* [PATCH bpf-next] selftests: bpf: test_lwt_ip_encap: add negative tests.
From: Peter Oskolkov @ 2019-02-15 23:49 UTC (permalink / raw)
To: Alexei Starovoitov, Daniel Borkmann, netdev
Cc: Peter Oskolkov, David Ahern, Peter Oskolkov
As requested by David Ahern:
- add negative tests (no routes, explicitly unreachable destinations)
to exercize error handling code paths;
- do not exit on test failures, but instead print a summary of
passed/failed tests at the end.
Future patches will add TSO and VRF tests.
Signed-off-by: Peter Oskolkov <posk@google.com>
---
.../selftests/bpf/test_lwt_ip_encap.sh | 111 ++++++++++++++----
1 file changed, 88 insertions(+), 23 deletions(-)
diff --git a/tools/testing/selftests/bpf/test_lwt_ip_encap.sh b/tools/testing/selftests/bpf/test_lwt_ip_encap.sh
index 4ca714e23ab0..612632c1425f 100755
--- a/tools/testing/selftests/bpf/test_lwt_ip_encap.sh
+++ b/tools/testing/selftests/bpf/test_lwt_ip_encap.sh
@@ -38,8 +38,6 @@
# ping: SRC->[encap at veth2:ingress]->GRE:decap->DST
# ping replies go DST->SRC directly
-set -e # exit on error
-
if [[ $EUID -ne 0 ]]; then
echo "This script must be run as root"
echo "FAIL"
@@ -76,8 +74,37 @@ readonly IPv6_GRE="fb10::1"
readonly IPv6_SRC=$IPv6_1
readonly IPv6_DST=$IPv6_4
-setup() {
-set -e # exit on error
+TEST_STATUS=0
+TESTS_SUCCEEDED=0
+TESTS_FAILED=0
+
+process_test_results()
+{
+ if [[ "${TEST_STATUS}" -eq 0 ]] ; then
+ echo "PASS"
+ TESTS_SUCCEEDED=$((TESTS_SUCCEEDED+1))
+ else
+ echo "FAIL"
+ TESTS_FAILED=$((TESTS_FAILED+1))
+ fi
+}
+
+print_test_summary_and_exit()
+{
+ echo "passed tests: ${TESTS_SUCCEEDED}"
+ echo "failed tests: ${TESTS_FAILED}"
+ if [ "${TESTS_FAILED}" -eq "0" ] ; then
+ exit 0
+ else
+ exit 1
+ fi
+}
+
+setup()
+{
+ set -e # exit on error
+ TEST_STATUS=0
+
# create devices and namespaces
ip netns add "${NS1}"
ip netns add "${NS2}"
@@ -178,7 +205,7 @@ set -e # exit on error
# configure IPv4 GRE device in NS3, and a route to it via the "bottom" route
ip -netns ${NS3} tunnel add gre_dev mode gre remote ${IPv4_1} local ${IPv4_GRE} ttl 255
ip -netns ${NS3} link set gre_dev up
- ip -netns ${NS3} addr add ${IPv4_GRE} dev gre_dev
+ ip -netns ${NS3} addr add ${IPv4_GRE} nodad dev gre_dev
ip -netns ${NS1} route add ${IPv4_GRE}/32 dev veth5 via ${IPv4_6}
ip -netns ${NS2} route add ${IPv4_GRE}/32 dev veth7 via ${IPv4_8}
@@ -194,9 +221,13 @@ set -e # exit on error
ip netns exec ${NS1} sysctl -wq net.ipv4.conf.all.rp_filter=0
ip netns exec ${NS2} sysctl -wq net.ipv4.conf.all.rp_filter=0
ip netns exec ${NS3} sysctl -wq net.ipv4.conf.all.rp_filter=0
+
+ sleep 1 # reduce flakiness
+ set +e
}
-cleanup() {
+cleanup()
+{
ip netns del ${NS1} 2> /dev/null
ip netns del ${NS2} 2> /dev/null
ip netns del ${NS3} 2> /dev/null
@@ -204,12 +235,28 @@ cleanup() {
trap cleanup EXIT
-test_ping() {
+remove_routes_to_gredev()
+{
+ ip -netns ${NS1} route del ${IPv4_GRE} dev veth5
+ ip -netns ${NS2} route del ${IPv4_GRE} dev veth7
+ ip -netns ${NS1} -6 route del ${IPv6_GRE}/128 dev veth5
+ ip -netns ${NS2} -6 route del ${IPv6_GRE}/128 dev veth7
+}
+
+add_unreachable_routes_to_gredev()
+{
+ ip -netns ${NS1} route add unreachable ${IPv4_GRE}/32
+ ip -netns ${NS2} route add unreachable ${IPv4_GRE}/32
+ ip -netns ${NS1} -6 route add unreachable ${IPv6_GRE}/128
+ ip -netns ${NS2} -6 route add unreachable ${IPv6_GRE}/128
+}
+
+test_ping()
+{
local readonly PROTO=$1
local readonly EXPECTED=$2
local RET=0
- set +e
if [ "${PROTO}" == "IPv4" ] ; then
ip netns exec ${NS1} ping -c 1 -W 1 -I ${IPv4_SRC} ${IPv4_DST} 2>&1 > /dev/null
RET=$?
@@ -217,29 +264,26 @@ test_ping() {
ip netns exec ${NS1} ping6 -c 1 -W 6 -I ${IPv6_SRC} ${IPv6_DST} 2>&1 > /dev/null
RET=$?
else
- echo "test_ping: unknown PROTO: ${PROTO}"
- exit 1
+ echo " test_ping: unknown PROTO: ${PROTO}"
+ TEST_STATUS=1
fi
- set -e
if [ "0" != "${RET}" ]; then
RET=1
fi
if [ "${EXPECTED}" != "${RET}" ] ; then
- echo "FAIL: test_ping: ${RET}"
- exit 1
+ echo " test_ping failed: expected: ${EXPECTED}; got ${RET}"
+ TEST_STATUS=1
fi
}
-test_egress() {
+test_egress()
+{
local readonly ENCAP=$1
echo "starting egress ${ENCAP} encap test"
setup
- # need to wait a bit for IPv6 to autoconf, otherwise
- # ping6 sometimes fails with "unable to bind to address"
-
# by default, pings work
test_ping IPv4 0
test_ping IPv6 0
@@ -258,16 +302,28 @@ test_egress() {
ip -netns ${NS1} route add ${IPv4_DST} encap bpf xmit obj test_lwt_ip_encap.o sec encap_gre6 dev veth1
ip -netns ${NS1} -6 route add ${IPv6_DST} encap bpf xmit obj test_lwt_ip_encap.o sec encap_gre6 dev veth1
else
- echo "FAIL: unknown encap ${ENCAP}"
+ echo " unknown encap ${ENCAP}"
+ TEST_STATUS=1
fi
test_ping IPv4 0
test_ping IPv6 0
+ # a negative test: remove routes to GRE devices: ping fails
+ remove_routes_to_gredev
+ test_ping IPv4 1
+ test_ping IPv6 1
+
+ # another negative test
+ add_unreachable_routes_to_gredev
+ test_ping IPv4 1
+ test_ping IPv6 1
+
cleanup
- echo "PASS"
+ process_test_results
}
-test_ingress() {
+test_ingress()
+{
local readonly ENCAP=$1
echo "starting ingress ${ENCAP} encap test"
setup
@@ -298,14 +354,23 @@ test_ingress() {
test_ping IPv4 0
test_ping IPv6 0
+ # a negative test: remove routes to GRE devices: ping fails
+ remove_routes_to_gredev
+ test_ping IPv4 1
+ test_ping IPv6 1
+
+ # another negative test
+ add_unreachable_routes_to_gredev
+ test_ping IPv4 1
+ test_ping IPv6 1
+
cleanup
- echo "PASS"
+ process_test_results
}
test_egress IPv4
test_egress IPv6
-
test_ingress IPv4
test_ingress IPv6
-echo "all tests passed"
+print_test_summary_and_exit
--
2.21.0.rc0.258.g878e2cd30e-goog
^ permalink raw reply related
* Re: [PATCH net-next v5] ipmr: ip6mr: Create new sockopt to clear mfc cache or vifs
From: Nikolay Aleksandrov @ 2019-02-16 0:21 UTC (permalink / raw)
To: Callum Sinclair, davem, kuznet, yoshfuji, netdev, linux-kernel
Cc: nicolas.dichtel
In-Reply-To: <20190214024418.21490-2-callum.sinclair@alliedtelesis.co.nz>
On 14/02/2019 04:44, Callum Sinclair wrote:
> Currently the only way to clear the forwarding cache was to delete the
> entries one by one using the MRT_DEL_MFC socket option or to destroy and
> recreate the socket.
>
> Create a new socket option which with the use of optional flags can
> clear any combination of multicast entries (static or not static) and
> multicast vifs (static or not static).
>
> Calling the new socket option MRT_FLUSH with the flags MRT_FLUSH_MFC and
> MRT_FLUSH_VIFS will clear all entries and vifs on the socket except for
> static entries.
>
> Signed-off-by: Callum Sinclair <callum.sinclair@alliedtelesis.co.nz>
> ---
> v1 -> v2:
> Implemented additional flags for static entries
> v2 -> v3:
> Cleaned up flag logic so any combination of routes can be cleared.
> Fixed style errors
> Fixed incorrect flag values
> v3 -> v4:
> Fixed style errors
> Fixed incorrect flag (MRT_FLUSH was used instead of MRT_FLUSH_VIFS)
> v4 -> v5:
> Only clear the unresolved queue when MRT_FLUSH_MFC flag is set.
>
> include/uapi/linux/mroute.h | 9 ++++-
> include/uapi/linux/mroute6.h | 9 ++++-
> net/ipv4/ipmr.c | 75 +++++++++++++++++++++-------------
> net/ipv6/ip6mr.c | 78 +++++++++++++++++++++++-------------
> 4 files changed, 115 insertions(+), 56 deletions(-)
>
+1 about Nicolas' comments, other than that looks good:
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
> diff --git a/include/uapi/linux/mroute.h b/include/uapi/linux/mroute.h
> index 5d37a9ccce63..11c8c1fc1124 100644
> --- a/include/uapi/linux/mroute.h
> +++ b/include/uapi/linux/mroute.h
> @@ -28,12 +28,19 @@
> #define MRT_TABLE (MRT_BASE+9) /* Specify mroute table ID */
> #define MRT_ADD_MFC_PROXY (MRT_BASE+10) /* Add a (*,*|G) mfc entry */
> #define MRT_DEL_MFC_PROXY (MRT_BASE+11) /* Del a (*,*|G) mfc entry */
> -#define MRT_MAX (MRT_BASE+11)
> +#define MRT_FLUSH (MRT_BASE+12) /* Flush all mfc entries and/or vifs */
> +#define MRT_MAX (MRT_BASE+12)
>
> #define SIOCGETVIFCNT SIOCPROTOPRIVATE /* IP protocol privates */
> #define SIOCGETSGCNT (SIOCPROTOPRIVATE+1)
> #define SIOCGETRPF (SIOCPROTOPRIVATE+2)
>
> +/* MRT_FLUSH optional flags */
> +#define MRT_FLUSH_MFC 1 /* Flush multicast entries */
> +#define MRT_FLUSH_MFC_STATIC 2 /* Flush static multicast entries */
> +#define MRT_FLUSH_VIFS 4 /* Flush multicast vifs */
> +#define MRT_FLUSH_VIFS_STATIC 8 /* Flush static multicast vifs */
> +
> #define MAXVIFS 32
> typedef unsigned long vifbitmap_t; /* User mode code depends on this lot */
> typedef unsigned short vifi_t;
> diff --git a/include/uapi/linux/mroute6.h b/include/uapi/linux/mroute6.h
> index 9999cc006390..ac84ef11b29c 100644
> --- a/include/uapi/linux/mroute6.h
> +++ b/include/uapi/linux/mroute6.h
> @@ -31,12 +31,19 @@
> #define MRT6_TABLE (MRT6_BASE+9) /* Specify mroute table ID */
> #define MRT6_ADD_MFC_PROXY (MRT6_BASE+10) /* Add a (*,*|G) mfc entry */
> #define MRT6_DEL_MFC_PROXY (MRT6_BASE+11) /* Del a (*,*|G) mfc entry */
> -#define MRT6_MAX (MRT6_BASE+11)
> +#define MRT6_FLUSH (MRT6_BASE+12) /* Flush all mfc entries and/or vifs */
> +#define MRT6_MAX (MRT6_BASE+12)
>
> #define SIOCGETMIFCNT_IN6 SIOCPROTOPRIVATE /* IP protocol privates */
> #define SIOCGETSGCNT_IN6 (SIOCPROTOPRIVATE+1)
> #define SIOCGETRPF (SIOCPROTOPRIVATE+2)
>
> +/* MRT6_FLUSH optional flags */
> +#define MRT6_FLUSH_MFC 1 /* Flush multicast entries */
> +#define MRT6_FLUSH_MFC_STATIC 2 /* Flush static multicast entries */
> +#define MRT6_FLUSH_VIFS 4 /* Flushing multicast vifs */
> +#define MRT6_FLUSH_VIFS_STATIC 8 /* Flush static multicast vifs */
> +
> #define MAXMIFS 32
> typedef unsigned long mifbitmap_t; /* User mode code depends on this lot */
> typedef unsigned short mifi_t;
> diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
> index e536970557dd..53869779af74 100644
> --- a/net/ipv4/ipmr.c
> +++ b/net/ipv4/ipmr.c
> @@ -110,7 +110,7 @@ static int ipmr_cache_report(struct mr_table *mrt,
> static void mroute_netlink_event(struct mr_table *mrt, struct mfc_cache *mfc,
> int cmd);
> static void igmpmsg_netlink_event(struct mr_table *mrt, struct sk_buff *pkt);
> -static void mroute_clean_tables(struct mr_table *mrt, bool all);
> +static void mroute_clean_tables(struct mr_table *mrt, int flags);
> static void ipmr_expire_process(struct timer_list *t);
>
> #ifdef CONFIG_IP_MROUTE_MULTIPLE_TABLES
> @@ -415,7 +415,8 @@ static struct mr_table *ipmr_new_table(struct net *net, u32 id)
> static void ipmr_free_table(struct mr_table *mrt)
> {
> del_timer_sync(&mrt->ipmr_expire_timer);
> - mroute_clean_tables(mrt, true);
> + mroute_clean_tables(mrt, MRT_FLUSH_VIFS | MRT_FLUSH_VIFS_STATIC |
> + MRT_FLUSH_MFC | MRT_FLUSH_MFC_STATIC);
> rhltable_destroy(&mrt->mfc_hash);
> kfree(mrt);
> }
> @@ -1296,7 +1297,7 @@ static int ipmr_mfc_add(struct net *net, struct mr_table *mrt,
> }
>
> /* Close the multicast socket, and clear the vif tables etc */
> -static void mroute_clean_tables(struct mr_table *mrt, bool all)
> +static void mroute_clean_tables(struct mr_table *mrt, int flags)
> {
> struct net *net = read_pnet(&mrt->net);
> struct mr_mfc *c, *tmp;
> @@ -1305,35 +1306,44 @@ static void mroute_clean_tables(struct mr_table *mrt, bool all)
> int i;
>
> /* Shut down all active vif entries */
> - for (i = 0; i < mrt->maxvif; i++) {
> - if (!all && (mrt->vif_table[i].flags & VIFF_STATIC))
> - continue;
> - vif_delete(mrt, i, 0, &list);
> + if (flags & (MRT_FLUSH_VIFS | MRT_FLUSH_VIFS_STATIC)) {
> + for (i = 0; i < mrt->maxvif; i++) {
> + if (((mrt->vif_table[i].flags & VIFF_STATIC) &&
> + !(flags & MRT_FLUSH_VIFS_STATIC)) ||
> + (!(mrt->vif_table[i].flags & VIFF_STATIC) && !(flags & MRT_FLUSH_VIFS)))
> + continue;
> + vif_delete(mrt, i, 0, &list);
> + }
> + unregister_netdevice_many(&list);
> }
> - unregister_netdevice_many(&list);
>
> /* Wipe the cache */
> - list_for_each_entry_safe(c, tmp, &mrt->mfc_cache_list, list) {
> - if (!all && (c->mfc_flags & MFC_STATIC))
> - continue;
> - rhltable_remove(&mrt->mfc_hash, &c->mnode, ipmr_rht_params);
> - list_del_rcu(&c->list);
> - cache = (struct mfc_cache *)c;
> - call_ipmr_mfc_entry_notifiers(net, FIB_EVENT_ENTRY_DEL, cache,
> - mrt->id);
> - mroute_netlink_event(mrt, cache, RTM_DELROUTE);
> - mr_cache_put(c);
> - }
> -
> - if (atomic_read(&mrt->cache_resolve_queue_len) != 0) {
> - spin_lock_bh(&mfc_unres_lock);
> - list_for_each_entry_safe(c, tmp, &mrt->mfc_unres_queue, list) {
> - list_del(&c->list);
> + if (flags & (MRT_FLUSH_MFC | MRT_FLUSH_MFC_STATIC)) {
> + list_for_each_entry_safe(c, tmp, &mrt->mfc_cache_list, list) {
> + if (((c->mfc_flags & MFC_STATIC) && !(flags & MRT_FLUSH_MFC_STATIC)) ||
> + (!(c->mfc_flags & MFC_STATIC) && !(flags & MRT_FLUSH_MFC)))
> + continue;
> + rhltable_remove(&mrt->mfc_hash, &c->mnode, ipmr_rht_params);
> + list_del_rcu(&c->list);
> cache = (struct mfc_cache *)c;
> + call_ipmr_mfc_entry_notifiers(net, FIB_EVENT_ENTRY_DEL, cache,
> + mrt->id);
> mroute_netlink_event(mrt, cache, RTM_DELROUTE);
> - ipmr_destroy_unres(mrt, cache);
> + mr_cache_put(c);
> + }
> + }
> +
> + if (flags & MRT_FLUSH_MFC) {
> + if (atomic_read(&mrt->cache_resolve_queue_len) != 0) {
> + spin_lock_bh(&mfc_unres_lock);
> + list_for_each_entry_safe(c, tmp, &mrt->mfc_unres_queue, list) {
> + list_del(&c->list);
> + cache = (struct mfc_cache *)c;
> + mroute_netlink_event(mrt, cache, RTM_DELROUTE);
> + ipmr_destroy_unres(mrt, cache);
> + }
> + spin_unlock_bh(&mfc_unres_lock);
> }
> - spin_unlock_bh(&mfc_unres_lock);
> }
> }
>
> @@ -1354,7 +1364,7 @@ static void mrtsock_destruct(struct sock *sk)
> NETCONFA_IFINDEX_ALL,
> net->ipv4.devconf_all);
> RCU_INIT_POINTER(mrt->mroute_sk, NULL);
> - mroute_clean_tables(mrt, false);
> + mroute_clean_tables(mrt, MRT_FLUSH_VIFS | MRT_FLUSH_MFC);
> }
> }
> rtnl_unlock();
> @@ -1479,6 +1489,17 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval,
> sk == rtnl_dereference(mrt->mroute_sk),
> parent);
> break;
> + case MRT_FLUSH:
> + if (optlen != sizeof(val)) {
> + ret = -EINVAL;
> + break;
> + }
> + if (get_user(val, (int __user *)optval)) {
> + ret = -EFAULT;
> + break;
> + }
> + mroute_clean_tables(mrt, val);
> + break;
> /* Control PIM assert. */
> case MRT_ASSERT:
> if (optlen != sizeof(val)) {
> diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c
> index cc01aa3f2b5e..b67a7c1e3615 100644
> --- a/net/ipv6/ip6mr.c
> +++ b/net/ipv6/ip6mr.c
> @@ -97,7 +97,7 @@ static void mr6_netlink_event(struct mr_table *mrt, struct mfc6_cache *mfc,
> static void mrt6msg_netlink_event(struct mr_table *mrt, struct sk_buff *pkt);
> static int ip6mr_rtm_dumproute(struct sk_buff *skb,
> struct netlink_callback *cb);
> -static void mroute_clean_tables(struct mr_table *mrt, bool all);
> +static void mroute_clean_tables(struct mr_table *mrt, int flags);
> static void ipmr_expire_process(struct timer_list *t);
>
> #ifdef CONFIG_IPV6_MROUTE_MULTIPLE_TABLES
> @@ -393,7 +393,8 @@ static struct mr_table *ip6mr_new_table(struct net *net, u32 id)
> static void ip6mr_free_table(struct mr_table *mrt)
> {
> del_timer_sync(&mrt->ipmr_expire_timer);
> - mroute_clean_tables(mrt, true);
> + mroute_clean_tables(mrt, MRT6_FLUSH_VIFS | MRT6_FLUSH_VIFS_STATIC |
> + MRT6_FLUSH_MFC | MRT6_FLUSH_MFC_STATIC);
> rhltable_destroy(&mrt->mfc_hash);
> kfree(mrt);
> }
> @@ -1496,42 +1497,51 @@ static int ip6mr_mfc_add(struct net *net, struct mr_table *mrt,
> * Close the multicast socket, and clear the vif tables etc
> */
>
> -static void mroute_clean_tables(struct mr_table *mrt, bool all)
> +static void mroute_clean_tables(struct mr_table *mrt, int flags)
> {
> struct mr_mfc *c, *tmp;
> LIST_HEAD(list);
> int i;
>
> /* Shut down all active vif entries */
> - for (i = 0; i < mrt->maxvif; i++) {
> - if (!all && (mrt->vif_table[i].flags & VIFF_STATIC))
> - continue;
> - mif6_delete(mrt, i, 0, &list);
> + if (flags & (MRT6_FLUSH_VIFS | MRT6_FLUSH_VIFS_STATIC)) {
> + for (i = 0; i < mrt->maxvif; i++) {
> + if (((mrt->vif_table[i].flags & VIFF_STATIC) &&
> + !(flags & MRT6_FLUSH_VIFS_STATIC)) ||
> + (!(mrt->vif_table[i].flags & VIFF_STATIC) && !(flags & MRT6_FLUSH_VIFS)))
> + continue;
> + mif6_delete(mrt, i, 0, &list);
> + }
> + unregister_netdevice_many(&list);
> }
> - unregister_netdevice_many(&list);
>
> /* Wipe the cache */
> - list_for_each_entry_safe(c, tmp, &mrt->mfc_cache_list, list) {
> - if (!all && (c->mfc_flags & MFC_STATIC))
> - continue;
> - rhltable_remove(&mrt->mfc_hash, &c->mnode, ip6mr_rht_params);
> - list_del_rcu(&c->list);
> - call_ip6mr_mfc_entry_notifiers(read_pnet(&mrt->net),
> - FIB_EVENT_ENTRY_DEL,
> - (struct mfc6_cache *)c, mrt->id);
> - mr6_netlink_event(mrt, (struct mfc6_cache *)c, RTM_DELROUTE);
> - mr_cache_put(c);
> + if (flags & (MRT6_FLUSH_MFC | MRT6_FLUSH_MFC_STATIC)) {
> + list_for_each_entry_safe(c, tmp, &mrt->mfc_cache_list, list) {
> + if (((c->mfc_flags & MFC_STATIC) && !(flags & MRT6_FLUSH_MFC_STATIC)) ||
> + (!(c->mfc_flags & MFC_STATIC) && !(flags & MRT6_FLUSH_MFC)))
> + continue;
> + rhltable_remove(&mrt->mfc_hash, &c->mnode, ip6mr_rht_params);
> + list_del_rcu(&c->list);
> + call_ip6mr_mfc_entry_notifiers(read_pnet(&mrt->net),
> + FIB_EVENT_ENTRY_DEL,
> + (struct mfc6_cache *)c, mrt->id);
> + mr6_netlink_event(mrt, (struct mfc6_cache *)c, RTM_DELROUTE);
> + mr_cache_put(c);
> + }
> }
>
> - if (atomic_read(&mrt->cache_resolve_queue_len) != 0) {
> - spin_lock_bh(&mfc_unres_lock);
> - list_for_each_entry_safe(c, tmp, &mrt->mfc_unres_queue, list) {
> - list_del(&c->list);
> - mr6_netlink_event(mrt, (struct mfc6_cache *)c,
> - RTM_DELROUTE);
> - ip6mr_destroy_unres(mrt, (struct mfc6_cache *)c);
> + if (flags & MRT6_FLUSH_MFC) {
> + if (atomic_read(&mrt->cache_resolve_queue_len) != 0) {
> + spin_lock_bh(&mfc_unres_lock);
> + list_for_each_entry_safe(c, tmp, &mrt->mfc_unres_queue, list) {
> + list_del(&c->list);
> + mr6_netlink_event(mrt, (struct mfc6_cache *)c,
> + RTM_DELROUTE);
> + ip6mr_destroy_unres(mrt, (struct mfc6_cache *)c);
> + }
> + spin_unlock_bh(&mfc_unres_lock);
> }
> - spin_unlock_bh(&mfc_unres_lock);
> }
> }
>
> @@ -1587,7 +1597,7 @@ int ip6mr_sk_done(struct sock *sk)
> NETCONFA_IFINDEX_ALL,
> net->ipv6.devconf_all);
>
> - mroute_clean_tables(mrt, false);
> + mroute_clean_tables(mrt, MRT6_FLUSH_VIFS | MRT6_FLUSH_MFC);
> err = 0;
> break;
> }
> @@ -1703,6 +1713,20 @@ int ip6_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, uns
> rtnl_unlock();
> return ret;
>
> + case MRT6_FLUSH:
> + {
> + int flags;
> +
> + if (optlen != sizeof(flags))
> + return -EINVAL;
> + if (get_user(flags, (int __user *)optval))
> + return -EFAULT;
> + rtnl_lock();
> + mroute_clean_tables(mrt, flags);
> + rtnl_unlock();
> + return 0;
> + }
> +
> /*
> * Control PIM assert (to activate pim will activate assert)
> */
>
^ permalink raw reply
* Re: [PATCH] net: sched: matchall: verify that filter is not NULL in mall_walk()
From: Cong Wang @ 2019-02-16 0:24 UTC (permalink / raw)
To: Vlad Buslov
Cc: Ido Schimmel, Linux Kernel Network Developers, Jamal Hadi Salim,
Jiri Pirko, David Miller
In-Reply-To: <20190215121120.4971-1-vladbu@mellanox.com>
On Fri, Feb 15, 2019 at 4:11 AM Vlad Buslov <vladbu@mellanox.com> wrote:
>
> Check that filter is not NULL before passing it to tcf_walker->fn()
> callback. This can happen when mall_change() failed to offload filter to
> hardware.
>
> Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
> ---
> net/sched/cls_matchall.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/net/sched/cls_matchall.c b/net/sched/cls_matchall.c
> index a37137430e61..1f9d481b0fbb 100644
> --- a/net/sched/cls_matchall.c
> +++ b/net/sched/cls_matchall.c
> @@ -247,6 +247,9 @@ static void mall_walk(struct tcf_proto *tp, struct tcf_walker *arg,
>
> if (arg->count < arg->skip)
> goto skip;
> +
> + if (!head)
> + return;
So head==NULL still counts one given that you check NULL after
checking arg->count. Is this expected?
^ permalink raw reply
* Re: [PATCH bpf-next v2] bpf: make LWTUNNEL_BPF dependent on INET
From: Daniel Borkmann @ 2019-02-16 0:16 UTC (permalink / raw)
To: Randy Dunlap, Peter Oskolkov, Alexei Starovoitov, netdev; +Cc: Peter Oskolkov
In-Reply-To: <7bbc8f63-3033-44ba-5b8e-ae597677bbe6@infradead.org>
On 02/15/2019 11:55 PM, Randy Dunlap wrote:
> On 2/15/19 9:51 AM, Peter Oskolkov wrote:
>> Lightweight tunnels are L3 constructs that are used with IP/IP6.
>>
>> For example, lwtunnel_xmit is called from ip_output.c and
>> ip6_output.c only.
>>
>> Make the dependency explicit at least for LWT-BPF, as now they
>> call into IP routing.
>>
>> V2: added "Reported-by" below.
>>
>> Reported-by: Randy Dunlap <rdunlap@infradead.org>
>> Signed-off-by: Peter Oskolkov <posk@google.com>
>
> Yes, that works. Thanks.
>
> Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested
Applied, thanks!
^ permalink raw reply
* Re: [PATCH net-next 1/3] net: stmmac: Fix NAPI poll in TX path when in multi-queue
From: Florian Fainelli @ 2019-02-16 0:35 UTC (permalink / raw)
To: Jose Abreu, netdev, linux-kernel
Cc: Joao Pinto, David S . Miller, Giuseppe Cavallaro,
Alexandre Torgue
In-Reply-To: <76da803a662214b024bfdb95731c38e45aef426d.1550237884.git.joabreu@synopsys.com>
On 2/15/19 5:42 AM, Jose Abreu wrote:
> Commit 8fce33317023 introduced the concept of NAPI per-channel and
> independent cleaning of TX path.
>
> This is currently breaking performance in some cases. The scenario
> happens when all packets are being received in Queue 0 but the TX is
> performed in Queue != 0.
>
> Fix this by using different NAPI instances per each TX and RX queue, as
> suggested by Florian.
>
> Signed-off-by: Jose Abreu <joabreu@synopsys.com>
> Cc: Florian Fainelli <f.fainelli@gmail.com>
> Cc: Joao Pinto <jpinto@synopsys.com>
> Cc: David S. Miller <davem@davemloft.net>
> Cc: Giuseppe Cavallaro <peppe.cavallaro@st.com>
> Cc: Alexandre Torgue <alexandre.torgue@st.com>
> ---
[snip]
> - if (work_done < budget && napi_complete_done(napi, work_done)) {
> - int stat;
> + priv->xstats.napi_poll++;
>
> + work_done = stmmac_tx_clean(priv, budget, chan);
> + if (work_done < budget && napi_complete_done(napi, work_done))
You should not be bounding your TX queue against the NAPI budge, it
should run unbound and clean as much as it can, which could be the
entire ring size if that is how many packets you pushed between
interrupts. That could be the cause of poor performance as well.
--
Florian
^ permalink raw reply
* Re: [PATCH net-next v2 0/9] net: Get rid of switchdev_port_attr_get()
From: Florian Fainelli @ 2019-02-16 0:37 UTC (permalink / raw)
To: netdev, David S. Miller
Cc: Ido Schimmel, open list, open list:STAGING SUBSYSTEM,
moderated list:ETHERNET BRIDGE, jiri, andrew, vivien.didelot
In-Reply-To: <20190215225313.32303-1-f.fainelli@gmail.com>
On 2/15/19 2:53 PM, Florian Fainelli wrote:
> Hi all,
>
> This patch series splits the removal of the switchdev_ops that was
> proposed a few times before and first tackles the easy part which is the
> removal of the single call to switchdev_port_attr_get() within the
> bridge code.
>
> As suggestd by Ido, this patch series adds a
> SWITCHDEV_ATTR_ID_PORT_PRE_BRIDGE_FLAGS which is used in the same
> context as the caller of switchdev_port_attr_set(), so not deferred, and
> then the operation is carried out in deferred context with setting a
> support bridge port flag.
>
> Follow-up patches will do the switchdev_ops removal after introducing
> the proper helpers for the switchdev blocking notifier to work across
> stacked devices (unlike the previous submissions).
David, please ignore this version, I will repost one that actually
builds, need to keep mangling with my kernel configuration and keep
those drivers enabled...
>
> Changes in v2:
>
> - differentiate callers not supporting switchdev_port_attr_set() from
> the driver not being able to support specific bridge flags
>
> - pass "mask" instead of "flags" for the PRE_BRIDGE_FLAGS check
>
> - skip prepare phase for PRE_BRIDGE_FLAGS
>
> - corrected documentation a bit more
>
> - tested bridge_vlan_aware.sh with veth/VRF
>
> Florian Fainelli (9):
> Documentation: networking: switchdev: Update port parent ID section
> net: switchdev: Add PORT_PRE_BRIDGE_FLAGS
> mlxsw: spectrum: Handle PORT_PRE_BRIDGE_FLAGS
> staging: fsl-dpaa2: ethsw: Handle PORT_PRE_BRIDGE_FLAGS
> net: dsa: Add setter for SWITCHDEV_ATTR_ID_PORT_BRIDGE_FLAGS
> rocker: Check Handle PORT_PRE_BRIDGE_FLAGS
> net: bridge: Stop calling switchdev_port_attr_get()
> net: Remove SWITCHDEV_ATTR_ID_PORT_BRIDGE_FLAGS_SUPPORT
> net: Get rid of switchdev_port_attr_get()
>
> Documentation/networking/switchdev.txt | 16 ++--
> .../mellanox/mlxsw/spectrum_switchdev.c | 38 +++++-----
> drivers/net/ethernet/rocker/rocker_main.c | 75 +++++++++++--------
> drivers/staging/fsl-dpaa2/ethsw/ethsw.c | 32 ++++----
> include/net/switchdev.h | 13 +---
> net/bridge/br_switchdev.c | 11 ++-
> net/dsa/dsa_priv.h | 6 ++
> net/dsa/port.c | 17 +++++
> net/dsa/slave.c | 24 +++---
> 9 files changed, 128 insertions(+), 104 deletions(-)
>
--
Florian
^ permalink raw reply
* [PATCH net-next] ip_tunnel: Fix DST_METADATA dst_entry handle in tnl_update_pmtu
From: wenxu @ 2019-02-16 0:58 UTC (permalink / raw)
To: davem, rong.a.chen, netdev; +Cc: sfr, lkp
From: wenxu <wenxu@ucloud.cn>
BUG report in selftests: bpf: test_tunnel.sh
Testing IPIP tunnel...
BUG: unable to handle kernel NULL pointer dereference at 0000000000000000
PGD 0 P4D 0
Oops: 0010 [#1] SMP PTI
CPU: 0 PID: 16822 Comm: ping Not tainted 5.0.0-rc3-00352-gc8b34e6 #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014
RIP: 0010: (null)
Code: Bad RIP value.
RSP: 0018:ffffc9000104f9c8 EFLAGS: 00010286
RAX: 0000000000000000 RBX: ffffe8ffffc071a8 RCX: 0000000000000000
RDX: ffff888054e33000 RSI: ffff88807796f500 RDI: ffffe8ffffc07130
RBP: ffff88807796f500 R08: ffff88806da4f0a0 R09: 0000000000000000
R10: 0000000000000004 R11: ffff888054e33000 R12: 0000000000000054
R13: ffff88805e714000 R14: ffff88806da4f0a0 R15: 0000000000000000
FS: 00007f4c00431500(0000) GS:ffff88813fc00000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: ffffffffffffffd6 CR3: 000000008276e000 CR4: 00000000000406f0
Call Trace:
? tnl_update_pmtu+0x21b/0x250 [ip_tunnel]
? ip_md_tunnel_xmit+0x1b7/0xdc0 [ip_tunnel]
? ipip_tunnel_xmit+0x90/0xc0 [ipip]
? dev_hard_start_xmit+0x98/0x210
? __dev_queue_xmit+0x6a9/0x8e0
The bpf program set tunnel_key through bpf_skb_set_tunnel_key which will
drop the old dst_entry and create a DST_METADATA dst_entry. It will lead
the tunnel_update_pmtu operator the dst_entry incorrect. So It should be
check the dst_entry is valid.
Fixes: c8b34e680a09 ("ip_tunnel: Add tnl_update_pmtu in ip_md_tunnel_xmit")
Signed-off-by: wenxu <wenxu@ucloud.cn>
---
net/ipv4/ip_tunnel.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/ipv4/ip_tunnel.c b/net/ipv4/ip_tunnel.c
index 893f013..a665f11 100644
--- a/net/ipv4/ip_tunnel.c
+++ b/net/ipv4/ip_tunnel.c
@@ -515,7 +515,7 @@ static int tnl_update_pmtu(struct net_device *dev, struct sk_buff *skb,
mtu = dst_mtu(&rt->dst) - dev->hard_header_len
- sizeof(struct iphdr) - tunnel_hlen;
else
- mtu = skb_dst(skb) ? dst_mtu(skb_dst(skb)) : dev->mtu;
+ mtu = skb_valid_dst(skb) ? dst_mtu(skb_dst(skb)) : dev->mtu;
skb_dst_update_pmtu(skb, mtu);
@@ -530,7 +530,7 @@ static int tnl_update_pmtu(struct net_device *dev, struct sk_buff *skb,
}
#if IS_ENABLED(CONFIG_IPV6)
else if (skb->protocol == htons(ETH_P_IPV6)) {
- struct rt6_info *rt6 = (struct rt6_info *)skb_dst(skb);
+ struct rt6_info *rt6 = (struct rt6_info *)skb_valid_dst(skb);
__be32 daddr;
daddr = md ? dst : tunnel->parms.iph.daddr;
--
1.8.3.1
^ permalink raw reply related
* Re: [MERGE HELP] cls_tcindex.c
From: Cong Wang @ 2019-02-16 1:31 UTC (permalink / raw)
To: David Miller; +Cc: Linux Kernel Network Developers, Vlad Buslov
In-Reply-To: <20190215.124125.1781326145216021027.davem@davemloft.net>
On Fri, Feb 15, 2019 at 12:41 PM David Miller <davem@davemloft.net> wrote:
>
>
> I've merged net into net-next.
>
> The worst conflict was cls_tcindex.c as Cong's fixes collided heavily
> with Vlad's work.
>
> The interim solution I used for this merge was to revert back to RCU.
>
> Please take a look at what I did and send me followups because I am
> absolutely certain that some are necessary :-)))
Well, there is nothing fundamentally changed w.r.t. the race I have fixed.
In tcindex_delete(), after all Vlad's patches, it still calls tcf_queue_work()
therefore still races with call_rcu(). This has nothing to do with whether
it is RTNL or another mutex.
For the first memory leak, it is same, nothing changes, tcindex_delete()
still skips r->res.class==0 therefore we still need to fix it.
For the rest two memory leaks, it looks like you already merged the fix
(patch 3/3) correctly.
I will cherry-pick the first two patches for net-next and send them to you.
Thanks.
^ permalink raw reply
* Re: [PATCH bpf 1/2] bpf/test_run: fix unkillable BPF_PROG_TEST_RUN
From: Daniel Borkmann @ 2019-02-16 1:17 UTC (permalink / raw)
To: Stanislav Fomichev, netdev; +Cc: davem, ast, syzbot
In-Reply-To: <20190212234239.174386-1-sdf@google.com>
On 02/13/2019 12:42 AM, Stanislav Fomichev wrote:
> Syzbot found out that running BPF_PROG_TEST_RUN with repeat=0xffffffff
> makes process unkillable. The problem is that when CONFIG_PREEMPT is
> enabled, we never see need_resched() return true. This is due to the
> fact that preempt_enable() (which we do in bpf_test_run_one on each
> iteration) now handles resched if it's needed.
>
> Let's disable preemption for the whole run, not per test. In this case
> we can properly see whether resched is needed.
> Let's also properly return -EINTR to the userspace in case of a signal
> interrupt.
>
> See recent discussion:
> http://lore.kernel.org/netdev/CAH3MdRWHr4N8jei8jxDppXjmw-Nw=puNDLbu1dQOFQHxfU2onA@mail.gmail.com
>
> I'll follow up with the same fix bpf_prog_test_run_flow_dissector in
> bpf-next.
>
> Reported-by: syzbot <syzkaller@googlegroups.com>
> Signed-off-by: Stanislav Fomichev <sdf@google.com>
> ---
> net/bpf/test_run.c | 45 ++++++++++++++++++++++++---------------------
> 1 file changed, 24 insertions(+), 21 deletions(-)
>
> diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c
> index fa2644d276ef..e31e1b20f7f4 100644
> --- a/net/bpf/test_run.c
> +++ b/net/bpf/test_run.c
> @@ -13,27 +13,13 @@
> #include <net/sock.h>
> #include <net/tcp.h>
>
> -static __always_inline u32 bpf_test_run_one(struct bpf_prog *prog, void *ctx,
> - struct bpf_cgroup_storage *storage[MAX_BPF_CGROUP_STORAGE_TYPE])
> -{
> - u32 ret;
> -
> - preempt_disable();
> - rcu_read_lock();
> - bpf_cgroup_storage_set(storage);
> - ret = BPF_PROG_RUN(prog, ctx);
> - rcu_read_unlock();
> - preempt_enable();
> -
> - return ret;
> -}
> -
> -static int bpf_test_run(struct bpf_prog *prog, void *ctx, u32 repeat, u32 *ret,
> - u32 *time)
> +static int bpf_test_run(struct bpf_prog *prog, void *ctx, u32 repeat,
> + u32 *retval, u32 *time)
> {
> struct bpf_cgroup_storage *storage[MAX_BPF_CGROUP_STORAGE_TYPE] = { 0 };
> enum bpf_cgroup_storage_type stype;
> u64 time_start, time_spent = 0;
> + int ret = 0;
> u32 i;
>
> for_each_cgroup_storage_type(stype) {
> @@ -48,25 +34,42 @@ static int bpf_test_run(struct bpf_prog *prog, void *ctx, u32 repeat, u32 *ret,
>
> if (!repeat)
> repeat = 1;
> +
> + rcu_read_lock();
> + preempt_disable();
> time_start = ktime_get_ns();
> for (i = 0; i < repeat; i++) {
> - *ret = bpf_test_run_one(prog, ctx, storage);
> + bpf_cgroup_storage_set(storage);
> + *retval = BPF_PROG_RUN(prog, ctx);
> +
> + if (signal_pending(current)) {
> + ret = -EINTR;
> + break;
> + }
Wouldn't it be enough to just move the signal_pending() test to
the above as you did to actually fix the unkillable issue? For
CONFIG_PREEMPT the below need_resched() is never triggered as you
mention as preempt_enable() handles rescheduling internally in
this situation, so moving it only out should suffice.
The rationale for disabling preemption for the whole run is imho
a bit different, namely that you would not screw up the ktime
measurements due to rescheduling happening in between otherwise.
But then, once preemption is disabled for the whole run, is there
a need to move out the extra signal_pending() test (presumably as
need_resched() does not handle TIF_SIGPENDING but only TIF_NEED_RESCHED
but we still wouldn't get into a unkillable situation here, no)?
> if (need_resched()) {
> - if (signal_pending(current))
> - break;
> time_spent += ktime_get_ns() - time_start;
> + preempt_enable();
> + rcu_read_unlock();
> +
> cond_resched();
> +
> + rcu_read_lock();
> + preempt_disable();
> time_start = ktime_get_ns();
> }
> }
> time_spent += ktime_get_ns() - time_start;
> + preempt_enable();
> + rcu_read_unlock();
> +
> do_div(time_spent, repeat);
> *time = time_spent > U32_MAX ? U32_MAX : (u32)time_spent;
>
> for_each_cgroup_storage_type(stype)
> bpf_cgroup_storage_free(storage[stype]);
>
> - return 0;
> + return ret;
> }
>
> static int bpf_test_finish(const union bpf_attr *kattr,
>
^ permalink raw reply
* [PATCH net-next] net: sgi: use GFP_ATOMIC under spin lock
From: Wei Yongjun @ 2019-02-16 1:48 UTC (permalink / raw)
To: David S . Miller, Yang Wei, Luis Chamberlain, YueHaibing,
Christoph Hellwig
Cc: Wei Yongjun, netdev, kernel-janitors
The function meth_init_tx_ring() is called from meth_tx_timeout(),
in which spin_lock is held, so we should use GFP_ATOMIC instead.
Fixes: 8d4c28fbc284 ("meth: pass struct device to DMA API functions")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
---
drivers/net/ethernet/sgi/meth.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/sgi/meth.c b/drivers/net/ethernet/sgi/meth.c
index f425ab528224..f1271402ca21 100644
--- a/drivers/net/ethernet/sgi/meth.c
+++ b/drivers/net/ethernet/sgi/meth.c
@@ -214,7 +214,7 @@ static int meth_init_tx_ring(struct meth_private *priv)
{
/* Init TX ring */
priv->tx_ring = dma_alloc_coherent(&priv->pdev->dev,
- TX_RING_BUFFER_SIZE, &priv->tx_ring_dma, GFP_KERNEL);
+ TX_RING_BUFFER_SIZE, &priv->tx_ring_dma, GFP_ATOMIC);
if (!priv->tx_ring)
return -ENOMEM;
^ permalink raw reply related
* [pull request][net-next 00/13] Mellanox, BlueField SmartNIC
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Saeed Mahameed
Hi Dave,
This series adds the support for Melanox BlueField SmartNIC.
For more information please see tag log below.
Please note the merge commit of mlx5-next at the base of the pull request:
259fae5a2cff ("Merge branch 'mlx5-next' of git://git.kernel.org/.../mellanox/linux")
Please pull and let me know if there is any problem.
Thanks,
Saeed.
---
The following changes since commit 259fae5a2cff72e19f82094fb73e2149f8d64396:
Merge branch 'mlx5-next' of git://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux (2019-02-15 16:45:31 -0800)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-updates-2019-02-15
for you to fetch changes up to c96692fb8f3d0a161c6892ab9cc51a5e9992ccf2:
net/mlx5: E-Switch, Allow transition to offloads mode for ECPF (2019-02-15 17:25:58 -0800)
----------------------------------------------------------------
Support Mellanox BlueField SmartNIC (mlx5-updates-2019-02-15)
Bodong Wang says,
BlueField device is a multi-core ARM processor in a highly integrated
system on chip coupled with the ConnectX interconnect controller.
BlueField device can be presented in one out of two modes:
- SEPARATED_HOST: ARM processors as a separated and orthogonal host
like any other external host in the multi-host virtualization model.
- EMBEDDED_CPU: ARM processors as Embedded CPU (EC) and part of the
external hosts virtualization model.
While existing driver already supports the device on separated_host
mode, this patch series focus on the functionalities of embedded_cpu
mode.
On embedded_cpu mode, BlueField device exposes regular network
controller PCI function in the BlueField host(e.g, x86). However, a
separate PCI function called Embedded CPU Physical Function(ECPF) is
also added to the ARM host side, where standard Linux distributions is
able to run on the ARM cores. Depends on the NV configuration from
firmware, ECPF can be the e-switch manager and firmware pages supplier.
If ECPF is configured as e-switch manager and page supplier, it will
take over the responsibilities from the PF on BlueField host includes:
- Owns, controls and manages all e-switch parts, and takes e-switch
traffic by default. It also should perform ENABLE_HCA for the host
PF just like a PF does for its VFs.
- Provides and manages the ICM host memory required for the HCA to
store various contexts for itself, the PF and VFs belong the
e-switch it manages.
The PF on BlueField host side is still responsible for:
- Control its own permanent MAC.
- PCI and SRIOV configurations and perform ENABLE_HCA for its VFs.
The ECPF can also retrieve information about the external host it
controls, like host identifier, PCI BDF and number of virtual functions.
As these parameters may be changed dynamically, an event will be triggered
to the driver on ECPF side.
----------------------------------------------------------------
Bodong Wang (13):
net/mlx5: Correctly set LAG mode for ECPF
net/mlx5: E-Switch, Properly refer to the esw manager vport
net/mlx5: E-Switch, Properly refer to host PF vport as other vport
net/mlx5: E-Switch, Refactor offloads flow steering init/cleanup
net/mlx5: E-Switch, Split VF and special vports for offloads mode
net/mlx5: E-Switch, Use getter and iterator to access vport/rep
net/mlx5: E-Switch, Add state to eswitch vport representors
net/mlx5: E-Switch, Support load/unload reps of specific vport types
net/mlx5: E-Switch, Centralize repersentor reg/unreg to eswitch driver
net/mlx5: E-Switch, Assign a different position for uplink rep and vport
net/mlx5: E-Switch, Consider ECPF vport depends on eswitch ownership
net/mlx5: E-Switch, Load/unload VF reps according to event from host PF
net/mlx5: E-Switch, Allow transition to offloads mode for ECPF
drivers/infiniband/hw/mlx5/ib_rep.c | 20 +-
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 25 +-
drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 175 +++++---
drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 65 ++-
.../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 470 +++++++++++++++++----
drivers/net/ethernet/mellanox/mlx5/core/lag.c | 5 +
drivers/net/ethernet/mellanox/mlx5/core/vport.c | 10 +-
include/linux/mlx5/driver.h | 5 +
include/linux/mlx5/eswitch.h | 19 +-
include/linux/mlx5/mlx5_ifc.h | 3 +-
include/linux/mlx5/vport.h | 20 +-
11 files changed, 617 insertions(+), 200 deletions(-)
^ permalink raw reply
* [net-next 01/13] net/mlx5: Correctly set LAG mode for ECPF
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Bodong Wang, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
When bonding is added, driver assumes that it's RoCE LAG if no VF is
enabled. This is not enough for ECPF as the VF is enabled in host PF
side. LAG should only choose RoCE mode when both slave devices meet
conditions below:
1. E-Switch offloads mode is NONE.
2. No VF is enabled.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/lag.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
index 2d223385dc81..04c5aca7f8c5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
@@ -343,6 +343,11 @@ static void mlx5_do_bond(struct mlx5_lag *ldev)
roce_lag = !mlx5_sriov_is_enabled(dev0) &&
!mlx5_sriov_is_enabled(dev1);
+#ifdef CONFIG_MLX5_ESWITCH
+ roce_lag &= dev0->priv.eswitch->mode == SRIOV_NONE &&
+ dev1->priv.eswitch->mode == SRIOV_NONE;
+#endif
+
if (roce_lag)
mlx5_lag_remove_ib_devices(ldev);
--
2.20.1
^ permalink raw reply related
* [net-next 03/13] net/mlx5: E-Switch, Properly refer to host PF vport as other vport
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Bodong Wang, Eli Cohen, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
Commands referring to vports use the following scheme:
1. When referring to my own vport, put 0 in vport and 0 in other_vport.
2. When referring to another vport, put the vport number of the
referred vport and put 1 in other_vport. It was assumed that driver
is accessing other vport when vport number is greater than 0.
With the above scheme, the case that ECPF eswitch manager is trying
to access host PF vport will fall over with scheme 1 as the vport
number is 0. This is apparently wrong as driver is trying to refer
other vport.
As such usage can only happen in the eswitch context, change relevant
functions to provide other vport input properly.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 6 ++++--
drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 13 ++++++-------
drivers/net/ethernet/mellanox/mlx5/core/vport.c | 10 ++++------
include/linux/mlx5/vport.h | 4 ++--
4 files changed, 16 insertions(+), 17 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index 685f1975be58..f84889bbe2a0 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -1083,7 +1083,8 @@ static int mlx5e_vf_rep_open(struct net_device *dev)
if (!mlx5_modify_vport_admin_state(priv->mdev,
MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
- rep->vport, MLX5_VPORT_ADMIN_STATE_UP))
+ rep->vport, 1,
+ MLX5_VPORT_ADMIN_STATE_UP))
netif_carrier_on(dev);
unlock:
@@ -1101,7 +1102,8 @@ static int mlx5e_vf_rep_close(struct net_device *dev)
mutex_lock(&priv->state_lock);
mlx5_modify_vport_admin_state(priv->mdev,
MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
- rep->vport, MLX5_VPORT_ADMIN_STATE_DOWN);
+ rep->vport, 1,
+ MLX5_VPORT_ADMIN_STATE_DOWN);
ret = mlx5e_close_locked(dev);
mutex_unlock(&priv->state_lock);
return ret;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 9c622749dbde..648c743cc947 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1462,7 +1462,7 @@ static void esw_apply_vport_conf(struct mlx5_eswitch *esw,
mlx5_modify_vport_admin_state(esw->dev,
MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
- vport_num,
+ vport_num, 1,
vport->info.link_state);
/* Host PF has its own mac/guid. */
@@ -1581,10 +1581,10 @@ static void esw_disable_vport(struct mlx5_eswitch *esw, int vport_num)
esw_vport_change_handle_locked(vport);
vport->enabled_events = 0;
esw_vport_disable_qos(esw, vport_num);
- if (vport_num && esw->mode == SRIOV_LEGACY) {
+ if (esw->mode == SRIOV_LEGACY) {
mlx5_modify_vport_admin_state(esw->dev,
MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
- vport_num,
+ vport_num, 1,
MLX5_VPORT_ADMIN_STATE_DOWN);
esw_vport_disable_egress_acl(esw, vport);
esw_vport_disable_ingress_acl(esw, vport);
@@ -1875,7 +1875,7 @@ int mlx5_eswitch_set_vport_state(struct mlx5_eswitch *esw,
err = mlx5_modify_vport_admin_state(esw->dev,
MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
- vport, link_state);
+ vport, 1, link_state);
if (err) {
mlx5_core_warn(esw->dev,
"Failed to set vport %d link state, err = %d",
@@ -2137,7 +2137,7 @@ static int mlx5_eswitch_query_vport_drop_stats(struct mlx5_core_dev *dev,
!MLX5_CAP_GEN(dev, transmit_discard_vport_down))
return 0;
- err = mlx5_query_vport_down_stats(dev, vport_idx,
+ err = mlx5_query_vport_down_stats(dev, vport_idx, 1,
&rx_discard_vport_down,
&tx_discard_vport_down);
if (err)
@@ -2174,8 +2174,7 @@ int mlx5_eswitch_get_vport_stats(struct mlx5_eswitch *esw,
MLX5_CMD_OP_QUERY_VPORT_COUNTER);
MLX5_SET(query_vport_counter_in, in, op_mod, 0);
MLX5_SET(query_vport_counter_in, in, vport_number, vport);
- if (vport)
- MLX5_SET(query_vport_counter_in, in, other_vport, 1);
+ MLX5_SET(query_vport_counter_in, in, other_vport, 1);
memset(out, 0, outlen);
err = mlx5_cmd_exec(esw->dev, in, sizeof(in), out, outlen);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/vport.c b/drivers/net/ethernet/mellanox/mlx5/core/vport.c
index 9a928eb48522..ef95feca9961 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/vport.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/vport.c
@@ -64,7 +64,7 @@ u8 mlx5_query_vport_state(struct mlx5_core_dev *mdev, u8 opmod, u16 vport)
}
int mlx5_modify_vport_admin_state(struct mlx5_core_dev *mdev, u8 opmod,
- u16 vport, u8 state)
+ u16 vport, u8 other_vport, u8 state)
{
u32 in[MLX5_ST_SZ_DW(modify_vport_state_in)] = {0};
u32 out[MLX5_ST_SZ_DW(modify_vport_state_out)] = {0};
@@ -73,8 +73,7 @@ int mlx5_modify_vport_admin_state(struct mlx5_core_dev *mdev, u8 opmod,
MLX5_CMD_OP_MODIFY_VPORT_STATE);
MLX5_SET(modify_vport_state_in, in, op_mod, opmod);
MLX5_SET(modify_vport_state_in, in, vport_number, vport);
- if (vport)
- MLX5_SET(modify_vport_state_in, in, other_vport, 1);
+ MLX5_SET(modify_vport_state_in, in, other_vport, other_vport);
MLX5_SET(modify_vport_state_in, in, admin_state, state);
return mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
@@ -1057,7 +1056,7 @@ int mlx5_core_query_vport_counter(struct mlx5_core_dev *dev, u8 other_vport,
EXPORT_SYMBOL_GPL(mlx5_core_query_vport_counter);
int mlx5_query_vport_down_stats(struct mlx5_core_dev *mdev, u16 vport,
- u64 *rx_discard_vport_down,
+ u8 other_vport, u64 *rx_discard_vport_down,
u64 *tx_discard_vport_down)
{
u32 out[MLX5_ST_SZ_DW(query_vnic_env_out)] = {0};
@@ -1068,8 +1067,7 @@ int mlx5_query_vport_down_stats(struct mlx5_core_dev *mdev, u16 vport,
MLX5_CMD_OP_QUERY_VNIC_ENV);
MLX5_SET(query_vnic_env_in, in, op_mod, 0);
MLX5_SET(query_vnic_env_in, in, vport_number, vport);
- if (vport)
- MLX5_SET(query_vnic_env_in, in, other_vport, 1);
+ MLX5_SET(query_vnic_env_in, in, other_vport, other_vport);
err = mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
if (err)
diff --git a/include/linux/mlx5/vport.h b/include/linux/mlx5/vport.h
index b67bcc95ab5d..b7edcb1dadd8 100644
--- a/include/linux/mlx5/vport.h
+++ b/include/linux/mlx5/vport.h
@@ -59,7 +59,7 @@ enum {
u8 mlx5_query_vport_state(struct mlx5_core_dev *mdev, u8 opmod, u16 vport);
int mlx5_modify_vport_admin_state(struct mlx5_core_dev *mdev, u8 opmod,
- u16 vport, u8 state);
+ u16 vport, u8 other_vport, u8 state);
int mlx5_query_nic_vport_mac_address(struct mlx5_core_dev *mdev,
u16 vport, u8 *addr);
int mlx5_query_nic_vport_min_inline(struct mlx5_core_dev *mdev,
@@ -121,7 +121,7 @@ int mlx5_modify_nic_vport_vlans(struct mlx5_core_dev *dev,
int mlx5_nic_vport_enable_roce(struct mlx5_core_dev *mdev);
int mlx5_nic_vport_disable_roce(struct mlx5_core_dev *mdev);
int mlx5_query_vport_down_stats(struct mlx5_core_dev *mdev, u16 vport,
- u64 *rx_discard_vport_down,
+ u8 other_vport, u64 *rx_discard_vport_down,
u64 *tx_discard_vport_down);
int mlx5_core_query_vport_counter(struct mlx5_core_dev *dev, u8 other_vport,
int vf, u8 port_num, void *out,
--
2.20.1
^ permalink raw reply related
* [net-next 02/13] net/mlx5: E-Switch, Properly refer to the esw manager vport
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Bodong Wang, Eli Cohen, Or Gerlitz, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
In SmartNIC mode, the eswitch manager is not necessarily the PF
(vport 0). Use a helper function to get the correct eswitch manager
vport number and cache on the eswitch instance for fast reference.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/eswitch.c | 35 ++++++++++++-------
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 10 ++++++
.../mellanox/mlx5/core/eswitch_offloads.c | 7 ++--
include/linux/mlx5/vport.h | 2 ++
4 files changed, 39 insertions(+), 15 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 05830696abd8..9c622749dbde 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -378,16 +378,16 @@ static int esw_add_uc_addr(struct mlx5_eswitch *esw, struct vport_addr *vaddr)
u16 vport = vaddr->vport;
int err;
- /* Skip mlx5_mpfs_add_mac for PFs,
- * it is already done by the PF netdev in mlx5e_execute_l2_action
+ /* Skip mlx5_mpfs_add_mac for eswitch_managers,
+ * it is already done by its netdev in mlx5e_execute_l2_action
*/
- if (!vport)
+ if (esw->manager_vport == vport)
goto fdb_add;
err = mlx5_mpfs_add_mac(esw->dev, mac);
if (err) {
esw_warn(esw->dev,
- "Failed to add L2 table mac(%pM) for vport(%d), err(%d)\n",
+ "Failed to add L2 table mac(%pM) for vport(0x%x), err(%d)\n",
mac, vport, err);
return err;
}
@@ -410,10 +410,10 @@ static int esw_del_uc_addr(struct mlx5_eswitch *esw, struct vport_addr *vaddr)
u16 vport = vaddr->vport;
int err = 0;
- /* Skip mlx5_mpfs_del_mac for PFs,
- * it is already done by the PF netdev in mlx5e_execute_l2_action
+ /* Skip mlx5_mpfs_del_mac for eswitch managerss,
+ * it is already done by its netdev in mlx5e_execute_l2_action
*/
- if (!vport || !vaddr->mpfs)
+ if (!vaddr->mpfs || esw->manager_vport == vport)
goto fdb_del;
err = mlx5_mpfs_del_mac(esw->dev, mac);
@@ -1457,15 +1457,22 @@ static void esw_apply_vport_conf(struct mlx5_eswitch *esw,
{
int vport_num = vport->vport;
- if (!vport_num)
+ if (esw->manager_vport == vport_num)
return;
mlx5_modify_vport_admin_state(esw->dev,
MLX5_VPORT_STATE_OP_MOD_ESW_VPORT,
vport_num,
vport->info.link_state);
- mlx5_modify_nic_vport_mac_address(esw->dev, vport_num, vport->info.mac);
- mlx5_modify_nic_vport_node_guid(esw->dev, vport_num, vport->info.node_guid);
+
+ /* Host PF has its own mac/guid. */
+ if (vport_num) {
+ mlx5_modify_nic_vport_mac_address(esw->dev, vport_num,
+ vport->info.mac);
+ mlx5_modify_nic_vport_node_guid(esw->dev, vport_num,
+ vport->info.node_guid);
+ }
+
modify_esw_vport_cvlan(esw->dev, vport_num, vport->info.vlan, vport->info.qos,
(vport->info.vlan || vport->info.qos));
@@ -1537,8 +1544,11 @@ static void esw_enable_vport(struct mlx5_eswitch *esw, int vport_num,
vport->enabled_events = enable_events;
vport->enabled = true;
- /* only PF is trusted by default */
- if (!vport_num)
+ /* Esw manager is trusted by default. Host PF (vport 0) is trusted as well
+ * in smartNIC as it's a vport group manager.
+ */
+ if (esw->manager_vport == vport_num ||
+ (!vport_num && mlx5_core_is_ecpf(esw->dev)))
vport->info.trusted = true;
esw_vport_change_handle_locked(vport);
@@ -1733,6 +1743,7 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
return -ENOMEM;
esw->dev = dev;
+ esw->manager_vport = mlx5_eswitch_manager_vport(dev);
esw->work_queue = create_singlethread_workqueue("mlx5_esw_wq");
if (!esw->work_queue) {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index 0a3eee8746c1..959a9e28d08f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -38,6 +38,7 @@
#include <net/devlink.h>
#include <linux/mlx5/device.h>
#include <linux/mlx5/eswitch.h>
+#include <linux/mlx5/vport.h>
#include <linux/mlx5/fs.h>
#include "lib/mpfs.h"
@@ -204,6 +205,7 @@ struct mlx5_eswitch {
struct mlx5_esw_offload offloads;
int mode;
int nvports;
+ u16 manager_vport;
};
void esw_offloads_cleanup(struct mlx5_eswitch *esw, int nvports);
@@ -363,6 +365,14 @@ bool mlx5_esw_lag_prereq(struct mlx5_core_dev *dev0,
#define esw_debug(dev, format, ...) \
mlx5_core_dbg_mask(dev, MLX5_DEBUG_ESWITCH_MASK, format, ##__VA_ARGS__)
+
+/* The returned number is valid only when the dev is eswitch manager. */
+static inline u16 mlx5_eswitch_manager_vport(struct mlx5_core_dev *dev)
+{
+ return mlx5_core_is_ecpf_esw_manager(dev) ?
+ MLX5_VPORT_ECPF : MLX5_VPORT_PF;
+}
+
#else /* CONFIG_MLX5_ESWITCH */
/* eswitch API stubs */
static inline int mlx5_eswitch_init(struct mlx5_core_dev *dev) { return 0; }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 9128b45f3f37..af2c44d31357 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -522,7 +522,8 @@ mlx5_eswitch_add_send_to_vport_rule(struct mlx5_eswitch *esw, int vport, u32 sqn
misc = MLX5_ADDR_OF(fte_match_param, spec->match_value, misc_parameters);
MLX5_SET(fte_match_set_misc, misc, source_sqn, sqn);
- MLX5_SET(fte_match_set_misc, misc, source_port, 0x0); /* source vport is 0 */
+ /* source vport is the esw manager */
+ MLX5_SET(fte_match_set_misc, misc, source_port, esw->manager_vport);
misc = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, misc_parameters);
MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_sqn);
@@ -567,7 +568,7 @@ static void peer_miss_rules_setup(struct mlx5_core_dev *peer_dev,
source_eswitch_owner_vhca_id);
dest->type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
- dest->vport.num = 0;
+ dest->vport.num = peer_dev->priv.eswitch->manager_vport;
dest->vport.vhca_id = MLX5_CAP_GEN(peer_dev, vhca_id);
dest->vport.flags |= MLX5_FLOW_DEST_VPORT_VHCA_ID;
}
@@ -666,7 +667,7 @@ static int esw_add_fdb_miss_rule(struct mlx5_eswitch *esw)
dmac_c[0] = 0x01;
dest.type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
- dest.vport.num = 0;
+ dest.vport.num = esw->manager_vport;
flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb, spec,
diff --git a/include/linux/mlx5/vport.h b/include/linux/mlx5/vport.h
index 3bc05449ac39..b67bcc95ab5d 100644
--- a/include/linux/mlx5/vport.h
+++ b/include/linux/mlx5/vport.h
@@ -52,6 +52,8 @@ enum {
};
enum {
+ MLX5_VPORT_PF = 0x0,
+ MLX5_VPORT_ECPF = 0xfffe,
MLX5_VPORT_UPLINK = 0xffff
};
--
2.20.1
^ permalink raw reply related
* [net-next 05/13] net/mlx5: E-Switch, Split VF and special vports for offloads mode
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Bodong Wang, Or Gerlitz, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
When driver is entering offloads mode, there are two major tasks to
do: initialize flow steering and create representors. Flow steering
should make sure enough flow table/group spaces are reserved for all
reps. Representors will be created in a group, all or none.
With the introduction of ECPF, flow steering should still reserve the
same spaces. But, the representors are not always loaded/unloaded in a
single piece. Once ECPF is in offloads mode, it will get the number
of VF changing event from host PF. In such scenario, only the VF reps
should be loaded/unloaded, not the reps for special vports (such as
the uplink vport).
Thus, when entering offloads mode, driver should specify the total
number of reps, and the number of VF reps separately. When leaving
offloads mode, the cleanup should use the information self-contained
in eswitch such as number of VFs.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/eswitch.c | 7 +--
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 5 +-
.../mellanox/mlx5/core/eswitch_offloads.c | 57 +++++++++++++------
include/linux/mlx5/vport.h | 1 +
4 files changed, 48 insertions(+), 22 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 648c743cc947..be6c2931d2a0 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1641,7 +1641,8 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode)
} else {
mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_ETH);
mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
- err = esw_offloads_init(esw, nvfs + MLX5_SPECIAL_VPORTS);
+ err = esw_offloads_init(esw, nvfs,
+ nvfs + MLX5_SPECIAL_VPORTS);
}
if (err)
@@ -1683,7 +1684,6 @@ void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw)
{
struct esw_mc_addr *mc_promisc;
int old_mode;
- int nvports;
int i;
if (!ESW_ALLOWED(esw) || esw->mode == SRIOV_NONE)
@@ -1693,7 +1693,6 @@ void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw)
esw->enabled_vports, esw->mode);
mc_promisc = &esw->mc_promisc;
- nvports = esw->enabled_vports;
if (esw->mode == SRIOV_LEGACY)
mlx5_eq_notifier_unregister(esw->dev, &esw->nb);
@@ -1709,7 +1708,7 @@ void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw)
if (esw->mode == SRIOV_LEGACY)
esw_destroy_legacy_fdb_table(esw);
else if (esw->mode == SRIOV_OFFLOADS)
- esw_offloads_cleanup(esw, nvports);
+ esw_offloads_cleanup(esw);
old_mode = esw->mode;
esw->mode = SRIOV_NONE;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index 959a9e28d08f..fd845e6c44d5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -208,8 +208,9 @@ struct mlx5_eswitch {
u16 manager_vport;
};
-void esw_offloads_cleanup(struct mlx5_eswitch *esw, int nvports);
-int esw_offloads_init(struct mlx5_eswitch *esw, int nvports);
+void esw_offloads_cleanup(struct mlx5_eswitch *esw);
+int esw_offloads_init(struct mlx5_eswitch *esw, int vf_nvports,
+ int total_nvports);
void esw_offloads_cleanup_reps(struct mlx5_eswitch *esw);
int esw_offloads_init_reps(struct mlx5_eswitch *esw);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 19969d487a01..14f7ad67cfe4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -54,6 +54,8 @@ enum {
#define fdb_prio_table(esw, chain, prio, level) \
(esw)->fdb_table.offloads.fdb_prio[(chain)][(prio)][(level)]
+#define UPLINK_REP_INDEX 0
+
static struct mlx5_flow_table *
esw_get_prio_table(struct mlx5_eswitch *esw, u32 chain, u16 prio, int level);
static void
@@ -1239,19 +1241,28 @@ int esw_offloads_init_reps(struct mlx5_eswitch *esw)
return 0;
}
+static void __esw_offloads_unload_rep(struct mlx5_eswitch *esw,
+ struct mlx5_eswitch_rep *rep, u8 rep_type)
+{
+ if (!rep->rep_if[rep_type].valid)
+ return;
+
+ rep->rep_if[rep_type].unload(rep);
+}
+
static void esw_offloads_unload_reps_type(struct mlx5_eswitch *esw, int nvports,
u8 rep_type)
{
struct mlx5_eswitch_rep *rep;
int vport;
- for (vport = nvports - 1; vport >= 0; vport--) {
+ for (vport = nvports; vport >= MLX5_VPORT_FIRST_VF; vport--) {
rep = &esw->offloads.vport_reps[vport];
- if (!rep->rep_if[rep_type].valid)
- continue;
-
- rep->rep_if[rep_type].unload(rep);
+ __esw_offloads_unload_rep(esw, rep, rep_type);
}
+
+ rep = &esw->offloads.vport_reps[UPLINK_REP_INDEX];
+ __esw_offloads_unload_rep(esw, rep, rep_type);
}
static void esw_offloads_unload_reps(struct mlx5_eswitch *esw, int nvports)
@@ -1262,6 +1273,15 @@ static void esw_offloads_unload_reps(struct mlx5_eswitch *esw, int nvports)
esw_offloads_unload_reps_type(esw, nvports, rep_type);
}
+static int __esw_offloads_load_rep(struct mlx5_eswitch *esw,
+ struct mlx5_eswitch_rep *rep, u8 rep_type)
+{
+ if (!rep->rep_if[rep_type].valid)
+ return 0;
+
+ return rep->rep_if[rep_type].load(esw->dev, rep);
+}
+
static int esw_offloads_load_reps_type(struct mlx5_eswitch *esw, int nvports,
u8 rep_type)
{
@@ -1269,12 +1289,14 @@ static int esw_offloads_load_reps_type(struct mlx5_eswitch *esw, int nvports,
int vport;
int err;
- for (vport = 0; vport < nvports; vport++) {
- rep = &esw->offloads.vport_reps[vport];
- if (!rep->rep_if[rep_type].valid)
- continue;
+ rep = &esw->offloads.vport_reps[UPLINK_REP_INDEX];
+ err = __esw_offloads_load_rep(esw, rep, rep_type);
+ if (err)
+ goto out;
- err = rep->rep_if[rep_type].load(esw->dev, rep);
+ for (vport = MLX5_VPORT_FIRST_VF; vport <= nvports; vport++) {
+ rep = &esw->offloads.vport_reps[vport];
+ err = __esw_offloads_load_rep(esw, rep, rep_type);
if (err)
goto err_reps;
}
@@ -1283,6 +1305,7 @@ static int esw_offloads_load_reps_type(struct mlx5_eswitch *esw, int nvports,
err_reps:
esw_offloads_unload_reps_type(esw, vport, rep_type);
+out:
return err;
}
@@ -1440,17 +1463,18 @@ static void esw_offloads_steering_cleanup(struct mlx5_eswitch *esw)
esw_destroy_offloads_fdb_tables(esw);
}
-int esw_offloads_init(struct mlx5_eswitch *esw, int nvports)
+int esw_offloads_init(struct mlx5_eswitch *esw, int vf_nvports,
+ int total_nvports)
{
int err;
mutex_init(&esw->fdb_table.offloads.fdb_prio_lock);
- err = esw_offloads_steering_init(esw, nvports);
+ err = esw_offloads_steering_init(esw, total_nvports);
if (err)
return err;
- err = esw_offloads_load_reps(esw, nvports);
+ err = esw_offloads_load_reps(esw, vf_nvports);
if (err)
goto err_reps;
@@ -1481,10 +1505,12 @@ static int esw_offloads_stop(struct mlx5_eswitch *esw,
return err;
}
-void esw_offloads_cleanup(struct mlx5_eswitch *esw, int nvports)
+void esw_offloads_cleanup(struct mlx5_eswitch *esw)
{
+ u16 num_vfs = esw->dev->priv.sriov.num_vfs;
+
esw_offloads_devcom_cleanup(esw);
- esw_offloads_unload_reps(esw, nvports);
+ esw_offloads_unload_reps(esw, num_vfs);
esw_offloads_steering_cleanup(esw);
}
@@ -1822,7 +1848,6 @@ EXPORT_SYMBOL(mlx5_eswitch_unregister_vport_rep);
void *mlx5_eswitch_get_uplink_priv(struct mlx5_eswitch *esw, u8 rep_type)
{
-#define UPLINK_REP_INDEX 0
struct mlx5_esw_offload *offloads = &esw->offloads;
struct mlx5_eswitch_rep *rep;
diff --git a/include/linux/mlx5/vport.h b/include/linux/mlx5/vport.h
index b7edcb1dadd8..755aeea19e1c 100644
--- a/include/linux/mlx5/vport.h
+++ b/include/linux/mlx5/vport.h
@@ -53,6 +53,7 @@ enum {
enum {
MLX5_VPORT_PF = 0x0,
+ MLX5_VPORT_FIRST_VF = 0x1,
MLX5_VPORT_ECPF = 0xfffe,
MLX5_VPORT_UPLINK = 0xffff
};
--
2.20.1
^ permalink raw reply related
* [net-next 04/13] net/mlx5: E-Switch, Refactor offloads flow steering init/cleanup
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Bodong Wang, Or Gerlitz, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
E-switch offloads mode initialize/cleanup multiple steering related
entities (flow table/group). Refactor these operations to internal
helper functions for better block design.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../mellanox/mlx5/core/eswitch_offloads.c | 43 +++++++++++++------
1 file changed, 31 insertions(+), 12 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index af2c44d31357..19969d487a01 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -1404,7 +1404,7 @@ static void esw_offloads_devcom_cleanup(struct mlx5_eswitch *esw)
mlx5_devcom_unregister_component(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
}
-int esw_offloads_init(struct mlx5_eswitch *esw, int nvports)
+static int esw_offloads_steering_init(struct mlx5_eswitch *esw, int nvports)
{
int err;
@@ -1422,16 +1422,8 @@ int esw_offloads_init(struct mlx5_eswitch *esw, int nvports)
if (err)
goto create_fg_err;
- err = esw_offloads_load_reps(esw, nvports);
- if (err)
- goto err_reps;
-
- esw_offloads_devcom_init(esw);
return 0;
-err_reps:
- esw_destroy_vport_rx_group(esw);
-
create_fg_err:
esw_destroy_offloads_table(esw);
@@ -1441,6 +1433,35 @@ int esw_offloads_init(struct mlx5_eswitch *esw, int nvports)
return err;
}
+static void esw_offloads_steering_cleanup(struct mlx5_eswitch *esw)
+{
+ esw_destroy_vport_rx_group(esw);
+ esw_destroy_offloads_table(esw);
+ esw_destroy_offloads_fdb_tables(esw);
+}
+
+int esw_offloads_init(struct mlx5_eswitch *esw, int nvports)
+{
+ int err;
+
+ mutex_init(&esw->fdb_table.offloads.fdb_prio_lock);
+
+ err = esw_offloads_steering_init(esw, nvports);
+ if (err)
+ return err;
+
+ err = esw_offloads_load_reps(esw, nvports);
+ if (err)
+ goto err_reps;
+
+ esw_offloads_devcom_init(esw);
+ return 0;
+
+err_reps:
+ esw_offloads_steering_cleanup(esw);
+ return err;
+}
+
static int esw_offloads_stop(struct mlx5_eswitch *esw,
struct netlink_ext_ack *extack)
{
@@ -1464,9 +1485,7 @@ void esw_offloads_cleanup(struct mlx5_eswitch *esw, int nvports)
{
esw_offloads_devcom_cleanup(esw);
esw_offloads_unload_reps(esw, nvports);
- esw_destroy_vport_rx_group(esw);
- esw_destroy_offloads_table(esw);
- esw_destroy_offloads_fdb_tables(esw);
+ esw_offloads_steering_cleanup(esw);
}
static int esw_mode_from_devlink(u16 mode, u16 *mlx5_mode)
--
2.20.1
^ permalink raw reply related
* [net-next 07/13] net/mlx5: E-Switch, Add state to eswitch vport representors
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Bodong Wang, Or Gerlitz, Mark Bloch, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
Currently the eswitch vport reps have a valid indicator, which is
set on register and unset on unregister. However, a rep can be loaded
or not loaded when doing unregister, current driver checks if the
vport of that rep is enabled as a flag to imply the rep is loaded.
However, for ECPF, this is not valid as the host PF will enable the
vports for its VFs instead.
Add three states: {unregistered, registered, loaded}, with the
following state changes across different operations:
create: (none) -> unregistered
reg: unregistered -> registered
load: registered -> loaded
unload: loaded -> registered
unreg: registered -> unregistered
Note that the state shall only be updated inside eswitch driver rather
than individual drivers such as ETH or IB.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Suggested-by: Mark Bloch <markb@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../mellanox/mlx5/core/eswitch_offloads.c | 31 +++++++++++++------
include/linux/mlx5/eswitch.h | 8 ++++-
2 files changed, 29 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 4979c7ee0ad7..c6c9dad69ba8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -364,7 +364,7 @@ static int esw_set_global_vlan_pop(struct mlx5_eswitch *esw, u8 val)
esw_debug(esw->dev, "%s applying global %s policy\n", __func__, val ? "pop" : "none");
for (vf_vport = 1; vf_vport < esw->enabled_vports; vf_vport++) {
rep = &esw->offloads.vport_reps[vf_vport];
- if (!rep->rep_if[REP_ETH].valid)
+ if (rep->rep_if[REP_ETH].state != REP_LOADED)
continue;
err = __mlx5_eswitch_set_vport_vlan(esw, rep->vport, 0, 0, val);
@@ -1256,7 +1256,7 @@ int esw_offloads_init_reps(struct mlx5_eswitch *esw)
struct mlx5_core_dev *dev = esw->dev;
struct mlx5_esw_offload *offloads;
struct mlx5_eswitch_rep *rep;
- u8 hw_id[ETH_ALEN];
+ u8 hw_id[ETH_ALEN], rep_type;
int vport;
esw->offloads.vport_reps = kcalloc(total_vfs,
@@ -1271,6 +1271,9 @@ int esw_offloads_init_reps(struct mlx5_eswitch *esw)
mlx5_esw_for_all_reps(esw, vport, rep) {
rep->vport = vport;
ether_addr_copy(rep->hw_id, hw_id);
+
+ for (rep_type = 0; rep_type < NUM_REP_TYPES; rep_type++)
+ rep->rep_if[rep_type].state = REP_UNREGISTERED;
}
offloads->vport_reps[0].vport = MLX5_VPORT_UPLINK;
@@ -1281,10 +1284,11 @@ int esw_offloads_init_reps(struct mlx5_eswitch *esw)
static void __esw_offloads_unload_rep(struct mlx5_eswitch *esw,
struct mlx5_eswitch_rep *rep, u8 rep_type)
{
- if (!rep->rep_if[rep_type].valid)
+ if (rep->rep_if[rep_type].state != REP_LOADED)
return;
rep->rep_if[rep_type].unload(rep);
+ rep->rep_if[rep_type].state = REP_REGISTERED;
}
static void esw_offloads_unload_reps_type(struct mlx5_eswitch *esw, int nvports,
@@ -1311,10 +1315,18 @@ static void esw_offloads_unload_reps(struct mlx5_eswitch *esw, int nvports)
static int __esw_offloads_load_rep(struct mlx5_eswitch *esw,
struct mlx5_eswitch_rep *rep, u8 rep_type)
{
- if (!rep->rep_if[rep_type].valid)
+ int err = 0;
+
+ if (rep->rep_if[rep_type].state != REP_REGISTERED)
return 0;
- return rep->rep_if[rep_type].load(esw->dev, rep);
+ err = rep->rep_if[rep_type].load(esw->dev, rep);
+ if (err)
+ return err;
+
+ rep->rep_if[rep_type].state = REP_LOADED;
+
+ return 0;
}
static int esw_offloads_load_reps_type(struct mlx5_eswitch *esw, int nvports,
@@ -1861,7 +1873,7 @@ void mlx5_eswitch_register_vport_rep(struct mlx5_eswitch *esw,
rep_if->get_proto_dev = __rep_if->get_proto_dev;
rep_if->priv = __rep_if->priv;
- rep_if->valid = true;
+ rep_if->state = REP_REGISTERED;
}
EXPORT_SYMBOL(mlx5_eswitch_register_vport_rep);
@@ -1873,10 +1885,11 @@ void mlx5_eswitch_unregister_vport_rep(struct mlx5_eswitch *esw,
rep = &offloads->vport_reps[vport_index];
- if (esw->mode == SRIOV_OFFLOADS && esw->vports[vport_index].enabled)
+ if (esw->mode == SRIOV_OFFLOADS &&
+ rep->rep_if[rep_type].state == REP_LOADED)
rep->rep_if[rep_type].unload(rep);
- rep->rep_if[rep_type].valid = false;
+ rep->rep_if[rep_type].state = REP_UNREGISTERED;
}
EXPORT_SYMBOL(mlx5_eswitch_unregister_vport_rep);
@@ -1896,7 +1909,7 @@ void *mlx5_eswitch_get_proto_dev(struct mlx5_eswitch *esw,
rep = mlx5_eswitch_get_rep(esw, vport);
- if (rep->rep_if[rep_type].valid &&
+ if (rep->rep_if[rep_type].state == REP_LOADED &&
rep->rep_if[rep_type].get_proto_dev)
return rep->rep_if[rep_type].get_proto_dev(rep);
return NULL;
diff --git a/include/linux/mlx5/eswitch.h b/include/linux/mlx5/eswitch.h
index fab5121ffb8f..e3dbc1bc0917 100644
--- a/include/linux/mlx5/eswitch.h
+++ b/include/linux/mlx5/eswitch.h
@@ -22,6 +22,12 @@ enum {
NUM_REP_TYPES,
};
+enum {
+ REP_UNREGISTERED,
+ REP_REGISTERED,
+ REP_LOADED,
+};
+
struct mlx5_eswitch_rep;
struct mlx5_eswitch_rep_if {
int (*load)(struct mlx5_core_dev *dev,
@@ -29,7 +35,7 @@ struct mlx5_eswitch_rep_if {
void (*unload)(struct mlx5_eswitch_rep *rep);
void *(*get_proto_dev)(struct mlx5_eswitch_rep *rep);
void *priv;
- bool valid;
+ u8 state;
};
struct mlx5_eswitch_rep {
--
2.20.1
^ permalink raw reply related
* [net-next 06/13] net/mlx5: E-Switch, Use getter and iterator to access vport/rep
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Bodong Wang, Saeed Mahameed, Or Gerlitz
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
With only PF and VF, it is sufficient to have the vport/rep array
index as the vport number. This is because PF and VF vports numbers
are consecutive serial numbers. In downstream patches with
introducing of ECPF and UPLINK vports, it's not consecutive any more.
Use getter to get specific vport/rep, and use iterator to traversal
a list of vport/rep. This hides the translation between array index
and vport number, and provides flexibility of using different
translation mechanism in the future.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Suggested-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/eswitch.c | 94 ++++++++++++-------
.../mellanox/mlx5/core/eswitch_offloads.c | 75 ++++++++++-----
2 files changed, 113 insertions(+), 56 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index be6c2931d2a0..d1454f18c0a7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -68,6 +68,26 @@ enum {
MC_ADDR_CHANGE | \
PROMISC_CHANGE)
+/* The vport getter/iterator are only valid after esw->total_vports
+ * and vport->vport are initialized in mlx5_eswitch_init.
+ */
+#define mlx5_esw_for_all_vports(esw, i, vport) \
+ for ((i) = MLX5_VPORT_PF; \
+ (vport) = &(esw)->vports[i], \
+ (i) < (esw)->total_vports; (i)++)
+
+#define mlx5_esw_for_each_vf_vport(esw, i, vport, nvfs) \
+ for ((i) = MLX5_VPORT_FIRST_VF; \
+ (vport) = &(esw)->vports[i], \
+ (i) <= (nvfs); (i)++)
+
+static struct mlx5_vport *mlx5_eswitch_get_vport(struct mlx5_eswitch *esw,
+ u16 vport_num)
+{
+ WARN_ON(vport_num > esw->total_vports - 1);
+ return &esw->vports[vport_num];
+}
+
static int arm_vport_context_events_cmd(struct mlx5_core_dev *dev, u16 vport,
u32 events_mask)
{
@@ -436,17 +456,18 @@ static void update_allmulti_vports(struct mlx5_eswitch *esw,
struct esw_mc_addr *esw_mc)
{
u8 *mac = vaddr->node.addr;
- u16 vport_idx = 0;
+ struct mlx5_vport *vport;
+ u16 i, vport_num;
- for (vport_idx = 0; vport_idx < esw->total_vports; vport_idx++) {
- struct mlx5_vport *vport = &esw->vports[vport_idx];
+ mlx5_esw_for_all_vports(esw, i, vport) {
struct hlist_head *vport_hash = vport->mc_list;
struct vport_addr *iter_vaddr =
l2addr_hash_find(vport_hash,
mac,
struct vport_addr);
+ vport_num = vport->vport;
if (IS_ERR_OR_NULL(vport->allmulti_rule) ||
- vaddr->vport == vport_idx)
+ vaddr->vport == vport_num)
continue;
switch (vaddr->action) {
case MLX5_ACTION_ADD:
@@ -458,14 +479,14 @@ static void update_allmulti_vports(struct mlx5_eswitch *esw,
if (!iter_vaddr) {
esw_warn(esw->dev,
"ALL-MULTI: Failed to add MAC(%pM) to vport[%d] DB\n",
- mac, vport_idx);
+ mac, vport_num);
continue;
}
- iter_vaddr->vport = vport_idx;
+ iter_vaddr->vport = vport_num;
iter_vaddr->flow_rule =
esw_fdb_set_vport_rule(esw,
mac,
- vport_idx);
+ vport_num);
iter_vaddr->mc_promisc = true;
break;
case MLX5_ACTION_DEL:
@@ -564,7 +585,7 @@ static int esw_del_mc_addr(struct mlx5_eswitch *esw, struct vport_addr *vaddr)
static void esw_apply_vport_addr_list(struct mlx5_eswitch *esw,
u16 vport_num, int list_type)
{
- struct mlx5_vport *vport = &esw->vports[vport_num];
+ struct mlx5_vport *vport = mlx5_eswitch_get_vport(esw, vport_num);
bool is_uc = list_type == MLX5_NVPRT_LIST_TYPE_UC;
vport_addr_action vport_addr_add;
vport_addr_action vport_addr_del;
@@ -599,7 +620,7 @@ static void esw_apply_vport_addr_list(struct mlx5_eswitch *esw,
static void esw_update_vport_addr_list(struct mlx5_eswitch *esw,
u16 vport_num, int list_type)
{
- struct mlx5_vport *vport = &esw->vports[vport_num];
+ struct mlx5_vport *vport = mlx5_eswitch_get_vport(esw, vport_num);
bool is_uc = list_type == MLX5_NVPRT_LIST_TYPE_UC;
u8 (*mac_list)[ETH_ALEN];
struct l2addr_node *node;
@@ -686,7 +707,7 @@ static void esw_update_vport_addr_list(struct mlx5_eswitch *esw,
*/
static void esw_update_vport_mc_promisc(struct mlx5_eswitch *esw, u16 vport_num)
{
- struct mlx5_vport *vport = &esw->vports[vport_num];
+ struct mlx5_vport *vport = mlx5_eswitch_get_vport(esw, vport_num);
struct l2addr_node *node;
struct vport_addr *addr;
struct hlist_head *hash;
@@ -722,8 +743,8 @@ static void esw_update_vport_mc_promisc(struct mlx5_eswitch *esw, u16 vport_num)
static void esw_apply_vport_rx_mode(struct mlx5_eswitch *esw, u16 vport_num,
bool promisc, bool mc_promisc)
{
+ struct mlx5_vport *vport = mlx5_eswitch_get_vport(esw, vport_num);
struct esw_mc_addr *allmulti_addr = &esw->mc_promisc;
- struct mlx5_vport *vport = &esw->vports[vport_num];
if (IS_ERR_OR_NULL(vport->allmulti_rule) != mc_promisc)
goto promisc;
@@ -764,7 +785,7 @@ static void esw_apply_vport_rx_mode(struct mlx5_eswitch *esw, u16 vport_num,
/* Sync vport rx mode from vport context */
static void esw_update_vport_rx_mode(struct mlx5_eswitch *esw, u16 vport_num)
{
- struct mlx5_vport *vport = &esw->vports[vport_num];
+ struct mlx5_vport *vport = mlx5_eswitch_get_vport(esw, vport_num);
int promisc_all = 0;
int promisc_uc = 0;
int promisc_mc = 0;
@@ -1341,8 +1362,8 @@ static void esw_destroy_tsar(struct mlx5_eswitch *esw)
static int esw_vport_enable_qos(struct mlx5_eswitch *esw, int vport_num,
u32 initial_max_rate, u32 initial_bw_share)
{
+ struct mlx5_vport *vport = mlx5_eswitch_get_vport(esw, vport_num);
u32 sched_ctx[MLX5_ST_SZ_DW(scheduling_context)] = {0};
- struct mlx5_vport *vport = &esw->vports[vport_num];
struct mlx5_core_dev *dev = esw->dev;
void *vport_elem;
int err = 0;
@@ -1381,7 +1402,7 @@ static int esw_vport_enable_qos(struct mlx5_eswitch *esw, int vport_num,
static void esw_vport_disable_qos(struct mlx5_eswitch *esw, int vport_num)
{
- struct mlx5_vport *vport = &esw->vports[vport_num];
+ struct mlx5_vport *vport = mlx5_eswitch_get_vport(esw, vport_num);
int err = 0;
if (!vport->qos.enabled)
@@ -1400,8 +1421,8 @@ static void esw_vport_disable_qos(struct mlx5_eswitch *esw, int vport_num)
static int esw_vport_qos_config(struct mlx5_eswitch *esw, int vport_num,
u32 max_rate, u32 bw_share)
{
+ struct mlx5_vport *vport = mlx5_eswitch_get_vport(esw, vport_num);
u32 sched_ctx[MLX5_ST_SZ_DW(scheduling_context)] = {0};
- struct mlx5_vport *vport = &esw->vports[vport_num];
struct mlx5_core_dev *dev = esw->dev;
void *vport_elem;
u32 bitmask = 0;
@@ -1518,10 +1539,10 @@ static void esw_vport_destroy_drop_counters(struct mlx5_vport *vport)
mlx5_fc_destroy(dev, vport->egress.drop_counter);
}
-static void esw_enable_vport(struct mlx5_eswitch *esw, int vport_num,
+static void esw_enable_vport(struct mlx5_eswitch *esw, struct mlx5_vport *vport,
int enable_events)
{
- struct mlx5_vport *vport = &esw->vports[vport_num];
+ u16 vport_num = vport->vport;
mutex_lock(&esw->state_lock);
WARN_ON(vport->enabled);
@@ -1558,9 +1579,10 @@ static void esw_enable_vport(struct mlx5_eswitch *esw, int vport_num,
mutex_unlock(&esw->state_lock);
}
-static void esw_disable_vport(struct mlx5_eswitch *esw, int vport_num)
+static void esw_disable_vport(struct mlx5_eswitch *esw,
+ struct mlx5_vport *vport)
{
- struct mlx5_vport *vport = &esw->vports[vport_num];
+ u16 vport_num = vport->vport;
if (!vport->enabled)
return;
@@ -1603,7 +1625,7 @@ static int eswitch_vport_event(struct notifier_block *nb,
u16 vport_num;
vport_num = be16_to_cpu(eqe->data.vport_change.vport_num);
- vport = &esw->vports[vport_num];
+ vport = mlx5_eswitch_get_vport(esw, vport_num);
if (vport->enabled)
queue_work(esw->work_queue, &vport->vport_change_handler);
@@ -1615,6 +1637,7 @@ static int eswitch_vport_event(struct notifier_block *nb,
int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode)
{
+ struct mlx5_vport *vport;
int err;
int i, enabled_events;
@@ -1657,8 +1680,14 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode)
* 2. FDB/Eswitch is programmed by user space tools
*/
enabled_events = (mode == SRIOV_LEGACY) ? SRIOV_VPORT_EVENTS : 0;
- for (i = 0; i <= nvfs; i++)
- esw_enable_vport(esw, i, enabled_events);
+
+ /* Enable PF vport */
+ vport = mlx5_eswitch_get_vport(esw, MLX5_VPORT_PF);
+ esw_enable_vport(esw, vport, enabled_events);
+
+ /* Enable VF vports */
+ mlx5_esw_for_each_vf_vport(esw, i, vport, nvfs)
+ esw_enable_vport(esw, vport, enabled_events);
if (mode == SRIOV_LEGACY) {
MLX5_NB_INIT(&esw->nb, eswitch_vport_event, NIC_VPORT_CHANGE);
@@ -1683,6 +1712,7 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode)
void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw)
{
struct esw_mc_addr *mc_promisc;
+ struct mlx5_vport *vport;
int old_mode;
int i;
@@ -1697,8 +1727,8 @@ void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw)
if (esw->mode == SRIOV_LEGACY)
mlx5_eq_notifier_unregister(esw->dev, &esw->nb);
- for (i = 0; i < esw->total_vports; i++)
- esw_disable_vport(esw, i);
+ mlx5_esw_for_all_vports(esw, i, vport)
+ esw_disable_vport(esw, vport);
if (mc_promisc && mc_promisc->uplink_rule)
mlx5_del_flow_rules(mc_promisc->uplink_rule);
@@ -1725,6 +1755,7 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
{
int total_vports = MLX5_TOTAL_VPORTS(dev);
struct mlx5_eswitch *esw;
+ struct mlx5_vport *vport;
int vport_num;
int err;
@@ -1757,6 +1788,8 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
goto abort;
}
+ esw->total_vports = total_vports;
+
err = esw_offloads_init_reps(esw);
if (err)
goto abort;
@@ -1765,9 +1798,7 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
hash_init(esw->offloads.mod_hdr_tbl);
mutex_init(&esw->state_lock);
- for (vport_num = 0; vport_num < total_vports; vport_num++) {
- struct mlx5_vport *vport = &esw->vports[vport_num];
-
+ mlx5_esw_for_all_vports(esw, vport_num, vport) {
vport->vport = vport_num;
vport->info.link_state = MLX5_VPORT_ADMIN_STATE_AUTO;
vport->dev = dev;
@@ -1775,7 +1806,6 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
esw_vport_change_handler);
}
- esw->total_vports = total_vports;
esw->enabled_vports = 0;
esw->mode = SRIOV_NONE;
esw->offloads.inline_mode = MLX5_INLINE_MODE_NONE;
@@ -2017,8 +2047,7 @@ static u32 calculate_vports_min_rate_divider(struct mlx5_eswitch *esw)
u32 max_guarantee = 0;
int i;
- for (i = 0; i < esw->total_vports; i++) {
- evport = &esw->vports[i];
+ mlx5_esw_for_all_vports(esw, i, evport) {
if (!evport->enabled || evport->info.min_rate < max_guarantee)
continue;
max_guarantee = evport->info.min_rate;
@@ -2037,8 +2066,7 @@ static int normalize_vports_min_rate(struct mlx5_eswitch *esw, u32 divider)
int err;
int i;
- for (i = 0; i < esw->total_vports; i++) {
- evport = &esw->vports[i];
+ mlx5_esw_for_all_vports(esw, i, evport) {
if (!evport->enabled)
continue;
vport_min_rate = evport->info.min_rate;
@@ -2053,7 +2081,7 @@ static int normalize_vports_min_rate(struct mlx5_eswitch *esw, u32 divider)
if (bw_share == evport->qos.bw_share)
continue;
- err = esw_vport_qos_config(esw, i, vport_max_rate,
+ err = esw_vport_qos_config(esw, evport->vport, vport_max_rate,
bw_share);
if (!err)
evport->qos.bw_share = bw_share;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 14f7ad67cfe4..4979c7ee0ad7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -56,6 +56,44 @@ enum {
#define UPLINK_REP_INDEX 0
+/* The rep getter/iterator are only valid after esw->total_vports
+ * and vport->vport are initialized in mlx5_eswitch_init.
+ */
+#define mlx5_esw_for_all_reps(esw, i, rep) \
+ for ((i) = MLX5_VPORT_PF; \
+ (rep) = &(esw)->offloads.vport_reps[i], \
+ (i) < (esw)->total_vports; (i)++)
+
+#define mlx5_esw_for_each_vf_rep(esw, i, rep, nvfs) \
+ for ((i) = MLX5_VPORT_FIRST_VF; \
+ (rep) = &(esw)->offloads.vport_reps[i], \
+ (i) <= (nvfs); (i)++)
+
+#define mlx5_esw_for_each_vf_rep_reverse(esw, i, rep, nvfs) \
+ for ((i) = (nvfs); \
+ (rep) = &(esw)->offloads.vport_reps[i], \
+ (i) >= MLX5_VPORT_FIRST_VF; (i)--)
+
+#define mlx5_esw_for_each_vf_vport(esw, vport, nvfs) \
+ for ((vport) = MLX5_VPORT_FIRST_VF; \
+ (vport) <= (nvfs); (vport)++)
+
+#define mlx5_esw_for_each_vf_vport_reverse(esw, vport, nvfs) \
+ for ((vport) = (nvfs); \
+ (vport) >= MLX5_VPORT_FIRST_VF; (vport)--)
+
+static struct mlx5_eswitch_rep *mlx5_eswitch_get_rep(struct mlx5_eswitch *esw,
+ u16 vport_num)
+{
+ u16 idx = vport_num;
+
+ if (vport_num == MLX5_VPORT_UPLINK)
+ idx = UPLINK_REP_INDEX;
+
+ WARN_ON(idx > esw->total_vports - 1);
+ return &esw->offloads.vport_reps[idx];
+}
+
static struct mlx5_flow_table *
esw_get_prio_table(struct mlx5_eswitch *esw, u32 chain, u16 prio, int level);
static void
@@ -604,7 +642,7 @@ static int esw_add_fdb_peer_miss_rules(struct mlx5_eswitch *esw,
misc = MLX5_ADDR_OF(fte_match_param, spec->match_value,
misc_parameters);
- for (i = 1; i < nvports; i++) {
+ mlx5_esw_for_each_vf_vport(esw, i, mlx5_core_max_vfs(esw->dev)) {
MLX5_SET(fte_match_set_misc, misc, source_port, i);
flow = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb,
spec, &flow_act, &dest, 1);
@@ -622,7 +660,8 @@ static int esw_add_fdb_peer_miss_rules(struct mlx5_eswitch *esw,
return 0;
add_flow_err:
- for (i--; i > 0; i--)
+ nvports = --i;
+ mlx5_esw_for_each_vf_vport_reverse(esw, i, nvports)
mlx5_del_flow_rules(flows[i]);
kvfree(flows);
alloc_flows_err:
@@ -637,7 +676,7 @@ static void esw_del_fdb_peer_miss_rules(struct mlx5_eswitch *esw)
flows = esw->fdb_table.offloads.peer_miss_rules;
- for (i = 1; i < esw->total_vports; i++)
+ mlx5_esw_for_each_vf_vport_reverse(esw, i, mlx5_core_max_vfs(esw->dev))
mlx5_del_flow_rules(flows[i]);
kvfree(flows);
@@ -1229,9 +1268,7 @@ int esw_offloads_init_reps(struct mlx5_eswitch *esw)
offloads = &esw->offloads;
mlx5_query_nic_vport_mac_address(dev, 0, hw_id);
- for (vport = 0; vport < total_vfs; vport++) {
- rep = &offloads->vport_reps[vport];
-
+ mlx5_esw_for_all_reps(esw, vport, rep) {
rep->vport = vport;
ether_addr_copy(rep->hw_id, hw_id);
}
@@ -1256,12 +1293,10 @@ static void esw_offloads_unload_reps_type(struct mlx5_eswitch *esw, int nvports,
struct mlx5_eswitch_rep *rep;
int vport;
- for (vport = nvports; vport >= MLX5_VPORT_FIRST_VF; vport--) {
- rep = &esw->offloads.vport_reps[vport];
+ mlx5_esw_for_each_vf_rep_reverse(esw, vport, rep, nvports)
__esw_offloads_unload_rep(esw, rep, rep_type);
- }
- rep = &esw->offloads.vport_reps[UPLINK_REP_INDEX];
+ rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_UPLINK);
__esw_offloads_unload_rep(esw, rep, rep_type);
}
@@ -1289,13 +1324,12 @@ static int esw_offloads_load_reps_type(struct mlx5_eswitch *esw, int nvports,
int vport;
int err;
- rep = &esw->offloads.vport_reps[UPLINK_REP_INDEX];
+ rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_UPLINK);
err = __esw_offloads_load_rep(esw, rep, rep_type);
if (err)
goto out;
- for (vport = MLX5_VPORT_FIRST_VF; vport <= nvports; vport++) {
- rep = &esw->offloads.vport_reps[vport];
+ mlx5_esw_for_each_vf_rep(esw, vport, rep, nvports) {
err = __esw_offloads_load_rep(esw, rep, rep_type);
if (err)
goto err_reps;
@@ -1304,7 +1338,7 @@ static int esw_offloads_load_reps_type(struct mlx5_eswitch *esw, int nvports,
return 0;
err_reps:
- esw_offloads_unload_reps_type(esw, vport, rep_type);
+ esw_offloads_unload_reps_type(esw, --vport, rep_type);
out:
return err;
}
@@ -1848,10 +1882,9 @@ EXPORT_SYMBOL(mlx5_eswitch_unregister_vport_rep);
void *mlx5_eswitch_get_uplink_priv(struct mlx5_eswitch *esw, u8 rep_type)
{
- struct mlx5_esw_offload *offloads = &esw->offloads;
struct mlx5_eswitch_rep *rep;
- rep = &offloads->vport_reps[UPLINK_REP_INDEX];
+ rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_UPLINK);
return rep->rep_if[rep_type].priv;
}
@@ -1859,13 +1892,9 @@ void *mlx5_eswitch_get_proto_dev(struct mlx5_eswitch *esw,
int vport,
u8 rep_type)
{
- struct mlx5_esw_offload *offloads = &esw->offloads;
struct mlx5_eswitch_rep *rep;
- if (vport == MLX5_VPORT_UPLINK)
- vport = UPLINK_REP_INDEX;
-
- rep = &offloads->vport_reps[vport];
+ rep = mlx5_eswitch_get_rep(esw, vport);
if (rep->rep_if[rep_type].valid &&
rep->rep_if[rep_type].get_proto_dev)
@@ -1876,13 +1905,13 @@ EXPORT_SYMBOL(mlx5_eswitch_get_proto_dev);
void *mlx5_eswitch_uplink_get_proto_dev(struct mlx5_eswitch *esw, u8 rep_type)
{
- return mlx5_eswitch_get_proto_dev(esw, UPLINK_REP_INDEX, rep_type);
+ return mlx5_eswitch_get_proto_dev(esw, MLX5_VPORT_UPLINK, rep_type);
}
EXPORT_SYMBOL(mlx5_eswitch_uplink_get_proto_dev);
struct mlx5_eswitch_rep *mlx5_eswitch_vport_rep(struct mlx5_eswitch *esw,
int vport)
{
- return &esw->offloads.vport_reps[vport];
+ return mlx5_eswitch_get_rep(esw, vport);
}
EXPORT_SYMBOL(mlx5_eswitch_vport_rep);
--
2.20.1
^ permalink raw reply related
* [net-next 09/13] net/mlx5: E-Switch, Centralize repersentor reg/unreg to eswitch driver
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Bodong Wang, Or Gerlitz, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
Eswitch has two users: IB and ETH. They both register repersentors
when mlx5 interface is added, and unregister the repersentors when
mlx5 interface is removed. Ideally, each driver should only deal with
the entities which are unique to itself. However, current IB and ETH
drivers have to perform the following eswitch operations:
1. When registering, specify how many vports to register. This number
is the same for both drivers which is the total available vport
numbers.
2. When unregistering, specify the number of registered vports to do
unregister. Also, unload the repersentors which are already loaded.
It's unnecessary for eswitch driver to hands out the control of above
operations to individual driver users, as they're not unique to each
driver. Instead, such operations should be centralized to eswitch
driver. This consolidates eswitch control flow, and simplified IB and
ETH driver.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/infiniband/hw/mlx5/ib_rep.c | 20 +++------
.../net/ethernet/mellanox/mlx5/core/en_rep.c | 19 +++-----
.../mellanox/mlx5/core/eswitch_offloads.c | 45 +++++++++----------
include/linux/mlx5/eswitch.h | 11 ++---
4 files changed, 39 insertions(+), 56 deletions(-)
diff --git a/drivers/infiniband/hw/mlx5/ib_rep.c b/drivers/infiniband/hw/mlx5/ib_rep.c
index 99cae9a10195..4700cffb5a00 100644
--- a/drivers/infiniband/hw/mlx5/ib_rep.c
+++ b/drivers/infiniband/hw/mlx5/ib_rep.c
@@ -95,26 +95,20 @@ static void *mlx5_ib_vport_get_proto_dev(struct mlx5_eswitch_rep *rep)
void mlx5_ib_register_vport_reps(struct mlx5_core_dev *mdev)
{
struct mlx5_eswitch *esw = mdev->priv.eswitch;
- int total_vports = MLX5_TOTAL_VPORTS(mdev);
struct mlx5_eswitch_rep_if rep_if = {};
- int vport;
-
- for (vport = 0; vport < total_vports; vport++) {
- rep_if.load = mlx5_ib_vport_rep_load;
- rep_if.unload = mlx5_ib_vport_rep_unload;
- rep_if.get_proto_dev = mlx5_ib_vport_get_proto_dev;
- mlx5_eswitch_register_vport_rep(esw, vport, &rep_if, REP_IB);
- }
+
+ rep_if.load = mlx5_ib_vport_rep_load;
+ rep_if.unload = mlx5_ib_vport_rep_unload;
+ rep_if.get_proto_dev = mlx5_ib_vport_get_proto_dev;
+
+ mlx5_eswitch_register_vport_reps(esw, &rep_if, REP_IB);
}
void mlx5_ib_unregister_vport_reps(struct mlx5_core_dev *mdev)
{
struct mlx5_eswitch *esw = mdev->priv.eswitch;
- int total_vports = MLX5_TOTAL_VPORTS(mdev);
- int vport;
- for (vport = total_vports - 1; vport >= 0; vport--)
- mlx5_eswitch_unregister_vport_rep(esw, vport, REP_IB);
+ mlx5_eswitch_unregister_vport_reps(esw, REP_IB);
}
u8 mlx5_ib_eswitch_mode(struct mlx5_eswitch *esw)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index f84889bbe2a0..287d48e5b073 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -1798,25 +1798,18 @@ static void *mlx5e_vport_rep_get_proto_dev(struct mlx5_eswitch_rep *rep)
void mlx5e_rep_register_vport_reps(struct mlx5_core_dev *mdev)
{
struct mlx5_eswitch *esw = mdev->priv.eswitch;
- int total_vfs = MLX5_TOTAL_VPORTS(mdev);
- int vport;
+ struct mlx5_eswitch_rep_if rep_if = {};
- for (vport = 0; vport < total_vfs; vport++) {
- struct mlx5_eswitch_rep_if rep_if = {};
+ rep_if.load = mlx5e_vport_rep_load;
+ rep_if.unload = mlx5e_vport_rep_unload;
+ rep_if.get_proto_dev = mlx5e_vport_rep_get_proto_dev;
- rep_if.load = mlx5e_vport_rep_load;
- rep_if.unload = mlx5e_vport_rep_unload;
- rep_if.get_proto_dev = mlx5e_vport_rep_get_proto_dev;
- mlx5_eswitch_register_vport_rep(esw, vport, &rep_if, REP_ETH);
- }
+ mlx5_eswitch_register_vport_reps(esw, &rep_if, REP_ETH);
}
void mlx5e_rep_unregister_vport_reps(struct mlx5_core_dev *mdev)
{
struct mlx5_eswitch *esw = mdev->priv.eswitch;
- int total_vfs = MLX5_TOTAL_VPORTS(mdev);
- int vport;
- for (vport = total_vfs - 1; vport >= 0; vport--)
- mlx5_eswitch_unregister_vport_rep(esw, vport, REP_ETH);
+ mlx5_eswitch_unregister_vport_reps(esw, REP_ETH);
}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 7131d41796fb..b702b56c457e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -1923,40 +1923,39 @@ int mlx5_devlink_eswitch_encap_mode_get(struct devlink *devlink, u8 *encap)
return 0;
}
-void mlx5_eswitch_register_vport_rep(struct mlx5_eswitch *esw,
- int vport_index,
- struct mlx5_eswitch_rep_if *__rep_if,
- u8 rep_type)
+void mlx5_eswitch_register_vport_reps(struct mlx5_eswitch *esw,
+ struct mlx5_eswitch_rep_if *__rep_if,
+ u8 rep_type)
{
- struct mlx5_esw_offload *offloads = &esw->offloads;
struct mlx5_eswitch_rep_if *rep_if;
+ struct mlx5_eswitch_rep *rep;
+ int i;
- rep_if = &offloads->vport_reps[vport_index].rep_if[rep_type];
-
- rep_if->load = __rep_if->load;
- rep_if->unload = __rep_if->unload;
- rep_if->get_proto_dev = __rep_if->get_proto_dev;
- rep_if->priv = __rep_if->priv;
+ mlx5_esw_for_all_reps(esw, i, rep) {
+ rep_if = &rep->rep_if[rep_type];
+ rep_if->load = __rep_if->load;
+ rep_if->unload = __rep_if->unload;
+ rep_if->get_proto_dev = __rep_if->get_proto_dev;
+ rep_if->priv = __rep_if->priv;
- rep_if->state = REP_REGISTERED;
+ rep_if->state = REP_REGISTERED;
+ }
}
-EXPORT_SYMBOL(mlx5_eswitch_register_vport_rep);
+EXPORT_SYMBOL(mlx5_eswitch_register_vport_reps);
-void mlx5_eswitch_unregister_vport_rep(struct mlx5_eswitch *esw,
- int vport_index, u8 rep_type)
+void mlx5_eswitch_unregister_vport_reps(struct mlx5_eswitch *esw, u8 rep_type)
{
- struct mlx5_esw_offload *offloads = &esw->offloads;
+ u16 max_vf = mlx5_core_max_vfs(esw->dev);
struct mlx5_eswitch_rep *rep;
+ int i;
- rep = &offloads->vport_reps[vport_index];
-
- if (esw->mode == SRIOV_OFFLOADS &&
- rep->rep_if[rep_type].state == REP_LOADED)
- rep->rep_if[rep_type].unload(rep);
+ if (esw->mode == SRIOV_OFFLOADS)
+ __unload_reps_all_vport(esw, max_vf, rep_type);
- rep->rep_if[rep_type].state = REP_UNREGISTERED;
+ mlx5_esw_for_all_reps(esw, i, rep)
+ rep->rep_if[rep_type].state = REP_UNREGISTERED;
}
-EXPORT_SYMBOL(mlx5_eswitch_unregister_vport_rep);
+EXPORT_SYMBOL(mlx5_eswitch_unregister_vport_reps);
void *mlx5_eswitch_get_uplink_priv(struct mlx5_eswitch *esw, u8 rep_type)
{
diff --git a/include/linux/mlx5/eswitch.h b/include/linux/mlx5/eswitch.h
index e3dbc1bc0917..96d8435421de 100644
--- a/include/linux/mlx5/eswitch.h
+++ b/include/linux/mlx5/eswitch.h
@@ -46,13 +46,10 @@ struct mlx5_eswitch_rep {
u32 vlan_refcount;
};
-void mlx5_eswitch_register_vport_rep(struct mlx5_eswitch *esw,
- int vport_index,
- struct mlx5_eswitch_rep_if *rep_if,
- u8 rep_type);
-void mlx5_eswitch_unregister_vport_rep(struct mlx5_eswitch *esw,
- int vport_index,
- u8 rep_type);
+void mlx5_eswitch_register_vport_reps(struct mlx5_eswitch *esw,
+ struct mlx5_eswitch_rep_if *rep_if,
+ u8 rep_type);
+void mlx5_eswitch_unregister_vport_reps(struct mlx5_eswitch *esw, u8 rep_type);
void *mlx5_eswitch_get_proto_dev(struct mlx5_eswitch *esw,
int vport,
u8 rep_type);
--
2.20.1
^ permalink raw reply related
* [net-next 08/13] net/mlx5: E-Switch, Support load/unload reps of specific vport types
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Bodong Wang, Or Gerlitz, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
Currently the driver loads and unloads all reps in an unbreakable
group. However, with ECPF, the reps of special vports such as uplink
and host PF should always be loaded in switchdev mode where the reps
for VFs will be loaded on-demand and unloaded on no-demand. This is
a pre-step for that change.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../mellanox/mlx5/core/eswitch_offloads.c | 109 ++++++++++++++----
1 file changed, 87 insertions(+), 22 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index c6c9dad69ba8..7131d41796fb 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -1291,25 +1291,47 @@ static void __esw_offloads_unload_rep(struct mlx5_eswitch *esw,
rep->rep_if[rep_type].state = REP_REGISTERED;
}
-static void esw_offloads_unload_reps_type(struct mlx5_eswitch *esw, int nvports,
- u8 rep_type)
+static void __unload_reps_special_vport(struct mlx5_eswitch *esw, u8 rep_type)
{
struct mlx5_eswitch_rep *rep;
- int vport;
-
- mlx5_esw_for_each_vf_rep_reverse(esw, vport, rep, nvports)
- __esw_offloads_unload_rep(esw, rep, rep_type);
rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_UPLINK);
__esw_offloads_unload_rep(esw, rep, rep_type);
}
-static void esw_offloads_unload_reps(struct mlx5_eswitch *esw, int nvports)
+static void __unload_reps_vf_vport(struct mlx5_eswitch *esw, int nvports,
+ u8 rep_type)
+{
+ struct mlx5_eswitch_rep *rep;
+ int i;
+
+ mlx5_esw_for_each_vf_rep_reverse(esw, i, rep, nvports)
+ __esw_offloads_unload_rep(esw, rep, rep_type);
+}
+
+static void esw_offloads_unload_vf_reps(struct mlx5_eswitch *esw, int nvports)
+{
+ u8 rep_type = NUM_REP_TYPES;
+
+ while (rep_type-- > 0)
+ __unload_reps_vf_vport(esw, nvports, rep_type);
+}
+
+static void __unload_reps_all_vport(struct mlx5_eswitch *esw, int nvports,
+ u8 rep_type)
+{
+ __unload_reps_vf_vport(esw, nvports, rep_type);
+
+ /* Special vports must be the last to unload. */
+ __unload_reps_special_vport(esw, rep_type);
+}
+
+static void esw_offloads_unload_all_reps(struct mlx5_eswitch *esw, int nvports)
{
u8 rep_type = NUM_REP_TYPES;
while (rep_type-- > 0)
- esw_offloads_unload_reps_type(esw, nvports, rep_type);
+ __unload_reps_all_vport(esw, nvports, rep_type);
}
static int __esw_offloads_load_rep(struct mlx5_eswitch *esw,
@@ -1329,39 +1351,82 @@ static int __esw_offloads_load_rep(struct mlx5_eswitch *esw,
return 0;
}
-static int esw_offloads_load_reps_type(struct mlx5_eswitch *esw, int nvports,
- u8 rep_type)
+static int __load_reps_special_vport(struct mlx5_eswitch *esw, u8 rep_type)
{
struct mlx5_eswitch_rep *rep;
- int vport;
int err;
rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_UPLINK);
err = __esw_offloads_load_rep(esw, rep, rep_type);
- if (err)
- goto out;
+ return err;
+}
- mlx5_esw_for_each_vf_rep(esw, vport, rep, nvports) {
+static int __load_reps_vf_vport(struct mlx5_eswitch *esw, int nvports,
+ u8 rep_type)
+{
+ struct mlx5_eswitch_rep *rep;
+ int err, i;
+
+ mlx5_esw_for_each_vf_rep(esw, i, rep, nvports) {
err = __esw_offloads_load_rep(esw, rep, rep_type);
if (err)
- goto err_reps;
+ goto err_vf;
}
return 0;
+err_vf:
+ __unload_reps_vf_vport(esw, --i, rep_type);
+ return err;
+}
+
+static int esw_offloads_load_vf_reps(struct mlx5_eswitch *esw, int nvports)
+{
+ u8 rep_type = 0;
+ int err;
+
+ for (rep_type = 0; rep_type < NUM_REP_TYPES; rep_type++) {
+ err = __load_reps_vf_vport(esw, nvports, rep_type);
+ if (err)
+ goto err_reps;
+ }
+
+ return err;
+
err_reps:
- esw_offloads_unload_reps_type(esw, --vport, rep_type);
-out:
+ while (rep_type-- > 0)
+ __unload_reps_vf_vport(esw, nvports, rep_type);
+ return err;
+}
+
+static int __load_reps_all_vport(struct mlx5_eswitch *esw, int nvports,
+ u8 rep_type)
+{
+ int err;
+
+ /* Special vports must be loaded first. */
+ err = __load_reps_special_vport(esw, rep_type);
+ if (err)
+ return err;
+
+ err = __load_reps_vf_vport(esw, nvports, rep_type);
+ if (err)
+ goto err_vfs;
+
+ return 0;
+
+err_vfs:
+ __unload_reps_special_vport(esw, rep_type);
return err;
}
-static int esw_offloads_load_reps(struct mlx5_eswitch *esw, int nvports)
+static int esw_offloads_load_all_reps(struct mlx5_eswitch *esw, int nvports)
{
u8 rep_type = 0;
int err;
for (rep_type = 0; rep_type < NUM_REP_TYPES; rep_type++) {
- err = esw_offloads_load_reps_type(esw, nvports, rep_type);
+ err = __load_reps_all_vport(esw, nvports, rep_type);
if (err)
goto err_reps;
}
@@ -1370,7 +1435,7 @@ static int esw_offloads_load_reps(struct mlx5_eswitch *esw, int nvports)
err_reps:
while (rep_type-- > 0)
- esw_offloads_unload_reps_type(esw, nvports, rep_type);
+ __unload_reps_all_vport(esw, nvports, rep_type);
return err;
}
@@ -1520,7 +1585,7 @@ int esw_offloads_init(struct mlx5_eswitch *esw, int vf_nvports,
if (err)
return err;
- err = esw_offloads_load_reps(esw, vf_nvports);
+ err = esw_offloads_load_all_reps(esw, vf_nvports);
if (err)
goto err_reps;
@@ -1556,7 +1621,7 @@ void esw_offloads_cleanup(struct mlx5_eswitch *esw)
u16 num_vfs = esw->dev->priv.sriov.num_vfs;
esw_offloads_devcom_cleanup(esw);
- esw_offloads_unload_reps(esw, num_vfs);
+ esw_offloads_unload_all_reps(esw, num_vfs);
esw_offloads_steering_cleanup(esw);
}
--
2.20.1
^ permalink raw reply related
* [net-next 10/13] net/mlx5: E-Switch, Assign a different position for uplink rep and vport
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Bodong Wang, Eli Cohen, Or Gerlitz, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
In offloads mode, the current implementation puts the uplink
representor at index zero of the vport reps array. It is not "natural"
to place it at index 0 since we want to put the representor for vport
0 at index 0 with the introduction of SmartNIC. A separate patch will
handle the case whether a rep is needed for vport 0 (PF vport).
So, we want to have a different placeholder for uplink vport and
representor. It was placed at the end of vport and rep array. Since
vport number can no longer act as an index into the vport or
representors arrays, use functions to map vport numbers to indices
when accessing the vports or representors arrays, and vice versa.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/eswitch.c | 11 +++++----
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 24 +++++++++++++++++++
.../mellanox/mlx5/core/eswitch_offloads.c | 11 ++-------
include/linux/mlx5/vport.h | 11 ++++++---
4 files changed, 40 insertions(+), 17 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index d1454f18c0a7..bb7f72467df9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -84,8 +84,10 @@ enum {
static struct mlx5_vport *mlx5_eswitch_get_vport(struct mlx5_eswitch *esw,
u16 vport_num)
{
+ u16 idx = mlx5_eswitch_vport_num_to_index(esw, vport_num);
+
WARN_ON(vport_num > esw->total_vports - 1);
- return &esw->vports[vport_num];
+ return &esw->vports[idx];
}
static int arm_vport_context_events_cmd(struct mlx5_core_dev *dev, u16 vport,
@@ -1756,8 +1758,7 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
int total_vports = MLX5_TOTAL_VPORTS(dev);
struct mlx5_eswitch *esw;
struct mlx5_vport *vport;
- int vport_num;
- int err;
+ int err, i;
if (!MLX5_VPORT_MANAGER(dev))
return 0;
@@ -1798,8 +1799,8 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
hash_init(esw->offloads.mod_hdr_tbl);
mutex_init(&esw->state_lock);
- mlx5_esw_for_all_vports(esw, vport_num, vport) {
- vport->vport = vport_num;
+ mlx5_esw_for_all_vports(esw, i, vport) {
+ vport->vport = mlx5_eswitch_index_to_vport_num(esw, i);
vport->info.link_state = MLX5_VPORT_ADMIN_STATE_AUTO;
vport->dev = dev;
INIT_WORK(&vport->vport_change_handler,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index fd845e6c44d5..2951c1296c3e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -374,6 +374,30 @@ static inline u16 mlx5_eswitch_manager_vport(struct mlx5_core_dev *dev)
MLX5_VPORT_ECPF : MLX5_VPORT_PF;
}
+static inline int mlx5_eswitch_uplink_idx(struct mlx5_eswitch *esw)
+{
+ /* Uplink always locate at the last element of the array.*/
+ return esw->total_vports - 1;
+}
+
+static inline int mlx5_eswitch_vport_num_to_index(struct mlx5_eswitch *esw,
+ u16 vport_num)
+{
+ if (vport_num == MLX5_VPORT_UPLINK)
+ return mlx5_eswitch_uplink_idx(esw);
+
+ return vport_num;
+}
+
+static inline int mlx5_eswitch_index_to_vport_num(struct mlx5_eswitch *esw,
+ int index)
+{
+ if (index == mlx5_eswitch_uplink_idx(esw))
+ return MLX5_VPORT_UPLINK;
+
+ return index;
+}
+
#else /* CONFIG_MLX5_ESWITCH */
/* eswitch API stubs */
static inline int mlx5_eswitch_init(struct mlx5_core_dev *dev) { return 0; }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index b702b56c457e..e787e9212174 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -85,10 +85,7 @@ enum {
static struct mlx5_eswitch_rep *mlx5_eswitch_get_rep(struct mlx5_eswitch *esw,
u16 vport_num)
{
- u16 idx = vport_num;
-
- if (vport_num == MLX5_VPORT_UPLINK)
- idx = UPLINK_REP_INDEX;
+ u16 idx = mlx5_eswitch_vport_num_to_index(esw, vport_num);
WARN_ON(idx > esw->total_vports - 1);
return &esw->offloads.vport_reps[idx];
@@ -1254,7 +1251,6 @@ int esw_offloads_init_reps(struct mlx5_eswitch *esw)
{
int total_vfs = MLX5_TOTAL_VPORTS(esw->dev);
struct mlx5_core_dev *dev = esw->dev;
- struct mlx5_esw_offload *offloads;
struct mlx5_eswitch_rep *rep;
u8 hw_id[ETH_ALEN], rep_type;
int vport;
@@ -1265,19 +1261,16 @@ int esw_offloads_init_reps(struct mlx5_eswitch *esw)
if (!esw->offloads.vport_reps)
return -ENOMEM;
- offloads = &esw->offloads;
mlx5_query_nic_vport_mac_address(dev, 0, hw_id);
mlx5_esw_for_all_reps(esw, vport, rep) {
- rep->vport = vport;
+ rep->vport = mlx5_eswitch_index_to_vport_num(esw, vport);
ether_addr_copy(rep->hw_id, hw_id);
for (rep_type = 0; rep_type < NUM_REP_TYPES; rep_type++)
rep->rep_if[rep_type].state = REP_UNREGISTERED;
}
- offloads->vport_reps[0].vport = MLX5_VPORT_UPLINK;
-
return 0;
}
diff --git a/include/linux/mlx5/vport.h b/include/linux/mlx5/vport.h
index 755aeea19e1c..134248c02786 100644
--- a/include/linux/mlx5/vport.h
+++ b/include/linux/mlx5/vport.h
@@ -36,9 +36,14 @@
#include <linux/mlx5/driver.h>
#include <linux/mlx5/device.h>
-#define MLX5_VPORT_PF_PLACEHOLDER (1u)
-#define MLX5_SPECIAL_VPORTS (MLX5_VPORT_PF_PLACEHOLDER)
-#define MLX5_TOTAL_VPORTS(mdev) (MLX5_SPECIAL_VPORTS + mlx5_core_max_vfs(mdev))
+#define MLX5_VPORT_PF_PLACEHOLDER (1u)
+#define MLX5_VPORT_UPLINK_PLACEHOLDER (1u)
+
+#define MLX5_SPECIAL_VPORTS (MLX5_VPORT_PF_PLACEHOLDER + \
+ MLX5_VPORT_UPLINK_PLACEHOLDER)
+
+#define MLX5_TOTAL_VPORTS(mdev) (MLX5_SPECIAL_VPORTS + \
+ mlx5_core_max_vfs(mdev))
#define MLX5_VPORT_MANAGER(mdev) \
(MLX5_CAP_GEN(mdev, vport_group_manager) && \
--
2.20.1
^ permalink raw reply related
* [net-next 11/13] net/mlx5: E-Switch, Consider ECPF vport depends on eswitch ownership
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Bodong Wang, Or Gerlitz, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
ECPF connects to the eswitch through vport 0xfffe. ECPF may or may
not be the eswitch manager depending on firmware configuration.
1. If ECPF is eswitch manager: ECPF will take over the eswitch manager
responsibility. A rep of the host PF shall be created at the ECPF
side for the eswitch manager to control.
2. If ECPF is not eswitch manager: host PF will be the eswitch manager,
ECPF acts similar as a VF to the host PF. Host PF will be aware
of the ECPF vport presence and control it's rep.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/eswitch.c | 8 +-
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 15 ++++
.../mellanox/mlx5/core/eswitch_offloads.c | 79 ++++++++++++++++++-
include/linux/mlx5/driver.h | 5 ++
include/linux/mlx5/mlx5_ifc.h | 3 +-
include/linux/mlx5/vport.h | 8 +-
6 files changed, 110 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index bb7f72467df9..d2ab1ee19b2a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1667,7 +1667,7 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode)
mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_ETH);
mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
err = esw_offloads_init(esw, nvfs,
- nvfs + MLX5_SPECIAL_VPORTS);
+ nvfs + MLX5_SPECIAL_VPORTS(esw->dev));
}
if (err)
@@ -1687,6 +1687,12 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode)
vport = mlx5_eswitch_get_vport(esw, MLX5_VPORT_PF);
esw_enable_vport(esw, vport, enabled_events);
+ /* Enable ECPF vports */
+ if (mlx5_ecpf_vport_exists(esw->dev)) {
+ vport = mlx5_eswitch_get_vport(esw, MLX5_VPORT_ECPF);
+ esw_enable_vport(esw, vport, enabled_events);
+ }
+
/* Enable VF vports */
mlx5_esw_for_each_vf_vport(esw, i, vport, nvfs)
esw_enable_vport(esw, vport, enabled_events);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index 2951c1296c3e..2baa0d71380c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -380,9 +380,20 @@ static inline int mlx5_eswitch_uplink_idx(struct mlx5_eswitch *esw)
return esw->total_vports - 1;
}
+static inline int mlx5_eswitch_ecpf_idx(struct mlx5_eswitch *esw)
+{
+ return esw->total_vports - 2;
+}
+
static inline int mlx5_eswitch_vport_num_to_index(struct mlx5_eswitch *esw,
u16 vport_num)
{
+ if (vport_num == MLX5_VPORT_ECPF) {
+ if (!mlx5_ecpf_vport_exists(esw->dev))
+ esw_warn(esw->dev, "ECPF vport doesn't exist!\n");
+ return mlx5_eswitch_ecpf_idx(esw);
+ }
+
if (vport_num == MLX5_VPORT_UPLINK)
return mlx5_eswitch_uplink_idx(esw);
@@ -392,6 +403,10 @@ static inline int mlx5_eswitch_vport_num_to_index(struct mlx5_eswitch *esw,
static inline int mlx5_eswitch_index_to_vport_num(struct mlx5_eswitch *esw,
int index)
{
+ if (index == mlx5_eswitch_ecpf_idx(esw) &&
+ mlx5_ecpf_vport_exists(esw->dev))
+ return MLX5_VPORT_ECPF;
+
if (index == mlx5_eswitch_uplink_idx(esw))
return MLX5_VPORT_UPLINK;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index e787e9212174..84a33f8e3350 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -639,14 +639,35 @@ static int esw_add_fdb_peer_miss_rules(struct mlx5_eswitch *esw,
misc = MLX5_ADDR_OF(fte_match_param, spec->match_value,
misc_parameters);
+ if (mlx5_core_is_ecpf_esw_manager(esw->dev)) {
+ MLX5_SET(fte_match_set_misc, misc, source_port, MLX5_VPORT_PF);
+ flow = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb,
+ spec, &flow_act, &dest, 1);
+ if (IS_ERR(flow)) {
+ err = PTR_ERR(flow);
+ goto add_pf_flow_err;
+ }
+ flows[MLX5_VPORT_PF] = flow;
+ }
+
+ if (mlx5_ecpf_vport_exists(esw->dev)) {
+ MLX5_SET(fte_match_set_misc, misc, source_port, MLX5_VPORT_ECPF);
+ flow = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb,
+ spec, &flow_act, &dest, 1);
+ if (IS_ERR(flow)) {
+ err = PTR_ERR(flow);
+ goto add_ecpf_flow_err;
+ }
+ flows[mlx5_eswitch_ecpf_idx(esw)] = flow;
+ }
+
mlx5_esw_for_each_vf_vport(esw, i, mlx5_core_max_vfs(esw->dev)) {
MLX5_SET(fte_match_set_misc, misc, source_port, i);
flow = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb,
spec, &flow_act, &dest, 1);
if (IS_ERR(flow)) {
err = PTR_ERR(flow);
- esw_warn(esw->dev, "FDB: Failed to add peer miss flow rule err %d\n", err);
- goto add_flow_err;
+ goto add_vf_flow_err;
}
flows[i] = flow;
}
@@ -656,10 +677,18 @@ static int esw_add_fdb_peer_miss_rules(struct mlx5_eswitch *esw,
kvfree(spec);
return 0;
-add_flow_err:
+add_vf_flow_err:
nvports = --i;
mlx5_esw_for_each_vf_vport_reverse(esw, i, nvports)
mlx5_del_flow_rules(flows[i]);
+
+ if (mlx5_ecpf_vport_exists(esw->dev))
+ mlx5_del_flow_rules(flows[mlx5_eswitch_ecpf_idx(esw)]);
+add_ecpf_flow_err:
+ if (mlx5_core_is_ecpf_esw_manager(esw->dev))
+ mlx5_del_flow_rules(flows[MLX5_VPORT_PF]);
+add_pf_flow_err:
+ esw_warn(esw->dev, "FDB: Failed to add peer miss flow rule err %d\n", err);
kvfree(flows);
alloc_flows_err:
kvfree(spec);
@@ -676,6 +705,12 @@ static void esw_del_fdb_peer_miss_rules(struct mlx5_eswitch *esw)
mlx5_esw_for_each_vf_vport_reverse(esw, i, mlx5_core_max_vfs(esw->dev))
mlx5_del_flow_rules(flows[i]);
+ if (mlx5_ecpf_vport_exists(esw->dev))
+ mlx5_del_flow_rules(flows[mlx5_eswitch_ecpf_idx(esw)]);
+
+ if (mlx5_core_is_ecpf_esw_manager(esw->dev))
+ mlx5_del_flow_rules(flows[MLX5_VPORT_PF]);
+
kvfree(flows);
}
@@ -1288,6 +1323,16 @@ static void __unload_reps_special_vport(struct mlx5_eswitch *esw, u8 rep_type)
{
struct mlx5_eswitch_rep *rep;
+ if (mlx5_ecpf_vport_exists(esw->dev)) {
+ rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_ECPF);
+ __esw_offloads_unload_rep(esw, rep, rep_type);
+ }
+
+ if (mlx5_core_is_ecpf_esw_manager(esw->dev)) {
+ rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_PF);
+ __esw_offloads_unload_rep(esw, rep, rep_type);
+ }
+
rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_UPLINK);
__esw_offloads_unload_rep(esw, rep, rep_type);
}
@@ -1351,6 +1396,34 @@ static int __load_reps_special_vport(struct mlx5_eswitch *esw, u8 rep_type)
rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_UPLINK);
err = __esw_offloads_load_rep(esw, rep, rep_type);
+ if (err)
+ return err;
+
+ if (mlx5_core_is_ecpf_esw_manager(esw->dev)) {
+ rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_PF);
+ err = __esw_offloads_load_rep(esw, rep, rep_type);
+ if (err)
+ goto err_pf;
+ }
+
+ if (mlx5_ecpf_vport_exists(esw->dev)) {
+ rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_ECPF);
+ err = __esw_offloads_load_rep(esw, rep, rep_type);
+ if (err)
+ goto err_ecpf;
+ }
+
+ return 0;
+
+err_ecpf:
+ if (mlx5_core_is_ecpf_esw_manager(esw->dev)) {
+ rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_PF);
+ __esw_offloads_unload_rep(esw, rep, rep_type);
+ }
+
+err_pf:
+ rep = mlx5_eswitch_get_rep(esw, MLX5_VPORT_UPLINK);
+ __esw_offloads_unload_rep(esw, rep, rep_type);
return err;
}
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index c5454f985e1d..c2de50f02b33 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -1088,6 +1088,11 @@ static inline bool mlx5_core_is_ecpf_esw_manager(struct mlx5_core_dev *dev)
return dev->caps.embedded_cpu && MLX5_CAP_GEN(dev, eswitch_manager);
}
+static inline bool mlx5_ecpf_vport_exists(struct mlx5_core_dev *dev)
+{
+ return mlx5_core_is_pf(dev) && MLX5_CAP_ESW(dev, ecpf_vport_exists);
+}
+
#define MLX5_HOST_PF_MAX_VFS (127u)
static inline u16 mlx5_core_max_vfs(struct mlx5_core_dev *dev)
{
diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h
index 5decffe565fb..b7bb774b57b0 100644
--- a/include/linux/mlx5/mlx5_ifc.h
+++ b/include/linux/mlx5/mlx5_ifc.h
@@ -631,7 +631,8 @@ struct mlx5_ifc_e_switch_cap_bits {
u8 vport_svlan_insert[0x1];
u8 vport_cvlan_insert_if_not_exist[0x1];
u8 vport_cvlan_insert_overwrite[0x1];
- u8 reserved_at_5[0x17];
+ u8 reserved_at_5[0x16];
+ u8 ecpf_vport_exists[0x1];
u8 counter_eswitch_affinity[0x1];
u8 merged_eswitch[0x1];
u8 nic_vport_node_guid_modify[0x1];
diff --git a/include/linux/mlx5/vport.h b/include/linux/mlx5/vport.h
index 134248c02786..0eef548b9946 100644
--- a/include/linux/mlx5/vport.h
+++ b/include/linux/mlx5/vport.h
@@ -38,11 +38,13 @@
#define MLX5_VPORT_PF_PLACEHOLDER (1u)
#define MLX5_VPORT_UPLINK_PLACEHOLDER (1u)
+#define MLX5_VPORT_ECPF_PLACEHOLDER(mdev) (mlx5_ecpf_vport_exists(mdev))
-#define MLX5_SPECIAL_VPORTS (MLX5_VPORT_PF_PLACEHOLDER + \
- MLX5_VPORT_UPLINK_PLACEHOLDER)
+#define MLX5_SPECIAL_VPORTS(mdev) (MLX5_VPORT_PF_PLACEHOLDER + \
+ MLX5_VPORT_UPLINK_PLACEHOLDER + \
+ MLX5_VPORT_ECPF_PLACEHOLDER(mdev))
-#define MLX5_TOTAL_VPORTS(mdev) (MLX5_SPECIAL_VPORTS + \
+#define MLX5_TOTAL_VPORTS(mdev) (MLX5_SPECIAL_VPORTS(mdev) + \
mlx5_core_max_vfs(mdev))
#define MLX5_VPORT_MANAGER(mdev) \
--
2.20.1
^ permalink raw reply related
* [net-next 13/13] net/mlx5: E-Switch, Allow transition to offloads mode for ECPF
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Bodong Wang, Or Gerlitz, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
Currently, the e-switch driver requires going to legacy mode before
changing to the offloads mode. This makes sense for regular case as
the legacy mode is done by creating VFs.
However, it's problematic when ECPF is the eswitch manager. In such
case, ECPF will control the vports on peer host including the peer
PF and VFs. But ECPF doesn't need and shall not create VFs as the
VFs are created in the peer PF host.
Grant ECPF the ability to change from none to the offloads mode. Note
that currently the only way to go back to none mode is by unloading
the ECPF driver.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 91c4095ac79e..f2260391be5b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -1250,7 +1250,8 @@ static int esw_offloads_start(struct mlx5_eswitch *esw,
{
int err, err1, num_vfs = esw->dev->priv.sriov.num_vfs;
- if (esw->mode != SRIOV_LEGACY) {
+ if (esw->mode != SRIOV_LEGACY &&
+ !mlx5_core_is_ecpf_esw_manager(esw->dev)) {
NL_SET_ERR_MSG_MOD(extack,
"Can't set offloads mode, SRIOV legacy not enabled");
return -EINVAL;
@@ -1846,7 +1847,8 @@ static int mlx5_devlink_eswitch_check(struct devlink *devlink)
if(!MLX5_ESWITCH_MANAGER(dev))
return -EPERM;
- if (dev->priv.eswitch->mode == SRIOV_NONE)
+ if (dev->priv.eswitch->mode == SRIOV_NONE &&
+ !mlx5_core_is_ecpf_esw_manager(dev))
return -EOPNOTSUPP;
return 0;
--
2.20.1
^ permalink raw reply related
* [net-next 12/13] net/mlx5: E-Switch, Load/unload VF reps according to event from host PF
From: Saeed Mahameed @ 2019-02-16 1:34 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Bodong Wang, Or Gerlitz, Saeed Mahameed
In-Reply-To: <20190216013452.21131-1-saeedm@mellanox.com>
From: Bodong Wang <bodong@mellanox.com>
When host PF changes the number of VFs, the ECPF esw driver will get
a FW event. It should query the number of VFs enabled by host PF and
update the VF reps accordingly. Note that host PF can't change the
number of VFs dynamically, it has to reset the number of VFs to 0
before changing to a new positive number.
The host event is registered when driver is moving to switchdev mode,
and it's the last step to do in esw_offloads_init. It's unregistered
and the work queue is flushed when driver quits from switchdev mode.
In this way, the host event and devlink command are serialized.
When driver is enabling switchdev mode, pay attention to the following
two facts:
1. Host PF must not have VF initialized as the flow table in ECPF has
ENCAP enabled as default. Such flow table can't be created with
existing initialized VFs.
2. ECPF doesn't know how many VFs the host PF will enable, ECPF
offloads flow steering shall create the flow table/groups based on
the max number of VFs possibly supported by host PF.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/eswitch.c | 17 ++++-
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 11 +++
.../mellanox/mlx5/core/eswitch_offloads.c | 71 ++++++++++++++++++-
3 files changed, 96 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index d2ab1ee19b2a..e18af31336e6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -39,6 +39,7 @@
#include "lib/eq.h"
#include "eswitch.h"
#include "fs_core.h"
+#include "ecpf.h"
enum {
MLX5_ACTION_NONE = 0,
@@ -1639,6 +1640,7 @@ static int eswitch_vport_event(struct notifier_block *nb,
int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode)
{
+ int vf_nvports = 0, total_nvports = 0;
struct mlx5_vport *vport;
int err;
int i, enabled_events;
@@ -1657,6 +1659,18 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode)
esw_info(esw->dev, "E-Switch enable SRIOV: nvfs(%d) mode (%d)\n", nvfs, mode);
+ if (mode == SRIOV_OFFLOADS) {
+ if (mlx5_core_is_ecpf_esw_manager(esw->dev)) {
+ err = mlx5_query_host_params_num_vfs(esw->dev, &vf_nvports);
+ if (err)
+ return err;
+ total_nvports = esw->total_vports;
+ } else {
+ vf_nvports = nvfs;
+ total_nvports = nvfs + MLX5_SPECIAL_VPORTS(esw->dev);
+ }
+ }
+
esw->mode = mode;
mlx5_lag_update(esw->dev);
@@ -1666,8 +1680,7 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode)
} else {
mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_ETH);
mlx5_reload_interface(esw->dev, MLX5_INTERFACE_PROTOCOL_IB);
- err = esw_offloads_init(esw, nvfs,
- nvfs + MLX5_SPECIAL_VPORTS(esw->dev));
+ err = esw_offloads_init(esw, vf_nvports, total_nvports);
}
if (err)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index 2baa0d71380c..af5581a57e56 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -182,6 +182,16 @@ struct esw_mc_addr { /* SRIOV only */
u32 refcnt;
};
+struct mlx5_host_work {
+ struct work_struct work;
+ struct mlx5_eswitch *esw;
+};
+
+struct mlx5_host_info {
+ struct mlx5_nb nb;
+ u16 num_vfs;
+};
+
struct mlx5_eswitch {
struct mlx5_core_dev *dev;
struct mlx5_nb nb;
@@ -206,6 +216,7 @@ struct mlx5_eswitch {
int mode;
int nvports;
u16 manager_vport;
+ struct mlx5_host_info host_info;
};
void esw_offloads_cleanup(struct mlx5_eswitch *esw);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 84a33f8e3350..91c4095ac79e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -40,6 +40,8 @@
#include "en.h"
#include "fs_core.h"
#include "lib/devcom.h"
+#include "ecpf.h"
+#include "lib/eq.h"
enum {
FDB_FAST_PATH = 0,
@@ -1640,6 +1642,57 @@ static void esw_offloads_steering_cleanup(struct mlx5_eswitch *esw)
esw_destroy_offloads_fdb_tables(esw);
}
+static void esw_host_params_event_handler(struct work_struct *work)
+{
+ struct mlx5_host_work *host_work;
+ struct mlx5_eswitch *esw;
+ int err, num_vf = 0;
+
+ host_work = container_of(work, struct mlx5_host_work, work);
+ esw = host_work->esw;
+
+ err = mlx5_query_host_params_num_vfs(esw->dev, &num_vf);
+ if (err || num_vf == esw->host_info.num_vfs)
+ goto out;
+
+ /* Number of VFs can only change from "0 to x" or "x to 0". */
+ if (esw->host_info.num_vfs > 0) {
+ esw_offloads_unload_vf_reps(esw, esw->host_info.num_vfs);
+ } else {
+ err = esw_offloads_load_vf_reps(esw, num_vf);
+
+ if (err)
+ goto out;
+ }
+
+ esw->host_info.num_vfs = num_vf;
+
+out:
+ kfree(host_work);
+}
+
+static int esw_host_params_event(struct notifier_block *nb,
+ unsigned long type, void *data)
+{
+ struct mlx5_host_work *host_work;
+ struct mlx5_host_info *host_info;
+ struct mlx5_eswitch *esw;
+
+ host_work = kzalloc(sizeof(*host_work), GFP_ATOMIC);
+ if (!host_work)
+ return NOTIFY_DONE;
+
+ host_info = mlx5_nb_cof(nb, struct mlx5_host_info, nb);
+ esw = container_of(host_info, struct mlx5_eswitch, host_info);
+
+ host_work->esw = esw;
+
+ INIT_WORK(&host_work->work, esw_host_params_event_handler);
+ queue_work(esw->work_queue, &host_work->work);
+
+ return NOTIFY_OK;
+}
+
int esw_offloads_init(struct mlx5_eswitch *esw, int vf_nvports,
int total_nvports)
{
@@ -1656,6 +1709,14 @@ int esw_offloads_init(struct mlx5_eswitch *esw, int vf_nvports,
goto err_reps;
esw_offloads_devcom_init(esw);
+
+ if (mlx5_core_is_ecpf_esw_manager(esw->dev)) {
+ MLX5_NB_INIT(&esw->host_info.nb, esw_host_params_event,
+ HOST_PARAMS_CHANGE);
+ mlx5_eq_notifier_register(esw->dev, &esw->host_info.nb);
+ esw->host_info.num_vfs = vf_nvports;
+ }
+
return 0;
err_reps:
@@ -1684,7 +1745,15 @@ static int esw_offloads_stop(struct mlx5_eswitch *esw,
void esw_offloads_cleanup(struct mlx5_eswitch *esw)
{
- u16 num_vfs = esw->dev->priv.sriov.num_vfs;
+ u16 num_vfs;
+
+ if (mlx5_core_is_ecpf_esw_manager(esw->dev)) {
+ mlx5_eq_notifier_unregister(esw->dev, &esw->host_info.nb);
+ flush_workqueue(esw->work_queue);
+ num_vfs = esw->host_info.num_vfs;
+ } else {
+ num_vfs = esw->dev->priv.sriov.num_vfs;
+ }
esw_offloads_devcom_cleanup(esw);
esw_offloads_unload_all_reps(esw, num_vfs);
--
2.20.1
^ permalink raw reply related
* [PATCH net-next] igc: Make function igc_write_rss_indir_tbl() static
From: Wei Yongjun @ 2019-02-16 2:04 UTC (permalink / raw)
To: Jeff Kirsher, Sasha Neftin
Cc: Wei Yongjun, intel-wired-lan, netdev, kernel-janitors
Fixes the following sparse warning:
drivers/net/ethernet/intel/igc/igc_ethtool.c:646:6: warning:
symbol 'igc_write_rss_indir_tbl' was not declared. Should it be static?
Fixes: 8c5ad0dae93c ("igc: Add ethtool support")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
---
drivers/net/ethernet/intel/igc/igc_ethtool.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/igc/igc_ethtool.c b/drivers/net/ethernet/intel/igc/igc_ethtool.c
index eff37a6c0afa..544239422577 100644
--- a/drivers/net/ethernet/intel/igc/igc_ethtool.c
+++ b/drivers/net/ethernet/intel/igc/igc_ethtool.c
@@ -643,7 +643,7 @@ static int igc_set_coalesce(struct net_device *netdev,
return 0;
}
-void igc_write_rss_indir_tbl(struct igc_adapter *adapter)
+static void igc_write_rss_indir_tbl(struct igc_adapter *adapter)
{
struct igc_hw *hw = &adapter->hw;
u32 reg = IGC_RETA(0);
^ 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