 @@ struct roc_npc_flow {
#define ROC_NPC_MIRROR_LIST_SIZE 2
uint16_t mcast_pf_funcs[ROC_NPC_MIRROR_LIST_SIZE];
uint16_t mcast_channels[ROC_NPC_MIRROR_LIST_SIZE];
+ uint16_t rep_pf_func;
+ uint16_t rep_channel;
+ struct mbox *rep_mbox;
+ bool has_rep;
+ bool is_rep_vf;
+ struct npc *rep_npc;
+ int port_id;
TAILQ_ENTRY(roc_npc_flow) next;
};
@@ -407,6 +415,9 @@ struct roc_npc {
uint16_t sdp_channel;
uint16_t sdp_channel_mask;
struct roc_npc_flow_age flow_age;
+ struct roc_npc *rep_npc;
+ uint16_t rep_pf_func;
+ int rep_port_id;
#define ROC_NPC_MEM_SZ (6 * 1024)
uint8_t reserved[ROC_NPC_MEM_SZ];
@@ -448,7 +459,7 @@ int __roc_api roc_npc_get_free_mcam_entry(struct roc_npc *roc_npc, struct roc_np
int __roc_api roc_npc_inl_mcam_read_counter(uint32_t ctr_id, uint64_t *count);
int __roc_api roc_npc_inl_mcam_clear_counter(uint32_t ctr_id);
int __roc_api roc_npc_mcam_free_all_resources(struct roc_npc *roc_npc);
-void __roc_api roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc);
+void __roc_api roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc, int rep_port_id);
void __roc_api roc_npc_flow_mcam_dump(FILE *file, struct roc_npc *roc_npc,
struct roc_npc_flow *mcam);
int __roc_api roc_npc_mark_actions_get(struct roc_npc *roc_npc);
diff --git a/drivers/common/cnxk/roc_npc_mcam.c b/drivers/common/cnxk/roc_npc_mcam.c
index 2de988a44b..f2d5004c78 100644
--- a/drivers/common/cnxk/roc_npc_mcam.c
+++ b/drivers/common/cnxk/roc_npc_mcam.c
@@ -143,8 +143,8 @@ npc_lid_lt_in_kex(struct npc *npc, uint8_t lid, uint8_t lt)
}
static void
-npc_construct_ldata_mask(struct npc *npc, struct plt_bitmap *bmap, uint8_t lid,
- uint8_t lt, uint8_t ld)
+npc_construct_ldata_mask(struct npc *npc, struct plt_bitmap *bmap, uint8_t lid, uint8_t lt,
+ uint8_t ld)
{
struct npc_xtract_info *x_info, *infoflag;
int hdr_off, keylen;
@@ -197,8 +197,7 @@ npc_construct_ldata_mask(struct npc *npc, struct plt_bitmap *bmap, uint8_t lid,
* @param len length of the match
*/
static bool
-npc_is_kex_enabled(struct npc *npc, uint8_t lid, uint8_t lt, int offset,
- int len)
+npc_is_kex_enabled(struct npc *npc, uint8_t lid, uint8_t lt, int offset, int len)
{
struct plt_bitmap *bmap;
uint32_t bmap_sz;
@@ -349,8 +348,8 @@ npc_mcam_alloc_entries(struct mbox *mbox, int ref_mcam, int *alloc_entry, int re
}
int
-npc_mcam_alloc_entry(struct npc *npc, struct roc_npc_flow *mcam,
- struct roc_npc_flow *ref_mcam, int prio, int *resp_count)
+npc_mcam_alloc_entry(struct npc *npc, struct roc_npc_flow *mcam, struct roc_npc_flow *ref_mcam,
+ int prio, int *resp_count)
{
struct npc_mcam_alloc_entry_req *req;
struct npc_mcam_alloc_entry_rsp *rsp;
@@ -450,22 +449,17 @@ npc_mcam_write_entry(struct mbox *mbox, struct roc_npc_flow *mcam)
static void
npc_mcam_process_mkex_cfg(struct npc *npc, struct npc_get_kex_cfg_rsp *kex_rsp)
{
- volatile uint64_t(
- *q)[NPC_MAX_INTF][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD];
+ volatile uint64_t(*q)[NPC_MAX_INTF][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD];
struct npc_xtract_info *x_info = NULL;
int lid, lt, ld, fl, ix;
npc_dxcfg_t *p;
uint64_t keyw;
uint64_t val;
- npc->keyx_supp_nmask[NPC_MCAM_RX] =
- kex_rsp->rx_keyx_cfg & 0x7fffffffULL;
- npc->keyx_supp_nmask[NPC_MCAM_TX] =
- kex_rsp->tx_keyx_cfg & 0x7fffffffULL;
- npc->keyx_len[NPC_MCAM_RX] =
- npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_RX]);
- npc->keyx_len[NPC_MCAM_TX] =
- npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_TX]);
+ npc->keyx_supp_nmask[NPC_MCAM_RX] = kex_rsp->rx_keyx_cfg & 0x7fffffffULL;
+ npc->keyx_supp_nmask[NPC_MCAM_TX] = kex_rsp->tx_keyx_cfg & 0x7fffffffULL;
+ npc->keyx_len[NPC_MCAM_RX] = npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_RX]);
+ npc->keyx_len[NPC_MCAM_TX] = npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_TX]);
keyw = (kex_rsp->rx_keyx_cfg >> 32) & 0x7ULL;
npc->keyw[NPC_MCAM_RX] = keyw;
@@ -485,8 +479,7 @@ npc_mcam_process_mkex_cfg(struct npc *npc, struct npc_get_kex_cfg_rsp *kex_rsp)
/* Update LID, LT and LDATA cfg */
p = &npc->prx_dxcfg;
- q = (volatile uint64_t(*)[][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD])(
- &kex_rsp->intf_lid_lt_ld);
+ q = (volatile uint64_t(*)[][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD])(&kex_rsp->intf_lid_lt_ld);
for (ix = 0; ix < NPC_MAX_INTF; ix++) {
for (lid = 0; lid < NPC_MAX_LID; lid++) {
for (lt = 0; lt < NPC_MAX_LT; lt++) {
@@ -539,8 +532,7 @@ npc_mcam_fetch_kex_cfg(struct npc *npc)
goto done;
}
- mbox_memcpy((char *)npc->profile_name, kex_rsp->mkex_pfl_name,
- MKEX_NAME_LEN);
+ mbox_memcpy((char *)npc->profile_name, kex_rsp->mkex_pfl_name, MKEX_NAME_LEN);
npc->exact_match_ena = (kex_rsp->rx_keyx_cfg >> 40) & 0xF;
npc_mcam_process_mkex_cfg(npc, kex_rsp);
@@ -551,9 +543,8 @@ npc_mcam_fetch_kex_cfg(struct npc *npc)
}
static void
-npc_mcam_set_channel(struct roc_npc_flow *flow,
- struct npc_mcam_write_entry_req *req, uint16_t channel,
- uint16_t chan_mask, bool is_second_pass)
+npc_mcam_set_channel(struct roc_npc_flow *flow, struct npc_mcam_write_entry_req *req,
+ uint16_t channel, uint16_t chan_mask, bool is_second_pass)
{
uint16_t chan = 0, mask = 0;
@@ -683,6 +674,9 @@ npc_mcam_alloc_and_write(struct npc *npc, struct roc_npc_flow *flow, struct npc_
if (flow->nix_intf == NIX_INTF_TX) {
uint16_t pf_func = (flow->npc_action >> 4) & 0xffff;
+ if (flow->has_rep)
+ pf_func = flow->rep_pf_func;
+
pf_func = plt_cpu_to_be_16(pf_func);
rc = npc_mcam_set_pf_func(npc, flow, pf_func);
@@ -759,6 +753,14 @@ npc_mcam_alloc_and_write(struct npc *npc, struct roc_npc_flow *flow, struct npc_
npc_mcam_set_channel(flow, req, inl_dev->channel, inl_dev->chan_mask,
false);
+ } else if (flow->has_rep) {
+ pf_func = flow->rep_pf_func;
+ req->entry_data.action &= ~(GENMASK(19, 4));
+ req->entry_data.action |= (uint64_t)pf_func << 4;
+ flow->npc_action &= ~(GENMASK(19, 4));
+ flow->npc_action |= (uint64_t)pf_func << 4;
+ npc_mcam_set_channel(flow, req, flow->rep_channel, (BIT_ULL(12) - 1),
+ false);
} else if (npc->is_sdp_link) {
npc_mcam_set_channel(flow, req, npc->sdp_channel, npc->sdp_channel_mask,
pst->is_second_pass_rule);
@@ -932,13 +934,11 @@ npc_program_mcam(struct npc *npc, struct npc_parse_state *pst, bool mcam_alloc)
data_off = 0;
index++;
}
- key_data[index] |=
- ((uint64_t)data << data_off);
+ key_data[index] |= ((uint64_t)data << data_off);
if (lt == 0)
mask = 0;
- key_mask[index] |=
- ((uint64_t)mask << data_off);
+ key_mask[index] |= ((uint64_t)mask << data_off);
data_off += 4;
}
}
@@ -963,8 +963,12 @@ npc_program_mcam(struct npc *npc, struct npc_parse_state *pst, bool mcam_alloc)
(pst->flow->npc_action & NIX_RX_ACTIONOP_UCAST_IPSEC))
skip_base_rule = true;
- if (pst->is_vf && pst->flow->nix_intf == NIX_INTF_RX && !skip_base_rule) {
- mbox = mbox_get(npc->mbox);
+ if ((pst->is_vf || pst->flow->is_rep_vf) && pst->flow->nix_intf == NIX_INTF_RX &&
+ !skip_base_rule) {
+ if (pst->flow->has_rep)
+ mbox = mbox_get(pst->flow->rep_mbox);
+ else
+ mbox = mbox_get(npc->mbox);
(void)mbox_alloc_msg_npc_read_base_steer_rule(mbox);
rc = mbox_process_msg(mbox, (void *)&base_rule_rsp);
if (rc) {
diff --git a/drivers/common/cnxk/roc_npc_parse.c b/drivers/common/cnxk/roc_npc_parse.c
index 9ceb707ebb..af1b9f79dd 100644
--- a/drivers/common/cnxk/roc_npc_parse.c
+++ b/drivers/common/cnxk/roc_npc_parse.c
@@ -35,11 +35,35 @@ npc_parse_mark_item(struct npc_parse_state *pst)
return 0;
}
+int
+npc_parse_port_representor_id(struct npc_parse_state *pst)
+{
+ if (pst->pattern->type != ROC_NPC_ITEM_TYPE_REPRESENTED_PORT)
+ return 0;
+
+ pst->pattern++;
+
+ return 0;
+}
+
+int
+npc_parse_represented_port_id(struct npc_parse_state *pst)
+{
+ if (pst->pattern->type != ROC_NPC_ITEM_TYPE_REPRESENTED_PORT)
+ return 0;
+
+ if (pst->flow->nix_intf != NIX_INTF_RX)
+ return -EINVAL;
+
+ pst->pattern++;
+
+ return 0;
+}
+
static int
npc_flow_raw_item_prepare(const struct roc_npc_flow_item_raw *raw_spec,
const struct roc_npc_flow_item_raw *raw_mask,
- struct npc_parse_item_info *info, uint8_t *spec_buf,
- uint8_t *mask_buf)
+ struct npc_parse_item_info *info, uint8_t *spec_buf, uint8_t *mask_buf)
{
memset(spec_buf, 0, NPC_MAX_RAW_ITEM_LEN);
diff --git a/drivers/common/cnxk/roc_npc_priv.h b/drivers/common/cnxk/roc_npc_priv.h
index 50b62b1244..069c625911 100644
--- a/drivers/common/cnxk/roc_npc_priv.h
+++ b/drivers/common/cnxk/roc_npc_priv.h
@@ -457,6 +457,8 @@ int npc_mask_is_supported(const char *mask, const char *hw_mask, int len);
int npc_parse_item_basic(const struct roc_npc_item_info *item, struct npc_parse_item_info *info);
int npc_parse_meta_items(struct npc_parse_state *pst);
int npc_parse_mark_item(struct npc_parse_state *pst);
+int npc_parse_port_representor_id(struct npc_parse_state *pst);
+int npc_parse_represented_port_id(struct npc_parse_state *pst);
int npc_parse_pre_l2(struct npc_parse_state *pst);
int npc_parse_higig2_hdr(struct npc_parse_state *pst);
int npc_parse_cpt_hdr(struct npc_parse_state *pst);
diff --git a/drivers/net/cnxk/cnxk_flow.c b/drivers/net/cnxk/cnxk_flow.c
index a92b61c332..5f74c356b1 100644
--- a/drivers/net/cnxk/cnxk_flow.c
+++ b/drivers/net/cnxk/cnxk_flow.c
@@ -594,7 +594,7 @@ cnxk_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
return -EINVAL;
}
- roc_npc_flow_dump(file, npc);
+ roc_npc_flow_dump(file, npc, -1);
return 0;
}
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 19/24] net/cnxk: add represented port pattern and action
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (17 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 18/24] common/cnxk: support port representor and represented port Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:39 ` [PATCH v2 20/24] net/cnxk: add port representor " Harman Kalra
` (4 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Adding support for represented_port item matching and action.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_flow.c | 107 +++++++++++++++++++----------------
1 file changed, 57 insertions(+), 50 deletions(-)
diff --git a/drivers/net/cnxk/cnxk_flow.c b/drivers/net/cnxk/cnxk_flow.c
index 5f74c356b1..a3b21f761f 100644
--- a/drivers/net/cnxk/cnxk_flow.c
+++ b/drivers/net/cnxk/cnxk_flow.c
@@ -4,67 +4,48 @@
#include <cnxk_flow.h>
const struct cnxk_rte_flow_term_info term[] = {
- [RTE_FLOW_ITEM_TYPE_ETH] = {ROC_NPC_ITEM_TYPE_ETH,
- sizeof(struct rte_flow_item_eth)},
- [RTE_FLOW_ITEM_TYPE_VLAN] = {ROC_NPC_ITEM_TYPE_VLAN,
- sizeof(struct rte_flow_item_vlan)},
- [RTE_FLOW_ITEM_TYPE_E_TAG] = {ROC_NPC_ITEM_TYPE_E_TAG,
- sizeof(struct rte_flow_item_e_tag)},
- [RTE_FLOW_ITEM_TYPE_IPV4] = {ROC_NPC_ITEM_TYPE_IPV4,
- sizeof(struct rte_flow_item_ipv4)},
- [RTE_FLOW_ITEM_TYPE_IPV6] = {ROC_NPC_ITEM_TYPE_IPV6,
- sizeof(struct rte_flow_item_ipv6)},
- [RTE_FLOW_ITEM_TYPE_IPV6_FRAG_EXT] = {
- ROC_NPC_ITEM_TYPE_IPV6_FRAG_EXT,
- sizeof(struct rte_flow_item_ipv6_frag_ext)},
- [RTE_FLOW_ITEM_TYPE_ARP_ETH_IPV4] = {
- ROC_NPC_ITEM_TYPE_ARP_ETH_IPV4,
- sizeof(struct rte_flow_item_arp_eth_ipv4)},
- [RTE_FLOW_ITEM_TYPE_MPLS] = {ROC_NPC_ITEM_TYPE_MPLS,
- sizeof(struct rte_flow_item_mpls)},
- [RTE_FLOW_ITEM_TYPE_ICMP] = {ROC_NPC_ITEM_TYPE_ICMP,
- sizeof(struct rte_flow_item_icmp)},
- [RTE_FLOW_ITEM_TYPE_UDP] = {ROC_NPC_ITEM_TYPE_UDP,
- sizeof(struct rte_flow_item_udp)},
- [RTE_FLOW_ITEM_TYPE_TCP] = {ROC_NPC_ITEM_TYPE_TCP,
- sizeof(struct rte_flow_item_tcp)},
- [RTE_FLOW_ITEM_TYPE_SCTP] = {ROC_NPC_ITEM_TYPE_SCTP,
- sizeof(struct rte_flow_item_sctp)},
- [RTE_FLOW_ITEM_TYPE_ESP] = {ROC_NPC_ITEM_TYPE_ESP,
- sizeof(struct rte_flow_item_esp)},
- [RTE_FLOW_ITEM_TYPE_GRE] = {ROC_NPC_ITEM_TYPE_GRE,
- sizeof(struct rte_flow_item_gre)},
- [RTE_FLOW_ITEM_TYPE_NVGRE] = {ROC_NPC_ITEM_TYPE_NVGRE,
- sizeof(struct rte_flow_item_nvgre)},
- [RTE_FLOW_ITEM_TYPE_VXLAN] = {ROC_NPC_ITEM_TYPE_VXLAN,
- sizeof(struct rte_flow_item_vxlan)},
- [RTE_FLOW_ITEM_TYPE_GTPC] = {ROC_NPC_ITEM_TYPE_GTPC,
- sizeof(struct rte_flow_item_gtp)},
- [RTE_FLOW_ITEM_TYPE_GTPU] = {ROC_NPC_ITEM_TYPE_GTPU,
- sizeof(struct rte_flow_item_gtp)},
+ [RTE_FLOW_ITEM_TYPE_ETH] = {ROC_NPC_ITEM_TYPE_ETH, sizeof(struct rte_flow_item_eth)},
+ [RTE_FLOW_ITEM_TYPE_VLAN] = {ROC_NPC_ITEM_TYPE_VLAN, sizeof(struct rte_flow_item_vlan)},
+ [RTE_FLOW_ITEM_TYPE_E_TAG] = {ROC_NPC_ITEM_TYPE_E_TAG, sizeof(struct rte_flow_item_e_tag)},
+ [RTE_FLOW_ITEM_TYPE_IPV4] = {ROC_NPC_ITEM_TYPE_IPV4, sizeof(struct rte_flow_item_ipv4)},
+ [RTE_FLOW_ITEM_TYPE_IPV6] = {ROC_NPC_ITEM_TYPE_IPV6, sizeof(struct rte_flow_item_ipv6)},
+ [RTE_FLOW_ITEM_TYPE_IPV6_FRAG_EXT] = {ROC_NPC_ITEM_TYPE_IPV6_FRAG_EXT,
+ sizeof(struct rte_flow_item_ipv6_frag_ext)},
+ [RTE_FLOW_ITEM_TYPE_ARP_ETH_IPV4] = {ROC_NPC_ITEM_TYPE_ARP_ETH_IPV4,
+ sizeof(struct rte_flow_item_arp_eth_ipv4)},
+ [RTE_FLOW_ITEM_TYPE_MPLS] = {ROC_NPC_ITEM_TYPE_MPLS, sizeof(struct rte_flow_item_mpls)},
+ [RTE_FLOW_ITEM_TYPE_ICMP] = {ROC_NPC_ITEM_TYPE_ICMP, sizeof(struct rte_flow_item_icmp)},
+ [RTE_FLOW_ITEM_TYPE_UDP] = {ROC_NPC_ITEM_TYPE_UDP, sizeof(struct rte_flow_item_udp)},
+ [RTE_FLOW_ITEM_TYPE_TCP] = {ROC_NPC_ITEM_TYPE_TCP, sizeof(struct rte_flow_item_tcp)},
+ [RTE_FLOW_ITEM_TYPE_SCTP] = {ROC_NPC_ITEM_TYPE_SCTP, sizeof(struct rte_flow_item_sctp)},
+ [RTE_FLOW_ITEM_TYPE_ESP] = {ROC_NPC_ITEM_TYPE_ESP, sizeof(struct rte_flow_item_esp)},
+ [RTE_FLOW_ITEM_TYPE_GRE] = {ROC_NPC_ITEM_TYPE_GRE, sizeof(struct rte_flow_item_gre)},
+ [RTE_FLOW_ITEM_TYPE_NVGRE] = {ROC_NPC_ITEM_TYPE_NVGRE, sizeof(struct rte_flow_item_nvgre)},
+ [RTE_FLOW_ITEM_TYPE_VXLAN] = {ROC_NPC_ITEM_TYPE_VXLAN, sizeof(struct rte_flow_item_vxlan)},
+ [RTE_FLOW_ITEM_TYPE_GTPC] = {ROC_NPC_ITEM_TYPE_GTPC, sizeof(struct rte_flow_item_gtp)},
+ [RTE_FLOW_ITEM_TYPE_GTPU] = {ROC_NPC_ITEM_TYPE_GTPU, sizeof(struct rte_flow_item_gtp)},
[RTE_FLOW_ITEM_TYPE_GENEVE] = {ROC_NPC_ITEM_TYPE_GENEVE,
sizeof(struct rte_flow_item_geneve)},
- [RTE_FLOW_ITEM_TYPE_VXLAN_GPE] = {
- ROC_NPC_ITEM_TYPE_VXLAN_GPE,
- sizeof(struct rte_flow_item_vxlan_gpe)},
+ [RTE_FLOW_ITEM_TYPE_VXLAN_GPE] = {ROC_NPC_ITEM_TYPE_VXLAN_GPE,
+ sizeof(struct rte_flow_item_vxlan_gpe)},
[RTE_FLOW_ITEM_TYPE_IPV6_EXT] = {ROC_NPC_ITEM_TYPE_IPV6_EXT,
sizeof(struct rte_flow_item_ipv6_ext)},
[RTE_FLOW_ITEM_TYPE_VOID] = {ROC_NPC_ITEM_TYPE_VOID, 0},
[RTE_FLOW_ITEM_TYPE_ANY] = {ROC_NPC_ITEM_TYPE_ANY, 0},
- [RTE_FLOW_ITEM_TYPE_GRE_KEY] = {ROC_NPC_ITEM_TYPE_GRE_KEY,
- sizeof(uint32_t)},
+ [RTE_FLOW_ITEM_TYPE_GRE_KEY] = {ROC_NPC_ITEM_TYPE_GRE_KEY, sizeof(uint32_t)},
[RTE_FLOW_ITEM_TYPE_HIGIG2] = {ROC_NPC_ITEM_TYPE_HIGIG2,
sizeof(struct rte_flow_item_higig2_hdr)},
- [RTE_FLOW_ITEM_TYPE_RAW] = {ROC_NPC_ITEM_TYPE_RAW,
- sizeof(struct rte_flow_item_raw)},
- [RTE_FLOW_ITEM_TYPE_MARK] = {ROC_NPC_ITEM_TYPE_MARK,
- sizeof(struct rte_flow_item_mark)},
+ [RTE_FLOW_ITEM_TYPE_RAW] = {ROC_NPC_ITEM_TYPE_RAW, sizeof(struct rte_flow_item_raw)},
+ [RTE_FLOW_ITEM_TYPE_MARK] = {ROC_NPC_ITEM_TYPE_MARK, sizeof(struct rte_flow_item_mark)},
[RTE_FLOW_ITEM_TYPE_IPV6_ROUTING_EXT] = {ROC_NPC_ITEM_TYPE_IPV6_ROUTING_EXT,
- sizeof(struct rte_flow_item_ipv6_routing_ext)},
+ sizeof(struct rte_flow_item_ipv6_routing_ext)},
[RTE_FLOW_ITEM_TYPE_TX_QUEUE] = {ROC_NPC_ITEM_TYPE_TX_QUEUE,
- sizeof(struct rte_flow_item_tx_queue)},
+ sizeof(struct rte_flow_item_tx_queue)},
+ [RTE_FLOW_ITEM_TYPE_REPRESENTED_PORT] = {ROC_NPC_ITEM_TYPE_REPRESENTED_PORT,
+ sizeof(struct rte_flow_item_ethdev)},
[RTE_FLOW_ITEM_TYPE_PPPOES] = {ROC_NPC_ITEM_TYPE_PPPOES,
- sizeof(struct rte_flow_item_pppoe)}};
+ sizeof(struct rte_flow_item_pppoe)}
+};
static int
npc_rss_action_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
@@ -372,6 +353,11 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
uint16_t *dst_pf_func)
{
+ struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
+ const struct rte_flow_item_ethdev *rep_eth_dev;
+ struct rte_eth_dev *portid_eth_dev;
+ char if_name[RTE_ETH_NAME_MAX_LEN];
+ struct cnxk_eth_dev *hw_dst;
int i = 0;
in_attr->priority = attr->priority;
@@ -384,6 +370,27 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
in_pattern[i].mask = pattern->mask;
in_pattern[i].type = term[pattern->type].item_type;
in_pattern[i].size = term[pattern->type].item_size;
+ if (pattern->type == RTE_FLOW_ITEM_TYPE_REPRESENTED_PORT) {
+ rep_eth_dev = (const struct rte_flow_item_ethdev *)pattern->spec;
+ if (rte_eth_dev_get_name_by_port(rep_eth_dev->port_id, if_name)) {
+ plt_err("Name not found for output port id");
+ return -EINVAL;
+ }
+ portid_eth_dev = rte_eth_dev_allocated(if_name);
+ if (!portid_eth_dev) {
+ plt_err("eth_dev not found for output port id");
+ return -EINVAL;
+ }
+ if (strcmp(portid_eth_dev->device->driver->name,
+ eth_dev->device->driver->name) != 0) {
+ plt_err("Output port not under same driver");
+ return -EINVAL;
+ }
+ hw_dst = portid_eth_dev->data->dev_private;
+ dev->npc.rep_npc = &hw_dst->npc;
+ dev->npc.rep_port_id = rep_eth_dev->port_id;
+ dev->npc.rep_pf_func = hw_dst->npc.pf_func;
+ }
pattern++;
i++;
}
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 20/24] net/cnxk: add port representor pattern and action
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (18 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 19/24] net/cnxk: add represented port pattern and action Harman Kalra
@ 2023-12-19 17:39 ` Harman Kalra
2023-12-19 17:40 ` [PATCH v2 21/24] net/cnxk: generalize flow operation APIs Harman Kalra
` (3 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:39 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Adding support for port_representor as item matching and action.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_flow.c | 224 +++++++++++++++++++++++++++++++----
drivers/net/cnxk/cnxk_rep.h | 14 +++
2 files changed, 212 insertions(+), 26 deletions(-)
diff --git a/drivers/net/cnxk/cnxk_flow.c b/drivers/net/cnxk/cnxk_flow.c
index a3b21f761f..959d773513 100644
--- a/drivers/net/cnxk/cnxk_flow.c
+++ b/drivers/net/cnxk/cnxk_flow.c
@@ -2,6 +2,7 @@
* Copyright(C) 2021 Marvell.
*/
#include <cnxk_flow.h>
+#include <cnxk_rep.h>
const struct cnxk_rte_flow_term_info term[] = {
[RTE_FLOW_ITEM_TYPE_ETH] = {ROC_NPC_ITEM_TYPE_ETH, sizeof(struct rte_flow_item_eth)},
@@ -185,11 +186,44 @@ roc_npc_parse_sample_subaction(struct rte_eth_dev *eth_dev, const struct rte_flo
return 0;
}
+static int
+representor_portid_action(struct roc_npc_action *in_actions, struct rte_eth_dev *portid_eth_dev,
+ uint16_t *dst_pf_func, uint8_t has_tunnel_pattern, int *act_cnt)
+{
+ struct rte_eth_dev *rep_eth_dev = portid_eth_dev;
+ struct rte_flow_action_mark *act_mark;
+ struct cnxk_rep_dev *rep_dev;
+ /* For inserting an action in the list */
+ int i = *act_cnt;
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ *dst_pf_func = rep_dev->hw_func;
+
+ /* Add Mark action */
+ i++;
+ act_mark = plt_zmalloc(sizeof(struct rte_flow_action_mark), 0);
+ if (!act_mark) {
+ plt_err("Error allocation memory");
+ return -ENOMEM;
+ }
+
+ /* Mark ID format: (tunnel type - VxLAN, Geneve << 6) | Tunnel decap */
+ act_mark->id = has_tunnel_pattern ? ((has_tunnel_pattern << 6) | 5) : 1;
+ in_actions[i].type = ROC_NPC_ACTION_TYPE_MARK;
+ in_actions[i].conf = (struct rte_flow_action_mark *)act_mark;
+
+ *act_cnt = i;
+ plt_rep_dbg("Rep port %d ID %d mark ID is %d rep_dev->hw_func 0x%x", rep_dev->port_id,
+ rep_dev->rep_id, act_mark->id, rep_dev->hw_func);
+
+ return 0;
+}
+
static int
cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
const struct rte_flow_action actions[], struct roc_npc_action in_actions[],
struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
- uint16_t *dst_pf_func)
+ uint16_t *dst_pf_func, uint8_t has_tunnel_pattern)
{
struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
const struct rte_flow_action_queue *act_q = NULL;
@@ -256,14 +290,27 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
plt_err("eth_dev not found for output port id");
goto err_exit;
}
- if (strcmp(portid_eth_dev->device->driver->name,
- eth_dev->device->driver->name) != 0) {
- plt_err("Output port not under same driver");
- goto err_exit;
+
+ if (cnxk_ethdev_is_representor(if_name)) {
+ plt_rep_dbg("Representor port %d act port %d", port_act->id,
+ act_ethdev->port_id);
+ if (representor_portid_action(in_actions, portid_eth_dev,
+ dst_pf_func, has_tunnel_pattern,
+ &i)) {
+ plt_err("Representor port action set failed");
+ goto err_exit;
+ }
+ } else {
+ if (strcmp(portid_eth_dev->device->driver->name,
+ eth_dev->device->driver->name) != 0) {
+ plt_err("Output port not under same driver");
+ goto err_exit;
+ }
+
+ hw_dst = portid_eth_dev->data->dev_private;
+ roc_npc_dst = &hw_dst->npc;
+ *dst_pf_func = roc_npc_dst->pf_func;
}
- hw_dst = portid_eth_dev->data->dev_private;
- roc_npc_dst = &hw_dst->npc;
- *dst_pf_func = roc_npc_dst->pf_func;
break;
case RTE_FLOW_ACTION_TYPE_QUEUE:
@@ -324,6 +371,8 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
in_actions[i].type = ROC_NPC_ACTION_TYPE_SAMPLE;
in_actions[i].conf = in_sample_actions;
break;
+ case RTE_FLOW_ACTION_TYPE_VXLAN_DECAP:
+ continue;
default:
plt_npc_dbg("Action is not supported = %d", actions->type);
goto err_exit;
@@ -346,12 +395,8 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
}
static int
-cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
- const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
- struct roc_npc_attr *in_attr, struct roc_npc_item_info in_pattern[],
- struct roc_npc_action in_actions[],
- struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
- uint16_t *dst_pf_func)
+cnxk_map_pattern(struct rte_eth_dev *eth_dev, const struct rte_flow_item pattern[],
+ struct roc_npc_item_info in_pattern[], uint8_t *has_tunnel_pattern)
{
struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
const struct rte_flow_item_ethdev *rep_eth_dev;
@@ -360,10 +405,6 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
struct cnxk_eth_dev *hw_dst;
int i = 0;
- in_attr->priority = attr->priority;
- in_attr->ingress = attr->ingress;
- in_attr->egress = attr->egress;
-
while (pattern->type != RTE_FLOW_ITEM_TYPE_END) {
in_pattern[i].spec = pattern->spec;
in_pattern[i].last = pattern->last;
@@ -374,30 +415,81 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
rep_eth_dev = (const struct rte_flow_item_ethdev *)pattern->spec;
if (rte_eth_dev_get_name_by_port(rep_eth_dev->port_id, if_name)) {
plt_err("Name not found for output port id");
- return -EINVAL;
+ goto fail;
}
portid_eth_dev = rte_eth_dev_allocated(if_name);
if (!portid_eth_dev) {
plt_err("eth_dev not found for output port id");
- return -EINVAL;
+ goto fail;
}
if (strcmp(portid_eth_dev->device->driver->name,
eth_dev->device->driver->name) != 0) {
plt_err("Output port not under same driver");
- return -EINVAL;
+ goto fail;
+ }
+ if (cnxk_ethdev_is_representor(if_name)) {
+ /* Case where represented port not part of same
+ * app and represented by a representor port.
+ */
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eswitch_dev *eswitch_dev;
+
+ rep_dev = cnxk_rep_pmd_priv(portid_eth_dev);
+ eswitch_dev = rep_dev->parent_dev;
+ dev->npc.rep_npc = &eswitch_dev->npc;
+ dev->npc.rep_port_id = rep_eth_dev->port_id;
+ dev->npc.rep_pf_func = rep_dev->hw_func;
+ plt_rep_dbg("Represented port %d act port %d rep_dev->hw_func 0x%x",
+ rep_eth_dev->port_id, eth_dev->data->port_id,
+ rep_dev->hw_func);
+ } else {
+ /* Case where represented port part of same app
+ * as PF.
+ */
+ hw_dst = portid_eth_dev->data->dev_private;
+ dev->npc.rep_npc = &hw_dst->npc;
+ dev->npc.rep_port_id = rep_eth_dev->port_id;
+ dev->npc.rep_pf_func = hw_dst->npc.pf_func;
}
- hw_dst = portid_eth_dev->data->dev_private;
- dev->npc.rep_npc = &hw_dst->npc;
- dev->npc.rep_port_id = rep_eth_dev->port_id;
- dev->npc.rep_pf_func = hw_dst->npc.pf_func;
}
+
+ if (pattern->type == RTE_FLOW_ITEM_TYPE_VXLAN ||
+ pattern->type == RTE_FLOW_ITEM_TYPE_VXLAN_GPE ||
+ pattern->type == RTE_FLOW_ITEM_TYPE_GRE)
+ *has_tunnel_pattern = pattern->type;
+
pattern++;
i++;
}
in_pattern[i].type = ROC_NPC_ITEM_TYPE_END;
+ return 0;
+fail:
+ return -EINVAL;
+}
+
+static int
+cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct roc_npc_attr *in_attr, struct roc_npc_item_info in_pattern[],
+ struct roc_npc_action in_actions[],
+ struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
+ uint16_t *dst_pf_func)
+{
+ uint8_t has_tunnel_pattern = 0;
+ int rc;
+
+ in_attr->priority = attr->priority;
+ in_attr->ingress = attr->ingress;
+ in_attr->egress = attr->egress;
+
+ rc = cnxk_map_pattern(eth_dev, pattern, in_pattern, &has_tunnel_pattern);
+ if (rc) {
+ plt_err("Failed to map pattern list");
+ return rc;
+ }
return cnxk_map_actions(eth_dev, attr, actions, in_actions, in_sample_actions, flowkey_cfg,
- dst_pf_func);
+ dst_pf_func, has_tunnel_pattern);
}
static int
@@ -461,6 +553,7 @@ cnxk_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
int rc;
memset(&in_sample_action, 0, sizeof(in_sample_action));
+ memset(&in_attr, 0, sizeof(struct roc_npc_attr));
rc = cnxk_map_flow_data(eth_dev, attr, pattern, actions, &in_attr, in_pattern, in_actions,
&in_sample_action, &npc->flowkey_cfg_state, &dst_pf_func);
if (rc) {
@@ -646,6 +739,81 @@ cnxk_flow_get_aged_flows(struct rte_eth_dev *eth_dev, void **context,
return cnt;
}
+static int
+cnxk_flow_tunnel_decap_set(__rte_unused struct rte_eth_dev *dev, struct rte_flow_tunnel *tunnel,
+ struct rte_flow_action **pmd_actions, uint32_t *num_of_actions,
+ __rte_unused struct rte_flow_error *err)
+{
+ struct rte_flow_action *nfp_action;
+
+ nfp_action = rte_zmalloc("nfp_tun_action", sizeof(struct rte_flow_action), 0);
+ if (nfp_action == NULL) {
+ plt_err("Alloc memory for nfp tunnel action failed.");
+ return -ENOMEM;
+ }
+
+ if (tunnel->is_ipv6)
+ nfp_action->conf = (void *)~0;
+
+ switch (tunnel->type) {
+ case RTE_FLOW_ITEM_TYPE_VXLAN:
+ nfp_action->type = RTE_FLOW_ACTION_TYPE_VXLAN_DECAP;
+ *pmd_actions = nfp_action;
+ *num_of_actions = 1;
+ break;
+ case RTE_FLOW_ITEM_TYPE_GENEVE:
+ case RTE_FLOW_ITEM_TYPE_GRE:
+ nfp_action->type = RTE_FLOW_ACTION_TYPE_RAW_DECAP;
+ *pmd_actions = nfp_action;
+ *num_of_actions = 1;
+ break;
+ default:
+ *pmd_actions = NULL;
+ *num_of_actions = 0;
+ rte_free(nfp_action);
+ break;
+ }
+
+ return 0;
+}
+
+static int
+cnxk_flow_tunnel_action_decap_release(__rte_unused struct rte_eth_dev *dev,
+ struct rte_flow_action *pmd_actions, uint32_t num_of_actions,
+ __rte_unused struct rte_flow_error *err)
+{
+ uint32_t i;
+ struct rte_flow_action *nfp_action;
+
+ for (i = 0; i < num_of_actions; i++) {
+ nfp_action = &pmd_actions[i];
+ nfp_action->conf = NULL;
+ rte_free(nfp_action);
+ }
+
+ return 0;
+}
+
+static int
+cnxk_flow_tunnel_match(__rte_unused struct rte_eth_dev *dev,
+ __rte_unused struct rte_flow_tunnel *tunnel,
+ __rte_unused struct rte_flow_item **pmd_items, uint32_t *num_of_items,
+ __rte_unused struct rte_flow_error *err)
+{
+ *num_of_items = 0;
+
+ return 0;
+}
+
+static int
+cnxk_flow_tunnel_item_release(__rte_unused struct rte_eth_dev *dev,
+ __rte_unused struct rte_flow_item *pmd_items,
+ __rte_unused uint32_t num_of_items,
+ __rte_unused struct rte_flow_error *err)
+{
+ return 0;
+}
+
struct rte_flow_ops cnxk_flow_ops = {
.validate = cnxk_flow_validate,
.flush = cnxk_flow_flush,
@@ -653,4 +821,8 @@ struct rte_flow_ops cnxk_flow_ops = {
.isolate = cnxk_flow_isolate,
.dev_dump = cnxk_flow_dev_dump,
.get_aged_flows = cnxk_flow_get_aged_flows,
+ .tunnel_match = cnxk_flow_tunnel_match,
+ .tunnel_item_release = cnxk_flow_tunnel_item_release,
+ .tunnel_decap_set = cnxk_flow_tunnel_decap_set,
+ .tunnel_action_decap_release = cnxk_flow_tunnel_action_decap_release,
};
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index 266dd4a688..9ac675426e 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -1,6 +1,9 @@
/* SPDX-License-Identifier: BSD-3-Clause
* Copyright(C) 2023 Marvell.
*/
+
+#include <regex.h>
+
#include <cnxk_eswitch.h>
#include <cnxk_ethdev.h>
@@ -90,6 +93,17 @@ cnxk_rep_pool_buffer_stats(struct rte_mempool *pool)
pool->size, rte_mempool_in_use_count(pool), rte_mempool_avail_count(pool));
}
+static inline int
+cnxk_ethdev_is_representor(const char *if_name)
+{
+ regex_t regex;
+ int val;
+
+ val = regcomp(®ex, "net_.*_representor_.*", 0);
+ val = regexec(®ex, if_name, 0, NULL, 0);
+ return (val == 0);
+}
+
/* Prototypes */
int cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
int cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev);
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 21/24] net/cnxk: generalize flow operation APIs
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (19 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 20/24] net/cnxk: add port representor " Harman Kalra
@ 2023-12-19 17:40 ` Harman Kalra
2023-12-19 17:40 ` [PATCH v2 22/24] net/cnxk: flow create on representor ports Harman Kalra
` (2 subsequent siblings)
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:40 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Flow operations can be performed on cnxk ports as well as representor
ports. Since representor ports are not cnxk ports but have eswitch as
base device underneath, special handling is required to align with base
infra. Introducing a flag to generic flow APIs to discriminate if the
operation request made on normal or representor ports.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_flow.c | 240 +++++++++++++++++++++++++++--------
drivers/net/cnxk/cnxk_flow.h | 19 +++
2 files changed, 205 insertions(+), 54 deletions(-)
diff --git a/drivers/net/cnxk/cnxk_flow.c b/drivers/net/cnxk/cnxk_flow.c
index 959d773513..7959f2ed6b 100644
--- a/drivers/net/cnxk/cnxk_flow.c
+++ b/drivers/net/cnxk/cnxk_flow.c
@@ -223,7 +223,7 @@ static int
cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
const struct rte_flow_action actions[], struct roc_npc_action in_actions[],
struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
- uint16_t *dst_pf_func, uint8_t has_tunnel_pattern)
+ uint16_t *dst_pf_func, uint8_t has_tunnel_pattern, bool is_rep)
{
struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
const struct rte_flow_action_queue *act_q = NULL;
@@ -273,6 +273,9 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
case RTE_FLOW_ACTION_TYPE_REPRESENTED_PORT:
case RTE_FLOW_ACTION_TYPE_PORT_ID:
+ /* No port ID action on representor ethdevs */
+ if (is_rep)
+ continue;
in_actions[i].type = ROC_NPC_ACTION_TYPE_PORT_ID;
in_actions[i].conf = actions->conf;
act_ethdev = (const struct rte_flow_action_ethdev *)
@@ -320,6 +323,9 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
break;
case RTE_FLOW_ACTION_TYPE_RSS:
+ /* No RSS action on representor ethdevs */
+ if (is_rep)
+ continue;
rc = npc_rss_action_validate(eth_dev, attr, actions);
if (rc)
goto err_exit;
@@ -396,22 +402,37 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
static int
cnxk_map_pattern(struct rte_eth_dev *eth_dev, const struct rte_flow_item pattern[],
- struct roc_npc_item_info in_pattern[], uint8_t *has_tunnel_pattern)
+ struct roc_npc_item_info in_pattern[], uint8_t *has_tunnel_pattern, bool is_rep)
{
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
const struct rte_flow_item_ethdev *rep_eth_dev;
struct rte_eth_dev *portid_eth_dev;
char if_name[RTE_ETH_NAME_MAX_LEN];
struct cnxk_eth_dev *hw_dst;
+ struct cnxk_rep_dev *rdev;
+ struct cnxk_eth_dev *dev;
+ struct roc_npc *npc;
int i = 0;
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rdev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rdev->parent_dev->npc;
+
+ npc->rep_npc = npc;
+ npc->rep_port_id = rdev->port_id;
+ npc->rep_pf_func = rdev->hw_func;
+ }
+
while (pattern->type != RTE_FLOW_ITEM_TYPE_END) {
in_pattern[i].spec = pattern->spec;
in_pattern[i].last = pattern->last;
in_pattern[i].mask = pattern->mask;
in_pattern[i].type = term[pattern->type].item_type;
in_pattern[i].size = term[pattern->type].item_size;
- if (pattern->type == RTE_FLOW_ITEM_TYPE_REPRESENTED_PORT) {
+ if (pattern->type == RTE_FLOW_ITEM_TYPE_REPRESENTED_PORT ||
+ pattern->type == RTE_FLOW_ITEM_TYPE_PORT_REPRESENTOR) {
rep_eth_dev = (const struct rte_flow_item_ethdev *)pattern->spec;
if (rte_eth_dev_get_name_by_port(rep_eth_dev->port_id, if_name)) {
plt_err("Name not found for output port id");
@@ -422,11 +443,6 @@ cnxk_map_pattern(struct rte_eth_dev *eth_dev, const struct rte_flow_item pattern
plt_err("eth_dev not found for output port id");
goto fail;
}
- if (strcmp(portid_eth_dev->device->driver->name,
- eth_dev->device->driver->name) != 0) {
- plt_err("Output port not under same driver");
- goto fail;
- }
if (cnxk_ethdev_is_representor(if_name)) {
/* Case where represented port not part of same
* app and represented by a representor port.
@@ -436,20 +452,25 @@ cnxk_map_pattern(struct rte_eth_dev *eth_dev, const struct rte_flow_item pattern
rep_dev = cnxk_rep_pmd_priv(portid_eth_dev);
eswitch_dev = rep_dev->parent_dev;
- dev->npc.rep_npc = &eswitch_dev->npc;
- dev->npc.rep_port_id = rep_eth_dev->port_id;
- dev->npc.rep_pf_func = rep_dev->hw_func;
+ npc->rep_npc = &eswitch_dev->npc;
+ npc->rep_port_id = rep_eth_dev->port_id;
+ npc->rep_pf_func = rep_dev->hw_func;
plt_rep_dbg("Represented port %d act port %d rep_dev->hw_func 0x%x",
rep_eth_dev->port_id, eth_dev->data->port_id,
rep_dev->hw_func);
} else {
+ if (strcmp(portid_eth_dev->device->driver->name,
+ eth_dev->device->driver->name) != 0) {
+ plt_err("Output port not under same driver");
+ goto fail;
+ }
/* Case where represented port part of same app
* as PF.
*/
hw_dst = portid_eth_dev->data->dev_private;
- dev->npc.rep_npc = &hw_dst->npc;
- dev->npc.rep_port_id = rep_eth_dev->port_id;
- dev->npc.rep_pf_func = hw_dst->npc.pf_func;
+ npc->rep_npc = &hw_dst->npc;
+ npc->rep_port_id = rep_eth_dev->port_id;
+ npc->rep_pf_func = hw_dst->npc.pf_func;
}
}
@@ -473,7 +494,7 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
struct roc_npc_attr *in_attr, struct roc_npc_item_info in_pattern[],
struct roc_npc_action in_actions[],
struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
- uint16_t *dst_pf_func)
+ uint16_t *dst_pf_func, bool is_rep)
{
uint8_t has_tunnel_pattern = 0;
int rc;
@@ -481,44 +502,61 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
in_attr->priority = attr->priority;
in_attr->ingress = attr->ingress;
in_attr->egress = attr->egress;
+ if (attr->transfer) {
+ /* For representor ethdevs transfer attribute corresponds to egress rule */
+ if (is_rep)
+ in_attr->egress = attr->transfer;
+ else
+ in_attr->ingress = attr->transfer;
+ }
- rc = cnxk_map_pattern(eth_dev, pattern, in_pattern, &has_tunnel_pattern);
+ rc = cnxk_map_pattern(eth_dev, pattern, in_pattern, &has_tunnel_pattern, is_rep);
if (rc) {
plt_err("Failed to map pattern list");
return rc;
}
return cnxk_map_actions(eth_dev, attr, actions, in_actions, in_sample_actions, flowkey_cfg,
- dst_pf_func, has_tunnel_pattern);
+ dst_pf_func, has_tunnel_pattern, is_rep);
}
-static int
-cnxk_flow_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
- const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
- struct rte_flow_error *error)
+int
+cnxk_flow_validate_internal(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[],
+ const struct rte_flow_action actions[], struct rte_flow_error *error,
+ bool is_rep)
{
struct roc_npc_item_info in_pattern[ROC_NPC_ITEM_TYPE_END + 1];
struct roc_npc_action in_actions[ROC_NPC_MAX_ACTION_COUNT];
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
struct roc_npc_action_sample in_sample_action;
- struct roc_npc *npc = &dev->npc;
+ struct cnxk_rep_dev *rep_dev;
struct roc_npc_attr in_attr;
+ struct cnxk_eth_dev *dev;
struct roc_npc_flow flow;
uint32_t flowkey_cfg = 0;
uint16_t dst_pf_func = 0;
+ struct roc_npc *npc;
int rc;
- /* Skip flow validation for MACsec. */
- if (actions[0].type == RTE_FLOW_ACTION_TYPE_SECURITY &&
- cnxk_eth_macsec_sess_get_by_sess(dev, actions[0].conf) != NULL)
- return 0;
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ /* Skip flow validation for MACsec. */
+ if (actions[0].type == RTE_FLOW_ACTION_TYPE_SECURITY &&
+ cnxk_eth_macsec_sess_get_by_sess(dev, actions[0].conf) != NULL)
+ return 0;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
memset(&flow, 0, sizeof(flow));
memset(&in_sample_action, 0, sizeof(in_sample_action));
flow.is_validate = true;
rc = cnxk_map_flow_data(eth_dev, attr, pattern, actions, &in_attr, in_pattern, in_actions,
- &in_sample_action, &flowkey_cfg, &dst_pf_func);
+ &in_sample_action, &flowkey_cfg, &dst_pf_func, is_rep);
if (rc) {
rte_flow_error_set(error, 0, RTE_FLOW_ERROR_TYPE_ACTION_NUM, NULL,
"Failed to map flow data");
@@ -535,27 +573,45 @@ cnxk_flow_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
return 0;
}
+static int
+cnxk_flow_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ return cnxk_flow_validate_internal(eth_dev, attr, pattern, actions, error, false);
+}
+
struct roc_npc_flow *
-cnxk_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
- const struct rte_flow_item pattern[],
- const struct rte_flow_action actions[],
- struct rte_flow_error *error)
+cnxk_flow_create_internal(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[],
+ const struct rte_flow_action actions[], struct rte_flow_error *error,
+ bool is_rep)
{
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
struct roc_npc_item_info in_pattern[ROC_NPC_ITEM_TYPE_END + 1];
struct roc_npc_action in_actions[ROC_NPC_MAX_ACTION_COUNT];
struct roc_npc_action_sample in_sample_action;
- struct roc_npc *npc = &dev->npc;
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct cnxk_eth_dev *dev = NULL;
struct roc_npc_attr in_attr;
struct roc_npc_flow *flow;
uint16_t dst_pf_func = 0;
+ struct roc_npc *npc;
int errcode = 0;
int rc;
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
+
memset(&in_sample_action, 0, sizeof(in_sample_action));
memset(&in_attr, 0, sizeof(struct roc_npc_attr));
rc = cnxk_map_flow_data(eth_dev, attr, pattern, actions, &in_attr, in_pattern, in_actions,
- &in_sample_action, &npc->flowkey_cfg_state, &dst_pf_func);
+ &in_sample_action, &npc->flowkey_cfg_state, &dst_pf_func, is_rep);
if (rc) {
rte_flow_error_set(error, 0, RTE_FLOW_ERROR_TYPE_ACTION_NUM, NULL,
"Failed to map flow data");
@@ -571,14 +627,32 @@ cnxk_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
return flow;
}
+struct roc_npc_flow *
+cnxk_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ return cnxk_flow_create_internal(eth_dev, attr, pattern, actions, error, false);
+}
+
int
-cnxk_flow_destroy(struct rte_eth_dev *eth_dev, struct roc_npc_flow *flow,
- struct rte_flow_error *error)
+cnxk_flow_destroy_internal(struct rte_eth_dev *eth_dev, struct roc_npc_flow *flow,
+ struct rte_flow_error *error, bool is_rep)
{
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
- struct roc_npc *npc = &dev->npc;
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eth_dev *dev;
+ struct roc_npc *npc;
int rc;
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
+
rc = roc_npc_flow_destroy(npc, flow);
if (rc)
rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
@@ -586,13 +660,30 @@ cnxk_flow_destroy(struct rte_eth_dev *eth_dev, struct roc_npc_flow *flow,
return rc;
}
-static int
-cnxk_flow_flush(struct rte_eth_dev *eth_dev, struct rte_flow_error *error)
+int
+cnxk_flow_destroy(struct rte_eth_dev *eth_dev, struct roc_npc_flow *flow,
+ struct rte_flow_error *error)
{
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
- struct roc_npc *npc = &dev->npc;
+ return cnxk_flow_destroy_internal(eth_dev, flow, error, false);
+}
+
+int
+cnxk_flow_flush_internal(struct rte_eth_dev *eth_dev, struct rte_flow_error *error, bool is_rep)
+{
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eth_dev *dev;
+ struct roc_npc *npc;
int rc;
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
+
rc = roc_npc_mcam_free_all_resources(npc);
if (rc) {
rte_flow_error_set(error, EIO, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
@@ -604,14 +695,21 @@ cnxk_flow_flush(struct rte_eth_dev *eth_dev, struct rte_flow_error *error)
}
static int
-cnxk_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
- const struct rte_flow_action *action, void *data,
- struct rte_flow_error *error)
+cnxk_flow_flush(struct rte_eth_dev *eth_dev, struct rte_flow_error *error)
+{
+ return cnxk_flow_flush_internal(eth_dev, error, false);
+}
+
+int
+cnxk_flow_query_internal(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data,
+ struct rte_flow_error *error, bool is_rep)
{
struct roc_npc_flow *in_flow = (struct roc_npc_flow *)flow;
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
- struct roc_npc *npc = &dev->npc;
struct rte_flow_query_count *query = data;
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eth_dev *dev;
+ struct roc_npc *npc;
const char *errmsg = NULL;
int errcode = ENOTSUP;
int rc;
@@ -626,6 +724,15 @@ cnxk_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
goto err_exit;
}
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
+
if (in_flow->use_pre_alloc)
rc = roc_npc_inl_mcam_read_counter(in_flow->ctr_id, &query->hits);
else
@@ -658,6 +765,14 @@ cnxk_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
return -rte_errno;
}
+static int
+cnxk_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data,
+ struct rte_flow_error *error)
+{
+ return cnxk_flow_query_internal(eth_dev, flow, action, data, error, false);
+}
+
static int
cnxk_flow_isolate(struct rte_eth_dev *eth_dev __rte_unused,
int enable __rte_unused, struct rte_flow_error *error)
@@ -672,12 +787,22 @@ cnxk_flow_isolate(struct rte_eth_dev *eth_dev __rte_unused,
return -rte_errno;
}
-static int
-cnxk_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
- FILE *file, struct rte_flow_error *error)
+int
+cnxk_flow_dev_dump_internal(struct rte_eth_dev *eth_dev, struct rte_flow *flow, FILE *file,
+ struct rte_flow_error *error, bool is_rep)
{
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
- struct roc_npc *npc = &dev->npc;
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eth_dev *dev;
+ struct roc_npc *npc;
+
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
if (file == NULL) {
rte_flow_error_set(error, EINVAL,
@@ -699,6 +824,13 @@ cnxk_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
return 0;
}
+static int
+cnxk_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ FILE *file, struct rte_flow_error *error)
+{
+ return cnxk_flow_dev_dump_internal(eth_dev, flow, file, error, false);
+}
+
static int
cnxk_flow_get_aged_flows(struct rte_eth_dev *eth_dev, void **context,
uint32_t nb_contexts, struct rte_flow_error *err)
diff --git a/drivers/net/cnxk/cnxk_flow.h b/drivers/net/cnxk/cnxk_flow.h
index bb23629819..84333e7f9d 100644
--- a/drivers/net/cnxk/cnxk_flow.h
+++ b/drivers/net/cnxk/cnxk_flow.h
@@ -24,4 +24,23 @@ struct roc_npc_flow *cnxk_flow_create(struct rte_eth_dev *dev,
int cnxk_flow_destroy(struct rte_eth_dev *dev, struct roc_npc_flow *flow,
struct rte_flow_error *error);
+struct roc_npc_flow *cnxk_flow_create_internal(struct rte_eth_dev *eth_dev,
+ const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[],
+ const struct rte_flow_action actions[],
+ struct rte_flow_error *error, bool is_rep);
+int cnxk_flow_validate_internal(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[],
+ const struct rte_flow_action actions[], struct rte_flow_error *error,
+ bool is_rep);
+int cnxk_flow_destroy_internal(struct rte_eth_dev *eth_dev, struct roc_npc_flow *flow,
+ struct rte_flow_error *error, bool is_rep);
+int cnxk_flow_flush_internal(struct rte_eth_dev *eth_dev, struct rte_flow_error *error,
+ bool is_rep);
+int cnxk_flow_query_internal(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data,
+ struct rte_flow_error *error, bool is_rep);
+int cnxk_flow_dev_dump_internal(struct rte_eth_dev *eth_dev, struct rte_flow *flow, FILE *file,
+ struct rte_flow_error *error, bool is_rep);
+
#endif /* __CNXK_RTE_FLOW_H__ */
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 22/24] net/cnxk: flow create on representor ports
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (20 preceding siblings ...)
2023-12-19 17:40 ` [PATCH v2 21/24] net/cnxk: generalize flow operation APIs Harman Kalra
@ 2023-12-19 17:40 ` Harman Kalra
2023-12-19 17:40 ` [PATCH v2 23/24] net/cnxk: other flow operations Harman Kalra
2023-12-19 17:40 ` [PATCH v2 24/24] doc: port representors in cnxk Harman Kalra
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:40 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
- Implementing base infra for handling flow operations performed on
representor ports, where these representor ports may be representing
native representees or part of companian apps.
- Handling flow create operation
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_flow.h | 9 +-
drivers/net/cnxk/cnxk_rep.h | 3 +
drivers/net/cnxk/cnxk_rep_flow.c | 399 +++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 27 +++
drivers/net/cnxk/cnxk_rep_ops.c | 3 +-
drivers/net/cnxk/meson.build | 1 +
6 files changed, 439 insertions(+), 3 deletions(-)
create mode 100644 drivers/net/cnxk/cnxk_rep_flow.c
diff --git a/drivers/net/cnxk/cnxk_flow.h b/drivers/net/cnxk/cnxk_flow.h
index 84333e7f9d..26384400c1 100644
--- a/drivers/net/cnxk/cnxk_flow.h
+++ b/drivers/net/cnxk/cnxk_flow.h
@@ -16,8 +16,13 @@ struct cnxk_rte_flow_term_info {
uint16_t item_size;
};
-struct roc_npc_flow *cnxk_flow_create(struct rte_eth_dev *dev,
- const struct rte_flow_attr *attr,
+struct cnxk_rte_flow_action_info {
+ uint16_t conf_size;
+};
+
+extern const struct cnxk_rte_flow_term_info term[];
+
+struct roc_npc_flow *cnxk_flow_create(struct rte_eth_dev *dev, const struct rte_flow_attr *attr,
const struct rte_flow_item pattern[],
const struct rte_flow_action actions[],
struct rte_flow_error *error);
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index 9ac675426e..2b850e7e59 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -20,6 +20,9 @@
/* Common ethdev ops */
extern struct eth_dev_ops cnxk_rep_dev_ops;
+/* Flow ops for representor ports */
+extern struct rte_flow_ops cnxk_rep_flow_ops;
+
struct cnxk_rep_queue_stats {
uint64_t pkts;
uint64_t bytes;
diff --git a/drivers/net/cnxk/cnxk_rep_flow.c b/drivers/net/cnxk/cnxk_rep_flow.c
new file mode 100644
index 0000000000..ab9ced6ece
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_flow.c
@@ -0,0 +1,399 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2023 Marvell.
+ */
+
+#include <fcntl.h>
+#include <unistd.h>
+
+#include <cnxk_flow.h>
+#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
+
+#define DEFAULT_DUMP_FILE_NAME "/tmp/fdump"
+#define MAX_BUFFER_SIZE 1500
+
+const struct cnxk_rte_flow_action_info action_info[] = {
+ [RTE_FLOW_ACTION_TYPE_MARK] = {sizeof(struct rte_flow_action_mark)},
+ [RTE_FLOW_ACTION_TYPE_VF] = {sizeof(struct rte_flow_action_vf)},
+ [RTE_FLOW_ACTION_TYPE_REPRESENTED_PORT] = {sizeof(struct rte_flow_action_port_id)},
+ [RTE_FLOW_ACTION_TYPE_PORT_ID] = {sizeof(struct rte_flow_action_port_id)},
+ [RTE_FLOW_ACTION_TYPE_QUEUE] = {sizeof(struct rte_flow_action_queue)},
+ [RTE_FLOW_ACTION_TYPE_RSS] = {sizeof(struct rte_flow_action_rss)},
+ [RTE_FLOW_ACTION_TYPE_SECURITY] = {sizeof(struct rte_flow_action_security)},
+ [RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID] = {sizeof(struct rte_flow_action_of_set_vlan_vid)},
+ [RTE_FLOW_ACTION_TYPE_OF_PUSH_VLAN] = {sizeof(struct rte_flow_action_of_push_vlan)},
+ [RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_PCP] = {sizeof(struct rte_flow_action_of_set_vlan_pcp)},
+ [RTE_FLOW_ACTION_TYPE_METER] = {sizeof(struct rte_flow_action_meter)},
+ [RTE_FLOW_ACTION_TYPE_OF_POP_MPLS] = {sizeof(struct rte_flow_action_of_pop_mpls)},
+ [RTE_FLOW_ACTION_TYPE_OF_PUSH_MPLS] = {sizeof(struct rte_flow_action_of_push_mpls)},
+ [RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP] = {sizeof(struct rte_flow_action_vxlan_encap)},
+ [RTE_FLOW_ACTION_TYPE_NVGRE_ENCAP] = {sizeof(struct rte_flow_action_nvgre_encap)},
+ [RTE_FLOW_ACTION_TYPE_RAW_ENCAP] = {sizeof(struct rte_flow_action_raw_encap)},
+ [RTE_FLOW_ACTION_TYPE_RAW_DECAP] = {sizeof(struct rte_flow_action_raw_decap)},
+ [RTE_FLOW_ACTION_TYPE_COUNT] = {sizeof(struct rte_flow_action_count)},
+};
+
+static void
+cnxk_flow_params_count(const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ uint16_t *n_pattern, uint16_t *n_action)
+{
+ int i = 0;
+
+ for (; pattern->type != RTE_FLOW_ITEM_TYPE_END; pattern++)
+ i++;
+
+ *n_pattern = ++i;
+ plt_rep_dbg("Total patterns is %d", *n_pattern);
+
+ i = 0;
+ for (; actions->type != RTE_FLOW_ACTION_TYPE_END; actions++)
+ i++;
+ *n_action = ++i;
+ plt_rep_dbg("Total actions is %d", *n_action);
+}
+
+static void
+populate_attr_data(void *buffer, uint32_t *length, const struct rte_flow_attr *attr)
+{
+ uint32_t sz = sizeof(struct rte_flow_attr);
+ uint32_t len;
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_ATTR, sz);
+
+ len = *length;
+ /* Populate the attribute data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), attr, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static uint16_t
+prepare_pattern_data(const struct rte_flow_item *pattern, uint16_t nb_pattern,
+ uint64_t *pattern_data)
+{
+ cnxk_pattern_hdr_t hdr;
+ uint16_t len = 0;
+ int i = 0;
+
+ for (i = 0; i < nb_pattern; i++) {
+ /* Populate the pattern type hdr */
+ memset(&hdr, 0, sizeof(cnxk_pattern_hdr_t));
+ hdr.type = pattern->type;
+ if (pattern->spec) {
+ hdr.spec_sz = term[pattern->type].item_size;
+ hdr.last_sz = 0;
+ hdr.mask_sz = term[pattern->type].item_size;
+ }
+
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), &hdr, sizeof(cnxk_pattern_hdr_t));
+ len += sizeof(cnxk_pattern_hdr_t);
+
+ /* Copy pattern spec data */
+ if (pattern->spec) {
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), pattern->spec,
+ term[pattern->type].item_size);
+ len += term[pattern->type].item_size;
+ }
+
+ /* Copy pattern last data */
+ if (pattern->last) {
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), pattern->last,
+ term[pattern->type].item_size);
+ len += term[pattern->type].item_size;
+ }
+
+ /* Copy pattern mask data */
+ if (pattern->mask) {
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), pattern->mask,
+ term[pattern->type].item_size);
+ len += term[pattern->type].item_size;
+ }
+ pattern++;
+ }
+
+ return len;
+}
+
+static void
+populate_pattern_data(void *buffer, uint32_t *length, const struct rte_flow_item *pattern,
+ uint16_t nb_pattern)
+{
+ uint64_t pattern_data[BUFSIZ];
+ uint32_t len;
+ uint32_t sz;
+
+ /* Prepare pattern_data */
+ sz = prepare_pattern_data(pattern, nb_pattern, pattern_data);
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_PATTERN, sz);
+
+ len = *length;
+ /* Populate the pattern data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), pattern_data, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static uint16_t
+populate_rss_action_conf(const struct rte_flow_action_rss *conf, void *rss_action_conf)
+{
+ int len, sz;
+
+ len = sizeof(struct rte_flow_action_rss) - sizeof(conf->key) - sizeof(conf->queue);
+
+ if (rss_action_conf)
+ rte_memcpy(rss_action_conf, conf, len);
+
+ if (conf->key) {
+ sz = conf->key_len;
+ if (rss_action_conf)
+ rte_memcpy(RTE_PTR_ADD(rss_action_conf, len), conf->key, sz);
+ len += sz;
+ }
+
+ if (conf->queue) {
+ sz = conf->queue_num * sizeof(conf->queue);
+ if (rss_action_conf)
+ rte_memcpy(RTE_PTR_ADD(rss_action_conf, len), conf->queue, sz);
+ len += sz;
+ }
+
+ return len;
+}
+
+static uint16_t
+populate_vxlan_encap_action_conf(const struct rte_flow_action_vxlan_encap *vxlan_conf,
+ void *vxlan_encap_action_data)
+{
+ const struct rte_flow_item *pattern;
+ uint64_t nb_patterns = 0;
+ uint16_t len, sz;
+
+ pattern = vxlan_conf->definition;
+ for (; pattern->type != RTE_FLOW_ITEM_TYPE_END; pattern++)
+ nb_patterns++;
+
+ len = sizeof(uint64_t);
+ rte_memcpy(vxlan_encap_action_data, &nb_patterns, len);
+ pattern = vxlan_conf->definition;
+ /* Prepare pattern_data */
+ sz = prepare_pattern_data(pattern, nb_patterns, RTE_PTR_ADD(vxlan_encap_action_data, len));
+
+ len += sz;
+ if (len > BUFSIZ) {
+ plt_err("Incomplete item definition loaded, len %d", len);
+ return 0;
+ }
+
+ return len;
+}
+
+static uint16_t
+prepare_action_data(const struct rte_flow_action *action, uint16_t nb_action, uint64_t *action_data)
+{
+ void *action_conf_data = NULL;
+ cnxk_action_hdr_t hdr;
+ uint16_t len = 0, sz = 0;
+ int i = 0;
+
+ for (i = 0; i < nb_action; i++) {
+ if (action->conf) {
+ switch (action->type) {
+ case RTE_FLOW_ACTION_TYPE_RSS:
+ sz = populate_rss_action_conf(action->conf, NULL);
+ action_conf_data = plt_zmalloc(sz, 0);
+ if (populate_rss_action_conf(action->conf, action_conf_data) !=
+ sz) {
+ plt_err("Populating RSS action config failed");
+ return 0;
+ }
+ break;
+ case RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP:
+ action_conf_data = plt_zmalloc(BUFSIZ, 0);
+ sz = populate_vxlan_encap_action_conf(action->conf,
+ action_conf_data);
+ if (!sz) {
+ plt_err("Populating vxlan action action config failed");
+ return 0;
+ }
+ break;
+ default:
+ sz = action_info[action->type].conf_size;
+ action_conf_data = plt_zmalloc(sz, 0);
+ rte_memcpy(action_conf_data, action->conf, sz);
+ break;
+ };
+ }
+
+ /* Populate the action type hdr */
+ memset(&hdr, 0, sizeof(cnxk_action_hdr_t));
+ hdr.type = action->type;
+ hdr.conf_sz = sz;
+
+ rte_memcpy(RTE_PTR_ADD(action_data, len), &hdr, sizeof(cnxk_action_hdr_t));
+ len += sizeof(cnxk_action_hdr_t);
+
+ /* Copy action conf data */
+ if (action_conf_data) {
+ rte_memcpy(RTE_PTR_ADD(action_data, len), action_conf_data, sz);
+ len += sz;
+ plt_free(action_conf_data);
+ action_conf_data = NULL;
+ }
+
+ action++;
+ }
+
+ return len;
+}
+
+static void
+populate_action_data(void *buffer, uint32_t *length, const struct rte_flow_action *action,
+ uint16_t nb_action)
+{
+ uint64_t action_data[BUFSIZ];
+ uint32_t len;
+ uint32_t sz;
+
+ /* Prepare action_data */
+ sz = prepare_action_data(action, nb_action, action_data);
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_ACTION, sz);
+
+ len = *length;
+ /* Populate the action data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), action_data, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static int
+process_flow_rule(struct cnxk_rep_dev *rep_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ cnxk_rep_msg_ack_data_t *adata, cnxk_rep_msg_t msg)
+{
+ cnxk_rep_msg_flow_create_meta_t msg_fc_meta;
+ uint16_t n_pattern, n_action;
+ uint32_t len = 0, rc = 0;
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ /* Get no of actions and patterns */
+ cnxk_flow_params_count(pattern, actions, &n_pattern, &n_action);
+
+ /* Adding the header */
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ /* Representor port identified as rep_xport queue */
+ msg_fc_meta.portid = rep_dev->rep_id;
+ msg_fc_meta.nb_pattern = n_pattern;
+ msg_fc_meta.nb_action = n_action;
+
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_fc_meta,
+ sizeof(cnxk_rep_msg_flow_create_meta_t), msg);
+
+ /* Populate flow create parameters data */
+ populate_attr_data(buffer, &len, attr);
+ populate_pattern_data(buffer, &len, pattern, n_pattern);
+ populate_action_data(buffer, &len, actions, n_action);
+
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static struct rte_flow *
+cnxk_rep_flow_create_native(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[],
+ const struct rte_flow_action actions[], struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ struct roc_npc_flow *flow;
+ uint16_t new_entry;
+ int rc;
+
+ flow = cnxk_flow_create_internal(eth_dev, attr, pattern, actions, error, true);
+ /* Shifting the rules with higher priority than exception path rules */
+ new_entry = (uint16_t)flow->mcam_id;
+ rc = cnxk_eswitch_flow_rule_shift(rep_dev->hw_func, &new_entry);
+ if (rc) {
+ plt_err("Failed to shift the flow rule entry, err %d", rc);
+ goto fail;
+ }
+
+ flow->mcam_id = new_entry;
+
+ return (struct rte_flow *)flow;
+fail:
+ return NULL;
+}
+
+static struct rte_flow *
+cnxk_rep_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ struct rte_flow *flow = NULL;
+ cnxk_rep_msg_ack_data_t adata;
+ int rc = 0;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_rep_flow_create_native(eth_dev, attr, pattern, actions, error);
+
+ rc = process_flow_rule(rep_dev, attr, pattern, actions, &adata, CNXK_REP_MSG_FLOW_CREATE);
+ if (!rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0) {
+ rc = (int)adata.u.sval;
+ rte_flow_error_set(error, adata.u.sval, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+ NULL, "Failed to validate flow");
+ goto fail;
+ }
+
+ flow = adata.u.data;
+ if (!flow) {
+ rte_flow_error_set(error, adata.u.sval, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+ NULL, "Failed to create flow");
+ goto fail;
+ }
+ } else {
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to create flow");
+ goto fail;
+ }
+ plt_rep_dbg("Flow %p created successfully", adata.u.data);
+
+ return flow;
+fail:
+ return NULL;
+}
+
+struct rte_flow_ops cnxk_rep_flow_ops = {
+ .create = cnxk_rep_flow_create,
+};
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
index 3236de50ad..2a7b5e3bc5 100644
--- a/drivers/net/cnxk/cnxk_rep_msg.h
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -12,6 +12,10 @@
typedef enum CNXK_TYPE {
CNXK_TYPE_HEADER = 0,
CNXK_TYPE_MSG,
+ CNXK_TYPE_ATTR,
+ CNXK_TYPE_PATTERN,
+ CNXK_TYPE_ACTION,
+ CNXK_TYPE_FLOW
} cnxk_type_t;
typedef enum CNXK_REP_MSG {
@@ -23,6 +27,8 @@ typedef enum CNXK_REP_MSG {
CNXK_REP_MSG_ETH_SET_MAC,
CNXK_REP_MSG_ETH_STATS_GET,
CNXK_REP_MSG_ETH_STATS_CLEAR,
+ /* Flow operation msgs */
+ CNXK_REP_MSG_FLOW_CREATE,
/* End of messaging sequence */
CNXK_REP_MSG_END,
} cnxk_rep_msg_t;
@@ -96,6 +102,27 @@ typedef struct cnxk_rep_msg_eth_stats_meta {
uint16_t portid;
} __rte_packed cnxk_rep_msg_eth_stats_meta_t;
+/* Flow create msg meta */
+typedef struct cnxk_rep_msg_flow_create_meta {
+ uint16_t portid;
+ uint16_t nb_pattern;
+ uint16_t nb_action;
+} __rte_packed cnxk_rep_msg_flow_create_meta_t;
+
+/* Type pattern meta */
+typedef struct cnxk_pattern_hdr {
+ uint16_t type;
+ uint16_t spec_sz;
+ uint16_t last_sz;
+ uint16_t mask_sz;
+} __rte_packed cnxk_pattern_hdr_t;
+
+/* Type action meta */
+typedef struct cnxk_action_hdr {
+ uint16_t type;
+ uint16_t conf_sz;
+} __rte_packed cnxk_action_hdr_t;
+
void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
uint32_t size);
void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
index e07c63dcb2..a461ae1dc3 100644
--- a/drivers/net/cnxk/cnxk_rep_ops.c
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -637,7 +637,8 @@ int
cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops)
{
PLT_SET_USED(ethdev);
- PLT_SET_USED(ops);
+ *ops = &cnxk_rep_flow_ops;
+
return 0;
}
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index 9ca7732713..8cc06f4967 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -39,6 +39,7 @@ sources = files(
'cnxk_rep.c',
'cnxk_rep_msg.c',
'cnxk_rep_ops.c',
+ 'cnxk_rep_flow.c',
'cnxk_stats.c',
'cnxk_tm.c',
)
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 23/24] net/cnxk: other flow operations
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (21 preceding siblings ...)
2023-12-19 17:40 ` [PATCH v2 22/24] net/cnxk: flow create on representor ports Harman Kalra
@ 2023-12-19 17:40 ` Harman Kalra
2023-12-19 17:40 ` [PATCH v2 24/24] doc: port representors in cnxk Harman Kalra
23 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:40 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, jerinj
Implementing other flow operations - validate, destroy, query,
flush, dump for representor ports
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_rep_flow.c | 414 +++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 32 +++
2 files changed, 446 insertions(+)
diff --git a/drivers/net/cnxk/cnxk_rep_flow.c b/drivers/net/cnxk/cnxk_rep_flow.c
index ab9ced6ece..2abec485bc 100644
--- a/drivers/net/cnxk/cnxk_rep_flow.c
+++ b/drivers/net/cnxk/cnxk_rep_flow.c
@@ -270,6 +270,221 @@ populate_action_data(void *buffer, uint32_t *length, const struct rte_flow_actio
*length = len;
}
+static int
+process_flow_destroy(struct cnxk_rep_dev *rep_dev, void *flow, cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_destroy_meta_t msg_fd_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_fd_meta.portid = rep_dev->rep_id;
+ msg_fd_meta.flow = (uint64_t)flow;
+ plt_rep_dbg("Flow Destroy: flow 0x%" PRIu64 ", portid %d", msg_fd_meta.flow,
+ msg_fd_meta.portid);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_fd_meta,
+ sizeof(cnxk_rep_msg_flow_destroy_meta_t),
+ CNXK_REP_MSG_FLOW_DESTROY);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+copy_flow_dump_file(FILE *target)
+{
+ FILE *source = NULL;
+ int pos;
+ char ch;
+
+ source = fopen(DEFAULT_DUMP_FILE_NAME, "r");
+ if (source == NULL) {
+ plt_err("Failed to read default dump file: %s, err %d", DEFAULT_DUMP_FILE_NAME,
+ errno);
+ return errno;
+ }
+
+ fseek(source, 0L, SEEK_END);
+ pos = ftell(source);
+ fseek(source, 0L, SEEK_SET);
+ while (pos--) {
+ ch = fgetc(source);
+ fputc(ch, target);
+ }
+
+ fclose(source);
+
+ /* Remove the default file after reading */
+ remove(DEFAULT_DUMP_FILE_NAME);
+
+ return 0;
+}
+
+static int
+process_flow_dump(struct cnxk_rep_dev *rep_dev, struct rte_flow *flow, FILE *file,
+ cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_dump_meta_t msg_fp_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_fp_meta.portid = rep_dev->rep_id;
+ msg_fp_meta.flow = (uint64_t)flow;
+ msg_fp_meta.is_stdout = (file == stdout) ? 1 : 0;
+
+ plt_rep_dbg("Flow Dump: flow 0x%" PRIu64 ", portid %d stdout %d", msg_fp_meta.flow,
+ msg_fp_meta.portid, msg_fp_meta.is_stdout);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_fp_meta,
+ sizeof(cnxk_rep_msg_flow_dump_meta_t),
+ CNXK_REP_MSG_FLOW_DUMP);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ /* Copy contents from default file to user file */
+ if (file != stdout)
+ copy_flow_dump_file(file);
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+process_flow_flush(struct cnxk_rep_dev *rep_dev, cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_flush_meta_t msg_ff_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_ff_meta.portid = rep_dev->rep_id;
+ plt_rep_dbg("Flow Flush: portid %d", msg_ff_meta.portid);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_ff_meta,
+ sizeof(cnxk_rep_msg_flow_flush_meta_t),
+ CNXK_REP_MSG_FLOW_FLUSH);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+process_flow_query(struct cnxk_rep_dev *rep_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data, cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_query_meta_t *msg_fq_meta;
+ struct rte_flow_query_count *query = data;
+ uint32_t len = 0, rc, sz, total_sz;
+ uint64_t action_data[BUFSIZ];
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ sz = prepare_action_data(action, 1, action_data);
+ total_sz = sz + sizeof(cnxk_rep_msg_flow_query_meta_t);
+
+ msg_fq_meta = plt_zmalloc(total_sz, 0);
+ if (!msg_fq_meta) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ msg_fq_meta->portid = rep_dev->rep_id;
+ msg_fq_meta->reset = query->reset;
+ ;
+ msg_fq_meta->flow = (uint64_t)flow;
+ /* Populate the action data */
+ rte_memcpy(msg_fq_meta->action_data, action_data, sz);
+ msg_fq_meta->action_data_sz = sz;
+
+ plt_rep_dbg("Flow query: flow 0x%" PRIu64 ", portid %d, action type %d total sz %d "
+ "action sz %d", msg_fq_meta->flow, msg_fq_meta->portid, action->type, total_sz,
+ sz);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, msg_fq_meta, total_sz,
+ CNXK_REP_MSG_FLOW_QUERY);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto free;
+ }
+
+ rte_free(msg_fq_meta);
+
+ return 0;
+
+free:
+ rte_free(msg_fq_meta);
+fail:
+ return rc;
+}
+
static int
process_flow_rule(struct cnxk_rep_dev *rep_dev, const struct rte_flow_attr *attr,
const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
@@ -394,6 +609,205 @@ cnxk_rep_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *at
return NULL;
}
+static int
+cnxk_rep_flow_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc = 0;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_flow_validate_internal(eth_dev, attr, pattern, actions, error, true);
+
+ rc = process_flow_rule(rep_dev, attr, pattern, actions, &adata, CNXK_REP_MSG_FLOW_VALIDATE);
+ if (!rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0) {
+ rc = (int)adata.u.sval;
+ rte_flow_error_set(error, adata.u.sval, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+ NULL, "Failed to validate flow");
+ goto fail;
+ }
+ } else {
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to validate flow");
+ }
+
+ plt_rep_dbg("Flow %p validated successfully", adata.u.data);
+
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_destroy(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_flow_destroy_internal(eth_dev, (struct roc_npc_flow *)flow, error,
+ true);
+
+ rc = process_flow_destroy(rep_dev, flow, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to destroy flow");
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data, struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (action->type != RTE_FLOW_ACTION_TYPE_COUNT) {
+ rc = -ENOTSUP;
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Only COUNT is supported in query");
+ goto fail;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_flow_query_internal(eth_dev, flow, action, data, error, true);
+
+ rc = process_flow_query(rep_dev, flow, action, data, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to query the flow");
+ goto fail;
+ }
+
+ rte_memcpy(data, adata.u.data, adata.size);
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_flush(struct rte_eth_dev *eth_dev, struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_flow_flush_internal(eth_dev, error, true);
+
+ rc = process_flow_flush(rep_dev, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to destroy flow");
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow, FILE *file,
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_flow_dev_dump_internal(eth_dev, flow, file, error, true);
+
+ rc = process_flow_dump(rep_dev, flow, file, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to destroy flow");
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_isolate(struct rte_eth_dev *eth_dev __rte_unused, int enable __rte_unused,
+ struct rte_flow_error *error)
+{
+ /* If we support, we need to un-install the default mcam
+ * entry for this port.
+ */
+
+ rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Flow isolation not supported");
+
+ return -rte_errno;
+}
+
struct rte_flow_ops cnxk_rep_flow_ops = {
+ .validate = cnxk_rep_flow_validate,
.create = cnxk_rep_flow_create,
+ .destroy = cnxk_rep_flow_destroy,
+ .query = cnxk_rep_flow_query,
+ .flush = cnxk_rep_flow_flush,
+ .isolate = cnxk_rep_flow_isolate,
+ .dev_dump = cnxk_rep_flow_dev_dump,
};
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
index 2a7b5e3bc5..837eb55ba6 100644
--- a/drivers/net/cnxk/cnxk_rep_msg.h
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -29,6 +29,11 @@ typedef enum CNXK_REP_MSG {
CNXK_REP_MSG_ETH_STATS_CLEAR,
/* Flow operation msgs */
CNXK_REP_MSG_FLOW_CREATE,
+ CNXK_REP_MSG_FLOW_DESTROY,
+ CNXK_REP_MSG_FLOW_VALIDATE,
+ CNXK_REP_MSG_FLOW_FLUSH,
+ CNXK_REP_MSG_FLOW_DUMP,
+ CNXK_REP_MSG_FLOW_QUERY,
/* End of messaging sequence */
CNXK_REP_MSG_END,
} cnxk_rep_msg_t;
@@ -109,6 +114,33 @@ typedef struct cnxk_rep_msg_flow_create_meta {
uint16_t nb_action;
} __rte_packed cnxk_rep_msg_flow_create_meta_t;
+/* Flow destroy msg meta */
+typedef struct cnxk_rep_msg_flow_destroy_meta {
+ uint64_t flow;
+ uint16_t portid;
+} __rte_packed cnxk_rep_msg_flow_destroy_meta_t;
+
+/* Flow flush msg meta */
+typedef struct cnxk_rep_msg_flow_flush_meta {
+ uint16_t portid;
+} __rte_packed cnxk_rep_msg_flow_flush_meta_t;
+
+/* Flow dump msg meta */
+typedef struct cnxk_rep_msg_flow_dump_meta {
+ uint64_t flow;
+ uint16_t portid;
+ uint8_t is_stdout;
+} __rte_packed cnxk_rep_msg_flow_dump_meta_t;
+
+/* Flow query msg meta */
+typedef struct cnxk_rep_msg_flow_query_meta {
+ uint64_t flow;
+ uint16_t portid;
+ uint8_t reset;
+ uint32_t action_data_sz;
+ uint8_t action_data[];
+} __rte_packed cnxk_rep_msg_flow_query_meta_t;
+
/* Type pattern meta */
typedef struct cnxk_pattern_hdr {
uint16_t type;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v2 24/24] doc: port representors in cnxk
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
` (22 preceding siblings ...)
2023-12-19 17:40 ` [PATCH v2 23/24] net/cnxk: other flow operations Harman Kalra
@ 2023-12-19 17:40 ` Harman Kalra
2023-12-20 9:37 ` Thomas Monjalon
23 siblings, 1 reply; 142+ messages in thread
From: Harman Kalra @ 2023-12-19 17:40 UTC (permalink / raw)
To: Thomas Monjalon, Nithin Dabilpuram, Kiran Kumar K,
Sunil Kumar Kori, Satha Rao, Harman Kalra
Cc: dev, jerinj
Updating the CNXK PMD documentation with the added support
for port representors.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
MAINTAINERS | 1 +
doc/guides/nics/cnxk.rst | 58 ++++++++++++++++++++++++++++
doc/guides/nics/features/cnxk.ini | 3 ++
doc/guides/nics/features/cnxk_vf.ini | 4 ++
4 files changed, 66 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index 0d1c8126e3..2716178e18 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -827,6 +827,7 @@ M: Nithin Dabilpuram <ndabilpuram@marvell.com>
M: Kiran Kumar K <kirankumark@marvell.com>
M: Sunil Kumar Kori <skori@marvell.com>
M: Satha Rao <skoteshwar@marvell.com>
+M: Harman Kalra <hkalra@marvell.com>
T: git://dpdk.org/next/dpdk-next-net-mrvl
F: drivers/common/cnxk/
F: drivers/net/cnxk/
diff --git a/doc/guides/nics/cnxk.rst b/doc/guides/nics/cnxk.rst
index 9ec52e380f..5fd1f6513a 100644
--- a/doc/guides/nics/cnxk.rst
+++ b/doc/guides/nics/cnxk.rst
@@ -37,6 +37,9 @@ Features of the CNXK Ethdev PMD are:
- Inline IPsec processing support
- Ingress meter support
- Queue based priority flow control support
+- Port representors
+- Represented port pattern matching and action
+- Port representor pattern matching and action
Prerequisites
-------------
@@ -613,6 +616,57 @@ Runtime Config Options for inline device
With the above configuration, driver would poll for aging flows every 50
seconds.
+Port Representors
+-----------------
+
+The CNXK driver supports port representor model by adding virtual ethernet
+ports providing a logical representation in DPDK for physical function(PF) or
+SR-IOV virtual function (VF) devices for control and monitoring.
+
+Base device or parent device underneath these representor ports is a eswitch
+device which is not a cnxk ethernet device but has NIC RX and TX capabilities.
+Each representor port is represented by a RQ and SQ pair of this eswitch
+device.
+
+Current implementation supports representors for both physical function and
+virtual function.
+
+These port representor ethdev instances can be spawned on an as needed basis
+through configuration parameters passed to the driver of the underlying
+base device using devargs ``-a <base PCI BDF>,representor=pf*vf*``
+
+.. note::
+
+ Representor ports to be created for respective representees should be
+ defined via these representor devargs.
+ Eg. To create a representor for representee PF1VF0, devargs to be passed
+ is ``-a <base PCI BDF>,representor=pf0vf0``
+
+ For PF representor
+ ``-a <base PCI BDF>,representor=pf2``
+
+ For defining range of vfs, say 5 representor ports under a PF
+ ``-a <base PCI BDF>,representor=pf0vf[0-4]``
+
+ For representing different VFs under different PFs
+ ``-a <base PCI BDF>,representor=pf0vf[1,2],representor=pf1vf[2-5]``
+
+In case of exception path (i.e. until the flow definition is offloaded to the
+hardware), packets transmitted by the VFs shall be received by these
+representor port, while packets transmitted by representor ports shall be
+received by respective VFs.
+
+On receiving the VF traffic via these representor ports, applications holding
+these representor ports can decide to offload the traffic flow into the HW.
+Henceforth the matching traffic shall be directly steered to the respective
+VFs without being received by the application.
+
+Current virtual representor port PMD supports following operations:
+
+- Get and clear VF statistics
+- Set mac address
+- Flow operations - create, validate, destroy, query, flush, dump
+
Debugging Options
-----------------
@@ -627,3 +681,7 @@ Debugging Options
+---+------------+-------------------------------------------------------+
| 2 | NPC | --log-level='pmd\.net.cnxk\.flow,8' |
+---+------------+-------------------------------------------------------+
+ | 3 | REP | --log-level='pmd\.net.cnxk\.rep,8' |
+ +---+------------+-------------------------------------------------------+
+ | 4 | ESW | --log-level='pmd\.net.cnxk\.esw,8' |
+ +---+------------+-------------------------------------------------------+
diff --git a/doc/guides/nics/features/cnxk.ini b/doc/guides/nics/features/cnxk.ini
index 94e7a6ab8d..88d5aaaa4e 100644
--- a/doc/guides/nics/features/cnxk.ini
+++ b/doc/guides/nics/features/cnxk.ini
@@ -73,6 +73,8 @@ mpls = Y
nvgre = Y
pppoes = Y
raw = Y
+represented_port = Y
+port_representor = Y
sctp = Y
tcp = Y
tx_queue = Y
@@ -96,6 +98,7 @@ pf = Y
port_id = Y
queue = Y
represented_port = Y
+port_representor = Y
rss = Y
sample = Y
security = Y
diff --git a/doc/guides/nics/features/cnxk_vf.ini b/doc/guides/nics/features/cnxk_vf.ini
index 53aa2a3d0c..7d7a1cad1b 100644
--- a/doc/guides/nics/features/cnxk_vf.ini
+++ b/doc/guides/nics/features/cnxk_vf.ini
@@ -64,6 +64,8 @@ mpls = Y
nvgre = Y
pppoes = Y
raw = Y
+represented_port = Y
+port_representor = Y
sctp = Y
tcp = Y
tx_queue = Y
@@ -85,6 +87,8 @@ of_set_vlan_pcp = Y
of_set_vlan_vid = Y
pf = Y
queue = Y
+represented port = Y
+port_representor = Y
rss = Y
security = Y
skip_cman = Y
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* Re: [PATCH v2 24/24] doc: port representors in cnxk
2023-12-19 17:40 ` [PATCH v2 24/24] doc: port representors in cnxk Harman Kalra
@ 2023-12-20 9:37 ` Thomas Monjalon
2023-12-21 13:28 ` [EXT] " Harman Kalra
0 siblings, 1 reply; 142+ messages in thread
From: Thomas Monjalon @ 2023-12-20 9:37 UTC (permalink / raw)
To: Harman Kalra
Cc: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
dev, jerinj
19/12/2023 18:40, Harman Kalra:
> +The CNXK driver supports port representor model by adding virtual ethernet
> +ports providing a logical representation in DPDK for physical function(PF) or
> +SR-IOV virtual function (VF) devices for control and monitoring.
> +
> +Base device or parent device underneath these representor ports is a eswitch
> +device which is not a cnxk ethernet device but has NIC RX and TX capabilities.
> +Each representor port is represented by a RQ and SQ pair of this eswitch
> +device.
> +
> +Current implementation supports representors for both physical function and
> +virtual function.
A doc comes with its implementation, so no need to say "current implementation".
> +
> +These port representor ethdev instances can be spawned on an as needed basis
> +through configuration parameters passed to the driver of the underlying
> +base device using devargs ``-a <base PCI BDF>,representor=pf*vf*``
> +
> +.. note::
> +
> + Representor ports to be created for respective representees should be
> + defined via these representor devargs.
> + Eg. To create a representor for representee PF1VF0, devargs to be passed
> + is ``-a <base PCI BDF>,representor=pf0vf0``
> +
> + For PF representor
> + ``-a <base PCI BDF>,representor=pf2``
> +
> + For defining range of vfs, say 5 representor ports under a PF
> + ``-a <base PCI BDF>,representor=pf0vf[0-4]``
> +
> + For representing different VFs under different PFs
> + ``-a <base PCI BDF>,representor=pf0vf[1,2],representor=pf1vf[2-5]``
It looks like something we should describe globally for ethdev,
instead of driver documentation.
> +In case of exception path (i.e. until the flow definition is offloaded to the
> +hardware), packets transmitted by the VFs shall be received by these
> +representor port, while packets transmitted by representor ports shall be
> +received by respective VFs.
Not clear. How is it related to any offload?
> +On receiving the VF traffic via these representor ports, applications holding
> +these representor ports can decide to offload the traffic flow into the HW.
> +Henceforth the matching traffic shall be directly steered to the respective
> +VFs without being received by the application.
Using "these" makes no sense here. Please prefer "the representor ports".
> +Current virtual representor port PMD supports following operations:
Again, no need of "current".
[...]
> +---+------------+-------------------------------------------------------+
> | 2 | NPC | --log-level='pmd\.net.cnxk\.flow,8' |
> +---+------------+-------------------------------------------------------+
> + | 3 | REP | --log-level='pmd\.net.cnxk\.rep,8' |
> + +---+------------+-------------------------------------------------------+
> + | 4 | ESW | --log-level='pmd\.net.cnxk\.esw,8' |
> + +---+------------+-------------------------------------------------------+
Why it is not aligned?
> --- a/doc/guides/nics/features/cnxk_vf.ini
> +++ b/doc/guides/nics/features/cnxk_vf.ini
> @@ -64,6 +64,8 @@ mpls = Y
> nvgre = Y
> pppoes = Y
> raw = Y
> +represented_port = Y
> +port_representor = Y
> sctp = Y
It should be in alphabetical order.
^ permalink raw reply [flat|nested] 142+ messages in thread
* RE: [EXT] Re: [PATCH v2 24/24] doc: port representors in cnxk
2023-12-20 9:37 ` Thomas Monjalon
@ 2023-12-21 13:28 ` Harman Kalra
2023-12-21 18:33 ` Thomas Monjalon
0 siblings, 1 reply; 142+ messages in thread
From: Harman Kalra @ 2023-12-21 13:28 UTC (permalink / raw)
To: Thomas Monjalon
Cc: Nithin Kumar Dabilpuram, Kiran Kumar Kokkilagadda,
Sunil Kumar Kori, Satha Koteswara Rao Kottidi, dev,
Jerin Jacob Kollanukkaran
Hi Thomas,
Thanks for reviewing.
Please find responses inline.
Thanks
Harman
> -----Original Message-----
> From: Thomas Monjalon <thomas@monjalon.net>
> Sent: Wednesday, December 20, 2023 3:08 PM
> To: Harman Kalra <hkalra@marvell.com>
> Cc: Nithin Kumar Dabilpuram <ndabilpuram@marvell.com>; Kiran Kumar
> Kokkilagadda <kirankumark@marvell.com>; Sunil Kumar Kori
> <skori@marvell.com>; Satha Koteswara Rao Kottidi
> <skoteshwar@marvell.com>; dev@dpdk.org; Jerin Jacob Kollanukkaran
> <jerinj@marvell.com>
> Subject: [EXT] Re: [PATCH v2 24/24] doc: port representors in cnxk
>
> External Email
>
> ----------------------------------------------------------------------
> 19/12/2023 18:40, Harman Kalra:
> > +The CNXK driver supports port representor model by adding virtual
> > +ethernet ports providing a logical representation in DPDK for
> > +physical function(PF) or SR-IOV virtual function (VF) devices for control
> and monitoring.
> > +
> > +Base device or parent device underneath these representor ports is a
> > +eswitch device which is not a cnxk ethernet device but has NIC RX and TX
> capabilities.
> > +Each representor port is represented by a RQ and SQ pair of this
> > +eswitch device.
> > +
> > +Current implementation supports representors for both physical
> > +function and virtual function.
>
> A doc comes with its implementation, so no need to say "current
> implementation".
Ack, I will fix this.
>
> > +
> > +These port representor ethdev instances can be spawned on an as
> > +needed basis through configuration parameters passed to the driver of
> > +the underlying base device using devargs ``-a <base PCI
> > +BDF>,representor=pf*vf*``
> > +
> > +.. note::
> > +
> > + Representor ports to be created for respective representees should be
> > + defined via these representor devargs.
> > + Eg. To create a representor for representee PF1VF0, devargs to be
> passed
> > + is ``-a <base PCI BDF>,representor=pf0vf0``
> > +
> > + For PF representor
> > + ``-a <base PCI BDF>,representor=pf2``
> > +
> > + For defining range of vfs, say 5 representor ports under a PF
> > + ``-a <base PCI BDF>,representor=pf0vf[0-4]``
> > +
> > + For representing different VFs under different PFs
> > + ``-a <base PCI
> > + BDF>,representor=pf0vf[1,2],representor=pf1vf[2-5]``
>
> It looks like something we should describe globally for ethdev, instead of
> driver documentation.
DPDK generic representor devarg parser (rte_eth_devargs_parse_representor_ports()) can parse first
3 cases i.e. a <base PCI BDF>,representor=pf0vf0 .... ``-a <base PCI BDF>,representor=pf0vf[0-4]``,
while 4 case was a special case which our PMD needs.
Representor devargs are processed as part of new device (eswitch) PMD only, normal CNXK
PMD won't accept representor as a devarg. Hence all devargs we define under eswitch PCI device
and all the required representors are created while probing eswitch device probing.
In the following format we are defining representors for which PFs and VFs should be created:
Eg.
-a <base PCI BDF >,representor=pf0vf[1,2],representor=pf1vf[2-5]
Here
VF representor will be created only for PF0VF1, PF2VF2, PF1VF2.....PF1VF5
Although there may be n no of PF VF combinations but user wants representors for this devices only.
Please let us know your opinion if "-a <base PCI BDF >,representor=pf0vf[1,2],representor=pf1vf[2-5]"
format handling can also be handled in common code. We can push a separate patch for it.
>
> > +In case of exception path (i.e. until the flow definition is
> > +offloaded to the hardware), packets transmitted by the VFs shall be
> > +received by these representor port, while packets transmitted by
> > +representor ports shall be received by respective VFs.
>
> Not clear. How is it related to any offload?
Point what I wanted to highlight here is, until the flow rule for a fastpath is identified
and installed (offloaded) to the HW, packet flow will take the slow path (exception path)
i.e. for every packet sent out via VF should be received by its representor port and
vice versa.
Once the application installs the rule packets can take fast path i.e. directly
from VF to destination (wire or other VF), representors will not come in the
datapath for fast processing.
>
> > +On receiving the VF traffic via these representor ports, applications
> > +holding these representor ports can decide to offload the traffic flow into
> the HW.
> > +Henceforth the matching traffic shall be directly steered to the
> > +respective VFs without being received by the application.
>
> Using "these" makes no sense here. Please prefer "the representor ports".
Ack, will fix this
>
> > +Current virtual representor port PMD supports following operations:
>
> Again, no need of "current".
Ack, will fix this
>
> [...]
> > +---+------------+-------------------------------------------------------+
> > | 2 | NPC | --log-level='pmd\.net.cnxk\.flow,8' |
> >
> > +---+------------+----------------------------------------------------
> > ---+
> > + | 3 | REP | --log-level='pmd\.net.cnxk\.rep,8' |
> > + +---+------------+-------------------------------------------------------+
> > + | 4 | ESW | --log-level='pmd\.net.cnxk\.esw,8' |
> > +
> > + +---+------------+--------------------------------------------------
> > + -----+
>
> Why it is not aligned?
Sorry, my bad I will fix this
>
> > --- a/doc/guides/nics/features/cnxk_vf.ini
> > +++ b/doc/guides/nics/features/cnxk_vf.ini
> > @@ -64,6 +64,8 @@ mpls = Y
> > nvgre = Y
> > pppoes = Y
> > raw = Y
> > +represented_port = Y
> > +port_representor = Y
> > sctp = Y
>
> It should be in alphabetical order.
Ack, will fix this
>
>
^ permalink raw reply [flat|nested] 142+ messages in thread
* Re: [EXT] Re: [PATCH v2 24/24] doc: port representors in cnxk
2023-12-21 13:28 ` [EXT] " Harman Kalra
@ 2023-12-21 18:33 ` Thomas Monjalon
2024-01-11 6:48 ` Harman Kalra
0 siblings, 1 reply; 142+ messages in thread
From: Thomas Monjalon @ 2023-12-21 18:33 UTC (permalink / raw)
To: Harman Kalra
Cc: Nithin Kumar Dabilpuram, Kiran Kumar Kokkilagadda,
Sunil Kumar Kori, Satha Koteswara Rao Kottidi, dev,
Jerin Jacob Kollanukkaran
21/12/2023 14:28, Harman Kalra:
> From: Thomas Monjalon <thomas@monjalon.net>
> > 19/12/2023 18:40, Harman Kalra:
> > > + Representor ports to be created for respective representees should be
> > > + defined via these representor devargs.
> > > + Eg. To create a representor for representee PF1VF0, devargs to be
> > passed
> > > + is ``-a <base PCI BDF>,representor=pf0vf0``
> > > +
> > > + For PF representor
> > > + ``-a <base PCI BDF>,representor=pf2``
> > > +
> > > + For defining range of vfs, say 5 representor ports under a PF
> > > + ``-a <base PCI BDF>,representor=pf0vf[0-4]``
> > > +
> > > + For representing different VFs under different PFs
> > > + ``-a <base PCI
> > > + BDF>,representor=pf0vf[1,2],representor=pf1vf[2-5]``
> >
> > It looks like something we should describe globally for ethdev, instead of
> > driver documentation.
>
> DPDK generic representor devarg parser (rte_eth_devargs_parse_representor_ports()) can parse first
> 3 cases i.e. a <base PCI BDF>,representor=pf0vf0 .... ``-a <base PCI BDF>,representor=pf0vf[0-4]``,
> while 4 case was a special case which our PMD needs.
>
> Representor devargs are processed as part of new device (eswitch) PMD only, normal CNXK
> PMD won't accept representor as a devarg. Hence all devargs we define under eswitch PCI device
> and all the required representors are created while probing eswitch device probing.
>
> In the following format we are defining representors for which PFs and VFs should be created:
> Eg.
> -a <base PCI BDF >,representor=pf0vf[1,2],representor=pf1vf[2-5]
> Here
> VF representor will be created only for PF0VF1, PF2VF2, PF1VF2.....PF1VF5
> Although there may be n no of PF VF combinations but user wants representors for this devices only.
>
> Please let us know your opinion if "-a <base PCI BDF >,representor=pf0vf[1,2],representor=pf1vf[2-5]"
> format handling can also be handled in common code. We can push a separate patch for it.
I think yes it could be moved to common code in ethdev.
> > > +In case of exception path (i.e. until the flow definition is
> > > +offloaded to the hardware), packets transmitted by the VFs shall be
> > > +received by these representor port, while packets transmitted by
> > > +representor ports shall be received by respective VFs.
> >
> > Not clear. How is it related to any offload?
>
> Point what I wanted to highlight here is, until the flow rule for a fastpath is identified
> and installed (offloaded) to the HW, packet flow will take the slow path (exception path)
> i.e. for every packet sent out via VF should be received by its representor port and
> vice versa.
That's the case for any flow rule, right?
I don't think it is specific to VF and representors.
> Once the application installs the rule packets can take fast path i.e. directly
> from VF to destination (wire or other VF), representors will not come in the
> datapath for fast processing.
You probably need to rephrase to explain what happens in VF scenario
without being something which looks like an exception.
^ permalink raw reply [flat|nested] 142+ messages in thread
* Re: [PATCH v2 02/24] net/cnxk: implementing eswitch device
2023-12-19 17:39 ` [PATCH v2 02/24] net/cnxk: implementing eswitch device Harman Kalra
@ 2024-01-04 12:30 ` Jerin Jacob
0 siblings, 0 replies; 142+ messages in thread
From: Jerin Jacob @ 2024-01-04 12:30 UTC (permalink / raw)
To: Harman Kalra
Cc: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Anatoly Burakov, dev, jerinj
On Wed, Dec 20, 2023 at 12:53 AM Harman Kalra <hkalra@marvell.com> wrote:
>
> Eswitch device is a parent or base device behind all the representors,
> acting as transport layer between representors and representees
>
> Signed-off-by: Harman Kalra <hkalra@marvell.com>
> ---
> drivers/net/cnxk/cnxk_eswitch.c | 465 ++++++++++++++++++++++++++++++++
> drivers/net/cnxk/cnxk_eswitch.h | 103 +++++++
> drivers/net/cnxk/meson.build | 1 +
> 3 files changed, 569 insertions(+)
> create mode 100644 drivers/net/cnxk/cnxk_eswitch.c
> create mode 100644 drivers/net/cnxk/cnxk_eswitch.h
>
> diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
> new file mode 100644
> index 0000000000..51110a762d
> --- /dev/null
> +++ b/drivers/net/cnxk/cnxk_eswitch.c
> @@ -0,0 +1,465 @@
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(C) 2023 Marvell.
Change to 2024 for new files in this series.
> +static int
> +eswitch_dev_nix_flow_ctrl_set(struct cnxk_eswitch_dev *eswitch_dev)
> +{
> +
> + rc = roc_nix_fc_mode_set(nix, mode_map[ROC_NIX_FC_FULL]);
> + if (rc)
> + return rc;
> +
> + return rc;
same as return roc_nix_fc_mode_set(nix, mode_map[ROC_NIX_FC_FULL]);
^ permalink raw reply [flat|nested] 142+ messages in thread
* Re: [PATCH v2 03/24] net/cnxk: eswitch HW resource configuration
2023-12-19 17:39 ` [PATCH v2 03/24] net/cnxk: eswitch HW resource configuration Harman Kalra
@ 2024-01-04 12:34 ` Jerin Jacob
0 siblings, 0 replies; 142+ messages in thread
From: Jerin Jacob @ 2024-01-04 12:34 UTC (permalink / raw)
To: Harman Kalra
Cc: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
dev, jerinj
On Wed, Dec 20, 2023 at 12:58 AM Harman Kalra <hkalra@marvell.com> wrote:
>
> Configuring the hardware resources used by the eswitch device.
>
> Signed-off-by: Harman Kalra <hkalra@marvell.com>
> ---
> drivers/net/cnxk/cnxk_eswitch.c | 206 ++++++++++++++++++++++++++++++++
> 1 file changed, 206 insertions(+)
>
> +
> static int
> cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
> {
> @@ -433,6 +630,12 @@ cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pc
> return rc;
> +free_mem:
> + if (mz)
Not needed as rte_memzone_free has the check
> + rte_memzone_free(mz);
> fail:
> return rc;
> }
> --
> 2.18.0
>
^ permalink raw reply [flat|nested] 142+ messages in thread
* Re: [PATCH v2 07/24] common/cnxk: interface to update VLAN TPID
2023-12-19 17:39 ` [PATCH v2 07/24] common/cnxk: interface to update VLAN TPID Harman Kalra
@ 2024-01-04 12:47 ` Jerin Jacob
0 siblings, 0 replies; 142+ messages in thread
From: Jerin Jacob @ 2024-01-04 12:47 UTC (permalink / raw)
To: Harman Kalra
Cc: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
dev, jerinj
On Wed, Dec 20, 2023 at 12:53 AM Harman Kalra <hkalra@marvell.com> wrote:
>
> Introducing eswitch variant of set vlan tpid api which can be
> using for PF and VF
>
> Signed-off-by: Harman Kalra <hkalra@marvell.com>
> +
> +int
> +roc_eswitch_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t tpid, bool is_vf)
> +{
> + struct nix *nix = roc_nix_to_nix_priv(roc_nix);
> + struct dev *dev = &nix->dev;
> + int rc = 0;
Across the series, Please check the need for initializing to zero for rc.
In this case, it is not needed.
> +
> + /* Configuring for PF/VF */
> + rc = nix_vlan_tpid_set(dev->mbox, dev->pf_func | is_vf, type, tpid);
> + if (rc)
> + plt_err("Failed to set tpid for PF, rc %d", rc);
> +
> + return rc;
> +}
^ permalink raw reply [flat|nested] 142+ messages in thread
* RE: [EXT] Re: [PATCH v2 24/24] doc: port representors in cnxk
2023-12-21 18:33 ` Thomas Monjalon
@ 2024-01-11 6:48 ` Harman Kalra
0 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-01-11 6:48 UTC (permalink / raw)
To: Thomas Monjalon
Cc: Nithin Kumar Dabilpuram, Kiran Kumar Kokkilagadda,
Sunil Kumar Kori, Satha Koteswara Rao Kottidi, dev,
Jerin Jacob Kollanukkaran
Hi Thomas
Thanks for the review
Please see inline
Thanks
Harman
> -----Original Message-----
> From: Thomas Monjalon <thomas@monjalon.net>
> Sent: Friday, December 22, 2023 12:04 AM
> To: Harman Kalra <hkalra@marvell.com>
> Cc: Nithin Kumar Dabilpuram <ndabilpuram@marvell.com>; Kiran Kumar
> Kokkilagadda <kirankumark@marvell.com>; Sunil Kumar Kori
> <skori@marvell.com>; Satha Koteswara Rao Kottidi
> <skoteshwar@marvell.com>; dev@dpdk.org; Jerin Jacob Kollanukkaran
> <jerinj@marvell.com>
> Subject: Re: [EXT] Re: [PATCH v2 24/24] doc: port representors in cnxk
>
<snip>
> >
> > In the following format we are defining representors for which PFs and VFs
> should be created:
> > Eg.
> > -a <base PCI BDF >,representor=pf0vf[1,2],representor=pf1vf[2-5]
> > Here
> > VF representor will be created only for PF0VF1, PF2VF2,
> > PF1VF2.....PF1VF5 Although there may be n no of PF VF combinations but
> user wants representors for this devices only.
> >
> > Please let us know your opinion if "-a <base PCI BDF
> >,representor=pf0vf[1,2],representor=pf1vf[2-5]"
> > format handling can also be handled in common code. We can push a
> separate patch for it.
>
> I think yes it could be moved to common code in ethdev.
I have pushed a series for the change:
https://patches.dpdk.org/project/dpdk/list/?series=30781
>
>
> > > > +In case of exception path (i.e. until the flow definition is
> > > > +offloaded to the hardware), packets transmitted by the VFs shall
> > > > +be received by these representor port, while packets transmitted
> > > > +by representor ports shall be received by respective VFs.
> > >
> > > Not clear. How is it related to any offload?
> >
> > Point what I wanted to highlight here is, until the flow rule for a
> > fastpath is identified and installed (offloaded) to the HW, packet
> > flow will take the slow path (exception path) i.e. for every packet
> > sent out via VF should be received by its representor port and vice versa.
>
> That's the case for any flow rule, right?
> I don't think it is specific to VF and representors.
Yes, will remove generic point
>
> > Once the application installs the rule packets can take fast path i.e.
> > directly from VF to destination (wire or other VF), representors will
> > not come in the datapath for fast processing.
>
> You probably need to rephrase to explain what happens in VF scenario
> without being something which looks like an exception.
Sure, will reword in next series.
>
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 00/23] net/cnxk: support for port representors
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
` (9 preceding siblings ...)
2023-12-19 17:39 ` [PATCH v2 00/24] net/cnxk: support for port representors Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 01/23] common/cnxk: add support for representors Harman Kalra
` (22 more replies)
2024-02-27 19:15 ` [PATCH v4 00/23] net/cnxk: support for port representors Harman Kalra
` (2 subsequent siblings)
13 siblings, 23 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
Cc: dev, Harman Kalra
Introducing port representor support to CNXK drivers by adding virtual ethernet
ports providing a logical representation in DPDK for physical function(PF) or
SR-IOV virtual function (VF) devices for control and monitoring.
These port representor ethdev instances can be spawned on an as needed basis
through configuration parameters passed to the driver of the underlying
base device using devargs ``-a <base PCI BDF>,representor=pf*vf*``
In case of exception path (i.e. until the flow definition is offloaded to the
hardware), packets transmitted by the VFs shall be received by these
representor port, while packets transmitted by representor ports shall be
received by respective VFs.
On receiving the VF traffic via these representor ports, applications holding
these representor ports can decide to offload the traffic flow into the HW.
Henceforth the matching traffic shall be directly steered to the respective
VFs without being received by the application.
Current virtual representor port PMD supports following operations:
- Get represented port statistics
- Flow operations - create, validate, destroy, query, flush, dump
Changes since V2:
* Moved devargs parsing logic to common code and sent as separate series
* Documentation updated
* Addressed comments from V2
Changes since V1:
* Updated communication layer between representor and represented port.
* Added support for native represented ports
* Port representor and represented port item and action support
* Build failure fixes
Harman Kalra (21):
common/cnxk: add support for representors
net/cnxk: implementing eswitch device
net/cnxk: eswitch HW resource configuration
net/cnxk: eswitch devargs parsing
net/cnxk: probing representor ports
common/cnxk: common NPC changes for eswitch
common/cnxk: interface to update VLAN TPID
net/cnxk: eswitch flow configurations
net/cnxk: eswitch fastpath routines
net/cnxk: add representor control plane
common/cnxk: representee notification callback
net/cnxk: handling representee notification
net/cnxk: representor ethdev ops
common/cnxk: get representees ethernet stats
net/cnxk: ethernet statistic for representor
common/cnxk: base support for eswitch VF
net/cnxk: eswitch VF as ethernet device
net/cnxk: add representor port pattern and action
net/cnxk: generalise flow operation APIs
net/cnxk: flow create on representor ports
net/cnxk: other flow operations
Kiran Kumar K (2):
common/cnxk: support port representor and represented port
net/cnxk: add represented port pattern and action
MAINTAINERS | 1 +
doc/guides/nics/cnxk.rst | 41 ++
doc/guides/nics/features/cnxk.ini | 5 +
doc/guides/nics/features/cnxk_vec.ini | 6 +
doc/guides/nics/features/cnxk_vf.ini | 6 +
drivers/common/cnxk/meson.build | 1 +
drivers/common/cnxk/roc_api.h | 3 +
drivers/common/cnxk/roc_constants.h | 2 +
drivers/common/cnxk/roc_dev.c | 25 +
drivers/common/cnxk/roc_dev_priv.h | 3 +
drivers/common/cnxk/roc_eswitch.c | 389 +++++++++++
drivers/common/cnxk/roc_eswitch.h | 34 +
drivers/common/cnxk/roc_mbox.c | 2 +
drivers/common/cnxk/roc_mbox.h | 82 ++-
drivers/common/cnxk/roc_nix.c | 46 +-
drivers/common/cnxk/roc_nix.h | 4 +
drivers/common/cnxk/roc_nix_priv.h | 12 +-
drivers/common/cnxk/roc_nix_vlan.c | 23 +-
drivers/common/cnxk/roc_npc.c | 89 ++-
drivers/common/cnxk/roc_npc.h | 18 +-
drivers/common/cnxk/roc_npc_mcam.c | 64 +-
drivers/common/cnxk/roc_npc_parse.c | 28 +-
drivers/common/cnxk/roc_npc_priv.h | 5 +-
drivers/common/cnxk/roc_platform.c | 2 +
drivers/common/cnxk/roc_platform.h | 4 +
drivers/common/cnxk/version.map | 14 +
drivers/net/cnxk/cn10k_ethdev.c | 1 +
drivers/net/cnxk/cnxk_eswitch.c | 807 +++++++++++++++++++++++
drivers/net/cnxk/cnxk_eswitch.h | 213 ++++++
drivers/net/cnxk/cnxk_eswitch_devargs.c | 125 ++++
drivers/net/cnxk/cnxk_eswitch_flow.c | 454 +++++++++++++
drivers/net/cnxk/cnxk_eswitch_rxtx.c | 211 ++++++
drivers/net/cnxk/cnxk_ethdev.c | 41 +-
drivers/net/cnxk/cnxk_ethdev.h | 3 +
drivers/net/cnxk/cnxk_ethdev_ops.c | 4 +
drivers/net/cnxk/cnxk_flow.c | 546 ++++++++++++----
drivers/net/cnxk/cnxk_flow.h | 27 +-
drivers/net/cnxk/cnxk_link.c | 3 +-
drivers/net/cnxk/cnxk_rep.c | 555 ++++++++++++++++
drivers/net/cnxk/cnxk_rep.h | 141 ++++
drivers/net/cnxk/cnxk_rep_flow.c | 815 +++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.c | 827 ++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 169 +++++
drivers/net/cnxk/cnxk_rep_ops.c | 715 ++++++++++++++++++++
drivers/net/cnxk/meson.build | 8 +
45 files changed, 6371 insertions(+), 203 deletions(-)
create mode 100644 drivers/common/cnxk/roc_eswitch.c
create mode 100644 drivers/common/cnxk/roc_eswitch.h
create mode 100644 drivers/net/cnxk/cnxk_eswitch.c
create mode 100644 drivers/net/cnxk/cnxk_eswitch.h
create mode 100644 drivers/net/cnxk/cnxk_eswitch_devargs.c
create mode 100644 drivers/net/cnxk/cnxk_eswitch_flow.c
create mode 100644 drivers/net/cnxk/cnxk_eswitch_rxtx.c
create mode 100644 drivers/net/cnxk/cnxk_rep.c
create mode 100644 drivers/net/cnxk/cnxk_rep.h
create mode 100644 drivers/net/cnxk/cnxk_rep_flow.c
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.c
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.h
create mode 100644 drivers/net/cnxk/cnxk_rep_ops.c
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 01/23] common/cnxk: add support for representors
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 02/23] net/cnxk: implementing eswitch device Harman Kalra
` (21 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Introducing a new Mailbox for registering base device behind
all representors and also registering debug log type for representors
and base device driver.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
doc/guides/nics/cnxk.rst | 4 ++++
drivers/common/cnxk/roc_constants.h | 1 +
drivers/common/cnxk/roc_mbox.h | 8 ++++++++
drivers/common/cnxk/roc_nix.c | 31 +++++++++++++++++++++++++++++
drivers/common/cnxk/roc_nix.h | 3 +++
drivers/common/cnxk/roc_platform.c | 2 ++
drivers/common/cnxk/roc_platform.h | 4 ++++
drivers/common/cnxk/version.map | 3 +++
8 files changed, 56 insertions(+)
diff --git a/doc/guides/nics/cnxk.rst b/doc/guides/nics/cnxk.rst
index 9ec52e380f..58cb8e2283 100644
--- a/doc/guides/nics/cnxk.rst
+++ b/doc/guides/nics/cnxk.rst
@@ -627,3 +627,7 @@ Debugging Options
+---+------------+-------------------------------------------------------+
| 2 | NPC | --log-level='pmd\.net.cnxk\.flow,8' |
+---+------------+-------------------------------------------------------+
+ | 3 | REP | --log-level='pmd\.net.cnxk\.rep,8' |
+ +---+------------+-------------------------------------------------------+
+ | 4 | ESW | --log-level='pmd\.net.cnxk\.esw,8' |
+ +---+------------+-------------------------------------------------------+
diff --git a/drivers/common/cnxk/roc_constants.h b/drivers/common/cnxk/roc_constants.h
index 291b6a4bc9..cb4edbea58 100644
--- a/drivers/common/cnxk/roc_constants.h
+++ b/drivers/common/cnxk/roc_constants.h
@@ -43,6 +43,7 @@
#define PCI_DEVID_CNXK_RVU_NIX_INL_VF 0xA0F1
#define PCI_DEVID_CNXK_RVU_REE_PF 0xA0f4
#define PCI_DEVID_CNXK_RVU_REE_VF 0xA0f5
+#define PCI_DEVID_CNXK_RVU_ESWITCH_PF 0xA0E0
#define PCI_DEVID_CN9K_CGX 0xA059
#define PCI_DEVID_CN10K_RPM 0xA060
diff --git a/drivers/common/cnxk/roc_mbox.h b/drivers/common/cnxk/roc_mbox.h
index 3257a370bc..b7e2f43d45 100644
--- a/drivers/common/cnxk/roc_mbox.h
+++ b/drivers/common/cnxk/roc_mbox.h
@@ -68,6 +68,7 @@ struct mbox_msghdr {
M(NDC_SYNC_OP, 0x009, ndc_sync_op, ndc_sync_op, msg_rsp) \
M(LMTST_TBL_SETUP, 0x00a, lmtst_tbl_setup, lmtst_tbl_setup_req, \
msg_rsp) \
+ M(GET_REP_CNT, 0x00d, get_rep_cnt, msg_req, get_rep_cnt_rsp) \
/* CGX mbox IDs (range 0x200 - 0x3FF) */ \
M(CGX_START_RXTX, 0x200, cgx_start_rxtx, msg_req, msg_rsp) \
M(CGX_STOP_RXTX, 0x201, cgx_stop_rxtx, msg_req, msg_rsp) \
@@ -546,6 +547,13 @@ struct lmtst_tbl_setup_req {
uint64_t __io rsvd[2]; /* Future use */
};
+#define MAX_PFVF_REP 64
+struct get_rep_cnt_rsp {
+ struct mbox_msghdr hdr;
+ uint16_t __io rep_cnt;
+ uint16_t __io rep_pfvf_map[MAX_PFVF_REP];
+};
+
/* CGX mbox message formats */
/* CGX mailbox error codes
* Range 1101 - 1200.
diff --git a/drivers/common/cnxk/roc_nix.c b/drivers/common/cnxk/roc_nix.c
index f64933a1d9..7e327a7e6e 100644
--- a/drivers/common/cnxk/roc_nix.c
+++ b/drivers/common/cnxk/roc_nix.c
@@ -531,3 +531,34 @@ roc_nix_dev_fini(struct roc_nix *roc_nix)
rc |= dev_fini(&nix->dev, nix->pci_dev);
return rc;
}
+
+int
+roc_nix_max_rep_count(struct roc_nix *roc_nix)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+ struct mbox *mbox = mbox_get(dev->mbox);
+ struct get_rep_cnt_rsp *rsp;
+ struct msg_req *req;
+ int rc, i;
+
+ req = mbox_alloc_msg_get_rep_cnt(mbox);
+ if (!req) {
+ rc = -ENOSPC;
+ goto exit;
+ }
+
+ req->hdr.pcifunc = roc_nix_get_pf_func(roc_nix);
+
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ roc_nix->rep_cnt = rsp->rep_cnt;
+ for (i = 0; i < rsp->rep_cnt; i++)
+ roc_nix->rep_pfvf_map[i] = rsp->rep_pfvf_map[i];
+
+exit:
+ mbox_put(mbox);
+ return rc;
+}
diff --git a/drivers/common/cnxk/roc_nix.h b/drivers/common/cnxk/roc_nix.h
index 84e6fc3df5..b369335fc4 100644
--- a/drivers/common/cnxk/roc_nix.h
+++ b/drivers/common/cnxk/roc_nix.h
@@ -483,6 +483,8 @@ struct roc_nix {
uint32_t buf_sz;
uint64_t meta_aura_handle;
uintptr_t meta_mempool;
+ uint16_t rep_cnt;
+ uint16_t rep_pfvf_map[MAX_PFVF_REP];
TAILQ_ENTRY(roc_nix) next;
#define ROC_NIX_MEM_SZ (6 * 1070)
@@ -1013,4 +1015,5 @@ int __roc_api roc_nix_mcast_list_setup(struct mbox *mbox, uint8_t intf, int nb_e
uint16_t *pf_funcs, uint16_t *channels, uint32_t *rqs,
uint32_t *grp_index, uint32_t *start_index);
int __roc_api roc_nix_mcast_list_free(struct mbox *mbox, uint32_t mcast_grp_index);
+int __roc_api roc_nix_max_rep_count(struct roc_nix *roc_nix);
#endif /* _ROC_NIX_H_ */
diff --git a/drivers/common/cnxk/roc_platform.c b/drivers/common/cnxk/roc_platform.c
index 15cbb6d68f..181902a585 100644
--- a/drivers/common/cnxk/roc_platform.c
+++ b/drivers/common/cnxk/roc_platform.c
@@ -96,4 +96,6 @@ RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_sso, NOTICE);
RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_tim, NOTICE);
RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_tm, NOTICE);
RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_dpi, NOTICE);
+RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_rep, NOTICE);
+RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_esw, NOTICE);
RTE_LOG_REGISTER_DEFAULT(cnxk_logtype_ree, NOTICE);
diff --git a/drivers/common/cnxk/roc_platform.h b/drivers/common/cnxk/roc_platform.h
index ba23b2e0d7..e08eb7f6ba 100644
--- a/drivers/common/cnxk/roc_platform.h
+++ b/drivers/common/cnxk/roc_platform.h
@@ -264,6 +264,8 @@ extern int cnxk_logtype_tim;
extern int cnxk_logtype_tm;
extern int cnxk_logtype_ree;
extern int cnxk_logtype_dpi;
+extern int cnxk_logtype_rep;
+extern int cnxk_logtype_esw;
#define plt_err(fmt, args...) \
RTE_LOG(ERR, PMD, "%s():%u " fmt "\n", __func__, __LINE__, ##args)
@@ -293,6 +295,8 @@ extern int cnxk_logtype_dpi;
#define plt_tm_dbg(fmt, ...) plt_dbg(tm, fmt, ##__VA_ARGS__)
#define plt_ree_dbg(fmt, ...) plt_dbg(ree, fmt, ##__VA_ARGS__)
#define plt_dpi_dbg(fmt, ...) plt_dbg(dpi, fmt, ##__VA_ARGS__)
+#define plt_rep_dbg(fmt, ...) plt_dbg(rep, fmt, ##__VA_ARGS__)
+#define plt_esw_dbg(fmt, ...) plt_dbg(esw, fmt, ##__VA_ARGS__)
/* Datapath logs */
#define plt_dp_err(fmt, args...) \
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index 7b6afa63a9..bd28803013 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -8,12 +8,14 @@ INTERNAL {
cnxk_logtype_base;
cnxk_logtype_cpt;
cnxk_logtype_dpi;
+ cnxk_logtype_esw;
cnxk_logtype_mbox;
cnxk_logtype_ml;
cnxk_logtype_nix;
cnxk_logtype_npa;
cnxk_logtype_npc;
cnxk_logtype_ree;
+ cnxk_logtype_rep;
cnxk_logtype_sso;
cnxk_logtype_tim;
cnxk_logtype_tm;
@@ -216,6 +218,7 @@ INTERNAL {
roc_nix_get_base_chan;
roc_nix_get_pf;
roc_nix_get_pf_func;
+ roc_nix_max_rep_count;
roc_nix_get_rx_chan_cnt;
roc_nix_get_vf;
roc_nix_get_vwqe_interval;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 02/23] net/cnxk: implementing eswitch device
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
2024-02-01 13:07 ` [PATCH v3 01/23] common/cnxk: add support for representors Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 03/23] net/cnxk: eswitch HW resource configuration Harman Kalra
` (20 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra, Anatoly Burakov
Cc: dev
Eswitch device is a parent or base device behind all the representors,
acting as transport layer between representors and representees
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 378 ++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_eswitch.h | 103 +++++++++
drivers/net/cnxk/meson.build | 1 +
3 files changed, 482 insertions(+)
create mode 100644 drivers/net/cnxk/cnxk_eswitch.c
create mode 100644 drivers/net/cnxk/cnxk_eswitch.h
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
new file mode 100644
index 0000000000..c4ea3063ae
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -0,0 +1,378 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#include <cnxk_eswitch.h>
+
+#define CNXK_NIX_DEF_SQ_COUNT 512
+
+static int
+cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
+{
+ struct cnxk_eswitch_dev *eswitch_dev;
+ int rc = 0;
+
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return 0;
+
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ if (!eswitch_dev) {
+ rc = -EINVAL;
+ goto exit;
+ }
+
+ rte_free(eswitch_dev);
+exit:
+ return rc;
+}
+
+int
+cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ int rc;
+
+ /* Enable Rx in NPC */
+ rc = roc_nix_npc_rx_ena_dis(&eswitch_dev->nix, true);
+ if (rc) {
+ plt_err("Failed to enable NPC rx %d", rc);
+ goto done;
+ }
+
+ rc = roc_npc_mcam_enable_all_entries(&eswitch_dev->npc, 1);
+ if (rc) {
+ plt_err("Failed to enable NPC entries %d", rc);
+ goto done;
+ }
+
+done:
+ return 0;
+}
+
+int
+cnxk_eswitch_txq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_sq *sq = &eswitch_dev->txq[qid].sqs;
+ int rc = -EINVAL;
+
+ if (eswitch_dev->txq[qid].state == CNXK_ESWITCH_QUEUE_STATE_STARTED)
+ return 0;
+
+ if (eswitch_dev->txq[qid].state != CNXK_ESWITCH_QUEUE_STATE_CONFIGURED) {
+ plt_err("Eswitch txq %d not configured yet", qid);
+ goto done;
+ }
+
+ rc = roc_nix_sq_ena_dis(sq, true);
+ if (rc) {
+ plt_err("Failed to enable sq aura fc, txq=%u, rc=%d", qid, rc);
+ goto done;
+ }
+
+ eswitch_dev->txq[qid].state = CNXK_ESWITCH_QUEUE_STATE_STARTED;
+done:
+ return rc;
+}
+
+int
+cnxk_eswitch_txq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_sq *sq = &eswitch_dev->txq[qid].sqs;
+ int rc = -EINVAL;
+
+ if (eswitch_dev->txq[qid].state == CNXK_ESWITCH_QUEUE_STATE_STOPPED ||
+ eswitch_dev->txq[qid].state == CNXK_ESWITCH_QUEUE_STATE_RELEASED)
+ return 0;
+
+ if (eswitch_dev->txq[qid].state != CNXK_ESWITCH_QUEUE_STATE_STARTED) {
+ plt_err("Eswitch txq %d not started", qid);
+ goto done;
+ }
+
+ rc = roc_nix_sq_ena_dis(sq, false);
+ if (rc) {
+ plt_err("Failed to disable sqb aura fc, txq=%u, rc=%d", qid, rc);
+ goto done;
+ }
+
+ eswitch_dev->txq[qid].state = CNXK_ESWITCH_QUEUE_STATE_STOPPED;
+done:
+ return rc;
+}
+
+int
+cnxk_eswitch_rxq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_rq *rq = &eswitch_dev->rxq[qid].rqs;
+ int rc = -EINVAL;
+
+ if (eswitch_dev->rxq[qid].state == CNXK_ESWITCH_QUEUE_STATE_STARTED)
+ return 0;
+
+ if (eswitch_dev->rxq[qid].state != CNXK_ESWITCH_QUEUE_STATE_CONFIGURED) {
+ plt_err("Eswitch rxq %d not configured yet", qid);
+ goto done;
+ }
+
+ rc = roc_nix_rq_ena_dis(rq, true);
+ if (rc) {
+ plt_err("Failed to enable rxq=%u, rc=%d", qid, rc);
+ goto done;
+ }
+
+ eswitch_dev->rxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_STARTED;
+done:
+ return rc;
+}
+
+int
+cnxk_eswitch_rxq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_rq *rq = &eswitch_dev->rxq[qid].rqs;
+ int rc = -EINVAL;
+
+ if (eswitch_dev->rxq[qid].state == CNXK_ESWITCH_QUEUE_STATE_STOPPED ||
+ eswitch_dev->rxq[qid].state == CNXK_ESWITCH_QUEUE_STATE_RELEASED)
+ return 0;
+
+ if (eswitch_dev->rxq[qid].state != CNXK_ESWITCH_QUEUE_STATE_STARTED) {
+ plt_err("Eswitch rxq %d not started", qid);
+ goto done;
+ }
+
+ rc = roc_nix_rq_ena_dis(rq, false);
+ if (rc) {
+ plt_err("Failed to disable rxq=%u, rc=%d", qid, rc);
+ goto done;
+ }
+
+ eswitch_dev->rxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_STOPPED;
+done:
+ return rc;
+}
+
+int
+cnxk_eswitch_rxq_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_rq *rq;
+ struct roc_nix_cq *cq;
+ int rc;
+
+ if (eswitch_dev->rxq[qid].state == CNXK_ESWITCH_QUEUE_STATE_RELEASED)
+ return 0;
+
+ /* Cleanup ROC SQ */
+ rq = &eswitch_dev->rxq[qid].rqs;
+ rc = roc_nix_rq_fini(rq);
+ if (rc) {
+ plt_err("Failed to cleanup sq, rc=%d", rc);
+ goto fail;
+ }
+
+ eswitch_dev->rxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_RELEASED;
+
+ /* Cleanup ROC CQ */
+ cq = &eswitch_dev->cxq[qid].cqs;
+ rc = roc_nix_cq_fini(cq);
+ if (rc) {
+ plt_err("Failed to cleanup cq, rc=%d", rc);
+ goto fail;
+ }
+
+ eswitch_dev->cxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_RELEASED;
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_rxq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint16_t nb_desc,
+ const struct rte_eth_rxconf *rx_conf, struct rte_mempool *mp)
+{
+ struct roc_nix *nix = &eswitch_dev->nix;
+ struct rte_mempool *lpb_pool = mp;
+ struct rte_mempool_ops *ops;
+ const char *platform_ops;
+ struct roc_nix_rq *rq;
+ struct roc_nix_cq *cq;
+ uint16_t first_skip;
+ int rc = -EINVAL;
+
+ if (eswitch_dev->rxq[qid].state != CNXK_ESWITCH_QUEUE_STATE_RELEASED ||
+ eswitch_dev->cxq[qid].state != CNXK_ESWITCH_QUEUE_STATE_RELEASED) {
+ plt_err("Queue %d is in invalid state %d, cannot be setup", qid,
+ eswitch_dev->txq[qid].state);
+ goto fail;
+ }
+
+ RTE_SET_USED(rx_conf);
+ platform_ops = rte_mbuf_platform_mempool_ops();
+ /* This driver needs cnxk_npa mempool ops to work */
+ ops = rte_mempool_get_ops(lpb_pool->ops_index);
+ if (strncmp(ops->name, platform_ops, RTE_MEMPOOL_OPS_NAMESIZE)) {
+ plt_err("mempool ops should be of cnxk_npa type");
+ goto fail;
+ }
+
+ if (lpb_pool->pool_id == 0) {
+ plt_err("Invalid pool_id");
+ goto fail;
+ }
+
+ /* Setup ROC CQ */
+ cq = &eswitch_dev->cxq[qid].cqs;
+ memset(cq, 0, sizeof(struct roc_nix_cq));
+ cq->qid = qid;
+ cq->nb_desc = nb_desc;
+ rc = roc_nix_cq_init(nix, cq);
+ if (rc) {
+ plt_err("Failed to init roc cq for rq=%d, rc=%d", qid, rc);
+ goto fail;
+ }
+ eswitch_dev->cxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_CONFIGURED;
+
+ /* Setup ROC RQ */
+ rq = &eswitch_dev->rxq[qid].rqs;
+ memset(rq, 0, sizeof(struct roc_nix_rq));
+ rq->qid = qid;
+ rq->cqid = cq->qid;
+ rq->aura_handle = lpb_pool->pool_id;
+ rq->flow_tag_width = 32;
+ rq->sso_ena = false;
+
+ /* Calculate first mbuf skip */
+ first_skip = (sizeof(struct rte_mbuf));
+ first_skip += RTE_PKTMBUF_HEADROOM;
+ first_skip += rte_pktmbuf_priv_size(lpb_pool);
+ rq->first_skip = first_skip;
+ rq->later_skip = sizeof(struct rte_mbuf) + rte_pktmbuf_priv_size(lpb_pool);
+ rq->lpb_size = lpb_pool->elt_size;
+ if (roc_errata_nix_no_meta_aura())
+ rq->lpb_drop_ena = true;
+
+ rc = roc_nix_rq_init(nix, rq, true);
+ if (rc) {
+ plt_err("Failed to init roc rq for rq=%d, rc=%d", qid, rc);
+ goto cq_fini;
+ }
+ eswitch_dev->rxq[qid].state = CNXK_ESWITCH_QUEUE_STATE_CONFIGURED;
+
+ return 0;
+cq_fini:
+ rc |= roc_nix_cq_fini(cq);
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_txq_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid)
+{
+ struct roc_nix_sq *sq;
+ int rc = 0;
+
+ if (eswitch_dev->txq[qid].state == CNXK_ESWITCH_QUEUE_STATE_RELEASED)
+ return 0;
+
+ /* Cleanup ROC SQ */
+ sq = &eswitch_dev->txq[qid].sqs;
+ rc = roc_nix_sq_fini(sq);
+ if (rc) {
+ plt_err("Failed to cleanup sq, rc=%d", rc);
+ goto fail;
+ }
+
+ eswitch_dev->txq[qid].state = CNXK_ESWITCH_QUEUE_STATE_RELEASED;
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_txq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint16_t nb_desc,
+ const struct rte_eth_txconf *tx_conf)
+{
+ struct roc_nix_sq *sq;
+ int rc = 0;
+
+ if (eswitch_dev->txq[qid].state != CNXK_ESWITCH_QUEUE_STATE_RELEASED) {
+ plt_err("Queue %d is in invalid state %d, cannot be setup", qid,
+ eswitch_dev->txq[qid].state);
+ rc = -EINVAL;
+ goto fail;
+ }
+ RTE_SET_USED(tx_conf);
+ /* Setup ROC SQ */
+ sq = &eswitch_dev->txq[qid].sqs;
+ memset(sq, 0, sizeof(struct roc_nix_sq));
+ sq->qid = qid;
+ sq->nb_desc = nb_desc;
+ sq->max_sqe_sz = NIX_MAXSQESZ_W8;
+ if (sq->nb_desc >= CNXK_NIX_DEF_SQ_COUNT)
+ sq->fc_hyst_bits = 0x1;
+
+ rc = roc_nix_sq_init(&eswitch_dev->nix, sq);
+ if (rc)
+ plt_err("Failed to init sq=%d, rc=%d", qid, rc);
+
+ eswitch_dev->txq[qid].state = CNXK_ESWITCH_QUEUE_STATE_CONFIGURED;
+
+fail:
+ return rc;
+}
+
+static int
+cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
+{
+ struct cnxk_eswitch_dev *eswitch_dev;
+ const struct rte_memzone *mz = NULL;
+ int rc = -ENOMEM;
+
+ RTE_SET_USED(pci_drv);
+
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ if (!eswitch_dev) {
+ rc = roc_plt_init();
+ if (rc) {
+ plt_err("Failed to initialize platform model, rc=%d", rc);
+ return rc;
+ }
+
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return 0;
+
+ mz = rte_memzone_reserve_aligned(CNXK_REP_ESWITCH_DEV_MZ, sizeof(*eswitch_dev),
+ SOCKET_ID_ANY, 0, RTE_CACHE_LINE_SIZE);
+ if (mz == NULL) {
+ plt_err("Failed to reserve a memzone");
+ goto fail;
+ }
+
+ eswitch_dev = mz->addr;
+ eswitch_dev->pci_dev = pci_dev;
+ }
+
+ /* Spinlock for synchronization between representors traffic and control
+ * messages
+ */
+ rte_spinlock_init(&eswitch_dev->rep_lock);
+
+ return rc;
+fail:
+ return rc;
+}
+
+static const struct rte_pci_id cnxk_eswitch_pci_map[] = {
+ {RTE_PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVID_CNXK_RVU_ESWITCH_PF)},
+ {
+ .vendor_id = 0,
+ },
+};
+
+static struct rte_pci_driver cnxk_eswitch_pci = {
+ .id_table = cnxk_eswitch_pci_map,
+ .drv_flags =
+ RTE_PCI_DRV_NEED_MAPPING | RTE_PCI_DRV_NEED_IOVA_AS_VA | RTE_PCI_DRV_PROBE_AGAIN,
+ .probe = cnxk_eswitch_dev_probe,
+ .remove = cnxk_eswitch_dev_remove,
+};
+
+RTE_PMD_REGISTER_PCI(cnxk_eswitch, cnxk_eswitch_pci);
+RTE_PMD_REGISTER_PCI_TABLE(cnxk_eswitch, cnxk_eswitch_pci_map);
+RTE_PMD_REGISTER_KMOD_DEP(cnxk_eswitch, "vfio-pci");
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
new file mode 100644
index 0000000000..d1b4fa8761
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -0,0 +1,103 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#ifndef __CNXK_ESWITCH_H__
+#define __CNXK_ESWITCH_H__
+
+#include <sys/socket.h>
+#include <sys/un.h>
+
+#include <cnxk_ethdev.h>
+
+#include "cn10k_tx.h"
+
+#define CNXK_ESWITCH_CTRL_MSG_SOCK_PATH "/tmp/cxk_rep_ctrl_msg_sock"
+#define CNXK_REP_ESWITCH_DEV_MZ "cnxk_eswitch_dev"
+#define CNXK_ESWITCH_VLAN_TPID 0x8100
+#define CNXK_ESWITCH_MAX_TXQ 256
+#define CNXK_ESWITCH_MAX_RXQ 256
+#define CNXK_ESWITCH_LBK_CHAN 63
+#define CNXK_ESWITCH_VFPF_SHIFT 8
+
+#define CNXK_ESWITCH_QUEUE_STATE_RELEASED 0
+#define CNXK_ESWITCH_QUEUE_STATE_CONFIGURED 1
+#define CNXK_ESWITCH_QUEUE_STATE_STARTED 2
+#define CNXK_ESWITCH_QUEUE_STATE_STOPPED 3
+
+struct cnxk_rep_info {
+ struct rte_eth_dev *rep_eth_dev;
+};
+
+struct cnxk_eswitch_txq {
+ struct roc_nix_sq sqs;
+ uint8_t state;
+};
+
+struct cnxk_eswitch_rxq {
+ struct roc_nix_rq rqs;
+ uint8_t state;
+};
+
+struct cnxk_eswitch_cxq {
+ struct roc_nix_cq cqs;
+ uint8_t state;
+};
+
+TAILQ_HEAD(eswitch_flow_list, roc_npc_flow);
+struct cnxk_eswitch_dev {
+ /* Input parameters */
+ struct plt_pci_device *pci_dev;
+ /* ROC NIX */
+ struct roc_nix nix;
+
+ /* ROC NPC */
+ struct roc_npc npc;
+
+ /* ROC NPA */
+ struct rte_mempool *ctrl_chan_pool;
+ const struct plt_memzone *pktmem_mz;
+ uint64_t pkt_aura;
+
+ /* Eswitch RQs, SQs and CQs */
+ struct cnxk_eswitch_txq *txq;
+ struct cnxk_eswitch_rxq *rxq;
+ struct cnxk_eswitch_cxq *cxq;
+
+ /* Configured queue count */
+ uint16_t nb_rxq;
+ uint16_t nb_txq;
+ uint16_t rep_cnt;
+ uint8_t configured;
+
+ /* Port representor fields */
+ rte_spinlock_t rep_lock;
+ uint16_t switch_domain_id;
+ uint16_t eswitch_vdev;
+ struct cnxk_rep_info *rep_info;
+};
+
+static inline struct cnxk_eswitch_dev *
+cnxk_eswitch_pmd_priv(void)
+{
+ const struct rte_memzone *mz;
+
+ mz = rte_memzone_lookup(CNXK_REP_ESWITCH_DEV_MZ);
+ if (!mz)
+ return NULL;
+
+ return mz->addr;
+}
+
+int cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_eswitch_txq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint16_t nb_desc,
+ const struct rte_eth_txconf *tx_conf);
+int cnxk_eswitch_txq_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+int cnxk_eswitch_rxq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint16_t nb_desc,
+ const struct rte_eth_rxconf *rx_conf, struct rte_mempool *mp);
+int cnxk_eswitch_rxq_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+int cnxk_eswitch_rxq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+int cnxk_eswitch_rxq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+int cnxk_eswitch_txq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+int cnxk_eswitch_txq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+#endif /* __CNXK_ESWITCH_H__ */
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index e83f3c9050..012d098f80 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -28,6 +28,7 @@ sources = files(
'cnxk_ethdev_sec.c',
'cnxk_ethdev_telemetry.c',
'cnxk_ethdev_sec_telemetry.c',
+ 'cnxk_eswitch.c',
'cnxk_link.c',
'cnxk_lookup.c',
'cnxk_ptp.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 03/23] net/cnxk: eswitch HW resource configuration
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
2024-02-01 13:07 ` [PATCH v3 01/23] common/cnxk: add support for representors Harman Kalra
2024-02-01 13:07 ` [PATCH v3 02/23] net/cnxk: implementing eswitch device Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 04/23] net/cnxk: eswitch devargs parsing Harman Kalra
` (19 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Configuring the hardware resources used by the eswitch device.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 223 ++++++++++++++++++++++++++++++++
1 file changed, 223 insertions(+)
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index c4ea3063ae..5712b71c3b 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -6,6 +6,47 @@
#define CNXK_NIX_DEF_SQ_COUNT 512
+static int
+eswitch_hw_rsrc_cleanup(struct cnxk_eswitch_dev *eswitch_dev, struct rte_pci_device *pci_dev)
+{
+ struct roc_nix *nix;
+ int rc = 0;
+
+ nix = &eswitch_dev->nix;
+
+ roc_nix_unregister_queue_irqs(nix);
+ roc_nix_tm_fini(nix);
+ rc = roc_nix_lf_free(nix);
+ if (rc) {
+ plt_err("Failed to cleanup sq, rc %d", rc);
+ goto exit;
+ }
+
+ /* Check if this device is hosting common resource */
+ nix = roc_idev_npa_nix_get();
+ if (!nix || nix->pci_dev != pci_dev) {
+ rc = 0;
+ goto exit;
+ }
+
+ /* Try nix fini now */
+ rc = roc_nix_dev_fini(nix);
+ if (rc == -EAGAIN) {
+ plt_info("Common resource in use by other devices %s", pci_dev->name);
+ goto exit;
+ } else if (rc) {
+ plt_err("Failed in nix dev fini, rc=%d", rc);
+ goto exit;
+ }
+
+ rte_free(eswitch_dev->txq);
+ rte_free(eswitch_dev->rxq);
+ rte_free(eswitch_dev->cxq);
+
+exit:
+ return rc;
+}
+
static int
cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
{
@@ -21,6 +62,9 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
goto exit;
}
+ /* Cleanup HW resources */
+ eswitch_hw_rsrc_cleanup(eswitch_dev, pci_dev);
+
rte_free(eswitch_dev);
exit:
return rc;
@@ -317,6 +361,177 @@ cnxk_eswitch_txq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint1
return rc;
}
+static int
+nix_lf_setup(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ uint16_t nb_rxq, nb_txq, nb_cq;
+ struct roc_nix_fc_cfg fc_cfg;
+ struct roc_nix *nix;
+ uint64_t rx_cfg;
+ void *qs;
+ int rc;
+
+ /* Initialize base roc nix */
+ nix = &eswitch_dev->nix;
+ nix->pci_dev = eswitch_dev->pci_dev;
+ nix->hw_vlan_ins = true;
+ nix->reta_sz = ROC_NIX_RSS_RETA_SZ_256;
+ rc = roc_nix_dev_init(nix);
+ if (rc) {
+ plt_err("Failed to init nix eswitch device, rc=%d(%s)", rc, roc_error_msg_get(rc));
+ goto fail;
+ }
+
+ /* Get the representors count */
+ rc = roc_nix_max_rep_count(&eswitch_dev->nix);
+ if (rc) {
+ plt_err("Failed to get rep cnt, rc=%d(%s)", rc, roc_error_msg_get(rc));
+ goto free_cqs;
+ }
+
+ /* Allocating an NIX LF */
+ nb_rxq = CNXK_ESWITCH_MAX_RXQ;
+ nb_txq = CNXK_ESWITCH_MAX_TXQ;
+ nb_cq = CNXK_ESWITCH_MAX_RXQ;
+ rx_cfg = ROC_NIX_LF_RX_CFG_DIS_APAD;
+ rc = roc_nix_lf_alloc(nix, nb_rxq, nb_txq, rx_cfg);
+ if (rc) {
+ plt_err("lf alloc failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto dev_fini;
+ }
+
+ if (nb_rxq) {
+ /* Allocate memory for eswitch rq's and cq's */
+ qs = plt_zmalloc(sizeof(struct cnxk_eswitch_rxq) * nb_rxq, 0);
+ if (!qs) {
+ plt_err("Failed to alloc eswitch rxq");
+ goto lf_free;
+ }
+ eswitch_dev->rxq = qs;
+ }
+
+ if (nb_txq) {
+ /* Allocate memory for roc sq's */
+ qs = plt_zmalloc(sizeof(struct cnxk_eswitch_txq) * nb_txq, 0);
+ if (!qs) {
+ plt_err("Failed to alloc eswitch txq");
+ goto free_rqs;
+ }
+ eswitch_dev->txq = qs;
+ }
+
+ if (nb_cq) {
+ qs = plt_zmalloc(sizeof(struct cnxk_eswitch_cxq) * nb_cq, 0);
+ if (!qs) {
+ plt_err("Failed to alloc eswitch cxq");
+ goto free_sqs;
+ }
+ eswitch_dev->cxq = qs;
+ }
+
+ eswitch_dev->nb_rxq = nb_rxq;
+ eswitch_dev->nb_txq = nb_txq;
+
+ /* Re-enable NIX LF error interrupts */
+ roc_nix_err_intr_ena_dis(nix, true);
+ roc_nix_ras_intr_ena_dis(nix, true);
+
+ rc = roc_nix_lso_fmt_setup(nix);
+ if (rc) {
+ plt_err("lso setup failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto free_cqs;
+ }
+
+ rc = roc_nix_switch_hdr_set(nix, 0, 0, 0, 0);
+ if (rc) {
+ plt_err("switch hdr set failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto free_cqs;
+ }
+
+ rc = roc_nix_rss_default_setup(nix,
+ FLOW_KEY_TYPE_IPV4 | FLOW_KEY_TYPE_TCP | FLOW_KEY_TYPE_UDP);
+ if (rc) {
+ plt_err("rss default setup failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto free_cqs;
+ }
+
+ rc = roc_nix_tm_init(nix);
+ if (rc) {
+ plt_err("tm failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto free_cqs;
+ }
+
+ /* Register queue IRQs */
+ rc = roc_nix_register_queue_irqs(nix);
+ if (rc) {
+ plt_err("Failed to register queue interrupts rc=%d", rc);
+ goto tm_fini;
+ }
+
+ /* Enable default tree */
+ rc = roc_nix_tm_hierarchy_enable(nix, ROC_NIX_TM_DEFAULT, false);
+ if (rc) {
+ plt_err("tm default hierarchy enable failed = %s(%d)", roc_error_msg_get(rc), rc);
+ goto q_irq_fini;
+ }
+
+ memset(&fc_cfg, 0, sizeof(struct roc_nix_fc_cfg));
+ fc_cfg.rxchan_cfg.enable = false;
+ rc = roc_nix_fc_config_set(nix, &fc_cfg);
+ if (rc) {
+ plt_err("Failed to setup flow control, rc=%d(%s)", rc, roc_error_msg_get(rc));
+ goto q_irq_fini;
+ }
+
+ roc_nix_fc_mode_get(nix);
+
+ return rc;
+q_irq_fini:
+ roc_nix_unregister_queue_irqs(nix);
+tm_fini:
+ roc_nix_tm_fini(nix);
+free_cqs:
+ rte_free(eswitch_dev->cxq);
+free_sqs:
+ rte_free(eswitch_dev->txq);
+free_rqs:
+ rte_free(eswitch_dev->rxq);
+lf_free:
+ roc_nix_lf_free(nix);
+dev_fini:
+ roc_nix_dev_fini(nix);
+fail:
+ return rc;
+}
+
+static int
+eswitch_hw_rsrc_setup(struct cnxk_eswitch_dev *eswitch_dev, struct rte_pci_device *pci_dev)
+{
+ struct roc_nix *nix;
+ int rc;
+
+ nix = &eswitch_dev->nix;
+ rc = nix_lf_setup(eswitch_dev);
+ if (rc) {
+ plt_err("Failed to setup hw rsrc, rc=%d(%s)", rc, roc_error_msg_get(rc));
+ goto fail;
+ }
+
+ /* Initialize roc npc */
+ eswitch_dev->npc.roc_nix = nix;
+ eswitch_dev->npc.flow_max_priority = 3;
+ eswitch_dev->npc.flow_prealloc_size = 1;
+ rc = roc_npc_init(&eswitch_dev->npc);
+ if (rc)
+ goto rsrc_cleanup;
+
+ return rc;
+rsrc_cleanup:
+ eswitch_hw_rsrc_cleanup(eswitch_dev, pci_dev);
+fail:
+ return rc;
+}
+
static int
cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
{
@@ -346,6 +561,12 @@ cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pc
eswitch_dev = mz->addr;
eswitch_dev->pci_dev = pci_dev;
+
+ rc = eswitch_hw_rsrc_setup(eswitch_dev, pci_dev);
+ if (rc) {
+ plt_err("Failed to setup hw rsrc, rc=%d(%s)", rc, roc_error_msg_get(rc));
+ goto free_mem;
+ }
}
/* Spinlock for synchronization between representors traffic and control
@@ -354,6 +575,8 @@ cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pc
rte_spinlock_init(&eswitch_dev->rep_lock);
return rc;
+free_mem:
+ rte_memzone_free(mz);
fail:
return rc;
}
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 04/23] net/cnxk: eswitch devargs parsing
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (2 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 03/23] net/cnxk: eswitch HW resource configuration Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 05/23] net/cnxk: probing representor ports Harman Kalra
` (18 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Implementing the devargs parsing logic via which the representors
pattern is provided. These patterns define for which representies
representors shall be created.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 88 +++++++++++++++++
drivers/net/cnxk/cnxk_eswitch.h | 52 ++++++++++
drivers/net/cnxk/cnxk_eswitch_devargs.c | 124 ++++++++++++++++++++++++
drivers/net/cnxk/meson.build | 1 +
4 files changed, 265 insertions(+)
create mode 100644 drivers/net/cnxk/cnxk_eswitch_devargs.c
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index 5712b71c3b..df1011cf7a 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -388,6 +388,7 @@ nix_lf_setup(struct cnxk_eswitch_dev *eswitch_dev)
plt_err("Failed to get rep cnt, rc=%d(%s)", rc, roc_error_msg_get(rc));
goto free_cqs;
}
+ eswitch_dev->repr_cnt.max_repr = eswitch_dev->nix.rep_cnt;
/* Allocating an NIX LF */
nb_rxq = CNXK_ESWITCH_MAX_RXQ;
@@ -532,11 +533,73 @@ eswitch_hw_rsrc_setup(struct cnxk_eswitch_dev *eswitch_dev, struct rte_pci_devic
return rc;
}
+int
+cnxk_eswitch_representor_info_get(struct cnxk_eswitch_dev *eswitch_dev,
+ struct rte_eth_representor_info *info)
+{
+ struct cnxk_eswitch_devargs *esw_da;
+ int rc = 0, n_entries, i, j = 0, k = 0;
+
+ for (i = 0; i < eswitch_dev->nb_esw_da; i++) {
+ for (j = 0; j < eswitch_dev->esw_da[i].nb_repr_ports; j++)
+ k++;
+ }
+ n_entries = k;
+
+ if (info == NULL)
+ goto out;
+
+ if ((uint32_t)n_entries > info->nb_ranges_alloc)
+ n_entries = info->nb_ranges_alloc;
+
+ k = 0;
+ info->controller = 0;
+ info->pf = 0;
+ for (i = 0; i < eswitch_dev->nb_esw_da; i++) {
+ esw_da = &eswitch_dev->esw_da[i];
+ info->ranges[k].type = esw_da->da.type;
+ switch (esw_da->da.type) {
+ case RTE_ETH_REPRESENTOR_PF:
+ info->ranges[k].controller = 0;
+ info->ranges[k].pf = esw_da->repr_hw_info[0].pfvf;
+ info->ranges[k].vf = 0;
+ info->ranges[k].id_base = info->ranges[i].pf;
+ info->ranges[k].id_end = info->ranges[i].pf;
+ snprintf(info->ranges[k].name, sizeof(info->ranges[k].name), "pf%d",
+ info->ranges[k].pf);
+ k++;
+ break;
+ case RTE_ETH_REPRESENTOR_VF:
+ for (j = 0; j < esw_da->nb_repr_ports; j++) {
+ info->ranges[k].controller = 0;
+ info->ranges[k].pf = esw_da->da.ports[0];
+ info->ranges[k].vf = esw_da->repr_hw_info[j].pfvf;
+ info->ranges[k].id_base = esw_da->repr_hw_info[j].port_id;
+ info->ranges[k].id_end = esw_da->repr_hw_info[j].port_id;
+ snprintf(info->ranges[k].name, sizeof(info->ranges[k].name),
+ "pf%dvf%d", info->ranges[k].pf, info->ranges[k].vf);
+ k++;
+ }
+ break;
+ default:
+ plt_err("Invalid type %d", esw_da->da.type);
+ rc = 0;
+ goto fail;
+ };
+ }
+ info->nb_ranges = k;
+fail:
+ return rc;
+out:
+ return n_entries;
+}
+
static int
cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
{
struct cnxk_eswitch_dev *eswitch_dev;
const struct rte_memzone *mz = NULL;
+ uint16_t num_reps;
int rc = -ENOMEM;
RTE_SET_USED(pci_drv);
@@ -569,12 +632,37 @@ cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pc
}
}
+ if (pci_dev->device.devargs) {
+ rc = cnxk_eswitch_repr_devargs(pci_dev, eswitch_dev);
+ if (rc)
+ goto rsrc_cleanup;
+ }
+
+ if (eswitch_dev->repr_cnt.nb_repr_created > eswitch_dev->repr_cnt.max_repr) {
+ plt_err("Representors to be created %d can be greater than max allowed %d",
+ eswitch_dev->repr_cnt.nb_repr_created, eswitch_dev->repr_cnt.max_repr);
+ rc = -EINVAL;
+ goto rsrc_cleanup;
+ }
+
+ num_reps = eswitch_dev->repr_cnt.nb_repr_created;
+ if (!num_reps) {
+ plt_err("No representors enabled");
+ goto fail;
+ }
+
+ plt_esw_dbg("Max no of reps %d reps to be created %d Eswtch pfunc %x",
+ eswitch_dev->repr_cnt.max_repr, eswitch_dev->repr_cnt.nb_repr_created,
+ roc_nix_get_pf_func(&eswitch_dev->nix));
+
/* Spinlock for synchronization between representors traffic and control
* messages
*/
rte_spinlock_init(&eswitch_dev->rep_lock);
return rc;
+rsrc_cleanup:
+ eswitch_hw_rsrc_cleanup(eswitch_dev, pci_dev);
free_mem:
rte_memzone_free(mz);
fail:
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index d1b4fa8761..6ff296399e 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -25,6 +25,47 @@
#define CNXK_ESWITCH_QUEUE_STATE_STARTED 2
#define CNXK_ESWITCH_QUEUE_STATE_STOPPED 3
+enum cnxk_esw_da_pattern_type {
+ CNXK_ESW_DA_TYPE_LIST = 0,
+ CNXK_ESW_DA_TYPE_PFVF,
+};
+
+struct cnxk_esw_repr_hw_info {
+ /* Representee pcifunc value */
+ uint16_t hw_func;
+ /* rep id in sync with kernel */
+ uint16_t rep_id;
+ /* pf or vf id */
+ uint16_t pfvf;
+ /* representor port id assigned to representee */
+ uint16_t port_id;
+};
+
+/* Structure representing per devarg information - this can be per representee
+ * or range of representee
+ */
+struct cnxk_eswitch_devargs {
+ /* Devargs populated */
+ struct rte_eth_devargs da;
+ /* HW info of representee */
+ struct cnxk_esw_repr_hw_info *repr_hw_info;
+ /* No of representor ports */
+ uint16_t nb_repr_ports;
+ /* Devargs pattern type */
+ enum cnxk_esw_da_pattern_type type;
+};
+
+struct cnxk_eswitch_repr_cnt {
+ /* Max possible representors */
+ uint16_t max_repr;
+ /* Representors to be created as per devargs passed */
+ uint16_t nb_repr_created;
+ /* Representors probed successfully */
+ uint16_t nb_repr_probed;
+ /* Representors started representing a representee */
+ uint16_t nb_repr_started;
+};
+
struct cnxk_rep_info {
struct rte_eth_dev *rep_eth_dev;
};
@@ -70,6 +111,14 @@ struct cnxk_eswitch_dev {
uint16_t rep_cnt;
uint8_t configured;
+ /* Eswitch Representors Devargs */
+ uint16_t nb_esw_da;
+ uint16_t last_probed;
+ struct cnxk_eswitch_devargs esw_da[RTE_MAX_ETHPORTS];
+
+ /* No of representors */
+ struct cnxk_eswitch_repr_cnt repr_cnt;
+
/* Port representor fields */
rte_spinlock_t rep_lock;
uint16_t switch_domain_id;
@@ -90,6 +139,9 @@ cnxk_eswitch_pmd_priv(void)
}
int cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_eswitch_repr_devargs(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_eswitch_representor_info_get(struct cnxk_eswitch_dev *eswitch_dev,
+ struct rte_eth_representor_info *info);
int cnxk_eswitch_txq_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid, uint16_t nb_desc,
const struct rte_eth_txconf *tx_conf);
int cnxk_eswitch_txq_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
diff --git a/drivers/net/cnxk/cnxk_eswitch_devargs.c b/drivers/net/cnxk/cnxk_eswitch_devargs.c
new file mode 100644
index 0000000000..b5f28e185c
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_eswitch_devargs.c
@@ -0,0 +1,124 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#include <cnxk_eswitch.h>
+
+#define PF_SHIFT 10
+static inline int
+get_hw_func(uint16_t pf, uint16_t vf)
+{
+ return (pf << PF_SHIFT) | vf;
+}
+
+static int
+populate_repr_hw_info(struct cnxk_eswitch_dev *eswitch_dev, struct rte_eth_devargs *eth_da,
+ uint16_t idx)
+{
+ struct cnxk_eswitch_devargs *esw_da = &eswitch_dev->esw_da[idx];
+ uint16_t nb_repr_ports, hw_func;
+ int rc, i, j;
+
+ if (eth_da->type == RTE_ETH_REPRESENTOR_NONE) {
+ plt_err("No representor type found");
+ return -EINVAL;
+ }
+
+ if (eth_da->type != RTE_ETH_REPRESENTOR_VF && eth_da->type != RTE_ETH_REPRESENTOR_PF &&
+ eth_da->type != RTE_ETH_REPRESENTOR_SF) {
+ plt_err("unsupported representor type %d\n", eth_da->type);
+ return -ENOTSUP;
+ }
+
+ nb_repr_ports = (eth_da->type == RTE_ETH_REPRESENTOR_PF) ? eth_da->nb_ports :
+ eth_da->nb_representor_ports;
+ esw_da->nb_repr_ports = nb_repr_ports;
+ /* If plain list is provided as representor pattern */
+ if (eth_da->nb_ports == 0)
+ return 0;
+
+ esw_da->repr_hw_info = plt_zmalloc(nb_repr_ports * sizeof(struct cnxk_esw_repr_hw_info), 0);
+ if (!esw_da->repr_hw_info) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ plt_esw_dbg("Representor param %d has %d pfvf", idx, nb_repr_ports);
+ /* Check if representor can be created for PFVF and populating HW func list */
+ for (i = 0; i < nb_repr_ports; i++) {
+ if (eth_da->type == RTE_ETH_REPRESENTOR_PF)
+ hw_func = get_hw_func(eth_da->ports[i], 0);
+ else
+ hw_func = get_hw_func(eth_da->ports[0], eth_da->representor_ports[i] + 1);
+
+ for (j = 0; j < eswitch_dev->repr_cnt.max_repr; j++) {
+ if (eswitch_dev->nix.rep_pfvf_map[j] == hw_func)
+ break;
+ }
+
+ /* HW func which doesn not match the map table received from AF, no
+ * representor port is assigned.
+ */
+ if (j == eswitch_dev->repr_cnt.max_repr) {
+ plt_err("Representor port cant be created for PF%dVF%d", eth_da->ports[0],
+ eth_da->representor_ports[i]);
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ esw_da->repr_hw_info[i].hw_func = hw_func;
+ esw_da->repr_hw_info[i].rep_id = j;
+ esw_da->repr_hw_info[i].pfvf = (eth_da->type == RTE_ETH_REPRESENTOR_PF) ?
+ eth_da->ports[0] :
+ eth_da->representor_ports[i];
+ plt_esw_dbg(" HW func %x index %d type %d", hw_func, j, eth_da->type);
+ }
+
+ esw_da->type = CNXK_ESW_DA_TYPE_PFVF;
+
+ return 0;
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_repr_devargs(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev)
+{
+ struct rte_devargs *devargs = pci_dev->device.devargs;
+ struct rte_eth_devargs eth_da[RTE_MAX_ETHPORTS];
+ int rc, i, j, count;
+
+ if (devargs == NULL) {
+ plt_err("No devargs passed");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Parse devargs passed to ESW device */
+ rc = rte_eth_devargs_parse(devargs->args, eth_da, RTE_MAX_ETHPORTS);
+ if (rc < 0) {
+ plt_err("Failed to parse devargs, err %d", rc);
+ goto fail;
+ }
+
+ count = rc;
+ j = eswitch_dev->nb_esw_da;
+ for (i = 0; i < count; i++) {
+ rc = populate_repr_hw_info(eswitch_dev, ð_da[i], j);
+ if (rc) {
+ plt_err("Failed to populate representer hw funcs, err %d", rc);
+ goto fail;
+ }
+
+ rte_memcpy(&eswitch_dev->esw_da[j].da, ð_da[i], sizeof(struct rte_eth_devargs));
+ /* No of representor ports to be created */
+ eswitch_dev->repr_cnt.nb_repr_created += eswitch_dev->esw_da[j].nb_repr_ports;
+ j++;
+ }
+ eswitch_dev->nb_esw_da += count;
+
+ return 0;
+fail:
+ return rc;
+}
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index 012d098f80..ea7e363e89 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -29,6 +29,7 @@ sources = files(
'cnxk_ethdev_telemetry.c',
'cnxk_ethdev_sec_telemetry.c',
'cnxk_eswitch.c',
+ 'cnxk_eswitch_devargs.c',
'cnxk_link.c',
'cnxk_lookup.c',
'cnxk_ptp.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 05/23] net/cnxk: probing representor ports
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (3 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 04/23] net/cnxk: eswitch devargs parsing Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 06/23] common/cnxk: common NPC changes for eswitch Harman Kalra
` (17 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Thomas Monjalon, Nithin Dabilpuram, Kiran Kumar K,
Sunil Kumar Kori, Satha Rao, Harman Kalra, Anatoly Burakov
Cc: dev
Basic skeleton for probing representor devices. If PF device is
passed with "representor" devargs, representor ports gets probed
as a separate ethdev device.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
MAINTAINERS | 1 +
doc/guides/nics/cnxk.rst | 35 +++++
drivers/net/cnxk/cnxk_eswitch.c | 12 ++
drivers/net/cnxk/cnxk_eswitch.h | 8 +-
drivers/net/cnxk/cnxk_rep.c | 256 ++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep.h | 50 +++++++
drivers/net/cnxk/cnxk_rep_ops.c | 129 ++++++++++++++++
drivers/net/cnxk/meson.build | 2 +
8 files changed, 492 insertions(+), 1 deletion(-)
create mode 100644 drivers/net/cnxk/cnxk_rep.c
create mode 100644 drivers/net/cnxk/cnxk_rep.h
create mode 100644 drivers/net/cnxk/cnxk_rep_ops.c
diff --git a/MAINTAINERS b/MAINTAINERS
index 0d1c8126e3..2716178e18 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -827,6 +827,7 @@ M: Nithin Dabilpuram <ndabilpuram@marvell.com>
M: Kiran Kumar K <kirankumark@marvell.com>
M: Sunil Kumar Kori <skori@marvell.com>
M: Satha Rao <skoteshwar@marvell.com>
+M: Harman Kalra <hkalra@marvell.com>
T: git://dpdk.org/next/dpdk-next-net-mrvl
F: drivers/common/cnxk/
F: drivers/net/cnxk/
diff --git a/doc/guides/nics/cnxk.rst b/doc/guides/nics/cnxk.rst
index 58cb8e2283..496474913f 100644
--- a/doc/guides/nics/cnxk.rst
+++ b/doc/guides/nics/cnxk.rst
@@ -37,6 +37,7 @@ Features of the CNXK Ethdev PMD are:
- Inline IPsec processing support
- Ingress meter support
- Queue based priority flow control support
+- Port representors
Prerequisites
-------------
@@ -613,6 +614,40 @@ Runtime Config Options for inline device
With the above configuration, driver would poll for aging flows every 50
seconds.
+Port Representors
+-----------------
+
+The CNXK driver supports port representor model by adding virtual ethernet
+ports providing a logical representation in DPDK for physical function(PF) or
+SR-IOV virtual function (VF) devices for control and monitoring.
+
+Base device or parent device underneath the representor ports is a eswitch
+device which is not a cnxk ethernet device but has NIC RX and TX capabilities.
+Each representor port is represented by a RQ and SQ pair of this eswitch
+device.
+
+Implementation supports representors for both physical function and virtual
+function.
+
+Port representor ethdev instances can be spawned on an as needed basis
+through configuration parameters passed to the driver of the underlying
+base device using devargs ``-a <base PCI BDF>,representor=pf*vf*``
+
+.. note::
+
+ Representor ports to be created for respective representees should be
+ defined via standard representor devargs patterns
+ Eg. To create a representor for representee PF1VF0, devargs to be passed
+ is ``-a <base PCI BDF>,representor=pf01vf0``
+
+ Implementation supports creation of multiple port representors with pattern:
+ ``-a <base PCI BDF>,representor=[pf0vf[1,2],pf1vf[2-5]]``
+
+Port representor PMD supports following operations:
+
+- Get PF/VF statistics
+- Flow operations - create, validate, destroy, query, flush, dump
+
Debugging Options
-----------------
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index df1011cf7a..4b2c907f9f 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -3,6 +3,7 @@
*/
#include <cnxk_eswitch.h>
+#include <cnxk_rep.h>
#define CNXK_NIX_DEF_SQ_COUNT 512
@@ -62,6 +63,10 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
goto exit;
}
+ /* Remove representor devices associated with PF */
+ if (eswitch_dev->repr_cnt.nb_repr_created)
+ cnxk_rep_dev_remove(eswitch_dev);
+
/* Cleanup HW resources */
eswitch_hw_rsrc_cleanup(eswitch_dev, pci_dev);
@@ -655,6 +660,13 @@ cnxk_eswitch_dev_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pc
eswitch_dev->repr_cnt.max_repr, eswitch_dev->repr_cnt.nb_repr_created,
roc_nix_get_pf_func(&eswitch_dev->nix));
+ /* Probe representor ports */
+ rc = cnxk_rep_dev_probe(pci_dev, eswitch_dev);
+ if (rc) {
+ plt_err("Failed to probe representor ports");
+ goto rsrc_cleanup;
+ }
+
/* Spinlock for synchronization between representors traffic and control
* messages
*/
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index 6ff296399e..dcd5add6d0 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -66,6 +66,11 @@ struct cnxk_eswitch_repr_cnt {
uint16_t nb_repr_started;
};
+struct cnxk_eswitch_switch_domain {
+ uint16_t switch_domain_id;
+ uint16_t pf;
+};
+
struct cnxk_rep_info {
struct rte_eth_dev *rep_eth_dev;
};
@@ -121,7 +126,8 @@ struct cnxk_eswitch_dev {
/* Port representor fields */
rte_spinlock_t rep_lock;
- uint16_t switch_domain_id;
+ uint16_t nb_switch_domain;
+ struct cnxk_eswitch_switch_domain sw_dom[RTE_MAX_ETHPORTS];
uint16_t eswitch_vdev;
struct cnxk_rep_info *rep_info;
};
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
new file mode 100644
index 0000000000..55156f5b56
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -0,0 +1,256 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+#include <cnxk_rep.h>
+
+#define PF_SHIFT 10
+#define PF_MASK 0x3F
+
+static uint16_t
+get_pf(uint16_t hw_func)
+{
+ return (hw_func >> PF_SHIFT) & PF_MASK;
+}
+
+static uint16_t
+switch_domain_id_allocate(struct cnxk_eswitch_dev *eswitch_dev, uint16_t pf)
+{
+ int i = 0;
+
+ for (i = 0; i < eswitch_dev->nb_switch_domain; i++) {
+ if (eswitch_dev->sw_dom[i].pf == pf)
+ return eswitch_dev->sw_dom[i].switch_domain_id;
+ }
+
+ return RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID;
+}
+
+int
+cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev)
+{
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return 0;
+
+ plt_rep_dbg("Representor port:%d uninit", ethdev->data->port_id);
+ rte_free(ethdev->data->mac_addrs);
+ ethdev->data->mac_addrs = NULL;
+
+ return 0;
+}
+
+int
+cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ int i, rc = 0;
+
+ for (i = 0; i < eswitch_dev->nb_switch_domain; i++) {
+ rc = rte_eth_switch_domain_free(eswitch_dev->sw_dom[i].switch_domain_id);
+ if (rc)
+ plt_err("Failed to alloc switch domain: %d", rc);
+ }
+
+ return rc;
+}
+
+static int
+cnxk_rep_parent_setup(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ uint16_t pf, prev_pf = 0, switch_domain_id;
+ int rc, i, j = 0;
+
+ if (eswitch_dev->rep_info)
+ return 0;
+
+ eswitch_dev->rep_info =
+ plt_zmalloc(sizeof(eswitch_dev->rep_info[0]) * eswitch_dev->repr_cnt.max_repr, 0);
+ if (!eswitch_dev->rep_info) {
+ plt_err("Failed to alloc memory for rep info");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ /* Allocate switch domain for all PFs (VFs will be under same domain as PF) */
+ for (i = 0; i < eswitch_dev->repr_cnt.max_repr; i++) {
+ pf = get_pf(eswitch_dev->nix.rep_pfvf_map[i]);
+ if (pf == prev_pf)
+ continue;
+
+ rc = rte_eth_switch_domain_alloc(&switch_domain_id);
+ if (rc) {
+ plt_err("Failed to alloc switch domain: %d", rc);
+ goto fail;
+ }
+ plt_rep_dbg("Allocated switch domain id %d for pf %d\n", switch_domain_id, pf);
+ eswitch_dev->sw_dom[j].switch_domain_id = switch_domain_id;
+ eswitch_dev->sw_dom[j].pf = pf;
+ prev_pf = pf;
+ j++;
+ }
+ eswitch_dev->nb_switch_domain = j;
+
+ return 0;
+fail:
+ return rc;
+}
+
+static uint16_t
+cnxk_rep_tx_burst(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(tx_queue);
+ PLT_SET_USED(tx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
+
+static uint16_t
+cnxk_rep_rx_burst(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(rx_queue);
+ PLT_SET_USED(rx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
+
+static int
+cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
+{
+ struct cnxk_rep_dev *rep_params = (struct cnxk_rep_dev *)params;
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+
+ rep_dev->port_id = rep_params->port_id;
+ rep_dev->switch_domain_id = rep_params->switch_domain_id;
+ rep_dev->parent_dev = rep_params->parent_dev;
+ rep_dev->hw_func = rep_params->hw_func;
+ rep_dev->rep_id = rep_params->rep_id;
+
+ eth_dev->data->dev_flags |= RTE_ETH_DEV_REPRESENTOR;
+ eth_dev->data->representor_id = rep_params->port_id;
+ eth_dev->data->backer_port_id = eth_dev->data->port_id;
+
+ eth_dev->data->mac_addrs = plt_zmalloc(RTE_ETHER_ADDR_LEN, 0);
+ if (!eth_dev->data->mac_addrs) {
+ plt_err("Failed to allocate memory for mac addr");
+ return -ENOMEM;
+ }
+
+ rte_eth_random_addr(rep_dev->mac_addr);
+ memcpy(eth_dev->data->mac_addrs, rep_dev->mac_addr, RTE_ETHER_ADDR_LEN);
+
+ /* Set the device operations */
+ eth_dev->dev_ops = &cnxk_rep_dev_ops;
+
+ /* Rx/Tx functions stubs to avoid crashing */
+ eth_dev->rx_pkt_burst = cnxk_rep_rx_burst;
+ eth_dev->tx_pkt_burst = cnxk_rep_tx_burst;
+
+ /* Only single queues for representor devices */
+ eth_dev->data->nb_rx_queues = 1;
+ eth_dev->data->nb_tx_queues = 1;
+
+ eth_dev->data->dev_link.link_speed = RTE_ETH_SPEED_NUM_NONE;
+ eth_dev->data->dev_link.link_duplex = RTE_ETH_LINK_FULL_DUPLEX;
+ eth_dev->data->dev_link.link_status = RTE_ETH_LINK_UP;
+ eth_dev->data->dev_link.link_autoneg = RTE_ETH_LINK_FIXED;
+
+ return 0;
+}
+
+static int
+create_representor_ethdev(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev,
+ struct cnxk_eswitch_devargs *esw_da, int idx)
+{
+ char name[RTE_ETH_NAME_MAX_LEN];
+ struct rte_eth_dev *rep_eth_dev;
+ uint16_t hw_func;
+ int rc = 0;
+
+ struct cnxk_rep_dev rep = {.port_id = eswitch_dev->repr_cnt.nb_repr_probed,
+ .parent_dev = eswitch_dev};
+
+ if (esw_da->type == CNXK_ESW_DA_TYPE_PFVF) {
+ hw_func = esw_da->repr_hw_info[idx].hw_func;
+ rep.switch_domain_id = switch_domain_id_allocate(eswitch_dev, get_pf(hw_func));
+ if (rep.switch_domain_id == RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID) {
+ plt_err("Failed to get a valid switch domain id");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ esw_da->repr_hw_info[idx].port_id = rep.port_id;
+ /* Representor port net_bdf_port */
+ snprintf(name, sizeof(name), "net_%s_hw_%x_representor_%d", pci_dev->device.name,
+ hw_func, rep.port_id);
+
+ rep.hw_func = hw_func;
+ rep.rep_id = esw_da->repr_hw_info[idx].rep_id;
+
+ } else {
+ snprintf(name, sizeof(name), "net_%s_representor_%d", pci_dev->device.name,
+ rep.port_id);
+ rep.switch_domain_id = RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID;
+ }
+
+ rc = rte_eth_dev_create(&pci_dev->device, name, sizeof(struct cnxk_rep_dev), NULL, NULL,
+ cnxk_rep_dev_init, &rep);
+ if (rc) {
+ plt_err("Failed to create cnxk vf representor %s", name);
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ rep_eth_dev = rte_eth_dev_allocated(name);
+ if (!rep_eth_dev) {
+ plt_err("Failed to find the eth_dev for VF-Rep: %s.", name);
+ rc = -ENODEV;
+ goto fail;
+ }
+
+ plt_rep_dbg("Representor portid %d (%s) type %d probe done", rep_eth_dev->data->port_id,
+ name, esw_da->da.type);
+ eswitch_dev->rep_info[rep.port_id].rep_eth_dev = rep_eth_dev;
+ eswitch_dev->repr_cnt.nb_repr_probed++;
+
+ return 0;
+fail:
+ return rc;
+}
+
+int
+cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev)
+{
+ struct cnxk_eswitch_devargs *esw_da;
+ uint16_t num_rep;
+ int i, j, rc;
+
+ if (eswitch_dev->repr_cnt.nb_repr_created > RTE_MAX_ETHPORTS) {
+ plt_err("nb_representor_ports %d > %d MAX ETHPORTS\n",
+ eswitch_dev->repr_cnt.nb_repr_created, RTE_MAX_ETHPORTS);
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Initialize the internals of representor ports */
+ rc = cnxk_rep_parent_setup(eswitch_dev);
+ if (rc) {
+ plt_err("Failed to setup the parent device, err %d", rc);
+ goto fail;
+ }
+
+ for (i = eswitch_dev->last_probed; i < eswitch_dev->nb_esw_da; i++) {
+ esw_da = &eswitch_dev->esw_da[i];
+ /* Check the representor devargs */
+ num_rep = esw_da->nb_repr_ports;
+ for (j = 0; j < num_rep; j++) {
+ rc = create_representor_ethdev(pci_dev, eswitch_dev, esw_da, j);
+ if (rc)
+ goto fail;
+ }
+ }
+ eswitch_dev->last_probed = i;
+
+ return 0;
+fail:
+ return rc;
+}
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
new file mode 100644
index 0000000000..b802c44b33
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+#include <cnxk_eswitch.h>
+#include <cnxk_ethdev.h>
+
+#ifndef __CNXK_REP_H__
+#define __CNXK_REP_H__
+
+/* Common ethdev ops */
+extern struct eth_dev_ops cnxk_rep_dev_ops;
+
+struct cnxk_rep_dev {
+ uint16_t port_id;
+ uint16_t rep_id;
+ uint16_t switch_domain_id;
+ struct cnxk_eswitch_dev *parent_dev;
+ uint16_t hw_func;
+ uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
+};
+
+static inline struct cnxk_rep_dev *
+cnxk_rep_pmd_priv(const struct rte_eth_dev *eth_dev)
+{
+ return eth_dev->data->dev_private;
+}
+
+int cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev);
+int cnxk_rep_dev_info_get(struct rte_eth_dev *eth_dev, struct rte_eth_dev_info *dev_info);
+int cnxk_rep_representor_info_get(struct rte_eth_dev *dev, struct rte_eth_representor_info *info);
+int cnxk_rep_dev_configure(struct rte_eth_dev *eth_dev);
+
+int cnxk_rep_link_update(struct rte_eth_dev *eth_dev, int wait_to_compl);
+int cnxk_rep_dev_start(struct rte_eth_dev *eth_dev);
+int cnxk_rep_rx_queue_setup(struct rte_eth_dev *eth_dev, uint16_t queue_idx, uint16_t nb_desc,
+ unsigned int socket_id, const struct rte_eth_rxconf *rx_conf,
+ struct rte_mempool *mp);
+int cnxk_rep_tx_queue_setup(struct rte_eth_dev *eth_dev, uint16_t queue_idx, uint16_t nb_desc,
+ unsigned int socket_id, const struct rte_eth_txconf *tx_conf);
+void cnxk_rep_rx_queue_release(struct rte_eth_dev *dev, uint16_t queue_idx);
+void cnxk_rep_tx_queue_release(struct rte_eth_dev *dev, uint16_t queue_idx);
+int cnxk_rep_dev_stop(struct rte_eth_dev *eth_dev);
+int cnxk_rep_dev_close(struct rte_eth_dev *eth_dev);
+int cnxk_rep_stats_get(struct rte_eth_dev *eth_dev, struct rte_eth_stats *stats);
+int cnxk_rep_stats_reset(struct rte_eth_dev *eth_dev);
+int cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops);
+
+#endif /* __CNXK_REP_H__ */
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
new file mode 100644
index 0000000000..15448688ce
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -0,0 +1,129 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#include <cnxk_rep.h>
+
+int
+cnxk_rep_link_update(struct rte_eth_dev *ethdev, int wait_to_complete)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(wait_to_complete);
+ return 0;
+}
+
+int
+cnxk_rep_dev_info_get(struct rte_eth_dev *ethdev, struct rte_eth_dev_info *devinfo)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(devinfo);
+ return 0;
+}
+
+int
+cnxk_rep_dev_configure(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_dev_start(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_dev_close(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_dev_stop(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_rx_queue_setup(struct rte_eth_dev *ethdev, uint16_t rx_queue_id, uint16_t nb_rx_desc,
+ unsigned int socket_id, const struct rte_eth_rxconf *rx_conf,
+ struct rte_mempool *mb_pool)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(rx_queue_id);
+ PLT_SET_USED(nb_rx_desc);
+ PLT_SET_USED(socket_id);
+ PLT_SET_USED(rx_conf);
+ PLT_SET_USED(mb_pool);
+ return 0;
+}
+
+void
+cnxk_rep_rx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(queue_id);
+}
+
+int
+cnxk_rep_tx_queue_setup(struct rte_eth_dev *ethdev, uint16_t tx_queue_id, uint16_t nb_tx_desc,
+ unsigned int socket_id, const struct rte_eth_txconf *tx_conf)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(tx_queue_id);
+ PLT_SET_USED(nb_tx_desc);
+ PLT_SET_USED(socket_id);
+ PLT_SET_USED(tx_conf);
+ return 0;
+}
+
+void
+cnxk_rep_tx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(queue_id);
+}
+
+int
+cnxk_rep_stats_get(struct rte_eth_dev *ethdev, struct rte_eth_stats *stats)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(stats);
+ return 0;
+}
+
+int
+cnxk_rep_stats_reset(struct rte_eth_dev *ethdev)
+{
+ PLT_SET_USED(ethdev);
+ return 0;
+}
+
+int
+cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops)
+{
+ PLT_SET_USED(ethdev);
+ PLT_SET_USED(ops);
+ return 0;
+}
+
+/* CNXK platform representor dev ops */
+struct eth_dev_ops cnxk_rep_dev_ops = {
+ .dev_infos_get = cnxk_rep_dev_info_get,
+ .dev_configure = cnxk_rep_dev_configure,
+ .dev_start = cnxk_rep_dev_start,
+ .rx_queue_setup = cnxk_rep_rx_queue_setup,
+ .rx_queue_release = cnxk_rep_rx_queue_release,
+ .tx_queue_setup = cnxk_rep_tx_queue_setup,
+ .tx_queue_release = cnxk_rep_tx_queue_release,
+ .link_update = cnxk_rep_link_update,
+ .dev_close = cnxk_rep_dev_close,
+ .dev_stop = cnxk_rep_dev_stop,
+ .stats_get = cnxk_rep_stats_get,
+ .stats_reset = cnxk_rep_stats_reset,
+ .flow_ops_get = cnxk_rep_flow_ops_get
+};
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index ea7e363e89..fcd5d3d569 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -34,6 +34,8 @@ sources = files(
'cnxk_lookup.c',
'cnxk_ptp.c',
'cnxk_flow.c',
+ 'cnxk_rep.c',
+ 'cnxk_rep_ops.c',
'cnxk_stats.c',
'cnxk_tm.c',
)
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 06/23] common/cnxk: common NPC changes for eswitch
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (4 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 05/23] net/cnxk: probing representor ports Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 07/23] common/cnxk: interface to update VLAN TPID Harman Kalra
` (16 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Adding new MCAM API for installing flow using generic npc_install_flow
mbox and other helper APIs. Also adding rss action configuration for
eswitch.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/meson.build | 1 +
drivers/common/cnxk/roc_api.h | 3 +
drivers/common/cnxk/roc_eswitch.c | 306 +++++++++++++++++++++++++++++
drivers/common/cnxk/roc_eswitch.h | 22 +++
drivers/common/cnxk/roc_mbox.h | 33 ++++
drivers/common/cnxk/roc_npc.c | 26 ++-
drivers/common/cnxk/roc_npc.h | 5 +-
drivers/common/cnxk/roc_npc_mcam.c | 2 +-
drivers/common/cnxk/roc_npc_priv.h | 3 +-
drivers/common/cnxk/version.map | 6 +
10 files changed, 398 insertions(+), 9 deletions(-)
create mode 100644 drivers/common/cnxk/roc_eswitch.c
create mode 100644 drivers/common/cnxk/roc_eswitch.h
diff --git a/drivers/common/cnxk/meson.build b/drivers/common/cnxk/meson.build
index 56eea52909..e0e4600989 100644
--- a/drivers/common/cnxk/meson.build
+++ b/drivers/common/cnxk/meson.build
@@ -20,6 +20,7 @@ sources = files(
'roc_cpt_debug.c',
'roc_dev.c',
'roc_dpi.c',
+ 'roc_eswitch.c',
'roc_hash.c',
'roc_idev.c',
'roc_irq.c',
diff --git a/drivers/common/cnxk/roc_api.h b/drivers/common/cnxk/roc_api.h
index f630853088..6a86863c57 100644
--- a/drivers/common/cnxk/roc_api.h
+++ b/drivers/common/cnxk/roc_api.h
@@ -117,4 +117,7 @@
/* MACsec */
#include "roc_mcs.h"
+/* Eswitch */
+#include "roc_eswitch.h"
+
#endif /* _ROC_API_H_ */
diff --git a/drivers/common/cnxk/roc_eswitch.c b/drivers/common/cnxk/roc_eswitch.c
new file mode 100644
index 0000000000..e480ab1046
--- /dev/null
+++ b/drivers/common/cnxk/roc_eswitch.c
@@ -0,0 +1,306 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#include <arpa/inet.h>
+
+#include "roc_api.h"
+#include "roc_priv.h"
+
+static int
+eswitch_vlan_rx_cfg(uint16_t pcifunc, struct mbox *mbox)
+{
+ struct nix_vtag_config *vtag_cfg;
+ int rc;
+
+ vtag_cfg = mbox_alloc_msg_nix_vtag_cfg(mbox_get(mbox));
+ if (!vtag_cfg) {
+ rc = -EINVAL;
+ goto exit;
+ }
+
+ /* config strip, capture and size */
+ vtag_cfg->hdr.pcifunc = pcifunc;
+ vtag_cfg->vtag_size = NIX_VTAGSIZE_T4;
+ vtag_cfg->cfg_type = VTAG_RX; /* rx vlan cfg */
+ vtag_cfg->rx.vtag_type = NIX_RX_VTAG_TYPE0;
+ vtag_cfg->rx.strip_vtag = true;
+ vtag_cfg->rx.capture_vtag = true;
+
+ rc = mbox_process(mbox);
+ if (rc)
+ goto exit;
+
+ rc = 0;
+exit:
+ mbox_put(mbox);
+ return rc;
+}
+
+static int
+eswitch_vlan_tx_cfg(struct roc_npc_flow *flow, uint16_t pcifunc, struct mbox *mbox,
+ uint16_t vlan_tci, uint16_t *vidx)
+{
+ struct nix_vtag_config *vtag_cfg;
+ struct nix_vtag_config_rsp *rsp;
+ int rc;
+
+ union {
+ uint64_t reg;
+ struct nix_tx_vtag_action_s act;
+ } tx_vtag_action;
+
+ vtag_cfg = mbox_alloc_msg_nix_vtag_cfg(mbox_get(mbox));
+ if (!vtag_cfg) {
+ rc = -EINVAL;
+ goto exit;
+ }
+
+ /* Insert vlan tag */
+ vtag_cfg->hdr.pcifunc = pcifunc;
+ vtag_cfg->vtag_size = NIX_VTAGSIZE_T4;
+ vtag_cfg->cfg_type = VTAG_TX; /* tx vlan cfg */
+ vtag_cfg->tx.cfg_vtag0 = true;
+ vtag_cfg->tx.vtag0 = (((uint32_t)ROC_ESWITCH_VLAN_TPID << 16) | vlan_tci);
+
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ if (rsp->vtag0_idx < 0) {
+ plt_err("Failed to config TX VTAG action");
+ rc = -EINVAL;
+ goto exit;
+ }
+
+ *vidx = rsp->vtag0_idx;
+ tx_vtag_action.reg = 0;
+ tx_vtag_action.act.vtag0_def = rsp->vtag0_idx;
+ tx_vtag_action.act.vtag0_lid = NPC_LID_LA;
+ tx_vtag_action.act.vtag0_op = NIX_TX_VTAGOP_INSERT;
+ tx_vtag_action.act.vtag0_relptr = NIX_TX_VTAGACTION_VTAG0_RELPTR;
+
+ flow->vtag_action = tx_vtag_action.reg;
+
+ rc = 0;
+exit:
+ mbox_put(mbox);
+ return rc;
+}
+
+int
+roc_eswitch_npc_mcam_tx_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow, uint16_t pcifunc,
+ uint32_t vlan_tci)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+ struct npc_install_flow_req *req;
+ struct npc_install_flow_rsp *rsp;
+ struct mbox *mbox = npc->mbox;
+ uint16_t vidx = 0, lbkid;
+ int rc;
+
+ rc = eswitch_vlan_tx_cfg(flow, roc_npc->pf_func, mbox, vlan_tci, &vidx);
+ if (rc) {
+ plt_err("Failed to configure VLAN TX, err %d", rc);
+ goto fail;
+ }
+
+ req = mbox_alloc_msg_npc_install_flow(mbox_get(mbox));
+ if (!req) {
+ rc = -EINVAL;
+ goto exit;
+ }
+
+ lbkid = 0;
+ req->hdr.pcifunc = roc_npc->pf_func; /* Eswitch PF is requester */
+ req->vf = pcifunc;
+ req->entry = flow->mcam_id;
+ req->intf = NPC_MCAM_TX;
+ req->op = NIX_TX_ACTIONOP_UCAST_CHAN;
+ req->index = (lbkid << 8) | ROC_ESWITCH_LBK_CHAN;
+ req->set_cntr = 1;
+ req->vtag0_def = vidx;
+ req->vtag0_op = 1;
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ flow->nix_intf = NIX_INTF_TX;
+exit:
+ mbox_put(mbox);
+fail:
+ return rc;
+}
+
+static int
+eswitch_vtag_cfg_delete(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+ struct nix_vtag_config *vtag_cfg;
+ struct nix_vtag_config_rsp *rsp;
+ struct mbox *mbox = npc->mbox;
+ int rc = 0;
+
+ union {
+ uint64_t reg;
+ struct nix_tx_vtag_action_s act;
+ } tx_vtag_action;
+
+ tx_vtag_action.reg = flow->vtag_action;
+ vtag_cfg = mbox_alloc_msg_nix_vtag_cfg(mbox_get(mbox));
+
+ if (vtag_cfg == NULL) {
+ rc = -ENOSPC;
+ goto exit;
+ }
+
+ vtag_cfg->cfg_type = VTAG_TX;
+ vtag_cfg->vtag_size = NIX_VTAGSIZE_T4;
+ vtag_cfg->tx.vtag0_idx = tx_vtag_action.act.vtag0_def;
+ vtag_cfg->tx.free_vtag0 = true;
+
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ rc = rsp->hdr.rc;
+exit:
+ mbox_put(mbox);
+ return rc;
+}
+
+int
+roc_eswitch_npc_mcam_delete_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow,
+ uint16_t pcifunc)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+ struct npc_delete_flow_req *req;
+ struct msg_rsp *rsp;
+ struct mbox *mbox = npc->mbox;
+ int rc = 0;
+
+ /* Removing the VLAN TX config */
+ if (flow->nix_intf == NIX_INTF_TX) {
+ rc = eswitch_vtag_cfg_delete(roc_npc, flow);
+ if (rc)
+ plt_err("Failed to delete TX vtag config");
+ }
+
+ req = mbox_alloc_msg_npc_delete_flow(mbox_get(mbox));
+ if (!req) {
+ rc = -EINVAL;
+ goto exit;
+ }
+
+ req->entry = flow->mcam_id;
+ req->vf = pcifunc;
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ rc = rsp->hdr.rc;
+exit:
+ mbox_put(mbox);
+ return rc;
+}
+
+int
+roc_eswitch_npc_mcam_rx_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow, uint16_t pcifunc,
+ uint16_t vlan_tci, uint16_t vlan_tci_mask)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+ struct npc_install_flow_req *req;
+ struct npc_install_flow_rsp *rsp;
+ struct mbox *mbox = npc->mbox;
+ bool is_esw_dev;
+ int rc;
+
+ /* For ESW PF/VF */
+ is_esw_dev = (dev_get_pf(roc_npc->pf_func) == dev_get_pf(pcifunc));
+ /* VLAN Rx config */
+ if (is_esw_dev) {
+ rc = eswitch_vlan_rx_cfg(roc_npc->pf_func, mbox);
+ if (rc) {
+ plt_err("Failed to configure VLAN RX rule, err %d", rc);
+ goto fail;
+ }
+ }
+
+ req = mbox_alloc_msg_npc_install_flow(mbox_get(mbox));
+ if (!req) {
+ rc = -EINVAL;
+ goto exit;
+ }
+
+ req->vf = pcifunc;
+ /* Action */
+ req->op = NIX_RX_ACTIONOP_DEFAULT;
+ req->index = 0;
+ req->entry = flow->mcam_id;
+ req->hdr.pcifunc = roc_npc->pf_func; /* Eswitch PF is requester */
+ req->features = BIT_ULL(NPC_OUTER_VID) | BIT_ULL(NPC_VLAN_ETYPE_CTAG);
+ req->vtag0_valid = true;
+ /* For ESW PF/VF using configured vlan rx cfg while for other
+ * representees using standard vlan_type = 7 which is strip.
+ */
+ req->vtag0_type = is_esw_dev ? NIX_RX_VTAG_TYPE0 : NIX_RX_VTAG_TYPE7;
+ req->packet.vlan_etype = ROC_ESWITCH_VLAN_TPID;
+ req->mask.vlan_etype = 0xFFFF;
+ req->packet.vlan_tci = ntohs(vlan_tci & 0xFFFF);
+ req->mask.vlan_tci = ntohs(vlan_tci_mask);
+
+ req->channel = ROC_ESWITCH_LBK_CHAN;
+ req->chan_mask = 0xffff;
+ req->intf = NPC_MCAM_RX;
+ req->set_cntr = 1;
+ req->cntr_val = flow->ctr_id;
+
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ flow->nix_intf = NIX_INTF_RX;
+exit:
+ mbox_put(mbox);
+fail:
+ return rc;
+}
+
+int
+roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc, struct roc_npc_flow *flow,
+ uint32_t flowkey_cfg, uint16_t *reta_tbl)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+ struct roc_nix *roc_nix = roc_npc->roc_nix;
+ uint32_t rss_grp_idx;
+ uint8_t flowkey_algx;
+ int rc;
+
+ rc = npc_rss_free_grp_get(npc, &rss_grp_idx);
+ /* RSS group :0 is not usable for flow rss action */
+ if (rc < 0 || rss_grp_idx == 0)
+ return -ENOSPC;
+
+ /* Populating reta table for the specific RSS group */
+ rc = roc_nix_rss_reta_set(roc_nix, rss_grp_idx, reta_tbl);
+ if (rc) {
+ plt_err("Failed to init rss table rc = %d", rc);
+ return rc;
+ }
+
+ rc = roc_nix_rss_flowkey_set(roc_nix, &flowkey_algx, flowkey_cfg, rss_grp_idx,
+ flow->mcam_id);
+ if (rc) {
+ plt_err("Failed to set rss hash function rc = %d", rc);
+ return rc;
+ }
+
+ plt_bitmap_set(npc->rss_grp_entries, rss_grp_idx);
+
+ flow->npc_action &= (~(0xfULL));
+ flow->npc_action |= NIX_RX_ACTIONOP_RSS;
+ flow->npc_action |=
+ ((uint64_t)(flowkey_algx & NPC_RSS_ACT_ALG_MASK) << NPC_RSS_ACT_ALG_OFFSET) |
+ ((uint64_t)(rss_grp_idx & NPC_RSS_ACT_GRP_MASK) << NPC_RSS_ACT_GRP_OFFSET);
+ return 0;
+}
diff --git a/drivers/common/cnxk/roc_eswitch.h b/drivers/common/cnxk/roc_eswitch.h
new file mode 100644
index 0000000000..cdbe808a71
--- /dev/null
+++ b/drivers/common/cnxk/roc_eswitch.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#ifndef __ROC_ESWITCH_H__
+#define __ROC_ESWITCH_H__
+
+#define ROC_ESWITCH_VLAN_TPID 0x8100
+#define ROC_ESWITCH_LBK_CHAN 63
+
+/* NPC */
+int __roc_api roc_eswitch_npc_mcam_rx_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow,
+ uint16_t pcifunc, uint16_t vlan_tci,
+ uint16_t vlan_tci_mask);
+int __roc_api roc_eswitch_npc_mcam_tx_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow,
+ uint16_t pcifunc, uint32_t vlan_tci);
+int __roc_api roc_eswitch_npc_mcam_delete_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow,
+ uint16_t pcifunc);
+int __roc_api roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc,
+ struct roc_npc_flow *flow, uint32_t flowkey_cfg,
+ uint16_t *reta_tbl);
+#endif /* __ROC_ESWITCH_H__ */
diff --git a/drivers/common/cnxk/roc_mbox.h b/drivers/common/cnxk/roc_mbox.h
index b7e2f43d45..e903db391c 100644
--- a/drivers/common/cnxk/roc_mbox.h
+++ b/drivers/common/cnxk/roc_mbox.h
@@ -386,6 +386,18 @@ enum rvu_af_status {
RVU_INVALID_VF_ID = -256,
};
+/* For NIX RX vtag action */
+enum nix_rx_vtag0_type {
+ NIX_RX_VTAG_TYPE0,
+ NIX_RX_VTAG_TYPE1,
+ NIX_RX_VTAG_TYPE2,
+ NIX_RX_VTAG_TYPE3,
+ NIX_RX_VTAG_TYPE4,
+ NIX_RX_VTAG_TYPE5,
+ NIX_RX_VTAG_TYPE6,
+ NIX_RX_VTAG_TYPE7,
+};
+
struct ready_msg_rsp {
struct mbox_msghdr hdr;
uint16_t __io sclk_freq; /* SCLK frequency */
@@ -2442,6 +2454,8 @@ enum header_fields {
NPC_DMAC,
NPC_SMAC,
NPC_ETYPE,
+ NPC_VLAN_ETYPE_CTAG, /* 0x8100 */
+ NPC_VLAN_ETYPE_STAG, /* 0x88A8 */
NPC_OUTER_VID,
NPC_TOS,
NPC_SIP_IPV4,
@@ -2470,12 +2484,27 @@ struct flow_msg {
uint32_t __io ip4dst;
uint32_t __io ip6dst[4];
};
+ union {
+ uint32_t spi;
+ };
uint8_t __io tos;
uint8_t __io ip_ver;
uint8_t __io ip_proto;
uint8_t __io tc;
uint16_t __io sport;
uint16_t __io dport;
+ union {
+ uint8_t __io ip_flag;
+ uint8_t __io next_header;
+ };
+ uint16_t __io vlan_itci;
+ uint8_t __io icmp_type;
+ uint8_t __io icmp_code;
+ uint16_t __io tcp_flags;
+ uint32_t __io gtpu_teid;
+ uint32_t __io gtpc_teid;
+ uint32_t __io mpls_lse[4];
+ uint16_t __io sq_id;
};
struct npc_install_flow_req {
@@ -2485,6 +2514,7 @@ struct npc_install_flow_req {
uint64_t __io features;
uint16_t __io entry;
uint16_t __io channel;
+ uint16_t __io chan_mask;
uint8_t __io intf;
uint8_t __io set_cntr;
uint8_t __io default_rule;
@@ -2507,6 +2537,8 @@ struct npc_install_flow_req {
uint8_t __io vtag0_op;
uint16_t __io vtag1_def;
uint8_t __io vtag1_op;
+ /* old counter value */
+ uint16_t __io cntr_val;
};
struct npc_install_flow_rsp {
@@ -2521,6 +2553,7 @@ struct npc_delete_flow_req {
uint16_t __io start; /*Disable range of entries */
uint16_t __io end;
uint8_t __io all; /* PF + VFs */
+ uint16_t __io vf; /* Requesting VF */
};
struct npc_mcam_read_entry_req {
diff --git a/drivers/common/cnxk/roc_npc.c b/drivers/common/cnxk/roc_npc.c
index 9a0fe5f4e2..67a660a2bc 100644
--- a/drivers/common/cnxk/roc_npc.c
+++ b/drivers/common/cnxk/roc_npc.c
@@ -77,8 +77,23 @@ roc_npc_inl_mcam_clear_counter(uint32_t ctr_id)
}
int
-roc_npc_mcam_read_counter(struct roc_npc *roc_npc, uint32_t ctr_id,
- uint64_t *count)
+roc_npc_mcam_alloc_counter(struct roc_npc *roc_npc, uint16_t *ctr_id)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+
+ return npc_mcam_alloc_counter(npc->mbox, ctr_id);
+}
+
+int
+roc_npc_get_free_mcam_entry(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
+{
+ struct npc *npc = roc_npc_to_npc_priv(roc_npc);
+
+ return npc_get_free_mcam_entry(npc->mbox, flow, npc);
+}
+
+int
+roc_npc_mcam_read_counter(struct roc_npc *roc_npc, uint32_t ctr_id, uint64_t *count)
{
struct npc *npc = roc_npc_to_npc_priv(roc_npc);
@@ -157,14 +172,13 @@ roc_npc_mcam_free_all_resources(struct roc_npc *roc_npc)
}
int
-roc_npc_mcam_alloc_entries(struct roc_npc *roc_npc, int ref_entry,
- int *alloc_entry, int req_count, int priority,
- int *resp_count)
+roc_npc_mcam_alloc_entries(struct roc_npc *roc_npc, int ref_entry, int *alloc_entry, int req_count,
+ int priority, int *resp_count, bool is_conti)
{
struct npc *npc = roc_npc_to_npc_priv(roc_npc);
return npc_mcam_alloc_entries(npc->mbox, ref_entry, alloc_entry, req_count, priority,
- resp_count, 0);
+ resp_count, is_conti);
}
int
diff --git a/drivers/common/cnxk/roc_npc.h b/drivers/common/cnxk/roc_npc.h
index e880a7fa67..349c7f9d22 100644
--- a/drivers/common/cnxk/roc_npc.h
+++ b/drivers/common/cnxk/roc_npc.h
@@ -431,7 +431,8 @@ int __roc_api roc_npc_mcam_enable_all_entries(struct roc_npc *roc_npc, bool enab
int __roc_api roc_npc_mcam_alloc_entry(struct roc_npc *roc_npc, struct roc_npc_flow *mcam,
struct roc_npc_flow *ref_mcam, int prio, int *resp_count);
int __roc_api roc_npc_mcam_alloc_entries(struct roc_npc *roc_npc, int ref_entry, int *alloc_entry,
- int req_count, int priority, int *resp_count);
+ int req_count, int priority, int *resp_count,
+ bool is_conti);
int __roc_api roc_npc_mcam_ena_dis_entry(struct roc_npc *roc_npc, struct roc_npc_flow *mcam,
bool enable);
int __roc_api roc_npc_mcam_write_entry(struct roc_npc *roc_npc, struct roc_npc_flow *mcam);
@@ -442,6 +443,8 @@ int __roc_api roc_npc_get_low_priority_mcam(struct roc_npc *roc_npc);
int __roc_api roc_npc_mcam_free_counter(struct roc_npc *roc_npc, uint16_t ctr_id);
int __roc_api roc_npc_mcam_read_counter(struct roc_npc *roc_npc, uint32_t ctr_id, uint64_t *count);
int __roc_api roc_npc_mcam_clear_counter(struct roc_npc *roc_npc, uint32_t ctr_id);
+int __roc_api roc_npc_mcam_alloc_counter(struct roc_npc *roc_npc, uint16_t *ctr_id);
+int __roc_api roc_npc_get_free_mcam_entry(struct roc_npc *roc_npc, struct roc_npc_flow *flow);
int __roc_api roc_npc_inl_mcam_read_counter(uint32_t ctr_id, uint64_t *count);
int __roc_api roc_npc_inl_mcam_clear_counter(uint32_t ctr_id);
int __roc_api roc_npc_mcam_free_all_resources(struct roc_npc *roc_npc);
diff --git a/drivers/common/cnxk/roc_npc_mcam.c b/drivers/common/cnxk/roc_npc_mcam.c
index 3ef189e184..2de988a44b 100644
--- a/drivers/common/cnxk/roc_npc_mcam.c
+++ b/drivers/common/cnxk/roc_npc_mcam.c
@@ -4,7 +4,7 @@
#include "roc_api.h"
#include "roc_priv.h"
-static int
+int
npc_mcam_alloc_counter(struct mbox *mbox, uint16_t *ctr)
{
struct npc_mcam_alloc_counter_req *req;
diff --git a/drivers/common/cnxk/roc_npc_priv.h b/drivers/common/cnxk/roc_npc_priv.h
index c0809407a6..50b62b1244 100644
--- a/drivers/common/cnxk/roc_npc_priv.h
+++ b/drivers/common/cnxk/roc_npc_priv.h
@@ -432,6 +432,7 @@ roc_npc_to_npc_priv(struct roc_npc *npc)
return (struct npc *)npc->reserved;
}
+int npc_mcam_alloc_counter(struct mbox *mbox, uint16_t *ctr);
int npc_mcam_free_counter(struct mbox *mbox, uint16_t ctr_id);
int npc_mcam_read_counter(struct mbox *mbox, uint32_t ctr_id, uint64_t *count);
int npc_mcam_clear_counter(struct mbox *mbox, uint32_t ctr_id);
@@ -480,7 +481,6 @@ uint64_t npc_get_kex_capability(struct npc *npc);
int npc_process_ipv6_field_hash(const struct roc_npc_flow_item_ipv6 *ipv6_spec,
const struct roc_npc_flow_item_ipv6 *ipv6_mask,
struct npc_parse_state *pst, uint8_t type);
-int npc_rss_free_grp_get(struct npc *npc, uint32_t *grp);
int npc_rss_action_configure(struct roc_npc *roc_npc, const struct roc_npc_action_rss *rss,
uint8_t *alg_idx, uint32_t *rss_grp, uint32_t mcam_id);
int npc_rss_action_program(struct roc_npc *roc_npc, const struct roc_npc_action actions[],
@@ -496,4 +496,5 @@ void npc_aged_flows_bitmap_free(struct roc_npc *roc_npc);
int npc_aging_ctrl_thread_create(struct roc_npc *roc_npc, const struct roc_npc_action_age *age,
struct roc_npc_flow *flow);
void npc_aging_ctrl_thread_destroy(struct roc_npc *roc_npc);
+int npc_rss_free_grp_get(struct npc *npc, uint32_t *pos);
#endif /* _ROC_NPC_PRIV_H_ */
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index bd28803013..feda34b852 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -91,6 +91,10 @@ INTERNAL {
roc_dpi_disable;
roc_dpi_enable;
roc_error_msg_get;
+ roc_eswitch_npc_mcam_delete_rule;
+ roc_eswitch_npc_mcam_rx_rule;
+ roc_eswitch_npc_mcam_tx_rule;
+ roc_eswitch_npc_rss_action_configure;
roc_hash_md5_gen;
roc_hash_sha1_gen;
roc_hash_sha256_gen;
@@ -443,6 +447,7 @@ INTERNAL {
roc_npc_flow_dump;
roc_npc_flow_mcam_dump;
roc_npc_flow_parse;
+ roc_npc_get_free_mcam_entry;
roc_npc_get_low_priority_mcam;
roc_npc_init;
roc_npc_kex_capa_get;
@@ -450,6 +455,7 @@ INTERNAL {
roc_npc_mark_actions_sub_return;
roc_npc_vtag_actions_get;
roc_npc_vtag_actions_sub_return;
+ roc_npc_mcam_alloc_counter;
roc_npc_mcam_alloc_entries;
roc_npc_mcam_alloc_entry;
roc_npc_mcam_clear_counter;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 07/23] common/cnxk: interface to update VLAN TPID
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (5 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 06/23] common/cnxk: common NPC changes for eswitch Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 08/23] net/cnxk: eswitch flow configurations Harman Kalra
` (15 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Introducing eswitch variant of set vlan tpid api which can be
using for PF and VF
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_eswitch.c | 15 +++++++++++++++
drivers/common/cnxk/roc_eswitch.h | 4 ++++
drivers/common/cnxk/roc_nix_priv.h | 11 +++++------
drivers/common/cnxk/roc_nix_vlan.c | 23 ++++++++++++++++++-----
drivers/common/cnxk/version.map | 1 +
5 files changed, 43 insertions(+), 11 deletions(-)
diff --git a/drivers/common/cnxk/roc_eswitch.c b/drivers/common/cnxk/roc_eswitch.c
index e480ab1046..020a891a32 100644
--- a/drivers/common/cnxk/roc_eswitch.c
+++ b/drivers/common/cnxk/roc_eswitch.c
@@ -304,3 +304,18 @@ roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc, struct roc_npc_flo
((uint64_t)(rss_grp_idx & NPC_RSS_ACT_GRP_MASK) << NPC_RSS_ACT_GRP_OFFSET);
return 0;
}
+
+int
+roc_eswitch_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t tpid, bool is_vf)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+ int rc;
+
+ /* Configuring for PF/VF */
+ rc = nix_vlan_tpid_set(dev->mbox, dev->pf_func | is_vf, type, tpid);
+ if (rc)
+ plt_err("Failed to set tpid for PF, rc %d", rc);
+
+ return rc;
+}
diff --git a/drivers/common/cnxk/roc_eswitch.h b/drivers/common/cnxk/roc_eswitch.h
index cdbe808a71..34b75d10ac 100644
--- a/drivers/common/cnxk/roc_eswitch.h
+++ b/drivers/common/cnxk/roc_eswitch.h
@@ -19,4 +19,8 @@ int __roc_api roc_eswitch_npc_mcam_delete_rule(struct roc_npc *roc_npc, struct r
int __roc_api roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc,
struct roc_npc_flow *flow, uint32_t flowkey_cfg,
uint16_t *reta_tbl);
+
+/* NIX */
+int __roc_api roc_eswitch_nix_vlan_tpid_set(struct roc_nix *nix, uint32_t type, uint16_t tpid,
+ bool is_vf);
#endif /* __ROC_ESWITCH_H__ */
diff --git a/drivers/common/cnxk/roc_nix_priv.h b/drivers/common/cnxk/roc_nix_priv.h
index a582b9df33..3d99ade2b4 100644
--- a/drivers/common/cnxk/roc_nix_priv.h
+++ b/drivers/common/cnxk/roc_nix_priv.h
@@ -469,13 +469,12 @@ void nix_tm_shaper_profile_free(struct nix_tm_shaper_profile *profile);
uint64_t nix_get_blkaddr(struct dev *dev);
void nix_lf_rq_dump(__io struct nix_cn10k_rq_ctx_s *ctx, FILE *file);
int nix_lf_gen_reg_dump(uintptr_t nix_lf_base, uint64_t *data);
-int nix_lf_stat_reg_dump(uintptr_t nix_lf_base, uint64_t *data,
- uint8_t lf_tx_stats, uint8_t lf_rx_stats);
-int nix_lf_int_reg_dump(uintptr_t nix_lf_base, uint64_t *data, uint16_t qints,
- uint16_t cints);
-int nix_q_ctx_get(struct dev *dev, uint8_t ctype, uint16_t qid,
- __io void **ctx_p);
+int nix_lf_stat_reg_dump(uintptr_t nix_lf_base, uint64_t *data, uint8_t lf_tx_stats,
+ uint8_t lf_rx_stats);
+int nix_lf_int_reg_dump(uintptr_t nix_lf_base, uint64_t *data, uint16_t qints, uint16_t cints);
+int nix_q_ctx_get(struct dev *dev, uint8_t ctype, uint16_t qid, __io void **ctx_p);
uint8_t nix_tm_lbk_relchan_get(struct nix *nix);
+int nix_vlan_tpid_set(struct mbox *mbox, uint16_t pcifunc, uint32_t type, uint16_t tpid);
/*
* Telemetry
diff --git a/drivers/common/cnxk/roc_nix_vlan.c b/drivers/common/cnxk/roc_nix_vlan.c
index abd2eb0571..db218593ad 100644
--- a/drivers/common/cnxk/roc_nix_vlan.c
+++ b/drivers/common/cnxk/roc_nix_vlan.c
@@ -211,18 +211,17 @@ roc_nix_vlan_insert_ena_dis(struct roc_nix *roc_nix,
}
int
-roc_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t tpid)
+nix_vlan_tpid_set(struct mbox *mbox, uint16_t pcifunc, uint32_t type, uint16_t tpid)
{
- struct nix *nix = roc_nix_to_nix_priv(roc_nix);
- struct dev *dev = &nix->dev;
- struct mbox *mbox = mbox_get(dev->mbox);
struct nix_set_vlan_tpid *tpid_cfg;
int rc = -ENOSPC;
- tpid_cfg = mbox_alloc_msg_nix_set_vlan_tpid(mbox);
+ /* Configuring for PF */
+ tpid_cfg = mbox_alloc_msg_nix_set_vlan_tpid(mbox_get(mbox));
if (tpid_cfg == NULL)
goto exit;
tpid_cfg->tpid = tpid;
+ tpid_cfg->hdr.pcifunc = pcifunc;
if (type & ROC_NIX_VLAN_TYPE_OUTER)
tpid_cfg->vlan_type = NIX_VLAN_TYPE_OUTER;
@@ -234,3 +233,17 @@ roc_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t tpid)
mbox_put(mbox);
return rc;
}
+
+int
+roc_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t tpid)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+ int rc;
+
+ rc = nix_vlan_tpid_set(dev->mbox, dev->pf_func, type, tpid);
+ if (rc)
+ plt_err("Failed to set tpid for PF, rc %d", rc);
+
+ return rc;
+}
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index feda34b852..78c421677d 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -91,6 +91,7 @@ INTERNAL {
roc_dpi_disable;
roc_dpi_enable;
roc_error_msg_get;
+ roc_eswitch_nix_vlan_tpid_set;
roc_eswitch_npc_mcam_delete_rule;
roc_eswitch_npc_mcam_rx_rule;
roc_eswitch_npc_mcam_tx_rule;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 08/23] net/cnxk: eswitch flow configurations
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (6 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 07/23] common/cnxk: interface to update VLAN TPID Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 09/23] net/cnxk: eswitch fastpath routines Harman Kalra
` (14 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Adding flow rules for eswitch PF and VF and implementing
interfaces to delete, shift flow rules
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 44 +++
drivers/net/cnxk/cnxk_eswitch.h | 25 +-
drivers/net/cnxk/cnxk_eswitch_devargs.c | 1 +
drivers/net/cnxk/cnxk_eswitch_flow.c | 454 ++++++++++++++++++++++++
drivers/net/cnxk/meson.build | 1 +
5 files changed, 522 insertions(+), 3 deletions(-)
create mode 100644 drivers/net/cnxk/cnxk_eswitch_flow.c
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index 4b2c907f9f..6f651e0857 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -2,11 +2,33 @@
* Copyright(C) 2024 Marvell.
*/
+#include <rte_thash.h>
+
#include <cnxk_eswitch.h>
#include <cnxk_rep.h>
#define CNXK_NIX_DEF_SQ_COUNT 512
+struct cnxk_esw_repr_hw_info *
+cnxk_eswitch_representor_hw_info(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func)
+{
+ struct cnxk_eswitch_devargs *esw_da;
+ int i, j;
+
+ if (!eswitch_dev)
+ return NULL;
+
+ /* Traversing the initialized represented list */
+ for (i = 0; i < eswitch_dev->nb_esw_da; i++) {
+ esw_da = &eswitch_dev->esw_da[i];
+ for (j = 0; j < esw_da->nb_repr_ports; j++) {
+ if (esw_da->repr_hw_info[j].hw_func == hw_func)
+ return &esw_da->repr_hw_info[j];
+ }
+ }
+ return NULL;
+}
+
static int
eswitch_hw_rsrc_cleanup(struct cnxk_eswitch_dev *eswitch_dev, struct rte_pci_device *pci_dev)
{
@@ -67,6 +89,10 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
if (eswitch_dev->repr_cnt.nb_repr_created)
cnxk_rep_dev_remove(eswitch_dev);
+ /* Cleanup NPC rxtx flow rules */
+ cnxk_eswitch_flow_rules_remove_list(eswitch_dev, &eswitch_dev->esw_flow_list,
+ eswitch_dev->npc.pf_func);
+
/* Cleanup HW resources */
eswitch_hw_rsrc_cleanup(eswitch_dev, pci_dev);
@@ -87,6 +113,21 @@ cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev)
goto done;
}
+ /* Install eswitch PF mcam rules */
+ rc = cnxk_eswitch_pfvf_flow_rules_install(eswitch_dev, false);
+ if (rc) {
+ plt_err("Failed to install rxtx rules, rc %d", rc);
+ goto done;
+ }
+
+ /* Configure TPID for Eswitch PF LFs */
+ rc = roc_eswitch_nix_vlan_tpid_set(&eswitch_dev->nix, ROC_NIX_VLAN_TYPE_OUTER,
+ CNXK_ESWITCH_VLAN_TPID, false);
+ if (rc) {
+ plt_err("Failed to configure tpid, rc %d", rc);
+ goto done;
+ }
+
rc = roc_npc_mcam_enable_all_entries(&eswitch_dev->npc, 1);
if (rc) {
plt_err("Failed to enable NPC entries %d", rc);
@@ -531,6 +572,9 @@ eswitch_hw_rsrc_setup(struct cnxk_eswitch_dev *eswitch_dev, struct rte_pci_devic
if (rc)
goto rsrc_cleanup;
+ /* List for eswitch default flows */
+ TAILQ_INIT(&eswitch_dev->esw_flow_list);
+
return rc;
rsrc_cleanup:
eswitch_hw_rsrc_cleanup(eswitch_dev, pci_dev);
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index dcd5add6d0..5b4e1b0a71 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -13,11 +13,10 @@
#include "cn10k_tx.h"
#define CNXK_ESWITCH_CTRL_MSG_SOCK_PATH "/tmp/cxk_rep_ctrl_msg_sock"
+#define CNXK_ESWITCH_VLAN_TPID ROC_ESWITCH_VLAN_TPID
#define CNXK_REP_ESWITCH_DEV_MZ "cnxk_eswitch_dev"
-#define CNXK_ESWITCH_VLAN_TPID 0x8100
#define CNXK_ESWITCH_MAX_TXQ 256
#define CNXK_ESWITCH_MAX_RXQ 256
-#define CNXK_ESWITCH_LBK_CHAN 63
#define CNXK_ESWITCH_VFPF_SHIFT 8
#define CNXK_ESWITCH_QUEUE_STATE_RELEASED 0
@@ -25,6 +24,7 @@
#define CNXK_ESWITCH_QUEUE_STATE_STARTED 2
#define CNXK_ESWITCH_QUEUE_STATE_STOPPED 3
+TAILQ_HEAD(eswitch_flow_list, roc_npc_flow);
enum cnxk_esw_da_pattern_type {
CNXK_ESW_DA_TYPE_LIST = 0,
CNXK_ESW_DA_TYPE_PFVF,
@@ -39,6 +39,9 @@ struct cnxk_esw_repr_hw_info {
uint16_t pfvf;
/* representor port id assigned to representee */
uint16_t port_id;
+ uint16_t num_flow_entries;
+
+ TAILQ_HEAD(flow_list, roc_npc_flow) repr_flow_list;
};
/* Structure representing per devarg information - this can be per representee
@@ -90,7 +93,6 @@ struct cnxk_eswitch_cxq {
uint8_t state;
};
-TAILQ_HEAD(eswitch_flow_list, roc_npc_flow);
struct cnxk_eswitch_dev {
/* Input parameters */
struct plt_pci_device *pci_dev;
@@ -116,6 +118,13 @@ struct cnxk_eswitch_dev {
uint16_t rep_cnt;
uint8_t configured;
+ /* NPC rxtx rules */
+ struct flow_list esw_flow_list;
+ uint16_t num_entries;
+ bool eswitch_vf_rules_setup;
+ uint16_t esw_pf_entry;
+ uint16_t esw_vf_entry;
+
/* Eswitch Representors Devargs */
uint16_t nb_esw_da;
uint16_t last_probed;
@@ -144,7 +153,10 @@ cnxk_eswitch_pmd_priv(void)
return mz->addr;
}
+/* HW Resources */
int cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev);
+struct cnxk_esw_repr_hw_info *cnxk_eswitch_representor_hw_info(struct cnxk_eswitch_dev *eswitch_dev,
+ uint16_t hw_func);
int cnxk_eswitch_repr_devargs(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
int cnxk_eswitch_representor_info_get(struct cnxk_eswitch_dev *eswitch_dev,
struct rte_eth_representor_info *info);
@@ -158,4 +170,11 @@ int cnxk_eswitch_rxq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
int cnxk_eswitch_rxq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
int cnxk_eswitch_txq_start(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
int cnxk_eswitch_txq_stop(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid);
+/* Flow Rules */
+int cnxk_eswitch_flow_rules_install(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func);
+int cnxk_eswitch_flow_rules_delete(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func);
+int cnxk_eswitch_pfvf_flow_rules_install(struct cnxk_eswitch_dev *eswitch_dev, bool is_vf);
+int cnxk_eswitch_flow_rule_shift(uint16_t hw_func, uint16_t *new_entry);
+int cnxk_eswitch_flow_rules_remove_list(struct cnxk_eswitch_dev *eswitch_dev,
+ struct flow_list *list, uint16_t hw_func);
#endif /* __CNXK_ESWITCH_H__ */
diff --git a/drivers/net/cnxk/cnxk_eswitch_devargs.c b/drivers/net/cnxk/cnxk_eswitch_devargs.c
index b5f28e185c..19165d1357 100644
--- a/drivers/net/cnxk/cnxk_eswitch_devargs.c
+++ b/drivers/net/cnxk/cnxk_eswitch_devargs.c
@@ -72,6 +72,7 @@ populate_repr_hw_info(struct cnxk_eswitch_dev *eswitch_dev, struct rte_eth_devar
esw_da->repr_hw_info[i].pfvf = (eth_da->type == RTE_ETH_REPRESENTOR_PF) ?
eth_da->ports[0] :
eth_da->representor_ports[i];
+ TAILQ_INIT(&esw_da->repr_hw_info[i].repr_flow_list);
plt_esw_dbg(" HW func %x index %d type %d", hw_func, j, eth_da->type);
}
diff --git a/drivers/net/cnxk/cnxk_eswitch_flow.c b/drivers/net/cnxk/cnxk_eswitch_flow.c
new file mode 100644
index 0000000000..06077bfe92
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_eswitch_flow.c
@@ -0,0 +1,454 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#include <rte_thash.h>
+
+#include <cnxk_eswitch.h>
+
+const uint8_t eswitch_vlan_rss_key[ROC_NIX_RSS_KEY_LEN] = {
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE,
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE,
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE,
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE,
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE,
+ 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE};
+
+int
+cnxk_eswitch_flow_rules_remove_list(struct cnxk_eswitch_dev *eswitch_dev, struct flow_list *list,
+ uint16_t hw_func)
+{
+ struct roc_npc_flow *flow, *tvar;
+ int rc = 0;
+
+ RTE_TAILQ_FOREACH_SAFE(flow, list, next, tvar) {
+ plt_esw_dbg("Removing flow %d", flow->mcam_id);
+ rc = roc_eswitch_npc_mcam_delete_rule(&eswitch_dev->npc, flow,
+ hw_func);
+ if (rc)
+ plt_err("Failed to delete rule %d", flow->mcam_id);
+ rc = roc_npc_mcam_free(&eswitch_dev->npc, flow);
+ if (rc)
+ plt_err("Failed to free entry %d", flow->mcam_id);
+ TAILQ_REMOVE(list, flow, next);
+ rte_free(flow);
+ }
+
+ return rc;
+}
+
+static int
+eswitch_npc_vlan_rss_configure(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
+{
+ struct roc_nix *roc_nix = roc_npc->roc_nix;
+ uint32_t qid, idx, hash, vlan_tci;
+ uint16_t *reta, reta_sz, id;
+ int rc = 0;
+
+ id = flow->mcam_id;
+ /* Setting up the key */
+ roc_nix_rss_key_set(roc_nix, eswitch_vlan_rss_key);
+
+ reta_sz = roc_nix->reta_sz;
+ reta = plt_zmalloc(reta_sz * sizeof(uint16_t), 0);
+ if (!reta) {
+ plt_err("Failed to allocate mem for reta table");
+ rc = -ENOMEM;
+ goto fail;
+ }
+ for (qid = 0; qid < reta_sz; qid++) {
+ vlan_tci = (1 << CNXK_ESWITCH_VFPF_SHIFT) | qid;
+ hash = rte_softrss(&vlan_tci, 1, eswitch_vlan_rss_key);
+ idx = hash & 0xFF;
+ reta[idx] = qid;
+ }
+ flow->mcam_id = id;
+ rc = roc_eswitch_npc_rss_action_configure(roc_npc, flow, FLOW_KEY_TYPE_VLAN, reta);
+ if (rc) {
+ plt_err("Failed to configure rss action, err %d", rc);
+ goto done;
+ }
+
+done:
+ plt_free(reta);
+fail:
+ return rc;
+}
+
+static int
+eswitch_pfvf_mcam_install_rules(struct cnxk_eswitch_dev *eswitch_dev, struct roc_npc_flow *flow,
+ bool is_vf)
+{
+ uint16_t vlan_tci = 0, hw_func;
+ int rc;
+
+ hw_func = eswitch_dev->npc.pf_func | is_vf;
+ if (!is_vf) {
+ /* Eswitch PF RX VLAN rule */
+ vlan_tci = 1ULL << CNXK_ESWITCH_VFPF_SHIFT;
+ rc = roc_eswitch_npc_mcam_rx_rule(&eswitch_dev->npc, flow, hw_func, vlan_tci,
+ 0xFF00);
+ if (rc) {
+ plt_err("Failed to install RX rule for ESW PF to ESW VF, rc %d", rc);
+ goto exit;
+ }
+ plt_esw_dbg("Installed eswitch PF RX rule %d", flow->mcam_id);
+ rc = eswitch_npc_vlan_rss_configure(&eswitch_dev->npc, flow);
+ if (rc)
+ goto exit;
+ flow->enable = true;
+ } else {
+ /* Eswitch VF RX VLAN rule */
+ rc = roc_eswitch_npc_mcam_rx_rule(&eswitch_dev->npc, flow, hw_func, vlan_tci,
+ 0xFF00);
+ if (rc) {
+ plt_err("Failed to install RX rule for ESW VF to ESW PF, rc %d", rc);
+ goto exit;
+ }
+ flow->enable = true;
+ plt_esw_dbg("Installed eswitch PF RX rule %d", flow->mcam_id);
+ }
+
+ return 0;
+exit:
+ return rc;
+}
+
+static int
+eswitch_npc_get_counter(struct roc_npc *npc, struct roc_npc_flow *flow)
+{
+ uint16_t ctr_id;
+ int rc;
+
+ rc = roc_npc_mcam_alloc_counter(npc, &ctr_id);
+ if (rc < 0) {
+ plt_err("Failed to allocate counter, rc %d", rc);
+ goto fail;
+ }
+ flow->ctr_id = ctr_id;
+ flow->use_ctr = true;
+
+ rc = roc_npc_mcam_clear_counter(npc, flow->ctr_id);
+ if (rc < 0) {
+ plt_err("Failed to clear counter idx %d, rc %d", flow->ctr_id, rc);
+ goto free;
+ }
+ return 0;
+free:
+ roc_npc_mcam_free_counter(npc, ctr_id);
+fail:
+ return rc;
+}
+
+static int
+eswitch_npc_get_counter_entry_ref(struct roc_npc *npc, struct roc_npc_flow *flow,
+ struct roc_npc_flow *ref_flow)
+{
+ int rc = 0, resp_count;
+
+ rc = eswitch_npc_get_counter(npc, flow);
+ if (rc)
+ goto free;
+
+ /* Allocate an entry viz higher priority than ref flow */
+ rc = roc_npc_mcam_alloc_entry(npc, flow, ref_flow, NPC_MCAM_HIGHER_PRIO, &resp_count);
+ if (rc) {
+ plt_err("Failed to allocate entry, err %d", rc);
+ goto free;
+ }
+ plt_esw_dbg("New entry %d ref entry %d resp_count %d", flow->mcam_id, ref_flow->mcam_id,
+ resp_count);
+
+ return 0;
+free:
+ roc_npc_mcam_free_counter(npc, flow->ctr_id);
+ return rc;
+}
+
+int
+cnxk_eswitch_flow_rule_shift(uint16_t hw_func, uint16_t *entry)
+{
+ struct cnxk_esw_repr_hw_info *repr_info;
+ struct cnxk_eswitch_dev *eswitch_dev;
+ struct roc_npc_flow *ref_flow, *flow;
+ uint16_t curr_entry, new_entry;
+ int rc = 0, resp_count;
+
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ if (!eswitch_dev) {
+ plt_err("Invalid eswitch_dev handle");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ repr_info = cnxk_eswitch_representor_hw_info(eswitch_dev, hw_func);
+ if (!repr_info) {
+ plt_warn("Failed to get representor group for %x", hw_func);
+ rc = -ENOENT;
+ goto fail;
+ }
+
+ ref_flow = TAILQ_FIRST(&repr_info->repr_flow_list);
+ if (*entry > ref_flow->mcam_id) {
+ flow = plt_zmalloc(sizeof(struct roc_npc_flow), 0);
+ if (!flow) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ /* Allocate a higher priority flow rule */
+ rc = roc_npc_mcam_alloc_entry(&eswitch_dev->npc, flow, ref_flow,
+ NPC_MCAM_HIGHER_PRIO, &resp_count);
+ if (rc < 0) {
+ plt_err("Failed to allocate a newmcam entry, rc %d", rc);
+ goto fail;
+ }
+
+ if (flow->mcam_id > ref_flow->mcam_id) {
+ plt_err("New flow %d is still at higher priority than ref_flow %d",
+ flow->mcam_id, ref_flow->mcam_id);
+ rc = -EINVAL;
+ goto free_entry;
+ }
+
+ plt_info("Before shift: HW_func %x curr_entry %d ref flow id %d new_entry %d",
+ hw_func, *entry, ref_flow->mcam_id, flow->mcam_id);
+
+ curr_entry = *entry;
+ new_entry = flow->mcam_id;
+
+ rc = roc_npc_mcam_move(&eswitch_dev->npc, curr_entry, new_entry);
+ if (rc) {
+ plt_err("Failed to shift the new index %d to curr index %d, err %d", *entry,
+ curr_entry, rc);
+ goto free_entry;
+ }
+ *entry = flow->mcam_id;
+
+ /* Freeing the current entry */
+ rc = roc_npc_mcam_free_entry(&eswitch_dev->npc, curr_entry);
+ if (rc) {
+ plt_err("Failed to free the old entry. err %d", rc);
+ goto free_entry;
+ }
+
+ plt_free(flow);
+ plt_info("After shift: HW_func %x old_entry %d new_entry %d", hw_func, curr_entry,
+ *entry);
+ }
+
+ return 0;
+free_entry:
+
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_flow_rules_delete(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func)
+{
+ struct cnxk_esw_repr_hw_info *repr_info;
+ struct flow_list *list;
+ int rc = 0;
+
+ repr_info = cnxk_eswitch_representor_hw_info(eswitch_dev, hw_func);
+ if (!repr_info) {
+ plt_warn("Failed to get representor group for %x", hw_func);
+ rc = -ENOENT;
+ goto fail;
+ }
+ list = &repr_info->repr_flow_list;
+
+ plt_esw_dbg("Deleting flows for %x", hw_func);
+ rc = cnxk_eswitch_flow_rules_remove_list(eswitch_dev, list, hw_func);
+ if (rc)
+ plt_err("Failed to delete rules for hw func %x", hw_func);
+
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_flow_rules_install(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func)
+{
+ struct roc_npc_flow *rx_flow, *tx_flow, *flow_iter, *esw_pf_flow = NULL;
+ struct cnxk_esw_repr_hw_info *repr_info;
+ struct flow_list *list;
+ uint16_t vlan_tci;
+ int rc = 0;
+
+ repr_info = cnxk_eswitch_representor_hw_info(eswitch_dev, hw_func);
+ if (!repr_info) {
+ plt_err("Failed to get representor group for %x", hw_func);
+ rc = -EINVAL;
+ goto fail;
+ }
+ list = &repr_info->repr_flow_list;
+
+ /* Taking ESW PF as reference entry for installing new rules */
+ TAILQ_FOREACH(flow_iter, &eswitch_dev->esw_flow_list, next) {
+ if (flow_iter->mcam_id == eswitch_dev->esw_pf_entry) {
+ esw_pf_flow = flow_iter;
+ break;
+ }
+ }
+
+ if (!esw_pf_flow) {
+ plt_err("Failed to get the ESW PF flow");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Installing RX rule */
+ rx_flow = plt_zmalloc(sizeof(struct roc_npc_flow), 0);
+ if (!rx_flow) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ rc = eswitch_npc_get_counter_entry_ref(&eswitch_dev->npc, rx_flow, esw_pf_flow);
+ if (rc) {
+ plt_err("Failed to get counter and mcam entry, rc %d", rc);
+ goto free_rx_flow;
+ }
+
+ /* VLAN TCI value for this representee is the rep id from AF driver */
+ vlan_tci = repr_info->rep_id;
+ rc = roc_eswitch_npc_mcam_rx_rule(&eswitch_dev->npc, rx_flow, hw_func, vlan_tci, 0xFFFF);
+ if (rc) {
+ plt_err("Failed to install RX rule for ESW PF to ESW VF, rc %d", rc);
+ goto free_rx_entry;
+ }
+ rx_flow->enable = true;
+ /* List in ascending order of mcam entries */
+ TAILQ_FOREACH(flow_iter, list, next) {
+ if (flow_iter->mcam_id > rx_flow->mcam_id) {
+ TAILQ_INSERT_BEFORE(flow_iter, rx_flow, next);
+ goto done_rx;
+ }
+ }
+ TAILQ_INSERT_TAIL(list, rx_flow, next);
+done_rx:
+ repr_info->num_flow_entries++;
+ plt_esw_dbg("Installed RX flow rule %d for representee %x with vlan tci %x MCAM id %d",
+ eswitch_dev->num_entries, hw_func, vlan_tci, rx_flow->mcam_id);
+
+ /* Installing TX rule */
+ tx_flow = plt_zmalloc(sizeof(struct roc_npc_flow), 0);
+ if (!tx_flow) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto remove_rx_rule;
+ }
+
+ rc = eswitch_npc_get_counter_entry_ref(&eswitch_dev->npc, tx_flow, esw_pf_flow);
+ if (rc) {
+ plt_err("Failed to get counter and mcam entry, rc %d", rc);
+ goto free_tx_flow;
+ }
+
+ vlan_tci = (1ULL << CNXK_ESWITCH_VFPF_SHIFT) | repr_info->rep_id;
+ rc = roc_eswitch_npc_mcam_tx_rule(&eswitch_dev->npc, tx_flow, hw_func, vlan_tci);
+ if (rc) {
+ plt_err("Failed to install RX rule for ESW PF to ESW VF, rc %d", rc);
+ goto free_tx_entry;
+ }
+ tx_flow->enable = true;
+ /* List in ascending order of mcam entries */
+ TAILQ_FOREACH(flow_iter, list, next) {
+ if (flow_iter->mcam_id > tx_flow->mcam_id) {
+ TAILQ_INSERT_BEFORE(flow_iter, tx_flow, next);
+ goto done_tx;
+ }
+ }
+ TAILQ_INSERT_TAIL(list, tx_flow, next);
+done_tx:
+ repr_info->num_flow_entries++;
+ plt_esw_dbg("Installed TX flow rule %d for representee %x with vlan tci %x MCAM id %d",
+ repr_info->num_flow_entries, hw_func, vlan_tci, tx_flow->mcam_id);
+
+ return 0;
+free_tx_entry:
+ roc_npc_mcam_free(&eswitch_dev->npc, tx_flow);
+free_tx_flow:
+ rte_free(tx_flow);
+remove_rx_rule:
+ TAILQ_REMOVE(list, rx_flow, next);
+free_rx_entry:
+ roc_npc_mcam_free(&eswitch_dev->npc, rx_flow);
+free_rx_flow:
+ rte_free(rx_flow);
+fail:
+ return rc;
+}
+
+int
+cnxk_eswitch_pfvf_flow_rules_install(struct cnxk_eswitch_dev *eswitch_dev, bool is_vf)
+{
+ struct roc_npc_flow *flow, *flow_iter;
+ struct flow_list *list;
+ int rc = 0;
+
+ list = &eswitch_dev->esw_flow_list;
+ flow = plt_zmalloc(sizeof(struct roc_npc_flow), 0);
+ if (!flow) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ rc = eswitch_npc_get_counter(&eswitch_dev->npc, flow);
+ if (rc) {
+ plt_err("Failed to get counter and mcam entry, rc %d", rc);
+ goto free_flow;
+ }
+ if (!is_vf) {
+ /* Reserving an entry for esw VF but will not be installed */
+ rc = roc_npc_get_free_mcam_entry(&eswitch_dev->npc, flow);
+ if (rc < 0) {
+ plt_err("Failed to allocate entry for vf, err %d", rc);
+ goto free_flow;
+ }
+ eswitch_dev->esw_vf_entry = flow->mcam_id;
+ /* Allocate an entry for esw PF */
+ rc = eswitch_npc_get_counter_entry_ref(&eswitch_dev->npc, flow, flow);
+ if (rc) {
+ plt_err("Failed to allocate entry for pf, err %d", rc);
+ goto free_flow;
+ }
+ eswitch_dev->esw_pf_entry = flow->mcam_id;
+ plt_esw_dbg("Allocated entries for esw: PF %d and VF %d", eswitch_dev->esw_pf_entry,
+ eswitch_dev->esw_vf_entry);
+ } else {
+ flow->mcam_id = eswitch_dev->esw_vf_entry;
+ }
+
+ rc = eswitch_pfvf_mcam_install_rules(eswitch_dev, flow, is_vf);
+ if (rc) {
+ plt_err("Failed to install entries, rc %d", rc);
+ goto free_flow;
+ }
+
+ /* List in ascending order of mcam entries */
+ TAILQ_FOREACH(flow_iter, list, next) {
+ if (flow_iter->mcam_id > flow->mcam_id) {
+ TAILQ_INSERT_BEFORE(flow_iter, flow, next);
+ goto done;
+ }
+ }
+ TAILQ_INSERT_TAIL(list, flow, next);
+done:
+ eswitch_dev->num_entries++;
+ plt_esw_dbg("Installed new eswitch flow rule %d with MCAM id %d", eswitch_dev->num_entries,
+ flow->mcam_id);
+
+ return 0;
+
+free_flow:
+ cnxk_eswitch_flow_rules_remove_list(eswitch_dev, &eswitch_dev->esw_flow_list,
+ eswitch_dev->npc.pf_func);
+fail:
+ return rc;
+}
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index fcd5d3d569..488e89253d 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -30,6 +30,7 @@ sources = files(
'cnxk_ethdev_sec_telemetry.c',
'cnxk_eswitch.c',
'cnxk_eswitch_devargs.c',
+ 'cnxk_eswitch_flow.c',
'cnxk_link.c',
'cnxk_lookup.c',
'cnxk_ptp.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 09/23] net/cnxk: eswitch fastpath routines
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (7 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 08/23] net/cnxk: eswitch flow configurations Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 10/23] net/cnxk: add representor control plane Harman Kalra
` (13 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Implementing fastpath RX and TX fast path routines which can be
invoked from respective representors rx burst and tx burst
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.h | 5 +
drivers/net/cnxk/cnxk_eswitch_rxtx.c | 211 +++++++++++++++++++++++++++
drivers/net/cnxk/meson.build | 1 +
3 files changed, 217 insertions(+)
create mode 100644 drivers/net/cnxk/cnxk_eswitch_rxtx.c
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index 5b4e1b0a71..4edfa91bdc 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -177,4 +177,9 @@ int cnxk_eswitch_pfvf_flow_rules_install(struct cnxk_eswitch_dev *eswitch_dev, b
int cnxk_eswitch_flow_rule_shift(uint16_t hw_func, uint16_t *new_entry);
int cnxk_eswitch_flow_rules_remove_list(struct cnxk_eswitch_dev *eswitch_dev,
struct flow_list *list, uint16_t hw_func);
+/* RX TX fastpath routines */
+uint16_t cnxk_eswitch_dev_tx_burst(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid,
+ struct rte_mbuf **pkts, uint16_t nb_tx, const uint16_t flags);
+uint16_t cnxk_eswitch_dev_rx_burst(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid,
+ struct rte_mbuf **pkts, uint16_t nb_pkts);
#endif /* __CNXK_ESWITCH_H__ */
diff --git a/drivers/net/cnxk/cnxk_eswitch_rxtx.c b/drivers/net/cnxk/cnxk_eswitch_rxtx.c
new file mode 100644
index 0000000000..d57e32b091
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_eswitch_rxtx.c
@@ -0,0 +1,211 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#include <cnxk_eswitch.h>
+
+static __rte_always_inline struct rte_mbuf *
+eswitch_nix_get_mbuf_from_cqe(void *cq, const uint64_t data_off)
+{
+ rte_iova_t buff;
+
+ /* Skip CQE, NIX_RX_PARSE_S and SG HDR(9 DWORDs) and peek buff addr */
+ buff = *((rte_iova_t *)((uint64_t *)cq + 9));
+ return (struct rte_mbuf *)(buff - data_off);
+}
+
+static inline uint64_t
+eswitch_nix_rx_nb_pkts(struct roc_nix_cq *cq, const uint64_t wdata, const uint32_t qmask)
+{
+ uint64_t reg, head, tail;
+ uint32_t available;
+
+ /* Update the available count if cached value is not enough */
+
+ /* Use LDADDA version to avoid reorder */
+ reg = roc_atomic64_add_sync(wdata, cq->status);
+ /* CQ_OP_STATUS operation error */
+ if (reg & BIT_ULL(NIX_CQ_OP_STAT_OP_ERR) || reg & BIT_ULL(NIX_CQ_OP_STAT_CQ_ERR))
+ return 0;
+
+ tail = reg & 0xFFFFF;
+ head = (reg >> 20) & 0xFFFFF;
+ if (tail < head)
+ available = tail - head + qmask + 1;
+ else
+ available = tail - head;
+
+ return available;
+}
+
+static inline void
+nix_cn9k_xmit_one(uint64_t *cmd, void *lmt_addr, const plt_iova_t io_addr)
+{
+ uint64_t lmt_status;
+
+ do {
+ roc_lmt_mov(lmt_addr, cmd, 0);
+ lmt_status = roc_lmt_submit_ldeor(io_addr);
+ } while (lmt_status == 0);
+}
+
+uint16_t
+cnxk_eswitch_dev_tx_burst(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid,
+ struct rte_mbuf **pkts, uint16_t nb_xmit, const uint16_t flags)
+{
+ struct roc_nix_sq *sq = &eswitch_dev->txq[qid].sqs;
+ struct roc_nix_rq *rq = &eswitch_dev->rxq[qid].rqs;
+ uint64_t aura_handle, cmd[6], data = 0;
+ uint16_t lmt_id, pkt = 0, nb_tx = 0;
+ struct nix_send_ext_s *send_hdr_ext;
+ struct nix_send_hdr_s *send_hdr;
+ uint16_t vlan_tci = qid;
+ union nix_send_sg_s *sg;
+ uintptr_t lmt_base, pa;
+ int64_t fc_pkts, dw_m1;
+ rte_iova_t io_addr;
+
+ if (unlikely(eswitch_dev->txq[qid].state != CNXK_ESWITCH_QUEUE_STATE_STARTED))
+ return 0;
+
+ lmt_base = sq->roc_nix->lmt_base;
+ io_addr = sq->io_addr;
+ aura_handle = rq->aura_handle;
+ /* Get LMT base address and LMT ID as per thread ID */
+ lmt_id = roc_plt_control_lmt_id_get();
+ lmt_base += ((uint64_t)lmt_id << ROC_LMT_LINE_SIZE_LOG2);
+ /* Double word minus 1: LMTST size-1 in units of 128 bits */
+ /* 2(HDR) + 2(EXT_HDR) + 1(SG) + 1(IOVA) = 6/2 - 1 = 2 */
+ dw_m1 = cn10k_nix_tx_ext_subs(flags) + 1;
+
+ memset(cmd, 0, sizeof(cmd));
+ send_hdr = (struct nix_send_hdr_s *)&cmd[0];
+ send_hdr->w0.sizem1 = dw_m1;
+ send_hdr->w0.sq = sq->qid;
+
+ if (dw_m1 >= 2) {
+ send_hdr_ext = (struct nix_send_ext_s *)&cmd[2];
+ send_hdr_ext->w0.subdc = NIX_SUBDC_EXT;
+ if (flags & NIX_TX_OFFLOAD_VLAN_QINQ_F) {
+ send_hdr_ext->w1.vlan0_ins_ena = true;
+ /* 2B before end of l2 header */
+ send_hdr_ext->w1.vlan0_ins_ptr = 12;
+ send_hdr_ext->w1.vlan0_ins_tci = 0;
+ }
+ sg = (union nix_send_sg_s *)&cmd[4];
+ } else {
+ sg = (union nix_send_sg_s *)&cmd[2];
+ }
+
+ sg->subdc = NIX_SUBDC_SG;
+ sg->segs = 1;
+ sg->ld_type = NIX_SENDLDTYPE_LDD;
+
+ /* Tx */
+ fc_pkts = ((int64_t)sq->nb_sqb_bufs_adj - *((uint64_t *)sq->fc)) << sq->sqes_per_sqb_log2;
+
+ if (fc_pkts < 0)
+ nb_tx = 0;
+ else
+ nb_tx = PLT_MIN(nb_xmit, (uint64_t)fc_pkts);
+
+ for (pkt = 0; pkt < nb_tx; pkt++) {
+ send_hdr->w0.total = pkts[pkt]->pkt_len;
+ if (pkts[pkt]->pool) {
+ aura_handle = pkts[pkt]->pool->pool_id;
+ send_hdr->w0.aura = roc_npa_aura_handle_to_aura(aura_handle);
+ } else {
+ send_hdr->w0.df = 1;
+ }
+ if (dw_m1 >= 2 && flags & NIX_TX_OFFLOAD_VLAN_QINQ_F)
+ send_hdr_ext->w1.vlan0_ins_tci = vlan_tci;
+ sg->seg1_size = pkts[pkt]->pkt_len;
+ *(plt_iova_t *)(sg + 1) = rte_mbuf_data_iova(pkts[pkt]);
+
+ plt_esw_dbg("Transmitting pkt %d (%p) vlan tci %x on sq %d esw qid %d", pkt,
+ pkts[pkt], vlan_tci, sq->qid, qid);
+ if (roc_model_is_cn9k()) {
+ nix_cn9k_xmit_one(cmd, sq->lmt_addr, sq->io_addr);
+ } else {
+ cn10k_nix_xmit_mv_lmt_base(lmt_base, cmd, flags);
+ /* PA<6:4> = LMTST size-1 in units of 128 bits. Size of the first LMTST in
+ * burst.
+ */
+ pa = io_addr | (dw_m1 << 4);
+ data &= ~0x7ULL;
+ /*<15:12> = CNTM1: Count minus one of LMTSTs in the burst */
+ data = (0ULL << 12);
+ /* *<10:0> = LMT_ID: Identifies which LMT line is used for the first LMTST
+ */
+ data |= (uint64_t)lmt_id;
+
+ /* STEOR0 */
+ roc_lmt_submit_steorl(data, pa);
+ rte_io_wmb();
+ }
+ }
+
+ return nb_tx;
+}
+
+uint16_t
+cnxk_eswitch_dev_rx_burst(struct cnxk_eswitch_dev *eswitch_dev, uint16_t qid,
+ struct rte_mbuf **pkts, uint16_t nb_pkts)
+{
+ struct roc_nix_rq *rq = &eswitch_dev->rxq[qid].rqs;
+ struct roc_nix_cq *cq = &eswitch_dev->cxq[qid].cqs;
+ const union nix_rx_parse_u *rx;
+ struct nix_cqe_hdr_s *cqe;
+ uint64_t pkt = 0, nb_rx;
+ struct rte_mbuf *mbuf;
+ uint64_t wdata;
+ uint32_t qmask;
+ uintptr_t desc;
+ uint32_t head;
+
+ if (unlikely(eswitch_dev->rxq[qid].state != CNXK_ESWITCH_QUEUE_STATE_STARTED))
+ return 0;
+
+ wdata = cq->wdata;
+ qmask = cq->qmask;
+ desc = (uintptr_t)cq->desc_base;
+ nb_rx = eswitch_nix_rx_nb_pkts(cq, wdata, qmask);
+ nb_rx = RTE_MIN(nb_rx, nb_pkts);
+ head = cq->head;
+
+ /* Nothing to receive */
+ if (!nb_rx)
+ return 0;
+
+ /* Rx */
+ for (pkt = 0; pkt < nb_rx; pkt++) {
+ /* Prefetch N desc ahead */
+ rte_prefetch_non_temporal((void *)(desc + (CQE_SZ((head + 2) & qmask))));
+ cqe = (struct nix_cqe_hdr_s *)(desc + CQE_SZ(head));
+ rx = (const union nix_rx_parse_u *)((const uint64_t *)cqe + 1);
+
+ /* Skip QE, NIX_RX_PARSE_S and SG HDR(9 DWORDs) and peek buff addr */
+ mbuf = eswitch_nix_get_mbuf_from_cqe(cqe, rq->first_skip);
+ mbuf->pkt_len = rx->pkt_lenm1 + 1;
+ mbuf->data_len = rx->pkt_lenm1 + 1;
+ mbuf->data_off = 128;
+ /* Rx parse to capture vlan info */
+ if (rx->vtag0_valid)
+ mbuf->vlan_tci = rx->vtag0_tci;
+ /* Populate RSS hash */
+ mbuf->hash.rss = cqe->tag;
+ mbuf->ol_flags |= RTE_MBUF_F_RX_RSS_HASH;
+ pkts[pkt] = mbuf;
+ roc_prefetch_store_keep(mbuf);
+ plt_esw_dbg("Packet %d rec on queue %d esw qid %d hash %x mbuf %p vlan tci %d",
+ (uint32_t)pkt, rq->qid, qid, mbuf->hash.rss, mbuf, mbuf->vlan_tci);
+ head++;
+ head &= qmask;
+ }
+
+ /* Free all the CQs that we've processed */
+ rte_write64_relaxed((wdata | nb_rx), (void *)cq->door);
+ cq->head = head;
+
+ return nb_rx;
+}
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index 488e89253d..7121845dc6 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -31,6 +31,7 @@ sources = files(
'cnxk_eswitch.c',
'cnxk_eswitch_devargs.c',
'cnxk_eswitch_flow.c',
+ 'cnxk_eswitch_rxtx.c',
'cnxk_link.c',
'cnxk_lookup.c',
'cnxk_ptp.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 10/23] net/cnxk: add representor control plane
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (8 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 09/23] net/cnxk: eswitch fastpath routines Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 11/23] common/cnxk: representee notification callback Harman Kalra
` (12 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Implementing the control path for representor ports, where represented
ports can be configured using TLV messaging.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 70 ++-
drivers/net/cnxk/cnxk_eswitch.h | 8 +
drivers/net/cnxk/cnxk_rep.c | 52 ++
drivers/net/cnxk/cnxk_rep.h | 3 +
drivers/net/cnxk/cnxk_rep_msg.c | 827 ++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 95 ++++
drivers/net/cnxk/meson.build | 1 +
7 files changed, 1048 insertions(+), 8 deletions(-)
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.c
create mode 100644 drivers/net/cnxk/cnxk_rep_msg.h
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index 6f651e0857..ad6834410d 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -9,6 +9,27 @@
#define CNXK_NIX_DEF_SQ_COUNT 512
+int
+cnxk_eswitch_representor_id(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func,
+ uint16_t *rep_id)
+{
+ struct cnxk_esw_repr_hw_info *repr_info;
+ int rc = 0;
+
+ repr_info = cnxk_eswitch_representor_hw_info(eswitch_dev, hw_func);
+ if (!repr_info) {
+ plt_warn("Failed to get representor group for %x", hw_func);
+ rc = -ENOENT;
+ goto fail;
+ }
+
+ *rep_id = repr_info->rep_id;
+
+ return 0;
+fail:
+ return rc;
+}
+
struct cnxk_esw_repr_hw_info *
cnxk_eswitch_representor_hw_info(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func)
{
@@ -86,8 +107,41 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
}
/* Remove representor devices associated with PF */
- if (eswitch_dev->repr_cnt.nb_repr_created)
+ if (eswitch_dev->repr_cnt.nb_repr_created) {
+ /* Exiting the rep msg ctrl thread */
+ if (eswitch_dev->start_ctrl_msg_thrd) {
+ uint32_t sunlen;
+ struct sockaddr_un sun = {0};
+ int sock_fd = 0;
+
+ eswitch_dev->start_ctrl_msg_thrd = false;
+ if (!eswitch_dev->client_connected) {
+ plt_esw_dbg("Establishing connection for teardown");
+ sock_fd = socket(AF_UNIX, SOCK_STREAM, 0);
+ if (sock_fd == -1) {
+ plt_err("Failed to open socket. err %d", -errno);
+ return -errno;
+ }
+ sun.sun_family = AF_UNIX;
+ sunlen = sizeof(struct sockaddr_un);
+ strncpy(sun.sun_path, CNXK_ESWITCH_CTRL_MSG_SOCK_PATH,
+ sizeof(sun.sun_path) - 1);
+
+ if (connect(sock_fd, (struct sockaddr *)&sun, sunlen) < 0) {
+ plt_err("Failed to connect socket: %s, err %d",
+ CNXK_ESWITCH_CTRL_MSG_SOCK_PATH, errno);
+ close(sock_fd);
+ return -errno;
+ }
+ }
+ rte_thread_join(eswitch_dev->rep_ctrl_msg_thread, NULL);
+ if (!eswitch_dev->client_connected)
+ close(sock_fd);
+ }
+
+ /* Remove representor devices associated with PF */
cnxk_rep_dev_remove(eswitch_dev);
+ }
/* Cleanup NPC rxtx flow rules */
cnxk_eswitch_flow_rules_remove_list(eswitch_dev, &eswitch_dev->esw_flow_list,
@@ -106,13 +160,6 @@ cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev)
{
int rc;
- /* Enable Rx in NPC */
- rc = roc_nix_npc_rx_ena_dis(&eswitch_dev->nix, true);
- if (rc) {
- plt_err("Failed to enable NPC rx %d", rc);
- goto done;
- }
-
/* Install eswitch PF mcam rules */
rc = cnxk_eswitch_pfvf_flow_rules_install(eswitch_dev, false);
if (rc) {
@@ -128,6 +175,13 @@ cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev)
goto done;
}
+ /* Enable Rx in NPC */
+ rc = roc_nix_npc_rx_ena_dis(&eswitch_dev->nix, true);
+ if (rc) {
+ plt_err("Failed to enable NPC rx %d", rc);
+ goto done;
+ }
+
rc = roc_npc_mcam_enable_all_entries(&eswitch_dev->npc, 1);
if (rc) {
plt_err("Failed to enable NPC entries %d", rc);
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index 4edfa91bdc..ecf10a8e08 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -133,6 +133,12 @@ struct cnxk_eswitch_dev {
/* No of representors */
struct cnxk_eswitch_repr_cnt repr_cnt;
+ /* Representor control channel field */
+ bool start_ctrl_msg_thrd;
+ rte_thread_t rep_ctrl_msg_thread;
+ bool client_connected;
+ int sock_fd;
+
/* Port representor fields */
rte_spinlock_t rep_lock;
uint16_t nb_switch_domain;
@@ -155,6 +161,8 @@ cnxk_eswitch_pmd_priv(void)
/* HW Resources */
int cnxk_eswitch_nix_rsrc_start(struct cnxk_eswitch_dev *eswitch_dev);
+int cnxk_eswitch_representor_id(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func,
+ uint16_t *rep_id);
struct cnxk_esw_repr_hw_info *cnxk_eswitch_representor_hw_info(struct cnxk_eswitch_dev *eswitch_dev,
uint16_t hw_func);
int cnxk_eswitch_repr_devargs(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
index 55156f5b56..5b619ebb9e 100644
--- a/drivers/net/cnxk/cnxk_rep.c
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -2,6 +2,7 @@
* Copyright(C) 2024 Marvell.
*/
#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
#define PF_SHIFT 10
#define PF_MASK 0x3F
@@ -25,6 +26,48 @@ switch_domain_id_allocate(struct cnxk_eswitch_dev *eswitch_dev, uint16_t pf)
return RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID;
}
+int
+cnxk_rep_state_update(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, uint16_t *rep_id)
+{
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct rte_eth_dev *rep_eth_dev;
+ int i, rc = 0;
+
+ /* Delete the individual PFVF flows as common eswitch VF rule will be used. */
+ rc = cnxk_eswitch_flow_rules_delete(eswitch_dev, hw_func);
+ if (rc) {
+ if (rc != -ENOENT) {
+ plt_err("Failed to delete %x flow rules", hw_func);
+ goto fail;
+ }
+ }
+ /* Rep ID for respective HW func */
+ rc = cnxk_eswitch_representor_id(eswitch_dev, hw_func, rep_id);
+ if (rc) {
+ if (rc != -ENOENT) {
+ plt_err("Failed to get rep info for %x", hw_func);
+ goto fail;
+ }
+ }
+ /* Update the state - representee is standalone or part of companian app */
+ for (i = 0; i < eswitch_dev->repr_cnt.nb_repr_probed; i++) {
+ rep_eth_dev = eswitch_dev->rep_info[i].rep_eth_dev;
+ if (!rep_eth_dev) {
+ plt_err("Failed to get rep ethdev handle");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ if (rep_dev->hw_func == hw_func && rep_dev->is_vf_active)
+ rep_dev->native_repte = false;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
int
cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev)
{
@@ -250,6 +293,15 @@ cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswi
}
eswitch_dev->last_probed = i;
+ /* Launch a thread to handle control messages */
+ if (!eswitch_dev->start_ctrl_msg_thrd) {
+ rc = cnxk_rep_msg_control_thread_launch(eswitch_dev);
+ if (rc) {
+ plt_err("Failed to launch message ctrl thread");
+ goto fail;
+ }
+ }
+
return 0;
fail:
return rc;
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index b802c44b33..da298823a7 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -16,6 +16,8 @@ struct cnxk_rep_dev {
uint16_t switch_domain_id;
struct cnxk_eswitch_dev *parent_dev;
uint16_t hw_func;
+ bool is_vf_active;
+ bool native_repte;
uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
};
@@ -46,5 +48,6 @@ int cnxk_rep_dev_close(struct rte_eth_dev *eth_dev);
int cnxk_rep_stats_get(struct rte_eth_dev *eth_dev, struct rte_eth_stats *stats);
int cnxk_rep_stats_reset(struct rte_eth_dev *eth_dev);
int cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops);
+int cnxk_rep_state_update(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, uint16_t *rep_id);
#endif /* __CNXK_REP_H__ */
diff --git a/drivers/net/cnxk/cnxk_rep_msg.c b/drivers/net/cnxk/cnxk_rep_msg.c
new file mode 100644
index 0000000000..0af87f0169
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_msg.c
@@ -0,0 +1,827 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
+
+#define CTRL_MSG_RCV_TIMEOUT_MS 2000
+#define CTRL_MSG_READY_WAIT_US 2000
+#define CTRL_MSG_THRD_NAME_LEN 35
+#define CTRL_MSG_BUFFER_SZ 1500
+#define CTRL_MSG_SIGNATURE 0xcdacdeadbeefcadc
+
+static void
+close_socket(int fd)
+{
+ close(fd);
+ unlink(CNXK_ESWITCH_CTRL_MSG_SOCK_PATH);
+}
+
+static int
+receive_control_message(int socketfd, void *data, uint32_t len)
+{
+ char ctl[CMSG_SPACE(sizeof(int)) + CMSG_SPACE(sizeof(struct ucred))] = {0};
+ struct ucred *cr __rte_unused;
+ struct msghdr mh = {0};
+ struct cmsghdr *cmsg;
+ static uint64_t rec;
+ struct iovec iov[1];
+ ssize_t size;
+ int afd = -1;
+
+ iov[0].iov_base = data;
+ iov[0].iov_len = len;
+ mh.msg_iov = iov;
+ mh.msg_iovlen = 1;
+ mh.msg_control = ctl;
+ mh.msg_controllen = sizeof(ctl);
+
+ size = recvmsg(socketfd, &mh, MSG_DONTWAIT);
+ if (size < 0) {
+ if (errno == EAGAIN)
+ return 0;
+ plt_err("recvmsg err %d size %ld", errno, size);
+ return -errno;
+ } else if (size == 0) {
+ return 0;
+ }
+
+ rec++;
+ plt_rep_dbg("Packet %" PRId64 " Received %" PRId64 " bytes over socketfd %d",
+ rec, size, socketfd);
+
+ cr = 0;
+ cmsg = CMSG_FIRSTHDR(&mh);
+ while (cmsg) {
+ if (cmsg->cmsg_level == SOL_SOCKET) {
+ if (cmsg->cmsg_type == SCM_CREDENTIALS) {
+ cr = (struct ucred *)CMSG_DATA(cmsg);
+ } else if (cmsg->cmsg_type == SCM_RIGHTS) {
+ rte_memcpy(&afd, CMSG_DATA(cmsg), sizeof(int));
+ plt_rep_dbg("afd %d", afd);
+ }
+ }
+ cmsg = CMSG_NXTHDR(&mh, cmsg);
+ }
+ return size;
+}
+
+static int
+send_message_on_socket(int socketfd, void *data, uint32_t len, int afd)
+{
+ char ctl[CMSG_SPACE(sizeof(int))];
+ struct msghdr mh = {0};
+ struct cmsghdr *cmsg;
+ static uint64_t sent;
+ struct iovec iov[1];
+ int size;
+
+ iov[0].iov_base = data;
+ iov[0].iov_len = len;
+ mh.msg_iov = iov;
+ mh.msg_iovlen = 1;
+
+ if (afd > 0) {
+ memset(&ctl, 0, sizeof(ctl));
+ mh.msg_control = ctl;
+ mh.msg_controllen = sizeof(ctl);
+ cmsg = CMSG_FIRSTHDR(&mh);
+ cmsg->cmsg_len = CMSG_LEN(sizeof(int));
+ cmsg->cmsg_level = SOL_SOCKET;
+ cmsg->cmsg_type = SCM_RIGHTS;
+ rte_memcpy(CMSG_DATA(cmsg), &afd, sizeof(int));
+ }
+
+ size = sendmsg(socketfd, &mh, MSG_DONTWAIT);
+ if (size < 0) {
+ if (errno == EAGAIN)
+ return 0;
+ plt_err("Failed to send message, err %d", -errno);
+ return -errno;
+ } else if (size == 0) {
+ return 0;
+ }
+ sent++;
+ plt_rep_dbg("Sent %" PRId64 " packets of size %d on socketfd %d", sent, size, socketfd);
+
+ return size;
+}
+
+static int
+open_socket_ctrl_channel(void)
+{
+ struct sockaddr_un un;
+ int sock_fd;
+
+ sock_fd = socket(AF_UNIX, SOCK_STREAM, 0);
+ if (sock_fd < 0) {
+ RTE_LOG(ERR, EAL, "failed to create unix socket\n");
+ return -1;
+ }
+
+ /* Set unix socket path and bind */
+ memset(&un, 0, sizeof(un));
+ un.sun_family = AF_UNIX;
+
+ if (strlen(CNXK_ESWITCH_CTRL_MSG_SOCK_PATH) > sizeof(un.sun_path) - 1) {
+ plt_err("Server socket path too long: %s", CNXK_ESWITCH_CTRL_MSG_SOCK_PATH);
+ close(sock_fd);
+ return -E2BIG;
+ }
+
+ if (remove(CNXK_ESWITCH_CTRL_MSG_SOCK_PATH) == -1 && errno != ENOENT) {
+ plt_err("remove-%s", CNXK_ESWITCH_CTRL_MSG_SOCK_PATH);
+ close(sock_fd);
+ return -errno;
+ }
+
+ memset(&un, 0, sizeof(struct sockaddr_un));
+ un.sun_family = AF_UNIX;
+ strncpy(un.sun_path, CNXK_ESWITCH_CTRL_MSG_SOCK_PATH, sizeof(un.sun_path) - 1);
+
+ if (bind(sock_fd, (struct sockaddr *)&un, sizeof(un)) < 0) {
+ plt_err("Failed to bind %s: %s", un.sun_path, strerror(errno));
+ close(sock_fd);
+ return -errno;
+ }
+
+ if (listen(sock_fd, 1) < 0) {
+ plt_err("Failed to listen, err %s", strerror(errno));
+ close(sock_fd);
+ return -errno;
+ }
+
+ plt_rep_dbg("Unix socket path %s", un.sun_path);
+ return sock_fd;
+}
+
+static int
+send_control_message(struct cnxk_eswitch_dev *eswitch_dev, void *buffer, uint32_t len)
+{
+ int sz;
+ int rc = 0;
+
+ sz = send_message_on_socket(eswitch_dev->sock_fd, buffer, len, 0);
+ if (sz < 0) {
+ plt_err("Error sending message, err %d", sz);
+ rc = sz;
+ goto done;
+ }
+
+ /* Ensuring entire message has been processed */
+ if (sz != (int)len) {
+ plt_err("Out of %d bytes only %d bytes sent", sz, len);
+ rc = -EFAULT;
+ goto done;
+ }
+ plt_rep_dbg("Sent %d bytes of buffer", sz);
+done:
+ return rc;
+}
+
+void
+cnxk_rep_msg_populate_msg_end(void *buffer, uint32_t *length)
+{
+ cnxk_rep_msg_populate_command(buffer, length, CNXK_REP_MSG_END, 0);
+}
+
+void
+cnxk_rep_msg_populate_type(void *buffer, uint32_t *length, cnxk_type_t type, uint32_t sz)
+{
+ uint32_t len = *length;
+ cnxk_type_data_t data;
+
+ memset(&data, 0, sizeof(cnxk_type_data_t));
+ /* Prepare type data */
+ data.type = type;
+ data.length = sz;
+
+ /* Populate the type data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), &data, sizeof(cnxk_type_data_t));
+ len += sizeof(cnxk_type_data_t);
+
+ *length = len;
+}
+
+void
+cnxk_rep_msg_populate_header(void *buffer, uint32_t *length)
+{
+ cnxk_header_t hdr;
+ int len;
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_HEADER, sizeof(cnxk_header_t));
+
+ memset(&hdr, 0, sizeof(cnxk_header_t));
+ len = *length;
+ /* Prepare header data */
+ hdr.signature = CTRL_MSG_SIGNATURE;
+
+ /* Populate header data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), &hdr, sizeof(cnxk_header_t));
+ len += sizeof(cnxk_header_t);
+
+ *length = len;
+}
+
+void
+cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type, uint32_t size)
+{
+ cnxk_rep_msg_data_t msg_data;
+ uint32_t len;
+ uint16_t sz = sizeof(cnxk_rep_msg_data_t);
+
+ memset(&msg_data, 0, sz);
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_MSG, sz);
+
+ len = *length;
+ /* Prepare command data */
+ msg_data.type = type;
+ msg_data.length = size;
+
+ /* Populate the command */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), &msg_data, sz);
+ len += sz;
+
+ *length = len;
+}
+
+void
+cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
+ cnxk_rep_msg_t msg)
+{
+ uint32_t len;
+
+ cnxk_rep_msg_populate_command(buffer, length, msg, sz);
+
+ len = *length;
+ /* Populate command data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), msg_meta, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static int
+parse_validate_header(void *msg_buf, uint32_t *buf_trav_len)
+{
+ cnxk_type_data_t *tdata = NULL;
+ cnxk_header_t *hdr = NULL;
+ void *data = NULL;
+ uint16_t len = 0;
+
+ /* Read first bytes of type data */
+ data = msg_buf;
+ tdata = (cnxk_type_data_t *)data;
+ if (tdata->type != CNXK_TYPE_HEADER) {
+ plt_err("Invalid type %d, type header expected", tdata->type);
+ goto fail;
+ }
+
+ /* Get the header value */
+ data = RTE_PTR_ADD(msg_buf, sizeof(cnxk_type_data_t));
+ len += sizeof(cnxk_type_data_t);
+
+ /* Validate the header */
+ hdr = (cnxk_header_t *)data;
+ if (hdr->signature != CTRL_MSG_SIGNATURE) {
+ plt_err("Invalid signature %" PRIu64 " detected", hdr->signature);
+ goto fail;
+ }
+
+ /* Update length read till point */
+ len += tdata->length;
+
+ *buf_trav_len = len;
+ return 0;
+fail:
+ return errno;
+}
+
+static cnxk_rep_msg_data_t *
+message_data_extract(void *msg_buf, uint32_t *buf_trav_len)
+{
+ cnxk_type_data_t *tdata = NULL;
+ cnxk_rep_msg_data_t *msg = NULL;
+ uint16_t len = *buf_trav_len;
+ void *data;
+
+ tdata = (cnxk_type_data_t *)RTE_PTR_ADD(msg_buf, len);
+ if (tdata->type != CNXK_TYPE_MSG) {
+ plt_err("Invalid type %d, type MSG expected", tdata->type);
+ goto fail;
+ }
+
+ /* Get the message type */
+ len += sizeof(cnxk_type_data_t);
+ data = RTE_PTR_ADD(msg_buf, len);
+ msg = (cnxk_rep_msg_data_t *)data;
+
+ /* Advance to actual message data */
+ len += tdata->length;
+ *buf_trav_len = len;
+
+ return msg;
+fail:
+ return NULL;
+}
+
+static void
+process_ack_message(void *msg_buf, uint32_t *buf_trav_len, uint32_t msg_len, void *data)
+{
+ cnxk_rep_msg_ack_data_t *adata = (cnxk_rep_msg_ack_data_t *)data;
+ uint16_t len = *buf_trav_len;
+ void *buf;
+
+ /* Get the message type data viz ack data */
+ buf = RTE_PTR_ADD(msg_buf, len);
+ adata->u.data = rte_zmalloc("Ack data", msg_len, 0);
+ adata->size = msg_len;
+ if (adata->size == sizeof(uint64_t))
+ rte_memcpy(&adata->u.data, buf, msg_len);
+ else
+ rte_memcpy(adata->u.data, buf, msg_len);
+ plt_rep_dbg("Address %p val 0x%" PRIu64 " sval %" PRId64 " msg_len %d",
+ adata->u.data, adata->u.val, adata->u.sval, msg_len);
+
+ /* Advance length to nex message */
+ len += msg_len;
+ *buf_trav_len = len;
+}
+
+static int
+notify_rep_dev_ready(cnxk_rep_msg_ready_data_t *rdata, void *data,
+ cnxk_rep_msg_ack_data1_t **padata)
+{
+ struct cnxk_eswitch_dev *eswitch_dev;
+ uint64_t rep_id_arr[RTE_MAX_ETHPORTS];
+ cnxk_rep_msg_ack_data1_t *adata;
+ uint16_t rep_id, sz, total_sz;
+ int rc, i, j = 0;
+
+ PLT_SET_USED(data);
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ if (!eswitch_dev) {
+ plt_err("Failed to get PF ethdev handle");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ memset(rep_id_arr, 0, RTE_MAX_ETHPORTS * sizeof(uint64_t));
+ /* For ready state */
+ if ((rdata->nb_ports / 2) > eswitch_dev->repr_cnt.nb_repr_probed) {
+ rc = CNXK_REP_CTRL_MSG_NACK_INV_REP_CNT;
+ goto fail;
+ }
+
+ for (i = 0; i < rdata->nb_ports / 2; i++) {
+ rep_id = UINT16_MAX;
+ rc = cnxk_rep_state_update(eswitch_dev, rdata->data[i], &rep_id);
+ if (rc) {
+ rc = CNXK_REP_CTRL_MSG_NACK_REP_STAT_UP_FAIL;
+ goto fail;
+ }
+ if (rep_id != UINT16_MAX)
+ rep_id_arr[j++] = rep_id;
+ }
+
+ /* Send Rep Id array to companian app */
+ sz = j * sizeof(uint64_t);
+ total_sz = sizeof(cnxk_rep_msg_ack_data1_t) + sz;
+ adata = plt_zmalloc(total_sz, 0);
+ rte_memcpy(adata->data, rep_id_arr, sz);
+ adata->size = sz;
+ *padata = adata;
+
+ plt_rep_dbg("Installing NPC rules for Eswitch VF");
+ /* Install RX VLAN rule for eswitch VF */
+ if (!eswitch_dev->eswitch_vf_rules_setup) {
+ rc = cnxk_eswitch_pfvf_flow_rules_install(eswitch_dev, true);
+ if (rc) {
+ plt_err("Failed to install rxtx rules, rc %d", rc);
+ goto fail;
+ }
+
+ /* Configure TPID for Eswitch PF LFs */
+ rc = roc_eswitch_nix_vlan_tpid_set(&eswitch_dev->nix, ROC_NIX_VLAN_TYPE_OUTER,
+ CNXK_ESWITCH_VLAN_TPID, true);
+ if (rc) {
+ plt_err("Failed to configure tpid, rc %d", rc);
+ goto fail;
+ }
+ eswitch_dev->eswitch_vf_rules_setup = true;
+ }
+
+ return 0;
+fail:
+ sz = sizeof(cnxk_rep_msg_ack_data1_t) + sizeof(uint64_t);
+ adata = plt_zmalloc(sz, 0);
+ adata->data[0] = rc;
+ adata->size = sizeof(uint64_t);
+ *padata = adata;
+
+ return rc;
+}
+
+static int
+process_ready_message(void *msg_buf, uint32_t *buf_trav_len, uint32_t msg_len, void *data,
+ cnxk_rep_msg_ack_data1_t **padata)
+{
+ cnxk_rep_msg_ready_data_t *rdata = NULL;
+ cnxk_rep_msg_ack_data1_t *adata;
+ uint16_t len = *buf_trav_len;
+ void *buf;
+ int rc = 0, sz;
+
+ /* Get the message type data viz ready data */
+ buf = RTE_PTR_ADD(msg_buf, len);
+ rdata = (cnxk_rep_msg_ready_data_t *)buf;
+
+ plt_rep_dbg("Ready data received %d, nb_ports %d", rdata->val, rdata->nb_ports);
+
+ /* Wait required to ensure other side ready for receiving the ack */
+ usleep(CTRL_MSG_READY_WAIT_US);
+
+ /* Update all representor about ready message */
+ if (rdata->val) {
+ rc = notify_rep_dev_ready(rdata, data, padata);
+ } else {
+ sz = sizeof(cnxk_rep_msg_ack_data1_t) + sizeof(uint64_t);
+ adata = plt_zmalloc(sz, 0);
+ adata->data[0] = CNXK_REP_CTRL_MSG_NACK_INV_RDY_DATA;
+ adata->size = sizeof(uint64_t);
+ *padata = adata;
+ }
+
+ /* Advance length to nex message */
+ len += msg_len;
+ *buf_trav_len = len;
+
+ return rc;
+}
+
+static int
+notify_rep_dev_exit(cnxk_rep_msg_exit_data_t *edata, void *data)
+{
+ struct cnxk_eswitch_dev *eswitch_dev;
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct rte_eth_dev *rep_eth_dev;
+ int i, rc = 0;
+
+ PLT_SET_USED(data);
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ if (!eswitch_dev) {
+ plt_err("Failed to get PF ethdev handle");
+ rc = -EINVAL;
+ goto fail;
+ }
+ if ((edata->nb_ports / 2) > eswitch_dev->repr_cnt.nb_repr_probed) {
+ rc = CNXK_REP_CTRL_MSG_NACK_INV_REP_CNT;
+ goto fail;
+ }
+
+ for (i = 0; i < eswitch_dev->repr_cnt.nb_repr_probed; i++) {
+ rep_eth_dev = eswitch_dev->rep_info[i].rep_eth_dev;
+ if (!rep_eth_dev) {
+ plt_err("Failed to get rep ethdev handle");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ if (!rep_dev->native_repte)
+ rep_dev->is_vf_active = false;
+ }
+ /* For Exit message */
+ eswitch_dev->client_connected = false;
+ return 0;
+fail:
+ return rc;
+}
+
+static void
+process_exit_message(void *msg_buf, uint32_t *buf_trav_len, uint32_t msg_len, void *data)
+{
+ cnxk_rep_msg_exit_data_t *edata = NULL;
+ uint16_t len = *buf_trav_len;
+ void *buf;
+
+ /* Get the message type data viz exit data */
+ buf = RTE_PTR_ADD(msg_buf, len);
+ edata = (cnxk_rep_msg_exit_data_t *)buf;
+
+ plt_rep_dbg("Exit data received %d", edata->val);
+
+ /* Update all representor about ready/exit message */
+ if (edata->val)
+ notify_rep_dev_exit(edata, data);
+
+ /* Advance length to nex message */
+ len += msg_len;
+ *buf_trav_len = len;
+}
+
+static void
+populate_ack_msg(void *buffer, uint32_t *length, cnxk_rep_msg_ack_data1_t *adata)
+{
+ uint32_t sz = sizeof(cnxk_rep_msg_ack_data1_t) + adata->size;
+ uint32_t len;
+
+ cnxk_rep_msg_populate_command(buffer, length, CNXK_REP_MSG_ACK, sz);
+
+ len = *length;
+
+ /* Populate ACK message data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), adata, sz);
+
+ len += sz;
+
+ *length = len;
+}
+
+static int
+send_ack_message(void *data, cnxk_rep_msg_ack_data1_t *adata)
+{
+ struct cnxk_eswitch_dev *eswitch_dev = (struct cnxk_eswitch_dev *)data;
+ uint32_t len = 0, size;
+ void *buffer;
+ int rc = 0;
+
+ /* Allocate memory for preparing a message */
+ size = CTRL_MSG_BUFFER_SZ;
+ buffer = rte_zmalloc("ACK msg", size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ return -ENOMEM;
+ }
+
+ /* Prepare the ACK message */
+ cnxk_rep_msg_populate_header(buffer, &len);
+ populate_ack_msg(buffer, &len, adata);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ /* Length check to avoid buffer overflow */
+ if (len > CTRL_MSG_BUFFER_SZ) {
+ plt_err("Invalid length %d for max sized buffer %d", len, CTRL_MSG_BUFFER_SZ);
+ rc = -EFAULT;
+ goto done;
+ }
+
+ /* Send it to the peer */
+ rc = send_control_message(eswitch_dev, buffer, len);
+ if (rc)
+ plt_err("Failed send ack");
+
+done:
+ return rc;
+}
+
+static int
+process_message(void *msg_buf, uint32_t *buf_trav_len, void *data)
+{
+ cnxk_rep_msg_data_t *msg = NULL;
+ cnxk_rep_msg_ack_data1_t *adata = NULL;
+ bool send_ack;
+ int rc = 0, sz;
+
+ /* Get the message data */
+ msg = message_data_extract(msg_buf, buf_trav_len);
+ if (!msg) {
+ plt_err("Failed to get message data");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Different message type processing */
+ while (msg->type != CNXK_REP_MSG_END) {
+ send_ack = true;
+ switch (msg->type) {
+ case CNXK_REP_MSG_ACK:
+ plt_rep_dbg("Received ack response");
+ process_ack_message(msg_buf, buf_trav_len, msg->length, data);
+ send_ack = false;
+ break;
+ case CNXK_REP_MSG_READY:
+ plt_rep_dbg("Received ready message");
+ process_ready_message(msg_buf, buf_trav_len, msg->length, data, &adata);
+ adata->type = CNXK_REP_MSG_READY;
+ break;
+ case CNXK_REP_MSG_EXIT:
+ plt_rep_dbg("Received exit message");
+ process_exit_message(msg_buf, buf_trav_len, msg->length, data);
+ sz = sizeof(cnxk_rep_msg_ack_data1_t) + sizeof(uint64_t);
+ adata = plt_zmalloc(sz, 0);
+ adata->type = CNXK_REP_MSG_EXIT;
+ adata->data[0] = 0;
+ adata->size = sizeof(uint64_t);
+ break;
+ default:
+ send_ack = false;
+ plt_err("Invalid message type: %d", msg->type);
+ rc = -EINVAL;
+ };
+
+ /* Send ACK */
+ if (send_ack)
+ send_ack_message(data, adata);
+
+ /* Advance to next message */
+ msg = message_data_extract(msg_buf, buf_trav_len);
+ if (!msg) {
+ plt_err("Failed to get message data");
+ rc = -EINVAL;
+ goto fail;
+ }
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+process_control_message(void *msg_buf, void *data, size_t sz)
+{
+ uint32_t buf_trav_len = 0;
+ int rc;
+
+ /* Validate the validity of the received message */
+ parse_validate_header(msg_buf, &buf_trav_len);
+
+ /* Detect message and process */
+ rc = process_message(msg_buf, &buf_trav_len, data);
+ if (rc) {
+ plt_err("Failed to process message");
+ goto fail;
+ }
+
+ /* Ensuring entire message has been processed */
+ if (sz != buf_trav_len) {
+ plt_err("Out of %" PRId64 " bytes %d bytes of msg_buf processed", sz, buf_trav_len);
+ rc = -EFAULT;
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+receive_control_msg_resp(struct cnxk_eswitch_dev *eswitch_dev, void *data)
+{
+ uint32_t wait_us = CTRL_MSG_RCV_TIMEOUT_MS * 1000;
+ uint32_t timeout = 0, sleep = 1;
+ int sz = 0;
+ int rc = -1;
+ uint32_t len = BUFSIZ;
+ void *msg_buf;
+
+ msg_buf = plt_zmalloc(len, 0);
+
+ do {
+ sz = receive_control_message(eswitch_dev->sock_fd, msg_buf, len);
+ if (sz != 0)
+ break;
+
+ /* Timeout after CTRL_MSG_RCV_TIMEOUT_MS */
+ if (timeout >= wait_us) {
+ plt_err("Control message wait timedout");
+ return -ETIMEDOUT;
+ }
+
+ plt_delay_us(sleep);
+ timeout += sleep;
+ } while ((sz == 0) || (timeout < wait_us));
+
+ if (sz > 0) {
+ plt_rep_dbg("Received %d sized response packet", sz);
+ rc = process_control_message(msg_buf, data, sz);
+ plt_free(msg_buf);
+ }
+
+ return rc;
+}
+
+int
+cnxk_rep_msg_send_process(struct cnxk_rep_dev *rep_dev, void *buffer, uint32_t len,
+ cnxk_rep_msg_ack_data_t *adata)
+{
+ struct cnxk_eswitch_dev *eswitch_dev;
+ int rc = 0;
+
+ eswitch_dev = rep_dev->parent_dev;
+ if (!eswitch_dev) {
+ plt_err("Failed to get parent eswitch handle");
+ rc = -1;
+ goto fail;
+ }
+
+ plt_spinlock_lock(&eswitch_dev->rep_lock);
+ rc = send_control_message(eswitch_dev, buffer, len);
+ if (rc) {
+ plt_err("Failed to send the message, err %d", rc);
+ goto free;
+ }
+
+ /* Get response of the command sent */
+ rc = receive_control_msg_resp(eswitch_dev, adata);
+ if (rc) {
+ plt_err("Failed to receive the response, err %d", rc);
+ goto free;
+ }
+ plt_spinlock_unlock(&eswitch_dev->rep_lock);
+
+ return 0;
+free:
+ plt_spinlock_unlock(&eswitch_dev->rep_lock);
+fail:
+ return rc;
+}
+
+static void
+poll_for_control_msg(void *data)
+{
+ struct cnxk_eswitch_dev *eswitch_dev = (struct cnxk_eswitch_dev *)data;
+ uint32_t len = BUFSIZ;
+ int sz = 0;
+ void *msg_buf;
+
+ while (eswitch_dev->client_connected) {
+ msg_buf = plt_zmalloc(len, 0);
+ do {
+ plt_spinlock_lock(&eswitch_dev->rep_lock);
+ sz = receive_control_message(eswitch_dev->sock_fd, msg_buf, len);
+ plt_spinlock_unlock(&eswitch_dev->rep_lock);
+ if (sz != 0)
+ break;
+ plt_delay_us(2000);
+ } while (sz == 0);
+
+ if (sz > 0) {
+ plt_rep_dbg("Received new %d bytes control message", sz);
+ plt_spinlock_lock(&eswitch_dev->rep_lock);
+ process_control_message(msg_buf, data, sz);
+ plt_spinlock_unlock(&eswitch_dev->rep_lock);
+ plt_free(msg_buf);
+ }
+ }
+ plt_rep_dbg("Exiting poll for control message loop");
+}
+
+static uint32_t
+rep_ctrl_msg_thread_main(void *arg)
+{
+ struct cnxk_eswitch_dev *eswitch_dev = (struct cnxk_eswitch_dev *)arg;
+ struct sockaddr_un client;
+ int addr_len;
+ int ssock_fd;
+ int sock_fd;
+
+ ssock_fd = open_socket_ctrl_channel();
+ if (ssock_fd < 0) {
+ plt_err("Failed to open socket for ctrl channel, err %d", ssock_fd);
+ return UINT32_MAX;
+ }
+
+ addr_len = sizeof(client);
+ while (eswitch_dev->start_ctrl_msg_thrd) {
+ /* Accept client connection until the thread is running */
+ sock_fd = accept(ssock_fd, (struct sockaddr *)&client, (socklen_t *)&addr_len);
+ if (sock_fd < 0) {
+ plt_err("Failed to accept connection request on socket fd %d", ssock_fd);
+ break;
+ }
+
+ plt_rep_dbg("Client %s: Connection request accepted.", client.sun_path);
+ eswitch_dev->sock_fd = sock_fd;
+ if (eswitch_dev->start_ctrl_msg_thrd) {
+ eswitch_dev->client_connected = true;
+ poll_for_control_msg(eswitch_dev);
+ }
+ eswitch_dev->sock_fd = -1;
+ close(sock_fd);
+ }
+
+ /* Closing the opened socket */
+ close_socket(ssock_fd);
+ plt_rep_dbg("Exiting representor ctrl thread");
+
+ return 0;
+}
+
+int
+cnxk_rep_msg_control_thread_launch(struct cnxk_eswitch_dev *eswitch_dev)
+{
+ char name[CTRL_MSG_THRD_NAME_LEN];
+ int rc = 0;
+
+ rte_strscpy(name, "rep_ctrl_msg_hndlr", CTRL_MSG_THRD_NAME_LEN);
+ eswitch_dev->start_ctrl_msg_thrd = true;
+ rc = rte_thread_create_internal_control(&eswitch_dev->rep_ctrl_msg_thread, name,
+ rep_ctrl_msg_thread_main, eswitch_dev);
+ if (rc)
+ plt_err("Failed to create rep control message handling");
+
+ return rc;
+}
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
new file mode 100644
index 0000000000..0543805148
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -0,0 +1,95 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#ifndef __CNXK_REP_MSG_H__
+#define __CNXK_REP_MSG_H__
+
+#include <stdint.h>
+
+#define CNXK_REP_MSG_MAX_BUFFER_SZ 1500
+
+typedef enum CNXK_TYPE {
+ CNXK_TYPE_HEADER = 0,
+ CNXK_TYPE_MSG,
+} cnxk_type_t;
+
+typedef enum CNXK_REP_MSG {
+ /* General sync messages */
+ CNXK_REP_MSG_READY = 0,
+ CNXK_REP_MSG_ACK,
+ CNXK_REP_MSG_EXIT,
+ /* End of messaging sequence */
+ CNXK_REP_MSG_END,
+} cnxk_rep_msg_t;
+
+typedef enum CNXK_NACK_CODE {
+ CNXK_REP_CTRL_MSG_NACK_INV_RDY_DATA = 0x501,
+ CNXK_REP_CTRL_MSG_NACK_INV_REP_CNT = 0x502,
+ CNXK_REP_CTRL_MSG_NACK_REP_STAT_UP_FAIL = 0x503,
+} cnxk_nack_code_t;
+
+/* Types */
+typedef struct cnxk_type_data {
+ cnxk_type_t type;
+ uint32_t length;
+ uint64_t data[];
+} __rte_packed cnxk_type_data_t;
+
+/* Header */
+typedef struct cnxk_header {
+ uint64_t signature;
+ uint16_t nb_hops;
+} __rte_packed cnxk_header_t;
+
+/* Message meta */
+typedef struct cnxk_rep_msg_data {
+ cnxk_rep_msg_t type;
+ uint32_t length;
+ uint64_t data[];
+} __rte_packed cnxk_rep_msg_data_t;
+
+/* Ack msg */
+typedef struct cnxk_rep_msg_ack_data {
+ cnxk_rep_msg_t type;
+ uint32_t size;
+ union {
+ void *data;
+ uint64_t val;
+ int64_t sval;
+ } u;
+} __rte_packed cnxk_rep_msg_ack_data_t;
+
+/* Ack msg */
+typedef struct cnxk_rep_msg_ack_data1 {
+ cnxk_rep_msg_t type;
+ uint32_t size;
+ uint64_t data[];
+} __rte_packed cnxk_rep_msg_ack_data1_t;
+
+/* Ready msg */
+typedef struct cnxk_rep_msg_ready_data {
+ uint8_t val;
+ uint16_t nb_ports;
+ uint16_t data[];
+} __rte_packed cnxk_rep_msg_ready_data_t;
+
+/* Exit msg */
+typedef struct cnxk_rep_msg_exit_data {
+ uint8_t val;
+ uint16_t nb_ports;
+ uint16_t data[];
+} __rte_packed cnxk_rep_msg_exit_data_t;
+
+void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
+ uint32_t size);
+void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
+ cnxk_rep_msg_t msg);
+void cnxk_rep_msg_populate_msg_end(void *buffer, uint32_t *length);
+void cnxk_rep_msg_populate_type(void *buffer, uint32_t *length, cnxk_type_t type, uint32_t sz);
+void cnxk_rep_msg_populate_header(void *buffer, uint32_t *length);
+int cnxk_rep_msg_send_process(struct cnxk_rep_dev *rep_dev, void *buffer, uint32_t len,
+ cnxk_rep_msg_ack_data_t *adata);
+int cnxk_rep_msg_control_thread_launch(struct cnxk_eswitch_dev *eswitch_dev);
+
+#endif /* __CNXK_REP_MSG_H__ */
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index 7121845dc6..9ca7732713 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -37,6 +37,7 @@ sources = files(
'cnxk_ptp.c',
'cnxk_flow.c',
'cnxk_rep.c',
+ 'cnxk_rep_msg.c',
'cnxk_rep_ops.c',
'cnxk_stats.c',
'cnxk_tm.c',
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 11/23] common/cnxk: representee notification callback
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (9 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 10/23] net/cnxk: add representor control plane Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 12/23] net/cnxk: handling representee notification Harman Kalra
` (11 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Setting up a callback which gets invoked every time a representee
comes up or goes down. Later this callback gets handled by network
conterpart.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_dev.c | 24 ++++++++++++++++++++++++
drivers/common/cnxk/roc_dev_priv.h | 3 +++
drivers/common/cnxk/roc_eswitch.c | 23 +++++++++++++++++++++++
drivers/common/cnxk/roc_eswitch.h | 6 ++++++
drivers/common/cnxk/roc_mbox.c | 2 ++
drivers/common/cnxk/roc_mbox.h | 11 ++++++++++-
drivers/common/cnxk/version.map | 2 ++
7 files changed, 70 insertions(+), 1 deletion(-)
diff --git a/drivers/common/cnxk/roc_dev.c b/drivers/common/cnxk/roc_dev.c
index 084343c3b4..3564dbe647 100644
--- a/drivers/common/cnxk/roc_dev.c
+++ b/drivers/common/cnxk/roc_dev.c
@@ -537,6 +537,29 @@ pf_vf_mbox_send_up_msg(struct dev *dev, void *rec_msg)
}
}
+static int
+mbox_up_handler_rep_repte_notify(struct dev *dev, struct rep_repte_req *req, struct msg_rsp *rsp)
+{
+ int rc = 0;
+
+ plt_base_dbg("pf:%d/vf:%d msg id 0x%x (%s) from: pf:%d/vf:%d", dev_get_pf(dev->pf_func),
+ dev_get_vf(dev->pf_func), req->hdr.id, mbox_id2name(req->hdr.id),
+ dev_get_pf(req->hdr.pcifunc), dev_get_vf(req->hdr.pcifunc));
+
+ plt_base_dbg("repte pcifunc %x, enable %d", req->repte_pcifunc, req->enable);
+
+ if (dev->ops && dev->ops->repte_notify) {
+ rc = dev->ops->repte_notify(dev->roc_nix, req->repte_pcifunc,
+ req->enable);
+ if (rc < 0)
+ plt_err("Failed to sent new representee %x notification to %s",
+ req->repte_pcifunc, (req->enable == true) ? "enable" : "disable");
+ }
+
+ rsp->hdr.rc = rc;
+ return rc;
+}
+
static int
mbox_up_handler_mcs_intr_notify(struct dev *dev, struct mcs_intr_info *info, struct msg_rsp *rsp)
{
@@ -711,6 +734,7 @@ mbox_process_msgs_up(struct dev *dev, struct mbox_msghdr *req)
}
MBOX_UP_CGX_MESSAGES
MBOX_UP_MCS_MESSAGES
+ MBOX_UP_REP_MESSAGES
#undef M
}
diff --git a/drivers/common/cnxk/roc_dev_priv.h b/drivers/common/cnxk/roc_dev_priv.h
index 5b2c5096f8..dd694b8572 100644
--- a/drivers/common/cnxk/roc_dev_priv.h
+++ b/drivers/common/cnxk/roc_dev_priv.h
@@ -36,12 +36,15 @@ typedef void (*q_err_cb_t)(void *roc_nix, void *data);
/* Link status get callback */
typedef void (*link_status_get_t)(void *roc_nix,
struct cgx_link_user_info *link);
+/* Representee notification callback */
+typedef int (*repte_notify_t)(void *roc_nix, uint16_t pf_func, bool enable);
struct dev_ops {
link_info_t link_status_update;
ptp_info_t ptp_info_update;
link_status_get_t link_status_get;
q_err_cb_t q_err_cb;
+ repte_notify_t repte_notify;
};
#define dev_is_vf(dev) ((dev)->hwcap & DEV_HWCAP_F_VF)
diff --git a/drivers/common/cnxk/roc_eswitch.c b/drivers/common/cnxk/roc_eswitch.c
index 020a891a32..14819bad75 100644
--- a/drivers/common/cnxk/roc_eswitch.c
+++ b/drivers/common/cnxk/roc_eswitch.c
@@ -319,3 +319,26 @@ roc_eswitch_nix_vlan_tpid_set(struct roc_nix *roc_nix, uint32_t type, uint16_t t
return rc;
}
+
+int
+roc_eswitch_nix_process_repte_notify_cb_register(struct roc_nix *roc_nix,
+ process_repte_notify_t proc_repte_nt)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+
+ if (proc_repte_nt == NULL)
+ return NIX_ERR_PARAM;
+
+ dev->ops->repte_notify = (repte_notify_t)proc_repte_nt;
+ return 0;
+}
+
+void
+roc_eswitch_nix_process_repte_notify_cb_unregister(struct roc_nix *roc_nix)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+
+ dev->ops->repte_notify = NULL;
+}
diff --git a/drivers/common/cnxk/roc_eswitch.h b/drivers/common/cnxk/roc_eswitch.h
index 34b75d10ac..f99f7e2981 100644
--- a/drivers/common/cnxk/roc_eswitch.h
+++ b/drivers/common/cnxk/roc_eswitch.h
@@ -8,6 +8,9 @@
#define ROC_ESWITCH_VLAN_TPID 0x8100
#define ROC_ESWITCH_LBK_CHAN 63
+/* Process representee notification callback */
+typedef int (*process_repte_notify_t)(void *roc_nix, uint16_t pf_func, bool enable);
+
/* NPC */
int __roc_api roc_eswitch_npc_mcam_rx_rule(struct roc_npc *roc_npc, struct roc_npc_flow *flow,
uint16_t pcifunc, uint16_t vlan_tci,
@@ -23,4 +26,7 @@ int __roc_api roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc,
/* NIX */
int __roc_api roc_eswitch_nix_vlan_tpid_set(struct roc_nix *nix, uint32_t type, uint16_t tpid,
bool is_vf);
+int __roc_api roc_eswitch_nix_process_repte_notify_cb_register(struct roc_nix *roc_nix,
+ process_repte_notify_t proc_repte_nt);
+void __roc_api roc_eswitch_nix_process_repte_notify_cb_unregister(struct roc_nix *roc_nix);
#endif /* __ROC_ESWITCH_H__ */
diff --git a/drivers/common/cnxk/roc_mbox.c b/drivers/common/cnxk/roc_mbox.c
index 7b734fcd24..10cdbc4d13 100644
--- a/drivers/common/cnxk/roc_mbox.c
+++ b/drivers/common/cnxk/roc_mbox.c
@@ -499,6 +499,7 @@ mbox_id2name(uint16_t id)
return #_name;
MBOX_MESSAGES
MBOX_UP_CGX_MESSAGES
+ MBOX_UP_REP_MESSAGES
#undef M
}
}
@@ -514,6 +515,7 @@ mbox_id2size(uint16_t id)
return sizeof(struct _req_type);
MBOX_MESSAGES
MBOX_UP_CGX_MESSAGES
+ MBOX_UP_REP_MESSAGES
#undef M
}
}
diff --git a/drivers/common/cnxk/roc_mbox.h b/drivers/common/cnxk/roc_mbox.h
index e903db391c..9de48c1907 100644
--- a/drivers/common/cnxk/roc_mbox.h
+++ b/drivers/common/cnxk/roc_mbox.h
@@ -355,9 +355,12 @@ struct mbox_msghdr {
#define MBOX_UP_MCS_MESSAGES M(MCS_INTR_NOTIFY, 0xE00, mcs_intr_notify, mcs_intr_info, msg_rsp)
+#define MBOX_UP_REP_MESSAGES \
+M(REP_REPTE_NOTIFY, 0xEF1, rep_repte_notify, rep_repte_req, msg_rsp)
+
enum {
#define M(_name, _id, _1, _2, _3) MBOX_MSG_##_name = _id,
- MBOX_MESSAGES MBOX_UP_CGX_MESSAGES MBOX_UP_MCS_MESSAGES
+ MBOX_MESSAGES MBOX_UP_CGX_MESSAGES MBOX_UP_MCS_MESSAGES MBOX_UP_REP_MESSAGES
#undef M
};
@@ -2786,4 +2789,10 @@ struct nix_spi_to_sa_delete_req {
uint16_t __io hash_index;
uint8_t __io way;
};
+
+struct rep_repte_req {
+ struct mbox_msghdr hdr;
+ uint16_t __io repte_pcifunc;
+ bool __io enable;
+};
#endif /* __ROC_MBOX_H__ */
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index 78c421677d..e170a6a63a 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -91,6 +91,8 @@ INTERNAL {
roc_dpi_disable;
roc_dpi_enable;
roc_error_msg_get;
+ roc_eswitch_nix_process_repte_notify_cb_register;
+ roc_eswitch_nix_process_repte_notify_cb_unregister;
roc_eswitch_nix_vlan_tpid_set;
roc_eswitch_npc_mcam_delete_rule;
roc_eswitch_npc_mcam_rx_rule;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 12/23] net/cnxk: handling representee notification
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (10 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 11/23] common/cnxk: representee notification callback Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 13/23] net/cnxk: representor ethdev ops Harman Kalra
` (10 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
In case of any representee coming up or going down, kernel sends a
mbox up call which signals a thread to process these messages and
enable/disable HW resources accordingly.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_eswitch.c | 8 +
drivers/net/cnxk/cnxk_eswitch.h | 20 +++
drivers/net/cnxk/cnxk_rep.c | 263 ++++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep.h | 36 +++++
4 files changed, 327 insertions(+)
diff --git a/drivers/net/cnxk/cnxk_eswitch.c b/drivers/net/cnxk/cnxk_eswitch.c
index ad6834410d..79f44de06e 100644
--- a/drivers/net/cnxk/cnxk_eswitch.c
+++ b/drivers/net/cnxk/cnxk_eswitch.c
@@ -139,6 +139,14 @@ cnxk_eswitch_dev_remove(struct rte_pci_device *pci_dev)
close(sock_fd);
}
+ if (eswitch_dev->repte_msg_proc.start_thread) {
+ eswitch_dev->repte_msg_proc.start_thread = false;
+ pthread_cond_signal(&eswitch_dev->repte_msg_proc.repte_msg_cond);
+ rte_thread_join(eswitch_dev->repte_msg_proc.repte_msg_thread, NULL);
+ pthread_mutex_destroy(&eswitch_dev->repte_msg_proc.mutex);
+ pthread_cond_destroy(&eswitch_dev->repte_msg_proc.repte_msg_cond);
+ }
+
/* Remove representor devices associated with PF */
cnxk_rep_dev_remove(eswitch_dev);
}
diff --git a/drivers/net/cnxk/cnxk_eswitch.h b/drivers/net/cnxk/cnxk_eswitch.h
index ecf10a8e08..1baf198d72 100644
--- a/drivers/net/cnxk/cnxk_eswitch.h
+++ b/drivers/net/cnxk/cnxk_eswitch.h
@@ -30,6 +30,23 @@ enum cnxk_esw_da_pattern_type {
CNXK_ESW_DA_TYPE_PFVF,
};
+struct cnxk_esw_repte_msg {
+ uint16_t hw_func;
+ bool enable;
+
+ TAILQ_ENTRY(cnxk_esw_repte_msg) next;
+};
+
+struct cnxk_esw_repte_msg_proc {
+ bool start_thread;
+ uint8_t msg_avail;
+ rte_thread_t repte_msg_thread;
+ pthread_cond_t repte_msg_cond;
+ pthread_mutex_t mutex;
+
+ TAILQ_HEAD(esw_repte_msg_list, cnxk_esw_repte_msg) msg_list;
+};
+
struct cnxk_esw_repr_hw_info {
/* Representee pcifunc value */
uint16_t hw_func;
@@ -139,6 +156,9 @@ struct cnxk_eswitch_dev {
bool client_connected;
int sock_fd;
+ /* Representee notification */
+ struct cnxk_esw_repte_msg_proc repte_msg_proc;
+
/* Port representor fields */
rte_spinlock_t rep_lock;
uint16_t nb_switch_domain;
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
index 5b619ebb9e..11901dac87 100644
--- a/drivers/net/cnxk/cnxk_rep.c
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -4,6 +4,8 @@
#include <cnxk_rep.h>
#include <cnxk_rep_msg.h>
+#define REPTE_MSG_PROC_THRD_NAME_MAX_LEN 30
+
#define PF_SHIFT 10
#define PF_MASK 0x3F
@@ -86,6 +88,7 @@ cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev)
{
int i, rc = 0;
+ roc_eswitch_nix_process_repte_notify_cb_unregister(&eswitch_dev->nix);
for (i = 0; i < eswitch_dev->nb_switch_domain; i++) {
rc = rte_eth_switch_domain_free(eswitch_dev->sw_dom[i].switch_domain_id);
if (rc)
@@ -95,6 +98,236 @@ cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev)
return rc;
}
+static int
+cnxk_representee_release(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func)
+{
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct rte_eth_dev *rep_eth_dev;
+ int i, rc = 0;
+
+ for (i = 0; i < eswitch_dev->repr_cnt.nb_repr_probed; i++) {
+ rep_eth_dev = eswitch_dev->rep_info[i].rep_eth_dev;
+ if (!rep_eth_dev) {
+ plt_err("Failed to get rep ethdev handle");
+ rc = -EINVAL;
+ goto done;
+ }
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ if (rep_dev->hw_func == hw_func &&
+ (!rep_dev->native_repte || rep_dev->is_vf_active)) {
+ rep_dev->is_vf_active = false;
+ rc = cnxk_rep_dev_stop(rep_eth_dev);
+ if (rc) {
+ plt_err("Failed to stop repr port %d, rep id %d", rep_dev->port_id,
+ rep_dev->rep_id);
+ goto done;
+ }
+
+ cnxk_rep_rx_queue_release(rep_eth_dev, 0);
+ cnxk_rep_tx_queue_release(rep_eth_dev, 0);
+ plt_rep_dbg("Released representor ID %d representing %x", rep_dev->rep_id,
+ hw_func);
+ break;
+ }
+ }
+done:
+ return rc;
+}
+
+static int
+cnxk_representee_setup(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, uint16_t rep_id)
+{
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct rte_eth_dev *rep_eth_dev;
+ int i, rc = 0;
+
+ for (i = 0; i < eswitch_dev->repr_cnt.nb_repr_probed; i++) {
+ rep_eth_dev = eswitch_dev->rep_info[i].rep_eth_dev;
+ if (!rep_eth_dev) {
+ plt_err("Failed to get rep ethdev handle");
+ rc = -EINVAL;
+ goto done;
+ }
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ if (rep_dev->hw_func == hw_func && !rep_dev->is_vf_active) {
+ rep_dev->is_vf_active = true;
+ rep_dev->native_repte = true;
+ if (rep_dev->rep_id != rep_id) {
+ plt_err("Rep ID assigned during init %d does not match %d",
+ rep_dev->rep_id, rep_id);
+ rc = -EINVAL;
+ goto done;
+ }
+
+ rc = cnxk_rep_rx_queue_setup(rep_eth_dev, rep_dev->rxq->qid,
+ rep_dev->rxq->nb_desc, 0,
+ rep_dev->rxq->rx_conf, rep_dev->rxq->mpool);
+ if (rc) {
+ plt_err("Failed to setup rxq repr port %d, rep id %d",
+ rep_dev->port_id, rep_dev->rep_id);
+ goto done;
+ }
+
+ rc = cnxk_rep_tx_queue_setup(rep_eth_dev, rep_dev->txq->qid,
+ rep_dev->txq->nb_desc, 0,
+ rep_dev->txq->tx_conf);
+ if (rc) {
+ plt_err("Failed to setup txq repr port %d, rep id %d",
+ rep_dev->port_id, rep_dev->rep_id);
+ goto done;
+ }
+
+ rc = cnxk_rep_dev_start(rep_eth_dev);
+ if (rc) {
+ plt_err("Failed to start repr port %d, rep id %d", rep_dev->port_id,
+ rep_dev->rep_id);
+ goto done;
+ }
+ break;
+ }
+ }
+done:
+ return rc;
+}
+
+static int
+cnxk_representee_msg_process(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, bool enable)
+{
+ struct cnxk_eswitch_devargs *esw_da;
+ uint16_t rep_id = UINT16_MAX;
+ int rc = 0, i, j;
+
+ /* Traversing the initialized represented list */
+ for (i = 0; i < eswitch_dev->nb_esw_da; i++) {
+ esw_da = &eswitch_dev->esw_da[i];
+ for (j = 0; j < esw_da->nb_repr_ports; j++) {
+ if (esw_da->repr_hw_info[j].hw_func == hw_func) {
+ rep_id = esw_da->repr_hw_info[j].rep_id;
+ break;
+ }
+ }
+ if (rep_id != UINT16_MAX)
+ break;
+ }
+ /* No action on PF func for which representor has not been created */
+ if (rep_id == UINT16_MAX)
+ goto done;
+
+ if (enable) {
+ rc = cnxk_representee_setup(eswitch_dev, hw_func, rep_id);
+ if (rc) {
+ plt_err("Failed to setup representee, err %d", rc);
+ goto fail;
+ }
+ plt_rep_dbg(" Representor ID %d representing %x", rep_id, hw_func);
+ rc = cnxk_eswitch_flow_rules_install(eswitch_dev, hw_func);
+ if (rc) {
+ plt_err("Failed to install rxtx flow rules for %x", hw_func);
+ goto fail;
+ }
+ } else {
+ rc = cnxk_eswitch_flow_rules_delete(eswitch_dev, hw_func);
+ if (rc) {
+ plt_err("Failed to delete flow rules for %x", hw_func);
+ goto fail;
+ }
+ rc = cnxk_representee_release(eswitch_dev, hw_func);
+ if (rc) {
+ plt_err("Failed to release representee, err %d", rc);
+ goto fail;
+ }
+ }
+
+done:
+ return 0;
+fail:
+ return rc;
+}
+
+static uint32_t
+cnxk_representee_msg_thread_main(void *arg)
+{
+ struct cnxk_eswitch_dev *eswitch_dev = (struct cnxk_eswitch_dev *)arg;
+ struct cnxk_esw_repte_msg_proc *repte_msg_proc;
+ struct cnxk_esw_repte_msg *msg, *next_msg;
+ int count, rc;
+
+ repte_msg_proc = &eswitch_dev->repte_msg_proc;
+ pthread_mutex_lock(&eswitch_dev->repte_msg_proc.mutex);
+ while (eswitch_dev->repte_msg_proc.start_thread) {
+ do {
+ rc = pthread_cond_wait(&eswitch_dev->repte_msg_proc.repte_msg_cond,
+ &eswitch_dev->repte_msg_proc.mutex);
+ } while (rc != 0);
+
+ /* Go through list pushed from interrupt context and process each message */
+ next_msg = TAILQ_FIRST(&repte_msg_proc->msg_list);
+ count = 0;
+ while (next_msg) {
+ msg = next_msg;
+ count++;
+ plt_rep_dbg(" Processing msg %d: hw_func %x action %s", count,
+ msg->hw_func, msg->enable ? "enable" : "disable");
+
+ /* Unlocking for interrupt thread to grab lock
+ * while thread process the message.
+ */
+ pthread_mutex_unlock(&eswitch_dev->repte_msg_proc.mutex);
+ /* Processing the message */
+ cnxk_representee_msg_process(eswitch_dev, msg->hw_func, msg->enable);
+ /* Locking as cond wait will unlock before wait */
+ pthread_mutex_lock(&eswitch_dev->repte_msg_proc.mutex);
+ next_msg = TAILQ_NEXT(msg, next);
+ TAILQ_REMOVE(&repte_msg_proc->msg_list, msg, next);
+ rte_free(msg);
+ }
+ }
+
+ pthread_mutex_unlock(&eswitch_dev->repte_msg_proc.mutex);
+
+ return 0;
+}
+
+static int
+cnxk_representee_notification(void *roc_nix, uint16_t hw_func, bool enable)
+{
+ struct cnxk_esw_repte_msg_proc *repte_msg_proc;
+ struct cnxk_eswitch_dev *eswitch_dev;
+ struct cnxk_esw_repte_msg *msg;
+ int rc = 0;
+
+ RTE_SET_USED(roc_nix);
+ eswitch_dev = cnxk_eswitch_pmd_priv();
+ if (!eswitch_dev) {
+ plt_err("Failed to get PF ethdev handle");
+ rc = -EINVAL;
+ goto done;
+ }
+
+ repte_msg_proc = &eswitch_dev->repte_msg_proc;
+ msg = rte_zmalloc("msg", sizeof(struct cnxk_esw_repte_msg), 0);
+ if (!msg) {
+ plt_err("Failed to allocate memory for repte msg");
+ rc = -ENOMEM;
+ goto done;
+ }
+
+ msg->hw_func = hw_func;
+ msg->enable = enable;
+
+ plt_rep_dbg("Pushing new notification : hw_func %x enable %d\n", msg->hw_func, enable);
+ pthread_mutex_lock(&eswitch_dev->repte_msg_proc.mutex);
+ TAILQ_INSERT_TAIL(&repte_msg_proc->msg_list, msg, next);
+ /* Signal vf message handler thread */
+ pthread_cond_signal(&eswitch_dev->repte_msg_proc.repte_msg_cond);
+ pthread_mutex_unlock(&eswitch_dev->repte_msg_proc.mutex);
+
+done:
+ return rc;
+}
+
static int
cnxk_rep_parent_setup(struct cnxk_eswitch_dev *eswitch_dev)
{
@@ -263,6 +496,7 @@ create_representor_ethdev(struct rte_pci_device *pci_dev, struct cnxk_eswitch_de
int
cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev)
{
+ char name[REPTE_MSG_PROC_THRD_NAME_MAX_LEN];
struct cnxk_eswitch_devargs *esw_da;
uint16_t num_rep;
int i, j, rc;
@@ -302,7 +536,36 @@ cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswi
}
}
+ if (!eswitch_dev->repte_msg_proc.start_thread) {
+ /* Register callback for representee notification */
+ if (roc_eswitch_nix_process_repte_notify_cb_register(&eswitch_dev->nix,
+ cnxk_representee_notification)) {
+ plt_err("Failed to register callback for representee notification");
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Create a thread for handling msgs from VFs */
+ TAILQ_INIT(&eswitch_dev->repte_msg_proc.msg_list);
+ pthread_cond_init(&eswitch_dev->repte_msg_proc.repte_msg_cond, NULL);
+ pthread_mutex_init(&eswitch_dev->repte_msg_proc.mutex, NULL);
+
+ rte_strscpy(name, "repte_msg_proc_thrd", REPTE_MSG_PROC_THRD_NAME_MAX_LEN);
+ eswitch_dev->repte_msg_proc.start_thread = true;
+ rc =
+ rte_thread_create_internal_control(&eswitch_dev->repte_msg_proc.repte_msg_thread,
+ name, cnxk_representee_msg_thread_main,
+ eswitch_dev);
+ if (rc != 0) {
+ plt_err("Failed to create thread for VF mbox handling\n");
+ goto thread_fail;
+ }
+ }
+
return 0;
+thread_fail:
+ pthread_mutex_destroy(&eswitch_dev->repte_msg_proc.mutex);
+ pthread_cond_destroy(&eswitch_dev->repte_msg_proc.repte_msg_cond);
fail:
return rc;
}
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index da298823a7..bee141e25b 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -10,6 +10,40 @@
/* Common ethdev ops */
extern struct eth_dev_ops cnxk_rep_dev_ops;
+struct cnxk_rep_queue_stats {
+ uint64_t pkts;
+ uint64_t bytes;
+};
+
+struct cnxk_rep_rxq {
+ /* Parent rep device */
+ struct cnxk_rep_dev *rep_dev;
+ /* Queue ID */
+ uint16_t qid;
+ /* No of desc */
+ uint16_t nb_desc;
+ /* mempool handle */
+ struct rte_mempool *mpool;
+ /* RX config parameters */
+ const struct rte_eth_rxconf *rx_conf;
+ /* Per queue TX statistics */
+ struct cnxk_rep_queue_stats stats;
+};
+
+struct cnxk_rep_txq {
+ /* Parent rep device */
+ struct cnxk_rep_dev *rep_dev;
+ /* Queue ID */
+ uint16_t qid;
+ /* No of desc */
+ uint16_t nb_desc;
+ /* TX config parameters */
+ const struct rte_eth_txconf *tx_conf;
+ /* Per queue TX statistics */
+ struct cnxk_rep_queue_stats stats;
+};
+
+/* Representor port configurations */
struct cnxk_rep_dev {
uint16_t port_id;
uint16_t rep_id;
@@ -18,6 +52,8 @@ struct cnxk_rep_dev {
uint16_t hw_func;
bool is_vf_active;
bool native_repte;
+ struct cnxk_rep_rxq *rxq;
+ struct cnxk_rep_txq *txq;
uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
};
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 13/23] net/cnxk: representor ethdev ops
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (11 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 12/23] net/cnxk: handling representee notification Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 14/23] common/cnxk: get representees ethernet stats Harman Kalra
` (9 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Implementing ethernet device operation callbacks for
port representors PMD
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_rep.c | 28 +-
drivers/net/cnxk/cnxk_rep.h | 35 +++
drivers/net/cnxk/cnxk_rep_msg.h | 8 +
drivers/net/cnxk/cnxk_rep_ops.c | 495 ++++++++++++++++++++++++++++++--
4 files changed, 523 insertions(+), 43 deletions(-)
diff --git a/drivers/net/cnxk/cnxk_rep.c b/drivers/net/cnxk/cnxk_rep.c
index 11901dac87..27940e7e97 100644
--- a/drivers/net/cnxk/cnxk_rep.c
+++ b/drivers/net/cnxk/cnxk_rep.c
@@ -73,6 +73,8 @@ cnxk_rep_state_update(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, ui
int
cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev)
{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+
if (rte_eal_process_type() != RTE_PROC_PRIMARY)
return 0;
@@ -80,6 +82,8 @@ cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev)
rte_free(ethdev->data->mac_addrs);
ethdev->data->mac_addrs = NULL;
+ rep_dev->parent_dev->repr_cnt.nb_repr_probed--;
+
return 0;
}
@@ -369,26 +373,6 @@ cnxk_rep_parent_setup(struct cnxk_eswitch_dev *eswitch_dev)
return rc;
}
-static uint16_t
-cnxk_rep_tx_burst(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
-{
- PLT_SET_USED(tx_queue);
- PLT_SET_USED(tx_pkts);
- PLT_SET_USED(nb_pkts);
-
- return 0;
-}
-
-static uint16_t
-cnxk_rep_rx_burst(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
-{
- PLT_SET_USED(rx_queue);
- PLT_SET_USED(rx_pkts);
- PLT_SET_USED(nb_pkts);
-
- return 0;
-}
-
static int
cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
{
@@ -418,8 +402,8 @@ cnxk_rep_dev_init(struct rte_eth_dev *eth_dev, void *params)
eth_dev->dev_ops = &cnxk_rep_dev_ops;
/* Rx/Tx functions stubs to avoid crashing */
- eth_dev->rx_pkt_burst = cnxk_rep_rx_burst;
- eth_dev->tx_pkt_burst = cnxk_rep_tx_burst;
+ eth_dev->rx_pkt_burst = cnxk_rep_rx_burst_dummy;
+ eth_dev->tx_pkt_burst = cnxk_rep_tx_burst_dummy;
/* Only single queues for representor devices */
eth_dev->data->nb_rx_queues = 1;
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index bee141e25b..b57b56cd12 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -7,6 +7,13 @@
#ifndef __CNXK_REP_H__
#define __CNXK_REP_H__
+#define CNXK_REP_TX_OFFLOAD_CAPA \
+ (RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE | RTE_ETH_TX_OFFLOAD_VLAN_INSERT | \
+ RTE_ETH_TX_OFFLOAD_MULTI_SEGS)
+
+#define CNXK_REP_RX_OFFLOAD_CAPA \
+ (RTE_ETH_RX_OFFLOAD_SCATTER | RTE_ETH_RX_OFFLOAD_RSS_HASH | RTE_ETH_RX_OFFLOAD_VLAN_STRIP)
+
/* Common ethdev ops */
extern struct eth_dev_ops cnxk_rep_dev_ops;
@@ -57,12 +64,33 @@ struct cnxk_rep_dev {
uint8_t mac_addr[RTE_ETHER_ADDR_LEN];
};
+/* Inline functions */
+static inline void
+cnxk_rep_lock(struct cnxk_rep_dev *rep)
+{
+ rte_spinlock_lock(&rep->parent_dev->rep_lock);
+}
+
+static inline void
+cnxk_rep_unlock(struct cnxk_rep_dev *rep)
+{
+ rte_spinlock_unlock(&rep->parent_dev->rep_lock);
+}
+
static inline struct cnxk_rep_dev *
cnxk_rep_pmd_priv(const struct rte_eth_dev *eth_dev)
{
return eth_dev->data->dev_private;
}
+static __rte_always_inline void
+cnxk_rep_pool_buffer_stats(struct rte_mempool *pool)
+{
+ plt_rep_dbg(" pool %s size %d buffer count in use %d available %d\n", pool->name,
+ pool->size, rte_mempool_in_use_count(pool), rte_mempool_avail_count(pool));
+}
+
+/* Prototypes */
int cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
int cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev);
int cnxk_rep_dev_uninit(struct rte_eth_dev *ethdev);
@@ -85,5 +113,12 @@ int cnxk_rep_stats_get(struct rte_eth_dev *eth_dev, struct rte_eth_stats *stats)
int cnxk_rep_stats_reset(struct rte_eth_dev *eth_dev);
int cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops);
int cnxk_rep_state_update(struct cnxk_eswitch_dev *eswitch_dev, uint16_t hw_func, uint16_t *rep_id);
+int cnxk_rep_promiscuous_enable(struct rte_eth_dev *ethdev);
+int cnxk_rep_promiscuous_disable(struct rte_eth_dev *ethdev);
+int cnxk_rep_mac_addr_set(struct rte_eth_dev *eth_dev, struct rte_ether_addr *addr);
+uint16_t cnxk_rep_tx_burst_dummy(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts);
+uint16_t cnxk_rep_rx_burst_dummy(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts);
+void cnxk_rep_tx_queue_stop(struct rte_eth_dev *ethdev, uint16_t queue_id);
+void cnxk_rep_rx_queue_stop(struct rte_eth_dev *ethdev, uint16_t queue_id);
#endif /* __CNXK_REP_H__ */
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
index 0543805148..63cfbe3f19 100644
--- a/drivers/net/cnxk/cnxk_rep_msg.h
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -19,6 +19,8 @@ typedef enum CNXK_REP_MSG {
CNXK_REP_MSG_READY = 0,
CNXK_REP_MSG_ACK,
CNXK_REP_MSG_EXIT,
+ /* Ethernet operation msgs */
+ CNXK_REP_MSG_ETH_SET_MAC,
/* End of messaging sequence */
CNXK_REP_MSG_END,
} cnxk_rep_msg_t;
@@ -81,6 +83,12 @@ typedef struct cnxk_rep_msg_exit_data {
uint16_t data[];
} __rte_packed cnxk_rep_msg_exit_data_t;
+/* Ethernet op - set mac */
+typedef struct cnxk_rep_msg_eth_mac_set_meta {
+ uint16_t portid;
+ uint8_t addr_bytes[RTE_ETHER_ADDR_LEN];
+} __rte_packed cnxk_rep_msg_eth_set_mac_meta_t;
+
void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
uint32_t size);
void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
index 15448688ce..97643a50f2 100644
--- a/drivers/net/cnxk/cnxk_rep_ops.c
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -3,25 +3,221 @@
*/
#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
+
+#define MEMPOOL_CACHE_SIZE 256
+#define TX_DESC_PER_QUEUE 512
+#define RX_DESC_PER_QUEUE 256
+#define NB_REP_VDEV_MBUF 1024
+
+static uint16_t
+cnxk_rep_tx_burst(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
+{
+ struct cnxk_rep_txq *txq = tx_queue;
+ struct cnxk_rep_dev *rep_dev;
+ uint16_t n_tx;
+
+ if (unlikely(!txq))
+ return 0;
+
+ rep_dev = txq->rep_dev;
+ plt_rep_dbg("Transmitting %d packets on eswitch queue %d", nb_pkts, txq->qid);
+ n_tx = cnxk_eswitch_dev_tx_burst(rep_dev->parent_dev, txq->qid, tx_pkts, nb_pkts,
+ NIX_TX_OFFLOAD_VLAN_QINQ_F);
+ return n_tx;
+}
+
+static uint16_t
+cnxk_rep_rx_burst(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
+{
+ struct cnxk_rep_rxq *rxq = rx_queue;
+ struct cnxk_rep_dev *rep_dev;
+ uint16_t n_rx;
+
+ if (unlikely(!rxq))
+ return 0;
+
+ rep_dev = rxq->rep_dev;
+ n_rx = cnxk_eswitch_dev_rx_burst(rep_dev->parent_dev, rxq->qid, rx_pkts, nb_pkts);
+ if (n_rx == 0)
+ return 0;
+
+ plt_rep_dbg("Received %d packets on eswitch queue %d", n_rx, rxq->qid);
+ return n_rx;
+}
+
+uint16_t
+cnxk_rep_tx_burst_dummy(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(tx_queue);
+ PLT_SET_USED(tx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
+
+uint16_t
+cnxk_rep_rx_burst_dummy(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
+{
+ PLT_SET_USED(rx_queue);
+ PLT_SET_USED(rx_pkts);
+ PLT_SET_USED(nb_pkts);
+
+ return 0;
+}
int
cnxk_rep_link_update(struct rte_eth_dev *ethdev, int wait_to_complete)
{
- PLT_SET_USED(ethdev);
+ struct rte_eth_link link;
PLT_SET_USED(wait_to_complete);
+
+ memset(&link, 0, sizeof(link));
+ if (ethdev->data->dev_started)
+ link.link_status = RTE_ETH_LINK_UP;
+ else
+ link.link_status = RTE_ETH_LINK_DOWN;
+
+ link.link_duplex = RTE_ETH_LINK_FULL_DUPLEX;
+ link.link_autoneg = RTE_ETH_LINK_FIXED;
+ link.link_speed = RTE_ETH_SPEED_NUM_UNKNOWN;
+
+ return rte_eth_linkstatus_set(ethdev, &link);
+}
+
+int
+cnxk_rep_dev_info_get(struct rte_eth_dev *ethdev, struct rte_eth_dev_info *dev_info)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ uint32_t max_rx_pktlen;
+
+ max_rx_pktlen = (roc_nix_max_pkt_len(&rep_dev->parent_dev->nix) + RTE_ETHER_CRC_LEN -
+ CNXK_NIX_MAX_VTAG_ACT_SIZE);
+
+ dev_info->min_rx_bufsize = NIX_MIN_HW_FRS + RTE_ETHER_CRC_LEN;
+ dev_info->max_rx_pktlen = max_rx_pktlen;
+ dev_info->max_mac_addrs = roc_nix_mac_max_entries_get(&rep_dev->parent_dev->nix);
+
+ dev_info->rx_offload_capa = CNXK_REP_RX_OFFLOAD_CAPA;
+ dev_info->tx_offload_capa = CNXK_REP_TX_OFFLOAD_CAPA;
+ dev_info->rx_queue_offload_capa = 0;
+ dev_info->tx_queue_offload_capa = 0;
+
+ /* For the sake of symmetry, max_rx_queues = max_tx_queues */
+ dev_info->max_rx_queues = 1;
+ dev_info->max_tx_queues = 1;
+
+ /* MTU specifics */
+ dev_info->max_mtu = dev_info->max_rx_pktlen - (RTE_ETHER_HDR_LEN + RTE_ETHER_CRC_LEN);
+ dev_info->min_mtu = dev_info->min_rx_bufsize - CNXK_NIX_L2_OVERHEAD;
+
+ /* Switch info specific */
+ dev_info->switch_info.name = ethdev->device->name;
+ dev_info->switch_info.domain_id = rep_dev->switch_domain_id;
+ dev_info->switch_info.port_id = rep_dev->port_id;
+
return 0;
}
int
-cnxk_rep_dev_info_get(struct rte_eth_dev *ethdev, struct rte_eth_dev_info *devinfo)
+cnxk_rep_representor_info_get(struct rte_eth_dev *ethdev, struct rte_eth_representor_info *info)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+
+ return cnxk_eswitch_representor_info_get(rep_dev->parent_dev, info);
+}
+
+static int
+rep_eth_conf_chk(const struct rte_eth_conf *conf, uint16_t nb_rx_queues)
+{
+ const struct rte_eth_rss_conf *rss_conf;
+ int ret = 0;
+
+ if (conf->link_speeds != 0) {
+ plt_err("specific link speeds not supported");
+ ret = -EINVAL;
+ }
+
+ switch (conf->rxmode.mq_mode) {
+ case RTE_ETH_MQ_RX_RSS:
+ if (nb_rx_queues != 1) {
+ plt_err("Rx RSS is not supported with %u queues", nb_rx_queues);
+ ret = -EINVAL;
+ break;
+ }
+
+ rss_conf = &conf->rx_adv_conf.rss_conf;
+ if (rss_conf->rss_key != NULL || rss_conf->rss_key_len != 0 ||
+ rss_conf->rss_hf != 0) {
+ plt_err("Rx RSS configuration is not supported");
+ ret = -EINVAL;
+ }
+ break;
+ case RTE_ETH_MQ_RX_NONE:
+ break;
+ default:
+ plt_err("Rx mode MQ modes other than RSS not supported");
+ ret = -EINVAL;
+ break;
+ }
+
+ if (conf->txmode.mq_mode != RTE_ETH_MQ_TX_NONE) {
+ plt_err("Tx mode MQ modes not supported");
+ ret = -EINVAL;
+ }
+
+ if (conf->lpbk_mode != 0) {
+ plt_err("loopback not supported");
+ ret = -EINVAL;
+ }
+
+ if (conf->dcb_capability_en != 0) {
+ plt_err("priority-based flow control not supported");
+ ret = -EINVAL;
+ }
+
+ if (conf->intr_conf.lsc != 0) {
+ plt_err("link status change interrupt not supported");
+ ret = -EINVAL;
+ }
+
+ if (conf->intr_conf.rxq != 0) {
+ plt_err("receive queue interrupt not supported");
+ ret = -EINVAL;
+ }
+
+ if (conf->intr_conf.rmv != 0) {
+ plt_err("remove interrupt not supported");
+ ret = -EINVAL;
+ }
+
+ return ret;
+}
+
+int
+cnxk_rep_dev_configure(struct rte_eth_dev *ethdev)
+{
+ struct rte_eth_dev_data *ethdev_data = ethdev->data;
+ int rc = -1;
+
+ rc = rep_eth_conf_chk(ðdev_data->dev_conf, ethdev_data->nb_rx_queues);
+ if (rc)
+ goto fail;
+
+ return 0;
+fail:
+ return rc;
+}
+
+int
+cnxk_rep_promiscuous_enable(struct rte_eth_dev *ethdev)
{
PLT_SET_USED(ethdev);
- PLT_SET_USED(devinfo);
return 0;
}
int
-cnxk_rep_dev_configure(struct rte_eth_dev *ethdev)
+cnxk_rep_promiscuous_disable(struct rte_eth_dev *ethdev)
{
PLT_SET_USED(ethdev);
return 0;
@@ -30,21 +226,73 @@ cnxk_rep_dev_configure(struct rte_eth_dev *ethdev)
int
cnxk_rep_dev_start(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ int rc = 0, qid;
+
+ ethdev->rx_pkt_burst = cnxk_rep_rx_burst;
+ ethdev->tx_pkt_burst = cnxk_rep_tx_burst;
+
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ if (!rep_dev->rxq || !rep_dev->txq) {
+ plt_err("Invalid rxq or txq for representor id %d", rep_dev->rep_id);
+ rc = -EINVAL;
+ goto fail;
+ }
+
+ /* Start rx queues */
+ qid = rep_dev->rxq->qid;
+ rc = cnxk_eswitch_rxq_start(rep_dev->parent_dev, qid);
+ if (rc) {
+ plt_err("Failed to start rxq %d, rc=%d", qid, rc);
+ goto fail;
+ }
+
+ /* Start tx queues */
+ qid = rep_dev->txq->qid;
+ rc = cnxk_eswitch_txq_start(rep_dev->parent_dev, qid);
+ if (rc) {
+ plt_err("Failed to start txq %d, rc=%d", qid, rc);
+ goto fail;
+ }
+
+ /* Start rep_xport device only once after first representor gets active */
+ if (!rep_dev->parent_dev->repr_cnt.nb_repr_started) {
+ rc = cnxk_eswitch_nix_rsrc_start(rep_dev->parent_dev);
+ if (rc) {
+ plt_err("Failed to start nix dev, rc %d", rc);
+ goto fail;
+ }
+ }
+
+ ethdev->data->tx_queue_state[0] = RTE_ETH_QUEUE_STATE_STARTED;
+ ethdev->data->rx_queue_state[0] = RTE_ETH_QUEUE_STATE_STARTED;
+
+ rep_dev->parent_dev->repr_cnt.nb_repr_started++;
+
return 0;
+fail:
+ return rc;
}
int
cnxk_rep_dev_close(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
- return 0;
+ return cnxk_rep_dev_uninit(ethdev);
}
int
cnxk_rep_dev_stop(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+
+ ethdev->rx_pkt_burst = cnxk_rep_rx_burst_dummy;
+ ethdev->tx_pkt_burst = cnxk_rep_tx_burst_dummy;
+ cnxk_rep_rx_queue_stop(ethdev, 0);
+ cnxk_rep_tx_queue_stop(ethdev, 0);
+ rep_dev->parent_dev->repr_cnt.nb_repr_started--;
+
return 0;
}
@@ -53,39 +301,189 @@ cnxk_rep_rx_queue_setup(struct rte_eth_dev *ethdev, uint16_t rx_queue_id, uint16
unsigned int socket_id, const struct rte_eth_rxconf *rx_conf,
struct rte_mempool *mb_pool)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(rx_queue_id);
- PLT_SET_USED(nb_rx_desc);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ struct cnxk_rep_rxq *rxq = NULL;
+ uint16_t qid = 0;
+ int rc;
+
PLT_SET_USED(socket_id);
- PLT_SET_USED(rx_conf);
- PLT_SET_USED(mb_pool);
+ /* If no representee assigned, store the respective rxq parameters */
+ if (!rep_dev->is_vf_active && !rep_dev->rxq) {
+ rxq = plt_zmalloc(sizeof(*rxq), RTE_CACHE_LINE_SIZE);
+ if (!rxq) {
+ rc = -ENOMEM;
+ plt_err("Failed to alloc RxQ for rep id %d", rep_dev->rep_id);
+ goto fail;
+ }
+
+ rxq->qid = qid;
+ rxq->nb_desc = nb_rx_desc;
+ rxq->rep_dev = rep_dev;
+ rxq->mpool = mb_pool;
+ rxq->rx_conf = rx_conf;
+ rep_dev->rxq = rxq;
+ ethdev->data->rx_queues[rx_queue_id] = NULL;
+
+ return 0;
+ }
+
+ qid = rep_dev->rep_id;
+ rc = cnxk_eswitch_rxq_setup(rep_dev->parent_dev, qid, nb_rx_desc, rx_conf, mb_pool);
+ if (rc) {
+ plt_err("failed to setup eswitch queue id %d", qid);
+ goto fail;
+ }
+
+ rxq = rep_dev->rxq;
+ if (!rxq) {
+ plt_err("Invalid RXQ handle for representor port %d rep id %d", rep_dev->port_id,
+ rep_dev->rep_id);
+ goto free_queue;
+ }
+
+ rxq->qid = qid;
+ ethdev->data->rx_queues[rx_queue_id] = rxq;
+ ethdev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
+ plt_rep_dbg("representor id %d portid %d rxq id %d", rep_dev->port_id,
+ ethdev->data->port_id, rxq->qid);
+
return 0;
+free_queue:
+ cnxk_eswitch_rxq_release(rep_dev->parent_dev, qid);
+fail:
+ return rc;
+}
+
+void
+cnxk_rep_rx_queue_stop(struct rte_eth_dev *ethdev, uint16_t queue_id)
+{
+ struct cnxk_rep_rxq *rxq = ethdev->data->rx_queues[queue_id];
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ int rc;
+
+ if (!rxq)
+ return;
+
+ plt_rep_dbg("Stopping rxq %u", rxq->qid);
+
+ rc = cnxk_eswitch_rxq_stop(rep_dev->parent_dev, rxq->qid);
+ if (rc)
+ plt_err("Failed to stop rxq %d, rc=%d", rc, rxq->qid);
+
+ ethdev->data->rx_queue_state[queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
}
void
cnxk_rep_rx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(queue_id);
+ struct cnxk_rep_rxq *rxq = ethdev->data->rx_queues[queue_id];
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ int rc;
+
+ if (!rxq) {
+ plt_err("Invalid rxq retrieved for rep_id %d", rep_dev->rep_id);
+ return;
+ }
+
+ plt_rep_dbg("Releasing rxq %u", rxq->qid);
+
+ rc = cnxk_eswitch_rxq_release(rep_dev->parent_dev, rxq->qid);
+ if (rc)
+ plt_err("Failed to release rxq %d, rc=%d", rc, rxq->qid);
}
int
cnxk_rep_tx_queue_setup(struct rte_eth_dev *ethdev, uint16_t tx_queue_id, uint16_t nb_tx_desc,
unsigned int socket_id, const struct rte_eth_txconf *tx_conf)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(tx_queue_id);
- PLT_SET_USED(nb_tx_desc);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ struct cnxk_rep_txq *txq = NULL;
+ int rc = 0, qid = 0;
+
PLT_SET_USED(socket_id);
- PLT_SET_USED(tx_conf);
+ /* If no representee assigned, store the respective rxq parameters */
+ if (!rep_dev->is_vf_active && !rep_dev->txq) {
+ txq = plt_zmalloc(sizeof(*txq), RTE_CACHE_LINE_SIZE);
+ if (!txq) {
+ rc = -ENOMEM;
+ plt_err("failed to alloc txq for rep id %d", rep_dev->rep_id);
+ goto free_queue;
+ }
+
+ txq->qid = qid;
+ txq->nb_desc = nb_tx_desc;
+ txq->tx_conf = tx_conf;
+ txq->rep_dev = rep_dev;
+ rep_dev->txq = txq;
+
+ ethdev->data->tx_queues[tx_queue_id] = NULL;
+
+ return 0;
+ }
+
+ qid = rep_dev->rep_id;
+ rc = cnxk_eswitch_txq_setup(rep_dev->parent_dev, qid, nb_tx_desc, tx_conf);
+ if (rc) {
+ plt_err("failed to setup eswitch queue id %d", qid);
+ goto fail;
+ }
+
+ txq = rep_dev->txq;
+ if (!txq) {
+ plt_err("Invalid TXQ handle for representor port %d rep id %d", rep_dev->port_id,
+ rep_dev->rep_id);
+ goto free_queue;
+ }
+
+ txq->qid = qid;
+ ethdev->data->tx_queues[tx_queue_id] = txq;
+ ethdev->data->tx_queue_state[tx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
+ plt_rep_dbg("representor id %d portid %d txq id %d", rep_dev->port_id,
+ ethdev->data->port_id, txq->qid);
+
return 0;
+free_queue:
+ cnxk_eswitch_txq_release(rep_dev->parent_dev, qid);
+fail:
+ return rc;
+}
+
+void
+cnxk_rep_tx_queue_stop(struct rte_eth_dev *ethdev, uint16_t queue_id)
+{
+ struct cnxk_rep_txq *txq = ethdev->data->tx_queues[queue_id];
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ int rc;
+
+ if (!txq)
+ return;
+
+ plt_rep_dbg("Releasing txq %u", txq->qid);
+
+ rc = cnxk_eswitch_txq_stop(rep_dev->parent_dev, txq->qid);
+ if (rc)
+ plt_err("Failed to stop txq %d, rc=%d", rc, txq->qid);
+
+ ethdev->data->tx_queue_state[queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
}
void
cnxk_rep_tx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(queue_id);
+ struct cnxk_rep_txq *txq = ethdev->data->tx_queues[queue_id];
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ int rc;
+
+ if (!txq) {
+ plt_err("Invalid txq retrieved for rep_id %d", rep_dev->rep_id);
+ return;
+ }
+
+ plt_rep_dbg("Releasing txq %u", txq->qid);
+
+ rc = cnxk_eswitch_txq_release(rep_dev->parent_dev, txq->qid);
+ if (rc)
+ plt_err("Failed to release txq %d, rc=%d", rc, txq->qid);
}
int
@@ -111,15 +509,70 @@ cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **op
return 0;
}
+int
+cnxk_rep_mac_addr_set(struct rte_eth_dev *eth_dev, struct rte_ether_addr *addr)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_eth_set_mac_meta_t msg_sm_meta;
+ cnxk_rep_msg_ack_data_t adata;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ /* If representor not representing any VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ size = CNXK_REP_MSG_MAX_BUFFER_SZ;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_sm_meta.portid = rep_dev->rep_id;
+ rte_memcpy(&msg_sm_meta.addr_bytes, addr->addr_bytes, RTE_ETHER_ADDR_LEN);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_sm_meta,
+ sizeof(cnxk_rep_msg_eth_set_mac_meta_t),
+ CNXK_REP_MSG_ETH_SET_MAC);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, &adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ if (adata.u.sval < 0) {
+ rc = adata.u.sval;
+ plt_err("Failed to set mac address, err %d", rc);
+ goto fail;
+ }
+
+ rte_free(buffer);
+
+ return 0;
+fail:
+ rte_free(buffer);
+ return rc;
+}
+
/* CNXK platform representor dev ops */
struct eth_dev_ops cnxk_rep_dev_ops = {
.dev_infos_get = cnxk_rep_dev_info_get,
+ .representor_info_get = cnxk_rep_representor_info_get,
.dev_configure = cnxk_rep_dev_configure,
.dev_start = cnxk_rep_dev_start,
.rx_queue_setup = cnxk_rep_rx_queue_setup,
.rx_queue_release = cnxk_rep_rx_queue_release,
.tx_queue_setup = cnxk_rep_tx_queue_setup,
.tx_queue_release = cnxk_rep_tx_queue_release,
+ .promiscuous_enable = cnxk_rep_promiscuous_enable,
+ .promiscuous_disable = cnxk_rep_promiscuous_disable,
+ .mac_addr_set = cnxk_rep_mac_addr_set,
.link_update = cnxk_rep_link_update,
.dev_close = cnxk_rep_dev_close,
.dev_stop = cnxk_rep_dev_stop,
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 14/23] common/cnxk: get representees ethernet stats
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (12 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 13/23] net/cnxk: representor ethdev ops Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 15/23] net/cnxk: ethernet statistic for representor Harman Kalra
` (8 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Implementing an mbox interface to fetch the representees's ethernet
stats from the kernel.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_eswitch.c | 45 +++++++++++++++++++++++++++++++
drivers/common/cnxk/roc_eswitch.h | 2 ++
drivers/common/cnxk/roc_mbox.h | 30 +++++++++++++++++++++
drivers/common/cnxk/version.map | 1 +
4 files changed, 78 insertions(+)
diff --git a/drivers/common/cnxk/roc_eswitch.c b/drivers/common/cnxk/roc_eswitch.c
index 14819bad75..c67b4090a5 100644
--- a/drivers/common/cnxk/roc_eswitch.c
+++ b/drivers/common/cnxk/roc_eswitch.c
@@ -342,3 +342,48 @@ roc_eswitch_nix_process_repte_notify_cb_unregister(struct roc_nix *roc_nix)
dev->ops->repte_notify = NULL;
}
+
+int
+roc_eswitch_nix_repte_stats(struct roc_nix *roc_nix, uint16_t pf_func, struct roc_nix_stats *stats)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+ struct dev *dev = &nix->dev;
+ struct nix_get_lf_stats_req *req;
+ struct nix_lf_stats_rsp *rsp;
+ struct mbox *mbox;
+ int rc;
+
+ mbox = mbox_get(dev->mbox);
+ req = mbox_alloc_msg_nix_get_lf_stats(mbox);
+ if (!req) {
+ rc = -ENOSPC;
+ goto exit;
+ }
+
+ req->hdr.pcifunc = roc_nix_get_pf_func(roc_nix);
+ req->pcifunc = pf_func;
+
+ rc = mbox_process_msg(mbox, (void *)&rsp);
+ if (rc)
+ goto exit;
+
+ stats->rx_octs = rsp->rx.octs;
+ stats->rx_ucast = rsp->rx.ucast;
+ stats->rx_bcast = rsp->rx.bcast;
+ stats->rx_mcast = rsp->rx.mcast;
+ stats->rx_drop = rsp->rx.drop;
+ stats->rx_drop_octs = rsp->rx.drop_octs;
+ stats->rx_drop_bcast = rsp->rx.drop_bcast;
+ stats->rx_drop_mcast = rsp->rx.drop_mcast;
+ stats->rx_err = rsp->rx.err;
+
+ stats->tx_ucast = rsp->tx.ucast;
+ stats->tx_bcast = rsp->tx.bcast;
+ stats->tx_mcast = rsp->tx.mcast;
+ stats->tx_drop = rsp->tx.drop;
+ stats->tx_octs = rsp->tx.octs;
+
+exit:
+ mbox_put(mbox);
+ return rc;
+}
diff --git a/drivers/common/cnxk/roc_eswitch.h b/drivers/common/cnxk/roc_eswitch.h
index f99f7e2981..5a0b7419c8 100644
--- a/drivers/common/cnxk/roc_eswitch.h
+++ b/drivers/common/cnxk/roc_eswitch.h
@@ -26,6 +26,8 @@ int __roc_api roc_eswitch_npc_rss_action_configure(struct roc_npc *roc_npc,
/* NIX */
int __roc_api roc_eswitch_nix_vlan_tpid_set(struct roc_nix *nix, uint32_t type, uint16_t tpid,
bool is_vf);
+int __roc_api roc_eswitch_nix_repte_stats(struct roc_nix *roc_nix, uint16_t pf_func,
+ struct roc_nix_stats *stats);
int __roc_api roc_eswitch_nix_process_repte_notify_cb_register(struct roc_nix *roc_nix,
process_repte_notify_t proc_repte_nt);
void __roc_api roc_eswitch_nix_process_repte_notify_cb_unregister(struct roc_nix *roc_nix);
diff --git a/drivers/common/cnxk/roc_mbox.h b/drivers/common/cnxk/roc_mbox.h
index 9de48c1907..39c1132792 100644
--- a/drivers/common/cnxk/roc_mbox.h
+++ b/drivers/common/cnxk/roc_mbox.h
@@ -304,6 +304,7 @@ struct mbox_msghdr {
M(NIX_MCAST_GRP_DESTROY, 0x802c, nix_mcast_grp_destroy, nix_mcast_grp_destroy_req, msg_rsp)\
M(NIX_MCAST_GRP_UPDATE, 0x802d, nix_mcast_grp_update, nix_mcast_grp_update_req, \
nix_mcast_grp_update_rsp) \
+ M(NIX_GET_LF_STATS, 0x802e, nix_get_lf_stats, nix_get_lf_stats_req, nix_lf_stats_rsp) \
/* MCS mbox IDs (range 0xa000 - 0xbFFF) */ \
M(MCS_ALLOC_RESOURCES, 0xa000, mcs_alloc_resources, mcs_alloc_rsrc_req, \
mcs_alloc_rsrc_rsp) \
@@ -1847,6 +1848,35 @@ struct nix_mcast_grp_update_rsp {
uint32_t __io mce_start_index;
};
+struct nix_get_lf_stats_req {
+ struct mbox_msghdr hdr;
+ uint16_t __io pcifunc;
+ uint64_t __io rsvd;
+};
+
+struct nix_lf_stats_rsp {
+ struct mbox_msghdr hdr;
+ struct {
+ uint64_t __io octs;
+ uint64_t __io ucast;
+ uint64_t __io bcast;
+ uint64_t __io mcast;
+ uint64_t __io drop;
+ uint64_t __io drop_octs;
+ uint64_t __io drop_mcast;
+ uint64_t __io drop_bcast;
+ uint64_t __io err;
+ uint64_t __io rsvd[5];
+ } rx;
+ struct {
+ uint64_t __io ucast;
+ uint64_t __io bcast;
+ uint64_t __io mcast;
+ uint64_t __io drop;
+ uint64_t __io octs;
+ } tx;
+};
+
/* Global NIX inline IPSec configuration */
struct nix_inline_ipsec_cfg {
struct mbox_msghdr hdr;
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index e170a6a63a..87c9d7511f 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -93,6 +93,7 @@ INTERNAL {
roc_error_msg_get;
roc_eswitch_nix_process_repte_notify_cb_register;
roc_eswitch_nix_process_repte_notify_cb_unregister;
+ roc_eswitch_nix_repte_stats;
roc_eswitch_nix_vlan_tpid_set;
roc_eswitch_npc_mcam_delete_rule;
roc_eswitch_npc_mcam_rx_rule;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 15/23] net/cnxk: ethernet statistic for representor
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (13 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 14/23] common/cnxk: get representees ethernet stats Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 16/23] common/cnxk: base support for eswitch VF Harman Kalra
` (7 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Adding representor ethernet statistics support which can fetch stats
for representees which are operating independently or part of
companian app.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_rep_msg.h | 7 ++
drivers/net/cnxk/cnxk_rep_ops.c | 140 +++++++++++++++++++++++++++++++-
2 files changed, 143 insertions(+), 4 deletions(-)
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
index 63cfbe3f19..277e25d92a 100644
--- a/drivers/net/cnxk/cnxk_rep_msg.h
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -21,6 +21,8 @@ typedef enum CNXK_REP_MSG {
CNXK_REP_MSG_EXIT,
/* Ethernet operation msgs */
CNXK_REP_MSG_ETH_SET_MAC,
+ CNXK_REP_MSG_ETH_STATS_GET,
+ CNXK_REP_MSG_ETH_STATS_CLEAR,
/* End of messaging sequence */
CNXK_REP_MSG_END,
} cnxk_rep_msg_t;
@@ -89,6 +91,11 @@ typedef struct cnxk_rep_msg_eth_mac_set_meta {
uint8_t addr_bytes[RTE_ETHER_ADDR_LEN];
} __rte_packed cnxk_rep_msg_eth_set_mac_meta_t;
+/* Ethernet op - get/clear stats */
+typedef struct cnxk_rep_msg_eth_stats_meta {
+ uint16_t portid;
+} __rte_packed cnxk_rep_msg_eth_stats_meta_t;
+
void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
uint32_t size);
void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
index 97643a50f2..6728b3e265 100644
--- a/drivers/net/cnxk/cnxk_rep_ops.c
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -486,19 +486,151 @@ cnxk_rep_tx_queue_release(struct rte_eth_dev *ethdev, uint16_t queue_id)
plt_err("Failed to release txq %d, rc=%d", rc, txq->qid);
}
+static int
+process_eth_stats(struct cnxk_rep_dev *rep_dev, cnxk_rep_msg_ack_data_t *adata, cnxk_rep_msg_t msg)
+{
+ cnxk_rep_msg_eth_stats_meta_t msg_st_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ size = CNXK_REP_MSG_MAX_BUFFER_SZ;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_st_meta.portid = rep_dev->rep_id;
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_st_meta,
+ sizeof(cnxk_rep_msg_eth_stats_meta_t), msg);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ rte_free(buffer);
+
+ return 0;
+fail:
+ rte_free(buffer);
+ return rc;
+}
+
+static int
+native_repte_eth_stats(struct cnxk_rep_dev *rep_dev, struct rte_eth_stats *stats)
+{
+ struct roc_nix_stats nix_stats;
+ int rc = 0;
+
+ rc = roc_eswitch_nix_repte_stats(&rep_dev->parent_dev->nix, rep_dev->hw_func, &nix_stats);
+ if (rc) {
+ plt_err("Failed to get stats for representee %x, err %d", rep_dev->hw_func, rc);
+ goto fail;
+ }
+
+ memset(stats, 0, sizeof(struct rte_eth_stats));
+ stats->opackets = nix_stats.tx_ucast;
+ stats->opackets += nix_stats.tx_mcast;
+ stats->opackets += nix_stats.tx_bcast;
+ stats->oerrors = nix_stats.tx_drop;
+ stats->obytes = nix_stats.tx_octs;
+
+ stats->ipackets = nix_stats.rx_ucast;
+ stats->ipackets += nix_stats.rx_mcast;
+ stats->ipackets += nix_stats.rx_bcast;
+ stats->imissed = nix_stats.rx_drop;
+ stats->ibytes = nix_stats.rx_octs;
+ stats->ierrors = nix_stats.rx_err;
+
+ return 0;
+fail:
+ return rc;
+}
+
int
cnxk_rep_stats_get(struct rte_eth_dev *ethdev, struct rte_eth_stats *stats)
{
- PLT_SET_USED(ethdev);
- PLT_SET_USED(stats);
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ struct rte_eth_stats vf_stats;
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ if (rep_dev->native_repte) {
+ /* For representees which are independent */
+ rc = native_repte_eth_stats(rep_dev, &vf_stats);
+ if (rc) {
+ plt_err("Failed to get stats for vf rep %x (hw_func %x), err %d",
+ rep_dev->port_id, rep_dev->hw_func, rc);
+ goto fail;
+ }
+ } else {
+ /* For representees which are part of companian app */
+ rc = process_eth_stats(rep_dev, &adata, CNXK_REP_MSG_ETH_STATS_GET);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ plt_err("Failed to get stats for vf rep %x, err %d", rep_dev->port_id, rc);
+ }
+
+ if (adata.size != sizeof(struct rte_eth_stats)) {
+ rc = -EINVAL;
+ plt_err("Incomplete stats received for vf rep %d", rep_dev->port_id);
+ goto fail;
+ }
+
+ rte_memcpy(&vf_stats, adata.u.data, adata.size);
+ }
+
+ stats->q_ipackets[0] = vf_stats.ipackets;
+ stats->q_ibytes[0] = vf_stats.ibytes;
+ stats->ipackets = vf_stats.ipackets;
+ stats->ibytes = vf_stats.ibytes;
+
+ stats->q_opackets[0] = vf_stats.opackets;
+ stats->q_obytes[0] = vf_stats.obytes;
+ stats->opackets = vf_stats.opackets;
+ stats->obytes = vf_stats.obytes;
+
+ plt_rep_dbg("Input packets %" PRId64 " Output packets %" PRId64 "", stats->ipackets,
+ stats->opackets);
+
return 0;
+fail:
+ return rc;
}
int
cnxk_rep_stats_reset(struct rte_eth_dev *ethdev)
{
- PLT_SET_USED(ethdev);
- return 0;
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(ethdev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc = 0;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ rc = process_eth_stats(rep_dev, &adata, CNXK_REP_MSG_ETH_STATS_CLEAR);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ plt_err("Failed to clear stats for vf rep %x, err %d", rep_dev->port_id, rc);
+ }
+
+ return rc;
}
int
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 16/23] common/cnxk: base support for eswitch VF
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (14 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 15/23] net/cnxk: ethernet statistic for representor Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 17/23] net/cnxk: eswitch VF as ethernet device Harman Kalra
` (6 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Base ROC layer changes for supporting eswitch VF and NIX lbk
changes for ESW
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_constants.h | 1 +
drivers/common/cnxk/roc_dev.c | 1 +
drivers/common/cnxk/roc_nix.c | 15 +++++++++++++--
drivers/common/cnxk/roc_nix.h | 1 +
drivers/common/cnxk/roc_nix_priv.h | 1 +
drivers/common/cnxk/version.map | 1 +
6 files changed, 18 insertions(+), 2 deletions(-)
diff --git a/drivers/common/cnxk/roc_constants.h b/drivers/common/cnxk/roc_constants.h
index cb4edbea58..21b3998cee 100644
--- a/drivers/common/cnxk/roc_constants.h
+++ b/drivers/common/cnxk/roc_constants.h
@@ -44,6 +44,7 @@
#define PCI_DEVID_CNXK_RVU_REE_PF 0xA0f4
#define PCI_DEVID_CNXK_RVU_REE_VF 0xA0f5
#define PCI_DEVID_CNXK_RVU_ESWITCH_PF 0xA0E0
+#define PCI_DEVID_CNXK_RVU_ESWITCH_VF 0xA0E1
#define PCI_DEVID_CN9K_CGX 0xA059
#define PCI_DEVID_CN10K_RPM 0xA060
diff --git a/drivers/common/cnxk/roc_dev.c b/drivers/common/cnxk/roc_dev.c
index 3564dbe647..0e83a92c43 100644
--- a/drivers/common/cnxk/roc_dev.c
+++ b/drivers/common/cnxk/roc_dev.c
@@ -1224,6 +1224,7 @@ dev_vf_hwcap_update(struct plt_pci_device *pci_dev, struct dev *dev)
case PCI_DEVID_CNXK_RVU_VF:
case PCI_DEVID_CNXK_RVU_SDP_VF:
case PCI_DEVID_CNXK_RVU_NIX_INL_VF:
+ case PCI_DEVID_CNXK_RVU_ESWITCH_VF:
dev->hwcap |= DEV_HWCAP_F_VF;
break;
}
diff --git a/drivers/common/cnxk/roc_nix.c b/drivers/common/cnxk/roc_nix.c
index 7e327a7e6e..f1eaca3ab4 100644
--- a/drivers/common/cnxk/roc_nix.c
+++ b/drivers/common/cnxk/roc_nix.c
@@ -13,6 +13,14 @@ roc_nix_is_lbk(struct roc_nix *roc_nix)
return nix->lbk_link;
}
+bool
+roc_nix_is_esw(struct roc_nix *roc_nix)
+{
+ struct nix *nix = roc_nix_to_nix_priv(roc_nix);
+
+ return nix->esw_link;
+}
+
int
roc_nix_get_base_chan(struct roc_nix *roc_nix)
{
@@ -156,7 +164,7 @@ roc_nix_max_pkt_len(struct roc_nix *roc_nix)
if (roc_model_is_cn9k())
return NIX_CN9K_MAX_HW_FRS;
- if (nix->lbk_link)
+ if (nix->lbk_link || nix->esw_link)
return NIX_LBK_MAX_HW_FRS;
return NIX_RPM_MAX_HW_FRS;
@@ -349,7 +357,7 @@ roc_nix_get_hw_info(struct roc_nix *roc_nix)
rc = mbox_process_msg(mbox, (void *)&hw_info);
if (rc == 0) {
nix->vwqe_interval = hw_info->vwqe_delay;
- if (nix->lbk_link)
+ if (nix->lbk_link || nix->esw_link)
roc_nix->dwrr_mtu = hw_info->lbk_dwrr_mtu;
else if (nix->sdp_link)
roc_nix->dwrr_mtu = hw_info->sdp_dwrr_mtu;
@@ -366,6 +374,7 @@ sdp_lbk_id_update(struct plt_pci_device *pci_dev, struct nix *nix)
{
nix->sdp_link = false;
nix->lbk_link = false;
+ nix->esw_link = false;
/* Update SDP/LBK link based on PCI device id */
switch (pci_dev->id.device_id) {
@@ -374,7 +383,9 @@ sdp_lbk_id_update(struct plt_pci_device *pci_dev, struct nix *nix)
nix->sdp_link = true;
break;
case PCI_DEVID_CNXK_RVU_AF_VF:
+ case PCI_DEVID_CNXK_RVU_ESWITCH_VF:
nix->lbk_link = true;
+ nix->esw_link = true;
break;
default:
break;
diff --git a/drivers/common/cnxk/roc_nix.h b/drivers/common/cnxk/roc_nix.h
index b369335fc4..ffea84dae8 100644
--- a/drivers/common/cnxk/roc_nix.h
+++ b/drivers/common/cnxk/roc_nix.h
@@ -527,6 +527,7 @@ int __roc_api roc_nix_dev_fini(struct roc_nix *roc_nix);
/* Type */
bool __roc_api roc_nix_is_lbk(struct roc_nix *roc_nix);
+bool __roc_api roc_nix_is_esw(struct roc_nix *roc_nix);
bool __roc_api roc_nix_is_sdp(struct roc_nix *roc_nix);
bool __roc_api roc_nix_is_pf(struct roc_nix *roc_nix);
bool __roc_api roc_nix_is_vf_or_sdp(struct roc_nix *roc_nix);
diff --git a/drivers/common/cnxk/roc_nix_priv.h b/drivers/common/cnxk/roc_nix_priv.h
index 3d99ade2b4..275ffc8ea3 100644
--- a/drivers/common/cnxk/roc_nix_priv.h
+++ b/drivers/common/cnxk/roc_nix_priv.h
@@ -170,6 +170,7 @@ struct nix {
uintptr_t base;
bool sdp_link;
bool lbk_link;
+ bool esw_link;
bool ptp_en;
bool is_nix1;
diff --git a/drivers/common/cnxk/version.map b/drivers/common/cnxk/version.map
index 87c9d7511f..cdb46d8739 100644
--- a/drivers/common/cnxk/version.map
+++ b/drivers/common/cnxk/version.map
@@ -276,6 +276,7 @@ INTERNAL {
roc_nix_inl_outb_cpt_lfs_dump;
roc_nix_cpt_ctx_cache_sync;
roc_nix_is_lbk;
+ roc_nix_is_esw;
roc_nix_is_pf;
roc_nix_is_sdp;
roc_nix_is_vf_or_sdp;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 17/23] net/cnxk: eswitch VF as ethernet device
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (15 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 16/23] common/cnxk: base support for eswitch VF Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 18/23] common/cnxk: support port representor and represented port Harman Kalra
` (5 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Adding support for eswitch VF to probe as normal cnxk ethernet device
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cn10k_ethdev.c | 1 +
drivers/net/cnxk/cnxk_ethdev.c | 41 +++++++++++++++++++++---------
drivers/net/cnxk/cnxk_ethdev.h | 3 +++
drivers/net/cnxk/cnxk_ethdev_ops.c | 4 +++
drivers/net/cnxk/cnxk_link.c | 3 ++-
5 files changed, 39 insertions(+), 13 deletions(-)
diff --git a/drivers/net/cnxk/cn10k_ethdev.c b/drivers/net/cnxk/cn10k_ethdev.c
index a2e943a3d0..9a072b72a7 100644
--- a/drivers/net/cnxk/cn10k_ethdev.c
+++ b/drivers/net/cnxk/cn10k_ethdev.c
@@ -963,6 +963,7 @@ static const struct rte_pci_id cn10k_pci_nix_map[] = {
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CN10KB, PCI_DEVID_CNXK_RVU_PF),
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CNF10KB, PCI_DEVID_CNXK_RVU_PF),
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CN10KA, PCI_DEVID_CNXK_RVU_VF),
+ CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CN10KA, PCI_DEVID_CNXK_RVU_ESWITCH_VF),
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CN10KAS, PCI_DEVID_CNXK_RVU_VF),
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CNF10KA, PCI_DEVID_CNXK_RVU_VF),
CNXK_PCI_ID(PCI_SUBSYSTEM_DEVID_CN10KB, PCI_DEVID_CNXK_RVU_VF),
diff --git a/drivers/net/cnxk/cnxk_ethdev.c b/drivers/net/cnxk/cnxk_ethdev.c
index 2372a4e793..ec43de6221 100644
--- a/drivers/net/cnxk/cnxk_ethdev.c
+++ b/drivers/net/cnxk/cnxk_ethdev.c
@@ -390,7 +390,7 @@ nix_update_flow_ctrl_config(struct rte_eth_dev *eth_dev)
struct cnxk_fc_cfg *fc = &dev->fc_cfg;
struct rte_eth_fc_conf fc_cfg = {0};
- if (roc_nix_is_sdp(&dev->nix))
+ if (roc_nix_is_sdp(&dev->nix) || roc_nix_is_esw(&dev->nix))
return 0;
/* Don't do anything if PFC is enabled */
@@ -1449,12 +1449,14 @@ cnxk_nix_configure(struct rte_eth_dev *eth_dev)
goto cq_fini;
/* Init flow control configuration */
- fc_cfg.type = ROC_NIX_FC_RXCHAN_CFG;
- fc_cfg.rxchan_cfg.enable = true;
- rc = roc_nix_fc_config_set(nix, &fc_cfg);
- if (rc) {
- plt_err("Failed to initialize flow control rc=%d", rc);
- goto cq_fini;
+ if (!roc_nix_is_esw(nix)) {
+ fc_cfg.type = ROC_NIX_FC_RXCHAN_CFG;
+ fc_cfg.rxchan_cfg.enable = true;
+ rc = roc_nix_fc_config_set(nix, &fc_cfg);
+ if (rc) {
+ plt_err("Failed to initialize flow control rc=%d", rc);
+ goto cq_fini;
+ }
}
/* Update flow control configuration to PMD */
@@ -1976,11 +1978,21 @@ cnxk_eth_dev_init(struct rte_eth_dev *eth_dev)
TAILQ_INIT(&dev->mcs_list);
}
- plt_nix_dbg("Port=%d pf=%d vf=%d ver=%s hwcap=0x%" PRIx64
- " rxoffload_capa=0x%" PRIx64 " txoffload_capa=0x%" PRIx64,
- eth_dev->data->port_id, roc_nix_get_pf(nix),
- roc_nix_get_vf(nix), CNXK_ETH_DEV_PMD_VERSION, dev->hwcap,
- dev->rx_offload_capa, dev->tx_offload_capa);
+ /* Reserve a switch domain for eswitch device */
+ if (pci_dev->id.device_id == PCI_DEVID_CNXK_RVU_ESWITCH_VF) {
+ eth_dev->data->dev_flags |= RTE_ETH_DEV_REPRESENTOR;
+ rc = rte_eth_switch_domain_alloc(&dev->switch_domain_id);
+ if (rc) {
+ plt_err("Failed to alloc switch domain: %d", rc);
+ goto free_mac_addrs;
+ }
+ }
+
+ plt_nix_dbg("Port=%d pf=%d vf=%d ver=%s hwcap=0x%" PRIx64 " rxoffload_capa=0x%" PRIx64
+ " txoffload_capa=0x%" PRIx64,
+ eth_dev->data->port_id, roc_nix_get_pf(nix), roc_nix_get_vf(nix),
+ CNXK_ETH_DEV_PMD_VERSION, dev->hwcap, dev->rx_offload_capa,
+ dev->tx_offload_capa);
return 0;
free_mac_addrs:
@@ -2046,6 +2058,11 @@ cnxk_eth_dev_uninit(struct rte_eth_dev *eth_dev, bool reset)
}
}
+ /* Free switch domain ID reserved for eswitch device */
+ if ((eth_dev->data->dev_flags & RTE_ETH_DEV_REPRESENTOR) &&
+ rte_eth_switch_domain_free(dev->switch_domain_id))
+ plt_err("Failed to free switch domain");
+
/* Disable and free rte_meter entries */
nix_meter_fini(dev);
diff --git a/drivers/net/cnxk/cnxk_ethdev.h b/drivers/net/cnxk/cnxk_ethdev.h
index 4d3ebf123b..d8eba5e1dd 100644
--- a/drivers/net/cnxk/cnxk_ethdev.h
+++ b/drivers/net/cnxk/cnxk_ethdev.h
@@ -424,6 +424,9 @@ struct cnxk_eth_dev {
/* MCS device */
struct cnxk_mcs_dev *mcs_dev;
struct cnxk_macsec_sess_list mcs_list;
+
+ /* Eswitch domain ID */
+ uint16_t switch_domain_id;
};
struct cnxk_eth_rxq_sp {
diff --git a/drivers/net/cnxk/cnxk_ethdev_ops.c b/drivers/net/cnxk/cnxk_ethdev_ops.c
index 5de2919047..67fbf7c269 100644
--- a/drivers/net/cnxk/cnxk_ethdev_ops.c
+++ b/drivers/net/cnxk/cnxk_ethdev_ops.c
@@ -71,6 +71,10 @@ cnxk_nix_info_get(struct rte_eth_dev *eth_dev, struct rte_eth_dev_info *devinfo)
RTE_ETH_DEV_CAPA_FLOW_RULE_KEEP;
devinfo->max_rx_mempools = CNXK_NIX_NUM_POOLS_MAX;
+ if (eth_dev->data->dev_flags & RTE_ETH_DEV_REPRESENTOR) {
+ devinfo->switch_info.name = eth_dev->device->name;
+ devinfo->switch_info.domain_id = dev->switch_domain_id;
+ }
return 0;
}
diff --git a/drivers/net/cnxk/cnxk_link.c b/drivers/net/cnxk/cnxk_link.c
index 127c9e72e7..903b44de2c 100644
--- a/drivers/net/cnxk/cnxk_link.c
+++ b/drivers/net/cnxk/cnxk_link.c
@@ -13,7 +13,8 @@ cnxk_nix_toggle_flag_link_cfg(struct cnxk_eth_dev *dev, bool set)
dev->flags &= ~CNXK_LINK_CFG_IN_PROGRESS_F;
/* Update link info for LBK */
- if (!set && (roc_nix_is_lbk(&dev->nix) || roc_nix_is_sdp(&dev->nix))) {
+ if (!set &&
+ (roc_nix_is_lbk(&dev->nix) || roc_nix_is_sdp(&dev->nix) || roc_nix_is_esw(&dev->nix))) {
struct rte_eth_link link;
link.link_status = RTE_ETH_LINK_UP;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 18/23] common/cnxk: support port representor and represented port
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (16 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 17/23] net/cnxk: eswitch VF as ethernet device Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 19/23] net/cnxk: add represented port pattern and action Harman Kalra
` (4 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, Satheesh Paul
From: Kiran Kumar K <kirankumark@marvell.com>
Implementing the common infrastructural changes for supporting port
representors and represented ports used as action and pattern in net
layer.
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Satheesh Paul <psatheesh@marvell.com>
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/common/cnxk/roc_npc.c | 63 +++++++++++++++++++++++------
drivers/common/cnxk/roc_npc.h | 13 +++++-
drivers/common/cnxk/roc_npc_mcam.c | 62 +++++++++++++++-------------
drivers/common/cnxk/roc_npc_parse.c | 28 ++++++++++++-
drivers/common/cnxk/roc_npc_priv.h | 2 +
drivers/net/cnxk/cnxk_flow.c | 2 +-
6 files changed, 125 insertions(+), 45 deletions(-)
diff --git a/drivers/common/cnxk/roc_npc.c b/drivers/common/cnxk/roc_npc.c
index 67a660a2bc..f4fb8c7660 100644
--- a/drivers/common/cnxk/roc_npc.c
+++ b/drivers/common/cnxk/roc_npc.c
@@ -570,6 +570,8 @@ npc_parse_actions(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
flow->ctr_id = NPC_COUNTER_NONE;
flow->mtr_id = ROC_NIX_MTR_ID_INVALID;
pf_func = npc->pf_func;
+ if (flow->has_rep)
+ pf_func = flow->rep_pf_func;
for (; actions->type != ROC_NPC_ACTION_TYPE_END; actions++) {
switch (actions->type) {
@@ -898,10 +900,14 @@ npc_parse_pattern(struct npc *npc, const struct roc_npc_item_info pattern[],
struct roc_npc_flow *flow, struct npc_parse_state *pst)
{
npc_parse_stage_func_t parse_stage_funcs[] = {
- npc_parse_meta_items, npc_parse_mark_item, npc_parse_pre_l2, npc_parse_cpt_hdr,
- npc_parse_higig2_hdr, npc_parse_tx_queue, npc_parse_la, npc_parse_lb,
- npc_parse_lc, npc_parse_ld, npc_parse_le, npc_parse_lf,
- npc_parse_lg, npc_parse_lh,
+ npc_parse_meta_items, npc_parse_port_representor_id,
+ npc_parse_mark_item, npc_parse_pre_l2,
+ npc_parse_cpt_hdr, npc_parse_higig2_hdr,
+ npc_parse_tx_queue, npc_parse_la,
+ npc_parse_lb, npc_parse_lc,
+ npc_parse_ld, npc_parse_le,
+ npc_parse_lf, npc_parse_lg,
+ npc_parse_lh,
};
uint8_t layer = 0;
int key_offset;
@@ -1140,15 +1146,20 @@ npc_rss_action_program(struct roc_npc *roc_npc,
struct roc_npc_flow *flow)
{
const struct roc_npc_action_rss *rss;
+ struct roc_npc *npc = roc_npc;
uint32_t rss_grp;
uint8_t alg_idx;
int rc;
+ if (flow->has_rep) {
+ npc = roc_npc->rep_npc;
+ npc->flowkey_cfg_state = roc_npc->flowkey_cfg_state;
+ }
+
for (; actions->type != ROC_NPC_ACTION_TYPE_END; actions++) {
if (actions->type == ROC_NPC_ACTION_TYPE_RSS) {
rss = (const struct roc_npc_action_rss *)actions->conf;
- rc = npc_rss_action_configure(roc_npc, rss, &alg_idx,
- &rss_grp, flow->mcam_id);
+ rc = npc_rss_action_configure(npc, rss, &alg_idx, &rss_grp, flow->mcam_id);
if (rc)
return rc;
@@ -1171,7 +1182,7 @@ npc_vtag_cfg_delete(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
struct roc_nix *roc_nix = roc_npc->roc_nix;
struct nix_vtag_config *vtag_cfg;
struct nix_vtag_config_rsp *rsp;
- struct mbox *mbox;
+ struct mbox *mbox, *ombox;
struct nix *nix;
int rc = 0;
@@ -1181,7 +1192,10 @@ npc_vtag_cfg_delete(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
} tx_vtag_action;
nix = roc_nix_to_nix_priv(roc_nix);
- mbox = mbox_get((&nix->dev)->mbox);
+ ombox = (&nix->dev)->mbox;
+ if (flow->has_rep)
+ ombox = flow->rep_mbox;
+ mbox = mbox_get(ombox);
tx_vtag_action.reg = flow->vtag_action;
vtag_cfg = mbox_alloc_msg_nix_vtag_cfg(mbox);
@@ -1400,6 +1414,7 @@ npc_vtag_strip_action_configure(struct mbox *mbox,
rx_vtag_action |= (NIX_RX_VTAGACTION_VTAG_VALID << 15);
rx_vtag_action |= ((uint64_t)NPC_LID_LB << 8);
+ rx_vtag_action |= ((uint64_t)NIX_RX_VTAG_TYPE6 << 12);
rx_vtag_action |= NIX_RX_VTAGACTION_VTAG0_RELPTR;
if (*strip_cnt == 2) {
@@ -1432,6 +1447,8 @@ npc_vtag_action_program(struct roc_npc *roc_npc,
nix = roc_nix_to_nix_priv(roc_nix);
mbox = (&nix->dev)->mbox;
+ if (flow->has_rep)
+ mbox = flow->rep_mbox;
memset(vlan_info, 0, sizeof(vlan_info));
@@ -1448,6 +1465,7 @@ npc_vtag_action_program(struct roc_npc *roc_npc,
if (rc)
return rc;
+ plt_npc_dbg("VLAN strip action, strip_cnt %d", strip_cnt);
if (strip_cnt == 2)
actions++;
@@ -1587,6 +1605,17 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
memset(flow, 0, sizeof(*flow));
memset(&parse_state, 0, sizeof(parse_state));
+ flow->port_id = -1;
+ if (roc_npc->rep_npc) {
+ flow->rep_channel = roc_nix_to_nix_priv(roc_npc->rep_npc->roc_nix)->rx_chan_base;
+ flow->rep_pf_func = roc_npc->rep_pf_func;
+ flow->rep_mbox = roc_npc_to_npc_priv(roc_npc->rep_npc)->mbox;
+ flow->has_rep = true;
+ flow->is_rep_vf = !roc_nix_is_pf(roc_npc->rep_npc->roc_nix);
+ flow->port_id = roc_npc->rep_port_id;
+ flow->rep_npc = roc_npc_to_npc_priv(roc_npc->rep_npc);
+ }
+
parse_state.dst_pf_func = dst_pf_func;
rc = npc_parse_rule(roc_npc, attr, pattern, actions, flow, &parse_state);
@@ -1629,6 +1658,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
*errcode = rc;
goto set_rss_failed;
}
+ roc_npc->rep_npc = NULL;
if (flow->has_age_action)
npc_age_flow_list_entry_add(roc_npc, flow);
@@ -1641,6 +1671,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
TAILQ_FOREACH(flow_iter, list, next) {
if (flow_iter->mcam_id > flow->mcam_id) {
TAILQ_INSERT_BEFORE(flow_iter, flow, next);
+ roc_npc->rep_npc = NULL;
return flow;
}
}
@@ -1649,6 +1680,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
return flow;
set_rss_failed:
+ roc_npc->rep_npc = NULL;
if (flow->use_pre_alloc == 0) {
rc = roc_npc_mcam_free_entry(roc_npc, flow->mcam_id);
if (rc != 0) {
@@ -1660,6 +1692,7 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
npc_inline_dev_ipsec_action_free(npc, flow);
}
err_exit:
+ roc_npc->rep_npc = NULL;
plt_free(flow);
return NULL;
}
@@ -1667,15 +1700,19 @@ roc_npc_flow_create(struct roc_npc *roc_npc, const struct roc_npc_attr *attr,
int
npc_rss_group_free(struct npc *npc, struct roc_npc_flow *flow)
{
+ struct npc *lnpc = npc;
uint32_t rss_grp;
+ if (flow->has_rep)
+ lnpc = flow->rep_npc;
+
if ((flow->npc_action & 0xF) == NIX_RX_ACTIONOP_RSS) {
rss_grp = (flow->npc_action >> NPC_RSS_ACT_GRP_OFFSET) &
NPC_RSS_ACT_GRP_MASK;
if (rss_grp == 0 || rss_grp >= npc->rss_grps)
return -EINVAL;
- plt_bitmap_clear(npc->rss_grp_entries, rss_grp);
+ plt_bitmap_clear(lnpc->rss_grp_entries, rss_grp);
}
return 0;
@@ -1770,7 +1807,7 @@ roc_npc_flow_destroy(struct roc_npc *roc_npc, struct roc_npc_flow *flow)
}
void
-roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc)
+roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc, int rep_port_id)
{
struct npc *npc = roc_npc_to_npc_priv(roc_npc);
struct roc_npc_flow *flow_iter;
@@ -1784,12 +1821,14 @@ roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc)
/* List in ascending order of mcam entries */
TAILQ_FOREACH(flow_iter, list, next) {
- roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
+ if (rep_port_id == -1 || rep_port_id == flow_iter->port_id)
+ roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
}
}
TAILQ_FOREACH(flow_iter, &npc->ipsec_list, next) {
- roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
+ if (rep_port_id == -1 || rep_port_id == flow_iter->port_id)
+ roc_npc_flow_mcam_dump(file, roc_npc, flow_iter);
}
}
diff --git a/drivers/common/cnxk/roc_npc.h b/drivers/common/cnxk/roc_npc.h
index 349c7f9d22..4d962abe50 100644
--- a/drivers/common/cnxk/roc_npc.h
+++ b/drivers/common/cnxk/roc_npc.h
@@ -42,6 +42,7 @@ enum roc_npc_item_type {
ROC_NPC_ITEM_TYPE_MARK,
ROC_NPC_ITEM_TYPE_TX_QUEUE,
ROC_NPC_ITEM_TYPE_IPV6_ROUTING_EXT,
+ ROC_NPC_ITEM_TYPE_REPRESENTED_PORT,
ROC_NPC_ITEM_TYPE_END,
};
@@ -332,6 +333,13 @@ struct roc_npc_flow {
void *age_context;
uint32_t timeout;
bool has_age_action;
+ uint16_t rep_pf_func;
+ uint16_t rep_channel;
+ struct mbox *rep_mbox;
+ bool has_rep;
+ bool is_rep_vf;
+ struct npc *rep_npc;
+ int port_id;
bool is_sampling_rule;
uint32_t recv_queue;
uint32_t mcast_grp_index;
@@ -407,6 +415,9 @@ struct roc_npc {
uint16_t sdp_channel;
uint16_t sdp_channel_mask;
struct roc_npc_flow_age flow_age;
+ struct roc_npc *rep_npc;
+ uint16_t rep_pf_func;
+ int rep_port_id;
#define ROC_NPC_MEM_SZ (6 * 1024)
uint8_t reserved[ROC_NPC_MEM_SZ];
@@ -448,7 +459,7 @@ int __roc_api roc_npc_get_free_mcam_entry(struct roc_npc *roc_npc, struct roc_np
int __roc_api roc_npc_inl_mcam_read_counter(uint32_t ctr_id, uint64_t *count);
int __roc_api roc_npc_inl_mcam_clear_counter(uint32_t ctr_id);
int __roc_api roc_npc_mcam_free_all_resources(struct roc_npc *roc_npc);
-void __roc_api roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc);
+void __roc_api roc_npc_flow_dump(FILE *file, struct roc_npc *roc_npc, int rep_port_id);
void __roc_api roc_npc_flow_mcam_dump(FILE *file, struct roc_npc *roc_npc,
struct roc_npc_flow *mcam);
int __roc_api roc_npc_mark_actions_get(struct roc_npc *roc_npc);
diff --git a/drivers/common/cnxk/roc_npc_mcam.c b/drivers/common/cnxk/roc_npc_mcam.c
index 2de988a44b..f2d5004c78 100644
--- a/drivers/common/cnxk/roc_npc_mcam.c
+++ b/drivers/common/cnxk/roc_npc_mcam.c
@@ -143,8 +143,8 @@ npc_lid_lt_in_kex(struct npc *npc, uint8_t lid, uint8_t lt)
}
static void
-npc_construct_ldata_mask(struct npc *npc, struct plt_bitmap *bmap, uint8_t lid,
- uint8_t lt, uint8_t ld)
+npc_construct_ldata_mask(struct npc *npc, struct plt_bitmap *bmap, uint8_t lid, uint8_t lt,
+ uint8_t ld)
{
struct npc_xtract_info *x_info, *infoflag;
int hdr_off, keylen;
@@ -197,8 +197,7 @@ npc_construct_ldata_mask(struct npc *npc, struct plt_bitmap *bmap, uint8_t lid,
* @param len length of the match
*/
static bool
-npc_is_kex_enabled(struct npc *npc, uint8_t lid, uint8_t lt, int offset,
- int len)
+npc_is_kex_enabled(struct npc *npc, uint8_t lid, uint8_t lt, int offset, int len)
{
struct plt_bitmap *bmap;
uint32_t bmap_sz;
@@ -349,8 +348,8 @@ npc_mcam_alloc_entries(struct mbox *mbox, int ref_mcam, int *alloc_entry, int re
}
int
-npc_mcam_alloc_entry(struct npc *npc, struct roc_npc_flow *mcam,
- struct roc_npc_flow *ref_mcam, int prio, int *resp_count)
+npc_mcam_alloc_entry(struct npc *npc, struct roc_npc_flow *mcam, struct roc_npc_flow *ref_mcam,
+ int prio, int *resp_count)
{
struct npc_mcam_alloc_entry_req *req;
struct npc_mcam_alloc_entry_rsp *rsp;
@@ -450,22 +449,17 @@ npc_mcam_write_entry(struct mbox *mbox, struct roc_npc_flow *mcam)
static void
npc_mcam_process_mkex_cfg(struct npc *npc, struct npc_get_kex_cfg_rsp *kex_rsp)
{
- volatile uint64_t(
- *q)[NPC_MAX_INTF][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD];
+ volatile uint64_t(*q)[NPC_MAX_INTF][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD];
struct npc_xtract_info *x_info = NULL;
int lid, lt, ld, fl, ix;
npc_dxcfg_t *p;
uint64_t keyw;
uint64_t val;
- npc->keyx_supp_nmask[NPC_MCAM_RX] =
- kex_rsp->rx_keyx_cfg & 0x7fffffffULL;
- npc->keyx_supp_nmask[NPC_MCAM_TX] =
- kex_rsp->tx_keyx_cfg & 0x7fffffffULL;
- npc->keyx_len[NPC_MCAM_RX] =
- npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_RX]);
- npc->keyx_len[NPC_MCAM_TX] =
- npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_TX]);
+ npc->keyx_supp_nmask[NPC_MCAM_RX] = kex_rsp->rx_keyx_cfg & 0x7fffffffULL;
+ npc->keyx_supp_nmask[NPC_MCAM_TX] = kex_rsp->tx_keyx_cfg & 0x7fffffffULL;
+ npc->keyx_len[NPC_MCAM_RX] = npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_RX]);
+ npc->keyx_len[NPC_MCAM_TX] = npc_supp_key_len(npc->keyx_supp_nmask[NPC_MCAM_TX]);
keyw = (kex_rsp->rx_keyx_cfg >> 32) & 0x7ULL;
npc->keyw[NPC_MCAM_RX] = keyw;
@@ -485,8 +479,7 @@ npc_mcam_process_mkex_cfg(struct npc *npc, struct npc_get_kex_cfg_rsp *kex_rsp)
/* Update LID, LT and LDATA cfg */
p = &npc->prx_dxcfg;
- q = (volatile uint64_t(*)[][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD])(
- &kex_rsp->intf_lid_lt_ld);
+ q = (volatile uint64_t(*)[][NPC_MAX_LID][NPC_MAX_LT][NPC_MAX_LD])(&kex_rsp->intf_lid_lt_ld);
for (ix = 0; ix < NPC_MAX_INTF; ix++) {
for (lid = 0; lid < NPC_MAX_LID; lid++) {
for (lt = 0; lt < NPC_MAX_LT; lt++) {
@@ -539,8 +532,7 @@ npc_mcam_fetch_kex_cfg(struct npc *npc)
goto done;
}
- mbox_memcpy((char *)npc->profile_name, kex_rsp->mkex_pfl_name,
- MKEX_NAME_LEN);
+ mbox_memcpy((char *)npc->profile_name, kex_rsp->mkex_pfl_name, MKEX_NAME_LEN);
npc->exact_match_ena = (kex_rsp->rx_keyx_cfg >> 40) & 0xF;
npc_mcam_process_mkex_cfg(npc, kex_rsp);
@@ -551,9 +543,8 @@ npc_mcam_fetch_kex_cfg(struct npc *npc)
}
static void
-npc_mcam_set_channel(struct roc_npc_flow *flow,
- struct npc_mcam_write_entry_req *req, uint16_t channel,
- uint16_t chan_mask, bool is_second_pass)
+npc_mcam_set_channel(struct roc_npc_flow *flow, struct npc_mcam_write_entry_req *req,
+ uint16_t channel, uint16_t chan_mask, bool is_second_pass)
{
uint16_t chan = 0, mask = 0;
@@ -683,6 +674,9 @@ npc_mcam_alloc_and_write(struct npc *npc, struct roc_npc_flow *flow, struct npc_
if (flow->nix_intf == NIX_INTF_TX) {
uint16_t pf_func = (flow->npc_action >> 4) & 0xffff;
+ if (flow->has_rep)
+ pf_func = flow->rep_pf_func;
+
pf_func = plt_cpu_to_be_16(pf_func);
rc = npc_mcam_set_pf_func(npc, flow, pf_func);
@@ -759,6 +753,14 @@ npc_mcam_alloc_and_write(struct npc *npc, struct roc_npc_flow *flow, struct npc_
npc_mcam_set_channel(flow, req, inl_dev->channel, inl_dev->chan_mask,
false);
+ } else if (flow->has_rep) {
+ pf_func = flow->rep_pf_func;
+ req->entry_data.action &= ~(GENMASK(19, 4));
+ req->entry_data.action |= (uint64_t)pf_func << 4;
+ flow->npc_action &= ~(GENMASK(19, 4));
+ flow->npc_action |= (uint64_t)pf_func << 4;
+ npc_mcam_set_channel(flow, req, flow->rep_channel, (BIT_ULL(12) - 1),
+ false);
} else if (npc->is_sdp_link) {
npc_mcam_set_channel(flow, req, npc->sdp_channel, npc->sdp_channel_mask,
pst->is_second_pass_rule);
@@ -932,13 +934,11 @@ npc_program_mcam(struct npc *npc, struct npc_parse_state *pst, bool mcam_alloc)
data_off = 0;
index++;
}
- key_data[index] |=
- ((uint64_t)data << data_off);
+ key_data[index] |= ((uint64_t)data << data_off);
if (lt == 0)
mask = 0;
- key_mask[index] |=
- ((uint64_t)mask << data_off);
+ key_mask[index] |= ((uint64_t)mask << data_off);
data_off += 4;
}
}
@@ -963,8 +963,12 @@ npc_program_mcam(struct npc *npc, struct npc_parse_state *pst, bool mcam_alloc)
(pst->flow->npc_action & NIX_RX_ACTIONOP_UCAST_IPSEC))
skip_base_rule = true;
- if (pst->is_vf && pst->flow->nix_intf == NIX_INTF_RX && !skip_base_rule) {
- mbox = mbox_get(npc->mbox);
+ if ((pst->is_vf || pst->flow->is_rep_vf) && pst->flow->nix_intf == NIX_INTF_RX &&
+ !skip_base_rule) {
+ if (pst->flow->has_rep)
+ mbox = mbox_get(pst->flow->rep_mbox);
+ else
+ mbox = mbox_get(npc->mbox);
(void)mbox_alloc_msg_npc_read_base_steer_rule(mbox);
rc = mbox_process_msg(mbox, (void *)&base_rule_rsp);
if (rc) {
diff --git a/drivers/common/cnxk/roc_npc_parse.c b/drivers/common/cnxk/roc_npc_parse.c
index 571d6b8e5d..0aaf86c768 100644
--- a/drivers/common/cnxk/roc_npc_parse.c
+++ b/drivers/common/cnxk/roc_npc_parse.c
@@ -35,11 +35,35 @@ npc_parse_mark_item(struct npc_parse_state *pst)
return 0;
}
+int
+npc_parse_port_representor_id(struct npc_parse_state *pst)
+{
+ if (pst->pattern->type != ROC_NPC_ITEM_TYPE_REPRESENTED_PORT)
+ return 0;
+
+ pst->pattern++;
+
+ return 0;
+}
+
+int
+npc_parse_represented_port_id(struct npc_parse_state *pst)
+{
+ if (pst->pattern->type != ROC_NPC_ITEM_TYPE_REPRESENTED_PORT)
+ return 0;
+
+ if (pst->flow->nix_intf != NIX_INTF_RX)
+ return -EINVAL;
+
+ pst->pattern++;
+
+ return 0;
+}
+
static int
npc_flow_raw_item_prepare(const struct roc_npc_flow_item_raw *raw_spec,
const struct roc_npc_flow_item_raw *raw_mask,
- struct npc_parse_item_info *info, uint8_t *spec_buf,
- uint8_t *mask_buf)
+ struct npc_parse_item_info *info, uint8_t *spec_buf, uint8_t *mask_buf)
{
memset(spec_buf, 0, NPC_MAX_RAW_ITEM_LEN);
diff --git a/drivers/common/cnxk/roc_npc_priv.h b/drivers/common/cnxk/roc_npc_priv.h
index 50b62b1244..069c625911 100644
--- a/drivers/common/cnxk/roc_npc_priv.h
+++ b/drivers/common/cnxk/roc_npc_priv.h
@@ -457,6 +457,8 @@ int npc_mask_is_supported(const char *mask, const char *hw_mask, int len);
int npc_parse_item_basic(const struct roc_npc_item_info *item, struct npc_parse_item_info *info);
int npc_parse_meta_items(struct npc_parse_state *pst);
int npc_parse_mark_item(struct npc_parse_state *pst);
+int npc_parse_port_representor_id(struct npc_parse_state *pst);
+int npc_parse_represented_port_id(struct npc_parse_state *pst);
int npc_parse_pre_l2(struct npc_parse_state *pst);
int npc_parse_higig2_hdr(struct npc_parse_state *pst);
int npc_parse_cpt_hdr(struct npc_parse_state *pst);
diff --git a/drivers/net/cnxk/cnxk_flow.c b/drivers/net/cnxk/cnxk_flow.c
index a92b61c332..5f74c356b1 100644
--- a/drivers/net/cnxk/cnxk_flow.c
+++ b/drivers/net/cnxk/cnxk_flow.c
@@ -594,7 +594,7 @@ cnxk_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
return -EINVAL;
}
- roc_npc_flow_dump(file, npc);
+ roc_npc_flow_dump(file, npc, -1);
return 0;
}
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 19/23] net/cnxk: add represented port pattern and action
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (17 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 18/23] common/cnxk: support port representor and represented port Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 20/23] net/cnxk: add representor " Harman Kalra
` (3 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev, Satheesh Paul
From: Kiran Kumar K <kirankumark@marvell.com>
Adding support for represented_port item matching and action.
Signed-off-by: Kiran Kumar K <kirankumark@marvell.com>
Signed-off-by: Satheesh Paul <psatheesh@marvell.com>
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
doc/guides/nics/cnxk.rst | 1 +
doc/guides/nics/features/cnxk.ini | 1 +
doc/guides/nics/features/cnxk_vec.ini | 2 +
doc/guides/nics/features/cnxk_vf.ini | 2 +
drivers/net/cnxk/cnxk_flow.c | 107 ++++++++++++++------------
5 files changed, 63 insertions(+), 50 deletions(-)
diff --git a/doc/guides/nics/cnxk.rst b/doc/guides/nics/cnxk.rst
index 496474913f..0f80d65562 100644
--- a/doc/guides/nics/cnxk.rst
+++ b/doc/guides/nics/cnxk.rst
@@ -38,6 +38,7 @@ Features of the CNXK Ethdev PMD are:
- Ingress meter support
- Queue based priority flow control support
- Port representors
+- Represented port pattern matching and action
Prerequisites
-------------
diff --git a/doc/guides/nics/features/cnxk.ini b/doc/guides/nics/features/cnxk.ini
index 94e7a6ab8d..8957eba46b 100644
--- a/doc/guides/nics/features/cnxk.ini
+++ b/doc/guides/nics/features/cnxk.ini
@@ -73,6 +73,7 @@ mpls = Y
nvgre = Y
pppoes = Y
raw = Y
+represented_port = Y
sctp = Y
tcp = Y
tx_queue = Y
diff --git a/doc/guides/nics/features/cnxk_vec.ini b/doc/guides/nics/features/cnxk_vec.ini
index 6086b3d73f..193fd14fbb 100644
--- a/doc/guides/nics/features/cnxk_vec.ini
+++ b/doc/guides/nics/features/cnxk_vec.ini
@@ -67,6 +67,7 @@ mpls = Y
nvgre = Y
pppoes = Y
raw = Y
+represented_port = Y
sctp = Y
tcp = Y
tx_queue = Y
@@ -86,6 +87,7 @@ of_set_vlan_pcp = Y
of_set_vlan_vid = Y
pf = Y
queue = Y
+represented_port = Y
rss = Y
security = Y
vf = Y
diff --git a/doc/guides/nics/features/cnxk_vf.ini b/doc/guides/nics/features/cnxk_vf.ini
index 53aa2a3d0c..7ae49b8107 100644
--- a/doc/guides/nics/features/cnxk_vf.ini
+++ b/doc/guides/nics/features/cnxk_vf.ini
@@ -64,6 +64,7 @@ mpls = Y
nvgre = Y
pppoes = Y
raw = Y
+represented_port = Y
sctp = Y
tcp = Y
tx_queue = Y
@@ -85,6 +86,7 @@ of_set_vlan_pcp = Y
of_set_vlan_vid = Y
pf = Y
queue = Y
+represented_port = Y
rss = Y
security = Y
skip_cman = Y
diff --git a/drivers/net/cnxk/cnxk_flow.c b/drivers/net/cnxk/cnxk_flow.c
index 5f74c356b1..a3b21f761f 100644
--- a/drivers/net/cnxk/cnxk_flow.c
+++ b/drivers/net/cnxk/cnxk_flow.c
@@ -4,67 +4,48 @@
#include <cnxk_flow.h>
const struct cnxk_rte_flow_term_info term[] = {
- [RTE_FLOW_ITEM_TYPE_ETH] = {ROC_NPC_ITEM_TYPE_ETH,
- sizeof(struct rte_flow_item_eth)},
- [RTE_FLOW_ITEM_TYPE_VLAN] = {ROC_NPC_ITEM_TYPE_VLAN,
- sizeof(struct rte_flow_item_vlan)},
- [RTE_FLOW_ITEM_TYPE_E_TAG] = {ROC_NPC_ITEM_TYPE_E_TAG,
- sizeof(struct rte_flow_item_e_tag)},
- [RTE_FLOW_ITEM_TYPE_IPV4] = {ROC_NPC_ITEM_TYPE_IPV4,
- sizeof(struct rte_flow_item_ipv4)},
- [RTE_FLOW_ITEM_TYPE_IPV6] = {ROC_NPC_ITEM_TYPE_IPV6,
- sizeof(struct rte_flow_item_ipv6)},
- [RTE_FLOW_ITEM_TYPE_IPV6_FRAG_EXT] = {
- ROC_NPC_ITEM_TYPE_IPV6_FRAG_EXT,
- sizeof(struct rte_flow_item_ipv6_frag_ext)},
- [RTE_FLOW_ITEM_TYPE_ARP_ETH_IPV4] = {
- ROC_NPC_ITEM_TYPE_ARP_ETH_IPV4,
- sizeof(struct rte_flow_item_arp_eth_ipv4)},
- [RTE_FLOW_ITEM_TYPE_MPLS] = {ROC_NPC_ITEM_TYPE_MPLS,
- sizeof(struct rte_flow_item_mpls)},
- [RTE_FLOW_ITEM_TYPE_ICMP] = {ROC_NPC_ITEM_TYPE_ICMP,
- sizeof(struct rte_flow_item_icmp)},
- [RTE_FLOW_ITEM_TYPE_UDP] = {ROC_NPC_ITEM_TYPE_UDP,
- sizeof(struct rte_flow_item_udp)},
- [RTE_FLOW_ITEM_TYPE_TCP] = {ROC_NPC_ITEM_TYPE_TCP,
- sizeof(struct rte_flow_item_tcp)},
- [RTE_FLOW_ITEM_TYPE_SCTP] = {ROC_NPC_ITEM_TYPE_SCTP,
- sizeof(struct rte_flow_item_sctp)},
- [RTE_FLOW_ITEM_TYPE_ESP] = {ROC_NPC_ITEM_TYPE_ESP,
- sizeof(struct rte_flow_item_esp)},
- [RTE_FLOW_ITEM_TYPE_GRE] = {ROC_NPC_ITEM_TYPE_GRE,
- sizeof(struct rte_flow_item_gre)},
- [RTE_FLOW_ITEM_TYPE_NVGRE] = {ROC_NPC_ITEM_TYPE_NVGRE,
- sizeof(struct rte_flow_item_nvgre)},
- [RTE_FLOW_ITEM_TYPE_VXLAN] = {ROC_NPC_ITEM_TYPE_VXLAN,
- sizeof(struct rte_flow_item_vxlan)},
- [RTE_FLOW_ITEM_TYPE_GTPC] = {ROC_NPC_ITEM_TYPE_GTPC,
- sizeof(struct rte_flow_item_gtp)},
- [RTE_FLOW_ITEM_TYPE_GTPU] = {ROC_NPC_ITEM_TYPE_GTPU,
- sizeof(struct rte_flow_item_gtp)},
+ [RTE_FLOW_ITEM_TYPE_ETH] = {ROC_NPC_ITEM_TYPE_ETH, sizeof(struct rte_flow_item_eth)},
+ [RTE_FLOW_ITEM_TYPE_VLAN] = {ROC_NPC_ITEM_TYPE_VLAN, sizeof(struct rte_flow_item_vlan)},
+ [RTE_FLOW_ITEM_TYPE_E_TAG] = {ROC_NPC_ITEM_TYPE_E_TAG, sizeof(struct rte_flow_item_e_tag)},
+ [RTE_FLOW_ITEM_TYPE_IPV4] = {ROC_NPC_ITEM_TYPE_IPV4, sizeof(struct rte_flow_item_ipv4)},
+ [RTE_FLOW_ITEM_TYPE_IPV6] = {ROC_NPC_ITEM_TYPE_IPV6, sizeof(struct rte_flow_item_ipv6)},
+ [RTE_FLOW_ITEM_TYPE_IPV6_FRAG_EXT] = {ROC_NPC_ITEM_TYPE_IPV6_FRAG_EXT,
+ sizeof(struct rte_flow_item_ipv6_frag_ext)},
+ [RTE_FLOW_ITEM_TYPE_ARP_ETH_IPV4] = {ROC_NPC_ITEM_TYPE_ARP_ETH_IPV4,
+ sizeof(struct rte_flow_item_arp_eth_ipv4)},
+ [RTE_FLOW_ITEM_TYPE_MPLS] = {ROC_NPC_ITEM_TYPE_MPLS, sizeof(struct rte_flow_item_mpls)},
+ [RTE_FLOW_ITEM_TYPE_ICMP] = {ROC_NPC_ITEM_TYPE_ICMP, sizeof(struct rte_flow_item_icmp)},
+ [RTE_FLOW_ITEM_TYPE_UDP] = {ROC_NPC_ITEM_TYPE_UDP, sizeof(struct rte_flow_item_udp)},
+ [RTE_FLOW_ITEM_TYPE_TCP] = {ROC_NPC_ITEM_TYPE_TCP, sizeof(struct rte_flow_item_tcp)},
+ [RTE_FLOW_ITEM_TYPE_SCTP] = {ROC_NPC_ITEM_TYPE_SCTP, sizeof(struct rte_flow_item_sctp)},
+ [RTE_FLOW_ITEM_TYPE_ESP] = {ROC_NPC_ITEM_TYPE_ESP, sizeof(struct rte_flow_item_esp)},
+ [RTE_FLOW_ITEM_TYPE_GRE] = {ROC_NPC_ITEM_TYPE_GRE, sizeof(struct rte_flow_item_gre)},
+ [RTE_FLOW_ITEM_TYPE_NVGRE] = {ROC_NPC_ITEM_TYPE_NVGRE, sizeof(struct rte_flow_item_nvgre)},
+ [RTE_FLOW_ITEM_TYPE_VXLAN] = {ROC_NPC_ITEM_TYPE_VXLAN, sizeof(struct rte_flow_item_vxlan)},
+ [RTE_FLOW_ITEM_TYPE_GTPC] = {ROC_NPC_ITEM_TYPE_GTPC, sizeof(struct rte_flow_item_gtp)},
+ [RTE_FLOW_ITEM_TYPE_GTPU] = {ROC_NPC_ITEM_TYPE_GTPU, sizeof(struct rte_flow_item_gtp)},
[RTE_FLOW_ITEM_TYPE_GENEVE] = {ROC_NPC_ITEM_TYPE_GENEVE,
sizeof(struct rte_flow_item_geneve)},
- [RTE_FLOW_ITEM_TYPE_VXLAN_GPE] = {
- ROC_NPC_ITEM_TYPE_VXLAN_GPE,
- sizeof(struct rte_flow_item_vxlan_gpe)},
+ [RTE_FLOW_ITEM_TYPE_VXLAN_GPE] = {ROC_NPC_ITEM_TYPE_VXLAN_GPE,
+ sizeof(struct rte_flow_item_vxlan_gpe)},
[RTE_FLOW_ITEM_TYPE_IPV6_EXT] = {ROC_NPC_ITEM_TYPE_IPV6_EXT,
sizeof(struct rte_flow_item_ipv6_ext)},
[RTE_FLOW_ITEM_TYPE_VOID] = {ROC_NPC_ITEM_TYPE_VOID, 0},
[RTE_FLOW_ITEM_TYPE_ANY] = {ROC_NPC_ITEM_TYPE_ANY, 0},
- [RTE_FLOW_ITEM_TYPE_GRE_KEY] = {ROC_NPC_ITEM_TYPE_GRE_KEY,
- sizeof(uint32_t)},
+ [RTE_FLOW_ITEM_TYPE_GRE_KEY] = {ROC_NPC_ITEM_TYPE_GRE_KEY, sizeof(uint32_t)},
[RTE_FLOW_ITEM_TYPE_HIGIG2] = {ROC_NPC_ITEM_TYPE_HIGIG2,
sizeof(struct rte_flow_item_higig2_hdr)},
- [RTE_FLOW_ITEM_TYPE_RAW] = {ROC_NPC_ITEM_TYPE_RAW,
- sizeof(struct rte_flow_item_raw)},
- [RTE_FLOW_ITEM_TYPE_MARK] = {ROC_NPC_ITEM_TYPE_MARK,
- sizeof(struct rte_flow_item_mark)},
+ [RTE_FLOW_ITEM_TYPE_RAW] = {ROC_NPC_ITEM_TYPE_RAW, sizeof(struct rte_flow_item_raw)},
+ [RTE_FLOW_ITEM_TYPE_MARK] = {ROC_NPC_ITEM_TYPE_MARK, sizeof(struct rte_flow_item_mark)},
[RTE_FLOW_ITEM_TYPE_IPV6_ROUTING_EXT] = {ROC_NPC_ITEM_TYPE_IPV6_ROUTING_EXT,
- sizeof(struct rte_flow_item_ipv6_routing_ext)},
+ sizeof(struct rte_flow_item_ipv6_routing_ext)},
[RTE_FLOW_ITEM_TYPE_TX_QUEUE] = {ROC_NPC_ITEM_TYPE_TX_QUEUE,
- sizeof(struct rte_flow_item_tx_queue)},
+ sizeof(struct rte_flow_item_tx_queue)},
+ [RTE_FLOW_ITEM_TYPE_REPRESENTED_PORT] = {ROC_NPC_ITEM_TYPE_REPRESENTED_PORT,
+ sizeof(struct rte_flow_item_ethdev)},
[RTE_FLOW_ITEM_TYPE_PPPOES] = {ROC_NPC_ITEM_TYPE_PPPOES,
- sizeof(struct rte_flow_item_pppoe)}};
+ sizeof(struct rte_flow_item_pppoe)}
+};
static int
npc_rss_action_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
@@ -372,6 +353,11 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
uint16_t *dst_pf_func)
{
+ struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
+ const struct rte_flow_item_ethdev *rep_eth_dev;
+ struct rte_eth_dev *portid_eth_dev;
+ char if_name[RTE_ETH_NAME_MAX_LEN];
+ struct cnxk_eth_dev *hw_dst;
int i = 0;
in_attr->priority = attr->priority;
@@ -384,6 +370,27 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
in_pattern[i].mask = pattern->mask;
in_pattern[i].type = term[pattern->type].item_type;
in_pattern[i].size = term[pattern->type].item_size;
+ if (pattern->type == RTE_FLOW_ITEM_TYPE_REPRESENTED_PORT) {
+ rep_eth_dev = (const struct rte_flow_item_ethdev *)pattern->spec;
+ if (rte_eth_dev_get_name_by_port(rep_eth_dev->port_id, if_name)) {
+ plt_err("Name not found for output port id");
+ return -EINVAL;
+ }
+ portid_eth_dev = rte_eth_dev_allocated(if_name);
+ if (!portid_eth_dev) {
+ plt_err("eth_dev not found for output port id");
+ return -EINVAL;
+ }
+ if (strcmp(portid_eth_dev->device->driver->name,
+ eth_dev->device->driver->name) != 0) {
+ plt_err("Output port not under same driver");
+ return -EINVAL;
+ }
+ hw_dst = portid_eth_dev->data->dev_private;
+ dev->npc.rep_npc = &hw_dst->npc;
+ dev->npc.rep_port_id = rep_eth_dev->port_id;
+ dev->npc.rep_pf_func = hw_dst->npc.pf_func;
+ }
pattern++;
i++;
}
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 20/23] net/cnxk: add representor port pattern and action
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (18 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 19/23] net/cnxk: add represented port pattern and action Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 21/23] net/cnxk: generalise flow operation APIs Harman Kalra
` (2 subsequent siblings)
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Adding support for representor port as item matching and action.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
doc/guides/nics/cnxk.rst | 1 +
doc/guides/nics/features/cnxk.ini | 4 +
doc/guides/nics/features/cnxk_vec.ini | 4 +
doc/guides/nics/features/cnxk_vf.ini | 4 +
drivers/net/cnxk/cnxk_flow.c | 219 +++++++++++++++++++++++---
drivers/net/cnxk/cnxk_rep.h | 14 ++
6 files changed, 220 insertions(+), 26 deletions(-)
diff --git a/doc/guides/nics/cnxk.rst b/doc/guides/nics/cnxk.rst
index 0f80d65562..f0d5418581 100644
--- a/doc/guides/nics/cnxk.rst
+++ b/doc/guides/nics/cnxk.rst
@@ -39,6 +39,7 @@ Features of the CNXK Ethdev PMD are:
- Queue based priority flow control support
- Port representors
- Represented port pattern matching and action
+- Port representor pattern matching and action
Prerequisites
-------------
diff --git a/doc/guides/nics/features/cnxk.ini b/doc/guides/nics/features/cnxk.ini
index 8957eba46b..b5d9f7e579 100644
--- a/doc/guides/nics/features/cnxk.ini
+++ b/doc/guides/nics/features/cnxk.ini
@@ -71,6 +71,7 @@ ipv6_routing_ext = Y
mark = Y
mpls = Y
nvgre = Y
+port_representor = Y
pppoes = Y
raw = Y
represented_port = Y
@@ -95,6 +96,7 @@ of_set_vlan_pcp = Y
of_set_vlan_vid = Y
pf = Y
port_id = Y
+port_representor = Y
queue = Y
represented_port = Y
rss = Y
@@ -102,3 +104,5 @@ sample = Y
security = Y
skip_cman = Y
vf = Y
+vxlan_decap = I
+vxlan_encap = I
diff --git a/doc/guides/nics/features/cnxk_vec.ini b/doc/guides/nics/features/cnxk_vec.ini
index 193fd14fbb..92a486664f 100644
--- a/doc/guides/nics/features/cnxk_vec.ini
+++ b/doc/guides/nics/features/cnxk_vec.ini
@@ -65,6 +65,7 @@ ipv6_frag_ext = Y
mark = Y
mpls = Y
nvgre = Y
+port_representor = Y
pppoes = Y
raw = Y
represented_port = Y
@@ -86,8 +87,11 @@ of_push_vlan = Y
of_set_vlan_pcp = Y
of_set_vlan_vid = Y
pf = Y
+port_representor = Y
queue = Y
represented_port = Y
rss = Y
security = Y
vf = Y
+vxlan_decap = I
+vxlan_encap = I
diff --git a/doc/guides/nics/features/cnxk_vf.ini b/doc/guides/nics/features/cnxk_vf.ini
index 7ae49b8107..a55f0e7ce5 100644
--- a/doc/guides/nics/features/cnxk_vf.ini
+++ b/doc/guides/nics/features/cnxk_vf.ini
@@ -62,6 +62,7 @@ ipv6_routing_ext = Y
mark = Y
mpls = Y
nvgre = Y
+port_representor = Y
pppoes = Y
raw = Y
represented_port = Y
@@ -85,9 +86,12 @@ of_push_vlan = Y
of_set_vlan_pcp = Y
of_set_vlan_vid = Y
pf = Y
+port_representor = Y
queue = Y
represented_port = Y
rss = Y
security = Y
skip_cman = Y
vf = Y
+vxlan_decap = I
+vxlan_encap = I
diff --git a/drivers/net/cnxk/cnxk_flow.c b/drivers/net/cnxk/cnxk_flow.c
index a3b21f761f..261915d84e 100644
--- a/drivers/net/cnxk/cnxk_flow.c
+++ b/drivers/net/cnxk/cnxk_flow.c
@@ -2,6 +2,7 @@
* Copyright(C) 2021 Marvell.
*/
#include <cnxk_flow.h>
+#include <cnxk_rep.h>
const struct cnxk_rte_flow_term_info term[] = {
[RTE_FLOW_ITEM_TYPE_ETH] = {ROC_NPC_ITEM_TYPE_ETH, sizeof(struct rte_flow_item_eth)},
@@ -185,11 +186,44 @@ roc_npc_parse_sample_subaction(struct rte_eth_dev *eth_dev, const struct rte_flo
return 0;
}
+static int
+representor_portid_action(struct roc_npc_action *in_actions, struct rte_eth_dev *portid_eth_dev,
+ uint16_t *dst_pf_func, uint8_t has_tunnel_pattern, int *act_cnt)
+{
+ struct rte_eth_dev *rep_eth_dev = portid_eth_dev;
+ struct rte_flow_action_mark *act_mark;
+ struct cnxk_rep_dev *rep_dev;
+ /* For inserting an action in the list */
+ int i = *act_cnt;
+
+ rep_dev = cnxk_rep_pmd_priv(rep_eth_dev);
+ *dst_pf_func = rep_dev->hw_func;
+
+ /* Add Mark action */
+ i++;
+ act_mark = plt_zmalloc(sizeof(struct rte_flow_action_mark), 0);
+ if (!act_mark) {
+ plt_err("Error allocation memory");
+ return -ENOMEM;
+ }
+
+ /* Mark ID format: (tunnel type - VxLAN, Geneve << 6) | Tunnel decap */
+ act_mark->id = has_tunnel_pattern ? ((has_tunnel_pattern << 6) | 5) : 1;
+ in_actions[i].type = ROC_NPC_ACTION_TYPE_MARK;
+ in_actions[i].conf = (struct rte_flow_action_mark *)act_mark;
+
+ *act_cnt = i;
+ plt_rep_dbg("Rep port %d ID %d mark ID is %d rep_dev->hw_func 0x%x", rep_dev->port_id,
+ rep_dev->rep_id, act_mark->id, rep_dev->hw_func);
+
+ return 0;
+}
+
static int
cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
const struct rte_flow_action actions[], struct roc_npc_action in_actions[],
struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
- uint16_t *dst_pf_func)
+ uint16_t *dst_pf_func, uint8_t has_tunnel_pattern)
{
struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
const struct rte_flow_action_queue *act_q = NULL;
@@ -238,6 +272,7 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
break;
case RTE_FLOW_ACTION_TYPE_REPRESENTED_PORT:
+ case RTE_FLOW_ACTION_TYPE_PORT_REPRESENTOR:
case RTE_FLOW_ACTION_TYPE_PORT_ID:
in_actions[i].type = ROC_NPC_ACTION_TYPE_PORT_ID;
in_actions[i].conf = actions->conf;
@@ -256,14 +291,27 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
plt_err("eth_dev not found for output port id");
goto err_exit;
}
- if (strcmp(portid_eth_dev->device->driver->name,
- eth_dev->device->driver->name) != 0) {
- plt_err("Output port not under same driver");
- goto err_exit;
+
+ if (cnxk_ethdev_is_representor(if_name)) {
+ plt_rep_dbg("Representor port %d act port %d", port_act->id,
+ act_ethdev->port_id);
+ if (representor_portid_action(in_actions, portid_eth_dev,
+ dst_pf_func, has_tunnel_pattern,
+ &i)) {
+ plt_err("Representor port action set failed");
+ goto err_exit;
+ }
+ } else {
+ if (strcmp(portid_eth_dev->device->driver->name,
+ eth_dev->device->driver->name) != 0) {
+ plt_err("Output port not under same driver");
+ goto err_exit;
+ }
+
+ hw_dst = portid_eth_dev->data->dev_private;
+ roc_npc_dst = &hw_dst->npc;
+ *dst_pf_func = roc_npc_dst->pf_func;
}
- hw_dst = portid_eth_dev->data->dev_private;
- roc_npc_dst = &hw_dst->npc;
- *dst_pf_func = roc_npc_dst->pf_func;
break;
case RTE_FLOW_ACTION_TYPE_QUEUE:
@@ -324,6 +372,8 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
in_actions[i].type = ROC_NPC_ACTION_TYPE_SAMPLE;
in_actions[i].conf = in_sample_actions;
break;
+ case RTE_FLOW_ACTION_TYPE_VXLAN_DECAP:
+ continue;
default:
plt_npc_dbg("Action is not supported = %d", actions->type);
goto err_exit;
@@ -346,12 +396,8 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
}
static int
-cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
- const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
- struct roc_npc_attr *in_attr, struct roc_npc_item_info in_pattern[],
- struct roc_npc_action in_actions[],
- struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
- uint16_t *dst_pf_func)
+cnxk_map_pattern(struct rte_eth_dev *eth_dev, const struct rte_flow_item pattern[],
+ struct roc_npc_item_info in_pattern[], uint8_t *has_tunnel_pattern)
{
struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
const struct rte_flow_item_ethdev *rep_eth_dev;
@@ -360,10 +406,6 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
struct cnxk_eth_dev *hw_dst;
int i = 0;
- in_attr->priority = attr->priority;
- in_attr->ingress = attr->ingress;
- in_attr->egress = attr->egress;
-
while (pattern->type != RTE_FLOW_ITEM_TYPE_END) {
in_pattern[i].spec = pattern->spec;
in_pattern[i].last = pattern->last;
@@ -374,30 +416,81 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
rep_eth_dev = (const struct rte_flow_item_ethdev *)pattern->spec;
if (rte_eth_dev_get_name_by_port(rep_eth_dev->port_id, if_name)) {
plt_err("Name not found for output port id");
- return -EINVAL;
+ goto fail;
}
portid_eth_dev = rte_eth_dev_allocated(if_name);
if (!portid_eth_dev) {
plt_err("eth_dev not found for output port id");
- return -EINVAL;
+ goto fail;
}
if (strcmp(portid_eth_dev->device->driver->name,
eth_dev->device->driver->name) != 0) {
plt_err("Output port not under same driver");
- return -EINVAL;
+ goto fail;
+ }
+ if (cnxk_ethdev_is_representor(if_name)) {
+ /* Case where represented port not part of same
+ * app and represented by a representor port.
+ */
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eswitch_dev *eswitch_dev;
+
+ rep_dev = cnxk_rep_pmd_priv(portid_eth_dev);
+ eswitch_dev = rep_dev->parent_dev;
+ dev->npc.rep_npc = &eswitch_dev->npc;
+ dev->npc.rep_port_id = rep_eth_dev->port_id;
+ dev->npc.rep_pf_func = rep_dev->hw_func;
+ plt_rep_dbg("Represented port %d act port %d rep_dev->hw_func 0x%x",
+ rep_eth_dev->port_id, eth_dev->data->port_id,
+ rep_dev->hw_func);
+ } else {
+ /* Case where represented port part of same app
+ * as PF.
+ */
+ hw_dst = portid_eth_dev->data->dev_private;
+ dev->npc.rep_npc = &hw_dst->npc;
+ dev->npc.rep_port_id = rep_eth_dev->port_id;
+ dev->npc.rep_pf_func = hw_dst->npc.pf_func;
}
- hw_dst = portid_eth_dev->data->dev_private;
- dev->npc.rep_npc = &hw_dst->npc;
- dev->npc.rep_port_id = rep_eth_dev->port_id;
- dev->npc.rep_pf_func = hw_dst->npc.pf_func;
}
+
+ if (pattern->type == RTE_FLOW_ITEM_TYPE_VXLAN ||
+ pattern->type == RTE_FLOW_ITEM_TYPE_VXLAN_GPE ||
+ pattern->type == RTE_FLOW_ITEM_TYPE_GRE)
+ *has_tunnel_pattern = pattern->type;
+
pattern++;
i++;
}
in_pattern[i].type = ROC_NPC_ITEM_TYPE_END;
+ return 0;
+fail:
+ return -EINVAL;
+}
+
+static int
+cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct roc_npc_attr *in_attr, struct roc_npc_item_info in_pattern[],
+ struct roc_npc_action in_actions[],
+ struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
+ uint16_t *dst_pf_func)
+{
+ uint8_t has_tunnel_pattern = 0;
+ int rc;
+
+ in_attr->priority = attr->priority;
+ in_attr->ingress = attr->ingress;
+ in_attr->egress = attr->egress;
+
+ rc = cnxk_map_pattern(eth_dev, pattern, in_pattern, &has_tunnel_pattern);
+ if (rc) {
+ plt_err("Failed to map pattern list");
+ return rc;
+ }
return cnxk_map_actions(eth_dev, attr, actions, in_actions, in_sample_actions, flowkey_cfg,
- dst_pf_func);
+ dst_pf_func, has_tunnel_pattern);
}
static int
@@ -461,6 +554,7 @@ cnxk_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
int rc;
memset(&in_sample_action, 0, sizeof(in_sample_action));
+ memset(&in_attr, 0, sizeof(struct roc_npc_attr));
rc = cnxk_map_flow_data(eth_dev, attr, pattern, actions, &in_attr, in_pattern, in_actions,
&in_sample_action, &npc->flowkey_cfg_state, &dst_pf_func);
if (rc) {
@@ -646,6 +740,75 @@ cnxk_flow_get_aged_flows(struct rte_eth_dev *eth_dev, void **context,
return cnt;
}
+static int
+cnxk_flow_tunnel_decap_set(__rte_unused struct rte_eth_dev *dev, struct rte_flow_tunnel *tunnel,
+ struct rte_flow_action **pmd_actions, uint32_t *num_of_actions,
+ __rte_unused struct rte_flow_error *err)
+{
+ struct rte_flow_action *nfp_action;
+
+ nfp_action = rte_zmalloc("nfp_tun_action", sizeof(struct rte_flow_action), 0);
+ if (nfp_action == NULL) {
+ plt_err("Alloc memory for nfp tunnel action failed.");
+ return -ENOMEM;
+ }
+
+ if (tunnel->is_ipv6)
+ nfp_action->conf = (void *)~0;
+
+ switch (tunnel->type) {
+ case RTE_FLOW_ITEM_TYPE_VXLAN:
+ nfp_action->type = RTE_FLOW_ACTION_TYPE_VXLAN_DECAP;
+ *pmd_actions = nfp_action;
+ *num_of_actions = 1;
+ break;
+ default:
+ *pmd_actions = NULL;
+ *num_of_actions = 0;
+ rte_free(nfp_action);
+ break;
+ }
+
+ return 0;
+}
+
+static int
+cnxk_flow_tunnel_action_decap_release(__rte_unused struct rte_eth_dev *dev,
+ struct rte_flow_action *pmd_actions, uint32_t num_of_actions,
+ __rte_unused struct rte_flow_error *err)
+{
+ uint32_t i;
+ struct rte_flow_action *nfp_action;
+
+ for (i = 0; i < num_of_actions; i++) {
+ nfp_action = &pmd_actions[i];
+ nfp_action->conf = NULL;
+ rte_free(nfp_action);
+ }
+
+ return 0;
+}
+
+static int
+cnxk_flow_tunnel_match(__rte_unused struct rte_eth_dev *dev,
+ __rte_unused struct rte_flow_tunnel *tunnel,
+ __rte_unused struct rte_flow_item **pmd_items, uint32_t *num_of_items,
+ __rte_unused struct rte_flow_error *err)
+{
+ *num_of_items = 0;
+
+ return 0;
+}
+
+static int
+cnxk_flow_tunnel_item_release(__rte_unused struct rte_eth_dev *dev,
+ __rte_unused struct rte_flow_item *pmd_items,
+ __rte_unused uint32_t num_of_items,
+ __rte_unused struct rte_flow_error *err)
+{
+ return 0;
+}
+
struct rte_flow_ops cnxk_flow_ops = {
.validate = cnxk_flow_validate,
.flush = cnxk_flow_flush,
@@ -653,4 +816,8 @@ struct rte_flow_ops cnxk_flow_ops = {
.isolate = cnxk_flow_isolate,
.dev_dump = cnxk_flow_dev_dump,
.get_aged_flows = cnxk_flow_get_aged_flows,
+ .tunnel_match = cnxk_flow_tunnel_match,
+ .tunnel_item_release = cnxk_flow_tunnel_item_release,
+ .tunnel_decap_set = cnxk_flow_tunnel_decap_set,
+ .tunnel_action_decap_release = cnxk_flow_tunnel_action_decap_release,
};
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index b57b56cd12..b0cf3340a2 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -1,6 +1,9 @@
/* SPDX-License-Identifier: BSD-3-Clause
* Copyright(C) 2024 Marvell.
*/
+
+#include <regex.h>
+
#include <cnxk_eswitch.h>
#include <cnxk_ethdev.h>
@@ -90,6 +93,17 @@ cnxk_rep_pool_buffer_stats(struct rte_mempool *pool)
pool->size, rte_mempool_in_use_count(pool), rte_mempool_avail_count(pool));
}
+static inline int
+cnxk_ethdev_is_representor(const char *if_name)
+{
+ regex_t regex;
+ int val;
+
+ val = regcomp(®ex, "net_.*_representor_.*", 0);
+ val = regexec(®ex, if_name, 0, NULL, 0);
+ return (val == 0);
+}
+
/* Prototypes */
int cnxk_rep_dev_probe(struct rte_pci_device *pci_dev, struct cnxk_eswitch_dev *eswitch_dev);
int cnxk_rep_dev_remove(struct cnxk_eswitch_dev *eswitch_dev);
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 21/23] net/cnxk: generalise flow operation APIs
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (19 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 20/23] net/cnxk: add representor " Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 22/23] net/cnxk: flow create on representor ports Harman Kalra
2024-02-01 13:07 ` [PATCH v3 23/23] net/cnxk: other flow operations Harman Kalra
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Flow operations can be performed on cnxk ports as well as representor
ports. Since representor ports are not cnxk ports but have eswitch as
base device underneath, special handling is required to align with base
infra. Introducing a flag to generic flow APIs to discriminate if the
operation request made on normal or representor ports.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_flow.c | 270 ++++++++++++++++++++++++++---------
drivers/net/cnxk/cnxk_flow.h | 18 +++
2 files changed, 218 insertions(+), 70 deletions(-)
diff --git a/drivers/net/cnxk/cnxk_flow.c b/drivers/net/cnxk/cnxk_flow.c
index 261915d84e..ff89004cc0 100644
--- a/drivers/net/cnxk/cnxk_flow.c
+++ b/drivers/net/cnxk/cnxk_flow.c
@@ -223,7 +223,7 @@ static int
cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
const struct rte_flow_action actions[], struct roc_npc_action in_actions[],
struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
- uint16_t *dst_pf_func, uint8_t has_tunnel_pattern)
+ uint16_t *dst_pf_func, uint8_t has_tunnel_pattern, bool is_rep)
{
struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
const struct rte_flow_action_queue *act_q = NULL;
@@ -274,15 +274,18 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
case RTE_FLOW_ACTION_TYPE_REPRESENTED_PORT:
case RTE_FLOW_ACTION_TYPE_PORT_REPRESENTOR:
case RTE_FLOW_ACTION_TYPE_PORT_ID:
+ /* No port ID action on representor ethdevs */
+ if (is_rep)
+ continue;
in_actions[i].type = ROC_NPC_ACTION_TYPE_PORT_ID;
in_actions[i].conf = actions->conf;
- act_ethdev = (const struct rte_flow_action_ethdev *)
- actions->conf;
- port_act = (const struct rte_flow_action_port_id *)
- actions->conf;
+ act_ethdev = (const struct rte_flow_action_ethdev *)actions->conf;
+ port_act = (const struct rte_flow_action_port_id *)actions->conf;
if (rte_eth_dev_get_name_by_port(
- actions->type != RTE_FLOW_ACTION_TYPE_PORT_ID ?
- act_ethdev->port_id : port_act->id, if_name)) {
+ actions->type != RTE_FLOW_ACTION_TYPE_PORT_ID ?
+ act_ethdev->port_id :
+ port_act->id,
+ if_name)) {
plt_err("Name not found for output port id");
goto err_exit;
}
@@ -321,6 +324,9 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
break;
case RTE_FLOW_ACTION_TYPE_RSS:
+ /* No RSS action on representor ethdevs */
+ if (is_rep)
+ continue;
rc = npc_rss_action_validate(eth_dev, attr, actions);
if (rc)
goto err_exit;
@@ -397,22 +403,37 @@ cnxk_map_actions(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
static int
cnxk_map_pattern(struct rte_eth_dev *eth_dev, const struct rte_flow_item pattern[],
- struct roc_npc_item_info in_pattern[], uint8_t *has_tunnel_pattern)
+ struct roc_npc_item_info in_pattern[], uint8_t *has_tunnel_pattern, bool is_rep)
{
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
const struct rte_flow_item_ethdev *rep_eth_dev;
struct rte_eth_dev *portid_eth_dev;
char if_name[RTE_ETH_NAME_MAX_LEN];
struct cnxk_eth_dev *hw_dst;
+ struct cnxk_rep_dev *rdev;
+ struct cnxk_eth_dev *dev;
+ struct roc_npc *npc;
int i = 0;
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rdev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rdev->parent_dev->npc;
+
+ npc->rep_npc = npc;
+ npc->rep_port_id = rdev->port_id;
+ npc->rep_pf_func = rdev->hw_func;
+ }
+
while (pattern->type != RTE_FLOW_ITEM_TYPE_END) {
in_pattern[i].spec = pattern->spec;
in_pattern[i].last = pattern->last;
in_pattern[i].mask = pattern->mask;
in_pattern[i].type = term[pattern->type].item_type;
in_pattern[i].size = term[pattern->type].item_size;
- if (pattern->type == RTE_FLOW_ITEM_TYPE_REPRESENTED_PORT) {
+ if (pattern->type == RTE_FLOW_ITEM_TYPE_REPRESENTED_PORT ||
+ pattern->type == RTE_FLOW_ITEM_TYPE_PORT_REPRESENTOR) {
rep_eth_dev = (const struct rte_flow_item_ethdev *)pattern->spec;
if (rte_eth_dev_get_name_by_port(rep_eth_dev->port_id, if_name)) {
plt_err("Name not found for output port id");
@@ -423,11 +444,6 @@ cnxk_map_pattern(struct rte_eth_dev *eth_dev, const struct rte_flow_item pattern
plt_err("eth_dev not found for output port id");
goto fail;
}
- if (strcmp(portid_eth_dev->device->driver->name,
- eth_dev->device->driver->name) != 0) {
- plt_err("Output port not under same driver");
- goto fail;
- }
if (cnxk_ethdev_is_representor(if_name)) {
/* Case where represented port not part of same
* app and represented by a representor port.
@@ -437,20 +453,25 @@ cnxk_map_pattern(struct rte_eth_dev *eth_dev, const struct rte_flow_item pattern
rep_dev = cnxk_rep_pmd_priv(portid_eth_dev);
eswitch_dev = rep_dev->parent_dev;
- dev->npc.rep_npc = &eswitch_dev->npc;
- dev->npc.rep_port_id = rep_eth_dev->port_id;
- dev->npc.rep_pf_func = rep_dev->hw_func;
+ npc->rep_npc = &eswitch_dev->npc;
+ npc->rep_port_id = rep_eth_dev->port_id;
+ npc->rep_pf_func = rep_dev->hw_func;
plt_rep_dbg("Represented port %d act port %d rep_dev->hw_func 0x%x",
rep_eth_dev->port_id, eth_dev->data->port_id,
rep_dev->hw_func);
} else {
+ if (strcmp(portid_eth_dev->device->driver->name,
+ eth_dev->device->driver->name) != 0) {
+ plt_err("Output port not under same driver");
+ goto fail;
+ }
/* Case where represented port part of same app
* as PF.
*/
hw_dst = portid_eth_dev->data->dev_private;
- dev->npc.rep_npc = &hw_dst->npc;
- dev->npc.rep_port_id = rep_eth_dev->port_id;
- dev->npc.rep_pf_func = hw_dst->npc.pf_func;
+ npc->rep_npc = &hw_dst->npc;
+ npc->rep_port_id = rep_eth_dev->port_id;
+ npc->rep_pf_func = hw_dst->npc.pf_func;
}
}
@@ -474,7 +495,7 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
struct roc_npc_attr *in_attr, struct roc_npc_item_info in_pattern[],
struct roc_npc_action in_actions[],
struct roc_npc_action_sample *in_sample_actions, uint32_t *flowkey_cfg,
- uint16_t *dst_pf_func)
+ uint16_t *dst_pf_func, bool is_rep)
{
uint8_t has_tunnel_pattern = 0;
int rc;
@@ -482,44 +503,61 @@ cnxk_map_flow_data(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
in_attr->priority = attr->priority;
in_attr->ingress = attr->ingress;
in_attr->egress = attr->egress;
+ if (attr->transfer) {
+ /* For representor ethdevs transfer attribute corresponds to egress rule */
+ if (is_rep)
+ in_attr->egress = attr->transfer;
+ else
+ in_attr->ingress = attr->transfer;
+ }
- rc = cnxk_map_pattern(eth_dev, pattern, in_pattern, &has_tunnel_pattern);
+ rc = cnxk_map_pattern(eth_dev, pattern, in_pattern, &has_tunnel_pattern, is_rep);
if (rc) {
plt_err("Failed to map pattern list");
return rc;
}
return cnxk_map_actions(eth_dev, attr, actions, in_actions, in_sample_actions, flowkey_cfg,
- dst_pf_func, has_tunnel_pattern);
+ dst_pf_func, has_tunnel_pattern, is_rep);
}
-static int
-cnxk_flow_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
- const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
- struct rte_flow_error *error)
+int
+cnxk_flow_validate_common(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[],
+ const struct rte_flow_action actions[], struct rte_flow_error *error,
+ bool is_rep)
{
struct roc_npc_item_info in_pattern[ROC_NPC_ITEM_TYPE_END + 1];
struct roc_npc_action in_actions[ROC_NPC_MAX_ACTION_COUNT];
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
struct roc_npc_action_sample in_sample_action;
- struct roc_npc *npc = &dev->npc;
+ struct cnxk_rep_dev *rep_dev;
struct roc_npc_attr in_attr;
+ struct cnxk_eth_dev *dev;
struct roc_npc_flow flow;
uint32_t flowkey_cfg = 0;
uint16_t dst_pf_func = 0;
+ struct roc_npc *npc;
int rc;
- /* Skip flow validation for MACsec. */
- if (actions[0].type == RTE_FLOW_ACTION_TYPE_SECURITY &&
- cnxk_eth_macsec_sess_get_by_sess(dev, actions[0].conf) != NULL)
- return 0;
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ /* Skip flow validation for MACsec. */
+ if (actions[0].type == RTE_FLOW_ACTION_TYPE_SECURITY &&
+ cnxk_eth_macsec_sess_get_by_sess(dev, actions[0].conf) != NULL)
+ return 0;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
memset(&flow, 0, sizeof(flow));
memset(&in_sample_action, 0, sizeof(in_sample_action));
flow.is_validate = true;
rc = cnxk_map_flow_data(eth_dev, attr, pattern, actions, &in_attr, in_pattern, in_actions,
- &in_sample_action, &flowkey_cfg, &dst_pf_func);
+ &in_sample_action, &flowkey_cfg, &dst_pf_func, is_rep);
if (rc) {
rte_flow_error_set(error, 0, RTE_FLOW_ERROR_TYPE_ACTION_NUM, NULL,
"Failed to map flow data");
@@ -536,27 +574,45 @@ cnxk_flow_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr
return 0;
}
+static int
+cnxk_flow_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ return cnxk_flow_validate_common(eth_dev, attr, pattern, actions, error, false);
+}
+
struct roc_npc_flow *
-cnxk_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
- const struct rte_flow_item pattern[],
- const struct rte_flow_action actions[],
- struct rte_flow_error *error)
+cnxk_flow_create_common(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[],
+ const struct rte_flow_action actions[], struct rte_flow_error *error,
+ bool is_rep)
{
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
struct roc_npc_item_info in_pattern[ROC_NPC_ITEM_TYPE_END + 1];
struct roc_npc_action in_actions[ROC_NPC_MAX_ACTION_COUNT];
struct roc_npc_action_sample in_sample_action;
- struct roc_npc *npc = &dev->npc;
+ struct cnxk_rep_dev *rep_dev = NULL;
+ struct cnxk_eth_dev *dev = NULL;
struct roc_npc_attr in_attr;
struct roc_npc_flow *flow;
uint16_t dst_pf_func = 0;
+ struct roc_npc *npc;
int errcode = 0;
int rc;
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
+
memset(&in_sample_action, 0, sizeof(in_sample_action));
memset(&in_attr, 0, sizeof(struct roc_npc_attr));
rc = cnxk_map_flow_data(eth_dev, attr, pattern, actions, &in_attr, in_pattern, in_actions,
- &in_sample_action, &npc->flowkey_cfg_state, &dst_pf_func);
+ &in_sample_action, &npc->flowkey_cfg_state, &dst_pf_func, is_rep);
if (rc) {
rte_flow_error_set(error, 0, RTE_FLOW_ERROR_TYPE_ACTION_NUM, NULL,
"Failed to map flow data");
@@ -572,32 +628,67 @@ cnxk_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
return flow;
}
+struct roc_npc_flow *
+cnxk_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ return cnxk_flow_create_common(eth_dev, attr, pattern, actions, error, false);
+}
+
int
-cnxk_flow_destroy(struct rte_eth_dev *eth_dev, struct roc_npc_flow *flow,
- struct rte_flow_error *error)
+cnxk_flow_destroy_common(struct rte_eth_dev *eth_dev, struct roc_npc_flow *flow,
+ struct rte_flow_error *error, bool is_rep)
{
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
- struct roc_npc *npc = &dev->npc;
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eth_dev *dev;
+ struct roc_npc *npc;
int rc;
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
+
rc = roc_npc_flow_destroy(npc, flow);
if (rc)
- rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
- NULL, "Flow Destroy failed");
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Flow Destroy failed");
return rc;
}
-static int
-cnxk_flow_flush(struct rte_eth_dev *eth_dev, struct rte_flow_error *error)
+int
+cnxk_flow_destroy(struct rte_eth_dev *eth_dev, struct roc_npc_flow *flow,
+ struct rte_flow_error *error)
{
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
- struct roc_npc *npc = &dev->npc;
+ return cnxk_flow_destroy_common(eth_dev, flow, error, false);
+}
+
+int
+cnxk_flow_flush_common(struct rte_eth_dev *eth_dev, struct rte_flow_error *error, bool is_rep)
+{
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eth_dev *dev;
+ struct roc_npc *npc;
int rc;
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
+
rc = roc_npc_mcam_free_all_resources(npc);
if (rc) {
- rte_flow_error_set(error, EIO, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
- NULL, "Failed to flush filter");
+ rte_flow_error_set(error, EIO, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to flush filter");
return -rte_errno;
}
@@ -605,14 +696,21 @@ cnxk_flow_flush(struct rte_eth_dev *eth_dev, struct rte_flow_error *error)
}
static int
-cnxk_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
- const struct rte_flow_action *action, void *data,
- struct rte_flow_error *error)
+cnxk_flow_flush(struct rte_eth_dev *eth_dev, struct rte_flow_error *error)
+{
+ return cnxk_flow_flush_common(eth_dev, error, false);
+}
+
+int
+cnxk_flow_query_common(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data,
+ struct rte_flow_error *error, bool is_rep)
{
struct roc_npc_flow *in_flow = (struct roc_npc_flow *)flow;
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
- struct roc_npc *npc = &dev->npc;
struct rte_flow_query_count *query = data;
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eth_dev *dev;
+ struct roc_npc *npc;
const char *errmsg = NULL;
int errcode = ENOTSUP;
int rc;
@@ -627,6 +725,15 @@ cnxk_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
goto err_exit;
}
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
+
if (in_flow->use_pre_alloc)
rc = roc_npc_inl_mcam_read_counter(in_flow->ctr_id, &query->hits);
else
@@ -660,8 +767,15 @@ cnxk_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
}
static int
-cnxk_flow_isolate(struct rte_eth_dev *eth_dev __rte_unused,
- int enable __rte_unused, struct rte_flow_error *error)
+cnxk_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data, struct rte_flow_error *error)
+{
+ return cnxk_flow_query_common(eth_dev, flow, action, data, error, false);
+}
+
+static int
+cnxk_flow_isolate(struct rte_eth_dev *eth_dev __rte_unused, int enable __rte_unused,
+ struct rte_flow_error *error)
{
/* If we support, we need to un-install the default mcam
* entry for this port.
@@ -673,16 +787,25 @@ cnxk_flow_isolate(struct rte_eth_dev *eth_dev __rte_unused,
return -rte_errno;
}
-static int
-cnxk_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
- FILE *file, struct rte_flow_error *error)
+int
+cnxk_flow_dev_dump_common(struct rte_eth_dev *eth_dev, struct rte_flow *flow, FILE *file,
+ struct rte_flow_error *error, bool is_rep)
{
- struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
- struct roc_npc *npc = &dev->npc;
+ struct cnxk_rep_dev *rep_dev;
+ struct cnxk_eth_dev *dev;
+ struct roc_npc *npc;
+
+ /* is_rep set for operation performed via representor ports */
+ if (!is_rep) {
+ dev = cnxk_eth_pmd_priv(eth_dev);
+ npc = &dev->npc;
+ } else {
+ rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ npc = &rep_dev->parent_dev->npc;
+ }
if (file == NULL) {
- rte_flow_error_set(error, EINVAL,
- RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ rte_flow_error_set(error, EINVAL, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
"Invalid file");
return -rte_errno;
}
@@ -701,8 +824,15 @@ cnxk_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
}
static int
-cnxk_flow_get_aged_flows(struct rte_eth_dev *eth_dev, void **context,
- uint32_t nb_contexts, struct rte_flow_error *err)
+cnxk_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow, FILE *file,
+ struct rte_flow_error *error)
+{
+ return cnxk_flow_dev_dump_common(eth_dev, flow, file, error, false);
+}
+
+static int
+cnxk_flow_get_aged_flows(struct rte_eth_dev *eth_dev, void **context, uint32_t nb_contexts,
+ struct rte_flow_error *err)
{
struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev);
struct roc_npc *roc_npc = &dev->npc;
diff --git a/drivers/net/cnxk/cnxk_flow.h b/drivers/net/cnxk/cnxk_flow.h
index bb23629819..226694fbed 100644
--- a/drivers/net/cnxk/cnxk_flow.h
+++ b/drivers/net/cnxk/cnxk_flow.h
@@ -24,4 +24,22 @@ struct roc_npc_flow *cnxk_flow_create(struct rte_eth_dev *dev,
int cnxk_flow_destroy(struct rte_eth_dev *dev, struct roc_npc_flow *flow,
struct rte_flow_error *error);
+struct roc_npc_flow *cnxk_flow_create_common(struct rte_eth_dev *eth_dev,
+ const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[],
+ const struct rte_flow_action actions[],
+ struct rte_flow_error *error, bool is_rep);
+int cnxk_flow_validate_common(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[],
+ const struct rte_flow_action actions[], struct rte_flow_error *error,
+ bool is_rep);
+int cnxk_flow_destroy_common(struct rte_eth_dev *eth_dev, struct roc_npc_flow *flow,
+ struct rte_flow_error *error, bool is_rep);
+int cnxk_flow_flush_common(struct rte_eth_dev *eth_dev, struct rte_flow_error *error, bool is_rep);
+int cnxk_flow_query_common(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data,
+ struct rte_flow_error *error, bool is_rep);
+int cnxk_flow_dev_dump_common(struct rte_eth_dev *eth_dev, struct rte_flow *flow, FILE *file,
+ struct rte_flow_error *error, bool is_rep);
+
#endif /* __CNXK_RTE_FLOW_H__ */
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 22/23] net/cnxk: flow create on representor ports
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (20 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 21/23] net/cnxk: generalise flow operation APIs Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
2024-02-01 13:07 ` [PATCH v3 23/23] net/cnxk: other flow operations Harman Kalra
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Implementing base infra for handling flow operations performed on
representor ports, where these representor ports may be representing
native representees or part of companian apps. Also added support for
handling flow create operation
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_flow.h | 9 +-
drivers/net/cnxk/cnxk_rep.h | 3 +
drivers/net/cnxk/cnxk_rep_flow.c | 401 +++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 27 +++
drivers/net/cnxk/cnxk_rep_ops.c | 3 +-
drivers/net/cnxk/meson.build | 1 +
6 files changed, 441 insertions(+), 3 deletions(-)
create mode 100644 drivers/net/cnxk/cnxk_rep_flow.c
diff --git a/drivers/net/cnxk/cnxk_flow.h b/drivers/net/cnxk/cnxk_flow.h
index 226694fbed..e51d04b2c9 100644
--- a/drivers/net/cnxk/cnxk_flow.h
+++ b/drivers/net/cnxk/cnxk_flow.h
@@ -16,8 +16,13 @@ struct cnxk_rte_flow_term_info {
uint16_t item_size;
};
-struct roc_npc_flow *cnxk_flow_create(struct rte_eth_dev *dev,
- const struct rte_flow_attr *attr,
+struct cnxk_rte_flow_action_info {
+ uint16_t conf_size;
+};
+
+extern const struct cnxk_rte_flow_term_info term[];
+
+struct roc_npc_flow *cnxk_flow_create(struct rte_eth_dev *dev, const struct rte_flow_attr *attr,
const struct rte_flow_item pattern[],
const struct rte_flow_action actions[],
struct rte_flow_error *error);
diff --git a/drivers/net/cnxk/cnxk_rep.h b/drivers/net/cnxk/cnxk_rep.h
index b0cf3340a2..1b46e202a8 100644
--- a/drivers/net/cnxk/cnxk_rep.h
+++ b/drivers/net/cnxk/cnxk_rep.h
@@ -20,6 +20,9 @@
/* Common ethdev ops */
extern struct eth_dev_ops cnxk_rep_dev_ops;
+/* Flow ops for representor ports */
+extern struct rte_flow_ops cnxk_rep_flow_ops;
+
struct cnxk_rep_queue_stats {
uint64_t pkts;
uint64_t bytes;
diff --git a/drivers/net/cnxk/cnxk_rep_flow.c b/drivers/net/cnxk/cnxk_rep_flow.c
new file mode 100644
index 0000000000..2613be5b9e
--- /dev/null
+++ b/drivers/net/cnxk/cnxk_rep_flow.c
@@ -0,0 +1,401 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2024 Marvell.
+ */
+
+#include <fcntl.h>
+#include <unistd.h>
+
+#include <cnxk_flow.h>
+#include <cnxk_rep.h>
+#include <cnxk_rep_msg.h>
+
+#define DEFAULT_DUMP_FILE_NAME "/tmp/fdump"
+#define MAX_BUFFER_SIZE 1500
+
+const struct cnxk_rte_flow_action_info action_info[] = {
+ [RTE_FLOW_ACTION_TYPE_MARK] = {sizeof(struct rte_flow_action_mark)},
+ [RTE_FLOW_ACTION_TYPE_VF] = {sizeof(struct rte_flow_action_vf)},
+ [RTE_FLOW_ACTION_TYPE_REPRESENTED_PORT] = {sizeof(struct rte_flow_action_port_id)},
+ [RTE_FLOW_ACTION_TYPE_PORT_ID] = {sizeof(struct rte_flow_action_port_id)},
+ [RTE_FLOW_ACTION_TYPE_QUEUE] = {sizeof(struct rte_flow_action_queue)},
+ [RTE_FLOW_ACTION_TYPE_RSS] = {sizeof(struct rte_flow_action_rss)},
+ [RTE_FLOW_ACTION_TYPE_SECURITY] = {sizeof(struct rte_flow_action_security)},
+ [RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID] = {sizeof(struct rte_flow_action_of_set_vlan_vid)},
+ [RTE_FLOW_ACTION_TYPE_OF_PUSH_VLAN] = {sizeof(struct rte_flow_action_of_push_vlan)},
+ [RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_PCP] = {sizeof(struct rte_flow_action_of_set_vlan_pcp)},
+ [RTE_FLOW_ACTION_TYPE_METER] = {sizeof(struct rte_flow_action_meter)},
+ [RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP] = {sizeof(struct rte_flow_action_vxlan_encap)},
+ [RTE_FLOW_ACTION_TYPE_COUNT] = {sizeof(struct rte_flow_action_count)},
+};
+
+static void
+cnxk_flow_params_count(const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ uint16_t *n_pattern, uint16_t *n_action)
+{
+ int i = 0;
+
+ for (; pattern->type != RTE_FLOW_ITEM_TYPE_END; pattern++)
+ i++;
+
+ *n_pattern = ++i;
+ plt_rep_dbg("Total patterns is %d", *n_pattern);
+
+ i = 0;
+ for (; actions->type != RTE_FLOW_ACTION_TYPE_END; actions++)
+ i++;
+ *n_action = ++i;
+ plt_rep_dbg("Total actions is %d", *n_action);
+}
+
+static void
+populate_attr_data(void *buffer, uint32_t *length, const struct rte_flow_attr *attr)
+{
+ uint32_t sz = sizeof(struct rte_flow_attr);
+ uint32_t len;
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_ATTR, sz);
+
+ len = *length;
+ /* Populate the attribute data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), attr, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static uint16_t
+prepare_pattern_data(const struct rte_flow_item *pattern, uint16_t nb_pattern,
+ uint64_t *pattern_data)
+{
+ cnxk_pattern_hdr_t hdr;
+ uint16_t len = 0;
+ int i = 0;
+
+ for (i = 0; i < nb_pattern; i++) {
+ /* Populate the pattern type hdr */
+ memset(&hdr, 0, sizeof(cnxk_pattern_hdr_t));
+ hdr.type = pattern->type;
+ if (pattern->spec) {
+ hdr.spec_sz = term[pattern->type].item_size;
+ hdr.last_sz = 0;
+ hdr.mask_sz = term[pattern->type].item_size;
+ }
+
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), &hdr, sizeof(cnxk_pattern_hdr_t));
+ len += sizeof(cnxk_pattern_hdr_t);
+
+ /* Copy pattern spec data */
+ if (pattern->spec) {
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), pattern->spec,
+ term[pattern->type].item_size);
+ len += term[pattern->type].item_size;
+ }
+
+ /* Copy pattern last data */
+ if (pattern->last) {
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), pattern->last,
+ term[pattern->type].item_size);
+ len += term[pattern->type].item_size;
+ }
+
+ /* Copy pattern mask data */
+ if (pattern->mask) {
+ rte_memcpy(RTE_PTR_ADD(pattern_data, len), pattern->mask,
+ term[pattern->type].item_size);
+ len += term[pattern->type].item_size;
+ }
+ pattern++;
+ }
+
+ return len;
+}
+
+static void
+populate_pattern_data(void *buffer, uint32_t *length, const struct rte_flow_item *pattern,
+ uint16_t nb_pattern)
+{
+ uint64_t pattern_data[BUFSIZ];
+ uint32_t len;
+ uint32_t sz;
+
+ memset(pattern_data, 0, BUFSIZ * sizeof(uint64_t));
+ /* Prepare pattern_data */
+ sz = prepare_pattern_data(pattern, nb_pattern, pattern_data);
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_PATTERN, sz);
+
+ len = *length;
+ /* Populate the pattern data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), pattern_data, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static uint16_t
+populate_rss_action_conf(const struct rte_flow_action_rss *conf, void *rss_action_conf)
+{
+ int len, sz;
+
+ len = sizeof(struct rte_flow_action_rss) - sizeof(conf->key) - sizeof(conf->queue);
+
+ if (rss_action_conf)
+ rte_memcpy(rss_action_conf, conf, len);
+
+ if (conf->key) {
+ sz = conf->key_len;
+ if (rss_action_conf)
+ rte_memcpy(RTE_PTR_ADD(rss_action_conf, len), conf->key, sz);
+ len += sz;
+ }
+
+ if (conf->queue) {
+ sz = conf->queue_num * sizeof(conf->queue);
+ if (rss_action_conf)
+ rte_memcpy(RTE_PTR_ADD(rss_action_conf, len), conf->queue, sz);
+ len += sz;
+ }
+
+ return len;
+}
+
+static uint16_t
+populate_vxlan_encap_action_conf(const struct rte_flow_action_vxlan_encap *vxlan_conf,
+ void *vxlan_encap_action_data)
+{
+ const struct rte_flow_item *pattern;
+ uint64_t nb_patterns = 0;
+ uint16_t len, sz;
+
+ pattern = vxlan_conf->definition;
+ for (; pattern->type != RTE_FLOW_ITEM_TYPE_END; pattern++)
+ nb_patterns++;
+
+ len = sizeof(uint64_t);
+ rte_memcpy(vxlan_encap_action_data, &nb_patterns, len);
+ pattern = vxlan_conf->definition;
+ /* Prepare pattern_data */
+ sz = prepare_pattern_data(pattern, nb_patterns, RTE_PTR_ADD(vxlan_encap_action_data, len));
+
+ len += sz;
+ if (len > BUFSIZ) {
+ plt_err("Incomplete item definition loaded, len %d", len);
+ return 0;
+ }
+
+ return len;
+}
+
+static uint16_t
+prepare_action_data(const struct rte_flow_action *action, uint16_t nb_action, uint64_t *action_data)
+{
+ void *action_conf_data = NULL;
+ cnxk_action_hdr_t hdr;
+ uint16_t len = 0, sz = 0;
+ int i = 0;
+
+ for (i = 0; i < nb_action; i++) {
+ if (action->conf) {
+ switch (action->type) {
+ case RTE_FLOW_ACTION_TYPE_RSS:
+ sz = populate_rss_action_conf(action->conf, NULL);
+ action_conf_data = plt_zmalloc(sz, 0);
+ if (populate_rss_action_conf(action->conf, action_conf_data) !=
+ sz) {
+ plt_err("Populating RSS action config failed");
+ return 0;
+ }
+ break;
+ case RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP:
+ action_conf_data = plt_zmalloc(BUFSIZ, 0);
+ sz = populate_vxlan_encap_action_conf(action->conf,
+ action_conf_data);
+ if (!sz) {
+ plt_err("Populating vxlan action config failed");
+ return 0;
+ }
+ break;
+ default:
+ sz = action_info[action->type].conf_size;
+ action_conf_data = plt_zmalloc(sz, 0);
+ rte_memcpy(action_conf_data, action->conf, sz);
+ break;
+ };
+ }
+
+ /* Populate the action type hdr */
+ memset(&hdr, 0, sizeof(cnxk_action_hdr_t));
+ hdr.type = action->type;
+ hdr.conf_sz = sz;
+
+ rte_memcpy(RTE_PTR_ADD(action_data, len), &hdr, sizeof(cnxk_action_hdr_t));
+ len += sizeof(cnxk_action_hdr_t);
+
+ /* Copy action conf data */
+ if (action_conf_data) {
+ rte_memcpy(RTE_PTR_ADD(action_data, len), action_conf_data, sz);
+ len += sz;
+ plt_free(action_conf_data);
+ action_conf_data = NULL;
+ }
+
+ action++;
+ }
+
+ return len;
+}
+
+static void
+populate_action_data(void *buffer, uint32_t *length, const struct rte_flow_action *action,
+ uint16_t nb_action)
+{
+ uint64_t action_data[BUFSIZ];
+ uint32_t len;
+ uint32_t sz;
+
+ memset(action_data, 0, BUFSIZ * sizeof(uint64_t));
+ /* Prepare action_data */
+ sz = prepare_action_data(action, nb_action, action_data);
+
+ cnxk_rep_msg_populate_type(buffer, length, CNXK_TYPE_ACTION, sz);
+
+ len = *length;
+ /* Populate the action data */
+ rte_memcpy(RTE_PTR_ADD(buffer, len), action_data, sz);
+ len += sz;
+
+ *length = len;
+}
+
+static int
+process_flow_rule(struct cnxk_rep_dev *rep_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ cnxk_rep_msg_ack_data_t *adata, cnxk_rep_msg_t msg)
+{
+ cnxk_rep_msg_flow_create_meta_t msg_fc_meta;
+ uint16_t n_pattern, n_action;
+ uint32_t len = 0, rc = 0;
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ /* Get no of actions and patterns */
+ cnxk_flow_params_count(pattern, actions, &n_pattern, &n_action);
+
+ /* Adding the header */
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ /* Representor port identified as rep_xport queue */
+ msg_fc_meta.portid = rep_dev->rep_id;
+ msg_fc_meta.nb_pattern = n_pattern;
+ msg_fc_meta.nb_action = n_action;
+
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_fc_meta,
+ sizeof(cnxk_rep_msg_flow_create_meta_t), msg);
+
+ /* Populate flow create parameters data */
+ populate_attr_data(buffer, &len, attr);
+ populate_pattern_data(buffer, &len, pattern, n_pattern);
+ populate_action_data(buffer, &len, actions, n_action);
+
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static struct rte_flow *
+cnxk_rep_flow_create_native(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[],
+ const struct rte_flow_action actions[], struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ struct roc_npc_flow *flow;
+ uint16_t new_entry;
+ int rc;
+
+ flow = cnxk_flow_create_common(eth_dev, attr, pattern, actions, error, true);
+ if (!flow) {
+ plt_err("Fail to create flow");
+ goto fail;
+ }
+
+ /* Shifting the rules with higher priority than exception path rules */
+ new_entry = (uint16_t)flow->mcam_id;
+ rc = cnxk_eswitch_flow_rule_shift(rep_dev->hw_func, &new_entry);
+ if (rc) {
+ plt_err("Failed to shift the flow rule entry, err %d", rc);
+ goto fail;
+ }
+
+ flow->mcam_id = new_entry;
+
+ return (struct rte_flow *)flow;
+fail:
+ return NULL;
+}
+
+static struct rte_flow *
+cnxk_rep_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ struct rte_flow *flow = NULL;
+ cnxk_rep_msg_ack_data_t adata;
+ int rc = 0;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_rep_flow_create_native(eth_dev, attr, pattern, actions, error);
+
+ rc = process_flow_rule(rep_dev, attr, pattern, actions, &adata, CNXK_REP_MSG_FLOW_CREATE);
+ if (!rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0) {
+ rc = (int)adata.u.sval;
+ rte_flow_error_set(error, adata.u.sval, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+ NULL, "Failed to validate flow");
+ goto fail;
+ }
+
+ flow = adata.u.data;
+ if (!flow) {
+ rte_flow_error_set(error, adata.u.sval, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+ NULL, "Failed to create flow");
+ goto fail;
+ }
+ } else {
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to create flow");
+ goto fail;
+ }
+ plt_rep_dbg("Flow %p created successfully", adata.u.data);
+
+ return flow;
+fail:
+ return NULL;
+}
+
+struct rte_flow_ops cnxk_rep_flow_ops = {
+ .create = cnxk_rep_flow_create,
+};
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
index 277e25d92a..d27a234e48 100644
--- a/drivers/net/cnxk/cnxk_rep_msg.h
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -12,6 +12,10 @@
typedef enum CNXK_TYPE {
CNXK_TYPE_HEADER = 0,
CNXK_TYPE_MSG,
+ CNXK_TYPE_ATTR,
+ CNXK_TYPE_PATTERN,
+ CNXK_TYPE_ACTION,
+ CNXK_TYPE_FLOW
} cnxk_type_t;
typedef enum CNXK_REP_MSG {
@@ -23,6 +27,8 @@ typedef enum CNXK_REP_MSG {
CNXK_REP_MSG_ETH_SET_MAC,
CNXK_REP_MSG_ETH_STATS_GET,
CNXK_REP_MSG_ETH_STATS_CLEAR,
+ /* Flow operation msgs */
+ CNXK_REP_MSG_FLOW_CREATE,
/* End of messaging sequence */
CNXK_REP_MSG_END,
} cnxk_rep_msg_t;
@@ -96,6 +102,27 @@ typedef struct cnxk_rep_msg_eth_stats_meta {
uint16_t portid;
} __rte_packed cnxk_rep_msg_eth_stats_meta_t;
+/* Flow create msg meta */
+typedef struct cnxk_rep_msg_flow_create_meta {
+ uint16_t portid;
+ uint16_t nb_pattern;
+ uint16_t nb_action;
+} __rte_packed cnxk_rep_msg_flow_create_meta_t;
+
+/* Type pattern meta */
+typedef struct cnxk_pattern_hdr {
+ uint16_t type;
+ uint16_t spec_sz;
+ uint16_t last_sz;
+ uint16_t mask_sz;
+} __rte_packed cnxk_pattern_hdr_t;
+
+/* Type action meta */
+typedef struct cnxk_action_hdr {
+ uint16_t type;
+ uint16_t conf_sz;
+} __rte_packed cnxk_action_hdr_t;
+
void cnxk_rep_msg_populate_command(void *buffer, uint32_t *length, cnxk_rep_msg_t type,
uint32_t size);
void cnxk_rep_msg_populate_command_meta(void *buffer, uint32_t *length, void *msg_meta, uint32_t sz,
diff --git a/drivers/net/cnxk/cnxk_rep_ops.c b/drivers/net/cnxk/cnxk_rep_ops.c
index 6728b3e265..80322a0cfd 100644
--- a/drivers/net/cnxk/cnxk_rep_ops.c
+++ b/drivers/net/cnxk/cnxk_rep_ops.c
@@ -637,7 +637,8 @@ int
cnxk_rep_flow_ops_get(struct rte_eth_dev *ethdev, const struct rte_flow_ops **ops)
{
PLT_SET_USED(ethdev);
- PLT_SET_USED(ops);
+ *ops = &cnxk_rep_flow_ops;
+
return 0;
}
diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build
index 9ca7732713..8cc06f4967 100644
--- a/drivers/net/cnxk/meson.build
+++ b/drivers/net/cnxk/meson.build
@@ -39,6 +39,7 @@ sources = files(
'cnxk_rep.c',
'cnxk_rep_msg.c',
'cnxk_rep_ops.c',
+ 'cnxk_rep_flow.c',
'cnxk_stats.c',
'cnxk_tm.c',
)
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v3 23/23] net/cnxk: other flow operations
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
` (21 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 22/23] net/cnxk: flow create on representor ports Harman Kalra
@ 2024-02-01 13:07 ` Harman Kalra
22 siblings, 0 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-01 13:07 UTC (permalink / raw)
To: Nithin Dabilpuram, Kiran Kumar K, Sunil Kumar Kori, Satha Rao,
Harman Kalra
Cc: dev
Implementing other flow operations - validate, destroy, query,
flush, dump for representor ports
Signed-off-by: Harman Kalra <hkalra@marvell.com>
---
drivers/net/cnxk/cnxk_rep_flow.c | 414 +++++++++++++++++++++++++++++++
drivers/net/cnxk/cnxk_rep_msg.h | 32 +++
2 files changed, 446 insertions(+)
diff --git a/drivers/net/cnxk/cnxk_rep_flow.c b/drivers/net/cnxk/cnxk_rep_flow.c
index 2613be5b9e..d26f5aa12c 100644
--- a/drivers/net/cnxk/cnxk_rep_flow.c
+++ b/drivers/net/cnxk/cnxk_rep_flow.c
@@ -267,6 +267,222 @@ populate_action_data(void *buffer, uint32_t *length, const struct rte_flow_actio
*length = len;
}
+static int
+process_flow_destroy(struct cnxk_rep_dev *rep_dev, void *flow, cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_destroy_meta_t msg_fd_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active)
+ return 0;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_fd_meta.portid = rep_dev->rep_id;
+ msg_fd_meta.flow = (uint64_t)flow;
+ plt_rep_dbg("Flow Destroy: flow 0x%" PRIu64 ", portid %d", msg_fd_meta.flow,
+ msg_fd_meta.portid);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_fd_meta,
+ sizeof(cnxk_rep_msg_flow_destroy_meta_t),
+ CNXK_REP_MSG_FLOW_DESTROY);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+copy_flow_dump_file(FILE *target)
+{
+ FILE *source = NULL;
+ int pos;
+ char ch;
+
+ source = fopen(DEFAULT_DUMP_FILE_NAME, "r");
+ if (source == NULL) {
+ plt_err("Failed to read default dump file: %s, err %d", DEFAULT_DUMP_FILE_NAME,
+ errno);
+ return errno;
+ }
+
+ fseek(source, 0L, SEEK_END);
+ pos = ftell(source);
+ fseek(source, 0L, SEEK_SET);
+ while (pos--) {
+ ch = fgetc(source);
+ fputc(ch, target);
+ }
+
+ fclose(source);
+
+ /* Remove the default file after reading */
+ remove(DEFAULT_DUMP_FILE_NAME);
+
+ return 0;
+}
+
+static int
+process_flow_dump(struct cnxk_rep_dev *rep_dev, struct rte_flow *flow, FILE *file,
+ cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_dump_meta_t msg_fp_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_fp_meta.portid = rep_dev->rep_id;
+ msg_fp_meta.flow = (uint64_t)flow;
+ msg_fp_meta.is_stdout = (file == stdout) ? 1 : 0;
+
+ plt_rep_dbg("Flow Dump: flow 0x%" PRIu64 ", portid %d stdout %d", msg_fp_meta.flow,
+ msg_fp_meta.portid, msg_fp_meta.is_stdout);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_fp_meta,
+ sizeof(cnxk_rep_msg_flow_dump_meta_t),
+ CNXK_REP_MSG_FLOW_DUMP);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ /* Copy contents from default file to user file */
+ if (file != stdout)
+ copy_flow_dump_file(file);
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+process_flow_flush(struct cnxk_rep_dev *rep_dev, cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_flush_meta_t msg_ff_meta;
+ uint32_t len = 0, rc;
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ msg_ff_meta.portid = rep_dev->rep_id;
+ plt_rep_dbg("Flow Flush: portid %d", msg_ff_meta.portid);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, &msg_ff_meta,
+ sizeof(cnxk_rep_msg_flow_flush_meta_t),
+ CNXK_REP_MSG_FLOW_FLUSH);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+process_flow_query(struct cnxk_rep_dev *rep_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data, cnxk_rep_msg_ack_data_t *adata)
+{
+ cnxk_rep_msg_flow_query_meta_t *msg_fq_meta;
+ struct rte_flow_query_count *query = data;
+ uint32_t len = 0, rc, sz, total_sz;
+ uint64_t action_data[BUFSIZ];
+ void *buffer;
+ size_t size;
+
+ size = MAX_BUFFER_SIZE;
+ buffer = plt_zmalloc(size, 0);
+ if (!buffer) {
+ plt_err("Failed to allocate mem");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ cnxk_rep_msg_populate_header(buffer, &len);
+
+ memset(action_data, 0, BUFSIZ * sizeof(uint64_t));
+ sz = prepare_action_data(action, 1, action_data);
+ total_sz = sz + sizeof(cnxk_rep_msg_flow_query_meta_t);
+
+ msg_fq_meta = plt_zmalloc(total_sz, 0);
+ if (!msg_fq_meta) {
+ plt_err("Failed to allocate memory");
+ rc = -ENOMEM;
+ goto fail;
+ }
+
+ msg_fq_meta->portid = rep_dev->rep_id;
+ msg_fq_meta->reset = query->reset;
+ ;
+ msg_fq_meta->flow = (uint64_t)flow;
+ /* Populate the action data */
+ rte_memcpy(msg_fq_meta->action_data, action_data, sz);
+ msg_fq_meta->action_data_sz = sz;
+
+ plt_rep_dbg("Flow query: flow 0x%" PRIu64 ", portid %d, action type %d total sz %d "
+ "action sz %d", msg_fq_meta->flow, msg_fq_meta->portid, action->type, total_sz,
+ sz);
+ cnxk_rep_msg_populate_command_meta(buffer, &len, msg_fq_meta, total_sz,
+ CNXK_REP_MSG_FLOW_QUERY);
+ cnxk_rep_msg_populate_msg_end(buffer, &len);
+
+ rc = cnxk_rep_msg_send_process(rep_dev, buffer, len, adata);
+ if (rc) {
+ plt_err("Failed to process the message, err %d", rc);
+ goto free;
+ }
+
+ rte_free(msg_fq_meta);
+
+ return 0;
+
+free:
+ rte_free(msg_fq_meta);
+fail:
+ return rc;
+}
+
static int
process_flow_rule(struct cnxk_rep_dev *rep_dev, const struct rte_flow_attr *attr,
const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
@@ -396,6 +612,204 @@ cnxk_rep_flow_create(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *at
return NULL;
}
+static int
+cnxk_rep_flow_validate(struct rte_eth_dev *eth_dev, const struct rte_flow_attr *attr,
+ const struct rte_flow_item pattern[], const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc = 0;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_flow_validate_common(eth_dev, attr, pattern, actions, error, true);
+
+ rc = process_flow_rule(rep_dev, attr, pattern, actions, &adata, CNXK_REP_MSG_FLOW_VALIDATE);
+ if (!rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0) {
+ rc = (int)adata.u.sval;
+ rte_flow_error_set(error, adata.u.sval, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+ NULL, "Failed to validate flow");
+ goto fail;
+ }
+ } else {
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to validate flow");
+ }
+
+ plt_rep_dbg("Flow %p validated successfully", adata.u.data);
+
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_destroy(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_flow_destroy_common(eth_dev, (struct roc_npc_flow *)flow, error, true);
+
+ rc = process_flow_destroy(rep_dev, flow, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to destroy flow");
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_query(struct rte_eth_dev *eth_dev, struct rte_flow *flow,
+ const struct rte_flow_action *action, void *data, struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (action->type != RTE_FLOW_ACTION_TYPE_COUNT) {
+ rc = -ENOTSUP;
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Only COUNT is supported in query");
+ goto fail;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_flow_query_common(eth_dev, flow, action, data, error, true);
+
+ rc = process_flow_query(rep_dev, flow, action, data, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to query the flow");
+ goto fail;
+ }
+
+ rte_memcpy(data, adata.u.data, adata.size);
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_flush(struct rte_eth_dev *eth_dev, struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_flow_flush_common(eth_dev, error, true);
+
+ rc = process_flow_flush(rep_dev, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to destroy flow");
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_dev_dump(struct rte_eth_dev *eth_dev, struct rte_flow *flow, FILE *file,
+ struct rte_flow_error *error)
+{
+ struct cnxk_rep_dev *rep_dev = cnxk_rep_pmd_priv(eth_dev);
+ cnxk_rep_msg_ack_data_t adata;
+ int rc;
+
+ /* If representor not representing any active VF, return 0 */
+ if (!rep_dev->is_vf_active) {
+ rte_flow_error_set(error, -EAGAIN, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Represented VF not active yet");
+ return 0;
+ }
+
+ if (rep_dev->native_repte)
+ return cnxk_flow_dev_dump_common(eth_dev, flow, file, error, true);
+
+ rc = process_flow_dump(rep_dev, flow, file, &adata);
+ if (rc || adata.u.sval < 0) {
+ if (adata.u.sval < 0)
+ rc = adata.u.sval;
+
+ rte_flow_error_set(error, rc, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Failed to destroy flow");
+ goto fail;
+ }
+
+ return 0;
+fail:
+ return rc;
+}
+
+static int
+cnxk_rep_flow_isolate(struct rte_eth_dev *eth_dev __rte_unused, int enable __rte_unused,
+ struct rte_flow_error *error)
+{
+ /* If we support, we need to un-install the default mcam
+ * entry for this port.
+ */
+
+ rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+ "Flow isolation not supported");
+
+ return -rte_errno;
+}
+
struct rte_flow_ops cnxk_rep_flow_ops = {
+ .validate = cnxk_rep_flow_validate,
.create = cnxk_rep_flow_create,
+ .destroy = cnxk_rep_flow_destroy,
+ .query = cnxk_rep_flow_query,
+ .flush = cnxk_rep_flow_flush,
+ .isolate = cnxk_rep_flow_isolate,
+ .dev_dump = cnxk_rep_flow_dev_dump,
};
diff --git a/drivers/net/cnxk/cnxk_rep_msg.h b/drivers/net/cnxk/cnxk_rep_msg.h
index d27a234e48..bfd9ce9c7b 100644
--- a/drivers/net/cnxk/cnxk_rep_msg.h
+++ b/drivers/net/cnxk/cnxk_rep_msg.h
@@ -29,6 +29,11 @@ typedef enum CNXK_REP_MSG {
CNXK_REP_MSG_ETH_STATS_CLEAR,
/* Flow operation msgs */
CNXK_REP_MSG_FLOW_CREATE,
+ CNXK_REP_MSG_FLOW_DESTROY,
+ CNXK_REP_MSG_FLOW_VALIDATE,
+ CNXK_REP_MSG_FLOW_FLUSH,
+ CNXK_REP_MSG_FLOW_DUMP,
+ CNXK_REP_MSG_FLOW_QUERY,
/* End of messaging sequence */
CNXK_REP_MSG_END,
} cnxk_rep_msg_t;
@@ -109,6 +114,33 @@ typedef struct cnxk_rep_msg_flow_create_meta {
uint16_t nb_action;
} __rte_packed cnxk_rep_msg_flow_create_meta_t;
+/* Flow destroy msg meta */
+typedef struct cnxk_rep_msg_flow_destroy_meta {
+ uint64_t flow;
+ uint16_t portid;
+} __rte_packed cnxk_rep_msg_flow_destroy_meta_t;
+
+/* Flow flush msg meta */
+typedef struct cnxk_rep_msg_flow_flush_meta {
+ uint16_t portid;
+} __rte_packed cnxk_rep_msg_flow_flush_meta_t;
+
+/* Flow dump msg meta */
+typedef struct cnxk_rep_msg_flow_dump_meta {
+ uint64_t flow;
+ uint16_t portid;
+ uint8_t is_stdout;
+} __rte_packed cnxk_rep_msg_flow_dump_meta_t;
+
+/* Flow query msg meta */
+typedef struct cnxk_rep_msg_flow_query_meta {
+ uint64_t flow;
+ uint16_t portid;
+ uint8_t reset;
+ uint32_t action_data_sz;
+ uint8_t action_data[];
+} __rte_packed cnxk_rep_msg_flow_query_meta_t;
+
/* Type pattern meta */
typedef struct cnxk_pattern_hdr {
uint16_t type;
--
2.18.0
^ permalink raw reply [flat|nested] 142+ messages in thread
* [PATCH v4 00/23] net/cnxk: support for port representors
2023-08-11 16:34 [PATCH 0/9] net/cnxk: support for port representors Harman Kalra
` (10 preceding siblings ...)
2024-02-01 13:07 ` [PATCH v3 00/23] net/cnxk: support for port representors Harman Kalra
@ 2024-02-27 19:15 ` Harman Kalra
2024-02-27 19:15 ` [PATCH v4 01/23] common/cnxk: add support for representors Harman Kalra
` (22 more replies)
2024-03-01 19:14 ` [PATCH v5 00/23] net/cnxk: support for port representors Harman Kalra
2024-03-03 17:38 ` [PATCH v6 00/23] net/cnxk: support for port representors Harman Kalra
13 siblings, 23 replies; 142+ messages in thread
From: Harman Kalra @ 2024-02-27 19:15 UTC (permalink / raw)
Cc: dev, Harman Kalra
Introducing port representor support to CNXK drivers by adding virtual ethernet
ports providing a logical representation in DPDK for physical function(PF) or
SR-IOV virtual function (VF) devices for control and monitoring.
These port representor ethdev instances can be spawned on an as needed basis
through configuration parameters passed to the driver of the underlying
base device using devargs ``-a <base PCI BDF>,representor=pf*vf*``
In case of ex