* [PATCH 0/1] XDP Program for Ip forward
From: cjacob @ 2017-10-03 7:37 UTC (permalink / raw)
To: netdev; +Cc: Christina.Jacob, linux-kernel, linux-arm-kernel
The patch below implements port to port forwarding through route table and arp
table lookup for ipv4 packets using bpf_redirect helper function and lpm_trie
map. This has an improved performance over the normal kernel stack ip forward.
Implementation details.
-----------------------
The program uses one map each for arp table, route table and packet count.
The number of entries the program can process is limited by the size of the
map used.
In the xdp3_user.c,
initially, the routing table is read and stored in an lpm trie map.
The arp table is read and stored in an array map. There are two netlink sockets
that listen to any change in the route table and arp table.
There are two types of changes to the route table.
1.New
The new entries are added to the lpm trie with proper key and prefix
length. If there is a another entry in the route table with a different
metric(only metric is considered), then the values are compared and the
one with lowest metric is added to the trie node.
2.Deletion
On deletion from the route table, the particular node is removed and the
entire route table is read again to check if there is another entry with
the same key and prefix length but a different metric. If it exists it
is added to the lpm trie.
This implementation depends on Patch bpf: Implement map_delete_elem for
BPF_MAP_TYPE_LPM_TRIE which is not yet upstreamed.
There are two types of changes to the route table
1.New
The new arp entries are added in the in the array map directly with the
ip address as the key and the destination mac address as the value.
2.Delete
The entry corresponding to the particular ip is deleted from the
arp table map.
Another map is maintained for entries in the route table having 32 bit mask.
Such entries can have a corresponding arp entry which if stored together with
the route entry in an array map and can be accessed in O(1) time, thereby
eliminating the trie lookup and arp lookup.
In the xdp3_kern.c,
The array map for the 32 bit mask entries is checked to see if there is a key
that exactly matches the destination ip. If it has a non zero destination mac
entry then the xdp data is updated accordingly. Otherwise a proper route and
arp table lookup is done using the lpm_trie and the arp table array map.
Usage: ./xdp3 [-S] <ifindex1...ifindexn>
-S to choose generic xdp implementation
[Default is driver xdp implementation]
ifindex - the index of the interface to which
the xdp program has to be attached.
in 4.14-rc3 kernel.
cjacob (1):
xdp: Sample xdp program implementing ip forward
samples/bpf/Makefile | 4 +
samples/bpf/xdp3_kern.c | 204 +++++++++++++++
samples/bpf/xdp3_user.c | 649 +++++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 857 insertions(+), 0 deletions(-)
create mode 100644 samples/bpf/xdp3_kern.c
create mode 100644 samples/bpf/xdp3_user.c
^ permalink raw reply
* [PATCH 1/1] xdp: Sample xdp program implementing ip forward
From: cjacob @ 2017-10-03 7:37 UTC (permalink / raw)
To: netdev; +Cc: Christina.Jacob, linux-kernel, linux-arm-kernel
In-Reply-To: <1507016225-319-1-git-send-email-Christina.Jacob@cavium.com>
Implements port to port forwarding with route table and arp table
lookup for ipv4 packets using bpf_redirect helper function and
lpm_trie map.
Signed-off-by: cjacob <Christina.Jacob@cavium.com>
---
samples/bpf/Makefile | 4 +
samples/bpf/xdp3_kern.c | 204 +++++++++++++++
samples/bpf/xdp3_user.c | 649 +++++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 857 insertions(+), 0 deletions(-)
diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
index cf17c79..cc9cc0b 100644
--- a/samples/bpf/Makefile
+++ b/samples/bpf/Makefile
@@ -28,6 +28,7 @@ hostprogs-y += test_cgrp2_sock
hostprogs-y += test_cgrp2_sock2
hostprogs-y += xdp1
hostprogs-y += xdp2
+hostprogs-y += xdp3
hostprogs-y += test_current_task_under_cgroup
hostprogs-y += trace_event
hostprogs-y += sampleip
@@ -73,6 +74,7 @@ test_cgrp2_sock2-objs := bpf_load.o $(LIBBPF) test_cgrp2_sock2.o
xdp1-objs := bpf_load.o $(LIBBPF) xdp1_user.o
# reuse xdp1 source intentionally
xdp2-objs := bpf_load.o $(LIBBPF) xdp1_user.o
+xdp3-objs := bpf_load.o $(LIBBPF) xdp3_user.o
test_current_task_under_cgroup-objs := bpf_load.o $(LIBBPF) cgroup_helpers.o \
test_current_task_under_cgroup_user.o
trace_event-objs := bpf_load.o $(LIBBPF) trace_event_user.o
@@ -114,6 +116,7 @@ always += parse_varlen.o parse_simple.o parse_ldabs.o
always += test_cgrp2_tc_kern.o
always += xdp1_kern.o
always += xdp2_kern.o
+always += xdp3_kern.o
always += test_current_task_under_cgroup_kern.o
always += trace_event_kern.o
always += sampleip_kern.o
@@ -160,6 +163,7 @@ HOSTLOADLIBES_map_perf_test += -lelf -lrt
HOSTLOADLIBES_test_overhead += -lelf -lrt
HOSTLOADLIBES_xdp1 += -lelf
HOSTLOADLIBES_xdp2 += -lelf
+HOSTLOADLIBES_xdp3 += -lelf
HOSTLOADLIBES_test_current_task_under_cgroup += -lelf
HOSTLOADLIBES_trace_event += -lelf
HOSTLOADLIBES_sampleip += -lelf
diff --git a/samples/bpf/xdp3_kern.c b/samples/bpf/xdp3_kern.c
new file mode 100644
index 0000000..62d905d
--- /dev/null
+++ b/samples/bpf/xdp3_kern.c
@@ -0,0 +1,204 @@
+/* Copyright (c) 2016 PLUMgrid
+ *
+ * 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.
+ */
+#define KBUILD_MODNAME "foo"
+#include <uapi/linux/bpf.h>
+#include <linux/in.h>
+#include <linux/if_ether.h>
+#include <linux/if_packet.h>
+#include <linux/if_vlan.h>
+#include <linux/ip.h>
+#include <linux/ipv6.h>
+#include "bpf_helpers.h"
+#include <linux/slab.h>
+#include <net/ip_fib.h>
+
+struct trie_value {
+ __u8 prefix[4];
+ long value;
+ int gw;
+ int ifindex;
+ int metric;
+};
+
+union key_4 {
+ u32 b32[2];
+ u8 b8[8];
+};
+
+struct arp_entry {
+ int dst;
+ long mac;
+};
+
+struct direct_map {
+ long mac;
+ int ifindex;
+ struct arp_entry arp;
+};
+
+/* Map for trie implementation*/
+struct bpf_map_def SEC("maps") lpm_map = {
+ .type = BPF_MAP_TYPE_LPM_TRIE,
+ .key_size = 8,
+ .value_size =
+ sizeof(struct trie_value),
+ .max_entries = 50,
+ .map_flags = BPF_F_NO_PREALLOC,
+};
+
+/* Map for counter*/
+struct bpf_map_def SEC("maps") rxcnt = {
+ .type = BPF_MAP_TYPE_PERCPU_ARRAY,
+ .key_size = sizeof(u32),
+ .value_size = sizeof(long),
+ .max_entries = 256,
+};
+
+/* Map for ARP table*/
+struct bpf_map_def SEC("maps") arp_table = {
+ .type = BPF_MAP_TYPE_HASH,
+ .key_size = sizeof(int),
+ .value_size = sizeof(long),
+ .max_entries = 50,
+};
+
+/* Map to keep the exact match entries in the route table*/
+struct bpf_map_def SEC("maps") exact_match = {
+ .type = BPF_MAP_TYPE_HASH,
+ .key_size = sizeof(int),
+ .value_size = sizeof(struct direct_map),
+ .max_entries = 50,
+};
+
+/**
+ * Function to set source and destination mac of the packet
+ */
+static inline void set_src_dst_mac(void *data, void *src, void *dst)
+{
+ unsigned short *p = data;
+ unsigned short *dest = dst;
+ unsigned short *source = src;
+
+ p[3] = source[0];
+ p[4] = source[1];
+ p[5] = source[2];
+ p[0] = dest[0];
+ p[1] = dest[1];
+ p[2] = dest[2];
+}
+
+/**
+ * Parse IPV4 packet to get SRC, DST IP and protocol
+ */
+static inline int parse_ipv4(void *data, u64 nh_off, void *data_end,
+ unsigned int *src, unsigned int *dest)
+{
+ struct iphdr *iph = data + nh_off;
+
+ if (iph + 1 > data_end)
+ return 0;
+ *src = (unsigned int)iph->saddr;
+ *dest = (unsigned int)iph->daddr;
+ return iph->protocol;
+}
+
+SEC("xdp3")
+int xdp_prog3(struct xdp_md *ctx)
+{
+ void *data_end = (void *)(long)ctx->data_end;
+ void *data = (void *)(long)ctx->data;
+ struct ethhdr *eth = data;
+ int rc = XDP_DROP, forward_to;
+ long *value;
+ struct trie_value *prefix_value;
+ long *dest_mac = NULL, *src_mac = NULL;
+ u16 h_proto;
+ u64 nh_off;
+ u32 ipproto;
+ union key_4 key4;
+
+ nh_off = sizeof(*eth);
+ if (data + nh_off > data_end)
+ return rc;
+
+ h_proto = eth->h_proto;
+
+ if (h_proto == htons(ETH_P_8021Q) || h_proto == htons(ETH_P_8021AD)) {
+ struct vlan_hdr *vhdr;
+
+ vhdr = data + nh_off;
+ nh_off += sizeof(struct vlan_hdr);
+ if (data + nh_off > data_end)
+ return rc;
+ h_proto = vhdr->h_vlan_encapsulated_proto;
+ }
+ if (h_proto == htons(ETH_P_ARP)) {
+ return XDP_PASS;
+ } else if (h_proto == htons(ETH_P_IP)) {
+ int src_ip = 0, dest_ip = 0;
+ struct direct_map *direct_entry;
+
+ ipproto = parse_ipv4(data, nh_off, data_end, &src_ip, &dest_ip);
+ direct_entry = (struct direct_map *)bpf_map_lookup_elem
+ (&exact_match, &dest_ip);
+ /*check for exact match, this would give a faster lookup*/
+ if (direct_entry && direct_entry->mac &&
+ direct_entry->arp.mac) {
+ src_mac = &direct_entry->mac;
+ dest_mac = &direct_entry->arp.mac;
+ forward_to = direct_entry->ifindex;
+ } else {
+ /*Look up in the trie for lpm*/
+ // Key for trie
+ key4.b32[0] = 32;
+ key4.b8[4] = dest_ip % 0x100;
+ key4.b8[5] = (dest_ip >> 8) % 0x100;
+ key4.b8[6] = (dest_ip >> 16) % 0x100;
+ key4.b8[7] = (dest_ip >> 24) % 0x100;
+ prefix_value =
+ ((struct trie_value *)bpf_map_lookup_elem
+ (&lpm_map, &key4));
+ if (!prefix_value) {
+ return XDP_DROP;
+ } else {
+ src_mac = &prefix_value->value;
+ if (src_mac) {
+ dest_mac = (long *)bpf_map_lookup_elem
+ (&arp_table, &dest_ip);
+ if (!dest_mac) {
+ if (prefix_value->gw) {
+ dest_ip = *(unsigned int *)(&(prefix_value->gw));
+ dest_mac = (long *)bpf_map_lookup_elem
+ (&arp_table, &dest_ip);
+ } else {
+ return XDP_DROP;
+ }
+ }
+ forward_to = prefix_value->ifindex;
+ } else {
+ return XDP_DROP;
+ }
+ }
+ }
+ } else {
+ ipproto = 0;
+ }
+ if (src_mac && dest_mac) {
+ set_src_dst_mac(data, src_mac,
+ dest_mac);
+ value = bpf_map_lookup_elem
+ (&rxcnt, &ipproto);
+ if (value)
+ *value += 1;
+ return bpf_redirect(
+ forward_to,
+ 0);
+ }
+ return rc;
+}
+
+char _license[] SEC("license") = "GPL";
diff --git a/samples/bpf/xdp3_user.c b/samples/bpf/xdp3_user.c
new file mode 100644
index 0000000..451b522
--- /dev/null
+++ b/samples/bpf/xdp3_user.c
@@ -0,0 +1,649 @@
+/* Copyright (c) 2016 PLUMgrid
+ *
+ * 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.
+ */
+#include <linux/bpf.h>
+#include <linux/netlink.h>
+#include <linux/rtnetlink.h>
+#include <assert.h>
+#include <errno.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/socket.h>
+#include <unistd.h>
+#include "bpf_load.h"
+#include "libbpf.h"
+#include <arpa/inet.h>
+#include <fcntl.h>
+#include <poll.h>
+#include <net/if.h>
+#include <netdb.h>
+#include <sys/ioctl.h>
+#include "bpf_util.h"
+#include <sys/syscall.h>
+
+int sock, sock_arp, flags = 0;
+char buf[8192];
+static int total_ifindex;
+char **index_list;
+
+static int get_route_table(int rtm_family);
+static void int_exit(int sig)
+{
+ int i = 0, index;
+
+ for (i = 0; i < total_ifindex; i++) {
+ index = strtoul(index_list[i], NULL, 0);
+ set_link_xdp_fd(index, -1, flags);
+ }
+ exit(0);
+}
+
+static void close_and_exit(int sig)
+{
+ int i = 0, index;
+
+ close(sock);
+ close(sock_arp);
+
+ for (i = 0; i < total_ifindex; i++) {
+ index = strtoul(index_list[i], NULL, 0);
+ set_link_xdp_fd(index, -1, flags);
+ }
+ exit(0);
+}
+
+/* Get the mac address of the interface given interface name */
+static long *getmac(char *iface)
+{
+ int fd;
+ struct ifreq ifr;
+ long *mac = NULL;
+
+ fd = socket(AF_INET, SOCK_DGRAM, 0);
+ ifr.ifr_addr.sa_family = AF_INET;
+ strncpy(ifr.ifr_name, iface, IFNAMSIZ - 1);
+ ioctl(fd, SIOCGIFHWADDR, &ifr);
+ mac = (long *)ifr.ifr_hwaddr.sa_data;
+ close(fd);
+ return mac;
+}
+
+static int recv_msg(struct sockaddr_nl sock_addr, int sock)
+{
+ char *buf_ptr;
+ struct nlmsghdr *nh;
+ int len, nll = 0;
+
+ buf_ptr = buf;
+ while (1) {
+ len = recv(sock, buf_ptr, sizeof(buf) - nll, 0);
+ if (len < 0)
+ return len;
+
+ nh = (struct nlmsghdr *)buf_ptr;
+
+ if (nh->nlmsg_type == NLMSG_DONE)
+ break;
+ buf_ptr += len;
+ nll += len;
+ if ((sock_addr.nl_groups & RTMGRP_NEIGH) == RTMGRP_NEIGH)
+ break;
+
+ if ((sock_addr.nl_groups & RTMGRP_IPV4_ROUTE) ==
+ RTMGRP_IPV4_ROUTE)
+ break;
+ }
+ return nll;
+}
+
+/* Function to parse the route entry returned by netlink
+ * Updates the route entry related map entries
+ */
+static void read_route(struct nlmsghdr *nh, int nll)
+{
+ struct route_table {
+ int dst, gw, dst_len, iface, metric;
+ long *mac;
+ char iface_name[IFNAMSIZ];
+ } route;
+ struct arp_table {
+ int dst;
+ long mac;
+ };
+
+ struct direct_map {
+ long mac;
+ int ifindex;
+ struct arp_table arp;
+ } direct_entry;
+ int i;
+ int rtm_family;
+ struct bpf_lpm_trie_key *prefix_key;
+ char dsts[24], gws[24], ifs[16], dsts_len[24], metrics[24];
+ struct rtmsg *rt_msg;
+ int rtl;
+ struct rtattr *rt_attr;
+
+ if (nh->nlmsg_type == RTM_DELROUTE)
+ printf("DELETING Route entry\n");
+ else if (nh->nlmsg_type == RTM_GETROUTE)
+ printf("READING Route entry\n");
+ else if (nh->nlmsg_type == RTM_NEWROUTE)
+ printf("NEW Route entry\n");
+ else
+ printf("%d\n", nh->nlmsg_type);
+
+ bzero(&route, sizeof(route));
+ printf("Destination\tGateway\t\tGenmask\tMetric\tIface\n");
+ for (; NLMSG_OK(nh, nll); nh = NLMSG_NEXT(nh, nll)) {
+ rt_msg = (struct rtmsg *)NLMSG_DATA(nh);
+ rtm_family = rt_msg->rtm_family;
+ if (rtm_family == AF_INET)
+ if (rt_msg->rtm_table != RT_TABLE_MAIN)
+ continue;
+ rt_attr = (struct rtattr *)RTM_RTA(rt_msg);
+ rtl = RTM_PAYLOAD(nh);
+
+ for (; RTA_OK(rt_attr, rtl); rt_attr = RTA_NEXT(rt_attr, rtl)) {
+ switch (rt_attr->rta_type) {
+ case NDA_DST:
+ sprintf(dsts, "%d",
+ *((int *)RTA_DATA(rt_attr)));
+ break;
+ case RTA_GATEWAY:
+ sprintf(gws, "%d", *((int *)RTA_DATA(rt_attr)));
+ break;
+ case RTA_OIF:
+ sprintf(ifs, "%d", *((int *)RTA_DATA(rt_attr)));
+ break;
+ case RTA_METRICS:
+ sprintf(metrics, "%d",
+ *((int *)RTA_DATA(rt_attr)));
+ default:
+ break;
+ }
+ }
+ sprintf(dsts_len, "%d", rt_msg->rtm_dst_len);
+
+ route.dst = atoi(dsts);
+ route.dst_len = atoi(dsts_len);
+ route.gw = atoi(gws);
+ route.iface = atoi(ifs);
+ route.metric = atoi(metrics);
+ if_indextoname(route.iface, route.iface_name);
+ route.mac = getmac(route.iface_name);
+ printf("%x\t\t%x\t\t%d\t%d\t%d\n", route.dst, route.gw,
+ route.dst_len, route.metric, route.iface);
+ if (rtm_family == AF_INET) {
+ struct trie_value {
+ __u8 prefix[4];
+ long value;
+ int gw;
+ int ifindex;
+ int metric;
+ } *prefix_value;
+
+ prefix_key = alloca(sizeof(*prefix_key) + 3);
+ prefix_value = alloca(sizeof(*prefix_value));
+
+ prefix_key->prefixlen = 32;
+ prefix_key->prefixlen = route.dst_len;
+ direct_entry.mac = *route.mac & 0xffffffffffff;
+ direct_entry.ifindex = route.iface;
+ direct_entry.arp.mac = 0;
+ direct_entry.arp.dst = 0;
+ if (route.dst_len == 32) {
+ if (nh->nlmsg_type == RTM_DELROUTE) {
+ assert(bpf_map_delete_elem(
+ map_fd[3],
+ &route.dst
+ ) == 0);
+ } else {
+ if (bpf_map_lookup_elem(map_fd[2],
+ &route.dst,
+ &direct_entry.arp.mac
+ ) == 0)
+ direct_entry.arp.dst = route.dst;
+
+ assert(bpf_map_update_elem(map_fd[3],
+ &route.dst,
+ &direct_entry,
+ 0) == 0);
+ }
+ }
+ for (i = 0; i < 4; i++)
+ prefix_key->data[i] =
+ (route.dst >> i * 8) % 0x100;
+ if (bpf_map_lookup_elem(map_fd[0], prefix_key,
+ prefix_value) < 0) {
+ for (i = 0; i < 4; i++)
+ prefix_value->prefix[i] =
+ prefix_key->data[i];
+ prefix_value->value =
+ *route.mac & 0xffffffffffff;
+ prefix_value->ifindex = route.iface;
+ prefix_value->gw = route.gw;
+ prefix_value->metric = route.metric;
+
+ assert(bpf_map_update_elem(map_fd[0],
+ prefix_key,
+ prefix_value, 0
+ ) == 0);
+ } else {
+ if (nh->nlmsg_type == RTM_DELROUTE) {
+ printf("deleting entry\n");
+ printf("prefix key=%d.%d.%d.%d/%d",
+ prefix_key->data[0],
+ prefix_key->data[1],
+ prefix_key->data[2],
+ prefix_key->data[3],
+ prefix_key->prefixlen);
+ assert(bpf_map_delete_elem(map_fd[0],
+ prefix_key
+ ) == 0);
+ /* Rereading the route table to check if
+ * there is an entry with the same
+ * prefix but a different metric as the
+ * deleted enty.
+ */
+ get_route_table(AF_INET);
+ } else if (prefix_key->data[0] ==
+ prefix_value->prefix[0] &&
+ prefix_key->data[1] ==
+ prefix_value->prefix[1] &&
+ prefix_key->data[2] ==
+ prefix_value->prefix[2] &&
+ prefix_key->data[3] ==
+ prefix_value->prefix[3] &&
+ route.metric >= prefix_value->metric) {
+ continue;
+ } else {
+ for (i = 0; i < 4; i++)
+ prefix_value->prefix[i] =
+ prefix_key->data[i];
+ prefix_value->value =
+ *route.mac & 0xffffffffffff;
+ prefix_value->ifindex = route.iface;
+ prefix_value->gw = route.gw;
+ prefix_value->metric = route.metric;
+ assert(bpf_map_update_elem(
+ map_fd[0],
+ prefix_key,
+ prefix_value,
+ 0) == 0);
+ }
+ }
+ }
+ bzero(&route, sizeof(route));
+ bzero(dsts, sizeof(dsts));
+ bzero(dsts_len, sizeof(dsts_len));
+ bzero(gws, sizeof(gws));
+ bzero(ifs, sizeof(ifs));
+ bzero(&route, sizeof(route));
+ }
+}
+
+/* Function to read the existing route table when the process is launched*/
+static int get_route_table(int rtm_family)
+{
+ struct {
+ struct nlmsghdr nl;
+ struct rtmsg rt;
+ char buf[8192];
+ } req;
+
+ int sock, seq = 0;
+ struct sockaddr_nl sa;
+ struct msghdr msg;
+ struct iovec iov;
+ int ret = 0;
+ struct nlmsghdr *nh;
+ int nll;
+
+ sock = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
+ if (sock < 0) {
+ printf("open netlink socket: %s\n", strerror(errno));
+ return -1;
+ }
+ bzero(&sa, sizeof(sa));
+ sa.nl_family = AF_NETLINK;
+ if (bind(sock, (struct sockaddr *)&sa, sizeof(sa)) < 0) {
+ printf("bind to netlink: %s\n", strerror(errno));
+ ret = -1;
+ goto cleanup;
+ }
+ bzero(&req, sizeof(req));
+ req.nl.nlmsg_len = NLMSG_LENGTH(sizeof(struct rtmsg));
+ req.nl.nlmsg_flags = NLM_F_REQUEST | NLM_F_DUMP;
+ req.nl.nlmsg_type = RTM_GETROUTE;
+
+ req.rt.rtm_family = rtm_family;
+ req.rt.rtm_table = RT_TABLE_MAIN;
+ req.nl.nlmsg_pid = 0;
+ req.nl.nlmsg_seq = ++seq;
+ bzero(&msg, sizeof(msg));
+ iov.iov_base = (void *)&req.nl;
+ iov.iov_len = req.nl.nlmsg_len;
+ msg.msg_iov = &iov;
+ msg.msg_iovlen = 1;
+ ret = sendmsg(sock, &msg, 0);
+ if (ret < 0) {
+ printf("send to netlink: %s\n", strerror(errno));
+ ret = -1;
+ goto cleanup;
+ }
+ bzero(buf, sizeof(buf));
+ nll = recv_msg(sa, sock);
+ if (nll < 0) {
+ printf("recv from netlink: %s\n", strerror(nll));
+ ret = -1;
+ goto cleanup;
+ }
+ nh = (struct nlmsghdr *)buf;
+ read_route(nh, nll);
+cleanup:
+ close(sock);
+ return ret;
+}
+
+/* Function to parse the arp entry returned by netlink
+ * Updates the arp entry related map entries
+ */
+static void read_arp(struct nlmsghdr *nh, int nll)
+{
+ struct arp_table {
+ int dst;
+ long mac;
+ } arp_entry;
+ struct direct_map {
+ long mac;
+ int ifindex;
+ struct arp_table arp;
+ } direct_entry;
+
+ char dsts[24], mac[24];
+ struct ndmsg *rt_msg;
+ int rtl, i = 0, ndm_family;
+ struct rtattr *rt_attr;
+
+ if (nh->nlmsg_type == RTM_GETNEIGH)
+ printf("READING arp entry\n");
+ printf("Address\tHwAddress\n");
+ for (; NLMSG_OK(nh, nll); nh = NLMSG_NEXT(nh, nll)) {
+ i++;
+ rt_msg = (struct ndmsg *)NLMSG_DATA(nh);
+ rt_attr = (struct rtattr *)RTM_RTA(rt_msg);
+ ndm_family = rt_msg->ndm_family;
+ rtl = RTM_PAYLOAD(nh);
+ for (; RTA_OK(rt_attr, rtl); rt_attr = RTA_NEXT(rt_attr, rtl)) {
+ switch (rt_attr->rta_type) {
+ case NDA_DST:
+ sprintf(dsts, "%d",
+ *((int *)RTA_DATA(rt_attr)));
+ break;
+ case NDA_LLADDR:
+ sprintf(mac, "%ld",
+ *((long *)RTA_DATA(rt_attr)));
+ break;
+ default:
+ break;
+ }
+ }
+ arp_entry.dst = atoi(dsts);
+ arp_entry.mac = atol(mac);
+ printf("%x\t\t%lx\n", arp_entry.dst, arp_entry.mac);
+ if (ndm_family == AF_INET) {
+ if (bpf_map_lookup_elem(map_fd[3], &arp_entry.dst,
+ &direct_entry) == 0) {
+ if (nh->nlmsg_type == RTM_DELNEIGH) {
+ direct_entry.arp.dst = 0;
+ direct_entry.arp.mac = 0;
+ } else if (nh->nlmsg_type == RTM_NEWNEIGH) {
+ direct_entry.arp.dst = arp_entry.dst;
+ direct_entry.arp.mac = arp_entry.mac;
+ }
+ assert(bpf_map_update_elem(map_fd[3],
+ &arp_entry.dst,
+ &direct_entry, 0
+ ) == 0);
+ bzero(&direct_entry, sizeof(direct_entry));
+ }
+ if (nh->nlmsg_type == RTM_DELNEIGH) {
+ assert(bpf_map_delete_elem(map_fd[2],
+ &arp_entry.dst) == 0);
+ } else if (nh->nlmsg_type == RTM_NEWNEIGH) {
+ assert(bpf_map_update_elem(map_fd[2],
+ &arp_entry.dst,
+ &arp_entry.mac, 0
+ ) == 0);
+ }
+ }
+ bzero(&arp_entry, sizeof(arp_entry));
+ bzero(dsts, sizeof(dsts));
+ }
+}
+
+/* Function to read the existing arp table when the process is launched*/
+static int get_arp_table(int rtm_family)
+{
+ struct {
+ struct nlmsghdr nl;
+ struct ndmsg rt;
+ char buf[8192];
+ } req;
+
+ int sock, seq = 0;
+ struct sockaddr_nl sa;
+ struct msghdr msg;
+ struct iovec iov;
+ int ret = 0;
+ struct nlmsghdr *nh;
+ int nll;
+
+ sock = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
+ if (sock < 0) {
+ printf("open netlink socket: %s\n", strerror(errno));
+ return -1;
+ }
+ bzero(&sa, sizeof(sa));
+ sa.nl_family = AF_NETLINK;
+ if (bind(sock, (struct sockaddr *)&sa, sizeof(sa)) < 0) {
+ printf("bind to netlink: %s\n", strerror(errno));
+ ret = -1;
+ goto cleanup;
+ }
+ bzero(&req, sizeof(req));
+ req.nl.nlmsg_len = NLMSG_LENGTH(sizeof(struct rtmsg));
+ req.nl.nlmsg_flags = NLM_F_REQUEST | NLM_F_DUMP;
+ req.nl.nlmsg_type = RTM_GETNEIGH;
+ req.rt.ndm_state = NUD_REACHABLE;
+ req.rt.ndm_family = rtm_family;
+ req.nl.nlmsg_pid = 0;
+ req.nl.nlmsg_seq = ++seq;
+ bzero(&msg, sizeof(msg));
+ iov.iov_base = (void *)&req.nl;
+ iov.iov_len = req.nl.nlmsg_len;
+ msg.msg_iov = &iov;
+ msg.msg_iovlen = 1;
+ ret = sendmsg(sock, &msg, 0);
+ if (ret < 0) {
+ printf("send to netlink: %s\n", strerror(errno));
+ ret = -1;
+ goto cleanup;
+ }
+ bzero(buf, sizeof(buf));
+ nll = recv_msg(sa, sock);
+ if (nll < 0) {
+ printf("recv from netlink: %s\n", strerror(nll));
+ ret = -1;
+ goto cleanup;
+ }
+ nh = (struct nlmsghdr *)buf;
+ read_arp(nh, nll);
+cleanup:
+ close(sock);
+ return ret;
+}
+
+/* Function to keep track and update changes in route and arp table
+ * Give regular statistics of packets forwarded
+ */
+static int monitor_route(void)
+{
+ struct sockaddr_nl la, lr;
+ struct nlmsghdr *nh;
+ int nll, ret = 0;
+ const unsigned int nr_keys = 256;
+ int interval = 5;
+ unsigned int nr_cpus = bpf_num_possible_cpus();
+ __u64 values[nr_cpus], prev[nr_keys][nr_cpus];
+ __u32 key;
+ int i;
+ struct pollfd fds_route, fds_arp;
+
+ sock = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
+ if (sock < 0) {
+ printf("open netlink socket: %s\n", strerror(errno));
+ return -1;
+ }
+
+ fcntl(sock, F_SETFL, O_NONBLOCK);
+ bzero(&lr, sizeof(lr));
+ lr.nl_family = AF_NETLINK;
+ lr.nl_groups = RTMGRP_IPV6_ROUTE | RTMGRP_IPV4_ROUTE | RTMGRP_NOTIFY;
+ if (bind(sock, (struct sockaddr *)&lr, sizeof(lr)) < 0) {
+ printf("bind to netlink: %s\n", strerror(errno));
+ ret = -1;
+ goto cleanup;
+ }
+ fds_route.fd = sock;
+ fds_route.events = POLL_IN;
+
+ sock_arp = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
+ if (sock_arp < 0) {
+ printf("open netlink socket: %s\n", strerror(errno));
+ return -1;
+ }
+
+ fcntl(sock_arp, F_SETFL, O_NONBLOCK);
+ bzero(&la, sizeof(la));
+ la.nl_family = AF_NETLINK;
+ la.nl_groups = RTMGRP_NEIGH | RTMGRP_NOTIFY;
+ if (bind(sock_arp, (struct sockaddr *)&la, sizeof(la)) < 0) {
+ printf("bind to netlink: %s\n", strerror(errno));
+ ret = -1;
+ goto cleanup;
+ }
+ fds_arp.fd = sock_arp;
+ fds_arp.events = POLL_IN;
+
+ memset(prev, 0, sizeof(prev));
+ do {
+ signal(SIGINT, close_and_exit);
+ signal(SIGTERM, close_and_exit);
+
+ sleep(interval);
+ for (key = 0; key < nr_keys; key++) {
+ __u64 sum = 0;
+
+ assert(bpf_map_lookup_elem(map_fd[1], &key, values) == 0);
+ for (i = 0; i < nr_cpus; i++)
+ sum += (values[i] - prev[key][i]);
+ if (sum)
+ printf("proto %u: %10llu pkt/s\n",
+ key, sum / interval);
+ memcpy(prev[key], values, sizeof(values));
+ }
+
+ bzero(buf, sizeof(buf));
+ if (poll(&fds_route, 1, 3) == POLL_IN) {
+ nll = recv_msg(lr, sock);
+ if (nll < 0) {
+ printf("recv from netlink: %s\n",
+ strerror(nll));
+ ret = -1;
+ goto cleanup;
+ }
+
+ nh = (struct nlmsghdr *)buf;
+ printf("Routing table updated.\n");
+ read_route(nh, nll);
+ }
+ bzero(buf, sizeof(buf));
+ if (poll(&fds_arp, 1, 3) == POLL_IN) {
+ nll = recv_msg(la, sock_arp);
+ if (nll < 0) {
+ printf("recv from netlink: %s\n",
+ strerror(nll));
+ ret = -1;
+ goto cleanup;
+ }
+
+ nh = (struct nlmsghdr *)buf;
+ read_arp(nh, nll);
+ }
+
+ } while (1);
+cleanup:
+ close(sock);
+ return ret;
+}
+
+int main(int ac, char **argv)
+{
+ char filename[256];
+ int i = 1, index;
+
+ snprintf(filename, sizeof(filename), "%s_kern.o", argv[0]);
+ printf("Entering user program\n");
+ if (ac < 2) {
+ printf("usage: %s [-S] IFINDEX\n", argv[0]);
+ return 1;
+ }
+ if (!strcmp(argv[1], "-S")) {
+ flags = XDP_FLAGS_SKB_MODE;
+ total_ifindex = ac - 2;
+ index_list = (argv + 2);
+ } else {
+ flags = 0;
+ total_ifindex = ac - 1;
+ index_list = (argv + 1);
+ }
+printf("Loading bpf program\n");
+ if (load_bpf_file(filename)) {
+ printf("%s", bpf_log_buf);
+ return 1;
+ }
+ printf("\n**************loading bpf file*********************\n\n\n");
+ if (!prog_fd[0]) {
+ printf("load_bpf_file: %s\n", strerror(errno));
+ return 1;
+ }
+
+ for (i = 0; i < total_ifindex; i++) {
+ index = strtoul(index_list[i], NULL, 0);
+ if (set_link_xdp_fd(index, prog_fd[0], flags) < 0) {
+ printf("link set xdp fd failed\n");
+ return 1;
+ }
+ printf("Attached to %d\n", index);
+ }
+ signal(SIGINT, int_exit);
+ signal(SIGTERM, int_exit);
+
+ printf("*******************ROUTE TABLE*************************\n\n\n");
+ get_route_table(AF_INET);
+ printf("*******************ARP TABLE***************************\n\n\n");
+ get_arp_table(AF_INET);
+ if (monitor_route() < 0) {
+ printf("Error in receiving route update");
+ return 1;
+ }
+
+ return 0;
+}
--
1.7.1
^ permalink raw reply related
* BUG in free_netdev() on ppp link deletion
From: Beniamino Galvani @ 2017-10-03 7:44 UTC (permalink / raw)
To: linux-ppp, netdev, Paul Mackerras, Guillaume Nault, David Ahern,
Gao Feng
Hi,
I see the following BUG on 4.14-rc2 and previous versions (reproduced
with 4.11 as well):
------------[ cut here ]------------
kernel BUG at net/core/dev.c:8141!
invalid opcode: 0000 [#1] SMP
Modules linked in: pppoe pppox ppp_generic slhc cfg80211 rfkill joydev uinput tun sunrpc snd_hda_codec_generic snd_hda_intel snd_hda_codec snd_hda_core snd_hwdep snd_seq snd_seq_device crct10dif_pclmul crc32_pclmul snd_pcm ghash_clmulni_intel snd_timer ppdev snd parport_pc pcspkr soundcore parport virtio_balloon i2c_piix4 8139too virtio_console qxl drm_kms_helper ttm drm crc32c_intel serio_raw e1000 virtio_pci 8139cp floppy virtio_ring qemu_fw_cfg mii virtio ata_generic pata_acpi
CPU: 1 PID: 1557 Comm: pppd Not tainted 4.14.0-rc2+ #4
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1.fc26 04/01/2014
task: ffff9cd4ba5e4c80 task.stack: ffffc28a40570000
RIP: 0010:free_netdev+0x107/0x110
RSP: 0018:ffffc28a40573d88 EFLAGS: 00010297
RAX: 0000000000000002 RBX: ffff9cd4bb0178c0 RCX: 000000000001e94c
RDX: 0000000000000001 RSI: 0000000000000286 RDI: 0000000000000000
RBP: ffffc28a40573da0 R08: 0000000000000001 R09: 000000000000019b
R10: 000000000000059b R11: 0000000000000001 R12: ffff9cd4bb017000
R13: ffff9cd4bb017060 R14: ffff9cd4bb01793c R15: 0000000000000000
FS: 00007fa720aa2840(0000) GS:ffff9cd4bfc80000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007fffeefa3000 CR3: 0000000072b34002 CR4: 00000000001606e0
Call Trace:
ppp_destroy_interface+0xd8/0xe0 [ppp_generic]
ppp_disconnect_channel+0xda/0x110 [ppp_generic]
ppp_unregister_channel+0x5e/0x110 [ppp_generic]
pppox_unbind_sock+0x23/0x30 [pppox]
pppoe_connect+0x130/0x440 [pppoe]
SYSC_connect+0x98/0x110
? do_fcntl+0x2c0/0x5d0
SyS_connect+0xe/0x10
entry_SYSCALL_64_fastpath+0x1a/0xa5
RIP: 0033:0x7fa71f4af840
RSP: 002b:00007ffe4ea40bf8 EFLAGS: 00000246 ORIG_RAX: 000000000000002a
RAX: ffffffffffffffda RBX: 0000556d37ae0538 RCX: 00007fa71f4af840
RDX: 000000000000001e RSI: 00007ffe4ea40c00 RDI: 0000000000000008
RBP: 0000556d37b2a1b0 R08: 0000556d396e95b0 R09: 0000000000000008
R10: 00000000aaaaaaab R11: 0000000000000246 R12: 0000556d37adc008
R13: 0000556d37adc004 R14: 0000556d37b2a1a4 R15: 0000000000000000
Code: 04 00 00 04 e8 cb 52 e3 ff 5b 41 5c 41 5d 5d c3 41 0f b7 84 24 32 02 00 00 4c 89 e7 48 29 c7 e8 80 8b aa ff 5b 41 5c 41 5d 5d c3 <0f> 0b 0f 1f 80 00 00 00 00 0f 1f 44 00 00 55 48 89 e5 41 57 41
RIP: free_netdev+0x107/0x110 RSP: ffffc28a40573d88
---[ end trace ed294ff0cc40eeff ]---
To reproduce this, establish a PPP connection through pppd, then bring
down and delete the ppp interface:
# pppd nodetach lock user client plugin rp-pppoe.so ens11 noauth nodeflate password password &
Plugin rp-pppoe.so loaded.
RP-PPPoE plugin version 3.8p compiled against pppd 2.4.7
PPP session is 16
Connected to fe:54:00:5f:04:13 via interface ens11
Using interface ppp0
Connect: ppp0 <--> ens11
CHAP authentication succeeded: Access granted
CHAP authentication succeeded
peer from calling number FE:54:00:5F:04:13 authorized
local IP address 3.1.1.10
remote IP address 3.1.1.1
# ip l set ppp0 down
# ip l del ppp0
It does not happen every time but only when ppp_destroy_interface() is
called with dev->reg_state = UNREGISTERING, set by the concurrent
rtnl_delete_link().
Beniamino
^ permalink raw reply
* Re: [PATCH v4 net-next 0/8] flow_dissector: Protocol specific flow dissector offload
From: Jiri Pirko @ 2017-10-03 7:46 UTC (permalink / raw)
To: Tom Herbert
Cc: David Miller, Hannes Frederic Sowa, Tom Herbert,
Linux Kernel Network Developers, Rohit Seth
In-Reply-To: <CALx6S37DhT095fY92HWZAKQvMkqjxqq08nxJqL=wTP8JorHHDg@mail.gmail.com>
Fri, Sep 29, 2017 at 07:59:35PM CEST, tom@herbertland.com wrote:
>On Fri, Sep 29, 2017 at 10:42 AM, David Miller <davem@davemloft.net> wrote:
>> From: Tom Herbert <tom@herbertland.com>
>> Date: Fri, 29 Sep 2017 08:48:55 -0700
>>
>>> The flow_dissector interface is not a uAPI.
>>
>> That's not true, insofar as cls_flower.c uses the flow_dissector
>> therefore if you change the flow_dissector in certain ways then
>> cls_flower.c might have it's behavior changed and that is in fact UAPI
>> facing.
>
>Then I would suggest adding another flag like FLOW_DISSECTOR_F_FLOWER
>and when anyone puts new code into flow_dissector they can wrap it
>with "if !(flags & FLOW_DISSECTOR_F_FLOWER)". If the flower uAPI is
>subsequently update then the conditional can be removed. This way
>flower can support maintain its APIs, but we can still still extend
>and improve flow_dissector for othersuse cases.
This is not flower-specific problem. Flow_dissector is a servant of many.
As such, it is instructed what should it do. If you want to
change the way inner headers are parsed, you should either:
1) change the callers so they are behaving the same as before
2) make the flow_dissection change optional so the caller can say if he
wants original or new behaviour.
^ permalink raw reply
* Re: [PATCH 1/2] net: phonet: mark header_ops as const
From: kbuild test robot @ 2017-10-03 7:46 UTC (permalink / raw)
To: Lin Zhang; +Cc: kbuild-all, davem, netdev, courmisch, Lin Zhang
In-Reply-To: <1506762519-4709-1-git-send-email-xiaolou4617@gmail.com>
[-- Attachment #1: Type: text/plain, Size: 1396 bytes --]
Hi Lin,
[auto build test ERROR on net-next/master]
[also build test ERROR on v4.14-rc3 next-20170929]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
url: https://github.com/0day-ci/linux/commits/Lin-Zhang/net-phonet-mark-header_ops-as-const/20171003-145726
config: x86_64-randconfig-x000-201740 (attached as .config)
compiler: gcc-6 (Debian 6.2.0-3) 6.2.0 20160901
reproduce:
# save the attached .config to linux build tree
make ARCH=x86_64
All errors (new ones prefixed by >>):
>> net//phonet/af_phonet.c:152:25: error: conflicting type qualifiers for 'phonet_header_ops'
const struct header_ops phonet_header_ops = {
^~~~~~~~~~~~~~~~~
In file included from net//phonet/af_phonet.c:32:0:
include/linux/if_phonet.h:13:26: note: previous declaration of 'phonet_header_ops' was here
extern struct header_ops phonet_header_ops;
^~~~~~~~~~~~~~~~~
vim +/phonet_header_ops +152 net//phonet/af_phonet.c
151
> 152 const struct header_ops phonet_header_ops = {
153 .create = pn_header_create,
154 .parse = pn_header_parse,
155 };
156 EXPORT_SYMBOL(phonet_header_ops);
157
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 27699 bytes --]
^ permalink raw reply
* [patch net-next v2 0/7] mlxsw: Add support for partial multicast route offload
From: Jiri Pirko @ 2017-10-03 7:58 UTC (permalink / raw)
To: netdev
Cc: davem, yotamg, idosch, mlxsw, nikolay, andrew, dsa, edumazet,
willemb, johannes.berg, dcaratti, pabeni, daniel, f.fainelli, fw,
gfree.wind
From: Jiri Pirko <jiri@mellanox.com>
Yotam says:
Previous patchset introduced support for offloading multicast MFC routes to
the Spectrum hardware. As described in that patchset, no partial offloading
is supported, i.e if a route has one output interface which is not a valid
offloadable device (e.g. pimreg device, dummy device, management NIC), the
route is trapped to the CPU and the forwarding is done in slow-path.
Add support for partial offloading of multicast routes, by letting the
hardware to forward the packet to all the in-hardware devices, while the
kernel ipmr module will continue forwarding to all other interfaces.
Similarly to the bridge, the kernel ipmr module will forward a marked
packet to an interface only if the interface has a different parent ID than
the packet's ingress interfaces.
The first patch introduces the offload_mr_fwd_mark skb field, which can be
used by offloading drivers to indicate that a packet had already gone
through multicast forwarding in hardware, similarly to the offload_fwd_mark
field that indicates that a packet had already gone through L2 forwarding
in hardware.
Patches 2 and 3 change the ipmr module to not forward packets that had
already been forwarded by the hardware, i.e. packets that are marked with
offload_mr_fwd_mark and the ingress VIF shares the same parent ID with the
egress VIF.
Patches 4, 5, 6 and 7 add the support in the mlxsw Spectrum driver for trap
and forward routes, while marking the trapped packets with the
offload_mr_fwd_mark.
Yotam Gigi (7):
skbuff: Add the offload_mr_fwd_mark field
ipv4: ipmr: Add the parent ID field to VIF struct
ipv4: ipmr: Don't forward packets already forwarded by hardware
mlxsw: acl: Introduce ACL trap and forward action
mlxsw: spectrum: Add trap for multicast trap-and-forward routes
mlxsw: spectrum: mr_tcam: Add trap-and-forward multicast route
mlxsw: spectrum: mr: Support trap-and-forward routes
.../mellanox/mlxsw/core_acl_flex_actions.c | 17 ++++++++
.../mellanox/mlxsw/core_acl_flex_actions.h | 2 +
drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 13 ++++++
drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.c | 17 ++++----
drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.h | 1 +
.../net/ethernet/mellanox/mlxsw/spectrum_mr_tcam.c | 8 ++++
drivers/net/ethernet/mellanox/mlxsw/trap.h | 2 +
include/linux/mroute.h | 1 +
include/linux/skbuff.h | 1 +
net/ipv4/ipmr.c | 48 +++++++++++++++++++---
10 files changed, 96 insertions(+), 14 deletions(-)
--
2.9.5
^ permalink raw reply
* [patch net-next v2 1/7] skbuff: Add the offload_mr_fwd_mark field
From: Jiri Pirko @ 2017-10-03 7:58 UTC (permalink / raw)
To: netdev
Cc: davem, yotamg, idosch, mlxsw, nikolay, andrew, dsa, edumazet,
willemb, johannes.berg, dcaratti, pabeni, daniel, f.fainelli, fw,
gfree.wind
In-Reply-To: <20171003075812.1540-1-jiri@resnulli.us>
From: Yotam Gigi <yotamg@mellanox.com>
Similarly to the offload_fwd_mark field, the offload_mr_fwd_mark field is
used to allow partial offloading of MFC multicast routes.
Switchdev drivers can offload MFC multicast routes to the hardware by
registering to the FIB notification chain. When one of the route output
interfaces is not offload-able, i.e. has different parent ID, the route
cannot be fully offloaded by the hardware. Examples to non-offload-able
devices are a management NIC, dummy device, pimreg device, etc.
Similar problem exists in the bridge module, as one bridge can hold
interfaces with different parent IDs. At the bridge, the problem is solved
by the offload_fwd_mark skb field.
Currently, when a route cannot go through full offload, the only solution
for a switchdev driver is not to offload it at all and let the packet go
through slow path.
Using the offload_mr_fwd_mark field, a driver can indicate that a packet
was already forwarded by hardware to all the devices with the same parent
ID as the input device. Further patches in this patch-set are going to
enhance ipmr to skip multicast forwarding to devices with the same parent
ID if a packets is marked with that field.
The reason why the already existing "offload_fwd_mark" bit cannot be used
is that a switchdev driver would want to make the distinction between a
packet that has already gone through L2 forwarding but did not go through
multicast forwarding, and a packet that has already gone through both L2
and multicast forwarding.
For example: when a packet is ingressing from a switchport enslaved to a
bridge, which is configured with multicast forwarding, the following
scenarios are possible:
- The packet can be trapped to the CPU due to exception while multicast
forwarding (for example, MTU error). In that case, it had already gone
through L2 forwarding in the hardware, thus A switchdev driver would
want to set the skb->offload_fwd_mark and not the
skb->offload_mr_fwd_mark.
- The packet can also be trapped due to a pimreg/dummy device used as one
of the output interfaces. In that case, it can go through both L2 and
(partial) multicast forwarding inside the hardware, thus a switchdev
driver would want to set both the skb->offload_fwd_mark and
skb->offload_mr_fwd_mark.
Signed-off-by: Yotam Gigi <yotamg@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellaox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
include/linux/skbuff.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 19e64bf..ada8214 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -772,6 +772,7 @@ struct sk_buff {
__u8 remcsum_offload:1;
#ifdef CONFIG_NET_SWITCHDEV
__u8 offload_fwd_mark:1;
+ __u8 offload_mr_fwd_mark:1;
#endif
#ifdef CONFIG_NET_CLS_ACT
__u8 tc_skip_classify:1;
--
2.9.5
^ permalink raw reply related
* [patch net-next v2 2/7] ipv4: ipmr: Add the parent ID field to VIF struct
From: Jiri Pirko @ 2017-10-03 7:58 UTC (permalink / raw)
To: netdev
Cc: davem, yotamg, idosch, mlxsw, nikolay, andrew, dsa, edumazet,
willemb, johannes.berg, dcaratti, pabeni, daniel, f.fainelli, fw,
gfree.wind
In-Reply-To: <20171003075812.1540-1-jiri@resnulli.us>
From: Yotam Gigi <yotamg@mellanox.com>
In order to allow the ipmr module to do partial multicast forwarding
according to the device parent ID, add the device parent ID field to the
VIF struct. This way, the forwarding path can use the parent ID field
without invoking switchdev calls, which requires the RTNL lock.
When a new VIF is added, set the device parent ID field in it by invoking
the switchdev_port_attr_get call.
Signed-off-by: Yotam Gigi <yotamg@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
v1->v2:
- Set the vif->dev_parent_id.len field
- Remove the vif->dev_parent_id_valid field and use len instead
- Set the vif->dev_parent_id to invalid if the device does not support the
switchdev PARENT_ID attribute
---
include/linux/mroute.h | 1 +
net/ipv4/ipmr.c | 11 +++++++++++
2 files changed, 12 insertions(+)
diff --git a/include/linux/mroute.h b/include/linux/mroute.h
index b072a84..8242d05 100644
--- a/include/linux/mroute.h
+++ b/include/linux/mroute.h
@@ -57,6 +57,7 @@ static inline bool ipmr_rule_default(const struct fib_rule *rule)
struct vif_device {
struct net_device *dev; /* Device we are using */
+ struct netdev_phys_item_id dev_parent_id; /* Device parent ID */
unsigned long bytes_in,bytes_out;
unsigned long pkt_in,pkt_out; /* Statistics */
unsigned long rate_limit; /* Traffic shaping (NI) */
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index a844738..1b161ad 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -67,6 +67,7 @@
#include <net/fib_rules.h>
#include <linux/netconf.h>
#include <net/nexthop.h>
+#include <net/switchdev.h>
struct ipmr_rule {
struct fib_rule common;
@@ -868,6 +869,9 @@ static int vif_add(struct net *net, struct mr_table *mrt,
struct vifctl *vifc, int mrtsock)
{
int vifi = vifc->vifc_vifi;
+ struct switchdev_attr attr = {
+ .id = SWITCHDEV_ATTR_ID_PORT_PARENT_ID,
+ };
struct vif_device *v = &mrt->vif_table[vifi];
struct net_device *dev;
struct in_device *in_dev;
@@ -942,6 +946,13 @@ static int vif_add(struct net *net, struct mr_table *mrt,
/* Fill in the VIF structures */
+ attr.orig_dev = dev;
+ if (!switchdev_port_attr_get(dev, &attr)) {
+ memcpy(v->dev_parent_id.id, attr.u.ppid.id, attr.u.ppid.id_len);
+ v->dev_parent_id.id_len = attr.u.ppid.id_len;
+ } else {
+ v->dev_parent_id.id_len = 0;
+ }
v->rate_limit = vifc->vifc_rate_limit;
v->local = vifc->vifc_lcl_addr.s_addr;
v->remote = vifc->vifc_rmt_addr.s_addr;
--
2.9.5
^ permalink raw reply related
* [patch net-next v2 3/7] ipv4: ipmr: Don't forward packets already forwarded by hardware
From: Jiri Pirko @ 2017-10-03 7:58 UTC (permalink / raw)
To: netdev
Cc: davem, yotamg, idosch, mlxsw, nikolay, andrew, dsa, edumazet,
willemb, johannes.berg, dcaratti, pabeni, daniel, f.fainelli, fw,
gfree.wind
In-Reply-To: <20171003075812.1540-1-jiri@resnulli.us>
From: Yotam Gigi <yotamg@mellanox.com>
Change the ipmr module to not forward packets if:
- The packet is marked with the offload_mr_fwd_mark, and
- Both input interface and output interface share the same parent ID.
This way, a packet can go through partial multicast forwarding in the
hardware, where it will be forwarded only to the devices that share the
same parent ID (AKA, reside inside the same hardware). The kernel will
forward the packet to all other interfaces.
To do this, add the ipmr_offload_forward helper, which per skb, ingress VIF
and egress VIF, returns whether the forwarding was offloaded to hardware.
The ipmr_queue_xmit frees the skb and does not forward it if the result is
a true value.
All the forwarding path code compiles out when the CONFIG_NET_SWITCHDEV is
not set.
Signed-off-by: Yotam Gigi <yotamg@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
v1->v2:
- Use dev_parent_id.len field instead of the dev_parent_id_valid field
---
net/ipv4/ipmr.c | 37 ++++++++++++++++++++++++++++++++-----
1 file changed, 32 insertions(+), 5 deletions(-)
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index 1b161ad..b3ee01b 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -1859,10 +1859,33 @@ static inline int ipmr_forward_finish(struct net *net, struct sock *sk,
return dst_output(net, sk, skb);
}
+#ifdef CONFIG_NET_SWITCHDEV
+static bool ipmr_forward_offloaded(struct sk_buff *skb, struct mr_table *mrt,
+ int in_vifi, int out_vifi)
+{
+ struct vif_device *out_vif = &mrt->vif_table[out_vifi];
+ struct vif_device *in_vif = &mrt->vif_table[in_vifi];
+
+ if (!skb->offload_mr_fwd_mark)
+ return false;
+ if (!out_vif->dev_parent_id.id_len || !in_vif->dev_parent_id.id_len)
+ return false;
+ return netdev_phys_item_id_same(&out_vif->dev_parent_id,
+ &in_vif->dev_parent_id);
+}
+#else
+static bool ipmr_forward_offloaded(struct sk_buff *skb, struct mr_table *mrt,
+ int in_vifi, int out_vifi)
+{
+ return false;
+}
+#endif
+
/* Processing handlers for ipmr_forward */
static void ipmr_queue_xmit(struct net *net, struct mr_table *mrt,
- struct sk_buff *skb, struct mfc_cache *c, int vifi)
+ int in_vifi, struct sk_buff *skb,
+ struct mfc_cache *c, int vifi)
{
const struct iphdr *iph = ip_hdr(skb);
struct vif_device *vif = &mrt->vif_table[vifi];
@@ -1883,6 +1906,9 @@ static void ipmr_queue_xmit(struct net *net, struct mr_table *mrt,
goto out_free;
}
+ if (ipmr_forward_offloaded(skb, mrt, in_vifi, vifi))
+ goto out_free;
+
if (vif->flags & VIFF_TUNNEL) {
rt = ip_route_output_ports(net, &fl4, NULL,
vif->remote, vif->local,
@@ -2060,8 +2086,8 @@ static void ip_mr_forward(struct net *net, struct mr_table *mrt,
struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
if (skb2)
- ipmr_queue_xmit(net, mrt, skb2, cache,
- psend);
+ ipmr_queue_xmit(net, mrt, true_vifi,
+ skb2, cache, psend);
}
psend = ct;
}
@@ -2072,9 +2098,10 @@ static void ip_mr_forward(struct net *net, struct mr_table *mrt,
struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
if (skb2)
- ipmr_queue_xmit(net, mrt, skb2, cache, psend);
+ ipmr_queue_xmit(net, mrt, true_vifi, skb2,
+ cache, psend);
} else {
- ipmr_queue_xmit(net, mrt, skb, cache, psend);
+ ipmr_queue_xmit(net, mrt, true_vifi, skb, cache, psend);
return;
}
}
--
2.9.5
^ permalink raw reply related
* [patch net-next v2 4/7] mlxsw: acl: Introduce ACL trap and forward action
From: Jiri Pirko @ 2017-10-03 7:58 UTC (permalink / raw)
To: netdev
Cc: davem, yotamg, idosch, mlxsw, nikolay, andrew, dsa, edumazet,
willemb, johannes.berg, dcaratti, pabeni, daniel, f.fainelli, fw,
gfree.wind
In-Reply-To: <20171003075812.1540-1-jiri@resnulli.us>
From: Yotam Gigi <yotamg@mellanox.com>
Use trap/discard flex action to implement trap and forward. The action will
later be used for multicast routing, as the multicast routing mechanism is
done using ACL flexible actions in Spectrum hardware. Using that action, it
will be possible to implement a trap-and-forward route.
Signed-off-by: Yotam Gigi <yotamg@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
.../net/ethernet/mellanox/mlxsw/core_acl_flex_actions.c | 17 +++++++++++++++++
.../net/ethernet/mellanox/mlxsw/core_acl_flex_actions.h | 2 ++
2 files changed, 19 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_actions.c b/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_actions.c
index bc55d0e..6a979a0 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_actions.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_actions.c
@@ -676,6 +676,7 @@ enum mlxsw_afa_trapdisc_trap_action {
MLXSW_ITEM32(afa, trapdisc, trap_action, 0x00, 24, 4);
enum mlxsw_afa_trapdisc_forward_action {
+ MLXSW_AFA_TRAPDISC_FORWARD_ACTION_FORWARD = 1,
MLXSW_AFA_TRAPDISC_FORWARD_ACTION_DISCARD = 3,
};
@@ -729,6 +730,22 @@ int mlxsw_afa_block_append_trap(struct mlxsw_afa_block *block, u16 trap_id)
}
EXPORT_SYMBOL(mlxsw_afa_block_append_trap);
+int mlxsw_afa_block_append_trap_and_forward(struct mlxsw_afa_block *block,
+ u16 trap_id)
+{
+ char *act = mlxsw_afa_block_append_action(block,
+ MLXSW_AFA_TRAPDISC_CODE,
+ MLXSW_AFA_TRAPDISC_SIZE);
+
+ if (!act)
+ return -ENOBUFS;
+ mlxsw_afa_trapdisc_pack(act, MLXSW_AFA_TRAPDISC_TRAP_ACTION_TRAP,
+ MLXSW_AFA_TRAPDISC_FORWARD_ACTION_FORWARD,
+ trap_id);
+ return 0;
+}
+EXPORT_SYMBOL(mlxsw_afa_block_append_trap_and_forward);
+
/* Forwarding Action
* -----------------
* Forwarding Action can be used to implement Policy Based Switching (PBS)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_actions.h b/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_actions.h
index 06b0be4..a8d3314 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_actions.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/core_acl_flex_actions.h
@@ -61,6 +61,8 @@ int mlxsw_afa_block_continue(struct mlxsw_afa_block *block);
int mlxsw_afa_block_jump(struct mlxsw_afa_block *block, u16 group_id);
int mlxsw_afa_block_append_drop(struct mlxsw_afa_block *block);
int mlxsw_afa_block_append_trap(struct mlxsw_afa_block *block, u16 trap_id);
+int mlxsw_afa_block_append_trap_and_forward(struct mlxsw_afa_block *block,
+ u16 trap_id);
int mlxsw_afa_block_append_fwd(struct mlxsw_afa_block *block,
u8 local_port, bool in_port);
int mlxsw_afa_block_append_vlan_modify(struct mlxsw_afa_block *block,
--
2.9.5
^ permalink raw reply related
* [patch net-next v2 6/7] mlxsw: spectrum: mr_tcam: Add trap-and-forward multicast route
From: Jiri Pirko @ 2017-10-03 7:58 UTC (permalink / raw)
To: netdev
Cc: davem, yotamg, idosch, mlxsw, nikolay, andrew, dsa, edumazet,
willemb, johannes.berg, dcaratti, pabeni, daniel, f.fainelli, fw,
gfree.wind
In-Reply-To: <20171003075812.1540-1-jiri@resnulli.us>
From: Yotam Gigi <yotamg@mellanox.com>
In addition to the current multicast route actions, which include trap
route action and a forward route action, add the trap-and-forward multicast
route action, and implement it in the multicast routing hardware logic.
To implement that, add a trap-and-forward ACL action as the last action in
the route flexible action set. The used trap is the ACL2 trap, which marks
the packets with offload_mr_forward_mark, to prevent the packet from being
forwarded again by the kernel.
Note: At that stage the offloading logic does not support trap-and-forward
multicast routes. This patch adds the support only in the hardware logic.
Signed-off-by: Yotam Gigi <yotamg@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.h | 1 +
drivers/net/ethernet/mellanox/mlxsw/spectrum_mr_tcam.c | 8 ++++++++
2 files changed, 9 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.h
index c851b23..5d26a12 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.h
@@ -42,6 +42,7 @@
enum mlxsw_sp_mr_route_action {
MLXSW_SP_MR_ROUTE_ACTION_FORWARD,
MLXSW_SP_MR_ROUTE_ACTION_TRAP,
+ MLXSW_SP_MR_ROUTE_ACTION_TRAP_AND_FORWARD,
};
enum mlxsw_sp_mr_route_prio {
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr_tcam.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr_tcam.c
index cda9e9a..3ffb28d 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr_tcam.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr_tcam.c
@@ -253,6 +253,7 @@ mlxsw_sp_mr_tcam_afa_block_create(struct mlxsw_sp *mlxsw_sp,
if (err)
goto err;
break;
+ case MLXSW_SP_MR_ROUTE_ACTION_TRAP_AND_FORWARD:
case MLXSW_SP_MR_ROUTE_ACTION_FORWARD:
/* If we are about to append a multicast router action, commit
* the erif_list.
@@ -266,6 +267,13 @@ mlxsw_sp_mr_tcam_afa_block_create(struct mlxsw_sp *mlxsw_sp,
erif_list->kvdl_index);
if (err)
goto err;
+
+ if (route_action == MLXSW_SP_MR_ROUTE_ACTION_TRAP_AND_FORWARD) {
+ err = mlxsw_afa_block_append_trap_and_forward(afa_block,
+ MLXSW_TRAP_ID_ACL2);
+ if (err)
+ goto err;
+ }
break;
default:
err = -EINVAL;
--
2.9.5
^ permalink raw reply related
* [patch net-next v2 5/7] mlxsw: spectrum: Add trap for multicast trap-and-forward routes
From: Jiri Pirko @ 2017-10-03 7:58 UTC (permalink / raw)
To: netdev
Cc: davem, yotamg, idosch, mlxsw, nikolay, andrew, dsa, edumazet,
willemb, johannes.berg, dcaratti, pabeni, daniel, f.fainelli, fw,
gfree.wind
In-Reply-To: <20171003075812.1540-1-jiri@resnulli.us>
From: Yotam Gigi <yotamg@mellanox.com>
When a multicast route is configured with trap-and-forward action, the
packets should be marked with skb->offload_mr_fwd_mark, in order to prevent
the packets from being forwarded again by the kernel ipmr module.
Due to this, it is not possible to use the already existing multicast trap
(MLXSW_TRAP_ID_ACL1) as the packet should be marked differently. Add the
MLXSW_TRAP_ID_ACL2 which is for trap-and-forward multicast routes, and set
the offload_mr_fwd_mark skb field in its handler.
Signed-off-by: Yotam Gigi <yotamg@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 13 +++++++++++++
drivers/net/ethernet/mellanox/mlxsw/trap.h | 2 ++
2 files changed, 15 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index e9b9443..3adf237 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -3312,6 +3312,14 @@ static void mlxsw_sp_rx_listener_mark_func(struct sk_buff *skb, u8 local_port,
return mlxsw_sp_rx_listener_no_mark_func(skb, local_port, priv);
}
+static void mlxsw_sp_rx_listener_mr_mark_func(struct sk_buff *skb,
+ u8 local_port, void *priv)
+{
+ skb->offload_mr_fwd_mark = 1;
+ skb->offload_fwd_mark = 1;
+ return mlxsw_sp_rx_listener_no_mark_func(skb, local_port, priv);
+}
+
static void mlxsw_sp_rx_listener_sample_func(struct sk_buff *skb, u8 local_port,
void *priv)
{
@@ -3355,6 +3363,10 @@ static void mlxsw_sp_rx_listener_sample_func(struct sk_buff *skb, u8 local_port,
MLXSW_RXL(mlxsw_sp_rx_listener_mark_func, _trap_id, _action, \
_is_ctrl, SP_##_trap_group, DISCARD)
+#define MLXSW_SP_RXL_MR_MARK(_trap_id, _action, _trap_group, _is_ctrl) \
+ MLXSW_RXL(mlxsw_sp_rx_listener_mr_mark_func, _trap_id, _action, \
+ _is_ctrl, SP_##_trap_group, DISCARD)
+
#define MLXSW_SP_EVENTL(_func, _trap_id) \
MLXSW_EVENTL(_func, _trap_id, SP_EVENT)
@@ -3425,6 +3437,7 @@ static const struct mlxsw_listener mlxsw_sp_listener[] = {
MLXSW_SP_RXL_MARK(IPV4_PIM, TRAP_TO_CPU, PIM, false),
MLXSW_SP_RXL_MARK(RPF, TRAP_TO_CPU, RPF, false),
MLXSW_SP_RXL_MARK(ACL1, TRAP_TO_CPU, MULTICAST, false),
+ MLXSW_SP_RXL_MR_MARK(ACL2, TRAP_TO_CPU, MULTICAST, false),
};
static int mlxsw_sp_cpu_policers_set(struct mlxsw_core *mlxsw_core)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/trap.h b/drivers/net/ethernet/mellanox/mlxsw/trap.h
index a981035..ec6cef8 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/trap.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/trap.h
@@ -93,6 +93,8 @@ enum {
MLXSW_TRAP_ID_ACL0 = 0x1C0,
/* Multicast trap used for routes with trap action */
MLXSW_TRAP_ID_ACL1 = 0x1C1,
+ /* Multicast trap used for routes with trap-and-forward action */
+ MLXSW_TRAP_ID_ACL2 = 0x1C2,
MLXSW_TRAP_ID_MAX = 0x1FF
};
--
2.9.5
^ permalink raw reply related
* [patch net-next v2 7/7] mlxsw: spectrum: mr: Support trap-and-forward routes
From: Jiri Pirko @ 2017-10-03 7:58 UTC (permalink / raw)
To: netdev
Cc: davem, yotamg, idosch, mlxsw, nikolay, andrew, dsa, edumazet,
willemb, johannes.berg, dcaratti, pabeni, daniel, f.fainelli, fw,
gfree.wind
In-Reply-To: <20171003075812.1540-1-jiri@resnulli.us>
From: Yotam Gigi <yotamg@mellanox.com>
Add the support of trap-and-forward route action in the multicast routing
offloading logic. A route will be set to trap-and-forward action if one (or
more) of its output interfaces is not offload-able, i.e. does not have a
valid Spectrum RIF.
This way, a route with mixed output VIFs list, which contains both
offload-able and un-offload-able devices can go through partial offloading
in hardware, and the rest will be done in the kernel ipmr module.
Signed-off-by: Yotam Gigi <yotamg@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.c | 17 ++++++++---------
1 file changed, 8 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.c
index 4aaf6ca..1f84bb8 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_mr.c
@@ -114,9 +114,9 @@ static bool mlxsw_sp_mr_vif_valid(const struct mlxsw_sp_mr_vif *vif)
return mlxsw_sp_mr_vif_regular(vif) && vif->dev && vif->rif;
}
-static bool mlxsw_sp_mr_vif_rif_invalid(const struct mlxsw_sp_mr_vif *vif)
+static bool mlxsw_sp_mr_vif_exists(const struct mlxsw_sp_mr_vif *vif)
{
- return mlxsw_sp_mr_vif_regular(vif) && vif->dev && !vif->rif;
+ return vif->dev;
}
static bool
@@ -182,14 +182,13 @@ mlxsw_sp_mr_route_action(const struct mlxsw_sp_mr_route *mr_route)
if (!mlxsw_sp_mr_route_valid_evifs_num(mr_route))
return MLXSW_SP_MR_ROUTE_ACTION_TRAP;
- /* If either one of the eVIFs is not regular (VIF of type pimreg or
- * tunnel) or one of the VIFs has no matching RIF, trap the packet.
+ /* If one of the eVIFs has no RIF, trap-and-forward the route as there
+ * is some more routing to do in software too.
*/
- list_for_each_entry(rve, &mr_route->evif_list, route_node) {
- if (!mlxsw_sp_mr_vif_regular(rve->mr_vif) ||
- mlxsw_sp_mr_vif_rif_invalid(rve->mr_vif))
- return MLXSW_SP_MR_ROUTE_ACTION_TRAP;
- }
+ list_for_each_entry(rve, &mr_route->evif_list, route_node)
+ if (mlxsw_sp_mr_vif_exists(rve->mr_vif) && !rve->mr_vif->rif)
+ return MLXSW_SP_MR_ROUTE_ACTION_TRAP_AND_FORWARD;
+
return MLXSW_SP_MR_ROUTE_ACTION_FORWARD;
}
--
2.9.5
^ permalink raw reply related
* Re: [PATCH 05/18] net: use ARRAY_SIZE
From: Andy Shevchenko @ 2017-10-03 8:09 UTC (permalink / raw)
To: Jérémy Lefaure
Cc: Sathya Perla, Ajit Khaparde, Sriharsha Basavapatna, Somnath Kotur,
Jeff Kirsher, Arend van Spriel, Franky Lin, Hante Meuleman,
Chi-Hsien Lin, Wright Feng, Kalle Valo, Larry Finger, Chaoming Li,
David S. Miller, Alexey Kuznetsov, Hideaki YOSHIFUJI, netdev,
"linux-kernel@vger.kernel.org" <
In-Reply-To: <20171002212214.30d7013a@blatinox-laptop.localdomain>
On Tue, Oct 3, 2017 at 4:22 AM, Jérémy Lefaure
<jeremy.lefaure@lse.epita.fr> wrote:
> On Mon, 2 Oct 2017 16:07:36 +0300
> Andy Shevchenko <andy.shevchenko@gmail.com> wrote:
>
>> > + {&gainctrl_lut_core0_rev0, ARRAY_SIZE(gainctrl_lut_core0_rev0), 26, 192,
>> > + 32},
>>
>> For all such cases I would rather put on one line disregard checkpatch
>> warning for better readability.
> I agree that it would be better. I didn't know that it was possible to
> not follow this rule for anything else than a string.
IMO, it increases readability quite enough to overrule checkpatch recomendation.
--
With Best Regards,
Andy Shevchenko
^ permalink raw reply
* Re: [v2] cdc-ether: divorce initialisation with a filter reset and a generic method
From: Vignesh R @ 2017-10-03 8:09 UTC (permalink / raw)
To: davem; +Cc: Oliver Neukum, netdev, bjorn
In-Reply-To: <20170522125030.21266-1-oneukum@suse.com>
Hi Dave,
On Monday 22 May 2017 06:20 PM, Oliver Neukum wrote:
> Some devices need their multicast filter reset but others are crashed by that.
> So the methods need to be separated.
>
> Signed-off-by: Oliver Neukum <oneukum@suse.com>
> Reported-by: "Ridgway, Keith" <kridgway@harris.com>
> ---
I see this patch was merged and said to be queued for -stable, but I
don't see this fix in any of the stable kernels. Would you please ask
for this patch to be backported to -stable kernels?
Regards
Vignesh
> drivers/net/usb/cdc_ether.c | 31 ++++++++++++++++++++++++-------
> include/linux/usb/usbnet.h | 1 +
> 2 files changed, 25 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/net/usb/cdc_ether.c b/drivers/net/usb/cdc_ether.c
> index f3ae88fdf332..8ab281b478f2 100644
> --- a/drivers/net/usb/cdc_ether.c
> +++ b/drivers/net/usb/cdc_ether.c
> @@ -310,6 +310,26 @@ int usbnet_generic_cdc_bind(struct usbnet *dev, struct usb_interface *intf)
> return -ENODEV;
> }
>
> + return 0;
> +
> +bad_desc:
> + dev_info(&dev->udev->dev, "bad CDC descriptors\n");
> + return -ENODEV;
> +}
> +EXPORT_SYMBOL_GPL(usbnet_generic_cdc_bind);
> +
> +
> +/* like usbnet_generic_cdc_bind() but handles filter initialization
> + * correctly
> + */
> +int usbnet_ether_cdc_bind(struct usbnet *dev, struct usb_interface *intf)
> +{
> + int rv;
> +
> + rv = usbnet_generic_cdc_bind(dev, intf);
> + if (rv < 0)
> + goto bail_out;
> +
> /* Some devices don't initialise properly. In particular
> * the packet filter is not reset. There are devices that
> * don't do reset all the way. So the packet filter should
> @@ -317,13 +337,10 @@ int usbnet_generic_cdc_bind(struct usbnet *dev, struct usb_interface *intf)
> */
> usbnet_cdc_update_filter(dev);
>
> - return 0;
> -
> -bad_desc:
> - dev_info(&dev->udev->dev, "bad CDC descriptors\n");
> - return -ENODEV;
> +bail_out:
> + return rv;
> }
> -EXPORT_SYMBOL_GPL(usbnet_generic_cdc_bind);
> +EXPORT_SYMBOL_GPL(usbnet_ether_cdc_bind);
>
> void usbnet_cdc_unbind(struct usbnet *dev, struct usb_interface *intf)
> {
> @@ -417,7 +434,7 @@ int usbnet_cdc_bind(struct usbnet *dev, struct usb_interface *intf)
> BUILD_BUG_ON((sizeof(((struct usbnet *)0)->data)
> < sizeof(struct cdc_state)));
>
> - status = usbnet_generic_cdc_bind(dev, intf);
> + status = usbnet_ether_cdc_bind(dev, intf);
> if (status < 0)
> return status;
>
> diff --git a/include/linux/usb/usbnet.h b/include/linux/usb/usbnet.h
> index 7dffa5624ea6..97116379db5f 100644
> --- a/include/linux/usb/usbnet.h
> +++ b/include/linux/usb/usbnet.h
> @@ -206,6 +206,7 @@ struct cdc_state {
> };
>
> extern int usbnet_generic_cdc_bind(struct usbnet *, struct usb_interface *);
> +extern int usbnet_ether_cdc_bind(struct usbnet *dev, struct usb_interface *intf);
> extern int usbnet_cdc_bind(struct usbnet *, struct usb_interface *);
> extern void usbnet_cdc_unbind(struct usbnet *, struct usb_interface *);
> extern void usbnet_cdc_status(struct usbnet *, struct urb *);
>
--
Regards
Vignesh
^ permalink raw reply
* (unknown),
From: morice.diane @ 2017-10-03 8:16 UTC (permalink / raw)
To: netdev
[-- Attachment #1: 747452.zip --]
[-- Type: application/zip, Size: 7300 bytes --]
^ permalink raw reply
* RE: [PATCH] fsl/fman: remove of_node
From: Madalin-cristian Bucur @ 2017-10-03 8:49 UTC (permalink / raw)
To: David Miller
Cc: netdev@vger.kernel.org, andrew@lunn.ch, f.fainelli@gmail.com,
linux-kernel@vger.kernel.org
In-Reply-To: <20171002.160445.737065995396076298.davem@davemloft.net>
> -----Original Message-----
> From: David Miller [mailto:davem@davemloft.net]
> Sent: Tuesday, October 03, 2017 2:05 AM
> To: Madalin-cristian Bucur <madalin.bucur@nxp.com>
> Subject: Re: [PATCH] fsl/fman: remove of_node
>
> From: Madalin Bucur <madalin.bucur@nxp.com>
> Date: Mon, 2 Oct 2017 13:31:37 +0300
>
> > The FMan MAC driver allocates a platform device for the Ethernet
> > driver to probe on. Setting pdev->dev.of_node with the MAC node
> > triggers the MAC driver probing of the new platform device. While
> > this fails quickly and does not affect the functionality of the
> > drivers, it is incorrect and must be removed. This was added to
> > address a report that DSA code using of_find_net_device_by_node()
> > is unable to use the DPAA interfaces. Error message seen before
> > this fix:
> >
> > fsl_mac dpaa-ethernet.0: __devm_request_mem_region(mac) failed
> > fsl_mac: probe of dpaa-ethernet.0 failed with error -16
> >
> > Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
>
> Is the DSA issue no longer something we need to be concerned
> about? If not, why? You have to explain this.
My patch removes the of_node that was set to a device that was not an
of_device, preventing duplicated probing of both the real of_device
and the "fake" one created through this assignment.
I understand that the DSA issue that triggered the initial change
was related to DSA finding the network devices using
of_find_net_device_by_node(), something that will not work for the
DPAA case where the netdevice does not have an of_node. I do not know
enough about DSA to come up with a solution for this problem now.
Andrew, Florian, can you please comment on this?
Thanks,
Madalin
^ permalink raw reply
* [PATCH v2 net-next 00/12] qed: Add iWARP support for unaligned MPA packets
From: Michal Kalderon @ 2017-10-03 8:54 UTC (permalink / raw)
To: davem-fT/PcQaiUtIeIZ0/mPfg9Q
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
dledford-H+wXaHxf7aLQT0dZR+AlfA, Michal Kalderon, Ariel Elior
This patch series adds support for handling unaligned MPA packets.
(FPDUs split over more than one tcp packet).
When FW detects a packet is unaligned it fowards the packet to
the driver via a light l2 dedicated connection. The driver then
stores this packet until the remainder of the packet is received.
Once the driver reconstructs the full FPDU, it sends it down
to fw via the ll2 connection. Driver also breaks down any packed
PDUs into separate packets for FW.
Patches 1-6 are all slight modifications to ll2 to support additional
requirements for the unaligned MPA ll2 client.
Patch 7 opens the additional ll2 connection for iWARP.
Patches 8-12 contain the algorithm for aligning packets.
Signed-off-by: Michal Kalderon <Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
Changes from v1:
patch #1: tx descq_array changed to void pointer and name
changed accordingly to descq_mem.
patch #2: modify boolean variables to be of type bool.
Michal Kalderon (12):
qed: Add ll2 option to limit the number of bds per packet
qed: Add ll2 ability of opening a secondary queue
qed: Add ll2 option for dropping a tx packet
qed: Fix initialization of ll2 offload feature
qed: Add the source of a packet sent on an iWARP ll2 connection
qed: Add LL2 slowpath handling
qed: Add ll2 connection for processing unaligned MPA packets
qed: Add mpa buffer descriptors for storing and processing mpa fpdus
qed: Add unaligned and packed packet processing
qed: Add support for freeing two ll2 buffers for corner cases
qed: Add support for MPA header being split over two tcp packets
qed: Add iWARP support for fpdu spanned over more than two tcp packets
drivers/net/ethernet/qlogic/qed/qed_iwarp.c | 690 ++++++++++++++++++++++++++++
drivers/net/ethernet/qlogic/qed/qed_iwarp.h | 33 ++
drivers/net/ethernet/qlogic/qed/qed_ll2.c | 102 +++-
drivers/net/ethernet/qlogic/qed/qed_ll2.h | 10 +-
include/linux/qed/qed_ll2_if.h | 7 +
5 files changed, 822 insertions(+), 20 deletions(-)
--
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
* [PATCH v2 net-next 02/12] qed: Add ll2 ability of opening a secondary queue
From: Michal Kalderon @ 2017-10-03 8:54 UTC (permalink / raw)
To: davem-fT/PcQaiUtIeIZ0/mPfg9Q
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
dledford-H+wXaHxf7aLQT0dZR+AlfA, Michal Kalderon, Ariel Elior
In-Reply-To: <1507020902-4952-1-git-send-email-Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
When more than one ll2 queue is opened ( that is not an OOO queue )
ll2 code does not have enough information to determine whether
the queue is the main one or not, so a new field is added to the
acquire input data to expose the control of determining whether
the queue is the main queue or a secondary queue.
Signed-off-by: Michal Kalderon <Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
drivers/net/ethernet/qlogic/qed/qed_ll2.c | 7 ++++++-
drivers/net/ethernet/qlogic/qed/qed_ll2.h | 1 +
include/linux/qed/qed_ll2_if.h | 1 +
3 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.c b/drivers/net/ethernet/qlogic/qed/qed_ll2.c
index 75af40a..3c695da 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_ll2.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.c
@@ -894,7 +894,7 @@ static int qed_sp_ll2_rx_queue_start(struct qed_hwfn *p_hwfn,
p_ramrod->drop_ttl0_flg = p_ll2_conn->input.rx_drop_ttl0_flg;
p_ramrod->inner_vlan_removal_en = p_ll2_conn->input.rx_vlan_removal_en;
p_ramrod->queue_id = p_ll2_conn->queue_id;
- p_ramrod->main_func_queue = (conn_type == QED_LL2_TYPE_OOO) ? 0 : 1;
+ p_ramrod->main_func_queue = p_ll2_conn->main_func_queue ? 1 : 0;
if ((IS_MF_DEFAULT(p_hwfn) || IS_MF_SI(p_hwfn)) &&
p_ramrod->main_func_queue && (conn_type != QED_LL2_TYPE_ROCE) &&
@@ -1265,6 +1265,11 @@ int qed_ll2_acquire_connection(void *cxt, struct qed_ll2_acquire_data *data)
p_ll2_info->tx_dest = (data->input.tx_dest == QED_LL2_TX_DEST_NW) ?
CORE_TX_DEST_NW : CORE_TX_DEST_LB;
+ if (data->input.conn_type == QED_LL2_TYPE_OOO ||
+ data->input.secondary_queue)
+ p_ll2_info->main_func_queue = false;
+ else
+ p_ll2_info->main_func_queue = true;
/* Correct maximum number of Tx BDs */
p_tx_max = &p_ll2_info->input.tx_max_bds_per_packet;
diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.h b/drivers/net/ethernet/qlogic/qed/qed_ll2.h
index 9bdd08f..f658170 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_ll2.h
+++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.h
@@ -121,6 +121,7 @@ struct qed_ll2_info {
bool b_active;
enum core_tx_dest tx_dest;
u8 tx_stats_en;
+ bool main_func_queue;
struct qed_ll2_rx_queue rx_queue;
struct qed_ll2_tx_queue tx_queue;
struct qed_ll2_cbs cbs;
diff --git a/include/linux/qed/qed_ll2_if.h b/include/linux/qed/qed_ll2_if.h
index 89fa0bb..d7cca59 100644
--- a/include/linux/qed/qed_ll2_if.h
+++ b/include/linux/qed/qed_ll2_if.h
@@ -171,6 +171,7 @@ struct qed_ll2_acquire_data_inputs {
enum qed_ll2_tx_dest tx_dest;
enum qed_ll2_error_handle ai_err_packet_too_big;
enum qed_ll2_error_handle ai_err_no_buf;
+ bool secondary_queue;
u8 gsi_enable;
};
--
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
* [PATCH v2 net-next 04/12] qed: Fix initialization of ll2 offload feature
From: Michal Kalderon @ 2017-10-03 8:54 UTC (permalink / raw)
To: davem-fT/PcQaiUtIeIZ0/mPfg9Q
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
dledford-H+wXaHxf7aLQT0dZR+AlfA, Michal Kalderon, Ariel Elior
In-Reply-To: <1507020902-4952-1-git-send-email-Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
enable_ip_cksum, enable_l4_cksum, calc_ip_len were added in
commit stated below but not passed through to FW. This was OK
until now as it wasn't used, but is required for the iWARP
unaligned flow
Fixes:7c7973b2ae27 ("qed: LL2 to use packed information for tx")
Signed-off-by: Michal Kalderon <Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
drivers/net/ethernet/qlogic/qed/qed_ll2.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.c b/drivers/net/ethernet/qlogic/qed/qed_ll2.c
index ad67d36..6d14474 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_ll2.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.c
@@ -1621,6 +1621,9 @@ static void qed_ll2_prepare_tx_packet_set(struct qed_hwfn *p_hwfn,
SET_FIELD(bd_data, CORE_TX_BD_DATA_START_BD, 0x1);
SET_FIELD(bd_data, CORE_TX_BD_DATA_NBDS, pkt->num_of_bds);
SET_FIELD(bd_data, CORE_TX_BD_DATA_ROCE_FLAV, roce_flavor);
+ SET_FIELD(bd_data, CORE_TX_BD_DATA_IP_CSUM, !!(pkt->enable_ip_cksum));
+ SET_FIELD(bd_data, CORE_TX_BD_DATA_L4_CSUM, !!(pkt->enable_l4_cksum));
+ SET_FIELD(bd_data, CORE_TX_BD_DATA_IP_LEN, !!(pkt->calc_ip_len));
start_bd->bd_data.as_bitfield = cpu_to_le16(bd_data);
DMA_REGPAIR_LE(start_bd->addr, pkt->first_frag);
start_bd->nbytes = cpu_to_le16(pkt->first_frag_len);
--
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
* [PATCH v2 net-next 07/12] qed: Add ll2 connection for processing unaligned MPA packets
From: Michal Kalderon @ 2017-10-03 8:54 UTC (permalink / raw)
To: davem-fT/PcQaiUtIeIZ0/mPfg9Q
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
dledford-H+wXaHxf7aLQT0dZR+AlfA, Michal Kalderon, Ariel Elior
In-Reply-To: <1507020902-4952-1-git-send-email-Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
This patch adds only the establishment and termination of the
ll2 connection that handles unaligned MPA packets.
Signed-off-by: Michal Kalderon <Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
drivers/net/ethernet/qlogic/qed/qed_iwarp.c | 65 +++++++++++++++++++++++++++++
drivers/net/ethernet/qlogic/qed/qed_iwarp.h | 1 +
2 files changed, 66 insertions(+)
diff --git a/drivers/net/ethernet/qlogic/qed/qed_iwarp.c b/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
index 8fc9c811..f413621 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
@@ -1713,6 +1713,19 @@ int qed_iwarp_reject(void *rdma_cxt, struct qed_iwarp_reject_in *iparams)
return 0;
}
+/* fpdu can be fragmented over maximum 3 bds: header, partial mpa, unaligned */
+#define QED_IWARP_MAX_BDS_PER_FPDU 3
+static void
+qed_iwarp_ll2_comp_mpa_pkt(void *cxt, struct qed_ll2_comp_rx_data *data)
+{
+ struct qed_iwarp_info *iwarp_info;
+ struct qed_hwfn *p_hwfn = cxt;
+
+ iwarp_info = &p_hwfn->p_rdma_info->iwarp;
+ qed_iwarp_ll2_post_rx(p_hwfn, data->cookie,
+ iwarp_info->ll2_mpa_handle);
+}
+
static void
qed_iwarp_ll2_comp_syn_pkt(void *cxt, struct qed_ll2_comp_rx_data *data)
{
@@ -1877,6 +1890,13 @@ static void qed_iwarp_ll2_rel_tx_pkt(void *cxt, u8 connection_handle,
kfree(buffer);
}
+void
+qed_iwarp_ll2_slowpath(void *cxt,
+ u8 connection_handle,
+ u32 opaque_data_0, u32 opaque_data_1)
+{
+}
+
static int qed_iwarp_ll2_stop(struct qed_hwfn *p_hwfn, struct qed_ptt *p_ptt)
{
struct qed_iwarp_info *iwarp_info = &p_hwfn->p_rdma_info->iwarp;
@@ -1902,6 +1922,16 @@ static int qed_iwarp_ll2_stop(struct qed_hwfn *p_hwfn, struct qed_ptt *p_ptt)
iwarp_info->ll2_ooo_handle = QED_IWARP_HANDLE_INVAL;
}
+ if (iwarp_info->ll2_mpa_handle != QED_IWARP_HANDLE_INVAL) {
+ rc = qed_ll2_terminate_connection(p_hwfn,
+ iwarp_info->ll2_mpa_handle);
+ if (rc)
+ DP_INFO(p_hwfn, "Failed to terminate mpa connection\n");
+
+ qed_ll2_release_connection(p_hwfn, iwarp_info->ll2_mpa_handle);
+ iwarp_info->ll2_mpa_handle = QED_IWARP_HANDLE_INVAL;
+ }
+
qed_llh_remove_mac_filter(p_hwfn,
p_ptt, p_hwfn->p_rdma_info->iwarp.mac_addr);
return rc;
@@ -1953,12 +1983,14 @@ static int qed_iwarp_ll2_stop(struct qed_hwfn *p_hwfn, struct qed_ptt *p_ptt)
struct qed_iwarp_info *iwarp_info;
struct qed_ll2_acquire_data data;
struct qed_ll2_cbs cbs;
+ u32 mpa_buff_size;
u16 n_ooo_bufs;
int rc = 0;
iwarp_info = &p_hwfn->p_rdma_info->iwarp;
iwarp_info->ll2_syn_handle = QED_IWARP_HANDLE_INVAL;
iwarp_info->ll2_ooo_handle = QED_IWARP_HANDLE_INVAL;
+ iwarp_info->ll2_mpa_handle = QED_IWARP_HANDLE_INVAL;
iwarp_info->max_mtu = params->max_mtu;
@@ -2029,6 +2061,39 @@ static int qed_iwarp_ll2_stop(struct qed_hwfn *p_hwfn, struct qed_ptt *p_ptt)
if (rc)
goto err;
+ /* Start Unaligned MPA connection */
+ cbs.rx_comp_cb = qed_iwarp_ll2_comp_mpa_pkt;
+ cbs.slowpath_cb = qed_iwarp_ll2_slowpath;
+
+ memset(&data, 0, sizeof(data));
+ data.input.conn_type = QED_LL2_TYPE_IWARP;
+ data.input.mtu = params->max_mtu;
+ /* FW requires that once a packet arrives OOO, it must have at
+ * least 2 rx buffers available on the unaligned connection
+ * for handling the case that it is a partial fpdu.
+ */
+ data.input.rx_num_desc = n_ooo_bufs * 2;
+ data.input.tx_num_desc = data.input.rx_num_desc;
+ data.input.tx_max_bds_per_packet = QED_IWARP_MAX_BDS_PER_FPDU;
+ data.p_connection_handle = &iwarp_info->ll2_mpa_handle;
+ data.input.secondary_queue = true;
+ data.cbs = &cbs;
+
+ rc = qed_ll2_acquire_connection(p_hwfn, &data);
+ if (rc)
+ goto err;
+
+ rc = qed_ll2_establish_connection(p_hwfn, iwarp_info->ll2_mpa_handle);
+ if (rc)
+ goto err;
+
+ mpa_buff_size = QED_IWARP_MAX_BUF_SIZE(params->max_mtu);
+ rc = qed_iwarp_ll2_alloc_buffers(p_hwfn,
+ data.input.rx_num_desc,
+ mpa_buff_size,
+ iwarp_info->ll2_mpa_handle);
+ if (rc)
+ goto err;
return rc;
err:
qed_iwarp_ll2_stop(p_hwfn, p_ptt);
diff --git a/drivers/net/ethernet/qlogic/qed/qed_iwarp.h b/drivers/net/ethernet/qlogic/qed/qed_iwarp.h
index 9e2bfde..9d33a1f 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_iwarp.h
+++ b/drivers/net/ethernet/qlogic/qed/qed_iwarp.h
@@ -73,6 +73,7 @@ struct qed_iwarp_info {
u8 tcp_flags;
u8 ll2_syn_handle;
u8 ll2_ooo_handle;
+ u8 ll2_mpa_handle;
u8 peer2peer;
enum mpa_negotiation_mode mpa_rev;
enum mpa_rtr_type rtr_type;
--
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
* [PATCH v2 net-next 10/12] qed: Add support for freeing two ll2 buffers for corner cases
From: Michal Kalderon @ 2017-10-03 8:55 UTC (permalink / raw)
To: davem-fT/PcQaiUtIeIZ0/mPfg9Q
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
dledford-H+wXaHxf7aLQT0dZR+AlfA, Michal Kalderon, Ariel Elior
In-Reply-To: <1507020902-4952-1-git-send-email-Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
When posting a packet on the ll2 tx, we can provide a cookie that
will be returned upon tx completion. This cookie is the ll2 iwarp buffer
which is then reposted to the rx ring. Part of the unaligned mpa flow
is determining when a buffer can be reposted. Each buffer needs to be
sent only once as a cookie for on the tx ring. In packed fpdu case, only
the last packet will be sent with the buffer, meaning we need to handle the
case that a cookie can be NULL on tx complete. In addition, when a fpdu
splits over two buffers, but there are no more fpdus on the second buffer,
two buffers need to be provided as a cookie. To avoid changing the ll2
interface to provide two cookies, we introduce a piggy buf pointer,
relevant for iWARP only, that holds a pointer to a second buffer that
needs to be released during tx completion.
Signed-off-by: Michal Kalderon <Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
drivers/net/ethernet/qlogic/qed/qed_iwarp.c | 25 +++++++++++++++++++++++++
drivers/net/ethernet/qlogic/qed/qed_iwarp.h | 1 +
2 files changed, 26 insertions(+)
diff --git a/drivers/net/ethernet/qlogic/qed/qed_iwarp.c b/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
index 83b147f..8b17369 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
@@ -1846,6 +1846,12 @@ enum qed_iwarp_mpa_pkt_type {
/* vlan overload with enum iwarp_ll2_tx_queues */
tx_pkt.vlan = IWARP_LL2_ALIGNED_TX_QUEUE;
+ /* special case of unaligned packet and not packed, need to send
+ * both buffers as cookie to release.
+ */
+ if (tcp_payload_size == fpdu->incomplete_bytes)
+ fpdu->mpa_buf->piggy_buf = buf;
+
ll2_handle = p_hwfn->p_rdma_info->iwarp.ll2_mpa_handle;
/* Set first fragment to header */
@@ -2195,9 +2201,19 @@ static void qed_iwarp_ll2_comp_tx_pkt(void *cxt, u8 connection_handle,
bool b_last_fragment, bool b_last_packet)
{
struct qed_iwarp_ll2_buff *buffer = cookie;
+ struct qed_iwarp_ll2_buff *piggy;
struct qed_hwfn *p_hwfn = cxt;
+ if (!buffer) /* can happen in packed mpa unaligned... */
+ return;
+
/* this was originally an rx packet, post it back */
+ piggy = buffer->piggy_buf;
+ if (piggy) {
+ buffer->piggy_buf = NULL;
+ qed_iwarp_ll2_post_rx(p_hwfn, piggy, connection_handle);
+ }
+
qed_iwarp_ll2_post_rx(p_hwfn, buffer, connection_handle);
if (connection_handle == p_hwfn->p_rdma_info->iwarp.ll2_mpa_handle)
@@ -2216,6 +2232,15 @@ static void qed_iwarp_ll2_rel_tx_pkt(void *cxt, u8 connection_handle,
if (!buffer)
return;
+ if (buffer->piggy_buf) {
+ dma_free_coherent(&p_hwfn->cdev->pdev->dev,
+ buffer->piggy_buf->buff_size,
+ buffer->piggy_buf->data,
+ buffer->piggy_buf->data_phys_addr);
+
+ kfree(buffer->piggy_buf);
+ }
+
dma_free_coherent(&p_hwfn->cdev->pdev->dev, buffer->buff_size,
buffer->data, buffer->data_phys_addr);
diff --git a/drivers/net/ethernet/qlogic/qed/qed_iwarp.h b/drivers/net/ethernet/qlogic/qed/qed_iwarp.h
index 858755c..58db51a 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_iwarp.h
+++ b/drivers/net/ethernet/qlogic/qed/qed_iwarp.h
@@ -55,6 +55,7 @@ enum qed_iwarp_qp_state {
#define QED_IWARP_HANDLE_INVAL (0xff)
struct qed_iwarp_ll2_buff {
+ struct qed_iwarp_ll2_buff *piggy_buf;
void *data;
dma_addr_t data_phys_addr;
u32 buff_size;
--
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
* [PATCH v2 net-next 12/12] qed: Add iWARP support for fpdu spanned over more than two tcp packets
From: Michal Kalderon @ 2017-10-03 8:55 UTC (permalink / raw)
To: davem-fT/PcQaiUtIeIZ0/mPfg9Q
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
dledford-H+wXaHxf7aLQT0dZR+AlfA, Michal Kalderon, Ariel Elior
In-Reply-To: <1507020902-4952-1-git-send-email-Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
We continue to maintain a maximum of three buffers per fpdu, to ensure
that there are enough buffers for additional unaligned mpa packets.
To support this, if a fpdu is split over more than two tcp packets, we
use an intermediate buffer to copy the data to the previous buffer, then
we can release the data. We need an intermediate buffer as the initial
buffer partial packet could be located at the end of the packet, not
leaving room for additional data. This is a corner case, and will usually
not be the case.
Signed-off-by: Michal Kalderon <Michal.Kalderon-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
Signed-off-by: Ariel Elior <Ariel.Elior-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
---
drivers/net/ethernet/qlogic/qed/qed_iwarp.c | 193 ++++++++++++++++++++++++++++
drivers/net/ethernet/qlogic/qed/qed_iwarp.h | 1 +
2 files changed, 194 insertions(+)
diff --git a/drivers/net/ethernet/qlogic/qed/qed_iwarp.c b/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
index 2994942..b2b1f87 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_iwarp.c
@@ -1420,6 +1420,7 @@ void qed_iwarp_resc_free(struct qed_hwfn *p_hwfn)
qed_rdma_bmap_free(p_hwfn, &p_hwfn->p_rdma_info->tcp_cid_map, 1);
kfree(iwarp_info->mpa_bufs);
kfree(iwarp_info->partial_fpdus);
+ kfree(iwarp_info->mpa_intermediate_buf);
}
int qed_iwarp_accept(void *rdma_cxt, struct qed_iwarp_accept_in *iparams)
@@ -1762,6 +1763,11 @@ enum qed_iwarp_mpa_pkt_type {
"QED_IWARP_MPA_PKT_UNALIGNED"
};
+static int
+qed_iwarp_recycle_pkt(struct qed_hwfn *p_hwfn,
+ struct qed_iwarp_fpdu *fpdu,
+ struct qed_iwarp_ll2_buff *buf);
+
static enum qed_iwarp_mpa_pkt_type
qed_iwarp_mpa_classify(struct qed_hwfn *p_hwfn,
struct qed_iwarp_fpdu *fpdu,
@@ -1822,6 +1828,68 @@ enum qed_iwarp_mpa_pkt_type {
fpdu->mpa_frag_len = fpdu->fpdu_length - fpdu->incomplete_bytes;
}
+static int
+qed_iwarp_cp_pkt(struct qed_hwfn *p_hwfn,
+ struct qed_iwarp_fpdu *fpdu,
+ struct unaligned_opaque_data *pkt_data,
+ struct qed_iwarp_ll2_buff *buf, u16 tcp_payload_size)
+{
+ u8 *tmp_buf = p_hwfn->p_rdma_info->iwarp.mpa_intermediate_buf;
+ int rc;
+
+ /* need to copy the data from the partial packet stored in fpdu
+ * to the new buf, for this we also need to move the data currently
+ * placed on the buf. The assumption is that the buffer is big enough
+ * since fpdu_length <= mss, we use an intermediate buffer since
+ * we may need to copy the new data to an overlapping location
+ */
+ if ((fpdu->mpa_frag_len + tcp_payload_size) > (u16)buf->buff_size) {
+ DP_ERR(p_hwfn,
+ "MPA ALIGN: Unexpected: buffer is not large enough for split fpdu buff_size = %d mpa_frag_len = %d, tcp_payload_size = %d, incomplete_bytes = %d\n",
+ buf->buff_size, fpdu->mpa_frag_len,
+ tcp_payload_size, fpdu->incomplete_bytes);
+ return -EINVAL;
+ }
+
+ DP_VERBOSE(p_hwfn, QED_MSG_RDMA,
+ "MPA ALIGN Copying fpdu: [%p, %d] [%p, %d]\n",
+ fpdu->mpa_frag_virt, fpdu->mpa_frag_len,
+ (u8 *)(buf->data) + pkt_data->first_mpa_offset,
+ tcp_payload_size);
+
+ memcpy(tmp_buf, fpdu->mpa_frag_virt, fpdu->mpa_frag_len);
+ memcpy(tmp_buf + fpdu->mpa_frag_len,
+ (u8 *)(buf->data) + pkt_data->first_mpa_offset,
+ tcp_payload_size);
+
+ rc = qed_iwarp_recycle_pkt(p_hwfn, fpdu, fpdu->mpa_buf);
+ if (rc)
+ return rc;
+
+ /* If we managed to post the buffer copy the data to the new buffer
+ * o/w this will occur in the next round...
+ */
+ memcpy((u8 *)(buf->data), tmp_buf,
+ fpdu->mpa_frag_len + tcp_payload_size);
+
+ fpdu->mpa_buf = buf;
+ /* fpdu->pkt_hdr remains as is */
+ /* fpdu->mpa_frag is overridden with new buf */
+ fpdu->mpa_frag = buf->data_phys_addr;
+ fpdu->mpa_frag_virt = buf->data;
+ fpdu->mpa_frag_len += tcp_payload_size;
+
+ fpdu->incomplete_bytes -= tcp_payload_size;
+
+ DP_VERBOSE(p_hwfn,
+ QED_MSG_RDMA,
+ "MPA ALIGN: split fpdu buff_size = %d mpa_frag_len = %d, tcp_payload_size = %d, incomplete_bytes = %d\n",
+ buf->buff_size, fpdu->mpa_frag_len, tcp_payload_size,
+ fpdu->incomplete_bytes);
+
+ return 0;
+}
+
static void
qed_iwarp_update_fpdu_length(struct qed_hwfn *p_hwfn,
struct qed_iwarp_fpdu *fpdu, u8 *mpa_data)
@@ -1843,6 +1911,90 @@ enum qed_iwarp_mpa_pkt_type {
}
}
+#define QED_IWARP_IS_RIGHT_EDGE(_curr_pkt) \
+ (GET_FIELD((_curr_pkt)->flags, \
+ UNALIGNED_OPAQUE_DATA_PKT_REACHED_WIN_RIGHT_EDGE))
+
+/* This function is used to recycle a buffer using the ll2 drop option. It
+ * uses the mechanism to ensure that all buffers posted to tx before this one
+ * were completed. The buffer sent here will be sent as a cookie in the tx
+ * completion function and can then be reposted to rx chain when done. The flow
+ * that requires this is the flow where a FPDU splits over more than 3 tcp
+ * segments. In this case the driver needs to re-post a rx buffer instead of
+ * the one received, but driver can't simply repost a buffer it copied from
+ * as there is a case where the buffer was originally a packed FPDU, and is
+ * partially posted to FW. Driver needs to ensure FW is done with it.
+ */
+static int
+qed_iwarp_recycle_pkt(struct qed_hwfn *p_hwfn,
+ struct qed_iwarp_fpdu *fpdu,
+ struct qed_iwarp_ll2_buff *buf)
+{
+ struct qed_ll2_tx_pkt_info tx_pkt;
+ u8 ll2_handle;
+ int rc;
+
+ memset(&tx_pkt, 0, sizeof(tx_pkt));
+ tx_pkt.num_of_bds = 1;
+ tx_pkt.tx_dest = QED_LL2_TX_DEST_DROP;
+ tx_pkt.l4_hdr_offset_w = fpdu->pkt_hdr_size >> 2;
+ tx_pkt.first_frag = fpdu->pkt_hdr;
+ tx_pkt.first_frag_len = fpdu->pkt_hdr_size;
+ buf->piggy_buf = NULL;
+ tx_pkt.cookie = buf;
+
+ ll2_handle = p_hwfn->p_rdma_info->iwarp.ll2_mpa_handle;
+
+ rc = qed_ll2_prepare_tx_packet(p_hwfn, ll2_handle, &tx_pkt, true);
+ if (rc)
+ DP_VERBOSE(p_hwfn, QED_MSG_RDMA,
+ "Can't drop packet rc=%d\n", rc);
+
+ DP_VERBOSE(p_hwfn,
+ QED_MSG_RDMA,
+ "MPA_ALIGN: send drop tx packet [%lx, 0x%x], buf=%p, rc=%d\n",
+ (unsigned long int)tx_pkt.first_frag,
+ tx_pkt.first_frag_len, buf, rc);
+
+ return rc;
+}
+
+static int
+qed_iwarp_win_right_edge(struct qed_hwfn *p_hwfn, struct qed_iwarp_fpdu *fpdu)
+{
+ struct qed_ll2_tx_pkt_info tx_pkt;
+ u8 ll2_handle;
+ int rc;
+
+ memset(&tx_pkt, 0, sizeof(tx_pkt));
+ tx_pkt.num_of_bds = 1;
+ tx_pkt.tx_dest = QED_LL2_TX_DEST_LB;
+ tx_pkt.l4_hdr_offset_w = fpdu->pkt_hdr_size >> 2;
+
+ tx_pkt.first_frag = fpdu->pkt_hdr;
+ tx_pkt.first_frag_len = fpdu->pkt_hdr_size;
+ tx_pkt.enable_ip_cksum = true;
+ tx_pkt.enable_l4_cksum = true;
+ tx_pkt.calc_ip_len = true;
+ /* vlan overload with enum iwarp_ll2_tx_queues */
+ tx_pkt.vlan = IWARP_LL2_ALIGNED_RIGHT_TRIMMED_TX_QUEUE;
+
+ ll2_handle = p_hwfn->p_rdma_info->iwarp.ll2_mpa_handle;
+
+ rc = qed_ll2_prepare_tx_packet(p_hwfn, ll2_handle, &tx_pkt, true);
+ if (rc)
+ DP_VERBOSE(p_hwfn, QED_MSG_RDMA,
+ "Can't send right edge rc=%d\n", rc);
+ DP_VERBOSE(p_hwfn,
+ QED_MSG_RDMA,
+ "MPA_ALIGN: Sent right edge FPDU num_bds=%d [%lx, 0x%x], rc=%d\n",
+ tx_pkt.num_of_bds,
+ (unsigned long int)tx_pkt.first_frag,
+ tx_pkt.first_frag_len, rc);
+
+ return rc;
+}
+
static int
qed_iwarp_send_fpdu(struct qed_hwfn *p_hwfn,
struct qed_iwarp_fpdu *fpdu,
@@ -1971,6 +2123,20 @@ enum qed_iwarp_mpa_pkt_type {
mpa_buf->tcp_payload_len,
mpa_buf->placement_offset);
+ if (!QED_IWARP_IS_RIGHT_EDGE(curr_pkt)) {
+ mpa_buf->tcp_payload_len = 0;
+ break;
+ }
+
+ rc = qed_iwarp_win_right_edge(p_hwfn, fpdu);
+
+ if (rc) {
+ DP_VERBOSE(p_hwfn, QED_MSG_RDMA,
+ "Can't send FPDU:reset rc=%d\n", rc);
+ memset(fpdu, 0, sizeof(*fpdu));
+ break;
+ }
+
mpa_buf->tcp_payload_len = 0;
break;
case QED_IWARP_MPA_PKT_PACKED:
@@ -1994,6 +2160,28 @@ enum qed_iwarp_mpa_pkt_type {
break;
case QED_IWARP_MPA_PKT_UNALIGNED:
qed_iwarp_update_fpdu_length(p_hwfn, fpdu, mpa_data);
+ if (mpa_buf->tcp_payload_len < fpdu->incomplete_bytes) {
+ /* special handling of fpdu split over more
+ * than 2 segments
+ */
+ if (QED_IWARP_IS_RIGHT_EDGE(curr_pkt)) {
+ rc = qed_iwarp_win_right_edge(p_hwfn,
+ fpdu);
+ /* packet will be re-processed later */
+ if (rc)
+ return rc;
+ }
+
+ rc = qed_iwarp_cp_pkt(p_hwfn, fpdu, curr_pkt,
+ buf,
+ mpa_buf->tcp_payload_len);
+ if (rc) /* packet will be re-processed later */
+ return rc;
+
+ mpa_buf->tcp_payload_len = 0;
+ break;
+ }
+
rc = qed_iwarp_send_fpdu(p_hwfn, fpdu, curr_pkt, buf,
mpa_buf->tcp_payload_len,
pkt_type);
@@ -2510,6 +2698,11 @@ static int qed_iwarp_ll2_stop(struct qed_hwfn *p_hwfn, struct qed_ptt *p_ptt)
goto err;
iwarp_info->max_num_partial_fpdus = (u16)p_hwfn->p_rdma_info->num_qps;
+
+ iwarp_info->mpa_intermediate_buf = kzalloc(mpa_buff_size, GFP_KERNEL);
+ if (!iwarp_info->mpa_intermediate_buf)
+ goto err;
+
/* The mpa_bufs array serves for pending RX packets received on the
* mpa ll2 that don't have place on the tx ring and require later
* processing. We can't fail on allocation of such a struct therefore
diff --git a/drivers/net/ethernet/qlogic/qed/qed_iwarp.h b/drivers/net/ethernet/qlogic/qed/qed_iwarp.h
index c58793a..c1ecd74 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_iwarp.h
+++ b/drivers/net/ethernet/qlogic/qed/qed_iwarp.h
@@ -107,6 +107,7 @@ struct qed_iwarp_info {
enum mpa_rtr_type rtr_type;
struct qed_iwarp_fpdu *partial_fpdus;
struct qed_iwarp_ll2_mpa_buf *mpa_bufs;
+ u8 *mpa_intermediate_buf;
u16 max_num_partial_fpdus;
};
--
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
* [PATCH v2 net-next 01/12] qed: Add ll2 option to limit the number of bds per packet
From: Michal Kalderon @ 2017-10-03 8:54 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-rdma, dledford, Michal Kalderon, Ariel Elior
In-Reply-To: <1507020902-4952-1-git-send-email-Michal.Kalderon@cavium.com>
iWARP uses 3 ll2 connections, the maximum number of bds is known
during connection setup. This patch modifies the static array in
the ll2_tx_packet descriptor to be a flexible array and
significantlly reduces memory size.
In addition, some redundant fields in the ll2_tx_packet were
removed, which also contributed to decreasing the descriptor size.
Signed-off-by: Michal Kalderon <Michal.Kalderon@cavium.com>
Signed-off-by: Ariel Elior <Ariel.Elior@cavium.com>
---
drivers/net/ethernet/qlogic/qed/qed_ll2.c | 29 +++++++++++++++++++++--------
drivers/net/ethernet/qlogic/qed/qed_ll2.h | 9 +++------
2 files changed, 24 insertions(+), 14 deletions(-)
diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.c b/drivers/net/ethernet/qlogic/qed/qed_ll2.c
index 250afa5..75af40a 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_ll2.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.c
@@ -1105,6 +1105,7 @@ static int qed_ll2_acquire_connection_tx(struct qed_hwfn *p_hwfn,
struct qed_ll2_info *p_ll2_info)
{
struct qed_ll2_tx_packet *p_descq;
+ u32 desc_size;
u32 capacity;
int rc = 0;
@@ -1122,13 +1123,17 @@ static int qed_ll2_acquire_connection_tx(struct qed_hwfn *p_hwfn,
goto out;
capacity = qed_chain_get_capacity(&p_ll2_info->tx_queue.txq_chain);
- p_descq = kcalloc(capacity, sizeof(struct qed_ll2_tx_packet),
- GFP_KERNEL);
+ /* First element is part of the packet, rest are flexibly added */
+ desc_size = (sizeof(*p_descq) +
+ (p_ll2_info->input.tx_max_bds_per_packet - 1) *
+ sizeof(p_descq->bds_set));
+
+ p_descq = kcalloc(capacity, desc_size, GFP_KERNEL);
if (!p_descq) {
rc = -ENOMEM;
goto out;
}
- p_ll2_info->tx_queue.descq_array = p_descq;
+ p_ll2_info->tx_queue.descq_mem = p_descq;
DP_VERBOSE(p_hwfn, QED_MSG_LL2,
"Allocated LL2 Txq [Type %08x] with 0x%08x buffers\n",
@@ -1359,11 +1364,13 @@ int qed_ll2_establish_connection(void *cxt, u8 connection_handle)
{
struct qed_hwfn *p_hwfn = cxt;
struct qed_ll2_info *p_ll2_conn;
+ struct qed_ll2_tx_packet *p_pkt;
struct qed_ll2_rx_queue *p_rx;
struct qed_ll2_tx_queue *p_tx;
struct qed_ptt *p_ptt;
int rc = -EINVAL;
u32 i, capacity;
+ u32 desc_size;
u8 qid;
p_ptt = qed_ptt_acquire(p_hwfn);
@@ -1397,9 +1404,15 @@ int qed_ll2_establish_connection(void *cxt, u8 connection_handle)
INIT_LIST_HEAD(&p_tx->sending_descq);
spin_lock_init(&p_tx->lock);
capacity = qed_chain_get_capacity(&p_tx->txq_chain);
- for (i = 0; i < capacity; i++)
- list_add_tail(&p_tx->descq_array[i].list_entry,
- &p_tx->free_descq);
+ /* First element is part of the packet, rest are flexibly added */
+ desc_size = (sizeof(*p_pkt) +
+ (p_ll2_conn->input.tx_max_bds_per_packet - 1) *
+ sizeof(p_pkt->bds_set));
+
+ for (i = 0; i < capacity; i++) {
+ p_pkt = p_tx->descq_mem + desc_size * i;
+ list_add_tail(&p_pkt->list_entry, &p_tx->free_descq);
+ }
p_tx->cur_completing_bd_idx = 0;
p_tx->bds_idx = 0;
p_tx->b_completing_packet = false;
@@ -1698,7 +1711,7 @@ int qed_ll2_prepare_tx_packet(void *cxt,
p_tx = &p_ll2_conn->tx_queue;
p_tx_chain = &p_tx->txq_chain;
- if (pkt->num_of_bds > CORE_LL2_TX_MAX_BDS_PER_PACKET)
+ if (pkt->num_of_bds > p_ll2_conn->input.tx_max_bds_per_packet)
return -EIO;
spin_lock_irqsave(&p_tx->lock, flags);
@@ -1858,7 +1871,7 @@ void qed_ll2_release_connection(void *cxt, u8 connection_handle)
qed_int_unregister_cb(p_hwfn, p_ll2_conn->tx_queue.tx_sb_index);
}
- kfree(p_ll2_conn->tx_queue.descq_array);
+ kfree(p_ll2_conn->tx_queue.descq_mem);
qed_chain_free(p_hwfn->cdev, &p_ll2_conn->tx_queue.txq_chain);
kfree(p_ll2_conn->rx_queue.descq_array);
diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.h b/drivers/net/ethernet/qlogic/qed/qed_ll2.h
index a822528..9bdd08f 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_ll2.h
+++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.h
@@ -63,17 +63,14 @@ struct qed_ll2_rx_packet {
struct qed_ll2_tx_packet {
struct list_head list_entry;
u16 bd_used;
- u16 vlan;
- u16 l4_hdr_offset_w;
- u8 bd_flags;
bool notify_fw;
void *cookie;
-
+ /* Flexible Array of bds_set determined by max_bds_per_packet */
struct {
struct core_tx_bd *txq_bd;
dma_addr_t tx_frag;
u16 frag_len;
- } bds_set[ETH_TX_MAX_BDS_PER_NON_LSO_PACKET];
+ } bds_set[1];
};
struct qed_ll2_rx_queue {
@@ -101,7 +98,7 @@ struct qed_ll2_tx_queue {
struct list_head active_descq;
struct list_head free_descq;
struct list_head sending_descq;
- struct qed_ll2_tx_packet *descq_array;
+ void *descq_mem; /* memory for variable sized qed_ll2_tx_packet*/
struct qed_ll2_tx_packet *cur_send_packet;
struct qed_ll2_tx_packet cur_completing_packet;
u16 cur_completing_bd_idx;
--
1.8.3.1
^ permalink raw reply related
* [PATCH v2 net-next 03/12] qed: Add ll2 option for dropping a tx packet
From: Michal Kalderon @ 2017-10-03 8:54 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-rdma, dledford, Michal Kalderon, Ariel Elior
In-Reply-To: <1507020902-4952-1-git-send-email-Michal.Kalderon@cavium.com>
The option of sending a packet on the ll2 and dropping it exists in
hardware and was not used until now, thus not exposed.
The iWARP unaligned MPA flow requires this functionality for
flushing the tx queue.
Signed-off-by: Michal Kalderon <Michal.Kalderon@cavium.com>
Signed-off-by: Ariel Elior <Ariel.Elior@cavium.com>
---
drivers/net/ethernet/qlogic/qed/qed_ll2.c | 16 ++++++++++++++--
include/linux/qed/qed_ll2_if.h | 1 +
2 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.c b/drivers/net/ethernet/qlogic/qed/qed_ll2.c
index 3c695da..ad67d36 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_ll2.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.c
@@ -1597,8 +1597,20 @@ static void qed_ll2_prepare_tx_packet_set(struct qed_hwfn *p_hwfn,
roce_flavor = (pkt->qed_roce_flavor == QED_LL2_ROCE) ? CORE_ROCE
: CORE_RROCE;
- tx_dest = (pkt->tx_dest == QED_LL2_TX_DEST_NW) ? CORE_TX_DEST_NW
- : CORE_TX_DEST_LB;
+ switch (pkt->tx_dest) {
+ case QED_LL2_TX_DEST_NW:
+ tx_dest = CORE_TX_DEST_NW;
+ break;
+ case QED_LL2_TX_DEST_LB:
+ tx_dest = CORE_TX_DEST_LB;
+ break;
+ case QED_LL2_TX_DEST_DROP:
+ tx_dest = CORE_TX_DEST_DROP;
+ break;
+ default:
+ tx_dest = CORE_TX_DEST_LB;
+ break;
+ }
start_bd = (struct core_tx_bd *)qed_chain_produce(p_tx_chain);
start_bd->nw_vlan_or_lb_echo = cpu_to_le16(pkt->vlan);
diff --git a/include/linux/qed/qed_ll2_if.h b/include/linux/qed/qed_ll2_if.h
index d7cca59..95fdf02 100644
--- a/include/linux/qed/qed_ll2_if.h
+++ b/include/linux/qed/qed_ll2_if.h
@@ -64,6 +64,7 @@ enum qed_ll2_roce_flavor_type {
enum qed_ll2_tx_dest {
QED_LL2_TX_DEST_NW, /* Light L2 TX Destination to the Network */
QED_LL2_TX_DEST_LB, /* Light L2 TX Destination to the Loopback */
+ QED_LL2_TX_DEST_DROP, /* Light L2 Drop the TX packet */
QED_LL2_TX_DEST_MAX
};
--
1.8.3.1
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox