[PATCH net-next 00/13] nfp: abm: add basic support for advanced buffering NIC

[PATCH net-next 00/13] nfp: abm: add basic support for advanced buffering NIC

[Jakub Kicinski]

Hi!

This series lays groundwork for advanced buffer management NIC feature.
It makes necessary NFP core changes, spawns representors and adds devlink
glue.  Following series will add the actual buffering configuration (patch
series size limit).

First three patches add support for configuring NFP buffer pools via a
mailbox.  The existing devlink APIs are used for the purpose.

Third patch allows us to perform small reads from the NFP memory.

The rest of the patch set adds eswitch mode change support and makes
the driver spawn appropriate representors.

Jakub Kicinski (13):
  nfp: move rtsym helpers to pf code
  nfp: add support for per-PCI PF mailbox
  nfp: add shared buffer configuration
  nfp: core: allow 4-byte aligned accesses to Memory Units
  nfp: abm: add initial active buffer management NIC skeleton
  nfp: abm: create project-specific vNIC structure
  nfp: add app pointer to port representors
  devlink: don't take instance lock around eswitch mode set
  nfp: add devlink_eswitch_mode_set callback
  nfp: abm: spawn port netdevs
  nfp: abm: force Ethternet port up
  nfp: use split in naming of PCIe PF ports
  nfp: assign vNIC id as phys_port_name of vNICs which are not ports

 drivers/net/ethernet/netronome/Kconfig        |  10 +
 drivers/net/ethernet/netronome/nfp/Makefile   |   7 +
 drivers/net/ethernet/netronome/nfp/abm/ctrl.c |  58 +++
 drivers/net/ethernet/netronome/nfp/abm/main.c | 399 ++++++++++++++++++
 drivers/net/ethernet/netronome/nfp/abm/main.h |  75 ++++
 drivers/net/ethernet/netronome/nfp/nfp_abi.h  | 129 ++++++
 drivers/net/ethernet/netronome/nfp/nfp_app.c  |   3 +
 drivers/net/ethernet/netronome/nfp/nfp_app.h  |  13 +
 .../net/ethernet/netronome/nfp/nfp_app_nic.c  |   5 +-
 .../net/ethernet/netronome/nfp/nfp_devlink.c  |  35 ++
 drivers/net/ethernet/netronome/nfp/nfp_main.c | 140 ++++++
 drivers/net/ethernet/netronome/nfp/nfp_main.h |  27 +-
 drivers/net/ethernet/netronome/nfp/nfp_net.h  |   7 +-
 .../ethernet/netronome/nfp/nfp_net_common.c   |  20 +-
 .../ethernet/netronome/nfp/nfp_net_debugfs.c  |   4 +-
 .../net/ethernet/netronome/nfp/nfp_net_main.c |  72 +---
 .../net/ethernet/netronome/nfp/nfp_net_repr.c |   2 +-
 .../net/ethernet/netronome/nfp/nfp_net_repr.h |   3 +
 .../ethernet/netronome/nfp/nfp_netvf_main.c   |   2 +-
 drivers/net/ethernet/netronome/nfp/nfp_port.c |   8 +-
 drivers/net/ethernet/netronome/nfp/nfp_port.h |   6 +
 .../ethernet/netronome/nfp/nfp_shared_buf.c   | 180 ++++++++
 .../netronome/nfp/nfpcore/nfp6000_pcie.c      |  94 ++---
 net/core/devlink.c                            |   3 +-
 24 files changed, 1190 insertions(+), 112 deletions(-)
 create mode 100644 drivers/net/ethernet/netronome/nfp/abm/ctrl.c
 create mode 100644 drivers/net/ethernet/netronome/nfp/abm/main.c
 create mode 100644 drivers/net/ethernet/netronome/nfp/abm/main.h
 create mode 100644 drivers/net/ethernet/netronome/nfp/nfp_abi.h
 create mode 100644 drivers/net/ethernet/netronome/nfp/nfp_shared_buf.c

-- 
2.17.0

[PATCH net-next 01/13] nfp: move rtsym helpers to pf code

[Jakub Kicinski]

nfp_net_pf_rtsym_read_optional() and nfp_net_pf_map_rtsym() are not
really related to networking code.  Move them to the PF code and
remove the net from their names.  They will soon be needed by code
outside of nfp_net_main.c anyway.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
 drivers/net/ethernet/netronome/nfp/nfp_main.c | 32 ++++++++++
 drivers/net/ethernet/netronome/nfp/nfp_main.h |  6 ++
 .../net/ethernet/netronome/nfp/nfp_net_main.c | 61 ++++---------------
 3 files changed, 51 insertions(+), 48 deletions(-)

diff --git a/drivers/net/ethernet/netronome/nfp/nfp_main.c b/drivers/net/ethernet/netronome/nfp/nfp_main.c
index 0ade122805ad..1ef3623c6e1c 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_main.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_main.c
@@ -75,6 +75,38 @@ static const struct pci_device_id nfp_pci_device_ids[] = {
 };
 MODULE_DEVICE_TABLE(pci, nfp_pci_device_ids);
 
+int nfp_pf_rtsym_read_optional(struct nfp_pf *pf, const char *format,
+			       unsigned int default_val)
+{
+	char name[256];
+	int err = 0;
+	u64 val;
+
+	snprintf(name, sizeof(name), format, nfp_cppcore_pcie_unit(pf->cpp));
+
+	val = nfp_rtsym_read_le(pf->rtbl, name, &err);
+	if (err) {
+		if (err == -ENOENT)
+			return default_val;
+		nfp_err(pf->cpp, "Unable to read symbol %s\n", name);
+		return err;
+	}
+
+	return val;
+}
+
+u8 __iomem *
+nfp_pf_map_rtsym(struct nfp_pf *pf, const char *name, const char *sym_fmt,
+		 unsigned int min_size, struct nfp_cpp_area **area)
+{
+	char pf_symbol[256];
+
+	snprintf(pf_symbol, sizeof(pf_symbol), sym_fmt,
+		 nfp_cppcore_pcie_unit(pf->cpp));
+
+	return nfp_rtsym_map(pf->rtbl, pf_symbol, name, min_size, area);
+}
+
 static bool nfp_board_ready(struct nfp_pf *pf)
 {
 	const char *cp;
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_main.h b/drivers/net/ethernet/netronome/nfp/nfp_main.h
index 42211083b51f..5b028486c894 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_main.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_main.h
@@ -177,6 +177,12 @@ nfp_net_get_mac_addr(struct nfp_pf *pf, struct net_device *netdev,
 
 bool nfp_ctrl_tx(struct nfp_net *nn, struct sk_buff *skb);
 
+int nfp_pf_rtsym_read_optional(struct nfp_pf *pf, const char *format,
+			       unsigned int default_val);
+u8 __iomem *
+nfp_pf_map_rtsym(struct nfp_pf *pf, const char *name, const char *sym_fmt,
+		 unsigned int min_size, struct nfp_cpp_area **area);
+
 enum nfp_dump_diag {
 	NFP_DUMP_NSP_DIAG = 0,
 };
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_main.c b/drivers/net/ethernet/netronome/nfp/nfp_net_main.c
index 45cd2092e498..f8abb4cd9cef 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_main.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_main.c
@@ -101,48 +101,15 @@ nfp_net_find_port(struct nfp_eth_table *eth_tbl, unsigned int index)
 	return NULL;
 }
 
-static int
-nfp_net_pf_rtsym_read_optional(struct nfp_pf *pf, const char *format,
-			       unsigned int default_val)
-{
-	char name[256];
-	int err = 0;
-	u64 val;
-
-	snprintf(name, sizeof(name), format, nfp_cppcore_pcie_unit(pf->cpp));
-
-	val = nfp_rtsym_read_le(pf->rtbl, name, &err);
-	if (err) {
-		if (err == -ENOENT)
-			return default_val;
-		nfp_err(pf->cpp, "Unable to read symbol %s\n", name);
-		return err;
-	}
-
-	return val;
-}
-
 static int nfp_net_pf_get_num_ports(struct nfp_pf *pf)
 {
-	return nfp_net_pf_rtsym_read_optional(pf, "nfd_cfg_pf%u_num_ports", 1);
+	return nfp_pf_rtsym_read_optional(pf, "nfd_cfg_pf%u_num_ports", 1);
 }
 
 static int nfp_net_pf_get_app_id(struct nfp_pf *pf)
 {
-	return nfp_net_pf_rtsym_read_optional(pf, "_pf%u_net_app_id",
-					      NFP_APP_CORE_NIC);
-}
-
-static u8 __iomem *
-nfp_net_pf_map_rtsym(struct nfp_pf *pf, const char *name, const char *sym_fmt,
-		     unsigned int min_size, struct nfp_cpp_area **area)
-{
-	char pf_symbol[256];
-
-	snprintf(pf_symbol, sizeof(pf_symbol), sym_fmt,
-		 nfp_cppcore_pcie_unit(pf->cpp));
-
-	return nfp_rtsym_map(pf->rtbl, pf_symbol, name, min_size, area);
+	return nfp_pf_rtsym_read_optional(pf, "_pf%u_net_app_id",
+					  NFP_APP_CORE_NIC);
 }
 
 static void nfp_net_pf_free_vnic(struct nfp_pf *pf, struct nfp_net *nn)
@@ -379,9 +346,8 @@ nfp_net_pf_app_init(struct nfp_pf *pf, u8 __iomem *qc_bar, unsigned int stride)
 	if (!nfp_app_needs_ctrl_vnic(pf->app))
 		return 0;
 
-	ctrl_bar = nfp_net_pf_map_rtsym(pf, "net.ctrl", "_pf%u_net_ctrl_bar",
-					NFP_PF_CSR_SLICE_SIZE,
-					&pf->ctrl_vnic_bar);
+	ctrl_bar = nfp_pf_map_rtsym(pf, "net.ctrl", "_pf%u_net_ctrl_bar",
+				    NFP_PF_CSR_SLICE_SIZE, &pf->ctrl_vnic_bar);
 	if (IS_ERR(ctrl_bar)) {
 		nfp_err(pf->cpp, "Failed to find ctrl vNIC memory symbol\n");
 		err = PTR_ERR(ctrl_bar);
@@ -507,8 +473,8 @@ static int nfp_net_pci_map_mem(struct nfp_pf *pf)
 	int err;
 
 	min_size = pf->max_data_vnics * NFP_PF_CSR_SLICE_SIZE;
-	mem = nfp_net_pf_map_rtsym(pf, "net.bar0", "_pf%d_net_bar0",
-				   min_size, &pf->data_vnic_bar);
+	mem = nfp_pf_map_rtsym(pf, "net.bar0", "_pf%d_net_bar0",
+			       min_size, &pf->data_vnic_bar);
 	if (IS_ERR(mem)) {
 		nfp_err(pf->cpp, "Failed to find data vNIC memory symbol\n");
 		return PTR_ERR(mem);
@@ -528,10 +494,9 @@ static int nfp_net_pci_map_mem(struct nfp_pf *pf)
 		}
 	}
 
-	pf->vf_cfg_mem = nfp_net_pf_map_rtsym(pf, "net.vfcfg",
-					      "_pf%d_net_vf_bar",
-					      NFP_NET_CFG_BAR_SZ *
-					      pf->limit_vfs, &pf->vf_cfg_bar);
+	pf->vf_cfg_mem = nfp_pf_map_rtsym(pf, "net.vfcfg", "_pf%d_net_vf_bar",
+					  NFP_NET_CFG_BAR_SZ * pf->limit_vfs,
+					  &pf->vf_cfg_bar);
 	if (IS_ERR(pf->vf_cfg_mem)) {
 		if (PTR_ERR(pf->vf_cfg_mem) != -ENOENT) {
 			err = PTR_ERR(pf->vf_cfg_mem);
@@ -541,9 +506,9 @@ static int nfp_net_pci_map_mem(struct nfp_pf *pf)
 	}
 
 	min_size = NFP_NET_VF_CFG_SZ * pf->limit_vfs + NFP_NET_VF_CFG_MB_SZ;
-	pf->vfcfg_tbl2 = nfp_net_pf_map_rtsym(pf, "net.vfcfg_tbl2",
-					      "_pf%d_net_vf_cfg2",
-					      min_size, &pf->vfcfg_tbl2_area);
+	pf->vfcfg_tbl2 = nfp_pf_map_rtsym(pf, "net.vfcfg_tbl2",
+					  "_pf%d_net_vf_cfg2",
+					  min_size, &pf->vfcfg_tbl2_area);
 	if (IS_ERR(pf->vfcfg_tbl2)) {
 		if (PTR_ERR(pf->vfcfg_tbl2) != -ENOENT) {
 			err = PTR_ERR(pf->vfcfg_tbl2);
-- 
2.17.0

[PATCH net-next 02/13] nfp: add support for per-PCI PF mailbox

[Jakub Kicinski]

When working with devlink-related functionality for locking reasons
it's easier to create a new mailbox per-PCI PF device than try to
use one of the netdev/vNIC mailboxes.

Define new mailbox structure and resolve its symbol during probe.
For forward compatibility allow silent truncation of mailbox command
data.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
 drivers/net/ethernet/netronome/nfp/nfp_abi.h  |  59 ++++++++++
 drivers/net/ethernet/netronome/nfp/nfp_main.c | 108 ++++++++++++++++++
 drivers/net/ethernet/netronome/nfp/nfp_main.h |   6 +
 3 files changed, 173 insertions(+)
 create mode 100644 drivers/net/ethernet/netronome/nfp/nfp_abi.h

diff --git a/drivers/net/ethernet/netronome/nfp/nfp_abi.h b/drivers/net/ethernet/netronome/nfp/nfp_abi.h
new file mode 100644
index 000000000000..ce3935d6729b
--- /dev/null
+++ b/drivers/net/ethernet/netronome/nfp/nfp_abi.h
@@ -0,0 +1,59 @@
+/* SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause) */
+/*
+ * Copyright (C) 2018 Netronome Systems, Inc.
+ *
+ * This software is dual licensed under the GNU General License Version 2,
+ * June 1991 as shown in the file COPYING in the top-level directory of this
+ * source tree or the BSD 2-Clause License provided below.  You have the
+ * option to license this software under the complete terms of either license.
+ *
+ * The BSD 2-Clause License:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      1. Redistributions of source code must retain the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer.
+ *
+ *      2. Redistributions in binary form must reproduce the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer in the documentation and/or other materials
+ *         provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __NFP_ABI__
+#define __NFP_ABI__ 1
+
+#include <linux/types.h>
+
+#define NFP_MBOX_SYM_NAME		"_abi_nfd_pf%u_mbox"
+#define NFP_MBOX_SYM_MIN_SIZE		16 /* When no data needed */
+
+#define NFP_MBOX_CMD		0x00
+#define NFP_MBOX_RET		0x04
+#define NFP_MBOX_DATA_LEN	0x08
+#define NFP_MBOX_RESERVED	0x0c
+#define NFP_MBOX_DATA		0x10
+
+/**
+ * enum nfp_mbox_cmd - PF mailbox commands
+ *
+ * @NFP_MBOX_NO_CMD:	null command
+ * Used to indicate previous command has finished.
+ */
+enum nfp_mbox_cmd {
+	NFP_MBOX_NO_CMD			= 0x00,
+};
+
+#endif
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_main.c b/drivers/net/ethernet/netronome/nfp/nfp_main.c
index 1ef3623c6e1c..46b76d5a726c 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_main.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_main.c
@@ -55,6 +55,7 @@
 
 #include "nfpcore/nfp6000_pcie.h"
 
+#include "nfp_abi.h"
 #include "nfp_app.h"
 #include "nfp_main.h"
 #include "nfp_net.h"
@@ -107,6 +108,90 @@ nfp_pf_map_rtsym(struct nfp_pf *pf, const char *name, const char *sym_fmt,
 	return nfp_rtsym_map(pf->rtbl, pf_symbol, name, min_size, area);
 }
 
+/* Callers should hold the devlink instance lock */
+int nfp_mbox_cmd(struct nfp_pf *pf, u32 cmd, void *in_data, u64 in_length,
+		 void *out_data, u64 out_length)
+{
+	unsigned long long addr;
+	unsigned long err_at;
+	u64 max_data_sz;
+	u32 val = 0;
+	u32 cpp_id;
+	int n, err;
+
+	if (!pf->mbox)
+		return -EOPNOTSUPP;
+
+	cpp_id = NFP_CPP_ISLAND_ID(pf->mbox->target, NFP_CPP_ACTION_RW, 0,
+				   pf->mbox->domain);
+	addr = pf->mbox->addr;
+	max_data_sz = pf->mbox->size - NFP_MBOX_SYM_MIN_SIZE;
+
+	/* Check if cmd field is clear */
+	err = nfp_cpp_readl(pf->cpp, cpp_id, addr + NFP_MBOX_CMD, &val);
+	if (err || val) {
+		nfp_warn(pf->cpp, "failed to issue command (%u): %u, err: %d\n",
+			 cmd, val, err);
+		return err ?: -EBUSY;
+	}
+
+	in_length = min(in_length, max_data_sz);
+	n = nfp_cpp_write(pf->cpp, cpp_id, addr + NFP_MBOX_DATA,
+			  in_data, in_length);
+	if (n != in_length)
+		return -EIO;
+	/* Write data_len and wipe reserved */
+	err = nfp_cpp_writeq(pf->cpp, cpp_id, addr + NFP_MBOX_DATA_LEN,
+			     in_length);
+	if (err)
+		return err;
+
+	/* Read back for ordering */
+	err = nfp_cpp_readl(pf->cpp, cpp_id, addr + NFP_MBOX_DATA_LEN, &val);
+	if (err)
+		return err;
+
+	/* Write cmd and wipe return value */
+	err = nfp_cpp_writeq(pf->cpp, cpp_id, addr + NFP_MBOX_CMD, cmd);
+	if (err)
+		return err;
+
+	err_at = jiffies + 5 * HZ;
+	while (true) {
+		/* Wait for command to go to 0 (NFP_MBOX_NO_CMD) */
+		err = nfp_cpp_readl(pf->cpp, cpp_id, addr + NFP_MBOX_CMD, &val);
+		if (err)
+			return err;
+		if (!val)
+			break;
+
+		if (time_is_before_eq_jiffies(err_at))
+			return -ETIMEDOUT;
+
+		msleep(5);
+	}
+
+	/* Copy output if any (could be error info, do it before reading ret) */
+	err = nfp_cpp_readl(pf->cpp, cpp_id, addr + NFP_MBOX_DATA_LEN, &val);
+	if (err)
+		return err;
+
+	out_length = min_t(u32, val, min(out_length, max_data_sz));
+	n = nfp_cpp_read(pf->cpp, cpp_id, addr + NFP_MBOX_DATA,
+			 out_data, out_length);
+	if (n != out_length)
+		return -EIO;
+
+	/* Check if there is an error */
+	err = nfp_cpp_readl(pf->cpp, cpp_id, addr + NFP_MBOX_RET, &val);
+	if (err)
+		return err;
+	if (val)
+		return -val;
+
+	return out_length;
+}
+
 static bool nfp_board_ready(struct nfp_pf *pf)
 {
 	const char *cp;
@@ -468,6 +553,25 @@ static void nfp_fw_unload(struct nfp_pf *pf)
 	nfp_nsp_close(nsp);
 }
 
+static int nfp_pf_find_rtsyms(struct nfp_pf *pf)
+{
+	char pf_symbol[256];
+	unsigned int pf_id;
+
+	pf_id = nfp_cppcore_pcie_unit(pf->cpp);
+
+	/* Optional per-PCI PF mailbox */
+	snprintf(pf_symbol, sizeof(pf_symbol), NFP_MBOX_SYM_NAME, pf_id);
+	pf->mbox = nfp_rtsym_lookup(pf->rtbl, pf_symbol);
+	if (pf->mbox && pf->mbox->size < NFP_MBOX_SYM_MIN_SIZE) {
+		nfp_err(pf->cpp, "PF mailbox symbol too small: %llu < %d\n",
+			pf->mbox->size, NFP_MBOX_SYM_MIN_SIZE);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static int nfp_pci_probe(struct pci_dev *pdev,
 			 const struct pci_device_id *pci_id)
 {
@@ -542,6 +646,10 @@ static int nfp_pci_probe(struct pci_dev *pdev,
 	pf->mip = nfp_mip_open(pf->cpp);
 	pf->rtbl = __nfp_rtsym_table_read(pf->cpp, pf->mip);
 
+	err = nfp_pf_find_rtsyms(pf);
+	if (err)
+		goto err_fw_unload;
+
 	pf->dump_flag = NFP_DUMP_NSP_DIAG;
 	pf->dumpspec = nfp_net_dump_load_dumpspec(pf->cpp, pf->rtbl);
 
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_main.h b/drivers/net/ethernet/netronome/nfp/nfp_main.h
index 5b028486c894..e3e1fa84ccd7 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_main.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_main.h
@@ -60,6 +60,7 @@ struct nfp_mip;
 struct nfp_net;
 struct nfp_nsp_identify;
 struct nfp_port;
+struct nfp_rtsym;
 struct nfp_rtsym_table;
 
 /**
@@ -87,6 +88,7 @@ struct nfp_dumpspec {
  * @vf_cfg_mem:		Pointer to mapped VF configuration area
  * @vfcfg_tbl2_area:	Pointer to the CPP area for the VF config table
  * @vfcfg_tbl2:		Pointer to mapped VF config table
+ * @mbox:		RTSym of per-PCI PF mailbox (under devlink lock)
  * @irq_entries:	Array of MSI-X entries for all vNICs
  * @limit_vfs:		Number of VFs supported by firmware (~0 for PCI limit)
  * @num_vfs:		Number of SR-IOV VFs enabled
@@ -127,6 +129,8 @@ struct nfp_pf {
 	struct nfp_cpp_area *vfcfg_tbl2_area;
 	u8 __iomem *vfcfg_tbl2;
 
+	const struct nfp_rtsym *mbox;
+
 	struct msix_entry *irq_entries;
 
 	unsigned int limit_vfs;
@@ -182,6 +186,8 @@ int nfp_pf_rtsym_read_optional(struct nfp_pf *pf, const char *format,
 u8 __iomem *
 nfp_pf_map_rtsym(struct nfp_pf *pf, const char *name, const char *sym_fmt,
 		 unsigned int min_size, struct nfp_cpp_area **area);
+int nfp_mbox_cmd(struct nfp_pf *pf, u32 cmd, void *in_data, u64 in_length,
+		 void *out_data, u64 out_length);
 
 enum nfp_dump_diag {
 	NFP_DUMP_NSP_DIAG = 0,
-- 
2.17.0

[PATCH net-next 03/13] nfp: add shared buffer configuration

[Jakub Kicinski]

Allow app FW to advertise its shared buffer pool information.
Use the per-PF mailbox to configure them from devlink.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
 drivers/net/ethernet/netronome/nfp/Makefile   |   1 +
 drivers/net/ethernet/netronome/nfp/nfp_abi.h  |  70 +++++++
 .../net/ethernet/netronome/nfp/nfp_devlink.c  |  22 +++
 drivers/net/ethernet/netronome/nfp/nfp_main.h |  15 +-
 .../net/ethernet/netronome/nfp/nfp_net_main.c |   7 +
 .../ethernet/netronome/nfp/nfp_shared_buf.c   | 180 ++++++++++++++++++
 6 files changed, 294 insertions(+), 1 deletion(-)
 create mode 100644 drivers/net/ethernet/netronome/nfp/nfp_shared_buf.c

diff --git a/drivers/net/ethernet/netronome/nfp/Makefile b/drivers/net/ethernet/netronome/nfp/Makefile
index d5866d708dfa..f032d645911c 100644
--- a/drivers/net/ethernet/netronome/nfp/Makefile
+++ b/drivers/net/ethernet/netronome/nfp/Makefile
@@ -30,6 +30,7 @@ nfp-objs := \
 	    nfp_net_sriov.o \
 	    nfp_netvf_main.o \
 	    nfp_port.o \
+	    nfp_shared_buf.o \
 	    nic/main.o
 
 ifeq ($(CONFIG_NFP_APP_FLOWER),y)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_abi.h b/drivers/net/ethernet/netronome/nfp/nfp_abi.h
index ce3935d6729b..7ffa6e6a9d1c 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_abi.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_abi.h
@@ -51,9 +51,79 @@
  *
  * @NFP_MBOX_NO_CMD:	null command
  * Used to indicate previous command has finished.
+ *
+ * @NFP_MBOX_POOL_GET:	get shared buffer pool info/config
+ * Input  - struct nfp_shared_buf_pool_id
+ * Output - struct nfp_shared_buf_pool_info_get
+ *
+ * @NFP_MBOX_POOL_SET:	set shared buffer pool info/config
+ * Input  - struct nfp_shared_buf_pool_info_set
+ * Output - None
  */
 enum nfp_mbox_cmd {
 	NFP_MBOX_NO_CMD			= 0x00,
+
+	NFP_MBOX_POOL_GET		= 0x01,
+	NFP_MBOX_POOL_SET		= 0x02,
+};
+
+#define NFP_SHARED_BUF_COUNT_SYM_NAME	"_abi_nfd_pf%u_sb_cnt"
+#define NFP_SHARED_BUF_TABLE_SYM_NAME	"_abi_nfd_pf%u_sb_tbl"
+
+/**
+ * struct nfp_shared_buf - NFP shared buffer description
+ * @id:				numerical user-visible id of the shared buffer
+ * @size:			size in bytes of the buffer
+ * @ingress_pools_count:	number of ingress pools
+ * @egress_pools_count:		number of egress pools
+ * @ingress_tc_count:		number of ingress trafic classes
+ * @egress_tc_count:		number of egress trafic classes
+ * @pool_size_unit:		pool size may be in credits, each credit is
+ *				@pool_size_unit bytes
+ */
+struct nfp_shared_buf {
+	__le32 id;
+	__le32 size;
+	__le16 ingress_pools_count;
+	__le16 egress_pools_count;
+	__le16 ingress_tc_count;
+	__le16 egress_tc_count;
+
+	__le32 pool_size_unit;
+};
+
+/**
+ * struct nfp_shared_buf_pool_id - shared buffer pool identification
+ * @shared_buf:		shared buffer id
+ * @pool:		pool index
+ */
+struct nfp_shared_buf_pool_id {
+	__le32 shared_buf;
+	__le32 pool;
+};
+
+/**
+ * struct nfp_shared_buf_pool_info_get - struct devlink_sb_pool_info mirror
+ * @pool_type:		one of enum devlink_sb_pool_type
+ * @size:		pool size in units of SB's @pool_size_unit
+ * @threshold_type:	one of enum devlink_sb_threshold_type
+ */
+struct nfp_shared_buf_pool_info_get {
+	__le32 pool_type;
+	__le32 size;
+	__le32 threshold_type;
+};
+
+/**
+ * struct nfp_shared_buf_pool_info_set - packed args of sb_pool_set
+ * @id:			pool identification info
+ * @size:		pool size in units of SB's @pool_size_unit
+ * @threshold_type:	one of enum devlink_sb_threshold_type
+ */
+struct nfp_shared_buf_pool_info_set {
+	struct nfp_shared_buf_pool_id id;
+	__le32 size;
+	__le32 threshold_type;
 };
 
 #endif
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_devlink.c b/drivers/net/ethernet/netronome/nfp/nfp_devlink.c
index b1e67cf4257a..73c7fcc820ac 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_devlink.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_devlink.c
@@ -149,6 +149,26 @@ nfp_devlink_port_unsplit(struct devlink *devlink, unsigned int port_index)
 	return ret;
 }
 
+static int
+nfp_devlink_sb_pool_get(struct devlink *devlink, unsigned int sb_index,
+			u16 pool_index, struct devlink_sb_pool_info *pool_info)
+{
+	struct nfp_pf *pf = devlink_priv(devlink);
+
+	return nfp_shared_buf_pool_get(pf, sb_index, pool_index, pool_info);
+}
+
+static int
+nfp_devlink_sb_pool_set(struct devlink *devlink, unsigned int sb_index,
+			u16 pool_index,
+			u32 size, enum devlink_sb_threshold_type threshold_type)
+{
+	struct nfp_pf *pf = devlink_priv(devlink);
+
+	return nfp_shared_buf_pool_set(pf, sb_index, pool_index,
+				       size, threshold_type);
+}
+
 static int nfp_devlink_eswitch_mode_get(struct devlink *devlink, u16 *mode)
 {
 	struct nfp_pf *pf = devlink_priv(devlink);
@@ -159,6 +179,8 @@ static int nfp_devlink_eswitch_mode_get(struct devlink *devlink, u16 *mode)
 const struct devlink_ops nfp_devlink_ops = {
 	.port_split		= nfp_devlink_port_split,
 	.port_unsplit		= nfp_devlink_port_unsplit,
+	.sb_pool_get		= nfp_devlink_sb_pool_get,
+	.sb_pool_set		= nfp_devlink_sb_pool_set,
 	.eswitch_mode_get	= nfp_devlink_eswitch_mode_get,
 };
 
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_main.h b/drivers/net/ethernet/netronome/nfp/nfp_main.h
index e3e1fa84ccd7..595b3dc280e3 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_main.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_main.h
@@ -46,10 +46,10 @@
 #include <linux/mutex.h>
 #include <linux/pci.h>
 #include <linux/workqueue.h>
+#include <net/devlink.h>
 
 struct dentry;
 struct device;
-struct devlink_ops;
 struct pci_dev;
 
 struct nfp_cpp;
@@ -62,6 +62,7 @@ struct nfp_nsp_identify;
 struct nfp_port;
 struct nfp_rtsym;
 struct nfp_rtsym_table;
+struct nfp_shared_buf;
 
 /**
  * struct nfp_dumpspec - NFP FW dump specification structure
@@ -110,6 +111,8 @@ struct nfp_dumpspec {
  * @ports:		Linked list of port structures (struct nfp_port)
  * @wq:			Workqueue for running works which need to grab @lock
  * @port_refresh_work:	Work entry for taking netdevs out
+ * @shared_bufs:	Array of shared buffer structures if FW has any SBs
+ * @num_shared_bufs:	Number of elements in @shared_bufs
  * @lock:		Protects all fields which may change after probe
  */
 struct nfp_pf {
@@ -162,6 +165,9 @@ struct nfp_pf {
 	struct workqueue_struct *wq;
 	struct work_struct port_refresh_work;
 
+	struct nfp_shared_buf *shared_bufs;
+	unsigned int num_shared_bufs;
+
 	struct mutex lock;
 };
 
@@ -200,4 +206,11 @@ s64 nfp_net_dump_calculate_size(struct nfp_pf *pf, struct nfp_dumpspec *spec,
 int nfp_net_dump_populate_buffer(struct nfp_pf *pf, struct nfp_dumpspec *spec,
 				 struct ethtool_dump *dump_param, void *dest);
 
+int nfp_shared_buf_register(struct nfp_pf *pf);
+void nfp_shared_buf_unregister(struct nfp_pf *pf);
+int nfp_shared_buf_pool_get(struct nfp_pf *pf, unsigned int sb, u16 pool_index,
+			    struct devlink_sb_pool_info *pool_info);
+int nfp_shared_buf_pool_set(struct nfp_pf *pf, unsigned int sb,
+			    u16 pool_index, u32 size,
+			    enum devlink_sb_threshold_type threshold_type);
 #endif /* NFP_MAIN_H */
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_main.c b/drivers/net/ethernet/netronome/nfp/nfp_net_main.c
index f8abb4cd9cef..b98422112385 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_main.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_main.c
@@ -728,6 +728,10 @@ int nfp_net_pci_probe(struct nfp_pf *pf)
 	if (err)
 		goto err_app_clean;
 
+	err = nfp_shared_buf_register(pf);
+	if (err)
+		goto err_devlink_unreg;
+
 	mutex_lock(&pf->lock);
 	pf->ddir = nfp_net_debugfs_device_add(pf->pdev);
 
@@ -761,6 +765,8 @@ int nfp_net_pci_probe(struct nfp_pf *pf)
 err_clean_ddir:
 	nfp_net_debugfs_dir_clean(&pf->ddir);
 	mutex_unlock(&pf->lock);
+	nfp_shared_buf_unregister(pf);
+err_devlink_unreg:
 	cancel_work_sync(&pf->port_refresh_work);
 	devlink_unregister(devlink);
 err_app_clean:
@@ -788,6 +794,7 @@ void nfp_net_pci_remove(struct nfp_pf *pf)
 
 	mutex_unlock(&pf->lock);
 
+	nfp_shared_buf_unregister(pf);
 	devlink_unregister(priv_to_devlink(pf));
 
 	nfp_net_pf_free_irqs(pf);
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_shared_buf.c b/drivers/net/ethernet/netronome/nfp/nfp_shared_buf.c
new file mode 100644
index 000000000000..0ecd83705368
--- /dev/null
+++ b/drivers/net/ethernet/netronome/nfp/nfp_shared_buf.c
@@ -0,0 +1,180 @@
+// SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
+/*
+ * Copyright (C) 2018 Netronome Systems, Inc.
+ *
+ * This software is dual licensed under the GNU General License Version 2,
+ * June 1991 as shown in the file COPYING in the top-level directory of this
+ * source tree or the BSD 2-Clause License provided below.  You have the
+ * option to license this software under the complete terms of either license.
+ *
+ * The BSD 2-Clause License:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      1. Redistributions of source code must retain the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer.
+ *
+ *      2. Redistributions in binary form must reproduce the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer in the documentation and/or other materials
+ *         provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/kernel.h>
+#include <net/devlink.h>
+
+#include "nfpcore/nfp_cpp.h"
+#include "nfpcore/nfp_nffw.h"
+#include "nfp_abi.h"
+#include "nfp_app.h"
+#include "nfp_main.h"
+
+static u32 nfp_shared_buf_pool_unit(struct nfp_pf *pf, unsigned int sb)
+{
+	__le32 sb_id = cpu_to_le32(sb);
+	unsigned int i;
+
+	for (i = 0; i < pf->num_shared_bufs; i++)
+		if (pf->shared_bufs[i].id == sb_id)
+			return le32_to_cpu(pf->shared_bufs[i].pool_size_unit);
+
+	WARN_ON_ONCE(1);
+	return 0;
+}
+
+int nfp_shared_buf_pool_get(struct nfp_pf *pf, unsigned int sb, u16 pool_index,
+			    struct devlink_sb_pool_info *pool_info)
+{
+	struct nfp_shared_buf_pool_info_get get_data;
+	struct nfp_shared_buf_pool_id id = {
+		.shared_buf	= cpu_to_le32(sb),
+		.pool		= cpu_to_le32(pool_index),
+	};
+	unsigned int unit_size;
+	int n;
+
+	unit_size = nfp_shared_buf_pool_unit(pf, sb);
+	if (!unit_size)
+		return -EINVAL;
+
+	n = nfp_mbox_cmd(pf, NFP_MBOX_POOL_GET, &id, sizeof(id),
+			 &get_data, sizeof(get_data));
+	if (n < 0)
+		return n;
+	if (n < sizeof(get_data))
+		return -EIO;
+
+	pool_info->pool_type = le32_to_cpu(get_data.pool_type);
+	pool_info->threshold_type = le32_to_cpu(get_data.threshold_type);
+	pool_info->size = le32_to_cpu(get_data.size) * unit_size;
+
+	return 0;
+}
+
+int nfp_shared_buf_pool_set(struct nfp_pf *pf, unsigned int sb,
+			    u16 pool_index, u32 size,
+			    enum devlink_sb_threshold_type threshold_type)
+{
+	struct nfp_shared_buf_pool_info_set set_data = {
+		.id = {
+			.shared_buf	= cpu_to_le32(sb),
+			.pool		= cpu_to_le32(pool_index),
+		},
+		.threshold_type	= cpu_to_le32(threshold_type),
+	};
+	unsigned int unit_size;
+
+	unit_size = nfp_shared_buf_pool_unit(pf, sb);
+	if (!unit_size || size % unit_size)
+		return -EINVAL;
+	set_data.size = cpu_to_le32(size / unit_size);
+
+	return nfp_mbox_cmd(pf, NFP_MBOX_POOL_SET, &set_data, sizeof(set_data),
+			    NULL, 0);
+}
+
+int nfp_shared_buf_register(struct nfp_pf *pf)
+{
+	struct devlink *devlink = priv_to_devlink(pf);
+	unsigned int i, num_entries, entry_sz;
+	struct nfp_cpp_area *sb_desc_area;
+	u8 __iomem *sb_desc;
+	int n, err;
+
+	if (!pf->mbox)
+		return 0;
+
+	n = nfp_pf_rtsym_read_optional(pf, NFP_SHARED_BUF_COUNT_SYM_NAME, 0);
+	if (n <= 0)
+		return n;
+	num_entries = n;
+
+	sb_desc = nfp_pf_map_rtsym(pf, "sb_tbl", NFP_SHARED_BUF_TABLE_SYM_NAME,
+				   num_entries * sizeof(pf->shared_bufs[0]),
+				   &sb_desc_area);
+	if (IS_ERR(sb_desc))
+		return PTR_ERR(sb_desc);
+
+	entry_sz = nfp_cpp_area_size(sb_desc_area) / num_entries;
+
+	pf->shared_bufs = kmalloc_array(num_entries, sizeof(pf->shared_bufs[0]),
+					GFP_KERNEL);
+	if (!pf->shared_bufs) {
+		err = -ENOMEM;
+		goto err_release_area;
+	}
+
+	for (i = 0; i < num_entries; i++) {
+		struct nfp_shared_buf *sb = &pf->shared_bufs[i];
+
+		/* Entries may be larger in future FW */
+		memcpy_fromio(sb, sb_desc + i * entry_sz, sizeof(*sb));
+
+		err = devlink_sb_register(devlink,
+					  le32_to_cpu(sb->id),
+					  le32_to_cpu(sb->size),
+					  le16_to_cpu(sb->ingress_pools_count),
+					  le16_to_cpu(sb->egress_pools_count),
+					  le16_to_cpu(sb->ingress_tc_count),
+					  le16_to_cpu(sb->egress_tc_count));
+		if (err)
+			goto err_unreg_prev;
+	}
+	pf->num_shared_bufs = num_entries;
+
+	nfp_cpp_area_release_free(sb_desc_area);
+
+	return 0;
+
+err_unreg_prev:
+	while (i--)
+		devlink_sb_unregister(devlink,
+				      le32_to_cpu(pf->shared_bufs[i].id));
+	kfree(pf->shared_bufs);
+err_release_area:
+	nfp_cpp_area_release_free(sb_desc_area);
+	return err;
+}
+
+void nfp_shared_buf_unregister(struct nfp_pf *pf)
+{
+	struct devlink *devlink = priv_to_devlink(pf);
+	unsigned int i;
+
+	for (i = 0; i < pf->num_shared_bufs; i++)
+		devlink_sb_unregister(devlink,
+				      le32_to_cpu(pf->shared_bufs[i].id));
+	kfree(pf->shared_bufs);
+}
-- 
2.17.0

[PATCH net-next 04/13] nfp: core: allow 4-byte aligned accesses to Memory Units

[Jakub Kicinski]

Current code doesn't enforce length requirements on 32bit accesses
with action NFP_CPP_ACTION_RW to memory units, but if the access
is only aligned to 4 bytes as well we will fall into the explicit
access case and error out.  Such accesses are correct, allow them
by lowering the width earlier.

While at it use a switch statement to improve readability.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
---
 .../netronome/nfp/nfpcore/nfp6000_pcie.c      | 94 +++++++++----------
 1 file changed, 44 insertions(+), 50 deletions(-)

diff --git a/drivers/net/ethernet/netronome/nfp/nfpcore/nfp6000_pcie.c b/drivers/net/ethernet/netronome/nfp/nfpcore/nfp6000_pcie.c
index a0e336bd1d85..749655c329b2 100644
--- a/drivers/net/ethernet/netronome/nfp/nfpcore/nfp6000_pcie.c
+++ b/drivers/net/ethernet/netronome/nfp/nfpcore/nfp6000_pcie.c
@@ -933,7 +933,6 @@ static int nfp6000_area_read(struct nfp_cpp_area *area, void *kernel_vaddr,
 	u32 *wrptr32 = kernel_vaddr;
 	const u32 __iomem *rdptr32;
 	int n, width;
-	bool is_64;
 
 	priv = nfp_cpp_area_priv(area);
 	rdptr64 = priv->iomem + offset;
@@ -943,10 +942,15 @@ static int nfp6000_area_read(struct nfp_cpp_area *area, void *kernel_vaddr,
 		return -EFAULT;
 
 	width = priv->width.read;
-
 	if (width <= 0)
 		return -EINVAL;
 
+	/* MU reads via a PCIe2CPP BAR support 32bit (and other) lengths */
+	if (priv->target == (NFP_CPP_TARGET_MU & NFP_CPP_TARGET_ID_MASK) &&
+	    priv->action == NFP_CPP_ACTION_RW &&
+	    (offset % sizeof(u64) == 4 || length % sizeof(u64) == 4))
+		width = TARGET_WIDTH_32;
+
 	/* Unaligned? Translate to an explicit access */
 	if ((priv->offset + offset) & (width - 1))
 		return nfp_cpp_explicit_read(nfp_cpp_area_cpp(area),
@@ -956,36 +960,29 @@ static int nfp6000_area_read(struct nfp_cpp_area *area, void *kernel_vaddr,
 					     priv->offset + offset,
 					     kernel_vaddr, length, width);
 
-	is_64 = width == TARGET_WIDTH_64;
-
-	/* MU reads via a PCIe2CPP BAR supports 32bit (and other) lengths */
-	if (priv->target == (NFP_CPP_TARGET_ID_MASK & NFP_CPP_TARGET_MU) &&
-	    priv->action == NFP_CPP_ACTION_RW)
-		is_64 = false;
+	if (WARN_ON(!priv->bar))
+		return -EFAULT;
 
-	if (is_64) {
-		if (offset % sizeof(u64) != 0 || length % sizeof(u64) != 0)
-			return -EINVAL;
-	} else {
+	switch (width) {
+	case TARGET_WIDTH_32:
 		if (offset % sizeof(u32) != 0 || length % sizeof(u32) != 0)
 			return -EINVAL;
-	}
 
-	if (WARN_ON(!priv->bar))
-		return -EFAULT;
+		for (n = 0; n < length; n += sizeof(u32))
+			*wrptr32++ = __raw_readl(rdptr32++);
+		return n;
+#ifdef __raw_readq
+	case TARGET_WIDTH_64:
+		if (offset % sizeof(u64) != 0 || length % sizeof(u64) != 0)
+			return -EINVAL;
 
-	if (is_64)
-#ifndef __raw_readq
-		return -EINVAL;
-#else
 		for (n = 0; n < length; n += sizeof(u64))
 			*wrptr64++ = __raw_readq(rdptr64++);
+		return n;
 #endif
-	else
-		for (n = 0; n < length; n += sizeof(u32))
-			*wrptr32++ = __raw_readl(rdptr32++);
-
-	return n;
+	default:
+		return -EINVAL;
+	}
 }
 
 static int
@@ -999,7 +996,6 @@ nfp6000_area_write(struct nfp_cpp_area *area,
 	struct nfp6000_area_priv *priv;
 	u32 __iomem *wrptr32;
 	int n, width;
-	bool is_64;
 
 	priv = nfp_cpp_area_priv(area);
 	wrptr64 = priv->iomem + offset;
@@ -1009,10 +1005,15 @@ nfp6000_area_write(struct nfp_cpp_area *area,
 		return -EFAULT;
 
 	width = priv->width.write;
-
 	if (width <= 0)
 		return -EINVAL;
 
+	/* MU writes via a PCIe2CPP BAR support 32bit (and other) lengths */
+	if (priv->target == (NFP_CPP_TARGET_ID_MASK & NFP_CPP_TARGET_MU) &&
+	    priv->action == NFP_CPP_ACTION_RW &&
+	    (offset % sizeof(u64) == 4 || length % sizeof(u64) == 4))
+		width = TARGET_WIDTH_32;
+
 	/* Unaligned? Translate to an explicit access */
 	if ((priv->offset + offset) & (width - 1))
 		return nfp_cpp_explicit_write(nfp_cpp_area_cpp(area),
@@ -1022,40 +1023,33 @@ nfp6000_area_write(struct nfp_cpp_area *area,
 					      priv->offset + offset,
 					      kernel_vaddr, length, width);
 
-	is_64 = width == TARGET_WIDTH_64;
-
-	/* MU writes via a PCIe2CPP BAR supports 32bit (and other) lengths */
-	if (priv->target == (NFP_CPP_TARGET_ID_MASK & NFP_CPP_TARGET_MU) &&
-	    priv->action == NFP_CPP_ACTION_RW)
-		is_64 = false;
+	if (WARN_ON(!priv->bar))
+		return -EFAULT;
 
-	if (is_64) {
-		if (offset % sizeof(u64) != 0 || length % sizeof(u64) != 0)
-			return -EINVAL;
-	} else {
+	switch (width) {
+	case TARGET_WIDTH_32:
 		if (offset % sizeof(u32) != 0 || length % sizeof(u32) != 0)
 			return -EINVAL;
-	}
 
-	if (WARN_ON(!priv->bar))
-		return -EFAULT;
+		for (n = 0; n < length; n += sizeof(u32)) {
+			__raw_writel(*rdptr32++, wrptr32++);
+			wmb();
+		}
+		return n;
+#ifdef __raw_writeq
+	case TARGET_WIDTH_64:
+		if (offset % sizeof(u64) != 0 || length % sizeof(u64) != 0)
+			return -EINVAL;
 
-	if (is_64)
-#ifndef __raw_writeq
-		return -EINVAL;
-#else
 		for (n = 0; n < length; n += sizeof(u64)) {
 			__raw_writeq(*rdptr64++, wrptr64++);
 			wmb();
 		}
+		return n;
 #endif
-	else
-		for (n = 0; n < length; n += sizeof(u32)) {
-			__raw_writel(*rdptr32++, wrptr32++);
-			wmb();
-		}
-
-	return n;
+	default:
+		return -EINVAL;
+	}
 }
 
 struct nfp6000_explicit_priv {
-- 
2.17.0

[PATCH net-next 05/13] nfp: abm: add initial active buffer management NIC skeleton

[Jakub Kicinski]

Add a very rudimentary active buffer management NIC support.
For now it's like a core NIC without SR-IOV support.  Next
commits will extend its functionality.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
---
 drivers/net/ethernet/netronome/Kconfig        | 13 +++
 drivers/net/ethernet/netronome/nfp/Makefile   |  5 ++
 drivers/net/ethernet/netronome/nfp/abm/main.c | 85 +++++++++++++++++++
 drivers/net/ethernet/netronome/nfp/abm/main.h | 47 ++++++++++
 drivers/net/ethernet/netronome/nfp/nfp_app.c  |  3 +
 drivers/net/ethernet/netronome/nfp/nfp_app.h  |  2 +
 6 files changed, 155 insertions(+)
 create mode 100644 drivers/net/ethernet/netronome/nfp/abm/main.c
 create mode 100644 drivers/net/ethernet/netronome/nfp/abm/main.h

diff --git a/drivers/net/ethernet/netronome/Kconfig b/drivers/net/ethernet/netronome/Kconfig
index ae0c46ba7546..66f15b05b65e 100644
--- a/drivers/net/ethernet/netronome/Kconfig
+++ b/drivers/net/ethernet/netronome/Kconfig
@@ -36,6 +36,19 @@ config NFP_APP_FLOWER
 	  either directly, with Open vSwitch, or any other way.  Note that
 	  TC Flower offload requires specific FW to work.
 
+config NFP_APP_ABM_NIC
+	bool "NFP4000/NFP6000 Advanced buffer management NIC support"
+	depends on NFP
+	depends on NET_SWITCHDEV
+	default y
+	help
+	  Enable driver support for Advanced buffer management NIC on NFP.
+	  ABM NIC allows advanced configuration of queuing and scheduling
+	  of packets, including ECN marking. Say Y, if you are planning to
+	  use one of the NFP4000 and NFP6000 platforms which support this
+	  functionality.
+	  Code will be built into the nfp.ko driver.
+
 config NFP_DEBUG
 	bool "Debug support for Netronome(R) NFP4000/NFP6000 NIC drivers"
 	depends on NFP
diff --git a/drivers/net/ethernet/netronome/nfp/Makefile b/drivers/net/ethernet/netronome/nfp/Makefile
index f032d645911c..3d79b709d77a 100644
--- a/drivers/net/ethernet/netronome/nfp/Makefile
+++ b/drivers/net/ethernet/netronome/nfp/Makefile
@@ -53,4 +53,9 @@ nfp-objs += \
 	    bpf/jit.o
 endif
 
+ifeq ($(CONFIG_NFP_APP_ABM_NIC),y)
+nfp-objs += \
+	    abm/main.o
+endif
+
 nfp-$(CONFIG_NFP_DEBUG) += nfp_net_debugfs.o
diff --git a/drivers/net/ethernet/netronome/nfp/abm/main.c b/drivers/net/ethernet/netronome/nfp/abm/main.c
new file mode 100644
index 000000000000..7a0a807bea4a
--- /dev/null
+++ b/drivers/net/ethernet/netronome/nfp/abm/main.c
@@ -0,0 +1,85 @@
+// SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
+/*
+ * Copyright (C) 2018 Netronome Systems, Inc.
+ *
+ * This software is dual licensed under the GNU General License Version 2,
+ * June 1991 as shown in the file COPYING in the top-level directory of this
+ * source tree or the BSD 2-Clause License provided below.  You have the
+ * option to license this software under the complete terms of either license.
+ *
+ * The BSD 2-Clause License:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      1. Redistributions of source code must retain the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer.
+ *
+ *      2. Redistributions in binary form must reproduce the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer in the documentation and/or other materials
+ *         provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include "../nfpcore/nfp_cpp.h"
+#include "../nfpcore/nfp_nsp.h"
+#include "../nfp_app.h"
+#include "../nfp_main.h"
+#include "main.h"
+
+static int nfp_abm_init(struct nfp_app *app)
+{
+	struct nfp_pf *pf = app->pf;
+	struct nfp_abm *abm;
+
+	if (!pf->eth_tbl) {
+		nfp_err(pf->cpp, "ABM NIC requires ETH table\n");
+		return -EINVAL;
+	}
+	if (pf->max_data_vnics != pf->eth_tbl->count) {
+		nfp_err(pf->cpp, "ETH entries don't match vNICs (%d vs %d)\n",
+			pf->max_data_vnics, pf->eth_tbl->count);
+		return -EINVAL;
+	}
+	if (!pf->mac_stats_bar) {
+		nfp_warn(app->cpp, "ABM NIC requires mac_stats symbol\n");
+		return -EINVAL;
+	}
+
+	abm = kzalloc(sizeof(*abm), GFP_KERNEL);
+	if (!abm)
+		return -ENOMEM;
+	app->priv = abm;
+	abm->app = app;
+
+	return 0;
+}
+
+static void nfp_abm_clean(struct nfp_app *app)
+{
+	struct nfp_abm *abm = app->priv;
+
+	kfree(abm);
+	app->priv = NULL;
+}
+
+const struct nfp_app_type app_abm = {
+	.id		= NFP_APP_ACTIVE_BUFFER_MGMT_NIC,
+	.name		= "abm",
+
+	.init		= nfp_abm_init,
+	.clean		= nfp_abm_clean,
+
+	.vnic_alloc	= nfp_app_nic_vnic_alloc,
+};
diff --git a/drivers/net/ethernet/netronome/nfp/abm/main.h b/drivers/net/ethernet/netronome/nfp/abm/main.h
new file mode 100644
index 000000000000..e54b6f64ee10
--- /dev/null
+++ b/drivers/net/ethernet/netronome/nfp/abm/main.h
@@ -0,0 +1,47 @@
+/* SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause) */
+/*
+ * Copyright (C) 2018 Netronome Systems, Inc.
+ *
+ * This software is dual licensed under the GNU General License Version 2,
+ * June 1991 as shown in the file COPYING in the top-level directory of this
+ * source tree or the BSD 2-Clause License provided below.  You have the
+ * option to license this software under the complete terms of either license.
+ *
+ * The BSD 2-Clause License:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      1. Redistributions of source code must retain the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer.
+ *
+ *      2. Redistributions in binary form must reproduce the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer in the documentation and/or other materials
+ *         provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __NFP_ABM_H__
+#define __NFP_ABM_H__ 1
+
+struct nfp_app;
+
+/**
+ * struct nfp_abm - ABM NIC app structure
+ * @app:	back pointer to nfp_app
+ */
+struct nfp_abm {
+	struct nfp_app *app;
+};
+#endif
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_app.c b/drivers/net/ethernet/netronome/nfp/nfp_app.c
index 0e0253c7e17b..c9d8a7ab311e 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_app.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_app.c
@@ -54,6 +54,9 @@ static const struct nfp_app_type *apps[] = {
 #ifdef CONFIG_NFP_APP_FLOWER
 	[NFP_APP_FLOWER_NIC]	= &app_flower,
 #endif
+#ifdef CONFIG_NFP_APP_ABM_NIC
+	[NFP_APP_ACTIVE_BUFFER_MGMT_NIC] = &app_abm,
+#endif
 };
 
 struct nfp_app *nfp_app_from_netdev(struct net_device *netdev)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_app.h b/drivers/net/ethernet/netronome/nfp/nfp_app.h
index 2d9cb2528fc7..654cabd95ee4 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_app.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_app.h
@@ -57,11 +57,13 @@ enum nfp_app_id {
 	NFP_APP_CORE_NIC	= 0x1,
 	NFP_APP_BPF_NIC		= 0x2,
 	NFP_APP_FLOWER_NIC	= 0x3,
+	NFP_APP_ACTIVE_BUFFER_MGMT_NIC = 0x4,
 };
 
 extern const struct nfp_app_type app_nic;
 extern const struct nfp_app_type app_bpf;
 extern const struct nfp_app_type app_flower;
+extern const struct nfp_app_type app_abm;
 
 /**
  * struct nfp_app_type - application definition
-- 
2.17.0

[PATCH net-next 06/13] nfp: abm: create project-specific vNIC structure

[Jakub Kicinski]

ABM NIC requires more complex vNIC handling, allocate
per-vNIC structure.  Find out RX queue base and PCI PF id.
There will be multiple PFs sharing the same MAC port, therefore
the MAC address assigned to the vNIC must be looked up in the
HWInfo database.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
---
 drivers/net/ethernet/netronome/nfp/Makefile   |   1 +
 drivers/net/ethernet/netronome/nfp/abm/ctrl.c |  58 ++++++++++
 drivers/net/ethernet/netronome/nfp/abm/main.c | 104 +++++++++++++++++-
 drivers/net/ethernet/netronome/nfp/abm/main.h |  20 ++++
 drivers/net/ethernet/netronome/nfp/nfp_app.h  |   2 +
 .../net/ethernet/netronome/nfp/nfp_app_nic.c  |   5 +-
 6 files changed, 186 insertions(+), 4 deletions(-)
 create mode 100644 drivers/net/ethernet/netronome/nfp/abm/ctrl.c

diff --git a/drivers/net/ethernet/netronome/nfp/Makefile b/drivers/net/ethernet/netronome/nfp/Makefile
index 3d79b709d77a..6373f56205fd 100644
--- a/drivers/net/ethernet/netronome/nfp/Makefile
+++ b/drivers/net/ethernet/netronome/nfp/Makefile
@@ -55,6 +55,7 @@ endif
 
 ifeq ($(CONFIG_NFP_APP_ABM_NIC),y)
 nfp-objs += \
+	    abm/ctrl.o \
 	    abm/main.o
 endif
 
diff --git a/drivers/net/ethernet/netronome/nfp/abm/ctrl.c b/drivers/net/ethernet/netronome/nfp/abm/ctrl.c
new file mode 100644
index 000000000000..e40f6f06417b
--- /dev/null
+++ b/drivers/net/ethernet/netronome/nfp/abm/ctrl.c
@@ -0,0 +1,58 @@
+// SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
+/*
+ * Copyright (C) 2018 Netronome Systems, Inc.
+ *
+ * This software is dual licensed under the GNU General License Version 2,
+ * June 1991 as shown in the file COPYING in the top-level directory of this
+ * source tree or the BSD 2-Clause License provided below.  You have the
+ * option to license this software under the complete terms of either license.
+ *
+ * The BSD 2-Clause License:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      1. Redistributions of source code must retain the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer.
+ *
+ *      2. Redistributions in binary form must reproduce the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer in the documentation and/or other materials
+ *         provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/kernel.h>
+
+#include "../nfpcore/nfp_cpp.h"
+#include "../nfp_app.h"
+#include "../nfp_main.h"
+#include "../nfp_net.h"
+#include "main.h"
+
+void nfp_abm_ctrl_read_params(struct nfp_abm_link *alink)
+{
+	alink->queue_base = nn_readl(alink->vnic, NFP_NET_CFG_START_RXQ);
+	alink->queue_base /= alink->vnic->stride_rx;
+}
+
+int nfp_abm_ctrl_find_addrs(struct nfp_abm *abm)
+{
+	struct nfp_pf *pf = abm->app->pf;
+	unsigned int pf_id;
+
+	pf_id =	nfp_cppcore_pcie_unit(pf->cpp);
+	abm->pf_id = pf_id;
+
+	return 0;
+}
diff --git a/drivers/net/ethernet/netronome/nfp/abm/main.c b/drivers/net/ethernet/netronome/nfp/abm/main.c
index 7a0a807bea4a..1985e58ab0db 100644
--- a/drivers/net/ethernet/netronome/nfp/abm/main.c
+++ b/drivers/net/ethernet/netronome/nfp/abm/main.c
@@ -32,16 +32,108 @@
  * SOFTWARE.
  */
 
+#include <linux/etherdevice.h>
+
+#include "../nfpcore/nfp.h"
 #include "../nfpcore/nfp_cpp.h"
 #include "../nfpcore/nfp_nsp.h"
 #include "../nfp_app.h"
 #include "../nfp_main.h"
+#include "../nfp_net.h"
+#include "../nfp_port.h"
 #include "main.h"
 
+static void
+nfp_abm_vnic_set_mac(struct nfp_pf *pf, struct nfp_abm *abm, struct nfp_net *nn,
+		     unsigned int id)
+{
+	struct nfp_eth_table_port *eth_port = &pf->eth_tbl->ports[id];
+	u8 mac_addr[ETH_ALEN];
+	const char *mac_str;
+	char name[32];
+
+	if (id > pf->eth_tbl->count) {
+		nfp_warn(pf->cpp, "No entry for persistent MAC address\n");
+		eth_hw_addr_random(nn->dp.netdev);
+		return;
+	}
+
+	snprintf(name, sizeof(name), "eth%u.mac.pf%u",
+		 eth_port->eth_index, abm->pf_id);
+
+	mac_str = nfp_hwinfo_lookup(pf->hwinfo, name);
+	if (!mac_str) {
+		nfp_warn(pf->cpp, "Can't lookup persistent MAC address (%s)\n",
+			 name);
+		eth_hw_addr_random(nn->dp.netdev);
+		return;
+	}
+
+	if (sscanf(mac_str, "%02hhx:%02hhx:%02hhx:%02hhx:%02hhx:%02hhx",
+		   &mac_addr[0], &mac_addr[1], &mac_addr[2],
+		   &mac_addr[3], &mac_addr[4], &mac_addr[5]) != 6) {
+		nfp_warn(pf->cpp, "Can't parse persistent MAC address (%s)\n",
+			 mac_str);
+		eth_hw_addr_random(nn->dp.netdev);
+		return;
+	}
+
+	ether_addr_copy(nn->dp.netdev->dev_addr, mac_addr);
+	ether_addr_copy(nn->dp.netdev->perm_addr, mac_addr);
+}
+
+static int
+nfp_abm_vnic_alloc(struct nfp_app *app, struct nfp_net *nn, unsigned int id)
+{
+	struct nfp_abm *abm = app->priv;
+	struct nfp_abm_link *alink;
+	int err;
+
+	alink = kzalloc(sizeof(*alink), GFP_KERNEL);
+	if (!alink)
+		return -ENOMEM;
+	nn->app_priv = alink;
+	alink->abm = abm;
+	alink->vnic = nn;
+	alink->id = id;
+
+	nn->port = nfp_port_alloc(app, NFP_PORT_PHYS_PORT, nn->dp.netdev);
+	if (IS_ERR(nn->port)) {
+		err = PTR_ERR(nn->port);
+		goto err_free_alink;
+	}
+
+	err = nfp_app_nic_vnic_init_phy_port(app->pf, app, nn, id);
+	if (err < 0)
+		goto err_free_port;
+	if (nn->port->type == NFP_PORT_INVALID)
+		/* core will kill this vNIC */
+		return 0;
+
+	nfp_abm_vnic_set_mac(app->pf, abm, nn, id);
+	nfp_abm_ctrl_read_params(alink);
+
+	return 0;
+
+err_free_port:
+	nfp_port_free(nn->port);
+err_free_alink:
+	kfree(alink);
+	return err;
+}
+
+static void nfp_abm_vnic_free(struct nfp_app *app, struct nfp_net *nn)
+{
+	struct nfp_abm_link *alink = nn->app_priv;
+
+	kfree(alink);
+}
+
 static int nfp_abm_init(struct nfp_app *app)
 {
 	struct nfp_pf *pf = app->pf;
 	struct nfp_abm *abm;
+	int err;
 
 	if (!pf->eth_tbl) {
 		nfp_err(pf->cpp, "ABM NIC requires ETH table\n");
@@ -63,7 +155,16 @@ static int nfp_abm_init(struct nfp_app *app)
 	app->priv = abm;
 	abm->app = app;
 
+	err = nfp_abm_ctrl_find_addrs(abm);
+	if (err)
+		goto err_free_abm;
+
 	return 0;
+
+err_free_abm:
+	kfree(abm);
+	app->priv = NULL;
+	return err;
 }
 
 static void nfp_abm_clean(struct nfp_app *app)
@@ -81,5 +182,6 @@ const struct nfp_app_type app_abm = {
 	.init		= nfp_abm_init,
 	.clean		= nfp_abm_clean,
 
-	.vnic_alloc	= nfp_app_nic_vnic_alloc,
+	.vnic_alloc	= nfp_abm_vnic_alloc,
+	.vnic_free	= nfp_abm_vnic_free,
 };
diff --git a/drivers/net/ethernet/netronome/nfp/abm/main.h b/drivers/net/ethernet/netronome/nfp/abm/main.h
index e54b6f64ee10..3c5a01c96ecd 100644
--- a/drivers/net/ethernet/netronome/nfp/abm/main.h
+++ b/drivers/net/ethernet/netronome/nfp/abm/main.h
@@ -36,12 +36,32 @@
 #define __NFP_ABM_H__ 1
 
 struct nfp_app;
+struct nfp_net;
 
 /**
  * struct nfp_abm - ABM NIC app structure
  * @app:	back pointer to nfp_app
+ * @pf_id:	ID of our PF link
  */
 struct nfp_abm {
 	struct nfp_app *app;
+	unsigned int pf_id;
 };
+
+/**
+ * struct nfp_abm_link - port tuple of a ABM NIC
+ * @abm:	back pointer to nfp_abm
+ * @vnic:	data vNIC
+ * @id:		id of the data vNIC
+ * @queue_base:	id of base to host queue within PCIe (not QC idx)
+ */
+struct nfp_abm_link {
+	struct nfp_abm *abm;
+	struct nfp_net *vnic;
+	unsigned int id;
+	unsigned int queue_base;
+};
+
+void nfp_abm_ctrl_read_params(struct nfp_abm_link *alink);
+int nfp_abm_ctrl_find_addrs(struct nfp_abm *abm);
 #endif
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_app.h b/drivers/net/ethernet/netronome/nfp/nfp_app.h
index 654cabd95ee4..fdf2593ae151 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_app.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_app.h
@@ -412,5 +412,7 @@ void nfp_app_free(struct nfp_app *app);
 
 int nfp_app_nic_vnic_alloc(struct nfp_app *app, struct nfp_net *nn,
 			   unsigned int id);
+int nfp_app_nic_vnic_init_phy_port(struct nfp_pf *pf, struct nfp_app *app,
+				   struct nfp_net *nn, unsigned int id);
 
 #endif
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_app_nic.c b/drivers/net/ethernet/netronome/nfp/nfp_app_nic.c
index b9618c37403f..e2dfe4f168bb 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_app_nic.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_app_nic.c
@@ -38,9 +38,8 @@
 #include "nfp_net.h"
 #include "nfp_port.h"
 
-static int
-nfp_app_nic_vnic_init_phy_port(struct nfp_pf *pf, struct nfp_app *app,
-			       struct nfp_net *nn, unsigned int id)
+int nfp_app_nic_vnic_init_phy_port(struct nfp_pf *pf, struct nfp_app *app,
+				   struct nfp_net *nn, unsigned int id)
 {
 	int err;
 
-- 
2.17.0

[PATCH net-next 07/13] nfp: add app pointer to port representors

[Jakub Kicinski]

nfp_apps can currently associate their structures with vNICs but
not representors.  Add app priv pointer to representors as well.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
---
 drivers/net/ethernet/netronome/nfp/nfp_net_repr.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_repr.h b/drivers/net/ethernet/netronome/nfp/nfp_net_repr.h
index cd756a15445f..8dca283c05c3 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_repr.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_repr.h
@@ -76,6 +76,7 @@ struct nfp_repr_pcpu_stats {
  * @port:	Port of representor
  * @app:	APP handle
  * @stats:	Statistic of packets hitting CPU
+ * @app_priv:	Pointer for APP data
  */
 struct nfp_repr {
 	struct net_device *netdev;
@@ -83,6 +84,7 @@ struct nfp_repr {
 	struct nfp_port *port;
 	struct nfp_app *app;
 	struct nfp_repr_pcpu_stats __percpu *stats;
+	void *app_priv;
 };
 
 /**
-- 
2.17.0

[PATCH net-next 08/13] devlink: don't take instance lock around eswitch mode set

[Jakub Kicinski]

Changing switch mode may want to register and unregister devlink
ports.  Therefore similarly to DEVLINK_CMD_PORT_SPLIT/UNSPLIT it
should not take the instance lock.  Drivers don't depend on existing
locking since it's a very recent addition.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
CC: jiri@resnulli.us

 net/core/devlink.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/net/core/devlink.c b/net/core/devlink.c
index 5c8a40e1a01e..475246b355f0 100644
--- a/net/core/devlink.c
+++ b/net/core/devlink.c
@@ -2756,7 +2756,8 @@ static const struct genl_ops devlink_nl_ops[] = {
 		.doit = devlink_nl_cmd_eswitch_set_doit,
 		.policy = devlink_nl_policy,
 		.flags = GENL_ADMIN_PERM,
-		.internal_flags = DEVLINK_NL_FLAG_NEED_DEVLINK,
+		.internal_flags = DEVLINK_NL_FLAG_NEED_DEVLINK |
+				  DEVLINK_NL_FLAG_NO_LOCK,
 	},
 	{
 		.cmd = DEVLINK_CMD_DPIPE_TABLE_GET,
-- 
2.17.0

[PATCH net-next 09/13] nfp: add devlink_eswitch_mode_set callback

[Jakub Kicinski]

Our previous apps all assumed to use only one eswitch mode (legacy
or switchdev) without the ability to change it.  ABM NIC will
want to support the switch so plumb devlink_eswitch_mode_set through.
The devlink_eswitch_mode_set is expected to spawn representors and
potentially devlink ports so it's called under big devlink lock and
pf->lock.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
---
 drivers/net/ethernet/netronome/nfp/nfp_app.h     |  9 +++++++++
 drivers/net/ethernet/netronome/nfp/nfp_devlink.c | 13 +++++++++++++
 2 files changed, 22 insertions(+)

diff --git a/drivers/net/ethernet/netronome/nfp/nfp_app.h b/drivers/net/ethernet/netronome/nfp/nfp_app.h
index fdf2593ae151..23b99a4e05c2 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_app.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_app.h
@@ -97,6 +97,7 @@ extern const struct nfp_app_type app_abm;
  * @bpf:	BPF ndo offload-related calls
  * @xdp_offload:    offload an XDP program
  * @eswitch_mode_get:    get SR-IOV eswitch mode
+ * @eswitch_mode_set:    set SR-IOV eswitch mode (under pf->lock)
  * @sriov_enable: app-specific sriov initialisation
  * @sriov_disable: app-specific sriov clean-up
  * @repr_get:	get representor netdev
@@ -148,6 +149,7 @@ struct nfp_app_type {
 	void (*sriov_disable)(struct nfp_app *app);
 
 	enum devlink_eswitch_mode (*eswitch_mode_get)(struct nfp_app *app);
+	int (*eswitch_mode_set)(struct nfp_app *app, u16 mode);
 	struct net_device *(*repr_get)(struct nfp_app *app, u32 id);
 };
 
@@ -372,6 +374,13 @@ static inline int nfp_app_eswitch_mode_get(struct nfp_app *app, u16 *mode)
 	return 0;
 }
 
+static inline int nfp_app_eswitch_mode_set(struct nfp_app *app, u16 mode)
+{
+	if (!app->type->eswitch_mode_set)
+		return -EOPNOTSUPP;
+	return app->type->eswitch_mode_set(app, mode);
+}
+
 static inline int nfp_app_sriov_enable(struct nfp_app *app, int num_vfs)
 {
 	if (!app || !app->type->sriov_enable)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_devlink.c b/drivers/net/ethernet/netronome/nfp/nfp_devlink.c
index 73c7fcc820ac..71c2edd83031 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_devlink.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_devlink.c
@@ -176,12 +176,25 @@ static int nfp_devlink_eswitch_mode_get(struct devlink *devlink, u16 *mode)
 	return nfp_app_eswitch_mode_get(pf->app, mode);
 }
 
+static int nfp_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode)
+{
+	struct nfp_pf *pf = devlink_priv(devlink);
+	int ret;
+
+	mutex_lock(&pf->lock);
+	ret = nfp_app_eswitch_mode_set(pf->app, mode);
+	mutex_unlock(&pf->lock);
+
+	return ret;
+}
+
 const struct devlink_ops nfp_devlink_ops = {
 	.port_split		= nfp_devlink_port_split,
 	.port_unsplit		= nfp_devlink_port_unsplit,
 	.sb_pool_get		= nfp_devlink_sb_pool_get,
 	.sb_pool_set		= nfp_devlink_sb_pool_set,
 	.eswitch_mode_get	= nfp_devlink_eswitch_mode_get,
+	.eswitch_mode_set	= nfp_devlink_eswitch_mode_set,
 };
 
 int nfp_devlink_port_register(struct nfp_app *app, struct nfp_port *port)
-- 
2.17.0

[PATCH net-next 10/13] nfp: abm: spawn port netdevs

[Jakub Kicinski]

To configure buffering points we need full set of netdevs:

                              ASIC

 user netdev  -- | -- PCIe port   MAC port -- | --

Configuring egrees qdiscs on user netdev configures standard
Linux TC software qdiscs, configuring PCIe port qdiscs will
provide a way of setting ASIC queuing parameters for PCIe block.
MAC port netdev egress qdiscs correspond to ASIC MAC Traffic
Manager block.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
---
 drivers/net/ethernet/netronome/nfp/abm/main.c | 234 ++++++++++++++++--
 drivers/net/ethernet/netronome/nfp/abm/main.h |   8 +
 .../net/ethernet/netronome/nfp/nfp_net_repr.c |   2 +-
 .../net/ethernet/netronome/nfp/nfp_net_repr.h |   1 +
 4 files changed, 225 insertions(+), 20 deletions(-)

diff --git a/drivers/net/ethernet/netronome/nfp/abm/main.c b/drivers/net/ethernet/netronome/nfp/abm/main.c
index 1985e58ab0db..d8f1bdac02b8 100644
--- a/drivers/net/ethernet/netronome/nfp/abm/main.c
+++ b/drivers/net/ethernet/netronome/nfp/abm/main.c
@@ -32,7 +32,12 @@
  * SOFTWARE.
  */
 
+#include <linux/bitfield.h>
 #include <linux/etherdevice.h>
+#include <linux/lockdep.h>
+#include <linux/netdevice.h>
+#include <linux/rcupdate.h>
+#include <linux/slab.h>
 
 #include "../nfpcore/nfp.h"
 #include "../nfpcore/nfp_cpp.h"
@@ -40,9 +45,195 @@
 #include "../nfp_app.h"
 #include "../nfp_main.h"
 #include "../nfp_net.h"
+#include "../nfp_net_repr.h"
 #include "../nfp_port.h"
 #include "main.h"
 
+static u32 nfp_abm_portid(enum nfp_repr_type rtype, unsigned int id)
+{
+	return FIELD_PREP(NFP_ABM_PORTID_TYPE, rtype) |
+	       FIELD_PREP(NFP_ABM_PORTID_ID, id);
+}
+
+static struct net_device *nfp_abm_repr_get(struct nfp_app *app, u32 port_id)
+{
+	enum nfp_repr_type rtype;
+	struct nfp_reprs *reprs;
+	u8 port;
+
+	rtype = FIELD_GET(NFP_ABM_PORTID_TYPE, port_id);
+	port = FIELD_GET(NFP_ABM_PORTID_ID, port_id);
+
+	reprs = rcu_dereference(app->reprs[rtype]);
+	if (!reprs)
+		return NULL;
+
+	if (port >= reprs->num_reprs)
+		return NULL;
+
+	return rcu_dereference(reprs->reprs[port]);
+}
+
+static int
+nfp_abm_spawn_repr(struct nfp_app *app, struct nfp_abm_link *alink,
+		   enum nfp_port_type ptype)
+{
+	struct net_device *netdev;
+	enum nfp_repr_type rtype;
+	struct nfp_reprs *reprs;
+	struct nfp_repr *repr;
+	struct nfp_port *port;
+	int err;
+
+	if (ptype == NFP_PORT_PHYS_PORT)
+		rtype = NFP_REPR_TYPE_PHYS_PORT;
+	else
+		rtype = NFP_REPR_TYPE_PF;
+
+	netdev = nfp_repr_alloc(app);
+	if (!netdev)
+		return -ENOMEM;
+	repr = netdev_priv(netdev);
+	repr->app_priv = alink;
+
+	port = nfp_port_alloc(app, ptype, netdev);
+	if (IS_ERR(port)) {
+		err = PTR_ERR(port);
+		goto err_free_repr;
+	}
+
+	if (ptype == NFP_PORT_PHYS_PORT) {
+		err = nfp_port_init_phy_port(app->pf, app, port, alink->id);
+		if (err)
+			goto err_free_port;
+	} else {
+		port->pf_id = alink->abm->pf_id;
+		port->vnic = alink->vnic->dp.ctrl_bar;
+	}
+
+	SET_NETDEV_DEV(netdev, &alink->vnic->pdev->dev);
+	eth_hw_addr_random(netdev);
+
+	err = nfp_repr_init(app, netdev, nfp_abm_portid(rtype, alink->id),
+			    port, alink->vnic->dp.netdev);
+	if (err)
+		goto err_free_port;
+
+	reprs = nfp_reprs_get_locked(app, rtype);
+	WARN(nfp_repr_get_locked(app, reprs, alink->id), "duplicate repr");
+	rcu_assign_pointer(reprs->reprs[alink->id], netdev);
+
+	nfp_info(app->cpp, "%s Port %d Representor(%s) created\n",
+		 ptype == NFP_PORT_PF_PORT ? "PCIe" : "Phys",
+		 alink->id, netdev->name);
+
+	return 0;
+
+err_free_port:
+	nfp_port_free(port);
+err_free_repr:
+	nfp_repr_free(netdev);
+	return err;
+}
+
+static void
+nfp_abm_kill_repr(struct nfp_app *app, struct nfp_abm_link *alink,
+		  enum nfp_repr_type rtype)
+{
+	struct net_device *netdev;
+	struct nfp_reprs *reprs;
+
+	reprs = nfp_reprs_get_locked(app, rtype);
+	netdev = nfp_repr_get_locked(app, reprs, alink->id);
+	if (!netdev)
+		return;
+	rcu_assign_pointer(reprs->reprs[alink->id], NULL);
+	synchronize_rcu();
+	/* Cast to make sure nfp_repr_clean_and_free() takes a nfp_repr */
+	nfp_repr_clean_and_free((struct nfp_repr *)netdev_priv(netdev));
+}
+
+static void
+nfp_abm_kill_reprs(struct nfp_abm *abm, struct nfp_abm_link *alink)
+{
+	nfp_abm_kill_repr(abm->app, alink, NFP_REPR_TYPE_PF);
+	nfp_abm_kill_repr(abm->app, alink, NFP_REPR_TYPE_PHYS_PORT);
+}
+
+static void nfp_abm_kill_reprs_all(struct nfp_abm *abm)
+{
+	struct nfp_pf *pf = abm->app->pf;
+	struct nfp_net *nn;
+
+	list_for_each_entry(nn, &pf->vnics, vnic_list)
+		nfp_abm_kill_reprs(abm, (struct nfp_abm_link *)nn->app_priv);
+}
+
+static enum devlink_eswitch_mode nfp_abm_eswitch_mode_get(struct nfp_app *app)
+{
+	struct nfp_abm *abm = app->priv;
+
+	return abm->eswitch_mode;
+}
+
+static int nfp_abm_eswitch_set_legacy(struct nfp_abm *abm)
+{
+	nfp_abm_kill_reprs_all(abm);
+
+	abm->eswitch_mode = DEVLINK_ESWITCH_MODE_LEGACY;
+	return 0;
+}
+
+static void nfp_abm_eswitch_clean_up(struct nfp_abm *abm)
+{
+	if (abm->eswitch_mode != DEVLINK_ESWITCH_MODE_LEGACY)
+		WARN_ON(nfp_abm_eswitch_set_legacy(abm));
+}
+
+static int nfp_abm_eswitch_set_switchdev(struct nfp_abm *abm)
+{
+	struct nfp_app *app = abm->app;
+	struct nfp_pf *pf = app->pf;
+	struct nfp_net *nn;
+	int err;
+
+	list_for_each_entry(nn, &pf->vnics, vnic_list) {
+		struct nfp_abm_link *alink = nn->app_priv;
+
+		err = nfp_abm_spawn_repr(app, alink, NFP_PORT_PHYS_PORT);
+		if (err)
+			goto err_kill_all_reprs;
+
+		err = nfp_abm_spawn_repr(app, alink, NFP_PORT_PF_PORT);
+		if (err)
+			goto err_kill_all_reprs;
+	}
+
+	abm->eswitch_mode = DEVLINK_ESWITCH_MODE_SWITCHDEV;
+	return 0;
+
+err_kill_all_reprs:
+	nfp_abm_kill_reprs_all(abm);
+	return err;
+}
+
+static int nfp_abm_eswitch_mode_set(struct nfp_app *app, u16 mode)
+{
+	struct nfp_abm *abm = app->priv;
+
+	if (abm->eswitch_mode == mode)
+		return 0;
+
+	switch (mode) {
+	case DEVLINK_ESWITCH_MODE_LEGACY:
+		return nfp_abm_eswitch_set_legacy(abm);
+	case DEVLINK_ESWITCH_MODE_SWITCHDEV:
+		return nfp_abm_eswitch_set_switchdev(abm);
+	default:
+		return -EINVAL;
+	}
+}
+
 static void
 nfp_abm_vnic_set_mac(struct nfp_pf *pf, struct nfp_abm *abm, struct nfp_net *nn,
 		     unsigned int id)
@@ -87,7 +278,6 @@ nfp_abm_vnic_alloc(struct nfp_app *app, struct nfp_net *nn, unsigned int id)
 {
 	struct nfp_abm *abm = app->priv;
 	struct nfp_abm_link *alink;
-	int err;
 
 	alink = kzalloc(sizeof(*alink), GFP_KERNEL);
 	if (!alink)
@@ -97,41 +287,26 @@ nfp_abm_vnic_alloc(struct nfp_app *app, struct nfp_net *nn, unsigned int id)
 	alink->vnic = nn;
 	alink->id = id;
 
-	nn->port = nfp_port_alloc(app, NFP_PORT_PHYS_PORT, nn->dp.netdev);
-	if (IS_ERR(nn->port)) {
-		err = PTR_ERR(nn->port);
-		goto err_free_alink;
-	}
-
-	err = nfp_app_nic_vnic_init_phy_port(app->pf, app, nn, id);
-	if (err < 0)
-		goto err_free_port;
-	if (nn->port->type == NFP_PORT_INVALID)
-		/* core will kill this vNIC */
-		return 0;
+	netif_keep_dst(nn->dp.netdev);
 
 	nfp_abm_vnic_set_mac(app->pf, abm, nn, id);
 	nfp_abm_ctrl_read_params(alink);
 
 	return 0;
-
-err_free_port:
-	nfp_port_free(nn->port);
-err_free_alink:
-	kfree(alink);
-	return err;
 }
 
 static void nfp_abm_vnic_free(struct nfp_app *app, struct nfp_net *nn)
 {
 	struct nfp_abm_link *alink = nn->app_priv;
 
+	nfp_abm_kill_reprs(alink->abm, alink);
 	kfree(alink);
 }
 
 static int nfp_abm_init(struct nfp_app *app)
 {
 	struct nfp_pf *pf = app->pf;
+	struct nfp_reprs *reprs;
 	struct nfp_abm *abm;
 	int err;
 
@@ -159,8 +334,21 @@ static int nfp_abm_init(struct nfp_app *app)
 	if (err)
 		goto err_free_abm;
 
+	err = -ENOMEM;
+	reprs = nfp_reprs_alloc(pf->max_data_vnics);
+	if (!reprs)
+		goto err_free_abm;
+	RCU_INIT_POINTER(app->reprs[NFP_REPR_TYPE_PHYS_PORT], reprs);
+
+	reprs = nfp_reprs_alloc(pf->max_data_vnics);
+	if (!reprs)
+		goto err_free_phys;
+	RCU_INIT_POINTER(app->reprs[NFP_REPR_TYPE_PF], reprs);
+
 	return 0;
 
+err_free_phys:
+	nfp_reprs_clean_and_free_by_type(app, NFP_REPR_TYPE_PHYS_PORT);
 err_free_abm:
 	kfree(abm);
 	app->priv = NULL;
@@ -171,6 +359,9 @@ static void nfp_abm_clean(struct nfp_app *app)
 {
 	struct nfp_abm *abm = app->priv;
 
+	nfp_abm_eswitch_clean_up(abm);
+	nfp_reprs_clean_and_free_by_type(app, NFP_REPR_TYPE_PF);
+	nfp_reprs_clean_and_free_by_type(app, NFP_REPR_TYPE_PHYS_PORT);
 	kfree(abm);
 	app->priv = NULL;
 }
@@ -184,4 +375,9 @@ const struct nfp_app_type app_abm = {
 
 	.vnic_alloc	= nfp_abm_vnic_alloc,
 	.vnic_free	= nfp_abm_vnic_free,
+
+	.eswitch_mode_get	= nfp_abm_eswitch_mode_get,
+	.eswitch_mode_set	= nfp_abm_eswitch_mode_set,
+
+	.repr_get	= nfp_abm_repr_get,
 };
diff --git a/drivers/net/ethernet/netronome/nfp/abm/main.h b/drivers/net/ethernet/netronome/nfp/abm/main.h
index 3c5a01c96ecd..5938b69b8a84 100644
--- a/drivers/net/ethernet/netronome/nfp/abm/main.h
+++ b/drivers/net/ethernet/netronome/nfp/abm/main.h
@@ -35,17 +35,25 @@
 #ifndef __NFP_ABM_H__
 #define __NFP_ABM_H__ 1
 
+#include <net/devlink.h>
+
 struct nfp_app;
 struct nfp_net;
 
+#define NFP_ABM_PORTID_TYPE	GENMASK(23, 16)
+#define NFP_ABM_PORTID_ID	GENMASK(7, 0)
+
 /**
  * struct nfp_abm - ABM NIC app structure
  * @app:	back pointer to nfp_app
  * @pf_id:	ID of our PF link
+ * @eswitch_mode:	devlink eswitch mode, advanced functions only visible
+ *			in switchdev mode
  */
 struct nfp_abm {
 	struct nfp_app *app;
 	unsigned int pf_id;
+	enum devlink_eswitch_mode eswitch_mode;
 };
 
 /**
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c b/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c
index 6e79da91e475..09e87d5f4f72 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c
@@ -385,7 +385,7 @@ struct net_device *nfp_repr_alloc(struct nfp_app *app)
 	return NULL;
 }
 
-static void nfp_repr_clean_and_free(struct nfp_repr *repr)
+void nfp_repr_clean_and_free(struct nfp_repr *repr)
 {
 	nfp_info(repr->app->cpp, "Destroying Representor(%s)\n",
 		 repr->netdev->name);
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_repr.h b/drivers/net/ethernet/netronome/nfp/nfp_net_repr.h
index 8dca283c05c3..8366e4f3c623 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_repr.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_repr.h
@@ -127,6 +127,7 @@ int nfp_repr_init(struct nfp_app *app, struct net_device *netdev,
 		  struct net_device *pf_netdev);
 void nfp_repr_free(struct net_device *netdev);
 struct net_device *nfp_repr_alloc(struct nfp_app *app);
+void nfp_repr_clean_and_free(struct nfp_repr *repr);
 void nfp_reprs_clean_and_free(struct nfp_app *app, struct nfp_reprs *reprs);
 void nfp_reprs_clean_and_free_by_type(struct nfp_app *app,
 				      enum nfp_repr_type type);
-- 
2.17.0

[PATCH net-next 11/13] nfp: abm: force Ethternet port up

[Jakub Kicinski]

Current control firmware does not cater too well to multi-host
applications.  There is no way to check which hosts are up or
otherwise negotiate what the state of the external port (the
Ethernet port) should be.  Make sure the link is up when driver
loads, and don't take it down when Ethernet port netdev is
closed.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
---
 drivers/net/ethernet/netronome/nfp/abm/main.c | 14 ++++++++++++++
 drivers/net/ethernet/netronome/nfp/nfp_port.c |  2 ++
 drivers/net/ethernet/netronome/nfp/nfp_port.h |  2 ++
 3 files changed, 18 insertions(+)

diff --git a/drivers/net/ethernet/netronome/nfp/abm/main.c b/drivers/net/ethernet/netronome/nfp/abm/main.c
index d8f1bdac02b8..7afd24ce79a5 100644
--- a/drivers/net/ethernet/netronome/nfp/abm/main.c
+++ b/drivers/net/ethernet/netronome/nfp/abm/main.c
@@ -103,6 +103,7 @@ nfp_abm_spawn_repr(struct nfp_app *app, struct nfp_abm_link *alink,
 	}
 
 	if (ptype == NFP_PORT_PHYS_PORT) {
+		port->eth_forced = true;
 		err = nfp_port_init_phy_port(app->pf, app, port, alink->id);
 		if (err)
 			goto err_free_port;
@@ -276,8 +277,10 @@ nfp_abm_vnic_set_mac(struct nfp_pf *pf, struct nfp_abm *abm, struct nfp_net *nn,
 static int
 nfp_abm_vnic_alloc(struct nfp_app *app, struct nfp_net *nn, unsigned int id)
 {
+	struct nfp_eth_table_port *eth_port = &app->pf->eth_tbl->ports[id];
 	struct nfp_abm *abm = app->priv;
 	struct nfp_abm_link *alink;
+	int err;
 
 	alink = kzalloc(sizeof(*alink), GFP_KERNEL);
 	if (!alink)
@@ -287,12 +290,23 @@ nfp_abm_vnic_alloc(struct nfp_app *app, struct nfp_net *nn, unsigned int id)
 	alink->vnic = nn;
 	alink->id = id;
 
+	/* This is a multi-host app, make sure MAC/PHY is up, but don't
+	 * make the MAC/PHY state follow the state of any of the ports.
+	 */
+	err = nfp_eth_set_configured(app->cpp, eth_port->index, true);
+	if (err < 0)
+		goto err_free_alink;
+
 	netif_keep_dst(nn->dp.netdev);
 
 	nfp_abm_vnic_set_mac(app->pf, abm, nn, id);
 	nfp_abm_ctrl_read_params(alink);
 
 	return 0;
+
+err_free_alink:
+	kfree(alink);
+	return err;
 }
 
 static void nfp_abm_vnic_free(struct nfp_app *app, struct nfp_net *nn)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_port.c b/drivers/net/ethernet/netronome/nfp/nfp_port.c
index 7bd8be5c833b..a17f1ace6988 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_port.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_port.c
@@ -218,6 +218,8 @@ int nfp_port_configure(struct net_device *netdev, bool configed)
 	eth_port = __nfp_port_get_eth_port(port);
 	if (!eth_port)
 		return 0;
+	if (port->eth_forced)
+		return 0;
 
 	err = nfp_eth_set_configured(port->app->cpp, eth_port->index, configed);
 	return err < 0 && err != -EOPNOTSUPP ? err : 0;
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_port.h b/drivers/net/ethernet/netronome/nfp/nfp_port.h
index fa7e669a969c..1d9b2b47e27d 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_port.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_port.h
@@ -77,6 +77,7 @@ enum nfp_port_flags {
  * @app:	backpointer to the app structure
  * @dl_port:	devlink port structure
  * @eth_id:	for %NFP_PORT_PHYS_PORT port ID in NFP enumeration scheme
+ * @eth_forced:	for %NFP_PORT_PHYS_PORT port is forced UP or DOWN, don't change
  * @eth_port:	for %NFP_PORT_PHYS_PORT translated ETH Table port entry
  * @eth_stats:	for %NFP_PORT_PHYS_PORT MAC stats if available
  * @pf_id:	for %NFP_PORT_PF_PORT, %NFP_PORT_VF_PORT ID of the PCI PF (0-3)
@@ -99,6 +100,7 @@ struct nfp_port {
 		/* NFP_PORT_PHYS_PORT */
 		struct {
 			unsigned int eth_id;
+			bool eth_forced;
 			struct nfp_eth_table_port *eth_port;
 			u8 __iomem *eth_stats;
 		};
-- 
2.17.0

[PATCH net-next 12/13] nfp: use split in naming of PCIe PF ports

[Jakub Kicinski]

PCI PFs can host more than one logical endpoint.  In NFP terms
this means having more than one vNIC for PCIe PF.  The vNICs
are usually corresponding 1:1 to Ethernet ports.  In core NIC
we use the legacy idea of vNIC *being* the Ethernet port,
hence netdevs put pX(sY) in their phys_port_name, like Ethernet
ports would.  When ASIC ports are fully represented we need to
be able to name different PCIe PF ports, too.  Use a scheme
similar to Ethernet ports - pfXsY, for PCIe PF number X,
sub-port Y.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
---
 drivers/net/ethernet/netronome/nfp/abm/main.c | 2 ++
 drivers/net/ethernet/netronome/nfp/nfp_port.c | 6 +++++-
 drivers/net/ethernet/netronome/nfp/nfp_port.h | 4 ++++
 3 files changed, 11 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/netronome/nfp/abm/main.c b/drivers/net/ethernet/netronome/nfp/abm/main.c
index 7afd24ce79a5..5a12bb20bced 100644
--- a/drivers/net/ethernet/netronome/nfp/abm/main.c
+++ b/drivers/net/ethernet/netronome/nfp/abm/main.c
@@ -109,6 +109,8 @@ nfp_abm_spawn_repr(struct nfp_app *app, struct nfp_abm_link *alink,
 			goto err_free_port;
 	} else {
 		port->pf_id = alink->abm->pf_id;
+		port->pf_split = app->pf->max_data_vnics > 1;
+		port->pf_split_id = alink->id;
 		port->vnic = alink->vnic->dp.ctrl_bar;
 	}
 
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_port.c b/drivers/net/ethernet/netronome/nfp/nfp_port.c
index a17f1ace6988..9c1298114c70 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_port.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_port.c
@@ -181,7 +181,11 @@ nfp_port_get_phys_port_name(struct net_device *netdev, char *name, size_t len)
 				     eth_port->label_subport);
 		break;
 	case NFP_PORT_PF_PORT:
-		n = snprintf(name, len, "pf%d", port->pf_id);
+		if (!port->pf_split)
+			n = snprintf(name, len, "pf%d", port->pf_id);
+		else
+			n = snprintf(name, len, "pf%ds%d", port->pf_id,
+				     port->pf_split_id);
 		break;
 	case NFP_PORT_VF_PORT:
 		n = snprintf(name, len, "pf%dvf%d", port->pf_id, port->vf_id);
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_port.h b/drivers/net/ethernet/netronome/nfp/nfp_port.h
index 1d9b2b47e27d..18666750456e 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_port.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_port.h
@@ -82,6 +82,8 @@ enum nfp_port_flags {
  * @eth_stats:	for %NFP_PORT_PHYS_PORT MAC stats if available
  * @pf_id:	for %NFP_PORT_PF_PORT, %NFP_PORT_VF_PORT ID of the PCI PF (0-3)
  * @vf_id:	for %NFP_PORT_VF_PORT ID of the PCI VF within @pf_id
+ * @pf_split:	for %NFP_PORT_PF_PORT %true if PCI PF has more than one vNIC
+ * @pf_split_id:for %NFP_PORT_PF_PORT ID of PCI PF vNIC (valid if @pf_split)
  * @vnic:	for %NFP_PORT_PF_PORT, %NFP_PORT_VF_PORT vNIC ctrl memory
  * @port_list:	entry on pf's list of ports
  */
@@ -108,6 +110,8 @@ struct nfp_port {
 		struct {
 			unsigned int pf_id;
 			unsigned int vf_id;
+			bool pf_split;
+			unsigned int pf_split_id;
 			u8 __iomem *vnic;
 		};
 	};
-- 
2.17.0

[PATCH net-next 13/13] nfp: assign vNIC id as phys_port_name of vNICs which are not ports

[Jakub Kicinski]

When NFP is modelled as a switch we assign phys_port_name to respective
port(representor )s:

 vNIC0 - | - PF port (pf%d)     MAC/PHY (p%d[s%d]) - |E==

In most cases there is only one vNIC for communication with the switch.
If there is more than one we need to be able to identify them.  Use %d
as phys_port_name of the vNICs.

We don't have to pass ID to nfp_net_debugfs_vnic_add() separately any
more.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
---
 drivers/net/ethernet/netronome/nfp/nfp_net.h  |  7 +++++--
 .../ethernet/netronome/nfp/nfp_net_common.c   | 20 ++++++++++++++++++-
 .../ethernet/netronome/nfp/nfp_net_debugfs.c  |  4 ++--
 .../net/ethernet/netronome/nfp/nfp_net_main.c |  4 +++-
 .../ethernet/netronome/nfp/nfp_netvf_main.c   |  2 +-
 5 files changed, 30 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net.h b/drivers/net/ethernet/netronome/nfp/nfp_net.h
index bd7d8ae31e17..57cb035dcc6d 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net.h
@@ -545,6 +545,7 @@ struct nfp_net_dp {
 /**
  * struct nfp_net - NFP network device structure
  * @dp:			Datapath structure
+ * @id:			vNIC id within the PF (0 for VFs)
  * @fw_ver:		Firmware version
  * @cap:                Capabilities advertised by the Firmware
  * @max_mtu:            Maximum support MTU advertised by the Firmware
@@ -597,6 +598,8 @@ struct nfp_net {
 
 	struct nfp_net_fw_version fw_ver;
 
+	u32 id;
+
 	u32 cap;
 	u32 max_mtu;
 
@@ -909,7 +912,7 @@ int nfp_net_ring_reconfig(struct nfp_net *nn, struct nfp_net_dp *new,
 void nfp_net_debugfs_create(void);
 void nfp_net_debugfs_destroy(void);
 struct dentry *nfp_net_debugfs_device_add(struct pci_dev *pdev);
-void nfp_net_debugfs_vnic_add(struct nfp_net *nn, struct dentry *ddir, int id);
+void nfp_net_debugfs_vnic_add(struct nfp_net *nn, struct dentry *ddir);
 void nfp_net_debugfs_dir_clean(struct dentry **dir);
 #else
 static inline void nfp_net_debugfs_create(void)
@@ -926,7 +929,7 @@ static inline struct dentry *nfp_net_debugfs_device_add(struct pci_dev *pdev)
 }
 
 static inline void
-nfp_net_debugfs_vnic_add(struct nfp_net *nn, struct dentry *ddir, int id)
+nfp_net_debugfs_vnic_add(struct nfp_net *nn, struct dentry *ddir)
 {
 }
 
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index d9111c077699..eea11e881bf5 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -3277,6 +3277,24 @@ nfp_net_features_check(struct sk_buff *skb, struct net_device *dev,
 	return features;
 }
 
+static int
+nfp_net_get_phys_port_name(struct net_device *netdev, char *name, size_t len)
+{
+	struct nfp_net *nn = netdev_priv(netdev);
+	int n;
+
+	if (nn->port)
+		return nfp_port_get_phys_port_name(netdev, name, len);
+
+	if (!nn->dp.is_vf) {
+		n = snprintf(name, len, "%d", nn->id);
+		if (n >= len)
+			return -EINVAL;
+	}
+
+	return 0;
+}
+
 /**
  * nfp_net_set_vxlan_port() - set vxlan port in SW and reconfigure HW
  * @nn:   NFP Net device to reconfigure
@@ -3475,7 +3493,7 @@ const struct net_device_ops nfp_net_netdev_ops = {
 	.ndo_set_mac_address	= nfp_net_set_mac_address,
 	.ndo_set_features	= nfp_net_set_features,
 	.ndo_features_check	= nfp_net_features_check,
-	.ndo_get_phys_port_name	= nfp_port_get_phys_port_name,
+	.ndo_get_phys_port_name	= nfp_net_get_phys_port_name,
 	.ndo_udp_tunnel_add	= nfp_net_add_vxlan_port,
 	.ndo_udp_tunnel_del	= nfp_net_del_vxlan_port,
 	.ndo_bpf		= nfp_net_xdp,
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c b/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c
index 67cdd8330c59..099b63d67451 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c
@@ -201,7 +201,7 @@ static const struct file_operations nfp_xdp_q_fops = {
 	.llseek = seq_lseek
 };
 
-void nfp_net_debugfs_vnic_add(struct nfp_net *nn, struct dentry *ddir, int id)
+void nfp_net_debugfs_vnic_add(struct nfp_net *nn, struct dentry *ddir)
 {
 	struct dentry *queues, *tx, *rx, *xdp;
 	char name[20];
@@ -211,7 +211,7 @@ void nfp_net_debugfs_vnic_add(struct nfp_net *nn, struct dentry *ddir, int id)
 		return;
 
 	if (nfp_net_is_data_vnic(nn))
-		sprintf(name, "vnic%d", id);
+		sprintf(name, "vnic%d", nn->id);
 	else
 		strcpy(name, "ctrl-vnic");
 	nn->debugfs_dir = debugfs_create_dir(name, ddir);
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_main.c b/drivers/net/ethernet/netronome/nfp/nfp_net_main.c
index b98422112385..28516eecccc8 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_main.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_main.c
@@ -178,11 +178,13 @@ nfp_net_pf_init_vnic(struct nfp_pf *pf, struct nfp_net *nn, unsigned int id)
 {
 	int err;
 
+	nn->id = id;
+
 	err = nfp_net_init(nn);
 	if (err)
 		return err;
 
-	nfp_net_debugfs_vnic_add(nn, pf->ddir, id);
+	nfp_net_debugfs_vnic_add(nn, pf->ddir);
 
 	if (nn->port) {
 		err = nfp_devlink_port_register(pf->app, nn->port);
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_netvf_main.c b/drivers/net/ethernet/netronome/nfp/nfp_netvf_main.c
index b802a1d55449..68928c86b698 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_netvf_main.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_netvf_main.c
@@ -283,7 +283,7 @@ static int nfp_netvf_pci_probe(struct pci_dev *pdev,
 
 	nfp_net_info(nn);
 	vf->ddir = nfp_net_debugfs_device_add(pdev);
-	nfp_net_debugfs_vnic_add(nn, vf->ddir, 0);
+	nfp_net_debugfs_vnic_add(nn, vf->ddir);
 
 	return 0;
 
-- 
2.17.0

Re: [PATCH net-next 00/13] nfp: abm: add basic support for advanced buffering NIC

[Or Gerlitz]

On Tue, May 22, 2018 at 8:12 AM, Jakub Kicinski
<jakub.kicinski@netronome.com> wrote:
> Hi!
>
> This series lays groundwork for advanced buffer management NIC feature.
> It makes necessary NFP core changes, spawns representors and adds devlink
> glue.  Following series will add the actual buffering configuration (patch
> series size limit).
>
> First three patches add support for configuring NFP buffer pools via a
> mailbox.  The existing devlink APIs are used for the purpose.
>
> Third patch allows us to perform small reads from the NFP memory.
>
> The rest of the patch set adds eswitch mode change support and makes
> the driver spawn appropriate representors.

Hi Jakub,

Could you provide more higher level description on the abm use-case
and nature of these representors? I understand that under abm you are
modeling the nic as switch with vNIC ports, does vNIC port and vNIC
port rep have the same characteristics as VF and VF rep (xmit on one side
<--> send on 2nd side), does traffic is to be offloaded using TC, etc. What one
would be doing with vNIC instance, hand it to container ala the Intel
VMDQ concept?
can this be seen as veth HW offload? etc

Or.


> Jakub Kicinski (13):
>   nfp: move rtsym helpers to pf code
>   nfp: add support for per-PCI PF mailbox
>   nfp: add shared buffer configuration
>   nfp: core: allow 4-byte aligned accesses to Memory Units
>   nfp: abm: add initial active buffer management NIC skeleton
>   nfp: abm: create project-specific vNIC structure
>   nfp: add app pointer to port representors
>   devlink: don't take instance lock around eswitch mode set
>   nfp: add devlink_eswitch_mode_set callback
>   nfp: abm: spawn port netdevs
>   nfp: abm: force Ethternet port up
>   nfp: use split in naming of PCIe PF ports
>   nfp: assign vNIC id as phys_port_name of vNICs which are not ports

Re: [PATCH net-next 00/13] nfp: abm: add basic support for advanced buffering NIC

[Jakub Kicinski]

On Mon, May 21, 2018 at 11:32 PM, Or Gerlitz wrote:
> On Tue, May 22, 2018 at 8:12 AM, Jakub Kicinski wrote:
>> Hi!
>>
>> This series lays groundwork for advanced buffer management NIC feature.
>> It makes necessary NFP core changes, spawns representors and adds devlink
>> glue.  Following series will add the actual buffering configuration (patch
>> series size limit).
>>
>> First three patches add support for configuring NFP buffer pools via a
>> mailbox.  The existing devlink APIs are used for the purpose.
>>
>> Third patch allows us to perform small reads from the NFP memory.
>>
>> The rest of the patch set adds eswitch mode change support and makes
>> the driver spawn appropriate representors.
>
> Hi Jakub,
>
> Could you provide more higher level description on the abm use-case
> and nature of these representors? I understand that under abm you are
> modeling the nic as switch with vNIC ports, does vNIC port and vNIC
> port rep have the same characteristics as VF and VF rep (xmit on one side
> <--> send on 2nd side),  does traffic is to be offloaded using TC, etc.
> What one would be doing with vNIC instance, hand it to container ala the Intel
> VMDQ concept?
> can this be seen as veth HW offload? etc

Yes, the reprs can be used like VF reprs but that's not the main use
case. We are targeting container world with ABM, so no VFs and no
SR-IOV.  There is only one vNIC per port and no veth offload etc. In
the most basic scenario with 1 PF corresponding to 1 port there is no
real use for switching.

The main purpose here is that we want to setup the buffering and QoS
inside the NIC (both for TX and RX) and then use eBPF to perform
filtering, queue assignment and per-application RSS. That's pretty
much it at this point.

Switching if any will be a basic bridge offload.  QoS configuration
will all be done using TC qdisc offload, RED etc. exactly like mlxsw
:)

Re: [PATCH net-next 08/13] devlink: don't take instance lock around eswitch mode set

[Jiri Pirko]

Tue, May 22, 2018 at 07:12:50AM CEST, jakub.kicinski@netronome.com wrote:
>Changing switch mode may want to register and unregister devlink
>ports.  Therefore similarly to DEVLINK_CMD_PORT_SPLIT/UNSPLIT it
>should not take the instance lock.  Drivers don't depend on existing
>locking since it's a very recent addition.
>
>Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>

Acked-by: Jiri Pirko <jiri@mellanox.com>

Re: [PATCH net-next 00/13] nfp: abm: add basic support for advanced buffering NIC

[Or Gerlitz]

On Tue, May 22, 2018 at 10:56 AM, Jakub Kicinski
<jakub.kicinski@netronome.com> wrote:
> On Mon, May 21, 2018 at 11:32 PM, Or Gerlitz wrote:
>> On Tue, May 22, 2018 at 8:12 AM, Jakub Kicinski wrote:
>>> Hi!
>>>
>>> This series lays groundwork for advanced buffer management NIC feature.
>>> It makes necessary NFP core changes, spawns representors and adds devlink
>>> glue.  Following series will add the actual buffering configuration (patch
>>> series size limit).
>>>
>>> First three patches add support for configuring NFP buffer pools via a
>>> mailbox.  The existing devlink APIs are used for the purpose.
>>>
>>> Third patch allows us to perform small reads from the NFP memory.
>>>
>>> The rest of the patch set adds eswitch mode change support and makes
>>> the driver spawn appropriate representors.
>>
>> Hi Jakub,
>>
>> Could you provide more higher level description on the abm use-case
>> and nature of these representors? I understand that under abm you are
>> modeling the nic as switch with vNIC ports, does vNIC port and vNIC
>> port rep have the same characteristics as VF and VF rep (xmit on one side
>> <--> send on 2nd side),  does traffic is to be offloaded using TC, etc.
>> What one would be doing with vNIC instance, hand it to container ala the Intel
>> VMDQ concept?
>> can this be seen as veth HW offload? etc

> Yes, the reprs can be used like VF reprs but that's not the main use
> case. We are targeting container world with ABM, so no VFs and no
> SR-IOV.  There is only one vNIC per port and no veth offload etc. In

one vNIC for multiple containers? or you have a (v?) port per container?

> In the most basic scenario with 1 PF corresponding to 1 port there is no
> real use for switching.

multiple containers? please clarify it a little better

> The main purpose here is that we want to setup the buffering and QoS
> inside the NIC (both for TX and RX) and then use eBPF to perform
> filtering, queue assignment and per-application RSS. That's pretty
> much it at this point.

> Switching if any will be a basic bridge offload.  QoS configuration
> will all be done using TC qdisc offload, RED etc. exactly like mlxsw :)

I guess I'll understand it better once you clarify the multiple
containers thing,
thanks for the details and openness

Re: [PATCH net-next 00/13] nfp: abm: add basic support for advanced buffering NIC

[Jakub Kicinski]

On Tue, 22 May 2018 17:50:45 +0300, Or Gerlitz wrote:
> On Tue, May 22, 2018 at 10:56 AM, Jakub Kicinski wrote:
> > On Mon, May 21, 2018 at 11:32 PM, Or Gerlitz wrote:  
> >> On Tue, May 22, 2018 at 8:12 AM, Jakub Kicinski wrote:  
> >>> Hi!
> >>>
> >>> This series lays groundwork for advanced buffer management NIC feature.
> >>> It makes necessary NFP core changes, spawns representors and adds devlink
> >>> glue.  Following series will add the actual buffering configuration (patch
> >>> series size limit).
> >>>
> >>> First three patches add support for configuring NFP buffer pools via a
> >>> mailbox.  The existing devlink APIs are used for the purpose.
> >>>
> >>> Third patch allows us to perform small reads from the NFP memory.
> >>>
> >>> The rest of the patch set adds eswitch mode change support and makes
> >>> the driver spawn appropriate representors.  
> >>
> >> Hi Jakub,
> >>
> >> Could you provide more higher level description on the abm use-case
> >> and nature of these representors? I understand that under abm you are
> >> modeling the nic as switch with vNIC ports, does vNIC port and vNIC
> >> port rep have the same characteristics as VF and VF rep (xmit on one side
> >> <--> send on 2nd side),  does traffic is to be offloaded using TC, etc.
> >> What one would be doing with vNIC instance, hand it to container ala the Intel
> >> VMDQ concept?
> >> can this be seen as veth HW offload? etc  
> 
> > Yes, the reprs can be used like VF reprs but that's not the main use
> > case. We are targeting container world with ABM, so no VFs and no
> > SR-IOV.  There is only one vNIC per port and no veth offload etc. In  
> 
> one vNIC for multiple containers? or you have a (v?) port per container?

One vNIC with many queues for multiple containers.  If containers have
QoS requirements they should be pinned to specific CPUs and eBPF on the
card can be used to RSS only to the queues associated with those CPUs,
while TC qdisc offload can be used to manipulate the scheduling of
those queues (and ECN marking).

> > In the most basic scenario with 1 PF corresponding to 1 port there is no
> > real use for switching.  
> 
> multiple containers? please clarify it a little better

Yes, if we had a netdev per container that would mean we would need
switching.  But vetch/macvlan/ipvlan offload and therefore multiple
netdevs is not a requirement right now.

IOW we have the eBPF programmable RSS and we are trying to extend the
QoS to be able to make sure that critical/important containers' RX/TX
doesn't get disturbed by less important workloads.

> > The main purpose here is that we want to setup the buffering and QoS
> > inside the NIC (both for TX and RX) and then use eBPF to perform
> > filtering, queue assignment and per-application RSS. That's pretty
> > much it at this point.  
> >
> > Switching if any will be a basic bridge offload.  QoS configuration
> > will all be done using TC qdisc offload, RED etc. exactly like mlxsw :)  
> 
> I guess I'll understand it better once you clarify the multiple
> containers thing,
> thanks for the details and openness

I hope my clarifications help :)

Re: [PATCH net-next 00/13] nfp: abm: add basic support for advanced buffering NIC

[David Miller]

From: Jakub Kicinski <jakub.kicinski@netronome.com>
Date: Mon, 21 May 2018 22:12:42 -0700

> This series lays groundwork for advanced buffer management NIC feature.
> It makes necessary NFP core changes, spawns representors and adds devlink
> glue.  Following series will add the actual buffering configuration (patch
> series size limit).
> 
> First three patches add support for configuring NFP buffer pools via a
> mailbox.  The existing devlink APIs are used for the purpose.
> 
> Third patch allows us to perform small reads from the NFP memory.
> 
> The rest of the patch set adds eswitch mode change support and makes
> the driver spawn appropriate representors.

Series applied, thank you!