[dpdk-dev] [PATCH 04/13] cxgbe: add probe to initialize VF devices

[Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>]

From: Kumar Sanghvi <kumaras@chelsio.com>

Add probe to initialize VF devices.  Separate init/de-init paths
for PF and VF.  Do firmware state initialization wrt VF and retrieve
various operational parameters by querying firmware.  Finally configure
and initialize ports.

Signed-off-by: Kumar Sanghvi <kumaras@chelsio.com>
Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
---
 drivers/net/cxgbe/Makefile         |   1 +
 drivers/net/cxgbe/base/adapter.h   |   1 +
 drivers/net/cxgbe/base/t4_hw.c     |   2 +-
 drivers/net/cxgbe/cxgbe.h          |   3 +
 drivers/net/cxgbe/cxgbe_main.c     |  17 ++-
 drivers/net/cxgbe/cxgbevf_ethdev.c |  65 ++++++++-
 drivers/net/cxgbe/cxgbevf_main.c   | 280 +++++++++++++++++++++++++++++++++++++
 7 files changed, 359 insertions(+), 10 deletions(-)
 create mode 100644 drivers/net/cxgbe/cxgbevf_main.c

diff --git a/drivers/net/cxgbe/Makefile b/drivers/net/cxgbe/Makefile
index 66c445ea3..fe177ff90 100644
--- a/drivers/net/cxgbe/Makefile
+++ b/drivers/net/cxgbe/Makefile
@@ -82,6 +82,7 @@ VPATH += $(SRCDIR)/base
 SRCS-$(CONFIG_RTE_LIBRTE_CXGBE_PMD) += cxgbe_ethdev.c
 SRCS-$(CONFIG_RTE_LIBRTE_CXGBE_PMD) += cxgbevf_ethdev.c
 SRCS-$(CONFIG_RTE_LIBRTE_CXGBE_PMD) += cxgbe_main.c
+SRCS-$(CONFIG_RTE_LIBRTE_CXGBE_PMD) += cxgbevf_main.c
 SRCS-$(CONFIG_RTE_LIBRTE_CXGBE_PMD) += sge.c
 SRCS-$(CONFIG_RTE_LIBRTE_CXGBE_PMD) += t4_hw.c
 SRCS-$(CONFIG_RTE_LIBRTE_CXGBE_PMD) += t4vf_hw.c
diff --git a/drivers/net/cxgbe/base/adapter.h b/drivers/net/cxgbe/base/adapter.h
index 169402836..6b2fc8b34 100644
--- a/drivers/net/cxgbe/base/adapter.h
+++ b/drivers/net/cxgbe/base/adapter.h
@@ -68,6 +68,7 @@ struct port_info {
 	u8     port_type;               /* firmware port type */
 	u8     mod_type;                /* firmware module type */
 	u8     port_id;                 /* physical port ID */
+	u8     pidx;			/* port index for this PF */
 	u8     tx_chan;                 /* associated channel */
 
 	u8     n_rx_qsets;              /* # of rx qsets */
diff --git a/drivers/net/cxgbe/base/t4_hw.c b/drivers/net/cxgbe/base/t4_hw.c
index 436a9953d..b3938943c 100644
--- a/drivers/net/cxgbe/base/t4_hw.c
+++ b/drivers/net/cxgbe/base/t4_hw.c
@@ -4488,7 +4488,7 @@ static void t4_handle_get_port_info(struct port_info *pi, const __be64 *rpl)
 		lc->auto_fec = fec;
 		pi->port_type = port_type;
 		pi->mod_type = mod_type;
-		t4_os_portmod_changed(adapter, pi->port_id);
+		t4_os_portmod_changed(adapter, pi->pidx);
 	}
 	if (link_ok != lc->link_ok || speed != lc->speed ||
 	    fc != lc->fc || fec != lc->fec) { /* something changed */
diff --git a/drivers/net/cxgbe/cxgbe.h b/drivers/net/cxgbe/cxgbe.h
index 489e09dda..040aca423 100644
--- a/drivers/net/cxgbe/cxgbe.h
+++ b/drivers/net/cxgbe/cxgbe.h
@@ -49,6 +49,7 @@
 #define CXGBE_DEFAULT_RSS_KEY_LEN     40 /* 320-bits */
 
 int cxgbe_probe(struct adapter *adapter);
+int cxgbevf_probe(struct adapter *adapter);
 void cxgbe_get_speed_caps(struct port_info *pi, u32 *speed_caps);
 int cxgbe_up(struct adapter *adap);
 int cxgbe_down(struct port_info *pi);
@@ -63,5 +64,7 @@ void cfg_queues(struct rte_eth_dev *eth_dev);
 int cfg_queue_count(struct rte_eth_dev *eth_dev);
 int setup_rss(struct port_info *pi);
 void cxgbe_enable_rx_queues(struct port_info *pi);
+void print_port_info(struct adapter *adap);
+void print_adapter_info(struct adapter *adap);
 
 #endif /* _CXGBE_H_ */
diff --git a/drivers/net/cxgbe/cxgbe_main.c b/drivers/net/cxgbe/cxgbe_main.c
index 5792032e7..8d78ad427 100644
--- a/drivers/net/cxgbe/cxgbe_main.c
+++ b/drivers/net/cxgbe/cxgbe_main.c
@@ -370,7 +370,7 @@ static int init_rss(struct adapter *adap)
 /**
  * Dump basic information about the adapter.
  */
-static void print_adapter_info(struct adapter *adap)
+void print_adapter_info(struct adapter *adap)
 {
 	/**
 	 * Hardware/Firmware/etc. Version/Revision IDs.
@@ -378,7 +378,7 @@ static void print_adapter_info(struct adapter *adap)
 	t4_dump_version_info(adap);
 }
 
-static void print_port_info(struct adapter *adap)
+void print_port_info(struct adapter *adap)
 {
 	int i;
 	char buf[80];
@@ -917,7 +917,7 @@ int link_start(struct port_info *pi)
 			ret = 0;
 		}
 	}
-	if (ret == 0)
+	if (ret == 0 && is_pf4(adapter))
 		ret = t4_link_l1cfg(adapter, adapter->mbox, pi->tx_chan,
 				    &pi->link_cfg);
 	if (ret == 0) {
@@ -1196,7 +1196,8 @@ int cxgbe_up(struct adapter *adap)
 {
 	enable_rx(adap, &adap->sge.fw_evtq);
 	t4_sge_tx_monitor_start(adap);
-	t4_intr_enable(adap);
+	if (is_pf4(adap))
+		t4_intr_enable(adap);
 	adap->flags |= FULL_INIT_DONE;
 
 	/* TODO: deadman watchdog ?? */
@@ -1217,7 +1218,7 @@ int cxgbe_down(struct port_info *pi)
 		return err;
 	}
 
-	t4_reset_link_config(adapter, pi->port_id);
+	t4_reset_link_config(adapter, pi->pidx);
 	return 0;
 }
 
@@ -1230,7 +1231,8 @@ void cxgbe_close(struct adapter *adapter)
 	int i;
 
 	if (adapter->flags & FULL_INIT_DONE) {
-		t4_intr_disable(adapter);
+		if (is_pf4(adapter))
+			t4_intr_disable(adapter);
 		t4_sge_tx_monitor_stop(adapter);
 		t4_free_sge_resources(adapter);
 		for_each_port(adapter, i) {
@@ -1248,7 +1250,7 @@ void cxgbe_close(struct adapter *adapter)
 		adapter->flags &= ~FULL_INIT_DONE;
 	}
 
-	if (adapter->flags & FW_OK)
+	if (is_pf4(adapter) && (adapter->flags & FW_OK))
 		t4_fw_bye(adapter, adapter->mbox);
 }
 
@@ -1355,6 +1357,7 @@ int cxgbe_probe(struct adapter *adapter)
 		pi->adapter = adapter;
 		pi->xact_addr_filt = -1;
 		pi->port_id = i;
+		pi->pidx = i;
 
 		pi->eth_dev->device = &adapter->pdev->device;
 		pi->eth_dev->dev_ops = adapter->eth_dev->dev_ops;
diff --git a/drivers/net/cxgbe/cxgbevf_ethdev.c b/drivers/net/cxgbe/cxgbevf_ethdev.c
index 7d5b8c07f..3b1deac52 100644
--- a/drivers/net/cxgbe/cxgbevf_ethdev.c
+++ b/drivers/net/cxgbe/cxgbevf_ethdev.c
@@ -59,12 +59,73 @@ static const struct eth_dev_ops cxgbevf_eth_dev_ops = {
  */
 static int eth_cxgbevf_dev_init(struct rte_eth_dev *eth_dev)
 {
+	struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+	struct rte_pci_device *pci_dev;
+	char name[RTE_ETH_NAME_MAX_LEN];
+	struct adapter *adapter = NULL;
+	int err = 0;
+
 	CXGBE_FUNC_TRACE();
 
 	eth_dev->dev_ops = &cxgbevf_eth_dev_ops;
+	eth_dev->rx_pkt_burst = NULL;
+	eth_dev->tx_pkt_burst = NULL;
+	pci_dev = RTE_ETH_DEV_TO_PCI(eth_dev);
+
+	/* for secondary processes, we attach to ethdevs allocated by primary
+	 * and do minimal initialization.
+	 */
+	if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
+		int i;
+
+		for (i = 1; i < MAX_NPORTS; i++) {
+			struct rte_eth_dev *rest_eth_dev;
+			char namei[RTE_ETH_NAME_MAX_LEN];
+
+			snprintf(namei, sizeof(namei), "%s_%d",
+				 pci_dev->device.name, i);
+			rest_eth_dev = rte_eth_dev_attach_secondary(namei);
+			if (rest_eth_dev) {
+				rest_eth_dev->device = &pci_dev->device;
+				rest_eth_dev->dev_ops =
+					eth_dev->dev_ops;
+				rest_eth_dev->rx_pkt_burst =
+					eth_dev->rx_pkt_burst;
+				rest_eth_dev->tx_pkt_burst =
+					eth_dev->tx_pkt_burst;
+			}
+		}
+		return 0;
+	}
+
+	snprintf(name, sizeof(name), "cxgbevfadapter%d",
+		 eth_dev->data->port_id);
+	adapter = rte_zmalloc(name, sizeof(*adapter), 0);
+	if (!adapter)
+		return -1;
+
+	adapter->use_unpacked_mode = 1;
+	adapter->regs = (void *)pci_dev->mem_resource[0].addr;
+	if (!adapter->regs) {
+		dev_err(adapter, "%s: cannot map device registers\n", __func__);
+		err = -ENOMEM;
+		goto out_free_adapter;
+	}
+	adapter->pdev = pci_dev;
+	adapter->eth_dev = eth_dev;
+	pi->adapter = adapter;
+	err = cxgbevf_probe(adapter);
+	if (err) {
+		dev_err(adapter, "%s: cxgbevf probe failed with err %d\n",
+			__func__, err);
+		goto out_free_adapter;
+	}
+
+	return 0;
 
-	/* XXX: Do probe */
-	return -EIO;
+out_free_adapter:
+	rte_free(adapter);
+	return err;
 }
 
 static int eth_cxgbevf_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
diff --git a/drivers/net/cxgbe/cxgbevf_main.c b/drivers/net/cxgbe/cxgbevf_main.c
new file mode 100644
index 000000000..0624267c7
--- /dev/null
+++ b/drivers/net/cxgbe/cxgbevf_main.c
@@ -0,0 +1,280 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2018 Chelsio Communications.
+ * All rights reserved.
+ */
+
+#include <rte_ethdev_driver.h>
+#include <rte_ethdev_pci.h>
+#include <rte_malloc.h>
+
+#include "common.h"
+#include "t4_regs.h"
+#include "t4_msg.h"
+#include "cxgbe.h"
+
+/*
+ * Figure out how many Ports and Queue Sets we can support.  This depends on
+ * knowing our Virtual Function Resources and may be called a second time if
+ * we fall back from MSI-X to MSI Interrupt Mode.
+ */
+static void size_nports_qsets(struct adapter *adapter)
+{
+	struct vf_resources *vfres = &adapter->params.vfres;
+	unsigned int ethqsets, pmask_nports;
+
+	/*
+	 * The number of "ports" which we support is equal to the number of
+	 * Virtual Interfaces with which we've been provisioned.
+	 */
+	adapter->params.nports = vfres->nvi;
+	if (adapter->params.nports > MAX_NPORTS) {
+		dev_warn(adapter->pdev_dev, "only using %d of %d maximum"
+			 " allowed virtual interfaces\n", MAX_NPORTS,
+			 adapter->params.nports);
+		adapter->params.nports = MAX_NPORTS;
+	}
+
+	/*
+	 * We may have been provisioned with more VIs than the number of
+	 * ports we're allowed to access (our Port Access Rights Mask).
+	 * This is obviously a configuration conflict but we don't want to
+	 * do anything silly just because of that.
+	 */
+	pmask_nports = hweight32(adapter->params.vfres.pmask);
+	if (pmask_nports < adapter->params.nports) {
+		dev_warn(adapter->pdev_dev, "only using %d of %d provissioned"
+			 " virtual interfaces; limited by Port Access Rights"
+			 " mask %#x\n", pmask_nports, adapter->params.nports,
+			 adapter->params.vfres.pmask);
+		adapter->params.nports = pmask_nports;
+	}
+
+	/*
+	 * We need to reserve an Ingress Queue for the Asynchronous Firmware
+	 * Event Queue.
+	 *
+	 * For each Queue Set, we'll need the ability to allocate two Egress
+	 * Contexts -- one for the Ingress Queue Free List and one for the TX
+	 * Ethernet Queue.
+	 */
+	ethqsets = vfres->niqflint - 1;
+	if (vfres->nethctrl != ethqsets)
+		ethqsets = min(vfres->nethctrl, ethqsets);
+	if (vfres->neq < ethqsets * 2)
+		ethqsets = vfres->neq / 2;
+	if (ethqsets > MAX_ETH_QSETS)
+		ethqsets = MAX_ETH_QSETS;
+	adapter->sge.max_ethqsets = ethqsets;
+
+	if (adapter->sge.max_ethqsets < adapter->params.nports) {
+		dev_warn(adapter->pdev_dev, "only using %d of %d available"
+			 " virtual interfaces (too few Queue Sets)\n",
+			 adapter->sge.max_ethqsets, adapter->params.nports);
+		adapter->params.nports = adapter->sge.max_ethqsets;
+	}
+}
+
+static int adap_init0vf(struct adapter *adapter)
+{
+	u32 param, val = 0;
+	int err;
+
+	err = t4vf_fw_reset(adapter);
+	if (err < 0) {
+		dev_err(adapter->pdev_dev, "FW reset failed: err=%d\n", err);
+		return err;
+	}
+
+	/*
+	 * Grab basic operational parameters.  These will predominantly have
+	 * been set up by the Physical Function Driver or will be hard coded
+	 * into the adapter.  We just have to live with them ...  Note that
+	 * we _must_ get our VPD parameters before our SGE parameters because
+	 * we need to know the adapter's core clock from the VPD in order to
+	 * properly decode the SGE Timer Values.
+	 */
+	err = t4vf_get_dev_params(adapter);
+	if (err) {
+		dev_err(adapter->pdev_dev, "unable to retrieve adapter"
+			" device parameters: err=%d\n", err);
+		return err;
+	}
+
+	err = t4vf_get_vpd_params(adapter);
+	if (err) {
+		dev_err(adapter->pdev_dev, "unable to retrieve adapter"
+			" VPD parameters: err=%d\n", err);
+		return err;
+	}
+
+	adapter->pf = t4vf_get_pf_from_vf(adapter);
+
+	/* If we're running on newer firmware, let it know that we're
+	 * prepared to deal with encapsulated CPL messages.  Older
+	 * firmware won't understand this and we'll just get
+	 * unencapsulated messages ...
+	 */
+	param = V_FW_PARAMS_MNEM(FW_PARAMS_MNEM_PFVF) |
+		V_FW_PARAMS_PARAM_X(FW_PARAMS_PARAM_PFVF_CPLFW4MSG_ENCAP);
+	val = 1;
+	t4vf_set_params(adapter, 1, &param, &val);
+
+	/*
+	 * Grab our Virtual Interface resource allocation, extract the
+	 * features that we're interested in and do a bit of sanity testing on
+	 * what we discover.
+	 */
+	err = t4vf_get_vfres(adapter);
+	if (err) {
+		dev_err(adapter->pdev_dev, "unable to get virtual interface"
+			" resources: err=%d\n", err);
+		return err;
+	}
+
+	/*
+	 * Check for various parameter sanity issues.
+	 */
+	if (adapter->params.vfres.pmask == 0) {
+		dev_err(adapter->pdev_dev, "no port access configured\n"
+			"usable!\n");
+		return -EINVAL;
+	}
+	if (adapter->params.vfres.nvi == 0) {
+		dev_err(adapter->pdev_dev, "no virtual interfaces configured/"
+			"usable!\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * Initialize nports and max_ethqsets now that we have our Virtual
+	 * Function Resources.
+	 */
+	size_nports_qsets(adapter);
+	adapter->flags |= FW_OK;
+	return 0;
+}
+
+int cxgbevf_probe(struct adapter *adapter)
+{
+	struct port_info *pi;
+	unsigned int pmask;
+	int err = 0;
+	int i;
+
+	t4_os_lock_init(&adapter->mbox_lock);
+	TAILQ_INIT(&adapter->mbox_list);
+	err = t4vf_prep_adapter(adapter);
+	if (err)
+		return err;
+
+	if (!is_t4(adapter->params.chip)) {
+		adapter->bar2 = (void *)adapter->pdev->mem_resource[2].addr;
+		if (!adapter->bar2) {
+			dev_err(adapter, "cannot map device bar2 region\n");
+			err = -ENOMEM;
+			return err;
+		}
+	}
+
+	err = adap_init0vf(adapter);
+	if (err) {
+		dev_err(adapter, "%s: Adapter initialization failed, error %d\n",
+				__func__, err);
+		goto out_free;
+	}
+
+	pmask = adapter->params.vfres.pmask;
+	for_each_port(adapter, i) {
+		const unsigned int numa_node = rte_socket_id();
+		char name[RTE_ETH_NAME_MAX_LEN];
+		struct rte_eth_dev *eth_dev;
+		int port_id;
+
+		if (pmask == 0)
+			break;
+		port_id = ffs(pmask) - 1;
+		pmask &= ~(1 << port_id);
+
+		snprintf(name, sizeof(name), "%s_%d",
+			 adapter->pdev->device.name, i);
+
+		if (i == 0) {
+			/* First port is already allocated by DPDK */
+			eth_dev = adapter->eth_dev;
+			goto allocate_mac;
+		}
+
+		/*
+		 * now do all data allocation - for eth_dev structure,
+		 * and internal (private) data for the remaining ports
+		 */
+
+		/* reserve an ethdev entry */
+		eth_dev = rte_eth_dev_allocate(name);
+		if (!eth_dev) {
+			err = -ENOMEM;
+			goto out_free;
+		}
+		eth_dev->data->dev_private =
+			rte_zmalloc_socket(name, sizeof(struct port_info),
+					   RTE_CACHE_LINE_SIZE, numa_node);
+		if (!eth_dev->data->dev_private)
+			goto out_free;
+
+allocate_mac:
+		pi = (struct port_info *)eth_dev->data->dev_private;
+		adapter->port[i] = pi;
+		pi->eth_dev = eth_dev;
+		pi->adapter = adapter;
+		pi->xact_addr_filt = -1;
+		pi->port_id = port_id;
+		pi->pidx = i;
+
+		pi->eth_dev->device = &adapter->pdev->device;
+		pi->eth_dev->dev_ops = adapter->eth_dev->dev_ops;
+		pi->eth_dev->tx_pkt_burst = adapter->eth_dev->tx_pkt_burst;
+		pi->eth_dev->rx_pkt_burst = adapter->eth_dev->rx_pkt_burst;
+
+		rte_eth_copy_pci_info(pi->eth_dev, adapter->pdev);
+		pi->eth_dev->data->mac_addrs = rte_zmalloc(name,
+							   ETHER_ADDR_LEN, 0);
+		if (!pi->eth_dev->data->mac_addrs) {
+			dev_err(adapter, "%s: Mem allocation failed for storing mac addr, aborting\n",
+				__func__);
+			err = -ENOMEM;
+			goto out_free;
+		}
+	}
+
+	if (adapter->flags & FW_OK) {
+		err = t4vf_port_init(adapter);
+		if (err) {
+			dev_err(adapter, "%s: t4_port_init failed with err %d\n",
+				__func__, err);
+			goto out_free;
+		}
+	}
+
+	cfg_queues(adapter->eth_dev);
+	print_adapter_info(adapter);
+	print_port_info(adapter);
+
+	return 0;
+
+out_free:
+	for_each_port(adapter, i) {
+		pi = adap2pinfo(adapter, i);
+		if (pi->viid != 0)
+			t4_free_vi(adapter, adapter->mbox, adapter->pf,
+				   0, pi->viid);
+		/* Skip first port since it'll be de-allocated by DPDK */
+		if (i == 0)
+			continue;
+		if (pi->eth_dev) {
+			if (pi->eth_dev->data->dev_private)
+				rte_free(pi->eth_dev->data->dev_private);
+			rte_eth_dev_release_port(pi->eth_dev);
+		}
+	}
+	return -err;
+}
-- 
2.14.1