* [RFC v3 03/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev
2017-02-07 20:22 [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
2017-02-07 20:23 ` [RFC v3 05/11] IB/hfi-vnic: VNIC statistics support Vishwanathapura, Niranjana
` (7 subsequent siblings)
8 siblings, 0 replies; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura, Sadanand Warrier, Sudeep Dutt,
Tanya K Jajodia, Andrzej Kacprowski
HFI VNIC netdev function supports Ethernet functionality over Omni-Path
fabric by encapsulating Ethernet packets inside Omni-Path packet header.
It allocates a rdma netdev device and interfaces with the network stack to
provide standard Ethernet network interfaces. It overrides HFI device's
netdev operations where it is required.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Sadanand Warrier <sadanand.warrier@intel.com>
Signed-off-by: Sudeep Dutt <sudeep.dutt@intel.com>
Signed-off-by: Tanya K Jajodia <tanya.k.jajodia@intel.com>
Signed-off-by: Andrzej Kacprowski <andrzej.kacprowski@intel.com>
---
MAINTAINERS | 7 +
drivers/infiniband/Kconfig | 1 +
drivers/infiniband/ulp/Makefile | 1 +
drivers/infiniband/ulp/hfi_vnic/Kconfig | 8 +
drivers/infiniband/ulp/hfi_vnic/Makefile | 6 +
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.c | 239 +++++++++++++++++++++
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h | 58 +++++
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c | 65 ++++++
.../infiniband/ulp/hfi_vnic/hfi_vnic_internal.h | 186 ++++++++++++++++
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c | 224 +++++++++++++++++++
10 files changed, 795 insertions(+)
create mode 100644 drivers/infiniband/ulp/hfi_vnic/Kconfig
create mode 100644 drivers/infiniband/ulp/hfi_vnic/Makefile
create mode 100644 drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.c
create mode 100644 drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h
create mode 100644 drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c
create mode 100644 drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
create mode 100644 drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
diff --git a/MAINTAINERS b/MAINTAINERS
index 5f0420a..83a8cfa 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -5764,6 +5764,13 @@ F: drivers/block/cciss*
F: include/linux/cciss_ioctl.h
F: include/uapi/linux/cciss_ioctl.h
+HFI-VNIC DRIVER
+M: Dennis Dalessandro <dennis.dalessandro@intel.com>
+M: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
+L: linux-rdma@vger.kernel.org
+S: Supported
+F: drivers/infiniband/ulp/hfi_vnic
+
HFI1 DRIVER
M: Mike Marciniszyn <mike.marciniszyn@intel.com>
M: Dennis Dalessandro <dennis.dalessandro@intel.com>
diff --git a/drivers/infiniband/Kconfig b/drivers/infiniband/Kconfig
index 6709173..7d97e58 100644
--- a/drivers/infiniband/Kconfig
+++ b/drivers/infiniband/Kconfig
@@ -85,6 +85,7 @@ source "drivers/infiniband/ulp/srpt/Kconfig"
source "drivers/infiniband/ulp/iser/Kconfig"
source "drivers/infiniband/ulp/isert/Kconfig"
+source "drivers/infiniband/ulp/hfi_vnic/Kconfig"
source "drivers/infiniband/sw/rdmavt/Kconfig"
source "drivers/infiniband/sw/rxe/Kconfig"
diff --git a/drivers/infiniband/ulp/Makefile b/drivers/infiniband/ulp/Makefile
index f3c7dcf..9d20ac7 100644
--- a/drivers/infiniband/ulp/Makefile
+++ b/drivers/infiniband/ulp/Makefile
@@ -3,3 +3,4 @@ obj-$(CONFIG_INFINIBAND_SRP) += srp/
obj-$(CONFIG_INFINIBAND_SRPT) += srpt/
obj-$(CONFIG_INFINIBAND_ISER) += iser/
obj-$(CONFIG_INFINIBAND_ISERT) += isert/
+obj-$(CONFIG_INFINIBAND_HFI_VNIC) += hfi_vnic/
diff --git a/drivers/infiniband/ulp/hfi_vnic/Kconfig b/drivers/infiniband/ulp/hfi_vnic/Kconfig
new file mode 100644
index 0000000..cca6810
--- /dev/null
+++ b/drivers/infiniband/ulp/hfi_vnic/Kconfig
@@ -0,0 +1,8 @@
+config INFINIBAND_HFI_VNIC
+ tristate "Intel HFI VNIC support"
+ depends on X86_64 && INFINIBAND
+ ---help---
+ This is HFI Virtual Network Interface Controller (VNIC) driver
+ for Ethernet over HFI feature. It implements the HW independent
+ VNIC functionality. It interfaces with Linux stack for data path
+ and IB MAD for the control path.
diff --git a/drivers/infiniband/ulp/hfi_vnic/Makefile b/drivers/infiniband/ulp/hfi_vnic/Makefile
new file mode 100644
index 0000000..ad7db32
--- /dev/null
+++ b/drivers/infiniband/ulp/hfi_vnic/Makefile
@@ -0,0 +1,6 @@
+# Makefile - Intel HFI Virtual Network Controller driver
+# Copyright(c) 2017, Intel Corporation.
+#
+obj-$(CONFIG_INFINIBAND_HFI_VNIC) += hfi_vnic.o
+
+hfi_vnic-y := hfi_vnic_netdev.o hfi_vnic_encap.o hfi_vnic_ethtool.o
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.c
new file mode 100644
index 0000000..159172c
--- /dev/null
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.c
@@ -0,0 +1,239 @@
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * This file contains HFI VNIC encapsulation/decapsulation function.
+ */
+
+#include <linux/if_ether.h>
+#include <linux/if_vlan.h>
+
+#include "hfi_vnic_internal.h"
+
+/* OPA 16B Header fields */
+#define OPA_16B_LID_MASK 0xFFFFFull
+#define OPA_16B_SLID_HIGH_SHFT 8
+#define OPA_16B_SLID_MASK 0xF00ull
+#define OPA_16B_DLID_MASK 0xF000ull
+#define OPA_16B_DLID_HIGH_SHFT 12
+#define OPA_16B_LEN_SHFT 20
+#define OPA_16B_SC_SHFT 20
+#define OPA_16B_RC_SHFT 25
+#define OPA_16B_PKEY_SHFT 16
+
+#define HFI_VNIC_L4_HDR_SHFT 16
+
+/* L2+L4 hdr len is 20 bytes (5 quad words) */
+#define HFI_VNIC_HDR_QW_LEN 5
+
+static inline void hfi_vnic_make_header(u8 *hdr, u32 slid, u32 dlid, u16 len,
+ u16 pkey, u16 entropy, u8 sc, u8 rc,
+ u8 l4_type, u16 l4_hdr)
+{
+ /* h[1]: LT=1, 16B L2=10 */
+ u32 h[HFI_VNIC_HDR_QW_LEN] = {0, 0xc0000000, 0, 0, 0};
+
+ h[2] = l4_type;
+ h[3] = entropy;
+ h[4] = l4_hdr << HFI_VNIC_L4_HDR_SHFT;
+
+ /* Extract and set 4 upper bits and 20 lower bits of the lids */
+ h[0] |= (slid & OPA_16B_LID_MASK);
+ h[2] |= ((slid >> (20 - OPA_16B_SLID_HIGH_SHFT)) & OPA_16B_SLID_MASK);
+
+ h[1] |= (dlid & OPA_16B_LID_MASK);
+ h[2] |= ((dlid >> (20 - OPA_16B_DLID_HIGH_SHFT)) & OPA_16B_DLID_MASK);
+
+ h[0] |= (len << OPA_16B_LEN_SHFT);
+ h[1] |= (rc << OPA_16B_RC_SHFT);
+ h[1] |= (sc << OPA_16B_SC_SHFT);
+ h[2] |= ((u32)pkey << OPA_16B_PKEY_SHFT);
+
+ memcpy(hdr, h, HFI_VNIC_HDR_LEN);
+}
+
+/* hfi_vnic_get_dlid - find and return the DLID */
+static uint32_t hfi_vnic_get_dlid(struct hfi_vnic_adapter *adapter,
+ struct sk_buff *skb, u8 def_port)
+{
+ struct __hfi_veswport_info *info = &adapter->info;
+ struct ethhdr *mac_hdr = (struct ethhdr *)skb_mac_header(skb);
+ u32 dlid;
+
+ if (is_multicast_ether_addr(mac_hdr->h_dest)) {
+ dlid = info->vesw.u_mcast_dlid;
+ } else {
+ if (is_local_ether_addr(mac_hdr->h_dest)) {
+ dlid = ((uint32_t)mac_hdr->h_dest[5] << 16) |
+ ((uint32_t)mac_hdr->h_dest[4] << 8) |
+ mac_hdr->h_dest[3];
+ if (unlikely(!dlid))
+ v_warn("Null dlid in MAC address\n");
+ } else if (def_port != HFI_VNIC_INVALID_PORT) {
+ dlid = info->vesw.u_ucast_dlid[def_port];
+ }
+ }
+
+ return dlid;
+}
+
+/* hfi_vnic_get_sc - return the service class */
+static u8 hfi_vnic_get_sc(struct __hfi_veswport_info *info,
+ struct sk_buff *skb)
+{
+ struct ethhdr *mac_hdr = (struct ethhdr *)skb_mac_header(skb);
+ u16 vlan_tci;
+ u8 sc;
+
+ if (!__vlan_get_tag(skb, &vlan_tci)) {
+ u8 pcp = HFI_VNIC_VLAN_PCP(vlan_tci);
+
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ sc = info->vport.pcp_to_sc_mc[pcp];
+ else
+ sc = info->vport.pcp_to_sc_uc[pcp];
+ } else {
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ sc = info->vport.non_vlan_sc_mc;
+ else
+ sc = info->vport.non_vlan_sc_uc;
+ }
+
+ return sc;
+}
+
+u8 hfi_vnic_get_vl(struct hfi_vnic_adapter *adapter, struct sk_buff *skb)
+{
+ struct ethhdr *mac_hdr = (struct ethhdr *)skb_mac_header(skb);
+ struct __hfi_veswport_info *info = &adapter->info;
+ u8 vl;
+
+ if (skb_vlan_tag_present(skb)) {
+ u8 pcp = skb_vlan_tag_get(skb) >> VLAN_PRIO_SHIFT;
+
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ vl = info->vport.pcp_to_vl_mc[pcp];
+ else
+ vl = info->vport.pcp_to_vl_uc[pcp];
+ } else {
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ vl = info->vport.non_vlan_vl_mc;
+ else
+ vl = info->vport.non_vlan_vl_uc;
+ }
+
+ return vl;
+}
+
+/* hfi_vnic_calc_entropy - calculate the packet entropy */
+u8 hfi_vnic_calc_entropy(struct hfi_vnic_adapter *adapter, struct sk_buff *skb)
+{
+ u16 hash16;
+
+ /*
+ * Get flow based 16-bit hash and then XOR the upper and lower bytes
+ * to get the entropy.
+ * __skb_tx_hash limits qcount to 16 bits. Hence, get 15-bit hash.
+ */
+ hash16 = __skb_tx_hash(adapter->netdev, skb, BIT(15));
+ return (u8)((hash16 >> 8) ^ (hash16 & 0xff));
+}
+
+/* hfi_vnic_get_def_port - get default port based on entropy */
+static inline u8 hfi_vnic_get_def_port(struct hfi_vnic_adapter *adapter,
+ u8 entropy)
+{
+ u8 flow_id;
+
+ /* Add the upper and lower 4-bits of entropy to get the flow id */
+ flow_id = ((entropy & 0xf) + (entropy >> 4));
+ return adapter->flow_tbl[flow_id & (HFI_VNIC_FLOW_TBL_SIZE - 1)];
+}
+
+/* Calculate packet length including OPA header, crc and padding */
+static inline int hfi_vnic_wire_length(struct sk_buff *skb)
+{
+ u32 pad_len;
+
+ /* padding for 8 bytes size alignment */
+ pad_len = -(skb->len + HFI_VNIC_ICRC_TAIL_LEN) & 0x7;
+ pad_len += HFI_VNIC_ICRC_TAIL_LEN;
+
+ return (skb->len + pad_len) >> 3;
+}
+
+/* hfi_vnic_encap_skb - encapsulate skb packet with OPA header and meta data */
+void hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb)
+{
+ struct __hfi_veswport_info *info = &adapter->info;
+ struct hfi_vnic_skb_mdata *mdata;
+ u8 def_port, sc, entropy, *hdr;
+ u16 len, l4_hdr;
+ u32 dlid;
+
+ hdr = skb_push(skb, HFI_VNIC_HDR_LEN);
+
+ entropy = hfi_vnic_calc_entropy(adapter, skb);
+ def_port = hfi_vnic_get_def_port(adapter, entropy);
+ len = hfi_vnic_wire_length(skb);
+ dlid = hfi_vnic_get_dlid(adapter, skb, def_port);
+ sc = hfi_vnic_get_sc(info, skb);
+ l4_hdr = info->vesw.vesw_id;
+
+ mdata = (struct hfi_vnic_skb_mdata *)skb_push(skb, sizeof(*mdata));
+ mdata->vl = hfi_vnic_get_vl(adapter, skb);
+ mdata->entropy = entropy;
+ mdata->flags = 0;
+ if (unlikely(!dlid)) {
+ mdata->flags = HFI_VNIC_SKB_MDATA_ENCAP_ERR;
+ return;
+ }
+
+ hfi_vnic_make_header(hdr, info->vport.encap_slid, dlid, len,
+ info->vesw.pkey, entropy, sc, 0,
+ HFI_VNIC_L4_ETHR, l4_hdr);
+}
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h
new file mode 100644
index 0000000..aa93f62
--- /dev/null
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h
@@ -0,0 +1,58 @@
+#ifndef _HFI_VNIC_ENCAP_H
+#define _HFI_VNIC_ENCAP_H
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * This file contains all HFI VNIC declaration required for encapsulation
+ * and decapsulation of Ethernet packets
+ */
+
+#define HFI_VESW_MAX_NUM_DEF_PORT 16
+#define HFI_VNIC_MAX_NUM_PCP 8
+
+#endif /* _HFI_VNIC_ENCAP_H */
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c
new file mode 100644
index 0000000..ca058a4
--- /dev/null
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c
@@ -0,0 +1,65 @@
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * This file contains HFI VNIC ethtool functions
+ */
+
+#include <linux/ethtool.h>
+
+#include "hfi_vnic_internal.h"
+
+/* ethtool ops */
+static const struct ethtool_ops hfi_vnic_ethtool_ops = {
+ .get_link = ethtool_op_get_link,
+};
+
+/* hfi_vnic_set_ethtool_ops - set ethtool ops */
+void hfi_vnic_set_ethtool_ops(struct net_device *netdev)
+{
+ netdev->ethtool_ops = &hfi_vnic_ethtool_ops;
+}
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
new file mode 100644
index 0000000..106563d
--- /dev/null
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
@@ -0,0 +1,186 @@
+#ifndef _HFI_VNIC_INTERNAL_H
+#define _HFI_VNIC_INTERNAL_H
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * This file contains HFI VNIC driver internal declarations
+ */
+
+#include <linux/bitops.h>
+#include <linux/etherdevice.h>
+#include <linux/hashtable.h>
+#include <linux/sizes.h>
+#include <rdma/opa_hfi.h>
+
+#include "hfi_vnic_encap.h"
+
+#define HFI_VNIC_VLAN_PCP(vlan_tci) \
+ (((vlan_tci) & VLAN_PRIO_MASK) >> VLAN_PRIO_SHIFT)
+
+/* Flow to default port redirection table size */
+#define HFI_VNIC_FLOW_TBL_SIZE 32
+
+/* Invalid port number */
+#define HFI_VNIC_INVALID_PORT 0xff
+
+struct hfi_vnic_adapter;
+
+/**
+ * struct __hfi_vesw_info - HFI vnic virtual switch info
+ */
+struct __hfi_vesw_info {
+ u16 fabric_id;
+ u16 vesw_id;
+
+ u8 rsvd0[6];
+ u16 def_port_mask;
+
+ u8 rsvd1[2];
+ u16 pkey;
+
+ u8 rsvd2[4];
+ u32 u_mcast_dlid;
+ u32 u_ucast_dlid[HFI_VESW_MAX_NUM_DEF_PORT];
+
+ u8 rsvd3[44];
+ u16 eth_mtu[HFI_VNIC_MAX_NUM_PCP];
+ u16 eth_mtu_non_vlan;
+ u8 rsvd4[2];
+} __packed;
+
+/**
+ * struct __hfi_per_veswport_info - HFI vnic per port info
+ */
+struct __hfi_per_veswport_info {
+ u32 port_num;
+
+ u8 eth_link_status;
+ u8 rsvd0[3];
+
+ u8 base_mac_addr[ETH_ALEN];
+ u8 config_state;
+ u8 oper_state;
+
+ u16 max_mac_tbl_ent;
+ u16 max_smac_ent;
+ u32 mac_tbl_digest;
+ u8 rsvd1[4];
+
+ u32 encap_slid;
+
+ u8 pcp_to_sc_uc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_vl_uc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_sc_mc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_vl_mc[HFI_VNIC_MAX_NUM_PCP];
+
+ u8 non_vlan_sc_uc;
+ u8 non_vlan_vl_uc;
+ u8 non_vlan_sc_mc;
+ u8 non_vlan_vl_mc;
+
+ u8 rsvd2[48];
+
+ u16 uc_macs_gen_count;
+ u16 mc_macs_gen_count;
+
+ u8 rsvd3[8];
+} __packed;
+
+/**
+ * struct __hfi_veswport_info - HFI vnic port info
+ */
+struct __hfi_veswport_info {
+ struct __hfi_vesw_info vesw;
+ struct __hfi_per_veswport_info vport;
+};
+
+/**
+ * struct hfi_vnic_adapter - HFI VNIC netdev private data structure
+ * @netdev: pointer to associated netdev
+ * @ibdev: ib device
+ * @rn_ops: rdma netdev's net_device_ops
+ * @port_num: OPA port number
+ * @vport_num: vesw port number
+ * @lock: adapter lock
+ * @info: virtual ethernet switch port information
+ * @flow_tbl: flow to default port redirection table
+ */
+struct hfi_vnic_adapter {
+ struct net_device *netdev;
+ struct ib_device *ibdev;
+ const struct net_device_ops *rn_ops;
+
+ u8 port_num;
+ u8 vport_num;
+
+ /* Lock used around concurrent updates to netdev */
+ struct mutex lock;
+
+ struct __hfi_veswport_info info;
+
+ u8 flow_tbl[HFI_VNIC_FLOW_TBL_SIZE];
+};
+
+#define v_dbg(format, arg...) \
+ netdev_dbg(adapter->netdev, format, ## arg)
+#define v_err(format, arg...) \
+ netdev_err(adapter->netdev, format, ## arg)
+#define v_info(format, arg...) \
+ netdev_info(adapter->netdev, format, ## arg)
+#define v_warn(format, arg...) \
+ netdev_warn(adapter->netdev, format, ## arg)
+
+struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
+ u8 port_num, u8 vport_num);
+void hfi_vnic_rem_netdev(struct hfi_vnic_adapter *adapter);
+void hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
+u8 hfi_vnic_get_vl(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
+u8 hfi_vnic_calc_entropy(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
+void hfi_vnic_set_ethtool_ops(struct net_device *netdev);
+
+#endif /* _HFI_VNIC_INTERNAL_H */
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
new file mode 100644
index 0000000..f04c8a834
--- /dev/null
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
@@ -0,0 +1,224 @@
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * This file contains HFI Virtual Network Interface Controller (VNIC) driver
+ */
+
+#include <linux/module.h>
+#include <linux/if_vlan.h>
+
+#include "hfi_vnic_internal.h"
+
+#define HFI_TX_TIMEOUT_MS 1000
+
+#define HFI_VNIC_SKB_HEADROOM \
+ ALIGN((HFI_VNIC_HDR_LEN + HFI_VNIC_SKB_MDATA_LEN), 8)
+
+/* hfi_netdev_start_xmit - transmit function */
+static netdev_tx_t hfi_netdev_start_xmit(struct sk_buff *skb,
+ struct net_device *netdev)
+{
+ struct hfi_vnic_adapter *adapter = hfi_vnic_priv(netdev);
+
+ v_dbg("xmit: queue %d skb len %d\n", skb->queue_mapping, skb->len);
+ /* pad to ensure mininum ethernet packet length */
+ if (unlikely(skb->len < ETH_ZLEN)) {
+ if (skb_padto(skb, ETH_ZLEN))
+ return NETDEV_TX_OK;
+
+ skb_put(skb, ETH_ZLEN - skb->len);
+ }
+
+ hfi_vnic_encap_skb(adapter, skb);
+ return adapter->rn_ops->ndo_start_xmit(skb, netdev);
+}
+
+static u16 hfi_vnic_select_queue(struct net_device *netdev, struct sk_buff *skb,
+ void *accel_priv,
+ select_queue_fallback_t fallback)
+{
+ struct hfi_vnic_adapter *adapter = hfi_vnic_priv(netdev);
+ struct hfi_vnic_skb_mdata *mdata;
+ int rc;
+
+ /* pass entropy and vl as metadata in skb */
+ mdata = (struct hfi_vnic_skb_mdata *)skb_push(skb, sizeof(*mdata));
+ mdata->entropy = hfi_vnic_calc_entropy(adapter, skb);
+ mdata->vl = hfi_vnic_get_vl(adapter, skb);
+ rc = adapter->rn_ops->ndo_select_queue(netdev, skb,
+ accel_priv, fallback);
+ skb_pull(skb, sizeof(*mdata));
+ return rc;
+}
+
+/* hfi_vnic_set_mac_addr - change mac address */
+static int hfi_vnic_set_mac_addr(struct net_device *netdev, void *addr)
+{
+ struct hfi_vnic_adapter *adapter = hfi_vnic_priv(netdev);
+ struct sockaddr *sa = addr;
+ int rc;
+
+ if (!memcmp(netdev->dev_addr, sa->sa_data, ETH_ALEN))
+ return 0;
+
+ mutex_lock(&adapter->lock);
+ rc = eth_mac_addr(netdev, addr);
+ mutex_unlock(&adapter->lock);
+
+ return rc;
+}
+
+/* hfi_netdev_open - activate network interface */
+static int hfi_netdev_open(struct net_device *netdev)
+{
+ struct hfi_vnic_adapter *adapter = hfi_vnic_priv(netdev);
+ int rc;
+
+ rc = adapter->rn_ops->ndo_open(adapter->netdev);
+ if (rc) {
+ v_dbg("open failed %d\n", rc);
+ return rc;
+ }
+
+ v_info("opened\n");
+ return 0;
+}
+
+/* hfi_netdev_close - disable network interface */
+static int hfi_netdev_close(struct net_device *netdev)
+{
+ struct hfi_vnic_adapter *adapter = hfi_vnic_priv(netdev);
+ int rc;
+
+ rc = adapter->rn_ops->ndo_stop(adapter->netdev);
+ if (rc) {
+ v_dbg("close failed %d\n", rc);
+ return rc;
+ }
+
+ v_info("closed\n");
+ return 0;
+}
+
+/* netdev ops */
+static const struct net_device_ops hfi_netdev_ops = {
+ .ndo_open = hfi_netdev_open,
+ .ndo_stop = hfi_netdev_close,
+ .ndo_start_xmit = hfi_netdev_start_xmit,
+ .ndo_select_queue = hfi_vnic_select_queue,
+ .ndo_set_mac_address = hfi_vnic_set_mac_addr,
+};
+
+/* hfi_vnic_add_netdev - create vnic netdev interface */
+struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
+ u8 port_num, u8 vport_num)
+{
+ struct hfi_vnic_adapter *adapter;
+ struct net_device *netdev;
+ struct rdma_netdev *rn;
+ int rc;
+
+ netdev = ibdev->alloc_rdma_netdev(ibdev, port_num,
+ RDMA_NETDEV_HFI_VNIC,
+ "veth%d", NET_NAME_UNKNOWN,
+ ether_setup);
+ if (!netdev)
+ return ERR_PTR(-ENOMEM);
+
+ adapter = kzalloc(sizeof(*adapter), GFP_KERNEL);
+ if (!adapter) {
+ rc = -ENOMEM;
+ goto adapter_err;
+ }
+
+ rn = netdev_priv(netdev);
+ rn->clnt_priv = adapter;
+ adapter->netdev = netdev;
+ adapter->ibdev = ibdev;
+ adapter->port_num = port_num;
+ adapter->vport_num = vport_num;
+ adapter->rn_ops = netdev->netdev_ops;
+
+ netdev->netdev_ops = &hfi_netdev_ops;
+ netdev->priv_flags |= IFF_LIVE_ADDR_CHANGE;
+ netdev->hard_header_len += HFI_VNIC_SKB_HEADROOM;
+ mutex_init(&adapter->lock);
+
+ SET_NETDEV_DEV(netdev, ibdev->dma_device);
+
+ hfi_vnic_set_ethtool_ops(netdev);
+ rc = register_netdev(netdev);
+ if (rc)
+ goto netdev_err;
+
+ netif_carrier_off(netdev);
+ netif_dormant_on(netdev);
+ v_info("initialized\n");
+
+ return adapter;
+netdev_err:
+ mutex_destroy(&adapter->lock);
+ kfree(adapter);
+adapter_err:
+ ibdev->free_rdma_netdev(netdev);
+
+ return ERR_PTR(rc);
+}
+
+/* hfi_vnic_rem_netdev - remove vnic netdev interface */
+void hfi_vnic_rem_netdev(struct hfi_vnic_adapter *adapter)
+{
+ struct net_device *netdev = adapter->netdev;
+ struct ib_device *ibdev = adapter->ibdev;
+
+ v_info("removing\n");
+ unregister_netdev(netdev);
+ mutex_destroy(&adapter->lock);
+ kfree(adapter);
+ ibdev->free_rdma_netdev(netdev);
+}
--
1.8.3.1
^ permalink raw reply related [flat|nested] 40+ messages in thread* [RFC v3 05/11] IB/hfi-vnic: VNIC statistics support
2017-02-07 20:22 [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
2017-02-07 20:23 ` [RFC v3 03/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev Vishwanathapura, Niranjana
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
2017-02-07 20:23 ` [RFC v3 06/11] IB/hfi-vnic: VNIC MAC table support Vishwanathapura, Niranjana
` (6 subsequent siblings)
8 siblings, 0 replies; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura
HFI VNIC driver statistics support maintains various counters including
standard netdev counters and the Ethernet manager defined counters.
Add the Ethtool hook to read the counters.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
---
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c | 128 +++++++++++++++++++++
.../infiniband/ulp/hfi_vnic/hfi_vnic_internal.h | 4 +
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c | 20 ++++
3 files changed, 152 insertions(+)
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c
index ca058a4..a9925df 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c
@@ -53,9 +53,137 @@
#include "hfi_vnic_internal.h"
+enum {NETDEV_STATS, VNIC_STATS};
+
+struct vnic_stats {
+ char stat_string[ETH_GSTRING_LEN];
+ struct {
+ int type;
+ int sizeof_stat;
+ int stat_offset;
+ };
+};
+
+#define VNIC_STAT(m) { VNIC_STATS, \
+ FIELD_SIZEOF(struct hfi_vnic_stats, m), \
+ offsetof(struct hfi_vnic_stats, m) }
+#define VNIC_NETDEV_STAT(m) { NETDEV_STATS, \
+ FIELD_SIZEOF(struct net_device, m), \
+ offsetof(struct net_device, m) }
+
+static struct vnic_stats vnic_gstrings_stats[] = {
+ /* NETDEV stats */
+ {"rx_packets", VNIC_NETDEV_STAT(stats.rx_packets)},
+ {"tx_packets", VNIC_NETDEV_STAT(stats.tx_packets)},
+ {"rx_bytes", VNIC_NETDEV_STAT(stats.rx_bytes)},
+ {"tx_bytes", VNIC_NETDEV_STAT(stats.tx_bytes)},
+ {"rx_errors", VNIC_NETDEV_STAT(stats.rx_errors)},
+ {"tx_errors", VNIC_NETDEV_STAT(stats.tx_errors)},
+ {"rx_dropped", VNIC_NETDEV_STAT(stats.rx_dropped)},
+ {"tx_dropped", VNIC_NETDEV_STAT(stats.tx_dropped)},
+
+ /* SUMMARY counters */
+ {"tx_unicast", VNIC_STAT(tx_grp.unicast)},
+ {"tx_mcastbcast", VNIC_STAT(tx_grp.mcastbcast)},
+ {"tx_untagged", VNIC_STAT(tx_grp.untagged)},
+ {"tx_vlan", VNIC_STAT(tx_grp.vlan)},
+
+ {"tx_64_size", VNIC_STAT(tx_grp.s_64)},
+ {"tx_65_127", VNIC_STAT(tx_grp.s_65_127)},
+ {"tx_128_255", VNIC_STAT(tx_grp.s_128_255)},
+ {"tx_256_511", VNIC_STAT(tx_grp.s_256_511)},
+ {"tx_512_1023", VNIC_STAT(tx_grp.s_512_1023)},
+ {"tx_1024_1518", VNIC_STAT(tx_grp.s_1024_1518)},
+ {"tx_1519_max", VNIC_STAT(tx_grp.s_1519_max)},
+
+ {"rx_unicast", VNIC_STAT(rx_grp.unicast)},
+ {"rx_mcastbcast", VNIC_STAT(rx_grp.mcastbcast)},
+ {"rx_untagged", VNIC_STAT(rx_grp.untagged)},
+ {"rx_vlan", VNIC_STAT(rx_grp.vlan)},
+
+ {"rx_64_size", VNIC_STAT(rx_grp.s_64)},
+ {"rx_65_127", VNIC_STAT(rx_grp.s_65_127)},
+ {"rx_128_255", VNIC_STAT(rx_grp.s_128_255)},
+ {"rx_256_511", VNIC_STAT(rx_grp.s_256_511)},
+ {"rx_512_1023", VNIC_STAT(rx_grp.s_512_1023)},
+ {"rx_1024_1518", VNIC_STAT(rx_grp.s_1024_1518)},
+ {"rx_1519_max", VNIC_STAT(rx_grp.s_1519_max)},
+
+ /* ERROR counters */
+ {"rx_fifo_errors", VNIC_NETDEV_STAT(stats.rx_fifo_errors)},
+ {"rx_length_errors", VNIC_NETDEV_STAT(stats.rx_length_errors)},
+
+ {"tx_fifo_errors", VNIC_NETDEV_STAT(stats.tx_fifo_errors)},
+ {"tx_carrier_errors", VNIC_NETDEV_STAT(stats.tx_carrier_errors)},
+
+ {"tx_dlid_zero", VNIC_STAT(tx_dlid_zero)},
+ {"tx_drop_state", VNIC_STAT(tx_drop_state)},
+ {"rx_drop_state", VNIC_STAT(rx_drop_state)},
+ {"rx_oversize", VNIC_STAT(rx_oversize)},
+ {"rx_runt", VNIC_STAT(rx_runt)},
+};
+
+#define VNIC_STATS_LEN ARRAY_SIZE(vnic_gstrings_stats)
+
+/* vnic_get_sset_count - get string set count */
+static int vnic_get_sset_count(struct net_device *netdev, int sset)
+{
+ return (sset == ETH_SS_STATS) ? VNIC_STATS_LEN : -EOPNOTSUPP;
+}
+
+/* vnic_get_ethtool_stats - get statistics */
+static void vnic_get_ethtool_stats(struct net_device *netdev,
+ struct ethtool_stats *stats, u64 *data)
+{
+ struct hfi_vnic_adapter *adapter = hfi_vnic_priv(netdev);
+ struct hfi_vnic_stats vstats;
+ char *p = NULL;
+ int i;
+
+ memset(&vstats, 0, sizeof(vstats));
+ mutex_lock(&adapter->stats_lock);
+ adapter->rn_ops->ndo_get_stats64(netdev, &vstats.netstats);
+ for (i = 0; i < VNIC_STATS_LEN; i++) {
+ switch (vnic_gstrings_stats[i].type) {
+ case NETDEV_STATS:
+ p = (char *)netdev +
+ vnic_gstrings_stats[i].stat_offset;
+ break;
+ case VNIC_STATS:
+ p = (char *)&vstats +
+ vnic_gstrings_stats[i].stat_offset;
+ break;
+ default:
+ p = NULL;
+ }
+
+ if (p)
+ data[i] = (vnic_gstrings_stats[i].sizeof_stat ==
+ sizeof(u64)) ? *(u64 *)p : *(u32 *)p;
+ }
+ mutex_unlock(&adapter->stats_lock);
+}
+
+/* vnic_get_strings - get strings */
+static void vnic_get_strings(struct net_device *netdev, u32 stringset, u8 *data)
+{
+ int i;
+
+ if (stringset != ETH_SS_STATS)
+ return;
+
+ for (i = 0; i < VNIC_STATS_LEN; i++)
+ memcpy(data + i * ETH_GSTRING_LEN,
+ vnic_gstrings_stats[i].stat_string,
+ ETH_GSTRING_LEN);
+}
+
/* ethtool ops */
static const struct ethtool_ops hfi_vnic_ethtool_ops = {
.get_link = ethtool_op_get_link,
+ .get_strings = vnic_get_strings,
+ .get_sset_count = vnic_get_sset_count,
+ .get_ethtool_stats = vnic_get_ethtool_stats,
};
/* hfi_vnic_set_ethtool_ops - set ethtool ops */
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
index 5ce925f..d4fcf3b 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
@@ -169,6 +169,7 @@ struct __hfi_veswport_trap {
* @vport_num: vesw port number
* @lock: adapter lock
* @info: virtual ethernet switch port information
+ * @stats_lock: statistics lock
* @flow_tbl: flow to default port redirection table
*/
struct hfi_vnic_adapter {
@@ -184,6 +185,9 @@ struct hfi_vnic_adapter {
struct __hfi_veswport_info info;
+ /* Lock used to protect access to vnic counters */
+ struct mutex stats_lock;
+
u8 flow_tbl[HFI_VNIC_FLOW_TBL_SIZE];
};
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
index f04c8a834..ba045ac 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
@@ -59,6 +59,22 @@
#define HFI_VNIC_SKB_HEADROOM \
ALIGN((HFI_VNIC_HDR_LEN + HFI_VNIC_SKB_MDATA_LEN), 8)
+/* This function is overloaded for hfi_vnic specific implementation */
+static struct rtnl_link_stats64 *
+hfi_vnic_get_stats64(struct net_device *netdev,
+ struct rtnl_link_stats64 *stats)
+{
+ struct hfi_vnic_adapter *adapter = hfi_vnic_priv(netdev);
+ struct hfi_vnic_stats vstats;
+
+ memset(&vstats, 0, sizeof(vstats));
+ mutex_lock(&adapter->stats_lock);
+ adapter->rn_ops->ndo_get_stats64(netdev, &vstats.netstats);
+ mutex_unlock(&adapter->stats_lock);
+ memcpy(stats, &vstats.netstats, sizeof(*stats));
+ return stats;
+}
+
/* hfi_netdev_start_xmit - transmit function */
static netdev_tx_t hfi_netdev_start_xmit(struct sk_buff *skb,
struct net_device *netdev)
@@ -150,6 +166,7 @@ static int hfi_netdev_close(struct net_device *netdev)
.ndo_open = hfi_netdev_open,
.ndo_stop = hfi_netdev_close,
.ndo_start_xmit = hfi_netdev_start_xmit,
+ .ndo_get_stats64 = hfi_vnic_get_stats64,
.ndo_select_queue = hfi_vnic_select_queue,
.ndo_set_mac_address = hfi_vnic_set_mac_addr,
};
@@ -188,6 +205,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
netdev->priv_flags |= IFF_LIVE_ADDR_CHANGE;
netdev->hard_header_len += HFI_VNIC_SKB_HEADROOM;
mutex_init(&adapter->lock);
+ mutex_init(&adapter->stats_lock);
SET_NETDEV_DEV(netdev, ibdev->dma_device);
@@ -203,6 +221,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
return adapter;
netdev_err:
mutex_destroy(&adapter->lock);
+ mutex_destroy(&adapter->stats_lock);
kfree(adapter);
adapter_err:
ibdev->free_rdma_netdev(netdev);
@@ -219,6 +238,7 @@ void hfi_vnic_rem_netdev(struct hfi_vnic_adapter *adapter)
v_info("removing\n");
unregister_netdev(netdev);
mutex_destroy(&adapter->lock);
+ mutex_destroy(&adapter->stats_lock);
kfree(adapter);
ibdev->free_rdma_netdev(netdev);
}
--
1.8.3.1
^ permalink raw reply related [flat|nested] 40+ messages in thread* [RFC v3 06/11] IB/hfi-vnic: VNIC MAC table support
2017-02-07 20:22 [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
2017-02-07 20:23 ` [RFC v3 03/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev Vishwanathapura, Niranjana
2017-02-07 20:23 ` [RFC v3 05/11] IB/hfi-vnic: VNIC statistics support Vishwanathapura, Niranjana
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
2017-02-07 20:23 ` [RFC v3 07/11] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) interface Vishwanathapura, Niranjana
` (5 subsequent siblings)
8 siblings, 0 replies; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura, Sadanand Warrier
HFI VNIC MAC table contains the MAC address to DLID mappings provided by
the Ethernet manager. During transmission, the MAC table provides the MAC
address to DLID translation. Implement MAC table using simple hash list.
Also provide support to update/query the MAC table by Ethernet manager.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Sadanand Warrier <sadanand.warrier@intel.com>
---
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.c | 236 +++++++++++++++++++++
.../infiniband/ulp/hfi_vnic/hfi_vnic_internal.h | 53 ++++-
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c | 4 +
3 files changed, 292 insertions(+), 1 deletion(-)
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.c
index 159172c..dc4f97f 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.c
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.c
@@ -96,6 +96,238 @@ static inline void hfi_vnic_make_header(u8 *hdr, u32 slid, u32 dlid, u16 len,
memcpy(hdr, h, HFI_VNIC_HDR_LEN);
}
+/*
+ * Using a simple hash table for mac table implementation with the last octet
+ * of mac address as a key.
+ */
+static void hfi_vnic_free_mac_tbl(struct hlist_head *mactbl)
+{
+ struct hfi_vnic_mac_tbl_node *node;
+ struct hlist_node *tmp;
+ int bkt;
+
+ if (!mactbl)
+ return;
+
+ vnic_hash_for_each_safe(mactbl, bkt, tmp, node, hlist) {
+ hash_del(&node->hlist);
+ kfree(node);
+ }
+ kfree(mactbl);
+}
+
+static struct hlist_head *hfi_vnic_alloc_mac_tbl(void)
+{
+ u32 size = sizeof(struct hlist_head) * HFI_VNIC_MAC_TBL_SIZE;
+ struct hlist_head *mactbl;
+
+ mactbl = kzalloc(size, GFP_KERNEL);
+ if (!mactbl)
+ return ERR_PTR(-ENOMEM);
+
+ vnic_hash_init(mactbl);
+ return mactbl;
+}
+
+/* hfi_vnic_release_mac_tbl - empty and free the mac table */
+void hfi_vnic_release_mac_tbl(struct hfi_vnic_adapter *adapter)
+{
+ struct hlist_head *mactbl;
+
+ mutex_lock(&adapter->mactbl_lock);
+ mactbl = rcu_access_pointer(adapter->mactbl);
+ rcu_assign_pointer(adapter->mactbl, NULL);
+ synchronize_rcu();
+ hfi_vnic_free_mac_tbl(mactbl);
+ mutex_unlock(&adapter->mactbl_lock);
+}
+
+/*
+ * hfi_vnic_query_mac_tbl - query the mac table for a section
+ *
+ * This function implements query of specific function of the mac table.
+ * The function also expects the requested range to be valid.
+ */
+void hfi_vnic_query_mac_tbl(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_mactable *tbl)
+{
+ struct hfi_vnic_mac_tbl_node *node;
+ struct hlist_head *mactbl;
+ int bkt;
+ u16 loffset, lnum_entries;
+
+ rcu_read_lock();
+ mactbl = rcu_dereference(adapter->mactbl);
+ if (!mactbl)
+ goto get_mac_done;
+
+ loffset = be16_to_cpu(tbl->offset);
+ lnum_entries = be16_to_cpu(tbl->num_entries);
+
+ vnic_hash_for_each(mactbl, bkt, node, hlist) {
+ struct __hfi_vnic_mactable_entry *nentry = &node->entry;
+ struct hfi_veswport_mactable_entry *entry;
+
+ if ((node->index < loffset) ||
+ (node->index >= (loffset + lnum_entries)))
+ continue;
+
+ /* populate entry in the tbl corresponding to the index */
+ entry = &tbl->tbl_entries[node->index - loffset];
+ memcpy(entry->mac_addr, nentry->mac_addr,
+ ARRAY_SIZE(entry->mac_addr));
+ memcpy(entry->mac_addr_mask, nentry->mac_addr_mask,
+ ARRAY_SIZE(entry->mac_addr_mask));
+ entry->dlid_sd = cpu_to_be32(nentry->dlid_sd);
+ }
+ tbl->mac_tbl_digest = cpu_to_be32(adapter->info.vport.mac_tbl_digest);
+get_mac_done:
+ rcu_read_unlock();
+}
+
+/*
+ * hfi_vnic_update_mac_tbl - update mac table section
+ *
+ * This function updates the specified section of the mac table.
+ * The procedure includes following steps.
+ * - Allocate a new mac (hash) table.
+ * - Add the specified entries to the new table.
+ * (except the ones that are requested to be deleted).
+ * - Add all the other entries from the old mac table.
+ * - If there is a failure, free the new table and return.
+ * - Switch to the new table.
+ * - Free the old table and return.
+ *
+ * The function also expects the requested range to be valid.
+ */
+int hfi_vnic_update_mac_tbl(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_mactable *tbl)
+{
+ struct hfi_vnic_mac_tbl_node *node, *new_node;
+ struct hlist_head *new_mactbl, *old_mactbl;
+ int i, bkt, rc = 0;
+ u8 key;
+ u16 loffset, lnum_entries;
+
+ mutex_lock(&adapter->mactbl_lock);
+ /* allocate new mac table */
+ new_mactbl = hfi_vnic_alloc_mac_tbl();
+ if (IS_ERR(new_mactbl)) {
+ mutex_unlock(&adapter->mactbl_lock);
+ return PTR_ERR(new_mactbl);
+ }
+
+ loffset = be16_to_cpu(tbl->offset);
+ lnum_entries = be16_to_cpu(tbl->num_entries);
+
+ /* add updated entries to the new mac table */
+ for (i = 0; i < lnum_entries; i++) {
+ struct __hfi_vnic_mactable_entry *nentry;
+ struct hfi_veswport_mactable_entry *entry =
+ &tbl->tbl_entries[i];
+ u8 *mac_addr = entry->mac_addr;
+ u8 empty_mac[ETH_ALEN] = { 0 };
+
+ v_dbg("new mac entry %4d: %02x:%02x:%02x:%02x:%02x:%02x %x\n",
+ loffset + i, mac_addr[0], mac_addr[1], mac_addr[2],
+ mac_addr[3], mac_addr[4], mac_addr[5],
+ entry->dlid_sd);
+
+ /* if the entry is being removed, do not add it */
+ if (!memcmp(mac_addr, empty_mac, ARRAY_SIZE(empty_mac)))
+ continue;
+
+ node = kzalloc(sizeof(*node), GFP_KERNEL);
+ if (!node) {
+ rc = -ENOMEM;
+ goto updt_done;
+ }
+
+ node->index = loffset + i;
+ nentry = &node->entry;
+ memcpy(nentry->mac_addr, entry->mac_addr,
+ ARRAY_SIZE(nentry->mac_addr));
+ memcpy(nentry->mac_addr_mask, entry->mac_addr_mask,
+ ARRAY_SIZE(nentry->mac_addr_mask));
+ nentry->dlid_sd = be32_to_cpu(entry->dlid_sd);
+ key = node->entry.mac_addr[HFI_VNIC_MAC_HASH_IDX];
+ vnic_hash_add(new_mactbl, &node->hlist, key);
+ }
+
+ /* add other entries from current mac table to new mac table */
+ old_mactbl = rcu_access_pointer(adapter->mactbl);
+ if (!old_mactbl)
+ goto switch_tbl;
+
+ vnic_hash_for_each(old_mactbl, bkt, node, hlist) {
+ if ((node->index >= loffset) &&
+ (node->index < (loffset + lnum_entries)))
+ continue;
+
+ new_node = kzalloc(sizeof(*new_node), GFP_KERNEL);
+ if (!new_node) {
+ rc = -ENOMEM;
+ goto updt_done;
+ }
+
+ new_node->index = node->index;
+ memcpy(&new_node->entry, &node->entry, sizeof(node->entry));
+ key = new_node->entry.mac_addr[HFI_VNIC_MAC_HASH_IDX];
+ vnic_hash_add(new_mactbl, &new_node->hlist, key);
+ }
+
+switch_tbl:
+ /* switch to new table */
+ rcu_assign_pointer(adapter->mactbl, new_mactbl);
+ synchronize_rcu();
+
+ adapter->info.vport.mac_tbl_digest = be32_to_cpu(tbl->mac_tbl_digest);
+updt_done:
+ /* upon failure, free the new table; otherwise, free the old table */
+ if (rc)
+ hfi_vnic_free_mac_tbl(new_mactbl);
+ else
+ hfi_vnic_free_mac_tbl(old_mactbl);
+
+ mutex_unlock(&adapter->mactbl_lock);
+ return rc;
+}
+
+/* hfi_vnic_chk_mac_tbl - check mac table for dlid */
+static uint32_t hfi_vnic_chk_mac_tbl(struct hfi_vnic_adapter *adapter,
+ struct ethhdr *mac_hdr)
+{
+ struct hfi_vnic_mac_tbl_node *node;
+ struct hlist_head *mactbl;
+ u32 dlid = 0;
+ u8 key;
+
+ rcu_read_lock();
+ mactbl = rcu_dereference(adapter->mactbl);
+ if (!mactbl)
+ goto chk_done;
+
+ key = mac_hdr->h_dest[HFI_VNIC_MAC_HASH_IDX];
+ vnic_hash_for_each_possible(mactbl, node, hlist, key) {
+ struct __hfi_vnic_mactable_entry *entry = &node->entry;
+
+ /* if related to source mac, skip */
+ if (HFI_VNIC_DLID_SD_IS_SRC_MAC(entry->dlid_sd))
+ continue;
+
+ if (!memcmp(node->entry.mac_addr, mac_hdr->h_dest,
+ ARRAY_SIZE(node->entry.mac_addr))) {
+ /* mac address found */
+ dlid = HFI_VNIC_DLID_SD_GET_DLID(node->entry.dlid_sd);
+ break;
+ }
+ }
+
+chk_done:
+ rcu_read_unlock();
+ return dlid;
+}
+
/* hfi_vnic_get_dlid - find and return the DLID */
static uint32_t hfi_vnic_get_dlid(struct hfi_vnic_adapter *adapter,
struct sk_buff *skb, u8 def_port)
@@ -104,6 +336,10 @@ static uint32_t hfi_vnic_get_dlid(struct hfi_vnic_adapter *adapter,
struct ethhdr *mac_hdr = (struct ethhdr *)skb_mac_header(skb);
u32 dlid;
+ dlid = hfi_vnic_chk_mac_tbl(adapter, mac_hdr);
+ if (dlid)
+ return dlid;
+
if (is_multicast_ether_addr(mac_hdr->h_dest)) {
dlid = info->vesw.u_mcast_dlid;
} else {
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
index d4fcf3b..a94dd2a 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
@@ -169,6 +169,8 @@ struct __hfi_veswport_trap {
* @vport_num: vesw port number
* @lock: adapter lock
* @info: virtual ethernet switch port information
+ * @mactbl: hash table of MAC entries
+ * @mactbl_lock: mac table lock
* @stats_lock: statistics lock
* @flow_tbl: flow to default port redirection table
*/
@@ -183,7 +185,11 @@ struct hfi_vnic_adapter {
/* Lock used around concurrent updates to netdev */
struct mutex lock;
- struct __hfi_veswport_info info;
+ struct __hfi_veswport_info info;
+ struct hlist_head __rcu *mactbl;
+
+ /* Lock used to protect updates to mac table */
+ struct mutex mactbl_lock;
/* Lock used to protect access to vnic counters */
struct mutex stats_lock;
@@ -191,6 +197,25 @@ struct hfi_vnic_adapter {
u8 flow_tbl[HFI_VNIC_FLOW_TBL_SIZE];
};
+/* Same as hfi_veswport_mactable_entry, but without bitwise attribute */
+struct __hfi_vnic_mactable_entry {
+ u8 mac_addr[ETH_ALEN];
+ u8 mac_addr_mask[ETH_ALEN];
+ u32 dlid_sd;
+} __packed;
+
+/**
+ * struct hfi_vnic_mac_tbl_node - HFI VNIC mac table node
+ * @hlist: hash list handle
+ * @index: index of entry in the mac table
+ * @entry: entry in the table
+ */
+struct hfi_vnic_mac_tbl_node {
+ struct hlist_node hlist;
+ u16 index;
+ struct __hfi_vnic_mactable_entry entry;
+};
+
#define v_dbg(format, arg...) \
netdev_dbg(adapter->netdev, format, ## arg)
#define v_err(format, arg...) \
@@ -212,12 +237,38 @@ struct hfi_vnic_adapter {
#define HFI_VNIC_MAC_TBL_HASH_BITS 8
#define HFI_VNIC_MAC_TBL_SIZE BIT(HFI_VNIC_MAC_TBL_HASH_BITS)
+/* VNIC HASH MACROS */
+#define vnic_hash_init(hashtable) __hash_init(hashtable, HFI_VNIC_MAC_TBL_SIZE)
+
+#define vnic_hash_add(hashtable, node, key) \
+ hlist_add_head(node, \
+ &hashtable[hash_min(key, ilog2(HFI_VNIC_MAC_TBL_SIZE))])
+
+#define vnic_hash_for_each_safe(name, bkt, tmp, obj, member) \
+ for ((bkt) = 0, obj = NULL; \
+ !obj && (bkt) < HFI_VNIC_MAC_TBL_SIZE; (bkt)++) \
+ hlist_for_each_entry_safe(obj, tmp, &name[bkt], member)
+
+#define vnic_hash_for_each_possible(name, obj, member, key) \
+ hlist_for_each_entry(obj, \
+ &name[hash_min(key, ilog2(HFI_VNIC_MAC_TBL_SIZE))], member)
+
+#define vnic_hash_for_each(name, bkt, obj, member) \
+ for ((bkt) = 0, obj = NULL; \
+ !obj && (bkt) < HFI_VNIC_MAC_TBL_SIZE; (bkt)++) \
+ hlist_for_each_entry(obj, &name[bkt], member)
+
struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
u8 port_num, u8 vport_num);
void hfi_vnic_rem_netdev(struct hfi_vnic_adapter *adapter);
void hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
u8 hfi_vnic_get_vl(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
u8 hfi_vnic_calc_entropy(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
+void hfi_vnic_release_mac_tbl(struct hfi_vnic_adapter *adapter);
+void hfi_vnic_query_mac_tbl(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_mactable *tbl);
+int hfi_vnic_update_mac_tbl(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_mactable *tbl);
void hfi_vnic_set_ethtool_ops(struct net_device *netdev);
#endif /* _HFI_VNIC_INTERNAL_H */
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
index ba045ac..6b8054f 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
@@ -205,6 +205,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
netdev->priv_flags |= IFF_LIVE_ADDR_CHANGE;
netdev->hard_header_len += HFI_VNIC_SKB_HEADROOM;
mutex_init(&adapter->lock);
+ mutex_init(&adapter->mactbl_lock);
mutex_init(&adapter->stats_lock);
SET_NETDEV_DEV(netdev, ibdev->dma_device);
@@ -221,6 +222,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
return adapter;
netdev_err:
mutex_destroy(&adapter->lock);
+ mutex_destroy(&adapter->mactbl_lock);
mutex_destroy(&adapter->stats_lock);
kfree(adapter);
adapter_err:
@@ -237,7 +239,9 @@ void hfi_vnic_rem_netdev(struct hfi_vnic_adapter *adapter)
v_info("removing\n");
unregister_netdev(netdev);
+ hfi_vnic_release_mac_tbl(adapter);
mutex_destroy(&adapter->lock);
+ mutex_destroy(&adapter->mactbl_lock);
mutex_destroy(&adapter->stats_lock);
kfree(adapter);
ibdev->free_rdma_netdev(netdev);
--
1.8.3.1
^ permalink raw reply related [flat|nested] 40+ messages in thread* [RFC v3 07/11] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) interface
2017-02-07 20:22 [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
` (2 preceding siblings ...)
2017-02-07 20:23 ` [RFC v3 06/11] IB/hfi-vnic: VNIC MAC table support Vishwanathapura, Niranjana
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
2017-02-07 20:23 ` [RFC v3 08/11] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) function Vishwanathapura, Niranjana
` (4 subsequent siblings)
8 siblings, 0 replies; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura, Sadanand Warrier, Tanya K Jajodia
HFI VNIC EMA interface functions are the management interfaces to the HFI
VNIC netdev. Add support to add and remove VNIC ports. Implement the
required GET/SET management interface functions and processing of new
management information. Add support to send trap notifications upon various
events like interface status change, unicast/multicast mac list update and
mac address change.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Sadanand Warrier <sadanand.warrier@intel.com>
Signed-off-by: Tanya K Jajodia <tanya.k.jajodia@intel.com>
---
drivers/infiniband/ulp/hfi_vnic/Makefile | 3 +-
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h | 4 +
.../infiniband/ulp/hfi_vnic/hfi_vnic_internal.h | 44 +++
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c | 144 +++++++-
.../infiniband/ulp/hfi_vnic/hfi_vnic_vema_iface.c | 390 +++++++++++++++++++++
5 files changed, 582 insertions(+), 3 deletions(-)
create mode 100644 drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema_iface.c
diff --git a/drivers/infiniband/ulp/hfi_vnic/Makefile b/drivers/infiniband/ulp/hfi_vnic/Makefile
index ad7db32..32d6821 100644
--- a/drivers/infiniband/ulp/hfi_vnic/Makefile
+++ b/drivers/infiniband/ulp/hfi_vnic/Makefile
@@ -3,4 +3,5 @@
#
obj-$(CONFIG_INFINIBAND_HFI_VNIC) += hfi_vnic.o
-hfi_vnic-y := hfi_vnic_netdev.o hfi_vnic_encap.o hfi_vnic_ethtool.o
+hfi_vnic-y := hfi_vnic_netdev.o hfi_vnic_encap.o hfi_vnic_ethtool.o \
+ hfi_vnic_vema_iface.o
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h
index 73c2bdc..be6d870 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h
@@ -95,6 +95,10 @@
#define HFI_VNIC_DLID_SD_IS_SRC_MAC(dlid_sd) (!!((dlid_sd) & 0x20))
#define HFI_VNIC_DLID_SD_GET_DLID(dlid_sd) ((dlid_sd) >> 8)
+/* VNIC Ethernet link status */
+#define HFI_VNIC_ETH_LINK_UP 1
+#define HFI_VNIC_ETH_LINK_DOWN 2
+
/**
* struct hfi_vesw_info - HFI vnic switch information
* @fabric_id: 10-bit fabric id
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
index a94dd2a..2ee80a7 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
@@ -161,14 +161,28 @@ struct __hfi_veswport_trap {
} __packed;
/**
+ * struct hfi_vnic_ctrl_port - HFI virtual NIC control port
+ * @ibdev: pointer to ib device
+ * @ops: hfi vnic control operations
+ */
+struct hfi_vnic_ctrl_port {
+ struct ib_device *ibdev;
+ struct hfi_vnic_ctrl_ops *ops;
+};
+
+/**
* struct hfi_vnic_adapter - HFI VNIC netdev private data structure
* @netdev: pointer to associated netdev
* @ibdev: ib device
+ * @cport: pointer to hfi vnic control port
* @rn_ops: rdma netdev's net_device_ops
* @port_num: OPA port number
* @vport_num: vesw port number
* @lock: adapter lock
* @info: virtual ethernet switch port information
+ * @vema_mac_addr: mac address configured by vema
+ * @umac_hash: unicast maclist hash
+ * @mmac_hash: multicast maclist hash
* @mactbl: hash table of MAC entries
* @mactbl_lock: mac table lock
* @stats_lock: statistics lock
@@ -177,6 +191,7 @@ struct __hfi_veswport_trap {
struct hfi_vnic_adapter {
struct net_device *netdev;
struct ib_device *ibdev;
+ struct hfi_vnic_ctrl_port *cport;
const struct net_device_ops *rn_ops;
u8 port_num;
@@ -186,6 +201,9 @@ struct hfi_vnic_adapter {
struct mutex lock;
struct __hfi_veswport_info info;
+ u8 vema_mac_addr[ETH_ALEN];
+ u32 umac_hash;
+ u32 mmac_hash;
struct hlist_head __rcu *mactbl;
/* Lock used to protect updates to mac table */
@@ -225,6 +243,11 @@ struct hfi_vnic_mac_tbl_node {
#define v_warn(format, arg...) \
netdev_warn(adapter->netdev, format, ## arg)
+#define c_err(format, arg...) \
+ dev_err(&cport->ibdev->dev, format, ## arg)
+#define c_info(format, arg...) \
+ dev_info(&cport->ibdev->dev, format, ## arg)
+
/* The maximum allowed entries in the mac table */
#define HFI_VNIC_MAC_TBL_MAX_ENTRIES 2048
/* Limit of smac entries in mac table */
@@ -264,11 +287,32 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
void hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
u8 hfi_vnic_get_vl(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
u8 hfi_vnic_calc_entropy(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
+void hfi_vnic_process_vema_config(struct hfi_vnic_adapter *adapter);
void hfi_vnic_release_mac_tbl(struct hfi_vnic_adapter *adapter);
void hfi_vnic_query_mac_tbl(struct hfi_vnic_adapter *adapter,
struct hfi_veswport_mactable *tbl);
int hfi_vnic_update_mac_tbl(struct hfi_vnic_adapter *adapter,
struct hfi_veswport_mactable *tbl);
+void hfi_vnic_query_ucast_macs(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_iface_macs *macs);
+void hfi_vnic_query_mcast_macs(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_iface_macs *macs);
+void hfi_vnic_get_summary_counters(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_summary_counters *cntrs);
+void hfi_vnic_get_error_counters(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_error_counters *cntrs);
+void hfi_vnic_get_vesw_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_vesw_info *info);
+void hfi_vnic_set_vesw_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_vesw_info *info);
+void hfi_vnic_get_per_veswport_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_per_veswport_info *info);
+void hfi_vnic_set_per_veswport_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_per_veswport_info *info);
+void hfi_vnic_vema_report_event(struct hfi_vnic_adapter *adapter, u8 event);
+struct hfi_vnic_adapter *hfi_vnic_add_vport(struct hfi_vnic_ctrl_port *cport,
+ u8 port_num, u8 vport_num);
+void hfi_vnic_rem_vport(struct hfi_vnic_adapter *adapter);
void hfi_vnic_set_ethtool_ops(struct net_device *netdev);
#endif /* _HFI_VNIC_INTERNAL_H */
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
index 6b8054f..9a9167e 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_netdev.c
@@ -51,6 +51,7 @@
#include <linux/module.h>
#include <linux/if_vlan.h>
+#include <linux/crc32.h>
#include "hfi_vnic_internal.h"
@@ -112,7 +113,80 @@ static u16 hfi_vnic_select_queue(struct net_device *netdev, struct sk_buff *skb,
return rc;
}
-/* hfi_vnic_set_mac_addr - change mac address */
+/* hfi_vnic_process_vema_config - process vema configuration updates */
+void hfi_vnic_process_vema_config(struct hfi_vnic_adapter *adapter)
+{
+ struct __hfi_veswport_info *info = &adapter->info;
+ struct rdma_netdev *rn = netdev_priv(adapter->netdev);
+ u8 port_num[HFI_VESW_MAX_NUM_DEF_PORT] = { 0 };
+ struct net_device *netdev = adapter->netdev;
+ u8 i, port_count = 0;
+ u16 port_mask;
+
+ /* If the base_mac_addr is changed, update the interface mac address */
+ if (memcmp(info->vport.base_mac_addr, adapter->vema_mac_addr,
+ ARRAY_SIZE(info->vport.base_mac_addr))) {
+ struct sockaddr saddr;
+
+ memcpy(saddr.sa_data, info->vport.base_mac_addr,
+ ARRAY_SIZE(info->vport.base_mac_addr));
+ mutex_lock(&adapter->lock);
+ eth_mac_addr(netdev, &saddr);
+ memcpy(adapter->vema_mac_addr,
+ info->vport.base_mac_addr, ETH_ALEN);
+ mutex_unlock(&adapter->lock);
+ }
+
+ rn->set_id(netdev, info->vesw.vesw_id);
+
+ /* Handle MTU limit change */
+ rtnl_lock();
+ netdev->max_mtu = max_t(unsigned int, info->vesw.eth_mtu_non_vlan,
+ netdev->min_mtu);
+ if (netdev->mtu > netdev->max_mtu)
+ dev_set_mtu(netdev, netdev->max_mtu);
+ rtnl_unlock();
+
+ /* Update flow to default port redirection table */
+ port_mask = info->vesw.def_port_mask;
+ for (i = 0; i < HFI_VESW_MAX_NUM_DEF_PORT; i++) {
+ if (port_mask & 1)
+ port_num[port_count++] = i;
+ port_mask >>= 1;
+ }
+
+ /*
+ * Build the flow table. Flow table is required when destination LID
+ * is not available. Up to HFI_VNIC_FLOW_TBL_SIZE flows supported.
+ * Each flow need a default port number to get its dlid from the
+ * u_ucast_dlid array.
+ */
+ for (i = 0; i < HFI_VNIC_FLOW_TBL_SIZE; i++)
+ adapter->flow_tbl[i] = port_count ? port_num[i % port_count] :
+ HFI_VNIC_INVALID_PORT;
+
+ /* Operational state can only be DROP_ALL or FORWARDING */
+ if (info->vport.config_state == HFI_VNIC_STATE_FORWARDING) {
+ info->vport.oper_state = HFI_VNIC_STATE_FORWARDING;
+ netif_dormant_off(netdev);
+ } else {
+ info->vport.oper_state = HFI_VNIC_STATE_DROP_ALL;
+ netif_dormant_on(netdev);
+ }
+}
+
+/*
+ * Set the power on default values in adapter's vema interface structure.
+ */
+static inline void hfi_vnic_set_pod_values(struct hfi_vnic_adapter *adapter)
+{
+ adapter->info.vport.max_mac_tbl_ent = HFI_VNIC_MAC_TBL_MAX_ENTRIES;
+ adapter->info.vport.max_smac_ent = HFI_VNIC_MAX_SMAC_LIMIT;
+ adapter->info.vport.config_state = HFI_VNIC_STATE_DROP_ALL;
+ adapter->info.vport.eth_link_status = HFI_VNIC_ETH_LINK_DOWN;
+}
+
+/* hfi_vnic_set_mac_addr - change mac address and send trap */
static int hfi_vnic_set_mac_addr(struct net_device *netdev, void *addr)
{
struct hfi_vnic_adapter *adapter = hfi_vnic_priv(netdev);
@@ -125,8 +199,62 @@ static int hfi_vnic_set_mac_addr(struct net_device *netdev, void *addr)
mutex_lock(&adapter->lock);
rc = eth_mac_addr(netdev, addr);
mutex_unlock(&adapter->lock);
+ if (rc)
+ return rc;
- return rc;
+ adapter->info.vport.uc_macs_gen_count++;
+ hfi_vnic_vema_report_event(adapter,
+ HFI_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE);
+ return 0;
+}
+
+/*
+ * hfi_vnic_mac_send_event - post event on possible mac list exchange
+ * Send trap when digest from uc/mc mac list differs from previous run.
+ * Digest is evaluated similar to how cksum does.
+ */
+static void hfi_vnic_mac_send_event(struct net_device *netdev, u8 event)
+{
+ struct hfi_vnic_adapter *adapter = hfi_vnic_priv(netdev);
+ struct netdev_hw_addr *ha;
+ struct netdev_hw_addr_list *hw_list;
+ u32 *ref_crc;
+ u32 l, crc = 0;
+
+ switch (event) {
+ case HFI_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE:
+ hw_list = &netdev->uc;
+ adapter->info.vport.uc_macs_gen_count++;
+ ref_crc = &adapter->umac_hash;
+ break;
+ case HFI_VESWPORT_TRAP_IFACE_MCAST_MAC_CHANGE:
+ hw_list = &netdev->mc;
+ adapter->info.vport.mc_macs_gen_count++;
+ ref_crc = &adapter->mmac_hash;
+ break;
+ default:
+ return;
+ }
+ netdev_hw_addr_list_for_each(ha, hw_list) {
+ crc = crc32_le(crc, ha->addr, ETH_ALEN);
+ }
+ l = netdev_hw_addr_list_count(hw_list) * ETH_ALEN;
+ crc = ~crc32_le(crc, (void *)&l, sizeof(l));
+
+ if (crc != *ref_crc) {
+ *ref_crc = crc;
+ hfi_vnic_vema_report_event(adapter, event);
+ }
+}
+
+/* hfi_vnic_set_rx_mode - handle uc/mc mac list change */
+static void hfi_vnic_set_rx_mode(struct net_device *netdev)
+{
+ hfi_vnic_mac_send_event(netdev,
+ HFI_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE);
+
+ hfi_vnic_mac_send_event(netdev,
+ HFI_VESWPORT_TRAP_IFACE_MCAST_MAC_CHANGE);
}
/* hfi_netdev_open - activate network interface */
@@ -141,6 +269,10 @@ static int hfi_netdev_open(struct net_device *netdev)
return rc;
}
+ /* Update eth link status and send trap */
+ adapter->info.vport.eth_link_status = HFI_VNIC_ETH_LINK_UP;
+ hfi_vnic_vema_report_event(adapter,
+ HFI_VESWPORT_TRAP_ETH_LINK_STATUS_CHANGE);
v_info("opened\n");
return 0;
}
@@ -157,6 +289,10 @@ static int hfi_netdev_close(struct net_device *netdev)
return rc;
}
+ /* Update eth link status and send trap */
+ adapter->info.vport.eth_link_status = HFI_VNIC_ETH_LINK_DOWN;
+ hfi_vnic_vema_report_event(adapter,
+ HFI_VESWPORT_TRAP_ETH_LINK_STATUS_CHANGE);
v_info("closed\n");
return 0;
}
@@ -167,6 +303,7 @@ static int hfi_netdev_close(struct net_device *netdev)
.ndo_stop = hfi_netdev_close,
.ndo_start_xmit = hfi_netdev_start_xmit,
.ndo_get_stats64 = hfi_vnic_get_stats64,
+ .ndo_set_rx_mode = hfi_vnic_set_rx_mode,
.ndo_select_queue = hfi_vnic_select_queue,
.ndo_set_mac_address = hfi_vnic_set_mac_addr,
};
@@ -211,6 +348,9 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
SET_NETDEV_DEV(netdev, ibdev->dma_device);
hfi_vnic_set_ethtool_ops(netdev);
+
+ hfi_vnic_set_pod_values(adapter);
+
rc = register_netdev(netdev);
if (rc)
goto netdev_err;
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema_iface.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema_iface.c
new file mode 100644
index 0000000..fca3bab
--- /dev/null
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema_iface.c
@@ -0,0 +1,390 @@
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * This file contains HFI VNIC EMA Interface functions.
+ */
+
+#include "hfi_vnic_internal.h"
+
+/**
+ * hfi_vnic_vema_report_event - sent trap to report the specified event
+ * @adapter: vnic port adapter
+ * @event: event to be reported
+ *
+ * This function calls vema api to sent a trap for the given event.
+ */
+void hfi_vnic_vema_report_event(struct hfi_vnic_adapter *adapter, u8 event)
+{
+ struct __hfi_veswport_info *info = &adapter->info;
+ struct __hfi_veswport_trap trap_data;
+
+ trap_data.fabric_id = info->vesw.fabric_id;
+ trap_data.veswid = info->vesw.vesw_id;
+ trap_data.veswportnum = info->vport.port_num;
+ trap_data.hfiportnum = adapter->port_num;
+ trap_data.veswportindex = adapter->vport_num;
+ trap_data.opcode = event;
+
+ /* Need to send trap here */
+}
+
+/**
+ * hfi_vnic_get_error_counters - get summary counters
+ * @adapter: vnic port adapter
+ * @cntrs: pointer to destination summary counters structure
+ *
+ * This function populates the summary counters that is maintained by the
+ * given adapter to destination address provided.
+ */
+void hfi_vnic_get_summary_counters(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_summary_counters *cntrs)
+{
+ struct hfi_vnic_stats vstats;
+ __be64 *dst;
+ u64 *src;
+
+ memset(&vstats, 0, sizeof(vstats));
+ mutex_lock(&adapter->stats_lock);
+ adapter->rn_ops->ndo_get_stats64(adapter->netdev, &vstats.netstats);
+ mutex_unlock(&adapter->stats_lock);
+
+ cntrs->vp_instance = cpu_to_be16(adapter->vport_num);
+ cntrs->vesw_id = cpu_to_be16(adapter->info.vesw.vesw_id);
+ cntrs->veswport_num = cpu_to_be32(adapter->port_num);
+
+ cntrs->tx_errors = cpu_to_be64(vstats.netstats.tx_errors);
+ cntrs->rx_errors = cpu_to_be64(vstats.netstats.rx_errors);
+ cntrs->tx_packets = cpu_to_be64(vstats.netstats.tx_packets);
+ cntrs->rx_packets = cpu_to_be64(vstats.netstats.rx_packets);
+ cntrs->tx_bytes = cpu_to_be64(vstats.netstats.tx_bytes);
+ cntrs->rx_bytes = cpu_to_be64(vstats.netstats.rx_bytes);
+
+ /*
+ * This loop depends on layout of
+ * hfi_veswport_summary_counters hfi_vnic_stats structures.
+ */
+ for (dst = &cntrs->tx_unicast, src = &vstats.tx_grp.unicast;
+ dst < &cntrs->reserved[0]; dst++, src++) {
+ *dst = cpu_to_be64(*src);
+ }
+}
+
+/**
+ * hfi_vnic_get_error_counters - get error counters
+ * @adapter: vnic port adapter
+ * @cntrs: pointer to destination error counters structure
+ *
+ * This function populates the error counters that is maintained by the
+ * given adapter to destination address provided.
+ */
+void hfi_vnic_get_error_counters(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_error_counters *cntrs)
+{
+ struct hfi_vnic_stats vstats;
+
+ memset(&vstats, 0, sizeof(vstats));
+ mutex_lock(&adapter->stats_lock);
+ adapter->rn_ops->ndo_get_stats64(adapter->netdev, &vstats.netstats);
+ mutex_unlock(&adapter->stats_lock);
+
+ cntrs->vp_instance = cpu_to_be16(adapter->vport_num);
+ cntrs->vesw_id = cpu_to_be16(adapter->info.vesw.vesw_id);
+ cntrs->veswport_num = cpu_to_be32(adapter->port_num);
+
+ cntrs->tx_errors = cpu_to_be64(vstats.netstats.tx_errors);
+ cntrs->rx_errors = cpu_to_be64(vstats.netstats.rx_errors);
+ cntrs->tx_dlid_zero = cpu_to_be64(vstats.tx_dlid_zero);
+ cntrs->tx_drop_state = cpu_to_be64(vstats.tx_drop_state);
+ cntrs->tx_logic = cpu_to_be64(vstats.netstats.tx_fifo_errors +
+ vstats.netstats.tx_carrier_errors);
+
+ cntrs->rx_bad_veswid = cpu_to_be64(vstats.netstats.rx_nohandler);
+ cntrs->rx_runt = cpu_to_be64(vstats.rx_runt);
+ cntrs->rx_oversize = cpu_to_be64(vstats.rx_oversize);
+ cntrs->rx_drop_state = cpu_to_be64(vstats.rx_drop_state);
+ cntrs->rx_logic = cpu_to_be64(vstats.netstats.rx_fifo_errors);
+}
+
+/**
+ * hfi_vnic_get_vesw_info -- Get the vesw information
+ * @adapter: vnic port adapter
+ * @info: pointer to destination vesw info structure
+ *
+ * This function copies the vesw info that is maintained by the
+ * given adapter to destination address provided.
+ */
+void hfi_vnic_get_vesw_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_vesw_info *info)
+{
+ struct __hfi_vesw_info *src = &adapter->info.vesw;
+ int i;
+
+ info->fabric_id = cpu_to_be16(src->fabric_id);
+ info->vesw_id = cpu_to_be16(src->vesw_id);
+ memcpy(info->rsvd0, src->rsvd0, ARRAY_SIZE(src->rsvd0));
+ info->def_port_mask = cpu_to_be16(src->def_port_mask);
+ memcpy(info->rsvd1, src->rsvd1, ARRAY_SIZE(src->rsvd1));
+ info->pkey = cpu_to_be16(src->pkey);
+
+ memcpy(info->rsvd2, src->rsvd2, ARRAY_SIZE(src->rsvd2));
+ info->u_mcast_dlid = cpu_to_be32(src->u_mcast_dlid);
+ for (i = 0; i < HFI_VESW_MAX_NUM_DEF_PORT; i++)
+ info->u_ucast_dlid[i] = cpu_to_be32(src->u_ucast_dlid[i]);
+
+ memcpy(info->rsvd3, src->rsvd3, ARRAY_SIZE(src->rsvd3));
+ for (i = 0; i < HFI_VNIC_MAX_NUM_PCP; i++)
+ info->eth_mtu[i] = cpu_to_be16(src->eth_mtu[i]);
+
+ info->eth_mtu_non_vlan = cpu_to_be16(src->eth_mtu_non_vlan);
+ memcpy(info->rsvd4, src->rsvd4, ARRAY_SIZE(src->rsvd4));
+}
+
+/**
+ * hfi_vnic_set_vesw_info -- Set the vesw information
+ * @adapter: vnic port adapter
+ * @info: pointer to vesw info structure
+ *
+ * This function updates the vesw info that is maintained by the
+ * given adapter with vesw info provided. Reserved fields are stored
+ * and returned back to EM as is.
+ */
+void hfi_vnic_set_vesw_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_vesw_info *info)
+{
+ struct __hfi_vesw_info *dst = &adapter->info.vesw;
+ int i;
+
+ dst->fabric_id = be16_to_cpu(info->fabric_id);
+ dst->vesw_id = be16_to_cpu(info->vesw_id);
+ memcpy(dst->rsvd0, info->rsvd0, ARRAY_SIZE(info->rsvd0));
+ dst->def_port_mask = be16_to_cpu(info->def_port_mask);
+ memcpy(dst->rsvd1, info->rsvd1, ARRAY_SIZE(info->rsvd1));
+ dst->pkey = be16_to_cpu(info->pkey);
+
+ memcpy(dst->rsvd2, info->rsvd2, ARRAY_SIZE(info->rsvd2));
+ dst->u_mcast_dlid = be32_to_cpu(info->u_mcast_dlid);
+ for (i = 0; i < HFI_VESW_MAX_NUM_DEF_PORT; i++)
+ dst->u_ucast_dlid[i] = be32_to_cpu(info->u_ucast_dlid[i]);
+
+ memcpy(dst->rsvd3, info->rsvd3, ARRAY_SIZE(info->rsvd3));
+ for (i = 0; i < HFI_VNIC_MAX_NUM_PCP; i++)
+ dst->eth_mtu[i] = be16_to_cpu(info->eth_mtu[i]);
+
+ dst->eth_mtu_non_vlan = be16_to_cpu(info->eth_mtu_non_vlan);
+ memcpy(dst->rsvd4, info->rsvd4, ARRAY_SIZE(info->rsvd4));
+}
+
+/**
+ * hfi_vnic_get_per_veswport_info -- Get the vesw per port information
+ * @adapter: vnic port adapter
+ * @info: pointer to destination vport info structure
+ *
+ * This function copies the vesw per port info that is maintained by the
+ * given adapter to destination address provided.
+ * Note that the read only fields are not copied.
+ */
+void hfi_vnic_get_per_veswport_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_per_veswport_info *info)
+{
+ struct __hfi_per_veswport_info *src = &adapter->info.vport;
+
+ info->port_num = cpu_to_be32(src->port_num);
+ info->eth_link_status = src->eth_link_status;
+ memcpy(info->rsvd0, src->rsvd0, ARRAY_SIZE(src->rsvd0));
+
+ memcpy(info->base_mac_addr, src->base_mac_addr,
+ ARRAY_SIZE(info->base_mac_addr));
+ info->config_state = src->config_state;
+ info->oper_state = src->oper_state;
+ info->max_mac_tbl_ent = cpu_to_be16(src->max_mac_tbl_ent);
+ info->max_smac_ent = cpu_to_be16(src->max_smac_ent);
+ info->mac_tbl_digest = cpu_to_be32(src->mac_tbl_digest);
+ memcpy(info->rsvd1, src->rsvd1, ARRAY_SIZE(src->rsvd1));
+
+ info->encap_slid = cpu_to_be32(src->encap_slid);
+ memcpy(info->pcp_to_sc_uc, src->pcp_to_sc_uc,
+ ARRAY_SIZE(info->pcp_to_sc_uc));
+ memcpy(info->pcp_to_vl_uc, src->pcp_to_vl_uc,
+ ARRAY_SIZE(info->pcp_to_vl_uc));
+ memcpy(info->pcp_to_sc_mc, src->pcp_to_sc_mc,
+ ARRAY_SIZE(info->pcp_to_sc_mc));
+ memcpy(info->pcp_to_vl_mc, src->pcp_to_vl_mc,
+ ARRAY_SIZE(info->pcp_to_vl_mc));
+ info->non_vlan_sc_uc = src->non_vlan_sc_uc;
+ info->non_vlan_vl_uc = src->non_vlan_vl_uc;
+ info->non_vlan_sc_mc = src->non_vlan_sc_mc;
+ info->non_vlan_vl_mc = src->non_vlan_vl_mc;
+ memcpy(info->rsvd2, src->rsvd2, ARRAY_SIZE(src->rsvd2));
+
+ info->uc_macs_gen_count = cpu_to_be16(src->uc_macs_gen_count);
+ info->mc_macs_gen_count = cpu_to_be16(src->mc_macs_gen_count);
+ memcpy(info->rsvd3, src->rsvd3, ARRAY_SIZE(src->rsvd3));
+}
+
+/**
+ * hfi_vnic_set_per_veswport_info -- Set vesw per port information
+ * @adapter: vnic port adapter
+ * @info: pointer to vport info structure
+ *
+ * This function updates the vesw per port info that is maintained by the
+ * given adapter with vesw per port info provided. Reserved fields are
+ * stored and returned back to EM as is.
+ */
+void hfi_vnic_set_per_veswport_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_per_veswport_info *info)
+{
+ struct __hfi_per_veswport_info *dst = &adapter->info.vport;
+
+ dst->port_num = be32_to_cpu(info->port_num);
+ memcpy(dst->rsvd0, info->rsvd0, ARRAY_SIZE(info->rsvd0));
+
+ memcpy(dst->base_mac_addr, info->base_mac_addr,
+ ARRAY_SIZE(dst->base_mac_addr));
+ dst->config_state = info->config_state;
+ memcpy(dst->rsvd1, info->rsvd1, ARRAY_SIZE(info->rsvd1));
+
+ dst->encap_slid = be32_to_cpu(info->encap_slid);
+ memcpy(dst->pcp_to_sc_uc, info->pcp_to_sc_uc,
+ ARRAY_SIZE(dst->pcp_to_sc_uc));
+ memcpy(dst->pcp_to_vl_uc, info->pcp_to_vl_uc,
+ ARRAY_SIZE(dst->pcp_to_vl_uc));
+ memcpy(dst->pcp_to_sc_mc, info->pcp_to_sc_mc,
+ ARRAY_SIZE(dst->pcp_to_sc_mc));
+ memcpy(dst->pcp_to_vl_mc, info->pcp_to_vl_mc,
+ ARRAY_SIZE(dst->pcp_to_vl_mc));
+ dst->non_vlan_sc_uc = info->non_vlan_sc_uc;
+ dst->non_vlan_vl_uc = info->non_vlan_vl_uc;
+ dst->non_vlan_sc_mc = info->non_vlan_sc_mc;
+ dst->non_vlan_vl_mc = info->non_vlan_vl_mc;
+ memcpy(dst->rsvd2, info->rsvd2, ARRAY_SIZE(info->rsvd2));
+ memcpy(dst->rsvd3, info->rsvd3, ARRAY_SIZE(info->rsvd3));
+}
+
+/**
+ * hfi_vnic_query_mcast_macs - query multicast mac list
+ * @adapter: vnic port adapter
+ * @macs: pointer mac list
+ *
+ * This function populates the provided mac list with the configured
+ * multicast addresses in the adapter.
+ */
+void hfi_vnic_query_mcast_macs(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_iface_macs *macs)
+{
+ u16 start_idx, num_macs, idx = 0, count = 0;
+ struct netdev_hw_addr *ha;
+
+ start_idx = be16_to_cpu(macs->start_idx);
+ num_macs = be16_to_cpu(macs->num_macs_in_msg);
+ netdev_for_each_mc_addr(ha, adapter->netdev) {
+ struct hfi_vnic_iface_mac_entry *entry = &macs->entry[count];
+
+ if (start_idx > idx++)
+ continue;
+ else if (num_macs == count)
+ break;
+ memcpy(entry, ha->addr, sizeof(*entry));
+ count++;
+ }
+
+ macs->tot_macs_in_lst = cpu_to_be16(netdev_mc_count(adapter->netdev));
+ macs->num_macs_in_msg = cpu_to_be16(count);
+ macs->gen_count = cpu_to_be16(adapter->info.vport.mc_macs_gen_count);
+}
+
+/**
+ * hfi_vnic_query_ucast_macs - query unicast mac list
+ * @adapter: vnic port adapter
+ * @macs: pointer mac list
+ *
+ * This function populates the provided mac list with the configured
+ * unicast addresses in the adapter.
+ */
+void hfi_vnic_query_ucast_macs(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_iface_macs *macs)
+{
+ u16 start_idx, tot_macs, num_macs, idx = 0, count = 0;
+ struct netdev_hw_addr *ha;
+
+ start_idx = be16_to_cpu(macs->start_idx);
+ num_macs = be16_to_cpu(macs->num_macs_in_msg);
+ /* loop through dev_addrs list first */
+ for_each_dev_addr(adapter->netdev, ha) {
+ struct hfi_vnic_iface_mac_entry *entry = &macs->entry[count];
+
+ /* Do not include EM specified MAC address */
+ if (!memcmp(adapter->info.vport.base_mac_addr, ha->addr,
+ ARRAY_SIZE(adapter->info.vport.base_mac_addr)))
+ continue;
+
+ if (start_idx > idx++)
+ continue;
+ else if (num_macs == count)
+ break;
+ memcpy(entry, ha->addr, sizeof(*entry));
+ count++;
+ }
+
+ /* loop through uc list */
+ netdev_for_each_uc_addr(ha, adapter->netdev) {
+ struct hfi_vnic_iface_mac_entry *entry = &macs->entry[count];
+
+ if (start_idx > idx++)
+ continue;
+ else if (num_macs == count)
+ break;
+ memcpy(entry, ha->addr, sizeof(*entry));
+ count++;
+ }
+
+ tot_macs = netdev_hw_addr_list_count(&adapter->netdev->dev_addrs) +
+ netdev_uc_count(adapter->netdev);
+ macs->tot_macs_in_lst = cpu_to_be16(tot_macs);
+ macs->num_macs_in_msg = cpu_to_be16(count);
+ macs->gen_count = cpu_to_be16(adapter->info.vport.uc_macs_gen_count);
+}
--
1.8.3.1
^ permalink raw reply related [flat|nested] 40+ messages in thread* [RFC v3 08/11] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) function
2017-02-07 20:22 [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
` (3 preceding siblings ...)
2017-02-07 20:23 ` [RFC v3 07/11] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) interface Vishwanathapura, Niranjana
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
2017-02-07 20:23 ` [RFC v3 09/11] IB/hfi1: HFI_VNIC RDMA netdev support Vishwanathapura, Niranjana
` (3 subsequent siblings)
8 siblings, 0 replies; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Sadanand Warrier, Niranjana Vishwanathapura, Tanya K Jajodia,
Sudeep Dutt
HFI VEMA function interfaces with the Infiniband MAD stack to exchange the
management information packets with the Ethernet Manager (EM).
It interfaces with the HFI VNIC netdev function to SET/GET the management
information. The information exchanged with the EM includes class port
details, encapsulation configuration, various counters, unicast and
multicast MAC list and the MAC table. It also supports sending traps
to the EM.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Sadanand Warrier <sadanand.warrier@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Tanya K Jajodia <tanya.k.jajodia@intel.com>
Signed-off-by: Sudeep Dutt <sudeep.dutt@intel.com>
---
drivers/infiniband/ulp/hfi_vnic/Makefile | 2 +-
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c | 12 +
.../infiniband/ulp/hfi_vnic/hfi_vnic_internal.h | 17 +-
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema.c | 1071 ++++++++++++++++++++
.../infiniband/ulp/hfi_vnic/hfi_vnic_vema_iface.c | 2 +-
5 files changed, 1099 insertions(+), 5 deletions(-)
create mode 100644 drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema.c
diff --git a/drivers/infiniband/ulp/hfi_vnic/Makefile b/drivers/infiniband/ulp/hfi_vnic/Makefile
index 32d6821..9262b6c 100644
--- a/drivers/infiniband/ulp/hfi_vnic/Makefile
+++ b/drivers/infiniband/ulp/hfi_vnic/Makefile
@@ -4,4 +4,4 @@
obj-$(CONFIG_INFINIBAND_HFI_VNIC) += hfi_vnic.o
hfi_vnic-y := hfi_vnic_netdev.o hfi_vnic_encap.o hfi_vnic_ethtool.o \
- hfi_vnic_vema_iface.o
+ hfi_vnic_vema.o hfi_vnic_vema_iface.o
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c
index a9925df..aca54ca 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_ethtool.c
@@ -125,6 +125,17 @@ struct vnic_stats {
#define VNIC_STATS_LEN ARRAY_SIZE(vnic_gstrings_stats)
+/* vnic_get_drvinfo - get driver info */
+static void vnic_get_drvinfo(struct net_device *netdev,
+ struct ethtool_drvinfo *drvinfo)
+{
+ strlcpy(drvinfo->driver, hfi_vnic_driver_name, sizeof(drvinfo->driver));
+ strlcpy(drvinfo->version, hfi_vnic_driver_version,
+ sizeof(drvinfo->version));
+ strlcpy(drvinfo->bus_info, dev_name(netdev->dev.parent),
+ sizeof(drvinfo->bus_info));
+}
+
/* vnic_get_sset_count - get string set count */
static int vnic_get_sset_count(struct net_device *netdev, int sset)
{
@@ -180,6 +191,7 @@ static void vnic_get_strings(struct net_device *netdev, u32 stringset, u8 *data)
/* ethtool ops */
static const struct ethtool_ops hfi_vnic_ethtool_ops = {
+ .get_drvinfo = vnic_get_drvinfo,
.get_link = ethtool_op_get_link,
.get_strings = vnic_get_strings,
.get_sset_count = vnic_get_sset_count,
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
index 2ee80a7..ad61624 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
@@ -164,10 +164,12 @@ struct __hfi_veswport_trap {
* struct hfi_vnic_ctrl_port - HFI virtual NIC control port
* @ibdev: pointer to ib device
* @ops: hfi vnic control operations
+ * @num_ports: number of hfi ports
*/
struct hfi_vnic_ctrl_port {
struct ib_device *ibdev;
struct hfi_vnic_ctrl_ops *ops;
+ u8 num_ports;
};
/**
@@ -187,6 +189,8 @@ struct hfi_vnic_ctrl_port {
* @mactbl_lock: mac table lock
* @stats_lock: statistics lock
* @flow_tbl: flow to default port redirection table
+ * @trap_timeout: trap timeout
+ * @trap_count: no. of traps allowed within timeout period
*/
struct hfi_vnic_adapter {
struct net_device *netdev;
@@ -213,6 +217,9 @@ struct hfi_vnic_adapter {
struct mutex stats_lock;
u8 flow_tbl[HFI_VNIC_FLOW_TBL_SIZE];
+
+ unsigned long trap_timeout;
+ u8 trap_count;
};
/* Same as hfi_veswport_mactable_entry, but without bitwise attribute */
@@ -247,6 +254,8 @@ struct hfi_vnic_mac_tbl_node {
dev_err(&cport->ibdev->dev, format, ## arg)
#define c_info(format, arg...) \
dev_info(&cport->ibdev->dev, format, ## arg)
+#define c_dbg(format, arg...) \
+ dev_dbg(&cport->ibdev->dev, format, ## arg)
/* The maximum allowed entries in the mac table */
#define HFI_VNIC_MAC_TBL_MAX_ENTRIES 2048
@@ -281,6 +290,9 @@ struct hfi_vnic_mac_tbl_node {
!obj && (bkt) < HFI_VNIC_MAC_TBL_SIZE; (bkt)++) \
hlist_for_each_entry(obj, &name[bkt], member)
+extern char hfi_vnic_driver_name[];
+extern const char hfi_vnic_driver_version[];
+
struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
u8 port_num, u8 vport_num);
void hfi_vnic_rem_netdev(struct hfi_vnic_adapter *adapter);
@@ -310,9 +322,8 @@ void hfi_vnic_get_per_veswport_info(struct hfi_vnic_adapter *adapter,
void hfi_vnic_set_per_veswport_info(struct hfi_vnic_adapter *adapter,
struct hfi_per_veswport_info *info);
void hfi_vnic_vema_report_event(struct hfi_vnic_adapter *adapter, u8 event);
-struct hfi_vnic_adapter *hfi_vnic_add_vport(struct hfi_vnic_ctrl_port *cport,
- u8 port_num, u8 vport_num);
-void hfi_vnic_rem_vport(struct hfi_vnic_adapter *adapter);
void hfi_vnic_set_ethtool_ops(struct net_device *netdev);
+void hfi_vnic_vema_send_trap(struct hfi_vnic_adapter *adapter,
+ struct __hfi_veswport_trap *data, u32 lid);
#endif /* _HFI_VNIC_INTERNAL_H */
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema.c
new file mode 100644
index 0000000..c1a7c81
--- /dev/null
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema.c
@@ -0,0 +1,1071 @@
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * This file contains HFI Virtual Network Interface Controller (VNIC)
+ * Ethernet Management Agent (EMA) driver
+ */
+
+#include <linux/module.h>
+#include <rdma/ib_addr.h>
+#include <rdma/ib_smi.h>
+
+#include "hfi_vnic_internal.h"
+
+#define DRV_VERSION "1.0"
+char hfi_vnic_driver_name[] = "hfi_vnic";
+const char hfi_vnic_driver_version[] = DRV_VERSION;
+
+/*
+ * The trap service level is kept in bits 3 to 7 in the trap_sl_rsvd
+ * field in the class port info MAD.
+ */
+#define GET_TRAP_SL_FROM_CLASS_PORT_INFO(x) (((x) >> 3) & 0x1f)
+
+/* Cap trap bursts to a reasonable limit good for normal cases */
+#define HFI_VNIC_TRAP_BURST_LIMIT 4
+
+/*
+ * VNIC trap limit timeout.
+ * Inverse of cap2_mask response time out (1.0737 secs) = 0.9
+ * secs approx IB spec 13.4.6.2.1 PortInfoSubnetTimeout and
+ * 13.4.9 Traps.
+ */
+#define HFI_VNIC_TRAP_TIMEOUT ((4096 * (1UL << 18)) / 1000)
+
+#define HFI_VNIC_UNSUP_ATTR \
+ cpu_to_be16(IB_MGMT_MAD_STATUS_UNSUPPORTED_METHOD_ATTRIB)
+
+#define HFI_VNIC_INVAL_ATTR \
+ cpu_to_be16(IB_MGMT_MAD_STATUS_INVALID_ATTRIB_VALUE)
+
+#define HFI_VNIC_CLASS_CAP_TRAP cpu_to_be16(1 << 8)
+
+struct hfi_class_port_info {
+ u8 base_version;
+ u8 class_version;
+ __be16 cap_mask;
+ __be32 cap_mask2_resp_time;
+
+ u8 redirect_gid[16];
+ __be32 redirect_tc_fl;
+ __be32 redirect_lid;
+ __be32 redirect_sl_qp;
+ __be32 redirect_qkey;
+
+ u8 trap_gid[16];
+ __be32 trap_tc_fl;
+ __be32 trap_lid;
+ __be32 trap_hl_qp;
+ __be32 trap_qkey;
+
+ __be16 trap_pkey;
+ __be16 redirect_pkey;
+
+ u8 trap_sl_rsvd;
+ u8 reserved[3];
+} __packed;
+
+/**
+ * struct hfi_vnic_vema_port -- VNIC VEMA port details
+ * @cport: pointer to port
+ * @mad_agent: pointer to mad agent for port
+ * @class_port_info: Class port info information.
+ * @tid: Transaction id
+ * @port_num: HFI port number
+ * @vport_idr: vnic ports idr
+ * @event_handler: ib event handler
+ * @lock: adapter interface lock
+ */
+struct hfi_vnic_vema_port {
+ struct hfi_vnic_ctrl_port *cport;
+ struct ib_mad_agent *mad_agent;
+ struct hfi_class_port_info class_port_info;
+ u64 tid;
+ u8 port_num;
+ struct idr vport_idr;
+ struct ib_event_handler event_handler;
+
+ /* Lock to query/update network adapter */
+ struct mutex lock;
+};
+
+static void hfi_vnic_vema_add_one(struct ib_device *device);
+static void hfi_vnic_vema_rem_one(struct ib_device *device,
+ void *client_data);
+
+static struct ib_client hfi_vnic_client = {
+ .name = hfi_vnic_driver_name,
+ .add = hfi_vnic_vema_add_one,
+ .remove = hfi_vnic_vema_rem_one,
+};
+
+/**
+ * vema_get_vport_num -- Get the vnic from the mad
+ * @recvd_mad: Received mad
+ *
+ * Return: returns value of the vnic port number
+ */
+static inline u8 vema_get_vport_num(struct hfi_vnic_vema_mad *recvd_mad)
+{
+ return be32_to_cpu(recvd_mad->mad_hdr.attr_mod) >> 16 & 0xff;
+}
+
+/**
+ * vema_get_vport_adapter -- Get vnic port adapter from recvd mad
+ * @recvd_mad: received mad
+ * @port: ptr to port struct on which MAD was recvd
+ *
+ * Return: vnic adapter
+ */
+static inline struct hfi_vnic_adapter *
+vema_get_vport_adapter(struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_port *port)
+{
+ u8 vport_num = vema_get_vport_num(recvd_mad);
+
+ return idr_find(&port->vport_idr, vport_num);
+}
+
+/**
+ * vema_mac_tbl_req_ok -- Check if mac request has correct values
+ * @mac_tbl: mac table
+ *
+ * This function checks for the validity of the offset and number of
+ * entries required.
+ *
+ * Return: true if offset and num_entries are valid
+ */
+static inline bool vema_mac_tbl_req_ok(struct hfi_veswport_mactable *mac_tbl)
+{
+ u16 offset, num_entries;
+ u16 req_entries = ((HFI_VNIC_EMA_DATA - sizeof(*mac_tbl)) /
+ sizeof(mac_tbl->tbl_entries[0]));
+
+ offset = be16_to_cpu(mac_tbl->offset);
+ num_entries = be16_to_cpu(mac_tbl->num_entries);
+
+ return ((num_entries <= req_entries) &&
+ (offset + num_entries <= HFI_VNIC_MAC_TBL_MAX_ENTRIES));
+}
+
+/*
+ * Return the power on default values in the port info structure
+ * in big endian format as required by MAD.
+ */
+static inline void vema_get_pod_values(struct hfi_veswport_info *port_info)
+{
+ memset(port_info, 0, sizeof(*port_info));
+ port_info->vport.max_mac_tbl_ent =
+ cpu_to_be16(HFI_VNIC_MAC_TBL_MAX_ENTRIES);
+ port_info->vport.max_smac_ent =
+ cpu_to_be16(HFI_VNIC_MAX_SMAC_LIMIT);
+ port_info->vport.oper_state = HFI_VNIC_STATE_DROP_ALL;
+ port_info->vport.config_state = HFI_VNIC_STATE_DROP_ALL;
+}
+
+/**
+ * vema_add_vport -- Add a new vnic port
+ * @port: ptr to hfi_vnic_vema_port struct
+ * @vport_num: vnic port number (to be added)
+ *
+ * Return a pointer to the vnic adapter structure
+ */
+static struct hfi_vnic_adapter *vema_add_vport(struct hfi_vnic_vema_port *port,
+ u8 vport_num)
+{
+ struct hfi_vnic_ctrl_port *cport = port->cport;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = hfi_vnic_add_netdev(cport->ibdev, port->port_num, vport_num);
+ if (!IS_ERR(adapter)) {
+ int rc;
+
+ adapter->cport = cport;
+ rc = idr_alloc(&port->vport_idr, adapter, vport_num,
+ vport_num + 1, GFP_NOWAIT);
+ if (rc < 0) {
+ hfi_vnic_rem_netdev(adapter);
+ adapter = ERR_PTR(rc);
+ }
+ }
+
+ return adapter;
+}
+
+/**
+ * vema_get_class_port_info -- Get class info for port
+ * @port: Port on whic MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function copies the latest class port info value set for the
+ * port and stores it for generating traps
+ */
+static void vema_get_class_port_info(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_class_port_info *port_info;
+
+ port_info = (struct hfi_class_port_info *)rsp_mad->data;
+ memcpy(port_info, &port->class_port_info, sizeof(*port_info));
+ port_info->base_version = OPA_MGMT_BASE_VERSION,
+ port_info->class_version = HFI_EMA_CLASS_VERSION;
+
+ /* Agent generates traps */
+ port_info->cap_mask = HFI_VNIC_CLASS_CAP_TRAP;
+
+ /*
+ * Since a get routine is always sent by the EM first we
+ * set the expected response time to
+ * 4.096 usec * 2^18 == 1.0737 sec here.
+ */
+ port_info->cap_mask2_resp_time = cpu_to_be32(18);
+}
+
+/**
+ * vema_set_class_port_info -- Get class info for port
+ * @port: Port on whic MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function updates the port class info for the specific vnic
+ * and sets up the response mad data
+ */
+static void vema_set_class_port_info(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ memcpy(&port->class_port_info, recvd_mad->data,
+ sizeof(port->class_port_info));
+
+ vema_get_class_port_info(port, recvd_mad, rsp_mad);
+}
+
+/**
+ * vema_get_veswport_info -- Get veswport info
+ * @port: source port on which MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ */
+static void vema_get_veswport_info(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_info *port_info =
+ (struct hfi_veswport_info *)rsp_mad->data;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (adapter) {
+ memset(port_info, 0, sizeof(*port_info));
+ hfi_vnic_get_vesw_info(adapter, &port_info->vesw);
+ hfi_vnic_get_per_veswport_info(adapter,
+ &port_info->vport);
+ } else {
+ vema_get_pod_values(port_info);
+ }
+}
+
+/**
+ * vema_set_veswport_info -- Set veswport info
+ * @port: source port on which MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function gets the port class infor for vnic
+ */
+static void vema_set_veswport_info(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_vnic_ctrl_port *cport = port->cport;
+ struct hfi_veswport_info *port_info;
+ struct hfi_vnic_adapter *adapter;
+ u8 vport_num;
+
+ vport_num = vema_get_vport_num(recvd_mad);
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (!adapter) {
+ adapter = vema_add_vport(port, vport_num);
+ if (IS_ERR(adapter)) {
+ c_err("failed to add vport %d: %ld\n",
+ vport_num, PTR_ERR(adapter));
+ goto err_exit;
+ }
+ }
+
+ port_info = (struct hfi_veswport_info *)recvd_mad->data;
+ hfi_vnic_set_vesw_info(adapter, &port_info->vesw);
+ hfi_vnic_set_per_veswport_info(adapter, &port_info->vport);
+
+ /* Process the new config settings */
+ hfi_vnic_process_vema_config(adapter);
+
+ vema_get_veswport_info(port, recvd_mad, rsp_mad);
+ return;
+
+err_exit:
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+}
+
+/**
+ * vema_get_mac_entries -- Get MAC entries in VNIC MAC table
+ * @port: source port on which MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function gets the MAC entries that are programmed into
+ * the VNIC MAC forwarding table. It checks for the validity of
+ * the index into the MAC table and the number of entries that
+ * are to be retrieved.
+ */
+static void vema_get_mac_entries(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_mactable *mac_tbl_in, *mac_tbl_out;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (!adapter) {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ return;
+ }
+
+ mac_tbl_in = (struct hfi_veswport_mactable *)recvd_mad->data;
+ mac_tbl_out = (struct hfi_veswport_mactable *)rsp_mad->data;
+
+ if (vema_mac_tbl_req_ok(mac_tbl_in)) {
+ mac_tbl_out->offset = mac_tbl_in->offset;
+ mac_tbl_out->num_entries = mac_tbl_in->num_entries;
+ hfi_vnic_query_mac_tbl(adapter, mac_tbl_out);
+ } else {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ }
+}
+
+/**
+ * vema_set_mac_entries -- Set MAC entries in VNIC MAC table
+ * @port: source port on which MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function sets the MAC entries in the VNIC forwarding table
+ * It checks for the validity of the index and the number of forwarding
+ * table entries to be programmed.
+ */
+static void vema_set_mac_entries(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_mactable *mac_tbl;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (!adapter) {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ return;
+ }
+
+ mac_tbl = (struct hfi_veswport_mactable *)recvd_mad->data;
+ if (vema_mac_tbl_req_ok(mac_tbl)) {
+ if (hfi_vnic_update_mac_tbl(adapter, mac_tbl))
+ rsp_mad->mad_hdr.status = HFI_VNIC_UNSUP_ATTR;
+ } else {
+ rsp_mad->mad_hdr.status = HFI_VNIC_UNSUP_ATTR;
+ }
+ vema_get_mac_entries(port, recvd_mad, rsp_mad);
+}
+
+/**
+ * vema_set_delete_vesw -- Reset VESW info to POD values
+ * @port: source port on which MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function clears all the fields of veswport info for the requested vesw
+ * and sets them back to the power-on default values. It does not delete the
+ * vesw.
+ */
+static void vema_set_delete_vesw(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_info *port_info =
+ (struct hfi_veswport_info *)rsp_mad->data;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (!adapter) {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ return;
+ }
+
+ vema_get_pod_values(port_info);
+ hfi_vnic_set_vesw_info(adapter, &port_info->vesw);
+ hfi_vnic_set_per_veswport_info(adapter, &port_info->vport);
+
+ /* Process the new config settings */
+ hfi_vnic_process_vema_config(adapter);
+
+ hfi_vnic_release_mac_tbl(adapter);
+
+ vema_get_veswport_info(port, recvd_mad, rsp_mad);
+}
+
+/**
+ * vema_get_mac_list -- Get the unicast/multicast macs.
+ * @port: source port on which MAD was received
+ * @recvd_mad: Received mad contains fields to set vnic parameters
+ * @rsp_mad: Response mad to be built
+ * @attr_id: Attribute ID indicating multicast or unicast mac list
+ */
+static void vema_get_mac_list(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad,
+ u16 attr_id)
+{
+ struct hfi_veswport_iface_macs *macs_in, *macs_out;
+ int max_entries = (HFI_VNIC_EMA_DATA - sizeof(*macs_out)) / ETH_ALEN;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (!adapter) {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ return;
+ }
+
+ macs_in = (struct hfi_veswport_iface_macs *)recvd_mad->data;
+ macs_out = (struct hfi_veswport_iface_macs *)rsp_mad->data;
+
+ macs_out->start_idx = macs_in->start_idx;
+ if (macs_in->num_macs_in_msg)
+ macs_out->num_macs_in_msg = macs_in->num_macs_in_msg;
+ else
+ macs_out->num_macs_in_msg = cpu_to_be16(max_entries);
+
+ if (attr_id == HFI_EM_ATTR_IFACE_MCAST_MACS)
+ hfi_vnic_query_mcast_macs(adapter, macs_out);
+ else
+ hfi_vnic_query_ucast_macs(adapter, macs_out);
+}
+
+/**
+ * vema_get_summary_counters -- Gets summary counters.
+ * @port: source port on which MAD was received
+ * @recvd_mad: Received mad contains fields to set vnic parameters
+ * @rsp_mad: Response mad to be built
+ */
+static void vema_get_summary_counters(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_summary_counters *cntrs;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (adapter) {
+ cntrs = (struct hfi_veswport_summary_counters *)rsp_mad->data;
+ hfi_vnic_get_summary_counters(adapter, cntrs);
+ } else {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ }
+}
+
+/**
+ * vema_get_error_counters -- Gets summary counters.
+ * @port: source port on which MAD was received
+ * @recvd_mad: Received mad contains fields to set vnic parameters
+ * @rsp_mad: Response mad to be built
+ */
+static void vema_get_error_counters(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_error_counters *cntrs;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (adapter) {
+ cntrs = (struct hfi_veswport_error_counters *)rsp_mad->data;
+ hfi_vnic_get_error_counters(adapter, cntrs);
+ } else {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ }
+}
+
+/**
+ * vema_get -- Process received get MAD
+ * @port: source port on which MAD was received
+ * @recvd_mad: Received mad
+ * @rsp_mad: Response mad to be built
+ */
+static void vema_get(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ u16 attr_id = be16_to_cpu(recvd_mad->mad_hdr.attr_id);
+
+ switch (attr_id) {
+ case HFI_EM_ATTR_CLASS_PORT_INFO:
+ vema_get_class_port_info(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_VESWPORT_INFO:
+ vema_get_veswport_info(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_VESWPORT_MAC_ENTRIES:
+ vema_get_mac_entries(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_IFACE_UCAST_MACS:
+ /* fall through */
+ case HFI_EM_ATTR_IFACE_MCAST_MACS:
+ vema_get_mac_list(port, recvd_mad, rsp_mad, attr_id);
+ break;
+ case HFI_EM_ATTR_VESWPORT_SUMMARY_COUNTERS:
+ vema_get_summary_counters(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_VESWPORT_ERROR_COUNTERS:
+ vema_get_error_counters(port, recvd_mad, rsp_mad);
+ break;
+ default:
+ rsp_mad->mad_hdr.status = HFI_VNIC_UNSUP_ATTR;
+ break;
+ }
+}
+
+/**
+ * vema_set -- Process received set MAD
+ * @port: source port on which MAD was received
+ * @recvd_mad: Received mad contains fields to set vnic parameters
+ * @rsp_mad: Response mad to be built
+ */
+static void vema_set(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ u16 attr_id = be16_to_cpu(recvd_mad->mad_hdr.attr_id);
+
+ switch (attr_id) {
+ case HFI_EM_ATTR_CLASS_PORT_INFO:
+ vema_set_class_port_info(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_VESWPORT_INFO:
+ vema_set_veswport_info(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_VESWPORT_MAC_ENTRIES:
+ vema_set_mac_entries(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_DELETE_VESW:
+ vema_set_delete_vesw(port, recvd_mad, rsp_mad);
+ break;
+ default:
+ rsp_mad->mad_hdr.status = HFI_VNIC_UNSUP_ATTR;
+ break;
+ }
+}
+
+/**
+ * vema_send -- Send handler for VEMA MAD agent
+ * @mad_agent: pointer to the mad agent
+ * @mad_wc: pointer to mad send work completion information
+ *
+ * Free all the data structures associated with the sent MAD
+ */
+static void vema_send(struct ib_mad_agent *mad_agent,
+ struct ib_mad_send_wc *mad_wc)
+{
+ ib_destroy_ah(mad_wc->send_buf->ah);
+ ib_free_send_mad(mad_wc->send_buf);
+}
+
+/**
+ * vema_recv -- Recv handler for VEMA MAD agent
+ * @mad_agent: pointer to the mad agent
+ * @send_buf: Send buffer if found, else NULL
+ * @mad_wc: pointer to mad send work completion information
+ *
+ * Handle only set and get methods and respond to other methods
+ * as unsupported. Allocate response buffer and address handle
+ * for the response MAD.
+ */
+static void vema_recv(struct ib_mad_agent *mad_agent,
+ struct ib_mad_send_buf *send_buf,
+ struct ib_mad_recv_wc *mad_wc)
+{
+ struct hfi_vnic_vema_port *port;
+ struct ib_ah *ah;
+ struct ib_mad_send_buf *rsp;
+ struct hfi_vnic_vema_mad *vema_mad;
+
+ if (!mad_wc || !mad_wc->recv_buf.mad)
+ return;
+
+ port = mad_agent->context;
+ ah = ib_create_ah_from_wc(mad_agent->qp->pd, mad_wc->wc,
+ mad_wc->recv_buf.grh, mad_agent->port_num);
+ if (IS_ERR(ah))
+ goto free_recv_mad;
+
+ rsp = ib_create_send_mad(mad_agent, mad_wc->wc->src_qp,
+ mad_wc->wc->pkey_index, 0,
+ IB_MGMT_VENDOR_HDR, HFI_VNIC_EMA_DATA,
+ GFP_KERNEL, OPA_MGMT_BASE_VERSION);
+ if (IS_ERR(rsp))
+ goto err_rsp;
+
+ rsp->ah = ah;
+ vema_mad = rsp->mad;
+ memcpy(vema_mad, mad_wc->recv_buf.mad, IB_MGMT_VENDOR_HDR);
+ vema_mad->mad_hdr.method = IB_MGMT_METHOD_GET_RESP;
+ vema_mad->mad_hdr.status = 0;
+
+ /* Lock ensures network adapter is not removed */
+ mutex_lock(&port->lock);
+
+ switch (mad_wc->recv_buf.mad->mad_hdr.method) {
+ case IB_MGMT_METHOD_GET:
+ vema_get(port, (struct hfi_vnic_vema_mad *)mad_wc->recv_buf.mad,
+ vema_mad);
+ break;
+ case IB_MGMT_METHOD_SET:
+ vema_set(port, (struct hfi_vnic_vema_mad *)mad_wc->recv_buf.mad,
+ vema_mad);
+ break;
+ default:
+ vema_mad->mad_hdr.status = HFI_VNIC_UNSUP_ATTR;
+ break;
+ }
+ mutex_unlock(&port->lock);
+
+ if (!ib_post_send_mad(rsp, NULL)) {
+ /*
+ * with post send successful ah and send mad
+ * will be destroyed in send handler
+ */
+ goto free_recv_mad;
+ }
+
+ ib_free_send_mad(rsp);
+
+err_rsp:
+ ib_destroy_ah(ah);
+free_recv_mad:
+ ib_free_recv_mad(mad_wc);
+}
+
+/**
+ * vema_get_port -- Gets the hfi_vnic_vema_port
+ * @cport: pointer to control dev
+ * @port_num: Port number
+ *
+ * This function loops through the ports and returns
+ * the hfi_vnic_vema port structure that is associated
+ * with the HFI port number
+ *
+ * Return: ptr to requested hfi_vnic_vema_port strucure
+ * if success, NULL if not
+ */
+static struct hfi_vnic_vema_port *
+vema_get_port(struct hfi_vnic_ctrl_port *cport, u8 port_num)
+{
+ struct hfi_vnic_vema_port *port = (void *)cport + sizeof(*cport);
+
+ if (port_num > cport->num_ports)
+ return NULL;
+
+ return port + (port_num - 1);
+}
+
+/**
+ * hfi_vnic_vema_send_trap -- This function sends a trap to the EM
+ * @cport: pointer to vnic control port
+ * @data: pointer to trap data filled by calling function
+ * @lid: issuers lid (encap_slid from vesw_port_info)
+ *
+ * This function is called from the VNIC driver to send a trap if there
+ * is somethng the EM should be notified about. These events currently
+ * are
+ * 1) UNICAST INTERFACE MACADDRESS changes
+ * 2) MULTICAST INTERFACE MACADDRESS changes
+ * 3) ETHERNET LINK STATUS changes
+ * While allocating the send mad the remote site qpn used is 1
+ * as this is the well known QP.
+ *
+ */
+void hfi_vnic_vema_send_trap(struct hfi_vnic_adapter *adapter,
+ struct __hfi_veswport_trap *data, u32 lid)
+{
+ struct hfi_vnic_ctrl_port *cport = adapter->cport;
+ struct ib_mad_send_buf *send_buf;
+ struct hfi_vnic_vema_port *port;
+ struct ib_device *ibp;
+ struct hfi_vnic_vema_mad_trap *trap_mad;
+ struct hfi_class_port_info *class;
+ struct ib_ah_attr ah_attr;
+ struct ib_ah *ah;
+ struct hfi_veswport_trap *trap;
+ u32 trap_lid;
+ u16 pkey_idx;
+
+ if (!cport)
+ goto err_exit;
+ ibp = cport->ibdev;
+ port = vema_get_port(cport, data->hfiportnum);
+ if (!port || !port->mad_agent)
+ goto err_exit;
+
+ if (time_before(jiffies, adapter->trap_timeout)) {
+ if (adapter->trap_count == HFI_VNIC_TRAP_BURST_LIMIT) {
+ v_warn("Trap rate exceeded\n");
+ goto err_exit;
+ } else {
+ adapter->trap_count++;
+ }
+ } else {
+ adapter->trap_count = 0;
+ }
+
+ class = &port->class_port_info;
+ /* Set up address handle */
+ memset(&ah_attr, 0, sizeof(ah_attr));
+ ah_attr.sl = GET_TRAP_SL_FROM_CLASS_PORT_INFO(class->trap_sl_rsvd);
+ ah_attr.port_num = port->port_num;
+ trap_lid = be32_to_cpu(class->trap_lid);
+ /*
+ * check for trap lid validity, must not be zero
+ * The trap sink could change after we fashion the MAD but since traps
+ * are not guaranteed we won't use a lock as anyway the change will take
+ * place even with locking.
+ */
+ if (!trap_lid) {
+ c_err("%s: Invalid dlid\n", __func__);
+ goto err_exit;
+ }
+
+ ah_attr.dlid = trap_lid;
+ ah = ib_create_ah(port->mad_agent->qp->pd, &ah_attr);
+ if (IS_ERR(ah)) {
+ c_err("%s:Couldn't create new AH = %p\n", __func__, ah);
+ c_err("%s:dlid = %d, sl = %d, port = %d\n", __func__,
+ ah_attr.dlid, ah_attr.sl, ah_attr.port_num);
+ goto err_exit;
+ }
+
+ if (ib_find_pkey(ibp, data->hfiportnum, IB_DEFAULT_PKEY_FULL,
+ &pkey_idx) < 0) {
+ c_err("%s:full key not found, defaulting to partial\n",
+ __func__);
+ if (ib_find_pkey(ibp, data->hfiportnum, IB_DEFAULT_PKEY_PARTIAL,
+ &pkey_idx) < 0)
+ pkey_idx = 1;
+ }
+
+ send_buf = ib_create_send_mad(port->mad_agent, 1, pkey_idx, 0,
+ IB_MGMT_VENDOR_HDR, IB_MGMT_MAD_DATA,
+ GFP_KERNEL, OPA_MGMT_BASE_VERSION);
+ if (IS_ERR(send_buf)) {
+ c_err("%s:Couldn't allocate send buf\n", __func__);
+ goto err_sndbuf;
+ }
+
+ send_buf->ah = ah;
+
+ /* Set up common MAD hdr */
+ trap_mad = send_buf->mad;
+ trap_mad->mad_hdr.base_version = OPA_MGMT_BASE_VERSION;
+ trap_mad->mad_hdr.mgmt_class = HFI_MGMT_CLASS_INTEL_EMA;
+ trap_mad->mad_hdr.class_version = HFI_EMA_CLASS_VERSION;
+ trap_mad->mad_hdr.method = IB_MGMT_METHOD_TRAP;
+ port->tid++;
+ trap_mad->mad_hdr.tid = cpu_to_be64(port->tid);
+ trap_mad->mad_hdr.attr_id = IB_SMP_ATTR_NOTICE;
+
+ /* Set up vendor OUI */
+ trap_mad->oui[0] = INTEL_OUI_1;
+ trap_mad->oui[1] = INTEL_OUI_2;
+ trap_mad->oui[2] = INTEL_OUI_3;
+
+ /* Setup notice attribute portion */
+ trap_mad->notice.gen_type = HFI_INTEL_EMA_NOTICE_TYPE_INFO << 1;
+ trap_mad->notice.oui_1 = INTEL_OUI_1;
+ trap_mad->notice.oui_2 = INTEL_OUI_2;
+ trap_mad->notice.oui_3 = INTEL_OUI_3;
+ trap_mad->notice.issuer_lid = cpu_to_be32(lid);
+
+ /* copy the actual trap data */
+ trap = (struct hfi_veswport_trap *)trap_mad->notice.raw_data;
+ trap->fabric_id = cpu_to_be16(data->fabric_id);
+ trap->veswid = cpu_to_be16(data->veswid);
+ trap->veswportnum = cpu_to_be32(data->veswportnum);
+ trap->hfiportnum = cpu_to_be16(data->hfiportnum);
+ trap->veswportindex = data->veswportindex;
+ trap->opcode = data->opcode;
+
+ /* If successful send set up rate limit timeout else bail */
+ if (ib_post_send_mad(send_buf, NULL)) {
+ ib_free_send_mad(send_buf);
+ } else {
+ if (adapter->trap_count)
+ return;
+ adapter->trap_timeout = jiffies +
+ usecs_to_jiffies(HFI_VNIC_TRAP_TIMEOUT);
+ return;
+ }
+
+err_sndbuf:
+ ib_destroy_ah(ah);
+err_exit:
+ v_err("Aborting trap\n");
+}
+
+static int vema_rem_vport(int id, void *p, void *data)
+{
+ struct hfi_vnic_adapter *adapter = p;
+
+ hfi_vnic_rem_netdev(adapter);
+ return 0;
+}
+
+static int vema_enable_vport(int id, void *p, void *data)
+{
+ struct hfi_vnic_adapter *adapter = p;
+
+ netif_carrier_on(adapter->netdev);
+ return 0;
+}
+
+static int vema_disable_vport(int id, void *p, void *data)
+{
+ struct hfi_vnic_adapter *adapter = p;
+
+ netif_carrier_off(adapter->netdev);
+ return 0;
+}
+
+static void hfi_vnic_event(struct ib_event_handler *handler,
+ struct ib_event *record)
+{
+ struct hfi_vnic_vema_port *port =
+ container_of(handler, struct hfi_vnic_vema_port, event_handler);
+ struct hfi_vnic_ctrl_port *cport = port->cport;
+
+ if (record->element.port_num != port->port_num)
+ return;
+
+ c_dbg("HFI_VNIC received event %d on device %s port %d\n",
+ record->event, record->device->name, record->element.port_num);
+
+ if (record->event == IB_EVENT_PORT_ERR)
+ idr_for_each(&port->vport_idr, vema_disable_vport, NULL);
+ if (record->event == IB_EVENT_PORT_ACTIVE)
+ idr_for_each(&port->vport_idr, vema_enable_vport, NULL);
+}
+
+/**
+ * vema_unregister -- Unregisters agent
+ * @cport: pointer to control port
+ *
+ * This deletes the registration by VEMA for MADs
+ */
+static void vema_unregister(struct hfi_vnic_ctrl_port *cport)
+{
+ int i;
+
+ for (i = 1; i <= cport->num_ports; i++) {
+ struct hfi_vnic_vema_port *port = vema_get_port(cport, i);
+
+ if (!port->mad_agent)
+ continue;
+
+ /* Lock ensures no MAD is being processed */
+ mutex_lock(&port->lock);
+ idr_for_each(&port->vport_idr, vema_rem_vport, NULL);
+ mutex_unlock(&port->lock);
+
+ ib_unregister_mad_agent(port->mad_agent);
+ port->mad_agent = NULL;
+ mutex_destroy(&port->lock);
+ idr_destroy(&port->vport_idr);
+ ib_unregister_event_handler(&port->event_handler);
+ }
+}
+
+/**
+ * vema_register -- Registers agent
+ * @cport: pointer to control port
+ *
+ * This function registers the handlers for the VEMA MADs
+ *
+ * Return: returns 0 on success. non zero otherwise
+ */
+static int vema_register(struct hfi_vnic_ctrl_port *cport)
+{
+ struct ib_mad_reg_req reg_req = {
+ .mgmt_class = HFI_MGMT_CLASS_INTEL_EMA,
+ .mgmt_class_version = OPA_MGMT_BASE_VERSION,
+ .oui = { INTEL_OUI_1, INTEL_OUI_2, INTEL_OUI_3 }
+ };
+ int i;
+
+ set_bit(IB_MGMT_METHOD_GET, reg_req.method_mask);
+ set_bit(IB_MGMT_METHOD_SET, reg_req.method_mask);
+
+ /* register ib event handler and mad agent for each port on dev */
+ for (i = 1; i <= cport->num_ports; i++) {
+ struct hfi_vnic_vema_port *port = vema_get_port(cport, i);
+ int ret;
+
+ port->cport = cport;
+ port->port_num = i;
+
+ INIT_IB_EVENT_HANDLER(&port->event_handler,
+ cport->ibdev, hfi_vnic_event);
+ ret = ib_register_event_handler(&port->event_handler);
+ if (ret) {
+ c_err("port %d: event handler register failed\n", i);
+ vema_unregister(cport);
+ return ret;
+ }
+
+ idr_init(&port->vport_idr);
+ mutex_init(&port->lock);
+ port->mad_agent = ib_register_mad_agent(cport->ibdev, i,
+ IB_QPT_GSI, ®_req,
+ IB_MGMT_RMPP_VERSION,
+ vema_send, vema_recv,
+ port, 0);
+ if (IS_ERR(port->mad_agent)) {
+ ret = PTR_ERR(port->mad_agent);
+ port->mad_agent = NULL;
+ mutex_destroy(&port->lock);
+ idr_destroy(&port->vport_idr);
+ vema_unregister(cport);
+ return ret;
+ }
+ }
+
+ return 0;
+}
+
+/**
+ * hfi_vnic_vema_add_one -- Handle new ib device
+ * @device: ib device pointer
+ *
+ * Allocate the vnic control port and initialize it.
+ */
+static void hfi_vnic_vema_add_one(struct ib_device *device)
+{
+ struct hfi_vnic_ctrl_port *cport;
+ int rc, size = sizeof(*cport);
+
+ if (!rdma_cap_hfi_vnic(device))
+ return;
+
+ size += device->phys_port_cnt * sizeof(struct hfi_vnic_vema_port);
+ cport = kzalloc(size, GFP_KERNEL);
+ if (!cport)
+ return;
+
+ cport->num_ports = device->phys_port_cnt;
+ cport->ibdev = device;
+
+ /* Initialize hfi vnic management agent (vema) */
+ rc = vema_register(cport);
+ if (!rc)
+ c_info("VNIC client initialized\n");
+
+ ib_set_client_data(device, &hfi_vnic_client, cport);
+}
+
+/**
+ * hfi_vnic_vema_rem_one -- Handle ib device removal
+ * @device: ib device pointer
+ * @client_data: ib client data
+ *
+ * Uninitialize and free the vnic control port.
+ */
+static void hfi_vnic_vema_rem_one(struct ib_device *device,
+ void *client_data)
+{
+ struct hfi_vnic_ctrl_port *cport = client_data;
+
+ if (!cport)
+ return;
+
+ c_info("removing VNIC client\n");
+ vema_unregister(cport);
+ kfree(cport);
+}
+
+static int __init hfi_vnic_init(void)
+{
+ int rc;
+
+ pr_info("HFI Virtual Network Driver - v%s\n",
+ hfi_vnic_driver_version);
+
+ rc = ib_register_client(&hfi_vnic_client);
+ if (rc)
+ pr_err("VNIC driver register failed %d\n", rc);
+
+ return rc;
+}
+module_init(hfi_vnic_init);
+
+static void hfi_vnic_deinit(void)
+{
+ ib_unregister_client(&hfi_vnic_client);
+}
+module_exit(hfi_vnic_deinit);
+
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_AUTHOR("Intel Corporation");
+MODULE_DESCRIPTION("Intel HFI Virtual Network driver");
+MODULE_VERSION(DRV_VERSION);
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema_iface.c b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema_iface.c
index fca3bab..80b557b 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema_iface.c
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_vema_iface.c
@@ -70,7 +70,7 @@ void hfi_vnic_vema_report_event(struct hfi_vnic_adapter *adapter, u8 event)
trap_data.veswportindex = adapter->vport_num;
trap_data.opcode = event;
- /* Need to send trap here */
+ hfi_vnic_vema_send_trap(adapter, &trap_data, info->vport.encap_slid);
}
/**
--
1.8.3.1
^ permalink raw reply related [flat|nested] 40+ messages in thread* [RFC v3 09/11] IB/hfi1: HFI_VNIC RDMA netdev support
2017-02-07 20:22 [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
` (4 preceding siblings ...)
2017-02-07 20:23 ` [RFC v3 08/11] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) function Vishwanathapura, Niranjana
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
[not found] ` <1486498990-76562-10-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2017-02-07 20:23 ` [RFC v3 10/11] IB/hfi1: Virtual Network Interface Controller (VNIC) HW support Vishwanathapura, Niranjana
` (2 subsequent siblings)
8 siblings, 1 reply; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura, Andrzej Kacprowski
Add support to create and free HFI_VNIC rdma netdev devices.
Implement netstack interface functionality including xmit_skb,
receive side NAPI etc. Also implement rdma netdev control functions.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Andrzej Kacprowski <andrzej.kacprowski@intel.com>
---
drivers/infiniband/hw/hfi1/Makefile | 2 +-
drivers/infiniband/hw/hfi1/driver.c | 25 +-
drivers/infiniband/hw/hfi1/hfi.h | 27 +-
drivers/infiniband/hw/hfi1/init.c | 9 +-
drivers/infiniband/hw/hfi1/vnic.h | 153 ++++++++
drivers/infiniband/hw/hfi1/vnic_main.c | 646 +++++++++++++++++++++++++++++++++
6 files changed, 855 insertions(+), 7 deletions(-)
create mode 100644 drivers/infiniband/hw/hfi1/vnic.h
create mode 100644 drivers/infiniband/hw/hfi1/vnic_main.c
diff --git a/drivers/infiniband/hw/hfi1/Makefile b/drivers/infiniband/hw/hfi1/Makefile
index 0cf97a0..2280538 100644
--- a/drivers/infiniband/hw/hfi1/Makefile
+++ b/drivers/infiniband/hw/hfi1/Makefile
@@ -12,7 +12,7 @@ hfi1-y := affinity.o chip.o device.o driver.o efivar.o \
init.o intr.o mad.o mmu_rb.o pcie.o pio.o pio_copy.o platform.o \
qp.o qsfp.o rc.o ruc.o sdma.o sysfs.o trace.o \
uc.o ud.o user_exp_rcv.o user_pages.o user_sdma.o verbs.o \
- verbs_txreq.o
+ verbs_txreq.o vnic_main.o
hfi1-$(CONFIG_DEBUG_FS) += debugfs.o
CFLAGS_trace.o = -I$(src)
diff --git a/drivers/infiniband/hw/hfi1/driver.c b/drivers/infiniband/hw/hfi1/driver.c
index 4fbaee6..3c226a8 100644
--- a/drivers/infiniband/hw/hfi1/driver.c
+++ b/drivers/infiniband/hw/hfi1/driver.c
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -59,6 +59,7 @@
#include "trace.h"
#include "qp.h"
#include "sdma.h"
+#include "vnic.h"
#undef pr_fmt
#define pr_fmt(fmt) DRIVER_NAME ": " fmt
@@ -1359,15 +1360,31 @@ int process_receive_ib(struct hfi1_packet *packet)
return RHF_RCV_CONTINUE;
}
+static inline bool hfi1_is_vnic_packet(struct hfi1_packet *packet)
+{
+ /* Packet received in VNIC context via RSM */
+ if (packet->rcd->is_vnic)
+ return true;
+
+ if ((HFI1_GET_L2_TYPE(packet->ebuf) == HFI_VNIC_L2_TYPE) &&
+ (HFI1_GET_L4_TYPE(packet->ebuf) == HFI_VNIC_L4_ETHR))
+ return true;
+
+ return false;
+}
+
int process_receive_bypass(struct hfi1_packet *packet)
{
struct hfi1_devdata *dd = packet->rcd->dd;
- if (unlikely(rhf_err_flags(packet->rhf)))
+ if (unlikely(rhf_err_flags(packet->rhf))) {
handle_eflags(packet);
+ } else if (hfi1_is_vnic_packet(packet)) {
+ hfi1_vnic_bypass_rcv(packet);
+ return RHF_RCV_CONTINUE;
+ }
- dd_dev_err(dd,
- "Bypass packets are not supported in normal operation. Dropping\n");
+ dd_dev_err(dd, "Unsupported bypass packet. Dropping\n");
incr_cntr64(&dd->sw_rcv_bypass_packet_errors);
if (!(dd->err_info_rcvport.status_and_code & OPA_EI_STATUS_SMASK)) {
u64 *flits = packet->ebuf;
diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h
index 751a0fb..a3641fb 100644
--- a/drivers/infiniband/hw/hfi1/hfi.h
+++ b/drivers/infiniband/hw/hfi1/hfi.h
@@ -1,7 +1,7 @@
#ifndef _HFI1_KERNEL_H
#define _HFI1_KERNEL_H
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -337,6 +337,12 @@ struct hfi1_ctxtdata {
* packets with the wrong interrupt handler.
*/
int (*do_interrupt)(struct hfi1_ctxtdata *rcd, int threaded);
+
+ /* Indicates that this is vnic context */
+ bool is_vnic;
+
+ /* vnic queue index this context is mapped to */
+ u8 vnic_q_idx;
};
/*
@@ -809,6 +815,19 @@ struct hfi1_asic_data {
struct hfi1_i2c_bus *i2c_bus1;
};
+/*
+ * Number of VNIC contexts used. Ensure it is less than or equal to
+ * max queues supported by VNIC (HFI_VNIC_MAX_QUEUE).
+ */
+#define HFI1_NUM_VNIC_CTXT 8
+
+/* Virtual NIC information */
+struct hfi1_vnic_data {
+ struct idr vesw_idr;
+};
+
+struct hfi1_vnic_vport_info;
+
/* device data struct now contains only "general per-device" info.
* fields related to a physical IB port are in a hfi1_pportdata struct.
*/
@@ -1116,6 +1135,9 @@ struct hfi1_devdata {
send_routine process_dma_send;
void (*pio_inline_send)(struct hfi1_devdata *dd, struct pio_buf *pbuf,
u64 pbc, const void *from, size_t count);
+ int (*process_vnic_dma_send)(struct hfi1_devdata *dd, u8 q_idx,
+ struct hfi1_vnic_vport_info *vinfo,
+ struct sk_buff *skb, u64 pbc, u8 plen);
/* hfi1_pportdata, points to array of (physical) port-specific
* data structs, indexed by pidx (0..n-1)
*/
@@ -1171,6 +1193,9 @@ struct hfi1_devdata {
struct rhashtable sdma_rht;
struct kobject kobj;
+
+ /* vnic data */
+ struct hfi1_vnic_data vnic;
};
/* 8051 firmware version helper */
diff --git a/drivers/infiniband/hw/hfi1/init.c b/drivers/infiniband/hw/hfi1/init.c
index e3b5bc9..b1f48b1 100644
--- a/drivers/infiniband/hw/hfi1/init.c
+++ b/drivers/infiniband/hw/hfi1/init.c
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -65,6 +65,7 @@
#include "verbs.h"
#include "aspm.h"
#include "affinity.h"
+#include "vnic.h"
#undef pr_fmt
#define pr_fmt(fmt) DRIVER_NAME ": " fmt
@@ -1496,6 +1497,9 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
/* do the generic initialization */
initfail = hfi1_init(dd, 0);
+ /* setup vnic */
+ hfi1_vnic_setup(dd);
+
ret = hfi1_register_ib_device(dd);
/*
@@ -1573,6 +1577,9 @@ static void remove_one(struct pci_dev *pdev)
/* unregister from IB core */
hfi1_unregister_ib_device(dd);
+ /* cleanup vnic */
+ hfi1_vnic_cleanup(dd);
+
/*
* Disable the IB link, disable interrupts on the device,
* clear dma engines, etc.
diff --git a/drivers/infiniband/hw/hfi1/vnic.h b/drivers/infiniband/hw/hfi1/vnic.h
new file mode 100644
index 0000000..23b92d7
--- /dev/null
+++ b/drivers/infiniband/hw/hfi1/vnic.h
@@ -0,0 +1,153 @@
+#ifndef _HFI1_VNIC_H
+#define _HFI1_VNIC_H
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+#include <rdma/opa_hfi.h>
+#include "hfi.h"
+
+#define HFI1_VNIC_MAX_TXQ 16
+#define HFI1_VNIC_MAX_PAD 12
+
+/* L2 header definitions */
+#define HFI1_L2_TYPE_OFFSET 0x7
+#define HFI1_L2_TYPE_SHFT 0x5
+#define HFI1_L2_TYPE_MASK 0x3
+
+#define HFI1_GET_L2_TYPE(hdr) \
+ ((*((u8 *)(hdr) + HFI1_L2_TYPE_OFFSET) >> HFI1_L2_TYPE_SHFT) & \
+ HFI1_L2_TYPE_MASK)
+
+/* L4 type definitions */
+#define HFI1_L4_TYPE_OFFSET 8
+
+#define HFI1_GET_L4_TYPE(data) \
+ (*((u8 *)(data) + HFI1_L4_TYPE_OFFSET))
+
+/* L4 header definitions */
+#define HFI1_VNIC_L4_HDR_OFFSET HFI_VNIC_L2_HDR_LEN
+
+#define HFI1_VNIC_GET_L4_HDR(data) \
+ (*((u16 *)((u8 *)(data) + HFI1_VNIC_L4_HDR_OFFSET)))
+
+#define HFI1_VNIC_GET_VESWID(data) \
+ (HFI1_VNIC_GET_L4_HDR(data) & 0xFF)
+
+/* Service class */
+#define HFI1_VNIC_SC_OFFSET_LOW 6
+#define HFI1_VNIC_SC_OFFSET_HI 7
+#define HFI1_VNIC_SC_SHIFT 4
+
+#define HFI_VNIC_MAX_QUEUE 16
+
+/**
+ * struct hfi1_vnic_rx_queue - HFI1 VNIC receive queue
+ * @idx: queue index
+ * @vinfo: pointer to vport information
+ * @netdev: network device
+ * @napi: netdev napi structure
+ * @skbq: queue of received socket buffers
+ */
+struct hfi1_vnic_rx_queue {
+ u8 idx;
+ struct hfi1_vnic_vport_info *vinfo;
+ struct net_device *netdev;
+ struct napi_struct napi;
+ struct sk_buff_head skbq;
+};
+
+/**
+ * struct hfi1_vnic_vport_info - HFI1 VNIC virtual port information
+ * @dd: device data pointer
+ * @netdev: net device pointer
+ * @flags: state flags
+ * @lock: vport lock
+ * @num_tx_q: number of transmit queues
+ * @num_rx_q: number of receive queues
+ * @vesw_id: virtual switch id
+ * @rxq: Array of receive queues
+ * @stats: per queue stats
+ */
+struct hfi1_vnic_vport_info {
+ struct hfi1_devdata *dd;
+ struct net_device *netdev;
+ unsigned long flags;
+
+ /* Lock used around state updates */
+ struct mutex lock;
+
+ u8 num_tx_q;
+ u8 num_rx_q;
+ u16 vesw_id;
+ struct hfi1_vnic_rx_queue rxq[HFI1_NUM_VNIC_CTXT];
+
+ struct hfi_vnic_stats stats[HFI_VNIC_MAX_QUEUE];
+};
+
+#define v_dbg(format, arg...) \
+ netdev_dbg(vinfo->netdev, format, ## arg)
+#define v_err(format, arg...) \
+ netdev_err(vinfo->netdev, format, ## arg)
+#define v_info(format, arg...) \
+ netdev_info(vinfo->netdev, format, ## arg)
+
+/* vnic hfi1 internal functions */
+void hfi1_vnic_setup(struct hfi1_devdata *dd);
+void hfi1_vnic_cleanup(struct hfi1_devdata *dd);
+
+void hfi1_vnic_bypass_rcv(struct hfi1_packet *packet);
+
+/* vnic rdma netdev operations */
+struct net_device *hfi1_vnic_alloc_rn(struct ib_device *device,
+ u8 port_num,
+ enum rdma_netdev_t type,
+ const char *name,
+ unsigned char name_assign_type,
+ void (*setup)(struct net_device *));
+void hfi1_vnic_free_rn(struct net_device *netdev);
+
+#endif /* _HFI1_VNIC_H */
diff --git a/drivers/infiniband/hw/hfi1/vnic_main.c b/drivers/infiniband/hw/hfi1/vnic_main.c
new file mode 100644
index 0000000..3cc430a
--- /dev/null
+++ b/drivers/infiniband/hw/hfi1/vnic_main.c
@@ -0,0 +1,646 @@
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * This file contains HFI1 support for VNIC functionality
+ */
+
+#include <linux/io.h>
+#include <linux/if_vlan.h>
+
+#include "vnic.h"
+
+#define HFI_TX_TIMEOUT_MS 1000
+
+#define HFI1_VNIC_RCV_Q_SIZE 1024
+
+#define HFI1_VNIC_UP 0
+
+static DEFINE_SPINLOCK(vport_cntr_lock);
+
+void hfi1_vnic_setup(struct hfi1_devdata *dd)
+{
+ idr_init(&dd->vnic.vesw_idr);
+}
+
+void hfi1_vnic_cleanup(struct hfi1_devdata *dd)
+{
+ idr_destroy(&dd->vnic.vesw_idr);
+}
+
+#define SUM_GRP_COUNTERS(stats, qstats, x_grp) do { \
+ u64 *src64, *dst64; \
+ for (src64 = &qstats->x_grp.unicast, \
+ dst64 = &stats->x_grp.unicast; \
+ dst64 <= &stats->x_grp.s_1519_max;) { \
+ *dst64++ += *src64++; \
+ } \
+ } while (0)
+
+/* hfi_vnic_update_stats - update statistics */
+static void hfi_vnic_update_stats(struct hfi1_vnic_vport_info *vinfo,
+ struct hfi_vnic_stats *stats)
+{
+ struct net_device *netdev = vinfo->netdev;
+ u8 i;
+
+ /* add tx counters on different queues */
+ for (i = 0; i < vinfo->num_tx_q; i++) {
+ struct hfi_vnic_stats *qstats = &vinfo->stats[i];
+ struct rtnl_link_stats64 *qnstats = &vinfo->stats[i].netstats;
+
+ stats->netstats.tx_fifo_errors += qnstats->tx_fifo_errors;
+ stats->netstats.tx_carrier_errors += qnstats->tx_carrier_errors;
+ stats->tx_drop_state += qstats->tx_drop_state;
+ stats->tx_dlid_zero += qstats->tx_dlid_zero;
+
+ SUM_GRP_COUNTERS(stats, qstats, tx_grp);
+ stats->netstats.tx_packets += qnstats->tx_packets;
+ stats->netstats.tx_bytes += qnstats->tx_bytes;
+ }
+
+ /* add rx counters on different queues */
+ for (i = 0; i < vinfo->num_rx_q; i++) {
+ struct hfi_vnic_stats *qstats = &vinfo->stats[i];
+ struct rtnl_link_stats64 *qnstats = &vinfo->stats[i].netstats;
+
+ stats->netstats.rx_fifo_errors += qnstats->rx_fifo_errors;
+ stats->netstats.rx_nohandler += qnstats->rx_nohandler;
+ stats->rx_drop_state += qstats->rx_drop_state;
+ stats->rx_oversize += qstats->rx_oversize;
+ stats->rx_runt += qstats->rx_runt;
+
+ SUM_GRP_COUNTERS(stats, qstats, rx_grp);
+ stats->netstats.rx_packets += qnstats->rx_packets;
+ stats->netstats.rx_bytes += qnstats->rx_bytes;
+ }
+
+ stats->netstats.tx_errors = stats->netstats.tx_fifo_errors +
+ stats->netstats.tx_carrier_errors +
+ stats->tx_drop_state + stats->tx_dlid_zero;
+ stats->netstats.tx_dropped = stats->netstats.tx_errors;
+
+ stats->netstats.rx_errors = stats->netstats.rx_fifo_errors +
+ stats->netstats.rx_nohandler +
+ stats->rx_drop_state + stats->rx_oversize +
+ stats->rx_runt;
+ stats->netstats.rx_dropped = stats->netstats.rx_errors;
+
+ netdev->stats.tx_packets = stats->netstats.tx_packets;
+ netdev->stats.tx_bytes = stats->netstats.tx_bytes;
+ netdev->stats.tx_fifo_errors = stats->netstats.tx_fifo_errors;
+ netdev->stats.tx_carrier_errors = stats->netstats.tx_carrier_errors;
+ netdev->stats.tx_errors = stats->netstats.tx_errors;
+ netdev->stats.tx_dropped = stats->netstats.tx_dropped;
+
+ netdev->stats.rx_packets = stats->netstats.rx_packets;
+ netdev->stats.rx_bytes = stats->netstats.rx_bytes;
+ netdev->stats.rx_fifo_errors = stats->netstats.rx_fifo_errors;
+ netdev->stats.multicast = stats->rx_grp.mcastbcast;
+ netdev->stats.rx_length_errors = stats->rx_oversize + stats->rx_runt;
+ netdev->stats.rx_errors = stats->netstats.rx_errors;
+ netdev->stats.rx_dropped = stats->netstats.rx_dropped;
+}
+
+/* update_len_counters - update pkt's len histogram counters */
+static inline void update_len_counters(struct hfi_vnic_grp_stats *grp,
+ int len)
+{
+ /* account for 4 byte FCS */
+ if (len >= 1515)
+ grp->s_1519_max++;
+ else if (len >= 1020)
+ grp->s_1024_1518++;
+ else if (len >= 508)
+ grp->s_512_1023++;
+ else if (len >= 252)
+ grp->s_256_511++;
+ else if (len >= 124)
+ grp->s_128_255++;
+ else if (len >= 61)
+ grp->s_65_127++;
+ else
+ grp->s_64++;
+}
+
+/* hfi_vnic_update_tx_counters - update transmit counters */
+static void hfi_vnic_update_tx_counters(struct hfi1_vnic_vport_info *vinfo,
+ u8 q_idx, struct sk_buff *skb, int err)
+{
+ struct ethhdr *mac_hdr = (struct ethhdr *)skb_mac_header(skb);
+ struct hfi_vnic_stats *stats = &vinfo->stats[q_idx];
+ struct hfi_vnic_grp_stats *tx_grp = &stats->tx_grp;
+ u16 vlan_tci;
+
+ stats->netstats.tx_packets++;
+ stats->netstats.tx_bytes += skb->len + ETH_FCS_LEN;
+
+ update_len_counters(tx_grp, skb->len);
+
+ /* rest of the counts are for good packets only */
+ if (err)
+ return;
+
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ tx_grp->mcastbcast++;
+ else
+ tx_grp->unicast++;
+
+ if (!__vlan_get_tag(skb, &vlan_tci))
+ tx_grp->vlan++;
+ else
+ tx_grp->untagged++;
+}
+
+/* hfi_vnic_update_rx_counters - update receive counters */
+static void hfi_vnic_update_rx_counters(struct hfi1_vnic_vport_info *vinfo,
+ u8 q_idx, struct sk_buff *skb, int err)
+{
+ struct ethhdr *mac_hdr = (struct ethhdr *)skb->data;
+ struct hfi_vnic_stats *stats = &vinfo->stats[q_idx];
+ struct hfi_vnic_grp_stats *rx_grp = &stats->rx_grp;
+ u16 vlan_tci;
+
+ stats->netstats.rx_packets++;
+ stats->netstats.rx_bytes += skb->len + ETH_FCS_LEN;
+
+ update_len_counters(rx_grp, skb->len);
+
+ /* rest of the counts are for good packets only */
+ if (err)
+ return;
+
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ rx_grp->mcastbcast++;
+ else
+ rx_grp->unicast++;
+
+ if (!__vlan_get_tag(skb, &vlan_tci))
+ rx_grp->vlan++;
+ else
+ rx_grp->untagged++;
+}
+
+/* This function is overloaded for hfi_vnic specific implementation */
+static struct rtnl_link_stats64 *
+hfi1_vnic_get_stats64(struct net_device *netdev,
+ struct rtnl_link_stats64 *stats)
+{
+ struct hfi_vnic_stats *vstats = (struct hfi_vnic_stats *)stats;
+ struct hfi1_vnic_vport_info *vinfo = hfi_vnic_dev_priv(netdev);
+
+ hfi_vnic_update_stats(vinfo, vstats);
+ return stats;
+}
+
+static u64 create_bypass_pbc(u32 vl, u32 dw_len)
+{
+ u64 pbc;
+
+ pbc = ((u64)PBC_IHCRC_NONE << PBC_INSERT_HCRC_SHIFT)
+ | PBC_INSERT_BYPASS_ICRC | PBC_CREDIT_RETURN
+ | PBC_PACKET_BYPASS
+ | ((vl & PBC_VL_MASK) << PBC_VL_SHIFT)
+ | (dw_len & PBC_LENGTH_DWS_MASK) << PBC_LENGTH_DWS_SHIFT;
+
+ return pbc;
+}
+
+/* hfi1_vnic_maybe_stop_tx - stop tx queue if required */
+static void hfi1_vnic_maybe_stop_tx(struct hfi1_vnic_vport_info *vinfo,
+ u8 q_idx)
+{
+ netif_stop_subqueue(vinfo->netdev, q_idx);
+}
+
+static netdev_tx_t hfi1_netdev_start_xmit(struct sk_buff *skb,
+ struct net_device *netdev)
+{
+ struct hfi1_vnic_vport_info *vinfo = hfi_vnic_dev_priv(netdev);
+ u8 pad_len, q_idx = skb->queue_mapping;
+ struct hfi1_devdata *dd = vinfo->dd;
+ struct hfi_vnic_skb_mdata *mdata;
+ u32 pkt_len, total_len;
+ int err = -EINVAL;
+ u64 pbc;
+
+ v_dbg("xmit: queue %d skb len %d\n", q_idx, skb->len);
+ if (!netif_oper_up(netdev)) {
+ vinfo->stats[q_idx].tx_drop_state++;
+ goto tx_finish;
+ }
+
+ /* take out meta data */
+ mdata = (struct hfi_vnic_skb_mdata *)skb->data;
+ skb_pull(skb, sizeof(*mdata));
+ if (unlikely(mdata->flags & HFI_VNIC_SKB_MDATA_ENCAP_ERR)) {
+ vinfo->stats[q_idx].tx_dlid_zero++;
+ goto tx_finish;
+ }
+
+ /* add tail padding (for 8 bytes size alignment) and icrc */
+ pad_len = -(skb->len + HFI_VNIC_ICRC_TAIL_LEN) & 0x7;
+ pad_len += HFI_VNIC_ICRC_TAIL_LEN;
+
+ /*
+ * pkt_len is how much data we have to write, includes header and data.
+ * total_len is length of the packet in Dwords plus the PBC should not
+ * include the CRC.
+ */
+ pkt_len = (skb->len + pad_len) >> 2;
+ total_len = pkt_len + 2; /* PBC + packet */
+
+ pbc = create_bypass_pbc(mdata->vl, total_len);
+
+ skb_get(skb);
+ v_dbg("pbc 0x%016llX len %d pad_len %d\n", pbc, skb->len, pad_len);
+ err = dd->process_vnic_dma_send(dd, q_idx, vinfo, skb, pbc, pad_len);
+ if (err) {
+ if (err == -ENOMEM)
+ vinfo->stats[q_idx].netstats.tx_fifo_errors++;
+ else if (err != -EBUSY)
+ vinfo->stats[q_idx].netstats.tx_carrier_errors++;
+ }
+ /* remove the header before updating tx counters */
+ skb_pull(skb, HFI_VNIC_HDR_LEN);
+
+ if (unlikely(err == -EBUSY)) {
+ hfi1_vnic_maybe_stop_tx(vinfo, q_idx);
+ dev_kfree_skb_any(skb);
+ return NETDEV_TX_BUSY;
+ }
+
+tx_finish:
+ /* update tx counters */
+ hfi_vnic_update_tx_counters(vinfo, q_idx, skb, err);
+ dev_kfree_skb_any(skb);
+ return NETDEV_TX_OK;
+}
+
+static u16 hfi1_vnic_select_queue(struct net_device *netdev,
+ struct sk_buff *skb,
+ void *accel_priv,
+ select_queue_fallback_t fallback)
+{
+ return 0;
+}
+
+/* hfi1_vnic_decap_skb - strip OPA header from the skb (ethernet) packet */
+static inline int hfi_vnic_decap_skb(struct hfi1_vnic_rx_queue *rxq,
+ struct sk_buff *skb)
+{
+ struct hfi1_vnic_vport_info *vinfo = rxq->vinfo;
+ int max_len = vinfo->netdev->mtu + VLAN_ETH_HLEN;
+ int rc = -EFAULT;
+
+ skb_pull(skb, HFI_VNIC_HDR_LEN);
+
+ /* Validate Packet length */
+ if (skb->len > max_len)
+ vinfo->stats[rxq->idx].rx_oversize++;
+ else if (skb->len < ETH_ZLEN)
+ vinfo->stats[rxq->idx].rx_runt++;
+ else
+ rc = 0;
+ return rc;
+}
+
+static inline struct sk_buff *hfi1_vnic_get_skb(struct hfi1_vnic_rx_queue *rxq)
+{
+ unsigned char *pad_info;
+ struct sk_buff *skb;
+
+ skb = skb_dequeue(&rxq->skbq);
+ if (!skb)
+ return NULL;
+
+ /* remove tail padding and icrc */
+ pad_info = skb->data + skb->len - 1;
+ skb_trim(skb, (skb->len - HFI_VNIC_ICRC_TAIL_LEN -
+ ((*pad_info) & 0x7)));
+
+ return skb;
+}
+
+/* hfi1_vnic_handle_rx - handle skb receive */
+static void hfi1_vnic_handle_rx(struct hfi1_vnic_rx_queue *rxq,
+ int *work_done, int work_to_do)
+{
+ struct hfi1_vnic_vport_info *vinfo = rxq->vinfo;
+ struct sk_buff *skb;
+ int rc;
+
+ while (1) {
+ if (*work_done >= work_to_do)
+ break;
+
+ skb = hfi1_vnic_get_skb(rxq);
+ if (!skb)
+ break;
+
+ rc = hfi_vnic_decap_skb(rxq, skb);
+ /* update rx counters */
+ hfi_vnic_update_rx_counters(vinfo, rxq->idx, skb, rc);
+ if (rc) {
+ dev_kfree_skb_any(skb);
+ continue;
+ }
+
+ skb_checksum_none_assert(skb);
+ skb->protocol = eth_type_trans(skb, rxq->netdev);
+
+ napi_gro_receive(&rxq->napi, skb);
+ (*work_done)++;
+ }
+}
+
+/* hfi1_vnic_napi - napi receive polling callback function */
+static int hfi1_vnic_napi(struct napi_struct *napi, int budget)
+{
+ struct hfi1_vnic_rx_queue *rxq = container_of(napi,
+ struct hfi1_vnic_rx_queue, napi);
+ struct hfi1_vnic_vport_info *vinfo = rxq->vinfo;
+ int work_done = 0;
+
+ v_dbg("napi %d budget %d\n", rxq->idx, budget);
+ hfi1_vnic_handle_rx(rxq, &work_done, budget);
+
+ v_dbg("napi %d work_done %d\n", rxq->idx, work_done);
+ if (work_done < budget)
+ napi_complete(napi);
+
+ return work_done;
+}
+
+void hfi1_vnic_bypass_rcv(struct hfi1_packet *packet)
+{
+ struct hfi1_devdata *dd = packet->rcd->dd;
+ struct hfi1_vnic_vport_info *vinfo = NULL;
+ struct hfi1_vnic_rx_queue *rxq;
+ struct sk_buff *skb;
+ int l4_type, vesw_id = -1;
+ u8 q_idx;
+
+ l4_type = HFI1_GET_L4_TYPE(packet->ebuf);
+ if (l4_type == HFI_VNIC_L4_ETHR) {
+ vesw_id = HFI1_VNIC_GET_VESWID(packet->ebuf);
+ vinfo = idr_find(&dd->vnic.vesw_idr, vesw_id);
+
+ /*
+ * In case of invalid vesw id, count the error on
+ * the first available vport.
+ */
+ if (unlikely(!vinfo)) {
+ struct hfi1_vnic_vport_info *vinfo_tmp;
+ int id_tmp = 0;
+
+ vinfo_tmp = idr_get_next(&dd->vnic.vesw_idr, &id_tmp);
+ if (vinfo_tmp) {
+ spin_lock(&vport_cntr_lock);
+ vinfo_tmp->stats[0].netstats.rx_nohandler++;
+ spin_unlock(&vport_cntr_lock);
+ }
+ }
+ }
+
+ if (unlikely(!vinfo)) {
+ dd_dev_warn(dd, "vnic rcv err: l4 %d vesw id %d ctx %d\n",
+ l4_type, vesw_id, packet->rcd->ctxt);
+ return;
+ }
+
+ q_idx = packet->rcd->vnic_q_idx;
+ rxq = &vinfo->rxq[q_idx];
+ if (!netif_oper_up(vinfo->netdev)) {
+ vinfo->stats[q_idx].rx_drop_state++;
+ skb_queue_purge(&rxq->skbq);
+ return;
+ }
+
+ if (skb_queue_len(&rxq->skbq) > HFI1_VNIC_RCV_Q_SIZE) {
+ vinfo->stats[q_idx].netstats.rx_fifo_errors++;
+ return;
+ }
+
+ skb = netdev_alloc_skb(vinfo->netdev, packet->tlen);
+ if (!skb) {
+ vinfo->stats[q_idx].netstats.rx_fifo_errors++;
+ return;
+ }
+
+ memcpy(skb->data, packet->ebuf, packet->tlen);
+ skb_put(skb, packet->tlen);
+ skb_queue_tail(&rxq->skbq, skb);
+
+ if (napi_schedule_prep(&rxq->napi)) {
+ v_dbg("napi %d scheduling\n", q_idx);
+ __napi_schedule(&rxq->napi);
+ }
+}
+
+static int hfi1_vnic_up(struct hfi1_vnic_vport_info *vinfo)
+{
+ struct hfi1_devdata *dd = vinfo->dd;
+ struct net_device *netdev = vinfo->netdev;
+ int i, rc;
+
+ /* ensure virtual eth switch id is valid */
+ if (!vinfo->vesw_id)
+ return -EINVAL;
+
+ rc = idr_alloc(&dd->vnic.vesw_idr, vinfo, vinfo->vesw_id,
+ vinfo->vesw_id + 1, GFP_NOWAIT);
+ if (rc < 0)
+ return rc;
+
+ for (i = 0; i < vinfo->num_rx_q; i++) {
+ struct hfi1_vnic_rx_queue *rxq = &vinfo->rxq[i];
+
+ skb_queue_head_init(&rxq->skbq);
+ napi_enable(&rxq->napi);
+ }
+
+ netif_carrier_on(netdev);
+ netif_tx_start_all_queues(netdev);
+ set_bit(HFI1_VNIC_UP, &vinfo->flags);
+
+ return 0;
+}
+
+static void hfi1_vnic_down(struct hfi1_vnic_vport_info *vinfo)
+{
+ struct hfi1_devdata *dd = vinfo->dd;
+ u8 i;
+
+ clear_bit(HFI1_VNIC_UP, &vinfo->flags);
+ netif_carrier_off(vinfo->netdev);
+ netif_tx_disable(vinfo->netdev);
+ idr_remove(&dd->vnic.vesw_idr, vinfo->vesw_id);
+
+ /* remove unread skbs */
+ for (i = 0; i < vinfo->num_rx_q; i++) {
+ struct hfi1_vnic_rx_queue *rxq = &vinfo->rxq[i];
+
+ napi_disable(&rxq->napi);
+ skb_queue_purge(&rxq->skbq);
+ }
+}
+
+static int hfi1_netdev_open(struct net_device *netdev)
+{
+ struct hfi1_vnic_vport_info *vinfo = hfi_vnic_dev_priv(netdev);
+ int rc;
+
+ mutex_lock(&vinfo->lock);
+ rc = hfi1_vnic_up(vinfo);
+ mutex_unlock(&vinfo->lock);
+ return rc;
+}
+
+static int hfi1_netdev_close(struct net_device *netdev)
+{
+ struct hfi1_vnic_vport_info *vinfo = hfi_vnic_dev_priv(netdev);
+
+ mutex_lock(&vinfo->lock);
+ if (test_bit(HFI1_VNIC_UP, &vinfo->flags))
+ hfi1_vnic_down(vinfo);
+ mutex_unlock(&vinfo->lock);
+ return 0;
+}
+
+static void hfi1_vnic_set_vesw_id(struct net_device *netdev, int id)
+{
+ struct hfi1_vnic_vport_info *vinfo = hfi_vnic_dev_priv(netdev);
+ bool reopen = false;
+
+ /*
+ * If vesw_id is being changed, and if the vnic port is up,
+ * reset the vnic port to ensure new vesw_id gets picked up
+ */
+ if (id != vinfo->vesw_id) {
+ mutex_lock(&vinfo->lock);
+ if (test_bit(HFI1_VNIC_UP, &vinfo->flags)) {
+ hfi1_vnic_down(vinfo);
+ reopen = true;
+ }
+
+ vinfo->vesw_id = id;
+ if (reopen)
+ hfi1_vnic_up(vinfo);
+
+ mutex_unlock(&vinfo->lock);
+ }
+}
+
+/* netdev ops */
+static const struct net_device_ops hfi1_netdev_ops = {
+ .ndo_open = hfi1_netdev_open,
+ .ndo_stop = hfi1_netdev_close,
+ .ndo_start_xmit = hfi1_netdev_start_xmit,
+ .ndo_select_queue = hfi1_vnic_select_queue,
+ .ndo_get_stats64 = hfi1_vnic_get_stats64,
+};
+
+struct net_device *hfi1_vnic_alloc_rn(struct ib_device *device,
+ u8 port_num,
+ enum rdma_netdev_t type,
+ const char *name,
+ unsigned char name_assign_type,
+ void (*setup)(struct net_device *))
+{
+ struct hfi1_devdata *dd = dd_from_ibdev(device);
+ struct hfi1_vnic_vport_info *vinfo;
+ struct net_device *netdev;
+ struct rdma_netdev *rn;
+ int i, size;
+
+ if (!port_num || (port_num > dd->num_pports))
+ return ERR_PTR(-EINVAL);
+
+ if (type != RDMA_NETDEV_HFI_VNIC)
+ return ERR_PTR(-EINVAL);
+
+ size = sizeof(struct hfi_vnic_rdma_netdev) + sizeof(*vinfo);
+ netdev = alloc_netdev_mqs(size, name, name_assign_type, setup,
+ dd->chip_sdma_engines, HFI1_NUM_VNIC_CTXT);
+ if (!netdev)
+ return ERR_PTR(-ENOMEM);
+
+ rn = netdev_priv(netdev);
+ vinfo = hfi_vnic_dev_priv(netdev);
+ vinfo->dd = dd;
+ vinfo->num_tx_q = dd->chip_sdma_engines;
+ vinfo->num_rx_q = HFI1_NUM_VNIC_CTXT;
+ vinfo->netdev = netdev;
+ rn->set_id = hfi1_vnic_set_vesw_id;
+
+ netdev->features = NETIF_F_HIGHDMA | NETIF_F_SG;
+ netdev->hw_features = netdev->features;
+ netdev->vlan_features = netdev->features;
+ netdev->watchdog_timeo = msecs_to_jiffies(HFI_TX_TIMEOUT_MS);
+ netdev->netdev_ops = &hfi1_netdev_ops;
+ mutex_init(&vinfo->lock);
+
+ for (i = 0; i < vinfo->num_rx_q; i++) {
+ struct hfi1_vnic_rx_queue *rxq = &vinfo->rxq[i];
+
+ rxq->idx = i;
+ rxq->vinfo = vinfo;
+ rxq->netdev = netdev;
+ netif_napi_add(netdev, &rxq->napi, hfi1_vnic_napi, 64);
+ }
+
+ return netdev;
+}
+
+void hfi1_vnic_free_rn(struct net_device *netdev)
+{
+ struct hfi1_vnic_vport_info *vinfo = hfi_vnic_dev_priv(netdev);
+
+ mutex_destroy(&vinfo->lock);
+ free_netdev(netdev);
+}
--
1.8.3.1
^ permalink raw reply related [flat|nested] 40+ messages in thread* [RFC v3 10/11] IB/hfi1: Virtual Network Interface Controller (VNIC) HW support
2017-02-07 20:22 [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
` (5 preceding siblings ...)
2017-02-07 20:23 ` [RFC v3 09/11] IB/hfi1: HFI_VNIC RDMA netdev support Vishwanathapura, Niranjana
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
2017-02-07 20:23 ` [RFC v3 11/11] IB/hfi1: VNIC SDMA support Vishwanathapura, Niranjana
[not found] ` <1486498990-76562-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
8 siblings, 0 replies; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura, Andrzej Kacprowski
HFI1 HW specific support for VNIC functionality.
Dynamically allocate a set of contexts for VNIC when the first vnic
port is instantiated. Allocate VNIC contexts from user contexts pool
and return them back to the same pool while freeing up. Set aside
enough MSI-X interrupts for VNIC contexts and assign them when the
contexts are allocated. On the receive side, use an RSM rule to
spread TCP/UDP streams among VNIC contexts.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Andrzej Kacprowski <andrzej.kacprowski@intel.com>
---
drivers/infiniband/hw/hfi1/aspm.h | 15 +-
drivers/infiniband/hw/hfi1/chip.c | 293 +++++++++++++++++++++++++-----
drivers/infiniband/hw/hfi1/chip.h | 4 +-
drivers/infiniband/hw/hfi1/debugfs.c | 8 +-
drivers/infiniband/hw/hfi1/driver.c | 52 ++++--
drivers/infiniband/hw/hfi1/file_ops.c | 27 ++-
drivers/infiniband/hw/hfi1/hfi.h | 29 ++-
drivers/infiniband/hw/hfi1/init.c | 29 +--
drivers/infiniband/hw/hfi1/mad.c | 10 +-
drivers/infiniband/hw/hfi1/pio.c | 19 +-
drivers/infiniband/hw/hfi1/pio.h | 8 +-
drivers/infiniband/hw/hfi1/sysfs.c | 4 +-
drivers/infiniband/hw/hfi1/user_exp_rcv.c | 8 +-
drivers/infiniband/hw/hfi1/user_pages.c | 5 +-
drivers/infiniband/hw/hfi1/verbs.c | 8 +-
drivers/infiniband/hw/hfi1/vnic.h | 3 +
drivers/infiniband/hw/hfi1/vnic_main.c | 245 ++++++++++++++++++++++++-
include/rdma/opa_port_info.h | 4 +-
18 files changed, 663 insertions(+), 108 deletions(-)
diff --git a/drivers/infiniband/hw/hfi1/aspm.h b/drivers/infiniband/hw/hfi1/aspm.h
index 0d58fe3..794e681 100644
--- a/drivers/infiniband/hw/hfi1/aspm.h
+++ b/drivers/infiniband/hw/hfi1/aspm.h
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -229,14 +229,17 @@ static inline void aspm_ctx_timer_function(unsigned long data)
spin_unlock_irqrestore(&rcd->aspm_lock, flags);
}
-/* Disable interrupt processing for verbs contexts when PSM contexts are open */
+/*
+ * Disable interrupt processing for verbs contexts when PSM or VNIC contexts
+ * are open.
+ */
static inline void aspm_disable_all(struct hfi1_devdata *dd)
{
struct hfi1_ctxtdata *rcd;
unsigned long flags;
unsigned i;
- for (i = 0; i < dd->first_user_ctxt; i++) {
+ for (i = 0; i < dd->first_dyn_alloc_ctxt; i++) {
rcd = dd->rcd[i];
del_timer_sync(&rcd->aspm_timer);
spin_lock_irqsave(&rcd->aspm_lock, flags);
@@ -260,7 +263,7 @@ static inline void aspm_enable_all(struct hfi1_devdata *dd)
if (aspm_mode != ASPM_MODE_DYNAMIC)
return;
- for (i = 0; i < dd->first_user_ctxt; i++) {
+ for (i = 0; i < dd->first_dyn_alloc_ctxt; i++) {
rcd = dd->rcd[i];
spin_lock_irqsave(&rcd->aspm_lock, flags);
rcd->aspm_intr_enable = true;
@@ -276,7 +279,7 @@ static inline void aspm_ctx_init(struct hfi1_ctxtdata *rcd)
(unsigned long)rcd);
rcd->aspm_intr_supported = rcd->dd->aspm_supported &&
aspm_mode == ASPM_MODE_DYNAMIC &&
- rcd->ctxt < rcd->dd->first_user_ctxt;
+ rcd->ctxt < rcd->dd->first_dyn_alloc_ctxt;
}
static inline void aspm_init(struct hfi1_devdata *dd)
@@ -286,7 +289,7 @@ static inline void aspm_init(struct hfi1_devdata *dd)
spin_lock_init(&dd->aspm_lock);
dd->aspm_supported = aspm_hw_l1_supported(dd);
- for (i = 0; i < dd->first_user_ctxt; i++)
+ for (i = 0; i < dd->first_dyn_alloc_ctxt; i++)
aspm_ctx_init(dd->rcd[i]);
/* Start with ASPM disabled */
diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c
index ef72bc2..fe19066 100644
--- a/drivers/infiniband/hw/hfi1/chip.c
+++ b/drivers/infiniband/hw/hfi1/chip.c
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -125,9 +125,16 @@ struct flag_table {
#define DEFAULT_KRCVQS 2
#define MIN_KERNEL_KCTXTS 2
#define FIRST_KERNEL_KCTXT 1
-/* sizes for both the QP and RSM map tables */
-#define NUM_MAP_ENTRIES 256
-#define NUM_MAP_REGS 32
+
+/*
+ * RSM instance allocation
+ * 0 - Verbs
+ * 1 - User Fecn Handling
+ * 2 - Vnic
+ */
+#define RSM_INS_VERBS 0
+#define RSM_INS_FECN 1
+#define RSM_INS_VNIC 2
/* Bit offset into the GUID which carries HFI id information */
#define GUID_HFI_INDEX_SHIFT 39
@@ -138,8 +145,7 @@ struct flag_table {
#define is_emulator_p(dd) ((((dd)->irev) & 0xf) == 3)
#define is_emulator_s(dd) ((((dd)->irev) & 0xf) == 4)
-/* RSM fields */
-
+/* RSM fields for Verbs */
/* packet type */
#define IB_PACKET_TYPE 2ull
#define QW_SHIFT 6ull
@@ -169,6 +175,28 @@ struct flag_table {
/* QPN[m+n:1] QW 1, OFFSET 1 */
#define QPN_SELECT_OFFSET ((1ull << QW_SHIFT) | (1ull))
+/* RSM fields for Vnic */
+/* L2_TYPE: QW 0, OFFSET 61 - for match */
+#define L2_TYPE_QW 0ull
+#define L2_TYPE_BIT_OFFSET 61ull
+#define L2_TYPE_OFFSET(off) ((L2_TYPE_QW << QW_SHIFT) | (off))
+#define L2_TYPE_MATCH_OFFSET L2_TYPE_OFFSET(L2_TYPE_BIT_OFFSET)
+#define L2_TYPE_MASK 3ull
+#define L2_16B_VALUE 2ull
+
+/* L4_TYPE QW 1, OFFSET 0 - for match */
+#define L4_TYPE_QW 1ull
+#define L4_TYPE_BIT_OFFSET 0ull
+#define L4_TYPE_OFFSET(off) ((L4_TYPE_QW << QW_SHIFT) | (off))
+#define L4_TYPE_MATCH_OFFSET L4_TYPE_OFFSET(L4_TYPE_BIT_OFFSET)
+#define L4_16B_TYPE_MASK 0xFFull
+#define L4_16B_ETH_VALUE 0x78ull
+
+/* 16B VESWID - for select */
+#define L4_16B_HDR_VESWID_OFFSET ((2 << QW_SHIFT) | (16ull))
+/* 16B ENTROPY - for select */
+#define L2_16B_ENTROPY_OFFSET ((1 << QW_SHIFT) | (32ull))
+
/* defines to build power on SC2VL table */
#define SC2VL_VAL( \
num, \
@@ -1045,6 +1073,7 @@ static int wait_logical_linkstate(struct hfi1_pportdata *ppd, u32 state,
static int qos_rmt_entries(struct hfi1_devdata *dd, unsigned int *mp,
unsigned int *np);
static void clear_full_mgmt_pkey(struct hfi1_pportdata *ppd);
+static void clear_rsm_rule(struct hfi1_devdata *dd, u8 rule_index);
/*
* Error interrupt table entry. This is used as input to the interrupt
@@ -6701,7 +6730,13 @@ static void rxe_kernel_unfreeze(struct hfi1_devdata *dd)
int i;
/* enable all kernel contexts */
- for (i = 0; i < dd->n_krcv_queues; i++) {
+ for (i = 0; i < dd->num_rcv_contexts; i++) {
+ struct hfi1_ctxtdata *rcd = dd->rcd[i];
+
+ /* Ensure all non-user contexts(including vnic) are enabled */
+ if (!rcd || !rcd->sc || (rcd->sc->type == SC_USER))
+ continue;
+
rcvmask = HFI1_RCVCTRL_CTXT_ENB;
/* HFI1_RCVCTRL_TAILUPD_[ENB|DIS] needs to be set explicitly */
rcvmask |= HFI1_CAP_KGET_MASK(dd->rcd[i]->flags, DMA_RTAIL) ?
@@ -7993,7 +8028,9 @@ static void is_rcv_avail_int(struct hfi1_devdata *dd, unsigned int source)
if (likely(source < dd->num_rcv_contexts)) {
rcd = dd->rcd[source];
if (rcd) {
- if (source < dd->first_user_ctxt)
+ /* Check for non-user contexts, including vnic */
+ if ((source < dd->first_dyn_alloc_ctxt) ||
+ (rcd->sc && (rcd->sc->type == SC_KERNEL)))
rcd->do_interrupt(rcd, 0);
else
handle_user_interrupt(rcd);
@@ -8021,7 +8058,8 @@ static void is_rcv_urgent_int(struct hfi1_devdata *dd, unsigned int source)
rcd = dd->rcd[source];
if (rcd) {
/* only pay attention to user urgent interrupts */
- if (source >= dd->first_user_ctxt)
+ if ((source >= dd->first_dyn_alloc_ctxt) &&
+ (!rcd->sc || (rcd->sc->type == SC_USER)))
handle_user_interrupt(rcd);
return; /* OK */
}
@@ -12722,7 +12760,10 @@ static int request_msix_irqs(struct hfi1_devdata *dd)
first_sdma = last_general;
last_sdma = first_sdma + dd->num_sdma;
first_rx = last_sdma;
- last_rx = first_rx + dd->n_krcv_queues;
+ last_rx = first_rx + dd->n_krcv_queues + HFI1_NUM_VNIC_CTXT;
+
+ /* VNIC MSIx interrupts get mapped when VNIC contexts are created */
+ dd->first_dyn_msix_idx = first_rx + dd->n_krcv_queues;
/*
* Sanity check - the code expects all SDMA chip source
@@ -12736,7 +12777,7 @@ static int request_msix_irqs(struct hfi1_devdata *dd)
const char *err_info;
irq_handler_t handler;
irq_handler_t thread = NULL;
- void *arg;
+ void *arg = NULL;
int idx;
struct hfi1_ctxtdata *rcd = NULL;
struct sdma_engine *sde = NULL;
@@ -12763,24 +12804,25 @@ static int request_msix_irqs(struct hfi1_devdata *dd)
} else if (first_rx <= i && i < last_rx) {
idx = i - first_rx;
rcd = dd->rcd[idx];
- /* no interrupt if no rcd */
- if (!rcd)
- continue;
- /*
- * Set the interrupt register and mask for this
- * context's interrupt.
- */
- rcd->ireg = (IS_RCVAVAIL_START + idx) / 64;
- rcd->imask = ((u64)1) <<
- ((IS_RCVAVAIL_START + idx) % 64);
- handler = receive_context_interrupt;
- thread = receive_context_thread;
- arg = rcd;
- snprintf(me->name, sizeof(me->name),
- DRIVER_NAME "_%d kctxt%d", dd->unit, idx);
- err_info = "receive context";
- remap_intr(dd, IS_RCVAVAIL_START + idx, i);
- me->type = IRQ_RCVCTXT;
+ if (rcd) {
+ /*
+ * Set the interrupt register and mask for this
+ * context's interrupt.
+ */
+ rcd->ireg = (IS_RCVAVAIL_START + idx) / 64;
+ rcd->imask = ((u64)1) <<
+ ((IS_RCVAVAIL_START + idx) % 64);
+ handler = receive_context_interrupt;
+ thread = receive_context_thread;
+ arg = rcd;
+ snprintf(me->name, sizeof(me->name),
+ DRIVER_NAME "_%d kctxt%d",
+ dd->unit, idx);
+ err_info = "receive context";
+ remap_intr(dd, IS_RCVAVAIL_START + idx, i);
+ me->type = IRQ_RCVCTXT;
+ rcd->msix_intr = i;
+ }
} else {
/* not in our expected range - complain, then
* ignore it
@@ -12818,6 +12860,84 @@ static int request_msix_irqs(struct hfi1_devdata *dd)
return ret;
}
+void hfi1_vnic_synchronize_irq(struct hfi1_devdata *dd)
+{
+ int i;
+
+ if (!dd->num_msix_entries) {
+ synchronize_irq(dd->pcidev->irq);
+ return;
+ }
+
+ for (i = 0; i < dd->vnic.num_ctxt; i++) {
+ struct hfi1_ctxtdata *rcd = dd->vnic.ctxt[i];
+ struct hfi1_msix_entry *me = &dd->msix_entries[rcd->msix_intr];
+
+ synchronize_irq(me->msix.vector);
+ }
+}
+
+void hfi1_reset_vnic_msix_info(struct hfi1_ctxtdata *rcd)
+{
+ struct hfi1_devdata *dd = rcd->dd;
+ struct hfi1_msix_entry *me = &dd->msix_entries[rcd->msix_intr];
+
+ if (!me->arg) /* => no irq, no affinity */
+ return;
+
+ hfi1_put_irq_affinity(dd, me);
+ free_irq(me->msix.vector, me->arg);
+
+ me->arg = NULL;
+}
+
+void hfi1_set_vnic_msix_info(struct hfi1_ctxtdata *rcd)
+{
+ struct hfi1_devdata *dd = rcd->dd;
+ struct hfi1_msix_entry *me;
+ int idx = rcd->ctxt;
+ void *arg = rcd;
+ int ret;
+
+ rcd->msix_intr = dd->vnic.msix_idx++;
+ me = &dd->msix_entries[rcd->msix_intr];
+
+ /*
+ * Set the interrupt register and mask for this
+ * context's interrupt.
+ */
+ rcd->ireg = (IS_RCVAVAIL_START + idx) / 64;
+ rcd->imask = ((u64)1) <<
+ ((IS_RCVAVAIL_START + idx) % 64);
+
+ snprintf(me->name, sizeof(me->name),
+ DRIVER_NAME "_%d kctxt%d", dd->unit, idx);
+ me->name[sizeof(me->name) - 1] = 0;
+ me->type = IRQ_RCVCTXT;
+
+ remap_intr(dd, IS_RCVAVAIL_START + idx, rcd->msix_intr);
+
+ ret = request_threaded_irq(me->msix.vector, receive_context_interrupt,
+ receive_context_thread, 0, me->name, arg);
+ if (ret) {
+ dd_dev_err(dd, "vnic irq request (vector %d, idx %d) fail %d\n",
+ me->msix.vector, idx, ret);
+ return;
+ }
+ /*
+ * assign arg after request_irq call, so it will be
+ * cleaned up
+ */
+ me->arg = arg;
+
+ ret = hfi1_get_irq_affinity(dd, me);
+ if (ret) {
+ dd_dev_err(dd,
+ "unable to pin IRQ %d\n", ret);
+ free_irq(me->msix.vector, me->arg);
+ }
+}
+
/*
* Set the general handler to accept all interrupts, remap all
* chip interrupts back to MSI-X 0.
@@ -12849,7 +12969,7 @@ static int set_up_interrupts(struct hfi1_devdata *dd)
* N interrupts - one per used SDMA engine
* M interrupt - one per kernel receive context
*/
- total = 1 + dd->num_sdma + dd->n_krcv_queues;
+ total = 1 + dd->num_sdma + dd->n_krcv_queues + HFI1_NUM_VNIC_CTXT;
entries = kcalloc(total, sizeof(*entries), GFP_KERNEL);
if (!entries) {
@@ -12914,7 +13034,8 @@ static int set_up_interrupts(struct hfi1_devdata *dd)
*
* num_rcv_contexts - number of contexts being used
* n_krcv_queues - number of kernel contexts
- * first_user_ctxt - first non-kernel context in array of contexts
+ * first_dyn_alloc_ctxt - first dynamically allocated context
+ * in array of contexts
* freectxts - number of free user contexts
* num_send_contexts - number of PIO send contexts being used
*/
@@ -12991,10 +13112,14 @@ static int set_up_context_variables(struct hfi1_devdata *dd)
total_contexts = num_kernel_contexts + num_user_contexts;
}
- /* the first N are kernel contexts, the rest are user contexts */
+ /* Accommodate VNIC contexts */
+ if ((total_contexts + HFI1_NUM_VNIC_CTXT) <= dd->chip_rcv_contexts)
+ total_contexts += HFI1_NUM_VNIC_CTXT;
+
+ /* the first N are kernel contexts, the rest are user/vnic contexts */
dd->num_rcv_contexts = total_contexts;
dd->n_krcv_queues = num_kernel_contexts;
- dd->first_user_ctxt = num_kernel_contexts;
+ dd->first_dyn_alloc_ctxt = num_kernel_contexts;
dd->num_user_contexts = num_user_contexts;
dd->freectxts = num_user_contexts;
dd_dev_info(dd,
@@ -13450,11 +13575,8 @@ static void reset_rxe_csrs(struct hfi1_devdata *dd)
write_csr(dd, RCV_COUNTER_ARRAY32 + (8 * i), 0);
for (i = 0; i < RXE_NUM_64_BIT_COUNTERS; i++)
write_csr(dd, RCV_COUNTER_ARRAY64 + (8 * i), 0);
- for (i = 0; i < RXE_NUM_RSM_INSTANCES; i++) {
- write_csr(dd, RCV_RSM_CFG + (8 * i), 0);
- write_csr(dd, RCV_RSM_SELECT + (8 * i), 0);
- write_csr(dd, RCV_RSM_MATCH + (8 * i), 0);
- }
+ for (i = 0; i < RXE_NUM_RSM_INSTANCES; i++)
+ clear_rsm_rule(dd, i);
for (i = 0; i < 32; i++)
write_csr(dd, RCV_RSM_MAP_TABLE + (8 * i), 0);
@@ -13813,6 +13935,16 @@ static void add_rsm_rule(struct hfi1_devdata *dd, u8 rule_index,
(u64)rrd->value2 << RCV_RSM_MATCH_VALUE2_SHIFT);
}
+/*
+ * Clear a receive side mapping rule.
+ */
+static void clear_rsm_rule(struct hfi1_devdata *dd, u8 rule_index)
+{
+ write_csr(dd, RCV_RSM_CFG + (8 * rule_index), 0);
+ write_csr(dd, RCV_RSM_SELECT + (8 * rule_index), 0);
+ write_csr(dd, RCV_RSM_MATCH + (8 * rule_index), 0);
+}
+
/* return the number of RSM map table entries that will be used for QOS */
static int qos_rmt_entries(struct hfi1_devdata *dd, unsigned int *mp,
unsigned int *np)
@@ -13928,7 +14060,7 @@ static void init_qos(struct hfi1_devdata *dd, struct rsm_map_table *rmt)
rrd.value2 = LRH_SC_VALUE;
/* add rule 0 */
- add_rsm_rule(dd, 0, &rrd);
+ add_rsm_rule(dd, RSM_INS_VERBS, &rrd);
/* mark RSM map entries as used */
rmt->used += rmt_entries;
@@ -13958,7 +14090,7 @@ static void init_user_fecn_handling(struct hfi1_devdata *dd,
/*
* RSM will extract the destination context as an index into the
* map table. The destination contexts are a sequential block
- * in the range first_user_ctxt...num_rcv_contexts-1 (inclusive).
+ * in the range first_dyn_alloc_ctxt...num_rcv_contexts-1 (inclusive).
* Map entries are accessed as offset + extracted value. Adjust
* the added offset so this sequence can be placed anywhere in
* the table - as long as the entries themselves do not wrap.
@@ -13966,9 +14098,9 @@ static void init_user_fecn_handling(struct hfi1_devdata *dd,
* start with that to allow for a "negative" offset.
*/
offset = (u8)(NUM_MAP_ENTRIES + (int)rmt->used -
- (int)dd->first_user_ctxt);
+ (int)dd->first_dyn_alloc_ctxt);
- for (i = dd->first_user_ctxt, idx = rmt->used;
+ for (i = dd->first_dyn_alloc_ctxt, idx = rmt->used;
i < dd->num_rcv_contexts; i++, idx++) {
/* replace with identity mapping */
regoff = (idx % 8) * 8;
@@ -14002,11 +14134,84 @@ static void init_user_fecn_handling(struct hfi1_devdata *dd,
rrd.value2 = 1;
/* add rule 1 */
- add_rsm_rule(dd, 1, &rrd);
+ add_rsm_rule(dd, RSM_INS_FECN, &rrd);
rmt->used += dd->num_user_contexts;
}
+/* Initialize RSM for VNIC */
+void hfi1_init_vnic_rsm(struct hfi1_devdata *dd)
+{
+ u8 i, j;
+ u8 ctx_id = 0;
+ u64 reg;
+ u32 regoff;
+ struct rsm_rule_data rrd;
+
+ if (hfi1_vnic_is_rsm_full(dd, NUM_VNIC_MAP_ENTRIES)) {
+ dd_dev_err(dd, "Vnic RSM disabled, rmt entries used = %d\n",
+ dd->vnic.rmt_start);
+ return;
+ }
+
+ dev_dbg(&(dd)->pcidev->dev, "Vnic rsm start = %d, end %d\n",
+ dd->vnic.rmt_start,
+ dd->vnic.rmt_start + NUM_VNIC_MAP_ENTRIES);
+
+ /* Update RSM mapping table, 32 regs, 256 entries - 1 ctx per byte */
+ regoff = RCV_RSM_MAP_TABLE + (dd->vnic.rmt_start / 8) * 8;
+ reg = read_csr(dd, regoff);
+ for (i = 0; i < NUM_VNIC_MAP_ENTRIES; i++) {
+ /* Update map register with vnic context */
+ j = (dd->vnic.rmt_start + i) % 8;
+ reg &= ~(0xffllu << (j * 8));
+ reg |= (u64)dd->vnic.ctxt[ctx_id++]->ctxt << (j * 8);
+ /* Wrap up vnic ctx index */
+ ctx_id %= dd->vnic.num_ctxt;
+ /* Write back map register */
+ if (j == 7 || ((i + 1) == NUM_VNIC_MAP_ENTRIES)) {
+ dev_dbg(&(dd)->pcidev->dev,
+ "Vnic rsm map reg[%d] =0x%llx\n",
+ regoff - RCV_RSM_MAP_TABLE, reg);
+
+ write_csr(dd, regoff, reg);
+ regoff += 8;
+ if (i < (NUM_VNIC_MAP_ENTRIES - 1))
+ reg = read_csr(dd, regoff);
+ }
+ }
+
+ /* Add rule for vnic */
+ rrd.offset = dd->vnic.rmt_start;
+ rrd.pkt_type = 4;
+ /* Match 16B packets */
+ rrd.field1_off = L2_TYPE_MATCH_OFFSET;
+ rrd.mask1 = L2_TYPE_MASK;
+ rrd.value1 = L2_16B_VALUE;
+ /* Match ETH L4 packets */
+ rrd.field2_off = L4_TYPE_MATCH_OFFSET;
+ rrd.mask2 = L4_16B_TYPE_MASK;
+ rrd.value2 = L4_16B_ETH_VALUE;
+ /* Calc context from veswid and entropy */
+ rrd.index1_off = L4_16B_HDR_VESWID_OFFSET;
+ rrd.index1_width = ilog2(NUM_VNIC_MAP_ENTRIES);
+ rrd.index2_off = L2_16B_ENTROPY_OFFSET;
+ rrd.index2_width = ilog2(NUM_VNIC_MAP_ENTRIES);
+ add_rsm_rule(dd, RSM_INS_VNIC, &rrd);
+
+ /* Enable RSM if not already enabled */
+ add_rcvctrl(dd, RCV_CTRL_RCV_RSM_ENABLE_SMASK);
+}
+
+void hfi1_deinit_vnic_rsm(struct hfi1_devdata *dd)
+{
+ clear_rsm_rule(dd, RSM_INS_VNIC);
+
+ /* Disable RSM if used only by vnic */
+ if (dd->vnic.rmt_start == 0)
+ clear_rcvctrl(dd, RCV_CTRL_RCV_RSM_ENABLE_SMASK);
+}
+
static void init_rxe(struct hfi1_devdata *dd)
{
struct rsm_map_table *rmt;
@@ -14019,6 +14224,8 @@ static void init_rxe(struct hfi1_devdata *dd)
init_qos(dd, rmt);
init_user_fecn_handling(dd, rmt);
complete_rsm_map_table(dd, rmt);
+ /* record number of used rsm map entries for vnic */
+ dd->vnic.rmt_start = rmt->used;
kfree(rmt);
/*
diff --git a/drivers/infiniband/hw/hfi1/chip.h b/drivers/infiniband/hw/hfi1/chip.h
index 043fd21..11c2044 100644
--- a/drivers/infiniband/hw/hfi1/chip.h
+++ b/drivers/infiniband/hw/hfi1/chip.h
@@ -1,7 +1,7 @@
#ifndef _CHIP_H
#define _CHIP_H
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -1358,6 +1358,8 @@ void hfi1_put_tid(struct hfi1_devdata *dd, u32 index,
int hfi1_set_ctxt_pkey(struct hfi1_devdata *dd, unsigned ctxt, u16 pkey);
int hfi1_clear_ctxt_pkey(struct hfi1_devdata *dd, unsigned ctxt);
void hfi1_read_link_quality(struct hfi1_devdata *dd, u8 *link_quality);
+void hfi1_init_vnic_rsm(struct hfi1_devdata *dd);
+void hfi1_deinit_vnic_rsm(struct hfi1_devdata *dd);
/*
* Interrupt source table.
diff --git a/drivers/infiniband/hw/hfi1/debugfs.c b/drivers/infiniband/hw/hfi1/debugfs.c
index 8725f4c..f6da0b1 100644
--- a/drivers/infiniband/hw/hfi1/debugfs.c
+++ b/drivers/infiniband/hw/hfi1/debugfs.c
@@ -1,6 +1,6 @@
#ifdef CONFIG_DEBUG_FS
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -169,7 +169,7 @@ static int _opcode_stats_seq_show(struct seq_file *s, void *v)
struct hfi1_ibdev *ibd = (struct hfi1_ibdev *)s->private;
struct hfi1_devdata *dd = dd_from_dev(ibd);
- for (j = 0; j < dd->first_user_ctxt; j++) {
+ for (j = 0; j < dd->first_dyn_alloc_ctxt; j++) {
if (!dd->rcd[j])
continue;
n_packets += dd->rcd[j]->opstats->stats[i].n_packets;
@@ -195,7 +195,7 @@ static void *_ctx_stats_seq_start(struct seq_file *s, loff_t *pos)
if (!*pos)
return SEQ_START_TOKEN;
- if (*pos >= dd->first_user_ctxt)
+ if (*pos >= dd->first_dyn_alloc_ctxt)
return NULL;
return pos;
}
@@ -209,7 +209,7 @@ static void *_ctx_stats_seq_next(struct seq_file *s, void *v, loff_t *pos)
return pos;
++*pos;
- if (*pos >= dd->first_user_ctxt)
+ if (*pos >= dd->first_dyn_alloc_ctxt)
return NULL;
return pos;
}
diff --git a/drivers/infiniband/hw/hfi1/driver.c b/drivers/infiniband/hw/hfi1/driver.c
index 3c226a8..75b530e 100644
--- a/drivers/infiniband/hw/hfi1/driver.c
+++ b/drivers/infiniband/hw/hfi1/driver.c
@@ -861,20 +861,42 @@ int handle_receive_interrupt_dma_rtail(struct hfi1_ctxtdata *rcd, int thread)
return last;
}
-static inline void set_all_nodma_rtail(struct hfi1_devdata *dd)
+static inline void set_nodma_rtail(struct hfi1_devdata *dd, u8 ctxt)
{
int i;
- for (i = HFI1_CTRL_CTXT + 1; i < dd->first_user_ctxt; i++)
+ /*
+ * For dynamically allocated kernel contexts (like vnic) switch
+ * interrupt handler only for that context. Otherwise, switch
+ * interrupt handler for all statically allocated kernel contexts.
+ */
+ if (ctxt >= dd->first_dyn_alloc_ctxt) {
+ dd->rcd[ctxt]->do_interrupt =
+ &handle_receive_interrupt_nodma_rtail;
+ return;
+ }
+
+ for (i = HFI1_CTRL_CTXT + 1; i < dd->first_dyn_alloc_ctxt; i++)
dd->rcd[i]->do_interrupt =
&handle_receive_interrupt_nodma_rtail;
}
-static inline void set_all_dma_rtail(struct hfi1_devdata *dd)
+static inline void set_dma_rtail(struct hfi1_devdata *dd, u8 ctxt)
{
int i;
- for (i = HFI1_CTRL_CTXT + 1; i < dd->first_user_ctxt; i++)
+ /*
+ * For dynamically allocated kernel contexts (like vnic) switch
+ * interrupt handler only for that context. Otherwise, switch
+ * interrupt handler for all statically allocated kernel contexts.
+ */
+ if (ctxt >= dd->first_dyn_alloc_ctxt) {
+ dd->rcd[ctxt]->do_interrupt =
+ &handle_receive_interrupt_dma_rtail;
+ return;
+ }
+
+ for (i = HFI1_CTRL_CTXT + 1; i < dd->first_dyn_alloc_ctxt; i++)
dd->rcd[i]->do_interrupt =
&handle_receive_interrupt_dma_rtail;
}
@@ -884,8 +906,13 @@ void set_all_slowpath(struct hfi1_devdata *dd)
int i;
/* HFI1_CTRL_CTXT must always use the slow path interrupt handler */
- for (i = HFI1_CTRL_CTXT + 1; i < dd->first_user_ctxt; i++)
- dd->rcd[i]->do_interrupt = &handle_receive_interrupt;
+ for (i = HFI1_CTRL_CTXT + 1; i < dd->num_rcv_contexts; i++) {
+ struct hfi1_ctxtdata *rcd = dd->rcd[i];
+
+ if ((i < dd->first_dyn_alloc_ctxt) ||
+ (rcd && rcd->sc && (rcd->sc->type == SC_KERNEL)))
+ rcd->do_interrupt = &handle_receive_interrupt;
+ }
}
static inline int set_armed_to_active(struct hfi1_ctxtdata *rcd,
@@ -995,7 +1022,7 @@ int handle_receive_interrupt(struct hfi1_ctxtdata *rcd, int thread)
last = RCV_PKT_DONE;
if (needset) {
dd_dev_info(dd, "Switching to NO_DMA_RTAIL\n");
- set_all_nodma_rtail(dd);
+ set_nodma_rtail(dd, rcd->ctxt);
needset = 0;
}
} else {
@@ -1017,7 +1044,7 @@ int handle_receive_interrupt(struct hfi1_ctxtdata *rcd, int thread)
if (needset) {
dd_dev_info(dd,
"Switching to DMA_RTAIL\n");
- set_all_dma_rtail(dd);
+ set_dma_rtail(dd, rcd->ctxt);
needset = 0;
}
}
@@ -1065,10 +1092,10 @@ void receive_interrupt_work(struct work_struct *work)
set_link_state(ppd, HLS_UP_ACTIVE);
/*
- * Interrupt all kernel contexts that could have had an
- * interrupt during auto activation.
+ * Interrupt all statically allocated kernel contexts that could
+ * have had an interrupt during auto activation.
*/
- for (i = HFI1_CTRL_CTXT; i < dd->first_user_ctxt; i++)
+ for (i = HFI1_CTRL_CTXT; i < dd->first_dyn_alloc_ctxt; i++)
force_recv_intr(dd->rcd[i]);
}
@@ -1282,7 +1309,8 @@ int hfi1_reset_device(int unit)
spin_lock_irqsave(&dd->uctxt_lock, flags);
if (dd->rcd)
- for (i = dd->first_user_ctxt; i < dd->num_rcv_contexts; i++) {
+ for (i = dd->first_dyn_alloc_ctxt;
+ i < dd->num_rcv_contexts; i++) {
if (!dd->rcd[i] || !dd->rcd[i]->cnt)
continue;
spin_unlock_irqrestore(&dd->uctxt_lock, flags);
diff --git a/drivers/infiniband/hw/hfi1/file_ops.c b/drivers/infiniband/hw/hfi1/file_ops.c
index bd786b7..544d65e 100644
--- a/drivers/infiniband/hw/hfi1/file_ops.c
+++ b/drivers/infiniband/hw/hfi1/file_ops.c
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -585,8 +585,8 @@ static int hfi1_file_mmap(struct file *fp, struct vm_area_struct *vma)
* knows where it's own bitmap is within the page.
*/
memaddr = (unsigned long)(dd->events +
- ((uctxt->ctxt - dd->first_user_ctxt) *
- HFI1_MAX_SHARED_CTXTS)) & PAGE_MASK;
+ ((uctxt->ctxt - dd->first_dyn_alloc_ctxt) *
+ HFI1_MAX_SHARED_CTXTS)) & PAGE_MASK;
memlen = PAGE_SIZE;
/*
* v3.7 removes VM_RESERVED but the effect is kept by
@@ -755,7 +755,7 @@ static int hfi1_file_close(struct inode *inode, struct file *fp)
* Clear any left over, unhandled events so the next process that
* gets this context doesn't get confused.
*/
- ev = dd->events + ((uctxt->ctxt - dd->first_user_ctxt) *
+ ev = dd->events + ((uctxt->ctxt - dd->first_dyn_alloc_ctxt) *
HFI1_MAX_SHARED_CTXTS) + fdata->subctxt;
*ev = 0;
@@ -908,12 +908,18 @@ static int find_shared_ctxt(struct file *fp,
if (!(dd && (dd->flags & HFI1_PRESENT) && dd->kregbase))
continue;
- for (i = dd->first_user_ctxt; i < dd->num_rcv_contexts; i++) {
+ for (i = dd->first_dyn_alloc_ctxt;
+ i < dd->num_rcv_contexts; i++) {
struct hfi1_ctxtdata *uctxt = dd->rcd[i];
/* Skip ctxts which are not yet open */
if (!uctxt || !uctxt->cnt)
continue;
+
+ /* Skip dynamically allocted kernel contexts */
+ if (uctxt->sc && (uctxt->sc->type == SC_KERNEL))
+ continue;
+
/* Skip ctxt if it doesn't match the requested one */
if (memcmp(uctxt->uuid, uinfo->uuid,
sizeof(uctxt->uuid)) ||
@@ -959,7 +965,8 @@ static int allocate_ctxt(struct file *fp, struct hfi1_devdata *dd,
return -EIO;
}
- for (ctxt = dd->first_user_ctxt; ctxt < dd->num_rcv_contexts; ctxt++)
+ for (ctxt = dd->first_dyn_alloc_ctxt;
+ ctxt < dd->num_rcv_contexts; ctxt++)
if (!dd->rcd[ctxt])
break;
@@ -1305,7 +1312,7 @@ static int get_base_info(struct file *fp, void __user *ubase, __u32 len)
*/
binfo.user_regbase = HFI1_MMAP_TOKEN(UREGS, uctxt->ctxt,
fd->subctxt, 0);
- offset = offset_in_page((((uctxt->ctxt - dd->first_user_ctxt) *
+ offset = offset_in_page((((uctxt->ctxt - dd->first_dyn_alloc_ctxt) *
HFI1_MAX_SHARED_CTXTS) + fd->subctxt) *
sizeof(*dd->events));
binfo.events_bufbase = HFI1_MMAP_TOKEN(EVENTS, uctxt->ctxt,
@@ -1399,12 +1406,12 @@ int hfi1_set_uevent_bits(struct hfi1_pportdata *ppd, const int evtbit)
}
spin_lock_irqsave(&dd->uctxt_lock, flags);
- for (ctxt = dd->first_user_ctxt; ctxt < dd->num_rcv_contexts;
+ for (ctxt = dd->first_dyn_alloc_ctxt; ctxt < dd->num_rcv_contexts;
ctxt++) {
uctxt = dd->rcd[ctxt];
if (uctxt) {
unsigned long *evs = dd->events +
- (uctxt->ctxt - dd->first_user_ctxt) *
+ (uctxt->ctxt - dd->first_dyn_alloc_ctxt) *
HFI1_MAX_SHARED_CTXTS;
int i;
/*
@@ -1476,7 +1483,7 @@ static int user_event_ack(struct hfi1_ctxtdata *uctxt, int subctxt,
if (!dd->events)
return 0;
- evs = dd->events + ((uctxt->ctxt - dd->first_user_ctxt) *
+ evs = dd->events + ((uctxt->ctxt - dd->first_dyn_alloc_ctxt) *
HFI1_MAX_SHARED_CTXTS) + subctxt;
for (i = 0; i <= _HFI1_MAX_EVENT_BIT; i++) {
diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h
index a3641fb..39bcd46 100644
--- a/drivers/infiniband/hw/hfi1/hfi.h
+++ b/drivers/infiniband/hw/hfi1/hfi.h
@@ -54,6 +54,7 @@
#include <linux/list.h>
#include <linux/scatterlist.h>
#include <linux/slab.h>
+#include <linux/idr.h>
#include <linux/io.h>
#include <linux/fs.h>
#include <linux/completion.h>
@@ -66,6 +67,7 @@
#include <linux/i2c-algo-bit.h>
#include <rdma/ib_hdrs.h>
#include <linux/rhashtable.h>
+#include <linux/netdevice.h>
#include <rdma/rdma_vt.h>
#include "chip_registers.h"
@@ -278,6 +280,8 @@ struct hfi1_ctxtdata {
struct hfi1_devdata *dd;
/* so functions that need physical port can get it easily */
struct hfi1_pportdata *ppd;
+ /* associated msix interrupt */
+ u32 msix_intr;
/* A page of memory for rcvhdrhead, rcvegrhead, rcvegrtail * N */
void *subctxt_uregbase;
/* An array of pages for the eager receive buffers * N */
@@ -815,15 +819,27 @@ struct hfi1_asic_data {
struct hfi1_i2c_bus *i2c_bus1;
};
+/* sizes for both the QP and RSM map tables */
+#define NUM_MAP_ENTRIES 256
+#define NUM_MAP_REGS 32
+
/*
* Number of VNIC contexts used. Ensure it is less than or equal to
* max queues supported by VNIC (HFI_VNIC_MAX_QUEUE).
*/
#define HFI1_NUM_VNIC_CTXT 8
+/* Number of VNIC RSM entries */
+#define NUM_VNIC_MAP_ENTRIES 8
+
/* Virtual NIC information */
struct hfi1_vnic_data {
+ struct hfi1_ctxtdata *ctxt[HFI1_NUM_VNIC_CTXT];
+ u8 num_vports;
struct idr vesw_idr;
+ u8 rmt_start;
+ u8 num_ctxt;
+ u32 msix_idx;
};
struct hfi1_vnic_vport_info;
@@ -1051,6 +1067,7 @@ struct hfi1_devdata {
/* MSI-X information */
struct hfi1_msix_entry *msix_entries;
u32 num_msix_entries;
+ u32 first_dyn_msix_idx;
/* INTx information */
u32 requested_intx_irq; /* did we request one? */
@@ -1149,8 +1166,8 @@ struct hfi1_devdata {
u16 flags;
/* Number of physical ports available */
u8 num_pports;
- /* Lowest context number which can be used by user processes */
- u8 first_user_ctxt;
+ /* Lowest context number which can be used by user processes or VNIC */
+ u8 first_dyn_alloc_ctxt;
/* adding a new field here would make it part of this cacheline */
/* seqlock for sc2vl */
@@ -1198,6 +1215,11 @@ struct hfi1_devdata {
struct hfi1_vnic_data vnic;
};
+static inline bool hfi1_vnic_is_rsm_full(struct hfi1_devdata *dd, int spare)
+{
+ return (dd->vnic.rmt_start + spare) > NUM_MAP_ENTRIES;
+}
+
/* 8051 firmware version helper */
#define dc8051_ver(a, b) ((a) << 8 | (b))
#define dc8051_ver_maj(a) ((a & 0xff00) >> 8)
@@ -1261,6 +1283,9 @@ void hfi1_init_pportdata(struct pci_dev *, struct hfi1_pportdata *,
int handle_receive_interrupt_nodma_rtail(struct hfi1_ctxtdata *, int);
int handle_receive_interrupt_dma_rtail(struct hfi1_ctxtdata *, int);
void set_all_slowpath(struct hfi1_devdata *dd);
+void hfi1_vnic_synchronize_irq(struct hfi1_devdata *dd);
+void hfi1_set_vnic_msix_info(struct hfi1_ctxtdata *rcd);
+void hfi1_reset_vnic_msix_info(struct hfi1_ctxtdata *rcd);
extern const struct pci_device_id hfi1_pci_tbl[];
diff --git a/drivers/infiniband/hw/hfi1/init.c b/drivers/infiniband/hw/hfi1/init.c
index b1f48b1..e3a94d4 100644
--- a/drivers/infiniband/hw/hfi1/init.c
+++ b/drivers/infiniband/hw/hfi1/init.c
@@ -140,7 +140,7 @@ int hfi1_create_ctxts(struct hfi1_devdata *dd)
goto nomem;
/* create one or more kernel contexts */
- for (i = 0; i < dd->first_user_ctxt; ++i) {
+ for (i = 0; i < dd->first_dyn_alloc_ctxt; ++i) {
struct hfi1_pportdata *ppd;
struct hfi1_ctxtdata *rcd;
@@ -215,9 +215,9 @@ struct hfi1_ctxtdata *hfi1_create_ctxtdata(struct hfi1_pportdata *ppd, u32 ctxt,
u32 base;
if (dd->rcv_entries.nctxt_extra >
- dd->num_rcv_contexts - dd->first_user_ctxt)
+ dd->num_rcv_contexts - dd->first_dyn_alloc_ctxt)
kctxt_ngroups = (dd->rcv_entries.nctxt_extra -
- (dd->num_rcv_contexts - dd->first_user_ctxt));
+ (dd->num_rcv_contexts - dd->first_dyn_alloc_ctxt));
rcd = kzalloc_node(sizeof(*rcd), GFP_KERNEL, numa);
if (rcd) {
u32 rcvtids, max_entries;
@@ -239,10 +239,10 @@ struct hfi1_ctxtdata *hfi1_create_ctxtdata(struct hfi1_pportdata *ppd, u32 ctxt,
* Calculate the context's RcvArray entry starting point.
* We do this here because we have to take into account all
* the RcvArray entries that previous context would have
- * taken and we have to account for any extra groups
- * assigned to the kernel or user contexts.
+ * taken and we have to account for any extra groups assigned
+ * to the static (kernel) or dynamic (vnic/user) contexts.
*/
- if (ctxt < dd->first_user_ctxt) {
+ if (ctxt < dd->first_dyn_alloc_ctxt) {
if (ctxt < kctxt_ngroups) {
base = ctxt * (dd->rcv_entries.ngroups + 1);
rcd->rcv_array_groups++;
@@ -250,7 +250,7 @@ struct hfi1_ctxtdata *hfi1_create_ctxtdata(struct hfi1_pportdata *ppd, u32 ctxt,
base = kctxt_ngroups +
(ctxt * dd->rcv_entries.ngroups);
} else {
- u16 ct = ctxt - dd->first_user_ctxt;
+ u16 ct = ctxt - dd->first_dyn_alloc_ctxt;
base = ((dd->n_krcv_queues * dd->rcv_entries.ngroups) +
kctxt_ngroups);
@@ -322,7 +322,8 @@ struct hfi1_ctxtdata *hfi1_create_ctxtdata(struct hfi1_pportdata *ppd, u32 ctxt,
}
rcd->egrbufs.rcvtid_size = HFI1_MAX_EAGER_BUFFER_SIZE;
- if (ctxt < dd->first_user_ctxt) { /* N/A for PSM contexts */
+ /* Applicable only for statically created kernel contexts */
+ if (ctxt < dd->first_dyn_alloc_ctxt) {
rcd->opstats = kzalloc(sizeof(*rcd->opstats),
GFP_KERNEL);
if (!rcd->opstats)
@@ -585,7 +586,7 @@ static void enable_chip(struct hfi1_devdata *dd)
* Enable kernel ctxts' receive and receive interrupt.
* Other ctxts done as user opens and initializes them.
*/
- for (i = 0; i < dd->first_user_ctxt; ++i) {
+ for (i = 0; i < dd->first_dyn_alloc_ctxt; ++i) {
rcvmask = HFI1_RCVCTRL_CTXT_ENB | HFI1_RCVCTRL_INTRAVAIL_ENB;
rcvmask |= HFI1_CAP_KGET_MASK(dd->rcd[i]->flags, DMA_RTAIL) ?
HFI1_RCVCTRL_TAILUPD_ENB : HFI1_RCVCTRL_TAILUPD_DIS;
@@ -714,7 +715,7 @@ int hfi1_init(struct hfi1_devdata *dd, int reinit)
}
/* dd->rcd can be NULL if early initialization failed */
- for (i = 0; dd->rcd && i < dd->first_user_ctxt; ++i) {
+ for (i = 0; dd->rcd && i < dd->first_dyn_alloc_ctxt; ++i) {
/*
* Set up the (kernel) rcvhdr queue and egr TIDs. If doing
* re-init, the simplest way to handle this is to free
@@ -1533,6 +1534,7 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
hfi1_device_remove(dd);
if (!ret)
hfi1_unregister_ib_device(dd);
+ hfi1_vnic_cleanup(dd);
postinit_cleanup(dd);
if (initfail)
ret = initfail;
@@ -1619,8 +1621,11 @@ int hfi1_create_rcvhdrq(struct hfi1_devdata *dd, struct hfi1_ctxtdata *rcd)
amt = PAGE_ALIGN(rcd->rcvhdrq_cnt * rcd->rcvhdrqentsize *
sizeof(u32));
- gfp_flags = (rcd->ctxt >= dd->first_user_ctxt) ?
- GFP_USER : GFP_KERNEL;
+ if ((rcd->ctxt < dd->first_dyn_alloc_ctxt) ||
+ (rcd->sc && (rcd->sc->type == SC_KERNEL)))
+ gfp_flags = GFP_KERNEL;
+ else
+ gfp_flags = GFP_USER;
rcd->rcvhdrq = dma_zalloc_coherent(
&dd->pcidev->dev, amt, &rcd->rcvhdrq_dma,
gfp_flags | __GFP_COMP);
diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c
index 6e595af..0e48fb5 100644
--- a/drivers/infiniband/hw/hfi1/mad.c
+++ b/drivers/infiniband/hw/hfi1/mad.c
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -53,6 +53,7 @@
#include "mad.h"
#include "trace.h"
#include "qp.h"
+#include "vnic.h"
/* the reset value from the FM is supposed to be 0xffff, handle both */
#define OPA_LINK_WIDTH_RESET_OLD 0x0fff
@@ -650,9 +651,11 @@ static int __subn_get_opa_portinfo(struct opa_smp *smp, u32 am, u8 *data,
OPA_PI_MASK_PORT_ACTIVE_OPTOMIZE : 0);
pi->port_packet_format.supported =
- cpu_to_be16(OPA_PORT_PACKET_FORMAT_9B);
+ cpu_to_be16(OPA_PORT_PACKET_FORMAT_9B |
+ OPA_PORT_PACKET_FORMAT_16B);
pi->port_packet_format.enabled =
- cpu_to_be16(OPA_PORT_PACKET_FORMAT_9B);
+ cpu_to_be16(OPA_PORT_PACKET_FORMAT_9B |
+ OPA_PORT_PACKET_FORMAT_16B);
/* flit_control.interleave is (OPA V1, version .76):
* bits use
@@ -678,6 +681,7 @@ static int __subn_get_opa_portinfo(struct opa_smp *smp, u32 am, u8 *data,
pi->resptimevalue = 3;
pi->local_port_num = port;
+ pi->num_vesw_port_supported = HFI_MAX_NUM_VNICS;
/* buffer info for FM */
pi->overall_buffer_space = cpu_to_be16(dd->link_credits);
diff --git a/drivers/infiniband/hw/hfi1/pio.c b/drivers/infiniband/hw/hfi1/pio.c
index 615be68..ed72b5a 100644
--- a/drivers/infiniband/hw/hfi1/pio.c
+++ b/drivers/infiniband/hw/hfi1/pio.c
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -703,6 +703,7 @@ struct send_context *sc_alloc(struct hfi1_devdata *dd, int type,
{
struct send_context_info *sci;
struct send_context *sc = NULL;
+ int req_type = type;
dma_addr_t dma;
unsigned long flags;
u64 reg;
@@ -729,6 +730,13 @@ struct send_context *sc_alloc(struct hfi1_devdata *dd, int type,
return NULL;
}
+ /*
+ * VNIC contexts are dynamically allocated.
+ * Hence, pick a user context for VNIC.
+ */
+ if (type == SC_VNIC)
+ type = SC_USER;
+
spin_lock_irqsave(&dd->sc_lock, flags);
ret = sc_hw_alloc(dd, type, &sw_index, &hw_context);
if (ret) {
@@ -738,6 +746,15 @@ struct send_context *sc_alloc(struct hfi1_devdata *dd, int type,
return NULL;
}
+ /*
+ * VNIC contexts are used by kernel driver.
+ * Hence, mark them as kernel contexts.
+ */
+ if (req_type == SC_VNIC) {
+ dd->send_contexts[sw_index].type = SC_KERNEL;
+ type = SC_KERNEL;
+ }
+
sci = &dd->send_contexts[sw_index];
sci->sc = sc;
diff --git a/drivers/infiniband/hw/hfi1/pio.h b/drivers/infiniband/hw/hfi1/pio.h
index 867e5ff..a6fb700 100644
--- a/drivers/infiniband/hw/hfi1/pio.h
+++ b/drivers/infiniband/hw/hfi1/pio.h
@@ -1,7 +1,7 @@
#ifndef _PIO_H
#define _PIO_H
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -54,6 +54,12 @@
#define SC_USER 3 /* must be the last one: it may take all left */
#define SC_MAX 4 /* count of send context types */
+/*
+ * SC_VNIC types are allocated (dynamically) from the user context pool,
+ * (SC_USER) and used by kernel driver as kernel contexts (SC_KERNEL).
+ */
+#define SC_VNIC SC_MAX
+
/* invalid send context index */
#define INVALID_SCI 0xff
diff --git a/drivers/infiniband/hw/hfi1/sysfs.c b/drivers/infiniband/hw/hfi1/sysfs.c
index 919a547..50d140d 100644
--- a/drivers/infiniband/hw/hfi1/sysfs.c
+++ b/drivers/infiniband/hw/hfi1/sysfs.c
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -542,7 +542,7 @@ static ssize_t show_nctxts(struct device *device,
* give a more accurate picture of total contexts available.
*/
return scnprintf(buf, PAGE_SIZE, "%u\n",
- min(dd->num_rcv_contexts - dd->first_user_ctxt,
+ min(dd->num_rcv_contexts - dd->first_dyn_alloc_ctxt,
(u32)dd->sc_sizes[SC_USER].count));
}
diff --git a/drivers/infiniband/hw/hfi1/user_exp_rcv.c b/drivers/infiniband/hw/hfi1/user_exp_rcv.c
index 64d2652..a0a8d96 100644
--- a/drivers/infiniband/hw/hfi1/user_exp_rcv.c
+++ b/drivers/infiniband/hw/hfi1/user_exp_rcv.c
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -612,7 +612,7 @@ int hfi1_user_exp_rcv_invalid(struct file *fp, struct hfi1_tid_info *tinfo)
struct hfi1_filedata *fd = fp->private_data;
struct hfi1_ctxtdata *uctxt = fd->uctxt;
unsigned long *ev = uctxt->dd->events +
- (((uctxt->ctxt - uctxt->dd->first_user_ctxt) *
+ (((uctxt->ctxt - uctxt->dd->first_dyn_alloc_ctxt) *
HFI1_MAX_SHARED_CTXTS) + fd->subctxt);
u32 *array;
int ret = 0;
@@ -1016,8 +1016,8 @@ static int tid_rb_invalidate(void *arg, struct mmu_rb_node *mnode)
* process in question.
*/
ev = uctxt->dd->events +
- (((uctxt->ctxt - uctxt->dd->first_user_ctxt) *
- HFI1_MAX_SHARED_CTXTS) + fdata->subctxt);
+ (((uctxt->ctxt - uctxt->dd->first_dyn_alloc_ctxt) *
+ HFI1_MAX_SHARED_CTXTS) + fdata->subctxt);
set_bit(_HFI1_EVENT_TID_MMU_NOTIFY_BIT, ev);
}
fdata->invalid_tid_idx++;
diff --git a/drivers/infiniband/hw/hfi1/user_pages.c b/drivers/infiniband/hw/hfi1/user_pages.c
index 20f4ddc..efecd69 100644
--- a/drivers/infiniband/hw/hfi1/user_pages.c
+++ b/drivers/infiniband/hw/hfi1/user_pages.c
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -73,7 +73,8 @@ bool hfi1_can_pin_pages(struct hfi1_devdata *dd, struct mm_struct *mm,
{
unsigned long ulimit = rlimit(RLIMIT_MEMLOCK), pinned, cache_limit,
size = (cache_size * (1UL << 20)); /* convert to bytes */
- unsigned usr_ctxts = dd->num_rcv_contexts - dd->first_user_ctxt;
+ unsigned int usr_ctxts =
+ dd->num_rcv_contexts - dd->first_dyn_alloc_ctxt;
bool can_lock = capable(CAP_IPC_LOCK);
/*
diff --git a/drivers/infiniband/hw/hfi1/verbs.c b/drivers/infiniband/hw/hfi1/verbs.c
index 95ed4d6..4952781 100644
--- a/drivers/infiniband/hw/hfi1/verbs.c
+++ b/drivers/infiniband/hw/hfi1/verbs.c
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2015, 2016 Intel Corporation.
+ * Copyright(c) 2015-2017 Intel Corporation.
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
@@ -60,6 +60,7 @@
#include "trace.h"
#include "qp.h"
#include "verbs_txreq.h"
+#include "vnic.h"
static unsigned int hfi1_lkey_table_size = 16;
module_param_named(lkey_table_size, hfi1_lkey_table_size, uint,
@@ -1312,7 +1313,8 @@ static void hfi1_fill_device_attr(struct hfi1_devdata *dd)
IB_DEVICE_BAD_QKEY_CNTR | IB_DEVICE_SHUTDOWN_PORT |
IB_DEVICE_SYS_IMAGE_GUID | IB_DEVICE_RC_RNR_NAK_GEN |
IB_DEVICE_PORT_ACTIVE_EVENT | IB_DEVICE_SRQ_RESIZE |
- IB_DEVICE_MEM_MGT_EXTENSIONS;
+ IB_DEVICE_MEM_MGT_EXTENSIONS |
+ IB_DEVICE_RDMA_NETDEV_HFI_VNIC;
rdi->dparms.props.page_size_cap = PAGE_SIZE;
rdi->dparms.props.vendor_id = dd->oui1 << 16 | dd->oui2 << 8 | dd->oui3;
rdi->dparms.props.vendor_part_id = dd->pcidev->device;
@@ -1788,6 +1790,8 @@ int hfi1_register_ib_device(struct hfi1_devdata *dd)
ibdev->modify_device = modify_device;
ibdev->alloc_hw_stats = alloc_hw_stats;
ibdev->get_hw_stats = get_hw_stats;
+ ibdev->alloc_rdma_netdev = hfi1_vnic_alloc_rn;
+ ibdev->free_rdma_netdev = hfi1_vnic_free_rn;
/* keep process mad in the driver */
ibdev->process_mad = hfi1_process_mad;
diff --git a/drivers/infiniband/hw/hfi1/vnic.h b/drivers/infiniband/hw/hfi1/vnic.h
index 23b92d7..10ae2be 100644
--- a/drivers/infiniband/hw/hfi1/vnic.h
+++ b/drivers/infiniband/hw/hfi1/vnic.h
@@ -149,5 +149,8 @@ struct net_device *hfi1_vnic_alloc_rn(struct ib_device *device,
unsigned char name_assign_type,
void (*setup)(struct net_device *));
void hfi1_vnic_free_rn(struct net_device *netdev);
+int hfi1_vnic_send_dma(struct hfi1_devdata *dd, u8 q_idx,
+ struct hfi1_vnic_vport_info *vinfo,
+ struct sk_buff *skb, u64 pbc, u8 plen);
#endif /* _HFI1_VNIC_H */
diff --git a/drivers/infiniband/hw/hfi1/vnic_main.c b/drivers/infiniband/hw/hfi1/vnic_main.c
index 3cc430a..9952a04 100644
--- a/drivers/infiniband/hw/hfi1/vnic_main.c
+++ b/drivers/infiniband/hw/hfi1/vnic_main.c
@@ -62,6 +62,159 @@
static DEFINE_SPINLOCK(vport_cntr_lock);
+static int setup_vnic_ctxt(struct hfi1_devdata *dd, struct hfi1_ctxtdata *uctxt)
+{
+ unsigned int rcvctrl_ops = 0;
+ int ret;
+
+ ret = hfi1_init_ctxt(uctxt->sc);
+ if (ret)
+ goto done;
+
+ uctxt->do_interrupt = &handle_receive_interrupt;
+
+ /* Now allocate the RcvHdr queue and eager buffers. */
+ ret = hfi1_create_rcvhdrq(dd, uctxt);
+ if (ret)
+ goto done;
+
+ ret = hfi1_setup_eagerbufs(uctxt);
+ if (ret)
+ goto done;
+
+ set_bit(HFI1_CTXT_SETUP_DONE, &uctxt->event_flags);
+
+ if (uctxt->rcvhdrtail_kvaddr)
+ clear_rcvhdrtail(uctxt);
+
+ rcvctrl_ops = HFI1_RCVCTRL_CTXT_ENB;
+ rcvctrl_ops |= HFI1_RCVCTRL_INTRAVAIL_ENB;
+
+ if (!HFI1_CAP_KGET_MASK(uctxt->flags, MULTI_PKT_EGR))
+ rcvctrl_ops |= HFI1_RCVCTRL_ONE_PKT_EGR_ENB;
+ if (HFI1_CAP_KGET_MASK(uctxt->flags, NODROP_EGR_FULL))
+ rcvctrl_ops |= HFI1_RCVCTRL_NO_EGR_DROP_ENB;
+ if (HFI1_CAP_KGET_MASK(uctxt->flags, NODROP_RHQ_FULL))
+ rcvctrl_ops |= HFI1_RCVCTRL_NO_RHQ_DROP_ENB;
+ if (HFI1_CAP_KGET_MASK(uctxt->flags, DMA_RTAIL))
+ rcvctrl_ops |= HFI1_RCVCTRL_TAILUPD_ENB;
+
+ hfi1_rcvctrl(uctxt->dd, rcvctrl_ops, uctxt->ctxt);
+
+ uctxt->is_vnic = true;
+done:
+ return ret;
+}
+
+static int allocate_vnic_ctxt(struct hfi1_devdata *dd,
+ struct hfi1_ctxtdata **vnic_ctxt)
+{
+ struct hfi1_ctxtdata *uctxt;
+ unsigned int ctxt;
+ int ret;
+
+ if (dd->flags & HFI1_FROZEN)
+ return -EIO;
+
+ for (ctxt = dd->first_dyn_alloc_ctxt;
+ ctxt < dd->num_rcv_contexts; ctxt++)
+ if (!dd->rcd[ctxt])
+ break;
+
+ if (ctxt == dd->num_rcv_contexts)
+ return -EBUSY;
+
+ uctxt = hfi1_create_ctxtdata(dd->pport, ctxt, dd->node);
+ if (!uctxt) {
+ dd_dev_err(dd, "Unable to create ctxtdata, failing open\n");
+ return -ENOMEM;
+ }
+
+ uctxt->flags = HFI1_CAP_KGET(MULTI_PKT_EGR) |
+ HFI1_CAP_KGET(NODROP_RHQ_FULL) |
+ HFI1_CAP_KGET(NODROP_EGR_FULL) |
+ HFI1_CAP_KGET(DMA_RTAIL);
+ uctxt->seq_cnt = 1;
+
+ /* Allocate and enable a PIO send context */
+ uctxt->sc = sc_alloc(dd, SC_VNIC, uctxt->rcvhdrqentsize,
+ uctxt->numa_id);
+
+ ret = uctxt->sc ? 0 : -ENOMEM;
+ if (ret)
+ goto bail;
+
+ dd_dev_dbg(dd, "allocated vnic send context %u(%u)\n",
+ uctxt->sc->sw_index, uctxt->sc->hw_context);
+ ret = sc_enable(uctxt->sc);
+ if (ret)
+ goto bail;
+
+ if (dd->num_msix_entries)
+ hfi1_set_vnic_msix_info(uctxt);
+
+ hfi1_stats.sps_ctxts++;
+ dd_dev_dbg(dd, "created vnic context %d\n", uctxt->ctxt);
+ *vnic_ctxt = uctxt;
+
+ return ret;
+bail:
+ /*
+ * hfi1_free_ctxtdata() also releases send_context
+ * structure if uctxt->sc is not null
+ */
+ dd->rcd[uctxt->ctxt] = NULL;
+ hfi1_free_ctxtdata(dd, uctxt);
+ dd_dev_dbg(dd, "vnic allocation failed. rc %d\n", ret);
+ return ret;
+}
+
+static void deallocate_vnic_ctxt(struct hfi1_devdata *dd,
+ struct hfi1_ctxtdata *uctxt)
+{
+ unsigned long flags;
+
+ dd_dev_dbg(dd, "closing vnic context %d\n", uctxt->ctxt);
+ flush_wc();
+
+ if (dd->num_msix_entries)
+ hfi1_reset_vnic_msix_info(uctxt);
+
+ spin_lock_irqsave(&dd->uctxt_lock, flags);
+ /*
+ * Disable receive context and interrupt available, reset all
+ * RcvCtxtCtrl bits to default values.
+ */
+ hfi1_rcvctrl(dd, HFI1_RCVCTRL_CTXT_DIS |
+ HFI1_RCVCTRL_TIDFLOW_DIS |
+ HFI1_RCVCTRL_INTRAVAIL_DIS |
+ HFI1_RCVCTRL_ONE_PKT_EGR_DIS |
+ HFI1_RCVCTRL_NO_RHQ_DROP_DIS |
+ HFI1_RCVCTRL_NO_EGR_DROP_DIS, uctxt->ctxt);
+ /*
+ * VNIC contexts are allocated from user context pool.
+ * Release them back to user context pool.
+ *
+ * Reset context integrity checks to default.
+ * (writes to CSRs probably belong in chip.c)
+ */
+ write_kctxt_csr(dd, uctxt->sc->hw_context, SEND_CTXT_CHECK_ENABLE,
+ hfi1_pkt_default_send_ctxt_mask(dd, SC_USER));
+ sc_disable(uctxt->sc);
+
+ dd->send_contexts[uctxt->sc->sw_index].type = SC_USER;
+ spin_unlock_irqrestore(&dd->uctxt_lock, flags);
+
+ dd->rcd[uctxt->ctxt] = NULL;
+ uctxt->event_flags = 0;
+
+ hfi1_clear_tids(uctxt);
+ hfi1_clear_ctxt_pkey(dd, uctxt->ctxt);
+
+ hfi1_stats.sps_ctxts--;
+ hfi1_free_ctxtdata(dd, uctxt);
+}
+
void hfi1_vnic_setup(struct hfi1_devdata *dd)
{
idr_init(&dd->vnic.vesw_idr);
@@ -521,6 +674,9 @@ static void hfi1_vnic_down(struct hfi1_vnic_vport_info *vinfo)
netif_tx_disable(vinfo->netdev);
idr_remove(&dd->vnic.vesw_idr, vinfo->vesw_id);
+ /* ensure irqs see the change */
+ hfi1_vnic_synchronize_irq(dd);
+
/* remove unread skbs */
for (i = 0; i < vinfo->num_rx_q; i++) {
struct hfi1_vnic_rx_queue *rxq = &vinfo->rxq[i];
@@ -552,6 +708,84 @@ static int hfi1_netdev_close(struct net_device *netdev)
return 0;
}
+static int hfi1_vnic_allot_ctxt(struct hfi1_devdata *dd,
+ struct hfi1_ctxtdata **vnic_ctxt)
+{
+ int rc;
+
+ rc = allocate_vnic_ctxt(dd, vnic_ctxt);
+ if (rc) {
+ dd_dev_err(dd, "vnic ctxt alloc failed %d\n", rc);
+ return rc;
+ }
+
+ rc = setup_vnic_ctxt(dd, *vnic_ctxt);
+ if (rc) {
+ dd_dev_err(dd, "vnic ctxt setup failed %d\n", rc);
+ deallocate_vnic_ctxt(dd, *vnic_ctxt);
+ *vnic_ctxt = NULL;
+ }
+
+ return rc;
+}
+
+static int hfi1_vnic_init(struct hfi1_vnic_vport_info *vinfo)
+{
+ struct hfi1_devdata *dd = vinfo->dd;
+ int i, rc = 0;
+
+ mutex_lock(&hfi1_mutex);
+ if (!dd->vnic.num_vports)
+ dd->vnic.msix_idx = dd->first_dyn_msix_idx;
+
+ for (i = dd->vnic.num_ctxt; i < vinfo->num_rx_q; i++) {
+ rc = hfi1_vnic_allot_ctxt(dd, &dd->vnic.ctxt[i]);
+ if (rc)
+ break;
+ dd->vnic.ctxt[i]->vnic_q_idx = i;
+ }
+
+ if (i < vinfo->num_rx_q) {
+ /*
+ * If required amount of contexts is not
+ * allocated successfully then remaining contexts
+ * are released.
+ */
+ while (i-- > dd->vnic.num_ctxt) {
+ deallocate_vnic_ctxt(dd, dd->vnic.ctxt[i]);
+ dd->vnic.ctxt[i] = NULL;
+ }
+ goto alloc_fail;
+ }
+
+ if (dd->vnic.num_ctxt != i) {
+ dd->vnic.num_ctxt = i;
+ hfi1_init_vnic_rsm(dd);
+ }
+
+ dd->vnic.num_vports++;
+alloc_fail:
+ mutex_unlock(&hfi1_mutex);
+ return rc;
+}
+
+static void hfi1_vnic_deinit(struct hfi1_vnic_vport_info *vinfo)
+{
+ struct hfi1_devdata *dd = vinfo->dd;
+ int i;
+
+ mutex_lock(&hfi1_mutex);
+ if (--dd->vnic.num_vports == 0) {
+ for (i = 0; i < dd->vnic.num_ctxt; i++) {
+ deallocate_vnic_ctxt(dd, dd->vnic.ctxt[i]);
+ dd->vnic.ctxt[i] = NULL;
+ }
+ hfi1_deinit_vnic_rsm(dd);
+ dd->vnic.num_ctxt = 0;
+ }
+ mutex_unlock(&hfi1_mutex);
+}
+
static void hfi1_vnic_set_vesw_id(struct net_device *netdev, int id)
{
struct hfi1_vnic_vport_info *vinfo = hfi_vnic_dev_priv(netdev);
@@ -596,7 +830,7 @@ struct net_device *hfi1_vnic_alloc_rn(struct ib_device *device,
struct hfi1_vnic_vport_info *vinfo;
struct net_device *netdev;
struct rdma_netdev *rn;
- int i, size;
+ int i, size, rc;
if (!port_num || (port_num > dd->num_pports))
return ERR_PTR(-EINVAL);
@@ -634,13 +868,22 @@ struct net_device *hfi1_vnic_alloc_rn(struct ib_device *device,
netif_napi_add(netdev, &rxq->napi, hfi1_vnic_napi, 64);
}
+ rc = hfi1_vnic_init(vinfo);
+ if (rc)
+ goto init_fail;
+
return netdev;
+init_fail:
+ mutex_destroy(&vinfo->lock);
+ free_netdev(netdev);
+ return ERR_PTR(rc);
}
void hfi1_vnic_free_rn(struct net_device *netdev)
{
struct hfi1_vnic_vport_info *vinfo = hfi_vnic_dev_priv(netdev);
+ hfi1_vnic_deinit(vinfo);
mutex_destroy(&vinfo->lock);
free_netdev(netdev);
}
diff --git a/include/rdma/opa_port_info.h b/include/rdma/opa_port_info.h
index 9303e0e..541ef47 100644
--- a/include/rdma/opa_port_info.h
+++ b/include/rdma/opa_port_info.h
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2014 Intel Corporation. All rights reserved.
+ * Copyright (c) 2014-2017 Intel Corporation. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
@@ -410,7 +410,7 @@ struct opa_port_info {
u8 resptimevalue; /* 3 res, 5 bits */
u8 local_port_num;
- u8 reserved12;
+ u8 num_vesw_port_supported;
u8 reserved13; /* was guid_cap */
} __attribute__ ((packed));
--
1.8.3.1
^ permalink raw reply related [flat|nested] 40+ messages in thread* [RFC v3 11/11] IB/hfi1: VNIC SDMA support
2017-02-07 20:22 [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
` (6 preceding siblings ...)
2017-02-07 20:23 ` [RFC v3 10/11] IB/hfi1: Virtual Network Interface Controller (VNIC) HW support Vishwanathapura, Niranjana
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
[not found] ` <1486498990-76562-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
8 siblings, 0 replies; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura
HFI1 VNIC SDMA support enables transmission of VNIC packets over SDMA.
Map VNIC queues to SDMA engines and support halting and wakeup of the
VNIC queues.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
---
drivers/infiniband/hw/hfi1/Makefile | 2 +-
drivers/infiniband/hw/hfi1/hfi.h | 1 +
drivers/infiniband/hw/hfi1/init.c | 1 +
drivers/infiniband/hw/hfi1/vnic.h | 28 +++
drivers/infiniband/hw/hfi1/vnic_main.c | 24 ++-
drivers/infiniband/hw/hfi1/vnic_sdma.c | 323 +++++++++++++++++++++++++++++++++
6 files changed, 376 insertions(+), 3 deletions(-)
create mode 100644 drivers/infiniband/hw/hfi1/vnic_sdma.c
diff --git a/drivers/infiniband/hw/hfi1/Makefile b/drivers/infiniband/hw/hfi1/Makefile
index 2280538..88085f6 100644
--- a/drivers/infiniband/hw/hfi1/Makefile
+++ b/drivers/infiniband/hw/hfi1/Makefile
@@ -12,7 +12,7 @@ hfi1-y := affinity.o chip.o device.o driver.o efivar.o \
init.o intr.o mad.o mmu_rb.o pcie.o pio.o pio_copy.o platform.o \
qp.o qsfp.o rc.o ruc.o sdma.o sysfs.o trace.o \
uc.o ud.o user_exp_rcv.o user_pages.o user_sdma.o verbs.o \
- verbs_txreq.o vnic_main.o
+ verbs_txreq.o vnic_main.o vnic_sdma.o
hfi1-$(CONFIG_DEBUG_FS) += debugfs.o
CFLAGS_trace.o = -I$(src)
diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h
index 39bcd46..59109e4 100644
--- a/drivers/infiniband/hw/hfi1/hfi.h
+++ b/drivers/infiniband/hw/hfi1/hfi.h
@@ -835,6 +835,7 @@ struct hfi1_asic_data {
/* Virtual NIC information */
struct hfi1_vnic_data {
struct hfi1_ctxtdata *ctxt[HFI1_NUM_VNIC_CTXT];
+ struct kmem_cache *txreq_cache;
u8 num_vports;
struct idr vesw_idr;
u8 rmt_start;
diff --git a/drivers/infiniband/hw/hfi1/init.c b/drivers/infiniband/hw/hfi1/init.c
index e3a94d4..24e6e9b 100644
--- a/drivers/infiniband/hw/hfi1/init.c
+++ b/drivers/infiniband/hw/hfi1/init.c
@@ -680,6 +680,7 @@ int hfi1_init(struct hfi1_devdata *dd, int reinit)
dd->process_pio_send = hfi1_verbs_send_pio;
dd->process_dma_send = hfi1_verbs_send_dma;
dd->pio_inline_send = pio_copy;
+ dd->process_vnic_dma_send = hfi1_vnic_send_dma;
if (is_ax(dd)) {
atomic_set(&dd->drop_packet, DROP_PACKET_ON);
diff --git a/drivers/infiniband/hw/hfi1/vnic.h b/drivers/infiniband/hw/hfi1/vnic.h
index 10ae2be..72dce3b 100644
--- a/drivers/infiniband/hw/hfi1/vnic.h
+++ b/drivers/infiniband/hw/hfi1/vnic.h
@@ -49,6 +49,7 @@
#include <rdma/opa_hfi.h>
#include "hfi.h"
+#include "sdma.h"
#define HFI1_VNIC_MAX_TXQ 16
#define HFI1_VNIC_MAX_PAD 12
@@ -85,6 +86,26 @@
#define HFI_VNIC_MAX_QUEUE 16
/**
+ * struct hfi1_vnic_sdma - VNIC per Tx ring SDMA information
+ * @dd - device data pointer
+ * @sde - sdma engine
+ * @vinfo - vnic info pointer
+ * @wait - iowait structure
+ * @stx - sdma tx request
+ * @state - vnic Tx ring SDMA state
+ * @q_idx - vnic Tx queue index
+ */
+struct hfi1_vnic_sdma {
+ struct hfi1_devdata *dd;
+ struct sdma_engine *sde;
+ struct hfi1_vnic_vport_info *vinfo;
+ struct iowait wait;
+ struct sdma_txreq stx;
+ unsigned int state;
+ u8 q_idx;
+};
+
+/**
* struct hfi1_vnic_rx_queue - HFI1 VNIC receive queue
* @idx: queue index
* @vinfo: pointer to vport information
@@ -111,6 +132,7 @@ struct hfi1_vnic_rx_queue {
* @vesw_id: virtual switch id
* @rxq: Array of receive queues
* @stats: per queue stats
+ * @sdma: VNIC SDMA structure per TXQ
*/
struct hfi1_vnic_vport_info {
struct hfi1_devdata *dd;
@@ -126,6 +148,7 @@ struct hfi1_vnic_vport_info {
struct hfi1_vnic_rx_queue rxq[HFI1_NUM_VNIC_CTXT];
struct hfi_vnic_stats stats[HFI_VNIC_MAX_QUEUE];
+ struct hfi1_vnic_sdma sdma[HFI1_VNIC_MAX_TXQ];
};
#define v_dbg(format, arg...) \
@@ -138,8 +161,13 @@ struct hfi1_vnic_vport_info {
/* vnic hfi1 internal functions */
void hfi1_vnic_setup(struct hfi1_devdata *dd);
void hfi1_vnic_cleanup(struct hfi1_devdata *dd);
+int hfi1_vnic_txreq_init(struct hfi1_devdata *dd);
+void hfi1_vnic_txreq_deinit(struct hfi1_devdata *dd);
void hfi1_vnic_bypass_rcv(struct hfi1_packet *packet);
+void hfi1_vnic_sdma_init(struct hfi1_vnic_vport_info *vinfo);
+bool hfi1_vnic_sdma_write_avail(struct hfi1_vnic_vport_info *vinfo,
+ u8 q_idx);
/* vnic rdma netdev operations */
struct net_device *hfi1_vnic_alloc_rn(struct ib_device *device,
diff --git a/drivers/infiniband/hw/hfi1/vnic_main.c b/drivers/infiniband/hw/hfi1/vnic_main.c
index 9952a04..e859846 100644
--- a/drivers/infiniband/hw/hfi1/vnic_main.c
+++ b/drivers/infiniband/hw/hfi1/vnic_main.c
@@ -408,6 +408,10 @@ static void hfi1_vnic_maybe_stop_tx(struct hfi1_vnic_vport_info *vinfo,
u8 q_idx)
{
netif_stop_subqueue(vinfo->netdev, q_idx);
+ if (!hfi1_vnic_sdma_write_avail(vinfo, q_idx))
+ return;
+
+ netif_start_subqueue(vinfo->netdev, q_idx);
}
static netdev_tx_t hfi1_netdev_start_xmit(struct sk_buff *skb,
@@ -479,7 +483,13 @@ static u16 hfi1_vnic_select_queue(struct net_device *netdev,
void *accel_priv,
select_queue_fallback_t fallback)
{
- return 0;
+ struct hfi1_vnic_vport_info *vinfo = hfi_vnic_dev_priv(netdev);
+ struct hfi_vnic_skb_mdata *mdata;
+ struct sdma_engine *sde;
+
+ mdata = (struct hfi_vnic_skb_mdata *)skb->data;
+ sde = sdma_select_engine_vl(vinfo->dd, mdata->entropy, mdata->vl);
+ return sde->this_idx;
}
/* hfi1_vnic_decap_skb - strip OPA header from the skb (ethernet) packet */
@@ -735,8 +745,13 @@ static int hfi1_vnic_init(struct hfi1_vnic_vport_info *vinfo)
int i, rc = 0;
mutex_lock(&hfi1_mutex);
- if (!dd->vnic.num_vports)
+ if (!dd->vnic.num_vports) {
+ rc = hfi1_vnic_txreq_init(dd);
+ if (rc)
+ goto txreq_fail;
+
dd->vnic.msix_idx = dd->first_dyn_msix_idx;
+ }
for (i = dd->vnic.num_ctxt; i < vinfo->num_rx_q; i++) {
rc = hfi1_vnic_allot_ctxt(dd, &dd->vnic.ctxt[i]);
@@ -764,7 +779,11 @@ static int hfi1_vnic_init(struct hfi1_vnic_vport_info *vinfo)
}
dd->vnic.num_vports++;
+ hfi1_vnic_sdma_init(vinfo);
alloc_fail:
+ if (!dd->vnic.num_vports)
+ hfi1_vnic_txreq_deinit(dd);
+txreq_fail:
mutex_unlock(&hfi1_mutex);
return rc;
}
@@ -782,6 +801,7 @@ static void hfi1_vnic_deinit(struct hfi1_vnic_vport_info *vinfo)
}
hfi1_deinit_vnic_rsm(dd);
dd->vnic.num_ctxt = 0;
+ hfi1_vnic_txreq_deinit(dd);
}
mutex_unlock(&hfi1_mutex);
}
diff --git a/drivers/infiniband/hw/hfi1/vnic_sdma.c b/drivers/infiniband/hw/hfi1/vnic_sdma.c
new file mode 100644
index 0000000..482b42c
--- /dev/null
+++ b/drivers/infiniband/hw/hfi1/vnic_sdma.c
@@ -0,0 +1,323 @@
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * This file contains HFI1 support for VNIC SDMA functionality
+ */
+
+#include "sdma.h"
+#include "vnic.h"
+
+#define HFI1_VNIC_SDMA_Q_ACTIVE BIT(0)
+#define HFI1_VNIC_SDMA_Q_DEFERRED BIT(1)
+
+#define HFI1_VNIC_TXREQ_NAME_LEN 32
+#define HFI1_VNIC_SDMA_DESC_WTRMRK 64
+#define HFI1_VNIC_SDMA_RETRY_COUNT 1
+
+/*
+ * struct vnic_txreq - VNIC transmit descriptor
+ * @txreq: sdma transmit request
+ * @sdma: vnic sdma pointer
+ * @skb: skb to send
+ * @pad: pad buffer
+ * @plen: pad length
+ * @pbc_val: pbc value
+ * @retry_count: tx retry count
+ */
+struct vnic_txreq {
+ struct sdma_txreq txreq;
+ struct hfi1_vnic_sdma *sdma;
+
+ struct sk_buff *skb;
+ unsigned char pad[HFI1_VNIC_MAX_PAD];
+ u16 plen;
+ __le64 pbc_val;
+
+ u32 retry_count;
+};
+
+static void vnic_sdma_complete(struct sdma_txreq *txreq,
+ int status)
+{
+ struct vnic_txreq *tx = container_of(txreq, struct vnic_txreq, txreq);
+ struct hfi1_vnic_sdma *vnic_sdma = tx->sdma;
+
+ sdma_txclean(vnic_sdma->dd, txreq);
+ dev_kfree_skb_any(tx->skb);
+ kmem_cache_free(vnic_sdma->dd->vnic.txreq_cache, tx);
+}
+
+static noinline int build_vnic_ulp_payload(struct sdma_engine *sde,
+ struct vnic_txreq *tx)
+{
+ int i, ret = 0;
+
+ ret = sdma_txadd_kvaddr(
+ sde->dd,
+ &tx->txreq,
+ tx->skb->data,
+ skb_headlen(tx->skb));
+ if (ret)
+ goto bail_txadd;
+
+ for (i = 0; i < skb_shinfo(tx->skb)->nr_frags; i++) {
+ struct skb_frag_struct *frag = &skb_shinfo(tx->skb)->frags[i];
+
+ /* combine physically continuous fragments later? */
+ ret = sdma_txadd_page(sde->dd,
+ &tx->txreq,
+ skb_frag_page(frag),
+ frag->page_offset,
+ skb_frag_size(frag));
+ if (ret)
+ goto bail_txadd;
+ }
+
+ if (tx->plen)
+ ret = sdma_txadd_kvaddr(sde->dd, &tx->txreq,
+ tx->pad + HFI1_VNIC_MAX_PAD - tx->plen,
+ tx->plen);
+
+bail_txadd:
+ return ret;
+}
+
+static int build_vnic_tx_desc(struct sdma_engine *sde,
+ struct vnic_txreq *tx,
+ u64 pbc)
+{
+ int ret = 0;
+ u16 hdrbytes = 2 << 2; /* PBC */
+
+ ret = sdma_txinit_ahg(
+ &tx->txreq,
+ 0,
+ hdrbytes + tx->skb->len + tx->plen,
+ 0,
+ 0,
+ NULL,
+ 0,
+ vnic_sdma_complete);
+ if (ret)
+ goto bail_txadd;
+
+ /* add pbc */
+ tx->pbc_val = cpu_to_le64(pbc);
+ ret = sdma_txadd_kvaddr(
+ sde->dd,
+ &tx->txreq,
+ &tx->pbc_val,
+ hdrbytes);
+ if (ret)
+ goto bail_txadd;
+
+ /* add the ulp payload */
+ ret = build_vnic_ulp_payload(sde, tx);
+bail_txadd:
+ return ret;
+}
+
+/* setup the last plen bypes of pad */
+static inline void hfi1_vnic_update_pad(unsigned char *pad, u8 plen)
+{
+ pad[HFI1_VNIC_MAX_PAD - 1] = plen - HFI_VNIC_ICRC_TAIL_LEN;
+}
+
+int hfi1_vnic_send_dma(struct hfi1_devdata *dd, u8 q_idx,
+ struct hfi1_vnic_vport_info *vinfo,
+ struct sk_buff *skb, u64 pbc, u8 plen)
+{
+ struct hfi1_vnic_sdma *vnic_sdma = &vinfo->sdma[q_idx];
+ struct sdma_engine *sde = vnic_sdma->sde;
+ struct vnic_txreq *tx;
+ int ret = -ECOMM;
+
+ if (READ_ONCE(vnic_sdma->state) != HFI1_VNIC_SDMA_Q_ACTIVE)
+ goto tx_err;
+
+ if (!sde || !sdma_running(sde))
+ goto tx_err;
+
+ tx = kmem_cache_alloc(dd->vnic.txreq_cache, GFP_ATOMIC);
+ if (!tx) {
+ ret = -ENOMEM;
+ goto tx_err;
+ }
+
+ tx->sdma = vnic_sdma;
+ tx->skb = skb;
+ hfi1_vnic_update_pad(tx->pad, plen);
+ tx->plen = plen;
+ ret = build_vnic_tx_desc(sde, tx, pbc);
+ if (unlikely(ret))
+ goto free_desc;
+ tx->retry_count = 0;
+
+ ret = sdma_send_txreq(sde, &vnic_sdma->wait, &tx->txreq);
+ /* When -ECOMM, sdma callback will be called with ABORT status */
+ if (ret && unlikely(ret != -ECOMM))
+ goto free_desc;
+
+ return ret;
+
+free_desc:
+ sdma_txclean(dd, &tx->txreq);
+ kmem_cache_free(dd->vnic.txreq_cache, tx);
+tx_err:
+ if (ret != -EBUSY)
+ dev_kfree_skb_any(skb);
+ return ret;
+}
+
+/*
+ * hfi1_vnic_sdma_sleep - vnic sdma sleep function
+ *
+ * This function gets called from sdma_send_txreq() when there are not enough
+ * sdma descriptors available to send the packet. It adds Tx queue's wait
+ * structure to sdma engine's dmawait list to be woken up when descriptors
+ * become available.
+ */
+static int hfi1_vnic_sdma_sleep(struct sdma_engine *sde,
+ struct iowait *wait,
+ struct sdma_txreq *txreq,
+ unsigned int seq)
+{
+ struct hfi1_vnic_sdma *vnic_sdma =
+ container_of(wait, struct hfi1_vnic_sdma, wait);
+ struct hfi1_ibdev *dev = &vnic_sdma->dd->verbs_dev;
+ struct vnic_txreq *tx = container_of(txreq, struct vnic_txreq, txreq);
+
+ if (sdma_progress(sde, seq, txreq))
+ if (tx->retry_count++ < HFI1_VNIC_SDMA_RETRY_COUNT)
+ return -EAGAIN;
+
+ vnic_sdma->state = HFI1_VNIC_SDMA_Q_DEFERRED;
+ write_seqlock(&dev->iowait_lock);
+ if (list_empty(&vnic_sdma->wait.list))
+ list_add_tail(&vnic_sdma->wait.list, &sde->dmawait);
+ write_sequnlock(&dev->iowait_lock);
+ return -EBUSY;
+}
+
+/*
+ * hfi1_vnic_sdma_wakeup - vnic sdma wakeup function
+ *
+ * This function gets called when SDMA descriptors becomes available and Tx
+ * queue's wait structure was previously added to sdma engine's dmawait list.
+ * It notifies the upper driver about Tx queue wakeup.
+ */
+static void hfi1_vnic_sdma_wakeup(struct iowait *wait, int reason)
+{
+ struct hfi1_vnic_sdma *vnic_sdma =
+ container_of(wait, struct hfi1_vnic_sdma, wait);
+ struct hfi1_vnic_vport_info *vinfo = vnic_sdma->vinfo;
+
+ vnic_sdma->state = HFI1_VNIC_SDMA_Q_ACTIVE;
+ if (__netif_subqueue_stopped(vinfo->netdev, vnic_sdma->q_idx))
+ netif_wake_subqueue(vinfo->netdev, vnic_sdma->q_idx);
+};
+
+inline bool hfi1_vnic_sdma_write_avail(struct hfi1_vnic_vport_info *vinfo,
+ u8 q_idx)
+{
+ struct hfi1_vnic_sdma *vnic_sdma = &vinfo->sdma[q_idx];
+
+ return (READ_ONCE(vnic_sdma->state) == HFI1_VNIC_SDMA_Q_ACTIVE);
+}
+
+void hfi1_vnic_sdma_init(struct hfi1_vnic_vport_info *vinfo)
+{
+ int i;
+
+ for (i = 0; i < vinfo->num_tx_q; i++) {
+ struct hfi1_vnic_sdma *vnic_sdma = &vinfo->sdma[i];
+
+ iowait_init(&vnic_sdma->wait, 0, NULL, hfi1_vnic_sdma_sleep,
+ hfi1_vnic_sdma_wakeup, NULL);
+ vnic_sdma->sde = &vinfo->dd->per_sdma[i];
+ vnic_sdma->dd = vinfo->dd;
+ vnic_sdma->vinfo = vinfo;
+ vnic_sdma->q_idx = i;
+ vnic_sdma->state = HFI1_VNIC_SDMA_Q_ACTIVE;
+
+ /* Add a free descriptor watermark for wakeups */
+ if (vnic_sdma->sde->descq_cnt >= HFI1_VNIC_SDMA_DESC_WTRMRK) {
+ INIT_LIST_HEAD(&vnic_sdma->stx.list);
+ vnic_sdma->stx.num_desc = HFI1_VNIC_SDMA_DESC_WTRMRK;
+ list_add_tail(&vnic_sdma->stx.list,
+ &vnic_sdma->wait.tx_head);
+ }
+ }
+}
+
+static void hfi1_vnic_txreq_kmem_cache_ctor(void *obj)
+{
+ struct vnic_txreq *tx = (struct vnic_txreq *)obj;
+
+ memset(tx, 0, sizeof(*tx));
+}
+
+int hfi1_vnic_txreq_init(struct hfi1_devdata *dd)
+{
+ char buf[HFI1_VNIC_TXREQ_NAME_LEN];
+
+ snprintf(buf, sizeof(buf), "hfi1_%u_vnic_txreq_cache", dd->unit);
+ dd->vnic.txreq_cache = kmem_cache_create(buf,
+ sizeof(struct vnic_txreq),
+ 0, SLAB_HWCACHE_ALIGN,
+ hfi1_vnic_txreq_kmem_cache_ctor);
+ if (!dd->vnic.txreq_cache)
+ return -ENOMEM;
+ return 0;
+}
+
+void hfi1_vnic_txreq_deinit(struct hfi1_devdata *dd)
+{
+ kmem_cache_destroy(dd->vnic.txreq_cache);
+ dd->vnic.txreq_cache = NULL;
+}
--
1.8.3.1
^ permalink raw reply related [flat|nested] 40+ messages in thread[parent not found: <1486498990-76562-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>]
* [RFC v3 01/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) documentation
[not found] ` <1486498990-76562-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
[not found] ` <1486498990-76562-2-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2017-02-07 20:23 ` [RFC v3 02/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface Vishwanathapura, Niranjana
` (4 subsequent siblings)
5 siblings, 1 reply; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Niranjana Vishwanathapura
Add HFI VNIC design document explaining the VNIC architecture and the
driver design.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
Documentation/infiniband/hfi_vnic.txt | 102 ++++++++++++++++++++++++++++++++++
1 file changed, 102 insertions(+)
create mode 100644 Documentation/infiniband/hfi_vnic.txt
diff --git a/Documentation/infiniband/hfi_vnic.txt b/Documentation/infiniband/hfi_vnic.txt
new file mode 100644
index 0000000..c6c801e
--- /dev/null
+++ b/Documentation/infiniband/hfi_vnic.txt
@@ -0,0 +1,102 @@
+Intel Omni-Path Host Fabric Interface (HFI) Virtual Network Interface
+Controller (VNIC) feature supports Ethernet functionality over Omni-Path
+fabric by encapsulating the Ethernet packets between HFI nodes.
+
+The patterns of exchanges of Omni-Path encapsulated Ethernet packets
+involves one or more virtual Ethernet switches overlaid on the Omni-Path
+fabric topology. A subset of HFI nodes on the Omni-Path fabric are
+permitted to exchange encapsulated Ethernet packets across a particular
+virtual Ethernet switch. The virtual Ethernet switches are logical
+abstractions achieved by configuring the HFI nodes on the fabric for
+header generation and processing. In the simplest configuration all HFI
+nodes across the fabric exchange encapsulated Ethernet packets over a
+single virtual Ethernet switch. A virtual Ethernet switch, is effectively
+an independent Ethernet network. The configuration is performed by an
+Ethernet Manager (EM) which is part of the trusted Fabric Manager (FM)
+application. HFI nodes can have multiple VNICs each connected to a
+different virtual Ethernet switch. The below diagram presents a case
+of two virtual Ethernet switches with two HFI nodes.
+
+ +-------------------+
+ | Subnet/ |
+ | Ethernet |
+ | Manager |
+ +-------------------+
+ / /
+ / /
+ / /
+ / /
++-----------------------------+ +------------------------------+
+| Virtual Ethernet Switch | | Virtual Ethernet Switch |
+| +---------+ +---------+ | | +---------+ +---------+ |
+| | VPORT | | VPORT | | | | VPORT | | VPORT | |
++--+---------+----+---------+-+ +-+---------+----+---------+---+
+ | \ / |
+ | \ / |
+ | \/ |
+ | / \ |
+ | / \ |
+ +-----------+------------+ +-----------+------------+
+ | VNIC | VNIC | | VNIC | VNIC |
+ +-----------+------------+ +-----------+------------+
+ | HFI | | HFI |
+ +------------------------+ +------------------------+
+
+
+Intel HFI VNIC software design is presented in the below diagram.
+HFI VNIC functionality has a HW dependent component and a HW
+independent component.
+
+The support has been added for IB device to allocate and free the RDMA
+netdev devices. The RDMA netdev supports interfacing with the network
+stack thus creating standard network interfaces. HFI_VNIC is an RDMA
+netdev device type.
+
+The HW dependent VNIC functionality is part of the HFI1 driver. It
+implements the verbs to allocate and free the HFI_VNIC RDMA netdev.
+It involves HW resource allocation/management for VNIC functionality.
+It interfaces with the network stack and implements the required
+net_device_ops functions. It expects Omni-Path encapsulated Ethernet
+packets in the transmit path and provides HW access to them. It strips
+the Omni-Path header from the received packets before passing them up
+the network stack. It also implements the RDMA netdev control operations.
+
+The HFI VNIC module implements the HW independent VNIC functionality.
+It consists of two parts. The VNIC Ethernet Management Agent (VEMA)
+registers itself with IB core as an IB client and interfaces with the
+IB MAD stack. It exchanges the management information with the Ethernet
+Manager (EM) and the VNIC netdev. The VNIC netdev part allocates and frees
+the HFI_VNIC RDMA netdev devices. It overrides the net_device_ops functions
+set by HW dependent VNIC driver where required to accommodate any control
+operation. It also handles the encapsulation of Ethernet packets with an
+Omni-Path header in the transmit path. For each VNIC interface, the
+information required for encapsulation is configured by the EM via VEMA MAD
+interface. It also passes any control information to the HW dependent driver
+by invoking the RDMA netdev control operations.
+
+ +-------------------+ +----------------------+
+ | | | Linux |
+ | IB MAD | | Network |
+ | | | Stack |
+ +-------------------+ +----------------------+
+ | | |
+ | | |
+ +----------------------------+ |
+ | | |
+ | HFI VNIC Module | |
+ | (HFI VNIC RDMA Netdev | |
+ | & EMA functions) | |
+ | | |
+ +----------------------------+ |
+ | |
+ | |
+ +------------------+ |
+ | IB core | |
+ +------------------+ |
+ | |
+ | |
+ +--------------------------------------------+
+ | |
+ | HFI1 Driver with VNIC support |
+ | |
+ +--------------------------------------------+
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 40+ messages in thread* [RFC v3 02/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface
[not found] ` <1486498990-76562-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2017-02-07 20:23 ` [RFC v3 01/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) documentation Vishwanathapura, Niranjana
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
2017-02-07 21:19 ` Jason Gunthorpe
2017-02-08 0:43 ` Parav Pandit
2017-02-07 20:23 ` [RFC v3 04/11] IB/hfi-vnic: VNIC Ethernet Management (EM) structure definitions Vishwanathapura, Niranjana
` (3 subsequent siblings)
5 siblings, 2 replies; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Niranjana Vishwanathapura
Add rdma netdev interface to ib device structure allowing rdma netdev
devices to be allocated by ib clients.
Define HFI VNIC interface between hardware independent VNIC
functionality and the hardware dependent VNIC functionality.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
include/rdma/ib_verbs.h | 27 +++++++++
include/rdma/opa_hfi.h | 147 ++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 174 insertions(+)
create mode 100644 include/rdma/opa_hfi.h
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index b1ac973..b9897d9d 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -55,6 +55,7 @@
#include <net/ip.h>
#include <linux/string.h>
#include <linux/slab.h>
+#include <linux/netdevice.h>
#include <linux/if_link.h>
#include <linux/atomic.h>
@@ -221,6 +222,7 @@ enum ib_device_cap_flags {
IB_DEVICE_SG_GAPS_REG = (1ULL << 32),
IB_DEVICE_VIRTUAL_FUNCTION = (1ULL << 33),
IB_DEVICE_RAW_SCATTER_FCS = (1ULL << 34),
+ IB_DEVICE_RDMA_NETDEV_HFI_VNIC = (1ULL << 35),
};
enum ib_signature_prot_cap {
@@ -1844,6 +1846,22 @@ struct ib_port_immutable {
u32 max_mad_size;
};
+/* rdma netdev type - specifies protocol type */
+enum rdma_netdev_t {
+ RDMA_NETDEV_HFI_VNIC
+};
+
+/**
+ * struct rdma_netdev - rdma netdev
+ * For cases where netstack interfacing is required.
+ */
+struct rdma_netdev {
+ void *clnt_priv;
+
+ /* control functions */
+ void (*set_id)(struct net_device *netdev, int id);
+};
+
struct ib_device {
struct device *dma_device;
@@ -2096,6 +2114,15 @@ struct ib_device {
struct ib_rwq_ind_table_init_attr *init_attr,
struct ib_udata *udata);
int (*destroy_rwq_ind_table)(struct ib_rwq_ind_table *wq_ind_table);
+ /* rdma netdev operations */
+ struct net_device *(*alloc_rdma_netdev)(
+ struct ib_device *device,
+ u8 port_num,
+ enum rdma_netdev_t type,
+ const char *name,
+ unsigned char name_assign_type,
+ void (*setup)(struct net_device *));
+ void (*free_rdma_netdev)(struct net_device *netdev);
struct ib_dma_mapping_ops *dma_ops;
struct module *owner;
diff --git a/include/rdma/opa_hfi.h b/include/rdma/opa_hfi.h
new file mode 100644
index 0000000..f357d04
--- /dev/null
+++ b/include/rdma/opa_hfi.h
@@ -0,0 +1,147 @@
+#ifndef _OPA_HFI_H
+#define _OPA_HFI_H
+/*
+ * Copyright(c) 2017 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * - 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.
+ * - Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * This file contains Intel Omni-Path (OPA) Host Fabric Interface (HFI)
+ * specific declarations.
+ */
+
+#include <rdma/ib_verbs.h>
+
+#define HFI_MAX_NUM_VNICS 255
+
+/* VNIC uses 16B header format */
+#define HFI_VNIC_L2_TYPE 0x2
+
+/* 16 header bytes + 2 reserved bytes */
+#define HFI_VNIC_L2_HDR_LEN (16 + 2)
+
+#define HFI_VNIC_L4_HDR_LEN 2
+
+#define HFI_VNIC_HDR_LEN (HFI_VNIC_L2_HDR_LEN + \
+ HFI_VNIC_L4_HDR_LEN)
+
+#define HFI_VNIC_L4_ETHR 0x78
+
+#define HFI_VNIC_ICRC_LEN 4
+#define HFI_VNIC_TAIL_LEN 1
+#define HFI_VNIC_ICRC_TAIL_LEN (HFI_VNIC_ICRC_LEN + HFI_VNIC_TAIL_LEN)
+
+/* VNIC configured and operational state values */
+#define HFI_VNIC_STATE_DROP_ALL 0x1
+#define HFI_VNIC_STATE_FORWARDING 0x3
+
+#define HFI_VNIC_SKB_MDATA_LEN 4
+#define HFI_VNIC_SKB_MDATA_ENCAP_ERR 0x1
+
+/* hfi vnic rdma netdev's private data structure */
+struct hfi_vnic_rdma_netdev {
+ struct rdma_netdev rn; /* keep this first */
+ /* followed by device private data */
+ char *dev_priv[0];
+};
+
+static inline void *hfi_vnic_priv(const struct net_device *dev)
+{
+ struct rdma_netdev *rn = netdev_priv(dev);
+
+ return rn->clnt_priv;
+}
+
+static inline void *hfi_vnic_dev_priv(const struct net_device *dev)
+{
+ struct rdma_netdev *rn = netdev_priv(dev);
+
+ return rn + 1;
+}
+
+/* hfi_vnic skb meta data structrue */
+struct hfi_vnic_skb_mdata {
+ u8 vl;
+ u8 entropy;
+ u8 flags;
+ u8 rsvd;
+} __packed;
+
+/* HFI VNIC group statistics */
+struct hfi_vnic_grp_stats {
+ u64 unicast;
+ u64 mcastbcast;
+ u64 untagged;
+ u64 vlan;
+ u64 s_64;
+ u64 s_65_127;
+ u64 s_128_255;
+ u64 s_256_511;
+ u64 s_512_1023;
+ u64 s_1024_1518;
+ u64 s_1519_max;
+};
+
+struct hfi_vnic_stats {
+ /* standard netdev statistics */
+ struct rtnl_link_stats64 netstats;
+
+ /* HFI VNIC statistics */
+ struct hfi_vnic_grp_stats tx_grp;
+ struct hfi_vnic_grp_stats rx_grp;
+ u64 tx_dlid_zero;
+ u64 tx_drop_state;
+ u64 rx_drop_state;
+ u64 rx_runt;
+ u64 rx_oversize;
+};
+
+static inline bool rdma_cap_hfi_vnic(struct ib_device *device)
+{
+ return !!(device->attrs.device_cap_flags &
+ IB_DEVICE_RDMA_NETDEV_HFI_VNIC);
+}
+
+#endif /* _OPA_HFI_H */
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 40+ messages in thread* Re: [RFC v3 02/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface
2017-02-07 20:23 ` [RFC v3 02/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface Vishwanathapura, Niranjana
@ 2017-02-07 21:19 ` Jason Gunthorpe
[not found] ` <20170207211901.GB28922-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2017-02-08 0:43 ` Parav Pandit
1 sibling, 1 reply; 40+ messages in thread
From: Jason Gunthorpe @ 2017-02-07 21:19 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: dledford, linux-rdma, netdev, dennis.dalessandro, ira.weiny, Liran Liss
On Tue, Feb 07, 2017 at 12:23:01PM -0800, Vishwanathapura, Niranjana wrote:
> Add rdma netdev interface to ib device structure allowing rdma netdev
> devices to be allocated by ib clients.
> Define HFI VNIC interface between hardware independent VNIC
> functionality and the hardware dependent VNIC functionality.
This commit message could be a bit clearer.
The alloc_rdma_netdev multiplexer is inteded as a new general
interface and this adds a protocol definition for ethernet VNIC on
OPA.
The hope is that ipoib can follow the same example and use the same
alloc_rdma_netdev entry point. Hopefully Mellanox will look at this
patch as I have talked to them in the past about doing this...
It looks like HFI turned out fairly well, the driver code and higher
level code have a reasonably nice split in my quick look.
> IB_DEVICE_RAW_SCATTER_FCS = (1ULL << 34),
> + IB_DEVICE_RDMA_NETDEV_HFI_VNIC = (1ULL << 35),
What is this called HFI_VNIC anyhow? Shouldn't this be OPA_VNIC? There
is nothing really HFI specific, right?
> +/* hfi vnic rdma netdev's private data structure */
> +struct hfi_vnic_rdma_netdev {
> + struct rdma_netdev rn; /* keep this first */
> + /* followed by device private data */
> + char *dev_priv[0];
> +};
> +
> +static inline void *hfi_vnic_priv(const struct net_device *dev)
> +{
> + struct rdma_netdev *rn = netdev_priv(dev);
> +
> + return rn->clnt_priv;
> +}
> +
> +static inline void *hfi_vnic_dev_priv(const struct net_device *dev)
> +{
> + struct rdma_netdev *rn = netdev_priv(dev);
Shouldn't this be hfi_vnic_rdma_netdev ?
> + return rn + 1;
And this should be rn->dev_priv ?
Jason
^ permalink raw reply [flat|nested] 40+ messages in thread* RE: [RFC v3 02/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface
2017-02-07 20:23 ` [RFC v3 02/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface Vishwanathapura, Niranjana
2017-02-07 21:19 ` Jason Gunthorpe
@ 2017-02-08 0:43 ` Parav Pandit
[not found] ` <VI1PR0502MB3008911E9587F4C004BFEF13D1420-o1MPJYiShExKsLr+rGaxW8DSnupUy6xnnBOFsp37pqbUKgpGm//BTAC/G2K4zDHf@public.gmane.org>
1 sibling, 1 reply; 40+ messages in thread
From: Parav Pandit @ 2017-02-08 0:43 UTC (permalink / raw)
To: Vishwanathapura, Niranjana, dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny
Hi
> -----Original Message-----
> From: linux-rdma-owner@vger.kernel.org [mailto:linux-rdma-
> owner@vger.kernel.org] On Behalf Of Vishwanathapura, Niranjana
> Sent: Tuesday, February 7, 2017 2:23 PM
> To: dledford@redhat.com
> Cc: linux-rdma@vger.kernel.org; netdev@vger.kernel.org;
> dennis.dalessandro@intel.com; ira.weiny@intel.com; Niranjana
> Vishwanathapura <niranjana.vishwanathapura@intel.com>
> Subject: [RFC v3 02/11] IB/hfi-vnic: Virtual Network Interface Controller
> (VNIC) interface
>
> Add rdma netdev interface to ib device structure allowing rdma netdev
> devices to be allocated by ib clients.
> Define HFI VNIC interface between hardware independent VNIC
> functionality and the hardware dependent VNIC functionality.
>
> Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
> Reviewed-by: Ira Weiny <ira.weiny@intel.com>
> Signed-off-by: Niranjana Vishwanathapura
> <niranjana.vishwanathapura@intel.com>
> ---
> struct ib_device {
> struct device *dma_device;
>
> @@ -2096,6 +2114,15 @@ struct ib_device {
> struct
> ib_rwq_ind_table_init_attr *init_attr,
> struct ib_udata
> *udata);
> int (*destroy_rwq_ind_table)(struct ib_rwq_ind_table
> *wq_ind_table);
> + /* rdma netdev operations */
> + struct net_device *(*alloc_rdma_netdev)(
> + struct ib_device *device,
> + u8 port_num,
> + enum rdma_netdev_t type,
> + const char *name,
> + unsigned char name_assign_type,
> + void (*setup)(struct net_device *));
> + void (*free_rdma_netdev)(struct net_device *netdev);
> struct ib_dma_mapping_ops *dma_ops;
>
> struct module *owner;
As its clear from the cover letter and from the request to place this in drivers/infiniband/ulp,
Instead of increasing the ib_dev structure further,
Can you change the code to make use of ib_register_client() and friend functions to register vnic as ULP.
(similar to other ULP such as uverbs, srp, ipoib).
This will also allow you get to get notified for removing the vnic device when underlying rdma device gets removed.
Based on the property that gets exposed by the ibdev, vnic driver filters whether it needs to load its vnic to specific device or not.
This way modules are isolated between core and ULP little better.
Would it work for you?
^ permalink raw reply [flat|nested] 40+ messages in thread
* [RFC v3 04/11] IB/hfi-vnic: VNIC Ethernet Management (EM) structure definitions
[not found] ` <1486498990-76562-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2017-02-07 20:23 ` [RFC v3 01/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) documentation Vishwanathapura, Niranjana
2017-02-07 20:23 ` [RFC v3 02/11] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface Vishwanathapura, Niranjana
@ 2017-02-07 20:23 ` Vishwanathapura, Niranjana
2017-02-07 20:36 ` [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC) Leon Romanovsky
` (2 subsequent siblings)
5 siblings, 0 replies; 40+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-02-07 20:23 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Niranjana Vishwanathapura,
Sadanand Warrier, Tanya K Jajodia
Define VNIC EM MAD structures and the associated macros. These structures
are used for information exchange between VNIC EM agent (EMA) on the HFI
host and the Ethernet manager. These include the virtual ethernet switch
(vesw) port information, vesw port mac table, summay and error counters,
vesw port interface mac lists and the EMA trap.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Sadanand Warrier <sadanand.warrier-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Tanya K Jajodia <tanya.k.jajodia-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h | 423 +++++++++++++++++++++
.../infiniband/ulp/hfi_vnic/hfi_vnic_internal.h | 33 ++
2 files changed, 456 insertions(+)
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h
index aa93f62..73c2bdc 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_encap.h
@@ -52,7 +52,430 @@
* and decapsulation of Ethernet packets
*/
+#include <linux/types.h>
+#include <rdma/ib_mad.h>
+
+/* EMA class version */
+#define HFI_EMA_CLASS_VERSION 0x80
+
+/*
+ * Define the Intel vendor management class for HFI
+ * ETHERNET MANAGEMENT
+ */
+#define HFI_MGMT_CLASS_INTEL_EMA 0x34
+
+/* EM attribute IDs */
+#define HFI_EM_ATTR_CLASS_PORT_INFO 0x0001
+#define HFI_EM_ATTR_VESWPORT_INFO 0x0011
+#define HFI_EM_ATTR_VESWPORT_MAC_ENTRIES 0x0012
+#define HFI_EM_ATTR_IFACE_UCAST_MACS 0x0013
+#define HFI_EM_ATTR_IFACE_MCAST_MACS 0x0014
+#define HFI_EM_ATTR_DELETE_VESW 0x0015
+#define HFI_EM_ATTR_VESWPORT_SUMMARY_COUNTERS 0x0020
+#define HFI_EM_ATTR_VESWPORT_ERROR_COUNTERS 0x0022
+
#define HFI_VESW_MAX_NUM_DEF_PORT 16
#define HFI_VNIC_MAX_NUM_PCP 8
+#define HFI_VNIC_EMA_DATA (OPA_MGMT_MAD_SIZE - IB_MGMT_VENDOR_HDR)
+
+/* Defines for vendor specific notice(trap) attributes */
+#define HFI_INTEL_EMA_NOTICE_TYPE_INFO 0x04
+
+/* INTEL OUI */
+#define INTEL_OUI_1 0x00
+#define INTEL_OUI_2 0x06
+#define INTEL_OUI_3 0x6a
+
+/* Trap opcodes sent from VNIC */
+#define HFI_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE 0x1
+#define HFI_VESWPORT_TRAP_IFACE_MCAST_MAC_CHANGE 0x2
+#define HFI_VESWPORT_TRAP_ETH_LINK_STATUS_CHANGE 0x3
+
+#define HFI_VNIC_DLID_SD_IS_SRC_MAC(dlid_sd) (!!((dlid_sd) & 0x20))
+#define HFI_VNIC_DLID_SD_GET_DLID(dlid_sd) ((dlid_sd) >> 8)
+
+/**
+ * struct hfi_vesw_info - HFI vnic switch information
+ * @fabric_id: 10-bit fabric id
+ * @vesw_id: 12-bit virtual ethernet switch id
+ * @def_port_mask: bitmask of default ports
+ * @pkey: partition key
+ * @u_mcast_dlid: unknown multicast dlid
+ * @u_ucast_dlid: array of unknown unicast dlids
+ * @eth_mtu: MTUs for each vlan PCP
+ * @eth_mtu_non_vlan: MTU for non vlan packets
+ */
+struct hfi_vesw_info {
+ __be16 fabric_id;
+ __be16 vesw_id;
+
+ u8 rsvd0[6];
+ __be16 def_port_mask;
+
+ u8 rsvd1[2];
+ __be16 pkey;
+
+ u8 rsvd2[4];
+ __be32 u_mcast_dlid;
+ __be32 u_ucast_dlid[HFI_VESW_MAX_NUM_DEF_PORT];
+
+ u8 rsvd3[44];
+ __be16 eth_mtu[HFI_VNIC_MAX_NUM_PCP];
+ __be16 eth_mtu_non_vlan;
+ u8 rsvd4[2];
+} __packed;
+
+/**
+ * struct hfi_per_veswport_info - HFI vnic per port information
+ * @port_num: port number
+ * @eth_link_status: current ethernet link state
+ * @base_mac_addr: base mac address
+ * @config_state: configured port state
+ * @oper_state: operational port state
+ * @max_mac_tbl_ent: max number of mac table entries
+ * @max_smac_ent: max smac entries in mac table
+ * @mac_tbl_digest: mac table digest
+ * @encap_slid: base slid for the port
+ * @pcp_to_sc_uc: sc by pcp index for unicast ethernet packets
+ * @pcp_to_vl_uc: vl by pcp index for unicast ethernet packets
+ * @pcp_to_sc_mc: sc by pcp index for multicast ethernet packets
+ * @pcp_to_vl_mc: vl by pcp index for multicast ethernet packets
+ * @non_vlan_sc_uc: sc for non-vlan unicast ethernet packets
+ * @non_vlan_vl_uc: vl for non-vlan unicast ethernet packets
+ * @non_vlan_sc_mc: sc for non-vlan multicast ethernet packets
+ * @non_vlan_vl_mc: vl for non-vlan multicast ethernet packets
+ * @uc_macs_gen_count: generation count for unicast macs list
+ * @mc_macs_gen_count: generation count for multicast macs list
+ */
+struct hfi_per_veswport_info {
+ __be32 port_num;
+
+ u8 eth_link_status;
+ u8 rsvd0[3];
+
+ u8 base_mac_addr[ETH_ALEN];
+ u8 config_state;
+ u8 oper_state;
+
+ __be16 max_mac_tbl_ent;
+ __be16 max_smac_ent;
+ __be32 mac_tbl_digest;
+ u8 rsvd1[4];
+
+ __be32 encap_slid;
+
+ u8 pcp_to_sc_uc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_vl_uc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_sc_mc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_vl_mc[HFI_VNIC_MAX_NUM_PCP];
+
+ u8 non_vlan_sc_uc;
+ u8 non_vlan_vl_uc;
+ u8 non_vlan_sc_mc;
+ u8 non_vlan_vl_mc;
+
+ u8 rsvd2[48];
+
+ __be16 uc_macs_gen_count;
+ __be16 mc_macs_gen_count;
+
+ u8 rsvd3[8];
+} __packed;
+
+/**
+ * struct hfi_veswport_info - HFI vnic port information
+ * @vesw: HFI vnic switch information
+ * @vport: HFI vnic per port information
+ *
+ * On host, each of the virtual ethernet ports belongs
+ * to a different virtual ethernet switches.
+ */
+struct hfi_veswport_info {
+ struct hfi_vesw_info vesw;
+ struct hfi_per_veswport_info vport;
+};
+
+/**
+ * struct hfi_veswport_mactable_entry - single entry in the forwarding table
+ * @mac_addr: MAC address
+ * @mac_addr_mask: MAC address bit mask
+ * @dlid_sd: Matching DLID and side data
+ *
+ * On the host each virtual ethernet port will have
+ * a forwarding table. These tables are used to
+ * map a MAC to a LID and other data. For more
+ * details see struct hfi_veswport_mactable_entries.
+ * This is the structure of a single mactable entry
+ */
+struct hfi_veswport_mactable_entry {
+ u8 mac_addr[ETH_ALEN];
+ u8 mac_addr_mask[ETH_ALEN];
+ __be32 dlid_sd;
+} __packed;
+
+/**
+ * struct hfi_veswport_mactable - Forwarding table array
+ * @offset: mac table starting offset
+ * @num_entries: Number of entries to get or set
+ * @mac_tbl_digest: mac table digest
+ * @tbl_entries[]: Array of table entries
+ *
+ * The EM sends down this structure in a MAD indicating
+ * the starting offset in the forwarding table that this
+ * entry is to be loaded into and the number of entries
+ * that that this MAD instance contains
+ * The mac_tbl_digest has been added to this MAD structure. It will be set by
+ * the EM and it will be used by the EM to check if there are any
+ * discrepancies with this value and the value
+ * maintained by the EM in the case of VNIC port being deleted or unloaded
+ * A new instantiation of a VNIC will always have a value of zero.
+ * This value is stored as part of the vnic adapter structure and will be
+ * accessed by the GET and SET routines for both the mactable entries and the
+ * veswport info.
+ */
+struct hfi_veswport_mactable {
+ __be16 offset;
+ __be16 num_entries;
+ __be32 mac_tbl_digest;
+ struct hfi_veswport_mactable_entry tbl_entries[0];
+} __packed;
+
+/**
+ * struct hfi_veswport_summary_counters - summary counters
+ * @vp_instance: vport instance on the HFI port
+ * @vesw_id: virtual ethernet switch id
+ * @veswport_num: virtual ethernet switch port number
+ * @tx_errors: transmit errors
+ * @rx_errors: receive errors
+ * @tx_packets: transmit packets
+ * @rx_packets: receive packets
+ * @tx_bytes: transmit bytes
+ * @rx_bytes: receive bytes
+ * @tx_unicast: unicast packets transmitted
+ * @tx_mcastbcast: multicast/broadcast packets transmitted
+ * @tx_untagged: non-vlan packets transmitted
+ * @tx_vlan: vlan packets transmitted
+ * @tx_64_size: transmit packet length is 64 bytes
+ * @tx_65_127: transmit packet length is >=65 and < 127 bytes
+ * @tx_128_255: transmit packet length is >=128 and < 255 bytes
+ * @tx_256_511: transmit packet length is >=256 and < 511 bytes
+ * @tx_512_1023: transmit packet length is >=512 and < 1023 bytes
+ * @tx_1024_1518: transmit packet length is >=1024 and < 1518 bytes
+ * @tx_1519_max: transmit packet length >= 1519 bytes
+ * @rx_unicast: unicast packets received
+ * @rx_mcastbcast: multicast/broadcast packets received
+ * @rx_untagged: non-vlan packets received
+ * @rx_vlan: vlan packets received
+ * @rx_64_size: received packet length is 64 bytes
+ * @rx_65_127: received packet length is >=65 and < 127 bytes
+ * @rx_128_255: received packet length is >=128 and < 255 bytes
+ * @rx_256_511: received packet length is >=256 and < 511 bytes
+ * @rx_512_1023: received packet length is >=512 and < 1023 bytes
+ * @rx_1024_1518: received packet length is >=1024 and < 1518 bytes
+ * @rx_1519_max: received packet length >= 1519 bytes
+ *
+ * All the above are counters of corresponding conditions.
+ */
+struct hfi_veswport_summary_counters {
+ __be16 vp_instance;
+ __be16 vesw_id;
+ __be32 veswport_num;
+
+ __be64 tx_errors;
+ __be64 rx_errors;
+ __be64 tx_packets;
+ __be64 rx_packets;
+ __be64 tx_bytes;
+ __be64 rx_bytes;
+
+ __be64 tx_unicast;
+ __be64 tx_mcastbcast;
+
+ __be64 tx_untagged;
+ __be64 tx_vlan;
+
+ __be64 tx_64_size;
+ __be64 tx_65_127;
+ __be64 tx_128_255;
+ __be64 tx_256_511;
+ __be64 tx_512_1023;
+ __be64 tx_1024_1518;
+ __be64 tx_1519_max;
+
+ __be64 rx_unicast;
+ __be64 rx_mcastbcast;
+
+ __be64 rx_untagged;
+ __be64 rx_vlan;
+
+ __be64 rx_64_size;
+ __be64 rx_65_127;
+ __be64 rx_128_255;
+ __be64 rx_256_511;
+ __be64 rx_512_1023;
+ __be64 rx_1024_1518;
+ __be64 rx_1519_max;
+
+ __be64 reserved[16];
+} __packed;
+
+/**
+ * struct hfi_veswport_error_counters - error counters
+ * @vp_instance: vport instance on the HFI port
+ * @vesw_id: virtual ethernet switch id
+ * @veswport_num: virtual ethernet switch port number
+ * @tx_errors: transmit errors
+ * @rx_errors: receive errors
+ * @tx_smac_filt: smac filter errors
+ * @tx_dlid_zero: transmit packets with invalid dlid
+ * @tx_logic: other transmit errors
+ * @tx_drop_state: packet tansmission in non-forward port state
+ * @rx_bad_veswid: received packet with invalid vesw id
+ * @rx_runt: received ethernet packet with length < 64 bytes
+ * @rx_oversize: received ethernet packet with length > MTU size
+ * @rx_eth_down: received packets when interface is down
+ * @rx_drop_state: received packets in non-forwarding port state
+ * @rx_logic: other receive errors
+ *
+ * All the above are counters of corresponding erorr conditions.
+ */
+struct hfi_veswport_error_counters {
+ __be16 vp_instance;
+ __be16 vesw_id;
+ __be32 veswport_num;
+
+ __be64 tx_errors;
+ __be64 rx_errors;
+
+ __be64 rsvd0;
+ __be64 tx_smac_filt;
+ __be64 rsvd1;
+ __be64 rsvd2;
+ __be64 rsvd3;
+ __be64 tx_dlid_zero;
+ __be64 rsvd4;
+ __be64 tx_logic;
+ __be64 rsvd5;
+ __be64 tx_drop_state;
+
+ __be64 rx_bad_veswid;
+ __be64 rsvd6;
+ __be64 rx_runt;
+ __be64 rx_oversize;
+ __be64 rsvd7;
+ __be64 rx_eth_down;
+ __be64 rx_drop_state;
+ __be64 rx_logic;
+ __be64 rsvd8;
+
+ __be64 rsvd9[16];
+} __packed;
+
+/**
+ * struct hfi_veswport_trap - Trap message sent to EM by VNIC
+ * @fabric_id: 10 bit fabric id
+ * @veswid: 12 bit virtual ethernet switch id
+ * @veswportnum: logical port number on the Virtual switch
+ * @hfiportnum: physical port num (redundant on host)
+ * @veswportindex: switch port index on hfi port 0 based
+ * @opcode: operation
+ * @reserved: 32 bit for alignment
+ *
+ * The VNIC will send trap messages to the Ethernet manager to
+ * inform it about changes to the VNIC config, behaviour etc.
+ * This is the format of the trap payload.
+ */
+struct hfi_veswport_trap {
+ __be16 fabric_id;
+ __be16 veswid;
+ __be32 veswportnum;
+ __be16 hfiportnum;
+ u8 veswportindex;
+ u8 opcode;
+ __be32 reserved;
+} __packed;
+
+/**
+ * struct hfi_vnic_iface_macs_entry - single entry in the mac list
+ * @mac_addr: MAC address
+ */
+struct hfi_vnic_iface_mac_entry {
+ u8 mac_addr[ETH_ALEN];
+};
+
+/**
+ * struct hfi_veswport_iface_macs - Msg to set globally administered MAC
+ * @start_idx: position of first entry (0 based)
+ * @num_macs_in_msg: number of MACs in this message
+ * @tot_macs_in_lst: The total number of MACs the agent has
+ * @gen_count: gen_count to indicate change
+ * @entry: The mac list entry
+ *
+ * Same attribute IDS and attribute modifiers as in locally administered
+ * addresses used to set globally administered addresses
+ */
+struct hfi_veswport_iface_macs {
+ __be16 start_idx;
+ __be16 num_macs_in_msg;
+ __be16 tot_macs_in_lst;
+ __be16 gen_count;
+ struct hfi_vnic_iface_mac_entry entry[0];
+} __packed;
+
+/**
+ * struct hfi_vnic_vema_mad - Generic VEMA MAD
+ * @mad_hdr: Generic MAD header
+ * @rmpp_hdr: RMPP header for vendor specific MADs
+ * @oui: Unique org identifier
+ * @data: MAD data
+ */
+struct hfi_vnic_vema_mad {
+ struct ib_mad_hdr mad_hdr;
+ struct ib_rmpp_hdr rmpp_hdr;
+ u8 reserved;
+ u8 oui[3];
+ u8 data[HFI_VNIC_EMA_DATA];
+};
+
+/**
+ * struct hfi_vnic_notice_attr - Generic Notice MAD
+ * @gen_type: Generic/Specific bit and type of notice
+ * @oui_1: Vendor ID byte 1
+ * @oui_2: Vendor ID byte 2
+ * @oui_3: Vendor ID byte 3
+ * @trap_num: Trap number
+ * @toggle_count: Notice toggle bit and count value
+ * @issuer_lid: Trap issuer's lid
+ * @issuer_gid: Issuer GID (only if Report method)
+ * @raw_data: Trap message body
+ */
+struct hfi_vnic_notice_attr {
+ u8 gen_type;
+ u8 oui_1;
+ u8 oui_2;
+ u8 oui_3;
+ __be16 trap_num;
+ __be16 toggle_count;
+ __be32 issuer_lid;
+ __be32 reserved;
+ u8 issuer_gid[16];
+ u8 raw_data[64];
+} __packed;
+
+/**
+ * struct hfi_vnic_vema_mad_trap - Generic VEMA MAD Trap
+ * @mad_hdr: Generic MAD header
+ * @rmpp_hdr: RMPP header for vendor specific MADs
+ * @oui: Unique org identifier
+ * @notice: Notice structure
+ */
+struct hfi_vnic_vema_mad_trap {
+ struct ib_mad_hdr mad_hdr;
+ struct ib_rmpp_hdr rmpp_hdr;
+ u8 reserved;
+ u8 oui[3];
+ struct hfi_vnic_notice_attr notice;
+};
+
#endif /* _HFI_VNIC_ENCAP_H */
diff --git a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
index 106563d..5ce925f 100644
--- a/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
+++ b/drivers/infiniband/ulp/hfi_vnic/hfi_vnic_internal.h
@@ -72,6 +72,8 @@
/**
* struct __hfi_vesw_info - HFI vnic virtual switch info
+ *
+ * Same as hfi_vesw_info without bitwise attribute.
*/
struct __hfi_vesw_info {
u16 fabric_id;
@@ -95,6 +97,8 @@ struct __hfi_vesw_info {
/**
* struct __hfi_per_veswport_info - HFI vnic per port info
+ *
+ * Same as hfi_per_veswport_info without bitwise attribute.
*/
struct __hfi_per_veswport_info {
u32 port_num;
@@ -133,6 +137,8 @@ struct __hfi_per_veswport_info {
/**
* struct __hfi_veswport_info - HFI vnic port info
+ *
+ * Same as hfi_veswport_info without bitwise attribute.
*/
struct __hfi_veswport_info {
struct __hfi_vesw_info vesw;
@@ -140,6 +146,21 @@ struct __hfi_veswport_info {
};
/**
+ * struct __hfi_veswport_trap - HFI vnic trap info
+ *
+ * Same as hfi_veswport_trap without bitwise attribute.
+ */
+struct __hfi_veswport_trap {
+ u16 fabric_id;
+ u16 veswid;
+ u32 veswportnum;
+ u16 hfiportnum;
+ u8 veswportindex;
+ u8 opcode;
+ u32 reserved;
+} __packed;
+
+/**
* struct hfi_vnic_adapter - HFI VNIC netdev private data structure
* @netdev: pointer to associated netdev
* @ibdev: ib device
@@ -175,6 +196,18 @@ struct hfi_vnic_adapter {
#define v_warn(format, arg...) \
netdev_warn(adapter->netdev, format, ## arg)
+/* The maximum allowed entries in the mac table */
+#define HFI_VNIC_MAC_TBL_MAX_ENTRIES 2048
+/* Limit of smac entries in mac table */
+#define HFI_VNIC_MAX_SMAC_LIMIT 256
+
+/* The last octet of the MAC address is used as the key to the hash table */
+#define HFI_VNIC_MAC_HASH_IDX 5
+
+/* The VNIC MAC hash table is of size 2^8 */
+#define HFI_VNIC_MAC_TBL_HASH_BITS 8
+#define HFI_VNIC_MAC_TBL_SIZE BIT(HFI_VNIC_MAC_TBL_HASH_BITS)
+
struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct ib_device *ibdev,
u8 port_num, u8 vport_num);
void hfi_vnic_rem_netdev(struct hfi_vnic_adapter *adapter);
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 40+ messages in thread* Re: [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <1486498990-76562-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
` (2 preceding siblings ...)
2017-02-07 20:23 ` [RFC v3 04/11] IB/hfi-vnic: VNIC Ethernet Management (EM) structure definitions Vishwanathapura, Niranjana
@ 2017-02-07 20:36 ` Leon Romanovsky
[not found] ` <20170207203649.GW6005-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
2017-02-07 21:28 ` Bart Van Assche
2017-02-12 13:26 ` Liran Liss
5 siblings, 1 reply; 40+ messages in thread
From: Leon Romanovsky @ 2017-02-07 20:36 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: dledford-H+wXaHxf7aLQT0dZR+AlfA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w
[-- Attachment #1: Type: text/plain, Size: 563 bytes --]
On Tue, Feb 07, 2017 at 12:22:59PM -0800, Vishwanathapura, Niranjana wrote:
> ChangeLog:
> =========
> v2 => v3:
> a) Introduce and adopt generic RDMA netdev interface including,
> - having bottom hfi1 driver directly interfacing with netstack.
> - optimizing interface between hfi_vnic and hfi1 driver.
> b) Remove bitfield usage.
> c) Move hfi_vnic driver to drivers/infiniband/ulp folder.
I didn't read patches yet, and prefer to ask it in advance. Does this new ULP work with all
drivers/infiniband/hw/* devices as it is expected from ULP?
Thanks
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 40+ messages in thread* Re: [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <1486498990-76562-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
` (3 preceding siblings ...)
2017-02-07 20:36 ` [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC) Leon Romanovsky
@ 2017-02-07 21:28 ` Bart Van Assche
2017-02-07 21:44 ` Hefty, Sean
2017-02-12 13:26 ` Liran Liss
5 siblings, 1 reply; 40+ messages in thread
From: Bart Van Assche @ 2017-02-07 21:28 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA,
niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w
On Tue, 2017-02-07 at 12:22 -0800, Vishwanathapura, Niranjana wrote:
> Intel Omni-Path Host Fabric Interface (HFI) Virtual Network Interface
> Controller (VNIC) feature supports Ethernet functionality over Omni-Path
> fabric by encapsulating the Ethernet packets between HFI nodes.
This may have been stated before, but what is missing from this description
is an explanation of why accepting an Ethernet over RDMA driver in the
upstream kernel is considered useful. We already have an IPoIB driver, so
why to add another driver to the kernel tree for communicating IP packets
over an RDMA network?
Bart.--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 40+ messages in thread* RE: [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC)
2017-02-07 21:28 ` Bart Van Assche
@ 2017-02-07 21:44 ` Hefty, Sean
2017-02-07 21:58 ` Bart Van Assche
0 siblings, 1 reply; 40+ messages in thread
From: Hefty, Sean @ 2017-02-07 21:44 UTC (permalink / raw)
To: Bart Van Assche, dledford, Vishwanathapura, Niranjana
Cc: netdev, linux-rdma, Dalessandro, Dennis, Weiny, Ira
> This may have been stated before, but what is missing from this
> description
> is an explanation of why accepting an Ethernet over RDMA driver in the
> upstream kernel is considered useful. We already have an IPoIB driver,
> so
> why to add another driver to the kernel tree for communicating IP
> packets
> over an RDMA network?
This is Ethernet - not IP - encapsulation over a non-InfiniBand device/protocol.
^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC)
2017-02-07 21:44 ` Hefty, Sean
@ 2017-02-07 21:58 ` Bart Van Assche
[not found] ` <1486504716.2581.8.camel-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
0 siblings, 1 reply; 40+ messages in thread
From: Bart Van Assche @ 2017-02-07 21:58 UTC (permalink / raw)
To: dledford, sean.hefty, niranjana.vishwanathapura
Cc: netdev, linux-rdma, dennis.dalessandro, ira.weiny
On Tue, 2017-02-07 at 21:44 +0000, Hefty, Sean wrote:
> This is Ethernet - not IP - encapsulation over a non-InfiniBand device/protocol.
That's more than clear from the cover letter. In my opinion the cover letter
should explain why it is considered useful to have such a driver upstream
and what the use cases are of encapsulating Ethernet frames inside RDMA
packets.
Bart.
^ permalink raw reply [flat|nested] 40+ messages in thread
* RE: [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <1486498990-76562-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
` (4 preceding siblings ...)
2017-02-07 21:28 ` Bart Van Assche
@ 2017-02-12 13:26 ` Liran Liss
2017-02-13 17:09 ` Jason Gunthorpe
5 siblings, 1 reply; 40+ messages in thread
From: Liran Liss @ 2017-02-12 13:26 UTC (permalink / raw)
To: Vishwanathapura, Niranjana, dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w
> From: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org [mailto:linux-rdma-
> owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org] On Behalf Of Vishwanathapura, Niranjana
>
> ChangeLog:
> =========
> v2 => v3:
> a) Introduce and adopt generic RDMA netdev interface including,
> - having bottom hfi1 driver directly interfacing with netstack.
> - optimizing interface between hfi_vnic and hfi1 driver.
> b) Remove bitfield usage.
> c) Move hfi_vnic driver to drivers/infiniband/ulp folder.
The vnic driver should be placed under drivers/infiniband/hw/hfi1/* since it is HFI-specific.
--Liran
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 40+ messages in thread* Re: [RFC v3 00/11] HFI Virtual Network Interface Controller (VNIC)
2017-02-12 13:26 ` Liran Liss
@ 2017-02-13 17:09 ` Jason Gunthorpe
[not found] ` <20170213170935.GD25430-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
0 siblings, 1 reply; 40+ messages in thread
From: Jason Gunthorpe @ 2017-02-13 17:09 UTC (permalink / raw)
To: Liran Liss
Cc: Vishwanathapura, Niranjana, dledford, linux-rdma, netdev,
dennis.dalessandro, ira.weiny
On Sun, Feb 12, 2017 at 01:26:35PM +0000, Liran Liss wrote:
> > From: linux-rdma-owner@vger.kernel.org [mailto:linux-rdma-
> > owner@vger.kernel.org] On Behalf Of Vishwanathapura, Niranjana
>
> >
> > ChangeLog:
> > =========
> > v2 => v3:
> > a) Introduce and adopt generic RDMA netdev interface including,
> > - having bottom hfi1 driver directly interfacing with netstack.
> > - optimizing interface between hfi_vnic and hfi1 driver.
> > b) Remove bitfield usage.
> > c) Move hfi_vnic driver to drivers/infiniband/ulp folder.
>
> The vnic driver should be placed under drivers/infiniband/hw/hfi1/*
> since it is HFI-specific.
I think they should call it opa_vnic and keep it in ulp to avoid this
confusion.
Jason
^ permalink raw reply [flat|nested] 40+ messages in thread