DPDK patches and discussions
 help / color / mirror / Atom feed
From: Heinrich Kuhn <heinrich.kuhn@netronome.com>
To: dev@dpdk.org
Cc: Heinrich Kuhn <heinrich.kuhn@netronome.com>,
	Simon Horman <simon.horman@corigine.com>
Subject: [dpdk-dev] [PATCH v2 2/2] net/nfp: fix PF secondary process probing
Date: Wed,  9 Jun 2021 16:08:14 +0200	[thread overview]
Message-ID: <20210609140814.76977-2-heinrich.kuhn@netronome.com> (raw)
In-Reply-To: <20210609140814.76977-1-heinrich.kuhn@netronome.com>

This patch creates a new function for handling PF probing of a secondary
process. A CPP handle is obtained for the CPP bridge service and the
service itself is also registered during secondary process
initialization. DPDK services aren't shared between processes so it is
not enough to only have the primary register the service if it is also
needed in a secondary process. This implies that both the primary and
secondary will have their own copy of the bridge service.

Fixes: 5e15e799d697 ("net/nfp: create separate entity for PF device")

Signed-off-by: Heinrich Kuhn <heinrich.kuhn@netronome.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
---
 drivers/net/nfp/nfp_net.c | 147 +++++++++++++++++++++++++++-----------
 1 file changed, 105 insertions(+), 42 deletions(-)

diff --git a/drivers/net/nfp/nfp_net.c b/drivers/net/nfp/nfp_net.c
index d94c8dc727..b18edd8c7b 100644
--- a/drivers/net/nfp/nfp_net.c
+++ b/drivers/net/nfp/nfp_net.c
@@ -59,6 +59,7 @@ static int nfp_net_infos_get(struct rte_eth_dev *dev,
 			     struct rte_eth_dev_info *dev_info);
 static int nfp_net_init(struct rte_eth_dev *eth_dev);
 static int nfp_pf_init(struct rte_pci_device *pci_dev);
+static int nfp_pf_secondary_init(struct rte_pci_device *pci_dev);
 static int nfp_pci_uninit(struct rte_eth_dev *eth_dev);
 static int nfp_init_phyports(struct nfp_pf_dev *pf_dev);
 static int nfp_net_link_update(struct rte_eth_dev *dev, int wait_to_complete);
@@ -98,6 +99,7 @@ static int nfp_net_rss_hash_write(struct rte_eth_dev *dev,
 static int nfp_set_mac_addr(struct rte_eth_dev *dev,
 			     struct rte_ether_addr *mac_addr);
 static int32_t nfp_cpp_bridge_service_func(void *args);
+static void nfp_register_cpp_service(struct nfp_cpp *cpp);
 static int nfp_fw_setup(struct rte_pci_device *dev,
 			struct nfp_cpp *cpp,
 			struct nfp_eth_table *nfp_eth_table,
@@ -3516,20 +3518,6 @@ static int nfp_init_phyports(struct nfp_pf_dev *pf_dev)
 		snprintf(port_name, sizeof(port_name), "%s_port%d",
 			 pf_dev->pci_dev->device.name, i);
 
-		if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
-			eth_dev = rte_eth_dev_attach_secondary(port_name);
-			if (!eth_dev) {
-				RTE_LOG(ERR, EAL,
-				"secondary process attach failed, "
-				"ethdev doesn't exist");
-				ret = -ENODEV;
-				goto error;
-			}
-
-			eth_dev->process_private = pf_dev->cpp;
-			goto nfp_net_init;
-		}
-
 		/* Allocate a eth_dev for this phyport */
 		eth_dev = rte_eth_dev_allocate(port_name);
 		if (!eth_dev) {
@@ -3559,7 +3547,6 @@ static int nfp_init_phyports(struct nfp_pf_dev *pf_dev)
 		hw->nfp_idx = nfp_eth_table->ports[i].index;
 		hw->is_phyport = true;
 
-nfp_net_init:
 		eth_dev->device = &pf_dev->pci_dev->device;
 
 		/* ctrl/tx/rx BAR mappings and remaining init happens in
@@ -3593,6 +3580,23 @@ static int nfp_init_phyports(struct nfp_pf_dev *pf_dev)
 	return ret;
 }
 
+static void nfp_register_cpp_service(struct nfp_cpp *cpp)
+{
+	uint32_t *cpp_service_id = NULL;
+	struct rte_service_spec service;
+
+	memset(&service, 0, sizeof(struct rte_service_spec));
+	snprintf(service.name, sizeof(service.name), "nfp_cpp_service");
+	service.callback = nfp_cpp_bridge_service_func;
+	service.callback_userdata = (void *)cpp;
+
+	if (rte_service_component_register(&service,
+					   cpp_service_id))
+		RTE_LOG(WARNING, PMD, "NFP CPP bridge service register() failed");
+	else
+		RTE_LOG(DEBUG, PMD, "NFP CPP bridge service registered");
+}
+
 static int nfp_pf_init(struct rte_pci_device *pci_dev)
 {
 	struct nfp_pf_dev *pf_dev = NULL;
@@ -3600,8 +3604,6 @@ static int nfp_pf_init(struct rte_pci_device *pci_dev)
 	struct nfp_hwinfo *hwinfo;
 	struct nfp_rtsym_table *sym_tbl;
 	struct nfp_eth_table *nfp_eth_table = NULL;
-	struct rte_service_spec service;
-	uint32_t *nfp_cpp_service_id = NULL;
 	char name[RTE_ETH_NAME_MAX_LEN];
 	int total_ports;
 	int ret = -ENODEV;
@@ -3642,12 +3644,10 @@ static int nfp_pf_init(struct rte_pci_device *pci_dev)
 		goto hwinfo_cleanup;
 	}
 
-	if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
-		if (nfp_fw_setup(pci_dev, cpp, nfp_eth_table, hwinfo)) {
-			PMD_INIT_LOG(ERR, "Error when uploading firmware");
-			ret = -EIO;
-			goto eth_table_cleanup;
-		}
+	if (nfp_fw_setup(pci_dev, cpp, nfp_eth_table, hwinfo)) {
+		PMD_INIT_LOG(ERR, "Error when uploading firmware");
+		ret = -EIO;
+		goto eth_table_cleanup;
 	}
 
 	/* Now the symbol table should be there */
@@ -3726,24 +3726,8 @@ static int nfp_pf_init(struct rte_pci_device *pci_dev)
 		goto hwqueues_cleanup;
 	}
 
-	/*
-	 * The rte_service needs to be created just once per PMD.
-	 * And the cpp handler needs to be linked to the service.
-	 * Secondary processes will be used for debugging DPDK apps
-	 * when requiring to use the CPP interface for accessing NFP
-	 * components. And the cpp handler for secondary processes is
-	 * available at this point.
-	 */
-	memset(&service, 0, sizeof(struct rte_service_spec));
-	snprintf(service.name, sizeof(service.name), "nfp_cpp_service");
-	service.callback = nfp_cpp_bridge_service_func;
-	service.callback_userdata = (void *)cpp;
-
-	if (rte_service_component_register(&service,
-					   nfp_cpp_service_id))
-		RTE_LOG(ERR, PMD, "NFP CPP bridge service register() failed");
-	else
-		RTE_LOG(DEBUG, PMD, "NFP CPP bridge service registered");
+	/* register the CPP bridge service here for primary use */
+	nfp_register_cpp_service(pf_dev->cpp);
 
 	return 0;
 
@@ -3763,10 +3747,89 @@ static int nfp_pf_init(struct rte_pci_device *pci_dev)
 	return ret;
 }
 
+/*
+ * When attaching to the NFP4000/6000 PF on a secondary process there
+ * is no need to initialize the PF again. Only minimal work is required
+ * here
+ */
+static int nfp_pf_secondary_init(struct rte_pci_device *pci_dev)
+{
+	struct nfp_cpp *cpp;
+	struct nfp_rtsym_table *sym_tbl;
+	int total_ports;
+	int i;
+	int err;
+
+	if (!pci_dev)
+		return -ENODEV;
+
+	/*
+	 * When device bound to UIO, the device could be used, by mistake,
+	 * by two DPDK apps, and the UIO driver does not avoid it. This
+	 * could lead to a serious problem when configuring the NFP CPP
+	 * interface. Here we avoid this telling to the CPP init code to
+	 * use a lock file if UIO is being used.
+	 */
+	if (pci_dev->kdrv == RTE_PCI_KDRV_VFIO)
+		cpp = nfp_cpp_from_device_name(pci_dev, 0);
+	else
+		cpp = nfp_cpp_from_device_name(pci_dev, 1);
+
+	if (!cpp) {
+		PMD_INIT_LOG(ERR, "A CPP handle can not be obtained");
+		return -EIO;
+	}
+
+	/*
+	 * We don't have access to the PF created in the primary process
+	 * here so we have to read the number of ports from firmware
+	 */
+	sym_tbl = nfp_rtsym_table_read(cpp);
+	if (!sym_tbl) {
+		PMD_INIT_LOG(ERR, "Something is wrong with the firmware"
+				" symbol table");
+		return -EIO;
+	}
+
+	total_ports = nfp_rtsym_read_le(sym_tbl, "nfd_cfg_pf0_num_ports", &err);
+
+	for (i = 0; i < total_ports; i++) {
+		struct rte_eth_dev *eth_dev;
+		char port_name[RTE_ETH_NAME_MAX_LEN];
+
+		snprintf(port_name, sizeof(port_name), "%s_port%d",
+			 pci_dev->device.name, i);
+
+		PMD_DRV_LOG(DEBUG, "Secondary attaching to port %s",
+		    port_name);
+		eth_dev = rte_eth_dev_attach_secondary(port_name);
+		if (!eth_dev) {
+			RTE_LOG(ERR, EAL,
+			"secondary process attach failed, "
+			"ethdev doesn't exist");
+			return -ENODEV;
+		}
+		eth_dev->process_private = cpp;
+		eth_dev->dev_ops = &nfp_net_eth_dev_ops;
+		eth_dev->rx_queue_count = nfp_net_rx_queue_count;
+		eth_dev->rx_pkt_burst = &nfp_net_recv_pkts;
+		eth_dev->tx_pkt_burst = &nfp_net_xmit_pkts;
+		rte_eth_dev_probing_finish(eth_dev);
+	}
+
+	/* Register the CPP bridge service for the secondary too */
+	nfp_register_cpp_service(cpp);
+
+	return 0;
+}
+
 static int nfp_pf_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 			    struct rte_pci_device *dev)
 {
-	return nfp_pf_init(dev);
+	if (rte_eal_process_type() == RTE_PROC_PRIMARY)
+		return nfp_pf_init(dev);
+	else
+		return nfp_pf_secondary_init(dev);
 }
 
 static const struct rte_pci_id pci_id_nfp_pf_net_map[] = {
-- 
2.30.1 (Apple Git-130)


  reply	other threads:[~2021-06-09 14:08 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-09 14:08 [dpdk-dev] [PATCH v2 1/2] net/nfp: improve PF probing logic Heinrich Kuhn
2021-06-09 14:08 ` Heinrich Kuhn [this message]
2021-07-01 14:33   ` [dpdk-dev] [PATCH v2 2/2] net/nfp: fix PF secondary process probing Andrew Rybchenko
2021-07-01 14:33 ` [dpdk-dev] [PATCH v2 1/2] net/nfp: improve PF probing logic Andrew Rybchenko

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210609140814.76977-2-heinrich.kuhn@netronome.com \
    --to=heinrich.kuhn@netronome.com \
    --cc=dev@dpdk.org \
    --cc=simon.horman@corigine.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).